[
  {
    "epoch": 0,
    "batch_idx": 0,
    "loss": 0.9700422286987305,
    "learning_rate": 0.001,
    "timestep_mean": 599.03125,
    "epsilon_norm": 7.851838111877441,
    "prediction_norm": 0.24037578701972961,
    "timestamp": "2025-07-14T12:33:32.502891"
  },
  {
    "epoch": 0,
    "batch_idx": 32,
    "loss": 1.132689118385315,
    "learning_rate": 0.001,
    "timestep_mean": 453.09375,
    "epsilon_norm": 8.5994291305542,
    "prediction_norm": 0.22191643714904785,
    "timestamp": "2025-07-14T12:33:32.509522"
  },
  {
    "epoch": 0,
    "batch_idx": 64,
    "loss": 0.7530454397201538,
    "learning_rate": 0.001,
    "timestep_mean": 524.53125,
    "epsilon_norm": 7.095963478088379,
    "prediction_norm": 0.2986404001712799,
    "timestamp": "2025-07-14T12:33:32.515065"
  },
  {
    "epoch": 0,
    "batch_idx": 96,
    "loss": 0.7888463139533997,
    "learning_rate": 0.001,
    "timestep_mean": 485.5625,
    "epsilon_norm": 7.298041343688965,
    "prediction_norm": 0.49841606616973877,
    "timestamp": "2025-07-14T12:33:32.520460"
  },
  {
    "epoch": 0,
    "batch_idx": 128,
    "loss": 0.7321728467941284,
    "learning_rate": 0.001,
    "timestep_mean": 545.03125,
    "epsilon_norm": 7.2007951736450195,
    "prediction_norm": 0.7209557294845581,
    "timestamp": "2025-07-14T12:33:32.525744"
  },
  {
    "epoch": 0,
    "batch_idx": 160,
    "loss": 1.0484752655029297,
    "learning_rate": 0.001,
    "timestep_mean": 394.78125,
    "epsilon_norm": 8.549195289611816,
    "prediction_norm": 1.0322675704956055,
    "timestamp": "2025-07-14T12:33:32.530825"
  },
  {
    "epoch": 0,
    "batch_idx": 192,
    "loss": 0.4899333417415619,
    "learning_rate": 0.001,
    "timestep_mean": 549.375,
    "epsilon_norm": 2.98447847366333,
    "prediction_norm": 0.5180799961090088,
    "timestamp": "2025-07-14T12:33:33.382074"
  },
  {
    "epoch": 1,
    "batch_idx": 0,
    "loss": 1.0887856483459473,
    "learning_rate": 0.0009999993831498518,
    "timestep_mean": 490.3125,
    "epsilon_norm": 9.159858703613281,
    "prediction_norm": 1.5546916723251343,
    "timestamp": "2025-07-14T12:33:33.389510"
  },
  {
    "epoch": 1,
    "batch_idx": 32,
    "loss": 0.889093279838562,
    "learning_rate": 0.0009999993831498518,
    "timestep_mean": 499.84375,
    "epsilon_norm": 8.33272933959961,
    "prediction_norm": 1.6735107898712158,
    "timestamp": "2025-07-14T12:33:33.394604"
  },
  {
    "epoch": 1,
    "batch_idx": 64,
    "loss": 0.772878885269165,
    "learning_rate": 0.0009999993831498518,
    "timestep_mean": 489.53125,
    "epsilon_norm": 8.31364917755127,
    "prediction_norm": 2.264113426208496,
    "timestamp": "2025-07-14T12:33:33.399753"
  },
  {
    "epoch": 1,
    "batch_idx": 96,
    "loss": 0.70796799659729,
    "learning_rate": 0.0009999993831498518,
    "timestep_mean": 601.03125,
    "epsilon_norm": 8.096470832824707,
    "prediction_norm": 2.2811477184295654,
    "timestamp": "2025-07-14T12:33:33.404737"
  },
  {
    "epoch": 1,
    "batch_idx": 128,
    "loss": 0.5903416872024536,
    "learning_rate": 0.0009999993831498518,
    "timestep_mean": 549.15625,
    "epsilon_norm": 8.194428443908691,
    "prediction_norm": 3.2459022998809814,
    "timestamp": "2025-07-14T12:33:33.409735"
  },
  {
    "epoch": 1,
    "batch_idx": 160,
    "loss": 0.7101048827171326,
    "learning_rate": 0.0009999993831498518,
    "timestep_mean": 574.53125,
    "epsilon_norm": 8.83180046081543,
    "prediction_norm": 3.5004124641418457,
    "timestamp": "2025-07-14T12:33:33.414671"
  },
  {
    "epoch": 1,
    "batch_idx": 192,
    "loss": 0.5899169445037842,
    "learning_rate": 0.0009999993831498518,
    "timestep_mean": 436.375,
    "epsilon_norm": 4.88240909576416,
    "prediction_norm": 2.3542661666870117,
    "timestamp": "2025-07-14T12:33:33.420123"
  },
  {
    "epoch": 2,
    "batch_idx": 0,
    "loss": 0.4935750961303711,
    "learning_rate": 0.0009999975326009292,
    "timestep_mean": 560.4375,
    "epsilon_norm": 7.7659173011779785,
    "prediction_norm": 4.736020565032959,
    "timestamp": "2025-07-14T12:33:33.427185"
  },
  {
    "epoch": 2,
    "batch_idx": 32,
    "loss": 0.9713810682296753,
    "learning_rate": 0.0009999975326009292,
    "timestep_mean": 496.34375,
    "epsilon_norm": 7.991763114929199,
    "prediction_norm": 5.267053127288818,
    "timestamp": "2025-07-14T12:33:33.433538"
  },
  {
    "epoch": 2,
    "batch_idx": 64,
    "loss": 0.6177010536193848,
    "learning_rate": 0.0009999975326009292,
    "timestep_mean": 544.6875,
    "epsilon_norm": 8.499715805053711,
    "prediction_norm": 6.524951934814453,
    "timestamp": "2025-07-14T12:33:33.439243"
  },
  {
    "epoch": 2,
    "batch_idx": 96,
    "loss": 0.6836239099502563,
    "learning_rate": 0.0009999975326009292,
    "timestep_mean": 398.875,
    "epsilon_norm": 7.241173267364502,
    "prediction_norm": 6.523881912231445,
    "timestamp": "2025-07-14T12:33:33.444775"
  },
  {
    "epoch": 2,
    "batch_idx": 128,
    "loss": 0.4216773509979248,
    "learning_rate": 0.0009999975326009292,
    "timestep_mean": 514.5,
    "epsilon_norm": 7.7234368324279785,
    "prediction_norm": 7.041996479034424,
    "timestamp": "2025-07-14T12:33:33.451224"
  },
  {
    "epoch": 2,
    "batch_idx": 160,
    "loss": 0.7116035223007202,
    "learning_rate": 0.0009999975326009292,
    "timestep_mean": 456.90625,
    "epsilon_norm": 7.611542701721191,
    "prediction_norm": 6.7809672355651855,
    "timestamp": "2025-07-14T12:33:33.456309"
  },
  {
    "epoch": 2,
    "batch_idx": 192,
    "loss": 0.9126548767089844,
    "learning_rate": 0.0009999975326009292,
    "timestep_mean": 317.125,
    "epsilon_norm": 2.92578387260437,
    "prediction_norm": 3.187087059020996,
    "timestamp": "2025-07-14T12:33:33.461343"
  },
  {
    "epoch": 3,
    "batch_idx": 0,
    "loss": 0.8832083940505981,
    "learning_rate": 0.000999994448357798,
    "timestep_mean": 417.53125,
    "epsilon_norm": 8.111289978027344,
    "prediction_norm": 7.201618194580078,
    "timestamp": "2025-07-14T12:33:33.466457"
  },
  {
    "epoch": 3,
    "batch_idx": 32,
    "loss": 0.6728758811950684,
    "learning_rate": 0.000999994448357798,
    "timestep_mean": 494.0625,
    "epsilon_norm": 8.042093276977539,
    "prediction_norm": 6.529180526733398,
    "timestamp": "2025-07-14T12:33:33.471575"
  },
  {
    "epoch": 3,
    "batch_idx": 64,
    "loss": 0.8163208365440369,
    "learning_rate": 0.000999994448357798,
    "timestep_mean": 466.125,
    "epsilon_norm": 7.591386318206787,
    "prediction_norm": 5.886318683624268,
    "timestamp": "2025-07-14T12:33:33.476574"
  },
  {
    "epoch": 3,
    "batch_idx": 96,
    "loss": 0.7256250381469727,
    "learning_rate": 0.000999994448357798,
    "timestep_mean": 443.875,
    "epsilon_norm": 8.167080879211426,
    "prediction_norm": 5.46676778793335,
    "timestamp": "2025-07-14T12:33:33.481663"
  },
  {
    "epoch": 3,
    "batch_idx": 128,
    "loss": 0.5433101654052734,
    "learning_rate": 0.000999994448357798,
    "timestep_mean": 508.59375,
    "epsilon_norm": 7.116990566253662,
    "prediction_norm": 4.710265636444092,
    "timestamp": "2025-07-14T12:33:33.486600"
  },
  {
    "epoch": 3,
    "batch_idx": 160,
    "loss": 0.45692896842956543,
    "learning_rate": 0.000999994448357798,
    "timestep_mean": 510.875,
    "epsilon_norm": 7.781551361083984,
    "prediction_norm": 4.87439489364624,
    "timestamp": "2025-07-14T12:33:33.491617"
  },
  {
    "epoch": 3,
    "batch_idx": 192,
    "loss": 0.297704815864563,
    "learning_rate": 0.000999994448357798,
    "timestep_mean": 402.625,
    "epsilon_norm": 3.0506885051727295,
    "prediction_norm": 2.203547954559326,
    "timestamp": "2025-07-14T12:33:33.496596"
  },
  {
    "epoch": 4,
    "batch_idx": 0,
    "loss": 0.6957889795303345,
    "learning_rate": 0.0009999901304280686,
    "timestep_mean": 471.21875,
    "epsilon_norm": 7.447842121124268,
    "prediction_norm": 3.962705135345459,
    "timestamp": "2025-07-14T12:33:33.501634"
  },
  {
    "epoch": 4,
    "batch_idx": 32,
    "loss": 0.7237871885299683,
    "learning_rate": 0.0009999901304280686,
    "timestep_mean": 564.25,
    "epsilon_norm": 7.761163234710693,
    "prediction_norm": 3.897693157196045,
    "timestamp": "2025-07-14T12:33:33.506540"
  },
  {
    "epoch": 4,
    "batch_idx": 64,
    "loss": 0.6766878366470337,
    "learning_rate": 0.0009999901304280686,
    "timestep_mean": 520.625,
    "epsilon_norm": 7.896240711212158,
    "prediction_norm": 3.538952589035034,
    "timestamp": "2025-07-14T12:33:33.511458"
  },
  {
    "epoch": 4,
    "batch_idx": 96,
    "loss": 0.5710905194282532,
    "learning_rate": 0.0009999901304280686,
    "timestep_mean": 540.03125,
    "epsilon_norm": 8.416303634643555,
    "prediction_norm": 3.652374267578125,
    "timestamp": "2025-07-14T12:33:33.516460"
  },
  {
    "epoch": 4,
    "batch_idx": 128,
    "loss": 0.7442758679389954,
    "learning_rate": 0.0009999901304280686,
    "timestep_mean": 378.5,
    "epsilon_norm": 8.17918586730957,
    "prediction_norm": 4.208994388580322,
    "timestamp": "2025-07-14T12:33:33.521430"
  },
  {
    "epoch": 4,
    "batch_idx": 160,
    "loss": 0.9756968021392822,
    "learning_rate": 0.0009999901304280686,
    "timestep_mean": 395.90625,
    "epsilon_norm": 8.491357803344727,
    "prediction_norm": 4.303796768188477,
    "timestamp": "2025-07-14T12:33:33.526331"
  },
  {
    "epoch": 4,
    "batch_idx": 192,
    "loss": 0.28986114263534546,
    "learning_rate": 0.0009999901304280686,
    "timestep_mean": 487.5,
    "epsilon_norm": 3.3772616386413574,
    "prediction_norm": 2.487638235092163,
    "timestamp": "2025-07-14T12:33:33.531269"
  },
  {
    "epoch": 5,
    "batch_idx": 0,
    "loss": 0.5395950078964233,
    "learning_rate": 0.000999984578822395,
    "timestep_mean": 459.9375,
    "epsilon_norm": 7.586756706237793,
    "prediction_norm": 3.6229965686798096,
    "timestamp": "2025-07-14T12:33:33.536207"
  },
  {
    "epoch": 5,
    "batch_idx": 32,
    "loss": 0.6862319707870483,
    "learning_rate": 0.000999984578822395,
    "timestep_mean": 414.8125,
    "epsilon_norm": 7.661279678344727,
    "prediction_norm": 3.3490896224975586,
    "timestamp": "2025-07-14T12:33:33.541143"
  },
  {
    "epoch": 5,
    "batch_idx": 64,
    "loss": 0.7461056709289551,
    "learning_rate": 0.000999984578822395,
    "timestep_mean": 501.59375,
    "epsilon_norm": 8.308408737182617,
    "prediction_norm": 3.83489990234375,
    "timestamp": "2025-07-14T12:33:33.546136"
  },
  {
    "epoch": 5,
    "batch_idx": 96,
    "loss": 0.885836124420166,
    "learning_rate": 0.000999984578822395,
    "timestep_mean": 437.9375,
    "epsilon_norm": 9.1243257522583,
    "prediction_norm": 3.383096218109131,
    "timestamp": "2025-07-14T12:33:33.551244"
  },
  {
    "epoch": 5,
    "batch_idx": 128,
    "loss": 0.5103493928909302,
    "learning_rate": 0.000999984578822395,
    "timestep_mean": 556.625,
    "epsilon_norm": 6.469796180725098,
    "prediction_norm": 3.1589319705963135,
    "timestamp": "2025-07-14T12:33:33.556311"
  },
  {
    "epoch": 5,
    "batch_idx": 160,
    "loss": 0.7446567416191101,
    "learning_rate": 0.000999984578822395,
    "timestep_mean": 539.0,
    "epsilon_norm": 9.067743301391602,
    "prediction_norm": 3.818814277648926,
    "timestamp": "2025-07-14T12:33:33.561528"
  },
  {
    "epoch": 5,
    "batch_idx": 192,
    "loss": 0.727408766746521,
    "learning_rate": 0.000999984578822395,
    "timestep_mean": 448.375,
    "epsilon_norm": 3.4741899967193604,
    "prediction_norm": 2.0288937091827393,
    "timestamp": "2025-07-14T12:33:33.566960"
  },
  {
    "epoch": 6,
    "batch_idx": 0,
    "loss": 0.46000027656555176,
    "learning_rate": 0.000999977793554475,
    "timestep_mean": 512.15625,
    "epsilon_norm": 7.5226359367370605,
    "prediction_norm": 4.0169525146484375,
    "timestamp": "2025-07-14T12:33:33.572058"
  },
  {
    "epoch": 6,
    "batch_idx": 32,
    "loss": 0.5218579769134521,
    "learning_rate": 0.000999977793554475,
    "timestep_mean": 573.0,
    "epsilon_norm": 7.661363124847412,
    "prediction_norm": 3.7266175746917725,
    "timestamp": "2025-07-14T12:33:33.577254"
  },
  {
    "epoch": 6,
    "batch_idx": 64,
    "loss": 0.4392111599445343,
    "learning_rate": 0.000999977793554475,
    "timestep_mean": 597.6875,
    "epsilon_norm": 6.9777984619140625,
    "prediction_norm": 3.554457664489746,
    "timestamp": "2025-07-14T12:33:33.582208"
  },
  {
    "epoch": 6,
    "batch_idx": 96,
    "loss": 0.4422580599784851,
    "learning_rate": 0.000999977793554475,
    "timestep_mean": 518.5625,
    "epsilon_norm": 7.781024932861328,
    "prediction_norm": 4.1184000968933105,
    "timestamp": "2025-07-14T12:33:33.587415"
  },
  {
    "epoch": 6,
    "batch_idx": 128,
    "loss": 0.49404823780059814,
    "learning_rate": 0.000999977793554475,
    "timestep_mean": 526.34375,
    "epsilon_norm": 8.37983226776123,
    "prediction_norm": 4.06630802154541,
    "timestamp": "2025-07-14T12:33:33.592324"
  },
  {
    "epoch": 6,
    "batch_idx": 160,
    "loss": 0.5862434506416321,
    "learning_rate": 0.000999977793554475,
    "timestep_mean": 544.625,
    "epsilon_norm": 7.402741432189941,
    "prediction_norm": 4.1152472496032715,
    "timestamp": "2025-07-14T12:33:33.597149"
  },
  {
    "epoch": 6,
    "batch_idx": 192,
    "loss": 0.6368410587310791,
    "learning_rate": 0.000999977793554475,
    "timestep_mean": 595.375,
    "epsilon_norm": 3.909714698791504,
    "prediction_norm": 2.5229318141937256,
    "timestamp": "2025-07-14T12:33:33.602110"
  },
  {
    "epoch": 7,
    "batch_idx": 0,
    "loss": 0.8557472825050354,
    "learning_rate": 0.0009999697746410508,
    "timestep_mean": 503.4375,
    "epsilon_norm": 8.919309616088867,
    "prediction_norm": 4.827269077301025,
    "timestamp": "2025-07-14T12:33:33.608097"
  },
  {
    "epoch": 7,
    "batch_idx": 32,
    "loss": 0.7827671766281128,
    "learning_rate": 0.0009999697746410508,
    "timestep_mean": 493.15625,
    "epsilon_norm": 8.530517578125,
    "prediction_norm": 5.243468284606934,
    "timestamp": "2025-07-14T12:33:33.613140"
  },
  {
    "epoch": 7,
    "batch_idx": 64,
    "loss": 0.5424000024795532,
    "learning_rate": 0.0009999697746410508,
    "timestep_mean": 565.375,
    "epsilon_norm": 9.10316276550293,
    "prediction_norm": 5.545526027679443,
    "timestamp": "2025-07-14T12:33:33.618107"
  },
  {
    "epoch": 7,
    "batch_idx": 96,
    "loss": 0.6944930553436279,
    "learning_rate": 0.0009999697746410508,
    "timestep_mean": 497.8125,
    "epsilon_norm": 7.548821926116943,
    "prediction_norm": 4.744184494018555,
    "timestamp": "2025-07-14T12:33:33.623272"
  },
  {
    "epoch": 7,
    "batch_idx": 128,
    "loss": 0.5142151713371277,
    "learning_rate": 0.0009999697746410508,
    "timestep_mean": 483.9375,
    "epsilon_norm": 8.279521942138672,
    "prediction_norm": 5.9566240310668945,
    "timestamp": "2025-07-14T12:33:33.628341"
  },
  {
    "epoch": 7,
    "batch_idx": 160,
    "loss": 0.6727257370948792,
    "learning_rate": 0.0009999697746410508,
    "timestep_mean": 477.78125,
    "epsilon_norm": 7.879017353057861,
    "prediction_norm": 6.167893409729004,
    "timestamp": "2025-07-14T12:33:33.633456"
  },
  {
    "epoch": 7,
    "batch_idx": 192,
    "loss": 0.8981176614761353,
    "learning_rate": 0.0009999697746410508,
    "timestep_mean": 471.75,
    "epsilon_norm": 4.934704303741455,
    "prediction_norm": 3.616628885269165,
    "timestamp": "2025-07-14T12:33:33.638400"
  },
  {
    "epoch": 8,
    "batch_idx": 0,
    "loss": 0.982629120349884,
    "learning_rate": 0.0009999605221019082,
    "timestep_mean": 407.0625,
    "epsilon_norm": 8.237577438354492,
    "prediction_norm": 7.0051655769348145,
    "timestamp": "2025-07-14T12:33:33.643381"
  },
  {
    "epoch": 8,
    "batch_idx": 32,
    "loss": 0.6161530017852783,
    "learning_rate": 0.0009999605221019082,
    "timestep_mean": 447.84375,
    "epsilon_norm": 7.4444403648376465,
    "prediction_norm": 6.0372138023376465,
    "timestamp": "2025-07-14T12:33:33.648272"
  },
  {
    "epoch": 8,
    "batch_idx": 64,
    "loss": 0.6495618224143982,
    "learning_rate": 0.0009999605221019082,
    "timestep_mean": 488.875,
    "epsilon_norm": 7.765734672546387,
    "prediction_norm": 6.251465320587158,
    "timestamp": "2025-07-14T12:33:33.653232"
  },
  {
    "epoch": 8,
    "batch_idx": 96,
    "loss": 0.3848874568939209,
    "learning_rate": 0.0009999605221019082,
    "timestep_mean": 597.25,
    "epsilon_norm": 8.472570419311523,
    "prediction_norm": 5.528875350952148,
    "timestamp": "2025-07-14T12:33:33.658287"
  },
  {
    "epoch": 8,
    "batch_idx": 128,
    "loss": 0.5671136379241943,
    "learning_rate": 0.0009999605221019082,
    "timestep_mean": 465.625,
    "epsilon_norm": 8.276474952697754,
    "prediction_norm": 5.762028217315674,
    "timestamp": "2025-07-14T12:33:33.663230"
  },
  {
    "epoch": 8,
    "batch_idx": 160,
    "loss": 0.39150020480155945,
    "learning_rate": 0.0009999605221019082,
    "timestep_mean": 517.03125,
    "epsilon_norm": 7.933290958404541,
    "prediction_norm": 5.949329853057861,
    "timestamp": "2025-07-14T12:33:33.668096"
  },
  {
    "epoch": 8,
    "batch_idx": 192,
    "loss": 0.7710219621658325,
    "learning_rate": 0.0009999605221019082,
    "timestep_mean": 425.5,
    "epsilon_norm": 3.6561832427978516,
    "prediction_norm": 2.4652256965637207,
    "timestamp": "2025-07-14T12:33:33.672950"
  },
  {
    "epoch": 9,
    "batch_idx": 0,
    "loss": 0.425117552280426,
    "learning_rate": 0.0009999500359598769,
    "timestep_mean": 532.65625,
    "epsilon_norm": 7.541585922241211,
    "prediction_norm": 5.391725063323975,
    "timestamp": "2025-07-14T12:33:33.677962"
  },
  {
    "epoch": 9,
    "batch_idx": 32,
    "loss": 0.6783839464187622,
    "learning_rate": 0.0009999500359598769,
    "timestep_mean": 451.625,
    "epsilon_norm": 7.2379069328308105,
    "prediction_norm": 5.7680253982543945,
    "timestamp": "2025-07-14T12:33:33.682906"
  },
  {
    "epoch": 9,
    "batch_idx": 64,
    "loss": 0.6921409368515015,
    "learning_rate": 0.0009999500359598769,
    "timestep_mean": 532.125,
    "epsilon_norm": 8.753902435302734,
    "prediction_norm": 5.298496246337891,
    "timestamp": "2025-07-14T12:33:33.687882"
  },
  {
    "epoch": 9,
    "batch_idx": 96,
    "loss": 0.6213143467903137,
    "learning_rate": 0.0009999500359598769,
    "timestep_mean": 499.25,
    "epsilon_norm": 7.765399932861328,
    "prediction_norm": 4.313083648681641,
    "timestamp": "2025-07-14T12:33:33.692760"
  },
  {
    "epoch": 9,
    "batch_idx": 128,
    "loss": 0.8389173746109009,
    "learning_rate": 0.0009999500359598769,
    "timestep_mean": 458.5625,
    "epsilon_norm": 8.756967544555664,
    "prediction_norm": 4.824139595031738,
    "timestamp": "2025-07-14T12:33:33.697666"
  },
  {
    "epoch": 9,
    "batch_idx": 160,
    "loss": 0.5527212619781494,
    "learning_rate": 0.0009999500359598769,
    "timestep_mean": 564.09375,
    "epsilon_norm": 9.474010467529297,
    "prediction_norm": 5.508468151092529,
    "timestamp": "2025-07-14T12:33:33.702524"
  },
  {
    "epoch": 9,
    "batch_idx": 192,
    "loss": 0.29197466373443604,
    "learning_rate": 0.0009999500359598769,
    "timestep_mean": 526.125,
    "epsilon_norm": 3.9783565998077393,
    "prediction_norm": 2.2897424697875977,
    "timestamp": "2025-07-14T12:33:33.707476"
  },
  {
    "epoch": 10,
    "batch_idx": 0,
    "loss": 0.5834082365036011,
    "learning_rate": 0.0009999383162408303,
    "timestep_mean": 366.125,
    "epsilon_norm": 7.649212837219238,
    "prediction_norm": 4.499996662139893,
    "timestamp": "2025-07-14T12:33:33.712538"
  },
  {
    "epoch": 10,
    "batch_idx": 32,
    "loss": 0.641750693321228,
    "learning_rate": 0.0009999383162408303,
    "timestep_mean": 473.46875,
    "epsilon_norm": 7.088619709014893,
    "prediction_norm": 4.325552940368652,
    "timestamp": "2025-07-14T12:33:33.717471"
  },
  {
    "epoch": 10,
    "batch_idx": 64,
    "loss": 0.44879472255706787,
    "learning_rate": 0.0009999383162408303,
    "timestep_mean": 547.125,
    "epsilon_norm": 7.6620612144470215,
    "prediction_norm": 4.7789530754089355,
    "timestamp": "2025-07-14T12:33:33.722327"
  },
  {
    "epoch": 10,
    "batch_idx": 96,
    "loss": 0.8438753485679626,
    "learning_rate": 0.0009999383162408303,
    "timestep_mean": 385.1875,
    "epsilon_norm": 8.063220024108887,
    "prediction_norm": 5.011194705963135,
    "timestamp": "2025-07-14T12:33:33.727207"
  },
  {
    "epoch": 10,
    "batch_idx": 128,
    "loss": 0.653715968132019,
    "learning_rate": 0.0009999383162408303,
    "timestep_mean": 449.78125,
    "epsilon_norm": 8.624809265136719,
    "prediction_norm": 4.8665361404418945,
    "timestamp": "2025-07-14T12:33:33.732073"
  },
  {
    "epoch": 10,
    "batch_idx": 160,
    "loss": 0.5381975173950195,
    "learning_rate": 0.0009999383162408303,
    "timestep_mean": 502.5625,
    "epsilon_norm": 7.356329917907715,
    "prediction_norm": 4.127949237823486,
    "timestamp": "2025-07-14T12:33:33.736999"
  },
  {
    "epoch": 10,
    "batch_idx": 192,
    "loss": 0.42130929231643677,
    "learning_rate": 0.0009999383162408303,
    "timestep_mean": 593.75,
    "epsilon_norm": 3.8792502880096436,
    "prediction_norm": 2.3702685832977295,
    "timestamp": "2025-07-14T12:33:33.742023"
  },
  {
    "epoch": 11,
    "batch_idx": 0,
    "loss": 0.4512147605419159,
    "learning_rate": 0.000999925362973686,
    "timestep_mean": 527.125,
    "epsilon_norm": 7.644030570983887,
    "prediction_norm": 5.018707752227783,
    "timestamp": "2025-07-14T12:33:33.746976"
  },
  {
    "epoch": 11,
    "batch_idx": 32,
    "loss": 0.5587720274925232,
    "learning_rate": 0.000999925362973686,
    "timestep_mean": 549.53125,
    "epsilon_norm": 8.131930351257324,
    "prediction_norm": 4.598929405212402,
    "timestamp": "2025-07-14T12:33:33.751857"
  },
  {
    "epoch": 11,
    "batch_idx": 64,
    "loss": 0.4332159757614136,
    "learning_rate": 0.000999925362973686,
    "timestep_mean": 523.625,
    "epsilon_norm": 8.234941482543945,
    "prediction_norm": 5.29261589050293,
    "timestamp": "2025-07-14T12:33:33.756722"
  },
  {
    "epoch": 11,
    "batch_idx": 96,
    "loss": 0.4497251510620117,
    "learning_rate": 0.000999925362973686,
    "timestep_mean": 538.84375,
    "epsilon_norm": 8.190832138061523,
    "prediction_norm": 5.104594707489014,
    "timestamp": "2025-07-14T12:33:33.761591"
  },
  {
    "epoch": 11,
    "batch_idx": 128,
    "loss": 0.5729820728302002,
    "learning_rate": 0.000999925362973686,
    "timestep_mean": 468.8125,
    "epsilon_norm": 6.169023513793945,
    "prediction_norm": 4.358567237854004,
    "timestamp": "2025-07-14T12:33:33.766595"
  },
  {
    "epoch": 11,
    "batch_idx": 160,
    "loss": 0.5025966167449951,
    "learning_rate": 0.000999925362973686,
    "timestep_mean": 505.46875,
    "epsilon_norm": 7.2714972496032715,
    "prediction_norm": 5.566616535186768,
    "timestamp": "2025-07-14T12:33:33.771529"
  },
  {
    "epoch": 11,
    "batch_idx": 192,
    "loss": 0.10653799772262573,
    "learning_rate": 0.000999925362973686,
    "timestep_mean": 639.25,
    "epsilon_norm": 3.2331111431121826,
    "prediction_norm": 2.499969482421875,
    "timestamp": "2025-07-14T12:33:33.776418"
  },
  {
    "epoch": 12,
    "batch_idx": 0,
    "loss": 0.37976279854774475,
    "learning_rate": 0.0009999111761904045,
    "timestep_mean": 504.0,
    "epsilon_norm": 6.478513240814209,
    "prediction_norm": 5.269669532775879,
    "timestamp": "2025-07-14T12:33:33.781335"
  },
  {
    "epoch": 12,
    "batch_idx": 32,
    "loss": 0.24988111853599548,
    "learning_rate": 0.0009999111761904045,
    "timestep_mean": 595.1875,
    "epsilon_norm": 8.310455322265625,
    "prediction_norm": 6.481138229370117,
    "timestamp": "2025-07-14T12:33:33.786191"
  },
  {
    "epoch": 12,
    "batch_idx": 64,
    "loss": 0.4402084946632385,
    "learning_rate": 0.0009999111761904045,
    "timestep_mean": 466.71875,
    "epsilon_norm": 8.832971572875977,
    "prediction_norm": 6.213233470916748,
    "timestamp": "2025-07-14T12:33:33.791076"
  },
  {
    "epoch": 12,
    "batch_idx": 96,
    "loss": 0.46454739570617676,
    "learning_rate": 0.0009999111761904045,
    "timestep_mean": 481.0,
    "epsilon_norm": 7.489416599273682,
    "prediction_norm": 6.984618663787842,
    "timestamp": "2025-07-14T12:33:33.795992"
  },
  {
    "epoch": 12,
    "batch_idx": 128,
    "loss": 0.48872509598731995,
    "learning_rate": 0.0009999111761904045,
    "timestep_mean": 542.0,
    "epsilon_norm": 8.602438926696777,
    "prediction_norm": 7.084203243255615,
    "timestamp": "2025-07-14T12:33:33.800903"
  },
  {
    "epoch": 12,
    "batch_idx": 160,
    "loss": 0.6957836151123047,
    "learning_rate": 0.0009999111761904045,
    "timestep_mean": 487.21875,
    "epsilon_norm": 7.892971992492676,
    "prediction_norm": 6.242560863494873,
    "timestamp": "2025-07-14T12:33:33.805817"
  },
  {
    "epoch": 12,
    "batch_idx": 192,
    "loss": 0.4568595290184021,
    "learning_rate": 0.0009999111761904045,
    "timestep_mean": 404.0,
    "epsilon_norm": 3.822312116622925,
    "prediction_norm": 2.878160238265991,
    "timestamp": "2025-07-14T12:33:33.810779"
  },
  {
    "epoch": 13,
    "batch_idx": 0,
    "loss": 0.5659916996955872,
    "learning_rate": 0.0009998957559259906,
    "timestep_mean": 447.875,
    "epsilon_norm": 8.741583824157715,
    "prediction_norm": 7.142623424530029,
    "timestamp": "2025-07-14T12:33:33.815726"
  },
  {
    "epoch": 13,
    "batch_idx": 32,
    "loss": 0.40998244285583496,
    "learning_rate": 0.0009998957559259906,
    "timestep_mean": 524.375,
    "epsilon_norm": 7.101408958435059,
    "prediction_norm": 5.429689884185791,
    "timestamp": "2025-07-14T12:33:33.820652"
  },
  {
    "epoch": 13,
    "batch_idx": 64,
    "loss": 0.4382737874984741,
    "learning_rate": 0.0009998957559259906,
    "timestep_mean": 550.5,
    "epsilon_norm": 7.525432586669922,
    "prediction_norm": 6.170873165130615,
    "timestamp": "2025-07-14T12:33:33.825530"
  },
  {
    "epoch": 13,
    "batch_idx": 96,
    "loss": 0.4522160291671753,
    "learning_rate": 0.0009998957559259906,
    "timestep_mean": 389.125,
    "epsilon_norm": 7.164207458496094,
    "prediction_norm": 4.788224697113037,
    "timestamp": "2025-07-14T12:33:33.830474"
  },
  {
    "epoch": 13,
    "batch_idx": 128,
    "loss": 0.3430190682411194,
    "learning_rate": 0.0009998957559259906,
    "timestep_mean": 497.875,
    "epsilon_norm": 7.504207134246826,
    "prediction_norm": 5.273898124694824,
    "timestamp": "2025-07-14T12:33:33.835385"
  },
  {
    "epoch": 13,
    "batch_idx": 160,
    "loss": 0.37662792205810547,
    "learning_rate": 0.0009998957559259906,
    "timestep_mean": 546.15625,
    "epsilon_norm": 6.594894886016846,
    "prediction_norm": 3.8371617794036865,
    "timestamp": "2025-07-14T12:33:33.840327"
  },
  {
    "epoch": 13,
    "batch_idx": 192,
    "loss": 0.7771257162094116,
    "learning_rate": 0.0009998957559259906,
    "timestep_mean": 550.5,
    "epsilon_norm": 5.677261829376221,
    "prediction_norm": 3.1027071475982666,
    "timestamp": "2025-07-14T12:33:33.845161"
  },
  {
    "epoch": 14,
    "batch_idx": 0,
    "loss": 0.4913976788520813,
    "learning_rate": 0.000999879102218492,
    "timestep_mean": 582.1875,
    "epsilon_norm": 8.523456573486328,
    "prediction_norm": 5.711867809295654,
    "timestamp": "2025-07-14T12:33:33.850180"
  },
  {
    "epoch": 14,
    "batch_idx": 32,
    "loss": 0.3333273231983185,
    "learning_rate": 0.000999879102218492,
    "timestep_mean": 499.625,
    "epsilon_norm": 7.403754711151123,
    "prediction_norm": 4.410769462585449,
    "timestamp": "2025-07-14T12:33:33.855403"
  },
  {
    "epoch": 14,
    "batch_idx": 64,
    "loss": 0.37057071924209595,
    "learning_rate": 0.000999879102218492,
    "timestep_mean": 499.3125,
    "epsilon_norm": 7.5346856117248535,
    "prediction_norm": 5.082629203796387,
    "timestamp": "2025-07-14T12:33:33.860811"
  },
  {
    "epoch": 14,
    "batch_idx": 96,
    "loss": 0.26434293389320374,
    "learning_rate": 0.000999879102218492,
    "timestep_mean": 572.34375,
    "epsilon_norm": 6.563711166381836,
    "prediction_norm": 5.225111484527588,
    "timestamp": "2025-07-14T12:33:33.865976"
  },
  {
    "epoch": 14,
    "batch_idx": 128,
    "loss": 0.48977145552635193,
    "learning_rate": 0.000999879102218492,
    "timestep_mean": 544.625,
    "epsilon_norm": 7.851384162902832,
    "prediction_norm": 5.688259601593018,
    "timestamp": "2025-07-14T12:33:33.871347"
  },
  {
    "epoch": 14,
    "batch_idx": 160,
    "loss": 0.3852778673171997,
    "learning_rate": 0.000999879102218492,
    "timestep_mean": 464.34375,
    "epsilon_norm": 7.551600933074951,
    "prediction_norm": 6.434968948364258,
    "timestamp": "2025-07-14T12:33:33.876742"
  },
  {
    "epoch": 14,
    "batch_idx": 192,
    "loss": 0.49636760354042053,
    "learning_rate": 0.000999879102218492,
    "timestep_mean": 312.75,
    "epsilon_norm": 2.953035354614258,
    "prediction_norm": 1.318996548652649,
    "timestamp": "2025-07-14T12:33:33.881756"
  },
  {
    "epoch": 15,
    "batch_idx": 0,
    "loss": 0.2276337742805481,
    "learning_rate": 0.0009998612151090004,
    "timestep_mean": 546.0625,
    "epsilon_norm": 7.544327735900879,
    "prediction_norm": 6.022006511688232,
    "timestamp": "2025-07-14T12:33:33.886675"
  },
  {
    "epoch": 15,
    "batch_idx": 32,
    "loss": 0.4725828766822815,
    "learning_rate": 0.0009998612151090004,
    "timestep_mean": 456.65625,
    "epsilon_norm": 7.193017959594727,
    "prediction_norm": 6.435213565826416,
    "timestamp": "2025-07-14T12:33:33.891628"
  },
  {
    "epoch": 15,
    "batch_idx": 64,
    "loss": 0.4669017493724823,
    "learning_rate": 0.0009998612151090004,
    "timestep_mean": 495.8125,
    "epsilon_norm": 8.472579002380371,
    "prediction_norm": 6.805439472198486,
    "timestamp": "2025-07-14T12:33:33.896539"
  },
  {
    "epoch": 15,
    "batch_idx": 96,
    "loss": 0.23285718262195587,
    "learning_rate": 0.0009998612151090004,
    "timestep_mean": 583.28125,
    "epsilon_norm": 7.88778829574585,
    "prediction_norm": 6.765713691711426,
    "timestamp": "2025-07-14T12:33:33.901442"
  },
  {
    "epoch": 15,
    "batch_idx": 128,
    "loss": 0.6151019334793091,
    "learning_rate": 0.0009998612151090004,
    "timestep_mean": 509.25,
    "epsilon_norm": 8.8139009475708,
    "prediction_norm": 6.317246437072754,
    "timestamp": "2025-07-14T12:33:33.911023"
  },
  {
    "epoch": 15,
    "batch_idx": 160,
    "loss": 0.2381564825773239,
    "learning_rate": 0.0009998612151090004,
    "timestep_mean": 529.34375,
    "epsilon_norm": 7.011195182800293,
    "prediction_norm": 6.448070526123047,
    "timestamp": "2025-07-14T12:33:33.915904"
  },
  {
    "epoch": 15,
    "batch_idx": 192,
    "loss": 0.544330358505249,
    "learning_rate": 0.0009998612151090004,
    "timestep_mean": 551.625,
    "epsilon_norm": 5.0246686935424805,
    "prediction_norm": 3.0235986709594727,
    "timestamp": "2025-07-14T12:33:33.920783"
  },
  {
    "epoch": 16,
    "batch_idx": 0,
    "loss": 0.5151839256286621,
    "learning_rate": 0.0009998420946416501,
    "timestep_mean": 381.8125,
    "epsilon_norm": 7.436837196350098,
    "prediction_norm": 5.359647274017334,
    "timestamp": "2025-07-14T12:33:33.925671"
  },
  {
    "epoch": 16,
    "batch_idx": 32,
    "loss": 0.6511681079864502,
    "learning_rate": 0.0009998420946416501,
    "timestep_mean": 512.03125,
    "epsilon_norm": 8.959441184997559,
    "prediction_norm": 6.251518249511719,
    "timestamp": "2025-07-14T12:33:33.930578"
  },
  {
    "epoch": 16,
    "batch_idx": 64,
    "loss": 0.4326334595680237,
    "learning_rate": 0.0009998420946416501,
    "timestep_mean": 537.6875,
    "epsilon_norm": 7.499380588531494,
    "prediction_norm": 5.478096008300781,
    "timestamp": "2025-07-14T12:33:33.935468"
  },
  {
    "epoch": 16,
    "batch_idx": 96,
    "loss": 0.5283766388893127,
    "learning_rate": 0.0009998420946416501,
    "timestep_mean": 499.75,
    "epsilon_norm": 8.61760139465332,
    "prediction_norm": 5.406317710876465,
    "timestamp": "2025-07-14T12:33:33.940428"
  },
  {
    "epoch": 16,
    "batch_idx": 128,
    "loss": 0.6740720868110657,
    "learning_rate": 0.0009998420946416501,
    "timestep_mean": 427.53125,
    "epsilon_norm": 8.091166496276855,
    "prediction_norm": 5.2071146965026855,
    "timestamp": "2025-07-14T12:33:33.945339"
  },
  {
    "epoch": 16,
    "batch_idx": 160,
    "loss": 0.5747741460800171,
    "learning_rate": 0.0009998420946416501,
    "timestep_mean": 493.03125,
    "epsilon_norm": 8.735962867736816,
    "prediction_norm": 6.937133312225342,
    "timestamp": "2025-07-14T12:33:33.950501"
  },
  {
    "epoch": 16,
    "batch_idx": 192,
    "loss": 0.7373274564743042,
    "learning_rate": 0.0009998420946416501,
    "timestep_mean": 317.0,
    "epsilon_norm": 3.3624308109283447,
    "prediction_norm": 1.754718542098999,
    "timestamp": "2025-07-14T12:33:33.955355"
  },
  {
    "epoch": 17,
    "batch_idx": 0,
    "loss": 0.33964040875434875,
    "learning_rate": 0.000999821740863619,
    "timestep_mean": 516.125,
    "epsilon_norm": 8.12473201751709,
    "prediction_norm": 6.038122653961182,
    "timestamp": "2025-07-14T12:33:33.960260"
  },
  {
    "epoch": 17,
    "batch_idx": 32,
    "loss": 0.23274081945419312,
    "learning_rate": 0.000999821740863619,
    "timestep_mean": 586.09375,
    "epsilon_norm": 7.495913982391357,
    "prediction_norm": 5.462395191192627,
    "timestamp": "2025-07-14T12:33:33.965195"
  },
  {
    "epoch": 17,
    "batch_idx": 64,
    "loss": 0.5289161205291748,
    "learning_rate": 0.000999821740863619,
    "timestep_mean": 459.8125,
    "epsilon_norm": 9.566326141357422,
    "prediction_norm": 6.310145378112793,
    "timestamp": "2025-07-14T12:33:33.970161"
  },
  {
    "epoch": 17,
    "batch_idx": 96,
    "loss": 0.4814433455467224,
    "learning_rate": 0.000999821740863619,
    "timestep_mean": 517.59375,
    "epsilon_norm": 8.294769287109375,
    "prediction_norm": 6.187612056732178,
    "timestamp": "2025-07-14T12:33:33.975268"
  },
  {
    "epoch": 17,
    "batch_idx": 128,
    "loss": 0.32586127519607544,
    "learning_rate": 0.000999821740863619,
    "timestep_mean": 523.28125,
    "epsilon_norm": 7.246175765991211,
    "prediction_norm": 5.638160705566406,
    "timestamp": "2025-07-14T12:33:33.980272"
  },
  {
    "epoch": 17,
    "batch_idx": 160,
    "loss": 0.4124196767807007,
    "learning_rate": 0.000999821740863619,
    "timestep_mean": 486.09375,
    "epsilon_norm": 7.4677205085754395,
    "prediction_norm": 5.432328701019287,
    "timestamp": "2025-07-14T12:33:33.985394"
  },
  {
    "epoch": 17,
    "batch_idx": 192,
    "loss": 0.23983538150787354,
    "learning_rate": 0.000999821740863619,
    "timestep_mean": 613.875,
    "epsilon_norm": 3.347728729248047,
    "prediction_norm": 3.23370099067688,
    "timestamp": "2025-07-14T12:33:33.990477"
  },
  {
    "epoch": 18,
    "batch_idx": 0,
    "loss": 0.3252343535423279,
    "learning_rate": 0.0009998001538251283,
    "timestep_mean": 516.65625,
    "epsilon_norm": 8.107926368713379,
    "prediction_norm": 7.014497756958008,
    "timestamp": "2025-07-14T12:33:33.995652"
  },
  {
    "epoch": 18,
    "batch_idx": 32,
    "loss": 0.36902761459350586,
    "learning_rate": 0.0009998001538251283,
    "timestep_mean": 532.75,
    "epsilon_norm": 7.434083461761475,
    "prediction_norm": 5.316591739654541,
    "timestamp": "2025-07-14T12:33:34.000779"
  },
  {
    "epoch": 18,
    "batch_idx": 64,
    "loss": 0.46467962861061096,
    "learning_rate": 0.0009998001538251283,
    "timestep_mean": 500.1875,
    "epsilon_norm": 8.797338485717773,
    "prediction_norm": 7.130926609039307,
    "timestamp": "2025-07-14T12:33:34.005660"
  },
  {
    "epoch": 18,
    "batch_idx": 96,
    "loss": 0.3319936990737915,
    "learning_rate": 0.0009998001538251283,
    "timestep_mean": 520.625,
    "epsilon_norm": 7.6348114013671875,
    "prediction_norm": 6.429571151733398,
    "timestamp": "2025-07-14T12:33:34.010815"
  },
  {
    "epoch": 18,
    "batch_idx": 128,
    "loss": 0.3868323564529419,
    "learning_rate": 0.0009998001538251283,
    "timestep_mean": 462.6875,
    "epsilon_norm": 7.923191070556641,
    "prediction_norm": 6.37831974029541,
    "timestamp": "2025-07-14T12:33:34.016030"
  },
  {
    "epoch": 18,
    "batch_idx": 160,
    "loss": 0.4167326092720032,
    "learning_rate": 0.0009998001538251283,
    "timestep_mean": 457.78125,
    "epsilon_norm": 7.861793041229248,
    "prediction_norm": 5.6844868659973145,
    "timestamp": "2025-07-14T12:33:34.021250"
  },
  {
    "epoch": 18,
    "batch_idx": 192,
    "loss": 0.1644323468208313,
    "learning_rate": 0.0009998001538251283,
    "timestep_mean": 659.375,
    "epsilon_norm": 4.296873092651367,
    "prediction_norm": 3.9074113368988037,
    "timestamp": "2025-07-14T12:33:34.026282"
  },
  {
    "epoch": 19,
    "batch_idx": 0,
    "loss": 0.6965703368186951,
    "learning_rate": 0.0009997773335794417,
    "timestep_mean": 412.28125,
    "epsilon_norm": 9.344977378845215,
    "prediction_norm": 6.637321472167969,
    "timestamp": "2025-07-14T12:33:34.031482"
  },
  {
    "epoch": 19,
    "batch_idx": 32,
    "loss": 0.39011678099632263,
    "learning_rate": 0.0009997773335794417,
    "timestep_mean": 519.71875,
    "epsilon_norm": 8.698805809020996,
    "prediction_norm": 6.888340473175049,
    "timestamp": "2025-07-14T12:33:34.036725"
  },
  {
    "epoch": 19,
    "batch_idx": 64,
    "loss": 0.5390080213546753,
    "learning_rate": 0.0009997773335794417,
    "timestep_mean": 433.9375,
    "epsilon_norm": 7.907402992248535,
    "prediction_norm": 5.54091215133667,
    "timestamp": "2025-07-14T12:33:34.042119"
  },
  {
    "epoch": 19,
    "batch_idx": 96,
    "loss": 0.5638591051101685,
    "learning_rate": 0.0009997773335794417,
    "timestep_mean": 487.21875,
    "epsilon_norm": 8.180967330932617,
    "prediction_norm": 5.837348937988281,
    "timestamp": "2025-07-14T12:33:34.047132"
  },
  {
    "epoch": 19,
    "batch_idx": 128,
    "loss": 0.4933106303215027,
    "learning_rate": 0.0009997773335794417,
    "timestep_mean": 428.5,
    "epsilon_norm": 7.35344934463501,
    "prediction_norm": 5.334895610809326,
    "timestamp": "2025-07-14T12:33:34.052319"
  },
  {
    "epoch": 19,
    "batch_idx": 160,
    "loss": 0.3934708535671234,
    "learning_rate": 0.0009997773335794417,
    "timestep_mean": 501.03125,
    "epsilon_norm": 7.4499616622924805,
    "prediction_norm": 5.839881896972656,
    "timestamp": "2025-07-14T12:33:34.057565"
  },
  {
    "epoch": 19,
    "batch_idx": 192,
    "loss": 0.4757155776023865,
    "learning_rate": 0.0009997773335794417,
    "timestep_mean": 487.75,
    "epsilon_norm": 3.9928767681121826,
    "prediction_norm": 2.0490148067474365,
    "timestamp": "2025-07-14T12:33:34.062782"
  },
  {
    "epoch": 20,
    "batch_idx": 0,
    "loss": 0.366152822971344,
    "learning_rate": 0.0009997532801828658,
    "timestep_mean": 501.125,
    "epsilon_norm": 8.597433090209961,
    "prediction_norm": 6.222235202789307,
    "timestamp": "2025-07-14T12:33:34.067949"
  },
  {
    "epoch": 20,
    "batch_idx": 32,
    "loss": 0.4068688452243805,
    "learning_rate": 0.0009997532801828658,
    "timestep_mean": 569.875,
    "epsilon_norm": 8.245372772216797,
    "prediction_norm": 6.78097677230835,
    "timestamp": "2025-07-14T12:33:34.073062"
  },
  {
    "epoch": 20,
    "batch_idx": 64,
    "loss": 0.5871517062187195,
    "learning_rate": 0.0009997532801828658,
    "timestep_mean": 535.90625,
    "epsilon_norm": 7.79585599899292,
    "prediction_norm": 4.762428283691406,
    "timestamp": "2025-07-14T12:33:34.078248"
  },
  {
    "epoch": 20,
    "batch_idx": 96,
    "loss": 0.19351187348365784,
    "learning_rate": 0.0009997532801828658,
    "timestep_mean": 562.6875,
    "epsilon_norm": 6.293637275695801,
    "prediction_norm": 5.468353271484375,
    "timestamp": "2025-07-14T12:33:34.083530"
  },
  {
    "epoch": 20,
    "batch_idx": 128,
    "loss": 0.33949363231658936,
    "learning_rate": 0.0009997532801828658,
    "timestep_mean": 516.875,
    "epsilon_norm": 8.454008102416992,
    "prediction_norm": 6.672039031982422,
    "timestamp": "2025-07-14T12:33:34.088597"
  },
  {
    "epoch": 20,
    "batch_idx": 160,
    "loss": 0.5454520583152771,
    "learning_rate": 0.0009997532801828658,
    "timestep_mean": 470.375,
    "epsilon_norm": 8.269648551940918,
    "prediction_norm": 6.198558330535889,
    "timestamp": "2025-07-14T12:33:34.093838"
  },
  {
    "epoch": 20,
    "batch_idx": 192,
    "loss": 0.3618166744709015,
    "learning_rate": 0.0009997532801828658,
    "timestep_mean": 453.25,
    "epsilon_norm": 3.7212934494018555,
    "prediction_norm": 2.882957696914673,
    "timestamp": "2025-07-14T12:33:34.098918"
  },
  {
    "epoch": 21,
    "batch_idx": 0,
    "loss": 0.5713584423065186,
    "learning_rate": 0.00099972799369475,
    "timestep_mean": 449.1875,
    "epsilon_norm": 7.780467510223389,
    "prediction_norm": 5.395230293273926,
    "timestamp": "2025-07-14T12:33:34.104040"
  },
  {
    "epoch": 21,
    "batch_idx": 32,
    "loss": 0.4005660116672516,
    "learning_rate": 0.00099972799369475,
    "timestep_mean": 446.625,
    "epsilon_norm": 7.278397560119629,
    "prediction_norm": 6.197869777679443,
    "timestamp": "2025-07-14T12:33:34.109174"
  },
  {
    "epoch": 21,
    "batch_idx": 64,
    "loss": 0.29723215103149414,
    "learning_rate": 0.00099972799369475,
    "timestep_mean": 576.875,
    "epsilon_norm": 7.673173427581787,
    "prediction_norm": 7.528564929962158,
    "timestamp": "2025-07-14T12:33:34.114281"
  },
  {
    "epoch": 21,
    "batch_idx": 96,
    "loss": 0.44385790824890137,
    "learning_rate": 0.00099972799369475,
    "timestep_mean": 480.21875,
    "epsilon_norm": 9.239713668823242,
    "prediction_norm": 7.253747463226318,
    "timestamp": "2025-07-14T12:33:34.119578"
  },
  {
    "epoch": 21,
    "batch_idx": 128,
    "loss": 0.47282424569129944,
    "learning_rate": 0.00099972799369475,
    "timestep_mean": 484.9375,
    "epsilon_norm": 9.296870231628418,
    "prediction_norm": 6.840408802032471,
    "timestamp": "2025-07-14T12:33:34.124890"
  },
  {
    "epoch": 21,
    "batch_idx": 160,
    "loss": 0.5365657210350037,
    "learning_rate": 0.00099972799369475,
    "timestep_mean": 527.96875,
    "epsilon_norm": 8.574572563171387,
    "prediction_norm": 6.049302577972412,
    "timestamp": "2025-07-14T12:33:34.129932"
  },
  {
    "epoch": 21,
    "batch_idx": 192,
    "loss": 0.20437239110469818,
    "learning_rate": 0.00099972799369475,
    "timestep_mean": 601.5,
    "epsilon_norm": 3.9749867916107178,
    "prediction_norm": 4.389467239379883,
    "timestamp": "2025-07-14T12:33:34.134978"
  },
  {
    "epoch": 22,
    "batch_idx": 0,
    "loss": 0.5603615045547485,
    "learning_rate": 0.0009997014741774863,
    "timestep_mean": 545.71875,
    "epsilon_norm": 8.11863899230957,
    "prediction_norm": 6.452871322631836,
    "timestamp": "2025-07-14T12:33:34.140082"
  },
  {
    "epoch": 22,
    "batch_idx": 32,
    "loss": 0.38376569747924805,
    "learning_rate": 0.0009997014741774863,
    "timestep_mean": 497.71875,
    "epsilon_norm": 8.003401756286621,
    "prediction_norm": 6.161538124084473,
    "timestamp": "2025-07-14T12:33:34.145278"
  },
  {
    "epoch": 22,
    "batch_idx": 64,
    "loss": 0.44514796137809753,
    "learning_rate": 0.0009997014741774863,
    "timestep_mean": 553.59375,
    "epsilon_norm": 8.478750228881836,
    "prediction_norm": 6.494361877441406,
    "timestamp": "2025-07-14T12:33:34.150425"
  },
  {
    "epoch": 22,
    "batch_idx": 96,
    "loss": 0.3218551576137543,
    "learning_rate": 0.0009997014741774863,
    "timestep_mean": 461.125,
    "epsilon_norm": 7.437495231628418,
    "prediction_norm": 4.966414928436279,
    "timestamp": "2025-07-14T12:33:34.155490"
  },
  {
    "epoch": 22,
    "batch_idx": 128,
    "loss": 0.6149073839187622,
    "learning_rate": 0.0009997014741774863,
    "timestep_mean": 491.9375,
    "epsilon_norm": 8.348165512084961,
    "prediction_norm": 5.304932117462158,
    "timestamp": "2025-07-14T12:33:34.160685"
  },
  {
    "epoch": 22,
    "batch_idx": 160,
    "loss": 0.522584080696106,
    "learning_rate": 0.0009997014741774863,
    "timestep_mean": 525.0625,
    "epsilon_norm": 8.986251831054688,
    "prediction_norm": 6.032813549041748,
    "timestamp": "2025-07-14T12:33:34.165875"
  },
  {
    "epoch": 22,
    "batch_idx": 192,
    "loss": 0.6789526343345642,
    "learning_rate": 0.0009997014741774863,
    "timestep_mean": 412.25,
    "epsilon_norm": 4.707398891448975,
    "prediction_norm": 2.3138256072998047,
    "timestamp": "2025-07-14T12:33:34.171026"
  },
  {
    "epoch": 23,
    "batch_idx": 0,
    "loss": 0.2669428586959839,
    "learning_rate": 0.000999673721696509,
    "timestep_mean": 639.90625,
    "epsilon_norm": 8.152395248413086,
    "prediction_norm": 7.798718452453613,
    "timestamp": "2025-07-14T12:33:34.176197"
  },
  {
    "epoch": 23,
    "batch_idx": 32,
    "loss": 0.44384637475013733,
    "learning_rate": 0.000999673721696509,
    "timestep_mean": 526.59375,
    "epsilon_norm": 7.68580436706543,
    "prediction_norm": 5.232377052307129,
    "timestamp": "2025-07-14T12:33:34.181377"
  },
  {
    "epoch": 23,
    "batch_idx": 64,
    "loss": 0.5479772686958313,
    "learning_rate": 0.000999673721696509,
    "timestep_mean": 494.59375,
    "epsilon_norm": 6.489279270172119,
    "prediction_norm": 4.635098934173584,
    "timestamp": "2025-07-14T12:33:34.186655"
  },
  {
    "epoch": 23,
    "batch_idx": 96,
    "loss": 0.42915570735931396,
    "learning_rate": 0.000999673721696509,
    "timestep_mean": 463.875,
    "epsilon_norm": 7.94487190246582,
    "prediction_norm": 5.952270984649658,
    "timestamp": "2025-07-14T12:33:34.191841"
  },
  {
    "epoch": 23,
    "batch_idx": 128,
    "loss": 0.6002069711685181,
    "learning_rate": 0.000999673721696509,
    "timestep_mean": 441.0625,
    "epsilon_norm": 8.290504455566406,
    "prediction_norm": 5.848984718322754,
    "timestamp": "2025-07-14T12:33:34.196947"
  },
  {
    "epoch": 23,
    "batch_idx": 160,
    "loss": 0.508122444152832,
    "learning_rate": 0.000999673721696509,
    "timestep_mean": 474.09375,
    "epsilon_norm": 9.191600799560547,
    "prediction_norm": 7.161291599273682,
    "timestamp": "2025-07-14T12:33:34.202148"
  },
  {
    "epoch": 23,
    "batch_idx": 192,
    "loss": 0.4847111105918884,
    "learning_rate": 0.000999673721696509,
    "timestep_mean": 414.25,
    "epsilon_norm": 3.758592128753662,
    "prediction_norm": 3.1288671493530273,
    "timestamp": "2025-07-14T12:33:34.207281"
  },
  {
    "epoch": 24,
    "batch_idx": 0,
    "loss": 0.4054192900657654,
    "learning_rate": 0.0009996447363202943,
    "timestep_mean": 469.59375,
    "epsilon_norm": 8.373442649841309,
    "prediction_norm": 5.975142955780029,
    "timestamp": "2025-07-14T12:33:34.212517"
  },
  {
    "epoch": 24,
    "batch_idx": 32,
    "loss": 0.24749082326889038,
    "learning_rate": 0.0009996447363202943,
    "timestep_mean": 500.15625,
    "epsilon_norm": 6.4850592613220215,
    "prediction_norm": 5.016002178192139,
    "timestamp": "2025-07-14T12:33:34.217606"
  },
  {
    "epoch": 24,
    "batch_idx": 64,
    "loss": 0.4000982940196991,
    "learning_rate": 0.0009996447363202943,
    "timestep_mean": 422.90625,
    "epsilon_norm": 7.813763618469238,
    "prediction_norm": 5.52808952331543,
    "timestamp": "2025-07-14T12:33:34.222739"
  },
  {
    "epoch": 24,
    "batch_idx": 96,
    "loss": 0.7359340190887451,
    "learning_rate": 0.0009996447363202943,
    "timestep_mean": 481.78125,
    "epsilon_norm": 9.782489776611328,
    "prediction_norm": 6.5062456130981445,
    "timestamp": "2025-07-14T12:33:34.227985"
  },
  {
    "epoch": 24,
    "batch_idx": 128,
    "loss": 0.5611652135848999,
    "learning_rate": 0.0009996447363202943,
    "timestep_mean": 462.5625,
    "epsilon_norm": 8.302457809448242,
    "prediction_norm": 5.371086120605469,
    "timestamp": "2025-07-14T12:33:34.233101"
  },
  {
    "epoch": 24,
    "batch_idx": 160,
    "loss": 0.44795578718185425,
    "learning_rate": 0.0009996447363202943,
    "timestep_mean": 542.46875,
    "epsilon_norm": 8.591243743896484,
    "prediction_norm": 7.170034885406494,
    "timestamp": "2025-07-14T12:33:34.238188"
  },
  {
    "epoch": 24,
    "batch_idx": 192,
    "loss": 0.16467277705669403,
    "learning_rate": 0.0009996447363202943,
    "timestep_mean": 536.125,
    "epsilon_norm": 4.077601432800293,
    "prediction_norm": 3.3372700214385986,
    "timestamp": "2025-07-14T12:33:34.243324"
  },
  {
    "epoch": 25,
    "batch_idx": 0,
    "loss": 0.28238028287887573,
    "learning_rate": 0.0009996145181203613,
    "timestep_mean": 498.0625,
    "epsilon_norm": 7.649423599243164,
    "prediction_norm": 6.304291725158691,
    "timestamp": "2025-07-14T12:33:34.248432"
  },
  {
    "epoch": 25,
    "batch_idx": 32,
    "loss": 0.6457226276397705,
    "learning_rate": 0.0009996145181203613,
    "timestep_mean": 448.4375,
    "epsilon_norm": 9.589588165283203,
    "prediction_norm": 5.739870071411133,
    "timestamp": "2025-07-14T12:33:34.253598"
  },
  {
    "epoch": 25,
    "batch_idx": 64,
    "loss": 0.43837738037109375,
    "learning_rate": 0.0009996145181203613,
    "timestep_mean": 496.53125,
    "epsilon_norm": 8.352925300598145,
    "prediction_norm": 6.240912437438965,
    "timestamp": "2025-07-14T12:33:34.258789"
  },
  {
    "epoch": 25,
    "batch_idx": 96,
    "loss": 0.24841895699501038,
    "learning_rate": 0.0009996145181203613,
    "timestep_mean": 554.25,
    "epsilon_norm": 6.688072204589844,
    "prediction_norm": 6.897038459777832,
    "timestamp": "2025-07-14T12:33:34.263932"
  },
  {
    "epoch": 25,
    "batch_idx": 128,
    "loss": 0.45202380418777466,
    "learning_rate": 0.0009996145181203613,
    "timestep_mean": 473.71875,
    "epsilon_norm": 8.2609281539917,
    "prediction_norm": 6.62941312789917,
    "timestamp": "2025-07-14T12:33:34.269199"
  },
  {
    "epoch": 25,
    "batch_idx": 160,
    "loss": 0.45392894744873047,
    "learning_rate": 0.0009996145181203613,
    "timestep_mean": 534.71875,
    "epsilon_norm": 7.752140045166016,
    "prediction_norm": 6.484078884124756,
    "timestamp": "2025-07-14T12:33:34.274417"
  },
  {
    "epoch": 25,
    "batch_idx": 192,
    "loss": 0.21683824062347412,
    "learning_rate": 0.0009996145181203613,
    "timestep_mean": 537.25,
    "epsilon_norm": 4.266556739807129,
    "prediction_norm": 4.56600284576416,
    "timestamp": "2025-07-14T12:33:34.279633"
  },
  {
    "epoch": 26,
    "batch_idx": 0,
    "loss": 0.3765743374824524,
    "learning_rate": 0.0009995830671712697,
    "timestep_mean": 503.09375,
    "epsilon_norm": 7.588694095611572,
    "prediction_norm": 6.071847438812256,
    "timestamp": "2025-07-14T12:33:34.284874"
  },
  {
    "epoch": 26,
    "batch_idx": 32,
    "loss": 0.4660848081111908,
    "learning_rate": 0.0009995830671712697,
    "timestep_mean": 580.65625,
    "epsilon_norm": 8.696836471557617,
    "prediction_norm": 6.639126777648926,
    "timestamp": "2025-07-14T12:33:34.289869"
  },
  {
    "epoch": 26,
    "batch_idx": 64,
    "loss": 0.317588210105896,
    "learning_rate": 0.0009995830671712697,
    "timestep_mean": 472.21875,
    "epsilon_norm": 7.437856674194336,
    "prediction_norm": 5.886825084686279,
    "timestamp": "2025-07-14T12:33:34.295122"
  },
  {
    "epoch": 26,
    "batch_idx": 96,
    "loss": 0.3307645916938782,
    "learning_rate": 0.0009995830671712697,
    "timestep_mean": 503.375,
    "epsilon_norm": 7.7024078369140625,
    "prediction_norm": 5.858065605163574,
    "timestamp": "2025-07-14T12:33:34.300315"
  },
  {
    "epoch": 26,
    "batch_idx": 128,
    "loss": 0.6184642910957336,
    "learning_rate": 0.0009995830671712697,
    "timestep_mean": 446.375,
    "epsilon_norm": 8.45130443572998,
    "prediction_norm": 5.47928524017334,
    "timestamp": "2025-07-14T12:33:34.305452"
  },
  {
    "epoch": 26,
    "batch_idx": 160,
    "loss": 0.12237434089183807,
    "learning_rate": 0.0009995830671712697,
    "timestep_mean": 581.53125,
    "epsilon_norm": 7.145048141479492,
    "prediction_norm": 6.424263000488281,
    "timestamp": "2025-07-14T12:33:34.310635"
  },
  {
    "epoch": 26,
    "batch_idx": 192,
    "loss": 0.4684421420097351,
    "learning_rate": 0.0009995830671712697,
    "timestep_mean": 569.625,
    "epsilon_norm": 4.27100944519043,
    "prediction_norm": 3.574542999267578,
    "timestamp": "2025-07-14T12:33:34.315770"
  },
  {
    "epoch": 27,
    "batch_idx": 0,
    "loss": 0.33709782361984253,
    "learning_rate": 0.0009995503835506221,
    "timestep_mean": 604.28125,
    "epsilon_norm": 8.519180297851562,
    "prediction_norm": 6.76296329498291,
    "timestamp": "2025-07-14T12:33:34.320918"
  },
  {
    "epoch": 27,
    "batch_idx": 32,
    "loss": 0.5372114181518555,
    "learning_rate": 0.0009995503835506221,
    "timestep_mean": 409.9375,
    "epsilon_norm": 8.189906120300293,
    "prediction_norm": 5.724071025848389,
    "timestamp": "2025-07-14T12:33:34.326001"
  },
  {
    "epoch": 27,
    "batch_idx": 64,
    "loss": 0.2833407521247864,
    "learning_rate": 0.0009995503835506221,
    "timestep_mean": 524.09375,
    "epsilon_norm": 7.7379350662231445,
    "prediction_norm": 5.896979808807373,
    "timestamp": "2025-07-14T12:33:34.331087"
  },
  {
    "epoch": 27,
    "batch_idx": 96,
    "loss": 0.4787154495716095,
    "learning_rate": 0.0009995503835506221,
    "timestep_mean": 505.90625,
    "epsilon_norm": 9.155503273010254,
    "prediction_norm": 6.994099140167236,
    "timestamp": "2025-07-14T12:33:34.336424"
  },
  {
    "epoch": 27,
    "batch_idx": 128,
    "loss": 0.30440419912338257,
    "learning_rate": 0.0009995503835506221,
    "timestep_mean": 458.53125,
    "epsilon_norm": 8.126705169677734,
    "prediction_norm": 6.127561092376709,
    "timestamp": "2025-07-14T12:33:34.341701"
  },
  {
    "epoch": 27,
    "batch_idx": 160,
    "loss": 0.284413605928421,
    "learning_rate": 0.0009995503835506221,
    "timestep_mean": 527.09375,
    "epsilon_norm": 7.493990898132324,
    "prediction_norm": 6.089718341827393,
    "timestamp": "2025-07-14T12:33:34.346884"
  },
  {
    "epoch": 27,
    "batch_idx": 192,
    "loss": 0.40265804529190063,
    "learning_rate": 0.0009995503835506221,
    "timestep_mean": 491.0,
    "epsilon_norm": 3.643343925476074,
    "prediction_norm": 2.634305000305176,
    "timestamp": "2025-07-14T12:33:34.352016"
  },
  {
    "epoch": 28,
    "batch_idx": 0,
    "loss": 0.3860037922859192,
    "learning_rate": 0.000999516467339062,
    "timestep_mean": 508.28125,
    "epsilon_norm": 7.621757507324219,
    "prediction_norm": 5.53905725479126,
    "timestamp": "2025-07-14T12:33:34.357292"
  },
  {
    "epoch": 28,
    "batch_idx": 32,
    "loss": 0.4106122851371765,
    "learning_rate": 0.000999516467339062,
    "timestep_mean": 524.90625,
    "epsilon_norm": 7.702131271362305,
    "prediction_norm": 5.332468032836914,
    "timestamp": "2025-07-14T12:33:34.362486"
  },
  {
    "epoch": 28,
    "batch_idx": 64,
    "loss": 0.3048815429210663,
    "learning_rate": 0.000999516467339062,
    "timestep_mean": 504.15625,
    "epsilon_norm": 8.846587181091309,
    "prediction_norm": 7.599952220916748,
    "timestamp": "2025-07-14T12:33:34.367671"
  },
  {
    "epoch": 28,
    "batch_idx": 96,
    "loss": 0.4647827446460724,
    "learning_rate": 0.000999516467339062,
    "timestep_mean": 414.5,
    "epsilon_norm": 7.85323429107666,
    "prediction_norm": 5.452749729156494,
    "timestamp": "2025-07-14T12:33:34.372742"
  },
  {
    "epoch": 28,
    "batch_idx": 128,
    "loss": 0.4195484220981598,
    "learning_rate": 0.000999516467339062,
    "timestep_mean": 525.875,
    "epsilon_norm": 7.410060882568359,
    "prediction_norm": 5.787480354309082,
    "timestamp": "2025-07-14T12:33:34.378000"
  },
  {
    "epoch": 28,
    "batch_idx": 160,
    "loss": 0.4893982410430908,
    "learning_rate": 0.000999516467339062,
    "timestep_mean": 488.59375,
    "epsilon_norm": 7.409452438354492,
    "prediction_norm": 5.120156288146973,
    "timestamp": "2025-07-14T12:33:34.383257"
  },
  {
    "epoch": 28,
    "batch_idx": 192,
    "loss": 0.23825758695602417,
    "learning_rate": 0.000999516467339062,
    "timestep_mean": 472.5,
    "epsilon_norm": 2.3409979343414307,
    "prediction_norm": 1.9570542573928833,
    "timestamp": "2025-07-14T12:33:34.388296"
  },
  {
    "epoch": 29,
    "batch_idx": 0,
    "loss": 0.4667438566684723,
    "learning_rate": 0.0009994813186202742,
    "timestep_mean": 431.40625,
    "epsilon_norm": 8.359573364257812,
    "prediction_norm": 6.7189249992370605,
    "timestamp": "2025-07-14T12:33:34.393450"
  },
  {
    "epoch": 29,
    "batch_idx": 32,
    "loss": 0.2487085908651352,
    "learning_rate": 0.0009994813186202742,
    "timestep_mean": 580.34375,
    "epsilon_norm": 6.630135536193848,
    "prediction_norm": 5.867722511291504,
    "timestamp": "2025-07-14T12:33:34.398713"
  },
  {
    "epoch": 29,
    "batch_idx": 64,
    "loss": 0.46342986822128296,
    "learning_rate": 0.0009994813186202742,
    "timestep_mean": 460.21875,
    "epsilon_norm": 8.252580642700195,
    "prediction_norm": 6.625604629516602,
    "timestamp": "2025-07-14T12:33:34.403843"
  },
  {
    "epoch": 29,
    "batch_idx": 96,
    "loss": 0.15748366713523865,
    "learning_rate": 0.0009994813186202742,
    "timestep_mean": 584.28125,
    "epsilon_norm": 7.588766098022461,
    "prediction_norm": 7.1586151123046875,
    "timestamp": "2025-07-14T12:33:34.408841"
  },
  {
    "epoch": 29,
    "batch_idx": 128,
    "loss": 0.41073697805404663,
    "learning_rate": 0.0009994813186202742,
    "timestep_mean": 577.15625,
    "epsilon_norm": 8.29539966583252,
    "prediction_norm": 6.675278186798096,
    "timestamp": "2025-07-14T12:33:34.413768"
  },
  {
    "epoch": 29,
    "batch_idx": 160,
    "loss": 0.2805934250354767,
    "learning_rate": 0.0009994813186202742,
    "timestep_mean": 460.8125,
    "epsilon_norm": 7.4971466064453125,
    "prediction_norm": 6.097380638122559,
    "timestamp": "2025-07-14T12:33:34.419116"
  },
  {
    "epoch": 29,
    "batch_idx": 192,
    "loss": 0.22227339446544647,
    "learning_rate": 0.0009994813186202742,
    "timestep_mean": 498.625,
    "epsilon_norm": 4.170774459838867,
    "prediction_norm": 3.678279161453247,
    "timestamp": "2025-07-14T12:33:34.424349"
  },
  {
    "epoch": 30,
    "batch_idx": 0,
    "loss": 0.4324135184288025,
    "learning_rate": 0.0009994449374809847,
    "timestep_mean": 448.09375,
    "epsilon_norm": 7.721848487854004,
    "prediction_norm": 5.4665727615356445,
    "timestamp": "2025-07-14T12:33:34.429520"
  },
  {
    "epoch": 30,
    "batch_idx": 32,
    "loss": 0.3252415657043457,
    "learning_rate": 0.0009994449374809847,
    "timestep_mean": 546.96875,
    "epsilon_norm": 7.339338779449463,
    "prediction_norm": 5.119734287261963,
    "timestamp": "2025-07-14T12:33:34.434755"
  },
  {
    "epoch": 30,
    "batch_idx": 64,
    "loss": 0.34056714177131653,
    "learning_rate": 0.0009994449374809847,
    "timestep_mean": 543.9375,
    "epsilon_norm": 7.9726738929748535,
    "prediction_norm": 6.227763652801514,
    "timestamp": "2025-07-14T12:33:34.440051"
  },
  {
    "epoch": 30,
    "batch_idx": 96,
    "loss": 0.5945218801498413,
    "learning_rate": 0.0009994449374809847,
    "timestep_mean": 423.0625,
    "epsilon_norm": 8.374679565429688,
    "prediction_norm": 4.915669918060303,
    "timestamp": "2025-07-14T12:33:34.445129"
  },
  {
    "epoch": 30,
    "batch_idx": 128,
    "loss": 0.4593685567378998,
    "learning_rate": 0.0009994449374809847,
    "timestep_mean": 574.90625,
    "epsilon_norm": 8.102773666381836,
    "prediction_norm": 6.213034629821777,
    "timestamp": "2025-07-14T12:33:34.450516"
  },
  {
    "epoch": 30,
    "batch_idx": 160,
    "loss": 0.4675840139389038,
    "learning_rate": 0.0009994449374809847,
    "timestep_mean": 462.84375,
    "epsilon_norm": 8.060113906860352,
    "prediction_norm": 6.42482852935791,
    "timestamp": "2025-07-14T12:33:34.455990"
  },
  {
    "epoch": 30,
    "batch_idx": 192,
    "loss": 0.6817132234573364,
    "learning_rate": 0.0009994449374809847,
    "timestep_mean": 375.125,
    "epsilon_norm": 4.501597881317139,
    "prediction_norm": 2.750171422958374,
    "timestamp": "2025-07-14T12:33:34.461172"
  },
  {
    "epoch": 31,
    "batch_idx": 0,
    "loss": 0.472756564617157,
    "learning_rate": 0.0009994073240109603,
    "timestep_mean": 432.875,
    "epsilon_norm": 9.285785675048828,
    "prediction_norm": 7.628361701965332,
    "timestamp": "2025-07-14T12:33:34.466591"
  },
  {
    "epoch": 31,
    "batch_idx": 32,
    "loss": 0.5162158012390137,
    "learning_rate": 0.0009994073240109603,
    "timestep_mean": 446.96875,
    "epsilon_norm": 8.22655963897705,
    "prediction_norm": 5.76530647277832,
    "timestamp": "2025-07-14T12:33:34.471767"
  },
  {
    "epoch": 31,
    "batch_idx": 64,
    "loss": 0.4652702808380127,
    "learning_rate": 0.0009994073240109603,
    "timestep_mean": 473.125,
    "epsilon_norm": 7.515713214874268,
    "prediction_norm": 5.766819477081299,
    "timestamp": "2025-07-14T12:33:34.476982"
  },
  {
    "epoch": 31,
    "batch_idx": 96,
    "loss": 0.6770790815353394,
    "learning_rate": 0.0009994073240109603,
    "timestep_mean": 465.375,
    "epsilon_norm": 8.513998031616211,
    "prediction_norm": 6.217909336090088,
    "timestamp": "2025-07-14T12:33:34.482111"
  },
  {
    "epoch": 31,
    "batch_idx": 128,
    "loss": 0.5416494607925415,
    "learning_rate": 0.0009994073240109603,
    "timestep_mean": 434.53125,
    "epsilon_norm": 8.34414291381836,
    "prediction_norm": 6.480800628662109,
    "timestamp": "2025-07-14T12:33:34.487156"
  },
  {
    "epoch": 31,
    "batch_idx": 160,
    "loss": 0.4125344753265381,
    "learning_rate": 0.0009994073240109603,
    "timestep_mean": 471.34375,
    "epsilon_norm": 7.659249305725098,
    "prediction_norm": 5.97114896774292,
    "timestamp": "2025-07-14T12:33:34.492283"
  },
  {
    "epoch": 31,
    "batch_idx": 192,
    "loss": 0.37423962354660034,
    "learning_rate": 0.0009994073240109603,
    "timestep_mean": 412.75,
    "epsilon_norm": 3.174255847930908,
    "prediction_norm": 2.146176338195801,
    "timestamp": "2025-07-14T12:33:34.497370"
  },
  {
    "epoch": 32,
    "batch_idx": 0,
    "loss": 0.2390536665916443,
    "learning_rate": 0.0009993684783030086,
    "timestep_mean": 640.15625,
    "epsilon_norm": 8.661609649658203,
    "prediction_norm": 8.174717903137207,
    "timestamp": "2025-07-14T12:33:34.508464"
  },
  {
    "epoch": 32,
    "batch_idx": 32,
    "loss": 0.2973710298538208,
    "learning_rate": 0.0009993684783030086,
    "timestep_mean": 530.5625,
    "epsilon_norm": 8.158773422241211,
    "prediction_norm": 7.040084362030029,
    "timestamp": "2025-07-14T12:33:34.513326"
  },
  {
    "epoch": 32,
    "batch_idx": 64,
    "loss": 0.5519070029258728,
    "learning_rate": 0.0009993684783030086,
    "timestep_mean": 455.8125,
    "epsilon_norm": 7.788482666015625,
    "prediction_norm": 5.521537780761719,
    "timestamp": "2025-07-14T12:33:34.518417"
  },
  {
    "epoch": 32,
    "batch_idx": 96,
    "loss": 0.1759326159954071,
    "learning_rate": 0.0009993684783030086,
    "timestep_mean": 586.5625,
    "epsilon_norm": 6.211206436157227,
    "prediction_norm": 5.436280250549316,
    "timestamp": "2025-07-14T12:33:34.523504"
  },
  {
    "epoch": 32,
    "batch_idx": 128,
    "loss": 0.5993180871009827,
    "learning_rate": 0.0009993684783030086,
    "timestep_mean": 466.8125,
    "epsilon_norm": 8.867569923400879,
    "prediction_norm": 6.170494556427002,
    "timestamp": "2025-07-14T12:33:34.528410"
  },
  {
    "epoch": 32,
    "batch_idx": 160,
    "loss": 0.5010381937026978,
    "learning_rate": 0.0009993684783030086,
    "timestep_mean": 435.625,
    "epsilon_norm": 8.394979476928711,
    "prediction_norm": 6.134125709533691,
    "timestamp": "2025-07-14T12:33:34.533467"
  },
  {
    "epoch": 32,
    "batch_idx": 192,
    "loss": 0.9012390375137329,
    "learning_rate": 0.0009993684783030086,
    "timestep_mean": 359.625,
    "epsilon_norm": 3.8284215927124023,
    "prediction_norm": 2.416416883468628,
    "timestamp": "2025-07-14T12:33:34.538369"
  },
  {
    "epoch": 33,
    "batch_idx": 0,
    "loss": 0.4502248764038086,
    "learning_rate": 0.0009993284004529773,
    "timestep_mean": 420.15625,
    "epsilon_norm": 7.948049068450928,
    "prediction_norm": 4.789514541625977,
    "timestamp": "2025-07-14T12:33:34.543302"
  },
  {
    "epoch": 33,
    "batch_idx": 32,
    "loss": 0.4254531264305115,
    "learning_rate": 0.0009993284004529773,
    "timestep_mean": 407.59375,
    "epsilon_norm": 8.53255558013916,
    "prediction_norm": 5.582605361938477,
    "timestamp": "2025-07-14T12:33:34.548283"
  },
  {
    "epoch": 33,
    "batch_idx": 64,
    "loss": 0.39605993032455444,
    "learning_rate": 0.0009993284004529773,
    "timestep_mean": 501.65625,
    "epsilon_norm": 7.539606094360352,
    "prediction_norm": 6.30483865737915,
    "timestamp": "2025-07-14T12:33:34.553377"
  },
  {
    "epoch": 33,
    "batch_idx": 96,
    "loss": 0.47897788882255554,
    "learning_rate": 0.0009993284004529773,
    "timestep_mean": 490.21875,
    "epsilon_norm": 8.073798179626465,
    "prediction_norm": 5.058511734008789,
    "timestamp": "2025-07-14T12:33:34.558399"
  },
  {
    "epoch": 33,
    "batch_idx": 128,
    "loss": 0.33676475286483765,
    "learning_rate": 0.0009993284004529773,
    "timestep_mean": 500.375,
    "epsilon_norm": 7.86513614654541,
    "prediction_norm": 5.245326042175293,
    "timestamp": "2025-07-14T12:33:34.563247"
  },
  {
    "epoch": 33,
    "batch_idx": 160,
    "loss": 0.4072301685810089,
    "learning_rate": 0.0009993284004529773,
    "timestep_mean": 543.375,
    "epsilon_norm": 8.39310073852539,
    "prediction_norm": 6.365594387054443,
    "timestamp": "2025-07-14T12:33:34.568315"
  },
  {
    "epoch": 33,
    "batch_idx": 192,
    "loss": 0.39908167719841003,
    "learning_rate": 0.0009993284004529773,
    "timestep_mean": 496.75,
    "epsilon_norm": 3.983036994934082,
    "prediction_norm": 3.0064005851745605,
    "timestamp": "2025-07-14T12:33:34.573309"
  },
  {
    "epoch": 34,
    "batch_idx": 0,
    "loss": 0.46956145763397217,
    "learning_rate": 0.0009992870905597548,
    "timestep_mean": 484.59375,
    "epsilon_norm": 8.005860328674316,
    "prediction_norm": 5.76733922958374,
    "timestamp": "2025-07-14T12:33:34.578332"
  },
  {
    "epoch": 34,
    "batch_idx": 32,
    "loss": 0.3990848660469055,
    "learning_rate": 0.0009992870905597548,
    "timestep_mean": 541.90625,
    "epsilon_norm": 9.559623718261719,
    "prediction_norm": 7.45894718170166,
    "timestamp": "2025-07-14T12:33:34.583396"
  },
  {
    "epoch": 34,
    "batch_idx": 64,
    "loss": 0.18664692342281342,
    "learning_rate": 0.0009992870905597548,
    "timestep_mean": 553.09375,
    "epsilon_norm": 7.314888954162598,
    "prediction_norm": 5.928837776184082,
    "timestamp": "2025-07-14T12:33:34.588386"
  },
  {
    "epoch": 34,
    "batch_idx": 96,
    "loss": 0.37491482496261597,
    "learning_rate": 0.0009992870905597548,
    "timestep_mean": 582.5625,
    "epsilon_norm": 8.0082426071167,
    "prediction_norm": 7.082531452178955,
    "timestamp": "2025-07-14T12:33:34.593509"
  },
  {
    "epoch": 34,
    "batch_idx": 128,
    "loss": 0.19453129172325134,
    "learning_rate": 0.0009992870905597548,
    "timestep_mean": 590.125,
    "epsilon_norm": 7.1376471519470215,
    "prediction_norm": 6.5526041984558105,
    "timestamp": "2025-07-14T12:33:34.598553"
  },
  {
    "epoch": 34,
    "batch_idx": 160,
    "loss": 0.5354225635528564,
    "learning_rate": 0.0009992870905597548,
    "timestep_mean": 444.34375,
    "epsilon_norm": 8.063992500305176,
    "prediction_norm": 5.638119697570801,
    "timestamp": "2025-07-14T12:33:34.603583"
  },
  {
    "epoch": 34,
    "batch_idx": 192,
    "loss": 0.22279489040374756,
    "learning_rate": 0.0009992870905597548,
    "timestep_mean": 443.75,
    "epsilon_norm": 4.084190368652344,
    "prediction_norm": 3.7224340438842773,
    "timestamp": "2025-07-14T12:33:34.608729"
  },
  {
    "epoch": 35,
    "batch_idx": 0,
    "loss": 0.2408783882856369,
    "learning_rate": 0.0009992445487252688,
    "timestep_mean": 577.03125,
    "epsilon_norm": 7.947574138641357,
    "prediction_norm": 6.629696846008301,
    "timestamp": "2025-07-14T12:33:34.613688"
  },
  {
    "epoch": 35,
    "batch_idx": 32,
    "loss": 0.3789792060852051,
    "learning_rate": 0.0009992445487252688,
    "timestep_mean": 514.15625,
    "epsilon_norm": 7.278088569641113,
    "prediction_norm": 6.293733596801758,
    "timestamp": "2025-07-14T12:33:34.618871"
  },
  {
    "epoch": 35,
    "batch_idx": 64,
    "loss": 0.25902819633483887,
    "learning_rate": 0.0009992445487252688,
    "timestep_mean": 544.3125,
    "epsilon_norm": 7.37811803817749,
    "prediction_norm": 6.433288097381592,
    "timestamp": "2025-07-14T12:33:34.623924"
  },
  {
    "epoch": 35,
    "batch_idx": 96,
    "loss": 0.4501807689666748,
    "learning_rate": 0.0009992445487252688,
    "timestep_mean": 438.6875,
    "epsilon_norm": 7.684645652770996,
    "prediction_norm": 6.2453718185424805,
    "timestamp": "2025-07-14T12:33:34.628988"
  },
  {
    "epoch": 35,
    "batch_idx": 128,
    "loss": 0.3790264129638672,
    "learning_rate": 0.0009992445487252688,
    "timestep_mean": 508.8125,
    "epsilon_norm": 8.278120040893555,
    "prediction_norm": 7.400768756866455,
    "timestamp": "2025-07-14T12:33:34.634132"
  },
  {
    "epoch": 35,
    "batch_idx": 160,
    "loss": 0.4262533485889435,
    "learning_rate": 0.0009992445487252688,
    "timestep_mean": 425.8125,
    "epsilon_norm": 7.021524429321289,
    "prediction_norm": 5.993009090423584,
    "timestamp": "2025-07-14T12:33:34.639205"
  },
  {
    "epoch": 35,
    "batch_idx": 192,
    "loss": 0.6301412582397461,
    "learning_rate": 0.0009992445487252688,
    "timestep_mean": 434.75,
    "epsilon_norm": 3.355333089828491,
    "prediction_norm": 1.6536848545074463,
    "timestamp": "2025-07-14T12:33:34.644376"
  },
  {
    "epoch": 36,
    "batch_idx": 0,
    "loss": 0.3451022803783417,
    "learning_rate": 0.0009992007750544874,
    "timestep_mean": 436.71875,
    "epsilon_norm": 7.451237201690674,
    "prediction_norm": 6.482329845428467,
    "timestamp": "2025-07-14T12:33:34.649764"
  },
  {
    "epoch": 36,
    "batch_idx": 32,
    "loss": 0.44921785593032837,
    "learning_rate": 0.0009992007750544874,
    "timestep_mean": 465.3125,
    "epsilon_norm": 7.75155782699585,
    "prediction_norm": 5.2523322105407715,
    "timestamp": "2025-07-14T12:33:34.654858"
  },
  {
    "epoch": 36,
    "batch_idx": 64,
    "loss": 0.5132143497467041,
    "learning_rate": 0.0009992007750544874,
    "timestep_mean": 563.78125,
    "epsilon_norm": 8.646661758422852,
    "prediction_norm": 6.391757488250732,
    "timestamp": "2025-07-14T12:33:34.660101"
  },
  {
    "epoch": 36,
    "batch_idx": 96,
    "loss": 0.37323302030563354,
    "learning_rate": 0.0009992007750544874,
    "timestep_mean": 512.75,
    "epsilon_norm": 8.169767379760742,
    "prediction_norm": 5.752200603485107,
    "timestamp": "2025-07-14T12:33:34.665291"
  },
  {
    "epoch": 36,
    "batch_idx": 128,
    "loss": 0.24900785088539124,
    "learning_rate": 0.0009992007750544874,
    "timestep_mean": 581.59375,
    "epsilon_norm": 7.732736587524414,
    "prediction_norm": 5.670712947845459,
    "timestamp": "2025-07-14T12:33:34.670476"
  },
  {
    "epoch": 36,
    "batch_idx": 160,
    "loss": 0.32415971159935,
    "learning_rate": 0.0009992007750544874,
    "timestep_mean": 531.21875,
    "epsilon_norm": 8.474204063415527,
    "prediction_norm": 6.03059720993042,
    "timestamp": "2025-07-14T12:33:34.675649"
  },
  {
    "epoch": 36,
    "batch_idx": 192,
    "loss": 0.16578975319862366,
    "learning_rate": 0.0009992007750544874,
    "timestep_mean": 592.875,
    "epsilon_norm": 3.470107316970825,
    "prediction_norm": 2.9458932876586914,
    "timestamp": "2025-07-14T12:33:34.680788"
  },
  {
    "epoch": 37,
    "batch_idx": 0,
    "loss": 0.28419139981269836,
    "learning_rate": 0.0009991557696554174,
    "timestep_mean": 531.75,
    "epsilon_norm": 7.438095569610596,
    "prediction_norm": 6.052523136138916,
    "timestamp": "2025-07-14T12:33:34.686026"
  },
  {
    "epoch": 37,
    "batch_idx": 32,
    "loss": 0.24596866965293884,
    "learning_rate": 0.0009991557696554174,
    "timestep_mean": 480.625,
    "epsilon_norm": 7.782995700836182,
    "prediction_norm": 6.595653533935547,
    "timestamp": "2025-07-14T12:33:34.691210"
  },
  {
    "epoch": 37,
    "batch_idx": 64,
    "loss": 0.3018166422843933,
    "learning_rate": 0.0009991557696554174,
    "timestep_mean": 462.96875,
    "epsilon_norm": 6.634260654449463,
    "prediction_norm": 5.695091247558594,
    "timestamp": "2025-07-14T12:33:34.696428"
  },
  {
    "epoch": 37,
    "batch_idx": 96,
    "loss": 0.5334321856498718,
    "learning_rate": 0.0009991557696554174,
    "timestep_mean": 498.71875,
    "epsilon_norm": 8.039973258972168,
    "prediction_norm": 6.381376266479492,
    "timestamp": "2025-07-14T12:33:34.701596"
  },
  {
    "epoch": 37,
    "batch_idx": 128,
    "loss": 0.3689291477203369,
    "learning_rate": 0.0009991557696554174,
    "timestep_mean": 526.84375,
    "epsilon_norm": 7.775177478790283,
    "prediction_norm": 6.389034748077393,
    "timestamp": "2025-07-14T12:33:34.706800"
  },
  {
    "epoch": 37,
    "batch_idx": 160,
    "loss": 0.34393179416656494,
    "learning_rate": 0.0009991557696554174,
    "timestep_mean": 517.46875,
    "epsilon_norm": 7.943103790283203,
    "prediction_norm": 6.436698913574219,
    "timestamp": "2025-07-14T12:33:34.711948"
  },
  {
    "epoch": 37,
    "batch_idx": 192,
    "loss": 0.4079653024673462,
    "learning_rate": 0.0009991557696554174,
    "timestep_mean": 410.625,
    "epsilon_norm": 4.1933979988098145,
    "prediction_norm": 3.4981446266174316,
    "timestamp": "2025-07-14T12:33:34.717244"
  },
  {
    "epoch": 38,
    "batch_idx": 0,
    "loss": 0.32437199354171753,
    "learning_rate": 0.0009991095326391058,
    "timestep_mean": 519.75,
    "epsilon_norm": 7.521472930908203,
    "prediction_norm": 5.828990936279297,
    "timestamp": "2025-07-14T12:33:34.722335"
  },
  {
    "epoch": 38,
    "batch_idx": 32,
    "loss": 0.3205774426460266,
    "learning_rate": 0.0009991095326391058,
    "timestep_mean": 517.125,
    "epsilon_norm": 7.477350234985352,
    "prediction_norm": 6.086625099182129,
    "timestamp": "2025-07-14T12:33:34.727528"
  },
  {
    "epoch": 38,
    "batch_idx": 64,
    "loss": 0.4242611825466156,
    "learning_rate": 0.0009991095326391058,
    "timestep_mean": 489.65625,
    "epsilon_norm": 8.08752727508545,
    "prediction_norm": 6.615679740905762,
    "timestamp": "2025-07-14T12:33:34.732411"
  },
  {
    "epoch": 38,
    "batch_idx": 96,
    "loss": 0.45887047052383423,
    "learning_rate": 0.0009991095326391058,
    "timestep_mean": 427.125,
    "epsilon_norm": 7.288559436798096,
    "prediction_norm": 4.799052715301514,
    "timestamp": "2025-07-14T12:33:34.737450"
  },
  {
    "epoch": 38,
    "batch_idx": 128,
    "loss": 0.3045217990875244,
    "learning_rate": 0.0009991095326391058,
    "timestep_mean": 517.125,
    "epsilon_norm": 9.250239372253418,
    "prediction_norm": 7.553296089172363,
    "timestamp": "2025-07-14T12:33:34.742640"
  },
  {
    "epoch": 38,
    "batch_idx": 160,
    "loss": 0.3141452968120575,
    "learning_rate": 0.0009991095326391058,
    "timestep_mean": 566.375,
    "epsilon_norm": 8.689441680908203,
    "prediction_norm": 7.162557125091553,
    "timestamp": "2025-07-14T12:33:34.747622"
  },
  {
    "epoch": 38,
    "batch_idx": 192,
    "loss": 0.7102562189102173,
    "learning_rate": 0.0009991095326391058,
    "timestep_mean": 496.125,
    "epsilon_norm": 5.2509846687316895,
    "prediction_norm": 3.616913318634033,
    "timestamp": "2025-07-14T12:33:34.752466"
  },
  {
    "epoch": 39,
    "batch_idx": 0,
    "loss": 0.3860257863998413,
    "learning_rate": 0.0009990620641196371,
    "timestep_mean": 491.0,
    "epsilon_norm": 7.598736763000488,
    "prediction_norm": 5.53222131729126,
    "timestamp": "2025-07-14T12:33:34.757333"
  },
  {
    "epoch": 39,
    "batch_idx": 32,
    "loss": 0.2715175449848175,
    "learning_rate": 0.0009990620641196371,
    "timestep_mean": 472.28125,
    "epsilon_norm": 7.226045608520508,
    "prediction_norm": 6.340096950531006,
    "timestamp": "2025-07-14T12:33:34.762222"
  },
  {
    "epoch": 39,
    "batch_idx": 64,
    "loss": 0.5049973130226135,
    "learning_rate": 0.0009990620641196371,
    "timestep_mean": 483.125,
    "epsilon_norm": 7.64470100402832,
    "prediction_norm": 5.136837005615234,
    "timestamp": "2025-07-14T12:33:34.767191"
  },
  {
    "epoch": 39,
    "batch_idx": 96,
    "loss": 0.4246383309364319,
    "learning_rate": 0.0009990620641196371,
    "timestep_mean": 518.03125,
    "epsilon_norm": 7.498032569885254,
    "prediction_norm": 6.317548751831055,
    "timestamp": "2025-07-14T12:33:34.772070"
  },
  {
    "epoch": 39,
    "batch_idx": 128,
    "loss": 0.5083227157592773,
    "learning_rate": 0.0009990620641196371,
    "timestep_mean": 449.4375,
    "epsilon_norm": 7.827255725860596,
    "prediction_norm": 5.359033584594727,
    "timestamp": "2025-07-14T12:33:34.776920"
  },
  {
    "epoch": 39,
    "batch_idx": 160,
    "loss": 0.528469979763031,
    "learning_rate": 0.0009990620641196371,
    "timestep_mean": 429.53125,
    "epsilon_norm": 7.740441799163818,
    "prediction_norm": 5.793527603149414,
    "timestamp": "2025-07-14T12:33:34.781751"
  },
  {
    "epoch": 39,
    "batch_idx": 192,
    "loss": 0.5297278165817261,
    "learning_rate": 0.0009990620641196371,
    "timestep_mean": 262.75,
    "epsilon_norm": 4.139678955078125,
    "prediction_norm": 2.1847176551818848,
    "timestamp": "2025-07-14T12:33:34.786592"
  },
  {
    "epoch": 40,
    "batch_idx": 0,
    "loss": 0.3422243595123291,
    "learning_rate": 0.0009990133642141356,
    "timestep_mean": 561.34375,
    "epsilon_norm": 9.033785820007324,
    "prediction_norm": 8.034878730773926,
    "timestamp": "2025-07-14T12:33:34.791529"
  },
  {
    "epoch": 40,
    "batch_idx": 32,
    "loss": 0.3329728841781616,
    "learning_rate": 0.0009990133642141356,
    "timestep_mean": 459.625,
    "epsilon_norm": 6.71198844909668,
    "prediction_norm": 5.223973274230957,
    "timestamp": "2025-07-14T12:33:34.796458"
  },
  {
    "epoch": 40,
    "batch_idx": 64,
    "loss": 0.3418046236038208,
    "learning_rate": 0.0009990133642141356,
    "timestep_mean": 523.40625,
    "epsilon_norm": 7.31744384765625,
    "prediction_norm": 6.282910346984863,
    "timestamp": "2025-07-14T12:33:34.801374"
  },
  {
    "epoch": 40,
    "batch_idx": 96,
    "loss": 0.5157711505889893,
    "learning_rate": 0.0009990133642141356,
    "timestep_mean": 484.3125,
    "epsilon_norm": 8.177388191223145,
    "prediction_norm": 5.5810441970825195,
    "timestamp": "2025-07-14T12:33:34.806289"
  },
  {
    "epoch": 40,
    "batch_idx": 128,
    "loss": 0.5877048969268799,
    "learning_rate": 0.0009990133642141356,
    "timestep_mean": 401.40625,
    "epsilon_norm": 9.02190113067627,
    "prediction_norm": 6.434938907623291,
    "timestamp": "2025-07-14T12:33:34.811190"
  },
  {
    "epoch": 40,
    "batch_idx": 160,
    "loss": 0.592509388923645,
    "learning_rate": 0.0009990133642141356,
    "timestep_mean": 426.6875,
    "epsilon_norm": 8.797382354736328,
    "prediction_norm": 6.067084312438965,
    "timestamp": "2025-07-14T12:33:34.816095"
  },
  {
    "epoch": 40,
    "batch_idx": 192,
    "loss": 0.7392776608467102,
    "learning_rate": 0.0009990133642141356,
    "timestep_mean": 472.5,
    "epsilon_norm": 5.135665416717529,
    "prediction_norm": 3.225037097930908,
    "timestamp": "2025-07-14T12:33:34.820998"
  },
  {
    "epoch": 41,
    "batch_idx": 0,
    "loss": 0.4648030996322632,
    "learning_rate": 0.0009989634330427632,
    "timestep_mean": 544.96875,
    "epsilon_norm": 8.330305099487305,
    "prediction_norm": 7.371079444885254,
    "timestamp": "2025-07-14T12:33:34.825842"
  },
  {
    "epoch": 41,
    "batch_idx": 32,
    "loss": 0.43399927020072937,
    "learning_rate": 0.0009989634330427632,
    "timestep_mean": 569.75,
    "epsilon_norm": 7.810431957244873,
    "prediction_norm": 6.303925037384033,
    "timestamp": "2025-07-14T12:33:34.830703"
  },
  {
    "epoch": 41,
    "batch_idx": 64,
    "loss": 0.2878105342388153,
    "learning_rate": 0.0009989634330427632,
    "timestep_mean": 494.15625,
    "epsilon_norm": 8.674935340881348,
    "prediction_norm": 7.358588218688965,
    "timestamp": "2025-07-14T12:33:34.835587"
  },
  {
    "epoch": 41,
    "batch_idx": 96,
    "loss": 0.30956241488456726,
    "learning_rate": 0.0009989634330427632,
    "timestep_mean": 463.0625,
    "epsilon_norm": 6.770909786224365,
    "prediction_norm": 6.267727375030518,
    "timestamp": "2025-07-14T12:33:34.840417"
  },
  {
    "epoch": 41,
    "batch_idx": 128,
    "loss": 0.7262046337127686,
    "learning_rate": 0.0009989634330427632,
    "timestep_mean": 418.59375,
    "epsilon_norm": 9.125041961669922,
    "prediction_norm": 6.891912460327148,
    "timestamp": "2025-07-14T12:33:34.845266"
  },
  {
    "epoch": 41,
    "batch_idx": 160,
    "loss": 0.3480755686759949,
    "learning_rate": 0.0009989634330427632,
    "timestep_mean": 439.1875,
    "epsilon_norm": 8.089351654052734,
    "prediction_norm": 6.330726146697998,
    "timestamp": "2025-07-14T12:33:34.850198"
  },
  {
    "epoch": 41,
    "batch_idx": 192,
    "loss": 0.14513203501701355,
    "learning_rate": 0.0009989634330427632,
    "timestep_mean": 605.0,
    "epsilon_norm": 4.4354777336120605,
    "prediction_norm": 3.957418203353882,
    "timestamp": "2025-07-14T12:33:34.855033"
  },
  {
    "epoch": 42,
    "batch_idx": 0,
    "loss": 0.15997254848480225,
    "learning_rate": 0.0009989122707287205,
    "timestep_mean": 549.03125,
    "epsilon_norm": 7.5722975730896,
    "prediction_norm": 7.290008544921875,
    "timestamp": "2025-07-14T12:33:34.859935"
  },
  {
    "epoch": 42,
    "batch_idx": 32,
    "loss": 0.39608731865882874,
    "learning_rate": 0.0009989122707287205,
    "timestep_mean": 520.5625,
    "epsilon_norm": 8.06595516204834,
    "prediction_norm": 6.545332431793213,
    "timestamp": "2025-07-14T12:33:34.864748"
  },
  {
    "epoch": 42,
    "batch_idx": 64,
    "loss": 0.3246022164821625,
    "learning_rate": 0.0009989122707287205,
    "timestep_mean": 482.5625,
    "epsilon_norm": 7.414881706237793,
    "prediction_norm": 6.50063943862915,
    "timestamp": "2025-07-14T12:33:34.869611"
  },
  {
    "epoch": 42,
    "batch_idx": 96,
    "loss": 0.30456221103668213,
    "learning_rate": 0.0009989122707287205,
    "timestep_mean": 527.3125,
    "epsilon_norm": 8.966263771057129,
    "prediction_norm": 7.093771457672119,
    "timestamp": "2025-07-14T12:33:34.874516"
  },
  {
    "epoch": 42,
    "batch_idx": 128,
    "loss": 0.2715238332748413,
    "learning_rate": 0.0009989122707287205,
    "timestep_mean": 569.5625,
    "epsilon_norm": 7.865637302398682,
    "prediction_norm": 6.175574779510498,
    "timestamp": "2025-07-14T12:33:34.879497"
  },
  {
    "epoch": 42,
    "batch_idx": 160,
    "loss": 0.20359131693840027,
    "learning_rate": 0.0009989122707287205,
    "timestep_mean": 547.0625,
    "epsilon_norm": 7.008275508880615,
    "prediction_norm": 6.0440778732299805,
    "timestamp": "2025-07-14T12:33:34.884293"
  },
  {
    "epoch": 42,
    "batch_idx": 192,
    "loss": 0.2798592150211334,
    "learning_rate": 0.0009989122707287205,
    "timestep_mean": 384.625,
    "epsilon_norm": 2.7326087951660156,
    "prediction_norm": 1.7627042531967163,
    "timestamp": "2025-07-14T12:33:34.889117"
  },
  {
    "epoch": 43,
    "batch_idx": 0,
    "loss": 0.5796380043029785,
    "learning_rate": 0.000998859877398245,
    "timestep_mean": 453.1875,
    "epsilon_norm": 8.2965087890625,
    "prediction_norm": 5.283029556274414,
    "timestamp": "2025-07-14T12:33:34.894019"
  },
  {
    "epoch": 43,
    "batch_idx": 32,
    "loss": 0.24915575981140137,
    "learning_rate": 0.000998859877398245,
    "timestep_mean": 537.9375,
    "epsilon_norm": 7.368948936462402,
    "prediction_norm": 5.974125862121582,
    "timestamp": "2025-07-14T12:33:34.898888"
  },
  {
    "epoch": 43,
    "batch_idx": 64,
    "loss": 0.38283079862594604,
    "learning_rate": 0.000998859877398245,
    "timestep_mean": 468.09375,
    "epsilon_norm": 8.308125495910645,
    "prediction_norm": 6.353785991668701,
    "timestamp": "2025-07-14T12:33:34.903706"
  },
  {
    "epoch": 43,
    "batch_idx": 96,
    "loss": 0.4986576437950134,
    "learning_rate": 0.000998859877398245,
    "timestep_mean": 485.3125,
    "epsilon_norm": 9.106317520141602,
    "prediction_norm": 5.996029376983643,
    "timestamp": "2025-07-14T12:33:34.908716"
  },
  {
    "epoch": 43,
    "batch_idx": 128,
    "loss": 0.4048210084438324,
    "learning_rate": 0.000998859877398245,
    "timestep_mean": 475.8125,
    "epsilon_norm": 7.016775608062744,
    "prediction_norm": 5.320266246795654,
    "timestamp": "2025-07-14T12:33:34.913549"
  },
  {
    "epoch": 43,
    "batch_idx": 160,
    "loss": 0.4424152672290802,
    "learning_rate": 0.000998859877398245,
    "timestep_mean": 427.4375,
    "epsilon_norm": 7.517586708068848,
    "prediction_norm": 5.568967819213867,
    "timestamp": "2025-07-14T12:33:34.918373"
  },
  {
    "epoch": 43,
    "batch_idx": 192,
    "loss": 0.4225770831108093,
    "learning_rate": 0.000998859877398245,
    "timestep_mean": 463.875,
    "epsilon_norm": 2.917868137359619,
    "prediction_norm": 2.077566385269165,
    "timestamp": "2025-07-14T12:33:34.923188"
  },
  {
    "epoch": 44,
    "batch_idx": 0,
    "loss": 0.3410232961177826,
    "learning_rate": 0.0009988062531806122,
    "timestep_mean": 423.9375,
    "epsilon_norm": 6.822185039520264,
    "prediction_norm": 4.929376602172852,
    "timestamp": "2025-07-14T12:33:34.928086"
  },
  {
    "epoch": 44,
    "batch_idx": 32,
    "loss": 0.3467409014701843,
    "learning_rate": 0.0009988062531806122,
    "timestep_mean": 601.34375,
    "epsilon_norm": 8.324427604675293,
    "prediction_norm": 6.518067359924316,
    "timestamp": "2025-07-14T12:33:34.933082"
  },
  {
    "epoch": 44,
    "batch_idx": 64,
    "loss": 0.39187467098236084,
    "learning_rate": 0.0009988062531806122,
    "timestep_mean": 564.875,
    "epsilon_norm": 7.604124546051025,
    "prediction_norm": 6.149160385131836,
    "timestamp": "2025-07-14T12:33:34.938018"
  },
  {
    "epoch": 44,
    "batch_idx": 96,
    "loss": 0.29530325531959534,
    "learning_rate": 0.0009988062531806122,
    "timestep_mean": 486.84375,
    "epsilon_norm": 7.2873406410217285,
    "prediction_norm": 6.162162780761719,
    "timestamp": "2025-07-14T12:33:34.942906"
  },
  {
    "epoch": 44,
    "batch_idx": 128,
    "loss": 0.3326469659805298,
    "learning_rate": 0.0009988062531806122,
    "timestep_mean": 564.5625,
    "epsilon_norm": 8.101387977600098,
    "prediction_norm": 6.551890850067139,
    "timestamp": "2025-07-14T12:33:34.947790"
  },
  {
    "epoch": 44,
    "batch_idx": 160,
    "loss": 0.27917447686195374,
    "learning_rate": 0.0009988062531806122,
    "timestep_mean": 541.1875,
    "epsilon_norm": 8.21895980834961,
    "prediction_norm": 7.156492233276367,
    "timestamp": "2025-07-14T12:33:34.952609"
  },
  {
    "epoch": 44,
    "batch_idx": 192,
    "loss": 0.729499101638794,
    "learning_rate": 0.0009988062531806122,
    "timestep_mean": 481.75,
    "epsilon_norm": 4.487497329711914,
    "prediction_norm": 3.1316120624542236,
    "timestamp": "2025-07-14T12:33:34.957458"
  },
  {
    "epoch": 45,
    "batch_idx": 0,
    "loss": 0.41343003511428833,
    "learning_rate": 0.0009987513982081346,
    "timestep_mean": 393.03125,
    "epsilon_norm": 7.014330863952637,
    "prediction_norm": 5.2940826416015625,
    "timestamp": "2025-07-14T12:33:34.962352"
  },
  {
    "epoch": 45,
    "batch_idx": 32,
    "loss": 0.325890451669693,
    "learning_rate": 0.0009987513982081346,
    "timestep_mean": 523.15625,
    "epsilon_norm": 8.228034973144531,
    "prediction_norm": 6.679928302764893,
    "timestamp": "2025-07-14T12:33:34.967206"
  },
  {
    "epoch": 45,
    "batch_idx": 64,
    "loss": 0.2210569530725479,
    "learning_rate": 0.0009987513982081346,
    "timestep_mean": 571.0,
    "epsilon_norm": 9.06646728515625,
    "prediction_norm": 8.303709030151367,
    "timestamp": "2025-07-14T12:33:34.972039"
  },
  {
    "epoch": 45,
    "batch_idx": 96,
    "loss": 0.4893958270549774,
    "learning_rate": 0.0009987513982081346,
    "timestep_mean": 494.0,
    "epsilon_norm": 8.831473350524902,
    "prediction_norm": 5.981587886810303,
    "timestamp": "2025-07-14T12:33:34.976883"
  },
  {
    "epoch": 45,
    "batch_idx": 128,
    "loss": 0.2804880142211914,
    "learning_rate": 0.0009987513982081346,
    "timestep_mean": 459.4375,
    "epsilon_norm": 7.50545597076416,
    "prediction_norm": 6.126575946807861,
    "timestamp": "2025-07-14T12:33:34.981728"
  },
  {
    "epoch": 45,
    "batch_idx": 160,
    "loss": 0.41151130199432373,
    "learning_rate": 0.0009987513982081346,
    "timestep_mean": 547.28125,
    "epsilon_norm": 8.221288681030273,
    "prediction_norm": 6.945974349975586,
    "timestamp": "2025-07-14T12:33:34.986601"
  },
  {
    "epoch": 45,
    "batch_idx": 192,
    "loss": 0.19044256210327148,
    "learning_rate": 0.0009987513982081346,
    "timestep_mean": 511.875,
    "epsilon_norm": 3.7883379459381104,
    "prediction_norm": 3.671117067337036,
    "timestamp": "2025-07-14T12:33:34.991558"
  },
  {
    "epoch": 46,
    "batch_idx": 0,
    "loss": 0.29227322340011597,
    "learning_rate": 0.0009986953126161614,
    "timestep_mean": 455.125,
    "epsilon_norm": 7.516045570373535,
    "prediction_norm": 6.554896354675293,
    "timestamp": "2025-07-14T12:33:34.996372"
  },
  {
    "epoch": 46,
    "batch_idx": 32,
    "loss": 0.17260964214801788,
    "learning_rate": 0.0009986953126161614,
    "timestep_mean": 567.875,
    "epsilon_norm": 8.379145622253418,
    "prediction_norm": 7.5550360679626465,
    "timestamp": "2025-07-14T12:33:35.001158"
  },
  {
    "epoch": 46,
    "batch_idx": 64,
    "loss": 0.36643826961517334,
    "learning_rate": 0.0009986953126161614,
    "timestep_mean": 490.125,
    "epsilon_norm": 8.499449729919434,
    "prediction_norm": 7.241464138031006,
    "timestamp": "2025-07-14T12:33:35.005981"
  },
  {
    "epoch": 46,
    "batch_idx": 96,
    "loss": 0.2304174154996872,
    "learning_rate": 0.0009986953126161614,
    "timestep_mean": 611.0,
    "epsilon_norm": 8.953933715820312,
    "prediction_norm": 7.77431058883667,
    "timestamp": "2025-07-14T12:33:35.010901"
  },
  {
    "epoch": 46,
    "batch_idx": 128,
    "loss": 0.43675369024276733,
    "learning_rate": 0.0009986953126161614,
    "timestep_mean": 495.875,
    "epsilon_norm": 8.311171531677246,
    "prediction_norm": 6.803277015686035,
    "timestamp": "2025-07-14T12:33:35.015776"
  },
  {
    "epoch": 46,
    "batch_idx": 160,
    "loss": 0.5818874835968018,
    "learning_rate": 0.0009986953126161614,
    "timestep_mean": 384.28125,
    "epsilon_norm": 8.97548770904541,
    "prediction_norm": 5.408576965332031,
    "timestamp": "2025-07-14T12:33:35.020636"
  },
  {
    "epoch": 46,
    "batch_idx": 192,
    "loss": 0.21470755338668823,
    "learning_rate": 0.0009986953126161614,
    "timestep_mean": 427.875,
    "epsilon_norm": 2.656968832015991,
    "prediction_norm": 2.5434885025024414,
    "timestamp": "2025-07-14T12:33:35.025534"
  },
  {
    "epoch": 47,
    "batch_idx": 0,
    "loss": 0.3570939302444458,
    "learning_rate": 0.000998637996543078,
    "timestep_mean": 527.375,
    "epsilon_norm": 7.132916450500488,
    "prediction_norm": 6.0844645500183105,
    "timestamp": "2025-07-14T12:33:35.030483"
  },
  {
    "epoch": 47,
    "batch_idx": 32,
    "loss": 0.27322763204574585,
    "learning_rate": 0.000998637996543078,
    "timestep_mean": 503.25,
    "epsilon_norm": 6.8161187171936035,
    "prediction_norm": 6.402465343475342,
    "timestamp": "2025-07-14T12:33:35.035400"
  },
  {
    "epoch": 47,
    "batch_idx": 64,
    "loss": 0.3237885534763336,
    "learning_rate": 0.000998637996543078,
    "timestep_mean": 438.8125,
    "epsilon_norm": 7.097697734832764,
    "prediction_norm": 4.864665508270264,
    "timestamp": "2025-07-14T12:33:35.040238"
  },
  {
    "epoch": 47,
    "batch_idx": 96,
    "loss": 0.6140618324279785,
    "learning_rate": 0.000998637996543078,
    "timestep_mean": 508.0625,
    "epsilon_norm": 9.08145523071289,
    "prediction_norm": 6.558746337890625,
    "timestamp": "2025-07-14T12:33:35.045072"
  },
  {
    "epoch": 47,
    "batch_idx": 128,
    "loss": 0.354031503200531,
    "learning_rate": 0.000998637996543078,
    "timestep_mean": 466.1875,
    "epsilon_norm": 8.793930053710938,
    "prediction_norm": 7.124752521514893,
    "timestamp": "2025-07-14T12:33:35.050006"
  },
  {
    "epoch": 47,
    "batch_idx": 160,
    "loss": 0.3738112151622772,
    "learning_rate": 0.000998637996543078,
    "timestep_mean": 440.5625,
    "epsilon_norm": 7.917984962463379,
    "prediction_norm": 5.7040557861328125,
    "timestamp": "2025-07-14T12:33:35.054882"
  },
  {
    "epoch": 47,
    "batch_idx": 192,
    "loss": 0.0645100325345993,
    "learning_rate": 0.000998637996543078,
    "timestep_mean": 696.5,
    "epsilon_norm": 4.803517818450928,
    "prediction_norm": 4.237303256988525,
    "timestamp": "2025-07-14T12:33:35.059764"
  },
  {
    "epoch": 48,
    "batch_idx": 0,
    "loss": 0.45813634991645813,
    "learning_rate": 0.0009985794501303064,
    "timestep_mean": 423.90625,
    "epsilon_norm": 8.30678653717041,
    "prediction_norm": 5.785134315490723,
    "timestamp": "2025-07-14T12:33:35.064651"
  },
  {
    "epoch": 48,
    "batch_idx": 32,
    "loss": 0.3017937242984772,
    "learning_rate": 0.0009985794501303064,
    "timestep_mean": 559.375,
    "epsilon_norm": 7.677806377410889,
    "prediction_norm": 6.664174556732178,
    "timestamp": "2025-07-14T12:33:35.069579"
  },
  {
    "epoch": 48,
    "batch_idx": 64,
    "loss": 0.3378027081489563,
    "learning_rate": 0.0009985794501303064,
    "timestep_mean": 548.71875,
    "epsilon_norm": 7.183210372924805,
    "prediction_norm": 5.271036148071289,
    "timestamp": "2025-07-14T12:33:35.074447"
  },
  {
    "epoch": 48,
    "batch_idx": 96,
    "loss": 0.5508302450180054,
    "learning_rate": 0.0009985794501303064,
    "timestep_mean": 441.8125,
    "epsilon_norm": 8.509601593017578,
    "prediction_norm": 6.017840385437012,
    "timestamp": "2025-07-14T12:33:35.079327"
  },
  {
    "epoch": 48,
    "batch_idx": 128,
    "loss": 0.29630035161972046,
    "learning_rate": 0.0009985794501303064,
    "timestep_mean": 480.875,
    "epsilon_norm": 6.978164196014404,
    "prediction_norm": 5.187042236328125,
    "timestamp": "2025-07-14T12:33:35.084192"
  },
  {
    "epoch": 48,
    "batch_idx": 160,
    "loss": 0.4165631830692291,
    "learning_rate": 0.0009985794501303064,
    "timestep_mean": 454.96875,
    "epsilon_norm": 7.706240653991699,
    "prediction_norm": 5.922509670257568,
    "timestamp": "2025-07-14T12:33:35.089118"
  },
  {
    "epoch": 48,
    "batch_idx": 192,
    "loss": 0.44254547357559204,
    "learning_rate": 0.0009985794501303064,
    "timestep_mean": 416.0,
    "epsilon_norm": 4.352702617645264,
    "prediction_norm": 2.882211685180664,
    "timestamp": "2025-07-14T12:33:35.093965"
  },
  {
    "epoch": 49,
    "batch_idx": 0,
    "loss": 0.36447346210479736,
    "learning_rate": 0.000998519673522304,
    "timestep_mean": 483.84375,
    "epsilon_norm": 6.61187744140625,
    "prediction_norm": 5.32600736618042,
    "timestamp": "2025-07-14T12:33:35.098839"
  },
  {
    "epoch": 49,
    "batch_idx": 32,
    "loss": 0.3704289197921753,
    "learning_rate": 0.000998519673522304,
    "timestep_mean": 545.84375,
    "epsilon_norm": 8.580754280090332,
    "prediction_norm": 7.0326619148254395,
    "timestamp": "2025-07-14T12:33:35.109424"
  },
  {
    "epoch": 49,
    "batch_idx": 64,
    "loss": 0.4403783977031708,
    "learning_rate": 0.000998519673522304,
    "timestep_mean": 504.0625,
    "epsilon_norm": 8.599863052368164,
    "prediction_norm": 6.856172561645508,
    "timestamp": "2025-07-14T12:33:35.114176"
  },
  {
    "epoch": 49,
    "batch_idx": 96,
    "loss": 0.4633917808532715,
    "learning_rate": 0.000998519673522304,
    "timestep_mean": 473.65625,
    "epsilon_norm": 7.7190752029418945,
    "prediction_norm": 5.5728840827941895,
    "timestamp": "2025-07-14T12:33:35.118915"
  },
  {
    "epoch": 49,
    "batch_idx": 128,
    "loss": 0.5261250138282776,
    "learning_rate": 0.000998519673522304,
    "timestep_mean": 470.5625,
    "epsilon_norm": 8.753304481506348,
    "prediction_norm": 6.882094860076904,
    "timestamp": "2025-07-14T12:33:35.123637"
  },
  {
    "epoch": 49,
    "batch_idx": 160,
    "loss": 0.65000319480896,
    "learning_rate": 0.000998519673522304,
    "timestep_mean": 432.6875,
    "epsilon_norm": 8.054677963256836,
    "prediction_norm": 5.286375999450684,
    "timestamp": "2025-07-14T12:33:35.128377"
  },
  {
    "epoch": 49,
    "batch_idx": 192,
    "loss": 0.16262230277061462,
    "learning_rate": 0.000998519673522304,
    "timestep_mean": 499.125,
    "epsilon_norm": 3.7755537033081055,
    "prediction_norm": 3.483389377593994,
    "timestamp": "2025-07-14T12:33:35.133115"
  },
  {
    "epoch": 50,
    "batch_idx": 0,
    "loss": 0.16187652945518494,
    "learning_rate": 0.0009984586668665635,
    "timestep_mean": 626.25,
    "epsilon_norm": 8.157458305358887,
    "prediction_norm": 7.972425937652588,
    "timestamp": "2025-07-14T12:33:35.137823"
  },
  {
    "epoch": 50,
    "batch_idx": 32,
    "loss": 0.38967952132225037,
    "learning_rate": 0.0009984586668665635,
    "timestep_mean": 492.1875,
    "epsilon_norm": 7.2952656745910645,
    "prediction_norm": 6.417675495147705,
    "timestamp": "2025-07-14T12:33:35.142525"
  },
  {
    "epoch": 50,
    "batch_idx": 64,
    "loss": 0.27421116828918457,
    "learning_rate": 0.0009984586668665635,
    "timestep_mean": 591.6875,
    "epsilon_norm": 7.774867057800293,
    "prediction_norm": 7.007035732269287,
    "timestamp": "2025-07-14T12:33:35.147264"
  },
  {
    "epoch": 50,
    "batch_idx": 96,
    "loss": 0.2890916168689728,
    "learning_rate": 0.0009984586668665635,
    "timestep_mean": 519.34375,
    "epsilon_norm": 8.616049766540527,
    "prediction_norm": 6.9786272048950195,
    "timestamp": "2025-07-14T12:33:35.152005"
  },
  {
    "epoch": 50,
    "batch_idx": 128,
    "loss": 0.41743651032447815,
    "learning_rate": 0.0009984586668665635,
    "timestep_mean": 438.65625,
    "epsilon_norm": 7.293202877044678,
    "prediction_norm": 5.221107006072998,
    "timestamp": "2025-07-14T12:33:35.156720"
  },
  {
    "epoch": 50,
    "batch_idx": 160,
    "loss": 0.4527297914028168,
    "learning_rate": 0.0009984586668665635,
    "timestep_mean": 419.71875,
    "epsilon_norm": 7.757376194000244,
    "prediction_norm": 5.11814022064209,
    "timestamp": "2025-07-14T12:33:35.161557"
  },
  {
    "epoch": 50,
    "batch_idx": 192,
    "loss": 0.7218188643455505,
    "learning_rate": 0.0009984586668665635,
    "timestep_mean": 440.875,
    "epsilon_norm": 5.801154136657715,
    "prediction_norm": 4.29258394241333,
    "timestamp": "2025-07-14T12:33:35.166250"
  },
  {
    "epoch": 51,
    "batch_idx": 0,
    "loss": 0.38502705097198486,
    "learning_rate": 0.0009983964303136128,
    "timestep_mean": 408.46875,
    "epsilon_norm": 7.087558269500732,
    "prediction_norm": 5.396631240844727,
    "timestamp": "2025-07-14T12:33:35.171137"
  },
  {
    "epoch": 51,
    "batch_idx": 32,
    "loss": 0.3511538505554199,
    "learning_rate": 0.0009983964303136128,
    "timestep_mean": 510.09375,
    "epsilon_norm": 7.721251487731934,
    "prediction_norm": 7.045361042022705,
    "timestamp": "2025-07-14T12:33:35.175874"
  },
  {
    "epoch": 51,
    "batch_idx": 64,
    "loss": 0.4111039936542511,
    "learning_rate": 0.0009983964303136128,
    "timestep_mean": 523.34375,
    "epsilon_norm": 8.452075004577637,
    "prediction_norm": 6.305177688598633,
    "timestamp": "2025-07-14T12:33:35.180635"
  },
  {
    "epoch": 51,
    "batch_idx": 96,
    "loss": 0.375790536403656,
    "learning_rate": 0.0009983964303136128,
    "timestep_mean": 516.5625,
    "epsilon_norm": 7.147587299346924,
    "prediction_norm": 5.765370845794678,
    "timestamp": "2025-07-14T12:33:35.185448"
  },
  {
    "epoch": 51,
    "batch_idx": 128,
    "loss": 0.4806368052959442,
    "learning_rate": 0.0009983964303136128,
    "timestep_mean": 491.75,
    "epsilon_norm": 7.266295433044434,
    "prediction_norm": 4.981171607971191,
    "timestamp": "2025-07-14T12:33:35.190184"
  },
  {
    "epoch": 51,
    "batch_idx": 160,
    "loss": 0.2327648550271988,
    "learning_rate": 0.0009983964303136128,
    "timestep_mean": 550.75,
    "epsilon_norm": 8.727380752563477,
    "prediction_norm": 7.714968204498291,
    "timestamp": "2025-07-14T12:33:35.195044"
  },
  {
    "epoch": 51,
    "batch_idx": 192,
    "loss": 0.111151322722435,
    "learning_rate": 0.0009983964303136128,
    "timestep_mean": 513.375,
    "epsilon_norm": 2.286357879638672,
    "prediction_norm": 2.7578842639923096,
    "timestamp": "2025-07-14T12:33:35.199868"
  },
  {
    "epoch": 52,
    "batch_idx": 0,
    "loss": 0.2924414277076721,
    "learning_rate": 0.0009983329640170144,
    "timestep_mean": 576.75,
    "epsilon_norm": 8.549287796020508,
    "prediction_norm": 7.025096893310547,
    "timestamp": "2025-07-14T12:33:35.204627"
  },
  {
    "epoch": 52,
    "batch_idx": 32,
    "loss": 0.3128913640975952,
    "learning_rate": 0.0009983329640170144,
    "timestep_mean": 519.78125,
    "epsilon_norm": 7.686305046081543,
    "prediction_norm": 6.164067268371582,
    "timestamp": "2025-07-14T12:33:35.209351"
  },
  {
    "epoch": 52,
    "batch_idx": 64,
    "loss": 0.5618785619735718,
    "learning_rate": 0.0009983329640170144,
    "timestep_mean": 445.71875,
    "epsilon_norm": 8.167203903198242,
    "prediction_norm": 5.9178900718688965,
    "timestamp": "2025-07-14T12:33:35.214114"
  },
  {
    "epoch": 52,
    "batch_idx": 96,
    "loss": 0.63541579246521,
    "learning_rate": 0.0009983329640170144,
    "timestep_mean": 401.71875,
    "epsilon_norm": 9.299251556396484,
    "prediction_norm": 5.595658302307129,
    "timestamp": "2025-07-14T12:33:35.218878"
  },
  {
    "epoch": 52,
    "batch_idx": 128,
    "loss": 0.43450823426246643,
    "learning_rate": 0.0009983329640170144,
    "timestep_mean": 464.0,
    "epsilon_norm": 8.389745712280273,
    "prediction_norm": 5.615866184234619,
    "timestamp": "2025-07-14T12:33:35.223626"
  },
  {
    "epoch": 52,
    "batch_idx": 160,
    "loss": 0.21588844060897827,
    "learning_rate": 0.0009983329640170144,
    "timestep_mean": 509.1875,
    "epsilon_norm": 7.647216320037842,
    "prediction_norm": 6.717273235321045,
    "timestamp": "2025-07-14T12:33:35.228565"
  },
  {
    "epoch": 52,
    "batch_idx": 192,
    "loss": 0.32762411236763,
    "learning_rate": 0.0009983329640170144,
    "timestep_mean": 581.25,
    "epsilon_norm": 2.975687265396118,
    "prediction_norm": 2.586022138595581,
    "timestamp": "2025-07-14T12:33:35.233415"
  },
  {
    "epoch": 53,
    "batch_idx": 0,
    "loss": 0.530953049659729,
    "learning_rate": 0.0009982682681333652,
    "timestep_mean": 414.96875,
    "epsilon_norm": 8.063356399536133,
    "prediction_norm": 5.575169563293457,
    "timestamp": "2025-07-14T12:33:35.238246"
  },
  {
    "epoch": 53,
    "batch_idx": 32,
    "loss": 0.4662478566169739,
    "learning_rate": 0.0009982682681333652,
    "timestep_mean": 510.0,
    "epsilon_norm": 7.482006072998047,
    "prediction_norm": 5.445042610168457,
    "timestamp": "2025-07-14T12:33:35.243193"
  },
  {
    "epoch": 53,
    "batch_idx": 64,
    "loss": 0.46595776081085205,
    "learning_rate": 0.0009982682681333652,
    "timestep_mean": 476.90625,
    "epsilon_norm": 7.9071736335754395,
    "prediction_norm": 5.662255764007568,
    "timestamp": "2025-07-14T12:33:35.248049"
  },
  {
    "epoch": 53,
    "batch_idx": 96,
    "loss": 0.41557973623275757,
    "learning_rate": 0.0009982682681333652,
    "timestep_mean": 516.5,
    "epsilon_norm": 8.003863334655762,
    "prediction_norm": 6.4303297996521,
    "timestamp": "2025-07-14T12:33:35.253111"
  },
  {
    "epoch": 53,
    "batch_idx": 128,
    "loss": 0.3159761130809784,
    "learning_rate": 0.0009982682681333652,
    "timestep_mean": 591.21875,
    "epsilon_norm": 8.329804420471191,
    "prediction_norm": 7.2192463874816895,
    "timestamp": "2025-07-14T12:33:35.258129"
  },
  {
    "epoch": 53,
    "batch_idx": 160,
    "loss": 0.33171191811561584,
    "learning_rate": 0.0009982682681333652,
    "timestep_mean": 520.6875,
    "epsilon_norm": 8.192403793334961,
    "prediction_norm": 6.256832122802734,
    "timestamp": "2025-07-14T12:33:35.263068"
  },
  {
    "epoch": 53,
    "batch_idx": 192,
    "loss": 0.12219169735908508,
    "learning_rate": 0.0009982682681333652,
    "timestep_mean": 632.75,
    "epsilon_norm": 3.953547239303589,
    "prediction_norm": 3.457256317138672,
    "timestamp": "2025-07-14T12:33:35.268189"
  },
  {
    "epoch": 54,
    "batch_idx": 0,
    "loss": 0.4713066518306732,
    "learning_rate": 0.0009982023428222956,
    "timestep_mean": 534.46875,
    "epsilon_norm": 8.265470504760742,
    "prediction_norm": 6.0624589920043945,
    "timestamp": "2025-07-14T12:33:35.273267"
  },
  {
    "epoch": 54,
    "batch_idx": 32,
    "loss": 0.3271346688270569,
    "learning_rate": 0.0009982023428222956,
    "timestep_mean": 493.15625,
    "epsilon_norm": 6.95628547668457,
    "prediction_norm": 5.595195770263672,
    "timestamp": "2025-07-14T12:33:35.278455"
  },
  {
    "epoch": 54,
    "batch_idx": 64,
    "loss": 0.42034468054771423,
    "learning_rate": 0.0009982023428222956,
    "timestep_mean": 459.34375,
    "epsilon_norm": 8.974395751953125,
    "prediction_norm": 6.799103260040283,
    "timestamp": "2025-07-14T12:33:35.283753"
  },
  {
    "epoch": 54,
    "batch_idx": 96,
    "loss": 0.3476872444152832,
    "learning_rate": 0.0009982023428222956,
    "timestep_mean": 489.84375,
    "epsilon_norm": 7.416536331176758,
    "prediction_norm": 5.278743743896484,
    "timestamp": "2025-07-14T12:33:35.288816"
  },
  {
    "epoch": 54,
    "batch_idx": 128,
    "loss": 0.48824357986450195,
    "learning_rate": 0.0009982023428222956,
    "timestep_mean": 443.71875,
    "epsilon_norm": 8.485635757446289,
    "prediction_norm": 5.859688758850098,
    "timestamp": "2025-07-14T12:33:35.293939"
  },
  {
    "epoch": 54,
    "batch_idx": 160,
    "loss": 0.24828487634658813,
    "learning_rate": 0.0009982023428222956,
    "timestep_mean": 472.65625,
    "epsilon_norm": 6.263227462768555,
    "prediction_norm": 5.539325714111328,
    "timestamp": "2025-07-14T12:33:35.299061"
  },
  {
    "epoch": 54,
    "batch_idx": 192,
    "loss": 1.0341088771820068,
    "learning_rate": 0.0009982023428222956,
    "timestep_mean": 399.375,
    "epsilon_norm": 4.452439308166504,
    "prediction_norm": 2.18324875831604,
    "timestamp": "2025-07-14T12:33:35.304144"
  },
  {
    "epoch": 55,
    "batch_idx": 0,
    "loss": 0.35778796672821045,
    "learning_rate": 0.00099813518824647,
    "timestep_mean": 492.875,
    "epsilon_norm": 8.428531646728516,
    "prediction_norm": 7.478564262390137,
    "timestamp": "2025-07-14T12:33:35.309296"
  },
  {
    "epoch": 55,
    "batch_idx": 32,
    "loss": 0.25565317273139954,
    "learning_rate": 0.00099813518824647,
    "timestep_mean": 571.0,
    "epsilon_norm": 7.067202568054199,
    "prediction_norm": 6.7952985763549805,
    "timestamp": "2025-07-14T12:33:35.314311"
  },
  {
    "epoch": 55,
    "batch_idx": 64,
    "loss": 0.45107460021972656,
    "learning_rate": 0.00099813518824647,
    "timestep_mean": 596.15625,
    "epsilon_norm": 8.647636413574219,
    "prediction_norm": 6.907862663269043,
    "timestamp": "2025-07-14T12:33:35.319342"
  },
  {
    "epoch": 55,
    "batch_idx": 96,
    "loss": 0.33140426874160767,
    "learning_rate": 0.00099813518824647,
    "timestep_mean": 540.96875,
    "epsilon_norm": 8.180109024047852,
    "prediction_norm": 6.59371280670166,
    "timestamp": "2025-07-14T12:33:35.324401"
  },
  {
    "epoch": 55,
    "batch_idx": 128,
    "loss": 0.39689141511917114,
    "learning_rate": 0.00099813518824647,
    "timestep_mean": 532.5625,
    "epsilon_norm": 7.34657096862793,
    "prediction_norm": 5.4567060470581055,
    "timestamp": "2025-07-14T12:33:35.329494"
  },
  {
    "epoch": 55,
    "batch_idx": 160,
    "loss": 0.299556165933609,
    "learning_rate": 0.00099813518824647,
    "timestep_mean": 490.25,
    "epsilon_norm": 7.802881240844727,
    "prediction_norm": 6.000542163848877,
    "timestamp": "2025-07-14T12:33:35.334553"
  },
  {
    "epoch": 55,
    "batch_idx": 192,
    "loss": 0.5031599402427673,
    "learning_rate": 0.00099813518824647,
    "timestep_mean": 374.5,
    "epsilon_norm": 4.601601600646973,
    "prediction_norm": 3.4892337322235107,
    "timestamp": "2025-07-14T12:33:35.339438"
  },
  {
    "epoch": 56,
    "batch_idx": 0,
    "loss": 0.41727307438850403,
    "learning_rate": 0.0009980668045715857,
    "timestep_mean": 479.8125,
    "epsilon_norm": 9.630297660827637,
    "prediction_norm": 7.930419921875,
    "timestamp": "2025-07-14T12:33:35.344241"
  },
  {
    "epoch": 56,
    "batch_idx": 32,
    "loss": 0.264198362827301,
    "learning_rate": 0.0009980668045715857,
    "timestep_mean": 506.6875,
    "epsilon_norm": 8.148259162902832,
    "prediction_norm": 7.065402030944824,
    "timestamp": "2025-07-14T12:33:35.349065"
  },
  {
    "epoch": 56,
    "batch_idx": 64,
    "loss": 0.28045228123664856,
    "learning_rate": 0.0009980668045715857,
    "timestep_mean": 498.28125,
    "epsilon_norm": 7.514228343963623,
    "prediction_norm": 7.006279945373535,
    "timestamp": "2025-07-14T12:33:35.353899"
  },
  {
    "epoch": 56,
    "batch_idx": 96,
    "loss": 0.279470831155777,
    "learning_rate": 0.0009980668045715857,
    "timestep_mean": 524.9375,
    "epsilon_norm": 6.772041320800781,
    "prediction_norm": 5.883260726928711,
    "timestamp": "2025-07-14T12:33:35.358881"
  },
  {
    "epoch": 56,
    "batch_idx": 128,
    "loss": 0.38866791129112244,
    "learning_rate": 0.0009980668045715857,
    "timestep_mean": 457.46875,
    "epsilon_norm": 6.104854583740234,
    "prediction_norm": 4.912173748016357,
    "timestamp": "2025-07-14T12:33:35.363846"
  },
  {
    "epoch": 56,
    "batch_idx": 160,
    "loss": 0.3318464457988739,
    "learning_rate": 0.0009980668045715857,
    "timestep_mean": 492.8125,
    "epsilon_norm": 6.646341323852539,
    "prediction_norm": 4.792591571807861,
    "timestamp": "2025-07-14T12:33:35.368936"
  },
  {
    "epoch": 56,
    "batch_idx": 192,
    "loss": 0.2739667594432831,
    "learning_rate": 0.0009980668045715857,
    "timestep_mean": 405.375,
    "epsilon_norm": 3.6234781742095947,
    "prediction_norm": 2.776963949203491,
    "timestamp": "2025-07-14T12:33:35.373908"
  },
  {
    "epoch": 57,
    "batch_idx": 0,
    "loss": 0.3015681505203247,
    "learning_rate": 0.0009979971919663725,
    "timestep_mean": 494.34375,
    "epsilon_norm": 6.953224182128906,
    "prediction_norm": 4.882697582244873,
    "timestamp": "2025-07-14T12:33:35.379068"
  },
  {
    "epoch": 57,
    "batch_idx": 32,
    "loss": 0.6396154761314392,
    "learning_rate": 0.0009979971919663725,
    "timestep_mean": 442.15625,
    "epsilon_norm": 8.521398544311523,
    "prediction_norm": 5.320400238037109,
    "timestamp": "2025-07-14T12:33:35.384151"
  },
  {
    "epoch": 57,
    "batch_idx": 64,
    "loss": 0.40295982360839844,
    "learning_rate": 0.0009979971919663725,
    "timestep_mean": 533.375,
    "epsilon_norm": 7.647167205810547,
    "prediction_norm": 6.066166877746582,
    "timestamp": "2025-07-14T12:33:35.389246"
  },
  {
    "epoch": 57,
    "batch_idx": 96,
    "loss": 0.3396454453468323,
    "learning_rate": 0.0009979971919663725,
    "timestep_mean": 476.875,
    "epsilon_norm": 7.47354793548584,
    "prediction_norm": 5.837466716766357,
    "timestamp": "2025-07-14T12:33:35.394379"
  },
  {
    "epoch": 57,
    "batch_idx": 128,
    "loss": 0.308167964220047,
    "learning_rate": 0.0009979971919663725,
    "timestep_mean": 505.53125,
    "epsilon_norm": 7.672144412994385,
    "prediction_norm": 6.121908187866211,
    "timestamp": "2025-07-14T12:33:35.399491"
  },
  {
    "epoch": 57,
    "batch_idx": 160,
    "loss": 0.32378697395324707,
    "learning_rate": 0.0009979971919663725,
    "timestep_mean": 505.90625,
    "epsilon_norm": 8.04190731048584,
    "prediction_norm": 6.8774003982543945,
    "timestamp": "2025-07-14T12:33:35.404737"
  },
  {
    "epoch": 57,
    "batch_idx": 192,
    "loss": 0.47211021184921265,
    "learning_rate": 0.0009979971919663725,
    "timestep_mean": 534.0,
    "epsilon_norm": 4.177207946777344,
    "prediction_norm": 3.233280658721924,
    "timestamp": "2025-07-14T12:33:35.409558"
  },
  {
    "epoch": 58,
    "batch_idx": 0,
    "loss": 0.3160935640335083,
    "learning_rate": 0.0009979263506025924,
    "timestep_mean": 532.90625,
    "epsilon_norm": 7.619232177734375,
    "prediction_norm": 5.960778713226318,
    "timestamp": "2025-07-14T12:33:35.414477"
  },
  {
    "epoch": 58,
    "batch_idx": 32,
    "loss": 0.46481072902679443,
    "learning_rate": 0.0009979263506025924,
    "timestep_mean": 477.4375,
    "epsilon_norm": 8.765752792358398,
    "prediction_norm": 6.843213081359863,
    "timestamp": "2025-07-14T12:33:35.419812"
  },
  {
    "epoch": 58,
    "batch_idx": 64,
    "loss": 0.55511474609375,
    "learning_rate": 0.0009979263506025924,
    "timestep_mean": 446.65625,
    "epsilon_norm": 8.715107917785645,
    "prediction_norm": 5.510277271270752,
    "timestamp": "2025-07-14T12:33:35.424963"
  },
  {
    "epoch": 58,
    "batch_idx": 96,
    "loss": 0.4560585618019104,
    "learning_rate": 0.0009979263506025924,
    "timestep_mean": 488.40625,
    "epsilon_norm": 7.514154434204102,
    "prediction_norm": 5.156955242156982,
    "timestamp": "2025-07-14T12:33:35.430028"
  },
  {
    "epoch": 58,
    "batch_idx": 128,
    "loss": 0.38619786500930786,
    "learning_rate": 0.0009979263506025924,
    "timestep_mean": 573.75,
    "epsilon_norm": 8.217698097229004,
    "prediction_norm": 7.18221378326416,
    "timestamp": "2025-07-14T12:33:35.435174"
  },
  {
    "epoch": 58,
    "batch_idx": 160,
    "loss": 0.30626416206359863,
    "learning_rate": 0.0009979263506025924,
    "timestep_mean": 510.46875,
    "epsilon_norm": 7.651997089385986,
    "prediction_norm": 6.156233787536621,
    "timestamp": "2025-07-14T12:33:35.440327"
  },
  {
    "epoch": 58,
    "batch_idx": 192,
    "loss": 0.7076389789581299,
    "learning_rate": 0.0009979263506025924,
    "timestep_mean": 523.0,
    "epsilon_norm": 4.954768657684326,
    "prediction_norm": 3.418931484222412,
    "timestamp": "2025-07-14T12:33:35.445556"
  },
  {
    "epoch": 59,
    "batch_idx": 0,
    "loss": 0.4316561818122864,
    "learning_rate": 0.0009978542806550396,
    "timestep_mean": 520.75,
    "epsilon_norm": 8.722533226013184,
    "prediction_norm": 7.243459224700928,
    "timestamp": "2025-07-14T12:33:35.450811"
  },
  {
    "epoch": 59,
    "batch_idx": 32,
    "loss": 0.5296357870101929,
    "learning_rate": 0.0009978542806550396,
    "timestep_mean": 427.25,
    "epsilon_norm": 8.09256362915039,
    "prediction_norm": 5.831114292144775,
    "timestamp": "2025-07-14T12:33:35.455982"
  },
  {
    "epoch": 59,
    "batch_idx": 64,
    "loss": 0.2032133787870407,
    "learning_rate": 0.0009978542806550396,
    "timestep_mean": 658.03125,
    "epsilon_norm": 8.407334327697754,
    "prediction_norm": 7.296337604522705,
    "timestamp": "2025-07-14T12:33:35.461155"
  },
  {
    "epoch": 59,
    "batch_idx": 96,
    "loss": 0.22784873843193054,
    "learning_rate": 0.0009978542806550396,
    "timestep_mean": 503.3125,
    "epsilon_norm": 6.449656009674072,
    "prediction_norm": 5.8944292068481445,
    "timestamp": "2025-07-14T12:33:35.466477"
  },
  {
    "epoch": 59,
    "batch_idx": 128,
    "loss": 0.3354717195034027,
    "learning_rate": 0.0009978542806550396,
    "timestep_mean": 499.625,
    "epsilon_norm": 6.607952117919922,
    "prediction_norm": 5.963024139404297,
    "timestamp": "2025-07-14T12:33:35.471588"
  },
  {
    "epoch": 59,
    "batch_idx": 160,
    "loss": 0.5997485518455505,
    "learning_rate": 0.0009978542806550396,
    "timestep_mean": 460.125,
    "epsilon_norm": 8.962057113647461,
    "prediction_norm": 6.716922283172607,
    "timestamp": "2025-07-14T12:33:35.476779"
  },
  {
    "epoch": 59,
    "batch_idx": 192,
    "loss": 0.28064650297164917,
    "learning_rate": 0.0009978542806550396,
    "timestep_mean": 450.125,
    "epsilon_norm": 3.388765335083008,
    "prediction_norm": 2.986051082611084,
    "timestamp": "2025-07-14T12:33:35.481902"
  },
  {
    "epoch": 60,
    "batch_idx": 0,
    "loss": 0.253661572933197,
    "learning_rate": 0.0009977809823015396,
    "timestep_mean": 450.34375,
    "epsilon_norm": 7.418625354766846,
    "prediction_norm": 6.19527006149292,
    "timestamp": "2025-07-14T12:33:35.487065"
  },
  {
    "epoch": 60,
    "batch_idx": 32,
    "loss": 0.39855825901031494,
    "learning_rate": 0.0009977809823015396,
    "timestep_mean": 477.6875,
    "epsilon_norm": 8.060595512390137,
    "prediction_norm": 6.6503400802612305,
    "timestamp": "2025-07-14T12:33:35.492359"
  },
  {
    "epoch": 60,
    "batch_idx": 64,
    "loss": 0.2331591248512268,
    "learning_rate": 0.0009977809823015396,
    "timestep_mean": 548.875,
    "epsilon_norm": 7.505940914154053,
    "prediction_norm": 6.053282260894775,
    "timestamp": "2025-07-14T12:33:35.497518"
  },
  {
    "epoch": 60,
    "batch_idx": 96,
    "loss": 0.4487335681915283,
    "learning_rate": 0.0009977809823015396,
    "timestep_mean": 542.28125,
    "epsilon_norm": 8.560179710388184,
    "prediction_norm": 7.389504909515381,
    "timestamp": "2025-07-14T12:33:35.502682"
  },
  {
    "epoch": 60,
    "batch_idx": 128,
    "loss": 0.44899141788482666,
    "learning_rate": 0.0009977809823015396,
    "timestep_mean": 465.0625,
    "epsilon_norm": 7.75318717956543,
    "prediction_norm": 6.197990417480469,
    "timestamp": "2025-07-14T12:33:35.507822"
  },
  {
    "epoch": 60,
    "batch_idx": 160,
    "loss": 0.36551904678344727,
    "learning_rate": 0.0009977809823015396,
    "timestep_mean": 501.25,
    "epsilon_norm": 7.930178165435791,
    "prediction_norm": 6.324380874633789,
    "timestamp": "2025-07-14T12:33:35.513030"
  },
  {
    "epoch": 60,
    "batch_idx": 192,
    "loss": 0.6147680282592773,
    "learning_rate": 0.0009977809823015396,
    "timestep_mean": 377.625,
    "epsilon_norm": 4.106868743896484,
    "prediction_norm": 2.539008140563965,
    "timestamp": "2025-07-14T12:33:35.518148"
  },
  {
    "epoch": 61,
    "batch_idx": 0,
    "loss": 0.27317216992378235,
    "learning_rate": 0.0009977064557229486,
    "timestep_mean": 573.375,
    "epsilon_norm": 7.687648296356201,
    "prediction_norm": 5.985143184661865,
    "timestamp": "2025-07-14T12:33:35.523052"
  },
  {
    "epoch": 61,
    "batch_idx": 32,
    "loss": 0.3920038938522339,
    "learning_rate": 0.0009977064557229486,
    "timestep_mean": 518.09375,
    "epsilon_norm": 7.784027099609375,
    "prediction_norm": 5.983628749847412,
    "timestamp": "2025-07-14T12:33:35.528225"
  },
  {
    "epoch": 61,
    "batch_idx": 64,
    "loss": 0.4002222418785095,
    "learning_rate": 0.0009977064557229486,
    "timestep_mean": 469.09375,
    "epsilon_norm": 7.722041606903076,
    "prediction_norm": 5.608301162719727,
    "timestamp": "2025-07-14T12:33:35.533370"
  },
  {
    "epoch": 61,
    "batch_idx": 96,
    "loss": 0.497697651386261,
    "learning_rate": 0.0009977064557229486,
    "timestep_mean": 523.71875,
    "epsilon_norm": 8.82501220703125,
    "prediction_norm": 6.456358432769775,
    "timestamp": "2025-07-14T12:33:35.538433"
  },
  {
    "epoch": 61,
    "batch_idx": 128,
    "loss": 0.4957907199859619,
    "learning_rate": 0.0009977064557229486,
    "timestep_mean": 471.375,
    "epsilon_norm": 8.349644660949707,
    "prediction_norm": 5.504672527313232,
    "timestamp": "2025-07-14T12:33:35.543629"
  },
  {
    "epoch": 61,
    "batch_idx": 160,
    "loss": 0.5458956360816956,
    "learning_rate": 0.0009977064557229486,
    "timestep_mean": 447.78125,
    "epsilon_norm": 8.918803215026855,
    "prediction_norm": 5.638072967529297,
    "timestamp": "2025-07-14T12:33:35.548790"
  },
  {
    "epoch": 61,
    "batch_idx": 192,
    "loss": 0.15431943535804749,
    "learning_rate": 0.0009977064557229486,
    "timestep_mean": 590.875,
    "epsilon_norm": 4.2262678146362305,
    "prediction_norm": 3.4293322563171387,
    "timestamp": "2025-07-14T12:33:35.553882"
  },
  {
    "epoch": 62,
    "batch_idx": 0,
    "loss": 0.5432950258255005,
    "learning_rate": 0.0009976307011031538,
    "timestep_mean": 399.5,
    "epsilon_norm": 8.116588592529297,
    "prediction_norm": 5.315114974975586,
    "timestamp": "2025-07-14T12:33:35.559023"
  },
  {
    "epoch": 62,
    "batch_idx": 32,
    "loss": 0.4202460050582886,
    "learning_rate": 0.0009976307011031538,
    "timestep_mean": 382.0625,
    "epsilon_norm": 8.584787368774414,
    "prediction_norm": 6.5689568519592285,
    "timestamp": "2025-07-14T12:33:35.564129"
  },
  {
    "epoch": 62,
    "batch_idx": 64,
    "loss": 0.3816954791545868,
    "learning_rate": 0.0009976307011031538,
    "timestep_mean": 530.65625,
    "epsilon_norm": 7.682132720947266,
    "prediction_norm": 5.976955890655518,
    "timestamp": "2025-07-14T12:33:35.569284"
  },
  {
    "epoch": 62,
    "batch_idx": 96,
    "loss": 0.47873079776763916,
    "learning_rate": 0.0009976307011031538,
    "timestep_mean": 450.78125,
    "epsilon_norm": 8.612998962402344,
    "prediction_norm": 6.588570594787598,
    "timestamp": "2025-07-14T12:33:35.574512"
  },
  {
    "epoch": 62,
    "batch_idx": 128,
    "loss": 0.35255753993988037,
    "learning_rate": 0.0009976307011031538,
    "timestep_mean": 567.4375,
    "epsilon_norm": 8.43568229675293,
    "prediction_norm": 7.58745002746582,
    "timestamp": "2025-07-14T12:33:35.579747"
  },
  {
    "epoch": 62,
    "batch_idx": 160,
    "loss": 0.3555222749710083,
    "learning_rate": 0.0009976307011031538,
    "timestep_mean": 532.84375,
    "epsilon_norm": 8.636590003967285,
    "prediction_norm": 7.665837287902832,
    "timestamp": "2025-07-14T12:33:35.584945"
  },
  {
    "epoch": 62,
    "batch_idx": 192,
    "loss": 0.2268024981021881,
    "learning_rate": 0.0009976307011031538,
    "timestep_mean": 346.5,
    "epsilon_norm": 2.801265001296997,
    "prediction_norm": 2.1466622352600098,
    "timestamp": "2025-07-14T12:33:35.589926"
  },
  {
    "epoch": 63,
    "batch_idx": 0,
    "loss": 0.42570993304252625,
    "learning_rate": 0.000997553718629072,
    "timestep_mean": 446.65625,
    "epsilon_norm": 7.629116535186768,
    "prediction_norm": 7.187376976013184,
    "timestamp": "2025-07-14T12:33:35.595014"
  },
  {
    "epoch": 63,
    "batch_idx": 32,
    "loss": 0.3741728961467743,
    "learning_rate": 0.000997553718629072,
    "timestep_mean": 438.84375,
    "epsilon_norm": 7.065075397491455,
    "prediction_norm": 6.061850070953369,
    "timestamp": "2025-07-14T12:33:35.600218"
  },
  {
    "epoch": 63,
    "batch_idx": 64,
    "loss": 0.46212777495384216,
    "learning_rate": 0.000997553718629072,
    "timestep_mean": 491.4375,
    "epsilon_norm": 8.280192375183105,
    "prediction_norm": 7.2042670249938965,
    "timestamp": "2025-07-14T12:33:35.605205"
  },
  {
    "epoch": 63,
    "batch_idx": 96,
    "loss": 0.2849298417568207,
    "learning_rate": 0.000997553718629072,
    "timestep_mean": 390.53125,
    "epsilon_norm": 6.846533298492432,
    "prediction_norm": 6.120741844177246,
    "timestamp": "2025-07-14T12:33:35.610308"
  },
  {
    "epoch": 63,
    "batch_idx": 128,
    "loss": 0.3814074397087097,
    "learning_rate": 0.000997553718629072,
    "timestep_mean": 544.65625,
    "epsilon_norm": 8.498893737792969,
    "prediction_norm": 7.371804714202881,
    "timestamp": "2025-07-14T12:33:35.615358"
  },
  {
    "epoch": 63,
    "batch_idx": 160,
    "loss": 0.38439059257507324,
    "learning_rate": 0.000997553718629072,
    "timestep_mean": 465.4375,
    "epsilon_norm": 7.583632946014404,
    "prediction_norm": 5.950404167175293,
    "timestamp": "2025-07-14T12:33:35.620490"
  },
  {
    "epoch": 63,
    "batch_idx": 192,
    "loss": 0.02175269089639187,
    "learning_rate": 0.000997553718629072,
    "timestep_mean": 759.375,
    "epsilon_norm": 4.470927715301514,
    "prediction_norm": 4.269873142242432,
    "timestamp": "2025-07-14T12:33:35.625667"
  },
  {
    "epoch": 64,
    "batch_idx": 0,
    "loss": 0.3983690142631531,
    "learning_rate": 0.0009974755084906498,
    "timestep_mean": 430.5,
    "epsilon_norm": 7.578323841094971,
    "prediction_norm": 5.672607898712158,
    "timestamp": "2025-07-14T12:33:35.630755"
  },
  {
    "epoch": 64,
    "batch_idx": 32,
    "loss": 0.3859925866127014,
    "learning_rate": 0.0009974755084906498,
    "timestep_mean": 523.9375,
    "epsilon_norm": 7.410995006561279,
    "prediction_norm": 5.124393939971924,
    "timestamp": "2025-07-14T12:33:35.635887"
  },
  {
    "epoch": 64,
    "batch_idx": 64,
    "loss": 0.4780244827270508,
    "learning_rate": 0.0009974755084906498,
    "timestep_mean": 496.96875,
    "epsilon_norm": 8.257706642150879,
    "prediction_norm": 5.2942304611206055,
    "timestamp": "2025-07-14T12:33:35.640996"
  },
  {
    "epoch": 64,
    "batch_idx": 96,
    "loss": 0.298175573348999,
    "learning_rate": 0.0009974755084906498,
    "timestep_mean": 476.34375,
    "epsilon_norm": 7.813703536987305,
    "prediction_norm": 6.1123175621032715,
    "timestamp": "2025-07-14T12:33:35.646117"
  },
  {
    "epoch": 64,
    "batch_idx": 128,
    "loss": 0.46641871333122253,
    "learning_rate": 0.0009974755084906498,
    "timestep_mean": 480.75,
    "epsilon_norm": 7.451779842376709,
    "prediction_norm": 5.337629318237305,
    "timestamp": "2025-07-14T12:33:35.651222"
  },
  {
    "epoch": 64,
    "batch_idx": 160,
    "loss": 0.3846522569656372,
    "learning_rate": 0.0009974755084906498,
    "timestep_mean": 591.34375,
    "epsilon_norm": 7.880157947540283,
    "prediction_norm": 5.739945888519287,
    "timestamp": "2025-07-14T12:33:35.656273"
  },
  {
    "epoch": 64,
    "batch_idx": 192,
    "loss": 0.29492270946502686,
    "learning_rate": 0.0009974755084906498,
    "timestep_mean": 624.125,
    "epsilon_norm": 3.0578620433807373,
    "prediction_norm": 2.030395746231079,
    "timestamp": "2025-07-14T12:33:35.661418"
  },
  {
    "epoch": 65,
    "batch_idx": 0,
    "loss": 0.46757733821868896,
    "learning_rate": 0.0009973960708808628,
    "timestep_mean": 419.84375,
    "epsilon_norm": 7.859777450561523,
    "prediction_norm": 5.088408946990967,
    "timestamp": "2025-07-14T12:33:35.666622"
  },
  {
    "epoch": 65,
    "batch_idx": 32,
    "loss": 0.36426180601119995,
    "learning_rate": 0.0009973960708808628,
    "timestep_mean": 452.5625,
    "epsilon_norm": 6.757716655731201,
    "prediction_norm": 4.759342670440674,
    "timestamp": "2025-07-14T12:33:35.671641"
  },
  {
    "epoch": 65,
    "batch_idx": 64,
    "loss": 0.1878334879875183,
    "learning_rate": 0.0009973960708808628,
    "timestep_mean": 631.9375,
    "epsilon_norm": 7.227263450622559,
    "prediction_norm": 6.027932167053223,
    "timestamp": "2025-07-14T12:33:35.676847"
  },
  {
    "epoch": 65,
    "batch_idx": 96,
    "loss": 0.3418540954589844,
    "learning_rate": 0.0009973960708808628,
    "timestep_mean": 543.4375,
    "epsilon_norm": 8.631808280944824,
    "prediction_norm": 7.16747522354126,
    "timestamp": "2025-07-14T12:33:35.681839"
  },
  {
    "epoch": 65,
    "batch_idx": 128,
    "loss": 0.38733458518981934,
    "learning_rate": 0.0009973960708808628,
    "timestep_mean": 576.3125,
    "epsilon_norm": 7.223601818084717,
    "prediction_norm": 5.470914363861084,
    "timestamp": "2025-07-14T12:33:35.686949"
  },
  {
    "epoch": 65,
    "batch_idx": 160,
    "loss": 0.38371068239212036,
    "learning_rate": 0.0009973960708808628,
    "timestep_mean": 516.75,
    "epsilon_norm": 8.940473556518555,
    "prediction_norm": 7.161853313446045,
    "timestamp": "2025-07-14T12:33:35.692244"
  },
  {
    "epoch": 65,
    "batch_idx": 192,
    "loss": 0.4225117564201355,
    "learning_rate": 0.0009973960708808628,
    "timestep_mean": 378.875,
    "epsilon_norm": 4.029794692993164,
    "prediction_norm": 2.511880874633789,
    "timestamp": "2025-07-14T12:33:35.702236"
  },
  {
    "epoch": 66,
    "batch_idx": 0,
    "loss": 0.36585667729377747,
    "learning_rate": 0.0009973154059957157,
    "timestep_mean": 441.125,
    "epsilon_norm": 7.186144828796387,
    "prediction_norm": 5.8198699951171875,
    "timestamp": "2025-07-14T12:33:35.707386"
  },
  {
    "epoch": 66,
    "batch_idx": 32,
    "loss": 0.29904070496559143,
    "learning_rate": 0.0009973154059957157,
    "timestep_mean": 524.75,
    "epsilon_norm": 8.046570777893066,
    "prediction_norm": 6.758166790008545,
    "timestamp": "2025-07-14T12:33:35.712501"
  },
  {
    "epoch": 66,
    "batch_idx": 64,
    "loss": 0.1974959671497345,
    "learning_rate": 0.0009973154059957157,
    "timestep_mean": 603.125,
    "epsilon_norm": 7.9014058113098145,
    "prediction_norm": 7.119871139526367,
    "timestamp": "2025-07-14T12:33:35.717646"
  },
  {
    "epoch": 66,
    "batch_idx": 96,
    "loss": 0.32048916816711426,
    "learning_rate": 0.0009973154059957157,
    "timestep_mean": 462.875,
    "epsilon_norm": 6.922468185424805,
    "prediction_norm": 5.765222072601318,
    "timestamp": "2025-07-14T12:33:35.722838"
  },
  {
    "epoch": 66,
    "batch_idx": 128,
    "loss": 0.663560688495636,
    "learning_rate": 0.0009973154059957157,
    "timestep_mean": 412.3125,
    "epsilon_norm": 8.71666145324707,
    "prediction_norm": 6.435178279876709,
    "timestamp": "2025-07-14T12:33:35.727994"
  },
  {
    "epoch": 66,
    "batch_idx": 160,
    "loss": 0.43038564920425415,
    "learning_rate": 0.0009973154059957157,
    "timestep_mean": 456.28125,
    "epsilon_norm": 8.261205673217773,
    "prediction_norm": 6.254148483276367,
    "timestamp": "2025-07-14T12:33:35.733311"
  },
  {
    "epoch": 66,
    "batch_idx": 192,
    "loss": 0.17977239191532135,
    "learning_rate": 0.0009973154059957157,
    "timestep_mean": 540.875,
    "epsilon_norm": 3.9537172317504883,
    "prediction_norm": 3.38215708732605,
    "timestamp": "2025-07-14T12:33:35.738436"
  },
  {
    "epoch": 67,
    "batch_idx": 0,
    "loss": 0.4017961025238037,
    "learning_rate": 0.000997233514034241,
    "timestep_mean": 472.96875,
    "epsilon_norm": 8.130484580993652,
    "prediction_norm": 7.094439506530762,
    "timestamp": "2025-07-14T12:33:35.743626"
  },
  {
    "epoch": 67,
    "batch_idx": 32,
    "loss": 0.35278117656707764,
    "learning_rate": 0.000997233514034241,
    "timestep_mean": 528.9375,
    "epsilon_norm": 9.196707725524902,
    "prediction_norm": 8.3092622756958,
    "timestamp": "2025-07-14T12:33:35.748814"
  },
  {
    "epoch": 67,
    "batch_idx": 64,
    "loss": 0.3484402894973755,
    "learning_rate": 0.000997233514034241,
    "timestep_mean": 598.03125,
    "epsilon_norm": 8.682462692260742,
    "prediction_norm": 7.5440497398376465,
    "timestamp": "2025-07-14T12:33:35.753993"
  },
  {
    "epoch": 67,
    "batch_idx": 96,
    "loss": 0.4995907247066498,
    "learning_rate": 0.000997233514034241,
    "timestep_mean": 500.0,
    "epsilon_norm": 8.382730484008789,
    "prediction_norm": 6.4576735496521,
    "timestamp": "2025-07-14T12:33:35.759138"
  },
  {
    "epoch": 67,
    "batch_idx": 128,
    "loss": 0.30389100313186646,
    "learning_rate": 0.000997233514034241,
    "timestep_mean": 529.34375,
    "epsilon_norm": 7.356706619262695,
    "prediction_norm": 6.438066005706787,
    "timestamp": "2025-07-14T12:33:35.764241"
  },
  {
    "epoch": 67,
    "batch_idx": 160,
    "loss": 0.3695436716079712,
    "learning_rate": 0.000997233514034241,
    "timestep_mean": 496.28125,
    "epsilon_norm": 7.78070068359375,
    "prediction_norm": 5.584944248199463,
    "timestamp": "2025-07-14T12:33:35.769508"
  },
  {
    "epoch": 67,
    "batch_idx": 192,
    "loss": 0.995194673538208,
    "learning_rate": 0.000997233514034241,
    "timestep_mean": 419.5,
    "epsilon_norm": 5.22461462020874,
    "prediction_norm": 3.4785287380218506,
    "timestamp": "2025-07-14T12:33:35.774815"
  },
  {
    "epoch": 68,
    "batch_idx": 0,
    "loss": 0.32093364000320435,
    "learning_rate": 0.0009971503951984992,
    "timestep_mean": 537.0,
    "epsilon_norm": 8.65616226196289,
    "prediction_norm": 7.58226203918457,
    "timestamp": "2025-07-14T12:33:35.779887"
  },
  {
    "epoch": 68,
    "batch_idx": 32,
    "loss": 0.43964776396751404,
    "learning_rate": 0.0009971503951984992,
    "timestep_mean": 462.03125,
    "epsilon_norm": 8.018600463867188,
    "prediction_norm": 5.494295120239258,
    "timestamp": "2025-07-14T12:33:35.784912"
  },
  {
    "epoch": 68,
    "batch_idx": 64,
    "loss": 0.5104684829711914,
    "learning_rate": 0.0009971503951984992,
    "timestep_mean": 457.375,
    "epsilon_norm": 8.286520957946777,
    "prediction_norm": 6.674611568450928,
    "timestamp": "2025-07-14T12:33:35.790035"
  },
  {
    "epoch": 68,
    "batch_idx": 96,
    "loss": 0.4274768531322479,
    "learning_rate": 0.0009971503951984992,
    "timestep_mean": 537.125,
    "epsilon_norm": 8.677231788635254,
    "prediction_norm": 6.699729919433594,
    "timestamp": "2025-07-14T12:33:35.795167"
  },
  {
    "epoch": 68,
    "batch_idx": 128,
    "loss": 0.48350614309310913,
    "learning_rate": 0.0009971503951984992,
    "timestep_mean": 475.03125,
    "epsilon_norm": 7.604640007019043,
    "prediction_norm": 5.568926811218262,
    "timestamp": "2025-07-14T12:33:35.800261"
  },
  {
    "epoch": 68,
    "batch_idx": 160,
    "loss": 0.6249456405639648,
    "learning_rate": 0.0009971503951984992,
    "timestep_mean": 406.03125,
    "epsilon_norm": 8.21788215637207,
    "prediction_norm": 5.053614616394043,
    "timestamp": "2025-07-14T12:33:35.805329"
  },
  {
    "epoch": 68,
    "batch_idx": 192,
    "loss": 0.7093954086303711,
    "learning_rate": 0.0009971503951984992,
    "timestep_mean": 391.25,
    "epsilon_norm": 4.116548538208008,
    "prediction_norm": 2.122952938079834,
    "timestamp": "2025-07-14T12:33:35.810395"
  },
  {
    "epoch": 69,
    "batch_idx": 0,
    "loss": 0.19547224044799805,
    "learning_rate": 0.0009970660496935772,
    "timestep_mean": 571.53125,
    "epsilon_norm": 8.083656311035156,
    "prediction_norm": 7.282032489776611,
    "timestamp": "2025-07-14T12:33:35.815600"
  },
  {
    "epoch": 69,
    "batch_idx": 32,
    "loss": 0.3663930892944336,
    "learning_rate": 0.0009970660496935772,
    "timestep_mean": 495.5,
    "epsilon_norm": 7.6884074211120605,
    "prediction_norm": 5.816800594329834,
    "timestamp": "2025-07-14T12:33:35.820749"
  },
  {
    "epoch": 69,
    "batch_idx": 64,
    "loss": 0.36781826615333557,
    "learning_rate": 0.0009970660496935772,
    "timestep_mean": 556.75,
    "epsilon_norm": 8.461858749389648,
    "prediction_norm": 6.595249652862549,
    "timestamp": "2025-07-14T12:33:35.825898"
  },
  {
    "epoch": 69,
    "batch_idx": 96,
    "loss": 0.6697105169296265,
    "learning_rate": 0.0009970660496935772,
    "timestep_mean": 363.84375,
    "epsilon_norm": 8.860550880432129,
    "prediction_norm": 5.123664379119873,
    "timestamp": "2025-07-14T12:33:35.830995"
  },
  {
    "epoch": 69,
    "batch_idx": 128,
    "loss": 0.21402814984321594,
    "learning_rate": 0.0009970660496935772,
    "timestep_mean": 563.46875,
    "epsilon_norm": 6.929014682769775,
    "prediction_norm": 6.2387189865112305,
    "timestamp": "2025-07-14T12:33:35.836162"
  },
  {
    "epoch": 69,
    "batch_idx": 160,
    "loss": 0.2696759104728699,
    "learning_rate": 0.0009970660496935772,
    "timestep_mean": 476.59375,
    "epsilon_norm": 7.445709228515625,
    "prediction_norm": 6.148009777069092,
    "timestamp": "2025-07-14T12:33:35.841279"
  },
  {
    "epoch": 69,
    "batch_idx": 192,
    "loss": 0.46812236309051514,
    "learning_rate": 0.0009970660496935772,
    "timestep_mean": 422.75,
    "epsilon_norm": 3.581373453140259,
    "prediction_norm": 2.5963685512542725,
    "timestamp": "2025-07-14T12:33:35.846445"
  },
  {
    "epoch": 70,
    "batch_idx": 0,
    "loss": 0.4453624188899994,
    "learning_rate": 0.0009969804777275897,
    "timestep_mean": 473.4375,
    "epsilon_norm": 8.116089820861816,
    "prediction_norm": 6.314143180847168,
    "timestamp": "2025-07-14T12:33:35.851616"
  },
  {
    "epoch": 70,
    "batch_idx": 32,
    "loss": 0.23673872649669647,
    "learning_rate": 0.0009969804777275897,
    "timestep_mean": 515.8125,
    "epsilon_norm": 7.170258045196533,
    "prediction_norm": 6.393421649932861,
    "timestamp": "2025-07-14T12:33:35.856811"
  },
  {
    "epoch": 70,
    "batch_idx": 64,
    "loss": 0.3865090012550354,
    "learning_rate": 0.0009969804777275897,
    "timestep_mean": 447.0,
    "epsilon_norm": 7.309277057647705,
    "prediction_norm": 5.048390865325928,
    "timestamp": "2025-07-14T12:33:35.861960"
  },
  {
    "epoch": 70,
    "batch_idx": 96,
    "loss": 0.286643922328949,
    "learning_rate": 0.0009969804777275897,
    "timestep_mean": 529.28125,
    "epsilon_norm": 8.088911056518555,
    "prediction_norm": 6.958602428436279,
    "timestamp": "2025-07-14T12:33:35.867168"
  },
  {
    "epoch": 70,
    "batch_idx": 128,
    "loss": 0.6118505001068115,
    "learning_rate": 0.0009969804777275897,
    "timestep_mean": 439.1875,
    "epsilon_norm": 9.190605163574219,
    "prediction_norm": 6.728879451751709,
    "timestamp": "2025-07-14T12:33:35.872162"
  },
  {
    "epoch": 70,
    "batch_idx": 160,
    "loss": 0.20436692237854004,
    "learning_rate": 0.0009969804777275897,
    "timestep_mean": 563.875,
    "epsilon_norm": 7.484165668487549,
    "prediction_norm": 7.1283464431762695,
    "timestamp": "2025-07-14T12:33:35.877346"
  },
  {
    "epoch": 70,
    "batch_idx": 192,
    "loss": 1.0022507905960083,
    "learning_rate": 0.0009969804777275897,
    "timestep_mean": 563.0,
    "epsilon_norm": 5.2957234382629395,
    "prediction_norm": 4.0588555335998535,
    "timestamp": "2025-07-14T12:33:35.882370"
  },
  {
    "epoch": 71,
    "batch_idx": 0,
    "loss": 0.5064877271652222,
    "learning_rate": 0.0009968936795116765,
    "timestep_mean": 486.15625,
    "epsilon_norm": 7.790098190307617,
    "prediction_norm": 5.7044806480407715,
    "timestamp": "2025-07-14T12:33:35.887465"
  },
  {
    "epoch": 71,
    "batch_idx": 32,
    "loss": 0.2538284659385681,
    "learning_rate": 0.0009968936795116765,
    "timestep_mean": 536.78125,
    "epsilon_norm": 8.504935264587402,
    "prediction_norm": 7.806210517883301,
    "timestamp": "2025-07-14T12:33:35.892461"
  },
  {
    "epoch": 71,
    "batch_idx": 64,
    "loss": 0.2490912824869156,
    "learning_rate": 0.0009968936795116765,
    "timestep_mean": 569.90625,
    "epsilon_norm": 8.575786590576172,
    "prediction_norm": 7.4216790199279785,
    "timestamp": "2025-07-14T12:33:35.897697"
  },
  {
    "epoch": 71,
    "batch_idx": 96,
    "loss": 0.46363282203674316,
    "learning_rate": 0.0009968936795116765,
    "timestep_mean": 576.65625,
    "epsilon_norm": 8.560187339782715,
    "prediction_norm": 6.910623073577881,
    "timestamp": "2025-07-14T12:33:35.902933"
  },
  {
    "epoch": 71,
    "batch_idx": 128,
    "loss": 0.47618335485458374,
    "learning_rate": 0.0009968936795116765,
    "timestep_mean": 446.875,
    "epsilon_norm": 9.081960678100586,
    "prediction_norm": 6.639400005340576,
    "timestamp": "2025-07-14T12:33:35.908173"
  },
  {
    "epoch": 71,
    "batch_idx": 160,
    "loss": 0.5487894415855408,
    "learning_rate": 0.0009968936795116765,
    "timestep_mean": 392.25,
    "epsilon_norm": 7.792753219604492,
    "prediction_norm": 5.612519264221191,
    "timestamp": "2025-07-14T12:33:35.913216"
  },
  {
    "epoch": 71,
    "batch_idx": 192,
    "loss": 0.1613432765007019,
    "learning_rate": 0.0009968936795116765,
    "timestep_mean": 530.625,
    "epsilon_norm": 3.211463451385498,
    "prediction_norm": 3.1487832069396973,
    "timestamp": "2025-07-14T12:33:35.918381"
  },
  {
    "epoch": 72,
    "batch_idx": 0,
    "loss": 0.5249707102775574,
    "learning_rate": 0.000996805655260004,
    "timestep_mean": 412.1875,
    "epsilon_norm": 7.543299198150635,
    "prediction_norm": 4.742505073547363,
    "timestamp": "2025-07-14T12:33:35.923500"
  },
  {
    "epoch": 72,
    "batch_idx": 32,
    "loss": 0.3684375286102295,
    "learning_rate": 0.000996805655260004,
    "timestep_mean": 542.125,
    "epsilon_norm": 7.9995646476745605,
    "prediction_norm": 5.113036155700684,
    "timestamp": "2025-07-14T12:33:35.928663"
  },
  {
    "epoch": 72,
    "batch_idx": 64,
    "loss": 0.4172227382659912,
    "learning_rate": 0.000996805655260004,
    "timestep_mean": 467.3125,
    "epsilon_norm": 7.8199567794799805,
    "prediction_norm": 4.997867107391357,
    "timestamp": "2025-07-14T12:33:35.933841"
  },
  {
    "epoch": 72,
    "batch_idx": 96,
    "loss": 0.5406438112258911,
    "learning_rate": 0.000996805655260004,
    "timestep_mean": 389.8125,
    "epsilon_norm": 7.802121639251709,
    "prediction_norm": 5.460415363311768,
    "timestamp": "2025-07-14T12:33:35.938910"
  },
  {
    "epoch": 72,
    "batch_idx": 128,
    "loss": 0.3174068331718445,
    "learning_rate": 0.000996805655260004,
    "timestep_mean": 530.53125,
    "epsilon_norm": 7.9616875648498535,
    "prediction_norm": 5.8409013748168945,
    "timestamp": "2025-07-14T12:33:35.944118"
  },
  {
    "epoch": 72,
    "batch_idx": 160,
    "loss": 0.40685904026031494,
    "learning_rate": 0.000996805655260004,
    "timestep_mean": 490.15625,
    "epsilon_norm": 7.431028842926025,
    "prediction_norm": 5.570094108581543,
    "timestamp": "2025-07-14T12:33:35.949289"
  },
  {
    "epoch": 72,
    "batch_idx": 192,
    "loss": 0.3570707440376282,
    "learning_rate": 0.000996805655260004,
    "timestep_mean": 392.375,
    "epsilon_norm": 4.900972366333008,
    "prediction_norm": 3.5080835819244385,
    "timestamp": "2025-07-14T12:33:35.954519"
  },
  {
    "epoch": 73,
    "batch_idx": 0,
    "loss": 0.5591930150985718,
    "learning_rate": 0.000996716405189763,
    "timestep_mean": 528.8125,
    "epsilon_norm": 8.47113037109375,
    "prediction_norm": 6.14518928527832,
    "timestamp": "2025-07-14T12:33:35.959694"
  },
  {
    "epoch": 73,
    "batch_idx": 32,
    "loss": 0.42381051182746887,
    "learning_rate": 0.000996716405189763,
    "timestep_mean": 432.03125,
    "epsilon_norm": 8.214213371276855,
    "prediction_norm": 5.590367317199707,
    "timestamp": "2025-07-14T12:33:35.964841"
  },
  {
    "epoch": 73,
    "batch_idx": 64,
    "loss": 0.4346664547920227,
    "learning_rate": 0.000996716405189763,
    "timestep_mean": 461.03125,
    "epsilon_norm": 8.042475700378418,
    "prediction_norm": 6.200558662414551,
    "timestamp": "2025-07-14T12:33:35.970012"
  },
  {
    "epoch": 73,
    "batch_idx": 96,
    "loss": 0.18205295503139496,
    "learning_rate": 0.000996716405189763,
    "timestep_mean": 629.5,
    "epsilon_norm": 7.404941082000732,
    "prediction_norm": 7.113296031951904,
    "timestamp": "2025-07-14T12:33:35.975233"
  },
  {
    "epoch": 73,
    "batch_idx": 128,
    "loss": 0.19037100672721863,
    "learning_rate": 0.000996716405189763,
    "timestep_mean": 600.84375,
    "epsilon_norm": 6.482698440551758,
    "prediction_norm": 5.660314083099365,
    "timestamp": "2025-07-14T12:33:35.980210"
  },
  {
    "epoch": 73,
    "batch_idx": 160,
    "loss": 0.20412692427635193,
    "learning_rate": 0.000996716405189763,
    "timestep_mean": 572.78125,
    "epsilon_norm": 7.160956859588623,
    "prediction_norm": 6.406454563140869,
    "timestamp": "2025-07-14T12:33:35.985461"
  },
  {
    "epoch": 73,
    "batch_idx": 192,
    "loss": 0.21257703006267548,
    "learning_rate": 0.000996716405189763,
    "timestep_mean": 538.5,
    "epsilon_norm": 3.473243474960327,
    "prediction_norm": 3.1171584129333496,
    "timestamp": "2025-07-14T12:33:35.990422"
  },
  {
    "epoch": 74,
    "batch_idx": 0,
    "loss": 0.3501836061477661,
    "learning_rate": 0.0009966259295211694,
    "timestep_mean": 477.4375,
    "epsilon_norm": 6.965643405914307,
    "prediction_norm": 5.8159098625183105,
    "timestamp": "2025-07-14T12:33:35.995257"
  },
  {
    "epoch": 74,
    "batch_idx": 32,
    "loss": 0.44565147161483765,
    "learning_rate": 0.0009966259295211694,
    "timestep_mean": 480.5625,
    "epsilon_norm": 8.334395408630371,
    "prediction_norm": 6.633623123168945,
    "timestamp": "2025-07-14T12:33:36.000214"
  },
  {
    "epoch": 74,
    "batch_idx": 64,
    "loss": 0.37807896733283997,
    "learning_rate": 0.0009966259295211694,
    "timestep_mean": 560.6875,
    "epsilon_norm": 8.32777214050293,
    "prediction_norm": 6.655712604522705,
    "timestamp": "2025-07-14T12:33:36.005053"
  },
  {
    "epoch": 74,
    "batch_idx": 96,
    "loss": 0.5351210832595825,
    "learning_rate": 0.0009966259295211694,
    "timestep_mean": 510.4375,
    "epsilon_norm": 8.073041915893555,
    "prediction_norm": 5.797770977020264,
    "timestamp": "2025-07-14T12:33:36.009934"
  },
  {
    "epoch": 74,
    "batch_idx": 128,
    "loss": 0.27265042066574097,
    "learning_rate": 0.0009966259295211694,
    "timestep_mean": 515.15625,
    "epsilon_norm": 7.3446478843688965,
    "prediction_norm": 6.504502773284912,
    "timestamp": "2025-07-14T12:33:36.014780"
  },
  {
    "epoch": 74,
    "batch_idx": 160,
    "loss": 0.5531566143035889,
    "learning_rate": 0.0009966259295211694,
    "timestep_mean": 425.0,
    "epsilon_norm": 8.707377433776855,
    "prediction_norm": 6.044936180114746,
    "timestamp": "2025-07-14T12:33:36.019636"
  },
  {
    "epoch": 74,
    "batch_idx": 192,
    "loss": 1.1505115032196045,
    "learning_rate": 0.0009966259295211694,
    "timestep_mean": 351.5,
    "epsilon_norm": 5.354159355163574,
    "prediction_norm": 2.7228176593780518,
    "timestamp": "2025-07-14T12:33:36.024462"
  },
  {
    "epoch": 75,
    "batch_idx": 0,
    "loss": 0.45333677530288696,
    "learning_rate": 0.000996534228477463,
    "timestep_mean": 476.125,
    "epsilon_norm": 7.3873772621154785,
    "prediction_norm": 5.1820549964904785,
    "timestamp": "2025-07-14T12:33:36.029365"
  },
  {
    "epoch": 75,
    "batch_idx": 32,
    "loss": 0.3925010561943054,
    "learning_rate": 0.000996534228477463,
    "timestep_mean": 503.125,
    "epsilon_norm": 7.761870384216309,
    "prediction_norm": 5.480347156524658,
    "timestamp": "2025-07-14T12:33:36.034177"
  },
  {
    "epoch": 75,
    "batch_idx": 64,
    "loss": 0.24819506704807281,
    "learning_rate": 0.000996534228477463,
    "timestep_mean": 525.71875,
    "epsilon_norm": 7.783392906188965,
    "prediction_norm": 6.902253150939941,
    "timestamp": "2025-07-14T12:33:36.039040"
  },
  {
    "epoch": 75,
    "batch_idx": 96,
    "loss": 0.32627198100090027,
    "learning_rate": 0.000996534228477463,
    "timestep_mean": 517.34375,
    "epsilon_norm": 8.292522430419922,
    "prediction_norm": 7.115931510925293,
    "timestamp": "2025-07-14T12:33:36.043882"
  },
  {
    "epoch": 75,
    "batch_idx": 128,
    "loss": 0.5306358337402344,
    "learning_rate": 0.000996534228477463,
    "timestep_mean": 431.84375,
    "epsilon_norm": 8.107317924499512,
    "prediction_norm": 5.639224052429199,
    "timestamp": "2025-07-14T12:33:36.048793"
  },
  {
    "epoch": 75,
    "batch_idx": 160,
    "loss": 0.3569982051849365,
    "learning_rate": 0.000996534228477463,
    "timestep_mean": 468.5,
    "epsilon_norm": 7.438900470733643,
    "prediction_norm": 6.043072700500488,
    "timestamp": "2025-07-14T12:33:36.053644"
  },
  {
    "epoch": 75,
    "batch_idx": 192,
    "loss": 0.16464054584503174,
    "learning_rate": 0.000996534228477463,
    "timestep_mean": 588.0,
    "epsilon_norm": 3.8652126789093018,
    "prediction_norm": 3.8221583366394043,
    "timestamp": "2025-07-14T12:33:36.058539"
  },
  {
    "epoch": 76,
    "batch_idx": 0,
    "loss": 0.3521265685558319,
    "learning_rate": 0.0009964413022849066,
    "timestep_mean": 478.90625,
    "epsilon_norm": 8.548921585083008,
    "prediction_norm": 7.886898040771484,
    "timestamp": "2025-07-14T12:33:36.063388"
  },
  {
    "epoch": 76,
    "batch_idx": 32,
    "loss": 0.3310493528842926,
    "learning_rate": 0.0009964413022849066,
    "timestep_mean": 516.71875,
    "epsilon_norm": 8.01241397857666,
    "prediction_norm": 6.582779407501221,
    "timestamp": "2025-07-14T12:33:36.068206"
  },
  {
    "epoch": 76,
    "batch_idx": 64,
    "loss": 0.4240565896034241,
    "learning_rate": 0.0009964413022849066,
    "timestep_mean": 441.125,
    "epsilon_norm": 8.058753967285156,
    "prediction_norm": 6.085375785827637,
    "timestamp": "2025-07-14T12:33:36.073047"
  },
  {
    "epoch": 76,
    "batch_idx": 96,
    "loss": 0.2855614423751831,
    "learning_rate": 0.0009964413022849066,
    "timestep_mean": 568.375,
    "epsilon_norm": 7.7116522789001465,
    "prediction_norm": 6.6584038734436035,
    "timestamp": "2025-07-14T12:33:36.077911"
  },
  {
    "epoch": 76,
    "batch_idx": 128,
    "loss": 0.2865007817745209,
    "learning_rate": 0.0009964413022849066,
    "timestep_mean": 512.71875,
    "epsilon_norm": 8.14345645904541,
    "prediction_norm": 6.787342548370361,
    "timestamp": "2025-07-14T12:33:36.082758"
  },
  {
    "epoch": 76,
    "batch_idx": 160,
    "loss": 0.4507352411746979,
    "learning_rate": 0.0009964413022849066,
    "timestep_mean": 510.65625,
    "epsilon_norm": 7.837698459625244,
    "prediction_norm": 5.265155792236328,
    "timestamp": "2025-07-14T12:33:36.087668"
  },
  {
    "epoch": 76,
    "batch_idx": 192,
    "loss": 0.711286187171936,
    "learning_rate": 0.0009964413022849066,
    "timestep_mean": 391.875,
    "epsilon_norm": 6.149535179138184,
    "prediction_norm": 3.728339910507202,
    "timestamp": "2025-07-14T12:33:36.092484"
  },
  {
    "epoch": 77,
    "batch_idx": 0,
    "loss": 0.2176000475883484,
    "learning_rate": 0.0009963471511727866,
    "timestep_mean": 479.0625,
    "epsilon_norm": 6.826260566711426,
    "prediction_norm": 5.818026065826416,
    "timestamp": "2025-07-14T12:33:36.097344"
  },
  {
    "epoch": 77,
    "batch_idx": 32,
    "loss": 0.6962525844573975,
    "learning_rate": 0.0009963471511727866,
    "timestep_mean": 445.75,
    "epsilon_norm": 9.593816757202148,
    "prediction_norm": 6.870771884918213,
    "timestamp": "2025-07-14T12:33:36.102211"
  },
  {
    "epoch": 77,
    "batch_idx": 64,
    "loss": 0.33567386865615845,
    "learning_rate": 0.0009963471511727866,
    "timestep_mean": 546.78125,
    "epsilon_norm": 8.202771186828613,
    "prediction_norm": 6.590116024017334,
    "timestamp": "2025-07-14T12:33:36.107035"
  },
  {
    "epoch": 77,
    "batch_idx": 96,
    "loss": 0.23918429017066956,
    "learning_rate": 0.0009963471511727866,
    "timestep_mean": 512.375,
    "epsilon_norm": 7.63981294631958,
    "prediction_norm": 6.8636860847473145,
    "timestamp": "2025-07-14T12:33:36.111879"
  },
  {
    "epoch": 77,
    "batch_idx": 128,
    "loss": 0.5130612254142761,
    "learning_rate": 0.0009963471511727866,
    "timestep_mean": 433.03125,
    "epsilon_norm": 8.086710929870605,
    "prediction_norm": 6.501905918121338,
    "timestamp": "2025-07-14T12:33:36.116865"
  },
  {
    "epoch": 77,
    "batch_idx": 160,
    "loss": 0.21633800864219666,
    "learning_rate": 0.0009963471511727866,
    "timestep_mean": 582.4375,
    "epsilon_norm": 6.140625476837158,
    "prediction_norm": 5.983260631561279,
    "timestamp": "2025-07-14T12:33:36.121763"
  },
  {
    "epoch": 77,
    "batch_idx": 192,
    "loss": 0.10550467669963837,
    "learning_rate": 0.0009963471511727866,
    "timestep_mean": 612.5,
    "epsilon_norm": 4.609331130981445,
    "prediction_norm": 4.49946928024292,
    "timestamp": "2025-07-14T12:33:36.126622"
  },
  {
    "epoch": 78,
    "batch_idx": 0,
    "loss": 0.32457706332206726,
    "learning_rate": 0.0009962517753734116,
    "timestep_mean": 556.1875,
    "epsilon_norm": 8.446720123291016,
    "prediction_norm": 6.718835353851318,
    "timestamp": "2025-07-14T12:33:36.131448"
  },
  {
    "epoch": 78,
    "batch_idx": 32,
    "loss": 0.4256301522254944,
    "learning_rate": 0.0009962517753734116,
    "timestep_mean": 557.09375,
    "epsilon_norm": 8.407120704650879,
    "prediction_norm": 6.614457130432129,
    "timestamp": "2025-07-14T12:33:36.136285"
  },
  {
    "epoch": 78,
    "batch_idx": 64,
    "loss": 0.3324577212333679,
    "learning_rate": 0.0009962517753734116,
    "timestep_mean": 532.59375,
    "epsilon_norm": 6.597814083099365,
    "prediction_norm": 5.2715678215026855,
    "timestamp": "2025-07-14T12:33:36.141123"
  },
  {
    "epoch": 78,
    "batch_idx": 96,
    "loss": 0.24393129348754883,
    "learning_rate": 0.0009962517753734116,
    "timestep_mean": 550.125,
    "epsilon_norm": 8.805728912353516,
    "prediction_norm": 7.582371711730957,
    "timestamp": "2025-07-14T12:33:36.146072"
  },
  {
    "epoch": 78,
    "batch_idx": 128,
    "loss": 0.2593621015548706,
    "learning_rate": 0.0009962517753734116,
    "timestep_mean": 530.59375,
    "epsilon_norm": 7.064943313598633,
    "prediction_norm": 5.789179801940918,
    "timestamp": "2025-07-14T12:33:36.150885"
  },
  {
    "epoch": 78,
    "batch_idx": 160,
    "loss": 0.20369160175323486,
    "learning_rate": 0.0009962517753734116,
    "timestep_mean": 470.84375,
    "epsilon_norm": 6.163996696472168,
    "prediction_norm": 4.955624103546143,
    "timestamp": "2025-07-14T12:33:36.155717"
  },
  {
    "epoch": 78,
    "batch_idx": 192,
    "loss": 0.5590087175369263,
    "learning_rate": 0.0009962517753734116,
    "timestep_mean": 556.875,
    "epsilon_norm": 6.514659404754639,
    "prediction_norm": 5.273530960083008,
    "timestamp": "2025-07-14T12:33:36.160550"
  },
  {
    "epoch": 79,
    "batch_idx": 0,
    "loss": 0.3279922902584076,
    "learning_rate": 0.0009961551751221118,
    "timestep_mean": 543.375,
    "epsilon_norm": 7.790713787078857,
    "prediction_norm": 6.121979236602783,
    "timestamp": "2025-07-14T12:33:36.165405"
  },
  {
    "epoch": 79,
    "batch_idx": 32,
    "loss": 0.1737201064825058,
    "learning_rate": 0.0009961551751221118,
    "timestep_mean": 538.71875,
    "epsilon_norm": 8.152202606201172,
    "prediction_norm": 7.5213704109191895,
    "timestamp": "2025-07-14T12:33:36.170237"
  },
  {
    "epoch": 79,
    "batch_idx": 64,
    "loss": 0.43399280309677124,
    "learning_rate": 0.0009961551751221118,
    "timestep_mean": 532.1875,
    "epsilon_norm": 8.528038024902344,
    "prediction_norm": 6.608132362365723,
    "timestamp": "2025-07-14T12:33:36.175194"
  },
  {
    "epoch": 79,
    "batch_idx": 96,
    "loss": 0.45787549018859863,
    "learning_rate": 0.0009961551751221118,
    "timestep_mean": 436.6875,
    "epsilon_norm": 7.798862457275391,
    "prediction_norm": 5.973755836486816,
    "timestamp": "2025-07-14T12:33:36.180069"
  },
  {
    "epoch": 79,
    "batch_idx": 128,
    "loss": 0.3779512643814087,
    "learning_rate": 0.0009961551751221118,
    "timestep_mean": 518.28125,
    "epsilon_norm": 8.824989318847656,
    "prediction_norm": 7.3573222160339355,
    "timestamp": "2025-07-14T12:33:36.184921"
  },
  {
    "epoch": 79,
    "batch_idx": 160,
    "loss": 0.3771446645259857,
    "learning_rate": 0.0009961551751221118,
    "timestep_mean": 430.0,
    "epsilon_norm": 8.0455904006958,
    "prediction_norm": 6.668665885925293,
    "timestamp": "2025-07-14T12:33:36.189748"
  },
  {
    "epoch": 79,
    "batch_idx": 192,
    "loss": 0.6271024942398071,
    "learning_rate": 0.0009961551751221118,
    "timestep_mean": 362.0,
    "epsilon_norm": 3.698796033859253,
    "prediction_norm": 2.6524083614349365,
    "timestamp": "2025-07-14T12:33:36.194568"
  },
  {
    "epoch": 80,
    "batch_idx": 0,
    "loss": 0.3943827152252197,
    "learning_rate": 0.0009960573506572388,
    "timestep_mean": 482.625,
    "epsilon_norm": 6.770893096923828,
    "prediction_norm": 5.106751918792725,
    "timestamp": "2025-07-14T12:33:36.199481"
  },
  {
    "epoch": 80,
    "batch_idx": 32,
    "loss": 0.40244191884994507,
    "learning_rate": 0.0009960573506572388,
    "timestep_mean": 503.59375,
    "epsilon_norm": 7.619384288787842,
    "prediction_norm": 5.685543060302734,
    "timestamp": "2025-07-14T12:33:36.204504"
  },
  {
    "epoch": 80,
    "batch_idx": 64,
    "loss": 0.2530096173286438,
    "learning_rate": 0.0009960573506572388,
    "timestep_mean": 555.75,
    "epsilon_norm": 8.10655403137207,
    "prediction_norm": 7.354044437408447,
    "timestamp": "2025-07-14T12:33:36.209360"
  },
  {
    "epoch": 80,
    "batch_idx": 96,
    "loss": 0.38867682218551636,
    "learning_rate": 0.0009960573506572388,
    "timestep_mean": 534.71875,
    "epsilon_norm": 8.776373863220215,
    "prediction_norm": 7.604098320007324,
    "timestamp": "2025-07-14T12:33:36.214184"
  },
  {
    "epoch": 80,
    "batch_idx": 128,
    "loss": 0.2131594717502594,
    "learning_rate": 0.0009960573506572388,
    "timestep_mean": 582.375,
    "epsilon_norm": 8.002435684204102,
    "prediction_norm": 7.869243621826172,
    "timestamp": "2025-07-14T12:33:36.219328"
  },
  {
    "epoch": 80,
    "batch_idx": 160,
    "loss": 0.7613075971603394,
    "learning_rate": 0.0009960573506572388,
    "timestep_mean": 424.59375,
    "epsilon_norm": 9.909728050231934,
    "prediction_norm": 6.183533191680908,
    "timestamp": "2025-07-14T12:33:36.224429"
  },
  {
    "epoch": 80,
    "batch_idx": 192,
    "loss": 0.22792550921440125,
    "learning_rate": 0.0009960573506572388,
    "timestep_mean": 599.875,
    "epsilon_norm": 3.5592892169952393,
    "prediction_norm": 3.23419189453125,
    "timestamp": "2025-07-14T12:33:36.229692"
  },
  {
    "epoch": 81,
    "batch_idx": 0,
    "loss": 0.3450525999069214,
    "learning_rate": 0.0009959583022201644,
    "timestep_mean": 538.21875,
    "epsilon_norm": 7.313296318054199,
    "prediction_norm": 5.633719444274902,
    "timestamp": "2025-07-14T12:33:36.234882"
  },
  {
    "epoch": 81,
    "batch_idx": 32,
    "loss": 0.2661622166633606,
    "learning_rate": 0.0009959583022201644,
    "timestep_mean": 520.625,
    "epsilon_norm": 7.631133556365967,
    "prediction_norm": 6.480350494384766,
    "timestamp": "2025-07-14T12:33:36.240093"
  },
  {
    "epoch": 81,
    "batch_idx": 64,
    "loss": 0.44139352440834045,
    "learning_rate": 0.0009959583022201644,
    "timestep_mean": 465.4375,
    "epsilon_norm": 8.895678520202637,
    "prediction_norm": 6.418815612792969,
    "timestamp": "2025-07-14T12:33:36.245202"
  },
  {
    "epoch": 81,
    "batch_idx": 96,
    "loss": 0.5716961622238159,
    "learning_rate": 0.0009959583022201644,
    "timestep_mean": 436.96875,
    "epsilon_norm": 6.949334144592285,
    "prediction_norm": 3.51096773147583,
    "timestamp": "2025-07-14T12:33:36.250491"
  },
  {
    "epoch": 81,
    "batch_idx": 128,
    "loss": 0.27186155319213867,
    "learning_rate": 0.0009959583022201644,
    "timestep_mean": 462.15625,
    "epsilon_norm": 7.435543060302734,
    "prediction_norm": 6.110170841217041,
    "timestamp": "2025-07-14T12:33:36.255560"
  },
  {
    "epoch": 81,
    "batch_idx": 160,
    "loss": 0.3528137803077698,
    "learning_rate": 0.0009959583022201644,
    "timestep_mean": 521.03125,
    "epsilon_norm": 8.709997177124023,
    "prediction_norm": 7.182016849517822,
    "timestamp": "2025-07-14T12:33:36.260744"
  },
  {
    "epoch": 81,
    "batch_idx": 192,
    "loss": 0.2608186900615692,
    "learning_rate": 0.0009959583022201644,
    "timestep_mean": 528.0,
    "epsilon_norm": 4.023627281188965,
    "prediction_norm": 3.327615976333618,
    "timestamp": "2025-07-14T12:33:36.265796"
  },
  {
    "epoch": 82,
    "batch_idx": 0,
    "loss": 0.3261563181877136,
    "learning_rate": 0.0009958580300552813,
    "timestep_mean": 519.03125,
    "epsilon_norm": 7.407475471496582,
    "prediction_norm": 6.025360584259033,
    "timestamp": "2025-07-14T12:33:36.270950"
  },
  {
    "epoch": 82,
    "batch_idx": 32,
    "loss": 0.3980061411857605,
    "learning_rate": 0.0009958580300552813,
    "timestep_mean": 427.5625,
    "epsilon_norm": 7.671177864074707,
    "prediction_norm": 5.5067524909973145,
    "timestamp": "2025-07-14T12:33:36.276051"
  },
  {
    "epoch": 82,
    "batch_idx": 64,
    "loss": 0.31016457080841064,
    "learning_rate": 0.0009958580300552813,
    "timestep_mean": 587.03125,
    "epsilon_norm": 7.876150608062744,
    "prediction_norm": 6.490326881408691,
    "timestamp": "2025-07-14T12:33:36.281212"
  },
  {
    "epoch": 82,
    "batch_idx": 96,
    "loss": 0.45127832889556885,
    "learning_rate": 0.0009958580300552813,
    "timestep_mean": 437.59375,
    "epsilon_norm": 7.968522071838379,
    "prediction_norm": 6.338454723358154,
    "timestamp": "2025-07-14T12:33:36.286345"
  },
  {
    "epoch": 82,
    "batch_idx": 128,
    "loss": 0.19579900801181793,
    "learning_rate": 0.0009958580300552813,
    "timestep_mean": 529.78125,
    "epsilon_norm": 6.135337829589844,
    "prediction_norm": 5.597999095916748,
    "timestamp": "2025-07-14T12:33:36.291659"
  },
  {
    "epoch": 82,
    "batch_idx": 160,
    "loss": 0.2851136326789856,
    "learning_rate": 0.0009958580300552813,
    "timestep_mean": 508.5,
    "epsilon_norm": 6.746590614318848,
    "prediction_norm": 5.722084045410156,
    "timestamp": "2025-07-14T12:33:36.296780"
  },
  {
    "epoch": 82,
    "batch_idx": 192,
    "loss": 0.20947371423244476,
    "learning_rate": 0.0009958580300552813,
    "timestep_mean": 457.125,
    "epsilon_norm": 2.605660915374756,
    "prediction_norm": 1.955407738685608,
    "timestamp": "2025-07-14T12:33:36.301739"
  },
  {
    "epoch": 83,
    "batch_idx": 0,
    "loss": 0.36264362931251526,
    "learning_rate": 0.0009957565344100004,
    "timestep_mean": 478.71875,
    "epsilon_norm": 6.669804573059082,
    "prediction_norm": 4.662842750549316,
    "timestamp": "2025-07-14T12:33:36.310697"
  },
  {
    "epoch": 83,
    "batch_idx": 32,
    "loss": 0.6390745639801025,
    "learning_rate": 0.0009957565344100004,
    "timestep_mean": 389.03125,
    "epsilon_norm": 8.223284721374512,
    "prediction_norm": 4.902613162994385,
    "timestamp": "2025-07-14T12:33:36.315841"
  },
  {
    "epoch": 83,
    "batch_idx": 64,
    "loss": 0.2744235396385193,
    "learning_rate": 0.0009957565344100004,
    "timestep_mean": 473.59375,
    "epsilon_norm": 8.153011322021484,
    "prediction_norm": 6.4175004959106445,
    "timestamp": "2025-07-14T12:33:36.320939"
  },
  {
    "epoch": 83,
    "batch_idx": 96,
    "loss": 0.45101380348205566,
    "learning_rate": 0.0009957565344100004,
    "timestep_mean": 467.375,
    "epsilon_norm": 8.64987564086914,
    "prediction_norm": 5.827738285064697,
    "timestamp": "2025-07-14T12:33:36.326048"
  },
  {
    "epoch": 83,
    "batch_idx": 128,
    "loss": 0.434851735830307,
    "learning_rate": 0.0009957565344100004,
    "timestep_mean": 422.75,
    "epsilon_norm": 7.954872131347656,
    "prediction_norm": 5.646246433258057,
    "timestamp": "2025-07-14T12:33:36.331064"
  },
  {
    "epoch": 83,
    "batch_idx": 160,
    "loss": 0.44665423035621643,
    "learning_rate": 0.0009957565344100004,
    "timestep_mean": 397.0625,
    "epsilon_norm": 6.852954864501953,
    "prediction_norm": 4.13892126083374,
    "timestamp": "2025-07-14T12:33:36.336276"
  },
  {
    "epoch": 83,
    "batch_idx": 192,
    "loss": 0.23572084307670593,
    "learning_rate": 0.0009957565344100004,
    "timestep_mean": 619.625,
    "epsilon_norm": 3.638930320739746,
    "prediction_norm": 2.7683792114257812,
    "timestamp": "2025-07-14T12:33:36.341621"
  },
  {
    "epoch": 84,
    "batch_idx": 0,
    "loss": 0.2886325716972351,
    "learning_rate": 0.000995653815534753,
    "timestep_mean": 549.09375,
    "epsilon_norm": 7.990252494812012,
    "prediction_norm": 7.220165729522705,
    "timestamp": "2025-07-14T12:33:36.346800"
  },
  {
    "epoch": 84,
    "batch_idx": 32,
    "loss": 0.4047207832336426,
    "learning_rate": 0.000995653815534753,
    "timestep_mean": 452.125,
    "epsilon_norm": 9.245335578918457,
    "prediction_norm": 8.24968147277832,
    "timestamp": "2025-07-14T12:33:36.352063"
  },
  {
    "epoch": 84,
    "batch_idx": 64,
    "loss": 0.30003267526626587,
    "learning_rate": 0.000995653815534753,
    "timestep_mean": 531.375,
    "epsilon_norm": 7.752971172332764,
    "prediction_norm": 6.604158401489258,
    "timestamp": "2025-07-14T12:33:36.357140"
  },
  {
    "epoch": 84,
    "batch_idx": 96,
    "loss": 0.3684186339378357,
    "learning_rate": 0.000995653815534753,
    "timestep_mean": 501.34375,
    "epsilon_norm": 7.918209075927734,
    "prediction_norm": 7.054169654846191,
    "timestamp": "2025-07-14T12:33:36.362308"
  },
  {
    "epoch": 84,
    "batch_idx": 128,
    "loss": 0.4390615224838257,
    "learning_rate": 0.000995653815534753,
    "timestep_mean": 523.125,
    "epsilon_norm": 8.308943748474121,
    "prediction_norm": 7.004117965698242,
    "timestamp": "2025-07-14T12:33:36.367473"
  },
  {
    "epoch": 84,
    "batch_idx": 160,
    "loss": 0.4564509093761444,
    "learning_rate": 0.000995653815534753,
    "timestep_mean": 474.15625,
    "epsilon_norm": 8.380154609680176,
    "prediction_norm": 7.594479084014893,
    "timestamp": "2025-07-14T12:33:36.372558"
  },
  {
    "epoch": 84,
    "batch_idx": 192,
    "loss": 0.903453528881073,
    "learning_rate": 0.000995653815534753,
    "timestep_mean": 553.625,
    "epsilon_norm": 3.9197778701782227,
    "prediction_norm": 2.299614667892456,
    "timestamp": "2025-07-14T12:33:36.377751"
  },
  {
    "epoch": 85,
    "batch_idx": 0,
    "loss": 0.45032191276550293,
    "learning_rate": 0.0009955498736829872,
    "timestep_mean": 468.21875,
    "epsilon_norm": 8.020739555358887,
    "prediction_norm": 6.197582721710205,
    "timestamp": "2025-07-14T12:33:36.382909"
  },
  {
    "epoch": 85,
    "batch_idx": 32,
    "loss": 0.2587951421737671,
    "learning_rate": 0.0009955498736829872,
    "timestep_mean": 449.0625,
    "epsilon_norm": 7.578625679016113,
    "prediction_norm": 5.9930315017700195,
    "timestamp": "2025-07-14T12:33:36.388092"
  },
  {
    "epoch": 85,
    "batch_idx": 64,
    "loss": 0.40093764662742615,
    "learning_rate": 0.0009955498736829872,
    "timestep_mean": 457.1875,
    "epsilon_norm": 7.384009838104248,
    "prediction_norm": 5.577269554138184,
    "timestamp": "2025-07-14T12:33:36.393312"
  },
  {
    "epoch": 85,
    "batch_idx": 96,
    "loss": 0.5020498633384705,
    "learning_rate": 0.0009955498736829872,
    "timestep_mean": 407.34375,
    "epsilon_norm": 8.486074447631836,
    "prediction_norm": 6.129355430603027,
    "timestamp": "2025-07-14T12:33:36.398421"
  },
  {
    "epoch": 85,
    "batch_idx": 128,
    "loss": 0.22833094000816345,
    "learning_rate": 0.0009955498736829872,
    "timestep_mean": 454.53125,
    "epsilon_norm": 7.542516231536865,
    "prediction_norm": 5.987452507019043,
    "timestamp": "2025-07-14T12:33:36.403607"
  },
  {
    "epoch": 85,
    "batch_idx": 160,
    "loss": 0.2303440272808075,
    "learning_rate": 0.0009955498736829872,
    "timestep_mean": 600.09375,
    "epsilon_norm": 8.759512901306152,
    "prediction_norm": 7.446164131164551,
    "timestamp": "2025-07-14T12:33:36.408529"
  },
  {
    "epoch": 85,
    "batch_idx": 192,
    "loss": 0.8638477325439453,
    "learning_rate": 0.0009955498736829872,
    "timestep_mean": 425.125,
    "epsilon_norm": 3.4808080196380615,
    "prediction_norm": 1.4993802309036255,
    "timestamp": "2025-07-14T12:33:36.413385"
  },
  {
    "epoch": 86,
    "batch_idx": 0,
    "loss": 0.23734727501869202,
    "learning_rate": 0.0009954447091111691,
    "timestep_mean": 539.4375,
    "epsilon_norm": 7.495124816894531,
    "prediction_norm": 5.6216535568237305,
    "timestamp": "2025-07-14T12:33:36.418964"
  },
  {
    "epoch": 86,
    "batch_idx": 32,
    "loss": 0.4730536639690399,
    "learning_rate": 0.0009954447091111691,
    "timestep_mean": 442.96875,
    "epsilon_norm": 8.439220428466797,
    "prediction_norm": 5.634437561035156,
    "timestamp": "2025-07-14T12:33:36.424288"
  },
  {
    "epoch": 86,
    "batch_idx": 64,
    "loss": 0.49128252267837524,
    "learning_rate": 0.0009954447091111691,
    "timestep_mean": 484.375,
    "epsilon_norm": 8.91250991821289,
    "prediction_norm": 6.033281326293945,
    "timestamp": "2025-07-14T12:33:36.429555"
  },
  {
    "epoch": 86,
    "batch_idx": 96,
    "loss": 0.5532141923904419,
    "learning_rate": 0.0009954447091111691,
    "timestep_mean": 468.84375,
    "epsilon_norm": 8.326910018920898,
    "prediction_norm": 4.712965488433838,
    "timestamp": "2025-07-14T12:33:36.434641"
  },
  {
    "epoch": 86,
    "batch_idx": 128,
    "loss": 0.47127896547317505,
    "learning_rate": 0.0009954447091111691,
    "timestep_mean": 506.4375,
    "epsilon_norm": 7.499520778656006,
    "prediction_norm": 4.648345947265625,
    "timestamp": "2025-07-14T12:33:36.439678"
  },
  {
    "epoch": 86,
    "batch_idx": 160,
    "loss": 0.4244818687438965,
    "learning_rate": 0.0009954447091111691,
    "timestep_mean": 548.375,
    "epsilon_norm": 7.754502773284912,
    "prediction_norm": 6.0267767906188965,
    "timestamp": "2025-07-14T12:33:36.444881"
  },
  {
    "epoch": 86,
    "batch_idx": 192,
    "loss": 0.4885968565940857,
    "learning_rate": 0.0009954447091111691,
    "timestep_mean": 314.625,
    "epsilon_norm": 2.6450297832489014,
    "prediction_norm": 1.1289780139923096,
    "timestamp": "2025-07-14T12:33:36.450192"
  },
  {
    "epoch": 87,
    "batch_idx": 0,
    "loss": 0.48000869154930115,
    "learning_rate": 0.000995338322078782,
    "timestep_mean": 516.40625,
    "epsilon_norm": 7.501360893249512,
    "prediction_norm": 5.509984970092773,
    "timestamp": "2025-07-14T12:33:36.455296"
  },
  {
    "epoch": 87,
    "batch_idx": 32,
    "loss": 0.3923645317554474,
    "learning_rate": 0.000995338322078782,
    "timestep_mean": 483.65625,
    "epsilon_norm": 7.266393184661865,
    "prediction_norm": 5.873012065887451,
    "timestamp": "2025-07-14T12:33:36.460505"
  },
  {
    "epoch": 87,
    "batch_idx": 64,
    "loss": 0.3377560079097748,
    "learning_rate": 0.000995338322078782,
    "timestep_mean": 576.5625,
    "epsilon_norm": 8.1646728515625,
    "prediction_norm": 7.160357475280762,
    "timestamp": "2025-07-14T12:33:36.465670"
  },
  {
    "epoch": 87,
    "batch_idx": 96,
    "loss": 0.23903429508209229,
    "learning_rate": 0.000995338322078782,
    "timestep_mean": 535.3125,
    "epsilon_norm": 7.94504451751709,
    "prediction_norm": 7.214087963104248,
    "timestamp": "2025-07-14T12:33:36.470762"
  },
  {
    "epoch": 87,
    "batch_idx": 128,
    "loss": 0.3041566014289856,
    "learning_rate": 0.000995338322078782,
    "timestep_mean": 574.71875,
    "epsilon_norm": 8.814644813537598,
    "prediction_norm": 8.336444854736328,
    "timestamp": "2025-07-14T12:33:36.475655"
  },
  {
    "epoch": 87,
    "batch_idx": 160,
    "loss": 0.27848711609840393,
    "learning_rate": 0.000995338322078782,
    "timestep_mean": 558.78125,
    "epsilon_norm": 7.879665851593018,
    "prediction_norm": 6.820571422576904,
    "timestamp": "2025-07-14T12:33:36.480475"
  },
  {
    "epoch": 87,
    "batch_idx": 192,
    "loss": 0.15508852899074554,
    "learning_rate": 0.000995338322078782,
    "timestep_mean": 631.625,
    "epsilon_norm": 2.3842105865478516,
    "prediction_norm": 2.210392475128174,
    "timestamp": "2025-07-14T12:33:36.485312"
  },
  {
    "epoch": 88,
    "batch_idx": 0,
    "loss": 0.3392065167427063,
    "learning_rate": 0.0009952307128483252,
    "timestep_mean": 443.5,
    "epsilon_norm": 7.089841365814209,
    "prediction_norm": 5.406303405761719,
    "timestamp": "2025-07-14T12:33:36.490078"
  },
  {
    "epoch": 88,
    "batch_idx": 32,
    "loss": 0.22903195023536682,
    "learning_rate": 0.0009952307128483252,
    "timestep_mean": 495.96875,
    "epsilon_norm": 6.482456684112549,
    "prediction_norm": 5.347847938537598,
    "timestamp": "2025-07-14T12:33:36.494915"
  },
  {
    "epoch": 88,
    "batch_idx": 64,
    "loss": 0.3136409521102905,
    "learning_rate": 0.0009952307128483252,
    "timestep_mean": 547.03125,
    "epsilon_norm": 8.830754280090332,
    "prediction_norm": 7.172086238861084,
    "timestamp": "2025-07-14T12:33:36.499880"
  },
  {
    "epoch": 88,
    "batch_idx": 96,
    "loss": 0.40903371572494507,
    "learning_rate": 0.0009952307128483252,
    "timestep_mean": 521.625,
    "epsilon_norm": 7.65601921081543,
    "prediction_norm": 5.967591285705566,
    "timestamp": "2025-07-14T12:33:36.504704"
  },
  {
    "epoch": 88,
    "batch_idx": 128,
    "loss": 0.3422360122203827,
    "learning_rate": 0.0009952307128483252,
    "timestep_mean": 616.46875,
    "epsilon_norm": 8.147972106933594,
    "prediction_norm": 6.500747203826904,
    "timestamp": "2025-07-14T12:33:36.509586"
  },
  {
    "epoch": 88,
    "batch_idx": 160,
    "loss": 0.41183823347091675,
    "learning_rate": 0.0009952307128483252,
    "timestep_mean": 442.09375,
    "epsilon_norm": 7.46156644821167,
    "prediction_norm": 5.3514933586120605,
    "timestamp": "2025-07-14T12:33:36.514409"
  },
  {
    "epoch": 88,
    "batch_idx": 192,
    "loss": 0.08110179752111435,
    "learning_rate": 0.0009952307128483252,
    "timestep_mean": 609.375,
    "epsilon_norm": 4.319145202636719,
    "prediction_norm": 4.106696128845215,
    "timestamp": "2025-07-14T12:33:36.519331"
  },
  {
    "epoch": 89,
    "batch_idx": 0,
    "loss": 0.2782868444919586,
    "learning_rate": 0.000995121881685314,
    "timestep_mean": 552.6875,
    "epsilon_norm": 7.486007213592529,
    "prediction_norm": 6.207818984985352,
    "timestamp": "2025-07-14T12:33:36.524185"
  },
  {
    "epoch": 89,
    "batch_idx": 32,
    "loss": 0.48083049058914185,
    "learning_rate": 0.000995121881685314,
    "timestep_mean": 432.71875,
    "epsilon_norm": 8.880182266235352,
    "prediction_norm": 6.429576396942139,
    "timestamp": "2025-07-14T12:33:36.529113"
  },
  {
    "epoch": 89,
    "batch_idx": 64,
    "loss": 0.24283182621002197,
    "learning_rate": 0.000995121881685314,
    "timestep_mean": 501.0,
    "epsilon_norm": 7.541929721832275,
    "prediction_norm": 6.523571014404297,
    "timestamp": "2025-07-14T12:33:36.533924"
  },
  {
    "epoch": 89,
    "batch_idx": 96,
    "loss": 0.5025176405906677,
    "learning_rate": 0.000995121881685314,
    "timestep_mean": 474.28125,
    "epsilon_norm": 7.3117852210998535,
    "prediction_norm": 4.206206321716309,
    "timestamp": "2025-07-14T12:33:36.538757"
  },
  {
    "epoch": 89,
    "batch_idx": 128,
    "loss": 0.29628682136535645,
    "learning_rate": 0.000995121881685314,
    "timestep_mean": 552.96875,
    "epsilon_norm": 8.307861328125,
    "prediction_norm": 6.677133083343506,
    "timestamp": "2025-07-14T12:33:36.543593"
  },
  {
    "epoch": 89,
    "batch_idx": 160,
    "loss": 0.6291851997375488,
    "learning_rate": 0.000995121881685314,
    "timestep_mean": 407.75,
    "epsilon_norm": 8.218555450439453,
    "prediction_norm": 5.065184116363525,
    "timestamp": "2025-07-14T12:33:36.548445"
  },
  {
    "epoch": 89,
    "batch_idx": 192,
    "loss": 0.211685448884964,
    "learning_rate": 0.000995121881685314,
    "timestep_mean": 516.125,
    "epsilon_norm": 4.017679214477539,
    "prediction_norm": 3.255810260772705,
    "timestamp": "2025-07-14T12:33:36.553268"
  },
  {
    "epoch": 90,
    "batch_idx": 0,
    "loss": 0.44527244567871094,
    "learning_rate": 0.0009950118288582783,
    "timestep_mean": 479.34375,
    "epsilon_norm": 8.107833862304688,
    "prediction_norm": 7.2771806716918945,
    "timestamp": "2025-07-14T12:33:36.558227"
  },
  {
    "epoch": 90,
    "batch_idx": 32,
    "loss": 0.4024530351161957,
    "learning_rate": 0.0009950118288582783,
    "timestep_mean": 425.03125,
    "epsilon_norm": 7.545523166656494,
    "prediction_norm": 6.0275115966796875,
    "timestamp": "2025-07-14T12:33:36.563042"
  },
  {
    "epoch": 90,
    "batch_idx": 64,
    "loss": 0.24390769004821777,
    "learning_rate": 0.0009950118288582783,
    "timestep_mean": 478.8125,
    "epsilon_norm": 8.754861831665039,
    "prediction_norm": 7.889832019805908,
    "timestamp": "2025-07-14T12:33:36.567932"
  },
  {
    "epoch": 90,
    "batch_idx": 96,
    "loss": 0.2869304120540619,
    "learning_rate": 0.0009950118288582783,
    "timestep_mean": 607.375,
    "epsilon_norm": 9.43797779083252,
    "prediction_norm": 8.078756332397461,
    "timestamp": "2025-07-14T12:33:36.572761"
  },
  {
    "epoch": 90,
    "batch_idx": 128,
    "loss": 0.3250220715999603,
    "learning_rate": 0.0009950118288582783,
    "timestep_mean": 539.8125,
    "epsilon_norm": 9.068964004516602,
    "prediction_norm": 7.978925704956055,
    "timestamp": "2025-07-14T12:33:36.577622"
  },
  {
    "epoch": 90,
    "batch_idx": 160,
    "loss": 0.1980874091386795,
    "learning_rate": 0.0009950118288582783,
    "timestep_mean": 552.96875,
    "epsilon_norm": 7.871379852294922,
    "prediction_norm": 7.060305118560791,
    "timestamp": "2025-07-14T12:33:36.582494"
  },
  {
    "epoch": 90,
    "batch_idx": 192,
    "loss": 0.46973973512649536,
    "learning_rate": 0.0009950118288582783,
    "timestep_mean": 496.375,
    "epsilon_norm": 4.086599826812744,
    "prediction_norm": 3.2971580028533936,
    "timestamp": "2025-07-14T12:33:36.587415"
  },
  {
    "epoch": 91,
    "batch_idx": 0,
    "loss": 0.49183389544487,
    "learning_rate": 0.0009949005546387626,
    "timestep_mean": 405.09375,
    "epsilon_norm": 7.143603801727295,
    "prediction_norm": 4.913193702697754,
    "timestamp": "2025-07-14T12:33:36.592257"
  },
  {
    "epoch": 91,
    "batch_idx": 32,
    "loss": 0.2805904448032379,
    "learning_rate": 0.0009949005546387626,
    "timestep_mean": 529.90625,
    "epsilon_norm": 7.492563724517822,
    "prediction_norm": 6.141265392303467,
    "timestamp": "2025-07-14T12:33:36.597148"
  },
  {
    "epoch": 91,
    "batch_idx": 64,
    "loss": 0.3353915512561798,
    "learning_rate": 0.0009949005546387626,
    "timestep_mean": 549.75,
    "epsilon_norm": 8.002671241760254,
    "prediction_norm": 6.640048503875732,
    "timestamp": "2025-07-14T12:33:36.602026"
  },
  {
    "epoch": 91,
    "batch_idx": 96,
    "loss": 0.6279861927032471,
    "learning_rate": 0.0009949005546387626,
    "timestep_mean": 446.15625,
    "epsilon_norm": 8.974250793457031,
    "prediction_norm": 5.561717987060547,
    "timestamp": "2025-07-14T12:33:36.606897"
  },
  {
    "epoch": 91,
    "batch_idx": 128,
    "loss": 0.4289284646511078,
    "learning_rate": 0.0009949005546387626,
    "timestep_mean": 476.125,
    "epsilon_norm": 9.361955642700195,
    "prediction_norm": 7.650237083435059,
    "timestamp": "2025-07-14T12:33:36.611726"
  },
  {
    "epoch": 91,
    "batch_idx": 160,
    "loss": 0.27840864658355713,
    "learning_rate": 0.0009949005546387626,
    "timestep_mean": 523.875,
    "epsilon_norm": 6.827390670776367,
    "prediction_norm": 5.433884620666504,
    "timestamp": "2025-07-14T12:33:36.616645"
  },
  {
    "epoch": 91,
    "batch_idx": 192,
    "loss": 0.07109589874744415,
    "learning_rate": 0.0009949005546387626,
    "timestep_mean": 546.375,
    "epsilon_norm": 4.231659412384033,
    "prediction_norm": 4.6579108238220215,
    "timestamp": "2025-07-14T12:33:36.621569"
  },
  {
    "epoch": 92,
    "batch_idx": 0,
    "loss": 0.31413692235946655,
    "learning_rate": 0.000994788059301325,
    "timestep_mean": 536.8125,
    "epsilon_norm": 7.7336039543151855,
    "prediction_norm": 6.467280387878418,
    "timestamp": "2025-07-14T12:33:36.626471"
  },
  {
    "epoch": 92,
    "batch_idx": 32,
    "loss": 0.20682522654533386,
    "learning_rate": 0.000994788059301325,
    "timestep_mean": 567.90625,
    "epsilon_norm": 7.99931001663208,
    "prediction_norm": 7.192995548248291,
    "timestamp": "2025-07-14T12:33:36.631361"
  },
  {
    "epoch": 92,
    "batch_idx": 64,
    "loss": 0.42646878957748413,
    "learning_rate": 0.000994788059301325,
    "timestep_mean": 507.21875,
    "epsilon_norm": 8.76132583618164,
    "prediction_norm": 6.98457670211792,
    "timestamp": "2025-07-14T12:33:36.636274"
  },
  {
    "epoch": 92,
    "batch_idx": 96,
    "loss": 0.2990836501121521,
    "learning_rate": 0.000994788059301325,
    "timestep_mean": 413.03125,
    "epsilon_norm": 8.345949172973633,
    "prediction_norm": 7.236357688903809,
    "timestamp": "2025-07-14T12:33:36.641151"
  },
  {
    "epoch": 92,
    "batch_idx": 128,
    "loss": 0.44317689538002014,
    "learning_rate": 0.000994788059301325,
    "timestep_mean": 542.3125,
    "epsilon_norm": 8.677595138549805,
    "prediction_norm": 6.477882385253906,
    "timestamp": "2025-07-14T12:33:36.646284"
  },
  {
    "epoch": 92,
    "batch_idx": 160,
    "loss": 0.3834819197654724,
    "learning_rate": 0.000994788059301325,
    "timestep_mean": 463.0,
    "epsilon_norm": 7.988166809082031,
    "prediction_norm": 7.2041425704956055,
    "timestamp": "2025-07-14T12:33:36.651137"
  },
  {
    "epoch": 92,
    "batch_idx": 192,
    "loss": 0.39663293957710266,
    "learning_rate": 0.000994788059301325,
    "timestep_mean": 409.625,
    "epsilon_norm": 3.3966057300567627,
    "prediction_norm": 2.3062617778778076,
    "timestamp": "2025-07-14T12:33:36.655975"
  },
  {
    "epoch": 93,
    "batch_idx": 0,
    "loss": 0.3363357484340668,
    "learning_rate": 0.0009946743431235366,
    "timestep_mean": 510.40625,
    "epsilon_norm": 7.8334245681762695,
    "prediction_norm": 6.377856731414795,
    "timestamp": "2025-07-14T12:33:36.660832"
  },
  {
    "epoch": 93,
    "batch_idx": 32,
    "loss": 0.32007676362991333,
    "learning_rate": 0.0009946743431235366,
    "timestep_mean": 499.53125,
    "epsilon_norm": 7.7155985832214355,
    "prediction_norm": 6.513158798217773,
    "timestamp": "2025-07-14T12:33:36.665696"
  },
  {
    "epoch": 93,
    "batch_idx": 64,
    "loss": 0.34587690234184265,
    "learning_rate": 0.0009946743431235366,
    "timestep_mean": 437.71875,
    "epsilon_norm": 8.133333206176758,
    "prediction_norm": 7.033232688903809,
    "timestamp": "2025-07-14T12:33:36.670506"
  },
  {
    "epoch": 93,
    "batch_idx": 96,
    "loss": 0.4252506494522095,
    "learning_rate": 0.0009946743431235366,
    "timestep_mean": 439.9375,
    "epsilon_norm": 7.717460632324219,
    "prediction_norm": 5.934478282928467,
    "timestamp": "2025-07-14T12:33:36.675381"
  },
  {
    "epoch": 93,
    "batch_idx": 128,
    "loss": 0.4699650704860687,
    "learning_rate": 0.0009946743431235366,
    "timestep_mean": 504.4375,
    "epsilon_norm": 7.854462146759033,
    "prediction_norm": 4.853221893310547,
    "timestamp": "2025-07-14T12:33:36.680273"
  },
  {
    "epoch": 93,
    "batch_idx": 160,
    "loss": 0.3393353223800659,
    "learning_rate": 0.0009946743431235366,
    "timestep_mean": 453.125,
    "epsilon_norm": 7.951918601989746,
    "prediction_norm": 6.049301624298096,
    "timestamp": "2025-07-14T12:33:36.685109"
  },
  {
    "epoch": 93,
    "batch_idx": 192,
    "loss": 0.5024043321609497,
    "learning_rate": 0.0009946743431235366,
    "timestep_mean": 337.25,
    "epsilon_norm": 3.881716251373291,
    "prediction_norm": 2.3281760215759277,
    "timestamp": "2025-07-14T12:33:36.690015"
  },
  {
    "epoch": 94,
    "batch_idx": 0,
    "loss": 0.47741395235061646,
    "learning_rate": 0.0009945594063859805,
    "timestep_mean": 456.84375,
    "epsilon_norm": 9.171212196350098,
    "prediction_norm": 6.063540458679199,
    "timestamp": "2025-07-14T12:33:36.694930"
  },
  {
    "epoch": 94,
    "batch_idx": 32,
    "loss": 0.3556879758834839,
    "learning_rate": 0.0009945594063859805,
    "timestep_mean": 456.71875,
    "epsilon_norm": 7.539243698120117,
    "prediction_norm": 5.729331970214844,
    "timestamp": "2025-07-14T12:33:36.699956"
  },
  {
    "epoch": 94,
    "batch_idx": 64,
    "loss": 0.4314684271812439,
    "learning_rate": 0.0009945594063859805,
    "timestep_mean": 427.375,
    "epsilon_norm": 8.018998146057129,
    "prediction_norm": 5.620591163635254,
    "timestamp": "2025-07-14T12:33:36.704870"
  },
  {
    "epoch": 94,
    "batch_idx": 96,
    "loss": 0.22523263096809387,
    "learning_rate": 0.0009945594063859805,
    "timestep_mean": 539.0,
    "epsilon_norm": 6.939995765686035,
    "prediction_norm": 5.757425785064697,
    "timestamp": "2025-07-14T12:33:36.709773"
  },
  {
    "epoch": 94,
    "batch_idx": 128,
    "loss": 0.28931206464767456,
    "learning_rate": 0.0009945594063859805,
    "timestep_mean": 550.03125,
    "epsilon_norm": 7.407326698303223,
    "prediction_norm": 6.067388534545898,
    "timestamp": "2025-07-14T12:33:36.714652"
  },
  {
    "epoch": 94,
    "batch_idx": 160,
    "loss": 0.5652316808700562,
    "learning_rate": 0.0009945594063859805,
    "timestep_mean": 447.65625,
    "epsilon_norm": 7.447558403015137,
    "prediction_norm": 4.639014720916748,
    "timestamp": "2025-07-14T12:33:36.719529"
  },
  {
    "epoch": 94,
    "batch_idx": 192,
    "loss": 0.3043774366378784,
    "learning_rate": 0.0009945594063859805,
    "timestep_mean": 563.75,
    "epsilon_norm": 4.028321743011475,
    "prediction_norm": 3.5328738689422607,
    "timestamp": "2025-07-14T12:33:36.724433"
  },
  {
    "epoch": 95,
    "batch_idx": 0,
    "loss": 0.5101208686828613,
    "learning_rate": 0.000994443249372252,
    "timestep_mean": 448.65625,
    "epsilon_norm": 7.696125030517578,
    "prediction_norm": 5.724435329437256,
    "timestamp": "2025-07-14T12:33:36.729384"
  },
  {
    "epoch": 95,
    "batch_idx": 32,
    "loss": 0.567093014717102,
    "learning_rate": 0.000994443249372252,
    "timestep_mean": 447.125,
    "epsilon_norm": 8.16093921661377,
    "prediction_norm": 5.92333984375,
    "timestamp": "2025-07-14T12:33:36.734242"
  },
  {
    "epoch": 95,
    "batch_idx": 64,
    "loss": 0.5548479557037354,
    "learning_rate": 0.000994443249372252,
    "timestep_mean": 398.875,
    "epsilon_norm": 7.260884761810303,
    "prediction_norm": 5.162304401397705,
    "timestamp": "2025-07-14T12:33:36.739115"
  },
  {
    "epoch": 95,
    "batch_idx": 96,
    "loss": 0.5949661135673523,
    "learning_rate": 0.000994443249372252,
    "timestep_mean": 442.0625,
    "epsilon_norm": 8.13223934173584,
    "prediction_norm": 5.406032562255859,
    "timestamp": "2025-07-14T12:33:36.744018"
  },
  {
    "epoch": 95,
    "batch_idx": 128,
    "loss": 0.3970436453819275,
    "learning_rate": 0.000994443249372252,
    "timestep_mean": 449.3125,
    "epsilon_norm": 7.762072563171387,
    "prediction_norm": 6.095621585845947,
    "timestamp": "2025-07-14T12:33:36.748906"
  },
  {
    "epoch": 95,
    "batch_idx": 160,
    "loss": 0.4220578372478485,
    "learning_rate": 0.000994443249372252,
    "timestep_mean": 472.96875,
    "epsilon_norm": 8.612730979919434,
    "prediction_norm": 6.902568340301514,
    "timestamp": "2025-07-14T12:33:36.753816"
  },
  {
    "epoch": 95,
    "batch_idx": 192,
    "loss": 0.5651695728302002,
    "learning_rate": 0.000994443249372252,
    "timestep_mean": 418.125,
    "epsilon_norm": 4.418335437774658,
    "prediction_norm": 2.792278289794922,
    "timestamp": "2025-07-14T12:33:36.758699"
  },
  {
    "epoch": 96,
    "batch_idx": 0,
    "loss": 0.5334373712539673,
    "learning_rate": 0.0009943258723689567,
    "timestep_mean": 443.96875,
    "epsilon_norm": 7.684787273406982,
    "prediction_norm": 4.827340602874756,
    "timestamp": "2025-07-14T12:33:36.763601"
  },
  {
    "epoch": 96,
    "batch_idx": 32,
    "loss": 0.3576987087726593,
    "learning_rate": 0.0009943258723689567,
    "timestep_mean": 533.90625,
    "epsilon_norm": 7.566948890686035,
    "prediction_norm": 5.774721145629883,
    "timestamp": "2025-07-14T12:33:36.768534"
  },
  {
    "epoch": 96,
    "batch_idx": 64,
    "loss": 0.452669620513916,
    "learning_rate": 0.0009943258723689567,
    "timestep_mean": 478.90625,
    "epsilon_norm": 8.196468353271484,
    "prediction_norm": 5.744713306427002,
    "timestamp": "2025-07-14T12:33:36.773545"
  },
  {
    "epoch": 96,
    "batch_idx": 96,
    "loss": 0.3610934913158417,
    "learning_rate": 0.0009943258723689567,
    "timestep_mean": 574.75,
    "epsilon_norm": 8.204984664916992,
    "prediction_norm": 6.421931266784668,
    "timestamp": "2025-07-14T12:33:36.778443"
  },
  {
    "epoch": 96,
    "batch_idx": 128,
    "loss": 0.6469030976295471,
    "learning_rate": 0.0009943258723689567,
    "timestep_mean": 389.09375,
    "epsilon_norm": 8.607786178588867,
    "prediction_norm": 5.982041358947754,
    "timestamp": "2025-07-14T12:33:36.783470"
  },
  {
    "epoch": 96,
    "batch_idx": 160,
    "loss": 0.38033175468444824,
    "learning_rate": 0.0009943258723689567,
    "timestep_mean": 576.9375,
    "epsilon_norm": 7.537169456481934,
    "prediction_norm": 6.028802871704102,
    "timestamp": "2025-07-14T12:33:36.788383"
  },
  {
    "epoch": 96,
    "batch_idx": 192,
    "loss": 0.3279222249984741,
    "learning_rate": 0.0009943258723689567,
    "timestep_mean": 431.25,
    "epsilon_norm": 3.6978209018707275,
    "prediction_norm": 2.972092866897583,
    "timestamp": "2025-07-14T12:33:36.793252"
  },
  {
    "epoch": 97,
    "batch_idx": 0,
    "loss": 0.2552438974380493,
    "learning_rate": 0.000994207275665711,
    "timestep_mean": 508.71875,
    "epsilon_norm": 7.756275177001953,
    "prediction_norm": 6.287402629852295,
    "timestamp": "2025-07-14T12:33:36.798187"
  },
  {
    "epoch": 97,
    "batch_idx": 32,
    "loss": 0.473777711391449,
    "learning_rate": 0.000994207275665711,
    "timestep_mean": 456.875,
    "epsilon_norm": 8.601751327514648,
    "prediction_norm": 6.151224613189697,
    "timestamp": "2025-07-14T12:33:36.803214"
  },
  {
    "epoch": 97,
    "batch_idx": 64,
    "loss": 0.36334842443466187,
    "learning_rate": 0.000994207275665711,
    "timestep_mean": 518.34375,
    "epsilon_norm": 9.001164436340332,
    "prediction_norm": 7.66635274887085,
    "timestamp": "2025-07-14T12:33:36.808234"
  },
  {
    "epoch": 97,
    "batch_idx": 96,
    "loss": 0.25658372044563293,
    "learning_rate": 0.000994207275665711,
    "timestep_mean": 501.03125,
    "epsilon_norm": 7.184560775756836,
    "prediction_norm": 5.912685871124268,
    "timestamp": "2025-07-14T12:33:36.813120"
  },
  {
    "epoch": 97,
    "batch_idx": 128,
    "loss": 0.20777863264083862,
    "learning_rate": 0.000994207275665711,
    "timestep_mean": 531.09375,
    "epsilon_norm": 6.85400390625,
    "prediction_norm": 6.252521514892578,
    "timestamp": "2025-07-14T12:33:36.818073"
  },
  {
    "epoch": 97,
    "batch_idx": 160,
    "loss": 0.189926877617836,
    "learning_rate": 0.000994207275665711,
    "timestep_mean": 602.375,
    "epsilon_norm": 7.205810546875,
    "prediction_norm": 6.894397735595703,
    "timestamp": "2025-07-14T12:33:36.823117"
  },
  {
    "epoch": 97,
    "batch_idx": 192,
    "loss": 0.45718544721603394,
    "learning_rate": 0.000994207275665711,
    "timestep_mean": 471.875,
    "epsilon_norm": 4.028402328491211,
    "prediction_norm": 2.786032199859619,
    "timestamp": "2025-07-14T12:33:36.828276"
  },
  {
    "epoch": 98,
    "batch_idx": 0,
    "loss": 0.432761013507843,
    "learning_rate": 0.00099408745955514,
    "timestep_mean": 519.25,
    "epsilon_norm": 8.67878246307373,
    "prediction_norm": 7.009236812591553,
    "timestamp": "2025-07-14T12:33:36.833559"
  },
  {
    "epoch": 98,
    "batch_idx": 32,
    "loss": 0.3440079391002655,
    "learning_rate": 0.00099408745955514,
    "timestep_mean": 526.84375,
    "epsilon_norm": 7.524142742156982,
    "prediction_norm": 6.707104206085205,
    "timestamp": "2025-07-14T12:33:36.838669"
  },
  {
    "epoch": 98,
    "batch_idx": 64,
    "loss": 0.32256221771240234,
    "learning_rate": 0.00099408745955514,
    "timestep_mean": 531.40625,
    "epsilon_norm": 7.025570392608643,
    "prediction_norm": 5.700721740722656,
    "timestamp": "2025-07-14T12:33:36.843937"
  },
  {
    "epoch": 98,
    "batch_idx": 96,
    "loss": 0.366989403963089,
    "learning_rate": 0.00099408745955514,
    "timestep_mean": 558.1875,
    "epsilon_norm": 8.505144119262695,
    "prediction_norm": 7.46346378326416,
    "timestamp": "2025-07-14T12:33:36.849216"
  },
  {
    "epoch": 98,
    "batch_idx": 128,
    "loss": 0.19330787658691406,
    "learning_rate": 0.00099408745955514,
    "timestep_mean": 563.84375,
    "epsilon_norm": 6.875184059143066,
    "prediction_norm": 6.3273606300354,
    "timestamp": "2025-07-14T12:33:36.854348"
  },
  {
    "epoch": 98,
    "batch_idx": 160,
    "loss": 0.5487728118896484,
    "learning_rate": 0.00099408745955514,
    "timestep_mean": 466.3125,
    "epsilon_norm": 8.453117370605469,
    "prediction_norm": 5.889194488525391,
    "timestamp": "2025-07-14T12:33:36.859518"
  },
  {
    "epoch": 98,
    "batch_idx": 192,
    "loss": 0.8230843544006348,
    "learning_rate": 0.00099408745955514,
    "timestep_mean": 434.125,
    "epsilon_norm": 4.247632026672363,
    "prediction_norm": 2.4748427867889404,
    "timestamp": "2025-07-14T12:33:36.864628"
  },
  {
    "epoch": 99,
    "batch_idx": 0,
    "loss": 0.2608836889266968,
    "learning_rate": 0.0009939664243328785,
    "timestep_mean": 473.5,
    "epsilon_norm": 6.714778900146484,
    "prediction_norm": 5.409607410430908,
    "timestamp": "2025-07-14T12:33:36.869802"
  },
  {
    "epoch": 99,
    "batch_idx": 32,
    "loss": 0.3959690034389496,
    "learning_rate": 0.0009939664243328785,
    "timestep_mean": 531.09375,
    "epsilon_norm": 8.920614242553711,
    "prediction_norm": 7.156506538391113,
    "timestamp": "2025-07-14T12:33:36.875259"
  },
  {
    "epoch": 99,
    "batch_idx": 64,
    "loss": 0.5295666456222534,
    "learning_rate": 0.0009939664243328785,
    "timestep_mean": 463.71875,
    "epsilon_norm": 7.252027988433838,
    "prediction_norm": 4.903682231903076,
    "timestamp": "2025-07-14T12:33:36.880351"
  },
  {
    "epoch": 99,
    "batch_idx": 96,
    "loss": 0.1751599907875061,
    "learning_rate": 0.0009939664243328785,
    "timestep_mean": 494.6875,
    "epsilon_norm": 6.158676624298096,
    "prediction_norm": 5.199674129486084,
    "timestamp": "2025-07-14T12:33:36.885535"
  },
  {
    "epoch": 99,
    "batch_idx": 128,
    "loss": 0.4449681341648102,
    "learning_rate": 0.0009939664243328785,
    "timestep_mean": 409.03125,
    "epsilon_norm": 7.886930465698242,
    "prediction_norm": 5.021754264831543,
    "timestamp": "2025-07-14T12:33:36.890656"
  },
  {
    "epoch": 99,
    "batch_idx": 160,
    "loss": 0.7398954033851624,
    "learning_rate": 0.0009939664243328785,
    "timestep_mean": 458.28125,
    "epsilon_norm": 9.626720428466797,
    "prediction_norm": 5.695218563079834,
    "timestamp": "2025-07-14T12:33:36.901189"
  },
  {
    "epoch": 99,
    "batch_idx": 192,
    "loss": 0.36504319310188293,
    "learning_rate": 0.0009939664243328785,
    "timestep_mean": 568.125,
    "epsilon_norm": 3.821928024291992,
    "prediction_norm": 2.758923292160034,
    "timestamp": "2025-07-14T12:33:36.906334"
  },
  {
    "epoch": 100,
    "batch_idx": 0,
    "loss": 0.2772006392478943,
    "learning_rate": 0.0009938441702975686,
    "timestep_mean": 531.46875,
    "epsilon_norm": 8.723272323608398,
    "prediction_norm": 7.111536026000977,
    "timestamp": "2025-07-14T12:33:36.911486"
  },
  {
    "epoch": 100,
    "batch_idx": 32,
    "loss": 0.2712431252002716,
    "learning_rate": 0.0009938441702975686,
    "timestep_mean": 572.375,
    "epsilon_norm": 9.039788246154785,
    "prediction_norm": 7.510835647583008,
    "timestamp": "2025-07-14T12:33:36.916550"
  },
  {
    "epoch": 100,
    "batch_idx": 64,
    "loss": 0.46679916977882385,
    "learning_rate": 0.0009938441702975686,
    "timestep_mean": 359.78125,
    "epsilon_norm": 7.268196105957031,
    "prediction_norm": 4.211639404296875,
    "timestamp": "2025-07-14T12:33:36.921555"
  },
  {
    "epoch": 100,
    "batch_idx": 96,
    "loss": 0.38940954208374023,
    "learning_rate": 0.0009938441702975686,
    "timestep_mean": 433.125,
    "epsilon_norm": 8.101743698120117,
    "prediction_norm": 6.207546234130859,
    "timestamp": "2025-07-14T12:33:36.926564"
  },
  {
    "epoch": 100,
    "batch_idx": 128,
    "loss": 0.3505213260650635,
    "learning_rate": 0.0009938441702975686,
    "timestep_mean": 517.125,
    "epsilon_norm": 8.347238540649414,
    "prediction_norm": 7.680215358734131,
    "timestamp": "2025-07-14T12:33:36.931554"
  },
  {
    "epoch": 100,
    "batch_idx": 160,
    "loss": 0.35721033811569214,
    "learning_rate": 0.0009938441702975686,
    "timestep_mean": 528.25,
    "epsilon_norm": 8.35438346862793,
    "prediction_norm": 6.6481828689575195,
    "timestamp": "2025-07-14T12:33:36.936766"
  },
  {
    "epoch": 100,
    "batch_idx": 192,
    "loss": 0.7355152368545532,
    "learning_rate": 0.0009938441702975686,
    "timestep_mean": 188.625,
    "epsilon_norm": 3.998323440551758,
    "prediction_norm": 1.922577977180481,
    "timestamp": "2025-07-14T12:33:36.942014"
  },
  {
    "epoch": 101,
    "batch_idx": 0,
    "loss": 0.5455557107925415,
    "learning_rate": 0.0009937206977508601,
    "timestep_mean": 422.5625,
    "epsilon_norm": 7.587577819824219,
    "prediction_norm": 5.869652271270752,
    "timestamp": "2025-07-14T12:33:36.947028"
  },
  {
    "epoch": 101,
    "batch_idx": 32,
    "loss": 0.511725127696991,
    "learning_rate": 0.0009937206977508601,
    "timestep_mean": 483.125,
    "epsilon_norm": 9.029644012451172,
    "prediction_norm": 7.481060981750488,
    "timestamp": "2025-07-14T12:33:36.952280"
  },
  {
    "epoch": 101,
    "batch_idx": 64,
    "loss": 0.29942673444747925,
    "learning_rate": 0.0009937206977508601,
    "timestep_mean": 517.9375,
    "epsilon_norm": 7.221333026885986,
    "prediction_norm": 6.712826251983643,
    "timestamp": "2025-07-14T12:33:36.957438"
  },
  {
    "epoch": 101,
    "batch_idx": 96,
    "loss": 0.37993139028549194,
    "learning_rate": 0.0009937206977508601,
    "timestep_mean": 533.6875,
    "epsilon_norm": 7.221596717834473,
    "prediction_norm": 5.908322811126709,
    "timestamp": "2025-07-14T12:33:36.962736"
  },
  {
    "epoch": 101,
    "batch_idx": 128,
    "loss": 0.352159321308136,
    "learning_rate": 0.0009937206977508601,
    "timestep_mean": 501.625,
    "epsilon_norm": 8.534143447875977,
    "prediction_norm": 8.126705169677734,
    "timestamp": "2025-07-14T12:33:36.967784"
  },
  {
    "epoch": 101,
    "batch_idx": 160,
    "loss": 0.1860753744840622,
    "learning_rate": 0.0009937206977508601,
    "timestep_mean": 528.0625,
    "epsilon_norm": 6.751574993133545,
    "prediction_norm": 6.7020463943481445,
    "timestamp": "2025-07-14T12:33:36.972782"
  },
  {
    "epoch": 101,
    "batch_idx": 192,
    "loss": 0.5374491810798645,
    "learning_rate": 0.0009937206977508601,
    "timestep_mean": 531.625,
    "epsilon_norm": 3.499622106552124,
    "prediction_norm": 2.770064353942871,
    "timestamp": "2025-07-14T12:33:36.977988"
  },
  {
    "epoch": 102,
    "batch_idx": 0,
    "loss": 0.2631532847881317,
    "learning_rate": 0.0009935960069974094,
    "timestep_mean": 590.0,
    "epsilon_norm": 8.167551040649414,
    "prediction_norm": 6.895909786224365,
    "timestamp": "2025-07-14T12:33:36.983384"
  },
  {
    "epoch": 102,
    "batch_idx": 32,
    "loss": 0.334185928106308,
    "learning_rate": 0.0009935960069974094,
    "timestep_mean": 529.625,
    "epsilon_norm": 7.469237804412842,
    "prediction_norm": 5.963769435882568,
    "timestamp": "2025-07-14T12:33:36.988507"
  },
  {
    "epoch": 102,
    "batch_idx": 64,
    "loss": 0.3564962148666382,
    "learning_rate": 0.0009935960069974094,
    "timestep_mean": 437.09375,
    "epsilon_norm": 8.210273742675781,
    "prediction_norm": 5.989192962646484,
    "timestamp": "2025-07-14T12:33:36.993648"
  },
  {
    "epoch": 102,
    "batch_idx": 96,
    "loss": 0.36092546582221985,
    "learning_rate": 0.0009935960069974094,
    "timestep_mean": 594.8125,
    "epsilon_norm": 8.531179428100586,
    "prediction_norm": 6.3861823081970215,
    "timestamp": "2025-07-14T12:33:36.998654"
  },
  {
    "epoch": 102,
    "batch_idx": 128,
    "loss": 0.2820889949798584,
    "learning_rate": 0.0009935960069974094,
    "timestep_mean": 502.875,
    "epsilon_norm": 8.821943283081055,
    "prediction_norm": 6.406214714050293,
    "timestamp": "2025-07-14T12:33:37.003788"
  },
  {
    "epoch": 102,
    "batch_idx": 160,
    "loss": 0.3700854182243347,
    "learning_rate": 0.0009935960069974094,
    "timestep_mean": 512.46875,
    "epsilon_norm": 7.914233207702637,
    "prediction_norm": 5.285330295562744,
    "timestamp": "2025-07-14T12:33:37.008799"
  },
  {
    "epoch": 102,
    "batch_idx": 192,
    "loss": 0.7445805072784424,
    "learning_rate": 0.0009935960069974094,
    "timestep_mean": 366.375,
    "epsilon_norm": 3.8970069885253906,
    "prediction_norm": 1.8015291690826416,
    "timestamp": "2025-07-14T12:33:37.013808"
  },
  {
    "epoch": 103,
    "batch_idx": 0,
    "loss": 0.3264533281326294,
    "learning_rate": 0.0009934700983448782,
    "timestep_mean": 569.71875,
    "epsilon_norm": 8.948659896850586,
    "prediction_norm": 6.809743881225586,
    "timestamp": "2025-07-14T12:33:37.018992"
  },
  {
    "epoch": 103,
    "batch_idx": 32,
    "loss": 0.46428436040878296,
    "learning_rate": 0.0009934700983448782,
    "timestep_mean": 423.1875,
    "epsilon_norm": 7.752265930175781,
    "prediction_norm": 5.595555782318115,
    "timestamp": "2025-07-14T12:33:37.024211"
  },
  {
    "epoch": 103,
    "batch_idx": 64,
    "loss": 0.32817569375038147,
    "learning_rate": 0.0009934700983448782,
    "timestep_mean": 574.78125,
    "epsilon_norm": 7.931771278381348,
    "prediction_norm": 5.9285125732421875,
    "timestamp": "2025-07-14T12:33:37.029421"
  },
  {
    "epoch": 103,
    "batch_idx": 96,
    "loss": 0.34742236137390137,
    "learning_rate": 0.0009934700983448782,
    "timestep_mean": 452.09375,
    "epsilon_norm": 7.735149383544922,
    "prediction_norm": 5.68527889251709,
    "timestamp": "2025-07-14T12:33:37.034603"
  },
  {
    "epoch": 103,
    "batch_idx": 128,
    "loss": 0.21466407179832458,
    "learning_rate": 0.0009934700983448782,
    "timestep_mean": 497.6875,
    "epsilon_norm": 6.843139171600342,
    "prediction_norm": 6.070842742919922,
    "timestamp": "2025-07-14T12:33:37.039513"
  },
  {
    "epoch": 103,
    "batch_idx": 160,
    "loss": 0.2889866232872009,
    "learning_rate": 0.0009934700983448782,
    "timestep_mean": 510.6875,
    "epsilon_norm": 8.854406356811523,
    "prediction_norm": 7.18943977355957,
    "timestamp": "2025-07-14T12:33:37.044686"
  },
  {
    "epoch": 103,
    "batch_idx": 192,
    "loss": 1.5483696460723877,
    "learning_rate": 0.0009934700983448782,
    "timestep_mean": 406.375,
    "epsilon_norm": 5.771408557891846,
    "prediction_norm": 2.7569661140441895,
    "timestamp": "2025-07-14T12:33:37.049950"
  },
  {
    "epoch": 104,
    "batch_idx": 0,
    "loss": 0.5047998428344727,
    "learning_rate": 0.0009933429721039337,
    "timestep_mean": 471.78125,
    "epsilon_norm": 8.299660682678223,
    "prediction_norm": 6.754024028778076,
    "timestamp": "2025-07-14T12:33:37.055120"
  },
  {
    "epoch": 104,
    "batch_idx": 32,
    "loss": 0.49077308177948,
    "learning_rate": 0.0009933429721039337,
    "timestep_mean": 433.5625,
    "epsilon_norm": 8.774603843688965,
    "prediction_norm": 6.2383503913879395,
    "timestamp": "2025-07-14T12:33:37.060279"
  },
  {
    "epoch": 104,
    "batch_idx": 64,
    "loss": 0.39925307035446167,
    "learning_rate": 0.0009933429721039337,
    "timestep_mean": 522.15625,
    "epsilon_norm": 7.907754421234131,
    "prediction_norm": 7.326230049133301,
    "timestamp": "2025-07-14T12:33:37.065399"
  },
  {
    "epoch": 104,
    "batch_idx": 96,
    "loss": 0.4939652681350708,
    "learning_rate": 0.0009933429721039337,
    "timestep_mean": 432.46875,
    "epsilon_norm": 6.765961170196533,
    "prediction_norm": 5.050776958465576,
    "timestamp": "2025-07-14T12:33:37.070613"
  },
  {
    "epoch": 104,
    "batch_idx": 128,
    "loss": 0.31124359369277954,
    "learning_rate": 0.0009933429721039337,
    "timestep_mean": 542.65625,
    "epsilon_norm": 9.177377700805664,
    "prediction_norm": 9.203100204467773,
    "timestamp": "2025-07-14T12:33:37.075798"
  },
  {
    "epoch": 104,
    "batch_idx": 160,
    "loss": 0.20461241900920868,
    "learning_rate": 0.0009933429721039337,
    "timestep_mean": 549.46875,
    "epsilon_norm": 7.330206871032715,
    "prediction_norm": 7.184663772583008,
    "timestamp": "2025-07-14T12:33:37.080847"
  },
  {
    "epoch": 104,
    "batch_idx": 192,
    "loss": 0.3324352502822876,
    "learning_rate": 0.0009933429721039337,
    "timestep_mean": 508.5,
    "epsilon_norm": 4.53663444519043,
    "prediction_norm": 3.9737908840179443,
    "timestamp": "2025-07-14T12:33:37.086026"
  },
  {
    "epoch": 105,
    "batch_idx": 0,
    "loss": 0.29668140411376953,
    "learning_rate": 0.0009932146285882475,
    "timestep_mean": 531.125,
    "epsilon_norm": 8.309447288513184,
    "prediction_norm": 6.850283622741699,
    "timestamp": "2025-07-14T12:33:37.091192"
  },
  {
    "epoch": 105,
    "batch_idx": 32,
    "loss": 0.31115037202835083,
    "learning_rate": 0.0009932146285882475,
    "timestep_mean": 477.03125,
    "epsilon_norm": 7.698022842407227,
    "prediction_norm": 6.149684906005859,
    "timestamp": "2025-07-14T12:33:37.096516"
  },
  {
    "epoch": 105,
    "batch_idx": 64,
    "loss": 0.3817676305770874,
    "learning_rate": 0.0009932146285882475,
    "timestep_mean": 453.71875,
    "epsilon_norm": 6.825430870056152,
    "prediction_norm": 4.7921247482299805,
    "timestamp": "2025-07-14T12:33:37.101681"
  },
  {
    "epoch": 105,
    "batch_idx": 96,
    "loss": 0.2946900725364685,
    "learning_rate": 0.0009932146285882475,
    "timestep_mean": 487.15625,
    "epsilon_norm": 6.930069446563721,
    "prediction_norm": 5.104409217834473,
    "timestamp": "2025-07-14T12:33:37.106771"
  },
  {
    "epoch": 105,
    "batch_idx": 128,
    "loss": 0.13514181971549988,
    "learning_rate": 0.0009932146285882475,
    "timestep_mean": 645.25,
    "epsilon_norm": 7.617280006408691,
    "prediction_norm": 7.070394992828369,
    "timestamp": "2025-07-14T12:33:37.111990"
  },
  {
    "epoch": 105,
    "batch_idx": 160,
    "loss": 0.410347580909729,
    "learning_rate": 0.0009932146285882475,
    "timestep_mean": 435.40625,
    "epsilon_norm": 7.035642623901367,
    "prediction_norm": 4.881410121917725,
    "timestamp": "2025-07-14T12:33:37.117189"
  },
  {
    "epoch": 105,
    "batch_idx": 192,
    "loss": 1.0346225500106812,
    "learning_rate": 0.0009932146285882475,
    "timestep_mean": 439.75,
    "epsilon_norm": 5.251662731170654,
    "prediction_norm": 2.3445682525634766,
    "timestamp": "2025-07-14T12:33:37.122252"
  },
  {
    "epoch": 106,
    "batch_idx": 0,
    "loss": 0.6063963174819946,
    "learning_rate": 0.0009930850681144942,
    "timestep_mean": 414.875,
    "epsilon_norm": 8.703764915466309,
    "prediction_norm": 5.996312618255615,
    "timestamp": "2025-07-14T12:33:37.127498"
  },
  {
    "epoch": 106,
    "batch_idx": 32,
    "loss": 0.3632166087627411,
    "learning_rate": 0.0009930850681144942,
    "timestep_mean": 516.5,
    "epsilon_norm": 8.522354125976562,
    "prediction_norm": 6.837202072143555,
    "timestamp": "2025-07-14T12:33:37.132619"
  },
  {
    "epoch": 106,
    "batch_idx": 64,
    "loss": 0.25807201862335205,
    "learning_rate": 0.0009930850681144942,
    "timestep_mean": 447.5625,
    "epsilon_norm": 7.926948547363281,
    "prediction_norm": 6.94603157043457,
    "timestamp": "2025-07-14T12:33:37.137938"
  },
  {
    "epoch": 106,
    "batch_idx": 96,
    "loss": 0.3956391215324402,
    "learning_rate": 0.0009930850681144942,
    "timestep_mean": 521.875,
    "epsilon_norm": 9.292359352111816,
    "prediction_norm": 8.035126686096191,
    "timestamp": "2025-07-14T12:33:37.143054"
  },
  {
    "epoch": 106,
    "batch_idx": 128,
    "loss": 0.27241775393486023,
    "learning_rate": 0.0009930850681144942,
    "timestep_mean": 522.125,
    "epsilon_norm": 7.054620265960693,
    "prediction_norm": 6.223887920379639,
    "timestamp": "2025-07-14T12:33:37.148170"
  },
  {
    "epoch": 106,
    "batch_idx": 160,
    "loss": 0.5454466342926025,
    "learning_rate": 0.0009930850681144942,
    "timestep_mean": 438.0625,
    "epsilon_norm": 8.496689796447754,
    "prediction_norm": 6.073446750640869,
    "timestamp": "2025-07-14T12:33:37.153446"
  },
  {
    "epoch": 106,
    "batch_idx": 192,
    "loss": 0.258236825466156,
    "learning_rate": 0.0009930850681144942,
    "timestep_mean": 454.125,
    "epsilon_norm": 4.128582954406738,
    "prediction_norm": 3.4689512252807617,
    "timestamp": "2025-07-14T12:33:37.158646"
  },
  {
    "epoch": 107,
    "batch_idx": 0,
    "loss": 0.25020843744277954,
    "learning_rate": 0.0009929542910023514,
    "timestep_mean": 537.46875,
    "epsilon_norm": 8.132495880126953,
    "prediction_norm": 7.1865105628967285,
    "timestamp": "2025-07-14T12:33:37.163747"
  },
  {
    "epoch": 107,
    "batch_idx": 32,
    "loss": 0.4853644073009491,
    "learning_rate": 0.0009929542910023514,
    "timestep_mean": 507.8125,
    "epsilon_norm": 7.954204082489014,
    "prediction_norm": 5.951508045196533,
    "timestamp": "2025-07-14T12:33:37.168968"
  },
  {
    "epoch": 107,
    "batch_idx": 64,
    "loss": 0.38410526514053345,
    "learning_rate": 0.0009929542910023514,
    "timestep_mean": 475.84375,
    "epsilon_norm": 7.400446891784668,
    "prediction_norm": 5.549895286560059,
    "timestamp": "2025-07-14T12:33:37.174059"
  },
  {
    "epoch": 107,
    "batch_idx": 96,
    "loss": 0.4184889793395996,
    "learning_rate": 0.0009929542910023514,
    "timestep_mean": 476.375,
    "epsilon_norm": 7.214792251586914,
    "prediction_norm": 4.89225435256958,
    "timestamp": "2025-07-14T12:33:37.179194"
  },
  {
    "epoch": 107,
    "batch_idx": 128,
    "loss": 0.32638341188430786,
    "learning_rate": 0.0009929542910023514,
    "timestep_mean": 492.15625,
    "epsilon_norm": 9.310751914978027,
    "prediction_norm": 8.006524085998535,
    "timestamp": "2025-07-14T12:33:37.184344"
  },
  {
    "epoch": 107,
    "batch_idx": 160,
    "loss": 0.40304726362228394,
    "learning_rate": 0.0009929542910023514,
    "timestep_mean": 517.09375,
    "epsilon_norm": 8.236848831176758,
    "prediction_norm": 6.6166486740112305,
    "timestamp": "2025-07-14T12:33:37.189532"
  },
  {
    "epoch": 107,
    "batch_idx": 192,
    "loss": 0.15662415325641632,
    "learning_rate": 0.0009929542910023514,
    "timestep_mean": 590.875,
    "epsilon_norm": 4.389446258544922,
    "prediction_norm": 4.017996788024902,
    "timestamp": "2025-07-14T12:33:37.194736"
  },
  {
    "epoch": 108,
    "batch_idx": 0,
    "loss": 0.5127643942832947,
    "learning_rate": 0.0009928222975744989,
    "timestep_mean": 567.9375,
    "epsilon_norm": 9.056181907653809,
    "prediction_norm": 6.609339714050293,
    "timestamp": "2025-07-14T12:33:37.199921"
  },
  {
    "epoch": 108,
    "batch_idx": 32,
    "loss": 0.31634610891342163,
    "learning_rate": 0.0009928222975744989,
    "timestep_mean": 505.9375,
    "epsilon_norm": 8.538628578186035,
    "prediction_norm": 6.847753047943115,
    "timestamp": "2025-07-14T12:33:37.205012"
  },
  {
    "epoch": 108,
    "batch_idx": 64,
    "loss": 0.3374740481376648,
    "learning_rate": 0.0009928222975744989,
    "timestep_mean": 406.03125,
    "epsilon_norm": 7.850300312042236,
    "prediction_norm": 6.137430667877197,
    "timestamp": "2025-07-14T12:33:37.210124"
  },
  {
    "epoch": 108,
    "batch_idx": 96,
    "loss": 0.32109832763671875,
    "learning_rate": 0.0009928222975744989,
    "timestep_mean": 527.5625,
    "epsilon_norm": 7.578832149505615,
    "prediction_norm": 6.612275123596191,
    "timestamp": "2025-07-14T12:33:37.215295"
  },
  {
    "epoch": 108,
    "batch_idx": 128,
    "loss": 0.2970845699310303,
    "learning_rate": 0.0009928222975744989,
    "timestep_mean": 506.625,
    "epsilon_norm": 7.865098476409912,
    "prediction_norm": 6.6142168045043945,
    "timestamp": "2025-07-14T12:33:37.220520"
  },
  {
    "epoch": 108,
    "batch_idx": 160,
    "loss": 0.30928510427474976,
    "learning_rate": 0.0009928222975744989,
    "timestep_mean": 568.0,
    "epsilon_norm": 8.804108619689941,
    "prediction_norm": 7.35479211807251,
    "timestamp": "2025-07-14T12:33:37.225706"
  },
  {
    "epoch": 108,
    "batch_idx": 192,
    "loss": 0.5756258368492126,
    "learning_rate": 0.0009928222975744989,
    "timestep_mean": 460.75,
    "epsilon_norm": 4.2511820793151855,
    "prediction_norm": 3.610414505004883,
    "timestamp": "2025-07-14T12:33:37.230777"
  },
  {
    "epoch": 109,
    "batch_idx": 0,
    "loss": 0.4343138337135315,
    "learning_rate": 0.000992689088156617,
    "timestep_mean": 442.25,
    "epsilon_norm": 8.05171012878418,
    "prediction_norm": 5.718123435974121,
    "timestamp": "2025-07-14T12:33:37.236067"
  },
  {
    "epoch": 109,
    "batch_idx": 32,
    "loss": 0.29711824655532837,
    "learning_rate": 0.000992689088156617,
    "timestep_mean": 580.84375,
    "epsilon_norm": 7.659984111785889,
    "prediction_norm": 6.54967737197876,
    "timestamp": "2025-07-14T12:33:37.241183"
  },
  {
    "epoch": 109,
    "batch_idx": 64,
    "loss": 0.398522287607193,
    "learning_rate": 0.000992689088156617,
    "timestep_mean": 505.65625,
    "epsilon_norm": 8.356657028198242,
    "prediction_norm": 6.535600662231445,
    "timestamp": "2025-07-14T12:33:37.246466"
  },
  {
    "epoch": 109,
    "batch_idx": 96,
    "loss": 0.23815815150737762,
    "learning_rate": 0.000992689088156617,
    "timestep_mean": 560.875,
    "epsilon_norm": 7.353175640106201,
    "prediction_norm": 6.174593448638916,
    "timestamp": "2025-07-14T12:33:37.251656"
  },
  {
    "epoch": 109,
    "batch_idx": 128,
    "loss": 0.4708647131919861,
    "learning_rate": 0.000992689088156617,
    "timestep_mean": 423.59375,
    "epsilon_norm": 8.04297924041748,
    "prediction_norm": 5.538055419921875,
    "timestamp": "2025-07-14T12:33:37.256736"
  },
  {
    "epoch": 109,
    "batch_idx": 160,
    "loss": 0.39368724822998047,
    "learning_rate": 0.000992689088156617,
    "timestep_mean": 504.8125,
    "epsilon_norm": 7.7733659744262695,
    "prediction_norm": 5.954164505004883,
    "timestamp": "2025-07-14T12:33:37.261919"
  },
  {
    "epoch": 109,
    "batch_idx": 192,
    "loss": 0.5432388782501221,
    "learning_rate": 0.000992689088156617,
    "timestep_mean": 441.0,
    "epsilon_norm": 3.862323045730591,
    "prediction_norm": 2.2704482078552246,
    "timestamp": "2025-07-14T12:33:37.267243"
  },
  {
    "epoch": 110,
    "batch_idx": 0,
    "loss": 0.4254012107849121,
    "learning_rate": 0.0009925546630773867,
    "timestep_mean": 519.71875,
    "epsilon_norm": 7.867763042449951,
    "prediction_norm": 6.298001289367676,
    "timestamp": "2025-07-14T12:33:37.272366"
  },
  {
    "epoch": 110,
    "batch_idx": 32,
    "loss": 0.2657698392868042,
    "learning_rate": 0.0009925546630773867,
    "timestep_mean": 548.40625,
    "epsilon_norm": 7.705111980438232,
    "prediction_norm": 6.780420780181885,
    "timestamp": "2025-07-14T12:33:37.277598"
  },
  {
    "epoch": 110,
    "batch_idx": 64,
    "loss": 0.2948544919490814,
    "learning_rate": 0.0009925546630773867,
    "timestep_mean": 505.34375,
    "epsilon_norm": 7.744751453399658,
    "prediction_norm": 6.314627647399902,
    "timestamp": "2025-07-14T12:33:37.282706"
  },
  {
    "epoch": 110,
    "batch_idx": 96,
    "loss": 0.41088229417800903,
    "learning_rate": 0.0009925546630773867,
    "timestep_mean": 544.375,
    "epsilon_norm": 9.255325317382812,
    "prediction_norm": 8.048694610595703,
    "timestamp": "2025-07-14T12:33:37.287756"
  },
  {
    "epoch": 110,
    "batch_idx": 128,
    "loss": 0.4493602514266968,
    "learning_rate": 0.0009925546630773867,
    "timestep_mean": 530.15625,
    "epsilon_norm": 9.46688175201416,
    "prediction_norm": 8.57406997680664,
    "timestamp": "2025-07-14T12:33:37.292818"
  },
  {
    "epoch": 110,
    "batch_idx": 160,
    "loss": 0.585589587688446,
    "learning_rate": 0.0009925546630773867,
    "timestep_mean": 480.09375,
    "epsilon_norm": 8.607787132263184,
    "prediction_norm": 7.145943641662598,
    "timestamp": "2025-07-14T12:33:37.297658"
  },
  {
    "epoch": 110,
    "batch_idx": 192,
    "loss": 0.1702098250389099,
    "learning_rate": 0.0009925546630773867,
    "timestep_mean": 588.25,
    "epsilon_norm": 2.589287519454956,
    "prediction_norm": 2.089301586151123,
    "timestamp": "2025-07-14T12:33:37.302482"
  },
  {
    "epoch": 111,
    "batch_idx": 0,
    "loss": 0.4193319082260132,
    "learning_rate": 0.0009924190226684889,
    "timestep_mean": 526.9375,
    "epsilon_norm": 8.701858520507812,
    "prediction_norm": 6.729931354522705,
    "timestamp": "2025-07-14T12:33:37.307292"
  },
  {
    "epoch": 111,
    "batch_idx": 32,
    "loss": 0.19601774215698242,
    "learning_rate": 0.0009924190226684889,
    "timestep_mean": 582.375,
    "epsilon_norm": 6.73960542678833,
    "prediction_norm": 6.01040506362915,
    "timestamp": "2025-07-14T12:33:37.312124"
  },
  {
    "epoch": 111,
    "batch_idx": 64,
    "loss": 0.25983479619026184,
    "learning_rate": 0.0009924190226684889,
    "timestep_mean": 452.625,
    "epsilon_norm": 7.6268181800842285,
    "prediction_norm": 6.274754047393799,
    "timestamp": "2025-07-14T12:33:37.317379"
  },
  {
    "epoch": 111,
    "batch_idx": 96,
    "loss": 0.3074480891227722,
    "learning_rate": 0.0009924190226684889,
    "timestep_mean": 553.03125,
    "epsilon_norm": 8.904986381530762,
    "prediction_norm": 8.049039840698242,
    "timestamp": "2025-07-14T12:33:37.322490"
  },
  {
    "epoch": 111,
    "batch_idx": 128,
    "loss": 0.2714252471923828,
    "learning_rate": 0.0009924190226684889,
    "timestep_mean": 547.6875,
    "epsilon_norm": 8.446633338928223,
    "prediction_norm": 6.466481685638428,
    "timestamp": "2025-07-14T12:33:37.327591"
  },
  {
    "epoch": 111,
    "batch_idx": 160,
    "loss": 0.5095931887626648,
    "learning_rate": 0.0009924190226684889,
    "timestep_mean": 513.59375,
    "epsilon_norm": 8.577930450439453,
    "prediction_norm": 5.666226387023926,
    "timestamp": "2025-07-14T12:33:37.332746"
  },
  {
    "epoch": 111,
    "batch_idx": 192,
    "loss": 0.3715968132019043,
    "learning_rate": 0.0009924190226684889,
    "timestep_mean": 517.625,
    "epsilon_norm": 3.0740628242492676,
    "prediction_norm": 2.1539275646209717,
    "timestamp": "2025-07-14T12:33:37.338021"
  },
  {
    "epoch": 112,
    "batch_idx": 0,
    "loss": 0.3091588318347931,
    "learning_rate": 0.0009922821672646025,
    "timestep_mean": 501.375,
    "epsilon_norm": 8.261238098144531,
    "prediction_norm": 6.464503765106201,
    "timestamp": "2025-07-14T12:33:37.343098"
  },
  {
    "epoch": 112,
    "batch_idx": 32,
    "loss": 0.4388945698738098,
    "learning_rate": 0.0009922821672646025,
    "timestep_mean": 513.28125,
    "epsilon_norm": 9.101574897766113,
    "prediction_norm": 6.443770408630371,
    "timestamp": "2025-07-14T12:33:37.348251"
  },
  {
    "epoch": 112,
    "batch_idx": 64,
    "loss": 0.3206368088722229,
    "learning_rate": 0.0009922821672646025,
    "timestep_mean": 445.4375,
    "epsilon_norm": 6.790196418762207,
    "prediction_norm": 4.918881416320801,
    "timestamp": "2025-07-14T12:33:37.353434"
  },
  {
    "epoch": 112,
    "batch_idx": 96,
    "loss": 0.3646446466445923,
    "learning_rate": 0.0009922821672646025,
    "timestep_mean": 502.34375,
    "epsilon_norm": 8.67854118347168,
    "prediction_norm": 6.598727703094482,
    "timestamp": "2025-07-14T12:33:37.358703"
  },
  {
    "epoch": 112,
    "batch_idx": 128,
    "loss": 0.5300803780555725,
    "learning_rate": 0.0009922821672646025,
    "timestep_mean": 483.40625,
    "epsilon_norm": 8.86438274383545,
    "prediction_norm": 6.321949005126953,
    "timestamp": "2025-07-14T12:33:37.363826"
  },
  {
    "epoch": 112,
    "batch_idx": 160,
    "loss": 0.18469399213790894,
    "learning_rate": 0.0009922821672646025,
    "timestep_mean": 567.125,
    "epsilon_norm": 6.772212028503418,
    "prediction_norm": 6.061461925506592,
    "timestamp": "2025-07-14T12:33:37.369046"
  },
  {
    "epoch": 112,
    "batch_idx": 192,
    "loss": 0.18912239372730255,
    "learning_rate": 0.0009922821672646025,
    "timestep_mean": 565.75,
    "epsilon_norm": 4.744101524353027,
    "prediction_norm": 4.6634955406188965,
    "timestamp": "2025-07-14T12:33:37.374151"
  },
  {
    "epoch": 113,
    "batch_idx": 0,
    "loss": 0.4940939247608185,
    "learning_rate": 0.0009921440972034048,
    "timestep_mean": 459.46875,
    "epsilon_norm": 8.946327209472656,
    "prediction_norm": 6.864156723022461,
    "timestamp": "2025-07-14T12:33:37.379397"
  },
  {
    "epoch": 113,
    "batch_idx": 32,
    "loss": 0.35390424728393555,
    "learning_rate": 0.0009921440972034048,
    "timestep_mean": 552.78125,
    "epsilon_norm": 8.557182312011719,
    "prediction_norm": 7.288146495819092,
    "timestamp": "2025-07-14T12:33:37.384542"
  },
  {
    "epoch": 113,
    "batch_idx": 64,
    "loss": 0.5406705737113953,
    "learning_rate": 0.0009921440972034048,
    "timestep_mean": 378.9375,
    "epsilon_norm": 8.356328964233398,
    "prediction_norm": 6.300563335418701,
    "timestamp": "2025-07-14T12:33:37.389603"
  },
  {
    "epoch": 113,
    "batch_idx": 96,
    "loss": 0.30652183294296265,
    "learning_rate": 0.0009921440972034048,
    "timestep_mean": 456.25,
    "epsilon_norm": 6.856436729431152,
    "prediction_norm": 5.805269241333008,
    "timestamp": "2025-07-14T12:33:37.394816"
  },
  {
    "epoch": 113,
    "batch_idx": 128,
    "loss": 0.4445890784263611,
    "learning_rate": 0.0009921440972034048,
    "timestep_mean": 529.28125,
    "epsilon_norm": 8.223190307617188,
    "prediction_norm": 6.621375560760498,
    "timestamp": "2025-07-14T12:33:37.400089"
  },
  {
    "epoch": 113,
    "batch_idx": 160,
    "loss": 0.3367628753185272,
    "learning_rate": 0.0009921440972034048,
    "timestep_mean": 492.75,
    "epsilon_norm": 6.720889568328857,
    "prediction_norm": 5.570764541625977,
    "timestamp": "2025-07-14T12:33:37.405144"
  },
  {
    "epoch": 113,
    "batch_idx": 192,
    "loss": 0.4246712625026703,
    "learning_rate": 0.0009921440972034048,
    "timestep_mean": 559.0,
    "epsilon_norm": 4.47627067565918,
    "prediction_norm": 4.005417346954346,
    "timestamp": "2025-07-14T12:33:37.409941"
  },
  {
    "epoch": 114,
    "batch_idx": 0,
    "loss": 0.340474933385849,
    "learning_rate": 0.0009920048128255698,
    "timestep_mean": 531.0625,
    "epsilon_norm": 8.098459243774414,
    "prediction_norm": 6.848378658294678,
    "timestamp": "2025-07-14T12:33:37.414918"
  },
  {
    "epoch": 114,
    "batch_idx": 32,
    "loss": 0.38151904940605164,
    "learning_rate": 0.0009920048128255698,
    "timestep_mean": 570.9375,
    "epsilon_norm": 8.254520416259766,
    "prediction_norm": 6.782140254974365,
    "timestamp": "2025-07-14T12:33:37.420469"
  },
  {
    "epoch": 114,
    "batch_idx": 64,
    "loss": 0.3459824323654175,
    "learning_rate": 0.0009920048128255698,
    "timestep_mean": 548.46875,
    "epsilon_norm": 7.207214832305908,
    "prediction_norm": 5.538647651672363,
    "timestamp": "2025-07-14T12:33:37.425800"
  },
  {
    "epoch": 114,
    "batch_idx": 96,
    "loss": 0.2641832232475281,
    "learning_rate": 0.0009920048128255698,
    "timestep_mean": 492.03125,
    "epsilon_norm": 7.270598411560059,
    "prediction_norm": 6.49916934967041,
    "timestamp": "2025-07-14T12:33:37.430879"
  },
  {
    "epoch": 114,
    "batch_idx": 128,
    "loss": 0.2877682149410248,
    "learning_rate": 0.0009920048128255698,
    "timestep_mean": 552.53125,
    "epsilon_norm": 7.792731761932373,
    "prediction_norm": 5.710397243499756,
    "timestamp": "2025-07-14T12:33:37.436084"
  },
  {
    "epoch": 114,
    "batch_idx": 160,
    "loss": 0.4254797101020813,
    "learning_rate": 0.0009920048128255698,
    "timestep_mean": 422.3125,
    "epsilon_norm": 7.996566295623779,
    "prediction_norm": 5.711620330810547,
    "timestamp": "2025-07-14T12:33:37.441440"
  },
  {
    "epoch": 114,
    "batch_idx": 192,
    "loss": 0.07541526854038239,
    "learning_rate": 0.0009920048128255698,
    "timestep_mean": 683.0,
    "epsilon_norm": 3.3396217823028564,
    "prediction_norm": 3.1158158779144287,
    "timestamp": "2025-07-14T12:33:37.446608"
  },
  {
    "epoch": 115,
    "batch_idx": 0,
    "loss": 0.21599745750427246,
    "learning_rate": 0.000991864314474768,
    "timestep_mean": 512.78125,
    "epsilon_norm": 8.17432689666748,
    "prediction_norm": 6.3986334800720215,
    "timestamp": "2025-07-14T12:33:37.451810"
  },
  {
    "epoch": 115,
    "batch_idx": 32,
    "loss": 0.49933913350105286,
    "learning_rate": 0.000991864314474768,
    "timestep_mean": 411.90625,
    "epsilon_norm": 8.160754203796387,
    "prediction_norm": 5.234872341156006,
    "timestamp": "2025-07-14T12:33:37.457004"
  },
  {
    "epoch": 115,
    "batch_idx": 64,
    "loss": 0.21127715706825256,
    "learning_rate": 0.000991864314474768,
    "timestep_mean": 539.34375,
    "epsilon_norm": 8.416984558105469,
    "prediction_norm": 7.0650835037231445,
    "timestamp": "2025-07-14T12:33:37.462186"
  },
  {
    "epoch": 115,
    "batch_idx": 96,
    "loss": 0.5148974657058716,
    "learning_rate": 0.000991864314474768,
    "timestep_mean": 539.5,
    "epsilon_norm": 9.137349128723145,
    "prediction_norm": 6.569322109222412,
    "timestamp": "2025-07-14T12:33:37.467441"
  },
  {
    "epoch": 115,
    "batch_idx": 128,
    "loss": 0.5556954741477966,
    "learning_rate": 0.000991864314474768,
    "timestep_mean": 428.21875,
    "epsilon_norm": 8.191551208496094,
    "prediction_norm": 5.965640544891357,
    "timestamp": "2025-07-14T12:33:37.472505"
  },
  {
    "epoch": 115,
    "batch_idx": 160,
    "loss": 0.332843542098999,
    "learning_rate": 0.000991864314474768,
    "timestep_mean": 519.9375,
    "epsilon_norm": 8.615157127380371,
    "prediction_norm": 7.4255266189575195,
    "timestamp": "2025-07-14T12:33:37.477719"
  },
  {
    "epoch": 115,
    "batch_idx": 192,
    "loss": 0.1179361343383789,
    "learning_rate": 0.000991864314474768,
    "timestep_mean": 697.75,
    "epsilon_norm": 4.033980846405029,
    "prediction_norm": 3.427659273147583,
    "timestamp": "2025-07-14T12:33:37.482942"
  },
  {
    "epoch": 116,
    "batch_idx": 0,
    "loss": 0.37956851720809937,
    "learning_rate": 0.0009917226024976648,
    "timestep_mean": 538.0,
    "epsilon_norm": 8.660490989685059,
    "prediction_norm": 7.579656600952148,
    "timestamp": "2025-07-14T12:33:37.488058"
  },
  {
    "epoch": 116,
    "batch_idx": 32,
    "loss": 0.6275892853736877,
    "learning_rate": 0.0009917226024976648,
    "timestep_mean": 444.28125,
    "epsilon_norm": 8.296151161193848,
    "prediction_norm": 5.026976585388184,
    "timestamp": "2025-07-14T12:33:37.493320"
  },
  {
    "epoch": 116,
    "batch_idx": 64,
    "loss": 0.22018404304981232,
    "learning_rate": 0.0009917226024976648,
    "timestep_mean": 536.09375,
    "epsilon_norm": 7.355233669281006,
    "prediction_norm": 6.973972797393799,
    "timestamp": "2025-07-14T12:33:37.498414"
  },
  {
    "epoch": 116,
    "batch_idx": 96,
    "loss": 0.4221387505531311,
    "learning_rate": 0.0009917226024976648,
    "timestep_mean": 421.15625,
    "epsilon_norm": 9.10452938079834,
    "prediction_norm": 7.928186416625977,
    "timestamp": "2025-07-14T12:33:37.503543"
  },
  {
    "epoch": 116,
    "batch_idx": 128,
    "loss": 0.3154829144477844,
    "learning_rate": 0.0009917226024976648,
    "timestep_mean": 547.625,
    "epsilon_norm": 8.592336654663086,
    "prediction_norm": 8.302753448486328,
    "timestamp": "2025-07-14T12:33:37.510927"
  },
  {
    "epoch": 116,
    "batch_idx": 160,
    "loss": 0.275186687707901,
    "learning_rate": 0.0009917226024976648,
    "timestep_mean": 581.53125,
    "epsilon_norm": 8.70511531829834,
    "prediction_norm": 8.300097465515137,
    "timestamp": "2025-07-14T12:33:37.515750"
  },
  {
    "epoch": 116,
    "batch_idx": 192,
    "loss": 0.5290161371231079,
    "learning_rate": 0.0009917226024976648,
    "timestep_mean": 388.375,
    "epsilon_norm": 3.505037307739258,
    "prediction_norm": 2.753570795059204,
    "timestamp": "2025-07-14T12:33:37.520617"
  },
  {
    "epoch": 117,
    "batch_idx": 0,
    "loss": 0.7045832276344299,
    "learning_rate": 0.0009915796772439207,
    "timestep_mean": 483.0625,
    "epsilon_norm": 8.546338081359863,
    "prediction_norm": 6.408624172210693,
    "timestamp": "2025-07-14T12:33:37.525435"
  },
  {
    "epoch": 117,
    "batch_idx": 32,
    "loss": 0.30690932273864746,
    "learning_rate": 0.0009915796772439207,
    "timestep_mean": 465.5,
    "epsilon_norm": 7.409895896911621,
    "prediction_norm": 5.577513217926025,
    "timestamp": "2025-07-14T12:33:37.530394"
  },
  {
    "epoch": 117,
    "batch_idx": 64,
    "loss": 0.32167840003967285,
    "learning_rate": 0.0009915796772439207,
    "timestep_mean": 531.28125,
    "epsilon_norm": 8.278764724731445,
    "prediction_norm": 6.87249755859375,
    "timestamp": "2025-07-14T12:33:37.535287"
  },
  {
    "epoch": 117,
    "batch_idx": 96,
    "loss": 0.45652562379837036,
    "learning_rate": 0.0009915796772439207,
    "timestep_mean": 427.0625,
    "epsilon_norm": 7.981244087219238,
    "prediction_norm": 5.401159286499023,
    "timestamp": "2025-07-14T12:33:37.540312"
  },
  {
    "epoch": 117,
    "batch_idx": 128,
    "loss": 0.3196851909160614,
    "learning_rate": 0.0009915796772439207,
    "timestep_mean": 507.59375,
    "epsilon_norm": 7.5318803787231445,
    "prediction_norm": 5.812379360198975,
    "timestamp": "2025-07-14T12:33:37.545049"
  },
  {
    "epoch": 117,
    "batch_idx": 160,
    "loss": 0.630097508430481,
    "learning_rate": 0.0009915796772439207,
    "timestep_mean": 376.96875,
    "epsilon_norm": 8.467517852783203,
    "prediction_norm": 4.975898265838623,
    "timestamp": "2025-07-14T12:33:37.550096"
  },
  {
    "epoch": 117,
    "batch_idx": 192,
    "loss": 0.43956711888313293,
    "learning_rate": 0.0009915796772439207,
    "timestep_mean": 500.25,
    "epsilon_norm": 3.351869583129883,
    "prediction_norm": 2.2702653408050537,
    "timestamp": "2025-07-14T12:33:37.554962"
  },
  {
    "epoch": 118,
    "batch_idx": 0,
    "loss": 0.5379266738891602,
    "learning_rate": 0.0009914355390661896,
    "timestep_mean": 430.1875,
    "epsilon_norm": 7.483062744140625,
    "prediction_norm": 4.412417888641357,
    "timestamp": "2025-07-14T12:33:37.559842"
  },
  {
    "epoch": 118,
    "batch_idx": 32,
    "loss": 0.5218991637229919,
    "learning_rate": 0.0009914355390661896,
    "timestep_mean": 542.34375,
    "epsilon_norm": 7.953352451324463,
    "prediction_norm": 4.95432186126709,
    "timestamp": "2025-07-14T12:33:37.564806"
  },
  {
    "epoch": 118,
    "batch_idx": 64,
    "loss": 0.3965049386024475,
    "learning_rate": 0.0009914355390661896,
    "timestep_mean": 483.75,
    "epsilon_norm": 7.437748432159424,
    "prediction_norm": 4.808621883392334,
    "timestamp": "2025-07-14T12:33:37.569742"
  },
  {
    "epoch": 118,
    "batch_idx": 96,
    "loss": 0.25532060861587524,
    "learning_rate": 0.0009914355390661896,
    "timestep_mean": 544.21875,
    "epsilon_norm": 7.862595081329346,
    "prediction_norm": 6.058428764343262,
    "timestamp": "2025-07-14T12:33:37.574875"
  },
  {
    "epoch": 118,
    "batch_idx": 128,
    "loss": 0.3183661699295044,
    "learning_rate": 0.0009914355390661896,
    "timestep_mean": 517.25,
    "epsilon_norm": 7.87383508682251,
    "prediction_norm": 5.495029926300049,
    "timestamp": "2025-07-14T12:33:37.579834"
  },
  {
    "epoch": 118,
    "batch_idx": 160,
    "loss": 0.38554853200912476,
    "learning_rate": 0.0009914355390661896,
    "timestep_mean": 458.5,
    "epsilon_norm": 8.800315856933594,
    "prediction_norm": 6.9463419914245605,
    "timestamp": "2025-07-14T12:33:37.584747"
  },
  {
    "epoch": 118,
    "batch_idx": 192,
    "loss": 0.14244890213012695,
    "learning_rate": 0.0009914355390661896,
    "timestep_mean": 584.75,
    "epsilon_norm": 3.95837140083313,
    "prediction_norm": 3.2570958137512207,
    "timestamp": "2025-07-14T12:33:37.589807"
  },
  {
    "epoch": 119,
    "batch_idx": 0,
    "loss": 0.38486745953559875,
    "learning_rate": 0.000991290188320118,
    "timestep_mean": 443.125,
    "epsilon_norm": 8.782317161560059,
    "prediction_norm": 6.685527801513672,
    "timestamp": "2025-07-14T12:33:37.594782"
  },
  {
    "epoch": 119,
    "batch_idx": 32,
    "loss": 0.2704054117202759,
    "learning_rate": 0.000991290188320118,
    "timestep_mean": 546.78125,
    "epsilon_norm": 7.89938497543335,
    "prediction_norm": 6.683765888214111,
    "timestamp": "2025-07-14T12:33:37.599853"
  },
  {
    "epoch": 119,
    "batch_idx": 64,
    "loss": 0.25802871584892273,
    "learning_rate": 0.000991290188320118,
    "timestep_mean": 507.15625,
    "epsilon_norm": 7.24906587600708,
    "prediction_norm": 6.324376583099365,
    "timestamp": "2025-07-14T12:33:37.604581"
  },
  {
    "epoch": 119,
    "batch_idx": 96,
    "loss": 0.4118366837501526,
    "learning_rate": 0.000991290188320118,
    "timestep_mean": 451.03125,
    "epsilon_norm": 7.37276029586792,
    "prediction_norm": 6.638822555541992,
    "timestamp": "2025-07-14T12:33:37.609385"
  },
  {
    "epoch": 119,
    "batch_idx": 128,
    "loss": 0.4292304217815399,
    "learning_rate": 0.000991290188320118,
    "timestep_mean": 498.75,
    "epsilon_norm": 7.239608287811279,
    "prediction_norm": 6.609118938446045,
    "timestamp": "2025-07-14T12:33:37.614141"
  },
  {
    "epoch": 119,
    "batch_idx": 160,
    "loss": 0.4058070182800293,
    "learning_rate": 0.000991290188320118,
    "timestep_mean": 510.53125,
    "epsilon_norm": 8.916321754455566,
    "prediction_norm": 7.814456939697266,
    "timestamp": "2025-07-14T12:33:37.618958"
  },
  {
    "epoch": 119,
    "batch_idx": 192,
    "loss": 0.6979591846466064,
    "learning_rate": 0.000991290188320118,
    "timestep_mean": 449.875,
    "epsilon_norm": 3.4004523754119873,
    "prediction_norm": 2.2355995178222656,
    "timestamp": "2025-07-14T12:33:37.623966"
  },
  {
    "epoch": 120,
    "batch_idx": 0,
    "loss": 0.5504604578018188,
    "learning_rate": 0.0009911436253643446,
    "timestep_mean": 492.25,
    "epsilon_norm": 7.68098783493042,
    "prediction_norm": 7.221776485443115,
    "timestamp": "2025-07-14T12:33:37.628912"
  },
  {
    "epoch": 120,
    "batch_idx": 32,
    "loss": 0.30327147245407104,
    "learning_rate": 0.0009911436253643446,
    "timestep_mean": 483.65625,
    "epsilon_norm": 8.155720710754395,
    "prediction_norm": 7.753027439117432,
    "timestamp": "2025-07-14T12:33:37.633902"
  },
  {
    "epoch": 120,
    "batch_idx": 64,
    "loss": 0.3136581778526306,
    "learning_rate": 0.0009911436253643446,
    "timestep_mean": 524.90625,
    "epsilon_norm": 7.838966369628906,
    "prediction_norm": 6.773355960845947,
    "timestamp": "2025-07-14T12:33:37.638824"
  },
  {
    "epoch": 120,
    "batch_idx": 96,
    "loss": 0.25892651081085205,
    "learning_rate": 0.0009911436253643446,
    "timestep_mean": 575.375,
    "epsilon_norm": 6.866048812866211,
    "prediction_norm": 5.473714828491211,
    "timestamp": "2025-07-14T12:33:37.643615"
  },
  {
    "epoch": 120,
    "batch_idx": 128,
    "loss": 0.292054146528244,
    "learning_rate": 0.0009911436253643446,
    "timestep_mean": 488.6875,
    "epsilon_norm": 6.717254161834717,
    "prediction_norm": 4.8916425704956055,
    "timestamp": "2025-07-14T12:33:37.648487"
  },
  {
    "epoch": 120,
    "batch_idx": 160,
    "loss": 0.5235522985458374,
    "learning_rate": 0.0009911436253643446,
    "timestep_mean": 397.9375,
    "epsilon_norm": 8.411746978759766,
    "prediction_norm": 5.463203430175781,
    "timestamp": "2025-07-14T12:33:37.653356"
  },
  {
    "epoch": 120,
    "batch_idx": 192,
    "loss": 0.9057825803756714,
    "learning_rate": 0.0009911436253643446,
    "timestep_mean": 457.625,
    "epsilon_norm": 4.482018947601318,
    "prediction_norm": 2.030118465423584,
    "timestamp": "2025-07-14T12:33:37.658305"
  },
  {
    "epoch": 121,
    "batch_idx": 0,
    "loss": 0.3592689037322998,
    "learning_rate": 0.0009909958505604985,
    "timestep_mean": 425.0625,
    "epsilon_norm": 7.208661079406738,
    "prediction_norm": 4.782301425933838,
    "timestamp": "2025-07-14T12:33:37.663176"
  },
  {
    "epoch": 121,
    "batch_idx": 32,
    "loss": 0.40276962518692017,
    "learning_rate": 0.0009909958505604985,
    "timestep_mean": 538.65625,
    "epsilon_norm": 7.8452839851379395,
    "prediction_norm": 5.01453971862793,
    "timestamp": "2025-07-14T12:33:37.668067"
  },
  {
    "epoch": 121,
    "batch_idx": 64,
    "loss": 0.32718726992607117,
    "learning_rate": 0.0009909958505604985,
    "timestep_mean": 498.25,
    "epsilon_norm": 7.206490993499756,
    "prediction_norm": 4.923237323760986,
    "timestamp": "2025-07-14T12:33:37.672925"
  },
  {
    "epoch": 121,
    "batch_idx": 96,
    "loss": 0.21730774641036987,
    "learning_rate": 0.0009909958505604985,
    "timestep_mean": 547.71875,
    "epsilon_norm": 7.193985939025879,
    "prediction_norm": 5.788108825683594,
    "timestamp": "2025-07-14T12:33:37.677805"
  },
  {
    "epoch": 121,
    "batch_idx": 128,
    "loss": 0.1997320055961609,
    "learning_rate": 0.0009909958505604985,
    "timestep_mean": 596.9375,
    "epsilon_norm": 7.394113540649414,
    "prediction_norm": 6.569543361663818,
    "timestamp": "2025-07-14T12:33:37.682727"
  },
  {
    "epoch": 121,
    "batch_idx": 160,
    "loss": 0.26382434368133545,
    "learning_rate": 0.0009909958505604985,
    "timestep_mean": 570.6875,
    "epsilon_norm": 9.36764907836914,
    "prediction_norm": 7.6552934646606445,
    "timestamp": "2025-07-14T12:33:37.687636"
  },
  {
    "epoch": 121,
    "batch_idx": 192,
    "loss": 0.3948315978050232,
    "learning_rate": 0.0009909958505604985,
    "timestep_mean": 666.0,
    "epsilon_norm": 4.1509881019592285,
    "prediction_norm": 3.523709535598755,
    "timestamp": "2025-07-14T12:33:37.692540"
  },
  {
    "epoch": 122,
    "batch_idx": 0,
    "loss": 0.4105948805809021,
    "learning_rate": 0.0009908468642731998,
    "timestep_mean": 435.3125,
    "epsilon_norm": 7.752055644989014,
    "prediction_norm": 6.927311897277832,
    "timestamp": "2025-07-14T12:33:37.697349"
  },
  {
    "epoch": 122,
    "batch_idx": 32,
    "loss": 0.21379129588603973,
    "learning_rate": 0.0009908468642731998,
    "timestep_mean": 470.0,
    "epsilon_norm": 8.819382667541504,
    "prediction_norm": 8.052188873291016,
    "timestamp": "2025-07-14T12:33:37.702178"
  },
  {
    "epoch": 122,
    "batch_idx": 64,
    "loss": 0.5532704591751099,
    "learning_rate": 0.0009908468642731998,
    "timestep_mean": 504.875,
    "epsilon_norm": 9.131240844726562,
    "prediction_norm": 6.910861492156982,
    "timestamp": "2025-07-14T12:33:37.707016"
  },
  {
    "epoch": 122,
    "batch_idx": 96,
    "loss": 0.4066206216812134,
    "learning_rate": 0.0009908468642731998,
    "timestep_mean": 478.65625,
    "epsilon_norm": 7.991306304931641,
    "prediction_norm": 6.056802749633789,
    "timestamp": "2025-07-14T12:33:37.711850"
  },
  {
    "epoch": 122,
    "batch_idx": 128,
    "loss": 0.3657856583595276,
    "learning_rate": 0.0009908468642731998,
    "timestep_mean": 503.46875,
    "epsilon_norm": 7.590660572052002,
    "prediction_norm": 5.901922225952148,
    "timestamp": "2025-07-14T12:33:37.716817"
  },
  {
    "epoch": 122,
    "batch_idx": 160,
    "loss": 0.4196021556854248,
    "learning_rate": 0.0009908468642731998,
    "timestep_mean": 505.8125,
    "epsilon_norm": 7.9193949699401855,
    "prediction_norm": 6.188198566436768,
    "timestamp": "2025-07-14T12:33:37.721613"
  },
  {
    "epoch": 122,
    "batch_idx": 192,
    "loss": 0.25860702991485596,
    "learning_rate": 0.0009908468642731998,
    "timestep_mean": 444.75,
    "epsilon_norm": 4.369572639465332,
    "prediction_norm": 3.4287023544311523,
    "timestamp": "2025-07-14T12:33:37.726432"
  },
  {
    "epoch": 123,
    "batch_idx": 0,
    "loss": 0.281716912984848,
    "learning_rate": 0.000990696666870057,
    "timestep_mean": 511.84375,
    "epsilon_norm": 7.539241313934326,
    "prediction_norm": 6.329103946685791,
    "timestamp": "2025-07-14T12:33:37.731277"
  },
  {
    "epoch": 123,
    "batch_idx": 32,
    "loss": 0.44532567262649536,
    "learning_rate": 0.000990696666870057,
    "timestep_mean": 447.09375,
    "epsilon_norm": 7.96061372756958,
    "prediction_norm": 6.476134300231934,
    "timestamp": "2025-07-14T12:33:37.736151"
  },
  {
    "epoch": 123,
    "batch_idx": 64,
    "loss": 0.37185293436050415,
    "learning_rate": 0.000990696666870057,
    "timestep_mean": 503.15625,
    "epsilon_norm": 7.711949348449707,
    "prediction_norm": 7.384365081787109,
    "timestamp": "2025-07-14T12:33:37.740968"
  },
  {
    "epoch": 123,
    "batch_idx": 96,
    "loss": 0.3241903483867645,
    "learning_rate": 0.000990696666870057,
    "timestep_mean": 558.15625,
    "epsilon_norm": 8.059011459350586,
    "prediction_norm": 7.428547382354736,
    "timestamp": "2025-07-14T12:33:37.745802"
  },
  {
    "epoch": 123,
    "batch_idx": 128,
    "loss": 0.391734778881073,
    "learning_rate": 0.000990696666870057,
    "timestep_mean": 487.46875,
    "epsilon_norm": 7.701684951782227,
    "prediction_norm": 6.701570987701416,
    "timestamp": "2025-07-14T12:33:37.750638"
  },
  {
    "epoch": 123,
    "batch_idx": 160,
    "loss": 0.32219237089157104,
    "learning_rate": 0.000990696666870057,
    "timestep_mean": 488.8125,
    "epsilon_norm": 7.411844730377197,
    "prediction_norm": 5.891948223114014,
    "timestamp": "2025-07-14T12:33:37.755404"
  },
  {
    "epoch": 123,
    "batch_idx": 192,
    "loss": 0.4000953733921051,
    "learning_rate": 0.000990696666870057,
    "timestep_mean": 604.625,
    "epsilon_norm": 5.542208194732666,
    "prediction_norm": 5.033451557159424,
    "timestamp": "2025-07-14T12:33:37.760194"
  },
  {
    "epoch": 124,
    "batch_idx": 0,
    "loss": 0.2538461983203888,
    "learning_rate": 0.0009905452587216671,
    "timestep_mean": 612.15625,
    "epsilon_norm": 7.198777675628662,
    "prediction_norm": 5.75086784362793,
    "timestamp": "2025-07-14T12:33:37.765058"
  },
  {
    "epoch": 124,
    "batch_idx": 32,
    "loss": 0.4274328947067261,
    "learning_rate": 0.0009905452587216671,
    "timestep_mean": 586.71875,
    "epsilon_norm": 7.874880313873291,
    "prediction_norm": 5.507667541503906,
    "timestamp": "2025-07-14T12:33:37.769933"
  },
  {
    "epoch": 124,
    "batch_idx": 64,
    "loss": 0.48032301664352417,
    "learning_rate": 0.0009905452587216671,
    "timestep_mean": 478.375,
    "epsilon_norm": 8.418073654174805,
    "prediction_norm": 5.765039920806885,
    "timestamp": "2025-07-14T12:33:37.774786"
  },
  {
    "epoch": 124,
    "batch_idx": 96,
    "loss": 0.3861296772956848,
    "learning_rate": 0.0009905452587216671,
    "timestep_mean": 415.625,
    "epsilon_norm": 7.1739959716796875,
    "prediction_norm": 4.417407989501953,
    "timestamp": "2025-07-14T12:33:37.779719"
  },
  {
    "epoch": 124,
    "batch_idx": 128,
    "loss": 0.33216947317123413,
    "learning_rate": 0.0009905452587216671,
    "timestep_mean": 498.1875,
    "epsilon_norm": 6.5287346839904785,
    "prediction_norm": 4.279565811157227,
    "timestamp": "2025-07-14T12:33:37.784520"
  },
  {
    "epoch": 124,
    "batch_idx": 160,
    "loss": 0.27003592252731323,
    "learning_rate": 0.0009905452587216671,
    "timestep_mean": 582.5625,
    "epsilon_norm": 8.876022338867188,
    "prediction_norm": 7.045156478881836,
    "timestamp": "2025-07-14T12:33:37.789365"
  },
  {
    "epoch": 124,
    "batch_idx": 192,
    "loss": 0.38322606682777405,
    "learning_rate": 0.0009905452587216671,
    "timestep_mean": 453.875,
    "epsilon_norm": 2.9779717922210693,
    "prediction_norm": 2.251176595687866,
    "timestamp": "2025-07-14T12:33:37.794194"
  },
  {
    "epoch": 125,
    "batch_idx": 0,
    "loss": 0.541598916053772,
    "learning_rate": 0.0009903926402016153,
    "timestep_mean": 443.03125,
    "epsilon_norm": 8.739550590515137,
    "prediction_norm": 6.484620094299316,
    "timestamp": "2025-07-14T12:33:37.799057"
  },
  {
    "epoch": 125,
    "batch_idx": 32,
    "loss": 0.26428455114364624,
    "learning_rate": 0.0009903926402016153,
    "timestep_mean": 542.0,
    "epsilon_norm": 8.230023384094238,
    "prediction_norm": 6.985214710235596,
    "timestamp": "2025-07-14T12:33:37.803888"
  },
  {
    "epoch": 125,
    "batch_idx": 64,
    "loss": 0.409614235162735,
    "learning_rate": 0.0009903926402016153,
    "timestep_mean": 465.3125,
    "epsilon_norm": 7.701648712158203,
    "prediction_norm": 4.896946907043457,
    "timestamp": "2025-07-14T12:33:37.808717"
  },
  {
    "epoch": 125,
    "batch_idx": 96,
    "loss": 0.18507543206214905,
    "learning_rate": 0.0009903926402016153,
    "timestep_mean": 554.0,
    "epsilon_norm": 6.483545303344727,
    "prediction_norm": 5.5479416847229,
    "timestamp": "2025-07-14T12:33:37.813562"
  },
  {
    "epoch": 125,
    "batch_idx": 128,
    "loss": 0.3077320158481598,
    "learning_rate": 0.0009903926402016153,
    "timestep_mean": 483.0,
    "epsilon_norm": 6.760209083557129,
    "prediction_norm": 4.711238384246826,
    "timestamp": "2025-07-14T12:33:37.818439"
  },
  {
    "epoch": 125,
    "batch_idx": 160,
    "loss": 0.36333298683166504,
    "learning_rate": 0.0009903926402016153,
    "timestep_mean": 482.0,
    "epsilon_norm": 8.016545295715332,
    "prediction_norm": 6.1329345703125,
    "timestamp": "2025-07-14T12:33:37.823293"
  },
  {
    "epoch": 125,
    "batch_idx": 192,
    "loss": 0.2536838948726654,
    "learning_rate": 0.0009903926402016153,
    "timestep_mean": 467.625,
    "epsilon_norm": 3.9958488941192627,
    "prediction_norm": 2.6799628734588623,
    "timestamp": "2025-07-14T12:33:37.828130"
  },
  {
    "epoch": 126,
    "batch_idx": 0,
    "loss": 0.27219581604003906,
    "learning_rate": 0.0009902388116864722,
    "timestep_mean": 579.5625,
    "epsilon_norm": 8.020051956176758,
    "prediction_norm": 6.878695487976074,
    "timestamp": "2025-07-14T12:33:37.833089"
  },
  {
    "epoch": 126,
    "batch_idx": 32,
    "loss": 0.2977909743785858,
    "learning_rate": 0.0009902388116864722,
    "timestep_mean": 539.71875,
    "epsilon_norm": 7.212671756744385,
    "prediction_norm": 5.956045150756836,
    "timestamp": "2025-07-14T12:33:37.837938"
  },
  {
    "epoch": 126,
    "batch_idx": 64,
    "loss": 0.16400524973869324,
    "learning_rate": 0.0009902388116864722,
    "timestep_mean": 501.28125,
    "epsilon_norm": 7.3723063468933105,
    "prediction_norm": 7.457326889038086,
    "timestamp": "2025-07-14T12:33:37.842754"
  },
  {
    "epoch": 126,
    "batch_idx": 96,
    "loss": 0.2122589349746704,
    "learning_rate": 0.0009902388116864722,
    "timestep_mean": 652.9375,
    "epsilon_norm": 7.745668411254883,
    "prediction_norm": 6.877631664276123,
    "timestamp": "2025-07-14T12:33:37.847595"
  },
  {
    "epoch": 126,
    "batch_idx": 128,
    "loss": 0.29928678274154663,
    "learning_rate": 0.0009902388116864722,
    "timestep_mean": 527.5,
    "epsilon_norm": 9.031517028808594,
    "prediction_norm": 8.22368335723877,
    "timestamp": "2025-07-14T12:33:37.852426"
  },
  {
    "epoch": 126,
    "batch_idx": 160,
    "loss": 0.4311370253562927,
    "learning_rate": 0.0009902388116864722,
    "timestep_mean": 450.40625,
    "epsilon_norm": 9.356561660766602,
    "prediction_norm": 7.4609174728393555,
    "timestamp": "2025-07-14T12:33:37.857283"
  },
  {
    "epoch": 126,
    "batch_idx": 192,
    "loss": 0.45007455348968506,
    "learning_rate": 0.0009902388116864722,
    "timestep_mean": 482.375,
    "epsilon_norm": 4.437366962432861,
    "prediction_norm": 3.904228687286377,
    "timestamp": "2025-07-14T12:33:37.862205"
  },
  {
    "epoch": 127,
    "batch_idx": 0,
    "loss": 0.5907585620880127,
    "learning_rate": 0.0009900837735557946,
    "timestep_mean": 517.5,
    "epsilon_norm": 8.873037338256836,
    "prediction_norm": 7.267930030822754,
    "timestamp": "2025-07-14T12:33:37.867376"
  },
  {
    "epoch": 127,
    "batch_idx": 32,
    "loss": 0.3631727695465088,
    "learning_rate": 0.0009900837735557946,
    "timestep_mean": 496.53125,
    "epsilon_norm": 7.469221115112305,
    "prediction_norm": 6.623608112335205,
    "timestamp": "2025-07-14T12:33:37.872362"
  },
  {
    "epoch": 127,
    "batch_idx": 64,
    "loss": 0.44190269708633423,
    "learning_rate": 0.0009900837735557946,
    "timestep_mean": 520.125,
    "epsilon_norm": 8.084002494812012,
    "prediction_norm": 6.549149513244629,
    "timestamp": "2025-07-14T12:33:37.877308"
  },
  {
    "epoch": 127,
    "batch_idx": 96,
    "loss": 0.36402830481529236,
    "learning_rate": 0.0009900837735557946,
    "timestep_mean": 541.0,
    "epsilon_norm": 8.28555679321289,
    "prediction_norm": 6.9304351806640625,
    "timestamp": "2025-07-14T12:33:37.882231"
  },
  {
    "epoch": 127,
    "batch_idx": 128,
    "loss": 0.3341398537158966,
    "learning_rate": 0.0009900837735557946,
    "timestep_mean": 548.3125,
    "epsilon_norm": 8.711830139160156,
    "prediction_norm": 7.49447774887085,
    "timestamp": "2025-07-14T12:33:37.887145"
  },
  {
    "epoch": 127,
    "batch_idx": 160,
    "loss": 0.36975836753845215,
    "learning_rate": 0.0009900837735557946,
    "timestep_mean": 488.8125,
    "epsilon_norm": 8.158108711242676,
    "prediction_norm": 7.259411334991455,
    "timestamp": "2025-07-14T12:33:37.892048"
  },
  {
    "epoch": 127,
    "batch_idx": 192,
    "loss": 0.35470157861709595,
    "learning_rate": 0.0009900837735557946,
    "timestep_mean": 642.75,
    "epsilon_norm": 3.9181606769561768,
    "prediction_norm": 2.6092803478240967,
    "timestamp": "2025-07-14T12:33:37.896962"
  },
  {
    "epoch": 128,
    "batch_idx": 0,
    "loss": 0.40647122263908386,
    "learning_rate": 0.0009899275261921233,
    "timestep_mean": 451.75,
    "epsilon_norm": 7.642467498779297,
    "prediction_norm": 5.732439041137695,
    "timestamp": "2025-07-14T12:33:37.901832"
  },
  {
    "epoch": 128,
    "batch_idx": 32,
    "loss": 0.6415207982063293,
    "learning_rate": 0.0009899275261921233,
    "timestep_mean": 412.34375,
    "epsilon_norm": 8.619030952453613,
    "prediction_norm": 5.519336223602295,
    "timestamp": "2025-07-14T12:33:37.906706"
  },
  {
    "epoch": 128,
    "batch_idx": 64,
    "loss": 0.2292550653219223,
    "learning_rate": 0.0009899275261921233,
    "timestep_mean": 507.0,
    "epsilon_norm": 6.944864749908447,
    "prediction_norm": 5.3638153076171875,
    "timestamp": "2025-07-14T12:33:37.911564"
  },
  {
    "epoch": 128,
    "batch_idx": 96,
    "loss": 0.3166985511779785,
    "learning_rate": 0.0009899275261921233,
    "timestep_mean": 526.9375,
    "epsilon_norm": 7.455838203430176,
    "prediction_norm": 4.975937366485596,
    "timestamp": "2025-07-14T12:33:37.916402"
  },
  {
    "epoch": 128,
    "batch_idx": 128,
    "loss": 0.36005833745002747,
    "learning_rate": 0.0009899275261921233,
    "timestep_mean": 527.25,
    "epsilon_norm": 6.512921333312988,
    "prediction_norm": 4.529518127441406,
    "timestamp": "2025-07-14T12:33:37.921363"
  },
  {
    "epoch": 128,
    "batch_idx": 160,
    "loss": 0.2988072633743286,
    "learning_rate": 0.0009899275261921233,
    "timestep_mean": 543.1875,
    "epsilon_norm": 6.882297515869141,
    "prediction_norm": 5.136580944061279,
    "timestamp": "2025-07-14T12:33:37.926225"
  },
  {
    "epoch": 128,
    "batch_idx": 192,
    "loss": 0.49407896399497986,
    "learning_rate": 0.0009899275261921233,
    "timestep_mean": 388.625,
    "epsilon_norm": 3.7963602542877197,
    "prediction_norm": 2.787377119064331,
    "timestamp": "2025-07-14T12:33:37.931059"
  },
  {
    "epoch": 129,
    "batch_idx": 0,
    "loss": 0.4390731453895569,
    "learning_rate": 0.0009897700699809835,
    "timestep_mean": 474.25,
    "epsilon_norm": 7.8388590812683105,
    "prediction_norm": 5.25400972366333,
    "timestamp": "2025-07-14T12:33:37.935923"
  },
  {
    "epoch": 129,
    "batch_idx": 32,
    "loss": 0.5206271409988403,
    "learning_rate": 0.0009897700699809835,
    "timestep_mean": 515.0,
    "epsilon_norm": 8.423182487487793,
    "prediction_norm": 5.953405857086182,
    "timestamp": "2025-07-14T12:33:37.940762"
  },
  {
    "epoch": 129,
    "batch_idx": 64,
    "loss": 0.4008273184299469,
    "learning_rate": 0.0009897700699809835,
    "timestep_mean": 451.78125,
    "epsilon_norm": 7.015207767486572,
    "prediction_norm": 5.373541831970215,
    "timestamp": "2025-07-14T12:33:37.945637"
  },
  {
    "epoch": 129,
    "batch_idx": 96,
    "loss": 0.3488777279853821,
    "learning_rate": 0.0009897700699809835,
    "timestep_mean": 516.125,
    "epsilon_norm": 8.746654510498047,
    "prediction_norm": 6.782931327819824,
    "timestamp": "2025-07-14T12:33:37.950469"
  },
  {
    "epoch": 129,
    "batch_idx": 128,
    "loss": 0.3583483397960663,
    "learning_rate": 0.0009897700699809835,
    "timestep_mean": 509.46875,
    "epsilon_norm": 8.294775009155273,
    "prediction_norm": 6.575315475463867,
    "timestamp": "2025-07-14T12:33:37.955235"
  },
  {
    "epoch": 129,
    "batch_idx": 160,
    "loss": 0.37956348061561584,
    "learning_rate": 0.0009897700699809835,
    "timestep_mean": 437.8125,
    "epsilon_norm": 8.477044105529785,
    "prediction_norm": 6.91210412979126,
    "timestamp": "2025-07-14T12:33:37.960107"
  },
  {
    "epoch": 129,
    "batch_idx": 192,
    "loss": 0.3571981191635132,
    "learning_rate": 0.0009897700699809835,
    "timestep_mean": 484.75,
    "epsilon_norm": 3.1949257850646973,
    "prediction_norm": 2.9314277172088623,
    "timestamp": "2025-07-14T12:33:37.964931"
  },
  {
    "epoch": 130,
    "batch_idx": 0,
    "loss": 0.49334096908569336,
    "learning_rate": 0.0009896114053108828,
    "timestep_mean": 481.125,
    "epsilon_norm": 8.25217056274414,
    "prediction_norm": 5.95289421081543,
    "timestamp": "2025-07-14T12:33:37.969792"
  },
  {
    "epoch": 130,
    "batch_idx": 32,
    "loss": 0.374250590801239,
    "learning_rate": 0.0009896114053108828,
    "timestep_mean": 488.4375,
    "epsilon_norm": 7.03187370300293,
    "prediction_norm": 5.283829689025879,
    "timestamp": "2025-07-14T12:33:37.974656"
  },
  {
    "epoch": 130,
    "batch_idx": 64,
    "loss": 0.28227120637893677,
    "learning_rate": 0.0009896114053108828,
    "timestep_mean": 483.75,
    "epsilon_norm": 7.406398773193359,
    "prediction_norm": 6.063229084014893,
    "timestamp": "2025-07-14T12:33:37.979545"
  },
  {
    "epoch": 130,
    "batch_idx": 96,
    "loss": 0.2624848484992981,
    "learning_rate": 0.0009896114053108828,
    "timestep_mean": 624.90625,
    "epsilon_norm": 7.992262840270996,
    "prediction_norm": 6.672957420349121,
    "timestamp": "2025-07-14T12:33:37.984418"
  },
  {
    "epoch": 130,
    "batch_idx": 128,
    "loss": 0.35307490825653076,
    "learning_rate": 0.0009896114053108828,
    "timestep_mean": 492.28125,
    "epsilon_norm": 7.2156243324279785,
    "prediction_norm": 5.338123321533203,
    "timestamp": "2025-07-14T12:33:37.989265"
  },
  {
    "epoch": 130,
    "batch_idx": 160,
    "loss": 0.3856384754180908,
    "learning_rate": 0.0009896114053108828,
    "timestep_mean": 451.46875,
    "epsilon_norm": 7.458943843841553,
    "prediction_norm": 5.073027610778809,
    "timestamp": "2025-07-14T12:33:37.994092"
  },
  {
    "epoch": 130,
    "batch_idx": 192,
    "loss": 0.35669153928756714,
    "learning_rate": 0.0009896114053108828,
    "timestep_mean": 579.375,
    "epsilon_norm": 4.180965423583984,
    "prediction_norm": 3.269686222076416,
    "timestamp": "2025-07-14T12:33:37.998911"
  },
  {
    "epoch": 131,
    "batch_idx": 0,
    "loss": 0.5236492156982422,
    "learning_rate": 0.0009894515325733102,
    "timestep_mean": 440.40625,
    "epsilon_norm": 9.745505332946777,
    "prediction_norm": 7.735624313354492,
    "timestamp": "2025-07-14T12:33:38.003695"
  },
  {
    "epoch": 131,
    "batch_idx": 32,
    "loss": 0.450142502784729,
    "learning_rate": 0.0009894515325733102,
    "timestep_mean": 526.96875,
    "epsilon_norm": 8.3966703414917,
    "prediction_norm": 6.818564414978027,
    "timestamp": "2025-07-14T12:33:38.008648"
  },
  {
    "epoch": 131,
    "batch_idx": 64,
    "loss": 0.36445093154907227,
    "learning_rate": 0.0009894515325733102,
    "timestep_mean": 494.28125,
    "epsilon_norm": 7.729453086853027,
    "prediction_norm": 5.674661159515381,
    "timestamp": "2025-07-14T12:33:38.013459"
  },
  {
    "epoch": 131,
    "batch_idx": 96,
    "loss": 0.42175325751304626,
    "learning_rate": 0.0009894515325733102,
    "timestep_mean": 547.65625,
    "epsilon_norm": 8.662545204162598,
    "prediction_norm": 7.233233451843262,
    "timestamp": "2025-07-14T12:33:38.018313"
  },
  {
    "epoch": 131,
    "batch_idx": 128,
    "loss": 0.3396119475364685,
    "learning_rate": 0.0009894515325733102,
    "timestep_mean": 474.40625,
    "epsilon_norm": 6.919595241546631,
    "prediction_norm": 5.435597896575928,
    "timestamp": "2025-07-14T12:33:38.023149"
  },
  {
    "epoch": 131,
    "batch_idx": 160,
    "loss": 0.24552908539772034,
    "learning_rate": 0.0009894515325733102,
    "timestep_mean": 541.15625,
    "epsilon_norm": 8.233838081359863,
    "prediction_norm": 8.026461601257324,
    "timestamp": "2025-07-14T12:33:38.027964"
  },
  {
    "epoch": 131,
    "batch_idx": 192,
    "loss": 0.3579411804676056,
    "learning_rate": 0.0009894515325733102,
    "timestep_mean": 545.875,
    "epsilon_norm": 4.3465142250061035,
    "prediction_norm": 4.353713512420654,
    "timestamp": "2025-07-14T12:33:38.032830"
  },
  {
    "epoch": 132,
    "batch_idx": 0,
    "loss": 0.3685416579246521,
    "learning_rate": 0.000989290452162736,
    "timestep_mean": 541.0625,
    "epsilon_norm": 7.674282073974609,
    "prediction_norm": 6.254162311553955,
    "timestamp": "2025-07-14T12:33:38.037717"
  },
  {
    "epoch": 132,
    "batch_idx": 32,
    "loss": 0.366532564163208,
    "learning_rate": 0.000989290452162736,
    "timestep_mean": 500.0625,
    "epsilon_norm": 8.286336898803711,
    "prediction_norm": 6.220160007476807,
    "timestamp": "2025-07-14T12:33:38.042564"
  },
  {
    "epoch": 132,
    "batch_idx": 64,
    "loss": 0.24867741763591766,
    "learning_rate": 0.000989290452162736,
    "timestep_mean": 555.71875,
    "epsilon_norm": 8.588746070861816,
    "prediction_norm": 7.945389747619629,
    "timestamp": "2025-07-14T12:33:38.047397"
  },
  {
    "epoch": 132,
    "batch_idx": 96,
    "loss": 0.3414137363433838,
    "learning_rate": 0.000989290452162736,
    "timestep_mean": 512.96875,
    "epsilon_norm": 7.8024091720581055,
    "prediction_norm": 6.228686332702637,
    "timestamp": "2025-07-14T12:33:38.052259"
  },
  {
    "epoch": 132,
    "batch_idx": 128,
    "loss": 0.4587099254131317,
    "learning_rate": 0.000989290452162736,
    "timestep_mean": 485.875,
    "epsilon_norm": 8.750106811523438,
    "prediction_norm": 7.076203346252441,
    "timestamp": "2025-07-14T12:33:38.057112"
  },
  {
    "epoch": 132,
    "batch_idx": 160,
    "loss": 0.33114659786224365,
    "learning_rate": 0.000989290452162736,
    "timestep_mean": 464.21875,
    "epsilon_norm": 7.5585408210754395,
    "prediction_norm": 5.56190824508667,
    "timestamp": "2025-07-14T12:33:38.061977"
  },
  {
    "epoch": 132,
    "batch_idx": 192,
    "loss": 0.4007032513618469,
    "learning_rate": 0.000989290452162736,
    "timestep_mean": 338.875,
    "epsilon_norm": 3.4588818550109863,
    "prediction_norm": 2.4284260272979736,
    "timestamp": "2025-07-14T12:33:38.066809"
  },
  {
    "epoch": 133,
    "batch_idx": 0,
    "loss": 0.4251295328140259,
    "learning_rate": 0.00098912816447661,
    "timestep_mean": 528.40625,
    "epsilon_norm": 10.26699161529541,
    "prediction_norm": 8.023534774780273,
    "timestamp": "2025-07-14T12:33:38.071637"
  },
  {
    "epoch": 133,
    "batch_idx": 32,
    "loss": 0.34614309668540955,
    "learning_rate": 0.00098912816447661,
    "timestep_mean": 477.09375,
    "epsilon_norm": 7.495998859405518,
    "prediction_norm": 6.1500725746154785,
    "timestamp": "2025-07-14T12:33:38.076450"
  },
  {
    "epoch": 133,
    "batch_idx": 64,
    "loss": 0.49581438302993774,
    "learning_rate": 0.00098912816447661,
    "timestep_mean": 505.3125,
    "epsilon_norm": 9.675564765930176,
    "prediction_norm": 7.233155250549316,
    "timestamp": "2025-07-14T12:33:38.081261"
  },
  {
    "epoch": 133,
    "batch_idx": 96,
    "loss": 0.39312103390693665,
    "learning_rate": 0.00098912816447661,
    "timestep_mean": 499.65625,
    "epsilon_norm": 7.003270149230957,
    "prediction_norm": 5.456074237823486,
    "timestamp": "2025-07-14T12:33:38.086119"
  },
  {
    "epoch": 133,
    "batch_idx": 128,
    "loss": 0.3973202705383301,
    "learning_rate": 0.00098912816447661,
    "timestep_mean": 482.875,
    "epsilon_norm": 8.316731452941895,
    "prediction_norm": 6.180601596832275,
    "timestamp": "2025-07-14T12:33:38.090954"
  },
  {
    "epoch": 133,
    "batch_idx": 160,
    "loss": 0.39319145679473877,
    "learning_rate": 0.00098912816447661,
    "timestep_mean": 542.71875,
    "epsilon_norm": 7.576585292816162,
    "prediction_norm": 5.268013000488281,
    "timestamp": "2025-07-14T12:33:38.095823"
  },
  {
    "epoch": 133,
    "batch_idx": 192,
    "loss": 0.37318965792655945,
    "learning_rate": 0.00098912816447661,
    "timestep_mean": 484.75,
    "epsilon_norm": 4.058908939361572,
    "prediction_norm": 3.5048553943634033,
    "timestamp": "2025-07-14T12:33:38.105143"
  },
  {
    "epoch": 134,
    "batch_idx": 0,
    "loss": 0.5099559426307678,
    "learning_rate": 0.000988964669915361,
    "timestep_mean": 374.5,
    "epsilon_norm": 7.808992862701416,
    "prediction_norm": 5.323095321655273,
    "timestamp": "2025-07-14T12:33:38.110030"
  },
  {
    "epoch": 134,
    "batch_idx": 32,
    "loss": 0.21073198318481445,
    "learning_rate": 0.000988964669915361,
    "timestep_mean": 520.09375,
    "epsilon_norm": 7.943885803222656,
    "prediction_norm": 7.525602340698242,
    "timestamp": "2025-07-14T12:33:38.114785"
  },
  {
    "epoch": 134,
    "batch_idx": 64,
    "loss": 0.383453369140625,
    "learning_rate": 0.000988964669915361,
    "timestep_mean": 548.6875,
    "epsilon_norm": 9.512195587158203,
    "prediction_norm": 7.941426753997803,
    "timestamp": "2025-07-14T12:33:38.119477"
  },
  {
    "epoch": 134,
    "batch_idx": 96,
    "loss": 0.43060076236724854,
    "learning_rate": 0.000988964669915361,
    "timestep_mean": 521.21875,
    "epsilon_norm": 8.038308143615723,
    "prediction_norm": 6.352415084838867,
    "timestamp": "2025-07-14T12:33:38.124204"
  },
  {
    "epoch": 134,
    "batch_idx": 128,
    "loss": 0.8230355978012085,
    "learning_rate": 0.000988964669915361,
    "timestep_mean": 406.1875,
    "epsilon_norm": 9.314943313598633,
    "prediction_norm": 5.319003105163574,
    "timestamp": "2025-07-14T12:33:38.129011"
  },
  {
    "epoch": 134,
    "batch_idx": 160,
    "loss": 0.3050346076488495,
    "learning_rate": 0.000988964669915361,
    "timestep_mean": 505.25,
    "epsilon_norm": 8.631148338317871,
    "prediction_norm": 7.10536527633667,
    "timestamp": "2025-07-14T12:33:38.133737"
  },
  {
    "epoch": 134,
    "batch_idx": 192,
    "loss": 0.8107174634933472,
    "learning_rate": 0.000988964669915361,
    "timestep_mean": 401.0,
    "epsilon_norm": 4.1479172706604,
    "prediction_norm": 2.3994174003601074,
    "timestamp": "2025-07-14T12:33:38.138515"
  },
  {
    "epoch": 135,
    "batch_idx": 0,
    "loss": 0.32839760184288025,
    "learning_rate": 0.0009887999688823954,
    "timestep_mean": 592.65625,
    "epsilon_norm": 8.053442001342773,
    "prediction_norm": 6.388841152191162,
    "timestamp": "2025-07-14T12:33:38.143233"
  },
  {
    "epoch": 135,
    "batch_idx": 32,
    "loss": 0.32864177227020264,
    "learning_rate": 0.0009887999688823954,
    "timestep_mean": 517.65625,
    "epsilon_norm": 7.64168643951416,
    "prediction_norm": 6.178993225097656,
    "timestamp": "2025-07-14T12:33:38.147969"
  },
  {
    "epoch": 135,
    "batch_idx": 64,
    "loss": 0.6050982475280762,
    "learning_rate": 0.0009887999688823954,
    "timestep_mean": 420.59375,
    "epsilon_norm": 8.999377250671387,
    "prediction_norm": 6.948339462280273,
    "timestamp": "2025-07-14T12:33:38.152690"
  },
  {
    "epoch": 135,
    "batch_idx": 96,
    "loss": 0.418062686920166,
    "learning_rate": 0.0009887999688823954,
    "timestep_mean": 519.75,
    "epsilon_norm": 8.269076347351074,
    "prediction_norm": 6.713356971740723,
    "timestamp": "2025-07-14T12:33:38.157398"
  },
  {
    "epoch": 135,
    "batch_idx": 128,
    "loss": 0.3094150424003601,
    "learning_rate": 0.0009887999688823954,
    "timestep_mean": 474.375,
    "epsilon_norm": 7.681648254394531,
    "prediction_norm": 6.887864589691162,
    "timestamp": "2025-07-14T12:33:38.162099"
  },
  {
    "epoch": 135,
    "batch_idx": 160,
    "loss": 0.28137049078941345,
    "learning_rate": 0.0009887999688823954,
    "timestep_mean": 512.53125,
    "epsilon_norm": 7.74337100982666,
    "prediction_norm": 6.520987510681152,
    "timestamp": "2025-07-14T12:33:38.166967"
  },
  {
    "epoch": 135,
    "batch_idx": 192,
    "loss": 0.21395771205425262,
    "learning_rate": 0.0009887999688823954,
    "timestep_mean": 472.75,
    "epsilon_norm": 4.40038537979126,
    "prediction_norm": 3.693993091583252,
    "timestamp": "2025-07-14T12:33:38.171672"
  },
  {
    "epoch": 136,
    "batch_idx": 0,
    "loss": 0.21178001165390015,
    "learning_rate": 0.0009886340617840967,
    "timestep_mean": 537.5,
    "epsilon_norm": 6.835607528686523,
    "prediction_norm": 6.35192346572876,
    "timestamp": "2025-07-14T12:33:38.176392"
  },
  {
    "epoch": 136,
    "batch_idx": 32,
    "loss": 0.2965569794178009,
    "learning_rate": 0.0009886340617840967,
    "timestep_mean": 497.8125,
    "epsilon_norm": 7.734415531158447,
    "prediction_norm": 6.387564182281494,
    "timestamp": "2025-07-14T12:33:38.181174"
  },
  {
    "epoch": 136,
    "batch_idx": 64,
    "loss": 0.4496522545814514,
    "learning_rate": 0.0009886340617840967,
    "timestep_mean": 437.875,
    "epsilon_norm": 7.7870378494262695,
    "prediction_norm": 5.334712505340576,
    "timestamp": "2025-07-14T12:33:38.185886"
  },
  {
    "epoch": 136,
    "batch_idx": 96,
    "loss": 0.6277056932449341,
    "learning_rate": 0.0009886340617840967,
    "timestep_mean": 445.71875,
    "epsilon_norm": 8.778982162475586,
    "prediction_norm": 5.637824058532715,
    "timestamp": "2025-07-14T12:33:38.190681"
  },
  {
    "epoch": 136,
    "batch_idx": 128,
    "loss": 0.5210875272750854,
    "learning_rate": 0.0009886340617840967,
    "timestep_mean": 375.46875,
    "epsilon_norm": 7.757683277130127,
    "prediction_norm": 5.138906478881836,
    "timestamp": "2025-07-14T12:33:38.195458"
  },
  {
    "epoch": 136,
    "batch_idx": 160,
    "loss": 0.31598737835884094,
    "learning_rate": 0.0009886340617840967,
    "timestep_mean": 449.5625,
    "epsilon_norm": 7.634571075439453,
    "prediction_norm": 6.354966163635254,
    "timestamp": "2025-07-14T12:33:38.200349"
  },
  {
    "epoch": 136,
    "batch_idx": 192,
    "loss": 0.4283060133457184,
    "learning_rate": 0.0009886340617840967,
    "timestep_mean": 587.25,
    "epsilon_norm": 3.8928627967834473,
    "prediction_norm": 2.7730777263641357,
    "timestamp": "2025-07-14T12:33:38.205074"
  },
  {
    "epoch": 137,
    "batch_idx": 0,
    "loss": 0.45253998041152954,
    "learning_rate": 0.0009884669490298243,
    "timestep_mean": 455.4375,
    "epsilon_norm": 8.291522979736328,
    "prediction_norm": 6.578347206115723,
    "timestamp": "2025-07-14T12:33:38.209872"
  },
  {
    "epoch": 137,
    "batch_idx": 32,
    "loss": 0.31202811002731323,
    "learning_rate": 0.0009884669490298243,
    "timestep_mean": 577.8125,
    "epsilon_norm": 8.202225685119629,
    "prediction_norm": 6.805773735046387,
    "timestamp": "2025-07-14T12:33:38.214659"
  },
  {
    "epoch": 137,
    "batch_idx": 64,
    "loss": 0.37757036089897156,
    "learning_rate": 0.0009884669490298243,
    "timestep_mean": 528.59375,
    "epsilon_norm": 7.330275535583496,
    "prediction_norm": 5.943345546722412,
    "timestamp": "2025-07-14T12:33:38.219425"
  },
  {
    "epoch": 137,
    "batch_idx": 96,
    "loss": 0.37031063437461853,
    "learning_rate": 0.0009884669490298243,
    "timestep_mean": 498.9375,
    "epsilon_norm": 8.311749458312988,
    "prediction_norm": 6.935544967651367,
    "timestamp": "2025-07-14T12:33:38.224291"
  },
  {
    "epoch": 137,
    "batch_idx": 128,
    "loss": 0.4916289448738098,
    "learning_rate": 0.0009884669490298243,
    "timestep_mean": 457.84375,
    "epsilon_norm": 8.285567283630371,
    "prediction_norm": 6.308594226837158,
    "timestamp": "2025-07-14T12:33:38.229221"
  },
  {
    "epoch": 137,
    "batch_idx": 160,
    "loss": 0.3913017213344574,
    "learning_rate": 0.0009884669490298243,
    "timestep_mean": 472.59375,
    "epsilon_norm": 7.727879524230957,
    "prediction_norm": 6.229140758514404,
    "timestamp": "2025-07-14T12:33:38.234058"
  },
  {
    "epoch": 137,
    "batch_idx": 192,
    "loss": 0.4218443036079407,
    "learning_rate": 0.0009884669490298243,
    "timestep_mean": 579.25,
    "epsilon_norm": 4.574007511138916,
    "prediction_norm": 3.1754488945007324,
    "timestamp": "2025-07-14T12:33:38.238828"
  },
  {
    "epoch": 138,
    "batch_idx": 0,
    "loss": 0.25845348834991455,
    "learning_rate": 0.0009882986310319121,
    "timestep_mean": 457.9375,
    "epsilon_norm": 6.942399024963379,
    "prediction_norm": 6.187239170074463,
    "timestamp": "2025-07-14T12:33:38.243711"
  },
  {
    "epoch": 138,
    "batch_idx": 32,
    "loss": 0.2935720682144165,
    "learning_rate": 0.0009882986310319121,
    "timestep_mean": 509.5,
    "epsilon_norm": 7.74373197555542,
    "prediction_norm": 6.283901691436768,
    "timestamp": "2025-07-14T12:33:38.248595"
  },
  {
    "epoch": 138,
    "batch_idx": 64,
    "loss": 0.4715535044670105,
    "learning_rate": 0.0009882986310319121,
    "timestep_mean": 493.4375,
    "epsilon_norm": 8.679581642150879,
    "prediction_norm": 6.779518127441406,
    "timestamp": "2025-07-14T12:33:38.253523"
  },
  {
    "epoch": 138,
    "batch_idx": 96,
    "loss": 0.3279935121536255,
    "learning_rate": 0.0009882986310319121,
    "timestep_mean": 452.03125,
    "epsilon_norm": 6.936617851257324,
    "prediction_norm": 5.409893035888672,
    "timestamp": "2025-07-14T12:33:38.258478"
  },
  {
    "epoch": 138,
    "batch_idx": 128,
    "loss": 0.25319159030914307,
    "learning_rate": 0.0009882986310319121,
    "timestep_mean": 556.65625,
    "epsilon_norm": 7.208668231964111,
    "prediction_norm": 6.196298122406006,
    "timestamp": "2025-07-14T12:33:38.263318"
  },
  {
    "epoch": 138,
    "batch_idx": 160,
    "loss": 0.5262362360954285,
    "learning_rate": 0.0009882986310319121,
    "timestep_mean": 464.625,
    "epsilon_norm": 9.155436515808105,
    "prediction_norm": 7.573394298553467,
    "timestamp": "2025-07-14T12:33:38.268182"
  },
  {
    "epoch": 138,
    "batch_idx": 192,
    "loss": 0.14214351773262024,
    "learning_rate": 0.0009882986310319121,
    "timestep_mean": 590.0,
    "epsilon_norm": 4.16187047958374,
    "prediction_norm": 3.834340810775757,
    "timestamp": "2025-07-14T12:33:38.273186"
  },
  {
    "epoch": 139,
    "batch_idx": 0,
    "loss": 0.4782210886478424,
    "learning_rate": 0.0009881291082056683,
    "timestep_mean": 446.1875,
    "epsilon_norm": 7.6037678718566895,
    "prediction_norm": 5.49184513092041,
    "timestamp": "2025-07-14T12:33:38.278232"
  },
  {
    "epoch": 139,
    "batch_idx": 32,
    "loss": 0.44374793767929077,
    "learning_rate": 0.0009881291082056683,
    "timestep_mean": 515.1875,
    "epsilon_norm": 8.36764907836914,
    "prediction_norm": 6.3941969871521,
    "timestamp": "2025-07-14T12:33:38.283426"
  },
  {
    "epoch": 139,
    "batch_idx": 64,
    "loss": 0.49790671467781067,
    "learning_rate": 0.0009881291082056683,
    "timestep_mean": 466.9375,
    "epsilon_norm": 7.950918197631836,
    "prediction_norm": 5.5335822105407715,
    "timestamp": "2025-07-14T12:33:38.288436"
  },
  {
    "epoch": 139,
    "batch_idx": 96,
    "loss": 0.3732576072216034,
    "learning_rate": 0.0009881291082056683,
    "timestep_mean": 521.8125,
    "epsilon_norm": 8.015997886657715,
    "prediction_norm": 6.284470558166504,
    "timestamp": "2025-07-14T12:33:38.293535"
  },
  {
    "epoch": 139,
    "batch_idx": 128,
    "loss": 0.2532110810279846,
    "learning_rate": 0.0009881291082056683,
    "timestep_mean": 577.625,
    "epsilon_norm": 7.619199275970459,
    "prediction_norm": 5.936221122741699,
    "timestamp": "2025-07-14T12:33:38.298642"
  },
  {
    "epoch": 139,
    "batch_idx": 160,
    "loss": 0.5067905187606812,
    "learning_rate": 0.0009881291082056683,
    "timestep_mean": 465.5,
    "epsilon_norm": 7.902256488800049,
    "prediction_norm": 4.7753729820251465,
    "timestamp": "2025-07-14T12:33:38.303745"
  },
  {
    "epoch": 139,
    "batch_idx": 192,
    "loss": 0.5606168508529663,
    "learning_rate": 0.0009881291082056683,
    "timestep_mean": 403.125,
    "epsilon_norm": 4.062267303466797,
    "prediction_norm": 2.342392683029175,
    "timestamp": "2025-07-14T12:33:38.308943"
  },
  {
    "epoch": 140,
    "batch_idx": 0,
    "loss": 0.39623114466667175,
    "learning_rate": 0.0009879583809693736,
    "timestep_mean": 430.34375,
    "epsilon_norm": 6.935153484344482,
    "prediction_norm": 4.64206600189209,
    "timestamp": "2025-07-14T12:33:38.313961"
  },
  {
    "epoch": 140,
    "batch_idx": 32,
    "loss": 0.4268863797187805,
    "learning_rate": 0.0009879583809693736,
    "timestep_mean": 458.53125,
    "epsilon_norm": 8.46947193145752,
    "prediction_norm": 6.510613441467285,
    "timestamp": "2025-07-14T12:33:38.319051"
  },
  {
    "epoch": 140,
    "batch_idx": 64,
    "loss": 0.39004939794540405,
    "learning_rate": 0.0009879583809693736,
    "timestep_mean": 492.1875,
    "epsilon_norm": 8.37269401550293,
    "prediction_norm": 6.94039249420166,
    "timestamp": "2025-07-14T12:33:38.324224"
  },
  {
    "epoch": 140,
    "batch_idx": 96,
    "loss": 0.2617003917694092,
    "learning_rate": 0.0009879583809693736,
    "timestep_mean": 493.65625,
    "epsilon_norm": 8.018271446228027,
    "prediction_norm": 8.00015640258789,
    "timestamp": "2025-07-14T12:33:38.329474"
  },
  {
    "epoch": 140,
    "batch_idx": 128,
    "loss": 0.29852569103240967,
    "learning_rate": 0.0009879583809693736,
    "timestep_mean": 535.78125,
    "epsilon_norm": 8.38676929473877,
    "prediction_norm": 7.649125099182129,
    "timestamp": "2025-07-14T12:33:38.334619"
  },
  {
    "epoch": 140,
    "batch_idx": 160,
    "loss": 0.3760351836681366,
    "learning_rate": 0.0009879583809693736,
    "timestep_mean": 470.1875,
    "epsilon_norm": 7.985918998718262,
    "prediction_norm": 6.1278886795043945,
    "timestamp": "2025-07-14T12:33:38.339746"
  },
  {
    "epoch": 140,
    "batch_idx": 192,
    "loss": 0.5272355079650879,
    "learning_rate": 0.0009879583809693736,
    "timestep_mean": 434.375,
    "epsilon_norm": 3.958300828933716,
    "prediction_norm": 2.5235445499420166,
    "timestamp": "2025-07-14T12:33:38.344734"
  },
  {
    "epoch": 141,
    "batch_idx": 0,
    "loss": 0.33872318267822266,
    "learning_rate": 0.0009877864497442803,
    "timestep_mean": 595.5625,
    "epsilon_norm": 7.587008476257324,
    "prediction_norm": 6.246509552001953,
    "timestamp": "2025-07-14T12:33:38.349940"
  },
  {
    "epoch": 141,
    "batch_idx": 32,
    "loss": 0.4037342965602875,
    "learning_rate": 0.0009877864497442803,
    "timestep_mean": 476.5,
    "epsilon_norm": 8.287652015686035,
    "prediction_norm": 6.541762351989746,
    "timestamp": "2025-07-14T12:33:38.355039"
  },
  {
    "epoch": 141,
    "batch_idx": 64,
    "loss": 0.40098121762275696,
    "learning_rate": 0.0009877864497442803,
    "timestep_mean": 506.71875,
    "epsilon_norm": 7.387403964996338,
    "prediction_norm": 5.9848551750183105,
    "timestamp": "2025-07-14T12:33:38.360124"
  },
  {
    "epoch": 141,
    "batch_idx": 96,
    "loss": 0.390341579914093,
    "learning_rate": 0.0009877864497442803,
    "timestep_mean": 489.78125,
    "epsilon_norm": 8.397212982177734,
    "prediction_norm": 6.332882881164551,
    "timestamp": "2025-07-14T12:33:38.365348"
  },
  {
    "epoch": 141,
    "batch_idx": 128,
    "loss": 0.44941434264183044,
    "learning_rate": 0.0009877864497442803,
    "timestep_mean": 413.6875,
    "epsilon_norm": 7.436962127685547,
    "prediction_norm": 5.3649582862854,
    "timestamp": "2025-07-14T12:33:38.370652"
  },
  {
    "epoch": 141,
    "batch_idx": 160,
    "loss": 0.3208995461463928,
    "learning_rate": 0.0009877864497442803,
    "timestep_mean": 576.40625,
    "epsilon_norm": 8.391380310058594,
    "prediction_norm": 7.120680809020996,
    "timestamp": "2025-07-14T12:33:38.375785"
  },
  {
    "epoch": 141,
    "batch_idx": 192,
    "loss": 0.2931937575340271,
    "learning_rate": 0.0009877864497442803,
    "timestep_mean": 445.5,
    "epsilon_norm": 3.665423631668091,
    "prediction_norm": 2.490936279296875,
    "timestamp": "2025-07-14T12:33:38.380883"
  },
  {
    "epoch": 142,
    "batch_idx": 0,
    "loss": 0.3300848603248596,
    "learning_rate": 0.0009876133149546117,
    "timestep_mean": 494.5,
    "epsilon_norm": 8.19133186340332,
    "prediction_norm": 6.676671981811523,
    "timestamp": "2025-07-14T12:33:38.385877"
  },
  {
    "epoch": 142,
    "batch_idx": 32,
    "loss": 0.48677679896354675,
    "learning_rate": 0.0009876133149546117,
    "timestep_mean": 478.84375,
    "epsilon_norm": 8.591464042663574,
    "prediction_norm": 7.133718013763428,
    "timestamp": "2025-07-14T12:33:38.390865"
  },
  {
    "epoch": 142,
    "batch_idx": 64,
    "loss": 0.607094407081604,
    "learning_rate": 0.0009876133149546117,
    "timestep_mean": 484.21875,
    "epsilon_norm": 8.869564056396484,
    "prediction_norm": 7.195720195770264,
    "timestamp": "2025-07-14T12:33:38.396107"
  },
  {
    "epoch": 142,
    "batch_idx": 96,
    "loss": 0.5250273942947388,
    "learning_rate": 0.0009876133149546117,
    "timestep_mean": 476.40625,
    "epsilon_norm": 8.646220207214355,
    "prediction_norm": 6.6511969566345215,
    "timestamp": "2025-07-14T12:33:38.401033"
  },
  {
    "epoch": 142,
    "batch_idx": 128,
    "loss": 0.42697179317474365,
    "learning_rate": 0.0009876133149546117,
    "timestep_mean": 527.90625,
    "epsilon_norm": 6.872282981872559,
    "prediction_norm": 5.69917106628418,
    "timestamp": "2025-07-14T12:33:38.405899"
  },
  {
    "epoch": 142,
    "batch_idx": 160,
    "loss": 0.17829996347427368,
    "learning_rate": 0.0009876133149546117,
    "timestep_mean": 626.4375,
    "epsilon_norm": 7.060296535491943,
    "prediction_norm": 6.585159778594971,
    "timestamp": "2025-07-14T12:33:38.410710"
  },
  {
    "epoch": 142,
    "batch_idx": 192,
    "loss": 0.12953375279903412,
    "learning_rate": 0.0009876133149546117,
    "timestep_mean": 484.25,
    "epsilon_norm": 3.357912540435791,
    "prediction_norm": 2.9941225051879883,
    "timestamp": "2025-07-14T12:33:38.415906"
  },
  {
    "epoch": 143,
    "batch_idx": 0,
    "loss": 0.5323273539543152,
    "learning_rate": 0.0009874389770275606,
    "timestep_mean": 503.09375,
    "epsilon_norm": 9.159379959106445,
    "prediction_norm": 6.235047340393066,
    "timestamp": "2025-07-14T12:33:38.421089"
  },
  {
    "epoch": 143,
    "batch_idx": 32,
    "loss": 0.2898102402687073,
    "learning_rate": 0.0009874389770275606,
    "timestep_mean": 533.15625,
    "epsilon_norm": 7.557284355163574,
    "prediction_norm": 6.151182651519775,
    "timestamp": "2025-07-14T12:33:38.426156"
  },
  {
    "epoch": 143,
    "batch_idx": 64,
    "loss": 0.29406630992889404,
    "learning_rate": 0.0009874389770275606,
    "timestep_mean": 475.125,
    "epsilon_norm": 7.533824920654297,
    "prediction_norm": 6.21959114074707,
    "timestamp": "2025-07-14T12:33:38.431164"
  },
  {
    "epoch": 143,
    "batch_idx": 96,
    "loss": 0.26555192470550537,
    "learning_rate": 0.0009874389770275606,
    "timestep_mean": 575.625,
    "epsilon_norm": 7.803958892822266,
    "prediction_norm": 6.341370582580566,
    "timestamp": "2025-07-14T12:33:38.436286"
  },
  {
    "epoch": 143,
    "batch_idx": 128,
    "loss": 0.3497627377510071,
    "learning_rate": 0.0009874389770275606,
    "timestep_mean": 495.8125,
    "epsilon_norm": 8.489578247070312,
    "prediction_norm": 6.373632907867432,
    "timestamp": "2025-07-14T12:33:38.441371"
  },
  {
    "epoch": 143,
    "batch_idx": 160,
    "loss": 0.28527554869651794,
    "learning_rate": 0.0009874389770275606,
    "timestep_mean": 522.15625,
    "epsilon_norm": 7.157426834106445,
    "prediction_norm": 5.7528204917907715,
    "timestamp": "2025-07-14T12:33:38.446510"
  },
  {
    "epoch": 143,
    "batch_idx": 192,
    "loss": 0.21793778240680695,
    "learning_rate": 0.0009874389770275606,
    "timestep_mean": 560.625,
    "epsilon_norm": 4.033718109130859,
    "prediction_norm": 3.571589946746826,
    "timestamp": "2025-07-14T12:33:38.451594"
  },
  {
    "epoch": 144,
    "batch_idx": 0,
    "loss": 0.1964206099510193,
    "learning_rate": 0.0009872634363932886,
    "timestep_mean": 491.71875,
    "epsilon_norm": 6.8155598640441895,
    "prediction_norm": 5.708065986633301,
    "timestamp": "2025-07-14T12:33:38.456650"
  },
  {
    "epoch": 144,
    "batch_idx": 32,
    "loss": 0.5146068930625916,
    "learning_rate": 0.0009872634363932886,
    "timestep_mean": 438.9375,
    "epsilon_norm": 7.736423015594482,
    "prediction_norm": 5.700151443481445,
    "timestamp": "2025-07-14T12:33:38.461874"
  },
  {
    "epoch": 144,
    "batch_idx": 64,
    "loss": 0.42846500873565674,
    "learning_rate": 0.0009872634363932886,
    "timestep_mean": 518.21875,
    "epsilon_norm": 7.536994457244873,
    "prediction_norm": 5.656781196594238,
    "timestamp": "2025-07-14T12:33:38.467098"
  },
  {
    "epoch": 144,
    "batch_idx": 96,
    "loss": 0.2749977111816406,
    "learning_rate": 0.0009872634363932886,
    "timestep_mean": 522.90625,
    "epsilon_norm": 7.645637512207031,
    "prediction_norm": 6.252374649047852,
    "timestamp": "2025-07-14T12:33:38.472232"
  },
  {
    "epoch": 144,
    "batch_idx": 128,
    "loss": 0.39560404419898987,
    "learning_rate": 0.0009872634363932886,
    "timestep_mean": 453.40625,
    "epsilon_norm": 7.913628101348877,
    "prediction_norm": 5.552348613739014,
    "timestamp": "2025-07-14T12:33:38.477292"
  },
  {
    "epoch": 144,
    "batch_idx": 160,
    "loss": 0.44675102829933167,
    "learning_rate": 0.0009872634363932886,
    "timestep_mean": 380.21875,
    "epsilon_norm": 7.625168323516846,
    "prediction_norm": 4.96535062789917,
    "timestamp": "2025-07-14T12:33:38.482393"
  },
  {
    "epoch": 144,
    "batch_idx": 192,
    "loss": 0.5153401494026184,
    "learning_rate": 0.0009872634363932886,
    "timestep_mean": 491.625,
    "epsilon_norm": 4.98394250869751,
    "prediction_norm": 3.9278998374938965,
    "timestamp": "2025-07-14T12:33:38.487528"
  },
  {
    "epoch": 145,
    "batch_idx": 0,
    "loss": 0.3622555732727051,
    "learning_rate": 0.0009870866934849247,
    "timestep_mean": 494.6875,
    "epsilon_norm": 7.6224045753479,
    "prediction_norm": 6.013622760772705,
    "timestamp": "2025-07-14T12:33:38.492640"
  },
  {
    "epoch": 145,
    "batch_idx": 32,
    "loss": 0.17222416400909424,
    "learning_rate": 0.0009870866934849247,
    "timestep_mean": 528.125,
    "epsilon_norm": 5.677306175231934,
    "prediction_norm": 4.994548320770264,
    "timestamp": "2025-07-14T12:33:38.497666"
  },
  {
    "epoch": 145,
    "batch_idx": 64,
    "loss": 0.41665273904800415,
    "learning_rate": 0.0009870866934849247,
    "timestep_mean": 503.84375,
    "epsilon_norm": 8.052787780761719,
    "prediction_norm": 6.29526948928833,
    "timestamp": "2025-07-14T12:33:38.502882"
  },
  {
    "epoch": 145,
    "batch_idx": 96,
    "loss": 0.28145164251327515,
    "learning_rate": 0.0009870866934849247,
    "timestep_mean": 524.90625,
    "epsilon_norm": 8.228402137756348,
    "prediction_norm": 6.8072309494018555,
    "timestamp": "2025-07-14T12:33:38.508009"
  },
  {
    "epoch": 145,
    "batch_idx": 128,
    "loss": 0.37654420733451843,
    "learning_rate": 0.0009870866934849247,
    "timestep_mean": 520.8125,
    "epsilon_norm": 9.300224304199219,
    "prediction_norm": 6.928101539611816,
    "timestamp": "2025-07-14T12:33:38.513181"
  },
  {
    "epoch": 145,
    "batch_idx": 160,
    "loss": 0.4143930673599243,
    "learning_rate": 0.0009870866934849247,
    "timestep_mean": 478.625,
    "epsilon_norm": 7.98581075668335,
    "prediction_norm": 6.299632549285889,
    "timestamp": "2025-07-14T12:33:38.518252"
  },
  {
    "epoch": 145,
    "batch_idx": 192,
    "loss": 0.20851632952690125,
    "learning_rate": 0.0009870866934849247,
    "timestep_mean": 497.25,
    "epsilon_norm": 3.4822986125946045,
    "prediction_norm": 3.6303012371063232,
    "timestamp": "2025-07-14T12:33:38.523316"
  },
  {
    "epoch": 146,
    "batch_idx": 0,
    "loss": 0.5425331592559814,
    "learning_rate": 0.0009869087487385643,
    "timestep_mean": 449.65625,
    "epsilon_norm": 8.088418006896973,
    "prediction_norm": 5.782394886016846,
    "timestamp": "2025-07-14T12:33:38.528547"
  },
  {
    "epoch": 146,
    "batch_idx": 32,
    "loss": 0.5139948129653931,
    "learning_rate": 0.0009869087487385643,
    "timestep_mean": 460.4375,
    "epsilon_norm": 8.57253646850586,
    "prediction_norm": 6.548473834991455,
    "timestamp": "2025-07-14T12:33:38.533809"
  },
  {
    "epoch": 146,
    "batch_idx": 64,
    "loss": 0.18755453824996948,
    "learning_rate": 0.0009869087487385643,
    "timestep_mean": 533.09375,
    "epsilon_norm": 7.4386186599731445,
    "prediction_norm": 6.726804733276367,
    "timestamp": "2025-07-14T12:33:38.538875"
  },
  {
    "epoch": 146,
    "batch_idx": 96,
    "loss": 0.4886438250541687,
    "learning_rate": 0.0009869087487385643,
    "timestep_mean": 447.875,
    "epsilon_norm": 8.031147003173828,
    "prediction_norm": 5.556299686431885,
    "timestamp": "2025-07-14T12:33:38.544072"
  },
  {
    "epoch": 146,
    "batch_idx": 128,
    "loss": 0.5679891109466553,
    "learning_rate": 0.0009869087487385643,
    "timestep_mean": 436.375,
    "epsilon_norm": 8.475266456604004,
    "prediction_norm": 5.997188091278076,
    "timestamp": "2025-07-14T12:33:38.549202"
  },
  {
    "epoch": 146,
    "batch_idx": 160,
    "loss": 0.6021931767463684,
    "learning_rate": 0.0009869087487385643,
    "timestep_mean": 458.625,
    "epsilon_norm": 8.848882675170898,
    "prediction_norm": 6.6161017417907715,
    "timestamp": "2025-07-14T12:33:38.554538"
  },
  {
    "epoch": 146,
    "batch_idx": 192,
    "loss": 0.206993967294693,
    "learning_rate": 0.0009869087487385643,
    "timestep_mean": 438.25,
    "epsilon_norm": 2.2369372844696045,
    "prediction_norm": 1.7092275619506836,
    "timestamp": "2025-07-14T12:33:38.559403"
  },
  {
    "epoch": 147,
    "batch_idx": 0,
    "loss": 0.24298112094402313,
    "learning_rate": 0.0009867296025932687,
    "timestep_mean": 574.8125,
    "epsilon_norm": 7.7104172706604,
    "prediction_norm": 7.09229040145874,
    "timestamp": "2025-07-14T12:33:38.564413"
  },
  {
    "epoch": 147,
    "batch_idx": 32,
    "loss": 0.25164347887039185,
    "learning_rate": 0.0009867296025932687,
    "timestep_mean": 561.8125,
    "epsilon_norm": 7.341259479522705,
    "prediction_norm": 6.47219705581665,
    "timestamp": "2025-07-14T12:33:38.569586"
  },
  {
    "epoch": 147,
    "batch_idx": 64,
    "loss": 0.2997892498970032,
    "learning_rate": 0.0009867296025932687,
    "timestep_mean": 503.5,
    "epsilon_norm": 7.69356107711792,
    "prediction_norm": 7.140072822570801,
    "timestamp": "2025-07-14T12:33:38.574777"
  },
  {
    "epoch": 147,
    "batch_idx": 96,
    "loss": 0.5946504473686218,
    "learning_rate": 0.0009867296025932687,
    "timestep_mean": 401.59375,
    "epsilon_norm": 8.642583847045898,
    "prediction_norm": 5.648366451263428,
    "timestamp": "2025-07-14T12:33:38.580019"
  },
  {
    "epoch": 147,
    "batch_idx": 128,
    "loss": 0.23769769072532654,
    "learning_rate": 0.0009867296025932687,
    "timestep_mean": 457.03125,
    "epsilon_norm": 7.061295509338379,
    "prediction_norm": 5.943256378173828,
    "timestamp": "2025-07-14T12:33:38.585100"
  },
  {
    "epoch": 147,
    "batch_idx": 160,
    "loss": 0.22388646006584167,
    "learning_rate": 0.0009867296025932687,
    "timestep_mean": 509.78125,
    "epsilon_norm": 7.80853796005249,
    "prediction_norm": 6.798792839050293,
    "timestamp": "2025-07-14T12:33:38.590118"
  },
  {
    "epoch": 147,
    "batch_idx": 192,
    "loss": 0.08428899943828583,
    "learning_rate": 0.0009867296025932687,
    "timestep_mean": 660.0,
    "epsilon_norm": 3.233792543411255,
    "prediction_norm": 2.992619752883911,
    "timestamp": "2025-07-14T12:33:38.595271"
  },
  {
    "epoch": 148,
    "batch_idx": 0,
    "loss": 0.5060014724731445,
    "learning_rate": 0.0009865492554910632,
    "timestep_mean": 572.9375,
    "epsilon_norm": 7.996522426605225,
    "prediction_norm": 5.631877422332764,
    "timestamp": "2025-07-14T12:33:38.600418"
  },
  {
    "epoch": 148,
    "batch_idx": 32,
    "loss": 0.2163730263710022,
    "learning_rate": 0.0009865492554910632,
    "timestep_mean": 562.15625,
    "epsilon_norm": 8.300383567810059,
    "prediction_norm": 6.8687567710876465,
    "timestamp": "2025-07-14T12:33:38.605571"
  },
  {
    "epoch": 148,
    "batch_idx": 64,
    "loss": 0.3894375264644623,
    "learning_rate": 0.0009865492554910632,
    "timestep_mean": 463.90625,
    "epsilon_norm": 8.332986831665039,
    "prediction_norm": 5.969400405883789,
    "timestamp": "2025-07-14T12:33:38.610671"
  },
  {
    "epoch": 148,
    "batch_idx": 96,
    "loss": 0.1957298070192337,
    "learning_rate": 0.0009865492554910632,
    "timestep_mean": 564.71875,
    "epsilon_norm": 7.637295722961426,
    "prediction_norm": 6.254002094268799,
    "timestamp": "2025-07-14T12:33:38.615746"
  },
  {
    "epoch": 148,
    "batch_idx": 128,
    "loss": 0.33119097352027893,
    "learning_rate": 0.0009865492554910632,
    "timestep_mean": 402.46875,
    "epsilon_norm": 6.882775783538818,
    "prediction_norm": 4.798147201538086,
    "timestamp": "2025-07-14T12:33:38.621069"
  },
  {
    "epoch": 148,
    "batch_idx": 160,
    "loss": 0.2627117335796356,
    "learning_rate": 0.0009865492554910632,
    "timestep_mean": 477.40625,
    "epsilon_norm": 6.792723178863525,
    "prediction_norm": 5.94468355178833,
    "timestamp": "2025-07-14T12:33:38.626162"
  },
  {
    "epoch": 148,
    "batch_idx": 192,
    "loss": 0.8013147115707397,
    "learning_rate": 0.0009865492554910632,
    "timestep_mean": 385.875,
    "epsilon_norm": 4.741724014282227,
    "prediction_norm": 3.245074987411499,
    "timestamp": "2025-07-14T12:33:38.631306"
  },
  {
    "epoch": 149,
    "batch_idx": 0,
    "loss": 0.2815103530883789,
    "learning_rate": 0.000986367707876936,
    "timestep_mean": 493.15625,
    "epsilon_norm": 8.008723258972168,
    "prediction_norm": 7.009469985961914,
    "timestamp": "2025-07-14T12:33:38.636427"
  },
  {
    "epoch": 149,
    "batch_idx": 32,
    "loss": 0.480879545211792,
    "learning_rate": 0.000986367707876936,
    "timestep_mean": 473.125,
    "epsilon_norm": 8.668213844299316,
    "prediction_norm": 6.987866401672363,
    "timestamp": "2025-07-14T12:33:38.641502"
  },
  {
    "epoch": 149,
    "batch_idx": 64,
    "loss": 0.3620558977127075,
    "learning_rate": 0.000986367707876936,
    "timestep_mean": 466.4375,
    "epsilon_norm": 8.438366889953613,
    "prediction_norm": 6.6241865158081055,
    "timestamp": "2025-07-14T12:33:38.646688"
  },
  {
    "epoch": 149,
    "batch_idx": 96,
    "loss": 0.27260246872901917,
    "learning_rate": 0.000986367707876936,
    "timestep_mean": 569.65625,
    "epsilon_norm": 8.72255802154541,
    "prediction_norm": 7.781583309173584,
    "timestamp": "2025-07-14T12:33:38.651782"
  },
  {
    "epoch": 149,
    "batch_idx": 128,
    "loss": 0.4961681365966797,
    "learning_rate": 0.000986367707876936,
    "timestep_mean": 423.03125,
    "epsilon_norm": 8.265026092529297,
    "prediction_norm": 5.985037326812744,
    "timestamp": "2025-07-14T12:33:38.656774"
  },
  {
    "epoch": 149,
    "batch_idx": 160,
    "loss": 0.438347190618515,
    "learning_rate": 0.000986367707876936,
    "timestep_mean": 477.1875,
    "epsilon_norm": 7.312155723571777,
    "prediction_norm": 5.353862762451172,
    "timestamp": "2025-07-14T12:33:38.661934"
  },
  {
    "epoch": 149,
    "batch_idx": 192,
    "loss": 0.05459504574537277,
    "learning_rate": 0.000986367707876936,
    "timestep_mean": 720.875,
    "epsilon_norm": 3.9920523166656494,
    "prediction_norm": 3.9819841384887695,
    "timestamp": "2025-07-14T12:33:38.667024"
  },
  {
    "epoch": 150,
    "batch_idx": 0,
    "loss": 0.43939533829689026,
    "learning_rate": 0.0009861849601988382,
    "timestep_mean": 590.21875,
    "epsilon_norm": 8.089761734008789,
    "prediction_norm": 6.010458469390869,
    "timestamp": "2025-07-14T12:33:38.672136"
  },
  {
    "epoch": 150,
    "batch_idx": 32,
    "loss": 0.2804264724254608,
    "learning_rate": 0.0009861849601988382,
    "timestep_mean": 483.40625,
    "epsilon_norm": 6.76630973815918,
    "prediction_norm": 5.98147439956665,
    "timestamp": "2025-07-14T12:33:38.677226"
  },
  {
    "epoch": 150,
    "batch_idx": 64,
    "loss": 0.3078145980834961,
    "learning_rate": 0.0009861849601988382,
    "timestep_mean": 552.5625,
    "epsilon_norm": 9.676898002624512,
    "prediction_norm": 9.10290241241455,
    "timestamp": "2025-07-14T12:33:38.682245"
  },
  {
    "epoch": 150,
    "batch_idx": 96,
    "loss": 0.3728404641151428,
    "learning_rate": 0.0009861849601988382,
    "timestep_mean": 547.78125,
    "epsilon_norm": 7.36281156539917,
    "prediction_norm": 6.149544715881348,
    "timestamp": "2025-07-14T12:33:38.687555"
  },
  {
    "epoch": 150,
    "batch_idx": 128,
    "loss": 0.35746410489082336,
    "learning_rate": 0.0009861849601988382,
    "timestep_mean": 544.625,
    "epsilon_norm": 8.42874526977539,
    "prediction_norm": 6.040744304656982,
    "timestamp": "2025-07-14T12:33:38.692680"
  },
  {
    "epoch": 150,
    "batch_idx": 160,
    "loss": 0.4720931649208069,
    "learning_rate": 0.0009861849601988382,
    "timestep_mean": 461.5,
    "epsilon_norm": 8.235126495361328,
    "prediction_norm": 6.254645347595215,
    "timestamp": "2025-07-14T12:33:38.697637"
  },
  {
    "epoch": 150,
    "batch_idx": 192,
    "loss": 0.45632535219192505,
    "learning_rate": 0.0009861849601988382,
    "timestep_mean": 547.125,
    "epsilon_norm": 5.143272876739502,
    "prediction_norm": 4.893424987792969,
    "timestamp": "2025-07-14T12:33:38.702552"
  },
  {
    "epoch": 151,
    "batch_idx": 0,
    "loss": 0.6575772762298584,
    "learning_rate": 0.000986001012907681,
    "timestep_mean": 430.34375,
    "epsilon_norm": 8.558269500732422,
    "prediction_norm": 6.036045551300049,
    "timestamp": "2025-07-14T12:33:38.713412"
  },
  {
    "epoch": 151,
    "batch_idx": 32,
    "loss": 0.27090948820114136,
    "learning_rate": 0.000986001012907681,
    "timestep_mean": 458.40625,
    "epsilon_norm": 7.505290985107422,
    "prediction_norm": 7.327585697174072,
    "timestamp": "2025-07-14T12:33:38.718311"
  },
  {
    "epoch": 151,
    "batch_idx": 64,
    "loss": 0.3975353240966797,
    "learning_rate": 0.000986001012907681,
    "timestep_mean": 493.96875,
    "epsilon_norm": 7.9152374267578125,
    "prediction_norm": 5.3767547607421875,
    "timestamp": "2025-07-14T12:33:38.723144"
  },
  {
    "epoch": 151,
    "batch_idx": 96,
    "loss": 0.6041408777236938,
    "learning_rate": 0.000986001012907681,
    "timestep_mean": 429.875,
    "epsilon_norm": 8.746626853942871,
    "prediction_norm": 5.313225746154785,
    "timestamp": "2025-07-14T12:33:38.728102"
  },
  {
    "epoch": 151,
    "batch_idx": 128,
    "loss": 0.29978832602500916,
    "learning_rate": 0.000986001012907681,
    "timestep_mean": 496.09375,
    "epsilon_norm": 7.546264171600342,
    "prediction_norm": 5.157865524291992,
    "timestamp": "2025-07-14T12:33:38.733010"
  },
  {
    "epoch": 151,
    "batch_idx": 160,
    "loss": 0.2865351736545563,
    "learning_rate": 0.000986001012907681,
    "timestep_mean": 479.125,
    "epsilon_norm": 8.533666610717773,
    "prediction_norm": 7.035165309906006,
    "timestamp": "2025-07-14T12:33:38.738032"
  },
  {
    "epoch": 151,
    "batch_idx": 192,
    "loss": 0.5771982669830322,
    "learning_rate": 0.000986001012907681,
    "timestep_mean": 467.0,
    "epsilon_norm": 3.9985997676849365,
    "prediction_norm": 2.1253952980041504,
    "timestamp": "2025-07-14T12:33:38.742948"
  },
  {
    "epoch": 152,
    "batch_idx": 0,
    "loss": 0.3806661069393158,
    "learning_rate": 0.0009858158664573368,
    "timestep_mean": 560.53125,
    "epsilon_norm": 7.348079681396484,
    "prediction_norm": 5.110754013061523,
    "timestamp": "2025-07-14T12:33:38.747953"
  },
  {
    "epoch": 152,
    "batch_idx": 32,
    "loss": 0.4491039216518402,
    "learning_rate": 0.0009858158664573368,
    "timestep_mean": 565.125,
    "epsilon_norm": 8.045354843139648,
    "prediction_norm": 5.547123908996582,
    "timestamp": "2025-07-14T12:33:38.752811"
  },
  {
    "epoch": 152,
    "batch_idx": 64,
    "loss": 0.3747696280479431,
    "learning_rate": 0.0009858158664573368,
    "timestep_mean": 548.28125,
    "epsilon_norm": 8.666568756103516,
    "prediction_norm": 6.770116806030273,
    "timestamp": "2025-07-14T12:33:38.757716"
  },
  {
    "epoch": 152,
    "batch_idx": 96,
    "loss": 0.41006022691726685,
    "learning_rate": 0.0009858158664573368,
    "timestep_mean": 510.4375,
    "epsilon_norm": 7.3502984046936035,
    "prediction_norm": 5.321653842926025,
    "timestamp": "2025-07-14T12:33:38.762664"
  },
  {
    "epoch": 152,
    "batch_idx": 128,
    "loss": 0.2953334152698517,
    "learning_rate": 0.0009858158664573368,
    "timestep_mean": 516.71875,
    "epsilon_norm": 8.437026977539062,
    "prediction_norm": 7.652603626251221,
    "timestamp": "2025-07-14T12:33:38.767643"
  },
  {
    "epoch": 152,
    "batch_idx": 160,
    "loss": 0.3021753430366516,
    "learning_rate": 0.0009858158664573368,
    "timestep_mean": 505.46875,
    "epsilon_norm": 8.053959846496582,
    "prediction_norm": 6.971004009246826,
    "timestamp": "2025-07-14T12:33:38.772538"
  },
  {
    "epoch": 152,
    "batch_idx": 192,
    "loss": 0.13545697927474976,
    "learning_rate": 0.0009858158664573368,
    "timestep_mean": 648.125,
    "epsilon_norm": 4.437756538391113,
    "prediction_norm": 3.7970569133758545,
    "timestamp": "2025-07-14T12:33:38.777401"
  },
  {
    "epoch": 153,
    "batch_idx": 0,
    "loss": 0.44228631258010864,
    "learning_rate": 0.0009856295213046355,
    "timestep_mean": 433.8125,
    "epsilon_norm": 8.1118803024292,
    "prediction_norm": 6.352757930755615,
    "timestamp": "2025-07-14T12:33:38.782339"
  },
  {
    "epoch": 153,
    "batch_idx": 32,
    "loss": 0.23414874076843262,
    "learning_rate": 0.0009856295213046355,
    "timestep_mean": 534.8125,
    "epsilon_norm": 8.206802368164062,
    "prediction_norm": 7.517362117767334,
    "timestamp": "2025-07-14T12:33:38.787393"
  },
  {
    "epoch": 153,
    "batch_idx": 64,
    "loss": 0.730804443359375,
    "learning_rate": 0.0009856295213046355,
    "timestep_mean": 449.78125,
    "epsilon_norm": 8.97269344329834,
    "prediction_norm": 4.644003391265869,
    "timestamp": "2025-07-14T12:33:38.792412"
  },
  {
    "epoch": 153,
    "batch_idx": 96,
    "loss": 0.34396347403526306,
    "learning_rate": 0.0009856295213046355,
    "timestep_mean": 531.53125,
    "epsilon_norm": 7.081201076507568,
    "prediction_norm": 6.1448822021484375,
    "timestamp": "2025-07-14T12:33:38.797339"
  },
  {
    "epoch": 153,
    "batch_idx": 128,
    "loss": 0.5316182374954224,
    "learning_rate": 0.0009856295213046355,
    "timestep_mean": 441.625,
    "epsilon_norm": 8.701515197753906,
    "prediction_norm": 6.633801460266113,
    "timestamp": "2025-07-14T12:33:38.802205"
  },
  {
    "epoch": 153,
    "batch_idx": 160,
    "loss": 0.5158335566520691,
    "learning_rate": 0.0009856295213046355,
    "timestep_mean": 455.625,
    "epsilon_norm": 8.141378402709961,
    "prediction_norm": 6.277690887451172,
    "timestamp": "2025-07-14T12:33:38.807214"
  },
  {
    "epoch": 153,
    "batch_idx": 192,
    "loss": 0.20563504099845886,
    "learning_rate": 0.0009856295213046355,
    "timestep_mean": 533.375,
    "epsilon_norm": 3.311140298843384,
    "prediction_norm": 2.6792783737182617,
    "timestamp": "2025-07-14T12:33:38.812030"
  },
  {
    "epoch": 154,
    "batch_idx": 0,
    "loss": 0.2339998483657837,
    "learning_rate": 0.0009854419779093653,
    "timestep_mean": 601.28125,
    "epsilon_norm": 7.442619323730469,
    "prediction_norm": 7.281373023986816,
    "timestamp": "2025-07-14T12:33:38.817159"
  },
  {
    "epoch": 154,
    "batch_idx": 32,
    "loss": 0.48357364535331726,
    "learning_rate": 0.0009854419779093653,
    "timestep_mean": 386.0,
    "epsilon_norm": 8.185462951660156,
    "prediction_norm": 7.411280632019043,
    "timestamp": "2025-07-14T12:33:38.822051"
  },
  {
    "epoch": 154,
    "batch_idx": 64,
    "loss": 0.6784850358963013,
    "learning_rate": 0.0009854419779093653,
    "timestep_mean": 529.15625,
    "epsilon_norm": 8.690228462219238,
    "prediction_norm": 6.820396423339844,
    "timestamp": "2025-07-14T12:33:38.826985"
  },
  {
    "epoch": 154,
    "batch_idx": 96,
    "loss": 0.3771262764930725,
    "learning_rate": 0.0009854419779093653,
    "timestep_mean": 447.375,
    "epsilon_norm": 8.721537590026855,
    "prediction_norm": 8.042389869689941,
    "timestamp": "2025-07-14T12:33:38.832057"
  },
  {
    "epoch": 154,
    "batch_idx": 128,
    "loss": 0.25265029072761536,
    "learning_rate": 0.0009854419779093653,
    "timestep_mean": 528.25,
    "epsilon_norm": 7.452601432800293,
    "prediction_norm": 6.226083278656006,
    "timestamp": "2025-07-14T12:33:38.837071"
  },
  {
    "epoch": 154,
    "batch_idx": 160,
    "loss": 0.266196608543396,
    "learning_rate": 0.0009854419779093653,
    "timestep_mean": 483.65625,
    "epsilon_norm": 7.1939377784729,
    "prediction_norm": 6.482114315032959,
    "timestamp": "2025-07-14T12:33:38.842355"
  },
  {
    "epoch": 154,
    "batch_idx": 192,
    "loss": 0.681479811668396,
    "learning_rate": 0.0009854419779093653,
    "timestep_mean": 415.75,
    "epsilon_norm": 3.94431734085083,
    "prediction_norm": 1.801392674446106,
    "timestamp": "2025-07-14T12:33:38.847354"
  },
  {
    "epoch": 155,
    "batch_idx": 0,
    "loss": 0.16094812750816345,
    "learning_rate": 0.000985253236734271,
    "timestep_mean": 543.78125,
    "epsilon_norm": 7.175221920013428,
    "prediction_norm": 5.936110973358154,
    "timestamp": "2025-07-14T12:33:38.852332"
  },
  {
    "epoch": 155,
    "batch_idx": 32,
    "loss": 0.3266833424568176,
    "learning_rate": 0.000985253236734271,
    "timestep_mean": 543.625,
    "epsilon_norm": 7.887372970581055,
    "prediction_norm": 6.116148471832275,
    "timestamp": "2025-07-14T12:33:38.857556"
  },
  {
    "epoch": 155,
    "batch_idx": 64,
    "loss": 0.21573489904403687,
    "learning_rate": 0.000985253236734271,
    "timestep_mean": 504.59375,
    "epsilon_norm": 7.184277534484863,
    "prediction_norm": 5.189637660980225,
    "timestamp": "2025-07-14T12:33:38.862733"
  },
  {
    "epoch": 155,
    "batch_idx": 96,
    "loss": 0.5032693147659302,
    "learning_rate": 0.000985253236734271,
    "timestep_mean": 418.1875,
    "epsilon_norm": 7.647682189941406,
    "prediction_norm": 4.81705904006958,
    "timestamp": "2025-07-14T12:33:38.867819"
  },
  {
    "epoch": 155,
    "batch_idx": 128,
    "loss": 0.5000709295272827,
    "learning_rate": 0.000985253236734271,
    "timestep_mean": 573.46875,
    "epsilon_norm": 9.122058868408203,
    "prediction_norm": 5.698226451873779,
    "timestamp": "2025-07-14T12:33:38.872927"
  },
  {
    "epoch": 155,
    "batch_idx": 160,
    "loss": 0.3361998498439789,
    "learning_rate": 0.000985253236734271,
    "timestep_mean": 540.96875,
    "epsilon_norm": 7.799466133117676,
    "prediction_norm": 6.2444071769714355,
    "timestamp": "2025-07-14T12:33:38.877985"
  },
  {
    "epoch": 155,
    "batch_idx": 192,
    "loss": 0.2115495502948761,
    "learning_rate": 0.000985253236734271,
    "timestep_mean": 567.375,
    "epsilon_norm": 3.229111671447754,
    "prediction_norm": 2.753178834915161,
    "timestamp": "2025-07-14T12:33:38.883154"
  },
  {
    "epoch": 156,
    "batch_idx": 0,
    "loss": 0.45422855019569397,
    "learning_rate": 0.0009850632982450528,
    "timestep_mean": 501.46875,
    "epsilon_norm": 9.870389938354492,
    "prediction_norm": 8.059093475341797,
    "timestamp": "2025-07-14T12:33:38.888300"
  },
  {
    "epoch": 156,
    "batch_idx": 32,
    "loss": 0.37258046865463257,
    "learning_rate": 0.0009850632982450528,
    "timestep_mean": 460.03125,
    "epsilon_norm": 7.051971912384033,
    "prediction_norm": 4.8846049308776855,
    "timestamp": "2025-07-14T12:33:38.893404"
  },
  {
    "epoch": 156,
    "batch_idx": 64,
    "loss": 0.5684573650360107,
    "learning_rate": 0.0009850632982450528,
    "timestep_mean": 374.40625,
    "epsilon_norm": 8.276479721069336,
    "prediction_norm": 5.252096176147461,
    "timestamp": "2025-07-14T12:33:38.898595"
  },
  {
    "epoch": 156,
    "batch_idx": 96,
    "loss": 0.5074732303619385,
    "learning_rate": 0.0009850632982450528,
    "timestep_mean": 513.25,
    "epsilon_norm": 7.614117622375488,
    "prediction_norm": 5.091432571411133,
    "timestamp": "2025-07-14T12:33:38.903612"
  },
  {
    "epoch": 156,
    "batch_idx": 128,
    "loss": 0.37482795119285583,
    "learning_rate": 0.0009850632982450528,
    "timestep_mean": 500.28125,
    "epsilon_norm": 7.7243547439575195,
    "prediction_norm": 6.516963005065918,
    "timestamp": "2025-07-14T12:33:38.908823"
  },
  {
    "epoch": 156,
    "batch_idx": 160,
    "loss": 0.317229688167572,
    "learning_rate": 0.0009850632982450528,
    "timestep_mean": 507.1875,
    "epsilon_norm": 7.992811679840088,
    "prediction_norm": 6.458600044250488,
    "timestamp": "2025-07-14T12:33:38.913900"
  },
  {
    "epoch": 156,
    "batch_idx": 192,
    "loss": 0.138563334941864,
    "learning_rate": 0.0009850632982450528,
    "timestep_mean": 488.875,
    "epsilon_norm": 2.564673900604248,
    "prediction_norm": 2.3274147510528564,
    "timestamp": "2025-07-14T12:33:38.918865"
  },
  {
    "epoch": 157,
    "batch_idx": 0,
    "loss": 0.435524582862854,
    "learning_rate": 0.0009848721629103648,
    "timestep_mean": 454.375,
    "epsilon_norm": 7.669771671295166,
    "prediction_norm": 6.35849666595459,
    "timestamp": "2025-07-14T12:33:38.924036"
  },
  {
    "epoch": 157,
    "batch_idx": 32,
    "loss": 0.3123677372932434,
    "learning_rate": 0.0009848721629103648,
    "timestep_mean": 535.59375,
    "epsilon_norm": 8.45476245880127,
    "prediction_norm": 7.816832065582275,
    "timestamp": "2025-07-14T12:33:38.929114"
  },
  {
    "epoch": 157,
    "batch_idx": 64,
    "loss": 0.5261873006820679,
    "learning_rate": 0.0009848721629103648,
    "timestep_mean": 471.4375,
    "epsilon_norm": 9.476502418518066,
    "prediction_norm": 8.43504810333252,
    "timestamp": "2025-07-14T12:33:38.934223"
  },
  {
    "epoch": 157,
    "batch_idx": 96,
    "loss": 0.23424923419952393,
    "learning_rate": 0.0009848721629103648,
    "timestep_mean": 516.65625,
    "epsilon_norm": 7.138298034667969,
    "prediction_norm": 5.906457424163818,
    "timestamp": "2025-07-14T12:33:38.939276"
  },
  {
    "epoch": 157,
    "batch_idx": 128,
    "loss": 0.6169530153274536,
    "learning_rate": 0.0009848721629103648,
    "timestep_mean": 421.875,
    "epsilon_norm": 8.336381912231445,
    "prediction_norm": 5.511867523193359,
    "timestamp": "2025-07-14T12:33:38.944194"
  },
  {
    "epoch": 157,
    "batch_idx": 160,
    "loss": 0.48452314734458923,
    "learning_rate": 0.0009848721629103648,
    "timestep_mean": 399.75,
    "epsilon_norm": 7.043124198913574,
    "prediction_norm": 4.23573637008667,
    "timestamp": "2025-07-14T12:33:38.949343"
  },
  {
    "epoch": 157,
    "batch_idx": 192,
    "loss": 0.4286840260028839,
    "learning_rate": 0.0009848721629103648,
    "timestep_mean": 480.875,
    "epsilon_norm": 3.5161094665527344,
    "prediction_norm": 2.151641607284546,
    "timestamp": "2025-07-14T12:33:38.954410"
  },
  {
    "epoch": 158,
    "batch_idx": 0,
    "loss": 0.39549869298934937,
    "learning_rate": 0.0009846798312018144,
    "timestep_mean": 502.625,
    "epsilon_norm": 7.355246543884277,
    "prediction_norm": 4.737573146820068,
    "timestamp": "2025-07-14T12:33:38.959512"
  },
  {
    "epoch": 158,
    "batch_idx": 32,
    "loss": 0.3113819360733032,
    "learning_rate": 0.0009846798312018144,
    "timestep_mean": 511.6875,
    "epsilon_norm": 7.463840484619141,
    "prediction_norm": 5.393931865692139,
    "timestamp": "2025-07-14T12:33:38.964541"
  },
  {
    "epoch": 158,
    "batch_idx": 64,
    "loss": 0.3764622211456299,
    "learning_rate": 0.0009846798312018144,
    "timestep_mean": 519.6875,
    "epsilon_norm": 8.10257625579834,
    "prediction_norm": 6.100135803222656,
    "timestamp": "2025-07-14T12:33:38.969550"
  },
  {
    "epoch": 158,
    "batch_idx": 96,
    "loss": 0.2318858951330185,
    "learning_rate": 0.0009846798312018144,
    "timestep_mean": 501.9375,
    "epsilon_norm": 7.318490028381348,
    "prediction_norm": 5.681520938873291,
    "timestamp": "2025-07-14T12:33:38.974641"
  },
  {
    "epoch": 158,
    "batch_idx": 128,
    "loss": 0.6468706727027893,
    "learning_rate": 0.0009846798312018144,
    "timestep_mean": 509.59375,
    "epsilon_norm": 9.15330696105957,
    "prediction_norm": 5.956964015960693,
    "timestamp": "2025-07-14T12:33:38.979667"
  },
  {
    "epoch": 158,
    "batch_idx": 160,
    "loss": 0.5804962515830994,
    "learning_rate": 0.0009846798312018144,
    "timestep_mean": 508.375,
    "epsilon_norm": 9.648063659667969,
    "prediction_norm": 7.462937831878662,
    "timestamp": "2025-07-14T12:33:38.984787"
  },
  {
    "epoch": 158,
    "batch_idx": 192,
    "loss": 0.3410620093345642,
    "learning_rate": 0.0009846798312018144,
    "timestep_mean": 407.625,
    "epsilon_norm": 3.3729491233825684,
    "prediction_norm": 2.4444093704223633,
    "timestamp": "2025-07-14T12:33:38.989826"
  },
  {
    "epoch": 159,
    "batch_idx": 0,
    "loss": 0.38448625802993774,
    "learning_rate": 0.0009844863035939613,
    "timestep_mean": 600.21875,
    "epsilon_norm": 8.650656700134277,
    "prediction_norm": 7.443136215209961,
    "timestamp": "2025-07-14T12:33:38.994816"
  },
  {
    "epoch": 159,
    "batch_idx": 32,
    "loss": 0.4632452130317688,
    "learning_rate": 0.0009844863035939613,
    "timestep_mean": 425.0,
    "epsilon_norm": 8.443965911865234,
    "prediction_norm": 5.8740739822387695,
    "timestamp": "2025-07-14T12:33:38.999917"
  },
  {
    "epoch": 159,
    "batch_idx": 64,
    "loss": 0.34008586406707764,
    "learning_rate": 0.0009844863035939613,
    "timestep_mean": 511.0,
    "epsilon_norm": 6.493505001068115,
    "prediction_norm": 4.810292720794678,
    "timestamp": "2025-07-14T12:33:39.004956"
  },
  {
    "epoch": 159,
    "batch_idx": 96,
    "loss": 0.3265373706817627,
    "learning_rate": 0.0009844863035939613,
    "timestep_mean": 431.8125,
    "epsilon_norm": 7.047808647155762,
    "prediction_norm": 6.09136962890625,
    "timestamp": "2025-07-14T12:33:39.010009"
  },
  {
    "epoch": 159,
    "batch_idx": 128,
    "loss": 0.40771543979644775,
    "learning_rate": 0.0009844863035939613,
    "timestep_mean": 503.1875,
    "epsilon_norm": 7.731118202209473,
    "prediction_norm": 6.008951187133789,
    "timestamp": "2025-07-14T12:33:39.015112"
  },
  {
    "epoch": 159,
    "batch_idx": 160,
    "loss": 0.5688282251358032,
    "learning_rate": 0.0009844863035939613,
    "timestep_mean": 451.75,
    "epsilon_norm": 7.9770660400390625,
    "prediction_norm": 6.097033500671387,
    "timestamp": "2025-07-14T12:33:39.020056"
  },
  {
    "epoch": 159,
    "batch_idx": 192,
    "loss": 0.49511346220970154,
    "learning_rate": 0.0009844863035939613,
    "timestep_mean": 419.125,
    "epsilon_norm": 4.468086242675781,
    "prediction_norm": 3.7442009449005127,
    "timestamp": "2025-07-14T12:33:39.025265"
  },
  {
    "epoch": 160,
    "batch_idx": 0,
    "loss": 0.5754594802856445,
    "learning_rate": 0.0009842915805643154,
    "timestep_mean": 480.3125,
    "epsilon_norm": 8.440608024597168,
    "prediction_norm": 6.314897060394287,
    "timestamp": "2025-07-14T12:33:39.030275"
  },
  {
    "epoch": 160,
    "batch_idx": 32,
    "loss": 0.4268763065338135,
    "learning_rate": 0.0009842915805643154,
    "timestep_mean": 499.15625,
    "epsilon_norm": 8.641047477722168,
    "prediction_norm": 7.881438255310059,
    "timestamp": "2025-07-14T12:33:39.035305"
  },
  {
    "epoch": 160,
    "batch_idx": 64,
    "loss": 0.3683531582355499,
    "learning_rate": 0.0009842915805643154,
    "timestep_mean": 508.90625,
    "epsilon_norm": 7.55413293838501,
    "prediction_norm": 6.636831760406494,
    "timestamp": "2025-07-14T12:33:39.040275"
  },
  {
    "epoch": 160,
    "batch_idx": 96,
    "loss": 0.22442612051963806,
    "learning_rate": 0.0009842915805643154,
    "timestep_mean": 534.96875,
    "epsilon_norm": 7.704458236694336,
    "prediction_norm": 7.014163970947266,
    "timestamp": "2025-07-14T12:33:39.045101"
  },
  {
    "epoch": 160,
    "batch_idx": 128,
    "loss": 0.3561028242111206,
    "learning_rate": 0.0009842915805643154,
    "timestep_mean": 527.5625,
    "epsilon_norm": 8.602700233459473,
    "prediction_norm": 7.159002304077148,
    "timestamp": "2025-07-14T12:33:39.049994"
  },
  {
    "epoch": 160,
    "batch_idx": 160,
    "loss": 0.2262844741344452,
    "learning_rate": 0.0009842915805643154,
    "timestep_mean": 563.78125,
    "epsilon_norm": 7.430230617523193,
    "prediction_norm": 6.492309093475342,
    "timestamp": "2025-07-14T12:33:39.054895"
  },
  {
    "epoch": 160,
    "batch_idx": 192,
    "loss": 0.7754100561141968,
    "learning_rate": 0.0009842915805643154,
    "timestep_mean": 503.0,
    "epsilon_norm": 5.31984806060791,
    "prediction_norm": 2.9913556575775146,
    "timestamp": "2025-07-14T12:33:39.059699"
  },
  {
    "epoch": 161,
    "batch_idx": 0,
    "loss": 0.4922644793987274,
    "learning_rate": 0.0009840956625933363,
    "timestep_mean": 385.65625,
    "epsilon_norm": 7.782602310180664,
    "prediction_norm": 5.1819682121276855,
    "timestamp": "2025-07-14T12:33:39.064565"
  },
  {
    "epoch": 161,
    "batch_idx": 32,
    "loss": 0.3530464768409729,
    "learning_rate": 0.0009840956625933363,
    "timestep_mean": 541.28125,
    "epsilon_norm": 8.820243835449219,
    "prediction_norm": 6.864660263061523,
    "timestamp": "2025-07-14T12:33:39.069423"
  },
  {
    "epoch": 161,
    "batch_idx": 64,
    "loss": 0.5728086233139038,
    "learning_rate": 0.0009840956625933363,
    "timestep_mean": 371.375,
    "epsilon_norm": 7.209831237792969,
    "prediction_norm": 3.2706074714660645,
    "timestamp": "2025-07-14T12:33:39.074253"
  },
  {
    "epoch": 161,
    "batch_idx": 96,
    "loss": 0.44053882360458374,
    "learning_rate": 0.0009840956625933363,
    "timestep_mean": 458.5,
    "epsilon_norm": 8.032203674316406,
    "prediction_norm": 5.513313293457031,
    "timestamp": "2025-07-14T12:33:39.079202"
  },
  {
    "epoch": 161,
    "batch_idx": 128,
    "loss": 0.4136078655719757,
    "learning_rate": 0.0009840956625933363,
    "timestep_mean": 500.28125,
    "epsilon_norm": 8.657903671264648,
    "prediction_norm": 6.725905418395996,
    "timestamp": "2025-07-14T12:33:39.084070"
  },
  {
    "epoch": 161,
    "batch_idx": 160,
    "loss": 0.295743852853775,
    "learning_rate": 0.0009840956625933363,
    "timestep_mean": 515.0,
    "epsilon_norm": 8.172235488891602,
    "prediction_norm": 6.368993759155273,
    "timestamp": "2025-07-14T12:33:39.088916"
  },
  {
    "epoch": 161,
    "batch_idx": 192,
    "loss": 0.08678677678108215,
    "learning_rate": 0.0009840956625933363,
    "timestep_mean": 645.375,
    "epsilon_norm": 3.6410608291625977,
    "prediction_norm": 3.376866102218628,
    "timestamp": "2025-07-14T12:33:39.093740"
  },
  {
    "epoch": 162,
    "batch_idx": 0,
    "loss": 0.30095216631889343,
    "learning_rate": 0.0009838985501644324,
    "timestep_mean": 519.34375,
    "epsilon_norm": 8.740683555603027,
    "prediction_norm": 7.570684432983398,
    "timestamp": "2025-07-14T12:33:39.098597"
  },
  {
    "epoch": 162,
    "batch_idx": 32,
    "loss": 0.4033185839653015,
    "learning_rate": 0.0009838985501644324,
    "timestep_mean": 444.78125,
    "epsilon_norm": 8.335198402404785,
    "prediction_norm": 6.610716819763184,
    "timestamp": "2025-07-14T12:33:39.103455"
  },
  {
    "epoch": 162,
    "batch_idx": 64,
    "loss": 0.27642595767974854,
    "learning_rate": 0.0009838985501644324,
    "timestep_mean": 520.40625,
    "epsilon_norm": 6.993413925170898,
    "prediction_norm": 6.943432807922363,
    "timestamp": "2025-07-14T12:33:39.108329"
  },
  {
    "epoch": 162,
    "batch_idx": 96,
    "loss": 0.4345163404941559,
    "learning_rate": 0.0009838985501644324,
    "timestep_mean": 449.03125,
    "epsilon_norm": 8.516389846801758,
    "prediction_norm": 6.764902591705322,
    "timestamp": "2025-07-14T12:33:39.113156"
  },
  {
    "epoch": 162,
    "batch_idx": 128,
    "loss": 0.310546875,
    "learning_rate": 0.0009838985501644324,
    "timestep_mean": 498.71875,
    "epsilon_norm": 7.7142109870910645,
    "prediction_norm": 6.070683002471924,
    "timestamp": "2025-07-14T12:33:39.117945"
  },
  {
    "epoch": 162,
    "batch_idx": 160,
    "loss": 0.26142364740371704,
    "learning_rate": 0.0009838985501644324,
    "timestep_mean": 487.4375,
    "epsilon_norm": 7.298758029937744,
    "prediction_norm": 6.387974739074707,
    "timestamp": "2025-07-14T12:33:39.122789"
  },
  {
    "epoch": 162,
    "batch_idx": 192,
    "loss": 0.3015170693397522,
    "learning_rate": 0.0009838985501644324,
    "timestep_mean": 445.5,
    "epsilon_norm": 3.4515998363494873,
    "prediction_norm": 3.24074125289917,
    "timestamp": "2025-07-14T12:33:39.127642"
  },
  {
    "epoch": 163,
    "batch_idx": 0,
    "loss": 0.2502132058143616,
    "learning_rate": 0.000983700243763959,
    "timestep_mean": 524.5,
    "epsilon_norm": 8.362244606018066,
    "prediction_norm": 7.653672695159912,
    "timestamp": "2025-07-14T12:33:39.132481"
  },
  {
    "epoch": 163,
    "batch_idx": 32,
    "loss": 0.5373959541320801,
    "learning_rate": 0.000983700243763959,
    "timestep_mean": 433.5625,
    "epsilon_norm": 8.776032447814941,
    "prediction_norm": 5.983464241027832,
    "timestamp": "2025-07-14T12:33:39.137445"
  },
  {
    "epoch": 163,
    "batch_idx": 64,
    "loss": 0.2473355382680893,
    "learning_rate": 0.000983700243763959,
    "timestep_mean": 490.8125,
    "epsilon_norm": 7.824832916259766,
    "prediction_norm": 5.855844020843506,
    "timestamp": "2025-07-14T12:33:39.142294"
  },
  {
    "epoch": 163,
    "batch_idx": 96,
    "loss": 0.2558196187019348,
    "learning_rate": 0.000983700243763959,
    "timestep_mean": 475.15625,
    "epsilon_norm": 6.941914081573486,
    "prediction_norm": 5.352507591247559,
    "timestamp": "2025-07-14T12:33:39.147108"
  },
  {
    "epoch": 163,
    "batch_idx": 128,
    "loss": 0.2631734609603882,
    "learning_rate": 0.000983700243763959,
    "timestep_mean": 554.21875,
    "epsilon_norm": 7.7986674308776855,
    "prediction_norm": 6.025549411773682,
    "timestamp": "2025-07-14T12:33:39.151950"
  },
  {
    "epoch": 163,
    "batch_idx": 160,
    "loss": 0.3975895941257477,
    "learning_rate": 0.000983700243763959,
    "timestep_mean": 536.9375,
    "epsilon_norm": 8.487120628356934,
    "prediction_norm": 6.399445056915283,
    "timestamp": "2025-07-14T12:33:39.156766"
  },
  {
    "epoch": 163,
    "batch_idx": 192,
    "loss": 0.1229398250579834,
    "learning_rate": 0.000983700243763959,
    "timestep_mean": 769.0,
    "epsilon_norm": 4.93261194229126,
    "prediction_norm": 5.23917293548584,
    "timestamp": "2025-07-14T12:33:39.161573"
  },
  {
    "epoch": 164,
    "batch_idx": 0,
    "loss": 0.24986425042152405,
    "learning_rate": 0.0009835007438812173,
    "timestep_mean": 555.34375,
    "epsilon_norm": 8.66428279876709,
    "prediction_norm": 7.700439453125,
    "timestamp": "2025-07-14T12:33:39.166408"
  },
  {
    "epoch": 164,
    "batch_idx": 32,
    "loss": 0.35192185640335083,
    "learning_rate": 0.0009835007438812173,
    "timestep_mean": 477.625,
    "epsilon_norm": 8.181527137756348,
    "prediction_norm": 5.966598987579346,
    "timestamp": "2025-07-14T12:33:39.171245"
  },
  {
    "epoch": 164,
    "batch_idx": 64,
    "loss": 0.3284994661808014,
    "learning_rate": 0.0009835007438812173,
    "timestep_mean": 432.59375,
    "epsilon_norm": 7.232795715332031,
    "prediction_norm": 5.8018412590026855,
    "timestamp": "2025-07-14T12:33:39.176092"
  },
  {
    "epoch": 164,
    "batch_idx": 96,
    "loss": 0.19761985540390015,
    "learning_rate": 0.0009835007438812173,
    "timestep_mean": 511.78125,
    "epsilon_norm": 8.181103706359863,
    "prediction_norm": 7.547508716583252,
    "timestamp": "2025-07-14T12:33:39.180919"
  },
  {
    "epoch": 164,
    "batch_idx": 128,
    "loss": 0.4693451523780823,
    "learning_rate": 0.0009835007438812173,
    "timestep_mean": 494.40625,
    "epsilon_norm": 8.597024917602539,
    "prediction_norm": 6.627535820007324,
    "timestamp": "2025-07-14T12:33:39.185771"
  },
  {
    "epoch": 164,
    "batch_idx": 160,
    "loss": 0.5429067015647888,
    "learning_rate": 0.0009835007438812173,
    "timestep_mean": 418.0625,
    "epsilon_norm": 6.790272235870361,
    "prediction_norm": 4.280710220336914,
    "timestamp": "2025-07-14T12:33:39.190584"
  },
  {
    "epoch": 164,
    "batch_idx": 192,
    "loss": 0.19160689413547516,
    "learning_rate": 0.0009835007438812173,
    "timestep_mean": 379.375,
    "epsilon_norm": 3.5907747745513916,
    "prediction_norm": 3.467503547668457,
    "timestamp": "2025-07-14T12:33:39.195404"
  },
  {
    "epoch": 165,
    "batch_idx": 0,
    "loss": 0.20366144180297852,
    "learning_rate": 0.0009833000510084534,
    "timestep_mean": 659.9375,
    "epsilon_norm": 7.865975379943848,
    "prediction_norm": 7.876739978790283,
    "timestamp": "2025-07-14T12:33:39.200333"
  },
  {
    "epoch": 165,
    "batch_idx": 32,
    "loss": 0.41632479429244995,
    "learning_rate": 0.0009833000510084534,
    "timestep_mean": 532.84375,
    "epsilon_norm": 8.276433944702148,
    "prediction_norm": 6.847295761108398,
    "timestamp": "2025-07-14T12:33:39.205151"
  },
  {
    "epoch": 165,
    "batch_idx": 64,
    "loss": 0.5401289463043213,
    "learning_rate": 0.0009833000510084534,
    "timestep_mean": 472.375,
    "epsilon_norm": 8.331792831420898,
    "prediction_norm": 6.240783214569092,
    "timestamp": "2025-07-14T12:33:39.210045"
  },
  {
    "epoch": 165,
    "batch_idx": 96,
    "loss": 0.2593687176704407,
    "learning_rate": 0.0009833000510084534,
    "timestep_mean": 530.0,
    "epsilon_norm": 7.519252777099609,
    "prediction_norm": 6.593700885772705,
    "timestamp": "2025-07-14T12:33:39.214868"
  },
  {
    "epoch": 165,
    "batch_idx": 128,
    "loss": 0.35643458366394043,
    "learning_rate": 0.0009833000510084534,
    "timestep_mean": 539.90625,
    "epsilon_norm": 8.171401023864746,
    "prediction_norm": 7.581094741821289,
    "timestamp": "2025-07-14T12:33:39.219669"
  },
  {
    "epoch": 165,
    "batch_idx": 160,
    "loss": 0.3491877317428589,
    "learning_rate": 0.0009833000510084534,
    "timestep_mean": 455.75,
    "epsilon_norm": 6.659500598907471,
    "prediction_norm": 4.6916351318359375,
    "timestamp": "2025-07-14T12:33:39.224473"
  },
  {
    "epoch": 165,
    "batch_idx": 192,
    "loss": 0.39188796281814575,
    "learning_rate": 0.0009833000510084534,
    "timestep_mean": 593.375,
    "epsilon_norm": 3.8788959980010986,
    "prediction_norm": 2.218512773513794,
    "timestamp": "2025-07-14T12:33:39.229404"
  },
  {
    "epoch": 166,
    "batch_idx": 0,
    "loss": 0.27883976697921753,
    "learning_rate": 0.000983098165640857,
    "timestep_mean": 578.03125,
    "epsilon_norm": 8.518433570861816,
    "prediction_norm": 6.895758628845215,
    "timestamp": "2025-07-14T12:33:39.234257"
  },
  {
    "epoch": 166,
    "batch_idx": 32,
    "loss": 0.4168716073036194,
    "learning_rate": 0.000983098165640857,
    "timestep_mean": 467.78125,
    "epsilon_norm": 7.873891830444336,
    "prediction_norm": 5.653619766235352,
    "timestamp": "2025-07-14T12:33:39.239115"
  },
  {
    "epoch": 166,
    "batch_idx": 64,
    "loss": 0.36178457736968994,
    "learning_rate": 0.000983098165640857,
    "timestep_mean": 555.0625,
    "epsilon_norm": 9.31773567199707,
    "prediction_norm": 7.684148788452148,
    "timestamp": "2025-07-14T12:33:39.243972"
  },
  {
    "epoch": 166,
    "batch_idx": 96,
    "loss": 0.3620247542858124,
    "learning_rate": 0.000983098165640857,
    "timestep_mean": 519.28125,
    "epsilon_norm": 7.553893089294434,
    "prediction_norm": 5.751008987426758,
    "timestamp": "2025-07-14T12:33:39.248801"
  },
  {
    "epoch": 166,
    "batch_idx": 128,
    "loss": 0.5868680477142334,
    "learning_rate": 0.000983098165640857,
    "timestep_mean": 391.8125,
    "epsilon_norm": 7.887352466583252,
    "prediction_norm": 4.887171268463135,
    "timestamp": "2025-07-14T12:33:39.253631"
  },
  {
    "epoch": 166,
    "batch_idx": 160,
    "loss": 0.4148207902908325,
    "learning_rate": 0.000983098165640857,
    "timestep_mean": 496.59375,
    "epsilon_norm": 7.715336799621582,
    "prediction_norm": 5.2839460372924805,
    "timestamp": "2025-07-14T12:33:39.258485"
  },
  {
    "epoch": 166,
    "batch_idx": 192,
    "loss": 0.1422867625951767,
    "learning_rate": 0.000983098165640857,
    "timestep_mean": 662.875,
    "epsilon_norm": 4.7354559898376465,
    "prediction_norm": 4.116851329803467,
    "timestamp": "2025-07-14T12:33:39.263281"
  },
  {
    "epoch": 167,
    "batch_idx": 0,
    "loss": 0.6094413995742798,
    "learning_rate": 0.0009828950882765607,
    "timestep_mean": 466.46875,
    "epsilon_norm": 9.859436988830566,
    "prediction_norm": 7.303080081939697,
    "timestamp": "2025-07-14T12:33:39.268115"
  },
  {
    "epoch": 167,
    "batch_idx": 32,
    "loss": 0.16128835082054138,
    "learning_rate": 0.0009828950882765607,
    "timestep_mean": 603.5,
    "epsilon_norm": 6.265192031860352,
    "prediction_norm": 5.467442989349365,
    "timestamp": "2025-07-14T12:33:39.273036"
  },
  {
    "epoch": 167,
    "batch_idx": 64,
    "loss": 0.3449282646179199,
    "learning_rate": 0.0009828950882765607,
    "timestep_mean": 440.21875,
    "epsilon_norm": 8.656702041625977,
    "prediction_norm": 7.361875057220459,
    "timestamp": "2025-07-14T12:33:39.278201"
  },
  {
    "epoch": 167,
    "batch_idx": 96,
    "loss": 0.3050372302532196,
    "learning_rate": 0.0009828950882765607,
    "timestep_mean": 511.03125,
    "epsilon_norm": 8.386780738830566,
    "prediction_norm": 7.201180458068848,
    "timestamp": "2025-07-14T12:33:39.283270"
  },
  {
    "epoch": 167,
    "batch_idx": 128,
    "loss": 0.4007357954978943,
    "learning_rate": 0.0009828950882765607,
    "timestep_mean": 520.4375,
    "epsilon_norm": 8.517478942871094,
    "prediction_norm": 6.501434803009033,
    "timestamp": "2025-07-14T12:33:39.288432"
  },
  {
    "epoch": 167,
    "batch_idx": 160,
    "loss": 0.30363893508911133,
    "learning_rate": 0.0009828950882765607,
    "timestep_mean": 588.5625,
    "epsilon_norm": 8.052826881408691,
    "prediction_norm": 7.645110130310059,
    "timestamp": "2025-07-14T12:33:39.293559"
  },
  {
    "epoch": 167,
    "batch_idx": 192,
    "loss": 0.11232317239046097,
    "learning_rate": 0.0009828950882765607,
    "timestep_mean": 630.5,
    "epsilon_norm": 2.8522379398345947,
    "prediction_norm": 2.7096545696258545,
    "timestamp": "2025-07-14T12:33:39.298477"
  },
  {
    "epoch": 168,
    "batch_idx": 0,
    "loss": 0.2792856693267822,
    "learning_rate": 0.0009826908194166368,
    "timestep_mean": 561.71875,
    "epsilon_norm": 7.246521949768066,
    "prediction_norm": 5.828732967376709,
    "timestamp": "2025-07-14T12:33:39.306543"
  },
  {
    "epoch": 168,
    "batch_idx": 32,
    "loss": 0.5749422907829285,
    "learning_rate": 0.0009826908194166368,
    "timestep_mean": 486.46875,
    "epsilon_norm": 8.045799255371094,
    "prediction_norm": 6.20237398147583,
    "timestamp": "2025-07-14T12:33:39.311666"
  },
  {
    "epoch": 168,
    "batch_idx": 64,
    "loss": 0.3446577787399292,
    "learning_rate": 0.0009826908194166368,
    "timestep_mean": 472.75,
    "epsilon_norm": 7.5742716789245605,
    "prediction_norm": 6.620166301727295,
    "timestamp": "2025-07-14T12:33:39.316752"
  },
  {
    "epoch": 168,
    "batch_idx": 96,
    "loss": 0.5018186569213867,
    "learning_rate": 0.0009826908194166368,
    "timestep_mean": 420.0625,
    "epsilon_norm": 7.7875142097473145,
    "prediction_norm": 5.418042182922363,
    "timestamp": "2025-07-14T12:33:39.321651"
  },
  {
    "epoch": 168,
    "batch_idx": 128,
    "loss": 0.3453655242919922,
    "learning_rate": 0.0009826908194166368,
    "timestep_mean": 474.21875,
    "epsilon_norm": 7.400015830993652,
    "prediction_norm": 5.472236156463623,
    "timestamp": "2025-07-14T12:33:39.326581"
  },
  {
    "epoch": 168,
    "batch_idx": 160,
    "loss": 0.4251039922237396,
    "learning_rate": 0.0009826908194166368,
    "timestep_mean": 446.78125,
    "epsilon_norm": 8.499835014343262,
    "prediction_norm": 6.960320949554443,
    "timestamp": "2025-07-14T12:33:39.331459"
  },
  {
    "epoch": 168,
    "batch_idx": 192,
    "loss": 0.45863163471221924,
    "learning_rate": 0.0009826908194166368,
    "timestep_mean": 457.75,
    "epsilon_norm": 4.8719072341918945,
    "prediction_norm": 4.237868309020996,
    "timestamp": "2025-07-14T12:33:39.336543"
  },
  {
    "epoch": 169,
    "batch_idx": 0,
    "loss": 0.37883901596069336,
    "learning_rate": 0.000982485359565099,
    "timestep_mean": 480.125,
    "epsilon_norm": 7.244182586669922,
    "prediction_norm": 4.738745212554932,
    "timestamp": "2025-07-14T12:33:39.341407"
  },
  {
    "epoch": 169,
    "batch_idx": 32,
    "loss": 0.3454233407974243,
    "learning_rate": 0.000982485359565099,
    "timestep_mean": 448.75,
    "epsilon_norm": 7.221580505371094,
    "prediction_norm": 5.45013427734375,
    "timestamp": "2025-07-14T12:33:39.346334"
  },
  {
    "epoch": 169,
    "batch_idx": 64,
    "loss": 0.23372487723827362,
    "learning_rate": 0.000982485359565099,
    "timestep_mean": 501.3125,
    "epsilon_norm": 7.79603910446167,
    "prediction_norm": 6.769505023956299,
    "timestamp": "2025-07-14T12:33:39.351156"
  },
  {
    "epoch": 169,
    "batch_idx": 96,
    "loss": 0.26121988892555237,
    "learning_rate": 0.000982485359565099,
    "timestep_mean": 562.90625,
    "epsilon_norm": 7.817646503448486,
    "prediction_norm": 5.765377044677734,
    "timestamp": "2025-07-14T12:33:39.356051"
  },
  {
    "epoch": 169,
    "batch_idx": 128,
    "loss": 0.2764696180820465,
    "learning_rate": 0.000982485359565099,
    "timestep_mean": 507.9375,
    "epsilon_norm": 6.064265727996826,
    "prediction_norm": 3.9011871814727783,
    "timestamp": "2025-07-14T12:33:39.361103"
  },
  {
    "epoch": 169,
    "batch_idx": 160,
    "loss": 0.3708915114402771,
    "learning_rate": 0.000982485359565099,
    "timestep_mean": 580.3125,
    "epsilon_norm": 9.74899673461914,
    "prediction_norm": 7.889326572418213,
    "timestamp": "2025-07-14T12:33:39.366043"
  },
  {
    "epoch": 169,
    "batch_idx": 192,
    "loss": 0.6466395854949951,
    "learning_rate": 0.000982485359565099,
    "timestep_mean": 325.75,
    "epsilon_norm": 3.714088201522827,
    "prediction_norm": 1.713236689567566,
    "timestamp": "2025-07-14T12:33:39.371121"
  },
  {
    "epoch": 170,
    "batch_idx": 0,
    "loss": 0.3884698748588562,
    "learning_rate": 0.000982278709228899,
    "timestep_mean": 497.25,
    "epsilon_norm": 7.39548397064209,
    "prediction_norm": 5.694192409515381,
    "timestamp": "2025-07-14T12:33:39.376024"
  },
  {
    "epoch": 170,
    "batch_idx": 32,
    "loss": 0.5081044435501099,
    "learning_rate": 0.000982278709228899,
    "timestep_mean": 451.78125,
    "epsilon_norm": 8.489386558532715,
    "prediction_norm": 6.067110061645508,
    "timestamp": "2025-07-14T12:33:39.381029"
  },
  {
    "epoch": 170,
    "batch_idx": 64,
    "loss": 0.3505328893661499,
    "learning_rate": 0.000982278709228899,
    "timestep_mean": 508.4375,
    "epsilon_norm": 8.592788696289062,
    "prediction_norm": 7.324742794036865,
    "timestamp": "2025-07-14T12:33:39.386029"
  },
  {
    "epoch": 170,
    "batch_idx": 96,
    "loss": 0.43891793489456177,
    "learning_rate": 0.000982278709228899,
    "timestep_mean": 514.4375,
    "epsilon_norm": 7.381440162658691,
    "prediction_norm": 5.565734386444092,
    "timestamp": "2025-07-14T12:33:39.391103"
  },
  {
    "epoch": 170,
    "batch_idx": 128,
    "loss": 0.22875861823558807,
    "learning_rate": 0.000982278709228899,
    "timestep_mean": 513.78125,
    "epsilon_norm": 7.837432384490967,
    "prediction_norm": 7.536004066467285,
    "timestamp": "2025-07-14T12:33:39.396228"
  },
  {
    "epoch": 170,
    "batch_idx": 160,
    "loss": 0.3874083161354065,
    "learning_rate": 0.000982278709228899,
    "timestep_mean": 554.46875,
    "epsilon_norm": 7.870110988616943,
    "prediction_norm": 6.205276966094971,
    "timestamp": "2025-07-14T12:33:39.401127"
  },
  {
    "epoch": 170,
    "batch_idx": 192,
    "loss": 0.26888173818588257,
    "learning_rate": 0.000982278709228899,
    "timestep_mean": 484.125,
    "epsilon_norm": 3.6828107833862305,
    "prediction_norm": 3.19572114944458,
    "timestamp": "2025-07-14T12:33:39.405939"
  },
  {
    "epoch": 171,
    "batch_idx": 0,
    "loss": 0.4289851188659668,
    "learning_rate": 0.0009820708689179257,
    "timestep_mean": 411.625,
    "epsilon_norm": 7.486232757568359,
    "prediction_norm": 5.714501857757568,
    "timestamp": "2025-07-14T12:33:39.410686"
  },
  {
    "epoch": 171,
    "batch_idx": 32,
    "loss": 0.3482024669647217,
    "learning_rate": 0.0009820708689179257,
    "timestep_mean": 488.0,
    "epsilon_norm": 7.579895973205566,
    "prediction_norm": 6.440743446350098,
    "timestamp": "2025-07-14T12:33:39.415578"
  },
  {
    "epoch": 171,
    "batch_idx": 64,
    "loss": 0.3026631474494934,
    "learning_rate": 0.0009820708689179257,
    "timestep_mean": 603.96875,
    "epsilon_norm": 9.116048812866211,
    "prediction_norm": 8.178881645202637,
    "timestamp": "2025-07-14T12:33:39.420721"
  },
  {
    "epoch": 171,
    "batch_idx": 96,
    "loss": 0.5180021524429321,
    "learning_rate": 0.0009820708689179257,
    "timestep_mean": 438.75,
    "epsilon_norm": 8.237544059753418,
    "prediction_norm": 5.825473308563232,
    "timestamp": "2025-07-14T12:33:39.425743"
  },
  {
    "epoch": 171,
    "batch_idx": 128,
    "loss": 0.32065749168395996,
    "learning_rate": 0.0009820708689179257,
    "timestep_mean": 489.625,
    "epsilon_norm": 9.659438133239746,
    "prediction_norm": 8.243952751159668,
    "timestamp": "2025-07-14T12:33:39.430565"
  },
  {
    "epoch": 171,
    "batch_idx": 160,
    "loss": 0.37537094950675964,
    "learning_rate": 0.0009820708689179257,
    "timestep_mean": 473.40625,
    "epsilon_norm": 7.673277854919434,
    "prediction_norm": 6.008948802947998,
    "timestamp": "2025-07-14T12:33:39.435326"
  },
  {
    "epoch": 171,
    "batch_idx": 192,
    "loss": 0.3282703161239624,
    "learning_rate": 0.0009820708689179257,
    "timestep_mean": 427.0,
    "epsilon_norm": 4.110658645629883,
    "prediction_norm": 3.559800624847412,
    "timestamp": "2025-07-14T12:33:39.440248"
  },
  {
    "epoch": 172,
    "batch_idx": 0,
    "loss": 0.3247228264808655,
    "learning_rate": 0.0009818618391450047,
    "timestep_mean": 570.5625,
    "epsilon_norm": 8.582120895385742,
    "prediction_norm": 6.798023700714111,
    "timestamp": "2025-07-14T12:33:39.445100"
  },
  {
    "epoch": 172,
    "batch_idx": 32,
    "loss": 0.11809051036834717,
    "learning_rate": 0.0009818618391450047,
    "timestep_mean": 535.375,
    "epsilon_norm": 6.689433574676514,
    "prediction_norm": 6.011265754699707,
    "timestamp": "2025-07-14T12:33:39.450124"
  },
  {
    "epoch": 172,
    "batch_idx": 64,
    "loss": 0.5055235624313354,
    "learning_rate": 0.0009818618391450047,
    "timestep_mean": 520.375,
    "epsilon_norm": 8.034997940063477,
    "prediction_norm": 6.325371742248535,
    "timestamp": "2025-07-14T12:33:39.455046"
  },
  {
    "epoch": 172,
    "batch_idx": 96,
    "loss": 0.3040660321712494,
    "learning_rate": 0.0009818618391450047,
    "timestep_mean": 543.71875,
    "epsilon_norm": 8.013511657714844,
    "prediction_norm": 6.278163433074951,
    "timestamp": "2025-07-14T12:33:39.459913"
  },
  {
    "epoch": 172,
    "batch_idx": 128,
    "loss": 0.5051825046539307,
    "learning_rate": 0.0009818618391450047,
    "timestep_mean": 461.65625,
    "epsilon_norm": 7.413747310638428,
    "prediction_norm": 4.667425155639648,
    "timestamp": "2025-07-14T12:33:39.464744"
  },
  {
    "epoch": 172,
    "batch_idx": 160,
    "loss": 0.3842051923274994,
    "learning_rate": 0.0009818618391450047,
    "timestep_mean": 475.6875,
    "epsilon_norm": 8.516193389892578,
    "prediction_norm": 6.769611835479736,
    "timestamp": "2025-07-14T12:33:39.469661"
  },
  {
    "epoch": 172,
    "batch_idx": 192,
    "loss": 0.8450621366500854,
    "learning_rate": 0.0009818618391450047,
    "timestep_mean": 342.375,
    "epsilon_norm": 4.104984760284424,
    "prediction_norm": 2.130199670791626,
    "timestamp": "2025-07-14T12:33:39.474536"
  },
  {
    "epoch": 173,
    "batch_idx": 0,
    "loss": 0.3965105414390564,
    "learning_rate": 0.000981651620425896,
    "timestep_mean": 491.9375,
    "epsilon_norm": 8.271219253540039,
    "prediction_norm": 6.034837245941162,
    "timestamp": "2025-07-14T12:33:39.479374"
  },
  {
    "epoch": 173,
    "batch_idx": 32,
    "loss": 0.26292386651039124,
    "learning_rate": 0.000981651620425896,
    "timestep_mean": 499.25,
    "epsilon_norm": 7.252069473266602,
    "prediction_norm": 6.015021800994873,
    "timestamp": "2025-07-14T12:33:39.484240"
  },
  {
    "epoch": 173,
    "batch_idx": 64,
    "loss": 0.26001405715942383,
    "learning_rate": 0.000981651620425896,
    "timestep_mean": 488.625,
    "epsilon_norm": 7.832998752593994,
    "prediction_norm": 6.400143146514893,
    "timestamp": "2025-07-14T12:33:39.489123"
  },
  {
    "epoch": 173,
    "batch_idx": 96,
    "loss": 0.16652202606201172,
    "learning_rate": 0.000981651620425896,
    "timestep_mean": 567.875,
    "epsilon_norm": 7.29891300201416,
    "prediction_norm": 6.1943488121032715,
    "timestamp": "2025-07-14T12:33:39.493924"
  },
  {
    "epoch": 173,
    "batch_idx": 128,
    "loss": 0.30762428045272827,
    "learning_rate": 0.000981651620425896,
    "timestep_mean": 452.90625,
    "epsilon_norm": 7.6497650146484375,
    "prediction_norm": 6.127732753753662,
    "timestamp": "2025-07-14T12:33:39.498765"
  },
  {
    "epoch": 173,
    "batch_idx": 160,
    "loss": 0.25623422861099243,
    "learning_rate": 0.000981651620425896,
    "timestep_mean": 575.0,
    "epsilon_norm": 8.02786922454834,
    "prediction_norm": 7.473644256591797,
    "timestamp": "2025-07-14T12:33:39.503570"
  },
  {
    "epoch": 173,
    "batch_idx": 192,
    "loss": 0.18231113255023956,
    "learning_rate": 0.000981651620425896,
    "timestep_mean": 678.5,
    "epsilon_norm": 4.6004486083984375,
    "prediction_norm": 4.386364459991455,
    "timestamp": "2025-07-14T12:33:39.508522"
  },
  {
    "epoch": 174,
    "batch_idx": 0,
    "loss": 0.21083185076713562,
    "learning_rate": 0.0009814402132792936,
    "timestep_mean": 534.75,
    "epsilon_norm": 8.389742851257324,
    "prediction_norm": 8.253207206726074,
    "timestamp": "2025-07-14T12:33:39.513391"
  },
  {
    "epoch": 174,
    "batch_idx": 32,
    "loss": 0.31744053959846497,
    "learning_rate": 0.0009814402132792936,
    "timestep_mean": 462.5625,
    "epsilon_norm": 8.080623626708984,
    "prediction_norm": 6.015226364135742,
    "timestamp": "2025-07-14T12:33:39.518282"
  },
  {
    "epoch": 174,
    "batch_idx": 64,
    "loss": 0.31099966168403625,
    "learning_rate": 0.0009814402132792936,
    "timestep_mean": 504.5625,
    "epsilon_norm": 8.293122291564941,
    "prediction_norm": 7.325445175170898,
    "timestamp": "2025-07-14T12:33:39.523110"
  },
  {
    "epoch": 174,
    "batch_idx": 96,
    "loss": 0.4599640369415283,
    "learning_rate": 0.0009814402132792936,
    "timestep_mean": 520.84375,
    "epsilon_norm": 7.981637477874756,
    "prediction_norm": 6.5415544509887695,
    "timestamp": "2025-07-14T12:33:39.527939"
  },
  {
    "epoch": 174,
    "batch_idx": 128,
    "loss": 0.2535451650619507,
    "learning_rate": 0.0009814402132792936,
    "timestep_mean": 471.53125,
    "epsilon_norm": 7.8224196434021,
    "prediction_norm": 6.976755142211914,
    "timestamp": "2025-07-14T12:33:39.532787"
  },
  {
    "epoch": 174,
    "batch_idx": 160,
    "loss": 0.5735741853713989,
    "learning_rate": 0.0009814402132792936,
    "timestep_mean": 412.03125,
    "epsilon_norm": 8.118607521057129,
    "prediction_norm": 4.965949058532715,
    "timestamp": "2025-07-14T12:33:39.537745"
  },
  {
    "epoch": 174,
    "batch_idx": 192,
    "loss": 0.2486393302679062,
    "learning_rate": 0.0009814402132792936,
    "timestep_mean": 626.5,
    "epsilon_norm": 4.799699306488037,
    "prediction_norm": 4.080323696136475,
    "timestamp": "2025-07-14T12:33:39.542573"
  },
  {
    "epoch": 175,
    "batch_idx": 0,
    "loss": 0.40957117080688477,
    "learning_rate": 0.0009812276182268233,
    "timestep_mean": 497.0625,
    "epsilon_norm": 9.078442573547363,
    "prediction_norm": 8.246028900146484,
    "timestamp": "2025-07-14T12:33:39.547449"
  },
  {
    "epoch": 175,
    "batch_idx": 32,
    "loss": 0.4385968744754791,
    "learning_rate": 0.0009812276182268233,
    "timestep_mean": 486.75,
    "epsilon_norm": 8.706158638000488,
    "prediction_norm": 6.0584588050842285,
    "timestamp": "2025-07-14T12:33:39.552295"
  },
  {
    "epoch": 175,
    "batch_idx": 64,
    "loss": 0.40375620126724243,
    "learning_rate": 0.0009812276182268233,
    "timestep_mean": 424.21875,
    "epsilon_norm": 6.991126537322998,
    "prediction_norm": 5.208216190338135,
    "timestamp": "2025-07-14T12:33:39.557160"
  },
  {
    "epoch": 175,
    "batch_idx": 96,
    "loss": 0.34214094281196594,
    "learning_rate": 0.0009812276182268233,
    "timestep_mean": 565.0625,
    "epsilon_norm": 7.541213512420654,
    "prediction_norm": 5.9529337882995605,
    "timestamp": "2025-07-14T12:33:39.562022"
  },
  {
    "epoch": 175,
    "batch_idx": 128,
    "loss": 0.4224626123905182,
    "learning_rate": 0.0009812276182268233,
    "timestep_mean": 450.90625,
    "epsilon_norm": 8.69803237915039,
    "prediction_norm": 7.163569927215576,
    "timestamp": "2025-07-14T12:33:39.566967"
  },
  {
    "epoch": 175,
    "batch_idx": 160,
    "loss": 0.37173447012901306,
    "learning_rate": 0.0009812276182268233,
    "timestep_mean": 498.78125,
    "epsilon_norm": 8.429832458496094,
    "prediction_norm": 6.969112873077393,
    "timestamp": "2025-07-14T12:33:39.571765"
  },
  {
    "epoch": 175,
    "batch_idx": 192,
    "loss": 0.5863133668899536,
    "learning_rate": 0.0009812276182268233,
    "timestep_mean": 442.25,
    "epsilon_norm": 3.9908337593078613,
    "prediction_norm": 2.507891893386841,
    "timestamp": "2025-07-14T12:33:39.576664"
  },
  {
    "epoch": 176,
    "batch_idx": 0,
    "loss": 0.27810144424438477,
    "learning_rate": 0.0009810138357930426,
    "timestep_mean": 517.09375,
    "epsilon_norm": 7.78272008895874,
    "prediction_norm": 6.84654426574707,
    "timestamp": "2025-07-14T12:33:39.581544"
  },
  {
    "epoch": 176,
    "batch_idx": 32,
    "loss": 0.4051264524459839,
    "learning_rate": 0.0009810138357930426,
    "timestep_mean": 386.84375,
    "epsilon_norm": 7.816766262054443,
    "prediction_norm": 5.304885387420654,
    "timestamp": "2025-07-14T12:33:39.586458"
  },
  {
    "epoch": 176,
    "batch_idx": 64,
    "loss": 0.30212515592575073,
    "learning_rate": 0.0009810138357930426,
    "timestep_mean": 496.09375,
    "epsilon_norm": 7.664928913116455,
    "prediction_norm": 6.385034561157227,
    "timestamp": "2025-07-14T12:33:39.591330"
  },
  {
    "epoch": 176,
    "batch_idx": 96,
    "loss": 0.18643109500408173,
    "learning_rate": 0.0009810138357930426,
    "timestep_mean": 567.46875,
    "epsilon_norm": 6.4208664894104,
    "prediction_norm": 5.862259387969971,
    "timestamp": "2025-07-14T12:33:39.596189"
  },
  {
    "epoch": 176,
    "batch_idx": 128,
    "loss": 0.4047369360923767,
    "learning_rate": 0.0009810138357930426,
    "timestep_mean": 547.21875,
    "epsilon_norm": 8.064153671264648,
    "prediction_norm": 6.211160659790039,
    "timestamp": "2025-07-14T12:33:39.601009"
  },
  {
    "epoch": 176,
    "batch_idx": 160,
    "loss": 0.4482831656932831,
    "learning_rate": 0.0009810138357930426,
    "timestep_mean": 475.78125,
    "epsilon_norm": 8.439933776855469,
    "prediction_norm": 6.312402248382568,
    "timestamp": "2025-07-14T12:33:39.605837"
  },
  {
    "epoch": 176,
    "batch_idx": 192,
    "loss": 0.39557069540023804,
    "learning_rate": 0.0009810138357930426,
    "timestep_mean": 400.375,
    "epsilon_norm": 4.666389465332031,
    "prediction_norm": 3.9044270515441895,
    "timestamp": "2025-07-14T12:33:39.610713"
  },
  {
    "epoch": 177,
    "batch_idx": 0,
    "loss": 0.27158671617507935,
    "learning_rate": 0.0009807988665054383,
    "timestep_mean": 546.34375,
    "epsilon_norm": 7.704794883728027,
    "prediction_norm": 6.273433208465576,
    "timestamp": "2025-07-14T12:33:39.615557"
  },
  {
    "epoch": 177,
    "batch_idx": 32,
    "loss": 0.2402123361825943,
    "learning_rate": 0.0009807988665054383,
    "timestep_mean": 508.25,
    "epsilon_norm": 7.167256832122803,
    "prediction_norm": 5.970188617706299,
    "timestamp": "2025-07-14T12:33:39.620401"
  },
  {
    "epoch": 177,
    "batch_idx": 64,
    "loss": 0.3129746615886688,
    "learning_rate": 0.0009807988665054383,
    "timestep_mean": 524.40625,
    "epsilon_norm": 8.260685920715332,
    "prediction_norm": 7.250081539154053,
    "timestamp": "2025-07-14T12:33:39.625378"
  },
  {
    "epoch": 177,
    "batch_idx": 96,
    "loss": 0.3964511454105377,
    "learning_rate": 0.0009807988665054383,
    "timestep_mean": 515.59375,
    "epsilon_norm": 8.471002578735352,
    "prediction_norm": 6.257091522216797,
    "timestamp": "2025-07-14T12:33:39.630264"
  },
  {
    "epoch": 177,
    "batch_idx": 128,
    "loss": 0.4184057414531708,
    "learning_rate": 0.0009807988665054383,
    "timestep_mean": 533.78125,
    "epsilon_norm": 7.373198509216309,
    "prediction_norm": 5.545414447784424,
    "timestamp": "2025-07-14T12:33:39.635186"
  },
  {
    "epoch": 177,
    "batch_idx": 160,
    "loss": 0.3492205739021301,
    "learning_rate": 0.0009807988665054383,
    "timestep_mean": 541.125,
    "epsilon_norm": 7.262063980102539,
    "prediction_norm": 5.611822605133057,
    "timestamp": "2025-07-14T12:33:39.640043"
  },
  {
    "epoch": 177,
    "batch_idx": 192,
    "loss": 0.1777779459953308,
    "learning_rate": 0.0009807988665054383,
    "timestep_mean": 368.625,
    "epsilon_norm": 3.7064926624298096,
    "prediction_norm": 3.8253285884857178,
    "timestamp": "2025-07-14T12:33:39.644997"
  },
  {
    "epoch": 178,
    "batch_idx": 0,
    "loss": 0.25587886571884155,
    "learning_rate": 0.0009805827108944256,
    "timestep_mean": 553.40625,
    "epsilon_norm": 8.340118408203125,
    "prediction_norm": 7.49409294128418,
    "timestamp": "2025-07-14T12:33:39.649863"
  },
  {
    "epoch": 178,
    "batch_idx": 32,
    "loss": 0.30296510457992554,
    "learning_rate": 0.0009805827108944256,
    "timestep_mean": 551.6875,
    "epsilon_norm": 7.4776835441589355,
    "prediction_norm": 6.353960990905762,
    "timestamp": "2025-07-14T12:33:39.654726"
  },
  {
    "epoch": 178,
    "batch_idx": 64,
    "loss": 0.39369821548461914,
    "learning_rate": 0.0009805827108944256,
    "timestep_mean": 560.84375,
    "epsilon_norm": 8.993583679199219,
    "prediction_norm": 7.185943603515625,
    "timestamp": "2025-07-14T12:33:39.659542"
  },
  {
    "epoch": 178,
    "batch_idx": 96,
    "loss": 0.328580379486084,
    "learning_rate": 0.0009805827108944256,
    "timestep_mean": 448.25,
    "epsilon_norm": 7.971893310546875,
    "prediction_norm": 6.732362747192383,
    "timestamp": "2025-07-14T12:33:39.664380"
  },
  {
    "epoch": 178,
    "batch_idx": 128,
    "loss": 0.2294812947511673,
    "learning_rate": 0.0009805827108944256,
    "timestep_mean": 569.78125,
    "epsilon_norm": 8.372142791748047,
    "prediction_norm": 7.507235527038574,
    "timestamp": "2025-07-14T12:33:39.669231"
  },
  {
    "epoch": 178,
    "batch_idx": 160,
    "loss": 0.5838268995285034,
    "learning_rate": 0.0009805827108944256,
    "timestep_mean": 513.1875,
    "epsilon_norm": 9.943658828735352,
    "prediction_norm": 7.680267810821533,
    "timestamp": "2025-07-14T12:33:39.674051"
  },
  {
    "epoch": 178,
    "batch_idx": 192,
    "loss": 0.043553680181503296,
    "learning_rate": 0.0009805827108944256,
    "timestep_mean": 693.5,
    "epsilon_norm": 3.401089906692505,
    "prediction_norm": 3.3508074283599854,
    "timestamp": "2025-07-14T12:33:39.678995"
  },
  {
    "epoch": 179,
    "batch_idx": 0,
    "loss": 0.3651553988456726,
    "learning_rate": 0.0009803653694933471,
    "timestep_mean": 561.5625,
    "epsilon_norm": 8.966453552246094,
    "prediction_norm": 7.093358993530273,
    "timestamp": "2025-07-14T12:33:39.683860"
  },
  {
    "epoch": 179,
    "batch_idx": 32,
    "loss": 0.377763032913208,
    "learning_rate": 0.0009803653694933471,
    "timestep_mean": 472.40625,
    "epsilon_norm": 8.200332641601562,
    "prediction_norm": 6.662075042724609,
    "timestamp": "2025-07-14T12:33:39.688729"
  },
  {
    "epoch": 179,
    "batch_idx": 64,
    "loss": 0.215434268116951,
    "learning_rate": 0.0009803653694933471,
    "timestep_mean": 608.59375,
    "epsilon_norm": 8.302151679992676,
    "prediction_norm": 7.3043107986450195,
    "timestamp": "2025-07-14T12:33:39.693598"
  },
  {
    "epoch": 179,
    "batch_idx": 96,
    "loss": 0.5588961839675903,
    "learning_rate": 0.0009803653694933471,
    "timestep_mean": 500.71875,
    "epsilon_norm": 7.7577409744262695,
    "prediction_norm": 4.727611541748047,
    "timestamp": "2025-07-14T12:33:39.698459"
  },
  {
    "epoch": 179,
    "batch_idx": 128,
    "loss": 0.5922113656997681,
    "learning_rate": 0.0009803653694933471,
    "timestep_mean": 411.0625,
    "epsilon_norm": 8.412971496582031,
    "prediction_norm": 5.553033828735352,
    "timestamp": "2025-07-14T12:33:39.703302"
  },
  {
    "epoch": 179,
    "batch_idx": 160,
    "loss": 0.31469273567199707,
    "learning_rate": 0.0009803653694933471,
    "timestep_mean": 526.625,
    "epsilon_norm": 8.817230224609375,
    "prediction_norm": 8.001930236816406,
    "timestamp": "2025-07-14T12:33:39.708169"
  },
  {
    "epoch": 179,
    "batch_idx": 192,
    "loss": 0.5080353021621704,
    "learning_rate": 0.0009803653694933471,
    "timestep_mean": 475.375,
    "epsilon_norm": 3.5511670112609863,
    "prediction_norm": 1.9446252584457397,
    "timestamp": "2025-07-14T12:33:39.713072"
  },
  {
    "epoch": 180,
    "batch_idx": 0,
    "loss": 0.4354325532913208,
    "learning_rate": 0.0009801468428384712,
    "timestep_mean": 504.4375,
    "epsilon_norm": 8.02587890625,
    "prediction_norm": 6.6301093101501465,
    "timestamp": "2025-07-14T12:33:39.717959"
  },
  {
    "epoch": 180,
    "batch_idx": 32,
    "loss": 0.287356436252594,
    "learning_rate": 0.0009801468428384712,
    "timestep_mean": 485.28125,
    "epsilon_norm": 7.466351509094238,
    "prediction_norm": 5.384911060333252,
    "timestamp": "2025-07-14T12:33:39.722829"
  },
  {
    "epoch": 180,
    "batch_idx": 64,
    "loss": 0.22739958763122559,
    "learning_rate": 0.0009801468428384712,
    "timestep_mean": 565.34375,
    "epsilon_norm": 6.986822605133057,
    "prediction_norm": 6.704452991485596,
    "timestamp": "2025-07-14T12:33:39.727688"
  },
  {
    "epoch": 180,
    "batch_idx": 96,
    "loss": 0.27723830938339233,
    "learning_rate": 0.0009801468428384712,
    "timestep_mean": 535.21875,
    "epsilon_norm": 8.47897720336914,
    "prediction_norm": 7.358461856842041,
    "timestamp": "2025-07-14T12:33:39.732646"
  },
  {
    "epoch": 180,
    "batch_idx": 128,
    "loss": 0.25038808584213257,
    "learning_rate": 0.0009801468428384712,
    "timestep_mean": 509.28125,
    "epsilon_norm": 7.173384189605713,
    "prediction_norm": 6.7296600341796875,
    "timestamp": "2025-07-14T12:33:39.737646"
  },
  {
    "epoch": 180,
    "batch_idx": 160,
    "loss": 0.4310612082481384,
    "learning_rate": 0.0009801468428384712,
    "timestep_mean": 469.15625,
    "epsilon_norm": 8.078624725341797,
    "prediction_norm": 5.580563068389893,
    "timestamp": "2025-07-14T12:33:39.742549"
  },
  {
    "epoch": 180,
    "batch_idx": 192,
    "loss": 0.4434806704521179,
    "learning_rate": 0.0009801468428384712,
    "timestep_mean": 406.5,
    "epsilon_norm": 4.67334508895874,
    "prediction_norm": 3.8619208335876465,
    "timestamp": "2025-07-14T12:33:39.747457"
  },
  {
    "epoch": 181,
    "batch_idx": 0,
    "loss": 0.3072177767753601,
    "learning_rate": 0.0009799271314689904,
    "timestep_mean": 517.46875,
    "epsilon_norm": 7.714419364929199,
    "prediction_norm": 6.375462055206299,
    "timestamp": "2025-07-14T12:33:39.752353"
  },
  {
    "epoch": 181,
    "batch_idx": 32,
    "loss": 0.39657220244407654,
    "learning_rate": 0.0009799271314689904,
    "timestep_mean": 422.8125,
    "epsilon_norm": 7.165322303771973,
    "prediction_norm": 4.909125804901123,
    "timestamp": "2025-07-14T12:33:39.757293"
  },
  {
    "epoch": 181,
    "batch_idx": 64,
    "loss": 0.4308553636074066,
    "learning_rate": 0.0009799271314689904,
    "timestep_mean": 505.1875,
    "epsilon_norm": 8.525214195251465,
    "prediction_norm": 6.63927698135376,
    "timestamp": "2025-07-14T12:33:39.762279"
  },
  {
    "epoch": 181,
    "batch_idx": 96,
    "loss": 0.21542717516422272,
    "learning_rate": 0.0009799271314689904,
    "timestep_mean": 526.5,
    "epsilon_norm": 7.1015214920043945,
    "prediction_norm": 6.184960842132568,
    "timestamp": "2025-07-14T12:33:39.767187"
  },
  {
    "epoch": 181,
    "batch_idx": 128,
    "loss": 0.43567657470703125,
    "learning_rate": 0.0009799271314689904,
    "timestep_mean": 460.125,
    "epsilon_norm": 7.424083709716797,
    "prediction_norm": 5.531344413757324,
    "timestamp": "2025-07-14T12:33:39.772012"
  },
  {
    "epoch": 181,
    "batch_idx": 160,
    "loss": 0.2680552303791046,
    "learning_rate": 0.0009799271314689904,
    "timestep_mean": 569.34375,
    "epsilon_norm": 8.72679328918457,
    "prediction_norm": 7.979001998901367,
    "timestamp": "2025-07-14T12:33:39.776896"
  },
  {
    "epoch": 181,
    "batch_idx": 192,
    "loss": 0.4288795292377472,
    "learning_rate": 0.0009799271314689904,
    "timestep_mean": 479.875,
    "epsilon_norm": 3.456712245941162,
    "prediction_norm": 2.1040713787078857,
    "timestamp": "2025-07-14T12:33:39.781809"
  },
  {
    "epoch": 182,
    "batch_idx": 0,
    "loss": 0.6230726838111877,
    "learning_rate": 0.000979706235927021,
    "timestep_mean": 469.96875,
    "epsilon_norm": 8.934820175170898,
    "prediction_norm": 7.3855438232421875,
    "timestamp": "2025-07-14T12:33:39.786973"
  },
  {
    "epoch": 182,
    "batch_idx": 32,
    "loss": 0.38711607456207275,
    "learning_rate": 0.000979706235927021,
    "timestep_mean": 494.40625,
    "epsilon_norm": 8.464391708374023,
    "prediction_norm": 6.009127140045166,
    "timestamp": "2025-07-14T12:33:39.792171"
  },
  {
    "epoch": 182,
    "batch_idx": 64,
    "loss": 0.5403153896331787,
    "learning_rate": 0.000979706235927021,
    "timestep_mean": 487.375,
    "epsilon_norm": 9.117857933044434,
    "prediction_norm": 7.543418884277344,
    "timestamp": "2025-07-14T12:33:39.797300"
  },
  {
    "epoch": 182,
    "batch_idx": 96,
    "loss": 0.22193661332130432,
    "learning_rate": 0.000979706235927021,
    "timestep_mean": 614.59375,
    "epsilon_norm": 7.336548805236816,
    "prediction_norm": 6.034757614135742,
    "timestamp": "2025-07-14T12:33:39.802411"
  },
  {
    "epoch": 182,
    "batch_idx": 128,
    "loss": 0.28976738452911377,
    "learning_rate": 0.000979706235927021,
    "timestep_mean": 573.03125,
    "epsilon_norm": 7.994943618774414,
    "prediction_norm": 6.720212459564209,
    "timestamp": "2025-07-14T12:33:39.807516"
  },
  {
    "epoch": 182,
    "batch_idx": 160,
    "loss": 0.3412122130393982,
    "learning_rate": 0.000979706235927021,
    "timestep_mean": 474.125,
    "epsilon_norm": 7.236384391784668,
    "prediction_norm": 5.855807304382324,
    "timestamp": "2025-07-14T12:33:39.812555"
  },
  {
    "epoch": 182,
    "batch_idx": 192,
    "loss": 0.05868188291788101,
    "learning_rate": 0.000979706235927021,
    "timestep_mean": 605.375,
    "epsilon_norm": 3.6600558757781982,
    "prediction_norm": 3.1334428787231445,
    "timestamp": "2025-07-14T12:33:39.817731"
  },
  {
    "epoch": 183,
    "batch_idx": 0,
    "loss": 0.49834710359573364,
    "learning_rate": 0.0009794841567576007,
    "timestep_mean": 445.78125,
    "epsilon_norm": 7.185330867767334,
    "prediction_norm": 4.672514915466309,
    "timestamp": "2025-07-14T12:33:39.822913"
  },
  {
    "epoch": 183,
    "batch_idx": 32,
    "loss": 0.31449854373931885,
    "learning_rate": 0.0009794841567576007,
    "timestep_mean": 556.03125,
    "epsilon_norm": 7.2808380126953125,
    "prediction_norm": 6.563467979431152,
    "timestamp": "2025-07-14T12:33:39.828083"
  },
  {
    "epoch": 183,
    "batch_idx": 64,
    "loss": 0.2631315290927887,
    "learning_rate": 0.0009794841567576007,
    "timestep_mean": 466.0625,
    "epsilon_norm": 8.481683731079102,
    "prediction_norm": 8.058533668518066,
    "timestamp": "2025-07-14T12:33:39.833224"
  },
  {
    "epoch": 183,
    "batch_idx": 96,
    "loss": 0.5218937397003174,
    "learning_rate": 0.0009794841567576007,
    "timestep_mean": 506.1875,
    "epsilon_norm": 7.728343486785889,
    "prediction_norm": 5.852066993713379,
    "timestamp": "2025-07-14T12:33:39.838381"
  },
  {
    "epoch": 183,
    "batch_idx": 128,
    "loss": 0.33641210198402405,
    "learning_rate": 0.0009794841567576007,
    "timestep_mean": 506.15625,
    "epsilon_norm": 7.939610481262207,
    "prediction_norm": 6.556996822357178,
    "timestamp": "2025-07-14T12:33:39.843552"
  },
  {
    "epoch": 183,
    "batch_idx": 160,
    "loss": 0.5687718391418457,
    "learning_rate": 0.0009794841567576007,
    "timestep_mean": 451.875,
    "epsilon_norm": 8.374176979064941,
    "prediction_norm": 5.048730850219727,
    "timestamp": "2025-07-14T12:33:39.848716"
  },
  {
    "epoch": 183,
    "batch_idx": 192,
    "loss": 0.277981162071228,
    "learning_rate": 0.0009794841567576007,
    "timestep_mean": 574.875,
    "epsilon_norm": 4.070099830627441,
    "prediction_norm": 3.254218578338623,
    "timestamp": "2025-07-14T12:33:39.853958"
  },
  {
    "epoch": 184,
    "batch_idx": 0,
    "loss": 0.20545318722724915,
    "learning_rate": 0.0009792608945086876,
    "timestep_mean": 595.625,
    "epsilon_norm": 7.0857720375061035,
    "prediction_norm": 6.326624870300293,
    "timestamp": "2025-07-14T12:33:39.859146"
  },
  {
    "epoch": 184,
    "batch_idx": 32,
    "loss": 0.43864309787750244,
    "learning_rate": 0.0009792608945086876,
    "timestep_mean": 526.84375,
    "epsilon_norm": 9.273734092712402,
    "prediction_norm": 6.468393325805664,
    "timestamp": "2025-07-14T12:33:39.864228"
  },
  {
    "epoch": 184,
    "batch_idx": 64,
    "loss": 0.3830002546310425,
    "learning_rate": 0.0009792608945086876,
    "timestep_mean": 476.96875,
    "epsilon_norm": 7.160870552062988,
    "prediction_norm": 4.715700626373291,
    "timestamp": "2025-07-14T12:33:39.869366"
  },
  {
    "epoch": 184,
    "batch_idx": 96,
    "loss": 0.397623211145401,
    "learning_rate": 0.0009792608945086876,
    "timestep_mean": 498.25,
    "epsilon_norm": 8.475764274597168,
    "prediction_norm": 6.3310112953186035,
    "timestamp": "2025-07-14T12:33:39.874516"
  },
  {
    "epoch": 184,
    "batch_idx": 128,
    "loss": 0.27488434314727783,
    "learning_rate": 0.0009792608945086876,
    "timestep_mean": 577.53125,
    "epsilon_norm": 7.91878080368042,
    "prediction_norm": 5.592525005340576,
    "timestamp": "2025-07-14T12:33:39.879835"
  },
  {
    "epoch": 184,
    "batch_idx": 160,
    "loss": 0.2936609387397766,
    "learning_rate": 0.0009792608945086876,
    "timestep_mean": 553.03125,
    "epsilon_norm": 6.877253532409668,
    "prediction_norm": 5.12551212310791,
    "timestamp": "2025-07-14T12:33:39.884978"
  },
  {
    "epoch": 184,
    "batch_idx": 192,
    "loss": 0.3334496319293976,
    "learning_rate": 0.0009792608945086876,
    "timestep_mean": 509.875,
    "epsilon_norm": 2.8732123374938965,
    "prediction_norm": 2.2603700160980225,
    "timestamp": "2025-07-14T12:33:39.890001"
  },
  {
    "epoch": 185,
    "batch_idx": 0,
    "loss": 0.38441023230552673,
    "learning_rate": 0.0009790364497311594,
    "timestep_mean": 479.8125,
    "epsilon_norm": 8.343108177185059,
    "prediction_norm": 6.470534801483154,
    "timestamp": "2025-07-14T12:33:39.895184"
  },
  {
    "epoch": 185,
    "batch_idx": 32,
    "loss": 0.5070953965187073,
    "learning_rate": 0.0009790364497311594,
    "timestep_mean": 416.28125,
    "epsilon_norm": 8.027688980102539,
    "prediction_norm": 6.287008285522461,
    "timestamp": "2025-07-14T12:33:39.905739"
  },
  {
    "epoch": 185,
    "batch_idx": 64,
    "loss": 0.15490640699863434,
    "learning_rate": 0.0009790364497311594,
    "timestep_mean": 612.46875,
    "epsilon_norm": 8.193510055541992,
    "prediction_norm": 8.761298179626465,
    "timestamp": "2025-07-14T12:33:39.910993"
  },
  {
    "epoch": 185,
    "batch_idx": 96,
    "loss": 0.28511282801628113,
    "learning_rate": 0.0009790364497311594,
    "timestep_mean": 510.875,
    "epsilon_norm": 7.876699447631836,
    "prediction_norm": 7.142436981201172,
    "timestamp": "2025-07-14T12:33:39.916173"
  },
  {
    "epoch": 185,
    "batch_idx": 128,
    "loss": 0.39594200253486633,
    "learning_rate": 0.0009790364497311594,
    "timestep_mean": 542.25,
    "epsilon_norm": 7.264502048492432,
    "prediction_norm": 6.3727827072143555,
    "timestamp": "2025-07-14T12:33:39.921297"
  },
  {
    "epoch": 185,
    "batch_idx": 160,
    "loss": 0.2677960991859436,
    "learning_rate": 0.0009790364497311594,
    "timestep_mean": 489.90625,
    "epsilon_norm": 7.284134387969971,
    "prediction_norm": 6.018970489501953,
    "timestamp": "2025-07-14T12:33:39.926455"
  },
  {
    "epoch": 185,
    "batch_idx": 192,
    "loss": 0.6458698511123657,
    "learning_rate": 0.0009790364497311594,
    "timestep_mean": 460.375,
    "epsilon_norm": 3.629707098007202,
    "prediction_norm": 2.004883050918579,
    "timestamp": "2025-07-14T12:33:39.931701"
  },
  {
    "epoch": 186,
    "batch_idx": 0,
    "loss": 0.2725680470466614,
    "learning_rate": 0.0009788108229788108,
    "timestep_mean": 532.15625,
    "epsilon_norm": 7.754539489746094,
    "prediction_norm": 6.279142379760742,
    "timestamp": "2025-07-14T12:33:39.936909"
  },
  {
    "epoch": 186,
    "batch_idx": 32,
    "loss": 0.22211119532585144,
    "learning_rate": 0.0009788108229788108,
    "timestep_mean": 569.40625,
    "epsilon_norm": 7.758828639984131,
    "prediction_norm": 6.494571685791016,
    "timestamp": "2025-07-14T12:33:39.942226"
  },
  {
    "epoch": 186,
    "batch_idx": 64,
    "loss": 0.2865215539932251,
    "learning_rate": 0.0009788108229788108,
    "timestep_mean": 496.59375,
    "epsilon_norm": 7.725531101226807,
    "prediction_norm": 6.309596061706543,
    "timestamp": "2025-07-14T12:33:39.947305"
  },
  {
    "epoch": 186,
    "batch_idx": 96,
    "loss": 0.6015092134475708,
    "learning_rate": 0.0009788108229788108,
    "timestep_mean": 487.65625,
    "epsilon_norm": 8.583756446838379,
    "prediction_norm": 6.346527576446533,
    "timestamp": "2025-07-14T12:33:39.952431"
  },
  {
    "epoch": 186,
    "batch_idx": 128,
    "loss": 0.34771037101745605,
    "learning_rate": 0.0009788108229788108,
    "timestep_mean": 441.625,
    "epsilon_norm": 6.912590980529785,
    "prediction_norm": 4.764747142791748,
    "timestamp": "2025-07-14T12:33:39.957700"
  },
  {
    "epoch": 186,
    "batch_idx": 160,
    "loss": 0.49160706996917725,
    "learning_rate": 0.0009788108229788108,
    "timestep_mean": 497.375,
    "epsilon_norm": 9.62175178527832,
    "prediction_norm": 7.251364707946777,
    "timestamp": "2025-07-14T12:33:39.962807"
  },
  {
    "epoch": 186,
    "batch_idx": 192,
    "loss": 0.2685173451900482,
    "learning_rate": 0.0009788108229788108,
    "timestep_mean": 565.125,
    "epsilon_norm": 3.272137403488159,
    "prediction_norm": 2.4397876262664795,
    "timestamp": "2025-07-14T12:33:39.967947"
  },
  {
    "epoch": 187,
    "batch_idx": 0,
    "loss": 0.37986326217651367,
    "learning_rate": 0.000978584014808354,
    "timestep_mean": 451.4375,
    "epsilon_norm": 8.704962730407715,
    "prediction_norm": 7.178049564361572,
    "timestamp": "2025-07-14T12:33:39.973143"
  },
  {
    "epoch": 187,
    "batch_idx": 32,
    "loss": 0.30907773971557617,
    "learning_rate": 0.000978584014808354,
    "timestep_mean": 514.40625,
    "epsilon_norm": 7.529425621032715,
    "prediction_norm": 6.1076741218566895,
    "timestamp": "2025-07-14T12:33:39.978316"
  },
  {
    "epoch": 187,
    "batch_idx": 64,
    "loss": 0.5618243217468262,
    "learning_rate": 0.000978584014808354,
    "timestep_mean": 422.71875,
    "epsilon_norm": 8.453949928283691,
    "prediction_norm": 6.400681018829346,
    "timestamp": "2025-07-14T12:33:39.983523"
  },
  {
    "epoch": 187,
    "batch_idx": 96,
    "loss": 0.30220580101013184,
    "learning_rate": 0.000978584014808354,
    "timestep_mean": 518.09375,
    "epsilon_norm": 6.771994590759277,
    "prediction_norm": 5.455907821655273,
    "timestamp": "2025-07-14T12:33:39.988584"
  },
  {
    "epoch": 187,
    "batch_idx": 128,
    "loss": 0.42924487590789795,
    "learning_rate": 0.000978584014808354,
    "timestep_mean": 492.78125,
    "epsilon_norm": 7.234229564666748,
    "prediction_norm": 5.576676368713379,
    "timestamp": "2025-07-14T12:33:39.993813"
  },
  {
    "epoch": 187,
    "batch_idx": 160,
    "loss": 0.41890937089920044,
    "learning_rate": 0.000978584014808354,
    "timestep_mean": 468.84375,
    "epsilon_norm": 7.6920366287231445,
    "prediction_norm": 5.726891040802002,
    "timestamp": "2025-07-14T12:33:39.999048"
  },
  {
    "epoch": 187,
    "batch_idx": 192,
    "loss": 0.4540422558784485,
    "learning_rate": 0.000978584014808354,
    "timestep_mean": 466.625,
    "epsilon_norm": 3.4750401973724365,
    "prediction_norm": 2.335958957672119,
    "timestamp": "2025-07-14T12:33:40.004212"
  },
  {
    "epoch": 188,
    "batch_idx": 0,
    "loss": 0.1516120433807373,
    "learning_rate": 0.0009783560257794152,
    "timestep_mean": 638.15625,
    "epsilon_norm": 8.068498611450195,
    "prediction_norm": 7.305166721343994,
    "timestamp": "2025-07-14T12:33:40.009403"
  },
  {
    "epoch": 188,
    "batch_idx": 32,
    "loss": 0.3519686460494995,
    "learning_rate": 0.0009783560257794152,
    "timestep_mean": 576.375,
    "epsilon_norm": 8.885687828063965,
    "prediction_norm": 7.266564846038818,
    "timestamp": "2025-07-14T12:33:40.014546"
  },
  {
    "epoch": 188,
    "batch_idx": 64,
    "loss": 0.239063560962677,
    "learning_rate": 0.0009783560257794152,
    "timestep_mean": 582.0,
    "epsilon_norm": 6.891904354095459,
    "prediction_norm": 5.193813800811768,
    "timestamp": "2025-07-14T12:33:40.019754"
  },
  {
    "epoch": 188,
    "batch_idx": 96,
    "loss": 0.1675015389919281,
    "learning_rate": 0.0009783560257794152,
    "timestep_mean": 551.25,
    "epsilon_norm": 7.228625774383545,
    "prediction_norm": 6.466637134552002,
    "timestamp": "2025-07-14T12:33:40.024885"
  },
  {
    "epoch": 188,
    "batch_idx": 128,
    "loss": 0.3563874065876007,
    "learning_rate": 0.0009783560257794152,
    "timestep_mean": 513.5625,
    "epsilon_norm": 8.025548934936523,
    "prediction_norm": 6.40635347366333,
    "timestamp": "2025-07-14T12:33:40.029994"
  },
  {
    "epoch": 188,
    "batch_idx": 160,
    "loss": 0.1692546308040619,
    "learning_rate": 0.0009783560257794152,
    "timestep_mean": 470.1875,
    "epsilon_norm": 6.621224403381348,
    "prediction_norm": 5.73070764541626,
    "timestamp": "2025-07-14T12:33:40.035203"
  },
  {
    "epoch": 188,
    "batch_idx": 192,
    "loss": 0.1848030388355255,
    "learning_rate": 0.0009783560257794152,
    "timestep_mean": 506.875,
    "epsilon_norm": 3.2052156925201416,
    "prediction_norm": 2.5367891788482666,
    "timestamp": "2025-07-14T12:33:40.040400"
  },
  {
    "epoch": 189,
    "batch_idx": 0,
    "loss": 0.34210342168807983,
    "learning_rate": 0.0009781268564545346,
    "timestep_mean": 452.25,
    "epsilon_norm": 6.648603916168213,
    "prediction_norm": 4.770050525665283,
    "timestamp": "2025-07-14T12:33:40.045662"
  },
  {
    "epoch": 189,
    "batch_idx": 32,
    "loss": 0.3679826557636261,
    "learning_rate": 0.0009781268564545346,
    "timestep_mean": 477.40625,
    "epsilon_norm": 8.181939125061035,
    "prediction_norm": 5.905539035797119,
    "timestamp": "2025-07-14T12:33:40.050728"
  },
  {
    "epoch": 189,
    "batch_idx": 64,
    "loss": 0.4244447648525238,
    "learning_rate": 0.0009781268564545346,
    "timestep_mean": 477.6875,
    "epsilon_norm": 8.084488868713379,
    "prediction_norm": 5.866314888000488,
    "timestamp": "2025-07-14T12:33:40.055888"
  },
  {
    "epoch": 189,
    "batch_idx": 96,
    "loss": 0.2995337247848511,
    "learning_rate": 0.0009781268564545346,
    "timestep_mean": 564.9375,
    "epsilon_norm": 8.40334701538086,
    "prediction_norm": 7.822577953338623,
    "timestamp": "2025-07-14T12:33:40.061117"
  },
  {
    "epoch": 189,
    "batch_idx": 128,
    "loss": 0.5645056962966919,
    "learning_rate": 0.0009781268564545346,
    "timestep_mean": 493.96875,
    "epsilon_norm": 8.377948760986328,
    "prediction_norm": 5.808204650878906,
    "timestamp": "2025-07-14T12:33:40.066281"
  },
  {
    "epoch": 189,
    "batch_idx": 160,
    "loss": 0.15696915984153748,
    "learning_rate": 0.0009781268564545346,
    "timestep_mean": 552.59375,
    "epsilon_norm": 8.075418472290039,
    "prediction_norm": 7.4205451011657715,
    "timestamp": "2025-07-14T12:33:40.071600"
  },
  {
    "epoch": 189,
    "batch_idx": 192,
    "loss": 0.24055691063404083,
    "learning_rate": 0.0009781268564545346,
    "timestep_mean": 577.875,
    "epsilon_norm": 4.182394981384277,
    "prediction_norm": 3.679853916168213,
    "timestamp": "2025-07-14T12:33:40.076420"
  },
  {
    "epoch": 190,
    "batch_idx": 0,
    "loss": 0.5858229398727417,
    "learning_rate": 0.000977896507399165,
    "timestep_mean": 422.375,
    "epsilon_norm": 8.677468299865723,
    "prediction_norm": 6.3164238929748535,
    "timestamp": "2025-07-14T12:33:40.081532"
  },
  {
    "epoch": 190,
    "batch_idx": 32,
    "loss": 0.3042675256729126,
    "learning_rate": 0.000977896507399165,
    "timestep_mean": 522.03125,
    "epsilon_norm": 8.09389877319336,
    "prediction_norm": 7.423672199249268,
    "timestamp": "2025-07-14T12:33:40.086725"
  },
  {
    "epoch": 190,
    "batch_idx": 64,
    "loss": 0.439924955368042,
    "learning_rate": 0.000977896507399165,
    "timestep_mean": 537.5625,
    "epsilon_norm": 8.783512115478516,
    "prediction_norm": 7.131798267364502,
    "timestamp": "2025-07-14T12:33:40.092038"
  },
  {
    "epoch": 190,
    "batch_idx": 96,
    "loss": 0.5670223236083984,
    "learning_rate": 0.000977896507399165,
    "timestep_mean": 530.78125,
    "epsilon_norm": 8.966215133666992,
    "prediction_norm": 6.667352676391602,
    "timestamp": "2025-07-14T12:33:40.097157"
  },
  {
    "epoch": 190,
    "batch_idx": 128,
    "loss": 0.3594551682472229,
    "learning_rate": 0.000977896507399165,
    "timestep_mean": 605.28125,
    "epsilon_norm": 8.919718742370605,
    "prediction_norm": 7.968746185302734,
    "timestamp": "2025-07-14T12:33:40.102355"
  },
  {
    "epoch": 190,
    "batch_idx": 160,
    "loss": 0.21847623586654663,
    "learning_rate": 0.000977896507399165,
    "timestep_mean": 567.75,
    "epsilon_norm": 7.5184502601623535,
    "prediction_norm": 7.399594306945801,
    "timestamp": "2025-07-14T12:33:40.107560"
  },
  {
    "epoch": 190,
    "batch_idx": 192,
    "loss": 0.15746088325977325,
    "learning_rate": 0.000977896507399165,
    "timestep_mean": 459.5,
    "epsilon_norm": 4.466811180114746,
    "prediction_norm": 4.513491153717041,
    "timestamp": "2025-07-14T12:33:40.112837"
  },
  {
    "epoch": 191,
    "batch_idx": 0,
    "loss": 0.4263994097709656,
    "learning_rate": 0.0009776649791816698,
    "timestep_mean": 431.0,
    "epsilon_norm": 7.9144158363342285,
    "prediction_norm": 5.552617073059082,
    "timestamp": "2025-07-14T12:33:40.118010"
  },
  {
    "epoch": 191,
    "batch_idx": 32,
    "loss": 0.2010059952735901,
    "learning_rate": 0.0009776649791816698,
    "timestep_mean": 504.03125,
    "epsilon_norm": 7.483516693115234,
    "prediction_norm": 6.133314609527588,
    "timestamp": "2025-07-14T12:33:40.123768"
  },
  {
    "epoch": 191,
    "batch_idx": 64,
    "loss": 0.28077179193496704,
    "learning_rate": 0.0009776649791816698,
    "timestep_mean": 423.125,
    "epsilon_norm": 8.605081558227539,
    "prediction_norm": 7.675351619720459,
    "timestamp": "2025-07-14T12:33:40.129472"
  },
  {
    "epoch": 191,
    "batch_idx": 96,
    "loss": 0.4443581700325012,
    "learning_rate": 0.0009776649791816698,
    "timestep_mean": 504.5625,
    "epsilon_norm": 8.01301097869873,
    "prediction_norm": 4.882139682769775,
    "timestamp": "2025-07-14T12:33:40.135010"
  },
  {
    "epoch": 191,
    "batch_idx": 128,
    "loss": 0.5605607628822327,
    "learning_rate": 0.0009776649791816698,
    "timestep_mean": 420.84375,
    "epsilon_norm": 7.719837665557861,
    "prediction_norm": 4.737612724304199,
    "timestamp": "2025-07-14T12:33:40.140497"
  },
  {
    "epoch": 191,
    "batch_idx": 160,
    "loss": 0.604783296585083,
    "learning_rate": 0.0009776649791816698,
    "timestep_mean": 488.40625,
    "epsilon_norm": 8.29819107055664,
    "prediction_norm": 5.588621139526367,
    "timestamp": "2025-07-14T12:33:40.145917"
  },
  {
    "epoch": 191,
    "batch_idx": 192,
    "loss": 0.1657698154449463,
    "learning_rate": 0.0009776649791816698,
    "timestep_mean": 665.25,
    "epsilon_norm": 3.3241477012634277,
    "prediction_norm": 2.615811347961426,
    "timestamp": "2025-07-14T12:33:40.151074"
  },
  {
    "epoch": 192,
    "batch_idx": 0,
    "loss": 0.18320134282112122,
    "learning_rate": 0.0009774322723733217,
    "timestep_mean": 496.09375,
    "epsilon_norm": 7.526070594787598,
    "prediction_norm": 6.525278568267822,
    "timestamp": "2025-07-14T12:33:40.156305"
  },
  {
    "epoch": 192,
    "batch_idx": 32,
    "loss": 0.26825180649757385,
    "learning_rate": 0.0009774322723733217,
    "timestep_mean": 480.8125,
    "epsilon_norm": 8.963095664978027,
    "prediction_norm": 7.227916717529297,
    "timestamp": "2025-07-14T12:33:40.161563"
  },
  {
    "epoch": 192,
    "batch_idx": 64,
    "loss": 0.39665746688842773,
    "learning_rate": 0.0009774322723733217,
    "timestep_mean": 509.4375,
    "epsilon_norm": 7.181318283081055,
    "prediction_norm": 5.296255588531494,
    "timestamp": "2025-07-14T12:33:40.166829"
  },
  {
    "epoch": 192,
    "batch_idx": 96,
    "loss": 0.3114614188671112,
    "learning_rate": 0.0009774322723733217,
    "timestep_mean": 523.875,
    "epsilon_norm": 7.445291519165039,
    "prediction_norm": 6.790407657623291,
    "timestamp": "2025-07-14T12:33:40.171932"
  },
  {
    "epoch": 192,
    "batch_idx": 128,
    "loss": 0.18124042451381683,
    "learning_rate": 0.0009774322723733217,
    "timestep_mean": 627.28125,
    "epsilon_norm": 7.131467819213867,
    "prediction_norm": 6.75841760635376,
    "timestamp": "2025-07-14T12:33:40.177119"
  },
  {
    "epoch": 192,
    "batch_idx": 160,
    "loss": 0.4005388617515564,
    "learning_rate": 0.0009774322723733217,
    "timestep_mean": 478.0625,
    "epsilon_norm": 8.804417610168457,
    "prediction_norm": 7.787439346313477,
    "timestamp": "2025-07-14T12:33:40.182345"
  },
  {
    "epoch": 192,
    "batch_idx": 192,
    "loss": 0.38604575395584106,
    "learning_rate": 0.0009774322723733217,
    "timestep_mean": 430.75,
    "epsilon_norm": 3.5888006687164307,
    "prediction_norm": 2.214122772216797,
    "timestamp": "2025-07-14T12:33:40.187576"
  },
  {
    "epoch": 193,
    "batch_idx": 0,
    "loss": 0.37266045808792114,
    "learning_rate": 0.0009771983875483015,
    "timestep_mean": 458.40625,
    "epsilon_norm": 8.086339950561523,
    "prediction_norm": 6.050333023071289,
    "timestamp": "2025-07-14T12:33:40.192759"
  },
  {
    "epoch": 193,
    "batch_idx": 32,
    "loss": 0.36594557762145996,
    "learning_rate": 0.0009771983875483015,
    "timestep_mean": 450.125,
    "epsilon_norm": 7.09895133972168,
    "prediction_norm": 5.17276668548584,
    "timestamp": "2025-07-14T12:33:40.197964"
  },
  {
    "epoch": 193,
    "batch_idx": 64,
    "loss": 0.4682411253452301,
    "learning_rate": 0.0009771983875483015,
    "timestep_mean": 436.84375,
    "epsilon_norm": 7.271365642547607,
    "prediction_norm": 4.736492156982422,
    "timestamp": "2025-07-14T12:33:40.203197"
  },
  {
    "epoch": 193,
    "batch_idx": 96,
    "loss": 0.3066655397415161,
    "learning_rate": 0.0009771983875483015,
    "timestep_mean": 537.125,
    "epsilon_norm": 8.621427536010742,
    "prediction_norm": 7.31430196762085,
    "timestamp": "2025-07-14T12:33:40.208447"
  },
  {
    "epoch": 193,
    "batch_idx": 128,
    "loss": 0.45955032110214233,
    "learning_rate": 0.0009771983875483015,
    "timestep_mean": 473.15625,
    "epsilon_norm": 7.195794582366943,
    "prediction_norm": 4.817103385925293,
    "timestamp": "2025-07-14T12:33:40.213643"
  },
  {
    "epoch": 193,
    "batch_idx": 160,
    "loss": 0.32264572381973267,
    "learning_rate": 0.0009771983875483015,
    "timestep_mean": 515.1875,
    "epsilon_norm": 7.583091735839844,
    "prediction_norm": 5.937277793884277,
    "timestamp": "2025-07-14T12:33:40.218893"
  },
  {
    "epoch": 193,
    "batch_idx": 192,
    "loss": 0.22272713482379913,
    "learning_rate": 0.0009771983875483015,
    "timestep_mean": 621.0,
    "epsilon_norm": 4.703578472137451,
    "prediction_norm": 3.950286388397217,
    "timestamp": "2025-07-14T12:33:40.224042"
  },
  {
    "epoch": 194,
    "batch_idx": 0,
    "loss": 0.31575411558151245,
    "learning_rate": 0.000976963325283697,
    "timestep_mean": 573.875,
    "epsilon_norm": 7.726300239562988,
    "prediction_norm": 6.721187591552734,
    "timestamp": "2025-07-14T12:33:40.229331"
  },
  {
    "epoch": 194,
    "batch_idx": 32,
    "loss": 0.3606981039047241,
    "learning_rate": 0.000976963325283697,
    "timestep_mean": 422.4375,
    "epsilon_norm": 6.738026142120361,
    "prediction_norm": 5.5494489669799805,
    "timestamp": "2025-07-14T12:33:40.234469"
  },
  {
    "epoch": 194,
    "batch_idx": 64,
    "loss": 0.27156761288642883,
    "learning_rate": 0.000976963325283697,
    "timestep_mean": 603.5,
    "epsilon_norm": 7.753468036651611,
    "prediction_norm": 7.353527069091797,
    "timestamp": "2025-07-14T12:33:40.239381"
  },
  {
    "epoch": 194,
    "batch_idx": 96,
    "loss": 0.4479724168777466,
    "learning_rate": 0.000976963325283697,
    "timestep_mean": 430.6875,
    "epsilon_norm": 8.357159614562988,
    "prediction_norm": 7.1453166007995605,
    "timestamp": "2025-07-14T12:33:40.244359"
  },
  {
    "epoch": 194,
    "batch_idx": 128,
    "loss": 0.4141591191291809,
    "learning_rate": 0.000976963325283697,
    "timestep_mean": 545.75,
    "epsilon_norm": 6.834050178527832,
    "prediction_norm": 4.384133815765381,
    "timestamp": "2025-07-14T12:33:40.249593"
  },
  {
    "epoch": 194,
    "batch_idx": 160,
    "loss": 0.4529650807380676,
    "learning_rate": 0.000976963325283697,
    "timestep_mean": 455.46875,
    "epsilon_norm": 8.019445419311523,
    "prediction_norm": 6.370580673217773,
    "timestamp": "2025-07-14T12:33:40.254898"
  },
  {
    "epoch": 194,
    "batch_idx": 192,
    "loss": 0.3904588222503662,
    "learning_rate": 0.000976963325283697,
    "timestep_mean": 381.25,
    "epsilon_norm": 4.488091945648193,
    "prediction_norm": 2.739905595779419,
    "timestamp": "2025-07-14T12:33:40.260229"
  },
  {
    "epoch": 195,
    "batch_idx": 0,
    "loss": 0.4125570058822632,
    "learning_rate": 0.000976727086159501,
    "timestep_mean": 490.3125,
    "epsilon_norm": 8.005284309387207,
    "prediction_norm": 6.297941207885742,
    "timestamp": "2025-07-14T12:33:40.265482"
  },
  {
    "epoch": 195,
    "batch_idx": 32,
    "loss": 0.43467557430267334,
    "learning_rate": 0.000976727086159501,
    "timestep_mean": 481.59375,
    "epsilon_norm": 8.180755615234375,
    "prediction_norm": 5.833125591278076,
    "timestamp": "2025-07-14T12:33:40.270847"
  },
  {
    "epoch": 195,
    "batch_idx": 64,
    "loss": 0.3375810980796814,
    "learning_rate": 0.000976727086159501,
    "timestep_mean": 505.46875,
    "epsilon_norm": 6.999967098236084,
    "prediction_norm": 6.317732810974121,
    "timestamp": "2025-07-14T12:33:40.275994"
  },
  {
    "epoch": 195,
    "batch_idx": 96,
    "loss": 0.2194090634584427,
    "learning_rate": 0.000976727086159501,
    "timestep_mean": 512.75,
    "epsilon_norm": 7.603115558624268,
    "prediction_norm": 7.0420050621032715,
    "timestamp": "2025-07-14T12:33:40.281179"
  },
  {
    "epoch": 195,
    "batch_idx": 128,
    "loss": 0.263361394405365,
    "learning_rate": 0.000976727086159501,
    "timestep_mean": 552.03125,
    "epsilon_norm": 7.464987754821777,
    "prediction_norm": 6.023630142211914,
    "timestamp": "2025-07-14T12:33:40.286398"
  },
  {
    "epoch": 195,
    "batch_idx": 160,
    "loss": 0.14292316138744354,
    "learning_rate": 0.000976727086159501,
    "timestep_mean": 579.125,
    "epsilon_norm": 7.363768577575684,
    "prediction_norm": 6.722339153289795,
    "timestamp": "2025-07-14T12:33:40.291606"
  },
  {
    "epoch": 195,
    "batch_idx": 192,
    "loss": 0.32985720038414,
    "learning_rate": 0.000976727086159501,
    "timestep_mean": 510.875,
    "epsilon_norm": 3.410947561264038,
    "prediction_norm": 2.0231854915618896,
    "timestamp": "2025-07-14T12:33:40.296736"
  },
  {
    "epoch": 196,
    "batch_idx": 0,
    "loss": 0.3635089695453644,
    "learning_rate": 0.0009764896707586098,
    "timestep_mean": 502.75,
    "epsilon_norm": 8.754558563232422,
    "prediction_norm": 7.5965070724487305,
    "timestamp": "2025-07-14T12:33:40.301899"
  },
  {
    "epoch": 196,
    "batch_idx": 32,
    "loss": 0.4997653663158417,
    "learning_rate": 0.0009764896707586098,
    "timestep_mean": 412.15625,
    "epsilon_norm": 7.981117248535156,
    "prediction_norm": 5.993743419647217,
    "timestamp": "2025-07-14T12:33:40.307139"
  },
  {
    "epoch": 196,
    "batch_idx": 64,
    "loss": 0.3289697766304016,
    "learning_rate": 0.0009764896707586098,
    "timestep_mean": 437.75,
    "epsilon_norm": 7.699695110321045,
    "prediction_norm": 5.878843784332275,
    "timestamp": "2025-07-14T12:33:40.312344"
  },
  {
    "epoch": 196,
    "batch_idx": 96,
    "loss": 0.16873037815093994,
    "learning_rate": 0.0009764896707586098,
    "timestep_mean": 537.5625,
    "epsilon_norm": 7.732607841491699,
    "prediction_norm": 7.149656295776367,
    "timestamp": "2025-07-14T12:33:40.317522"
  },
  {
    "epoch": 196,
    "batch_idx": 128,
    "loss": 0.22526678442955017,
    "learning_rate": 0.0009764896707586098,
    "timestep_mean": 568.40625,
    "epsilon_norm": 8.62100601196289,
    "prediction_norm": 7.963403224945068,
    "timestamp": "2025-07-14T12:33:40.322697"
  },
  {
    "epoch": 196,
    "batch_idx": 160,
    "loss": 0.3315175175666809,
    "learning_rate": 0.0009764896707586098,
    "timestep_mean": 524.625,
    "epsilon_norm": 8.22562313079834,
    "prediction_norm": 6.303012371063232,
    "timestamp": "2025-07-14T12:33:40.327789"
  },
  {
    "epoch": 196,
    "batch_idx": 192,
    "loss": 0.7251247763633728,
    "learning_rate": 0.0009764896707586098,
    "timestep_mean": 322.625,
    "epsilon_norm": 5.499290943145752,
    "prediction_norm": 3.7972118854522705,
    "timestamp": "2025-07-14T12:33:40.332967"
  },
  {
    "epoch": 197,
    "batch_idx": 0,
    "loss": 0.5419880747795105,
    "learning_rate": 0.0009762510796668223,
    "timestep_mean": 424.875,
    "epsilon_norm": 8.691728591918945,
    "prediction_norm": 5.607192516326904,
    "timestamp": "2025-07-14T12:33:40.338178"
  },
  {
    "epoch": 197,
    "batch_idx": 32,
    "loss": 0.2498629242181778,
    "learning_rate": 0.0009762510796668223,
    "timestep_mean": 579.09375,
    "epsilon_norm": 7.96385383605957,
    "prediction_norm": 6.954916477203369,
    "timestamp": "2025-07-14T12:33:40.343323"
  },
  {
    "epoch": 197,
    "batch_idx": 64,
    "loss": 0.2952505052089691,
    "learning_rate": 0.0009762510796668223,
    "timestep_mean": 529.625,
    "epsilon_norm": 8.989331245422363,
    "prediction_norm": 8.302990913391113,
    "timestamp": "2025-07-14T12:33:40.348571"
  },
  {
    "epoch": 197,
    "batch_idx": 96,
    "loss": 0.34140342473983765,
    "learning_rate": 0.0009762510796668223,
    "timestep_mean": 516.90625,
    "epsilon_norm": 8.42202091217041,
    "prediction_norm": 7.189132213592529,
    "timestamp": "2025-07-14T12:33:40.353768"
  },
  {
    "epoch": 197,
    "batch_idx": 128,
    "loss": 0.22438129782676697,
    "learning_rate": 0.0009762510796668223,
    "timestep_mean": 587.4375,
    "epsilon_norm": 7.702288627624512,
    "prediction_norm": 7.85145902633667,
    "timestamp": "2025-07-14T12:33:40.359021"
  },
  {
    "epoch": 197,
    "batch_idx": 160,
    "loss": 0.28907668590545654,
    "learning_rate": 0.0009762510796668223,
    "timestep_mean": 458.96875,
    "epsilon_norm": 7.474889755249023,
    "prediction_norm": 6.6414666175842285,
    "timestamp": "2025-07-14T12:33:40.364227"
  },
  {
    "epoch": 197,
    "batch_idx": 192,
    "loss": 0.11921025067567825,
    "learning_rate": 0.0009762510796668223,
    "timestep_mean": 617.75,
    "epsilon_norm": 3.108381986618042,
    "prediction_norm": 2.879371404647827,
    "timestamp": "2025-07-14T12:33:40.369351"
  },
  {
    "epoch": 198,
    "batch_idx": 0,
    "loss": 0.2650718688964844,
    "learning_rate": 0.0009760113134728387,
    "timestep_mean": 540.5,
    "epsilon_norm": 7.199930667877197,
    "prediction_norm": 5.868649482727051,
    "timestamp": "2025-07-14T12:33:40.374580"
  },
  {
    "epoch": 198,
    "batch_idx": 32,
    "loss": 0.2643088400363922,
    "learning_rate": 0.0009760113134728387,
    "timestep_mean": 549.9375,
    "epsilon_norm": 7.785155773162842,
    "prediction_norm": 7.015934944152832,
    "timestamp": "2025-07-14T12:33:40.379845"
  },
  {
    "epoch": 198,
    "batch_idx": 64,
    "loss": 0.22965827584266663,
    "learning_rate": 0.0009760113134728387,
    "timestep_mean": 517.3125,
    "epsilon_norm": 7.59014368057251,
    "prediction_norm": 6.029500484466553,
    "timestamp": "2025-07-14T12:33:40.385008"
  },
  {
    "epoch": 198,
    "batch_idx": 96,
    "loss": 0.2944478690624237,
    "learning_rate": 0.0009760113134728387,
    "timestep_mean": 520.96875,
    "epsilon_norm": 7.915103435516357,
    "prediction_norm": 7.060166835784912,
    "timestamp": "2025-07-14T12:33:40.390212"
  },
  {
    "epoch": 198,
    "batch_idx": 128,
    "loss": 0.3121395409107208,
    "learning_rate": 0.0009760113134728387,
    "timestep_mean": 425.28125,
    "epsilon_norm": 8.025052070617676,
    "prediction_norm": 7.538192272186279,
    "timestamp": "2025-07-14T12:33:40.395402"
  },
  {
    "epoch": 198,
    "batch_idx": 160,
    "loss": 0.3826000392436981,
    "learning_rate": 0.0009760113134728387,
    "timestep_mean": 544.65625,
    "epsilon_norm": 8.550345420837402,
    "prediction_norm": 6.701201438903809,
    "timestamp": "2025-07-14T12:33:40.400677"
  },
  {
    "epoch": 198,
    "batch_idx": 192,
    "loss": 0.6066112518310547,
    "learning_rate": 0.0009760113134728387,
    "timestep_mean": 484.125,
    "epsilon_norm": 3.683579444885254,
    "prediction_norm": 2.3530099391937256,
    "timestamp": "2025-07-14T12:33:40.405610"
  },
  {
    "epoch": 199,
    "batch_idx": 0,
    "loss": 0.3904520869255066,
    "learning_rate": 0.0009757703727682577,
    "timestep_mean": 512.25,
    "epsilon_norm": 8.39681339263916,
    "prediction_norm": 6.367499828338623,
    "timestamp": "2025-07-14T12:33:40.410483"
  },
  {
    "epoch": 199,
    "batch_idx": 32,
    "loss": 0.4187729060649872,
    "learning_rate": 0.0009757703727682577,
    "timestep_mean": 345.84375,
    "epsilon_norm": 7.742539882659912,
    "prediction_norm": 5.1829376220703125,
    "timestamp": "2025-07-14T12:33:40.415774"
  },
  {
    "epoch": 199,
    "batch_idx": 64,
    "loss": 0.5239849090576172,
    "learning_rate": 0.0009757703727682577,
    "timestep_mean": 566.125,
    "epsilon_norm": 8.58084487915039,
    "prediction_norm": 6.370896339416504,
    "timestamp": "2025-07-14T12:33:40.421191"
  },
  {
    "epoch": 199,
    "batch_idx": 96,
    "loss": 0.24129197001457214,
    "learning_rate": 0.0009757703727682577,
    "timestep_mean": 574.5,
    "epsilon_norm": 7.606634616851807,
    "prediction_norm": 6.637239933013916,
    "timestamp": "2025-07-14T12:33:40.426263"
  },
  {
    "epoch": 199,
    "batch_idx": 128,
    "loss": 0.4880250096321106,
    "learning_rate": 0.0009757703727682577,
    "timestep_mean": 466.1875,
    "epsilon_norm": 8.725495338439941,
    "prediction_norm": 6.029301166534424,
    "timestamp": "2025-07-14T12:33:40.431508"
  },
  {
    "epoch": 199,
    "batch_idx": 160,
    "loss": 0.21568486094474792,
    "learning_rate": 0.0009757703727682577,
    "timestep_mean": 622.46875,
    "epsilon_norm": 6.982759952545166,
    "prediction_norm": 5.741062641143799,
    "timestamp": "2025-07-14T12:33:40.436875"
  },
  {
    "epoch": 199,
    "batch_idx": 192,
    "loss": 0.1413477510213852,
    "learning_rate": 0.0009757703727682577,
    "timestep_mean": 468.75,
    "epsilon_norm": 3.5440971851348877,
    "prediction_norm": 3.2977612018585205,
    "timestamp": "2025-07-14T12:33:40.442298"
  },
  {
    "epoch": 200,
    "batch_idx": 0,
    "loss": 0.3176347017288208,
    "learning_rate": 0.0009755282581475771,
    "timestep_mean": 503.65625,
    "epsilon_norm": 7.456384658813477,
    "prediction_norm": 5.855165481567383,
    "timestamp": "2025-07-14T12:33:40.447566"
  },
  {
    "epoch": 200,
    "batch_idx": 32,
    "loss": 0.2844318449497223,
    "learning_rate": 0.0009755282581475771,
    "timestep_mean": 494.375,
    "epsilon_norm": 7.048741340637207,
    "prediction_norm": 5.909933567047119,
    "timestamp": "2025-07-14T12:33:40.452797"
  },
  {
    "epoch": 200,
    "batch_idx": 64,
    "loss": 0.29738283157348633,
    "learning_rate": 0.0009755282581475771,
    "timestep_mean": 515.15625,
    "epsilon_norm": 7.387776851654053,
    "prediction_norm": 5.77857780456543,
    "timestamp": "2025-07-14T12:33:40.458014"
  },
  {
    "epoch": 200,
    "batch_idx": 96,
    "loss": 0.31898361444473267,
    "learning_rate": 0.0009755282581475771,
    "timestep_mean": 415.75,
    "epsilon_norm": 8.019205093383789,
    "prediction_norm": 7.389442443847656,
    "timestamp": "2025-07-14T12:33:40.463178"
  },
  {
    "epoch": 200,
    "batch_idx": 128,
    "loss": 0.3064457178115845,
    "learning_rate": 0.0009755282581475771,
    "timestep_mean": 532.96875,
    "epsilon_norm": 7.910617828369141,
    "prediction_norm": 6.9712138175964355,
    "timestamp": "2025-07-14T12:33:40.468397"
  },
  {
    "epoch": 200,
    "batch_idx": 160,
    "loss": 0.2510402500629425,
    "learning_rate": 0.0009755282581475771,
    "timestep_mean": 539.0625,
    "epsilon_norm": 7.228522300720215,
    "prediction_norm": 5.879055500030518,
    "timestamp": "2025-07-14T12:33:40.473631"
  },
  {
    "epoch": 200,
    "batch_idx": 192,
    "loss": 0.19691134989261627,
    "learning_rate": 0.0009755282581475771,
    "timestep_mean": 637.875,
    "epsilon_norm": 2.8641791343688965,
    "prediction_norm": 2.527850866317749,
    "timestamp": "2025-07-14T12:33:40.478796"
  },
  {
    "epoch": 201,
    "batch_idx": 0,
    "loss": 0.26905548572540283,
    "learning_rate": 0.0009752849702081904,
    "timestep_mean": 520.125,
    "epsilon_norm": 7.201982498168945,
    "prediction_norm": 5.413697242736816,
    "timestamp": "2025-07-14T12:33:40.484008"
  },
  {
    "epoch": 201,
    "batch_idx": 32,
    "loss": 0.19408053159713745,
    "learning_rate": 0.0009752849702081904,
    "timestep_mean": 539.21875,
    "epsilon_norm": 7.976449012756348,
    "prediction_norm": 6.29250431060791,
    "timestamp": "2025-07-14T12:33:40.489177"
  },
  {
    "epoch": 201,
    "batch_idx": 64,
    "loss": 0.3003019094467163,
    "learning_rate": 0.0009752849702081904,
    "timestep_mean": 462.3125,
    "epsilon_norm": 6.572289943695068,
    "prediction_norm": 4.508690357208252,
    "timestamp": "2025-07-14T12:33:40.494331"
  },
  {
    "epoch": 201,
    "batch_idx": 96,
    "loss": 0.3613472282886505,
    "learning_rate": 0.0009752849702081904,
    "timestep_mean": 511.4375,
    "epsilon_norm": 8.310772895812988,
    "prediction_norm": 6.809742450714111,
    "timestamp": "2025-07-14T12:33:40.499556"
  },
  {
    "epoch": 201,
    "batch_idx": 128,
    "loss": 0.22718290984630585,
    "learning_rate": 0.0009752849702081904,
    "timestep_mean": 469.375,
    "epsilon_norm": 7.8921380043029785,
    "prediction_norm": 6.471572399139404,
    "timestamp": "2025-07-14T12:33:40.509487"
  },
  {
    "epoch": 201,
    "batch_idx": 160,
    "loss": 0.27628132700920105,
    "learning_rate": 0.0009752849702081904,
    "timestep_mean": 508.84375,
    "epsilon_norm": 8.199832916259766,
    "prediction_norm": 6.560194969177246,
    "timestamp": "2025-07-14T12:33:40.514803"
  },
  {
    "epoch": 201,
    "batch_idx": 192,
    "loss": 0.6438822746276855,
    "learning_rate": 0.0009752849702081904,
    "timestep_mean": 586.875,
    "epsilon_norm": 4.204556465148926,
    "prediction_norm": 2.489504814147949,
    "timestamp": "2025-07-14T12:33:40.520178"
  },
  {
    "epoch": 202,
    "batch_idx": 0,
    "loss": 0.2671262323856354,
    "learning_rate": 0.0009750405095503861,
    "timestep_mean": 551.5,
    "epsilon_norm": 7.997896194458008,
    "prediction_norm": 6.8814802169799805,
    "timestamp": "2025-07-14T12:33:40.525493"
  },
  {
    "epoch": 202,
    "batch_idx": 32,
    "loss": 0.26599931716918945,
    "learning_rate": 0.0009750405095503861,
    "timestep_mean": 490.53125,
    "epsilon_norm": 6.983788013458252,
    "prediction_norm": 5.6193437576293945,
    "timestamp": "2025-07-14T12:33:40.530589"
  },
  {
    "epoch": 202,
    "batch_idx": 64,
    "loss": 0.3327549695968628,
    "learning_rate": 0.0009750405095503861,
    "timestep_mean": 503.0625,
    "epsilon_norm": 8.117570877075195,
    "prediction_norm": 7.145700454711914,
    "timestamp": "2025-07-14T12:33:40.535806"
  },
  {
    "epoch": 202,
    "batch_idx": 96,
    "loss": 0.1703263819217682,
    "learning_rate": 0.0009750405095503861,
    "timestep_mean": 581.40625,
    "epsilon_norm": 7.488602161407471,
    "prediction_norm": 6.780303955078125,
    "timestamp": "2025-07-14T12:33:40.541028"
  },
  {
    "epoch": 202,
    "batch_idx": 128,
    "loss": 0.2342589646577835,
    "learning_rate": 0.0009750405095503861,
    "timestep_mean": 568.71875,
    "epsilon_norm": 9.096485137939453,
    "prediction_norm": 9.725821495056152,
    "timestamp": "2025-07-14T12:33:40.546298"
  },
  {
    "epoch": 202,
    "batch_idx": 160,
    "loss": 0.4132508337497711,
    "learning_rate": 0.0009750405095503861,
    "timestep_mean": 409.9375,
    "epsilon_norm": 7.27598762512207,
    "prediction_norm": 6.355322360992432,
    "timestamp": "2025-07-14T12:33:40.551490"
  },
  {
    "epoch": 202,
    "batch_idx": 192,
    "loss": 1.0297884941101074,
    "learning_rate": 0.0009750405095503861,
    "timestep_mean": 358.625,
    "epsilon_norm": 4.239841461181641,
    "prediction_norm": 1.9965757131576538,
    "timestamp": "2025-07-14T12:33:40.556709"
  },
  {
    "epoch": 203,
    "batch_idx": 0,
    "loss": 0.3361584544181824,
    "learning_rate": 0.0009747948767773472,
    "timestep_mean": 570.40625,
    "epsilon_norm": 8.256802558898926,
    "prediction_norm": 6.98629903793335,
    "timestamp": "2025-07-14T12:33:40.562024"
  },
  {
    "epoch": 203,
    "batch_idx": 32,
    "loss": 0.32240724563598633,
    "learning_rate": 0.0009747948767773472,
    "timestep_mean": 566.4375,
    "epsilon_norm": 8.624112129211426,
    "prediction_norm": 6.930532932281494,
    "timestamp": "2025-07-14T12:33:40.567306"
  },
  {
    "epoch": 203,
    "batch_idx": 64,
    "loss": 0.3622993230819702,
    "learning_rate": 0.0009747948767773472,
    "timestep_mean": 453.1875,
    "epsilon_norm": 8.925848007202148,
    "prediction_norm": 6.962618827819824,
    "timestamp": "2025-07-14T12:33:40.572436"
  },
  {
    "epoch": 203,
    "batch_idx": 96,
    "loss": 0.34879541397094727,
    "learning_rate": 0.0009747948767773472,
    "timestep_mean": 498.90625,
    "epsilon_norm": 8.561050415039062,
    "prediction_norm": 7.0051350593566895,
    "timestamp": "2025-07-14T12:33:40.577634"
  },
  {
    "epoch": 203,
    "batch_idx": 128,
    "loss": 0.21323800086975098,
    "learning_rate": 0.0009747948767773472,
    "timestep_mean": 524.96875,
    "epsilon_norm": 6.92987060546875,
    "prediction_norm": 5.86483097076416,
    "timestamp": "2025-07-14T12:33:40.582492"
  },
  {
    "epoch": 203,
    "batch_idx": 160,
    "loss": 0.5143631100654602,
    "learning_rate": 0.0009747948767773472,
    "timestep_mean": 349.71875,
    "epsilon_norm": 8.320235252380371,
    "prediction_norm": 5.754872798919678,
    "timestamp": "2025-07-14T12:33:40.587442"
  },
  {
    "epoch": 203,
    "batch_idx": 192,
    "loss": 0.11555502563714981,
    "learning_rate": 0.0009747948767773472,
    "timestep_mean": 658.875,
    "epsilon_norm": 3.1632721424102783,
    "prediction_norm": 2.3563404083251953,
    "timestamp": "2025-07-14T12:33:40.592333"
  },
  {
    "epoch": 204,
    "batch_idx": 0,
    "loss": 0.4409986138343811,
    "learning_rate": 0.0009745480724951476,
    "timestep_mean": 455.96875,
    "epsilon_norm": 8.03881549835205,
    "prediction_norm": 5.783751010894775,
    "timestamp": "2025-07-14T12:33:40.597234"
  },
  {
    "epoch": 204,
    "batch_idx": 32,
    "loss": 0.34592634439468384,
    "learning_rate": 0.0009745480724951476,
    "timestep_mean": 507.28125,
    "epsilon_norm": 8.356675148010254,
    "prediction_norm": 6.667220592498779,
    "timestamp": "2025-07-14T12:33:40.602127"
  },
  {
    "epoch": 204,
    "batch_idx": 64,
    "loss": 0.4389164447784424,
    "learning_rate": 0.0009745480724951476,
    "timestep_mean": 467.5625,
    "epsilon_norm": 8.526164054870605,
    "prediction_norm": 7.639910697937012,
    "timestamp": "2025-07-14T12:33:40.606989"
  },
  {
    "epoch": 204,
    "batch_idx": 96,
    "loss": 0.29480215907096863,
    "learning_rate": 0.0009745480724951476,
    "timestep_mean": 472.25,
    "epsilon_norm": 7.544063568115234,
    "prediction_norm": 6.126587867736816,
    "timestamp": "2025-07-14T12:33:40.611865"
  },
  {
    "epoch": 204,
    "batch_idx": 128,
    "loss": 0.6180779337882996,
    "learning_rate": 0.0009745480724951476,
    "timestep_mean": 493.75,
    "epsilon_norm": 9.495746612548828,
    "prediction_norm": 6.884762763977051,
    "timestamp": "2025-07-14T12:33:40.616839"
  },
  {
    "epoch": 204,
    "batch_idx": 160,
    "loss": 0.20372387766838074,
    "learning_rate": 0.0009745480724951476,
    "timestep_mean": 520.53125,
    "epsilon_norm": 6.4571146965026855,
    "prediction_norm": 5.5437140464782715,
    "timestamp": "2025-07-14T12:33:40.621673"
  },
  {
    "epoch": 204,
    "batch_idx": 192,
    "loss": 0.17455849051475525,
    "learning_rate": 0.0009745480724951476,
    "timestep_mean": 535.0,
    "epsilon_norm": 4.051971912384033,
    "prediction_norm": 3.8233134746551514,
    "timestamp": "2025-07-14T12:33:40.626571"
  },
  {
    "epoch": 205,
    "batch_idx": 0,
    "loss": 0.30295851826667786,
    "learning_rate": 0.0009743000973127526,
    "timestep_mean": 574.375,
    "epsilon_norm": 7.353100299835205,
    "prediction_norm": 6.129764556884766,
    "timestamp": "2025-07-14T12:33:40.631424"
  },
  {
    "epoch": 205,
    "batch_idx": 32,
    "loss": 0.5834380984306335,
    "learning_rate": 0.0009743000973127526,
    "timestep_mean": 436.625,
    "epsilon_norm": 8.154417037963867,
    "prediction_norm": 5.661209583282471,
    "timestamp": "2025-07-14T12:33:40.636279"
  },
  {
    "epoch": 205,
    "batch_idx": 64,
    "loss": 0.18649140000343323,
    "learning_rate": 0.0009743000973127526,
    "timestep_mean": 566.25,
    "epsilon_norm": 7.197351932525635,
    "prediction_norm": 5.690145015716553,
    "timestamp": "2025-07-14T12:33:40.641153"
  },
  {
    "epoch": 205,
    "batch_idx": 96,
    "loss": 0.3712712228298187,
    "learning_rate": 0.0009743000973127526,
    "timestep_mean": 478.9375,
    "epsilon_norm": 8.514588356018066,
    "prediction_norm": 6.716677665710449,
    "timestamp": "2025-07-14T12:33:40.646079"
  },
  {
    "epoch": 205,
    "batch_idx": 128,
    "loss": 0.3854350745677948,
    "learning_rate": 0.0009743000973127526,
    "timestep_mean": 447.6875,
    "epsilon_norm": 7.77194881439209,
    "prediction_norm": 5.990550994873047,
    "timestamp": "2025-07-14T12:33:40.650994"
  },
  {
    "epoch": 205,
    "batch_idx": 160,
    "loss": 0.3041214942932129,
    "learning_rate": 0.0009743000973127526,
    "timestep_mean": 479.75,
    "epsilon_norm": 7.310948848724365,
    "prediction_norm": 6.0244550704956055,
    "timestamp": "2025-07-14T12:33:40.655835"
  },
  {
    "epoch": 205,
    "batch_idx": 192,
    "loss": 0.33459556102752686,
    "learning_rate": 0.0009743000973127526,
    "timestep_mean": 388.5,
    "epsilon_norm": 2.73220157623291,
    "prediction_norm": 2.292731523513794,
    "timestamp": "2025-07-14T12:33:40.660623"
  },
  {
    "epoch": 206,
    "batch_idx": 0,
    "loss": 0.20726090669631958,
    "learning_rate": 0.0009740509518420163,
    "timestep_mean": 467.28125,
    "epsilon_norm": 6.953883171081543,
    "prediction_norm": 6.239810466766357,
    "timestamp": "2025-07-14T12:33:40.665573"
  },
  {
    "epoch": 206,
    "batch_idx": 32,
    "loss": 0.2826575040817261,
    "learning_rate": 0.0009740509518420163,
    "timestep_mean": 509.96875,
    "epsilon_norm": 7.65444278717041,
    "prediction_norm": 6.751986503601074,
    "timestamp": "2025-07-14T12:33:40.670372"
  },
  {
    "epoch": 206,
    "batch_idx": 64,
    "loss": 0.5296647548675537,
    "learning_rate": 0.0009740509518420163,
    "timestep_mean": 451.625,
    "epsilon_norm": 8.481709480285645,
    "prediction_norm": 6.759113311767578,
    "timestamp": "2025-07-14T12:33:40.675282"
  },
  {
    "epoch": 206,
    "batch_idx": 96,
    "loss": 0.11347626149654388,
    "learning_rate": 0.0009740509518420163,
    "timestep_mean": 529.78125,
    "epsilon_norm": 7.1584014892578125,
    "prediction_norm": 6.8286519050598145,
    "timestamp": "2025-07-14T12:33:40.680064"
  },
  {
    "epoch": 206,
    "batch_idx": 128,
    "loss": 0.24838203191757202,
    "learning_rate": 0.0009740509518420163,
    "timestep_mean": 513.1875,
    "epsilon_norm": 7.068186283111572,
    "prediction_norm": 6.316379547119141,
    "timestamp": "2025-07-14T12:33:40.684910"
  },
  {
    "epoch": 206,
    "batch_idx": 160,
    "loss": 0.2086312174797058,
    "learning_rate": 0.0009740509518420163,
    "timestep_mean": 531.5,
    "epsilon_norm": 7.795516014099121,
    "prediction_norm": 6.897351264953613,
    "timestamp": "2025-07-14T12:33:40.689767"
  },
  {
    "epoch": 206,
    "batch_idx": 192,
    "loss": 0.3234865665435791,
    "learning_rate": 0.0009740509518420163,
    "timestep_mean": 416.5,
    "epsilon_norm": 3.0426979064941406,
    "prediction_norm": 1.6448332071304321,
    "timestamp": "2025-07-14T12:33:40.694600"
  },
  {
    "epoch": 207,
    "batch_idx": 0,
    "loss": 0.3871687650680542,
    "learning_rate": 0.0009738006366976804,
    "timestep_mean": 523.96875,
    "epsilon_norm": 8.825470924377441,
    "prediction_norm": 7.281988620758057,
    "timestamp": "2025-07-14T12:33:40.699492"
  },
  {
    "epoch": 207,
    "batch_idx": 32,
    "loss": 0.23349758982658386,
    "learning_rate": 0.0009738006366976804,
    "timestep_mean": 535.8125,
    "epsilon_norm": 7.905970096588135,
    "prediction_norm": 6.356802463531494,
    "timestamp": "2025-07-14T12:33:40.704397"
  },
  {
    "epoch": 207,
    "batch_idx": 64,
    "loss": 0.3310668170452118,
    "learning_rate": 0.0009738006366976804,
    "timestep_mean": 555.84375,
    "epsilon_norm": 8.9862060546875,
    "prediction_norm": 6.982590198516846,
    "timestamp": "2025-07-14T12:33:40.709260"
  },
  {
    "epoch": 207,
    "batch_idx": 96,
    "loss": 0.44552701711654663,
    "learning_rate": 0.0009738006366976804,
    "timestep_mean": 543.625,
    "epsilon_norm": 9.545424461364746,
    "prediction_norm": 7.706325531005859,
    "timestamp": "2025-07-14T12:33:40.714094"
  },
  {
    "epoch": 207,
    "batch_idx": 128,
    "loss": 0.2765967845916748,
    "learning_rate": 0.0009738006366976804,
    "timestep_mean": 476.65625,
    "epsilon_norm": 7.632850646972656,
    "prediction_norm": 5.971622943878174,
    "timestamp": "2025-07-14T12:33:40.718941"
  },
  {
    "epoch": 207,
    "batch_idx": 160,
    "loss": 0.29344066977500916,
    "learning_rate": 0.0009738006366976804,
    "timestep_mean": 477.21875,
    "epsilon_norm": 6.578460693359375,
    "prediction_norm": 5.143273830413818,
    "timestamp": "2025-07-14T12:33:40.723795"
  },
  {
    "epoch": 207,
    "batch_idx": 192,
    "loss": 0.6171592473983765,
    "learning_rate": 0.0009738006366976804,
    "timestep_mean": 319.625,
    "epsilon_norm": 4.767050266265869,
    "prediction_norm": 2.5605924129486084,
    "timestamp": "2025-07-14T12:33:40.728679"
  },
  {
    "epoch": 208,
    "batch_idx": 0,
    "loss": 0.35305723547935486,
    "learning_rate": 0.0009735491524973725,
    "timestep_mean": 526.78125,
    "epsilon_norm": 8.006266593933105,
    "prediction_norm": 6.662628173828125,
    "timestamp": "2025-07-14T12:33:40.733678"
  },
  {
    "epoch": 208,
    "batch_idx": 32,
    "loss": 0.33907029032707214,
    "learning_rate": 0.0009735491524973725,
    "timestep_mean": 472.90625,
    "epsilon_norm": 7.143332481384277,
    "prediction_norm": 5.248802185058594,
    "timestamp": "2025-07-14T12:33:40.738478"
  },
  {
    "epoch": 208,
    "batch_idx": 64,
    "loss": 0.25576716661453247,
    "learning_rate": 0.0009735491524973725,
    "timestep_mean": 464.46875,
    "epsilon_norm": 7.500953197479248,
    "prediction_norm": 6.290701866149902,
    "timestamp": "2025-07-14T12:33:40.743352"
  },
  {
    "epoch": 208,
    "batch_idx": 96,
    "loss": 0.2905774712562561,
    "learning_rate": 0.0009735491524973725,
    "timestep_mean": 467.71875,
    "epsilon_norm": 6.769448280334473,
    "prediction_norm": 5.555790901184082,
    "timestamp": "2025-07-14T12:33:40.748179"
  },
  {
    "epoch": 208,
    "batch_idx": 128,
    "loss": 0.3024430572986603,
    "learning_rate": 0.0009735491524973725,
    "timestep_mean": 523.71875,
    "epsilon_norm": 7.7215495109558105,
    "prediction_norm": 7.0143046379089355,
    "timestamp": "2025-07-14T12:33:40.753047"
  },
  {
    "epoch": 208,
    "batch_idx": 160,
    "loss": 0.43957874178886414,
    "learning_rate": 0.0009735491524973725,
    "timestep_mean": 490.8125,
    "epsilon_norm": 9.093117713928223,
    "prediction_norm": 7.617720603942871,
    "timestamp": "2025-07-14T12:33:40.757953"
  },
  {
    "epoch": 208,
    "batch_idx": 192,
    "loss": 0.3188101649284363,
    "learning_rate": 0.0009735491524973725,
    "timestep_mean": 543.0,
    "epsilon_norm": 3.309790849685669,
    "prediction_norm": 3.0142507553100586,
    "timestamp": "2025-07-14T12:33:40.762919"
  },
  {
    "epoch": 209,
    "batch_idx": 0,
    "loss": 0.3731388747692108,
    "learning_rate": 0.000973296499861605,
    "timestep_mean": 439.21875,
    "epsilon_norm": 7.029715061187744,
    "prediction_norm": 5.480068683624268,
    "timestamp": "2025-07-14T12:33:40.767787"
  },
  {
    "epoch": 209,
    "batch_idx": 32,
    "loss": 0.4657251834869385,
    "learning_rate": 0.000973296499861605,
    "timestep_mean": 408.0,
    "epsilon_norm": 8.997011184692383,
    "prediction_norm": 7.922183513641357,
    "timestamp": "2025-07-14T12:33:40.772624"
  },
  {
    "epoch": 209,
    "batch_idx": 64,
    "loss": 0.18015632033348083,
    "learning_rate": 0.000973296499861605,
    "timestep_mean": 545.15625,
    "epsilon_norm": 7.69529914855957,
    "prediction_norm": 6.929919719696045,
    "timestamp": "2025-07-14T12:33:40.777468"
  },
  {
    "epoch": 209,
    "batch_idx": 96,
    "loss": 0.3103601336479187,
    "learning_rate": 0.000973296499861605,
    "timestep_mean": 557.75,
    "epsilon_norm": 8.535611152648926,
    "prediction_norm": 7.09858512878418,
    "timestamp": "2025-07-14T12:33:40.782315"
  },
  {
    "epoch": 209,
    "batch_idx": 128,
    "loss": 0.24452275037765503,
    "learning_rate": 0.000973296499861605,
    "timestep_mean": 476.3125,
    "epsilon_norm": 7.023499965667725,
    "prediction_norm": 5.873510360717773,
    "timestamp": "2025-07-14T12:33:40.787265"
  },
  {
    "epoch": 209,
    "batch_idx": 160,
    "loss": 0.5457819700241089,
    "learning_rate": 0.000973296499861605,
    "timestep_mean": 486.8125,
    "epsilon_norm": 8.378409385681152,
    "prediction_norm": 5.727301597595215,
    "timestamp": "2025-07-14T12:33:40.792149"
  },
  {
    "epoch": 209,
    "batch_idx": 192,
    "loss": 0.2558711767196655,
    "learning_rate": 0.000973296499861605,
    "timestep_mean": 320.75,
    "epsilon_norm": 3.518608808517456,
    "prediction_norm": 3.1139023303985596,
    "timestamp": "2025-07-14T12:33:40.796952"
  },
  {
    "epoch": 210,
    "batch_idx": 0,
    "loss": 0.4370180368423462,
    "learning_rate": 0.0009730426794137731,
    "timestep_mean": 491.96875,
    "epsilon_norm": 7.625668048858643,
    "prediction_norm": 5.261023044586182,
    "timestamp": "2025-07-14T12:33:40.801839"
  },
  {
    "epoch": 210,
    "batch_idx": 32,
    "loss": 0.22065496444702148,
    "learning_rate": 0.0009730426794137731,
    "timestep_mean": 556.6875,
    "epsilon_norm": 7.130053997039795,
    "prediction_norm": 5.469358444213867,
    "timestamp": "2025-07-14T12:33:40.807015"
  },
  {
    "epoch": 210,
    "batch_idx": 64,
    "loss": 0.5593991875648499,
    "learning_rate": 0.0009730426794137731,
    "timestep_mean": 454.71875,
    "epsilon_norm": 9.239501953125,
    "prediction_norm": 7.067849159240723,
    "timestamp": "2025-07-14T12:33:40.812312"
  },
  {
    "epoch": 210,
    "batch_idx": 96,
    "loss": 0.348864883184433,
    "learning_rate": 0.0009730426794137731,
    "timestep_mean": 479.71875,
    "epsilon_norm": 8.534836769104004,
    "prediction_norm": 6.759067535400391,
    "timestamp": "2025-07-14T12:33:40.817430"
  },
  {
    "epoch": 210,
    "batch_idx": 128,
    "loss": 0.4151306748390198,
    "learning_rate": 0.0009730426794137731,
    "timestep_mean": 466.59375,
    "epsilon_norm": 7.530785083770752,
    "prediction_norm": 5.536104202270508,
    "timestamp": "2025-07-14T12:33:40.822640"
  },
  {
    "epoch": 210,
    "batch_idx": 160,
    "loss": 0.38517436385154724,
    "learning_rate": 0.0009730426794137731,
    "timestep_mean": 496.53125,
    "epsilon_norm": 9.017251014709473,
    "prediction_norm": 7.773351192474365,
    "timestamp": "2025-07-14T12:33:40.827838"
  },
  {
    "epoch": 210,
    "batch_idx": 192,
    "loss": 0.3501493036746979,
    "learning_rate": 0.0009730426794137731,
    "timestep_mean": 365.625,
    "epsilon_norm": 3.3896923065185547,
    "prediction_norm": 2.210296392440796,
    "timestamp": "2025-07-14T12:33:40.833020"
  },
  {
    "epoch": 211,
    "batch_idx": 0,
    "loss": 0.40092411637306213,
    "learning_rate": 0.0009727876917801535,
    "timestep_mean": 498.9375,
    "epsilon_norm": 7.543084144592285,
    "prediction_norm": 5.498741149902344,
    "timestamp": "2025-07-14T12:33:40.838178"
  },
  {
    "epoch": 211,
    "batch_idx": 32,
    "loss": 0.3063514828681946,
    "learning_rate": 0.0009727876917801535,
    "timestep_mean": 530.34375,
    "epsilon_norm": 8.903670310974121,
    "prediction_norm": 7.547702312469482,
    "timestamp": "2025-07-14T12:33:40.843269"
  },
  {
    "epoch": 211,
    "batch_idx": 64,
    "loss": 0.1781785786151886,
    "learning_rate": 0.0009727876917801535,
    "timestep_mean": 552.9375,
    "epsilon_norm": 6.221132278442383,
    "prediction_norm": 5.122361660003662,
    "timestamp": "2025-07-14T12:33:40.848411"
  },
  {
    "epoch": 211,
    "batch_idx": 96,
    "loss": 0.45471853017807007,
    "learning_rate": 0.0009727876917801535,
    "timestep_mean": 356.90625,
    "epsilon_norm": 7.575614929199219,
    "prediction_norm": 4.737189769744873,
    "timestamp": "2025-07-14T12:33:40.853562"
  },
  {
    "epoch": 211,
    "batch_idx": 128,
    "loss": 0.6006971597671509,
    "learning_rate": 0.0009727876917801535,
    "timestep_mean": 446.71875,
    "epsilon_norm": 8.828014373779297,
    "prediction_norm": 6.047687530517578,
    "timestamp": "2025-07-14T12:33:40.858799"
  },
  {
    "epoch": 211,
    "batch_idx": 160,
    "loss": 0.2392641007900238,
    "learning_rate": 0.0009727876917801535,
    "timestep_mean": 499.65625,
    "epsilon_norm": 8.075325965881348,
    "prediction_norm": 7.296314716339111,
    "timestamp": "2025-07-14T12:33:40.863756"
  },
  {
    "epoch": 211,
    "batch_idx": 192,
    "loss": 0.283770352602005,
    "learning_rate": 0.0009727876917801535,
    "timestep_mean": 528.75,
    "epsilon_norm": 4.062389373779297,
    "prediction_norm": 3.3433704376220703,
    "timestamp": "2025-07-14T12:33:40.868790"
  },
  {
    "epoch": 212,
    "batch_idx": 0,
    "loss": 0.479483425617218,
    "learning_rate": 0.0009725315375899028,
    "timestep_mean": 453.59375,
    "epsilon_norm": 9.083399772644043,
    "prediction_norm": 7.636087417602539,
    "timestamp": "2025-07-14T12:33:40.873996"
  },
  {
    "epoch": 212,
    "batch_idx": 32,
    "loss": 0.3185093402862549,
    "learning_rate": 0.0009725315375899028,
    "timestep_mean": 430.25,
    "epsilon_norm": 7.632769584655762,
    "prediction_norm": 6.740569591522217,
    "timestamp": "2025-07-14T12:33:40.879235"
  },
  {
    "epoch": 212,
    "batch_idx": 64,
    "loss": 0.3411821722984314,
    "learning_rate": 0.0009725315375899028,
    "timestep_mean": 545.53125,
    "epsilon_norm": 8.754539489746094,
    "prediction_norm": 7.9567108154296875,
    "timestamp": "2025-07-14T12:33:40.884436"
  },
  {
    "epoch": 212,
    "batch_idx": 96,
    "loss": 0.190697580575943,
    "learning_rate": 0.0009725315375899028,
    "timestep_mean": 616.46875,
    "epsilon_norm": 7.668626308441162,
    "prediction_norm": 6.864736080169678,
    "timestamp": "2025-07-14T12:33:40.889457"
  },
  {
    "epoch": 212,
    "batch_idx": 128,
    "loss": 0.3658325672149658,
    "learning_rate": 0.0009725315375899028,
    "timestep_mean": 399.4375,
    "epsilon_norm": 7.410592079162598,
    "prediction_norm": 5.611767292022705,
    "timestamp": "2025-07-14T12:33:40.894340"
  },
  {
    "epoch": 212,
    "batch_idx": 160,
    "loss": 0.33210429549217224,
    "learning_rate": 0.0009725315375899028,
    "timestep_mean": 487.875,
    "epsilon_norm": 8.336470603942871,
    "prediction_norm": 6.665705680847168,
    "timestamp": "2025-07-14T12:33:40.899475"
  },
  {
    "epoch": 212,
    "batch_idx": 192,
    "loss": 0.12329588085412979,
    "learning_rate": 0.0009725315375899028,
    "timestep_mean": 493.625,
    "epsilon_norm": 2.563483715057373,
    "prediction_norm": 2.4207351207733154,
    "timestamp": "2025-07-14T12:33:40.904663"
  },
  {
    "epoch": 213,
    "batch_idx": 0,
    "loss": 0.2526112496852875,
    "learning_rate": 0.0009722742174750563,
    "timestep_mean": 510.9375,
    "epsilon_norm": 6.986128330230713,
    "prediction_norm": 5.91839599609375,
    "timestamp": "2025-07-14T12:33:40.909813"
  },
  {
    "epoch": 213,
    "batch_idx": 32,
    "loss": 0.3148842751979828,
    "learning_rate": 0.0009722742174750563,
    "timestep_mean": 512.5,
    "epsilon_norm": 7.976715087890625,
    "prediction_norm": 6.226927280426025,
    "timestamp": "2025-07-14T12:33:40.915115"
  },
  {
    "epoch": 213,
    "batch_idx": 64,
    "loss": 0.24822255969047546,
    "learning_rate": 0.0009722742174750563,
    "timestep_mean": 499.21875,
    "epsilon_norm": 8.175324440002441,
    "prediction_norm": 7.044888973236084,
    "timestamp": "2025-07-14T12:33:40.920259"
  },
  {
    "epoch": 213,
    "batch_idx": 96,
    "loss": 0.42841118574142456,
    "learning_rate": 0.0009722742174750563,
    "timestep_mean": 543.71875,
    "epsilon_norm": 8.473883628845215,
    "prediction_norm": 6.836319923400879,
    "timestamp": "2025-07-14T12:33:40.925599"
  },
  {
    "epoch": 213,
    "batch_idx": 128,
    "loss": 0.23596912622451782,
    "learning_rate": 0.0009722742174750563,
    "timestep_mean": 567.8125,
    "epsilon_norm": 8.580604553222656,
    "prediction_norm": 7.53084659576416,
    "timestamp": "2025-07-14T12:33:40.930677"
  },
  {
    "epoch": 213,
    "batch_idx": 160,
    "loss": 0.4200858473777771,
    "learning_rate": 0.0009722742174750563,
    "timestep_mean": 493.71875,
    "epsilon_norm": 8.755523681640625,
    "prediction_norm": 6.660022258758545,
    "timestamp": "2025-07-14T12:33:40.935901"
  },
  {
    "epoch": 213,
    "batch_idx": 192,
    "loss": 0.48264509439468384,
    "learning_rate": 0.0009722742174750563,
    "timestep_mean": 456.875,
    "epsilon_norm": 3.218428611755371,
    "prediction_norm": 1.6147630214691162,
    "timestamp": "2025-07-14T12:33:40.941109"
  },
  {
    "epoch": 214,
    "batch_idx": 0,
    "loss": 0.2314845323562622,
    "learning_rate": 0.0009720157320705254,
    "timestep_mean": 543.25,
    "epsilon_norm": 7.743490695953369,
    "prediction_norm": 6.461840629577637,
    "timestamp": "2025-07-14T12:33:40.946345"
  },
  {
    "epoch": 214,
    "batch_idx": 32,
    "loss": 0.34886252880096436,
    "learning_rate": 0.0009720157320705254,
    "timestep_mean": 519.03125,
    "epsilon_norm": 7.809394836425781,
    "prediction_norm": 6.4469828605651855,
    "timestamp": "2025-07-14T12:33:40.951604"
  },
  {
    "epoch": 214,
    "batch_idx": 64,
    "loss": 0.3357508182525635,
    "learning_rate": 0.0009720157320705254,
    "timestep_mean": 474.90625,
    "epsilon_norm": 8.257670402526855,
    "prediction_norm": 6.354870796203613,
    "timestamp": "2025-07-14T12:33:40.956770"
  },
  {
    "epoch": 214,
    "batch_idx": 96,
    "loss": 0.3623031675815582,
    "learning_rate": 0.0009720157320705254,
    "timestep_mean": 447.5,
    "epsilon_norm": 8.055407524108887,
    "prediction_norm": 6.154675006866455,
    "timestamp": "2025-07-14T12:33:40.961958"
  },
  {
    "epoch": 214,
    "batch_idx": 128,
    "loss": 0.26389968395233154,
    "learning_rate": 0.0009720157320705254,
    "timestep_mean": 553.21875,
    "epsilon_norm": 8.683791160583496,
    "prediction_norm": 8.056198120117188,
    "timestamp": "2025-07-14T12:33:40.967009"
  },
  {
    "epoch": 214,
    "batch_idx": 160,
    "loss": 0.4318622052669525,
    "learning_rate": 0.0009720157320705254,
    "timestep_mean": 501.0,
    "epsilon_norm": 7.933224201202393,
    "prediction_norm": 6.14851188659668,
    "timestamp": "2025-07-14T12:33:40.971980"
  },
  {
    "epoch": 214,
    "batch_idx": 192,
    "loss": 0.6605227589607239,
    "learning_rate": 0.0009720157320705254,
    "timestep_mean": 518.625,
    "epsilon_norm": 4.432162284851074,
    "prediction_norm": 2.944262981414795,
    "timestamp": "2025-07-14T12:33:40.977023"
  },
  {
    "epoch": 215,
    "batch_idx": 0,
    "loss": 0.43316543102264404,
    "learning_rate": 0.0009717560820140973,
    "timestep_mean": 462.21875,
    "epsilon_norm": 7.844769477844238,
    "prediction_norm": 6.073803901672363,
    "timestamp": "2025-07-14T12:33:40.982231"
  },
  {
    "epoch": 215,
    "batch_idx": 32,
    "loss": 0.4211694002151489,
    "learning_rate": 0.0009717560820140973,
    "timestep_mean": 493.125,
    "epsilon_norm": 8.233272552490234,
    "prediction_norm": 6.6113739013671875,
    "timestamp": "2025-07-14T12:33:40.987419"
  },
  {
    "epoch": 215,
    "batch_idx": 64,
    "loss": 0.7354840040206909,
    "learning_rate": 0.0009717560820140973,
    "timestep_mean": 474.0,
    "epsilon_norm": 9.200237274169922,
    "prediction_norm": 6.443537712097168,
    "timestamp": "2025-07-14T12:33:40.992451"
  },
  {
    "epoch": 215,
    "batch_idx": 96,
    "loss": 0.14413225650787354,
    "learning_rate": 0.0009717560820140973,
    "timestep_mean": 583.375,
    "epsilon_norm": 5.9054107666015625,
    "prediction_norm": 5.348276138305664,
    "timestamp": "2025-07-14T12:33:40.997564"
  },
  {
    "epoch": 215,
    "batch_idx": 128,
    "loss": 0.43414008617401123,
    "learning_rate": 0.0009717560820140973,
    "timestep_mean": 544.90625,
    "epsilon_norm": 7.847953796386719,
    "prediction_norm": 6.492234230041504,
    "timestamp": "2025-07-14T12:33:41.002719"
  },
  {
    "epoch": 215,
    "batch_idx": 160,
    "loss": 0.29612088203430176,
    "learning_rate": 0.0009717560820140973,
    "timestep_mean": 501.78125,
    "epsilon_norm": 8.170215606689453,
    "prediction_norm": 6.623138904571533,
    "timestamp": "2025-07-14T12:33:41.007967"
  },
  {
    "epoch": 215,
    "batch_idx": 192,
    "loss": 0.3053930997848511,
    "learning_rate": 0.0009717560820140973,
    "timestep_mean": 397.25,
    "epsilon_norm": 3.5857486724853516,
    "prediction_norm": 2.5707967281341553,
    "timestamp": "2025-07-14T12:33:41.012989"
  },
  {
    "epoch": 216,
    "batch_idx": 0,
    "loss": 0.3642438054084778,
    "learning_rate": 0.0009714952679464328,
    "timestep_mean": 486.59375,
    "epsilon_norm": 7.174489498138428,
    "prediction_norm": 5.3094258308410645,
    "timestamp": "2025-07-14T12:33:41.018046"
  },
  {
    "epoch": 216,
    "batch_idx": 32,
    "loss": 0.346252977848053,
    "learning_rate": 0.0009714952679464328,
    "timestep_mean": 413.21875,
    "epsilon_norm": 6.832430362701416,
    "prediction_norm": 4.853271961212158,
    "timestamp": "2025-07-14T12:33:41.023228"
  },
  {
    "epoch": 216,
    "batch_idx": 64,
    "loss": 0.24859389662742615,
    "learning_rate": 0.0009714952679464328,
    "timestep_mean": 536.875,
    "epsilon_norm": 8.058785438537598,
    "prediction_norm": 6.948939323425293,
    "timestamp": "2025-07-14T12:33:41.028204"
  },
  {
    "epoch": 216,
    "batch_idx": 96,
    "loss": 0.3531762361526489,
    "learning_rate": 0.0009714952679464328,
    "timestep_mean": 459.8125,
    "epsilon_norm": 7.333559989929199,
    "prediction_norm": 5.437101364135742,
    "timestamp": "2025-07-14T12:33:41.033380"
  },
  {
    "epoch": 216,
    "batch_idx": 128,
    "loss": 0.2533641755580902,
    "learning_rate": 0.0009714952679464328,
    "timestep_mean": 491.875,
    "epsilon_norm": 7.1445512771606445,
    "prediction_norm": 5.607593059539795,
    "timestamp": "2025-07-14T12:33:41.038272"
  },
  {
    "epoch": 216,
    "batch_idx": 160,
    "loss": 0.27642104029655457,
    "learning_rate": 0.0009714952679464328,
    "timestep_mean": 495.28125,
    "epsilon_norm": 7.246965408325195,
    "prediction_norm": 6.109582901000977,
    "timestamp": "2025-07-14T12:33:41.043246"
  },
  {
    "epoch": 216,
    "batch_idx": 192,
    "loss": 0.6113734245300293,
    "learning_rate": 0.0009714952679464328,
    "timestep_mean": 363.625,
    "epsilon_norm": 3.5467617511749268,
    "prediction_norm": 2.354506015777588,
    "timestamp": "2025-07-14T12:33:41.048973"
  },
  {
    "epoch": 217,
    "batch_idx": 0,
    "loss": 0.3367953300476074,
    "learning_rate": 0.0009712332905110645,
    "timestep_mean": 473.5,
    "epsilon_norm": 7.9688639640808105,
    "prediction_norm": 6.036840438842773,
    "timestamp": "2025-07-14T12:33:41.054651"
  },
  {
    "epoch": 217,
    "batch_idx": 32,
    "loss": 0.42308834195137024,
    "learning_rate": 0.0009712332905110645,
    "timestep_mean": 455.375,
    "epsilon_norm": 7.7239155769348145,
    "prediction_norm": 5.245384216308594,
    "timestamp": "2025-07-14T12:33:41.060075"
  },
  {
    "epoch": 217,
    "batch_idx": 64,
    "loss": 0.31776025891304016,
    "learning_rate": 0.0009712332905110645,
    "timestep_mean": 467.53125,
    "epsilon_norm": 7.266770839691162,
    "prediction_norm": 5.927569389343262,
    "timestamp": "2025-07-14T12:33:41.066277"
  },
  {
    "epoch": 217,
    "batch_idx": 96,
    "loss": 0.4475162625312805,
    "learning_rate": 0.0009712332905110645,
    "timestep_mean": 398.65625,
    "epsilon_norm": 8.216972351074219,
    "prediction_norm": 6.601370811462402,
    "timestamp": "2025-07-14T12:33:41.071992"
  },
  {
    "epoch": 217,
    "batch_idx": 128,
    "loss": 0.22463294863700867,
    "learning_rate": 0.0009712332905110645,
    "timestep_mean": 488.3125,
    "epsilon_norm": 6.158415794372559,
    "prediction_norm": 4.9817962646484375,
    "timestamp": "2025-07-14T12:33:41.077471"
  },
  {
    "epoch": 217,
    "batch_idx": 160,
    "loss": 0.45659518241882324,
    "learning_rate": 0.0009712332905110645,
    "timestep_mean": 460.40625,
    "epsilon_norm": 7.989046573638916,
    "prediction_norm": 6.335320472717285,
    "timestamp": "2025-07-14T12:33:41.082878"
  },
  {
    "epoch": 217,
    "batch_idx": 192,
    "loss": 0.32693004608154297,
    "learning_rate": 0.0009712332905110645,
    "timestep_mean": 516.75,
    "epsilon_norm": 4.137400150299072,
    "prediction_norm": 3.231492042541504,
    "timestamp": "2025-07-14T12:33:41.088463"
  },
  {
    "epoch": 218,
    "batch_idx": 0,
    "loss": 0.24506664276123047,
    "learning_rate": 0.0009709701503543957,
    "timestep_mean": 500.96875,
    "epsilon_norm": 8.373664855957031,
    "prediction_norm": 7.170017242431641,
    "timestamp": "2025-07-14T12:33:41.094156"
  },
  {
    "epoch": 218,
    "batch_idx": 32,
    "loss": 0.272879958152771,
    "learning_rate": 0.0009709701503543957,
    "timestep_mean": 574.84375,
    "epsilon_norm": 7.993661403656006,
    "prediction_norm": 6.81988000869751,
    "timestamp": "2025-07-14T12:33:41.104895"
  },
  {
    "epoch": 218,
    "batch_idx": 64,
    "loss": 0.30821895599365234,
    "learning_rate": 0.0009709701503543957,
    "timestep_mean": 446.5,
    "epsilon_norm": 7.184412956237793,
    "prediction_norm": 5.842522621154785,
    "timestamp": "2025-07-14T12:33:41.110348"
  },
  {
    "epoch": 218,
    "batch_idx": 96,
    "loss": 0.23133584856987,
    "learning_rate": 0.0009709701503543957,
    "timestep_mean": 555.59375,
    "epsilon_norm": 7.76155948638916,
    "prediction_norm": 6.975061416625977,
    "timestamp": "2025-07-14T12:33:41.115522"
  },
  {
    "epoch": 218,
    "batch_idx": 128,
    "loss": 0.3778078556060791,
    "learning_rate": 0.0009709701503543957,
    "timestep_mean": 418.125,
    "epsilon_norm": 7.4348626136779785,
    "prediction_norm": 5.22147798538208,
    "timestamp": "2025-07-14T12:33:41.120847"
  },
  {
    "epoch": 218,
    "batch_idx": 160,
    "loss": 0.26544174551963806,
    "learning_rate": 0.0009709701503543957,
    "timestep_mean": 535.3125,
    "epsilon_norm": 8.070151329040527,
    "prediction_norm": 6.814970970153809,
    "timestamp": "2025-07-14T12:33:41.126467"
  },
  {
    "epoch": 218,
    "batch_idx": 192,
    "loss": 1.2118855714797974,
    "learning_rate": 0.0009709701503543957,
    "timestep_mean": 370.375,
    "epsilon_norm": 4.347548961639404,
    "prediction_norm": 1.3166844844818115,
    "timestamp": "2025-07-14T12:33:41.131691"
  },
  {
    "epoch": 219,
    "batch_idx": 0,
    "loss": 0.3235297203063965,
    "learning_rate": 0.0009707058481256989,
    "timestep_mean": 490.8125,
    "epsilon_norm": 8.229619026184082,
    "prediction_norm": 6.8144965171813965,
    "timestamp": "2025-07-14T12:33:41.136935"
  },
  {
    "epoch": 219,
    "batch_idx": 32,
    "loss": 0.20081907510757446,
    "learning_rate": 0.0009707058481256989,
    "timestep_mean": 501.625,
    "epsilon_norm": 8.255269050598145,
    "prediction_norm": 7.7608489990234375,
    "timestamp": "2025-07-14T12:33:41.142097"
  },
  {
    "epoch": 219,
    "batch_idx": 64,
    "loss": 0.4681142568588257,
    "learning_rate": 0.0009707058481256989,
    "timestep_mean": 434.1875,
    "epsilon_norm": 7.5881571769714355,
    "prediction_norm": 5.025596618652344,
    "timestamp": "2025-07-14T12:33:41.147032"
  },
  {
    "epoch": 219,
    "batch_idx": 96,
    "loss": 0.27930670976638794,
    "learning_rate": 0.0009707058481256989,
    "timestep_mean": 527.84375,
    "epsilon_norm": 8.354900360107422,
    "prediction_norm": 6.762811183929443,
    "timestamp": "2025-07-14T12:33:41.152054"
  },
  {
    "epoch": 219,
    "batch_idx": 128,
    "loss": 0.43684306740760803,
    "learning_rate": 0.0009707058481256989,
    "timestep_mean": 458.8125,
    "epsilon_norm": 7.6740403175354,
    "prediction_norm": 5.375432968139648,
    "timestamp": "2025-07-14T12:33:41.157147"
  },
  {
    "epoch": 219,
    "batch_idx": 160,
    "loss": 0.3490334153175354,
    "learning_rate": 0.0009707058481256989,
    "timestep_mean": 562.78125,
    "epsilon_norm": 7.989283084869385,
    "prediction_norm": 7.0659871101379395,
    "timestamp": "2025-07-14T12:33:41.162348"
  },
  {
    "epoch": 219,
    "batch_idx": 192,
    "loss": 0.11441044509410858,
    "learning_rate": 0.0009707058481256989,
    "timestep_mean": 613.75,
    "epsilon_norm": 3.654064416885376,
    "prediction_norm": 3.6669256687164307,
    "timestamp": "2025-07-14T12:33:41.167491"
  },
  {
    "epoch": 220,
    "batch_idx": 0,
    "loss": 0.41791820526123047,
    "learning_rate": 0.0009704403844771133,
    "timestep_mean": 464.375,
    "epsilon_norm": 8.45164680480957,
    "prediction_norm": 7.610401153564453,
    "timestamp": "2025-07-14T12:33:41.172487"
  },
  {
    "epoch": 220,
    "batch_idx": 32,
    "loss": 0.5482832193374634,
    "learning_rate": 0.0009704403844771133,
    "timestep_mean": 521.125,
    "epsilon_norm": 8.521142959594727,
    "prediction_norm": 6.873719692230225,
    "timestamp": "2025-07-14T12:33:41.177639"
  },
  {
    "epoch": 220,
    "batch_idx": 64,
    "loss": 0.3177177608013153,
    "learning_rate": 0.0009704403844771133,
    "timestep_mean": 514.46875,
    "epsilon_norm": 9.88440990447998,
    "prediction_norm": 8.581897735595703,
    "timestamp": "2025-07-14T12:33:41.182579"
  },
  {
    "epoch": 220,
    "batch_idx": 96,
    "loss": 0.41017305850982666,
    "learning_rate": 0.0009704403844771133,
    "timestep_mean": 521.46875,
    "epsilon_norm": 7.776589393615723,
    "prediction_norm": 7.400843143463135,
    "timestamp": "2025-07-14T12:33:41.187469"
  },
  {
    "epoch": 220,
    "batch_idx": 128,
    "loss": 0.3386151194572449,
    "learning_rate": 0.0009704403844771133,
    "timestep_mean": 515.90625,
    "epsilon_norm": 7.486819744110107,
    "prediction_norm": 6.46102237701416,
    "timestamp": "2025-07-14T12:33:41.192481"
  },
  {
    "epoch": 220,
    "batch_idx": 160,
    "loss": 0.5774587988853455,
    "learning_rate": 0.0009704403844771133,
    "timestep_mean": 413.15625,
    "epsilon_norm": 7.514098644256592,
    "prediction_norm": 5.307523250579834,
    "timestamp": "2025-07-14T12:33:41.197412"
  },
  {
    "epoch": 220,
    "batch_idx": 192,
    "loss": 0.12798652052879333,
    "learning_rate": 0.0009704403844771133,
    "timestep_mean": 520.375,
    "epsilon_norm": 3.9228219985961914,
    "prediction_norm": 3.680452346801758,
    "timestamp": "2025-07-14T12:33:41.202444"
  },
  {
    "epoch": 221,
    "batch_idx": 0,
    "loss": 0.39487770199775696,
    "learning_rate": 0.000970173760063644,
    "timestep_mean": 458.9375,
    "epsilon_norm": 9.168624877929688,
    "prediction_norm": 7.988271236419678,
    "timestamp": "2025-07-14T12:33:41.207391"
  },
  {
    "epoch": 221,
    "batch_idx": 32,
    "loss": 0.5139639377593994,
    "learning_rate": 0.000970173760063644,
    "timestep_mean": 472.25,
    "epsilon_norm": 8.817060470581055,
    "prediction_norm": 5.730572700500488,
    "timestamp": "2025-07-14T12:33:41.212312"
  },
  {
    "epoch": 221,
    "batch_idx": 64,
    "loss": 0.38031575083732605,
    "learning_rate": 0.000970173760063644,
    "timestep_mean": 426.21875,
    "epsilon_norm": 6.31685209274292,
    "prediction_norm": 3.7563893795013428,
    "timestamp": "2025-07-14T12:33:41.217234"
  },
  {
    "epoch": 221,
    "batch_idx": 96,
    "loss": 0.38245874643325806,
    "learning_rate": 0.000970173760063644,
    "timestep_mean": 544.8125,
    "epsilon_norm": 8.561737060546875,
    "prediction_norm": 6.83828067779541,
    "timestamp": "2025-07-14T12:33:41.222102"
  },
  {
    "epoch": 221,
    "batch_idx": 128,
    "loss": 0.3031230568885803,
    "learning_rate": 0.000970173760063644,
    "timestep_mean": 479.09375,
    "epsilon_norm": 8.425493240356445,
    "prediction_norm": 6.373313903808594,
    "timestamp": "2025-07-14T12:33:41.226937"
  },
  {
    "epoch": 221,
    "batch_idx": 160,
    "loss": 0.27247154712677,
    "learning_rate": 0.000970173760063644,
    "timestep_mean": 542.4375,
    "epsilon_norm": 7.183475494384766,
    "prediction_norm": 5.3974690437316895,
    "timestamp": "2025-07-14T12:33:41.231835"
  },
  {
    "epoch": 221,
    "batch_idx": 192,
    "loss": 0.3512199819087982,
    "learning_rate": 0.000970173760063644,
    "timestep_mean": 461.125,
    "epsilon_norm": 3.5077288150787354,
    "prediction_norm": 2.4593327045440674,
    "timestamp": "2025-07-14T12:33:41.236711"
  },
  {
    "epoch": 222,
    "batch_idx": 0,
    "loss": 0.319174200296402,
    "learning_rate": 0.0009699059755431603,
    "timestep_mean": 538.375,
    "epsilon_norm": 8.715664863586426,
    "prediction_norm": 6.672312259674072,
    "timestamp": "2025-07-14T12:33:41.241722"
  },
  {
    "epoch": 222,
    "batch_idx": 32,
    "loss": 0.4573338031768799,
    "learning_rate": 0.0009699059755431603,
    "timestep_mean": 480.40625,
    "epsilon_norm": 8.426885604858398,
    "prediction_norm": 5.97202205657959,
    "timestamp": "2025-07-14T12:33:41.246683"
  },
  {
    "epoch": 222,
    "batch_idx": 64,
    "loss": 0.1377314031124115,
    "learning_rate": 0.0009699059755431603,
    "timestep_mean": 571.8125,
    "epsilon_norm": 6.741800308227539,
    "prediction_norm": 5.706418991088867,
    "timestamp": "2025-07-14T12:33:41.251617"
  },
  {
    "epoch": 222,
    "batch_idx": 96,
    "loss": 0.3607771396636963,
    "learning_rate": 0.0009699059755431603,
    "timestep_mean": 489.5625,
    "epsilon_norm": 7.946305274963379,
    "prediction_norm": 6.299427509307861,
    "timestamp": "2025-07-14T12:33:41.256495"
  },
  {
    "epoch": 222,
    "batch_idx": 128,
    "loss": 0.4176241457462311,
    "learning_rate": 0.0009699059755431603,
    "timestep_mean": 488.5,
    "epsilon_norm": 8.106552124023438,
    "prediction_norm": 6.327889442443848,
    "timestamp": "2025-07-14T12:33:41.261434"
  },
  {
    "epoch": 222,
    "batch_idx": 160,
    "loss": 0.2115378975868225,
    "learning_rate": 0.0009699059755431603,
    "timestep_mean": 585.625,
    "epsilon_norm": 8.50744342803955,
    "prediction_norm": 7.383934020996094,
    "timestamp": "2025-07-14T12:33:41.266399"
  },
  {
    "epoch": 222,
    "batch_idx": 192,
    "loss": 0.10246855020523071,
    "learning_rate": 0.0009699059755431603,
    "timestep_mean": 573.625,
    "epsilon_norm": 2.7716264724731445,
    "prediction_norm": 2.554924488067627,
    "timestamp": "2025-07-14T12:33:41.271421"
  },
  {
    "epoch": 223,
    "batch_idx": 0,
    "loss": 0.3921490013599396,
    "learning_rate": 0.000969637031576394,
    "timestep_mean": 517.21875,
    "epsilon_norm": 7.182446002960205,
    "prediction_norm": 6.292383670806885,
    "timestamp": "2025-07-14T12:33:41.276362"
  },
  {
    "epoch": 223,
    "batch_idx": 32,
    "loss": 0.34850773215293884,
    "learning_rate": 0.000969637031576394,
    "timestep_mean": 481.375,
    "epsilon_norm": 8.172160148620605,
    "prediction_norm": 7.822056770324707,
    "timestamp": "2025-07-14T12:33:41.281316"
  },
  {
    "epoch": 223,
    "batch_idx": 64,
    "loss": 0.24400271475315094,
    "learning_rate": 0.000969637031576394,
    "timestep_mean": 460.6875,
    "epsilon_norm": 7.076143741607666,
    "prediction_norm": 6.611865520477295,
    "timestamp": "2025-07-14T12:33:41.286357"
  },
  {
    "epoch": 223,
    "batch_idx": 96,
    "loss": 0.3393212556838989,
    "learning_rate": 0.000969637031576394,
    "timestep_mean": 414.03125,
    "epsilon_norm": 7.891327381134033,
    "prediction_norm": 7.079727649688721,
    "timestamp": "2025-07-14T12:33:41.291206"
  },
  {
    "epoch": 223,
    "batch_idx": 128,
    "loss": 0.24537153542041779,
    "learning_rate": 0.000969637031576394,
    "timestep_mean": 563.8125,
    "epsilon_norm": 7.968796730041504,
    "prediction_norm": 6.877852916717529,
    "timestamp": "2025-07-14T12:33:41.296212"
  },
  {
    "epoch": 223,
    "batch_idx": 160,
    "loss": 0.40822118520736694,
    "learning_rate": 0.000969637031576394,
    "timestep_mean": 475.46875,
    "epsilon_norm": 8.171724319458008,
    "prediction_norm": 6.743762969970703,
    "timestamp": "2025-07-14T12:33:41.302468"
  },
  {
    "epoch": 223,
    "batch_idx": 192,
    "loss": 0.27727267146110535,
    "learning_rate": 0.000969637031576394,
    "timestep_mean": 595.625,
    "epsilon_norm": 3.719266891479492,
    "prediction_norm": 3.084533929824829,
    "timestamp": "2025-07-14T12:33:41.307497"
  },
  {
    "epoch": 224,
    "batch_idx": 0,
    "loss": 0.41616037487983704,
    "learning_rate": 0.0009693669288269375,
    "timestep_mean": 537.5625,
    "epsilon_norm": 8.687081336975098,
    "prediction_norm": 6.598635673522949,
    "timestamp": "2025-07-14T12:33:41.312507"
  },
  {
    "epoch": 224,
    "batch_idx": 32,
    "loss": 0.39942917227745056,
    "learning_rate": 0.0009693669288269375,
    "timestep_mean": 454.21875,
    "epsilon_norm": 6.916072368621826,
    "prediction_norm": 4.978482723236084,
    "timestamp": "2025-07-14T12:33:41.317459"
  },
  {
    "epoch": 224,
    "batch_idx": 64,
    "loss": 0.42732977867126465,
    "learning_rate": 0.0009693669288269375,
    "timestep_mean": 474.90625,
    "epsilon_norm": 8.285869598388672,
    "prediction_norm": 5.573022365570068,
    "timestamp": "2025-07-14T12:33:41.322374"
  },
  {
    "epoch": 224,
    "batch_idx": 96,
    "loss": 0.4280204474925995,
    "learning_rate": 0.0009693669288269375,
    "timestep_mean": 511.625,
    "epsilon_norm": 7.6966872215271,
    "prediction_norm": 5.752046585083008,
    "timestamp": "2025-07-14T12:33:41.327278"
  },
  {
    "epoch": 224,
    "batch_idx": 128,
    "loss": 0.25521454215049744,
    "learning_rate": 0.0009693669288269375,
    "timestep_mean": 519.21875,
    "epsilon_norm": 8.4108247756958,
    "prediction_norm": 6.827477931976318,
    "timestamp": "2025-07-14T12:33:41.332231"
  },
  {
    "epoch": 224,
    "batch_idx": 160,
    "loss": 0.25760236382484436,
    "learning_rate": 0.0009693669288269375,
    "timestep_mean": 484.875,
    "epsilon_norm": 6.8215250968933105,
    "prediction_norm": 5.517899036407471,
    "timestamp": "2025-07-14T12:33:41.337129"
  },
  {
    "epoch": 224,
    "batch_idx": 192,
    "loss": 0.4191780090332031,
    "learning_rate": 0.0009693669288269375,
    "timestep_mean": 365.375,
    "epsilon_norm": 4.190253734588623,
    "prediction_norm": 2.7721617221832275,
    "timestamp": "2025-07-14T12:33:41.342155"
  },
  {
    "epoch": 225,
    "batch_idx": 0,
    "loss": 0.39711835980415344,
    "learning_rate": 0.0009690956679612425,
    "timestep_mean": 555.84375,
    "epsilon_norm": 7.985769748687744,
    "prediction_norm": 5.838931083679199,
    "timestamp": "2025-07-14T12:33:41.347063"
  },
  {
    "epoch": 225,
    "batch_idx": 32,
    "loss": 0.2770857512950897,
    "learning_rate": 0.0009690956679612425,
    "timestep_mean": 526.1875,
    "epsilon_norm": 8.048355102539062,
    "prediction_norm": 6.483038425445557,
    "timestamp": "2025-07-14T12:33:41.352003"
  },
  {
    "epoch": 225,
    "batch_idx": 64,
    "loss": 0.4023550748825073,
    "learning_rate": 0.0009690956679612425,
    "timestep_mean": 453.8125,
    "epsilon_norm": 7.486982822418213,
    "prediction_norm": 4.638513565063477,
    "timestamp": "2025-07-14T12:33:41.356922"
  },
  {
    "epoch": 225,
    "batch_idx": 96,
    "loss": 0.23541410267353058,
    "learning_rate": 0.0009690956679612425,
    "timestep_mean": 492.875,
    "epsilon_norm": 8.01455020904541,
    "prediction_norm": 7.528802394866943,
    "timestamp": "2025-07-14T12:33:41.361889"
  },
  {
    "epoch": 225,
    "batch_idx": 128,
    "loss": 0.6141762733459473,
    "learning_rate": 0.0009690956679612425,
    "timestep_mean": 381.46875,
    "epsilon_norm": 8.136872291564941,
    "prediction_norm": 5.387387752532959,
    "timestamp": "2025-07-14T12:33:41.366913"
  },
  {
    "epoch": 225,
    "batch_idx": 160,
    "loss": 0.30263736844062805,
    "learning_rate": 0.0009690956679612425,
    "timestep_mean": 527.0,
    "epsilon_norm": 8.00533676147461,
    "prediction_norm": 6.55251407623291,
    "timestamp": "2025-07-14T12:33:41.371800"
  },
  {
    "epoch": 225,
    "batch_idx": 192,
    "loss": 0.5049403309822083,
    "learning_rate": 0.0009690956679612425,
    "timestep_mean": 382.0,
    "epsilon_norm": 3.9886844158172607,
    "prediction_norm": 2.7443177700042725,
    "timestamp": "2025-07-14T12:33:41.376737"
  },
  {
    "epoch": 226,
    "batch_idx": 0,
    "loss": 0.21055889129638672,
    "learning_rate": 0.0009688232496486181,
    "timestep_mean": 552.65625,
    "epsilon_norm": 8.217560768127441,
    "prediction_norm": 7.150450229644775,
    "timestamp": "2025-07-14T12:33:41.381665"
  },
  {
    "epoch": 226,
    "batch_idx": 32,
    "loss": 0.2722688317298889,
    "learning_rate": 0.0009688232496486181,
    "timestep_mean": 557.75,
    "epsilon_norm": 8.005709648132324,
    "prediction_norm": 6.473597526550293,
    "timestamp": "2025-07-14T12:33:41.386593"
  },
  {
    "epoch": 226,
    "batch_idx": 64,
    "loss": 0.26523295044898987,
    "learning_rate": 0.0009688232496486181,
    "timestep_mean": 552.03125,
    "epsilon_norm": 7.743731498718262,
    "prediction_norm": 7.374416828155518,
    "timestamp": "2025-07-14T12:33:41.391503"
  },
  {
    "epoch": 226,
    "batch_idx": 96,
    "loss": 0.36493250727653503,
    "learning_rate": 0.0009688232496486181,
    "timestep_mean": 532.96875,
    "epsilon_norm": 7.636924743652344,
    "prediction_norm": 6.838954925537109,
    "timestamp": "2025-07-14T12:33:41.396410"
  },
  {
    "epoch": 226,
    "batch_idx": 128,
    "loss": 0.21360968053340912,
    "learning_rate": 0.0009688232496486181,
    "timestep_mean": 490.53125,
    "epsilon_norm": 8.39876651763916,
    "prediction_norm": 8.417612075805664,
    "timestamp": "2025-07-14T12:33:41.401303"
  },
  {
    "epoch": 226,
    "batch_idx": 160,
    "loss": 0.614268958568573,
    "learning_rate": 0.0009688232496486181,
    "timestep_mean": 425.28125,
    "epsilon_norm": 8.951916694641113,
    "prediction_norm": 6.898895263671875,
    "timestamp": "2025-07-14T12:33:41.406168"
  },
  {
    "epoch": 226,
    "batch_idx": 192,
    "loss": 0.11903508007526398,
    "learning_rate": 0.0009688232496486181,
    "timestep_mean": 485.875,
    "epsilon_norm": 3.788684844970703,
    "prediction_norm": 3.7083499431610107,
    "timestamp": "2025-07-14T12:33:41.411117"
  },
  {
    "epoch": 227,
    "batch_idx": 0,
    "loss": 0.537155270576477,
    "learning_rate": 0.0009685496745612298,
    "timestep_mean": 471.90625,
    "epsilon_norm": 8.627957344055176,
    "prediction_norm": 6.041202545166016,
    "timestamp": "2025-07-14T12:33:41.416278"
  },
  {
    "epoch": 227,
    "batch_idx": 32,
    "loss": 0.3144018054008484,
    "learning_rate": 0.0009685496745612298,
    "timestep_mean": 515.125,
    "epsilon_norm": 7.349217891693115,
    "prediction_norm": 5.8411455154418945,
    "timestamp": "2025-07-14T12:33:41.421626"
  },
  {
    "epoch": 227,
    "batch_idx": 64,
    "loss": 0.1826610267162323,
    "learning_rate": 0.0009685496745612298,
    "timestep_mean": 605.25,
    "epsilon_norm": 8.66043472290039,
    "prediction_norm": 7.545737266540527,
    "timestamp": "2025-07-14T12:33:41.426657"
  },
  {
    "epoch": 227,
    "batch_idx": 96,
    "loss": 0.2291005402803421,
    "learning_rate": 0.0009685496745612298,
    "timestep_mean": 536.5,
    "epsilon_norm": 7.096254825592041,
    "prediction_norm": 5.766153812408447,
    "timestamp": "2025-07-14T12:33:41.431756"
  },
  {
    "epoch": 227,
    "batch_idx": 128,
    "loss": 0.32059839367866516,
    "learning_rate": 0.0009685496745612298,
    "timestep_mean": 495.71875,
    "epsilon_norm": 6.803040504455566,
    "prediction_norm": 5.216886520385742,
    "timestamp": "2025-07-14T12:33:41.436712"
  },
  {
    "epoch": 227,
    "batch_idx": 160,
    "loss": 0.2496212273836136,
    "learning_rate": 0.0009685496745612298,
    "timestep_mean": 504.5625,
    "epsilon_norm": 7.829354286193848,
    "prediction_norm": 6.827521800994873,
    "timestamp": "2025-07-14T12:33:41.441714"
  },
  {
    "epoch": 227,
    "batch_idx": 192,
    "loss": 0.6237369775772095,
    "learning_rate": 0.0009685496745612298,
    "timestep_mean": 506.875,
    "epsilon_norm": 4.387098789215088,
    "prediction_norm": 2.8414034843444824,
    "timestamp": "2025-07-14T12:33:41.446699"
  },
  {
    "epoch": 228,
    "batch_idx": 0,
    "loss": 0.4181319773197174,
    "learning_rate": 0.0009682749433740966,
    "timestep_mean": 439.96875,
    "epsilon_norm": 7.503528118133545,
    "prediction_norm": 5.820066928863525,
    "timestamp": "2025-07-14T12:33:41.451682"
  },
  {
    "epoch": 228,
    "batch_idx": 32,
    "loss": 0.3023146092891693,
    "learning_rate": 0.0009682749433740966,
    "timestep_mean": 521.84375,
    "epsilon_norm": 7.966155529022217,
    "prediction_norm": 7.08296537399292,
    "timestamp": "2025-07-14T12:33:41.456635"
  },
  {
    "epoch": 228,
    "batch_idx": 64,
    "loss": 0.43170228600502014,
    "learning_rate": 0.0009682749433740966,
    "timestep_mean": 434.65625,
    "epsilon_norm": 8.499787330627441,
    "prediction_norm": 6.732306480407715,
    "timestamp": "2025-07-14T12:33:41.461593"
  },
  {
    "epoch": 228,
    "batch_idx": 96,
    "loss": 0.3149155378341675,
    "learning_rate": 0.0009682749433740966,
    "timestep_mean": 482.15625,
    "epsilon_norm": 7.551005840301514,
    "prediction_norm": 6.174107551574707,
    "timestamp": "2025-07-14T12:33:41.466501"
  },
  {
    "epoch": 228,
    "batch_idx": 128,
    "loss": 0.23009121417999268,
    "learning_rate": 0.0009682749433740966,
    "timestep_mean": 497.40625,
    "epsilon_norm": 8.630313873291016,
    "prediction_norm": 8.001723289489746,
    "timestamp": "2025-07-14T12:33:41.471450"
  },
  {
    "epoch": 228,
    "batch_idx": 160,
    "loss": 0.2130919247865677,
    "learning_rate": 0.0009682749433740966,
    "timestep_mean": 466.875,
    "epsilon_norm": 7.72353458404541,
    "prediction_norm": 6.692541122436523,
    "timestamp": "2025-07-14T12:33:41.476318"
  },
  {
    "epoch": 228,
    "batch_idx": 192,
    "loss": 0.37891924381256104,
    "learning_rate": 0.0009682749433740966,
    "timestep_mean": 475.875,
    "epsilon_norm": 3.776791572570801,
    "prediction_norm": 2.8158133029937744,
    "timestamp": "2025-07-14T12:33:41.481303"
  },
  {
    "epoch": 229,
    "batch_idx": 0,
    "loss": 0.3392764627933502,
    "learning_rate": 0.0009679990567650903,
    "timestep_mean": 578.34375,
    "epsilon_norm": 8.525014877319336,
    "prediction_norm": 6.884515285491943,
    "timestamp": "2025-07-14T12:33:41.486269"
  },
  {
    "epoch": 229,
    "batch_idx": 32,
    "loss": 0.3387441337108612,
    "learning_rate": 0.0009679990567650903,
    "timestep_mean": 423.25,
    "epsilon_norm": 7.623795032501221,
    "prediction_norm": 5.51043701171875,
    "timestamp": "2025-07-14T12:33:41.491183"
  },
  {
    "epoch": 229,
    "batch_idx": 64,
    "loss": 0.3008601665496826,
    "learning_rate": 0.0009679990567650903,
    "timestep_mean": 531.0625,
    "epsilon_norm": 7.687475204467773,
    "prediction_norm": 5.757003307342529,
    "timestamp": "2025-07-14T12:33:41.496072"
  },
  {
    "epoch": 229,
    "batch_idx": 96,
    "loss": 0.21702653169631958,
    "learning_rate": 0.0009679990567650903,
    "timestep_mean": 513.90625,
    "epsilon_norm": 7.195687294006348,
    "prediction_norm": 6.160331726074219,
    "timestamp": "2025-07-14T12:33:41.500955"
  },
  {
    "epoch": 229,
    "batch_idx": 128,
    "loss": 0.21298298239707947,
    "learning_rate": 0.0009679990567650903,
    "timestep_mean": 551.1875,
    "epsilon_norm": 8.947185516357422,
    "prediction_norm": 7.894164562225342,
    "timestamp": "2025-07-14T12:33:41.505857"
  },
  {
    "epoch": 229,
    "batch_idx": 160,
    "loss": 0.29164254665374756,
    "learning_rate": 0.0009679990567650903,
    "timestep_mean": 471.0,
    "epsilon_norm": 6.988353729248047,
    "prediction_norm": 5.953579902648926,
    "timestamp": "2025-07-14T12:33:41.510811"
  },
  {
    "epoch": 229,
    "batch_idx": 192,
    "loss": 0.2435624748468399,
    "learning_rate": 0.0009679990567650903,
    "timestep_mean": 590.25,
    "epsilon_norm": 2.9618844985961914,
    "prediction_norm": 2.999514579772949,
    "timestamp": "2025-07-14T12:33:41.515708"
  },
  {
    "epoch": 230,
    "batch_idx": 0,
    "loss": 0.32070979475975037,
    "learning_rate": 0.0009677220154149339,
    "timestep_mean": 472.15625,
    "epsilon_norm": 8.339774131774902,
    "prediction_norm": 7.556278705596924,
    "timestamp": "2025-07-14T12:33:41.520888"
  },
  {
    "epoch": 230,
    "batch_idx": 32,
    "loss": 0.4673251211643219,
    "learning_rate": 0.0009677220154149339,
    "timestep_mean": 454.4375,
    "epsilon_norm": 8.192543983459473,
    "prediction_norm": 6.421334266662598,
    "timestamp": "2025-07-14T12:33:41.526071"
  },
  {
    "epoch": 230,
    "batch_idx": 64,
    "loss": 0.2525249123573303,
    "learning_rate": 0.0009677220154149339,
    "timestep_mean": 487.875,
    "epsilon_norm": 7.519744396209717,
    "prediction_norm": 7.0033860206604,
    "timestamp": "2025-07-14T12:33:41.531181"
  },
  {
    "epoch": 230,
    "batch_idx": 96,
    "loss": 0.3730056881904602,
    "learning_rate": 0.0009677220154149339,
    "timestep_mean": 480.53125,
    "epsilon_norm": 7.61333703994751,
    "prediction_norm": 6.037668228149414,
    "timestamp": "2025-07-14T12:33:41.536327"
  },
  {
    "epoch": 230,
    "batch_idx": 128,
    "loss": 0.17076143622398376,
    "learning_rate": 0.0009677220154149339,
    "timestep_mean": 608.71875,
    "epsilon_norm": 8.190381050109863,
    "prediction_norm": 8.397032737731934,
    "timestamp": "2025-07-14T12:33:41.541420"
  },
  {
    "epoch": 230,
    "batch_idx": 160,
    "loss": 0.25855058431625366,
    "learning_rate": 0.0009677220154149339,
    "timestep_mean": 557.59375,
    "epsilon_norm": 7.56463623046875,
    "prediction_norm": 6.768766403198242,
    "timestamp": "2025-07-14T12:33:41.546510"
  },
  {
    "epoch": 230,
    "batch_idx": 192,
    "loss": 0.17466427385807037,
    "learning_rate": 0.0009677220154149339,
    "timestep_mean": 542.75,
    "epsilon_norm": 3.3283419609069824,
    "prediction_norm": 3.2936489582061768,
    "timestamp": "2025-07-14T12:33:41.551601"
  },
  {
    "epoch": 231,
    "batch_idx": 0,
    "loss": 0.23260074853897095,
    "learning_rate": 0.0009674438200071994,
    "timestep_mean": 513.75,
    "epsilon_norm": 7.273692607879639,
    "prediction_norm": 5.5338454246521,
    "timestamp": "2025-07-14T12:33:41.556793"
  },
  {
    "epoch": 231,
    "batch_idx": 32,
    "loss": 0.28563228249549866,
    "learning_rate": 0.0009674438200071994,
    "timestep_mean": 520.90625,
    "epsilon_norm": 8.177473068237305,
    "prediction_norm": 6.377694129943848,
    "timestamp": "2025-07-14T12:33:41.561943"
  },
  {
    "epoch": 231,
    "batch_idx": 64,
    "loss": 0.2093067765235901,
    "learning_rate": 0.0009674438200071994,
    "timestep_mean": 567.78125,
    "epsilon_norm": 7.918065071105957,
    "prediction_norm": 6.752001762390137,
    "timestamp": "2025-07-14T12:33:41.567142"
  },
  {
    "epoch": 231,
    "batch_idx": 96,
    "loss": 0.27096426486968994,
    "learning_rate": 0.0009674438200071994,
    "timestep_mean": 482.125,
    "epsilon_norm": 6.32857608795166,
    "prediction_norm": 4.679280757904053,
    "timestamp": "2025-07-14T12:33:41.572119"
  },
  {
    "epoch": 231,
    "batch_idx": 128,
    "loss": 0.3430211544036865,
    "learning_rate": 0.0009674438200071994,
    "timestep_mean": 484.6875,
    "epsilon_norm": 6.914484024047852,
    "prediction_norm": 4.8177056312561035,
    "timestamp": "2025-07-14T12:33:41.577221"
  },
  {
    "epoch": 231,
    "batch_idx": 160,
    "loss": 0.2595892548561096,
    "learning_rate": 0.0009674438200071994,
    "timestep_mean": 549.40625,
    "epsilon_norm": 7.560962677001953,
    "prediction_norm": 5.4539923667907715,
    "timestamp": "2025-07-14T12:33:41.582565"
  },
  {
    "epoch": 231,
    "batch_idx": 192,
    "loss": 0.11451806128025055,
    "learning_rate": 0.0009674438200071994,
    "timestep_mean": 470.625,
    "epsilon_norm": 2.9648752212524414,
    "prediction_norm": 2.531780958175659,
    "timestamp": "2025-07-14T12:33:41.587753"
  },
  {
    "epoch": 232,
    "batch_idx": 0,
    "loss": 0.4854207932949066,
    "learning_rate": 0.0009671644712283063,
    "timestep_mean": 374.875,
    "epsilon_norm": 8.67119312286377,
    "prediction_norm": 5.535628318786621,
    "timestamp": "2025-07-14T12:33:41.592697"
  },
  {
    "epoch": 232,
    "batch_idx": 32,
    "loss": 0.34484273195266724,
    "learning_rate": 0.0009671644712283063,
    "timestep_mean": 599.65625,
    "epsilon_norm": 7.411604881286621,
    "prediction_norm": 5.62164306640625,
    "timestamp": "2025-07-14T12:33:41.597853"
  },
  {
    "epoch": 232,
    "batch_idx": 64,
    "loss": 0.13988631963729858,
    "learning_rate": 0.0009671644712283063,
    "timestep_mean": 632.75,
    "epsilon_norm": 6.988178730010986,
    "prediction_norm": 6.827081203460693,
    "timestamp": "2025-07-14T12:33:41.603038"
  },
  {
    "epoch": 232,
    "batch_idx": 96,
    "loss": 0.2941078841686249,
    "learning_rate": 0.0009671644712283063,
    "timestep_mean": 526.5,
    "epsilon_norm": 6.874649524688721,
    "prediction_norm": 5.7863240242004395,
    "timestamp": "2025-07-14T12:33:41.608242"
  },
  {
    "epoch": 232,
    "batch_idx": 128,
    "loss": 0.4634346067905426,
    "learning_rate": 0.0009671644712283063,
    "timestep_mean": 447.09375,
    "epsilon_norm": 8.223084449768066,
    "prediction_norm": 6.734621047973633,
    "timestamp": "2025-07-14T12:33:41.613237"
  },
  {
    "epoch": 232,
    "batch_idx": 160,
    "loss": 0.43872135877609253,
    "learning_rate": 0.0009671644712283063,
    "timestep_mean": 499.3125,
    "epsilon_norm": 8.409534454345703,
    "prediction_norm": 6.756033897399902,
    "timestamp": "2025-07-14T12:33:41.618390"
  },
  {
    "epoch": 232,
    "batch_idx": 192,
    "loss": 0.08491814136505127,
    "learning_rate": 0.0009671644712283063,
    "timestep_mean": 589.25,
    "epsilon_norm": 3.020329475402832,
    "prediction_norm": 2.9772963523864746,
    "timestamp": "2025-07-14T12:33:41.623702"
  },
  {
    "epoch": 233,
    "batch_idx": 0,
    "loss": 0.2506526708602905,
    "learning_rate": 0.0009668839697675198,
    "timestep_mean": 463.0625,
    "epsilon_norm": 7.121356964111328,
    "prediction_norm": 5.810676097869873,
    "timestamp": "2025-07-14T12:33:41.628991"
  },
  {
    "epoch": 233,
    "batch_idx": 32,
    "loss": 0.3896405100822449,
    "learning_rate": 0.0009668839697675198,
    "timestep_mean": 437.96875,
    "epsilon_norm": 7.596075534820557,
    "prediction_norm": 4.864956855773926,
    "timestamp": "2025-07-14T12:33:41.634185"
  },
  {
    "epoch": 233,
    "batch_idx": 64,
    "loss": 0.3696625828742981,
    "learning_rate": 0.0009668839697675198,
    "timestep_mean": 468.8125,
    "epsilon_norm": 7.603157043457031,
    "prediction_norm": 6.1635894775390625,
    "timestamp": "2025-07-14T12:33:41.639241"
  },
  {
    "epoch": 233,
    "batch_idx": 96,
    "loss": 0.296112060546875,
    "learning_rate": 0.0009668839697675198,
    "timestep_mean": 494.0625,
    "epsilon_norm": 7.590395450592041,
    "prediction_norm": 6.185445308685303,
    "timestamp": "2025-07-14T12:33:41.644395"
  },
  {
    "epoch": 233,
    "batch_idx": 128,
    "loss": 0.43297797441482544,
    "learning_rate": 0.0009668839697675198,
    "timestep_mean": 470.125,
    "epsilon_norm": 8.19424819946289,
    "prediction_norm": 5.754701614379883,
    "timestamp": "2025-07-14T12:33:41.649666"
  },
  {
    "epoch": 233,
    "batch_idx": 160,
    "loss": 0.4787116050720215,
    "learning_rate": 0.0009668839697675198,
    "timestep_mean": 416.96875,
    "epsilon_norm": 8.049593925476074,
    "prediction_norm": 5.46985387802124,
    "timestamp": "2025-07-14T12:33:41.654877"
  },
  {
    "epoch": 233,
    "batch_idx": 192,
    "loss": 0.2923770248889923,
    "learning_rate": 0.0009668839697675198,
    "timestep_mean": 398.125,
    "epsilon_norm": 3.01125168800354,
    "prediction_norm": 1.9558638334274292,
    "timestamp": "2025-07-14T12:33:41.660081"
  },
  {
    "epoch": 234,
    "batch_idx": 0,
    "loss": 0.19387446343898773,
    "learning_rate": 0.0009666023163169495,
    "timestep_mean": 485.4375,
    "epsilon_norm": 7.669543743133545,
    "prediction_norm": 6.529651165008545,
    "timestamp": "2025-07-14T12:33:41.665287"
  },
  {
    "epoch": 234,
    "batch_idx": 32,
    "loss": 0.3497943878173828,
    "learning_rate": 0.0009666023163169495,
    "timestep_mean": 517.1875,
    "epsilon_norm": 7.996325492858887,
    "prediction_norm": 6.25592041015625,
    "timestamp": "2025-07-14T12:33:41.670640"
  },
  {
    "epoch": 234,
    "batch_idx": 64,
    "loss": 0.3035372197628021,
    "learning_rate": 0.0009666023163169495,
    "timestep_mean": 474.15625,
    "epsilon_norm": 7.737728118896484,
    "prediction_norm": 6.0469889640808105,
    "timestamp": "2025-07-14T12:33:41.675808"
  },
  {
    "epoch": 234,
    "batch_idx": 96,
    "loss": 0.4238099753856659,
    "learning_rate": 0.0009666023163169495,
    "timestep_mean": 379.1875,
    "epsilon_norm": 6.413434028625488,
    "prediction_norm": 4.0389580726623535,
    "timestamp": "2025-07-14T12:33:41.681037"
  },
  {
    "epoch": 234,
    "batch_idx": 128,
    "loss": 0.342873215675354,
    "learning_rate": 0.0009666023163169495,
    "timestep_mean": 459.5625,
    "epsilon_norm": 7.204655647277832,
    "prediction_norm": 5.163255214691162,
    "timestamp": "2025-07-14T12:33:41.686298"
  },
  {
    "epoch": 234,
    "batch_idx": 160,
    "loss": 0.40627866983413696,
    "learning_rate": 0.0009666023163169495,
    "timestep_mean": 426.1875,
    "epsilon_norm": 7.319559097290039,
    "prediction_norm": 5.706838130950928,
    "timestamp": "2025-07-14T12:33:41.691479"
  },
  {
    "epoch": 234,
    "batch_idx": 192,
    "loss": 0.5037494897842407,
    "learning_rate": 0.0009666023163169495,
    "timestep_mean": 549.0,
    "epsilon_norm": 4.662459373474121,
    "prediction_norm": 3.2237842082977295,
    "timestamp": "2025-07-14T12:33:41.698729"
  },
  {
    "epoch": 235,
    "batch_idx": 0,
    "loss": 0.23998165130615234,
    "learning_rate": 0.0009663195115715473,
    "timestep_mean": 567.8125,
    "epsilon_norm": 7.073649883270264,
    "prediction_norm": 7.082988739013672,
    "timestamp": "2025-07-14T12:33:41.703759"
  },
  {
    "epoch": 235,
    "batch_idx": 32,
    "loss": 0.3822879195213318,
    "learning_rate": 0.0009663195115715473,
    "timestep_mean": 462.46875,
    "epsilon_norm": 7.294985294342041,
    "prediction_norm": 5.384134769439697,
    "timestamp": "2025-07-14T12:33:41.708880"
  },
  {
    "epoch": 235,
    "batch_idx": 64,
    "loss": 0.307561993598938,
    "learning_rate": 0.0009663195115715473,
    "timestep_mean": 514.125,
    "epsilon_norm": 8.039968490600586,
    "prediction_norm": 7.4323906898498535,
    "timestamp": "2025-07-14T12:33:41.713759"
  },
  {
    "epoch": 235,
    "batch_idx": 96,
    "loss": 0.45498934388160706,
    "learning_rate": 0.0009663195115715473,
    "timestep_mean": 463.65625,
    "epsilon_norm": 8.01576042175293,
    "prediction_norm": 5.896045207977295,
    "timestamp": "2025-07-14T12:33:41.718807"
  },
  {
    "epoch": 235,
    "batch_idx": 128,
    "loss": 0.2688899636268616,
    "learning_rate": 0.0009663195115715473,
    "timestep_mean": 538.96875,
    "epsilon_norm": 7.955757141113281,
    "prediction_norm": 6.9770121574401855,
    "timestamp": "2025-07-14T12:33:41.723796"
  },
  {
    "epoch": 235,
    "batch_idx": 160,
    "loss": 0.16752669215202332,
    "learning_rate": 0.0009663195115715473,
    "timestep_mean": 555.5625,
    "epsilon_norm": 7.429235458374023,
    "prediction_norm": 7.081950664520264,
    "timestamp": "2025-07-14T12:33:41.728766"
  },
  {
    "epoch": 235,
    "batch_idx": 192,
    "loss": 0.7121539115905762,
    "learning_rate": 0.0009663195115715473,
    "timestep_mean": 379.875,
    "epsilon_norm": 3.6750400066375732,
    "prediction_norm": 2.992094039916992,
    "timestamp": "2025-07-14T12:33:41.733847"
  },
  {
    "epoch": 236,
    "batch_idx": 0,
    "loss": 0.3021090626716614,
    "learning_rate": 0.0009660355562291056,
    "timestep_mean": 451.625,
    "epsilon_norm": 8.168499946594238,
    "prediction_norm": 7.081724643707275,
    "timestamp": "2025-07-14T12:33:41.738815"
  },
  {
    "epoch": 236,
    "batch_idx": 32,
    "loss": 0.20670709013938904,
    "learning_rate": 0.0009660355562291056,
    "timestep_mean": 514.875,
    "epsilon_norm": 8.802995681762695,
    "prediction_norm": 8.160670280456543,
    "timestamp": "2025-07-14T12:33:41.743790"
  },
  {
    "epoch": 236,
    "batch_idx": 64,
    "loss": 0.1704123169183731,
    "learning_rate": 0.0009660355562291056,
    "timestep_mean": 580.0625,
    "epsilon_norm": 7.945349216461182,
    "prediction_norm": 7.481636047363281,
    "timestamp": "2025-07-14T12:33:41.748809"
  },
  {
    "epoch": 236,
    "batch_idx": 96,
    "loss": 0.2437448352575302,
    "learning_rate": 0.0009660355562291056,
    "timestep_mean": 460.5625,
    "epsilon_norm": 7.236146926879883,
    "prediction_norm": 6.73375129699707,
    "timestamp": "2025-07-14T12:33:41.753985"
  },
  {
    "epoch": 236,
    "batch_idx": 128,
    "loss": 0.3087288737297058,
    "learning_rate": 0.0009660355562291056,
    "timestep_mean": 574.03125,
    "epsilon_norm": 8.356156349182129,
    "prediction_norm": 7.2649149894714355,
    "timestamp": "2025-07-14T12:33:41.758933"
  },
  {
    "epoch": 236,
    "batch_idx": 160,
    "loss": 0.2577207386493683,
    "learning_rate": 0.0009660355562291056,
    "timestep_mean": 542.1875,
    "epsilon_norm": 7.959635257720947,
    "prediction_norm": 6.798083782196045,
    "timestamp": "2025-07-14T12:33:41.764021"
  },
  {
    "epoch": 236,
    "batch_idx": 192,
    "loss": 0.059453707188367844,
    "learning_rate": 0.0009660355562291056,
    "timestep_mean": 598.25,
    "epsilon_norm": 4.774759769439697,
    "prediction_norm": 4.882011890411377,
    "timestamp": "2025-07-14T12:33:41.769026"
  },
  {
    "epoch": 237,
    "batch_idx": 0,
    "loss": 0.4771794080734253,
    "learning_rate": 0.0009657504509902563,
    "timestep_mean": 450.375,
    "epsilon_norm": 9.071758270263672,
    "prediction_norm": 6.456801414489746,
    "timestamp": "2025-07-14T12:33:41.774087"
  },
  {
    "epoch": 237,
    "batch_idx": 32,
    "loss": 0.6278703212738037,
    "learning_rate": 0.0009657504509902563,
    "timestep_mean": 437.0,
    "epsilon_norm": 9.477852821350098,
    "prediction_norm": 6.548376560211182,
    "timestamp": "2025-07-14T12:33:41.779159"
  },
  {
    "epoch": 237,
    "batch_idx": 64,
    "loss": 0.16977474093437195,
    "learning_rate": 0.0009657504509902563,
    "timestep_mean": 649.625,
    "epsilon_norm": 8.195375442504883,
    "prediction_norm": 6.892025470733643,
    "timestamp": "2025-07-14T12:33:41.784129"
  },
  {
    "epoch": 237,
    "batch_idx": 96,
    "loss": 0.3232262134552002,
    "learning_rate": 0.0009657504509902563,
    "timestep_mean": 465.40625,
    "epsilon_norm": 8.185328483581543,
    "prediction_norm": 5.881411552429199,
    "timestamp": "2025-07-14T12:33:41.789118"
  },
  {
    "epoch": 237,
    "batch_idx": 128,
    "loss": 0.2633022665977478,
    "learning_rate": 0.0009657504509902563,
    "timestep_mean": 556.65625,
    "epsilon_norm": 7.574088096618652,
    "prediction_norm": 5.907549858093262,
    "timestamp": "2025-07-14T12:33:41.794181"
  },
  {
    "epoch": 237,
    "batch_idx": 160,
    "loss": 0.23811335861682892,
    "learning_rate": 0.0009657504509902563,
    "timestep_mean": 555.0,
    "epsilon_norm": 8.321549415588379,
    "prediction_norm": 6.67724084854126,
    "timestamp": "2025-07-14T12:33:41.799230"
  },
  {
    "epoch": 237,
    "batch_idx": 192,
    "loss": 0.364858478307724,
    "learning_rate": 0.0009657504509902563,
    "timestep_mean": 449.5,
    "epsilon_norm": 4.862738132476807,
    "prediction_norm": 3.2551655769348145,
    "timestamp": "2025-07-14T12:33:41.804308"
  },
  {
    "epoch": 238,
    "batch_idx": 0,
    "loss": 0.3790520429611206,
    "learning_rate": 0.000965464196558468,
    "timestep_mean": 519.9375,
    "epsilon_norm": 8.617757797241211,
    "prediction_norm": 6.801211833953857,
    "timestamp": "2025-07-14T12:33:41.809267"
  },
  {
    "epoch": 238,
    "batch_idx": 32,
    "loss": 0.20241640508174896,
    "learning_rate": 0.000965464196558468,
    "timestep_mean": 550.28125,
    "epsilon_norm": 6.219956398010254,
    "prediction_norm": 5.435702800750732,
    "timestamp": "2025-07-14T12:33:41.814327"
  },
  {
    "epoch": 238,
    "batch_idx": 64,
    "loss": 0.31212615966796875,
    "learning_rate": 0.000965464196558468,
    "timestep_mean": 509.9375,
    "epsilon_norm": 7.718173027038574,
    "prediction_norm": 6.883274555206299,
    "timestamp": "2025-07-14T12:33:41.819359"
  },
  {
    "epoch": 238,
    "batch_idx": 96,
    "loss": 0.393383264541626,
    "learning_rate": 0.000965464196558468,
    "timestep_mean": 514.125,
    "epsilon_norm": 8.519023895263672,
    "prediction_norm": 7.279860019683838,
    "timestamp": "2025-07-14T12:33:41.824562"
  },
  {
    "epoch": 238,
    "batch_idx": 128,
    "loss": 0.4138905107975006,
    "learning_rate": 0.000965464196558468,
    "timestep_mean": 422.65625,
    "epsilon_norm": 6.6394829750061035,
    "prediction_norm": 5.368218898773193,
    "timestamp": "2025-07-14T12:33:41.829859"
  },
  {
    "epoch": 238,
    "batch_idx": 160,
    "loss": 0.28241944313049316,
    "learning_rate": 0.000965464196558468,
    "timestep_mean": 511.625,
    "epsilon_norm": 7.361552715301514,
    "prediction_norm": 7.022766590118408,
    "timestamp": "2025-07-14T12:33:41.834969"
  },
  {
    "epoch": 238,
    "batch_idx": 192,
    "loss": 0.4271950125694275,
    "learning_rate": 0.000965464196558468,
    "timestep_mean": 449.5,
    "epsilon_norm": 4.320983409881592,
    "prediction_norm": 3.6300594806671143,
    "timestamp": "2025-07-14T12:33:41.840134"
  },
  {
    "epoch": 239,
    "batch_idx": 0,
    "loss": 0.28934216499328613,
    "learning_rate": 0.0009651767936400453,
    "timestep_mean": 516.53125,
    "epsilon_norm": 7.3245744705200195,
    "prediction_norm": 6.439404487609863,
    "timestamp": "2025-07-14T12:33:41.845372"
  },
  {
    "epoch": 239,
    "batch_idx": 32,
    "loss": 0.3412807881832123,
    "learning_rate": 0.0009651767936400453,
    "timestep_mean": 565.6875,
    "epsilon_norm": 8.434442520141602,
    "prediction_norm": 6.920770645141602,
    "timestamp": "2025-07-14T12:33:41.850699"
  },
  {
    "epoch": 239,
    "batch_idx": 64,
    "loss": 0.22116462886333466,
    "learning_rate": 0.0009651767936400453,
    "timestep_mean": 550.625,
    "epsilon_norm": 7.833651542663574,
    "prediction_norm": 6.65255880355835,
    "timestamp": "2025-07-14T12:33:41.855799"
  },
  {
    "epoch": 239,
    "batch_idx": 96,
    "loss": 0.35866376757621765,
    "learning_rate": 0.0009651767936400453,
    "timestep_mean": 510.15625,
    "epsilon_norm": 8.295490264892578,
    "prediction_norm": 6.374754428863525,
    "timestamp": "2025-07-14T12:33:41.861047"
  },
  {
    "epoch": 239,
    "batch_idx": 128,
    "loss": 0.22023624181747437,
    "learning_rate": 0.0009651767936400453,
    "timestep_mean": 540.46875,
    "epsilon_norm": 8.28872299194336,
    "prediction_norm": 7.50404691696167,
    "timestamp": "2025-07-14T12:33:41.866230"
  },
  {
    "epoch": 239,
    "batch_idx": 160,
    "loss": 0.3078077733516693,
    "learning_rate": 0.0009651767936400453,
    "timestep_mean": 490.8125,
    "epsilon_norm": 7.6109137535095215,
    "prediction_norm": 6.387338638305664,
    "timestamp": "2025-07-14T12:33:41.871519"
  },
  {
    "epoch": 239,
    "batch_idx": 192,
    "loss": 0.41442522406578064,
    "learning_rate": 0.0009651767936400453,
    "timestep_mean": 479.25,
    "epsilon_norm": 3.5024943351745605,
    "prediction_norm": 2.367952823638916,
    "timestamp": "2025-07-14T12:33:41.876655"
  },
  {
    "epoch": 240,
    "batch_idx": 0,
    "loss": 0.3579743802547455,
    "learning_rate": 0.0009648882429441259,
    "timestep_mean": 432.4375,
    "epsilon_norm": 8.091290473937988,
    "prediction_norm": 6.014742374420166,
    "timestamp": "2025-07-14T12:33:41.881854"
  },
  {
    "epoch": 240,
    "batch_idx": 32,
    "loss": 0.3062225580215454,
    "learning_rate": 0.0009648882429441259,
    "timestep_mean": 438.875,
    "epsilon_norm": 7.177178859710693,
    "prediction_norm": 5.440199375152588,
    "timestamp": "2025-07-14T12:33:41.887087"
  },
  {
    "epoch": 240,
    "batch_idx": 64,
    "loss": 0.5374053716659546,
    "learning_rate": 0.0009648882429441259,
    "timestep_mean": 528.0,
    "epsilon_norm": 8.54964542388916,
    "prediction_norm": 6.271337032318115,
    "timestamp": "2025-07-14T12:33:41.892503"
  },
  {
    "epoch": 240,
    "batch_idx": 96,
    "loss": 0.6733458042144775,
    "learning_rate": 0.0009648882429441259,
    "timestep_mean": 449.96875,
    "epsilon_norm": 10.535775184631348,
    "prediction_norm": 8.226150512695312,
    "timestamp": "2025-07-14T12:33:41.897698"
  },
  {
    "epoch": 240,
    "batch_idx": 128,
    "loss": 0.35015803575515747,
    "learning_rate": 0.0009648882429441259,
    "timestep_mean": 503.1875,
    "epsilon_norm": 7.859155654907227,
    "prediction_norm": 6.098453044891357,
    "timestamp": "2025-07-14T12:33:41.902901"
  },
  {
    "epoch": 240,
    "batch_idx": 160,
    "loss": 0.22466734051704407,
    "learning_rate": 0.0009648882429441259,
    "timestep_mean": 473.625,
    "epsilon_norm": 6.765945911407471,
    "prediction_norm": 5.475675582885742,
    "timestamp": "2025-07-14T12:33:41.908120"
  },
  {
    "epoch": 240,
    "batch_idx": 192,
    "loss": 0.1502012461423874,
    "learning_rate": 0.0009648882429441259,
    "timestep_mean": 610.25,
    "epsilon_norm": 4.603833198547363,
    "prediction_norm": 4.300489902496338,
    "timestamp": "2025-07-14T12:33:41.913223"
  },
  {
    "epoch": 241,
    "batch_idx": 0,
    "loss": 0.2288193255662918,
    "learning_rate": 0.0009645985451826804,
    "timestep_mean": 527.34375,
    "epsilon_norm": 8.825468063354492,
    "prediction_norm": 7.410528659820557,
    "timestamp": "2025-07-14T12:33:41.918401"
  },
  {
    "epoch": 241,
    "batch_idx": 32,
    "loss": 0.3537129759788513,
    "learning_rate": 0.0009645985451826804,
    "timestep_mean": 595.96875,
    "epsilon_norm": 8.388753890991211,
    "prediction_norm": 7.162772178649902,
    "timestamp": "2025-07-14T12:33:41.923651"
  },
  {
    "epoch": 241,
    "batch_idx": 64,
    "loss": 0.21175053715705872,
    "learning_rate": 0.0009645985451826804,
    "timestep_mean": 587.53125,
    "epsilon_norm": 6.914588928222656,
    "prediction_norm": 6.137802600860596,
    "timestamp": "2025-07-14T12:33:41.928980"
  },
  {
    "epoch": 241,
    "batch_idx": 96,
    "loss": 0.5234667062759399,
    "learning_rate": 0.0009645985451826804,
    "timestep_mean": 401.40625,
    "epsilon_norm": 9.016725540161133,
    "prediction_norm": 6.216387748718262,
    "timestamp": "2025-07-14T12:33:41.934199"
  },
  {
    "epoch": 241,
    "batch_idx": 128,
    "loss": 0.518162190914154,
    "learning_rate": 0.0009645985451826804,
    "timestep_mean": 460.25,
    "epsilon_norm": 8.458207130432129,
    "prediction_norm": 6.409230709075928,
    "timestamp": "2025-07-14T12:33:41.939427"
  },
  {
    "epoch": 241,
    "batch_idx": 160,
    "loss": 0.35352352261543274,
    "learning_rate": 0.0009645985451826804,
    "timestep_mean": 497.71875,
    "epsilon_norm": 8.585700988769531,
    "prediction_norm": 7.707237243652344,
    "timestamp": "2025-07-14T12:33:41.944554"
  },
  {
    "epoch": 241,
    "batch_idx": 192,
    "loss": 0.4197842478752136,
    "learning_rate": 0.0009645985451826804,
    "timestep_mean": 595.875,
    "epsilon_norm": 4.817616939544678,
    "prediction_norm": 3.8580596446990967,
    "timestamp": "2025-07-14T12:33:41.949717"
  },
  {
    "epoch": 242,
    "batch_idx": 0,
    "loss": 0.4003905653953552,
    "learning_rate": 0.0009643077010705089,
    "timestep_mean": 460.28125,
    "epsilon_norm": 8.059843063354492,
    "prediction_norm": 6.443035125732422,
    "timestamp": "2025-07-14T12:33:41.954930"
  },
  {
    "epoch": 242,
    "batch_idx": 32,
    "loss": 0.3404315710067749,
    "learning_rate": 0.0009643077010705089,
    "timestep_mean": 466.875,
    "epsilon_norm": 8.502205848693848,
    "prediction_norm": 7.036999225616455,
    "timestamp": "2025-07-14T12:33:41.960154"
  },
  {
    "epoch": 242,
    "batch_idx": 64,
    "loss": 0.3607027530670166,
    "learning_rate": 0.0009643077010705089,
    "timestep_mean": 495.96875,
    "epsilon_norm": 8.041072845458984,
    "prediction_norm": 7.7498860359191895,
    "timestamp": "2025-07-14T12:33:41.965320"
  },
  {
    "epoch": 242,
    "batch_idx": 96,
    "loss": 0.2856985926628113,
    "learning_rate": 0.0009643077010705089,
    "timestep_mean": 569.03125,
    "epsilon_norm": 8.138517379760742,
    "prediction_norm": 7.258949279785156,
    "timestamp": "2025-07-14T12:33:41.970499"
  },
  {
    "epoch": 242,
    "batch_idx": 128,
    "loss": 0.12767185270786285,
    "learning_rate": 0.0009643077010705089,
    "timestep_mean": 585.53125,
    "epsilon_norm": 8.005585670471191,
    "prediction_norm": 7.79371452331543,
    "timestamp": "2025-07-14T12:33:41.975695"
  },
  {
    "epoch": 242,
    "batch_idx": 160,
    "loss": 0.3153156042098999,
    "learning_rate": 0.0009643077010705089,
    "timestep_mean": 489.21875,
    "epsilon_norm": 7.613100051879883,
    "prediction_norm": 6.5951666831970215,
    "timestamp": "2025-07-14T12:33:41.980835"
  },
  {
    "epoch": 242,
    "batch_idx": 192,
    "loss": 0.10073557496070862,
    "learning_rate": 0.0009643077010705089,
    "timestep_mean": 671.5,
    "epsilon_norm": 4.333009243011475,
    "prediction_norm": 4.223378658294678,
    "timestamp": "2025-07-14T12:33:41.986075"
  },
  {
    "epoch": 243,
    "batch_idx": 0,
    "loss": 0.45092183351516724,
    "learning_rate": 0.0009640157113252406,
    "timestep_mean": 409.25,
    "epsilon_norm": 9.252811431884766,
    "prediction_norm": 7.087477207183838,
    "timestamp": "2025-07-14T12:33:41.991464"
  },
  {
    "epoch": 243,
    "batch_idx": 32,
    "loss": 0.25208598375320435,
    "learning_rate": 0.0009640157113252406,
    "timestep_mean": 495.75,
    "epsilon_norm": 8.00679874420166,
    "prediction_norm": 6.90116548538208,
    "timestamp": "2025-07-14T12:33:41.996600"
  },
  {
    "epoch": 243,
    "batch_idx": 64,
    "loss": 0.278482586145401,
    "learning_rate": 0.0009640157113252406,
    "timestep_mean": 560.9375,
    "epsilon_norm": 8.537260055541992,
    "prediction_norm": 6.891568660736084,
    "timestamp": "2025-07-14T12:33:42.001781"
  },
  {
    "epoch": 243,
    "batch_idx": 96,
    "loss": 0.24043771624565125,
    "learning_rate": 0.0009640157113252406,
    "timestep_mean": 562.09375,
    "epsilon_norm": 7.701353073120117,
    "prediction_norm": 6.428760051727295,
    "timestamp": "2025-07-14T12:33:42.007154"
  },
  {
    "epoch": 243,
    "batch_idx": 128,
    "loss": 0.24219736456871033,
    "learning_rate": 0.0009640157113252406,
    "timestep_mean": 409.34375,
    "epsilon_norm": 6.66867733001709,
    "prediction_norm": 5.492269515991211,
    "timestamp": "2025-07-14T12:33:42.012407"
  },
  {
    "epoch": 243,
    "batch_idx": 160,
    "loss": 0.20694860816001892,
    "learning_rate": 0.0009640157113252406,
    "timestep_mean": 583.625,
    "epsilon_norm": 8.692288398742676,
    "prediction_norm": 8.118277549743652,
    "timestamp": "2025-07-14T12:33:42.017568"
  },
  {
    "epoch": 243,
    "batch_idx": 192,
    "loss": 0.18481570482254028,
    "learning_rate": 0.0009640157113252406,
    "timestep_mean": 588.875,
    "epsilon_norm": 4.753277778625488,
    "prediction_norm": 4.391918659210205,
    "timestamp": "2025-07-14T12:33:42.022747"
  },
  {
    "epoch": 244,
    "batch_idx": 0,
    "loss": 0.20528529584407806,
    "learning_rate": 0.0009637225766673308,
    "timestep_mean": 468.40625,
    "epsilon_norm": 7.738330841064453,
    "prediction_norm": 6.947273254394531,
    "timestamp": "2025-07-14T12:33:42.027947"
  },
  {
    "epoch": 244,
    "batch_idx": 32,
    "loss": 0.2696126699447632,
    "learning_rate": 0.0009637225766673308,
    "timestep_mean": 548.125,
    "epsilon_norm": 8.345588684082031,
    "prediction_norm": 7.351987838745117,
    "timestamp": "2025-07-14T12:33:42.033245"
  },
  {
    "epoch": 244,
    "batch_idx": 64,
    "loss": 0.26293352246284485,
    "learning_rate": 0.0009637225766673308,
    "timestep_mean": 565.375,
    "epsilon_norm": 8.217650413513184,
    "prediction_norm": 8.016704559326172,
    "timestamp": "2025-07-14T12:33:42.038426"
  },
  {
    "epoch": 244,
    "batch_idx": 96,
    "loss": 0.15120495855808258,
    "learning_rate": 0.0009637225766673308,
    "timestep_mean": 585.59375,
    "epsilon_norm": 8.413583755493164,
    "prediction_norm": 7.447567462921143,
    "timestamp": "2025-07-14T12:33:42.043694"
  },
  {
    "epoch": 244,
    "batch_idx": 128,
    "loss": 0.4659184217453003,
    "learning_rate": 0.0009637225766673308,
    "timestep_mean": 481.5625,
    "epsilon_norm": 7.548017978668213,
    "prediction_norm": 5.430371284484863,
    "timestamp": "2025-07-14T12:33:42.048905"
  },
  {
    "epoch": 244,
    "batch_idx": 160,
    "loss": 0.49863603711128235,
    "learning_rate": 0.0009637225766673308,
    "timestep_mean": 491.59375,
    "epsilon_norm": 8.450122833251953,
    "prediction_norm": 5.741940498352051,
    "timestamp": "2025-07-14T12:33:42.054146"
  },
  {
    "epoch": 244,
    "batch_idx": 192,
    "loss": 0.5305977463722229,
    "learning_rate": 0.0009637225766673308,
    "timestep_mean": 582.5,
    "epsilon_norm": 4.633618354797363,
    "prediction_norm": 3.761681079864502,
    "timestamp": "2025-07-14T12:33:42.059277"
  },
  {
    "epoch": 245,
    "batch_idx": 0,
    "loss": 0.25180941820144653,
    "learning_rate": 0.0009634282978200606,
    "timestep_mean": 526.03125,
    "epsilon_norm": 8.22990894317627,
    "prediction_norm": 6.720467567443848,
    "timestamp": "2025-07-14T12:33:42.064458"
  },
  {
    "epoch": 245,
    "batch_idx": 32,
    "loss": 0.22373375296592712,
    "learning_rate": 0.0009634282978200606,
    "timestep_mean": 591.15625,
    "epsilon_norm": 7.1331610679626465,
    "prediction_norm": 5.763077735900879,
    "timestamp": "2025-07-14T12:33:42.069681"
  },
  {
    "epoch": 245,
    "batch_idx": 64,
    "loss": 0.41676467657089233,
    "learning_rate": 0.0009634282978200606,
    "timestep_mean": 428.40625,
    "epsilon_norm": 7.696876525878906,
    "prediction_norm": 6.110773086547852,
    "timestamp": "2025-07-14T12:33:42.074846"
  },
  {
    "epoch": 245,
    "batch_idx": 96,
    "loss": 0.6368671655654907,
    "learning_rate": 0.0009634282978200606,
    "timestep_mean": 405.875,
    "epsilon_norm": 8.36279582977295,
    "prediction_norm": 5.579293251037598,
    "timestamp": "2025-07-14T12:33:42.080016"
  },
  {
    "epoch": 245,
    "batch_idx": 128,
    "loss": 0.2766377925872803,
    "learning_rate": 0.0009634282978200606,
    "timestep_mean": 462.3125,
    "epsilon_norm": 7.123586177825928,
    "prediction_norm": 5.633825778961182,
    "timestamp": "2025-07-14T12:33:42.085285"
  },
  {
    "epoch": 245,
    "batch_idx": 160,
    "loss": 0.31059175729751587,
    "learning_rate": 0.0009634282978200606,
    "timestep_mean": 575.84375,
    "epsilon_norm": 8.293327331542969,
    "prediction_norm": 6.9235076904296875,
    "timestamp": "2025-07-14T12:33:42.090534"
  },
  {
    "epoch": 245,
    "batch_idx": 192,
    "loss": 0.5907585024833679,
    "learning_rate": 0.0009634282978200606,
    "timestep_mean": 443.5,
    "epsilon_norm": 4.289231300354004,
    "prediction_norm": 2.9471473693847656,
    "timestamp": "2025-07-14T12:33:42.095607"
  },
  {
    "epoch": 246,
    "batch_idx": 0,
    "loss": 0.14519371092319489,
    "learning_rate": 0.0009631328755095337,
    "timestep_mean": 546.78125,
    "epsilon_norm": 6.606520175933838,
    "prediction_norm": 5.700242519378662,
    "timestamp": "2025-07-14T12:33:42.100744"
  },
  {
    "epoch": 246,
    "batch_idx": 32,
    "loss": 0.3366389274597168,
    "learning_rate": 0.0009631328755095337,
    "timestep_mean": 508.625,
    "epsilon_norm": 7.62266206741333,
    "prediction_norm": 5.480016708374023,
    "timestamp": "2025-07-14T12:33:42.105887"
  },
  {
    "epoch": 246,
    "batch_idx": 64,
    "loss": 0.5093318819999695,
    "learning_rate": 0.0009631328755095337,
    "timestep_mean": 486.0625,
    "epsilon_norm": 8.854898452758789,
    "prediction_norm": 6.9354753494262695,
    "timestamp": "2025-07-14T12:33:42.111078"
  },
  {
    "epoch": 246,
    "batch_idx": 96,
    "loss": 0.4454324245452881,
    "learning_rate": 0.0009631328755095337,
    "timestep_mean": 510.75,
    "epsilon_norm": 7.5648932456970215,
    "prediction_norm": 5.270310401916504,
    "timestamp": "2025-07-14T12:33:42.116223"
  },
  {
    "epoch": 246,
    "batch_idx": 128,
    "loss": 0.42435452342033386,
    "learning_rate": 0.0009631328755095337,
    "timestep_mean": 460.21875,
    "epsilon_norm": 9.08584976196289,
    "prediction_norm": 7.117251873016357,
    "timestamp": "2025-07-14T12:33:42.121529"
  },
  {
    "epoch": 246,
    "batch_idx": 160,
    "loss": 0.2733459770679474,
    "learning_rate": 0.0009631328755095337,
    "timestep_mean": 494.21875,
    "epsilon_norm": 7.858903408050537,
    "prediction_norm": 6.3923163414001465,
    "timestamp": "2025-07-14T12:33:42.126714"
  },
  {
    "epoch": 246,
    "batch_idx": 192,
    "loss": 0.6754292249679565,
    "learning_rate": 0.0009631328755095337,
    "timestep_mean": 470.625,
    "epsilon_norm": 4.70768404006958,
    "prediction_norm": 3.3363354206085205,
    "timestamp": "2025-07-14T12:33:42.131853"
  },
  {
    "epoch": 247,
    "batch_idx": 0,
    "loss": 0.15099528431892395,
    "learning_rate": 0.000962836310464675,
    "timestep_mean": 516.8125,
    "epsilon_norm": 6.400698661804199,
    "prediction_norm": 6.232865333557129,
    "timestamp": "2025-07-14T12:33:42.137108"
  },
  {
    "epoch": 247,
    "batch_idx": 32,
    "loss": 0.32260236144065857,
    "learning_rate": 0.000962836310464675,
    "timestep_mean": 459.46875,
    "epsilon_norm": 7.019588947296143,
    "prediction_norm": 5.6025190353393555,
    "timestamp": "2025-07-14T12:33:42.142427"
  },
  {
    "epoch": 247,
    "batch_idx": 64,
    "loss": 0.4027959406375885,
    "learning_rate": 0.000962836310464675,
    "timestep_mean": 536.75,
    "epsilon_norm": 8.887876510620117,
    "prediction_norm": 7.286297798156738,
    "timestamp": "2025-07-14T12:33:42.147663"
  },
  {
    "epoch": 247,
    "batch_idx": 96,
    "loss": 0.3292312026023865,
    "learning_rate": 0.000962836310464675,
    "timestep_mean": 508.5,
    "epsilon_norm": 7.444075107574463,
    "prediction_norm": 6.395957946777344,
    "timestamp": "2025-07-14T12:33:42.152866"
  },
  {
    "epoch": 247,
    "batch_idx": 128,
    "loss": 0.22006356716156006,
    "learning_rate": 0.000962836310464675,
    "timestep_mean": 546.8125,
    "epsilon_norm": 7.645473957061768,
    "prediction_norm": 6.864709377288818,
    "timestamp": "2025-07-14T12:33:42.158077"
  },
  {
    "epoch": 247,
    "batch_idx": 160,
    "loss": 0.32121965289115906,
    "learning_rate": 0.000962836310464675,
    "timestep_mean": 467.09375,
    "epsilon_norm": 7.813013553619385,
    "prediction_norm": 6.796579837799072,
    "timestamp": "2025-07-14T12:33:42.163261"
  },
  {
    "epoch": 247,
    "batch_idx": 192,
    "loss": 0.6319661140441895,
    "learning_rate": 0.000962836310464675,
    "timestep_mean": 572.25,
    "epsilon_norm": 4.844314098358154,
    "prediction_norm": 3.408491373062134,
    "timestamp": "2025-07-14T12:33:42.168332"
  },
  {
    "epoch": 248,
    "batch_idx": 0,
    "loss": 0.445720374584198,
    "learning_rate": 0.0009625386034172292,
    "timestep_mean": 442.75,
    "epsilon_norm": 7.876131057739258,
    "prediction_norm": 5.584854602813721,
    "timestamp": "2025-07-14T12:33:42.173577"
  },
  {
    "epoch": 248,
    "batch_idx": 32,
    "loss": 0.2836914658546448,
    "learning_rate": 0.0009625386034172292,
    "timestep_mean": 530.71875,
    "epsilon_norm": 8.043829917907715,
    "prediction_norm": 6.677227020263672,
    "timestamp": "2025-07-14T12:33:42.178775"
  },
  {
    "epoch": 248,
    "batch_idx": 64,
    "loss": 0.5104978084564209,
    "learning_rate": 0.0009625386034172292,
    "timestep_mean": 364.46875,
    "epsilon_norm": 8.75587272644043,
    "prediction_norm": 5.884483814239502,
    "timestamp": "2025-07-14T12:33:42.184095"
  },
  {
    "epoch": 248,
    "batch_idx": 96,
    "loss": 0.4536994397640228,
    "learning_rate": 0.0009625386034172292,
    "timestep_mean": 472.21875,
    "epsilon_norm": 7.923215389251709,
    "prediction_norm": 5.477219581604004,
    "timestamp": "2025-07-14T12:33:42.189324"
  },
  {
    "epoch": 248,
    "batch_idx": 128,
    "loss": 0.3465518355369568,
    "learning_rate": 0.0009625386034172292,
    "timestep_mean": 452.9375,
    "epsilon_norm": 6.68222713470459,
    "prediction_norm": 5.601228713989258,
    "timestamp": "2025-07-14T12:33:42.194520"
  },
  {
    "epoch": 248,
    "batch_idx": 160,
    "loss": 0.2786481976509094,
    "learning_rate": 0.0009625386034172292,
    "timestep_mean": 544.15625,
    "epsilon_norm": 8.239627838134766,
    "prediction_norm": 7.209095001220703,
    "timestamp": "2025-07-14T12:33:42.199705"
  },
  {
    "epoch": 248,
    "batch_idx": 192,
    "loss": 0.5800841450691223,
    "learning_rate": 0.0009625386034172292,
    "timestep_mean": 330.5,
    "epsilon_norm": 3.5927679538726807,
    "prediction_norm": 2.556485891342163,
    "timestamp": "2025-07-14T12:33:42.204913"
  },
  {
    "epoch": 249,
    "batch_idx": 0,
    "loss": 0.25866106152534485,
    "learning_rate": 0.0009622397551017595,
    "timestep_mean": 484.0625,
    "epsilon_norm": 7.060512065887451,
    "prediction_norm": 6.048542022705078,
    "timestamp": "2025-07-14T12:33:42.210146"
  },
  {
    "epoch": 249,
    "batch_idx": 32,
    "loss": 0.3321267366409302,
    "learning_rate": 0.0009622397551017595,
    "timestep_mean": 567.0625,
    "epsilon_norm": 8.572933197021484,
    "prediction_norm": 7.908308029174805,
    "timestamp": "2025-07-14T12:33:42.215397"
  },
  {
    "epoch": 249,
    "batch_idx": 64,
    "loss": 0.3370864987373352,
    "learning_rate": 0.0009622397551017595,
    "timestep_mean": 552.40625,
    "epsilon_norm": 8.329862594604492,
    "prediction_norm": 6.584588527679443,
    "timestamp": "2025-07-14T12:33:42.220703"
  },
  {
    "epoch": 249,
    "batch_idx": 96,
    "loss": 0.3087378740310669,
    "learning_rate": 0.0009622397551017595,
    "timestep_mean": 451.84375,
    "epsilon_norm": 7.369382858276367,
    "prediction_norm": 5.733143329620361,
    "timestamp": "2025-07-14T12:33:42.225906"
  },
  {
    "epoch": 249,
    "batch_idx": 128,
    "loss": 0.5199850797653198,
    "learning_rate": 0.0009622397551017595,
    "timestep_mean": 415.75,
    "epsilon_norm": 8.821061134338379,
    "prediction_norm": 5.564218997955322,
    "timestamp": "2025-07-14T12:33:42.230986"
  },
  {
    "epoch": 249,
    "batch_idx": 160,
    "loss": 0.3504268229007721,
    "learning_rate": 0.0009622397551017595,
    "timestep_mean": 442.46875,
    "epsilon_norm": 8.436003684997559,
    "prediction_norm": 6.6265177726745605,
    "timestamp": "2025-07-14T12:33:42.236253"
  },
  {
    "epoch": 249,
    "batch_idx": 192,
    "loss": 0.36788952350616455,
    "learning_rate": 0.0009622397551017595,
    "timestep_mean": 560.25,
    "epsilon_norm": 3.953789472579956,
    "prediction_norm": 3.0893688201904297,
    "timestamp": "2025-07-14T12:33:42.241352"
  },
  {
    "epoch": 250,
    "batch_idx": 0,
    "loss": 0.2295597344636917,
    "learning_rate": 0.0009619397662556437,
    "timestep_mean": 500.0,
    "epsilon_norm": 7.21528959274292,
    "prediction_norm": 6.1365251541137695,
    "timestamp": "2025-07-14T12:33:42.246507"
  },
  {
    "epoch": 250,
    "batch_idx": 32,
    "loss": 0.4242497682571411,
    "learning_rate": 0.0009619397662556437,
    "timestep_mean": 458.90625,
    "epsilon_norm": 9.145475387573242,
    "prediction_norm": 6.8108062744140625,
    "timestamp": "2025-07-14T12:33:42.251693"
  },
  {
    "epoch": 250,
    "batch_idx": 64,
    "loss": 0.2971702814102173,
    "learning_rate": 0.0009619397662556437,
    "timestep_mean": 466.65625,
    "epsilon_norm": 7.996716022491455,
    "prediction_norm": 6.932270526885986,
    "timestamp": "2025-07-14T12:33:42.256847"
  },
  {
    "epoch": 250,
    "batch_idx": 96,
    "loss": 0.24050991237163544,
    "learning_rate": 0.0009619397662556437,
    "timestep_mean": 535.5625,
    "epsilon_norm": 6.7403564453125,
    "prediction_norm": 6.226027965545654,
    "timestamp": "2025-07-14T12:33:42.262120"
  },
  {
    "epoch": 250,
    "batch_idx": 128,
    "loss": 0.5552016496658325,
    "learning_rate": 0.0009619397662556437,
    "timestep_mean": 330.875,
    "epsilon_norm": 8.304445266723633,
    "prediction_norm": 6.079875946044922,
    "timestamp": "2025-07-14T12:33:42.267387"
  },
  {
    "epoch": 250,
    "batch_idx": 160,
    "loss": 0.36839771270751953,
    "learning_rate": 0.0009619397662556437,
    "timestep_mean": 470.125,
    "epsilon_norm": 7.147552013397217,
    "prediction_norm": 5.983415126800537,
    "timestamp": "2025-07-14T12:33:42.272543"
  },
  {
    "epoch": 250,
    "batch_idx": 192,
    "loss": 0.3401331305503845,
    "learning_rate": 0.0009619397662556437,
    "timestep_mean": 558.5,
    "epsilon_norm": 3.8688273429870605,
    "prediction_norm": 3.110301971435547,
    "timestamp": "2025-07-14T12:33:42.277726"
  },
  {
    "epoch": 251,
    "batch_idx": 0,
    "loss": 0.2593400478363037,
    "learning_rate": 0.0009616386376190747,
    "timestep_mean": 532.28125,
    "epsilon_norm": 8.189072608947754,
    "prediction_norm": 7.070341110229492,
    "timestamp": "2025-07-14T12:33:42.282826"
  },
  {
    "epoch": 251,
    "batch_idx": 32,
    "loss": 0.327873170375824,
    "learning_rate": 0.0009616386376190747,
    "timestep_mean": 532.25,
    "epsilon_norm": 7.164290904998779,
    "prediction_norm": 5.360599994659424,
    "timestamp": "2025-07-14T12:33:42.287718"
  },
  {
    "epoch": 251,
    "batch_idx": 64,
    "loss": 0.4580843448638916,
    "learning_rate": 0.0009616386376190747,
    "timestep_mean": 474.21875,
    "epsilon_norm": 8.231637954711914,
    "prediction_norm": 5.628023624420166,
    "timestamp": "2025-07-14T12:33:42.292587"
  },
  {
    "epoch": 251,
    "batch_idx": 96,
    "loss": 0.3280595541000366,
    "learning_rate": 0.0009616386376190747,
    "timestep_mean": 506.84375,
    "epsilon_norm": 7.080455303192139,
    "prediction_norm": 4.837930679321289,
    "timestamp": "2025-07-14T12:33:42.302010"
  },
  {
    "epoch": 251,
    "batch_idx": 128,
    "loss": 0.4374769926071167,
    "learning_rate": 0.0009616386376190747,
    "timestep_mean": 425.28125,
    "epsilon_norm": 8.264566421508789,
    "prediction_norm": 6.238140106201172,
    "timestamp": "2025-07-14T12:33:42.307011"
  },
  {
    "epoch": 251,
    "batch_idx": 160,
    "loss": 0.2611355185508728,
    "learning_rate": 0.0009616386376190747,
    "timestep_mean": 522.34375,
    "epsilon_norm": 8.006803512573242,
    "prediction_norm": 7.359671592712402,
    "timestamp": "2025-07-14T12:33:42.312045"
  },
  {
    "epoch": 251,
    "batch_idx": 192,
    "loss": 0.4835083782672882,
    "learning_rate": 0.0009616386376190747,
    "timestep_mean": 283.0,
    "epsilon_norm": 3.9267594814300537,
    "prediction_norm": 2.7675280570983887,
    "timestamp": "2025-07-14T12:33:42.317044"
  },
  {
    "epoch": 252,
    "batch_idx": 0,
    "loss": 0.4187352657318115,
    "learning_rate": 0.0009613363699350578,
    "timestep_mean": 487.8125,
    "epsilon_norm": 8.212860107421875,
    "prediction_norm": 6.105140209197998,
    "timestamp": "2025-07-14T12:33:42.321875"
  },
  {
    "epoch": 252,
    "batch_idx": 32,
    "loss": 0.2520124316215515,
    "learning_rate": 0.0009613363699350578,
    "timestep_mean": 472.8125,
    "epsilon_norm": 7.275944232940674,
    "prediction_norm": 6.1948699951171875,
    "timestamp": "2025-07-14T12:33:42.326844"
  },
  {
    "epoch": 252,
    "batch_idx": 64,
    "loss": 0.17558452486991882,
    "learning_rate": 0.0009613363699350578,
    "timestep_mean": 575.5625,
    "epsilon_norm": 7.579800128936768,
    "prediction_norm": 7.214906692504883,
    "timestamp": "2025-07-14T12:33:42.331828"
  },
  {
    "epoch": 252,
    "batch_idx": 96,
    "loss": 0.3485753536224365,
    "learning_rate": 0.0009613363699350578,
    "timestep_mean": 473.53125,
    "epsilon_norm": 7.32499361038208,
    "prediction_norm": 6.419917583465576,
    "timestamp": "2025-07-14T12:33:42.336758"
  },
  {
    "epoch": 252,
    "batch_idx": 128,
    "loss": 0.2843817472457886,
    "learning_rate": 0.0009613363699350578,
    "timestep_mean": 560.0,
    "epsilon_norm": 8.528117179870605,
    "prediction_norm": 7.974977970123291,
    "timestamp": "2025-07-14T12:33:42.341687"
  },
  {
    "epoch": 252,
    "batch_idx": 160,
    "loss": 0.2279641032218933,
    "learning_rate": 0.0009613363699350578,
    "timestep_mean": 629.75,
    "epsilon_norm": 8.546510696411133,
    "prediction_norm": 8.48138427734375,
    "timestamp": "2025-07-14T12:33:42.346578"
  },
  {
    "epoch": 252,
    "batch_idx": 192,
    "loss": 0.16777443885803223,
    "learning_rate": 0.0009613363699350578,
    "timestep_mean": 674.375,
    "epsilon_norm": 3.7160706520080566,
    "prediction_norm": 3.7104201316833496,
    "timestamp": "2025-07-14T12:33:42.351401"
  },
  {
    "epoch": 253,
    "batch_idx": 0,
    "loss": 0.27742433547973633,
    "learning_rate": 0.0009610329639494079,
    "timestep_mean": 517.9375,
    "epsilon_norm": 8.022734642028809,
    "prediction_norm": 6.729760646820068,
    "timestamp": "2025-07-14T12:33:42.356354"
  },
  {
    "epoch": 253,
    "batch_idx": 32,
    "loss": 0.47236448526382446,
    "learning_rate": 0.0009610329639494079,
    "timestep_mean": 437.0,
    "epsilon_norm": 7.051414489746094,
    "prediction_norm": 5.539896488189697,
    "timestamp": "2025-07-14T12:33:42.361286"
  },
  {
    "epoch": 253,
    "batch_idx": 64,
    "loss": 0.2752629518508911,
    "learning_rate": 0.0009610329639494079,
    "timestep_mean": 515.34375,
    "epsilon_norm": 7.803750991821289,
    "prediction_norm": 6.069071292877197,
    "timestamp": "2025-07-14T12:33:42.366146"
  },
  {
    "epoch": 253,
    "batch_idx": 96,
    "loss": 0.4439041018486023,
    "learning_rate": 0.0009610329639494079,
    "timestep_mean": 433.9375,
    "epsilon_norm": 7.980780601501465,
    "prediction_norm": 5.497969150543213,
    "timestamp": "2025-07-14T12:33:42.371075"
  },
  {
    "epoch": 253,
    "batch_idx": 128,
    "loss": 0.32579565048217773,
    "learning_rate": 0.0009610329639494079,
    "timestep_mean": 459.6875,
    "epsilon_norm": 7.224947452545166,
    "prediction_norm": 5.620479106903076,
    "timestamp": "2025-07-14T12:33:42.375978"
  },
  {
    "epoch": 253,
    "batch_idx": 160,
    "loss": 0.39849987626075745,
    "learning_rate": 0.0009610329639494079,
    "timestep_mean": 461.34375,
    "epsilon_norm": 9.041139602661133,
    "prediction_norm": 6.347946643829346,
    "timestamp": "2025-07-14T12:33:42.380908"
  },
  {
    "epoch": 253,
    "batch_idx": 192,
    "loss": 0.14904987812042236,
    "learning_rate": 0.0009610329639494079,
    "timestep_mean": 602.375,
    "epsilon_norm": 3.792280912399292,
    "prediction_norm": 3.0838205814361572,
    "timestamp": "2025-07-14T12:33:42.385868"
  },
  {
    "epoch": 254,
    "batch_idx": 0,
    "loss": 0.3699157238006592,
    "learning_rate": 0.0009607284204107496,
    "timestep_mean": 568.21875,
    "epsilon_norm": 7.115604400634766,
    "prediction_norm": 4.9023661613464355,
    "timestamp": "2025-07-14T12:33:42.390790"
  },
  {
    "epoch": 254,
    "batch_idx": 32,
    "loss": 0.25303149223327637,
    "learning_rate": 0.0009607284204107496,
    "timestep_mean": 480.125,
    "epsilon_norm": 8.480204582214355,
    "prediction_norm": 7.295487880706787,
    "timestamp": "2025-07-14T12:33:42.395871"
  },
  {
    "epoch": 254,
    "batch_idx": 64,
    "loss": 0.38534873723983765,
    "learning_rate": 0.0009607284204107496,
    "timestep_mean": 433.6875,
    "epsilon_norm": 7.37945032119751,
    "prediction_norm": 5.253054618835449,
    "timestamp": "2025-07-14T12:33:42.400884"
  },
  {
    "epoch": 254,
    "batch_idx": 96,
    "loss": 0.3089044988155365,
    "learning_rate": 0.0009607284204107496,
    "timestep_mean": 546.125,
    "epsilon_norm": 8.793208122253418,
    "prediction_norm": 6.71982479095459,
    "timestamp": "2025-07-14T12:33:42.405860"
  },
  {
    "epoch": 254,
    "batch_idx": 128,
    "loss": 0.41258013248443604,
    "learning_rate": 0.0009607284204107496,
    "timestep_mean": 435.875,
    "epsilon_norm": 7.879859447479248,
    "prediction_norm": 5.631402492523193,
    "timestamp": "2025-07-14T12:33:42.410717"
  },
  {
    "epoch": 254,
    "batch_idx": 160,
    "loss": 0.3717595338821411,
    "learning_rate": 0.0009607284204107496,
    "timestep_mean": 528.25,
    "epsilon_norm": 8.404594421386719,
    "prediction_norm": 7.100460529327393,
    "timestamp": "2025-07-14T12:33:42.415883"
  },
  {
    "epoch": 254,
    "batch_idx": 192,
    "loss": 0.16116686165332794,
    "learning_rate": 0.0009607284204107496,
    "timestep_mean": 511.875,
    "epsilon_norm": 3.1001217365264893,
    "prediction_norm": 2.4234931468963623,
    "timestamp": "2025-07-14T12:33:42.421263"
  },
  {
    "epoch": 255,
    "batch_idx": 0,
    "loss": 0.2619541585445404,
    "learning_rate": 0.0009604227400705136,
    "timestep_mean": 511.71875,
    "epsilon_norm": 8.959439277648926,
    "prediction_norm": 8.55776309967041,
    "timestamp": "2025-07-14T12:33:42.426437"
  },
  {
    "epoch": 255,
    "batch_idx": 32,
    "loss": 0.3434302508831024,
    "learning_rate": 0.0009604227400705136,
    "timestep_mean": 484.75,
    "epsilon_norm": 7.718499660491943,
    "prediction_norm": 6.416794300079346,
    "timestamp": "2025-07-14T12:33:42.431602"
  },
  {
    "epoch": 255,
    "batch_idx": 64,
    "loss": 0.5933029651641846,
    "learning_rate": 0.0009604227400705136,
    "timestep_mean": 456.8125,
    "epsilon_norm": 8.43149471282959,
    "prediction_norm": 6.259721279144287,
    "timestamp": "2025-07-14T12:33:42.436531"
  },
  {
    "epoch": 255,
    "batch_idx": 96,
    "loss": 0.4220585227012634,
    "learning_rate": 0.0009604227400705136,
    "timestep_mean": 463.03125,
    "epsilon_norm": 8.536669731140137,
    "prediction_norm": 7.112143516540527,
    "timestamp": "2025-07-14T12:33:42.441447"
  },
  {
    "epoch": 255,
    "batch_idx": 128,
    "loss": 0.4034930467605591,
    "learning_rate": 0.0009604227400705136,
    "timestep_mean": 495.125,
    "epsilon_norm": 8.68695068359375,
    "prediction_norm": 6.762289047241211,
    "timestamp": "2025-07-14T12:33:42.446466"
  },
  {
    "epoch": 255,
    "batch_idx": 160,
    "loss": 0.36221933364868164,
    "learning_rate": 0.0009604227400705136,
    "timestep_mean": 478.25,
    "epsilon_norm": 7.5705108642578125,
    "prediction_norm": 5.954532146453857,
    "timestamp": "2025-07-14T12:33:42.451350"
  },
  {
    "epoch": 255,
    "batch_idx": 192,
    "loss": 0.7719384431838989,
    "learning_rate": 0.0009604227400705136,
    "timestep_mean": 245.75,
    "epsilon_norm": 5.036623001098633,
    "prediction_norm": 4.229990005493164,
    "timestamp": "2025-07-14T12:33:42.456306"
  },
  {
    "epoch": 256,
    "batch_idx": 0,
    "loss": 0.2653079330921173,
    "learning_rate": 0.0009601159236829356,
    "timestep_mean": 494.5,
    "epsilon_norm": 7.088785648345947,
    "prediction_norm": 6.41112756729126,
    "timestamp": "2025-07-14T12:33:42.461357"
  },
  {
    "epoch": 256,
    "batch_idx": 32,
    "loss": 0.2676044702529907,
    "learning_rate": 0.0009601159236829356,
    "timestep_mean": 419.28125,
    "epsilon_norm": 8.299417495727539,
    "prediction_norm": 6.97492790222168,
    "timestamp": "2025-07-14T12:33:42.466696"
  },
  {
    "epoch": 256,
    "batch_idx": 64,
    "loss": 0.22591477632522583,
    "learning_rate": 0.0009601159236829356,
    "timestep_mean": 512.75,
    "epsilon_norm": 7.29892110824585,
    "prediction_norm": 6.464282989501953,
    "timestamp": "2025-07-14T12:33:42.472920"
  },
  {
    "epoch": 256,
    "batch_idx": 96,
    "loss": 0.23404930531978607,
    "learning_rate": 0.0009601159236829356,
    "timestep_mean": 477.875,
    "epsilon_norm": 6.720057010650635,
    "prediction_norm": 5.775899887084961,
    "timestamp": "2025-07-14T12:33:42.478378"
  },
  {
    "epoch": 256,
    "batch_idx": 128,
    "loss": 0.2929326891899109,
    "learning_rate": 0.0009601159236829356,
    "timestep_mean": 457.5,
    "epsilon_norm": 7.121748447418213,
    "prediction_norm": 5.8067402839660645,
    "timestamp": "2025-07-14T12:33:42.483769"
  },
  {
    "epoch": 256,
    "batch_idx": 160,
    "loss": 0.27665868401527405,
    "learning_rate": 0.0009601159236829356,
    "timestep_mean": 468.34375,
    "epsilon_norm": 8.230920791625977,
    "prediction_norm": 7.146592617034912,
    "timestamp": "2025-07-14T12:33:42.488682"
  },
  {
    "epoch": 256,
    "batch_idx": 192,
    "loss": 0.2699699401855469,
    "learning_rate": 0.0009601159236829356,
    "timestep_mean": 569.5,
    "epsilon_norm": 3.323099136352539,
    "prediction_norm": 2.492489814758301,
    "timestamp": "2025-07-14T12:33:42.493782"
  },
  {
    "epoch": 257,
    "batch_idx": 0,
    "loss": 0.2937653660774231,
    "learning_rate": 0.0009598079720050547,
    "timestep_mean": 493.15625,
    "epsilon_norm": 7.701318264007568,
    "prediction_norm": 6.0223894119262695,
    "timestamp": "2025-07-14T12:33:42.498737"
  },
  {
    "epoch": 257,
    "batch_idx": 32,
    "loss": 0.33691680431365967,
    "learning_rate": 0.0009598079720050547,
    "timestep_mean": 490.1875,
    "epsilon_norm": 8.6082763671875,
    "prediction_norm": 6.979051113128662,
    "timestamp": "2025-07-14T12:33:42.503794"
  },
  {
    "epoch": 257,
    "batch_idx": 64,
    "loss": 0.308096706867218,
    "learning_rate": 0.0009598079720050547,
    "timestep_mean": 489.53125,
    "epsilon_norm": 7.117594242095947,
    "prediction_norm": 5.398326873779297,
    "timestamp": "2025-07-14T12:33:42.508758"
  },
  {
    "epoch": 257,
    "batch_idx": 96,
    "loss": 0.34079042077064514,
    "learning_rate": 0.0009598079720050547,
    "timestep_mean": 465.21875,
    "epsilon_norm": 7.776247978210449,
    "prediction_norm": 6.244919300079346,
    "timestamp": "2025-07-14T12:33:42.513675"
  },
  {
    "epoch": 257,
    "batch_idx": 128,
    "loss": 0.2896559536457062,
    "learning_rate": 0.0009598079720050547,
    "timestep_mean": 444.84375,
    "epsilon_norm": 7.119440078735352,
    "prediction_norm": 6.011839866638184,
    "timestamp": "2025-07-14T12:33:42.518810"
  },
  {
    "epoch": 257,
    "batch_idx": 160,
    "loss": 0.4783017933368683,
    "learning_rate": 0.0009598079720050547,
    "timestep_mean": 545.9375,
    "epsilon_norm": 7.72016716003418,
    "prediction_norm": 6.105885982513428,
    "timestamp": "2025-07-14T12:33:42.524020"
  },
  {
    "epoch": 257,
    "batch_idx": 192,
    "loss": 0.4527190625667572,
    "learning_rate": 0.0009598079720050547,
    "timestep_mean": 471.5,
    "epsilon_norm": 3.3398051261901855,
    "prediction_norm": 2.530893087387085,
    "timestamp": "2025-07-14T12:33:42.529197"
  },
  {
    "epoch": 258,
    "batch_idx": 0,
    "loss": 0.33240553736686707,
    "learning_rate": 0.000959498885796711,
    "timestep_mean": 539.4375,
    "epsilon_norm": 8.406532287597656,
    "prediction_norm": 7.380124092102051,
    "timestamp": "2025-07-14T12:33:42.534364"
  },
  {
    "epoch": 258,
    "batch_idx": 32,
    "loss": 0.24594084918498993,
    "learning_rate": 0.000959498885796711,
    "timestep_mean": 480.65625,
    "epsilon_norm": 7.194711685180664,
    "prediction_norm": 6.550598621368408,
    "timestamp": "2025-07-14T12:33:42.539426"
  },
  {
    "epoch": 258,
    "batch_idx": 64,
    "loss": 0.4852291941642761,
    "learning_rate": 0.000959498885796711,
    "timestep_mean": 409.09375,
    "epsilon_norm": 8.442928314208984,
    "prediction_norm": 6.213719367980957,
    "timestamp": "2025-07-14T12:33:42.544634"
  },
  {
    "epoch": 258,
    "batch_idx": 96,
    "loss": 0.2303948700428009,
    "learning_rate": 0.000959498885796711,
    "timestep_mean": 493.03125,
    "epsilon_norm": 7.819971084594727,
    "prediction_norm": 6.133723258972168,
    "timestamp": "2025-07-14T12:33:42.549706"
  },
  {
    "epoch": 258,
    "batch_idx": 128,
    "loss": 0.3449195325374603,
    "learning_rate": 0.000959498885796711,
    "timestep_mean": 495.15625,
    "epsilon_norm": 8.3511323928833,
    "prediction_norm": 6.401870250701904,
    "timestamp": "2025-07-14T12:33:42.554781"
  },
  {
    "epoch": 258,
    "batch_idx": 160,
    "loss": 0.2795366644859314,
    "learning_rate": 0.000959498885796711,
    "timestep_mean": 534.1875,
    "epsilon_norm": 8.619440078735352,
    "prediction_norm": 6.045011043548584,
    "timestamp": "2025-07-14T12:33:42.559675"
  },
  {
    "epoch": 258,
    "batch_idx": 192,
    "loss": 0.30910658836364746,
    "learning_rate": 0.000959498885796711,
    "timestep_mean": 460.125,
    "epsilon_norm": 3.222166061401367,
    "prediction_norm": 2.1238982677459717,
    "timestamp": "2025-07-14T12:33:42.564622"
  },
  {
    "epoch": 259,
    "batch_idx": 0,
    "loss": 0.508135199546814,
    "learning_rate": 0.0009591886658205441,
    "timestep_mean": 497.0,
    "epsilon_norm": 8.308491706848145,
    "prediction_norm": 5.601384162902832,
    "timestamp": "2025-07-14T12:33:42.569706"
  },
  {
    "epoch": 259,
    "batch_idx": 32,
    "loss": 0.2477201670408249,
    "learning_rate": 0.0009591886658205441,
    "timestep_mean": 481.71875,
    "epsilon_norm": 7.648263454437256,
    "prediction_norm": 6.224039554595947,
    "timestamp": "2025-07-14T12:33:42.574827"
  },
  {
    "epoch": 259,
    "batch_idx": 64,
    "loss": 0.2291359305381775,
    "learning_rate": 0.0009591886658205441,
    "timestep_mean": 565.21875,
    "epsilon_norm": 8.148090362548828,
    "prediction_norm": 6.8713226318359375,
    "timestamp": "2025-07-14T12:33:42.579898"
  },
  {
    "epoch": 259,
    "batch_idx": 96,
    "loss": 0.3659017086029053,
    "learning_rate": 0.0009591886658205441,
    "timestep_mean": 461.03125,
    "epsilon_norm": 8.266215324401855,
    "prediction_norm": 6.448995590209961,
    "timestamp": "2025-07-14T12:33:42.585015"
  },
  {
    "epoch": 259,
    "batch_idx": 128,
    "loss": 0.38319459557533264,
    "learning_rate": 0.0009591886658205441,
    "timestep_mean": 514.0,
    "epsilon_norm": 8.766351699829102,
    "prediction_norm": 6.968611717224121,
    "timestamp": "2025-07-14T12:33:42.590104"
  },
  {
    "epoch": 259,
    "batch_idx": 160,
    "loss": 0.2595463693141937,
    "learning_rate": 0.0009591886658205441,
    "timestep_mean": 533.5,
    "epsilon_norm": 7.306024074554443,
    "prediction_norm": 6.5672430992126465,
    "timestamp": "2025-07-14T12:33:42.595102"
  },
  {
    "epoch": 259,
    "batch_idx": 192,
    "loss": 0.8287380933761597,
    "learning_rate": 0.0009591886658205441,
    "timestep_mean": 428.875,
    "epsilon_norm": 4.6088337898254395,
    "prediction_norm": 2.313778877258301,
    "timestamp": "2025-07-14T12:33:42.600088"
  },
  {
    "epoch": 260,
    "batch_idx": 0,
    "loss": 0.2729886770248413,
    "learning_rate": 0.0009588773128419908,
    "timestep_mean": 546.34375,
    "epsilon_norm": 7.215595722198486,
    "prediction_norm": 6.405482769012451,
    "timestamp": "2025-07-14T12:33:42.605156"
  },
  {
    "epoch": 260,
    "batch_idx": 32,
    "loss": 0.2511187791824341,
    "learning_rate": 0.0009588773128419908,
    "timestep_mean": 520.5,
    "epsilon_norm": 8.677741050720215,
    "prediction_norm": 7.956971168518066,
    "timestamp": "2025-07-14T12:33:42.610042"
  },
  {
    "epoch": 260,
    "batch_idx": 64,
    "loss": 0.2724774181842804,
    "learning_rate": 0.0009588773128419908,
    "timestep_mean": 560.65625,
    "epsilon_norm": 7.48033332824707,
    "prediction_norm": 6.003373146057129,
    "timestamp": "2025-07-14T12:33:42.615183"
  },
  {
    "epoch": 260,
    "batch_idx": 96,
    "loss": 0.3323656916618347,
    "learning_rate": 0.0009588773128419908,
    "timestep_mean": 509.875,
    "epsilon_norm": 7.625157356262207,
    "prediction_norm": 6.278683662414551,
    "timestamp": "2025-07-14T12:33:42.621609"
  },
  {
    "epoch": 260,
    "batch_idx": 128,
    "loss": 0.2072032243013382,
    "learning_rate": 0.0009588773128419908,
    "timestep_mean": 475.5625,
    "epsilon_norm": 7.416119575500488,
    "prediction_norm": 6.354039669036865,
    "timestamp": "2025-07-14T12:33:42.626750"
  },
  {
    "epoch": 260,
    "batch_idx": 160,
    "loss": 0.3297930061817169,
    "learning_rate": 0.0009588773128419908,
    "timestep_mean": 502.21875,
    "epsilon_norm": 8.204290390014648,
    "prediction_norm": 7.073162078857422,
    "timestamp": "2025-07-14T12:33:42.631835"
  },
  {
    "epoch": 260,
    "batch_idx": 192,
    "loss": 0.41941043734550476,
    "learning_rate": 0.0009588773128419908,
    "timestep_mean": 557.125,
    "epsilon_norm": 3.5020859241485596,
    "prediction_norm": 3.079547643661499,
    "timestamp": "2025-07-14T12:33:42.638066"
  },
  {
    "epoch": 261,
    "batch_idx": 0,
    "loss": 0.4152694344520569,
    "learning_rate": 0.0009585648276292839,
    "timestep_mean": 463.96875,
    "epsilon_norm": 7.951239109039307,
    "prediction_norm": 6.678215980529785,
    "timestamp": "2025-07-14T12:33:42.643606"
  },
  {
    "epoch": 261,
    "batch_idx": 32,
    "loss": 0.2562418580055237,
    "learning_rate": 0.0009585648276292839,
    "timestep_mean": 616.5,
    "epsilon_norm": 7.773021697998047,
    "prediction_norm": 6.836628437042236,
    "timestamp": "2025-07-14T12:33:42.648942"
  },
  {
    "epoch": 261,
    "batch_idx": 64,
    "loss": 0.13415902853012085,
    "learning_rate": 0.0009585648276292839,
    "timestep_mean": 525.6875,
    "epsilon_norm": 7.947964668273926,
    "prediction_norm": 7.458927154541016,
    "timestamp": "2025-07-14T12:33:42.654447"
  },
  {
    "epoch": 261,
    "batch_idx": 96,
    "loss": 0.32758063077926636,
    "learning_rate": 0.0009585648276292839,
    "timestep_mean": 575.40625,
    "epsilon_norm": 8.153164863586426,
    "prediction_norm": 6.557496070861816,
    "timestamp": "2025-07-14T12:33:42.659870"
  },
  {
    "epoch": 261,
    "batch_idx": 128,
    "loss": 0.27255070209503174,
    "learning_rate": 0.0009585648276292839,
    "timestep_mean": 514.90625,
    "epsilon_norm": 8.055773735046387,
    "prediction_norm": 7.3368821144104,
    "timestamp": "2025-07-14T12:33:42.665078"
  },
  {
    "epoch": 261,
    "batch_idx": 160,
    "loss": 0.18338477611541748,
    "learning_rate": 0.0009585648276292839,
    "timestep_mean": 507.46875,
    "epsilon_norm": 6.520529747009277,
    "prediction_norm": 5.42018461227417,
    "timestamp": "2025-07-14T12:33:42.670410"
  },
  {
    "epoch": 261,
    "batch_idx": 192,
    "loss": 0.4676479399204254,
    "learning_rate": 0.0009585648276292839,
    "timestep_mean": 480.75,
    "epsilon_norm": 4.069939136505127,
    "prediction_norm": 3.1469922065734863,
    "timestamp": "2025-07-14T12:33:42.675762"
  },
  {
    "epoch": 262,
    "batch_idx": 0,
    "loss": 0.2040361613035202,
    "learning_rate": 0.0009582512109534492,
    "timestep_mean": 492.59375,
    "epsilon_norm": 7.308541774749756,
    "prediction_norm": 6.600910663604736,
    "timestamp": "2025-07-14T12:33:42.681097"
  },
  {
    "epoch": 262,
    "batch_idx": 32,
    "loss": 0.3849254846572876,
    "learning_rate": 0.0009582512109534492,
    "timestep_mean": 513.6875,
    "epsilon_norm": 9.021173477172852,
    "prediction_norm": 7.177058219909668,
    "timestamp": "2025-07-14T12:33:42.686410"
  },
  {
    "epoch": 262,
    "batch_idx": 64,
    "loss": 0.37891319394111633,
    "learning_rate": 0.0009582512109534492,
    "timestep_mean": 410.46875,
    "epsilon_norm": 7.446741580963135,
    "prediction_norm": 5.020070552825928,
    "timestamp": "2025-07-14T12:33:42.691547"
  },
  {
    "epoch": 262,
    "batch_idx": 96,
    "loss": 0.17970792949199677,
    "learning_rate": 0.0009582512109534492,
    "timestep_mean": 574.125,
    "epsilon_norm": 7.966298580169678,
    "prediction_norm": 6.728221416473389,
    "timestamp": "2025-07-14T12:33:42.696836"
  },
  {
    "epoch": 262,
    "batch_idx": 128,
    "loss": 0.20223365724086761,
    "learning_rate": 0.0009582512109534492,
    "timestep_mean": 542.875,
    "epsilon_norm": 6.655958652496338,
    "prediction_norm": 5.571084499359131,
    "timestamp": "2025-07-14T12:33:42.702106"
  },
  {
    "epoch": 262,
    "batch_idx": 160,
    "loss": 0.19250187277793884,
    "learning_rate": 0.0009582512109534492,
    "timestep_mean": 531.5,
    "epsilon_norm": 7.352254867553711,
    "prediction_norm": 6.0272698402404785,
    "timestamp": "2025-07-14T12:33:42.707311"
  },
  {
    "epoch": 262,
    "batch_idx": 192,
    "loss": 0.12579283118247986,
    "learning_rate": 0.0009582512109534492,
    "timestep_mean": 432.75,
    "epsilon_norm": 3.811613082885742,
    "prediction_norm": 3.5645289421081543,
    "timestamp": "2025-07-14T12:33:42.712568"
  },
  {
    "epoch": 263,
    "batch_idx": 0,
    "loss": 0.368137001991272,
    "learning_rate": 0.0009579364635883051,
    "timestep_mean": 491.875,
    "epsilon_norm": 8.048986434936523,
    "prediction_norm": 5.962050914764404,
    "timestamp": "2025-07-14T12:33:42.717681"
  },
  {
    "epoch": 263,
    "batch_idx": 32,
    "loss": 0.417609304189682,
    "learning_rate": 0.0009579364635883051,
    "timestep_mean": 570.53125,
    "epsilon_norm": 8.795953750610352,
    "prediction_norm": 7.065436840057373,
    "timestamp": "2025-07-14T12:33:42.722835"
  },
  {
    "epoch": 263,
    "batch_idx": 64,
    "loss": 0.2903348207473755,
    "learning_rate": 0.0009579364635883051,
    "timestep_mean": 625.5,
    "epsilon_norm": 8.277665138244629,
    "prediction_norm": 7.438764572143555,
    "timestamp": "2025-07-14T12:33:42.728097"
  },
  {
    "epoch": 263,
    "batch_idx": 96,
    "loss": 0.41260671615600586,
    "learning_rate": 0.0009579364635883051,
    "timestep_mean": 506.96875,
    "epsilon_norm": 8.054598808288574,
    "prediction_norm": 6.533071517944336,
    "timestamp": "2025-07-14T12:33:42.733359"
  },
  {
    "epoch": 263,
    "batch_idx": 128,
    "loss": 0.47449642419815063,
    "learning_rate": 0.0009579364635883051,
    "timestep_mean": 405.375,
    "epsilon_norm": 8.145397186279297,
    "prediction_norm": 6.903686046600342,
    "timestamp": "2025-07-14T12:33:42.738538"
  },
  {
    "epoch": 263,
    "batch_idx": 160,
    "loss": 0.31277310848236084,
    "learning_rate": 0.0009579364635883051,
    "timestep_mean": 534.125,
    "epsilon_norm": 8.844175338745117,
    "prediction_norm": 7.776297092437744,
    "timestamp": "2025-07-14T12:33:42.743763"
  },
  {
    "epoch": 263,
    "batch_idx": 192,
    "loss": 0.4478166103363037,
    "learning_rate": 0.0009579364635883051,
    "timestep_mean": 395.5,
    "epsilon_norm": 5.05605936050415,
    "prediction_norm": 3.295231819152832,
    "timestamp": "2025-07-14T12:33:42.748836"
  },
  {
    "epoch": 264,
    "batch_idx": 0,
    "loss": 0.39714860916137695,
    "learning_rate": 0.0009576205863104591,
    "timestep_mean": 524.28125,
    "epsilon_norm": 8.388583183288574,
    "prediction_norm": 6.558848857879639,
    "timestamp": "2025-07-14T12:33:42.754226"
  },
  {
    "epoch": 264,
    "batch_idx": 32,
    "loss": 0.2775366008281708,
    "learning_rate": 0.0009576205863104591,
    "timestep_mean": 490.46875,
    "epsilon_norm": 8.203421592712402,
    "prediction_norm": 7.647592067718506,
    "timestamp": "2025-07-14T12:33:42.759336"
  },
  {
    "epoch": 264,
    "batch_idx": 64,
    "loss": 0.6122229099273682,
    "learning_rate": 0.0009576205863104591,
    "timestep_mean": 484.96875,
    "epsilon_norm": 8.21510124206543,
    "prediction_norm": 6.082369327545166,
    "timestamp": "2025-07-14T12:33:42.764493"
  },
  {
    "epoch": 264,
    "batch_idx": 96,
    "loss": 0.32987546920776367,
    "learning_rate": 0.0009576205863104591,
    "timestep_mean": 515.8125,
    "epsilon_norm": 7.595113754272461,
    "prediction_norm": 5.9294843673706055,
    "timestamp": "2025-07-14T12:33:42.769787"
  },
  {
    "epoch": 264,
    "batch_idx": 128,
    "loss": 0.38397568464279175,
    "learning_rate": 0.0009576205863104591,
    "timestep_mean": 440.53125,
    "epsilon_norm": 7.797076225280762,
    "prediction_norm": 6.562893390655518,
    "timestamp": "2025-07-14T12:33:42.774944"
  },
  {
    "epoch": 264,
    "batch_idx": 160,
    "loss": 0.3492473065853119,
    "learning_rate": 0.0009576205863104591,
    "timestep_mean": 542.90625,
    "epsilon_norm": 7.6874566078186035,
    "prediction_norm": 5.69131326675415,
    "timestamp": "2025-07-14T12:33:42.780167"
  },
  {
    "epoch": 264,
    "batch_idx": 192,
    "loss": 0.14339911937713623,
    "learning_rate": 0.0009576205863104591,
    "timestep_mean": 465.375,
    "epsilon_norm": 4.125204086303711,
    "prediction_norm": 3.724910259246826,
    "timestamp": "2025-07-14T12:33:42.785372"
  },
  {
    "epoch": 265,
    "batch_idx": 0,
    "loss": 0.3729681372642517,
    "learning_rate": 0.0009573035798993072,
    "timestep_mean": 447.34375,
    "epsilon_norm": 7.982299327850342,
    "prediction_norm": 6.335365295410156,
    "timestamp": "2025-07-14T12:33:42.790602"
  },
  {
    "epoch": 265,
    "batch_idx": 32,
    "loss": 0.24050942063331604,
    "learning_rate": 0.0009573035798993072,
    "timestep_mean": 511.625,
    "epsilon_norm": 7.4099507331848145,
    "prediction_norm": 6.288759231567383,
    "timestamp": "2025-07-14T12:33:42.796046"
  },
  {
    "epoch": 265,
    "batch_idx": 64,
    "loss": 0.3097451627254486,
    "learning_rate": 0.0009573035798993072,
    "timestep_mean": 544.875,
    "epsilon_norm": 7.935401439666748,
    "prediction_norm": 6.943385124206543,
    "timestamp": "2025-07-14T12:33:42.801135"
  },
  {
    "epoch": 265,
    "batch_idx": 96,
    "loss": 0.45266062021255493,
    "learning_rate": 0.0009573035798993072,
    "timestep_mean": 500.3125,
    "epsilon_norm": 8.541009902954102,
    "prediction_norm": 6.7345757484436035,
    "timestamp": "2025-07-14T12:33:42.806226"
  },
  {
    "epoch": 265,
    "batch_idx": 128,
    "loss": 0.2412552833557129,
    "learning_rate": 0.0009573035798993072,
    "timestep_mean": 469.0625,
    "epsilon_norm": 7.493993759155273,
    "prediction_norm": 5.962330341339111,
    "timestamp": "2025-07-14T12:33:42.811336"
  },
  {
    "epoch": 265,
    "batch_idx": 160,
    "loss": 0.17156846821308136,
    "learning_rate": 0.0009573035798993072,
    "timestep_mean": 569.4375,
    "epsilon_norm": 8.191346168518066,
    "prediction_norm": 7.300162315368652,
    "timestamp": "2025-07-14T12:33:42.816437"
  },
  {
    "epoch": 265,
    "batch_idx": 192,
    "loss": 0.22606846690177917,
    "learning_rate": 0.0009573035798993072,
    "timestep_mean": 519.875,
    "epsilon_norm": 4.8501811027526855,
    "prediction_norm": 4.099292755126953,
    "timestamp": "2025-07-14T12:33:42.821559"
  },
  {
    "epoch": 266,
    "batch_idx": 0,
    "loss": 0.36687079071998596,
    "learning_rate": 0.000956985445137031,
    "timestep_mean": 447.03125,
    "epsilon_norm": 7.2100982666015625,
    "prediction_norm": 5.391615867614746,
    "timestamp": "2025-07-14T12:33:42.826776"
  },
  {
    "epoch": 266,
    "batch_idx": 32,
    "loss": 0.3198414742946625,
    "learning_rate": 0.000956985445137031,
    "timestep_mean": 587.0,
    "epsilon_norm": 8.185173034667969,
    "prediction_norm": 6.814352512359619,
    "timestamp": "2025-07-14T12:33:42.832018"
  },
  {
    "epoch": 266,
    "batch_idx": 64,
    "loss": 0.5030628442764282,
    "learning_rate": 0.000956985445137031,
    "timestep_mean": 430.5,
    "epsilon_norm": 7.971944332122803,
    "prediction_norm": 5.3079071044921875,
    "timestamp": "2025-07-14T12:33:42.837044"
  },
  {
    "epoch": 266,
    "batch_idx": 96,
    "loss": 0.3585372269153595,
    "learning_rate": 0.000956985445137031,
    "timestep_mean": 495.1875,
    "epsilon_norm": 7.220314979553223,
    "prediction_norm": 6.015036582946777,
    "timestamp": "2025-07-14T12:33:42.842039"
  },
  {
    "epoch": 266,
    "batch_idx": 128,
    "loss": 0.5491623878479004,
    "learning_rate": 0.000956985445137031,
    "timestep_mean": 479.3125,
    "epsilon_norm": 8.752286911010742,
    "prediction_norm": 6.127659797668457,
    "timestamp": "2025-07-14T12:33:42.846866"
  },
  {
    "epoch": 266,
    "batch_idx": 160,
    "loss": 0.3075769245624542,
    "learning_rate": 0.000956985445137031,
    "timestep_mean": 438.09375,
    "epsilon_norm": 8.462896347045898,
    "prediction_norm": 6.663125991821289,
    "timestamp": "2025-07-14T12:33:42.851739"
  },
  {
    "epoch": 266,
    "batch_idx": 192,
    "loss": 0.3083544969558716,
    "learning_rate": 0.000956985445137031,
    "timestep_mean": 588.25,
    "epsilon_norm": 4.368552207946777,
    "prediction_norm": 4.004289627075195,
    "timestamp": "2025-07-14T12:33:42.856584"
  },
  {
    "epoch": 267,
    "batch_idx": 0,
    "loss": 0.2366217076778412,
    "learning_rate": 0.0009566661828085965,
    "timestep_mean": 574.125,
    "epsilon_norm": 8.306273460388184,
    "prediction_norm": 7.399158477783203,
    "timestamp": "2025-07-14T12:33:42.861639"
  },
  {
    "epoch": 267,
    "batch_idx": 32,
    "loss": 0.35052788257598877,
    "learning_rate": 0.0009566661828085965,
    "timestep_mean": 482.5,
    "epsilon_norm": 7.865785598754883,
    "prediction_norm": 6.078075408935547,
    "timestamp": "2025-07-14T12:33:42.866519"
  },
  {
    "epoch": 267,
    "batch_idx": 64,
    "loss": 0.22753363847732544,
    "learning_rate": 0.0009566661828085965,
    "timestep_mean": 482.84375,
    "epsilon_norm": 7.0268874168396,
    "prediction_norm": 6.043872356414795,
    "timestamp": "2025-07-14T12:33:42.871400"
  },
  {
    "epoch": 267,
    "batch_idx": 96,
    "loss": 0.3668428063392639,
    "learning_rate": 0.0009566661828085965,
    "timestep_mean": 434.0,
    "epsilon_norm": 7.930110454559326,
    "prediction_norm": 5.927646636962891,
    "timestamp": "2025-07-14T12:33:42.876246"
  },
  {
    "epoch": 267,
    "batch_idx": 128,
    "loss": 0.3016645014286041,
    "learning_rate": 0.0009566661828085965,
    "timestep_mean": 473.34375,
    "epsilon_norm": 8.346257209777832,
    "prediction_norm": 6.621098518371582,
    "timestamp": "2025-07-14T12:33:42.881162"
  },
  {
    "epoch": 267,
    "batch_idx": 160,
    "loss": 0.28927701711654663,
    "learning_rate": 0.0009566661828085965,
    "timestep_mean": 528.90625,
    "epsilon_norm": 7.121756553649902,
    "prediction_norm": 5.681346416473389,
    "timestamp": "2025-07-14T12:33:42.886005"
  },
  {
    "epoch": 267,
    "batch_idx": 192,
    "loss": 0.2781168222427368,
    "learning_rate": 0.0009566661828085965,
    "timestep_mean": 444.0,
    "epsilon_norm": 3.5560896396636963,
    "prediction_norm": 2.96360445022583,
    "timestamp": "2025-07-14T12:33:42.890921"
  },
  {
    "epoch": 268,
    "batch_idx": 0,
    "loss": 0.22886478900909424,
    "learning_rate": 0.0009563457937017518,
    "timestep_mean": 466.0625,
    "epsilon_norm": 6.268468856811523,
    "prediction_norm": 6.266790390014648,
    "timestamp": "2025-07-14T12:33:42.895850"
  },
  {
    "epoch": 268,
    "batch_idx": 32,
    "loss": 0.2382570505142212,
    "learning_rate": 0.0009563457937017518,
    "timestep_mean": 565.46875,
    "epsilon_norm": 7.124578952789307,
    "prediction_norm": 6.24500036239624,
    "timestamp": "2025-07-14T12:33:42.900685"
  },
  {
    "epoch": 268,
    "batch_idx": 64,
    "loss": 0.38458600640296936,
    "learning_rate": 0.0009563457937017518,
    "timestep_mean": 529.125,
    "epsilon_norm": 8.074642181396484,
    "prediction_norm": 6.177724838256836,
    "timestamp": "2025-07-14T12:33:42.910351"
  },
  {
    "epoch": 268,
    "batch_idx": 96,
    "loss": 0.37435027956962585,
    "learning_rate": 0.0009563457937017518,
    "timestep_mean": 512.03125,
    "epsilon_norm": 7.629035949707031,
    "prediction_norm": 6.239115238189697,
    "timestamp": "2025-07-14T12:33:42.915427"
  },
  {
    "epoch": 268,
    "batch_idx": 128,
    "loss": 0.24594929814338684,
    "learning_rate": 0.0009563457937017518,
    "timestep_mean": 488.15625,
    "epsilon_norm": 7.9307355880737305,
    "prediction_norm": 6.618714809417725,
    "timestamp": "2025-07-14T12:33:42.920346"
  },
  {
    "epoch": 268,
    "batch_idx": 160,
    "loss": 0.31562408804893494,
    "learning_rate": 0.0009563457937017518,
    "timestep_mean": 531.03125,
    "epsilon_norm": 8.411787033081055,
    "prediction_norm": 7.850581645965576,
    "timestamp": "2025-07-14T12:33:42.925400"
  },
  {
    "epoch": 268,
    "batch_idx": 192,
    "loss": 0.07893136143684387,
    "learning_rate": 0.0009563457937017518,
    "timestep_mean": 552.25,
    "epsilon_norm": 3.80777907371521,
    "prediction_norm": 3.614226818084717,
    "timestamp": "2025-07-14T12:33:42.930588"
  },
  {
    "epoch": 269,
    "batch_idx": 0,
    "loss": 0.29219892621040344,
    "learning_rate": 0.0009560242786070251,
    "timestep_mean": 481.09375,
    "epsilon_norm": 8.356796264648438,
    "prediction_norm": 6.828084945678711,
    "timestamp": "2025-07-14T12:33:42.935740"
  },
  {
    "epoch": 269,
    "batch_idx": 32,
    "loss": 0.14655150473117828,
    "learning_rate": 0.0009560242786070251,
    "timestep_mean": 572.75,
    "epsilon_norm": 8.78822135925293,
    "prediction_norm": 7.900921821594238,
    "timestamp": "2025-07-14T12:33:42.940679"
  },
  {
    "epoch": 269,
    "batch_idx": 64,
    "loss": 0.30622124671936035,
    "learning_rate": 0.0009560242786070251,
    "timestep_mean": 446.71875,
    "epsilon_norm": 8.683597564697266,
    "prediction_norm": 7.3911452293396,
    "timestamp": "2025-07-14T12:33:42.945857"
  },
  {
    "epoch": 269,
    "batch_idx": 96,
    "loss": 0.30263015627861023,
    "learning_rate": 0.0009560242786070251,
    "timestep_mean": 534.0625,
    "epsilon_norm": 8.82755184173584,
    "prediction_norm": 7.763561248779297,
    "timestamp": "2025-07-14T12:33:42.950828"
  },
  {
    "epoch": 269,
    "batch_idx": 128,
    "loss": 0.1670912504196167,
    "learning_rate": 0.0009560242786070251,
    "timestep_mean": 573.8125,
    "epsilon_norm": 8.26970100402832,
    "prediction_norm": 7.328499794006348,
    "timestamp": "2025-07-14T12:33:42.955859"
  },
  {
    "epoch": 269,
    "batch_idx": 160,
    "loss": 0.38294345140457153,
    "learning_rate": 0.0009560242786070251,
    "timestep_mean": 591.84375,
    "epsilon_norm": 7.807260036468506,
    "prediction_norm": 6.5162248611450195,
    "timestamp": "2025-07-14T12:33:42.961056"
  },
  {
    "epoch": 269,
    "batch_idx": 192,
    "loss": 0.3141610622406006,
    "learning_rate": 0.0009560242786070251,
    "timestep_mean": 556.0,
    "epsilon_norm": 3.560698986053467,
    "prediction_norm": 2.7830827236175537,
    "timestamp": "2025-07-14T12:33:42.966228"
  },
  {
    "epoch": 270,
    "batch_idx": 0,
    "loss": 0.3750165104866028,
    "learning_rate": 0.0009557016383177229,
    "timestep_mean": 499.46875,
    "epsilon_norm": 7.758876800537109,
    "prediction_norm": 5.655979633331299,
    "timestamp": "2025-07-14T12:33:42.971426"
  },
  {
    "epoch": 270,
    "batch_idx": 32,
    "loss": 0.4016362428665161,
    "learning_rate": 0.0009557016383177229,
    "timestep_mean": 615.0,
    "epsilon_norm": 8.151334762573242,
    "prediction_norm": 6.85075044631958,
    "timestamp": "2025-07-14T12:33:42.976485"
  },
  {
    "epoch": 270,
    "batch_idx": 64,
    "loss": 0.29662981629371643,
    "learning_rate": 0.0009557016383177229,
    "timestep_mean": 458.0,
    "epsilon_norm": 8.335168838500977,
    "prediction_norm": 7.332855224609375,
    "timestamp": "2025-07-14T12:33:42.981485"
  },
  {
    "epoch": 270,
    "batch_idx": 96,
    "loss": 0.5657117366790771,
    "learning_rate": 0.0009557016383177229,
    "timestep_mean": 484.46875,
    "epsilon_norm": 8.423124313354492,
    "prediction_norm": 6.046850204467773,
    "timestamp": "2025-07-14T12:33:42.986624"
  },
  {
    "epoch": 270,
    "batch_idx": 128,
    "loss": 0.3898690342903137,
    "learning_rate": 0.0009557016383177229,
    "timestep_mean": 477.03125,
    "epsilon_norm": 7.800199508666992,
    "prediction_norm": 5.963641166687012,
    "timestamp": "2025-07-14T12:33:42.991689"
  },
  {
    "epoch": 270,
    "batch_idx": 160,
    "loss": 0.1821231245994568,
    "learning_rate": 0.0009557016383177229,
    "timestep_mean": 571.71875,
    "epsilon_norm": 7.390082359313965,
    "prediction_norm": 6.652356147766113,
    "timestamp": "2025-07-14T12:33:42.996709"
  },
  {
    "epoch": 270,
    "batch_idx": 192,
    "loss": 0.8219398260116577,
    "learning_rate": 0.0009557016383177229,
    "timestep_mean": 405.375,
    "epsilon_norm": 4.342014312744141,
    "prediction_norm": 2.2049033641815186,
    "timestamp": "2025-07-14T12:33:43.002084"
  },
  {
    "epoch": 271,
    "batch_idx": 0,
    "loss": 0.4797790050506592,
    "learning_rate": 0.0009553778736299281,
    "timestep_mean": 427.3125,
    "epsilon_norm": 8.594084739685059,
    "prediction_norm": 6.258948802947998,
    "timestamp": "2025-07-14T12:33:43.007108"
  },
  {
    "epoch": 271,
    "batch_idx": 32,
    "loss": 0.49690207839012146,
    "learning_rate": 0.0009553778736299281,
    "timestep_mean": 442.09375,
    "epsilon_norm": 8.731045722961426,
    "prediction_norm": 6.661929130554199,
    "timestamp": "2025-07-14T12:33:43.012670"
  },
  {
    "epoch": 271,
    "batch_idx": 64,
    "loss": 0.24653246998786926,
    "learning_rate": 0.0009553778736299281,
    "timestep_mean": 506.0,
    "epsilon_norm": 7.237330436706543,
    "prediction_norm": 6.3185882568359375,
    "timestamp": "2025-07-14T12:33:43.017753"
  },
  {
    "epoch": 271,
    "batch_idx": 96,
    "loss": 0.316866010427475,
    "learning_rate": 0.0009553778736299281,
    "timestep_mean": 453.03125,
    "epsilon_norm": 6.617225646972656,
    "prediction_norm": 4.903761386871338,
    "timestamp": "2025-07-14T12:33:43.022754"
  },
  {
    "epoch": 271,
    "batch_idx": 128,
    "loss": 0.3502839207649231,
    "learning_rate": 0.0009553778736299281,
    "timestep_mean": 541.125,
    "epsilon_norm": 9.557391166687012,
    "prediction_norm": 7.710070610046387,
    "timestamp": "2025-07-14T12:33:43.027681"
  },
  {
    "epoch": 271,
    "batch_idx": 160,
    "loss": 0.4730897843837738,
    "learning_rate": 0.0009553778736299281,
    "timestep_mean": 507.375,
    "epsilon_norm": 7.9861297607421875,
    "prediction_norm": 5.673940181732178,
    "timestamp": "2025-07-14T12:33:43.032654"
  },
  {
    "epoch": 271,
    "batch_idx": 192,
    "loss": 0.17502915859222412,
    "learning_rate": 0.0009553778736299281,
    "timestep_mean": 578.75,
    "epsilon_norm": 3.9402225017547607,
    "prediction_norm": 3.6704211235046387,
    "timestamp": "2025-07-14T12:33:43.037701"
  },
  {
    "epoch": 272,
    "batch_idx": 0,
    "loss": 0.5618453025817871,
    "learning_rate": 0.0009550529853424981,
    "timestep_mean": 359.96875,
    "epsilon_norm": 8.880282402038574,
    "prediction_norm": 5.819097995758057,
    "timestamp": "2025-07-14T12:33:43.042633"
  },
  {
    "epoch": 272,
    "batch_idx": 32,
    "loss": 0.36918869614601135,
    "learning_rate": 0.0009550529853424981,
    "timestep_mean": 516.8125,
    "epsilon_norm": 7.497654438018799,
    "prediction_norm": 5.464251518249512,
    "timestamp": "2025-07-14T12:33:43.047555"
  },
  {
    "epoch": 272,
    "batch_idx": 64,
    "loss": 0.38731124997138977,
    "learning_rate": 0.0009550529853424981,
    "timestep_mean": 532.09375,
    "epsilon_norm": 8.40816593170166,
    "prediction_norm": 6.449214458465576,
    "timestamp": "2025-07-14T12:33:43.052460"
  },
  {
    "epoch": 272,
    "batch_idx": 96,
    "loss": 0.3773483335971832,
    "learning_rate": 0.0009550529853424981,
    "timestep_mean": 437.96875,
    "epsilon_norm": 7.086436748504639,
    "prediction_norm": 5.133882999420166,
    "timestamp": "2025-07-14T12:33:43.057395"
  },
  {
    "epoch": 272,
    "batch_idx": 128,
    "loss": 0.38530880212783813,
    "learning_rate": 0.0009550529853424981,
    "timestep_mean": 416.5,
    "epsilon_norm": 8.157477378845215,
    "prediction_norm": 6.326211929321289,
    "timestamp": "2025-07-14T12:33:43.062313"
  },
  {
    "epoch": 272,
    "batch_idx": 160,
    "loss": 0.19876965880393982,
    "learning_rate": 0.0009550529853424981,
    "timestep_mean": 603.21875,
    "epsilon_norm": 7.980673313140869,
    "prediction_norm": 7.353129863739014,
    "timestamp": "2025-07-14T12:33:43.067286"
  },
  {
    "epoch": 272,
    "batch_idx": 192,
    "loss": 0.3593575656414032,
    "learning_rate": 0.0009550529853424981,
    "timestep_mean": 510.0,
    "epsilon_norm": 3.5281903743743896,
    "prediction_norm": 2.770850658416748,
    "timestamp": "2025-07-14T12:33:43.072235"
  },
  {
    "epoch": 273,
    "batch_idx": 0,
    "loss": 0.4965061545372009,
    "learning_rate": 0.0009547269742570621,
    "timestep_mean": 367.90625,
    "epsilon_norm": 8.002921104431152,
    "prediction_norm": 6.531871318817139,
    "timestamp": "2025-07-14T12:33:43.077201"
  },
  {
    "epoch": 273,
    "batch_idx": 32,
    "loss": 0.33780497312545776,
    "learning_rate": 0.0009547269742570621,
    "timestep_mean": 458.46875,
    "epsilon_norm": 7.818486213684082,
    "prediction_norm": 6.272164821624756,
    "timestamp": "2025-07-14T12:33:43.082161"
  },
  {
    "epoch": 273,
    "batch_idx": 64,
    "loss": 0.24443399906158447,
    "learning_rate": 0.0009547269742570621,
    "timestep_mean": 524.15625,
    "epsilon_norm": 9.280214309692383,
    "prediction_norm": 7.898434162139893,
    "timestamp": "2025-07-14T12:33:43.087122"
  },
  {
    "epoch": 273,
    "batch_idx": 96,
    "loss": 0.26430362462997437,
    "learning_rate": 0.0009547269742570621,
    "timestep_mean": 534.84375,
    "epsilon_norm": 6.97166109085083,
    "prediction_norm": 6.433035373687744,
    "timestamp": "2025-07-14T12:33:43.092074"
  },
  {
    "epoch": 273,
    "batch_idx": 128,
    "loss": 0.25747889280319214,
    "learning_rate": 0.0009547269742570621,
    "timestep_mean": 459.15625,
    "epsilon_norm": 7.466910362243652,
    "prediction_norm": 6.4608540534973145,
    "timestamp": "2025-07-14T12:33:43.097024"
  },
  {
    "epoch": 273,
    "batch_idx": 160,
    "loss": 0.16911956667900085,
    "learning_rate": 0.0009547269742570621,
    "timestep_mean": 545.15625,
    "epsilon_norm": 8.17175006866455,
    "prediction_norm": 7.079657554626465,
    "timestamp": "2025-07-14T12:33:43.101960"
  },
  {
    "epoch": 273,
    "batch_idx": 192,
    "loss": 0.2775319814682007,
    "learning_rate": 0.0009547269742570621,
    "timestep_mean": 523.25,
    "epsilon_norm": 4.290273666381836,
    "prediction_norm": 4.05177116394043,
    "timestamp": "2025-07-14T12:33:43.106898"
  },
  {
    "epoch": 274,
    "batch_idx": 0,
    "loss": 0.5172027945518494,
    "learning_rate": 0.0009543998411780203,
    "timestep_mean": 404.90625,
    "epsilon_norm": 7.911867618560791,
    "prediction_norm": 5.770598411560059,
    "timestamp": "2025-07-14T12:33:43.111860"
  },
  {
    "epoch": 274,
    "batch_idx": 32,
    "loss": 0.31982871890068054,
    "learning_rate": 0.0009543998411780203,
    "timestep_mean": 547.375,
    "epsilon_norm": 8.772747039794922,
    "prediction_norm": 7.032867431640625,
    "timestamp": "2025-07-14T12:33:43.116846"
  },
  {
    "epoch": 274,
    "batch_idx": 64,
    "loss": 0.20595988631248474,
    "learning_rate": 0.0009543998411780203,
    "timestep_mean": 565.625,
    "epsilon_norm": 7.313658714294434,
    "prediction_norm": 6.6029438972473145,
    "timestamp": "2025-07-14T12:33:43.121742"
  },
  {
    "epoch": 274,
    "batch_idx": 96,
    "loss": 0.274141401052475,
    "learning_rate": 0.0009543998411780203,
    "timestep_mean": 533.71875,
    "epsilon_norm": 7.797775745391846,
    "prediction_norm": 7.098731994628906,
    "timestamp": "2025-07-14T12:33:43.126645"
  },
  {
    "epoch": 274,
    "batch_idx": 128,
    "loss": 0.1993744671344757,
    "learning_rate": 0.0009543998411780203,
    "timestep_mean": 511.15625,
    "epsilon_norm": 7.109847068786621,
    "prediction_norm": 5.7387871742248535,
    "timestamp": "2025-07-14T12:33:43.131480"
  },
  {
    "epoch": 274,
    "batch_idx": 160,
    "loss": 0.48983609676361084,
    "learning_rate": 0.0009543998411780203,
    "timestep_mean": 425.75,
    "epsilon_norm": 8.659080505371094,
    "prediction_norm": 6.20819091796875,
    "timestamp": "2025-07-14T12:33:43.136365"
  },
  {
    "epoch": 274,
    "batch_idx": 192,
    "loss": 0.3972463011741638,
    "learning_rate": 0.0009543998411780203,
    "timestep_mean": 491.375,
    "epsilon_norm": 4.897994518280029,
    "prediction_norm": 3.7443785667419434,
    "timestamp": "2025-07-14T12:33:43.141459"
  },
  {
    "epoch": 275,
    "batch_idx": 0,
    "loss": 0.27764368057250977,
    "learning_rate": 0.0009540715869125409,
    "timestep_mean": 488.53125,
    "epsilon_norm": 8.422477722167969,
    "prediction_norm": 7.0207929611206055,
    "timestamp": "2025-07-14T12:33:43.146492"
  },
  {
    "epoch": 275,
    "batch_idx": 32,
    "loss": 0.3346928358078003,
    "learning_rate": 0.0009540715869125409,
    "timestep_mean": 451.5625,
    "epsilon_norm": 8.05932331085205,
    "prediction_norm": 6.687866687774658,
    "timestamp": "2025-07-14T12:33:43.151644"
  },
  {
    "epoch": 275,
    "batch_idx": 64,
    "loss": 0.3520004153251648,
    "learning_rate": 0.0009540715869125409,
    "timestep_mean": 516.625,
    "epsilon_norm": 7.60126256942749,
    "prediction_norm": 6.101322650909424,
    "timestamp": "2025-07-14T12:33:43.156800"
  },
  {
    "epoch": 275,
    "batch_idx": 96,
    "loss": 0.2864948809146881,
    "learning_rate": 0.0009540715869125409,
    "timestep_mean": 518.65625,
    "epsilon_norm": 9.417340278625488,
    "prediction_norm": 7.859977722167969,
    "timestamp": "2025-07-14T12:33:43.161988"
  },
  {
    "epoch": 275,
    "batch_idx": 128,
    "loss": 0.24816831946372986,
    "learning_rate": 0.0009540715869125409,
    "timestep_mean": 581.1875,
    "epsilon_norm": 7.943411350250244,
    "prediction_norm": 6.758279323577881,
    "timestamp": "2025-07-14T12:33:43.167050"
  },
  {
    "epoch": 275,
    "batch_idx": 160,
    "loss": 0.22538115084171295,
    "learning_rate": 0.0009540715869125409,
    "timestep_mean": 626.21875,
    "epsilon_norm": 7.4721245765686035,
    "prediction_norm": 6.251407623291016,
    "timestamp": "2025-07-14T12:33:43.172141"
  },
  {
    "epoch": 275,
    "batch_idx": 192,
    "loss": 0.3941110372543335,
    "learning_rate": 0.0009540715869125409,
    "timestep_mean": 419.5,
    "epsilon_norm": 4.861063003540039,
    "prediction_norm": 3.6360113620758057,
    "timestamp": "2025-07-14T12:33:43.177309"
  },
  {
    "epoch": 276,
    "batch_idx": 0,
    "loss": 0.2566916048526764,
    "learning_rate": 0.0009537422122705587,
    "timestep_mean": 512.59375,
    "epsilon_norm": 7.325244426727295,
    "prediction_norm": 6.417150974273682,
    "timestamp": "2025-07-14T12:33:43.182284"
  },
  {
    "epoch": 276,
    "batch_idx": 32,
    "loss": 0.26086539030075073,
    "learning_rate": 0.0009537422122705587,
    "timestep_mean": 554.71875,
    "epsilon_norm": 7.7557172775268555,
    "prediction_norm": 6.724920749664307,
    "timestamp": "2025-07-14T12:33:43.187276"
  },
  {
    "epoch": 276,
    "batch_idx": 64,
    "loss": 0.19626298546791077,
    "learning_rate": 0.0009537422122705587,
    "timestep_mean": 535.75,
    "epsilon_norm": 7.122483253479004,
    "prediction_norm": 6.823844909667969,
    "timestamp": "2025-07-14T12:33:43.192259"
  },
  {
    "epoch": 276,
    "batch_idx": 96,
    "loss": 0.2731602191925049,
    "learning_rate": 0.0009537422122705587,
    "timestep_mean": 481.6875,
    "epsilon_norm": 9.595301628112793,
    "prediction_norm": 8.389588356018066,
    "timestamp": "2025-07-14T12:33:43.197205"
  },
  {
    "epoch": 276,
    "batch_idx": 128,
    "loss": 0.4741371273994446,
    "learning_rate": 0.0009537422122705587,
    "timestep_mean": 430.03125,
    "epsilon_norm": 9.194351196289062,
    "prediction_norm": 7.2394609451293945,
    "timestamp": "2025-07-14T12:33:43.202156"
  },
  {
    "epoch": 276,
    "batch_idx": 160,
    "loss": 0.2602452039718628,
    "learning_rate": 0.0009537422122705587,
    "timestep_mean": 541.0625,
    "epsilon_norm": 7.528190612792969,
    "prediction_norm": 6.162415981292725,
    "timestamp": "2025-07-14T12:33:43.207071"
  },
  {
    "epoch": 276,
    "batch_idx": 192,
    "loss": 0.8170834183692932,
    "learning_rate": 0.0009537422122705587,
    "timestep_mean": 370.875,
    "epsilon_norm": 4.832016944885254,
    "prediction_norm": 3.326127529144287,
    "timestamp": "2025-07-14T12:33:43.212086"
  },
  {
    "epoch": 277,
    "batch_idx": 0,
    "loss": 0.4462498724460602,
    "learning_rate": 0.000953411718064773,
    "timestep_mean": 542.40625,
    "epsilon_norm": 8.885943412780762,
    "prediction_norm": 7.227989673614502,
    "timestamp": "2025-07-14T12:33:43.217086"
  },
  {
    "epoch": 277,
    "batch_idx": 32,
    "loss": 0.46744391322135925,
    "learning_rate": 0.000953411718064773,
    "timestep_mean": 565.6875,
    "epsilon_norm": 9.776216506958008,
    "prediction_norm": 8.19290542602539,
    "timestamp": "2025-07-14T12:33:43.222022"
  },
  {
    "epoch": 277,
    "batch_idx": 64,
    "loss": 0.36058542132377625,
    "learning_rate": 0.000953411718064773,
    "timestep_mean": 482.71875,
    "epsilon_norm": 7.872745037078857,
    "prediction_norm": 5.933739185333252,
    "timestamp": "2025-07-14T12:33:43.226927"
  },
  {
    "epoch": 277,
    "batch_idx": 96,
    "loss": 0.12499435245990753,
    "learning_rate": 0.000953411718064773,
    "timestep_mean": 614.84375,
    "epsilon_norm": 6.9317402839660645,
    "prediction_norm": 7.097598552703857,
    "timestamp": "2025-07-14T12:33:43.232126"
  },
  {
    "epoch": 277,
    "batch_idx": 128,
    "loss": 0.4033123254776001,
    "learning_rate": 0.000953411718064773,
    "timestep_mean": 492.875,
    "epsilon_norm": 7.552353382110596,
    "prediction_norm": 5.977241039276123,
    "timestamp": "2025-07-14T12:33:43.237146"
  },
  {
    "epoch": 277,
    "batch_idx": 160,
    "loss": 0.3172343075275421,
    "learning_rate": 0.000953411718064773,
    "timestep_mean": 423.8125,
    "epsilon_norm": 8.266775131225586,
    "prediction_norm": 6.599905490875244,
    "timestamp": "2025-07-14T12:33:43.242366"
  },
  {
    "epoch": 277,
    "batch_idx": 192,
    "loss": 0.19041697680950165,
    "learning_rate": 0.000953411718064773,
    "timestep_mean": 440.25,
    "epsilon_norm": 3.28715443611145,
    "prediction_norm": 3.315537929534912,
    "timestamp": "2025-07-14T12:33:43.247474"
  },
  {
    "epoch": 278,
    "batch_idx": 0,
    "loss": 0.3783823847770691,
    "learning_rate": 0.0009530801051106453,
    "timestep_mean": 491.96875,
    "epsilon_norm": 7.296532154083252,
    "prediction_norm": 5.779242038726807,
    "timestamp": "2025-07-14T12:33:43.252506"
  },
  {
    "epoch": 278,
    "batch_idx": 32,
    "loss": 0.22734004259109497,
    "learning_rate": 0.0009530801051106453,
    "timestep_mean": 493.1875,
    "epsilon_norm": 8.978155136108398,
    "prediction_norm": 8.34516429901123,
    "timestamp": "2025-07-14T12:33:43.257762"
  },
  {
    "epoch": 278,
    "batch_idx": 64,
    "loss": 0.7843325138092041,
    "learning_rate": 0.0009530801051106453,
    "timestep_mean": 395.40625,
    "epsilon_norm": 8.907567977905273,
    "prediction_norm": 5.071845054626465,
    "timestamp": "2025-07-14T12:33:43.262983"
  },
  {
    "epoch": 278,
    "batch_idx": 96,
    "loss": 0.6959280967712402,
    "learning_rate": 0.0009530801051106453,
    "timestep_mean": 417.4375,
    "epsilon_norm": 9.494704246520996,
    "prediction_norm": 6.338712215423584,
    "timestamp": "2025-07-14T12:33:43.267940"
  },
  {
    "epoch": 278,
    "batch_idx": 128,
    "loss": 0.43274036049842834,
    "learning_rate": 0.0009530801051106453,
    "timestep_mean": 486.71875,
    "epsilon_norm": 7.835443496704102,
    "prediction_norm": 5.812536716461182,
    "timestamp": "2025-07-14T12:33:43.273129"
  },
  {
    "epoch": 278,
    "batch_idx": 160,
    "loss": 0.38103291392326355,
    "learning_rate": 0.0009530801051106453,
    "timestep_mean": 480.6875,
    "epsilon_norm": 7.688238620758057,
    "prediction_norm": 5.899369239807129,
    "timestamp": "2025-07-14T12:33:43.278538"
  },
  {
    "epoch": 278,
    "batch_idx": 192,
    "loss": 0.2340206354856491,
    "learning_rate": 0.0009530801051106453,
    "timestep_mean": 564.875,
    "epsilon_norm": 3.894770383834839,
    "prediction_norm": 3.052163600921631,
    "timestamp": "2025-07-14T12:33:43.283714"
  },
  {
    "epoch": 279,
    "batch_idx": 0,
    "loss": 0.5404988527297974,
    "learning_rate": 0.0009527473742263975,
    "timestep_mean": 451.4375,
    "epsilon_norm": 8.836343765258789,
    "prediction_norm": 6.745689392089844,
    "timestamp": "2025-07-14T12:33:43.289100"
  },
  {
    "epoch": 279,
    "batch_idx": 32,
    "loss": 0.42256927490234375,
    "learning_rate": 0.0009527473742263975,
    "timestep_mean": 433.90625,
    "epsilon_norm": 8.434379577636719,
    "prediction_norm": 6.43220329284668,
    "timestamp": "2025-07-14T12:33:43.294308"
  },
  {
    "epoch": 279,
    "batch_idx": 64,
    "loss": 0.3648952841758728,
    "learning_rate": 0.0009527473742263975,
    "timestep_mean": 431.25,
    "epsilon_norm": 6.995096683502197,
    "prediction_norm": 5.25522518157959,
    "timestamp": "2025-07-14T12:33:43.299526"
  },
  {
    "epoch": 279,
    "batch_idx": 96,
    "loss": 0.2431793063879013,
    "learning_rate": 0.0009527473742263975,
    "timestep_mean": 603.1875,
    "epsilon_norm": 7.432921409606934,
    "prediction_norm": 6.0892815589904785,
    "timestamp": "2025-07-14T12:33:43.304724"
  },
  {
    "epoch": 279,
    "batch_idx": 128,
    "loss": 0.3298434615135193,
    "learning_rate": 0.0009527473742263975,
    "timestep_mean": 493.4375,
    "epsilon_norm": 7.656698703765869,
    "prediction_norm": 6.670475006103516,
    "timestamp": "2025-07-14T12:33:43.309857"
  },
  {
    "epoch": 279,
    "batch_idx": 160,
    "loss": 0.1394038200378418,
    "learning_rate": 0.0009527473742263975,
    "timestep_mean": 515.0,
    "epsilon_norm": 7.219043731689453,
    "prediction_norm": 6.960116386413574,
    "timestamp": "2025-07-14T12:33:43.315073"
  },
  {
    "epoch": 279,
    "batch_idx": 192,
    "loss": 0.10002560168504715,
    "learning_rate": 0.0009527473742263975,
    "timestep_mean": 531.75,
    "epsilon_norm": 2.7491679191589355,
    "prediction_norm": 2.4482333660125732,
    "timestamp": "2025-07-14T12:33:43.320159"
  },
  {
    "epoch": 280,
    "batch_idx": 0,
    "loss": 0.5036802887916565,
    "learning_rate": 0.00095241352623301,
    "timestep_mean": 435.875,
    "epsilon_norm": 8.153838157653809,
    "prediction_norm": 6.305902481079102,
    "timestamp": "2025-07-14T12:33:43.325273"
  },
  {
    "epoch": 280,
    "batch_idx": 32,
    "loss": 0.27821865677833557,
    "learning_rate": 0.00095241352623301,
    "timestep_mean": 497.1875,
    "epsilon_norm": 8.374788284301758,
    "prediction_norm": 7.368062496185303,
    "timestamp": "2025-07-14T12:33:43.330418"
  },
  {
    "epoch": 280,
    "batch_idx": 64,
    "loss": 0.4359850287437439,
    "learning_rate": 0.00095241352623301,
    "timestep_mean": 518.0,
    "epsilon_norm": 8.075910568237305,
    "prediction_norm": 6.4800028800964355,
    "timestamp": "2025-07-14T12:33:43.335459"
  },
  {
    "epoch": 280,
    "batch_idx": 96,
    "loss": 0.33493930101394653,
    "learning_rate": 0.00095241352623301,
    "timestep_mean": 576.96875,
    "epsilon_norm": 8.431829452514648,
    "prediction_norm": 6.87376070022583,
    "timestamp": "2025-07-14T12:33:43.340629"
  },
  {
    "epoch": 280,
    "batch_idx": 128,
    "loss": 0.3184651732444763,
    "learning_rate": 0.00095241352623301,
    "timestep_mean": 520.46875,
    "epsilon_norm": 7.0237908363342285,
    "prediction_norm": 6.275636672973633,
    "timestamp": "2025-07-14T12:33:43.345689"
  },
  {
    "epoch": 280,
    "batch_idx": 160,
    "loss": 0.20940838754177094,
    "learning_rate": 0.00095241352623301,
    "timestep_mean": 435.59375,
    "epsilon_norm": 8.212065696716309,
    "prediction_norm": 7.556963920593262,
    "timestamp": "2025-07-14T12:33:43.350764"
  },
  {
    "epoch": 280,
    "batch_idx": 192,
    "loss": 0.19391636550426483,
    "learning_rate": 0.00095241352623301,
    "timestep_mean": 466.625,
    "epsilon_norm": 3.363673210144043,
    "prediction_norm": 2.6812007427215576,
    "timestamp": "2025-07-14T12:33:43.355944"
  },
  {
    "epoch": 281,
    "batch_idx": 0,
    "loss": 0.31992295384407043,
    "learning_rate": 0.0009520785619542198,
    "timestep_mean": 575.0,
    "epsilon_norm": 8.434867858886719,
    "prediction_norm": 6.894613742828369,
    "timestamp": "2025-07-14T12:33:43.361101"
  },
  {
    "epoch": 281,
    "batch_idx": 32,
    "loss": 0.2938736081123352,
    "learning_rate": 0.0009520785619542198,
    "timestep_mean": 547.90625,
    "epsilon_norm": 7.595629692077637,
    "prediction_norm": 6.123859405517578,
    "timestamp": "2025-07-14T12:33:43.366206"
  },
  {
    "epoch": 281,
    "batch_idx": 64,
    "loss": 0.46140313148498535,
    "learning_rate": 0.0009520785619542198,
    "timestep_mean": 550.71875,
    "epsilon_norm": 9.018599510192871,
    "prediction_norm": 6.559987545013428,
    "timestamp": "2025-07-14T12:33:43.371415"
  },
  {
    "epoch": 281,
    "batch_idx": 96,
    "loss": 0.2874956727027893,
    "learning_rate": 0.0009520785619542198,
    "timestep_mean": 543.375,
    "epsilon_norm": 8.840948104858398,
    "prediction_norm": 7.098427772521973,
    "timestamp": "2025-07-14T12:33:43.376433"
  },
  {
    "epoch": 281,
    "batch_idx": 128,
    "loss": 0.23280391097068787,
    "learning_rate": 0.0009520785619542198,
    "timestep_mean": 540.96875,
    "epsilon_norm": 7.594395160675049,
    "prediction_norm": 6.365180015563965,
    "timestamp": "2025-07-14T12:33:43.381487"
  },
  {
    "epoch": 281,
    "batch_idx": 160,
    "loss": 0.21256804466247559,
    "learning_rate": 0.0009520785619542198,
    "timestep_mean": 459.34375,
    "epsilon_norm": 6.833194732666016,
    "prediction_norm": 5.306425094604492,
    "timestamp": "2025-07-14T12:33:43.386461"
  },
  {
    "epoch": 281,
    "batch_idx": 192,
    "loss": 0.4288645386695862,
    "learning_rate": 0.0009520785619542198,
    "timestep_mean": 395.375,
    "epsilon_norm": 3.9463863372802734,
    "prediction_norm": 2.809236526489258,
    "timestamp": "2025-07-14T12:33:43.391624"
  },
  {
    "epoch": 282,
    "batch_idx": 0,
    "loss": 0.30125942826271057,
    "learning_rate": 0.0009517424822165177,
    "timestep_mean": 562.5625,
    "epsilon_norm": 8.461594581604004,
    "prediction_norm": 7.545709609985352,
    "timestamp": "2025-07-14T12:33:43.396764"
  },
  {
    "epoch": 282,
    "batch_idx": 32,
    "loss": 0.3295402526855469,
    "learning_rate": 0.0009517424822165177,
    "timestep_mean": 528.40625,
    "epsilon_norm": 8.523642539978027,
    "prediction_norm": 6.993513107299805,
    "timestamp": "2025-07-14T12:33:43.401679"
  },
  {
    "epoch": 282,
    "batch_idx": 64,
    "loss": 0.29935625195503235,
    "learning_rate": 0.0009517424822165177,
    "timestep_mean": 478.75,
    "epsilon_norm": 6.728855133056641,
    "prediction_norm": 5.234987735748291,
    "timestamp": "2025-07-14T12:33:43.406633"
  },
  {
    "epoch": 282,
    "batch_idx": 96,
    "loss": 0.20410531759262085,
    "learning_rate": 0.0009517424822165177,
    "timestep_mean": 563.875,
    "epsilon_norm": 8.770517349243164,
    "prediction_norm": 8.581124305725098,
    "timestamp": "2025-07-14T12:33:43.411572"
  },
  {
    "epoch": 282,
    "batch_idx": 128,
    "loss": 0.396003782749176,
    "learning_rate": 0.0009517424822165177,
    "timestep_mean": 434.21875,
    "epsilon_norm": 8.404264450073242,
    "prediction_norm": 7.491061210632324,
    "timestamp": "2025-07-14T12:33:43.416949"
  },
  {
    "epoch": 282,
    "batch_idx": 160,
    "loss": 0.22342810034751892,
    "learning_rate": 0.0009517424822165177,
    "timestep_mean": 463.4375,
    "epsilon_norm": 7.635499477386475,
    "prediction_norm": 6.678894996643066,
    "timestamp": "2025-07-14T12:33:43.422259"
  },
  {
    "epoch": 282,
    "batch_idx": 192,
    "loss": 0.28102174401283264,
    "learning_rate": 0.0009517424822165177,
    "timestep_mean": 466.5,
    "epsilon_norm": 2.9370667934417725,
    "prediction_norm": 2.654582977294922,
    "timestamp": "2025-07-14T12:33:43.428877"
  },
  {
    "epoch": 283,
    "batch_idx": 0,
    "loss": 0.26926329731941223,
    "learning_rate": 0.000951405287849147,
    "timestep_mean": 514.21875,
    "epsilon_norm": 7.321929931640625,
    "prediction_norm": 5.999095439910889,
    "timestamp": "2025-07-14T12:33:43.435566"
  },
  {
    "epoch": 283,
    "batch_idx": 32,
    "loss": 0.2652513086795807,
    "learning_rate": 0.000951405287849147,
    "timestep_mean": 496.28125,
    "epsilon_norm": 7.927099704742432,
    "prediction_norm": 7.003142833709717,
    "timestamp": "2025-07-14T12:33:43.441315"
  },
  {
    "epoch": 283,
    "batch_idx": 64,
    "loss": 0.21343925595283508,
    "learning_rate": 0.000951405287849147,
    "timestep_mean": 558.5625,
    "epsilon_norm": 6.891580581665039,
    "prediction_norm": 6.520120620727539,
    "timestamp": "2025-07-14T12:33:43.446638"
  },
  {
    "epoch": 283,
    "batch_idx": 96,
    "loss": 0.26493167877197266,
    "learning_rate": 0.000951405287849147,
    "timestep_mean": 571.375,
    "epsilon_norm": 8.055170059204102,
    "prediction_norm": 6.808821201324463,
    "timestamp": "2025-07-14T12:33:43.451818"
  },
  {
    "epoch": 283,
    "batch_idx": 128,
    "loss": 0.3552946448326111,
    "learning_rate": 0.000951405287849147,
    "timestep_mean": 540.59375,
    "epsilon_norm": 8.804756164550781,
    "prediction_norm": 7.188752174377441,
    "timestamp": "2025-07-14T12:33:43.457057"
  },
  {
    "epoch": 283,
    "batch_idx": 160,
    "loss": 0.448713481426239,
    "learning_rate": 0.000951405287849147,
    "timestep_mean": 483.8125,
    "epsilon_norm": 8.573962211608887,
    "prediction_norm": 6.40022087097168,
    "timestamp": "2025-07-14T12:33:43.462292"
  },
  {
    "epoch": 283,
    "batch_idx": 192,
    "loss": 0.3047129213809967,
    "learning_rate": 0.000951405287849147,
    "timestep_mean": 506.75,
    "epsilon_norm": 4.476334571838379,
    "prediction_norm": 3.3484373092651367,
    "timestamp": "2025-07-14T12:33:43.467598"
  },
  {
    "epoch": 284,
    "batch_idx": 0,
    "loss": 0.3578745722770691,
    "learning_rate": 0.0009510669796841016,
    "timestep_mean": 513.40625,
    "epsilon_norm": 9.361444473266602,
    "prediction_norm": 7.474137783050537,
    "timestamp": "2025-07-14T12:33:43.472750"
  },
  {
    "epoch": 284,
    "batch_idx": 32,
    "loss": 0.2592999339103699,
    "learning_rate": 0.0009510669796841016,
    "timestep_mean": 522.71875,
    "epsilon_norm": 6.6596455574035645,
    "prediction_norm": 5.754810333251953,
    "timestamp": "2025-07-14T12:33:43.477977"
  },
  {
    "epoch": 284,
    "batch_idx": 64,
    "loss": 0.30478435754776,
    "learning_rate": 0.0009510669796841016,
    "timestep_mean": 538.84375,
    "epsilon_norm": 8.463028907775879,
    "prediction_norm": 7.485373020172119,
    "timestamp": "2025-07-14T12:33:43.483325"
  },
  {
    "epoch": 284,
    "batch_idx": 96,
    "loss": 0.2864181399345398,
    "learning_rate": 0.0009510669796841016,
    "timestep_mean": 488.1875,
    "epsilon_norm": 8.229435920715332,
    "prediction_norm": 7.233951091766357,
    "timestamp": "2025-07-14T12:33:43.488416"
  },
  {
    "epoch": 284,
    "batch_idx": 128,
    "loss": 0.2299785017967224,
    "learning_rate": 0.0009510669796841016,
    "timestep_mean": 562.53125,
    "epsilon_norm": 7.691765785217285,
    "prediction_norm": 7.126996040344238,
    "timestamp": "2025-07-14T12:33:43.493554"
  },
  {
    "epoch": 284,
    "batch_idx": 160,
    "loss": 0.25539952516555786,
    "learning_rate": 0.0009510669796841016,
    "timestep_mean": 515.34375,
    "epsilon_norm": 8.646953582763672,
    "prediction_norm": 8.02354907989502,
    "timestamp": "2025-07-14T12:33:43.498556"
  },
  {
    "epoch": 284,
    "batch_idx": 192,
    "loss": 0.032896511256694794,
    "learning_rate": 0.0009510669796841016,
    "timestep_mean": 595.875,
    "epsilon_norm": 4.782681465148926,
    "prediction_norm": 4.578314304351807,
    "timestamp": "2025-07-14T12:33:43.509030"
  },
  {
    "epoch": 285,
    "batch_idx": 0,
    "loss": 0.3941916227340698,
    "learning_rate": 0.0009507275585561231,
    "timestep_mean": 467.59375,
    "epsilon_norm": 8.315583229064941,
    "prediction_norm": 6.61884880065918,
    "timestamp": "2025-07-14T12:33:43.514637"
  },
  {
    "epoch": 285,
    "batch_idx": 32,
    "loss": 0.3172629773616791,
    "learning_rate": 0.0009507275585561231,
    "timestep_mean": 399.5625,
    "epsilon_norm": 8.62767219543457,
    "prediction_norm": 6.4317426681518555,
    "timestamp": "2025-07-14T12:33:43.519729"
  },
  {
    "epoch": 285,
    "batch_idx": 64,
    "loss": 0.2964518666267395,
    "learning_rate": 0.0009507275585561231,
    "timestep_mean": 500.3125,
    "epsilon_norm": 8.562232971191406,
    "prediction_norm": 7.581015110015869,
    "timestamp": "2025-07-14T12:33:43.524834"
  },
  {
    "epoch": 285,
    "batch_idx": 96,
    "loss": 0.32200947403907776,
    "learning_rate": 0.0009507275585561231,
    "timestep_mean": 519.28125,
    "epsilon_norm": 7.5167927742004395,
    "prediction_norm": 5.606142044067383,
    "timestamp": "2025-07-14T12:33:43.530040"
  },
  {
    "epoch": 285,
    "batch_idx": 128,
    "loss": 0.46026691794395447,
    "learning_rate": 0.0009507275585561231,
    "timestep_mean": 528.71875,
    "epsilon_norm": 8.734243392944336,
    "prediction_norm": 7.003402233123779,
    "timestamp": "2025-07-14T12:33:43.535084"
  },
  {
    "epoch": 285,
    "batch_idx": 160,
    "loss": 0.4098825752735138,
    "learning_rate": 0.0009507275585561231,
    "timestep_mean": 520.4375,
    "epsilon_norm": 7.804510116577148,
    "prediction_norm": 6.354818820953369,
    "timestamp": "2025-07-14T12:33:43.540171"
  },
  {
    "epoch": 285,
    "batch_idx": 192,
    "loss": 0.16490113735198975,
    "learning_rate": 0.0009507275585561231,
    "timestep_mean": 544.125,
    "epsilon_norm": 4.035801887512207,
    "prediction_norm": 3.4074654579162598,
    "timestamp": "2025-07-14T12:33:43.545360"
  },
  {
    "epoch": 286,
    "batch_idx": 0,
    "loss": 0.23371583223342896,
    "learning_rate": 0.0009503870253026994,
    "timestep_mean": 532.28125,
    "epsilon_norm": 8.444372177124023,
    "prediction_norm": 7.653898239135742,
    "timestamp": "2025-07-14T12:33:43.550496"
  },
  {
    "epoch": 286,
    "batch_idx": 32,
    "loss": 0.388875275850296,
    "learning_rate": 0.0009503870253026994,
    "timestep_mean": 484.84375,
    "epsilon_norm": 8.0057373046875,
    "prediction_norm": 6.153618812561035,
    "timestamp": "2025-07-14T12:33:43.555416"
  },
  {
    "epoch": 286,
    "batch_idx": 64,
    "loss": 0.14896494150161743,
    "learning_rate": 0.0009503870253026994,
    "timestep_mean": 629.40625,
    "epsilon_norm": 7.90508508682251,
    "prediction_norm": 7.4276275634765625,
    "timestamp": "2025-07-14T12:33:43.560422"
  },
  {
    "epoch": 286,
    "batch_idx": 96,
    "loss": 0.17150095105171204,
    "learning_rate": 0.0009503870253026994,
    "timestep_mean": 572.65625,
    "epsilon_norm": 7.759212970733643,
    "prediction_norm": 7.118258476257324,
    "timestamp": "2025-07-14T12:33:43.565447"
  },
  {
    "epoch": 286,
    "batch_idx": 128,
    "loss": 0.319935142993927,
    "learning_rate": 0.0009503870253026994,
    "timestep_mean": 434.96875,
    "epsilon_norm": 8.188996315002441,
    "prediction_norm": 7.150568962097168,
    "timestamp": "2025-07-14T12:33:43.570524"
  },
  {
    "epoch": 286,
    "batch_idx": 160,
    "loss": 0.33620762825012207,
    "learning_rate": 0.0009503870253026994,
    "timestep_mean": 523.65625,
    "epsilon_norm": 8.479475021362305,
    "prediction_norm": 7.119488716125488,
    "timestamp": "2025-07-14T12:33:43.575580"
  },
  {
    "epoch": 286,
    "batch_idx": 192,
    "loss": 0.37244412302970886,
    "learning_rate": 0.0009503870253026994,
    "timestep_mean": 476.625,
    "epsilon_norm": 3.8470122814178467,
    "prediction_norm": 2.941427707672119,
    "timestamp": "2025-07-14T12:33:43.580573"
  },
  {
    "epoch": 287,
    "batch_idx": 0,
    "loss": 0.5870616436004639,
    "learning_rate": 0.0009500453807640623,
    "timestep_mean": 488.53125,
    "epsilon_norm": 8.128312110900879,
    "prediction_norm": 5.635207653045654,
    "timestamp": "2025-07-14T12:33:43.585659"
  },
  {
    "epoch": 287,
    "batch_idx": 32,
    "loss": 0.3597393333911896,
    "learning_rate": 0.0009500453807640623,
    "timestep_mean": 498.75,
    "epsilon_norm": 7.945052146911621,
    "prediction_norm": 6.151861190795898,
    "timestamp": "2025-07-14T12:33:43.590646"
  },
  {
    "epoch": 287,
    "batch_idx": 64,
    "loss": 0.24719981849193573,
    "learning_rate": 0.0009500453807640623,
    "timestep_mean": 536.65625,
    "epsilon_norm": 7.391284465789795,
    "prediction_norm": 6.883116245269775,
    "timestamp": "2025-07-14T12:33:43.595977"
  },
  {
    "epoch": 287,
    "batch_idx": 96,
    "loss": 0.4635246992111206,
    "learning_rate": 0.0009500453807640623,
    "timestep_mean": 443.625,
    "epsilon_norm": 8.049601554870605,
    "prediction_norm": 6.826620101928711,
    "timestamp": "2025-07-14T12:33:43.601079"
  },
  {
    "epoch": 287,
    "batch_idx": 128,
    "loss": 0.48253098130226135,
    "learning_rate": 0.0009500453807640623,
    "timestep_mean": 494.8125,
    "epsilon_norm": 7.656102657318115,
    "prediction_norm": 6.648655414581299,
    "timestamp": "2025-07-14T12:33:43.606286"
  },
  {
    "epoch": 287,
    "batch_idx": 160,
    "loss": 0.4078180193901062,
    "learning_rate": 0.0009500453807640623,
    "timestep_mean": 495.15625,
    "epsilon_norm": 9.752081871032715,
    "prediction_norm": 8.390168190002441,
    "timestamp": "2025-07-14T12:33:43.611491"
  },
  {
    "epoch": 287,
    "batch_idx": 192,
    "loss": 0.3794407248497009,
    "learning_rate": 0.0009500453807640623,
    "timestep_mean": 406.0,
    "epsilon_norm": 3.8012142181396484,
    "prediction_norm": 2.9666147232055664,
    "timestamp": "2025-07-14T12:33:43.616481"
  },
  {
    "epoch": 288,
    "batch_idx": 0,
    "loss": 0.17981478571891785,
    "learning_rate": 0.0009497026257831859,
    "timestep_mean": 596.25,
    "epsilon_norm": 8.118450164794922,
    "prediction_norm": 7.572685241699219,
    "timestamp": "2025-07-14T12:33:43.621465"
  },
  {
    "epoch": 288,
    "batch_idx": 32,
    "loss": 0.30483168363571167,
    "learning_rate": 0.0009497026257831859,
    "timestep_mean": 452.5,
    "epsilon_norm": 7.10155725479126,
    "prediction_norm": 5.738879203796387,
    "timestamp": "2025-07-14T12:33:43.626549"
  },
  {
    "epoch": 288,
    "batch_idx": 64,
    "loss": 0.447457492351532,
    "learning_rate": 0.0009497026257831859,
    "timestep_mean": 477.4375,
    "epsilon_norm": 8.488900184631348,
    "prediction_norm": 6.131280899047852,
    "timestamp": "2025-07-14T12:33:43.631514"
  },
  {
    "epoch": 288,
    "batch_idx": 96,
    "loss": 0.4325283169746399,
    "learning_rate": 0.0009497026257831859,
    "timestep_mean": 512.71875,
    "epsilon_norm": 8.376795768737793,
    "prediction_norm": 6.158642768859863,
    "timestamp": "2025-07-14T12:33:43.636604"
  },
  {
    "epoch": 288,
    "batch_idx": 128,
    "loss": 0.2629166841506958,
    "learning_rate": 0.0009497026257831859,
    "timestep_mean": 565.75,
    "epsilon_norm": 8.121328353881836,
    "prediction_norm": 6.6672282218933105,
    "timestamp": "2025-07-14T12:33:43.641766"
  },
  {
    "epoch": 288,
    "batch_idx": 160,
    "loss": 0.4595460295677185,
    "learning_rate": 0.0009497026257831859,
    "timestep_mean": 471.09375,
    "epsilon_norm": 9.040013313293457,
    "prediction_norm": 6.84328556060791,
    "timestamp": "2025-07-14T12:33:43.647008"
  },
  {
    "epoch": 288,
    "batch_idx": 192,
    "loss": 0.36813250184059143,
    "learning_rate": 0.0009497026257831859,
    "timestep_mean": 629.875,
    "epsilon_norm": 4.448502540588379,
    "prediction_norm": 3.6191904544830322,
    "timestamp": "2025-07-14T12:33:43.652182"
  },
  {
    "epoch": 289,
    "batch_idx": 0,
    "loss": 0.35314223170280457,
    "learning_rate": 0.000949358761205784,
    "timestep_mean": 492.875,
    "epsilon_norm": 8.579522132873535,
    "prediction_norm": 6.51767110824585,
    "timestamp": "2025-07-14T12:33:43.657329"
  },
  {
    "epoch": 289,
    "batch_idx": 32,
    "loss": 0.25687575340270996,
    "learning_rate": 0.000949358761205784,
    "timestep_mean": 504.34375,
    "epsilon_norm": 7.894010066986084,
    "prediction_norm": 6.528560638427734,
    "timestamp": "2025-07-14T12:33:43.662385"
  },
  {
    "epoch": 289,
    "batch_idx": 64,
    "loss": 0.22489067912101746,
    "learning_rate": 0.000949358761205784,
    "timestep_mean": 502.84375,
    "epsilon_norm": 7.92274284362793,
    "prediction_norm": 6.955752849578857,
    "timestamp": "2025-07-14T12:33:43.667448"
  },
  {
    "epoch": 289,
    "batch_idx": 96,
    "loss": 0.3774206042289734,
    "learning_rate": 0.000949358761205784,
    "timestep_mean": 492.28125,
    "epsilon_norm": 7.260254859924316,
    "prediction_norm": 5.471286296844482,
    "timestamp": "2025-07-14T12:33:43.672628"
  },
  {
    "epoch": 289,
    "batch_idx": 128,
    "loss": 0.37467390298843384,
    "learning_rate": 0.000949358761205784,
    "timestep_mean": 478.59375,
    "epsilon_norm": 7.929901599884033,
    "prediction_norm": 6.171762943267822,
    "timestamp": "2025-07-14T12:33:43.677724"
  },
  {
    "epoch": 289,
    "batch_idx": 160,
    "loss": 0.24239349365234375,
    "learning_rate": 0.000949358761205784,
    "timestep_mean": 547.375,
    "epsilon_norm": 7.135112285614014,
    "prediction_norm": 5.986904144287109,
    "timestamp": "2025-07-14T12:33:43.682739"
  },
  {
    "epoch": 289,
    "batch_idx": 192,
    "loss": 0.480083703994751,
    "learning_rate": 0.000949358761205784,
    "timestep_mean": 561.0,
    "epsilon_norm": 3.519402265548706,
    "prediction_norm": 2.5641748905181885,
    "timestamp": "2025-07-14T12:33:43.687957"
  },
  {
    "epoch": 290,
    "batch_idx": 0,
    "loss": 0.3294666111469269,
    "learning_rate": 0.000949013787880308,
    "timestep_mean": 528.5625,
    "epsilon_norm": 9.915165901184082,
    "prediction_norm": 8.887137413024902,
    "timestamp": "2025-07-14T12:33:43.693086"
  },
  {
    "epoch": 290,
    "batch_idx": 32,
    "loss": 0.31010642647743225,
    "learning_rate": 0.000949013787880308,
    "timestep_mean": 529.96875,
    "epsilon_norm": 7.805123329162598,
    "prediction_norm": 6.657403945922852,
    "timestamp": "2025-07-14T12:33:43.698085"
  },
  {
    "epoch": 290,
    "batch_idx": 64,
    "loss": 0.28901809453964233,
    "learning_rate": 0.000949013787880308,
    "timestep_mean": 539.3125,
    "epsilon_norm": 8.482211112976074,
    "prediction_norm": 8.118890762329102,
    "timestamp": "2025-07-14T12:33:43.703281"
  },
  {
    "epoch": 290,
    "batch_idx": 96,
    "loss": 0.32834893465042114,
    "learning_rate": 0.000949013787880308,
    "timestep_mean": 560.84375,
    "epsilon_norm": 7.483584880828857,
    "prediction_norm": 7.357409477233887,
    "timestamp": "2025-07-14T12:33:43.708642"
  },
  {
    "epoch": 290,
    "batch_idx": 128,
    "loss": 0.2141311764717102,
    "learning_rate": 0.000949013787880308,
    "timestep_mean": 496.25,
    "epsilon_norm": 7.594937324523926,
    "prediction_norm": 6.776896953582764,
    "timestamp": "2025-07-14T12:33:43.713814"
  },
  {
    "epoch": 290,
    "batch_idx": 160,
    "loss": 0.6046381592750549,
    "learning_rate": 0.000949013787880308,
    "timestep_mean": 455.3125,
    "epsilon_norm": 9.301623344421387,
    "prediction_norm": 6.844699382781982,
    "timestamp": "2025-07-14T12:33:43.719148"
  },
  {
    "epoch": 290,
    "batch_idx": 192,
    "loss": 0.07565350830554962,
    "learning_rate": 0.000949013787880308,
    "timestep_mean": 668.0,
    "epsilon_norm": 3.336392879486084,
    "prediction_norm": 3.3505489826202393,
    "timestamp": "2025-07-14T12:33:43.724435"
  },
  {
    "epoch": 291,
    "batch_idx": 0,
    "loss": 0.1819155216217041,
    "learning_rate": 0.0009486677066579459,
    "timestep_mean": 566.40625,
    "epsilon_norm": 8.169462203979492,
    "prediction_norm": 7.32847261428833,
    "timestamp": "2025-07-14T12:33:43.729711"
  },
  {
    "epoch": 291,
    "batch_idx": 32,
    "loss": 0.3422015309333801,
    "learning_rate": 0.0009486677066579459,
    "timestep_mean": 405.96875,
    "epsilon_norm": 7.191867351531982,
    "prediction_norm": 4.792884349822998,
    "timestamp": "2025-07-14T12:33:43.735012"
  },
  {
    "epoch": 291,
    "batch_idx": 64,
    "loss": 0.358684778213501,
    "learning_rate": 0.0009486677066579459,
    "timestep_mean": 473.1875,
    "epsilon_norm": 7.721518039703369,
    "prediction_norm": 5.723184108734131,
    "timestamp": "2025-07-14T12:33:43.740093"
  },
  {
    "epoch": 291,
    "batch_idx": 96,
    "loss": 0.3247261643409729,
    "learning_rate": 0.0009486677066579459,
    "timestep_mean": 488.5,
    "epsilon_norm": 8.158528327941895,
    "prediction_norm": 6.682134628295898,
    "timestamp": "2025-07-14T12:33:43.745498"
  },
  {
    "epoch": 291,
    "batch_idx": 128,
    "loss": 0.5722085237503052,
    "learning_rate": 0.0009486677066579459,
    "timestep_mean": 444.3125,
    "epsilon_norm": 8.587498664855957,
    "prediction_norm": 6.075716495513916,
    "timestamp": "2025-07-14T12:33:43.750680"
  },
  {
    "epoch": 291,
    "batch_idx": 160,
    "loss": 0.41504979133605957,
    "learning_rate": 0.0009486677066579459,
    "timestep_mean": 475.78125,
    "epsilon_norm": 8.768874168395996,
    "prediction_norm": 6.274311065673828,
    "timestamp": "2025-07-14T12:33:43.755593"
  },
  {
    "epoch": 291,
    "batch_idx": 192,
    "loss": 0.9514535665512085,
    "learning_rate": 0.0009486677066579459,
    "timestep_mean": 512.125,
    "epsilon_norm": 5.7912821769714355,
    "prediction_norm": 4.057114601135254,
    "timestamp": "2025-07-14T12:33:43.760681"
  },
  {
    "epoch": 292,
    "batch_idx": 0,
    "loss": 0.3162926435470581,
    "learning_rate": 0.0009483205183926181,
    "timestep_mean": 517.40625,
    "epsilon_norm": 7.788922309875488,
    "prediction_norm": 6.533851623535156,
    "timestamp": "2025-07-14T12:33:43.765853"
  },
  {
    "epoch": 292,
    "batch_idx": 32,
    "loss": 0.340827077627182,
    "learning_rate": 0.0009483205183926181,
    "timestep_mean": 596.1875,
    "epsilon_norm": 8.16195011138916,
    "prediction_norm": 6.342129230499268,
    "timestamp": "2025-07-14T12:33:43.770877"
  },
  {
    "epoch": 292,
    "batch_idx": 64,
    "loss": 0.1744796633720398,
    "learning_rate": 0.0009483205183926181,
    "timestep_mean": 548.96875,
    "epsilon_norm": 8.418747901916504,
    "prediction_norm": 6.748696327209473,
    "timestamp": "2025-07-14T12:33:43.775855"
  },
  {
    "epoch": 292,
    "batch_idx": 96,
    "loss": 0.4037711024284363,
    "learning_rate": 0.0009483205183926181,
    "timestep_mean": 458.46875,
    "epsilon_norm": 8.26524829864502,
    "prediction_norm": 6.217696189880371,
    "timestamp": "2025-07-14T12:33:43.780958"
  },
  {
    "epoch": 292,
    "batch_idx": 128,
    "loss": 0.39895468950271606,
    "learning_rate": 0.0009483205183926181,
    "timestep_mean": 460.21875,
    "epsilon_norm": 8.116792678833008,
    "prediction_norm": 7.140500068664551,
    "timestamp": "2025-07-14T12:33:43.786180"
  },
  {
    "epoch": 292,
    "batch_idx": 160,
    "loss": 0.15225109457969666,
    "learning_rate": 0.0009483205183926181,
    "timestep_mean": 512.625,
    "epsilon_norm": 6.391591548919678,
    "prediction_norm": 5.968740940093994,
    "timestamp": "2025-07-14T12:33:43.791330"
  },
  {
    "epoch": 292,
    "batch_idx": 192,
    "loss": 0.2588142156600952,
    "learning_rate": 0.0009483205183926181,
    "timestep_mean": 475.5,
    "epsilon_norm": 4.5174994468688965,
    "prediction_norm": 4.404111385345459,
    "timestamp": "2025-07-14T12:33:43.796382"
  },
  {
    "epoch": 293,
    "batch_idx": 0,
    "loss": 0.2584855258464813,
    "learning_rate": 0.0009479722239409775,
    "timestep_mean": 524.71875,
    "epsilon_norm": 8.064905166625977,
    "prediction_norm": 6.537012577056885,
    "timestamp": "2025-07-14T12:33:43.801723"
  },
  {
    "epoch": 293,
    "batch_idx": 32,
    "loss": 0.44443196058273315,
    "learning_rate": 0.0009479722239409775,
    "timestep_mean": 510.875,
    "epsilon_norm": 9.072032928466797,
    "prediction_norm": 7.1451263427734375,
    "timestamp": "2025-07-14T12:33:43.806852"
  },
  {
    "epoch": 293,
    "batch_idx": 64,
    "loss": 0.2861091196537018,
    "learning_rate": 0.0009479722239409775,
    "timestep_mean": 540.34375,
    "epsilon_norm": 8.661910057067871,
    "prediction_norm": 7.404699325561523,
    "timestamp": "2025-07-14T12:33:43.811930"
  },
  {
    "epoch": 293,
    "batch_idx": 96,
    "loss": 0.21056130528450012,
    "learning_rate": 0.0009479722239409775,
    "timestep_mean": 488.15625,
    "epsilon_norm": 7.200410842895508,
    "prediction_norm": 6.2396240234375,
    "timestamp": "2025-07-14T12:33:43.816911"
  },
  {
    "epoch": 293,
    "batch_idx": 128,
    "loss": 0.34963852167129517,
    "learning_rate": 0.0009479722239409775,
    "timestep_mean": 481.375,
    "epsilon_norm": 8.628558158874512,
    "prediction_norm": 7.709107398986816,
    "timestamp": "2025-07-14T12:33:43.822043"
  },
  {
    "epoch": 293,
    "batch_idx": 160,
    "loss": 0.26501980423927307,
    "learning_rate": 0.0009479722239409775,
    "timestep_mean": 499.96875,
    "epsilon_norm": 7.622534275054932,
    "prediction_norm": 6.275333881378174,
    "timestamp": "2025-07-14T12:33:43.827311"
  },
  {
    "epoch": 293,
    "batch_idx": 192,
    "loss": 0.5604407787322998,
    "learning_rate": 0.0009479722239409775,
    "timestep_mean": 676.625,
    "epsilon_norm": 4.502326011657715,
    "prediction_norm": 3.729743003845215,
    "timestamp": "2025-07-14T12:33:43.832567"
  },
  {
    "epoch": 294,
    "batch_idx": 0,
    "loss": 0.3087858557701111,
    "learning_rate": 0.0009476228241624059,
    "timestep_mean": 508.59375,
    "epsilon_norm": 7.90500020980835,
    "prediction_norm": 7.386854648590088,
    "timestamp": "2025-07-14T12:33:43.837968"
  },
  {
    "epoch": 294,
    "batch_idx": 32,
    "loss": 0.40726810693740845,
    "learning_rate": 0.0009476228241624059,
    "timestep_mean": 558.125,
    "epsilon_norm": 8.073263168334961,
    "prediction_norm": 6.366892337799072,
    "timestamp": "2025-07-14T12:33:43.843219"
  },
  {
    "epoch": 294,
    "batch_idx": 64,
    "loss": 0.15424515306949615,
    "learning_rate": 0.0009476228241624059,
    "timestep_mean": 533.375,
    "epsilon_norm": 8.321022987365723,
    "prediction_norm": 8.117642402648926,
    "timestamp": "2025-07-14T12:33:43.848415"
  },
  {
    "epoch": 294,
    "batch_idx": 96,
    "loss": 0.438045859336853,
    "learning_rate": 0.0009476228241624059,
    "timestep_mean": 470.78125,
    "epsilon_norm": 8.05429458618164,
    "prediction_norm": 6.545435905456543,
    "timestamp": "2025-07-14T12:33:43.853420"
  },
  {
    "epoch": 294,
    "batch_idx": 128,
    "loss": 0.26511725783348083,
    "learning_rate": 0.0009476228241624059,
    "timestep_mean": 475.9375,
    "epsilon_norm": 7.318776607513428,
    "prediction_norm": 5.956703186035156,
    "timestamp": "2025-07-14T12:33:43.858797"
  },
  {
    "epoch": 294,
    "batch_idx": 160,
    "loss": 0.19552044570446014,
    "learning_rate": 0.0009476228241624059,
    "timestep_mean": 532.09375,
    "epsilon_norm": 8.305910110473633,
    "prediction_norm": 7.231509685516357,
    "timestamp": "2025-07-14T12:33:43.863998"
  },
  {
    "epoch": 294,
    "batch_idx": 192,
    "loss": 0.08486360311508179,
    "learning_rate": 0.0009476228241624059,
    "timestep_mean": 625.0,
    "epsilon_norm": 3.0726358890533447,
    "prediction_norm": 2.845607042312622,
    "timestamp": "2025-07-14T12:33:43.868997"
  },
  {
    "epoch": 295,
    "batch_idx": 0,
    "loss": 0.22994668781757355,
    "learning_rate": 0.0009472723199190126,
    "timestep_mean": 461.875,
    "epsilon_norm": 6.4048566818237305,
    "prediction_norm": 5.275302886962891,
    "timestamp": "2025-07-14T12:33:43.874202"
  },
  {
    "epoch": 295,
    "batch_idx": 32,
    "loss": 0.37447425723075867,
    "learning_rate": 0.0009472723199190126,
    "timestep_mean": 502.9375,
    "epsilon_norm": 8.39098834991455,
    "prediction_norm": 6.650120735168457,
    "timestamp": "2025-07-14T12:33:43.879460"
  },
  {
    "epoch": 295,
    "batch_idx": 64,
    "loss": 0.2828803062438965,
    "learning_rate": 0.0009472723199190126,
    "timestep_mean": 463.875,
    "epsilon_norm": 8.656293869018555,
    "prediction_norm": 7.14120626449585,
    "timestamp": "2025-07-14T12:33:43.884477"
  },
  {
    "epoch": 295,
    "batch_idx": 96,
    "loss": 0.2800787091255188,
    "learning_rate": 0.0009472723199190126,
    "timestep_mean": 597.84375,
    "epsilon_norm": 7.63676118850708,
    "prediction_norm": 6.242681503295898,
    "timestamp": "2025-07-14T12:33:43.889353"
  },
  {
    "epoch": 295,
    "batch_idx": 128,
    "loss": 0.2947509288787842,
    "learning_rate": 0.0009472723199190126,
    "timestep_mean": 494.8125,
    "epsilon_norm": 9.067557334899902,
    "prediction_norm": 7.207055568695068,
    "timestamp": "2025-07-14T12:33:43.894285"
  },
  {
    "epoch": 295,
    "batch_idx": 160,
    "loss": 0.3947070240974426,
    "learning_rate": 0.0009472723199190126,
    "timestep_mean": 478.78125,
    "epsilon_norm": 8.43501091003418,
    "prediction_norm": 7.09042501449585,
    "timestamp": "2025-07-14T12:33:43.899161"
  },
  {
    "epoch": 295,
    "batch_idx": 192,
    "loss": 0.44839614629745483,
    "learning_rate": 0.0009472723199190126,
    "timestep_mean": 410.625,
    "epsilon_norm": 3.7309188842773438,
    "prediction_norm": 2.765873432159424,
    "timestamp": "2025-07-14T12:33:43.903967"
  },
  {
    "epoch": 296,
    "batch_idx": 0,
    "loss": 0.3994876444339752,
    "learning_rate": 0.0009469207120756319,
    "timestep_mean": 493.875,
    "epsilon_norm": 7.984034061431885,
    "prediction_norm": 6.342176914215088,
    "timestamp": "2025-07-14T12:33:43.908952"
  },
  {
    "epoch": 296,
    "batch_idx": 32,
    "loss": 0.17606580257415771,
    "learning_rate": 0.0009469207120756319,
    "timestep_mean": 617.71875,
    "epsilon_norm": 6.981696128845215,
    "prediction_norm": 6.000242710113525,
    "timestamp": "2025-07-14T12:33:43.913880"
  },
  {
    "epoch": 296,
    "batch_idx": 64,
    "loss": 0.35127025842666626,
    "learning_rate": 0.0009469207120756319,
    "timestep_mean": 504.21875,
    "epsilon_norm": 8.319480895996094,
    "prediction_norm": 7.647485733032227,
    "timestamp": "2025-07-14T12:33:43.918780"
  },
  {
    "epoch": 296,
    "batch_idx": 96,
    "loss": 0.34703582525253296,
    "learning_rate": 0.0009469207120756319,
    "timestep_mean": 488.21875,
    "epsilon_norm": 7.327659606933594,
    "prediction_norm": 5.815189838409424,
    "timestamp": "2025-07-14T12:33:43.923684"
  },
  {
    "epoch": 296,
    "batch_idx": 128,
    "loss": 0.15733490884304047,
    "learning_rate": 0.0009469207120756319,
    "timestep_mean": 526.84375,
    "epsilon_norm": 8.666464805603027,
    "prediction_norm": 8.386492729187012,
    "timestamp": "2025-07-14T12:33:43.928637"
  },
  {
    "epoch": 296,
    "batch_idx": 160,
    "loss": 0.22636504471302032,
    "learning_rate": 0.0009469207120756319,
    "timestep_mean": 588.0,
    "epsilon_norm": 8.222555160522461,
    "prediction_norm": 7.146106243133545,
    "timestamp": "2025-07-14T12:33:43.933752"
  },
  {
    "epoch": 296,
    "batch_idx": 192,
    "loss": 0.2802479565143585,
    "learning_rate": 0.0009469207120756319,
    "timestep_mean": 590.125,
    "epsilon_norm": 5.522658348083496,
    "prediction_norm": 4.54299259185791,
    "timestamp": "2025-07-14T12:33:43.938824"
  },
  {
    "epoch": 297,
    "batch_idx": 0,
    "loss": 0.29025954008102417,
    "learning_rate": 0.0009465680014998213,
    "timestep_mean": 423.6875,
    "epsilon_norm": 7.109519004821777,
    "prediction_norm": 5.542810916900635,
    "timestamp": "2025-07-14T12:33:43.943987"
  },
  {
    "epoch": 297,
    "batch_idx": 32,
    "loss": 0.2503524422645569,
    "learning_rate": 0.0009465680014998213,
    "timestep_mean": 489.40625,
    "epsilon_norm": 8.970248222351074,
    "prediction_norm": 7.753088474273682,
    "timestamp": "2025-07-14T12:33:43.949112"
  },
  {
    "epoch": 297,
    "batch_idx": 64,
    "loss": 0.3785260319709778,
    "learning_rate": 0.0009465680014998213,
    "timestep_mean": 521.84375,
    "epsilon_norm": 8.02007007598877,
    "prediction_norm": 6.118301868438721,
    "timestamp": "2025-07-14T12:33:43.954225"
  },
  {
    "epoch": 297,
    "batch_idx": 96,
    "loss": 0.32244133949279785,
    "learning_rate": 0.0009465680014998213,
    "timestep_mean": 490.75,
    "epsilon_norm": 8.559189796447754,
    "prediction_norm": 6.8836259841918945,
    "timestamp": "2025-07-14T12:33:43.959340"
  },
  {
    "epoch": 297,
    "batch_idx": 128,
    "loss": 0.3893614411354065,
    "learning_rate": 0.0009465680014998213,
    "timestep_mean": 430.75,
    "epsilon_norm": 7.838315963745117,
    "prediction_norm": 6.426507949829102,
    "timestamp": "2025-07-14T12:33:43.964483"
  },
  {
    "epoch": 297,
    "batch_idx": 160,
    "loss": 0.2765830159187317,
    "learning_rate": 0.0009465680014998213,
    "timestep_mean": 530.75,
    "epsilon_norm": 7.172094821929932,
    "prediction_norm": 6.159479141235352,
    "timestamp": "2025-07-14T12:33:43.969753"
  },
  {
    "epoch": 297,
    "batch_idx": 192,
    "loss": 0.26456570625305176,
    "learning_rate": 0.0009465680014998213,
    "timestep_mean": 470.0,
    "epsilon_norm": 3.1299350261688232,
    "prediction_norm": 2.3368289470672607,
    "timestamp": "2025-07-14T12:33:43.975040"
  },
  {
    "epoch": 298,
    "batch_idx": 0,
    "loss": 0.3943202495574951,
    "learning_rate": 0.000946214189061859,
    "timestep_mean": 496.5,
    "epsilon_norm": 7.8874006271362305,
    "prediction_norm": 5.915721416473389,
    "timestamp": "2025-07-14T12:33:43.980317"
  },
  {
    "epoch": 298,
    "batch_idx": 32,
    "loss": 0.1437932252883911,
    "learning_rate": 0.000946214189061859,
    "timestep_mean": 501.34375,
    "epsilon_norm": 7.921999454498291,
    "prediction_norm": 7.231447219848633,
    "timestamp": "2025-07-14T12:33:43.985656"
  },
  {
    "epoch": 298,
    "batch_idx": 64,
    "loss": 0.2555316686630249,
    "learning_rate": 0.000946214189061859,
    "timestep_mean": 471.46875,
    "epsilon_norm": 7.264331817626953,
    "prediction_norm": 6.090349197387695,
    "timestamp": "2025-07-14T12:33:43.990788"
  },
  {
    "epoch": 298,
    "batch_idx": 96,
    "loss": 0.4237869381904602,
    "learning_rate": 0.000946214189061859,
    "timestep_mean": 554.8125,
    "epsilon_norm": 8.196137428283691,
    "prediction_norm": 6.5170063972473145,
    "timestamp": "2025-07-14T12:33:43.995782"
  },
  {
    "epoch": 298,
    "batch_idx": 128,
    "loss": 0.33687344193458557,
    "learning_rate": 0.000946214189061859,
    "timestep_mean": 502.96875,
    "epsilon_norm": 8.632125854492188,
    "prediction_norm": 6.983222007751465,
    "timestamp": "2025-07-14T12:33:44.000839"
  },
  {
    "epoch": 298,
    "batch_idx": 160,
    "loss": 0.31244516372680664,
    "learning_rate": 0.000946214189061859,
    "timestep_mean": 443.375,
    "epsilon_norm": 8.501773834228516,
    "prediction_norm": 7.173810005187988,
    "timestamp": "2025-07-14T12:33:44.005870"
  },
  {
    "epoch": 298,
    "batch_idx": 192,
    "loss": 0.12110153585672379,
    "learning_rate": 0.000946214189061859,
    "timestep_mean": 584.625,
    "epsilon_norm": 3.9524636268615723,
    "prediction_norm": 4.05687141418457,
    "timestamp": "2025-07-14T12:33:44.010999"
  },
  {
    "epoch": 299,
    "batch_idx": 0,
    "loss": 0.649905800819397,
    "learning_rate": 0.000945859275634742,
    "timestep_mean": 415.34375,
    "epsilon_norm": 9.112911224365234,
    "prediction_norm": 6.172330379486084,
    "timestamp": "2025-07-14T12:33:44.015951"
  },
  {
    "epoch": 299,
    "batch_idx": 32,
    "loss": 0.24472267925739288,
    "learning_rate": 0.000945859275634742,
    "timestep_mean": 377.1875,
    "epsilon_norm": 6.546050548553467,
    "prediction_norm": 5.314480781555176,
    "timestamp": "2025-07-14T12:33:44.020919"
  },
  {
    "epoch": 299,
    "batch_idx": 64,
    "loss": 0.506339430809021,
    "learning_rate": 0.000945859275634742,
    "timestep_mean": 493.09375,
    "epsilon_norm": 9.095796585083008,
    "prediction_norm": 7.042329788208008,
    "timestamp": "2025-07-14T12:33:44.025825"
  },
  {
    "epoch": 299,
    "batch_idx": 96,
    "loss": 0.35509419441223145,
    "learning_rate": 0.000945859275634742,
    "timestep_mean": 473.53125,
    "epsilon_norm": 8.643277168273926,
    "prediction_norm": 7.0981316566467285,
    "timestamp": "2025-07-14T12:33:44.030770"
  },
  {
    "epoch": 299,
    "batch_idx": 128,
    "loss": 0.0983162447810173,
    "learning_rate": 0.000945859275634742,
    "timestep_mean": 624.6875,
    "epsilon_norm": 9.183856010437012,
    "prediction_norm": 8.836013793945312,
    "timestamp": "2025-07-14T12:33:44.035929"
  },
  {
    "epoch": 299,
    "batch_idx": 160,
    "loss": 0.08645975589752197,
    "learning_rate": 0.000945859275634742,
    "timestep_mean": 625.96875,
    "epsilon_norm": 7.660700798034668,
    "prediction_norm": 7.307703971862793,
    "timestamp": "2025-07-14T12:33:44.041457"
  },
  {
    "epoch": 299,
    "batch_idx": 192,
    "loss": 0.3452412188053131,
    "learning_rate": 0.000945859275634742,
    "timestep_mean": 500.125,
    "epsilon_norm": 3.2554879188537598,
    "prediction_norm": 2.7982735633850098,
    "timestamp": "2025-07-14T12:33:44.046487"
  },
  {
    "epoch": 300,
    "batch_idx": 0,
    "loss": 0.4034804403781891,
    "learning_rate": 0.0009455032620941839,
    "timestep_mean": 416.90625,
    "epsilon_norm": 8.687883377075195,
    "prediction_norm": 6.5208306312561035,
    "timestamp": "2025-07-14T12:33:44.051600"
  },
  {
    "epoch": 300,
    "batch_idx": 32,
    "loss": 0.31714123487472534,
    "learning_rate": 0.0009455032620941839,
    "timestep_mean": 586.3125,
    "epsilon_norm": 7.4266462326049805,
    "prediction_norm": 5.84691858291626,
    "timestamp": "2025-07-14T12:33:44.056749"
  },
  {
    "epoch": 300,
    "batch_idx": 64,
    "loss": 0.2675716280937195,
    "learning_rate": 0.0009455032620941839,
    "timestep_mean": 598.875,
    "epsilon_norm": 6.7415690422058105,
    "prediction_norm": 5.661332130432129,
    "timestamp": "2025-07-14T12:33:44.061924"
  },
  {
    "epoch": 300,
    "batch_idx": 96,
    "loss": 0.32597750425338745,
    "learning_rate": 0.0009455032620941839,
    "timestep_mean": 519.5625,
    "epsilon_norm": 8.762388229370117,
    "prediction_norm": 7.825619697570801,
    "timestamp": "2025-07-14T12:33:44.067057"
  },
  {
    "epoch": 300,
    "batch_idx": 128,
    "loss": 0.3916771113872528,
    "learning_rate": 0.0009455032620941839,
    "timestep_mean": 492.59375,
    "epsilon_norm": 7.578874111175537,
    "prediction_norm": 6.154804706573486,
    "timestamp": "2025-07-14T12:33:44.072065"
  },
  {
    "epoch": 300,
    "batch_idx": 160,
    "loss": 0.16883420944213867,
    "learning_rate": 0.0009455032620941839,
    "timestep_mean": 555.0,
    "epsilon_norm": 6.337752342224121,
    "prediction_norm": 5.660130500793457,
    "timestamp": "2025-07-14T12:33:44.077142"
  },
  {
    "epoch": 300,
    "batch_idx": 192,
    "loss": 0.31765374541282654,
    "learning_rate": 0.0009455032620941839,
    "timestep_mean": 580.375,
    "epsilon_norm": 3.512927532196045,
    "prediction_norm": 3.704134464263916,
    "timestamp": "2025-07-14T12:33:44.082325"
  },
  {
    "epoch": 301,
    "batch_idx": 0,
    "loss": 0.3743511140346527,
    "learning_rate": 0.0009451461493186129,
    "timestep_mean": 408.34375,
    "epsilon_norm": 8.291205406188965,
    "prediction_norm": 7.128695487976074,
    "timestamp": "2025-07-14T12:33:44.087621"
  },
  {
    "epoch": 301,
    "batch_idx": 32,
    "loss": 0.25829771161079407,
    "learning_rate": 0.0009451461493186129,
    "timestep_mean": 571.15625,
    "epsilon_norm": 7.686369895935059,
    "prediction_norm": 6.863716125488281,
    "timestamp": "2025-07-14T12:33:44.098927"
  },
  {
    "epoch": 301,
    "batch_idx": 64,
    "loss": 0.497531920671463,
    "learning_rate": 0.0009451461493186129,
    "timestep_mean": 449.9375,
    "epsilon_norm": 8.20654296875,
    "prediction_norm": 5.939269542694092,
    "timestamp": "2025-07-14T12:33:44.109239"
  },
  {
    "epoch": 301,
    "batch_idx": 96,
    "loss": 0.3699810802936554,
    "learning_rate": 0.0009451461493186129,
    "timestep_mean": 484.875,
    "epsilon_norm": 8.0306396484375,
    "prediction_norm": 5.9098334312438965,
    "timestamp": "2025-07-14T12:33:44.115707"
  },
  {
    "epoch": 301,
    "batch_idx": 128,
    "loss": 0.38766393065452576,
    "learning_rate": 0.0009451461493186129,
    "timestep_mean": 500.3125,
    "epsilon_norm": 8.09057331085205,
    "prediction_norm": 6.315675258636475,
    "timestamp": "2025-07-14T12:33:44.121355"
  },
  {
    "epoch": 301,
    "batch_idx": 160,
    "loss": 0.4980757534503937,
    "learning_rate": 0.0009451461493186129,
    "timestep_mean": 518.1875,
    "epsilon_norm": 9.124058723449707,
    "prediction_norm": 6.547987937927246,
    "timestamp": "2025-07-14T12:33:44.127443"
  },
  {
    "epoch": 301,
    "batch_idx": 192,
    "loss": 0.29603374004364014,
    "learning_rate": 0.0009451461493186129,
    "timestep_mean": 634.375,
    "epsilon_norm": 4.099339008331299,
    "prediction_norm": 2.6534104347229004,
    "timestamp": "2025-07-14T12:33:44.132954"
  },
  {
    "epoch": 302,
    "batch_idx": 0,
    "loss": 0.5133895874023438,
    "learning_rate": 0.0009447879381891691,
    "timestep_mean": 471.78125,
    "epsilon_norm": 8.514453887939453,
    "prediction_norm": 6.461006164550781,
    "timestamp": "2025-07-14T12:33:44.138401"
  },
  {
    "epoch": 302,
    "batch_idx": 32,
    "loss": 0.24308235943317413,
    "learning_rate": 0.0009447879381891691,
    "timestep_mean": 491.4375,
    "epsilon_norm": 8.213787078857422,
    "prediction_norm": 7.261626720428467,
    "timestamp": "2025-07-14T12:33:44.144145"
  },
  {
    "epoch": 302,
    "batch_idx": 64,
    "loss": 0.40587589144706726,
    "learning_rate": 0.0009447879381891691,
    "timestep_mean": 452.03125,
    "epsilon_norm": 7.687921047210693,
    "prediction_norm": 6.058546543121338,
    "timestamp": "2025-07-14T12:33:44.149508"
  },
  {
    "epoch": 302,
    "batch_idx": 96,
    "loss": 0.43161922693252563,
    "learning_rate": 0.0009447879381891691,
    "timestep_mean": 495.40625,
    "epsilon_norm": 8.617172241210938,
    "prediction_norm": 6.43179988861084,
    "timestamp": "2025-07-14T12:33:44.155104"
  },
  {
    "epoch": 302,
    "batch_idx": 128,
    "loss": 0.23346397280693054,
    "learning_rate": 0.0009447879381891691,
    "timestep_mean": 574.03125,
    "epsilon_norm": 8.2391357421875,
    "prediction_norm": 6.9779133796691895,
    "timestamp": "2025-07-14T12:33:44.160870"
  },
  {
    "epoch": 302,
    "batch_idx": 160,
    "loss": 0.5246378779411316,
    "learning_rate": 0.0009447879381891691,
    "timestep_mean": 428.5,
    "epsilon_norm": 9.083999633789062,
    "prediction_norm": 6.584704399108887,
    "timestamp": "2025-07-14T12:33:44.166102"
  },
  {
    "epoch": 302,
    "batch_idx": 192,
    "loss": 0.16201119124889374,
    "learning_rate": 0.0009447879381891691,
    "timestep_mean": 717.75,
    "epsilon_norm": 4.367132186889648,
    "prediction_norm": 4.179272651672363,
    "timestamp": "2025-07-14T12:33:44.171905"
  },
  {
    "epoch": 303,
    "batch_idx": 0,
    "loss": 0.4923763871192932,
    "learning_rate": 0.0009444286295897027,
    "timestep_mean": 382.25,
    "epsilon_norm": 8.37208366394043,
    "prediction_norm": 5.845190525054932,
    "timestamp": "2025-07-14T12:33:44.177448"
  },
  {
    "epoch": 303,
    "batch_idx": 32,
    "loss": 0.33898353576660156,
    "learning_rate": 0.0009444286295897027,
    "timestep_mean": 492.75,
    "epsilon_norm": 8.996074676513672,
    "prediction_norm": 7.834615707397461,
    "timestamp": "2025-07-14T12:33:44.182557"
  },
  {
    "epoch": 303,
    "batch_idx": 64,
    "loss": 0.1904754787683487,
    "learning_rate": 0.0009444286295897027,
    "timestep_mean": 528.5,
    "epsilon_norm": 7.172072887420654,
    "prediction_norm": 6.427820205688477,
    "timestamp": "2025-07-14T12:33:44.187472"
  },
  {
    "epoch": 303,
    "batch_idx": 96,
    "loss": 0.4241962134838104,
    "learning_rate": 0.0009444286295897027,
    "timestep_mean": 559.78125,
    "epsilon_norm": 7.595330238342285,
    "prediction_norm": 6.375924587249756,
    "timestamp": "2025-07-14T12:33:44.192513"
  },
  {
    "epoch": 303,
    "batch_idx": 128,
    "loss": 0.3706049919128418,
    "learning_rate": 0.0009444286295897027,
    "timestep_mean": 535.0,
    "epsilon_norm": 8.02413272857666,
    "prediction_norm": 7.534745216369629,
    "timestamp": "2025-07-14T12:33:44.197774"
  },
  {
    "epoch": 303,
    "batch_idx": 160,
    "loss": 0.43259871006011963,
    "learning_rate": 0.0009444286295897027,
    "timestep_mean": 482.84375,
    "epsilon_norm": 8.881080627441406,
    "prediction_norm": 8.302698135375977,
    "timestamp": "2025-07-14T12:33:44.202751"
  },
  {
    "epoch": 303,
    "batch_idx": 192,
    "loss": 0.18162694573402405,
    "learning_rate": 0.0009444286295897027,
    "timestep_mean": 743.375,
    "epsilon_norm": 3.7714974880218506,
    "prediction_norm": 3.517674207687378,
    "timestamp": "2025-07-14T12:33:44.208007"
  },
  {
    "epoch": 304,
    "batch_idx": 0,
    "loss": 0.3185010552406311,
    "learning_rate": 0.0009440682244067724,
    "timestep_mean": 532.625,
    "epsilon_norm": 8.208925247192383,
    "prediction_norm": 7.243027210235596,
    "timestamp": "2025-07-14T12:33:44.214353"
  },
  {
    "epoch": 304,
    "batch_idx": 32,
    "loss": 0.2869075834751129,
    "learning_rate": 0.0009440682244067724,
    "timestep_mean": 498.28125,
    "epsilon_norm": 8.170647621154785,
    "prediction_norm": 7.09319543838501,
    "timestamp": "2025-07-14T12:33:44.220375"
  },
  {
    "epoch": 304,
    "batch_idx": 64,
    "loss": 0.3152880072593689,
    "learning_rate": 0.0009440682244067724,
    "timestep_mean": 561.6875,
    "epsilon_norm": 7.24281120300293,
    "prediction_norm": 5.334068775177002,
    "timestamp": "2025-07-14T12:33:44.225581"
  },
  {
    "epoch": 304,
    "batch_idx": 96,
    "loss": 0.477895051240921,
    "learning_rate": 0.0009440682244067724,
    "timestep_mean": 397.5,
    "epsilon_norm": 7.866332530975342,
    "prediction_norm": 5.70941162109375,
    "timestamp": "2025-07-14T12:33:44.230882"
  },
  {
    "epoch": 304,
    "batch_idx": 128,
    "loss": 0.2598576247692108,
    "learning_rate": 0.0009440682244067724,
    "timestep_mean": 507.3125,
    "epsilon_norm": 7.47178316116333,
    "prediction_norm": 6.13486909866333,
    "timestamp": "2025-07-14T12:33:44.235887"
  },
  {
    "epoch": 304,
    "batch_idx": 160,
    "loss": 0.23038248717784882,
    "learning_rate": 0.0009440682244067724,
    "timestep_mean": 490.125,
    "epsilon_norm": 6.9833760261535645,
    "prediction_norm": 5.711684226989746,
    "timestamp": "2025-07-14T12:33:44.240892"
  },
  {
    "epoch": 304,
    "batch_idx": 192,
    "loss": 0.17138135433197021,
    "learning_rate": 0.0009440682244067724,
    "timestep_mean": 466.75,
    "epsilon_norm": 3.1750152111053467,
    "prediction_norm": 2.6159424781799316,
    "timestamp": "2025-07-14T12:33:44.246380"
  },
  {
    "epoch": 305,
    "batch_idx": 0,
    "loss": 0.29888656735420227,
    "learning_rate": 0.0009437067235296418,
    "timestep_mean": 477.25,
    "epsilon_norm": 7.971816539764404,
    "prediction_norm": 5.924543857574463,
    "timestamp": "2025-07-14T12:33:44.252421"
  },
  {
    "epoch": 305,
    "batch_idx": 32,
    "loss": 0.35591235756874084,
    "learning_rate": 0.0009437067235296418,
    "timestep_mean": 490.65625,
    "epsilon_norm": 8.126527786254883,
    "prediction_norm": 6.178677558898926,
    "timestamp": "2025-07-14T12:33:44.257788"
  },
  {
    "epoch": 305,
    "batch_idx": 64,
    "loss": 0.3870609700679779,
    "learning_rate": 0.0009437067235296418,
    "timestep_mean": 562.96875,
    "epsilon_norm": 8.750605583190918,
    "prediction_norm": 7.1792893409729,
    "timestamp": "2025-07-14T12:33:44.262905"
  },
  {
    "epoch": 305,
    "batch_idx": 96,
    "loss": 0.4517001509666443,
    "learning_rate": 0.0009437067235296418,
    "timestep_mean": 435.0,
    "epsilon_norm": 8.978607177734375,
    "prediction_norm": 6.9421186447143555,
    "timestamp": "2025-07-14T12:33:44.267901"
  },
  {
    "epoch": 305,
    "batch_idx": 128,
    "loss": 0.5100349187850952,
    "learning_rate": 0.0009437067235296418,
    "timestep_mean": 468.40625,
    "epsilon_norm": 7.616387367248535,
    "prediction_norm": 4.829473495483398,
    "timestamp": "2025-07-14T12:33:44.272973"
  },
  {
    "epoch": 305,
    "batch_idx": 160,
    "loss": 0.32550618052482605,
    "learning_rate": 0.0009437067235296418,
    "timestep_mean": 484.6875,
    "epsilon_norm": 8.297393798828125,
    "prediction_norm": 6.835019111633301,
    "timestamp": "2025-07-14T12:33:44.279030"
  },
  {
    "epoch": 305,
    "batch_idx": 192,
    "loss": 0.2882809638977051,
    "learning_rate": 0.0009437067235296418,
    "timestep_mean": 468.0,
    "epsilon_norm": 3.967648506164551,
    "prediction_norm": 3.262216329574585,
    "timestamp": "2025-07-14T12:33:44.284259"
  },
  {
    "epoch": 306,
    "batch_idx": 0,
    "loss": 0.4267584979534149,
    "learning_rate": 0.0009433441278502784,
    "timestep_mean": 479.6875,
    "epsilon_norm": 8.392251014709473,
    "prediction_norm": 6.1239542961120605,
    "timestamp": "2025-07-14T12:33:44.289483"
  },
  {
    "epoch": 306,
    "batch_idx": 32,
    "loss": 0.41661781072616577,
    "learning_rate": 0.0009433441278502784,
    "timestep_mean": 500.59375,
    "epsilon_norm": 8.51495361328125,
    "prediction_norm": 6.631791591644287,
    "timestamp": "2025-07-14T12:33:44.295505"
  },
  {
    "epoch": 306,
    "batch_idx": 64,
    "loss": 0.3907891511917114,
    "learning_rate": 0.0009433441278502784,
    "timestep_mean": 467.40625,
    "epsilon_norm": 7.397355079650879,
    "prediction_norm": 6.022799968719482,
    "timestamp": "2025-07-14T12:33:44.300697"
  },
  {
    "epoch": 306,
    "batch_idx": 96,
    "loss": 0.3830035924911499,
    "learning_rate": 0.0009433441278502784,
    "timestep_mean": 500.5625,
    "epsilon_norm": 8.176933288574219,
    "prediction_norm": 6.170395374298096,
    "timestamp": "2025-07-14T12:33:44.306691"
  },
  {
    "epoch": 306,
    "batch_idx": 128,
    "loss": 0.42758986353874207,
    "learning_rate": 0.0009433441278502784,
    "timestep_mean": 436.5625,
    "epsilon_norm": 7.607638359069824,
    "prediction_norm": 6.315819263458252,
    "timestamp": "2025-07-14T12:33:44.312010"
  },
  {
    "epoch": 306,
    "batch_idx": 160,
    "loss": 0.21313664317131042,
    "learning_rate": 0.0009433441278502784,
    "timestep_mean": 536.46875,
    "epsilon_norm": 8.054346084594727,
    "prediction_norm": 8.117559432983398,
    "timestamp": "2025-07-14T12:33:44.317112"
  },
  {
    "epoch": 306,
    "batch_idx": 192,
    "loss": 0.22539402544498444,
    "learning_rate": 0.0009433441278502784,
    "timestep_mean": 427.5,
    "epsilon_norm": 4.484469890594482,
    "prediction_norm": 3.6070008277893066,
    "timestamp": "2025-07-14T12:33:44.322213"
  },
  {
    "epoch": 307,
    "batch_idx": 0,
    "loss": 0.2071193903684616,
    "learning_rate": 0.000942980438263351,
    "timestep_mean": 588.90625,
    "epsilon_norm": 6.850804805755615,
    "prediction_norm": 6.312331199645996,
    "timestamp": "2025-07-14T12:33:44.327453"
  },
  {
    "epoch": 307,
    "batch_idx": 32,
    "loss": 0.24257774651050568,
    "learning_rate": 0.000942980438263351,
    "timestep_mean": 567.0,
    "epsilon_norm": 7.8718743324279785,
    "prediction_norm": 7.486760139465332,
    "timestamp": "2025-07-14T12:33:44.332578"
  },
  {
    "epoch": 307,
    "batch_idx": 64,
    "loss": 0.19091558456420898,
    "learning_rate": 0.000942980438263351,
    "timestep_mean": 478.5,
    "epsilon_norm": 7.674900054931641,
    "prediction_norm": 6.897159576416016,
    "timestamp": "2025-07-14T12:33:44.337574"
  },
  {
    "epoch": 307,
    "batch_idx": 96,
    "loss": 0.34560805559158325,
    "learning_rate": 0.000942980438263351,
    "timestep_mean": 505.28125,
    "epsilon_norm": 7.678585529327393,
    "prediction_norm": 6.083215713500977,
    "timestamp": "2025-07-14T12:33:44.342556"
  },
  {
    "epoch": 307,
    "batch_idx": 128,
    "loss": 0.2520064115524292,
    "learning_rate": 0.000942980438263351,
    "timestep_mean": 494.15625,
    "epsilon_norm": 7.968954563140869,
    "prediction_norm": 7.440489292144775,
    "timestamp": "2025-07-14T12:33:44.347539"
  },
  {
    "epoch": 307,
    "batch_idx": 160,
    "loss": 0.3838580846786499,
    "learning_rate": 0.000942980438263351,
    "timestep_mean": 536.0625,
    "epsilon_norm": 7.969798564910889,
    "prediction_norm": 6.597183704376221,
    "timestamp": "2025-07-14T12:33:44.353033"
  },
  {
    "epoch": 307,
    "batch_idx": 192,
    "loss": 0.1508917212486267,
    "learning_rate": 0.000942980438263351,
    "timestep_mean": 497.375,
    "epsilon_norm": 3.030184745788574,
    "prediction_norm": 2.8116602897644043,
    "timestamp": "2025-07-14T12:33:44.358155"
  },
  {
    "epoch": 308,
    "batch_idx": 0,
    "loss": 0.2992423474788666,
    "learning_rate": 0.0009426156556662277,
    "timestep_mean": 585.65625,
    "epsilon_norm": 8.676684379577637,
    "prediction_norm": 7.574132442474365,
    "timestamp": "2025-07-14T12:33:44.363395"
  },
  {
    "epoch": 308,
    "batch_idx": 32,
    "loss": 0.33425435423851013,
    "learning_rate": 0.0009426156556662277,
    "timestep_mean": 480.6875,
    "epsilon_norm": 6.990876197814941,
    "prediction_norm": 5.361337661743164,
    "timestamp": "2025-07-14T12:33:44.368444"
  },
  {
    "epoch": 308,
    "batch_idx": 64,
    "loss": 0.36941027641296387,
    "learning_rate": 0.0009426156556662277,
    "timestep_mean": 432.4375,
    "epsilon_norm": 7.503459453582764,
    "prediction_norm": 6.360117435455322,
    "timestamp": "2025-07-14T12:33:44.374149"
  },
  {
    "epoch": 308,
    "batch_idx": 96,
    "loss": 0.40337318181991577,
    "learning_rate": 0.0009426156556662277,
    "timestep_mean": 538.3125,
    "epsilon_norm": 7.527911186218262,
    "prediction_norm": 5.682023525238037,
    "timestamp": "2025-07-14T12:33:44.379071"
  },
  {
    "epoch": 308,
    "batch_idx": 128,
    "loss": 0.3617030084133148,
    "learning_rate": 0.0009426156556662277,
    "timestep_mean": 512.15625,
    "epsilon_norm": 8.402844429016113,
    "prediction_norm": 6.3267645835876465,
    "timestamp": "2025-07-14T12:33:44.384124"
  },
  {
    "epoch": 308,
    "batch_idx": 160,
    "loss": 0.3804055452346802,
    "learning_rate": 0.0009426156556662277,
    "timestep_mean": 429.5625,
    "epsilon_norm": 8.214112281799316,
    "prediction_norm": 6.525462627410889,
    "timestamp": "2025-07-14T12:33:44.389269"
  },
  {
    "epoch": 308,
    "batch_idx": 192,
    "loss": 0.4018365740776062,
    "learning_rate": 0.0009426156556662277,
    "timestep_mean": 496.125,
    "epsilon_norm": 5.025627136230469,
    "prediction_norm": 4.318049907684326,
    "timestamp": "2025-07-14T12:33:44.395138"
  },
  {
    "epoch": 309,
    "batch_idx": 0,
    "loss": 0.48627305030822754,
    "learning_rate": 0.0009422497809589732,
    "timestep_mean": 475.78125,
    "epsilon_norm": 8.55827808380127,
    "prediction_norm": 6.079723834991455,
    "timestamp": "2025-07-14T12:33:44.400375"
  },
  {
    "epoch": 309,
    "batch_idx": 32,
    "loss": 0.23030400276184082,
    "learning_rate": 0.0009422497809589732,
    "timestep_mean": 526.5,
    "epsilon_norm": 7.69527006149292,
    "prediction_norm": 6.409314155578613,
    "timestamp": "2025-07-14T12:33:44.405379"
  },
  {
    "epoch": 309,
    "batch_idx": 64,
    "loss": 0.1610921025276184,
    "learning_rate": 0.0009422497809589732,
    "timestep_mean": 561.65625,
    "epsilon_norm": 9.41823959350586,
    "prediction_norm": 8.989935874938965,
    "timestamp": "2025-07-14T12:33:44.410317"
  },
  {
    "epoch": 309,
    "batch_idx": 96,
    "loss": 0.3375703990459442,
    "learning_rate": 0.0009422497809589732,
    "timestep_mean": 505.59375,
    "epsilon_norm": 8.628965377807617,
    "prediction_norm": 6.967066764831543,
    "timestamp": "2025-07-14T12:33:44.415797"
  },
  {
    "epoch": 309,
    "batch_idx": 128,
    "loss": 0.35332852602005005,
    "learning_rate": 0.0009422497809589732,
    "timestep_mean": 464.1875,
    "epsilon_norm": 8.305685043334961,
    "prediction_norm": 6.146960735321045,
    "timestamp": "2025-07-14T12:33:44.421196"
  },
  {
    "epoch": 309,
    "batch_idx": 160,
    "loss": 0.41356462240219116,
    "learning_rate": 0.0009422497809589732,
    "timestep_mean": 492.5625,
    "epsilon_norm": 7.144815444946289,
    "prediction_norm": 5.235233783721924,
    "timestamp": "2025-07-14T12:33:44.426453"
  },
  {
    "epoch": 309,
    "batch_idx": 192,
    "loss": 0.2695973813533783,
    "learning_rate": 0.0009422497809589732,
    "timestep_mean": 379.375,
    "epsilon_norm": 3.2443089485168457,
    "prediction_norm": 2.9164750576019287,
    "timestamp": "2025-07-14T12:33:44.431822"
  },
  {
    "epoch": 310,
    "batch_idx": 0,
    "loss": 0.34739816188812256,
    "learning_rate": 0.0009418828150443469,
    "timestep_mean": 543.375,
    "epsilon_norm": 7.4447479248046875,
    "prediction_norm": 5.554076194763184,
    "timestamp": "2025-07-14T12:33:44.436975"
  },
  {
    "epoch": 310,
    "batch_idx": 32,
    "loss": 0.392769992351532,
    "learning_rate": 0.0009418828150443469,
    "timestep_mean": 458.8125,
    "epsilon_norm": 8.397704124450684,
    "prediction_norm": 7.071004867553711,
    "timestamp": "2025-07-14T12:33:44.442319"
  },
  {
    "epoch": 310,
    "batch_idx": 64,
    "loss": 0.3510593771934509,
    "learning_rate": 0.0009418828150443469,
    "timestep_mean": 430.90625,
    "epsilon_norm": 7.791094779968262,
    "prediction_norm": 5.2693586349487305,
    "timestamp": "2025-07-14T12:33:44.448500"
  },
  {
    "epoch": 310,
    "batch_idx": 96,
    "loss": 0.34687161445617676,
    "learning_rate": 0.0009418828150443469,
    "timestep_mean": 488.6875,
    "epsilon_norm": 8.463309288024902,
    "prediction_norm": 6.724893093109131,
    "timestamp": "2025-07-14T12:33:44.455425"
  },
  {
    "epoch": 310,
    "batch_idx": 128,
    "loss": 0.2196124792098999,
    "learning_rate": 0.0009418828150443469,
    "timestep_mean": 559.0625,
    "epsilon_norm": 7.7352070808410645,
    "prediction_norm": 7.218239784240723,
    "timestamp": "2025-07-14T12:33:44.460799"
  },
  {
    "epoch": 310,
    "batch_idx": 160,
    "loss": 0.32723110914230347,
    "learning_rate": 0.0009418828150443469,
    "timestep_mean": 516.5,
    "epsilon_norm": 7.954757213592529,
    "prediction_norm": 6.80149507522583,
    "timestamp": "2025-07-14T12:33:44.465941"
  },
  {
    "epoch": 310,
    "batch_idx": 192,
    "loss": 0.25338929891586304,
    "learning_rate": 0.0009418828150443469,
    "timestep_mean": 568.75,
    "epsilon_norm": 4.12007999420166,
    "prediction_norm": 4.259413719177246,
    "timestamp": "2025-07-14T12:33:44.471238"
  },
  {
    "epoch": 311,
    "batch_idx": 0,
    "loss": 0.2489445060491562,
    "learning_rate": 0.0009415147588278007,
    "timestep_mean": 497.3125,
    "epsilon_norm": 7.339078903198242,
    "prediction_norm": 6.597434043884277,
    "timestamp": "2025-07-14T12:33:44.476130"
  },
  {
    "epoch": 311,
    "batch_idx": 32,
    "loss": 0.2842828035354614,
    "learning_rate": 0.0009415147588278007,
    "timestep_mean": 491.4375,
    "epsilon_norm": 8.1062593460083,
    "prediction_norm": 6.598108291625977,
    "timestamp": "2025-07-14T12:33:44.481099"
  },
  {
    "epoch": 311,
    "batch_idx": 64,
    "loss": 0.3177689015865326,
    "learning_rate": 0.0009415147588278007,
    "timestep_mean": 547.28125,
    "epsilon_norm": 8.477967262268066,
    "prediction_norm": 7.313542366027832,
    "timestamp": "2025-07-14T12:33:44.486117"
  },
  {
    "epoch": 311,
    "batch_idx": 96,
    "loss": 0.38460689783096313,
    "learning_rate": 0.0009415147588278007,
    "timestep_mean": 450.75,
    "epsilon_norm": 7.5407843589782715,
    "prediction_norm": 6.032519817352295,
    "timestamp": "2025-07-14T12:33:44.491218"
  },
  {
    "epoch": 311,
    "batch_idx": 128,
    "loss": 0.3432968258857727,
    "learning_rate": 0.0009415147588278007,
    "timestep_mean": 433.53125,
    "epsilon_norm": 7.94008207321167,
    "prediction_norm": 6.201416492462158,
    "timestamp": "2025-07-14T12:33:44.496229"
  },
  {
    "epoch": 311,
    "batch_idx": 160,
    "loss": 0.22931775450706482,
    "learning_rate": 0.0009415147588278007,
    "timestep_mean": 462.5,
    "epsilon_norm": 8.556777954101562,
    "prediction_norm": 7.6266560554504395,
    "timestamp": "2025-07-14T12:33:44.501170"
  },
  {
    "epoch": 311,
    "batch_idx": 192,
    "loss": 0.5058915615081787,
    "learning_rate": 0.0009415147588278007,
    "timestep_mean": 490.25,
    "epsilon_norm": 4.241489410400391,
    "prediction_norm": 2.9291510581970215,
    "timestamp": "2025-07-14T12:33:44.506107"
  },
  {
    "epoch": 312,
    "batch_idx": 0,
    "loss": 0.15864905714988708,
    "learning_rate": 0.0009411456132174768,
    "timestep_mean": 457.4375,
    "epsilon_norm": 6.348081111907959,
    "prediction_norm": 5.86568021774292,
    "timestamp": "2025-07-14T12:33:44.512190"
  },
  {
    "epoch": 312,
    "batch_idx": 32,
    "loss": 0.16611003875732422,
    "learning_rate": 0.0009411456132174768,
    "timestep_mean": 523.3125,
    "epsilon_norm": 7.47551155090332,
    "prediction_norm": 6.453419208526611,
    "timestamp": "2025-07-14T12:33:44.517254"
  },
  {
    "epoch": 312,
    "batch_idx": 64,
    "loss": 0.21579784154891968,
    "learning_rate": 0.0009411456132174768,
    "timestep_mean": 556.875,
    "epsilon_norm": 7.579047203063965,
    "prediction_norm": 6.141852378845215,
    "timestamp": "2025-07-14T12:33:44.522210"
  },
  {
    "epoch": 312,
    "batch_idx": 96,
    "loss": 0.35204723477363586,
    "learning_rate": 0.0009411456132174768,
    "timestep_mean": 477.0,
    "epsilon_norm": 7.27876091003418,
    "prediction_norm": 5.2754082679748535,
    "timestamp": "2025-07-14T12:33:44.527135"
  },
  {
    "epoch": 312,
    "batch_idx": 128,
    "loss": 0.5295079350471497,
    "learning_rate": 0.0009411456132174768,
    "timestep_mean": 463.84375,
    "epsilon_norm": 9.080291748046875,
    "prediction_norm": 5.921959400177002,
    "timestamp": "2025-07-14T12:33:44.532043"
  },
  {
    "epoch": 312,
    "batch_idx": 160,
    "loss": 0.2988135516643524,
    "learning_rate": 0.0009411456132174768,
    "timestep_mean": 389.875,
    "epsilon_norm": 7.127275466918945,
    "prediction_norm": 5.256661415100098,
    "timestamp": "2025-07-14T12:33:44.536949"
  },
  {
    "epoch": 312,
    "batch_idx": 192,
    "loss": 0.07216119766235352,
    "learning_rate": 0.0009411456132174768,
    "timestep_mean": 581.5,
    "epsilon_norm": 3.473041296005249,
    "prediction_norm": 3.143571615219116,
    "timestamp": "2025-07-14T12:33:44.541949"
  },
  {
    "epoch": 313,
    "batch_idx": 0,
    "loss": 0.23970448970794678,
    "learning_rate": 0.0009407753791242052,
    "timestep_mean": 543.875,
    "epsilon_norm": 7.202841758728027,
    "prediction_norm": 5.678368091583252,
    "timestamp": "2025-07-14T12:33:44.546830"
  },
  {
    "epoch": 313,
    "batch_idx": 32,
    "loss": 0.23423272371292114,
    "learning_rate": 0.0009407753791242052,
    "timestep_mean": 530.78125,
    "epsilon_norm": 8.14113998413086,
    "prediction_norm": 7.498908042907715,
    "timestamp": "2025-07-14T12:33:44.551766"
  },
  {
    "epoch": 313,
    "batch_idx": 64,
    "loss": 0.5642032623291016,
    "learning_rate": 0.0009407753791242052,
    "timestep_mean": 432.40625,
    "epsilon_norm": 8.658342361450195,
    "prediction_norm": 6.39337158203125,
    "timestamp": "2025-07-14T12:33:44.556745"
  },
  {
    "epoch": 313,
    "batch_idx": 96,
    "loss": 0.21605387330055237,
    "learning_rate": 0.0009407753791242052,
    "timestep_mean": 439.75,
    "epsilon_norm": 7.638027191162109,
    "prediction_norm": 7.356625080108643,
    "timestamp": "2025-07-14T12:33:44.561734"
  },
  {
    "epoch": 313,
    "batch_idx": 128,
    "loss": 0.3712928295135498,
    "learning_rate": 0.0009407753791242052,
    "timestep_mean": 556.78125,
    "epsilon_norm": 8.29919719696045,
    "prediction_norm": 7.475476264953613,
    "timestamp": "2025-07-14T12:33:44.566695"
  },
  {
    "epoch": 313,
    "batch_idx": 160,
    "loss": 0.24069000780582428,
    "learning_rate": 0.0009407753791242052,
    "timestep_mean": 496.40625,
    "epsilon_norm": 6.824608325958252,
    "prediction_norm": 5.816011905670166,
    "timestamp": "2025-07-14T12:33:44.571556"
  },
  {
    "epoch": 313,
    "batch_idx": 192,
    "loss": 0.3513469994068146,
    "learning_rate": 0.0009407753791242052,
    "timestep_mean": 492.25,
    "epsilon_norm": 3.5834131240844727,
    "prediction_norm": 3.413670063018799,
    "timestamp": "2025-07-14T12:33:44.576532"
  },
  {
    "epoch": 314,
    "batch_idx": 0,
    "loss": 0.2983747720718384,
    "learning_rate": 0.0009404040574615019,
    "timestep_mean": 615.3125,
    "epsilon_norm": 7.946205139160156,
    "prediction_norm": 7.357501029968262,
    "timestamp": "2025-07-14T12:33:44.581479"
  },
  {
    "epoch": 314,
    "batch_idx": 32,
    "loss": 0.32375532388687134,
    "learning_rate": 0.0009404040574615019,
    "timestep_mean": 471.46875,
    "epsilon_norm": 8.477621078491211,
    "prediction_norm": 7.875897407531738,
    "timestamp": "2025-07-14T12:33:44.586354"
  },
  {
    "epoch": 314,
    "batch_idx": 64,
    "loss": 0.2804136276245117,
    "learning_rate": 0.0009404040574615019,
    "timestep_mean": 454.90625,
    "epsilon_norm": 7.189303398132324,
    "prediction_norm": 5.785415172576904,
    "timestamp": "2025-07-14T12:33:44.591303"
  },
  {
    "epoch": 314,
    "batch_idx": 96,
    "loss": 0.3366519808769226,
    "learning_rate": 0.0009404040574615019,
    "timestep_mean": 469.03125,
    "epsilon_norm": 7.542979717254639,
    "prediction_norm": 5.979731559753418,
    "timestamp": "2025-07-14T12:33:44.596266"
  },
  {
    "epoch": 314,
    "batch_idx": 128,
    "loss": 0.48365986347198486,
    "learning_rate": 0.0009404040574615019,
    "timestep_mean": 485.125,
    "epsilon_norm": 8.396023750305176,
    "prediction_norm": 5.8728251457214355,
    "timestamp": "2025-07-14T12:33:44.601150"
  },
  {
    "epoch": 314,
    "batch_idx": 160,
    "loss": 0.37090760469436646,
    "learning_rate": 0.0009404040574615019,
    "timestep_mean": 467.5,
    "epsilon_norm": 8.83098030090332,
    "prediction_norm": 7.388851165771484,
    "timestamp": "2025-07-14T12:33:44.606008"
  },
  {
    "epoch": 314,
    "batch_idx": 192,
    "loss": 0.4403170943260193,
    "learning_rate": 0.0009404040574615019,
    "timestep_mean": 552.625,
    "epsilon_norm": 3.888488292694092,
    "prediction_norm": 2.9988932609558105,
    "timestamp": "2025-07-14T12:33:44.610907"
  },
  {
    "epoch": 315,
    "batch_idx": 0,
    "loss": 0.2576512098312378,
    "learning_rate": 0.0009400316491455661,
    "timestep_mean": 499.75,
    "epsilon_norm": 7.289783477783203,
    "prediction_norm": 5.212705612182617,
    "timestamp": "2025-07-14T12:33:44.615880"
  },
  {
    "epoch": 315,
    "batch_idx": 32,
    "loss": 0.2614319920539856,
    "learning_rate": 0.0009400316491455661,
    "timestep_mean": 509.375,
    "epsilon_norm": 7.6879377365112305,
    "prediction_norm": 6.455900192260742,
    "timestamp": "2025-07-14T12:33:44.620907"
  },
  {
    "epoch": 315,
    "batch_idx": 64,
    "loss": 0.4578295946121216,
    "learning_rate": 0.0009400316491455661,
    "timestep_mean": 482.78125,
    "epsilon_norm": 8.016833305358887,
    "prediction_norm": 5.286067962646484,
    "timestamp": "2025-07-14T12:33:44.625796"
  },
  {
    "epoch": 315,
    "batch_idx": 96,
    "loss": 0.24864670634269714,
    "learning_rate": 0.0009400316491455661,
    "timestep_mean": 474.75,
    "epsilon_norm": 7.427969932556152,
    "prediction_norm": 6.0099968910217285,
    "timestamp": "2025-07-14T12:33:44.630650"
  },
  {
    "epoch": 315,
    "batch_idx": 128,
    "loss": 0.22773626446723938,
    "learning_rate": 0.0009400316491455661,
    "timestep_mean": 459.25,
    "epsilon_norm": 7.907585620880127,
    "prediction_norm": 5.505805015563965,
    "timestamp": "2025-07-14T12:33:44.635558"
  },
  {
    "epoch": 315,
    "batch_idx": 160,
    "loss": 0.5144651532173157,
    "learning_rate": 0.0009400316491455661,
    "timestep_mean": 499.9375,
    "epsilon_norm": 8.144424438476562,
    "prediction_norm": 6.022611618041992,
    "timestamp": "2025-07-14T12:33:44.640485"
  },
  {
    "epoch": 315,
    "batch_idx": 192,
    "loss": 0.7554811835289001,
    "learning_rate": 0.0009400316491455661,
    "timestep_mean": 309.25,
    "epsilon_norm": 3.832308769226074,
    "prediction_norm": 1.9612971544265747,
    "timestamp": "2025-07-14T12:33:44.645389"
  },
  {
    "epoch": 316,
    "batch_idx": 0,
    "loss": 0.2949504256248474,
    "learning_rate": 0.0009396581550952781,
    "timestep_mean": 480.875,
    "epsilon_norm": 6.466949939727783,
    "prediction_norm": 5.194095134735107,
    "timestamp": "2025-07-14T12:33:44.650427"
  },
  {
    "epoch": 316,
    "batch_idx": 32,
    "loss": 0.2537625730037689,
    "learning_rate": 0.0009396581550952781,
    "timestep_mean": 517.75,
    "epsilon_norm": 7.444360256195068,
    "prediction_norm": 6.119564056396484,
    "timestamp": "2025-07-14T12:33:44.655271"
  },
  {
    "epoch": 316,
    "batch_idx": 64,
    "loss": 0.22668859362602234,
    "learning_rate": 0.0009396581550952781,
    "timestep_mean": 528.0,
    "epsilon_norm": 7.405768871307373,
    "prediction_norm": 6.4441704750061035,
    "timestamp": "2025-07-14T12:33:44.660166"
  },
  {
    "epoch": 316,
    "batch_idx": 96,
    "loss": 0.19553062319755554,
    "learning_rate": 0.0009396581550952781,
    "timestep_mean": 482.25,
    "epsilon_norm": 7.0808024406433105,
    "prediction_norm": 6.171265125274658,
    "timestamp": "2025-07-14T12:33:44.665030"
  },
  {
    "epoch": 316,
    "batch_idx": 128,
    "loss": 0.3195009231567383,
    "learning_rate": 0.0009396581550952781,
    "timestep_mean": 410.53125,
    "epsilon_norm": 7.511287212371826,
    "prediction_norm": 5.951711654663086,
    "timestamp": "2025-07-14T12:33:44.670063"
  },
  {
    "epoch": 316,
    "batch_idx": 160,
    "loss": 0.14498376846313477,
    "learning_rate": 0.0009396581550952781,
    "timestep_mean": 553.65625,
    "epsilon_norm": 6.562129497528076,
    "prediction_norm": 6.411725044250488,
    "timestamp": "2025-07-14T12:33:44.675049"
  },
  {
    "epoch": 316,
    "batch_idx": 192,
    "loss": 0.2804133892059326,
    "learning_rate": 0.0009396581550952781,
    "timestep_mean": 579.875,
    "epsilon_norm": 3.6217336654663086,
    "prediction_norm": 2.9244191646575928,
    "timestamp": "2025-07-14T12:33:44.679948"
  },
  {
    "epoch": 317,
    "batch_idx": 0,
    "loss": 0.3189537525177002,
    "learning_rate": 0.0009392835762321978,
    "timestep_mean": 407.65625,
    "epsilon_norm": 7.783290863037109,
    "prediction_norm": 6.011063098907471,
    "timestamp": "2025-07-14T12:33:44.684903"
  },
  {
    "epoch": 317,
    "batch_idx": 32,
    "loss": 0.3780250549316406,
    "learning_rate": 0.0009392835762321978,
    "timestep_mean": 536.84375,
    "epsilon_norm": 8.237127304077148,
    "prediction_norm": 6.838249683380127,
    "timestamp": "2025-07-14T12:33:44.689816"
  },
  {
    "epoch": 317,
    "batch_idx": 64,
    "loss": 0.24682527780532837,
    "learning_rate": 0.0009392835762321978,
    "timestep_mean": 529.5,
    "epsilon_norm": 7.755156517028809,
    "prediction_norm": 7.1732497215271,
    "timestamp": "2025-07-14T12:33:44.694811"
  },
  {
    "epoch": 317,
    "batch_idx": 96,
    "loss": 0.33602631092071533,
    "learning_rate": 0.0009392835762321978,
    "timestep_mean": 525.28125,
    "epsilon_norm": 8.147594451904297,
    "prediction_norm": 7.386172294616699,
    "timestamp": "2025-07-14T12:33:44.699867"
  },
  {
    "epoch": 317,
    "batch_idx": 128,
    "loss": 0.46860671043395996,
    "learning_rate": 0.0009392835762321978,
    "timestep_mean": 504.84375,
    "epsilon_norm": 8.305645942687988,
    "prediction_norm": 6.7826666831970215,
    "timestamp": "2025-07-14T12:33:44.704742"
  },
  {
    "epoch": 317,
    "batch_idx": 160,
    "loss": 0.45903050899505615,
    "learning_rate": 0.0009392835762321978,
    "timestep_mean": 500.3125,
    "epsilon_norm": 8.364349365234375,
    "prediction_norm": 6.4628586769104,
    "timestamp": "2025-07-14T12:33:44.709753"
  },
  {
    "epoch": 317,
    "batch_idx": 192,
    "loss": 0.24519428610801697,
    "learning_rate": 0.0009392835762321978,
    "timestep_mean": 526.625,
    "epsilon_norm": 3.8380446434020996,
    "prediction_norm": 3.4377710819244385,
    "timestamp": "2025-07-14T12:33:44.717246"
  },
  {
    "epoch": 318,
    "batch_idx": 0,
    "loss": 0.2704229950904846,
    "learning_rate": 0.0009389079134805609,
    "timestep_mean": 506.125,
    "epsilon_norm": 7.69659423828125,
    "prediction_norm": 6.221818923950195,
    "timestamp": "2025-07-14T12:33:44.722337"
  },
  {
    "epoch": 318,
    "batch_idx": 32,
    "loss": 0.3293595016002655,
    "learning_rate": 0.0009389079134805609,
    "timestep_mean": 459.8125,
    "epsilon_norm": 7.88912296295166,
    "prediction_norm": 6.249311447143555,
    "timestamp": "2025-07-14T12:33:44.727307"
  },
  {
    "epoch": 318,
    "batch_idx": 64,
    "loss": 0.27395838499069214,
    "learning_rate": 0.0009389079134805609,
    "timestep_mean": 518.5,
    "epsilon_norm": 7.80129337310791,
    "prediction_norm": 6.5641398429870605,
    "timestamp": "2025-07-14T12:33:44.732296"
  },
  {
    "epoch": 318,
    "batch_idx": 96,
    "loss": 0.2904775142669678,
    "learning_rate": 0.0009389079134805609,
    "timestep_mean": 496.09375,
    "epsilon_norm": 8.332097053527832,
    "prediction_norm": 6.840487003326416,
    "timestamp": "2025-07-14T12:33:44.737341"
  },
  {
    "epoch": 318,
    "batch_idx": 128,
    "loss": 0.5092045068740845,
    "learning_rate": 0.0009389079134805609,
    "timestep_mean": 546.5625,
    "epsilon_norm": 8.40394401550293,
    "prediction_norm": 5.809972286224365,
    "timestamp": "2025-07-14T12:33:44.742252"
  },
  {
    "epoch": 318,
    "batch_idx": 160,
    "loss": 0.2712084650993347,
    "learning_rate": 0.0009389079134805609,
    "timestep_mean": 484.75,
    "epsilon_norm": 7.2740936279296875,
    "prediction_norm": 5.8610334396362305,
    "timestamp": "2025-07-14T12:33:44.747258"
  },
  {
    "epoch": 318,
    "batch_idx": 192,
    "loss": 0.515049397945404,
    "learning_rate": 0.0009389079134805609,
    "timestep_mean": 523.875,
    "epsilon_norm": 4.878202438354492,
    "prediction_norm": 3.6683738231658936,
    "timestamp": "2025-07-14T12:33:44.752233"
  },
  {
    "epoch": 319,
    "batch_idx": 0,
    "loss": 0.4654548764228821,
    "learning_rate": 0.0009385311677672781,
    "timestep_mean": 445.1875,
    "epsilon_norm": 8.434076309204102,
    "prediction_norm": 6.120800018310547,
    "timestamp": "2025-07-14T12:33:44.757163"
  },
  {
    "epoch": 319,
    "batch_idx": 32,
    "loss": 0.21010226011276245,
    "learning_rate": 0.0009385311677672781,
    "timestep_mean": 587.53125,
    "epsilon_norm": 7.314823627471924,
    "prediction_norm": 5.839356899261475,
    "timestamp": "2025-07-14T12:33:44.762132"
  },
  {
    "epoch": 319,
    "batch_idx": 64,
    "loss": 0.48756125569343567,
    "learning_rate": 0.0009385311677672781,
    "timestep_mean": 474.3125,
    "epsilon_norm": 9.255435943603516,
    "prediction_norm": 7.174157619476318,
    "timestamp": "2025-07-14T12:33:44.767162"
  },
  {
    "epoch": 319,
    "batch_idx": 96,
    "loss": 0.31470444798469543,
    "learning_rate": 0.0009385311677672781,
    "timestep_mean": 470.5,
    "epsilon_norm": 7.865230560302734,
    "prediction_norm": 6.908642768859863,
    "timestamp": "2025-07-14T12:33:44.772176"
  },
  {
    "epoch": 319,
    "batch_idx": 128,
    "loss": 0.4242164194583893,
    "learning_rate": 0.0009385311677672781,
    "timestep_mean": 454.625,
    "epsilon_norm": 7.622800350189209,
    "prediction_norm": 6.0279340744018555,
    "timestamp": "2025-07-14T12:33:44.777110"
  },
  {
    "epoch": 319,
    "batch_idx": 160,
    "loss": 0.22906316816806793,
    "learning_rate": 0.0009385311677672781,
    "timestep_mean": 565.09375,
    "epsilon_norm": 8.711198806762695,
    "prediction_norm": 8.012642860412598,
    "timestamp": "2025-07-14T12:33:44.782078"
  },
  {
    "epoch": 319,
    "batch_idx": 192,
    "loss": 0.5367709398269653,
    "learning_rate": 0.0009385311677672781,
    "timestep_mean": 543.375,
    "epsilon_norm": 4.030792713165283,
    "prediction_norm": 2.901562213897705,
    "timestamp": "2025-07-14T12:33:44.787064"
  },
  {
    "epoch": 320,
    "batch_idx": 0,
    "loss": 0.8227490186691284,
    "learning_rate": 0.0009381533400219318,
    "timestep_mean": 535.6875,
    "epsilon_norm": 9.56025218963623,
    "prediction_norm": 6.024362087249756,
    "timestamp": "2025-07-14T12:33:44.792158"
  },
  {
    "epoch": 320,
    "batch_idx": 32,
    "loss": 0.2491491436958313,
    "learning_rate": 0.0009381533400219318,
    "timestep_mean": 483.96875,
    "epsilon_norm": 8.267199516296387,
    "prediction_norm": 6.766894817352295,
    "timestamp": "2025-07-14T12:33:44.797017"
  },
  {
    "epoch": 320,
    "batch_idx": 64,
    "loss": 0.3062819540500641,
    "learning_rate": 0.0009381533400219318,
    "timestep_mean": 475.65625,
    "epsilon_norm": 7.428168773651123,
    "prediction_norm": 6.650835990905762,
    "timestamp": "2025-07-14T12:33:44.802096"
  },
  {
    "epoch": 320,
    "batch_idx": 96,
    "loss": 0.31490784883499146,
    "learning_rate": 0.0009381533400219318,
    "timestep_mean": 561.34375,
    "epsilon_norm": 8.198983192443848,
    "prediction_norm": 7.443453788757324,
    "timestamp": "2025-07-14T12:33:44.807024"
  },
  {
    "epoch": 320,
    "batch_idx": 128,
    "loss": 0.40128445625305176,
    "learning_rate": 0.0009381533400219318,
    "timestep_mean": 500.84375,
    "epsilon_norm": 8.477516174316406,
    "prediction_norm": 7.4607768058776855,
    "timestamp": "2025-07-14T12:33:44.811962"
  },
  {
    "epoch": 320,
    "batch_idx": 160,
    "loss": 0.5919876098632812,
    "learning_rate": 0.0009381533400219318,
    "timestep_mean": 446.96875,
    "epsilon_norm": 8.183693885803223,
    "prediction_norm": 7.088517189025879,
    "timestamp": "2025-07-14T12:33:44.817174"
  },
  {
    "epoch": 320,
    "batch_idx": 192,
    "loss": 0.5183454751968384,
    "learning_rate": 0.0009381533400219318,
    "timestep_mean": 243.125,
    "epsilon_norm": 3.4731099605560303,
    "prediction_norm": 1.822806715965271,
    "timestamp": "2025-07-14T12:33:44.822176"
  },
  {
    "epoch": 321,
    "batch_idx": 0,
    "loss": 0.3299107253551483,
    "learning_rate": 0.0009377744311767747,
    "timestep_mean": 498.28125,
    "epsilon_norm": 8.385936737060547,
    "prediction_norm": 7.166971206665039,
    "timestamp": "2025-07-14T12:33:44.827309"
  },
  {
    "epoch": 321,
    "batch_idx": 32,
    "loss": 0.29740551114082336,
    "learning_rate": 0.0009377744311767747,
    "timestep_mean": 523.375,
    "epsilon_norm": 8.100316047668457,
    "prediction_norm": 7.086917877197266,
    "timestamp": "2025-07-14T12:33:44.832266"
  },
  {
    "epoch": 321,
    "batch_idx": 64,
    "loss": 0.36451637744903564,
    "learning_rate": 0.0009377744311767747,
    "timestep_mean": 470.71875,
    "epsilon_norm": 8.69484806060791,
    "prediction_norm": 6.337137222290039,
    "timestamp": "2025-07-14T12:33:44.837316"
  },
  {
    "epoch": 321,
    "batch_idx": 96,
    "loss": 0.22015099227428436,
    "learning_rate": 0.0009377744311767747,
    "timestep_mean": 476.0625,
    "epsilon_norm": 7.671747207641602,
    "prediction_norm": 6.232197284698486,
    "timestamp": "2025-07-14T12:33:44.842487"
  },
  {
    "epoch": 321,
    "batch_idx": 128,
    "loss": 0.2965056300163269,
    "learning_rate": 0.0009377744311767747,
    "timestep_mean": 464.84375,
    "epsilon_norm": 6.527348041534424,
    "prediction_norm": 5.180583477020264,
    "timestamp": "2025-07-14T12:33:44.847569"
  },
  {
    "epoch": 321,
    "batch_idx": 160,
    "loss": 0.4419811964035034,
    "learning_rate": 0.0009377744311767747,
    "timestep_mean": 539.625,
    "epsilon_norm": 8.47243881225586,
    "prediction_norm": 6.896434783935547,
    "timestamp": "2025-07-14T12:33:44.852690"
  },
  {
    "epoch": 321,
    "batch_idx": 192,
    "loss": 0.15465518832206726,
    "learning_rate": 0.0009377744311767747,
    "timestep_mean": 289.625,
    "epsilon_norm": 3.219763994216919,
    "prediction_norm": 2.421147108078003,
    "timestamp": "2025-07-14T12:33:44.857723"
  },
  {
    "epoch": 322,
    "batch_idx": 0,
    "loss": 0.3017786145210266,
    "learning_rate": 0.0009373944421667264,
    "timestep_mean": 489.9375,
    "epsilon_norm": 7.486268997192383,
    "prediction_norm": 6.138865947723389,
    "timestamp": "2025-07-14T12:33:44.863046"
  },
  {
    "epoch": 322,
    "batch_idx": 32,
    "loss": 0.29932236671447754,
    "learning_rate": 0.0009373944421667264,
    "timestep_mean": 502.125,
    "epsilon_norm": 7.397151947021484,
    "prediction_norm": 6.156217575073242,
    "timestamp": "2025-07-14T12:33:44.868194"
  },
  {
    "epoch": 322,
    "batch_idx": 64,
    "loss": 0.2925782799720764,
    "learning_rate": 0.0009373944421667264,
    "timestep_mean": 521.53125,
    "epsilon_norm": 7.543788433074951,
    "prediction_norm": 5.703526020050049,
    "timestamp": "2025-07-14T12:33:44.873401"
  },
  {
    "epoch": 322,
    "batch_idx": 96,
    "loss": 0.22304826974868774,
    "learning_rate": 0.0009373944421667264,
    "timestep_mean": 531.375,
    "epsilon_norm": 8.485624313354492,
    "prediction_norm": 7.182766914367676,
    "timestamp": "2025-07-14T12:33:44.878566"
  },
  {
    "epoch": 322,
    "batch_idx": 128,
    "loss": 0.36766546964645386,
    "learning_rate": 0.0009373944421667264,
    "timestep_mean": 473.40625,
    "epsilon_norm": 7.330257415771484,
    "prediction_norm": 5.52805233001709,
    "timestamp": "2025-07-14T12:33:44.883709"
  },
  {
    "epoch": 322,
    "batch_idx": 160,
    "loss": 0.24116432666778564,
    "learning_rate": 0.0009373944421667264,
    "timestep_mean": 513.15625,
    "epsilon_norm": 7.39544153213501,
    "prediction_norm": 6.344460964202881,
    "timestamp": "2025-07-14T12:33:44.888916"
  },
  {
    "epoch": 322,
    "batch_idx": 192,
    "loss": 0.5438330769538879,
    "learning_rate": 0.0009373944421667264,
    "timestep_mean": 495.0,
    "epsilon_norm": 4.346377849578857,
    "prediction_norm": 3.1725988388061523,
    "timestamp": "2025-07-14T12:33:44.894075"
  },
  {
    "epoch": 323,
    "batch_idx": 0,
    "loss": 0.25606316328048706,
    "learning_rate": 0.0009370133739293723,
    "timestep_mean": 602.5625,
    "epsilon_norm": 7.402236461639404,
    "prediction_norm": 6.11984395980835,
    "timestamp": "2025-07-14T12:33:44.899184"
  },
  {
    "epoch": 323,
    "batch_idx": 32,
    "loss": 0.4299777150154114,
    "learning_rate": 0.0009370133739293723,
    "timestep_mean": 525.34375,
    "epsilon_norm": 9.138660430908203,
    "prediction_norm": 7.113570690155029,
    "timestamp": "2025-07-14T12:33:44.904439"
  },
  {
    "epoch": 323,
    "batch_idx": 64,
    "loss": 0.4576312005519867,
    "learning_rate": 0.0009370133739293723,
    "timestep_mean": 396.84375,
    "epsilon_norm": 7.833112716674805,
    "prediction_norm": 5.6335320472717285,
    "timestamp": "2025-07-14T12:33:44.909612"
  },
  {
    "epoch": 323,
    "batch_idx": 96,
    "loss": 0.34681060910224915,
    "learning_rate": 0.0009370133739293723,
    "timestep_mean": 447.15625,
    "epsilon_norm": 7.265563488006592,
    "prediction_norm": 5.2406229972839355,
    "timestamp": "2025-07-14T12:33:44.914700"
  },
  {
    "epoch": 323,
    "batch_idx": 128,
    "loss": 0.2951222062110901,
    "learning_rate": 0.0009370133739293723,
    "timestep_mean": 577.0625,
    "epsilon_norm": 8.617317199707031,
    "prediction_norm": 7.299414157867432,
    "timestamp": "2025-07-14T12:33:44.919931"
  },
  {
    "epoch": 323,
    "batch_idx": 160,
    "loss": 0.428803026676178,
    "learning_rate": 0.0009370133739293723,
    "timestep_mean": 412.90625,
    "epsilon_norm": 7.198616981506348,
    "prediction_norm": 5.537113666534424,
    "timestamp": "2025-07-14T12:33:44.925106"
  },
  {
    "epoch": 323,
    "batch_idx": 192,
    "loss": 0.32293713092803955,
    "learning_rate": 0.0009370133739293723,
    "timestep_mean": 592.0,
    "epsilon_norm": 3.6128079891204834,
    "prediction_norm": 2.86466121673584,
    "timestamp": "2025-07-14T12:33:44.930190"
  },
  {
    "epoch": 324,
    "batch_idx": 0,
    "loss": 0.29382482171058655,
    "learning_rate": 0.0009366312274049601,
    "timestep_mean": 635.59375,
    "epsilon_norm": 8.470697402954102,
    "prediction_norm": 7.364697456359863,
    "timestamp": "2025-07-14T12:33:44.935322"
  },
  {
    "epoch": 324,
    "batch_idx": 32,
    "loss": 0.35874465107917786,
    "learning_rate": 0.0009366312274049601,
    "timestep_mean": 577.0,
    "epsilon_norm": 8.033785820007324,
    "prediction_norm": 6.9351725578308105,
    "timestamp": "2025-07-14T12:33:44.940472"
  },
  {
    "epoch": 324,
    "batch_idx": 64,
    "loss": 0.39008021354675293,
    "learning_rate": 0.0009366312274049601,
    "timestep_mean": 494.28125,
    "epsilon_norm": 6.845444202423096,
    "prediction_norm": 6.110713481903076,
    "timestamp": "2025-07-14T12:33:44.945634"
  },
  {
    "epoch": 324,
    "batch_idx": 96,
    "loss": 0.354007750749588,
    "learning_rate": 0.0009366312274049601,
    "timestep_mean": 397.96875,
    "epsilon_norm": 7.601259231567383,
    "prediction_norm": 6.172926902770996,
    "timestamp": "2025-07-14T12:33:44.950828"
  },
  {
    "epoch": 324,
    "batch_idx": 128,
    "loss": 0.26298773288726807,
    "learning_rate": 0.0009366312274049601,
    "timestep_mean": 538.4375,
    "epsilon_norm": 7.784610748291016,
    "prediction_norm": 6.962967872619629,
    "timestamp": "2025-07-14T12:33:44.955946"
  },
  {
    "epoch": 324,
    "batch_idx": 160,
    "loss": 0.31503671407699585,
    "learning_rate": 0.0009366312274049601,
    "timestep_mean": 529.03125,
    "epsilon_norm": 8.252153396606445,
    "prediction_norm": 6.842968940734863,
    "timestamp": "2025-07-14T12:33:44.961146"
  },
  {
    "epoch": 324,
    "batch_idx": 192,
    "loss": 0.07768993079662323,
    "learning_rate": 0.0009366312274049601,
    "timestep_mean": 643.25,
    "epsilon_norm": 3.22808575630188,
    "prediction_norm": 2.9750328063964844,
    "timestamp": "2025-07-14T12:33:44.966318"
  },
  {
    "epoch": 325,
    "batch_idx": 0,
    "loss": 0.2284817099571228,
    "learning_rate": 0.0009362480035363987,
    "timestep_mean": 457.90625,
    "epsilon_norm": 7.783531188964844,
    "prediction_norm": 6.713696002960205,
    "timestamp": "2025-07-14T12:33:44.971593"
  },
  {
    "epoch": 325,
    "batch_idx": 32,
    "loss": 0.19613555073738098,
    "learning_rate": 0.0009362480035363987,
    "timestep_mean": 535.96875,
    "epsilon_norm": 6.8160624504089355,
    "prediction_norm": 5.798940181732178,
    "timestamp": "2025-07-14T12:33:44.976605"
  },
  {
    "epoch": 325,
    "batch_idx": 64,
    "loss": 0.3432314991950989,
    "learning_rate": 0.0009362480035363987,
    "timestep_mean": 478.4375,
    "epsilon_norm": 6.898754119873047,
    "prediction_norm": 5.151538372039795,
    "timestamp": "2025-07-14T12:33:44.981733"
  },
  {
    "epoch": 325,
    "batch_idx": 96,
    "loss": 0.22392863035202026,
    "learning_rate": 0.0009362480035363987,
    "timestep_mean": 442.0625,
    "epsilon_norm": 6.53169059753418,
    "prediction_norm": 5.4844651222229,
    "timestamp": "2025-07-14T12:33:44.986930"
  },
  {
    "epoch": 325,
    "batch_idx": 128,
    "loss": 0.49857303500175476,
    "learning_rate": 0.0009362480035363987,
    "timestep_mean": 441.53125,
    "epsilon_norm": 8.409029006958008,
    "prediction_norm": 5.86059045791626,
    "timestamp": "2025-07-14T12:33:44.992048"
  },
  {
    "epoch": 325,
    "batch_idx": 160,
    "loss": 0.3299362063407898,
    "learning_rate": 0.0009362480035363987,
    "timestep_mean": 559.09375,
    "epsilon_norm": 7.684901237487793,
    "prediction_norm": 6.6695876121521,
    "timestamp": "2025-07-14T12:33:44.997273"
  },
  {
    "epoch": 325,
    "batch_idx": 192,
    "loss": 0.2733255624771118,
    "learning_rate": 0.0009362480035363987,
    "timestep_mean": 471.0,
    "epsilon_norm": 4.228118419647217,
    "prediction_norm": 3.485860586166382,
    "timestamp": "2025-07-14T12:33:45.002427"
  },
  {
    "epoch": 326,
    "batch_idx": 0,
    "loss": 0.38038575649261475,
    "learning_rate": 0.0009358637032692546,
    "timestep_mean": 441.71875,
    "epsilon_norm": 7.793579578399658,
    "prediction_norm": 5.4501566886901855,
    "timestamp": "2025-07-14T12:33:45.007420"
  },
  {
    "epoch": 326,
    "batch_idx": 32,
    "loss": 0.3603929281234741,
    "learning_rate": 0.0009358637032692546,
    "timestep_mean": 501.0625,
    "epsilon_norm": 7.078699111938477,
    "prediction_norm": 5.03217887878418,
    "timestamp": "2025-07-14T12:33:45.012646"
  },
  {
    "epoch": 326,
    "batch_idx": 64,
    "loss": 0.4232179522514343,
    "learning_rate": 0.0009358637032692546,
    "timestep_mean": 473.65625,
    "epsilon_norm": 8.55245304107666,
    "prediction_norm": 6.101275444030762,
    "timestamp": "2025-07-14T12:33:45.017741"
  },
  {
    "epoch": 326,
    "batch_idx": 96,
    "loss": 0.16433285176753998,
    "learning_rate": 0.0009358637032692546,
    "timestep_mean": 570.15625,
    "epsilon_norm": 6.699319839477539,
    "prediction_norm": 5.495124340057373,
    "timestamp": "2025-07-14T12:33:45.022814"
  },
  {
    "epoch": 326,
    "batch_idx": 128,
    "loss": 0.3544028401374817,
    "learning_rate": 0.0009358637032692546,
    "timestep_mean": 489.46875,
    "epsilon_norm": 8.98823356628418,
    "prediction_norm": 6.9446306228637695,
    "timestamp": "2025-07-14T12:33:45.028066"
  },
  {
    "epoch": 326,
    "batch_idx": 160,
    "loss": 0.3953145146369934,
    "learning_rate": 0.0009358637032692546,
    "timestep_mean": 451.625,
    "epsilon_norm": 7.402812480926514,
    "prediction_norm": 5.831839084625244,
    "timestamp": "2025-07-14T12:33:45.033384"
  },
  {
    "epoch": 326,
    "batch_idx": 192,
    "loss": 0.5893805623054504,
    "learning_rate": 0.0009358637032692546,
    "timestep_mean": 477.375,
    "epsilon_norm": 4.471737384796143,
    "prediction_norm": 2.7110462188720703,
    "timestamp": "2025-07-14T12:33:45.038406"
  },
  {
    "epoch": 327,
    "batch_idx": 0,
    "loss": 0.5166831016540527,
    "learning_rate": 0.0009354783275517505,
    "timestep_mean": 456.90625,
    "epsilon_norm": 8.183140754699707,
    "prediction_norm": 5.876658916473389,
    "timestamp": "2025-07-14T12:33:45.043640"
  },
  {
    "epoch": 327,
    "batch_idx": 32,
    "loss": 0.3401091694831848,
    "learning_rate": 0.0009354783275517505,
    "timestep_mean": 472.46875,
    "epsilon_norm": 7.135385990142822,
    "prediction_norm": 6.388131141662598,
    "timestamp": "2025-07-14T12:33:45.048794"
  },
  {
    "epoch": 327,
    "batch_idx": 64,
    "loss": 0.3456166684627533,
    "learning_rate": 0.0009354783275517505,
    "timestep_mean": 405.34375,
    "epsilon_norm": 6.560975551605225,
    "prediction_norm": 5.769069671630859,
    "timestamp": "2025-07-14T12:33:45.053950"
  },
  {
    "epoch": 327,
    "batch_idx": 96,
    "loss": 0.49359130859375,
    "learning_rate": 0.0009354783275517505,
    "timestep_mean": 440.4375,
    "epsilon_norm": 7.9947075843811035,
    "prediction_norm": 6.085263252258301,
    "timestamp": "2025-07-14T12:33:45.059111"
  },
  {
    "epoch": 327,
    "batch_idx": 128,
    "loss": 0.2875552177429199,
    "learning_rate": 0.0009354783275517505,
    "timestep_mean": 521.03125,
    "epsilon_norm": 10.158442497253418,
    "prediction_norm": 9.415653228759766,
    "timestamp": "2025-07-14T12:33:45.064238"
  },
  {
    "epoch": 327,
    "batch_idx": 160,
    "loss": 0.3254936933517456,
    "learning_rate": 0.0009354783275517505,
    "timestep_mean": 460.84375,
    "epsilon_norm": 8.027608871459961,
    "prediction_norm": 6.615546703338623,
    "timestamp": "2025-07-14T12:33:45.069375"
  },
  {
    "epoch": 327,
    "batch_idx": 192,
    "loss": 0.15696479380130768,
    "learning_rate": 0.0009354783275517505,
    "timestep_mean": 662.5,
    "epsilon_norm": 3.8459677696228027,
    "prediction_norm": 3.2490482330322266,
    "timestamp": "2025-07-14T12:33:45.074527"
  },
  {
    "epoch": 328,
    "batch_idx": 0,
    "loss": 0.1966431736946106,
    "learning_rate": 0.0009350918773347629,
    "timestep_mean": 620.34375,
    "epsilon_norm": 7.750726222991943,
    "prediction_norm": 7.0187883377075195,
    "timestamp": "2025-07-14T12:33:45.079905"
  },
  {
    "epoch": 328,
    "batch_idx": 32,
    "loss": 0.3299223780632019,
    "learning_rate": 0.0009350918773347629,
    "timestep_mean": 504.3125,
    "epsilon_norm": 8.029668807983398,
    "prediction_norm": 6.4541826248168945,
    "timestamp": "2025-07-14T12:33:45.085048"
  },
  {
    "epoch": 328,
    "batch_idx": 64,
    "loss": 0.3734380006790161,
    "learning_rate": 0.0009350918773347629,
    "timestep_mean": 518.25,
    "epsilon_norm": 9.309460639953613,
    "prediction_norm": 8.02552604675293,
    "timestamp": "2025-07-14T12:33:45.090307"
  },
  {
    "epoch": 328,
    "batch_idx": 96,
    "loss": 0.3837401866912842,
    "learning_rate": 0.0009350918773347629,
    "timestep_mean": 385.03125,
    "epsilon_norm": 7.407751083374023,
    "prediction_norm": 5.3663859367370605,
    "timestamp": "2025-07-14T12:33:45.095526"
  },
  {
    "epoch": 328,
    "batch_idx": 128,
    "loss": 0.23887158930301666,
    "learning_rate": 0.0009350918773347629,
    "timestep_mean": 510.0,
    "epsilon_norm": 7.233211517333984,
    "prediction_norm": 6.005202770233154,
    "timestamp": "2025-07-14T12:33:45.100692"
  },
  {
    "epoch": 328,
    "batch_idx": 160,
    "loss": 0.3782515525817871,
    "learning_rate": 0.0009350918773347629,
    "timestep_mean": 490.4375,
    "epsilon_norm": 8.524199485778809,
    "prediction_norm": 6.542079448699951,
    "timestamp": "2025-07-14T12:33:45.105944"
  },
  {
    "epoch": 328,
    "batch_idx": 192,
    "loss": 0.1459328979253769,
    "learning_rate": 0.0009350918773347629,
    "timestep_mean": 602.75,
    "epsilon_norm": 3.997588872909546,
    "prediction_norm": 3.426022529602051,
    "timestamp": "2025-07-14T12:33:45.111098"
  },
  {
    "epoch": 329,
    "batch_idx": 0,
    "loss": 0.3120006322860718,
    "learning_rate": 0.0009347043535718191,
    "timestep_mean": 435.3125,
    "epsilon_norm": 7.991665840148926,
    "prediction_norm": 6.65086030960083,
    "timestamp": "2025-07-14T12:33:45.116253"
  },
  {
    "epoch": 329,
    "batch_idx": 32,
    "loss": 0.2544616162776947,
    "learning_rate": 0.0009347043535718191,
    "timestep_mean": 509.25,
    "epsilon_norm": 8.550199508666992,
    "prediction_norm": 7.053972244262695,
    "timestamp": "2025-07-14T12:33:45.121569"
  },
  {
    "epoch": 329,
    "batch_idx": 64,
    "loss": 0.2957930564880371,
    "learning_rate": 0.0009347043535718191,
    "timestep_mean": 609.4375,
    "epsilon_norm": 8.245162010192871,
    "prediction_norm": 7.40421724319458,
    "timestamp": "2025-07-14T12:33:45.126749"
  },
  {
    "epoch": 329,
    "batch_idx": 96,
    "loss": 0.5791637897491455,
    "learning_rate": 0.0009347043535718191,
    "timestep_mean": 490.0,
    "epsilon_norm": 9.125340461730957,
    "prediction_norm": 6.586138725280762,
    "timestamp": "2025-07-14T12:33:45.131682"
  },
  {
    "epoch": 329,
    "batch_idx": 128,
    "loss": 0.3207286596298218,
    "learning_rate": 0.0009347043535718191,
    "timestep_mean": 543.125,
    "epsilon_norm": 7.587430477142334,
    "prediction_norm": 6.066394805908203,
    "timestamp": "2025-07-14T12:33:45.136808"
  },
  {
    "epoch": 329,
    "batch_idx": 160,
    "loss": 0.40116241574287415,
    "learning_rate": 0.0009347043535718191,
    "timestep_mean": 518.875,
    "epsilon_norm": 9.24863338470459,
    "prediction_norm": 8.621554374694824,
    "timestamp": "2025-07-14T12:33:45.141990"
  },
  {
    "epoch": 329,
    "batch_idx": 192,
    "loss": 0.2815100848674774,
    "learning_rate": 0.0009347043535718191,
    "timestep_mean": 657.0,
    "epsilon_norm": 4.134084701538086,
    "prediction_norm": 3.4250497817993164,
    "timestamp": "2025-07-14T12:33:45.147027"
  },
  {
    "epoch": 330,
    "batch_idx": 0,
    "loss": 0.48544278740882874,
    "learning_rate": 0.0009343157572190957,
    "timestep_mean": 395.6875,
    "epsilon_norm": 8.45034408569336,
    "prediction_norm": 6.0641093254089355,
    "timestamp": "2025-07-14T12:33:45.152150"
  },
  {
    "epoch": 330,
    "batch_idx": 32,
    "loss": 0.38654690980911255,
    "learning_rate": 0.0009343157572190957,
    "timestep_mean": 513.5625,
    "epsilon_norm": 7.507251262664795,
    "prediction_norm": 6.569231986999512,
    "timestamp": "2025-07-14T12:33:45.157351"
  },
  {
    "epoch": 330,
    "batch_idx": 64,
    "loss": 0.6290712356567383,
    "learning_rate": 0.0009343157572190957,
    "timestep_mean": 474.6875,
    "epsilon_norm": 8.438064575195312,
    "prediction_norm": 6.329038619995117,
    "timestamp": "2025-07-14T12:33:45.162487"
  },
  {
    "epoch": 330,
    "batch_idx": 96,
    "loss": 0.18355940282344818,
    "learning_rate": 0.0009343157572190957,
    "timestep_mean": 497.71875,
    "epsilon_norm": 7.224197864532471,
    "prediction_norm": 6.876110553741455,
    "timestamp": "2025-07-14T12:33:45.167641"
  },
  {
    "epoch": 330,
    "batch_idx": 128,
    "loss": 0.3738311529159546,
    "learning_rate": 0.0009343157572190957,
    "timestep_mean": 553.25,
    "epsilon_norm": 9.733384132385254,
    "prediction_norm": 8.691534996032715,
    "timestamp": "2025-07-14T12:33:45.172820"
  },
  {
    "epoch": 330,
    "batch_idx": 160,
    "loss": 0.2050565630197525,
    "learning_rate": 0.0009343157572190957,
    "timestep_mean": 473.375,
    "epsilon_norm": 7.505713939666748,
    "prediction_norm": 6.187203884124756,
    "timestamp": "2025-07-14T12:33:45.177968"
  },
  {
    "epoch": 330,
    "batch_idx": 192,
    "loss": 0.13285614550113678,
    "learning_rate": 0.0009343157572190957,
    "timestep_mean": 607.375,
    "epsilon_norm": 4.167892932891846,
    "prediction_norm": 4.432363510131836,
    "timestamp": "2025-07-14T12:33:45.183201"
  },
  {
    "epoch": 331,
    "batch_idx": 0,
    "loss": 0.19786173105239868,
    "learning_rate": 0.0009339260892354154,
    "timestep_mean": 583.46875,
    "epsilon_norm": 8.894761085510254,
    "prediction_norm": 8.143902778625488,
    "timestamp": "2025-07-14T12:33:45.188433"
  },
  {
    "epoch": 331,
    "batch_idx": 32,
    "loss": 0.19913974404335022,
    "learning_rate": 0.0009339260892354154,
    "timestep_mean": 495.46875,
    "epsilon_norm": 7.671364784240723,
    "prediction_norm": 7.119377136230469,
    "timestamp": "2025-07-14T12:33:45.193519"
  },
  {
    "epoch": 331,
    "batch_idx": 64,
    "loss": 0.44832247495651245,
    "learning_rate": 0.0009339260892354154,
    "timestep_mean": 479.34375,
    "epsilon_norm": 8.969779968261719,
    "prediction_norm": 7.234184265136719,
    "timestamp": "2025-07-14T12:33:45.198720"
  },
  {
    "epoch": 331,
    "batch_idx": 96,
    "loss": 0.3755531907081604,
    "learning_rate": 0.0009339260892354154,
    "timestep_mean": 526.0,
    "epsilon_norm": 8.670573234558105,
    "prediction_norm": 6.914809703826904,
    "timestamp": "2025-07-14T12:33:45.203959"
  },
  {
    "epoch": 331,
    "batch_idx": 128,
    "loss": 0.285478800535202,
    "learning_rate": 0.0009339260892354154,
    "timestep_mean": 471.28125,
    "epsilon_norm": 8.094794273376465,
    "prediction_norm": 7.122848033905029,
    "timestamp": "2025-07-14T12:33:45.209126"
  },
  {
    "epoch": 331,
    "batch_idx": 160,
    "loss": 0.18771538138389587,
    "learning_rate": 0.0009339260892354154,
    "timestep_mean": 613.1875,
    "epsilon_norm": 7.537577152252197,
    "prediction_norm": 6.746129512786865,
    "timestamp": "2025-07-14T12:33:45.214376"
  },
  {
    "epoch": 331,
    "batch_idx": 192,
    "loss": 0.20331613719463348,
    "learning_rate": 0.0009339260892354154,
    "timestep_mean": 383.375,
    "epsilon_norm": 3.450618028640747,
    "prediction_norm": 3.4581167697906494,
    "timestamp": "2025-07-14T12:33:45.219544"
  },
  {
    "epoch": 332,
    "batch_idx": 0,
    "loss": 0.3067319393157959,
    "learning_rate": 0.0009335353505822451,
    "timestep_mean": 504.6875,
    "epsilon_norm": 7.55999231338501,
    "prediction_norm": 5.774165630340576,
    "timestamp": "2025-07-14T12:33:45.224808"
  },
  {
    "epoch": 332,
    "batch_idx": 32,
    "loss": 0.2084396481513977,
    "learning_rate": 0.0009335353505822451,
    "timestep_mean": 540.375,
    "epsilon_norm": 7.289248466491699,
    "prediction_norm": 6.296610355377197,
    "timestamp": "2025-07-14T12:33:45.230010"
  },
  {
    "epoch": 332,
    "batch_idx": 64,
    "loss": 0.3893466889858246,
    "learning_rate": 0.0009335353505822451,
    "timestep_mean": 410.59375,
    "epsilon_norm": 8.319255828857422,
    "prediction_norm": 6.79318904876709,
    "timestamp": "2025-07-14T12:33:45.235206"
  },
  {
    "epoch": 332,
    "batch_idx": 96,
    "loss": 0.38753873109817505,
    "learning_rate": 0.0009335353505822451,
    "timestep_mean": 420.5,
    "epsilon_norm": 6.484818458557129,
    "prediction_norm": 5.266894340515137,
    "timestamp": "2025-07-14T12:33:45.240263"
  },
  {
    "epoch": 332,
    "batch_idx": 128,
    "loss": 0.6211955547332764,
    "learning_rate": 0.0009335353505822451,
    "timestep_mean": 428.34375,
    "epsilon_norm": 7.917386054992676,
    "prediction_norm": 4.87903356552124,
    "timestamp": "2025-07-14T12:33:45.245429"
  },
  {
    "epoch": 332,
    "batch_idx": 160,
    "loss": 0.3513447642326355,
    "learning_rate": 0.0009335353505822451,
    "timestep_mean": 494.65625,
    "epsilon_norm": 8.262202262878418,
    "prediction_norm": 6.343782901763916,
    "timestamp": "2025-07-14T12:33:45.250852"
  },
  {
    "epoch": 332,
    "batch_idx": 192,
    "loss": 0.5877410769462585,
    "learning_rate": 0.0009335353505822451,
    "timestep_mean": 529.875,
    "epsilon_norm": 4.169889450073242,
    "prediction_norm": 1.8385127782821655,
    "timestamp": "2025-07-14T12:33:45.255954"
  },
  {
    "epoch": 333,
    "batch_idx": 0,
    "loss": 0.33256736397743225,
    "learning_rate": 0.0009331435422236939,
    "timestep_mean": 566.46875,
    "epsilon_norm": 7.387651443481445,
    "prediction_norm": 5.2947678565979,
    "timestamp": "2025-07-14T12:33:45.261082"
  },
  {
    "epoch": 333,
    "batch_idx": 32,
    "loss": 0.2733416259288788,
    "learning_rate": 0.0009331435422236939,
    "timestep_mean": 511.125,
    "epsilon_norm": 7.570917129516602,
    "prediction_norm": 5.47385835647583,
    "timestamp": "2025-07-14T12:33:45.266303"
  },
  {
    "epoch": 333,
    "batch_idx": 64,
    "loss": 0.28451815247535706,
    "learning_rate": 0.0009331435422236939,
    "timestep_mean": 580.53125,
    "epsilon_norm": 8.252553939819336,
    "prediction_norm": 6.4172868728637695,
    "timestamp": "2025-07-14T12:33:45.271473"
  },
  {
    "epoch": 333,
    "batch_idx": 96,
    "loss": 0.38049647212028503,
    "learning_rate": 0.0009331435422236939,
    "timestep_mean": 496.4375,
    "epsilon_norm": 9.042162895202637,
    "prediction_norm": 7.416004180908203,
    "timestamp": "2025-07-14T12:33:45.276680"
  },
  {
    "epoch": 333,
    "batch_idx": 128,
    "loss": 0.22472064197063446,
    "learning_rate": 0.0009331435422236939,
    "timestep_mean": 467.09375,
    "epsilon_norm": 8.539217948913574,
    "prediction_norm": 6.7636332511901855,
    "timestamp": "2025-07-14T12:33:45.281901"
  },
  {
    "epoch": 333,
    "batch_idx": 160,
    "loss": 0.1735662817955017,
    "learning_rate": 0.0009331435422236939,
    "timestep_mean": 524.28125,
    "epsilon_norm": 8.663803100585938,
    "prediction_norm": 8.062294006347656,
    "timestamp": "2025-07-14T12:33:45.286934"
  },
  {
    "epoch": 333,
    "batch_idx": 192,
    "loss": 0.7752264738082886,
    "learning_rate": 0.0009331435422236939,
    "timestep_mean": 282.375,
    "epsilon_norm": 4.217559337615967,
    "prediction_norm": 1.945594310760498,
    "timestamp": "2025-07-14T12:33:45.291994"
  },
  {
    "epoch": 334,
    "batch_idx": 0,
    "loss": 0.22349777817726135,
    "learning_rate": 0.0009327506651265095,
    "timestep_mean": 563.375,
    "epsilon_norm": 8.77393913269043,
    "prediction_norm": 7.790497779846191,
    "timestamp": "2025-07-14T12:33:45.296948"
  },
  {
    "epoch": 334,
    "batch_idx": 32,
    "loss": 0.3201819658279419,
    "learning_rate": 0.0009327506651265095,
    "timestep_mean": 497.65625,
    "epsilon_norm": 8.798712730407715,
    "prediction_norm": 7.121084690093994,
    "timestamp": "2025-07-14T12:33:45.302804"
  },
  {
    "epoch": 334,
    "batch_idx": 64,
    "loss": 0.4270124137401581,
    "learning_rate": 0.0009327506651265095,
    "timestep_mean": 478.21875,
    "epsilon_norm": 7.710042476654053,
    "prediction_norm": 5.821093559265137,
    "timestamp": "2025-07-14T12:33:45.307837"
  },
  {
    "epoch": 334,
    "batch_idx": 96,
    "loss": 0.34602659940719604,
    "learning_rate": 0.0009327506651265095,
    "timestep_mean": 508.96875,
    "epsilon_norm": 8.725031852722168,
    "prediction_norm": 7.1315765380859375,
    "timestamp": "2025-07-14T12:33:45.312794"
  },
  {
    "epoch": 334,
    "batch_idx": 128,
    "loss": 0.16061799228191376,
    "learning_rate": 0.0009327506651265095,
    "timestep_mean": 637.03125,
    "epsilon_norm": 8.489923477172852,
    "prediction_norm": 7.990312576293945,
    "timestamp": "2025-07-14T12:33:45.317661"
  },
  {
    "epoch": 334,
    "batch_idx": 160,
    "loss": 0.27662166953086853,
    "learning_rate": 0.0009327506651265095,
    "timestep_mean": 484.75,
    "epsilon_norm": 7.736158847808838,
    "prediction_norm": 6.522719383239746,
    "timestamp": "2025-07-14T12:33:45.322519"
  },
  {
    "epoch": 334,
    "batch_idx": 192,
    "loss": 0.4825608432292938,
    "learning_rate": 0.0009327506651265095,
    "timestep_mean": 418.0,
    "epsilon_norm": 4.346455097198486,
    "prediction_norm": 3.1563549041748047,
    "timestamp": "2025-07-14T12:33:45.327585"
  },
  {
    "epoch": 335,
    "batch_idx": 0,
    "loss": 0.30425673723220825,
    "learning_rate": 0.0009323567202600776,
    "timestep_mean": 508.59375,
    "epsilon_norm": 8.687613487243652,
    "prediction_norm": 8.334211349487305,
    "timestamp": "2025-07-14T12:33:45.332588"
  },
  {
    "epoch": 335,
    "batch_idx": 32,
    "loss": 0.23588493466377258,
    "learning_rate": 0.0009323567202600776,
    "timestep_mean": 533.875,
    "epsilon_norm": 7.402196407318115,
    "prediction_norm": 6.71196985244751,
    "timestamp": "2025-07-14T12:33:45.337775"
  },
  {
    "epoch": 335,
    "batch_idx": 64,
    "loss": 0.49367380142211914,
    "learning_rate": 0.0009323567202600776,
    "timestep_mean": 455.4375,
    "epsilon_norm": 9.38419246673584,
    "prediction_norm": 8.367477416992188,
    "timestamp": "2025-07-14T12:33:45.342677"
  },
  {
    "epoch": 335,
    "batch_idx": 96,
    "loss": 0.26254525780677795,
    "learning_rate": 0.0009323567202600776,
    "timestep_mean": 502.625,
    "epsilon_norm": 7.976943016052246,
    "prediction_norm": 7.733051300048828,
    "timestamp": "2025-07-14T12:33:45.347680"
  },
  {
    "epoch": 335,
    "batch_idx": 128,
    "loss": 0.2840430736541748,
    "learning_rate": 0.0009323567202600776,
    "timestep_mean": 466.03125,
    "epsilon_norm": 7.555212020874023,
    "prediction_norm": 7.176472187042236,
    "timestamp": "2025-07-14T12:33:45.352632"
  },
  {
    "epoch": 335,
    "batch_idx": 160,
    "loss": 0.2801705598831177,
    "learning_rate": 0.0009323567202600776,
    "timestep_mean": 569.25,
    "epsilon_norm": 8.648835182189941,
    "prediction_norm": 8.240962982177734,
    "timestamp": "2025-07-14T12:33:45.357667"
  },
  {
    "epoch": 335,
    "batch_idx": 192,
    "loss": 0.38471150398254395,
    "learning_rate": 0.0009323567202600776,
    "timestep_mean": 444.125,
    "epsilon_norm": 4.11828088760376,
    "prediction_norm": 3.249150037765503,
    "timestamp": "2025-07-14T12:33:45.362697"
  },
  {
    "epoch": 336,
    "batch_idx": 0,
    "loss": 0.2175339162349701,
    "learning_rate": 0.0009319617085964176,
    "timestep_mean": 521.65625,
    "epsilon_norm": 7.785294055938721,
    "prediction_norm": 6.6975321769714355,
    "timestamp": "2025-07-14T12:33:45.367665"
  },
  {
    "epoch": 336,
    "batch_idx": 32,
    "loss": 0.37076711654663086,
    "learning_rate": 0.0009319617085964176,
    "timestep_mean": 506.96875,
    "epsilon_norm": 7.927225589752197,
    "prediction_norm": 6.006929397583008,
    "timestamp": "2025-07-14T12:33:45.372654"
  },
  {
    "epoch": 336,
    "batch_idx": 64,
    "loss": 0.3167089521884918,
    "learning_rate": 0.0009319617085964176,
    "timestep_mean": 553.3125,
    "epsilon_norm": 7.372682571411133,
    "prediction_norm": 6.3062005043029785,
    "timestamp": "2025-07-14T12:33:45.377650"
  },
  {
    "epoch": 336,
    "batch_idx": 96,
    "loss": 0.36239421367645264,
    "learning_rate": 0.0009319617085964176,
    "timestep_mean": 462.59375,
    "epsilon_norm": 7.829444885253906,
    "prediction_norm": 5.622237682342529,
    "timestamp": "2025-07-14T12:33:45.382580"
  },
  {
    "epoch": 336,
    "batch_idx": 128,
    "loss": 0.4955889880657196,
    "learning_rate": 0.0009319617085964176,
    "timestep_mean": 427.53125,
    "epsilon_norm": 7.627050399780273,
    "prediction_norm": 5.162503242492676,
    "timestamp": "2025-07-14T12:33:45.387784"
  },
  {
    "epoch": 336,
    "batch_idx": 160,
    "loss": 0.24443936347961426,
    "learning_rate": 0.0009319617085964176,
    "timestep_mean": 531.46875,
    "epsilon_norm": 6.987743377685547,
    "prediction_norm": 6.010362148284912,
    "timestamp": "2025-07-14T12:33:45.392706"
  },
  {
    "epoch": 336,
    "batch_idx": 192,
    "loss": 0.22246088087558746,
    "learning_rate": 0.0009319617085964176,
    "timestep_mean": 484.0,
    "epsilon_norm": 3.953246593475342,
    "prediction_norm": 3.171989917755127,
    "timestamp": "2025-07-14T12:33:45.397700"
  },
  {
    "epoch": 337,
    "batch_idx": 0,
    "loss": 0.21914426982402802,
    "learning_rate": 0.0009315656311101819,
    "timestep_mean": 496.34375,
    "epsilon_norm": 7.8253607749938965,
    "prediction_norm": 6.177182197570801,
    "timestamp": "2025-07-14T12:33:45.402689"
  },
  {
    "epoch": 337,
    "batch_idx": 32,
    "loss": 0.23116575181484222,
    "learning_rate": 0.0009315656311101819,
    "timestep_mean": 548.3125,
    "epsilon_norm": 7.807775974273682,
    "prediction_norm": 6.25148344039917,
    "timestamp": "2025-07-14T12:33:45.407543"
  },
  {
    "epoch": 337,
    "batch_idx": 64,
    "loss": 0.3217014670372009,
    "learning_rate": 0.0009315656311101819,
    "timestep_mean": 506.1875,
    "epsilon_norm": 8.421021461486816,
    "prediction_norm": 6.277462482452393,
    "timestamp": "2025-07-14T12:33:45.412340"
  },
  {
    "epoch": 337,
    "batch_idx": 96,
    "loss": 0.26008340716362,
    "learning_rate": 0.0009315656311101819,
    "timestep_mean": 452.375,
    "epsilon_norm": 6.949124336242676,
    "prediction_norm": 6.01174783706665,
    "timestamp": "2025-07-14T12:33:45.417581"
  },
  {
    "epoch": 337,
    "batch_idx": 128,
    "loss": 0.33387699723243713,
    "learning_rate": 0.0009315656311101819,
    "timestep_mean": 526.15625,
    "epsilon_norm": 8.174365997314453,
    "prediction_norm": 6.500499248504639,
    "timestamp": "2025-07-14T12:33:45.422581"
  },
  {
    "epoch": 337,
    "batch_idx": 160,
    "loss": 0.42867743968963623,
    "learning_rate": 0.0009315656311101819,
    "timestep_mean": 368.53125,
    "epsilon_norm": 7.442055702209473,
    "prediction_norm": 5.300850868225098,
    "timestamp": "2025-07-14T12:33:45.427733"
  },
  {
    "epoch": 337,
    "batch_idx": 192,
    "loss": 0.2629264295101166,
    "learning_rate": 0.0009315656311101819,
    "timestep_mean": 406.25,
    "epsilon_norm": 2.480229139328003,
    "prediction_norm": 2.4766018390655518,
    "timestamp": "2025-07-14T12:33:45.432812"
  },
  {
    "epoch": 338,
    "batch_idx": 0,
    "loss": 0.2689002752304077,
    "learning_rate": 0.000931168488778652,
    "timestep_mean": 470.65625,
    "epsilon_norm": 8.316020011901855,
    "prediction_norm": 7.133758068084717,
    "timestamp": "2025-07-14T12:33:45.438155"
  },
  {
    "epoch": 338,
    "batch_idx": 32,
    "loss": 0.3902928829193115,
    "learning_rate": 0.000931168488778652,
    "timestep_mean": 400.25,
    "epsilon_norm": 7.642480850219727,
    "prediction_norm": 5.0584940910339355,
    "timestamp": "2025-07-14T12:33:45.443425"
  },
  {
    "epoch": 338,
    "batch_idx": 64,
    "loss": 0.17469149827957153,
    "learning_rate": 0.000931168488778652,
    "timestep_mean": 590.25,
    "epsilon_norm": 8.299661636352539,
    "prediction_norm": 7.48076057434082,
    "timestamp": "2025-07-14T12:33:45.448654"
  },
  {
    "epoch": 338,
    "batch_idx": 96,
    "loss": 0.2953590452671051,
    "learning_rate": 0.000931168488778652,
    "timestep_mean": 491.0,
    "epsilon_norm": 7.611939907073975,
    "prediction_norm": 6.442070484161377,
    "timestamp": "2025-07-14T12:33:45.453990"
  },
  {
    "epoch": 338,
    "batch_idx": 128,
    "loss": 0.35436034202575684,
    "learning_rate": 0.000931168488778652,
    "timestep_mean": 512.53125,
    "epsilon_norm": 8.043869018554688,
    "prediction_norm": 6.125267028808594,
    "timestamp": "2025-07-14T12:33:45.459096"
  },
  {
    "epoch": 338,
    "batch_idx": 160,
    "loss": 0.18288655579090118,
    "learning_rate": 0.000931168488778652,
    "timestep_mean": 482.25,
    "epsilon_norm": 6.712825775146484,
    "prediction_norm": 6.323061943054199,
    "timestamp": "2025-07-14T12:33:45.464186"
  },
  {
    "epoch": 338,
    "batch_idx": 192,
    "loss": 0.5240607857704163,
    "learning_rate": 0.000931168488778652,
    "timestep_mean": 501.375,
    "epsilon_norm": 4.4868292808532715,
    "prediction_norm": 3.747230291366577,
    "timestamp": "2025-07-14T12:33:45.469036"
  },
  {
    "epoch": 339,
    "batch_idx": 0,
    "loss": 0.38643205165863037,
    "learning_rate": 0.0009307702825817372,
    "timestep_mean": 491.1875,
    "epsilon_norm": 7.7814249992370605,
    "prediction_norm": 5.738709926605225,
    "timestamp": "2025-07-14T12:33:45.473943"
  },
  {
    "epoch": 339,
    "batch_idx": 32,
    "loss": 0.4370597004890442,
    "learning_rate": 0.0009307702825817372,
    "timestep_mean": 541.5,
    "epsilon_norm": 8.545014381408691,
    "prediction_norm": 6.661876678466797,
    "timestamp": "2025-07-14T12:33:45.478829"
  },
  {
    "epoch": 339,
    "batch_idx": 64,
    "loss": 0.37025710940361023,
    "learning_rate": 0.0009307702825817372,
    "timestep_mean": 508.25,
    "epsilon_norm": 8.735931396484375,
    "prediction_norm": 7.405289173126221,
    "timestamp": "2025-07-14T12:33:45.483740"
  },
  {
    "epoch": 339,
    "batch_idx": 96,
    "loss": 0.2896130383014679,
    "learning_rate": 0.0009307702825817372,
    "timestep_mean": 537.5625,
    "epsilon_norm": 8.168874740600586,
    "prediction_norm": 6.58588981628418,
    "timestamp": "2025-07-14T12:33:45.488573"
  },
  {
    "epoch": 339,
    "batch_idx": 128,
    "loss": 0.15566006302833557,
    "learning_rate": 0.0009307702825817372,
    "timestep_mean": 621.75,
    "epsilon_norm": 7.148989677429199,
    "prediction_norm": 7.15526008605957,
    "timestamp": "2025-07-14T12:33:45.493548"
  },
  {
    "epoch": 339,
    "batch_idx": 160,
    "loss": 0.18076938390731812,
    "learning_rate": 0.0009307702825817372,
    "timestep_mean": 596.53125,
    "epsilon_norm": 7.730767250061035,
    "prediction_norm": 7.252462863922119,
    "timestamp": "2025-07-14T12:33:45.498414"
  },
  {
    "epoch": 339,
    "batch_idx": 192,
    "loss": 0.2104681432247162,
    "learning_rate": 0.0009307702825817372,
    "timestep_mean": 565.5,
    "epsilon_norm": 4.15019416809082,
    "prediction_norm": 3.8536038398742676,
    "timestamp": "2025-07-14T12:33:45.503359"
  },
  {
    "epoch": 340,
    "batch_idx": 0,
    "loss": 0.3049585819244385,
    "learning_rate": 0.0009303710135019718,
    "timestep_mean": 527.0,
    "epsilon_norm": 7.452011585235596,
    "prediction_norm": 5.819995403289795,
    "timestamp": "2025-07-14T12:33:45.508368"
  },
  {
    "epoch": 340,
    "batch_idx": 32,
    "loss": 0.420470267534256,
    "learning_rate": 0.0009303710135019718,
    "timestep_mean": 442.75,
    "epsilon_norm": 7.835904121398926,
    "prediction_norm": 6.254153728485107,
    "timestamp": "2025-07-14T12:33:45.513201"
  },
  {
    "epoch": 340,
    "batch_idx": 64,
    "loss": 0.36916130781173706,
    "learning_rate": 0.0009303710135019718,
    "timestep_mean": 490.40625,
    "epsilon_norm": 8.086276054382324,
    "prediction_norm": 6.70589017868042,
    "timestamp": "2025-07-14T12:33:45.518140"
  },
  {
    "epoch": 340,
    "batch_idx": 96,
    "loss": 0.2879962921142578,
    "learning_rate": 0.0009303710135019718,
    "timestep_mean": 546.03125,
    "epsilon_norm": 7.518360614776611,
    "prediction_norm": 5.930495262145996,
    "timestamp": "2025-07-14T12:33:45.522961"
  },
  {
    "epoch": 340,
    "batch_idx": 128,
    "loss": 0.5112946033477783,
    "learning_rate": 0.0009303710135019718,
    "timestep_mean": 479.09375,
    "epsilon_norm": 8.836670875549316,
    "prediction_norm": 6.35638427734375,
    "timestamp": "2025-07-14T12:33:45.527792"
  },
  {
    "epoch": 340,
    "batch_idx": 160,
    "loss": 0.32400739192962646,
    "learning_rate": 0.0009303710135019718,
    "timestep_mean": 518.5,
    "epsilon_norm": 7.902315139770508,
    "prediction_norm": 6.143378734588623,
    "timestamp": "2025-07-14T12:33:45.532638"
  },
  {
    "epoch": 340,
    "batch_idx": 192,
    "loss": 0.39381974935531616,
    "learning_rate": 0.0009303710135019718,
    "timestep_mean": 432.75,
    "epsilon_norm": 4.455095291137695,
    "prediction_norm": 3.2642366886138916,
    "timestamp": "2025-07-14T12:33:45.537594"
  },
  {
    "epoch": 341,
    "batch_idx": 0,
    "loss": 0.43745070695877075,
    "learning_rate": 0.0009299706825245125,
    "timestep_mean": 459.03125,
    "epsilon_norm": 8.718620300292969,
    "prediction_norm": 6.755401611328125,
    "timestamp": "2025-07-14T12:33:45.542444"
  },
  {
    "epoch": 341,
    "batch_idx": 32,
    "loss": 0.6046879291534424,
    "learning_rate": 0.0009299706825245125,
    "timestep_mean": 414.625,
    "epsilon_norm": 8.956587791442871,
    "prediction_norm": 6.107207298278809,
    "timestamp": "2025-07-14T12:33:45.547315"
  },
  {
    "epoch": 341,
    "batch_idx": 64,
    "loss": 0.38803830742836,
    "learning_rate": 0.0009299706825245125,
    "timestep_mean": 493.90625,
    "epsilon_norm": 8.890917778015137,
    "prediction_norm": 6.952581405639648,
    "timestamp": "2025-07-14T12:33:45.552247"
  },
  {
    "epoch": 341,
    "batch_idx": 96,
    "loss": 0.4862794280052185,
    "learning_rate": 0.0009299706825245125,
    "timestep_mean": 490.9375,
    "epsilon_norm": 8.637434959411621,
    "prediction_norm": 7.180347919464111,
    "timestamp": "2025-07-14T12:33:45.557111"
  },
  {
    "epoch": 341,
    "batch_idx": 128,
    "loss": 0.34793955087661743,
    "learning_rate": 0.0009299706825245125,
    "timestep_mean": 495.5625,
    "epsilon_norm": 8.81333065032959,
    "prediction_norm": 7.898257255554199,
    "timestamp": "2025-07-14T12:33:45.564533"
  },
  {
    "epoch": 341,
    "batch_idx": 160,
    "loss": 0.36023932695388794,
    "learning_rate": 0.0009299706825245125,
    "timestep_mean": 496.28125,
    "epsilon_norm": 8.298604011535645,
    "prediction_norm": 7.650144100189209,
    "timestamp": "2025-07-14T12:33:45.571288"
  },
  {
    "epoch": 341,
    "batch_idx": 192,
    "loss": 0.5333171486854553,
    "learning_rate": 0.0009299706825245125,
    "timestep_mean": 553.125,
    "epsilon_norm": 3.3255746364593506,
    "prediction_norm": 2.1569912433624268,
    "timestamp": "2025-07-14T12:33:45.577433"
  },
  {
    "epoch": 342,
    "batch_idx": 0,
    "loss": 0.2769220471382141,
    "learning_rate": 0.0009295692906371362,
    "timestep_mean": 531.6875,
    "epsilon_norm": 7.700050354003906,
    "prediction_norm": 6.3543009757995605,
    "timestamp": "2025-07-14T12:33:45.585135"
  },
  {
    "epoch": 342,
    "batch_idx": 32,
    "loss": 0.18197639286518097,
    "learning_rate": 0.0009295692906371362,
    "timestep_mean": 545.0625,
    "epsilon_norm": 7.320470333099365,
    "prediction_norm": 7.209074020385742,
    "timestamp": "2025-07-14T12:33:45.592232"
  },
  {
    "epoch": 342,
    "batch_idx": 64,
    "loss": 0.21140410006046295,
    "learning_rate": 0.0009295692906371362,
    "timestep_mean": 551.125,
    "epsilon_norm": 7.8269829750061035,
    "prediction_norm": 7.6425628662109375,
    "timestamp": "2025-07-14T12:33:45.598379"
  },
  {
    "epoch": 342,
    "batch_idx": 96,
    "loss": 0.383544385433197,
    "learning_rate": 0.0009295692906371362,
    "timestep_mean": 426.1875,
    "epsilon_norm": 7.662734031677246,
    "prediction_norm": 6.31130838394165,
    "timestamp": "2025-07-14T12:33:45.604640"
  },
  {
    "epoch": 342,
    "batch_idx": 128,
    "loss": 0.2939422130584717,
    "learning_rate": 0.0009295692906371362,
    "timestep_mean": 463.09375,
    "epsilon_norm": 7.886014461517334,
    "prediction_norm": 6.787796497344971,
    "timestamp": "2025-07-14T12:33:45.611077"
  },
  {
    "epoch": 342,
    "batch_idx": 160,
    "loss": 0.42439353466033936,
    "learning_rate": 0.0009295692906371362,
    "timestep_mean": 562.9375,
    "epsilon_norm": 9.41513729095459,
    "prediction_norm": 7.445720672607422,
    "timestamp": "2025-07-14T12:33:45.616731"
  },
  {
    "epoch": 342,
    "batch_idx": 192,
    "loss": 0.407412052154541,
    "learning_rate": 0.0009295692906371362,
    "timestep_mean": 517.125,
    "epsilon_norm": 3.484480142593384,
    "prediction_norm": 2.8556432723999023,
    "timestamp": "2025-07-14T12:33:45.621889"
  },
  {
    "epoch": 343,
    "batch_idx": 0,
    "loss": 0.4174284338951111,
    "learning_rate": 0.0009291668388302374,
    "timestep_mean": 410.53125,
    "epsilon_norm": 7.261399745941162,
    "prediction_norm": 5.603984832763672,
    "timestamp": "2025-07-14T12:33:45.627245"
  },
  {
    "epoch": 343,
    "batch_idx": 32,
    "loss": 0.3718053102493286,
    "learning_rate": 0.0009291668388302374,
    "timestep_mean": 504.375,
    "epsilon_norm": 8.888189315795898,
    "prediction_norm": 6.4374494552612305,
    "timestamp": "2025-07-14T12:33:45.632287"
  },
  {
    "epoch": 343,
    "batch_idx": 64,
    "loss": 0.4526597261428833,
    "learning_rate": 0.0009291668388302374,
    "timestep_mean": 450.3125,
    "epsilon_norm": 8.283814430236816,
    "prediction_norm": 5.639289379119873,
    "timestamp": "2025-07-14T12:33:45.637397"
  },
  {
    "epoch": 343,
    "batch_idx": 96,
    "loss": 0.2760603427886963,
    "learning_rate": 0.0009291668388302374,
    "timestep_mean": 486.5625,
    "epsilon_norm": 7.35312557220459,
    "prediction_norm": 6.206159591674805,
    "timestamp": "2025-07-14T12:33:45.642483"
  },
  {
    "epoch": 343,
    "batch_idx": 128,
    "loss": 0.20876988768577576,
    "learning_rate": 0.0009291668388302374,
    "timestep_mean": 540.9375,
    "epsilon_norm": 7.184624195098877,
    "prediction_norm": 5.784562110900879,
    "timestamp": "2025-07-14T12:33:45.647574"
  },
  {
    "epoch": 343,
    "batch_idx": 160,
    "loss": 0.27597129344940186,
    "learning_rate": 0.0009291668388302374,
    "timestep_mean": 521.40625,
    "epsilon_norm": 8.456976890563965,
    "prediction_norm": 7.040165424346924,
    "timestamp": "2025-07-14T12:33:45.652599"
  },
  {
    "epoch": 343,
    "batch_idx": 192,
    "loss": 0.8268498778343201,
    "learning_rate": 0.0009291668388302374,
    "timestep_mean": 338.875,
    "epsilon_norm": 4.5645952224731445,
    "prediction_norm": 2.5703048706054688,
    "timestamp": "2025-07-14T12:33:45.657835"
  },
  {
    "epoch": 344,
    "batch_idx": 0,
    "loss": 0.4983888268470764,
    "learning_rate": 0.0009287633280968261,
    "timestep_mean": 474.1875,
    "epsilon_norm": 8.327317237854004,
    "prediction_norm": 6.370543003082275,
    "timestamp": "2025-07-14T12:33:45.662890"
  },
  {
    "epoch": 344,
    "batch_idx": 32,
    "loss": 0.29531359672546387,
    "learning_rate": 0.0009287633280968261,
    "timestep_mean": 544.40625,
    "epsilon_norm": 8.589101791381836,
    "prediction_norm": 7.6201300621032715,
    "timestamp": "2025-07-14T12:33:45.667837"
  },
  {
    "epoch": 344,
    "batch_idx": 64,
    "loss": 0.39503785967826843,
    "learning_rate": 0.0009287633280968261,
    "timestep_mean": 487.78125,
    "epsilon_norm": 7.958360195159912,
    "prediction_norm": 6.452298641204834,
    "timestamp": "2025-07-14T12:33:45.672963"
  },
  {
    "epoch": 344,
    "batch_idx": 96,
    "loss": 0.387229323387146,
    "learning_rate": 0.0009287633280968261,
    "timestep_mean": 532.53125,
    "epsilon_norm": 8.633830070495605,
    "prediction_norm": 7.192997932434082,
    "timestamp": "2025-07-14T12:33:45.677925"
  },
  {
    "epoch": 344,
    "batch_idx": 128,
    "loss": 0.28998202085494995,
    "learning_rate": 0.0009287633280968261,
    "timestep_mean": 510.46875,
    "epsilon_norm": 8.711455345153809,
    "prediction_norm": 8.076580047607422,
    "timestamp": "2025-07-14T12:33:45.682833"
  },
  {
    "epoch": 344,
    "batch_idx": 160,
    "loss": 0.44956040382385254,
    "learning_rate": 0.0009287633280968261,
    "timestep_mean": 421.84375,
    "epsilon_norm": 8.55441951751709,
    "prediction_norm": 6.279787540435791,
    "timestamp": "2025-07-14T12:33:45.687950"
  },
  {
    "epoch": 344,
    "batch_idx": 192,
    "loss": 0.22316387295722961,
    "learning_rate": 0.0009287633280968261,
    "timestep_mean": 620.5,
    "epsilon_norm": 3.871630907058716,
    "prediction_norm": 3.3542778491973877,
    "timestamp": "2025-07-14T12:33:45.693144"
  },
  {
    "epoch": 345,
    "batch_idx": 0,
    "loss": 0.35950738191604614,
    "learning_rate": 0.0009283587594325248,
    "timestep_mean": 464.4375,
    "epsilon_norm": 8.4182710647583,
    "prediction_norm": 6.603404521942139,
    "timestamp": "2025-07-14T12:33:45.698481"
  },
  {
    "epoch": 345,
    "batch_idx": 32,
    "loss": 0.18282504379749298,
    "learning_rate": 0.0009283587594325248,
    "timestep_mean": 544.5625,
    "epsilon_norm": 8.57090950012207,
    "prediction_norm": 8.09542465209961,
    "timestamp": "2025-07-14T12:33:45.703792"
  },
  {
    "epoch": 345,
    "batch_idx": 64,
    "loss": 0.2220020592212677,
    "learning_rate": 0.0009283587594325248,
    "timestep_mean": 562.09375,
    "epsilon_norm": 8.220338821411133,
    "prediction_norm": 6.959495544433594,
    "timestamp": "2025-07-14T12:33:45.709134"
  },
  {
    "epoch": 345,
    "batch_idx": 96,
    "loss": 0.44684547185897827,
    "learning_rate": 0.0009283587594325248,
    "timestep_mean": 457.8125,
    "epsilon_norm": 8.355923652648926,
    "prediction_norm": 6.871048927307129,
    "timestamp": "2025-07-14T12:33:45.714343"
  },
  {
    "epoch": 345,
    "batch_idx": 128,
    "loss": 0.22419771552085876,
    "learning_rate": 0.0009283587594325248,
    "timestep_mean": 544.4375,
    "epsilon_norm": 7.776349067687988,
    "prediction_norm": 6.402248382568359,
    "timestamp": "2025-07-14T12:33:45.719762"
  },
  {
    "epoch": 345,
    "batch_idx": 160,
    "loss": 0.17871752381324768,
    "learning_rate": 0.0009283587594325248,
    "timestep_mean": 534.6875,
    "epsilon_norm": 7.6884446144104,
    "prediction_norm": 7.010536193847656,
    "timestamp": "2025-07-14T12:33:45.725285"
  },
  {
    "epoch": 345,
    "batch_idx": 192,
    "loss": 0.3496919274330139,
    "learning_rate": 0.0009283587594325248,
    "timestep_mean": 535.375,
    "epsilon_norm": 3.479255199432373,
    "prediction_norm": 2.9373080730438232,
    "timestamp": "2025-07-14T12:33:45.730648"
  },
  {
    "epoch": 346,
    "batch_idx": 0,
    "loss": 0.28434357047080994,
    "learning_rate": 0.0009279531338355665,
    "timestep_mean": 503.3125,
    "epsilon_norm": 7.467909336090088,
    "prediction_norm": 6.000118255615234,
    "timestamp": "2025-07-14T12:33:45.736028"
  },
  {
    "epoch": 346,
    "batch_idx": 32,
    "loss": 0.6578769683837891,
    "learning_rate": 0.0009279531338355665,
    "timestep_mean": 435.15625,
    "epsilon_norm": 7.674469947814941,
    "prediction_norm": 4.701651096343994,
    "timestamp": "2025-07-14T12:33:45.741320"
  },
  {
    "epoch": 346,
    "batch_idx": 64,
    "loss": 0.2769275903701782,
    "learning_rate": 0.0009279531338355665,
    "timestep_mean": 537.4375,
    "epsilon_norm": 8.288272857666016,
    "prediction_norm": 6.5998430252075195,
    "timestamp": "2025-07-14T12:33:45.746678"
  },
  {
    "epoch": 346,
    "batch_idx": 96,
    "loss": 0.18819482624530792,
    "learning_rate": 0.0009279531338355665,
    "timestep_mean": 611.53125,
    "epsilon_norm": 8.262740135192871,
    "prediction_norm": 7.3426127433776855,
    "timestamp": "2025-07-14T12:33:45.751978"
  },
  {
    "epoch": 346,
    "batch_idx": 128,
    "loss": 0.43251341581344604,
    "learning_rate": 0.0009279531338355665,
    "timestep_mean": 478.65625,
    "epsilon_norm": 8.115120887756348,
    "prediction_norm": 6.168483734130859,
    "timestamp": "2025-07-14T12:33:45.757058"
  },
  {
    "epoch": 346,
    "batch_idx": 160,
    "loss": 0.288064569234848,
    "learning_rate": 0.0009279531338355665,
    "timestep_mean": 586.375,
    "epsilon_norm": 9.05400562286377,
    "prediction_norm": 7.879845142364502,
    "timestamp": "2025-07-14T12:33:45.762263"
  },
  {
    "epoch": 346,
    "batch_idx": 192,
    "loss": 0.6350671052932739,
    "learning_rate": 0.0009279531338355665,
    "timestep_mean": 441.125,
    "epsilon_norm": 4.494472503662109,
    "prediction_norm": 2.732848644256592,
    "timestamp": "2025-07-14T12:33:45.767434"
  },
  {
    "epoch": 347,
    "batch_idx": 0,
    "loss": 0.3789244294166565,
    "learning_rate": 0.000927546452306792,
    "timestep_mean": 482.25,
    "epsilon_norm": 7.677286148071289,
    "prediction_norm": 5.882498741149902,
    "timestamp": "2025-07-14T12:33:45.772526"
  },
  {
    "epoch": 347,
    "batch_idx": 32,
    "loss": 0.19620172679424286,
    "learning_rate": 0.000927546452306792,
    "timestep_mean": 501.59375,
    "epsilon_norm": 7.779372215270996,
    "prediction_norm": 7.379083633422852,
    "timestamp": "2025-07-14T12:33:45.777766"
  },
  {
    "epoch": 347,
    "batch_idx": 64,
    "loss": 0.32789766788482666,
    "learning_rate": 0.000927546452306792,
    "timestep_mean": 547.78125,
    "epsilon_norm": 7.2854437828063965,
    "prediction_norm": 6.2524566650390625,
    "timestamp": "2025-07-14T12:33:45.782994"
  },
  {
    "epoch": 347,
    "batch_idx": 96,
    "loss": 0.20649778842926025,
    "learning_rate": 0.000927546452306792,
    "timestep_mean": 513.5625,
    "epsilon_norm": 7.237482070922852,
    "prediction_norm": 7.227352142333984,
    "timestamp": "2025-07-14T12:33:45.788205"
  },
  {
    "epoch": 347,
    "batch_idx": 128,
    "loss": 0.28904294967651367,
    "learning_rate": 0.000927546452306792,
    "timestep_mean": 574.6875,
    "epsilon_norm": 9.041394233703613,
    "prediction_norm": 7.652707099914551,
    "timestamp": "2025-07-14T12:33:45.793409"
  },
  {
    "epoch": 347,
    "batch_idx": 160,
    "loss": 0.2307884395122528,
    "learning_rate": 0.000927546452306792,
    "timestep_mean": 574.46875,
    "epsilon_norm": 8.947169303894043,
    "prediction_norm": 8.429522514343262,
    "timestamp": "2025-07-14T12:33:45.798595"
  },
  {
    "epoch": 347,
    "batch_idx": 192,
    "loss": 0.23045611381530762,
    "learning_rate": 0.000927546452306792,
    "timestep_mean": 588.875,
    "epsilon_norm": 3.677529811859131,
    "prediction_norm": 3.264944314956665,
    "timestamp": "2025-07-14T12:33:45.803676"
  },
  {
    "epoch": 348,
    "batch_idx": 0,
    "loss": 0.4472619891166687,
    "learning_rate": 0.0009271387158496476,
    "timestep_mean": 453.90625,
    "epsilon_norm": 7.852653980255127,
    "prediction_norm": 5.47428560256958,
    "timestamp": "2025-07-14T12:33:45.808842"
  },
  {
    "epoch": 348,
    "batch_idx": 32,
    "loss": 0.28979599475860596,
    "learning_rate": 0.0009271387158496476,
    "timestep_mean": 539.03125,
    "epsilon_norm": 8.676971435546875,
    "prediction_norm": 7.421919345855713,
    "timestamp": "2025-07-14T12:33:45.814038"
  },
  {
    "epoch": 348,
    "batch_idx": 64,
    "loss": 0.2123512625694275,
    "learning_rate": 0.0009271387158496476,
    "timestep_mean": 587.0,
    "epsilon_norm": 6.782299995422363,
    "prediction_norm": 6.092621326446533,
    "timestamp": "2025-07-14T12:33:45.819078"
  },
  {
    "epoch": 348,
    "batch_idx": 96,
    "loss": 0.3205406069755554,
    "learning_rate": 0.0009271387158496476,
    "timestep_mean": 458.71875,
    "epsilon_norm": 7.197538375854492,
    "prediction_norm": 5.75623893737793,
    "timestamp": "2025-07-14T12:33:45.824185"
  },
  {
    "epoch": 348,
    "batch_idx": 128,
    "loss": 0.3068937361240387,
    "learning_rate": 0.0009271387158496476,
    "timestep_mean": 437.46875,
    "epsilon_norm": 8.10871696472168,
    "prediction_norm": 6.750082969665527,
    "timestamp": "2025-07-14T12:33:45.829380"
  },
  {
    "epoch": 348,
    "batch_idx": 160,
    "loss": 0.44827163219451904,
    "learning_rate": 0.0009271387158496476,
    "timestep_mean": 478.59375,
    "epsilon_norm": 7.673686981201172,
    "prediction_norm": 5.868881702423096,
    "timestamp": "2025-07-14T12:33:45.834300"
  },
  {
    "epoch": 348,
    "batch_idx": 192,
    "loss": 0.4430253505706787,
    "learning_rate": 0.0009271387158496476,
    "timestep_mean": 350.875,
    "epsilon_norm": 4.401247024536133,
    "prediction_norm": 3.11020827293396,
    "timestamp": "2025-07-14T12:33:45.839186"
  },
  {
    "epoch": 349,
    "batch_idx": 0,
    "loss": 0.2516966462135315,
    "learning_rate": 0.0009267299254701824,
    "timestep_mean": 483.34375,
    "epsilon_norm": 7.337686538696289,
    "prediction_norm": 6.291324615478516,
    "timestamp": "2025-07-14T12:33:45.844115"
  },
  {
    "epoch": 349,
    "batch_idx": 32,
    "loss": 0.23579296469688416,
    "learning_rate": 0.0009267299254701824,
    "timestep_mean": 511.53125,
    "epsilon_norm": 7.458775043487549,
    "prediction_norm": 5.68925142288208,
    "timestamp": "2025-07-14T12:33:45.849019"
  },
  {
    "epoch": 349,
    "batch_idx": 64,
    "loss": 0.32757002115249634,
    "learning_rate": 0.0009267299254701824,
    "timestep_mean": 553.5625,
    "epsilon_norm": 8.495429992675781,
    "prediction_norm": 7.063802242279053,
    "timestamp": "2025-07-14T12:33:45.853929"
  },
  {
    "epoch": 349,
    "batch_idx": 96,
    "loss": 0.3759228587150574,
    "learning_rate": 0.0009267299254701824,
    "timestep_mean": 533.8125,
    "epsilon_norm": 8.571170806884766,
    "prediction_norm": 6.236736297607422,
    "timestamp": "2025-07-14T12:33:45.858917"
  },
  {
    "epoch": 349,
    "batch_idx": 128,
    "loss": 0.28426429629325867,
    "learning_rate": 0.0009267299254701824,
    "timestep_mean": 469.46875,
    "epsilon_norm": 7.779723644256592,
    "prediction_norm": 5.976394176483154,
    "timestamp": "2025-07-14T12:33:45.863837"
  },
  {
    "epoch": 349,
    "batch_idx": 160,
    "loss": 0.3021465837955475,
    "learning_rate": 0.0009267299254701824,
    "timestep_mean": 560.84375,
    "epsilon_norm": 8.294951438903809,
    "prediction_norm": 6.414536952972412,
    "timestamp": "2025-07-14T12:33:45.868763"
  },
  {
    "epoch": 349,
    "batch_idx": 192,
    "loss": 0.1624496579170227,
    "learning_rate": 0.0009267299254701824,
    "timestep_mean": 395.625,
    "epsilon_norm": 3.5655558109283447,
    "prediction_norm": 3.2932350635528564,
    "timestamp": "2025-07-14T12:33:45.873648"
  },
  {
    "epoch": 350,
    "batch_idx": 0,
    "loss": 0.28362128138542175,
    "learning_rate": 0.0009263200821770461,
    "timestep_mean": 503.78125,
    "epsilon_norm": 8.45587158203125,
    "prediction_norm": 7.224867343902588,
    "timestamp": "2025-07-14T12:33:45.878534"
  },
  {
    "epoch": 350,
    "batch_idx": 32,
    "loss": 0.2986801564693451,
    "learning_rate": 0.0009263200821770461,
    "timestep_mean": 410.15625,
    "epsilon_norm": 8.510860443115234,
    "prediction_norm": 7.555502414703369,
    "timestamp": "2025-07-14T12:33:45.883463"
  },
  {
    "epoch": 350,
    "batch_idx": 64,
    "loss": 0.1598212718963623,
    "learning_rate": 0.0009263200821770461,
    "timestep_mean": 498.90625,
    "epsilon_norm": 7.05086088180542,
    "prediction_norm": 6.265646457672119,
    "timestamp": "2025-07-14T12:33:45.888370"
  },
  {
    "epoch": 350,
    "batch_idx": 96,
    "loss": 0.1699843406677246,
    "learning_rate": 0.0009263200821770461,
    "timestep_mean": 559.34375,
    "epsilon_norm": 9.291553497314453,
    "prediction_norm": 8.597644805908203,
    "timestamp": "2025-07-14T12:33:45.893257"
  },
  {
    "epoch": 350,
    "batch_idx": 128,
    "loss": 0.2695990204811096,
    "learning_rate": 0.0009263200821770461,
    "timestep_mean": 463.4375,
    "epsilon_norm": 6.958752155303955,
    "prediction_norm": 5.7608561515808105,
    "timestamp": "2025-07-14T12:33:45.898105"
  },
  {
    "epoch": 350,
    "batch_idx": 160,
    "loss": 0.23729167878627777,
    "learning_rate": 0.0009263200821770461,
    "timestep_mean": 513.03125,
    "epsilon_norm": 8.86258602142334,
    "prediction_norm": 8.79459285736084,
    "timestamp": "2025-07-14T12:33:45.906830"
  },
  {
    "epoch": 350,
    "batch_idx": 192,
    "loss": 0.3834182322025299,
    "learning_rate": 0.0009263200821770461,
    "timestep_mean": 573.5,
    "epsilon_norm": 4.346140384674072,
    "prediction_norm": 3.1181931495666504,
    "timestamp": "2025-07-14T12:33:45.911752"
  },
  {
    "epoch": 351,
    "batch_idx": 0,
    "loss": 0.28416189551353455,
    "learning_rate": 0.0009259091869814862,
    "timestep_mean": 573.90625,
    "epsilon_norm": 7.492339134216309,
    "prediction_norm": 7.185687065124512,
    "timestamp": "2025-07-14T12:33:45.916704"
  },
  {
    "epoch": 351,
    "batch_idx": 32,
    "loss": 0.24578794836997986,
    "learning_rate": 0.0009259091869814862,
    "timestep_mean": 499.125,
    "epsilon_norm": 7.16976261138916,
    "prediction_norm": 6.966254234313965,
    "timestamp": "2025-07-14T12:33:45.921572"
  },
  {
    "epoch": 351,
    "batch_idx": 64,
    "loss": 0.4958178400993347,
    "learning_rate": 0.0009259091869814862,
    "timestep_mean": 506.96875,
    "epsilon_norm": 8.076253890991211,
    "prediction_norm": 6.122076988220215,
    "timestamp": "2025-07-14T12:33:45.926469"
  },
  {
    "epoch": 351,
    "batch_idx": 96,
    "loss": 0.3313004970550537,
    "learning_rate": 0.0009259091869814862,
    "timestep_mean": 479.1875,
    "epsilon_norm": 7.6975226402282715,
    "prediction_norm": 6.209826946258545,
    "timestamp": "2025-07-14T12:33:45.931391"
  },
  {
    "epoch": 351,
    "batch_idx": 128,
    "loss": 0.30046772956848145,
    "learning_rate": 0.0009259091869814862,
    "timestep_mean": 416.34375,
    "epsilon_norm": 8.13823413848877,
    "prediction_norm": 6.770497798919678,
    "timestamp": "2025-07-14T12:33:45.936274"
  },
  {
    "epoch": 351,
    "batch_idx": 160,
    "loss": 0.35389190912246704,
    "learning_rate": 0.0009259091869814862,
    "timestep_mean": 428.28125,
    "epsilon_norm": 8.136228561401367,
    "prediction_norm": 6.672805309295654,
    "timestamp": "2025-07-14T12:33:45.941148"
  },
  {
    "epoch": 351,
    "batch_idx": 192,
    "loss": 0.45606833696365356,
    "learning_rate": 0.0009259091869814862,
    "timestep_mean": 432.875,
    "epsilon_norm": 3.498289108276367,
    "prediction_norm": 2.8510475158691406,
    "timestamp": "2025-07-14T12:33:45.946042"
  },
  {
    "epoch": 352,
    "batch_idx": 0,
    "loss": 0.4183984696865082,
    "learning_rate": 0.000925497240897346,
    "timestep_mean": 405.59375,
    "epsilon_norm": 7.24595832824707,
    "prediction_norm": 4.8339762687683105,
    "timestamp": "2025-07-14T12:33:45.950941"
  },
  {
    "epoch": 352,
    "batch_idx": 32,
    "loss": 0.33042091131210327,
    "learning_rate": 0.000925497240897346,
    "timestep_mean": 543.28125,
    "epsilon_norm": 8.105301856994629,
    "prediction_norm": 6.403093338012695,
    "timestamp": "2025-07-14T12:33:45.955867"
  },
  {
    "epoch": 352,
    "batch_idx": 64,
    "loss": 0.16506752371788025,
    "learning_rate": 0.000925497240897346,
    "timestep_mean": 655.0625,
    "epsilon_norm": 8.310770034790039,
    "prediction_norm": 7.568541049957275,
    "timestamp": "2025-07-14T12:33:45.960725"
  },
  {
    "epoch": 352,
    "batch_idx": 96,
    "loss": 0.30065077543258667,
    "learning_rate": 0.000925497240897346,
    "timestep_mean": 453.09375,
    "epsilon_norm": 7.8946309089660645,
    "prediction_norm": 6.087364673614502,
    "timestamp": "2025-07-14T12:33:45.965616"
  },
  {
    "epoch": 352,
    "batch_idx": 128,
    "loss": 0.3529072403907776,
    "learning_rate": 0.000925497240897346,
    "timestep_mean": 499.1875,
    "epsilon_norm": 8.266913414001465,
    "prediction_norm": 6.721081256866455,
    "timestamp": "2025-07-14T12:33:45.970538"
  },
  {
    "epoch": 352,
    "batch_idx": 160,
    "loss": 0.44041430950164795,
    "learning_rate": 0.000925497240897346,
    "timestep_mean": 449.25,
    "epsilon_norm": 8.565659523010254,
    "prediction_norm": 6.712456703186035,
    "timestamp": "2025-07-14T12:33:45.975410"
  },
  {
    "epoch": 352,
    "batch_idx": 192,
    "loss": 0.505704402923584,
    "learning_rate": 0.000925497240897346,
    "timestep_mean": 531.375,
    "epsilon_norm": 4.2508416175842285,
    "prediction_norm": 3.0438289642333984,
    "timestamp": "2025-07-14T12:33:45.980225"
  },
  {
    "epoch": 353,
    "batch_idx": 0,
    "loss": 0.3801248073577881,
    "learning_rate": 0.0009250842449410611,
    "timestep_mean": 481.21875,
    "epsilon_norm": 7.85445499420166,
    "prediction_norm": 6.540495872497559,
    "timestamp": "2025-07-14T12:33:45.985094"
  },
  {
    "epoch": 353,
    "batch_idx": 32,
    "loss": 0.3773317337036133,
    "learning_rate": 0.0009250842449410611,
    "timestep_mean": 538.6875,
    "epsilon_norm": 8.441593170166016,
    "prediction_norm": 6.853285789489746,
    "timestamp": "2025-07-14T12:33:45.989926"
  },
  {
    "epoch": 353,
    "batch_idx": 64,
    "loss": 0.38211578130722046,
    "learning_rate": 0.0009250842449410611,
    "timestep_mean": 487.15625,
    "epsilon_norm": 8.256522178649902,
    "prediction_norm": 5.982324600219727,
    "timestamp": "2025-07-14T12:33:45.994813"
  },
  {
    "epoch": 353,
    "batch_idx": 96,
    "loss": 0.6010164022445679,
    "learning_rate": 0.0009250842449410611,
    "timestep_mean": 425.75,
    "epsilon_norm": 8.559761047363281,
    "prediction_norm": 6.360806465148926,
    "timestamp": "2025-07-14T12:33:45.999810"
  },
  {
    "epoch": 353,
    "batch_idx": 128,
    "loss": 0.18531757593154907,
    "learning_rate": 0.0009250842449410611,
    "timestep_mean": 573.125,
    "epsilon_norm": 7.477869033813477,
    "prediction_norm": 6.7126264572143555,
    "timestamp": "2025-07-14T12:33:46.004679"
  },
  {
    "epoch": 353,
    "batch_idx": 160,
    "loss": 0.2817617654800415,
    "learning_rate": 0.0009250842449410611,
    "timestep_mean": 553.71875,
    "epsilon_norm": 7.118169784545898,
    "prediction_norm": 5.363905906677246,
    "timestamp": "2025-07-14T12:33:46.009617"
  },
  {
    "epoch": 353,
    "batch_idx": 192,
    "loss": 0.22667458653450012,
    "learning_rate": 0.0009250842449410611,
    "timestep_mean": 509.375,
    "epsilon_norm": 3.795522928237915,
    "prediction_norm": 3.183624029159546,
    "timestamp": "2025-07-14T12:33:46.014511"
  },
  {
    "epoch": 354,
    "batch_idx": 0,
    "loss": 0.2404608428478241,
    "learning_rate": 0.0009246702001316581,
    "timestep_mean": 483.0,
    "epsilon_norm": 8.271102905273438,
    "prediction_norm": 7.6227874755859375,
    "timestamp": "2025-07-14T12:33:46.019400"
  },
  {
    "epoch": 354,
    "batch_idx": 32,
    "loss": 0.33729279041290283,
    "learning_rate": 0.0009246702001316581,
    "timestep_mean": 574.0625,
    "epsilon_norm": 8.583921432495117,
    "prediction_norm": 7.051599025726318,
    "timestamp": "2025-07-14T12:33:46.024353"
  },
  {
    "epoch": 354,
    "batch_idx": 64,
    "loss": 0.4062849283218384,
    "learning_rate": 0.0009246702001316581,
    "timestep_mean": 552.75,
    "epsilon_norm": 8.617365837097168,
    "prediction_norm": 6.7236480712890625,
    "timestamp": "2025-07-14T12:33:46.029290"
  },
  {
    "epoch": 354,
    "batch_idx": 96,
    "loss": 0.3330228626728058,
    "learning_rate": 0.0009246702001316581,
    "timestep_mean": 507.4375,
    "epsilon_norm": 7.7520647048950195,
    "prediction_norm": 6.194860458374023,
    "timestamp": "2025-07-14T12:33:46.034227"
  },
  {
    "epoch": 354,
    "batch_idx": 128,
    "loss": 0.328129380941391,
    "learning_rate": 0.0009246702001316581,
    "timestep_mean": 533.75,
    "epsilon_norm": 9.085224151611328,
    "prediction_norm": 8.032323837280273,
    "timestamp": "2025-07-14T12:33:46.039107"
  },
  {
    "epoch": 354,
    "batch_idx": 160,
    "loss": 0.1581132411956787,
    "learning_rate": 0.0009246702001316581,
    "timestep_mean": 520.21875,
    "epsilon_norm": 6.679025173187256,
    "prediction_norm": 6.346819877624512,
    "timestamp": "2025-07-14T12:33:46.044040"
  },
  {
    "epoch": 354,
    "batch_idx": 192,
    "loss": 0.29766616225242615,
    "learning_rate": 0.0009246702001316581,
    "timestep_mean": 501.375,
    "epsilon_norm": 3.4655964374542236,
    "prediction_norm": 3.0386786460876465,
    "timestamp": "2025-07-14T12:33:46.048922"
  },
  {
    "epoch": 355,
    "batch_idx": 0,
    "loss": 0.36495643854141235,
    "learning_rate": 0.0009242551074907518,
    "timestep_mean": 502.21875,
    "epsilon_norm": 9.064933776855469,
    "prediction_norm": 7.7690653800964355,
    "timestamp": "2025-07-14T12:33:46.053806"
  },
  {
    "epoch": 355,
    "batch_idx": 32,
    "loss": 0.2376464158296585,
    "learning_rate": 0.0009242551074907518,
    "timestep_mean": 506.125,
    "epsilon_norm": 7.858137130737305,
    "prediction_norm": 7.228971004486084,
    "timestamp": "2025-07-14T12:33:46.058763"
  },
  {
    "epoch": 355,
    "batch_idx": 64,
    "loss": 0.15475575625896454,
    "learning_rate": 0.0009242551074907518,
    "timestep_mean": 575.96875,
    "epsilon_norm": 8.812658309936523,
    "prediction_norm": 8.913000106811523,
    "timestamp": "2025-07-14T12:33:46.063669"
  },
  {
    "epoch": 355,
    "batch_idx": 96,
    "loss": 0.22614750266075134,
    "learning_rate": 0.0009242551074907518,
    "timestep_mean": 524.1875,
    "epsilon_norm": 7.991516590118408,
    "prediction_norm": 7.23611307144165,
    "timestamp": "2025-07-14T12:33:46.068600"
  },
  {
    "epoch": 355,
    "batch_idx": 128,
    "loss": 0.2559262812137604,
    "learning_rate": 0.0009242551074907518,
    "timestep_mean": 531.25,
    "epsilon_norm": 7.516300678253174,
    "prediction_norm": 7.044670581817627,
    "timestamp": "2025-07-14T12:33:46.073476"
  },
  {
    "epoch": 355,
    "batch_idx": 160,
    "loss": 0.26915767788887024,
    "learning_rate": 0.0009242551074907518,
    "timestep_mean": 539.9375,
    "epsilon_norm": 8.355189323425293,
    "prediction_norm": 7.383586406707764,
    "timestamp": "2025-07-14T12:33:46.078372"
  },
  {
    "epoch": 355,
    "batch_idx": 192,
    "loss": 0.2495191991329193,
    "learning_rate": 0.0009242551074907518,
    "timestep_mean": 564.625,
    "epsilon_norm": 4.162978172302246,
    "prediction_norm": 3.2782018184661865,
    "timestamp": "2025-07-14T12:33:46.083391"
  },
  {
    "epoch": 356,
    "batch_idx": 0,
    "loss": 0.3014598488807678,
    "learning_rate": 0.0009238389680425415,
    "timestep_mean": 598.8125,
    "epsilon_norm": 8.490276336669922,
    "prediction_norm": 6.75814962387085,
    "timestamp": "2025-07-14T12:33:46.088273"
  },
  {
    "epoch": 356,
    "batch_idx": 32,
    "loss": 0.3072460889816284,
    "learning_rate": 0.0009238389680425415,
    "timestep_mean": 481.4375,
    "epsilon_norm": 7.704914569854736,
    "prediction_norm": 6.447755813598633,
    "timestamp": "2025-07-14T12:33:46.093182"
  },
  {
    "epoch": 356,
    "batch_idx": 64,
    "loss": 0.34397557377815247,
    "learning_rate": 0.0009238389680425415,
    "timestep_mean": 460.96875,
    "epsilon_norm": 7.570664882659912,
    "prediction_norm": 6.0762410163879395,
    "timestamp": "2025-07-14T12:33:46.098083"
  },
  {
    "epoch": 356,
    "batch_idx": 96,
    "loss": 0.24425369501113892,
    "learning_rate": 0.0009238389680425415,
    "timestep_mean": 532.5625,
    "epsilon_norm": 6.739645957946777,
    "prediction_norm": 5.826333045959473,
    "timestamp": "2025-07-14T12:33:46.103012"
  },
  {
    "epoch": 356,
    "batch_idx": 128,
    "loss": 0.42343243956565857,
    "learning_rate": 0.0009238389680425415,
    "timestep_mean": 459.15625,
    "epsilon_norm": 8.622453689575195,
    "prediction_norm": 6.976572036743164,
    "timestamp": "2025-07-14T12:33:46.107891"
  },
  {
    "epoch": 356,
    "batch_idx": 160,
    "loss": 0.4629381000995636,
    "learning_rate": 0.0009238389680425415,
    "timestep_mean": 413.0625,
    "epsilon_norm": 8.639791488647461,
    "prediction_norm": 6.236761093139648,
    "timestamp": "2025-07-14T12:33:46.112880"
  },
  {
    "epoch": 356,
    "batch_idx": 192,
    "loss": 0.1399846374988556,
    "learning_rate": 0.0009238389680425415,
    "timestep_mean": 587.875,
    "epsilon_norm": 3.276784658432007,
    "prediction_norm": 2.8960628509521484,
    "timestamp": "2025-07-14T12:33:46.117720"
  },
  {
    "epoch": 357,
    "batch_idx": 0,
    "loss": 0.3500228226184845,
    "learning_rate": 0.0009234217828138102,
    "timestep_mean": 487.6875,
    "epsilon_norm": 8.675390243530273,
    "prediction_norm": 6.711639881134033,
    "timestamp": "2025-07-14T12:33:46.122579"
  },
  {
    "epoch": 357,
    "batch_idx": 32,
    "loss": 0.23746927082538605,
    "learning_rate": 0.0009234217828138102,
    "timestep_mean": 597.5625,
    "epsilon_norm": 7.829845905303955,
    "prediction_norm": 6.949378967285156,
    "timestamp": "2025-07-14T12:33:46.127480"
  },
  {
    "epoch": 357,
    "batch_idx": 64,
    "loss": 0.19940337538719177,
    "learning_rate": 0.0009234217828138102,
    "timestep_mean": 499.90625,
    "epsilon_norm": 7.272525310516357,
    "prediction_norm": 6.157571792602539,
    "timestamp": "2025-07-14T12:33:46.132379"
  },
  {
    "epoch": 357,
    "batch_idx": 96,
    "loss": 0.43341749906539917,
    "learning_rate": 0.0009234217828138102,
    "timestep_mean": 503.65625,
    "epsilon_norm": 8.541666030883789,
    "prediction_norm": 6.2663445472717285,
    "timestamp": "2025-07-14T12:33:46.137320"
  },
  {
    "epoch": 357,
    "batch_idx": 128,
    "loss": 0.27276611328125,
    "learning_rate": 0.0009234217828138102,
    "timestep_mean": 536.5,
    "epsilon_norm": 7.7380523681640625,
    "prediction_norm": 6.108105659484863,
    "timestamp": "2025-07-14T12:33:46.142271"
  },
  {
    "epoch": 357,
    "batch_idx": 160,
    "loss": 0.44531872868537903,
    "learning_rate": 0.0009234217828138102,
    "timestep_mean": 503.15625,
    "epsilon_norm": 8.001143455505371,
    "prediction_norm": 6.060911178588867,
    "timestamp": "2025-07-14T12:33:46.147151"
  },
  {
    "epoch": 357,
    "batch_idx": 192,
    "loss": 0.20772410929203033,
    "learning_rate": 0.0009234217828138102,
    "timestep_mean": 542.75,
    "epsilon_norm": 3.2903811931610107,
    "prediction_norm": 2.7101664543151855,
    "timestamp": "2025-07-14T12:33:46.152046"
  },
  {
    "epoch": 358,
    "batch_idx": 0,
    "loss": 0.393329918384552,
    "learning_rate": 0.0009230035528339209,
    "timestep_mean": 489.78125,
    "epsilon_norm": 8.69731330871582,
    "prediction_norm": 6.694550037384033,
    "timestamp": "2025-07-14T12:33:46.157020"
  },
  {
    "epoch": 358,
    "batch_idx": 32,
    "loss": 0.26382333040237427,
    "learning_rate": 0.0009230035528339209,
    "timestep_mean": 417.46875,
    "epsilon_norm": 7.794112682342529,
    "prediction_norm": 6.025036334991455,
    "timestamp": "2025-07-14T12:33:46.161912"
  },
  {
    "epoch": 358,
    "batch_idx": 64,
    "loss": 0.2937026917934418,
    "learning_rate": 0.0009230035528339209,
    "timestep_mean": 498.0625,
    "epsilon_norm": 8.279033660888672,
    "prediction_norm": 7.261772155761719,
    "timestamp": "2025-07-14T12:33:46.166936"
  },
  {
    "epoch": 358,
    "batch_idx": 96,
    "loss": 0.23430314660072327,
    "learning_rate": 0.0009230035528339209,
    "timestep_mean": 498.53125,
    "epsilon_norm": 7.728147506713867,
    "prediction_norm": 7.201255798339844,
    "timestamp": "2025-07-14T12:33:46.171811"
  },
  {
    "epoch": 358,
    "batch_idx": 128,
    "loss": 0.3297118544578552,
    "learning_rate": 0.0009230035528339209,
    "timestep_mean": 505.75,
    "epsilon_norm": 7.740627288818359,
    "prediction_norm": 6.6688432693481445,
    "timestamp": "2025-07-14T12:33:46.176702"
  },
  {
    "epoch": 358,
    "batch_idx": 160,
    "loss": 0.3262072801589966,
    "learning_rate": 0.0009230035528339209,
    "timestep_mean": 609.09375,
    "epsilon_norm": 8.399806022644043,
    "prediction_norm": 7.488250732421875,
    "timestamp": "2025-07-14T12:33:46.181671"
  },
  {
    "epoch": 358,
    "batch_idx": 192,
    "loss": 0.21335318684577942,
    "learning_rate": 0.0009230035528339209,
    "timestep_mean": 554.5,
    "epsilon_norm": 3.8859987258911133,
    "prediction_norm": 3.5142922401428223,
    "timestamp": "2025-07-14T12:33:46.186557"
  },
  {
    "epoch": 359,
    "batch_idx": 0,
    "loss": 0.43879541754722595,
    "learning_rate": 0.0009225842791348146,
    "timestep_mean": 426.59375,
    "epsilon_norm": 7.489935398101807,
    "prediction_norm": 4.863041400909424,
    "timestamp": "2025-07-14T12:33:46.191533"
  },
  {
    "epoch": 359,
    "batch_idx": 32,
    "loss": 0.2593190371990204,
    "learning_rate": 0.0009225842791348146,
    "timestep_mean": 541.3125,
    "epsilon_norm": 7.9186015129089355,
    "prediction_norm": 6.522952556610107,
    "timestamp": "2025-07-14T12:33:46.196456"
  },
  {
    "epoch": 359,
    "batch_idx": 64,
    "loss": 0.18822745978832245,
    "learning_rate": 0.0009225842791348146,
    "timestep_mean": 521.78125,
    "epsilon_norm": 6.321935653686523,
    "prediction_norm": 5.669466972351074,
    "timestamp": "2025-07-14T12:33:46.201327"
  },
  {
    "epoch": 359,
    "batch_idx": 96,
    "loss": 0.27074065804481506,
    "learning_rate": 0.0009225842791348146,
    "timestep_mean": 552.78125,
    "epsilon_norm": 9.468779563903809,
    "prediction_norm": 8.293654441833496,
    "timestamp": "2025-07-14T12:33:46.206238"
  },
  {
    "epoch": 359,
    "batch_idx": 128,
    "loss": 0.463692843914032,
    "learning_rate": 0.0009225842791348146,
    "timestep_mean": 488.59375,
    "epsilon_norm": 8.897542953491211,
    "prediction_norm": 7.025502681732178,
    "timestamp": "2025-07-14T12:33:46.211439"
  },
  {
    "epoch": 359,
    "batch_idx": 160,
    "loss": 0.29813525080680847,
    "learning_rate": 0.0009225842791348146,
    "timestep_mean": 464.65625,
    "epsilon_norm": 7.853890895843506,
    "prediction_norm": 6.2453765869140625,
    "timestamp": "2025-07-14T12:33:46.216704"
  },
  {
    "epoch": 359,
    "batch_idx": 192,
    "loss": 0.2552018165588379,
    "learning_rate": 0.0009225842791348146,
    "timestep_mean": 547.875,
    "epsilon_norm": 3.035229206085205,
    "prediction_norm": 2.9708759784698486,
    "timestamp": "2025-07-14T12:33:46.221730"
  },
  {
    "epoch": 360,
    "batch_idx": 0,
    "loss": 0.2575872838497162,
    "learning_rate": 0.0009221639627510073,
    "timestep_mean": 498.09375,
    "epsilon_norm": 7.206940650939941,
    "prediction_norm": 5.842960357666016,
    "timestamp": "2025-07-14T12:33:46.226942"
  },
  {
    "epoch": 360,
    "batch_idx": 32,
    "loss": 0.3682326376438141,
    "learning_rate": 0.0009221639627510073,
    "timestep_mean": 431.875,
    "epsilon_norm": 7.461300373077393,
    "prediction_norm": 5.679052829742432,
    "timestamp": "2025-07-14T12:33:46.232138"
  },
  {
    "epoch": 360,
    "batch_idx": 64,
    "loss": 0.23623394966125488,
    "learning_rate": 0.0009221639627510073,
    "timestep_mean": 510.125,
    "epsilon_norm": 6.718584060668945,
    "prediction_norm": 5.708479404449463,
    "timestamp": "2025-07-14T12:33:46.237266"
  },
  {
    "epoch": 360,
    "batch_idx": 96,
    "loss": 0.5502577424049377,
    "learning_rate": 0.0009221639627510073,
    "timestep_mean": 518.78125,
    "epsilon_norm": 7.998533725738525,
    "prediction_norm": 5.812041282653809,
    "timestamp": "2025-07-14T12:33:46.242488"
  },
  {
    "epoch": 360,
    "batch_idx": 128,
    "loss": 0.31708329916000366,
    "learning_rate": 0.0009221639627510073,
    "timestep_mean": 538.5625,
    "epsilon_norm": 7.805743217468262,
    "prediction_norm": 7.029172420501709,
    "timestamp": "2025-07-14T12:33:46.247696"
  },
  {
    "epoch": 360,
    "batch_idx": 160,
    "loss": 0.5652446746826172,
    "learning_rate": 0.0009221639627510073,
    "timestep_mean": 495.71875,
    "epsilon_norm": 8.718881607055664,
    "prediction_norm": 6.67781400680542,
    "timestamp": "2025-07-14T12:33:46.252771"
  },
  {
    "epoch": 360,
    "batch_idx": 192,
    "loss": 0.27753376960754395,
    "learning_rate": 0.0009221639627510073,
    "timestep_mean": 450.875,
    "epsilon_norm": 3.996856212615967,
    "prediction_norm": 3.4814319610595703,
    "timestamp": "2025-07-14T12:33:46.257844"
  },
  {
    "epoch": 361,
    "batch_idx": 0,
    "loss": 0.4538624882698059,
    "learning_rate": 0.0009217426047195881,
    "timestep_mean": 474.75,
    "epsilon_norm": 8.484498023986816,
    "prediction_norm": 6.56024169921875,
    "timestamp": "2025-07-14T12:33:46.263177"
  },
  {
    "epoch": 361,
    "batch_idx": 32,
    "loss": 0.23997290432453156,
    "learning_rate": 0.0009217426047195881,
    "timestep_mean": 490.0625,
    "epsilon_norm": 7.053502559661865,
    "prediction_norm": 5.569026947021484,
    "timestamp": "2025-07-14T12:33:46.268298"
  },
  {
    "epoch": 361,
    "batch_idx": 64,
    "loss": 0.6138637065887451,
    "learning_rate": 0.0009217426047195881,
    "timestep_mean": 432.625,
    "epsilon_norm": 8.446070671081543,
    "prediction_norm": 4.956024169921875,
    "timestamp": "2025-07-14T12:33:46.273543"
  },
  {
    "epoch": 361,
    "batch_idx": 96,
    "loss": 0.2640655040740967,
    "learning_rate": 0.0009217426047195881,
    "timestep_mean": 568.75,
    "epsilon_norm": 8.04747486114502,
    "prediction_norm": 6.899135112762451,
    "timestamp": "2025-07-14T12:33:46.278734"
  },
  {
    "epoch": 361,
    "batch_idx": 128,
    "loss": 0.4876057803630829,
    "learning_rate": 0.0009217426047195881,
    "timestep_mean": 421.0,
    "epsilon_norm": 8.567708969116211,
    "prediction_norm": 6.602552890777588,
    "timestamp": "2025-07-14T12:33:46.283984"
  },
  {
    "epoch": 361,
    "batch_idx": 160,
    "loss": 0.2941967248916626,
    "learning_rate": 0.0009217426047195881,
    "timestep_mean": 495.96875,
    "epsilon_norm": 7.303338050842285,
    "prediction_norm": 5.7605156898498535,
    "timestamp": "2025-07-14T12:33:46.289180"
  },
  {
    "epoch": 361,
    "batch_idx": 192,
    "loss": 0.6198558211326599,
    "learning_rate": 0.0009217426047195881,
    "timestep_mean": 344.125,
    "epsilon_norm": 4.62253999710083,
    "prediction_norm": 4.211756229400635,
    "timestamp": "2025-07-14T12:33:46.294191"
  },
  {
    "epoch": 362,
    "batch_idx": 0,
    "loss": 0.27824628353118896,
    "learning_rate": 0.0009213202060802159,
    "timestep_mean": 539.28125,
    "epsilon_norm": 7.751702308654785,
    "prediction_norm": 6.3112640380859375,
    "timestamp": "2025-07-14T12:33:46.299235"
  },
  {
    "epoch": 362,
    "batch_idx": 32,
    "loss": 0.058660756796598434,
    "learning_rate": 0.0009213202060802159,
    "timestep_mean": 660.03125,
    "epsilon_norm": 8.058781623840332,
    "prediction_norm": 7.916727542877197,
    "timestamp": "2025-07-14T12:33:46.304568"
  },
  {
    "epoch": 362,
    "batch_idx": 64,
    "loss": 0.5228232145309448,
    "learning_rate": 0.0009213202060802159,
    "timestep_mean": 439.25,
    "epsilon_norm": 7.7308349609375,
    "prediction_norm": 5.754303932189941,
    "timestamp": "2025-07-14T12:33:46.309684"
  },
  {
    "epoch": 362,
    "batch_idx": 96,
    "loss": 0.26563379168510437,
    "learning_rate": 0.0009213202060802159,
    "timestep_mean": 535.875,
    "epsilon_norm": 7.985123157501221,
    "prediction_norm": 7.165014266967773,
    "timestamp": "2025-07-14T12:33:46.314847"
  },
  {
    "epoch": 362,
    "batch_idx": 128,
    "loss": 0.26232898235321045,
    "learning_rate": 0.0009213202060802159,
    "timestep_mean": 579.3125,
    "epsilon_norm": 7.895740032196045,
    "prediction_norm": 6.991217613220215,
    "timestamp": "2025-07-14T12:33:46.319988"
  },
  {
    "epoch": 362,
    "batch_idx": 160,
    "loss": 0.23784273862838745,
    "learning_rate": 0.0009213202060802159,
    "timestep_mean": 467.625,
    "epsilon_norm": 7.173421382904053,
    "prediction_norm": 6.041637897491455,
    "timestamp": "2025-07-14T12:33:46.325104"
  },
  {
    "epoch": 362,
    "batch_idx": 192,
    "loss": 0.28740787506103516,
    "learning_rate": 0.0009213202060802159,
    "timestep_mean": 557.25,
    "epsilon_norm": 3.5020883083343506,
    "prediction_norm": 2.998335123062134,
    "timestamp": "2025-07-14T12:33:46.330235"
  },
  {
    "epoch": 363,
    "batch_idx": 0,
    "loss": 0.22273147106170654,
    "learning_rate": 0.0009208967678751174,
    "timestep_mean": 566.46875,
    "epsilon_norm": 7.842491149902344,
    "prediction_norm": 6.987246513366699,
    "timestamp": "2025-07-14T12:33:46.335385"
  },
  {
    "epoch": 363,
    "batch_idx": 32,
    "loss": 0.189549520611763,
    "learning_rate": 0.0009208967678751174,
    "timestep_mean": 626.78125,
    "epsilon_norm": 8.043227195739746,
    "prediction_norm": 6.9600324630737305,
    "timestamp": "2025-07-14T12:33:46.340599"
  },
  {
    "epoch": 363,
    "batch_idx": 64,
    "loss": 0.3506842255592346,
    "learning_rate": 0.0009208967678751174,
    "timestep_mean": 499.46875,
    "epsilon_norm": 8.615832328796387,
    "prediction_norm": 7.234225273132324,
    "timestamp": "2025-07-14T12:33:46.345750"
  },
  {
    "epoch": 363,
    "batch_idx": 96,
    "loss": 0.414409875869751,
    "learning_rate": 0.0009208967678751174,
    "timestep_mean": 478.71875,
    "epsilon_norm": 7.760412216186523,
    "prediction_norm": 5.758901119232178,
    "timestamp": "2025-07-14T12:33:46.350891"
  },
  {
    "epoch": 363,
    "batch_idx": 128,
    "loss": 0.4236958920955658,
    "learning_rate": 0.0009208967678751174,
    "timestep_mean": 401.3125,
    "epsilon_norm": 7.618948936462402,
    "prediction_norm": 5.916797161102295,
    "timestamp": "2025-07-14T12:33:46.356068"
  },
  {
    "epoch": 363,
    "batch_idx": 160,
    "loss": 0.33705735206604004,
    "learning_rate": 0.0009208967678751174,
    "timestep_mean": 441.84375,
    "epsilon_norm": 7.241689682006836,
    "prediction_norm": 5.387757301330566,
    "timestamp": "2025-07-14T12:33:46.361234"
  },
  {
    "epoch": 363,
    "batch_idx": 192,
    "loss": 1.294116735458374,
    "learning_rate": 0.0009208967678751174,
    "timestep_mean": 373.875,
    "epsilon_norm": 5.17889928817749,
    "prediction_norm": 2.3890955448150635,
    "timestamp": "2025-07-14T12:33:46.366497"
  },
  {
    "epoch": 364,
    "batch_idx": 0,
    "loss": 0.24714751541614532,
    "learning_rate": 0.0009204722911490842,
    "timestep_mean": 554.03125,
    "epsilon_norm": 7.925562381744385,
    "prediction_norm": 6.013763427734375,
    "timestamp": "2025-07-14T12:33:46.371738"
  },
  {
    "epoch": 364,
    "batch_idx": 32,
    "loss": 0.3386995196342468,
    "learning_rate": 0.0009204722911490842,
    "timestep_mean": 470.53125,
    "epsilon_norm": 7.753925323486328,
    "prediction_norm": 5.7094550132751465,
    "timestamp": "2025-07-14T12:33:46.376958"
  },
  {
    "epoch": 364,
    "batch_idx": 64,
    "loss": 0.40718597173690796,
    "learning_rate": 0.0009204722911490842,
    "timestep_mean": 449.65625,
    "epsilon_norm": 7.576210975646973,
    "prediction_norm": 5.262523174285889,
    "timestamp": "2025-07-14T12:33:46.382096"
  },
  {
    "epoch": 364,
    "batch_idx": 96,
    "loss": 0.3490986227989197,
    "learning_rate": 0.0009204722911490842,
    "timestep_mean": 502.90625,
    "epsilon_norm": 7.866333961486816,
    "prediction_norm": 6.09950065612793,
    "timestamp": "2025-07-14T12:33:46.387169"
  },
  {
    "epoch": 364,
    "batch_idx": 128,
    "loss": 0.2128659188747406,
    "learning_rate": 0.0009204722911490842,
    "timestep_mean": 447.90625,
    "epsilon_norm": 7.367937088012695,
    "prediction_norm": 6.572241306304932,
    "timestamp": "2025-07-14T12:33:46.392326"
  },
  {
    "epoch": 364,
    "batch_idx": 160,
    "loss": 0.38569626212120056,
    "learning_rate": 0.0009204722911490842,
    "timestep_mean": 472.46875,
    "epsilon_norm": 7.89537239074707,
    "prediction_norm": 6.436363220214844,
    "timestamp": "2025-07-14T12:33:46.397447"
  },
  {
    "epoch": 364,
    "batch_idx": 192,
    "loss": 0.1507037729024887,
    "learning_rate": 0.0009204722911490842,
    "timestep_mean": 414.375,
    "epsilon_norm": 3.609513521194458,
    "prediction_norm": 3.1003870964050293,
    "timestamp": "2025-07-14T12:33:46.402643"
  },
  {
    "epoch": 365,
    "batch_idx": 0,
    "loss": 0.27625924348831177,
    "learning_rate": 0.0009200467769494705,
    "timestep_mean": 473.9375,
    "epsilon_norm": 8.116816520690918,
    "prediction_norm": 7.024892330169678,
    "timestamp": "2025-07-14T12:33:46.407631"
  },
  {
    "epoch": 365,
    "batch_idx": 32,
    "loss": 0.34472447633743286,
    "learning_rate": 0.0009200467769494705,
    "timestep_mean": 466.375,
    "epsilon_norm": 7.580371379852295,
    "prediction_norm": 6.2576117515563965,
    "timestamp": "2025-07-14T12:33:46.412540"
  },
  {
    "epoch": 365,
    "batch_idx": 64,
    "loss": 0.3217446804046631,
    "learning_rate": 0.0009200467769494705,
    "timestep_mean": 565.53125,
    "epsilon_norm": 8.172175407409668,
    "prediction_norm": 6.800981044769287,
    "timestamp": "2025-07-14T12:33:46.417836"
  },
  {
    "epoch": 365,
    "batch_idx": 96,
    "loss": 0.47871267795562744,
    "learning_rate": 0.0009200467769494705,
    "timestep_mean": 444.03125,
    "epsilon_norm": 8.645017623901367,
    "prediction_norm": 7.045979022979736,
    "timestamp": "2025-07-14T12:33:46.423149"
  },
  {
    "epoch": 365,
    "batch_idx": 128,
    "loss": 0.5104832053184509,
    "learning_rate": 0.0009200467769494705,
    "timestep_mean": 495.65625,
    "epsilon_norm": 8.569397926330566,
    "prediction_norm": 6.678182601928711,
    "timestamp": "2025-07-14T12:33:46.428342"
  },
  {
    "epoch": 365,
    "batch_idx": 160,
    "loss": 0.388873428106308,
    "learning_rate": 0.0009200467769494705,
    "timestep_mean": 471.3125,
    "epsilon_norm": 7.431327819824219,
    "prediction_norm": 6.27978515625,
    "timestamp": "2025-07-14T12:33:46.433620"
  },
  {
    "epoch": 365,
    "batch_idx": 192,
    "loss": 0.6959378719329834,
    "learning_rate": 0.0009200467769494705,
    "timestep_mean": 387.75,
    "epsilon_norm": 4.415450096130371,
    "prediction_norm": 2.9211223125457764,
    "timestamp": "2025-07-14T12:33:46.438885"
  },
  {
    "epoch": 366,
    "batch_idx": 0,
    "loss": 0.2893417477607727,
    "learning_rate": 0.0009196202263261903,
    "timestep_mean": 491.84375,
    "epsilon_norm": 7.448940753936768,
    "prediction_norm": 6.31883430480957,
    "timestamp": "2025-07-14T12:33:46.444101"
  },
  {
    "epoch": 366,
    "batch_idx": 32,
    "loss": 0.320695161819458,
    "learning_rate": 0.0009196202263261903,
    "timestep_mean": 506.0,
    "epsilon_norm": 8.742114067077637,
    "prediction_norm": 7.098962783813477,
    "timestamp": "2025-07-14T12:33:46.449331"
  },
  {
    "epoch": 366,
    "batch_idx": 64,
    "loss": 0.27999475598335266,
    "learning_rate": 0.0009196202263261903,
    "timestep_mean": 510.375,
    "epsilon_norm": 7.843890190124512,
    "prediction_norm": 6.322033405303955,
    "timestamp": "2025-07-14T12:33:46.454693"
  },
  {
    "epoch": 366,
    "batch_idx": 96,
    "loss": 0.17455092072486877,
    "learning_rate": 0.0009196202263261903,
    "timestep_mean": 602.6875,
    "epsilon_norm": 8.803771018981934,
    "prediction_norm": 8.052502632141113,
    "timestamp": "2025-07-14T12:33:46.459763"
  },
  {
    "epoch": 366,
    "batch_idx": 128,
    "loss": 0.471916526556015,
    "learning_rate": 0.0009196202263261903,
    "timestep_mean": 440.0625,
    "epsilon_norm": 7.864318370819092,
    "prediction_norm": 5.538044452667236,
    "timestamp": "2025-07-14T12:33:46.464960"
  },
  {
    "epoch": 366,
    "batch_idx": 160,
    "loss": 0.26463988423347473,
    "learning_rate": 0.0009196202263261903,
    "timestep_mean": 516.0,
    "epsilon_norm": 7.808857440948486,
    "prediction_norm": 7.333944797515869,
    "timestamp": "2025-07-14T12:33:46.470168"
  },
  {
    "epoch": 366,
    "batch_idx": 192,
    "loss": 0.48025405406951904,
    "learning_rate": 0.0009196202263261903,
    "timestep_mean": 373.875,
    "epsilon_norm": 3.798550605773926,
    "prediction_norm": 2.1158156394958496,
    "timestamp": "2025-07-14T12:33:46.475913"
  },
  {
    "epoch": 367,
    "batch_idx": 0,
    "loss": 0.48467081785202026,
    "learning_rate": 0.0009191926403317149,
    "timestep_mean": 502.0625,
    "epsilon_norm": 8.697918891906738,
    "prediction_norm": 6.352405071258545,
    "timestamp": "2025-07-14T12:33:46.481544"
  },
  {
    "epoch": 367,
    "batch_idx": 32,
    "loss": 0.39774614572525024,
    "learning_rate": 0.0009191926403317149,
    "timestep_mean": 439.34375,
    "epsilon_norm": 8.652934074401855,
    "prediction_norm": 6.753415584564209,
    "timestamp": "2025-07-14T12:33:46.486871"
  },
  {
    "epoch": 367,
    "batch_idx": 64,
    "loss": 0.19382889568805695,
    "learning_rate": 0.0009191926403317149,
    "timestep_mean": 519.5,
    "epsilon_norm": 7.248839378356934,
    "prediction_norm": 6.233659267425537,
    "timestamp": "2025-07-14T12:33:46.492330"
  },
  {
    "epoch": 367,
    "batch_idx": 96,
    "loss": 0.31074434518814087,
    "learning_rate": 0.0009191926403317149,
    "timestep_mean": 484.0625,
    "epsilon_norm": 7.993813991546631,
    "prediction_norm": 6.626798152923584,
    "timestamp": "2025-07-14T12:33:46.497535"
  },
  {
    "epoch": 367,
    "batch_idx": 128,
    "loss": 0.3658154308795929,
    "learning_rate": 0.0009191926403317149,
    "timestep_mean": 422.28125,
    "epsilon_norm": 7.99564790725708,
    "prediction_norm": 6.322828769683838,
    "timestamp": "2025-07-14T12:33:46.502502"
  },
  {
    "epoch": 367,
    "batch_idx": 160,
    "loss": 0.434065580368042,
    "learning_rate": 0.0009191926403317149,
    "timestep_mean": 534.875,
    "epsilon_norm": 8.654597282409668,
    "prediction_norm": 6.464230060577393,
    "timestamp": "2025-07-14T12:33:46.509895"
  },
  {
    "epoch": 367,
    "batch_idx": 192,
    "loss": 0.1643097698688507,
    "learning_rate": 0.0009191926403317149,
    "timestep_mean": 472.5,
    "epsilon_norm": 3.0168802738189697,
    "prediction_norm": 3.0146849155426025,
    "timestamp": "2025-07-14T12:33:46.514942"
  },
  {
    "epoch": 368,
    "batch_idx": 0,
    "loss": 0.3609286844730377,
    "learning_rate": 0.0009187640200210704,
    "timestep_mean": 447.71875,
    "epsilon_norm": 7.886796951293945,
    "prediction_norm": 5.983949661254883,
    "timestamp": "2025-07-14T12:33:46.519987"
  },
  {
    "epoch": 368,
    "batch_idx": 32,
    "loss": 0.2416708767414093,
    "learning_rate": 0.0009187640200210704,
    "timestep_mean": 510.71875,
    "epsilon_norm": 7.801915168762207,
    "prediction_norm": 7.029874324798584,
    "timestamp": "2025-07-14T12:33:46.524933"
  },
  {
    "epoch": 368,
    "batch_idx": 64,
    "loss": 0.13584229350090027,
    "learning_rate": 0.0009187640200210704,
    "timestep_mean": 535.09375,
    "epsilon_norm": 7.657482147216797,
    "prediction_norm": 7.099218368530273,
    "timestamp": "2025-07-14T12:33:46.529898"
  },
  {
    "epoch": 368,
    "batch_idx": 96,
    "loss": 0.14788773655891418,
    "learning_rate": 0.0009187640200210704,
    "timestep_mean": 522.59375,
    "epsilon_norm": 8.167823791503906,
    "prediction_norm": 7.279317855834961,
    "timestamp": "2025-07-14T12:33:46.534854"
  },
  {
    "epoch": 368,
    "batch_idx": 128,
    "loss": 0.3130873441696167,
    "learning_rate": 0.0009187640200210704,
    "timestep_mean": 486.84375,
    "epsilon_norm": 7.404679298400879,
    "prediction_norm": 5.919886589050293,
    "timestamp": "2025-07-14T12:33:46.539874"
  },
  {
    "epoch": 368,
    "batch_idx": 160,
    "loss": 0.13667160272598267,
    "learning_rate": 0.0009187640200210704,
    "timestep_mean": 515.21875,
    "epsilon_norm": 6.129954814910889,
    "prediction_norm": 5.576681137084961,
    "timestamp": "2025-07-14T12:33:46.544796"
  },
  {
    "epoch": 368,
    "batch_idx": 192,
    "loss": 0.7736369371414185,
    "learning_rate": 0.0009187640200210704,
    "timestep_mean": 545.125,
    "epsilon_norm": 4.0547895431518555,
    "prediction_norm": 1.9900962114334106,
    "timestamp": "2025-07-14T12:33:46.549792"
  },
  {
    "epoch": 369,
    "batch_idx": 0,
    "loss": 0.5206841230392456,
    "learning_rate": 0.0009183343664518343,
    "timestep_mean": 491.53125,
    "epsilon_norm": 8.022323608398438,
    "prediction_norm": 5.270209312438965,
    "timestamp": "2025-07-14T12:33:46.554648"
  },
  {
    "epoch": 369,
    "batch_idx": 32,
    "loss": 0.23908483982086182,
    "learning_rate": 0.0009183343664518343,
    "timestep_mean": 514.46875,
    "epsilon_norm": 7.005189418792725,
    "prediction_norm": 6.2397236824035645,
    "timestamp": "2025-07-14T12:33:46.559610"
  },
  {
    "epoch": 369,
    "batch_idx": 64,
    "loss": 0.25536221265792847,
    "learning_rate": 0.0009183343664518343,
    "timestep_mean": 523.90625,
    "epsilon_norm": 7.945797443389893,
    "prediction_norm": 7.186140537261963,
    "timestamp": "2025-07-14T12:33:46.564601"
  },
  {
    "epoch": 369,
    "batch_idx": 96,
    "loss": 0.3620924651622772,
    "learning_rate": 0.0009183343664518343,
    "timestep_mean": 480.5625,
    "epsilon_norm": 8.660910606384277,
    "prediction_norm": 7.021625995635986,
    "timestamp": "2025-07-14T12:33:46.569615"
  },
  {
    "epoch": 369,
    "batch_idx": 128,
    "loss": 0.33279144763946533,
    "learning_rate": 0.0009183343664518343,
    "timestep_mean": 530.53125,
    "epsilon_norm": 8.481466293334961,
    "prediction_norm": 6.846768856048584,
    "timestamp": "2025-07-14T12:33:46.574849"
  },
  {
    "epoch": 369,
    "batch_idx": 160,
    "loss": 0.5367790460586548,
    "learning_rate": 0.0009183343664518343,
    "timestep_mean": 486.75,
    "epsilon_norm": 9.422301292419434,
    "prediction_norm": 6.9351301193237305,
    "timestamp": "2025-07-14T12:33:46.579885"
  },
  {
    "epoch": 369,
    "batch_idx": 192,
    "loss": 0.33004581928253174,
    "learning_rate": 0.0009183343664518343,
    "timestep_mean": 512.0,
    "epsilon_norm": 5.213177680969238,
    "prediction_norm": 4.485701560974121,
    "timestamp": "2025-07-14T12:33:46.585028"
  },
  {
    "epoch": 370,
    "batch_idx": 0,
    "loss": 0.23606623709201813,
    "learning_rate": 0.0009179036806841346,
    "timestep_mean": 590.71875,
    "epsilon_norm": 7.569699764251709,
    "prediction_norm": 6.2594380378723145,
    "timestamp": "2025-07-14T12:33:46.590013"
  },
  {
    "epoch": 370,
    "batch_idx": 32,
    "loss": 0.26788604259490967,
    "learning_rate": 0.0009179036806841346,
    "timestep_mean": 574.46875,
    "epsilon_norm": 7.172449588775635,
    "prediction_norm": 6.4140143394470215,
    "timestamp": "2025-07-14T12:33:46.595039"
  },
  {
    "epoch": 370,
    "batch_idx": 64,
    "loss": 0.24610167741775513,
    "learning_rate": 0.0009179036806841346,
    "timestep_mean": 576.71875,
    "epsilon_norm": 7.570621013641357,
    "prediction_norm": 6.803798198699951,
    "timestamp": "2025-07-14T12:33:46.600099"
  },
  {
    "epoch": 370,
    "batch_idx": 96,
    "loss": 0.32637640833854675,
    "learning_rate": 0.0009179036806841346,
    "timestep_mean": 509.0625,
    "epsilon_norm": 8.51824951171875,
    "prediction_norm": 7.286000728607178,
    "timestamp": "2025-07-14T12:33:46.605071"
  },
  {
    "epoch": 370,
    "batch_idx": 128,
    "loss": 0.26658016443252563,
    "learning_rate": 0.0009179036806841346,
    "timestep_mean": 577.375,
    "epsilon_norm": 6.832314968109131,
    "prediction_norm": 6.466300010681152,
    "timestamp": "2025-07-14T12:33:46.610037"
  },
  {
    "epoch": 370,
    "batch_idx": 160,
    "loss": 0.28628844022750854,
    "learning_rate": 0.0009179036806841346,
    "timestep_mean": 449.21875,
    "epsilon_norm": 8.582353591918945,
    "prediction_norm": 7.71307373046875,
    "timestamp": "2025-07-14T12:33:46.615083"
  },
  {
    "epoch": 370,
    "batch_idx": 192,
    "loss": 0.41694873571395874,
    "learning_rate": 0.0009179036806841346,
    "timestep_mean": 469.5,
    "epsilon_norm": 3.5242345333099365,
    "prediction_norm": 3.3905866146087646,
    "timestamp": "2025-07-14T12:33:46.620494"
  },
  {
    "epoch": 371,
    "batch_idx": 0,
    "loss": 0.2879617214202881,
    "learning_rate": 0.0009174719637806454,
    "timestep_mean": 484.0625,
    "epsilon_norm": 7.927184581756592,
    "prediction_norm": 6.741091251373291,
    "timestamp": "2025-07-14T12:33:46.625726"
  },
  {
    "epoch": 371,
    "batch_idx": 32,
    "loss": 0.3236260414123535,
    "learning_rate": 0.0009174719637806454,
    "timestep_mean": 465.46875,
    "epsilon_norm": 6.78944730758667,
    "prediction_norm": 5.277744770050049,
    "timestamp": "2025-07-14T12:33:46.630836"
  },
  {
    "epoch": 371,
    "batch_idx": 64,
    "loss": 0.14407619833946228,
    "learning_rate": 0.0009174719637806454,
    "timestep_mean": 537.375,
    "epsilon_norm": 7.724420547485352,
    "prediction_norm": 7.095118045806885,
    "timestamp": "2025-07-14T12:33:46.635942"
  },
  {
    "epoch": 371,
    "batch_idx": 96,
    "loss": 0.41370683908462524,
    "learning_rate": 0.0009174719637806454,
    "timestep_mean": 525.40625,
    "epsilon_norm": 7.797347068786621,
    "prediction_norm": 6.18491268157959,
    "timestamp": "2025-07-14T12:33:46.640911"
  },
  {
    "epoch": 371,
    "batch_idx": 128,
    "loss": 0.3799532651901245,
    "learning_rate": 0.0009174719637806454,
    "timestep_mean": 492.71875,
    "epsilon_norm": 8.010735511779785,
    "prediction_norm": 6.1236982345581055,
    "timestamp": "2025-07-14T12:33:46.646235"
  },
  {
    "epoch": 371,
    "batch_idx": 160,
    "loss": 0.18768900632858276,
    "learning_rate": 0.0009174719637806454,
    "timestep_mean": 565.25,
    "epsilon_norm": 7.498779773712158,
    "prediction_norm": 6.180719375610352,
    "timestamp": "2025-07-14T12:33:46.651081"
  },
  {
    "epoch": 371,
    "batch_idx": 192,
    "loss": 0.6028737425804138,
    "learning_rate": 0.0009174719637806454,
    "timestep_mean": 419.875,
    "epsilon_norm": 4.227015972137451,
    "prediction_norm": 2.6521453857421875,
    "timestamp": "2025-07-14T12:33:46.656157"
  },
  {
    "epoch": 372,
    "batch_idx": 0,
    "loss": 0.4731791317462921,
    "learning_rate": 0.0009170392168065851,
    "timestep_mean": 440.21875,
    "epsilon_norm": 7.813991546630859,
    "prediction_norm": 5.522091388702393,
    "timestamp": "2025-07-14T12:33:46.661360"
  },
  {
    "epoch": 372,
    "batch_idx": 32,
    "loss": 0.19883838295936584,
    "learning_rate": 0.0009170392168065851,
    "timestep_mean": 444.75,
    "epsilon_norm": 7.008399486541748,
    "prediction_norm": 5.659086227416992,
    "timestamp": "2025-07-14T12:33:46.666715"
  },
  {
    "epoch": 372,
    "batch_idx": 64,
    "loss": 0.2304278016090393,
    "learning_rate": 0.0009170392168065851,
    "timestep_mean": 407.15625,
    "epsilon_norm": 7.033030986785889,
    "prediction_norm": 5.30418062210083,
    "timestamp": "2025-07-14T12:33:46.671876"
  },
  {
    "epoch": 372,
    "batch_idx": 96,
    "loss": 0.3745972514152527,
    "learning_rate": 0.0009170392168065851,
    "timestep_mean": 473.0625,
    "epsilon_norm": 8.379281997680664,
    "prediction_norm": 6.579376697540283,
    "timestamp": "2025-07-14T12:33:46.676811"
  },
  {
    "epoch": 372,
    "batch_idx": 128,
    "loss": 0.24509909749031067,
    "learning_rate": 0.0009170392168065851,
    "timestep_mean": 513.84375,
    "epsilon_norm": 7.693875789642334,
    "prediction_norm": 6.336897373199463,
    "timestamp": "2025-07-14T12:33:46.681656"
  },
  {
    "epoch": 372,
    "batch_idx": 160,
    "loss": 0.3724125325679779,
    "learning_rate": 0.0009170392168065851,
    "timestep_mean": 493.65625,
    "epsilon_norm": 7.878705978393555,
    "prediction_norm": 5.926409721374512,
    "timestamp": "2025-07-14T12:33:46.686700"
  },
  {
    "epoch": 372,
    "batch_idx": 192,
    "loss": 0.2558761537075043,
    "learning_rate": 0.0009170392168065851,
    "timestep_mean": 507.5,
    "epsilon_norm": 5.300971031188965,
    "prediction_norm": 4.781431674957275,
    "timestamp": "2025-07-14T12:33:46.691992"
  },
  {
    "epoch": 373,
    "batch_idx": 0,
    "loss": 0.3169274926185608,
    "learning_rate": 0.000916605440829714,
    "timestep_mean": 485.1875,
    "epsilon_norm": 8.258638381958008,
    "prediction_norm": 6.314354419708252,
    "timestamp": "2025-07-14T12:33:46.697238"
  },
  {
    "epoch": 373,
    "batch_idx": 32,
    "loss": 0.37926238775253296,
    "learning_rate": 0.000916605440829714,
    "timestep_mean": 467.0625,
    "epsilon_norm": 7.600398540496826,
    "prediction_norm": 5.845625877380371,
    "timestamp": "2025-07-14T12:33:46.702505"
  },
  {
    "epoch": 373,
    "batch_idx": 64,
    "loss": 0.24033412337303162,
    "learning_rate": 0.000916605440829714,
    "timestep_mean": 513.15625,
    "epsilon_norm": 7.582303524017334,
    "prediction_norm": 6.348523139953613,
    "timestamp": "2025-07-14T12:33:46.707684"
  },
  {
    "epoch": 373,
    "batch_idx": 96,
    "loss": 0.3931902050971985,
    "learning_rate": 0.000916605440829714,
    "timestep_mean": 474.78125,
    "epsilon_norm": 7.766665935516357,
    "prediction_norm": 5.762588024139404,
    "timestamp": "2025-07-14T12:33:46.713029"
  },
  {
    "epoch": 373,
    "batch_idx": 128,
    "loss": 0.2842944860458374,
    "learning_rate": 0.000916605440829714,
    "timestep_mean": 487.375,
    "epsilon_norm": 8.413616180419922,
    "prediction_norm": 6.342679977416992,
    "timestamp": "2025-07-14T12:33:46.718304"
  },
  {
    "epoch": 373,
    "batch_idx": 160,
    "loss": 0.2851971387863159,
    "learning_rate": 0.000916605440829714,
    "timestep_mean": 452.59375,
    "epsilon_norm": 8.503626823425293,
    "prediction_norm": 6.881727695465088,
    "timestamp": "2025-07-14T12:33:46.723522"
  },
  {
    "epoch": 373,
    "batch_idx": 192,
    "loss": 0.2521740794181824,
    "learning_rate": 0.000916605440829714,
    "timestep_mean": 425.25,
    "epsilon_norm": 3.025357484817505,
    "prediction_norm": 1.933433175086975,
    "timestamp": "2025-07-14T12:33:46.728691"
  },
  {
    "epoch": 374,
    "batch_idx": 0,
    "loss": 0.2510175108909607,
    "learning_rate": 0.0009161706369203313,
    "timestep_mean": 545.25,
    "epsilon_norm": 7.675294399261475,
    "prediction_norm": 7.265194416046143,
    "timestamp": "2025-07-14T12:33:46.733853"
  },
  {
    "epoch": 374,
    "batch_idx": 32,
    "loss": 0.15005701780319214,
    "learning_rate": 0.0009161706369203313,
    "timestep_mean": 514.53125,
    "epsilon_norm": 7.545353889465332,
    "prediction_norm": 7.179527759552002,
    "timestamp": "2025-07-14T12:33:46.739074"
  },
  {
    "epoch": 374,
    "batch_idx": 64,
    "loss": 0.2675095796585083,
    "learning_rate": 0.0009161706369203313,
    "timestep_mean": 533.96875,
    "epsilon_norm": 7.204530715942383,
    "prediction_norm": 6.69991397857666,
    "timestamp": "2025-07-14T12:33:46.744273"
  },
  {
    "epoch": 374,
    "batch_idx": 96,
    "loss": 0.6236420273780823,
    "learning_rate": 0.0009161706369203313,
    "timestep_mean": 375.0625,
    "epsilon_norm": 8.387904167175293,
    "prediction_norm": 6.636289596557617,
    "timestamp": "2025-07-14T12:33:46.749429"
  },
  {
    "epoch": 374,
    "batch_idx": 128,
    "loss": 0.2123676985502243,
    "learning_rate": 0.0009161706369203313,
    "timestep_mean": 507.96875,
    "epsilon_norm": 8.404473304748535,
    "prediction_norm": 8.30901050567627,
    "timestamp": "2025-07-14T12:33:46.754714"
  },
  {
    "epoch": 374,
    "batch_idx": 160,
    "loss": 0.3458769917488098,
    "learning_rate": 0.0009161706369203313,
    "timestep_mean": 458.15625,
    "epsilon_norm": 8.2361421585083,
    "prediction_norm": 7.835917949676514,
    "timestamp": "2025-07-14T12:33:46.759979"
  },
  {
    "epoch": 374,
    "batch_idx": 192,
    "loss": 0.18183469772338867,
    "learning_rate": 0.0009161706369203313,
    "timestep_mean": 679.5,
    "epsilon_norm": 3.8830156326293945,
    "prediction_norm": 3.697998046875,
    "timestamp": "2025-07-14T12:33:46.765043"
  },
  {
    "epoch": 375,
    "batch_idx": 0,
    "loss": 0.3163454532623291,
    "learning_rate": 0.0009157348061512722,
    "timestep_mean": 591.375,
    "epsilon_norm": 8.864786148071289,
    "prediction_norm": 7.936357021331787,
    "timestamp": "2025-07-14T12:33:46.770265"
  },
  {
    "epoch": 375,
    "batch_idx": 32,
    "loss": 0.31805768609046936,
    "learning_rate": 0.0009157348061512722,
    "timestep_mean": 486.0,
    "epsilon_norm": 7.752194404602051,
    "prediction_norm": 6.484238624572754,
    "timestamp": "2025-07-14T12:33:46.775544"
  },
  {
    "epoch": 375,
    "batch_idx": 64,
    "loss": 0.3241788148880005,
    "learning_rate": 0.0009157348061512722,
    "timestep_mean": 561.5,
    "epsilon_norm": 8.536572456359863,
    "prediction_norm": 6.946804046630859,
    "timestamp": "2025-07-14T12:33:46.780665"
  },
  {
    "epoch": 375,
    "batch_idx": 96,
    "loss": 0.3840904235839844,
    "learning_rate": 0.0009157348061512722,
    "timestep_mean": 525.71875,
    "epsilon_norm": 8.920877456665039,
    "prediction_norm": 7.635203838348389,
    "timestamp": "2025-07-14T12:33:46.785987"
  },
  {
    "epoch": 375,
    "batch_idx": 128,
    "loss": 0.37024056911468506,
    "learning_rate": 0.0009157348061512722,
    "timestep_mean": 505.34375,
    "epsilon_norm": 8.757864952087402,
    "prediction_norm": 6.36260986328125,
    "timestamp": "2025-07-14T12:33:46.791211"
  },
  {
    "epoch": 375,
    "batch_idx": 160,
    "loss": 0.3682704269886017,
    "learning_rate": 0.0009157348061512722,
    "timestep_mean": 525.1875,
    "epsilon_norm": 8.897773742675781,
    "prediction_norm": 6.460442543029785,
    "timestamp": "2025-07-14T12:33:46.796533"
  },
  {
    "epoch": 375,
    "batch_idx": 192,
    "loss": 0.3225111663341522,
    "learning_rate": 0.0009157348061512722,
    "timestep_mean": 343.0,
    "epsilon_norm": 3.261033296585083,
    "prediction_norm": 2.2463881969451904,
    "timestamp": "2025-07-14T12:33:46.801725"
  },
  {
    "epoch": 376,
    "batch_idx": 0,
    "loss": 0.2722902297973633,
    "learning_rate": 0.0009152979495979059,
    "timestep_mean": 484.25,
    "epsilon_norm": 7.733835220336914,
    "prediction_norm": 6.603818416595459,
    "timestamp": "2025-07-14T12:33:46.806866"
  },
  {
    "epoch": 376,
    "batch_idx": 32,
    "loss": 0.2134196162223816,
    "learning_rate": 0.0009152979495979059,
    "timestep_mean": 515.09375,
    "epsilon_norm": 6.898220539093018,
    "prediction_norm": 5.699009418487549,
    "timestamp": "2025-07-14T12:33:46.811870"
  },
  {
    "epoch": 376,
    "batch_idx": 64,
    "loss": 0.3665927052497864,
    "learning_rate": 0.0009152979495979059,
    "timestep_mean": 454.40625,
    "epsilon_norm": 7.551220893859863,
    "prediction_norm": 5.335820198059082,
    "timestamp": "2025-07-14T12:33:46.817058"
  },
  {
    "epoch": 376,
    "batch_idx": 96,
    "loss": 0.4314733147621155,
    "learning_rate": 0.0009152979495979059,
    "timestep_mean": 460.15625,
    "epsilon_norm": 9.551142692565918,
    "prediction_norm": 7.531399726867676,
    "timestamp": "2025-07-14T12:33:46.822223"
  },
  {
    "epoch": 376,
    "batch_idx": 128,
    "loss": 0.30053064227104187,
    "learning_rate": 0.0009152979495979059,
    "timestep_mean": 546.8125,
    "epsilon_norm": 8.135720252990723,
    "prediction_norm": 6.9120049476623535,
    "timestamp": "2025-07-14T12:33:46.827404"
  },
  {
    "epoch": 376,
    "batch_idx": 160,
    "loss": 0.38662102818489075,
    "learning_rate": 0.0009152979495979059,
    "timestep_mean": 446.4375,
    "epsilon_norm": 9.413142204284668,
    "prediction_norm": 7.920718669891357,
    "timestamp": "2025-07-14T12:33:46.832617"
  },
  {
    "epoch": 376,
    "batch_idx": 192,
    "loss": 0.9840326905250549,
    "learning_rate": 0.0009152979495979059,
    "timestep_mean": 440.125,
    "epsilon_norm": 5.280327796936035,
    "prediction_norm": 2.536445140838623,
    "timestamp": "2025-07-14T12:33:46.837916"
  },
  {
    "epoch": 377,
    "batch_idx": 0,
    "loss": 0.2875902056694031,
    "learning_rate": 0.0009148600683381324,
    "timestep_mean": 476.4375,
    "epsilon_norm": 7.859764099121094,
    "prediction_norm": 6.464170932769775,
    "timestamp": "2025-07-14T12:33:46.843130"
  },
  {
    "epoch": 377,
    "batch_idx": 32,
    "loss": 0.3955591320991516,
    "learning_rate": 0.0009148600683381324,
    "timestep_mean": 498.375,
    "epsilon_norm": 8.394508361816406,
    "prediction_norm": 6.3588385581970215,
    "timestamp": "2025-07-14T12:33:46.848315"
  },
  {
    "epoch": 377,
    "batch_idx": 64,
    "loss": 0.7329883575439453,
    "learning_rate": 0.0009148600683381324,
    "timestep_mean": 410.40625,
    "epsilon_norm": 8.751277923583984,
    "prediction_norm": 6.6193718910217285,
    "timestamp": "2025-07-14T12:33:46.853542"
  },
  {
    "epoch": 377,
    "batch_idx": 96,
    "loss": 0.4601222574710846,
    "learning_rate": 0.0009148600683381324,
    "timestep_mean": 472.625,
    "epsilon_norm": 8.679444313049316,
    "prediction_norm": 7.07977819442749,
    "timestamp": "2025-07-14T12:33:46.858734"
  },
  {
    "epoch": 377,
    "batch_idx": 128,
    "loss": 0.3925711512565613,
    "learning_rate": 0.0009148600683381324,
    "timestep_mean": 368.09375,
    "epsilon_norm": 7.5134148597717285,
    "prediction_norm": 5.789811134338379,
    "timestamp": "2025-07-14T12:33:46.863940"
  },
  {
    "epoch": 377,
    "batch_idx": 160,
    "loss": 0.4136161804199219,
    "learning_rate": 0.0009148600683381324,
    "timestep_mean": 416.03125,
    "epsilon_norm": 9.03235149383545,
    "prediction_norm": 7.680203437805176,
    "timestamp": "2025-07-14T12:33:46.869172"
  },
  {
    "epoch": 377,
    "batch_idx": 192,
    "loss": 0.12091659009456635,
    "learning_rate": 0.0009148600683381324,
    "timestep_mean": 642.25,
    "epsilon_norm": 5.066678524017334,
    "prediction_norm": 4.825183868408203,
    "timestamp": "2025-07-14T12:33:46.874246"
  },
  {
    "epoch": 378,
    "batch_idx": 0,
    "loss": 0.2808122932910919,
    "learning_rate": 0.0009144211634523803,
    "timestep_mean": 552.125,
    "epsilon_norm": 9.228167533874512,
    "prediction_norm": 8.713068008422852,
    "timestamp": "2025-07-14T12:33:46.879509"
  },
  {
    "epoch": 378,
    "batch_idx": 32,
    "loss": 0.2314492166042328,
    "learning_rate": 0.0009144211634523803,
    "timestep_mean": 497.71875,
    "epsilon_norm": 7.282073497772217,
    "prediction_norm": 5.906676769256592,
    "timestamp": "2025-07-14T12:33:46.884717"
  },
  {
    "epoch": 378,
    "batch_idx": 64,
    "loss": 0.31797054409980774,
    "learning_rate": 0.0009144211634523803,
    "timestep_mean": 491.0625,
    "epsilon_norm": 7.870544910430908,
    "prediction_norm": 6.564328670501709,
    "timestamp": "2025-07-14T12:33:46.889776"
  },
  {
    "epoch": 378,
    "batch_idx": 96,
    "loss": 0.3535101115703583,
    "learning_rate": 0.0009144211634523803,
    "timestep_mean": 518.9375,
    "epsilon_norm": 7.571077346801758,
    "prediction_norm": 6.747438907623291,
    "timestamp": "2025-07-14T12:33:46.895127"
  },
  {
    "epoch": 378,
    "batch_idx": 128,
    "loss": 0.3692103624343872,
    "learning_rate": 0.0009144211634523803,
    "timestep_mean": 525.40625,
    "epsilon_norm": 7.579912185668945,
    "prediction_norm": 6.085991859436035,
    "timestamp": "2025-07-14T12:33:46.900382"
  },
  {
    "epoch": 378,
    "batch_idx": 160,
    "loss": 0.347423255443573,
    "learning_rate": 0.0009144211634523803,
    "timestep_mean": 506.875,
    "epsilon_norm": 7.34377384185791,
    "prediction_norm": 5.296910762786865,
    "timestamp": "2025-07-14T12:33:46.905479"
  },
  {
    "epoch": 378,
    "batch_idx": 192,
    "loss": 0.5906371474266052,
    "learning_rate": 0.0009144211634523803,
    "timestep_mean": 485.125,
    "epsilon_norm": 4.445722579956055,
    "prediction_norm": 3.3881728649139404,
    "timestamp": "2025-07-14T12:33:46.910706"
  },
  {
    "epoch": 379,
    "batch_idx": 0,
    "loss": 0.245950847864151,
    "learning_rate": 0.000913981236023604,
    "timestep_mean": 499.03125,
    "epsilon_norm": 7.785088539123535,
    "prediction_norm": 6.598616123199463,
    "timestamp": "2025-07-14T12:33:46.915904"
  },
  {
    "epoch": 379,
    "batch_idx": 32,
    "loss": 0.3645993173122406,
    "learning_rate": 0.000913981236023604,
    "timestep_mean": 522.0,
    "epsilon_norm": 8.299405097961426,
    "prediction_norm": 6.437527179718018,
    "timestamp": "2025-07-14T12:33:46.921160"
  },
  {
    "epoch": 379,
    "batch_idx": 64,
    "loss": 0.3036770820617676,
    "learning_rate": 0.000913981236023604,
    "timestep_mean": 482.75,
    "epsilon_norm": 8.020859718322754,
    "prediction_norm": 6.126039505004883,
    "timestamp": "2025-07-14T12:33:46.926283"
  },
  {
    "epoch": 379,
    "batch_idx": 96,
    "loss": 0.6278923749923706,
    "learning_rate": 0.000913981236023604,
    "timestep_mean": 431.8125,
    "epsilon_norm": 8.853232383728027,
    "prediction_norm": 6.649745941162109,
    "timestamp": "2025-07-14T12:33:46.931332"
  },
  {
    "epoch": 379,
    "batch_idx": 128,
    "loss": 0.3560621738433838,
    "learning_rate": 0.000913981236023604,
    "timestep_mean": 533.5,
    "epsilon_norm": 8.399866104125977,
    "prediction_norm": 6.724804878234863,
    "timestamp": "2025-07-14T12:33:46.936453"
  },
  {
    "epoch": 379,
    "batch_idx": 160,
    "loss": 0.3648810386657715,
    "learning_rate": 0.000913981236023604,
    "timestep_mean": 429.75,
    "epsilon_norm": 7.851602077484131,
    "prediction_norm": 6.104230880737305,
    "timestamp": "2025-07-14T12:33:46.941723"
  },
  {
    "epoch": 379,
    "batch_idx": 192,
    "loss": 0.2903828024864197,
    "learning_rate": 0.000913981236023604,
    "timestep_mean": 479.125,
    "epsilon_norm": 2.788893699645996,
    "prediction_norm": 1.4264123439788818,
    "timestamp": "2025-07-14T12:33:46.946869"
  },
  {
    "epoch": 380,
    "batch_idx": 0,
    "loss": 0.14287492632865906,
    "learning_rate": 0.0009135402871372803,
    "timestep_mean": 533.3125,
    "epsilon_norm": 8.100934982299805,
    "prediction_norm": 7.172347545623779,
    "timestamp": "2025-07-14T12:33:46.951950"
  },
  {
    "epoch": 380,
    "batch_idx": 32,
    "loss": 0.31567490100860596,
    "learning_rate": 0.0009135402871372803,
    "timestep_mean": 449.625,
    "epsilon_norm": 7.116452217102051,
    "prediction_norm": 5.738299369812012,
    "timestamp": "2025-07-14T12:33:46.957025"
  },
  {
    "epoch": 380,
    "batch_idx": 64,
    "loss": 0.2607179284095764,
    "learning_rate": 0.0009135402871372803,
    "timestep_mean": 551.0,
    "epsilon_norm": 8.042359352111816,
    "prediction_norm": 6.5541157722473145,
    "timestamp": "2025-07-14T12:33:46.961800"
  },
  {
    "epoch": 380,
    "batch_idx": 96,
    "loss": 0.5008677244186401,
    "learning_rate": 0.0009135402871372803,
    "timestep_mean": 498.90625,
    "epsilon_norm": 8.716548919677734,
    "prediction_norm": 7.312173366546631,
    "timestamp": "2025-07-14T12:33:46.966730"
  },
  {
    "epoch": 380,
    "batch_idx": 128,
    "loss": 0.3491700291633606,
    "learning_rate": 0.0009135402871372803,
    "timestep_mean": 531.34375,
    "epsilon_norm": 8.441939353942871,
    "prediction_norm": 6.890341758728027,
    "timestamp": "2025-07-14T12:33:46.971624"
  },
  {
    "epoch": 380,
    "batch_idx": 160,
    "loss": 0.17303353548049927,
    "learning_rate": 0.0009135402871372803,
    "timestep_mean": 610.90625,
    "epsilon_norm": 7.164989948272705,
    "prediction_norm": 6.5088934898376465,
    "timestamp": "2025-07-14T12:33:46.976553"
  },
  {
    "epoch": 380,
    "batch_idx": 192,
    "loss": 0.2788620889186859,
    "learning_rate": 0.0009135402871372803,
    "timestep_mean": 485.875,
    "epsilon_norm": 3.5158145427703857,
    "prediction_norm": 2.703547954559326,
    "timestamp": "2025-07-14T12:33:46.981413"
  },
  {
    "epoch": 381,
    "batch_idx": 0,
    "loss": 0.28712350130081177,
    "learning_rate": 0.0009130983178814071,
    "timestep_mean": 538.96875,
    "epsilon_norm": 8.556864738464355,
    "prediction_norm": 7.761315822601318,
    "timestamp": "2025-07-14T12:33:46.986272"
  },
  {
    "epoch": 381,
    "batch_idx": 32,
    "loss": 0.42169177532196045,
    "learning_rate": 0.0009130983178814071,
    "timestep_mean": 420.34375,
    "epsilon_norm": 7.645704746246338,
    "prediction_norm": 5.890979290008545,
    "timestamp": "2025-07-14T12:33:46.991173"
  },
  {
    "epoch": 381,
    "batch_idx": 64,
    "loss": 0.22264394164085388,
    "learning_rate": 0.0009130983178814071,
    "timestep_mean": 484.9375,
    "epsilon_norm": 8.047333717346191,
    "prediction_norm": 6.8376784324646,
    "timestamp": "2025-07-14T12:33:46.996096"
  },
  {
    "epoch": 381,
    "batch_idx": 96,
    "loss": 0.27358660101890564,
    "learning_rate": 0.0009130983178814071,
    "timestep_mean": 489.46875,
    "epsilon_norm": 8.12814712524414,
    "prediction_norm": 7.1763811111450195,
    "timestamp": "2025-07-14T12:33:47.001064"
  },
  {
    "epoch": 381,
    "batch_idx": 128,
    "loss": 0.15969257056713104,
    "learning_rate": 0.0009130983178814071,
    "timestep_mean": 491.78125,
    "epsilon_norm": 6.777032375335693,
    "prediction_norm": 6.306613922119141,
    "timestamp": "2025-07-14T12:33:47.005984"
  },
  {
    "epoch": 381,
    "batch_idx": 160,
    "loss": 0.4298820495605469,
    "learning_rate": 0.0009130983178814071,
    "timestep_mean": 436.4375,
    "epsilon_norm": 8.566910743713379,
    "prediction_norm": 7.284010410308838,
    "timestamp": "2025-07-14T12:33:47.010835"
  },
  {
    "epoch": 381,
    "batch_idx": 192,
    "loss": 0.059799082577228546,
    "learning_rate": 0.0009130983178814071,
    "timestep_mean": 544.75,
    "epsilon_norm": 3.8356707096099854,
    "prediction_norm": 4.369474411010742,
    "timestamp": "2025-07-14T12:33:47.015637"
  },
  {
    "epoch": 382,
    "batch_idx": 0,
    "loss": 0.22435542941093445,
    "learning_rate": 0.0009126553293464994,
    "timestep_mean": 507.9375,
    "epsilon_norm": 8.206829071044922,
    "prediction_norm": 7.0859575271606445,
    "timestamp": "2025-07-14T12:33:47.020452"
  },
  {
    "epoch": 382,
    "batch_idx": 32,
    "loss": 0.3205886483192444,
    "learning_rate": 0.0009126553293464994,
    "timestep_mean": 568.25,
    "epsilon_norm": 8.066937446594238,
    "prediction_norm": 6.264998435974121,
    "timestamp": "2025-07-14T12:33:47.025394"
  },
  {
    "epoch": 382,
    "batch_idx": 64,
    "loss": 0.32761847972869873,
    "learning_rate": 0.0009126553293464994,
    "timestep_mean": 526.09375,
    "epsilon_norm": 9.197253227233887,
    "prediction_norm": 8.07101058959961,
    "timestamp": "2025-07-14T12:33:47.030243"
  },
  {
    "epoch": 382,
    "batch_idx": 96,
    "loss": 0.32626962661743164,
    "learning_rate": 0.0009126553293464994,
    "timestep_mean": 602.25,
    "epsilon_norm": 7.4729084968566895,
    "prediction_norm": 5.998202323913574,
    "timestamp": "2025-07-14T12:33:47.035160"
  },
  {
    "epoch": 382,
    "batch_idx": 128,
    "loss": 0.2168334573507309,
    "learning_rate": 0.0009126553293464994,
    "timestep_mean": 598.375,
    "epsilon_norm": 8.141075134277344,
    "prediction_norm": 7.267999649047852,
    "timestamp": "2025-07-14T12:33:47.039978"
  },
  {
    "epoch": 382,
    "batch_idx": 160,
    "loss": 0.3257434666156769,
    "learning_rate": 0.0009126553293464994,
    "timestep_mean": 449.90625,
    "epsilon_norm": 7.359846115112305,
    "prediction_norm": 6.054855823516846,
    "timestamp": "2025-07-14T12:33:47.044825"
  },
  {
    "epoch": 382,
    "batch_idx": 192,
    "loss": 0.06174455210566521,
    "learning_rate": 0.0009126553293464994,
    "timestep_mean": 713.875,
    "epsilon_norm": 3.07962703704834,
    "prediction_norm": 2.9703009128570557,
    "timestamp": "2025-07-14T12:33:47.049638"
  },
  {
    "epoch": 383,
    "batch_idx": 0,
    "loss": 0.3391224145889282,
    "learning_rate": 0.0009122113226255872,
    "timestep_mean": 504.21875,
    "epsilon_norm": 8.538383483886719,
    "prediction_norm": 7.318571090698242,
    "timestamp": "2025-07-14T12:33:47.054588"
  },
  {
    "epoch": 383,
    "batch_idx": 32,
    "loss": 0.5075046420097351,
    "learning_rate": 0.0009122113226255872,
    "timestep_mean": 407.4375,
    "epsilon_norm": 8.220245361328125,
    "prediction_norm": 6.015698432922363,
    "timestamp": "2025-07-14T12:33:47.059401"
  },
  {
    "epoch": 383,
    "batch_idx": 64,
    "loss": 0.36984074115753174,
    "learning_rate": 0.0009122113226255872,
    "timestep_mean": 403.8125,
    "epsilon_norm": 8.420649528503418,
    "prediction_norm": 6.825329780578613,
    "timestamp": "2025-07-14T12:33:47.064250"
  },
  {
    "epoch": 383,
    "batch_idx": 96,
    "loss": 0.3894708752632141,
    "learning_rate": 0.0009122113226255872,
    "timestep_mean": 536.15625,
    "epsilon_norm": 8.911434173583984,
    "prediction_norm": 6.841189861297607,
    "timestamp": "2025-07-14T12:33:47.069087"
  },
  {
    "epoch": 383,
    "batch_idx": 128,
    "loss": 0.3391876816749573,
    "learning_rate": 0.0009122113226255872,
    "timestep_mean": 502.21875,
    "epsilon_norm": 7.419733047485352,
    "prediction_norm": 5.417155742645264,
    "timestamp": "2025-07-14T12:33:47.073927"
  },
  {
    "epoch": 383,
    "batch_idx": 160,
    "loss": 0.5440739393234253,
    "learning_rate": 0.0009122113226255872,
    "timestep_mean": 436.9375,
    "epsilon_norm": 8.627439498901367,
    "prediction_norm": 5.857480049133301,
    "timestamp": "2025-07-14T12:33:47.078802"
  },
  {
    "epoch": 383,
    "batch_idx": 192,
    "loss": 0.19256441295146942,
    "learning_rate": 0.0009122113226255872,
    "timestep_mean": 613.375,
    "epsilon_norm": 4.316714286804199,
    "prediction_norm": 3.9647161960601807,
    "timestamp": "2025-07-14T12:33:47.083798"
  },
  {
    "epoch": 384,
    "batch_idx": 0,
    "loss": 0.24243740737438202,
    "learning_rate": 0.0009117662988142131,
    "timestep_mean": 484.15625,
    "epsilon_norm": 6.983468055725098,
    "prediction_norm": 5.567881107330322,
    "timestamp": "2025-07-14T12:33:47.088650"
  },
  {
    "epoch": 384,
    "batch_idx": 32,
    "loss": 0.42282605171203613,
    "learning_rate": 0.0009117662988142131,
    "timestep_mean": 444.84375,
    "epsilon_norm": 7.47990083694458,
    "prediction_norm": 6.201026439666748,
    "timestamp": "2025-07-14T12:33:47.093502"
  },
  {
    "epoch": 384,
    "batch_idx": 64,
    "loss": 0.1951887160539627,
    "learning_rate": 0.0009117662988142131,
    "timestep_mean": 542.9375,
    "epsilon_norm": 8.3868408203125,
    "prediction_norm": 7.640333652496338,
    "timestamp": "2025-07-14T12:33:47.098342"
  },
  {
    "epoch": 384,
    "batch_idx": 96,
    "loss": 0.405902236700058,
    "learning_rate": 0.0009117662988142131,
    "timestep_mean": 509.375,
    "epsilon_norm": 9.262486457824707,
    "prediction_norm": 7.153860092163086,
    "timestamp": "2025-07-14T12:33:47.103142"
  },
  {
    "epoch": 384,
    "batch_idx": 128,
    "loss": 0.29826080799102783,
    "learning_rate": 0.0009117662988142131,
    "timestep_mean": 501.21875,
    "epsilon_norm": 8.613492012023926,
    "prediction_norm": 7.936779499053955,
    "timestamp": "2025-07-14T12:33:47.113924"
  },
  {
    "epoch": 384,
    "batch_idx": 160,
    "loss": 0.30765610933303833,
    "learning_rate": 0.0009117662988142131,
    "timestep_mean": 439.84375,
    "epsilon_norm": 7.438939094543457,
    "prediction_norm": 5.774415493011475,
    "timestamp": "2025-07-14T12:33:47.118789"
  },
  {
    "epoch": 384,
    "batch_idx": 192,
    "loss": 0.24321261048316956,
    "learning_rate": 0.0009117662988142131,
    "timestep_mean": 506.0,
    "epsilon_norm": 3.464845657348633,
    "prediction_norm": 3.228754997253418,
    "timestamp": "2025-07-14T12:33:47.123628"
  },
  {
    "epoch": 385,
    "batch_idx": 0,
    "loss": 0.34906846284866333,
    "learning_rate": 0.0009113202590104294,
    "timestep_mean": 515.875,
    "epsilon_norm": 7.708472728729248,
    "prediction_norm": 6.740666389465332,
    "timestamp": "2025-07-14T12:33:47.128559"
  },
  {
    "epoch": 385,
    "batch_idx": 32,
    "loss": 0.18182504177093506,
    "learning_rate": 0.0009113202590104294,
    "timestep_mean": 518.875,
    "epsilon_norm": 6.993478775024414,
    "prediction_norm": 6.647475242614746,
    "timestamp": "2025-07-14T12:33:47.133522"
  },
  {
    "epoch": 385,
    "batch_idx": 64,
    "loss": 0.20872218906879425,
    "learning_rate": 0.0009113202590104294,
    "timestep_mean": 473.90625,
    "epsilon_norm": 7.43433952331543,
    "prediction_norm": 7.140961647033691,
    "timestamp": "2025-07-14T12:33:47.138360"
  },
  {
    "epoch": 385,
    "batch_idx": 96,
    "loss": 0.3932303190231323,
    "learning_rate": 0.0009113202590104294,
    "timestep_mean": 460.0625,
    "epsilon_norm": 8.802886009216309,
    "prediction_norm": 8.208298683166504,
    "timestamp": "2025-07-14T12:33:47.143164"
  },
  {
    "epoch": 385,
    "batch_idx": 128,
    "loss": 0.4764167368412018,
    "learning_rate": 0.0009113202590104294,
    "timestep_mean": 421.65625,
    "epsilon_norm": 8.683103561401367,
    "prediction_norm": 6.512998104095459,
    "timestamp": "2025-07-14T12:33:47.148050"
  },
  {
    "epoch": 385,
    "batch_idx": 160,
    "loss": 0.21626752614974976,
    "learning_rate": 0.0009113202590104294,
    "timestep_mean": 494.09375,
    "epsilon_norm": 8.293495178222656,
    "prediction_norm": 7.555916786193848,
    "timestamp": "2025-07-14T12:33:47.152909"
  },
  {
    "epoch": 385,
    "batch_idx": 192,
    "loss": 0.22979086637496948,
    "learning_rate": 0.0009113202590104294,
    "timestep_mean": 580.5,
    "epsilon_norm": 4.281595230102539,
    "prediction_norm": 3.703968048095703,
    "timestamp": "2025-07-14T12:33:47.157720"
  },
  {
    "epoch": 386,
    "batch_idx": 0,
    "loss": 0.4288714528083801,
    "learning_rate": 0.0009108732043147947,
    "timestep_mean": 435.96875,
    "epsilon_norm": 8.007796287536621,
    "prediction_norm": 6.589771747589111,
    "timestamp": "2025-07-14T12:33:47.162761"
  },
  {
    "epoch": 386,
    "batch_idx": 32,
    "loss": 0.21944230794906616,
    "learning_rate": 0.0009108732043147947,
    "timestep_mean": 465.53125,
    "epsilon_norm": 7.427361488342285,
    "prediction_norm": 6.300756454467773,
    "timestamp": "2025-07-14T12:33:47.167570"
  },
  {
    "epoch": 386,
    "batch_idx": 64,
    "loss": 0.2377319037914276,
    "learning_rate": 0.0009108732043147947,
    "timestep_mean": 509.53125,
    "epsilon_norm": 7.46825647354126,
    "prediction_norm": 6.616880893707275,
    "timestamp": "2025-07-14T12:33:47.172396"
  },
  {
    "epoch": 386,
    "batch_idx": 96,
    "loss": 0.4021279513835907,
    "learning_rate": 0.0009108732043147947,
    "timestep_mean": 389.71875,
    "epsilon_norm": 7.448846817016602,
    "prediction_norm": 4.555451393127441,
    "timestamp": "2025-07-14T12:33:47.177222"
  },
  {
    "epoch": 386,
    "batch_idx": 128,
    "loss": 0.1439189612865448,
    "learning_rate": 0.0009108732043147947,
    "timestep_mean": 614.8125,
    "epsilon_norm": 6.341395378112793,
    "prediction_norm": 4.917256832122803,
    "timestamp": "2025-07-14T12:33:47.182003"
  },
  {
    "epoch": 386,
    "batch_idx": 160,
    "loss": 0.3286561369895935,
    "learning_rate": 0.0009108732043147947,
    "timestep_mean": 540.21875,
    "epsilon_norm": 7.7809576988220215,
    "prediction_norm": 5.5683512687683105,
    "timestamp": "2025-07-14T12:33:47.186877"
  },
  {
    "epoch": 386,
    "batch_idx": 192,
    "loss": 0.04959952458739281,
    "learning_rate": 0.0009108732043147947,
    "timestep_mean": 619.375,
    "epsilon_norm": 3.2351603507995605,
    "prediction_norm": 3.0061230659484863,
    "timestamp": "2025-07-14T12:33:47.191712"
  },
  {
    "epoch": 387,
    "batch_idx": 0,
    "loss": 0.24619275331497192,
    "learning_rate": 0.0009104251358303719,
    "timestep_mean": 502.96875,
    "epsilon_norm": 8.16820240020752,
    "prediction_norm": 6.265803813934326,
    "timestamp": "2025-07-14T12:33:47.196569"
  },
  {
    "epoch": 387,
    "batch_idx": 32,
    "loss": 0.3103325664997101,
    "learning_rate": 0.0009104251358303719,
    "timestep_mean": 549.75,
    "epsilon_norm": 8.947409629821777,
    "prediction_norm": 7.166922569274902,
    "timestamp": "2025-07-14T12:33:47.201413"
  },
  {
    "epoch": 387,
    "batch_idx": 64,
    "loss": 0.23559430241584778,
    "learning_rate": 0.0009104251358303719,
    "timestep_mean": 534.5625,
    "epsilon_norm": 7.936453342437744,
    "prediction_norm": 6.803822040557861,
    "timestamp": "2025-07-14T12:33:47.206233"
  },
  {
    "epoch": 387,
    "batch_idx": 96,
    "loss": 0.23767638206481934,
    "learning_rate": 0.0009104251358303719,
    "timestep_mean": 484.46875,
    "epsilon_norm": 7.335801601409912,
    "prediction_norm": 6.113001823425293,
    "timestamp": "2025-07-14T12:33:47.211125"
  },
  {
    "epoch": 387,
    "batch_idx": 128,
    "loss": 0.29366716742515564,
    "learning_rate": 0.0009104251358303719,
    "timestep_mean": 497.8125,
    "epsilon_norm": 7.446328163146973,
    "prediction_norm": 6.471634864807129,
    "timestamp": "2025-07-14T12:33:47.215964"
  },
  {
    "epoch": 387,
    "batch_idx": 160,
    "loss": 0.2181309163570404,
    "learning_rate": 0.0009104251358303719,
    "timestep_mean": 504.09375,
    "epsilon_norm": 8.13509750366211,
    "prediction_norm": 6.474783897399902,
    "timestamp": "2025-07-14T12:33:47.220880"
  },
  {
    "epoch": 387,
    "batch_idx": 192,
    "loss": 0.4461767077445984,
    "learning_rate": 0.0009104251358303719,
    "timestep_mean": 428.0,
    "epsilon_norm": 4.5279765129089355,
    "prediction_norm": 3.3227148056030273,
    "timestamp": "2025-07-14T12:33:47.225690"
  },
  {
    "epoch": 388,
    "batch_idx": 0,
    "loss": 0.3863016664981842,
    "learning_rate": 0.0009099760546627258,
    "timestep_mean": 488.03125,
    "epsilon_norm": 8.945290565490723,
    "prediction_norm": 7.150551795959473,
    "timestamp": "2025-07-14T12:33:47.230568"
  },
  {
    "epoch": 388,
    "batch_idx": 32,
    "loss": 0.26845139265060425,
    "learning_rate": 0.0009099760546627258,
    "timestep_mean": 462.78125,
    "epsilon_norm": 6.825475692749023,
    "prediction_norm": 6.7537055015563965,
    "timestamp": "2025-07-14T12:33:47.235432"
  },
  {
    "epoch": 388,
    "batch_idx": 64,
    "loss": 0.3306322693824768,
    "learning_rate": 0.0009099760546627258,
    "timestep_mean": 441.0625,
    "epsilon_norm": 7.740973472595215,
    "prediction_norm": 7.316278457641602,
    "timestamp": "2025-07-14T12:33:47.240238"
  },
  {
    "epoch": 388,
    "batch_idx": 96,
    "loss": 0.26419681310653687,
    "learning_rate": 0.0009099760546627258,
    "timestep_mean": 540.8125,
    "epsilon_norm": 8.51372241973877,
    "prediction_norm": 7.84295129776001,
    "timestamp": "2025-07-14T12:33:47.245079"
  },
  {
    "epoch": 388,
    "batch_idx": 128,
    "loss": 0.2634185552597046,
    "learning_rate": 0.0009099760546627258,
    "timestep_mean": 557.4375,
    "epsilon_norm": 7.9830169677734375,
    "prediction_norm": 7.900806427001953,
    "timestamp": "2025-07-14T12:33:47.250043"
  },
  {
    "epoch": 388,
    "batch_idx": 160,
    "loss": 0.3605384826660156,
    "learning_rate": 0.0009099760546627258,
    "timestep_mean": 545.28125,
    "epsilon_norm": 8.488308906555176,
    "prediction_norm": 8.082418441772461,
    "timestamp": "2025-07-14T12:33:47.254896"
  },
  {
    "epoch": 388,
    "batch_idx": 192,
    "loss": 0.47640809416770935,
    "learning_rate": 0.0009099760546627258,
    "timestep_mean": 545.5,
    "epsilon_norm": 4.270870208740234,
    "prediction_norm": 3.6858761310577393,
    "timestamp": "2025-07-14T12:33:47.259750"
  },
  {
    "epoch": 389,
    "batch_idx": 0,
    "loss": 0.25921526551246643,
    "learning_rate": 0.0009095259619199192,
    "timestep_mean": 585.375,
    "epsilon_norm": 7.410343170166016,
    "prediction_norm": 6.6940083503723145,
    "timestamp": "2025-07-14T12:33:47.264623"
  },
  {
    "epoch": 389,
    "batch_idx": 32,
    "loss": 0.12624140083789825,
    "learning_rate": 0.0009095259619199192,
    "timestep_mean": 585.5,
    "epsilon_norm": 7.809053897857666,
    "prediction_norm": 7.484401702880859,
    "timestamp": "2025-07-14T12:33:47.269480"
  },
  {
    "epoch": 389,
    "batch_idx": 64,
    "loss": 0.22737297415733337,
    "learning_rate": 0.0009095259619199192,
    "timestep_mean": 576.21875,
    "epsilon_norm": 6.909130573272705,
    "prediction_norm": 6.247427940368652,
    "timestamp": "2025-07-14T12:33:47.274336"
  },
  {
    "epoch": 389,
    "batch_idx": 96,
    "loss": 0.39232897758483887,
    "learning_rate": 0.0009095259619199192,
    "timestep_mean": 526.40625,
    "epsilon_norm": 9.19918155670166,
    "prediction_norm": 7.997061729431152,
    "timestamp": "2025-07-14T12:33:47.279210"
  },
  {
    "epoch": 389,
    "batch_idx": 128,
    "loss": 0.46179303526878357,
    "learning_rate": 0.0009095259619199192,
    "timestep_mean": 410.1875,
    "epsilon_norm": 7.54665994644165,
    "prediction_norm": 5.403060436248779,
    "timestamp": "2025-07-14T12:33:47.284049"
  },
  {
    "epoch": 389,
    "batch_idx": 160,
    "loss": 0.41538092494010925,
    "learning_rate": 0.0009095259619199192,
    "timestep_mean": 457.03125,
    "epsilon_norm": 8.448546409606934,
    "prediction_norm": 6.879824161529541,
    "timestamp": "2025-07-14T12:33:47.288888"
  },
  {
    "epoch": 389,
    "batch_idx": 192,
    "loss": 0.4208380877971649,
    "learning_rate": 0.0009095259619199192,
    "timestep_mean": 510.875,
    "epsilon_norm": 3.8390989303588867,
    "prediction_norm": 2.5334713459014893,
    "timestamp": "2025-07-14T12:33:47.293832"
  },
  {
    "epoch": 390,
    "batch_idx": 0,
    "loss": 0.2249654233455658,
    "learning_rate": 0.0009090748587125112,
    "timestep_mean": 448.375,
    "epsilon_norm": 7.635381698608398,
    "prediction_norm": 6.029047966003418,
    "timestamp": "2025-07-14T12:33:47.298738"
  },
  {
    "epoch": 390,
    "batch_idx": 32,
    "loss": 0.2703670859336853,
    "learning_rate": 0.0009090748587125112,
    "timestep_mean": 578.125,
    "epsilon_norm": 8.126022338867188,
    "prediction_norm": 6.358936786651611,
    "timestamp": "2025-07-14T12:33:47.303598"
  },
  {
    "epoch": 390,
    "batch_idx": 64,
    "loss": 0.24181683361530304,
    "learning_rate": 0.0009090748587125112,
    "timestep_mean": 536.03125,
    "epsilon_norm": 7.109699726104736,
    "prediction_norm": 5.352100372314453,
    "timestamp": "2025-07-14T12:33:47.308610"
  },
  {
    "epoch": 390,
    "batch_idx": 96,
    "loss": 0.312715083360672,
    "learning_rate": 0.0009090748587125112,
    "timestep_mean": 453.8125,
    "epsilon_norm": 7.03009033203125,
    "prediction_norm": 4.629850387573242,
    "timestamp": "2025-07-14T12:33:47.313455"
  },
  {
    "epoch": 390,
    "batch_idx": 128,
    "loss": 0.6085129976272583,
    "learning_rate": 0.0009090748587125112,
    "timestep_mean": 383.03125,
    "epsilon_norm": 8.627937316894531,
    "prediction_norm": 5.009953498840332,
    "timestamp": "2025-07-14T12:33:47.318248"
  },
  {
    "epoch": 390,
    "batch_idx": 160,
    "loss": 0.2664499282836914,
    "learning_rate": 0.0009090748587125112,
    "timestep_mean": 488.25,
    "epsilon_norm": 7.291323184967041,
    "prediction_norm": 5.7288618087768555,
    "timestamp": "2025-07-14T12:33:47.323135"
  },
  {
    "epoch": 390,
    "batch_idx": 192,
    "loss": 0.6178404092788696,
    "learning_rate": 0.0009090748587125112,
    "timestep_mean": 513.875,
    "epsilon_norm": 4.563313007354736,
    "prediction_norm": 3.052056312561035,
    "timestamp": "2025-07-14T12:33:47.327978"
  },
  {
    "epoch": 391,
    "batch_idx": 0,
    "loss": 0.20027124881744385,
    "learning_rate": 0.0009086227461535545,
    "timestep_mean": 610.75,
    "epsilon_norm": 7.270041465759277,
    "prediction_norm": 5.944759845733643,
    "timestamp": "2025-07-14T12:33:47.332811"
  },
  {
    "epoch": 391,
    "batch_idx": 32,
    "loss": 0.5130071640014648,
    "learning_rate": 0.0009086227461535545,
    "timestep_mean": 462.5625,
    "epsilon_norm": 8.834693908691406,
    "prediction_norm": 6.7827959060668945,
    "timestamp": "2025-07-14T12:33:47.337750"
  },
  {
    "epoch": 391,
    "batch_idx": 64,
    "loss": 0.2788015604019165,
    "learning_rate": 0.0009086227461535545,
    "timestep_mean": 455.6875,
    "epsilon_norm": 8.079535484313965,
    "prediction_norm": 6.518146514892578,
    "timestamp": "2025-07-14T12:33:47.342541"
  },
  {
    "epoch": 391,
    "batch_idx": 96,
    "loss": 0.18710193037986755,
    "learning_rate": 0.0009086227461535545,
    "timestep_mean": 527.03125,
    "epsilon_norm": 7.685517311096191,
    "prediction_norm": 6.73834753036499,
    "timestamp": "2025-07-14T12:33:47.347385"
  },
  {
    "epoch": 391,
    "batch_idx": 128,
    "loss": 0.347905695438385,
    "learning_rate": 0.0009086227461535545,
    "timestep_mean": 455.0625,
    "epsilon_norm": 6.570341110229492,
    "prediction_norm": 5.305156707763672,
    "timestamp": "2025-07-14T12:33:47.352310"
  },
  {
    "epoch": 391,
    "batch_idx": 160,
    "loss": 0.29956647753715515,
    "learning_rate": 0.0009086227461535545,
    "timestep_mean": 501.375,
    "epsilon_norm": 8.375587463378906,
    "prediction_norm": 6.7173614501953125,
    "timestamp": "2025-07-14T12:33:47.357183"
  },
  {
    "epoch": 391,
    "batch_idx": 192,
    "loss": 0.2528141438961029,
    "learning_rate": 0.0009086227461535545,
    "timestep_mean": 412.625,
    "epsilon_norm": 4.251312732696533,
    "prediction_norm": 4.007143020629883,
    "timestamp": "2025-07-14T12:33:47.362047"
  },
  {
    "epoch": 392,
    "batch_idx": 0,
    "loss": 0.2760295271873474,
    "learning_rate": 0.0009081696253585915,
    "timestep_mean": 476.21875,
    "epsilon_norm": 7.786892890930176,
    "prediction_norm": 7.183342456817627,
    "timestamp": "2025-07-14T12:33:47.366936"
  },
  {
    "epoch": 392,
    "batch_idx": 32,
    "loss": 0.2395719587802887,
    "learning_rate": 0.0009081696253585915,
    "timestep_mean": 459.625,
    "epsilon_norm": 7.321230411529541,
    "prediction_norm": 6.398129940032959,
    "timestamp": "2025-07-14T12:33:47.371835"
  },
  {
    "epoch": 392,
    "batch_idx": 64,
    "loss": 0.3847854733467102,
    "learning_rate": 0.0009081696253585915,
    "timestep_mean": 457.28125,
    "epsilon_norm": 8.376273155212402,
    "prediction_norm": 6.014871597290039,
    "timestamp": "2025-07-14T12:33:47.376690"
  },
  {
    "epoch": 392,
    "batch_idx": 96,
    "loss": 0.35181373357772827,
    "learning_rate": 0.0009081696253585915,
    "timestep_mean": 472.46875,
    "epsilon_norm": 7.503905296325684,
    "prediction_norm": 6.347714424133301,
    "timestamp": "2025-07-14T12:33:47.381569"
  },
  {
    "epoch": 392,
    "batch_idx": 128,
    "loss": 0.26569730043411255,
    "learning_rate": 0.0009081696253585915,
    "timestep_mean": 524.28125,
    "epsilon_norm": 7.650736331939697,
    "prediction_norm": 6.300340175628662,
    "timestamp": "2025-07-14T12:33:47.386383"
  },
  {
    "epoch": 392,
    "batch_idx": 160,
    "loss": 0.18231521546840668,
    "learning_rate": 0.0009081696253585915,
    "timestep_mean": 543.65625,
    "epsilon_norm": 8.261905670166016,
    "prediction_norm": 7.4865264892578125,
    "timestamp": "2025-07-14T12:33:47.391251"
  },
  {
    "epoch": 392,
    "batch_idx": 192,
    "loss": 0.06120119243860245,
    "learning_rate": 0.0009081696253585915,
    "timestep_mean": 693.125,
    "epsilon_norm": 3.2306065559387207,
    "prediction_norm": 3.5608255863189697,
    "timestamp": "2025-07-14T12:33:47.396198"
  },
  {
    "epoch": 393,
    "batch_idx": 0,
    "loss": 0.47172996401786804,
    "learning_rate": 0.0009077154974456529,
    "timestep_mean": 495.75,
    "epsilon_norm": 9.602818489074707,
    "prediction_norm": 7.66719913482666,
    "timestamp": "2025-07-14T12:33:47.401046"
  },
  {
    "epoch": 393,
    "batch_idx": 32,
    "loss": 0.32598620653152466,
    "learning_rate": 0.0009077154974456529,
    "timestep_mean": 412.5625,
    "epsilon_norm": 7.610756874084473,
    "prediction_norm": 6.567131042480469,
    "timestamp": "2025-07-14T12:33:47.405953"
  },
  {
    "epoch": 393,
    "batch_idx": 64,
    "loss": 0.29810547828674316,
    "learning_rate": 0.0009077154974456529,
    "timestep_mean": 469.25,
    "epsilon_norm": 7.871214866638184,
    "prediction_norm": 6.581181526184082,
    "timestamp": "2025-07-14T12:33:47.410808"
  },
  {
    "epoch": 393,
    "batch_idx": 96,
    "loss": 0.3879941701889038,
    "learning_rate": 0.0009077154974456529,
    "timestep_mean": 516.375,
    "epsilon_norm": 7.956246376037598,
    "prediction_norm": 6.607825756072998,
    "timestamp": "2025-07-14T12:33:47.415759"
  },
  {
    "epoch": 393,
    "batch_idx": 128,
    "loss": 0.1493780016899109,
    "learning_rate": 0.0009077154974456529,
    "timestep_mean": 552.0,
    "epsilon_norm": 8.014355659484863,
    "prediction_norm": 7.249086856842041,
    "timestamp": "2025-07-14T12:33:47.420828"
  },
  {
    "epoch": 393,
    "batch_idx": 160,
    "loss": 0.2764413356781006,
    "learning_rate": 0.0009077154974456529,
    "timestep_mean": 510.125,
    "epsilon_norm": 7.761258602142334,
    "prediction_norm": 6.793670177459717,
    "timestamp": "2025-07-14T12:33:47.425802"
  },
  {
    "epoch": 393,
    "batch_idx": 192,
    "loss": 0.21591226756572723,
    "learning_rate": 0.0009077154974456529,
    "timestep_mean": 557.375,
    "epsilon_norm": 3.611048460006714,
    "prediction_norm": 2.8729946613311768,
    "timestamp": "2025-07-14T12:33:47.430826"
  },
  {
    "epoch": 394,
    "batch_idx": 0,
    "loss": 0.5526614785194397,
    "learning_rate": 0.0009072603635352543,
    "timestep_mean": 447.375,
    "epsilon_norm": 9.800658226013184,
    "prediction_norm": 7.321316242218018,
    "timestamp": "2025-07-14T12:33:47.435988"
  },
  {
    "epoch": 394,
    "batch_idx": 32,
    "loss": 0.46243083477020264,
    "learning_rate": 0.0009072603635352543,
    "timestep_mean": 444.1875,
    "epsilon_norm": 8.695317268371582,
    "prediction_norm": 6.748047351837158,
    "timestamp": "2025-07-14T12:33:47.440840"
  },
  {
    "epoch": 394,
    "batch_idx": 64,
    "loss": 0.4416510760784149,
    "learning_rate": 0.0009072603635352543,
    "timestep_mean": 461.1875,
    "epsilon_norm": 7.990733623504639,
    "prediction_norm": 6.036722660064697,
    "timestamp": "2025-07-14T12:33:47.445718"
  },
  {
    "epoch": 394,
    "batch_idx": 96,
    "loss": 0.17241182923316956,
    "learning_rate": 0.0009072603635352543,
    "timestep_mean": 551.625,
    "epsilon_norm": 7.499109268188477,
    "prediction_norm": 6.707020282745361,
    "timestamp": "2025-07-14T12:33:47.450618"
  },
  {
    "epoch": 394,
    "batch_idx": 128,
    "loss": 0.33902138471603394,
    "learning_rate": 0.0009072603635352543,
    "timestep_mean": 544.125,
    "epsilon_norm": 7.644613265991211,
    "prediction_norm": 6.492349624633789,
    "timestamp": "2025-07-14T12:33:47.455607"
  },
  {
    "epoch": 394,
    "batch_idx": 160,
    "loss": 0.24924048781394958,
    "learning_rate": 0.0009072603635352543,
    "timestep_mean": 553.0625,
    "epsilon_norm": 7.728740215301514,
    "prediction_norm": 6.441935062408447,
    "timestamp": "2025-07-14T12:33:47.460565"
  },
  {
    "epoch": 394,
    "batch_idx": 192,
    "loss": 0.49968063831329346,
    "learning_rate": 0.0009072603635352543,
    "timestep_mean": 519.125,
    "epsilon_norm": 5.17191743850708,
    "prediction_norm": 3.2419626712799072,
    "timestamp": "2025-07-14T12:33:47.465440"
  },
  {
    "epoch": 395,
    "batch_idx": 0,
    "loss": 0.1877775490283966,
    "learning_rate": 0.0009068042247503932,
    "timestep_mean": 612.21875,
    "epsilon_norm": 7.915755748748779,
    "prediction_norm": 7.10808801651001,
    "timestamp": "2025-07-14T12:33:47.470359"
  },
  {
    "epoch": 395,
    "batch_idx": 32,
    "loss": 0.42521804571151733,
    "learning_rate": 0.0009068042247503932,
    "timestep_mean": 519.4375,
    "epsilon_norm": 8.296319007873535,
    "prediction_norm": 6.7279253005981445,
    "timestamp": "2025-07-14T12:33:47.475459"
  },
  {
    "epoch": 395,
    "batch_idx": 64,
    "loss": 0.25386375188827515,
    "learning_rate": 0.0009068042247503932,
    "timestep_mean": 573.9375,
    "epsilon_norm": 8.845603942871094,
    "prediction_norm": 7.9409589767456055,
    "timestamp": "2025-07-14T12:33:47.480386"
  },
  {
    "epoch": 395,
    "batch_idx": 96,
    "loss": 0.44698092341423035,
    "learning_rate": 0.0009068042247503932,
    "timestep_mean": 487.15625,
    "epsilon_norm": 8.558450698852539,
    "prediction_norm": 7.36390495300293,
    "timestamp": "2025-07-14T12:33:47.485257"
  },
  {
    "epoch": 395,
    "batch_idx": 128,
    "loss": 0.1470540165901184,
    "learning_rate": 0.0009068042247503932,
    "timestep_mean": 659.53125,
    "epsilon_norm": 6.878442287445068,
    "prediction_norm": 6.801534175872803,
    "timestamp": "2025-07-14T12:33:47.490161"
  },
  {
    "epoch": 395,
    "batch_idx": 160,
    "loss": 0.24676191806793213,
    "learning_rate": 0.0009068042247503932,
    "timestep_mean": 492.28125,
    "epsilon_norm": 7.255410194396973,
    "prediction_norm": 6.619253635406494,
    "timestamp": "2025-07-14T12:33:47.495041"
  },
  {
    "epoch": 395,
    "batch_idx": 192,
    "loss": 0.3362213373184204,
    "learning_rate": 0.0009068042247503932,
    "timestep_mean": 379.5,
    "epsilon_norm": 4.542241096496582,
    "prediction_norm": 4.562015533447266,
    "timestamp": "2025-07-14T12:33:47.500067"
  },
  {
    "epoch": 396,
    "batch_idx": 0,
    "loss": 0.4797883927822113,
    "learning_rate": 0.0009063470822165466,
    "timestep_mean": 350.84375,
    "epsilon_norm": 8.114091873168945,
    "prediction_norm": 6.508447170257568,
    "timestamp": "2025-07-14T12:33:47.504971"
  },
  {
    "epoch": 396,
    "batch_idx": 32,
    "loss": 0.25394541025161743,
    "learning_rate": 0.0009063470822165466,
    "timestep_mean": 558.1875,
    "epsilon_norm": 8.073986053466797,
    "prediction_norm": 7.185022354125977,
    "timestamp": "2025-07-14T12:33:47.509844"
  },
  {
    "epoch": 396,
    "batch_idx": 64,
    "loss": 0.3081636130809784,
    "learning_rate": 0.0009063470822165466,
    "timestep_mean": 470.9375,
    "epsilon_norm": 8.42592716217041,
    "prediction_norm": 7.018808364868164,
    "timestamp": "2025-07-14T12:33:47.514757"
  },
  {
    "epoch": 396,
    "batch_idx": 96,
    "loss": 0.21311165392398834,
    "learning_rate": 0.0009063470822165466,
    "timestep_mean": 424.15625,
    "epsilon_norm": 6.623401165008545,
    "prediction_norm": 5.463318347930908,
    "timestamp": "2025-07-14T12:33:47.519605"
  },
  {
    "epoch": 396,
    "batch_idx": 128,
    "loss": 0.35730648040771484,
    "learning_rate": 0.0009063470822165466,
    "timestep_mean": 408.09375,
    "epsilon_norm": 8.326004028320312,
    "prediction_norm": 6.719676494598389,
    "timestamp": "2025-07-14T12:33:47.524637"
  },
  {
    "epoch": 396,
    "batch_idx": 160,
    "loss": 0.2565886974334717,
    "learning_rate": 0.0009063470822165466,
    "timestep_mean": 468.5625,
    "epsilon_norm": 8.543129920959473,
    "prediction_norm": 7.260468482971191,
    "timestamp": "2025-07-14T12:33:47.529575"
  },
  {
    "epoch": 396,
    "batch_idx": 192,
    "loss": 0.25272682309150696,
    "learning_rate": 0.0009063470822165466,
    "timestep_mean": 539.125,
    "epsilon_norm": 3.480868339538574,
    "prediction_norm": 2.00956130027771,
    "timestamp": "2025-07-14T12:33:47.534340"
  },
  {
    "epoch": 397,
    "batch_idx": 0,
    "loss": 0.26478248834609985,
    "learning_rate": 0.0009058889370616685,
    "timestep_mean": 493.125,
    "epsilon_norm": 6.980388164520264,
    "prediction_norm": 5.427079200744629,
    "timestamp": "2025-07-14T12:33:47.539209"
  },
  {
    "epoch": 397,
    "batch_idx": 32,
    "loss": 0.26541972160339355,
    "learning_rate": 0.0009058889370616685,
    "timestep_mean": 581.375,
    "epsilon_norm": 7.496420860290527,
    "prediction_norm": 5.918099403381348,
    "timestamp": "2025-07-14T12:33:47.544054"
  },
  {
    "epoch": 397,
    "batch_idx": 64,
    "loss": 0.4127817153930664,
    "learning_rate": 0.0009058889370616685,
    "timestep_mean": 490.53125,
    "epsilon_norm": 7.51641845703125,
    "prediction_norm": 5.89547872543335,
    "timestamp": "2025-07-14T12:33:47.548927"
  },
  {
    "epoch": 397,
    "batch_idx": 96,
    "loss": 0.33986562490463257,
    "learning_rate": 0.0009058889370616685,
    "timestep_mean": 490.5,
    "epsilon_norm": 6.966545581817627,
    "prediction_norm": 4.851737022399902,
    "timestamp": "2025-07-14T12:33:47.553792"
  },
  {
    "epoch": 397,
    "batch_idx": 128,
    "loss": 0.4691251814365387,
    "learning_rate": 0.0009058889370616685,
    "timestep_mean": 486.03125,
    "epsilon_norm": 9.040081977844238,
    "prediction_norm": 6.729933261871338,
    "timestamp": "2025-07-14T12:33:47.559055"
  },
  {
    "epoch": 397,
    "batch_idx": 160,
    "loss": 0.3056657910346985,
    "learning_rate": 0.0009058889370616685,
    "timestep_mean": 563.90625,
    "epsilon_norm": 7.388952255249023,
    "prediction_norm": 5.619429111480713,
    "timestamp": "2025-07-14T12:33:47.564109"
  },
  {
    "epoch": 397,
    "batch_idx": 192,
    "loss": 0.37020957469940186,
    "learning_rate": 0.0009058889370616685,
    "timestep_mean": 452.125,
    "epsilon_norm": 3.4824280738830566,
    "prediction_norm": 3.223609447479248,
    "timestamp": "2025-07-14T12:33:47.569100"
  },
  {
    "epoch": 398,
    "batch_idx": 0,
    "loss": 0.2906031608581543,
    "learning_rate": 0.0009054297904161864,
    "timestep_mean": 512.40625,
    "epsilon_norm": 8.027027130126953,
    "prediction_norm": 6.9359025955200195,
    "timestamp": "2025-07-14T12:33:47.574223"
  },
  {
    "epoch": 398,
    "batch_idx": 32,
    "loss": 0.38935065269470215,
    "learning_rate": 0.0009054297904161864,
    "timestep_mean": 475.28125,
    "epsilon_norm": 8.854537963867188,
    "prediction_norm": 7.284171104431152,
    "timestamp": "2025-07-14T12:33:47.579541"
  },
  {
    "epoch": 398,
    "batch_idx": 64,
    "loss": 0.3433353900909424,
    "learning_rate": 0.0009054297904161864,
    "timestep_mean": 480.9375,
    "epsilon_norm": 7.533205509185791,
    "prediction_norm": 5.424672603607178,
    "timestamp": "2025-07-14T12:33:47.584558"
  },
  {
    "epoch": 398,
    "batch_idx": 96,
    "loss": 0.34049171209335327,
    "learning_rate": 0.0009054297904161864,
    "timestep_mean": 481.90625,
    "epsilon_norm": 7.367508888244629,
    "prediction_norm": 6.812683582305908,
    "timestamp": "2025-07-14T12:33:47.589777"
  },
  {
    "epoch": 398,
    "batch_idx": 128,
    "loss": 0.25536346435546875,
    "learning_rate": 0.0009054297904161864,
    "timestep_mean": 463.90625,
    "epsilon_norm": 7.573538780212402,
    "prediction_norm": 6.640827655792236,
    "timestamp": "2025-07-14T12:33:47.594819"
  },
  {
    "epoch": 398,
    "batch_idx": 160,
    "loss": 0.12406647205352783,
    "learning_rate": 0.0009054297904161864,
    "timestep_mean": 568.8125,
    "epsilon_norm": 6.935297966003418,
    "prediction_norm": 6.2088623046875,
    "timestamp": "2025-07-14T12:33:47.600015"
  },
  {
    "epoch": 398,
    "batch_idx": 192,
    "loss": 0.5648609399795532,
    "learning_rate": 0.0009054297904161864,
    "timestep_mean": 470.5,
    "epsilon_norm": 5.02621603012085,
    "prediction_norm": 3.4901626110076904,
    "timestamp": "2025-07-14T12:33:47.605075"
  },
  {
    "epoch": 399,
    "batch_idx": 0,
    "loss": 0.25119540095329285,
    "learning_rate": 0.0009049696434129989,
    "timestep_mean": 560.96875,
    "epsilon_norm": 8.628143310546875,
    "prediction_norm": 7.842659950256348,
    "timestamp": "2025-07-14T12:33:47.610269"
  },
  {
    "epoch": 399,
    "batch_idx": 32,
    "loss": 0.22960379719734192,
    "learning_rate": 0.0009049696434129989,
    "timestep_mean": 470.78125,
    "epsilon_norm": 7.597567558288574,
    "prediction_norm": 7.107944011688232,
    "timestamp": "2025-07-14T12:33:47.615375"
  },
  {
    "epoch": 399,
    "batch_idx": 64,
    "loss": 0.21427462995052338,
    "learning_rate": 0.0009049696434129989,
    "timestep_mean": 596.78125,
    "epsilon_norm": 8.785503387451172,
    "prediction_norm": 8.055252075195312,
    "timestamp": "2025-07-14T12:33:47.620447"
  },
  {
    "epoch": 399,
    "batch_idx": 96,
    "loss": 0.26537322998046875,
    "learning_rate": 0.0009049696434129989,
    "timestep_mean": 477.15625,
    "epsilon_norm": 7.102849006652832,
    "prediction_norm": 5.748462677001953,
    "timestamp": "2025-07-14T12:33:47.625786"
  },
  {
    "epoch": 399,
    "batch_idx": 128,
    "loss": 0.2053455114364624,
    "learning_rate": 0.0009049696434129989,
    "timestep_mean": 511.9375,
    "epsilon_norm": 7.876649856567383,
    "prediction_norm": 6.552175521850586,
    "timestamp": "2025-07-14T12:33:47.630919"
  },
  {
    "epoch": 399,
    "batch_idx": 160,
    "loss": 0.320387065410614,
    "learning_rate": 0.0009049696434129989,
    "timestep_mean": 503.125,
    "epsilon_norm": 8.75728988647461,
    "prediction_norm": 7.022495746612549,
    "timestamp": "2025-07-14T12:33:47.636001"
  },
  {
    "epoch": 399,
    "batch_idx": 192,
    "loss": 0.44102907180786133,
    "learning_rate": 0.0009049696434129989,
    "timestep_mean": 412.75,
    "epsilon_norm": 3.31745982170105,
    "prediction_norm": 2.166949510574341,
    "timestamp": "2025-07-14T12:33:47.641154"
  },
  {
    "epoch": 400,
    "batch_idx": 0,
    "loss": 0.3118014633655548,
    "learning_rate": 0.0009045084971874733,
    "timestep_mean": 513.9375,
    "epsilon_norm": 8.692638397216797,
    "prediction_norm": 7.220771789550781,
    "timestamp": "2025-07-14T12:33:47.646221"
  },
  {
    "epoch": 400,
    "batch_idx": 32,
    "loss": 0.16918203234672546,
    "learning_rate": 0.0009045084971874733,
    "timestep_mean": 556.6875,
    "epsilon_norm": 7.71619176864624,
    "prediction_norm": 6.969061374664307,
    "timestamp": "2025-07-14T12:33:47.651314"
  },
  {
    "epoch": 400,
    "batch_idx": 64,
    "loss": 0.31823164224624634,
    "learning_rate": 0.0009045084971874733,
    "timestep_mean": 538.0625,
    "epsilon_norm": 8.359238624572754,
    "prediction_norm": 6.999870777130127,
    "timestamp": "2025-07-14T12:33:47.656422"
  },
  {
    "epoch": 400,
    "batch_idx": 96,
    "loss": 0.28125232458114624,
    "learning_rate": 0.0009045084971874733,
    "timestep_mean": 516.875,
    "epsilon_norm": 7.848263263702393,
    "prediction_norm": 6.669520854949951,
    "timestamp": "2025-07-14T12:33:47.661359"
  },
  {
    "epoch": 400,
    "batch_idx": 128,
    "loss": 0.3739858567714691,
    "learning_rate": 0.0009045084971874733,
    "timestep_mean": 426.09375,
    "epsilon_norm": 10.13796329498291,
    "prediction_norm": 8.901857376098633,
    "timestamp": "2025-07-14T12:33:47.666326"
  },
  {
    "epoch": 400,
    "batch_idx": 160,
    "loss": 0.44929665327072144,
    "learning_rate": 0.0009045084971874733,
    "timestep_mean": 412.9375,
    "epsilon_norm": 9.076108932495117,
    "prediction_norm": 7.04762601852417,
    "timestamp": "2025-07-14T12:33:47.671557"
  },
  {
    "epoch": 400,
    "batch_idx": 192,
    "loss": 0.35058125853538513,
    "learning_rate": 0.0009045084971874733,
    "timestep_mean": 471.0,
    "epsilon_norm": 4.21588134765625,
    "prediction_norm": 2.6981892585754395,
    "timestamp": "2025-07-14T12:33:47.676773"
  },
  {
    "epoch": 401,
    "batch_idx": 0,
    "loss": 0.28990769386291504,
    "learning_rate": 0.0009040463528774419,
    "timestep_mean": 444.78125,
    "epsilon_norm": 7.425649642944336,
    "prediction_norm": 6.981701850891113,
    "timestamp": "2025-07-14T12:33:47.681934"
  },
  {
    "epoch": 401,
    "batch_idx": 32,
    "loss": 0.2855182886123657,
    "learning_rate": 0.0009040463528774419,
    "timestep_mean": 476.53125,
    "epsilon_norm": 6.709381580352783,
    "prediction_norm": 5.973288059234619,
    "timestamp": "2025-07-14T12:33:47.686993"
  },
  {
    "epoch": 401,
    "batch_idx": 64,
    "loss": 0.2797340750694275,
    "learning_rate": 0.0009040463528774419,
    "timestep_mean": 495.1875,
    "epsilon_norm": 7.909843921661377,
    "prediction_norm": 7.518128395080566,
    "timestamp": "2025-07-14T12:33:47.692240"
  },
  {
    "epoch": 401,
    "batch_idx": 96,
    "loss": 0.2299772948026657,
    "learning_rate": 0.0009040463528774419,
    "timestep_mean": 470.9375,
    "epsilon_norm": 6.78867769241333,
    "prediction_norm": 6.529448509216309,
    "timestamp": "2025-07-14T12:33:47.697332"
  },
  {
    "epoch": 401,
    "batch_idx": 128,
    "loss": 0.3524532914161682,
    "learning_rate": 0.0009040463528774419,
    "timestep_mean": 482.34375,
    "epsilon_norm": 8.214083671569824,
    "prediction_norm": 7.935303211212158,
    "timestamp": "2025-07-14T12:33:47.702905"
  },
  {
    "epoch": 401,
    "batch_idx": 160,
    "loss": 0.18418481945991516,
    "learning_rate": 0.0009040463528774419,
    "timestep_mean": 609.5625,
    "epsilon_norm": 8.16684341430664,
    "prediction_norm": 8.16152572631836,
    "timestamp": "2025-07-14T12:33:47.707894"
  },
  {
    "epoch": 401,
    "batch_idx": 192,
    "loss": 0.5426429510116577,
    "learning_rate": 0.0009040463528774419,
    "timestep_mean": 476.625,
    "epsilon_norm": 4.0121283531188965,
    "prediction_norm": 2.42494797706604,
    "timestamp": "2025-07-14T12:33:47.712934"
  },
  {
    "epoch": 402,
    "batch_idx": 0,
    "loss": 0.4351423978805542,
    "learning_rate": 0.0009035832116231998,
    "timestep_mean": 504.46875,
    "epsilon_norm": 7.667496681213379,
    "prediction_norm": 5.942230701446533,
    "timestamp": "2025-07-14T12:33:47.717882"
  },
  {
    "epoch": 402,
    "batch_idx": 32,
    "loss": 0.40461528301239014,
    "learning_rate": 0.0009035832116231998,
    "timestep_mean": 463.875,
    "epsilon_norm": 8.222492218017578,
    "prediction_norm": 6.62135648727417,
    "timestamp": "2025-07-14T12:33:47.722749"
  },
  {
    "epoch": 402,
    "batch_idx": 64,
    "loss": 0.37393829226493835,
    "learning_rate": 0.0009035832116231998,
    "timestep_mean": 560.25,
    "epsilon_norm": 7.669863700866699,
    "prediction_norm": 5.784738063812256,
    "timestamp": "2025-07-14T12:33:47.727727"
  },
  {
    "epoch": 402,
    "batch_idx": 96,
    "loss": 0.25021061301231384,
    "learning_rate": 0.0009035832116231998,
    "timestep_mean": 536.375,
    "epsilon_norm": 8.1422700881958,
    "prediction_norm": 6.907384395599365,
    "timestamp": "2025-07-14T12:33:47.732717"
  },
  {
    "epoch": 402,
    "batch_idx": 128,
    "loss": 0.45174556970596313,
    "learning_rate": 0.0009035832116231998,
    "timestep_mean": 482.28125,
    "epsilon_norm": 9.113763809204102,
    "prediction_norm": 7.135798931121826,
    "timestamp": "2025-07-14T12:33:47.737763"
  },
  {
    "epoch": 402,
    "batch_idx": 160,
    "loss": 0.30229276418685913,
    "learning_rate": 0.0009035832116231998,
    "timestep_mean": 541.0625,
    "epsilon_norm": 7.22559118270874,
    "prediction_norm": 6.270677089691162,
    "timestamp": "2025-07-14T12:33:47.742662"
  },
  {
    "epoch": 402,
    "batch_idx": 192,
    "loss": 0.570026695728302,
    "learning_rate": 0.0009035832116231998,
    "timestep_mean": 504.625,
    "epsilon_norm": 3.56538462638855,
    "prediction_norm": 2.7269296646118164,
    "timestamp": "2025-07-14T12:33:47.747681"
  },
  {
    "epoch": 403,
    "batch_idx": 0,
    "loss": 0.4414196014404297,
    "learning_rate": 0.0009031190745675021,
    "timestep_mean": 501.28125,
    "epsilon_norm": 8.122089385986328,
    "prediction_norm": 5.507215976715088,
    "timestamp": "2025-07-14T12:33:47.752553"
  },
  {
    "epoch": 403,
    "batch_idx": 32,
    "loss": 0.41208702325820923,
    "learning_rate": 0.0009031190745675021,
    "timestep_mean": 434.96875,
    "epsilon_norm": 7.987333297729492,
    "prediction_norm": 5.863386631011963,
    "timestamp": "2025-07-14T12:33:47.757560"
  },
  {
    "epoch": 403,
    "batch_idx": 64,
    "loss": 0.34564152359962463,
    "learning_rate": 0.0009031190745675021,
    "timestep_mean": 429.78125,
    "epsilon_norm": 7.884695529937744,
    "prediction_norm": 6.063139915466309,
    "timestamp": "2025-07-14T12:33:47.762642"
  },
  {
    "epoch": 403,
    "batch_idx": 96,
    "loss": 0.2016994059085846,
    "learning_rate": 0.0009031190745675021,
    "timestep_mean": 479.0,
    "epsilon_norm": 7.044763565063477,
    "prediction_norm": 6.277874946594238,
    "timestamp": "2025-07-14T12:33:47.767614"
  },
  {
    "epoch": 403,
    "batch_idx": 128,
    "loss": 0.20424354076385498,
    "learning_rate": 0.0009031190745675021,
    "timestep_mean": 529.59375,
    "epsilon_norm": 7.224498748779297,
    "prediction_norm": 6.195442199707031,
    "timestamp": "2025-07-14T12:33:47.772570"
  },
  {
    "epoch": 403,
    "batch_idx": 160,
    "loss": 0.22479690611362457,
    "learning_rate": 0.0009031190745675021,
    "timestep_mean": 461.65625,
    "epsilon_norm": 8.182451248168945,
    "prediction_norm": 6.715845108032227,
    "timestamp": "2025-07-14T12:33:47.777511"
  },
  {
    "epoch": 403,
    "batch_idx": 192,
    "loss": 0.24952906370162964,
    "learning_rate": 0.0009031190745675021,
    "timestep_mean": 544.375,
    "epsilon_norm": 4.528454780578613,
    "prediction_norm": 3.3773088455200195,
    "timestamp": "2025-07-14T12:33:47.782425"
  },
  {
    "epoch": 404,
    "batch_idx": 0,
    "loss": 0.27480942010879517,
    "learning_rate": 0.0009026539428555606,
    "timestep_mean": 490.28125,
    "epsilon_norm": 8.044146537780762,
    "prediction_norm": 6.509572982788086,
    "timestamp": "2025-07-14T12:33:47.787396"
  },
  {
    "epoch": 404,
    "batch_idx": 32,
    "loss": 0.22624097764492035,
    "learning_rate": 0.0009026539428555606,
    "timestep_mean": 562.28125,
    "epsilon_norm": 8.114245414733887,
    "prediction_norm": 6.831114292144775,
    "timestamp": "2025-07-14T12:33:47.792475"
  },
  {
    "epoch": 404,
    "batch_idx": 64,
    "loss": 0.30680882930755615,
    "learning_rate": 0.0009026539428555606,
    "timestep_mean": 524.15625,
    "epsilon_norm": 9.1437349319458,
    "prediction_norm": 7.64793586730957,
    "timestamp": "2025-07-14T12:33:47.797442"
  },
  {
    "epoch": 404,
    "batch_idx": 96,
    "loss": 0.19409441947937012,
    "learning_rate": 0.0009026539428555606,
    "timestep_mean": 576.0625,
    "epsilon_norm": 8.650134086608887,
    "prediction_norm": 7.813859939575195,
    "timestamp": "2025-07-14T12:33:47.802413"
  },
  {
    "epoch": 404,
    "batch_idx": 128,
    "loss": 0.5687633156776428,
    "learning_rate": 0.0009026539428555606,
    "timestep_mean": 430.4375,
    "epsilon_norm": 8.66015911102295,
    "prediction_norm": 6.527499198913574,
    "timestamp": "2025-07-14T12:33:47.807482"
  },
  {
    "epoch": 404,
    "batch_idx": 160,
    "loss": 0.20854619145393372,
    "learning_rate": 0.0009026539428555606,
    "timestep_mean": 514.6875,
    "epsilon_norm": 7.595152378082275,
    "prediction_norm": 6.448770523071289,
    "timestamp": "2025-07-14T12:33:47.812464"
  },
  {
    "epoch": 404,
    "batch_idx": 192,
    "loss": 0.31845730543136597,
    "learning_rate": 0.0009026539428555606,
    "timestep_mean": 290.375,
    "epsilon_norm": 3.768281936645508,
    "prediction_norm": 2.739941120147705,
    "timestamp": "2025-07-14T12:33:47.817497"
  },
  {
    "epoch": 405,
    "batch_idx": 0,
    "loss": 0.2045854926109314,
    "learning_rate": 0.0009021878176350419,
    "timestep_mean": 540.84375,
    "epsilon_norm": 8.425310134887695,
    "prediction_norm": 7.498685359954834,
    "timestamp": "2025-07-14T12:33:47.822574"
  },
  {
    "epoch": 405,
    "batch_idx": 32,
    "loss": 0.42195653915405273,
    "learning_rate": 0.0009021878176350419,
    "timestep_mean": 495.625,
    "epsilon_norm": 8.197935104370117,
    "prediction_norm": 6.437475204467773,
    "timestamp": "2025-07-14T12:33:47.827631"
  },
  {
    "epoch": 405,
    "batch_idx": 64,
    "loss": 0.32936352491378784,
    "learning_rate": 0.0009021878176350419,
    "timestep_mean": 447.78125,
    "epsilon_norm": 8.697378158569336,
    "prediction_norm": 7.28953218460083,
    "timestamp": "2025-07-14T12:33:47.832581"
  },
  {
    "epoch": 405,
    "batch_idx": 96,
    "loss": 0.42636770009994507,
    "learning_rate": 0.0009021878176350419,
    "timestep_mean": 444.15625,
    "epsilon_norm": 8.167937278747559,
    "prediction_norm": 6.073695182800293,
    "timestamp": "2025-07-14T12:33:47.837907"
  },
  {
    "epoch": 405,
    "batch_idx": 128,
    "loss": 0.24810495972633362,
    "learning_rate": 0.0009021878176350419,
    "timestep_mean": 534.71875,
    "epsilon_norm": 8.811448097229004,
    "prediction_norm": 7.608856678009033,
    "timestamp": "2025-07-14T12:33:47.843002"
  },
  {
    "epoch": 405,
    "batch_idx": 160,
    "loss": 0.3391626477241516,
    "learning_rate": 0.0009021878176350419,
    "timestep_mean": 455.15625,
    "epsilon_norm": 8.000920295715332,
    "prediction_norm": 6.265882968902588,
    "timestamp": "2025-07-14T12:33:47.848170"
  },
  {
    "epoch": 405,
    "batch_idx": 192,
    "loss": 0.4418880343437195,
    "learning_rate": 0.0009021878176350419,
    "timestep_mean": 402.875,
    "epsilon_norm": 2.9502952098846436,
    "prediction_norm": 1.717754602432251,
    "timestamp": "2025-07-14T12:33:47.853359"
  },
  {
    "epoch": 406,
    "batch_idx": 0,
    "loss": 0.32550048828125,
    "learning_rate": 0.0009017207000560635,
    "timestep_mean": 487.75,
    "epsilon_norm": 7.818685531616211,
    "prediction_norm": 6.830154895782471,
    "timestamp": "2025-07-14T12:33:47.858661"
  },
  {
    "epoch": 406,
    "batch_idx": 32,
    "loss": 0.2186683565378189,
    "learning_rate": 0.0009017207000560635,
    "timestep_mean": 525.21875,
    "epsilon_norm": 7.933127403259277,
    "prediction_norm": 7.04520320892334,
    "timestamp": "2025-07-14T12:33:47.863796"
  },
  {
    "epoch": 406,
    "batch_idx": 64,
    "loss": 0.3492003083229065,
    "learning_rate": 0.0009017207000560635,
    "timestep_mean": 439.65625,
    "epsilon_norm": 7.553896427154541,
    "prediction_norm": 6.423619747161865,
    "timestamp": "2025-07-14T12:33:47.869035"
  },
  {
    "epoch": 406,
    "batch_idx": 96,
    "loss": 0.4512191414833069,
    "learning_rate": 0.0009017207000560635,
    "timestep_mean": 421.5,
    "epsilon_norm": 8.97329044342041,
    "prediction_norm": 7.398744106292725,
    "timestamp": "2025-07-14T12:33:47.874130"
  },
  {
    "epoch": 406,
    "batch_idx": 128,
    "loss": 0.4905403256416321,
    "learning_rate": 0.0009017207000560635,
    "timestep_mean": 505.9375,
    "epsilon_norm": 8.446196556091309,
    "prediction_norm": 6.593331336975098,
    "timestamp": "2025-07-14T12:33:47.879283"
  },
  {
    "epoch": 406,
    "batch_idx": 160,
    "loss": 0.377943754196167,
    "learning_rate": 0.0009017207000560635,
    "timestep_mean": 561.28125,
    "epsilon_norm": 8.885019302368164,
    "prediction_norm": 7.368152141571045,
    "timestamp": "2025-07-14T12:33:47.884363"
  },
  {
    "epoch": 406,
    "batch_idx": 192,
    "loss": 0.6793354153633118,
    "learning_rate": 0.0009017207000560635,
    "timestep_mean": 423.875,
    "epsilon_norm": 4.798354148864746,
    "prediction_norm": 3.391151189804077,
    "timestamp": "2025-07-14T12:33:47.889479"
  },
  {
    "epoch": 407,
    "batch_idx": 0,
    "loss": 0.24643683433532715,
    "learning_rate": 0.0009012525912711916,
    "timestep_mean": 509.65625,
    "epsilon_norm": 7.559593677520752,
    "prediction_norm": 6.448104381561279,
    "timestamp": "2025-07-14T12:33:47.894665"
  },
  {
    "epoch": 407,
    "batch_idx": 32,
    "loss": 0.25868672132492065,
    "learning_rate": 0.0009012525912711916,
    "timestep_mean": 555.40625,
    "epsilon_norm": 6.992212772369385,
    "prediction_norm": 5.26556921005249,
    "timestamp": "2025-07-14T12:33:47.899911"
  },
  {
    "epoch": 407,
    "batch_idx": 64,
    "loss": 0.2954922914505005,
    "learning_rate": 0.0009012525912711916,
    "timestep_mean": 408.5625,
    "epsilon_norm": 7.652658939361572,
    "prediction_norm": 5.7209649085998535,
    "timestamp": "2025-07-14T12:33:47.904987"
  },
  {
    "epoch": 407,
    "batch_idx": 96,
    "loss": 0.5262587070465088,
    "learning_rate": 0.0009012525912711916,
    "timestep_mean": 524.59375,
    "epsilon_norm": 8.358901977539062,
    "prediction_norm": 5.957059860229492,
    "timestamp": "2025-07-14T12:33:47.910128"
  },
  {
    "epoch": 407,
    "batch_idx": 128,
    "loss": 0.5361099243164062,
    "learning_rate": 0.0009012525912711916,
    "timestep_mean": 382.46875,
    "epsilon_norm": 8.339980125427246,
    "prediction_norm": 5.2477335929870605,
    "timestamp": "2025-07-14T12:33:47.915301"
  },
  {
    "epoch": 407,
    "batch_idx": 160,
    "loss": 0.23424196243286133,
    "learning_rate": 0.0009012525912711916,
    "timestep_mean": 465.0625,
    "epsilon_norm": 8.058367729187012,
    "prediction_norm": 6.436001300811768,
    "timestamp": "2025-07-14T12:33:47.920437"
  },
  {
    "epoch": 407,
    "batch_idx": 192,
    "loss": 0.4364120364189148,
    "learning_rate": 0.0009012525912711916,
    "timestep_mean": 544.75,
    "epsilon_norm": 3.9026145935058594,
    "prediction_norm": 2.7733705043792725,
    "timestamp": "2025-07-14T12:33:47.925744"
  },
  {
    "epoch": 408,
    "batch_idx": 0,
    "loss": 0.49489325284957886,
    "learning_rate": 0.0009007834924354381,
    "timestep_mean": 431.0625,
    "epsilon_norm": 8.861262321472168,
    "prediction_norm": 6.719475746154785,
    "timestamp": "2025-07-14T12:33:47.930890"
  },
  {
    "epoch": 408,
    "batch_idx": 32,
    "loss": 0.33567142486572266,
    "learning_rate": 0.0009007834924354381,
    "timestep_mean": 526.875,
    "epsilon_norm": 8.033307075500488,
    "prediction_norm": 6.666247367858887,
    "timestamp": "2025-07-14T12:33:47.936032"
  },
  {
    "epoch": 408,
    "batch_idx": 64,
    "loss": 0.36276134848594666,
    "learning_rate": 0.0009007834924354381,
    "timestep_mean": 515.8125,
    "epsilon_norm": 7.851948261260986,
    "prediction_norm": 6.591492176055908,
    "timestamp": "2025-07-14T12:33:47.941197"
  },
  {
    "epoch": 408,
    "batch_idx": 96,
    "loss": 0.43483203649520874,
    "learning_rate": 0.0009007834924354381,
    "timestep_mean": 529.0,
    "epsilon_norm": 8.908193588256836,
    "prediction_norm": 7.4663801193237305,
    "timestamp": "2025-07-14T12:33:47.946486"
  },
  {
    "epoch": 408,
    "batch_idx": 128,
    "loss": 0.23505061864852905,
    "learning_rate": 0.0009007834924354381,
    "timestep_mean": 531.28125,
    "epsilon_norm": 8.184144973754883,
    "prediction_norm": 7.570288181304932,
    "timestamp": "2025-07-14T12:33:47.951684"
  },
  {
    "epoch": 408,
    "batch_idx": 160,
    "loss": 0.3156466484069824,
    "learning_rate": 0.0009007834924354381,
    "timestep_mean": 459.8125,
    "epsilon_norm": 8.128935813903809,
    "prediction_norm": 5.840479373931885,
    "timestamp": "2025-07-14T12:33:47.956861"
  },
  {
    "epoch": 408,
    "batch_idx": 192,
    "loss": 0.19046202301979065,
    "learning_rate": 0.0009007834924354381,
    "timestep_mean": 513.625,
    "epsilon_norm": 4.1478190422058105,
    "prediction_norm": 4.796164035797119,
    "timestamp": "2025-07-14T12:33:47.961903"
  },
  {
    "epoch": 409,
    "batch_idx": 0,
    "loss": 0.34366801381111145,
    "learning_rate": 0.0009003134047062577,
    "timestep_mean": 502.15625,
    "epsilon_norm": 7.160341739654541,
    "prediction_norm": 5.930466651916504,
    "timestamp": "2025-07-14T12:33:47.967343"
  },
  {
    "epoch": 409,
    "batch_idx": 32,
    "loss": 0.4514942169189453,
    "learning_rate": 0.0009003134047062577,
    "timestep_mean": 450.46875,
    "epsilon_norm": 8.892741203308105,
    "prediction_norm": 7.746070384979248,
    "timestamp": "2025-07-14T12:33:47.972490"
  },
  {
    "epoch": 409,
    "batch_idx": 64,
    "loss": 0.1663857400417328,
    "learning_rate": 0.0009003134047062577,
    "timestep_mean": 493.625,
    "epsilon_norm": 7.48400354385376,
    "prediction_norm": 6.4328131675720215,
    "timestamp": "2025-07-14T12:33:47.977668"
  },
  {
    "epoch": 409,
    "batch_idx": 96,
    "loss": 0.39392828941345215,
    "learning_rate": 0.0009003134047062577,
    "timestep_mean": 389.875,
    "epsilon_norm": 7.599849700927734,
    "prediction_norm": 6.011190414428711,
    "timestamp": "2025-07-14T12:33:47.982830"
  },
  {
    "epoch": 409,
    "batch_idx": 128,
    "loss": 0.34291619062423706,
    "learning_rate": 0.0009003134047062577,
    "timestep_mean": 462.15625,
    "epsilon_norm": 8.008703231811523,
    "prediction_norm": 6.201752185821533,
    "timestamp": "2025-07-14T12:33:47.988046"
  },
  {
    "epoch": 409,
    "batch_idx": 160,
    "loss": 0.39517146348953247,
    "learning_rate": 0.0009003134047062577,
    "timestep_mean": 457.25,
    "epsilon_norm": 8.100861549377441,
    "prediction_norm": 5.808765888214111,
    "timestamp": "2025-07-14T12:33:47.993259"
  },
  {
    "epoch": 409,
    "batch_idx": 192,
    "loss": 0.06236943230032921,
    "learning_rate": 0.0009003134047062577,
    "timestep_mean": 644.125,
    "epsilon_norm": 3.6815106868743896,
    "prediction_norm": 3.4843599796295166,
    "timestamp": "2025-07-14T12:33:47.998341"
  },
  {
    "epoch": 410,
    "batch_idx": 0,
    "loss": 0.18256354331970215,
    "learning_rate": 0.0008998423292435452,
    "timestep_mean": 497.09375,
    "epsilon_norm": 6.725271224975586,
    "prediction_norm": 5.341047763824463,
    "timestamp": "2025-07-14T12:33:48.003591"
  },
  {
    "epoch": 410,
    "batch_idx": 32,
    "loss": 0.30810797214508057,
    "learning_rate": 0.0008998423292435452,
    "timestep_mean": 481.21875,
    "epsilon_norm": 8.017989158630371,
    "prediction_norm": 6.662903785705566,
    "timestamp": "2025-07-14T12:33:48.008721"
  },
  {
    "epoch": 410,
    "batch_idx": 64,
    "loss": 0.2983311712741852,
    "learning_rate": 0.0008998423292435452,
    "timestep_mean": 493.0,
    "epsilon_norm": 8.558832168579102,
    "prediction_norm": 7.762862682342529,
    "timestamp": "2025-07-14T12:33:48.013811"
  },
  {
    "epoch": 410,
    "batch_idx": 96,
    "loss": 0.21253064274787903,
    "learning_rate": 0.0008998423292435452,
    "timestep_mean": 534.0,
    "epsilon_norm": 7.934514999389648,
    "prediction_norm": 6.280016899108887,
    "timestamp": "2025-07-14T12:33:48.018987"
  },
  {
    "epoch": 410,
    "batch_idx": 128,
    "loss": 0.28772860765457153,
    "learning_rate": 0.0008998423292435452,
    "timestep_mean": 464.4375,
    "epsilon_norm": 8.844560623168945,
    "prediction_norm": 7.019192218780518,
    "timestamp": "2025-07-14T12:33:48.024091"
  },
  {
    "epoch": 410,
    "batch_idx": 160,
    "loss": 0.18144303560256958,
    "learning_rate": 0.0008998423292435452,
    "timestep_mean": 474.4375,
    "epsilon_norm": 7.051270008087158,
    "prediction_norm": 6.339920997619629,
    "timestamp": "2025-07-14T12:33:48.029157"
  },
  {
    "epoch": 410,
    "batch_idx": 192,
    "loss": 0.12293518334627151,
    "learning_rate": 0.0008998423292435452,
    "timestep_mean": 765.125,
    "epsilon_norm": 3.1020352840423584,
    "prediction_norm": 2.497999429702759,
    "timestamp": "2025-07-14T12:33:48.034159"
  },
  {
    "epoch": 411,
    "batch_idx": 0,
    "loss": 0.35816317796707153,
    "learning_rate": 0.0008993702672096323,
    "timestep_mean": 491.9375,
    "epsilon_norm": 7.427440643310547,
    "prediction_norm": 5.335704803466797,
    "timestamp": "2025-07-14T12:33:48.039297"
  },
  {
    "epoch": 411,
    "batch_idx": 32,
    "loss": 0.3852503001689911,
    "learning_rate": 0.0008993702672096323,
    "timestep_mean": 492.1875,
    "epsilon_norm": 7.670564651489258,
    "prediction_norm": 6.3408203125,
    "timestamp": "2025-07-14T12:33:48.044564"
  },
  {
    "epoch": 411,
    "batch_idx": 64,
    "loss": 0.28818145394325256,
    "learning_rate": 0.0008993702672096323,
    "timestep_mean": 512.125,
    "epsilon_norm": 8.724568367004395,
    "prediction_norm": 7.315159320831299,
    "timestamp": "2025-07-14T12:33:48.049855"
  },
  {
    "epoch": 411,
    "batch_idx": 96,
    "loss": 0.4004572033882141,
    "learning_rate": 0.0008993702672096323,
    "timestep_mean": 501.53125,
    "epsilon_norm": 8.07726764678955,
    "prediction_norm": 6.871872901916504,
    "timestamp": "2025-07-14T12:33:48.055057"
  },
  {
    "epoch": 411,
    "batch_idx": 128,
    "loss": 0.3154934346675873,
    "learning_rate": 0.0008993702672096323,
    "timestep_mean": 465.59375,
    "epsilon_norm": 6.992151260375977,
    "prediction_norm": 6.717059135437012,
    "timestamp": "2025-07-14T12:33:48.060264"
  },
  {
    "epoch": 411,
    "batch_idx": 160,
    "loss": 0.14280839264392853,
    "learning_rate": 0.0008993702672096323,
    "timestep_mean": 554.09375,
    "epsilon_norm": 7.2027482986450195,
    "prediction_norm": 7.0705084800720215,
    "timestamp": "2025-07-14T12:33:48.065362"
  },
  {
    "epoch": 411,
    "batch_idx": 192,
    "loss": 0.3084966838359833,
    "learning_rate": 0.0008993702672096323,
    "timestep_mean": 366.5,
    "epsilon_norm": 2.827697992324829,
    "prediction_norm": 2.353602647781372,
    "timestamp": "2025-07-14T12:33:48.070506"
  },
  {
    "epoch": 412,
    "batch_idx": 0,
    "loss": 0.27187496423721313,
    "learning_rate": 0.0008988972197692854,
    "timestep_mean": 491.875,
    "epsilon_norm": 7.900460243225098,
    "prediction_norm": 7.4803571701049805,
    "timestamp": "2025-07-14T12:33:48.075742"
  },
  {
    "epoch": 412,
    "batch_idx": 32,
    "loss": 0.20019562542438507,
    "learning_rate": 0.0008988972197692854,
    "timestep_mean": 424.03125,
    "epsilon_norm": 7.232812881469727,
    "prediction_norm": 6.485391616821289,
    "timestamp": "2025-07-14T12:33:48.080821"
  },
  {
    "epoch": 412,
    "batch_idx": 64,
    "loss": 0.32948654890060425,
    "learning_rate": 0.0008988972197692854,
    "timestep_mean": 544.8125,
    "epsilon_norm": 7.754676818847656,
    "prediction_norm": 5.629140377044678,
    "timestamp": "2025-07-14T12:33:48.086071"
  },
  {
    "epoch": 412,
    "batch_idx": 96,
    "loss": 0.34161752462387085,
    "learning_rate": 0.0008988972197692854,
    "timestep_mean": 436.4375,
    "epsilon_norm": 6.98786735534668,
    "prediction_norm": 6.023682117462158,
    "timestamp": "2025-07-14T12:33:48.091029"
  },
  {
    "epoch": 412,
    "batch_idx": 128,
    "loss": 0.316529244184494,
    "learning_rate": 0.0008988972197692854,
    "timestep_mean": 505.40625,
    "epsilon_norm": 7.80363655090332,
    "prediction_norm": 6.075523853302002,
    "timestamp": "2025-07-14T12:33:48.095890"
  },
  {
    "epoch": 412,
    "batch_idx": 160,
    "loss": 0.14801296591758728,
    "learning_rate": 0.0008988972197692854,
    "timestep_mean": 556.75,
    "epsilon_norm": 8.31097412109375,
    "prediction_norm": 7.624771595001221,
    "timestamp": "2025-07-14T12:33:48.100764"
  },
  {
    "epoch": 412,
    "batch_idx": 192,
    "loss": 0.46149832010269165,
    "learning_rate": 0.0008988972197692854,
    "timestep_mean": 471.0,
    "epsilon_norm": 3.9954183101654053,
    "prediction_norm": 2.8909475803375244,
    "timestamp": "2025-07-14T12:33:48.105566"
  },
  {
    "epoch": 413,
    "batch_idx": 0,
    "loss": 0.48045557737350464,
    "learning_rate": 0.0008984231880897019,
    "timestep_mean": 453.40625,
    "epsilon_norm": 10.158014297485352,
    "prediction_norm": 8.053606986999512,
    "timestamp": "2025-07-14T12:33:48.110501"
  },
  {
    "epoch": 413,
    "batch_idx": 32,
    "loss": 0.20909002423286438,
    "learning_rate": 0.0008984231880897019,
    "timestep_mean": 473.65625,
    "epsilon_norm": 7.245824813842773,
    "prediction_norm": 6.287201404571533,
    "timestamp": "2025-07-14T12:33:48.115495"
  },
  {
    "epoch": 413,
    "batch_idx": 64,
    "loss": 0.3818513751029968,
    "learning_rate": 0.0008984231880897019,
    "timestep_mean": 522.625,
    "epsilon_norm": 7.474605560302734,
    "prediction_norm": 5.443531513214111,
    "timestamp": "2025-07-14T12:33:48.120592"
  },
  {
    "epoch": 413,
    "batch_idx": 96,
    "loss": 0.21549761295318604,
    "learning_rate": 0.0008984231880897019,
    "timestep_mean": 510.9375,
    "epsilon_norm": 7.9008965492248535,
    "prediction_norm": 7.338656425476074,
    "timestamp": "2025-07-14T12:33:48.125694"
  },
  {
    "epoch": 413,
    "batch_idx": 128,
    "loss": 0.26731356978416443,
    "learning_rate": 0.0008984231880897019,
    "timestep_mean": 498.34375,
    "epsilon_norm": 7.036350727081299,
    "prediction_norm": 5.780038833618164,
    "timestamp": "2025-07-14T12:33:48.130806"
  },
  {
    "epoch": 413,
    "batch_idx": 160,
    "loss": 0.2515662908554077,
    "learning_rate": 0.0008984231880897019,
    "timestep_mean": 547.4375,
    "epsilon_norm": 7.68224573135376,
    "prediction_norm": 6.2953643798828125,
    "timestamp": "2025-07-14T12:33:48.135901"
  },
  {
    "epoch": 413,
    "batch_idx": 192,
    "loss": 0.15789580345153809,
    "learning_rate": 0.0008984231880897019,
    "timestep_mean": 444.875,
    "epsilon_norm": 3.8568203449249268,
    "prediction_norm": 3.146486759185791,
    "timestamp": "2025-07-14T12:33:48.141001"
  },
  {
    "epoch": 414,
    "batch_idx": 0,
    "loss": 0.2137460559606552,
    "learning_rate": 0.0008979481733405078,
    "timestep_mean": 563.5,
    "epsilon_norm": 7.865836143493652,
    "prediction_norm": 7.23440408706665,
    "timestamp": "2025-07-14T12:33:48.146124"
  },
  {
    "epoch": 414,
    "batch_idx": 32,
    "loss": 0.23416531085968018,
    "learning_rate": 0.0008979481733405078,
    "timestep_mean": 526.1875,
    "epsilon_norm": 8.204627990722656,
    "prediction_norm": 7.353928089141846,
    "timestamp": "2025-07-14T12:33:48.151160"
  },
  {
    "epoch": 414,
    "batch_idx": 64,
    "loss": 0.3563404083251953,
    "learning_rate": 0.0008979481733405078,
    "timestep_mean": 587.90625,
    "epsilon_norm": 8.42585277557373,
    "prediction_norm": 6.466163158416748,
    "timestamp": "2025-07-14T12:33:48.156327"
  },
  {
    "epoch": 414,
    "batch_idx": 96,
    "loss": 0.4461548328399658,
    "learning_rate": 0.0008979481733405078,
    "timestep_mean": 522.5,
    "epsilon_norm": 8.430318832397461,
    "prediction_norm": 5.4970784187316895,
    "timestamp": "2025-07-14T12:33:48.161455"
  },
  {
    "epoch": 414,
    "batch_idx": 128,
    "loss": 0.39047378301620483,
    "learning_rate": 0.0008979481733405078,
    "timestep_mean": 539.6875,
    "epsilon_norm": 9.255330085754395,
    "prediction_norm": 7.500912189483643,
    "timestamp": "2025-07-14T12:33:48.166474"
  },
  {
    "epoch": 414,
    "batch_idx": 160,
    "loss": 0.3651798367500305,
    "learning_rate": 0.0008979481733405078,
    "timestep_mean": 438.21875,
    "epsilon_norm": 7.414696216583252,
    "prediction_norm": 5.952286720275879,
    "timestamp": "2025-07-14T12:33:48.171510"
  },
  {
    "epoch": 414,
    "batch_idx": 192,
    "loss": 0.20604348182678223,
    "learning_rate": 0.0008979481733405078,
    "timestep_mean": 603.5,
    "epsilon_norm": 3.2057113647460938,
    "prediction_norm": 3.3848135471343994,
    "timestamp": "2025-07-14T12:33:48.176520"
  },
  {
    "epoch": 415,
    "batch_idx": 0,
    "loss": 0.38866615295410156,
    "learning_rate": 0.000897472176693755,
    "timestep_mean": 467.5,
    "epsilon_norm": 8.170816421508789,
    "prediction_norm": 6.709630966186523,
    "timestamp": "2025-07-14T12:33:48.181553"
  },
  {
    "epoch": 415,
    "batch_idx": 32,
    "loss": 0.23421820998191833,
    "learning_rate": 0.000897472176693755,
    "timestep_mean": 556.90625,
    "epsilon_norm": 7.685722351074219,
    "prediction_norm": 6.8326520919799805,
    "timestamp": "2025-07-14T12:33:48.186779"
  },
  {
    "epoch": 415,
    "batch_idx": 64,
    "loss": 0.27421605587005615,
    "learning_rate": 0.000897472176693755,
    "timestep_mean": 459.25,
    "epsilon_norm": 7.665120601654053,
    "prediction_norm": 6.337751865386963,
    "timestamp": "2025-07-14T12:33:48.192044"
  },
  {
    "epoch": 415,
    "batch_idx": 96,
    "loss": 0.17570027709007263,
    "learning_rate": 0.000897472176693755,
    "timestep_mean": 549.5625,
    "epsilon_norm": 7.303263187408447,
    "prediction_norm": 7.380789756774902,
    "timestamp": "2025-07-14T12:33:48.197210"
  },
  {
    "epoch": 415,
    "batch_idx": 128,
    "loss": 0.41035544872283936,
    "learning_rate": 0.000897472176693755,
    "timestep_mean": 413.59375,
    "epsilon_norm": 8.266877174377441,
    "prediction_norm": 6.575463771820068,
    "timestamp": "2025-07-14T12:33:48.202352"
  },
  {
    "epoch": 415,
    "batch_idx": 160,
    "loss": 0.3403615951538086,
    "learning_rate": 0.000897472176693755,
    "timestep_mean": 509.53125,
    "epsilon_norm": 8.788643836975098,
    "prediction_norm": 8.143569946289062,
    "timestamp": "2025-07-14T12:33:48.207556"
  },
  {
    "epoch": 415,
    "batch_idx": 192,
    "loss": 0.689681351184845,
    "learning_rate": 0.000897472176693755,
    "timestep_mean": 442.625,
    "epsilon_norm": 4.098809242248535,
    "prediction_norm": 2.7410576343536377,
    "timestamp": "2025-07-14T12:33:48.212754"
  },
  {
    "epoch": 416,
    "batch_idx": 0,
    "loss": 0.3947261571884155,
    "learning_rate": 0.0008969951993239175,
    "timestep_mean": 471.1875,
    "epsilon_norm": 7.628300666809082,
    "prediction_norm": 6.191385746002197,
    "timestamp": "2025-07-14T12:33:48.217909"
  },
  {
    "epoch": 416,
    "batch_idx": 32,
    "loss": 0.4069090783596039,
    "learning_rate": 0.0008969951993239175,
    "timestep_mean": 435.59375,
    "epsilon_norm": 8.340414047241211,
    "prediction_norm": 7.306901931762695,
    "timestamp": "2025-07-14T12:33:48.222961"
  },
  {
    "epoch": 416,
    "batch_idx": 64,
    "loss": 0.43529438972473145,
    "learning_rate": 0.0008969951993239175,
    "timestep_mean": 492.5625,
    "epsilon_norm": 8.650203704833984,
    "prediction_norm": 6.837698459625244,
    "timestamp": "2025-07-14T12:33:48.228088"
  },
  {
    "epoch": 416,
    "batch_idx": 96,
    "loss": 0.29527682065963745,
    "learning_rate": 0.0008969951993239175,
    "timestep_mean": 508.75,
    "epsilon_norm": 7.446694374084473,
    "prediction_norm": 6.062970161437988,
    "timestamp": "2025-07-14T12:33:48.233247"
  },
  {
    "epoch": 416,
    "batch_idx": 128,
    "loss": 0.24388878047466278,
    "learning_rate": 0.0008969951993239175,
    "timestep_mean": 509.96875,
    "epsilon_norm": 7.186275482177734,
    "prediction_norm": 6.103186130523682,
    "timestamp": "2025-07-14T12:33:48.238490"
  },
  {
    "epoch": 416,
    "batch_idx": 160,
    "loss": 0.49674296379089355,
    "learning_rate": 0.0008969951993239175,
    "timestep_mean": 425.46875,
    "epsilon_norm": 8.033090591430664,
    "prediction_norm": 5.401211261749268,
    "timestamp": "2025-07-14T12:33:48.243655"
  },
  {
    "epoch": 416,
    "batch_idx": 192,
    "loss": 0.13744494318962097,
    "learning_rate": 0.0008969951993239175,
    "timestep_mean": 414.75,
    "epsilon_norm": 4.411439895629883,
    "prediction_norm": 3.976656198501587,
    "timestamp": "2025-07-14T12:33:48.248737"
  },
  {
    "epoch": 417,
    "batch_idx": 0,
    "loss": 0.31331679224967957,
    "learning_rate": 0.0008965172424078901,
    "timestep_mean": 489.71875,
    "epsilon_norm": 7.151201248168945,
    "prediction_norm": 5.561047077178955,
    "timestamp": "2025-07-14T12:33:48.253925"
  },
  {
    "epoch": 417,
    "batch_idx": 32,
    "loss": 0.2987270951271057,
    "learning_rate": 0.0008965172424078901,
    "timestep_mean": 525.625,
    "epsilon_norm": 7.939846992492676,
    "prediction_norm": 6.241045951843262,
    "timestamp": "2025-07-14T12:33:48.259059"
  },
  {
    "epoch": 417,
    "batch_idx": 64,
    "loss": 0.3221835792064667,
    "learning_rate": 0.0008965172424078901,
    "timestep_mean": 554.125,
    "epsilon_norm": 7.277718544006348,
    "prediction_norm": 5.8313212394714355,
    "timestamp": "2025-07-14T12:33:48.264136"
  },
  {
    "epoch": 417,
    "batch_idx": 96,
    "loss": 0.24498645961284637,
    "learning_rate": 0.0008965172424078901,
    "timestep_mean": 500.4375,
    "epsilon_norm": 6.547336578369141,
    "prediction_norm": 4.830390453338623,
    "timestamp": "2025-07-14T12:33:48.269156"
  },
  {
    "epoch": 417,
    "batch_idx": 128,
    "loss": 0.30826902389526367,
    "learning_rate": 0.0008965172424078901,
    "timestep_mean": 533.375,
    "epsilon_norm": 8.853267669677734,
    "prediction_norm": 6.774858474731445,
    "timestamp": "2025-07-14T12:33:48.274291"
  },
  {
    "epoch": 417,
    "batch_idx": 160,
    "loss": 0.3754478096961975,
    "learning_rate": 0.0008965172424078901,
    "timestep_mean": 505.125,
    "epsilon_norm": 8.197928428649902,
    "prediction_norm": 5.876392841339111,
    "timestamp": "2025-07-14T12:33:48.279553"
  },
  {
    "epoch": 417,
    "batch_idx": 192,
    "loss": 0.1172914206981659,
    "learning_rate": 0.0008965172424078901,
    "timestep_mean": 658.75,
    "epsilon_norm": 5.022191524505615,
    "prediction_norm": 4.587923049926758,
    "timestamp": "2025-07-14T12:33:48.284596"
  },
  {
    "epoch": 418,
    "batch_idx": 0,
    "loss": 0.2379148006439209,
    "learning_rate": 0.0008960383071249835,
    "timestep_mean": 465.0,
    "epsilon_norm": 7.632162570953369,
    "prediction_norm": 6.728840351104736,
    "timestamp": "2025-07-14T12:33:48.289673"
  },
  {
    "epoch": 418,
    "batch_idx": 32,
    "loss": 0.19466640055179596,
    "learning_rate": 0.0008960383071249835,
    "timestep_mean": 520.96875,
    "epsilon_norm": 6.966991901397705,
    "prediction_norm": 5.67929220199585,
    "timestamp": "2025-07-14T12:33:48.294685"
  },
  {
    "epoch": 418,
    "batch_idx": 64,
    "loss": 0.2692403197288513,
    "learning_rate": 0.0008960383071249835,
    "timestep_mean": 535.40625,
    "epsilon_norm": 8.501184463500977,
    "prediction_norm": 7.223643779754639,
    "timestamp": "2025-07-14T12:33:48.302576"
  },
  {
    "epoch": 418,
    "batch_idx": 96,
    "loss": 0.2962683439254761,
    "learning_rate": 0.0008960383071249835,
    "timestep_mean": 466.3125,
    "epsilon_norm": 7.99953556060791,
    "prediction_norm": 6.74130916595459,
    "timestamp": "2025-07-14T12:33:48.307299"
  },
  {
    "epoch": 418,
    "batch_idx": 128,
    "loss": 0.24693672358989716,
    "learning_rate": 0.0008960383071249835,
    "timestep_mean": 512.84375,
    "epsilon_norm": 8.426179885864258,
    "prediction_norm": 7.442724704742432,
    "timestamp": "2025-07-14T12:33:48.312010"
  },
  {
    "epoch": 418,
    "batch_idx": 160,
    "loss": 0.23590604960918427,
    "learning_rate": 0.0008960383071249835,
    "timestep_mean": 557.25,
    "epsilon_norm": 8.357577323913574,
    "prediction_norm": 7.4788689613342285,
    "timestamp": "2025-07-14T12:33:48.316848"
  },
  {
    "epoch": 418,
    "batch_idx": 192,
    "loss": 0.06092226505279541,
    "learning_rate": 0.0008960383071249835,
    "timestep_mean": 743.0,
    "epsilon_norm": 5.018163681030273,
    "prediction_norm": 5.062113285064697,
    "timestamp": "2025-07-14T12:33:48.321529"
  },
  {
    "epoch": 419,
    "batch_idx": 0,
    "loss": 0.3457697629928589,
    "learning_rate": 0.0008955583946569231,
    "timestep_mean": 437.15625,
    "epsilon_norm": 8.17546272277832,
    "prediction_norm": 7.253634452819824,
    "timestamp": "2025-07-14T12:33:48.326229"
  },
  {
    "epoch": 419,
    "batch_idx": 32,
    "loss": 0.229354128241539,
    "learning_rate": 0.0008955583946569231,
    "timestep_mean": 648.625,
    "epsilon_norm": 8.772315979003906,
    "prediction_norm": 8.62431812286377,
    "timestamp": "2025-07-14T12:33:48.330878"
  },
  {
    "epoch": 419,
    "batch_idx": 64,
    "loss": 0.3408880829811096,
    "learning_rate": 0.0008955583946569231,
    "timestep_mean": 384.15625,
    "epsilon_norm": 7.737819671630859,
    "prediction_norm": 6.1097493171691895,
    "timestamp": "2025-07-14T12:33:48.335536"
  },
  {
    "epoch": 419,
    "batch_idx": 96,
    "loss": 0.3148036599159241,
    "learning_rate": 0.0008955583946569231,
    "timestep_mean": 574.21875,
    "epsilon_norm": 7.705609321594238,
    "prediction_norm": 6.237236022949219,
    "timestamp": "2025-07-14T12:33:48.340249"
  },
  {
    "epoch": 419,
    "batch_idx": 128,
    "loss": 0.2755981683731079,
    "learning_rate": 0.0008955583946569231,
    "timestep_mean": 566.75,
    "epsilon_norm": 7.986855506896973,
    "prediction_norm": 7.970716953277588,
    "timestamp": "2025-07-14T12:33:48.344909"
  },
  {
    "epoch": 419,
    "batch_idx": 160,
    "loss": 0.2929280400276184,
    "learning_rate": 0.0008955583946569231,
    "timestep_mean": 565.9375,
    "epsilon_norm": 8.362358093261719,
    "prediction_norm": 8.096583366394043,
    "timestamp": "2025-07-14T12:33:48.349564"
  },
  {
    "epoch": 419,
    "batch_idx": 192,
    "loss": 0.08319956064224243,
    "learning_rate": 0.0008955583946569231,
    "timestep_mean": 523.25,
    "epsilon_norm": 3.90405535697937,
    "prediction_norm": 3.5626981258392334,
    "timestamp": "2025-07-14T12:33:48.354321"
  },
  {
    "epoch": 420,
    "batch_idx": 0,
    "loss": 0.2750675082206726,
    "learning_rate": 0.0008950775061878452,
    "timestep_mean": 503.6875,
    "epsilon_norm": 7.462965488433838,
    "prediction_norm": 6.965181350708008,
    "timestamp": "2025-07-14T12:33:48.359017"
  },
  {
    "epoch": 420,
    "batch_idx": 32,
    "loss": 0.30573657155036926,
    "learning_rate": 0.0008950775061878452,
    "timestep_mean": 480.40625,
    "epsilon_norm": 6.579963684082031,
    "prediction_norm": 5.719186782836914,
    "timestamp": "2025-07-14T12:33:48.363694"
  },
  {
    "epoch": 420,
    "batch_idx": 64,
    "loss": 0.4619656801223755,
    "learning_rate": 0.0008950775061878452,
    "timestep_mean": 439.0625,
    "epsilon_norm": 8.03273868560791,
    "prediction_norm": 5.07749080657959,
    "timestamp": "2025-07-14T12:33:48.368418"
  },
  {
    "epoch": 420,
    "batch_idx": 96,
    "loss": 0.43714654445648193,
    "learning_rate": 0.0008950775061878452,
    "timestep_mean": 519.5625,
    "epsilon_norm": 9.67473316192627,
    "prediction_norm": 7.107696533203125,
    "timestamp": "2025-07-14T12:33:48.373083"
  },
  {
    "epoch": 420,
    "batch_idx": 128,
    "loss": 0.2680095434188843,
    "learning_rate": 0.0008950775061878452,
    "timestep_mean": 439.0625,
    "epsilon_norm": 8.190850257873535,
    "prediction_norm": 6.8232526779174805,
    "timestamp": "2025-07-14T12:33:48.377787"
  },
  {
    "epoch": 420,
    "batch_idx": 160,
    "loss": 0.3293282389640808,
    "learning_rate": 0.0008950775061878452,
    "timestep_mean": 541.4375,
    "epsilon_norm": 8.780762672424316,
    "prediction_norm": 7.061250686645508,
    "timestamp": "2025-07-14T12:33:48.382466"
  },
  {
    "epoch": 420,
    "batch_idx": 192,
    "loss": 0.17008434236049652,
    "learning_rate": 0.0008950775061878452,
    "timestep_mean": 364.0,
    "epsilon_norm": 3.165510654449463,
    "prediction_norm": 3.01802134513855,
    "timestamp": "2025-07-14T12:33:48.387144"
  },
  {
    "epoch": 421,
    "batch_idx": 0,
    "loss": 0.4276816248893738,
    "learning_rate": 0.0008945956429042942,
    "timestep_mean": 470.96875,
    "epsilon_norm": 8.737136840820312,
    "prediction_norm": 6.868564128875732,
    "timestamp": "2025-07-14T12:33:48.391931"
  },
  {
    "epoch": 421,
    "batch_idx": 32,
    "loss": 0.2781282961368561,
    "learning_rate": 0.0008945956429042942,
    "timestep_mean": 467.96875,
    "epsilon_norm": 7.461306095123291,
    "prediction_norm": 6.2190470695495605,
    "timestamp": "2025-07-14T12:33:48.396645"
  },
  {
    "epoch": 421,
    "batch_idx": 64,
    "loss": 0.2566470205783844,
    "learning_rate": 0.0008945956429042942,
    "timestep_mean": 452.34375,
    "epsilon_norm": 6.485964298248291,
    "prediction_norm": 5.4908294677734375,
    "timestamp": "2025-07-14T12:33:48.401360"
  },
  {
    "epoch": 421,
    "batch_idx": 96,
    "loss": 0.28968650102615356,
    "learning_rate": 0.0008945956429042942,
    "timestep_mean": 558.1875,
    "epsilon_norm": 8.565633773803711,
    "prediction_norm": 7.417553901672363,
    "timestamp": "2025-07-14T12:33:48.406048"
  },
  {
    "epoch": 421,
    "batch_idx": 128,
    "loss": 0.5058626532554626,
    "learning_rate": 0.0008945956429042942,
    "timestep_mean": 511.90625,
    "epsilon_norm": 8.483142852783203,
    "prediction_norm": 6.609210014343262,
    "timestamp": "2025-07-14T12:33:48.410750"
  },
  {
    "epoch": 421,
    "batch_idx": 160,
    "loss": 0.2912415564060211,
    "learning_rate": 0.0008945956429042942,
    "timestep_mean": 462.5,
    "epsilon_norm": 8.68756103515625,
    "prediction_norm": 7.904200077056885,
    "timestamp": "2025-07-14T12:33:48.415901"
  },
  {
    "epoch": 421,
    "batch_idx": 192,
    "loss": 0.25792527198791504,
    "learning_rate": 0.0008945956429042942,
    "timestep_mean": 559.875,
    "epsilon_norm": 4.466324806213379,
    "prediction_norm": 3.686948299407959,
    "timestamp": "2025-07-14T12:33:48.421180"
  },
  {
    "epoch": 422,
    "batch_idx": 0,
    "loss": 0.410169780254364,
    "learning_rate": 0.0008941128059952199,
    "timestep_mean": 446.6875,
    "epsilon_norm": 7.881811141967773,
    "prediction_norm": 5.432734489440918,
    "timestamp": "2025-07-14T12:33:48.426142"
  },
  {
    "epoch": 422,
    "batch_idx": 32,
    "loss": 0.3642832934856415,
    "learning_rate": 0.0008941128059952199,
    "timestep_mean": 551.375,
    "epsilon_norm": 8.866366386413574,
    "prediction_norm": 6.8448710441589355,
    "timestamp": "2025-07-14T12:33:48.430963"
  },
  {
    "epoch": 422,
    "batch_idx": 64,
    "loss": 0.2700408697128296,
    "learning_rate": 0.0008941128059952199,
    "timestep_mean": 549.375,
    "epsilon_norm": 7.228437423706055,
    "prediction_norm": 5.420901775360107,
    "timestamp": "2025-07-14T12:33:48.435883"
  },
  {
    "epoch": 422,
    "batch_idx": 96,
    "loss": 0.23987460136413574,
    "learning_rate": 0.0008941128059952199,
    "timestep_mean": 529.96875,
    "epsilon_norm": 7.8647565841674805,
    "prediction_norm": 7.4012861251831055,
    "timestamp": "2025-07-14T12:33:48.440633"
  },
  {
    "epoch": 422,
    "batch_idx": 128,
    "loss": 0.3993147611618042,
    "learning_rate": 0.0008941128059952199,
    "timestep_mean": 459.6875,
    "epsilon_norm": 8.06896686553955,
    "prediction_norm": 5.82867431640625,
    "timestamp": "2025-07-14T12:33:48.445508"
  },
  {
    "epoch": 422,
    "batch_idx": 160,
    "loss": 0.310661643743515,
    "learning_rate": 0.0008941128059952199,
    "timestep_mean": 414.5625,
    "epsilon_norm": 7.42755126953125,
    "prediction_norm": 6.36312198638916,
    "timestamp": "2025-07-14T12:33:48.450470"
  },
  {
    "epoch": 422,
    "batch_idx": 192,
    "loss": 0.67326819896698,
    "learning_rate": 0.0008941128059952199,
    "timestep_mean": 310.5,
    "epsilon_norm": 4.427050590515137,
    "prediction_norm": 2.8057878017425537,
    "timestamp": "2025-07-14T12:33:48.455349"
  },
  {
    "epoch": 423,
    "batch_idx": 0,
    "loss": 0.20010660588741302,
    "learning_rate": 0.0008936289966519745,
    "timestep_mean": 556.625,
    "epsilon_norm": 8.828535079956055,
    "prediction_norm": 8.013099670410156,
    "timestamp": "2025-07-14T12:33:48.460227"
  },
  {
    "epoch": 423,
    "batch_idx": 32,
    "loss": 0.28668349981307983,
    "learning_rate": 0.0008936289966519745,
    "timestep_mean": 519.28125,
    "epsilon_norm": 7.948158264160156,
    "prediction_norm": 6.66749382019043,
    "timestamp": "2025-07-14T12:33:48.465100"
  },
  {
    "epoch": 423,
    "batch_idx": 64,
    "loss": 0.34677231311798096,
    "learning_rate": 0.0008936289966519745,
    "timestep_mean": 508.59375,
    "epsilon_norm": 8.675933837890625,
    "prediction_norm": 7.410365581512451,
    "timestamp": "2025-07-14T12:33:48.470017"
  },
  {
    "epoch": 423,
    "batch_idx": 96,
    "loss": 0.3142940104007721,
    "learning_rate": 0.0008936289966519745,
    "timestep_mean": 522.71875,
    "epsilon_norm": 7.425649642944336,
    "prediction_norm": 5.798241138458252,
    "timestamp": "2025-07-14T12:33:48.475013"
  },
  {
    "epoch": 423,
    "batch_idx": 128,
    "loss": 0.327121764421463,
    "learning_rate": 0.0008936289966519745,
    "timestep_mean": 547.1875,
    "epsilon_norm": 8.722643852233887,
    "prediction_norm": 7.292423725128174,
    "timestamp": "2025-07-14T12:33:48.479889"
  },
  {
    "epoch": 423,
    "batch_idx": 160,
    "loss": 0.3959028124809265,
    "learning_rate": 0.0008936289966519745,
    "timestep_mean": 571.4375,
    "epsilon_norm": 9.234820365905762,
    "prediction_norm": 9.054076194763184,
    "timestamp": "2025-07-14T12:33:48.484829"
  },
  {
    "epoch": 423,
    "batch_idx": 192,
    "loss": 0.28814345598220825,
    "learning_rate": 0.0008936289966519745,
    "timestep_mean": 361.25,
    "epsilon_norm": 3.09102201461792,
    "prediction_norm": 2.4566431045532227,
    "timestamp": "2025-07-14T12:33:48.489681"
  },
  {
    "epoch": 424,
    "batch_idx": 0,
    "loss": 0.340206116437912,
    "learning_rate": 0.0008931442160683093,
    "timestep_mean": 422.71875,
    "epsilon_norm": 7.700845718383789,
    "prediction_norm": 5.914405822753906,
    "timestamp": "2025-07-14T12:33:48.494549"
  },
  {
    "epoch": 424,
    "batch_idx": 32,
    "loss": 0.3880244195461273,
    "learning_rate": 0.0008931442160683093,
    "timestep_mean": 549.09375,
    "epsilon_norm": 8.325189590454102,
    "prediction_norm": 7.459073066711426,
    "timestamp": "2025-07-14T12:33:48.499419"
  },
  {
    "epoch": 424,
    "batch_idx": 64,
    "loss": 0.3733959496021271,
    "learning_rate": 0.0008931442160683093,
    "timestep_mean": 474.375,
    "epsilon_norm": 8.747276306152344,
    "prediction_norm": 7.4693989753723145,
    "timestamp": "2025-07-14T12:33:48.504426"
  },
  {
    "epoch": 424,
    "batch_idx": 96,
    "loss": 0.35012710094451904,
    "learning_rate": 0.0008931442160683093,
    "timestep_mean": 491.625,
    "epsilon_norm": 8.530752182006836,
    "prediction_norm": 7.040094375610352,
    "timestamp": "2025-07-14T12:33:48.509205"
  },
  {
    "epoch": 424,
    "batch_idx": 128,
    "loss": 0.3249579668045044,
    "learning_rate": 0.0008931442160683093,
    "timestep_mean": 540.875,
    "epsilon_norm": 8.078767776489258,
    "prediction_norm": 6.5674262046813965,
    "timestamp": "2025-07-14T12:33:48.514046"
  },
  {
    "epoch": 424,
    "batch_idx": 160,
    "loss": 0.34604209661483765,
    "learning_rate": 0.0008931442160683093,
    "timestep_mean": 523.125,
    "epsilon_norm": 9.903100967407227,
    "prediction_norm": 9.04344367980957,
    "timestamp": "2025-07-14T12:33:48.519012"
  },
  {
    "epoch": 424,
    "batch_idx": 192,
    "loss": 0.10907205939292908,
    "learning_rate": 0.0008931442160683093,
    "timestep_mean": 447.25,
    "epsilon_norm": 3.092665433883667,
    "prediction_norm": 3.4722628593444824,
    "timestamp": "2025-07-14T12:33:48.524133"
  },
  {
    "epoch": 425,
    "batch_idx": 0,
    "loss": 0.3135627508163452,
    "learning_rate": 0.0008926584654403723,
    "timestep_mean": 467.78125,
    "epsilon_norm": 7.795074462890625,
    "prediction_norm": 6.720922470092773,
    "timestamp": "2025-07-14T12:33:48.529289"
  },
  {
    "epoch": 425,
    "batch_idx": 32,
    "loss": 0.27964991331100464,
    "learning_rate": 0.0008926584654403723,
    "timestep_mean": 412.0,
    "epsilon_norm": 7.190287113189697,
    "prediction_norm": 6.385099411010742,
    "timestamp": "2025-07-14T12:33:48.534340"
  },
  {
    "epoch": 425,
    "batch_idx": 64,
    "loss": 0.5076051950454712,
    "learning_rate": 0.0008926584654403723,
    "timestep_mean": 469.28125,
    "epsilon_norm": 7.893585205078125,
    "prediction_norm": 6.655274868011475,
    "timestamp": "2025-07-14T12:33:48.539534"
  },
  {
    "epoch": 425,
    "batch_idx": 96,
    "loss": 0.40275511145591736,
    "learning_rate": 0.0008926584654403723,
    "timestep_mean": 484.75,
    "epsilon_norm": 8.45201301574707,
    "prediction_norm": 6.91738748550415,
    "timestamp": "2025-07-14T12:33:48.544665"
  },
  {
    "epoch": 425,
    "batch_idx": 128,
    "loss": 0.3185610771179199,
    "learning_rate": 0.0008926584654403723,
    "timestep_mean": 427.25,
    "epsilon_norm": 7.91813325881958,
    "prediction_norm": 6.409538269042969,
    "timestamp": "2025-07-14T12:33:48.549915"
  },
  {
    "epoch": 425,
    "batch_idx": 160,
    "loss": 0.47972625494003296,
    "learning_rate": 0.0008926584654403723,
    "timestep_mean": 516.125,
    "epsilon_norm": 8.76370906829834,
    "prediction_norm": 6.508824825286865,
    "timestamp": "2025-07-14T12:33:48.555017"
  },
  {
    "epoch": 425,
    "batch_idx": 192,
    "loss": 0.42490988969802856,
    "learning_rate": 0.0008926584654403723,
    "timestep_mean": 334.625,
    "epsilon_norm": 2.8055667877197266,
    "prediction_norm": 2.2664923667907715,
    "timestamp": "2025-07-14T12:33:48.560104"
  },
  {
    "epoch": 426,
    "batch_idx": 0,
    "loss": 0.3793511986732483,
    "learning_rate": 0.0008921717459667048,
    "timestep_mean": 502.96875,
    "epsilon_norm": 9.003044128417969,
    "prediction_norm": 6.947031497955322,
    "timestamp": "2025-07-14T12:33:48.565274"
  },
  {
    "epoch": 426,
    "batch_idx": 32,
    "loss": 0.21960356831550598,
    "learning_rate": 0.0008921717459667048,
    "timestep_mean": 521.6875,
    "epsilon_norm": 6.974788665771484,
    "prediction_norm": 5.839409828186035,
    "timestamp": "2025-07-14T12:33:48.570109"
  },
  {
    "epoch": 426,
    "batch_idx": 64,
    "loss": 0.1961442232131958,
    "learning_rate": 0.0008921717459667048,
    "timestep_mean": 503.96875,
    "epsilon_norm": 7.5538649559021,
    "prediction_norm": 6.5736918449401855,
    "timestamp": "2025-07-14T12:33:48.575296"
  },
  {
    "epoch": 426,
    "batch_idx": 96,
    "loss": 0.37295788526535034,
    "learning_rate": 0.0008921717459667048,
    "timestep_mean": 420.4375,
    "epsilon_norm": 7.34002161026001,
    "prediction_norm": 5.349989891052246,
    "timestamp": "2025-07-14T12:33:48.580216"
  },
  {
    "epoch": 426,
    "batch_idx": 128,
    "loss": 0.20109501481056213,
    "learning_rate": 0.0008921717459667048,
    "timestep_mean": 476.6875,
    "epsilon_norm": 8.165760040283203,
    "prediction_norm": 7.183745861053467,
    "timestamp": "2025-07-14T12:33:48.585107"
  },
  {
    "epoch": 426,
    "batch_idx": 160,
    "loss": 0.36843377351760864,
    "learning_rate": 0.0008921717459667048,
    "timestep_mean": 504.625,
    "epsilon_norm": 7.954537868499756,
    "prediction_norm": 5.2021026611328125,
    "timestamp": "2025-07-14T12:33:48.589935"
  },
  {
    "epoch": 426,
    "batch_idx": 192,
    "loss": 0.44176220893859863,
    "learning_rate": 0.0008921717459667048,
    "timestep_mean": 435.625,
    "epsilon_norm": 4.223081111907959,
    "prediction_norm": 3.235948324203491,
    "timestamp": "2025-07-14T12:33:48.594755"
  },
  {
    "epoch": 427,
    "batch_idx": 0,
    "loss": 0.2751598358154297,
    "learning_rate": 0.000891684058848239,
    "timestep_mean": 461.40625,
    "epsilon_norm": 7.747354507446289,
    "prediction_norm": 6.077392578125,
    "timestamp": "2025-07-14T12:33:48.599636"
  },
  {
    "epoch": 427,
    "batch_idx": 32,
    "loss": 0.32188093662261963,
    "learning_rate": 0.000891684058848239,
    "timestep_mean": 473.8125,
    "epsilon_norm": 9.407950401306152,
    "prediction_norm": 7.070483684539795,
    "timestamp": "2025-07-14T12:33:48.604579"
  },
  {
    "epoch": 427,
    "batch_idx": 64,
    "loss": 0.3963601589202881,
    "learning_rate": 0.000891684058848239,
    "timestep_mean": 411.5,
    "epsilon_norm": 7.906922340393066,
    "prediction_norm": 5.575963497161865,
    "timestamp": "2025-07-14T12:33:48.609463"
  },
  {
    "epoch": 427,
    "batch_idx": 96,
    "loss": 0.34714698791503906,
    "learning_rate": 0.000891684058848239,
    "timestep_mean": 499.03125,
    "epsilon_norm": 8.923895835876465,
    "prediction_norm": 6.391911029815674,
    "timestamp": "2025-07-14T12:33:48.614312"
  },
  {
    "epoch": 427,
    "batch_idx": 128,
    "loss": 0.26974964141845703,
    "learning_rate": 0.000891684058848239,
    "timestep_mean": 428.125,
    "epsilon_norm": 7.569489479064941,
    "prediction_norm": 6.337088108062744,
    "timestamp": "2025-07-14T12:33:48.619232"
  },
  {
    "epoch": 427,
    "batch_idx": 160,
    "loss": 0.44850990176200867,
    "learning_rate": 0.000891684058848239,
    "timestep_mean": 507.03125,
    "epsilon_norm": 8.516189575195312,
    "prediction_norm": 6.564535140991211,
    "timestamp": "2025-07-14T12:33:48.624158"
  },
  {
    "epoch": 427,
    "batch_idx": 192,
    "loss": 0.3269985318183899,
    "learning_rate": 0.000891684058848239,
    "timestep_mean": 527.25,
    "epsilon_norm": 3.637023448944092,
    "prediction_norm": 2.7615084648132324,
    "timestamp": "2025-07-14T12:33:48.629028"
  },
  {
    "epoch": 428,
    "batch_idx": 0,
    "loss": 0.4019063711166382,
    "learning_rate": 0.0008911954052882939,
    "timestep_mean": 470.0625,
    "epsilon_norm": 10.368478775024414,
    "prediction_norm": 9.387225151062012,
    "timestamp": "2025-07-14T12:33:48.633936"
  },
  {
    "epoch": 428,
    "batch_idx": 32,
    "loss": 0.30656614899635315,
    "learning_rate": 0.0008911954052882939,
    "timestep_mean": 450.84375,
    "epsilon_norm": 7.9288716316223145,
    "prediction_norm": 6.9721221923828125,
    "timestamp": "2025-07-14T12:33:48.638760"
  },
  {
    "epoch": 428,
    "batch_idx": 64,
    "loss": 0.20115971565246582,
    "learning_rate": 0.0008911954052882939,
    "timestep_mean": 536.3125,
    "epsilon_norm": 6.894926071166992,
    "prediction_norm": 6.142970085144043,
    "timestamp": "2025-07-14T12:33:48.643571"
  },
  {
    "epoch": 428,
    "batch_idx": 96,
    "loss": 0.2643454670906067,
    "learning_rate": 0.0008911954052882939,
    "timestep_mean": 503.21875,
    "epsilon_norm": 7.389537334442139,
    "prediction_norm": 7.416880130767822,
    "timestamp": "2025-07-14T12:33:48.648393"
  },
  {
    "epoch": 428,
    "batch_idx": 128,
    "loss": 0.14700183272361755,
    "learning_rate": 0.0008911954052882939,
    "timestep_mean": 526.9375,
    "epsilon_norm": 8.194632530212402,
    "prediction_norm": 7.451456069946289,
    "timestamp": "2025-07-14T12:33:48.653243"
  },
  {
    "epoch": 428,
    "batch_idx": 160,
    "loss": 0.28286707401275635,
    "learning_rate": 0.0008911954052882939,
    "timestep_mean": 535.8125,
    "epsilon_norm": 7.07090425491333,
    "prediction_norm": 5.765438556671143,
    "timestamp": "2025-07-14T12:33:48.658205"
  },
  {
    "epoch": 428,
    "batch_idx": 192,
    "loss": 0.2031518667936325,
    "learning_rate": 0.0008911954052882939,
    "timestep_mean": 512.0,
    "epsilon_norm": 2.6517059803009033,
    "prediction_norm": 2.472792863845825,
    "timestamp": "2025-07-14T12:33:48.663079"
  },
  {
    "epoch": 429,
    "batch_idx": 0,
    "loss": 0.34769076108932495,
    "learning_rate": 0.000890705786492574,
    "timestep_mean": 453.5,
    "epsilon_norm": 8.160701751708984,
    "prediction_norm": 7.134474277496338,
    "timestamp": "2025-07-14T12:33:48.667950"
  },
  {
    "epoch": 429,
    "batch_idx": 32,
    "loss": 0.2435736507177353,
    "learning_rate": 0.000890705786492574,
    "timestep_mean": 537.09375,
    "epsilon_norm": 7.905246734619141,
    "prediction_norm": 7.500735759735107,
    "timestamp": "2025-07-14T12:33:48.672754"
  },
  {
    "epoch": 429,
    "batch_idx": 64,
    "loss": 0.25178882479667664,
    "learning_rate": 0.000890705786492574,
    "timestep_mean": 404.6875,
    "epsilon_norm": 7.710984230041504,
    "prediction_norm": 6.29188346862793,
    "timestamp": "2025-07-14T12:33:48.677610"
  },
  {
    "epoch": 429,
    "batch_idx": 96,
    "loss": 0.22119122743606567,
    "learning_rate": 0.000890705786492574,
    "timestep_mean": 516.53125,
    "epsilon_norm": 7.045535564422607,
    "prediction_norm": 6.113900661468506,
    "timestamp": "2025-07-14T12:33:48.682487"
  },
  {
    "epoch": 429,
    "batch_idx": 128,
    "loss": 0.39284712076187134,
    "learning_rate": 0.000890705786492574,
    "timestep_mean": 541.5,
    "epsilon_norm": 8.322770118713379,
    "prediction_norm": 7.530933856964111,
    "timestamp": "2025-07-14T12:33:48.687382"
  },
  {
    "epoch": 429,
    "batch_idx": 160,
    "loss": 0.36352282762527466,
    "learning_rate": 0.000890705786492574,
    "timestep_mean": 471.75,
    "epsilon_norm": 7.713549613952637,
    "prediction_norm": 6.014718532562256,
    "timestamp": "2025-07-14T12:33:48.692320"
  },
  {
    "epoch": 429,
    "batch_idx": 192,
    "loss": 0.14308980107307434,
    "learning_rate": 0.000890705786492574,
    "timestep_mean": 554.875,
    "epsilon_norm": 3.6539528369903564,
    "prediction_norm": 3.572514772415161,
    "timestamp": "2025-07-14T12:33:48.697167"
  },
  {
    "epoch": 430,
    "batch_idx": 0,
    "loss": 0.27334529161453247,
    "learning_rate": 0.0008902152036691647,
    "timestep_mean": 544.53125,
    "epsilon_norm": 8.546409606933594,
    "prediction_norm": 7.677101135253906,
    "timestamp": "2025-07-14T12:33:48.702040"
  },
  {
    "epoch": 430,
    "batch_idx": 32,
    "loss": 0.2557797431945801,
    "learning_rate": 0.0008902152036691647,
    "timestep_mean": 551.5,
    "epsilon_norm": 7.539606094360352,
    "prediction_norm": 6.076610565185547,
    "timestamp": "2025-07-14T12:33:48.706876"
  },
  {
    "epoch": 430,
    "batch_idx": 64,
    "loss": 0.3235435485839844,
    "learning_rate": 0.0008902152036691647,
    "timestep_mean": 515.75,
    "epsilon_norm": 8.14832878112793,
    "prediction_norm": 6.536252021789551,
    "timestamp": "2025-07-14T12:33:48.711725"
  },
  {
    "epoch": 430,
    "batch_idx": 96,
    "loss": 0.2957686185836792,
    "learning_rate": 0.0008902152036691647,
    "timestep_mean": 486.46875,
    "epsilon_norm": 7.434238910675049,
    "prediction_norm": 5.391195297241211,
    "timestamp": "2025-07-14T12:33:48.716661"
  },
  {
    "epoch": 430,
    "batch_idx": 128,
    "loss": 0.3400188386440277,
    "learning_rate": 0.0008902152036691647,
    "timestep_mean": 397.125,
    "epsilon_norm": 9.106741905212402,
    "prediction_norm": 7.08209753036499,
    "timestamp": "2025-07-14T12:33:48.721537"
  },
  {
    "epoch": 430,
    "batch_idx": 160,
    "loss": 0.19198189675807953,
    "learning_rate": 0.0008902152036691647,
    "timestep_mean": 539.125,
    "epsilon_norm": 7.25101375579834,
    "prediction_norm": 6.4008049964904785,
    "timestamp": "2025-07-14T12:33:48.726482"
  },
  {
    "epoch": 430,
    "batch_idx": 192,
    "loss": 0.14355166256427765,
    "learning_rate": 0.0008902152036691647,
    "timestep_mean": 637.125,
    "epsilon_norm": 4.317273139953613,
    "prediction_norm": 3.7435219287872314,
    "timestamp": "2025-07-14T12:33:48.731316"
  },
  {
    "epoch": 431,
    "batch_idx": 0,
    "loss": 0.2406112551689148,
    "learning_rate": 0.0008897236580285306,
    "timestep_mean": 573.8125,
    "epsilon_norm": 7.876240253448486,
    "prediction_norm": 6.8129377365112305,
    "timestamp": "2025-07-14T12:33:48.736164"
  },
  {
    "epoch": 431,
    "batch_idx": 32,
    "loss": 0.2837787866592407,
    "learning_rate": 0.0008897236580285306,
    "timestep_mean": 571.21875,
    "epsilon_norm": 7.924516201019287,
    "prediction_norm": 6.587240219116211,
    "timestamp": "2025-07-14T12:33:48.741029"
  },
  {
    "epoch": 431,
    "batch_idx": 64,
    "loss": 0.5840481519699097,
    "learning_rate": 0.0008897236580285306,
    "timestep_mean": 387.5,
    "epsilon_norm": 8.770970344543457,
    "prediction_norm": 6.5118536949157715,
    "timestamp": "2025-07-14T12:33:48.745900"
  },
  {
    "epoch": 431,
    "batch_idx": 96,
    "loss": 0.3933514952659607,
    "learning_rate": 0.0008897236580285306,
    "timestep_mean": 455.3125,
    "epsilon_norm": 8.079695701599121,
    "prediction_norm": 5.510043144226074,
    "timestamp": "2025-07-14T12:33:48.750766"
  },
  {
    "epoch": 431,
    "batch_idx": 128,
    "loss": 0.29262468218803406,
    "learning_rate": 0.0008897236580285306,
    "timestep_mean": 618.5625,
    "epsilon_norm": 8.227492332458496,
    "prediction_norm": 7.047773361206055,
    "timestamp": "2025-07-14T12:33:48.755632"
  },
  {
    "epoch": 431,
    "batch_idx": 160,
    "loss": 0.235753133893013,
    "learning_rate": 0.0008897236580285306,
    "timestep_mean": 515.53125,
    "epsilon_norm": 7.45813512802124,
    "prediction_norm": 6.862444877624512,
    "timestamp": "2025-07-14T12:33:48.760502"
  },
  {
    "epoch": 431,
    "batch_idx": 192,
    "loss": 0.20914845168590546,
    "learning_rate": 0.0008897236580285306,
    "timestep_mean": 392.875,
    "epsilon_norm": 3.3606278896331787,
    "prediction_norm": 2.768472194671631,
    "timestamp": "2025-07-14T12:33:48.765409"
  },
  {
    "epoch": 432,
    "batch_idx": 0,
    "loss": 0.42034250497817993,
    "learning_rate": 0.0008892311507835116,
    "timestep_mean": 459.03125,
    "epsilon_norm": 7.651876449584961,
    "prediction_norm": 6.391684055328369,
    "timestamp": "2025-07-14T12:33:48.770443"
  },
  {
    "epoch": 432,
    "batch_idx": 32,
    "loss": 0.3589232861995697,
    "learning_rate": 0.0008892311507835116,
    "timestep_mean": 521.5,
    "epsilon_norm": 7.608226776123047,
    "prediction_norm": 6.235553741455078,
    "timestamp": "2025-07-14T12:33:48.775413"
  },
  {
    "epoch": 432,
    "batch_idx": 64,
    "loss": 0.3060142993927002,
    "learning_rate": 0.0008892311507835116,
    "timestep_mean": 570.84375,
    "epsilon_norm": 7.623175144195557,
    "prediction_norm": 7.320306777954102,
    "timestamp": "2025-07-14T12:33:48.780326"
  },
  {
    "epoch": 432,
    "batch_idx": 96,
    "loss": 0.36247462034225464,
    "learning_rate": 0.0008892311507835116,
    "timestep_mean": 469.71875,
    "epsilon_norm": 8.091955184936523,
    "prediction_norm": 7.109118938446045,
    "timestamp": "2025-07-14T12:33:48.785186"
  },
  {
    "epoch": 432,
    "batch_idx": 128,
    "loss": 0.28049758076667786,
    "learning_rate": 0.0008892311507835116,
    "timestep_mean": 493.15625,
    "epsilon_norm": 9.45162296295166,
    "prediction_norm": 9.138854026794434,
    "timestamp": "2025-07-14T12:33:48.790107"
  },
  {
    "epoch": 432,
    "batch_idx": 160,
    "loss": 0.41126519441604614,
    "learning_rate": 0.0008892311507835116,
    "timestep_mean": 547.5625,
    "epsilon_norm": 9.840004920959473,
    "prediction_norm": 8.341209411621094,
    "timestamp": "2025-07-14T12:33:48.794961"
  },
  {
    "epoch": 432,
    "batch_idx": 192,
    "loss": 0.47286513447761536,
    "learning_rate": 0.0008892311507835116,
    "timestep_mean": 515.0,
    "epsilon_norm": 4.401327133178711,
    "prediction_norm": 3.7018582820892334,
    "timestamp": "2025-07-14T12:33:48.799868"
  },
  {
    "epoch": 433,
    "batch_idx": 0,
    "loss": 0.22881607711315155,
    "learning_rate": 0.0008887376831493203,
    "timestep_mean": 522.46875,
    "epsilon_norm": 7.138621807098389,
    "prediction_norm": 6.455114841461182,
    "timestamp": "2025-07-14T12:33:48.804741"
  },
  {
    "epoch": 433,
    "batch_idx": 32,
    "loss": 0.2711865305900574,
    "learning_rate": 0.0008887376831493203,
    "timestep_mean": 524.09375,
    "epsilon_norm": 7.869759559631348,
    "prediction_norm": 7.116369724273682,
    "timestamp": "2025-07-14T12:33:48.809728"
  },
  {
    "epoch": 433,
    "batch_idx": 64,
    "loss": 0.1717405915260315,
    "learning_rate": 0.0008887376831493203,
    "timestep_mean": 646.03125,
    "epsilon_norm": 7.522247791290283,
    "prediction_norm": 7.32592248916626,
    "timestamp": "2025-07-14T12:33:48.814614"
  },
  {
    "epoch": 433,
    "batch_idx": 96,
    "loss": 0.15816131234169006,
    "learning_rate": 0.0008887376831493203,
    "timestep_mean": 512.84375,
    "epsilon_norm": 7.263710975646973,
    "prediction_norm": 6.261187553405762,
    "timestamp": "2025-07-14T12:33:48.819526"
  },
  {
    "epoch": 433,
    "batch_idx": 128,
    "loss": 0.4202183485031128,
    "learning_rate": 0.0008887376831493203,
    "timestep_mean": 528.90625,
    "epsilon_norm": 8.281670570373535,
    "prediction_norm": 6.65873908996582,
    "timestamp": "2025-07-14T12:33:48.824429"
  },
  {
    "epoch": 433,
    "batch_idx": 160,
    "loss": 0.2724342346191406,
    "learning_rate": 0.0008887376831493203,
    "timestep_mean": 467.90625,
    "epsilon_norm": 7.283137798309326,
    "prediction_norm": 5.7701311111450195,
    "timestamp": "2025-07-14T12:33:48.829419"
  },
  {
    "epoch": 433,
    "batch_idx": 192,
    "loss": 0.783814549446106,
    "learning_rate": 0.0008887376831493203,
    "timestep_mean": 280.125,
    "epsilon_norm": 5.1042022705078125,
    "prediction_norm": 2.8688390254974365,
    "timestamp": "2025-07-14T12:33:48.834341"
  },
  {
    "epoch": 434,
    "batch_idx": 0,
    "loss": 0.2305964231491089,
    "learning_rate": 0.0008882432563435392,
    "timestep_mean": 549.71875,
    "epsilon_norm": 8.028058052062988,
    "prediction_norm": 6.812018871307373,
    "timestamp": "2025-07-14T12:33:48.839325"
  },
  {
    "epoch": 434,
    "batch_idx": 32,
    "loss": 0.27508917450904846,
    "learning_rate": 0.0008882432563435392,
    "timestep_mean": 460.09375,
    "epsilon_norm": 8.237043380737305,
    "prediction_norm": 6.671871185302734,
    "timestamp": "2025-07-14T12:33:48.844190"
  },
  {
    "epoch": 434,
    "batch_idx": 64,
    "loss": 0.36757057905197144,
    "learning_rate": 0.0008882432563435392,
    "timestep_mean": 537.0,
    "epsilon_norm": 8.819687843322754,
    "prediction_norm": 7.019830226898193,
    "timestamp": "2025-07-14T12:33:48.849069"
  },
  {
    "epoch": 434,
    "batch_idx": 96,
    "loss": 0.2930482029914856,
    "learning_rate": 0.0008882432563435392,
    "timestep_mean": 502.0625,
    "epsilon_norm": 7.84828519821167,
    "prediction_norm": 5.428439140319824,
    "timestamp": "2025-07-14T12:33:48.854029"
  },
  {
    "epoch": 434,
    "batch_idx": 128,
    "loss": 0.21378520131111145,
    "learning_rate": 0.0008882432563435392,
    "timestep_mean": 455.375,
    "epsilon_norm": 8.174880981445312,
    "prediction_norm": 6.710346698760986,
    "timestamp": "2025-07-14T12:33:48.858998"
  },
  {
    "epoch": 434,
    "batch_idx": 160,
    "loss": 0.18367670476436615,
    "learning_rate": 0.0008882432563435392,
    "timestep_mean": 527.875,
    "epsilon_norm": 7.954965591430664,
    "prediction_norm": 6.71370792388916,
    "timestamp": "2025-07-14T12:33:48.863876"
  },
  {
    "epoch": 434,
    "batch_idx": 192,
    "loss": 0.2527492046356201,
    "learning_rate": 0.0008882432563435392,
    "timestep_mean": 543.75,
    "epsilon_norm": 4.065571308135986,
    "prediction_norm": 2.7478981018066406,
    "timestamp": "2025-07-14T12:33:48.868767"
  },
  {
    "epoch": 435,
    "batch_idx": 0,
    "loss": 0.34419265389442444,
    "learning_rate": 0.0008877478715861172,
    "timestep_mean": 439.90625,
    "epsilon_norm": 7.560461521148682,
    "prediction_norm": 5.740915298461914,
    "timestamp": "2025-07-14T12:33:48.873670"
  },
  {
    "epoch": 435,
    "batch_idx": 32,
    "loss": 0.3233972191810608,
    "learning_rate": 0.0008877478715861172,
    "timestep_mean": 492.0625,
    "epsilon_norm": 7.852683067321777,
    "prediction_norm": 6.386907577514648,
    "timestamp": "2025-07-14T12:33:48.878649"
  },
  {
    "epoch": 435,
    "batch_idx": 64,
    "loss": 0.2895878851413727,
    "learning_rate": 0.0008877478715861172,
    "timestep_mean": 553.96875,
    "epsilon_norm": 8.4710693359375,
    "prediction_norm": 7.351367950439453,
    "timestamp": "2025-07-14T12:33:48.883630"
  },
  {
    "epoch": 435,
    "batch_idx": 96,
    "loss": 0.1950286626815796,
    "learning_rate": 0.0008877478715861172,
    "timestep_mean": 514.90625,
    "epsilon_norm": 7.909698486328125,
    "prediction_norm": 6.928281784057617,
    "timestamp": "2025-07-14T12:33:48.888562"
  },
  {
    "epoch": 435,
    "batch_idx": 128,
    "loss": 0.48442378640174866,
    "learning_rate": 0.0008877478715861172,
    "timestep_mean": 539.03125,
    "epsilon_norm": 9.592876434326172,
    "prediction_norm": 7.682176113128662,
    "timestamp": "2025-07-14T12:33:48.893510"
  },
  {
    "epoch": 435,
    "batch_idx": 160,
    "loss": 0.10801525413990021,
    "learning_rate": 0.0008877478715861172,
    "timestep_mean": 561.28125,
    "epsilon_norm": 6.518592834472656,
    "prediction_norm": 7.1545257568359375,
    "timestamp": "2025-07-14T12:33:48.898510"
  },
  {
    "epoch": 435,
    "batch_idx": 192,
    "loss": 0.379381000995636,
    "learning_rate": 0.0008877478715861172,
    "timestep_mean": 504.875,
    "epsilon_norm": 4.362770080566406,
    "prediction_norm": 4.064977645874023,
    "timestamp": "2025-07-14T12:33:48.906676"
  },
  {
    "epoch": 436,
    "batch_idx": 0,
    "loss": 0.18186025321483612,
    "learning_rate": 0.0008872515300993669,
    "timestep_mean": 504.40625,
    "epsilon_norm": 7.074313163757324,
    "prediction_norm": 6.653547286987305,
    "timestamp": "2025-07-14T12:33:48.911571"
  },
  {
    "epoch": 436,
    "batch_idx": 32,
    "loss": 0.20799556374549866,
    "learning_rate": 0.0008872515300993669,
    "timestep_mean": 512.21875,
    "epsilon_norm": 9.175802230834961,
    "prediction_norm": 8.836913108825684,
    "timestamp": "2025-07-14T12:33:48.916528"
  },
  {
    "epoch": 436,
    "batch_idx": 64,
    "loss": 0.42271625995635986,
    "learning_rate": 0.0008872515300993669,
    "timestep_mean": 562.8125,
    "epsilon_norm": 7.938514232635498,
    "prediction_norm": 6.952149868011475,
    "timestamp": "2025-07-14T12:33:48.921378"
  },
  {
    "epoch": 436,
    "batch_idx": 96,
    "loss": 0.40594273805618286,
    "learning_rate": 0.0008872515300993669,
    "timestep_mean": 485.34375,
    "epsilon_norm": 8.84876537322998,
    "prediction_norm": 7.145166873931885,
    "timestamp": "2025-07-14T12:33:48.926379"
  },
  {
    "epoch": 436,
    "batch_idx": 128,
    "loss": 0.2994154691696167,
    "learning_rate": 0.0008872515300993669,
    "timestep_mean": 550.5625,
    "epsilon_norm": 10.055360794067383,
    "prediction_norm": 9.356038093566895,
    "timestamp": "2025-07-14T12:33:48.931336"
  },
  {
    "epoch": 436,
    "batch_idx": 160,
    "loss": 0.32900315523147583,
    "learning_rate": 0.0008872515300993669,
    "timestep_mean": 465.21875,
    "epsilon_norm": 8.2112398147583,
    "prediction_norm": 6.9929986000061035,
    "timestamp": "2025-07-14T12:33:48.936175"
  },
  {
    "epoch": 436,
    "batch_idx": 192,
    "loss": 1.1757780313491821,
    "learning_rate": 0.0008872515300993669,
    "timestep_mean": 472.125,
    "epsilon_norm": 4.854423522949219,
    "prediction_norm": 2.6346192359924316,
    "timestamp": "2025-07-14T12:33:48.941011"
  },
  {
    "epoch": 437,
    "batch_idx": 0,
    "loss": 0.22231358289718628,
    "learning_rate": 0.0008867542331079616,
    "timestep_mean": 471.25,
    "epsilon_norm": 7.025962829589844,
    "prediction_norm": 6.159201145172119,
    "timestamp": "2025-07-14T12:33:48.945931"
  },
  {
    "epoch": 437,
    "batch_idx": 32,
    "loss": 0.2910657823085785,
    "learning_rate": 0.0008867542331079616,
    "timestep_mean": 501.75,
    "epsilon_norm": 8.553153038024902,
    "prediction_norm": 6.702212810516357,
    "timestamp": "2025-07-14T12:33:48.950831"
  },
  {
    "epoch": 437,
    "batch_idx": 64,
    "loss": 0.24853390455245972,
    "learning_rate": 0.0008867542331079616,
    "timestep_mean": 511.65625,
    "epsilon_norm": 8.237789154052734,
    "prediction_norm": 7.466691493988037,
    "timestamp": "2025-07-14T12:33:48.955721"
  },
  {
    "epoch": 437,
    "batch_idx": 96,
    "loss": 0.3510624170303345,
    "learning_rate": 0.0008867542331079616,
    "timestep_mean": 486.875,
    "epsilon_norm": 8.464258193969727,
    "prediction_norm": 6.682155132293701,
    "timestamp": "2025-07-14T12:33:48.960721"
  },
  {
    "epoch": 437,
    "batch_idx": 128,
    "loss": 0.36074358224868774,
    "learning_rate": 0.0008867542331079616,
    "timestep_mean": 514.84375,
    "epsilon_norm": 7.9401631355285645,
    "prediction_norm": 5.948818206787109,
    "timestamp": "2025-07-14T12:33:48.965691"
  },
  {
    "epoch": 437,
    "batch_idx": 160,
    "loss": 0.6859301328659058,
    "learning_rate": 0.0008867542331079616,
    "timestep_mean": 490.78125,
    "epsilon_norm": 8.03842830657959,
    "prediction_norm": 5.1122145652771,
    "timestamp": "2025-07-14T12:33:48.970820"
  },
  {
    "epoch": 437,
    "batch_idx": 192,
    "loss": 0.21626386046409607,
    "learning_rate": 0.0008867542331079616,
    "timestep_mean": 487.25,
    "epsilon_norm": 3.1091766357421875,
    "prediction_norm": 2.375955820083618,
    "timestamp": "2025-07-14T12:33:48.975763"
  },
  {
    "epoch": 438,
    "batch_idx": 0,
    "loss": 0.43582063913345337,
    "learning_rate": 0.0008862559818389321,
    "timestep_mean": 505.5,
    "epsilon_norm": 8.034468650817871,
    "prediction_norm": 5.780572414398193,
    "timestamp": "2025-07-14T12:33:48.980811"
  },
  {
    "epoch": 438,
    "batch_idx": 32,
    "loss": 0.4426080584526062,
    "learning_rate": 0.0008862559818389321,
    "timestep_mean": 486.1875,
    "epsilon_norm": 8.657205581665039,
    "prediction_norm": 6.4330525398254395,
    "timestamp": "2025-07-14T12:33:48.985995"
  },
  {
    "epoch": 438,
    "batch_idx": 64,
    "loss": 0.18893781304359436,
    "learning_rate": 0.0008862559818389321,
    "timestep_mean": 576.5625,
    "epsilon_norm": 7.642154693603516,
    "prediction_norm": 6.6195292472839355,
    "timestamp": "2025-07-14T12:33:48.990981"
  },
  {
    "epoch": 438,
    "batch_idx": 96,
    "loss": 0.3272632360458374,
    "learning_rate": 0.0008862559818389321,
    "timestep_mean": 581.75,
    "epsilon_norm": 7.461287498474121,
    "prediction_norm": 4.650430679321289,
    "timestamp": "2025-07-14T12:33:48.996188"
  },
  {
    "epoch": 438,
    "batch_idx": 128,
    "loss": 0.41989970207214355,
    "learning_rate": 0.0008862559818389321,
    "timestep_mean": 469.71875,
    "epsilon_norm": 7.741237640380859,
    "prediction_norm": 5.827157497406006,
    "timestamp": "2025-07-14T12:33:49.001171"
  },
  {
    "epoch": 438,
    "batch_idx": 160,
    "loss": 0.44153764843940735,
    "learning_rate": 0.0008862559818389321,
    "timestep_mean": 505.15625,
    "epsilon_norm": 7.183114528656006,
    "prediction_norm": 5.244935035705566,
    "timestamp": "2025-07-14T12:33:49.006159"
  },
  {
    "epoch": 438,
    "batch_idx": 192,
    "loss": 0.6937592625617981,
    "learning_rate": 0.0008862559818389321,
    "timestep_mean": 443.75,
    "epsilon_norm": 4.545793056488037,
    "prediction_norm": 2.4324076175689697,
    "timestamp": "2025-07-14T12:33:49.011251"
  },
  {
    "epoch": 439,
    "batch_idx": 0,
    "loss": 0.3957856297492981,
    "learning_rate": 0.000885756777521664,
    "timestep_mean": 533.65625,
    "epsilon_norm": 8.372260093688965,
    "prediction_norm": 6.771801948547363,
    "timestamp": "2025-07-14T12:33:49.016562"
  },
  {
    "epoch": 439,
    "batch_idx": 32,
    "loss": 0.21750426292419434,
    "learning_rate": 0.000885756777521664,
    "timestep_mean": 556.78125,
    "epsilon_norm": 8.147852897644043,
    "prediction_norm": 8.0697603225708,
    "timestamp": "2025-07-14T12:33:49.021642"
  },
  {
    "epoch": 439,
    "batch_idx": 64,
    "loss": 0.21790874004364014,
    "learning_rate": 0.000885756777521664,
    "timestep_mean": 534.78125,
    "epsilon_norm": 8.63550853729248,
    "prediction_norm": 7.264214992523193,
    "timestamp": "2025-07-14T12:33:49.026726"
  },
  {
    "epoch": 439,
    "batch_idx": 96,
    "loss": 0.1910698115825653,
    "learning_rate": 0.000885756777521664,
    "timestep_mean": 560.09375,
    "epsilon_norm": 7.6149516105651855,
    "prediction_norm": 7.46766471862793,
    "timestamp": "2025-07-14T12:33:49.031753"
  },
  {
    "epoch": 439,
    "batch_idx": 128,
    "loss": 0.16706790030002594,
    "learning_rate": 0.000885756777521664,
    "timestep_mean": 578.4375,
    "epsilon_norm": 8.430310249328613,
    "prediction_norm": 7.678993225097656,
    "timestamp": "2025-07-14T12:33:49.036710"
  },
  {
    "epoch": 439,
    "batch_idx": 160,
    "loss": 0.44182974100112915,
    "learning_rate": 0.000885756777521664,
    "timestep_mean": 458.34375,
    "epsilon_norm": 8.176870346069336,
    "prediction_norm": 6.236240386962891,
    "timestamp": "2025-07-14T12:33:49.042067"
  },
  {
    "epoch": 439,
    "batch_idx": 192,
    "loss": 0.8025292158126831,
    "learning_rate": 0.000885756777521664,
    "timestep_mean": 480.75,
    "epsilon_norm": 4.78410530090332,
    "prediction_norm": 3.716566562652588,
    "timestamp": "2025-07-14T12:33:49.047181"
  },
  {
    "epoch": 440,
    "batch_idx": 0,
    "loss": 0.32369738817214966,
    "learning_rate": 0.0008852566213878943,
    "timestep_mean": 495.71875,
    "epsilon_norm": 6.640781402587891,
    "prediction_norm": 5.412577152252197,
    "timestamp": "2025-07-14T12:33:49.052489"
  },
  {
    "epoch": 440,
    "batch_idx": 32,
    "loss": 0.23375624418258667,
    "learning_rate": 0.0008852566213878943,
    "timestep_mean": 433.21875,
    "epsilon_norm": 7.38719367980957,
    "prediction_norm": 6.923376560211182,
    "timestamp": "2025-07-14T12:33:49.057605"
  },
  {
    "epoch": 440,
    "batch_idx": 64,
    "loss": 0.335527241230011,
    "learning_rate": 0.0008852566213878943,
    "timestep_mean": 521.84375,
    "epsilon_norm": 8.486425399780273,
    "prediction_norm": 8.812308311462402,
    "timestamp": "2025-07-14T12:33:49.062777"
  },
  {
    "epoch": 440,
    "batch_idx": 96,
    "loss": 0.39424842596054077,
    "learning_rate": 0.0008852566213878943,
    "timestep_mean": 482.6875,
    "epsilon_norm": 7.700058460235596,
    "prediction_norm": 6.55540657043457,
    "timestamp": "2025-07-14T12:33:49.067796"
  },
  {
    "epoch": 440,
    "batch_idx": 128,
    "loss": 0.2640683948993683,
    "learning_rate": 0.0008852566213878943,
    "timestep_mean": 534.65625,
    "epsilon_norm": 8.826411247253418,
    "prediction_norm": 7.662128448486328,
    "timestamp": "2025-07-14T12:33:49.072926"
  },
  {
    "epoch": 440,
    "batch_idx": 160,
    "loss": 0.13360312581062317,
    "learning_rate": 0.0008852566213878943,
    "timestep_mean": 571.5,
    "epsilon_norm": 7.020877838134766,
    "prediction_norm": 6.432757377624512,
    "timestamp": "2025-07-14T12:33:49.078049"
  },
  {
    "epoch": 440,
    "batch_idx": 192,
    "loss": 0.08081557601690292,
    "learning_rate": 0.0008852566213878943,
    "timestep_mean": 633.5,
    "epsilon_norm": 5.289231300354004,
    "prediction_norm": 5.014416694641113,
    "timestamp": "2025-07-14T12:33:49.083143"
  },
  {
    "epoch": 441,
    "batch_idx": 0,
    "loss": 0.21768300235271454,
    "learning_rate": 0.0008847555146717088,
    "timestep_mean": 576.71875,
    "epsilon_norm": 7.916326522827148,
    "prediction_norm": 6.549160957336426,
    "timestamp": "2025-07-14T12:33:49.088295"
  },
  {
    "epoch": 441,
    "batch_idx": 32,
    "loss": 0.2040572464466095,
    "learning_rate": 0.0008847555146717088,
    "timestep_mean": 558.21875,
    "epsilon_norm": 8.078499794006348,
    "prediction_norm": 6.838926315307617,
    "timestamp": "2025-07-14T12:33:49.093373"
  },
  {
    "epoch": 441,
    "batch_idx": 64,
    "loss": 0.6855658888816833,
    "learning_rate": 0.0008847555146717088,
    "timestep_mean": 425.96875,
    "epsilon_norm": 9.703933715820312,
    "prediction_norm": 7.122090816497803,
    "timestamp": "2025-07-14T12:33:49.098631"
  },
  {
    "epoch": 441,
    "batch_idx": 96,
    "loss": 0.2826583981513977,
    "learning_rate": 0.0008847555146717088,
    "timestep_mean": 489.5,
    "epsilon_norm": 7.738193035125732,
    "prediction_norm": 6.63223934173584,
    "timestamp": "2025-07-14T12:33:49.103729"
  },
  {
    "epoch": 441,
    "batch_idx": 128,
    "loss": 0.3691118359565735,
    "learning_rate": 0.0008847555146717088,
    "timestep_mean": 463.1875,
    "epsilon_norm": 7.781984806060791,
    "prediction_norm": 5.858293533325195,
    "timestamp": "2025-07-14T12:33:49.108919"
  },
  {
    "epoch": 441,
    "batch_idx": 160,
    "loss": 0.35595518350601196,
    "learning_rate": 0.0008847555146717088,
    "timestep_mean": 460.90625,
    "epsilon_norm": 7.349828243255615,
    "prediction_norm": 6.447614669799805,
    "timestamp": "2025-07-14T12:33:49.114083"
  },
  {
    "epoch": 441,
    "batch_idx": 192,
    "loss": 0.26676368713378906,
    "learning_rate": 0.0008847555146717088,
    "timestep_mean": 529.125,
    "epsilon_norm": 4.12507438659668,
    "prediction_norm": 3.359975576400757,
    "timestamp": "2025-07-14T12:33:49.119157"
  },
  {
    "epoch": 442,
    "batch_idx": 0,
    "loss": 0.6464172601699829,
    "learning_rate": 0.000884253458609538,
    "timestep_mean": 392.9375,
    "epsilon_norm": 8.958596229553223,
    "prediction_norm": 6.0570759773254395,
    "timestamp": "2025-07-14T12:33:49.124247"
  },
  {
    "epoch": 442,
    "batch_idx": 32,
    "loss": 0.20400899648666382,
    "learning_rate": 0.000884253458609538,
    "timestep_mean": 536.21875,
    "epsilon_norm": 7.613751411437988,
    "prediction_norm": 6.782995700836182,
    "timestamp": "2025-07-14T12:33:49.129418"
  },
  {
    "epoch": 442,
    "batch_idx": 64,
    "loss": 0.1435788869857788,
    "learning_rate": 0.000884253458609538,
    "timestep_mean": 544.3125,
    "epsilon_norm": 6.097810745239258,
    "prediction_norm": 6.0243964195251465,
    "timestamp": "2025-07-14T12:33:49.134524"
  },
  {
    "epoch": 442,
    "batch_idx": 96,
    "loss": 0.24438637495040894,
    "learning_rate": 0.000884253458609538,
    "timestep_mean": 497.71875,
    "epsilon_norm": 7.274527072906494,
    "prediction_norm": 5.813244342803955,
    "timestamp": "2025-07-14T12:33:49.139583"
  },
  {
    "epoch": 442,
    "batch_idx": 128,
    "loss": 0.3298448324203491,
    "learning_rate": 0.000884253458609538,
    "timestep_mean": 475.78125,
    "epsilon_norm": 8.322668075561523,
    "prediction_norm": 6.729963302612305,
    "timestamp": "2025-07-14T12:33:49.144743"
  },
  {
    "epoch": 442,
    "batch_idx": 160,
    "loss": 0.2908828854560852,
    "learning_rate": 0.000884253458609538,
    "timestep_mean": 527.03125,
    "epsilon_norm": 9.189210891723633,
    "prediction_norm": 7.551748752593994,
    "timestamp": "2025-07-14T12:33:49.149940"
  },
  {
    "epoch": 442,
    "batch_idx": 192,
    "loss": 0.18822965025901794,
    "learning_rate": 0.000884253458609538,
    "timestep_mean": 627.125,
    "epsilon_norm": 4.262662410736084,
    "prediction_norm": 3.62107515335083,
    "timestamp": "2025-07-14T12:33:49.155034"
  },
  {
    "epoch": 443,
    "batch_idx": 0,
    "loss": 0.3088805079460144,
    "learning_rate": 0.0008837504544401557,
    "timestep_mean": 514.125,
    "epsilon_norm": 8.251691818237305,
    "prediction_norm": 6.973348140716553,
    "timestamp": "2025-07-14T12:33:49.160150"
  },
  {
    "epoch": 443,
    "batch_idx": 32,
    "loss": 0.2927192449569702,
    "learning_rate": 0.0008837504544401557,
    "timestep_mean": 522.25,
    "epsilon_norm": 8.020339965820312,
    "prediction_norm": 6.173281192779541,
    "timestamp": "2025-07-14T12:33:49.165326"
  },
  {
    "epoch": 443,
    "batch_idx": 64,
    "loss": 0.40708062052726746,
    "learning_rate": 0.0008837504544401557,
    "timestep_mean": 417.21875,
    "epsilon_norm": 8.381632804870605,
    "prediction_norm": 6.112500190734863,
    "timestamp": "2025-07-14T12:33:49.170455"
  },
  {
    "epoch": 443,
    "batch_idx": 96,
    "loss": 0.4187037944793701,
    "learning_rate": 0.0008837504544401557,
    "timestep_mean": 535.34375,
    "epsilon_norm": 7.282094478607178,
    "prediction_norm": 5.312765598297119,
    "timestamp": "2025-07-14T12:33:49.175644"
  },
  {
    "epoch": 443,
    "batch_idx": 128,
    "loss": 0.3470819592475891,
    "learning_rate": 0.0008837504544401557,
    "timestep_mean": 452.6875,
    "epsilon_norm": 7.989263534545898,
    "prediction_norm": 6.382078170776367,
    "timestamp": "2025-07-14T12:33:49.180657"
  },
  {
    "epoch": 443,
    "batch_idx": 160,
    "loss": 0.32269537448883057,
    "learning_rate": 0.0008837504544401557,
    "timestep_mean": 485.78125,
    "epsilon_norm": 8.609885215759277,
    "prediction_norm": 6.925787448883057,
    "timestamp": "2025-07-14T12:33:49.185852"
  },
  {
    "epoch": 443,
    "batch_idx": 192,
    "loss": 0.840728759765625,
    "learning_rate": 0.0008837504544401557,
    "timestep_mean": 352.25,
    "epsilon_norm": 4.099148273468018,
    "prediction_norm": 1.352817416191101,
    "timestamp": "2025-07-14T12:33:49.191079"
  },
  {
    "epoch": 444,
    "batch_idx": 0,
    "loss": 0.42484548687934875,
    "learning_rate": 0.0008832465034046747,
    "timestep_mean": 554.125,
    "epsilon_norm": 8.761887550354004,
    "prediction_norm": 6.488715171813965,
    "timestamp": "2025-07-14T12:33:49.196337"
  },
  {
    "epoch": 444,
    "batch_idx": 32,
    "loss": 0.11964605748653412,
    "learning_rate": 0.0008832465034046747,
    "timestep_mean": 545.65625,
    "epsilon_norm": 7.253596305847168,
    "prediction_norm": 6.771503925323486,
    "timestamp": "2025-07-14T12:33:49.201444"
  },
  {
    "epoch": 444,
    "batch_idx": 64,
    "loss": 0.38555455207824707,
    "learning_rate": 0.0008832465034046747,
    "timestep_mean": 478.1875,
    "epsilon_norm": 8.790074348449707,
    "prediction_norm": 7.0263261795043945,
    "timestamp": "2025-07-14T12:33:49.206609"
  },
  {
    "epoch": 444,
    "batch_idx": 96,
    "loss": 0.39559298753738403,
    "learning_rate": 0.0008832465034046747,
    "timestep_mean": 464.1875,
    "epsilon_norm": 9.211739540100098,
    "prediction_norm": 7.699112415313721,
    "timestamp": "2025-07-14T12:33:49.211726"
  },
  {
    "epoch": 444,
    "batch_idx": 128,
    "loss": 0.4230116307735443,
    "learning_rate": 0.0008832465034046747,
    "timestep_mean": 541.8125,
    "epsilon_norm": 8.199201583862305,
    "prediction_norm": 6.933363437652588,
    "timestamp": "2025-07-14T12:33:49.217016"
  },
  {
    "epoch": 444,
    "batch_idx": 160,
    "loss": 0.32139524817466736,
    "learning_rate": 0.0008832465034046747,
    "timestep_mean": 514.03125,
    "epsilon_norm": 8.717000961303711,
    "prediction_norm": 7.54773473739624,
    "timestamp": "2025-07-14T12:33:49.222208"
  },
  {
    "epoch": 444,
    "batch_idx": 192,
    "loss": 0.18206733465194702,
    "learning_rate": 0.0008832465034046747,
    "timestep_mean": 676.75,
    "epsilon_norm": 4.021199703216553,
    "prediction_norm": 3.5978410243988037,
    "timestamp": "2025-07-14T12:33:49.227316"
  },
  {
    "epoch": 445,
    "batch_idx": 0,
    "loss": 0.5775725245475769,
    "learning_rate": 0.0008827416067465439,
    "timestep_mean": 458.21875,
    "epsilon_norm": 8.582853317260742,
    "prediction_norm": 6.6053242683410645,
    "timestamp": "2025-07-14T12:33:49.232353"
  },
  {
    "epoch": 445,
    "batch_idx": 32,
    "loss": 0.5202676057815552,
    "learning_rate": 0.0008827416067465439,
    "timestep_mean": 443.65625,
    "epsilon_norm": 7.651447772979736,
    "prediction_norm": 5.44642972946167,
    "timestamp": "2025-07-14T12:33:49.237551"
  },
  {
    "epoch": 445,
    "batch_idx": 64,
    "loss": 0.25695449113845825,
    "learning_rate": 0.0008827416067465439,
    "timestep_mean": 504.375,
    "epsilon_norm": 8.67768669128418,
    "prediction_norm": 8.533828735351562,
    "timestamp": "2025-07-14T12:33:49.242683"
  },
  {
    "epoch": 445,
    "batch_idx": 96,
    "loss": 0.468334436416626,
    "learning_rate": 0.0008827416067465439,
    "timestep_mean": 436.53125,
    "epsilon_norm": 8.103248596191406,
    "prediction_norm": 7.202175617218018,
    "timestamp": "2025-07-14T12:33:49.247944"
  },
  {
    "epoch": 445,
    "batch_idx": 128,
    "loss": 0.35517942905426025,
    "learning_rate": 0.0008827416067465439,
    "timestep_mean": 469.9375,
    "epsilon_norm": 8.366923332214355,
    "prediction_norm": 7.820867538452148,
    "timestamp": "2025-07-14T12:33:49.253174"
  },
  {
    "epoch": 445,
    "batch_idx": 160,
    "loss": 0.4039859473705292,
    "learning_rate": 0.0008827416067465439,
    "timestep_mean": 547.5,
    "epsilon_norm": 8.422183990478516,
    "prediction_norm": 6.701696872711182,
    "timestamp": "2025-07-14T12:33:49.258464"
  },
  {
    "epoch": 445,
    "batch_idx": 192,
    "loss": 0.3643079996109009,
    "learning_rate": 0.0008827416067465439,
    "timestep_mean": 357.25,
    "epsilon_norm": 3.8500845432281494,
    "prediction_norm": 2.9308605194091797,
    "timestamp": "2025-07-14T12:33:49.263423"
  },
  {
    "epoch": 446,
    "batch_idx": 0,
    "loss": 0.527113139629364,
    "learning_rate": 0.0008822357657115456,
    "timestep_mean": 445.90625,
    "epsilon_norm": 7.9123382568359375,
    "prediction_norm": 5.803274154663086,
    "timestamp": "2025-07-14T12:33:49.268618"
  },
  {
    "epoch": 446,
    "batch_idx": 32,
    "loss": 0.2983846962451935,
    "learning_rate": 0.0008822357657115456,
    "timestep_mean": 449.09375,
    "epsilon_norm": 7.935153961181641,
    "prediction_norm": 6.952359199523926,
    "timestamp": "2025-07-14T12:33:49.273578"
  },
  {
    "epoch": 446,
    "batch_idx": 64,
    "loss": 0.3533344268798828,
    "learning_rate": 0.0008822357657115456,
    "timestep_mean": 494.1875,
    "epsilon_norm": 8.9158353805542,
    "prediction_norm": 7.319838047027588,
    "timestamp": "2025-07-14T12:33:49.278598"
  },
  {
    "epoch": 446,
    "batch_idx": 96,
    "loss": 0.26943439245224,
    "learning_rate": 0.0008822357657115456,
    "timestep_mean": 537.65625,
    "epsilon_norm": 7.500217914581299,
    "prediction_norm": 7.101439476013184,
    "timestamp": "2025-07-14T12:33:49.283898"
  },
  {
    "epoch": 446,
    "batch_idx": 128,
    "loss": 0.16772781312465668,
    "learning_rate": 0.0008822357657115456,
    "timestep_mean": 534.6875,
    "epsilon_norm": 8.054715156555176,
    "prediction_norm": 6.999856472015381,
    "timestamp": "2025-07-14T12:33:49.289122"
  },
  {
    "epoch": 446,
    "batch_idx": 160,
    "loss": 0.31385868787765503,
    "learning_rate": 0.0008822357657115456,
    "timestep_mean": 512.21875,
    "epsilon_norm": 8.474666595458984,
    "prediction_norm": 7.151582717895508,
    "timestamp": "2025-07-14T12:33:49.294301"
  },
  {
    "epoch": 446,
    "batch_idx": 192,
    "loss": 0.1018601730465889,
    "learning_rate": 0.0008822357657115456,
    "timestep_mean": 664.25,
    "epsilon_norm": 3.652294874191284,
    "prediction_norm": 3.0080268383026123,
    "timestamp": "2025-07-14T12:33:49.299533"
  },
  {
    "epoch": 447,
    "batch_idx": 0,
    "loss": 0.34895059466362,
    "learning_rate": 0.0008817289815477924,
    "timestep_mean": 551.0625,
    "epsilon_norm": 8.399113655090332,
    "prediction_norm": 6.449570178985596,
    "timestamp": "2025-07-14T12:33:49.304767"
  },
  {
    "epoch": 447,
    "batch_idx": 32,
    "loss": 0.22537165880203247,
    "learning_rate": 0.0008817289815477924,
    "timestep_mean": 524.375,
    "epsilon_norm": 7.475079536437988,
    "prediction_norm": 6.314560413360596,
    "timestamp": "2025-07-14T12:33:49.309889"
  },
  {
    "epoch": 447,
    "batch_idx": 64,
    "loss": 0.3614380955696106,
    "learning_rate": 0.0008817289815477924,
    "timestep_mean": 475.0625,
    "epsilon_norm": 8.179046630859375,
    "prediction_norm": 5.43057107925415,
    "timestamp": "2025-07-14T12:33:49.314994"
  },
  {
    "epoch": 447,
    "batch_idx": 96,
    "loss": 0.42446160316467285,
    "learning_rate": 0.0008817289815477924,
    "timestep_mean": 452.375,
    "epsilon_norm": 7.79103946685791,
    "prediction_norm": 5.635120868682861,
    "timestamp": "2025-07-14T12:33:49.320150"
  },
  {
    "epoch": 447,
    "batch_idx": 128,
    "loss": 0.2829563617706299,
    "learning_rate": 0.0008817289815477924,
    "timestep_mean": 520.75,
    "epsilon_norm": 7.401810169219971,
    "prediction_norm": 5.627610206604004,
    "timestamp": "2025-07-14T12:33:49.325306"
  },
  {
    "epoch": 447,
    "batch_idx": 160,
    "loss": 0.3748626112937927,
    "learning_rate": 0.0008817289815477924,
    "timestep_mean": 553.90625,
    "epsilon_norm": 7.609420299530029,
    "prediction_norm": 5.403140068054199,
    "timestamp": "2025-07-14T12:33:49.330446"
  },
  {
    "epoch": 447,
    "batch_idx": 192,
    "loss": 0.06402428448200226,
    "learning_rate": 0.0008817289815477924,
    "timestep_mean": 547.125,
    "epsilon_norm": 4.100808620452881,
    "prediction_norm": 4.056556701660156,
    "timestamp": "2025-07-14T12:33:49.335547"
  },
  {
    "epoch": 448,
    "batch_idx": 0,
    "loss": 0.4471384286880493,
    "learning_rate": 0.0008812212555057236,
    "timestep_mean": 529.03125,
    "epsilon_norm": 8.034896850585938,
    "prediction_norm": 6.77198600769043,
    "timestamp": "2025-07-14T12:33:49.340735"
  },
  {
    "epoch": 448,
    "batch_idx": 32,
    "loss": 0.2711802124977112,
    "learning_rate": 0.0008812212555057236,
    "timestep_mean": 497.9375,
    "epsilon_norm": 7.4424147605896,
    "prediction_norm": 5.924081325531006,
    "timestamp": "2025-07-14T12:33:49.345917"
  },
  {
    "epoch": 448,
    "batch_idx": 64,
    "loss": 0.3198091983795166,
    "learning_rate": 0.0008812212555057236,
    "timestep_mean": 462.125,
    "epsilon_norm": 8.12444019317627,
    "prediction_norm": 6.897472858428955,
    "timestamp": "2025-07-14T12:33:49.351010"
  },
  {
    "epoch": 448,
    "batch_idx": 96,
    "loss": 0.3998318910598755,
    "learning_rate": 0.0008812212555057236,
    "timestep_mean": 384.25,
    "epsilon_norm": 7.902132511138916,
    "prediction_norm": 6.318171501159668,
    "timestamp": "2025-07-14T12:33:49.356222"
  },
  {
    "epoch": 448,
    "batch_idx": 128,
    "loss": 0.36021384596824646,
    "learning_rate": 0.0008812212555057236,
    "timestep_mean": 504.28125,
    "epsilon_norm": 8.608478546142578,
    "prediction_norm": 7.41687536239624,
    "timestamp": "2025-07-14T12:33:49.361437"
  },
  {
    "epoch": 448,
    "batch_idx": 160,
    "loss": 0.3010382056236267,
    "learning_rate": 0.0008812212555057236,
    "timestep_mean": 545.65625,
    "epsilon_norm": 8.934765815734863,
    "prediction_norm": 8.45395565032959,
    "timestamp": "2025-07-14T12:33:49.366536"
  },
  {
    "epoch": 448,
    "batch_idx": 192,
    "loss": 0.5096020102500916,
    "learning_rate": 0.0008812212555057236,
    "timestep_mean": 328.5,
    "epsilon_norm": 3.6112077236175537,
    "prediction_norm": 1.9518237113952637,
    "timestamp": "2025-07-14T12:33:49.371687"
  },
  {
    "epoch": 449,
    "batch_idx": 0,
    "loss": 0.18582898378372192,
    "learning_rate": 0.000880712588838103,
    "timestep_mean": 510.6875,
    "epsilon_norm": 7.37746000289917,
    "prediction_norm": 6.702426910400391,
    "timestamp": "2025-07-14T12:33:49.376862"
  },
  {
    "epoch": 449,
    "batch_idx": 32,
    "loss": 0.22035007178783417,
    "learning_rate": 0.000880712588838103,
    "timestep_mean": 499.03125,
    "epsilon_norm": 7.894087791442871,
    "prediction_norm": 6.989034175872803,
    "timestamp": "2025-07-14T12:33:49.382047"
  },
  {
    "epoch": 449,
    "batch_idx": 64,
    "loss": 0.23116573691368103,
    "learning_rate": 0.000880712588838103,
    "timestep_mean": 504.28125,
    "epsilon_norm": 6.675891876220703,
    "prediction_norm": 5.757907390594482,
    "timestamp": "2025-07-14T12:33:49.387277"
  },
  {
    "epoch": 449,
    "batch_idx": 96,
    "loss": 0.476662814617157,
    "learning_rate": 0.000880712588838103,
    "timestep_mean": 499.46875,
    "epsilon_norm": 8.349723815917969,
    "prediction_norm": 6.26164436340332,
    "timestamp": "2025-07-14T12:33:49.392439"
  },
  {
    "epoch": 449,
    "batch_idx": 128,
    "loss": 0.27399587631225586,
    "learning_rate": 0.000880712588838103,
    "timestep_mean": 491.375,
    "epsilon_norm": 7.387768745422363,
    "prediction_norm": 6.217624187469482,
    "timestamp": "2025-07-14T12:33:49.397530"
  },
  {
    "epoch": 449,
    "batch_idx": 160,
    "loss": 0.22234712541103363,
    "learning_rate": 0.000880712588838103,
    "timestep_mean": 497.125,
    "epsilon_norm": 7.335827827453613,
    "prediction_norm": 6.917712688446045,
    "timestamp": "2025-07-14T12:33:49.402764"
  },
  {
    "epoch": 449,
    "batch_idx": 192,
    "loss": 0.09613484144210815,
    "learning_rate": 0.000880712588838103,
    "timestep_mean": 635.5,
    "epsilon_norm": 4.802093982696533,
    "prediction_norm": 4.445836544036865,
    "timestamp": "2025-07-14T12:33:49.407625"
  },
  {
    "epoch": 450,
    "batch_idx": 0,
    "loss": 0.2692754566669464,
    "learning_rate": 0.0008802029828000153,
    "timestep_mean": 490.65625,
    "epsilon_norm": 6.763337135314941,
    "prediction_norm": 5.323863506317139,
    "timestamp": "2025-07-14T12:33:49.412577"
  },
  {
    "epoch": 450,
    "batch_idx": 32,
    "loss": 0.4297334551811218,
    "learning_rate": 0.0008802029828000153,
    "timestep_mean": 528.5625,
    "epsilon_norm": 8.981826782226562,
    "prediction_norm": 6.841155052185059,
    "timestamp": "2025-07-14T12:33:49.417789"
  },
  {
    "epoch": 450,
    "batch_idx": 64,
    "loss": 0.329389750957489,
    "learning_rate": 0.0008802029828000153,
    "timestep_mean": 479.5625,
    "epsilon_norm": 7.6921234130859375,
    "prediction_norm": 5.706376075744629,
    "timestamp": "2025-07-14T12:33:49.423039"
  },
  {
    "epoch": 450,
    "batch_idx": 96,
    "loss": 0.5223479270935059,
    "learning_rate": 0.0008802029828000153,
    "timestep_mean": 445.375,
    "epsilon_norm": 8.495811462402344,
    "prediction_norm": 6.284539699554443,
    "timestamp": "2025-07-14T12:33:49.428248"
  },
  {
    "epoch": 450,
    "batch_idx": 128,
    "loss": 0.2133837342262268,
    "learning_rate": 0.0008802029828000153,
    "timestep_mean": 487.59375,
    "epsilon_norm": 7.517237663269043,
    "prediction_norm": 6.204813480377197,
    "timestamp": "2025-07-14T12:33:49.433434"
  },
  {
    "epoch": 450,
    "batch_idx": 160,
    "loss": 0.1975640058517456,
    "learning_rate": 0.0008802029828000153,
    "timestep_mean": 531.25,
    "epsilon_norm": 8.200984954833984,
    "prediction_norm": 6.823150634765625,
    "timestamp": "2025-07-14T12:33:49.438624"
  },
  {
    "epoch": 450,
    "batch_idx": 192,
    "loss": 0.3040016293525696,
    "learning_rate": 0.0008802029828000153,
    "timestep_mean": 342.0,
    "epsilon_norm": 3.0912020206451416,
    "prediction_norm": 2.371474027633667,
    "timestamp": "2025-07-14T12:33:49.443781"
  },
  {
    "epoch": 451,
    "batch_idx": 0,
    "loss": 0.34191185235977173,
    "learning_rate": 0.000879692438648862,
    "timestep_mean": 492.40625,
    "epsilon_norm": 8.171806335449219,
    "prediction_norm": 6.557376384735107,
    "timestamp": "2025-07-14T12:33:49.449122"
  },
  {
    "epoch": 451,
    "batch_idx": 32,
    "loss": 0.31210994720458984,
    "learning_rate": 0.000879692438648862,
    "timestep_mean": 443.71875,
    "epsilon_norm": 7.155932903289795,
    "prediction_norm": 5.218804359436035,
    "timestamp": "2025-07-14T12:33:49.454279"
  },
  {
    "epoch": 451,
    "batch_idx": 64,
    "loss": 0.5335365533828735,
    "learning_rate": 0.000879692438648862,
    "timestep_mean": 492.0625,
    "epsilon_norm": 9.30517578125,
    "prediction_norm": 7.311218738555908,
    "timestamp": "2025-07-14T12:33:49.459291"
  },
  {
    "epoch": 451,
    "batch_idx": 96,
    "loss": 0.2449362874031067,
    "learning_rate": 0.000879692438648862,
    "timestep_mean": 482.0625,
    "epsilon_norm": 7.804734706878662,
    "prediction_norm": 6.906010627746582,
    "timestamp": "2025-07-14T12:33:49.464486"
  },
  {
    "epoch": 451,
    "batch_idx": 128,
    "loss": 0.37453293800354004,
    "learning_rate": 0.000879692438648862,
    "timestep_mean": 485.90625,
    "epsilon_norm": 8.480931282043457,
    "prediction_norm": 7.205756664276123,
    "timestamp": "2025-07-14T12:33:49.469677"
  },
  {
    "epoch": 451,
    "batch_idx": 160,
    "loss": 0.34660789370536804,
    "learning_rate": 0.000879692438648862,
    "timestep_mean": 465.53125,
    "epsilon_norm": 7.666535377502441,
    "prediction_norm": 6.853784561157227,
    "timestamp": "2025-07-14T12:33:49.475004"
  },
  {
    "epoch": 451,
    "batch_idx": 192,
    "loss": 0.2080363631248474,
    "learning_rate": 0.000879692438648862,
    "timestep_mean": 413.25,
    "epsilon_norm": 4.208456516265869,
    "prediction_norm": 4.1275858879089355,
    "timestamp": "2025-07-14T12:33:49.480207"
  },
  {
    "epoch": 452,
    "batch_idx": 0,
    "loss": 0.3498073220252991,
    "learning_rate": 0.0008791809576443608,
    "timestep_mean": 484.6875,
    "epsilon_norm": 8.149600982666016,
    "prediction_norm": 7.0371222496032715,
    "timestamp": "2025-07-14T12:33:49.485485"
  },
  {
    "epoch": 452,
    "batch_idx": 32,
    "loss": 0.20709599554538727,
    "learning_rate": 0.0008791809576443608,
    "timestep_mean": 525.3125,
    "epsilon_norm": 7.980923652648926,
    "prediction_norm": 7.408085823059082,
    "timestamp": "2025-07-14T12:33:49.490657"
  },
  {
    "epoch": 452,
    "batch_idx": 64,
    "loss": 0.2730843424797058,
    "learning_rate": 0.0008791809576443608,
    "timestep_mean": 569.25,
    "epsilon_norm": 8.919184684753418,
    "prediction_norm": 7.299042224884033,
    "timestamp": "2025-07-14T12:33:49.495829"
  },
  {
    "epoch": 452,
    "batch_idx": 96,
    "loss": 0.3246113657951355,
    "learning_rate": 0.0008791809576443608,
    "timestep_mean": 503.0625,
    "epsilon_norm": 7.588912487030029,
    "prediction_norm": 6.988216400146484,
    "timestamp": "2025-07-14T12:33:49.500828"
  },
  {
    "epoch": 452,
    "batch_idx": 128,
    "loss": 0.6494501829147339,
    "learning_rate": 0.0008791809576443608,
    "timestep_mean": 389.125,
    "epsilon_norm": 8.426901817321777,
    "prediction_norm": 5.968725204467773,
    "timestamp": "2025-07-14T12:33:49.510819"
  },
  {
    "epoch": 452,
    "batch_idx": 160,
    "loss": 0.27453166246414185,
    "learning_rate": 0.0008791809576443608,
    "timestep_mean": 494.3125,
    "epsilon_norm": 7.7733845710754395,
    "prediction_norm": 6.240138053894043,
    "timestamp": "2025-07-14T12:33:49.516030"
  },
  {
    "epoch": 452,
    "batch_idx": 192,
    "loss": 0.2293865829706192,
    "learning_rate": 0.0008791809576443608,
    "timestep_mean": 331.0,
    "epsilon_norm": 3.38580060005188,
    "prediction_norm": 2.6350905895233154,
    "timestamp": "2025-07-14T12:33:49.521341"
  },
  {
    "epoch": 453,
    "batch_idx": 0,
    "loss": 0.4179864525794983,
    "learning_rate": 0.0008786685410485395,
    "timestep_mean": 412.8125,
    "epsilon_norm": 9.258298873901367,
    "prediction_norm": 6.847846031188965,
    "timestamp": "2025-07-14T12:33:49.526504"
  },
  {
    "epoch": 453,
    "batch_idx": 32,
    "loss": 0.24027733504772186,
    "learning_rate": 0.0008786685410485395,
    "timestep_mean": 539.4375,
    "epsilon_norm": 7.872036457061768,
    "prediction_norm": 7.123576641082764,
    "timestamp": "2025-07-14T12:33:49.531794"
  },
  {
    "epoch": 453,
    "batch_idx": 64,
    "loss": 0.2628713846206665,
    "learning_rate": 0.0008786685410485395,
    "timestep_mean": 477.1875,
    "epsilon_norm": 7.147491455078125,
    "prediction_norm": 6.526827335357666,
    "timestamp": "2025-07-14T12:33:49.537046"
  },
  {
    "epoch": 453,
    "batch_idx": 96,
    "loss": 0.2551678419113159,
    "learning_rate": 0.0008786685410485395,
    "timestep_mean": 476.75,
    "epsilon_norm": 8.581358909606934,
    "prediction_norm": 6.982464790344238,
    "timestamp": "2025-07-14T12:33:49.542458"
  },
  {
    "epoch": 453,
    "batch_idx": 128,
    "loss": 0.17932501435279846,
    "learning_rate": 0.0008786685410485395,
    "timestep_mean": 485.75,
    "epsilon_norm": 7.40509033203125,
    "prediction_norm": 6.464811325073242,
    "timestamp": "2025-07-14T12:33:49.547630"
  },
  {
    "epoch": 453,
    "batch_idx": 160,
    "loss": 0.2107815146446228,
    "learning_rate": 0.0008786685410485395,
    "timestep_mean": 543.28125,
    "epsilon_norm": 6.83091926574707,
    "prediction_norm": 6.275549411773682,
    "timestamp": "2025-07-14T12:33:49.552754"
  },
  {
    "epoch": 453,
    "batch_idx": 192,
    "loss": 0.052063778042793274,
    "learning_rate": 0.0008786685410485395,
    "timestep_mean": 628.5,
    "epsilon_norm": 3.9251718521118164,
    "prediction_norm": 3.8699240684509277,
    "timestamp": "2025-07-14T12:33:49.557884"
  },
  {
    "epoch": 454,
    "batch_idx": 0,
    "loss": 0.572910726070404,
    "learning_rate": 0.0008781551901257358,
    "timestep_mean": 359.9375,
    "epsilon_norm": 7.5648512840271,
    "prediction_norm": 5.228195667266846,
    "timestamp": "2025-07-14T12:33:49.563213"
  },
  {
    "epoch": 454,
    "batch_idx": 32,
    "loss": 0.2506491541862488,
    "learning_rate": 0.0008781551901257358,
    "timestep_mean": 436.03125,
    "epsilon_norm": 7.652312278747559,
    "prediction_norm": 6.117155075073242,
    "timestamp": "2025-07-14T12:33:49.568431"
  },
  {
    "epoch": 454,
    "batch_idx": 64,
    "loss": 0.6265476942062378,
    "learning_rate": 0.0008781551901257358,
    "timestep_mean": 523.8125,
    "epsilon_norm": 9.182351112365723,
    "prediction_norm": 6.215310096740723,
    "timestamp": "2025-07-14T12:33:49.573619"
  },
  {
    "epoch": 454,
    "batch_idx": 96,
    "loss": 0.3590298295021057,
    "learning_rate": 0.0008781551901257358,
    "timestep_mean": 513.0,
    "epsilon_norm": 9.237278938293457,
    "prediction_norm": 7.863007545471191,
    "timestamp": "2025-07-14T12:33:49.578798"
  },
  {
    "epoch": 454,
    "batch_idx": 128,
    "loss": 0.328624427318573,
    "learning_rate": 0.0008781551901257358,
    "timestep_mean": 480.0,
    "epsilon_norm": 7.221337795257568,
    "prediction_norm": 6.464920520782471,
    "timestamp": "2025-07-14T12:33:49.584074"
  },
  {
    "epoch": 454,
    "batch_idx": 160,
    "loss": 0.33839812874794006,
    "learning_rate": 0.0008781551901257358,
    "timestep_mean": 470.25,
    "epsilon_norm": 7.714804649353027,
    "prediction_norm": 6.313314437866211,
    "timestamp": "2025-07-14T12:33:49.589179"
  },
  {
    "epoch": 454,
    "batch_idx": 192,
    "loss": 0.23204785585403442,
    "learning_rate": 0.0008781551901257358,
    "timestep_mean": 390.375,
    "epsilon_norm": 3.213386058807373,
    "prediction_norm": 3.3990890979766846,
    "timestamp": "2025-07-14T12:33:49.594197"
  },
  {
    "epoch": 455,
    "batch_idx": 0,
    "loss": 0.46844956278800964,
    "learning_rate": 0.0008776409061425916,
    "timestep_mean": 504.25,
    "epsilon_norm": 7.859203338623047,
    "prediction_norm": 5.9065375328063965,
    "timestamp": "2025-07-14T12:33:49.599525"
  },
  {
    "epoch": 455,
    "batch_idx": 32,
    "loss": 0.3785182237625122,
    "learning_rate": 0.0008776409061425916,
    "timestep_mean": 439.125,
    "epsilon_norm": 7.79528284072876,
    "prediction_norm": 6.13303279876709,
    "timestamp": "2025-07-14T12:33:49.604946"
  },
  {
    "epoch": 455,
    "batch_idx": 64,
    "loss": 0.2324114441871643,
    "learning_rate": 0.0008776409061425916,
    "timestep_mean": 504.96875,
    "epsilon_norm": 8.405913352966309,
    "prediction_norm": 7.78365421295166,
    "timestamp": "2025-07-14T12:33:49.610118"
  },
  {
    "epoch": 455,
    "batch_idx": 96,
    "loss": 0.2628238797187805,
    "learning_rate": 0.0008776409061425916,
    "timestep_mean": 572.71875,
    "epsilon_norm": 7.5283589363098145,
    "prediction_norm": 5.774229526519775,
    "timestamp": "2025-07-14T12:33:49.615408"
  },
  {
    "epoch": 455,
    "batch_idx": 128,
    "loss": 0.2153058499097824,
    "learning_rate": 0.0008776409061425916,
    "timestep_mean": 526.25,
    "epsilon_norm": 7.866270065307617,
    "prediction_norm": 6.628986358642578,
    "timestamp": "2025-07-14T12:33:49.620560"
  },
  {
    "epoch": 455,
    "batch_idx": 160,
    "loss": 0.25502026081085205,
    "learning_rate": 0.0008776409061425916,
    "timestep_mean": 452.53125,
    "epsilon_norm": 7.613289833068848,
    "prediction_norm": 6.776162624359131,
    "timestamp": "2025-07-14T12:33:49.625839"
  },
  {
    "epoch": 455,
    "batch_idx": 192,
    "loss": 0.2994926869869232,
    "learning_rate": 0.0008776409061425916,
    "timestep_mean": 579.125,
    "epsilon_norm": 4.2820587158203125,
    "prediction_norm": 3.456346273422241,
    "timestamp": "2025-07-14T12:33:49.630996"
  },
  {
    "epoch": 456,
    "batch_idx": 0,
    "loss": 0.2584552764892578,
    "learning_rate": 0.0008771256903680518,
    "timestep_mean": 463.6875,
    "epsilon_norm": 7.3816633224487305,
    "prediction_norm": 6.5680413246154785,
    "timestamp": "2025-07-14T12:33:49.636253"
  },
  {
    "epoch": 456,
    "batch_idx": 32,
    "loss": 0.24634288251399994,
    "learning_rate": 0.0008771256903680518,
    "timestep_mean": 470.625,
    "epsilon_norm": 10.13056755065918,
    "prediction_norm": 8.46495246887207,
    "timestamp": "2025-07-14T12:33:49.641555"
  },
  {
    "epoch": 456,
    "batch_idx": 64,
    "loss": 0.2413218468427658,
    "learning_rate": 0.0008771256903680518,
    "timestep_mean": 536.125,
    "epsilon_norm": 9.186793327331543,
    "prediction_norm": 8.018502235412598,
    "timestamp": "2025-07-14T12:33:49.646742"
  },
  {
    "epoch": 456,
    "batch_idx": 96,
    "loss": 0.3388284146785736,
    "learning_rate": 0.0008771256903680518,
    "timestep_mean": 498.3125,
    "epsilon_norm": 7.9652862548828125,
    "prediction_norm": 6.400528430938721,
    "timestamp": "2025-07-14T12:33:49.651919"
  },
  {
    "epoch": 456,
    "batch_idx": 128,
    "loss": 0.2754468321800232,
    "learning_rate": 0.0008771256903680518,
    "timestep_mean": 510.84375,
    "epsilon_norm": 7.944580554962158,
    "prediction_norm": 6.577991008758545,
    "timestamp": "2025-07-14T12:33:49.657230"
  },
  {
    "epoch": 456,
    "batch_idx": 160,
    "loss": 0.4076806306838989,
    "learning_rate": 0.0008771256903680518,
    "timestep_mean": 455.53125,
    "epsilon_norm": 8.499672889709473,
    "prediction_norm": 6.420517444610596,
    "timestamp": "2025-07-14T12:33:49.662407"
  },
  {
    "epoch": 456,
    "batch_idx": 192,
    "loss": 0.1809757947921753,
    "learning_rate": 0.0008771256903680518,
    "timestep_mean": 447.5,
    "epsilon_norm": 3.8696675300598145,
    "prediction_norm": 3.314884901046753,
    "timestamp": "2025-07-14T12:33:49.667562"
  },
  {
    "epoch": 457,
    "batch_idx": 0,
    "loss": 0.14795838296413422,
    "learning_rate": 0.0008766095440733598,
    "timestep_mean": 538.09375,
    "epsilon_norm": 8.165775299072266,
    "prediction_norm": 7.483279705047607,
    "timestamp": "2025-07-14T12:33:49.672888"
  },
  {
    "epoch": 457,
    "batch_idx": 32,
    "loss": 0.22772139310836792,
    "learning_rate": 0.0008766095440733598,
    "timestep_mean": 470.0625,
    "epsilon_norm": 6.541994094848633,
    "prediction_norm": 5.433615684509277,
    "timestamp": "2025-07-14T12:33:49.678151"
  },
  {
    "epoch": 457,
    "batch_idx": 64,
    "loss": 0.2534133791923523,
    "learning_rate": 0.0008766095440733598,
    "timestep_mean": 538.1875,
    "epsilon_norm": 7.821144104003906,
    "prediction_norm": 6.931856155395508,
    "timestamp": "2025-07-14T12:33:49.683300"
  },
  {
    "epoch": 457,
    "batch_idx": 96,
    "loss": 0.25870317220687866,
    "learning_rate": 0.0008766095440733598,
    "timestep_mean": 484.6875,
    "epsilon_norm": 8.928873062133789,
    "prediction_norm": 8.238809585571289,
    "timestamp": "2025-07-14T12:33:49.688183"
  },
  {
    "epoch": 457,
    "batch_idx": 128,
    "loss": 0.34167569875717163,
    "learning_rate": 0.0008766095440733598,
    "timestep_mean": 525.5625,
    "epsilon_norm": 8.974119186401367,
    "prediction_norm": 7.41075325012207,
    "timestamp": "2025-07-14T12:33:49.693102"
  },
  {
    "epoch": 457,
    "batch_idx": 160,
    "loss": 0.34176450967788696,
    "learning_rate": 0.0008766095440733598,
    "timestep_mean": 428.78125,
    "epsilon_norm": 8.42817211151123,
    "prediction_norm": 6.912989616394043,
    "timestamp": "2025-07-14T12:33:49.697933"
  },
  {
    "epoch": 457,
    "batch_idx": 192,
    "loss": 0.24253374338150024,
    "learning_rate": 0.0008766095440733598,
    "timestep_mean": 384.0,
    "epsilon_norm": 4.253870964050293,
    "prediction_norm": 3.524350166320801,
    "timestamp": "2025-07-14T12:33:49.702779"
  },
  {
    "epoch": 458,
    "batch_idx": 0,
    "loss": 0.19998939335346222,
    "learning_rate": 0.0008760924685320556,
    "timestep_mean": 538.125,
    "epsilon_norm": 7.722532749176025,
    "prediction_norm": 7.030046463012695,
    "timestamp": "2025-07-14T12:33:49.707657"
  },
  {
    "epoch": 458,
    "batch_idx": 32,
    "loss": 0.39164066314697266,
    "learning_rate": 0.0008760924685320556,
    "timestep_mean": 517.8125,
    "epsilon_norm": 9.782060623168945,
    "prediction_norm": 8.147074699401855,
    "timestamp": "2025-07-14T12:33:49.712722"
  },
  {
    "epoch": 458,
    "batch_idx": 64,
    "loss": 0.3463515639305115,
    "learning_rate": 0.0008760924685320556,
    "timestep_mean": 455.375,
    "epsilon_norm": 8.000442504882812,
    "prediction_norm": 6.505074977874756,
    "timestamp": "2025-07-14T12:33:49.717571"
  },
  {
    "epoch": 458,
    "batch_idx": 96,
    "loss": 0.2453068494796753,
    "learning_rate": 0.0008760924685320556,
    "timestep_mean": 549.625,
    "epsilon_norm": 8.108680725097656,
    "prediction_norm": 7.10120153427124,
    "timestamp": "2025-07-14T12:33:49.722499"
  },
  {
    "epoch": 458,
    "batch_idx": 128,
    "loss": 0.2746100425720215,
    "learning_rate": 0.0008760924685320556,
    "timestep_mean": 580.96875,
    "epsilon_norm": 7.346329689025879,
    "prediction_norm": 6.695780277252197,
    "timestamp": "2025-07-14T12:33:49.727311"
  },
  {
    "epoch": 458,
    "batch_idx": 160,
    "loss": 0.4555043578147888,
    "learning_rate": 0.0008760924685320556,
    "timestep_mean": 460.5,
    "epsilon_norm": 8.364243507385254,
    "prediction_norm": 6.706384658813477,
    "timestamp": "2025-07-14T12:33:49.732128"
  },
  {
    "epoch": 458,
    "batch_idx": 192,
    "loss": 0.8001914024353027,
    "learning_rate": 0.0008760924685320556,
    "timestep_mean": 372.0,
    "epsilon_norm": 3.7122743129730225,
    "prediction_norm": 1.6794184446334839,
    "timestamp": "2025-07-14T12:33:49.736968"
  },
  {
    "epoch": 459,
    "batch_idx": 0,
    "loss": 0.25183671712875366,
    "learning_rate": 0.0008755744650199715,
    "timestep_mean": 477.96875,
    "epsilon_norm": 8.17023754119873,
    "prediction_norm": 7.466386318206787,
    "timestamp": "2025-07-14T12:33:49.741912"
  },
  {
    "epoch": 459,
    "batch_idx": 32,
    "loss": 0.3516453504562378,
    "learning_rate": 0.0008755744650199715,
    "timestep_mean": 502.78125,
    "epsilon_norm": 8.894871711730957,
    "prediction_norm": 6.974818706512451,
    "timestamp": "2025-07-14T12:33:49.746775"
  },
  {
    "epoch": 459,
    "batch_idx": 64,
    "loss": 0.28147071599960327,
    "learning_rate": 0.0008755744650199715,
    "timestep_mean": 506.59375,
    "epsilon_norm": 7.906162261962891,
    "prediction_norm": 6.337948322296143,
    "timestamp": "2025-07-14T12:33:49.751646"
  },
  {
    "epoch": 459,
    "batch_idx": 96,
    "loss": 0.28234946727752686,
    "learning_rate": 0.0008755744650199715,
    "timestep_mean": 475.0625,
    "epsilon_norm": 7.890827655792236,
    "prediction_norm": 6.513803005218506,
    "timestamp": "2025-07-14T12:33:49.756502"
  },
  {
    "epoch": 459,
    "batch_idx": 128,
    "loss": 0.2838525176048279,
    "learning_rate": 0.0008755744650199715,
    "timestep_mean": 442.40625,
    "epsilon_norm": 7.107943058013916,
    "prediction_norm": 6.286503791809082,
    "timestamp": "2025-07-14T12:33:49.761407"
  },
  {
    "epoch": 459,
    "batch_idx": 160,
    "loss": 0.32858550548553467,
    "learning_rate": 0.0008755744650199715,
    "timestep_mean": 468.46875,
    "epsilon_norm": 7.249039649963379,
    "prediction_norm": 5.308745861053467,
    "timestamp": "2025-07-14T12:33:49.766280"
  },
  {
    "epoch": 459,
    "batch_idx": 192,
    "loss": 0.35702386498451233,
    "learning_rate": 0.0008755744650199715,
    "timestep_mean": 533.75,
    "epsilon_norm": 3.9957966804504395,
    "prediction_norm": 3.255833148956299,
    "timestamp": "2025-07-14T12:33:49.771259"
  },
  {
    "epoch": 460,
    "batch_idx": 0,
    "loss": 0.23406289517879486,
    "learning_rate": 0.0008750555348152297,
    "timestep_mean": 586.28125,
    "epsilon_norm": 7.5229644775390625,
    "prediction_norm": 6.6490583419799805,
    "timestamp": "2025-07-14T12:33:49.776306"
  },
  {
    "epoch": 460,
    "batch_idx": 32,
    "loss": 0.3519531190395355,
    "learning_rate": 0.0008750555348152297,
    "timestep_mean": 503.59375,
    "epsilon_norm": 8.692090034484863,
    "prediction_norm": 7.940867900848389,
    "timestamp": "2025-07-14T12:33:49.781141"
  },
  {
    "epoch": 460,
    "batch_idx": 64,
    "loss": 0.13231892883777618,
    "learning_rate": 0.0008750555348152297,
    "timestep_mean": 557.84375,
    "epsilon_norm": 7.865764141082764,
    "prediction_norm": 7.838144302368164,
    "timestamp": "2025-07-14T12:33:49.786151"
  },
  {
    "epoch": 460,
    "batch_idx": 96,
    "loss": 0.27285879850387573,
    "learning_rate": 0.0008750555348152297,
    "timestep_mean": 463.375,
    "epsilon_norm": 7.984342575073242,
    "prediction_norm": 6.549251079559326,
    "timestamp": "2025-07-14T12:33:49.790995"
  },
  {
    "epoch": 460,
    "batch_idx": 128,
    "loss": 0.4446534812450409,
    "learning_rate": 0.0008750555348152297,
    "timestep_mean": 346.53125,
    "epsilon_norm": 9.157654762268066,
    "prediction_norm": 7.062966823577881,
    "timestamp": "2025-07-14T12:33:49.795967"
  },
  {
    "epoch": 460,
    "batch_idx": 160,
    "loss": 0.47031140327453613,
    "learning_rate": 0.0008750555348152297,
    "timestep_mean": 465.71875,
    "epsilon_norm": 8.481025695800781,
    "prediction_norm": 6.502326965332031,
    "timestamp": "2025-07-14T12:33:49.800789"
  },
  {
    "epoch": 460,
    "batch_idx": 192,
    "loss": 0.08837655186653137,
    "learning_rate": 0.0008750555348152297,
    "timestep_mean": 531.25,
    "epsilon_norm": 3.1552228927612305,
    "prediction_norm": 2.672384738922119,
    "timestamp": "2025-07-14T12:33:49.805669"
  },
  {
    "epoch": 461,
    "batch_idx": 0,
    "loss": 0.5950208902359009,
    "learning_rate": 0.0008745356791982388,
    "timestep_mean": 315.5625,
    "epsilon_norm": 7.801357269287109,
    "prediction_norm": 4.623563289642334,
    "timestamp": "2025-07-14T12:33:49.810563"
  },
  {
    "epoch": 461,
    "batch_idx": 32,
    "loss": 0.3197459876537323,
    "learning_rate": 0.0008745356791982388,
    "timestep_mean": 533.5,
    "epsilon_norm": 7.977339267730713,
    "prediction_norm": 7.0128607749938965,
    "timestamp": "2025-07-14T12:33:49.815460"
  },
  {
    "epoch": 461,
    "batch_idx": 64,
    "loss": 0.359001100063324,
    "learning_rate": 0.0008745356791982388,
    "timestep_mean": 521.75,
    "epsilon_norm": 7.874016284942627,
    "prediction_norm": 6.92617654800415,
    "timestamp": "2025-07-14T12:33:49.820352"
  },
  {
    "epoch": 461,
    "batch_idx": 96,
    "loss": 0.1786448359489441,
    "learning_rate": 0.0008745356791982388,
    "timestep_mean": 582.25,
    "epsilon_norm": 7.372865676879883,
    "prediction_norm": 6.740109920501709,
    "timestamp": "2025-07-14T12:33:49.825262"
  },
  {
    "epoch": 461,
    "batch_idx": 128,
    "loss": 0.4418841302394867,
    "learning_rate": 0.0008745356791982388,
    "timestep_mean": 500.15625,
    "epsilon_norm": 7.431482791900635,
    "prediction_norm": 5.386451721191406,
    "timestamp": "2025-07-14T12:33:49.830210"
  },
  {
    "epoch": 461,
    "batch_idx": 160,
    "loss": 0.30565229058265686,
    "learning_rate": 0.0008745356791982388,
    "timestep_mean": 533.875,
    "epsilon_norm": 8.112859725952148,
    "prediction_norm": 7.041840553283691,
    "timestamp": "2025-07-14T12:33:49.835047"
  },
  {
    "epoch": 461,
    "batch_idx": 192,
    "loss": 0.07755911350250244,
    "learning_rate": 0.0008745356791982388,
    "timestep_mean": 476.5,
    "epsilon_norm": 4.0286102294921875,
    "prediction_norm": 3.964823007583618,
    "timestamp": "2025-07-14T12:33:49.839970"
  },
  {
    "epoch": 462,
    "batch_idx": 0,
    "loss": 0.16084720194339752,
    "learning_rate": 0.000874014899451691,
    "timestep_mean": 526.53125,
    "epsilon_norm": 8.017923355102539,
    "prediction_norm": 7.535689830780029,
    "timestamp": "2025-07-14T12:33:49.844838"
  },
  {
    "epoch": 462,
    "batch_idx": 32,
    "loss": 0.3343208432197571,
    "learning_rate": 0.000874014899451691,
    "timestep_mean": 453.125,
    "epsilon_norm": 8.12486457824707,
    "prediction_norm": 6.274914264678955,
    "timestamp": "2025-07-14T12:33:49.849803"
  },
  {
    "epoch": 462,
    "batch_idx": 64,
    "loss": 0.3355017304420471,
    "learning_rate": 0.000874014899451691,
    "timestep_mean": 528.625,
    "epsilon_norm": 7.239624500274658,
    "prediction_norm": 5.546292781829834,
    "timestamp": "2025-07-14T12:33:49.854710"
  },
  {
    "epoch": 462,
    "batch_idx": 96,
    "loss": 0.27460777759552,
    "learning_rate": 0.000874014899451691,
    "timestep_mean": 548.0,
    "epsilon_norm": 7.204378604888916,
    "prediction_norm": 5.240395545959473,
    "timestamp": "2025-07-14T12:33:49.859583"
  },
  {
    "epoch": 462,
    "batch_idx": 128,
    "loss": 0.5200434923171997,
    "learning_rate": 0.000874014899451691,
    "timestep_mean": 518.65625,
    "epsilon_norm": 7.819389820098877,
    "prediction_norm": 5.441172122955322,
    "timestamp": "2025-07-14T12:33:49.864390"
  },
  {
    "epoch": 462,
    "batch_idx": 160,
    "loss": 0.3100946545600891,
    "learning_rate": 0.000874014899451691,
    "timestep_mean": 454.03125,
    "epsilon_norm": 7.421916961669922,
    "prediction_norm": 5.788043022155762,
    "timestamp": "2025-07-14T12:33:49.869261"
  },
  {
    "epoch": 462,
    "batch_idx": 192,
    "loss": 0.7519489526748657,
    "learning_rate": 0.000874014899451691,
    "timestep_mean": 510.0,
    "epsilon_norm": 5.0163445472717285,
    "prediction_norm": 2.7558836936950684,
    "timestamp": "2025-07-14T12:33:49.874057"
  },
  {
    "epoch": 463,
    "batch_idx": 0,
    "loss": 0.31371304392814636,
    "learning_rate": 0.0008734931968605586,
    "timestep_mean": 510.84375,
    "epsilon_norm": 8.206391334533691,
    "prediction_norm": 6.350988388061523,
    "timestamp": "2025-07-14T12:33:49.878940"
  },
  {
    "epoch": 463,
    "batch_idx": 32,
    "loss": 0.2461012303829193,
    "learning_rate": 0.0008734931968605586,
    "timestep_mean": 571.6875,
    "epsilon_norm": 8.031256675720215,
    "prediction_norm": 6.52882194519043,
    "timestamp": "2025-07-14T12:33:49.883767"
  },
  {
    "epoch": 463,
    "batch_idx": 64,
    "loss": 0.3891352415084839,
    "learning_rate": 0.0008734931968605586,
    "timestep_mean": 437.4375,
    "epsilon_norm": 8.373466491699219,
    "prediction_norm": 6.791379451751709,
    "timestamp": "2025-07-14T12:33:49.888854"
  },
  {
    "epoch": 463,
    "batch_idx": 96,
    "loss": 0.24742171168327332,
    "learning_rate": 0.0008734931968605586,
    "timestep_mean": 408.4375,
    "epsilon_norm": 8.238405227661133,
    "prediction_norm": 7.394923210144043,
    "timestamp": "2025-07-14T12:33:49.893707"
  },
  {
    "epoch": 463,
    "batch_idx": 128,
    "loss": 0.236745685338974,
    "learning_rate": 0.0008734931968605586,
    "timestep_mean": 532.6875,
    "epsilon_norm": 7.244460582733154,
    "prediction_norm": 6.788948059082031,
    "timestamp": "2025-07-14T12:33:49.898560"
  },
  {
    "epoch": 463,
    "batch_idx": 160,
    "loss": 0.29136964678764343,
    "learning_rate": 0.0008734931968605586,
    "timestep_mean": 561.21875,
    "epsilon_norm": 8.66619873046875,
    "prediction_norm": 7.97808313369751,
    "timestamp": "2025-07-14T12:33:49.903368"
  },
  {
    "epoch": 463,
    "batch_idx": 192,
    "loss": 0.2605149745941162,
    "learning_rate": 0.0008734931968605586,
    "timestep_mean": 392.0,
    "epsilon_norm": 3.3792152404785156,
    "prediction_norm": 3.2469229698181152,
    "timestamp": "2025-07-14T12:33:49.908307"
  },
  {
    "epoch": 464,
    "batch_idx": 0,
    "loss": 0.4018743634223938,
    "learning_rate": 0.0008729705727120909,
    "timestep_mean": 539.5,
    "epsilon_norm": 8.091165542602539,
    "prediction_norm": 6.1057963371276855,
    "timestamp": "2025-07-14T12:33:49.913139"
  },
  {
    "epoch": 464,
    "batch_idx": 32,
    "loss": 0.2844087481498718,
    "learning_rate": 0.0008729705727120909,
    "timestep_mean": 511.40625,
    "epsilon_norm": 8.491654396057129,
    "prediction_norm": 7.758227348327637,
    "timestamp": "2025-07-14T12:33:49.918220"
  },
  {
    "epoch": 464,
    "batch_idx": 64,
    "loss": 0.3193395137786865,
    "learning_rate": 0.0008729705727120909,
    "timestep_mean": 490.53125,
    "epsilon_norm": 7.7373127937316895,
    "prediction_norm": 6.652754783630371,
    "timestamp": "2025-07-14T12:33:49.923047"
  },
  {
    "epoch": 464,
    "batch_idx": 96,
    "loss": 0.3145507574081421,
    "learning_rate": 0.0008729705727120909,
    "timestep_mean": 467.96875,
    "epsilon_norm": 7.8771162033081055,
    "prediction_norm": 7.187670707702637,
    "timestamp": "2025-07-14T12:33:49.928024"
  },
  {
    "epoch": 464,
    "batch_idx": 128,
    "loss": 0.25322261452674866,
    "learning_rate": 0.0008729705727120909,
    "timestep_mean": 539.3125,
    "epsilon_norm": 7.442564487457275,
    "prediction_norm": 6.094521522521973,
    "timestamp": "2025-07-14T12:33:49.932903"
  },
  {
    "epoch": 464,
    "batch_idx": 160,
    "loss": 0.2346833348274231,
    "learning_rate": 0.0008729705727120909,
    "timestep_mean": 460.375,
    "epsilon_norm": 8.288626670837402,
    "prediction_norm": 7.012665748596191,
    "timestamp": "2025-07-14T12:33:49.937856"
  },
  {
    "epoch": 464,
    "batch_idx": 192,
    "loss": 0.4436631202697754,
    "learning_rate": 0.0008729705727120909,
    "timestep_mean": 374.0,
    "epsilon_norm": 3.752225637435913,
    "prediction_norm": 2.170419931411743,
    "timestamp": "2025-07-14T12:33:49.942684"
  },
  {
    "epoch": 465,
    "batch_idx": 0,
    "loss": 0.23453229665756226,
    "learning_rate": 0.0008724470282958108,
    "timestep_mean": 498.4375,
    "epsilon_norm": 7.174046516418457,
    "prediction_norm": 5.957595348358154,
    "timestamp": "2025-07-14T12:33:49.947686"
  },
  {
    "epoch": 465,
    "batch_idx": 32,
    "loss": 0.23696643114089966,
    "learning_rate": 0.0008724470282958108,
    "timestep_mean": 457.65625,
    "epsilon_norm": 7.31289529800415,
    "prediction_norm": 6.2820940017700195,
    "timestamp": "2025-07-14T12:33:49.952510"
  },
  {
    "epoch": 465,
    "batch_idx": 64,
    "loss": 0.5208369493484497,
    "learning_rate": 0.0008724470282958108,
    "timestep_mean": 473.40625,
    "epsilon_norm": 10.061911582946777,
    "prediction_norm": 7.457589149475098,
    "timestamp": "2025-07-14T12:33:49.957345"
  },
  {
    "epoch": 465,
    "batch_idx": 96,
    "loss": 0.33170080184936523,
    "learning_rate": 0.0008724470282958108,
    "timestep_mean": 447.46875,
    "epsilon_norm": 8.028304100036621,
    "prediction_norm": 6.733082294464111,
    "timestamp": "2025-07-14T12:33:49.962154"
  },
  {
    "epoch": 465,
    "batch_idx": 128,
    "loss": 0.3638710379600525,
    "learning_rate": 0.0008724470282958108,
    "timestep_mean": 512.03125,
    "epsilon_norm": 7.143385887145996,
    "prediction_norm": 5.796713829040527,
    "timestamp": "2025-07-14T12:33:49.967109"
  },
  {
    "epoch": 465,
    "batch_idx": 160,
    "loss": 0.31362807750701904,
    "learning_rate": 0.0008724470282958108,
    "timestep_mean": 419.6875,
    "epsilon_norm": 7.709694862365723,
    "prediction_norm": 6.001672744750977,
    "timestamp": "2025-07-14T12:33:49.971952"
  },
  {
    "epoch": 465,
    "batch_idx": 192,
    "loss": 0.10752379149198532,
    "learning_rate": 0.0008724470282958108,
    "timestep_mean": 612.5,
    "epsilon_norm": 4.321716785430908,
    "prediction_norm": 3.6275830268859863,
    "timestamp": "2025-07-14T12:33:49.976842"
  },
  {
    "epoch": 466,
    "batch_idx": 0,
    "loss": 0.5115037560462952,
    "learning_rate": 0.0008719225649035123,
    "timestep_mean": 439.96875,
    "epsilon_norm": 8.267346382141113,
    "prediction_norm": 5.911312580108643,
    "timestamp": "2025-07-14T12:33:49.981712"
  },
  {
    "epoch": 466,
    "batch_idx": 32,
    "loss": 0.4886695146560669,
    "learning_rate": 0.0008719225649035123,
    "timestep_mean": 501.0625,
    "epsilon_norm": 9.869512557983398,
    "prediction_norm": 8.141582489013672,
    "timestamp": "2025-07-14T12:33:49.986526"
  },
  {
    "epoch": 466,
    "batch_idx": 64,
    "loss": 0.37021076679229736,
    "learning_rate": 0.0008719225649035123,
    "timestep_mean": 421.375,
    "epsilon_norm": 7.32914924621582,
    "prediction_norm": 5.282492637634277,
    "timestamp": "2025-07-14T12:33:49.991488"
  },
  {
    "epoch": 466,
    "batch_idx": 96,
    "loss": 0.2912188768386841,
    "learning_rate": 0.0008719225649035123,
    "timestep_mean": 464.4375,
    "epsilon_norm": 9.229649543762207,
    "prediction_norm": 7.53739595413208,
    "timestamp": "2025-07-14T12:33:49.996355"
  },
  {
    "epoch": 466,
    "batch_idx": 128,
    "loss": 0.3300357162952423,
    "learning_rate": 0.0008719225649035123,
    "timestep_mean": 545.21875,
    "epsilon_norm": 8.292885780334473,
    "prediction_norm": 6.883030414581299,
    "timestamp": "2025-07-14T12:33:50.001228"
  },
  {
    "epoch": 466,
    "batch_idx": 160,
    "loss": 0.28537261486053467,
    "learning_rate": 0.0008719225649035123,
    "timestep_mean": 484.59375,
    "epsilon_norm": 6.788822174072266,
    "prediction_norm": 5.533576011657715,
    "timestamp": "2025-07-14T12:33:50.006240"
  },
  {
    "epoch": 466,
    "batch_idx": 192,
    "loss": 0.4132443368434906,
    "learning_rate": 0.0008719225649035123,
    "timestep_mean": 460.875,
    "epsilon_norm": 4.108095169067383,
    "prediction_norm": 2.9320850372314453,
    "timestamp": "2025-07-14T12:33:50.011197"
  },
  {
    "epoch": 467,
    "batch_idx": 0,
    "loss": 0.3744833767414093,
    "learning_rate": 0.0008713971838292567,
    "timestep_mean": 408.9375,
    "epsilon_norm": 7.6163010597229,
    "prediction_norm": 6.371960639953613,
    "timestamp": "2025-07-14T12:33:50.016102"
  },
  {
    "epoch": 467,
    "batch_idx": 32,
    "loss": 0.27237075567245483,
    "learning_rate": 0.0008713971838292567,
    "timestep_mean": 552.375,
    "epsilon_norm": 7.92392635345459,
    "prediction_norm": 7.229348182678223,
    "timestamp": "2025-07-14T12:33:50.021172"
  },
  {
    "epoch": 467,
    "batch_idx": 64,
    "loss": 0.3947789669036865,
    "learning_rate": 0.0008713971838292567,
    "timestep_mean": 423.65625,
    "epsilon_norm": 8.169249534606934,
    "prediction_norm": 6.261952877044678,
    "timestamp": "2025-07-14T12:33:50.026075"
  },
  {
    "epoch": 467,
    "batch_idx": 96,
    "loss": 0.22751864790916443,
    "learning_rate": 0.0008713971838292567,
    "timestep_mean": 504.5,
    "epsilon_norm": 8.343321800231934,
    "prediction_norm": 8.098471641540527,
    "timestamp": "2025-07-14T12:33:50.030957"
  },
  {
    "epoch": 467,
    "batch_idx": 128,
    "loss": 0.42951011657714844,
    "learning_rate": 0.0008713971838292567,
    "timestep_mean": 476.71875,
    "epsilon_norm": 8.620223045349121,
    "prediction_norm": 6.798360347747803,
    "timestamp": "2025-07-14T12:33:50.035906"
  },
  {
    "epoch": 467,
    "batch_idx": 160,
    "loss": 0.26385509967803955,
    "learning_rate": 0.0008713971838292567,
    "timestep_mean": 461.0625,
    "epsilon_norm": 7.766257286071777,
    "prediction_norm": 6.861433029174805,
    "timestamp": "2025-07-14T12:33:50.040877"
  },
  {
    "epoch": 467,
    "batch_idx": 192,
    "loss": 0.2758483290672302,
    "learning_rate": 0.0008713971838292567,
    "timestep_mean": 544.875,
    "epsilon_norm": 4.073853015899658,
    "prediction_norm": 3.9938881397247314,
    "timestamp": "2025-07-14T12:33:50.045846"
  },
  {
    "epoch": 468,
    "batch_idx": 0,
    "loss": 0.41310596466064453,
    "learning_rate": 0.0008708708863693693,
    "timestep_mean": 438.90625,
    "epsilon_norm": 9.33961296081543,
    "prediction_norm": 6.977057933807373,
    "timestamp": "2025-07-14T12:33:50.050802"
  },
  {
    "epoch": 468,
    "batch_idx": 32,
    "loss": 0.25866323709487915,
    "learning_rate": 0.0008708708863693693,
    "timestep_mean": 433.21875,
    "epsilon_norm": 7.289346694946289,
    "prediction_norm": 6.136714935302734,
    "timestamp": "2025-07-14T12:33:50.055648"
  },
  {
    "epoch": 468,
    "batch_idx": 64,
    "loss": 0.3029854893684387,
    "learning_rate": 0.0008708708863693693,
    "timestep_mean": 528.9375,
    "epsilon_norm": 9.047152519226074,
    "prediction_norm": 7.609547138214111,
    "timestamp": "2025-07-14T12:33:50.060592"
  },
  {
    "epoch": 468,
    "batch_idx": 96,
    "loss": 0.3914767801761627,
    "learning_rate": 0.0008708708863693693,
    "timestep_mean": 379.34375,
    "epsilon_norm": 7.800879001617432,
    "prediction_norm": 5.797019958496094,
    "timestamp": "2025-07-14T12:33:50.065434"
  },
  {
    "epoch": 468,
    "batch_idx": 128,
    "loss": 0.4319908022880554,
    "learning_rate": 0.0008708708863693693,
    "timestep_mean": 458.34375,
    "epsilon_norm": 8.27796459197998,
    "prediction_norm": 7.077667236328125,
    "timestamp": "2025-07-14T12:33:50.070335"
  },
  {
    "epoch": 468,
    "batch_idx": 160,
    "loss": 0.2394104301929474,
    "learning_rate": 0.0008708708863693693,
    "timestep_mean": 452.75,
    "epsilon_norm": 8.514176368713379,
    "prediction_norm": 7.493717670440674,
    "timestamp": "2025-07-14T12:33:50.075244"
  },
  {
    "epoch": 468,
    "batch_idx": 192,
    "loss": 0.7978836297988892,
    "learning_rate": 0.0008708708863693693,
    "timestep_mean": 334.875,
    "epsilon_norm": 4.1271562576293945,
    "prediction_norm": 2.305987596511841,
    "timestamp": "2025-07-14T12:33:50.080102"
  },
  {
    "epoch": 469,
    "batch_idx": 0,
    "loss": 0.3981231451034546,
    "learning_rate": 0.0008703436738224372,
    "timestep_mean": 487.21875,
    "epsilon_norm": 7.484806060791016,
    "prediction_norm": 6.0817742347717285,
    "timestamp": "2025-07-14T12:33:50.085018"
  },
  {
    "epoch": 469,
    "batch_idx": 32,
    "loss": 0.21523486077785492,
    "learning_rate": 0.0008703436738224372,
    "timestep_mean": 571.65625,
    "epsilon_norm": 7.694062232971191,
    "prediction_norm": 7.155866622924805,
    "timestamp": "2025-07-14T12:33:50.089969"
  },
  {
    "epoch": 469,
    "batch_idx": 64,
    "loss": 0.43755966424942017,
    "learning_rate": 0.0008703436738224372,
    "timestep_mean": 557.46875,
    "epsilon_norm": 8.255148887634277,
    "prediction_norm": 6.1728291511535645,
    "timestamp": "2025-07-14T12:33:50.094889"
  },
  {
    "epoch": 469,
    "batch_idx": 96,
    "loss": 0.2869570851325989,
    "learning_rate": 0.0008703436738224372,
    "timestep_mean": 520.28125,
    "epsilon_norm": 8.616023063659668,
    "prediction_norm": 7.4374470710754395,
    "timestamp": "2025-07-14T12:33:50.099884"
  },
  {
    "epoch": 469,
    "batch_idx": 128,
    "loss": 0.14379659295082092,
    "learning_rate": 0.0008703436738224372,
    "timestep_mean": 530.34375,
    "epsilon_norm": 7.840397834777832,
    "prediction_norm": 6.856809139251709,
    "timestamp": "2025-07-14T12:33:50.109566"
  },
  {
    "epoch": 469,
    "batch_idx": 160,
    "loss": 0.3629191517829895,
    "learning_rate": 0.0008703436738224372,
    "timestep_mean": 552.21875,
    "epsilon_norm": 8.823677062988281,
    "prediction_norm": 7.183991432189941,
    "timestamp": "2025-07-14T12:33:50.114332"
  },
  {
    "epoch": 469,
    "batch_idx": 192,
    "loss": 0.43818050622940063,
    "learning_rate": 0.0008703436738224372,
    "timestep_mean": 588.75,
    "epsilon_norm": 4.609949111938477,
    "prediction_norm": 4.0452775955200195,
    "timestamp": "2025-07-14T12:33:50.119019"
  },
  {
    "epoch": 470,
    "batch_idx": 0,
    "loss": 0.42749860882759094,
    "learning_rate": 0.0008698155474893046,
    "timestep_mean": 431.1875,
    "epsilon_norm": 8.65583324432373,
    "prediction_norm": 6.696417331695557,
    "timestamp": "2025-07-14T12:33:50.123730"
  },
  {
    "epoch": 470,
    "batch_idx": 32,
    "loss": 0.2703523337841034,
    "learning_rate": 0.0008698155474893046,
    "timestep_mean": 488.21875,
    "epsilon_norm": 7.016371726989746,
    "prediction_norm": 6.070254802703857,
    "timestamp": "2025-07-14T12:33:50.128429"
  },
  {
    "epoch": 470,
    "batch_idx": 64,
    "loss": 0.4396288990974426,
    "learning_rate": 0.0008698155474893046,
    "timestep_mean": 484.125,
    "epsilon_norm": 8.237896919250488,
    "prediction_norm": 7.72207498550415,
    "timestamp": "2025-07-14T12:33:50.133352"
  },
  {
    "epoch": 470,
    "batch_idx": 96,
    "loss": 0.3167303800582886,
    "learning_rate": 0.0008698155474893046,
    "timestep_mean": 477.53125,
    "epsilon_norm": 8.105694770812988,
    "prediction_norm": 7.353856086730957,
    "timestamp": "2025-07-14T12:33:50.138043"
  },
  {
    "epoch": 470,
    "batch_idx": 128,
    "loss": 0.3243848383426666,
    "learning_rate": 0.0008698155474893046,
    "timestep_mean": 517.5,
    "epsilon_norm": 7.87830114364624,
    "prediction_norm": 6.980574131011963,
    "timestamp": "2025-07-14T12:33:50.142786"
  },
  {
    "epoch": 470,
    "batch_idx": 160,
    "loss": 0.22463633120059967,
    "learning_rate": 0.0008698155474893046,
    "timestep_mean": 524.78125,
    "epsilon_norm": 7.3559489250183105,
    "prediction_norm": 5.79324197769165,
    "timestamp": "2025-07-14T12:33:50.147497"
  },
  {
    "epoch": 470,
    "batch_idx": 192,
    "loss": 0.5017445087432861,
    "learning_rate": 0.0008698155474893046,
    "timestep_mean": 479.5,
    "epsilon_norm": 4.022158622741699,
    "prediction_norm": 3.1113266944885254,
    "timestamp": "2025-07-14T12:33:50.152310"
  },
  {
    "epoch": 471,
    "batch_idx": 0,
    "loss": 0.22948987782001495,
    "learning_rate": 0.0008692865086730709,
    "timestep_mean": 543.9375,
    "epsilon_norm": 8.015422821044922,
    "prediction_norm": 6.527677536010742,
    "timestamp": "2025-07-14T12:33:50.157018"
  },
  {
    "epoch": 471,
    "batch_idx": 32,
    "loss": 0.3995361030101776,
    "learning_rate": 0.0008692865086730709,
    "timestep_mean": 516.125,
    "epsilon_norm": 7.722955226898193,
    "prediction_norm": 5.685404300689697,
    "timestamp": "2025-07-14T12:33:50.161733"
  },
  {
    "epoch": 471,
    "batch_idx": 64,
    "loss": 0.23405303061008453,
    "learning_rate": 0.0008692865086730709,
    "timestep_mean": 474.03125,
    "epsilon_norm": 6.556753635406494,
    "prediction_norm": 5.364703178405762,
    "timestamp": "2025-07-14T12:33:50.166599"
  },
  {
    "epoch": 471,
    "batch_idx": 96,
    "loss": 0.3321622610092163,
    "learning_rate": 0.0008692865086730709,
    "timestep_mean": 516.9375,
    "epsilon_norm": 7.847441673278809,
    "prediction_norm": 6.302558422088623,
    "timestamp": "2025-07-14T12:33:50.171315"
  },
  {
    "epoch": 471,
    "batch_idx": 128,
    "loss": 0.23411785066127777,
    "learning_rate": 0.0008692865086730709,
    "timestep_mean": 480.5,
    "epsilon_norm": 6.816795825958252,
    "prediction_norm": 5.449267387390137,
    "timestamp": "2025-07-14T12:33:50.176100"
  },
  {
    "epoch": 471,
    "batch_idx": 160,
    "loss": 0.32656335830688477,
    "learning_rate": 0.0008692865086730709,
    "timestep_mean": 459.59375,
    "epsilon_norm": 6.946283340454102,
    "prediction_norm": 5.220087051391602,
    "timestamp": "2025-07-14T12:33:50.180897"
  },
  {
    "epoch": 471,
    "batch_idx": 192,
    "loss": 0.2233111709356308,
    "learning_rate": 0.0008692865086730709,
    "timestep_mean": 462.125,
    "epsilon_norm": 3.1083927154541016,
    "prediction_norm": 2.252617359161377,
    "timestamp": "2025-07-14T12:33:50.185620"
  },
  {
    "epoch": 472,
    "batch_idx": 0,
    "loss": 0.21677391231060028,
    "learning_rate": 0.0008687565586790868,
    "timestep_mean": 581.84375,
    "epsilon_norm": 9.279668807983398,
    "prediction_norm": 8.8069486618042,
    "timestamp": "2025-07-14T12:33:50.190367"
  },
  {
    "epoch": 472,
    "batch_idx": 32,
    "loss": 0.4425567388534546,
    "learning_rate": 0.0008687565586790868,
    "timestep_mean": 446.96875,
    "epsilon_norm": 7.668220520019531,
    "prediction_norm": 5.380705833435059,
    "timestamp": "2025-07-14T12:33:50.195291"
  },
  {
    "epoch": 472,
    "batch_idx": 64,
    "loss": 0.302365779876709,
    "learning_rate": 0.0008687565586790868,
    "timestep_mean": 498.46875,
    "epsilon_norm": 8.527923583984375,
    "prediction_norm": 7.254311561584473,
    "timestamp": "2025-07-14T12:33:50.200172"
  },
  {
    "epoch": 472,
    "batch_idx": 96,
    "loss": 0.35731035470962524,
    "learning_rate": 0.0008687565586790868,
    "timestep_mean": 454.5625,
    "epsilon_norm": 7.241676330566406,
    "prediction_norm": 5.380042552947998,
    "timestamp": "2025-07-14T12:33:50.204927"
  },
  {
    "epoch": 472,
    "batch_idx": 128,
    "loss": 0.3004232347011566,
    "learning_rate": 0.0008687565586790868,
    "timestep_mean": 528.6875,
    "epsilon_norm": 7.228137016296387,
    "prediction_norm": 5.789812088012695,
    "timestamp": "2025-07-14T12:33:50.209689"
  },
  {
    "epoch": 472,
    "batch_idx": 160,
    "loss": 0.5989596843719482,
    "learning_rate": 0.0008687565586790868,
    "timestep_mean": 489.15625,
    "epsilon_norm": 9.107905387878418,
    "prediction_norm": 6.2027716636657715,
    "timestamp": "2025-07-14T12:33:50.214455"
  },
  {
    "epoch": 472,
    "batch_idx": 192,
    "loss": 0.6192334294319153,
    "learning_rate": 0.0008687565586790868,
    "timestep_mean": 354.25,
    "epsilon_norm": 4.316745758056641,
    "prediction_norm": 2.6763315200805664,
    "timestamp": "2025-07-14T12:33:50.219158"
  },
  {
    "epoch": 473,
    "batch_idx": 0,
    "loss": 0.3606725335121155,
    "learning_rate": 0.000868225698814951,
    "timestep_mean": 536.59375,
    "epsilon_norm": 8.167487144470215,
    "prediction_norm": 6.179544448852539,
    "timestamp": "2025-07-14T12:33:50.223927"
  },
  {
    "epoch": 473,
    "batch_idx": 32,
    "loss": 0.44421452283859253,
    "learning_rate": 0.000868225698814951,
    "timestep_mean": 464.3125,
    "epsilon_norm": 7.271047592163086,
    "prediction_norm": 5.286990165710449,
    "timestamp": "2025-07-14T12:33:50.228749"
  },
  {
    "epoch": 473,
    "batch_idx": 64,
    "loss": 0.3893806040287018,
    "learning_rate": 0.000868225698814951,
    "timestep_mean": 482.3125,
    "epsilon_norm": 6.815140724182129,
    "prediction_norm": 4.817713260650635,
    "timestamp": "2025-07-14T12:33:50.233658"
  },
  {
    "epoch": 473,
    "batch_idx": 96,
    "loss": 0.19132736325263977,
    "learning_rate": 0.000868225698814951,
    "timestep_mean": 511.5625,
    "epsilon_norm": 7.446698188781738,
    "prediction_norm": 6.351417064666748,
    "timestamp": "2025-07-14T12:33:50.238562"
  },
  {
    "epoch": 473,
    "batch_idx": 128,
    "loss": 0.2621312141418457,
    "learning_rate": 0.000868225698814951,
    "timestep_mean": 481.75,
    "epsilon_norm": 6.804652690887451,
    "prediction_norm": 5.1342997550964355,
    "timestamp": "2025-07-14T12:33:50.243446"
  },
  {
    "epoch": 473,
    "batch_idx": 160,
    "loss": 0.3086332380771637,
    "learning_rate": 0.000868225698814951,
    "timestep_mean": 569.3125,
    "epsilon_norm": 7.948792934417725,
    "prediction_norm": 7.011631488800049,
    "timestamp": "2025-07-14T12:33:50.248403"
  },
  {
    "epoch": 473,
    "batch_idx": 192,
    "loss": 0.7782944440841675,
    "learning_rate": 0.000868225698814951,
    "timestep_mean": 503.125,
    "epsilon_norm": 3.770385980606079,
    "prediction_norm": 1.3186661005020142,
    "timestamp": "2025-07-14T12:33:50.253414"
  },
  {
    "epoch": 474,
    "batch_idx": 0,
    "loss": 0.15785983204841614,
    "learning_rate": 0.0008676939303905078,
    "timestep_mean": 580.125,
    "epsilon_norm": 7.473786354064941,
    "prediction_norm": 6.768843650817871,
    "timestamp": "2025-07-14T12:33:50.258694"
  },
  {
    "epoch": 474,
    "batch_idx": 32,
    "loss": 0.36945387721061707,
    "learning_rate": 0.0008676939303905078,
    "timestep_mean": 481.46875,
    "epsilon_norm": 8.185213088989258,
    "prediction_norm": 7.02690315246582,
    "timestamp": "2025-07-14T12:33:50.263906"
  },
  {
    "epoch": 474,
    "batch_idx": 64,
    "loss": 0.38178345561027527,
    "learning_rate": 0.0008676939303905078,
    "timestep_mean": 441.28125,
    "epsilon_norm": 8.457796096801758,
    "prediction_norm": 7.069308280944824,
    "timestamp": "2025-07-14T12:33:50.269100"
  },
  {
    "epoch": 474,
    "batch_idx": 96,
    "loss": 0.09804511070251465,
    "learning_rate": 0.0008676939303905078,
    "timestep_mean": 644.625,
    "epsilon_norm": 6.297600269317627,
    "prediction_norm": 5.928073883056641,
    "timestamp": "2025-07-14T12:33:50.274236"
  },
  {
    "epoch": 474,
    "batch_idx": 128,
    "loss": 0.2203839123249054,
    "learning_rate": 0.0008676939303905078,
    "timestep_mean": 562.65625,
    "epsilon_norm": 7.461888790130615,
    "prediction_norm": 6.763738632202148,
    "timestamp": "2025-07-14T12:33:50.279538"
  },
  {
    "epoch": 474,
    "batch_idx": 160,
    "loss": 0.3389427065849304,
    "learning_rate": 0.0008676939303905078,
    "timestep_mean": 428.65625,
    "epsilon_norm": 8.805913925170898,
    "prediction_norm": 7.535896301269531,
    "timestamp": "2025-07-14T12:33:50.284636"
  },
  {
    "epoch": 474,
    "batch_idx": 192,
    "loss": 0.1775350719690323,
    "learning_rate": 0.0008676939303905078,
    "timestep_mean": 318.875,
    "epsilon_norm": 3.7497000694274902,
    "prediction_norm": 2.907498598098755,
    "timestamp": "2025-07-14T12:33:50.289815"
  },
  {
    "epoch": 475,
    "batch_idx": 0,
    "loss": 0.29069340229034424,
    "learning_rate": 0.0008671612547178426,
    "timestep_mean": 561.625,
    "epsilon_norm": 8.746490478515625,
    "prediction_norm": 6.632490634918213,
    "timestamp": "2025-07-14T12:33:50.295007"
  },
  {
    "epoch": 475,
    "batch_idx": 32,
    "loss": 0.286828875541687,
    "learning_rate": 0.0008671612547178426,
    "timestep_mean": 546.21875,
    "epsilon_norm": 6.9479475021362305,
    "prediction_norm": 5.998127460479736,
    "timestamp": "2025-07-14T12:33:50.300324"
  },
  {
    "epoch": 475,
    "batch_idx": 64,
    "loss": 0.4407501816749573,
    "learning_rate": 0.0008671612547178426,
    "timestep_mean": 502.75,
    "epsilon_norm": 8.239235877990723,
    "prediction_norm": 6.322855472564697,
    "timestamp": "2025-07-14T12:33:50.305431"
  },
  {
    "epoch": 475,
    "batch_idx": 96,
    "loss": 0.2670108675956726,
    "learning_rate": 0.0008671612547178426,
    "timestep_mean": 505.34375,
    "epsilon_norm": 8.00799560546875,
    "prediction_norm": 7.090100288391113,
    "timestamp": "2025-07-14T12:33:50.310588"
  },
  {
    "epoch": 475,
    "batch_idx": 128,
    "loss": 0.42278558015823364,
    "learning_rate": 0.0008671612547178426,
    "timestep_mean": 441.53125,
    "epsilon_norm": 7.972682476043701,
    "prediction_norm": 6.706887722015381,
    "timestamp": "2025-07-14T12:33:50.315800"
  },
  {
    "epoch": 475,
    "batch_idx": 160,
    "loss": 0.4006485342979431,
    "learning_rate": 0.0008671612547178426,
    "timestep_mean": 398.28125,
    "epsilon_norm": 7.791654109954834,
    "prediction_norm": 5.89540958404541,
    "timestamp": "2025-07-14T12:33:50.320975"
  },
  {
    "epoch": 475,
    "batch_idx": 192,
    "loss": 0.6123420000076294,
    "learning_rate": 0.0008671612547178426,
    "timestep_mean": 486.375,
    "epsilon_norm": 4.764242649078369,
    "prediction_norm": 4.107194423675537,
    "timestamp": "2025-07-14T12:33:50.326034"
  },
  {
    "epoch": 476,
    "batch_idx": 0,
    "loss": 0.31375551223754883,
    "learning_rate": 0.0008666276731112798,
    "timestep_mean": 481.09375,
    "epsilon_norm": 6.979750633239746,
    "prediction_norm": 5.214075088500977,
    "timestamp": "2025-07-14T12:33:50.331168"
  },
  {
    "epoch": 476,
    "batch_idx": 32,
    "loss": 0.36773890256881714,
    "learning_rate": 0.0008666276731112798,
    "timestep_mean": 482.59375,
    "epsilon_norm": 9.274646759033203,
    "prediction_norm": 7.816603183746338,
    "timestamp": "2025-07-14T12:33:50.336376"
  },
  {
    "epoch": 476,
    "batch_idx": 64,
    "loss": 0.26768961548805237,
    "learning_rate": 0.0008666276731112798,
    "timestep_mean": 471.75,
    "epsilon_norm": 7.772522449493408,
    "prediction_norm": 6.6727495193481445,
    "timestamp": "2025-07-14T12:33:50.341737"
  },
  {
    "epoch": 476,
    "batch_idx": 96,
    "loss": 0.3134087920188904,
    "learning_rate": 0.0008666276731112798,
    "timestep_mean": 464.96875,
    "epsilon_norm": 6.7344841957092285,
    "prediction_norm": 5.949621200561523,
    "timestamp": "2025-07-14T12:33:50.347006"
  },
  {
    "epoch": 476,
    "batch_idx": 128,
    "loss": 0.3140534460544586,
    "learning_rate": 0.0008666276731112798,
    "timestep_mean": 572.53125,
    "epsilon_norm": 7.6604461669921875,
    "prediction_norm": 6.231040954589844,
    "timestamp": "2025-07-14T12:33:50.352272"
  },
  {
    "epoch": 476,
    "batch_idx": 160,
    "loss": 0.4065176248550415,
    "learning_rate": 0.0008666276731112798,
    "timestep_mean": 454.46875,
    "epsilon_norm": 7.261277675628662,
    "prediction_norm": 5.652075290679932,
    "timestamp": "2025-07-14T12:33:50.357610"
  },
  {
    "epoch": 476,
    "batch_idx": 192,
    "loss": 0.26183199882507324,
    "learning_rate": 0.0008666276731112798,
    "timestep_mean": 467.375,
    "epsilon_norm": 2.9930219650268555,
    "prediction_norm": 2.2931816577911377,
    "timestamp": "2025-07-14T12:33:50.362801"
  },
  {
    "epoch": 477,
    "batch_idx": 0,
    "loss": 0.45441287755966187,
    "learning_rate": 0.000866093186887379,
    "timestep_mean": 480.46875,
    "epsilon_norm": 8.664122581481934,
    "prediction_norm": 6.409491062164307,
    "timestamp": "2025-07-14T12:33:50.367943"
  },
  {
    "epoch": 477,
    "batch_idx": 32,
    "loss": 0.27404385805130005,
    "learning_rate": 0.000866093186887379,
    "timestep_mean": 532.0625,
    "epsilon_norm": 7.904651641845703,
    "prediction_norm": 6.138033866882324,
    "timestamp": "2025-07-14T12:33:50.373141"
  },
  {
    "epoch": 477,
    "batch_idx": 64,
    "loss": 0.3426889479160309,
    "learning_rate": 0.000866093186887379,
    "timestep_mean": 508.09375,
    "epsilon_norm": 8.058138847351074,
    "prediction_norm": 6.264453887939453,
    "timestamp": "2025-07-14T12:33:50.378346"
  },
  {
    "epoch": 477,
    "batch_idx": 96,
    "loss": 0.4537976086139679,
    "learning_rate": 0.000866093186887379,
    "timestep_mean": 515.9375,
    "epsilon_norm": 9.306066513061523,
    "prediction_norm": 7.308419704437256,
    "timestamp": "2025-07-14T12:33:50.383734"
  },
  {
    "epoch": 477,
    "batch_idx": 128,
    "loss": 0.1766022890806198,
    "learning_rate": 0.000866093186887379,
    "timestep_mean": 549.9375,
    "epsilon_norm": 8.344120025634766,
    "prediction_norm": 7.295503616333008,
    "timestamp": "2025-07-14T12:33:50.388918"
  },
  {
    "epoch": 477,
    "batch_idx": 160,
    "loss": 0.38721519708633423,
    "learning_rate": 0.000866093186887379,
    "timestep_mean": 521.96875,
    "epsilon_norm": 9.107487678527832,
    "prediction_norm": 7.128711223602295,
    "timestamp": "2025-07-14T12:33:50.394149"
  },
  {
    "epoch": 477,
    "batch_idx": 192,
    "loss": 0.5154204368591309,
    "learning_rate": 0.000866093186887379,
    "timestep_mean": 295.625,
    "epsilon_norm": 5.092311859130859,
    "prediction_norm": 4.191187381744385,
    "timestamp": "2025-07-14T12:33:50.399413"
  },
  {
    "epoch": 478,
    "batch_idx": 0,
    "loss": 0.24145402014255524,
    "learning_rate": 0.0008655577973649319,
    "timestep_mean": 510.625,
    "epsilon_norm": 8.474753379821777,
    "prediction_norm": 7.555180549621582,
    "timestamp": "2025-07-14T12:33:50.404535"
  },
  {
    "epoch": 478,
    "batch_idx": 32,
    "loss": 0.374596506357193,
    "learning_rate": 0.0008655577973649319,
    "timestep_mean": 547.21875,
    "epsilon_norm": 8.402411460876465,
    "prediction_norm": 6.714964866638184,
    "timestamp": "2025-07-14T12:33:50.409301"
  },
  {
    "epoch": 478,
    "batch_idx": 64,
    "loss": 0.47581058740615845,
    "learning_rate": 0.0008655577973649319,
    "timestep_mean": 479.09375,
    "epsilon_norm": 7.336705207824707,
    "prediction_norm": 5.524991035461426,
    "timestamp": "2025-07-14T12:33:50.414258"
  },
  {
    "epoch": 478,
    "batch_idx": 96,
    "loss": 0.3311106562614441,
    "learning_rate": 0.0008655577973649319,
    "timestep_mean": 472.53125,
    "epsilon_norm": 8.990689277648926,
    "prediction_norm": 7.508254528045654,
    "timestamp": "2025-07-14T12:33:50.419589"
  },
  {
    "epoch": 478,
    "batch_idx": 128,
    "loss": 0.2849142551422119,
    "learning_rate": 0.0008655577973649319,
    "timestep_mean": 571.4375,
    "epsilon_norm": 7.669301509857178,
    "prediction_norm": 6.1853790283203125,
    "timestamp": "2025-07-14T12:33:50.424907"
  },
  {
    "epoch": 478,
    "batch_idx": 160,
    "loss": 0.23847204446792603,
    "learning_rate": 0.0008655577973649319,
    "timestep_mean": 482.96875,
    "epsilon_norm": 7.35598087310791,
    "prediction_norm": 5.424883842468262,
    "timestamp": "2025-07-14T12:33:50.430182"
  },
  {
    "epoch": 478,
    "batch_idx": 192,
    "loss": 0.5274382829666138,
    "learning_rate": 0.0008655577973649319,
    "timestep_mean": 353.125,
    "epsilon_norm": 4.673506736755371,
    "prediction_norm": 3.2284719944000244,
    "timestamp": "2025-07-14T12:33:50.435393"
  },
  {
    "epoch": 479,
    "batch_idx": 0,
    "loss": 0.3386112451553345,
    "learning_rate": 0.0008650215058649586,
    "timestep_mean": 499.59375,
    "epsilon_norm": 8.181309700012207,
    "prediction_norm": 6.530240535736084,
    "timestamp": "2025-07-14T12:33:50.440531"
  },
  {
    "epoch": 479,
    "batch_idx": 32,
    "loss": 0.3555072546005249,
    "learning_rate": 0.0008650215058649586,
    "timestep_mean": 457.5,
    "epsilon_norm": 7.2648701667785645,
    "prediction_norm": 5.233362674713135,
    "timestamp": "2025-07-14T12:33:50.445934"
  },
  {
    "epoch": 479,
    "batch_idx": 64,
    "loss": 0.25768521428108215,
    "learning_rate": 0.0008650215058649586,
    "timestep_mean": 573.78125,
    "epsilon_norm": 7.915111064910889,
    "prediction_norm": 6.376054763793945,
    "timestamp": "2025-07-14T12:33:50.451024"
  },
  {
    "epoch": 479,
    "batch_idx": 96,
    "loss": 0.3884843587875366,
    "learning_rate": 0.0008650215058649586,
    "timestep_mean": 517.28125,
    "epsilon_norm": 7.619157314300537,
    "prediction_norm": 5.622419357299805,
    "timestamp": "2025-07-14T12:33:50.456195"
  },
  {
    "epoch": 479,
    "batch_idx": 128,
    "loss": 0.37383630871772766,
    "learning_rate": 0.0008650215058649586,
    "timestep_mean": 536.90625,
    "epsilon_norm": 8.757311820983887,
    "prediction_norm": 7.2249650955200195,
    "timestamp": "2025-07-14T12:33:50.461377"
  },
  {
    "epoch": 479,
    "batch_idx": 160,
    "loss": 0.42990997433662415,
    "learning_rate": 0.0008650215058649586,
    "timestep_mean": 396.15625,
    "epsilon_norm": 8.378865242004395,
    "prediction_norm": 6.30882453918457,
    "timestamp": "2025-07-14T12:33:50.466709"
  },
  {
    "epoch": 479,
    "batch_idx": 192,
    "loss": 0.6491521000862122,
    "learning_rate": 0.0008650215058649586,
    "timestep_mean": 495.25,
    "epsilon_norm": 4.34619665145874,
    "prediction_norm": 2.676433563232422,
    "timestamp": "2025-07-14T12:33:50.471947"
  },
  {
    "epoch": 480,
    "batch_idx": 0,
    "loss": 0.5754444599151611,
    "learning_rate": 0.0008644843137107056,
    "timestep_mean": 459.0625,
    "epsilon_norm": 8.606310844421387,
    "prediction_norm": 5.950738430023193,
    "timestamp": "2025-07-14T12:33:50.477151"
  },
  {
    "epoch": 480,
    "batch_idx": 32,
    "loss": 0.19447875022888184,
    "learning_rate": 0.0008644843137107056,
    "timestep_mean": 541.75,
    "epsilon_norm": 7.437073230743408,
    "prediction_norm": 6.7778191566467285,
    "timestamp": "2025-07-14T12:33:50.482278"
  },
  {
    "epoch": 480,
    "batch_idx": 64,
    "loss": 0.2844600975513458,
    "learning_rate": 0.0008644843137107056,
    "timestep_mean": 465.34375,
    "epsilon_norm": 7.945069789886475,
    "prediction_norm": 6.946829795837402,
    "timestamp": "2025-07-14T12:33:50.487369"
  },
  {
    "epoch": 480,
    "batch_idx": 96,
    "loss": 0.25595664978027344,
    "learning_rate": 0.0008644843137107056,
    "timestep_mean": 543.90625,
    "epsilon_norm": 7.669236660003662,
    "prediction_norm": 6.365880489349365,
    "timestamp": "2025-07-14T12:33:50.492420"
  },
  {
    "epoch": 480,
    "batch_idx": 128,
    "loss": 0.3767832815647125,
    "learning_rate": 0.0008644843137107056,
    "timestep_mean": 540.8125,
    "epsilon_norm": 8.463977813720703,
    "prediction_norm": 7.143777370452881,
    "timestamp": "2025-07-14T12:33:50.497606"
  },
  {
    "epoch": 480,
    "batch_idx": 160,
    "loss": 0.231656014919281,
    "learning_rate": 0.0008644843137107056,
    "timestep_mean": 499.25,
    "epsilon_norm": 6.223816394805908,
    "prediction_norm": 5.409963607788086,
    "timestamp": "2025-07-14T12:33:50.502701"
  },
  {
    "epoch": 480,
    "batch_idx": 192,
    "loss": 0.40107840299606323,
    "learning_rate": 0.0008644843137107056,
    "timestep_mean": 526.875,
    "epsilon_norm": 3.4428133964538574,
    "prediction_norm": 2.806908369064331,
    "timestamp": "2025-07-14T12:33:50.507835"
  },
  {
    "epoch": 481,
    "batch_idx": 0,
    "loss": 0.3029189109802246,
    "learning_rate": 0.0008639462222276407,
    "timestep_mean": 505.3125,
    "epsilon_norm": 8.599852561950684,
    "prediction_norm": 7.493921279907227,
    "timestamp": "2025-07-14T12:33:50.513056"
  },
  {
    "epoch": 481,
    "batch_idx": 32,
    "loss": 0.3804831802845001,
    "learning_rate": 0.0008639462222276407,
    "timestep_mean": 494.875,
    "epsilon_norm": 7.928156852722168,
    "prediction_norm": 6.08436393737793,
    "timestamp": "2025-07-14T12:33:50.518154"
  },
  {
    "epoch": 481,
    "batch_idx": 64,
    "loss": 0.32560378313064575,
    "learning_rate": 0.0008639462222276407,
    "timestep_mean": 422.90625,
    "epsilon_norm": 7.8596415519714355,
    "prediction_norm": 6.627910614013672,
    "timestamp": "2025-07-14T12:33:50.523470"
  },
  {
    "epoch": 481,
    "batch_idx": 96,
    "loss": 0.4034003019332886,
    "learning_rate": 0.0008639462222276407,
    "timestep_mean": 457.71875,
    "epsilon_norm": 8.37210464477539,
    "prediction_norm": 7.279458999633789,
    "timestamp": "2025-07-14T12:33:50.528653"
  },
  {
    "epoch": 481,
    "batch_idx": 128,
    "loss": 0.30595242977142334,
    "learning_rate": 0.0008639462222276407,
    "timestep_mean": 427.9375,
    "epsilon_norm": 7.631344318389893,
    "prediction_norm": 6.563309192657471,
    "timestamp": "2025-07-14T12:33:50.533910"
  },
  {
    "epoch": 481,
    "batch_idx": 160,
    "loss": 0.3448675870895386,
    "learning_rate": 0.0008639462222276407,
    "timestep_mean": 487.6875,
    "epsilon_norm": 7.559894561767578,
    "prediction_norm": 6.355727672576904,
    "timestamp": "2025-07-14T12:33:50.539168"
  },
  {
    "epoch": 481,
    "batch_idx": 192,
    "loss": 0.5517615675926208,
    "learning_rate": 0.0008639462222276407,
    "timestep_mean": 369.5,
    "epsilon_norm": 4.368862152099609,
    "prediction_norm": 4.190807819366455,
    "timestamp": "2025-07-14T12:33:50.544354"
  },
  {
    "epoch": 482,
    "batch_idx": 0,
    "loss": 0.4234052300453186,
    "learning_rate": 0.0008634072327434513,
    "timestep_mean": 456.4375,
    "epsilon_norm": 9.023333549499512,
    "prediction_norm": 8.517434120178223,
    "timestamp": "2025-07-14T12:33:50.549590"
  },
  {
    "epoch": 482,
    "batch_idx": 32,
    "loss": 0.26517537236213684,
    "learning_rate": 0.0008634072327434513,
    "timestep_mean": 547.1875,
    "epsilon_norm": 8.954773902893066,
    "prediction_norm": 8.30272388458252,
    "timestamp": "2025-07-14T12:33:50.554862"
  },
  {
    "epoch": 482,
    "batch_idx": 64,
    "loss": 0.3760635256767273,
    "learning_rate": 0.0008634072327434513,
    "timestep_mean": 500.4375,
    "epsilon_norm": 8.109829902648926,
    "prediction_norm": 6.3634419441223145,
    "timestamp": "2025-07-14T12:33:50.559941"
  },
  {
    "epoch": 482,
    "batch_idx": 96,
    "loss": 0.3009702265262604,
    "learning_rate": 0.0008634072327434513,
    "timestep_mean": 460.65625,
    "epsilon_norm": 7.775805473327637,
    "prediction_norm": 7.047311782836914,
    "timestamp": "2025-07-14T12:33:50.565160"
  },
  {
    "epoch": 482,
    "batch_idx": 128,
    "loss": 0.35859036445617676,
    "learning_rate": 0.0008634072327434513,
    "timestep_mean": 546.15625,
    "epsilon_norm": 8.961599349975586,
    "prediction_norm": 7.896459102630615,
    "timestamp": "2025-07-14T12:33:50.570370"
  },
  {
    "epoch": 482,
    "batch_idx": 160,
    "loss": 0.41606736183166504,
    "learning_rate": 0.0008634072327434513,
    "timestep_mean": 421.625,
    "epsilon_norm": 8.288036346435547,
    "prediction_norm": 6.266894817352295,
    "timestamp": "2025-07-14T12:33:50.575828"
  },
  {
    "epoch": 482,
    "batch_idx": 192,
    "loss": 0.1682748645544052,
    "learning_rate": 0.0008634072327434513,
    "timestep_mean": 617.5,
    "epsilon_norm": 4.220430850982666,
    "prediction_norm": 4.079795837402344,
    "timestamp": "2025-07-14T12:33:50.580981"
  },
  {
    "epoch": 483,
    "batch_idx": 0,
    "loss": 0.12804538011550903,
    "learning_rate": 0.0008628673465880403,
    "timestep_mean": 535.09375,
    "epsilon_norm": 7.021328926086426,
    "prediction_norm": 5.98304557800293,
    "timestamp": "2025-07-14T12:33:50.586186"
  },
  {
    "epoch": 483,
    "batch_idx": 32,
    "loss": 0.3044136166572571,
    "learning_rate": 0.0008628673465880403,
    "timestep_mean": 466.25,
    "epsilon_norm": 8.27219295501709,
    "prediction_norm": 6.506111145019531,
    "timestamp": "2025-07-14T12:33:50.591526"
  },
  {
    "epoch": 483,
    "batch_idx": 64,
    "loss": 0.10604951530694962,
    "learning_rate": 0.0008628673465880403,
    "timestep_mean": 564.875,
    "epsilon_norm": 7.553529262542725,
    "prediction_norm": 6.963963031768799,
    "timestamp": "2025-07-14T12:33:50.596823"
  },
  {
    "epoch": 483,
    "batch_idx": 96,
    "loss": 0.31863486766815186,
    "learning_rate": 0.0008628673465880403,
    "timestep_mean": 470.8125,
    "epsilon_norm": 7.88496208190918,
    "prediction_norm": 6.097822666168213,
    "timestamp": "2025-07-14T12:33:50.602023"
  },
  {
    "epoch": 483,
    "batch_idx": 128,
    "loss": 0.3680790364742279,
    "learning_rate": 0.0008628673465880403,
    "timestep_mean": 542.53125,
    "epsilon_norm": 7.734004020690918,
    "prediction_norm": 5.8375773429870605,
    "timestamp": "2025-07-14T12:33:50.607127"
  },
  {
    "epoch": 483,
    "batch_idx": 160,
    "loss": 0.3319612145423889,
    "learning_rate": 0.0008628673465880403,
    "timestep_mean": 456.78125,
    "epsilon_norm": 7.547186851501465,
    "prediction_norm": 5.970348358154297,
    "timestamp": "2025-07-14T12:33:50.612273"
  },
  {
    "epoch": 483,
    "batch_idx": 192,
    "loss": 0.359119713306427,
    "learning_rate": 0.0008628673465880403,
    "timestep_mean": 337.875,
    "epsilon_norm": 4.506099700927734,
    "prediction_norm": 3.3300554752349854,
    "timestamp": "2025-07-14T12:33:50.617468"
  },
  {
    "epoch": 484,
    "batch_idx": 0,
    "loss": 0.5041252374649048,
    "learning_rate": 0.0008623265650935231,
    "timestep_mean": 492.125,
    "epsilon_norm": 8.494192123413086,
    "prediction_norm": 6.129557132720947,
    "timestamp": "2025-07-14T12:33:50.622581"
  },
  {
    "epoch": 484,
    "batch_idx": 32,
    "loss": 0.29674232006073,
    "learning_rate": 0.0008623265650935231,
    "timestep_mean": 425.15625,
    "epsilon_norm": 8.188353538513184,
    "prediction_norm": 6.574093818664551,
    "timestamp": "2025-07-14T12:33:50.627816"
  },
  {
    "epoch": 484,
    "batch_idx": 64,
    "loss": 0.44885045289993286,
    "learning_rate": 0.0008623265650935231,
    "timestep_mean": 521.09375,
    "epsilon_norm": 8.297475814819336,
    "prediction_norm": 6.317780017852783,
    "timestamp": "2025-07-14T12:33:50.632924"
  },
  {
    "epoch": 484,
    "batch_idx": 96,
    "loss": 0.3700808882713318,
    "learning_rate": 0.0008623265650935231,
    "timestep_mean": 523.3125,
    "epsilon_norm": 7.707574844360352,
    "prediction_norm": 5.184891223907471,
    "timestamp": "2025-07-14T12:33:50.638094"
  },
  {
    "epoch": 484,
    "batch_idx": 128,
    "loss": 0.2715252637863159,
    "learning_rate": 0.0008623265650935231,
    "timestep_mean": 496.0625,
    "epsilon_norm": 8.78254508972168,
    "prediction_norm": 7.060117721557617,
    "timestamp": "2025-07-14T12:33:50.643276"
  },
  {
    "epoch": 484,
    "batch_idx": 160,
    "loss": 0.387945294380188,
    "learning_rate": 0.0008623265650935231,
    "timestep_mean": 475.09375,
    "epsilon_norm": 7.7991108894348145,
    "prediction_norm": 5.618727684020996,
    "timestamp": "2025-07-14T12:33:50.648424"
  },
  {
    "epoch": 484,
    "batch_idx": 192,
    "loss": 0.24570268392562866,
    "learning_rate": 0.0008623265650935231,
    "timestep_mean": 491.75,
    "epsilon_norm": 3.7567694187164307,
    "prediction_norm": 2.927061080932617,
    "timestamp": "2025-07-14T12:33:50.653622"
  },
  {
    "epoch": 485,
    "batch_idx": 0,
    "loss": 0.32165563106536865,
    "learning_rate": 0.0008617848895942244,
    "timestep_mean": 500.9375,
    "epsilon_norm": 8.343957901000977,
    "prediction_norm": 7.047532558441162,
    "timestamp": "2025-07-14T12:33:50.658904"
  },
  {
    "epoch": 485,
    "batch_idx": 32,
    "loss": 0.49753549695014954,
    "learning_rate": 0.0008617848895942244,
    "timestep_mean": 519.03125,
    "epsilon_norm": 7.755403518676758,
    "prediction_norm": 5.894956588745117,
    "timestamp": "2025-07-14T12:33:50.663991"
  },
  {
    "epoch": 485,
    "batch_idx": 64,
    "loss": 0.19166119396686554,
    "learning_rate": 0.0008617848895942244,
    "timestep_mean": 463.75,
    "epsilon_norm": 7.3991498947143555,
    "prediction_norm": 6.600238800048828,
    "timestamp": "2025-07-14T12:33:50.669228"
  },
  {
    "epoch": 485,
    "batch_idx": 96,
    "loss": 0.31785428524017334,
    "learning_rate": 0.0008617848895942244,
    "timestep_mean": 527.625,
    "epsilon_norm": 7.808377265930176,
    "prediction_norm": 6.7956156730651855,
    "timestamp": "2025-07-14T12:33:50.674316"
  },
  {
    "epoch": 485,
    "batch_idx": 128,
    "loss": 0.4385935664176941,
    "learning_rate": 0.0008617848895942244,
    "timestep_mean": 469.21875,
    "epsilon_norm": 7.586659908294678,
    "prediction_norm": 5.589197158813477,
    "timestamp": "2025-07-14T12:33:50.679604"
  },
  {
    "epoch": 485,
    "batch_idx": 160,
    "loss": 0.31247758865356445,
    "learning_rate": 0.0008617848895942244,
    "timestep_mean": 481.9375,
    "epsilon_norm": 8.263399124145508,
    "prediction_norm": 7.087164878845215,
    "timestamp": "2025-07-14T12:33:50.684793"
  },
  {
    "epoch": 485,
    "batch_idx": 192,
    "loss": 0.18405935168266296,
    "learning_rate": 0.0008617848895942244,
    "timestep_mean": 588.5,
    "epsilon_norm": 3.1675562858581543,
    "prediction_norm": 2.8583974838256836,
    "timestamp": "2025-07-14T12:33:50.689918"
  },
  {
    "epoch": 486,
    "batch_idx": 0,
    "loss": 0.3505718410015106,
    "learning_rate": 0.0008612423214266747,
    "timestep_mean": 517.90625,
    "epsilon_norm": 7.500035285949707,
    "prediction_norm": 6.336836814880371,
    "timestamp": "2025-07-14T12:33:50.700790"
  },
  {
    "epoch": 486,
    "batch_idx": 32,
    "loss": 0.23597028851509094,
    "learning_rate": 0.0008612423214266747,
    "timestep_mean": 535.65625,
    "epsilon_norm": 8.313985824584961,
    "prediction_norm": 7.157595157623291,
    "timestamp": "2025-07-14T12:33:50.705844"
  },
  {
    "epoch": 486,
    "batch_idx": 64,
    "loss": 0.2848537564277649,
    "learning_rate": 0.0008612423214266747,
    "timestep_mean": 511.5,
    "epsilon_norm": 7.281454086303711,
    "prediction_norm": 5.984560012817383,
    "timestamp": "2025-07-14T12:33:50.710883"
  },
  {
    "epoch": 486,
    "batch_idx": 96,
    "loss": 0.303779661655426,
    "learning_rate": 0.0008612423214266747,
    "timestep_mean": 475.46875,
    "epsilon_norm": 8.024267196655273,
    "prediction_norm": 6.902770519256592,
    "timestamp": "2025-07-14T12:33:50.715702"
  },
  {
    "epoch": 486,
    "batch_idx": 128,
    "loss": 0.32231271266937256,
    "learning_rate": 0.0008612423214266747,
    "timestep_mean": 474.875,
    "epsilon_norm": 8.70020580291748,
    "prediction_norm": 6.780557155609131,
    "timestamp": "2025-07-14T12:33:50.720679"
  },
  {
    "epoch": 486,
    "batch_idx": 160,
    "loss": 0.3546519875526428,
    "learning_rate": 0.0008612423214266747,
    "timestep_mean": 558.8125,
    "epsilon_norm": 9.094610214233398,
    "prediction_norm": 7.482278823852539,
    "timestamp": "2025-07-14T12:33:50.725633"
  },
  {
    "epoch": 486,
    "batch_idx": 192,
    "loss": 0.7755923271179199,
    "learning_rate": 0.0008612423214266747,
    "timestep_mean": 423.375,
    "epsilon_norm": 4.572943687438965,
    "prediction_norm": 2.5903563499450684,
    "timestamp": "2025-07-14T12:33:50.730667"
  },
  {
    "epoch": 487,
    "batch_idx": 0,
    "loss": 0.16631212830543518,
    "learning_rate": 0.0008606988619296069,
    "timestep_mean": 546.875,
    "epsilon_norm": 8.085738182067871,
    "prediction_norm": 8.07768440246582,
    "timestamp": "2025-07-14T12:33:50.735697"
  },
  {
    "epoch": 487,
    "batch_idx": 32,
    "loss": 0.6635468602180481,
    "learning_rate": 0.0008606988619296069,
    "timestep_mean": 496.875,
    "epsilon_norm": 8.126782417297363,
    "prediction_norm": 5.656069278717041,
    "timestamp": "2025-07-14T12:33:50.740660"
  },
  {
    "epoch": 487,
    "batch_idx": 64,
    "loss": 0.2072908580303192,
    "learning_rate": 0.0008606988619296069,
    "timestep_mean": 496.75,
    "epsilon_norm": 7.140161514282227,
    "prediction_norm": 6.7001872062683105,
    "timestamp": "2025-07-14T12:33:50.745593"
  },
  {
    "epoch": 487,
    "batch_idx": 96,
    "loss": 0.14674566686153412,
    "learning_rate": 0.0008606988619296069,
    "timestep_mean": 559.65625,
    "epsilon_norm": 7.426112174987793,
    "prediction_norm": 6.9270124435424805,
    "timestamp": "2025-07-14T12:33:50.750601"
  },
  {
    "epoch": 487,
    "batch_idx": 128,
    "loss": 0.17465931177139282,
    "learning_rate": 0.0008606988619296069,
    "timestep_mean": 587.625,
    "epsilon_norm": 8.35788631439209,
    "prediction_norm": 7.962863922119141,
    "timestamp": "2025-07-14T12:33:50.755622"
  },
  {
    "epoch": 487,
    "batch_idx": 160,
    "loss": 0.3836445212364197,
    "learning_rate": 0.0008606988619296069,
    "timestep_mean": 462.8125,
    "epsilon_norm": 8.849173545837402,
    "prediction_norm": 7.18843412399292,
    "timestamp": "2025-07-14T12:33:50.760611"
  },
  {
    "epoch": 487,
    "batch_idx": 192,
    "loss": 0.7879772782325745,
    "learning_rate": 0.0008606988619296069,
    "timestep_mean": 380.875,
    "epsilon_norm": 4.270082950592041,
    "prediction_norm": 1.890707015991211,
    "timestamp": "2025-07-14T12:33:50.765557"
  },
  {
    "epoch": 488,
    "batch_idx": 0,
    "loss": 0.17643031477928162,
    "learning_rate": 0.0008601545124439532,
    "timestep_mean": 554.625,
    "epsilon_norm": 8.049949645996094,
    "prediction_norm": 7.169726371765137,
    "timestamp": "2025-07-14T12:33:50.770604"
  },
  {
    "epoch": 488,
    "batch_idx": 32,
    "loss": 0.17654597759246826,
    "learning_rate": 0.0008601545124439532,
    "timestep_mean": 497.65625,
    "epsilon_norm": 7.489556789398193,
    "prediction_norm": 6.603344440460205,
    "timestamp": "2025-07-14T12:33:50.775646"
  },
  {
    "epoch": 488,
    "batch_idx": 64,
    "loss": 0.28988099098205566,
    "learning_rate": 0.0008601545124439532,
    "timestep_mean": 479.53125,
    "epsilon_norm": 8.392760276794434,
    "prediction_norm": 7.811473369598389,
    "timestamp": "2025-07-14T12:33:50.780588"
  },
  {
    "epoch": 488,
    "batch_idx": 96,
    "loss": 0.30355584621429443,
    "learning_rate": 0.0008601545124439532,
    "timestep_mean": 444.03125,
    "epsilon_norm": 8.039297103881836,
    "prediction_norm": 6.526747226715088,
    "timestamp": "2025-07-14T12:33:50.785665"
  },
  {
    "epoch": 488,
    "batch_idx": 128,
    "loss": 0.19773027300834656,
    "learning_rate": 0.0008601545124439532,
    "timestep_mean": 540.25,
    "epsilon_norm": 6.628207683563232,
    "prediction_norm": 6.061254024505615,
    "timestamp": "2025-07-14T12:33:50.790620"
  },
  {
    "epoch": 488,
    "batch_idx": 160,
    "loss": 0.44619351625442505,
    "learning_rate": 0.0008601545124439532,
    "timestep_mean": 494.1875,
    "epsilon_norm": 8.252483367919922,
    "prediction_norm": 6.136395454406738,
    "timestamp": "2025-07-14T12:33:50.795679"
  },
  {
    "epoch": 488,
    "batch_idx": 192,
    "loss": 0.313102126121521,
    "learning_rate": 0.0008601545124439532,
    "timestep_mean": 219.5,
    "epsilon_norm": 2.399277925491333,
    "prediction_norm": 2.000356912612915,
    "timestamp": "2025-07-14T12:33:50.800673"
  },
  {
    "epoch": 489,
    "batch_idx": 0,
    "loss": 0.3842281997203827,
    "learning_rate": 0.0008596092743128421,
    "timestep_mean": 494.1875,
    "epsilon_norm": 8.613842010498047,
    "prediction_norm": 6.262493133544922,
    "timestamp": "2025-07-14T12:33:50.805670"
  },
  {
    "epoch": 489,
    "batch_idx": 32,
    "loss": 0.266971230506897,
    "learning_rate": 0.0008596092743128421,
    "timestep_mean": 516.6875,
    "epsilon_norm": 8.053566932678223,
    "prediction_norm": 6.365889549255371,
    "timestamp": "2025-07-14T12:33:50.810721"
  },
  {
    "epoch": 489,
    "batch_idx": 64,
    "loss": 0.4076575040817261,
    "learning_rate": 0.0008596092743128421,
    "timestep_mean": 410.3125,
    "epsilon_norm": 8.44168472290039,
    "prediction_norm": 6.552252769470215,
    "timestamp": "2025-07-14T12:33:50.815758"
  },
  {
    "epoch": 489,
    "batch_idx": 96,
    "loss": 0.4119946360588074,
    "learning_rate": 0.0008596092743128421,
    "timestep_mean": 415.1875,
    "epsilon_norm": 7.019932270050049,
    "prediction_norm": 5.237268924713135,
    "timestamp": "2025-07-14T12:33:50.820782"
  },
  {
    "epoch": 489,
    "batch_idx": 128,
    "loss": 0.49200356006622314,
    "learning_rate": 0.0008596092743128421,
    "timestep_mean": 416.46875,
    "epsilon_norm": 8.57534408569336,
    "prediction_norm": 7.125566005706787,
    "timestamp": "2025-07-14T12:33:50.825974"
  },
  {
    "epoch": 489,
    "batch_idx": 160,
    "loss": 0.24127230048179626,
    "learning_rate": 0.0008596092743128421,
    "timestep_mean": 478.21875,
    "epsilon_norm": 7.9061665534973145,
    "prediction_norm": 7.196526527404785,
    "timestamp": "2025-07-14T12:33:50.830965"
  },
  {
    "epoch": 489,
    "batch_idx": 192,
    "loss": 0.11004621535539627,
    "learning_rate": 0.0008596092743128421,
    "timestep_mean": 517.5,
    "epsilon_norm": 3.560483455657959,
    "prediction_norm": 2.88613224029541,
    "timestamp": "2025-07-14T12:33:50.836067"
  },
  {
    "epoch": 490,
    "batch_idx": 0,
    "loss": 0.28138434886932373,
    "learning_rate": 0.0008590631488815943,
    "timestep_mean": 544.4375,
    "epsilon_norm": 8.352497100830078,
    "prediction_norm": 7.36609411239624,
    "timestamp": "2025-07-14T12:33:50.841172"
  },
  {
    "epoch": 490,
    "batch_idx": 32,
    "loss": 0.417388916015625,
    "learning_rate": 0.0008590631488815943,
    "timestep_mean": 450.65625,
    "epsilon_norm": 8.491584777832031,
    "prediction_norm": 6.243702411651611,
    "timestamp": "2025-07-14T12:33:50.846450"
  },
  {
    "epoch": 490,
    "batch_idx": 64,
    "loss": 0.14070549607276917,
    "learning_rate": 0.0008590631488815943,
    "timestep_mean": 531.90625,
    "epsilon_norm": 7.139372825622559,
    "prediction_norm": 6.872193813323975,
    "timestamp": "2025-07-14T12:33:50.851610"
  },
  {
    "epoch": 490,
    "batch_idx": 96,
    "loss": 0.40861034393310547,
    "learning_rate": 0.0008590631488815943,
    "timestep_mean": 547.8125,
    "epsilon_norm": 9.42366886138916,
    "prediction_norm": 7.498275279998779,
    "timestamp": "2025-07-14T12:33:50.856805"
  },
  {
    "epoch": 490,
    "batch_idx": 128,
    "loss": 0.418610155582428,
    "learning_rate": 0.0008590631488815943,
    "timestep_mean": 454.28125,
    "epsilon_norm": 7.839323997497559,
    "prediction_norm": 5.95016622543335,
    "timestamp": "2025-07-14T12:33:50.862011"
  },
  {
    "epoch": 490,
    "batch_idx": 160,
    "loss": 0.2234322428703308,
    "learning_rate": 0.0008590631488815943,
    "timestep_mean": 606.21875,
    "epsilon_norm": 7.522637367248535,
    "prediction_norm": 6.9694342613220215,
    "timestamp": "2025-07-14T12:33:50.867180"
  },
  {
    "epoch": 490,
    "batch_idx": 192,
    "loss": 0.45874306559562683,
    "learning_rate": 0.0008590631488815943,
    "timestep_mean": 323.875,
    "epsilon_norm": 3.01682710647583,
    "prediction_norm": 2.5278358459472656,
    "timestamp": "2025-07-14T12:33:50.872337"
  },
  {
    "epoch": 491,
    "batch_idx": 0,
    "loss": 0.2509349584579468,
    "learning_rate": 0.00085851613749772,
    "timestep_mean": 525.6875,
    "epsilon_norm": 6.891496658325195,
    "prediction_norm": 6.228269100189209,
    "timestamp": "2025-07-14T12:33:50.877550"
  },
  {
    "epoch": 491,
    "batch_idx": 32,
    "loss": 0.2894314229488373,
    "learning_rate": 0.00085851613749772,
    "timestep_mean": 520.21875,
    "epsilon_norm": 7.694441318511963,
    "prediction_norm": 6.659149169921875,
    "timestamp": "2025-07-14T12:33:50.882766"
  },
  {
    "epoch": 491,
    "batch_idx": 64,
    "loss": 0.3364461064338684,
    "learning_rate": 0.00085851613749772,
    "timestep_mean": 409.5625,
    "epsilon_norm": 6.698198318481445,
    "prediction_norm": 5.188449382781982,
    "timestamp": "2025-07-14T12:33:50.888068"
  },
  {
    "epoch": 491,
    "batch_idx": 96,
    "loss": 0.540387749671936,
    "learning_rate": 0.00085851613749772,
    "timestep_mean": 440.25,
    "epsilon_norm": 9.403613090515137,
    "prediction_norm": 7.208248615264893,
    "timestamp": "2025-07-14T12:33:50.893225"
  },
  {
    "epoch": 491,
    "batch_idx": 128,
    "loss": 0.23210129141807556,
    "learning_rate": 0.00085851613749772,
    "timestep_mean": 519.4375,
    "epsilon_norm": 6.95484733581543,
    "prediction_norm": 5.270964622497559,
    "timestamp": "2025-07-14T12:33:50.898577"
  },
  {
    "epoch": 491,
    "batch_idx": 160,
    "loss": 0.49036115407943726,
    "learning_rate": 0.00085851613749772,
    "timestep_mean": 455.28125,
    "epsilon_norm": 8.703160285949707,
    "prediction_norm": 6.696439266204834,
    "timestamp": "2025-07-14T12:33:50.903832"
  },
  {
    "epoch": 491,
    "batch_idx": 192,
    "loss": 0.14858824014663696,
    "learning_rate": 0.00085851613749772,
    "timestep_mean": 521.125,
    "epsilon_norm": 3.743896007537842,
    "prediction_norm": 2.811229944229126,
    "timestamp": "2025-07-14T12:33:50.909052"
  },
  {
    "epoch": 492,
    "batch_idx": 0,
    "loss": 0.31667083501815796,
    "learning_rate": 0.0008579682415109153,
    "timestep_mean": 522.25,
    "epsilon_norm": 7.835806369781494,
    "prediction_norm": 6.340260028839111,
    "timestamp": "2025-07-14T12:33:50.914223"
  },
  {
    "epoch": 492,
    "batch_idx": 32,
    "loss": 0.2567257285118103,
    "learning_rate": 0.0008579682415109153,
    "timestep_mean": 503.71875,
    "epsilon_norm": 6.437739849090576,
    "prediction_norm": 5.451585292816162,
    "timestamp": "2025-07-14T12:33:50.919390"
  },
  {
    "epoch": 492,
    "batch_idx": 64,
    "loss": 0.19699636101722717,
    "learning_rate": 0.0008579682415109153,
    "timestep_mean": 539.09375,
    "epsilon_norm": 7.423292636871338,
    "prediction_norm": 6.477919578552246,
    "timestamp": "2025-07-14T12:33:50.924557"
  },
  {
    "epoch": 492,
    "batch_idx": 96,
    "loss": 0.36275219917297363,
    "learning_rate": 0.0008579682415109153,
    "timestep_mean": 514.25,
    "epsilon_norm": 8.01561164855957,
    "prediction_norm": 5.588317394256592,
    "timestamp": "2025-07-14T12:33:50.929805"
  },
  {
    "epoch": 492,
    "batch_idx": 128,
    "loss": 0.3721974492073059,
    "learning_rate": 0.0008579682415109153,
    "timestep_mean": 539.71875,
    "epsilon_norm": 8.919438362121582,
    "prediction_norm": 6.456262588500977,
    "timestamp": "2025-07-14T12:33:50.934999"
  },
  {
    "epoch": 492,
    "batch_idx": 160,
    "loss": 0.3526419997215271,
    "learning_rate": 0.0008579682415109153,
    "timestep_mean": 476.46875,
    "epsilon_norm": 8.601901054382324,
    "prediction_norm": 6.74104642868042,
    "timestamp": "2025-07-14T12:33:50.940227"
  },
  {
    "epoch": 492,
    "batch_idx": 192,
    "loss": 0.37805062532424927,
    "learning_rate": 0.0008579682415109153,
    "timestep_mean": 577.75,
    "epsilon_norm": 4.218938827514648,
    "prediction_norm": 3.0150742530822754,
    "timestamp": "2025-07-14T12:33:50.945381"
  },
  {
    "epoch": 493,
    "batch_idx": 0,
    "loss": 0.36480459570884705,
    "learning_rate": 0.0008574194622730595,
    "timestep_mean": 444.375,
    "epsilon_norm": 7.528451919555664,
    "prediction_norm": 5.794751167297363,
    "timestamp": "2025-07-14T12:33:50.950756"
  },
  {
    "epoch": 493,
    "batch_idx": 32,
    "loss": 0.36389628052711487,
    "learning_rate": 0.0008574194622730595,
    "timestep_mean": 426.90625,
    "epsilon_norm": 8.233941078186035,
    "prediction_norm": 6.813314914703369,
    "timestamp": "2025-07-14T12:33:50.955750"
  },
  {
    "epoch": 493,
    "batch_idx": 64,
    "loss": 0.2982175946235657,
    "learning_rate": 0.0008574194622730595,
    "timestep_mean": 520.3125,
    "epsilon_norm": 9.3129301071167,
    "prediction_norm": 8.207341194152832,
    "timestamp": "2025-07-14T12:33:50.960641"
  },
  {
    "epoch": 493,
    "batch_idx": 96,
    "loss": 0.2067440152168274,
    "learning_rate": 0.0008574194622730595,
    "timestep_mean": 526.59375,
    "epsilon_norm": 8.20058822631836,
    "prediction_norm": 7.977107524871826,
    "timestamp": "2025-07-14T12:33:50.965887"
  },
  {
    "epoch": 493,
    "batch_idx": 128,
    "loss": 0.307241827249527,
    "learning_rate": 0.0008574194622730595,
    "timestep_mean": 458.65625,
    "epsilon_norm": 7.9983367919921875,
    "prediction_norm": 6.8654704093933105,
    "timestamp": "2025-07-14T12:33:50.971180"
  },
  {
    "epoch": 493,
    "batch_idx": 160,
    "loss": 0.3435947000980377,
    "learning_rate": 0.0008574194622730595,
    "timestep_mean": 505.25,
    "epsilon_norm": 6.537605285644531,
    "prediction_norm": 5.774803638458252,
    "timestamp": "2025-07-14T12:33:50.976369"
  },
  {
    "epoch": 493,
    "batch_idx": 192,
    "loss": 0.06828826665878296,
    "learning_rate": 0.0008574194622730595,
    "timestep_mean": 602.625,
    "epsilon_norm": 3.6564137935638428,
    "prediction_norm": 3.4068078994750977,
    "timestamp": "2025-07-14T12:33:50.981516"
  },
  {
    "epoch": 494,
    "batch_idx": 0,
    "loss": 0.2641909420490265,
    "learning_rate": 0.0008568698011382104,
    "timestep_mean": 488.65625,
    "epsilon_norm": 7.858497619628906,
    "prediction_norm": 6.49350118637085,
    "timestamp": "2025-07-14T12:33:50.986465"
  },
  {
    "epoch": 494,
    "batch_idx": 32,
    "loss": 0.41927996277809143,
    "learning_rate": 0.0008568698011382104,
    "timestep_mean": 516.375,
    "epsilon_norm": 8.331266403198242,
    "prediction_norm": 6.118144512176514,
    "timestamp": "2025-07-14T12:33:50.991349"
  },
  {
    "epoch": 494,
    "batch_idx": 64,
    "loss": 0.3936363756656647,
    "learning_rate": 0.0008568698011382104,
    "timestep_mean": 422.09375,
    "epsilon_norm": 8.606969833374023,
    "prediction_norm": 6.6938300132751465,
    "timestamp": "2025-07-14T12:33:50.996229"
  },
  {
    "epoch": 494,
    "batch_idx": 96,
    "loss": 0.32681459188461304,
    "learning_rate": 0.0008568698011382104,
    "timestep_mean": 523.1875,
    "epsilon_norm": 7.270155906677246,
    "prediction_norm": 6.125894069671631,
    "timestamp": "2025-07-14T12:33:51.001095"
  },
  {
    "epoch": 494,
    "batch_idx": 128,
    "loss": 0.12602770328521729,
    "learning_rate": 0.0008568698011382104,
    "timestep_mean": 623.1875,
    "epsilon_norm": 8.788703918457031,
    "prediction_norm": 7.8913373947143555,
    "timestamp": "2025-07-14T12:33:51.005996"
  },
  {
    "epoch": 494,
    "batch_idx": 160,
    "loss": 0.25582507252693176,
    "learning_rate": 0.0008568698011382104,
    "timestep_mean": 470.5625,
    "epsilon_norm": 8.24924373626709,
    "prediction_norm": 7.179214954376221,
    "timestamp": "2025-07-14T12:33:51.010870"
  },
  {
    "epoch": 494,
    "batch_idx": 192,
    "loss": 0.034418996423482895,
    "learning_rate": 0.0008568698011382104,
    "timestep_mean": 677.375,
    "epsilon_norm": 3.6004040241241455,
    "prediction_norm": 3.5651497840881348,
    "timestamp": "2025-07-14T12:33:51.015751"
  },
  {
    "epoch": 495,
    "batch_idx": 0,
    "loss": 0.2829053997993469,
    "learning_rate": 0.0008563192594626025,
    "timestep_mean": 457.84375,
    "epsilon_norm": 7.905414581298828,
    "prediction_norm": 6.3401713371276855,
    "timestamp": "2025-07-14T12:33:51.020664"
  },
  {
    "epoch": 495,
    "batch_idx": 32,
    "loss": 0.23064756393432617,
    "learning_rate": 0.0008563192594626025,
    "timestep_mean": 498.78125,
    "epsilon_norm": 8.597862243652344,
    "prediction_norm": 7.220711708068848,
    "timestamp": "2025-07-14T12:33:51.025671"
  },
  {
    "epoch": 495,
    "batch_idx": 64,
    "loss": 0.36839890480041504,
    "learning_rate": 0.0008563192594626025,
    "timestep_mean": 503.375,
    "epsilon_norm": 8.265405654907227,
    "prediction_norm": 6.703043460845947,
    "timestamp": "2025-07-14T12:33:51.030582"
  },
  {
    "epoch": 495,
    "batch_idx": 96,
    "loss": 0.43961429595947266,
    "learning_rate": 0.0008563192594626025,
    "timestep_mean": 474.15625,
    "epsilon_norm": 8.562023162841797,
    "prediction_norm": 6.574748992919922,
    "timestamp": "2025-07-14T12:33:51.035504"
  },
  {
    "epoch": 495,
    "batch_idx": 128,
    "loss": 0.11745156347751617,
    "learning_rate": 0.0008563192594626025,
    "timestep_mean": 573.0625,
    "epsilon_norm": 7.537048816680908,
    "prediction_norm": 7.181663513183594,
    "timestamp": "2025-07-14T12:33:51.040392"
  },
  {
    "epoch": 495,
    "batch_idx": 160,
    "loss": 0.2528432607650757,
    "learning_rate": 0.0008563192594626025,
    "timestep_mean": 477.15625,
    "epsilon_norm": 8.096385955810547,
    "prediction_norm": 7.436939239501953,
    "timestamp": "2025-07-14T12:33:51.045189"
  },
  {
    "epoch": 495,
    "batch_idx": 192,
    "loss": 0.17178329825401306,
    "learning_rate": 0.0008563192594626025,
    "timestep_mean": 612.5,
    "epsilon_norm": 3.2134552001953125,
    "prediction_norm": 3.1999545097351074,
    "timestamp": "2025-07-14T12:33:51.050054"
  },
  {
    "epoch": 496,
    "batch_idx": 0,
    "loss": 0.2737083435058594,
    "learning_rate": 0.0008557678386046425,
    "timestep_mean": 416.25,
    "epsilon_norm": 7.587289810180664,
    "prediction_norm": 5.94697904586792,
    "timestamp": "2025-07-14T12:33:51.054916"
  },
  {
    "epoch": 496,
    "batch_idx": 32,
    "loss": 0.3776068389415741,
    "learning_rate": 0.0008557678386046425,
    "timestep_mean": 581.53125,
    "epsilon_norm": 8.834610939025879,
    "prediction_norm": 7.381899356842041,
    "timestamp": "2025-07-14T12:33:51.059799"
  },
  {
    "epoch": 496,
    "batch_idx": 64,
    "loss": 0.2688295841217041,
    "learning_rate": 0.0008557678386046425,
    "timestep_mean": 532.53125,
    "epsilon_norm": 8.111968994140625,
    "prediction_norm": 7.833734035491943,
    "timestamp": "2025-07-14T12:33:51.064669"
  },
  {
    "epoch": 496,
    "batch_idx": 96,
    "loss": 0.16956931352615356,
    "learning_rate": 0.0008557678386046425,
    "timestep_mean": 629.0625,
    "epsilon_norm": 7.530919551849365,
    "prediction_norm": 7.2866950035095215,
    "timestamp": "2025-07-14T12:33:51.069568"
  },
  {
    "epoch": 496,
    "batch_idx": 128,
    "loss": 0.23138892650604248,
    "learning_rate": 0.0008557678386046425,
    "timestep_mean": 470.71875,
    "epsilon_norm": 8.12567138671875,
    "prediction_norm": 7.51487398147583,
    "timestamp": "2025-07-14T12:33:51.074354"
  },
  {
    "epoch": 496,
    "batch_idx": 160,
    "loss": 0.20298510789871216,
    "learning_rate": 0.0008557678386046425,
    "timestep_mean": 560.09375,
    "epsilon_norm": 7.920253276824951,
    "prediction_norm": 7.327599048614502,
    "timestamp": "2025-07-14T12:33:51.079275"
  },
  {
    "epoch": 496,
    "batch_idx": 192,
    "loss": 0.2314799576997757,
    "learning_rate": 0.0008557678386046425,
    "timestep_mean": 534.625,
    "epsilon_norm": 4.232894420623779,
    "prediction_norm": 3.46917462348938,
    "timestamp": "2025-07-14T12:33:51.084083"
  },
  {
    "epoch": 497,
    "batch_idx": 0,
    "loss": 0.09275761246681213,
    "learning_rate": 0.0008552155399249066,
    "timestep_mean": 483.65625,
    "epsilon_norm": 7.559925079345703,
    "prediction_norm": 7.286777973175049,
    "timestamp": "2025-07-14T12:33:51.089012"
  },
  {
    "epoch": 497,
    "batch_idx": 32,
    "loss": 0.23235845565795898,
    "learning_rate": 0.0008552155399249066,
    "timestep_mean": 540.65625,
    "epsilon_norm": 8.325772285461426,
    "prediction_norm": 7.040813446044922,
    "timestamp": "2025-07-14T12:33:51.093821"
  },
  {
    "epoch": 497,
    "batch_idx": 64,
    "loss": 0.32596278190612793,
    "learning_rate": 0.0008552155399249066,
    "timestep_mean": 463.84375,
    "epsilon_norm": 8.840017318725586,
    "prediction_norm": 6.97944974899292,
    "timestamp": "2025-07-14T12:33:51.098676"
  },
  {
    "epoch": 497,
    "batch_idx": 96,
    "loss": 0.30929023027420044,
    "learning_rate": 0.0008552155399249066,
    "timestep_mean": 461.03125,
    "epsilon_norm": 8.422719955444336,
    "prediction_norm": 7.183502197265625,
    "timestamp": "2025-07-14T12:33:51.103520"
  },
  {
    "epoch": 497,
    "batch_idx": 128,
    "loss": 0.17514656484127045,
    "learning_rate": 0.0008552155399249066,
    "timestep_mean": 512.1875,
    "epsilon_norm": 6.779819011688232,
    "prediction_norm": 6.05372428894043,
    "timestamp": "2025-07-14T12:33:51.108409"
  },
  {
    "epoch": 497,
    "batch_idx": 160,
    "loss": 0.37214940786361694,
    "learning_rate": 0.0008552155399249066,
    "timestep_mean": 601.53125,
    "epsilon_norm": 9.055928230285645,
    "prediction_norm": 7.236616134643555,
    "timestamp": "2025-07-14T12:33:51.113224"
  },
  {
    "epoch": 497,
    "batch_idx": 192,
    "loss": 0.818217933177948,
    "learning_rate": 0.0008552155399249066,
    "timestep_mean": 337.125,
    "epsilon_norm": 5.083693027496338,
    "prediction_norm": 2.836153268814087,
    "timestamp": "2025-07-14T12:33:51.118028"
  },
  {
    "epoch": 498,
    "batch_idx": 0,
    "loss": 0.26750069856643677,
    "learning_rate": 0.0008546623647861369,
    "timestep_mean": 582.09375,
    "epsilon_norm": 8.36980152130127,
    "prediction_norm": 7.538743019104004,
    "timestamp": "2025-07-14T12:33:51.122918"
  },
  {
    "epoch": 498,
    "batch_idx": 32,
    "loss": 0.26748937368392944,
    "learning_rate": 0.0008546623647861369,
    "timestep_mean": 555.5,
    "epsilon_norm": 8.367752075195312,
    "prediction_norm": 7.18582010269165,
    "timestamp": "2025-07-14T12:33:51.127757"
  },
  {
    "epoch": 498,
    "batch_idx": 64,
    "loss": 0.2549894154071808,
    "learning_rate": 0.0008546623647861369,
    "timestep_mean": 496.6875,
    "epsilon_norm": 8.476725578308105,
    "prediction_norm": 7.532102584838867,
    "timestamp": "2025-07-14T12:33:51.132635"
  },
  {
    "epoch": 498,
    "batch_idx": 96,
    "loss": 0.25639843940734863,
    "learning_rate": 0.0008546623647861369,
    "timestep_mean": 522.4375,
    "epsilon_norm": 8.235068321228027,
    "prediction_norm": 6.745726108551025,
    "timestamp": "2025-07-14T12:33:51.137579"
  },
  {
    "epoch": 498,
    "batch_idx": 128,
    "loss": 0.40242820978164673,
    "learning_rate": 0.0008546623647861369,
    "timestep_mean": 419.625,
    "epsilon_norm": 8.738119125366211,
    "prediction_norm": 6.49898624420166,
    "timestamp": "2025-07-14T12:33:51.142421"
  },
  {
    "epoch": 498,
    "batch_idx": 160,
    "loss": 0.2567780613899231,
    "learning_rate": 0.0008546623647861369,
    "timestep_mean": 534.53125,
    "epsilon_norm": 7.692916393280029,
    "prediction_norm": 6.239454746246338,
    "timestamp": "2025-07-14T12:33:51.147264"
  },
  {
    "epoch": 498,
    "batch_idx": 192,
    "loss": 0.5452595353126526,
    "learning_rate": 0.0008546623647861369,
    "timestep_mean": 506.0,
    "epsilon_norm": 4.758431911468506,
    "prediction_norm": 3.672407865524292,
    "timestamp": "2025-07-14T12:33:51.152099"
  },
  {
    "epoch": 499,
    "batch_idx": 0,
    "loss": 0.24895957112312317,
    "learning_rate": 0.0008541083145532379,
    "timestep_mean": 509.0625,
    "epsilon_norm": 8.470108032226562,
    "prediction_norm": 7.537006855010986,
    "timestamp": "2025-07-14T12:33:51.156990"
  },
  {
    "epoch": 499,
    "batch_idx": 32,
    "loss": 0.33312633633613586,
    "learning_rate": 0.0008541083145532379,
    "timestep_mean": 452.09375,
    "epsilon_norm": 7.495562553405762,
    "prediction_norm": 6.351821422576904,
    "timestamp": "2025-07-14T12:33:51.161912"
  },
  {
    "epoch": 499,
    "batch_idx": 64,
    "loss": 0.3330894112586975,
    "learning_rate": 0.0008541083145532379,
    "timestep_mean": 546.1875,
    "epsilon_norm": 7.931204795837402,
    "prediction_norm": 6.167198181152344,
    "timestamp": "2025-07-14T12:33:51.166906"
  },
  {
    "epoch": 499,
    "batch_idx": 96,
    "loss": 0.29473963379859924,
    "learning_rate": 0.0008541083145532379,
    "timestep_mean": 498.75,
    "epsilon_norm": 7.778460502624512,
    "prediction_norm": 6.3109822273254395,
    "timestamp": "2025-07-14T12:33:51.171804"
  },
  {
    "epoch": 499,
    "batch_idx": 128,
    "loss": 0.348934143781662,
    "learning_rate": 0.0008541083145532379,
    "timestep_mean": 496.5,
    "epsilon_norm": 8.44786548614502,
    "prediction_norm": 7.0790300369262695,
    "timestamp": "2025-07-14T12:33:51.176681"
  },
  {
    "epoch": 499,
    "batch_idx": 160,
    "loss": 0.2632649838924408,
    "learning_rate": 0.0008541083145532379,
    "timestep_mean": 469.75,
    "epsilon_norm": 6.807429313659668,
    "prediction_norm": 5.624873638153076,
    "timestamp": "2025-07-14T12:33:51.181488"
  },
  {
    "epoch": 499,
    "batch_idx": 192,
    "loss": 0.4580058157444,
    "learning_rate": 0.0008541083145532379,
    "timestep_mean": 601.25,
    "epsilon_norm": 4.14289665222168,
    "prediction_norm": 3.7932825088500977,
    "timestamp": "2025-07-14T12:33:51.186350"
  },
  {
    "epoch": 500,
    "batch_idx": 0,
    "loss": 0.3900134861469269,
    "learning_rate": 0.0008535533905932736,
    "timestep_mean": 461.71875,
    "epsilon_norm": 8.548665046691895,
    "prediction_norm": 6.380897521972656,
    "timestamp": "2025-07-14T12:33:51.191217"
  },
  {
    "epoch": 500,
    "batch_idx": 32,
    "loss": 0.4276255965232849,
    "learning_rate": 0.0008535533905932736,
    "timestep_mean": 419.625,
    "epsilon_norm": 7.175928115844727,
    "prediction_norm": 5.721790790557861,
    "timestamp": "2025-07-14T12:33:51.196162"
  },
  {
    "epoch": 500,
    "batch_idx": 64,
    "loss": 0.24857959151268005,
    "learning_rate": 0.0008535533905932736,
    "timestep_mean": 486.625,
    "epsilon_norm": 8.23208999633789,
    "prediction_norm": 7.405496597290039,
    "timestamp": "2025-07-14T12:33:51.200937"
  },
  {
    "epoch": 500,
    "batch_idx": 96,
    "loss": 0.16301599144935608,
    "learning_rate": 0.0008535533905932736,
    "timestep_mean": 525.21875,
    "epsilon_norm": 8.53414249420166,
    "prediction_norm": 8.080041885375977,
    "timestamp": "2025-07-14T12:33:51.205795"
  },
  {
    "epoch": 500,
    "batch_idx": 128,
    "loss": 0.2589738965034485,
    "learning_rate": 0.0008535533905932736,
    "timestep_mean": 527.5625,
    "epsilon_norm": 7.565157413482666,
    "prediction_norm": 6.425820827484131,
    "timestamp": "2025-07-14T12:33:51.210595"
  },
  {
    "epoch": 500,
    "batch_idx": 160,
    "loss": 0.24247880280017853,
    "learning_rate": 0.0008535533905932736,
    "timestep_mean": 546.1875,
    "epsilon_norm": 7.676827907562256,
    "prediction_norm": 6.429293155670166,
    "timestamp": "2025-07-14T12:33:51.215483"
  },
  {
    "epoch": 500,
    "batch_idx": 192,
    "loss": 0.3836631178855896,
    "learning_rate": 0.0008535533905932736,
    "timestep_mean": 428.125,
    "epsilon_norm": 4.90961217880249,
    "prediction_norm": 3.2202868461608887,
    "timestamp": "2025-07-14T12:33:51.220285"
  },
  {
    "epoch": 501,
    "batch_idx": 0,
    "loss": 0.31474393606185913,
    "learning_rate": 0.0008529975942754638,
    "timestep_mean": 453.9375,
    "epsilon_norm": 7.574865341186523,
    "prediction_norm": 6.240561008453369,
    "timestamp": "2025-07-14T12:33:51.225282"
  },
  {
    "epoch": 501,
    "batch_idx": 32,
    "loss": 0.5662617683410645,
    "learning_rate": 0.0008529975942754638,
    "timestep_mean": 554.90625,
    "epsilon_norm": 8.76009464263916,
    "prediction_norm": 6.638611316680908,
    "timestamp": "2025-07-14T12:33:51.230150"
  },
  {
    "epoch": 501,
    "batch_idx": 64,
    "loss": 0.40444105863571167,
    "learning_rate": 0.0008529975942754638,
    "timestep_mean": 426.5625,
    "epsilon_norm": 7.707460403442383,
    "prediction_norm": 5.854126453399658,
    "timestamp": "2025-07-14T12:33:51.235068"
  },
  {
    "epoch": 501,
    "batch_idx": 96,
    "loss": 0.2519383728504181,
    "learning_rate": 0.0008529975942754638,
    "timestep_mean": 528.96875,
    "epsilon_norm": 8.424210548400879,
    "prediction_norm": 7.69224214553833,
    "timestamp": "2025-07-14T12:33:51.239887"
  },
  {
    "epoch": 501,
    "batch_idx": 128,
    "loss": 0.3851169943809509,
    "learning_rate": 0.0008529975942754638,
    "timestep_mean": 464.03125,
    "epsilon_norm": 6.91864538192749,
    "prediction_norm": 5.185594081878662,
    "timestamp": "2025-07-14T12:33:51.244710"
  },
  {
    "epoch": 501,
    "batch_idx": 160,
    "loss": 0.3295100927352905,
    "learning_rate": 0.0008529975942754638,
    "timestep_mean": 512.9375,
    "epsilon_norm": 7.930196285247803,
    "prediction_norm": 5.816701889038086,
    "timestamp": "2025-07-14T12:33:51.249524"
  },
  {
    "epoch": 501,
    "batch_idx": 192,
    "loss": 0.5277502536773682,
    "learning_rate": 0.0008529975942754638,
    "timestep_mean": 466.375,
    "epsilon_norm": 4.91508150100708,
    "prediction_norm": 3.588064193725586,
    "timestamp": "2025-07-14T12:33:51.254353"
  },
  {
    "epoch": 502,
    "batch_idx": 0,
    "loss": 0.6046957969665527,
    "learning_rate": 0.0008524409269711805,
    "timestep_mean": 457.1875,
    "epsilon_norm": 8.238693237304688,
    "prediction_norm": 5.345643520355225,
    "timestamp": "2025-07-14T12:33:51.259181"
  },
  {
    "epoch": 502,
    "batch_idx": 32,
    "loss": 0.339651882648468,
    "learning_rate": 0.0008524409269711805,
    "timestep_mean": 492.59375,
    "epsilon_norm": 7.930695056915283,
    "prediction_norm": 5.69191312789917,
    "timestamp": "2025-07-14T12:33:51.264014"
  },
  {
    "epoch": 502,
    "batch_idx": 64,
    "loss": 0.19634009897708893,
    "learning_rate": 0.0008524409269711805,
    "timestep_mean": 474.0625,
    "epsilon_norm": 7.916443347930908,
    "prediction_norm": 6.827960968017578,
    "timestamp": "2025-07-14T12:33:51.268886"
  },
  {
    "epoch": 502,
    "batch_idx": 96,
    "loss": 0.21211589872837067,
    "learning_rate": 0.0008524409269711805,
    "timestep_mean": 553.8125,
    "epsilon_norm": 7.801105976104736,
    "prediction_norm": 6.4162068367004395,
    "timestamp": "2025-07-14T12:33:51.273777"
  },
  {
    "epoch": 502,
    "batch_idx": 128,
    "loss": 0.44719433784484863,
    "learning_rate": 0.0008524409269711805,
    "timestep_mean": 491.5625,
    "epsilon_norm": 8.662148475646973,
    "prediction_norm": 6.819577217102051,
    "timestamp": "2025-07-14T12:33:51.278664"
  },
  {
    "epoch": 502,
    "batch_idx": 160,
    "loss": 0.26815325021743774,
    "learning_rate": 0.0008524409269711805,
    "timestep_mean": 497.53125,
    "epsilon_norm": 8.109704971313477,
    "prediction_norm": 7.4671854972839355,
    "timestamp": "2025-07-14T12:33:51.283611"
  },
  {
    "epoch": 502,
    "batch_idx": 192,
    "loss": 0.41096484661102295,
    "learning_rate": 0.0008524409269711805,
    "timestep_mean": 378.375,
    "epsilon_norm": 4.050934314727783,
    "prediction_norm": 2.6326208114624023,
    "timestamp": "2025-07-14T12:33:51.288446"
  },
  {
    "epoch": 503,
    "batch_idx": 0,
    "loss": 0.3088376820087433,
    "learning_rate": 0.0008518833900539451,
    "timestep_mean": 562.6875,
    "epsilon_norm": 7.5706400871276855,
    "prediction_norm": 6.012054920196533,
    "timestamp": "2025-07-14T12:33:51.293324"
  },
  {
    "epoch": 503,
    "batch_idx": 32,
    "loss": 0.22625811398029327,
    "learning_rate": 0.0008518833900539451,
    "timestep_mean": 488.78125,
    "epsilon_norm": 7.384448051452637,
    "prediction_norm": 6.690738201141357,
    "timestamp": "2025-07-14T12:33:51.298152"
  },
  {
    "epoch": 503,
    "batch_idx": 64,
    "loss": 0.2140444964170456,
    "learning_rate": 0.0008518833900539451,
    "timestep_mean": 567.59375,
    "epsilon_norm": 7.822290897369385,
    "prediction_norm": 6.894629001617432,
    "timestamp": "2025-07-14T12:33:51.306875"
  },
  {
    "epoch": 503,
    "batch_idx": 96,
    "loss": 0.400147020816803,
    "learning_rate": 0.0008518833900539451,
    "timestep_mean": 445.40625,
    "epsilon_norm": 7.659946441650391,
    "prediction_norm": 5.960699558258057,
    "timestamp": "2025-07-14T12:33:51.311732"
  },
  {
    "epoch": 503,
    "batch_idx": 128,
    "loss": 0.279217392206192,
    "learning_rate": 0.0008518833900539451,
    "timestep_mean": 535.5625,
    "epsilon_norm": 7.990078926086426,
    "prediction_norm": 6.875929832458496,
    "timestamp": "2025-07-14T12:33:51.316658"
  },
  {
    "epoch": 503,
    "batch_idx": 160,
    "loss": 0.4334360957145691,
    "learning_rate": 0.0008518833900539451,
    "timestep_mean": 521.71875,
    "epsilon_norm": 8.400152206420898,
    "prediction_norm": 7.116607666015625,
    "timestamp": "2025-07-14T12:33:51.321526"
  },
  {
    "epoch": 503,
    "batch_idx": 192,
    "loss": 0.8151654005050659,
    "learning_rate": 0.0008518833900539451,
    "timestep_mean": 347.0,
    "epsilon_norm": 3.840911388397217,
    "prediction_norm": 1.2783286571502686,
    "timestamp": "2025-07-14T12:33:51.326353"
  },
  {
    "epoch": 504,
    "batch_idx": 0,
    "loss": 0.18940693140029907,
    "learning_rate": 0.0008513249848994244,
    "timestep_mean": 487.96875,
    "epsilon_norm": 6.448127746582031,
    "prediction_norm": 5.4553303718566895,
    "timestamp": "2025-07-14T12:33:51.331246"
  },
  {
    "epoch": 504,
    "batch_idx": 32,
    "loss": 0.27463194727897644,
    "learning_rate": 0.0008513249848994244,
    "timestep_mean": 432.75,
    "epsilon_norm": 7.273794651031494,
    "prediction_norm": 6.112168788909912,
    "timestamp": "2025-07-14T12:33:51.336124"
  },
  {
    "epoch": 504,
    "batch_idx": 64,
    "loss": 0.27131760120391846,
    "learning_rate": 0.0008513249848994244,
    "timestep_mean": 479.09375,
    "epsilon_norm": 7.645265579223633,
    "prediction_norm": 6.510519504547119,
    "timestamp": "2025-07-14T12:33:51.340992"
  },
  {
    "epoch": 504,
    "batch_idx": 96,
    "loss": 0.19878533482551575,
    "learning_rate": 0.0008513249848994244,
    "timestep_mean": 568.5625,
    "epsilon_norm": 6.947453022003174,
    "prediction_norm": 6.390366554260254,
    "timestamp": "2025-07-14T12:33:51.345918"
  },
  {
    "epoch": 504,
    "batch_idx": 128,
    "loss": 0.36809515953063965,
    "learning_rate": 0.0008513249848994244,
    "timestep_mean": 456.9375,
    "epsilon_norm": 7.938162803649902,
    "prediction_norm": 6.8059492111206055,
    "timestamp": "2025-07-14T12:33:51.350769"
  },
  {
    "epoch": 504,
    "batch_idx": 160,
    "loss": 0.276638925075531,
    "learning_rate": 0.0008513249848994244,
    "timestep_mean": 486.6875,
    "epsilon_norm": 9.383790016174316,
    "prediction_norm": 7.906965255737305,
    "timestamp": "2025-07-14T12:33:51.355669"
  },
  {
    "epoch": 504,
    "batch_idx": 192,
    "loss": 0.11374539136886597,
    "learning_rate": 0.0008513249848994244,
    "timestep_mean": 484.625,
    "epsilon_norm": 3.0705156326293945,
    "prediction_norm": 2.6235947608947754,
    "timestamp": "2025-07-14T12:33:51.360529"
  },
  {
    "epoch": 505,
    "batch_idx": 0,
    "loss": 0.2059962898492813,
    "learning_rate": 0.0008507657128854276,
    "timestep_mean": 607.3125,
    "epsilon_norm": 7.740170478820801,
    "prediction_norm": 6.33306884765625,
    "timestamp": "2025-07-14T12:33:51.365431"
  },
  {
    "epoch": 505,
    "batch_idx": 32,
    "loss": 0.2739560008049011,
    "learning_rate": 0.0008507657128854276,
    "timestep_mean": 438.28125,
    "epsilon_norm": 7.190876007080078,
    "prediction_norm": 5.981081008911133,
    "timestamp": "2025-07-14T12:33:51.370306"
  },
  {
    "epoch": 505,
    "batch_idx": 64,
    "loss": 0.29552456736564636,
    "learning_rate": 0.0008507657128854276,
    "timestep_mean": 611.75,
    "epsilon_norm": 7.361275672912598,
    "prediction_norm": 6.091343402862549,
    "timestamp": "2025-07-14T12:33:51.375275"
  },
  {
    "epoch": 505,
    "batch_idx": 96,
    "loss": 0.2963630259037018,
    "learning_rate": 0.0008507657128854276,
    "timestep_mean": 450.84375,
    "epsilon_norm": 7.036917686462402,
    "prediction_norm": 6.091546058654785,
    "timestamp": "2025-07-14T12:33:51.380105"
  },
  {
    "epoch": 505,
    "batch_idx": 128,
    "loss": 0.30475056171417236,
    "learning_rate": 0.0008507657128854276,
    "timestep_mean": 510.65625,
    "epsilon_norm": 8.393646240234375,
    "prediction_norm": 7.383905410766602,
    "timestamp": "2025-07-14T12:33:51.384934"
  },
  {
    "epoch": 505,
    "batch_idx": 160,
    "loss": 0.15747976303100586,
    "learning_rate": 0.0008507657128854276,
    "timestep_mean": 525.3125,
    "epsilon_norm": 6.688980579376221,
    "prediction_norm": 5.933604717254639,
    "timestamp": "2025-07-14T12:33:51.389874"
  },
  {
    "epoch": 505,
    "batch_idx": 192,
    "loss": 0.14899592101573944,
    "learning_rate": 0.0008507657128854276,
    "timestep_mean": 620.375,
    "epsilon_norm": 3.9461781978607178,
    "prediction_norm": 3.153224229812622,
    "timestamp": "2025-07-14T12:33:51.394777"
  },
  {
    "epoch": 506,
    "batch_idx": 0,
    "loss": 0.18741944432258606,
    "learning_rate": 0.0008502055753919028,
    "timestep_mean": 588.625,
    "epsilon_norm": 7.978530406951904,
    "prediction_norm": 7.381153583526611,
    "timestamp": "2025-07-14T12:33:51.399682"
  },
  {
    "epoch": 506,
    "batch_idx": 32,
    "loss": 0.10896605998277664,
    "learning_rate": 0.0008502055753919028,
    "timestep_mean": 566.03125,
    "epsilon_norm": 7.242958068847656,
    "prediction_norm": 7.098724842071533,
    "timestamp": "2025-07-14T12:33:51.404584"
  },
  {
    "epoch": 506,
    "batch_idx": 64,
    "loss": 0.19730331003665924,
    "learning_rate": 0.0008502055753919028,
    "timestep_mean": 559.21875,
    "epsilon_norm": 8.377022743225098,
    "prediction_norm": 7.803500175476074,
    "timestamp": "2025-07-14T12:33:51.409461"
  },
  {
    "epoch": 506,
    "batch_idx": 96,
    "loss": 0.22660736739635468,
    "learning_rate": 0.0008502055753919028,
    "timestep_mean": 480.71875,
    "epsilon_norm": 8.551958084106445,
    "prediction_norm": 7.279463768005371,
    "timestamp": "2025-07-14T12:33:51.414277"
  },
  {
    "epoch": 506,
    "batch_idx": 128,
    "loss": 0.5367985963821411,
    "learning_rate": 0.0008502055753919028,
    "timestep_mean": 487.96875,
    "epsilon_norm": 9.120203018188477,
    "prediction_norm": 6.785884380340576,
    "timestamp": "2025-07-14T12:33:51.419395"
  },
  {
    "epoch": 506,
    "batch_idx": 160,
    "loss": 0.3725232481956482,
    "learning_rate": 0.0008502055753919028,
    "timestep_mean": 484.34375,
    "epsilon_norm": 8.628433227539062,
    "prediction_norm": 7.3945183753967285,
    "timestamp": "2025-07-14T12:33:51.424312"
  },
  {
    "epoch": 506,
    "batch_idx": 192,
    "loss": 0.11848936975002289,
    "learning_rate": 0.0008502055753919028,
    "timestep_mean": 592.125,
    "epsilon_norm": 4.023580074310303,
    "prediction_norm": 3.821035861968994,
    "timestamp": "2025-07-14T12:33:51.429385"
  },
  {
    "epoch": 507,
    "batch_idx": 0,
    "loss": 0.42233535647392273,
    "learning_rate": 0.0008496445738009338,
    "timestep_mean": 488.875,
    "epsilon_norm": 8.21632194519043,
    "prediction_norm": 6.413162708282471,
    "timestamp": "2025-07-14T12:33:51.434358"
  },
  {
    "epoch": 507,
    "batch_idx": 32,
    "loss": 0.3125745952129364,
    "learning_rate": 0.0008496445738009338,
    "timestep_mean": 534.28125,
    "epsilon_norm": 8.198720932006836,
    "prediction_norm": 6.7076802253723145,
    "timestamp": "2025-07-14T12:33:51.439203"
  },
  {
    "epoch": 507,
    "batch_idx": 64,
    "loss": 0.1849658191204071,
    "learning_rate": 0.0008496445738009338,
    "timestep_mean": 587.71875,
    "epsilon_norm": 7.764237880706787,
    "prediction_norm": 7.37535285949707,
    "timestamp": "2025-07-14T12:33:51.444074"
  },
  {
    "epoch": 507,
    "batch_idx": 96,
    "loss": 0.16499526798725128,
    "learning_rate": 0.0008496445738009338,
    "timestep_mean": 527.90625,
    "epsilon_norm": 8.230863571166992,
    "prediction_norm": 7.611385345458984,
    "timestamp": "2025-07-14T12:33:51.449002"
  },
  {
    "epoch": 507,
    "batch_idx": 128,
    "loss": 0.5927271246910095,
    "learning_rate": 0.0008496445738009338,
    "timestep_mean": 460.03125,
    "epsilon_norm": 8.241199493408203,
    "prediction_norm": 5.686962604522705,
    "timestamp": "2025-07-14T12:33:51.453946"
  },
  {
    "epoch": 507,
    "batch_idx": 160,
    "loss": 0.3905310034751892,
    "learning_rate": 0.0008496445738009338,
    "timestep_mean": 523.75,
    "epsilon_norm": 8.943659782409668,
    "prediction_norm": 7.401174545288086,
    "timestamp": "2025-07-14T12:33:51.458938"
  },
  {
    "epoch": 507,
    "batch_idx": 192,
    "loss": 0.593897819519043,
    "learning_rate": 0.0008496445738009338,
    "timestep_mean": 453.75,
    "epsilon_norm": 4.0497965812683105,
    "prediction_norm": 2.9125754833221436,
    "timestamp": "2025-07-14T12:33:51.463818"
  },
  {
    "epoch": 508,
    "batch_idx": 0,
    "loss": 0.25199222564697266,
    "learning_rate": 0.000849082709496736,
    "timestep_mean": 570.78125,
    "epsilon_norm": 8.025507926940918,
    "prediction_norm": 6.757215976715088,
    "timestamp": "2025-07-14T12:33:51.468725"
  },
  {
    "epoch": 508,
    "batch_idx": 32,
    "loss": 0.28413695096969604,
    "learning_rate": 0.000849082709496736,
    "timestep_mean": 428.6875,
    "epsilon_norm": 7.5499982833862305,
    "prediction_norm": 6.1439900398254395,
    "timestamp": "2025-07-14T12:33:51.473676"
  },
  {
    "epoch": 508,
    "batch_idx": 64,
    "loss": 0.4510495662689209,
    "learning_rate": 0.000849082709496736,
    "timestep_mean": 475.21875,
    "epsilon_norm": 8.914546012878418,
    "prediction_norm": 7.503255844116211,
    "timestamp": "2025-07-14T12:33:51.478546"
  },
  {
    "epoch": 508,
    "batch_idx": 96,
    "loss": 0.3194626569747925,
    "learning_rate": 0.000849082709496736,
    "timestep_mean": 447.375,
    "epsilon_norm": 7.489073276519775,
    "prediction_norm": 5.8996453285217285,
    "timestamp": "2025-07-14T12:33:51.483529"
  },
  {
    "epoch": 508,
    "batch_idx": 128,
    "loss": 0.2823646664619446,
    "learning_rate": 0.000849082709496736,
    "timestep_mean": 425.09375,
    "epsilon_norm": 8.04762077331543,
    "prediction_norm": 6.850398063659668,
    "timestamp": "2025-07-14T12:33:51.488466"
  },
  {
    "epoch": 508,
    "batch_idx": 160,
    "loss": 0.2742045521736145,
    "learning_rate": 0.000849082709496736,
    "timestep_mean": 432.9375,
    "epsilon_norm": 7.976040363311768,
    "prediction_norm": 6.907200813293457,
    "timestamp": "2025-07-14T12:33:51.493368"
  },
  {
    "epoch": 508,
    "batch_idx": 192,
    "loss": 0.1122761219739914,
    "learning_rate": 0.000849082709496736,
    "timestep_mean": 567.5,
    "epsilon_norm": 3.0814015865325928,
    "prediction_norm": 2.6337594985961914,
    "timestamp": "2025-07-14T12:33:51.498161"
  },
  {
    "epoch": 509,
    "batch_idx": 0,
    "loss": 0.30728331208229065,
    "learning_rate": 0.0008485199838656538,
    "timestep_mean": 458.78125,
    "epsilon_norm": 9.041836738586426,
    "prediction_norm": 7.491349220275879,
    "timestamp": "2025-07-14T12:33:51.503084"
  },
  {
    "epoch": 509,
    "batch_idx": 32,
    "loss": 0.3307483196258545,
    "learning_rate": 0.0008485199838656538,
    "timestep_mean": 502.34375,
    "epsilon_norm": 8.085925102233887,
    "prediction_norm": 6.5445780754089355,
    "timestamp": "2025-07-14T12:33:51.507953"
  },
  {
    "epoch": 509,
    "batch_idx": 64,
    "loss": 0.5091614723205566,
    "learning_rate": 0.0008485199838656538,
    "timestep_mean": 403.0,
    "epsilon_norm": 8.275964736938477,
    "prediction_norm": 6.576063632965088,
    "timestamp": "2025-07-14T12:33:51.512977"
  },
  {
    "epoch": 509,
    "batch_idx": 96,
    "loss": 0.2543877363204956,
    "learning_rate": 0.0008485199838656538,
    "timestep_mean": 515.3125,
    "epsilon_norm": 7.994571208953857,
    "prediction_norm": 6.97989559173584,
    "timestamp": "2025-07-14T12:33:51.517826"
  },
  {
    "epoch": 509,
    "batch_idx": 128,
    "loss": 0.30288827419281006,
    "learning_rate": 0.0008485199838656538,
    "timestep_mean": 520.03125,
    "epsilon_norm": 8.389616966247559,
    "prediction_norm": 7.174736022949219,
    "timestamp": "2025-07-14T12:33:51.522702"
  },
  {
    "epoch": 509,
    "batch_idx": 160,
    "loss": 0.2843846082687378,
    "learning_rate": 0.0008485199838656538,
    "timestep_mean": 500.4375,
    "epsilon_norm": 7.650330543518066,
    "prediction_norm": 5.9737114906311035,
    "timestamp": "2025-07-14T12:33:51.527549"
  },
  {
    "epoch": 509,
    "batch_idx": 192,
    "loss": 0.39389413595199585,
    "learning_rate": 0.0008485199838656538,
    "timestep_mean": 618.125,
    "epsilon_norm": 3.9504013061523438,
    "prediction_norm": 2.350316047668457,
    "timestamp": "2025-07-14T12:33:51.532378"
  },
  {
    "epoch": 510,
    "batch_idx": 0,
    "loss": 0.11991126835346222,
    "learning_rate": 0.0008479563982961568,
    "timestep_mean": 556.375,
    "epsilon_norm": 7.661069869995117,
    "prediction_norm": 7.131797790527344,
    "timestamp": "2025-07-14T12:33:51.537229"
  },
  {
    "epoch": 510,
    "batch_idx": 32,
    "loss": 0.37970197200775146,
    "learning_rate": 0.0008479563982961568,
    "timestep_mean": 490.4375,
    "epsilon_norm": 8.984437942504883,
    "prediction_norm": 6.915032386779785,
    "timestamp": "2025-07-14T12:33:51.542105"
  },
  {
    "epoch": 510,
    "batch_idx": 64,
    "loss": 0.3317949175834656,
    "learning_rate": 0.0008479563982961568,
    "timestep_mean": 496.71875,
    "epsilon_norm": 6.8308892250061035,
    "prediction_norm": 5.896763324737549,
    "timestamp": "2025-07-14T12:33:51.546976"
  },
  {
    "epoch": 510,
    "batch_idx": 96,
    "loss": 0.27952542901039124,
    "learning_rate": 0.0008479563982961568,
    "timestep_mean": 495.46875,
    "epsilon_norm": 6.685575485229492,
    "prediction_norm": 5.3338823318481445,
    "timestamp": "2025-07-14T12:33:51.551814"
  },
  {
    "epoch": 510,
    "batch_idx": 128,
    "loss": 0.24979469180107117,
    "learning_rate": 0.0008479563982961568,
    "timestep_mean": 510.875,
    "epsilon_norm": 7.5701470375061035,
    "prediction_norm": 6.68892765045166,
    "timestamp": "2025-07-14T12:33:51.556694"
  },
  {
    "epoch": 510,
    "batch_idx": 160,
    "loss": 0.245992049574852,
    "learning_rate": 0.0008479563982961568,
    "timestep_mean": 576.46875,
    "epsilon_norm": 6.9827399253845215,
    "prediction_norm": 5.982144355773926,
    "timestamp": "2025-07-14T12:33:51.561610"
  },
  {
    "epoch": 510,
    "batch_idx": 192,
    "loss": 0.24070420861244202,
    "learning_rate": 0.0008479563982961568,
    "timestep_mean": 485.75,
    "epsilon_norm": 4.040390968322754,
    "prediction_norm": 3.465284824371338,
    "timestamp": "2025-07-14T12:33:51.566537"
  },
  {
    "epoch": 511,
    "batch_idx": 0,
    "loss": 0.6345882415771484,
    "learning_rate": 0.0008473919541788364,
    "timestep_mean": 425.59375,
    "epsilon_norm": 8.483851432800293,
    "prediction_norm": 5.31809663772583,
    "timestamp": "2025-07-14T12:33:51.571382"
  },
  {
    "epoch": 511,
    "batch_idx": 32,
    "loss": 0.48962870240211487,
    "learning_rate": 0.0008473919541788364,
    "timestep_mean": 530.9375,
    "epsilon_norm": 9.69830322265625,
    "prediction_norm": 8.263620376586914,
    "timestamp": "2025-07-14T12:33:51.576209"
  },
  {
    "epoch": 511,
    "batch_idx": 64,
    "loss": 0.26294153928756714,
    "learning_rate": 0.0008473919541788364,
    "timestep_mean": 379.6875,
    "epsilon_norm": 7.004594326019287,
    "prediction_norm": 6.4584245681762695,
    "timestamp": "2025-07-14T12:33:51.581024"
  },
  {
    "epoch": 511,
    "batch_idx": 96,
    "loss": 0.3326282501220703,
    "learning_rate": 0.0008473919541788364,
    "timestep_mean": 444.5,
    "epsilon_norm": 7.586187362670898,
    "prediction_norm": 6.028090953826904,
    "timestamp": "2025-07-14T12:33:51.585894"
  },
  {
    "epoch": 511,
    "batch_idx": 128,
    "loss": 0.24711757898330688,
    "learning_rate": 0.0008473919541788364,
    "timestep_mean": 482.34375,
    "epsilon_norm": 8.115449905395508,
    "prediction_norm": 7.2168450355529785,
    "timestamp": "2025-07-14T12:33:51.591009"
  },
  {
    "epoch": 511,
    "batch_idx": 160,
    "loss": 0.3105599582195282,
    "learning_rate": 0.0008473919541788364,
    "timestep_mean": 501.53125,
    "epsilon_norm": 8.755049705505371,
    "prediction_norm": 7.300869464874268,
    "timestamp": "2025-07-14T12:33:51.596100"
  },
  {
    "epoch": 511,
    "batch_idx": 192,
    "loss": 0.9127191305160522,
    "learning_rate": 0.0008473919541788364,
    "timestep_mean": 502.875,
    "epsilon_norm": 4.518418312072754,
    "prediction_norm": 3.11002254486084,
    "timestamp": "2025-07-14T12:33:51.601137"
  },
  {
    "epoch": 512,
    "batch_idx": 0,
    "loss": 0.3062838912010193,
    "learning_rate": 0.0008468266529064022,
    "timestep_mean": 519.625,
    "epsilon_norm": 8.28431510925293,
    "prediction_norm": 6.509503364562988,
    "timestamp": "2025-07-14T12:33:51.606367"
  },
  {
    "epoch": 512,
    "batch_idx": 32,
    "loss": 0.21067452430725098,
    "learning_rate": 0.0008468266529064022,
    "timestep_mean": 596.15625,
    "epsilon_norm": 8.759186744689941,
    "prediction_norm": 8.001708030700684,
    "timestamp": "2025-07-14T12:33:51.611633"
  },
  {
    "epoch": 512,
    "batch_idx": 64,
    "loss": 0.40469890832901,
    "learning_rate": 0.0008468266529064022,
    "timestep_mean": 446.5,
    "epsilon_norm": 8.996100425720215,
    "prediction_norm": 8.017934799194336,
    "timestamp": "2025-07-14T12:33:51.616753"
  },
  {
    "epoch": 512,
    "batch_idx": 96,
    "loss": 0.3395262658596039,
    "learning_rate": 0.0008468266529064022,
    "timestep_mean": 405.4375,
    "epsilon_norm": 9.127949714660645,
    "prediction_norm": 7.1127190589904785,
    "timestamp": "2025-07-14T12:33:51.622087"
  },
  {
    "epoch": 512,
    "batch_idx": 128,
    "loss": 0.25231462717056274,
    "learning_rate": 0.0008468266529064022,
    "timestep_mean": 543.0,
    "epsilon_norm": 8.425091743469238,
    "prediction_norm": 7.529365062713623,
    "timestamp": "2025-07-14T12:33:51.627339"
  },
  {
    "epoch": 512,
    "batch_idx": 160,
    "loss": 0.32165879011154175,
    "learning_rate": 0.0008468266529064022,
    "timestep_mean": 594.59375,
    "epsilon_norm": 8.040216445922852,
    "prediction_norm": 6.754864692687988,
    "timestamp": "2025-07-14T12:33:51.632601"
  },
  {
    "epoch": 512,
    "batch_idx": 192,
    "loss": 0.6381323337554932,
    "learning_rate": 0.0008468266529064022,
    "timestep_mean": 210.625,
    "epsilon_norm": 4.496805191040039,
    "prediction_norm": 2.3618650436401367,
    "timestamp": "2025-07-14T12:33:51.637927"
  },
  {
    "epoch": 513,
    "batch_idx": 0,
    "loss": 0.2952430546283722,
    "learning_rate": 0.000846260495873679,
    "timestep_mean": 488.125,
    "epsilon_norm": 6.9902849197387695,
    "prediction_norm": 6.042089462280273,
    "timestamp": "2025-07-14T12:33:51.643168"
  },
  {
    "epoch": 513,
    "batch_idx": 32,
    "loss": 0.35855820775032043,
    "learning_rate": 0.000846260495873679,
    "timestep_mean": 474.53125,
    "epsilon_norm": 8.8707857131958,
    "prediction_norm": 7.404160976409912,
    "timestamp": "2025-07-14T12:33:51.648421"
  },
  {
    "epoch": 513,
    "batch_idx": 64,
    "loss": 0.3361247777938843,
    "learning_rate": 0.000846260495873679,
    "timestep_mean": 467.71875,
    "epsilon_norm": 8.413331031799316,
    "prediction_norm": 6.803485870361328,
    "timestamp": "2025-07-14T12:33:51.653654"
  },
  {
    "epoch": 513,
    "batch_idx": 96,
    "loss": 0.34116923809051514,
    "learning_rate": 0.000846260495873679,
    "timestep_mean": 419.875,
    "epsilon_norm": 7.739381790161133,
    "prediction_norm": 5.92331600189209,
    "timestamp": "2025-07-14T12:33:51.658900"
  },
  {
    "epoch": 513,
    "batch_idx": 128,
    "loss": 0.37499165534973145,
    "learning_rate": 0.000846260495873679,
    "timestep_mean": 425.625,
    "epsilon_norm": 8.323809623718262,
    "prediction_norm": 6.428933143615723,
    "timestamp": "2025-07-14T12:33:51.664036"
  },
  {
    "epoch": 513,
    "batch_idx": 160,
    "loss": 0.37910014390945435,
    "learning_rate": 0.000846260495873679,
    "timestep_mean": 433.34375,
    "epsilon_norm": 7.793877124786377,
    "prediction_norm": 5.760988235473633,
    "timestamp": "2025-07-14T12:33:51.669286"
  },
  {
    "epoch": 513,
    "batch_idx": 192,
    "loss": 0.17028215527534485,
    "learning_rate": 0.000846260495873679,
    "timestep_mean": 533.0,
    "epsilon_norm": 3.649414539337158,
    "prediction_norm": 3.1344645023345947,
    "timestamp": "2025-07-14T12:33:51.674466"
  },
  {
    "epoch": 514,
    "batch_idx": 0,
    "loss": 0.26602938771247864,
    "learning_rate": 0.0008456934844776029,
    "timestep_mean": 440.34375,
    "epsilon_norm": 7.318490982055664,
    "prediction_norm": 5.762434005737305,
    "timestamp": "2025-07-14T12:33:51.679756"
  },
  {
    "epoch": 514,
    "batch_idx": 32,
    "loss": 0.27715927362442017,
    "learning_rate": 0.0008456934844776029,
    "timestep_mean": 491.6875,
    "epsilon_norm": 7.2107133865356445,
    "prediction_norm": 6.1101603507995605,
    "timestamp": "2025-07-14T12:33:51.684995"
  },
  {
    "epoch": 514,
    "batch_idx": 64,
    "loss": 0.28306829929351807,
    "learning_rate": 0.0008456934844776029,
    "timestep_mean": 580.40625,
    "epsilon_norm": 8.249229431152344,
    "prediction_norm": 6.8387770652771,
    "timestamp": "2025-07-14T12:33:51.690098"
  },
  {
    "epoch": 514,
    "batch_idx": 96,
    "loss": 0.2557263970375061,
    "learning_rate": 0.0008456934844776029,
    "timestep_mean": 537.71875,
    "epsilon_norm": 7.352387428283691,
    "prediction_norm": 6.023082256317139,
    "timestamp": "2025-07-14T12:33:51.695362"
  },
  {
    "epoch": 514,
    "batch_idx": 128,
    "loss": 0.28976869583129883,
    "learning_rate": 0.0008456934844776029,
    "timestep_mean": 513.03125,
    "epsilon_norm": 8.62508773803711,
    "prediction_norm": 6.889240264892578,
    "timestamp": "2025-07-14T12:33:51.700758"
  },
  {
    "epoch": 514,
    "batch_idx": 160,
    "loss": 0.26219114661216736,
    "learning_rate": 0.0008456934844776029,
    "timestep_mean": 555.6875,
    "epsilon_norm": 7.737520217895508,
    "prediction_norm": 6.580883502960205,
    "timestamp": "2025-07-14T12:33:51.705999"
  },
  {
    "epoch": 514,
    "batch_idx": 192,
    "loss": 0.2822685241699219,
    "learning_rate": 0.0008456934844776029,
    "timestep_mean": 482.75,
    "epsilon_norm": 4.743034362792969,
    "prediction_norm": 4.204122066497803,
    "timestamp": "2025-07-14T12:33:51.710842"
  },
  {
    "epoch": 515,
    "batch_idx": 0,
    "loss": 0.21527448296546936,
    "learning_rate": 0.0008451256201172184,
    "timestep_mean": 563.875,
    "epsilon_norm": 7.0042405128479,
    "prediction_norm": 5.994886875152588,
    "timestamp": "2025-07-14T12:33:51.716041"
  },
  {
    "epoch": 515,
    "batch_idx": 32,
    "loss": 0.46762028336524963,
    "learning_rate": 0.0008451256201172184,
    "timestep_mean": 446.28125,
    "epsilon_norm": 8.085871696472168,
    "prediction_norm": 7.181272029876709,
    "timestamp": "2025-07-14T12:33:51.721367"
  },
  {
    "epoch": 515,
    "batch_idx": 64,
    "loss": 0.19004294276237488,
    "learning_rate": 0.0008451256201172184,
    "timestep_mean": 499.375,
    "epsilon_norm": 6.9679856300354,
    "prediction_norm": 6.7961955070495605,
    "timestamp": "2025-07-14T12:33:51.726620"
  },
  {
    "epoch": 515,
    "batch_idx": 96,
    "loss": 0.33462050557136536,
    "learning_rate": 0.0008451256201172184,
    "timestep_mean": 407.9375,
    "epsilon_norm": 7.560042381286621,
    "prediction_norm": 6.318342685699463,
    "timestamp": "2025-07-14T12:33:51.731840"
  },
  {
    "epoch": 515,
    "batch_idx": 128,
    "loss": 0.15796762704849243,
    "learning_rate": 0.0008451256201172184,
    "timestep_mean": 511.34375,
    "epsilon_norm": 6.954206943511963,
    "prediction_norm": 6.642051696777344,
    "timestamp": "2025-07-14T12:33:51.737002"
  },
  {
    "epoch": 515,
    "batch_idx": 160,
    "loss": 0.12817391753196716,
    "learning_rate": 0.0008451256201172184,
    "timestep_mean": 568.90625,
    "epsilon_norm": 7.304054260253906,
    "prediction_norm": 6.678903102874756,
    "timestamp": "2025-07-14T12:33:51.742459"
  },
  {
    "epoch": 515,
    "batch_idx": 192,
    "loss": 0.3138258457183838,
    "learning_rate": 0.0008451256201172184,
    "timestep_mean": 474.375,
    "epsilon_norm": 4.1647233963012695,
    "prediction_norm": 3.097757577896118,
    "timestamp": "2025-07-14T12:33:51.747570"
  },
  {
    "epoch": 516,
    "batch_idx": 0,
    "loss": 0.36747461557388306,
    "learning_rate": 0.000844556904193674,
    "timestep_mean": 438.53125,
    "epsilon_norm": 7.773983001708984,
    "prediction_norm": 6.181323051452637,
    "timestamp": "2025-07-14T12:33:51.752789"
  },
  {
    "epoch": 516,
    "batch_idx": 32,
    "loss": 0.3471260368824005,
    "learning_rate": 0.000844556904193674,
    "timestep_mean": 476.71875,
    "epsilon_norm": 8.109976768493652,
    "prediction_norm": 7.106834411621094,
    "timestamp": "2025-07-14T12:33:51.758153"
  },
  {
    "epoch": 516,
    "batch_idx": 64,
    "loss": 0.19782567024230957,
    "learning_rate": 0.000844556904193674,
    "timestep_mean": 420.6875,
    "epsilon_norm": 7.833287239074707,
    "prediction_norm": 6.531392574310303,
    "timestamp": "2025-07-14T12:33:51.763372"
  },
  {
    "epoch": 516,
    "batch_idx": 96,
    "loss": 0.582112729549408,
    "learning_rate": 0.000844556904193674,
    "timestep_mean": 525.625,
    "epsilon_norm": 9.321745872497559,
    "prediction_norm": 6.890419006347656,
    "timestamp": "2025-07-14T12:33:51.768541"
  },
  {
    "epoch": 516,
    "batch_idx": 128,
    "loss": 0.2982369065284729,
    "learning_rate": 0.000844556904193674,
    "timestep_mean": 473.59375,
    "epsilon_norm": 8.397254943847656,
    "prediction_norm": 6.611259937286377,
    "timestamp": "2025-07-14T12:33:51.773769"
  },
  {
    "epoch": 516,
    "batch_idx": 160,
    "loss": 0.24898147583007812,
    "learning_rate": 0.000844556904193674,
    "timestep_mean": 543.96875,
    "epsilon_norm": 8.820730209350586,
    "prediction_norm": 7.5651350021362305,
    "timestamp": "2025-07-14T12:33:51.778957"
  },
  {
    "epoch": 516,
    "batch_idx": 192,
    "loss": 0.03492351621389389,
    "learning_rate": 0.000844556904193674,
    "timestep_mean": 608.0,
    "epsilon_norm": 4.057490348815918,
    "prediction_norm": 4.046261787414551,
    "timestamp": "2025-07-14T12:33:51.784050"
  },
  {
    "epoch": 517,
    "batch_idx": 0,
    "loss": 0.5377817153930664,
    "learning_rate": 0.00084398733811022,
    "timestep_mean": 411.71875,
    "epsilon_norm": 7.8996806144714355,
    "prediction_norm": 5.509714603424072,
    "timestamp": "2025-07-14T12:33:51.789284"
  },
  {
    "epoch": 517,
    "batch_idx": 32,
    "loss": 0.31887632608413696,
    "learning_rate": 0.00084398733811022,
    "timestep_mean": 517.875,
    "epsilon_norm": 8.293811798095703,
    "prediction_norm": 6.838642597198486,
    "timestamp": "2025-07-14T12:33:51.794481"
  },
  {
    "epoch": 517,
    "batch_idx": 64,
    "loss": 0.21882691979408264,
    "learning_rate": 0.00084398733811022,
    "timestep_mean": 582.5,
    "epsilon_norm": 8.777810096740723,
    "prediction_norm": 7.962713718414307,
    "timestamp": "2025-07-14T12:33:51.799882"
  },
  {
    "epoch": 517,
    "batch_idx": 96,
    "loss": 0.21927720308303833,
    "learning_rate": 0.00084398733811022,
    "timestep_mean": 592.78125,
    "epsilon_norm": 7.7997965812683105,
    "prediction_norm": 6.519583702087402,
    "timestamp": "2025-07-14T12:33:51.805110"
  },
  {
    "epoch": 517,
    "batch_idx": 128,
    "loss": 0.5313162207603455,
    "learning_rate": 0.00084398733811022,
    "timestep_mean": 485.0625,
    "epsilon_norm": 9.952766418457031,
    "prediction_norm": 7.88631010055542,
    "timestamp": "2025-07-14T12:33:51.810198"
  },
  {
    "epoch": 517,
    "batch_idx": 160,
    "loss": 0.1847309172153473,
    "learning_rate": 0.00084398733811022,
    "timestep_mean": 453.15625,
    "epsilon_norm": 7.0699944496154785,
    "prediction_norm": 7.232804298400879,
    "timestamp": "2025-07-14T12:33:51.815382"
  },
  {
    "epoch": 517,
    "batch_idx": 192,
    "loss": 0.23010332882404327,
    "learning_rate": 0.00084398733811022,
    "timestep_mean": 497.0,
    "epsilon_norm": 3.1927552223205566,
    "prediction_norm": 3.519185781478882,
    "timestamp": "2025-07-14T12:33:51.820572"
  },
  {
    "epoch": 518,
    "batch_idx": 0,
    "loss": 0.31414562463760376,
    "learning_rate": 0.000843416923272204,
    "timestep_mean": 500.28125,
    "epsilon_norm": 7.573074817657471,
    "prediction_norm": 6.299376010894775,
    "timestamp": "2025-07-14T12:33:51.825722"
  },
  {
    "epoch": 518,
    "batch_idx": 32,
    "loss": 0.23180431127548218,
    "learning_rate": 0.000843416923272204,
    "timestep_mean": 495.03125,
    "epsilon_norm": 6.933512210845947,
    "prediction_norm": 6.255309581756592,
    "timestamp": "2025-07-14T12:33:51.830889"
  },
  {
    "epoch": 518,
    "batch_idx": 64,
    "loss": 0.32270509004592896,
    "learning_rate": 0.000843416923272204,
    "timestep_mean": 446.0,
    "epsilon_norm": 8.361351013183594,
    "prediction_norm": 6.406042098999023,
    "timestamp": "2025-07-14T12:33:51.835941"
  },
  {
    "epoch": 518,
    "batch_idx": 96,
    "loss": 0.3003435730934143,
    "learning_rate": 0.000843416923272204,
    "timestep_mean": 540.0625,
    "epsilon_norm": 7.431343078613281,
    "prediction_norm": 6.457558631896973,
    "timestamp": "2025-07-14T12:33:51.841053"
  },
  {
    "epoch": 518,
    "batch_idx": 128,
    "loss": 0.29694831371307373,
    "learning_rate": 0.000843416923272204,
    "timestep_mean": 456.59375,
    "epsilon_norm": 9.063858032226562,
    "prediction_norm": 8.042713165283203,
    "timestamp": "2025-07-14T12:33:51.846389"
  },
  {
    "epoch": 518,
    "batch_idx": 160,
    "loss": 0.29758089780807495,
    "learning_rate": 0.000843416923272204,
    "timestep_mean": 435.78125,
    "epsilon_norm": 7.99298620223999,
    "prediction_norm": 6.471994876861572,
    "timestamp": "2025-07-14T12:33:51.851533"
  },
  {
    "epoch": 518,
    "batch_idx": 192,
    "loss": 0.2667216658592224,
    "learning_rate": 0.000843416923272204,
    "timestep_mean": 470.0,
    "epsilon_norm": 4.017093181610107,
    "prediction_norm": 3.2101621627807617,
    "timestamp": "2025-07-14T12:33:51.856664"
  },
  {
    "epoch": 519,
    "batch_idx": 0,
    "loss": 0.37282323837280273,
    "learning_rate": 0.0008428456610870677,
    "timestep_mean": 455.28125,
    "epsilon_norm": 7.536995887756348,
    "prediction_norm": 5.631012439727783,
    "timestamp": "2025-07-14T12:33:51.861878"
  },
  {
    "epoch": 519,
    "batch_idx": 32,
    "loss": 0.22875113785266876,
    "learning_rate": 0.0008428456610870677,
    "timestep_mean": 550.625,
    "epsilon_norm": 7.589481830596924,
    "prediction_norm": 6.551477909088135,
    "timestamp": "2025-07-14T12:33:51.867170"
  },
  {
    "epoch": 519,
    "batch_idx": 64,
    "loss": 0.48035258054733276,
    "learning_rate": 0.0008428456610870677,
    "timestep_mean": 464.65625,
    "epsilon_norm": 9.136331558227539,
    "prediction_norm": 6.558859348297119,
    "timestamp": "2025-07-14T12:33:51.872343"
  },
  {
    "epoch": 519,
    "batch_idx": 96,
    "loss": 0.2077617645263672,
    "learning_rate": 0.0008428456610870677,
    "timestep_mean": 498.75,
    "epsilon_norm": 8.172748565673828,
    "prediction_norm": 6.64646053314209,
    "timestamp": "2025-07-14T12:33:51.877519"
  },
  {
    "epoch": 519,
    "batch_idx": 128,
    "loss": 0.17493775486946106,
    "learning_rate": 0.0008428456610870677,
    "timestep_mean": 435.3125,
    "epsilon_norm": 8.402372360229492,
    "prediction_norm": 7.582658767700195,
    "timestamp": "2025-07-14T12:33:51.882712"
  },
  {
    "epoch": 519,
    "batch_idx": 160,
    "loss": 0.16255402565002441,
    "learning_rate": 0.0008428456610870677,
    "timestep_mean": 580.5625,
    "epsilon_norm": 8.225317001342773,
    "prediction_norm": 7.127468109130859,
    "timestamp": "2025-07-14T12:33:51.887977"
  },
  {
    "epoch": 519,
    "batch_idx": 192,
    "loss": 0.2512328028678894,
    "learning_rate": 0.0008428456610870677,
    "timestep_mean": 342.875,
    "epsilon_norm": 3.378528356552124,
    "prediction_norm": 2.234653949737549,
    "timestamp": "2025-07-14T12:33:51.893109"
  },
  {
    "epoch": 520,
    "batch_idx": 0,
    "loss": 0.35921135544776917,
    "learning_rate": 0.0008422735529643442,
    "timestep_mean": 475.15625,
    "epsilon_norm": 7.440422534942627,
    "prediction_norm": 5.403580665588379,
    "timestamp": "2025-07-14T12:33:51.898392"
  },
  {
    "epoch": 520,
    "batch_idx": 32,
    "loss": 0.21611148118972778,
    "learning_rate": 0.0008422735529643442,
    "timestep_mean": 507.4375,
    "epsilon_norm": 6.970862865447998,
    "prediction_norm": 5.813869953155518,
    "timestamp": "2025-07-14T12:33:51.908946"
  },
  {
    "epoch": 520,
    "batch_idx": 64,
    "loss": 0.2795858681201935,
    "learning_rate": 0.0008422735529643442,
    "timestep_mean": 517.34375,
    "epsilon_norm": 9.461140632629395,
    "prediction_norm": 8.175951957702637,
    "timestamp": "2025-07-14T12:33:51.914068"
  },
  {
    "epoch": 520,
    "batch_idx": 96,
    "loss": 0.1715523898601532,
    "learning_rate": 0.0008422735529643442,
    "timestep_mean": 595.3125,
    "epsilon_norm": 8.403545379638672,
    "prediction_norm": 7.977405071258545,
    "timestamp": "2025-07-14T12:33:51.919107"
  },
  {
    "epoch": 520,
    "batch_idx": 128,
    "loss": 0.27701589465141296,
    "learning_rate": 0.0008422735529643442,
    "timestep_mean": 479.5625,
    "epsilon_norm": 7.950539588928223,
    "prediction_norm": 7.487881183624268,
    "timestamp": "2025-07-14T12:33:51.924058"
  },
  {
    "epoch": 520,
    "batch_idx": 160,
    "loss": 0.2875164747238159,
    "learning_rate": 0.0008422735529643442,
    "timestep_mean": 400.3125,
    "epsilon_norm": 7.675097465515137,
    "prediction_norm": 6.347157001495361,
    "timestamp": "2025-07-14T12:33:51.929096"
  },
  {
    "epoch": 520,
    "batch_idx": 192,
    "loss": 0.5433855652809143,
    "learning_rate": 0.0008422735529643442,
    "timestep_mean": 482.25,
    "epsilon_norm": 4.455175399780273,
    "prediction_norm": 3.7500085830688477,
    "timestamp": "2025-07-14T12:33:51.934090"
  },
  {
    "epoch": 521,
    "batch_idx": 0,
    "loss": 0.31769585609436035,
    "learning_rate": 0.000841700600315653,
    "timestep_mean": 483.46875,
    "epsilon_norm": 7.636806964874268,
    "prediction_norm": 6.533432960510254,
    "timestamp": "2025-07-14T12:33:51.939120"
  },
  {
    "epoch": 521,
    "batch_idx": 32,
    "loss": 0.3302791118621826,
    "learning_rate": 0.000841700600315653,
    "timestep_mean": 420.0625,
    "epsilon_norm": 7.171535491943359,
    "prediction_norm": 5.7764410972595215,
    "timestamp": "2025-07-14T12:33:51.944117"
  },
  {
    "epoch": 521,
    "batch_idx": 64,
    "loss": 0.2579019069671631,
    "learning_rate": 0.000841700600315653,
    "timestep_mean": 580.28125,
    "epsilon_norm": 7.273426055908203,
    "prediction_norm": 6.706763744354248,
    "timestamp": "2025-07-14T12:33:51.949071"
  },
  {
    "epoch": 521,
    "batch_idx": 96,
    "loss": 0.3206450939178467,
    "learning_rate": 0.000841700600315653,
    "timestep_mean": 559.5,
    "epsilon_norm": 7.019341468811035,
    "prediction_norm": 6.53986120223999,
    "timestamp": "2025-07-14T12:33:51.954061"
  },
  {
    "epoch": 521,
    "batch_idx": 128,
    "loss": 0.3188028633594513,
    "learning_rate": 0.000841700600315653,
    "timestep_mean": 537.34375,
    "epsilon_norm": 8.489995002746582,
    "prediction_norm": 7.181880950927734,
    "timestamp": "2025-07-14T12:33:51.959027"
  },
  {
    "epoch": 521,
    "batch_idx": 160,
    "loss": 0.3428574204444885,
    "learning_rate": 0.000841700600315653,
    "timestep_mean": 492.96875,
    "epsilon_norm": 8.613565444946289,
    "prediction_norm": 6.751492500305176,
    "timestamp": "2025-07-14T12:33:51.963955"
  },
  {
    "epoch": 521,
    "batch_idx": 192,
    "loss": 0.349653422832489,
    "learning_rate": 0.000841700600315653,
    "timestep_mean": 496.375,
    "epsilon_norm": 3.08937406539917,
    "prediction_norm": 1.8827731609344482,
    "timestamp": "2025-07-14T12:33:51.968917"
  },
  {
    "epoch": 522,
    "batch_idx": 0,
    "loss": 0.2958662211894989,
    "learning_rate": 0.000841126804554698,
    "timestep_mean": 539.21875,
    "epsilon_norm": 9.320459365844727,
    "prediction_norm": 8.183819770812988,
    "timestamp": "2025-07-14T12:33:51.973937"
  },
  {
    "epoch": 522,
    "batch_idx": 32,
    "loss": 0.20768707990646362,
    "learning_rate": 0.000841126804554698,
    "timestep_mean": 547.96875,
    "epsilon_norm": 10.407564163208008,
    "prediction_norm": 10.033430099487305,
    "timestamp": "2025-07-14T12:33:51.978958"
  },
  {
    "epoch": 522,
    "batch_idx": 64,
    "loss": 0.25171297788619995,
    "learning_rate": 0.000841126804554698,
    "timestep_mean": 442.875,
    "epsilon_norm": 6.3470964431762695,
    "prediction_norm": 5.896969318389893,
    "timestamp": "2025-07-14T12:33:51.983970"
  },
  {
    "epoch": 522,
    "batch_idx": 96,
    "loss": 0.19371701776981354,
    "learning_rate": 0.000841126804554698,
    "timestep_mean": 479.0,
    "epsilon_norm": 8.317272186279297,
    "prediction_norm": 7.250835418701172,
    "timestamp": "2025-07-14T12:33:51.989002"
  },
  {
    "epoch": 522,
    "batch_idx": 128,
    "loss": 0.1263296902179718,
    "learning_rate": 0.000841126804554698,
    "timestep_mean": 534.78125,
    "epsilon_norm": 6.954568386077881,
    "prediction_norm": 6.086551189422607,
    "timestamp": "2025-07-14T12:33:51.994099"
  },
  {
    "epoch": 522,
    "batch_idx": 160,
    "loss": 0.3289009630680084,
    "learning_rate": 0.000841126804554698,
    "timestep_mean": 431.9375,
    "epsilon_norm": 7.656630039215088,
    "prediction_norm": 6.240488529205322,
    "timestamp": "2025-07-14T12:33:51.999084"
  },
  {
    "epoch": 522,
    "batch_idx": 192,
    "loss": 0.16702237725257874,
    "learning_rate": 0.000841126804554698,
    "timestep_mean": 644.0,
    "epsilon_norm": 4.246315002441406,
    "prediction_norm": 4.136777400970459,
    "timestamp": "2025-07-14T12:33:52.004085"
  },
  {
    "epoch": 523,
    "batch_idx": 0,
    "loss": 0.21966753900051117,
    "learning_rate": 0.0008405521670972633,
    "timestep_mean": 524.125,
    "epsilon_norm": 7.745222568511963,
    "prediction_norm": 6.32288932800293,
    "timestamp": "2025-07-14T12:33:52.009158"
  },
  {
    "epoch": 523,
    "batch_idx": 32,
    "loss": 0.2663382589817047,
    "learning_rate": 0.0008405521670972633,
    "timestep_mean": 463.03125,
    "epsilon_norm": 6.630814552307129,
    "prediction_norm": 5.732498645782471,
    "timestamp": "2025-07-14T12:33:52.014180"
  },
  {
    "epoch": 523,
    "batch_idx": 64,
    "loss": 0.26065361499786377,
    "learning_rate": 0.0008405521670972633,
    "timestep_mean": 476.59375,
    "epsilon_norm": 8.240202903747559,
    "prediction_norm": 6.797255516052246,
    "timestamp": "2025-07-14T12:33:52.019208"
  },
  {
    "epoch": 523,
    "batch_idx": 96,
    "loss": 0.3396299481391907,
    "learning_rate": 0.0008405521670972633,
    "timestep_mean": 480.84375,
    "epsilon_norm": 7.424260139465332,
    "prediction_norm": 6.992956161499023,
    "timestamp": "2025-07-14T12:33:52.024233"
  },
  {
    "epoch": 523,
    "batch_idx": 128,
    "loss": 0.1633985936641693,
    "learning_rate": 0.0008405521670972633,
    "timestep_mean": 482.0,
    "epsilon_norm": 7.7051849365234375,
    "prediction_norm": 7.037768363952637,
    "timestamp": "2025-07-14T12:33:52.029405"
  },
  {
    "epoch": 523,
    "batch_idx": 160,
    "loss": 0.2353135645389557,
    "learning_rate": 0.0008405521670972633,
    "timestep_mean": 465.46875,
    "epsilon_norm": 8.613481521606445,
    "prediction_norm": 7.544132709503174,
    "timestamp": "2025-07-14T12:33:52.034472"
  },
  {
    "epoch": 523,
    "batch_idx": 192,
    "loss": 0.2820887565612793,
    "learning_rate": 0.0008405521670972633,
    "timestep_mean": 455.5,
    "epsilon_norm": 3.420306444168091,
    "prediction_norm": 3.3533527851104736,
    "timestamp": "2025-07-14T12:33:52.039502"
  },
  {
    "epoch": 524,
    "batch_idx": 0,
    "loss": 0.32451802492141724,
    "learning_rate": 0.0008399766893612094,
    "timestep_mean": 502.59375,
    "epsilon_norm": 9.280303001403809,
    "prediction_norm": 8.333014488220215,
    "timestamp": "2025-07-14T12:33:52.044613"
  },
  {
    "epoch": 524,
    "batch_idx": 32,
    "loss": 0.25636160373687744,
    "learning_rate": 0.0008399766893612094,
    "timestep_mean": 500.0,
    "epsilon_norm": 8.496230125427246,
    "prediction_norm": 7.711301803588867,
    "timestamp": "2025-07-14T12:33:52.049661"
  },
  {
    "epoch": 524,
    "batch_idx": 64,
    "loss": 0.23063164949417114,
    "learning_rate": 0.0008399766893612094,
    "timestep_mean": 423.03125,
    "epsilon_norm": 6.2903876304626465,
    "prediction_norm": 5.226603031158447,
    "timestamp": "2025-07-14T12:33:52.055036"
  },
  {
    "epoch": 524,
    "batch_idx": 96,
    "loss": 0.19076058268547058,
    "learning_rate": 0.0008399766893612094,
    "timestep_mean": 532.1875,
    "epsilon_norm": 7.174780368804932,
    "prediction_norm": 5.885661602020264,
    "timestamp": "2025-07-14T12:33:52.060165"
  },
  {
    "epoch": 524,
    "batch_idx": 128,
    "loss": 0.24225442111492157,
    "learning_rate": 0.0008399766893612094,
    "timestep_mean": 408.4375,
    "epsilon_norm": 7.057789325714111,
    "prediction_norm": 5.463741302490234,
    "timestamp": "2025-07-14T12:33:52.065333"
  },
  {
    "epoch": 524,
    "batch_idx": 160,
    "loss": 0.3237372040748596,
    "learning_rate": 0.0008399766893612094,
    "timestep_mean": 398.84375,
    "epsilon_norm": 6.482713222503662,
    "prediction_norm": 5.159402847290039,
    "timestamp": "2025-07-14T12:33:52.070506"
  },
  {
    "epoch": 524,
    "batch_idx": 192,
    "loss": 0.33391955494880676,
    "learning_rate": 0.0008399766893612094,
    "timestep_mean": 333.25,
    "epsilon_norm": 4.1367597579956055,
    "prediction_norm": 2.727252721786499,
    "timestamp": "2025-07-14T12:33:52.075753"
  },
  {
    "epoch": 525,
    "batch_idx": 0,
    "loss": 0.30338457226753235,
    "learning_rate": 0.0008394003727664708,
    "timestep_mean": 476.46875,
    "epsilon_norm": 8.519425392150879,
    "prediction_norm": 6.598670959472656,
    "timestamp": "2025-07-14T12:33:52.080955"
  },
  {
    "epoch": 525,
    "batch_idx": 32,
    "loss": 0.338346004486084,
    "learning_rate": 0.0008394003727664708,
    "timestep_mean": 447.28125,
    "epsilon_norm": 9.213512420654297,
    "prediction_norm": 7.223980903625488,
    "timestamp": "2025-07-14T12:33:52.086096"
  },
  {
    "epoch": 525,
    "batch_idx": 64,
    "loss": 0.1473519504070282,
    "learning_rate": 0.0008394003727664708,
    "timestep_mean": 551.8125,
    "epsilon_norm": 6.859052658081055,
    "prediction_norm": 5.5952982902526855,
    "timestamp": "2025-07-14T12:33:52.091164"
  },
  {
    "epoch": 525,
    "batch_idx": 96,
    "loss": 0.3139372766017914,
    "learning_rate": 0.0008394003727664708,
    "timestep_mean": 479.1875,
    "epsilon_norm": 7.4858784675598145,
    "prediction_norm": 5.395910263061523,
    "timestamp": "2025-07-14T12:33:52.096401"
  },
  {
    "epoch": 525,
    "batch_idx": 128,
    "loss": 0.20260125398635864,
    "learning_rate": 0.0008394003727664708,
    "timestep_mean": 582.5625,
    "epsilon_norm": 8.69979190826416,
    "prediction_norm": 7.280420303344727,
    "timestamp": "2025-07-14T12:33:52.101641"
  },
  {
    "epoch": 525,
    "batch_idx": 160,
    "loss": 0.20781418681144714,
    "learning_rate": 0.0008394003727664708,
    "timestep_mean": 531.0625,
    "epsilon_norm": 7.082242965698242,
    "prediction_norm": 5.943947792053223,
    "timestamp": "2025-07-14T12:33:52.106853"
  },
  {
    "epoch": 525,
    "batch_idx": 192,
    "loss": 0.15544039011001587,
    "learning_rate": 0.0008394003727664708,
    "timestep_mean": 606.75,
    "epsilon_norm": 3.6843650341033936,
    "prediction_norm": 3.410378932952881,
    "timestamp": "2025-07-14T12:33:52.112008"
  },
  {
    "epoch": 526,
    "batch_idx": 0,
    "loss": 0.41206103563308716,
    "learning_rate": 0.000838823218735051,
    "timestep_mean": 473.53125,
    "epsilon_norm": 8.063704490661621,
    "prediction_norm": 6.601579666137695,
    "timestamp": "2025-07-14T12:33:52.117265"
  },
  {
    "epoch": 526,
    "batch_idx": 32,
    "loss": 0.34525904059410095,
    "learning_rate": 0.000838823218735051,
    "timestep_mean": 480.8125,
    "epsilon_norm": 9.376534461975098,
    "prediction_norm": 7.938535690307617,
    "timestamp": "2025-07-14T12:33:52.122426"
  },
  {
    "epoch": 526,
    "batch_idx": 64,
    "loss": 0.36619997024536133,
    "learning_rate": 0.000838823218735051,
    "timestep_mean": 449.625,
    "epsilon_norm": 7.598244667053223,
    "prediction_norm": 5.0674943923950195,
    "timestamp": "2025-07-14T12:33:52.127642"
  },
  {
    "epoch": 526,
    "batch_idx": 96,
    "loss": 0.23432204127311707,
    "learning_rate": 0.000838823218735051,
    "timestep_mean": 532.84375,
    "epsilon_norm": 9.174925804138184,
    "prediction_norm": 8.31879711151123,
    "timestamp": "2025-07-14T12:33:52.132874"
  },
  {
    "epoch": 526,
    "batch_idx": 128,
    "loss": 0.39410948753356934,
    "learning_rate": 0.000838823218735051,
    "timestep_mean": 503.09375,
    "epsilon_norm": 6.886312007904053,
    "prediction_norm": 5.4195637702941895,
    "timestamp": "2025-07-14T12:33:52.138179"
  },
  {
    "epoch": 526,
    "batch_idx": 160,
    "loss": 0.3270808756351471,
    "learning_rate": 0.000838823218735051,
    "timestep_mean": 443.03125,
    "epsilon_norm": 8.31410026550293,
    "prediction_norm": 7.30116605758667,
    "timestamp": "2025-07-14T12:33:52.143341"
  },
  {
    "epoch": 526,
    "batch_idx": 192,
    "loss": 0.4901210069656372,
    "learning_rate": 0.000838823218735051,
    "timestep_mean": 425.875,
    "epsilon_norm": 4.007627487182617,
    "prediction_norm": 3.1198952198028564,
    "timestamp": "2025-07-14T12:33:52.148517"
  },
  {
    "epoch": 527,
    "batch_idx": 0,
    "loss": 0.3215678036212921,
    "learning_rate": 0.0008382452286910205,
    "timestep_mean": 442.6875,
    "epsilon_norm": 7.612481117248535,
    "prediction_norm": 6.206309795379639,
    "timestamp": "2025-07-14T12:33:52.153756"
  },
  {
    "epoch": 527,
    "batch_idx": 32,
    "loss": 0.2709117829799652,
    "learning_rate": 0.0008382452286910205,
    "timestep_mean": 529.125,
    "epsilon_norm": 7.788471698760986,
    "prediction_norm": 7.614226341247559,
    "timestamp": "2025-07-14T12:33:52.159068"
  },
  {
    "epoch": 527,
    "batch_idx": 64,
    "loss": 0.2600591778755188,
    "learning_rate": 0.0008382452286910205,
    "timestep_mean": 558.90625,
    "epsilon_norm": 8.002154350280762,
    "prediction_norm": 7.903932094573975,
    "timestamp": "2025-07-14T12:33:52.164322"
  },
  {
    "epoch": 527,
    "batch_idx": 96,
    "loss": 0.21520546078681946,
    "learning_rate": 0.0008382452286910205,
    "timestep_mean": 561.15625,
    "epsilon_norm": 7.808770179748535,
    "prediction_norm": 7.141592025756836,
    "timestamp": "2025-07-14T12:33:52.169429"
  },
  {
    "epoch": 527,
    "batch_idx": 128,
    "loss": 0.22082629799842834,
    "learning_rate": 0.0008382452286910205,
    "timestep_mean": 458.875,
    "epsilon_norm": 7.373908996582031,
    "prediction_norm": 6.93440580368042,
    "timestamp": "2025-07-14T12:33:52.174626"
  },
  {
    "epoch": 527,
    "batch_idx": 160,
    "loss": 0.35565319657325745,
    "learning_rate": 0.0008382452286910205,
    "timestep_mean": 525.65625,
    "epsilon_norm": 7.452085494995117,
    "prediction_norm": 6.8759660720825195,
    "timestamp": "2025-07-14T12:33:52.179978"
  },
  {
    "epoch": 527,
    "batch_idx": 192,
    "loss": 0.4408766031265259,
    "learning_rate": 0.0008382452286910205,
    "timestep_mean": 365.375,
    "epsilon_norm": 3.9788994789123535,
    "prediction_norm": 3.084705352783203,
    "timestamp": "2025-07-14T12:33:52.185155"
  },
  {
    "epoch": 528,
    "batch_idx": 0,
    "loss": 0.17509512603282928,
    "learning_rate": 0.0008376664040605121,
    "timestep_mean": 549.1875,
    "epsilon_norm": 8.159902572631836,
    "prediction_norm": 7.444177627563477,
    "timestamp": "2025-07-14T12:33:52.190399"
  },
  {
    "epoch": 528,
    "batch_idx": 32,
    "loss": 0.2729312777519226,
    "learning_rate": 0.0008376664040605121,
    "timestep_mean": 602.125,
    "epsilon_norm": 7.636791229248047,
    "prediction_norm": 6.481707572937012,
    "timestamp": "2025-07-14T12:33:52.195624"
  },
  {
    "epoch": 528,
    "batch_idx": 64,
    "loss": 0.260184645652771,
    "learning_rate": 0.0008376664040605121,
    "timestep_mean": 447.1875,
    "epsilon_norm": 7.638185501098633,
    "prediction_norm": 6.283962726593018,
    "timestamp": "2025-07-14T12:33:52.200843"
  },
  {
    "epoch": 528,
    "batch_idx": 96,
    "loss": 0.31550565361976624,
    "learning_rate": 0.0008376664040605121,
    "timestep_mean": 526.34375,
    "epsilon_norm": 8.738543510437012,
    "prediction_norm": 7.252481937408447,
    "timestamp": "2025-07-14T12:33:52.206008"
  },
  {
    "epoch": 528,
    "batch_idx": 128,
    "loss": 0.20060300827026367,
    "learning_rate": 0.0008376664040605121,
    "timestep_mean": 540.46875,
    "epsilon_norm": 7.266801357269287,
    "prediction_norm": 6.641214847564697,
    "timestamp": "2025-07-14T12:33:52.211216"
  },
  {
    "epoch": 528,
    "batch_idx": 160,
    "loss": 0.3113388121128082,
    "learning_rate": 0.0008376664040605121,
    "timestep_mean": 552.40625,
    "epsilon_norm": 8.18474006652832,
    "prediction_norm": 6.668787956237793,
    "timestamp": "2025-07-14T12:33:52.216203"
  },
  {
    "epoch": 528,
    "batch_idx": 192,
    "loss": 0.4243975281715393,
    "learning_rate": 0.0008376664040605121,
    "timestep_mean": 428.75,
    "epsilon_norm": 3.222080707550049,
    "prediction_norm": 1.4636726379394531,
    "timestamp": "2025-07-14T12:33:52.221251"
  },
  {
    "epoch": 529,
    "batch_idx": 0,
    "loss": 0.28180795907974243,
    "learning_rate": 0.0008370867462717182,
    "timestep_mean": 467.84375,
    "epsilon_norm": 7.8800859451293945,
    "prediction_norm": 6.1013712882995605,
    "timestamp": "2025-07-14T12:33:52.226365"
  },
  {
    "epoch": 529,
    "batch_idx": 32,
    "loss": 0.32281821966171265,
    "learning_rate": 0.0008370867462717182,
    "timestep_mean": 405.8125,
    "epsilon_norm": 8.562814712524414,
    "prediction_norm": 7.879406452178955,
    "timestamp": "2025-07-14T12:33:52.231592"
  },
  {
    "epoch": 529,
    "batch_idx": 64,
    "loss": 0.41674625873565674,
    "learning_rate": 0.0008370867462717182,
    "timestep_mean": 430.65625,
    "epsilon_norm": 7.355148792266846,
    "prediction_norm": 4.637368679046631,
    "timestamp": "2025-07-14T12:33:52.236791"
  },
  {
    "epoch": 529,
    "batch_idx": 96,
    "loss": 0.14898405969142914,
    "learning_rate": 0.0008370867462717182,
    "timestep_mean": 605.8125,
    "epsilon_norm": 7.716498851776123,
    "prediction_norm": 6.7943572998046875,
    "timestamp": "2025-07-14T12:33:52.242053"
  },
  {
    "epoch": 529,
    "batch_idx": 128,
    "loss": 0.3180360198020935,
    "learning_rate": 0.0008370867462717182,
    "timestep_mean": 572.34375,
    "epsilon_norm": 7.8976030349731445,
    "prediction_norm": 5.9948906898498535,
    "timestamp": "2025-07-14T12:33:52.247264"
  },
  {
    "epoch": 529,
    "batch_idx": 160,
    "loss": 0.30170148611068726,
    "learning_rate": 0.0008370867462717182,
    "timestep_mean": 512.84375,
    "epsilon_norm": 8.489013671875,
    "prediction_norm": 6.44303560256958,
    "timestamp": "2025-07-14T12:33:52.252340"
  },
  {
    "epoch": 529,
    "batch_idx": 192,
    "loss": 0.7004818916320801,
    "learning_rate": 0.0008370867462717182,
    "timestep_mean": 408.5,
    "epsilon_norm": 4.157259941101074,
    "prediction_norm": 2.3109261989593506,
    "timestamp": "2025-07-14T12:33:52.257438"
  },
  {
    "epoch": 530,
    "batch_idx": 0,
    "loss": 0.4211270213127136,
    "learning_rate": 0.0008365062567548864,
    "timestep_mean": 528.6875,
    "epsilon_norm": 8.168675422668457,
    "prediction_norm": 5.609532356262207,
    "timestamp": "2025-07-14T12:33:52.262664"
  },
  {
    "epoch": 530,
    "batch_idx": 32,
    "loss": 0.3094230592250824,
    "learning_rate": 0.0008365062567548864,
    "timestep_mean": 540.96875,
    "epsilon_norm": 8.470288276672363,
    "prediction_norm": 6.995694160461426,
    "timestamp": "2025-07-14T12:33:52.267661"
  },
  {
    "epoch": 530,
    "batch_idx": 64,
    "loss": 0.1558210551738739,
    "learning_rate": 0.0008365062567548864,
    "timestep_mean": 524.90625,
    "epsilon_norm": 8.682201385498047,
    "prediction_norm": 8.095404624938965,
    "timestamp": "2025-07-14T12:33:52.272846"
  },
  {
    "epoch": 530,
    "batch_idx": 96,
    "loss": 0.20753565430641174,
    "learning_rate": 0.0008365062567548864,
    "timestep_mean": 456.03125,
    "epsilon_norm": 8.37094783782959,
    "prediction_norm": 7.23136043548584,
    "timestamp": "2025-07-14T12:33:52.278112"
  },
  {
    "epoch": 530,
    "batch_idx": 128,
    "loss": 0.14348457753658295,
    "learning_rate": 0.0008365062567548864,
    "timestep_mean": 585.5,
    "epsilon_norm": 7.787761211395264,
    "prediction_norm": 7.108295917510986,
    "timestamp": "2025-07-14T12:33:52.283503"
  },
  {
    "epoch": 530,
    "batch_idx": 160,
    "loss": 0.21715721487998962,
    "learning_rate": 0.0008365062567548864,
    "timestep_mean": 503.34375,
    "epsilon_norm": 8.125531196594238,
    "prediction_norm": 7.285629749298096,
    "timestamp": "2025-07-14T12:33:52.288702"
  },
  {
    "epoch": 530,
    "batch_idx": 192,
    "loss": 0.2554613947868347,
    "learning_rate": 0.0008365062567548864,
    "timestep_mean": 564.5,
    "epsilon_norm": 3.5384433269500732,
    "prediction_norm": 3.4447591304779053,
    "timestamp": "2025-07-14T12:33:52.293789"
  },
  {
    "epoch": 531,
    "batch_idx": 0,
    "loss": 0.3221036195755005,
    "learning_rate": 0.0008359249369423174,
    "timestep_mean": 508.875,
    "epsilon_norm": 7.913827896118164,
    "prediction_norm": 6.6470866203308105,
    "timestamp": "2025-07-14T12:33:52.298959"
  },
  {
    "epoch": 531,
    "batch_idx": 32,
    "loss": 0.32849961519241333,
    "learning_rate": 0.0008359249369423174,
    "timestep_mean": 478.125,
    "epsilon_norm": 7.998474597930908,
    "prediction_norm": 7.667579174041748,
    "timestamp": "2025-07-14T12:33:52.304126"
  },
  {
    "epoch": 531,
    "batch_idx": 64,
    "loss": 0.31976282596588135,
    "learning_rate": 0.0008359249369423174,
    "timestep_mean": 495.09375,
    "epsilon_norm": 9.30974006652832,
    "prediction_norm": 8.828751564025879,
    "timestamp": "2025-07-14T12:33:52.309274"
  },
  {
    "epoch": 531,
    "batch_idx": 96,
    "loss": 0.2196706235408783,
    "learning_rate": 0.0008359249369423174,
    "timestep_mean": 550.34375,
    "epsilon_norm": 7.093867778778076,
    "prediction_norm": 6.85682487487793,
    "timestamp": "2025-07-14T12:33:52.314440"
  },
  {
    "epoch": 531,
    "batch_idx": 128,
    "loss": 0.2198629081249237,
    "learning_rate": 0.0008359249369423174,
    "timestep_mean": 553.0625,
    "epsilon_norm": 8.36153507232666,
    "prediction_norm": 7.999627113342285,
    "timestamp": "2025-07-14T12:33:52.319474"
  },
  {
    "epoch": 531,
    "batch_idx": 160,
    "loss": 0.24546602368354797,
    "learning_rate": 0.0008359249369423174,
    "timestep_mean": 489.03125,
    "epsilon_norm": 7.011900901794434,
    "prediction_norm": 6.475665092468262,
    "timestamp": "2025-07-14T12:33:52.324303"
  },
  {
    "epoch": 531,
    "batch_idx": 192,
    "loss": 0.09621007740497589,
    "learning_rate": 0.0008359249369423174,
    "timestep_mean": 570.875,
    "epsilon_norm": 2.9461402893066406,
    "prediction_norm": 2.5170459747314453,
    "timestamp": "2025-07-14T12:33:52.329226"
  },
  {
    "epoch": 532,
    "batch_idx": 0,
    "loss": 0.1509159505367279,
    "learning_rate": 0.0008353427882683599,
    "timestep_mean": 623.25,
    "epsilon_norm": 9.085874557495117,
    "prediction_norm": 8.259134292602539,
    "timestamp": "2025-07-14T12:33:52.334133"
  },
  {
    "epoch": 532,
    "batch_idx": 32,
    "loss": 0.26717409491539,
    "learning_rate": 0.0008353427882683599,
    "timestep_mean": 494.5,
    "epsilon_norm": 7.126208782196045,
    "prediction_norm": 6.733205318450928,
    "timestamp": "2025-07-14T12:33:52.338998"
  },
  {
    "epoch": 532,
    "batch_idx": 64,
    "loss": 0.2191016674041748,
    "learning_rate": 0.0008353427882683599,
    "timestep_mean": 547.875,
    "epsilon_norm": 8.701619148254395,
    "prediction_norm": 7.193819522857666,
    "timestamp": "2025-07-14T12:33:52.343862"
  },
  {
    "epoch": 532,
    "batch_idx": 96,
    "loss": 0.48899713158607483,
    "learning_rate": 0.0008353427882683599,
    "timestep_mean": 470.5,
    "epsilon_norm": 7.6466240882873535,
    "prediction_norm": 5.40996789932251,
    "timestamp": "2025-07-14T12:33:52.348688"
  },
  {
    "epoch": 532,
    "batch_idx": 128,
    "loss": 0.24317361414432526,
    "learning_rate": 0.0008353427882683599,
    "timestep_mean": 539.53125,
    "epsilon_norm": 8.415589332580566,
    "prediction_norm": 6.866358757019043,
    "timestamp": "2025-07-14T12:33:52.353532"
  },
  {
    "epoch": 532,
    "batch_idx": 160,
    "loss": 0.17750602960586548,
    "learning_rate": 0.0008353427882683599,
    "timestep_mean": 572.6875,
    "epsilon_norm": 8.208626747131348,
    "prediction_norm": 7.206963539123535,
    "timestamp": "2025-07-14T12:33:52.358416"
  },
  {
    "epoch": 532,
    "batch_idx": 192,
    "loss": 0.09730245172977448,
    "learning_rate": 0.0008353427882683599,
    "timestep_mean": 642.5,
    "epsilon_norm": 3.493955135345459,
    "prediction_norm": 3.0765326023101807,
    "timestamp": "2025-07-14T12:33:52.363240"
  },
  {
    "epoch": 533,
    "batch_idx": 0,
    "loss": 0.4051516056060791,
    "learning_rate": 0.0008347598121694076,
    "timestep_mean": 384.53125,
    "epsilon_norm": 8.440962791442871,
    "prediction_norm": 5.513869762420654,
    "timestamp": "2025-07-14T12:33:52.368178"
  },
  {
    "epoch": 533,
    "batch_idx": 32,
    "loss": 0.3845905065536499,
    "learning_rate": 0.0008347598121694076,
    "timestep_mean": 481.53125,
    "epsilon_norm": 8.567743301391602,
    "prediction_norm": 6.531632900238037,
    "timestamp": "2025-07-14T12:33:52.373037"
  },
  {
    "epoch": 533,
    "batch_idx": 64,
    "loss": 0.271743506193161,
    "learning_rate": 0.0008347598121694076,
    "timestep_mean": 559.3125,
    "epsilon_norm": 7.490133285522461,
    "prediction_norm": 6.133842945098877,
    "timestamp": "2025-07-14T12:33:52.377862"
  },
  {
    "epoch": 533,
    "batch_idx": 96,
    "loss": 0.40979766845703125,
    "learning_rate": 0.0008347598121694076,
    "timestep_mean": 556.40625,
    "epsilon_norm": 8.08508586883545,
    "prediction_norm": 5.801601409912109,
    "timestamp": "2025-07-14T12:33:52.382673"
  },
  {
    "epoch": 533,
    "batch_idx": 128,
    "loss": 0.2669743597507477,
    "learning_rate": 0.0008347598121694076,
    "timestep_mean": 386.15625,
    "epsilon_norm": 8.135239601135254,
    "prediction_norm": 6.9810791015625,
    "timestamp": "2025-07-14T12:33:52.387577"
  },
  {
    "epoch": 533,
    "batch_idx": 160,
    "loss": 0.21661438047885895,
    "learning_rate": 0.0008347598121694076,
    "timestep_mean": 536.46875,
    "epsilon_norm": 8.835989952087402,
    "prediction_norm": 7.854270935058594,
    "timestamp": "2025-07-14T12:33:52.392406"
  },
  {
    "epoch": 533,
    "batch_idx": 192,
    "loss": 0.2988683879375458,
    "learning_rate": 0.0008347598121694076,
    "timestep_mean": 424.875,
    "epsilon_norm": 4.054152488708496,
    "prediction_norm": 3.65409517288208,
    "timestamp": "2025-07-14T12:33:52.397172"
  },
  {
    "epoch": 534,
    "batch_idx": 0,
    "loss": 0.23344919085502625,
    "learning_rate": 0.0008341760100838964,
    "timestep_mean": 540.34375,
    "epsilon_norm": 9.108226776123047,
    "prediction_norm": 8.484611511230469,
    "timestamp": "2025-07-14T12:33:52.401979"
  },
  {
    "epoch": 534,
    "batch_idx": 32,
    "loss": 0.5291018486022949,
    "learning_rate": 0.0008341760100838964,
    "timestep_mean": 370.21875,
    "epsilon_norm": 8.580574989318848,
    "prediction_norm": 5.473289489746094,
    "timestamp": "2025-07-14T12:33:52.406790"
  },
  {
    "epoch": 534,
    "batch_idx": 64,
    "loss": 0.5618777871131897,
    "learning_rate": 0.0008341760100838964,
    "timestep_mean": 427.71875,
    "epsilon_norm": 9.655828475952148,
    "prediction_norm": 7.09073543548584,
    "timestamp": "2025-07-14T12:33:52.411678"
  },
  {
    "epoch": 534,
    "batch_idx": 96,
    "loss": 0.2826679050922394,
    "learning_rate": 0.0008341760100838964,
    "timestep_mean": 477.40625,
    "epsilon_norm": 8.299686431884766,
    "prediction_norm": 7.055700302124023,
    "timestamp": "2025-07-14T12:33:52.416896"
  },
  {
    "epoch": 534,
    "batch_idx": 128,
    "loss": 0.5204580426216125,
    "learning_rate": 0.0008341760100838964,
    "timestep_mean": 473.03125,
    "epsilon_norm": 8.711882591247559,
    "prediction_norm": 6.779055595397949,
    "timestamp": "2025-07-14T12:33:52.421968"
  },
  {
    "epoch": 534,
    "batch_idx": 160,
    "loss": 0.46809571981430054,
    "learning_rate": 0.0008341760100838964,
    "timestep_mean": 417.34375,
    "epsilon_norm": 7.401791095733643,
    "prediction_norm": 5.223972797393799,
    "timestamp": "2025-07-14T12:33:52.427003"
  },
  {
    "epoch": 534,
    "batch_idx": 192,
    "loss": 0.5177828073501587,
    "learning_rate": 0.0008341760100838964,
    "timestep_mean": 325.375,
    "epsilon_norm": 3.1437900066375732,
    "prediction_norm": 2.4544007778167725,
    "timestamp": "2025-07-14T12:33:52.432010"
  },
  {
    "epoch": 535,
    "batch_idx": 0,
    "loss": 0.43206316232681274,
    "learning_rate": 0.0008335913834522998,
    "timestep_mean": 472.4375,
    "epsilon_norm": 7.944831848144531,
    "prediction_norm": 6.482742786407471,
    "timestamp": "2025-07-14T12:33:52.436936"
  },
  {
    "epoch": 535,
    "batch_idx": 32,
    "loss": 0.38525354862213135,
    "learning_rate": 0.0008335913834522998,
    "timestep_mean": 471.4375,
    "epsilon_norm": 8.12172794342041,
    "prediction_norm": 7.197832107543945,
    "timestamp": "2025-07-14T12:33:52.441902"
  },
  {
    "epoch": 535,
    "batch_idx": 64,
    "loss": 0.264254093170166,
    "learning_rate": 0.0008335913834522998,
    "timestep_mean": 528.71875,
    "epsilon_norm": 8.136259078979492,
    "prediction_norm": 7.175394535064697,
    "timestamp": "2025-07-14T12:33:52.446840"
  },
  {
    "epoch": 535,
    "batch_idx": 96,
    "loss": 0.3772798776626587,
    "learning_rate": 0.0008335913834522998,
    "timestep_mean": 513.34375,
    "epsilon_norm": 7.0763020515441895,
    "prediction_norm": 5.864112854003906,
    "timestamp": "2025-07-14T12:33:52.451730"
  },
  {
    "epoch": 535,
    "batch_idx": 128,
    "loss": 0.47192394733428955,
    "learning_rate": 0.0008335913834522998,
    "timestep_mean": 378.59375,
    "epsilon_norm": 8.844780921936035,
    "prediction_norm": 6.867459774017334,
    "timestamp": "2025-07-14T12:33:52.456569"
  },
  {
    "epoch": 535,
    "batch_idx": 160,
    "loss": 0.15575507283210754,
    "learning_rate": 0.0008335913834522998,
    "timestep_mean": 538.75,
    "epsilon_norm": 7.881137371063232,
    "prediction_norm": 7.490925312042236,
    "timestamp": "2025-07-14T12:33:52.461399"
  },
  {
    "epoch": 535,
    "batch_idx": 192,
    "loss": 0.09629923850297928,
    "learning_rate": 0.0008335913834522998,
    "timestep_mean": 418.625,
    "epsilon_norm": 2.67787766456604,
    "prediction_norm": 2.8573760986328125,
    "timestamp": "2025-07-14T12:33:52.466285"
  },
  {
    "epoch": 536,
    "batch_idx": 0,
    "loss": 0.22707730531692505,
    "learning_rate": 0.0008330059337171257,
    "timestep_mean": 526.71875,
    "epsilon_norm": 7.5422892570495605,
    "prediction_norm": 6.481544017791748,
    "timestamp": "2025-07-14T12:33:52.471384"
  },
  {
    "epoch": 536,
    "batch_idx": 32,
    "loss": 0.12950687110424042,
    "learning_rate": 0.0008330059337171257,
    "timestep_mean": 550.5,
    "epsilon_norm": 7.872735023498535,
    "prediction_norm": 7.431615352630615,
    "timestamp": "2025-07-14T12:33:52.476231"
  },
  {
    "epoch": 536,
    "batch_idx": 64,
    "loss": 0.3285813331604004,
    "learning_rate": 0.0008330059337171257,
    "timestep_mean": 579.6875,
    "epsilon_norm": 8.960834503173828,
    "prediction_norm": 7.971649169921875,
    "timestamp": "2025-07-14T12:33:52.481080"
  },
  {
    "epoch": 536,
    "batch_idx": 96,
    "loss": 0.29171764850616455,
    "learning_rate": 0.0008330059337171257,
    "timestep_mean": 520.625,
    "epsilon_norm": 8.258468627929688,
    "prediction_norm": 6.443027019500732,
    "timestamp": "2025-07-14T12:33:52.485937"
  },
  {
    "epoch": 536,
    "batch_idx": 128,
    "loss": 0.22018256783485413,
    "learning_rate": 0.0008330059337171257,
    "timestep_mean": 542.40625,
    "epsilon_norm": 7.632692337036133,
    "prediction_norm": 7.2272515296936035,
    "timestamp": "2025-07-14T12:33:52.490793"
  },
  {
    "epoch": 536,
    "batch_idx": 160,
    "loss": 0.16266292333602905,
    "learning_rate": 0.0008330059337171257,
    "timestep_mean": 548.6875,
    "epsilon_norm": 6.724484920501709,
    "prediction_norm": 5.755990028381348,
    "timestamp": "2025-07-14T12:33:52.495677"
  },
  {
    "epoch": 536,
    "batch_idx": 192,
    "loss": 0.27177584171295166,
    "learning_rate": 0.0008330059337171257,
    "timestep_mean": 484.875,
    "epsilon_norm": 3.4222629070281982,
    "prediction_norm": 2.8577067852020264,
    "timestamp": "2025-07-14T12:33:52.500631"
  },
  {
    "epoch": 537,
    "batch_idx": 0,
    "loss": 0.6052963137626648,
    "learning_rate": 0.0008324196623229134,
    "timestep_mean": 315.09375,
    "epsilon_norm": 7.9021897315979,
    "prediction_norm": 4.996262550354004,
    "timestamp": "2025-07-14T12:33:52.505527"
  },
  {
    "epoch": 537,
    "batch_idx": 32,
    "loss": 0.22394314408302307,
    "learning_rate": 0.0008324196623229134,
    "timestep_mean": 529.34375,
    "epsilon_norm": 7.836503982543945,
    "prediction_norm": 6.4645233154296875,
    "timestamp": "2025-07-14T12:33:52.511267"
  },
  {
    "epoch": 537,
    "batch_idx": 64,
    "loss": 0.3783859610557556,
    "learning_rate": 0.0008324196623229134,
    "timestep_mean": 440.25,
    "epsilon_norm": 7.417535305023193,
    "prediction_norm": 5.7483930587768555,
    "timestamp": "2025-07-14T12:33:52.515955"
  },
  {
    "epoch": 537,
    "batch_idx": 96,
    "loss": 0.24920541048049927,
    "learning_rate": 0.0008324196623229134,
    "timestep_mean": 499.5,
    "epsilon_norm": 8.766390800476074,
    "prediction_norm": 6.952662467956543,
    "timestamp": "2025-07-14T12:33:52.520779"
  },
  {
    "epoch": 537,
    "batch_idx": 128,
    "loss": 0.254915714263916,
    "learning_rate": 0.0008324196623229134,
    "timestep_mean": 479.75,
    "epsilon_norm": 8.940386772155762,
    "prediction_norm": 7.175955772399902,
    "timestamp": "2025-07-14T12:33:52.525516"
  },
  {
    "epoch": 537,
    "batch_idx": 160,
    "loss": 0.4307202100753784,
    "learning_rate": 0.0008324196623229134,
    "timestep_mean": 477.84375,
    "epsilon_norm": 8.132617950439453,
    "prediction_norm": 6.757593154907227,
    "timestamp": "2025-07-14T12:33:52.530220"
  },
  {
    "epoch": 537,
    "batch_idx": 192,
    "loss": 0.07999981939792633,
    "learning_rate": 0.0008324196623229134,
    "timestep_mean": 665.5,
    "epsilon_norm": 5.668647766113281,
    "prediction_norm": 5.244875907897949,
    "timestamp": "2025-07-14T12:33:52.534932"
  },
  {
    "epoch": 538,
    "batch_idx": 0,
    "loss": 0.33527544140815735,
    "learning_rate": 0.0008318325707162292,
    "timestep_mean": 491.40625,
    "epsilon_norm": 8.252907752990723,
    "prediction_norm": 7.650301933288574,
    "timestamp": "2025-07-14T12:33:52.539629"
  },
  {
    "epoch": 538,
    "batch_idx": 32,
    "loss": 0.29269227385520935,
    "learning_rate": 0.0008318325707162292,
    "timestep_mean": 432.53125,
    "epsilon_norm": 6.845277309417725,
    "prediction_norm": 6.017983913421631,
    "timestamp": "2025-07-14T12:33:52.544626"
  },
  {
    "epoch": 538,
    "batch_idx": 64,
    "loss": 0.3073444366455078,
    "learning_rate": 0.0008318325707162292,
    "timestep_mean": 508.625,
    "epsilon_norm": 6.620617389678955,
    "prediction_norm": 4.9281511306762695,
    "timestamp": "2025-07-14T12:33:52.549583"
  },
  {
    "epoch": 538,
    "batch_idx": 96,
    "loss": 0.2605902850627899,
    "learning_rate": 0.0008318325707162292,
    "timestep_mean": 540.84375,
    "epsilon_norm": 7.786811351776123,
    "prediction_norm": 6.1865129470825195,
    "timestamp": "2025-07-14T12:33:52.554698"
  },
  {
    "epoch": 538,
    "batch_idx": 128,
    "loss": 0.21232718229293823,
    "learning_rate": 0.0008318325707162292,
    "timestep_mean": 486.15625,
    "epsilon_norm": 7.485489368438721,
    "prediction_norm": 6.825383186340332,
    "timestamp": "2025-07-14T12:33:52.559706"
  },
  {
    "epoch": 538,
    "batch_idx": 160,
    "loss": 0.5234391689300537,
    "learning_rate": 0.0008318325707162292,
    "timestep_mean": 422.375,
    "epsilon_norm": 7.7663140296936035,
    "prediction_norm": 6.321539878845215,
    "timestamp": "2025-07-14T12:33:52.564628"
  },
  {
    "epoch": 538,
    "batch_idx": 192,
    "loss": 0.19938355684280396,
    "learning_rate": 0.0008318325707162292,
    "timestep_mean": 486.375,
    "epsilon_norm": 4.254465103149414,
    "prediction_norm": 3.745941400527954,
    "timestamp": "2025-07-14T12:33:52.569595"
  },
  {
    "epoch": 539,
    "batch_idx": 0,
    "loss": 0.2691141963005066,
    "learning_rate": 0.0008312446603456632,
    "timestep_mean": 466.09375,
    "epsilon_norm": 6.993662357330322,
    "prediction_norm": 5.718502521514893,
    "timestamp": "2025-07-14T12:33:52.574530"
  },
  {
    "epoch": 539,
    "batch_idx": 32,
    "loss": 0.28295472264289856,
    "learning_rate": 0.0008312446603456632,
    "timestep_mean": 525.34375,
    "epsilon_norm": 8.153265953063965,
    "prediction_norm": 6.34879732131958,
    "timestamp": "2025-07-14T12:33:52.579606"
  },
  {
    "epoch": 539,
    "batch_idx": 64,
    "loss": 0.33967381715774536,
    "learning_rate": 0.0008312446603456632,
    "timestep_mean": 436.21875,
    "epsilon_norm": 7.541039943695068,
    "prediction_norm": 7.032586574554443,
    "timestamp": "2025-07-14T12:33:52.584511"
  },
  {
    "epoch": 539,
    "batch_idx": 96,
    "loss": 0.39150112867355347,
    "learning_rate": 0.0008312446603456632,
    "timestep_mean": 396.0,
    "epsilon_norm": 7.114458084106445,
    "prediction_norm": 5.271183490753174,
    "timestamp": "2025-07-14T12:33:52.589327"
  },
  {
    "epoch": 539,
    "batch_idx": 128,
    "loss": 0.29827070236206055,
    "learning_rate": 0.0008312446603456632,
    "timestep_mean": 464.28125,
    "epsilon_norm": 7.82318639755249,
    "prediction_norm": 6.6539812088012695,
    "timestamp": "2025-07-14T12:33:52.594049"
  },
  {
    "epoch": 539,
    "batch_idx": 160,
    "loss": 0.37606748938560486,
    "learning_rate": 0.0008312446603456632,
    "timestep_mean": 448.0,
    "epsilon_norm": 8.041088104248047,
    "prediction_norm": 5.74648904800415,
    "timestamp": "2025-07-14T12:33:52.598723"
  },
  {
    "epoch": 539,
    "batch_idx": 192,
    "loss": 0.11565300822257996,
    "learning_rate": 0.0008312446603456632,
    "timestep_mean": 572.75,
    "epsilon_norm": 2.5279712677001953,
    "prediction_norm": 2.105950355529785,
    "timestamp": "2025-07-14T12:33:52.603447"
  },
  {
    "epoch": 540,
    "batch_idx": 0,
    "loss": 0.22266167402267456,
    "learning_rate": 0.0008306559326618258,
    "timestep_mean": 469.59375,
    "epsilon_norm": 5.720142841339111,
    "prediction_norm": 5.229923248291016,
    "timestamp": "2025-07-14T12:33:52.608295"
  },
  {
    "epoch": 540,
    "batch_idx": 32,
    "loss": 0.3809134066104889,
    "learning_rate": 0.0008306559326618258,
    "timestep_mean": 403.34375,
    "epsilon_norm": 7.809660911560059,
    "prediction_norm": 5.954322338104248,
    "timestamp": "2025-07-14T12:33:52.613012"
  },
  {
    "epoch": 540,
    "batch_idx": 64,
    "loss": 0.33913835883140564,
    "learning_rate": 0.0008306559326618258,
    "timestep_mean": 457.03125,
    "epsilon_norm": 7.243037700653076,
    "prediction_norm": 5.337432384490967,
    "timestamp": "2025-07-14T12:33:52.617813"
  },
  {
    "epoch": 540,
    "batch_idx": 96,
    "loss": 0.29842472076416016,
    "learning_rate": 0.0008306559326618258,
    "timestep_mean": 494.84375,
    "epsilon_norm": 7.133727550506592,
    "prediction_norm": 5.106333255767822,
    "timestamp": "2025-07-14T12:33:52.622592"
  },
  {
    "epoch": 540,
    "batch_idx": 128,
    "loss": 0.2403424084186554,
    "learning_rate": 0.0008306559326618258,
    "timestep_mean": 502.8125,
    "epsilon_norm": 7.47146463394165,
    "prediction_norm": 5.821852684020996,
    "timestamp": "2025-07-14T12:33:52.627441"
  },
  {
    "epoch": 540,
    "batch_idx": 160,
    "loss": 0.27110055088996887,
    "learning_rate": 0.0008306559326618258,
    "timestep_mean": 506.25,
    "epsilon_norm": 8.244985580444336,
    "prediction_norm": 6.645422458648682,
    "timestamp": "2025-07-14T12:33:52.632191"
  },
  {
    "epoch": 540,
    "batch_idx": 192,
    "loss": 0.4916340708732605,
    "learning_rate": 0.0008306559326618258,
    "timestep_mean": 476.75,
    "epsilon_norm": 4.786910533905029,
    "prediction_norm": 3.6766483783721924,
    "timestamp": "2025-07-14T12:33:52.636983"
  },
  {
    "epoch": 541,
    "batch_idx": 0,
    "loss": 0.303280234336853,
    "learning_rate": 0.0008300663891173442,
    "timestep_mean": 486.0625,
    "epsilon_norm": 7.5859785079956055,
    "prediction_norm": 6.177152633666992,
    "timestamp": "2025-07-14T12:33:52.641868"
  },
  {
    "epoch": 541,
    "batch_idx": 32,
    "loss": 0.2967832088470459,
    "learning_rate": 0.0008300663891173442,
    "timestep_mean": 505.25,
    "epsilon_norm": 7.668684959411621,
    "prediction_norm": 6.143614292144775,
    "timestamp": "2025-07-14T12:33:52.646756"
  },
  {
    "epoch": 541,
    "batch_idx": 64,
    "loss": 0.34314271807670593,
    "learning_rate": 0.0008300663891173442,
    "timestep_mean": 413.84375,
    "epsilon_norm": 7.3320207595825195,
    "prediction_norm": 5.619996070861816,
    "timestamp": "2025-07-14T12:33:52.651577"
  },
  {
    "epoch": 541,
    "batch_idx": 96,
    "loss": 0.23846036195755005,
    "learning_rate": 0.0008300663891173442,
    "timestep_mean": 484.5,
    "epsilon_norm": 8.19950008392334,
    "prediction_norm": 7.32240104675293,
    "timestamp": "2025-07-14T12:33:52.656482"
  },
  {
    "epoch": 541,
    "batch_idx": 128,
    "loss": 0.4327477812767029,
    "learning_rate": 0.0008300663891173442,
    "timestep_mean": 483.625,
    "epsilon_norm": 8.858965873718262,
    "prediction_norm": 6.745444297790527,
    "timestamp": "2025-07-14T12:33:52.661339"
  },
  {
    "epoch": 541,
    "batch_idx": 160,
    "loss": 0.2654668688774109,
    "learning_rate": 0.0008300663891173442,
    "timestep_mean": 453.03125,
    "epsilon_norm": 7.599301815032959,
    "prediction_norm": 6.912647724151611,
    "timestamp": "2025-07-14T12:33:52.666116"
  },
  {
    "epoch": 541,
    "batch_idx": 192,
    "loss": 0.09808027744293213,
    "learning_rate": 0.0008300663891173442,
    "timestep_mean": 384.125,
    "epsilon_norm": 3.8976633548736572,
    "prediction_norm": 3.85032057762146,
    "timestamp": "2025-07-14T12:33:52.671054"
  },
  {
    "epoch": 542,
    "batch_idx": 0,
    "loss": 0.33186566829681396,
    "learning_rate": 0.0008294760311668584,
    "timestep_mean": 395.0,
    "epsilon_norm": 7.846661567687988,
    "prediction_norm": 6.211153507232666,
    "timestamp": "2025-07-14T12:33:52.675849"
  },
  {
    "epoch": 542,
    "batch_idx": 32,
    "loss": 0.4954933524131775,
    "learning_rate": 0.0008294760311668584,
    "timestep_mean": 424.40625,
    "epsilon_norm": 7.752534866333008,
    "prediction_norm": 5.349117755889893,
    "timestamp": "2025-07-14T12:33:52.680757"
  },
  {
    "epoch": 542,
    "batch_idx": 64,
    "loss": 0.39528533816337585,
    "learning_rate": 0.0008294760311668584,
    "timestep_mean": 516.125,
    "epsilon_norm": 8.397403717041016,
    "prediction_norm": 6.534680366516113,
    "timestamp": "2025-07-14T12:33:52.685634"
  },
  {
    "epoch": 542,
    "batch_idx": 96,
    "loss": 0.30075937509536743,
    "learning_rate": 0.0008294760311668584,
    "timestep_mean": 570.1875,
    "epsilon_norm": 8.342193603515625,
    "prediction_norm": 6.799661159515381,
    "timestamp": "2025-07-14T12:33:52.690436"
  },
  {
    "epoch": 542,
    "batch_idx": 128,
    "loss": 0.30069297552108765,
    "learning_rate": 0.0008294760311668584,
    "timestep_mean": 423.125,
    "epsilon_norm": 7.782064437866211,
    "prediction_norm": 6.947736740112305,
    "timestamp": "2025-07-14T12:33:52.695312"
  },
  {
    "epoch": 542,
    "batch_idx": 160,
    "loss": 0.24344246089458466,
    "learning_rate": 0.0008294760311668584,
    "timestep_mean": 480.40625,
    "epsilon_norm": 8.58816909790039,
    "prediction_norm": 7.727373123168945,
    "timestamp": "2025-07-14T12:33:52.700275"
  },
  {
    "epoch": 542,
    "batch_idx": 192,
    "loss": 0.16672325134277344,
    "learning_rate": 0.0008294760311668584,
    "timestep_mean": 528.625,
    "epsilon_norm": 3.5479207038879395,
    "prediction_norm": 3.5220115184783936,
    "timestamp": "2025-07-14T12:33:52.705063"
  },
  {
    "epoch": 543,
    "batch_idx": 0,
    "loss": 0.3057418465614319,
    "learning_rate": 0.0008288848602670179,
    "timestep_mean": 421.28125,
    "epsilon_norm": 6.705413341522217,
    "prediction_norm": 5.738607883453369,
    "timestamp": "2025-07-14T12:33:52.709987"
  },
  {
    "epoch": 543,
    "batch_idx": 32,
    "loss": 0.3721877634525299,
    "learning_rate": 0.0008288848602670179,
    "timestep_mean": 539.9375,
    "epsilon_norm": 7.790609359741211,
    "prediction_norm": 6.177402973175049,
    "timestamp": "2025-07-14T12:33:52.714803"
  },
  {
    "epoch": 543,
    "batch_idx": 64,
    "loss": 0.26648491621017456,
    "learning_rate": 0.0008288848602670179,
    "timestep_mean": 493.625,
    "epsilon_norm": 7.332150936126709,
    "prediction_norm": 7.283907413482666,
    "timestamp": "2025-07-14T12:33:52.719651"
  },
  {
    "epoch": 543,
    "batch_idx": 96,
    "loss": 0.37532609701156616,
    "learning_rate": 0.0008288848602670179,
    "timestep_mean": 451.78125,
    "epsilon_norm": 7.427873611450195,
    "prediction_norm": 6.4242143630981445,
    "timestamp": "2025-07-14T12:33:52.724502"
  },
  {
    "epoch": 543,
    "batch_idx": 128,
    "loss": 0.27622246742248535,
    "learning_rate": 0.0008288848602670179,
    "timestep_mean": 492.5,
    "epsilon_norm": 7.147800922393799,
    "prediction_norm": 6.08174467086792,
    "timestamp": "2025-07-14T12:33:52.729474"
  },
  {
    "epoch": 543,
    "batch_idx": 160,
    "loss": 0.2965796887874603,
    "learning_rate": 0.0008288848602670179,
    "timestep_mean": 520.46875,
    "epsilon_norm": 8.051615715026855,
    "prediction_norm": 6.799569129943848,
    "timestamp": "2025-07-14T12:33:52.734275"
  },
  {
    "epoch": 543,
    "batch_idx": 192,
    "loss": 0.22412309050559998,
    "learning_rate": 0.0008288848602670179,
    "timestep_mean": 414.0,
    "epsilon_norm": 3.4354970455169678,
    "prediction_norm": 2.8407235145568848,
    "timestamp": "2025-07-14T12:33:52.739105"
  },
  {
    "epoch": 544,
    "batch_idx": 0,
    "loss": 0.42166566848754883,
    "learning_rate": 0.0008282928778764781,
    "timestep_mean": 531.28125,
    "epsilon_norm": 7.768196105957031,
    "prediction_norm": 6.33132266998291,
    "timestamp": "2025-07-14T12:33:52.743956"
  },
  {
    "epoch": 544,
    "batch_idx": 32,
    "loss": 0.20672976970672607,
    "learning_rate": 0.0008282928778764781,
    "timestep_mean": 530.0,
    "epsilon_norm": 7.646562099456787,
    "prediction_norm": 6.648271560668945,
    "timestamp": "2025-07-14T12:33:52.748883"
  },
  {
    "epoch": 544,
    "batch_idx": 64,
    "loss": 0.5079541206359863,
    "learning_rate": 0.0008282928778764781,
    "timestep_mean": 405.125,
    "epsilon_norm": 7.99378776550293,
    "prediction_norm": 5.357324123382568,
    "timestamp": "2025-07-14T12:33:52.753754"
  },
  {
    "epoch": 544,
    "batch_idx": 96,
    "loss": 0.31876468658447266,
    "learning_rate": 0.0008282928778764781,
    "timestep_mean": 459.3125,
    "epsilon_norm": 10.119962692260742,
    "prediction_norm": 8.678132057189941,
    "timestamp": "2025-07-14T12:33:52.758623"
  },
  {
    "epoch": 544,
    "batch_idx": 128,
    "loss": 0.1414133459329605,
    "learning_rate": 0.0008282928778764781,
    "timestep_mean": 578.125,
    "epsilon_norm": 7.868107795715332,
    "prediction_norm": 6.841431617736816,
    "timestamp": "2025-07-14T12:33:52.763522"
  },
  {
    "epoch": 544,
    "batch_idx": 160,
    "loss": 0.18987113237380981,
    "learning_rate": 0.0008282928778764781,
    "timestep_mean": 560.46875,
    "epsilon_norm": 8.981481552124023,
    "prediction_norm": 7.659695625305176,
    "timestamp": "2025-07-14T12:33:52.768379"
  },
  {
    "epoch": 544,
    "batch_idx": 192,
    "loss": 0.1451532542705536,
    "learning_rate": 0.0008282928778764781,
    "timestep_mean": 496.875,
    "epsilon_norm": 4.05851411819458,
    "prediction_norm": 3.789524793624878,
    "timestamp": "2025-07-14T12:33:52.773285"
  },
  {
    "epoch": 545,
    "batch_idx": 0,
    "loss": 0.4173038601875305,
    "learning_rate": 0.0008277000854558968,
    "timestep_mean": 480.03125,
    "epsilon_norm": 9.178007125854492,
    "prediction_norm": 7.905139446258545,
    "timestamp": "2025-07-14T12:33:52.778147"
  },
  {
    "epoch": 545,
    "batch_idx": 32,
    "loss": 0.2756223678588867,
    "learning_rate": 0.0008277000854558968,
    "timestep_mean": 571.46875,
    "epsilon_norm": 7.418406009674072,
    "prediction_norm": 6.097850799560547,
    "timestamp": "2025-07-14T12:33:52.782971"
  },
  {
    "epoch": 545,
    "batch_idx": 64,
    "loss": 0.39420855045318604,
    "learning_rate": 0.0008277000854558968,
    "timestep_mean": 429.4375,
    "epsilon_norm": 8.243245124816895,
    "prediction_norm": 6.449786186218262,
    "timestamp": "2025-07-14T12:33:52.787976"
  },
  {
    "epoch": 545,
    "batch_idx": 96,
    "loss": 0.32679283618927,
    "learning_rate": 0.0008277000854558968,
    "timestep_mean": 415.09375,
    "epsilon_norm": 7.439220428466797,
    "prediction_norm": 5.80004358291626,
    "timestamp": "2025-07-14T12:33:52.792798"
  },
  {
    "epoch": 545,
    "batch_idx": 128,
    "loss": 0.546543538570404,
    "learning_rate": 0.0008277000854558968,
    "timestep_mean": 545.0625,
    "epsilon_norm": 9.38393497467041,
    "prediction_norm": 7.609503269195557,
    "timestamp": "2025-07-14T12:33:52.797679"
  },
  {
    "epoch": 545,
    "batch_idx": 160,
    "loss": 0.20860734581947327,
    "learning_rate": 0.0008277000854558968,
    "timestep_mean": 444.96875,
    "epsilon_norm": 7.7297043800354,
    "prediction_norm": 6.690979957580566,
    "timestamp": "2025-07-14T12:33:52.802530"
  },
  {
    "epoch": 545,
    "batch_idx": 192,
    "loss": 0.24983453750610352,
    "learning_rate": 0.0008277000854558968,
    "timestep_mean": 548.875,
    "epsilon_norm": 3.5043253898620605,
    "prediction_norm": 2.717806100845337,
    "timestamp": "2025-07-14T12:33:52.807405"
  },
  {
    "epoch": 546,
    "batch_idx": 0,
    "loss": 0.17697884142398834,
    "learning_rate": 0.0008271064844679305,
    "timestep_mean": 580.34375,
    "epsilon_norm": 7.531233310699463,
    "prediction_norm": 7.321969985961914,
    "timestamp": "2025-07-14T12:33:52.812352"
  },
  {
    "epoch": 546,
    "batch_idx": 32,
    "loss": 0.23679709434509277,
    "learning_rate": 0.0008271064844679305,
    "timestep_mean": 467.53125,
    "epsilon_norm": 8.353509902954102,
    "prediction_norm": 7.641147136688232,
    "timestamp": "2025-07-14T12:33:52.817335"
  },
  {
    "epoch": 546,
    "batch_idx": 64,
    "loss": 0.4121468663215637,
    "learning_rate": 0.0008271064844679305,
    "timestep_mean": 532.78125,
    "epsilon_norm": 7.6023101806640625,
    "prediction_norm": 6.0009379386901855,
    "timestamp": "2025-07-14T12:33:52.822247"
  },
  {
    "epoch": 546,
    "batch_idx": 96,
    "loss": 0.24951545894145966,
    "learning_rate": 0.0008271064844679305,
    "timestep_mean": 531.59375,
    "epsilon_norm": 8.118390083312988,
    "prediction_norm": 7.1881537437438965,
    "timestamp": "2025-07-14T12:33:52.827118"
  },
  {
    "epoch": 546,
    "batch_idx": 128,
    "loss": 0.5662369132041931,
    "learning_rate": 0.0008271064844679305,
    "timestep_mean": 428.5,
    "epsilon_norm": 7.620622158050537,
    "prediction_norm": 5.41257381439209,
    "timestamp": "2025-07-14T12:33:52.832019"
  },
  {
    "epoch": 546,
    "batch_idx": 160,
    "loss": 0.2539081871509552,
    "learning_rate": 0.0008271064844679305,
    "timestep_mean": 533.5625,
    "epsilon_norm": 8.123580932617188,
    "prediction_norm": 7.014669418334961,
    "timestamp": "2025-07-14T12:33:52.836904"
  },
  {
    "epoch": 546,
    "batch_idx": 192,
    "loss": 0.37623172998428345,
    "learning_rate": 0.0008271064844679305,
    "timestep_mean": 515.25,
    "epsilon_norm": 3.973360538482666,
    "prediction_norm": 3.5406253337860107,
    "timestamp": "2025-07-14T12:33:52.841858"
  },
  {
    "epoch": 547,
    "batch_idx": 0,
    "loss": 0.46760281920433044,
    "learning_rate": 0.0008265120763772304,
    "timestep_mean": 439.90625,
    "epsilon_norm": 9.7317533493042,
    "prediction_norm": 7.770279407501221,
    "timestamp": "2025-07-14T12:33:52.846858"
  },
  {
    "epoch": 547,
    "batch_idx": 32,
    "loss": 0.43250665068626404,
    "learning_rate": 0.0008265120763772304,
    "timestep_mean": 416.5,
    "epsilon_norm": 8.747934341430664,
    "prediction_norm": 6.887958526611328,
    "timestamp": "2025-07-14T12:33:52.851767"
  },
  {
    "epoch": 547,
    "batch_idx": 64,
    "loss": 0.3981080949306488,
    "learning_rate": 0.0008265120763772304,
    "timestep_mean": 551.4375,
    "epsilon_norm": 8.944707870483398,
    "prediction_norm": 7.519352912902832,
    "timestamp": "2025-07-14T12:33:52.856675"
  },
  {
    "epoch": 547,
    "batch_idx": 96,
    "loss": 0.2425062358379364,
    "learning_rate": 0.0008265120763772304,
    "timestep_mean": 458.21875,
    "epsilon_norm": 7.203482151031494,
    "prediction_norm": 6.555600643157959,
    "timestamp": "2025-07-14T12:33:52.861628"
  },
  {
    "epoch": 547,
    "batch_idx": 128,
    "loss": 0.2856036126613617,
    "learning_rate": 0.0008265120763772304,
    "timestep_mean": 505.6875,
    "epsilon_norm": 8.07113265991211,
    "prediction_norm": 6.665631294250488,
    "timestamp": "2025-07-14T12:33:52.866673"
  },
  {
    "epoch": 547,
    "batch_idx": 160,
    "loss": 0.18258745968341827,
    "learning_rate": 0.0008265120763772304,
    "timestep_mean": 536.4375,
    "epsilon_norm": 7.350720405578613,
    "prediction_norm": 6.563277244567871,
    "timestamp": "2025-07-14T12:33:52.871618"
  },
  {
    "epoch": 547,
    "batch_idx": 192,
    "loss": 0.15635423362255096,
    "learning_rate": 0.0008265120763772304,
    "timestep_mean": 560.125,
    "epsilon_norm": 4.862567901611328,
    "prediction_norm": 4.448986053466797,
    "timestamp": "2025-07-14T12:33:52.876525"
  },
  {
    "epoch": 548,
    "batch_idx": 0,
    "loss": 0.3594135642051697,
    "learning_rate": 0.0008259168626504394,
    "timestep_mean": 503.5625,
    "epsilon_norm": 8.555274963378906,
    "prediction_norm": 6.655765056610107,
    "timestamp": "2025-07-14T12:33:52.881430"
  },
  {
    "epoch": 548,
    "batch_idx": 32,
    "loss": 0.3860677480697632,
    "learning_rate": 0.0008259168626504394,
    "timestep_mean": 520.125,
    "epsilon_norm": 8.254118919372559,
    "prediction_norm": 6.3394455909729,
    "timestamp": "2025-07-14T12:33:52.886319"
  },
  {
    "epoch": 548,
    "batch_idx": 64,
    "loss": 0.26918888092041016,
    "learning_rate": 0.0008259168626504394,
    "timestep_mean": 391.21875,
    "epsilon_norm": 6.351973533630371,
    "prediction_norm": 4.916591167449951,
    "timestamp": "2025-07-14T12:33:52.891216"
  },
  {
    "epoch": 548,
    "batch_idx": 96,
    "loss": 0.26495832204818726,
    "learning_rate": 0.0008259168626504394,
    "timestep_mean": 459.34375,
    "epsilon_norm": 6.982990741729736,
    "prediction_norm": 5.953994274139404,
    "timestamp": "2025-07-14T12:33:52.896272"
  },
  {
    "epoch": 548,
    "batch_idx": 128,
    "loss": 0.3918512165546417,
    "learning_rate": 0.0008259168626504394,
    "timestep_mean": 481.75,
    "epsilon_norm": 8.197420120239258,
    "prediction_norm": 6.235473155975342,
    "timestamp": "2025-07-14T12:33:52.901147"
  },
  {
    "epoch": 548,
    "batch_idx": 160,
    "loss": 0.310811847448349,
    "learning_rate": 0.0008259168626504394,
    "timestep_mean": 484.8125,
    "epsilon_norm": 8.18880844116211,
    "prediction_norm": 6.730571746826172,
    "timestamp": "2025-07-14T12:33:52.906094"
  },
  {
    "epoch": 548,
    "batch_idx": 192,
    "loss": 0.04756491258740425,
    "learning_rate": 0.0008259168626504394,
    "timestep_mean": 676.375,
    "epsilon_norm": 2.3360073566436768,
    "prediction_norm": 2.428274154663086,
    "timestamp": "2025-07-14T12:33:52.910970"
  },
  {
    "epoch": 549,
    "batch_idx": 0,
    "loss": 0.5296825170516968,
    "learning_rate": 0.0008253208447561882,
    "timestep_mean": 484.8125,
    "epsilon_norm": 8.731006622314453,
    "prediction_norm": 6.66266393661499,
    "timestamp": "2025-07-14T12:33:52.915877"
  },
  {
    "epoch": 549,
    "batch_idx": 32,
    "loss": 0.33706212043762207,
    "learning_rate": 0.0008253208447561882,
    "timestep_mean": 437.84375,
    "epsilon_norm": 8.099559783935547,
    "prediction_norm": 6.936229228973389,
    "timestamp": "2025-07-14T12:33:52.920745"
  },
  {
    "epoch": 549,
    "batch_idx": 64,
    "loss": 0.24512989819049835,
    "learning_rate": 0.0008253208447561882,
    "timestep_mean": 590.90625,
    "epsilon_norm": 8.960221290588379,
    "prediction_norm": 7.895504951477051,
    "timestamp": "2025-07-14T12:33:52.925594"
  },
  {
    "epoch": 549,
    "batch_idx": 96,
    "loss": 0.2877141833305359,
    "learning_rate": 0.0008253208447561882,
    "timestep_mean": 472.71875,
    "epsilon_norm": 6.897711277008057,
    "prediction_norm": 5.819803714752197,
    "timestamp": "2025-07-14T12:33:52.930456"
  },
  {
    "epoch": 549,
    "batch_idx": 128,
    "loss": 0.22922009229660034,
    "learning_rate": 0.0008253208447561882,
    "timestep_mean": 570.6875,
    "epsilon_norm": 7.340730667114258,
    "prediction_norm": 6.320517063140869,
    "timestamp": "2025-07-14T12:33:52.935369"
  },
  {
    "epoch": 549,
    "batch_idx": 160,
    "loss": 0.4199260175228119,
    "learning_rate": 0.0008253208447561882,
    "timestep_mean": 405.28125,
    "epsilon_norm": 7.0504150390625,
    "prediction_norm": 5.736003875732422,
    "timestamp": "2025-07-14T12:33:52.940314"
  },
  {
    "epoch": 549,
    "batch_idx": 192,
    "loss": 0.6405441761016846,
    "learning_rate": 0.0008253208447561882,
    "timestep_mean": 307.5,
    "epsilon_norm": 4.020500659942627,
    "prediction_norm": 2.9303126335144043,
    "timestamp": "2025-07-14T12:33:52.945311"
  },
  {
    "epoch": 550,
    "batch_idx": 0,
    "loss": 0.4354551136493683,
    "learning_rate": 0.0008247240241650919,
    "timestep_mean": 560.59375,
    "epsilon_norm": 9.262311935424805,
    "prediction_norm": 7.262179374694824,
    "timestamp": "2025-07-14T12:33:52.950431"
  },
  {
    "epoch": 550,
    "batch_idx": 32,
    "loss": 0.29573529958724976,
    "learning_rate": 0.0008247240241650919,
    "timestep_mean": 519.28125,
    "epsilon_norm": 8.06901741027832,
    "prediction_norm": 6.618895530700684,
    "timestamp": "2025-07-14T12:33:52.955516"
  },
  {
    "epoch": 550,
    "batch_idx": 64,
    "loss": 0.1922341138124466,
    "learning_rate": 0.0008247240241650919,
    "timestep_mean": 561.375,
    "epsilon_norm": 8.506338119506836,
    "prediction_norm": 7.043019771575928,
    "timestamp": "2025-07-14T12:33:52.960567"
  },
  {
    "epoch": 550,
    "batch_idx": 96,
    "loss": 0.5699830055236816,
    "learning_rate": 0.0008247240241650919,
    "timestep_mean": 494.90625,
    "epsilon_norm": 8.056102752685547,
    "prediction_norm": 5.865272521972656,
    "timestamp": "2025-07-14T12:33:52.965653"
  },
  {
    "epoch": 550,
    "batch_idx": 128,
    "loss": 0.22509117424488068,
    "learning_rate": 0.0008247240241650919,
    "timestep_mean": 520.6875,
    "epsilon_norm": 7.642202377319336,
    "prediction_norm": 6.815115451812744,
    "timestamp": "2025-07-14T12:33:52.970800"
  },
  {
    "epoch": 550,
    "batch_idx": 160,
    "loss": 0.3021514117717743,
    "learning_rate": 0.0008247240241650919,
    "timestep_mean": 488.5,
    "epsilon_norm": 7.637898921966553,
    "prediction_norm": 5.701916694641113,
    "timestamp": "2025-07-14T12:33:52.975866"
  },
  {
    "epoch": 550,
    "batch_idx": 192,
    "loss": 0.24488796293735504,
    "learning_rate": 0.0008247240241650919,
    "timestep_mean": 474.375,
    "epsilon_norm": 3.377638101577759,
    "prediction_norm": 2.3648221492767334,
    "timestamp": "2025-07-14T12:33:52.981085"
  },
  {
    "epoch": 551,
    "batch_idx": 0,
    "loss": 0.16528239846229553,
    "learning_rate": 0.0008241264023497458,
    "timestep_mean": 532.5625,
    "epsilon_norm": 7.115818023681641,
    "prediction_norm": 6.248463153839111,
    "timestamp": "2025-07-14T12:33:52.986278"
  },
  {
    "epoch": 551,
    "batch_idx": 32,
    "loss": 0.2281767725944519,
    "learning_rate": 0.0008241264023497458,
    "timestep_mean": 542.9375,
    "epsilon_norm": 7.301271438598633,
    "prediction_norm": 6.496294975280762,
    "timestamp": "2025-07-14T12:33:52.991578"
  },
  {
    "epoch": 551,
    "batch_idx": 64,
    "loss": 0.20148798823356628,
    "learning_rate": 0.0008241264023497458,
    "timestep_mean": 589.96875,
    "epsilon_norm": 6.586304664611816,
    "prediction_norm": 5.5469865798950195,
    "timestamp": "2025-07-14T12:33:52.996933"
  },
  {
    "epoch": 551,
    "batch_idx": 96,
    "loss": 0.22032448649406433,
    "learning_rate": 0.0008241264023497458,
    "timestep_mean": 499.5625,
    "epsilon_norm": 6.776030540466309,
    "prediction_norm": 5.73349666595459,
    "timestamp": "2025-07-14T12:33:53.002163"
  },
  {
    "epoch": 551,
    "batch_idx": 128,
    "loss": 0.2885952889919281,
    "learning_rate": 0.0008241264023497458,
    "timestep_mean": 506.40625,
    "epsilon_norm": 7.562016010284424,
    "prediction_norm": 5.667175769805908,
    "timestamp": "2025-07-14T12:33:53.007477"
  },
  {
    "epoch": 551,
    "batch_idx": 160,
    "loss": 0.15782803297042847,
    "learning_rate": 0.0008241264023497458,
    "timestep_mean": 558.53125,
    "epsilon_norm": 7.982126712799072,
    "prediction_norm": 7.504553318023682,
    "timestamp": "2025-07-14T12:33:53.012734"
  },
  {
    "epoch": 551,
    "batch_idx": 192,
    "loss": 0.3821827471256256,
    "learning_rate": 0.0008241264023497458,
    "timestep_mean": 447.625,
    "epsilon_norm": 3.261448621749878,
    "prediction_norm": 2.299071788787842,
    "timestamp": "2025-07-14T12:33:53.017992"
  },
  {
    "epoch": 552,
    "batch_idx": 0,
    "loss": 0.19987091422080994,
    "learning_rate": 0.0008235279807847223,
    "timestep_mean": 554.21875,
    "epsilon_norm": 9.065234184265137,
    "prediction_norm": 8.051752090454102,
    "timestamp": "2025-07-14T12:33:53.023225"
  },
  {
    "epoch": 552,
    "batch_idx": 32,
    "loss": 0.5136911869049072,
    "learning_rate": 0.0008235279807847223,
    "timestep_mean": 448.0625,
    "epsilon_norm": 8.676536560058594,
    "prediction_norm": 5.8251142501831055,
    "timestamp": "2025-07-14T12:33:53.028473"
  },
  {
    "epoch": 552,
    "batch_idx": 64,
    "loss": 0.3698459267616272,
    "learning_rate": 0.0008235279807847223,
    "timestep_mean": 444.46875,
    "epsilon_norm": 7.6795125007629395,
    "prediction_norm": 5.860923767089844,
    "timestamp": "2025-07-14T12:33:53.033889"
  },
  {
    "epoch": 552,
    "batch_idx": 96,
    "loss": 0.2949379086494446,
    "learning_rate": 0.0008235279807847223,
    "timestep_mean": 489.4375,
    "epsilon_norm": 7.608392715454102,
    "prediction_norm": 5.95156717300415,
    "timestamp": "2025-07-14T12:33:53.039108"
  },
  {
    "epoch": 552,
    "batch_idx": 128,
    "loss": 0.24452951550483704,
    "learning_rate": 0.0008235279807847223,
    "timestep_mean": 494.96875,
    "epsilon_norm": 6.865818977355957,
    "prediction_norm": 5.564493179321289,
    "timestamp": "2025-07-14T12:33:53.044570"
  },
  {
    "epoch": 552,
    "batch_idx": 160,
    "loss": 0.31956636905670166,
    "learning_rate": 0.0008235279807847223,
    "timestep_mean": 589.1875,
    "epsilon_norm": 8.331421852111816,
    "prediction_norm": 6.775394439697266,
    "timestamp": "2025-07-14T12:33:53.049888"
  },
  {
    "epoch": 552,
    "batch_idx": 192,
    "loss": 0.40023207664489746,
    "learning_rate": 0.0008235279807847223,
    "timestep_mean": 517.875,
    "epsilon_norm": 3.9183349609375,
    "prediction_norm": 3.1092216968536377,
    "timestamp": "2025-07-14T12:33:53.055181"
  },
  {
    "epoch": 553,
    "batch_idx": 0,
    "loss": 0.09856810420751572,
    "learning_rate": 0.0008229287609465672,
    "timestep_mean": 588.21875,
    "epsilon_norm": 7.322017192840576,
    "prediction_norm": 7.045004367828369,
    "timestamp": "2025-07-14T12:33:53.060566"
  },
  {
    "epoch": 553,
    "batch_idx": 32,
    "loss": 0.18742501735687256,
    "learning_rate": 0.0008229287609465672,
    "timestep_mean": 501.75,
    "epsilon_norm": 8.41867733001709,
    "prediction_norm": 7.679063320159912,
    "timestamp": "2025-07-14T12:33:53.065718"
  },
  {
    "epoch": 553,
    "batch_idx": 64,
    "loss": 0.37587642669677734,
    "learning_rate": 0.0008229287609465672,
    "timestep_mean": 516.75,
    "epsilon_norm": 8.873506546020508,
    "prediction_norm": 7.010201930999756,
    "timestamp": "2025-07-14T12:33:53.070888"
  },
  {
    "epoch": 553,
    "batch_idx": 96,
    "loss": 0.35963284969329834,
    "learning_rate": 0.0008229287609465672,
    "timestep_mean": 492.71875,
    "epsilon_norm": 8.459373474121094,
    "prediction_norm": 7.030096054077148,
    "timestamp": "2025-07-14T12:33:53.076126"
  },
  {
    "epoch": 553,
    "batch_idx": 128,
    "loss": 0.346750944852829,
    "learning_rate": 0.0008229287609465672,
    "timestep_mean": 414.625,
    "epsilon_norm": 7.128200054168701,
    "prediction_norm": 5.234160900115967,
    "timestamp": "2025-07-14T12:33:53.081411"
  },
  {
    "epoch": 553,
    "batch_idx": 160,
    "loss": 0.18397748470306396,
    "learning_rate": 0.0008229287609465672,
    "timestep_mean": 501.625,
    "epsilon_norm": 6.700435638427734,
    "prediction_norm": 5.837625026702881,
    "timestamp": "2025-07-14T12:33:53.086707"
  },
  {
    "epoch": 553,
    "batch_idx": 192,
    "loss": 0.149656742811203,
    "learning_rate": 0.0008229287609465672,
    "timestep_mean": 545.0,
    "epsilon_norm": 3.549811840057373,
    "prediction_norm": 3.300983190536499,
    "timestamp": "2025-07-14T12:33:53.092061"
  },
  {
    "epoch": 554,
    "batch_idx": 0,
    "loss": 0.3314182758331299,
    "learning_rate": 0.0008223287443137958,
    "timestep_mean": 457.0,
    "epsilon_norm": 8.312447547912598,
    "prediction_norm": 6.646414279937744,
    "timestamp": "2025-07-14T12:33:53.097232"
  },
  {
    "epoch": 554,
    "batch_idx": 32,
    "loss": 0.1660267561674118,
    "learning_rate": 0.0008223287443137958,
    "timestep_mean": 499.65625,
    "epsilon_norm": 8.221845626831055,
    "prediction_norm": 7.69728422164917,
    "timestamp": "2025-07-14T12:33:53.102247"
  },
  {
    "epoch": 554,
    "batch_idx": 64,
    "loss": 0.3960540294647217,
    "learning_rate": 0.0008223287443137958,
    "timestep_mean": 477.71875,
    "epsilon_norm": 8.594257354736328,
    "prediction_norm": 6.6950249671936035,
    "timestamp": "2025-07-14T12:33:53.111469"
  },
  {
    "epoch": 554,
    "batch_idx": 96,
    "loss": 0.3699922561645508,
    "learning_rate": 0.0008223287443137958,
    "timestep_mean": 443.5625,
    "epsilon_norm": 8.764044761657715,
    "prediction_norm": 7.451245307922363,
    "timestamp": "2025-07-14T12:33:53.116920"
  },
  {
    "epoch": 554,
    "batch_idx": 128,
    "loss": 0.14981909096240997,
    "learning_rate": 0.0008223287443137958,
    "timestep_mean": 625.53125,
    "epsilon_norm": 8.839249610900879,
    "prediction_norm": 8.571807861328125,
    "timestamp": "2025-07-14T12:33:53.122109"
  },
  {
    "epoch": 554,
    "batch_idx": 160,
    "loss": 0.3673016428947449,
    "learning_rate": 0.0008223287443137958,
    "timestep_mean": 508.53125,
    "epsilon_norm": 7.947742462158203,
    "prediction_norm": 6.684020042419434,
    "timestamp": "2025-07-14T12:33:53.127326"
  },
  {
    "epoch": 554,
    "batch_idx": 192,
    "loss": 0.17029981315135956,
    "learning_rate": 0.0008223287443137958,
    "timestep_mean": 495.0,
    "epsilon_norm": 3.9368956089019775,
    "prediction_norm": 3.4983971118927,
    "timestamp": "2025-07-14T12:33:53.132601"
  },
  {
    "epoch": 555,
    "batch_idx": 0,
    "loss": 0.19403062760829926,
    "learning_rate": 0.0008217279323668896,
    "timestep_mean": 502.09375,
    "epsilon_norm": 7.42411470413208,
    "prediction_norm": 7.577823638916016,
    "timestamp": "2025-07-14T12:33:53.138031"
  },
  {
    "epoch": 555,
    "batch_idx": 32,
    "loss": 0.23807057738304138,
    "learning_rate": 0.0008217279323668896,
    "timestep_mean": 546.03125,
    "epsilon_norm": 6.842319965362549,
    "prediction_norm": 5.777703762054443,
    "timestamp": "2025-07-14T12:33:53.143244"
  },
  {
    "epoch": 555,
    "batch_idx": 64,
    "loss": 0.48980939388275146,
    "learning_rate": 0.0008217279323668896,
    "timestep_mean": 498.1875,
    "epsilon_norm": 8.021520614624023,
    "prediction_norm": 5.718495845794678,
    "timestamp": "2025-07-14T12:33:53.148584"
  },
  {
    "epoch": 555,
    "batch_idx": 96,
    "loss": 0.2510954737663269,
    "learning_rate": 0.0008217279323668896,
    "timestep_mean": 493.78125,
    "epsilon_norm": 7.540036678314209,
    "prediction_norm": 7.0595622062683105,
    "timestamp": "2025-07-14T12:33:53.153761"
  },
  {
    "epoch": 555,
    "batch_idx": 128,
    "loss": 0.3515564203262329,
    "learning_rate": 0.0008217279323668896,
    "timestep_mean": 521.1875,
    "epsilon_norm": 7.34415340423584,
    "prediction_norm": 5.830999374389648,
    "timestamp": "2025-07-14T12:33:53.158900"
  },
  {
    "epoch": 555,
    "batch_idx": 160,
    "loss": 0.38939547538757324,
    "learning_rate": 0.0008217279323668896,
    "timestep_mean": 406.15625,
    "epsilon_norm": 7.928624153137207,
    "prediction_norm": 5.775920867919922,
    "timestamp": "2025-07-14T12:33:53.164133"
  },
  {
    "epoch": 555,
    "batch_idx": 192,
    "loss": 0.30627185106277466,
    "learning_rate": 0.0008217279323668896,
    "timestep_mean": 528.75,
    "epsilon_norm": 3.6816279888153076,
    "prediction_norm": 3.6423401832580566,
    "timestamp": "2025-07-14T12:33:53.169379"
  },
  {
    "epoch": 556,
    "batch_idx": 0,
    "loss": 0.21952995657920837,
    "learning_rate": 0.0008211263265882924,
    "timestep_mean": 494.40625,
    "epsilon_norm": 8.343269348144531,
    "prediction_norm": 7.141726016998291,
    "timestamp": "2025-07-14T12:33:53.174687"
  },
  {
    "epoch": 556,
    "batch_idx": 32,
    "loss": 0.299471914768219,
    "learning_rate": 0.0008211263265882924,
    "timestep_mean": 490.84375,
    "epsilon_norm": 8.784005165100098,
    "prediction_norm": 7.332727432250977,
    "timestamp": "2025-07-14T12:33:53.179963"
  },
  {
    "epoch": 556,
    "batch_idx": 64,
    "loss": 0.2891981303691864,
    "learning_rate": 0.0008211263265882924,
    "timestep_mean": 546.6875,
    "epsilon_norm": 8.323959350585938,
    "prediction_norm": 6.959736347198486,
    "timestamp": "2025-07-14T12:33:53.185208"
  },
  {
    "epoch": 556,
    "batch_idx": 96,
    "loss": 0.2763429284095764,
    "learning_rate": 0.0008211263265882924,
    "timestep_mean": 582.0625,
    "epsilon_norm": 8.31722640991211,
    "prediction_norm": 6.843637466430664,
    "timestamp": "2025-07-14T12:33:53.190597"
  },
  {
    "epoch": 556,
    "batch_idx": 128,
    "loss": 0.2507368326187134,
    "learning_rate": 0.0008211263265882924,
    "timestep_mean": 436.28125,
    "epsilon_norm": 7.9641523361206055,
    "prediction_norm": 6.367751121520996,
    "timestamp": "2025-07-14T12:33:53.195737"
  },
  {
    "epoch": 556,
    "batch_idx": 160,
    "loss": 0.28013020753860474,
    "learning_rate": 0.0008211263265882924,
    "timestep_mean": 484.0625,
    "epsilon_norm": 8.472195625305176,
    "prediction_norm": 6.684982776641846,
    "timestamp": "2025-07-14T12:33:53.201005"
  },
  {
    "epoch": 556,
    "batch_idx": 192,
    "loss": 0.0514591708779335,
    "learning_rate": 0.0008211263265882924,
    "timestep_mean": 510.875,
    "epsilon_norm": 3.233576536178589,
    "prediction_norm": 3.283658027648926,
    "timestamp": "2025-07-14T12:33:53.206244"
  },
  {
    "epoch": 557,
    "batch_idx": 0,
    "loss": 0.33520227670669556,
    "learning_rate": 0.0008205239284624063,
    "timestep_mean": 458.21875,
    "epsilon_norm": 7.764044284820557,
    "prediction_norm": 6.236258029937744,
    "timestamp": "2025-07-14T12:33:53.211454"
  },
  {
    "epoch": 557,
    "batch_idx": 32,
    "loss": 0.2372879981994629,
    "learning_rate": 0.0008205239284624063,
    "timestep_mean": 617.90625,
    "epsilon_norm": 7.725386619567871,
    "prediction_norm": 5.986668586730957,
    "timestamp": "2025-07-14T12:33:53.216837"
  },
  {
    "epoch": 557,
    "batch_idx": 64,
    "loss": 0.5859376192092896,
    "learning_rate": 0.0008205239284624063,
    "timestep_mean": 396.0,
    "epsilon_norm": 9.56737232208252,
    "prediction_norm": 6.675210475921631,
    "timestamp": "2025-07-14T12:33:53.222091"
  },
  {
    "epoch": 557,
    "batch_idx": 96,
    "loss": 0.23561206459999084,
    "learning_rate": 0.0008205239284624063,
    "timestep_mean": 567.75,
    "epsilon_norm": 7.391024589538574,
    "prediction_norm": 6.108021259307861,
    "timestamp": "2025-07-14T12:33:53.227076"
  },
  {
    "epoch": 557,
    "batch_idx": 128,
    "loss": 0.29712963104248047,
    "learning_rate": 0.0008205239284624063,
    "timestep_mean": 465.15625,
    "epsilon_norm": 7.481122016906738,
    "prediction_norm": 6.6050920486450195,
    "timestamp": "2025-07-14T12:33:53.232222"
  },
  {
    "epoch": 557,
    "batch_idx": 160,
    "loss": 0.25368526577949524,
    "learning_rate": 0.0008205239284624063,
    "timestep_mean": 477.75,
    "epsilon_norm": 7.082399845123291,
    "prediction_norm": 5.712671279907227,
    "timestamp": "2025-07-14T12:33:53.237414"
  },
  {
    "epoch": 557,
    "batch_idx": 192,
    "loss": 0.7576262950897217,
    "learning_rate": 0.0008205239284624063,
    "timestep_mean": 498.0,
    "epsilon_norm": 4.669229984283447,
    "prediction_norm": 2.5337436199188232,
    "timestamp": "2025-07-14T12:33:53.242568"
  },
  {
    "epoch": 558,
    "batch_idx": 0,
    "loss": 0.5338668823242188,
    "learning_rate": 0.0008199207394755894,
    "timestep_mean": 459.96875,
    "epsilon_norm": 9.459280014038086,
    "prediction_norm": 7.500687599182129,
    "timestamp": "2025-07-14T12:33:53.247725"
  },
  {
    "epoch": 558,
    "batch_idx": 32,
    "loss": 0.2645385265350342,
    "learning_rate": 0.0008199207394755894,
    "timestep_mean": 559.0,
    "epsilon_norm": 8.41883373260498,
    "prediction_norm": 7.20559549331665,
    "timestamp": "2025-07-14T12:33:53.252960"
  },
  {
    "epoch": 558,
    "batch_idx": 64,
    "loss": 0.17915236949920654,
    "learning_rate": 0.0008199207394755894,
    "timestep_mean": 498.46875,
    "epsilon_norm": 8.055355072021484,
    "prediction_norm": 7.041849136352539,
    "timestamp": "2025-07-14T12:33:53.258235"
  },
  {
    "epoch": 558,
    "batch_idx": 96,
    "loss": 0.2540016174316406,
    "learning_rate": 0.0008199207394755894,
    "timestep_mean": 572.34375,
    "epsilon_norm": 8.595242500305176,
    "prediction_norm": 7.860304355621338,
    "timestamp": "2025-07-14T12:33:53.263450"
  },
  {
    "epoch": 558,
    "batch_idx": 128,
    "loss": 0.31438493728637695,
    "learning_rate": 0.0008199207394755894,
    "timestep_mean": 523.03125,
    "epsilon_norm": 7.6119489669799805,
    "prediction_norm": 6.964836120605469,
    "timestamp": "2025-07-14T12:33:53.268729"
  },
  {
    "epoch": 558,
    "batch_idx": 160,
    "loss": 0.47173649072647095,
    "learning_rate": 0.0008199207394755894,
    "timestep_mean": 466.6875,
    "epsilon_norm": 8.936506271362305,
    "prediction_norm": 7.140841484069824,
    "timestamp": "2025-07-14T12:33:53.273955"
  },
  {
    "epoch": 558,
    "batch_idx": 192,
    "loss": 0.36690640449523926,
    "learning_rate": 0.0008199207394755894,
    "timestep_mean": 515.875,
    "epsilon_norm": 4.812328338623047,
    "prediction_norm": 3.4931561946868896,
    "timestamp": "2025-07-14T12:33:53.279128"
  },
  {
    "epoch": 559,
    "batch_idx": 0,
    "loss": 0.4125440716743469,
    "learning_rate": 0.0008193167611161502,
    "timestep_mean": 484.34375,
    "epsilon_norm": 8.462117195129395,
    "prediction_norm": 6.246316432952881,
    "timestamp": "2025-07-14T12:33:53.284366"
  },
  {
    "epoch": 559,
    "batch_idx": 32,
    "loss": 0.2049105167388916,
    "learning_rate": 0.0008193167611161502,
    "timestep_mean": 544.5,
    "epsilon_norm": 7.847578525543213,
    "prediction_norm": 7.487967491149902,
    "timestamp": "2025-07-14T12:33:53.289369"
  },
  {
    "epoch": 559,
    "batch_idx": 64,
    "loss": 0.4119790494441986,
    "learning_rate": 0.0008193167611161502,
    "timestep_mean": 528.78125,
    "epsilon_norm": 8.976351737976074,
    "prediction_norm": 7.922760009765625,
    "timestamp": "2025-07-14T12:33:53.294500"
  },
  {
    "epoch": 559,
    "batch_idx": 96,
    "loss": 0.20897378027439117,
    "learning_rate": 0.0008193167611161502,
    "timestep_mean": 493.625,
    "epsilon_norm": 8.169455528259277,
    "prediction_norm": 7.399107933044434,
    "timestamp": "2025-07-14T12:33:53.299679"
  },
  {
    "epoch": 559,
    "batch_idx": 128,
    "loss": 0.1940748393535614,
    "learning_rate": 0.0008193167611161502,
    "timestep_mean": 534.21875,
    "epsilon_norm": 8.272628784179688,
    "prediction_norm": 7.8007659912109375,
    "timestamp": "2025-07-14T12:33:53.304864"
  },
  {
    "epoch": 559,
    "batch_idx": 160,
    "loss": 0.2605684995651245,
    "learning_rate": 0.0008193167611161502,
    "timestep_mean": 429.90625,
    "epsilon_norm": 7.020839691162109,
    "prediction_norm": 6.476832866668701,
    "timestamp": "2025-07-14T12:33:53.309986"
  },
  {
    "epoch": 559,
    "batch_idx": 192,
    "loss": 0.43847090005874634,
    "learning_rate": 0.0008193167611161502,
    "timestep_mean": 569.25,
    "epsilon_norm": 4.622138500213623,
    "prediction_norm": 3.168020486831665,
    "timestamp": "2025-07-14T12:33:53.315137"
  },
  {
    "epoch": 560,
    "batch_idx": 0,
    "loss": 0.3261224031448364,
    "learning_rate": 0.0008187119948743451,
    "timestep_mean": 477.9375,
    "epsilon_norm": 8.541855812072754,
    "prediction_norm": 7.659853458404541,
    "timestamp": "2025-07-14T12:33:53.320270"
  },
  {
    "epoch": 560,
    "batch_idx": 32,
    "loss": 0.27163374423980713,
    "learning_rate": 0.0008187119948743451,
    "timestep_mean": 487.4375,
    "epsilon_norm": 7.653929710388184,
    "prediction_norm": 6.973300933837891,
    "timestamp": "2025-07-14T12:33:53.325418"
  },
  {
    "epoch": 560,
    "batch_idx": 64,
    "loss": 0.17759354412555695,
    "learning_rate": 0.0008187119948743451,
    "timestep_mean": 512.03125,
    "epsilon_norm": 8.198724746704102,
    "prediction_norm": 7.668247222900391,
    "timestamp": "2025-07-14T12:33:53.330507"
  },
  {
    "epoch": 560,
    "batch_idx": 96,
    "loss": 0.5130598545074463,
    "learning_rate": 0.0008187119948743451,
    "timestep_mean": 417.75,
    "epsilon_norm": 9.38175106048584,
    "prediction_norm": 6.730436325073242,
    "timestamp": "2025-07-14T12:33:53.335493"
  },
  {
    "epoch": 560,
    "batch_idx": 128,
    "loss": 0.2942969799041748,
    "learning_rate": 0.0008187119948743451,
    "timestep_mean": 557.78125,
    "epsilon_norm": 9.553853034973145,
    "prediction_norm": 7.531108856201172,
    "timestamp": "2025-07-14T12:33:53.340702"
  },
  {
    "epoch": 560,
    "batch_idx": 160,
    "loss": 0.40652114152908325,
    "learning_rate": 0.0008187119948743451,
    "timestep_mean": 500.34375,
    "epsilon_norm": 8.679854393005371,
    "prediction_norm": 7.201955795288086,
    "timestamp": "2025-07-14T12:33:53.345806"
  },
  {
    "epoch": 560,
    "batch_idx": 192,
    "loss": 0.45879387855529785,
    "learning_rate": 0.0008187119948743451,
    "timestep_mean": 431.5,
    "epsilon_norm": 3.776095151901245,
    "prediction_norm": 2.926165819168091,
    "timestamp": "2025-07-14T12:33:53.350965"
  },
  {
    "epoch": 561,
    "batch_idx": 0,
    "loss": 0.5138311386108398,
    "learning_rate": 0.0008181064422423749,
    "timestep_mean": 443.4375,
    "epsilon_norm": 8.473165512084961,
    "prediction_norm": 6.4726881980896,
    "timestamp": "2025-07-14T12:33:53.356141"
  },
  {
    "epoch": 561,
    "batch_idx": 32,
    "loss": 0.16548475623130798,
    "learning_rate": 0.0008181064422423749,
    "timestep_mean": 504.875,
    "epsilon_norm": 9.321707725524902,
    "prediction_norm": 8.356208801269531,
    "timestamp": "2025-07-14T12:33:53.361275"
  },
  {
    "epoch": 561,
    "batch_idx": 64,
    "loss": 0.34029361605644226,
    "learning_rate": 0.0008181064422423749,
    "timestep_mean": 441.09375,
    "epsilon_norm": 7.71460485458374,
    "prediction_norm": 6.40395450592041,
    "timestamp": "2025-07-14T12:33:53.366592"
  },
  {
    "epoch": 561,
    "batch_idx": 96,
    "loss": 0.3714491128921509,
    "learning_rate": 0.0008181064422423749,
    "timestep_mean": 396.03125,
    "epsilon_norm": 6.949014663696289,
    "prediction_norm": 4.740902900695801,
    "timestamp": "2025-07-14T12:33:53.371684"
  },
  {
    "epoch": 561,
    "batch_idx": 128,
    "loss": 0.31793147325515747,
    "learning_rate": 0.0008181064422423749,
    "timestep_mean": 487.3125,
    "epsilon_norm": 8.805159568786621,
    "prediction_norm": 7.008829116821289,
    "timestamp": "2025-07-14T12:33:53.376864"
  },
  {
    "epoch": 561,
    "batch_idx": 160,
    "loss": 0.3853030502796173,
    "learning_rate": 0.0008181064422423749,
    "timestep_mean": 551.28125,
    "epsilon_norm": 9.049766540527344,
    "prediction_norm": 7.498266696929932,
    "timestamp": "2025-07-14T12:33:53.381915"
  },
  {
    "epoch": 561,
    "batch_idx": 192,
    "loss": 0.25956180691719055,
    "learning_rate": 0.0008181064422423749,
    "timestep_mean": 444.0,
    "epsilon_norm": 3.803403854370117,
    "prediction_norm": 2.771862030029297,
    "timestamp": "2025-07-14T12:33:53.387050"
  },
  {
    "epoch": 562,
    "batch_idx": 0,
    "loss": 0.24130156636238098,
    "learning_rate": 0.0008175001047143805,
    "timestep_mean": 459.625,
    "epsilon_norm": 7.551267623901367,
    "prediction_norm": 6.570672035217285,
    "timestamp": "2025-07-14T12:33:53.392076"
  },
  {
    "epoch": 562,
    "batch_idx": 32,
    "loss": 0.37179315090179443,
    "learning_rate": 0.0008175001047143805,
    "timestep_mean": 400.625,
    "epsilon_norm": 7.908262729644775,
    "prediction_norm": 5.674473285675049,
    "timestamp": "2025-07-14T12:33:53.397786"
  },
  {
    "epoch": 562,
    "batch_idx": 64,
    "loss": 0.20359282195568085,
    "learning_rate": 0.0008175001047143805,
    "timestep_mean": 556.5625,
    "epsilon_norm": 8.099939346313477,
    "prediction_norm": 7.613663673400879,
    "timestamp": "2025-07-14T12:33:53.403461"
  },
  {
    "epoch": 562,
    "batch_idx": 96,
    "loss": 0.4063378572463989,
    "learning_rate": 0.0008175001047143805,
    "timestep_mean": 564.28125,
    "epsilon_norm": 8.542601585388184,
    "prediction_norm": 7.843437194824219,
    "timestamp": "2025-07-14T12:33:53.408795"
  },
  {
    "epoch": 562,
    "batch_idx": 128,
    "loss": 0.23664724826812744,
    "learning_rate": 0.0008175001047143805,
    "timestep_mean": 387.75,
    "epsilon_norm": 7.628000736236572,
    "prediction_norm": 7.147702693939209,
    "timestamp": "2025-07-14T12:33:53.413886"
  },
  {
    "epoch": 562,
    "batch_idx": 160,
    "loss": 0.6466197371482849,
    "learning_rate": 0.0008175001047143805,
    "timestep_mean": 407.875,
    "epsilon_norm": 8.327065467834473,
    "prediction_norm": 5.477272033691406,
    "timestamp": "2025-07-14T12:33:53.419566"
  },
  {
    "epoch": 562,
    "batch_idx": 192,
    "loss": 0.13147494196891785,
    "learning_rate": 0.0008175001047143805,
    "timestep_mean": 484.75,
    "epsilon_norm": 3.705357074737549,
    "prediction_norm": 3.8326570987701416,
    "timestamp": "2025-07-14T12:33:53.424934"
  },
  {
    "epoch": 563,
    "batch_idx": 0,
    "loss": 0.2129727005958557,
    "learning_rate": 0.0008168929837864394,
    "timestep_mean": 486.84375,
    "epsilon_norm": 7.668203830718994,
    "prediction_norm": 7.139510631561279,
    "timestamp": "2025-07-14T12:33:53.431085"
  },
  {
    "epoch": 563,
    "batch_idx": 32,
    "loss": 0.21437150239944458,
    "learning_rate": 0.0008168929837864394,
    "timestep_mean": 499.59375,
    "epsilon_norm": 7.806020259857178,
    "prediction_norm": 7.45466423034668,
    "timestamp": "2025-07-14T12:33:53.437184"
  },
  {
    "epoch": 563,
    "batch_idx": 64,
    "loss": 0.3224414587020874,
    "learning_rate": 0.0008168929837864394,
    "timestep_mean": 535.53125,
    "epsilon_norm": 7.6375041007995605,
    "prediction_norm": 6.035741329193115,
    "timestamp": "2025-07-14T12:33:53.442963"
  },
  {
    "epoch": 563,
    "batch_idx": 96,
    "loss": 0.33746016025543213,
    "learning_rate": 0.0008168929837864394,
    "timestep_mean": 455.34375,
    "epsilon_norm": 9.377824783325195,
    "prediction_norm": 7.2925591468811035,
    "timestamp": "2025-07-14T12:33:53.448347"
  },
  {
    "epoch": 563,
    "batch_idx": 128,
    "loss": 0.2544950842857361,
    "learning_rate": 0.0008168929837864394,
    "timestep_mean": 539.0625,
    "epsilon_norm": 7.393306732177734,
    "prediction_norm": 6.243963241577148,
    "timestamp": "2025-07-14T12:33:53.453559"
  },
  {
    "epoch": 563,
    "batch_idx": 160,
    "loss": 0.35675060749053955,
    "learning_rate": 0.0008168929837864394,
    "timestep_mean": 454.875,
    "epsilon_norm": 8.656366348266602,
    "prediction_norm": 6.768298149108887,
    "timestamp": "2025-07-14T12:33:53.458934"
  },
  {
    "epoch": 563,
    "batch_idx": 192,
    "loss": 0.2544415593147278,
    "learning_rate": 0.0008168929837864394,
    "timestep_mean": 649.25,
    "epsilon_norm": 4.208225727081299,
    "prediction_norm": 3.3561863899230957,
    "timestamp": "2025-07-14T12:33:53.464273"
  },
  {
    "epoch": 564,
    "batch_idx": 0,
    "loss": 0.419672429561615,
    "learning_rate": 0.0008162850809565622,
    "timestep_mean": 440.1875,
    "epsilon_norm": 8.068584442138672,
    "prediction_norm": 5.979682445526123,
    "timestamp": "2025-07-14T12:33:53.469523"
  },
  {
    "epoch": 564,
    "batch_idx": 32,
    "loss": 0.26795583963394165,
    "learning_rate": 0.0008162850809565622,
    "timestep_mean": 458.75,
    "epsilon_norm": 7.881250381469727,
    "prediction_norm": 6.1212382316589355,
    "timestamp": "2025-07-14T12:33:53.475170"
  },
  {
    "epoch": 564,
    "batch_idx": 64,
    "loss": 0.23843777179718018,
    "learning_rate": 0.0008162850809565622,
    "timestep_mean": 493.125,
    "epsilon_norm": 8.293322563171387,
    "prediction_norm": 6.352236747741699,
    "timestamp": "2025-07-14T12:33:53.480443"
  },
  {
    "epoch": 564,
    "batch_idx": 96,
    "loss": 0.35230568051338196,
    "learning_rate": 0.0008162850809565622,
    "timestep_mean": 498.75,
    "epsilon_norm": 7.670251846313477,
    "prediction_norm": 6.118503093719482,
    "timestamp": "2025-07-14T12:33:53.485767"
  },
  {
    "epoch": 564,
    "batch_idx": 128,
    "loss": 0.2755669057369232,
    "learning_rate": 0.0008162850809565622,
    "timestep_mean": 483.53125,
    "epsilon_norm": 9.080896377563477,
    "prediction_norm": 7.752753257751465,
    "timestamp": "2025-07-14T12:33:53.491014"
  },
  {
    "epoch": 564,
    "batch_idx": 160,
    "loss": 0.2515285611152649,
    "learning_rate": 0.0008162850809565622,
    "timestep_mean": 500.3125,
    "epsilon_norm": 8.087906837463379,
    "prediction_norm": 6.840210914611816,
    "timestamp": "2025-07-14T12:33:53.496361"
  },
  {
    "epoch": 564,
    "batch_idx": 192,
    "loss": 0.14758741855621338,
    "learning_rate": 0.0008162850809565622,
    "timestep_mean": 544.375,
    "epsilon_norm": 3.5143308639526367,
    "prediction_norm": 3.8074395656585693,
    "timestamp": "2025-07-14T12:33:53.501614"
  },
  {
    "epoch": 565,
    "batch_idx": 0,
    "loss": 0.3076188564300537,
    "learning_rate": 0.0008156763977246888,
    "timestep_mean": 519.21875,
    "epsilon_norm": 7.599605083465576,
    "prediction_norm": 6.900252342224121,
    "timestamp": "2025-07-14T12:33:53.506854"
  },
  {
    "epoch": 565,
    "batch_idx": 32,
    "loss": 0.3666646480560303,
    "learning_rate": 0.0008156763977246888,
    "timestep_mean": 453.0,
    "epsilon_norm": 9.102126121520996,
    "prediction_norm": 7.60386848449707,
    "timestamp": "2025-07-14T12:33:53.512078"
  },
  {
    "epoch": 565,
    "batch_idx": 64,
    "loss": 0.39247098565101624,
    "learning_rate": 0.0008156763977246888,
    "timestep_mean": 486.28125,
    "epsilon_norm": 8.80671215057373,
    "prediction_norm": 7.372004508972168,
    "timestamp": "2025-07-14T12:33:53.517341"
  },
  {
    "epoch": 565,
    "batch_idx": 96,
    "loss": 0.34404513239860535,
    "learning_rate": 0.0008156763977246888,
    "timestep_mean": 546.53125,
    "epsilon_norm": 9.2333345413208,
    "prediction_norm": 7.785501480102539,
    "timestamp": "2025-07-14T12:33:53.522549"
  },
  {
    "epoch": 565,
    "batch_idx": 128,
    "loss": 0.23356255888938904,
    "learning_rate": 0.0008156763977246888,
    "timestep_mean": 538.90625,
    "epsilon_norm": 7.572683334350586,
    "prediction_norm": 7.270052433013916,
    "timestamp": "2025-07-14T12:33:53.527734"
  },
  {
    "epoch": 565,
    "batch_idx": 160,
    "loss": 0.508662760257721,
    "learning_rate": 0.0008156763977246888,
    "timestep_mean": 425.21875,
    "epsilon_norm": 8.17966079711914,
    "prediction_norm": 6.452948093414307,
    "timestamp": "2025-07-14T12:33:53.532946"
  },
  {
    "epoch": 565,
    "batch_idx": 192,
    "loss": 0.20904244482517242,
    "learning_rate": 0.0008156763977246888,
    "timestep_mean": 453.875,
    "epsilon_norm": 2.7646422386169434,
    "prediction_norm": 2.873318672180176,
    "timestamp": "2025-07-14T12:33:53.538361"
  },
  {
    "epoch": 566,
    "batch_idx": 0,
    "loss": 0.38783520460128784,
    "learning_rate": 0.0008150669355926845,
    "timestep_mean": 454.1875,
    "epsilon_norm": 8.50633716583252,
    "prediction_norm": 6.72434139251709,
    "timestamp": "2025-07-14T12:33:53.543572"
  },
  {
    "epoch": 566,
    "batch_idx": 32,
    "loss": 0.38444292545318604,
    "learning_rate": 0.0008150669355926845,
    "timestep_mean": 464.1875,
    "epsilon_norm": 7.54771089553833,
    "prediction_norm": 5.939326286315918,
    "timestamp": "2025-07-14T12:33:53.548681"
  },
  {
    "epoch": 566,
    "batch_idx": 64,
    "loss": 0.3799446225166321,
    "learning_rate": 0.0008150669355926845,
    "timestep_mean": 478.0,
    "epsilon_norm": 8.068451881408691,
    "prediction_norm": 7.003300189971924,
    "timestamp": "2025-07-14T12:33:53.553965"
  },
  {
    "epoch": 566,
    "batch_idx": 96,
    "loss": 0.2224498987197876,
    "learning_rate": 0.0008150669355926845,
    "timestep_mean": 439.65625,
    "epsilon_norm": 7.0586748123168945,
    "prediction_norm": 5.911192893981934,
    "timestamp": "2025-07-14T12:33:53.559116"
  },
  {
    "epoch": 566,
    "batch_idx": 128,
    "loss": 0.35441693663597107,
    "learning_rate": 0.0008150669355926845,
    "timestep_mean": 515.875,
    "epsilon_norm": 7.669662952423096,
    "prediction_norm": 6.462503910064697,
    "timestamp": "2025-07-14T12:33:53.564352"
  },
  {
    "epoch": 566,
    "batch_idx": 160,
    "loss": 0.49132394790649414,
    "learning_rate": 0.0008150669355926845,
    "timestep_mean": 429.78125,
    "epsilon_norm": 8.48075008392334,
    "prediction_norm": 6.064507007598877,
    "timestamp": "2025-07-14T12:33:53.569560"
  },
  {
    "epoch": 566,
    "batch_idx": 192,
    "loss": 0.1295202374458313,
    "learning_rate": 0.0008150669355926845,
    "timestep_mean": 532.375,
    "epsilon_norm": 3.1013033390045166,
    "prediction_norm": 2.8942954540252686,
    "timestamp": "2025-07-14T12:33:53.574680"
  },
  {
    "epoch": 567,
    "batch_idx": 0,
    "loss": 0.14914275705814362,
    "learning_rate": 0.0008144566960643366,
    "timestep_mean": 517.25,
    "epsilon_norm": 7.309062957763672,
    "prediction_norm": 7.117672443389893,
    "timestamp": "2025-07-14T12:33:53.580005"
  },
  {
    "epoch": 567,
    "batch_idx": 32,
    "loss": 0.4552580714225769,
    "learning_rate": 0.0008144566960643366,
    "timestep_mean": 523.5625,
    "epsilon_norm": 7.857323169708252,
    "prediction_norm": 5.891040802001953,
    "timestamp": "2025-07-14T12:33:53.585216"
  },
  {
    "epoch": 567,
    "batch_idx": 64,
    "loss": 0.2158692330121994,
    "learning_rate": 0.0008144566960643366,
    "timestep_mean": 549.4375,
    "epsilon_norm": 7.982527732849121,
    "prediction_norm": 6.466280460357666,
    "timestamp": "2025-07-14T12:33:53.590374"
  },
  {
    "epoch": 567,
    "batch_idx": 96,
    "loss": 0.2865809202194214,
    "learning_rate": 0.0008144566960643366,
    "timestep_mean": 451.71875,
    "epsilon_norm": 8.811307907104492,
    "prediction_norm": 7.81320858001709,
    "timestamp": "2025-07-14T12:33:53.595496"
  },
  {
    "epoch": 567,
    "batch_idx": 128,
    "loss": 0.4489961862564087,
    "learning_rate": 0.0008144566960643366,
    "timestep_mean": 539.1875,
    "epsilon_norm": 8.31196403503418,
    "prediction_norm": 6.252905368804932,
    "timestamp": "2025-07-14T12:33:53.600756"
  },
  {
    "epoch": 567,
    "batch_idx": 160,
    "loss": 0.2810174822807312,
    "learning_rate": 0.0008144566960643366,
    "timestep_mean": 558.84375,
    "epsilon_norm": 8.097382545471191,
    "prediction_norm": 6.950619697570801,
    "timestamp": "2025-07-14T12:33:53.605848"
  },
  {
    "epoch": 567,
    "batch_idx": 192,
    "loss": 0.7456563711166382,
    "learning_rate": 0.0008144566960643366,
    "timestep_mean": 449.125,
    "epsilon_norm": 5.4004669189453125,
    "prediction_norm": 4.133697986602783,
    "timestamp": "2025-07-14T12:33:53.611052"
  },
  {
    "epoch": 568,
    "batch_idx": 0,
    "loss": 0.4337823987007141,
    "learning_rate": 0.0008138456806453503,
    "timestep_mean": 409.0625,
    "epsilon_norm": 8.680498123168945,
    "prediction_norm": 6.440194129943848,
    "timestamp": "2025-07-14T12:33:53.616322"
  },
  {
    "epoch": 568,
    "batch_idx": 32,
    "loss": 0.3685101866722107,
    "learning_rate": 0.0008138456806453503,
    "timestep_mean": 481.0625,
    "epsilon_norm": 8.267621994018555,
    "prediction_norm": 6.66534423828125,
    "timestamp": "2025-07-14T12:33:53.621653"
  },
  {
    "epoch": 568,
    "batch_idx": 64,
    "loss": 0.5270348191261292,
    "learning_rate": 0.0008138456806453503,
    "timestep_mean": 452.59375,
    "epsilon_norm": 8.901652336120605,
    "prediction_norm": 6.888351917266846,
    "timestamp": "2025-07-14T12:33:53.626731"
  },
  {
    "epoch": 568,
    "batch_idx": 96,
    "loss": 0.39926570653915405,
    "learning_rate": 0.0008138456806453503,
    "timestep_mean": 491.46875,
    "epsilon_norm": 9.675747871398926,
    "prediction_norm": 7.865435600280762,
    "timestamp": "2025-07-14T12:33:53.631980"
  },
  {
    "epoch": 568,
    "batch_idx": 128,
    "loss": 0.3643171787261963,
    "learning_rate": 0.0008138456806453503,
    "timestep_mean": 540.5625,
    "epsilon_norm": 10.158541679382324,
    "prediction_norm": 8.185770988464355,
    "timestamp": "2025-07-14T12:33:53.637154"
  },
  {
    "epoch": 568,
    "batch_idx": 160,
    "loss": 0.3458923101425171,
    "learning_rate": 0.0008138456806453503,
    "timestep_mean": 503.9375,
    "epsilon_norm": 8.10109806060791,
    "prediction_norm": 6.694194793701172,
    "timestamp": "2025-07-14T12:33:53.642424"
  },
  {
    "epoch": 568,
    "batch_idx": 192,
    "loss": 0.11374012380838394,
    "learning_rate": 0.0008138456806453503,
    "timestep_mean": 462.375,
    "epsilon_norm": 3.584618091583252,
    "prediction_norm": 3.3243300914764404,
    "timestamp": "2025-07-14T12:33:53.647463"
  },
  {
    "epoch": 569,
    "batch_idx": 0,
    "loss": 0.38238054513931274,
    "learning_rate": 0.0008132338908433455,
    "timestep_mean": 462.84375,
    "epsilon_norm": 9.559161186218262,
    "prediction_norm": 7.193698883056641,
    "timestamp": "2025-07-14T12:33:53.652580"
  },
  {
    "epoch": 569,
    "batch_idx": 32,
    "loss": 0.22272741794586182,
    "learning_rate": 0.0008132338908433455,
    "timestep_mean": 444.75,
    "epsilon_norm": 6.3298211097717285,
    "prediction_norm": 4.871914863586426,
    "timestamp": "2025-07-14T12:33:53.657468"
  },
  {
    "epoch": 569,
    "batch_idx": 64,
    "loss": 0.3111111521720886,
    "learning_rate": 0.0008132338908433455,
    "timestep_mean": 516.96875,
    "epsilon_norm": 8.984405517578125,
    "prediction_norm": 8.247069358825684,
    "timestamp": "2025-07-14T12:33:53.662332"
  },
  {
    "epoch": 569,
    "batch_idx": 96,
    "loss": 0.2462380826473236,
    "learning_rate": 0.0008132338908433455,
    "timestep_mean": 515.3125,
    "epsilon_norm": 8.237253189086914,
    "prediction_norm": 7.902799606323242,
    "timestamp": "2025-07-14T12:33:53.667322"
  },
  {
    "epoch": 569,
    "batch_idx": 128,
    "loss": 0.30180490016937256,
    "learning_rate": 0.0008132338908433455,
    "timestep_mean": 511.9375,
    "epsilon_norm": 8.217087745666504,
    "prediction_norm": 7.654120445251465,
    "timestamp": "2025-07-14T12:33:53.672226"
  },
  {
    "epoch": 569,
    "batch_idx": 160,
    "loss": 0.242714986205101,
    "learning_rate": 0.0008132338908433455,
    "timestep_mean": 581.5,
    "epsilon_norm": 7.865047454833984,
    "prediction_norm": 6.756139278411865,
    "timestamp": "2025-07-14T12:33:53.677176"
  },
  {
    "epoch": 569,
    "batch_idx": 192,
    "loss": 0.31165263056755066,
    "learning_rate": 0.0008132338908433455,
    "timestep_mean": 537.875,
    "epsilon_norm": 3.5109338760375977,
    "prediction_norm": 2.7831764221191406,
    "timestamp": "2025-07-14T12:33:53.682061"
  },
  {
    "epoch": 570,
    "batch_idx": 0,
    "loss": 0.3144237697124481,
    "learning_rate": 0.0008126213281678527,
    "timestep_mean": 483.65625,
    "epsilon_norm": 8.876824378967285,
    "prediction_norm": 7.851962566375732,
    "timestamp": "2025-07-14T12:33:53.686961"
  },
  {
    "epoch": 570,
    "batch_idx": 32,
    "loss": 0.24112173914909363,
    "learning_rate": 0.0008126213281678527,
    "timestep_mean": 529.5,
    "epsilon_norm": 7.395845413208008,
    "prediction_norm": 6.44694709777832,
    "timestamp": "2025-07-14T12:33:53.691923"
  },
  {
    "epoch": 570,
    "batch_idx": 64,
    "loss": 0.26816970109939575,
    "learning_rate": 0.0008126213281678527,
    "timestep_mean": 527.5,
    "epsilon_norm": 7.710269927978516,
    "prediction_norm": 6.613955497741699,
    "timestamp": "2025-07-14T12:33:53.696821"
  },
  {
    "epoch": 570,
    "batch_idx": 96,
    "loss": 0.306934654712677,
    "learning_rate": 0.0008126213281678527,
    "timestep_mean": 579.15625,
    "epsilon_norm": 9.08157730102539,
    "prediction_norm": 7.408523082733154,
    "timestamp": "2025-07-14T12:33:53.701752"
  },
  {
    "epoch": 570,
    "batch_idx": 128,
    "loss": 0.2146984338760376,
    "learning_rate": 0.0008126213281678527,
    "timestep_mean": 515.3125,
    "epsilon_norm": 8.706578254699707,
    "prediction_norm": 7.352734565734863,
    "timestamp": "2025-07-14T12:33:53.711070"
  },
  {
    "epoch": 570,
    "batch_idx": 160,
    "loss": 0.24537886679172516,
    "learning_rate": 0.0008126213281678527,
    "timestep_mean": 485.21875,
    "epsilon_norm": 8.662443161010742,
    "prediction_norm": 6.860342979431152,
    "timestamp": "2025-07-14T12:33:53.715965"
  },
  {
    "epoch": 570,
    "batch_idx": 192,
    "loss": 0.415048748254776,
    "learning_rate": 0.0008126213281678527,
    "timestep_mean": 509.125,
    "epsilon_norm": 4.742568016052246,
    "prediction_norm": 3.4523491859436035,
    "timestamp": "2025-07-14T12:33:53.720937"
  },
  {
    "epoch": 571,
    "batch_idx": 0,
    "loss": 0.4150998890399933,
    "learning_rate": 0.0008120079941303094,
    "timestep_mean": 447.125,
    "epsilon_norm": 8.859380722045898,
    "prediction_norm": 6.414712429046631,
    "timestamp": "2025-07-14T12:33:53.725824"
  },
  {
    "epoch": 571,
    "batch_idx": 32,
    "loss": 0.2492794394493103,
    "learning_rate": 0.0008120079941303094,
    "timestep_mean": 423.84375,
    "epsilon_norm": 6.640286445617676,
    "prediction_norm": 5.094588756561279,
    "timestamp": "2025-07-14T12:33:53.730686"
  },
  {
    "epoch": 571,
    "batch_idx": 64,
    "loss": 0.2476186603307724,
    "learning_rate": 0.0008120079941303094,
    "timestep_mean": 428.28125,
    "epsilon_norm": 8.143290519714355,
    "prediction_norm": 7.038103103637695,
    "timestamp": "2025-07-14T12:33:53.735556"
  },
  {
    "epoch": 571,
    "batch_idx": 96,
    "loss": 0.41102439165115356,
    "learning_rate": 0.0008120079941303094,
    "timestep_mean": 485.375,
    "epsilon_norm": 8.146637916564941,
    "prediction_norm": 5.867094039916992,
    "timestamp": "2025-07-14T12:33:53.740388"
  },
  {
    "epoch": 571,
    "batch_idx": 128,
    "loss": 0.17765718698501587,
    "learning_rate": 0.0008120079941303094,
    "timestep_mean": 466.8125,
    "epsilon_norm": 7.908930778503418,
    "prediction_norm": 6.981315612792969,
    "timestamp": "2025-07-14T12:33:53.745261"
  },
  {
    "epoch": 571,
    "batch_idx": 160,
    "loss": 0.2084142416715622,
    "learning_rate": 0.0008120079941303094,
    "timestep_mean": 502.1875,
    "epsilon_norm": 6.811840534210205,
    "prediction_norm": 6.068652153015137,
    "timestamp": "2025-07-14T12:33:53.750236"
  },
  {
    "epoch": 571,
    "batch_idx": 192,
    "loss": 0.2737712562084198,
    "learning_rate": 0.0008120079941303094,
    "timestep_mean": 472.625,
    "epsilon_norm": 3.2503159046173096,
    "prediction_norm": 2.3594765663146973,
    "timestamp": "2025-07-14T12:33:53.755159"
  },
  {
    "epoch": 572,
    "batch_idx": 0,
    "loss": 0.16246716678142548,
    "learning_rate": 0.0008113938902440564,
    "timestep_mean": 519.0,
    "epsilon_norm": 7.616713523864746,
    "prediction_norm": 7.313882827758789,
    "timestamp": "2025-07-14T12:33:53.760054"
  },
  {
    "epoch": 572,
    "batch_idx": 32,
    "loss": 0.46035242080688477,
    "learning_rate": 0.0008113938902440564,
    "timestep_mean": 484.0625,
    "epsilon_norm": 8.448758125305176,
    "prediction_norm": 5.947676658630371,
    "timestamp": "2025-07-14T12:33:53.765023"
  },
  {
    "epoch": 572,
    "batch_idx": 64,
    "loss": 0.17401280999183655,
    "learning_rate": 0.0008113938902440564,
    "timestep_mean": 509.25,
    "epsilon_norm": 8.412418365478516,
    "prediction_norm": 7.675103664398193,
    "timestamp": "2025-07-14T12:33:53.769914"
  },
  {
    "epoch": 572,
    "batch_idx": 96,
    "loss": 0.24948793649673462,
    "learning_rate": 0.0008113938902440564,
    "timestep_mean": 485.78125,
    "epsilon_norm": 7.38577938079834,
    "prediction_norm": 7.149113178253174,
    "timestamp": "2025-07-14T12:33:53.774916"
  },
  {
    "epoch": 572,
    "batch_idx": 128,
    "loss": 0.3348791003227234,
    "learning_rate": 0.0008113938902440564,
    "timestep_mean": 599.0625,
    "epsilon_norm": 8.526751518249512,
    "prediction_norm": 7.793784141540527,
    "timestamp": "2025-07-14T12:33:53.779904"
  },
  {
    "epoch": 572,
    "batch_idx": 160,
    "loss": 0.2880650758743286,
    "learning_rate": 0.0008113938902440564,
    "timestep_mean": 483.5,
    "epsilon_norm": 8.86264705657959,
    "prediction_norm": 7.338775157928467,
    "timestamp": "2025-07-14T12:33:53.784826"
  },
  {
    "epoch": 572,
    "batch_idx": 192,
    "loss": 0.3822298049926758,
    "learning_rate": 0.0008113938902440564,
    "timestep_mean": 372.375,
    "epsilon_norm": 4.594961643218994,
    "prediction_norm": 2.703747272491455,
    "timestamp": "2025-07-14T12:33:53.789707"
  },
  {
    "epoch": 573,
    "batch_idx": 0,
    "loss": 0.425663560628891,
    "learning_rate": 0.000810779018024334,
    "timestep_mean": 403.125,
    "epsilon_norm": 7.476966857910156,
    "prediction_norm": 5.91892671585083,
    "timestamp": "2025-07-14T12:33:53.794559"
  },
  {
    "epoch": 573,
    "batch_idx": 32,
    "loss": 0.2736194133758545,
    "learning_rate": 0.000810779018024334,
    "timestep_mean": 556.25,
    "epsilon_norm": 8.218759536743164,
    "prediction_norm": 6.790764331817627,
    "timestamp": "2025-07-14T12:33:53.799502"
  },
  {
    "epoch": 573,
    "batch_idx": 64,
    "loss": 0.3014352321624756,
    "learning_rate": 0.000810779018024334,
    "timestep_mean": 503.875,
    "epsilon_norm": 7.425297737121582,
    "prediction_norm": 6.668481826782227,
    "timestamp": "2025-07-14T12:33:53.804471"
  },
  {
    "epoch": 573,
    "batch_idx": 96,
    "loss": 0.2856751084327698,
    "learning_rate": 0.000810779018024334,
    "timestep_mean": 470.78125,
    "epsilon_norm": 7.661370277404785,
    "prediction_norm": 6.737109661102295,
    "timestamp": "2025-07-14T12:33:53.809314"
  },
  {
    "epoch": 573,
    "batch_idx": 128,
    "loss": 0.2622523009777069,
    "learning_rate": 0.000810779018024334,
    "timestep_mean": 516.375,
    "epsilon_norm": 7.768551349639893,
    "prediction_norm": 6.704810619354248,
    "timestamp": "2025-07-14T12:33:53.814218"
  },
  {
    "epoch": 573,
    "batch_idx": 160,
    "loss": 0.41143009066581726,
    "learning_rate": 0.000810779018024334,
    "timestep_mean": 474.125,
    "epsilon_norm": 8.962276458740234,
    "prediction_norm": 7.212262153625488,
    "timestamp": "2025-07-14T12:33:53.819136"
  },
  {
    "epoch": 573,
    "batch_idx": 192,
    "loss": 0.33399778604507446,
    "learning_rate": 0.000810779018024334,
    "timestep_mean": 365.375,
    "epsilon_norm": 3.621565103530884,
    "prediction_norm": 3.3252758979797363,
    "timestamp": "2025-07-14T12:33:53.824063"
  },
  {
    "epoch": 574,
    "batch_idx": 0,
    "loss": 0.18516570329666138,
    "learning_rate": 0.0008101633789882782,
    "timestep_mean": 503.09375,
    "epsilon_norm": 7.3999223709106445,
    "prediction_norm": 6.641232013702393,
    "timestamp": "2025-07-14T12:33:53.828951"
  },
  {
    "epoch": 574,
    "batch_idx": 32,
    "loss": 0.47458213567733765,
    "learning_rate": 0.0008101633789882782,
    "timestep_mean": 526.53125,
    "epsilon_norm": 8.465802192687988,
    "prediction_norm": 5.850622177124023,
    "timestamp": "2025-07-14T12:33:53.834081"
  },
  {
    "epoch": 574,
    "batch_idx": 64,
    "loss": 0.3638361692428589,
    "learning_rate": 0.0008101633789882782,
    "timestep_mean": 532.125,
    "epsilon_norm": 8.16311264038086,
    "prediction_norm": 6.738574981689453,
    "timestamp": "2025-07-14T12:33:53.839043"
  },
  {
    "epoch": 574,
    "batch_idx": 96,
    "loss": 0.2934033274650574,
    "learning_rate": 0.0008101633789882782,
    "timestep_mean": 494.90625,
    "epsilon_norm": 7.320430755615234,
    "prediction_norm": 6.409369945526123,
    "timestamp": "2025-07-14T12:33:53.843889"
  },
  {
    "epoch": 574,
    "batch_idx": 128,
    "loss": 0.44191402196884155,
    "learning_rate": 0.0008101633789882782,
    "timestep_mean": 508.5625,
    "epsilon_norm": 9.18226432800293,
    "prediction_norm": 7.8059892654418945,
    "timestamp": "2025-07-14T12:33:53.848749"
  },
  {
    "epoch": 574,
    "batch_idx": 160,
    "loss": 0.2792525291442871,
    "learning_rate": 0.0008101633789882782,
    "timestep_mean": 526.46875,
    "epsilon_norm": 7.156010150909424,
    "prediction_norm": 6.715934753417969,
    "timestamp": "2025-07-14T12:33:53.853582"
  },
  {
    "epoch": 574,
    "batch_idx": 192,
    "loss": 0.48221173882484436,
    "learning_rate": 0.0008101633789882782,
    "timestep_mean": 404.375,
    "epsilon_norm": 4.257048606872559,
    "prediction_norm": 3.1325736045837402,
    "timestamp": "2025-07-14T12:33:53.858414"
  },
  {
    "epoch": 575,
    "batch_idx": 0,
    "loss": 0.17818523943424225,
    "learning_rate": 0.0008095469746549172,
    "timestep_mean": 556.5625,
    "epsilon_norm": 7.399807929992676,
    "prediction_norm": 6.991397380828857,
    "timestamp": "2025-07-14T12:33:53.863354"
  },
  {
    "epoch": 575,
    "batch_idx": 32,
    "loss": 0.18815305829048157,
    "learning_rate": 0.0008095469746549172,
    "timestep_mean": 529.34375,
    "epsilon_norm": 8.492024421691895,
    "prediction_norm": 7.481624603271484,
    "timestamp": "2025-07-14T12:33:53.868228"
  },
  {
    "epoch": 575,
    "batch_idx": 64,
    "loss": 0.1901351511478424,
    "learning_rate": 0.0008095469746549172,
    "timestep_mean": 520.09375,
    "epsilon_norm": 7.965873718261719,
    "prediction_norm": 6.744185447692871,
    "timestamp": "2025-07-14T12:33:53.873106"
  },
  {
    "epoch": 575,
    "batch_idx": 96,
    "loss": 0.3602188229560852,
    "learning_rate": 0.0008095469746549172,
    "timestep_mean": 506.9375,
    "epsilon_norm": 8.932666778564453,
    "prediction_norm": 7.172345161437988,
    "timestamp": "2025-07-14T12:33:53.877990"
  },
  {
    "epoch": 575,
    "batch_idx": 128,
    "loss": 0.11771354079246521,
    "learning_rate": 0.0008095469746549172,
    "timestep_mean": 556.625,
    "epsilon_norm": 7.057221412658691,
    "prediction_norm": 5.84817361831665,
    "timestamp": "2025-07-14T12:33:53.882879"
  },
  {
    "epoch": 575,
    "batch_idx": 160,
    "loss": 0.2825498878955841,
    "learning_rate": 0.0008095469746549172,
    "timestep_mean": 483.40625,
    "epsilon_norm": 7.975479602813721,
    "prediction_norm": 6.352144241333008,
    "timestamp": "2025-07-14T12:33:53.888063"
  },
  {
    "epoch": 575,
    "batch_idx": 192,
    "loss": 0.0832783430814743,
    "learning_rate": 0.0008095469746549172,
    "timestep_mean": 767.625,
    "epsilon_norm": 3.655113935470581,
    "prediction_norm": 2.8197576999664307,
    "timestamp": "2025-07-14T12:33:53.893114"
  },
  {
    "epoch": 576,
    "batch_idx": 0,
    "loss": 0.3508465886116028,
    "learning_rate": 0.0008089298065451672,
    "timestep_mean": 429.0625,
    "epsilon_norm": 7.527942180633545,
    "prediction_norm": 6.2405829429626465,
    "timestamp": "2025-07-14T12:33:53.898284"
  },
  {
    "epoch": 576,
    "batch_idx": 32,
    "loss": 0.387658953666687,
    "learning_rate": 0.0008089298065451672,
    "timestep_mean": 403.96875,
    "epsilon_norm": 7.816605567932129,
    "prediction_norm": 5.822672367095947,
    "timestamp": "2025-07-14T12:33:53.903600"
  },
  {
    "epoch": 576,
    "batch_idx": 64,
    "loss": 0.24766665697097778,
    "learning_rate": 0.0008089298065451672,
    "timestep_mean": 495.6875,
    "epsilon_norm": 8.030773162841797,
    "prediction_norm": 7.406518459320068,
    "timestamp": "2025-07-14T12:33:53.908754"
  },
  {
    "epoch": 576,
    "batch_idx": 96,
    "loss": 0.19665414094924927,
    "learning_rate": 0.0008089298065451672,
    "timestep_mean": 438.0625,
    "epsilon_norm": 7.623734474182129,
    "prediction_norm": 7.245415687561035,
    "timestamp": "2025-07-14T12:33:53.913850"
  },
  {
    "epoch": 576,
    "batch_idx": 128,
    "loss": 0.261510968208313,
    "learning_rate": 0.0008089298065451672,
    "timestep_mean": 520.8125,
    "epsilon_norm": 7.82246208190918,
    "prediction_norm": 6.524365425109863,
    "timestamp": "2025-07-14T12:33:53.919018"
  },
  {
    "epoch": 576,
    "batch_idx": 160,
    "loss": 0.2521553039550781,
    "learning_rate": 0.0008089298065451672,
    "timestep_mean": 452.1875,
    "epsilon_norm": 6.9626784324646,
    "prediction_norm": 6.029189109802246,
    "timestamp": "2025-07-14T12:33:53.924232"
  },
  {
    "epoch": 576,
    "batch_idx": 192,
    "loss": 0.42764952778816223,
    "learning_rate": 0.0008089298065451672,
    "timestep_mean": 372.25,
    "epsilon_norm": 3.8728067874908447,
    "prediction_norm": 3.0341038703918457,
    "timestamp": "2025-07-14T12:33:53.929526"
  },
  {
    "epoch": 577,
    "batch_idx": 0,
    "loss": 0.2048494964838028,
    "learning_rate": 0.0008083118761818297,
    "timestep_mean": 593.15625,
    "epsilon_norm": 8.899614334106445,
    "prediction_norm": 8.889716148376465,
    "timestamp": "2025-07-14T12:33:53.934695"
  },
  {
    "epoch": 577,
    "batch_idx": 32,
    "loss": 0.27999210357666016,
    "learning_rate": 0.0008083118761818297,
    "timestep_mean": 519.0625,
    "epsilon_norm": 8.512027740478516,
    "prediction_norm": 7.5086541175842285,
    "timestamp": "2025-07-14T12:33:53.939879"
  },
  {
    "epoch": 577,
    "batch_idx": 64,
    "loss": 0.17462104558944702,
    "learning_rate": 0.0008083118761818297,
    "timestep_mean": 515.46875,
    "epsilon_norm": 7.125322341918945,
    "prediction_norm": 6.4466447830200195,
    "timestamp": "2025-07-14T12:33:53.945081"
  },
  {
    "epoch": 577,
    "batch_idx": 96,
    "loss": 0.35864686965942383,
    "learning_rate": 0.0008083118761818297,
    "timestep_mean": 488.78125,
    "epsilon_norm": 9.20193099975586,
    "prediction_norm": 8.121782302856445,
    "timestamp": "2025-07-14T12:33:53.950422"
  },
  {
    "epoch": 577,
    "batch_idx": 128,
    "loss": 0.18179479241371155,
    "learning_rate": 0.0008083118761818297,
    "timestep_mean": 499.9375,
    "epsilon_norm": 6.101785659790039,
    "prediction_norm": 5.304047584533691,
    "timestamp": "2025-07-14T12:33:53.955592"
  },
  {
    "epoch": 577,
    "batch_idx": 160,
    "loss": 0.3945467472076416,
    "learning_rate": 0.0008083118761818297,
    "timestep_mean": 491.3125,
    "epsilon_norm": 7.3812103271484375,
    "prediction_norm": 5.795259475708008,
    "timestamp": "2025-07-14T12:33:53.960831"
  },
  {
    "epoch": 577,
    "batch_idx": 192,
    "loss": 0.1906992495059967,
    "learning_rate": 0.0008083118761818297,
    "timestep_mean": 520.875,
    "epsilon_norm": 2.6922826766967773,
    "prediction_norm": 2.3690507411956787,
    "timestamp": "2025-07-14T12:33:53.966019"
  },
  {
    "epoch": 578,
    "batch_idx": 0,
    "loss": 0.22029834985733032,
    "learning_rate": 0.0008076931850895859,
    "timestep_mean": 538.375,
    "epsilon_norm": 7.47007942199707,
    "prediction_norm": 7.1098432540893555,
    "timestamp": "2025-07-14T12:33:53.971223"
  },
  {
    "epoch": 578,
    "batch_idx": 32,
    "loss": 0.4334854781627655,
    "learning_rate": 0.0008076931850895859,
    "timestep_mean": 470.53125,
    "epsilon_norm": 8.29016399383545,
    "prediction_norm": 6.801222801208496,
    "timestamp": "2025-07-14T12:33:53.976236"
  },
  {
    "epoch": 578,
    "batch_idx": 64,
    "loss": 0.176448792219162,
    "learning_rate": 0.0008076931850895859,
    "timestep_mean": 583.53125,
    "epsilon_norm": 6.086092472076416,
    "prediction_norm": 5.234414577484131,
    "timestamp": "2025-07-14T12:33:53.981342"
  },
  {
    "epoch": 578,
    "batch_idx": 96,
    "loss": 0.40382498502731323,
    "learning_rate": 0.0008076931850895859,
    "timestep_mean": 500.28125,
    "epsilon_norm": 9.904806137084961,
    "prediction_norm": 7.96783971786499,
    "timestamp": "2025-07-14T12:33:53.986440"
  },
  {
    "epoch": 578,
    "batch_idx": 128,
    "loss": 0.3415493667125702,
    "learning_rate": 0.0008076931850895859,
    "timestep_mean": 512.21875,
    "epsilon_norm": 8.14952278137207,
    "prediction_norm": 6.557600021362305,
    "timestamp": "2025-07-14T12:33:53.991376"
  },
  {
    "epoch": 578,
    "batch_idx": 160,
    "loss": 0.17373530566692352,
    "learning_rate": 0.0008076931850895859,
    "timestep_mean": 617.125,
    "epsilon_norm": 9.071824073791504,
    "prediction_norm": 8.110057830810547,
    "timestamp": "2025-07-14T12:33:53.996377"
  },
  {
    "epoch": 578,
    "batch_idx": 192,
    "loss": 0.1403319537639618,
    "learning_rate": 0.0008076931850895859,
    "timestep_mean": 568.125,
    "epsilon_norm": 4.884602069854736,
    "prediction_norm": 4.367465972900391,
    "timestamp": "2025-07-14T12:33:54.001160"
  },
  {
    "epoch": 579,
    "batch_idx": 0,
    "loss": 0.636387825012207,
    "learning_rate": 0.0008070737347949948,
    "timestep_mean": 373.46875,
    "epsilon_norm": 8.772751808166504,
    "prediction_norm": 5.939367294311523,
    "timestamp": "2025-07-14T12:33:54.006070"
  },
  {
    "epoch": 579,
    "batch_idx": 32,
    "loss": 0.24978092312812805,
    "learning_rate": 0.0008070737347949948,
    "timestep_mean": 503.9375,
    "epsilon_norm": 6.957542419433594,
    "prediction_norm": 5.577349662780762,
    "timestamp": "2025-07-14T12:33:54.010994"
  },
  {
    "epoch": 579,
    "batch_idx": 64,
    "loss": 0.19239448010921478,
    "learning_rate": 0.0008070737347949948,
    "timestep_mean": 516.03125,
    "epsilon_norm": 7.9259114265441895,
    "prediction_norm": 7.039395809173584,
    "timestamp": "2025-07-14T12:33:54.015874"
  },
  {
    "epoch": 579,
    "batch_idx": 96,
    "loss": 0.3439749479293823,
    "learning_rate": 0.0008070737347949948,
    "timestep_mean": 476.84375,
    "epsilon_norm": 7.455793380737305,
    "prediction_norm": 6.219176769256592,
    "timestamp": "2025-07-14T12:33:54.020709"
  },
  {
    "epoch": 579,
    "batch_idx": 128,
    "loss": 0.39036738872528076,
    "learning_rate": 0.0008070737347949948,
    "timestep_mean": 445.75,
    "epsilon_norm": 8.521141052246094,
    "prediction_norm": 6.938745498657227,
    "timestamp": "2025-07-14T12:33:54.025647"
  },
  {
    "epoch": 579,
    "batch_idx": 160,
    "loss": 0.39084967970848083,
    "learning_rate": 0.0008070737347949948,
    "timestep_mean": 488.8125,
    "epsilon_norm": 7.689652919769287,
    "prediction_norm": 6.834331512451172,
    "timestamp": "2025-07-14T12:33:54.030497"
  },
  {
    "epoch": 579,
    "batch_idx": 192,
    "loss": 0.238142192363739,
    "learning_rate": 0.0008070737347949948,
    "timestep_mean": 437.5,
    "epsilon_norm": 3.347618579864502,
    "prediction_norm": 3.161968946456909,
    "timestamp": "2025-07-14T12:33:54.035294"
  },
  {
    "epoch": 580,
    "batch_idx": 0,
    "loss": 0.3352936804294586,
    "learning_rate": 0.0008064535268264884,
    "timestep_mean": 458.25,
    "epsilon_norm": 7.626201629638672,
    "prediction_norm": 6.243826389312744,
    "timestamp": "2025-07-14T12:33:54.040128"
  },
  {
    "epoch": 580,
    "batch_idx": 32,
    "loss": 0.35022401809692383,
    "learning_rate": 0.0008064535268264884,
    "timestep_mean": 449.875,
    "epsilon_norm": 8.794903755187988,
    "prediction_norm": 7.425783157348633,
    "timestamp": "2025-07-14T12:33:54.045008"
  },
  {
    "epoch": 580,
    "batch_idx": 64,
    "loss": 0.37120193243026733,
    "learning_rate": 0.0008064535268264884,
    "timestep_mean": 496.34375,
    "epsilon_norm": 8.017006874084473,
    "prediction_norm": 6.106851100921631,
    "timestamp": "2025-07-14T12:33:54.049905"
  },
  {
    "epoch": 580,
    "batch_idx": 96,
    "loss": 0.4875362515449524,
    "learning_rate": 0.0008064535268264884,
    "timestep_mean": 442.375,
    "epsilon_norm": 8.658427238464355,
    "prediction_norm": 6.532862186431885,
    "timestamp": "2025-07-14T12:33:54.054825"
  },
  {
    "epoch": 580,
    "batch_idx": 128,
    "loss": 0.40626204013824463,
    "learning_rate": 0.0008064535268264884,
    "timestep_mean": 484.5,
    "epsilon_norm": 7.202991962432861,
    "prediction_norm": 6.034703254699707,
    "timestamp": "2025-07-14T12:33:54.059698"
  },
  {
    "epoch": 580,
    "batch_idx": 160,
    "loss": 0.20258383452892303,
    "learning_rate": 0.0008064535268264884,
    "timestep_mean": 520.78125,
    "epsilon_norm": 6.685196399688721,
    "prediction_norm": 6.059629440307617,
    "timestamp": "2025-07-14T12:33:54.064560"
  },
  {
    "epoch": 580,
    "batch_idx": 192,
    "loss": 0.49352896213531494,
    "learning_rate": 0.0008064535268264884,
    "timestep_mean": 325.375,
    "epsilon_norm": 3.7159266471862793,
    "prediction_norm": 2.564462423324585,
    "timestamp": "2025-07-14T12:33:54.069410"
  },
  {
    "epoch": 581,
    "batch_idx": 0,
    "loss": 0.2685278058052063,
    "learning_rate": 0.0008058325627143682,
    "timestep_mean": 596.21875,
    "epsilon_norm": 7.961605548858643,
    "prediction_norm": 6.763531684875488,
    "timestamp": "2025-07-14T12:33:54.074345"
  },
  {
    "epoch": 581,
    "batch_idx": 32,
    "loss": 0.18903477489948273,
    "learning_rate": 0.0008058325627143682,
    "timestep_mean": 580.78125,
    "epsilon_norm": 8.577322959899902,
    "prediction_norm": 7.202025413513184,
    "timestamp": "2025-07-14T12:33:54.079303"
  },
  {
    "epoch": 581,
    "batch_idx": 64,
    "loss": 0.45558395981788635,
    "learning_rate": 0.0008058325627143682,
    "timestep_mean": 476.84375,
    "epsilon_norm": 8.175600051879883,
    "prediction_norm": 6.515568733215332,
    "timestamp": "2025-07-14T12:33:54.084211"
  },
  {
    "epoch": 581,
    "batch_idx": 96,
    "loss": 0.23713016510009766,
    "learning_rate": 0.0008058325627143682,
    "timestep_mean": 485.8125,
    "epsilon_norm": 7.245066165924072,
    "prediction_norm": 5.800107002258301,
    "timestamp": "2025-07-14T12:33:54.089127"
  },
  {
    "epoch": 581,
    "batch_idx": 128,
    "loss": 0.28099900484085083,
    "learning_rate": 0.0008058325627143682,
    "timestep_mean": 498.75,
    "epsilon_norm": 7.746548652648926,
    "prediction_norm": 6.675172805786133,
    "timestamp": "2025-07-14T12:33:54.093988"
  },
  {
    "epoch": 581,
    "batch_idx": 160,
    "loss": 0.25205692648887634,
    "learning_rate": 0.0008058325627143682,
    "timestep_mean": 468.6875,
    "epsilon_norm": 8.084968566894531,
    "prediction_norm": 7.02780294418335,
    "timestamp": "2025-07-14T12:33:54.098829"
  },
  {
    "epoch": 581,
    "batch_idx": 192,
    "loss": 0.022482385858893394,
    "learning_rate": 0.0008058325627143682,
    "timestep_mean": 735.75,
    "epsilon_norm": 4.173730850219727,
    "prediction_norm": 3.9053874015808105,
    "timestamp": "2025-07-14T12:33:54.103697"
  },
  {
    "epoch": 582,
    "batch_idx": 0,
    "loss": 0.2143665999174118,
    "learning_rate": 0.0008052108439908013,
    "timestep_mean": 543.46875,
    "epsilon_norm": 7.908277988433838,
    "prediction_norm": 6.324498176574707,
    "timestamp": "2025-07-14T12:33:54.108632"
  },
  {
    "epoch": 582,
    "batch_idx": 32,
    "loss": 0.6368143558502197,
    "learning_rate": 0.0008052108439908013,
    "timestep_mean": 458.09375,
    "epsilon_norm": 9.255864143371582,
    "prediction_norm": 6.507952690124512,
    "timestamp": "2025-07-14T12:33:54.113509"
  },
  {
    "epoch": 582,
    "batch_idx": 64,
    "loss": 0.23376186192035675,
    "learning_rate": 0.0008052108439908013,
    "timestep_mean": 546.15625,
    "epsilon_norm": 7.426056861877441,
    "prediction_norm": 5.899637222290039,
    "timestamp": "2025-07-14T12:33:54.118403"
  },
  {
    "epoch": 582,
    "batch_idx": 96,
    "loss": 0.11034909635782242,
    "learning_rate": 0.0008052108439908013,
    "timestep_mean": 543.5625,
    "epsilon_norm": 8.260696411132812,
    "prediction_norm": 7.4878315925598145,
    "timestamp": "2025-07-14T12:33:54.123252"
  },
  {
    "epoch": 582,
    "batch_idx": 128,
    "loss": 0.2156040072441101,
    "learning_rate": 0.0008052108439908013,
    "timestep_mean": 532.875,
    "epsilon_norm": 7.607922554016113,
    "prediction_norm": 6.749401569366455,
    "timestamp": "2025-07-14T12:33:54.128178"
  },
  {
    "epoch": 582,
    "batch_idx": 160,
    "loss": 0.14883063733577728,
    "learning_rate": 0.0008052108439908013,
    "timestep_mean": 535.1875,
    "epsilon_norm": 7.002678871154785,
    "prediction_norm": 6.585697174072266,
    "timestamp": "2025-07-14T12:33:54.133006"
  },
  {
    "epoch": 582,
    "batch_idx": 192,
    "loss": 0.1497741937637329,
    "learning_rate": 0.0008052108439908013,
    "timestep_mean": 657.25,
    "epsilon_norm": 3.0371718406677246,
    "prediction_norm": 2.850886821746826,
    "timestamp": "2025-07-14T12:33:54.137919"
  },
  {
    "epoch": 583,
    "batch_idx": 0,
    "loss": 0.19496504962444305,
    "learning_rate": 0.0008045883721898171,
    "timestep_mean": 591.5625,
    "epsilon_norm": 7.789240837097168,
    "prediction_norm": 6.413294315338135,
    "timestamp": "2025-07-14T12:33:54.142788"
  },
  {
    "epoch": 583,
    "batch_idx": 32,
    "loss": 0.23029164969921112,
    "learning_rate": 0.0008045883721898171,
    "timestep_mean": 578.59375,
    "epsilon_norm": 9.790892601013184,
    "prediction_norm": 8.024553298950195,
    "timestamp": "2025-07-14T12:33:54.147675"
  },
  {
    "epoch": 583,
    "batch_idx": 64,
    "loss": 0.3029314875602722,
    "learning_rate": 0.0008045883721898171,
    "timestep_mean": 485.59375,
    "epsilon_norm": 7.529836654663086,
    "prediction_norm": 6.245214939117432,
    "timestamp": "2025-07-14T12:33:54.152585"
  },
  {
    "epoch": 583,
    "batch_idx": 96,
    "loss": 0.282518208026886,
    "learning_rate": 0.0008045883721898171,
    "timestep_mean": 522.96875,
    "epsilon_norm": 8.296460151672363,
    "prediction_norm": 6.758415699005127,
    "timestamp": "2025-07-14T12:33:54.157543"
  },
  {
    "epoch": 583,
    "batch_idx": 128,
    "loss": 0.26509708166122437,
    "learning_rate": 0.0008045883721898171,
    "timestep_mean": 475.71875,
    "epsilon_norm": 7.625598430633545,
    "prediction_norm": 5.941205024719238,
    "timestamp": "2025-07-14T12:33:54.162385"
  },
  {
    "epoch": 583,
    "batch_idx": 160,
    "loss": 0.18983738124370575,
    "learning_rate": 0.0008045883721898171,
    "timestep_mean": 476.84375,
    "epsilon_norm": 7.015228271484375,
    "prediction_norm": 6.821292877197266,
    "timestamp": "2025-07-14T12:33:54.167375"
  },
  {
    "epoch": 583,
    "batch_idx": 192,
    "loss": 0.04314345866441727,
    "learning_rate": 0.0008045883721898171,
    "timestep_mean": 524.25,
    "epsilon_norm": 3.623124361038208,
    "prediction_norm": 3.4631216526031494,
    "timestamp": "2025-07-14T12:33:54.172180"
  },
  {
    "epoch": 584,
    "batch_idx": 0,
    "loss": 0.32448482513427734,
    "learning_rate": 0.0008039651488473028,
    "timestep_mean": 452.53125,
    "epsilon_norm": 6.715845584869385,
    "prediction_norm": 5.833698749542236,
    "timestamp": "2025-07-14T12:33:54.177076"
  },
  {
    "epoch": 584,
    "batch_idx": 32,
    "loss": 0.2907941937446594,
    "learning_rate": 0.0008039651488473028,
    "timestep_mean": 362.3125,
    "epsilon_norm": 7.491757869720459,
    "prediction_norm": 5.330745697021484,
    "timestamp": "2025-07-14T12:33:54.181913"
  },
  {
    "epoch": 584,
    "batch_idx": 64,
    "loss": 0.263034462928772,
    "learning_rate": 0.0008039651488473028,
    "timestep_mean": 524.4375,
    "epsilon_norm": 6.689336776733398,
    "prediction_norm": 5.372795581817627,
    "timestamp": "2025-07-14T12:33:54.186790"
  },
  {
    "epoch": 584,
    "batch_idx": 96,
    "loss": 0.19599147140979767,
    "learning_rate": 0.0008039651488473028,
    "timestep_mean": 539.90625,
    "epsilon_norm": 8.112831115722656,
    "prediction_norm": 7.874556541442871,
    "timestamp": "2025-07-14T12:33:54.191736"
  },
  {
    "epoch": 584,
    "batch_idx": 128,
    "loss": 0.35150617361068726,
    "learning_rate": 0.0008039651488473028,
    "timestep_mean": 415.25,
    "epsilon_norm": 7.359872341156006,
    "prediction_norm": 5.838387966156006,
    "timestamp": "2025-07-14T12:33:54.196620"
  },
  {
    "epoch": 584,
    "batch_idx": 160,
    "loss": 0.3614129424095154,
    "learning_rate": 0.0008039651488473028,
    "timestep_mean": 475.25,
    "epsilon_norm": 7.492652893066406,
    "prediction_norm": 6.552816867828369,
    "timestamp": "2025-07-14T12:33:54.201576"
  },
  {
    "epoch": 584,
    "batch_idx": 192,
    "loss": 0.4505203664302826,
    "learning_rate": 0.0008039651488473028,
    "timestep_mean": 464.125,
    "epsilon_norm": 3.3252458572387695,
    "prediction_norm": 2.878647804260254,
    "timestamp": "2025-07-14T12:33:54.206448"
  },
  {
    "epoch": 585,
    "batch_idx": 0,
    "loss": 0.20522867143154144,
    "learning_rate": 0.0008033411755009999,
    "timestep_mean": 529.0625,
    "epsilon_norm": 7.06742525100708,
    "prediction_norm": 6.676483154296875,
    "timestamp": "2025-07-14T12:33:54.211365"
  },
  {
    "epoch": 585,
    "batch_idx": 32,
    "loss": 0.33840686082839966,
    "learning_rate": 0.0008033411755009999,
    "timestep_mean": 458.21875,
    "epsilon_norm": 7.796037673950195,
    "prediction_norm": 5.928811073303223,
    "timestamp": "2025-07-14T12:33:54.216278"
  },
  {
    "epoch": 585,
    "batch_idx": 64,
    "loss": 0.3857945203781128,
    "learning_rate": 0.0008033411755009999,
    "timestep_mean": 460.71875,
    "epsilon_norm": 7.468967914581299,
    "prediction_norm": 5.506938457489014,
    "timestamp": "2025-07-14T12:33:54.221361"
  },
  {
    "epoch": 585,
    "batch_idx": 96,
    "loss": 0.49164626002311707,
    "learning_rate": 0.0008033411755009999,
    "timestep_mean": 494.5,
    "epsilon_norm": 9.400124549865723,
    "prediction_norm": 7.11655330657959,
    "timestamp": "2025-07-14T12:33:54.226193"
  },
  {
    "epoch": 585,
    "batch_idx": 128,
    "loss": 0.30946141481399536,
    "learning_rate": 0.0008033411755009999,
    "timestep_mean": 536.09375,
    "epsilon_norm": 8.87509536743164,
    "prediction_norm": 7.2620720863342285,
    "timestamp": "2025-07-14T12:33:54.231055"
  },
  {
    "epoch": 585,
    "batch_idx": 160,
    "loss": 0.1423424631357193,
    "learning_rate": 0.0008033411755009999,
    "timestep_mean": 540.03125,
    "epsilon_norm": 6.723874092102051,
    "prediction_norm": 6.054367542266846,
    "timestamp": "2025-07-14T12:33:54.235901"
  },
  {
    "epoch": 585,
    "batch_idx": 192,
    "loss": 0.5300909876823425,
    "learning_rate": 0.0008033411755009999,
    "timestep_mean": 511.625,
    "epsilon_norm": 4.184975624084473,
    "prediction_norm": 3.047590970993042,
    "timestamp": "2025-07-14T12:33:54.240785"
  },
  {
    "epoch": 586,
    "batch_idx": 0,
    "loss": 0.31774014234542847,
    "learning_rate": 0.0008027164536905007,
    "timestep_mean": 471.28125,
    "epsilon_norm": 7.867706298828125,
    "prediction_norm": 6.656895160675049,
    "timestamp": "2025-07-14T12:33:54.245657"
  },
  {
    "epoch": 586,
    "batch_idx": 32,
    "loss": 0.2829166054725647,
    "learning_rate": 0.0008027164536905007,
    "timestep_mean": 431.15625,
    "epsilon_norm": 8.667102813720703,
    "prediction_norm": 6.782639503479004,
    "timestamp": "2025-07-14T12:33:54.250518"
  },
  {
    "epoch": 586,
    "batch_idx": 64,
    "loss": 0.1778179407119751,
    "learning_rate": 0.0008027164536905007,
    "timestep_mean": 487.34375,
    "epsilon_norm": 6.826539993286133,
    "prediction_norm": 6.00369930267334,
    "timestamp": "2025-07-14T12:33:54.255374"
  },
  {
    "epoch": 586,
    "batch_idx": 96,
    "loss": 0.2427518665790558,
    "learning_rate": 0.0008027164536905007,
    "timestep_mean": 439.78125,
    "epsilon_norm": 7.587223052978516,
    "prediction_norm": 6.634093761444092,
    "timestamp": "2025-07-14T12:33:54.260265"
  },
  {
    "epoch": 586,
    "batch_idx": 128,
    "loss": 0.2979792356491089,
    "learning_rate": 0.0008027164536905007,
    "timestep_mean": 444.15625,
    "epsilon_norm": 8.589706420898438,
    "prediction_norm": 7.003293037414551,
    "timestamp": "2025-07-14T12:33:54.265137"
  },
  {
    "epoch": 586,
    "batch_idx": 160,
    "loss": 0.44959738850593567,
    "learning_rate": 0.0008027164536905007,
    "timestep_mean": 458.75,
    "epsilon_norm": 8.401555061340332,
    "prediction_norm": 6.711781978607178,
    "timestamp": "2025-07-14T12:33:54.270012"
  },
  {
    "epoch": 586,
    "batch_idx": 192,
    "loss": 0.4738423228263855,
    "learning_rate": 0.0008027164536905007,
    "timestep_mean": 277.375,
    "epsilon_norm": 3.2822208404541016,
    "prediction_norm": 2.3678033351898193,
    "timestamp": "2025-07-14T12:33:54.274928"
  },
  {
    "epoch": 587,
    "batch_idx": 0,
    "loss": 0.2937212586402893,
    "learning_rate": 0.0008020909849572442,
    "timestep_mean": 468.0625,
    "epsilon_norm": 6.344388484954834,
    "prediction_norm": 5.434142589569092,
    "timestamp": "2025-07-14T12:33:54.279806"
  },
  {
    "epoch": 587,
    "batch_idx": 32,
    "loss": 0.23208431899547577,
    "learning_rate": 0.0008020909849572442,
    "timestep_mean": 554.125,
    "epsilon_norm": 7.9512481689453125,
    "prediction_norm": 6.769199848175049,
    "timestamp": "2025-07-14T12:33:54.284683"
  },
  {
    "epoch": 587,
    "batch_idx": 64,
    "loss": 0.26851022243499756,
    "learning_rate": 0.0008020909849572442,
    "timestep_mean": 555.9375,
    "epsilon_norm": 7.956432342529297,
    "prediction_norm": 7.137068748474121,
    "timestamp": "2025-07-14T12:33:54.289552"
  },
  {
    "epoch": 587,
    "batch_idx": 96,
    "loss": 0.2352381944656372,
    "learning_rate": 0.0008020909849572442,
    "timestep_mean": 477.0,
    "epsilon_norm": 7.63233757019043,
    "prediction_norm": 7.306801795959473,
    "timestamp": "2025-07-14T12:33:54.294462"
  },
  {
    "epoch": 587,
    "batch_idx": 128,
    "loss": 0.3139568567276001,
    "learning_rate": 0.0008020909849572442,
    "timestep_mean": 502.3125,
    "epsilon_norm": 8.227558135986328,
    "prediction_norm": 7.831916332244873,
    "timestamp": "2025-07-14T12:33:54.299362"
  },
  {
    "epoch": 587,
    "batch_idx": 160,
    "loss": 0.2894286513328552,
    "learning_rate": 0.0008020909849572442,
    "timestep_mean": 420.375,
    "epsilon_norm": 7.048558235168457,
    "prediction_norm": 5.528077602386475,
    "timestamp": "2025-07-14T12:33:54.304351"
  },
  {
    "epoch": 587,
    "batch_idx": 192,
    "loss": 0.3536284267902374,
    "learning_rate": 0.0008020909849572442,
    "timestep_mean": 493.25,
    "epsilon_norm": 3.881580352783203,
    "prediction_norm": 2.5032951831817627,
    "timestamp": "2025-07-14T12:33:54.311245"
  },
  {
    "epoch": 588,
    "batch_idx": 0,
    "loss": 0.4216307997703552,
    "learning_rate": 0.0008014647708445125,
    "timestep_mean": 475.5,
    "epsilon_norm": 8.472672462463379,
    "prediction_norm": 6.640277862548828,
    "timestamp": "2025-07-14T12:33:54.315978"
  },
  {
    "epoch": 588,
    "batch_idx": 32,
    "loss": 0.3333929777145386,
    "learning_rate": 0.0008014647708445125,
    "timestep_mean": 407.25,
    "epsilon_norm": 7.970365047454834,
    "prediction_norm": 6.515722274780273,
    "timestamp": "2025-07-14T12:33:54.320796"
  },
  {
    "epoch": 588,
    "batch_idx": 64,
    "loss": 0.15025199949741364,
    "learning_rate": 0.0008014647708445125,
    "timestep_mean": 512.9375,
    "epsilon_norm": 7.2912797927856445,
    "prediction_norm": 6.026322364807129,
    "timestamp": "2025-07-14T12:33:54.325548"
  },
  {
    "epoch": 588,
    "batch_idx": 96,
    "loss": 0.1764051765203476,
    "learning_rate": 0.0008014647708445125,
    "timestep_mean": 521.0625,
    "epsilon_norm": 8.015347480773926,
    "prediction_norm": 6.559041976928711,
    "timestamp": "2025-07-14T12:33:54.330252"
  },
  {
    "epoch": 588,
    "batch_idx": 128,
    "loss": 0.3020779490470886,
    "learning_rate": 0.0008014647708445125,
    "timestep_mean": 364.0,
    "epsilon_norm": 7.455297946929932,
    "prediction_norm": 5.793741226196289,
    "timestamp": "2025-07-14T12:33:54.334975"
  },
  {
    "epoch": 588,
    "batch_idx": 160,
    "loss": 0.4439450204372406,
    "learning_rate": 0.0008014647708445125,
    "timestep_mean": 469.65625,
    "epsilon_norm": 9.4586820602417,
    "prediction_norm": 7.327541828155518,
    "timestamp": "2025-07-14T12:33:54.339717"
  },
  {
    "epoch": 588,
    "batch_idx": 192,
    "loss": 0.46243584156036377,
    "learning_rate": 0.0008014647708445125,
    "timestep_mean": 454.0,
    "epsilon_norm": 4.957906723022461,
    "prediction_norm": 4.167214870452881,
    "timestamp": "2025-07-14T12:33:54.344485"
  },
  {
    "epoch": 589,
    "batch_idx": 0,
    "loss": 0.2455292046070099,
    "learning_rate": 0.0008008378128974262,
    "timestep_mean": 531.40625,
    "epsilon_norm": 8.247477531433105,
    "prediction_norm": 7.240451335906982,
    "timestamp": "2025-07-14T12:33:54.349239"
  },
  {
    "epoch": 589,
    "batch_idx": 32,
    "loss": 0.3656123876571655,
    "learning_rate": 0.0008008378128974262,
    "timestep_mean": 490.3125,
    "epsilon_norm": 8.432402610778809,
    "prediction_norm": 6.590733528137207,
    "timestamp": "2025-07-14T12:33:54.353972"
  },
  {
    "epoch": 589,
    "batch_idx": 64,
    "loss": 0.27031409740448,
    "learning_rate": 0.0008008378128974262,
    "timestep_mean": 479.375,
    "epsilon_norm": 7.200237274169922,
    "prediction_norm": 6.856029033660889,
    "timestamp": "2025-07-14T12:33:54.358837"
  },
  {
    "epoch": 589,
    "batch_idx": 96,
    "loss": 0.2015705108642578,
    "learning_rate": 0.0008008378128974262,
    "timestep_mean": 532.0,
    "epsilon_norm": 6.924928665161133,
    "prediction_norm": 7.124802112579346,
    "timestamp": "2025-07-14T12:33:54.363561"
  },
  {
    "epoch": 589,
    "batch_idx": 128,
    "loss": 0.326458215713501,
    "learning_rate": 0.0008008378128974262,
    "timestep_mean": 498.5,
    "epsilon_norm": 8.754297256469727,
    "prediction_norm": 7.35050630569458,
    "timestamp": "2025-07-14T12:33:54.368268"
  },
  {
    "epoch": 589,
    "batch_idx": 160,
    "loss": 0.19001264870166779,
    "learning_rate": 0.0008008378128974262,
    "timestep_mean": 529.15625,
    "epsilon_norm": 8.87238883972168,
    "prediction_norm": 8.58545207977295,
    "timestamp": "2025-07-14T12:33:54.372984"
  },
  {
    "epoch": 589,
    "batch_idx": 192,
    "loss": 0.49371853470802307,
    "learning_rate": 0.0008008378128974262,
    "timestep_mean": 359.5,
    "epsilon_norm": 4.580071926116943,
    "prediction_norm": 2.953580379486084,
    "timestamp": "2025-07-14T12:33:54.377689"
  },
  {
    "epoch": 590,
    "batch_idx": 0,
    "loss": 0.3485438823699951,
    "learning_rate": 0.000800210112662942,
    "timestep_mean": 472.71875,
    "epsilon_norm": 7.217624187469482,
    "prediction_norm": 5.938055515289307,
    "timestamp": "2025-07-14T12:33:54.382704"
  },
  {
    "epoch": 590,
    "batch_idx": 32,
    "loss": 0.3390220105648041,
    "learning_rate": 0.000800210112662942,
    "timestep_mean": 416.0,
    "epsilon_norm": 7.175054550170898,
    "prediction_norm": 5.946816444396973,
    "timestamp": "2025-07-14T12:33:54.387642"
  },
  {
    "epoch": 590,
    "batch_idx": 64,
    "loss": 0.42802539467811584,
    "learning_rate": 0.000800210112662942,
    "timestep_mean": 432.8125,
    "epsilon_norm": 8.482123374938965,
    "prediction_norm": 6.96275520324707,
    "timestamp": "2025-07-14T12:33:54.392373"
  },
  {
    "epoch": 590,
    "batch_idx": 96,
    "loss": 0.19221365451812744,
    "learning_rate": 0.000800210112662942,
    "timestep_mean": 601.21875,
    "epsilon_norm": 7.088610649108887,
    "prediction_norm": 6.3102030754089355,
    "timestamp": "2025-07-14T12:33:54.397053"
  },
  {
    "epoch": 590,
    "batch_idx": 128,
    "loss": 0.36647629737854004,
    "learning_rate": 0.000800210112662942,
    "timestep_mean": 526.5625,
    "epsilon_norm": 7.477072238922119,
    "prediction_norm": 5.923842430114746,
    "timestamp": "2025-07-14T12:33:54.401818"
  },
  {
    "epoch": 590,
    "batch_idx": 160,
    "loss": 0.35715121030807495,
    "learning_rate": 0.000800210112662942,
    "timestep_mean": 480.75,
    "epsilon_norm": 7.678944110870361,
    "prediction_norm": 5.470360279083252,
    "timestamp": "2025-07-14T12:33:54.406515"
  },
  {
    "epoch": 590,
    "batch_idx": 192,
    "loss": 0.1790814995765686,
    "learning_rate": 0.000800210112662942,
    "timestep_mean": 592.75,
    "epsilon_norm": 4.356313228607178,
    "prediction_norm": 3.3407888412475586,
    "timestamp": "2025-07-14T12:33:54.411260"
  },
  {
    "epoch": 591,
    "batch_idx": 0,
    "loss": 0.2171194851398468,
    "learning_rate": 0.000799581671689848,
    "timestep_mean": 458.5625,
    "epsilon_norm": 8.20460319519043,
    "prediction_norm": 6.989102363586426,
    "timestamp": "2025-07-14T12:33:54.416344"
  },
  {
    "epoch": 591,
    "batch_idx": 32,
    "loss": 0.20610439777374268,
    "learning_rate": 0.000799581671689848,
    "timestep_mean": 510.3125,
    "epsilon_norm": 7.7643351554870605,
    "prediction_norm": 6.48557710647583,
    "timestamp": "2025-07-14T12:33:54.421582"
  },
  {
    "epoch": 591,
    "batch_idx": 64,
    "loss": 0.20522889494895935,
    "learning_rate": 0.000799581671689848,
    "timestep_mean": 478.8125,
    "epsilon_norm": 6.370408058166504,
    "prediction_norm": 4.622952938079834,
    "timestamp": "2025-07-14T12:33:54.426678"
  },
  {
    "epoch": 591,
    "batch_idx": 96,
    "loss": 0.3918086886405945,
    "learning_rate": 0.000799581671689848,
    "timestep_mean": 442.3125,
    "epsilon_norm": 8.162698745727539,
    "prediction_norm": 5.642830848693848,
    "timestamp": "2025-07-14T12:33:54.431713"
  },
  {
    "epoch": 591,
    "batch_idx": 128,
    "loss": 0.3113970160484314,
    "learning_rate": 0.000799581671689848,
    "timestep_mean": 437.125,
    "epsilon_norm": 9.361176490783691,
    "prediction_norm": 7.912267684936523,
    "timestamp": "2025-07-14T12:33:54.436805"
  },
  {
    "epoch": 591,
    "batch_idx": 160,
    "loss": 0.49010205268859863,
    "learning_rate": 0.000799581671689848,
    "timestep_mean": 479.5,
    "epsilon_norm": 7.692274570465088,
    "prediction_norm": 5.477819442749023,
    "timestamp": "2025-07-14T12:33:54.441892"
  },
  {
    "epoch": 591,
    "batch_idx": 192,
    "loss": 0.5666530728340149,
    "learning_rate": 0.000799581671689848,
    "timestep_mean": 564.75,
    "epsilon_norm": 4.920832633972168,
    "prediction_norm": 3.894676923751831,
    "timestamp": "2025-07-14T12:33:54.447436"
  },
  {
    "epoch": 592,
    "batch_idx": 0,
    "loss": 0.3077928423881531,
    "learning_rate": 0.0007989524915287596,
    "timestep_mean": 562.03125,
    "epsilon_norm": 9.41061782836914,
    "prediction_norm": 8.566438674926758,
    "timestamp": "2025-07-14T12:33:54.452586"
  },
  {
    "epoch": 592,
    "batch_idx": 32,
    "loss": 0.1274995505809784,
    "learning_rate": 0.0007989524915287596,
    "timestep_mean": 550.84375,
    "epsilon_norm": 6.583618640899658,
    "prediction_norm": 5.999009132385254,
    "timestamp": "2025-07-14T12:33:54.457728"
  },
  {
    "epoch": 592,
    "batch_idx": 64,
    "loss": 0.3720812201499939,
    "learning_rate": 0.0007989524915287596,
    "timestep_mean": 554.40625,
    "epsilon_norm": 8.125293731689453,
    "prediction_norm": 6.222607135772705,
    "timestamp": "2025-07-14T12:33:54.462913"
  },
  {
    "epoch": 592,
    "batch_idx": 96,
    "loss": 0.29303058981895447,
    "learning_rate": 0.0007989524915287596,
    "timestep_mean": 498.0,
    "epsilon_norm": 8.432564735412598,
    "prediction_norm": 7.610702991485596,
    "timestamp": "2025-07-14T12:33:54.467981"
  },
  {
    "epoch": 592,
    "batch_idx": 128,
    "loss": 0.4285441040992737,
    "learning_rate": 0.0007989524915287596,
    "timestep_mean": 430.3125,
    "epsilon_norm": 7.507527828216553,
    "prediction_norm": 6.798493385314941,
    "timestamp": "2025-07-14T12:33:54.473132"
  },
  {
    "epoch": 592,
    "batch_idx": 160,
    "loss": 0.36571037769317627,
    "learning_rate": 0.0007989524915287596,
    "timestep_mean": 445.5,
    "epsilon_norm": 8.323561668395996,
    "prediction_norm": 7.415691375732422,
    "timestamp": "2025-07-14T12:33:54.478389"
  },
  {
    "epoch": 592,
    "batch_idx": 192,
    "loss": 0.28051596879959106,
    "learning_rate": 0.0007989524915287596,
    "timestep_mean": 453.25,
    "epsilon_norm": 3.8746962547302246,
    "prediction_norm": 4.144207954406738,
    "timestamp": "2025-07-14T12:33:54.483946"
  },
  {
    "epoch": 593,
    "batch_idx": 0,
    "loss": 0.4741050899028778,
    "learning_rate": 0.0007983225737321163,
    "timestep_mean": 474.1875,
    "epsilon_norm": 8.230843544006348,
    "prediction_norm": 6.291380405426025,
    "timestamp": "2025-07-14T12:33:54.489246"
  },
  {
    "epoch": 593,
    "batch_idx": 32,
    "loss": 0.2695971429347992,
    "learning_rate": 0.0007983225737321163,
    "timestep_mean": 477.21875,
    "epsilon_norm": 7.6796112060546875,
    "prediction_norm": 6.899258613586426,
    "timestamp": "2025-07-14T12:33:54.494207"
  },
  {
    "epoch": 593,
    "batch_idx": 64,
    "loss": 0.17738458514213562,
    "learning_rate": 0.0007983225737321163,
    "timestep_mean": 517.46875,
    "epsilon_norm": 6.945209503173828,
    "prediction_norm": 7.095204830169678,
    "timestamp": "2025-07-14T12:33:54.499084"
  },
  {
    "epoch": 593,
    "batch_idx": 96,
    "loss": 0.5043244361877441,
    "learning_rate": 0.0007983225737321163,
    "timestep_mean": 417.40625,
    "epsilon_norm": 8.214821815490723,
    "prediction_norm": 7.168853759765625,
    "timestamp": "2025-07-14T12:33:54.503966"
  },
  {
    "epoch": 593,
    "batch_idx": 128,
    "loss": 0.3440381586551666,
    "learning_rate": 0.0007983225737321163,
    "timestep_mean": 497.875,
    "epsilon_norm": 8.908353805541992,
    "prediction_norm": 7.121870517730713,
    "timestamp": "2025-07-14T12:33:54.508801"
  },
  {
    "epoch": 593,
    "batch_idx": 160,
    "loss": 0.37990516424179077,
    "learning_rate": 0.0007983225737321163,
    "timestep_mean": 483.71875,
    "epsilon_norm": 8.54076099395752,
    "prediction_norm": 7.3914103507995605,
    "timestamp": "2025-07-14T12:33:54.513614"
  },
  {
    "epoch": 593,
    "batch_idx": 192,
    "loss": 0.15924465656280518,
    "learning_rate": 0.0007983225737321163,
    "timestep_mean": 504.375,
    "epsilon_norm": 4.629871368408203,
    "prediction_norm": 4.15053129196167,
    "timestamp": "2025-07-14T12:33:54.518469"
  },
  {
    "epoch": 594,
    "batch_idx": 0,
    "loss": 0.33624768257141113,
    "learning_rate": 0.0007976919198541775,
    "timestep_mean": 431.5,
    "epsilon_norm": 7.581874847412109,
    "prediction_norm": 6.356462478637695,
    "timestamp": "2025-07-14T12:33:54.523446"
  },
  {
    "epoch": 594,
    "batch_idx": 32,
    "loss": 0.39304274320602417,
    "learning_rate": 0.0007976919198541775,
    "timestep_mean": 473.15625,
    "epsilon_norm": 8.041156768798828,
    "prediction_norm": 6.367915153503418,
    "timestamp": "2025-07-14T12:33:54.528344"
  },
  {
    "epoch": 594,
    "batch_idx": 64,
    "loss": 0.45315268635749817,
    "learning_rate": 0.0007976919198541775,
    "timestep_mean": 439.71875,
    "epsilon_norm": 8.226703643798828,
    "prediction_norm": 5.69043493270874,
    "timestamp": "2025-07-14T12:33:54.533441"
  },
  {
    "epoch": 594,
    "batch_idx": 96,
    "loss": 0.27508407831192017,
    "learning_rate": 0.0007976919198541775,
    "timestep_mean": 537.84375,
    "epsilon_norm": 8.25632095336914,
    "prediction_norm": 6.99239444732666,
    "timestamp": "2025-07-14T12:33:54.538272"
  },
  {
    "epoch": 594,
    "batch_idx": 128,
    "loss": 0.2563897669315338,
    "learning_rate": 0.0007976919198541775,
    "timestep_mean": 587.15625,
    "epsilon_norm": 7.758695602416992,
    "prediction_norm": 6.871661186218262,
    "timestamp": "2025-07-14T12:33:54.543214"
  },
  {
    "epoch": 594,
    "batch_idx": 160,
    "loss": 0.2066033035516739,
    "learning_rate": 0.0007976919198541775,
    "timestep_mean": 535.875,
    "epsilon_norm": 7.4915242195129395,
    "prediction_norm": 6.06214714050293,
    "timestamp": "2025-07-14T12:33:54.548096"
  },
  {
    "epoch": 594,
    "batch_idx": 192,
    "loss": 0.27204659581184387,
    "learning_rate": 0.0007976919198541775,
    "timestep_mean": 428.625,
    "epsilon_norm": 4.375885009765625,
    "prediction_norm": 3.491825580596924,
    "timestamp": "2025-07-14T12:33:54.552969"
  },
  {
    "epoch": 595,
    "batch_idx": 0,
    "loss": 0.28270789980888367,
    "learning_rate": 0.0007970605314510194,
    "timestep_mean": 478.84375,
    "epsilon_norm": 7.650748252868652,
    "prediction_norm": 5.998465061187744,
    "timestamp": "2025-07-14T12:33:54.557863"
  },
  {
    "epoch": 595,
    "batch_idx": 32,
    "loss": 0.35491490364074707,
    "learning_rate": 0.0007970605314510194,
    "timestep_mean": 559.375,
    "epsilon_norm": 8.312040328979492,
    "prediction_norm": 6.780356407165527,
    "timestamp": "2025-07-14T12:33:54.563020"
  },
  {
    "epoch": 595,
    "batch_idx": 64,
    "loss": 0.15786847472190857,
    "learning_rate": 0.0007970605314510194,
    "timestep_mean": 541.4375,
    "epsilon_norm": 7.100879669189453,
    "prediction_norm": 6.143548488616943,
    "timestamp": "2025-07-14T12:33:54.568035"
  },
  {
    "epoch": 595,
    "batch_idx": 96,
    "loss": 0.3631921410560608,
    "learning_rate": 0.0007970605314510194,
    "timestep_mean": 595.625,
    "epsilon_norm": 8.365558624267578,
    "prediction_norm": 6.869409561157227,
    "timestamp": "2025-07-14T12:33:54.573137"
  },
  {
    "epoch": 595,
    "batch_idx": 128,
    "loss": 0.3038678467273712,
    "learning_rate": 0.0007970605314510194,
    "timestep_mean": 505.34375,
    "epsilon_norm": 8.687161445617676,
    "prediction_norm": 7.903497219085693,
    "timestamp": "2025-07-14T12:33:54.578195"
  },
  {
    "epoch": 595,
    "batch_idx": 160,
    "loss": 0.32517245411872864,
    "learning_rate": 0.0007970605314510194,
    "timestep_mean": 587.3125,
    "epsilon_norm": 8.108672142028809,
    "prediction_norm": 6.708189487457275,
    "timestamp": "2025-07-14T12:33:54.583466"
  },
  {
    "epoch": 595,
    "batch_idx": 192,
    "loss": 0.16330982744693756,
    "learning_rate": 0.0007970605314510194,
    "timestep_mean": 661.0,
    "epsilon_norm": 3.4891421794891357,
    "prediction_norm": 3.3952558040618896,
    "timestamp": "2025-07-14T12:33:54.588563"
  },
  {
    "epoch": 596,
    "batch_idx": 0,
    "loss": 0.31065747141838074,
    "learning_rate": 0.0007964284100805298,
    "timestep_mean": 526.75,
    "epsilon_norm": 8.910533905029297,
    "prediction_norm": 7.751968860626221,
    "timestamp": "2025-07-14T12:33:54.593650"
  },
  {
    "epoch": 596,
    "batch_idx": 32,
    "loss": 0.23962226510047913,
    "learning_rate": 0.0007964284100805298,
    "timestep_mean": 501.46875,
    "epsilon_norm": 8.997477531433105,
    "prediction_norm": 7.365325450897217,
    "timestamp": "2025-07-14T12:33:54.599023"
  },
  {
    "epoch": 596,
    "batch_idx": 64,
    "loss": 0.2349691540002823,
    "learning_rate": 0.0007964284100805298,
    "timestep_mean": 527.4375,
    "epsilon_norm": 7.097275257110596,
    "prediction_norm": 6.170819282531738,
    "timestamp": "2025-07-14T12:33:54.604166"
  },
  {
    "epoch": 596,
    "batch_idx": 96,
    "loss": 0.27887821197509766,
    "learning_rate": 0.0007964284100805298,
    "timestep_mean": 524.34375,
    "epsilon_norm": 8.241400718688965,
    "prediction_norm": 7.11134147644043,
    "timestamp": "2025-07-14T12:33:54.609283"
  },
  {
    "epoch": 596,
    "batch_idx": 128,
    "loss": 0.26970022916793823,
    "learning_rate": 0.0007964284100805298,
    "timestep_mean": 413.1875,
    "epsilon_norm": 6.805300712585449,
    "prediction_norm": 6.093843936920166,
    "timestamp": "2025-07-14T12:33:54.614485"
  },
  {
    "epoch": 596,
    "batch_idx": 160,
    "loss": 0.37235596776008606,
    "learning_rate": 0.0007964284100805298,
    "timestep_mean": 520.03125,
    "epsilon_norm": 8.62051773071289,
    "prediction_norm": 7.7470245361328125,
    "timestamp": "2025-07-14T12:33:54.619715"
  },
  {
    "epoch": 596,
    "batch_idx": 192,
    "loss": 0.24250586330890656,
    "learning_rate": 0.0007964284100805298,
    "timestep_mean": 579.125,
    "epsilon_norm": 4.8823161125183105,
    "prediction_norm": 3.943131923675537,
    "timestamp": "2025-07-14T12:33:54.624869"
  },
  {
    "epoch": 597,
    "batch_idx": 0,
    "loss": 0.3655821681022644,
    "learning_rate": 0.0007957955573024054,
    "timestep_mean": 487.0,
    "epsilon_norm": 6.664597511291504,
    "prediction_norm": 5.213394641876221,
    "timestamp": "2025-07-14T12:33:54.629984"
  },
  {
    "epoch": 597,
    "batch_idx": 32,
    "loss": 0.36927133798599243,
    "learning_rate": 0.0007957955573024054,
    "timestep_mean": 531.25,
    "epsilon_norm": 7.622805118560791,
    "prediction_norm": 5.788527965545654,
    "timestamp": "2025-07-14T12:33:54.635166"
  },
  {
    "epoch": 597,
    "batch_idx": 64,
    "loss": 0.18690434098243713,
    "learning_rate": 0.0007957955573024054,
    "timestep_mean": 540.0625,
    "epsilon_norm": 7.216451644897461,
    "prediction_norm": 6.242405891418457,
    "timestamp": "2025-07-14T12:33:54.640436"
  },
  {
    "epoch": 597,
    "batch_idx": 96,
    "loss": 0.19112619757652283,
    "learning_rate": 0.0007957955573024054,
    "timestep_mean": 518.75,
    "epsilon_norm": 7.522788047790527,
    "prediction_norm": 5.750361442565918,
    "timestamp": "2025-07-14T12:33:54.645560"
  },
  {
    "epoch": 597,
    "batch_idx": 128,
    "loss": 0.1887553632259369,
    "learning_rate": 0.0007957955573024054,
    "timestep_mean": 618.84375,
    "epsilon_norm": 9.29450798034668,
    "prediction_norm": 7.944473743438721,
    "timestamp": "2025-07-14T12:33:54.650926"
  },
  {
    "epoch": 597,
    "batch_idx": 160,
    "loss": 0.26124146580696106,
    "learning_rate": 0.0007957955573024054,
    "timestep_mean": 521.375,
    "epsilon_norm": 7.898778915405273,
    "prediction_norm": 6.736999034881592,
    "timestamp": "2025-07-14T12:33:54.656047"
  },
  {
    "epoch": 597,
    "batch_idx": 192,
    "loss": 0.3597587049007416,
    "learning_rate": 0.0007957955573024054,
    "timestep_mean": 547.75,
    "epsilon_norm": 3.4193625450134277,
    "prediction_norm": 1.8215267658233643,
    "timestamp": "2025-07-14T12:33:54.661223"
  },
  {
    "epoch": 598,
    "batch_idx": 0,
    "loss": 0.24238181114196777,
    "learning_rate": 0.0007951619746781475,
    "timestep_mean": 447.3125,
    "epsilon_norm": 7.464756965637207,
    "prediction_norm": 5.722575664520264,
    "timestamp": "2025-07-14T12:33:54.666578"
  },
  {
    "epoch": 598,
    "batch_idx": 32,
    "loss": 0.507819652557373,
    "learning_rate": 0.0007951619746781475,
    "timestep_mean": 521.21875,
    "epsilon_norm": 8.877065658569336,
    "prediction_norm": 7.194950580596924,
    "timestamp": "2025-07-14T12:33:54.671873"
  },
  {
    "epoch": 598,
    "batch_idx": 64,
    "loss": 0.293911337852478,
    "learning_rate": 0.0007951619746781475,
    "timestep_mean": 460.53125,
    "epsilon_norm": 7.710027694702148,
    "prediction_norm": 6.4088521003723145,
    "timestamp": "2025-07-14T12:33:54.677075"
  },
  {
    "epoch": 598,
    "batch_idx": 96,
    "loss": 0.4028116762638092,
    "learning_rate": 0.0007951619746781475,
    "timestep_mean": 508.875,
    "epsilon_norm": 9.09226131439209,
    "prediction_norm": 7.168190002441406,
    "timestamp": "2025-07-14T12:33:54.682540"
  },
  {
    "epoch": 598,
    "batch_idx": 128,
    "loss": 0.2625892758369446,
    "learning_rate": 0.0007951619746781475,
    "timestep_mean": 546.875,
    "epsilon_norm": 8.475560188293457,
    "prediction_norm": 7.8022871017456055,
    "timestamp": "2025-07-14T12:33:54.687591"
  },
  {
    "epoch": 598,
    "batch_idx": 160,
    "loss": 0.2112700194120407,
    "learning_rate": 0.0007951619746781475,
    "timestep_mean": 490.1875,
    "epsilon_norm": 7.403312683105469,
    "prediction_norm": 7.490826606750488,
    "timestamp": "2025-07-14T12:33:54.692793"
  },
  {
    "epoch": 598,
    "batch_idx": 192,
    "loss": 0.7333288192749023,
    "learning_rate": 0.0007951619746781475,
    "timestep_mean": 252.25,
    "epsilon_norm": 4.429341793060303,
    "prediction_norm": 2.775050640106201,
    "timestamp": "2025-07-14T12:33:54.697951"
  },
  {
    "epoch": 599,
    "batch_idx": 0,
    "loss": 0.08762761950492859,
    "learning_rate": 0.0007945276637710581,
    "timestep_mean": 610.21875,
    "epsilon_norm": 7.822150707244873,
    "prediction_norm": 7.8095221519470215,
    "timestamp": "2025-07-14T12:33:54.703105"
  },
  {
    "epoch": 599,
    "batch_idx": 32,
    "loss": 0.14981229603290558,
    "learning_rate": 0.0007945276637710581,
    "timestep_mean": 602.875,
    "epsilon_norm": 8.724898338317871,
    "prediction_norm": 8.172800064086914,
    "timestamp": "2025-07-14T12:33:54.708494"
  },
  {
    "epoch": 599,
    "batch_idx": 64,
    "loss": 0.15743571519851685,
    "learning_rate": 0.0007945276637710581,
    "timestep_mean": 562.0625,
    "epsilon_norm": 7.934651851654053,
    "prediction_norm": 7.473768711090088,
    "timestamp": "2025-07-14T12:33:54.713870"
  },
  {
    "epoch": 599,
    "batch_idx": 96,
    "loss": 0.31554973125457764,
    "learning_rate": 0.0007945276637710581,
    "timestep_mean": 482.28125,
    "epsilon_norm": 9.22367000579834,
    "prediction_norm": 7.546209812164307,
    "timestamp": "2025-07-14T12:33:54.718974"
  },
  {
    "epoch": 599,
    "batch_idx": 128,
    "loss": 0.25479254126548767,
    "learning_rate": 0.0007945276637710581,
    "timestep_mean": 513.21875,
    "epsilon_norm": 8.348548889160156,
    "prediction_norm": 6.441202640533447,
    "timestamp": "2025-07-14T12:33:54.724235"
  },
  {
    "epoch": 599,
    "batch_idx": 160,
    "loss": 0.34684959053993225,
    "learning_rate": 0.0007945276637710581,
    "timestep_mean": 461.9375,
    "epsilon_norm": 8.607911109924316,
    "prediction_norm": 7.153061389923096,
    "timestamp": "2025-07-14T12:33:54.729651"
  },
  {
    "epoch": 599,
    "batch_idx": 192,
    "loss": 0.22011518478393555,
    "learning_rate": 0.0007945276637710581,
    "timestep_mean": 449.0,
    "epsilon_norm": 4.27928352355957,
    "prediction_norm": 3.3441364765167236,
    "timestamp": "2025-07-14T12:33:54.734866"
  },
  {
    "epoch": 600,
    "batch_idx": 0,
    "loss": 0.4086081385612488,
    "learning_rate": 0.0007938926261462368,
    "timestep_mean": 443.84375,
    "epsilon_norm": 8.461177825927734,
    "prediction_norm": 7.175861835479736,
    "timestamp": "2025-07-14T12:33:54.740117"
  },
  {
    "epoch": 600,
    "batch_idx": 32,
    "loss": 0.3054317831993103,
    "learning_rate": 0.0007938926261462368,
    "timestep_mean": 528.75,
    "epsilon_norm": 8.918789863586426,
    "prediction_norm": 7.702774524688721,
    "timestamp": "2025-07-14T12:33:54.745013"
  },
  {
    "epoch": 600,
    "batch_idx": 64,
    "loss": 0.23234930634498596,
    "learning_rate": 0.0007938926261462368,
    "timestep_mean": 529.625,
    "epsilon_norm": 7.252071380615234,
    "prediction_norm": 6.126571178436279,
    "timestamp": "2025-07-14T12:33:54.750142"
  },
  {
    "epoch": 600,
    "batch_idx": 96,
    "loss": 0.24165970087051392,
    "learning_rate": 0.0007938926261462368,
    "timestep_mean": 461.0625,
    "epsilon_norm": 7.611763954162598,
    "prediction_norm": 6.470850944519043,
    "timestamp": "2025-07-14T12:33:54.755341"
  },
  {
    "epoch": 600,
    "batch_idx": 128,
    "loss": 0.3496386408805847,
    "learning_rate": 0.0007938926261462368,
    "timestep_mean": 529.4375,
    "epsilon_norm": 8.653106689453125,
    "prediction_norm": 7.837969779968262,
    "timestamp": "2025-07-14T12:33:54.760644"
  },
  {
    "epoch": 600,
    "batch_idx": 160,
    "loss": 0.46185487508773804,
    "learning_rate": 0.0007938926261462368,
    "timestep_mean": 502.25,
    "epsilon_norm": 8.999746322631836,
    "prediction_norm": 7.288740158081055,
    "timestamp": "2025-07-14T12:33:54.765856"
  },
  {
    "epoch": 600,
    "batch_idx": 192,
    "loss": 0.556061327457428,
    "learning_rate": 0.0007938926261462368,
    "timestep_mean": 352.625,
    "epsilon_norm": 4.904259204864502,
    "prediction_norm": 4.414536952972412,
    "timestamp": "2025-07-14T12:33:54.771372"
  },
  {
    "epoch": 601,
    "batch_idx": 0,
    "loss": 0.25435060262680054,
    "learning_rate": 0.0007932568633705753,
    "timestep_mean": 529.75,
    "epsilon_norm": 7.9456281661987305,
    "prediction_norm": 6.981668949127197,
    "timestamp": "2025-07-14T12:33:54.776812"
  },
  {
    "epoch": 601,
    "batch_idx": 32,
    "loss": 0.4003484845161438,
    "learning_rate": 0.0007932568633705753,
    "timestep_mean": 459.15625,
    "epsilon_norm": 7.961057662963867,
    "prediction_norm": 6.607573986053467,
    "timestamp": "2025-07-14T12:33:54.782000"
  },
  {
    "epoch": 601,
    "batch_idx": 64,
    "loss": 0.3053574562072754,
    "learning_rate": 0.0007932568633705753,
    "timestep_mean": 482.21875,
    "epsilon_norm": 8.071113586425781,
    "prediction_norm": 6.986741065979004,
    "timestamp": "2025-07-14T12:33:54.787221"
  },
  {
    "epoch": 601,
    "batch_idx": 96,
    "loss": 0.1589682698249817,
    "learning_rate": 0.0007932568633705753,
    "timestep_mean": 563.03125,
    "epsilon_norm": 7.645876407623291,
    "prediction_norm": 7.134720325469971,
    "timestamp": "2025-07-14T12:33:54.792472"
  },
  {
    "epoch": 601,
    "batch_idx": 128,
    "loss": 0.2738525867462158,
    "learning_rate": 0.0007932568633705753,
    "timestep_mean": 546.5625,
    "epsilon_norm": 8.171435356140137,
    "prediction_norm": 6.816774368286133,
    "timestamp": "2025-07-14T12:33:54.797592"
  },
  {
    "epoch": 601,
    "batch_idx": 160,
    "loss": 0.369479775428772,
    "learning_rate": 0.0007932568633705753,
    "timestep_mean": 481.125,
    "epsilon_norm": 8.572795867919922,
    "prediction_norm": 7.022752285003662,
    "timestamp": "2025-07-14T12:33:54.802830"
  },
  {
    "epoch": 601,
    "batch_idx": 192,
    "loss": 0.27341538667678833,
    "learning_rate": 0.0007932568633705753,
    "timestep_mean": 491.625,
    "epsilon_norm": 3.813300371170044,
    "prediction_norm": 2.7696104049682617,
    "timestamp": "2025-07-14T12:33:54.807848"
  },
  {
    "epoch": 602,
    "batch_idx": 0,
    "loss": 0.3390073776245117,
    "learning_rate": 0.0007926203770127553,
    "timestep_mean": 404.90625,
    "epsilon_norm": 7.100414752960205,
    "prediction_norm": 6.139986991882324,
    "timestamp": "2025-07-14T12:33:54.812988"
  },
  {
    "epoch": 602,
    "batch_idx": 32,
    "loss": 0.19739671051502228,
    "learning_rate": 0.0007926203770127553,
    "timestep_mean": 502.59375,
    "epsilon_norm": 7.484036922454834,
    "prediction_norm": 6.383188247680664,
    "timestamp": "2025-07-14T12:33:54.818326"
  },
  {
    "epoch": 602,
    "batch_idx": 64,
    "loss": 0.30935513973236084,
    "learning_rate": 0.0007926203770127553,
    "timestep_mean": 419.46875,
    "epsilon_norm": 8.526888847351074,
    "prediction_norm": 6.934589862823486,
    "timestamp": "2025-07-14T12:33:54.823622"
  },
  {
    "epoch": 602,
    "batch_idx": 96,
    "loss": 0.28912627696990967,
    "learning_rate": 0.0007926203770127553,
    "timestep_mean": 443.03125,
    "epsilon_norm": 7.8900675773620605,
    "prediction_norm": 7.225907325744629,
    "timestamp": "2025-07-14T12:33:54.828984"
  },
  {
    "epoch": 602,
    "batch_idx": 128,
    "loss": 0.1911277174949646,
    "learning_rate": 0.0007926203770127553,
    "timestep_mean": 615.03125,
    "epsilon_norm": 8.918596267700195,
    "prediction_norm": 8.594955444335938,
    "timestamp": "2025-07-14T12:33:54.834308"
  },
  {
    "epoch": 602,
    "batch_idx": 160,
    "loss": 0.26557114720344543,
    "learning_rate": 0.0007926203770127553,
    "timestep_mean": 498.6875,
    "epsilon_norm": 7.853238582611084,
    "prediction_norm": 7.161367416381836,
    "timestamp": "2025-07-14T12:33:54.839598"
  },
  {
    "epoch": 602,
    "batch_idx": 192,
    "loss": 0.43496888875961304,
    "learning_rate": 0.0007926203770127553,
    "timestep_mean": 518.0,
    "epsilon_norm": 3.3375678062438965,
    "prediction_norm": 3.10239839553833,
    "timestamp": "2025-07-14T12:33:54.844816"
  },
  {
    "epoch": 603,
    "batch_idx": 0,
    "loss": 0.21325817704200745,
    "learning_rate": 0.0007919831686432435,
    "timestep_mean": 508.8125,
    "epsilon_norm": 7.256530284881592,
    "prediction_norm": 6.633516788482666,
    "timestamp": "2025-07-14T12:33:54.850112"
  },
  {
    "epoch": 603,
    "batch_idx": 32,
    "loss": 0.21160683035850525,
    "learning_rate": 0.0007919831686432435,
    "timestep_mean": 507.25,
    "epsilon_norm": 8.6687593460083,
    "prediction_norm": 6.687179088592529,
    "timestamp": "2025-07-14T12:33:54.855274"
  },
  {
    "epoch": 603,
    "batch_idx": 64,
    "loss": 0.11805886030197144,
    "learning_rate": 0.0007919831686432435,
    "timestep_mean": 575.21875,
    "epsilon_norm": 7.50013542175293,
    "prediction_norm": 6.73386287689209,
    "timestamp": "2025-07-14T12:33:54.860504"
  },
  {
    "epoch": 603,
    "batch_idx": 96,
    "loss": 0.42548033595085144,
    "learning_rate": 0.0007919831686432435,
    "timestep_mean": 469.03125,
    "epsilon_norm": 7.986146926879883,
    "prediction_norm": 6.43992805480957,
    "timestamp": "2025-07-14T12:33:54.865617"
  },
  {
    "epoch": 603,
    "batch_idx": 128,
    "loss": 0.30522292852401733,
    "learning_rate": 0.0007919831686432435,
    "timestep_mean": 520.375,
    "epsilon_norm": 8.546314239501953,
    "prediction_norm": 7.073338508605957,
    "timestamp": "2025-07-14T12:33:54.870892"
  },
  {
    "epoch": 603,
    "batch_idx": 160,
    "loss": 0.24321402609348297,
    "learning_rate": 0.0007919831686432435,
    "timestep_mean": 521.625,
    "epsilon_norm": 7.433801174163818,
    "prediction_norm": 6.576996803283691,
    "timestamp": "2025-07-14T12:33:54.876061"
  },
  {
    "epoch": 603,
    "batch_idx": 192,
    "loss": 0.15142622590065002,
    "learning_rate": 0.0007919831686432435,
    "timestep_mean": 296.5,
    "epsilon_norm": 2.9564175605773926,
    "prediction_norm": 2.77070951461792,
    "timestamp": "2025-07-14T12:33:54.881254"
  },
  {
    "epoch": 604,
    "batch_idx": 0,
    "loss": 0.1292024403810501,
    "learning_rate": 0.0007913452398342883,
    "timestep_mean": 502.6875,
    "epsilon_norm": 6.8986310958862305,
    "prediction_norm": 6.7462239265441895,
    "timestamp": "2025-07-14T12:33:54.886523"
  },
  {
    "epoch": 604,
    "batch_idx": 32,
    "loss": 0.4945594072341919,
    "learning_rate": 0.0007913452398342883,
    "timestep_mean": 453.78125,
    "epsilon_norm": 7.900171756744385,
    "prediction_norm": 4.672304630279541,
    "timestamp": "2025-07-14T12:33:54.891751"
  },
  {
    "epoch": 604,
    "batch_idx": 64,
    "loss": 0.3565825819969177,
    "learning_rate": 0.0007913452398342883,
    "timestep_mean": 438.125,
    "epsilon_norm": 7.627243518829346,
    "prediction_norm": 5.513579845428467,
    "timestamp": "2025-07-14T12:33:54.896951"
  },
  {
    "epoch": 604,
    "batch_idx": 96,
    "loss": 0.3053266406059265,
    "learning_rate": 0.0007913452398342883,
    "timestep_mean": 437.25,
    "epsilon_norm": 8.067147254943848,
    "prediction_norm": 6.297876358032227,
    "timestamp": "2025-07-14T12:33:54.906473"
  },
  {
    "epoch": 604,
    "batch_idx": 128,
    "loss": 0.24098819494247437,
    "learning_rate": 0.0007913452398342883,
    "timestep_mean": 515.03125,
    "epsilon_norm": 8.78857707977295,
    "prediction_norm": 7.293684482574463,
    "timestamp": "2025-07-14T12:33:54.911589"
  },
  {
    "epoch": 604,
    "batch_idx": 160,
    "loss": 0.44287949800491333,
    "learning_rate": 0.0007913452398342883,
    "timestep_mean": 468.84375,
    "epsilon_norm": 8.557027816772461,
    "prediction_norm": 6.581034183502197,
    "timestamp": "2025-07-14T12:33:54.916823"
  },
  {
    "epoch": 604,
    "batch_idx": 192,
    "loss": 0.34206563234329224,
    "learning_rate": 0.0007913452398342883,
    "timestep_mean": 475.875,
    "epsilon_norm": 3.89345383644104,
    "prediction_norm": 3.1691734790802,
    "timestamp": "2025-07-14T12:33:54.922089"
  },
  {
    "epoch": 605,
    "batch_idx": 0,
    "loss": 0.37499892711639404,
    "learning_rate": 0.0007907065921599156,
    "timestep_mean": 514.15625,
    "epsilon_norm": 8.227828979492188,
    "prediction_norm": 6.212306976318359,
    "timestamp": "2025-07-14T12:33:54.927385"
  },
  {
    "epoch": 605,
    "batch_idx": 32,
    "loss": 0.3303370475769043,
    "learning_rate": 0.0007907065921599156,
    "timestep_mean": 489.09375,
    "epsilon_norm": 6.320742607116699,
    "prediction_norm": 4.940533638000488,
    "timestamp": "2025-07-14T12:33:54.932586"
  },
  {
    "epoch": 605,
    "batch_idx": 64,
    "loss": 0.20955230295658112,
    "learning_rate": 0.0007907065921599156,
    "timestep_mean": 538.3125,
    "epsilon_norm": 7.158727645874023,
    "prediction_norm": 6.092035293579102,
    "timestamp": "2025-07-14T12:33:54.937892"
  },
  {
    "epoch": 605,
    "batch_idx": 96,
    "loss": 0.2658945322036743,
    "learning_rate": 0.0007907065921599156,
    "timestep_mean": 466.4375,
    "epsilon_norm": 7.8710527420043945,
    "prediction_norm": 6.455411911010742,
    "timestamp": "2025-07-14T12:33:54.943030"
  },
  {
    "epoch": 605,
    "batch_idx": 128,
    "loss": 0.451109915971756,
    "learning_rate": 0.0007907065921599156,
    "timestep_mean": 467.15625,
    "epsilon_norm": 7.781208515167236,
    "prediction_norm": 5.176970958709717,
    "timestamp": "2025-07-14T12:33:54.948182"
  },
  {
    "epoch": 605,
    "batch_idx": 160,
    "loss": 0.20271652936935425,
    "learning_rate": 0.0007907065921599156,
    "timestep_mean": 530.34375,
    "epsilon_norm": 7.412535190582275,
    "prediction_norm": 6.637954235076904,
    "timestamp": "2025-07-14T12:33:54.953362"
  },
  {
    "epoch": 605,
    "batch_idx": 192,
    "loss": 0.1447390913963318,
    "learning_rate": 0.0007907065921599156,
    "timestep_mean": 539.75,
    "epsilon_norm": 3.286151647567749,
    "prediction_norm": 3.3734498023986816,
    "timestamp": "2025-07-14T12:33:54.958668"
  },
  {
    "epoch": 606,
    "batch_idx": 0,
    "loss": 0.23118339478969574,
    "learning_rate": 0.000790067227195925,
    "timestep_mean": 446.9375,
    "epsilon_norm": 7.107758522033691,
    "prediction_norm": 6.5723347663879395,
    "timestamp": "2025-07-14T12:33:54.963936"
  },
  {
    "epoch": 606,
    "batch_idx": 32,
    "loss": 0.45088550448417664,
    "learning_rate": 0.000790067227195925,
    "timestep_mean": 517.4375,
    "epsilon_norm": 10.436634063720703,
    "prediction_norm": 8.57888412475586,
    "timestamp": "2025-07-14T12:33:54.969022"
  },
  {
    "epoch": 606,
    "batch_idx": 64,
    "loss": 0.15575116872787476,
    "learning_rate": 0.000790067227195925,
    "timestep_mean": 552.03125,
    "epsilon_norm": 6.755055904388428,
    "prediction_norm": 5.746099948883057,
    "timestamp": "2025-07-14T12:33:54.974459"
  },
  {
    "epoch": 606,
    "batch_idx": 96,
    "loss": 0.23915566504001617,
    "learning_rate": 0.000790067227195925,
    "timestep_mean": 539.1875,
    "epsilon_norm": 8.208972930908203,
    "prediction_norm": 7.107175827026367,
    "timestamp": "2025-07-14T12:33:54.979571"
  },
  {
    "epoch": 606,
    "batch_idx": 128,
    "loss": 0.3669397532939911,
    "learning_rate": 0.000790067227195925,
    "timestep_mean": 372.125,
    "epsilon_norm": 8.055620193481445,
    "prediction_norm": 6.425983428955078,
    "timestamp": "2025-07-14T12:33:54.984567"
  },
  {
    "epoch": 606,
    "batch_idx": 160,
    "loss": 0.21375176310539246,
    "learning_rate": 0.000790067227195925,
    "timestep_mean": 507.0625,
    "epsilon_norm": 7.82171106338501,
    "prediction_norm": 6.874068737030029,
    "timestamp": "2025-07-14T12:33:54.990149"
  },
  {
    "epoch": 606,
    "batch_idx": 192,
    "loss": 0.07186806201934814,
    "learning_rate": 0.000790067227195925,
    "timestep_mean": 677.25,
    "epsilon_norm": 3.148437261581421,
    "prediction_norm": 3.0359294414520264,
    "timestamp": "2025-07-14T12:33:54.995688"
  },
  {
    "epoch": 607,
    "batch_idx": 0,
    "loss": 0.1867290735244751,
    "learning_rate": 0.000789427146519886,
    "timestep_mean": 542.46875,
    "epsilon_norm": 8.479857444763184,
    "prediction_norm": 7.642116069793701,
    "timestamp": "2025-07-14T12:33:55.001102"
  },
  {
    "epoch": 607,
    "batch_idx": 32,
    "loss": 0.1877480447292328,
    "learning_rate": 0.000789427146519886,
    "timestep_mean": 537.34375,
    "epsilon_norm": 7.526780128479004,
    "prediction_norm": 7.078815460205078,
    "timestamp": "2025-07-14T12:33:55.006586"
  },
  {
    "epoch": 607,
    "batch_idx": 64,
    "loss": 0.2804937958717346,
    "learning_rate": 0.000789427146519886,
    "timestep_mean": 513.78125,
    "epsilon_norm": 8.690750122070312,
    "prediction_norm": 8.006239891052246,
    "timestamp": "2025-07-14T12:33:55.011819"
  },
  {
    "epoch": 607,
    "batch_idx": 96,
    "loss": 0.3158456087112427,
    "learning_rate": 0.000789427146519886,
    "timestep_mean": 516.40625,
    "epsilon_norm": 8.540168762207031,
    "prediction_norm": 7.002948760986328,
    "timestamp": "2025-07-14T12:33:55.017102"
  },
  {
    "epoch": 607,
    "batch_idx": 128,
    "loss": 0.26131612062454224,
    "learning_rate": 0.000789427146519886,
    "timestep_mean": 445.21875,
    "epsilon_norm": 8.048184394836426,
    "prediction_norm": 6.503544807434082,
    "timestamp": "2025-07-14T12:33:55.022114"
  },
  {
    "epoch": 607,
    "batch_idx": 160,
    "loss": 0.3803409934043884,
    "learning_rate": 0.000789427146519886,
    "timestep_mean": 429.28125,
    "epsilon_norm": 7.357088565826416,
    "prediction_norm": 5.795218467712402,
    "timestamp": "2025-07-14T12:33:55.027422"
  },
  {
    "epoch": 607,
    "batch_idx": 192,
    "loss": 0.07202187925577164,
    "learning_rate": 0.000789427146519886,
    "timestep_mean": 538.375,
    "epsilon_norm": 3.866682529449463,
    "prediction_norm": 3.7625396251678467,
    "timestamp": "2025-07-14T12:33:55.032738"
  },
  {
    "epoch": 608,
    "batch_idx": 0,
    "loss": 0.23153971135616302,
    "learning_rate": 0.000788786351711134,
    "timestep_mean": 553.46875,
    "epsilon_norm": 6.928187370300293,
    "prediction_norm": 5.854788303375244,
    "timestamp": "2025-07-14T12:33:55.037962"
  },
  {
    "epoch": 608,
    "batch_idx": 32,
    "loss": 0.14392536878585815,
    "learning_rate": 0.000788786351711134,
    "timestep_mean": 562.65625,
    "epsilon_norm": 7.612607955932617,
    "prediction_norm": 6.70250940322876,
    "timestamp": "2025-07-14T12:33:55.043169"
  },
  {
    "epoch": 608,
    "batch_idx": 64,
    "loss": 0.17559954524040222,
    "learning_rate": 0.000788786351711134,
    "timestep_mean": 545.34375,
    "epsilon_norm": 7.264011859893799,
    "prediction_norm": 6.71235466003418,
    "timestamp": "2025-07-14T12:33:55.048444"
  },
  {
    "epoch": 608,
    "batch_idx": 96,
    "loss": 0.290169894695282,
    "learning_rate": 0.000788786351711134,
    "timestep_mean": 470.0625,
    "epsilon_norm": 7.862423896789551,
    "prediction_norm": 6.833017349243164,
    "timestamp": "2025-07-14T12:33:55.053743"
  },
  {
    "epoch": 608,
    "batch_idx": 128,
    "loss": 0.2720085680484772,
    "learning_rate": 0.000788786351711134,
    "timestep_mean": 532.125,
    "epsilon_norm": 10.077635765075684,
    "prediction_norm": 8.459218978881836,
    "timestamp": "2025-07-14T12:33:55.058963"
  },
  {
    "epoch": 608,
    "batch_idx": 160,
    "loss": 0.1486334204673767,
    "learning_rate": 0.000788786351711134,
    "timestep_mean": 550.03125,
    "epsilon_norm": 7.636003494262695,
    "prediction_norm": 6.994219779968262,
    "timestamp": "2025-07-14T12:33:55.063975"
  },
  {
    "epoch": 608,
    "batch_idx": 192,
    "loss": 0.28868865966796875,
    "learning_rate": 0.000788786351711134,
    "timestep_mean": 526.75,
    "epsilon_norm": 3.5219664573669434,
    "prediction_norm": 3.06540846824646,
    "timestamp": "2025-07-14T12:33:55.069158"
  },
  {
    "epoch": 609,
    "batch_idx": 0,
    "loss": 0.22854246199131012,
    "learning_rate": 0.0007881448443507667,
    "timestep_mean": 486.875,
    "epsilon_norm": 7.451025009155273,
    "prediction_norm": 6.600052356719971,
    "timestamp": "2025-07-14T12:33:55.074363"
  },
  {
    "epoch": 609,
    "batch_idx": 32,
    "loss": 0.37539124488830566,
    "learning_rate": 0.0007881448443507667,
    "timestep_mean": 517.3125,
    "epsilon_norm": 8.080249786376953,
    "prediction_norm": 6.290846824645996,
    "timestamp": "2025-07-14T12:33:55.079429"
  },
  {
    "epoch": 609,
    "batch_idx": 64,
    "loss": 0.3575126826763153,
    "learning_rate": 0.0007881448443507667,
    "timestep_mean": 481.75,
    "epsilon_norm": 8.509963989257812,
    "prediction_norm": 6.8585028648376465,
    "timestamp": "2025-07-14T12:33:55.084608"
  },
  {
    "epoch": 609,
    "batch_idx": 96,
    "loss": 0.2535092234611511,
    "learning_rate": 0.0007881448443507667,
    "timestep_mean": 538.21875,
    "epsilon_norm": 9.537671089172363,
    "prediction_norm": 8.281283378601074,
    "timestamp": "2025-07-14T12:33:55.089822"
  },
  {
    "epoch": 609,
    "batch_idx": 128,
    "loss": 0.26882702112197876,
    "learning_rate": 0.0007881448443507667,
    "timestep_mean": 544.59375,
    "epsilon_norm": 8.42625617980957,
    "prediction_norm": 7.17034387588501,
    "timestamp": "2025-07-14T12:33:55.094848"
  },
  {
    "epoch": 609,
    "batch_idx": 160,
    "loss": 0.20479252934455872,
    "learning_rate": 0.0007881448443507667,
    "timestep_mean": 539.90625,
    "epsilon_norm": 7.705943584442139,
    "prediction_norm": 6.901565074920654,
    "timestamp": "2025-07-14T12:33:55.099721"
  },
  {
    "epoch": 609,
    "batch_idx": 192,
    "loss": 0.25500115752220154,
    "learning_rate": 0.0007881448443507667,
    "timestep_mean": 593.75,
    "epsilon_norm": 4.394140720367432,
    "prediction_norm": 2.9546959400177,
    "timestamp": "2025-07-14T12:33:55.104687"
  },
  {
    "epoch": 610,
    "batch_idx": 0,
    "loss": 0.10763882100582123,
    "learning_rate": 0.0007875026260216396,
    "timestep_mean": 598.90625,
    "epsilon_norm": 9.32903003692627,
    "prediction_norm": 8.653157234191895,
    "timestamp": "2025-07-14T12:33:55.109561"
  },
  {
    "epoch": 610,
    "batch_idx": 32,
    "loss": 0.2987273931503296,
    "learning_rate": 0.0007875026260216396,
    "timestep_mean": 538.53125,
    "epsilon_norm": 7.837971210479736,
    "prediction_norm": 7.080177307128906,
    "timestamp": "2025-07-14T12:33:55.114467"
  },
  {
    "epoch": 610,
    "batch_idx": 64,
    "loss": 0.191432923078537,
    "learning_rate": 0.0007875026260216396,
    "timestep_mean": 556.8125,
    "epsilon_norm": 7.564914703369141,
    "prediction_norm": 7.067963123321533,
    "timestamp": "2025-07-14T12:33:55.119471"
  },
  {
    "epoch": 610,
    "batch_idx": 96,
    "loss": 0.18954387307167053,
    "learning_rate": 0.0007875026260216396,
    "timestep_mean": 534.21875,
    "epsilon_norm": 7.562222003936768,
    "prediction_norm": 6.512276649475098,
    "timestamp": "2025-07-14T12:33:55.124338"
  },
  {
    "epoch": 610,
    "batch_idx": 128,
    "loss": 0.3710859417915344,
    "learning_rate": 0.0007875026260216396,
    "timestep_mean": 530.0,
    "epsilon_norm": 8.23963451385498,
    "prediction_norm": 7.356166362762451,
    "timestamp": "2025-07-14T12:33:55.129334"
  },
  {
    "epoch": 610,
    "batch_idx": 160,
    "loss": 0.5802881717681885,
    "learning_rate": 0.0007875026260216396,
    "timestep_mean": 489.125,
    "epsilon_norm": 8.964024543762207,
    "prediction_norm": 6.937014579772949,
    "timestamp": "2025-07-14T12:33:55.134258"
  },
  {
    "epoch": 610,
    "batch_idx": 192,
    "loss": 0.48952269554138184,
    "learning_rate": 0.0007875026260216396,
    "timestep_mean": 394.375,
    "epsilon_norm": 4.368307590484619,
    "prediction_norm": 4.0234503746032715,
    "timestamp": "2025-07-14T12:33:55.139461"
  },
  {
    "epoch": 611,
    "batch_idx": 0,
    "loss": 0.31817957758903503,
    "learning_rate": 0.0007868596983083625,
    "timestep_mean": 502.0625,
    "epsilon_norm": 7.7604079246521,
    "prediction_norm": 6.353594779968262,
    "timestamp": "2025-07-14T12:33:55.144453"
  },
  {
    "epoch": 611,
    "batch_idx": 32,
    "loss": 0.24236589670181274,
    "learning_rate": 0.0007868596983083625,
    "timestep_mean": 548.125,
    "epsilon_norm": 8.467033386230469,
    "prediction_norm": 7.60391902923584,
    "timestamp": "2025-07-14T12:33:55.149298"
  },
  {
    "epoch": 611,
    "batch_idx": 64,
    "loss": 0.21178144216537476,
    "learning_rate": 0.0007868596983083625,
    "timestep_mean": 505.46875,
    "epsilon_norm": 8.917804718017578,
    "prediction_norm": 8.548151969909668,
    "timestamp": "2025-07-14T12:33:55.154339"
  },
  {
    "epoch": 611,
    "batch_idx": 96,
    "loss": 0.1399340033531189,
    "learning_rate": 0.0007868596983083625,
    "timestep_mean": 550.09375,
    "epsilon_norm": 7.235214710235596,
    "prediction_norm": 6.9634690284729,
    "timestamp": "2025-07-14T12:33:55.159197"
  },
  {
    "epoch": 611,
    "batch_idx": 128,
    "loss": 0.21885982155799866,
    "learning_rate": 0.0007868596983083625,
    "timestep_mean": 548.25,
    "epsilon_norm": 8.294015884399414,
    "prediction_norm": 7.570694923400879,
    "timestamp": "2025-07-14T12:33:55.164087"
  },
  {
    "epoch": 611,
    "batch_idx": 160,
    "loss": 0.38886624574661255,
    "learning_rate": 0.0007868596983083625,
    "timestep_mean": 494.125,
    "epsilon_norm": 7.946965217590332,
    "prediction_norm": 5.743090629577637,
    "timestamp": "2025-07-14T12:33:55.168958"
  },
  {
    "epoch": 611,
    "batch_idx": 192,
    "loss": 0.10934537649154663,
    "learning_rate": 0.0007868596983083625,
    "timestep_mean": 677.5,
    "epsilon_norm": 3.513178586959839,
    "prediction_norm": 3.425208806991577,
    "timestamp": "2025-07-14T12:33:55.173769"
  },
  {
    "epoch": 612,
    "batch_idx": 0,
    "loss": 0.24656382203102112,
    "learning_rate": 0.0007862160627972957,
    "timestep_mean": 495.15625,
    "epsilon_norm": 6.665102958679199,
    "prediction_norm": 6.0909504890441895,
    "timestamp": "2025-07-14T12:33:55.178697"
  },
  {
    "epoch": 612,
    "batch_idx": 32,
    "loss": 0.34801381826400757,
    "learning_rate": 0.0007862160627972957,
    "timestep_mean": 507.96875,
    "epsilon_norm": 8.628400802612305,
    "prediction_norm": 7.408700942993164,
    "timestamp": "2025-07-14T12:33:55.183668"
  },
  {
    "epoch": 612,
    "batch_idx": 64,
    "loss": 0.3394480347633362,
    "learning_rate": 0.0007862160627972957,
    "timestep_mean": 590.71875,
    "epsilon_norm": 9.146567344665527,
    "prediction_norm": 7.5339813232421875,
    "timestamp": "2025-07-14T12:33:55.188526"
  },
  {
    "epoch": 612,
    "batch_idx": 96,
    "loss": 0.2597561180591583,
    "learning_rate": 0.0007862160627972957,
    "timestep_mean": 532.90625,
    "epsilon_norm": 7.619558811187744,
    "prediction_norm": 6.940369606018066,
    "timestamp": "2025-07-14T12:33:55.193386"
  },
  {
    "epoch": 612,
    "batch_idx": 128,
    "loss": 0.3311445116996765,
    "learning_rate": 0.0007862160627972957,
    "timestep_mean": 485.0,
    "epsilon_norm": 7.77280855178833,
    "prediction_norm": 6.290544033050537,
    "timestamp": "2025-07-14T12:33:55.198266"
  },
  {
    "epoch": 612,
    "batch_idx": 160,
    "loss": 0.2206854522228241,
    "learning_rate": 0.0007862160627972957,
    "timestep_mean": 487.8125,
    "epsilon_norm": 7.784180164337158,
    "prediction_norm": 6.954102993011475,
    "timestamp": "2025-07-14T12:33:55.203153"
  },
  {
    "epoch": 612,
    "batch_idx": 192,
    "loss": 0.2991976737976074,
    "learning_rate": 0.0007862160627972957,
    "timestep_mean": 580.625,
    "epsilon_norm": 4.361965179443359,
    "prediction_norm": 3.924088954925537,
    "timestamp": "2025-07-14T12:33:55.208036"
  },
  {
    "epoch": 613,
    "batch_idx": 0,
    "loss": 0.27105170488357544,
    "learning_rate": 0.0007855717210765459,
    "timestep_mean": 561.375,
    "epsilon_norm": 8.423852920532227,
    "prediction_norm": 7.160859107971191,
    "timestamp": "2025-07-14T12:33:55.213054"
  },
  {
    "epoch": 613,
    "batch_idx": 32,
    "loss": 0.6289896368980408,
    "learning_rate": 0.0007855717210765459,
    "timestep_mean": 471.84375,
    "epsilon_norm": 7.91322135925293,
    "prediction_norm": 5.2394256591796875,
    "timestamp": "2025-07-14T12:33:55.217852"
  },
  {
    "epoch": 613,
    "batch_idx": 64,
    "loss": 0.30986595153808594,
    "learning_rate": 0.0007855717210765459,
    "timestep_mean": 496.15625,
    "epsilon_norm": 7.301361083984375,
    "prediction_norm": 5.97946834564209,
    "timestamp": "2025-07-14T12:33:55.222747"
  },
  {
    "epoch": 613,
    "batch_idx": 96,
    "loss": 0.13669759035110474,
    "learning_rate": 0.0007855717210765459,
    "timestep_mean": 510.875,
    "epsilon_norm": 7.487174987792969,
    "prediction_norm": 7.249037742614746,
    "timestamp": "2025-07-14T12:33:55.227645"
  },
  {
    "epoch": 613,
    "batch_idx": 128,
    "loss": 0.16816338896751404,
    "learning_rate": 0.0007855717210765459,
    "timestep_mean": 600.21875,
    "epsilon_norm": 6.996500492095947,
    "prediction_norm": 6.4472808837890625,
    "timestamp": "2025-07-14T12:33:55.232483"
  },
  {
    "epoch": 613,
    "batch_idx": 160,
    "loss": 0.33292078971862793,
    "learning_rate": 0.0007855717210765459,
    "timestep_mean": 458.4375,
    "epsilon_norm": 7.825644493103027,
    "prediction_norm": 6.128693580627441,
    "timestamp": "2025-07-14T12:33:55.237416"
  },
  {
    "epoch": 613,
    "batch_idx": 192,
    "loss": 0.19287462532520294,
    "learning_rate": 0.0007855717210765459,
    "timestep_mean": 613.5,
    "epsilon_norm": 3.925504684448242,
    "prediction_norm": 3.370361328125,
    "timestamp": "2025-07-14T12:33:55.242339"
  },
  {
    "epoch": 614,
    "batch_idx": 0,
    "loss": 0.29234403371810913,
    "learning_rate": 0.0007849266747359622,
    "timestep_mean": 498.65625,
    "epsilon_norm": 8.420679092407227,
    "prediction_norm": 6.33388090133667,
    "timestamp": "2025-07-14T12:33:55.247338"
  },
  {
    "epoch": 614,
    "batch_idx": 32,
    "loss": 0.2963094711303711,
    "learning_rate": 0.0007849266747359622,
    "timestep_mean": 467.65625,
    "epsilon_norm": 7.202146053314209,
    "prediction_norm": 5.3981194496154785,
    "timestamp": "2025-07-14T12:33:55.252233"
  },
  {
    "epoch": 614,
    "batch_idx": 64,
    "loss": 0.2894558906555176,
    "learning_rate": 0.0007849266747359622,
    "timestep_mean": 447.09375,
    "epsilon_norm": 7.975290775299072,
    "prediction_norm": 6.328225612640381,
    "timestamp": "2025-07-14T12:33:55.257124"
  },
  {
    "epoch": 614,
    "batch_idx": 96,
    "loss": 0.3916932940483093,
    "learning_rate": 0.0007849266747359622,
    "timestep_mean": 406.25,
    "epsilon_norm": 8.674825668334961,
    "prediction_norm": 7.230764389038086,
    "timestamp": "2025-07-14T12:33:55.262016"
  },
  {
    "epoch": 614,
    "batch_idx": 128,
    "loss": 0.20619583129882812,
    "learning_rate": 0.0007849266747359622,
    "timestep_mean": 476.90625,
    "epsilon_norm": 8.679937362670898,
    "prediction_norm": 7.260034084320068,
    "timestamp": "2025-07-14T12:33:55.267000"
  },
  {
    "epoch": 614,
    "batch_idx": 160,
    "loss": 0.2049328237771988,
    "learning_rate": 0.0007849266747359622,
    "timestep_mean": 512.125,
    "epsilon_norm": 7.741691589355469,
    "prediction_norm": 6.970935821533203,
    "timestamp": "2025-07-14T12:33:55.271845"
  },
  {
    "epoch": 614,
    "batch_idx": 192,
    "loss": 0.14514346420764923,
    "learning_rate": 0.0007849266747359622,
    "timestep_mean": 535.5,
    "epsilon_norm": 4.1372151374816895,
    "prediction_norm": 3.210872173309326,
    "timestamp": "2025-07-14T12:33:55.276723"
  },
  {
    "epoch": 615,
    "batch_idx": 0,
    "loss": 0.2443757951259613,
    "learning_rate": 0.0007842809253671323,
    "timestep_mean": 511.53125,
    "epsilon_norm": 7.349569797515869,
    "prediction_norm": 6.336715221405029,
    "timestamp": "2025-07-14T12:33:55.281597"
  },
  {
    "epoch": 615,
    "batch_idx": 32,
    "loss": 0.1585022211074829,
    "learning_rate": 0.0007842809253671323,
    "timestep_mean": 566.96875,
    "epsilon_norm": 6.66223669052124,
    "prediction_norm": 5.697731971740723,
    "timestamp": "2025-07-14T12:33:55.286523"
  },
  {
    "epoch": 615,
    "batch_idx": 64,
    "loss": 0.20916981995105743,
    "learning_rate": 0.0007842809253671323,
    "timestep_mean": 537.1875,
    "epsilon_norm": 7.9100117683410645,
    "prediction_norm": 6.997907638549805,
    "timestamp": "2025-07-14T12:33:55.291559"
  },
  {
    "epoch": 615,
    "batch_idx": 96,
    "loss": 0.2757253050804138,
    "learning_rate": 0.0007842809253671323,
    "timestep_mean": 544.78125,
    "epsilon_norm": 6.940810680389404,
    "prediction_norm": 5.2752299308776855,
    "timestamp": "2025-07-14T12:33:55.296469"
  },
  {
    "epoch": 615,
    "batch_idx": 128,
    "loss": 0.28149333596229553,
    "learning_rate": 0.0007842809253671323,
    "timestep_mean": 535.96875,
    "epsilon_norm": 7.966217517852783,
    "prediction_norm": 7.623057842254639,
    "timestamp": "2025-07-14T12:33:55.301354"
  },
  {
    "epoch": 615,
    "batch_idx": 160,
    "loss": 0.20362520217895508,
    "learning_rate": 0.0007842809253671323,
    "timestep_mean": 477.84375,
    "epsilon_norm": 7.264965057373047,
    "prediction_norm": 5.977475166320801,
    "timestamp": "2025-07-14T12:33:55.306279"
  },
  {
    "epoch": 615,
    "batch_idx": 192,
    "loss": 0.13630300760269165,
    "learning_rate": 0.0007842809253671323,
    "timestep_mean": 567.125,
    "epsilon_norm": 3.9598097801208496,
    "prediction_norm": 3.556105852127075,
    "timestamp": "2025-07-14T12:33:55.311193"
  },
  {
    "epoch": 616,
    "batch_idx": 0,
    "loss": 0.3161156177520752,
    "learning_rate": 0.0007836344745633786,
    "timestep_mean": 487.28125,
    "epsilon_norm": 7.766277313232422,
    "prediction_norm": 5.936014175415039,
    "timestamp": "2025-07-14T12:33:55.316113"
  },
  {
    "epoch": 616,
    "batch_idx": 32,
    "loss": 0.23867538571357727,
    "learning_rate": 0.0007836344745633786,
    "timestep_mean": 464.84375,
    "epsilon_norm": 8.481680870056152,
    "prediction_norm": 7.017793655395508,
    "timestamp": "2025-07-14T12:33:55.321170"
  },
  {
    "epoch": 616,
    "batch_idx": 64,
    "loss": 0.20380061864852905,
    "learning_rate": 0.0007836344745633786,
    "timestep_mean": 568.03125,
    "epsilon_norm": 6.613635063171387,
    "prediction_norm": 6.016664981842041,
    "timestamp": "2025-07-14T12:33:55.326174"
  },
  {
    "epoch": 616,
    "batch_idx": 96,
    "loss": 0.2077314257621765,
    "learning_rate": 0.0007836344745633786,
    "timestep_mean": 504.1875,
    "epsilon_norm": 8.516359329223633,
    "prediction_norm": 7.841699600219727,
    "timestamp": "2025-07-14T12:33:55.331135"
  },
  {
    "epoch": 616,
    "batch_idx": 128,
    "loss": 0.349118173122406,
    "learning_rate": 0.0007836344745633786,
    "timestep_mean": 524.84375,
    "epsilon_norm": 8.303243637084961,
    "prediction_norm": 6.665735721588135,
    "timestamp": "2025-07-14T12:33:55.336088"
  },
  {
    "epoch": 616,
    "batch_idx": 160,
    "loss": 0.1941039115190506,
    "learning_rate": 0.0007836344745633786,
    "timestep_mean": 575.0625,
    "epsilon_norm": 8.082134246826172,
    "prediction_norm": 7.6923346519470215,
    "timestamp": "2025-07-14T12:33:55.341019"
  },
  {
    "epoch": 616,
    "batch_idx": 192,
    "loss": 0.33697018027305603,
    "learning_rate": 0.0007836344745633786,
    "timestep_mean": 446.625,
    "epsilon_norm": 4.205406665802002,
    "prediction_norm": 3.802870273590088,
    "timestamp": "2025-07-14T12:33:55.346035"
  },
  {
    "epoch": 617,
    "batch_idx": 0,
    "loss": 0.24965131282806396,
    "learning_rate": 0.000782987323919754,
    "timestep_mean": 514.09375,
    "epsilon_norm": 6.5673747062683105,
    "prediction_norm": 5.353375434875488,
    "timestamp": "2025-07-14T12:33:55.350951"
  },
  {
    "epoch": 617,
    "batch_idx": 32,
    "loss": 0.27327919006347656,
    "learning_rate": 0.000782987323919754,
    "timestep_mean": 479.5,
    "epsilon_norm": 7.205744743347168,
    "prediction_norm": 6.11168098449707,
    "timestamp": "2025-07-14T12:33:55.355826"
  },
  {
    "epoch": 617,
    "batch_idx": 64,
    "loss": 0.24490855634212494,
    "learning_rate": 0.000782987323919754,
    "timestep_mean": 475.03125,
    "epsilon_norm": 8.360203742980957,
    "prediction_norm": 7.200743675231934,
    "timestamp": "2025-07-14T12:33:55.360669"
  },
  {
    "epoch": 617,
    "batch_idx": 96,
    "loss": 0.23230496048927307,
    "learning_rate": 0.000782987323919754,
    "timestep_mean": 575.15625,
    "epsilon_norm": 7.472414493560791,
    "prediction_norm": 6.315943241119385,
    "timestamp": "2025-07-14T12:33:55.365572"
  },
  {
    "epoch": 617,
    "batch_idx": 128,
    "loss": 0.3497897982597351,
    "learning_rate": 0.000782987323919754,
    "timestep_mean": 476.0625,
    "epsilon_norm": 8.332833290100098,
    "prediction_norm": 7.04056453704834,
    "timestamp": "2025-07-14T12:33:55.370516"
  },
  {
    "epoch": 617,
    "batch_idx": 160,
    "loss": 0.20974823832511902,
    "learning_rate": 0.000782987323919754,
    "timestep_mean": 491.0,
    "epsilon_norm": 7.033666610717773,
    "prediction_norm": 5.857180118560791,
    "timestamp": "2025-07-14T12:33:55.375503"
  },
  {
    "epoch": 617,
    "batch_idx": 192,
    "loss": 0.28169679641723633,
    "learning_rate": 0.000782987323919754,
    "timestep_mean": 502.75,
    "epsilon_norm": 4.428422927856445,
    "prediction_norm": 4.00337553024292,
    "timestamp": "2025-07-14T12:33:55.380389"
  },
  {
    "epoch": 618,
    "batch_idx": 0,
    "loss": 0.12452998757362366,
    "learning_rate": 0.0007823394750330388,
    "timestep_mean": 548.21875,
    "epsilon_norm": 7.600606441497803,
    "prediction_norm": 6.912797927856445,
    "timestamp": "2025-07-14T12:33:55.385307"
  },
  {
    "epoch": 618,
    "batch_idx": 32,
    "loss": 0.08473207801580429,
    "learning_rate": 0.0007823394750330388,
    "timestep_mean": 556.375,
    "epsilon_norm": 6.966324329376221,
    "prediction_norm": 6.619941711425781,
    "timestamp": "2025-07-14T12:33:55.390175"
  },
  {
    "epoch": 618,
    "batch_idx": 64,
    "loss": 0.5870987176895142,
    "learning_rate": 0.0007823394750330388,
    "timestep_mean": 331.375,
    "epsilon_norm": 8.095146179199219,
    "prediction_norm": 5.758589744567871,
    "timestamp": "2025-07-14T12:33:55.395069"
  },
  {
    "epoch": 618,
    "batch_idx": 96,
    "loss": 0.4415152072906494,
    "learning_rate": 0.0007823394750330388,
    "timestep_mean": 492.3125,
    "epsilon_norm": 7.706903457641602,
    "prediction_norm": 5.836564540863037,
    "timestamp": "2025-07-14T12:33:55.400036"
  },
  {
    "epoch": 618,
    "batch_idx": 128,
    "loss": 0.1804654449224472,
    "learning_rate": 0.0007823394750330388,
    "timestep_mean": 515.0625,
    "epsilon_norm": 8.18445873260498,
    "prediction_norm": 7.309074878692627,
    "timestamp": "2025-07-14T12:33:55.404903"
  },
  {
    "epoch": 618,
    "batch_idx": 160,
    "loss": 0.20212844014167786,
    "learning_rate": 0.0007823394750330388,
    "timestep_mean": 580.09375,
    "epsilon_norm": 7.257700443267822,
    "prediction_norm": 6.486915111541748,
    "timestamp": "2025-07-14T12:33:55.409809"
  },
  {
    "epoch": 618,
    "batch_idx": 192,
    "loss": 0.20556238293647766,
    "learning_rate": 0.0007823394750330388,
    "timestep_mean": 443.875,
    "epsilon_norm": 3.972750186920166,
    "prediction_norm": 3.827641010284424,
    "timestamp": "2025-07-14T12:33:55.414678"
  },
  {
    "epoch": 619,
    "batch_idx": 0,
    "loss": 0.22011712193489075,
    "learning_rate": 0.0007816909295017354,
    "timestep_mean": 551.875,
    "epsilon_norm": 8.077543258666992,
    "prediction_norm": 7.5805583000183105,
    "timestamp": "2025-07-14T12:33:55.419806"
  },
  {
    "epoch": 619,
    "batch_idx": 32,
    "loss": 0.12725402414798737,
    "learning_rate": 0.0007816909295017354,
    "timestep_mean": 612.375,
    "epsilon_norm": 8.800785064697266,
    "prediction_norm": 8.301945686340332,
    "timestamp": "2025-07-14T12:33:55.424836"
  },
  {
    "epoch": 619,
    "batch_idx": 64,
    "loss": 0.26717042922973633,
    "learning_rate": 0.0007816909295017354,
    "timestep_mean": 506.28125,
    "epsilon_norm": 8.502039909362793,
    "prediction_norm": 6.931204795837402,
    "timestamp": "2025-07-14T12:33:55.429894"
  },
  {
    "epoch": 619,
    "batch_idx": 96,
    "loss": 0.32645314931869507,
    "learning_rate": 0.0007816909295017354,
    "timestep_mean": 481.71875,
    "epsilon_norm": 7.6192169189453125,
    "prediction_norm": 6.010093688964844,
    "timestamp": "2025-07-14T12:33:55.434836"
  },
  {
    "epoch": 619,
    "batch_idx": 128,
    "loss": 0.19768302142620087,
    "learning_rate": 0.0007816909295017354,
    "timestep_mean": 473.1875,
    "epsilon_norm": 7.511440753936768,
    "prediction_norm": 6.643062114715576,
    "timestamp": "2025-07-14T12:33:55.439704"
  },
  {
    "epoch": 619,
    "batch_idx": 160,
    "loss": 0.3836963176727295,
    "learning_rate": 0.0007816909295017354,
    "timestep_mean": 495.125,
    "epsilon_norm": 8.622896194458008,
    "prediction_norm": 6.8022050857543945,
    "timestamp": "2025-07-14T12:33:55.444583"
  },
  {
    "epoch": 619,
    "batch_idx": 192,
    "loss": 0.12659981846809387,
    "learning_rate": 0.0007816909295017354,
    "timestep_mean": 615.375,
    "epsilon_norm": 3.263603448867798,
    "prediction_norm": 2.8255727291107178,
    "timestamp": "2025-07-14T12:33:55.449655"
  },
  {
    "epoch": 620,
    "batch_idx": 0,
    "loss": 0.20381978154182434,
    "learning_rate": 0.0007810416889260655,
    "timestep_mean": 438.75,
    "epsilon_norm": 7.978601932525635,
    "prediction_norm": 6.663152694702148,
    "timestamp": "2025-07-14T12:33:55.454688"
  },
  {
    "epoch": 620,
    "batch_idx": 32,
    "loss": 0.20441538095474243,
    "learning_rate": 0.0007810416889260655,
    "timestep_mean": 511.125,
    "epsilon_norm": 8.096878051757812,
    "prediction_norm": 7.422311782836914,
    "timestamp": "2025-07-14T12:33:55.459560"
  },
  {
    "epoch": 620,
    "batch_idx": 64,
    "loss": 0.24488970637321472,
    "learning_rate": 0.0007810416889260655,
    "timestep_mean": 445.125,
    "epsilon_norm": 7.399048805236816,
    "prediction_norm": 6.2039008140563965,
    "timestamp": "2025-07-14T12:33:55.464363"
  },
  {
    "epoch": 620,
    "batch_idx": 96,
    "loss": 0.28849494457244873,
    "learning_rate": 0.0007810416889260655,
    "timestep_mean": 420.375,
    "epsilon_norm": 8.0128755569458,
    "prediction_norm": 6.272402286529541,
    "timestamp": "2025-07-14T12:33:55.469213"
  },
  {
    "epoch": 620,
    "batch_idx": 128,
    "loss": 0.3329898715019226,
    "learning_rate": 0.0007810416889260655,
    "timestep_mean": 483.03125,
    "epsilon_norm": 9.456500053405762,
    "prediction_norm": 7.891369819641113,
    "timestamp": "2025-07-14T12:33:55.474082"
  },
  {
    "epoch": 620,
    "batch_idx": 160,
    "loss": 0.1805921196937561,
    "learning_rate": 0.0007810416889260655,
    "timestep_mean": 482.625,
    "epsilon_norm": 7.479738235473633,
    "prediction_norm": 7.317142486572266,
    "timestamp": "2025-07-14T12:33:55.478951"
  },
  {
    "epoch": 620,
    "batch_idx": 192,
    "loss": 0.12916019558906555,
    "learning_rate": 0.0007810416889260655,
    "timestep_mean": 544.625,
    "epsilon_norm": 3.1421542167663574,
    "prediction_norm": 2.8808727264404297,
    "timestamp": "2025-07-14T12:33:55.483973"
  },
  {
    "epoch": 621,
    "batch_idx": 0,
    "loss": 0.14518359303474426,
    "learning_rate": 0.0007803917549079657,
    "timestep_mean": 525.96875,
    "epsilon_norm": 8.109651565551758,
    "prediction_norm": 7.2100958824157715,
    "timestamp": "2025-07-14T12:33:55.488990"
  },
  {
    "epoch": 621,
    "batch_idx": 32,
    "loss": 0.25204700231552124,
    "learning_rate": 0.0007803917549079657,
    "timestep_mean": 498.875,
    "epsilon_norm": 7.650232791900635,
    "prediction_norm": 6.681551933288574,
    "timestamp": "2025-07-14T12:33:55.493869"
  },
  {
    "epoch": 621,
    "batch_idx": 64,
    "loss": 0.6499111652374268,
    "learning_rate": 0.0007803917549079657,
    "timestep_mean": 420.3125,
    "epsilon_norm": 8.086498260498047,
    "prediction_norm": 5.755841255187988,
    "timestamp": "2025-07-14T12:33:55.498748"
  },
  {
    "epoch": 621,
    "batch_idx": 96,
    "loss": 0.21141888201236725,
    "learning_rate": 0.0007803917549079657,
    "timestep_mean": 592.0625,
    "epsilon_norm": 8.326367378234863,
    "prediction_norm": 6.9785308837890625,
    "timestamp": "2025-07-14T12:33:55.509251"
  },
  {
    "epoch": 621,
    "batch_idx": 128,
    "loss": 0.19979825615882874,
    "learning_rate": 0.0007803917549079657,
    "timestep_mean": 505.53125,
    "epsilon_norm": 7.694628715515137,
    "prediction_norm": 6.6018524169921875,
    "timestamp": "2025-07-14T12:33:55.514069"
  },
  {
    "epoch": 621,
    "batch_idx": 160,
    "loss": 0.33830493688583374,
    "learning_rate": 0.0007803917549079657,
    "timestep_mean": 423.09375,
    "epsilon_norm": 8.344527244567871,
    "prediction_norm": 6.9652838706970215,
    "timestamp": "2025-07-14T12:33:55.518856"
  },
  {
    "epoch": 621,
    "batch_idx": 192,
    "loss": 0.1702091097831726,
    "learning_rate": 0.0007803917549079657,
    "timestep_mean": 625.375,
    "epsilon_norm": 3.8343679904937744,
    "prediction_norm": 3.581620216369629,
    "timestamp": "2025-07-14T12:33:55.523605"
  },
  {
    "epoch": 622,
    "batch_idx": 0,
    "loss": 0.36853474378585815,
    "learning_rate": 0.0007797411290510838,
    "timestep_mean": 431.90625,
    "epsilon_norm": 8.490532875061035,
    "prediction_norm": 6.974099159240723,
    "timestamp": "2025-07-14T12:33:55.528353"
  },
  {
    "epoch": 622,
    "batch_idx": 32,
    "loss": 0.2012537270784378,
    "learning_rate": 0.0007797411290510838,
    "timestep_mean": 535.96875,
    "epsilon_norm": 8.305135726928711,
    "prediction_norm": 6.992110729217529,
    "timestamp": "2025-07-14T12:33:55.533177"
  },
  {
    "epoch": 622,
    "batch_idx": 64,
    "loss": 0.29981690645217896,
    "learning_rate": 0.0007797411290510838,
    "timestep_mean": 539.96875,
    "epsilon_norm": 7.999209880828857,
    "prediction_norm": 7.253680229187012,
    "timestamp": "2025-07-14T12:33:55.537929"
  },
  {
    "epoch": 622,
    "batch_idx": 96,
    "loss": 0.12039527297019958,
    "learning_rate": 0.0007797411290510838,
    "timestep_mean": 612.0,
    "epsilon_norm": 8.231719970703125,
    "prediction_norm": 8.064508438110352,
    "timestamp": "2025-07-14T12:33:55.542646"
  },
  {
    "epoch": 622,
    "batch_idx": 128,
    "loss": 0.3935086131095886,
    "learning_rate": 0.0007797411290510838,
    "timestep_mean": 523.21875,
    "epsilon_norm": 8.376976013183594,
    "prediction_norm": 6.420001029968262,
    "timestamp": "2025-07-14T12:33:55.547429"
  },
  {
    "epoch": 622,
    "batch_idx": 160,
    "loss": 0.2670997381210327,
    "learning_rate": 0.0007797411290510838,
    "timestep_mean": 528.75,
    "epsilon_norm": 8.606590270996094,
    "prediction_norm": 8.104151725769043,
    "timestamp": "2025-07-14T12:33:55.552182"
  },
  {
    "epoch": 622,
    "batch_idx": 192,
    "loss": 0.30703145265579224,
    "learning_rate": 0.0007797411290510838,
    "timestep_mean": 524.5,
    "epsilon_norm": 4.7371826171875,
    "prediction_norm": 3.6395208835601807,
    "timestamp": "2025-07-14T12:33:55.556910"
  },
  {
    "epoch": 623,
    "batch_idx": 0,
    "loss": 0.3171941637992859,
    "learning_rate": 0.000779089812960774,
    "timestep_mean": 455.25,
    "epsilon_norm": 8.24280071258545,
    "prediction_norm": 7.371577262878418,
    "timestamp": "2025-07-14T12:33:55.561696"
  },
  {
    "epoch": 623,
    "batch_idx": 32,
    "loss": 0.3277568519115448,
    "learning_rate": 0.000779089812960774,
    "timestep_mean": 574.9375,
    "epsilon_norm": 9.829867362976074,
    "prediction_norm": 9.119353294372559,
    "timestamp": "2025-07-14T12:33:55.566629"
  },
  {
    "epoch": 623,
    "batch_idx": 64,
    "loss": 0.3280458152294159,
    "learning_rate": 0.000779089812960774,
    "timestep_mean": 450.1875,
    "epsilon_norm": 8.871705055236816,
    "prediction_norm": 8.016160011291504,
    "timestamp": "2025-07-14T12:33:55.571397"
  },
  {
    "epoch": 623,
    "batch_idx": 96,
    "loss": 0.22884199023246765,
    "learning_rate": 0.000779089812960774,
    "timestep_mean": 457.4375,
    "epsilon_norm": 8.118664741516113,
    "prediction_norm": 6.936599254608154,
    "timestamp": "2025-07-14T12:33:55.576157"
  },
  {
    "epoch": 623,
    "batch_idx": 128,
    "loss": 0.38080811500549316,
    "learning_rate": 0.000779089812960774,
    "timestep_mean": 536.71875,
    "epsilon_norm": 8.447075843811035,
    "prediction_norm": 7.252736568450928,
    "timestamp": "2025-07-14T12:33:55.580907"
  },
  {
    "epoch": 623,
    "batch_idx": 160,
    "loss": 0.2767174243927002,
    "learning_rate": 0.000779089812960774,
    "timestep_mean": 401.34375,
    "epsilon_norm": 7.4213337898254395,
    "prediction_norm": 6.1203789710998535,
    "timestamp": "2025-07-14T12:33:55.585719"
  },
  {
    "epoch": 623,
    "batch_idx": 192,
    "loss": 0.15710116922855377,
    "learning_rate": 0.000779089812960774,
    "timestep_mean": 520.375,
    "epsilon_norm": 4.30610990524292,
    "prediction_norm": 3.8356773853302,
    "timestamp": "2025-07-14T12:33:55.590485"
  },
  {
    "epoch": 624,
    "batch_idx": 0,
    "loss": 0.5156851410865784,
    "learning_rate": 0.0007784378082440943,
    "timestep_mean": 528.46875,
    "epsilon_norm": 9.662932395935059,
    "prediction_norm": 7.086239337921143,
    "timestamp": "2025-07-14T12:33:55.595359"
  },
  {
    "epoch": 624,
    "batch_idx": 32,
    "loss": 0.21657493710517883,
    "learning_rate": 0.0007784378082440943,
    "timestep_mean": 533.625,
    "epsilon_norm": 7.528666019439697,
    "prediction_norm": 6.21589994430542,
    "timestamp": "2025-07-14T12:33:55.600236"
  },
  {
    "epoch": 624,
    "batch_idx": 64,
    "loss": 0.19973596930503845,
    "learning_rate": 0.0007784378082440943,
    "timestep_mean": 470.75,
    "epsilon_norm": 8.569511413574219,
    "prediction_norm": 7.797336101531982,
    "timestamp": "2025-07-14T12:33:55.605088"
  },
  {
    "epoch": 624,
    "batch_idx": 96,
    "loss": 0.2491733878850937,
    "learning_rate": 0.0007784378082440943,
    "timestep_mean": 573.78125,
    "epsilon_norm": 7.343823432922363,
    "prediction_norm": 5.579619884490967,
    "timestamp": "2025-07-14T12:33:55.610070"
  },
  {
    "epoch": 624,
    "batch_idx": 128,
    "loss": 0.19955703616142273,
    "learning_rate": 0.0007784378082440943,
    "timestep_mean": 541.78125,
    "epsilon_norm": 8.298103332519531,
    "prediction_norm": 7.164498329162598,
    "timestamp": "2025-07-14T12:33:55.614885"
  },
  {
    "epoch": 624,
    "batch_idx": 160,
    "loss": 0.36180418729782104,
    "learning_rate": 0.0007784378082440943,
    "timestep_mean": 438.15625,
    "epsilon_norm": 8.176676750183105,
    "prediction_norm": 6.641210556030273,
    "timestamp": "2025-07-14T12:33:55.619683"
  },
  {
    "epoch": 624,
    "batch_idx": 192,
    "loss": 0.11519259214401245,
    "learning_rate": 0.0007784378082440943,
    "timestep_mean": 613.25,
    "epsilon_norm": 3.6850855350494385,
    "prediction_norm": 3.0620553493499756,
    "timestamp": "2025-07-14T12:33:55.624428"
  },
  {
    "epoch": 625,
    "batch_idx": 0,
    "loss": 0.3874729871749878,
    "learning_rate": 0.0007777851165098013,
    "timestep_mean": 457.15625,
    "epsilon_norm": 8.81126594543457,
    "prediction_norm": 6.896322727203369,
    "timestamp": "2025-07-14T12:33:55.629297"
  },
  {
    "epoch": 625,
    "batch_idx": 32,
    "loss": 0.2669721841812134,
    "learning_rate": 0.0007777851165098013,
    "timestep_mean": 544.6875,
    "epsilon_norm": 8.274686813354492,
    "prediction_norm": 8.090360641479492,
    "timestamp": "2025-07-14T12:33:55.634163"
  },
  {
    "epoch": 625,
    "batch_idx": 64,
    "loss": 0.4100411534309387,
    "learning_rate": 0.0007777851165098013,
    "timestep_mean": 396.96875,
    "epsilon_norm": 8.268519401550293,
    "prediction_norm": 7.613272666931152,
    "timestamp": "2025-07-14T12:33:55.639035"
  },
  {
    "epoch": 625,
    "batch_idx": 96,
    "loss": 0.27224332094192505,
    "learning_rate": 0.0007777851165098013,
    "timestep_mean": 504.0625,
    "epsilon_norm": 7.512964725494385,
    "prediction_norm": 6.694843769073486,
    "timestamp": "2025-07-14T12:33:55.643866"
  },
  {
    "epoch": 625,
    "batch_idx": 128,
    "loss": 0.2618717551231384,
    "learning_rate": 0.0007777851165098013,
    "timestep_mean": 414.78125,
    "epsilon_norm": 5.532772064208984,
    "prediction_norm": 4.456562042236328,
    "timestamp": "2025-07-14T12:33:55.648822"
  },
  {
    "epoch": 625,
    "batch_idx": 160,
    "loss": 0.3788629472255707,
    "learning_rate": 0.0007777851165098013,
    "timestep_mean": 545.03125,
    "epsilon_norm": 7.6837992668151855,
    "prediction_norm": 6.889869213104248,
    "timestamp": "2025-07-14T12:33:55.653697"
  },
  {
    "epoch": 625,
    "batch_idx": 192,
    "loss": 0.4383370876312256,
    "learning_rate": 0.0007777851165098013,
    "timestep_mean": 566.75,
    "epsilon_norm": 4.313314437866211,
    "prediction_norm": 3.092485189437866,
    "timestamp": "2025-07-14T12:33:55.658642"
  },
  {
    "epoch": 626,
    "batch_idx": 0,
    "loss": 0.38048744201660156,
    "learning_rate": 0.0007771317393683472,
    "timestep_mean": 595.90625,
    "epsilon_norm": 8.153388977050781,
    "prediction_norm": 6.927546977996826,
    "timestamp": "2025-07-14T12:33:55.663570"
  },
  {
    "epoch": 626,
    "batch_idx": 32,
    "loss": 0.18528272211551666,
    "learning_rate": 0.0007771317393683472,
    "timestep_mean": 501.21875,
    "epsilon_norm": 5.911881446838379,
    "prediction_norm": 5.393021583557129,
    "timestamp": "2025-07-14T12:33:55.668467"
  },
  {
    "epoch": 626,
    "batch_idx": 64,
    "loss": 0.20511090755462646,
    "learning_rate": 0.0007771317393683472,
    "timestep_mean": 412.0,
    "epsilon_norm": 6.733705043792725,
    "prediction_norm": 5.991199493408203,
    "timestamp": "2025-07-14T12:33:55.673381"
  },
  {
    "epoch": 626,
    "batch_idx": 96,
    "loss": 0.13051488995552063,
    "learning_rate": 0.0007771317393683472,
    "timestep_mean": 619.625,
    "epsilon_norm": 6.8610334396362305,
    "prediction_norm": 5.925032138824463,
    "timestamp": "2025-07-14T12:33:55.678272"
  },
  {
    "epoch": 626,
    "batch_idx": 128,
    "loss": 0.4279578924179077,
    "learning_rate": 0.0007771317393683472,
    "timestep_mean": 474.84375,
    "epsilon_norm": 8.46571159362793,
    "prediction_norm": 6.689525604248047,
    "timestamp": "2025-07-14T12:33:55.683262"
  },
  {
    "epoch": 626,
    "batch_idx": 160,
    "loss": 0.40618935227394104,
    "learning_rate": 0.0007771317393683472,
    "timestep_mean": 461.65625,
    "epsilon_norm": 7.763603687286377,
    "prediction_norm": 6.825643539428711,
    "timestamp": "2025-07-14T12:33:55.688165"
  },
  {
    "epoch": 626,
    "batch_idx": 192,
    "loss": 0.2761788070201874,
    "learning_rate": 0.0007771317393683472,
    "timestep_mean": 615.125,
    "epsilon_norm": 3.5326759815216064,
    "prediction_norm": 3.2722396850585938,
    "timestamp": "2025-07-14T12:33:55.693144"
  },
  {
    "epoch": 627,
    "batch_idx": 0,
    "loss": 0.3038785755634308,
    "learning_rate": 0.0007764776784318752,
    "timestep_mean": 520.0625,
    "epsilon_norm": 7.275808811187744,
    "prediction_norm": 5.493159770965576,
    "timestamp": "2025-07-14T12:33:55.698087"
  },
  {
    "epoch": 627,
    "batch_idx": 32,
    "loss": 0.3117843270301819,
    "learning_rate": 0.0007764776784318752,
    "timestep_mean": 545.125,
    "epsilon_norm": 8.75198745727539,
    "prediction_norm": 7.54418420791626,
    "timestamp": "2025-07-14T12:33:55.702985"
  },
  {
    "epoch": 627,
    "batch_idx": 64,
    "loss": 0.23841992020606995,
    "learning_rate": 0.0007764776784318752,
    "timestep_mean": 508.71875,
    "epsilon_norm": 8.912198066711426,
    "prediction_norm": 7.263509750366211,
    "timestamp": "2025-07-14T12:33:55.707855"
  },
  {
    "epoch": 627,
    "batch_idx": 96,
    "loss": 0.5032238960266113,
    "learning_rate": 0.0007764776784318752,
    "timestep_mean": 425.875,
    "epsilon_norm": 9.29175853729248,
    "prediction_norm": 6.818057060241699,
    "timestamp": "2025-07-14T12:33:55.712793"
  },
  {
    "epoch": 627,
    "batch_idx": 128,
    "loss": 0.24042214453220367,
    "learning_rate": 0.0007764776784318752,
    "timestep_mean": 611.78125,
    "epsilon_norm": 8.321283340454102,
    "prediction_norm": 6.79504919052124,
    "timestamp": "2025-07-14T12:33:55.717624"
  },
  {
    "epoch": 627,
    "batch_idx": 160,
    "loss": 0.36342039704322815,
    "learning_rate": 0.0007764776784318752,
    "timestep_mean": 423.53125,
    "epsilon_norm": 8.40601634979248,
    "prediction_norm": 6.514266490936279,
    "timestamp": "2025-07-14T12:33:55.722492"
  },
  {
    "epoch": 627,
    "batch_idx": 192,
    "loss": 0.21372976899147034,
    "learning_rate": 0.0007764776784318752,
    "timestep_mean": 538.0,
    "epsilon_norm": 4.072576522827148,
    "prediction_norm": 3.176537275314331,
    "timestamp": "2025-07-14T12:33:55.727408"
  },
  {
    "epoch": 628,
    "batch_idx": 0,
    "loss": 0.2029903084039688,
    "learning_rate": 0.0007758229353142153,
    "timestep_mean": 574.625,
    "epsilon_norm": 9.220015525817871,
    "prediction_norm": 8.018269538879395,
    "timestamp": "2025-07-14T12:33:55.732340"
  },
  {
    "epoch": 628,
    "batch_idx": 32,
    "loss": 0.2928563058376312,
    "learning_rate": 0.0007758229353142153,
    "timestep_mean": 490.8125,
    "epsilon_norm": 8.014179229736328,
    "prediction_norm": 7.376698970794678,
    "timestamp": "2025-07-14T12:33:55.737314"
  },
  {
    "epoch": 628,
    "batch_idx": 64,
    "loss": 0.19434994459152222,
    "learning_rate": 0.0007758229353142153,
    "timestep_mean": 587.34375,
    "epsilon_norm": 7.713389873504639,
    "prediction_norm": 6.864293098449707,
    "timestamp": "2025-07-14T12:33:55.742310"
  },
  {
    "epoch": 628,
    "batch_idx": 96,
    "loss": 0.2676665186882019,
    "learning_rate": 0.0007758229353142153,
    "timestep_mean": 470.96875,
    "epsilon_norm": 8.247305870056152,
    "prediction_norm": 6.801772117614746,
    "timestamp": "2025-07-14T12:33:55.747255"
  },
  {
    "epoch": 628,
    "batch_idx": 128,
    "loss": 0.12405648827552795,
    "learning_rate": 0.0007758229353142153,
    "timestep_mean": 677.90625,
    "epsilon_norm": 8.58091926574707,
    "prediction_norm": 8.322132110595703,
    "timestamp": "2025-07-14T12:33:55.752190"
  },
  {
    "epoch": 628,
    "batch_idx": 160,
    "loss": 0.24570372700691223,
    "learning_rate": 0.0007758229353142153,
    "timestep_mean": 595.46875,
    "epsilon_norm": 7.575616836547852,
    "prediction_norm": 6.968838214874268,
    "timestamp": "2025-07-14T12:33:55.757073"
  },
  {
    "epoch": 628,
    "batch_idx": 192,
    "loss": 0.20322288572788239,
    "learning_rate": 0.0007758229353142153,
    "timestep_mean": 464.0,
    "epsilon_norm": 4.28308629989624,
    "prediction_norm": 3.9565460681915283,
    "timestamp": "2025-07-14T12:33:55.761958"
  },
  {
    "epoch": 629,
    "batch_idx": 0,
    "loss": 0.25402164459228516,
    "learning_rate": 0.0007751675116308812,
    "timestep_mean": 501.46875,
    "epsilon_norm": 7.655596733093262,
    "prediction_norm": 6.12814474105835,
    "timestamp": "2025-07-14T12:33:55.767228"
  },
  {
    "epoch": 629,
    "batch_idx": 32,
    "loss": 0.3175796866416931,
    "learning_rate": 0.0007751675116308812,
    "timestep_mean": 559.84375,
    "epsilon_norm": 8.549030303955078,
    "prediction_norm": 6.714244842529297,
    "timestamp": "2025-07-14T12:33:55.772303"
  },
  {
    "epoch": 629,
    "batch_idx": 64,
    "loss": 0.40365394949913025,
    "learning_rate": 0.0007751675116308812,
    "timestep_mean": 429.875,
    "epsilon_norm": 8.164704322814941,
    "prediction_norm": 6.5903778076171875,
    "timestamp": "2025-07-14T12:33:55.777345"
  },
  {
    "epoch": 629,
    "batch_idx": 96,
    "loss": 0.19438979029655457,
    "learning_rate": 0.0007751675116308812,
    "timestep_mean": 546.21875,
    "epsilon_norm": 6.802876949310303,
    "prediction_norm": 6.065086841583252,
    "timestamp": "2025-07-14T12:33:55.782575"
  },
  {
    "epoch": 629,
    "batch_idx": 128,
    "loss": 0.25491464138031006,
    "learning_rate": 0.0007751675116308812,
    "timestep_mean": 516.6875,
    "epsilon_norm": 8.451601028442383,
    "prediction_norm": 7.613572597503662,
    "timestamp": "2025-07-14T12:33:55.787693"
  },
  {
    "epoch": 629,
    "batch_idx": 160,
    "loss": 0.12259931117296219,
    "learning_rate": 0.0007751675116308812,
    "timestep_mean": 576.09375,
    "epsilon_norm": 7.909121036529541,
    "prediction_norm": 7.628907680511475,
    "timestamp": "2025-07-14T12:33:55.792939"
  },
  {
    "epoch": 629,
    "batch_idx": 192,
    "loss": 0.15575240552425385,
    "learning_rate": 0.0007751675116308812,
    "timestep_mean": 433.125,
    "epsilon_norm": 2.485705614089966,
    "prediction_norm": 2.5003113746643066,
    "timestamp": "2025-07-14T12:33:55.798080"
  },
  {
    "epoch": 630,
    "batch_idx": 0,
    "loss": 0.2512456178665161,
    "learning_rate": 0.000774511408999066,
    "timestep_mean": 537.9375,
    "epsilon_norm": 7.666942596435547,
    "prediction_norm": 6.166594505310059,
    "timestamp": "2025-07-14T12:33:55.803186"
  },
  {
    "epoch": 630,
    "batch_idx": 32,
    "loss": 0.3192206621170044,
    "learning_rate": 0.000774511408999066,
    "timestep_mean": 511.71875,
    "epsilon_norm": 8.598945617675781,
    "prediction_norm": 6.848229885101318,
    "timestamp": "2025-07-14T12:33:55.808439"
  },
  {
    "epoch": 630,
    "batch_idx": 64,
    "loss": 0.28503966331481934,
    "learning_rate": 0.000774511408999066,
    "timestep_mean": 572.5625,
    "epsilon_norm": 9.2028169631958,
    "prediction_norm": 8.27136516571045,
    "timestamp": "2025-07-14T12:33:55.813567"
  },
  {
    "epoch": 630,
    "batch_idx": 96,
    "loss": 0.2069547176361084,
    "learning_rate": 0.000774511408999066,
    "timestep_mean": 587.6875,
    "epsilon_norm": 7.938050746917725,
    "prediction_norm": 7.0538811683654785,
    "timestamp": "2025-07-14T12:33:55.818714"
  },
  {
    "epoch": 630,
    "batch_idx": 128,
    "loss": 0.3163830637931824,
    "learning_rate": 0.000774511408999066,
    "timestep_mean": 551.25,
    "epsilon_norm": 7.869532108306885,
    "prediction_norm": 6.436375617980957,
    "timestamp": "2025-07-14T12:33:55.823827"
  },
  {
    "epoch": 630,
    "batch_idx": 160,
    "loss": 0.14964361488819122,
    "learning_rate": 0.000774511408999066,
    "timestep_mean": 558.34375,
    "epsilon_norm": 7.165193557739258,
    "prediction_norm": 6.7930989265441895,
    "timestamp": "2025-07-14T12:33:55.829113"
  },
  {
    "epoch": 630,
    "batch_idx": 192,
    "loss": 0.4892449975013733,
    "learning_rate": 0.000774511408999066,
    "timestep_mean": 277.875,
    "epsilon_norm": 4.227692127227783,
    "prediction_norm": 2.1708016395568848,
    "timestamp": "2025-07-14T12:33:55.834265"
  },
  {
    "epoch": 631,
    "batch_idx": 0,
    "loss": 0.2214801013469696,
    "learning_rate": 0.0007738546290376373,
    "timestep_mean": 524.5625,
    "epsilon_norm": 7.138473033905029,
    "prediction_norm": 6.48324728012085,
    "timestamp": "2025-07-14T12:33:55.839374"
  },
  {
    "epoch": 631,
    "batch_idx": 32,
    "loss": 0.2594212293624878,
    "learning_rate": 0.0007738546290376373,
    "timestep_mean": 520.5625,
    "epsilon_norm": 7.879105091094971,
    "prediction_norm": 7.127758979797363,
    "timestamp": "2025-07-14T12:33:55.844575"
  },
  {
    "epoch": 631,
    "batch_idx": 64,
    "loss": 0.3110690414905548,
    "learning_rate": 0.0007738546290376373,
    "timestep_mean": 537.6875,
    "epsilon_norm": 7.971896171569824,
    "prediction_norm": 6.120950222015381,
    "timestamp": "2025-07-14T12:33:55.849729"
  },
  {
    "epoch": 631,
    "batch_idx": 96,
    "loss": 0.334253191947937,
    "learning_rate": 0.0007738546290376373,
    "timestep_mean": 535.0625,
    "epsilon_norm": 8.143119812011719,
    "prediction_norm": 7.080721378326416,
    "timestamp": "2025-07-14T12:33:55.854905"
  },
  {
    "epoch": 631,
    "batch_idx": 128,
    "loss": 0.3743303418159485,
    "learning_rate": 0.0007738546290376373,
    "timestep_mean": 420.25,
    "epsilon_norm": 7.161596775054932,
    "prediction_norm": 6.145355224609375,
    "timestamp": "2025-07-14T12:33:55.860148"
  },
  {
    "epoch": 631,
    "batch_idx": 160,
    "loss": 0.2558766007423401,
    "learning_rate": 0.0007738546290376373,
    "timestep_mean": 469.90625,
    "epsilon_norm": 6.068190097808838,
    "prediction_norm": 5.589697360992432,
    "timestamp": "2025-07-14T12:33:55.865431"
  },
  {
    "epoch": 631,
    "batch_idx": 192,
    "loss": 0.1529540717601776,
    "learning_rate": 0.0007738546290376373,
    "timestep_mean": 438.75,
    "epsilon_norm": 3.7260758876800537,
    "prediction_norm": 3.7978482246398926,
    "timestamp": "2025-07-14T12:33:55.870499"
  },
  {
    "epoch": 632,
    "batch_idx": 0,
    "loss": 0.3684534430503845,
    "learning_rate": 0.0007731971733671347,
    "timestep_mean": 560.75,
    "epsilon_norm": 8.025836944580078,
    "prediction_norm": 6.691573619842529,
    "timestamp": "2025-07-14T12:33:55.875817"
  },
  {
    "epoch": 632,
    "batch_idx": 32,
    "loss": 0.19288530945777893,
    "learning_rate": 0.0007731971733671347,
    "timestep_mean": 411.84375,
    "epsilon_norm": 7.502342700958252,
    "prediction_norm": 6.350095748901367,
    "timestamp": "2025-07-14T12:33:55.880843"
  },
  {
    "epoch": 632,
    "batch_idx": 64,
    "loss": 0.27384424209594727,
    "learning_rate": 0.0007731971733671347,
    "timestep_mean": 489.09375,
    "epsilon_norm": 7.195037364959717,
    "prediction_norm": 6.1361918449401855,
    "timestamp": "2025-07-14T12:33:55.885728"
  },
  {
    "epoch": 632,
    "batch_idx": 96,
    "loss": 0.33129245042800903,
    "learning_rate": 0.0007731971733671347,
    "timestep_mean": 518.9375,
    "epsilon_norm": 7.721014499664307,
    "prediction_norm": 6.142897129058838,
    "timestamp": "2025-07-14T12:33:55.890671"
  },
  {
    "epoch": 632,
    "batch_idx": 128,
    "loss": 0.40658989548683167,
    "learning_rate": 0.0007731971733671347,
    "timestep_mean": 507.28125,
    "epsilon_norm": 9.116591453552246,
    "prediction_norm": 8.024158477783203,
    "timestamp": "2025-07-14T12:33:55.895549"
  },
  {
    "epoch": 632,
    "batch_idx": 160,
    "loss": 0.36291801929473877,
    "learning_rate": 0.0007731971733671347,
    "timestep_mean": 481.9375,
    "epsilon_norm": 6.467769145965576,
    "prediction_norm": 4.777113437652588,
    "timestamp": "2025-07-14T12:33:55.900507"
  },
  {
    "epoch": 632,
    "batch_idx": 192,
    "loss": 0.05695924162864685,
    "learning_rate": 0.0007731971733671347,
    "timestep_mean": 419.75,
    "epsilon_norm": 3.5478055477142334,
    "prediction_norm": 3.639404058456421,
    "timestamp": "2025-07-14T12:33:55.905618"
  },
  {
    "epoch": 633,
    "batch_idx": 0,
    "loss": 0.22709237039089203,
    "learning_rate": 0.0007725390436097644,
    "timestep_mean": 491.34375,
    "epsilon_norm": 7.7121806144714355,
    "prediction_norm": 6.231161117553711,
    "timestamp": "2025-07-14T12:33:55.910713"
  },
  {
    "epoch": 633,
    "batch_idx": 32,
    "loss": 0.20814761519432068,
    "learning_rate": 0.0007725390436097644,
    "timestep_mean": 470.59375,
    "epsilon_norm": 8.420174598693848,
    "prediction_norm": 7.6320271492004395,
    "timestamp": "2025-07-14T12:33:55.915873"
  },
  {
    "epoch": 633,
    "batch_idx": 64,
    "loss": 0.37603944540023804,
    "learning_rate": 0.0007725390436097644,
    "timestep_mean": 438.09375,
    "epsilon_norm": 8.581755638122559,
    "prediction_norm": 6.188243389129639,
    "timestamp": "2025-07-14T12:33:55.920992"
  },
  {
    "epoch": 633,
    "batch_idx": 96,
    "loss": 0.5042405724525452,
    "learning_rate": 0.0007725390436097644,
    "timestep_mean": 398.875,
    "epsilon_norm": 7.669864654541016,
    "prediction_norm": 5.36316442489624,
    "timestamp": "2025-07-14T12:33:55.926055"
  },
  {
    "epoch": 633,
    "batch_idx": 128,
    "loss": 0.18308213353157043,
    "learning_rate": 0.0007725390436097644,
    "timestep_mean": 412.03125,
    "epsilon_norm": 6.585567951202393,
    "prediction_norm": 4.841373920440674,
    "timestamp": "2025-07-14T12:33:55.931288"
  },
  {
    "epoch": 633,
    "batch_idx": 160,
    "loss": 0.2541467547416687,
    "learning_rate": 0.0007725390436097644,
    "timestep_mean": 545.15625,
    "epsilon_norm": 8.344436645507812,
    "prediction_norm": 7.1857757568359375,
    "timestamp": "2025-07-14T12:33:55.936350"
  },
  {
    "epoch": 633,
    "batch_idx": 192,
    "loss": 0.24982309341430664,
    "learning_rate": 0.0007725390436097644,
    "timestep_mean": 434.0,
    "epsilon_norm": 4.403172969818115,
    "prediction_norm": 3.553065299987793,
    "timestamp": "2025-07-14T12:33:55.941400"
  },
  {
    "epoch": 634,
    "batch_idx": 0,
    "loss": 0.4492323398590088,
    "learning_rate": 0.0007718802413893964,
    "timestep_mean": 396.25,
    "epsilon_norm": 7.432992458343506,
    "prediction_norm": 4.877532482147217,
    "timestamp": "2025-07-14T12:33:55.946580"
  },
  {
    "epoch": 634,
    "batch_idx": 32,
    "loss": 0.21863451600074768,
    "learning_rate": 0.0007718802413893964,
    "timestep_mean": 499.6875,
    "epsilon_norm": 7.892181873321533,
    "prediction_norm": 6.849778175354004,
    "timestamp": "2025-07-14T12:33:55.951716"
  },
  {
    "epoch": 634,
    "batch_idx": 64,
    "loss": 0.2948455214500427,
    "learning_rate": 0.0007718802413893964,
    "timestep_mean": 499.5625,
    "epsilon_norm": 7.881026268005371,
    "prediction_norm": 6.815621376037598,
    "timestamp": "2025-07-14T12:33:55.956890"
  },
  {
    "epoch": 634,
    "batch_idx": 96,
    "loss": 0.26977333426475525,
    "learning_rate": 0.0007718802413893964,
    "timestep_mean": 477.125,
    "epsilon_norm": 7.4747796058654785,
    "prediction_norm": 6.921372413635254,
    "timestamp": "2025-07-14T12:33:55.962111"
  },
  {
    "epoch": 634,
    "batch_idx": 128,
    "loss": 0.19979852437973022,
    "learning_rate": 0.0007718802413893964,
    "timestep_mean": 554.71875,
    "epsilon_norm": 7.683176040649414,
    "prediction_norm": 7.157772541046143,
    "timestamp": "2025-07-14T12:33:55.967302"
  },
  {
    "epoch": 634,
    "batch_idx": 160,
    "loss": 0.30552029609680176,
    "learning_rate": 0.0007718802413893964,
    "timestep_mean": 479.34375,
    "epsilon_norm": 8.480378150939941,
    "prediction_norm": 7.2048444747924805,
    "timestamp": "2025-07-14T12:33:55.972329"
  },
  {
    "epoch": 634,
    "batch_idx": 192,
    "loss": 0.3143565356731415,
    "learning_rate": 0.0007718802413893964,
    "timestep_mean": 493.25,
    "epsilon_norm": 2.564652681350708,
    "prediction_norm": 2.1166598796844482,
    "timestamp": "2025-07-14T12:33:55.977549"
  },
  {
    "epoch": 635,
    "batch_idx": 0,
    "loss": 0.32509928941726685,
    "learning_rate": 0.0007712207683315596,
    "timestep_mean": 571.3125,
    "epsilon_norm": 9.601000785827637,
    "prediction_norm": 7.844747066497803,
    "timestamp": "2025-07-14T12:33:55.982749"
  },
  {
    "epoch": 635,
    "batch_idx": 32,
    "loss": 0.4780339002609253,
    "learning_rate": 0.0007712207683315596,
    "timestep_mean": 465.78125,
    "epsilon_norm": 8.309236526489258,
    "prediction_norm": 6.226583957672119,
    "timestamp": "2025-07-14T12:33:55.987942"
  },
  {
    "epoch": 635,
    "batch_idx": 64,
    "loss": 0.2330116331577301,
    "learning_rate": 0.0007712207683315596,
    "timestep_mean": 464.875,
    "epsilon_norm": 7.766391754150391,
    "prediction_norm": 6.654778003692627,
    "timestamp": "2025-07-14T12:33:55.993153"
  },
  {
    "epoch": 635,
    "batch_idx": 96,
    "loss": 0.2072772979736328,
    "learning_rate": 0.0007712207683315596,
    "timestep_mean": 415.46875,
    "epsilon_norm": 8.725129127502441,
    "prediction_norm": 7.712265968322754,
    "timestamp": "2025-07-14T12:33:55.998387"
  },
  {
    "epoch": 635,
    "batch_idx": 128,
    "loss": 0.2023601084947586,
    "learning_rate": 0.0007712207683315596,
    "timestep_mean": 535.53125,
    "epsilon_norm": 7.335774898529053,
    "prediction_norm": 6.462559700012207,
    "timestamp": "2025-07-14T12:33:56.003525"
  },
  {
    "epoch": 635,
    "batch_idx": 160,
    "loss": 0.36802899837493896,
    "learning_rate": 0.0007712207683315596,
    "timestep_mean": 528.8125,
    "epsilon_norm": 8.0632963180542,
    "prediction_norm": 6.692541122436523,
    "timestamp": "2025-07-14T12:33:56.008708"
  },
  {
    "epoch": 635,
    "batch_idx": 192,
    "loss": 0.3023284673690796,
    "learning_rate": 0.0007712207683315596,
    "timestep_mean": 690.5,
    "epsilon_norm": 4.6034040451049805,
    "prediction_norm": 4.165930271148682,
    "timestamp": "2025-07-14T12:33:56.013848"
  },
  {
    "epoch": 636,
    "batch_idx": 0,
    "loss": 0.20130395889282227,
    "learning_rate": 0.0007705606260634381,
    "timestep_mean": 519.59375,
    "epsilon_norm": 8.968851089477539,
    "prediction_norm": 7.667372703552246,
    "timestamp": "2025-07-14T12:33:56.018884"
  },
  {
    "epoch": 636,
    "batch_idx": 32,
    "loss": 0.2895268201828003,
    "learning_rate": 0.0007705606260634381,
    "timestep_mean": 377.21875,
    "epsilon_norm": 6.853681564331055,
    "prediction_norm": 5.9086012840271,
    "timestamp": "2025-07-14T12:33:56.024080"
  },
  {
    "epoch": 636,
    "batch_idx": 64,
    "loss": 0.31323492527008057,
    "learning_rate": 0.0007705606260634381,
    "timestep_mean": 498.5,
    "epsilon_norm": 7.361472129821777,
    "prediction_norm": 5.644984722137451,
    "timestamp": "2025-07-14T12:33:56.029266"
  },
  {
    "epoch": 636,
    "batch_idx": 96,
    "loss": 0.37109899520874023,
    "learning_rate": 0.0007705606260634381,
    "timestep_mean": 509.09375,
    "epsilon_norm": 8.640768051147461,
    "prediction_norm": 7.225833892822266,
    "timestamp": "2025-07-14T12:33:56.034263"
  },
  {
    "epoch": 636,
    "batch_idx": 128,
    "loss": 0.10548004508018494,
    "learning_rate": 0.0007705606260634381,
    "timestep_mean": 631.4375,
    "epsilon_norm": 8.670578956604004,
    "prediction_norm": 8.16926097869873,
    "timestamp": "2025-07-14T12:33:56.039317"
  },
  {
    "epoch": 636,
    "batch_idx": 160,
    "loss": 0.196374773979187,
    "learning_rate": 0.0007705606260634381,
    "timestep_mean": 574.40625,
    "epsilon_norm": 8.065383911132812,
    "prediction_norm": 6.557712554931641,
    "timestamp": "2025-07-14T12:33:56.044537"
  },
  {
    "epoch": 636,
    "batch_idx": 192,
    "loss": 0.35954082012176514,
    "learning_rate": 0.0007705606260634381,
    "timestep_mean": 595.375,
    "epsilon_norm": 4.772645473480225,
    "prediction_norm": 3.816375732421875,
    "timestamp": "2025-07-14T12:33:56.049610"
  },
  {
    "epoch": 637,
    "batch_idx": 0,
    "loss": 0.3649696111679077,
    "learning_rate": 0.0007698998162138674,
    "timestep_mean": 487.75,
    "epsilon_norm": 8.403812408447266,
    "prediction_norm": 6.2370381355285645,
    "timestamp": "2025-07-14T12:33:56.055032"
  },
  {
    "epoch": 637,
    "batch_idx": 32,
    "loss": 0.13109545409679413,
    "learning_rate": 0.0007698998162138674,
    "timestep_mean": 480.5625,
    "epsilon_norm": 6.889039039611816,
    "prediction_norm": 6.191630840301514,
    "timestamp": "2025-07-14T12:33:56.060270"
  },
  {
    "epoch": 637,
    "batch_idx": 64,
    "loss": 0.3444804549217224,
    "learning_rate": 0.0007698998162138674,
    "timestep_mean": 543.6875,
    "epsilon_norm": 9.518841743469238,
    "prediction_norm": 7.832849502563477,
    "timestamp": "2025-07-14T12:33:56.065340"
  },
  {
    "epoch": 637,
    "batch_idx": 96,
    "loss": 0.4240258038043976,
    "learning_rate": 0.0007698998162138674,
    "timestep_mean": 510.9375,
    "epsilon_norm": 8.259622573852539,
    "prediction_norm": 6.676515102386475,
    "timestamp": "2025-07-14T12:33:56.070464"
  },
  {
    "epoch": 637,
    "batch_idx": 128,
    "loss": 0.33183664083480835,
    "learning_rate": 0.0007698998162138674,
    "timestep_mean": 479.46875,
    "epsilon_norm": 7.287464618682861,
    "prediction_norm": 5.483103275299072,
    "timestamp": "2025-07-14T12:33:56.075646"
  },
  {
    "epoch": 637,
    "batch_idx": 160,
    "loss": 0.24357964098453522,
    "learning_rate": 0.0007698998162138674,
    "timestep_mean": 480.21875,
    "epsilon_norm": 7.914938926696777,
    "prediction_norm": 6.684289932250977,
    "timestamp": "2025-07-14T12:33:56.080819"
  },
  {
    "epoch": 637,
    "batch_idx": 192,
    "loss": 0.1635507494211197,
    "learning_rate": 0.0007698998162138674,
    "timestep_mean": 443.5,
    "epsilon_norm": 3.587632179260254,
    "prediction_norm": 3.3006889820098877,
    "timestamp": "2025-07-14T12:33:56.085959"
  },
  {
    "epoch": 638,
    "batch_idx": 0,
    "loss": 0.2920815050601959,
    "learning_rate": 0.0007692383404133302,
    "timestep_mean": 482.15625,
    "epsilon_norm": 7.85476016998291,
    "prediction_norm": 7.298560619354248,
    "timestamp": "2025-07-14T12:33:56.091103"
  },
  {
    "epoch": 638,
    "batch_idx": 32,
    "loss": 0.4531466066837311,
    "learning_rate": 0.0007692383404133302,
    "timestep_mean": 480.0625,
    "epsilon_norm": 8.820042610168457,
    "prediction_norm": 7.195842742919922,
    "timestamp": "2025-07-14T12:33:56.096548"
  },
  {
    "epoch": 638,
    "batch_idx": 64,
    "loss": 0.2204444706439972,
    "learning_rate": 0.0007692383404133302,
    "timestep_mean": 479.28125,
    "epsilon_norm": 7.884718894958496,
    "prediction_norm": 7.380302906036377,
    "timestamp": "2025-07-14T12:33:56.101793"
  },
  {
    "epoch": 638,
    "batch_idx": 96,
    "loss": 0.21665158867835999,
    "learning_rate": 0.0007692383404133302,
    "timestep_mean": 562.5,
    "epsilon_norm": 8.147886276245117,
    "prediction_norm": 7.281417369842529,
    "timestamp": "2025-07-14T12:33:56.106932"
  },
  {
    "epoch": 638,
    "batch_idx": 128,
    "loss": 0.2090102732181549,
    "learning_rate": 0.0007692383404133302,
    "timestep_mean": 539.65625,
    "epsilon_norm": 7.814888954162598,
    "prediction_norm": 6.912275314331055,
    "timestamp": "2025-07-14T12:33:56.118612"
  },
  {
    "epoch": 638,
    "batch_idx": 160,
    "loss": 0.21569760143756866,
    "learning_rate": 0.0007692383404133302,
    "timestep_mean": 561.75,
    "epsilon_norm": 8.628491401672363,
    "prediction_norm": 7.59541654586792,
    "timestamp": "2025-07-14T12:33:56.123820"
  },
  {
    "epoch": 638,
    "batch_idx": 192,
    "loss": 0.14165376126766205,
    "learning_rate": 0.0007692383404133302,
    "timestep_mean": 285.75,
    "epsilon_norm": 2.580859422683716,
    "prediction_norm": 2.3983566761016846,
    "timestamp": "2025-07-14T12:33:56.128800"
  },
  {
    "epoch": 639,
    "batch_idx": 0,
    "loss": 0.20013344287872314,
    "learning_rate": 0.0007685762002939524,
    "timestep_mean": 652.1875,
    "epsilon_norm": 8.358504295349121,
    "prediction_norm": 8.271716117858887,
    "timestamp": "2025-07-14T12:33:56.133740"
  },
  {
    "epoch": 639,
    "batch_idx": 32,
    "loss": 0.3082271218299866,
    "learning_rate": 0.0007685762002939524,
    "timestep_mean": 490.28125,
    "epsilon_norm": 7.324692249298096,
    "prediction_norm": 5.8474345207214355,
    "timestamp": "2025-07-14T12:33:56.138667"
  },
  {
    "epoch": 639,
    "batch_idx": 64,
    "loss": 0.36000725626945496,
    "learning_rate": 0.0007685762002939524,
    "timestep_mean": 391.90625,
    "epsilon_norm": 6.976168632507324,
    "prediction_norm": 5.640900611877441,
    "timestamp": "2025-07-14T12:33:56.143632"
  },
  {
    "epoch": 639,
    "batch_idx": 96,
    "loss": 0.21665289998054504,
    "learning_rate": 0.0007685762002939524,
    "timestep_mean": 567.40625,
    "epsilon_norm": 7.600551128387451,
    "prediction_norm": 6.9932122230529785,
    "timestamp": "2025-07-14T12:33:56.148626"
  },
  {
    "epoch": 639,
    "batch_idx": 128,
    "loss": 0.4390428960323334,
    "learning_rate": 0.0007685762002939524,
    "timestep_mean": 426.6875,
    "epsilon_norm": 7.934222221374512,
    "prediction_norm": 6.705890655517578,
    "timestamp": "2025-07-14T12:33:56.153601"
  },
  {
    "epoch": 639,
    "batch_idx": 160,
    "loss": 0.21671468019485474,
    "learning_rate": 0.0007685762002939524,
    "timestep_mean": 521.8125,
    "epsilon_norm": 7.907815456390381,
    "prediction_norm": 6.502933025360107,
    "timestamp": "2025-07-14T12:33:56.158735"
  },
  {
    "epoch": 639,
    "batch_idx": 192,
    "loss": 0.3778313100337982,
    "learning_rate": 0.0007685762002939524,
    "timestep_mean": 453.5,
    "epsilon_norm": 4.0151543617248535,
    "prediction_norm": 3.360464096069336,
    "timestamp": "2025-07-14T12:33:56.163736"
  },
  {
    "epoch": 640,
    "batch_idx": 0,
    "loss": 0.25063300132751465,
    "learning_rate": 0.0007679133974894984,
    "timestep_mean": 546.9375,
    "epsilon_norm": 6.628600597381592,
    "prediction_norm": 5.4378814697265625,
    "timestamp": "2025-07-14T12:33:56.168785"
  },
  {
    "epoch": 640,
    "batch_idx": 32,
    "loss": 0.30468568205833435,
    "learning_rate": 0.0007679133974894984,
    "timestep_mean": 532.84375,
    "epsilon_norm": 7.858932971954346,
    "prediction_norm": 6.512387275695801,
    "timestamp": "2025-07-14T12:33:56.173742"
  },
  {
    "epoch": 640,
    "batch_idx": 64,
    "loss": 0.28930842876434326,
    "learning_rate": 0.0007679133974894984,
    "timestep_mean": 536.5625,
    "epsilon_norm": 8.097529411315918,
    "prediction_norm": 6.840206146240234,
    "timestamp": "2025-07-14T12:33:56.178757"
  },
  {
    "epoch": 640,
    "batch_idx": 96,
    "loss": 0.38951635360717773,
    "learning_rate": 0.0007679133974894984,
    "timestep_mean": 428.46875,
    "epsilon_norm": 8.367815971374512,
    "prediction_norm": 6.527542591094971,
    "timestamp": "2025-07-14T12:33:56.183823"
  },
  {
    "epoch": 640,
    "batch_idx": 128,
    "loss": 0.24101176857948303,
    "learning_rate": 0.0007679133974894984,
    "timestep_mean": 523.125,
    "epsilon_norm": 8.456491470336914,
    "prediction_norm": 7.417337417602539,
    "timestamp": "2025-07-14T12:33:56.188789"
  },
  {
    "epoch": 640,
    "batch_idx": 160,
    "loss": 0.35275065898895264,
    "learning_rate": 0.0007679133974894984,
    "timestep_mean": 449.03125,
    "epsilon_norm": 8.031689643859863,
    "prediction_norm": 7.037400245666504,
    "timestamp": "2025-07-14T12:33:56.193837"
  },
  {
    "epoch": 640,
    "batch_idx": 192,
    "loss": 0.3567683696746826,
    "learning_rate": 0.0007679133974894984,
    "timestep_mean": 409.375,
    "epsilon_norm": 3.5120019912719727,
    "prediction_norm": 2.632014513015747,
    "timestamp": "2025-07-14T12:33:56.198805"
  },
  {
    "epoch": 641,
    "batch_idx": 0,
    "loss": 0.3144807815551758,
    "learning_rate": 0.0007672499336353688,
    "timestep_mean": 455.0625,
    "epsilon_norm": 7.801919937133789,
    "prediction_norm": 6.447849273681641,
    "timestamp": "2025-07-14T12:33:56.203856"
  },
  {
    "epoch": 641,
    "batch_idx": 32,
    "loss": 0.15886960923671722,
    "learning_rate": 0.0007672499336353688,
    "timestep_mean": 541.78125,
    "epsilon_norm": 6.060384273529053,
    "prediction_norm": 5.820659637451172,
    "timestamp": "2025-07-14T12:33:56.208878"
  },
  {
    "epoch": 641,
    "batch_idx": 64,
    "loss": 0.5274243354797363,
    "learning_rate": 0.0007672499336353688,
    "timestep_mean": 510.90625,
    "epsilon_norm": 8.780083656311035,
    "prediction_norm": 7.067659378051758,
    "timestamp": "2025-07-14T12:33:56.213826"
  },
  {
    "epoch": 641,
    "batch_idx": 96,
    "loss": 0.27845343947410583,
    "learning_rate": 0.0007672499336353688,
    "timestep_mean": 513.875,
    "epsilon_norm": 7.635231018066406,
    "prediction_norm": 6.171726226806641,
    "timestamp": "2025-07-14T12:33:56.219005"
  },
  {
    "epoch": 641,
    "batch_idx": 128,
    "loss": 0.2850129008293152,
    "learning_rate": 0.0007672499336353688,
    "timestep_mean": 516.65625,
    "epsilon_norm": 9.520528793334961,
    "prediction_norm": 8.906042098999023,
    "timestamp": "2025-07-14T12:33:56.224044"
  },
  {
    "epoch": 641,
    "batch_idx": 160,
    "loss": 0.20017290115356445,
    "learning_rate": 0.0007672499336353688,
    "timestep_mean": 565.15625,
    "epsilon_norm": 8.51699161529541,
    "prediction_norm": 7.091610908508301,
    "timestamp": "2025-07-14T12:33:56.229046"
  },
  {
    "epoch": 641,
    "batch_idx": 192,
    "loss": 0.4462488889694214,
    "learning_rate": 0.0007672499336353688,
    "timestep_mean": 399.875,
    "epsilon_norm": 3.713923215866089,
    "prediction_norm": 2.374471426010132,
    "timestamp": "2025-07-14T12:33:56.234238"
  },
  {
    "epoch": 642,
    "batch_idx": 0,
    "loss": 0.3132615387439728,
    "learning_rate": 0.0007665858103685945,
    "timestep_mean": 533.125,
    "epsilon_norm": 7.998593807220459,
    "prediction_norm": 6.453216075897217,
    "timestamp": "2025-07-14T12:33:56.239230"
  },
  {
    "epoch": 642,
    "batch_idx": 32,
    "loss": 0.35808077454566956,
    "learning_rate": 0.0007665858103685945,
    "timestep_mean": 491.8125,
    "epsilon_norm": 7.003857612609863,
    "prediction_norm": 5.123208045959473,
    "timestamp": "2025-07-14T12:33:56.244311"
  },
  {
    "epoch": 642,
    "batch_idx": 64,
    "loss": 0.2409297525882721,
    "learning_rate": 0.0007665858103685945,
    "timestep_mean": 524.5,
    "epsilon_norm": 7.514325141906738,
    "prediction_norm": 6.2905592918396,
    "timestamp": "2025-07-14T12:33:56.249377"
  },
  {
    "epoch": 642,
    "batch_idx": 96,
    "loss": 0.5140782594680786,
    "learning_rate": 0.0007665858103685945,
    "timestep_mean": 462.21875,
    "epsilon_norm": 9.360386848449707,
    "prediction_norm": 6.348939418792725,
    "timestamp": "2025-07-14T12:33:56.254540"
  },
  {
    "epoch": 642,
    "batch_idx": 128,
    "loss": 0.5601446628570557,
    "learning_rate": 0.0007665858103685945,
    "timestep_mean": 496.0,
    "epsilon_norm": 9.26929759979248,
    "prediction_norm": 7.267959117889404,
    "timestamp": "2025-07-14T12:33:56.259596"
  },
  {
    "epoch": 642,
    "batch_idx": 160,
    "loss": 0.3122951090335846,
    "learning_rate": 0.0007665858103685945,
    "timestep_mean": 508.3125,
    "epsilon_norm": 8.92949390411377,
    "prediction_norm": 7.285760402679443,
    "timestamp": "2025-07-14T12:33:56.264555"
  },
  {
    "epoch": 642,
    "batch_idx": 192,
    "loss": 0.1012183278799057,
    "learning_rate": 0.0007665858103685945,
    "timestep_mean": 610.75,
    "epsilon_norm": 4.048186779022217,
    "prediction_norm": 3.680544853210449,
    "timestamp": "2025-07-14T12:33:56.269658"
  },
  {
    "epoch": 643,
    "batch_idx": 0,
    "loss": 0.2368488907814026,
    "learning_rate": 0.0007659210293278335,
    "timestep_mean": 543.53125,
    "epsilon_norm": 8.49898624420166,
    "prediction_norm": 7.145923614501953,
    "timestamp": "2025-07-14T12:33:56.274874"
  },
  {
    "epoch": 643,
    "batch_idx": 32,
    "loss": 0.25431889295578003,
    "learning_rate": 0.0007659210293278335,
    "timestep_mean": 555.0,
    "epsilon_norm": 9.495402336120605,
    "prediction_norm": 7.742218494415283,
    "timestamp": "2025-07-14T12:33:56.279952"
  },
  {
    "epoch": 643,
    "batch_idx": 64,
    "loss": 0.1632404625415802,
    "learning_rate": 0.0007659210293278335,
    "timestep_mean": 520.96875,
    "epsilon_norm": 7.748581886291504,
    "prediction_norm": 7.025643825531006,
    "timestamp": "2025-07-14T12:33:56.285092"
  },
  {
    "epoch": 643,
    "batch_idx": 96,
    "loss": 0.2285872995853424,
    "learning_rate": 0.0007659210293278335,
    "timestep_mean": 486.46875,
    "epsilon_norm": 7.993507385253906,
    "prediction_norm": 6.819548606872559,
    "timestamp": "2025-07-14T12:33:56.290202"
  },
  {
    "epoch": 643,
    "batch_idx": 128,
    "loss": 0.20852801203727722,
    "learning_rate": 0.0007659210293278335,
    "timestep_mean": 533.53125,
    "epsilon_norm": 8.350833892822266,
    "prediction_norm": 7.833771705627441,
    "timestamp": "2025-07-14T12:33:56.295394"
  },
  {
    "epoch": 643,
    "batch_idx": 160,
    "loss": 0.2965606451034546,
    "learning_rate": 0.0007659210293278335,
    "timestep_mean": 566.59375,
    "epsilon_norm": 7.800375461578369,
    "prediction_norm": 7.163290500640869,
    "timestamp": "2025-07-14T12:33:56.300665"
  },
  {
    "epoch": 643,
    "batch_idx": 192,
    "loss": 0.18284641206264496,
    "learning_rate": 0.0007659210293278335,
    "timestep_mean": 557.0,
    "epsilon_norm": 3.7976150512695312,
    "prediction_norm": 3.6077468395233154,
    "timestamp": "2025-07-14T12:33:56.305819"
  },
  {
    "epoch": 644,
    "batch_idx": 0,
    "loss": 0.21553748846054077,
    "learning_rate": 0.0007652555921533672,
    "timestep_mean": 455.46875,
    "epsilon_norm": 8.030715942382812,
    "prediction_norm": 6.89270544052124,
    "timestamp": "2025-07-14T12:33:56.310993"
  },
  {
    "epoch": 644,
    "batch_idx": 32,
    "loss": 0.5132657289505005,
    "learning_rate": 0.0007652555921533672,
    "timestep_mean": 472.375,
    "epsilon_norm": 8.02306079864502,
    "prediction_norm": 6.56771183013916,
    "timestamp": "2025-07-14T12:33:56.316171"
  },
  {
    "epoch": 644,
    "batch_idx": 64,
    "loss": 0.3165983557701111,
    "learning_rate": 0.0007652555921533672,
    "timestep_mean": 536.59375,
    "epsilon_norm": 8.719611167907715,
    "prediction_norm": 7.306617736816406,
    "timestamp": "2025-07-14T12:33:56.321404"
  },
  {
    "epoch": 644,
    "batch_idx": 96,
    "loss": 0.1911967396736145,
    "learning_rate": 0.0007652555921533672,
    "timestep_mean": 534.125,
    "epsilon_norm": 7.83561897277832,
    "prediction_norm": 6.803727626800537,
    "timestamp": "2025-07-14T12:33:56.326574"
  },
  {
    "epoch": 644,
    "batch_idx": 128,
    "loss": 0.3173612654209137,
    "learning_rate": 0.0007652555921533672,
    "timestep_mean": 426.625,
    "epsilon_norm": 8.088347434997559,
    "prediction_norm": 6.436386585235596,
    "timestamp": "2025-07-14T12:33:56.331613"
  },
  {
    "epoch": 644,
    "batch_idx": 160,
    "loss": 0.3935908079147339,
    "learning_rate": 0.0007652555921533672,
    "timestep_mean": 514.71875,
    "epsilon_norm": 8.139508247375488,
    "prediction_norm": 6.077547073364258,
    "timestamp": "2025-07-14T12:33:56.336770"
  },
  {
    "epoch": 644,
    "batch_idx": 192,
    "loss": 0.3997633755207062,
    "learning_rate": 0.0007652555921533672,
    "timestep_mean": 475.5,
    "epsilon_norm": 3.8444864749908447,
    "prediction_norm": 2.893115997314453,
    "timestamp": "2025-07-14T12:33:56.342068"
  },
  {
    "epoch": 645,
    "batch_idx": 0,
    "loss": 0.3905363976955414,
    "learning_rate": 0.0007645895004870954,
    "timestep_mean": 411.125,
    "epsilon_norm": 8.436396598815918,
    "prediction_norm": 6.66930627822876,
    "timestamp": "2025-07-14T12:33:56.347217"
  },
  {
    "epoch": 645,
    "batch_idx": 32,
    "loss": 0.24450090527534485,
    "learning_rate": 0.0007645895004870954,
    "timestep_mean": 482.625,
    "epsilon_norm": 6.744765758514404,
    "prediction_norm": 6.065915107727051,
    "timestamp": "2025-07-14T12:33:56.352411"
  },
  {
    "epoch": 645,
    "batch_idx": 64,
    "loss": 0.21027871966362,
    "learning_rate": 0.0007645895004870954,
    "timestep_mean": 564.53125,
    "epsilon_norm": 7.790566444396973,
    "prediction_norm": 7.300525188446045,
    "timestamp": "2025-07-14T12:33:56.357597"
  },
  {
    "epoch": 645,
    "batch_idx": 96,
    "loss": 0.34327274560928345,
    "learning_rate": 0.0007645895004870954,
    "timestep_mean": 490.0,
    "epsilon_norm": 8.534135818481445,
    "prediction_norm": 7.284397602081299,
    "timestamp": "2025-07-14T12:33:56.362957"
  },
  {
    "epoch": 645,
    "batch_idx": 128,
    "loss": 0.20500272512435913,
    "learning_rate": 0.0007645895004870954,
    "timestep_mean": 503.75,
    "epsilon_norm": 8.009263038635254,
    "prediction_norm": 7.656444549560547,
    "timestamp": "2025-07-14T12:33:56.368075"
  },
  {
    "epoch": 645,
    "batch_idx": 160,
    "loss": 0.22429968416690826,
    "learning_rate": 0.0007645895004870954,
    "timestep_mean": 603.59375,
    "epsilon_norm": 8.353876113891602,
    "prediction_norm": 7.073904514312744,
    "timestamp": "2025-07-14T12:33:56.373260"
  },
  {
    "epoch": 645,
    "batch_idx": 192,
    "loss": 0.44868117570877075,
    "learning_rate": 0.0007645895004870954,
    "timestep_mean": 419.625,
    "epsilon_norm": 4.708799839019775,
    "prediction_norm": 3.9499125480651855,
    "timestamp": "2025-07-14T12:33:56.378387"
  },
  {
    "epoch": 646,
    "batch_idx": 0,
    "loss": 0.2848058342933655,
    "learning_rate": 0.0007639227559725334,
    "timestep_mean": 556.40625,
    "epsilon_norm": 8.033105850219727,
    "prediction_norm": 7.172730445861816,
    "timestamp": "2025-07-14T12:33:56.383628"
  },
  {
    "epoch": 646,
    "batch_idx": 32,
    "loss": 0.3173692226409912,
    "learning_rate": 0.0007639227559725334,
    "timestep_mean": 491.46875,
    "epsilon_norm": 7.315510272979736,
    "prediction_norm": 6.231038570404053,
    "timestamp": "2025-07-14T12:33:56.388637"
  },
  {
    "epoch": 646,
    "batch_idx": 64,
    "loss": 0.4452386498451233,
    "learning_rate": 0.0007639227559725334,
    "timestep_mean": 475.21875,
    "epsilon_norm": 8.757743835449219,
    "prediction_norm": 7.533439636230469,
    "timestamp": "2025-07-14T12:33:56.393530"
  },
  {
    "epoch": 646,
    "batch_idx": 96,
    "loss": 0.23899421095848083,
    "learning_rate": 0.0007639227559725334,
    "timestep_mean": 533.78125,
    "epsilon_norm": 6.951777458190918,
    "prediction_norm": 5.545475006103516,
    "timestamp": "2025-07-14T12:33:56.398366"
  },
  {
    "epoch": 646,
    "batch_idx": 128,
    "loss": 0.15030565857887268,
    "learning_rate": 0.0007639227559725334,
    "timestep_mean": 520.84375,
    "epsilon_norm": 8.767779350280762,
    "prediction_norm": 7.976378917694092,
    "timestamp": "2025-07-14T12:33:56.403282"
  },
  {
    "epoch": 646,
    "batch_idx": 160,
    "loss": 0.34310632944107056,
    "learning_rate": 0.0007639227559725334,
    "timestep_mean": 490.875,
    "epsilon_norm": 7.146115303039551,
    "prediction_norm": 6.099623680114746,
    "timestamp": "2025-07-14T12:33:56.408261"
  },
  {
    "epoch": 646,
    "batch_idx": 192,
    "loss": 0.35805532336235046,
    "learning_rate": 0.0007639227559725334,
    "timestep_mean": 432.375,
    "epsilon_norm": 3.9435977935791016,
    "prediction_norm": 4.271261692047119,
    "timestamp": "2025-07-14T12:33:56.413159"
  },
  {
    "epoch": 647,
    "batch_idx": 0,
    "loss": 0.30714210867881775,
    "learning_rate": 0.0007632553602548066,
    "timestep_mean": 484.4375,
    "epsilon_norm": 7.863589286804199,
    "prediction_norm": 6.74052619934082,
    "timestamp": "2025-07-14T12:33:56.418745"
  },
  {
    "epoch": 647,
    "batch_idx": 32,
    "loss": 0.2808490991592407,
    "learning_rate": 0.0007632553602548066,
    "timestep_mean": 552.625,
    "epsilon_norm": 7.096898078918457,
    "prediction_norm": 6.256129741668701,
    "timestamp": "2025-07-14T12:33:56.423830"
  },
  {
    "epoch": 647,
    "batch_idx": 64,
    "loss": 0.2545602023601532,
    "learning_rate": 0.0007632553602548066,
    "timestep_mean": 543.4375,
    "epsilon_norm": 7.872561454772949,
    "prediction_norm": 6.637471675872803,
    "timestamp": "2025-07-14T12:33:56.428841"
  },
  {
    "epoch": 647,
    "batch_idx": 96,
    "loss": 0.23457005620002747,
    "learning_rate": 0.0007632553602548066,
    "timestep_mean": 495.59375,
    "epsilon_norm": 7.2841362953186035,
    "prediction_norm": 6.400679111480713,
    "timestamp": "2025-07-14T12:33:56.434032"
  },
  {
    "epoch": 647,
    "batch_idx": 128,
    "loss": 0.3324359059333801,
    "learning_rate": 0.0007632553602548066,
    "timestep_mean": 461.09375,
    "epsilon_norm": 8.191312789916992,
    "prediction_norm": 6.460608005523682,
    "timestamp": "2025-07-14T12:33:56.438954"
  },
  {
    "epoch": 647,
    "batch_idx": 160,
    "loss": 0.4289602041244507,
    "learning_rate": 0.0007632553602548066,
    "timestep_mean": 422.4375,
    "epsilon_norm": 7.9868879318237305,
    "prediction_norm": 5.608627796173096,
    "timestamp": "2025-07-14T12:33:56.443876"
  },
  {
    "epoch": 647,
    "batch_idx": 192,
    "loss": 0.8217569589614868,
    "learning_rate": 0.0007632553602548066,
    "timestep_mean": 611.125,
    "epsilon_norm": 5.218880653381348,
    "prediction_norm": 2.7726757526397705,
    "timestamp": "2025-07-14T12:33:56.448831"
  },
  {
    "epoch": 648,
    "batch_idx": 0,
    "loss": 0.3316243290901184,
    "learning_rate": 0.000762587314980648,
    "timestep_mean": 410.375,
    "epsilon_norm": 7.5736212730407715,
    "prediction_norm": 4.992012023925781,
    "timestamp": "2025-07-14T12:33:56.453876"
  },
  {
    "epoch": 648,
    "batch_idx": 32,
    "loss": 0.24165946245193481,
    "learning_rate": 0.000762587314980648,
    "timestep_mean": 479.34375,
    "epsilon_norm": 7.177943229675293,
    "prediction_norm": 5.868902206420898,
    "timestamp": "2025-07-14T12:33:56.458982"
  },
  {
    "epoch": 648,
    "batch_idx": 64,
    "loss": 0.3246441185474396,
    "learning_rate": 0.000762587314980648,
    "timestep_mean": 563.53125,
    "epsilon_norm": 8.39729118347168,
    "prediction_norm": 6.677609920501709,
    "timestamp": "2025-07-14T12:33:56.463937"
  },
  {
    "epoch": 648,
    "batch_idx": 96,
    "loss": 0.3237403631210327,
    "learning_rate": 0.000762587314980648,
    "timestep_mean": 486.125,
    "epsilon_norm": 8.943215370178223,
    "prediction_norm": 7.356393337249756,
    "timestamp": "2025-07-14T12:33:56.468832"
  },
  {
    "epoch": 648,
    "batch_idx": 128,
    "loss": 0.26791316270828247,
    "learning_rate": 0.000762587314980648,
    "timestep_mean": 570.5,
    "epsilon_norm": 7.795736312866211,
    "prediction_norm": 6.700916767120361,
    "timestamp": "2025-07-14T12:33:56.473681"
  },
  {
    "epoch": 648,
    "batch_idx": 160,
    "loss": 0.3839603364467621,
    "learning_rate": 0.000762587314980648,
    "timestep_mean": 449.46875,
    "epsilon_norm": 8.15011215209961,
    "prediction_norm": 6.090091705322266,
    "timestamp": "2025-07-14T12:33:56.478552"
  },
  {
    "epoch": 648,
    "batch_idx": 192,
    "loss": 0.15881836414337158,
    "learning_rate": 0.000762587314980648,
    "timestep_mean": 594.75,
    "epsilon_norm": 3.2676241397857666,
    "prediction_norm": 3.205697536468506,
    "timestamp": "2025-07-14T12:33:56.483563"
  },
  {
    "epoch": 649,
    "batch_idx": 0,
    "loss": 0.27116158604621887,
    "learning_rate": 0.0007619186217983925,
    "timestep_mean": 499.375,
    "epsilon_norm": 8.015677452087402,
    "prediction_norm": 6.368290901184082,
    "timestamp": "2025-07-14T12:33:56.488559"
  },
  {
    "epoch": 649,
    "batch_idx": 32,
    "loss": 0.3741832375526428,
    "learning_rate": 0.0007619186217983925,
    "timestep_mean": 500.8125,
    "epsilon_norm": 9.203438758850098,
    "prediction_norm": 8.159893989562988,
    "timestamp": "2025-07-14T12:33:56.493435"
  },
  {
    "epoch": 649,
    "batch_idx": 64,
    "loss": 0.21126818656921387,
    "learning_rate": 0.0007619186217983925,
    "timestep_mean": 481.375,
    "epsilon_norm": 8.544310569763184,
    "prediction_norm": 8.020750999450684,
    "timestamp": "2025-07-14T12:33:56.498394"
  },
  {
    "epoch": 649,
    "batch_idx": 96,
    "loss": 0.19456642866134644,
    "learning_rate": 0.0007619186217983925,
    "timestep_mean": 491.96875,
    "epsilon_norm": 7.3994269371032715,
    "prediction_norm": 7.079113006591797,
    "timestamp": "2025-07-14T12:33:56.503362"
  },
  {
    "epoch": 649,
    "batch_idx": 128,
    "loss": 0.3233974575996399,
    "learning_rate": 0.0007619186217983925,
    "timestep_mean": 464.1875,
    "epsilon_norm": 8.148552894592285,
    "prediction_norm": 6.081429958343506,
    "timestamp": "2025-07-14T12:33:56.508367"
  },
  {
    "epoch": 649,
    "batch_idx": 160,
    "loss": 0.27129995822906494,
    "learning_rate": 0.0007619186217983925,
    "timestep_mean": 511.9375,
    "epsilon_norm": 8.765135765075684,
    "prediction_norm": 7.2297587394714355,
    "timestamp": "2025-07-14T12:33:56.513267"
  },
  {
    "epoch": 649,
    "batch_idx": 192,
    "loss": 0.15148110687732697,
    "learning_rate": 0.0007619186217983925,
    "timestep_mean": 478.625,
    "epsilon_norm": 4.064060211181641,
    "prediction_norm": 3.702565908432007,
    "timestamp": "2025-07-14T12:33:56.518243"
  },
  {
    "epoch": 650,
    "batch_idx": 0,
    "loss": 0.23215660452842712,
    "learning_rate": 0.0007612492823579744,
    "timestep_mean": 602.0625,
    "epsilon_norm": 7.869878768920898,
    "prediction_norm": 6.920848369598389,
    "timestamp": "2025-07-14T12:33:56.523128"
  },
  {
    "epoch": 650,
    "batch_idx": 32,
    "loss": 0.24202987551689148,
    "learning_rate": 0.0007612492823579744,
    "timestep_mean": 506.4375,
    "epsilon_norm": 7.994778633117676,
    "prediction_norm": 6.800198078155518,
    "timestamp": "2025-07-14T12:33:56.528027"
  },
  {
    "epoch": 650,
    "batch_idx": 64,
    "loss": 0.3175536096096039,
    "learning_rate": 0.0007612492823579744,
    "timestep_mean": 470.9375,
    "epsilon_norm": 8.853371620178223,
    "prediction_norm": 7.4099602699279785,
    "timestamp": "2025-07-14T12:33:56.532957"
  },
  {
    "epoch": 650,
    "batch_idx": 96,
    "loss": 0.27173447608947754,
    "learning_rate": 0.0007612492823579744,
    "timestep_mean": 524.21875,
    "epsilon_norm": 8.942248344421387,
    "prediction_norm": 8.65333080291748,
    "timestamp": "2025-07-14T12:33:56.537953"
  },
  {
    "epoch": 650,
    "batch_idx": 128,
    "loss": 0.16122321784496307,
    "learning_rate": 0.0007612492823579744,
    "timestep_mean": 533.65625,
    "epsilon_norm": 6.726527690887451,
    "prediction_norm": 6.129166603088379,
    "timestamp": "2025-07-14T12:33:56.542812"
  },
  {
    "epoch": 650,
    "batch_idx": 160,
    "loss": 0.05539480596780777,
    "learning_rate": 0.0007612492823579744,
    "timestep_mean": 636.25,
    "epsilon_norm": 7.89682674407959,
    "prediction_norm": 7.999429702758789,
    "timestamp": "2025-07-14T12:33:56.547650"
  },
  {
    "epoch": 650,
    "batch_idx": 192,
    "loss": 0.4335595369338989,
    "learning_rate": 0.0007612492823579744,
    "timestep_mean": 302.625,
    "epsilon_norm": 3.3920435905456543,
    "prediction_norm": 2.701435089111328,
    "timestamp": "2025-07-14T12:33:56.552511"
  },
  {
    "epoch": 651,
    "batch_idx": 0,
    "loss": 0.3239936828613281,
    "learning_rate": 0.0007605792983109223,
    "timestep_mean": 568.625,
    "epsilon_norm": 7.645554065704346,
    "prediction_norm": 6.22144889831543,
    "timestamp": "2025-07-14T12:33:56.557385"
  },
  {
    "epoch": 651,
    "batch_idx": 32,
    "loss": 0.29975733160972595,
    "learning_rate": 0.0007605792983109223,
    "timestep_mean": 507.65625,
    "epsilon_norm": 8.253280639648438,
    "prediction_norm": 7.299086093902588,
    "timestamp": "2025-07-14T12:33:56.562277"
  },
  {
    "epoch": 651,
    "batch_idx": 64,
    "loss": 0.39881670475006104,
    "learning_rate": 0.0007605792983109223,
    "timestep_mean": 489.5,
    "epsilon_norm": 8.859770774841309,
    "prediction_norm": 6.52504825592041,
    "timestamp": "2025-07-14T12:33:56.567161"
  },
  {
    "epoch": 651,
    "batch_idx": 96,
    "loss": 0.3138972520828247,
    "learning_rate": 0.0007605792983109223,
    "timestep_mean": 526.59375,
    "epsilon_norm": 7.987207889556885,
    "prediction_norm": 6.441268444061279,
    "timestamp": "2025-07-14T12:33:56.572016"
  },
  {
    "epoch": 651,
    "batch_idx": 128,
    "loss": 0.269340455532074,
    "learning_rate": 0.0007605792983109223,
    "timestep_mean": 452.28125,
    "epsilon_norm": 7.423282146453857,
    "prediction_norm": 6.613715648651123,
    "timestamp": "2025-07-14T12:33:56.576911"
  },
  {
    "epoch": 651,
    "batch_idx": 160,
    "loss": 0.347960889339447,
    "learning_rate": 0.0007605792983109223,
    "timestep_mean": 476.28125,
    "epsilon_norm": 8.196120262145996,
    "prediction_norm": 7.11952543258667,
    "timestamp": "2025-07-14T12:33:56.581879"
  },
  {
    "epoch": 651,
    "batch_idx": 192,
    "loss": 0.32257965207099915,
    "learning_rate": 0.0007605792983109223,
    "timestep_mean": 546.125,
    "epsilon_norm": 3.876582622528076,
    "prediction_norm": 3.267021894454956,
    "timestamp": "2025-07-14T12:33:56.586725"
  },
  {
    "epoch": 652,
    "batch_idx": 0,
    "loss": 0.34115853905677795,
    "learning_rate": 0.0007599086713103547,
    "timestep_mean": 525.59375,
    "epsilon_norm": 7.731894016265869,
    "prediction_norm": 6.036222457885742,
    "timestamp": "2025-07-14T12:33:56.591696"
  },
  {
    "epoch": 652,
    "batch_idx": 32,
    "loss": 0.29058700799942017,
    "learning_rate": 0.0007599086713103547,
    "timestep_mean": 454.0,
    "epsilon_norm": 8.73102855682373,
    "prediction_norm": 7.239794731140137,
    "timestamp": "2025-07-14T12:33:56.596642"
  },
  {
    "epoch": 652,
    "batch_idx": 64,
    "loss": 0.20365649461746216,
    "learning_rate": 0.0007599086713103547,
    "timestep_mean": 541.96875,
    "epsilon_norm": 7.473026275634766,
    "prediction_norm": 7.39298152923584,
    "timestamp": "2025-07-14T12:33:56.601529"
  },
  {
    "epoch": 652,
    "batch_idx": 96,
    "loss": 0.2291761189699173,
    "learning_rate": 0.0007599086713103547,
    "timestep_mean": 564.65625,
    "epsilon_norm": 7.608014106750488,
    "prediction_norm": 6.720310688018799,
    "timestamp": "2025-07-14T12:33:56.606404"
  },
  {
    "epoch": 652,
    "batch_idx": 128,
    "loss": 0.32523658871650696,
    "learning_rate": 0.0007599086713103547,
    "timestep_mean": 469.59375,
    "epsilon_norm": 7.40668249130249,
    "prediction_norm": 6.229211807250977,
    "timestamp": "2025-07-14T12:33:56.611351"
  },
  {
    "epoch": 652,
    "batch_idx": 160,
    "loss": 0.34405821561813354,
    "learning_rate": 0.0007599086713103547,
    "timestep_mean": 603.84375,
    "epsilon_norm": 7.657544136047363,
    "prediction_norm": 6.264958381652832,
    "timestamp": "2025-07-14T12:33:56.616262"
  },
  {
    "epoch": 652,
    "batch_idx": 192,
    "loss": 0.055815499275922775,
    "learning_rate": 0.0007599086713103547,
    "timestep_mean": 534.875,
    "epsilon_norm": 3.5376598834991455,
    "prediction_norm": 3.745274543762207,
    "timestamp": "2025-07-14T12:33:56.621521"
  },
  {
    "epoch": 653,
    "batch_idx": 0,
    "loss": 0.19275939464569092,
    "learning_rate": 0.0007592374030109777,
    "timestep_mean": 516.25,
    "epsilon_norm": 8.01553726196289,
    "prediction_norm": 7.04543399810791,
    "timestamp": "2025-07-14T12:33:56.626695"
  },
  {
    "epoch": 653,
    "batch_idx": 32,
    "loss": 0.37136244773864746,
    "learning_rate": 0.0007592374030109777,
    "timestep_mean": 493.09375,
    "epsilon_norm": 7.830696105957031,
    "prediction_norm": 6.598687648773193,
    "timestamp": "2025-07-14T12:33:56.631915"
  },
  {
    "epoch": 653,
    "batch_idx": 64,
    "loss": 0.22191613912582397,
    "learning_rate": 0.0007592374030109777,
    "timestep_mean": 520.84375,
    "epsilon_norm": 7.884151935577393,
    "prediction_norm": 6.3963942527771,
    "timestamp": "2025-07-14T12:33:56.636990"
  },
  {
    "epoch": 653,
    "batch_idx": 96,
    "loss": 0.24688522517681122,
    "learning_rate": 0.0007592374030109777,
    "timestep_mean": 497.96875,
    "epsilon_norm": 7.3010149002075195,
    "prediction_norm": 6.110013961791992,
    "timestamp": "2025-07-14T12:33:56.642168"
  },
  {
    "epoch": 653,
    "batch_idx": 128,
    "loss": 0.13297244906425476,
    "learning_rate": 0.0007592374030109777,
    "timestep_mean": 523.5,
    "epsilon_norm": 7.129900932312012,
    "prediction_norm": 5.831419944763184,
    "timestamp": "2025-07-14T12:33:56.647302"
  },
  {
    "epoch": 653,
    "batch_idx": 160,
    "loss": 0.2752838134765625,
    "learning_rate": 0.0007592374030109777,
    "timestep_mean": 540.25,
    "epsilon_norm": 8.038589477539062,
    "prediction_norm": 5.952642440795898,
    "timestamp": "2025-07-14T12:33:56.652494"
  },
  {
    "epoch": 653,
    "batch_idx": 192,
    "loss": 0.4086565375328064,
    "learning_rate": 0.0007592374030109777,
    "timestep_mean": 399.625,
    "epsilon_norm": 2.7189202308654785,
    "prediction_norm": 1.7213104963302612,
    "timestamp": "2025-07-14T12:33:56.657687"
  },
  {
    "epoch": 654,
    "batch_idx": 0,
    "loss": 0.21331852674484253,
    "learning_rate": 0.0007585654950690787,
    "timestep_mean": 463.875,
    "epsilon_norm": 7.286875247955322,
    "prediction_norm": 6.3824896812438965,
    "timestamp": "2025-07-14T12:33:56.662881"
  },
  {
    "epoch": 654,
    "batch_idx": 32,
    "loss": 0.31139445304870605,
    "learning_rate": 0.0007585654950690787,
    "timestep_mean": 487.4375,
    "epsilon_norm": 8.365384101867676,
    "prediction_norm": 6.350221633911133,
    "timestamp": "2025-07-14T12:33:56.668025"
  },
  {
    "epoch": 654,
    "batch_idx": 64,
    "loss": 0.2564592659473419,
    "learning_rate": 0.0007585654950690787,
    "timestep_mean": 495.15625,
    "epsilon_norm": 7.421371936798096,
    "prediction_norm": 5.668842315673828,
    "timestamp": "2025-07-14T12:33:56.673126"
  },
  {
    "epoch": 654,
    "batch_idx": 96,
    "loss": 0.4760441780090332,
    "learning_rate": 0.0007585654950690787,
    "timestep_mean": 524.6875,
    "epsilon_norm": 7.9616851806640625,
    "prediction_norm": 5.762088775634766,
    "timestamp": "2025-07-14T12:33:56.678343"
  },
  {
    "epoch": 654,
    "batch_idx": 128,
    "loss": 0.23124109208583832,
    "learning_rate": 0.0007585654950690787,
    "timestep_mean": 513.09375,
    "epsilon_norm": 9.01987075805664,
    "prediction_norm": 7.6784586906433105,
    "timestamp": "2025-07-14T12:33:56.683694"
  },
  {
    "epoch": 654,
    "batch_idx": 160,
    "loss": 0.26342928409576416,
    "learning_rate": 0.0007585654950690787,
    "timestep_mean": 502.71875,
    "epsilon_norm": 8.10275650024414,
    "prediction_norm": 6.909658432006836,
    "timestamp": "2025-07-14T12:33:56.688866"
  },
  {
    "epoch": 654,
    "batch_idx": 192,
    "loss": 0.43568146228790283,
    "learning_rate": 0.0007585654950690787,
    "timestep_mean": 465.0,
    "epsilon_norm": 3.523303747177124,
    "prediction_norm": 2.6382384300231934,
    "timestamp": "2025-07-14T12:33:56.693962"
  },
  {
    "epoch": 655,
    "batch_idx": 0,
    "loss": 0.33347854018211365,
    "learning_rate": 0.0007578929491425238,
    "timestep_mean": 444.96875,
    "epsilon_norm": 6.869692802429199,
    "prediction_norm": 5.730037212371826,
    "timestamp": "2025-07-14T12:33:56.699145"
  },
  {
    "epoch": 655,
    "batch_idx": 32,
    "loss": 0.32531946897506714,
    "learning_rate": 0.0007578929491425238,
    "timestep_mean": 543.6875,
    "epsilon_norm": 8.634956359863281,
    "prediction_norm": 7.322362899780273,
    "timestamp": "2025-07-14T12:33:56.709878"
  },
  {
    "epoch": 655,
    "batch_idx": 64,
    "loss": 0.2675991654396057,
    "learning_rate": 0.0007578929491425238,
    "timestep_mean": 497.9375,
    "epsilon_norm": 8.394970893859863,
    "prediction_norm": 8.241466522216797,
    "timestamp": "2025-07-14T12:33:56.715119"
  },
  {
    "epoch": 655,
    "batch_idx": 96,
    "loss": 0.2502947151660919,
    "learning_rate": 0.0007578929491425238,
    "timestep_mean": 519.8125,
    "epsilon_norm": 8.65047836303711,
    "prediction_norm": 7.968364715576172,
    "timestamp": "2025-07-14T12:33:56.720395"
  },
  {
    "epoch": 655,
    "batch_idx": 128,
    "loss": 0.2914716303348541,
    "learning_rate": 0.0007578929491425238,
    "timestep_mean": 594.53125,
    "epsilon_norm": 8.207167625427246,
    "prediction_norm": 7.524071216583252,
    "timestamp": "2025-07-14T12:33:56.725615"
  },
  {
    "epoch": 655,
    "batch_idx": 160,
    "loss": 0.4341653287410736,
    "learning_rate": 0.0007578929491425238,
    "timestep_mean": 470.90625,
    "epsilon_norm": 8.088562965393066,
    "prediction_norm": 6.729377746582031,
    "timestamp": "2025-07-14T12:33:56.730796"
  },
  {
    "epoch": 655,
    "batch_idx": 192,
    "loss": 0.2333914339542389,
    "learning_rate": 0.0007578929491425238,
    "timestep_mean": 538.25,
    "epsilon_norm": 3.651455879211426,
    "prediction_norm": 3.045086622238159,
    "timestamp": "2025-07-14T12:33:56.735985"
  },
  {
    "epoch": 656,
    "batch_idx": 0,
    "loss": 0.4984291195869446,
    "learning_rate": 0.0007572197668907534,
    "timestep_mean": 478.84375,
    "epsilon_norm": 8.56256103515625,
    "prediction_norm": 7.146698951721191,
    "timestamp": "2025-07-14T12:33:56.741205"
  },
  {
    "epoch": 656,
    "batch_idx": 32,
    "loss": 0.27669018507003784,
    "learning_rate": 0.0007572197668907534,
    "timestep_mean": 573.03125,
    "epsilon_norm": 7.330681324005127,
    "prediction_norm": 6.625913619995117,
    "timestamp": "2025-07-14T12:33:56.746323"
  },
  {
    "epoch": 656,
    "batch_idx": 64,
    "loss": 0.20745790004730225,
    "learning_rate": 0.0007572197668907534,
    "timestep_mean": 451.8125,
    "epsilon_norm": 7.365075588226318,
    "prediction_norm": 7.0602006912231445,
    "timestamp": "2025-07-14T12:33:56.751420"
  },
  {
    "epoch": 656,
    "batch_idx": 96,
    "loss": 0.21922102570533752,
    "learning_rate": 0.0007572197668907534,
    "timestep_mean": 468.09375,
    "epsilon_norm": 7.3121724128723145,
    "prediction_norm": 6.573883533477783,
    "timestamp": "2025-07-14T12:33:56.756633"
  },
  {
    "epoch": 656,
    "batch_idx": 128,
    "loss": 0.239660382270813,
    "learning_rate": 0.0007572197668907534,
    "timestep_mean": 499.75,
    "epsilon_norm": 8.87191104888916,
    "prediction_norm": 7.905396938323975,
    "timestamp": "2025-07-14T12:33:56.761794"
  },
  {
    "epoch": 656,
    "batch_idx": 160,
    "loss": 0.39676547050476074,
    "learning_rate": 0.0007572197668907534,
    "timestep_mean": 384.625,
    "epsilon_norm": 8.048383712768555,
    "prediction_norm": 5.515355110168457,
    "timestamp": "2025-07-14T12:33:56.766890"
  },
  {
    "epoch": 656,
    "batch_idx": 192,
    "loss": 0.1311635673046112,
    "learning_rate": 0.0007572197668907534,
    "timestep_mean": 684.375,
    "epsilon_norm": 4.514763832092285,
    "prediction_norm": 3.956089973449707,
    "timestamp": "2025-07-14T12:33:56.771804"
  },
  {
    "epoch": 657,
    "batch_idx": 0,
    "loss": 0.28292378783226013,
    "learning_rate": 0.0007565459499747775,
    "timestep_mean": 431.125,
    "epsilon_norm": 6.861475944519043,
    "prediction_norm": 5.271775722503662,
    "timestamp": "2025-07-14T12:33:56.776708"
  },
  {
    "epoch": 657,
    "batch_idx": 32,
    "loss": 0.2536929249763489,
    "learning_rate": 0.0007565459499747775,
    "timestep_mean": 501.5625,
    "epsilon_norm": 7.353944301605225,
    "prediction_norm": 6.012862205505371,
    "timestamp": "2025-07-14T12:33:56.781571"
  },
  {
    "epoch": 657,
    "batch_idx": 64,
    "loss": 0.2154337763786316,
    "learning_rate": 0.0007565459499747775,
    "timestep_mean": 508.3125,
    "epsilon_norm": 8.07831859588623,
    "prediction_norm": 6.831132888793945,
    "timestamp": "2025-07-14T12:33:56.786455"
  },
  {
    "epoch": 657,
    "batch_idx": 96,
    "loss": 0.4976065754890442,
    "learning_rate": 0.0007565459499747775,
    "timestep_mean": 440.09375,
    "epsilon_norm": 8.080889701843262,
    "prediction_norm": 5.106652736663818,
    "timestamp": "2025-07-14T12:33:56.791348"
  },
  {
    "epoch": 657,
    "batch_idx": 128,
    "loss": 0.17264202237129211,
    "learning_rate": 0.0007565459499747775,
    "timestep_mean": 604.96875,
    "epsilon_norm": 8.362710952758789,
    "prediction_norm": 7.157307147979736,
    "timestamp": "2025-07-14T12:33:56.796309"
  },
  {
    "epoch": 657,
    "batch_idx": 160,
    "loss": 0.22492823004722595,
    "learning_rate": 0.0007565459499747775,
    "timestep_mean": 590.0,
    "epsilon_norm": 8.20622444152832,
    "prediction_norm": 7.241015434265137,
    "timestamp": "2025-07-14T12:33:56.801158"
  },
  {
    "epoch": 657,
    "batch_idx": 192,
    "loss": 0.22150534391403198,
    "learning_rate": 0.0007565459499747775,
    "timestep_mean": 393.625,
    "epsilon_norm": 4.004940986633301,
    "prediction_norm": 3.142150402069092,
    "timestamp": "2025-07-14T12:33:56.805949"
  },
  {
    "epoch": 658,
    "batch_idx": 0,
    "loss": 0.4281688332557678,
    "learning_rate": 0.0007558715000571726,
    "timestep_mean": 398.875,
    "epsilon_norm": 8.08753490447998,
    "prediction_norm": 6.163640022277832,
    "timestamp": "2025-07-14T12:33:56.810858"
  },
  {
    "epoch": 658,
    "batch_idx": 32,
    "loss": 0.18885616958141327,
    "learning_rate": 0.0007558715000571726,
    "timestep_mean": 502.6875,
    "epsilon_norm": 8.096415519714355,
    "prediction_norm": 6.858583450317383,
    "timestamp": "2025-07-14T12:33:56.815763"
  },
  {
    "epoch": 658,
    "batch_idx": 64,
    "loss": 0.11440590769052505,
    "learning_rate": 0.0007558715000571726,
    "timestep_mean": 579.5,
    "epsilon_norm": 8.219388961791992,
    "prediction_norm": 7.493962287902832,
    "timestamp": "2025-07-14T12:33:56.820607"
  },
  {
    "epoch": 658,
    "batch_idx": 96,
    "loss": 0.4145386815071106,
    "learning_rate": 0.0007558715000571726,
    "timestep_mean": 438.1875,
    "epsilon_norm": 8.857643127441406,
    "prediction_norm": 7.21046257019043,
    "timestamp": "2025-07-14T12:33:56.825496"
  },
  {
    "epoch": 658,
    "batch_idx": 128,
    "loss": 0.12989547848701477,
    "learning_rate": 0.0007558715000571726,
    "timestep_mean": 603.75,
    "epsilon_norm": 7.9338555335998535,
    "prediction_norm": 8.038113594055176,
    "timestamp": "2025-07-14T12:33:56.830403"
  },
  {
    "epoch": 658,
    "batch_idx": 160,
    "loss": 0.42690712213516235,
    "learning_rate": 0.0007558715000571726,
    "timestep_mean": 396.4375,
    "epsilon_norm": 9.109481811523438,
    "prediction_norm": 7.366216659545898,
    "timestamp": "2025-07-14T12:33:56.835319"
  },
  {
    "epoch": 658,
    "batch_idx": 192,
    "loss": 0.12375202029943466,
    "learning_rate": 0.0007558715000571726,
    "timestep_mean": 454.5,
    "epsilon_norm": 3.9060091972351074,
    "prediction_norm": 3.4665238857269287,
    "timestamp": "2025-07-14T12:33:56.840193"
  },
  {
    "epoch": 659,
    "batch_idx": 0,
    "loss": 0.29175660014152527,
    "learning_rate": 0.0007551964188020766,
    "timestep_mean": 470.5,
    "epsilon_norm": 7.139786243438721,
    "prediction_norm": 5.848505020141602,
    "timestamp": "2025-07-14T12:33:56.845141"
  },
  {
    "epoch": 659,
    "batch_idx": 32,
    "loss": 0.3001645505428314,
    "learning_rate": 0.0007551964188020766,
    "timestep_mean": 397.40625,
    "epsilon_norm": 8.103605270385742,
    "prediction_norm": 6.699901103973389,
    "timestamp": "2025-07-14T12:33:56.850071"
  },
  {
    "epoch": 659,
    "batch_idx": 64,
    "loss": 0.1455160677433014,
    "learning_rate": 0.0007551964188020766,
    "timestep_mean": 526.78125,
    "epsilon_norm": 7.320567607879639,
    "prediction_norm": 7.22662353515625,
    "timestamp": "2025-07-14T12:33:56.854983"
  },
  {
    "epoch": 659,
    "batch_idx": 96,
    "loss": 0.39053890109062195,
    "learning_rate": 0.0007551964188020766,
    "timestep_mean": 526.40625,
    "epsilon_norm": 8.463397026062012,
    "prediction_norm": 6.881779670715332,
    "timestamp": "2025-07-14T12:33:56.859927"
  },
  {
    "epoch": 659,
    "batch_idx": 128,
    "loss": 0.12556937336921692,
    "learning_rate": 0.0007551964188020766,
    "timestep_mean": 492.3125,
    "epsilon_norm": 7.9975056648254395,
    "prediction_norm": 7.345346927642822,
    "timestamp": "2025-07-14T12:33:56.864814"
  },
  {
    "epoch": 659,
    "batch_idx": 160,
    "loss": 0.31417182087898254,
    "learning_rate": 0.0007551964188020766,
    "timestep_mean": 531.65625,
    "epsilon_norm": 8.301419258117676,
    "prediction_norm": 7.862356185913086,
    "timestamp": "2025-07-14T12:33:56.869603"
  },
  {
    "epoch": 659,
    "batch_idx": 192,
    "loss": 0.36857712268829346,
    "learning_rate": 0.0007551964188020766,
    "timestep_mean": 566.375,
    "epsilon_norm": 4.458182334899902,
    "prediction_norm": 4.381822109222412,
    "timestamp": "2025-07-14T12:33:56.874418"
  },
  {
    "epoch": 660,
    "batch_idx": 0,
    "loss": 0.24116802215576172,
    "learning_rate": 0.0007545207078751858,
    "timestep_mean": 389.125,
    "epsilon_norm": 7.20411491394043,
    "prediction_norm": 5.919898986816406,
    "timestamp": "2025-07-14T12:33:56.879301"
  },
  {
    "epoch": 660,
    "batch_idx": 32,
    "loss": 0.2504032850265503,
    "learning_rate": 0.0007545207078751858,
    "timestep_mean": 529.34375,
    "epsilon_norm": 8.379629135131836,
    "prediction_norm": 7.591984748840332,
    "timestamp": "2025-07-14T12:33:56.884170"
  },
  {
    "epoch": 660,
    "batch_idx": 64,
    "loss": 0.2512235641479492,
    "learning_rate": 0.0007545207078751858,
    "timestep_mean": 549.34375,
    "epsilon_norm": 7.585609436035156,
    "prediction_norm": 6.062295436859131,
    "timestamp": "2025-07-14T12:33:56.889128"
  },
  {
    "epoch": 660,
    "batch_idx": 96,
    "loss": 0.35145676136016846,
    "learning_rate": 0.0007545207078751858,
    "timestep_mean": 479.8125,
    "epsilon_norm": 7.701045036315918,
    "prediction_norm": 5.996349811553955,
    "timestamp": "2025-07-14T12:33:56.893986"
  },
  {
    "epoch": 660,
    "batch_idx": 128,
    "loss": 0.2442249357700348,
    "learning_rate": 0.0007545207078751858,
    "timestep_mean": 486.21875,
    "epsilon_norm": 7.512790679931641,
    "prediction_norm": 6.416060447692871,
    "timestamp": "2025-07-14T12:33:56.898874"
  },
  {
    "epoch": 660,
    "batch_idx": 160,
    "loss": 0.27607399225234985,
    "learning_rate": 0.0007545207078751858,
    "timestep_mean": 474.8125,
    "epsilon_norm": 8.968850135803223,
    "prediction_norm": 7.944507122039795,
    "timestamp": "2025-07-14T12:33:56.903758"
  },
  {
    "epoch": 660,
    "batch_idx": 192,
    "loss": 0.1426810324192047,
    "learning_rate": 0.0007545207078751858,
    "timestep_mean": 539.875,
    "epsilon_norm": 4.030832767486572,
    "prediction_norm": 3.9579694271087646,
    "timestamp": "2025-07-14T12:33:56.908865"
  },
  {
    "epoch": 661,
    "batch_idx": 0,
    "loss": 0.22774939239025116,
    "learning_rate": 0.0007538443689437493,
    "timestep_mean": 563.09375,
    "epsilon_norm": 8.487019538879395,
    "prediction_norm": 7.628133773803711,
    "timestamp": "2025-07-14T12:33:56.913764"
  },
  {
    "epoch": 661,
    "batch_idx": 32,
    "loss": 0.22025220096111298,
    "learning_rate": 0.0007538443689437493,
    "timestep_mean": 533.5625,
    "epsilon_norm": 7.738463878631592,
    "prediction_norm": 6.772290229797363,
    "timestamp": "2025-07-14T12:33:56.918666"
  },
  {
    "epoch": 661,
    "batch_idx": 64,
    "loss": 0.32942867279052734,
    "learning_rate": 0.0007538443689437493,
    "timestep_mean": 429.0,
    "epsilon_norm": 8.455859184265137,
    "prediction_norm": 6.91723108291626,
    "timestamp": "2025-07-14T12:33:56.923501"
  },
  {
    "epoch": 661,
    "batch_idx": 96,
    "loss": 0.36954641342163086,
    "learning_rate": 0.0007538443689437493,
    "timestep_mean": 428.34375,
    "epsilon_norm": 7.938462257385254,
    "prediction_norm": 5.738821506500244,
    "timestamp": "2025-07-14T12:33:56.928407"
  },
  {
    "epoch": 661,
    "batch_idx": 128,
    "loss": 0.27338922023773193,
    "learning_rate": 0.0007538443689437493,
    "timestep_mean": 413.3125,
    "epsilon_norm": 7.317031383514404,
    "prediction_norm": 5.904715061187744,
    "timestamp": "2025-07-14T12:33:56.933372"
  },
  {
    "epoch": 661,
    "batch_idx": 160,
    "loss": 0.3228236734867096,
    "learning_rate": 0.0007538443689437493,
    "timestep_mean": 568.0,
    "epsilon_norm": 8.110023498535156,
    "prediction_norm": 6.692657947540283,
    "timestamp": "2025-07-14T12:33:56.938299"
  },
  {
    "epoch": 661,
    "batch_idx": 192,
    "loss": 0.11147375404834747,
    "learning_rate": 0.0007538443689437493,
    "timestep_mean": 499.125,
    "epsilon_norm": 4.6279778480529785,
    "prediction_norm": 3.915379285812378,
    "timestamp": "2025-07-14T12:33:56.943189"
  },
  {
    "epoch": 662,
    "batch_idx": 0,
    "loss": 0.2299506962299347,
    "learning_rate": 0.0007531674036765665,
    "timestep_mean": 508.1875,
    "epsilon_norm": 6.878459930419922,
    "prediction_norm": 5.932253837585449,
    "timestamp": "2025-07-14T12:33:56.948160"
  },
  {
    "epoch": 662,
    "batch_idx": 32,
    "loss": 0.297787070274353,
    "learning_rate": 0.0007531674036765665,
    "timestep_mean": 454.90625,
    "epsilon_norm": 8.832953453063965,
    "prediction_norm": 8.161592483520508,
    "timestamp": "2025-07-14T12:33:56.953035"
  },
  {
    "epoch": 662,
    "batch_idx": 64,
    "loss": 0.48653432726860046,
    "learning_rate": 0.0007531674036765665,
    "timestep_mean": 456.40625,
    "epsilon_norm": 9.057602882385254,
    "prediction_norm": 6.136070728302002,
    "timestamp": "2025-07-14T12:33:56.957951"
  },
  {
    "epoch": 662,
    "batch_idx": 96,
    "loss": 0.29679635167121887,
    "learning_rate": 0.0007531674036765665,
    "timestep_mean": 474.5,
    "epsilon_norm": 8.108924865722656,
    "prediction_norm": 6.26950740814209,
    "timestamp": "2025-07-14T12:33:56.962927"
  },
  {
    "epoch": 662,
    "batch_idx": 128,
    "loss": 0.31994304060935974,
    "learning_rate": 0.0007531674036765665,
    "timestep_mean": 559.1875,
    "epsilon_norm": 8.33780574798584,
    "prediction_norm": 6.418641567230225,
    "timestamp": "2025-07-14T12:33:56.967812"
  },
  {
    "epoch": 662,
    "batch_idx": 160,
    "loss": 0.2006562352180481,
    "learning_rate": 0.0007531674036765665,
    "timestep_mean": 518.1875,
    "epsilon_norm": 7.71318244934082,
    "prediction_norm": 6.978662967681885,
    "timestamp": "2025-07-14T12:33:56.972614"
  },
  {
    "epoch": 662,
    "batch_idx": 192,
    "loss": 0.1877208948135376,
    "learning_rate": 0.0007531674036765665,
    "timestep_mean": 619.0,
    "epsilon_norm": 3.2322680950164795,
    "prediction_norm": 2.9753341674804688,
    "timestamp": "2025-07-14T12:33:56.977458"
  },
  {
    "epoch": 663,
    "batch_idx": 0,
    "loss": 0.3369370698928833,
    "learning_rate": 0.0007524898137439817,
    "timestep_mean": 490.75,
    "epsilon_norm": 6.896925926208496,
    "prediction_norm": 5.663956165313721,
    "timestamp": "2025-07-14T12:33:56.982329"
  },
  {
    "epoch": 663,
    "batch_idx": 32,
    "loss": 0.34791162610054016,
    "learning_rate": 0.0007524898137439817,
    "timestep_mean": 427.5625,
    "epsilon_norm": 7.374258518218994,
    "prediction_norm": 5.673062801361084,
    "timestamp": "2025-07-14T12:33:56.987207"
  },
  {
    "epoch": 663,
    "batch_idx": 64,
    "loss": 0.19908291101455688,
    "learning_rate": 0.0007524898137439817,
    "timestep_mean": 580.59375,
    "epsilon_norm": 6.908322334289551,
    "prediction_norm": 6.139412879943848,
    "timestamp": "2025-07-14T12:33:56.992184"
  },
  {
    "epoch": 663,
    "batch_idx": 96,
    "loss": 0.36347219347953796,
    "learning_rate": 0.0007524898137439817,
    "timestep_mean": 549.28125,
    "epsilon_norm": 9.53429889678955,
    "prediction_norm": 7.776726245880127,
    "timestamp": "2025-07-14T12:33:56.997075"
  },
  {
    "epoch": 663,
    "batch_idx": 128,
    "loss": 0.4749288856983185,
    "learning_rate": 0.0007524898137439817,
    "timestep_mean": 439.4375,
    "epsilon_norm": 7.138655185699463,
    "prediction_norm": 5.444093704223633,
    "timestamp": "2025-07-14T12:33:57.001915"
  },
  {
    "epoch": 663,
    "batch_idx": 160,
    "loss": 0.21538503468036652,
    "learning_rate": 0.0007524898137439817,
    "timestep_mean": 588.46875,
    "epsilon_norm": 7.2037858963012695,
    "prediction_norm": 6.1669020652771,
    "timestamp": "2025-07-14T12:33:57.006785"
  },
  {
    "epoch": 663,
    "batch_idx": 192,
    "loss": 0.0924203097820282,
    "learning_rate": 0.0007524898137439817,
    "timestep_mean": 601.875,
    "epsilon_norm": 3.6655116081237793,
    "prediction_norm": 4.140848159790039,
    "timestamp": "2025-07-14T12:33:57.011632"
  },
  {
    "epoch": 664,
    "batch_idx": 0,
    "loss": 0.2300148457288742,
    "learning_rate": 0.0007518116008178805,
    "timestep_mean": 506.3125,
    "epsilon_norm": 8.55940055847168,
    "prediction_norm": 7.3867597579956055,
    "timestamp": "2025-07-14T12:33:57.016666"
  },
  {
    "epoch": 664,
    "batch_idx": 32,
    "loss": 0.27753371000289917,
    "learning_rate": 0.0007518116008178805,
    "timestep_mean": 388.84375,
    "epsilon_norm": 7.871447563171387,
    "prediction_norm": 6.748367786407471,
    "timestamp": "2025-07-14T12:33:57.021588"
  },
  {
    "epoch": 664,
    "batch_idx": 64,
    "loss": 0.19230887293815613,
    "learning_rate": 0.0007518116008178805,
    "timestep_mean": 490.0,
    "epsilon_norm": 6.477114200592041,
    "prediction_norm": 5.2944488525390625,
    "timestamp": "2025-07-14T12:33:57.026504"
  },
  {
    "epoch": 664,
    "batch_idx": 96,
    "loss": 0.40744462609291077,
    "learning_rate": 0.0007518116008178805,
    "timestep_mean": 481.625,
    "epsilon_norm": 8.483725547790527,
    "prediction_norm": 6.468848705291748,
    "timestamp": "2025-07-14T12:33:57.031427"
  },
  {
    "epoch": 664,
    "batch_idx": 128,
    "loss": 0.4257373511791229,
    "learning_rate": 0.0007518116008178805,
    "timestep_mean": 530.90625,
    "epsilon_norm": 8.85273265838623,
    "prediction_norm": 6.814853191375732,
    "timestamp": "2025-07-14T12:33:57.036386"
  },
  {
    "epoch": 664,
    "batch_idx": 160,
    "loss": 0.27733293175697327,
    "learning_rate": 0.0007518116008178805,
    "timestep_mean": 461.1875,
    "epsilon_norm": 6.9330644607543945,
    "prediction_norm": 5.5962653160095215,
    "timestamp": "2025-07-14T12:33:57.041350"
  },
  {
    "epoch": 664,
    "batch_idx": 192,
    "loss": 0.16313736140727997,
    "learning_rate": 0.0007518116008178805,
    "timestep_mean": 498.625,
    "epsilon_norm": 2.4799723625183105,
    "prediction_norm": 1.8423582315444946,
    "timestamp": "2025-07-14T12:33:57.046302"
  },
  {
    "epoch": 665,
    "batch_idx": 0,
    "loss": 0.22645150125026703,
    "learning_rate": 0.0007511327665716864,
    "timestep_mean": 436.71875,
    "epsilon_norm": 8.140690803527832,
    "prediction_norm": 6.276859283447266,
    "timestamp": "2025-07-14T12:33:57.051148"
  },
  {
    "epoch": 665,
    "batch_idx": 32,
    "loss": 0.16493193805217743,
    "learning_rate": 0.0007511327665716864,
    "timestep_mean": 548.625,
    "epsilon_norm": 7.114961624145508,
    "prediction_norm": 6.495859622955322,
    "timestamp": "2025-07-14T12:33:57.056014"
  },
  {
    "epoch": 665,
    "batch_idx": 64,
    "loss": 0.4033324718475342,
    "learning_rate": 0.0007511327665716864,
    "timestep_mean": 490.5625,
    "epsilon_norm": 9.021688461303711,
    "prediction_norm": 7.5743184089660645,
    "timestamp": "2025-07-14T12:33:57.060910"
  },
  {
    "epoch": 665,
    "batch_idx": 96,
    "loss": 0.10600543022155762,
    "learning_rate": 0.0007511327665716864,
    "timestep_mean": 536.6875,
    "epsilon_norm": 6.911126136779785,
    "prediction_norm": 6.771701812744141,
    "timestamp": "2025-07-14T12:33:57.065826"
  },
  {
    "epoch": 665,
    "batch_idx": 128,
    "loss": 0.4160156846046448,
    "learning_rate": 0.0007511327665716864,
    "timestep_mean": 454.15625,
    "epsilon_norm": 8.771634101867676,
    "prediction_norm": 6.874095916748047,
    "timestamp": "2025-07-14T12:33:57.070682"
  },
  {
    "epoch": 665,
    "batch_idx": 160,
    "loss": 0.2662164568901062,
    "learning_rate": 0.0007511327665716864,
    "timestep_mean": 450.4375,
    "epsilon_norm": 7.310688495635986,
    "prediction_norm": 6.1407976150512695,
    "timestamp": "2025-07-14T12:33:57.075678"
  },
  {
    "epoch": 665,
    "batch_idx": 192,
    "loss": 0.2216653823852539,
    "learning_rate": 0.0007511327665716864,
    "timestep_mean": 697.75,
    "epsilon_norm": 3.9505224227905273,
    "prediction_norm": 3.2221505641937256,
    "timestamp": "2025-07-14T12:33:57.080564"
  },
  {
    "epoch": 666,
    "batch_idx": 0,
    "loss": 0.36702901124954224,
    "learning_rate": 0.0007504533126803551,
    "timestep_mean": 500.0,
    "epsilon_norm": 8.162056922912598,
    "prediction_norm": 6.620540142059326,
    "timestamp": "2025-07-14T12:33:57.085505"
  },
  {
    "epoch": 666,
    "batch_idx": 32,
    "loss": 0.3611189126968384,
    "learning_rate": 0.0007504533126803551,
    "timestep_mean": 487.625,
    "epsilon_norm": 7.765607833862305,
    "prediction_norm": 6.057906150817871,
    "timestamp": "2025-07-14T12:33:57.090495"
  },
  {
    "epoch": 666,
    "batch_idx": 64,
    "loss": 0.24229595065116882,
    "learning_rate": 0.0007504533126803551,
    "timestep_mean": 566.90625,
    "epsilon_norm": 8.349908828735352,
    "prediction_norm": 7.3548583984375,
    "timestamp": "2025-07-14T12:33:57.095421"
  },
  {
    "epoch": 666,
    "batch_idx": 96,
    "loss": 0.20416510105133057,
    "learning_rate": 0.0007504533126803551,
    "timestep_mean": 527.90625,
    "epsilon_norm": 8.814374923706055,
    "prediction_norm": 7.933847904205322,
    "timestamp": "2025-07-14T12:33:57.100322"
  },
  {
    "epoch": 666,
    "batch_idx": 128,
    "loss": 0.2609141170978546,
    "learning_rate": 0.0007504533126803551,
    "timestep_mean": 574.34375,
    "epsilon_norm": 8.580775260925293,
    "prediction_norm": 7.292761325836182,
    "timestamp": "2025-07-14T12:33:57.105204"
  },
  {
    "epoch": 666,
    "batch_idx": 160,
    "loss": 0.2128821164369583,
    "learning_rate": 0.0007504533126803551,
    "timestep_mean": 523.46875,
    "epsilon_norm": 7.334185600280762,
    "prediction_norm": 6.415692329406738,
    "timestamp": "2025-07-14T12:33:57.110186"
  },
  {
    "epoch": 666,
    "batch_idx": 192,
    "loss": 0.19036024808883667,
    "learning_rate": 0.0007504533126803551,
    "timestep_mean": 646.5,
    "epsilon_norm": 3.981233835220337,
    "prediction_norm": 3.5432448387145996,
    "timestamp": "2025-07-14T12:33:57.115066"
  },
  {
    "epoch": 667,
    "batch_idx": 0,
    "loss": 0.1657007336616516,
    "learning_rate": 0.0007497732408203716,
    "timestep_mean": 588.96875,
    "epsilon_norm": 7.375016212463379,
    "prediction_norm": 7.098101615905762,
    "timestamp": "2025-07-14T12:33:57.119949"
  },
  {
    "epoch": 667,
    "batch_idx": 32,
    "loss": 0.20833346247673035,
    "learning_rate": 0.0007497732408203716,
    "timestep_mean": 492.15625,
    "epsilon_norm": 8.759801864624023,
    "prediction_norm": 7.828646183013916,
    "timestamp": "2025-07-14T12:33:57.124915"
  },
  {
    "epoch": 667,
    "batch_idx": 64,
    "loss": 0.44365447759628296,
    "learning_rate": 0.0007497732408203716,
    "timestep_mean": 453.9375,
    "epsilon_norm": 7.964358329772949,
    "prediction_norm": 6.522460460662842,
    "timestamp": "2025-07-14T12:33:57.129840"
  },
  {
    "epoch": 667,
    "batch_idx": 96,
    "loss": 0.3866310715675354,
    "learning_rate": 0.0007497732408203716,
    "timestep_mean": 482.8125,
    "epsilon_norm": 8.436976432800293,
    "prediction_norm": 6.915698528289795,
    "timestamp": "2025-07-14T12:33:57.134746"
  },
  {
    "epoch": 667,
    "batch_idx": 128,
    "loss": 0.23062032461166382,
    "learning_rate": 0.0007497732408203716,
    "timestep_mean": 580.0625,
    "epsilon_norm": 7.967150688171387,
    "prediction_norm": 6.359705924987793,
    "timestamp": "2025-07-14T12:33:57.139675"
  },
  {
    "epoch": 667,
    "batch_idx": 160,
    "loss": 0.34885674715042114,
    "learning_rate": 0.0007497732408203716,
    "timestep_mean": 502.84375,
    "epsilon_norm": 8.356653213500977,
    "prediction_norm": 6.994970798492432,
    "timestamp": "2025-07-14T12:33:57.144624"
  },
  {
    "epoch": 667,
    "batch_idx": 192,
    "loss": 0.11176647245883942,
    "learning_rate": 0.0007497732408203716,
    "timestep_mean": 618.375,
    "epsilon_norm": 4.303778171539307,
    "prediction_norm": 4.105515003204346,
    "timestamp": "2025-07-14T12:33:57.149556"
  },
  {
    "epoch": 668,
    "batch_idx": 0,
    "loss": 0.2182348668575287,
    "learning_rate": 0.0007490925526697455,
    "timestep_mean": 470.4375,
    "epsilon_norm": 6.788089752197266,
    "prediction_norm": 5.822624206542969,
    "timestamp": "2025-07-14T12:33:57.154521"
  },
  {
    "epoch": 668,
    "batch_idx": 32,
    "loss": 0.20076023042201996,
    "learning_rate": 0.0007490925526697455,
    "timestep_mean": 486.03125,
    "epsilon_norm": 7.2355122566223145,
    "prediction_norm": 6.247697353363037,
    "timestamp": "2025-07-14T12:33:57.159382"
  },
  {
    "epoch": 668,
    "batch_idx": 64,
    "loss": 0.4213217794895172,
    "learning_rate": 0.0007490925526697455,
    "timestep_mean": 424.15625,
    "epsilon_norm": 8.314133644104004,
    "prediction_norm": 6.378006935119629,
    "timestamp": "2025-07-14T12:33:57.164282"
  },
  {
    "epoch": 668,
    "batch_idx": 96,
    "loss": 0.24336159229278564,
    "learning_rate": 0.0007490925526697455,
    "timestep_mean": 551.96875,
    "epsilon_norm": 9.052107810974121,
    "prediction_norm": 8.531235694885254,
    "timestamp": "2025-07-14T12:33:57.169190"
  },
  {
    "epoch": 668,
    "batch_idx": 128,
    "loss": 0.28968554735183716,
    "learning_rate": 0.0007490925526697455,
    "timestep_mean": 440.9375,
    "epsilon_norm": 8.293524742126465,
    "prediction_norm": 7.4292988777160645,
    "timestamp": "2025-07-14T12:33:57.174127"
  },
  {
    "epoch": 668,
    "batch_idx": 160,
    "loss": 0.23131859302520752,
    "learning_rate": 0.0007490925526697455,
    "timestep_mean": 548.09375,
    "epsilon_norm": 7.6977338790893555,
    "prediction_norm": 7.051453590393066,
    "timestamp": "2025-07-14T12:33:57.179022"
  },
  {
    "epoch": 668,
    "batch_idx": 192,
    "loss": 0.4987574517726898,
    "learning_rate": 0.0007490925526697455,
    "timestep_mean": 496.0,
    "epsilon_norm": 3.9720232486724854,
    "prediction_norm": 2.4566187858581543,
    "timestamp": "2025-07-14T12:33:57.184105"
  },
  {
    "epoch": 669,
    "batch_idx": 0,
    "loss": 0.3873438835144043,
    "learning_rate": 0.0007484112499080074,
    "timestep_mean": 482.4375,
    "epsilon_norm": 7.64494514465332,
    "prediction_norm": 5.546629428863525,
    "timestamp": "2025-07-14T12:33:57.189194"
  },
  {
    "epoch": 669,
    "batch_idx": 32,
    "loss": 0.34811127185821533,
    "learning_rate": 0.0007484112499080074,
    "timestep_mean": 458.125,
    "epsilon_norm": 8.15125846862793,
    "prediction_norm": 6.224278926849365,
    "timestamp": "2025-07-14T12:33:57.194396"
  },
  {
    "epoch": 669,
    "batch_idx": 64,
    "loss": 0.2188446819782257,
    "learning_rate": 0.0007484112499080074,
    "timestep_mean": 495.875,
    "epsilon_norm": 7.773622989654541,
    "prediction_norm": 6.277700901031494,
    "timestamp": "2025-07-14T12:33:57.199554"
  },
  {
    "epoch": 669,
    "batch_idx": 96,
    "loss": 0.306896448135376,
    "learning_rate": 0.0007484112499080074,
    "timestep_mean": 487.84375,
    "epsilon_norm": 7.634940147399902,
    "prediction_norm": 6.7440690994262695,
    "timestamp": "2025-07-14T12:33:57.204682"
  },
  {
    "epoch": 669,
    "batch_idx": 128,
    "loss": 0.3604865074157715,
    "learning_rate": 0.0007484112499080074,
    "timestep_mean": 425.84375,
    "epsilon_norm": 8.05133056640625,
    "prediction_norm": 6.354687213897705,
    "timestamp": "2025-07-14T12:33:57.209938"
  },
  {
    "epoch": 669,
    "batch_idx": 160,
    "loss": 0.12352659553289413,
    "learning_rate": 0.0007484112499080074,
    "timestep_mean": 612.375,
    "epsilon_norm": 7.1819963455200195,
    "prediction_norm": 6.975585460662842,
    "timestamp": "2025-07-14T12:33:57.215129"
  },
  {
    "epoch": 669,
    "batch_idx": 192,
    "loss": 0.11248428374528885,
    "learning_rate": 0.0007484112499080074,
    "timestep_mean": 669.5,
    "epsilon_norm": 3.3804118633270264,
    "prediction_norm": 3.014940023422241,
    "timestamp": "2025-07-14T12:33:57.220176"
  },
  {
    "epoch": 670,
    "batch_idx": 0,
    "loss": 0.368733286857605,
    "learning_rate": 0.0007477293342162039,
    "timestep_mean": 500.625,
    "epsilon_norm": 7.732820987701416,
    "prediction_norm": 5.937407970428467,
    "timestamp": "2025-07-14T12:33:57.225466"
  },
  {
    "epoch": 670,
    "batch_idx": 32,
    "loss": 0.18812979757785797,
    "learning_rate": 0.0007477293342162039,
    "timestep_mean": 534.625,
    "epsilon_norm": 6.71026086807251,
    "prediction_norm": 6.122768878936768,
    "timestamp": "2025-07-14T12:33:57.230653"
  },
  {
    "epoch": 670,
    "batch_idx": 64,
    "loss": 0.5503231287002563,
    "learning_rate": 0.0007477293342162039,
    "timestep_mean": 391.78125,
    "epsilon_norm": 8.614516258239746,
    "prediction_norm": 6.537160873413086,
    "timestamp": "2025-07-14T12:33:57.235852"
  },
  {
    "epoch": 670,
    "batch_idx": 96,
    "loss": 0.4746706187725067,
    "learning_rate": 0.0007477293342162039,
    "timestep_mean": 448.96875,
    "epsilon_norm": 8.99733829498291,
    "prediction_norm": 6.631110668182373,
    "timestamp": "2025-07-14T12:33:57.241132"
  },
  {
    "epoch": 670,
    "batch_idx": 128,
    "loss": 0.45491713285446167,
    "learning_rate": 0.0007477293342162039,
    "timestep_mean": 484.15625,
    "epsilon_norm": 8.732019424438477,
    "prediction_norm": 6.874205112457275,
    "timestamp": "2025-07-14T12:33:57.246437"
  },
  {
    "epoch": 670,
    "batch_idx": 160,
    "loss": 0.3919975459575653,
    "learning_rate": 0.0007477293342162039,
    "timestep_mean": 451.96875,
    "epsilon_norm": 7.305694103240967,
    "prediction_norm": 4.767740726470947,
    "timestamp": "2025-07-14T12:33:57.251552"
  },
  {
    "epoch": 670,
    "batch_idx": 192,
    "loss": 0.8354693651199341,
    "learning_rate": 0.0007477293342162039,
    "timestep_mean": 360.75,
    "epsilon_norm": 5.569028377532959,
    "prediction_norm": 3.5646095275878906,
    "timestamp": "2025-07-14T12:33:57.256733"
  },
  {
    "epoch": 671,
    "batch_idx": 0,
    "loss": 0.2116730809211731,
    "learning_rate": 0.0007470468072768943,
    "timestep_mean": 517.53125,
    "epsilon_norm": 6.8704752922058105,
    "prediction_norm": 6.419886589050293,
    "timestamp": "2025-07-14T12:33:57.261939"
  },
  {
    "epoch": 671,
    "batch_idx": 32,
    "loss": 0.32764214277267456,
    "learning_rate": 0.0007470468072768943,
    "timestep_mean": 504.21875,
    "epsilon_norm": 8.511658668518066,
    "prediction_norm": 6.864818572998047,
    "timestamp": "2025-07-14T12:33:57.267031"
  },
  {
    "epoch": 671,
    "batch_idx": 64,
    "loss": 0.17641335725784302,
    "learning_rate": 0.0007470468072768943,
    "timestep_mean": 526.03125,
    "epsilon_norm": 8.401901245117188,
    "prediction_norm": 7.486472129821777,
    "timestamp": "2025-07-14T12:33:57.272197"
  },
  {
    "epoch": 671,
    "batch_idx": 96,
    "loss": 0.147468701004982,
    "learning_rate": 0.0007470468072768943,
    "timestep_mean": 527.25,
    "epsilon_norm": 7.935412406921387,
    "prediction_norm": 7.245453357696533,
    "timestamp": "2025-07-14T12:33:57.277052"
  },
  {
    "epoch": 671,
    "batch_idx": 128,
    "loss": 0.39072883129119873,
    "learning_rate": 0.0007470468072768943,
    "timestep_mean": 490.8125,
    "epsilon_norm": 8.011479377746582,
    "prediction_norm": 6.604682445526123,
    "timestamp": "2025-07-14T12:33:57.282281"
  },
  {
    "epoch": 671,
    "batch_idx": 160,
    "loss": 0.23159627616405487,
    "learning_rate": 0.0007470468072768943,
    "timestep_mean": 518.84375,
    "epsilon_norm": 8.646330833435059,
    "prediction_norm": 7.494166851043701,
    "timestamp": "2025-07-14T12:33:57.287410"
  },
  {
    "epoch": 671,
    "batch_idx": 192,
    "loss": 0.23400627076625824,
    "learning_rate": 0.0007470468072768943,
    "timestep_mean": 423.375,
    "epsilon_norm": 2.8049144744873047,
    "prediction_norm": 2.188072919845581,
    "timestamp": "2025-07-14T12:33:57.292554"
  },
  {
    "epoch": 672,
    "batch_idx": 0,
    "loss": 0.23882794380187988,
    "learning_rate": 0.0007463636707741459,
    "timestep_mean": 549.59375,
    "epsilon_norm": 8.162364959716797,
    "prediction_norm": 7.182445526123047,
    "timestamp": "2025-07-14T12:33:57.297786"
  },
  {
    "epoch": 672,
    "batch_idx": 32,
    "loss": 0.3904862403869629,
    "learning_rate": 0.0007463636707741459,
    "timestep_mean": 458.40625,
    "epsilon_norm": 7.408009052276611,
    "prediction_norm": 5.8423171043396,
    "timestamp": "2025-07-14T12:33:57.308074"
  },
  {
    "epoch": 672,
    "batch_idx": 64,
    "loss": 0.20293614268302917,
    "learning_rate": 0.0007463636707741459,
    "timestep_mean": 497.4375,
    "epsilon_norm": 7.83595609664917,
    "prediction_norm": 6.850552082061768,
    "timestamp": "2025-07-14T12:33:57.313450"
  },
  {
    "epoch": 672,
    "batch_idx": 96,
    "loss": 0.33829230070114136,
    "learning_rate": 0.0007463636707741459,
    "timestep_mean": 431.96875,
    "epsilon_norm": 8.214964866638184,
    "prediction_norm": 7.396937847137451,
    "timestamp": "2025-07-14T12:33:57.318370"
  },
  {
    "epoch": 672,
    "batch_idx": 128,
    "loss": 0.4636611342430115,
    "learning_rate": 0.0007463636707741459,
    "timestep_mean": 420.3125,
    "epsilon_norm": 8.386100769042969,
    "prediction_norm": 6.384766101837158,
    "timestamp": "2025-07-14T12:33:57.323325"
  },
  {
    "epoch": 672,
    "batch_idx": 160,
    "loss": 0.2682027220726013,
    "learning_rate": 0.0007463636707741459,
    "timestep_mean": 487.84375,
    "epsilon_norm": 7.446418285369873,
    "prediction_norm": 6.054783821105957,
    "timestamp": "2025-07-14T12:33:57.328190"
  },
  {
    "epoch": 672,
    "batch_idx": 192,
    "loss": 0.22598201036453247,
    "learning_rate": 0.0007463636707741459,
    "timestep_mean": 548.125,
    "epsilon_norm": 4.423681735992432,
    "prediction_norm": 4.780529022216797,
    "timestamp": "2025-07-14T12:33:57.333210"
  },
  {
    "epoch": 673,
    "batch_idx": 0,
    "loss": 0.21505215764045715,
    "learning_rate": 0.0007456799263935301,
    "timestep_mean": 564.3125,
    "epsilon_norm": 7.426815509796143,
    "prediction_norm": 6.26831579208374,
    "timestamp": "2025-07-14T12:33:57.338191"
  },
  {
    "epoch": 673,
    "batch_idx": 32,
    "loss": 0.32033371925354004,
    "learning_rate": 0.0007456799263935301,
    "timestep_mean": 528.25,
    "epsilon_norm": 8.821684837341309,
    "prediction_norm": 6.862504959106445,
    "timestamp": "2025-07-14T12:33:57.343116"
  },
  {
    "epoch": 673,
    "batch_idx": 64,
    "loss": 0.3246452510356903,
    "learning_rate": 0.0007456799263935301,
    "timestep_mean": 530.625,
    "epsilon_norm": 8.503987312316895,
    "prediction_norm": 8.017045974731445,
    "timestamp": "2025-07-14T12:33:57.347972"
  },
  {
    "epoch": 673,
    "batch_idx": 96,
    "loss": 0.5362831354141235,
    "learning_rate": 0.0007456799263935301,
    "timestep_mean": 381.8125,
    "epsilon_norm": 7.806972026824951,
    "prediction_norm": 5.712399482727051,
    "timestamp": "2025-07-14T12:33:57.352929"
  },
  {
    "epoch": 673,
    "batch_idx": 128,
    "loss": 0.4442656934261322,
    "learning_rate": 0.0007456799263935301,
    "timestep_mean": 428.25,
    "epsilon_norm": 8.162309646606445,
    "prediction_norm": 6.45346736907959,
    "timestamp": "2025-07-14T12:33:57.357893"
  },
  {
    "epoch": 673,
    "batch_idx": 160,
    "loss": 0.29847362637519836,
    "learning_rate": 0.0007456799263935301,
    "timestep_mean": 542.4375,
    "epsilon_norm": 8.906928062438965,
    "prediction_norm": 7.988521575927734,
    "timestamp": "2025-07-14T12:33:57.362933"
  },
  {
    "epoch": 673,
    "batch_idx": 192,
    "loss": 0.0866868793964386,
    "learning_rate": 0.0007456799263935301,
    "timestep_mean": 603.375,
    "epsilon_norm": 3.7985923290252686,
    "prediction_norm": 3.6248745918273926,
    "timestamp": "2025-07-14T12:33:57.367861"
  },
  {
    "epoch": 674,
    "batch_idx": 0,
    "loss": 0.36213165521621704,
    "learning_rate": 0.0007449955758221184,
    "timestep_mean": 458.28125,
    "epsilon_norm": 8.187104225158691,
    "prediction_norm": 7.229116439819336,
    "timestamp": "2025-07-14T12:33:57.372841"
  },
  {
    "epoch": 674,
    "batch_idx": 32,
    "loss": 0.12361396849155426,
    "learning_rate": 0.0007449955758221184,
    "timestep_mean": 517.375,
    "epsilon_norm": 5.946052074432373,
    "prediction_norm": 5.6000189781188965,
    "timestamp": "2025-07-14T12:33:57.377852"
  },
  {
    "epoch": 674,
    "batch_idx": 64,
    "loss": 0.28431639075279236,
    "learning_rate": 0.0007449955758221184,
    "timestep_mean": 515.40625,
    "epsilon_norm": 8.355910301208496,
    "prediction_norm": 7.685142993927002,
    "timestamp": "2025-07-14T12:33:57.382831"
  },
  {
    "epoch": 674,
    "batch_idx": 96,
    "loss": 0.22202074527740479,
    "learning_rate": 0.0007449955758221184,
    "timestep_mean": 473.03125,
    "epsilon_norm": 9.00415325164795,
    "prediction_norm": 8.361567497253418,
    "timestamp": "2025-07-14T12:33:57.387902"
  },
  {
    "epoch": 674,
    "batch_idx": 128,
    "loss": 0.25253400206565857,
    "learning_rate": 0.0007449955758221184,
    "timestep_mean": 466.5625,
    "epsilon_norm": 7.416473388671875,
    "prediction_norm": 6.122687816619873,
    "timestamp": "2025-07-14T12:33:57.392921"
  },
  {
    "epoch": 674,
    "batch_idx": 160,
    "loss": 0.3652908205986023,
    "learning_rate": 0.0007449955758221184,
    "timestep_mean": 486.375,
    "epsilon_norm": 7.248540878295898,
    "prediction_norm": 5.779375076293945,
    "timestamp": "2025-07-14T12:33:57.397877"
  },
  {
    "epoch": 674,
    "batch_idx": 192,
    "loss": 0.10476137697696686,
    "learning_rate": 0.0007449955758221184,
    "timestep_mean": 784.5,
    "epsilon_norm": 3.6078312397003174,
    "prediction_norm": 3.085580587387085,
    "timestamp": "2025-07-14T12:33:57.402799"
  },
  {
    "epoch": 675,
    "batch_idx": 0,
    "loss": 0.2328546643257141,
    "learning_rate": 0.0007443106207484776,
    "timestep_mean": 520.0,
    "epsilon_norm": 7.859004020690918,
    "prediction_norm": 6.388484001159668,
    "timestamp": "2025-07-14T12:33:57.407609"
  },
  {
    "epoch": 675,
    "batch_idx": 32,
    "loss": 0.5475090742111206,
    "learning_rate": 0.0007443106207484776,
    "timestep_mean": 388.6875,
    "epsilon_norm": 9.738485336303711,
    "prediction_norm": 7.300070285797119,
    "timestamp": "2025-07-14T12:33:57.412374"
  },
  {
    "epoch": 675,
    "batch_idx": 64,
    "loss": 0.20555399358272552,
    "learning_rate": 0.0007443106207484776,
    "timestep_mean": 560.625,
    "epsilon_norm": 8.372879981994629,
    "prediction_norm": 7.1296610832214355,
    "timestamp": "2025-07-14T12:33:57.417697"
  },
  {
    "epoch": 675,
    "batch_idx": 96,
    "loss": 0.350566029548645,
    "learning_rate": 0.0007443106207484776,
    "timestep_mean": 524.46875,
    "epsilon_norm": 9.191536903381348,
    "prediction_norm": 8.128047943115234,
    "timestamp": "2025-07-14T12:33:57.422733"
  },
  {
    "epoch": 675,
    "batch_idx": 128,
    "loss": 0.24235907196998596,
    "learning_rate": 0.0007443106207484776,
    "timestep_mean": 571.78125,
    "epsilon_norm": 7.655516147613525,
    "prediction_norm": 6.530214309692383,
    "timestamp": "2025-07-14T12:33:57.427813"
  },
  {
    "epoch": 675,
    "batch_idx": 160,
    "loss": 0.279499351978302,
    "learning_rate": 0.0007443106207484776,
    "timestep_mean": 465.125,
    "epsilon_norm": 7.697009086608887,
    "prediction_norm": 6.589016914367676,
    "timestamp": "2025-07-14T12:33:57.432874"
  },
  {
    "epoch": 675,
    "batch_idx": 192,
    "loss": 0.2350928783416748,
    "learning_rate": 0.0007443106207484776,
    "timestep_mean": 364.0,
    "epsilon_norm": 4.499508857727051,
    "prediction_norm": 4.597968578338623,
    "timestamp": "2025-07-14T12:33:57.438060"
  },
  {
    "epoch": 676,
    "batch_idx": 0,
    "loss": 0.3065366744995117,
    "learning_rate": 0.0007436250628626662,
    "timestep_mean": 483.21875,
    "epsilon_norm": 7.840745449066162,
    "prediction_norm": 6.51275110244751,
    "timestamp": "2025-07-14T12:33:57.443208"
  },
  {
    "epoch": 676,
    "batch_idx": 32,
    "loss": 0.2410777509212494,
    "learning_rate": 0.0007436250628626662,
    "timestep_mean": 542.34375,
    "epsilon_norm": 8.330218315124512,
    "prediction_norm": 6.995594501495361,
    "timestamp": "2025-07-14T12:33:57.448438"
  },
  {
    "epoch": 676,
    "batch_idx": 64,
    "loss": 0.18633554875850677,
    "learning_rate": 0.0007436250628626662,
    "timestep_mean": 497.34375,
    "epsilon_norm": 6.956140995025635,
    "prediction_norm": 5.475959777832031,
    "timestamp": "2025-07-14T12:33:57.453780"
  },
  {
    "epoch": 676,
    "batch_idx": 96,
    "loss": 0.2706677317619324,
    "learning_rate": 0.0007436250628626662,
    "timestep_mean": 551.96875,
    "epsilon_norm": 8.516294479370117,
    "prediction_norm": 7.656140327453613,
    "timestamp": "2025-07-14T12:33:57.458983"
  },
  {
    "epoch": 676,
    "batch_idx": 128,
    "loss": 0.24842479825019836,
    "learning_rate": 0.0007436250628626662,
    "timestep_mean": 443.6875,
    "epsilon_norm": 7.572749614715576,
    "prediction_norm": 6.147344589233398,
    "timestamp": "2025-07-14T12:33:57.464137"
  },
  {
    "epoch": 676,
    "batch_idx": 160,
    "loss": 0.24661478400230408,
    "learning_rate": 0.0007436250628626662,
    "timestep_mean": 502.03125,
    "epsilon_norm": 8.309821128845215,
    "prediction_norm": 7.307368278503418,
    "timestamp": "2025-07-14T12:33:57.469240"
  },
  {
    "epoch": 676,
    "batch_idx": 192,
    "loss": 0.3069468140602112,
    "learning_rate": 0.0007436250628626662,
    "timestep_mean": 460.75,
    "epsilon_norm": 4.537806034088135,
    "prediction_norm": 3.6070713996887207,
    "timestamp": "2025-07-14T12:33:57.474360"
  },
  {
    "epoch": 677,
    "batch_idx": 0,
    "loss": 0.26142704486846924,
    "learning_rate": 0.0007429389038562304,
    "timestep_mean": 489.78125,
    "epsilon_norm": 7.58730411529541,
    "prediction_norm": 6.483573913574219,
    "timestamp": "2025-07-14T12:33:57.479657"
  },
  {
    "epoch": 677,
    "batch_idx": 32,
    "loss": 0.16146501898765564,
    "learning_rate": 0.0007429389038562304,
    "timestep_mean": 510.25,
    "epsilon_norm": 8.180035591125488,
    "prediction_norm": 7.203035354614258,
    "timestamp": "2025-07-14T12:33:57.484786"
  },
  {
    "epoch": 677,
    "batch_idx": 64,
    "loss": 0.21965111792087555,
    "learning_rate": 0.0007429389038562304,
    "timestep_mean": 485.75,
    "epsilon_norm": 8.472527503967285,
    "prediction_norm": 6.903227806091309,
    "timestamp": "2025-07-14T12:33:57.489891"
  },
  {
    "epoch": 677,
    "batch_idx": 96,
    "loss": 0.2992126941680908,
    "learning_rate": 0.0007429389038562304,
    "timestep_mean": 531.78125,
    "epsilon_norm": 8.02269458770752,
    "prediction_norm": 6.538666248321533,
    "timestamp": "2025-07-14T12:33:57.495120"
  },
  {
    "epoch": 677,
    "batch_idx": 128,
    "loss": 0.366496205329895,
    "learning_rate": 0.0007429389038562304,
    "timestep_mean": 488.75,
    "epsilon_norm": 7.614270210266113,
    "prediction_norm": 6.072378158569336,
    "timestamp": "2025-07-14T12:33:57.500418"
  },
  {
    "epoch": 677,
    "batch_idx": 160,
    "loss": 0.15501709282398224,
    "learning_rate": 0.0007429389038562304,
    "timestep_mean": 561.375,
    "epsilon_norm": 7.895906925201416,
    "prediction_norm": 6.711522579193115,
    "timestamp": "2025-07-14T12:33:57.505544"
  },
  {
    "epoch": 677,
    "batch_idx": 192,
    "loss": 0.48160049319267273,
    "learning_rate": 0.0007429389038562304,
    "timestep_mean": 398.75,
    "epsilon_norm": 3.2856688499450684,
    "prediction_norm": 2.214151620864868,
    "timestamp": "2025-07-14T12:33:57.510704"
  },
  {
    "epoch": 678,
    "batch_idx": 0,
    "loss": 0.3264067769050598,
    "learning_rate": 0.0007422521454221991,
    "timestep_mean": 433.75,
    "epsilon_norm": 7.83232307434082,
    "prediction_norm": 6.231173038482666,
    "timestamp": "2025-07-14T12:33:57.516024"
  },
  {
    "epoch": 678,
    "batch_idx": 32,
    "loss": 0.22622092068195343,
    "learning_rate": 0.0007422521454221991,
    "timestep_mean": 550.28125,
    "epsilon_norm": 7.718677043914795,
    "prediction_norm": 6.7667555809021,
    "timestamp": "2025-07-14T12:33:57.521224"
  },
  {
    "epoch": 678,
    "batch_idx": 64,
    "loss": 0.1868782937526703,
    "learning_rate": 0.0007422521454221991,
    "timestep_mean": 505.15625,
    "epsilon_norm": 9.145925521850586,
    "prediction_norm": 7.84481954574585,
    "timestamp": "2025-07-14T12:33:57.526363"
  },
  {
    "epoch": 678,
    "batch_idx": 96,
    "loss": 0.31041425466537476,
    "learning_rate": 0.0007422521454221991,
    "timestep_mean": 543.8125,
    "epsilon_norm": 8.91334056854248,
    "prediction_norm": 8.426624298095703,
    "timestamp": "2025-07-14T12:33:57.531523"
  },
  {
    "epoch": 678,
    "batch_idx": 128,
    "loss": 0.38001859188079834,
    "learning_rate": 0.0007422521454221991,
    "timestep_mean": 587.96875,
    "epsilon_norm": 8.817169189453125,
    "prediction_norm": 7.577378749847412,
    "timestamp": "2025-07-14T12:33:57.536610"
  },
  {
    "epoch": 678,
    "batch_idx": 160,
    "loss": 0.3217180371284485,
    "learning_rate": 0.0007422521454221991,
    "timestep_mean": 463.53125,
    "epsilon_norm": 8.064714431762695,
    "prediction_norm": 6.711030006408691,
    "timestamp": "2025-07-14T12:33:57.541846"
  },
  {
    "epoch": 678,
    "batch_idx": 192,
    "loss": 0.0489603653550148,
    "learning_rate": 0.0007422521454221991,
    "timestep_mean": 599.625,
    "epsilon_norm": 4.715402126312256,
    "prediction_norm": 4.210780620574951,
    "timestamp": "2025-07-14T12:33:57.547049"
  },
  {
    "epoch": 679,
    "batch_idx": 0,
    "loss": 0.2938203811645508,
    "learning_rate": 0.0007415647892550805,
    "timestep_mean": 443.59375,
    "epsilon_norm": 7.972865581512451,
    "prediction_norm": 6.195690631866455,
    "timestamp": "2025-07-14T12:33:57.552209"
  },
  {
    "epoch": 679,
    "batch_idx": 32,
    "loss": 0.358812153339386,
    "learning_rate": 0.0007415647892550805,
    "timestep_mean": 522.78125,
    "epsilon_norm": 7.600549697875977,
    "prediction_norm": 6.607015609741211,
    "timestamp": "2025-07-14T12:33:57.557475"
  },
  {
    "epoch": 679,
    "batch_idx": 64,
    "loss": 0.3278222382068634,
    "learning_rate": 0.0007415647892550805,
    "timestep_mean": 546.28125,
    "epsilon_norm": 8.453417778015137,
    "prediction_norm": 7.775610446929932,
    "timestamp": "2025-07-14T12:33:57.562677"
  },
  {
    "epoch": 679,
    "batch_idx": 96,
    "loss": 0.3189823627471924,
    "learning_rate": 0.0007415647892550805,
    "timestep_mean": 480.5,
    "epsilon_norm": 7.293793678283691,
    "prediction_norm": 6.080776214599609,
    "timestamp": "2025-07-14T12:33:57.567751"
  },
  {
    "epoch": 679,
    "batch_idx": 128,
    "loss": 0.3452475666999817,
    "learning_rate": 0.0007415647892550805,
    "timestep_mean": 340.40625,
    "epsilon_norm": 6.7561869621276855,
    "prediction_norm": 4.548829078674316,
    "timestamp": "2025-07-14T12:33:57.573009"
  },
  {
    "epoch": 679,
    "batch_idx": 160,
    "loss": 0.17402872443199158,
    "learning_rate": 0.0007415647892550805,
    "timestep_mean": 514.9375,
    "epsilon_norm": 6.657937049865723,
    "prediction_norm": 6.5729265213012695,
    "timestamp": "2025-07-14T12:33:57.578191"
  },
  {
    "epoch": 679,
    "batch_idx": 192,
    "loss": 0.13885284960269928,
    "learning_rate": 0.0007415647892550805,
    "timestep_mean": 468.0,
    "epsilon_norm": 3.958225727081299,
    "prediction_norm": 3.770986795425415,
    "timestamp": "2025-07-14T12:33:57.583408"
  },
  {
    "epoch": 680,
    "batch_idx": 0,
    "loss": 0.46324971318244934,
    "learning_rate": 0.0007408768370508577,
    "timestep_mean": 477.6875,
    "epsilon_norm": 8.257814407348633,
    "prediction_norm": 7.005237579345703,
    "timestamp": "2025-07-14T12:33:57.588636"
  },
  {
    "epoch": 680,
    "batch_idx": 32,
    "loss": 0.2784690260887146,
    "learning_rate": 0.0007408768370508577,
    "timestep_mean": 551.5625,
    "epsilon_norm": 8.47753620147705,
    "prediction_norm": 7.563153266906738,
    "timestamp": "2025-07-14T12:33:57.593777"
  },
  {
    "epoch": 680,
    "batch_idx": 64,
    "loss": 0.16682168841362,
    "learning_rate": 0.0007408768370508577,
    "timestep_mean": 560.46875,
    "epsilon_norm": 8.357035636901855,
    "prediction_norm": 7.728221893310547,
    "timestamp": "2025-07-14T12:33:57.598894"
  },
  {
    "epoch": 680,
    "batch_idx": 96,
    "loss": 0.2256009578704834,
    "learning_rate": 0.0007408768370508577,
    "timestep_mean": 501.0,
    "epsilon_norm": 7.566349029541016,
    "prediction_norm": 5.860668659210205,
    "timestamp": "2025-07-14T12:33:57.604031"
  },
  {
    "epoch": 680,
    "batch_idx": 128,
    "loss": 0.14965900778770447,
    "learning_rate": 0.0007408768370508577,
    "timestep_mean": 595.75,
    "epsilon_norm": 7.460659503936768,
    "prediction_norm": 7.051390171051025,
    "timestamp": "2025-07-14T12:33:57.609183"
  },
  {
    "epoch": 680,
    "batch_idx": 160,
    "loss": 0.1939733624458313,
    "learning_rate": 0.0007408768370508577,
    "timestep_mean": 423.1875,
    "epsilon_norm": 6.561879634857178,
    "prediction_norm": 5.528894901275635,
    "timestamp": "2025-07-14T12:33:57.614320"
  },
  {
    "epoch": 680,
    "batch_idx": 192,
    "loss": 0.3163134753704071,
    "learning_rate": 0.0007408768370508577,
    "timestep_mean": 495.375,
    "epsilon_norm": 3.1948835849761963,
    "prediction_norm": 2.2967259883880615,
    "timestamp": "2025-07-14T12:33:57.619404"
  },
  {
    "epoch": 681,
    "batch_idx": 0,
    "loss": 0.2720831036567688,
    "learning_rate": 0.0007401882905069844,
    "timestep_mean": 435.375,
    "epsilon_norm": 7.438694000244141,
    "prediction_norm": 6.380342960357666,
    "timestamp": "2025-07-14T12:33:57.624693"
  },
  {
    "epoch": 681,
    "batch_idx": 32,
    "loss": 0.19560253620147705,
    "learning_rate": 0.0007401882905069844,
    "timestep_mean": 420.03125,
    "epsilon_norm": 8.115968704223633,
    "prediction_norm": 7.173490524291992,
    "timestamp": "2025-07-14T12:33:57.629968"
  },
  {
    "epoch": 681,
    "batch_idx": 64,
    "loss": 0.2604696452617645,
    "learning_rate": 0.0007401882905069844,
    "timestep_mean": 501.3125,
    "epsilon_norm": 8.376150131225586,
    "prediction_norm": 7.175858974456787,
    "timestamp": "2025-07-14T12:33:57.635112"
  },
  {
    "epoch": 681,
    "batch_idx": 96,
    "loss": 0.1988157331943512,
    "learning_rate": 0.0007401882905069844,
    "timestep_mean": 519.40625,
    "epsilon_norm": 6.674993991851807,
    "prediction_norm": 6.1745829582214355,
    "timestamp": "2025-07-14T12:33:57.640311"
  },
  {
    "epoch": 681,
    "batch_idx": 128,
    "loss": 0.2458990216255188,
    "learning_rate": 0.0007401882905069844,
    "timestep_mean": 486.0625,
    "epsilon_norm": 8.352887153625488,
    "prediction_norm": 7.098373889923096,
    "timestamp": "2025-07-14T12:33:57.645454"
  },
  {
    "epoch": 681,
    "batch_idx": 160,
    "loss": 0.22413799166679382,
    "learning_rate": 0.0007401882905069844,
    "timestep_mean": 511.625,
    "epsilon_norm": 7.840031623840332,
    "prediction_norm": 6.152707576751709,
    "timestamp": "2025-07-14T12:33:57.650630"
  },
  {
    "epoch": 681,
    "batch_idx": 192,
    "loss": 0.23692366480827332,
    "learning_rate": 0.0007401882905069844,
    "timestep_mean": 567.625,
    "epsilon_norm": 4.039456367492676,
    "prediction_norm": 3.425895929336548,
    "timestamp": "2025-07-14T12:33:57.655767"
  },
  {
    "epoch": 682,
    "batch_idx": 0,
    "loss": 0.27060365676879883,
    "learning_rate": 0.0007394991513223806,
    "timestep_mean": 526.28125,
    "epsilon_norm": 7.6299920082092285,
    "prediction_norm": 6.820623397827148,
    "timestamp": "2025-07-14T12:33:57.660897"
  },
  {
    "epoch": 682,
    "batch_idx": 32,
    "loss": 0.31901097297668457,
    "learning_rate": 0.0007394991513223806,
    "timestep_mean": 400.75,
    "epsilon_norm": 7.253231048583984,
    "prediction_norm": 6.385590553283691,
    "timestamp": "2025-07-14T12:33:57.666031"
  },
  {
    "epoch": 682,
    "batch_idx": 64,
    "loss": 0.274558961391449,
    "learning_rate": 0.0007394991513223806,
    "timestep_mean": 443.0625,
    "epsilon_norm": 7.612577438354492,
    "prediction_norm": 6.137753963470459,
    "timestamp": "2025-07-14T12:33:57.671035"
  },
  {
    "epoch": 682,
    "batch_idx": 96,
    "loss": 0.25025057792663574,
    "learning_rate": 0.0007394991513223806,
    "timestep_mean": 504.84375,
    "epsilon_norm": 7.687594413757324,
    "prediction_norm": 6.398525714874268,
    "timestamp": "2025-07-14T12:33:57.676213"
  },
  {
    "epoch": 682,
    "batch_idx": 128,
    "loss": 0.17370420694351196,
    "learning_rate": 0.0007394991513223806,
    "timestep_mean": 540.78125,
    "epsilon_norm": 7.1544718742370605,
    "prediction_norm": 6.488955497741699,
    "timestamp": "2025-07-14T12:33:57.681368"
  },
  {
    "epoch": 682,
    "batch_idx": 160,
    "loss": 0.1769910305738449,
    "learning_rate": 0.0007394991513223806,
    "timestep_mean": 576.9375,
    "epsilon_norm": 8.573237419128418,
    "prediction_norm": 7.466385841369629,
    "timestamp": "2025-07-14T12:33:57.686473"
  },
  {
    "epoch": 682,
    "batch_idx": 192,
    "loss": 0.2310233861207962,
    "learning_rate": 0.0007394991513223806,
    "timestep_mean": 496.625,
    "epsilon_norm": 4.175018787384033,
    "prediction_norm": 4.0806355476379395,
    "timestamp": "2025-07-14T12:33:57.691398"
  },
  {
    "epoch": 683,
    "batch_idx": 0,
    "loss": 0.26341909170150757,
    "learning_rate": 0.0007388094211974289,
    "timestep_mean": 478.65625,
    "epsilon_norm": 8.548891067504883,
    "prediction_norm": 7.061498165130615,
    "timestamp": "2025-07-14T12:33:57.696353"
  },
  {
    "epoch": 683,
    "batch_idx": 32,
    "loss": 0.1992138922214508,
    "learning_rate": 0.0007388094211974289,
    "timestep_mean": 510.40625,
    "epsilon_norm": 7.903172969818115,
    "prediction_norm": 6.98750114440918,
    "timestamp": "2025-07-14T12:33:57.701153"
  },
  {
    "epoch": 683,
    "batch_idx": 64,
    "loss": 0.3516595959663391,
    "learning_rate": 0.0007388094211974289,
    "timestep_mean": 453.3125,
    "epsilon_norm": 8.919113159179688,
    "prediction_norm": 7.0502753257751465,
    "timestamp": "2025-07-14T12:33:57.706089"
  },
  {
    "epoch": 683,
    "batch_idx": 96,
    "loss": 0.3604544997215271,
    "learning_rate": 0.0007388094211974289,
    "timestep_mean": 515.125,
    "epsilon_norm": 8.722258567810059,
    "prediction_norm": 7.099595069885254,
    "timestamp": "2025-07-14T12:33:57.711234"
  },
  {
    "epoch": 683,
    "batch_idx": 128,
    "loss": 0.2734891176223755,
    "learning_rate": 0.0007388094211974289,
    "timestep_mean": 511.9375,
    "epsilon_norm": 8.91490364074707,
    "prediction_norm": 7.512182712554932,
    "timestamp": "2025-07-14T12:33:57.716331"
  },
  {
    "epoch": 683,
    "batch_idx": 160,
    "loss": 0.2940511703491211,
    "learning_rate": 0.0007388094211974289,
    "timestep_mean": 485.96875,
    "epsilon_norm": 7.780551433563232,
    "prediction_norm": 6.582086086273193,
    "timestamp": "2025-07-14T12:33:57.721613"
  },
  {
    "epoch": 683,
    "batch_idx": 192,
    "loss": 0.1061113029718399,
    "learning_rate": 0.0007388094211974289,
    "timestep_mean": 444.25,
    "epsilon_norm": 3.1165807247161865,
    "prediction_norm": 2.9553911685943604,
    "timestamp": "2025-07-14T12:33:57.726729"
  },
  {
    "epoch": 684,
    "batch_idx": 0,
    "loss": 0.3010552227497101,
    "learning_rate": 0.0007381191018339697,
    "timestep_mean": 455.375,
    "epsilon_norm": 7.747618675231934,
    "prediction_norm": 6.137242317199707,
    "timestamp": "2025-07-14T12:33:57.731877"
  },
  {
    "epoch": 684,
    "batch_idx": 32,
    "loss": 0.23038095235824585,
    "learning_rate": 0.0007381191018339697,
    "timestep_mean": 551.90625,
    "epsilon_norm": 7.494445323944092,
    "prediction_norm": 7.219111442565918,
    "timestamp": "2025-07-14T12:33:57.737045"
  },
  {
    "epoch": 684,
    "batch_idx": 64,
    "loss": 0.19951188564300537,
    "learning_rate": 0.0007381191018339697,
    "timestep_mean": 545.8125,
    "epsilon_norm": 7.37641716003418,
    "prediction_norm": 7.245425701141357,
    "timestamp": "2025-07-14T12:33:57.742402"
  },
  {
    "epoch": 684,
    "batch_idx": 96,
    "loss": 0.3831441402435303,
    "learning_rate": 0.0007381191018339697,
    "timestep_mean": 422.75,
    "epsilon_norm": 8.15511417388916,
    "prediction_norm": 6.000944137573242,
    "timestamp": "2025-07-14T12:33:57.747569"
  },
  {
    "epoch": 684,
    "batch_idx": 128,
    "loss": 0.2730725407600403,
    "learning_rate": 0.0007381191018339697,
    "timestep_mean": 462.15625,
    "epsilon_norm": 7.353637218475342,
    "prediction_norm": 5.903276443481445,
    "timestamp": "2025-07-14T12:33:57.752756"
  },
  {
    "epoch": 684,
    "batch_idx": 160,
    "loss": 0.5523374676704407,
    "learning_rate": 0.0007381191018339697,
    "timestep_mean": 458.03125,
    "epsilon_norm": 9.243672370910645,
    "prediction_norm": 6.531119346618652,
    "timestamp": "2025-07-14T12:33:57.757982"
  },
  {
    "epoch": 684,
    "batch_idx": 192,
    "loss": 0.25899916887283325,
    "learning_rate": 0.0007381191018339697,
    "timestep_mean": 630.875,
    "epsilon_norm": 4.3232421875,
    "prediction_norm": 3.8125998973846436,
    "timestamp": "2025-07-14T12:33:57.763205"
  },
  {
    "epoch": 685,
    "batch_idx": 0,
    "loss": 0.31042367219924927,
    "learning_rate": 0.0007374281949352974,
    "timestep_mean": 411.5625,
    "epsilon_norm": 8.391976356506348,
    "prediction_norm": 7.175376892089844,
    "timestamp": "2025-07-14T12:33:57.768437"
  },
  {
    "epoch": 685,
    "batch_idx": 32,
    "loss": 0.29141002893447876,
    "learning_rate": 0.0007374281949352974,
    "timestep_mean": 485.6875,
    "epsilon_norm": 6.994468688964844,
    "prediction_norm": 5.590514659881592,
    "timestamp": "2025-07-14T12:33:57.773594"
  },
  {
    "epoch": 685,
    "batch_idx": 64,
    "loss": 0.21170783042907715,
    "learning_rate": 0.0007374281949352974,
    "timestep_mean": 486.1875,
    "epsilon_norm": 7.098293304443359,
    "prediction_norm": 5.737707138061523,
    "timestamp": "2025-07-14T12:33:57.778783"
  },
  {
    "epoch": 685,
    "batch_idx": 96,
    "loss": 0.19498544931411743,
    "learning_rate": 0.0007374281949352974,
    "timestep_mean": 484.8125,
    "epsilon_norm": 7.401895046234131,
    "prediction_norm": 6.720247268676758,
    "timestamp": "2025-07-14T12:33:57.783949"
  },
  {
    "epoch": 685,
    "batch_idx": 128,
    "loss": 0.11844281107187271,
    "learning_rate": 0.0007374281949352974,
    "timestep_mean": 631.96875,
    "epsilon_norm": 7.70032262802124,
    "prediction_norm": 7.13504695892334,
    "timestamp": "2025-07-14T12:33:57.789017"
  },
  {
    "epoch": 685,
    "batch_idx": 160,
    "loss": 0.30599308013916016,
    "learning_rate": 0.0007374281949352974,
    "timestep_mean": 540.34375,
    "epsilon_norm": 8.112893104553223,
    "prediction_norm": 6.84115743637085,
    "timestamp": "2025-07-14T12:33:57.794193"
  },
  {
    "epoch": 685,
    "batch_idx": 192,
    "loss": 0.3235268294811249,
    "learning_rate": 0.0007374281949352974,
    "timestep_mean": 422.75,
    "epsilon_norm": 3.9313571453094482,
    "prediction_norm": 3.427877902984619,
    "timestamp": "2025-07-14T12:33:57.799281"
  },
  {
    "epoch": 686,
    "batch_idx": 0,
    "loss": 0.462001770734787,
    "learning_rate": 0.0007367367022061562,
    "timestep_mean": 503.03125,
    "epsilon_norm": 8.723786354064941,
    "prediction_norm": 7.223417282104492,
    "timestamp": "2025-07-14T12:33:57.804588"
  },
  {
    "epoch": 686,
    "batch_idx": 32,
    "loss": 0.3620974123477936,
    "learning_rate": 0.0007367367022061562,
    "timestep_mean": 519.84375,
    "epsilon_norm": 8.195452690124512,
    "prediction_norm": 7.027097225189209,
    "timestamp": "2025-07-14T12:33:57.809724"
  },
  {
    "epoch": 686,
    "batch_idx": 64,
    "loss": 0.1438262164592743,
    "learning_rate": 0.0007367367022061562,
    "timestep_mean": 573.5625,
    "epsilon_norm": 8.243537902832031,
    "prediction_norm": 7.857112407684326,
    "timestamp": "2025-07-14T12:33:57.814800"
  },
  {
    "epoch": 686,
    "batch_idx": 96,
    "loss": 0.19794940948486328,
    "learning_rate": 0.0007367367022061562,
    "timestep_mean": 469.375,
    "epsilon_norm": 9.256714820861816,
    "prediction_norm": 7.527924537658691,
    "timestamp": "2025-07-14T12:33:57.819925"
  },
  {
    "epoch": 686,
    "batch_idx": 128,
    "loss": 0.20789998769760132,
    "learning_rate": 0.0007367367022061562,
    "timestep_mean": 516.65625,
    "epsilon_norm": 9.124373435974121,
    "prediction_norm": 7.945028305053711,
    "timestamp": "2025-07-14T12:33:57.825169"
  },
  {
    "epoch": 686,
    "batch_idx": 160,
    "loss": 0.38128507137298584,
    "learning_rate": 0.0007367367022061562,
    "timestep_mean": 486.5625,
    "epsilon_norm": 7.245733261108398,
    "prediction_norm": 5.489073753356934,
    "timestamp": "2025-07-14T12:33:57.830301"
  },
  {
    "epoch": 686,
    "batch_idx": 192,
    "loss": 0.4221653640270233,
    "learning_rate": 0.0007367367022061562,
    "timestep_mean": 472.0,
    "epsilon_norm": 3.760784864425659,
    "prediction_norm": 2.7183520793914795,
    "timestamp": "2025-07-14T12:33:57.835407"
  },
  {
    "epoch": 687,
    "batch_idx": 0,
    "loss": 0.3177652359008789,
    "learning_rate": 0.0007360446253527356,
    "timestep_mean": 486.28125,
    "epsilon_norm": 7.634947776794434,
    "prediction_norm": 6.3392415046691895,
    "timestamp": "2025-07-14T12:33:57.840670"
  },
  {
    "epoch": 687,
    "batch_idx": 32,
    "loss": 0.30459335446357727,
    "learning_rate": 0.0007360446253527356,
    "timestep_mean": 464.09375,
    "epsilon_norm": 8.429039001464844,
    "prediction_norm": 6.936007976531982,
    "timestamp": "2025-07-14T12:33:57.845862"
  },
  {
    "epoch": 687,
    "batch_idx": 64,
    "loss": 0.357917845249176,
    "learning_rate": 0.0007360446253527356,
    "timestep_mean": 439.53125,
    "epsilon_norm": 8.550825119018555,
    "prediction_norm": 6.625646114349365,
    "timestamp": "2025-07-14T12:33:57.850981"
  },
  {
    "epoch": 687,
    "batch_idx": 96,
    "loss": 0.26639044284820557,
    "learning_rate": 0.0007360446253527356,
    "timestep_mean": 529.65625,
    "epsilon_norm": 8.51323413848877,
    "prediction_norm": 7.573938369750977,
    "timestamp": "2025-07-14T12:33:57.856144"
  },
  {
    "epoch": 687,
    "batch_idx": 128,
    "loss": 0.28034257888793945,
    "learning_rate": 0.0007360446253527356,
    "timestep_mean": 543.34375,
    "epsilon_norm": 8.577093124389648,
    "prediction_norm": 6.960946083068848,
    "timestamp": "2025-07-14T12:33:57.861258"
  },
  {
    "epoch": 687,
    "batch_idx": 160,
    "loss": 0.2101946771144867,
    "learning_rate": 0.0007360446253527356,
    "timestep_mean": 483.03125,
    "epsilon_norm": 7.888156890869141,
    "prediction_norm": 6.897160053253174,
    "timestamp": "2025-07-14T12:33:57.866108"
  },
  {
    "epoch": 687,
    "batch_idx": 192,
    "loss": 0.13975471258163452,
    "learning_rate": 0.0007360446253527356,
    "timestep_mean": 270.625,
    "epsilon_norm": 3.4505975246429443,
    "prediction_norm": 2.9538497924804688,
    "timestamp": "2025-07-14T12:33:57.871085"
  },
  {
    "epoch": 688,
    "batch_idx": 0,
    "loss": 0.32241880893707275,
    "learning_rate": 0.0007353519660826663,
    "timestep_mean": 451.03125,
    "epsilon_norm": 7.98237943649292,
    "prediction_norm": 6.839458465576172,
    "timestamp": "2025-07-14T12:33:57.875921"
  },
  {
    "epoch": 688,
    "batch_idx": 32,
    "loss": 0.23528200387954712,
    "learning_rate": 0.0007353519660826663,
    "timestep_mean": 532.3125,
    "epsilon_norm": 7.566122531890869,
    "prediction_norm": 6.867816925048828,
    "timestamp": "2025-07-14T12:33:57.880841"
  },
  {
    "epoch": 688,
    "batch_idx": 64,
    "loss": 0.2674564719200134,
    "learning_rate": 0.0007353519660826663,
    "timestep_mean": 452.125,
    "epsilon_norm": 8.35758113861084,
    "prediction_norm": 7.9077677726745605,
    "timestamp": "2025-07-14T12:33:57.885741"
  },
  {
    "epoch": 688,
    "batch_idx": 96,
    "loss": 0.37110215425491333,
    "learning_rate": 0.0007353519660826663,
    "timestep_mean": 394.25,
    "epsilon_norm": 7.7580246925354,
    "prediction_norm": 6.376158237457275,
    "timestamp": "2025-07-14T12:33:57.890643"
  },
  {
    "epoch": 688,
    "batch_idx": 128,
    "loss": 0.19417645037174225,
    "learning_rate": 0.0007353519660826663,
    "timestep_mean": 545.34375,
    "epsilon_norm": 6.72613525390625,
    "prediction_norm": 5.869868755340576,
    "timestamp": "2025-07-14T12:33:57.901005"
  },
  {
    "epoch": 688,
    "batch_idx": 160,
    "loss": 0.21538802981376648,
    "learning_rate": 0.0007353519660826663,
    "timestep_mean": 469.625,
    "epsilon_norm": 7.301545143127441,
    "prediction_norm": 6.98159122467041,
    "timestamp": "2025-07-14T12:33:57.905718"
  },
  {
    "epoch": 688,
    "batch_idx": 192,
    "loss": 0.3653015196323395,
    "learning_rate": 0.0007353519660826663,
    "timestep_mean": 504.625,
    "epsilon_norm": 4.490545272827148,
    "prediction_norm": 3.5200839042663574,
    "timestamp": "2025-07-14T12:33:57.910423"
  },
  {
    "epoch": 689,
    "batch_idx": 0,
    "loss": 0.2513998746871948,
    "learning_rate": 0.0007346587261050165,
    "timestep_mean": 537.03125,
    "epsilon_norm": 10.060656547546387,
    "prediction_norm": 8.654939651489258,
    "timestamp": "2025-07-14T12:33:57.915102"
  },
  {
    "epoch": 689,
    "batch_idx": 32,
    "loss": 0.4500212073326111,
    "learning_rate": 0.0007346587261050165,
    "timestep_mean": 480.625,
    "epsilon_norm": 8.931846618652344,
    "prediction_norm": 6.832881927490234,
    "timestamp": "2025-07-14T12:33:57.919758"
  },
  {
    "epoch": 689,
    "batch_idx": 64,
    "loss": 0.32884490489959717,
    "learning_rate": 0.0007346587261050165,
    "timestep_mean": 491.625,
    "epsilon_norm": 7.447397708892822,
    "prediction_norm": 6.176425933837891,
    "timestamp": "2025-07-14T12:33:57.924472"
  },
  {
    "epoch": 689,
    "batch_idx": 96,
    "loss": 0.31227922439575195,
    "learning_rate": 0.0007346587261050165,
    "timestep_mean": 453.375,
    "epsilon_norm": 8.959206581115723,
    "prediction_norm": 7.27318000793457,
    "timestamp": "2025-07-14T12:33:57.929267"
  },
  {
    "epoch": 689,
    "batch_idx": 128,
    "loss": 0.23657934367656708,
    "learning_rate": 0.0007346587261050165,
    "timestep_mean": 435.96875,
    "epsilon_norm": 7.194814682006836,
    "prediction_norm": 6.138340950012207,
    "timestamp": "2025-07-14T12:33:57.933920"
  },
  {
    "epoch": 689,
    "batch_idx": 160,
    "loss": 0.6854512095451355,
    "learning_rate": 0.0007346587261050165,
    "timestep_mean": 453.25,
    "epsilon_norm": 9.654060363769531,
    "prediction_norm": 7.067469596862793,
    "timestamp": "2025-07-14T12:33:57.938617"
  },
  {
    "epoch": 689,
    "batch_idx": 192,
    "loss": 0.2577691972255707,
    "learning_rate": 0.0007346587261050165,
    "timestep_mean": 436.375,
    "epsilon_norm": 3.115957260131836,
    "prediction_norm": 1.828224778175354,
    "timestamp": "2025-07-14T12:33:57.943283"
  },
  {
    "epoch": 690,
    "batch_idx": 0,
    "loss": 0.3255695402622223,
    "learning_rate": 0.0007339649071302867,
    "timestep_mean": 436.03125,
    "epsilon_norm": 7.47683572769165,
    "prediction_norm": 6.378402233123779,
    "timestamp": "2025-07-14T12:33:57.948061"
  },
  {
    "epoch": 690,
    "batch_idx": 32,
    "loss": 0.32673412561416626,
    "learning_rate": 0.0007339649071302867,
    "timestep_mean": 495.875,
    "epsilon_norm": 7.930936336517334,
    "prediction_norm": 6.765379428863525,
    "timestamp": "2025-07-14T12:33:57.952790"
  },
  {
    "epoch": 690,
    "batch_idx": 64,
    "loss": 0.42996275424957275,
    "learning_rate": 0.0007339649071302867,
    "timestep_mean": 400.96875,
    "epsilon_norm": 7.494848251342773,
    "prediction_norm": 6.279111385345459,
    "timestamp": "2025-07-14T12:33:57.957510"
  },
  {
    "epoch": 690,
    "batch_idx": 96,
    "loss": 0.49469488859176636,
    "learning_rate": 0.0007339649071302867,
    "timestep_mean": 437.34375,
    "epsilon_norm": 7.5767741203308105,
    "prediction_norm": 5.944107532501221,
    "timestamp": "2025-07-14T12:33:57.962219"
  },
  {
    "epoch": 690,
    "batch_idx": 128,
    "loss": 0.2545081377029419,
    "learning_rate": 0.0007339649071302867,
    "timestep_mean": 540.96875,
    "epsilon_norm": 8.64868450164795,
    "prediction_norm": 7.763833045959473,
    "timestamp": "2025-07-14T12:33:57.966985"
  },
  {
    "epoch": 690,
    "batch_idx": 160,
    "loss": 0.1666884422302246,
    "learning_rate": 0.0007339649071302867,
    "timestep_mean": 481.4375,
    "epsilon_norm": 6.9814558029174805,
    "prediction_norm": 6.218928337097168,
    "timestamp": "2025-07-14T12:33:57.971699"
  },
  {
    "epoch": 690,
    "batch_idx": 192,
    "loss": 0.10351847112178802,
    "learning_rate": 0.0007339649071302867,
    "timestep_mean": 591.125,
    "epsilon_norm": 5.358413219451904,
    "prediction_norm": 5.179907321929932,
    "timestamp": "2025-07-14T12:33:57.976439"
  },
  {
    "epoch": 691,
    "batch_idx": 0,
    "loss": 0.34797149896621704,
    "learning_rate": 0.0007332705108704064,
    "timestep_mean": 420.34375,
    "epsilon_norm": 8.575292587280273,
    "prediction_norm": 5.966057777404785,
    "timestamp": "2025-07-14T12:33:57.981187"
  },
  {
    "epoch": 691,
    "batch_idx": 32,
    "loss": 0.2327662706375122,
    "learning_rate": 0.0007332705108704064,
    "timestep_mean": 485.5,
    "epsilon_norm": 7.2327752113342285,
    "prediction_norm": 5.894917011260986,
    "timestamp": "2025-07-14T12:33:57.985939"
  },
  {
    "epoch": 691,
    "batch_idx": 64,
    "loss": 0.3268430233001709,
    "learning_rate": 0.0007332705108704064,
    "timestep_mean": 520.5,
    "epsilon_norm": 8.101399421691895,
    "prediction_norm": 6.610337257385254,
    "timestamp": "2025-07-14T12:33:57.990648"
  },
  {
    "epoch": 691,
    "batch_idx": 96,
    "loss": 0.28084707260131836,
    "learning_rate": 0.0007332705108704064,
    "timestep_mean": 418.25,
    "epsilon_norm": 8.308019638061523,
    "prediction_norm": 7.808736801147461,
    "timestamp": "2025-07-14T12:33:57.995398"
  },
  {
    "epoch": 691,
    "batch_idx": 128,
    "loss": 0.23914143443107605,
    "learning_rate": 0.0007332705108704064,
    "timestep_mean": 543.34375,
    "epsilon_norm": 8.75780963897705,
    "prediction_norm": 7.927838325500488,
    "timestamp": "2025-07-14T12:33:58.000087"
  },
  {
    "epoch": 691,
    "batch_idx": 160,
    "loss": 0.3598794639110565,
    "learning_rate": 0.0007332705108704064,
    "timestep_mean": 522.1875,
    "epsilon_norm": 8.401962280273438,
    "prediction_norm": 6.313910484313965,
    "timestamp": "2025-07-14T12:33:58.004867"
  },
  {
    "epoch": 691,
    "batch_idx": 192,
    "loss": 0.6519439220428467,
    "learning_rate": 0.0007332705108704064,
    "timestep_mean": 469.5,
    "epsilon_norm": 4.251901149749756,
    "prediction_norm": 2.76544189453125,
    "timestamp": "2025-07-14T12:33:58.009621"
  },
  {
    "epoch": 692,
    "batch_idx": 0,
    "loss": 0.3151647746562958,
    "learning_rate": 0.0007325755390387292,
    "timestep_mean": 491.125,
    "epsilon_norm": 8.085478782653809,
    "prediction_norm": 6.570241451263428,
    "timestamp": "2025-07-14T12:33:58.014469"
  },
  {
    "epoch": 692,
    "batch_idx": 32,
    "loss": 0.4289757013320923,
    "learning_rate": 0.0007325755390387292,
    "timestep_mean": 521.125,
    "epsilon_norm": 9.202473640441895,
    "prediction_norm": 7.638974666595459,
    "timestamp": "2025-07-14T12:33:58.019288"
  },
  {
    "epoch": 692,
    "batch_idx": 64,
    "loss": 0.26159048080444336,
    "learning_rate": 0.0007325755390387292,
    "timestep_mean": 550.96875,
    "epsilon_norm": 8.446860313415527,
    "prediction_norm": 7.439999103546143,
    "timestamp": "2025-07-14T12:33:58.024030"
  },
  {
    "epoch": 692,
    "batch_idx": 96,
    "loss": 0.2566365599632263,
    "learning_rate": 0.0007325755390387292,
    "timestep_mean": 461.125,
    "epsilon_norm": 7.346886157989502,
    "prediction_norm": 6.145574569702148,
    "timestamp": "2025-07-14T12:33:58.028855"
  },
  {
    "epoch": 692,
    "batch_idx": 128,
    "loss": 0.2851472496986389,
    "learning_rate": 0.0007325755390387292,
    "timestep_mean": 474.78125,
    "epsilon_norm": 7.165074825286865,
    "prediction_norm": 5.540445327758789,
    "timestamp": "2025-07-14T12:33:58.033720"
  },
  {
    "epoch": 692,
    "batch_idx": 160,
    "loss": 0.20635737478733063,
    "learning_rate": 0.0007325755390387292,
    "timestep_mean": 548.46875,
    "epsilon_norm": 6.866789817810059,
    "prediction_norm": 6.304154396057129,
    "timestamp": "2025-07-14T12:33:58.038518"
  },
  {
    "epoch": 692,
    "batch_idx": 192,
    "loss": 0.24923977255821228,
    "learning_rate": 0.0007325755390387292,
    "timestep_mean": 345.75,
    "epsilon_norm": 2.771589517593384,
    "prediction_norm": 2.0346312522888184,
    "timestamp": "2025-07-14T12:33:58.043321"
  },
  {
    "epoch": 693,
    "batch_idx": 0,
    "loss": 0.24099409580230713,
    "learning_rate": 0.0007318799933500291,
    "timestep_mean": 443.78125,
    "epsilon_norm": 7.817798614501953,
    "prediction_norm": 6.880293369293213,
    "timestamp": "2025-07-14T12:33:58.048220"
  },
  {
    "epoch": 693,
    "batch_idx": 32,
    "loss": 0.24121218919754028,
    "learning_rate": 0.0007318799933500291,
    "timestep_mean": 453.90625,
    "epsilon_norm": 7.783764362335205,
    "prediction_norm": 7.185712814331055,
    "timestamp": "2025-07-14T12:33:58.053080"
  },
  {
    "epoch": 693,
    "batch_idx": 64,
    "loss": 0.29028743505477905,
    "learning_rate": 0.0007318799933500291,
    "timestep_mean": 520.1875,
    "epsilon_norm": 7.574498653411865,
    "prediction_norm": 5.948590278625488,
    "timestamp": "2025-07-14T12:33:58.057953"
  },
  {
    "epoch": 693,
    "batch_idx": 96,
    "loss": 0.46774983406066895,
    "learning_rate": 0.0007318799933500291,
    "timestep_mean": 455.5625,
    "epsilon_norm": 7.806473255157471,
    "prediction_norm": 5.481211185455322,
    "timestamp": "2025-07-14T12:33:58.062942"
  },
  {
    "epoch": 693,
    "batch_idx": 128,
    "loss": 0.2685903310775757,
    "learning_rate": 0.0007318799933500291,
    "timestep_mean": 366.75,
    "epsilon_norm": 8.046588897705078,
    "prediction_norm": 7.6843743324279785,
    "timestamp": "2025-07-14T12:33:58.067783"
  },
  {
    "epoch": 693,
    "batch_idx": 160,
    "loss": 0.29372283816337585,
    "learning_rate": 0.0007318799933500291,
    "timestep_mean": 438.6875,
    "epsilon_norm": 7.716843128204346,
    "prediction_norm": 6.426339626312256,
    "timestamp": "2025-07-14T12:33:58.072618"
  },
  {
    "epoch": 693,
    "batch_idx": 192,
    "loss": 0.05594497546553612,
    "learning_rate": 0.0007318799933500291,
    "timestep_mean": 746.875,
    "epsilon_norm": 4.249956130981445,
    "prediction_norm": 4.461178302764893,
    "timestamp": "2025-07-14T12:33:58.077473"
  },
  {
    "epoch": 694,
    "batch_idx": 0,
    "loss": 0.3340628147125244,
    "learning_rate": 0.0007311838755204957,
    "timestep_mean": 488.71875,
    "epsilon_norm": 7.673023700714111,
    "prediction_norm": 6.453580856323242,
    "timestamp": "2025-07-14T12:33:58.082306"
  },
  {
    "epoch": 694,
    "batch_idx": 32,
    "loss": 0.1835033893585205,
    "learning_rate": 0.0007311838755204957,
    "timestep_mean": 542.1875,
    "epsilon_norm": 8.101311683654785,
    "prediction_norm": 6.587826728820801,
    "timestamp": "2025-07-14T12:33:58.087420"
  },
  {
    "epoch": 694,
    "batch_idx": 64,
    "loss": 0.33461493253707886,
    "learning_rate": 0.0007311838755204957,
    "timestep_mean": 433.40625,
    "epsilon_norm": 8.669220924377441,
    "prediction_norm": 7.166617393493652,
    "timestamp": "2025-07-14T12:33:58.092580"
  },
  {
    "epoch": 694,
    "batch_idx": 96,
    "loss": 0.3326309323310852,
    "learning_rate": 0.0007311838755204957,
    "timestep_mean": 433.96875,
    "epsilon_norm": 7.4828410148620605,
    "prediction_norm": 6.209640979766846,
    "timestamp": "2025-07-14T12:33:58.097712"
  },
  {
    "epoch": 694,
    "batch_idx": 128,
    "loss": 0.24596035480499268,
    "learning_rate": 0.0007311838755204957,
    "timestep_mean": 575.75,
    "epsilon_norm": 7.5071821212768555,
    "prediction_norm": 6.196455478668213,
    "timestamp": "2025-07-14T12:33:58.102903"
  },
  {
    "epoch": 694,
    "batch_idx": 160,
    "loss": 0.18950165808200836,
    "learning_rate": 0.0007311838755204957,
    "timestep_mean": 528.15625,
    "epsilon_norm": 8.017122268676758,
    "prediction_norm": 7.1842265129089355,
    "timestamp": "2025-07-14T12:33:58.108229"
  },
  {
    "epoch": 694,
    "batch_idx": 192,
    "loss": 0.6919382810592651,
    "learning_rate": 0.0007311838755204957,
    "timestep_mean": 411.75,
    "epsilon_norm": 4.89389181137085,
    "prediction_norm": 3.2128183841705322,
    "timestamp": "2025-07-14T12:33:58.113249"
  },
  {
    "epoch": 695,
    "batch_idx": 0,
    "loss": 0.1740177869796753,
    "learning_rate": 0.0007304871872677311,
    "timestep_mean": 508.59375,
    "epsilon_norm": 7.1290740966796875,
    "prediction_norm": 6.462034225463867,
    "timestamp": "2025-07-14T12:33:58.118402"
  },
  {
    "epoch": 695,
    "batch_idx": 32,
    "loss": 0.28984183073043823,
    "learning_rate": 0.0007304871872677311,
    "timestep_mean": 597.65625,
    "epsilon_norm": 8.481461524963379,
    "prediction_norm": 7.303550720214844,
    "timestamp": "2025-07-14T12:33:58.123509"
  },
  {
    "epoch": 695,
    "batch_idx": 64,
    "loss": 0.2670609951019287,
    "learning_rate": 0.0007304871872677311,
    "timestep_mean": 449.9375,
    "epsilon_norm": 7.754383563995361,
    "prediction_norm": 6.175677299499512,
    "timestamp": "2025-07-14T12:33:58.128672"
  },
  {
    "epoch": 695,
    "batch_idx": 96,
    "loss": 0.33088573813438416,
    "learning_rate": 0.0007304871872677311,
    "timestep_mean": 524.25,
    "epsilon_norm": 8.524518966674805,
    "prediction_norm": 6.725390911102295,
    "timestamp": "2025-07-14T12:33:58.133944"
  },
  {
    "epoch": 695,
    "batch_idx": 128,
    "loss": 0.41029155254364014,
    "learning_rate": 0.0007304871872677311,
    "timestep_mean": 438.9375,
    "epsilon_norm": 7.804800510406494,
    "prediction_norm": 6.278250694274902,
    "timestamp": "2025-07-14T12:33:58.139070"
  },
  {
    "epoch": 695,
    "batch_idx": 160,
    "loss": 0.1777508705854416,
    "learning_rate": 0.0007304871872677311,
    "timestep_mean": 458.90625,
    "epsilon_norm": 6.804531097412109,
    "prediction_norm": 6.111474990844727,
    "timestamp": "2025-07-14T12:33:58.144202"
  },
  {
    "epoch": 695,
    "batch_idx": 192,
    "loss": 0.3211989104747772,
    "learning_rate": 0.0007304871872677311,
    "timestep_mean": 489.0,
    "epsilon_norm": 3.5725057125091553,
    "prediction_norm": 2.962202787399292,
    "timestamp": "2025-07-14T12:33:58.149325"
  },
  {
    "epoch": 696,
    "batch_idx": 0,
    "loss": 0.26349973678588867,
    "learning_rate": 0.0007297899303107439,
    "timestep_mean": 514.34375,
    "epsilon_norm": 8.912883758544922,
    "prediction_norm": 7.544497966766357,
    "timestamp": "2025-07-14T12:33:58.154560"
  },
  {
    "epoch": 696,
    "batch_idx": 32,
    "loss": 0.17683041095733643,
    "learning_rate": 0.0007297899303107439,
    "timestep_mean": 549.125,
    "epsilon_norm": 7.340088367462158,
    "prediction_norm": 6.8321356773376465,
    "timestamp": "2025-07-14T12:33:58.159641"
  },
  {
    "epoch": 696,
    "batch_idx": 64,
    "loss": 0.3180287480354309,
    "learning_rate": 0.0007297899303107439,
    "timestep_mean": 482.9375,
    "epsilon_norm": 7.871426582336426,
    "prediction_norm": 6.177069187164307,
    "timestamp": "2025-07-14T12:33:58.164723"
  },
  {
    "epoch": 696,
    "batch_idx": 96,
    "loss": 0.16942816972732544,
    "learning_rate": 0.0007297899303107439,
    "timestep_mean": 521.6875,
    "epsilon_norm": 8.038710594177246,
    "prediction_norm": 7.342776298522949,
    "timestamp": "2025-07-14T12:33:58.169881"
  },
  {
    "epoch": 696,
    "batch_idx": 128,
    "loss": 0.2547982335090637,
    "learning_rate": 0.0007297899303107439,
    "timestep_mean": 487.3125,
    "epsilon_norm": 8.91712474822998,
    "prediction_norm": 7.946012020111084,
    "timestamp": "2025-07-14T12:33:58.175164"
  },
  {
    "epoch": 696,
    "batch_idx": 160,
    "loss": 0.14500126242637634,
    "learning_rate": 0.0007297899303107439,
    "timestep_mean": 580.5625,
    "epsilon_norm": 7.095844745635986,
    "prediction_norm": 7.389588832855225,
    "timestamp": "2025-07-14T12:33:58.180260"
  },
  {
    "epoch": 696,
    "batch_idx": 192,
    "loss": 0.5978673696517944,
    "learning_rate": 0.0007297899303107439,
    "timestep_mean": 496.0,
    "epsilon_norm": 4.689656734466553,
    "prediction_norm": 3.389655113220215,
    "timestamp": "2025-07-14T12:33:58.185108"
  },
  {
    "epoch": 697,
    "batch_idx": 0,
    "loss": 0.44238099455833435,
    "learning_rate": 0.0007290921063699462,
    "timestep_mean": 542.6875,
    "epsilon_norm": 8.771265029907227,
    "prediction_norm": 7.008030414581299,
    "timestamp": "2025-07-14T12:33:58.190256"
  },
  {
    "epoch": 697,
    "batch_idx": 32,
    "loss": 0.26770880818367004,
    "learning_rate": 0.0007290921063699462,
    "timestep_mean": 530.9375,
    "epsilon_norm": 7.816702365875244,
    "prediction_norm": 6.098689079284668,
    "timestamp": "2025-07-14T12:33:58.195118"
  },
  {
    "epoch": 697,
    "batch_idx": 64,
    "loss": 0.26533401012420654,
    "learning_rate": 0.0007290921063699462,
    "timestep_mean": 479.5625,
    "epsilon_norm": 9.015960693359375,
    "prediction_norm": 7.184396266937256,
    "timestamp": "2025-07-14T12:33:58.200123"
  },
  {
    "epoch": 697,
    "batch_idx": 96,
    "loss": 0.19408275187015533,
    "learning_rate": 0.0007290921063699462,
    "timestep_mean": 497.0625,
    "epsilon_norm": 8.222640037536621,
    "prediction_norm": 7.890372276306152,
    "timestamp": "2025-07-14T12:33:58.204951"
  },
  {
    "epoch": 697,
    "batch_idx": 128,
    "loss": 0.12523585557937622,
    "learning_rate": 0.0007290921063699462,
    "timestep_mean": 526.03125,
    "epsilon_norm": 7.919952392578125,
    "prediction_norm": 7.393045425415039,
    "timestamp": "2025-07-14T12:33:58.209852"
  },
  {
    "epoch": 697,
    "batch_idx": 160,
    "loss": 0.19804377853870392,
    "learning_rate": 0.0007290921063699462,
    "timestep_mean": 594.875,
    "epsilon_norm": 6.824826717376709,
    "prediction_norm": 6.297140121459961,
    "timestamp": "2025-07-14T12:33:58.214695"
  },
  {
    "epoch": 697,
    "batch_idx": 192,
    "loss": 0.15109485387802124,
    "learning_rate": 0.0007290921063699462,
    "timestep_mean": 383.25,
    "epsilon_norm": 3.5353877544403076,
    "prediction_norm": 3.5379421710968018,
    "timestamp": "2025-07-14T12:33:58.219527"
  },
  {
    "epoch": 698,
    "batch_idx": 0,
    "loss": 0.6244339346885681,
    "learning_rate": 0.0007283937171671497,
    "timestep_mean": 432.0,
    "epsilon_norm": 8.294034957885742,
    "prediction_norm": 5.662026882171631,
    "timestamp": "2025-07-14T12:33:58.224432"
  },
  {
    "epoch": 698,
    "batch_idx": 32,
    "loss": 0.28609776496887207,
    "learning_rate": 0.0007283937171671497,
    "timestep_mean": 456.03125,
    "epsilon_norm": 6.638619899749756,
    "prediction_norm": 5.60569429397583,
    "timestamp": "2025-07-14T12:33:58.229402"
  },
  {
    "epoch": 698,
    "batch_idx": 64,
    "loss": 0.3494429588317871,
    "learning_rate": 0.0007283937171671497,
    "timestep_mean": 450.875,
    "epsilon_norm": 7.585149765014648,
    "prediction_norm": 6.235799789428711,
    "timestamp": "2025-07-14T12:33:58.234318"
  },
  {
    "epoch": 698,
    "batch_idx": 96,
    "loss": 0.16967836022377014,
    "learning_rate": 0.0007283937171671497,
    "timestep_mean": 594.34375,
    "epsilon_norm": 8.087845802307129,
    "prediction_norm": 7.26338529586792,
    "timestamp": "2025-07-14T12:33:58.239140"
  },
  {
    "epoch": 698,
    "batch_idx": 128,
    "loss": 0.1534317135810852,
    "learning_rate": 0.0007283937171671497,
    "timestep_mean": 611.71875,
    "epsilon_norm": 7.079888343811035,
    "prediction_norm": 6.093320846557617,
    "timestamp": "2025-07-14T12:33:58.243975"
  },
  {
    "epoch": 698,
    "batch_idx": 160,
    "loss": 0.19867104291915894,
    "learning_rate": 0.0007283937171671497,
    "timestep_mean": 522.78125,
    "epsilon_norm": 8.160553932189941,
    "prediction_norm": 7.484668731689453,
    "timestamp": "2025-07-14T12:33:58.248835"
  },
  {
    "epoch": 698,
    "batch_idx": 192,
    "loss": 0.24001917243003845,
    "learning_rate": 0.0007283937171671497,
    "timestep_mean": 453.375,
    "epsilon_norm": 3.961695909500122,
    "prediction_norm": 3.4241831302642822,
    "timestamp": "2025-07-14T12:33:58.253638"
  },
  {
    "epoch": 699,
    "batch_idx": 0,
    "loss": 0.33075571060180664,
    "learning_rate": 0.00072769476442556,
    "timestep_mean": 461.34375,
    "epsilon_norm": 7.706783294677734,
    "prediction_norm": 6.296471118927002,
    "timestamp": "2025-07-14T12:33:58.258595"
  },
  {
    "epoch": 699,
    "batch_idx": 32,
    "loss": 0.2929225265979767,
    "learning_rate": 0.00072769476442556,
    "timestep_mean": 453.875,
    "epsilon_norm": 8.011741638183594,
    "prediction_norm": 6.476785182952881,
    "timestamp": "2025-07-14T12:33:58.263504"
  },
  {
    "epoch": 699,
    "batch_idx": 64,
    "loss": 0.40814733505249023,
    "learning_rate": 0.00072769476442556,
    "timestep_mean": 482.0,
    "epsilon_norm": 7.701401233673096,
    "prediction_norm": 5.531338214874268,
    "timestamp": "2025-07-14T12:33:58.268400"
  },
  {
    "epoch": 699,
    "batch_idx": 96,
    "loss": 0.4276607632637024,
    "learning_rate": 0.00072769476442556,
    "timestep_mean": 498.625,
    "epsilon_norm": 8.949551582336426,
    "prediction_norm": 6.534580707550049,
    "timestamp": "2025-07-14T12:33:58.273249"
  },
  {
    "epoch": 699,
    "batch_idx": 128,
    "loss": 0.21625494956970215,
    "learning_rate": 0.00072769476442556,
    "timestep_mean": 561.0,
    "epsilon_norm": 8.30355167388916,
    "prediction_norm": 6.709305286407471,
    "timestamp": "2025-07-14T12:33:58.278134"
  },
  {
    "epoch": 699,
    "batch_idx": 160,
    "loss": 0.3572952449321747,
    "learning_rate": 0.00072769476442556,
    "timestep_mean": 481.625,
    "epsilon_norm": 7.378661632537842,
    "prediction_norm": 5.637676239013672,
    "timestamp": "2025-07-14T12:33:58.283074"
  },
  {
    "epoch": 699,
    "batch_idx": 192,
    "loss": 0.22801992297172546,
    "learning_rate": 0.00072769476442556,
    "timestep_mean": 617.625,
    "epsilon_norm": 3.3953888416290283,
    "prediction_norm": 2.8494949340820312,
    "timestamp": "2025-07-14T12:33:58.287902"
  },
  {
    "epoch": 700,
    "batch_idx": 0,
    "loss": 0.1571107655763626,
    "learning_rate": 0.0007269952498697733,
    "timestep_mean": 573.84375,
    "epsilon_norm": 7.75475549697876,
    "prediction_norm": 7.245110511779785,
    "timestamp": "2025-07-14T12:33:58.292713"
  },
  {
    "epoch": 700,
    "batch_idx": 32,
    "loss": 0.1518867462873459,
    "learning_rate": 0.0007269952498697733,
    "timestep_mean": 553.90625,
    "epsilon_norm": 6.607979774475098,
    "prediction_norm": 6.237055778503418,
    "timestamp": "2025-07-14T12:33:58.297581"
  },
  {
    "epoch": 700,
    "batch_idx": 64,
    "loss": 0.3122628331184387,
    "learning_rate": 0.0007269952498697733,
    "timestep_mean": 462.53125,
    "epsilon_norm": 9.149871826171875,
    "prediction_norm": 7.5569376945495605,
    "timestamp": "2025-07-14T12:33:58.302472"
  },
  {
    "epoch": 700,
    "batch_idx": 96,
    "loss": 0.36333662271499634,
    "learning_rate": 0.0007269952498697733,
    "timestep_mean": 465.375,
    "epsilon_norm": 8.044027328491211,
    "prediction_norm": 6.527718544006348,
    "timestamp": "2025-07-14T12:33:58.307292"
  },
  {
    "epoch": 700,
    "batch_idx": 128,
    "loss": 0.35195010900497437,
    "learning_rate": 0.0007269952498697733,
    "timestep_mean": 520.90625,
    "epsilon_norm": 7.866118431091309,
    "prediction_norm": 6.647599220275879,
    "timestamp": "2025-07-14T12:33:58.312145"
  },
  {
    "epoch": 700,
    "batch_idx": 160,
    "loss": 0.3552209734916687,
    "learning_rate": 0.0007269952498697733,
    "timestep_mean": 475.46875,
    "epsilon_norm": 9.253350257873535,
    "prediction_norm": 8.175055503845215,
    "timestamp": "2025-07-14T12:33:58.317016"
  },
  {
    "epoch": 700,
    "batch_idx": 192,
    "loss": 0.19447505474090576,
    "learning_rate": 0.0007269952498697733,
    "timestep_mean": 435.125,
    "epsilon_norm": 3.5326602458953857,
    "prediction_norm": 3.0426676273345947,
    "timestamp": "2025-07-14T12:33:58.321828"
  },
  {
    "epoch": 701,
    "batch_idx": 0,
    "loss": 0.20343202352523804,
    "learning_rate": 0.0007262951752257727,
    "timestep_mean": 487.46875,
    "epsilon_norm": 7.530087947845459,
    "prediction_norm": 6.8821868896484375,
    "timestamp": "2025-07-14T12:33:58.326754"
  },
  {
    "epoch": 701,
    "batch_idx": 32,
    "loss": 0.21742203831672668,
    "learning_rate": 0.0007262951752257727,
    "timestep_mean": 441.65625,
    "epsilon_norm": 8.242959022521973,
    "prediction_norm": 7.386083602905273,
    "timestamp": "2025-07-14T12:33:58.331626"
  },
  {
    "epoch": 701,
    "batch_idx": 64,
    "loss": 0.3101736903190613,
    "learning_rate": 0.0007262951752257727,
    "timestep_mean": 494.625,
    "epsilon_norm": 8.614511489868164,
    "prediction_norm": 7.165955543518066,
    "timestamp": "2025-07-14T12:33:58.336484"
  },
  {
    "epoch": 701,
    "batch_idx": 96,
    "loss": 0.3078107237815857,
    "learning_rate": 0.0007262951752257727,
    "timestep_mean": 448.1875,
    "epsilon_norm": 6.842981815338135,
    "prediction_norm": 5.498386859893799,
    "timestamp": "2025-07-14T12:33:58.341383"
  },
  {
    "epoch": 701,
    "batch_idx": 128,
    "loss": 0.3264678120613098,
    "learning_rate": 0.0007262951752257727,
    "timestep_mean": 548.875,
    "epsilon_norm": 7.273016452789307,
    "prediction_norm": 6.258686065673828,
    "timestamp": "2025-07-14T12:33:58.346320"
  },
  {
    "epoch": 701,
    "batch_idx": 160,
    "loss": 0.43403539061546326,
    "learning_rate": 0.0007262951752257727,
    "timestep_mean": 382.1875,
    "epsilon_norm": 7.970625400543213,
    "prediction_norm": 6.40268087387085,
    "timestamp": "2025-07-14T12:33:58.351200"
  },
  {
    "epoch": 701,
    "batch_idx": 192,
    "loss": 0.34574419260025024,
    "learning_rate": 0.0007262951752257727,
    "timestep_mean": 492.625,
    "epsilon_norm": 4.100465774536133,
    "prediction_norm": 3.888040781021118,
    "timestamp": "2025-07-14T12:33:58.356030"
  },
  {
    "epoch": 702,
    "batch_idx": 0,
    "loss": 0.27687758207321167,
    "learning_rate": 0.0007255945422209225,
    "timestep_mean": 556.0625,
    "epsilon_norm": 7.623223304748535,
    "prediction_norm": 6.470264911651611,
    "timestamp": "2025-07-14T12:33:58.360882"
  },
  {
    "epoch": 702,
    "batch_idx": 32,
    "loss": 0.3001948893070221,
    "learning_rate": 0.0007255945422209225,
    "timestep_mean": 639.3125,
    "epsilon_norm": 8.165058135986328,
    "prediction_norm": 6.993256092071533,
    "timestamp": "2025-07-14T12:33:58.365768"
  },
  {
    "epoch": 702,
    "batch_idx": 64,
    "loss": 0.26442351937294006,
    "learning_rate": 0.0007255945422209225,
    "timestep_mean": 499.875,
    "epsilon_norm": 7.99967622756958,
    "prediction_norm": 6.773791790008545,
    "timestamp": "2025-07-14T12:33:58.370598"
  },
  {
    "epoch": 702,
    "batch_idx": 96,
    "loss": 0.2751343846321106,
    "learning_rate": 0.0007255945422209225,
    "timestep_mean": 572.0625,
    "epsilon_norm": 8.601043701171875,
    "prediction_norm": 7.368783473968506,
    "timestamp": "2025-07-14T12:33:58.375488"
  },
  {
    "epoch": 702,
    "batch_idx": 128,
    "loss": 0.37063032388687134,
    "learning_rate": 0.0007255945422209225,
    "timestep_mean": 555.0625,
    "epsilon_norm": 9.048720359802246,
    "prediction_norm": 7.950344562530518,
    "timestamp": "2025-07-14T12:33:58.380367"
  },
  {
    "epoch": 702,
    "batch_idx": 160,
    "loss": 0.2771691083908081,
    "learning_rate": 0.0007255945422209225,
    "timestep_mean": 492.25,
    "epsilon_norm": 8.541585922241211,
    "prediction_norm": 7.259130001068115,
    "timestamp": "2025-07-14T12:33:58.385237"
  },
  {
    "epoch": 702,
    "batch_idx": 192,
    "loss": 0.6300097703933716,
    "learning_rate": 0.0007255945422209225,
    "timestep_mean": 523.625,
    "epsilon_norm": 4.088455677032471,
    "prediction_norm": 2.763932943344116,
    "timestamp": "2025-07-14T12:33:58.390122"
  },
  {
    "epoch": 703,
    "batch_idx": 0,
    "loss": 0.2467808574438095,
    "learning_rate": 0.0007248933525839651,
    "timestep_mean": 481.375,
    "epsilon_norm": 7.330109119415283,
    "prediction_norm": 5.746142387390137,
    "timestamp": "2025-07-14T12:33:58.394997"
  },
  {
    "epoch": 703,
    "batch_idx": 32,
    "loss": 0.27732348442077637,
    "learning_rate": 0.0007248933525839651,
    "timestep_mean": 531.65625,
    "epsilon_norm": 7.177150726318359,
    "prediction_norm": 5.417337894439697,
    "timestamp": "2025-07-14T12:33:58.399946"
  },
  {
    "epoch": 703,
    "batch_idx": 64,
    "loss": 0.3486790657043457,
    "learning_rate": 0.0007248933525839651,
    "timestep_mean": 463.3125,
    "epsilon_norm": 8.362993240356445,
    "prediction_norm": 6.709488391876221,
    "timestamp": "2025-07-14T12:33:58.404832"
  },
  {
    "epoch": 703,
    "batch_idx": 96,
    "loss": 0.18824827671051025,
    "learning_rate": 0.0007248933525839651,
    "timestep_mean": 516.9375,
    "epsilon_norm": 6.535276412963867,
    "prediction_norm": 5.523053169250488,
    "timestamp": "2025-07-14T12:33:58.409673"
  },
  {
    "epoch": 703,
    "batch_idx": 128,
    "loss": 0.24982339143753052,
    "learning_rate": 0.0007248933525839651,
    "timestep_mean": 441.34375,
    "epsilon_norm": 8.021463394165039,
    "prediction_norm": 6.246750354766846,
    "timestamp": "2025-07-14T12:33:58.414598"
  },
  {
    "epoch": 703,
    "batch_idx": 160,
    "loss": 0.3703494668006897,
    "learning_rate": 0.0007248933525839651,
    "timestep_mean": 503.0,
    "epsilon_norm": 8.199952125549316,
    "prediction_norm": 6.65310001373291,
    "timestamp": "2025-07-14T12:33:58.419708"
  },
  {
    "epoch": 703,
    "batch_idx": 192,
    "loss": 0.09250333160161972,
    "learning_rate": 0.0007248933525839651,
    "timestep_mean": 510.75,
    "epsilon_norm": 2.885538101196289,
    "prediction_norm": 2.647599935531616,
    "timestamp": "2025-07-14T12:33:58.424676"
  },
  {
    "epoch": 704,
    "batch_idx": 0,
    "loss": 0.3100575804710388,
    "learning_rate": 0.0007241916080450161,
    "timestep_mean": 432.90625,
    "epsilon_norm": 7.452956676483154,
    "prediction_norm": 6.478898525238037,
    "timestamp": "2025-07-14T12:33:58.430158"
  },
  {
    "epoch": 704,
    "batch_idx": 32,
    "loss": 0.20120185613632202,
    "learning_rate": 0.0007241916080450161,
    "timestep_mean": 521.96875,
    "epsilon_norm": 7.521719932556152,
    "prediction_norm": 6.549854755401611,
    "timestamp": "2025-07-14T12:33:58.435081"
  },
  {
    "epoch": 704,
    "batch_idx": 64,
    "loss": 0.30320173501968384,
    "learning_rate": 0.0007241916080450161,
    "timestep_mean": 561.625,
    "epsilon_norm": 7.922633171081543,
    "prediction_norm": 6.754322052001953,
    "timestamp": "2025-07-14T12:33:58.440223"
  },
  {
    "epoch": 704,
    "batch_idx": 96,
    "loss": 0.17626206576824188,
    "learning_rate": 0.0007241916080450161,
    "timestep_mean": 496.5625,
    "epsilon_norm": 6.240403175354004,
    "prediction_norm": 5.306860446929932,
    "timestamp": "2025-07-14T12:33:58.445140"
  },
  {
    "epoch": 704,
    "batch_idx": 128,
    "loss": 0.1421399861574173,
    "learning_rate": 0.0007241916080450161,
    "timestep_mean": 455.03125,
    "epsilon_norm": 7.812759876251221,
    "prediction_norm": 7.091733455657959,
    "timestamp": "2025-07-14T12:33:58.450111"
  },
  {
    "epoch": 704,
    "batch_idx": 160,
    "loss": 0.17882953584194183,
    "learning_rate": 0.0007241916080450161,
    "timestep_mean": 510.5,
    "epsilon_norm": 7.2285590171813965,
    "prediction_norm": 6.290648937225342,
    "timestamp": "2025-07-14T12:33:58.454992"
  },
  {
    "epoch": 704,
    "batch_idx": 192,
    "loss": 0.2620672881603241,
    "learning_rate": 0.0007241916080450161,
    "timestep_mean": 534.75,
    "epsilon_norm": 2.8603177070617676,
    "prediction_norm": 1.5395876169204712,
    "timestamp": "2025-07-14T12:33:58.459891"
  },
  {
    "epoch": 705,
    "batch_idx": 0,
    "loss": 0.3937848210334778,
    "learning_rate": 0.0007234893103355605,
    "timestep_mean": 511.9375,
    "epsilon_norm": 7.367333889007568,
    "prediction_norm": 5.526620388031006,
    "timestamp": "2025-07-14T12:33:58.464801"
  },
  {
    "epoch": 705,
    "batch_idx": 32,
    "loss": 0.29583168029785156,
    "learning_rate": 0.0007234893103355605,
    "timestep_mean": 568.28125,
    "epsilon_norm": 8.574728012084961,
    "prediction_norm": 7.035070896148682,
    "timestamp": "2025-07-14T12:33:58.469895"
  },
  {
    "epoch": 705,
    "batch_idx": 64,
    "loss": 0.24631859362125397,
    "learning_rate": 0.0007234893103355605,
    "timestep_mean": 520.1875,
    "epsilon_norm": 6.951186180114746,
    "prediction_norm": 5.952493667602539,
    "timestamp": "2025-07-14T12:33:58.474896"
  },
  {
    "epoch": 705,
    "batch_idx": 96,
    "loss": 0.21452337503433228,
    "learning_rate": 0.0007234893103355605,
    "timestep_mean": 495.03125,
    "epsilon_norm": 8.073051452636719,
    "prediction_norm": 6.14164924621582,
    "timestamp": "2025-07-14T12:33:58.479774"
  },
  {
    "epoch": 705,
    "batch_idx": 128,
    "loss": 0.2780962586402893,
    "learning_rate": 0.0007234893103355605,
    "timestep_mean": 480.375,
    "epsilon_norm": 7.338830471038818,
    "prediction_norm": 6.153400897979736,
    "timestamp": "2025-07-14T12:33:58.484624"
  },
  {
    "epoch": 705,
    "batch_idx": 160,
    "loss": 0.22252224385738373,
    "learning_rate": 0.0007234893103355605,
    "timestep_mean": 535.53125,
    "epsilon_norm": 7.6748046875,
    "prediction_norm": 6.6717705726623535,
    "timestamp": "2025-07-14T12:33:58.489489"
  },
  {
    "epoch": 705,
    "batch_idx": 192,
    "loss": 0.258340060710907,
    "learning_rate": 0.0007234893103355605,
    "timestep_mean": 474.625,
    "epsilon_norm": 2.6900336742401123,
    "prediction_norm": 2.196096897125244,
    "timestamp": "2025-07-14T12:33:58.494340"
  },
  {
    "epoch": 706,
    "batch_idx": 0,
    "loss": 0.19414769113063812,
    "learning_rate": 0.0007227864611884481,
    "timestep_mean": 543.21875,
    "epsilon_norm": 7.605129241943359,
    "prediction_norm": 6.888883113861084,
    "timestamp": "2025-07-14T12:33:58.507252"
  },
  {
    "epoch": 706,
    "batch_idx": 32,
    "loss": 0.2674003839492798,
    "learning_rate": 0.0007227864611884481,
    "timestep_mean": 611.28125,
    "epsilon_norm": 7.9747209548950195,
    "prediction_norm": 7.275387287139893,
    "timestamp": "2025-07-14T12:33:58.512177"
  },
  {
    "epoch": 706,
    "batch_idx": 64,
    "loss": 0.3391718566417694,
    "learning_rate": 0.0007227864611884481,
    "timestep_mean": 564.40625,
    "epsilon_norm": 7.735576152801514,
    "prediction_norm": 6.9073486328125,
    "timestamp": "2025-07-14T12:33:58.517113"
  },
  {
    "epoch": 706,
    "batch_idx": 96,
    "loss": 0.3181611895561218,
    "learning_rate": 0.0007227864611884481,
    "timestep_mean": 392.6875,
    "epsilon_norm": 8.969284057617188,
    "prediction_norm": 7.7428812980651855,
    "timestamp": "2025-07-14T12:33:58.521961"
  },
  {
    "epoch": 706,
    "batch_idx": 128,
    "loss": 0.37669888138771057,
    "learning_rate": 0.0007227864611884481,
    "timestep_mean": 508.6875,
    "epsilon_norm": 7.310427665710449,
    "prediction_norm": 6.789506435394287,
    "timestamp": "2025-07-14T12:33:58.527027"
  },
  {
    "epoch": 706,
    "batch_idx": 160,
    "loss": 0.14958402514457703,
    "learning_rate": 0.0007227864611884481,
    "timestep_mean": 645.71875,
    "epsilon_norm": 8.094779014587402,
    "prediction_norm": 7.374687194824219,
    "timestamp": "2025-07-14T12:33:58.531855"
  },
  {
    "epoch": 706,
    "batch_idx": 192,
    "loss": 0.25468945503234863,
    "learning_rate": 0.0007227864611884481,
    "timestep_mean": 617.875,
    "epsilon_norm": 4.219775199890137,
    "prediction_norm": 3.5934693813323975,
    "timestamp": "2025-07-14T12:33:58.536800"
  },
  {
    "epoch": 707,
    "batch_idx": 0,
    "loss": 0.2777567505836487,
    "learning_rate": 0.0007220830623378893,
    "timestep_mean": 487.96875,
    "epsilon_norm": 6.999749660491943,
    "prediction_norm": 6.826963424682617,
    "timestamp": "2025-07-14T12:33:58.541687"
  },
  {
    "epoch": 707,
    "batch_idx": 32,
    "loss": 0.32483047246932983,
    "learning_rate": 0.0007220830623378893,
    "timestep_mean": 512.59375,
    "epsilon_norm": 8.705292701721191,
    "prediction_norm": 7.952414035797119,
    "timestamp": "2025-07-14T12:33:58.546429"
  },
  {
    "epoch": 707,
    "batch_idx": 64,
    "loss": 0.14529821276664734,
    "learning_rate": 0.0007220830623378893,
    "timestep_mean": 562.5,
    "epsilon_norm": 6.914387226104736,
    "prediction_norm": 6.9184417724609375,
    "timestamp": "2025-07-14T12:33:58.551471"
  },
  {
    "epoch": 707,
    "batch_idx": 96,
    "loss": 0.33334076404571533,
    "learning_rate": 0.0007220830623378893,
    "timestep_mean": 460.34375,
    "epsilon_norm": 7.757781982421875,
    "prediction_norm": 7.365514755249023,
    "timestamp": "2025-07-14T12:33:58.556282"
  },
  {
    "epoch": 707,
    "batch_idx": 128,
    "loss": 0.2012416124343872,
    "learning_rate": 0.0007220830623378893,
    "timestep_mean": 519.625,
    "epsilon_norm": 7.31065034866333,
    "prediction_norm": 6.934901237487793,
    "timestamp": "2025-07-14T12:33:58.561209"
  },
  {
    "epoch": 707,
    "batch_idx": 160,
    "loss": 0.20866122841835022,
    "learning_rate": 0.0007220830623378893,
    "timestep_mean": 534.03125,
    "epsilon_norm": 7.624945163726807,
    "prediction_norm": 6.562987804412842,
    "timestamp": "2025-07-14T12:33:58.565964"
  },
  {
    "epoch": 707,
    "batch_idx": 192,
    "loss": 0.8378815650939941,
    "learning_rate": 0.0007220830623378893,
    "timestep_mean": 358.625,
    "epsilon_norm": 4.920280456542969,
    "prediction_norm": 3.400996446609497,
    "timestamp": "2025-07-14T12:33:58.571020"
  },
  {
    "epoch": 708,
    "batch_idx": 0,
    "loss": 0.21417683362960815,
    "learning_rate": 0.0007213791155194507,
    "timestep_mean": 542.59375,
    "epsilon_norm": 8.582295417785645,
    "prediction_norm": 7.660714149475098,
    "timestamp": "2025-07-14T12:33:58.575782"
  },
  {
    "epoch": 708,
    "batch_idx": 32,
    "loss": 0.29568248987197876,
    "learning_rate": 0.0007213791155194507,
    "timestep_mean": 449.34375,
    "epsilon_norm": 7.720324516296387,
    "prediction_norm": 6.026623249053955,
    "timestamp": "2025-07-14T12:33:58.580521"
  },
  {
    "epoch": 708,
    "batch_idx": 64,
    "loss": 0.3322007358074188,
    "learning_rate": 0.0007213791155194507,
    "timestep_mean": 482.0625,
    "epsilon_norm": 8.325562477111816,
    "prediction_norm": 6.012332439422607,
    "timestamp": "2025-07-14T12:33:58.585335"
  },
  {
    "epoch": 708,
    "batch_idx": 96,
    "loss": 0.4788714349269867,
    "learning_rate": 0.0007213791155194507,
    "timestep_mean": 489.15625,
    "epsilon_norm": 8.44093132019043,
    "prediction_norm": 5.755306720733643,
    "timestamp": "2025-07-14T12:33:58.590082"
  },
  {
    "epoch": 708,
    "batch_idx": 128,
    "loss": 0.34946393966674805,
    "learning_rate": 0.0007213791155194507,
    "timestep_mean": 483.3125,
    "epsilon_norm": 8.826422691345215,
    "prediction_norm": 7.3295440673828125,
    "timestamp": "2025-07-14T12:33:58.594854"
  },
  {
    "epoch": 708,
    "batch_idx": 160,
    "loss": 0.24202527105808258,
    "learning_rate": 0.0007213791155194507,
    "timestep_mean": 544.71875,
    "epsilon_norm": 7.425210952758789,
    "prediction_norm": 5.912367343902588,
    "timestamp": "2025-07-14T12:33:58.599620"
  },
  {
    "epoch": 708,
    "batch_idx": 192,
    "loss": 0.41631028056144714,
    "learning_rate": 0.0007213791155194507,
    "timestep_mean": 330.875,
    "epsilon_norm": 3.9482357501983643,
    "prediction_norm": 3.291987657546997,
    "timestamp": "2025-07-14T12:33:58.604497"
  },
  {
    "epoch": 709,
    "batch_idx": 0,
    "loss": 0.2139761745929718,
    "learning_rate": 0.0007206746224700513,
    "timestep_mean": 566.0,
    "epsilon_norm": 7.819302558898926,
    "prediction_norm": 6.516278266906738,
    "timestamp": "2025-07-14T12:33:58.609311"
  },
  {
    "epoch": 709,
    "batch_idx": 32,
    "loss": 0.16577771306037903,
    "learning_rate": 0.0007206746224700513,
    "timestep_mean": 549.1875,
    "epsilon_norm": 6.723856449127197,
    "prediction_norm": 5.940332889556885,
    "timestamp": "2025-07-14T12:33:58.614238"
  },
  {
    "epoch": 709,
    "batch_idx": 64,
    "loss": 0.564293384552002,
    "learning_rate": 0.0007206746224700513,
    "timestep_mean": 489.875,
    "epsilon_norm": 8.685395240783691,
    "prediction_norm": 6.395021438598633,
    "timestamp": "2025-07-14T12:33:58.619024"
  },
  {
    "epoch": 709,
    "batch_idx": 96,
    "loss": 0.29448264837265015,
    "learning_rate": 0.0007206746224700513,
    "timestep_mean": 472.8125,
    "epsilon_norm": 8.585908889770508,
    "prediction_norm": 6.748438835144043,
    "timestamp": "2025-07-14T12:33:58.623790"
  },
  {
    "epoch": 709,
    "batch_idx": 128,
    "loss": 0.5701632499694824,
    "learning_rate": 0.0007206746224700513,
    "timestep_mean": 521.4375,
    "epsilon_norm": 8.558859825134277,
    "prediction_norm": 6.718552589416504,
    "timestamp": "2025-07-14T12:33:58.628726"
  },
  {
    "epoch": 709,
    "batch_idx": 160,
    "loss": 0.25325000286102295,
    "learning_rate": 0.0007206746224700513,
    "timestep_mean": 615.59375,
    "epsilon_norm": 8.191967010498047,
    "prediction_norm": 7.210397720336914,
    "timestamp": "2025-07-14T12:33:58.633659"
  },
  {
    "epoch": 709,
    "batch_idx": 192,
    "loss": 0.11689931154251099,
    "learning_rate": 0.0007206746224700513,
    "timestep_mean": 479.125,
    "epsilon_norm": 3.6602909564971924,
    "prediction_norm": 3.1018388271331787,
    "timestamp": "2025-07-14T12:33:58.638833"
  },
  {
    "epoch": 710,
    "batch_idx": 0,
    "loss": 0.4122236371040344,
    "learning_rate": 0.0007199695849279576,
    "timestep_mean": 426.125,
    "epsilon_norm": 7.65323543548584,
    "prediction_norm": 5.347025394439697,
    "timestamp": "2025-07-14T12:33:58.643786"
  },
  {
    "epoch": 710,
    "batch_idx": 32,
    "loss": 0.29670917987823486,
    "learning_rate": 0.0007199695849279576,
    "timestep_mean": 551.65625,
    "epsilon_norm": 7.609105587005615,
    "prediction_norm": 5.989418029785156,
    "timestamp": "2025-07-14T12:33:58.649095"
  },
  {
    "epoch": 710,
    "batch_idx": 64,
    "loss": 0.15146026015281677,
    "learning_rate": 0.0007199695849279576,
    "timestep_mean": 586.0,
    "epsilon_norm": 8.549932479858398,
    "prediction_norm": 7.027929782867432,
    "timestamp": "2025-07-14T12:33:58.654207"
  },
  {
    "epoch": 710,
    "batch_idx": 96,
    "loss": 0.45876115560531616,
    "learning_rate": 0.0007199695849279576,
    "timestep_mean": 477.03125,
    "epsilon_norm": 8.528135299682617,
    "prediction_norm": 6.644417762756348,
    "timestamp": "2025-07-14T12:33:58.659190"
  },
  {
    "epoch": 710,
    "batch_idx": 128,
    "loss": 0.23712441325187683,
    "learning_rate": 0.0007199695849279576,
    "timestep_mean": 566.875,
    "epsilon_norm": 8.399924278259277,
    "prediction_norm": 7.3332109451293945,
    "timestamp": "2025-07-14T12:33:58.664327"
  },
  {
    "epoch": 710,
    "batch_idx": 160,
    "loss": 0.2865622937679291,
    "learning_rate": 0.0007199695849279576,
    "timestep_mean": 488.09375,
    "epsilon_norm": 8.794514656066895,
    "prediction_norm": 7.06041145324707,
    "timestamp": "2025-07-14T12:33:58.669436"
  },
  {
    "epoch": 710,
    "batch_idx": 192,
    "loss": 0.09495183080434799,
    "learning_rate": 0.0007199695849279576,
    "timestep_mean": 638.875,
    "epsilon_norm": 3.492941379547119,
    "prediction_norm": 3.431661367416382,
    "timestamp": "2025-07-14T12:33:58.674468"
  },
  {
    "epoch": 711,
    "batch_idx": 0,
    "loss": 0.2515530288219452,
    "learning_rate": 0.0007192640046327797,
    "timestep_mean": 482.4375,
    "epsilon_norm": 8.199182510375977,
    "prediction_norm": 6.879876613616943,
    "timestamp": "2025-07-14T12:33:58.679775"
  },
  {
    "epoch": 711,
    "batch_idx": 32,
    "loss": 0.3064877986907959,
    "learning_rate": 0.0007192640046327797,
    "timestep_mean": 503.8125,
    "epsilon_norm": 8.320209503173828,
    "prediction_norm": 7.308848857879639,
    "timestamp": "2025-07-14T12:33:58.684834"
  },
  {
    "epoch": 711,
    "batch_idx": 64,
    "loss": 0.30041229724884033,
    "learning_rate": 0.0007192640046327797,
    "timestep_mean": 548.6875,
    "epsilon_norm": 7.5917768478393555,
    "prediction_norm": 6.367720127105713,
    "timestamp": "2025-07-14T12:33:58.689870"
  },
  {
    "epoch": 711,
    "batch_idx": 96,
    "loss": 0.2595120072364807,
    "learning_rate": 0.0007192640046327797,
    "timestep_mean": 546.03125,
    "epsilon_norm": 8.455231666564941,
    "prediction_norm": 7.383439064025879,
    "timestamp": "2025-07-14T12:33:58.695175"
  },
  {
    "epoch": 711,
    "batch_idx": 128,
    "loss": 0.3518888056278229,
    "learning_rate": 0.0007192640046327797,
    "timestep_mean": 497.90625,
    "epsilon_norm": 8.41879653930664,
    "prediction_norm": 6.338743686676025,
    "timestamp": "2025-07-14T12:33:58.700435"
  },
  {
    "epoch": 711,
    "batch_idx": 160,
    "loss": 0.2700846791267395,
    "learning_rate": 0.0007192640046327797,
    "timestep_mean": 524.15625,
    "epsilon_norm": 8.303581237792969,
    "prediction_norm": 7.169771671295166,
    "timestamp": "2025-07-14T12:33:58.705645"
  },
  {
    "epoch": 711,
    "batch_idx": 192,
    "loss": 0.27949246764183044,
    "learning_rate": 0.0007192640046327797,
    "timestep_mean": 528.75,
    "epsilon_norm": 3.7563438415527344,
    "prediction_norm": 3.069169044494629,
    "timestamp": "2025-07-14T12:33:58.710783"
  },
  {
    "epoch": 712,
    "batch_idx": 0,
    "loss": 0.45488476753234863,
    "learning_rate": 0.0007185578833254665,
    "timestep_mean": 421.65625,
    "epsilon_norm": 8.180270195007324,
    "prediction_norm": 5.933549404144287,
    "timestamp": "2025-07-14T12:33:58.715975"
  },
  {
    "epoch": 712,
    "batch_idx": 32,
    "loss": 0.16232958436012268,
    "learning_rate": 0.0007185578833254665,
    "timestep_mean": 526.90625,
    "epsilon_norm": 7.891732215881348,
    "prediction_norm": 7.309345245361328,
    "timestamp": "2025-07-14T12:33:58.721291"
  },
  {
    "epoch": 712,
    "batch_idx": 64,
    "loss": 0.2026020884513855,
    "learning_rate": 0.0007185578833254665,
    "timestep_mean": 570.75,
    "epsilon_norm": 8.248090744018555,
    "prediction_norm": 7.190162658691406,
    "timestamp": "2025-07-14T12:33:58.726445"
  },
  {
    "epoch": 712,
    "batch_idx": 96,
    "loss": 0.3787666857242584,
    "learning_rate": 0.0007185578833254665,
    "timestep_mean": 495.1875,
    "epsilon_norm": 9.111288070678711,
    "prediction_norm": 7.447572231292725,
    "timestamp": "2025-07-14T12:33:58.731730"
  },
  {
    "epoch": 712,
    "batch_idx": 128,
    "loss": 0.41538718342781067,
    "learning_rate": 0.0007185578833254665,
    "timestep_mean": 458.09375,
    "epsilon_norm": 7.646220684051514,
    "prediction_norm": 6.145707130432129,
    "timestamp": "2025-07-14T12:33:58.736847"
  },
  {
    "epoch": 712,
    "batch_idx": 160,
    "loss": 0.3183387815952301,
    "learning_rate": 0.0007185578833254665,
    "timestep_mean": 424.90625,
    "epsilon_norm": 7.47648811340332,
    "prediction_norm": 6.197359561920166,
    "timestamp": "2025-07-14T12:33:58.742105"
  },
  {
    "epoch": 712,
    "batch_idx": 192,
    "loss": 0.15400080382823944,
    "learning_rate": 0.0007185578833254665,
    "timestep_mean": 447.25,
    "epsilon_norm": 2.313457727432251,
    "prediction_norm": 2.0722877979278564,
    "timestamp": "2025-07-14T12:33:58.747168"
  },
  {
    "epoch": 713,
    "batch_idx": 0,
    "loss": 0.3597206771373749,
    "learning_rate": 0.0007178512227483025,
    "timestep_mean": 521.4375,
    "epsilon_norm": 9.478570938110352,
    "prediction_norm": 8.363513946533203,
    "timestamp": "2025-07-14T12:33:58.752420"
  },
  {
    "epoch": 713,
    "batch_idx": 32,
    "loss": 0.22771702706813812,
    "learning_rate": 0.0007178512227483025,
    "timestep_mean": 564.4375,
    "epsilon_norm": 8.24101734161377,
    "prediction_norm": 7.617284297943115,
    "timestamp": "2025-07-14T12:33:58.757546"
  },
  {
    "epoch": 713,
    "batch_idx": 64,
    "loss": 0.23742279410362244,
    "learning_rate": 0.0007178512227483025,
    "timestep_mean": 464.0,
    "epsilon_norm": 7.729972839355469,
    "prediction_norm": 6.683274745941162,
    "timestamp": "2025-07-14T12:33:58.762745"
  },
  {
    "epoch": 713,
    "batch_idx": 96,
    "loss": 0.33752596378326416,
    "learning_rate": 0.0007178512227483025,
    "timestep_mean": 514.59375,
    "epsilon_norm": 7.599262237548828,
    "prediction_norm": 6.239521026611328,
    "timestamp": "2025-07-14T12:33:58.767883"
  },
  {
    "epoch": 713,
    "batch_idx": 128,
    "loss": 0.19464629888534546,
    "learning_rate": 0.0007178512227483025,
    "timestep_mean": 604.0625,
    "epsilon_norm": 8.322944641113281,
    "prediction_norm": 7.353000640869141,
    "timestamp": "2025-07-14T12:33:58.773089"
  },
  {
    "epoch": 713,
    "batch_idx": 160,
    "loss": 0.11616280674934387,
    "learning_rate": 0.0007178512227483025,
    "timestep_mean": 571.5625,
    "epsilon_norm": 9.22998332977295,
    "prediction_norm": 8.219167709350586,
    "timestamp": "2025-07-14T12:33:58.778156"
  },
  {
    "epoch": 713,
    "batch_idx": 192,
    "loss": 0.3299587070941925,
    "learning_rate": 0.0007178512227483025,
    "timestep_mean": 569.375,
    "epsilon_norm": 4.640188217163086,
    "prediction_norm": 3.4618802070617676,
    "timestamp": "2025-07-14T12:33:58.783507"
  },
  {
    "epoch": 714,
    "batch_idx": 0,
    "loss": 0.32412388920783997,
    "learning_rate": 0.0007171440246449023,
    "timestep_mean": 465.4375,
    "epsilon_norm": 8.533116340637207,
    "prediction_norm": 6.564481735229492,
    "timestamp": "2025-07-14T12:33:58.788663"
  },
  {
    "epoch": 714,
    "batch_idx": 32,
    "loss": 0.2415444403886795,
    "learning_rate": 0.0007171440246449023,
    "timestep_mean": 428.375,
    "epsilon_norm": 7.247196197509766,
    "prediction_norm": 6.334211349487305,
    "timestamp": "2025-07-14T12:33:58.793609"
  },
  {
    "epoch": 714,
    "batch_idx": 64,
    "loss": 0.2653777301311493,
    "learning_rate": 0.0007171440246449023,
    "timestep_mean": 509.4375,
    "epsilon_norm": 8.020550727844238,
    "prediction_norm": 6.689027786254883,
    "timestamp": "2025-07-14T12:33:58.798654"
  },
  {
    "epoch": 714,
    "batch_idx": 96,
    "loss": 0.50739586353302,
    "learning_rate": 0.0007171440246449023,
    "timestep_mean": 505.5,
    "epsilon_norm": 9.604691505432129,
    "prediction_norm": 7.071983814239502,
    "timestamp": "2025-07-14T12:33:58.803842"
  },
  {
    "epoch": 714,
    "batch_idx": 128,
    "loss": 0.3048100769519806,
    "learning_rate": 0.0007171440246449023,
    "timestep_mean": 456.3125,
    "epsilon_norm": 8.395647048950195,
    "prediction_norm": 7.116523742675781,
    "timestamp": "2025-07-14T12:33:58.809002"
  },
  {
    "epoch": 714,
    "batch_idx": 160,
    "loss": 0.20019911229610443,
    "learning_rate": 0.0007171440246449023,
    "timestep_mean": 537.875,
    "epsilon_norm": 7.690849781036377,
    "prediction_norm": 6.523160457611084,
    "timestamp": "2025-07-14T12:33:58.814129"
  },
  {
    "epoch": 714,
    "batch_idx": 192,
    "loss": 0.28626152873039246,
    "learning_rate": 0.0007171440246449023,
    "timestep_mean": 528.0,
    "epsilon_norm": 4.278429985046387,
    "prediction_norm": 3.7104334831237793,
    "timestamp": "2025-07-14T12:33:58.819244"
  },
  {
    "epoch": 715,
    "batch_idx": 0,
    "loss": 0.40761101245880127,
    "learning_rate": 0.000716436290760207,
    "timestep_mean": 524.6875,
    "epsilon_norm": 7.746930122375488,
    "prediction_norm": 5.924557209014893,
    "timestamp": "2025-07-14T12:33:58.824361"
  },
  {
    "epoch": 715,
    "batch_idx": 32,
    "loss": 0.5251151323318481,
    "learning_rate": 0.000716436290760207,
    "timestep_mean": 584.34375,
    "epsilon_norm": 8.88046932220459,
    "prediction_norm": 7.621762275695801,
    "timestamp": "2025-07-14T12:33:58.829570"
  },
  {
    "epoch": 715,
    "batch_idx": 64,
    "loss": 0.34961962699890137,
    "learning_rate": 0.000716436290760207,
    "timestep_mean": 447.46875,
    "epsilon_norm": 7.915564060211182,
    "prediction_norm": 6.710988998413086,
    "timestamp": "2025-07-14T12:33:58.834659"
  },
  {
    "epoch": 715,
    "batch_idx": 96,
    "loss": 0.38164639472961426,
    "learning_rate": 0.000716436290760207,
    "timestep_mean": 492.0625,
    "epsilon_norm": 7.227053642272949,
    "prediction_norm": 5.946262836456299,
    "timestamp": "2025-07-14T12:33:58.839804"
  },
  {
    "epoch": 715,
    "batch_idx": 128,
    "loss": 0.3461979031562805,
    "learning_rate": 0.000716436290760207,
    "timestep_mean": 413.15625,
    "epsilon_norm": 7.845835208892822,
    "prediction_norm": 6.574276447296143,
    "timestamp": "2025-07-14T12:33:58.845104"
  },
  {
    "epoch": 715,
    "batch_idx": 160,
    "loss": 0.37925565242767334,
    "learning_rate": 0.000716436290760207,
    "timestep_mean": 531.90625,
    "epsilon_norm": 9.260435104370117,
    "prediction_norm": 8.746917724609375,
    "timestamp": "2025-07-14T12:33:58.850424"
  },
  {
    "epoch": 715,
    "batch_idx": 192,
    "loss": 0.10470091551542282,
    "learning_rate": 0.000716436290760207,
    "timestep_mean": 360.375,
    "epsilon_norm": 3.117973566055298,
    "prediction_norm": 2.393033504486084,
    "timestamp": "2025-07-14T12:33:58.855633"
  },
  {
    "epoch": 716,
    "batch_idx": 0,
    "loss": 0.30683767795562744,
    "learning_rate": 0.0007157280228404796,
    "timestep_mean": 494.84375,
    "epsilon_norm": 7.078813076019287,
    "prediction_norm": 5.6129069328308105,
    "timestamp": "2025-07-14T12:33:58.860757"
  },
  {
    "epoch": 716,
    "batch_idx": 32,
    "loss": 0.3098263442516327,
    "learning_rate": 0.0007157280228404796,
    "timestep_mean": 500.6875,
    "epsilon_norm": 7.535358905792236,
    "prediction_norm": 6.8320441246032715,
    "timestamp": "2025-07-14T12:33:58.865719"
  },
  {
    "epoch": 716,
    "batch_idx": 64,
    "loss": 0.32781749963760376,
    "learning_rate": 0.0007157280228404796,
    "timestep_mean": 496.71875,
    "epsilon_norm": 8.374560356140137,
    "prediction_norm": 7.794520854949951,
    "timestamp": "2025-07-14T12:33:58.870872"
  },
  {
    "epoch": 716,
    "batch_idx": 96,
    "loss": 0.18555936217308044,
    "learning_rate": 0.0007157280228404796,
    "timestep_mean": 528.84375,
    "epsilon_norm": 8.290787696838379,
    "prediction_norm": 7.855147361755371,
    "timestamp": "2025-07-14T12:33:58.875926"
  },
  {
    "epoch": 716,
    "batch_idx": 128,
    "loss": 0.38603460788726807,
    "learning_rate": 0.0007157280228404796,
    "timestep_mean": 452.84375,
    "epsilon_norm": 7.2517008781433105,
    "prediction_norm": 5.258768081665039,
    "timestamp": "2025-07-14T12:33:58.880898"
  },
  {
    "epoch": 716,
    "batch_idx": 160,
    "loss": 0.4130040407180786,
    "learning_rate": 0.0007157280228404796,
    "timestep_mean": 487.71875,
    "epsilon_norm": 8.30051326751709,
    "prediction_norm": 6.474512100219727,
    "timestamp": "2025-07-14T12:33:58.886012"
  },
  {
    "epoch": 716,
    "batch_idx": 192,
    "loss": 0.11513134837150574,
    "learning_rate": 0.0007157280228404796,
    "timestep_mean": 578.75,
    "epsilon_norm": 4.053394317626953,
    "prediction_norm": 4.214474201202393,
    "timestamp": "2025-07-14T12:33:58.891268"
  },
  {
    "epoch": 717,
    "batch_idx": 0,
    "loss": 0.5013995170593262,
    "learning_rate": 0.0007150192226333008,
    "timestep_mean": 509.28125,
    "epsilon_norm": 8.307562828063965,
    "prediction_norm": 5.760359287261963,
    "timestamp": "2025-07-14T12:33:58.896333"
  },
  {
    "epoch": 717,
    "batch_idx": 32,
    "loss": 0.3713052272796631,
    "learning_rate": 0.0007150192226333008,
    "timestep_mean": 542.625,
    "epsilon_norm": 8.817614555358887,
    "prediction_norm": 7.890945911407471,
    "timestamp": "2025-07-14T12:33:58.901449"
  },
  {
    "epoch": 717,
    "batch_idx": 64,
    "loss": 0.23426376283168793,
    "learning_rate": 0.0007150192226333008,
    "timestep_mean": 513.09375,
    "epsilon_norm": 8.436675071716309,
    "prediction_norm": 7.130530834197998,
    "timestamp": "2025-07-14T12:33:58.906660"
  },
  {
    "epoch": 717,
    "batch_idx": 96,
    "loss": 0.3019406497478485,
    "learning_rate": 0.0007150192226333008,
    "timestep_mean": 476.375,
    "epsilon_norm": 7.959503650665283,
    "prediction_norm": 6.0099897384643555,
    "timestamp": "2025-07-14T12:33:58.911678"
  },
  {
    "epoch": 717,
    "batch_idx": 128,
    "loss": 0.32571524381637573,
    "learning_rate": 0.0007150192226333008,
    "timestep_mean": 504.5,
    "epsilon_norm": 8.344780921936035,
    "prediction_norm": 6.792454242706299,
    "timestamp": "2025-07-14T12:33:58.916804"
  },
  {
    "epoch": 717,
    "batch_idx": 160,
    "loss": 0.4382486045360565,
    "learning_rate": 0.0007150192226333008,
    "timestep_mean": 458.96875,
    "epsilon_norm": 8.345723152160645,
    "prediction_norm": 6.117888450622559,
    "timestamp": "2025-07-14T12:33:58.921924"
  },
  {
    "epoch": 717,
    "batch_idx": 192,
    "loss": 0.06460248678922653,
    "learning_rate": 0.0007150192226333008,
    "timestep_mean": 744.375,
    "epsilon_norm": 3.914433717727661,
    "prediction_norm": 3.489776372909546,
    "timestamp": "2025-07-14T12:33:58.926932"
  },
  {
    "epoch": 718,
    "batch_idx": 0,
    "loss": 0.27485302090644836,
    "learning_rate": 0.0007143098918875643,
    "timestep_mean": 504.9375,
    "epsilon_norm": 7.690449237823486,
    "prediction_norm": 6.016412734985352,
    "timestamp": "2025-07-14T12:33:58.932030"
  },
  {
    "epoch": 718,
    "batch_idx": 32,
    "loss": 0.20692035555839539,
    "learning_rate": 0.0007143098918875643,
    "timestep_mean": 452.84375,
    "epsilon_norm": 7.444101333618164,
    "prediction_norm": 5.8371686935424805,
    "timestamp": "2025-07-14T12:33:58.937156"
  },
  {
    "epoch": 718,
    "batch_idx": 64,
    "loss": 0.3134737014770508,
    "learning_rate": 0.0007143098918875643,
    "timestep_mean": 526.46875,
    "epsilon_norm": 8.70478343963623,
    "prediction_norm": 6.752562522888184,
    "timestamp": "2025-07-14T12:33:58.942476"
  },
  {
    "epoch": 718,
    "batch_idx": 96,
    "loss": 0.22390583157539368,
    "learning_rate": 0.0007143098918875643,
    "timestep_mean": 494.6875,
    "epsilon_norm": 7.6447062492370605,
    "prediction_norm": 6.413750648498535,
    "timestamp": "2025-07-14T12:33:58.947617"
  },
  {
    "epoch": 718,
    "batch_idx": 128,
    "loss": 0.31097936630249023,
    "learning_rate": 0.0007143098918875643,
    "timestep_mean": 555.15625,
    "epsilon_norm": 8.234563827514648,
    "prediction_norm": 6.392703056335449,
    "timestamp": "2025-07-14T12:33:58.952695"
  },
  {
    "epoch": 718,
    "batch_idx": 160,
    "loss": 0.22951224446296692,
    "learning_rate": 0.0007143098918875643,
    "timestep_mean": 465.125,
    "epsilon_norm": 7.502571105957031,
    "prediction_norm": 5.886164665222168,
    "timestamp": "2025-07-14T12:33:58.957730"
  },
  {
    "epoch": 718,
    "batch_idx": 192,
    "loss": 0.209749236702919,
    "learning_rate": 0.0007143098918875643,
    "timestep_mean": 511.75,
    "epsilon_norm": 3.129166603088379,
    "prediction_norm": 2.3325822353363037,
    "timestamp": "2025-07-14T12:33:58.963067"
  },
  {
    "epoch": 719,
    "batch_idx": 0,
    "loss": 0.2797974944114685,
    "learning_rate": 0.0007136000323534736,
    "timestep_mean": 464.75,
    "epsilon_norm": 8.027216911315918,
    "prediction_norm": 6.185622215270996,
    "timestamp": "2025-07-14T12:33:58.968012"
  },
  {
    "epoch": 719,
    "batch_idx": 32,
    "loss": 0.14931824803352356,
    "learning_rate": 0.0007136000323534736,
    "timestep_mean": 561.53125,
    "epsilon_norm": 7.370213031768799,
    "prediction_norm": 6.672801971435547,
    "timestamp": "2025-07-14T12:33:58.973204"
  },
  {
    "epoch": 719,
    "batch_idx": 64,
    "loss": 0.20686620473861694,
    "learning_rate": 0.0007136000323534736,
    "timestep_mean": 474.09375,
    "epsilon_norm": 6.963998794555664,
    "prediction_norm": 5.791626453399658,
    "timestamp": "2025-07-14T12:33:58.978331"
  },
  {
    "epoch": 719,
    "batch_idx": 96,
    "loss": 0.2297494113445282,
    "learning_rate": 0.0007136000323534736,
    "timestep_mean": 503.90625,
    "epsilon_norm": 8.039173126220703,
    "prediction_norm": 7.305006504058838,
    "timestamp": "2025-07-14T12:33:58.983390"
  },
  {
    "epoch": 719,
    "batch_idx": 128,
    "loss": 0.30911514163017273,
    "learning_rate": 0.0007136000323534736,
    "timestep_mean": 457.71875,
    "epsilon_norm": 6.294522285461426,
    "prediction_norm": 5.288009166717529,
    "timestamp": "2025-07-14T12:33:58.988511"
  },
  {
    "epoch": 719,
    "batch_idx": 160,
    "loss": 0.5005289316177368,
    "learning_rate": 0.0007136000323534736,
    "timestep_mean": 436.34375,
    "epsilon_norm": 8.607104301452637,
    "prediction_norm": 6.7148284912109375,
    "timestamp": "2025-07-14T12:33:58.993619"
  },
  {
    "epoch": 719,
    "batch_idx": 192,
    "loss": 0.13747817277908325,
    "learning_rate": 0.0007136000323534736,
    "timestep_mean": 653.375,
    "epsilon_norm": 4.578871250152588,
    "prediction_norm": 4.723270416259766,
    "timestamp": "2025-07-14T12:33:58.998646"
  },
  {
    "epoch": 720,
    "batch_idx": 0,
    "loss": 0.0966024249792099,
    "learning_rate": 0.0007128896457825365,
    "timestep_mean": 580.0625,
    "epsilon_norm": 6.845454216003418,
    "prediction_norm": 7.044750213623047,
    "timestamp": "2025-07-14T12:33:59.003851"
  },
  {
    "epoch": 720,
    "batch_idx": 32,
    "loss": 0.17339786887168884,
    "learning_rate": 0.0007128896457825365,
    "timestep_mean": 570.90625,
    "epsilon_norm": 8.296174049377441,
    "prediction_norm": 7.599496364593506,
    "timestamp": "2025-07-14T12:33:59.009193"
  },
  {
    "epoch": 720,
    "batch_idx": 64,
    "loss": 0.17542944848537445,
    "learning_rate": 0.0007128896457825365,
    "timestep_mean": 451.0,
    "epsilon_norm": 7.165689468383789,
    "prediction_norm": 6.569559574127197,
    "timestamp": "2025-07-14T12:33:59.014146"
  },
  {
    "epoch": 720,
    "batch_idx": 96,
    "loss": 0.19059397280216217,
    "learning_rate": 0.0007128896457825365,
    "timestep_mean": 589.375,
    "epsilon_norm": 8.444119453430176,
    "prediction_norm": 8.31130313873291,
    "timestamp": "2025-07-14T12:33:59.019232"
  },
  {
    "epoch": 720,
    "batch_idx": 128,
    "loss": 0.2940574288368225,
    "learning_rate": 0.0007128896457825365,
    "timestep_mean": 488.21875,
    "epsilon_norm": 8.415812492370605,
    "prediction_norm": 7.073695182800293,
    "timestamp": "2025-07-14T12:33:59.024309"
  },
  {
    "epoch": 720,
    "batch_idx": 160,
    "loss": 0.16743265092372894,
    "learning_rate": 0.0007128896457825365,
    "timestep_mean": 464.84375,
    "epsilon_norm": 6.872969627380371,
    "prediction_norm": 5.89417028427124,
    "timestamp": "2025-07-14T12:33:59.029478"
  },
  {
    "epoch": 720,
    "batch_idx": 192,
    "loss": 0.25828099250793457,
    "learning_rate": 0.0007128896457825365,
    "timestep_mean": 445.5,
    "epsilon_norm": 3.107813835144043,
    "prediction_norm": 2.3176026344299316,
    "timestamp": "2025-07-14T12:33:59.034561"
  },
  {
    "epoch": 721,
    "batch_idx": 0,
    "loss": 0.19119566679000854,
    "learning_rate": 0.0007121787339275611,
    "timestep_mean": 481.9375,
    "epsilon_norm": 6.795310974121094,
    "prediction_norm": 6.192459583282471,
    "timestamp": "2025-07-14T12:33:59.039661"
  },
  {
    "epoch": 721,
    "batch_idx": 32,
    "loss": 0.5446075201034546,
    "learning_rate": 0.0007121787339275611,
    "timestep_mean": 523.875,
    "epsilon_norm": 8.138826370239258,
    "prediction_norm": 5.0802998542785645,
    "timestamp": "2025-07-14T12:33:59.044723"
  },
  {
    "epoch": 721,
    "batch_idx": 64,
    "loss": 0.3017546534538269,
    "learning_rate": 0.0007121787339275611,
    "timestep_mean": 427.4375,
    "epsilon_norm": 6.798635482788086,
    "prediction_norm": 5.2412495613098145,
    "timestamp": "2025-07-14T12:33:59.049978"
  },
  {
    "epoch": 721,
    "batch_idx": 96,
    "loss": 0.4737449884414673,
    "learning_rate": 0.0007121787339275611,
    "timestep_mean": 445.4375,
    "epsilon_norm": 7.599409103393555,
    "prediction_norm": 5.6032938957214355,
    "timestamp": "2025-07-14T12:33:59.055117"
  },
  {
    "epoch": 721,
    "batch_idx": 128,
    "loss": 0.36482053995132446,
    "learning_rate": 0.0007121787339275611,
    "timestep_mean": 439.15625,
    "epsilon_norm": 8.022509574890137,
    "prediction_norm": 5.6067986488342285,
    "timestamp": "2025-07-14T12:33:59.060120"
  },
  {
    "epoch": 721,
    "batch_idx": 160,
    "loss": 0.29507842659950256,
    "learning_rate": 0.0007121787339275611,
    "timestep_mean": 498.5,
    "epsilon_norm": 6.712803840637207,
    "prediction_norm": 4.6976799964904785,
    "timestamp": "2025-07-14T12:33:59.065271"
  },
  {
    "epoch": 721,
    "batch_idx": 192,
    "loss": 0.07277534157037735,
    "learning_rate": 0.0007121787339275611,
    "timestep_mean": 715.5,
    "epsilon_norm": 3.253708600997925,
    "prediction_norm": 3.3430120944976807,
    "timestamp": "2025-07-14T12:33:59.070302"
  },
  {
    "epoch": 722,
    "batch_idx": 0,
    "loss": 0.3603670597076416,
    "learning_rate": 0.000711467298542652,
    "timestep_mean": 476.46875,
    "epsilon_norm": 8.503808975219727,
    "prediction_norm": 6.737642765045166,
    "timestamp": "2025-07-14T12:33:59.075533"
  },
  {
    "epoch": 722,
    "batch_idx": 32,
    "loss": 0.36288177967071533,
    "learning_rate": 0.000711467298542652,
    "timestep_mean": 401.5,
    "epsilon_norm": 7.757282733917236,
    "prediction_norm": 5.555115222930908,
    "timestamp": "2025-07-14T12:33:59.080673"
  },
  {
    "epoch": 722,
    "batch_idx": 64,
    "loss": 0.13777925074100494,
    "learning_rate": 0.000711467298542652,
    "timestep_mean": 589.59375,
    "epsilon_norm": 7.187222003936768,
    "prediction_norm": 7.0816121101379395,
    "timestamp": "2025-07-14T12:33:59.085812"
  },
  {
    "epoch": 722,
    "batch_idx": 96,
    "loss": 0.49960893392562866,
    "learning_rate": 0.000711467298542652,
    "timestep_mean": 398.4375,
    "epsilon_norm": 7.818810939788818,
    "prediction_norm": 5.449707508087158,
    "timestamp": "2025-07-14T12:33:59.090950"
  },
  {
    "epoch": 722,
    "batch_idx": 128,
    "loss": 0.49949538707733154,
    "learning_rate": 0.000711467298542652,
    "timestep_mean": 396.9375,
    "epsilon_norm": 9.205594062805176,
    "prediction_norm": 7.416499614715576,
    "timestamp": "2025-07-14T12:33:59.095923"
  },
  {
    "epoch": 722,
    "batch_idx": 160,
    "loss": 0.1853509545326233,
    "learning_rate": 0.000711467298542652,
    "timestep_mean": 503.875,
    "epsilon_norm": 8.056251525878906,
    "prediction_norm": 7.052068710327148,
    "timestamp": "2025-07-14T12:33:59.100810"
  },
  {
    "epoch": 722,
    "batch_idx": 192,
    "loss": 0.27789318561553955,
    "learning_rate": 0.000711467298542652,
    "timestep_mean": 689.25,
    "epsilon_norm": 4.235372066497803,
    "prediction_norm": 3.619319438934326,
    "timestamp": "2025-07-14T12:33:59.110980"
  },
  {
    "epoch": 723,
    "batch_idx": 0,
    "loss": 0.16707202792167664,
    "learning_rate": 0.0007107553413832048,
    "timestep_mean": 540.90625,
    "epsilon_norm": 7.028529167175293,
    "prediction_norm": 6.287486553192139,
    "timestamp": "2025-07-14T12:33:59.116315"
  },
  {
    "epoch": 723,
    "batch_idx": 32,
    "loss": 0.13408634066581726,
    "learning_rate": 0.0007107553413832048,
    "timestep_mean": 545.6875,
    "epsilon_norm": 8.1674222946167,
    "prediction_norm": 7.91895866394043,
    "timestamp": "2025-07-14T12:33:59.121434"
  },
  {
    "epoch": 723,
    "batch_idx": 64,
    "loss": 0.6294342279434204,
    "learning_rate": 0.0007107553413832048,
    "timestep_mean": 489.59375,
    "epsilon_norm": 8.755121231079102,
    "prediction_norm": 6.182326316833496,
    "timestamp": "2025-07-14T12:33:59.126567"
  },
  {
    "epoch": 723,
    "batch_idx": 96,
    "loss": 0.29044219851493835,
    "learning_rate": 0.0007107553413832048,
    "timestep_mean": 441.90625,
    "epsilon_norm": 6.867800712585449,
    "prediction_norm": 5.3989152908325195,
    "timestamp": "2025-07-14T12:33:59.131796"
  },
  {
    "epoch": 723,
    "batch_idx": 128,
    "loss": 0.22007282078266144,
    "learning_rate": 0.0007107553413832048,
    "timestep_mean": 496.0625,
    "epsilon_norm": 7.149811267852783,
    "prediction_norm": 6.4641499519348145,
    "timestamp": "2025-07-14T12:33:59.137086"
  },
  {
    "epoch": 723,
    "batch_idx": 160,
    "loss": 0.39890754222869873,
    "learning_rate": 0.0007107553413832048,
    "timestep_mean": 540.6875,
    "epsilon_norm": 7.672801494598389,
    "prediction_norm": 5.91523551940918,
    "timestamp": "2025-07-14T12:33:59.142269"
  },
  {
    "epoch": 723,
    "batch_idx": 192,
    "loss": 0.07600704580545425,
    "learning_rate": 0.0007107553413832048,
    "timestep_mean": 571.375,
    "epsilon_norm": 4.072873592376709,
    "prediction_norm": 3.7552998065948486,
    "timestamp": "2025-07-14T12:33:59.147577"
  },
  {
    "epoch": 724,
    "batch_idx": 0,
    "loss": 0.12617811560630798,
    "learning_rate": 0.0007100428642059033,
    "timestep_mean": 609.84375,
    "epsilon_norm": 7.328172206878662,
    "prediction_norm": 6.819637298583984,
    "timestamp": "2025-07-14T12:33:59.152732"
  },
  {
    "epoch": 724,
    "batch_idx": 32,
    "loss": 0.25512510538101196,
    "learning_rate": 0.0007100428642059033,
    "timestep_mean": 543.78125,
    "epsilon_norm": 8.108868598937988,
    "prediction_norm": 6.513264179229736,
    "timestamp": "2025-07-14T12:33:59.157856"
  },
  {
    "epoch": 724,
    "batch_idx": 64,
    "loss": 0.2474842071533203,
    "learning_rate": 0.0007100428642059033,
    "timestep_mean": 579.78125,
    "epsilon_norm": 8.99514102935791,
    "prediction_norm": 7.719152450561523,
    "timestamp": "2025-07-14T12:33:59.163103"
  },
  {
    "epoch": 724,
    "batch_idx": 96,
    "loss": 0.2894447445869446,
    "learning_rate": 0.0007100428642059033,
    "timestep_mean": 453.78125,
    "epsilon_norm": 7.600831508636475,
    "prediction_norm": 6.641906261444092,
    "timestamp": "2025-07-14T12:33:59.168339"
  },
  {
    "epoch": 724,
    "batch_idx": 128,
    "loss": 0.3164448142051697,
    "learning_rate": 0.0007100428642059033,
    "timestep_mean": 515.28125,
    "epsilon_norm": 8.566035270690918,
    "prediction_norm": 7.096412181854248,
    "timestamp": "2025-07-14T12:33:59.173460"
  },
  {
    "epoch": 724,
    "batch_idx": 160,
    "loss": 0.4264855682849884,
    "learning_rate": 0.0007100428642059033,
    "timestep_mean": 430.5625,
    "epsilon_norm": 7.974578857421875,
    "prediction_norm": 5.792981147766113,
    "timestamp": "2025-07-14T12:33:59.178654"
  },
  {
    "epoch": 724,
    "batch_idx": 192,
    "loss": 0.9526410698890686,
    "learning_rate": 0.0007100428642059033,
    "timestep_mean": 206.375,
    "epsilon_norm": 3.9108850955963135,
    "prediction_norm": 2.2679898738861084,
    "timestamp": "2025-07-14T12:33:59.183981"
  },
  {
    "epoch": 725,
    "batch_idx": 0,
    "loss": 0.3273511230945587,
    "learning_rate": 0.0007093298687687143,
    "timestep_mean": 428.46875,
    "epsilon_norm": 7.480594158172607,
    "prediction_norm": 5.799675464630127,
    "timestamp": "2025-07-14T12:33:59.189111"
  },
  {
    "epoch": 725,
    "batch_idx": 32,
    "loss": 0.3084166347980499,
    "learning_rate": 0.0007093298687687143,
    "timestep_mean": 469.84375,
    "epsilon_norm": 7.066646575927734,
    "prediction_norm": 5.279489994049072,
    "timestamp": "2025-07-14T12:33:59.194050"
  },
  {
    "epoch": 725,
    "batch_idx": 64,
    "loss": 0.587490439414978,
    "learning_rate": 0.0007093298687687143,
    "timestep_mean": 418.3125,
    "epsilon_norm": 8.5175199508667,
    "prediction_norm": 5.505988597869873,
    "timestamp": "2025-07-14T12:33:59.198834"
  },
  {
    "epoch": 725,
    "batch_idx": 96,
    "loss": 0.3382597863674164,
    "learning_rate": 0.0007093298687687143,
    "timestep_mean": 521.5,
    "epsilon_norm": 8.86902904510498,
    "prediction_norm": 7.485049247741699,
    "timestamp": "2025-07-14T12:33:59.203682"
  },
  {
    "epoch": 725,
    "batch_idx": 128,
    "loss": 0.2662004828453064,
    "learning_rate": 0.0007093298687687143,
    "timestep_mean": 367.625,
    "epsilon_norm": 7.4496331214904785,
    "prediction_norm": 6.623532295227051,
    "timestamp": "2025-07-14T12:33:59.208649"
  },
  {
    "epoch": 725,
    "batch_idx": 160,
    "loss": 0.23127105832099915,
    "learning_rate": 0.0007093298687687143,
    "timestep_mean": 569.65625,
    "epsilon_norm": 7.256282329559326,
    "prediction_norm": 6.1784749031066895,
    "timestamp": "2025-07-14T12:33:59.213546"
  },
  {
    "epoch": 725,
    "batch_idx": 192,
    "loss": 0.20475320518016815,
    "learning_rate": 0.0007093298687687143,
    "timestep_mean": 500.625,
    "epsilon_norm": 3.076662063598633,
    "prediction_norm": 2.371441125869751,
    "timestamp": "2025-07-14T12:33:59.218390"
  },
  {
    "epoch": 726,
    "batch_idx": 0,
    "loss": 0.41668564081192017,
    "learning_rate": 0.0007086163568308829,
    "timestep_mean": 520.46875,
    "epsilon_norm": 7.668398857116699,
    "prediction_norm": 5.910931587219238,
    "timestamp": "2025-07-14T12:33:59.223267"
  },
  {
    "epoch": 726,
    "batch_idx": 32,
    "loss": 0.43063127994537354,
    "learning_rate": 0.0007086163568308829,
    "timestep_mean": 559.0625,
    "epsilon_norm": 7.939570903778076,
    "prediction_norm": 6.061634540557861,
    "timestamp": "2025-07-14T12:33:59.228121"
  },
  {
    "epoch": 726,
    "batch_idx": 64,
    "loss": 0.37108081579208374,
    "learning_rate": 0.0007086163568308829,
    "timestep_mean": 480.96875,
    "epsilon_norm": 7.172943592071533,
    "prediction_norm": 5.296690940856934,
    "timestamp": "2025-07-14T12:33:59.232984"
  },
  {
    "epoch": 726,
    "batch_idx": 96,
    "loss": 0.18285106122493744,
    "learning_rate": 0.0007086163568308829,
    "timestep_mean": 532.0,
    "epsilon_norm": 6.647216796875,
    "prediction_norm": 5.628781318664551,
    "timestamp": "2025-07-14T12:33:59.238058"
  },
  {
    "epoch": 726,
    "batch_idx": 128,
    "loss": 0.27159708738327026,
    "learning_rate": 0.0007086163568308829,
    "timestep_mean": 528.5,
    "epsilon_norm": 6.936359405517578,
    "prediction_norm": 5.2743635177612305,
    "timestamp": "2025-07-14T12:33:59.242863"
  },
  {
    "epoch": 726,
    "batch_idx": 160,
    "loss": 0.18705153465270996,
    "learning_rate": 0.0007086163568308829,
    "timestep_mean": 580.28125,
    "epsilon_norm": 7.6471991539001465,
    "prediction_norm": 6.584229946136475,
    "timestamp": "2025-07-14T12:33:59.247694"
  },
  {
    "epoch": 726,
    "batch_idx": 192,
    "loss": 0.25832653045654297,
    "learning_rate": 0.0007086163568308829,
    "timestep_mean": 489.125,
    "epsilon_norm": 3.7164595127105713,
    "prediction_norm": 3.399958372116089,
    "timestamp": "2025-07-14T12:33:59.252520"
  },
  {
    "epoch": 727,
    "batch_idx": 0,
    "loss": 0.20896467566490173,
    "learning_rate": 0.000707902330152929,
    "timestep_mean": 485.65625,
    "epsilon_norm": 7.788443088531494,
    "prediction_norm": 6.018687725067139,
    "timestamp": "2025-07-14T12:33:59.257412"
  },
  {
    "epoch": 727,
    "batch_idx": 32,
    "loss": 0.16922509670257568,
    "learning_rate": 0.000707902330152929,
    "timestep_mean": 507.5,
    "epsilon_norm": 6.05771541595459,
    "prediction_norm": 5.431266784667969,
    "timestamp": "2025-07-14T12:33:59.262377"
  },
  {
    "epoch": 727,
    "batch_idx": 64,
    "loss": 0.20427632331848145,
    "learning_rate": 0.000707902330152929,
    "timestep_mean": 572.375,
    "epsilon_norm": 7.542120456695557,
    "prediction_norm": 7.526307106018066,
    "timestamp": "2025-07-14T12:33:59.267326"
  },
  {
    "epoch": 727,
    "batch_idx": 96,
    "loss": 0.27944621443748474,
    "learning_rate": 0.000707902330152929,
    "timestep_mean": 473.96875,
    "epsilon_norm": 7.740318775177002,
    "prediction_norm": 6.943536758422852,
    "timestamp": "2025-07-14T12:33:59.272176"
  },
  {
    "epoch": 727,
    "batch_idx": 128,
    "loss": 0.2391941100358963,
    "learning_rate": 0.000707902330152929,
    "timestep_mean": 505.125,
    "epsilon_norm": 7.317426681518555,
    "prediction_norm": 7.3284220695495605,
    "timestamp": "2025-07-14T12:33:59.277066"
  },
  {
    "epoch": 727,
    "batch_idx": 160,
    "loss": 0.28688305616378784,
    "learning_rate": 0.000707902330152929,
    "timestep_mean": 509.9375,
    "epsilon_norm": 8.46591854095459,
    "prediction_norm": 7.460269927978516,
    "timestamp": "2025-07-14T12:33:59.281981"
  },
  {
    "epoch": 727,
    "batch_idx": 192,
    "loss": 0.22551949322223663,
    "learning_rate": 0.000707902330152929,
    "timestep_mean": 425.0,
    "epsilon_norm": 3.740173816680908,
    "prediction_norm": 3.54126238822937,
    "timestamp": "2025-07-14T12:33:59.286838"
  },
  {
    "epoch": 728,
    "batch_idx": 0,
    "loss": 0.3381269872188568,
    "learning_rate": 0.0007071877904966425,
    "timestep_mean": 448.40625,
    "epsilon_norm": 7.566246032714844,
    "prediction_norm": 5.356151580810547,
    "timestamp": "2025-07-14T12:33:59.291782"
  },
  {
    "epoch": 728,
    "batch_idx": 32,
    "loss": 0.276786208152771,
    "learning_rate": 0.0007071877904966425,
    "timestep_mean": 430.75,
    "epsilon_norm": 8.964849472045898,
    "prediction_norm": 7.370620250701904,
    "timestamp": "2025-07-14T12:33:59.296622"
  },
  {
    "epoch": 728,
    "batch_idx": 64,
    "loss": 0.16995248198509216,
    "learning_rate": 0.0007071877904966425,
    "timestep_mean": 461.84375,
    "epsilon_norm": 7.435530662536621,
    "prediction_norm": 6.773738861083984,
    "timestamp": "2025-07-14T12:33:59.301468"
  },
  {
    "epoch": 728,
    "batch_idx": 96,
    "loss": 0.4308488368988037,
    "learning_rate": 0.0007071877904966425,
    "timestep_mean": 450.40625,
    "epsilon_norm": 8.993744850158691,
    "prediction_norm": 6.918970108032227,
    "timestamp": "2025-07-14T12:33:59.306315"
  },
  {
    "epoch": 728,
    "batch_idx": 128,
    "loss": 0.2695837616920471,
    "learning_rate": 0.0007071877904966425,
    "timestep_mean": 570.9375,
    "epsilon_norm": 8.385825157165527,
    "prediction_norm": 6.460786819458008,
    "timestamp": "2025-07-14T12:33:59.311234"
  },
  {
    "epoch": 728,
    "batch_idx": 160,
    "loss": 0.31163063645362854,
    "learning_rate": 0.0007071877904966425,
    "timestep_mean": 470.84375,
    "epsilon_norm": 7.69844388961792,
    "prediction_norm": 5.651548862457275,
    "timestamp": "2025-07-14T12:33:59.316156"
  },
  {
    "epoch": 728,
    "batch_idx": 192,
    "loss": 0.6210423707962036,
    "learning_rate": 0.0007071877904966425,
    "timestep_mean": 336.0,
    "epsilon_norm": 5.36527681350708,
    "prediction_norm": 3.740997791290283,
    "timestamp": "2025-07-14T12:33:59.321174"
  },
  {
    "epoch": 729,
    "batch_idx": 0,
    "loss": 0.19434717297554016,
    "learning_rate": 0.0007064727396250787,
    "timestep_mean": 475.03125,
    "epsilon_norm": 7.849376678466797,
    "prediction_norm": 6.649007320404053,
    "timestamp": "2025-07-14T12:33:59.325988"
  },
  {
    "epoch": 729,
    "batch_idx": 32,
    "loss": 0.4234180748462677,
    "learning_rate": 0.0007064727396250787,
    "timestep_mean": 406.9375,
    "epsilon_norm": 8.812946319580078,
    "prediction_norm": 6.7743120193481445,
    "timestamp": "2025-07-14T12:33:59.330800"
  },
  {
    "epoch": 729,
    "batch_idx": 64,
    "loss": 0.4193167984485626,
    "learning_rate": 0.0007064727396250787,
    "timestep_mean": 595.6875,
    "epsilon_norm": 8.765959739685059,
    "prediction_norm": 6.782097816467285,
    "timestamp": "2025-07-14T12:33:59.335647"
  },
  {
    "epoch": 729,
    "batch_idx": 96,
    "loss": 0.20012930035591125,
    "learning_rate": 0.0007064727396250787,
    "timestep_mean": 542.28125,
    "epsilon_norm": 7.9568586349487305,
    "prediction_norm": 6.766480445861816,
    "timestamp": "2025-07-14T12:33:59.340479"
  },
  {
    "epoch": 729,
    "batch_idx": 128,
    "loss": 0.5683116912841797,
    "learning_rate": 0.0007064727396250787,
    "timestep_mean": 445.5,
    "epsilon_norm": 7.600695610046387,
    "prediction_norm": 5.952857971191406,
    "timestamp": "2025-07-14T12:33:59.345338"
  },
  {
    "epoch": 729,
    "batch_idx": 160,
    "loss": 0.24800889194011688,
    "learning_rate": 0.0007064727396250787,
    "timestep_mean": 507.1875,
    "epsilon_norm": 8.247426986694336,
    "prediction_norm": 7.727251052856445,
    "timestamp": "2025-07-14T12:33:59.350273"
  },
  {
    "epoch": 729,
    "batch_idx": 192,
    "loss": 0.14222902059555054,
    "learning_rate": 0.0007064727396250787,
    "timestep_mean": 527.0,
    "epsilon_norm": 3.835926055908203,
    "prediction_norm": 3.4810173511505127,
    "timestamp": "2025-07-14T12:33:59.355075"
  },
  {
    "epoch": 730,
    "batch_idx": 0,
    "loss": 0.24323627352714539,
    "learning_rate": 0.0007057571793025547,
    "timestep_mean": 565.71875,
    "epsilon_norm": 8.066305160522461,
    "prediction_norm": 7.4092864990234375,
    "timestamp": "2025-07-14T12:33:59.360017"
  },
  {
    "epoch": 730,
    "batch_idx": 32,
    "loss": 0.3353845179080963,
    "learning_rate": 0.0007057571793025547,
    "timestep_mean": 478.875,
    "epsilon_norm": 8.131171226501465,
    "prediction_norm": 6.830340385437012,
    "timestamp": "2025-07-14T12:33:59.364922"
  },
  {
    "epoch": 730,
    "batch_idx": 64,
    "loss": 0.1880088746547699,
    "learning_rate": 0.0007057571793025547,
    "timestep_mean": 553.0625,
    "epsilon_norm": 8.063501358032227,
    "prediction_norm": 7.1342549324035645,
    "timestamp": "2025-07-14T12:33:59.369921"
  },
  {
    "epoch": 730,
    "batch_idx": 96,
    "loss": 0.2910200357437134,
    "learning_rate": 0.0007057571793025547,
    "timestep_mean": 430.6875,
    "epsilon_norm": 7.3136677742004395,
    "prediction_norm": 6.284046173095703,
    "timestamp": "2025-07-14T12:33:59.374880"
  },
  {
    "epoch": 730,
    "batch_idx": 128,
    "loss": 0.4076913595199585,
    "learning_rate": 0.0007057571793025547,
    "timestep_mean": 534.125,
    "epsilon_norm": 7.739346027374268,
    "prediction_norm": 6.527527809143066,
    "timestamp": "2025-07-14T12:33:59.379796"
  },
  {
    "epoch": 730,
    "batch_idx": 160,
    "loss": 0.40585243701934814,
    "learning_rate": 0.0007057571793025547,
    "timestep_mean": 459.15625,
    "epsilon_norm": 7.67803430557251,
    "prediction_norm": 5.636929988861084,
    "timestamp": "2025-07-14T12:33:59.384599"
  },
  {
    "epoch": 730,
    "batch_idx": 192,
    "loss": 0.06344452500343323,
    "learning_rate": 0.0007057571793025547,
    "timestep_mean": 594.625,
    "epsilon_norm": 3.6029715538024902,
    "prediction_norm": 3.360807180404663,
    "timestamp": "2025-07-14T12:33:59.389395"
  },
  {
    "epoch": 731,
    "batch_idx": 0,
    "loss": 0.40397128462791443,
    "learning_rate": 0.0007050411112946446,
    "timestep_mean": 460.96875,
    "epsilon_norm": 8.86995792388916,
    "prediction_norm": 7.207489013671875,
    "timestamp": "2025-07-14T12:33:59.394268"
  },
  {
    "epoch": 731,
    "batch_idx": 32,
    "loss": 0.4658081531524658,
    "learning_rate": 0.0007050411112946446,
    "timestep_mean": 488.0,
    "epsilon_norm": 8.001577377319336,
    "prediction_norm": 7.031224727630615,
    "timestamp": "2025-07-14T12:33:59.399141"
  },
  {
    "epoch": 731,
    "batch_idx": 64,
    "loss": 0.20156043767929077,
    "learning_rate": 0.0007050411112946446,
    "timestep_mean": 509.84375,
    "epsilon_norm": 7.454356670379639,
    "prediction_norm": 7.1705803871154785,
    "timestamp": "2025-07-14T12:33:59.404096"
  },
  {
    "epoch": 731,
    "batch_idx": 96,
    "loss": 0.31344491243362427,
    "learning_rate": 0.0007050411112946446,
    "timestep_mean": 461.71875,
    "epsilon_norm": 6.896486759185791,
    "prediction_norm": 6.091495513916016,
    "timestamp": "2025-07-14T12:33:59.409032"
  },
  {
    "epoch": 731,
    "batch_idx": 128,
    "loss": 0.1666417419910431,
    "learning_rate": 0.0007050411112946446,
    "timestep_mean": 525.875,
    "epsilon_norm": 7.448153972625732,
    "prediction_norm": 6.722683429718018,
    "timestamp": "2025-07-14T12:33:59.413865"
  },
  {
    "epoch": 731,
    "batch_idx": 160,
    "loss": 0.21675068140029907,
    "learning_rate": 0.0007050411112946446,
    "timestep_mean": 538.28125,
    "epsilon_norm": 7.574352741241455,
    "prediction_norm": 6.98933219909668,
    "timestamp": "2025-07-14T12:33:59.418897"
  },
  {
    "epoch": 731,
    "batch_idx": 192,
    "loss": 0.05955533683300018,
    "learning_rate": 0.0007050411112946446,
    "timestep_mean": 635.625,
    "epsilon_norm": 3.415618658065796,
    "prediction_norm": 3.1581332683563232,
    "timestamp": "2025-07-14T12:33:59.424041"
  },
  {
    "epoch": 732,
    "batch_idx": 0,
    "loss": 0.24480277299880981,
    "learning_rate": 0.000704324537368175,
    "timestep_mean": 490.6875,
    "epsilon_norm": 7.155170917510986,
    "prediction_norm": 6.3258514404296875,
    "timestamp": "2025-07-14T12:33:59.429333"
  },
  {
    "epoch": 732,
    "batch_idx": 32,
    "loss": 0.15280061960220337,
    "learning_rate": 0.000704324537368175,
    "timestep_mean": 500.75,
    "epsilon_norm": 7.479772567749023,
    "prediction_norm": 6.749338626861572,
    "timestamp": "2025-07-14T12:33:59.434611"
  },
  {
    "epoch": 732,
    "batch_idx": 64,
    "loss": 0.2971615195274353,
    "learning_rate": 0.000704324537368175,
    "timestep_mean": 457.5625,
    "epsilon_norm": 7.697431564331055,
    "prediction_norm": 5.899953842163086,
    "timestamp": "2025-07-14T12:33:59.439772"
  },
  {
    "epoch": 732,
    "batch_idx": 96,
    "loss": 0.15991650521755219,
    "learning_rate": 0.000704324537368175,
    "timestep_mean": 459.0625,
    "epsilon_norm": 7.404354095458984,
    "prediction_norm": 6.242382526397705,
    "timestamp": "2025-07-14T12:33:59.444971"
  },
  {
    "epoch": 732,
    "batch_idx": 128,
    "loss": 0.38825270533561707,
    "learning_rate": 0.000704324537368175,
    "timestep_mean": 459.71875,
    "epsilon_norm": 7.821486473083496,
    "prediction_norm": 6.1534600257873535,
    "timestamp": "2025-07-14T12:33:59.450321"
  },
  {
    "epoch": 732,
    "batch_idx": 160,
    "loss": 0.19818148016929626,
    "learning_rate": 0.000704324537368175,
    "timestep_mean": 479.03125,
    "epsilon_norm": 6.6924028396606445,
    "prediction_norm": 5.697198390960693,
    "timestamp": "2025-07-14T12:33:59.455484"
  },
  {
    "epoch": 732,
    "batch_idx": 192,
    "loss": 0.5370355844497681,
    "learning_rate": 0.000704324537368175,
    "timestep_mean": 404.625,
    "epsilon_norm": 4.347012996673584,
    "prediction_norm": 3.188265085220337,
    "timestamp": "2025-07-14T12:33:59.460677"
  },
  {
    "epoch": 733,
    "batch_idx": 0,
    "loss": 0.26547732949256897,
    "learning_rate": 0.0007036074592912207,
    "timestep_mean": 452.125,
    "epsilon_norm": 7.142210960388184,
    "prediction_norm": 6.046498775482178,
    "timestamp": "2025-07-14T12:33:59.465816"
  },
  {
    "epoch": 733,
    "batch_idx": 32,
    "loss": 0.2671314477920532,
    "learning_rate": 0.0007036074592912207,
    "timestep_mean": 482.75,
    "epsilon_norm": 7.40390157699585,
    "prediction_norm": 6.401738166809082,
    "timestamp": "2025-07-14T12:33:59.470974"
  },
  {
    "epoch": 733,
    "batch_idx": 64,
    "loss": 0.2348472774028778,
    "learning_rate": 0.0007036074592912207,
    "timestep_mean": 556.21875,
    "epsilon_norm": 7.533138751983643,
    "prediction_norm": 6.800741195678711,
    "timestamp": "2025-07-14T12:33:59.476106"
  },
  {
    "epoch": 733,
    "batch_idx": 96,
    "loss": 0.23881447315216064,
    "learning_rate": 0.0007036074592912207,
    "timestep_mean": 489.53125,
    "epsilon_norm": 8.075435638427734,
    "prediction_norm": 6.800831317901611,
    "timestamp": "2025-07-14T12:33:59.481328"
  },
  {
    "epoch": 733,
    "batch_idx": 128,
    "loss": 0.2986215353012085,
    "learning_rate": 0.0007036074592912207,
    "timestep_mean": 485.46875,
    "epsilon_norm": 8.226881980895996,
    "prediction_norm": 7.315464973449707,
    "timestamp": "2025-07-14T12:33:59.486544"
  },
  {
    "epoch": 733,
    "batch_idx": 160,
    "loss": 0.2616995573043823,
    "learning_rate": 0.0007036074592912207,
    "timestep_mean": 484.65625,
    "epsilon_norm": 8.518718719482422,
    "prediction_norm": 7.991205215454102,
    "timestamp": "2025-07-14T12:33:59.491867"
  },
  {
    "epoch": 733,
    "batch_idx": 192,
    "loss": 0.12465405464172363,
    "learning_rate": 0.0007036074592912207,
    "timestep_mean": 534.0,
    "epsilon_norm": 3.5619094371795654,
    "prediction_norm": 3.5849883556365967,
    "timestamp": "2025-07-14T12:33:59.497014"
  },
  {
    "epoch": 734,
    "batch_idx": 0,
    "loss": 0.18261851370334625,
    "learning_rate": 0.0007028898788331004,
    "timestep_mean": 541.5,
    "epsilon_norm": 8.456212997436523,
    "prediction_norm": 6.991888999938965,
    "timestamp": "2025-07-14T12:33:59.502037"
  },
  {
    "epoch": 734,
    "batch_idx": 32,
    "loss": 0.3805871903896332,
    "learning_rate": 0.0007028898788331004,
    "timestep_mean": 453.375,
    "epsilon_norm": 8.479592323303223,
    "prediction_norm": 7.142617702484131,
    "timestamp": "2025-07-14T12:33:59.507277"
  },
  {
    "epoch": 734,
    "batch_idx": 64,
    "loss": 0.33176007866859436,
    "learning_rate": 0.0007028898788331004,
    "timestep_mean": 447.90625,
    "epsilon_norm": 8.601987838745117,
    "prediction_norm": 6.959305286407471,
    "timestamp": "2025-07-14T12:33:59.512491"
  },
  {
    "epoch": 734,
    "batch_idx": 96,
    "loss": 0.17278967797756195,
    "learning_rate": 0.0007028898788331004,
    "timestep_mean": 543.59375,
    "epsilon_norm": 8.67889404296875,
    "prediction_norm": 7.687067985534668,
    "timestamp": "2025-07-14T12:33:59.517723"
  },
  {
    "epoch": 734,
    "batch_idx": 128,
    "loss": 0.14070113003253937,
    "learning_rate": 0.0007028898788331004,
    "timestep_mean": 576.125,
    "epsilon_norm": 7.375453948974609,
    "prediction_norm": 6.695241928100586,
    "timestamp": "2025-07-14T12:33:59.522810"
  },
  {
    "epoch": 734,
    "batch_idx": 160,
    "loss": 0.614294171333313,
    "learning_rate": 0.0007028898788331004,
    "timestep_mean": 419.71875,
    "epsilon_norm": 7.758327484130859,
    "prediction_norm": 5.23914098739624,
    "timestamp": "2025-07-14T12:33:59.527971"
  },
  {
    "epoch": 734,
    "batch_idx": 192,
    "loss": 0.1758173555135727,
    "learning_rate": 0.0007028898788331004,
    "timestep_mean": 661.25,
    "epsilon_norm": 4.654642105102539,
    "prediction_norm": 3.855825662612915,
    "timestamp": "2025-07-14T12:33:59.533406"
  },
  {
    "epoch": 735,
    "batch_idx": 0,
    "loss": 0.22339937090873718,
    "learning_rate": 0.000702171797764373,
    "timestep_mean": 513.40625,
    "epsilon_norm": 7.934299945831299,
    "prediction_norm": 6.460108280181885,
    "timestamp": "2025-07-14T12:33:59.538713"
  },
  {
    "epoch": 735,
    "batch_idx": 32,
    "loss": 0.329458087682724,
    "learning_rate": 0.000702171797764373,
    "timestep_mean": 537.59375,
    "epsilon_norm": 9.132698059082031,
    "prediction_norm": 7.1406025886535645,
    "timestamp": "2025-07-14T12:33:59.543846"
  },
  {
    "epoch": 735,
    "batch_idx": 64,
    "loss": 0.2849527895450592,
    "learning_rate": 0.000702171797764373,
    "timestep_mean": 522.15625,
    "epsilon_norm": 7.7070207595825195,
    "prediction_norm": 6.135343074798584,
    "timestamp": "2025-07-14T12:33:59.548938"
  },
  {
    "epoch": 735,
    "batch_idx": 96,
    "loss": 0.35185831785202026,
    "learning_rate": 0.000702171797764373,
    "timestep_mean": 561.53125,
    "epsilon_norm": 8.2840576171875,
    "prediction_norm": 6.383473873138428,
    "timestamp": "2025-07-14T12:33:59.554110"
  },
  {
    "epoch": 735,
    "batch_idx": 128,
    "loss": 0.17825651168823242,
    "learning_rate": 0.000702171797764373,
    "timestep_mean": 493.375,
    "epsilon_norm": 7.572638034820557,
    "prediction_norm": 7.047796726226807,
    "timestamp": "2025-07-14T12:33:59.559247"
  },
  {
    "epoch": 735,
    "batch_idx": 160,
    "loss": 0.35658693313598633,
    "learning_rate": 0.000702171797764373,
    "timestep_mean": 500.0625,
    "epsilon_norm": 8.000983238220215,
    "prediction_norm": 6.280588150024414,
    "timestamp": "2025-07-14T12:33:59.564388"
  },
  {
    "epoch": 735,
    "batch_idx": 192,
    "loss": 0.10469207167625427,
    "learning_rate": 0.000702171797764373,
    "timestep_mean": 559.375,
    "epsilon_norm": 4.011107921600342,
    "prediction_norm": 4.115179538726807,
    "timestamp": "2025-07-14T12:33:59.569568"
  },
  {
    "epoch": 736,
    "batch_idx": 0,
    "loss": 0.37250667810440063,
    "learning_rate": 0.0007014532178568319,
    "timestep_mean": 463.0,
    "epsilon_norm": 9.337756156921387,
    "prediction_norm": 7.344729423522949,
    "timestamp": "2025-07-14T12:33:59.574872"
  },
  {
    "epoch": 736,
    "batch_idx": 32,
    "loss": 0.21820634603500366,
    "learning_rate": 0.0007014532178568319,
    "timestep_mean": 591.40625,
    "epsilon_norm": 7.861833572387695,
    "prediction_norm": 7.145846843719482,
    "timestamp": "2025-07-14T12:33:59.579776"
  },
  {
    "epoch": 736,
    "batch_idx": 64,
    "loss": 0.23331904411315918,
    "learning_rate": 0.0007014532178568319,
    "timestep_mean": 506.71875,
    "epsilon_norm": 7.961989879608154,
    "prediction_norm": 6.906602382659912,
    "timestamp": "2025-07-14T12:33:59.584558"
  },
  {
    "epoch": 736,
    "batch_idx": 96,
    "loss": 0.284293532371521,
    "learning_rate": 0.0007014532178568319,
    "timestep_mean": 542.71875,
    "epsilon_norm": 7.408029556274414,
    "prediction_norm": 5.8773298263549805,
    "timestamp": "2025-07-14T12:33:59.589412"
  },
  {
    "epoch": 736,
    "batch_idx": 128,
    "loss": 0.4195294976234436,
    "learning_rate": 0.0007014532178568319,
    "timestep_mean": 469.78125,
    "epsilon_norm": 8.971753120422363,
    "prediction_norm": 7.231907844543457,
    "timestamp": "2025-07-14T12:33:59.594245"
  },
  {
    "epoch": 736,
    "batch_idx": 160,
    "loss": 0.24278944730758667,
    "learning_rate": 0.0007014532178568319,
    "timestep_mean": 516.9375,
    "epsilon_norm": 8.152462005615234,
    "prediction_norm": 7.5087175369262695,
    "timestamp": "2025-07-14T12:33:59.599067"
  },
  {
    "epoch": 736,
    "batch_idx": 192,
    "loss": 0.5085647106170654,
    "learning_rate": 0.0007014532178568319,
    "timestep_mean": 689.25,
    "epsilon_norm": 4.297138214111328,
    "prediction_norm": 3.5168943405151367,
    "timestamp": "2025-07-14T12:33:59.603924"
  },
  {
    "epoch": 737,
    "batch_idx": 0,
    "loss": 0.32048749923706055,
    "learning_rate": 0.0007007341408835014,
    "timestep_mean": 532.5,
    "epsilon_norm": 7.001840591430664,
    "prediction_norm": 6.3002471923828125,
    "timestamp": "2025-07-14T12:33:59.608830"
  },
  {
    "epoch": 737,
    "batch_idx": 32,
    "loss": 0.12210620939731598,
    "learning_rate": 0.0007007341408835014,
    "timestep_mean": 534.75,
    "epsilon_norm": 7.3000168800354,
    "prediction_norm": 6.379007339477539,
    "timestamp": "2025-07-14T12:33:59.613653"
  },
  {
    "epoch": 737,
    "batch_idx": 64,
    "loss": 0.364745169878006,
    "learning_rate": 0.0007007341408835014,
    "timestep_mean": 451.8125,
    "epsilon_norm": 8.372113227844238,
    "prediction_norm": 6.6112494468688965,
    "timestamp": "2025-07-14T12:33:59.618537"
  },
  {
    "epoch": 737,
    "batch_idx": 96,
    "loss": 0.3710223436355591,
    "learning_rate": 0.0007007341408835014,
    "timestep_mean": 439.0625,
    "epsilon_norm": 7.829428672790527,
    "prediction_norm": 6.245638847351074,
    "timestamp": "2025-07-14T12:33:59.623536"
  },
  {
    "epoch": 737,
    "batch_idx": 128,
    "loss": 0.6425491571426392,
    "learning_rate": 0.0007007341408835014,
    "timestep_mean": 483.3125,
    "epsilon_norm": 8.375768661499023,
    "prediction_norm": 6.131192684173584,
    "timestamp": "2025-07-14T12:33:59.628381"
  },
  {
    "epoch": 737,
    "batch_idx": 160,
    "loss": 0.2232755720615387,
    "learning_rate": 0.0007007341408835014,
    "timestep_mean": 531.0,
    "epsilon_norm": 8.159982681274414,
    "prediction_norm": 7.680403709411621,
    "timestamp": "2025-07-14T12:33:59.633342"
  },
  {
    "epoch": 737,
    "batch_idx": 192,
    "loss": 0.2913503646850586,
    "learning_rate": 0.0007007341408835014,
    "timestep_mean": 429.875,
    "epsilon_norm": 3.4303882122039795,
    "prediction_norm": 3.0308260917663574,
    "timestamp": "2025-07-14T12:33:59.638279"
  },
  {
    "epoch": 738,
    "batch_idx": 0,
    "loss": 0.15502770245075226,
    "learning_rate": 0.0007000145686186328,
    "timestep_mean": 594.0,
    "epsilon_norm": 8.44895076751709,
    "prediction_norm": 7.82278299331665,
    "timestamp": "2025-07-14T12:33:59.643184"
  },
  {
    "epoch": 738,
    "batch_idx": 32,
    "loss": 0.352760910987854,
    "learning_rate": 0.0007000145686186328,
    "timestep_mean": 534.59375,
    "epsilon_norm": 7.037411212921143,
    "prediction_norm": 5.712863445281982,
    "timestamp": "2025-07-14T12:33:59.647987"
  },
  {
    "epoch": 738,
    "batch_idx": 64,
    "loss": 0.3019944727420807,
    "learning_rate": 0.0007000145686186328,
    "timestep_mean": 485.21875,
    "epsilon_norm": 7.745538711547852,
    "prediction_norm": 6.0213942527771,
    "timestamp": "2025-07-14T12:33:59.652826"
  },
  {
    "epoch": 738,
    "batch_idx": 96,
    "loss": 0.1888134628534317,
    "learning_rate": 0.0007000145686186328,
    "timestep_mean": 522.0625,
    "epsilon_norm": 7.581334590911865,
    "prediction_norm": 6.430419445037842,
    "timestamp": "2025-07-14T12:33:59.657637"
  },
  {
    "epoch": 738,
    "batch_idx": 128,
    "loss": 0.18619300425052643,
    "learning_rate": 0.0007000145686186328,
    "timestep_mean": 550.78125,
    "epsilon_norm": 7.625493049621582,
    "prediction_norm": 6.875778675079346,
    "timestamp": "2025-07-14T12:33:59.662570"
  },
  {
    "epoch": 738,
    "batch_idx": 160,
    "loss": 0.16013380885124207,
    "learning_rate": 0.0007000145686186328,
    "timestep_mean": 533.40625,
    "epsilon_norm": 7.7273077964782715,
    "prediction_norm": 6.792153358459473,
    "timestamp": "2025-07-14T12:33:59.667409"
  },
  {
    "epoch": 738,
    "batch_idx": 192,
    "loss": 0.035013917833566666,
    "learning_rate": 0.0007000145686186328,
    "timestep_mean": 567.375,
    "epsilon_norm": 1.7986150979995728,
    "prediction_norm": 1.6844356060028076,
    "timestamp": "2025-07-14T12:33:59.672351"
  },
  {
    "epoch": 739,
    "batch_idx": 0,
    "loss": 0.21365238726139069,
    "learning_rate": 0.0006992945028376991,
    "timestep_mean": 492.78125,
    "epsilon_norm": 8.273883819580078,
    "prediction_norm": 7.835653305053711,
    "timestamp": "2025-07-14T12:33:59.677240"
  },
  {
    "epoch": 739,
    "batch_idx": 32,
    "loss": 0.439361035823822,
    "learning_rate": 0.0006992945028376991,
    "timestep_mean": 484.21875,
    "epsilon_norm": 7.421995162963867,
    "prediction_norm": 6.191606044769287,
    "timestamp": "2025-07-14T12:33:59.682106"
  },
  {
    "epoch": 739,
    "batch_idx": 64,
    "loss": 0.34706854820251465,
    "learning_rate": 0.0006992945028376991,
    "timestep_mean": 560.75,
    "epsilon_norm": 8.143729209899902,
    "prediction_norm": 6.616661548614502,
    "timestamp": "2025-07-14T12:33:59.687050"
  },
  {
    "epoch": 739,
    "batch_idx": 96,
    "loss": 0.4113733768463135,
    "learning_rate": 0.0006992945028376991,
    "timestep_mean": 385.71875,
    "epsilon_norm": 6.919094085693359,
    "prediction_norm": 5.127596378326416,
    "timestamp": "2025-07-14T12:33:59.692004"
  },
  {
    "epoch": 739,
    "batch_idx": 128,
    "loss": 0.25121912360191345,
    "learning_rate": 0.0006992945028376991,
    "timestep_mean": 577.09375,
    "epsilon_norm": 7.788191795349121,
    "prediction_norm": 6.454682350158691,
    "timestamp": "2025-07-14T12:33:59.696917"
  },
  {
    "epoch": 739,
    "batch_idx": 160,
    "loss": 0.2632474899291992,
    "learning_rate": 0.0006992945028376991,
    "timestep_mean": 489.59375,
    "epsilon_norm": 7.829163551330566,
    "prediction_norm": 6.890789031982422,
    "timestamp": "2025-07-14T12:33:59.706512"
  },
  {
    "epoch": 739,
    "batch_idx": 192,
    "loss": 0.32616108655929565,
    "learning_rate": 0.0006992945028376991,
    "timestep_mean": 449.625,
    "epsilon_norm": 2.7894551753997803,
    "prediction_norm": 2.4633004665374756,
    "timestamp": "2025-07-14T12:33:59.711476"
  },
  {
    "epoch": 740,
    "batch_idx": 0,
    "loss": 0.2392062097787857,
    "learning_rate": 0.0006985739453173908,
    "timestep_mean": 570.78125,
    "epsilon_norm": 6.994523048400879,
    "prediction_norm": 5.999776363372803,
    "timestamp": "2025-07-14T12:33:59.716365"
  },
  {
    "epoch": 740,
    "batch_idx": 32,
    "loss": 0.14194530248641968,
    "learning_rate": 0.0006985739453173908,
    "timestep_mean": 553.84375,
    "epsilon_norm": 7.965203762054443,
    "prediction_norm": 7.0148515701293945,
    "timestamp": "2025-07-14T12:33:59.721298"
  },
  {
    "epoch": 740,
    "batch_idx": 64,
    "loss": 0.26723235845565796,
    "learning_rate": 0.0006985739453173908,
    "timestep_mean": 501.375,
    "epsilon_norm": 8.86974811553955,
    "prediction_norm": 8.231979370117188,
    "timestamp": "2025-07-14T12:33:59.726126"
  },
  {
    "epoch": 740,
    "batch_idx": 96,
    "loss": 0.22099685668945312,
    "learning_rate": 0.0006985739453173908,
    "timestep_mean": 478.9375,
    "epsilon_norm": 6.920773029327393,
    "prediction_norm": 5.404926300048828,
    "timestamp": "2025-07-14T12:33:59.730978"
  },
  {
    "epoch": 740,
    "batch_idx": 128,
    "loss": 0.7698825597763062,
    "learning_rate": 0.0006985739453173908,
    "timestep_mean": 451.5,
    "epsilon_norm": 8.738604545593262,
    "prediction_norm": 6.179070949554443,
    "timestamp": "2025-07-14T12:33:59.735907"
  },
  {
    "epoch": 740,
    "batch_idx": 160,
    "loss": 0.32915985584259033,
    "learning_rate": 0.0006985739453173908,
    "timestep_mean": 456.0,
    "epsilon_norm": 9.099261283874512,
    "prediction_norm": 7.568121910095215,
    "timestamp": "2025-07-14T12:33:59.740844"
  },
  {
    "epoch": 740,
    "batch_idx": 192,
    "loss": 0.15298712253570557,
    "learning_rate": 0.0006985739453173908,
    "timestep_mean": 523.625,
    "epsilon_norm": 5.100596904754639,
    "prediction_norm": 4.953334808349609,
    "timestamp": "2025-07-14T12:33:59.745711"
  },
  {
    "epoch": 741,
    "batch_idx": 0,
    "loss": 0.16359767317771912,
    "learning_rate": 0.0006978528978356122,
    "timestep_mean": 519.625,
    "epsilon_norm": 7.7408576011657715,
    "prediction_norm": 6.801822662353516,
    "timestamp": "2025-07-14T12:33:59.750679"
  },
  {
    "epoch": 741,
    "batch_idx": 32,
    "loss": 0.40230557322502136,
    "learning_rate": 0.0006978528978356122,
    "timestep_mean": 490.34375,
    "epsilon_norm": 7.2301106452941895,
    "prediction_norm": 5.338983535766602,
    "timestamp": "2025-07-14T12:33:59.755596"
  },
  {
    "epoch": 741,
    "batch_idx": 64,
    "loss": 0.1800755262374878,
    "learning_rate": 0.0006978528978356122,
    "timestep_mean": 472.0,
    "epsilon_norm": 8.088545799255371,
    "prediction_norm": 6.732752799987793,
    "timestamp": "2025-07-14T12:33:59.760403"
  },
  {
    "epoch": 741,
    "batch_idx": 96,
    "loss": 0.27905717492103577,
    "learning_rate": 0.0006978528978356122,
    "timestep_mean": 450.15625,
    "epsilon_norm": 8.33420467376709,
    "prediction_norm": 7.034770488739014,
    "timestamp": "2025-07-14T12:33:59.765299"
  },
  {
    "epoch": 741,
    "batch_idx": 128,
    "loss": 0.3187044858932495,
    "learning_rate": 0.0006978528978356122,
    "timestep_mean": 473.84375,
    "epsilon_norm": 8.647778511047363,
    "prediction_norm": 6.811402797698975,
    "timestamp": "2025-07-14T12:33:59.770156"
  },
  {
    "epoch": 741,
    "batch_idx": 160,
    "loss": 0.2811838388442993,
    "learning_rate": 0.0006978528978356122,
    "timestep_mean": 500.15625,
    "epsilon_norm": 8.552783012390137,
    "prediction_norm": 6.9990339279174805,
    "timestamp": "2025-07-14T12:33:59.775102"
  },
  {
    "epoch": 741,
    "batch_idx": 192,
    "loss": 0.05261215567588806,
    "learning_rate": 0.0006978528978356122,
    "timestep_mean": 577.375,
    "epsilon_norm": 4.329008102416992,
    "prediction_norm": 3.969064235687256,
    "timestamp": "2025-07-14T12:33:59.780012"
  },
  {
    "epoch": 742,
    "batch_idx": 0,
    "loss": 0.30522626638412476,
    "learning_rate": 0.0006971313621714759,
    "timestep_mean": 419.9375,
    "epsilon_norm": 7.5155792236328125,
    "prediction_norm": 5.825839996337891,
    "timestamp": "2025-07-14T12:33:59.784941"
  },
  {
    "epoch": 742,
    "batch_idx": 32,
    "loss": 0.3197675049304962,
    "learning_rate": 0.0006971313621714759,
    "timestep_mean": 475.9375,
    "epsilon_norm": 7.364101886749268,
    "prediction_norm": 5.75106143951416,
    "timestamp": "2025-07-14T12:33:59.789863"
  },
  {
    "epoch": 742,
    "batch_idx": 64,
    "loss": 0.4482550024986267,
    "learning_rate": 0.0006971313621714759,
    "timestep_mean": 438.65625,
    "epsilon_norm": 9.13268756866455,
    "prediction_norm": 6.821033477783203,
    "timestamp": "2025-07-14T12:33:59.794751"
  },
  {
    "epoch": 742,
    "batch_idx": 96,
    "loss": 0.2822769284248352,
    "learning_rate": 0.0006971313621714759,
    "timestep_mean": 486.0,
    "epsilon_norm": 6.701696872711182,
    "prediction_norm": 5.431859493255615,
    "timestamp": "2025-07-14T12:33:59.799610"
  },
  {
    "epoch": 742,
    "batch_idx": 128,
    "loss": 0.3357696533203125,
    "learning_rate": 0.0006971313621714759,
    "timestep_mean": 512.53125,
    "epsilon_norm": 8.217985153198242,
    "prediction_norm": 6.2851243019104,
    "timestamp": "2025-07-14T12:33:59.804584"
  },
  {
    "epoch": 742,
    "batch_idx": 160,
    "loss": 0.2394225150346756,
    "learning_rate": 0.0006971313621714759,
    "timestep_mean": 470.5,
    "epsilon_norm": 6.708562850952148,
    "prediction_norm": 5.3452606201171875,
    "timestamp": "2025-07-14T12:33:59.809392"
  },
  {
    "epoch": 742,
    "batch_idx": 192,
    "loss": 0.027451813220977783,
    "learning_rate": 0.0006971313621714759,
    "timestep_mean": 722.25,
    "epsilon_norm": 3.3378586769104004,
    "prediction_norm": 3.3459360599517822,
    "timestamp": "2025-07-14T12:33:59.814228"
  },
  {
    "epoch": 743,
    "batch_idx": 0,
    "loss": 0.25030517578125,
    "learning_rate": 0.0006964093401053,
    "timestep_mean": 477.84375,
    "epsilon_norm": 7.231486797332764,
    "prediction_norm": 6.234717845916748,
    "timestamp": "2025-07-14T12:33:59.819147"
  },
  {
    "epoch": 743,
    "batch_idx": 32,
    "loss": 0.15785565972328186,
    "learning_rate": 0.0006964093401053,
    "timestep_mean": 586.15625,
    "epsilon_norm": 7.3414154052734375,
    "prediction_norm": 6.842960834503174,
    "timestamp": "2025-07-14T12:33:59.824034"
  },
  {
    "epoch": 743,
    "batch_idx": 64,
    "loss": 0.1540481150150299,
    "learning_rate": 0.0006964093401053,
    "timestep_mean": 575.1875,
    "epsilon_norm": 8.626209259033203,
    "prediction_norm": 8.023711204528809,
    "timestamp": "2025-07-14T12:33:59.828967"
  },
  {
    "epoch": 743,
    "batch_idx": 96,
    "loss": 0.2076120674610138,
    "learning_rate": 0.0006964093401053,
    "timestep_mean": 496.4375,
    "epsilon_norm": 8.686685562133789,
    "prediction_norm": 8.111474990844727,
    "timestamp": "2025-07-14T12:33:59.833859"
  },
  {
    "epoch": 743,
    "batch_idx": 128,
    "loss": 0.2139953076839447,
    "learning_rate": 0.0006964093401053,
    "timestep_mean": 564.84375,
    "epsilon_norm": 7.967579364776611,
    "prediction_norm": 6.875943183898926,
    "timestamp": "2025-07-14T12:33:59.838797"
  },
  {
    "epoch": 743,
    "batch_idx": 160,
    "loss": 0.32617172598838806,
    "learning_rate": 0.0006964093401053,
    "timestep_mean": 511.0625,
    "epsilon_norm": 7.4861578941345215,
    "prediction_norm": 6.115450382232666,
    "timestamp": "2025-07-14T12:33:59.843674"
  },
  {
    "epoch": 743,
    "batch_idx": 192,
    "loss": 0.8610376119613647,
    "learning_rate": 0.0006964093401053,
    "timestep_mean": 341.375,
    "epsilon_norm": 4.866621971130371,
    "prediction_norm": 2.673710823059082,
    "timestamp": "2025-07-14T12:33:59.848516"
  },
  {
    "epoch": 744,
    "batch_idx": 0,
    "loss": 0.29642003774642944,
    "learning_rate": 0.0006956868334186018,
    "timestep_mean": 524.4375,
    "epsilon_norm": 7.529481887817383,
    "prediction_norm": 6.570550441741943,
    "timestamp": "2025-07-14T12:33:59.853383"
  },
  {
    "epoch": 744,
    "batch_idx": 32,
    "loss": 0.36164242029190063,
    "learning_rate": 0.0006956868334186018,
    "timestep_mean": 434.4375,
    "epsilon_norm": 9.20988655090332,
    "prediction_norm": 8.161225318908691,
    "timestamp": "2025-07-14T12:33:59.858384"
  },
  {
    "epoch": 744,
    "batch_idx": 64,
    "loss": 0.3315008580684662,
    "learning_rate": 0.0006956868334186018,
    "timestep_mean": 520.5,
    "epsilon_norm": 8.02260684967041,
    "prediction_norm": 7.190608024597168,
    "timestamp": "2025-07-14T12:33:59.863230"
  },
  {
    "epoch": 744,
    "batch_idx": 96,
    "loss": 0.22610481083393097,
    "learning_rate": 0.0006956868334186018,
    "timestep_mean": 475.1875,
    "epsilon_norm": 6.478628158569336,
    "prediction_norm": 5.5446624755859375,
    "timestamp": "2025-07-14T12:33:59.868090"
  },
  {
    "epoch": 744,
    "batch_idx": 128,
    "loss": 0.2496538907289505,
    "learning_rate": 0.0006956868334186018,
    "timestep_mean": 474.0625,
    "epsilon_norm": 7.8051676750183105,
    "prediction_norm": 7.094040393829346,
    "timestamp": "2025-07-14T12:33:59.872965"
  },
  {
    "epoch": 744,
    "batch_idx": 160,
    "loss": 0.1970292180776596,
    "learning_rate": 0.0006956868334186018,
    "timestep_mean": 539.90625,
    "epsilon_norm": 8.3163480758667,
    "prediction_norm": 7.696073055267334,
    "timestamp": "2025-07-14T12:33:59.877876"
  },
  {
    "epoch": 744,
    "batch_idx": 192,
    "loss": 0.23295491933822632,
    "learning_rate": 0.0006956868334186018,
    "timestep_mean": 466.125,
    "epsilon_norm": 3.438551664352417,
    "prediction_norm": 2.425489664077759,
    "timestamp": "2025-07-14T12:33:59.882728"
  },
  {
    "epoch": 745,
    "batch_idx": 0,
    "loss": 0.5718801021575928,
    "learning_rate": 0.0006949638438940946,
    "timestep_mean": 403.84375,
    "epsilon_norm": 8.822307586669922,
    "prediction_norm": 6.622082233428955,
    "timestamp": "2025-07-14T12:33:59.887708"
  },
  {
    "epoch": 745,
    "batch_idx": 32,
    "loss": 0.2233290821313858,
    "learning_rate": 0.0006949638438940946,
    "timestep_mean": 479.3125,
    "epsilon_norm": 8.158475875854492,
    "prediction_norm": 6.759502410888672,
    "timestamp": "2025-07-14T12:33:59.892587"
  },
  {
    "epoch": 745,
    "batch_idx": 64,
    "loss": 0.31390875577926636,
    "learning_rate": 0.0006949638438940946,
    "timestep_mean": 485.15625,
    "epsilon_norm": 8.350996971130371,
    "prediction_norm": 6.973779201507568,
    "timestamp": "2025-07-14T12:33:59.897538"
  },
  {
    "epoch": 745,
    "batch_idx": 96,
    "loss": 0.1911117285490036,
    "learning_rate": 0.0006949638438940946,
    "timestep_mean": 534.5,
    "epsilon_norm": 7.867158889770508,
    "prediction_norm": 6.519618511199951,
    "timestamp": "2025-07-14T12:33:59.902419"
  },
  {
    "epoch": 745,
    "batch_idx": 128,
    "loss": 0.37109410762786865,
    "learning_rate": 0.0006949638438940946,
    "timestep_mean": 496.6875,
    "epsilon_norm": 9.336570739746094,
    "prediction_norm": 8.010372161865234,
    "timestamp": "2025-07-14T12:33:59.907325"
  },
  {
    "epoch": 745,
    "batch_idx": 160,
    "loss": 0.32839155197143555,
    "learning_rate": 0.0006949638438940946,
    "timestep_mean": 475.84375,
    "epsilon_norm": 7.796741962432861,
    "prediction_norm": 6.408232688903809,
    "timestamp": "2025-07-14T12:33:59.912228"
  },
  {
    "epoch": 745,
    "batch_idx": 192,
    "loss": 0.18719199299812317,
    "learning_rate": 0.0006949638438940946,
    "timestep_mean": 362.625,
    "epsilon_norm": 3.354463577270508,
    "prediction_norm": 2.5140395164489746,
    "timestamp": "2025-07-14T12:33:59.917189"
  },
  {
    "epoch": 746,
    "batch_idx": 0,
    "loss": 0.20624038577079773,
    "learning_rate": 0.0006942403733156835,
    "timestep_mean": 596.71875,
    "epsilon_norm": 7.684670925140381,
    "prediction_norm": 6.5155768394470215,
    "timestamp": "2025-07-14T12:33:59.922114"
  },
  {
    "epoch": 746,
    "batch_idx": 32,
    "loss": 0.3530765175819397,
    "learning_rate": 0.0006942403733156835,
    "timestep_mean": 426.375,
    "epsilon_norm": 7.705456733703613,
    "prediction_norm": 6.415536880493164,
    "timestamp": "2025-07-14T12:33:59.926916"
  },
  {
    "epoch": 746,
    "batch_idx": 64,
    "loss": 0.29050081968307495,
    "learning_rate": 0.0006942403733156835,
    "timestep_mean": 466.40625,
    "epsilon_norm": 7.309947490692139,
    "prediction_norm": 5.9264936447143555,
    "timestamp": "2025-07-14T12:33:59.931836"
  },
  {
    "epoch": 746,
    "batch_idx": 96,
    "loss": 0.5090221166610718,
    "learning_rate": 0.0006942403733156835,
    "timestep_mean": 458.65625,
    "epsilon_norm": 8.005852699279785,
    "prediction_norm": 6.475186347961426,
    "timestamp": "2025-07-14T12:33:59.936783"
  },
  {
    "epoch": 746,
    "batch_idx": 128,
    "loss": 0.3087034225463867,
    "learning_rate": 0.0006942403733156835,
    "timestep_mean": 507.8125,
    "epsilon_norm": 7.513553142547607,
    "prediction_norm": 6.442535877227783,
    "timestamp": "2025-07-14T12:33:59.941814"
  },
  {
    "epoch": 746,
    "batch_idx": 160,
    "loss": 0.24057403206825256,
    "learning_rate": 0.0006942403733156835,
    "timestep_mean": 477.4375,
    "epsilon_norm": 8.624598503112793,
    "prediction_norm": 7.43643856048584,
    "timestamp": "2025-07-14T12:33:59.946719"
  },
  {
    "epoch": 746,
    "batch_idx": 192,
    "loss": 0.15028680860996246,
    "learning_rate": 0.0006942403733156835,
    "timestep_mean": 509.5,
    "epsilon_norm": 3.824725866317749,
    "prediction_norm": 3.2947306632995605,
    "timestamp": "2025-07-14T12:33:59.951603"
  },
  {
    "epoch": 747,
    "batch_idx": 0,
    "loss": 0.42586323618888855,
    "learning_rate": 0.0006935164234684602,
    "timestep_mean": 481.71875,
    "epsilon_norm": 7.8832197189331055,
    "prediction_norm": 6.42658805847168,
    "timestamp": "2025-07-14T12:33:59.956529"
  },
  {
    "epoch": 747,
    "batch_idx": 32,
    "loss": 0.2081858217716217,
    "learning_rate": 0.0006935164234684602,
    "timestep_mean": 426.875,
    "epsilon_norm": 7.322075366973877,
    "prediction_norm": 6.127760410308838,
    "timestamp": "2025-07-14T12:33:59.961428"
  },
  {
    "epoch": 747,
    "batch_idx": 64,
    "loss": 0.24498175084590912,
    "learning_rate": 0.0006935164234684602,
    "timestep_mean": 532.84375,
    "epsilon_norm": 7.7332963943481445,
    "prediction_norm": 6.4066057205200195,
    "timestamp": "2025-07-14T12:33:59.966297"
  },
  {
    "epoch": 747,
    "batch_idx": 96,
    "loss": 0.260880708694458,
    "learning_rate": 0.0006935164234684602,
    "timestep_mean": 554.0625,
    "epsilon_norm": 9.091285705566406,
    "prediction_norm": 8.053641319274902,
    "timestamp": "2025-07-14T12:33:59.971297"
  },
  {
    "epoch": 747,
    "batch_idx": 128,
    "loss": 0.3413112759590149,
    "learning_rate": 0.0006935164234684602,
    "timestep_mean": 513.21875,
    "epsilon_norm": 8.737039566040039,
    "prediction_norm": 7.037387847900391,
    "timestamp": "2025-07-14T12:33:59.976137"
  },
  {
    "epoch": 747,
    "batch_idx": 160,
    "loss": 0.31157809495925903,
    "learning_rate": 0.0006935164234684602,
    "timestep_mean": 464.78125,
    "epsilon_norm": 8.385238647460938,
    "prediction_norm": 7.1304192543029785,
    "timestamp": "2025-07-14T12:33:59.981138"
  },
  {
    "epoch": 747,
    "batch_idx": 192,
    "loss": 0.12381885945796967,
    "learning_rate": 0.0006935164234684602,
    "timestep_mean": 504.75,
    "epsilon_norm": 4.225733280181885,
    "prediction_norm": 4.046024322509766,
    "timestamp": "2025-07-14T12:33:59.986299"
  },
  {
    "epoch": 748,
    "batch_idx": 0,
    "loss": 0.4040175676345825,
    "learning_rate": 0.0006927919961386987,
    "timestep_mean": 485.3125,
    "epsilon_norm": 8.629755973815918,
    "prediction_norm": 6.575131893157959,
    "timestamp": "2025-07-14T12:33:59.991634"
  },
  {
    "epoch": 748,
    "batch_idx": 32,
    "loss": 0.26848384737968445,
    "learning_rate": 0.0006927919961386987,
    "timestep_mean": 568.03125,
    "epsilon_norm": 9.622514724731445,
    "prediction_norm": 8.858237266540527,
    "timestamp": "2025-07-14T12:33:59.996738"
  },
  {
    "epoch": 748,
    "batch_idx": 64,
    "loss": 0.31647026538848877,
    "learning_rate": 0.0006927919961386987,
    "timestep_mean": 465.03125,
    "epsilon_norm": 8.05984115600586,
    "prediction_norm": 6.402967929840088,
    "timestamp": "2025-07-14T12:34:00.001776"
  },
  {
    "epoch": 748,
    "batch_idx": 96,
    "loss": 0.3641531467437744,
    "learning_rate": 0.0006927919961386987,
    "timestep_mean": 564.46875,
    "epsilon_norm": 8.145549774169922,
    "prediction_norm": 6.77419900894165,
    "timestamp": "2025-07-14T12:34:00.007275"
  },
  {
    "epoch": 748,
    "batch_idx": 128,
    "loss": 0.4064788818359375,
    "learning_rate": 0.0006927919961386987,
    "timestep_mean": 460.40625,
    "epsilon_norm": 8.245457649230957,
    "prediction_norm": 6.956646919250488,
    "timestamp": "2025-07-14T12:34:00.012291"
  },
  {
    "epoch": 748,
    "batch_idx": 160,
    "loss": 0.3821560740470886,
    "learning_rate": 0.0006927919961386987,
    "timestep_mean": 413.78125,
    "epsilon_norm": 9.225370407104492,
    "prediction_norm": 7.8162641525268555,
    "timestamp": "2025-07-14T12:34:00.017446"
  },
  {
    "epoch": 748,
    "batch_idx": 192,
    "loss": 0.20819348096847534,
    "learning_rate": 0.0006927919961386987,
    "timestep_mean": 458.25,
    "epsilon_norm": 3.6982526779174805,
    "prediction_norm": 2.84865665435791,
    "timestamp": "2025-07-14T12:34:00.022557"
  },
  {
    "epoch": 749,
    "batch_idx": 0,
    "loss": 0.15104684233665466,
    "learning_rate": 0.0006920670931138517,
    "timestep_mean": 525.8125,
    "epsilon_norm": 7.133618354797363,
    "prediction_norm": 7.2475152015686035,
    "timestamp": "2025-07-14T12:34:00.027751"
  },
  {
    "epoch": 749,
    "batch_idx": 32,
    "loss": 0.349892258644104,
    "learning_rate": 0.0006920670931138517,
    "timestep_mean": 502.09375,
    "epsilon_norm": 8.498212814331055,
    "prediction_norm": 7.284247398376465,
    "timestamp": "2025-07-14T12:34:00.032896"
  },
  {
    "epoch": 749,
    "batch_idx": 64,
    "loss": 0.2480369359254837,
    "learning_rate": 0.0006920670931138517,
    "timestep_mean": 536.0,
    "epsilon_norm": 8.922290802001953,
    "prediction_norm": 8.133062362670898,
    "timestamp": "2025-07-14T12:34:00.038102"
  },
  {
    "epoch": 749,
    "batch_idx": 96,
    "loss": 0.16496844589710236,
    "learning_rate": 0.0006920670931138517,
    "timestep_mean": 537.4375,
    "epsilon_norm": 6.876428127288818,
    "prediction_norm": 6.201807022094727,
    "timestamp": "2025-07-14T12:34:00.043306"
  },
  {
    "epoch": 749,
    "batch_idx": 128,
    "loss": 0.6331501603126526,
    "learning_rate": 0.0006920670931138517,
    "timestep_mean": 383.90625,
    "epsilon_norm": 7.9285807609558105,
    "prediction_norm": 5.531793594360352,
    "timestamp": "2025-07-14T12:34:00.048464"
  },
  {
    "epoch": 749,
    "batch_idx": 160,
    "loss": 0.21305543184280396,
    "learning_rate": 0.0006920670931138517,
    "timestep_mean": 535.0,
    "epsilon_norm": 6.97083044052124,
    "prediction_norm": 6.080755233764648,
    "timestamp": "2025-07-14T12:34:00.053703"
  },
  {
    "epoch": 749,
    "batch_idx": 192,
    "loss": 0.25777360796928406,
    "learning_rate": 0.0006920670931138517,
    "timestep_mean": 443.625,
    "epsilon_norm": 4.134942531585693,
    "prediction_norm": 3.0583553314208984,
    "timestamp": "2025-07-14T12:34:00.058959"
  },
  {
    "epoch": 750,
    "batch_idx": 0,
    "loss": 0.2827247977256775,
    "learning_rate": 0.0006913417161825454,
    "timestep_mean": 417.125,
    "epsilon_norm": 8.610978126525879,
    "prediction_norm": 7.589603424072266,
    "timestamp": "2025-07-14T12:34:00.064031"
  },
  {
    "epoch": 750,
    "batch_idx": 32,
    "loss": 0.250152051448822,
    "learning_rate": 0.0006913417161825454,
    "timestep_mean": 490.75,
    "epsilon_norm": 8.710386276245117,
    "prediction_norm": 7.852663516998291,
    "timestamp": "2025-07-14T12:34:00.069251"
  },
  {
    "epoch": 750,
    "batch_idx": 64,
    "loss": 0.1749686300754547,
    "learning_rate": 0.0006913417161825454,
    "timestep_mean": 480.8125,
    "epsilon_norm": 8.05138874053955,
    "prediction_norm": 6.9740891456604,
    "timestamp": "2025-07-14T12:34:00.074410"
  },
  {
    "epoch": 750,
    "batch_idx": 96,
    "loss": 0.1577179729938507,
    "learning_rate": 0.0006913417161825454,
    "timestep_mean": 526.75,
    "epsilon_norm": 7.104442596435547,
    "prediction_norm": 6.476675510406494,
    "timestamp": "2025-07-14T12:34:00.079755"
  },
  {
    "epoch": 750,
    "batch_idx": 128,
    "loss": 0.11558595299720764,
    "learning_rate": 0.0006913417161825454,
    "timestep_mean": 555.0,
    "epsilon_norm": 7.382700443267822,
    "prediction_norm": 6.776037693023682,
    "timestamp": "2025-07-14T12:34:00.084971"
  },
  {
    "epoch": 750,
    "batch_idx": 160,
    "loss": 0.5650748014450073,
    "learning_rate": 0.0006913417161825454,
    "timestep_mean": 402.78125,
    "epsilon_norm": 8.684063911437988,
    "prediction_norm": 5.791816711425781,
    "timestamp": "2025-07-14T12:34:00.090150"
  },
  {
    "epoch": 750,
    "batch_idx": 192,
    "loss": 0.15836849808692932,
    "learning_rate": 0.0006913417161825454,
    "timestep_mean": 444.25,
    "epsilon_norm": 3.25347638130188,
    "prediction_norm": 3.1828861236572266,
    "timestamp": "2025-07-14T12:34:00.095230"
  },
  {
    "epoch": 751,
    "batch_idx": 0,
    "loss": 0.19920974969863892,
    "learning_rate": 0.0006906158671345751,
    "timestep_mean": 450.96875,
    "epsilon_norm": 7.438549518585205,
    "prediction_norm": 6.508028984069824,
    "timestamp": "2025-07-14T12:34:00.100436"
  },
  {
    "epoch": 751,
    "batch_idx": 32,
    "loss": 0.283825546503067,
    "learning_rate": 0.0006906158671345751,
    "timestep_mean": 490.78125,
    "epsilon_norm": 8.046797752380371,
    "prediction_norm": 7.112633228302002,
    "timestamp": "2025-07-14T12:34:00.105620"
  },
  {
    "epoch": 751,
    "batch_idx": 64,
    "loss": 0.20273268222808838,
    "learning_rate": 0.0006906158671345751,
    "timestep_mean": 479.0,
    "epsilon_norm": 7.371939182281494,
    "prediction_norm": 6.364645481109619,
    "timestamp": "2025-07-14T12:34:00.110780"
  },
  {
    "epoch": 751,
    "batch_idx": 96,
    "loss": 0.3229551315307617,
    "learning_rate": 0.0006906158671345751,
    "timestep_mean": 540.6875,
    "epsilon_norm": 8.630426406860352,
    "prediction_norm": 7.157614707946777,
    "timestamp": "2025-07-14T12:34:00.115973"
  },
  {
    "epoch": 751,
    "batch_idx": 128,
    "loss": 0.633960485458374,
    "learning_rate": 0.0006906158671345751,
    "timestep_mean": 404.59375,
    "epsilon_norm": 9.52485179901123,
    "prediction_norm": 6.650822162628174,
    "timestamp": "2025-07-14T12:34:00.121308"
  },
  {
    "epoch": 751,
    "batch_idx": 160,
    "loss": 0.36875563859939575,
    "learning_rate": 0.0006906158671345751,
    "timestep_mean": 467.75,
    "epsilon_norm": 8.474211692810059,
    "prediction_norm": 7.01563024520874,
    "timestamp": "2025-07-14T12:34:00.126421"
  },
  {
    "epoch": 751,
    "batch_idx": 192,
    "loss": 0.6733106374740601,
    "learning_rate": 0.0006906158671345751,
    "timestep_mean": 385.125,
    "epsilon_norm": 4.169654369354248,
    "prediction_norm": 2.3716368675231934,
    "timestamp": "2025-07-14T12:34:00.131615"
  },
  {
    "epoch": 752,
    "batch_idx": 0,
    "loss": 0.24636498093605042,
    "learning_rate": 0.0006898895477609012,
    "timestep_mean": 518.3125,
    "epsilon_norm": 8.028030395507812,
    "prediction_norm": 6.977594375610352,
    "timestamp": "2025-07-14T12:34:00.136796"
  },
  {
    "epoch": 752,
    "batch_idx": 32,
    "loss": 0.3023231029510498,
    "learning_rate": 0.0006898895477609012,
    "timestep_mean": 524.1875,
    "epsilon_norm": 7.545277118682861,
    "prediction_norm": 6.575173854827881,
    "timestamp": "2025-07-14T12:34:00.142064"
  },
  {
    "epoch": 752,
    "batch_idx": 64,
    "loss": 0.16390013694763184,
    "learning_rate": 0.0006898895477609012,
    "timestep_mean": 495.28125,
    "epsilon_norm": 8.85409927368164,
    "prediction_norm": 7.957536220550537,
    "timestamp": "2025-07-14T12:34:00.147320"
  },
  {
    "epoch": 752,
    "batch_idx": 96,
    "loss": 0.28985968232154846,
    "learning_rate": 0.0006898895477609012,
    "timestep_mean": 425.3125,
    "epsilon_norm": 7.720755577087402,
    "prediction_norm": 6.247060298919678,
    "timestamp": "2025-07-14T12:34:00.152582"
  },
  {
    "epoch": 752,
    "batch_idx": 128,
    "loss": 0.2117139995098114,
    "learning_rate": 0.0006898895477609012,
    "timestep_mean": 530.03125,
    "epsilon_norm": 7.682403564453125,
    "prediction_norm": 6.7003302574157715,
    "timestamp": "2025-07-14T12:34:00.157788"
  },
  {
    "epoch": 752,
    "batch_idx": 160,
    "loss": 0.31747502088546753,
    "learning_rate": 0.0006898895477609012,
    "timestep_mean": 483.46875,
    "epsilon_norm": 8.243123054504395,
    "prediction_norm": 7.291342258453369,
    "timestamp": "2025-07-14T12:34:00.163208"
  },
  {
    "epoch": 752,
    "batch_idx": 192,
    "loss": 1.0477900505065918,
    "learning_rate": 0.0006898895477609012,
    "timestep_mean": 449.875,
    "epsilon_norm": 5.578405857086182,
    "prediction_norm": 3.852064847946167,
    "timestamp": "2025-07-14T12:34:00.168432"
  },
  {
    "epoch": 753,
    "batch_idx": 0,
    "loss": 0.3708476126194,
    "learning_rate": 0.0006891627598536445,
    "timestep_mean": 565.53125,
    "epsilon_norm": 8.674689292907715,
    "prediction_norm": 7.150520324707031,
    "timestamp": "2025-07-14T12:34:00.173555"
  },
  {
    "epoch": 753,
    "batch_idx": 32,
    "loss": 0.2004997730255127,
    "learning_rate": 0.0006891627598536445,
    "timestep_mean": 552.21875,
    "epsilon_norm": 8.057435035705566,
    "prediction_norm": 6.94464635848999,
    "timestamp": "2025-07-14T12:34:00.178804"
  },
  {
    "epoch": 753,
    "batch_idx": 64,
    "loss": 0.287986159324646,
    "learning_rate": 0.0006891627598536445,
    "timestep_mean": 551.21875,
    "epsilon_norm": 8.664434432983398,
    "prediction_norm": 7.493192195892334,
    "timestamp": "2025-07-14T12:34:00.184182"
  },
  {
    "epoch": 753,
    "batch_idx": 96,
    "loss": 0.2248123288154602,
    "learning_rate": 0.0006891627598536445,
    "timestep_mean": 546.40625,
    "epsilon_norm": 8.115249633789062,
    "prediction_norm": 6.498839855194092,
    "timestamp": "2025-07-14T12:34:00.189367"
  },
  {
    "epoch": 753,
    "batch_idx": 128,
    "loss": 0.3362908959388733,
    "learning_rate": 0.0006891627598536445,
    "timestep_mean": 558.34375,
    "epsilon_norm": 8.256958961486816,
    "prediction_norm": 6.773054599761963,
    "timestamp": "2025-07-14T12:34:00.194544"
  },
  {
    "epoch": 753,
    "batch_idx": 160,
    "loss": 0.24700410664081573,
    "learning_rate": 0.0006891627598536445,
    "timestep_mean": 526.0,
    "epsilon_norm": 8.154114723205566,
    "prediction_norm": 7.038292407989502,
    "timestamp": "2025-07-14T12:34:00.199731"
  },
  {
    "epoch": 753,
    "batch_idx": 192,
    "loss": 0.13311398029327393,
    "learning_rate": 0.0006891627598536445,
    "timestep_mean": 692.0,
    "epsilon_norm": 3.919980764389038,
    "prediction_norm": 3.3381505012512207,
    "timestamp": "2025-07-14T12:34:00.204891"
  },
  {
    "epoch": 754,
    "batch_idx": 0,
    "loss": 0.39801791310310364,
    "learning_rate": 0.0006884355052060819,
    "timestep_mean": 560.40625,
    "epsilon_norm": 8.48791790008545,
    "prediction_norm": 6.9296112060546875,
    "timestamp": "2025-07-14T12:34:00.210063"
  },
  {
    "epoch": 754,
    "batch_idx": 32,
    "loss": 0.32326245307922363,
    "learning_rate": 0.0006884355052060819,
    "timestep_mean": 455.3125,
    "epsilon_norm": 6.9077935218811035,
    "prediction_norm": 5.4554033279418945,
    "timestamp": "2025-07-14T12:34:00.215293"
  },
  {
    "epoch": 754,
    "batch_idx": 64,
    "loss": 0.2631897032260895,
    "learning_rate": 0.0006884355052060819,
    "timestep_mean": 547.75,
    "epsilon_norm": 8.799450874328613,
    "prediction_norm": 7.831597805023193,
    "timestamp": "2025-07-14T12:34:00.220507"
  },
  {
    "epoch": 754,
    "batch_idx": 96,
    "loss": 0.20646198093891144,
    "learning_rate": 0.0006884355052060819,
    "timestep_mean": 545.875,
    "epsilon_norm": 8.645868301391602,
    "prediction_norm": 7.226813316345215,
    "timestamp": "2025-07-14T12:34:00.225854"
  },
  {
    "epoch": 754,
    "batch_idx": 128,
    "loss": 0.34312841296195984,
    "learning_rate": 0.0006884355052060819,
    "timestep_mean": 497.03125,
    "epsilon_norm": 6.904964447021484,
    "prediction_norm": 6.054078578948975,
    "timestamp": "2025-07-14T12:34:00.231084"
  },
  {
    "epoch": 754,
    "batch_idx": 160,
    "loss": 0.22655650973320007,
    "learning_rate": 0.0006884355052060819,
    "timestep_mean": 558.75,
    "epsilon_norm": 7.977053642272949,
    "prediction_norm": 6.765519142150879,
    "timestamp": "2025-07-14T12:34:00.236196"
  },
  {
    "epoch": 754,
    "batch_idx": 192,
    "loss": 0.47118908166885376,
    "learning_rate": 0.0006884355052060819,
    "timestep_mean": 526.875,
    "epsilon_norm": 4.62803840637207,
    "prediction_norm": 3.59529972076416,
    "timestamp": "2025-07-14T12:34:00.241407"
  },
  {
    "epoch": 755,
    "batch_idx": 0,
    "loss": 0.2632891535758972,
    "learning_rate": 0.0006877077856126421,
    "timestep_mean": 510.5,
    "epsilon_norm": 9.099077224731445,
    "prediction_norm": 8.167033195495605,
    "timestamp": "2025-07-14T12:34:00.246584"
  },
  {
    "epoch": 755,
    "batch_idx": 32,
    "loss": 0.23091387748718262,
    "learning_rate": 0.0006877077856126421,
    "timestep_mean": 479.28125,
    "epsilon_norm": 7.541460037231445,
    "prediction_norm": 6.833068370819092,
    "timestamp": "2025-07-14T12:34:00.251784"
  },
  {
    "epoch": 755,
    "batch_idx": 64,
    "loss": 0.2356388419866562,
    "learning_rate": 0.0006877077856126421,
    "timestep_mean": 558.375,
    "epsilon_norm": 9.831572532653809,
    "prediction_norm": 9.141583442687988,
    "timestamp": "2025-07-14T12:34:00.256934"
  },
  {
    "epoch": 755,
    "batch_idx": 96,
    "loss": 0.4920646846294403,
    "learning_rate": 0.0006877077856126421,
    "timestep_mean": 467.96875,
    "epsilon_norm": 8.878438949584961,
    "prediction_norm": 6.959094047546387,
    "timestamp": "2025-07-14T12:34:00.262102"
  },
  {
    "epoch": 755,
    "batch_idx": 128,
    "loss": 0.28422439098358154,
    "learning_rate": 0.0006877077856126421,
    "timestep_mean": 545.625,
    "epsilon_norm": 7.931662082672119,
    "prediction_norm": 7.431674480438232,
    "timestamp": "2025-07-14T12:34:00.267182"
  },
  {
    "epoch": 755,
    "batch_idx": 160,
    "loss": 0.3622138500213623,
    "learning_rate": 0.0006877077856126421,
    "timestep_mean": 577.0,
    "epsilon_norm": 9.26666259765625,
    "prediction_norm": 8.507786750793457,
    "timestamp": "2025-07-14T12:34:00.272312"
  },
  {
    "epoch": 755,
    "batch_idx": 192,
    "loss": 0.2089216560125351,
    "learning_rate": 0.0006877077856126421,
    "timestep_mean": 646.75,
    "epsilon_norm": 3.013202667236328,
    "prediction_norm": 2.6348726749420166,
    "timestamp": "2025-07-14T12:34:00.277457"
  },
  {
    "epoch": 756,
    "batch_idx": 0,
    "loss": 0.2704434096813202,
    "learning_rate": 0.0006869796028689006,
    "timestep_mean": 569.75,
    "epsilon_norm": 8.698359489440918,
    "prediction_norm": 7.551875114440918,
    "timestamp": "2025-07-14T12:34:00.282654"
  },
  {
    "epoch": 756,
    "batch_idx": 32,
    "loss": 0.26554983854293823,
    "learning_rate": 0.0006869796028689006,
    "timestep_mean": 394.46875,
    "epsilon_norm": 6.471426963806152,
    "prediction_norm": 5.165889263153076,
    "timestamp": "2025-07-14T12:34:00.287967"
  },
  {
    "epoch": 756,
    "batch_idx": 64,
    "loss": 0.19751930236816406,
    "learning_rate": 0.0006869796028689006,
    "timestep_mean": 468.53125,
    "epsilon_norm": 7.843505382537842,
    "prediction_norm": 6.882437229156494,
    "timestamp": "2025-07-14T12:34:00.293123"
  },
  {
    "epoch": 756,
    "batch_idx": 96,
    "loss": 0.19683781266212463,
    "learning_rate": 0.0006869796028689006,
    "timestep_mean": 417.0,
    "epsilon_norm": 6.979611396789551,
    "prediction_norm": 6.225618839263916,
    "timestamp": "2025-07-14T12:34:00.298084"
  },
  {
    "epoch": 756,
    "batch_idx": 128,
    "loss": 0.28842589259147644,
    "learning_rate": 0.0006869796028689006,
    "timestep_mean": 526.40625,
    "epsilon_norm": 8.652109146118164,
    "prediction_norm": 7.2948503494262695,
    "timestamp": "2025-07-14T12:34:00.307292"
  },
  {
    "epoch": 756,
    "batch_idx": 160,
    "loss": 0.28081411123275757,
    "learning_rate": 0.0006869796028689006,
    "timestep_mean": 546.09375,
    "epsilon_norm": 6.560567855834961,
    "prediction_norm": 5.888656139373779,
    "timestamp": "2025-07-14T12:34:00.312254"
  },
  {
    "epoch": 756,
    "batch_idx": 192,
    "loss": 0.4252753257751465,
    "learning_rate": 0.0006869796028689006,
    "timestep_mean": 435.375,
    "epsilon_norm": 3.993616819381714,
    "prediction_norm": 3.6327881813049316,
    "timestamp": "2025-07-14T12:34:00.317892"
  },
  {
    "epoch": 757,
    "batch_idx": 0,
    "loss": 0.28265976905822754,
    "learning_rate": 0.0006862509587715765,
    "timestep_mean": 479.53125,
    "epsilon_norm": 6.973104953765869,
    "prediction_norm": 5.718372344970703,
    "timestamp": "2025-07-14T12:34:00.323334"
  },
  {
    "epoch": 757,
    "batch_idx": 32,
    "loss": 0.4053038954734802,
    "learning_rate": 0.0006862509587715765,
    "timestep_mean": 534.46875,
    "epsilon_norm": 8.366137504577637,
    "prediction_norm": 7.166070938110352,
    "timestamp": "2025-07-14T12:34:00.328484"
  },
  {
    "epoch": 757,
    "batch_idx": 64,
    "loss": 0.304686963558197,
    "learning_rate": 0.0006862509587715765,
    "timestep_mean": 507.625,
    "epsilon_norm": 7.426486492156982,
    "prediction_norm": 5.260166645050049,
    "timestamp": "2025-07-14T12:34:00.333971"
  },
  {
    "epoch": 757,
    "batch_idx": 96,
    "loss": 0.34362220764160156,
    "learning_rate": 0.0006862509587715765,
    "timestep_mean": 492.15625,
    "epsilon_norm": 7.685445308685303,
    "prediction_norm": 6.419299602508545,
    "timestamp": "2025-07-14T12:34:00.339137"
  },
  {
    "epoch": 757,
    "batch_idx": 128,
    "loss": 0.19199903309345245,
    "learning_rate": 0.0006862509587715765,
    "timestep_mean": 540.84375,
    "epsilon_norm": 7.256570339202881,
    "prediction_norm": 5.560619831085205,
    "timestamp": "2025-07-14T12:34:00.344339"
  },
  {
    "epoch": 757,
    "batch_idx": 160,
    "loss": 0.2393396496772766,
    "learning_rate": 0.0006862509587715765,
    "timestep_mean": 456.15625,
    "epsilon_norm": 7.887454032897949,
    "prediction_norm": 6.7814764976501465,
    "timestamp": "2025-07-14T12:34:00.349479"
  },
  {
    "epoch": 757,
    "batch_idx": 192,
    "loss": 0.48631593585014343,
    "learning_rate": 0.0006862509587715765,
    "timestep_mean": 535.75,
    "epsilon_norm": 4.642507076263428,
    "prediction_norm": 3.378000497817993,
    "timestamp": "2025-07-14T12:34:00.354756"
  },
  {
    "epoch": 758,
    "batch_idx": 0,
    "loss": 0.49090567231178284,
    "learning_rate": 0.0006855218551185261,
    "timestep_mean": 476.28125,
    "epsilon_norm": 8.08925724029541,
    "prediction_norm": 5.864712238311768,
    "timestamp": "2025-07-14T12:34:00.360003"
  },
  {
    "epoch": 758,
    "batch_idx": 32,
    "loss": 0.2814182937145233,
    "learning_rate": 0.0006855218551185261,
    "timestep_mean": 487.875,
    "epsilon_norm": 8.300003051757812,
    "prediction_norm": 7.337600231170654,
    "timestamp": "2025-07-14T12:34:00.365238"
  },
  {
    "epoch": 758,
    "batch_idx": 64,
    "loss": 0.41070279479026794,
    "learning_rate": 0.0006855218551185261,
    "timestep_mean": 416.46875,
    "epsilon_norm": 8.515212059020996,
    "prediction_norm": 6.218814849853516,
    "timestamp": "2025-07-14T12:34:00.370512"
  },
  {
    "epoch": 758,
    "batch_idx": 96,
    "loss": 0.29895079135894775,
    "learning_rate": 0.0006855218551185261,
    "timestep_mean": 545.03125,
    "epsilon_norm": 8.237577438354492,
    "prediction_norm": 7.213611602783203,
    "timestamp": "2025-07-14T12:34:00.375807"
  },
  {
    "epoch": 758,
    "batch_idx": 128,
    "loss": 0.27830424904823303,
    "learning_rate": 0.0006855218551185261,
    "timestep_mean": 452.4375,
    "epsilon_norm": 8.185298919677734,
    "prediction_norm": 6.839956760406494,
    "timestamp": "2025-07-14T12:34:00.381017"
  },
  {
    "epoch": 758,
    "batch_idx": 160,
    "loss": 0.33104416728019714,
    "learning_rate": 0.0006855218551185261,
    "timestep_mean": 418.34375,
    "epsilon_norm": 7.445013999938965,
    "prediction_norm": 5.844623565673828,
    "timestamp": "2025-07-14T12:34:00.386206"
  },
  {
    "epoch": 758,
    "batch_idx": 192,
    "loss": 0.28518903255462646,
    "learning_rate": 0.0006855218551185261,
    "timestep_mean": 461.625,
    "epsilon_norm": 4.141830921173096,
    "prediction_norm": 3.4939045906066895,
    "timestamp": "2025-07-14T12:34:00.391420"
  },
  {
    "epoch": 759,
    "batch_idx": 0,
    "loss": 0.1883586198091507,
    "learning_rate": 0.0006847922937087404,
    "timestep_mean": 648.5,
    "epsilon_norm": 7.47743034362793,
    "prediction_norm": 6.471220016479492,
    "timestamp": "2025-07-14T12:34:00.396707"
  },
  {
    "epoch": 759,
    "batch_idx": 32,
    "loss": 0.2236834019422531,
    "learning_rate": 0.0006847922937087404,
    "timestep_mean": 461.125,
    "epsilon_norm": 8.054644584655762,
    "prediction_norm": 7.169726848602295,
    "timestamp": "2025-07-14T12:34:00.402210"
  },
  {
    "epoch": 759,
    "batch_idx": 64,
    "loss": 0.23203587532043457,
    "learning_rate": 0.0006847922937087404,
    "timestep_mean": 458.40625,
    "epsilon_norm": 9.20173168182373,
    "prediction_norm": 8.103805541992188,
    "timestamp": "2025-07-14T12:34:00.407498"
  },
  {
    "epoch": 759,
    "batch_idx": 96,
    "loss": 0.2731964588165283,
    "learning_rate": 0.0006847922937087404,
    "timestep_mean": 518.625,
    "epsilon_norm": 8.234203338623047,
    "prediction_norm": 6.805638313293457,
    "timestamp": "2025-07-14T12:34:00.412371"
  },
  {
    "epoch": 759,
    "batch_idx": 128,
    "loss": 0.13827145099639893,
    "learning_rate": 0.0006847922937087404,
    "timestep_mean": 537.59375,
    "epsilon_norm": 8.673457145690918,
    "prediction_norm": 7.867798805236816,
    "timestamp": "2025-07-14T12:34:00.417785"
  },
  {
    "epoch": 759,
    "batch_idx": 160,
    "loss": 0.17382359504699707,
    "learning_rate": 0.0006847922937087404,
    "timestep_mean": 555.3125,
    "epsilon_norm": 8.49698543548584,
    "prediction_norm": 7.276407241821289,
    "timestamp": "2025-07-14T12:34:00.423306"
  },
  {
    "epoch": 759,
    "batch_idx": 192,
    "loss": 0.15508702397346497,
    "learning_rate": 0.0006847922937087404,
    "timestep_mean": 575.375,
    "epsilon_norm": 3.4725182056427,
    "prediction_norm": 2.879624605178833,
    "timestamp": "2025-07-14T12:34:00.428719"
  },
  {
    "epoch": 760,
    "batch_idx": 0,
    "loss": 0.13515378534793854,
    "learning_rate": 0.0006840622763423396,
    "timestep_mean": 511.28125,
    "epsilon_norm": 7.664071083068848,
    "prediction_norm": 6.591329574584961,
    "timestamp": "2025-07-14T12:34:00.434161"
  },
  {
    "epoch": 760,
    "batch_idx": 32,
    "loss": 0.27699214220046997,
    "learning_rate": 0.0006840622763423396,
    "timestep_mean": 473.8125,
    "epsilon_norm": 7.7711501121521,
    "prediction_norm": 6.672968864440918,
    "timestamp": "2025-07-14T12:34:00.439383"
  },
  {
    "epoch": 760,
    "batch_idx": 64,
    "loss": 0.2574191093444824,
    "learning_rate": 0.0006840622763423396,
    "timestep_mean": 500.375,
    "epsilon_norm": 7.248721599578857,
    "prediction_norm": 6.06232213973999,
    "timestamp": "2025-07-14T12:34:00.444572"
  },
  {
    "epoch": 760,
    "batch_idx": 96,
    "loss": 0.18875333666801453,
    "learning_rate": 0.0006840622763423396,
    "timestep_mean": 580.0625,
    "epsilon_norm": 8.121109962463379,
    "prediction_norm": 7.28546142578125,
    "timestamp": "2025-07-14T12:34:00.449984"
  },
  {
    "epoch": 760,
    "batch_idx": 128,
    "loss": 0.42775142192840576,
    "learning_rate": 0.0006840622763423396,
    "timestep_mean": 441.59375,
    "epsilon_norm": 7.10413122177124,
    "prediction_norm": 5.262937068939209,
    "timestamp": "2025-07-14T12:34:00.455260"
  },
  {
    "epoch": 760,
    "batch_idx": 160,
    "loss": 0.4818965196609497,
    "learning_rate": 0.0006840622763423396,
    "timestep_mean": 439.8125,
    "epsilon_norm": 8.606507301330566,
    "prediction_norm": 6.58779239654541,
    "timestamp": "2025-07-14T12:34:00.460503"
  },
  {
    "epoch": 760,
    "batch_idx": 192,
    "loss": 0.2661372721195221,
    "learning_rate": 0.0006840622763423396,
    "timestep_mean": 396.5,
    "epsilon_norm": 3.4976372718811035,
    "prediction_norm": 2.734276533126831,
    "timestamp": "2025-07-14T12:34:00.465671"
  },
  {
    "epoch": 761,
    "batch_idx": 0,
    "loss": 0.29071080684661865,
    "learning_rate": 0.000683331804820569,
    "timestep_mean": 550.125,
    "epsilon_norm": 7.505110740661621,
    "prediction_norm": 6.513011455535889,
    "timestamp": "2025-07-14T12:34:00.470869"
  },
  {
    "epoch": 761,
    "batch_idx": 32,
    "loss": 0.26245033740997314,
    "learning_rate": 0.000683331804820569,
    "timestep_mean": 468.28125,
    "epsilon_norm": 6.721161365509033,
    "prediction_norm": 5.795951843261719,
    "timestamp": "2025-07-14T12:34:00.476106"
  },
  {
    "epoch": 761,
    "batch_idx": 64,
    "loss": 0.33668404817581177,
    "learning_rate": 0.000683331804820569,
    "timestep_mean": 417.40625,
    "epsilon_norm": 6.314907550811768,
    "prediction_norm": 4.920339107513428,
    "timestamp": "2025-07-14T12:34:00.481299"
  },
  {
    "epoch": 761,
    "batch_idx": 96,
    "loss": 0.5050286054611206,
    "learning_rate": 0.000683331804820569,
    "timestep_mean": 495.125,
    "epsilon_norm": 9.526799201965332,
    "prediction_norm": 7.546884536743164,
    "timestamp": "2025-07-14T12:34:00.486434"
  },
  {
    "epoch": 761,
    "batch_idx": 128,
    "loss": 0.18266957998275757,
    "learning_rate": 0.000683331804820569,
    "timestep_mean": 536.40625,
    "epsilon_norm": 7.967593669891357,
    "prediction_norm": 7.171773433685303,
    "timestamp": "2025-07-14T12:34:00.491803"
  },
  {
    "epoch": 761,
    "batch_idx": 160,
    "loss": 0.21825197339057922,
    "learning_rate": 0.000683331804820569,
    "timestep_mean": 460.96875,
    "epsilon_norm": 7.064614295959473,
    "prediction_norm": 6.234253883361816,
    "timestamp": "2025-07-14T12:34:00.497060"
  },
  {
    "epoch": 761,
    "batch_idx": 192,
    "loss": 0.026347851380705833,
    "learning_rate": 0.000683331804820569,
    "timestep_mean": 768.5,
    "epsilon_norm": 3.631228446960449,
    "prediction_norm": 4.0321044921875,
    "timestamp": "2025-07-14T12:34:00.502227"
  },
  {
    "epoch": 762,
    "batch_idx": 0,
    "loss": 0.2590547204017639,
    "learning_rate": 0.0006826008809457945,
    "timestep_mean": 520.46875,
    "epsilon_norm": 7.973249435424805,
    "prediction_norm": 6.612236022949219,
    "timestamp": "2025-07-14T12:34:00.507508"
  },
  {
    "epoch": 762,
    "batch_idx": 32,
    "loss": 0.26768720149993896,
    "learning_rate": 0.0006826008809457945,
    "timestep_mean": 474.3125,
    "epsilon_norm": 7.312922477722168,
    "prediction_norm": 7.0054707527160645,
    "timestamp": "2025-07-14T12:34:00.512732"
  },
  {
    "epoch": 762,
    "batch_idx": 64,
    "loss": 0.2967383563518524,
    "learning_rate": 0.0006826008809457945,
    "timestep_mean": 441.875,
    "epsilon_norm": 7.673308372497559,
    "prediction_norm": 6.686188697814941,
    "timestamp": "2025-07-14T12:34:00.517955"
  },
  {
    "epoch": 762,
    "batch_idx": 96,
    "loss": 0.25042062997817993,
    "learning_rate": 0.0006826008809457945,
    "timestep_mean": 568.03125,
    "epsilon_norm": 8.929605484008789,
    "prediction_norm": 7.295935153961182,
    "timestamp": "2025-07-14T12:34:00.523156"
  },
  {
    "epoch": 762,
    "batch_idx": 128,
    "loss": 0.4716404974460602,
    "learning_rate": 0.0006826008809457945,
    "timestep_mean": 479.71875,
    "epsilon_norm": 8.40628433227539,
    "prediction_norm": 6.449341773986816,
    "timestamp": "2025-07-14T12:34:00.528245"
  },
  {
    "epoch": 762,
    "batch_idx": 160,
    "loss": 0.4988769590854645,
    "learning_rate": 0.0006826008809457945,
    "timestep_mean": 505.15625,
    "epsilon_norm": 7.757968425750732,
    "prediction_norm": 5.885504722595215,
    "timestamp": "2025-07-14T12:34:00.533549"
  },
  {
    "epoch": 762,
    "batch_idx": 192,
    "loss": 0.23479962348937988,
    "learning_rate": 0.0006826008809457945,
    "timestep_mean": 565.25,
    "epsilon_norm": 3.4280142784118652,
    "prediction_norm": 3.0622718334198,
    "timestamp": "2025-07-14T12:34:00.538711"
  },
  {
    "epoch": 763,
    "batch_idx": 0,
    "loss": 0.28917378187179565,
    "learning_rate": 0.000681869506521498,
    "timestep_mean": 481.65625,
    "epsilon_norm": 7.522876262664795,
    "prediction_norm": 6.287927627563477,
    "timestamp": "2025-07-14T12:34:00.543872"
  },
  {
    "epoch": 763,
    "batch_idx": 32,
    "loss": 0.2341376543045044,
    "learning_rate": 0.000681869506521498,
    "timestep_mean": 467.875,
    "epsilon_norm": 7.582867622375488,
    "prediction_norm": 6.388352870941162,
    "timestamp": "2025-07-14T12:34:00.549145"
  },
  {
    "epoch": 763,
    "batch_idx": 64,
    "loss": 0.35872843861579895,
    "learning_rate": 0.000681869506521498,
    "timestep_mean": 497.09375,
    "epsilon_norm": 8.299418449401855,
    "prediction_norm": 6.364987850189209,
    "timestamp": "2025-07-14T12:34:00.554283"
  },
  {
    "epoch": 763,
    "batch_idx": 96,
    "loss": 0.5255984663963318,
    "learning_rate": 0.000681869506521498,
    "timestep_mean": 496.875,
    "epsilon_norm": 7.71177339553833,
    "prediction_norm": 5.239797592163086,
    "timestamp": "2025-07-14T12:34:00.559367"
  },
  {
    "epoch": 763,
    "batch_idx": 128,
    "loss": 0.6334778070449829,
    "learning_rate": 0.000681869506521498,
    "timestep_mean": 413.71875,
    "epsilon_norm": 8.398163795471191,
    "prediction_norm": 5.263477325439453,
    "timestamp": "2025-07-14T12:34:00.564638"
  },
  {
    "epoch": 763,
    "batch_idx": 160,
    "loss": 0.33364081382751465,
    "learning_rate": 0.000681869506521498,
    "timestep_mean": 471.8125,
    "epsilon_norm": 8.453032493591309,
    "prediction_norm": 6.6137847900390625,
    "timestamp": "2025-07-14T12:34:00.569792"
  },
  {
    "epoch": 763,
    "batch_idx": 192,
    "loss": 0.30506038665771484,
    "learning_rate": 0.000681869506521498,
    "timestep_mean": 460.375,
    "epsilon_norm": 3.6052234172821045,
    "prediction_norm": 2.461052417755127,
    "timestamp": "2025-07-14T12:34:00.575115"
  },
  {
    "epoch": 764,
    "batch_idx": 0,
    "loss": 0.23086753487586975,
    "learning_rate": 0.0006811376833522734,
    "timestep_mean": 464.3125,
    "epsilon_norm": 7.412353038787842,
    "prediction_norm": 6.358001232147217,
    "timestamp": "2025-07-14T12:34:00.580401"
  },
  {
    "epoch": 764,
    "batch_idx": 32,
    "loss": 0.2801600694656372,
    "learning_rate": 0.0006811376833522734,
    "timestep_mean": 579.84375,
    "epsilon_norm": 8.701725959777832,
    "prediction_norm": 7.480426788330078,
    "timestamp": "2025-07-14T12:34:00.585550"
  },
  {
    "epoch": 764,
    "batch_idx": 64,
    "loss": 0.2694008946418762,
    "learning_rate": 0.0006811376833522734,
    "timestep_mean": 410.4375,
    "epsilon_norm": 7.087764739990234,
    "prediction_norm": 5.920346736907959,
    "timestamp": "2025-07-14T12:34:00.590728"
  },
  {
    "epoch": 764,
    "batch_idx": 96,
    "loss": 0.23097796738147736,
    "learning_rate": 0.0006811376833522734,
    "timestep_mean": 526.75,
    "epsilon_norm": 8.339038848876953,
    "prediction_norm": 7.546273708343506,
    "timestamp": "2025-07-14T12:34:00.595961"
  },
  {
    "epoch": 764,
    "batch_idx": 128,
    "loss": 0.38106459379196167,
    "learning_rate": 0.0006811376833522734,
    "timestep_mean": 504.625,
    "epsilon_norm": 9.25262451171875,
    "prediction_norm": 7.819614410400391,
    "timestamp": "2025-07-14T12:34:00.601096"
  },
  {
    "epoch": 764,
    "batch_idx": 160,
    "loss": 0.21347549557685852,
    "learning_rate": 0.0006811376833522734,
    "timestep_mean": 574.71875,
    "epsilon_norm": 7.443282127380371,
    "prediction_norm": 6.676885604858398,
    "timestamp": "2025-07-14T12:34:00.606279"
  },
  {
    "epoch": 764,
    "batch_idx": 192,
    "loss": 0.7318053245544434,
    "learning_rate": 0.0006811376833522734,
    "timestep_mean": 653.75,
    "epsilon_norm": 4.491795063018799,
    "prediction_norm": 3.107403039932251,
    "timestamp": "2025-07-14T12:34:00.611422"
  },
  {
    "epoch": 765,
    "batch_idx": 0,
    "loss": 0.2833103537559509,
    "learning_rate": 0.0006804054132438215,
    "timestep_mean": 548.1875,
    "epsilon_norm": 8.322599411010742,
    "prediction_norm": 7.616994380950928,
    "timestamp": "2025-07-14T12:34:00.616810"
  },
  {
    "epoch": 765,
    "batch_idx": 32,
    "loss": 0.20071721076965332,
    "learning_rate": 0.0006804054132438215,
    "timestep_mean": 485.90625,
    "epsilon_norm": 7.384419918060303,
    "prediction_norm": 7.155956745147705,
    "timestamp": "2025-07-14T12:34:00.622009"
  },
  {
    "epoch": 765,
    "batch_idx": 64,
    "loss": 0.45141446590423584,
    "learning_rate": 0.0006804054132438215,
    "timestep_mean": 481.34375,
    "epsilon_norm": 8.56185245513916,
    "prediction_norm": 7.490810394287109,
    "timestamp": "2025-07-14T12:34:00.627196"
  },
  {
    "epoch": 765,
    "batch_idx": 96,
    "loss": 0.30181291699409485,
    "learning_rate": 0.0006804054132438215,
    "timestep_mean": 635.21875,
    "epsilon_norm": 7.782972812652588,
    "prediction_norm": 7.050706386566162,
    "timestamp": "2025-07-14T12:34:00.632379"
  },
  {
    "epoch": 765,
    "batch_idx": 128,
    "loss": 0.2852363586425781,
    "learning_rate": 0.0006804054132438215,
    "timestep_mean": 494.5,
    "epsilon_norm": 7.255921363830566,
    "prediction_norm": 5.474769115447998,
    "timestamp": "2025-07-14T12:34:00.637531"
  },
  {
    "epoch": 765,
    "batch_idx": 160,
    "loss": 0.4625736474990845,
    "learning_rate": 0.0006804054132438215,
    "timestep_mean": 515.59375,
    "epsilon_norm": 7.7600603103637695,
    "prediction_norm": 5.645328998565674,
    "timestamp": "2025-07-14T12:34:00.642729"
  },
  {
    "epoch": 765,
    "batch_idx": 192,
    "loss": 0.28811174631118774,
    "learning_rate": 0.0006804054132438215,
    "timestep_mean": 322.75,
    "epsilon_norm": 4.371129989624023,
    "prediction_norm": 2.6248631477355957,
    "timestamp": "2025-07-14T12:34:00.647947"
  },
  {
    "epoch": 766,
    "batch_idx": 0,
    "loss": 0.13996900618076324,
    "learning_rate": 0.0006796726980029459,
    "timestep_mean": 546.1875,
    "epsilon_norm": 7.327193260192871,
    "prediction_norm": 6.221517562866211,
    "timestamp": "2025-07-14T12:34:00.652998"
  },
  {
    "epoch": 766,
    "batch_idx": 32,
    "loss": 0.22782443463802338,
    "learning_rate": 0.0006796726980029459,
    "timestep_mean": 458.125,
    "epsilon_norm": 7.54335880279541,
    "prediction_norm": 6.678886890411377,
    "timestamp": "2025-07-14T12:34:00.657916"
  },
  {
    "epoch": 766,
    "batch_idx": 64,
    "loss": 0.6274237632751465,
    "learning_rate": 0.0006796726980029459,
    "timestep_mean": 382.1875,
    "epsilon_norm": 9.423897743225098,
    "prediction_norm": 6.820736408233643,
    "timestamp": "2025-07-14T12:34:00.662759"
  },
  {
    "epoch": 766,
    "batch_idx": 96,
    "loss": 0.3071362376213074,
    "learning_rate": 0.0006796726980029459,
    "timestep_mean": 458.65625,
    "epsilon_norm": 7.162008285522461,
    "prediction_norm": 6.214064598083496,
    "timestamp": "2025-07-14T12:34:00.667654"
  },
  {
    "epoch": 766,
    "batch_idx": 128,
    "loss": 0.16040751338005066,
    "learning_rate": 0.0006796726980029459,
    "timestep_mean": 477.875,
    "epsilon_norm": 8.023370742797852,
    "prediction_norm": 7.522610187530518,
    "timestamp": "2025-07-14T12:34:00.672527"
  },
  {
    "epoch": 766,
    "batch_idx": 160,
    "loss": 0.4517066776752472,
    "learning_rate": 0.0006796726980029459,
    "timestep_mean": 452.9375,
    "epsilon_norm": 8.035139083862305,
    "prediction_norm": 5.948617458343506,
    "timestamp": "2025-07-14T12:34:00.677436"
  },
  {
    "epoch": 766,
    "batch_idx": 192,
    "loss": 0.17081135511398315,
    "learning_rate": 0.0006796726980029459,
    "timestep_mean": 622.625,
    "epsilon_norm": 4.685177803039551,
    "prediction_norm": 4.2453484535217285,
    "timestamp": "2025-07-14T12:34:00.682267"
  },
  {
    "epoch": 767,
    "batch_idx": 0,
    "loss": 0.3550143241882324,
    "learning_rate": 0.0006789395394375486,
    "timestep_mean": 457.875,
    "epsilon_norm": 8.616722106933594,
    "prediction_norm": 7.179376602172852,
    "timestamp": "2025-07-14T12:34:00.687165"
  },
  {
    "epoch": 767,
    "batch_idx": 32,
    "loss": 0.3091662526130676,
    "learning_rate": 0.0006789395394375486,
    "timestep_mean": 454.0,
    "epsilon_norm": 7.9546942710876465,
    "prediction_norm": 6.513420581817627,
    "timestamp": "2025-07-14T12:34:00.692142"
  },
  {
    "epoch": 767,
    "batch_idx": 64,
    "loss": 0.3058120608329773,
    "learning_rate": 0.0006789395394375486,
    "timestep_mean": 561.78125,
    "epsilon_norm": 8.192017555236816,
    "prediction_norm": 7.933198928833008,
    "timestamp": "2025-07-14T12:34:00.696985"
  },
  {
    "epoch": 767,
    "batch_idx": 96,
    "loss": 0.35120052099227905,
    "learning_rate": 0.0006789395394375486,
    "timestep_mean": 505.34375,
    "epsilon_norm": 7.470240592956543,
    "prediction_norm": 6.2296223640441895,
    "timestamp": "2025-07-14T12:34:00.701851"
  },
  {
    "epoch": 767,
    "batch_idx": 128,
    "loss": 0.302950918674469,
    "learning_rate": 0.0006789395394375486,
    "timestep_mean": 518.5625,
    "epsilon_norm": 9.153650283813477,
    "prediction_norm": 7.419779300689697,
    "timestamp": "2025-07-14T12:34:00.706685"
  },
  {
    "epoch": 767,
    "batch_idx": 160,
    "loss": 0.33410096168518066,
    "learning_rate": 0.0006789395394375486,
    "timestep_mean": 526.46875,
    "epsilon_norm": 7.679688930511475,
    "prediction_norm": 5.817961692810059,
    "timestamp": "2025-07-14T12:34:00.711583"
  },
  {
    "epoch": 767,
    "batch_idx": 192,
    "loss": 0.1621684730052948,
    "learning_rate": 0.0006789395394375486,
    "timestep_mean": 534.75,
    "epsilon_norm": 3.9229578971862793,
    "prediction_norm": 3.669187545776367,
    "timestamp": "2025-07-14T12:34:00.716378"
  },
  {
    "epoch": 768,
    "batch_idx": 0,
    "loss": 0.518855631351471,
    "learning_rate": 0.0006782059393566258,
    "timestep_mean": 453.84375,
    "epsilon_norm": 8.131281852722168,
    "prediction_norm": 5.748144149780273,
    "timestamp": "2025-07-14T12:34:00.721291"
  },
  {
    "epoch": 768,
    "batch_idx": 32,
    "loss": 0.4466109871864319,
    "learning_rate": 0.0006782059393566258,
    "timestep_mean": 507.9375,
    "epsilon_norm": 8.425948143005371,
    "prediction_norm": 6.822482109069824,
    "timestamp": "2025-07-14T12:34:00.726032"
  },
  {
    "epoch": 768,
    "batch_idx": 64,
    "loss": 0.19094523787498474,
    "learning_rate": 0.0006782059393566258,
    "timestep_mean": 509.34375,
    "epsilon_norm": 8.026504516601562,
    "prediction_norm": 6.531182765960693,
    "timestamp": "2025-07-14T12:34:00.730916"
  },
  {
    "epoch": 768,
    "batch_idx": 96,
    "loss": 0.30961617827415466,
    "learning_rate": 0.0006782059393566258,
    "timestep_mean": 513.75,
    "epsilon_norm": 7.125604629516602,
    "prediction_norm": 6.238638401031494,
    "timestamp": "2025-07-14T12:34:00.735780"
  },
  {
    "epoch": 768,
    "batch_idx": 128,
    "loss": 0.3846457004547119,
    "learning_rate": 0.0006782059393566258,
    "timestep_mean": 543.4375,
    "epsilon_norm": 8.012465476989746,
    "prediction_norm": 5.4766526222229,
    "timestamp": "2025-07-14T12:34:00.740608"
  },
  {
    "epoch": 768,
    "batch_idx": 160,
    "loss": 0.3283265233039856,
    "learning_rate": 0.0006782059393566258,
    "timestep_mean": 510.21875,
    "epsilon_norm": 7.179832935333252,
    "prediction_norm": 5.887435436248779,
    "timestamp": "2025-07-14T12:34:00.745441"
  },
  {
    "epoch": 768,
    "batch_idx": 192,
    "loss": 0.4014700949192047,
    "learning_rate": 0.0006782059393566258,
    "timestep_mean": 326.5,
    "epsilon_norm": 3.168858528137207,
    "prediction_norm": 2.6439208984375,
    "timestamp": "2025-07-14T12:34:00.750321"
  },
  {
    "epoch": 769,
    "batch_idx": 0,
    "loss": 0.3058556318283081,
    "learning_rate": 0.0006774718995702626,
    "timestep_mean": 485.40625,
    "epsilon_norm": 7.521402835845947,
    "prediction_norm": 6.3321709632873535,
    "timestamp": "2025-07-14T12:34:00.755152"
  },
  {
    "epoch": 769,
    "batch_idx": 32,
    "loss": 0.41804301738739014,
    "learning_rate": 0.0006774718995702626,
    "timestep_mean": 475.6875,
    "epsilon_norm": 7.558708667755127,
    "prediction_norm": 5.569374084472656,
    "timestamp": "2025-07-14T12:34:00.759995"
  },
  {
    "epoch": 769,
    "batch_idx": 64,
    "loss": 0.26908254623413086,
    "learning_rate": 0.0006774718995702626,
    "timestep_mean": 411.3125,
    "epsilon_norm": 7.818443775177002,
    "prediction_norm": 6.186042308807373,
    "timestamp": "2025-07-14T12:34:00.764852"
  },
  {
    "epoch": 769,
    "batch_idx": 96,
    "loss": 0.6137735843658447,
    "learning_rate": 0.0006774718995702626,
    "timestep_mean": 499.5,
    "epsilon_norm": 9.552529335021973,
    "prediction_norm": 6.639481067657471,
    "timestamp": "2025-07-14T12:34:00.769683"
  },
  {
    "epoch": 769,
    "batch_idx": 128,
    "loss": 0.1774873584508896,
    "learning_rate": 0.0006774718995702626,
    "timestep_mean": 560.375,
    "epsilon_norm": 7.211103439331055,
    "prediction_norm": 7.056578159332275,
    "timestamp": "2025-07-14T12:34:00.774537"
  },
  {
    "epoch": 769,
    "batch_idx": 160,
    "loss": 0.2898373007774353,
    "learning_rate": 0.0006774718995702626,
    "timestep_mean": 540.25,
    "epsilon_norm": 7.837486267089844,
    "prediction_norm": 6.376238822937012,
    "timestamp": "2025-07-14T12:34:00.779508"
  },
  {
    "epoch": 769,
    "batch_idx": 192,
    "loss": 0.884296178817749,
    "learning_rate": 0.0006774718995702626,
    "timestep_mean": 406.625,
    "epsilon_norm": 5.119576454162598,
    "prediction_norm": 3.2234954833984375,
    "timestamp": "2025-07-14T12:34:00.784399"
  },
  {
    "epoch": 770,
    "batch_idx": 0,
    "loss": 0.40686047077178955,
    "learning_rate": 0.0006767374218896291,
    "timestep_mean": 488.84375,
    "epsilon_norm": 7.920162200927734,
    "prediction_norm": 6.179548740386963,
    "timestamp": "2025-07-14T12:34:00.789309"
  },
  {
    "epoch": 770,
    "batch_idx": 32,
    "loss": 0.2432096302509308,
    "learning_rate": 0.0006767374218896291,
    "timestep_mean": 532.34375,
    "epsilon_norm": 7.887821197509766,
    "prediction_norm": 6.815384387969971,
    "timestamp": "2025-07-14T12:34:00.794194"
  },
  {
    "epoch": 770,
    "batch_idx": 64,
    "loss": 0.21205158531665802,
    "learning_rate": 0.0006767374218896291,
    "timestep_mean": 445.375,
    "epsilon_norm": 7.801424026489258,
    "prediction_norm": 7.071230411529541,
    "timestamp": "2025-07-14T12:34:00.799057"
  },
  {
    "epoch": 770,
    "batch_idx": 96,
    "loss": 0.28500375151634216,
    "learning_rate": 0.0006767374218896291,
    "timestep_mean": 470.28125,
    "epsilon_norm": 7.903038501739502,
    "prediction_norm": 6.733908176422119,
    "timestamp": "2025-07-14T12:34:00.803913"
  },
  {
    "epoch": 770,
    "batch_idx": 128,
    "loss": 0.29353421926498413,
    "learning_rate": 0.0006767374218896291,
    "timestep_mean": 494.65625,
    "epsilon_norm": 9.205822944641113,
    "prediction_norm": 7.508460521697998,
    "timestamp": "2025-07-14T12:34:00.808852"
  },
  {
    "epoch": 770,
    "batch_idx": 160,
    "loss": 0.45114684104919434,
    "learning_rate": 0.0006767374218896291,
    "timestep_mean": 532.09375,
    "epsilon_norm": 8.027658462524414,
    "prediction_norm": 6.1127400398254395,
    "timestamp": "2025-07-14T12:34:00.813676"
  },
  {
    "epoch": 770,
    "batch_idx": 192,
    "loss": 0.6299835443496704,
    "learning_rate": 0.0006767374218896291,
    "timestep_mean": 437.375,
    "epsilon_norm": 4.789022922515869,
    "prediction_norm": 3.953490972518921,
    "timestamp": "2025-07-14T12:34:00.818532"
  },
  {
    "epoch": 771,
    "batch_idx": 0,
    "loss": 0.3026584982872009,
    "learning_rate": 0.0006760025081269761,
    "timestep_mean": 432.34375,
    "epsilon_norm": 7.251044273376465,
    "prediction_norm": 5.832332611083984,
    "timestamp": "2025-07-14T12:34:00.823387"
  },
  {
    "epoch": 771,
    "batch_idx": 32,
    "loss": 0.1849319040775299,
    "learning_rate": 0.0006760025081269761,
    "timestep_mean": 550.0625,
    "epsilon_norm": 8.398734092712402,
    "prediction_norm": 7.768378734588623,
    "timestamp": "2025-07-14T12:34:00.828249"
  },
  {
    "epoch": 771,
    "batch_idx": 64,
    "loss": 0.35215291380882263,
    "learning_rate": 0.0006760025081269761,
    "timestep_mean": 413.21875,
    "epsilon_norm": 8.6207275390625,
    "prediction_norm": 6.655646324157715,
    "timestamp": "2025-07-14T12:34:00.833250"
  },
  {
    "epoch": 771,
    "batch_idx": 96,
    "loss": 0.20466631650924683,
    "learning_rate": 0.0006760025081269761,
    "timestep_mean": 602.09375,
    "epsilon_norm": 7.649155139923096,
    "prediction_norm": 6.293452262878418,
    "timestamp": "2025-07-14T12:34:00.838084"
  },
  {
    "epoch": 771,
    "batch_idx": 128,
    "loss": 0.6074598431587219,
    "learning_rate": 0.0006760025081269761,
    "timestep_mean": 494.53125,
    "epsilon_norm": 8.915994644165039,
    "prediction_norm": 6.924543380737305,
    "timestamp": "2025-07-14T12:34:00.842867"
  },
  {
    "epoch": 771,
    "batch_idx": 160,
    "loss": 0.38256388902664185,
    "learning_rate": 0.0006760025081269761,
    "timestep_mean": 500.15625,
    "epsilon_norm": 7.917591571807861,
    "prediction_norm": 6.85955810546875,
    "timestamp": "2025-07-14T12:34:00.847751"
  },
  {
    "epoch": 771,
    "batch_idx": 192,
    "loss": 1.1697678565979004,
    "learning_rate": 0.0006760025081269761,
    "timestep_mean": 411.625,
    "epsilon_norm": 4.7204270362854,
    "prediction_norm": 3.4156494140625,
    "timestamp": "2025-07-14T12:34:00.852524"
  },
  {
    "epoch": 772,
    "batch_idx": 0,
    "loss": 0.16430440545082092,
    "learning_rate": 0.00067526716009563,
    "timestep_mean": 578.5,
    "epsilon_norm": 8.150582313537598,
    "prediction_norm": 7.646988391876221,
    "timestamp": "2025-07-14T12:34:00.857362"
  },
  {
    "epoch": 772,
    "batch_idx": 32,
    "loss": 0.18577840924263,
    "learning_rate": 0.00067526716009563,
    "timestep_mean": 536.59375,
    "epsilon_norm": 6.998836517333984,
    "prediction_norm": 6.242587089538574,
    "timestamp": "2025-07-14T12:34:00.862277"
  },
  {
    "epoch": 772,
    "batch_idx": 64,
    "loss": 0.21979978680610657,
    "learning_rate": 0.00067526716009563,
    "timestep_mean": 633.75,
    "epsilon_norm": 9.284226417541504,
    "prediction_norm": 7.926877975463867,
    "timestamp": "2025-07-14T12:34:00.867170"
  },
  {
    "epoch": 772,
    "batch_idx": 96,
    "loss": 0.4141610860824585,
    "learning_rate": 0.00067526716009563,
    "timestep_mean": 454.15625,
    "epsilon_norm": 7.902661323547363,
    "prediction_norm": 5.653197288513184,
    "timestamp": "2025-07-14T12:34:00.872003"
  },
  {
    "epoch": 772,
    "batch_idx": 128,
    "loss": 0.11098356544971466,
    "learning_rate": 0.00067526716009563,
    "timestep_mean": 609.78125,
    "epsilon_norm": 7.448870658874512,
    "prediction_norm": 7.1334614753723145,
    "timestamp": "2025-07-14T12:34:00.876856"
  },
  {
    "epoch": 772,
    "batch_idx": 160,
    "loss": 0.2272820621728897,
    "learning_rate": 0.00067526716009563,
    "timestep_mean": 476.65625,
    "epsilon_norm": 8.110603332519531,
    "prediction_norm": 6.275060176849365,
    "timestamp": "2025-07-14T12:34:00.881949"
  },
  {
    "epoch": 772,
    "batch_idx": 192,
    "loss": 0.21184414625167847,
    "learning_rate": 0.00067526716009563,
    "timestep_mean": 497.375,
    "epsilon_norm": 4.019181728363037,
    "prediction_norm": 2.9533843994140625,
    "timestamp": "2025-07-14T12:34:00.887088"
  },
  {
    "epoch": 773,
    "batch_idx": 0,
    "loss": 0.2945951819419861,
    "learning_rate": 0.0006745313796099893,
    "timestep_mean": 525.53125,
    "epsilon_norm": 7.685843467712402,
    "prediction_norm": 6.406893253326416,
    "timestamp": "2025-07-14T12:34:00.892276"
  },
  {
    "epoch": 773,
    "batch_idx": 32,
    "loss": 0.19051921367645264,
    "learning_rate": 0.0006745313796099893,
    "timestep_mean": 573.5,
    "epsilon_norm": 8.396506309509277,
    "prediction_norm": 7.70584774017334,
    "timestamp": "2025-07-14T12:34:00.897509"
  },
  {
    "epoch": 773,
    "batch_idx": 64,
    "loss": 0.23583467304706573,
    "learning_rate": 0.0006745313796099893,
    "timestep_mean": 463.375,
    "epsilon_norm": 8.147857666015625,
    "prediction_norm": 7.375539302825928,
    "timestamp": "2025-07-14T12:34:00.902497"
  },
  {
    "epoch": 773,
    "batch_idx": 96,
    "loss": 0.13829465210437775,
    "learning_rate": 0.0006745313796099893,
    "timestep_mean": 529.40625,
    "epsilon_norm": 7.207621097564697,
    "prediction_norm": 6.977088451385498,
    "timestamp": "2025-07-14T12:34:00.911535"
  },
  {
    "epoch": 773,
    "batch_idx": 128,
    "loss": 0.3220707178115845,
    "learning_rate": 0.0006745313796099893,
    "timestep_mean": 487.9375,
    "epsilon_norm": 8.21557903289795,
    "prediction_norm": 6.807068347930908,
    "timestamp": "2025-07-14T12:34:00.916772"
  },
  {
    "epoch": 773,
    "batch_idx": 160,
    "loss": 0.27091410756111145,
    "learning_rate": 0.0006745313796099893,
    "timestep_mean": 530.25,
    "epsilon_norm": 7.702857971191406,
    "prediction_norm": 7.3246989250183105,
    "timestamp": "2025-07-14T12:34:00.921915"
  },
  {
    "epoch": 773,
    "batch_idx": 192,
    "loss": 0.25970107316970825,
    "learning_rate": 0.0006745313796099893,
    "timestep_mean": 587.375,
    "epsilon_norm": 4.872201442718506,
    "prediction_norm": 4.7933735847473145,
    "timestamp": "2025-07-14T12:34:00.927067"
  },
  {
    "epoch": 774,
    "batch_idx": 0,
    "loss": 0.3950503468513489,
    "learning_rate": 0.000673795168485519,
    "timestep_mean": 506.84375,
    "epsilon_norm": 7.856884479522705,
    "prediction_norm": 7.405538558959961,
    "timestamp": "2025-07-14T12:34:00.932283"
  },
  {
    "epoch": 774,
    "batch_idx": 32,
    "loss": 0.3401866853237152,
    "learning_rate": 0.000673795168485519,
    "timestep_mean": 504.25,
    "epsilon_norm": 8.080753326416016,
    "prediction_norm": 6.982304573059082,
    "timestamp": "2025-07-14T12:34:00.937411"
  },
  {
    "epoch": 774,
    "batch_idx": 64,
    "loss": 0.3096863627433777,
    "learning_rate": 0.000673795168485519,
    "timestep_mean": 439.65625,
    "epsilon_norm": 7.527021408081055,
    "prediction_norm": 6.153636455535889,
    "timestamp": "2025-07-14T12:34:00.942417"
  },
  {
    "epoch": 774,
    "batch_idx": 96,
    "loss": 0.3976234197616577,
    "learning_rate": 0.000673795168485519,
    "timestep_mean": 561.15625,
    "epsilon_norm": 8.382200241088867,
    "prediction_norm": 7.1644182205200195,
    "timestamp": "2025-07-14T12:34:00.947577"
  },
  {
    "epoch": 774,
    "batch_idx": 128,
    "loss": 0.1841634213924408,
    "learning_rate": 0.000673795168485519,
    "timestep_mean": 622.09375,
    "epsilon_norm": 8.802535057067871,
    "prediction_norm": 7.956564903259277,
    "timestamp": "2025-07-14T12:34:00.952705"
  },
  {
    "epoch": 774,
    "batch_idx": 160,
    "loss": 0.24820396304130554,
    "learning_rate": 0.000673795168485519,
    "timestep_mean": 420.0625,
    "epsilon_norm": 7.996700286865234,
    "prediction_norm": 6.737068176269531,
    "timestamp": "2025-07-14T12:34:00.957889"
  },
  {
    "epoch": 774,
    "batch_idx": 192,
    "loss": 0.1940264105796814,
    "learning_rate": 0.000673795168485519,
    "timestep_mean": 503.0,
    "epsilon_norm": 4.547476291656494,
    "prediction_norm": 4.387920379638672,
    "timestamp": "2025-07-14T12:34:00.963077"
  },
  {
    "epoch": 775,
    "batch_idx": 0,
    "loss": 0.3713032603263855,
    "learning_rate": 0.000673058528538747,
    "timestep_mean": 526.5,
    "epsilon_norm": 9.312052726745605,
    "prediction_norm": 7.620825290679932,
    "timestamp": "2025-07-14T12:34:00.968211"
  },
  {
    "epoch": 775,
    "batch_idx": 32,
    "loss": 0.19981172680854797,
    "learning_rate": 0.000673058528538747,
    "timestep_mean": 510.21875,
    "epsilon_norm": 7.35378360748291,
    "prediction_norm": 6.103830337524414,
    "timestamp": "2025-07-14T12:34:00.973400"
  },
  {
    "epoch": 775,
    "batch_idx": 64,
    "loss": 0.23423124849796295,
    "learning_rate": 0.000673058528538747,
    "timestep_mean": 459.09375,
    "epsilon_norm": 7.300588607788086,
    "prediction_norm": 5.926000118255615,
    "timestamp": "2025-07-14T12:34:00.978449"
  },
  {
    "epoch": 775,
    "batch_idx": 96,
    "loss": 0.5185193419456482,
    "learning_rate": 0.000673058528538747,
    "timestep_mean": 473.59375,
    "epsilon_norm": 8.657238960266113,
    "prediction_norm": 6.002519130706787,
    "timestamp": "2025-07-14T12:34:00.983466"
  },
  {
    "epoch": 775,
    "batch_idx": 128,
    "loss": 0.3267725110054016,
    "learning_rate": 0.000673058528538747,
    "timestep_mean": 500.5625,
    "epsilon_norm": 8.932024002075195,
    "prediction_norm": 7.1734466552734375,
    "timestamp": "2025-07-14T12:34:00.988301"
  },
  {
    "epoch": 775,
    "batch_idx": 160,
    "loss": 0.44791269302368164,
    "learning_rate": 0.000673058528538747,
    "timestep_mean": 453.6875,
    "epsilon_norm": 7.786929130554199,
    "prediction_norm": 6.462318420410156,
    "timestamp": "2025-07-14T12:34:00.993170"
  },
  {
    "epoch": 775,
    "batch_idx": 192,
    "loss": 0.309305876493454,
    "learning_rate": 0.000673058528538747,
    "timestep_mean": 409.5,
    "epsilon_norm": 3.569234848022461,
    "prediction_norm": 2.740623712539673,
    "timestamp": "2025-07-14T12:34:00.998030"
  },
  {
    "epoch": 776,
    "batch_idx": 0,
    "loss": 0.3347020149230957,
    "learning_rate": 0.000672321461587259,
    "timestep_mean": 472.09375,
    "epsilon_norm": 8.874320030212402,
    "prediction_norm": 7.223389625549316,
    "timestamp": "2025-07-14T12:34:01.002956"
  },
  {
    "epoch": 776,
    "batch_idx": 32,
    "loss": 0.42806291580200195,
    "learning_rate": 0.000672321461587259,
    "timestep_mean": 463.09375,
    "epsilon_norm": 7.614438533782959,
    "prediction_norm": 5.729492664337158,
    "timestamp": "2025-07-14T12:34:01.007797"
  },
  {
    "epoch": 776,
    "batch_idx": 64,
    "loss": 0.3429839015007019,
    "learning_rate": 0.000672321461587259,
    "timestep_mean": 358.40625,
    "epsilon_norm": 7.36329984664917,
    "prediction_norm": 5.942210674285889,
    "timestamp": "2025-07-14T12:34:01.012724"
  },
  {
    "epoch": 776,
    "batch_idx": 96,
    "loss": 0.36237210035324097,
    "learning_rate": 0.000672321461587259,
    "timestep_mean": 485.34375,
    "epsilon_norm": 8.414745330810547,
    "prediction_norm": 6.50822114944458,
    "timestamp": "2025-07-14T12:34:01.017568"
  },
  {
    "epoch": 776,
    "batch_idx": 128,
    "loss": 0.4932650625705719,
    "learning_rate": 0.000672321461587259,
    "timestep_mean": 518.0625,
    "epsilon_norm": 8.524234771728516,
    "prediction_norm": 6.568222999572754,
    "timestamp": "2025-07-14T12:34:01.022405"
  },
  {
    "epoch": 776,
    "batch_idx": 160,
    "loss": 0.1579011082649231,
    "learning_rate": 0.000672321461587259,
    "timestep_mean": 543.59375,
    "epsilon_norm": 6.9854021072387695,
    "prediction_norm": 6.146418571472168,
    "timestamp": "2025-07-14T12:34:01.027281"
  },
  {
    "epoch": 776,
    "batch_idx": 192,
    "loss": 0.029805805534124374,
    "learning_rate": 0.000672321461587259,
    "timestep_mean": 765.625,
    "epsilon_norm": 4.887668132781982,
    "prediction_norm": 4.781601428985596,
    "timestamp": "2025-07-14T12:34:01.032082"
  },
  {
    "epoch": 777,
    "batch_idx": 0,
    "loss": 0.3146163821220398,
    "learning_rate": 0.0006715839694496947,
    "timestep_mean": 428.59375,
    "epsilon_norm": 7.947925567626953,
    "prediction_norm": 5.905974388122559,
    "timestamp": "2025-07-14T12:34:01.036906"
  },
  {
    "epoch": 777,
    "batch_idx": 32,
    "loss": 0.38530126214027405,
    "learning_rate": 0.0006715839694496947,
    "timestep_mean": 454.125,
    "epsilon_norm": 8.349003791809082,
    "prediction_norm": 6.78541898727417,
    "timestamp": "2025-07-14T12:34:01.041756"
  },
  {
    "epoch": 777,
    "batch_idx": 64,
    "loss": 0.3112379014492035,
    "learning_rate": 0.0006715839694496947,
    "timestep_mean": 498.40625,
    "epsilon_norm": 8.128177642822266,
    "prediction_norm": 7.164653301239014,
    "timestamp": "2025-07-14T12:34:01.046662"
  },
  {
    "epoch": 777,
    "batch_idx": 96,
    "loss": 0.18577265739440918,
    "learning_rate": 0.0006715839694496947,
    "timestep_mean": 519.6875,
    "epsilon_norm": 7.524924278259277,
    "prediction_norm": 6.2064361572265625,
    "timestamp": "2025-07-14T12:34:01.051508"
  },
  {
    "epoch": 777,
    "batch_idx": 128,
    "loss": 0.3688119351863861,
    "learning_rate": 0.0006715839694496947,
    "timestep_mean": 484.09375,
    "epsilon_norm": 10.700857162475586,
    "prediction_norm": 9.129083633422852,
    "timestamp": "2025-07-14T12:34:01.056428"
  },
  {
    "epoch": 777,
    "batch_idx": 160,
    "loss": 0.3955153524875641,
    "learning_rate": 0.0006715839694496947,
    "timestep_mean": 466.625,
    "epsilon_norm": 9.263489723205566,
    "prediction_norm": 7.756608486175537,
    "timestamp": "2025-07-14T12:34:01.061276"
  },
  {
    "epoch": 777,
    "batch_idx": 192,
    "loss": 0.12346654385328293,
    "learning_rate": 0.0006715839694496947,
    "timestep_mean": 416.375,
    "epsilon_norm": 3.6531717777252197,
    "prediction_norm": 2.940910577774048,
    "timestamp": "2025-07-14T12:34:01.066046"
  },
  {
    "epoch": 778,
    "batch_idx": 0,
    "loss": 0.2505086660385132,
    "learning_rate": 0.0006708460539457421,
    "timestep_mean": 408.15625,
    "epsilon_norm": 7.61446475982666,
    "prediction_norm": 6.604176998138428,
    "timestamp": "2025-07-14T12:34:01.070940"
  },
  {
    "epoch": 778,
    "batch_idx": 32,
    "loss": 0.6747812032699585,
    "learning_rate": 0.0006708460539457421,
    "timestep_mean": 430.90625,
    "epsilon_norm": 7.939788818359375,
    "prediction_norm": 5.535750865936279,
    "timestamp": "2025-07-14T12:34:01.075790"
  },
  {
    "epoch": 778,
    "batch_idx": 64,
    "loss": 0.3711979389190674,
    "learning_rate": 0.0006708460539457421,
    "timestep_mean": 486.15625,
    "epsilon_norm": 7.682202339172363,
    "prediction_norm": 6.562282085418701,
    "timestamp": "2025-07-14T12:34:01.080693"
  },
  {
    "epoch": 778,
    "batch_idx": 96,
    "loss": 0.4299139976501465,
    "learning_rate": 0.0006708460539457421,
    "timestep_mean": 410.96875,
    "epsilon_norm": 7.1814188957214355,
    "prediction_norm": 5.700707912445068,
    "timestamp": "2025-07-14T12:34:01.085546"
  },
  {
    "epoch": 778,
    "batch_idx": 128,
    "loss": 0.30921268463134766,
    "learning_rate": 0.0006708460539457421,
    "timestep_mean": 524.03125,
    "epsilon_norm": 7.706587314605713,
    "prediction_norm": 6.812013149261475,
    "timestamp": "2025-07-14T12:34:01.090421"
  },
  {
    "epoch": 778,
    "batch_idx": 160,
    "loss": 0.21465584635734558,
    "learning_rate": 0.0006708460539457421,
    "timestep_mean": 507.90625,
    "epsilon_norm": 6.551901340484619,
    "prediction_norm": 6.012977600097656,
    "timestamp": "2025-07-14T12:34:01.095311"
  },
  {
    "epoch": 778,
    "batch_idx": 192,
    "loss": 0.4209069609642029,
    "learning_rate": 0.0006708460539457421,
    "timestep_mean": 375.375,
    "epsilon_norm": 4.456404685974121,
    "prediction_norm": 3.1460812091827393,
    "timestamp": "2025-07-14T12:34:01.100244"
  },
  {
    "epoch": 779,
    "batch_idx": 0,
    "loss": 0.308218777179718,
    "learning_rate": 0.0006701077168961349,
    "timestep_mean": 440.59375,
    "epsilon_norm": 7.6139984130859375,
    "prediction_norm": 6.564878940582275,
    "timestamp": "2025-07-14T12:34:01.105098"
  },
  {
    "epoch": 779,
    "batch_idx": 32,
    "loss": 0.245905339717865,
    "learning_rate": 0.0006701077168961349,
    "timestep_mean": 570.46875,
    "epsilon_norm": 7.953955173492432,
    "prediction_norm": 7.666914939880371,
    "timestamp": "2025-07-14T12:34:01.109977"
  },
  {
    "epoch": 779,
    "batch_idx": 64,
    "loss": 0.2744686007499695,
    "learning_rate": 0.0006701077168961349,
    "timestep_mean": 414.34375,
    "epsilon_norm": 7.257026195526123,
    "prediction_norm": 6.0907368659973145,
    "timestamp": "2025-07-14T12:34:01.114831"
  },
  {
    "epoch": 779,
    "batch_idx": 96,
    "loss": 0.1426929235458374,
    "learning_rate": 0.0006701077168961349,
    "timestep_mean": 557.5,
    "epsilon_norm": 7.587140083312988,
    "prediction_norm": 6.866144180297852,
    "timestamp": "2025-07-14T12:34:01.119664"
  },
  {
    "epoch": 779,
    "batch_idx": 128,
    "loss": 0.28450343012809753,
    "learning_rate": 0.0006701077168961349,
    "timestep_mean": 390.5625,
    "epsilon_norm": 7.528977394104004,
    "prediction_norm": 6.110707759857178,
    "timestamp": "2025-07-14T12:34:01.124531"
  },
  {
    "epoch": 779,
    "batch_idx": 160,
    "loss": 0.27263331413269043,
    "learning_rate": 0.0006701077168961349,
    "timestep_mean": 515.84375,
    "epsilon_norm": 6.683640003204346,
    "prediction_norm": 5.123011589050293,
    "timestamp": "2025-07-14T12:34:01.129457"
  },
  {
    "epoch": 779,
    "batch_idx": 192,
    "loss": 0.4788534343242645,
    "learning_rate": 0.0006701077168961349,
    "timestep_mean": 455.0,
    "epsilon_norm": 5.118138790130615,
    "prediction_norm": 3.6123952865600586,
    "timestamp": "2025-07-14T12:34:01.134265"
  },
  {
    "epoch": 780,
    "batch_idx": 0,
    "loss": 0.2868403196334839,
    "learning_rate": 0.0006693689601226463,
    "timestep_mean": 473.3125,
    "epsilon_norm": 7.022474765777588,
    "prediction_norm": 5.6537322998046875,
    "timestamp": "2025-07-14T12:34:01.139099"
  },
  {
    "epoch": 780,
    "batch_idx": 32,
    "loss": 0.1613360047340393,
    "learning_rate": 0.0006693689601226463,
    "timestep_mean": 520.625,
    "epsilon_norm": 7.05734920501709,
    "prediction_norm": 6.179283618927002,
    "timestamp": "2025-07-14T12:34:01.143957"
  },
  {
    "epoch": 780,
    "batch_idx": 64,
    "loss": 0.20420093834400177,
    "learning_rate": 0.0006693689601226463,
    "timestep_mean": 593.65625,
    "epsilon_norm": 7.485966205596924,
    "prediction_norm": 6.241004467010498,
    "timestamp": "2025-07-14T12:34:01.148788"
  },
  {
    "epoch": 780,
    "batch_idx": 96,
    "loss": 0.3140197992324829,
    "learning_rate": 0.0006693689601226463,
    "timestep_mean": 485.15625,
    "epsilon_norm": 7.1186323165893555,
    "prediction_norm": 5.647608757019043,
    "timestamp": "2025-07-14T12:34:01.153607"
  },
  {
    "epoch": 780,
    "batch_idx": 128,
    "loss": 0.1862148642539978,
    "learning_rate": 0.0006693689601226463,
    "timestep_mean": 528.5,
    "epsilon_norm": 7.072301387786865,
    "prediction_norm": 6.850555896759033,
    "timestamp": "2025-07-14T12:34:01.158459"
  },
  {
    "epoch": 780,
    "batch_idx": 160,
    "loss": 0.19752775132656097,
    "learning_rate": 0.0006693689601226463,
    "timestep_mean": 456.34375,
    "epsilon_norm": 7.769069194793701,
    "prediction_norm": 6.616793155670166,
    "timestamp": "2025-07-14T12:34:01.163322"
  },
  {
    "epoch": 780,
    "batch_idx": 192,
    "loss": 0.24449506402015686,
    "learning_rate": 0.0006693689601226463,
    "timestep_mean": 551.875,
    "epsilon_norm": 3.805224895477295,
    "prediction_norm": 3.1600828170776367,
    "timestamp": "2025-07-14T12:34:01.168194"
  },
  {
    "epoch": 781,
    "batch_idx": 0,
    "loss": 0.2641853988170624,
    "learning_rate": 0.0006686297854480847,
    "timestep_mean": 629.78125,
    "epsilon_norm": 8.805180549621582,
    "prediction_norm": 7.350062847137451,
    "timestamp": "2025-07-14T12:34:01.173050"
  },
  {
    "epoch": 781,
    "batch_idx": 32,
    "loss": 0.3475249707698822,
    "learning_rate": 0.0006686297854480847,
    "timestep_mean": 468.5625,
    "epsilon_norm": 7.717282772064209,
    "prediction_norm": 6.466390609741211,
    "timestamp": "2025-07-14T12:34:01.177922"
  },
  {
    "epoch": 781,
    "batch_idx": 64,
    "loss": 0.31649988889694214,
    "learning_rate": 0.0006686297854480847,
    "timestep_mean": 465.1875,
    "epsilon_norm": 7.205334186553955,
    "prediction_norm": 5.568127155303955,
    "timestamp": "2025-07-14T12:34:01.182809"
  },
  {
    "epoch": 781,
    "batch_idx": 96,
    "loss": 0.372577041387558,
    "learning_rate": 0.0006686297854480847,
    "timestep_mean": 508.46875,
    "epsilon_norm": 8.384940147399902,
    "prediction_norm": 6.163424015045166,
    "timestamp": "2025-07-14T12:34:01.187804"
  },
  {
    "epoch": 781,
    "batch_idx": 128,
    "loss": 0.3372517228126526,
    "learning_rate": 0.0006686297854480847,
    "timestep_mean": 545.75,
    "epsilon_norm": 8.55443286895752,
    "prediction_norm": 7.058516502380371,
    "timestamp": "2025-07-14T12:34:01.192715"
  },
  {
    "epoch": 781,
    "batch_idx": 160,
    "loss": 0.24579624831676483,
    "learning_rate": 0.0006686297854480847,
    "timestep_mean": 550.28125,
    "epsilon_norm": 7.874287128448486,
    "prediction_norm": 6.728898525238037,
    "timestamp": "2025-07-14T12:34:01.197587"
  },
  {
    "epoch": 781,
    "batch_idx": 192,
    "loss": 0.2933318316936493,
    "learning_rate": 0.0006686297854480847,
    "timestep_mean": 600.25,
    "epsilon_norm": 5.07674503326416,
    "prediction_norm": 5.098637580871582,
    "timestamp": "2025-07-14T12:34:01.202434"
  },
  {
    "epoch": 782,
    "batch_idx": 0,
    "loss": 0.3162297308444977,
    "learning_rate": 0.0006678901946962908,
    "timestep_mean": 482.71875,
    "epsilon_norm": 7.228177070617676,
    "prediction_norm": 6.318871021270752,
    "timestamp": "2025-07-14T12:34:01.207319"
  },
  {
    "epoch": 782,
    "batch_idx": 32,
    "loss": 0.27609360218048096,
    "learning_rate": 0.0006678901946962908,
    "timestep_mean": 496.9375,
    "epsilon_norm": 6.721283435821533,
    "prediction_norm": 5.21114444732666,
    "timestamp": "2025-07-14T12:34:01.212214"
  },
  {
    "epoch": 782,
    "batch_idx": 64,
    "loss": 0.3986307978630066,
    "learning_rate": 0.0006678901946962908,
    "timestep_mean": 447.53125,
    "epsilon_norm": 7.991375923156738,
    "prediction_norm": 6.52043342590332,
    "timestamp": "2025-07-14T12:34:01.217190"
  },
  {
    "epoch": 782,
    "batch_idx": 96,
    "loss": 0.2709435522556305,
    "learning_rate": 0.0006678901946962908,
    "timestep_mean": 536.125,
    "epsilon_norm": 8.280716896057129,
    "prediction_norm": 7.09640645980835,
    "timestamp": "2025-07-14T12:34:01.222071"
  },
  {
    "epoch": 782,
    "batch_idx": 128,
    "loss": 0.20954766869544983,
    "learning_rate": 0.0006678901946962908,
    "timestep_mean": 550.25,
    "epsilon_norm": 7.145319938659668,
    "prediction_norm": 6.3276543617248535,
    "timestamp": "2025-07-14T12:34:01.226964"
  },
  {
    "epoch": 782,
    "batch_idx": 160,
    "loss": 0.2003423124551773,
    "learning_rate": 0.0006678901946962908,
    "timestep_mean": 547.34375,
    "epsilon_norm": 7.5230793952941895,
    "prediction_norm": 6.307770729064941,
    "timestamp": "2025-07-14T12:34:01.231881"
  },
  {
    "epoch": 782,
    "batch_idx": 192,
    "loss": 0.1612679362297058,
    "learning_rate": 0.0006678901946962908,
    "timestep_mean": 722.125,
    "epsilon_norm": 4.3564252853393555,
    "prediction_norm": 3.6592156887054443,
    "timestamp": "2025-07-14T12:34:01.236751"
  },
  {
    "epoch": 783,
    "batch_idx": 0,
    "loss": 0.33835017681121826,
    "learning_rate": 0.0006671501896921309,
    "timestep_mean": 432.78125,
    "epsilon_norm": 7.652511119842529,
    "prediction_norm": 5.809495449066162,
    "timestamp": "2025-07-14T12:34:01.241731"
  },
  {
    "epoch": 783,
    "batch_idx": 32,
    "loss": 0.22451232373714447,
    "learning_rate": 0.0006671501896921309,
    "timestep_mean": 568.40625,
    "epsilon_norm": 8.218171119689941,
    "prediction_norm": 7.723814010620117,
    "timestamp": "2025-07-14T12:34:01.246604"
  },
  {
    "epoch": 783,
    "batch_idx": 64,
    "loss": 0.2568361163139343,
    "learning_rate": 0.0006671501896921309,
    "timestep_mean": 542.9375,
    "epsilon_norm": 8.15749454498291,
    "prediction_norm": 7.237539291381836,
    "timestamp": "2025-07-14T12:34:01.251487"
  },
  {
    "epoch": 783,
    "batch_idx": 96,
    "loss": 0.20953641831874847,
    "learning_rate": 0.0006671501896921309,
    "timestep_mean": 617.9375,
    "epsilon_norm": 7.938832759857178,
    "prediction_norm": 7.232602596282959,
    "timestamp": "2025-07-14T12:34:01.256338"
  },
  {
    "epoch": 783,
    "batch_idx": 128,
    "loss": 0.2985398769378662,
    "learning_rate": 0.0006671501896921309,
    "timestep_mean": 504.90625,
    "epsilon_norm": 6.646096706390381,
    "prediction_norm": 5.987393379211426,
    "timestamp": "2025-07-14T12:34:01.261245"
  },
  {
    "epoch": 783,
    "batch_idx": 160,
    "loss": 0.30471155047416687,
    "learning_rate": 0.0006671501896921309,
    "timestep_mean": 504.15625,
    "epsilon_norm": 7.776108741760254,
    "prediction_norm": 6.763390064239502,
    "timestamp": "2025-07-14T12:34:01.266184"
  },
  {
    "epoch": 783,
    "batch_idx": 192,
    "loss": 0.17830301821231842,
    "learning_rate": 0.0006671501896921309,
    "timestep_mean": 603.625,
    "epsilon_norm": 5.091731071472168,
    "prediction_norm": 4.227892875671387,
    "timestamp": "2025-07-14T12:34:01.271217"
  },
  {
    "epoch": 784,
    "batch_idx": 0,
    "loss": 0.2102794498205185,
    "learning_rate": 0.0006664097722614938,
    "timestep_mean": 572.6875,
    "epsilon_norm": 7.673713207244873,
    "prediction_norm": 6.788773059844971,
    "timestamp": "2025-07-14T12:34:01.276074"
  },
  {
    "epoch": 784,
    "batch_idx": 32,
    "loss": 0.3161223232746124,
    "learning_rate": 0.0006664097722614938,
    "timestep_mean": 401.875,
    "epsilon_norm": 7.440103054046631,
    "prediction_norm": 5.597384452819824,
    "timestamp": "2025-07-14T12:34:01.280969"
  },
  {
    "epoch": 784,
    "batch_idx": 64,
    "loss": 0.5572435259819031,
    "learning_rate": 0.0006664097722614938,
    "timestep_mean": 410.75,
    "epsilon_norm": 9.441514015197754,
    "prediction_norm": 7.013061046600342,
    "timestamp": "2025-07-14T12:34:01.285827"
  },
  {
    "epoch": 784,
    "batch_idx": 96,
    "loss": 0.2286054491996765,
    "learning_rate": 0.0006664097722614938,
    "timestep_mean": 438.3125,
    "epsilon_norm": 6.868591785430908,
    "prediction_norm": 6.077041149139404,
    "timestamp": "2025-07-14T12:34:01.290754"
  },
  {
    "epoch": 784,
    "batch_idx": 128,
    "loss": 0.2787936329841614,
    "learning_rate": 0.0006664097722614938,
    "timestep_mean": 515.3125,
    "epsilon_norm": 8.818778991699219,
    "prediction_norm": 7.346440315246582,
    "timestamp": "2025-07-14T12:34:01.295671"
  },
  {
    "epoch": 784,
    "batch_idx": 160,
    "loss": 0.2557951807975769,
    "learning_rate": 0.0006664097722614938,
    "timestep_mean": 503.6875,
    "epsilon_norm": 7.355800628662109,
    "prediction_norm": 5.697983741760254,
    "timestamp": "2025-07-14T12:34:01.300686"
  },
  {
    "epoch": 784,
    "batch_idx": 192,
    "loss": 0.08854995667934418,
    "learning_rate": 0.0006664097722614938,
    "timestep_mean": 703.625,
    "epsilon_norm": 3.950916290283203,
    "prediction_norm": 3.510270833969116,
    "timestamp": "2025-07-14T12:34:01.305618"
  },
  {
    "epoch": 785,
    "batch_idx": 0,
    "loss": 0.24475355446338654,
    "learning_rate": 0.0006656689442312858,
    "timestep_mean": 577.84375,
    "epsilon_norm": 9.662976264953613,
    "prediction_norm": 9.093342781066895,
    "timestamp": "2025-07-14T12:34:01.310508"
  },
  {
    "epoch": 785,
    "batch_idx": 32,
    "loss": 0.3537406921386719,
    "learning_rate": 0.0006656689442312858,
    "timestep_mean": 492.40625,
    "epsilon_norm": 9.426508903503418,
    "prediction_norm": 8.006451606750488,
    "timestamp": "2025-07-14T12:34:01.315381"
  },
  {
    "epoch": 785,
    "batch_idx": 64,
    "loss": 0.3220076560974121,
    "learning_rate": 0.0006656689442312858,
    "timestep_mean": 452.21875,
    "epsilon_norm": 7.440953254699707,
    "prediction_norm": 6.173100471496582,
    "timestamp": "2025-07-14T12:34:01.320232"
  },
  {
    "epoch": 785,
    "batch_idx": 96,
    "loss": 0.2283553034067154,
    "learning_rate": 0.0006656689442312858,
    "timestep_mean": 454.875,
    "epsilon_norm": 7.616853713989258,
    "prediction_norm": 6.5695624351501465,
    "timestamp": "2025-07-14T12:34:01.325141"
  },
  {
    "epoch": 785,
    "batch_idx": 128,
    "loss": 0.26481711864471436,
    "learning_rate": 0.0006656689442312858,
    "timestep_mean": 484.40625,
    "epsilon_norm": 7.99247407913208,
    "prediction_norm": 6.715317726135254,
    "timestamp": "2025-07-14T12:34:01.330006"
  },
  {
    "epoch": 785,
    "batch_idx": 160,
    "loss": 0.18625086545944214,
    "learning_rate": 0.0006656689442312858,
    "timestep_mean": 475.78125,
    "epsilon_norm": 7.6403093338012695,
    "prediction_norm": 6.474528789520264,
    "timestamp": "2025-07-14T12:34:01.334941"
  },
  {
    "epoch": 785,
    "batch_idx": 192,
    "loss": 0.1527421921491623,
    "learning_rate": 0.0006656689442312858,
    "timestep_mean": 511.25,
    "epsilon_norm": 3.866124391555786,
    "prediction_norm": 3.353952646255493,
    "timestamp": "2025-07-14T12:34:01.339780"
  },
  {
    "epoch": 786,
    "batch_idx": 0,
    "loss": 0.13706254959106445,
    "learning_rate": 0.0006649277074294268,
    "timestep_mean": 542.3125,
    "epsilon_norm": 7.761649131774902,
    "prediction_norm": 6.9619574546813965,
    "timestamp": "2025-07-14T12:34:01.344688"
  },
  {
    "epoch": 786,
    "batch_idx": 32,
    "loss": 0.4732425808906555,
    "learning_rate": 0.0006649277074294268,
    "timestep_mean": 488.46875,
    "epsilon_norm": 10.353922843933105,
    "prediction_norm": 8.150346755981445,
    "timestamp": "2025-07-14T12:34:01.349580"
  },
  {
    "epoch": 786,
    "batch_idx": 64,
    "loss": 0.32700079679489136,
    "learning_rate": 0.0006649277074294268,
    "timestep_mean": 498.0625,
    "epsilon_norm": 8.040029525756836,
    "prediction_norm": 6.024259567260742,
    "timestamp": "2025-07-14T12:34:01.354475"
  },
  {
    "epoch": 786,
    "batch_idx": 96,
    "loss": 0.2935327887535095,
    "learning_rate": 0.0006649277074294268,
    "timestep_mean": 476.59375,
    "epsilon_norm": 7.471320629119873,
    "prediction_norm": 5.888762474060059,
    "timestamp": "2025-07-14T12:34:01.359293"
  },
  {
    "epoch": 786,
    "batch_idx": 128,
    "loss": 0.314216285943985,
    "learning_rate": 0.0006649277074294268,
    "timestep_mean": 552.71875,
    "epsilon_norm": 8.068561553955078,
    "prediction_norm": 6.877972602844238,
    "timestamp": "2025-07-14T12:34:01.364158"
  },
  {
    "epoch": 786,
    "batch_idx": 160,
    "loss": 0.23365351557731628,
    "learning_rate": 0.0006649277074294268,
    "timestep_mean": 482.59375,
    "epsilon_norm": 8.538723945617676,
    "prediction_norm": 7.931885242462158,
    "timestamp": "2025-07-14T12:34:01.369046"
  },
  {
    "epoch": 786,
    "batch_idx": 192,
    "loss": 0.0354723259806633,
    "learning_rate": 0.0006649277074294268,
    "timestep_mean": 763.875,
    "epsilon_norm": 4.510036945343018,
    "prediction_norm": 4.447174549102783,
    "timestamp": "2025-07-14T12:34:01.373985"
  },
  {
    "epoch": 787,
    "batch_idx": 0,
    "loss": 0.6192619800567627,
    "learning_rate": 0.0006641860636848447,
    "timestep_mean": 466.6875,
    "epsilon_norm": 9.89846420288086,
    "prediction_norm": 7.582927703857422,
    "timestamp": "2025-07-14T12:34:01.378907"
  },
  {
    "epoch": 787,
    "batch_idx": 32,
    "loss": 0.18551799654960632,
    "learning_rate": 0.0006641860636848447,
    "timestep_mean": 500.3125,
    "epsilon_norm": 7.599833965301514,
    "prediction_norm": 6.944958209991455,
    "timestamp": "2025-07-14T12:34:01.383848"
  },
  {
    "epoch": 787,
    "batch_idx": 64,
    "loss": 0.14550206065177917,
    "learning_rate": 0.0006641860636848447,
    "timestep_mean": 560.5,
    "epsilon_norm": 8.388447761535645,
    "prediction_norm": 7.609072685241699,
    "timestamp": "2025-07-14T12:34:01.388916"
  },
  {
    "epoch": 787,
    "batch_idx": 96,
    "loss": 0.36899441480636597,
    "learning_rate": 0.0006641860636848447,
    "timestep_mean": 434.6875,
    "epsilon_norm": 8.155628204345703,
    "prediction_norm": 6.375521183013916,
    "timestamp": "2025-07-14T12:34:01.394111"
  },
  {
    "epoch": 787,
    "batch_idx": 128,
    "loss": 0.21950288116931915,
    "learning_rate": 0.0006641860636848447,
    "timestep_mean": 552.90625,
    "epsilon_norm": 7.722672939300537,
    "prediction_norm": 6.93550443649292,
    "timestamp": "2025-07-14T12:34:01.399299"
  },
  {
    "epoch": 787,
    "batch_idx": 160,
    "loss": 0.22738036513328552,
    "learning_rate": 0.0006641860636848447,
    "timestep_mean": 452.9375,
    "epsilon_norm": 7.116744041442871,
    "prediction_norm": 5.260550498962402,
    "timestamp": "2025-07-14T12:34:01.404611"
  },
  {
    "epoch": 787,
    "batch_idx": 192,
    "loss": 0.774534285068512,
    "learning_rate": 0.0006641860636848447,
    "timestep_mean": 444.5,
    "epsilon_norm": 4.481581687927246,
    "prediction_norm": 3.464892625808716,
    "timestamp": "2025-07-14T12:34:01.409502"
  },
  {
    "epoch": 788,
    "batch_idx": 0,
    "loss": 0.46226605772972107,
    "learning_rate": 0.0006634440148274718,
    "timestep_mean": 438.84375,
    "epsilon_norm": 9.51671314239502,
    "prediction_norm": 7.479925155639648,
    "timestamp": "2025-07-14T12:34:01.414425"
  },
  {
    "epoch": 788,
    "batch_idx": 32,
    "loss": 0.2632765769958496,
    "learning_rate": 0.0006634440148274718,
    "timestep_mean": 477.4375,
    "epsilon_norm": 8.188541412353516,
    "prediction_norm": 7.058535099029541,
    "timestamp": "2025-07-14T12:34:01.419653"
  },
  {
    "epoch": 788,
    "batch_idx": 64,
    "loss": 0.2935752272605896,
    "learning_rate": 0.0006634440148274718,
    "timestep_mean": 455.25,
    "epsilon_norm": 7.436379909515381,
    "prediction_norm": 5.815278053283691,
    "timestamp": "2025-07-14T12:34:01.424950"
  },
  {
    "epoch": 788,
    "batch_idx": 96,
    "loss": 0.3650803565979004,
    "learning_rate": 0.0006634440148274718,
    "timestep_mean": 502.34375,
    "epsilon_norm": 8.717979431152344,
    "prediction_norm": 7.658267974853516,
    "timestamp": "2025-07-14T12:34:01.430220"
  },
  {
    "epoch": 788,
    "batch_idx": 128,
    "loss": 0.17716966569423676,
    "learning_rate": 0.0006634440148274718,
    "timestep_mean": 553.25,
    "epsilon_norm": 7.040099143981934,
    "prediction_norm": 6.997079372406006,
    "timestamp": "2025-07-14T12:34:01.435432"
  },
  {
    "epoch": 788,
    "batch_idx": 160,
    "loss": 0.35340601205825806,
    "learning_rate": 0.0006634440148274718,
    "timestep_mean": 417.65625,
    "epsilon_norm": 7.105958938598633,
    "prediction_norm": 5.379652500152588,
    "timestamp": "2025-07-14T12:34:01.440627"
  },
  {
    "epoch": 788,
    "batch_idx": 192,
    "loss": 0.22308452427387238,
    "learning_rate": 0.0006634440148274718,
    "timestep_mean": 594.25,
    "epsilon_norm": 3.3166990280151367,
    "prediction_norm": 2.577749252319336,
    "timestamp": "2025-07-14T12:34:01.445960"
  },
  {
    "epoch": 789,
    "batch_idx": 0,
    "loss": 0.233993798494339,
    "learning_rate": 0.0006627015626882396,
    "timestep_mean": 576.25,
    "epsilon_norm": 8.261990547180176,
    "prediction_norm": 7.552079200744629,
    "timestamp": "2025-07-14T12:34:01.451216"
  },
  {
    "epoch": 789,
    "batch_idx": 32,
    "loss": 0.26625776290893555,
    "learning_rate": 0.0006627015626882396,
    "timestep_mean": 480.125,
    "epsilon_norm": 7.998293399810791,
    "prediction_norm": 6.973394393920898,
    "timestamp": "2025-07-14T12:34:01.456412"
  },
  {
    "epoch": 789,
    "batch_idx": 64,
    "loss": 0.2554161548614502,
    "learning_rate": 0.0006627015626882396,
    "timestep_mean": 508.03125,
    "epsilon_norm": 7.420166015625,
    "prediction_norm": 6.727182865142822,
    "timestamp": "2025-07-14T12:34:01.461605"
  },
  {
    "epoch": 789,
    "batch_idx": 96,
    "loss": 0.36802154779434204,
    "learning_rate": 0.0006627015626882396,
    "timestep_mean": 498.65625,
    "epsilon_norm": 9.071508407592773,
    "prediction_norm": 8.078621864318848,
    "timestamp": "2025-07-14T12:34:01.466800"
  },
  {
    "epoch": 789,
    "batch_idx": 128,
    "loss": 0.3206806480884552,
    "learning_rate": 0.0006627015626882396,
    "timestep_mean": 502.875,
    "epsilon_norm": 8.850257873535156,
    "prediction_norm": 7.4681501388549805,
    "timestamp": "2025-07-14T12:34:01.471884"
  },
  {
    "epoch": 789,
    "batch_idx": 160,
    "loss": 0.33016663789749146,
    "learning_rate": 0.0006627015626882396,
    "timestep_mean": 496.5625,
    "epsilon_norm": 7.9850897789001465,
    "prediction_norm": 7.592645168304443,
    "timestamp": "2025-07-14T12:34:01.477081"
  },
  {
    "epoch": 789,
    "batch_idx": 192,
    "loss": 0.12588265538215637,
    "learning_rate": 0.0006627015626882396,
    "timestep_mean": 475.875,
    "epsilon_norm": 3.7129459381103516,
    "prediction_norm": 3.8340368270874023,
    "timestamp": "2025-07-14T12:34:01.482364"
  },
  {
    "epoch": 790,
    "batch_idx": 0,
    "loss": 0.2995215654373169,
    "learning_rate": 0.0006619587090990752,
    "timestep_mean": 451.40625,
    "epsilon_norm": 7.302713871002197,
    "prediction_norm": 6.271256446838379,
    "timestamp": "2025-07-14T12:34:01.487492"
  },
  {
    "epoch": 790,
    "batch_idx": 32,
    "loss": 0.29362037777900696,
    "learning_rate": 0.0006619587090990752,
    "timestep_mean": 441.78125,
    "epsilon_norm": 7.627053260803223,
    "prediction_norm": 6.540469646453857,
    "timestamp": "2025-07-14T12:34:01.492724"
  },
  {
    "epoch": 790,
    "batch_idx": 64,
    "loss": 0.25242623686790466,
    "learning_rate": 0.0006619587090990752,
    "timestep_mean": 517.53125,
    "epsilon_norm": 7.808633327484131,
    "prediction_norm": 6.403500080108643,
    "timestamp": "2025-07-14T12:34:01.497844"
  },
  {
    "epoch": 790,
    "batch_idx": 96,
    "loss": 0.3332621455192566,
    "learning_rate": 0.0006619587090990752,
    "timestep_mean": 494.78125,
    "epsilon_norm": 8.210613250732422,
    "prediction_norm": 7.0567474365234375,
    "timestamp": "2025-07-14T12:34:01.503667"
  },
  {
    "epoch": 790,
    "batch_idx": 128,
    "loss": 0.4246533513069153,
    "learning_rate": 0.0006619587090990752,
    "timestep_mean": 486.3125,
    "epsilon_norm": 8.960712432861328,
    "prediction_norm": 7.506664276123047,
    "timestamp": "2025-07-14T12:34:01.510814"
  },
  {
    "epoch": 790,
    "batch_idx": 160,
    "loss": 0.19869637489318848,
    "learning_rate": 0.0006619587090990752,
    "timestep_mean": 475.25,
    "epsilon_norm": 7.7992143630981445,
    "prediction_norm": 6.8936991691589355,
    "timestamp": "2025-07-14T12:34:01.515757"
  },
  {
    "epoch": 790,
    "batch_idx": 192,
    "loss": 0.324676513671875,
    "learning_rate": 0.0006619587090990752,
    "timestep_mean": 454.0,
    "epsilon_norm": 3.8521058559417725,
    "prediction_norm": 3.6775028705596924,
    "timestamp": "2025-07-14T12:34:01.520781"
  },
  {
    "epoch": 791,
    "batch_idx": 0,
    "loss": 0.10214442014694214,
    "learning_rate": 0.000661215455892896,
    "timestep_mean": 588.28125,
    "epsilon_norm": 8.385165214538574,
    "prediction_norm": 7.990106105804443,
    "timestamp": "2025-07-14T12:34:01.525664"
  },
  {
    "epoch": 791,
    "batch_idx": 32,
    "loss": 0.3496518135070801,
    "learning_rate": 0.000661215455892896,
    "timestep_mean": 477.8125,
    "epsilon_norm": 7.091405391693115,
    "prediction_norm": 4.9052252769470215,
    "timestamp": "2025-07-14T12:34:01.530688"
  },
  {
    "epoch": 791,
    "batch_idx": 64,
    "loss": 0.3561248779296875,
    "learning_rate": 0.000661215455892896,
    "timestep_mean": 514.59375,
    "epsilon_norm": 8.81529426574707,
    "prediction_norm": 7.42904806137085,
    "timestamp": "2025-07-14T12:34:01.535740"
  },
  {
    "epoch": 791,
    "batch_idx": 96,
    "loss": 0.11696535348892212,
    "learning_rate": 0.000661215455892896,
    "timestep_mean": 541.71875,
    "epsilon_norm": 7.104636192321777,
    "prediction_norm": 6.864995002746582,
    "timestamp": "2025-07-14T12:34:01.540706"
  },
  {
    "epoch": 791,
    "batch_idx": 128,
    "loss": 0.27079638838768005,
    "learning_rate": 0.000661215455892896,
    "timestep_mean": 485.21875,
    "epsilon_norm": 8.126850128173828,
    "prediction_norm": 6.646753311157227,
    "timestamp": "2025-07-14T12:34:01.545606"
  },
  {
    "epoch": 791,
    "batch_idx": 160,
    "loss": 0.477010041475296,
    "learning_rate": 0.000661215455892896,
    "timestep_mean": 493.65625,
    "epsilon_norm": 9.343406677246094,
    "prediction_norm": 6.989210605621338,
    "timestamp": "2025-07-14T12:34:01.550757"
  },
  {
    "epoch": 791,
    "batch_idx": 192,
    "loss": 0.11882926523685455,
    "learning_rate": 0.000661215455892896,
    "timestep_mean": 670.375,
    "epsilon_norm": 6.014462947845459,
    "prediction_norm": 5.377333641052246,
    "timestamp": "2025-07-14T12:34:01.555698"
  },
  {
    "epoch": 792,
    "batch_idx": 0,
    "loss": 0.34948647022247314,
    "learning_rate": 0.0006604718049036051,
    "timestep_mean": 448.625,
    "epsilon_norm": 7.0796003341674805,
    "prediction_norm": 5.000588893890381,
    "timestamp": "2025-07-14T12:34:01.560711"
  },
  {
    "epoch": 792,
    "batch_idx": 32,
    "loss": 0.16437256336212158,
    "learning_rate": 0.0006604718049036051,
    "timestep_mean": 552.125,
    "epsilon_norm": 7.547889232635498,
    "prediction_norm": 6.521944046020508,
    "timestamp": "2025-07-14T12:34:01.565630"
  },
  {
    "epoch": 792,
    "batch_idx": 64,
    "loss": 0.2707712650299072,
    "learning_rate": 0.0006604718049036051,
    "timestep_mean": 442.28125,
    "epsilon_norm": 7.7016401290893555,
    "prediction_norm": 6.48767614364624,
    "timestamp": "2025-07-14T12:34:01.570652"
  },
  {
    "epoch": 792,
    "batch_idx": 96,
    "loss": 0.2615663409233093,
    "learning_rate": 0.0006604718049036051,
    "timestep_mean": 479.21875,
    "epsilon_norm": 8.336822509765625,
    "prediction_norm": 6.761175632476807,
    "timestamp": "2025-07-14T12:34:01.575709"
  },
  {
    "epoch": 792,
    "batch_idx": 128,
    "loss": 0.28606706857681274,
    "learning_rate": 0.0006604718049036051,
    "timestep_mean": 432.875,
    "epsilon_norm": 7.424535274505615,
    "prediction_norm": 5.731436729431152,
    "timestamp": "2025-07-14T12:34:01.580730"
  },
  {
    "epoch": 792,
    "batch_idx": 160,
    "loss": 0.3210112452507019,
    "learning_rate": 0.0006604718049036051,
    "timestep_mean": 395.75,
    "epsilon_norm": 7.5163373947143555,
    "prediction_norm": 5.746405124664307,
    "timestamp": "2025-07-14T12:34:01.585722"
  },
  {
    "epoch": 792,
    "batch_idx": 192,
    "loss": 0.17744001746177673,
    "learning_rate": 0.0006604718049036051,
    "timestep_mean": 625.375,
    "epsilon_norm": 2.5321767330169678,
    "prediction_norm": 2.006746292114258,
    "timestamp": "2025-07-14T12:34:01.590708"
  },
  {
    "epoch": 793,
    "batch_idx": 0,
    "loss": 0.27590253949165344,
    "learning_rate": 0.0006597277579660879,
    "timestep_mean": 451.53125,
    "epsilon_norm": 7.973306655883789,
    "prediction_norm": 6.804870128631592,
    "timestamp": "2025-07-14T12:34:01.595773"
  },
  {
    "epoch": 793,
    "batch_idx": 32,
    "loss": 0.2726467251777649,
    "learning_rate": 0.0006597277579660879,
    "timestep_mean": 512.96875,
    "epsilon_norm": 8.418917655944824,
    "prediction_norm": 6.874958038330078,
    "timestamp": "2025-07-14T12:34:01.600797"
  },
  {
    "epoch": 793,
    "batch_idx": 64,
    "loss": 0.184658944606781,
    "learning_rate": 0.0006597277579660879,
    "timestep_mean": 597.3125,
    "epsilon_norm": 9.194835662841797,
    "prediction_norm": 8.177836418151855,
    "timestamp": "2025-07-14T12:34:01.605925"
  },
  {
    "epoch": 793,
    "batch_idx": 96,
    "loss": 0.35165107250213623,
    "learning_rate": 0.0006597277579660879,
    "timestep_mean": 480.53125,
    "epsilon_norm": 7.945113182067871,
    "prediction_norm": 6.14178466796875,
    "timestamp": "2025-07-14T12:34:01.610951"
  },
  {
    "epoch": 793,
    "batch_idx": 128,
    "loss": 0.19472704827785492,
    "learning_rate": 0.0006597277579660879,
    "timestep_mean": 582.09375,
    "epsilon_norm": 7.898184776306152,
    "prediction_norm": 6.630216121673584,
    "timestamp": "2025-07-14T12:34:01.615958"
  },
  {
    "epoch": 793,
    "batch_idx": 160,
    "loss": 0.2169855535030365,
    "learning_rate": 0.0006597277579660879,
    "timestep_mean": 535.125,
    "epsilon_norm": 7.96765661239624,
    "prediction_norm": 7.208416938781738,
    "timestamp": "2025-07-14T12:34:01.621132"
  },
  {
    "epoch": 793,
    "batch_idx": 192,
    "loss": 0.21053948998451233,
    "learning_rate": 0.0006597277579660879,
    "timestep_mean": 519.25,
    "epsilon_norm": 5.263453483581543,
    "prediction_norm": 4.983222961425781,
    "timestamp": "2025-07-14T12:34:01.626228"
  },
  {
    "epoch": 794,
    "batch_idx": 0,
    "loss": 0.12543250620365143,
    "learning_rate": 0.0006589833169162061,
    "timestep_mean": 581.375,
    "epsilon_norm": 8.151534080505371,
    "prediction_norm": 7.360270023345947,
    "timestamp": "2025-07-14T12:34:01.631274"
  },
  {
    "epoch": 794,
    "batch_idx": 32,
    "loss": 0.4141051173210144,
    "learning_rate": 0.0006589833169162061,
    "timestep_mean": 517.9375,
    "epsilon_norm": 8.295108795166016,
    "prediction_norm": 6.5042524337768555,
    "timestamp": "2025-07-14T12:34:01.636518"
  },
  {
    "epoch": 794,
    "batch_idx": 64,
    "loss": 0.44204181432724,
    "learning_rate": 0.0006589833169162061,
    "timestep_mean": 479.71875,
    "epsilon_norm": 8.820441246032715,
    "prediction_norm": 7.1058268547058105,
    "timestamp": "2025-07-14T12:34:01.641666"
  },
  {
    "epoch": 794,
    "batch_idx": 96,
    "loss": 0.33336925506591797,
    "learning_rate": 0.0006589833169162061,
    "timestep_mean": 510.53125,
    "epsilon_norm": 7.696302890777588,
    "prediction_norm": 6.80722188949585,
    "timestamp": "2025-07-14T12:34:01.646841"
  },
  {
    "epoch": 794,
    "batch_idx": 128,
    "loss": 0.21787463128566742,
    "learning_rate": 0.0006589833169162061,
    "timestep_mean": 567.28125,
    "epsilon_norm": 8.812420845031738,
    "prediction_norm": 8.625842094421387,
    "timestamp": "2025-07-14T12:34:01.652070"
  },
  {
    "epoch": 794,
    "batch_idx": 160,
    "loss": 0.3306763172149658,
    "learning_rate": 0.0006589833169162061,
    "timestep_mean": 479.0,
    "epsilon_norm": 8.099503517150879,
    "prediction_norm": 7.19922399520874,
    "timestamp": "2025-07-14T12:34:01.657305"
  },
  {
    "epoch": 794,
    "batch_idx": 192,
    "loss": 0.12981674075126648,
    "learning_rate": 0.0006589833169162061,
    "timestep_mean": 670.25,
    "epsilon_norm": 3.809814691543579,
    "prediction_norm": 3.7491931915283203,
    "timestamp": "2025-07-14T12:34:01.662452"
  },
  {
    "epoch": 795,
    "batch_idx": 0,
    "loss": 0.2653172016143799,
    "learning_rate": 0.0006582384835907936,
    "timestep_mean": 449.59375,
    "epsilon_norm": 8.081375122070312,
    "prediction_norm": 7.104587554931641,
    "timestamp": "2025-07-14T12:34:01.667519"
  },
  {
    "epoch": 795,
    "batch_idx": 32,
    "loss": 0.23710200190544128,
    "learning_rate": 0.0006582384835907936,
    "timestep_mean": 462.21875,
    "epsilon_norm": 6.960428237915039,
    "prediction_norm": 5.939416408538818,
    "timestamp": "2025-07-14T12:34:01.672678"
  },
  {
    "epoch": 795,
    "batch_idx": 64,
    "loss": 0.43145307898521423,
    "learning_rate": 0.0006582384835907936,
    "timestep_mean": 385.90625,
    "epsilon_norm": 8.193671226501465,
    "prediction_norm": 5.9779791831970215,
    "timestamp": "2025-07-14T12:34:01.677884"
  },
  {
    "epoch": 795,
    "batch_idx": 96,
    "loss": 0.36919182538986206,
    "learning_rate": 0.0006582384835907936,
    "timestep_mean": 454.8125,
    "epsilon_norm": 7.799267292022705,
    "prediction_norm": 5.824078559875488,
    "timestamp": "2025-07-14T12:34:01.683248"
  },
  {
    "epoch": 795,
    "batch_idx": 128,
    "loss": 0.21455547213554382,
    "learning_rate": 0.0006582384835907936,
    "timestep_mean": 498.15625,
    "epsilon_norm": 6.792245864868164,
    "prediction_norm": 6.488816738128662,
    "timestamp": "2025-07-14T12:34:01.688369"
  },
  {
    "epoch": 795,
    "batch_idx": 160,
    "loss": 0.31686702370643616,
    "learning_rate": 0.0006582384835907936,
    "timestep_mean": 479.78125,
    "epsilon_norm": 7.935456275939941,
    "prediction_norm": 7.175462245941162,
    "timestamp": "2025-07-14T12:34:01.693536"
  },
  {
    "epoch": 795,
    "batch_idx": 192,
    "loss": 0.2798501253128052,
    "learning_rate": 0.0006582384835907936,
    "timestep_mean": 407.5,
    "epsilon_norm": 4.198564052581787,
    "prediction_norm": 3.708284854888916,
    "timestamp": "2025-07-14T12:34:01.698776"
  },
  {
    "epoch": 796,
    "batch_idx": 0,
    "loss": 0.1837121695280075,
    "learning_rate": 0.0006574932598276529,
    "timestep_mean": 540.78125,
    "epsilon_norm": 9.053457260131836,
    "prediction_norm": 8.482770919799805,
    "timestamp": "2025-07-14T12:34:01.703945"
  },
  {
    "epoch": 796,
    "batch_idx": 32,
    "loss": 0.2933546006679535,
    "learning_rate": 0.0006574932598276529,
    "timestep_mean": 536.78125,
    "epsilon_norm": 7.747808456420898,
    "prediction_norm": 5.619182109832764,
    "timestamp": "2025-07-14T12:34:01.709145"
  },
  {
    "epoch": 796,
    "batch_idx": 64,
    "loss": 0.5320430994033813,
    "learning_rate": 0.0006574932598276529,
    "timestep_mean": 424.90625,
    "epsilon_norm": 7.891677379608154,
    "prediction_norm": 6.294033050537109,
    "timestamp": "2025-07-14T12:34:01.714376"
  },
  {
    "epoch": 796,
    "batch_idx": 96,
    "loss": 0.26935526728630066,
    "learning_rate": 0.0006574932598276529,
    "timestep_mean": 552.90625,
    "epsilon_norm": 7.910986423492432,
    "prediction_norm": 7.298193454742432,
    "timestamp": "2025-07-14T12:34:01.719705"
  },
  {
    "epoch": 796,
    "batch_idx": 128,
    "loss": 0.40200769901275635,
    "learning_rate": 0.0006574932598276529,
    "timestep_mean": 504.6875,
    "epsilon_norm": 8.407843589782715,
    "prediction_norm": 7.241059303283691,
    "timestamp": "2025-07-14T12:34:01.724950"
  },
  {
    "epoch": 796,
    "batch_idx": 160,
    "loss": 0.17389991879463196,
    "learning_rate": 0.0006574932598276529,
    "timestep_mean": 545.28125,
    "epsilon_norm": 7.513613224029541,
    "prediction_norm": 6.578157901763916,
    "timestamp": "2025-07-14T12:34:01.730183"
  },
  {
    "epoch": 796,
    "batch_idx": 192,
    "loss": 0.5411033034324646,
    "learning_rate": 0.0006574932598276529,
    "timestep_mean": 382.875,
    "epsilon_norm": 4.434426784515381,
    "prediction_norm": 2.7686233520507812,
    "timestamp": "2025-07-14T12:34:01.735286"
  },
  {
    "epoch": 797,
    "batch_idx": 0,
    "loss": 0.3668200969696045,
    "learning_rate": 0.0006567476474655496,
    "timestep_mean": 523.8125,
    "epsilon_norm": 7.9454026222229,
    "prediction_norm": 6.020479679107666,
    "timestamp": "2025-07-14T12:34:01.740457"
  },
  {
    "epoch": 797,
    "batch_idx": 32,
    "loss": 0.39265841245651245,
    "learning_rate": 0.0006567476474655496,
    "timestep_mean": 517.5,
    "epsilon_norm": 7.864656925201416,
    "prediction_norm": 6.695263862609863,
    "timestamp": "2025-07-14T12:34:01.745701"
  },
  {
    "epoch": 797,
    "batch_idx": 64,
    "loss": 0.2870244085788727,
    "learning_rate": 0.0006567476474655496,
    "timestep_mean": 536.625,
    "epsilon_norm": 8.089778900146484,
    "prediction_norm": 6.9312520027160645,
    "timestamp": "2025-07-14T12:34:01.750866"
  },
  {
    "epoch": 797,
    "batch_idx": 96,
    "loss": 0.6045885682106018,
    "learning_rate": 0.0006567476474655496,
    "timestep_mean": 376.9375,
    "epsilon_norm": 8.527566909790039,
    "prediction_norm": 5.152925968170166,
    "timestamp": "2025-07-14T12:34:01.755952"
  },
  {
    "epoch": 797,
    "batch_idx": 128,
    "loss": 0.2831956744194031,
    "learning_rate": 0.0006567476474655496,
    "timestep_mean": 492.03125,
    "epsilon_norm": 7.743203163146973,
    "prediction_norm": 6.158047676086426,
    "timestamp": "2025-07-14T12:34:01.761159"
  },
  {
    "epoch": 797,
    "batch_idx": 160,
    "loss": 0.3076806366443634,
    "learning_rate": 0.0006567476474655496,
    "timestep_mean": 552.09375,
    "epsilon_norm": 8.731287956237793,
    "prediction_norm": 6.622101783752441,
    "timestamp": "2025-07-14T12:34:01.766388"
  },
  {
    "epoch": 797,
    "batch_idx": 192,
    "loss": 0.4197004437446594,
    "learning_rate": 0.0006567476474655496,
    "timestep_mean": 432.5,
    "epsilon_norm": 3.534738540649414,
    "prediction_norm": 2.837307929992676,
    "timestamp": "2025-07-14T12:34:01.771575"
  },
  {
    "epoch": 798,
    "batch_idx": 0,
    "loss": 0.4617311656475067,
    "learning_rate": 0.000656001648344208,
    "timestep_mean": 423.0625,
    "epsilon_norm": 9.092823028564453,
    "prediction_norm": 6.254375457763672,
    "timestamp": "2025-07-14T12:34:01.776817"
  },
  {
    "epoch": 798,
    "batch_idx": 32,
    "loss": 0.5978293418884277,
    "learning_rate": 0.000656001648344208,
    "timestep_mean": 401.15625,
    "epsilon_norm": 8.052434921264648,
    "prediction_norm": 4.8145270347595215,
    "timestamp": "2025-07-14T12:34:01.782009"
  },
  {
    "epoch": 798,
    "batch_idx": 64,
    "loss": 0.21523645520210266,
    "learning_rate": 0.000656001648344208,
    "timestep_mean": 487.75,
    "epsilon_norm": 8.356019973754883,
    "prediction_norm": 7.029616355895996,
    "timestamp": "2025-07-14T12:34:01.787203"
  },
  {
    "epoch": 798,
    "batch_idx": 96,
    "loss": 0.2034142166376114,
    "learning_rate": 0.000656001648344208,
    "timestep_mean": 475.25,
    "epsilon_norm": 7.022299766540527,
    "prediction_norm": 5.42695951461792,
    "timestamp": "2025-07-14T12:34:01.792545"
  },
  {
    "epoch": 798,
    "batch_idx": 128,
    "loss": 0.3017444908618927,
    "learning_rate": 0.000656001648344208,
    "timestep_mean": 503.0,
    "epsilon_norm": 8.25990104675293,
    "prediction_norm": 6.644515037536621,
    "timestamp": "2025-07-14T12:34:01.797729"
  },
  {
    "epoch": 798,
    "batch_idx": 160,
    "loss": 0.3904212713241577,
    "learning_rate": 0.000656001648344208,
    "timestep_mean": 575.5,
    "epsilon_norm": 8.858582496643066,
    "prediction_norm": 6.541060447692871,
    "timestamp": "2025-07-14T12:34:01.803059"
  },
  {
    "epoch": 798,
    "batch_idx": 192,
    "loss": 0.2971939146518707,
    "learning_rate": 0.000656001648344208,
    "timestep_mean": 409.75,
    "epsilon_norm": 3.3150148391723633,
    "prediction_norm": 1.9904652833938599,
    "timestamp": "2025-07-14T12:34:01.808078"
  },
  {
    "epoch": 799,
    "batch_idx": 0,
    "loss": 0.20292583107948303,
    "learning_rate": 0.0006552552643043067,
    "timestep_mean": 524.5625,
    "epsilon_norm": 8.471861839294434,
    "prediction_norm": 7.847574710845947,
    "timestamp": "2025-07-14T12:34:01.813384"
  },
  {
    "epoch": 799,
    "batch_idx": 32,
    "loss": 0.26462382078170776,
    "learning_rate": 0.0006552552643043067,
    "timestep_mean": 500.21875,
    "epsilon_norm": 7.062175273895264,
    "prediction_norm": 6.235843181610107,
    "timestamp": "2025-07-14T12:34:01.818596"
  },
  {
    "epoch": 799,
    "batch_idx": 64,
    "loss": 0.3879489302635193,
    "learning_rate": 0.0006552552643043067,
    "timestep_mean": 463.59375,
    "epsilon_norm": 8.965867042541504,
    "prediction_norm": 7.318151950836182,
    "timestamp": "2025-07-14T12:34:01.823701"
  },
  {
    "epoch": 799,
    "batch_idx": 96,
    "loss": 0.6041929721832275,
    "learning_rate": 0.0006552552643043067,
    "timestep_mean": 461.625,
    "epsilon_norm": 9.395976066589355,
    "prediction_norm": 6.800534248352051,
    "timestamp": "2025-07-14T12:34:01.828635"
  },
  {
    "epoch": 799,
    "batch_idx": 128,
    "loss": 0.3172004222869873,
    "learning_rate": 0.0006552552643043067,
    "timestep_mean": 521.8125,
    "epsilon_norm": 7.425589084625244,
    "prediction_norm": 5.63462495803833,
    "timestamp": "2025-07-14T12:34:01.833806"
  },
  {
    "epoch": 799,
    "batch_idx": 160,
    "loss": 0.4520505368709564,
    "learning_rate": 0.0006552552643043067,
    "timestep_mean": 542.75,
    "epsilon_norm": 8.379383087158203,
    "prediction_norm": 6.769801616668701,
    "timestamp": "2025-07-14T12:34:01.839013"
  },
  {
    "epoch": 799,
    "batch_idx": 192,
    "loss": 0.20439909398555756,
    "learning_rate": 0.0006552552643043067,
    "timestep_mean": 463.75,
    "epsilon_norm": 3.3039674758911133,
    "prediction_norm": 3.1879794597625732,
    "timestamp": "2025-07-14T12:34:01.844289"
  },
  {
    "epoch": 800,
    "batch_idx": 0,
    "loss": 0.21141596138477325,
    "learning_rate": 0.0006545084971874742,
    "timestep_mean": 566.53125,
    "epsilon_norm": 7.909607887268066,
    "prediction_norm": 6.805902004241943,
    "timestamp": "2025-07-14T12:34:01.849433"
  },
  {
    "epoch": 800,
    "batch_idx": 32,
    "loss": 0.2505382001399994,
    "learning_rate": 0.0006545084971874742,
    "timestep_mean": 402.9375,
    "epsilon_norm": 7.444465637207031,
    "prediction_norm": 6.427174091339111,
    "timestamp": "2025-07-14T12:34:01.854733"
  },
  {
    "epoch": 800,
    "batch_idx": 64,
    "loss": 0.19981469213962555,
    "learning_rate": 0.0006545084971874742,
    "timestep_mean": 505.0625,
    "epsilon_norm": 7.503239154815674,
    "prediction_norm": 6.451706409454346,
    "timestamp": "2025-07-14T12:34:01.859919"
  },
  {
    "epoch": 800,
    "batch_idx": 96,
    "loss": 0.1857375055551529,
    "learning_rate": 0.0006545084971874742,
    "timestep_mean": 475.46875,
    "epsilon_norm": 7.240138053894043,
    "prediction_norm": 6.484846115112305,
    "timestamp": "2025-07-14T12:34:01.865157"
  },
  {
    "epoch": 800,
    "batch_idx": 128,
    "loss": 0.31799858808517456,
    "learning_rate": 0.0006545084971874742,
    "timestep_mean": 483.25,
    "epsilon_norm": 9.254694938659668,
    "prediction_norm": 7.787622928619385,
    "timestamp": "2025-07-14T12:34:01.870384"
  },
  {
    "epoch": 800,
    "batch_idx": 160,
    "loss": 0.5012243986129761,
    "learning_rate": 0.0006545084971874742,
    "timestep_mean": 537.09375,
    "epsilon_norm": 8.734164237976074,
    "prediction_norm": 7.344667911529541,
    "timestamp": "2025-07-14T12:34:01.875646"
  },
  {
    "epoch": 800,
    "batch_idx": 192,
    "loss": 0.2821844220161438,
    "learning_rate": 0.0006545084971874742,
    "timestep_mean": 504.5,
    "epsilon_norm": 4.019606113433838,
    "prediction_norm": 4.036633014678955,
    "timestamp": "2025-07-14T12:34:01.880790"
  },
  {
    "epoch": 801,
    "batch_idx": 0,
    "loss": 0.3039488196372986,
    "learning_rate": 0.0006537613488362843,
    "timestep_mean": 441.59375,
    "epsilon_norm": 7.577295303344727,
    "prediction_norm": 5.838149070739746,
    "timestamp": "2025-07-14T12:34:01.886051"
  },
  {
    "epoch": 801,
    "batch_idx": 32,
    "loss": 0.19586369395256042,
    "learning_rate": 0.0006537613488362843,
    "timestep_mean": 556.09375,
    "epsilon_norm": 8.117363929748535,
    "prediction_norm": 7.9423041343688965,
    "timestamp": "2025-07-14T12:34:01.891278"
  },
  {
    "epoch": 801,
    "batch_idx": 64,
    "loss": 0.3229852318763733,
    "learning_rate": 0.0006537613488362843,
    "timestep_mean": 495.75,
    "epsilon_norm": 8.279662132263184,
    "prediction_norm": 8.04849624633789,
    "timestamp": "2025-07-14T12:34:01.896740"
  },
  {
    "epoch": 801,
    "batch_idx": 96,
    "loss": 0.1879810094833374,
    "learning_rate": 0.0006537613488362843,
    "timestep_mean": 469.53125,
    "epsilon_norm": 7.6518473625183105,
    "prediction_norm": 6.791036128997803,
    "timestamp": "2025-07-14T12:34:01.901951"
  },
  {
    "epoch": 801,
    "batch_idx": 128,
    "loss": 0.36609989404678345,
    "learning_rate": 0.0006537613488362843,
    "timestep_mean": 502.3125,
    "epsilon_norm": 8.148956298828125,
    "prediction_norm": 7.464693069458008,
    "timestamp": "2025-07-14T12:34:01.907138"
  },
  {
    "epoch": 801,
    "batch_idx": 160,
    "loss": 0.27841299772262573,
    "learning_rate": 0.0006537613488362843,
    "timestep_mean": 569.59375,
    "epsilon_norm": 7.6288676261901855,
    "prediction_norm": 7.153961181640625,
    "timestamp": "2025-07-14T12:34:01.912304"
  },
  {
    "epoch": 801,
    "batch_idx": 192,
    "loss": 0.2083490788936615,
    "learning_rate": 0.0006537613488362843,
    "timestep_mean": 597.0,
    "epsilon_norm": 3.7347941398620605,
    "prediction_norm": 3.6505656242370605,
    "timestamp": "2025-07-14T12:34:01.917510"
  },
  {
    "epoch": 802,
    "batch_idx": 0,
    "loss": 0.264371782541275,
    "learning_rate": 0.000653013821094251,
    "timestep_mean": 483.21875,
    "epsilon_norm": 8.49468994140625,
    "prediction_norm": 7.379434108734131,
    "timestamp": "2025-07-14T12:34:01.922774"
  },
  {
    "epoch": 802,
    "batch_idx": 32,
    "loss": 0.48243290185928345,
    "learning_rate": 0.000653013821094251,
    "timestep_mean": 428.5625,
    "epsilon_norm": 8.087361335754395,
    "prediction_norm": 6.545325756072998,
    "timestamp": "2025-07-14T12:34:01.928035"
  },
  {
    "epoch": 802,
    "batch_idx": 64,
    "loss": 0.19299568235874176,
    "learning_rate": 0.000653013821094251,
    "timestep_mean": 503.84375,
    "epsilon_norm": 7.920388221740723,
    "prediction_norm": 7.324620723724365,
    "timestamp": "2025-07-14T12:34:01.933472"
  },
  {
    "epoch": 802,
    "batch_idx": 96,
    "loss": 0.253645122051239,
    "learning_rate": 0.000653013821094251,
    "timestep_mean": 534.6875,
    "epsilon_norm": 8.098917007446289,
    "prediction_norm": 6.986657619476318,
    "timestamp": "2025-07-14T12:34:01.938673"
  },
  {
    "epoch": 802,
    "batch_idx": 128,
    "loss": 0.316034197807312,
    "learning_rate": 0.000653013821094251,
    "timestep_mean": 459.1875,
    "epsilon_norm": 7.891958713531494,
    "prediction_norm": 6.70518684387207,
    "timestamp": "2025-07-14T12:34:01.943851"
  },
  {
    "epoch": 802,
    "batch_idx": 160,
    "loss": 0.2808154821395874,
    "learning_rate": 0.000653013821094251,
    "timestep_mean": 484.65625,
    "epsilon_norm": 8.082984924316406,
    "prediction_norm": 6.982532024383545,
    "timestamp": "2025-07-14T12:34:01.949097"
  },
  {
    "epoch": 802,
    "batch_idx": 192,
    "loss": 0.9158711433410645,
    "learning_rate": 0.000653013821094251,
    "timestep_mean": 374.625,
    "epsilon_norm": 4.045973300933838,
    "prediction_norm": 3.166607141494751,
    "timestamp": "2025-07-14T12:34:01.954223"
  },
  {
    "epoch": 803,
    "batch_idx": 0,
    "loss": 0.34871596097946167,
    "learning_rate": 0.0006522659158058247,
    "timestep_mean": 463.625,
    "epsilon_norm": 8.529452323913574,
    "prediction_norm": 6.034793376922607,
    "timestamp": "2025-07-14T12:34:01.959434"
  },
  {
    "epoch": 803,
    "batch_idx": 32,
    "loss": 0.24238087236881256,
    "learning_rate": 0.0006522659158058247,
    "timestep_mean": 504.8125,
    "epsilon_norm": 8.320111274719238,
    "prediction_norm": 6.916386604309082,
    "timestamp": "2025-07-14T12:34:01.964676"
  },
  {
    "epoch": 803,
    "batch_idx": 64,
    "loss": 0.27846387028694153,
    "learning_rate": 0.0006522659158058247,
    "timestep_mean": 505.34375,
    "epsilon_norm": 7.705268383026123,
    "prediction_norm": 6.400585174560547,
    "timestamp": "2025-07-14T12:34:01.969853"
  },
  {
    "epoch": 803,
    "batch_idx": 96,
    "loss": 0.20191690325737,
    "learning_rate": 0.0006522659158058247,
    "timestep_mean": 514.625,
    "epsilon_norm": 8.357216835021973,
    "prediction_norm": 7.308615207672119,
    "timestamp": "2025-07-14T12:34:01.975203"
  },
  {
    "epoch": 803,
    "batch_idx": 128,
    "loss": 0.24871429800987244,
    "learning_rate": 0.0006522659158058247,
    "timestep_mean": 496.875,
    "epsilon_norm": 6.83428430557251,
    "prediction_norm": 5.500020980834961,
    "timestamp": "2025-07-14T12:34:01.980392"
  },
  {
    "epoch": 803,
    "batch_idx": 160,
    "loss": 0.29260513186454773,
    "learning_rate": 0.0006522659158058247,
    "timestep_mean": 470.25,
    "epsilon_norm": 7.929239273071289,
    "prediction_norm": 6.372204780578613,
    "timestamp": "2025-07-14T12:34:01.985581"
  },
  {
    "epoch": 803,
    "batch_idx": 192,
    "loss": 0.09479504823684692,
    "learning_rate": 0.0006522659158058247,
    "timestep_mean": 541.125,
    "epsilon_norm": 4.623711109161377,
    "prediction_norm": 4.017772674560547,
    "timestamp": "2025-07-14T12:34:01.990756"
  },
  {
    "epoch": 804,
    "batch_idx": 0,
    "loss": 0.5184312462806702,
    "learning_rate": 0.0006515176348163874,
    "timestep_mean": 498.15625,
    "epsilon_norm": 7.61539363861084,
    "prediction_norm": 5.254964351654053,
    "timestamp": "2025-07-14T12:34:01.995982"
  },
  {
    "epoch": 804,
    "batch_idx": 32,
    "loss": 0.20042824745178223,
    "learning_rate": 0.0006515176348163874,
    "timestep_mean": 552.5625,
    "epsilon_norm": 6.97430419921875,
    "prediction_norm": 5.806369304656982,
    "timestamp": "2025-07-14T12:34:02.001146"
  },
  {
    "epoch": 804,
    "batch_idx": 64,
    "loss": 0.20348519086837769,
    "learning_rate": 0.0006515176348163874,
    "timestep_mean": 503.125,
    "epsilon_norm": 6.30039119720459,
    "prediction_norm": 5.153372764587402,
    "timestamp": "2025-07-14T12:34:02.006377"
  },
  {
    "epoch": 804,
    "batch_idx": 96,
    "loss": 0.2896466851234436,
    "learning_rate": 0.0006515176348163874,
    "timestep_mean": 515.5625,
    "epsilon_norm": 7.857169151306152,
    "prediction_norm": 6.091118335723877,
    "timestamp": "2025-07-14T12:34:02.011694"
  },
  {
    "epoch": 804,
    "batch_idx": 128,
    "loss": 0.23317669332027435,
    "learning_rate": 0.0006515176348163874,
    "timestep_mean": 592.46875,
    "epsilon_norm": 8.15446949005127,
    "prediction_norm": 6.6567487716674805,
    "timestamp": "2025-07-14T12:34:02.017036"
  },
  {
    "epoch": 804,
    "batch_idx": 160,
    "loss": 0.2592674493789673,
    "learning_rate": 0.0006515176348163874,
    "timestep_mean": 487.96875,
    "epsilon_norm": 7.86294412612915,
    "prediction_norm": 6.179494857788086,
    "timestamp": "2025-07-14T12:34:02.022143"
  },
  {
    "epoch": 804,
    "batch_idx": 192,
    "loss": 0.7348623275756836,
    "learning_rate": 0.0006515176348163874,
    "timestep_mean": 531.375,
    "epsilon_norm": 3.911478281021118,
    "prediction_norm": 2.699941873550415,
    "timestamp": "2025-07-14T12:34:02.027272"
  },
  {
    "epoch": 805,
    "batch_idx": 0,
    "loss": 0.6148914098739624,
    "learning_rate": 0.0006507689799722483,
    "timestep_mean": 472.46875,
    "epsilon_norm": 8.5003662109375,
    "prediction_norm": 6.162399768829346,
    "timestamp": "2025-07-14T12:34:02.032577"
  },
  {
    "epoch": 805,
    "batch_idx": 32,
    "loss": 0.3163471817970276,
    "learning_rate": 0.0006507689799722483,
    "timestep_mean": 561.46875,
    "epsilon_norm": 9.322014808654785,
    "prediction_norm": 7.3235650062561035,
    "timestamp": "2025-07-14T12:34:02.037945"
  },
  {
    "epoch": 805,
    "batch_idx": 64,
    "loss": 0.14311820268630981,
    "learning_rate": 0.0006507689799722483,
    "timestep_mean": 551.40625,
    "epsilon_norm": 7.459771633148193,
    "prediction_norm": 6.5079169273376465,
    "timestamp": "2025-07-14T12:34:02.043158"
  },
  {
    "epoch": 805,
    "batch_idx": 96,
    "loss": 0.2830383777618408,
    "learning_rate": 0.0006507689799722483,
    "timestep_mean": 540.0625,
    "epsilon_norm": 6.9063825607299805,
    "prediction_norm": 5.445450782775879,
    "timestamp": "2025-07-14T12:34:02.048344"
  },
  {
    "epoch": 805,
    "batch_idx": 128,
    "loss": 0.2597595453262329,
    "learning_rate": 0.0006507689799722483,
    "timestep_mean": 529.71875,
    "epsilon_norm": 8.195950508117676,
    "prediction_norm": 6.799923419952393,
    "timestamp": "2025-07-14T12:34:02.053570"
  },
  {
    "epoch": 805,
    "batch_idx": 160,
    "loss": 0.29470014572143555,
    "learning_rate": 0.0006507689799722483,
    "timestep_mean": 513.53125,
    "epsilon_norm": 8.043654441833496,
    "prediction_norm": 6.505262851715088,
    "timestamp": "2025-07-14T12:34:02.058893"
  },
  {
    "epoch": 805,
    "batch_idx": 192,
    "loss": 0.3806828558444977,
    "learning_rate": 0.0006507689799722483,
    "timestep_mean": 487.75,
    "epsilon_norm": 3.688227415084839,
    "prediction_norm": 2.652770519256592,
    "timestamp": "2025-07-14T12:34:02.063847"
  },
  {
    "epoch": 806,
    "batch_idx": 0,
    "loss": 0.2914409935474396,
    "learning_rate": 0.0006500199531206387,
    "timestep_mean": 470.65625,
    "epsilon_norm": 7.920267581939697,
    "prediction_norm": 7.009460926055908,
    "timestamp": "2025-07-14T12:34:02.068730"
  },
  {
    "epoch": 806,
    "batch_idx": 32,
    "loss": 0.5138649940490723,
    "learning_rate": 0.0006500199531206387,
    "timestep_mean": 457.1875,
    "epsilon_norm": 7.986525535583496,
    "prediction_norm": 5.720874786376953,
    "timestamp": "2025-07-14T12:34:02.073541"
  },
  {
    "epoch": 806,
    "batch_idx": 64,
    "loss": 0.27848345041275024,
    "learning_rate": 0.0006500199531206387,
    "timestep_mean": 467.4375,
    "epsilon_norm": 8.045854568481445,
    "prediction_norm": 7.567841529846191,
    "timestamp": "2025-07-14T12:34:02.078350"
  },
  {
    "epoch": 806,
    "batch_idx": 96,
    "loss": 0.4912572503089905,
    "learning_rate": 0.0006500199531206387,
    "timestep_mean": 403.40625,
    "epsilon_norm": 8.521934509277344,
    "prediction_norm": 6.219618320465088,
    "timestamp": "2025-07-14T12:34:02.083249"
  },
  {
    "epoch": 806,
    "batch_idx": 128,
    "loss": 0.28925079107284546,
    "learning_rate": 0.0006500199531206387,
    "timestep_mean": 535.0625,
    "epsilon_norm": 8.834184646606445,
    "prediction_norm": 7.630832195281982,
    "timestamp": "2025-07-14T12:34:02.088140"
  },
  {
    "epoch": 806,
    "batch_idx": 160,
    "loss": 0.17330992221832275,
    "learning_rate": 0.0006500199531206387,
    "timestep_mean": 575.125,
    "epsilon_norm": 8.460824966430664,
    "prediction_norm": 8.569632530212402,
    "timestamp": "2025-07-14T12:34:02.092962"
  },
  {
    "epoch": 806,
    "batch_idx": 192,
    "loss": 0.19926488399505615,
    "learning_rate": 0.0006500199531206387,
    "timestep_mean": 488.625,
    "epsilon_norm": 3.4770870208740234,
    "prediction_norm": 2.926938772201538,
    "timestamp": "2025-07-14T12:34:02.097795"
  },
  {
    "epoch": 807,
    "batch_idx": 0,
    "loss": 0.1980466991662979,
    "learning_rate": 0.0006492705561097076,
    "timestep_mean": 497.9375,
    "epsilon_norm": 8.129674911499023,
    "prediction_norm": 7.785244464874268,
    "timestamp": "2025-07-14T12:34:02.103785"
  },
  {
    "epoch": 807,
    "batch_idx": 32,
    "loss": 0.62508624792099,
    "learning_rate": 0.0006492705561097076,
    "timestep_mean": 418.21875,
    "epsilon_norm": 9.560147285461426,
    "prediction_norm": 7.460549354553223,
    "timestamp": "2025-07-14T12:34:02.111806"
  },
  {
    "epoch": 807,
    "batch_idx": 64,
    "loss": 0.26681432127952576,
    "learning_rate": 0.0006492705561097076,
    "timestep_mean": 499.96875,
    "epsilon_norm": 7.8274359703063965,
    "prediction_norm": 7.737584114074707,
    "timestamp": "2025-07-14T12:34:02.116643"
  },
  {
    "epoch": 807,
    "batch_idx": 96,
    "loss": 0.2194390445947647,
    "learning_rate": 0.0006492705561097076,
    "timestep_mean": 484.15625,
    "epsilon_norm": 8.134075164794922,
    "prediction_norm": 7.496742248535156,
    "timestamp": "2025-07-14T12:34:02.121332"
  },
  {
    "epoch": 807,
    "batch_idx": 128,
    "loss": 0.21653953194618225,
    "learning_rate": 0.0006492705561097076,
    "timestep_mean": 488.21875,
    "epsilon_norm": 7.1511311531066895,
    "prediction_norm": 6.216290473937988,
    "timestamp": "2025-07-14T12:34:02.125996"
  },
  {
    "epoch": 807,
    "batch_idx": 160,
    "loss": 0.4121299386024475,
    "learning_rate": 0.0006492705561097076,
    "timestep_mean": 532.9375,
    "epsilon_norm": 8.203878402709961,
    "prediction_norm": 6.673356056213379,
    "timestamp": "2025-07-14T12:34:02.130725"
  },
  {
    "epoch": 807,
    "batch_idx": 192,
    "loss": 1.1171455383300781,
    "learning_rate": 0.0006492705561097076,
    "timestep_mean": 331.875,
    "epsilon_norm": 5.5659027099609375,
    "prediction_norm": 2.636416435241699,
    "timestamp": "2025-07-14T12:34:02.135375"
  },
  {
    "epoch": 808,
    "batch_idx": 0,
    "loss": 0.4166867136955261,
    "learning_rate": 0.0006485207907885179,
    "timestep_mean": 448.625,
    "epsilon_norm": 8.076786994934082,
    "prediction_norm": 6.189014434814453,
    "timestamp": "2025-07-14T12:34:02.140095"
  },
  {
    "epoch": 808,
    "batch_idx": 32,
    "loss": 0.1450384259223938,
    "learning_rate": 0.0006485207907885179,
    "timestep_mean": 496.5,
    "epsilon_norm": 8.023625373840332,
    "prediction_norm": 7.079674243927002,
    "timestamp": "2025-07-14T12:34:02.144748"
  },
  {
    "epoch": 808,
    "batch_idx": 64,
    "loss": 0.22281666100025177,
    "learning_rate": 0.0006485207907885179,
    "timestep_mean": 444.40625,
    "epsilon_norm": 7.8275580406188965,
    "prediction_norm": 6.77864408493042,
    "timestamp": "2025-07-14T12:34:02.149402"
  },
  {
    "epoch": 808,
    "batch_idx": 96,
    "loss": 0.23483321070671082,
    "learning_rate": 0.0006485207907885179,
    "timestep_mean": 551.4375,
    "epsilon_norm": 6.8005595207214355,
    "prediction_norm": 5.62003231048584,
    "timestamp": "2025-07-14T12:34:02.154232"
  },
  {
    "epoch": 808,
    "batch_idx": 128,
    "loss": 0.21918576955795288,
    "learning_rate": 0.0006485207907885179,
    "timestep_mean": 492.78125,
    "epsilon_norm": 6.992062091827393,
    "prediction_norm": 5.676955699920654,
    "timestamp": "2025-07-14T12:34:02.158899"
  },
  {
    "epoch": 808,
    "batch_idx": 160,
    "loss": 0.23040050268173218,
    "learning_rate": 0.0006485207907885179,
    "timestep_mean": 518.5,
    "epsilon_norm": 6.895893096923828,
    "prediction_norm": 5.945461273193359,
    "timestamp": "2025-07-14T12:34:02.163553"
  },
  {
    "epoch": 808,
    "batch_idx": 192,
    "loss": 0.16008810698986053,
    "learning_rate": 0.0006485207907885179,
    "timestep_mean": 669.375,
    "epsilon_norm": 3.2616593837738037,
    "prediction_norm": 2.8207476139068604,
    "timestamp": "2025-07-14T12:34:02.168195"
  },
  {
    "epoch": 809,
    "batch_idx": 0,
    "loss": 0.3543316721916199,
    "learning_rate": 0.000647770659007041,
    "timestep_mean": 449.9375,
    "epsilon_norm": 7.568753242492676,
    "prediction_norm": 5.339426040649414,
    "timestamp": "2025-07-14T12:34:02.172901"
  },
  {
    "epoch": 809,
    "batch_idx": 32,
    "loss": 0.40054404735565186,
    "learning_rate": 0.000647770659007041,
    "timestep_mean": 393.03125,
    "epsilon_norm": 6.992609977722168,
    "prediction_norm": 5.2926225662231445,
    "timestamp": "2025-07-14T12:34:02.177642"
  },
  {
    "epoch": 809,
    "batch_idx": 64,
    "loss": 0.3537752032279968,
    "learning_rate": 0.000647770659007041,
    "timestep_mean": 503.625,
    "epsilon_norm": 7.932925224304199,
    "prediction_norm": 6.430522918701172,
    "timestamp": "2025-07-14T12:34:02.182316"
  },
  {
    "epoch": 809,
    "batch_idx": 96,
    "loss": 0.11531755328178406,
    "learning_rate": 0.000647770659007041,
    "timestep_mean": 601.90625,
    "epsilon_norm": 7.502579689025879,
    "prediction_norm": 7.321255207061768,
    "timestamp": "2025-07-14T12:34:02.187053"
  },
  {
    "epoch": 809,
    "batch_idx": 128,
    "loss": 0.18880745768547058,
    "learning_rate": 0.000647770659007041,
    "timestep_mean": 531.75,
    "epsilon_norm": 6.324785232543945,
    "prediction_norm": 5.197040557861328,
    "timestamp": "2025-07-14T12:34:02.191778"
  },
  {
    "epoch": 809,
    "batch_idx": 160,
    "loss": 0.41208210587501526,
    "learning_rate": 0.000647770659007041,
    "timestep_mean": 428.21875,
    "epsilon_norm": 7.681751728057861,
    "prediction_norm": 6.036311626434326,
    "timestamp": "2025-07-14T12:34:02.196476"
  },
  {
    "epoch": 809,
    "batch_idx": 192,
    "loss": 0.3014373183250427,
    "learning_rate": 0.000647770659007041,
    "timestep_mean": 479.625,
    "epsilon_norm": 5.298951625823975,
    "prediction_norm": 4.8196845054626465,
    "timestamp": "2025-07-14T12:34:02.201147"
  },
  {
    "epoch": 810,
    "batch_idx": 0,
    "loss": 0.15581881999969482,
    "learning_rate": 0.0006470201626161524,
    "timestep_mean": 600.15625,
    "epsilon_norm": 7.696084499359131,
    "prediction_norm": 6.754247188568115,
    "timestamp": "2025-07-14T12:34:02.205934"
  },
  {
    "epoch": 810,
    "batch_idx": 32,
    "loss": 0.16043147444725037,
    "learning_rate": 0.0006470201626161524,
    "timestep_mean": 551.34375,
    "epsilon_norm": 6.827479362487793,
    "prediction_norm": 6.081067085266113,
    "timestamp": "2025-07-14T12:34:02.210673"
  },
  {
    "epoch": 810,
    "batch_idx": 64,
    "loss": 0.3063507676124573,
    "learning_rate": 0.0006470201626161524,
    "timestep_mean": 429.25,
    "epsilon_norm": 8.183406829833984,
    "prediction_norm": 6.80269718170166,
    "timestamp": "2025-07-14T12:34:02.215414"
  },
  {
    "epoch": 810,
    "batch_idx": 96,
    "loss": 0.4350782632827759,
    "learning_rate": 0.0006470201626161524,
    "timestep_mean": 508.03125,
    "epsilon_norm": 9.455588340759277,
    "prediction_norm": 7.511597633361816,
    "timestamp": "2025-07-14T12:34:02.220140"
  },
  {
    "epoch": 810,
    "batch_idx": 128,
    "loss": 0.24228443205356598,
    "learning_rate": 0.0006470201626161524,
    "timestep_mean": 435.0625,
    "epsilon_norm": 8.25490951538086,
    "prediction_norm": 8.024225234985352,
    "timestamp": "2025-07-14T12:34:02.224937"
  },
  {
    "epoch": 810,
    "batch_idx": 160,
    "loss": 0.33246099948883057,
    "learning_rate": 0.0006470201626161524,
    "timestep_mean": 472.75,
    "epsilon_norm": 7.230940818786621,
    "prediction_norm": 5.744312763214111,
    "timestamp": "2025-07-14T12:34:02.229743"
  },
  {
    "epoch": 810,
    "batch_idx": 192,
    "loss": 0.08441894501447678,
    "learning_rate": 0.0006470201626161524,
    "timestep_mean": 579.0,
    "epsilon_norm": 4.698655605316162,
    "prediction_norm": 4.689125061035156,
    "timestamp": "2025-07-14T12:34:02.234520"
  },
  {
    "epoch": 811,
    "batch_idx": 0,
    "loss": 0.3104667067527771,
    "learning_rate": 0.0006462693034676274,
    "timestep_mean": 532.0,
    "epsilon_norm": 7.968264579772949,
    "prediction_norm": 6.311712741851807,
    "timestamp": "2025-07-14T12:34:02.239299"
  },
  {
    "epoch": 811,
    "batch_idx": 32,
    "loss": 0.18210026621818542,
    "learning_rate": 0.0006462693034676274,
    "timestep_mean": 558.59375,
    "epsilon_norm": 7.892883777618408,
    "prediction_norm": 7.263368129730225,
    "timestamp": "2025-07-14T12:34:02.244166"
  },
  {
    "epoch": 811,
    "batch_idx": 64,
    "loss": 0.35653263330459595,
    "learning_rate": 0.0006462693034676274,
    "timestep_mean": 508.875,
    "epsilon_norm": 8.980022430419922,
    "prediction_norm": 7.373626708984375,
    "timestamp": "2025-07-14T12:34:02.248958"
  },
  {
    "epoch": 811,
    "batch_idx": 96,
    "loss": 0.5353724956512451,
    "learning_rate": 0.0006462693034676274,
    "timestep_mean": 458.71875,
    "epsilon_norm": 9.125024795532227,
    "prediction_norm": 6.741667747497559,
    "timestamp": "2025-07-14T12:34:02.253803"
  },
  {
    "epoch": 811,
    "batch_idx": 128,
    "loss": 0.1835031807422638,
    "learning_rate": 0.0006462693034676274,
    "timestep_mean": 430.0625,
    "epsilon_norm": 7.1335883140563965,
    "prediction_norm": 6.504061698913574,
    "timestamp": "2025-07-14T12:34:02.258691"
  },
  {
    "epoch": 811,
    "batch_idx": 160,
    "loss": 0.19237935543060303,
    "learning_rate": 0.0006462693034676274,
    "timestep_mean": 411.8125,
    "epsilon_norm": 7.138692855834961,
    "prediction_norm": 6.23099422454834,
    "timestamp": "2025-07-14T12:34:02.263526"
  },
  {
    "epoch": 811,
    "batch_idx": 192,
    "loss": 0.22783717513084412,
    "learning_rate": 0.0006462693034676274,
    "timestep_mean": 424.125,
    "epsilon_norm": 3.244101047515869,
    "prediction_norm": 2.6274774074554443,
    "timestamp": "2025-07-14T12:34:02.268404"
  },
  {
    "epoch": 812,
    "batch_idx": 0,
    "loss": 0.151508629322052,
    "learning_rate": 0.0006455180834141364,
    "timestep_mean": 612.4375,
    "epsilon_norm": 7.488368511199951,
    "prediction_norm": 6.860332012176514,
    "timestamp": "2025-07-14T12:34:02.273234"
  },
  {
    "epoch": 812,
    "batch_idx": 32,
    "loss": 0.33326607942581177,
    "learning_rate": 0.0006455180834141364,
    "timestep_mean": 470.84375,
    "epsilon_norm": 9.3839750289917,
    "prediction_norm": 8.363147735595703,
    "timestamp": "2025-07-14T12:34:02.278091"
  },
  {
    "epoch": 812,
    "batch_idx": 64,
    "loss": 0.318783164024353,
    "learning_rate": 0.0006455180834141364,
    "timestep_mean": 555.75,
    "epsilon_norm": 8.161530494689941,
    "prediction_norm": 6.732138156890869,
    "timestamp": "2025-07-14T12:34:02.282989"
  },
  {
    "epoch": 812,
    "batch_idx": 96,
    "loss": 0.25124427676200867,
    "learning_rate": 0.0006455180834141364,
    "timestep_mean": 448.65625,
    "epsilon_norm": 6.623193740844727,
    "prediction_norm": 6.535636901855469,
    "timestamp": "2025-07-14T12:34:02.288126"
  },
  {
    "epoch": 812,
    "batch_idx": 128,
    "loss": 0.1620347648859024,
    "learning_rate": 0.0006455180834141364,
    "timestep_mean": 534.125,
    "epsilon_norm": 8.4708251953125,
    "prediction_norm": 7.752806186676025,
    "timestamp": "2025-07-14T12:34:02.293332"
  },
  {
    "epoch": 812,
    "batch_idx": 160,
    "loss": 0.22174830734729767,
    "learning_rate": 0.0006455180834141364,
    "timestep_mean": 481.1875,
    "epsilon_norm": 7.686550617218018,
    "prediction_norm": 6.721316337585449,
    "timestamp": "2025-07-14T12:34:02.298490"
  },
  {
    "epoch": 812,
    "batch_idx": 192,
    "loss": 0.20101912319660187,
    "learning_rate": 0.0006455180834141364,
    "timestep_mean": 637.625,
    "epsilon_norm": 4.807067394256592,
    "prediction_norm": 4.309231281280518,
    "timestamp": "2025-07-14T12:34:02.304928"
  },
  {
    "epoch": 813,
    "batch_idx": 0,
    "loss": 0.15336382389068604,
    "learning_rate": 0.00064476650430924,
    "timestep_mean": 492.6875,
    "epsilon_norm": 7.4829936027526855,
    "prediction_norm": 7.2814860343933105,
    "timestamp": "2025-07-14T12:34:02.310110"
  },
  {
    "epoch": 813,
    "batch_idx": 32,
    "loss": 0.2768949568271637,
    "learning_rate": 0.00064476650430924,
    "timestep_mean": 479.4375,
    "epsilon_norm": 8.207890510559082,
    "prediction_norm": 7.117623805999756,
    "timestamp": "2025-07-14T12:34:02.315191"
  },
  {
    "epoch": 813,
    "batch_idx": 64,
    "loss": 0.32839539647102356,
    "learning_rate": 0.00064476650430924,
    "timestep_mean": 486.9375,
    "epsilon_norm": 7.539921760559082,
    "prediction_norm": 6.175673961639404,
    "timestamp": "2025-07-14T12:34:02.320330"
  },
  {
    "epoch": 813,
    "batch_idx": 96,
    "loss": 0.17830930650234222,
    "learning_rate": 0.00064476650430924,
    "timestep_mean": 445.65625,
    "epsilon_norm": 8.39754867553711,
    "prediction_norm": 7.952093124389648,
    "timestamp": "2025-07-14T12:34:02.325617"
  },
  {
    "epoch": 813,
    "batch_idx": 128,
    "loss": 0.183661550283432,
    "learning_rate": 0.00064476650430924,
    "timestep_mean": 508.1875,
    "epsilon_norm": 6.857404708862305,
    "prediction_norm": 5.237751483917236,
    "timestamp": "2025-07-14T12:34:02.330679"
  },
  {
    "epoch": 813,
    "batch_idx": 160,
    "loss": 0.46664929389953613,
    "learning_rate": 0.00064476650430924,
    "timestep_mean": 438.6875,
    "epsilon_norm": 8.748891830444336,
    "prediction_norm": 6.0976948738098145,
    "timestamp": "2025-07-14T12:34:02.335679"
  },
  {
    "epoch": 813,
    "batch_idx": 192,
    "loss": 0.32512369751930237,
    "learning_rate": 0.00064476650430924,
    "timestep_mean": 619.125,
    "epsilon_norm": 5.348898887634277,
    "prediction_norm": 4.748053550720215,
    "timestamp": "2025-07-14T12:34:02.340789"
  },
  {
    "epoch": 814,
    "batch_idx": 0,
    "loss": 0.26724275946617126,
    "learning_rate": 0.000644014568007385,
    "timestep_mean": 552.34375,
    "epsilon_norm": 8.008853912353516,
    "prediction_norm": 6.996530055999756,
    "timestamp": "2025-07-14T12:34:02.345980"
  },
  {
    "epoch": 814,
    "batch_idx": 32,
    "loss": 0.3598681092262268,
    "learning_rate": 0.000644014568007385,
    "timestep_mean": 522.53125,
    "epsilon_norm": 8.139256477355957,
    "prediction_norm": 6.011812210083008,
    "timestamp": "2025-07-14T12:34:02.351270"
  },
  {
    "epoch": 814,
    "batch_idx": 64,
    "loss": 0.4615197777748108,
    "learning_rate": 0.000644014568007385,
    "timestep_mean": 490.71875,
    "epsilon_norm": 8.755722045898438,
    "prediction_norm": 6.217837810516357,
    "timestamp": "2025-07-14T12:34:02.356439"
  },
  {
    "epoch": 814,
    "batch_idx": 96,
    "loss": 0.2366911768913269,
    "learning_rate": 0.000644014568007385,
    "timestep_mean": 458.03125,
    "epsilon_norm": 8.483570098876953,
    "prediction_norm": 7.40541934967041,
    "timestamp": "2025-07-14T12:34:02.361531"
  },
  {
    "epoch": 814,
    "batch_idx": 128,
    "loss": 0.2488025575876236,
    "learning_rate": 0.000644014568007385,
    "timestep_mean": 416.21875,
    "epsilon_norm": 8.094294548034668,
    "prediction_norm": 6.712893486022949,
    "timestamp": "2025-07-14T12:34:02.366687"
  },
  {
    "epoch": 814,
    "batch_idx": 160,
    "loss": 0.1950247883796692,
    "learning_rate": 0.000644014568007385,
    "timestep_mean": 537.65625,
    "epsilon_norm": 9.69729232788086,
    "prediction_norm": 8.669432640075684,
    "timestamp": "2025-07-14T12:34:02.371791"
  },
  {
    "epoch": 814,
    "batch_idx": 192,
    "loss": 0.2531575560569763,
    "learning_rate": 0.000644014568007385,
    "timestep_mean": 646.125,
    "epsilon_norm": 3.491222620010376,
    "prediction_norm": 3.5081284046173096,
    "timestamp": "2025-07-14T12:34:02.376757"
  },
  {
    "epoch": 815,
    "batch_idx": 0,
    "loss": 0.26010650396347046,
    "learning_rate": 0.0006432622763638996,
    "timestep_mean": 543.0625,
    "epsilon_norm": 7.592711448669434,
    "prediction_norm": 6.79888916015625,
    "timestamp": "2025-07-14T12:34:02.381591"
  },
  {
    "epoch": 815,
    "batch_idx": 32,
    "loss": 0.29123908281326294,
    "learning_rate": 0.0006432622763638996,
    "timestep_mean": 497.09375,
    "epsilon_norm": 7.895913600921631,
    "prediction_norm": 7.308589458465576,
    "timestamp": "2025-07-14T12:34:02.386385"
  },
  {
    "epoch": 815,
    "batch_idx": 64,
    "loss": 0.27073630690574646,
    "learning_rate": 0.0006432622763638996,
    "timestep_mean": 453.75,
    "epsilon_norm": 7.587109088897705,
    "prediction_norm": 6.294862747192383,
    "timestamp": "2025-07-14T12:34:02.391279"
  },
  {
    "epoch": 815,
    "batch_idx": 96,
    "loss": 0.31636562943458557,
    "learning_rate": 0.0006432622763638996,
    "timestep_mean": 505.75,
    "epsilon_norm": 9.413114547729492,
    "prediction_norm": 8.007403373718262,
    "timestamp": "2025-07-14T12:34:02.396302"
  },
  {
    "epoch": 815,
    "batch_idx": 128,
    "loss": 0.13350394368171692,
    "learning_rate": 0.0006432622763638996,
    "timestep_mean": 562.28125,
    "epsilon_norm": 7.30978536605835,
    "prediction_norm": 7.492107391357422,
    "timestamp": "2025-07-14T12:34:02.401146"
  },
  {
    "epoch": 815,
    "batch_idx": 160,
    "loss": 0.3181757926940918,
    "learning_rate": 0.0006432622763638996,
    "timestep_mean": 504.75,
    "epsilon_norm": 7.355043411254883,
    "prediction_norm": 6.58603048324585,
    "timestamp": "2025-07-14T12:34:02.406124"
  },
  {
    "epoch": 815,
    "batch_idx": 192,
    "loss": 0.1962514966726303,
    "learning_rate": 0.0006432622763638996,
    "timestep_mean": 542.125,
    "epsilon_norm": 4.4612321853637695,
    "prediction_norm": 3.8171985149383545,
    "timestamp": "2025-07-14T12:34:02.410992"
  },
  {
    "epoch": 816,
    "batch_idx": 0,
    "loss": 0.33695918321609497,
    "learning_rate": 0.0006425096312349885,
    "timestep_mean": 546.59375,
    "epsilon_norm": 10.18409538269043,
    "prediction_norm": 8.381742477416992,
    "timestamp": "2025-07-14T12:34:02.415988"
  },
  {
    "epoch": 816,
    "batch_idx": 32,
    "loss": 0.5668501853942871,
    "learning_rate": 0.0006425096312349885,
    "timestep_mean": 532.3125,
    "epsilon_norm": 8.532628059387207,
    "prediction_norm": 5.986341953277588,
    "timestamp": "2025-07-14T12:34:02.421151"
  },
  {
    "epoch": 816,
    "batch_idx": 64,
    "loss": 0.13129106163978577,
    "learning_rate": 0.0006425096312349885,
    "timestep_mean": 559.5625,
    "epsilon_norm": 7.220187664031982,
    "prediction_norm": 6.789241313934326,
    "timestamp": "2025-07-14T12:34:02.426085"
  },
  {
    "epoch": 816,
    "batch_idx": 96,
    "loss": 0.2088651955127716,
    "learning_rate": 0.0006425096312349885,
    "timestep_mean": 541.78125,
    "epsilon_norm": 9.26222038269043,
    "prediction_norm": 7.9471821784973145,
    "timestamp": "2025-07-14T12:34:02.431011"
  },
  {
    "epoch": 816,
    "batch_idx": 128,
    "loss": 0.2834419906139374,
    "learning_rate": 0.0006425096312349885,
    "timestep_mean": 518.28125,
    "epsilon_norm": 6.964008808135986,
    "prediction_norm": 4.977497577667236,
    "timestamp": "2025-07-14T12:34:02.435946"
  },
  {
    "epoch": 816,
    "batch_idx": 160,
    "loss": 0.12897494435310364,
    "learning_rate": 0.0006425096312349885,
    "timestep_mean": 588.4375,
    "epsilon_norm": 8.034679412841797,
    "prediction_norm": 7.287677764892578,
    "timestamp": "2025-07-14T12:34:02.440856"
  },
  {
    "epoch": 816,
    "batch_idx": 192,
    "loss": 0.6466503143310547,
    "learning_rate": 0.0006425096312349885,
    "timestep_mean": 437.625,
    "epsilon_norm": 4.162496566772461,
    "prediction_norm": 2.791588544845581,
    "timestamp": "2025-07-14T12:34:02.445829"
  },
  {
    "epoch": 817,
    "batch_idx": 0,
    "loss": 0.311972975730896,
    "learning_rate": 0.000641756634477729,
    "timestep_mean": 492.375,
    "epsilon_norm": 7.575748443603516,
    "prediction_norm": 5.845742702484131,
    "timestamp": "2025-07-14T12:34:02.450715"
  },
  {
    "epoch": 817,
    "batch_idx": 32,
    "loss": 0.19542455673217773,
    "learning_rate": 0.000641756634477729,
    "timestep_mean": 512.59375,
    "epsilon_norm": 7.43574857711792,
    "prediction_norm": 6.698965072631836,
    "timestamp": "2025-07-14T12:34:02.455549"
  },
  {
    "epoch": 817,
    "batch_idx": 64,
    "loss": 0.5876585245132446,
    "learning_rate": 0.000641756634477729,
    "timestep_mean": 406.09375,
    "epsilon_norm": 8.678873062133789,
    "prediction_norm": 5.308974266052246,
    "timestamp": "2025-07-14T12:34:02.460438"
  },
  {
    "epoch": 817,
    "batch_idx": 96,
    "loss": 0.2411951720714569,
    "learning_rate": 0.000641756634477729,
    "timestep_mean": 455.5,
    "epsilon_norm": 7.338451385498047,
    "prediction_norm": 6.120917320251465,
    "timestamp": "2025-07-14T12:34:02.465328"
  },
  {
    "epoch": 817,
    "batch_idx": 128,
    "loss": 0.3699111342430115,
    "learning_rate": 0.000641756634477729,
    "timestep_mean": 447.09375,
    "epsilon_norm": 8.064077377319336,
    "prediction_norm": 6.32972526550293,
    "timestamp": "2025-07-14T12:34:02.470206"
  },
  {
    "epoch": 817,
    "batch_idx": 160,
    "loss": 0.29707610607147217,
    "learning_rate": 0.000641756634477729,
    "timestep_mean": 449.0,
    "epsilon_norm": 8.162906646728516,
    "prediction_norm": 6.969390869140625,
    "timestamp": "2025-07-14T12:34:02.475116"
  },
  {
    "epoch": 817,
    "batch_idx": 192,
    "loss": 0.10941049456596375,
    "learning_rate": 0.000641756634477729,
    "timestep_mean": 460.25,
    "epsilon_norm": 2.3230865001678467,
    "prediction_norm": 1.9388740062713623,
    "timestamp": "2025-07-14T12:34:02.480039"
  },
  {
    "epoch": 818,
    "batch_idx": 0,
    "loss": 0.1587219387292862,
    "learning_rate": 0.0006410032879500652,
    "timestep_mean": 497.0,
    "epsilon_norm": 7.766716480255127,
    "prediction_norm": 7.7886176109313965,
    "timestamp": "2025-07-14T12:34:02.484890"
  },
  {
    "epoch": 818,
    "batch_idx": 32,
    "loss": 0.35897427797317505,
    "learning_rate": 0.0006410032879500652,
    "timestep_mean": 405.15625,
    "epsilon_norm": 7.953383445739746,
    "prediction_norm": 6.112918376922607,
    "timestamp": "2025-07-14T12:34:02.489801"
  },
  {
    "epoch": 818,
    "batch_idx": 64,
    "loss": 0.2246755063533783,
    "learning_rate": 0.0006410032879500652,
    "timestep_mean": 518.96875,
    "epsilon_norm": 9.302214622497559,
    "prediction_norm": 9.252885818481445,
    "timestamp": "2025-07-14T12:34:02.494698"
  },
  {
    "epoch": 818,
    "batch_idx": 96,
    "loss": 0.3506840467453003,
    "learning_rate": 0.0006410032879500652,
    "timestep_mean": 454.8125,
    "epsilon_norm": 8.253729820251465,
    "prediction_norm": 6.959712028503418,
    "timestamp": "2025-07-14T12:34:02.499636"
  },
  {
    "epoch": 818,
    "batch_idx": 128,
    "loss": 0.15138250589370728,
    "learning_rate": 0.0006410032879500652,
    "timestep_mean": 443.25,
    "epsilon_norm": 7.2521491050720215,
    "prediction_norm": 7.054692268371582,
    "timestamp": "2025-07-14T12:34:02.504599"
  },
  {
    "epoch": 818,
    "batch_idx": 160,
    "loss": 0.2232079803943634,
    "learning_rate": 0.0006410032879500652,
    "timestep_mean": 545.96875,
    "epsilon_norm": 7.897329330444336,
    "prediction_norm": 7.6661505699157715,
    "timestamp": "2025-07-14T12:34:02.509432"
  },
  {
    "epoch": 818,
    "batch_idx": 192,
    "loss": 1.1934984922409058,
    "learning_rate": 0.0006410032879500652,
    "timestep_mean": 368.75,
    "epsilon_norm": 4.948817253112793,
    "prediction_norm": 2.558048963546753,
    "timestamp": "2025-07-14T12:34:02.514298"
  },
  {
    "epoch": 819,
    "batch_idx": 0,
    "loss": 0.2708583474159241,
    "learning_rate": 0.0006402495935108052,
    "timestep_mean": 490.59375,
    "epsilon_norm": 8.519153594970703,
    "prediction_norm": 7.467766284942627,
    "timestamp": "2025-07-14T12:34:02.519202"
  },
  {
    "epoch": 819,
    "batch_idx": 32,
    "loss": 0.2864284813404083,
    "learning_rate": 0.0006402495935108052,
    "timestep_mean": 517.46875,
    "epsilon_norm": 6.725096225738525,
    "prediction_norm": 5.260704517364502,
    "timestamp": "2025-07-14T12:34:02.524115"
  },
  {
    "epoch": 819,
    "batch_idx": 64,
    "loss": 0.48208358883857727,
    "learning_rate": 0.0006402495935108052,
    "timestep_mean": 547.09375,
    "epsilon_norm": 8.891767501831055,
    "prediction_norm": 6.812483310699463,
    "timestamp": "2025-07-14T12:34:02.529020"
  },
  {
    "epoch": 819,
    "batch_idx": 96,
    "loss": 0.15951675176620483,
    "learning_rate": 0.0006402495935108052,
    "timestep_mean": 574.0625,
    "epsilon_norm": 7.995936393737793,
    "prediction_norm": 7.290929317474365,
    "timestamp": "2025-07-14T12:34:02.534003"
  },
  {
    "epoch": 819,
    "batch_idx": 128,
    "loss": 0.3488968014717102,
    "learning_rate": 0.0006402495935108052,
    "timestep_mean": 402.71875,
    "epsilon_norm": 7.3036603927612305,
    "prediction_norm": 6.004515171051025,
    "timestamp": "2025-07-14T12:34:02.538860"
  },
  {
    "epoch": 819,
    "batch_idx": 160,
    "loss": 0.21593411266803741,
    "learning_rate": 0.0006402495935108052,
    "timestep_mean": 545.96875,
    "epsilon_norm": 7.815529823303223,
    "prediction_norm": 7.157112121582031,
    "timestamp": "2025-07-14T12:34:02.543695"
  },
  {
    "epoch": 819,
    "batch_idx": 192,
    "loss": 0.013356588780879974,
    "learning_rate": 0.0006402495935108052,
    "timestep_mean": 711.0,
    "epsilon_norm": 3.7887909412384033,
    "prediction_norm": 3.766507387161255,
    "timestamp": "2025-07-14T12:34:02.548539"
  },
  {
    "epoch": 820,
    "batch_idx": 0,
    "loss": 0.19320419430732727,
    "learning_rate": 0.0006394955530196151,
    "timestep_mean": 522.75,
    "epsilon_norm": 7.5504913330078125,
    "prediction_norm": 6.675360202789307,
    "timestamp": "2025-07-14T12:34:02.553453"
  },
  {
    "epoch": 820,
    "batch_idx": 32,
    "loss": 0.24963238835334778,
    "learning_rate": 0.0006394955530196151,
    "timestep_mean": 449.09375,
    "epsilon_norm": 6.688511848449707,
    "prediction_norm": 5.961224555969238,
    "timestamp": "2025-07-14T12:34:02.558415"
  },
  {
    "epoch": 820,
    "batch_idx": 64,
    "loss": 0.22993320226669312,
    "learning_rate": 0.0006394955530196151,
    "timestep_mean": 470.84375,
    "epsilon_norm": 7.746649742126465,
    "prediction_norm": 6.968139171600342,
    "timestamp": "2025-07-14T12:34:02.563334"
  },
  {
    "epoch": 820,
    "batch_idx": 96,
    "loss": 0.16654053330421448,
    "learning_rate": 0.0006394955530196151,
    "timestep_mean": 452.53125,
    "epsilon_norm": 7.524998188018799,
    "prediction_norm": 6.586517333984375,
    "timestamp": "2025-07-14T12:34:02.568217"
  },
  {
    "epoch": 820,
    "batch_idx": 128,
    "loss": 0.3652299642562866,
    "learning_rate": 0.0006394955530196151,
    "timestep_mean": 480.34375,
    "epsilon_norm": 8.773778915405273,
    "prediction_norm": 7.417643070220947,
    "timestamp": "2025-07-14T12:34:02.573105"
  },
  {
    "epoch": 820,
    "batch_idx": 160,
    "loss": 0.29314935207366943,
    "learning_rate": 0.0006394955530196151,
    "timestep_mean": 544.9375,
    "epsilon_norm": 8.18967056274414,
    "prediction_norm": 6.396864414215088,
    "timestamp": "2025-07-14T12:34:02.578000"
  },
  {
    "epoch": 820,
    "batch_idx": 192,
    "loss": 0.4122251868247986,
    "learning_rate": 0.0006394955530196151,
    "timestep_mean": 502.25,
    "epsilon_norm": 4.6176276206970215,
    "prediction_norm": 3.638521671295166,
    "timestamp": "2025-07-14T12:34:02.582817"
  },
  {
    "epoch": 821,
    "batch_idx": 0,
    "loss": 0.40188461542129517,
    "learning_rate": 0.0006387411683370148,
    "timestep_mean": 529.75,
    "epsilon_norm": 7.8805646896362305,
    "prediction_norm": 5.337419509887695,
    "timestamp": "2025-07-14T12:34:02.587777"
  },
  {
    "epoch": 821,
    "batch_idx": 32,
    "loss": 0.3610841631889343,
    "learning_rate": 0.0006387411683370148,
    "timestep_mean": 491.1875,
    "epsilon_norm": 7.307289123535156,
    "prediction_norm": 5.602691650390625,
    "timestamp": "2025-07-14T12:34:02.592631"
  },
  {
    "epoch": 821,
    "batch_idx": 64,
    "loss": 0.3306385278701782,
    "learning_rate": 0.0006387411683370148,
    "timestep_mean": 524.34375,
    "epsilon_norm": 8.379877090454102,
    "prediction_norm": 6.911374568939209,
    "timestamp": "2025-07-14T12:34:02.597586"
  },
  {
    "epoch": 821,
    "batch_idx": 96,
    "loss": 0.17059390246868134,
    "learning_rate": 0.0006387411683370148,
    "timestep_mean": 490.4375,
    "epsilon_norm": 7.659287929534912,
    "prediction_norm": 6.698213577270508,
    "timestamp": "2025-07-14T12:34:02.602480"
  },
  {
    "epoch": 821,
    "batch_idx": 128,
    "loss": 0.29491540789604187,
    "learning_rate": 0.0006387411683370148,
    "timestep_mean": 481.84375,
    "epsilon_norm": 8.012574195861816,
    "prediction_norm": 6.724267482757568,
    "timestamp": "2025-07-14T12:34:02.607365"
  },
  {
    "epoch": 821,
    "batch_idx": 160,
    "loss": 0.2916371822357178,
    "learning_rate": 0.0006387411683370148,
    "timestep_mean": 462.21875,
    "epsilon_norm": 7.7078537940979,
    "prediction_norm": 6.078029632568359,
    "timestamp": "2025-07-14T12:34:02.612283"
  },
  {
    "epoch": 821,
    "batch_idx": 192,
    "loss": 0.5327258706092834,
    "learning_rate": 0.0006387411683370148,
    "timestep_mean": 475.5,
    "epsilon_norm": 4.41227388381958,
    "prediction_norm": 2.93615984916687,
    "timestamp": "2025-07-14T12:34:02.617177"
  },
  {
    "epoch": 822,
    "batch_idx": 0,
    "loss": 0.1785198450088501,
    "learning_rate": 0.0006379864413243733,
    "timestep_mean": 562.75,
    "epsilon_norm": 7.677400588989258,
    "prediction_norm": 6.943445205688477,
    "timestamp": "2025-07-14T12:34:02.622137"
  },
  {
    "epoch": 822,
    "batch_idx": 32,
    "loss": 0.22177261114120483,
    "learning_rate": 0.0006379864413243733,
    "timestep_mean": 553.84375,
    "epsilon_norm": 8.04407024383545,
    "prediction_norm": 6.933412551879883,
    "timestamp": "2025-07-14T12:34:02.627021"
  },
  {
    "epoch": 822,
    "batch_idx": 64,
    "loss": 0.3485771715641022,
    "learning_rate": 0.0006379864413243733,
    "timestep_mean": 506.0625,
    "epsilon_norm": 7.951786994934082,
    "prediction_norm": 6.433498859405518,
    "timestamp": "2025-07-14T12:34:02.631986"
  },
  {
    "epoch": 822,
    "batch_idx": 96,
    "loss": 0.22222068905830383,
    "learning_rate": 0.0006379864413243733,
    "timestep_mean": 375.9375,
    "epsilon_norm": 8.99329662322998,
    "prediction_norm": 7.87267541885376,
    "timestamp": "2025-07-14T12:34:02.637005"
  },
  {
    "epoch": 822,
    "batch_idx": 128,
    "loss": 0.27526870369911194,
    "learning_rate": 0.0006379864413243733,
    "timestep_mean": 431.5625,
    "epsilon_norm": 7.896076202392578,
    "prediction_norm": 6.315722942352295,
    "timestamp": "2025-07-14T12:34:02.642004"
  },
  {
    "epoch": 822,
    "batch_idx": 160,
    "loss": 0.38050350546836853,
    "learning_rate": 0.0006379864413243733,
    "timestep_mean": 501.90625,
    "epsilon_norm": 8.676177978515625,
    "prediction_norm": 7.009548187255859,
    "timestamp": "2025-07-14T12:34:02.646862"
  },
  {
    "epoch": 822,
    "batch_idx": 192,
    "loss": 0.20846056938171387,
    "learning_rate": 0.0006379864413243733,
    "timestep_mean": 620.5,
    "epsilon_norm": 4.594844818115234,
    "prediction_norm": 3.74619722366333,
    "timestamp": "2025-07-14T12:34:02.651792"
  },
  {
    "epoch": 823,
    "batch_idx": 0,
    "loss": 0.5443356037139893,
    "learning_rate": 0.0006372313738439048,
    "timestep_mean": 338.625,
    "epsilon_norm": 8.379352569580078,
    "prediction_norm": 5.9245381355285645,
    "timestamp": "2025-07-14T12:34:02.656724"
  },
  {
    "epoch": 823,
    "batch_idx": 32,
    "loss": 0.39905846118927,
    "learning_rate": 0.0006372313738439048,
    "timestep_mean": 483.53125,
    "epsilon_norm": 8.958727836608887,
    "prediction_norm": 7.633352756500244,
    "timestamp": "2025-07-14T12:34:02.661577"
  },
  {
    "epoch": 823,
    "batch_idx": 64,
    "loss": 0.4771541357040405,
    "learning_rate": 0.0006372313738439048,
    "timestep_mean": 481.1875,
    "epsilon_norm": 8.367022514343262,
    "prediction_norm": 6.289551258087158,
    "timestamp": "2025-07-14T12:34:02.666416"
  },
  {
    "epoch": 823,
    "batch_idx": 96,
    "loss": 0.29061007499694824,
    "learning_rate": 0.0006372313738439048,
    "timestep_mean": 423.875,
    "epsilon_norm": 7.987207412719727,
    "prediction_norm": 6.854248046875,
    "timestamp": "2025-07-14T12:34:02.671274"
  },
  {
    "epoch": 823,
    "batch_idx": 128,
    "loss": 0.2376582771539688,
    "learning_rate": 0.0006372313738439048,
    "timestep_mean": 447.75,
    "epsilon_norm": 7.081725120544434,
    "prediction_norm": 6.504429340362549,
    "timestamp": "2025-07-14T12:34:02.676120"
  },
  {
    "epoch": 823,
    "batch_idx": 160,
    "loss": 0.3551504611968994,
    "learning_rate": 0.0006372313738439048,
    "timestep_mean": 484.75,
    "epsilon_norm": 8.095494270324707,
    "prediction_norm": 6.641974925994873,
    "timestamp": "2025-07-14T12:34:02.681035"
  },
  {
    "epoch": 823,
    "batch_idx": 192,
    "loss": 0.7836132645606995,
    "learning_rate": 0.0006372313738439048,
    "timestep_mean": 366.75,
    "epsilon_norm": 4.072246074676514,
    "prediction_norm": 2.454413890838623,
    "timestamp": "2025-07-14T12:34:02.685904"
  },
  {
    "epoch": 824,
    "batch_idx": 0,
    "loss": 0.16562160849571228,
    "learning_rate": 0.0006364759677586631,
    "timestep_mean": 534.96875,
    "epsilon_norm": 6.423922061920166,
    "prediction_norm": 6.091685771942139,
    "timestamp": "2025-07-14T12:34:02.690792"
  },
  {
    "epoch": 824,
    "batch_idx": 32,
    "loss": 0.3069342374801636,
    "learning_rate": 0.0006364759677586631,
    "timestep_mean": 440.21875,
    "epsilon_norm": 8.312809944152832,
    "prediction_norm": 7.5287394523620605,
    "timestamp": "2025-07-14T12:34:02.695714"
  },
  {
    "epoch": 824,
    "batch_idx": 64,
    "loss": 0.16363544762134552,
    "learning_rate": 0.0006364759677586631,
    "timestep_mean": 493.21875,
    "epsilon_norm": 8.002480506896973,
    "prediction_norm": 7.247742176055908,
    "timestamp": "2025-07-14T12:34:02.700688"
  },
  {
    "epoch": 824,
    "batch_idx": 96,
    "loss": 0.3660391867160797,
    "learning_rate": 0.0006364759677586631,
    "timestep_mean": 482.28125,
    "epsilon_norm": 7.8850908279418945,
    "prediction_norm": 6.942089080810547,
    "timestamp": "2025-07-14T12:34:02.709723"
  },
  {
    "epoch": 824,
    "batch_idx": 128,
    "loss": 0.2445056438446045,
    "learning_rate": 0.0006364759677586631,
    "timestep_mean": 560.25,
    "epsilon_norm": 7.208876132965088,
    "prediction_norm": 6.510278224945068,
    "timestamp": "2025-07-14T12:34:02.714450"
  },
  {
    "epoch": 824,
    "batch_idx": 160,
    "loss": 0.21278561651706696,
    "learning_rate": 0.0006364759677586631,
    "timestep_mean": 520.90625,
    "epsilon_norm": 7.71278715133667,
    "prediction_norm": 6.683032512664795,
    "timestamp": "2025-07-14T12:34:02.719148"
  },
  {
    "epoch": 824,
    "batch_idx": 192,
    "loss": 0.2789517641067505,
    "learning_rate": 0.0006364759677586631,
    "timestep_mean": 467.75,
    "epsilon_norm": 3.6787636280059814,
    "prediction_norm": 3.683598756790161,
    "timestamp": "2025-07-14T12:34:02.723822"
  },
  {
    "epoch": 825,
    "batch_idx": 0,
    "loss": 0.2773803472518921,
    "learning_rate": 0.0006357202249325375,
    "timestep_mean": 470.59375,
    "epsilon_norm": 8.05951976776123,
    "prediction_norm": 7.067568778991699,
    "timestamp": "2025-07-14T12:34:02.728530"
  },
  {
    "epoch": 825,
    "batch_idx": 32,
    "loss": 0.1976465880870819,
    "learning_rate": 0.0006357202249325375,
    "timestep_mean": 567.71875,
    "epsilon_norm": 8.894318580627441,
    "prediction_norm": 7.989524841308594,
    "timestamp": "2025-07-14T12:34:02.733334"
  },
  {
    "epoch": 825,
    "batch_idx": 64,
    "loss": 0.3212294280529022,
    "learning_rate": 0.0006357202249325375,
    "timestep_mean": 446.9375,
    "epsilon_norm": 8.356119155883789,
    "prediction_norm": 7.446075439453125,
    "timestamp": "2025-07-14T12:34:02.738033"
  },
  {
    "epoch": 825,
    "batch_idx": 96,
    "loss": 0.4167447090148926,
    "learning_rate": 0.0006357202249325375,
    "timestep_mean": 454.71875,
    "epsilon_norm": 8.865126609802246,
    "prediction_norm": 7.639034748077393,
    "timestamp": "2025-07-14T12:34:02.742719"
  },
  {
    "epoch": 825,
    "batch_idx": 128,
    "loss": 0.23025420308113098,
    "learning_rate": 0.0006357202249325375,
    "timestep_mean": 443.625,
    "epsilon_norm": 7.223183631896973,
    "prediction_norm": 6.213339805603027,
    "timestamp": "2025-07-14T12:34:02.747459"
  },
  {
    "epoch": 825,
    "batch_idx": 160,
    "loss": 0.22061477601528168,
    "learning_rate": 0.0006357202249325375,
    "timestep_mean": 561.6875,
    "epsilon_norm": 6.576729774475098,
    "prediction_norm": 6.171441078186035,
    "timestamp": "2025-07-14T12:34:02.752155"
  },
  {
    "epoch": 825,
    "batch_idx": 192,
    "loss": 0.35068878531455994,
    "learning_rate": 0.0006357202249325375,
    "timestep_mean": 496.625,
    "epsilon_norm": 3.7774062156677246,
    "prediction_norm": 2.8433127403259277,
    "timestamp": "2025-07-14T12:34:02.756830"
  },
  {
    "epoch": 826,
    "batch_idx": 0,
    "loss": 0.21346689760684967,
    "learning_rate": 0.0006349641472302485,
    "timestep_mean": 578.75,
    "epsilon_norm": 7.012495040893555,
    "prediction_norm": 6.396857261657715,
    "timestamp": "2025-07-14T12:34:02.761612"
  },
  {
    "epoch": 826,
    "batch_idx": 32,
    "loss": 0.3197977542877197,
    "learning_rate": 0.0006349641472302485,
    "timestep_mean": 481.65625,
    "epsilon_norm": 7.625764846801758,
    "prediction_norm": 5.9224748611450195,
    "timestamp": "2025-07-14T12:34:02.766315"
  },
  {
    "epoch": 826,
    "batch_idx": 64,
    "loss": 0.3588228225708008,
    "learning_rate": 0.0006349641472302485,
    "timestep_mean": 454.125,
    "epsilon_norm": 7.222875118255615,
    "prediction_norm": 5.3250861167907715,
    "timestamp": "2025-07-14T12:34:02.771237"
  },
  {
    "epoch": 826,
    "batch_idx": 96,
    "loss": 0.3875557780265808,
    "learning_rate": 0.0006349641472302485,
    "timestep_mean": 407.625,
    "epsilon_norm": 7.927419185638428,
    "prediction_norm": 5.702438831329346,
    "timestamp": "2025-07-14T12:34:02.775955"
  },
  {
    "epoch": 826,
    "batch_idx": 128,
    "loss": 0.26835083961486816,
    "learning_rate": 0.0006349641472302485,
    "timestep_mean": 486.9375,
    "epsilon_norm": 7.572254180908203,
    "prediction_norm": 5.961175441741943,
    "timestamp": "2025-07-14T12:34:02.780687"
  },
  {
    "epoch": 826,
    "batch_idx": 160,
    "loss": 0.20575745403766632,
    "learning_rate": 0.0006349641472302485,
    "timestep_mean": 518.625,
    "epsilon_norm": 6.687437057495117,
    "prediction_norm": 5.373758792877197,
    "timestamp": "2025-07-14T12:34:02.785425"
  },
  {
    "epoch": 826,
    "batch_idx": 192,
    "loss": 0.3732065260410309,
    "learning_rate": 0.0006349641472302485,
    "timestep_mean": 617.25,
    "epsilon_norm": 3.9662551879882812,
    "prediction_norm": 2.0848896503448486,
    "timestamp": "2025-07-14T12:34:02.790093"
  },
  {
    "epoch": 827,
    "batch_idx": 0,
    "loss": 0.28168562054634094,
    "learning_rate": 0.0006342077365173428,
    "timestep_mean": 509.25,
    "epsilon_norm": 7.403321743011475,
    "prediction_norm": 5.564582824707031,
    "timestamp": "2025-07-14T12:34:02.794865"
  },
  {
    "epoch": 827,
    "batch_idx": 32,
    "loss": 0.21758896112442017,
    "learning_rate": 0.0006342077365173428,
    "timestep_mean": 512.0625,
    "epsilon_norm": 7.473426342010498,
    "prediction_norm": 6.598607063293457,
    "timestamp": "2025-07-14T12:34:02.799892"
  },
  {
    "epoch": 827,
    "batch_idx": 64,
    "loss": 0.21875162422657013,
    "learning_rate": 0.0006342077365173428,
    "timestep_mean": 521.90625,
    "epsilon_norm": 8.519909858703613,
    "prediction_norm": 7.1257100105285645,
    "timestamp": "2025-07-14T12:34:02.804906"
  },
  {
    "epoch": 827,
    "batch_idx": 96,
    "loss": 0.3635767102241516,
    "learning_rate": 0.0006342077365173428,
    "timestep_mean": 483.96875,
    "epsilon_norm": 7.467823505401611,
    "prediction_norm": 6.05072021484375,
    "timestamp": "2025-07-14T12:34:02.809804"
  },
  {
    "epoch": 827,
    "batch_idx": 128,
    "loss": 0.15618523955345154,
    "learning_rate": 0.0006342077365173428,
    "timestep_mean": 527.15625,
    "epsilon_norm": 7.874871253967285,
    "prediction_norm": 6.8318915367126465,
    "timestamp": "2025-07-14T12:34:02.814876"
  },
  {
    "epoch": 827,
    "batch_idx": 160,
    "loss": 0.16727614402770996,
    "learning_rate": 0.0006342077365173428,
    "timestep_mean": 518.71875,
    "epsilon_norm": 7.513646602630615,
    "prediction_norm": 6.558372497558594,
    "timestamp": "2025-07-14T12:34:02.819869"
  },
  {
    "epoch": 827,
    "batch_idx": 192,
    "loss": 0.26387813687324524,
    "learning_rate": 0.0006342077365173428,
    "timestep_mean": 401.75,
    "epsilon_norm": 3.1885995864868164,
    "prediction_norm": 2.212268114089966,
    "timestamp": "2025-07-14T12:34:02.825025"
  },
  {
    "epoch": 828,
    "batch_idx": 0,
    "loss": 0.3363601565361023,
    "learning_rate": 0.0006334509946601883,
    "timestep_mean": 413.34375,
    "epsilon_norm": 7.556926250457764,
    "prediction_norm": 5.725842475891113,
    "timestamp": "2025-07-14T12:34:02.829993"
  },
  {
    "epoch": 828,
    "batch_idx": 32,
    "loss": 0.34826430678367615,
    "learning_rate": 0.0006334509946601883,
    "timestep_mean": 484.0625,
    "epsilon_norm": 7.8097310066223145,
    "prediction_norm": 6.096954822540283,
    "timestamp": "2025-07-14T12:34:02.835054"
  },
  {
    "epoch": 828,
    "batch_idx": 64,
    "loss": 0.2714642882347107,
    "learning_rate": 0.0006334509946601883,
    "timestep_mean": 498.8125,
    "epsilon_norm": 6.846990585327148,
    "prediction_norm": 6.0207624435424805,
    "timestamp": "2025-07-14T12:34:02.839976"
  },
  {
    "epoch": 828,
    "batch_idx": 96,
    "loss": 0.38672658801078796,
    "learning_rate": 0.0006334509946601883,
    "timestep_mean": 509.15625,
    "epsilon_norm": 6.8332295417785645,
    "prediction_norm": 5.311321258544922,
    "timestamp": "2025-07-14T12:34:02.844988"
  },
  {
    "epoch": 828,
    "batch_idx": 128,
    "loss": 0.20780298113822937,
    "learning_rate": 0.0006334509946601883,
    "timestep_mean": 560.4375,
    "epsilon_norm": 8.527536392211914,
    "prediction_norm": 6.96732234954834,
    "timestamp": "2025-07-14T12:34:02.850176"
  },
  {
    "epoch": 828,
    "batch_idx": 160,
    "loss": 0.26318278908729553,
    "learning_rate": 0.0006334509946601883,
    "timestep_mean": 470.46875,
    "epsilon_norm": 9.469664573669434,
    "prediction_norm": 8.26744556427002,
    "timestamp": "2025-07-14T12:34:02.855332"
  },
  {
    "epoch": 828,
    "batch_idx": 192,
    "loss": 0.18876001238822937,
    "learning_rate": 0.0006334509946601883,
    "timestep_mean": 451.25,
    "epsilon_norm": 3.5207481384277344,
    "prediction_norm": 2.84879207611084,
    "timestamp": "2025-07-14T12:34:02.860532"
  },
  {
    "epoch": 829,
    "batch_idx": 0,
    "loss": 0.3018895387649536,
    "learning_rate": 0.0006326939235259704,
    "timestep_mean": 523.71875,
    "epsilon_norm": 9.193563461303711,
    "prediction_norm": 7.21894645690918,
    "timestamp": "2025-07-14T12:34:02.865775"
  },
  {
    "epoch": 829,
    "batch_idx": 32,
    "loss": 0.36592787504196167,
    "learning_rate": 0.0006326939235259704,
    "timestep_mean": 489.0625,
    "epsilon_norm": 7.891282081604004,
    "prediction_norm": 6.498189449310303,
    "timestamp": "2025-07-14T12:34:02.870914"
  },
  {
    "epoch": 829,
    "batch_idx": 64,
    "loss": 0.18634559214115143,
    "learning_rate": 0.0006326939235259704,
    "timestep_mean": 415.78125,
    "epsilon_norm": 7.331772804260254,
    "prediction_norm": 6.041213512420654,
    "timestamp": "2025-07-14T12:34:02.875895"
  },
  {
    "epoch": 829,
    "batch_idx": 96,
    "loss": 0.516633152961731,
    "learning_rate": 0.0006326939235259704,
    "timestep_mean": 453.78125,
    "epsilon_norm": 8.000077247619629,
    "prediction_norm": 5.957062721252441,
    "timestamp": "2025-07-14T12:34:02.881016"
  },
  {
    "epoch": 829,
    "batch_idx": 128,
    "loss": 0.29918980598449707,
    "learning_rate": 0.0006326939235259704,
    "timestep_mean": 421.46875,
    "epsilon_norm": 8.438643455505371,
    "prediction_norm": 7.1167216300964355,
    "timestamp": "2025-07-14T12:34:02.886099"
  },
  {
    "epoch": 829,
    "batch_idx": 160,
    "loss": 0.30085816979408264,
    "learning_rate": 0.0006326939235259704,
    "timestep_mean": 518.34375,
    "epsilon_norm": 7.736724376678467,
    "prediction_norm": 6.9163818359375,
    "timestamp": "2025-07-14T12:34:02.891146"
  },
  {
    "epoch": 829,
    "batch_idx": 192,
    "loss": 0.266649454832077,
    "learning_rate": 0.0006326939235259704,
    "timestep_mean": 355.375,
    "epsilon_norm": 3.9602630138397217,
    "prediction_norm": 3.679025888442993,
    "timestamp": "2025-07-14T12:34:02.896638"
  },
  {
    "epoch": 830,
    "batch_idx": 0,
    "loss": 0.21110491454601288,
    "learning_rate": 0.0006319365249826869,
    "timestep_mean": 583.75,
    "epsilon_norm": 7.817829608917236,
    "prediction_norm": 6.436022758483887,
    "timestamp": "2025-07-14T12:34:02.901766"
  },
  {
    "epoch": 830,
    "batch_idx": 32,
    "loss": 0.46539437770843506,
    "learning_rate": 0.0006319365249826869,
    "timestep_mean": 486.96875,
    "epsilon_norm": 8.503002166748047,
    "prediction_norm": 7.489010334014893,
    "timestamp": "2025-07-14T12:34:02.907067"
  },
  {
    "epoch": 830,
    "batch_idx": 64,
    "loss": 0.2471422553062439,
    "learning_rate": 0.0006319365249826869,
    "timestep_mean": 482.8125,
    "epsilon_norm": 8.035091400146484,
    "prediction_norm": 7.687289714813232,
    "timestamp": "2025-07-14T12:34:02.912231"
  },
  {
    "epoch": 830,
    "batch_idx": 96,
    "loss": 0.2622174322605133,
    "learning_rate": 0.0006319365249826869,
    "timestep_mean": 512.21875,
    "epsilon_norm": 8.182238578796387,
    "prediction_norm": 7.8315582275390625,
    "timestamp": "2025-07-14T12:34:02.917498"
  },
  {
    "epoch": 830,
    "batch_idx": 128,
    "loss": 0.27639007568359375,
    "learning_rate": 0.0006319365249826869,
    "timestep_mean": 463.75,
    "epsilon_norm": 7.288690567016602,
    "prediction_norm": 6.790013790130615,
    "timestamp": "2025-07-14T12:34:02.922645"
  },
  {
    "epoch": 830,
    "batch_idx": 160,
    "loss": 0.4694899618625641,
    "learning_rate": 0.0006319365249826869,
    "timestep_mean": 455.625,
    "epsilon_norm": 8.184807777404785,
    "prediction_norm": 6.811655521392822,
    "timestamp": "2025-07-14T12:34:02.927829"
  },
  {
    "epoch": 830,
    "batch_idx": 192,
    "loss": 0.19473539292812347,
    "learning_rate": 0.0006319365249826869,
    "timestep_mean": 426.0,
    "epsilon_norm": 3.6356117725372314,
    "prediction_norm": 3.5314033031463623,
    "timestamp": "2025-07-14T12:34:02.932885"
  },
  {
    "epoch": 831,
    "batch_idx": 0,
    "loss": 0.24676768481731415,
    "learning_rate": 0.0006311788008991436,
    "timestep_mean": 522.6875,
    "epsilon_norm": 8.098118782043457,
    "prediction_norm": 6.566852569580078,
    "timestamp": "2025-07-14T12:34:02.938104"
  },
  {
    "epoch": 831,
    "batch_idx": 32,
    "loss": 0.2693306803703308,
    "learning_rate": 0.0006311788008991436,
    "timestep_mean": 480.84375,
    "epsilon_norm": 6.508223533630371,
    "prediction_norm": 5.525000095367432,
    "timestamp": "2025-07-14T12:34:02.943286"
  },
  {
    "epoch": 831,
    "batch_idx": 64,
    "loss": 0.33268624544143677,
    "learning_rate": 0.0006311788008991436,
    "timestep_mean": 471.34375,
    "epsilon_norm": 8.618088722229004,
    "prediction_norm": 6.782747268676758,
    "timestamp": "2025-07-14T12:34:02.948490"
  },
  {
    "epoch": 831,
    "batch_idx": 96,
    "loss": 0.23396499454975128,
    "learning_rate": 0.0006311788008991436,
    "timestep_mean": 502.09375,
    "epsilon_norm": 7.28792667388916,
    "prediction_norm": 5.913156986236572,
    "timestamp": "2025-07-14T12:34:02.953628"
  },
  {
    "epoch": 831,
    "batch_idx": 128,
    "loss": 0.3762015104293823,
    "learning_rate": 0.0006311788008991436,
    "timestep_mean": 454.6875,
    "epsilon_norm": 7.478992462158203,
    "prediction_norm": 6.403337001800537,
    "timestamp": "2025-07-14T12:34:02.958679"
  },
  {
    "epoch": 831,
    "batch_idx": 160,
    "loss": 0.3075878620147705,
    "learning_rate": 0.0006311788008991436,
    "timestep_mean": 401.0625,
    "epsilon_norm": 6.003478527069092,
    "prediction_norm": 4.719420909881592,
    "timestamp": "2025-07-14T12:34:02.963507"
  },
  {
    "epoch": 831,
    "batch_idx": 192,
    "loss": 0.38059714436531067,
    "learning_rate": 0.0006311788008991436,
    "timestep_mean": 413.125,
    "epsilon_norm": 4.352624893188477,
    "prediction_norm": 3.705693483352661,
    "timestamp": "2025-07-14T12:34:02.972516"
  },
  {
    "epoch": 832,
    "batch_idx": 0,
    "loss": 0.26105570793151855,
    "learning_rate": 0.0006304207531449488,
    "timestep_mean": 485.6875,
    "epsilon_norm": 9.095332145690918,
    "prediction_norm": 7.569969654083252,
    "timestamp": "2025-07-14T12:34:02.978955"
  },
  {
    "epoch": 832,
    "batch_idx": 32,
    "loss": 0.4718647003173828,
    "learning_rate": 0.0006304207531449488,
    "timestep_mean": 429.84375,
    "epsilon_norm": 10.012648582458496,
    "prediction_norm": 7.377501487731934,
    "timestamp": "2025-07-14T12:34:02.985448"
  },
  {
    "epoch": 832,
    "batch_idx": 64,
    "loss": 0.26371026039123535,
    "learning_rate": 0.0006304207531449488,
    "timestep_mean": 551.65625,
    "epsilon_norm": 7.430582046508789,
    "prediction_norm": 5.558877944946289,
    "timestamp": "2025-07-14T12:34:02.991928"
  },
  {
    "epoch": 832,
    "batch_idx": 96,
    "loss": 0.306122750043869,
    "learning_rate": 0.0006304207531449488,
    "timestep_mean": 477.125,
    "epsilon_norm": 8.710041046142578,
    "prediction_norm": 6.581117630004883,
    "timestamp": "2025-07-14T12:34:03.000421"
  },
  {
    "epoch": 832,
    "batch_idx": 128,
    "loss": 0.3238368630409241,
    "learning_rate": 0.0006304207531449488,
    "timestep_mean": 537.5625,
    "epsilon_norm": 8.475375175476074,
    "prediction_norm": 7.121154308319092,
    "timestamp": "2025-07-14T12:34:03.007061"
  },
  {
    "epoch": 832,
    "batch_idx": 160,
    "loss": 0.24132102727890015,
    "learning_rate": 0.0006304207531449488,
    "timestep_mean": 494.28125,
    "epsilon_norm": 8.253922462463379,
    "prediction_norm": 6.7092390060424805,
    "timestamp": "2025-07-14T12:34:03.013500"
  },
  {
    "epoch": 832,
    "batch_idx": 192,
    "loss": 0.2335985004901886,
    "learning_rate": 0.0006304207531449488,
    "timestep_mean": 536.625,
    "epsilon_norm": 3.7365546226501465,
    "prediction_norm": 3.0107102394104004,
    "timestamp": "2025-07-14T12:34:03.019607"
  },
  {
    "epoch": 833,
    "batch_idx": 0,
    "loss": 0.3341246247291565,
    "learning_rate": 0.0006296623835905106,
    "timestep_mean": 519.25,
    "epsilon_norm": 7.4748358726501465,
    "prediction_norm": 6.46685266494751,
    "timestamp": "2025-07-14T12:34:03.025201"
  },
  {
    "epoch": 833,
    "batch_idx": 32,
    "loss": 0.39997774362564087,
    "learning_rate": 0.0006296623835905106,
    "timestep_mean": 456.96875,
    "epsilon_norm": 8.307595252990723,
    "prediction_norm": 7.048362731933594,
    "timestamp": "2025-07-14T12:34:03.030558"
  },
  {
    "epoch": 833,
    "batch_idx": 64,
    "loss": 0.26469364762306213,
    "learning_rate": 0.0006296623835905106,
    "timestep_mean": 436.40625,
    "epsilon_norm": 7.607583522796631,
    "prediction_norm": 7.237154006958008,
    "timestamp": "2025-07-14T12:34:03.036033"
  },
  {
    "epoch": 833,
    "batch_idx": 96,
    "loss": 0.3039470911026001,
    "learning_rate": 0.0006296623835905106,
    "timestep_mean": 425.125,
    "epsilon_norm": 7.857478618621826,
    "prediction_norm": 6.926450252532959,
    "timestamp": "2025-07-14T12:34:03.041340"
  },
  {
    "epoch": 833,
    "batch_idx": 128,
    "loss": 0.2866399884223938,
    "learning_rate": 0.0006296623835905106,
    "timestep_mean": 427.59375,
    "epsilon_norm": 8.472729682922363,
    "prediction_norm": 8.176007270812988,
    "timestamp": "2025-07-14T12:34:03.046895"
  },
  {
    "epoch": 833,
    "batch_idx": 160,
    "loss": 0.42962646484375,
    "learning_rate": 0.0006296623835905106,
    "timestep_mean": 459.25,
    "epsilon_norm": 8.734947204589844,
    "prediction_norm": 7.096916198730469,
    "timestamp": "2025-07-14T12:34:03.052131"
  },
  {
    "epoch": 833,
    "batch_idx": 192,
    "loss": 0.4439502954483032,
    "learning_rate": 0.0006296623835905106,
    "timestep_mean": 343.125,
    "epsilon_norm": 4.129273414611816,
    "prediction_norm": 4.348424434661865,
    "timestamp": "2025-07-14T12:34:03.057435"
  },
  {
    "epoch": 834,
    "batch_idx": 0,
    "loss": 0.2682040333747864,
    "learning_rate": 0.0006289036941070304,
    "timestep_mean": 555.46875,
    "epsilon_norm": 8.45838737487793,
    "prediction_norm": 7.385776996612549,
    "timestamp": "2025-07-14T12:34:03.062772"
  },
  {
    "epoch": 834,
    "batch_idx": 32,
    "loss": 0.22597694396972656,
    "learning_rate": 0.0006289036941070304,
    "timestep_mean": 480.71875,
    "epsilon_norm": 8.416276931762695,
    "prediction_norm": 7.436191082000732,
    "timestamp": "2025-07-14T12:34:03.068176"
  },
  {
    "epoch": 834,
    "batch_idx": 64,
    "loss": 0.29661116003990173,
    "learning_rate": 0.0006289036941070304,
    "timestep_mean": 483.96875,
    "epsilon_norm": 7.633673667907715,
    "prediction_norm": 6.016092777252197,
    "timestamp": "2025-07-14T12:34:03.073719"
  },
  {
    "epoch": 834,
    "batch_idx": 96,
    "loss": 0.4023192524909973,
    "learning_rate": 0.0006289036941070304,
    "timestep_mean": 488.8125,
    "epsilon_norm": 8.133363723754883,
    "prediction_norm": 6.79403018951416,
    "timestamp": "2025-07-14T12:34:03.079189"
  },
  {
    "epoch": 834,
    "batch_idx": 128,
    "loss": 0.36192604899406433,
    "learning_rate": 0.0006289036941070304,
    "timestep_mean": 336.3125,
    "epsilon_norm": 8.213638305664062,
    "prediction_norm": 6.599580764770508,
    "timestamp": "2025-07-14T12:34:03.084359"
  },
  {
    "epoch": 834,
    "batch_idx": 160,
    "loss": 0.259538859128952,
    "learning_rate": 0.0006289036941070304,
    "timestep_mean": 454.9375,
    "epsilon_norm": 7.30271577835083,
    "prediction_norm": 6.9497575759887695,
    "timestamp": "2025-07-14T12:34:03.089519"
  },
  {
    "epoch": 834,
    "batch_idx": 192,
    "loss": 0.1875978410243988,
    "learning_rate": 0.0006289036941070304,
    "timestep_mean": 640.875,
    "epsilon_norm": 2.5706303119659424,
    "prediction_norm": 2.0896897315979004,
    "timestamp": "2025-07-14T12:34:03.094772"
  },
  {
    "epoch": 835,
    "batch_idx": 0,
    "loss": 0.23309651017189026,
    "learning_rate": 0.0006281446865664988,
    "timestep_mean": 513.125,
    "epsilon_norm": 7.422321796417236,
    "prediction_norm": 6.5399017333984375,
    "timestamp": "2025-07-14T12:34:03.100359"
  },
  {
    "epoch": 835,
    "batch_idx": 32,
    "loss": 0.34927934408187866,
    "learning_rate": 0.0006281446865664988,
    "timestep_mean": 500.0625,
    "epsilon_norm": 7.82409143447876,
    "prediction_norm": 6.284338474273682,
    "timestamp": "2025-07-14T12:34:03.105710"
  },
  {
    "epoch": 835,
    "batch_idx": 64,
    "loss": 0.19888240098953247,
    "learning_rate": 0.0006281446865664988,
    "timestep_mean": 498.25,
    "epsilon_norm": 7.372838973999023,
    "prediction_norm": 6.205239772796631,
    "timestamp": "2025-07-14T12:34:03.111048"
  },
  {
    "epoch": 835,
    "batch_idx": 96,
    "loss": 0.18395668268203735,
    "learning_rate": 0.0006281446865664988,
    "timestep_mean": 556.1875,
    "epsilon_norm": 8.054399490356445,
    "prediction_norm": 6.845965385437012,
    "timestamp": "2025-07-14T12:34:03.116313"
  },
  {
    "epoch": 835,
    "batch_idx": 128,
    "loss": 0.37367451190948486,
    "learning_rate": 0.0006281446865664988,
    "timestep_mean": 442.6875,
    "epsilon_norm": 7.563314914703369,
    "prediction_norm": 5.365499973297119,
    "timestamp": "2025-07-14T12:34:03.121606"
  },
  {
    "epoch": 835,
    "batch_idx": 160,
    "loss": 0.24431458115577698,
    "learning_rate": 0.0006281446865664988,
    "timestep_mean": 472.90625,
    "epsilon_norm": 6.641883373260498,
    "prediction_norm": 5.505634307861328,
    "timestamp": "2025-07-14T12:34:03.126961"
  },
  {
    "epoch": 835,
    "batch_idx": 192,
    "loss": 0.2106955349445343,
    "learning_rate": 0.0006281446865664988,
    "timestep_mean": 391.75,
    "epsilon_norm": 4.03250789642334,
    "prediction_norm": 3.3527588844299316,
    "timestamp": "2025-07-14T12:34:03.132269"
  },
  {
    "epoch": 836,
    "batch_idx": 0,
    "loss": 0.19271275401115417,
    "learning_rate": 0.0006273853628416916,
    "timestep_mean": 536.5,
    "epsilon_norm": 6.807317733764648,
    "prediction_norm": 5.962086200714111,
    "timestamp": "2025-07-14T12:34:03.137557"
  },
  {
    "epoch": 836,
    "batch_idx": 32,
    "loss": 0.3371095657348633,
    "learning_rate": 0.0006273853628416916,
    "timestep_mean": 504.90625,
    "epsilon_norm": 8.594247817993164,
    "prediction_norm": 6.566903591156006,
    "timestamp": "2025-07-14T12:34:03.143250"
  },
  {
    "epoch": 836,
    "batch_idx": 64,
    "loss": 0.23031893372535706,
    "learning_rate": 0.0006273853628416916,
    "timestep_mean": 425.5,
    "epsilon_norm": 6.530264854431152,
    "prediction_norm": 5.4393839836120605,
    "timestamp": "2025-07-14T12:34:03.148892"
  },
  {
    "epoch": 836,
    "batch_idx": 96,
    "loss": 0.35194557905197144,
    "learning_rate": 0.0006273853628416916,
    "timestep_mean": 543.75,
    "epsilon_norm": 9.234796524047852,
    "prediction_norm": 7.785411357879639,
    "timestamp": "2025-07-14T12:34:03.154086"
  },
  {
    "epoch": 836,
    "batch_idx": 128,
    "loss": 0.14435160160064697,
    "learning_rate": 0.0006273853628416916,
    "timestep_mean": 520.75,
    "epsilon_norm": 8.054667472839355,
    "prediction_norm": 7.316811561584473,
    "timestamp": "2025-07-14T12:34:03.159291"
  },
  {
    "epoch": 836,
    "batch_idx": 160,
    "loss": 0.3970864415168762,
    "learning_rate": 0.0006273853628416916,
    "timestep_mean": 442.03125,
    "epsilon_norm": 6.911874771118164,
    "prediction_norm": 4.841146469116211,
    "timestamp": "2025-07-14T12:34:03.164484"
  },
  {
    "epoch": 836,
    "batch_idx": 192,
    "loss": 0.3207095265388489,
    "learning_rate": 0.0006273853628416916,
    "timestep_mean": 346.375,
    "epsilon_norm": 4.725308895111084,
    "prediction_norm": 4.203103542327881,
    "timestamp": "2025-07-14T12:34:03.169711"
  },
  {
    "epoch": 837,
    "batch_idx": 0,
    "loss": 0.281266450881958,
    "learning_rate": 0.0006266257248061646,
    "timestep_mean": 516.46875,
    "epsilon_norm": 7.95183801651001,
    "prediction_norm": 6.713303565979004,
    "timestamp": "2025-07-14T12:34:03.174965"
  },
  {
    "epoch": 837,
    "batch_idx": 32,
    "loss": 0.3415405750274658,
    "learning_rate": 0.0006266257248061646,
    "timestep_mean": 435.375,
    "epsilon_norm": 8.593008995056152,
    "prediction_norm": 7.096613883972168,
    "timestamp": "2025-07-14T12:34:03.180224"
  },
  {
    "epoch": 837,
    "batch_idx": 64,
    "loss": 0.19944748282432556,
    "learning_rate": 0.0006266257248061646,
    "timestep_mean": 505.40625,
    "epsilon_norm": 7.211553573608398,
    "prediction_norm": 5.815295696258545,
    "timestamp": "2025-07-14T12:34:03.185438"
  },
  {
    "epoch": 837,
    "batch_idx": 96,
    "loss": 0.33835920691490173,
    "learning_rate": 0.0006266257248061646,
    "timestep_mean": 482.9375,
    "epsilon_norm": 7.73238468170166,
    "prediction_norm": 5.798999309539795,
    "timestamp": "2025-07-14T12:34:03.190665"
  },
  {
    "epoch": 837,
    "batch_idx": 128,
    "loss": 0.261851966381073,
    "learning_rate": 0.0006266257248061646,
    "timestep_mean": 525.53125,
    "epsilon_norm": 7.823847770690918,
    "prediction_norm": 6.678645610809326,
    "timestamp": "2025-07-14T12:34:03.195939"
  },
  {
    "epoch": 837,
    "batch_idx": 160,
    "loss": 0.35766682028770447,
    "learning_rate": 0.0006266257248061646,
    "timestep_mean": 555.53125,
    "epsilon_norm": 8.924553871154785,
    "prediction_norm": 7.879831790924072,
    "timestamp": "2025-07-14T12:34:03.201104"
  },
  {
    "epoch": 837,
    "batch_idx": 192,
    "loss": 0.1718367636203766,
    "learning_rate": 0.0006266257248061646,
    "timestep_mean": 613.75,
    "epsilon_norm": 3.20239520072937,
    "prediction_norm": 2.78407621383667,
    "timestamp": "2025-07-14T12:34:03.206193"
  },
  {
    "epoch": 838,
    "batch_idx": 0,
    "loss": 0.20376573503017426,
    "learning_rate": 0.0006258657743342491,
    "timestep_mean": 527.25,
    "epsilon_norm": 7.916862964630127,
    "prediction_norm": 6.8528947830200195,
    "timestamp": "2025-07-14T12:34:03.211339"
  },
  {
    "epoch": 838,
    "batch_idx": 32,
    "loss": 0.14788022637367249,
    "learning_rate": 0.0006258657743342491,
    "timestep_mean": 483.84375,
    "epsilon_norm": 6.651601791381836,
    "prediction_norm": 5.924656867980957,
    "timestamp": "2025-07-14T12:34:03.216630"
  },
  {
    "epoch": 838,
    "batch_idx": 64,
    "loss": 0.351864218711853,
    "learning_rate": 0.0006258657743342491,
    "timestep_mean": 439.78125,
    "epsilon_norm": 7.753241539001465,
    "prediction_norm": 6.042378902435303,
    "timestamp": "2025-07-14T12:34:03.221809"
  },
  {
    "epoch": 838,
    "batch_idx": 96,
    "loss": 0.3906664550304413,
    "learning_rate": 0.0006258657743342491,
    "timestep_mean": 482.25,
    "epsilon_norm": 7.836554527282715,
    "prediction_norm": 6.060096263885498,
    "timestamp": "2025-07-14T12:34:03.227013"
  },
  {
    "epoch": 838,
    "batch_idx": 128,
    "loss": 0.20986470580101013,
    "learning_rate": 0.0006258657743342491,
    "timestep_mean": 510.25,
    "epsilon_norm": 8.504548072814941,
    "prediction_norm": 7.4752912521362305,
    "timestamp": "2025-07-14T12:34:03.232386"
  },
  {
    "epoch": 838,
    "batch_idx": 160,
    "loss": 0.3488812744617462,
    "learning_rate": 0.0006258657743342491,
    "timestep_mean": 529.6875,
    "epsilon_norm": 6.659400939941406,
    "prediction_norm": 5.697475433349609,
    "timestamp": "2025-07-14T12:34:03.237572"
  },
  {
    "epoch": 838,
    "batch_idx": 192,
    "loss": 0.2655904293060303,
    "learning_rate": 0.0006258657743342491,
    "timestep_mean": 732.25,
    "epsilon_norm": 4.243963241577148,
    "prediction_norm": 4.0137739181518555,
    "timestamp": "2025-07-14T12:34:03.242828"
  },
  {
    "epoch": 839,
    "batch_idx": 0,
    "loss": 0.1655527949333191,
    "learning_rate": 0.0006251055133010474,
    "timestep_mean": 537.09375,
    "epsilon_norm": 8.142289161682129,
    "prediction_norm": 7.37855863571167,
    "timestamp": "2025-07-14T12:34:03.248047"
  },
  {
    "epoch": 839,
    "batch_idx": 32,
    "loss": 0.33880776166915894,
    "learning_rate": 0.0006251055133010474,
    "timestep_mean": 407.625,
    "epsilon_norm": 7.4280853271484375,
    "prediction_norm": 6.736242771148682,
    "timestamp": "2025-07-14T12:34:03.253132"
  },
  {
    "epoch": 839,
    "batch_idx": 64,
    "loss": 0.41134196519851685,
    "learning_rate": 0.0006251055133010474,
    "timestep_mean": 515.3125,
    "epsilon_norm": 7.887697219848633,
    "prediction_norm": 6.159121990203857,
    "timestamp": "2025-07-14T12:34:03.258310"
  },
  {
    "epoch": 839,
    "batch_idx": 96,
    "loss": 0.21000826358795166,
    "learning_rate": 0.0006251055133010474,
    "timestep_mean": 581.0625,
    "epsilon_norm": 7.054594993591309,
    "prediction_norm": 6.106844902038574,
    "timestamp": "2025-07-14T12:34:03.263539"
  },
  {
    "epoch": 839,
    "batch_idx": 128,
    "loss": 0.31926876306533813,
    "learning_rate": 0.0006251055133010474,
    "timestep_mean": 549.28125,
    "epsilon_norm": 8.783839225769043,
    "prediction_norm": 7.435664653778076,
    "timestamp": "2025-07-14T12:34:03.268824"
  },
  {
    "epoch": 839,
    "batch_idx": 160,
    "loss": 0.3033784031867981,
    "learning_rate": 0.0006251055133010474,
    "timestep_mean": 573.125,
    "epsilon_norm": 7.520227432250977,
    "prediction_norm": 6.018953800201416,
    "timestamp": "2025-07-14T12:34:03.273979"
  },
  {
    "epoch": 839,
    "batch_idx": 192,
    "loss": 0.2808758020401001,
    "learning_rate": 0.0006251055133010474,
    "timestep_mean": 539.5,
    "epsilon_norm": 4.304712772369385,
    "prediction_norm": 3.6033170223236084,
    "timestamp": "2025-07-14T12:34:03.279099"
  },
  {
    "epoch": 840,
    "batch_idx": 0,
    "loss": 0.15090793371200562,
    "learning_rate": 0.0006243449435824279,
    "timestep_mean": 493.625,
    "epsilon_norm": 8.417122840881348,
    "prediction_norm": 7.436885833740234,
    "timestamp": "2025-07-14T12:34:03.284320"
  },
  {
    "epoch": 840,
    "batch_idx": 32,
    "loss": 0.22829905152320862,
    "learning_rate": 0.0006243449435824279,
    "timestep_mean": 523.96875,
    "epsilon_norm": 7.44700813293457,
    "prediction_norm": 6.6667985916137695,
    "timestamp": "2025-07-14T12:34:03.289390"
  },
  {
    "epoch": 840,
    "batch_idx": 64,
    "loss": 0.41371193528175354,
    "learning_rate": 0.0006243449435824279,
    "timestep_mean": 412.375,
    "epsilon_norm": 8.214803695678711,
    "prediction_norm": 6.86101770401001,
    "timestamp": "2025-07-14T12:34:03.294591"
  },
  {
    "epoch": 840,
    "batch_idx": 96,
    "loss": 0.18537291884422302,
    "learning_rate": 0.0006243449435824279,
    "timestep_mean": 487.4375,
    "epsilon_norm": 8.552743911743164,
    "prediction_norm": 7.9424872398376465,
    "timestamp": "2025-07-14T12:34:03.299968"
  },
  {
    "epoch": 840,
    "batch_idx": 128,
    "loss": 0.32989197969436646,
    "learning_rate": 0.0006243449435824279,
    "timestep_mean": 443.625,
    "epsilon_norm": 8.145100593566895,
    "prediction_norm": 6.227133274078369,
    "timestamp": "2025-07-14T12:34:03.305051"
  },
  {
    "epoch": 840,
    "batch_idx": 160,
    "loss": 0.2970508635044098,
    "learning_rate": 0.0006243449435824279,
    "timestep_mean": 451.3125,
    "epsilon_norm": 8.001167297363281,
    "prediction_norm": 6.113877773284912,
    "timestamp": "2025-07-14T12:34:03.311382"
  },
  {
    "epoch": 840,
    "batch_idx": 192,
    "loss": 0.12950348854064941,
    "learning_rate": 0.0006243449435824279,
    "timestep_mean": 438.375,
    "epsilon_norm": 3.316481590270996,
    "prediction_norm": 3.0471839904785156,
    "timestamp": "2025-07-14T12:34:03.316289"
  },
  {
    "epoch": 841,
    "batch_idx": 0,
    "loss": 0.19886884093284607,
    "learning_rate": 0.000623584067055021,
    "timestep_mean": 582.71875,
    "epsilon_norm": 8.077973365783691,
    "prediction_norm": 7.1415252685546875,
    "timestamp": "2025-07-14T12:34:03.321349"
  },
  {
    "epoch": 841,
    "batch_idx": 32,
    "loss": 0.27252620458602905,
    "learning_rate": 0.000623584067055021,
    "timestep_mean": 580.75,
    "epsilon_norm": 8.860904693603516,
    "prediction_norm": 7.644340515136719,
    "timestamp": "2025-07-14T12:34:03.326193"
  },
  {
    "epoch": 841,
    "batch_idx": 64,
    "loss": 0.36602067947387695,
    "learning_rate": 0.000623584067055021,
    "timestep_mean": 386.5,
    "epsilon_norm": 7.335783004760742,
    "prediction_norm": 6.208039283752441,
    "timestamp": "2025-07-14T12:34:03.331283"
  },
  {
    "epoch": 841,
    "batch_idx": 96,
    "loss": 0.23103013634681702,
    "learning_rate": 0.000623584067055021,
    "timestep_mean": 496.0625,
    "epsilon_norm": 7.998506546020508,
    "prediction_norm": 7.009728908538818,
    "timestamp": "2025-07-14T12:34:03.336321"
  },
  {
    "epoch": 841,
    "batch_idx": 128,
    "loss": 0.3412292003631592,
    "learning_rate": 0.000623584067055021,
    "timestep_mean": 495.65625,
    "epsilon_norm": 9.266558647155762,
    "prediction_norm": 8.183870315551758,
    "timestamp": "2025-07-14T12:34:03.341356"
  },
  {
    "epoch": 841,
    "batch_idx": 160,
    "loss": 0.25028860569000244,
    "learning_rate": 0.000623584067055021,
    "timestep_mean": 607.40625,
    "epsilon_norm": 8.782610893249512,
    "prediction_norm": 7.413491725921631,
    "timestamp": "2025-07-14T12:34:03.346317"
  },
  {
    "epoch": 841,
    "batch_idx": 192,
    "loss": 0.5067667365074158,
    "learning_rate": 0.000623584067055021,
    "timestep_mean": 489.625,
    "epsilon_norm": 4.3667497634887695,
    "prediction_norm": 3.840442657470703,
    "timestamp": "2025-07-14T12:34:03.351107"
  },
  {
    "epoch": 842,
    "batch_idx": 0,
    "loss": 0.361842542886734,
    "learning_rate": 0.0006228228855962137,
    "timestep_mean": 478.0,
    "epsilon_norm": 7.958192825317383,
    "prediction_norm": 5.872201442718506,
    "timestamp": "2025-07-14T12:34:03.356105"
  },
  {
    "epoch": 842,
    "batch_idx": 32,
    "loss": 0.39964109659194946,
    "learning_rate": 0.0006228228855962137,
    "timestep_mean": 370.125,
    "epsilon_norm": 8.158780097961426,
    "prediction_norm": 5.941639423370361,
    "timestamp": "2025-07-14T12:34:03.361047"
  },
  {
    "epoch": 842,
    "batch_idx": 64,
    "loss": 0.31840842962265015,
    "learning_rate": 0.0006228228855962137,
    "timestep_mean": 489.75,
    "epsilon_norm": 8.42294979095459,
    "prediction_norm": 6.860009670257568,
    "timestamp": "2025-07-14T12:34:03.366066"
  },
  {
    "epoch": 842,
    "batch_idx": 96,
    "loss": 0.18822026252746582,
    "learning_rate": 0.0006228228855962137,
    "timestep_mean": 590.71875,
    "epsilon_norm": 8.940618515014648,
    "prediction_norm": 8.509629249572754,
    "timestamp": "2025-07-14T12:34:03.371100"
  },
  {
    "epoch": 842,
    "batch_idx": 128,
    "loss": 0.20173531770706177,
    "learning_rate": 0.0006228228855962137,
    "timestep_mean": 491.0625,
    "epsilon_norm": 7.095849990844727,
    "prediction_norm": 6.310610294342041,
    "timestamp": "2025-07-14T12:34:03.375985"
  },
  {
    "epoch": 842,
    "batch_idx": 160,
    "loss": 0.24756719172000885,
    "learning_rate": 0.0006228228855962137,
    "timestep_mean": 439.21875,
    "epsilon_norm": 6.280992031097412,
    "prediction_norm": 6.1179375648498535,
    "timestamp": "2025-07-14T12:34:03.381029"
  },
  {
    "epoch": 842,
    "batch_idx": 192,
    "loss": 0.2568238377571106,
    "learning_rate": 0.0006228228855962137,
    "timestep_mean": 453.625,
    "epsilon_norm": 3.3011841773986816,
    "prediction_norm": 2.3695943355560303,
    "timestamp": "2025-07-14T12:34:03.385985"
  },
  {
    "epoch": 843,
    "batch_idx": 0,
    "loss": 0.29332566261291504,
    "learning_rate": 0.0006220614010841457,
    "timestep_mean": 550.4375,
    "epsilon_norm": 8.381284713745117,
    "prediction_norm": 6.885555267333984,
    "timestamp": "2025-07-14T12:34:03.390967"
  },
  {
    "epoch": 843,
    "batch_idx": 32,
    "loss": 0.1960933804512024,
    "learning_rate": 0.0006220614010841457,
    "timestep_mean": 616.09375,
    "epsilon_norm": 9.376511573791504,
    "prediction_norm": 8.536791801452637,
    "timestamp": "2025-07-14T12:34:03.395989"
  },
  {
    "epoch": 843,
    "batch_idx": 64,
    "loss": 0.329214483499527,
    "learning_rate": 0.0006220614010841457,
    "timestep_mean": 458.0,
    "epsilon_norm": 7.3871588706970215,
    "prediction_norm": 5.809594631195068,
    "timestamp": "2025-07-14T12:34:03.400864"
  },
  {
    "epoch": 843,
    "batch_idx": 96,
    "loss": 0.4609043300151825,
    "learning_rate": 0.0006220614010841457,
    "timestep_mean": 478.75,
    "epsilon_norm": 9.938244819641113,
    "prediction_norm": 7.3487019538879395,
    "timestamp": "2025-07-14T12:34:03.405732"
  },
  {
    "epoch": 843,
    "batch_idx": 128,
    "loss": 0.2517033815383911,
    "learning_rate": 0.0006220614010841457,
    "timestep_mean": 619.75,
    "epsilon_norm": 7.097125053405762,
    "prediction_norm": 5.925156593322754,
    "timestamp": "2025-07-14T12:34:03.410498"
  },
  {
    "epoch": 843,
    "batch_idx": 160,
    "loss": 0.4478873908519745,
    "learning_rate": 0.0006220614010841457,
    "timestep_mean": 528.6875,
    "epsilon_norm": 7.4684014320373535,
    "prediction_norm": 6.395077228546143,
    "timestamp": "2025-07-14T12:34:03.415421"
  },
  {
    "epoch": 843,
    "batch_idx": 192,
    "loss": 0.5659047365188599,
    "learning_rate": 0.0006220614010841457,
    "timestep_mean": 324.375,
    "epsilon_norm": 3.465954303741455,
    "prediction_norm": 1.8552124500274658,
    "timestamp": "2025-07-14T12:34:03.421519"
  },
  {
    "epoch": 844,
    "batch_idx": 0,
    "loss": 0.2805643081665039,
    "learning_rate": 0.0006212996153977042,
    "timestep_mean": 514.78125,
    "epsilon_norm": 6.710041522979736,
    "prediction_norm": 5.539139270782471,
    "timestamp": "2025-07-14T12:34:03.427186"
  },
  {
    "epoch": 844,
    "batch_idx": 32,
    "loss": 0.505354642868042,
    "learning_rate": 0.0006212996153977042,
    "timestep_mean": 468.0,
    "epsilon_norm": 9.440716743469238,
    "prediction_norm": 6.613284587860107,
    "timestamp": "2025-07-14T12:34:03.433082"
  },
  {
    "epoch": 844,
    "batch_idx": 64,
    "loss": 0.21824896335601807,
    "learning_rate": 0.0006212996153977042,
    "timestep_mean": 491.375,
    "epsilon_norm": 7.126712322235107,
    "prediction_norm": 5.575402736663818,
    "timestamp": "2025-07-14T12:34:03.440427"
  },
  {
    "epoch": 844,
    "batch_idx": 96,
    "loss": 0.18902462720870972,
    "learning_rate": 0.0006212996153977042,
    "timestep_mean": 472.875,
    "epsilon_norm": 7.896561622619629,
    "prediction_norm": 6.861081123352051,
    "timestamp": "2025-07-14T12:34:03.446597"
  },
  {
    "epoch": 844,
    "batch_idx": 128,
    "loss": 0.3463153839111328,
    "learning_rate": 0.0006212996153977042,
    "timestep_mean": 461.53125,
    "epsilon_norm": 6.911752223968506,
    "prediction_norm": 5.462928771972656,
    "timestamp": "2025-07-14T12:34:03.452012"
  },
  {
    "epoch": 844,
    "batch_idx": 160,
    "loss": 0.20869871973991394,
    "learning_rate": 0.0006212996153977042,
    "timestep_mean": 561.15625,
    "epsilon_norm": 7.606029033660889,
    "prediction_norm": 6.946201801300049,
    "timestamp": "2025-07-14T12:34:03.457394"
  },
  {
    "epoch": 844,
    "batch_idx": 192,
    "loss": 0.08999896049499512,
    "learning_rate": 0.0006212996153977042,
    "timestep_mean": 660.5,
    "epsilon_norm": 3.4781036376953125,
    "prediction_norm": 2.893005609512329,
    "timestamp": "2025-07-14T12:34:03.462628"
  },
  {
    "epoch": 845,
    "batch_idx": 0,
    "loss": 0.23682737350463867,
    "learning_rate": 0.0006205375304165198,
    "timestep_mean": 530.53125,
    "epsilon_norm": 7.495232105255127,
    "prediction_norm": 7.062004566192627,
    "timestamp": "2025-07-14T12:34:03.467713"
  },
  {
    "epoch": 845,
    "batch_idx": 32,
    "loss": 0.34969374537467957,
    "learning_rate": 0.0006205375304165198,
    "timestep_mean": 434.3125,
    "epsilon_norm": 7.646072864532471,
    "prediction_norm": 6.122629642486572,
    "timestamp": "2025-07-14T12:34:03.472721"
  },
  {
    "epoch": 845,
    "batch_idx": 64,
    "loss": 0.1583624631166458,
    "learning_rate": 0.0006205375304165198,
    "timestep_mean": 448.625,
    "epsilon_norm": 7.9851484298706055,
    "prediction_norm": 7.95512056350708,
    "timestamp": "2025-07-14T12:34:03.477865"
  },
  {
    "epoch": 845,
    "batch_idx": 96,
    "loss": 0.24577705562114716,
    "learning_rate": 0.0006205375304165198,
    "timestep_mean": 462.28125,
    "epsilon_norm": 7.351690292358398,
    "prediction_norm": 6.726013660430908,
    "timestamp": "2025-07-14T12:34:03.483030"
  },
  {
    "epoch": 845,
    "batch_idx": 128,
    "loss": 0.3113119602203369,
    "learning_rate": 0.0006205375304165198,
    "timestep_mean": 464.5,
    "epsilon_norm": 7.1662797927856445,
    "prediction_norm": 4.798799514770508,
    "timestamp": "2025-07-14T12:34:03.488081"
  },
  {
    "epoch": 845,
    "batch_idx": 160,
    "loss": 0.3408541679382324,
    "learning_rate": 0.0006205375304165198,
    "timestep_mean": 378.1875,
    "epsilon_norm": 7.959474086761475,
    "prediction_norm": 6.442774295806885,
    "timestamp": "2025-07-14T12:34:03.493066"
  },
  {
    "epoch": 845,
    "batch_idx": 192,
    "loss": 0.7065831422805786,
    "learning_rate": 0.0006205375304165198,
    "timestep_mean": 537.25,
    "epsilon_norm": 4.087852954864502,
    "prediction_norm": 2.585411310195923,
    "timestamp": "2025-07-14T12:34:03.497977"
  },
  {
    "epoch": 846,
    "batch_idx": 0,
    "loss": 0.39931637048721313,
    "learning_rate": 0.0006197751480209615,
    "timestep_mean": 501.125,
    "epsilon_norm": 8.836277961730957,
    "prediction_norm": 6.961362361907959,
    "timestamp": "2025-07-14T12:34:03.502918"
  },
  {
    "epoch": 846,
    "batch_idx": 32,
    "loss": 0.11025127023458481,
    "learning_rate": 0.0006197751480209615,
    "timestep_mean": 522.28125,
    "epsilon_norm": 8.06560230255127,
    "prediction_norm": 7.48166561126709,
    "timestamp": "2025-07-14T12:34:03.507959"
  },
  {
    "epoch": 846,
    "batch_idx": 64,
    "loss": 0.20758017897605896,
    "learning_rate": 0.0006197751480209615,
    "timestep_mean": 503.21875,
    "epsilon_norm": 7.588281631469727,
    "prediction_norm": 6.756884574890137,
    "timestamp": "2025-07-14T12:34:03.513002"
  },
  {
    "epoch": 846,
    "batch_idx": 96,
    "loss": 0.4827333092689514,
    "learning_rate": 0.0006197751480209615,
    "timestep_mean": 497.03125,
    "epsilon_norm": 7.873631477355957,
    "prediction_norm": 6.3271636962890625,
    "timestamp": "2025-07-14T12:34:03.517871"
  },
  {
    "epoch": 846,
    "batch_idx": 128,
    "loss": 0.17381355166435242,
    "learning_rate": 0.0006197751480209615,
    "timestep_mean": 510.78125,
    "epsilon_norm": 7.914297580718994,
    "prediction_norm": 7.065426826477051,
    "timestamp": "2025-07-14T12:34:03.522810"
  },
  {
    "epoch": 846,
    "batch_idx": 160,
    "loss": 0.28929078578948975,
    "learning_rate": 0.0006197751480209615,
    "timestep_mean": 532.875,
    "epsilon_norm": 9.185202598571777,
    "prediction_norm": 8.587584495544434,
    "timestamp": "2025-07-14T12:34:03.527694"
  },
  {
    "epoch": 846,
    "batch_idx": 192,
    "loss": 0.8072289228439331,
    "learning_rate": 0.0006197751480209615,
    "timestep_mean": 397.5,
    "epsilon_norm": 4.958916664123535,
    "prediction_norm": 3.5818021297454834,
    "timestamp": "2025-07-14T12:34:03.532591"
  },
  {
    "epoch": 847,
    "batch_idx": 0,
    "loss": 0.5289876461029053,
    "learning_rate": 0.0006190124700921317,
    "timestep_mean": 512.8125,
    "epsilon_norm": 8.7341947555542,
    "prediction_norm": 6.609306812286377,
    "timestamp": "2025-07-14T12:34:03.537563"
  },
  {
    "epoch": 847,
    "batch_idx": 32,
    "loss": 0.3213885426521301,
    "learning_rate": 0.0006190124700921317,
    "timestep_mean": 490.125,
    "epsilon_norm": 9.233960151672363,
    "prediction_norm": 7.761419773101807,
    "timestamp": "2025-07-14T12:34:03.542444"
  },
  {
    "epoch": 847,
    "batch_idx": 64,
    "loss": 0.3762770891189575,
    "learning_rate": 0.0006190124700921317,
    "timestep_mean": 455.875,
    "epsilon_norm": 8.322815895080566,
    "prediction_norm": 6.944413661956787,
    "timestamp": "2025-07-14T12:34:03.547303"
  },
  {
    "epoch": 847,
    "batch_idx": 96,
    "loss": 0.3123185634613037,
    "learning_rate": 0.0006190124700921317,
    "timestep_mean": 551.0,
    "epsilon_norm": 9.084932327270508,
    "prediction_norm": 8.575881958007812,
    "timestamp": "2025-07-14T12:34:03.552192"
  },
  {
    "epoch": 847,
    "batch_idx": 128,
    "loss": 0.18308186531066895,
    "learning_rate": 0.0006190124700921317,
    "timestep_mean": 522.0,
    "epsilon_norm": 7.590123176574707,
    "prediction_norm": 7.232614994049072,
    "timestamp": "2025-07-14T12:34:03.557083"
  },
  {
    "epoch": 847,
    "batch_idx": 160,
    "loss": 0.16931135952472687,
    "learning_rate": 0.0006190124700921317,
    "timestep_mean": 478.71875,
    "epsilon_norm": 6.946680545806885,
    "prediction_norm": 6.190296649932861,
    "timestamp": "2025-07-14T12:34:03.562011"
  },
  {
    "epoch": 847,
    "batch_idx": 192,
    "loss": 0.21573901176452637,
    "learning_rate": 0.0006190124700921317,
    "timestep_mean": 596.375,
    "epsilon_norm": 4.057839393615723,
    "prediction_norm": 3.156114339828491,
    "timestamp": "2025-07-14T12:34:03.567057"
  },
  {
    "epoch": 848,
    "batch_idx": 0,
    "loss": 0.2684691548347473,
    "learning_rate": 0.0006182494985118628,
    "timestep_mean": 525.09375,
    "epsilon_norm": 7.952744960784912,
    "prediction_norm": 6.340198040008545,
    "timestamp": "2025-07-14T12:34:03.571998"
  },
  {
    "epoch": 848,
    "batch_idx": 32,
    "loss": 0.2427254021167755,
    "learning_rate": 0.0006182494985118628,
    "timestep_mean": 428.6875,
    "epsilon_norm": 7.341256618499756,
    "prediction_norm": 6.38803768157959,
    "timestamp": "2025-07-14T12:34:03.576899"
  },
  {
    "epoch": 848,
    "batch_idx": 64,
    "loss": 0.2443910837173462,
    "learning_rate": 0.0006182494985118628,
    "timestep_mean": 572.25,
    "epsilon_norm": 7.99338436126709,
    "prediction_norm": 7.207540512084961,
    "timestamp": "2025-07-14T12:34:03.581744"
  },
  {
    "epoch": 848,
    "batch_idx": 96,
    "loss": 0.42283472418785095,
    "learning_rate": 0.0006182494985118628,
    "timestep_mean": 505.15625,
    "epsilon_norm": 9.089932441711426,
    "prediction_norm": 7.641907215118408,
    "timestamp": "2025-07-14T12:34:03.586671"
  },
  {
    "epoch": 848,
    "batch_idx": 128,
    "loss": 0.2686343789100647,
    "learning_rate": 0.0006182494985118628,
    "timestep_mean": 365.09375,
    "epsilon_norm": 7.009836673736572,
    "prediction_norm": 5.928531646728516,
    "timestamp": "2025-07-14T12:34:03.591622"
  },
  {
    "epoch": 848,
    "batch_idx": 160,
    "loss": 0.4171617925167084,
    "learning_rate": 0.0006182494985118628,
    "timestep_mean": 468.46875,
    "epsilon_norm": 8.329120635986328,
    "prediction_norm": 7.550257205963135,
    "timestamp": "2025-07-14T12:34:03.596523"
  },
  {
    "epoch": 848,
    "batch_idx": 192,
    "loss": 0.6493402719497681,
    "learning_rate": 0.0006182494985118628,
    "timestep_mean": 352.25,
    "epsilon_norm": 3.9393935203552246,
    "prediction_norm": 1.9661847352981567,
    "timestamp": "2025-07-14T12:34:03.601356"
  },
  {
    "epoch": 849,
    "batch_idx": 0,
    "loss": 0.28246521949768066,
    "learning_rate": 0.0006174862351627111,
    "timestep_mean": 424.9375,
    "epsilon_norm": 6.836928844451904,
    "prediction_norm": 6.002205848693848,
    "timestamp": "2025-07-14T12:34:03.606251"
  },
  {
    "epoch": 849,
    "batch_idx": 32,
    "loss": 0.30138659477233887,
    "learning_rate": 0.0006174862351627111,
    "timestep_mean": 555.0,
    "epsilon_norm": 7.194441318511963,
    "prediction_norm": 5.963679790496826,
    "timestamp": "2025-07-14T12:34:03.611112"
  },
  {
    "epoch": 849,
    "batch_idx": 64,
    "loss": 0.15042826533317566,
    "learning_rate": 0.0006174862351627111,
    "timestep_mean": 556.25,
    "epsilon_norm": 7.081010341644287,
    "prediction_norm": 7.062247276306152,
    "timestamp": "2025-07-14T12:34:03.616019"
  },
  {
    "epoch": 849,
    "batch_idx": 96,
    "loss": 0.2470446079969406,
    "learning_rate": 0.0006174862351627111,
    "timestep_mean": 494.1875,
    "epsilon_norm": 7.197326183319092,
    "prediction_norm": 6.28074312210083,
    "timestamp": "2025-07-14T12:34:03.620995"
  },
  {
    "epoch": 849,
    "batch_idx": 128,
    "loss": 0.5237860083580017,
    "learning_rate": 0.0006174862351627111,
    "timestep_mean": 460.78125,
    "epsilon_norm": 8.336776733398438,
    "prediction_norm": 6.400306224822998,
    "timestamp": "2025-07-14T12:34:03.625855"
  },
  {
    "epoch": 849,
    "batch_idx": 160,
    "loss": 0.2770463228225708,
    "learning_rate": 0.0006174862351627111,
    "timestep_mean": 505.09375,
    "epsilon_norm": 8.906384468078613,
    "prediction_norm": 7.664565563201904,
    "timestamp": "2025-07-14T12:34:03.630718"
  },
  {
    "epoch": 849,
    "batch_idx": 192,
    "loss": 0.3274165987968445,
    "learning_rate": 0.0006174862351627111,
    "timestep_mean": 531.125,
    "epsilon_norm": 5.033968925476074,
    "prediction_norm": 3.9312586784362793,
    "timestamp": "2025-07-14T12:34:03.635621"
  },
  {
    "epoch": 850,
    "batch_idx": 0,
    "loss": 0.32938840985298157,
    "learning_rate": 0.0006167226819279531,
    "timestep_mean": 458.75,
    "epsilon_norm": 8.456338882446289,
    "prediction_norm": 6.195706844329834,
    "timestamp": "2025-07-14T12:34:03.640515"
  },
  {
    "epoch": 850,
    "batch_idx": 32,
    "loss": 0.5847762227058411,
    "learning_rate": 0.0006167226819279531,
    "timestep_mean": 445.21875,
    "epsilon_norm": 9.27276611328125,
    "prediction_norm": 6.714612007141113,
    "timestamp": "2025-07-14T12:34:03.645372"
  },
  {
    "epoch": 850,
    "batch_idx": 64,
    "loss": 0.16576749086380005,
    "learning_rate": 0.0006167226819279531,
    "timestep_mean": 488.375,
    "epsilon_norm": 7.860654830932617,
    "prediction_norm": 7.403237342834473,
    "timestamp": "2025-07-14T12:34:03.650371"
  },
  {
    "epoch": 850,
    "batch_idx": 96,
    "loss": 0.2383994460105896,
    "learning_rate": 0.0006167226819279531,
    "timestep_mean": 553.0,
    "epsilon_norm": 7.157716751098633,
    "prediction_norm": 5.886836528778076,
    "timestamp": "2025-07-14T12:34:03.655223"
  },
  {
    "epoch": 850,
    "batch_idx": 128,
    "loss": 0.46288856863975525,
    "learning_rate": 0.0006167226819279531,
    "timestep_mean": 438.9375,
    "epsilon_norm": 8.832847595214844,
    "prediction_norm": 7.066629886627197,
    "timestamp": "2025-07-14T12:34:03.660127"
  },
  {
    "epoch": 850,
    "batch_idx": 160,
    "loss": 0.21200953423976898,
    "learning_rate": 0.0006167226819279531,
    "timestep_mean": 467.875,
    "epsilon_norm": 5.949865818023682,
    "prediction_norm": 5.013354778289795,
    "timestamp": "2025-07-14T12:34:03.665011"
  },
  {
    "epoch": 850,
    "batch_idx": 192,
    "loss": 0.16399943828582764,
    "learning_rate": 0.0006167226819279531,
    "timestep_mean": 631.625,
    "epsilon_norm": 3.6816749572753906,
    "prediction_norm": 2.6748976707458496,
    "timestamp": "2025-07-14T12:34:03.669855"
  },
  {
    "epoch": 851,
    "batch_idx": 0,
    "loss": 0.17511993646621704,
    "learning_rate": 0.0006159588406915805,
    "timestep_mean": 583.90625,
    "epsilon_norm": 6.6238017082214355,
    "prediction_norm": 6.142673969268799,
    "timestamp": "2025-07-14T12:34:03.674900"
  },
  {
    "epoch": 851,
    "batch_idx": 32,
    "loss": 0.29170796275138855,
    "learning_rate": 0.0006159588406915805,
    "timestep_mean": 444.09375,
    "epsilon_norm": 8.446574211120605,
    "prediction_norm": 6.49600076675415,
    "timestamp": "2025-07-14T12:34:03.679764"
  },
  {
    "epoch": 851,
    "batch_idx": 64,
    "loss": 0.24008288979530334,
    "learning_rate": 0.0006159588406915805,
    "timestep_mean": 562.5,
    "epsilon_norm": 6.672126293182373,
    "prediction_norm": 5.935582637786865,
    "timestamp": "2025-07-14T12:34:03.684826"
  },
  {
    "epoch": 851,
    "batch_idx": 96,
    "loss": 0.23941922187805176,
    "learning_rate": 0.0006159588406915805,
    "timestep_mean": 467.4375,
    "epsilon_norm": 8.45594310760498,
    "prediction_norm": 7.507897853851318,
    "timestamp": "2025-07-14T12:34:03.689963"
  },
  {
    "epoch": 851,
    "batch_idx": 128,
    "loss": 0.4344479441642761,
    "learning_rate": 0.0006159588406915805,
    "timestep_mean": 419.59375,
    "epsilon_norm": 8.240022659301758,
    "prediction_norm": 6.162822723388672,
    "timestamp": "2025-07-14T12:34:03.695206"
  },
  {
    "epoch": 851,
    "batch_idx": 160,
    "loss": 0.13691461086273193,
    "learning_rate": 0.0006159588406915805,
    "timestep_mean": 562.96875,
    "epsilon_norm": 8.617216110229492,
    "prediction_norm": 8.020004272460938,
    "timestamp": "2025-07-14T12:34:03.700608"
  },
  {
    "epoch": 851,
    "batch_idx": 192,
    "loss": 0.1844550222158432,
    "learning_rate": 0.0006159588406915805,
    "timestep_mean": 549.375,
    "epsilon_norm": 2.700493335723877,
    "prediction_norm": 2.485198736190796,
    "timestamp": "2025-07-14T12:34:03.705769"
  },
  {
    "epoch": 852,
    "batch_idx": 0,
    "loss": 0.30376023054122925,
    "learning_rate": 0.0006151947133382957,
    "timestep_mean": 367.53125,
    "epsilon_norm": 7.284754276275635,
    "prediction_norm": 5.649720191955566,
    "timestamp": "2025-07-14T12:34:03.710998"
  },
  {
    "epoch": 852,
    "batch_idx": 32,
    "loss": 0.20318928360939026,
    "learning_rate": 0.0006151947133382957,
    "timestep_mean": 515.84375,
    "epsilon_norm": 7.132256507873535,
    "prediction_norm": 5.90558385848999,
    "timestamp": "2025-07-14T12:34:03.716110"
  },
  {
    "epoch": 852,
    "batch_idx": 64,
    "loss": 0.5698972940444946,
    "learning_rate": 0.0006151947133382957,
    "timestep_mean": 453.5625,
    "epsilon_norm": 8.345024108886719,
    "prediction_norm": 5.718793869018555,
    "timestamp": "2025-07-14T12:34:03.721381"
  },
  {
    "epoch": 852,
    "batch_idx": 96,
    "loss": 0.23258918523788452,
    "learning_rate": 0.0006151947133382957,
    "timestep_mean": 522.15625,
    "epsilon_norm": 7.062675952911377,
    "prediction_norm": 6.2454023361206055,
    "timestamp": "2025-07-14T12:34:03.726643"
  },
  {
    "epoch": 852,
    "batch_idx": 128,
    "loss": 0.32980796694755554,
    "learning_rate": 0.0006151947133382957,
    "timestep_mean": 385.09375,
    "epsilon_norm": 7.272281646728516,
    "prediction_norm": 5.617065906524658,
    "timestamp": "2025-07-14T12:34:03.731846"
  },
  {
    "epoch": 852,
    "batch_idx": 160,
    "loss": 0.31617724895477295,
    "learning_rate": 0.0006151947133382957,
    "timestep_mean": 506.78125,
    "epsilon_norm": 8.132899284362793,
    "prediction_norm": 6.735416412353516,
    "timestamp": "2025-07-14T12:34:03.737058"
  },
  {
    "epoch": 852,
    "batch_idx": 192,
    "loss": 0.30281853675842285,
    "learning_rate": 0.0006151947133382957,
    "timestep_mean": 339.125,
    "epsilon_norm": 4.068906784057617,
    "prediction_norm": 3.2729599475860596,
    "timestamp": "2025-07-14T12:34:03.742313"
  },
  {
    "epoch": 853,
    "batch_idx": 0,
    "loss": 0.22432109713554382,
    "learning_rate": 0.0006144303017535068,
    "timestep_mean": 493.65625,
    "epsilon_norm": 7.480733871459961,
    "prediction_norm": 6.38728141784668,
    "timestamp": "2025-07-14T12:34:03.747541"
  },
  {
    "epoch": 853,
    "batch_idx": 32,
    "loss": 0.49256160855293274,
    "learning_rate": 0.0006144303017535068,
    "timestep_mean": 452.0,
    "epsilon_norm": 7.671748161315918,
    "prediction_norm": 6.158130168914795,
    "timestamp": "2025-07-14T12:34:03.752657"
  },
  {
    "epoch": 853,
    "batch_idx": 64,
    "loss": 0.36732417345046997,
    "learning_rate": 0.0006144303017535068,
    "timestep_mean": 462.1875,
    "epsilon_norm": 7.099970817565918,
    "prediction_norm": 5.932095527648926,
    "timestamp": "2025-07-14T12:34:03.757809"
  },
  {
    "epoch": 853,
    "batch_idx": 96,
    "loss": 0.133113294839859,
    "learning_rate": 0.0006144303017535068,
    "timestep_mean": 542.125,
    "epsilon_norm": 8.139015197753906,
    "prediction_norm": 7.40056037902832,
    "timestamp": "2025-07-14T12:34:03.763091"
  },
  {
    "epoch": 853,
    "batch_idx": 128,
    "loss": 0.37315863370895386,
    "learning_rate": 0.0006144303017535068,
    "timestep_mean": 460.4375,
    "epsilon_norm": 7.012815952301025,
    "prediction_norm": 5.351832389831543,
    "timestamp": "2025-07-14T12:34:03.768243"
  },
  {
    "epoch": 853,
    "batch_idx": 160,
    "loss": 0.3232119679450989,
    "learning_rate": 0.0006144303017535068,
    "timestep_mean": 550.1875,
    "epsilon_norm": 8.728275299072266,
    "prediction_norm": 7.288044452667236,
    "timestamp": "2025-07-14T12:34:03.773447"
  },
  {
    "epoch": 853,
    "batch_idx": 192,
    "loss": 0.2833889424800873,
    "learning_rate": 0.0006144303017535068,
    "timestep_mean": 545.625,
    "epsilon_norm": 3.0095839500427246,
    "prediction_norm": 2.3774516582489014,
    "timestamp": "2025-07-14T12:34:03.778700"
  },
  {
    "epoch": 854,
    "batch_idx": 0,
    "loss": 0.33880627155303955,
    "learning_rate": 0.0006136656078233236,
    "timestep_mean": 556.15625,
    "epsilon_norm": 8.408651351928711,
    "prediction_norm": 7.117166519165039,
    "timestamp": "2025-07-14T12:34:03.784028"
  },
  {
    "epoch": 854,
    "batch_idx": 32,
    "loss": 0.16998344659805298,
    "learning_rate": 0.0006136656078233236,
    "timestep_mean": 526.125,
    "epsilon_norm": 7.828731060028076,
    "prediction_norm": 6.989126682281494,
    "timestamp": "2025-07-14T12:34:03.789151"
  },
  {
    "epoch": 854,
    "batch_idx": 64,
    "loss": 0.23011033236980438,
    "learning_rate": 0.0006136656078233236,
    "timestep_mean": 530.90625,
    "epsilon_norm": 7.626675605773926,
    "prediction_norm": 6.413277626037598,
    "timestamp": "2025-07-14T12:34:03.794374"
  },
  {
    "epoch": 854,
    "batch_idx": 96,
    "loss": 0.32924503087997437,
    "learning_rate": 0.0006136656078233236,
    "timestep_mean": 543.65625,
    "epsilon_norm": 7.998308181762695,
    "prediction_norm": 6.273464202880859,
    "timestamp": "2025-07-14T12:34:03.799538"
  },
  {
    "epoch": 854,
    "batch_idx": 128,
    "loss": 0.4608115255832672,
    "learning_rate": 0.0006136656078233236,
    "timestep_mean": 512.0,
    "epsilon_norm": 9.775060653686523,
    "prediction_norm": 7.110431671142578,
    "timestamp": "2025-07-14T12:34:03.804796"
  },
  {
    "epoch": 854,
    "batch_idx": 160,
    "loss": 0.30729877948760986,
    "learning_rate": 0.0006136656078233236,
    "timestep_mean": 489.9375,
    "epsilon_norm": 7.297199726104736,
    "prediction_norm": 5.509218692779541,
    "timestamp": "2025-07-14T12:34:03.810015"
  },
  {
    "epoch": 854,
    "batch_idx": 192,
    "loss": 0.4267697036266327,
    "learning_rate": 0.0006136656078233236,
    "timestep_mean": 297.875,
    "epsilon_norm": 4.167718887329102,
    "prediction_norm": 3.0709991455078125,
    "timestamp": "2025-07-14T12:34:03.815209"
  },
  {
    "epoch": 855,
    "batch_idx": 0,
    "loss": 0.44237497448921204,
    "learning_rate": 0.0006129006334345523,
    "timestep_mean": 552.21875,
    "epsilon_norm": 10.030816078186035,
    "prediction_norm": 8.382026672363281,
    "timestamp": "2025-07-14T12:34:03.820389"
  },
  {
    "epoch": 855,
    "batch_idx": 32,
    "loss": 0.1366647183895111,
    "learning_rate": 0.0006129006334345523,
    "timestep_mean": 556.375,
    "epsilon_norm": 7.862054824829102,
    "prediction_norm": 7.246452808380127,
    "timestamp": "2025-07-14T12:34:03.825637"
  },
  {
    "epoch": 855,
    "batch_idx": 64,
    "loss": 0.36028149724006653,
    "learning_rate": 0.0006129006334345523,
    "timestep_mean": 552.09375,
    "epsilon_norm": 8.745952606201172,
    "prediction_norm": 6.894902229309082,
    "timestamp": "2025-07-14T12:34:03.830847"
  },
  {
    "epoch": 855,
    "batch_idx": 96,
    "loss": 0.3720753788948059,
    "learning_rate": 0.0006129006334345523,
    "timestep_mean": 513.96875,
    "epsilon_norm": 8.544442176818848,
    "prediction_norm": 6.638117790222168,
    "timestamp": "2025-07-14T12:34:03.835801"
  },
  {
    "epoch": 855,
    "batch_idx": 128,
    "loss": 0.3098551034927368,
    "learning_rate": 0.0006129006334345523,
    "timestep_mean": 465.15625,
    "epsilon_norm": 7.942939281463623,
    "prediction_norm": 6.331608295440674,
    "timestamp": "2025-07-14T12:34:03.840737"
  },
  {
    "epoch": 855,
    "batch_idx": 160,
    "loss": 0.23421134054660797,
    "learning_rate": 0.0006129006334345523,
    "timestep_mean": 542.625,
    "epsilon_norm": 6.319738864898682,
    "prediction_norm": 6.0339555740356445,
    "timestamp": "2025-07-14T12:34:03.845917"
  },
  {
    "epoch": 855,
    "batch_idx": 192,
    "loss": 0.2486152946949005,
    "learning_rate": 0.0006129006334345523,
    "timestep_mean": 401.375,
    "epsilon_norm": 3.059985637664795,
    "prediction_norm": 2.667628049850464,
    "timestamp": "2025-07-14T12:34:03.850869"
  },
  {
    "epoch": 856,
    "batch_idx": 0,
    "loss": 0.19206354022026062,
    "learning_rate": 0.000612135380474691,
    "timestep_mean": 453.5625,
    "epsilon_norm": 7.35055685043335,
    "prediction_norm": 7.478143215179443,
    "timestamp": "2025-07-14T12:34:03.855850"
  },
  {
    "epoch": 856,
    "batch_idx": 32,
    "loss": 0.23443453013896942,
    "learning_rate": 0.000612135380474691,
    "timestep_mean": 549.5625,
    "epsilon_norm": 8.555689811706543,
    "prediction_norm": 7.6999711990356445,
    "timestamp": "2025-07-14T12:34:03.860987"
  },
  {
    "epoch": 856,
    "batch_idx": 64,
    "loss": 0.39043286442756653,
    "learning_rate": 0.000612135380474691,
    "timestep_mean": 576.40625,
    "epsilon_norm": 8.231073379516602,
    "prediction_norm": 7.071655750274658,
    "timestamp": "2025-07-14T12:34:03.866192"
  },
  {
    "epoch": 856,
    "batch_idx": 96,
    "loss": 0.14787323772907257,
    "learning_rate": 0.000612135380474691,
    "timestep_mean": 558.34375,
    "epsilon_norm": 8.156142234802246,
    "prediction_norm": 8.006422996520996,
    "timestamp": "2025-07-14T12:34:03.871392"
  },
  {
    "epoch": 856,
    "batch_idx": 128,
    "loss": 0.22232134640216827,
    "learning_rate": 0.000612135380474691,
    "timestep_mean": 564.1875,
    "epsilon_norm": 7.912175178527832,
    "prediction_norm": 7.611886501312256,
    "timestamp": "2025-07-14T12:34:03.876559"
  },
  {
    "epoch": 856,
    "batch_idx": 160,
    "loss": 0.3264092803001404,
    "learning_rate": 0.000612135380474691,
    "timestep_mean": 459.6875,
    "epsilon_norm": 8.085152626037598,
    "prediction_norm": 6.8270416259765625,
    "timestamp": "2025-07-14T12:34:03.881723"
  },
  {
    "epoch": 856,
    "batch_idx": 192,
    "loss": 0.3207317292690277,
    "learning_rate": 0.000612135380474691,
    "timestep_mean": 625.375,
    "epsilon_norm": 3.7686593532562256,
    "prediction_norm": 3.386061429977417,
    "timestamp": "2025-07-14T12:34:03.886836"
  },
  {
    "epoch": 857,
    "batch_idx": 0,
    "loss": 0.26088714599609375,
    "learning_rate": 0.0006113698508319255,
    "timestep_mean": 551.09375,
    "epsilon_norm": 7.487618923187256,
    "prediction_norm": 6.460686683654785,
    "timestamp": "2025-07-14T12:34:03.892032"
  },
  {
    "epoch": 857,
    "batch_idx": 32,
    "loss": 0.3515739440917969,
    "learning_rate": 0.0006113698508319255,
    "timestep_mean": 471.6875,
    "epsilon_norm": 8.4739990234375,
    "prediction_norm": 7.234567642211914,
    "timestamp": "2025-07-14T12:34:03.897195"
  },
  {
    "epoch": 857,
    "batch_idx": 64,
    "loss": 0.3177516758441925,
    "learning_rate": 0.0006113698508319255,
    "timestep_mean": 535.40625,
    "epsilon_norm": 8.423478126525879,
    "prediction_norm": 7.430878162384033,
    "timestamp": "2025-07-14T12:34:03.902397"
  },
  {
    "epoch": 857,
    "batch_idx": 96,
    "loss": 0.37312883138656616,
    "learning_rate": 0.0006113698508319255,
    "timestep_mean": 431.59375,
    "epsilon_norm": 7.10825777053833,
    "prediction_norm": 5.147620677947998,
    "timestamp": "2025-07-14T12:34:03.909050"
  },
  {
    "epoch": 857,
    "batch_idx": 128,
    "loss": 0.20292767882347107,
    "learning_rate": 0.0006113698508319255,
    "timestep_mean": 492.34375,
    "epsilon_norm": 6.769712924957275,
    "prediction_norm": 6.127074241638184,
    "timestamp": "2025-07-14T12:34:03.917341"
  },
  {
    "epoch": 857,
    "batch_idx": 160,
    "loss": 0.2691854238510132,
    "learning_rate": 0.0006113698508319255,
    "timestep_mean": 458.0,
    "epsilon_norm": 7.84381103515625,
    "prediction_norm": 6.360086441040039,
    "timestamp": "2025-07-14T12:34:03.922575"
  },
  {
    "epoch": 857,
    "batch_idx": 192,
    "loss": 0.30248257517814636,
    "learning_rate": 0.0006113698508319255,
    "timestep_mean": 566.875,
    "epsilon_norm": 4.306270122528076,
    "prediction_norm": 3.109144926071167,
    "timestamp": "2025-07-14T12:34:03.927621"
  },
  {
    "epoch": 858,
    "batch_idx": 0,
    "loss": 0.31062349677085876,
    "learning_rate": 0.000610604046395124,
    "timestep_mean": 544.4375,
    "epsilon_norm": 7.416293144226074,
    "prediction_norm": 5.788955211639404,
    "timestamp": "2025-07-14T12:34:03.932813"
  },
  {
    "epoch": 858,
    "batch_idx": 32,
    "loss": 0.36447593569755554,
    "learning_rate": 0.000610604046395124,
    "timestep_mean": 484.5,
    "epsilon_norm": 8.646695137023926,
    "prediction_norm": 6.7590861320495605,
    "timestamp": "2025-07-14T12:34:03.938176"
  },
  {
    "epoch": 858,
    "batch_idx": 64,
    "loss": 0.17002825438976288,
    "learning_rate": 0.000610604046395124,
    "timestep_mean": 465.0625,
    "epsilon_norm": 8.163789749145508,
    "prediction_norm": 7.448318958282471,
    "timestamp": "2025-07-14T12:34:03.943450"
  },
  {
    "epoch": 858,
    "batch_idx": 96,
    "loss": 0.32298505306243896,
    "learning_rate": 0.000610604046395124,
    "timestep_mean": 479.78125,
    "epsilon_norm": 8.478666305541992,
    "prediction_norm": 6.564938545227051,
    "timestamp": "2025-07-14T12:34:03.948626"
  },
  {
    "epoch": 858,
    "batch_idx": 128,
    "loss": 0.13391074538230896,
    "learning_rate": 0.000610604046395124,
    "timestep_mean": 593.25,
    "epsilon_norm": 7.512141227722168,
    "prediction_norm": 6.7598443031311035,
    "timestamp": "2025-07-14T12:34:03.953763"
  },
  {
    "epoch": 858,
    "batch_idx": 160,
    "loss": 0.18665319681167603,
    "learning_rate": 0.000610604046395124,
    "timestep_mean": 568.0,
    "epsilon_norm": 8.134465217590332,
    "prediction_norm": 6.760143280029297,
    "timestamp": "2025-07-14T12:34:03.959002"
  },
  {
    "epoch": 858,
    "batch_idx": 192,
    "loss": 0.19981901347637177,
    "learning_rate": 0.000610604046395124,
    "timestep_mean": 450.5,
    "epsilon_norm": 4.090298175811768,
    "prediction_norm": 3.2348506450653076,
    "timestamp": "2025-07-14T12:34:03.964017"
  },
  {
    "epoch": 859,
    "batch_idx": 0,
    "loss": 0.2931891977787018,
    "learning_rate": 0.000609837969053833,
    "timestep_mean": 511.625,
    "epsilon_norm": 8.065227508544922,
    "prediction_norm": 6.712500095367432,
    "timestamp": "2025-07-14T12:34:03.969276"
  },
  {
    "epoch": 859,
    "batch_idx": 32,
    "loss": 0.3006073832511902,
    "learning_rate": 0.000609837969053833,
    "timestep_mean": 503.40625,
    "epsilon_norm": 7.049145698547363,
    "prediction_norm": 5.229019641876221,
    "timestamp": "2025-07-14T12:34:03.974453"
  },
  {
    "epoch": 859,
    "batch_idx": 64,
    "loss": 0.4193599820137024,
    "learning_rate": 0.000609837969053833,
    "timestep_mean": 439.90625,
    "epsilon_norm": 7.06697940826416,
    "prediction_norm": 5.043257713317871,
    "timestamp": "2025-07-14T12:34:03.979772"
  },
  {
    "epoch": 859,
    "batch_idx": 96,
    "loss": 0.2103290557861328,
    "learning_rate": 0.000609837969053833,
    "timestep_mean": 488.6875,
    "epsilon_norm": 7.08164644241333,
    "prediction_norm": 6.040266990661621,
    "timestamp": "2025-07-14T12:34:03.984990"
  },
  {
    "epoch": 859,
    "batch_idx": 128,
    "loss": 0.3183017671108246,
    "learning_rate": 0.000609837969053833,
    "timestep_mean": 510.375,
    "epsilon_norm": 8.520130157470703,
    "prediction_norm": 6.811656475067139,
    "timestamp": "2025-07-14T12:34:03.990040"
  },
  {
    "epoch": 859,
    "batch_idx": 160,
    "loss": 0.3637232482433319,
    "learning_rate": 0.000609837969053833,
    "timestep_mean": 504.3125,
    "epsilon_norm": 7.559401512145996,
    "prediction_norm": 5.504824638366699,
    "timestamp": "2025-07-14T12:34:03.994876"
  },
  {
    "epoch": 859,
    "batch_idx": 192,
    "loss": 0.7109675407409668,
    "learning_rate": 0.000609837969053833,
    "timestep_mean": 393.625,
    "epsilon_norm": 4.707465648651123,
    "prediction_norm": 3.2804415225982666,
    "timestamp": "2025-07-14T12:34:03.999690"
  },
  {
    "epoch": 860,
    "batch_idx": 0,
    "loss": 0.2914610505104065,
    "learning_rate": 0.0006090716206982718,
    "timestep_mean": 537.4375,
    "epsilon_norm": 7.6971001625061035,
    "prediction_norm": 6.105113983154297,
    "timestamp": "2025-07-14T12:34:04.004641"
  },
  {
    "epoch": 860,
    "batch_idx": 32,
    "loss": 0.22707988321781158,
    "learning_rate": 0.0006090716206982718,
    "timestep_mean": 481.625,
    "epsilon_norm": 7.213721752166748,
    "prediction_norm": 6.285219192504883,
    "timestamp": "2025-07-14T12:34:04.009482"
  },
  {
    "epoch": 860,
    "batch_idx": 64,
    "loss": 0.1394367665052414,
    "learning_rate": 0.0006090716206982718,
    "timestep_mean": 529.03125,
    "epsilon_norm": 7.657589912414551,
    "prediction_norm": 6.774622917175293,
    "timestamp": "2025-07-14T12:34:04.014445"
  },
  {
    "epoch": 860,
    "batch_idx": 96,
    "loss": 0.19966533780097961,
    "learning_rate": 0.0006090716206982718,
    "timestep_mean": 620.65625,
    "epsilon_norm": 6.783871173858643,
    "prediction_norm": 5.741869926452637,
    "timestamp": "2025-07-14T12:34:04.019531"
  },
  {
    "epoch": 860,
    "batch_idx": 128,
    "loss": 0.4455307126045227,
    "learning_rate": 0.0006090716206982718,
    "timestep_mean": 515.28125,
    "epsilon_norm": 8.363274574279785,
    "prediction_norm": 6.585512161254883,
    "timestamp": "2025-07-14T12:34:04.024669"
  },
  {
    "epoch": 860,
    "batch_idx": 160,
    "loss": 0.4178180694580078,
    "learning_rate": 0.0006090716206982718,
    "timestep_mean": 403.9375,
    "epsilon_norm": 8.364005088806152,
    "prediction_norm": 7.231619358062744,
    "timestamp": "2025-07-14T12:34:04.029964"
  },
  {
    "epoch": 860,
    "batch_idx": 192,
    "loss": 0.3438633978366852,
    "learning_rate": 0.0006090716206982718,
    "timestep_mean": 441.375,
    "epsilon_norm": 5.423467636108398,
    "prediction_norm": 4.218672275543213,
    "timestamp": "2025-07-14T12:34:04.035046"
  },
  {
    "epoch": 861,
    "batch_idx": 0,
    "loss": 0.1772925853729248,
    "learning_rate": 0.0006083050032193292,
    "timestep_mean": 463.25,
    "epsilon_norm": 7.447628498077393,
    "prediction_norm": 6.442004680633545,
    "timestamp": "2025-07-14T12:34:04.040196"
  },
  {
    "epoch": 861,
    "batch_idx": 32,
    "loss": 0.37247413396835327,
    "learning_rate": 0.0006083050032193292,
    "timestep_mean": 442.5625,
    "epsilon_norm": 7.47213888168335,
    "prediction_norm": 6.5857415199279785,
    "timestamp": "2025-07-14T12:34:04.045400"
  },
  {
    "epoch": 861,
    "batch_idx": 64,
    "loss": 0.21186482906341553,
    "learning_rate": 0.0006083050032193292,
    "timestep_mean": 519.6875,
    "epsilon_norm": 7.711405277252197,
    "prediction_norm": 6.801644325256348,
    "timestamp": "2025-07-14T12:34:04.050595"
  },
  {
    "epoch": 861,
    "batch_idx": 96,
    "loss": 0.3086118698120117,
    "learning_rate": 0.0006083050032193292,
    "timestep_mean": 493.40625,
    "epsilon_norm": 7.797002792358398,
    "prediction_norm": 6.662252902984619,
    "timestamp": "2025-07-14T12:34:04.055809"
  },
  {
    "epoch": 861,
    "batch_idx": 128,
    "loss": 0.1291675567626953,
    "learning_rate": 0.0006083050032193292,
    "timestep_mean": 592.6875,
    "epsilon_norm": 7.89914083480835,
    "prediction_norm": 8.065590858459473,
    "timestamp": "2025-07-14T12:34:04.060975"
  },
  {
    "epoch": 861,
    "batch_idx": 160,
    "loss": 0.22608694434165955,
    "learning_rate": 0.0006083050032193292,
    "timestep_mean": 486.75,
    "epsilon_norm": 6.682717800140381,
    "prediction_norm": 5.816085338592529,
    "timestamp": "2025-07-14T12:34:04.066168"
  },
  {
    "epoch": 861,
    "batch_idx": 192,
    "loss": 0.07496882975101471,
    "learning_rate": 0.0006083050032193292,
    "timestep_mean": 653.125,
    "epsilon_norm": 3.2662353515625,
    "prediction_norm": 3.4996280670166016,
    "timestamp": "2025-07-14T12:34:04.071542"
  },
  {
    "epoch": 862,
    "batch_idx": 0,
    "loss": 0.36227303743362427,
    "learning_rate": 0.0006075381185085572,
    "timestep_mean": 486.90625,
    "epsilon_norm": 8.079440116882324,
    "prediction_norm": 7.5019330978393555,
    "timestamp": "2025-07-14T12:34:04.076619"
  },
  {
    "epoch": 862,
    "batch_idx": 32,
    "loss": 0.25729185342788696,
    "learning_rate": 0.0006075381185085572,
    "timestep_mean": 419.03125,
    "epsilon_norm": 7.697131156921387,
    "prediction_norm": 6.542961597442627,
    "timestamp": "2025-07-14T12:34:04.081708"
  },
  {
    "epoch": 862,
    "batch_idx": 64,
    "loss": 0.25225257873535156,
    "learning_rate": 0.0006075381185085572,
    "timestep_mean": 496.8125,
    "epsilon_norm": 7.721860408782959,
    "prediction_norm": 5.875545978546143,
    "timestamp": "2025-07-14T12:34:04.086808"
  },
  {
    "epoch": 862,
    "batch_idx": 96,
    "loss": 0.3089556396007538,
    "learning_rate": 0.0006075381185085572,
    "timestep_mean": 528.9375,
    "epsilon_norm": 7.10037899017334,
    "prediction_norm": 5.419345378875732,
    "timestamp": "2025-07-14T12:34:04.092100"
  },
  {
    "epoch": 862,
    "batch_idx": 128,
    "loss": 0.1556118279695511,
    "learning_rate": 0.0006075381185085572,
    "timestep_mean": 547.21875,
    "epsilon_norm": 7.523608207702637,
    "prediction_norm": 6.2675886154174805,
    "timestamp": "2025-07-14T12:34:04.097255"
  },
  {
    "epoch": 862,
    "batch_idx": 160,
    "loss": 0.43010130524635315,
    "learning_rate": 0.0006075381185085572,
    "timestep_mean": 429.625,
    "epsilon_norm": 7.998387336730957,
    "prediction_norm": 5.904062271118164,
    "timestamp": "2025-07-14T12:34:04.102403"
  },
  {
    "epoch": 862,
    "batch_idx": 192,
    "loss": 0.1884356141090393,
    "learning_rate": 0.0006075381185085572,
    "timestep_mean": 465.25,
    "epsilon_norm": 2.8660995960235596,
    "prediction_norm": 2.1539814472198486,
    "timestamp": "2025-07-14T12:34:04.107536"
  },
  {
    "epoch": 863,
    "batch_idx": 0,
    "loss": 0.3069305121898651,
    "learning_rate": 0.000606770968458168,
    "timestep_mean": 458.125,
    "epsilon_norm": 7.344196796417236,
    "prediction_norm": 5.609867095947266,
    "timestamp": "2025-07-14T12:34:04.112712"
  },
  {
    "epoch": 863,
    "batch_idx": 32,
    "loss": 0.22909346222877502,
    "learning_rate": 0.000606770968458168,
    "timestep_mean": 500.96875,
    "epsilon_norm": 8.1444091796875,
    "prediction_norm": 6.843929290771484,
    "timestamp": "2025-07-14T12:34:04.117853"
  },
  {
    "epoch": 863,
    "batch_idx": 64,
    "loss": 0.23334118723869324,
    "learning_rate": 0.000606770968458168,
    "timestep_mean": 525.53125,
    "epsilon_norm": 7.854078769683838,
    "prediction_norm": 6.442335605621338,
    "timestamp": "2025-07-14T12:34:04.123038"
  },
  {
    "epoch": 863,
    "batch_idx": 96,
    "loss": 0.36688148975372314,
    "learning_rate": 0.000606770968458168,
    "timestep_mean": 440.9375,
    "epsilon_norm": 7.973243236541748,
    "prediction_norm": 5.940864562988281,
    "timestamp": "2025-07-14T12:34:04.128235"
  },
  {
    "epoch": 863,
    "batch_idx": 128,
    "loss": 0.32717636227607727,
    "learning_rate": 0.000606770968458168,
    "timestep_mean": 497.09375,
    "epsilon_norm": 9.30366325378418,
    "prediction_norm": 7.3611931800842285,
    "timestamp": "2025-07-14T12:34:04.133465"
  },
  {
    "epoch": 863,
    "batch_idx": 160,
    "loss": 0.22695910930633545,
    "learning_rate": 0.000606770968458168,
    "timestep_mean": 515.15625,
    "epsilon_norm": 7.311010360717773,
    "prediction_norm": 6.5128607749938965,
    "timestamp": "2025-07-14T12:34:04.138572"
  },
  {
    "epoch": 863,
    "batch_idx": 192,
    "loss": 0.0905497819185257,
    "learning_rate": 0.000606770968458168,
    "timestep_mean": 352.375,
    "epsilon_norm": 2.877527952194214,
    "prediction_norm": 3.064121723175049,
    "timestamp": "2025-07-14T12:34:04.143516"
  },
  {
    "epoch": 864,
    "batch_idx": 0,
    "loss": 0.2846643328666687,
    "learning_rate": 0.000606003554961028,
    "timestep_mean": 502.0625,
    "epsilon_norm": 8.368102073669434,
    "prediction_norm": 7.341275691986084,
    "timestamp": "2025-07-14T12:34:04.148533"
  },
  {
    "epoch": 864,
    "batch_idx": 32,
    "loss": 0.2018185704946518,
    "learning_rate": 0.000606003554961028,
    "timestep_mean": 608.25,
    "epsilon_norm": 7.822696685791016,
    "prediction_norm": 6.980366230010986,
    "timestamp": "2025-07-14T12:34:04.153542"
  },
  {
    "epoch": 864,
    "batch_idx": 64,
    "loss": 0.35664528608322144,
    "learning_rate": 0.000606003554961028,
    "timestep_mean": 451.59375,
    "epsilon_norm": 8.817410469055176,
    "prediction_norm": 7.890940189361572,
    "timestamp": "2025-07-14T12:34:04.158789"
  },
  {
    "epoch": 864,
    "batch_idx": 96,
    "loss": 0.40217167139053345,
    "learning_rate": 0.000606003554961028,
    "timestep_mean": 442.96875,
    "epsilon_norm": 8.682291030883789,
    "prediction_norm": 6.7337965965271,
    "timestamp": "2025-07-14T12:34:04.163733"
  },
  {
    "epoch": 864,
    "batch_idx": 128,
    "loss": 0.2556096911430359,
    "learning_rate": 0.000606003554961028,
    "timestep_mean": 518.9375,
    "epsilon_norm": 6.801145076751709,
    "prediction_norm": 6.037487506866455,
    "timestamp": "2025-07-14T12:34:04.168634"
  },
  {
    "epoch": 864,
    "batch_idx": 160,
    "loss": 0.31579017639160156,
    "learning_rate": 0.000606003554961028,
    "timestep_mean": 542.0625,
    "epsilon_norm": 8.331003189086914,
    "prediction_norm": 7.172677516937256,
    "timestamp": "2025-07-14T12:34:04.173533"
  },
  {
    "epoch": 864,
    "batch_idx": 192,
    "loss": 0.39792513847351074,
    "learning_rate": 0.000606003554961028,
    "timestep_mean": 386.0,
    "epsilon_norm": 2.884235382080078,
    "prediction_norm": 2.280673027038574,
    "timestamp": "2025-07-14T12:34:04.178370"
  },
  {
    "epoch": 865,
    "batch_idx": 0,
    "loss": 0.47121649980545044,
    "learning_rate": 0.0006052358799106533,
    "timestep_mean": 526.375,
    "epsilon_norm": 9.4434175491333,
    "prediction_norm": 7.452001571655273,
    "timestamp": "2025-07-14T12:34:04.183288"
  },
  {
    "epoch": 865,
    "batch_idx": 32,
    "loss": 0.19278393685817719,
    "learning_rate": 0.0006052358799106533,
    "timestep_mean": 598.9375,
    "epsilon_norm": 8.232413291931152,
    "prediction_norm": 7.476521015167236,
    "timestamp": "2025-07-14T12:34:04.188184"
  },
  {
    "epoch": 865,
    "batch_idx": 64,
    "loss": 0.26547926664352417,
    "learning_rate": 0.0006052358799106533,
    "timestep_mean": 431.125,
    "epsilon_norm": 7.748887062072754,
    "prediction_norm": 6.343480110168457,
    "timestamp": "2025-07-14T12:34:04.193064"
  },
  {
    "epoch": 865,
    "batch_idx": 96,
    "loss": 0.2434227466583252,
    "learning_rate": 0.0006052358799106533,
    "timestep_mean": 448.0,
    "epsilon_norm": 7.535167217254639,
    "prediction_norm": 6.076927185058594,
    "timestamp": "2025-07-14T12:34:04.197913"
  },
  {
    "epoch": 865,
    "batch_idx": 128,
    "loss": 0.2246396839618683,
    "learning_rate": 0.0006052358799106533,
    "timestep_mean": 535.0625,
    "epsilon_norm": 8.049482345581055,
    "prediction_norm": 7.25341272354126,
    "timestamp": "2025-07-14T12:34:04.202799"
  },
  {
    "epoch": 865,
    "batch_idx": 160,
    "loss": 0.33199042081832886,
    "learning_rate": 0.0006052358799106533,
    "timestep_mean": 554.0625,
    "epsilon_norm": 8.520712852478027,
    "prediction_norm": 6.115126609802246,
    "timestamp": "2025-07-14T12:34:04.207721"
  },
  {
    "epoch": 865,
    "batch_idx": 192,
    "loss": 0.22580520808696747,
    "learning_rate": 0.0006052358799106533,
    "timestep_mean": 541.5,
    "epsilon_norm": 3.124506711959839,
    "prediction_norm": 2.856097459793091,
    "timestamp": "2025-07-14T12:34:04.212699"
  },
  {
    "epoch": 866,
    "batch_idx": 0,
    "loss": 0.23008085787296295,
    "learning_rate": 0.0006044679452012064,
    "timestep_mean": 468.96875,
    "epsilon_norm": 7.137782096862793,
    "prediction_norm": 5.918162822723389,
    "timestamp": "2025-07-14T12:34:04.217601"
  },
  {
    "epoch": 866,
    "batch_idx": 32,
    "loss": 0.21093836426734924,
    "learning_rate": 0.0006044679452012064,
    "timestep_mean": 512.4375,
    "epsilon_norm": 7.961483478546143,
    "prediction_norm": 6.998222827911377,
    "timestamp": "2025-07-14T12:34:04.222448"
  },
  {
    "epoch": 866,
    "batch_idx": 64,
    "loss": 0.36607033014297485,
    "learning_rate": 0.0006044679452012064,
    "timestep_mean": 374.1875,
    "epsilon_norm": 7.944620609283447,
    "prediction_norm": 6.590587615966797,
    "timestamp": "2025-07-14T12:34:04.227296"
  },
  {
    "epoch": 866,
    "batch_idx": 96,
    "loss": 0.3008865714073181,
    "learning_rate": 0.0006044679452012064,
    "timestep_mean": 456.15625,
    "epsilon_norm": 7.870365142822266,
    "prediction_norm": 6.469836235046387,
    "timestamp": "2025-07-14T12:34:04.232186"
  },
  {
    "epoch": 866,
    "batch_idx": 128,
    "loss": 0.13124015927314758,
    "learning_rate": 0.0006044679452012064,
    "timestep_mean": 705.21875,
    "epsilon_norm": 8.961592674255371,
    "prediction_norm": 8.311285972595215,
    "timestamp": "2025-07-14T12:34:04.237018"
  },
  {
    "epoch": 866,
    "batch_idx": 160,
    "loss": 0.18987709283828735,
    "learning_rate": 0.0006044679452012064,
    "timestep_mean": 536.6875,
    "epsilon_norm": 7.4749040603637695,
    "prediction_norm": 6.717467784881592,
    "timestamp": "2025-07-14T12:34:04.241984"
  },
  {
    "epoch": 866,
    "batch_idx": 192,
    "loss": 0.2173592746257782,
    "learning_rate": 0.0006044679452012064,
    "timestep_mean": 649.125,
    "epsilon_norm": 3.9480502605438232,
    "prediction_norm": 3.402954578399658,
    "timestamp": "2025-07-14T12:34:04.246843"
  },
  {
    "epoch": 867,
    "batch_idx": 0,
    "loss": 0.19517475366592407,
    "learning_rate": 0.0006036997527274896,
    "timestep_mean": 533.5625,
    "epsilon_norm": 7.313024520874023,
    "prediction_norm": 6.6828179359436035,
    "timestamp": "2025-07-14T12:34:04.251719"
  },
  {
    "epoch": 867,
    "batch_idx": 32,
    "loss": 0.2636374533176422,
    "learning_rate": 0.0006036997527274896,
    "timestep_mean": 490.59375,
    "epsilon_norm": 8.34088134765625,
    "prediction_norm": 7.310413837432861,
    "timestamp": "2025-07-14T12:34:04.256600"
  },
  {
    "epoch": 867,
    "batch_idx": 64,
    "loss": 0.2411094307899475,
    "learning_rate": 0.0006036997527274896,
    "timestep_mean": 428.375,
    "epsilon_norm": 6.871287822723389,
    "prediction_norm": 5.613307476043701,
    "timestamp": "2025-07-14T12:34:04.261444"
  },
  {
    "epoch": 867,
    "batch_idx": 96,
    "loss": 0.46917009353637695,
    "learning_rate": 0.0006036997527274896,
    "timestep_mean": 393.78125,
    "epsilon_norm": 8.092164993286133,
    "prediction_norm": 6.477256774902344,
    "timestamp": "2025-07-14T12:34:04.266317"
  },
  {
    "epoch": 867,
    "batch_idx": 128,
    "loss": 0.31212127208709717,
    "learning_rate": 0.0006036997527274896,
    "timestep_mean": 469.84375,
    "epsilon_norm": 6.310975551605225,
    "prediction_norm": 4.804286479949951,
    "timestamp": "2025-07-14T12:34:04.271373"
  },
  {
    "epoch": 867,
    "batch_idx": 160,
    "loss": 0.2436455637216568,
    "learning_rate": 0.0006036997527274896,
    "timestep_mean": 486.0625,
    "epsilon_norm": 7.3848876953125,
    "prediction_norm": 6.548411846160889,
    "timestamp": "2025-07-14T12:34:04.276219"
  },
  {
    "epoch": 867,
    "batch_idx": 192,
    "loss": 0.08201810717582703,
    "learning_rate": 0.0006036997527274896,
    "timestep_mean": 670.375,
    "epsilon_norm": 4.247315406799316,
    "prediction_norm": 4.067704677581787,
    "timestamp": "2025-07-14T12:34:04.281044"
  },
  {
    "epoch": 868,
    "batch_idx": 0,
    "loss": 0.2944106459617615,
    "learning_rate": 0.0006029313043849412,
    "timestep_mean": 566.03125,
    "epsilon_norm": 8.62108039855957,
    "prediction_norm": 7.005315780639648,
    "timestamp": "2025-07-14T12:34:04.285957"
  },
  {
    "epoch": 868,
    "batch_idx": 32,
    "loss": 0.16707076132297516,
    "learning_rate": 0.0006029313043849412,
    "timestep_mean": 527.3125,
    "epsilon_norm": 6.77486515045166,
    "prediction_norm": 6.244350910186768,
    "timestamp": "2025-07-14T12:34:04.290854"
  },
  {
    "epoch": 868,
    "batch_idx": 64,
    "loss": 0.36581480503082275,
    "learning_rate": 0.0006029313043849412,
    "timestep_mean": 465.84375,
    "epsilon_norm": 7.919800281524658,
    "prediction_norm": 5.421345233917236,
    "timestamp": "2025-07-14T12:34:04.295712"
  },
  {
    "epoch": 868,
    "batch_idx": 96,
    "loss": 0.23609787225723267,
    "learning_rate": 0.0006029313043849412,
    "timestep_mean": 500.53125,
    "epsilon_norm": 6.9521403312683105,
    "prediction_norm": 6.095463752746582,
    "timestamp": "2025-07-14T12:34:04.300668"
  },
  {
    "epoch": 868,
    "batch_idx": 128,
    "loss": 0.2536938786506653,
    "learning_rate": 0.0006029313043849412,
    "timestep_mean": 497.96875,
    "epsilon_norm": 7.362762451171875,
    "prediction_norm": 6.083246231079102,
    "timestamp": "2025-07-14T12:34:04.305532"
  },
  {
    "epoch": 868,
    "batch_idx": 160,
    "loss": 0.3520489037036896,
    "learning_rate": 0.0006029313043849412,
    "timestep_mean": 443.96875,
    "epsilon_norm": 7.480865001678467,
    "prediction_norm": 5.943732738494873,
    "timestamp": "2025-07-14T12:34:04.310415"
  },
  {
    "epoch": 868,
    "batch_idx": 192,
    "loss": 0.3377273976802826,
    "learning_rate": 0.0006029313043849412,
    "timestep_mean": 333.0,
    "epsilon_norm": 3.766184091567993,
    "prediction_norm": 3.1060523986816406,
    "timestamp": "2025-07-14T12:34:04.315267"
  },
  {
    "epoch": 869,
    "batch_idx": 0,
    "loss": 0.1910572350025177,
    "learning_rate": 0.0006021626020696314,
    "timestep_mean": 454.25,
    "epsilon_norm": 7.66657018661499,
    "prediction_norm": 6.604300498962402,
    "timestamp": "2025-07-14T12:34:04.320163"
  },
  {
    "epoch": 869,
    "batch_idx": 32,
    "loss": 0.33222195506095886,
    "learning_rate": 0.0006021626020696314,
    "timestep_mean": 614.84375,
    "epsilon_norm": 8.22673225402832,
    "prediction_norm": 7.070071220397949,
    "timestamp": "2025-07-14T12:34:04.325158"
  },
  {
    "epoch": 869,
    "batch_idx": 64,
    "loss": 0.2758095860481262,
    "learning_rate": 0.0006021626020696314,
    "timestep_mean": 500.71875,
    "epsilon_norm": 8.072166442871094,
    "prediction_norm": 6.221305847167969,
    "timestamp": "2025-07-14T12:34:04.330030"
  },
  {
    "epoch": 869,
    "batch_idx": 96,
    "loss": 0.13731837272644043,
    "learning_rate": 0.0006021626020696314,
    "timestep_mean": 464.5,
    "epsilon_norm": 6.9696269035339355,
    "prediction_norm": 6.183681488037109,
    "timestamp": "2025-07-14T12:34:04.334955"
  },
  {
    "epoch": 869,
    "batch_idx": 128,
    "loss": 0.13553224503993988,
    "learning_rate": 0.0006021626020696314,
    "timestep_mean": 572.3125,
    "epsilon_norm": 7.860003471374512,
    "prediction_norm": 7.1495466232299805,
    "timestamp": "2025-07-14T12:34:04.339843"
  },
  {
    "epoch": 869,
    "batch_idx": 160,
    "loss": 0.28435730934143066,
    "learning_rate": 0.0006021626020696314,
    "timestep_mean": 448.09375,
    "epsilon_norm": 7.754717826843262,
    "prediction_norm": 6.154643535614014,
    "timestamp": "2025-07-14T12:34:04.344697"
  },
  {
    "epoch": 869,
    "batch_idx": 192,
    "loss": 0.35539519786834717,
    "learning_rate": 0.0006021626020696314,
    "timestep_mean": 353.875,
    "epsilon_norm": 3.719144344329834,
    "prediction_norm": 2.828913688659668,
    "timestamp": "2025-07-14T12:34:04.349574"
  },
  {
    "epoch": 870,
    "batch_idx": 0,
    "loss": 0.43254518508911133,
    "learning_rate": 0.0006013936476782567,
    "timestep_mean": 483.53125,
    "epsilon_norm": 8.023724555969238,
    "prediction_norm": 5.948825836181641,
    "timestamp": "2025-07-14T12:34:04.354585"
  },
  {
    "epoch": 870,
    "batch_idx": 32,
    "loss": 0.28953152894973755,
    "learning_rate": 0.0006013936476782567,
    "timestep_mean": 466.09375,
    "epsilon_norm": 7.058306694030762,
    "prediction_norm": 5.683963775634766,
    "timestamp": "2025-07-14T12:34:04.359416"
  },
  {
    "epoch": 870,
    "batch_idx": 64,
    "loss": 0.33326154947280884,
    "learning_rate": 0.0006013936476782567,
    "timestep_mean": 535.09375,
    "epsilon_norm": 6.536019802093506,
    "prediction_norm": 5.5117292404174805,
    "timestamp": "2025-07-14T12:34:04.364269"
  },
  {
    "epoch": 870,
    "batch_idx": 96,
    "loss": 0.35796087980270386,
    "learning_rate": 0.0006013936476782567,
    "timestep_mean": 407.53125,
    "epsilon_norm": 7.886646270751953,
    "prediction_norm": 6.425854206085205,
    "timestamp": "2025-07-14T12:34:04.369158"
  },
  {
    "epoch": 870,
    "batch_idx": 128,
    "loss": 0.21537499129772186,
    "learning_rate": 0.0006013936476782567,
    "timestep_mean": 518.96875,
    "epsilon_norm": 7.849514007568359,
    "prediction_norm": 7.075221538543701,
    "timestamp": "2025-07-14T12:34:04.374041"
  },
  {
    "epoch": 870,
    "batch_idx": 160,
    "loss": 0.44480371475219727,
    "learning_rate": 0.0006013936476782567,
    "timestep_mean": 473.4375,
    "epsilon_norm": 7.8826141357421875,
    "prediction_norm": 5.902328968048096,
    "timestamp": "2025-07-14T12:34:04.378954"
  },
  {
    "epoch": 870,
    "batch_idx": 192,
    "loss": 0.39780622720718384,
    "learning_rate": 0.0006013936476782567,
    "timestep_mean": 552.875,
    "epsilon_norm": 3.5266497135162354,
    "prediction_norm": 2.9284181594848633,
    "timestamp": "2025-07-14T12:34:04.383909"
  },
  {
    "epoch": 871,
    "batch_idx": 0,
    "loss": 0.42545294761657715,
    "learning_rate": 0.0006006244431081357,
    "timestep_mean": 450.15625,
    "epsilon_norm": 7.247486114501953,
    "prediction_norm": 5.600124835968018,
    "timestamp": "2025-07-14T12:34:04.388831"
  },
  {
    "epoch": 871,
    "batch_idx": 32,
    "loss": 0.3784976005554199,
    "learning_rate": 0.0006006244431081357,
    "timestep_mean": 400.1875,
    "epsilon_norm": 7.187625408172607,
    "prediction_norm": 6.388879776000977,
    "timestamp": "2025-07-14T12:34:04.393695"
  },
  {
    "epoch": 871,
    "batch_idx": 64,
    "loss": 0.14385822415351868,
    "learning_rate": 0.0006006244431081357,
    "timestep_mean": 586.875,
    "epsilon_norm": 7.134423732757568,
    "prediction_norm": 6.354145526885986,
    "timestamp": "2025-07-14T12:34:04.398616"
  },
  {
    "epoch": 871,
    "batch_idx": 96,
    "loss": 0.40886592864990234,
    "learning_rate": 0.0006006244431081357,
    "timestep_mean": 430.25,
    "epsilon_norm": 9.136826515197754,
    "prediction_norm": 7.479814529418945,
    "timestamp": "2025-07-14T12:34:04.403507"
  },
  {
    "epoch": 871,
    "batch_idx": 128,
    "loss": 0.4531228542327881,
    "learning_rate": 0.0006006244431081357,
    "timestep_mean": 508.375,
    "epsilon_norm": 8.734681129455566,
    "prediction_norm": 6.594300746917725,
    "timestamp": "2025-07-14T12:34:04.409112"
  },
  {
    "epoch": 871,
    "batch_idx": 160,
    "loss": 0.45812419056892395,
    "learning_rate": 0.0006006244431081357,
    "timestep_mean": 370.125,
    "epsilon_norm": 7.910684108734131,
    "prediction_norm": 5.650044918060303,
    "timestamp": "2025-07-14T12:34:04.415158"
  },
  {
    "epoch": 871,
    "batch_idx": 192,
    "loss": 0.2961413562297821,
    "learning_rate": 0.0006006244431081357,
    "timestep_mean": 435.875,
    "epsilon_norm": 3.676145076751709,
    "prediction_norm": 2.477281332015991,
    "timestamp": "2025-07-14T12:34:04.420419"
  },
  {
    "epoch": 872,
    "batch_idx": 0,
    "loss": 0.5236262679100037,
    "learning_rate": 0.0005998549902572041,
    "timestep_mean": 424.125,
    "epsilon_norm": 9.004857063293457,
    "prediction_norm": 6.569106578826904,
    "timestamp": "2025-07-14T12:34:04.425918"
  },
  {
    "epoch": 872,
    "batch_idx": 32,
    "loss": 0.39329373836517334,
    "learning_rate": 0.0005998549902572041,
    "timestep_mean": 449.875,
    "epsilon_norm": 8.487093925476074,
    "prediction_norm": 7.4339094161987305,
    "timestamp": "2025-07-14T12:34:04.431261"
  },
  {
    "epoch": 872,
    "batch_idx": 64,
    "loss": 0.42579221725463867,
    "learning_rate": 0.0005998549902572041,
    "timestep_mean": 519.46875,
    "epsilon_norm": 8.726336479187012,
    "prediction_norm": 7.2765936851501465,
    "timestamp": "2025-07-14T12:34:04.436423"
  },
  {
    "epoch": 872,
    "batch_idx": 96,
    "loss": 0.1954362690448761,
    "learning_rate": 0.0005998549902572041,
    "timestep_mean": 467.0625,
    "epsilon_norm": 7.7514424324035645,
    "prediction_norm": 7.04330587387085,
    "timestamp": "2025-07-14T12:34:04.450258"
  },
  {
    "epoch": 872,
    "batch_idx": 128,
    "loss": 0.41769230365753174,
    "learning_rate": 0.0005998549902572041,
    "timestep_mean": 444.78125,
    "epsilon_norm": 8.241570472717285,
    "prediction_norm": 6.671504974365234,
    "timestamp": "2025-07-14T12:34:04.456633"
  },
  {
    "epoch": 872,
    "batch_idx": 160,
    "loss": 0.3757401704788208,
    "learning_rate": 0.0005998549902572041,
    "timestep_mean": 461.5,
    "epsilon_norm": 9.013040542602539,
    "prediction_norm": 7.504664897918701,
    "timestamp": "2025-07-14T12:34:04.461726"
  },
  {
    "epoch": 872,
    "batch_idx": 192,
    "loss": 0.27993884682655334,
    "learning_rate": 0.0005998549902572041,
    "timestep_mean": 547.5,
    "epsilon_norm": 3.5832042694091797,
    "prediction_norm": 3.0157172679901123,
    "timestamp": "2025-07-14T12:34:04.466867"
  },
  {
    "epoch": 873,
    "batch_idx": 0,
    "loss": 0.5716372132301331,
    "learning_rate": 0.0005990852910240102,
    "timestep_mean": 458.875,
    "epsilon_norm": 8.647539138793945,
    "prediction_norm": 6.758501052856445,
    "timestamp": "2025-07-14T12:34:04.471802"
  },
  {
    "epoch": 873,
    "batch_idx": 32,
    "loss": 0.5812454223632812,
    "learning_rate": 0.0005990852910240102,
    "timestep_mean": 439.625,
    "epsilon_norm": 7.7881622314453125,
    "prediction_norm": 4.869889736175537,
    "timestamp": "2025-07-14T12:34:04.476888"
  },
  {
    "epoch": 873,
    "batch_idx": 64,
    "loss": 0.24206674098968506,
    "learning_rate": 0.0005990852910240102,
    "timestep_mean": 461.65625,
    "epsilon_norm": 7.87725830078125,
    "prediction_norm": 6.531732559204102,
    "timestamp": "2025-07-14T12:34:04.482663"
  },
  {
    "epoch": 873,
    "batch_idx": 96,
    "loss": 0.2131389081478119,
    "learning_rate": 0.0005990852910240102,
    "timestep_mean": 556.71875,
    "epsilon_norm": 6.772494316101074,
    "prediction_norm": 6.10390043258667,
    "timestamp": "2025-07-14T12:34:04.488202"
  },
  {
    "epoch": 873,
    "batch_idx": 128,
    "loss": 0.31534266471862793,
    "learning_rate": 0.0005990852910240102,
    "timestep_mean": 443.84375,
    "epsilon_norm": 8.190598487854004,
    "prediction_norm": 6.553474426269531,
    "timestamp": "2025-07-14T12:34:04.494323"
  },
  {
    "epoch": 873,
    "batch_idx": 160,
    "loss": 0.3575890064239502,
    "learning_rate": 0.0005990852910240102,
    "timestep_mean": 483.65625,
    "epsilon_norm": 8.422761917114258,
    "prediction_norm": 7.131953716278076,
    "timestamp": "2025-07-14T12:34:04.500530"
  },
  {
    "epoch": 873,
    "batch_idx": 192,
    "loss": 0.3934074342250824,
    "learning_rate": 0.0005990852910240102,
    "timestep_mean": 378.25,
    "epsilon_norm": 2.8273086547851562,
    "prediction_norm": 2.6823031902313232,
    "timestamp": "2025-07-14T12:34:04.507132"
  },
  {
    "epoch": 874,
    "batch_idx": 0,
    "loss": 0.36397433280944824,
    "learning_rate": 0.0005983153473077103,
    "timestep_mean": 513.15625,
    "epsilon_norm": 8.82901382446289,
    "prediction_norm": 7.0528950691223145,
    "timestamp": "2025-07-14T12:34:04.512660"
  },
  {
    "epoch": 874,
    "batch_idx": 32,
    "loss": 0.2608790993690491,
    "learning_rate": 0.0005983153473077103,
    "timestep_mean": 576.90625,
    "epsilon_norm": 8.511512756347656,
    "prediction_norm": 7.017621994018555,
    "timestamp": "2025-07-14T12:34:04.517926"
  },
  {
    "epoch": 874,
    "batch_idx": 64,
    "loss": 0.26048392057418823,
    "learning_rate": 0.0005983153473077103,
    "timestep_mean": 525.46875,
    "epsilon_norm": 7.512396812438965,
    "prediction_norm": 5.819273948669434,
    "timestamp": "2025-07-14T12:34:04.523459"
  },
  {
    "epoch": 874,
    "batch_idx": 96,
    "loss": 0.35289785265922546,
    "learning_rate": 0.0005983153473077103,
    "timestep_mean": 465.0,
    "epsilon_norm": 8.848163604736328,
    "prediction_norm": 7.175539970397949,
    "timestamp": "2025-07-14T12:34:04.528813"
  },
  {
    "epoch": 874,
    "batch_idx": 128,
    "loss": 0.24700897932052612,
    "learning_rate": 0.0005983153473077103,
    "timestep_mean": 533.03125,
    "epsilon_norm": 8.607337951660156,
    "prediction_norm": 7.30062198638916,
    "timestamp": "2025-07-14T12:34:04.534268"
  },
  {
    "epoch": 874,
    "batch_idx": 160,
    "loss": 0.2845134139060974,
    "learning_rate": 0.0005983153473077103,
    "timestep_mean": 544.40625,
    "epsilon_norm": 8.421741485595703,
    "prediction_norm": 7.055161952972412,
    "timestamp": "2025-07-14T12:34:04.539788"
  },
  {
    "epoch": 874,
    "batch_idx": 192,
    "loss": 0.16774582862854004,
    "learning_rate": 0.0005983153473077103,
    "timestep_mean": 498.0,
    "epsilon_norm": 2.5369369983673096,
    "prediction_norm": 2.7797138690948486,
    "timestamp": "2025-07-14T12:34:04.545218"
  },
  {
    "epoch": 875,
    "batch_idx": 0,
    "loss": 0.13153305649757385,
    "learning_rate": 0.0005975451610080645,
    "timestep_mean": 546.09375,
    "epsilon_norm": 7.941908359527588,
    "prediction_norm": 6.731374263763428,
    "timestamp": "2025-07-14T12:34:04.550545"
  },
  {
    "epoch": 875,
    "batch_idx": 32,
    "loss": 0.2942652702331543,
    "learning_rate": 0.0005975451610080645,
    "timestep_mean": 491.9375,
    "epsilon_norm": 8.511236190795898,
    "prediction_norm": 7.1376566886901855,
    "timestamp": "2025-07-14T12:34:04.555437"
  },
  {
    "epoch": 875,
    "batch_idx": 64,
    "loss": 0.20611514151096344,
    "learning_rate": 0.0005975451610080645,
    "timestep_mean": 562.3125,
    "epsilon_norm": 7.522027015686035,
    "prediction_norm": 6.2255449295043945,
    "timestamp": "2025-07-14T12:34:04.560409"
  },
  {
    "epoch": 875,
    "batch_idx": 96,
    "loss": 0.3057236671447754,
    "learning_rate": 0.0005975451610080645,
    "timestep_mean": 478.28125,
    "epsilon_norm": 7.627872467041016,
    "prediction_norm": 6.240911960601807,
    "timestamp": "2025-07-14T12:34:04.566115"
  },
  {
    "epoch": 875,
    "batch_idx": 128,
    "loss": 0.3121848404407501,
    "learning_rate": 0.0005975451610080645,
    "timestep_mean": 484.53125,
    "epsilon_norm": 8.040487289428711,
    "prediction_norm": 6.686014175415039,
    "timestamp": "2025-07-14T12:34:04.571043"
  },
  {
    "epoch": 875,
    "batch_idx": 160,
    "loss": 0.30724722146987915,
    "learning_rate": 0.0005975451610080645,
    "timestep_mean": 511.875,
    "epsilon_norm": 7.786394119262695,
    "prediction_norm": 6.594461441040039,
    "timestamp": "2025-07-14T12:34:04.576033"
  },
  {
    "epoch": 875,
    "batch_idx": 192,
    "loss": 0.30188190937042236,
    "learning_rate": 0.0005975451610080645,
    "timestep_mean": 389.625,
    "epsilon_norm": 4.250234603881836,
    "prediction_norm": 3.129370927810669,
    "timestamp": "2025-07-14T12:34:04.580945"
  },
  {
    "epoch": 876,
    "batch_idx": 0,
    "loss": 0.27138322591781616,
    "learning_rate": 0.0005967747340254305,
    "timestep_mean": 487.3125,
    "epsilon_norm": 8.136119842529297,
    "prediction_norm": 7.443118095397949,
    "timestamp": "2025-07-14T12:34:04.586930"
  },
  {
    "epoch": 876,
    "batch_idx": 32,
    "loss": 0.19057011604309082,
    "learning_rate": 0.0005967747340254305,
    "timestep_mean": 557.40625,
    "epsilon_norm": 8.27188777923584,
    "prediction_norm": 6.942891597747803,
    "timestamp": "2025-07-14T12:34:04.591866"
  },
  {
    "epoch": 876,
    "batch_idx": 64,
    "loss": 0.3437979817390442,
    "learning_rate": 0.0005967747340254305,
    "timestep_mean": 468.625,
    "epsilon_norm": 7.4831929206848145,
    "prediction_norm": 5.703573703765869,
    "timestamp": "2025-07-14T12:34:04.596734"
  },
  {
    "epoch": 876,
    "batch_idx": 96,
    "loss": 0.27633529901504517,
    "learning_rate": 0.0005967747340254305,
    "timestep_mean": 545.96875,
    "epsilon_norm": 7.805829048156738,
    "prediction_norm": 6.981980800628662,
    "timestamp": "2025-07-14T12:34:04.606338"
  },
  {
    "epoch": 876,
    "batch_idx": 128,
    "loss": 0.2706936001777649,
    "learning_rate": 0.0005967747340254305,
    "timestep_mean": 504.125,
    "epsilon_norm": 7.309593677520752,
    "prediction_norm": 6.101138591766357,
    "timestamp": "2025-07-14T12:34:04.611875"
  },
  {
    "epoch": 876,
    "batch_idx": 160,
    "loss": 0.3370193839073181,
    "learning_rate": 0.0005967747340254305,
    "timestep_mean": 454.875,
    "epsilon_norm": 8.284589767456055,
    "prediction_norm": 6.784708023071289,
    "timestamp": "2025-07-14T12:34:04.616973"
  },
  {
    "epoch": 876,
    "batch_idx": 192,
    "loss": 0.14194487035274506,
    "learning_rate": 0.0005967747340254305,
    "timestep_mean": 518.5,
    "epsilon_norm": 3.0474417209625244,
    "prediction_norm": 3.019174575805664,
    "timestamp": "2025-07-14T12:34:04.622273"
  },
  {
    "epoch": 877,
    "batch_idx": 0,
    "loss": 0.21798336505889893,
    "learning_rate": 0.0005960040682607603,
    "timestep_mean": 547.03125,
    "epsilon_norm": 7.253439903259277,
    "prediction_norm": 5.809142112731934,
    "timestamp": "2025-07-14T12:34:04.627723"
  },
  {
    "epoch": 877,
    "batch_idx": 32,
    "loss": 0.26160821318626404,
    "learning_rate": 0.0005960040682607603,
    "timestep_mean": 567.84375,
    "epsilon_norm": 8.02279281616211,
    "prediction_norm": 6.558597087860107,
    "timestamp": "2025-07-14T12:34:04.633034"
  },
  {
    "epoch": 877,
    "batch_idx": 64,
    "loss": 0.3182348608970642,
    "learning_rate": 0.0005960040682607603,
    "timestep_mean": 529.03125,
    "epsilon_norm": 8.01220417022705,
    "prediction_norm": 5.934548854827881,
    "timestamp": "2025-07-14T12:34:04.638459"
  },
  {
    "epoch": 877,
    "batch_idx": 96,
    "loss": 0.20628474652767181,
    "learning_rate": 0.0005960040682607603,
    "timestep_mean": 539.875,
    "epsilon_norm": 8.730560302734375,
    "prediction_norm": 8.197406768798828,
    "timestamp": "2025-07-14T12:34:04.644416"
  },
  {
    "epoch": 877,
    "batch_idx": 128,
    "loss": 0.1620400846004486,
    "learning_rate": 0.0005960040682607603,
    "timestep_mean": 452.28125,
    "epsilon_norm": 7.244598388671875,
    "prediction_norm": 6.487936496734619,
    "timestamp": "2025-07-14T12:34:04.649782"
  },
  {
    "epoch": 877,
    "batch_idx": 160,
    "loss": 0.34882500767707825,
    "learning_rate": 0.0005960040682607603,
    "timestep_mean": 439.40625,
    "epsilon_norm": 6.110475540161133,
    "prediction_norm": 4.556323051452637,
    "timestamp": "2025-07-14T12:34:04.655008"
  },
  {
    "epoch": 877,
    "batch_idx": 192,
    "loss": 0.2295791506767273,
    "learning_rate": 0.0005960040682607603,
    "timestep_mean": 507.625,
    "epsilon_norm": 2.365696430206299,
    "prediction_norm": 2.6848108768463135,
    "timestamp": "2025-07-14T12:34:04.660088"
  },
  {
    "epoch": 878,
    "batch_idx": 0,
    "loss": 0.2928806245326996,
    "learning_rate": 0.0005952331656155954,
    "timestep_mean": 513.71875,
    "epsilon_norm": 8.360697746276855,
    "prediction_norm": 6.860109806060791,
    "timestamp": "2025-07-14T12:34:04.666294"
  },
  {
    "epoch": 878,
    "batch_idx": 32,
    "loss": 0.20392626523971558,
    "learning_rate": 0.0005952331656155954,
    "timestep_mean": 498.25,
    "epsilon_norm": 7.3214240074157715,
    "prediction_norm": 6.5270233154296875,
    "timestamp": "2025-07-14T12:34:04.671360"
  },
  {
    "epoch": 878,
    "batch_idx": 64,
    "loss": 0.21836887300014496,
    "learning_rate": 0.0005952331656155954,
    "timestep_mean": 483.34375,
    "epsilon_norm": 8.892023086547852,
    "prediction_norm": 7.680006504058838,
    "timestamp": "2025-07-14T12:34:04.676300"
  },
  {
    "epoch": 878,
    "batch_idx": 96,
    "loss": 0.14803855121135712,
    "learning_rate": 0.0005952331656155954,
    "timestep_mean": 512.03125,
    "epsilon_norm": 7.124700546264648,
    "prediction_norm": 6.314462661743164,
    "timestamp": "2025-07-14T12:34:04.681650"
  },
  {
    "epoch": 878,
    "batch_idx": 128,
    "loss": 0.21934527158737183,
    "learning_rate": 0.0005952331656155954,
    "timestep_mean": 544.71875,
    "epsilon_norm": 8.522716522216797,
    "prediction_norm": 6.982369899749756,
    "timestamp": "2025-07-14T12:34:04.686997"
  },
  {
    "epoch": 878,
    "batch_idx": 160,
    "loss": 0.21537981927394867,
    "learning_rate": 0.0005952331656155954,
    "timestep_mean": 512.78125,
    "epsilon_norm": 8.849898338317871,
    "prediction_norm": 7.838700771331787,
    "timestamp": "2025-07-14T12:34:04.692198"
  },
  {
    "epoch": 878,
    "batch_idx": 192,
    "loss": 0.2252904623746872,
    "learning_rate": 0.0005952331656155954,
    "timestep_mean": 508.5,
    "epsilon_norm": 4.314293384552002,
    "prediction_norm": 3.203684091567993,
    "timestamp": "2025-07-14T12:34:04.697267"
  },
  {
    "epoch": 879,
    "batch_idx": 0,
    "loss": 0.13467815518379211,
    "learning_rate": 0.0005944620279920611,
    "timestep_mean": 513.59375,
    "epsilon_norm": 6.94494104385376,
    "prediction_norm": 5.924540996551514,
    "timestamp": "2025-07-14T12:34:04.702514"
  },
  {
    "epoch": 879,
    "batch_idx": 32,
    "loss": 0.15639379620552063,
    "learning_rate": 0.0005944620279920611,
    "timestep_mean": 582.6875,
    "epsilon_norm": 8.983144760131836,
    "prediction_norm": 8.702973365783691,
    "timestamp": "2025-07-14T12:34:04.707892"
  },
  {
    "epoch": 879,
    "batch_idx": 64,
    "loss": 0.2583826780319214,
    "learning_rate": 0.0005944620279920611,
    "timestep_mean": 589.46875,
    "epsilon_norm": 9.664891242980957,
    "prediction_norm": 8.842604637145996,
    "timestamp": "2025-07-14T12:34:04.713091"
  },
  {
    "epoch": 879,
    "batch_idx": 96,
    "loss": 0.18860986828804016,
    "learning_rate": 0.0005944620279920611,
    "timestep_mean": 561.90625,
    "epsilon_norm": 7.4078168869018555,
    "prediction_norm": 6.826412677764893,
    "timestamp": "2025-07-14T12:34:04.718724"
  },
  {
    "epoch": 879,
    "batch_idx": 128,
    "loss": 0.24867300689220428,
    "learning_rate": 0.0005944620279920611,
    "timestep_mean": 516.65625,
    "epsilon_norm": 7.8040900230407715,
    "prediction_norm": 6.269689559936523,
    "timestamp": "2025-07-14T12:34:04.724195"
  },
  {
    "epoch": 879,
    "batch_idx": 160,
    "loss": 0.5554059743881226,
    "learning_rate": 0.0005944620279920611,
    "timestep_mean": 470.625,
    "epsilon_norm": 8.46070384979248,
    "prediction_norm": 6.282062530517578,
    "timestamp": "2025-07-14T12:34:04.729266"
  },
  {
    "epoch": 879,
    "batch_idx": 192,
    "loss": 0.36994457244873047,
    "learning_rate": 0.0005944620279920611,
    "timestep_mean": 360.125,
    "epsilon_norm": 3.418628215789795,
    "prediction_norm": 2.4792869091033936,
    "timestamp": "2025-07-14T12:34:04.734370"
  },
  {
    "epoch": 880,
    "batch_idx": 0,
    "loss": 0.18502390384674072,
    "learning_rate": 0.0005936906572928626,
    "timestep_mean": 484.0,
    "epsilon_norm": 7.311773777008057,
    "prediction_norm": 6.582951068878174,
    "timestamp": "2025-07-14T12:34:04.739643"
  },
  {
    "epoch": 880,
    "batch_idx": 32,
    "loss": 0.32603567838668823,
    "learning_rate": 0.0005936906572928626,
    "timestep_mean": 534.59375,
    "epsilon_norm": 8.194174766540527,
    "prediction_norm": 7.356791973114014,
    "timestamp": "2025-07-14T12:34:04.744715"
  },
  {
    "epoch": 880,
    "batch_idx": 64,
    "loss": 0.1954077035188675,
    "learning_rate": 0.0005936906572928626,
    "timestep_mean": 581.15625,
    "epsilon_norm": 7.6260576248168945,
    "prediction_norm": 7.051142692565918,
    "timestamp": "2025-07-14T12:34:04.749993"
  },
  {
    "epoch": 880,
    "batch_idx": 96,
    "loss": 0.16980111598968506,
    "learning_rate": 0.0005936906572928626,
    "timestep_mean": 481.78125,
    "epsilon_norm": 7.254738807678223,
    "prediction_norm": 7.0245232582092285,
    "timestamp": "2025-07-14T12:34:04.755167"
  },
  {
    "epoch": 880,
    "batch_idx": 128,
    "loss": 0.2953186631202698,
    "learning_rate": 0.0005936906572928626,
    "timestep_mean": 536.9375,
    "epsilon_norm": 9.140475273132324,
    "prediction_norm": 8.673762321472168,
    "timestamp": "2025-07-14T12:34:04.760176"
  },
  {
    "epoch": 880,
    "batch_idx": 160,
    "loss": 0.23070411384105682,
    "learning_rate": 0.0005936906572928626,
    "timestep_mean": 487.78125,
    "epsilon_norm": 8.462884902954102,
    "prediction_norm": 8.320418357849121,
    "timestamp": "2025-07-14T12:34:04.765321"
  },
  {
    "epoch": 880,
    "batch_idx": 192,
    "loss": 0.21160058677196503,
    "learning_rate": 0.0005936906572928626,
    "timestep_mean": 493.375,
    "epsilon_norm": 4.877245903015137,
    "prediction_norm": 4.879324436187744,
    "timestamp": "2025-07-14T12:34:04.770512"
  },
  {
    "epoch": 881,
    "batch_idx": 0,
    "loss": 0.3548499345779419,
    "learning_rate": 0.000592919055421281,
    "timestep_mean": 531.125,
    "epsilon_norm": 8.22065544128418,
    "prediction_norm": 6.959688186645508,
    "timestamp": "2025-07-14T12:34:04.775757"
  },
  {
    "epoch": 881,
    "batch_idx": 32,
    "loss": 0.3148602247238159,
    "learning_rate": 0.000592919055421281,
    "timestep_mean": 451.34375,
    "epsilon_norm": 8.572036743164062,
    "prediction_norm": 7.19788932800293,
    "timestamp": "2025-07-14T12:34:04.780996"
  },
  {
    "epoch": 881,
    "batch_idx": 64,
    "loss": 0.180312842130661,
    "learning_rate": 0.000592919055421281,
    "timestep_mean": 586.0,
    "epsilon_norm": 7.491597652435303,
    "prediction_norm": 6.784143447875977,
    "timestamp": "2025-07-14T12:34:04.785915"
  },
  {
    "epoch": 881,
    "batch_idx": 96,
    "loss": 0.27622541785240173,
    "learning_rate": 0.000592919055421281,
    "timestep_mean": 482.65625,
    "epsilon_norm": 7.973775386810303,
    "prediction_norm": 6.859677314758301,
    "timestamp": "2025-07-14T12:34:04.790956"
  },
  {
    "epoch": 881,
    "batch_idx": 128,
    "loss": 0.1090923398733139,
    "learning_rate": 0.000592919055421281,
    "timestep_mean": 508.21875,
    "epsilon_norm": 6.788036823272705,
    "prediction_norm": 6.362069129943848,
    "timestamp": "2025-07-14T12:34:04.796022"
  },
  {
    "epoch": 881,
    "batch_idx": 160,
    "loss": 0.22058862447738647,
    "learning_rate": 0.000592919055421281,
    "timestep_mean": 478.90625,
    "epsilon_norm": 7.122143268585205,
    "prediction_norm": 5.627779960632324,
    "timestamp": "2025-07-14T12:34:04.801229"
  },
  {
    "epoch": 881,
    "batch_idx": 192,
    "loss": 0.6209877133369446,
    "learning_rate": 0.000592919055421281,
    "timestep_mean": 539.125,
    "epsilon_norm": 4.437323570251465,
    "prediction_norm": 2.8416223526000977,
    "timestamp": "2025-07-14T12:34:04.809044"
  },
  {
    "epoch": 882,
    "batch_idx": 0,
    "loss": 0.39187875390052795,
    "learning_rate": 0.000592147224281167,
    "timestep_mean": 517.65625,
    "epsilon_norm": 8.612687110900879,
    "prediction_norm": 6.755123615264893,
    "timestamp": "2025-07-14T12:34:04.814646"
  },
  {
    "epoch": 882,
    "batch_idx": 32,
    "loss": 0.42549556493759155,
    "learning_rate": 0.000592147224281167,
    "timestep_mean": 405.71875,
    "epsilon_norm": 7.948147773742676,
    "prediction_norm": 5.45650577545166,
    "timestamp": "2025-07-14T12:34:04.820175"
  },
  {
    "epoch": 882,
    "batch_idx": 64,
    "loss": 0.19593694806098938,
    "learning_rate": 0.000592147224281167,
    "timestep_mean": 486.40625,
    "epsilon_norm": 7.693824768066406,
    "prediction_norm": 6.73314094543457,
    "timestamp": "2025-07-14T12:34:04.825615"
  },
  {
    "epoch": 882,
    "batch_idx": 96,
    "loss": 0.3276313543319702,
    "learning_rate": 0.000592147224281167,
    "timestep_mean": 457.78125,
    "epsilon_norm": 8.344167709350586,
    "prediction_norm": 6.604665756225586,
    "timestamp": "2025-07-14T12:34:04.831274"
  },
  {
    "epoch": 882,
    "batch_idx": 128,
    "loss": 0.18982873857021332,
    "learning_rate": 0.000592147224281167,
    "timestep_mean": 519.40625,
    "epsilon_norm": 7.482949733734131,
    "prediction_norm": 6.47163724899292,
    "timestamp": "2025-07-14T12:34:04.844123"
  },
  {
    "epoch": 882,
    "batch_idx": 160,
    "loss": 0.20012468099594116,
    "learning_rate": 0.000592147224281167,
    "timestep_mean": 566.0625,
    "epsilon_norm": 8.35251522064209,
    "prediction_norm": 6.646849155426025,
    "timestamp": "2025-07-14T12:34:04.849905"
  },
  {
    "epoch": 882,
    "batch_idx": 192,
    "loss": 0.0933016836643219,
    "learning_rate": 0.000592147224281167,
    "timestep_mean": 465.25,
    "epsilon_norm": 3.3779876232147217,
    "prediction_norm": 3.2141366004943848,
    "timestamp": "2025-07-14T12:34:04.855087"
  },
  {
    "epoch": 883,
    "batch_idx": 0,
    "loss": 0.2784728407859802,
    "learning_rate": 0.0005913751657769371,
    "timestep_mean": 518.78125,
    "epsilon_norm": 8.165013313293457,
    "prediction_norm": 6.497565746307373,
    "timestamp": "2025-07-14T12:34:04.860565"
  },
  {
    "epoch": 883,
    "batch_idx": 32,
    "loss": 0.09915634989738464,
    "learning_rate": 0.0005913751657769371,
    "timestep_mean": 614.71875,
    "epsilon_norm": 8.662100791931152,
    "prediction_norm": 7.993589878082275,
    "timestamp": "2025-07-14T12:34:04.866074"
  },
  {
    "epoch": 883,
    "batch_idx": 64,
    "loss": 0.1887303739786148,
    "learning_rate": 0.0005913751657769371,
    "timestep_mean": 474.03125,
    "epsilon_norm": 8.382662773132324,
    "prediction_norm": 7.087419033050537,
    "timestamp": "2025-07-14T12:34:04.871331"
  },
  {
    "epoch": 883,
    "batch_idx": 96,
    "loss": 0.2876685857772827,
    "learning_rate": 0.0005913751657769371,
    "timestep_mean": 443.8125,
    "epsilon_norm": 7.993464469909668,
    "prediction_norm": 6.622313976287842,
    "timestamp": "2025-07-14T12:34:04.876504"
  },
  {
    "epoch": 883,
    "batch_idx": 128,
    "loss": 0.37966442108154297,
    "learning_rate": 0.0005913751657769371,
    "timestep_mean": 477.21875,
    "epsilon_norm": 7.82361364364624,
    "prediction_norm": 6.149179458618164,
    "timestamp": "2025-07-14T12:34:04.881676"
  },
  {
    "epoch": 883,
    "batch_idx": 160,
    "loss": 0.2736767530441284,
    "learning_rate": 0.0005913751657769371,
    "timestep_mean": 488.9375,
    "epsilon_norm": 6.450148105621338,
    "prediction_norm": 4.8393874168396,
    "timestamp": "2025-07-14T12:34:04.886871"
  },
  {
    "epoch": 883,
    "batch_idx": 192,
    "loss": 0.7577924728393555,
    "learning_rate": 0.0005913751657769371,
    "timestep_mean": 404.375,
    "epsilon_norm": 5.668593883514404,
    "prediction_norm": 4.246513366699219,
    "timestamp": "2025-07-14T12:34:04.892017"
  },
  {
    "epoch": 884,
    "batch_idx": 0,
    "loss": 0.30160588026046753,
    "learning_rate": 0.000590602881813569,
    "timestep_mean": 502.84375,
    "epsilon_norm": 9.03055191040039,
    "prediction_norm": 8.1432523727417,
    "timestamp": "2025-07-14T12:34:04.897135"
  },
  {
    "epoch": 884,
    "batch_idx": 32,
    "loss": 0.4287962317466736,
    "learning_rate": 0.000590602881813569,
    "timestep_mean": 422.71875,
    "epsilon_norm": 7.101417064666748,
    "prediction_norm": 5.153651237487793,
    "timestamp": "2025-07-14T12:34:04.902303"
  },
  {
    "epoch": 884,
    "batch_idx": 64,
    "loss": 0.3043827712535858,
    "learning_rate": 0.000590602881813569,
    "timestep_mean": 514.46875,
    "epsilon_norm": 7.955622673034668,
    "prediction_norm": 7.155517578125,
    "timestamp": "2025-07-14T12:34:04.907403"
  },
  {
    "epoch": 884,
    "batch_idx": 96,
    "loss": 0.26986807584762573,
    "learning_rate": 0.000590602881813569,
    "timestep_mean": 591.59375,
    "epsilon_norm": 8.058032035827637,
    "prediction_norm": 6.880631923675537,
    "timestamp": "2025-07-14T12:34:04.912449"
  },
  {
    "epoch": 884,
    "batch_idx": 128,
    "loss": 0.2908209264278412,
    "learning_rate": 0.000590602881813569,
    "timestep_mean": 468.15625,
    "epsilon_norm": 8.003705978393555,
    "prediction_norm": 6.4341535568237305,
    "timestamp": "2025-07-14T12:34:04.917582"
  },
  {
    "epoch": 884,
    "batch_idx": 160,
    "loss": 0.31318825483322144,
    "learning_rate": 0.000590602881813569,
    "timestep_mean": 453.71875,
    "epsilon_norm": 8.4938325881958,
    "prediction_norm": 6.929808616638184,
    "timestamp": "2025-07-14T12:34:04.922620"
  },
  {
    "epoch": 884,
    "batch_idx": 192,
    "loss": 0.2605206370353699,
    "learning_rate": 0.000590602881813569,
    "timestep_mean": 509.625,
    "epsilon_norm": 3.0763750076293945,
    "prediction_norm": 2.4125208854675293,
    "timestamp": "2025-07-14T12:34:04.927801"
  },
  {
    "epoch": 885,
    "batch_idx": 0,
    "loss": 0.30825182795524597,
    "learning_rate": 0.0005898303742965966,
    "timestep_mean": 496.5,
    "epsilon_norm": 7.738811016082764,
    "prediction_norm": 7.4361348152160645,
    "timestamp": "2025-07-14T12:34:04.932950"
  },
  {
    "epoch": 885,
    "batch_idx": 32,
    "loss": 0.33224642276763916,
    "learning_rate": 0.0005898303742965966,
    "timestep_mean": 498.4375,
    "epsilon_norm": 8.558693885803223,
    "prediction_norm": 7.91837215423584,
    "timestamp": "2025-07-14T12:34:04.938321"
  },
  {
    "epoch": 885,
    "batch_idx": 64,
    "loss": 0.4006165862083435,
    "learning_rate": 0.0005898303742965966,
    "timestep_mean": 415.0,
    "epsilon_norm": 7.332511901855469,
    "prediction_norm": 6.57857084274292,
    "timestamp": "2025-07-14T12:34:04.943356"
  },
  {
    "epoch": 885,
    "batch_idx": 96,
    "loss": 0.46751970052719116,
    "learning_rate": 0.0005898303742965966,
    "timestep_mean": 425.8125,
    "epsilon_norm": 7.458524227142334,
    "prediction_norm": 5.97392463684082,
    "timestamp": "2025-07-14T12:34:04.948520"
  },
  {
    "epoch": 885,
    "batch_idx": 128,
    "loss": 0.28927165269851685,
    "learning_rate": 0.0005898303742965966,
    "timestep_mean": 525.03125,
    "epsilon_norm": 7.896065711975098,
    "prediction_norm": 6.835989952087402,
    "timestamp": "2025-07-14T12:34:04.953645"
  },
  {
    "epoch": 885,
    "batch_idx": 160,
    "loss": 0.1778649240732193,
    "learning_rate": 0.0005898303742965966,
    "timestep_mean": 507.84375,
    "epsilon_norm": 6.945063591003418,
    "prediction_norm": 5.992074012756348,
    "timestamp": "2025-07-14T12:34:04.958832"
  },
  {
    "epoch": 885,
    "batch_idx": 192,
    "loss": 0.1329771876335144,
    "learning_rate": 0.0005898303742965966,
    "timestep_mean": 508.0,
    "epsilon_norm": 5.242624759674072,
    "prediction_norm": 4.549074649810791,
    "timestamp": "2025-07-14T12:34:04.964002"
  },
  {
    "epoch": 886,
    "batch_idx": 0,
    "loss": 0.30887046456336975,
    "learning_rate": 0.0005890576451321052,
    "timestep_mean": 448.5,
    "epsilon_norm": 8.275439262390137,
    "prediction_norm": 6.939919471740723,
    "timestamp": "2025-07-14T12:34:04.969030"
  },
  {
    "epoch": 886,
    "batch_idx": 32,
    "loss": 0.3875574469566345,
    "learning_rate": 0.0005890576451321052,
    "timestep_mean": 546.28125,
    "epsilon_norm": 8.045605659484863,
    "prediction_norm": 6.599250793457031,
    "timestamp": "2025-07-14T12:34:04.974249"
  },
  {
    "epoch": 886,
    "batch_idx": 64,
    "loss": 0.36101001501083374,
    "learning_rate": 0.0005890576451321052,
    "timestep_mean": 471.15625,
    "epsilon_norm": 8.620912551879883,
    "prediction_norm": 7.132010459899902,
    "timestamp": "2025-07-14T12:34:04.979380"
  },
  {
    "epoch": 886,
    "batch_idx": 96,
    "loss": 0.22158633172512054,
    "learning_rate": 0.0005890576451321052,
    "timestep_mean": 485.84375,
    "epsilon_norm": 7.337554931640625,
    "prediction_norm": 7.15015983581543,
    "timestamp": "2025-07-14T12:34:04.984547"
  },
  {
    "epoch": 886,
    "batch_idx": 128,
    "loss": 0.3687300384044647,
    "learning_rate": 0.0005890576451321052,
    "timestep_mean": 487.0,
    "epsilon_norm": 8.611424446105957,
    "prediction_norm": 6.658236980438232,
    "timestamp": "2025-07-14T12:34:04.989733"
  },
  {
    "epoch": 886,
    "batch_idx": 160,
    "loss": 0.2628432512283325,
    "learning_rate": 0.0005890576451321052,
    "timestep_mean": 571.75,
    "epsilon_norm": 8.36428165435791,
    "prediction_norm": 8.327762603759766,
    "timestamp": "2025-07-14T12:34:04.994984"
  },
  {
    "epoch": 886,
    "batch_idx": 192,
    "loss": 0.11184835433959961,
    "learning_rate": 0.0005890576451321052,
    "timestep_mean": 480.875,
    "epsilon_norm": 3.491736650466919,
    "prediction_norm": 2.865813732147217,
    "timestamp": "2025-07-14T12:34:05.000135"
  },
  {
    "epoch": 887,
    "batch_idx": 0,
    "loss": 0.5303994417190552,
    "learning_rate": 0.0005882846962267276,
    "timestep_mean": 555.5,
    "epsilon_norm": 9.136672019958496,
    "prediction_norm": 6.486862659454346,
    "timestamp": "2025-07-14T12:34:05.005173"
  },
  {
    "epoch": 887,
    "batch_idx": 32,
    "loss": 0.20706695318222046,
    "learning_rate": 0.0005882846962267276,
    "timestep_mean": 583.75,
    "epsilon_norm": 8.148114204406738,
    "prediction_norm": 7.323721408843994,
    "timestamp": "2025-07-14T12:34:05.010383"
  },
  {
    "epoch": 887,
    "batch_idx": 64,
    "loss": 0.21355406939983368,
    "learning_rate": 0.0005882846962267276,
    "timestep_mean": 552.875,
    "epsilon_norm": 7.781936168670654,
    "prediction_norm": 6.888298511505127,
    "timestamp": "2025-07-14T12:34:05.015597"
  },
  {
    "epoch": 887,
    "batch_idx": 96,
    "loss": 0.27261456847190857,
    "learning_rate": 0.0005882846962267276,
    "timestep_mean": 488.25,
    "epsilon_norm": 7.018293380737305,
    "prediction_norm": 6.773589134216309,
    "timestamp": "2025-07-14T12:34:05.020942"
  },
  {
    "epoch": 887,
    "batch_idx": 128,
    "loss": 0.21067266166210175,
    "learning_rate": 0.0005882846962267276,
    "timestep_mean": 498.875,
    "epsilon_norm": 8.162873268127441,
    "prediction_norm": 7.07875394821167,
    "timestamp": "2025-07-14T12:34:05.026146"
  },
  {
    "epoch": 887,
    "batch_idx": 160,
    "loss": 0.12700222432613373,
    "learning_rate": 0.0005882846962267276,
    "timestep_mean": 587.125,
    "epsilon_norm": 7.097792148590088,
    "prediction_norm": 6.320326328277588,
    "timestamp": "2025-07-14T12:34:05.031558"
  },
  {
    "epoch": 887,
    "batch_idx": 192,
    "loss": 0.6995911598205566,
    "learning_rate": 0.0005882846962267276,
    "timestep_mean": 542.625,
    "epsilon_norm": 4.392315864562988,
    "prediction_norm": 2.7280595302581787,
    "timestamp": "2025-07-14T12:34:05.037161"
  },
  {
    "epoch": 888,
    "batch_idx": 0,
    "loss": 0.39833226799964905,
    "learning_rate": 0.0005875115294876383,
    "timestep_mean": 425.59375,
    "epsilon_norm": 7.706345558166504,
    "prediction_norm": 5.226324558258057,
    "timestamp": "2025-07-14T12:34:05.042639"
  },
  {
    "epoch": 888,
    "batch_idx": 32,
    "loss": 0.354148805141449,
    "learning_rate": 0.0005875115294876383,
    "timestep_mean": 565.5,
    "epsilon_norm": 9.081560134887695,
    "prediction_norm": 7.448801040649414,
    "timestamp": "2025-07-14T12:34:05.048213"
  },
  {
    "epoch": 888,
    "batch_idx": 64,
    "loss": 0.4315609931945801,
    "learning_rate": 0.0005875115294876383,
    "timestep_mean": 468.25,
    "epsilon_norm": 8.816122055053711,
    "prediction_norm": 7.356345176696777,
    "timestamp": "2025-07-14T12:34:05.053392"
  },
  {
    "epoch": 888,
    "batch_idx": 96,
    "loss": 0.24637183547019958,
    "learning_rate": 0.0005875115294876383,
    "timestep_mean": 542.78125,
    "epsilon_norm": 7.525887489318848,
    "prediction_norm": 6.073697566986084,
    "timestamp": "2025-07-14T12:34:05.058670"
  },
  {
    "epoch": 888,
    "batch_idx": 128,
    "loss": 0.22287069261074066,
    "learning_rate": 0.0005875115294876383,
    "timestep_mean": 506.1875,
    "epsilon_norm": 7.021078109741211,
    "prediction_norm": 5.608766555786133,
    "timestamp": "2025-07-14T12:34:05.063783"
  },
  {
    "epoch": 888,
    "batch_idx": 160,
    "loss": 0.35279321670532227,
    "learning_rate": 0.0005875115294876383,
    "timestep_mean": 531.78125,
    "epsilon_norm": 7.534111976623535,
    "prediction_norm": 5.610611915588379,
    "timestamp": "2025-07-14T12:34:05.068827"
  },
  {
    "epoch": 888,
    "batch_idx": 192,
    "loss": 0.23090983927249908,
    "learning_rate": 0.0005875115294876383,
    "timestep_mean": 609.25,
    "epsilon_norm": 5.233855247497559,
    "prediction_norm": 4.434727668762207,
    "timestamp": "2025-07-14T12:34:05.074069"
  },
  {
    "epoch": 889,
    "batch_idx": 0,
    "loss": 0.2331991195678711,
    "learning_rate": 0.0005867381468225491,
    "timestep_mean": 546.4375,
    "epsilon_norm": 6.449524879455566,
    "prediction_norm": 4.824872016906738,
    "timestamp": "2025-07-14T12:34:05.079361"
  },
  {
    "epoch": 889,
    "batch_idx": 32,
    "loss": 0.27446112036705017,
    "learning_rate": 0.0005867381468225491,
    "timestep_mean": 478.4375,
    "epsilon_norm": 7.884093761444092,
    "prediction_norm": 5.9418535232543945,
    "timestamp": "2025-07-14T12:34:05.084378"
  },
  {
    "epoch": 889,
    "batch_idx": 64,
    "loss": 0.2457299530506134,
    "learning_rate": 0.0005867381468225491,
    "timestep_mean": 450.84375,
    "epsilon_norm": 6.2534356117248535,
    "prediction_norm": 4.390660762786865,
    "timestamp": "2025-07-14T12:34:05.089466"
  },
  {
    "epoch": 889,
    "batch_idx": 96,
    "loss": 0.31923019886016846,
    "learning_rate": 0.0005867381468225491,
    "timestep_mean": 508.71875,
    "epsilon_norm": 7.6210174560546875,
    "prediction_norm": 6.107636451721191,
    "timestamp": "2025-07-14T12:34:05.094582"
  },
  {
    "epoch": 889,
    "batch_idx": 128,
    "loss": 0.143783837556839,
    "learning_rate": 0.0005867381468225491,
    "timestep_mean": 550.46875,
    "epsilon_norm": 7.362100601196289,
    "prediction_norm": 6.356136322021484,
    "timestamp": "2025-07-14T12:34:05.105482"
  },
  {
    "epoch": 889,
    "batch_idx": 160,
    "loss": 0.5075531005859375,
    "learning_rate": 0.0005867381468225491,
    "timestep_mean": 413.46875,
    "epsilon_norm": 8.887984275817871,
    "prediction_norm": 6.917126178741455,
    "timestamp": "2025-07-14T12:34:05.110967"
  },
  {
    "epoch": 889,
    "batch_idx": 192,
    "loss": 0.2587636709213257,
    "learning_rate": 0.0005867381468225491,
    "timestep_mean": 434.75,
    "epsilon_norm": 4.209715843200684,
    "prediction_norm": 3.8824007511138916,
    "timestamp": "2025-07-14T12:34:05.117906"
  },
  {
    "epoch": 890,
    "batch_idx": 0,
    "loss": 0.31448614597320557,
    "learning_rate": 0.0005859645501397051,
    "timestep_mean": 537.0,
    "epsilon_norm": 7.780911922454834,
    "prediction_norm": 5.96928596496582,
    "timestamp": "2025-07-14T12:34:05.123541"
  },
  {
    "epoch": 890,
    "batch_idx": 32,
    "loss": 0.1583288162946701,
    "learning_rate": 0.0005859645501397051,
    "timestep_mean": 575.0625,
    "epsilon_norm": 7.6689066886901855,
    "prediction_norm": 7.600767135620117,
    "timestamp": "2025-07-14T12:34:05.128665"
  },
  {
    "epoch": 890,
    "batch_idx": 64,
    "loss": 0.2934854030609131,
    "learning_rate": 0.0005859645501397051,
    "timestep_mean": 556.0625,
    "epsilon_norm": 8.889450073242188,
    "prediction_norm": 7.41943883895874,
    "timestamp": "2025-07-14T12:34:05.134005"
  },
  {
    "epoch": 890,
    "batch_idx": 96,
    "loss": 0.1597100794315338,
    "learning_rate": 0.0005859645501397051,
    "timestep_mean": 562.625,
    "epsilon_norm": 6.810886383056641,
    "prediction_norm": 6.099906921386719,
    "timestamp": "2025-07-14T12:34:05.139471"
  },
  {
    "epoch": 890,
    "batch_idx": 128,
    "loss": 0.4225274920463562,
    "learning_rate": 0.0005859645501397051,
    "timestep_mean": 491.46875,
    "epsilon_norm": 7.867416858673096,
    "prediction_norm": 6.303689479827881,
    "timestamp": "2025-07-14T12:34:05.145169"
  },
  {
    "epoch": 890,
    "batch_idx": 160,
    "loss": 0.3024173676967621,
    "learning_rate": 0.0005859645501397051,
    "timestep_mean": 495.125,
    "epsilon_norm": 8.684216499328613,
    "prediction_norm": 7.368504524230957,
    "timestamp": "2025-07-14T12:34:05.150360"
  },
  {
    "epoch": 890,
    "batch_idx": 192,
    "loss": 0.14442838728427887,
    "learning_rate": 0.0005859645501397051,
    "timestep_mean": 473.125,
    "epsilon_norm": 3.7456982135772705,
    "prediction_norm": 3.3176262378692627,
    "timestamp": "2025-07-14T12:34:05.155618"
  },
  {
    "epoch": 891,
    "batch_idx": 0,
    "loss": 0.23659035563468933,
    "learning_rate": 0.000585190741347879,
    "timestep_mean": 426.65625,
    "epsilon_norm": 7.040618419647217,
    "prediction_norm": 6.292492866516113,
    "timestamp": "2025-07-14T12:34:05.160799"
  },
  {
    "epoch": 891,
    "batch_idx": 32,
    "loss": 0.1589544266462326,
    "learning_rate": 0.000585190741347879,
    "timestep_mean": 470.90625,
    "epsilon_norm": 7.231451511383057,
    "prediction_norm": 6.958710670471191,
    "timestamp": "2025-07-14T12:34:05.166052"
  },
  {
    "epoch": 891,
    "batch_idx": 64,
    "loss": 0.23015423119068146,
    "learning_rate": 0.000585190741347879,
    "timestep_mean": 523.75,
    "epsilon_norm": 8.546052932739258,
    "prediction_norm": 6.736761093139648,
    "timestamp": "2025-07-14T12:34:05.171341"
  },
  {
    "epoch": 891,
    "batch_idx": 96,
    "loss": 0.3940798044204712,
    "learning_rate": 0.000585190741347879,
    "timestep_mean": 500.25,
    "epsilon_norm": 8.56589412689209,
    "prediction_norm": 6.770724773406982,
    "timestamp": "2025-07-14T12:34:05.176742"
  },
  {
    "epoch": 891,
    "batch_idx": 128,
    "loss": 0.16252973675727844,
    "learning_rate": 0.000585190741347879,
    "timestep_mean": 498.90625,
    "epsilon_norm": 7.109250545501709,
    "prediction_norm": 7.0282063484191895,
    "timestamp": "2025-07-14T12:34:05.181968"
  },
  {
    "epoch": 891,
    "batch_idx": 160,
    "loss": 0.2084863781929016,
    "learning_rate": 0.000585190741347879,
    "timestep_mean": 465.71875,
    "epsilon_norm": 7.662379264831543,
    "prediction_norm": 6.4958953857421875,
    "timestamp": "2025-07-14T12:34:05.187149"
  },
  {
    "epoch": 891,
    "batch_idx": 192,
    "loss": 0.29018542170524597,
    "learning_rate": 0.000585190741347879,
    "timestep_mean": 426.5,
    "epsilon_norm": 3.611952066421509,
    "prediction_norm": 2.902029514312744,
    "timestamp": "2025-07-14T12:34:05.192486"
  },
  {
    "epoch": 892,
    "batch_idx": 0,
    "loss": 0.43437090516090393,
    "learning_rate": 0.0005844167223563671,
    "timestep_mean": 420.96875,
    "epsilon_norm": 8.72285270690918,
    "prediction_norm": 7.612954139709473,
    "timestamp": "2025-07-14T12:34:05.197642"
  },
  {
    "epoch": 892,
    "batch_idx": 32,
    "loss": 0.25077706575393677,
    "learning_rate": 0.0005844167223563671,
    "timestep_mean": 502.5625,
    "epsilon_norm": 7.782905578613281,
    "prediction_norm": 7.307784080505371,
    "timestamp": "2025-07-14T12:34:05.202636"
  },
  {
    "epoch": 892,
    "batch_idx": 64,
    "loss": 0.2629462480545044,
    "learning_rate": 0.0005844167223563671,
    "timestep_mean": 561.03125,
    "epsilon_norm": 8.254851341247559,
    "prediction_norm": 7.697389125823975,
    "timestamp": "2025-07-14T12:34:05.207533"
  },
  {
    "epoch": 892,
    "batch_idx": 96,
    "loss": 0.3146703243255615,
    "learning_rate": 0.0005844167223563671,
    "timestep_mean": 502.625,
    "epsilon_norm": 7.825161933898926,
    "prediction_norm": 7.541354656219482,
    "timestamp": "2025-07-14T12:34:05.212404"
  },
  {
    "epoch": 892,
    "batch_idx": 128,
    "loss": 0.3504866063594818,
    "learning_rate": 0.0005844167223563671,
    "timestep_mean": 510.0625,
    "epsilon_norm": 9.75282096862793,
    "prediction_norm": 7.825176239013672,
    "timestamp": "2025-07-14T12:34:05.217290"
  },
  {
    "epoch": 892,
    "batch_idx": 160,
    "loss": 0.24150754511356354,
    "learning_rate": 0.0005844167223563671,
    "timestep_mean": 538.0,
    "epsilon_norm": 8.51364517211914,
    "prediction_norm": 7.7921929359436035,
    "timestamp": "2025-07-14T12:34:05.222326"
  },
  {
    "epoch": 892,
    "batch_idx": 192,
    "loss": 0.28498831391334534,
    "learning_rate": 0.0005844167223563671,
    "timestep_mean": 441.75,
    "epsilon_norm": 3.89152193069458,
    "prediction_norm": 3.1244070529937744,
    "timestamp": "2025-07-14T12:34:05.227378"
  },
  {
    "epoch": 893,
    "batch_idx": 0,
    "loss": 0.32751819491386414,
    "learning_rate": 0.0005836424950749846,
    "timestep_mean": 410.84375,
    "epsilon_norm": 9.260709762573242,
    "prediction_norm": 7.364098072052002,
    "timestamp": "2025-07-14T12:34:05.232314"
  },
  {
    "epoch": 893,
    "batch_idx": 32,
    "loss": 0.21293838322162628,
    "learning_rate": 0.0005836424950749846,
    "timestep_mean": 588.5,
    "epsilon_norm": 7.748279094696045,
    "prediction_norm": 6.41322135925293,
    "timestamp": "2025-07-14T12:34:05.237259"
  },
  {
    "epoch": 893,
    "batch_idx": 64,
    "loss": 0.20440053939819336,
    "learning_rate": 0.0005836424950749846,
    "timestep_mean": 500.75,
    "epsilon_norm": 7.139408111572266,
    "prediction_norm": 6.495112895965576,
    "timestamp": "2025-07-14T12:34:05.242272"
  },
  {
    "epoch": 893,
    "batch_idx": 96,
    "loss": 0.22918462753295898,
    "learning_rate": 0.0005836424950749846,
    "timestep_mean": 687.4375,
    "epsilon_norm": 9.264312744140625,
    "prediction_norm": 8.33031177520752,
    "timestamp": "2025-07-14T12:34:05.247276"
  },
  {
    "epoch": 893,
    "batch_idx": 128,
    "loss": 0.2867598533630371,
    "learning_rate": 0.0005836424950749846,
    "timestep_mean": 566.15625,
    "epsilon_norm": 8.210738182067871,
    "prediction_norm": 6.919022560119629,
    "timestamp": "2025-07-14T12:34:05.252161"
  },
  {
    "epoch": 893,
    "batch_idx": 160,
    "loss": 0.3182740807533264,
    "learning_rate": 0.0005836424950749846,
    "timestep_mean": 460.75,
    "epsilon_norm": 8.109736442565918,
    "prediction_norm": 6.212646484375,
    "timestamp": "2025-07-14T12:34:05.257125"
  },
  {
    "epoch": 893,
    "batch_idx": 192,
    "loss": 0.1032647043466568,
    "learning_rate": 0.0005836424950749846,
    "timestep_mean": 538.5,
    "epsilon_norm": 3.4290242195129395,
    "prediction_norm": 3.1804351806640625,
    "timestamp": "2025-07-14T12:34:05.262110"
  },
  {
    "epoch": 894,
    "batch_idx": 0,
    "loss": 0.35233795642852783,
    "learning_rate": 0.0005828680614140601,
    "timestep_mean": 486.75,
    "epsilon_norm": 7.645921230316162,
    "prediction_norm": 6.2217278480529785,
    "timestamp": "2025-07-14T12:34:05.267031"
  },
  {
    "epoch": 894,
    "batch_idx": 32,
    "loss": 0.20508699119091034,
    "learning_rate": 0.0005828680614140601,
    "timestep_mean": 503.59375,
    "epsilon_norm": 8.031801223754883,
    "prediction_norm": 7.3422770500183105,
    "timestamp": "2025-07-14T12:34:05.271866"
  },
  {
    "epoch": 894,
    "batch_idx": 64,
    "loss": 0.36138367652893066,
    "learning_rate": 0.0005828680614140601,
    "timestep_mean": 481.75,
    "epsilon_norm": 8.113863945007324,
    "prediction_norm": 6.599432945251465,
    "timestamp": "2025-07-14T12:34:05.276809"
  },
  {
    "epoch": 894,
    "batch_idx": 96,
    "loss": 0.2973199188709259,
    "learning_rate": 0.0005828680614140601,
    "timestep_mean": 472.8125,
    "epsilon_norm": 8.972424507141113,
    "prediction_norm": 7.53361701965332,
    "timestamp": "2025-07-14T12:34:05.281698"
  },
  {
    "epoch": 894,
    "batch_idx": 128,
    "loss": 0.29864591360092163,
    "learning_rate": 0.0005828680614140601,
    "timestep_mean": 504.78125,
    "epsilon_norm": 8.057175636291504,
    "prediction_norm": 7.280864715576172,
    "timestamp": "2025-07-14T12:34:05.286546"
  },
  {
    "epoch": 894,
    "batch_idx": 160,
    "loss": 0.2815012037754059,
    "learning_rate": 0.0005828680614140601,
    "timestep_mean": 511.90625,
    "epsilon_norm": 7.697985649108887,
    "prediction_norm": 6.824623107910156,
    "timestamp": "2025-07-14T12:34:05.291502"
  },
  {
    "epoch": 894,
    "batch_idx": 192,
    "loss": 0.3534512221813202,
    "learning_rate": 0.0005828680614140601,
    "timestep_mean": 616.625,
    "epsilon_norm": 4.64102029800415,
    "prediction_norm": 4.146145820617676,
    "timestamp": "2025-07-14T12:34:05.296584"
  },
  {
    "epoch": 895,
    "batch_idx": 0,
    "loss": 0.26584935188293457,
    "learning_rate": 0.0005820934232844317,
    "timestep_mean": 565.5625,
    "epsilon_norm": 8.842329025268555,
    "prediction_norm": 7.455718994140625,
    "timestamp": "2025-07-14T12:34:05.301533"
  },
  {
    "epoch": 895,
    "batch_idx": 32,
    "loss": 0.32216325402259827,
    "learning_rate": 0.0005820934232844317,
    "timestep_mean": 455.5625,
    "epsilon_norm": 8.260780334472656,
    "prediction_norm": 7.29219913482666,
    "timestamp": "2025-07-14T12:34:05.306450"
  },
  {
    "epoch": 895,
    "batch_idx": 64,
    "loss": 0.2478577345609665,
    "learning_rate": 0.0005820934232844317,
    "timestep_mean": 546.5,
    "epsilon_norm": 8.0167818069458,
    "prediction_norm": 7.17781400680542,
    "timestamp": "2025-07-14T12:34:05.311375"
  },
  {
    "epoch": 895,
    "batch_idx": 96,
    "loss": 0.3021039366722107,
    "learning_rate": 0.0005820934232844317,
    "timestep_mean": 458.96875,
    "epsilon_norm": 7.885797500610352,
    "prediction_norm": 6.024424076080322,
    "timestamp": "2025-07-14T12:34:05.316360"
  },
  {
    "epoch": 895,
    "batch_idx": 128,
    "loss": 0.3542068600654602,
    "learning_rate": 0.0005820934232844317,
    "timestep_mean": 410.28125,
    "epsilon_norm": 9.036458015441895,
    "prediction_norm": 7.551814556121826,
    "timestamp": "2025-07-14T12:34:05.321341"
  },
  {
    "epoch": 895,
    "batch_idx": 160,
    "loss": 0.17225243151187897,
    "learning_rate": 0.0005820934232844317,
    "timestep_mean": 509.28125,
    "epsilon_norm": 6.892306327819824,
    "prediction_norm": 6.729895114898682,
    "timestamp": "2025-07-14T12:34:05.326308"
  },
  {
    "epoch": 895,
    "batch_idx": 192,
    "loss": 0.6454496383666992,
    "learning_rate": 0.0005820934232844317,
    "timestep_mean": 475.0,
    "epsilon_norm": 3.9936649799346924,
    "prediction_norm": 3.959282636642456,
    "timestamp": "2025-07-14T12:34:05.331253"
  },
  {
    "epoch": 896,
    "batch_idx": 0,
    "loss": 0.5894690752029419,
    "learning_rate": 0.0005813185825974421,
    "timestep_mean": 375.40625,
    "epsilon_norm": 8.437535285949707,
    "prediction_norm": 5.5627899169921875,
    "timestamp": "2025-07-14T12:34:05.336206"
  },
  {
    "epoch": 896,
    "batch_idx": 32,
    "loss": 0.10044722259044647,
    "learning_rate": 0.0005813185825974421,
    "timestep_mean": 549.3125,
    "epsilon_norm": 7.686664581298828,
    "prediction_norm": 6.650269508361816,
    "timestamp": "2025-07-14T12:34:05.341113"
  },
  {
    "epoch": 896,
    "batch_idx": 64,
    "loss": 0.302858829498291,
    "learning_rate": 0.0005813185825974421,
    "timestep_mean": 463.15625,
    "epsilon_norm": 6.913346767425537,
    "prediction_norm": 6.160654067993164,
    "timestamp": "2025-07-14T12:34:05.346130"
  },
  {
    "epoch": 896,
    "batch_idx": 96,
    "loss": 0.36458641290664673,
    "learning_rate": 0.0005813185825974421,
    "timestep_mean": 510.15625,
    "epsilon_norm": 8.350432395935059,
    "prediction_norm": 6.544620990753174,
    "timestamp": "2025-07-14T12:34:05.350988"
  },
  {
    "epoch": 896,
    "batch_idx": 128,
    "loss": 0.29496973752975464,
    "learning_rate": 0.0005813185825974421,
    "timestep_mean": 495.59375,
    "epsilon_norm": 8.382867813110352,
    "prediction_norm": 6.126250267028809,
    "timestamp": "2025-07-14T12:34:05.355865"
  },
  {
    "epoch": 896,
    "batch_idx": 160,
    "loss": 0.37059861421585083,
    "learning_rate": 0.0005813185825974421,
    "timestep_mean": 586.5,
    "epsilon_norm": 8.270523071289062,
    "prediction_norm": 6.572021484375,
    "timestamp": "2025-07-14T12:34:05.360767"
  },
  {
    "epoch": 896,
    "batch_idx": 192,
    "loss": 0.07080133259296417,
    "learning_rate": 0.0005813185825974421,
    "timestep_mean": 590.75,
    "epsilon_norm": 3.4231865406036377,
    "prediction_norm": 2.939075231552124,
    "timestamp": "2025-07-14T12:34:05.365804"
  },
  {
    "epoch": 897,
    "batch_idx": 0,
    "loss": 0.3988260328769684,
    "learning_rate": 0.0005805435412649337,
    "timestep_mean": 509.5625,
    "epsilon_norm": 9.546026229858398,
    "prediction_norm": 7.35116720199585,
    "timestamp": "2025-07-14T12:34:05.370810"
  },
  {
    "epoch": 897,
    "batch_idx": 32,
    "loss": 0.22740408778190613,
    "learning_rate": 0.0005805435412649337,
    "timestep_mean": 498.375,
    "epsilon_norm": 8.776392936706543,
    "prediction_norm": 7.792647361755371,
    "timestamp": "2025-07-14T12:34:05.375720"
  },
  {
    "epoch": 897,
    "batch_idx": 64,
    "loss": 0.23920466005802155,
    "learning_rate": 0.0005805435412649337,
    "timestep_mean": 485.34375,
    "epsilon_norm": 7.8200883865356445,
    "prediction_norm": 6.454146862030029,
    "timestamp": "2025-07-14T12:34:05.380607"
  },
  {
    "epoch": 897,
    "batch_idx": 96,
    "loss": 0.3479939103126526,
    "learning_rate": 0.0005805435412649337,
    "timestep_mean": 409.65625,
    "epsilon_norm": 8.265732765197754,
    "prediction_norm": 7.055017471313477,
    "timestamp": "2025-07-14T12:34:05.385602"
  },
  {
    "epoch": 897,
    "batch_idx": 128,
    "loss": 0.23586398363113403,
    "learning_rate": 0.0005805435412649337,
    "timestep_mean": 386.96875,
    "epsilon_norm": 6.889242649078369,
    "prediction_norm": 5.90189790725708,
    "timestamp": "2025-07-14T12:34:05.390429"
  },
  {
    "epoch": 897,
    "batch_idx": 160,
    "loss": 0.5987491607666016,
    "learning_rate": 0.0005805435412649337,
    "timestep_mean": 426.46875,
    "epsilon_norm": 8.703533172607422,
    "prediction_norm": 6.03316068649292,
    "timestamp": "2025-07-14T12:34:05.395421"
  },
  {
    "epoch": 897,
    "batch_idx": 192,
    "loss": 0.6463764905929565,
    "learning_rate": 0.0005805435412649337,
    "timestep_mean": 404.875,
    "epsilon_norm": 4.436162948608398,
    "prediction_norm": 3.175565242767334,
    "timestamp": "2025-07-14T12:34:05.400434"
  },
  {
    "epoch": 898,
    "batch_idx": 0,
    "loss": 0.20995797216892242,
    "learning_rate": 0.0005797683011992434,
    "timestep_mean": 542.1875,
    "epsilon_norm": 7.588816165924072,
    "prediction_norm": 7.360837459564209,
    "timestamp": "2025-07-14T12:34:05.405379"
  },
  {
    "epoch": 898,
    "batch_idx": 32,
    "loss": 0.22894971072673798,
    "learning_rate": 0.0005797683011992434,
    "timestep_mean": 537.1875,
    "epsilon_norm": 7.357650279998779,
    "prediction_norm": 6.819366455078125,
    "timestamp": "2025-07-14T12:34:05.410260"
  },
  {
    "epoch": 898,
    "batch_idx": 64,
    "loss": 0.18984144926071167,
    "learning_rate": 0.0005797683011992434,
    "timestep_mean": 500.28125,
    "epsilon_norm": 6.603163719177246,
    "prediction_norm": 5.670839309692383,
    "timestamp": "2025-07-14T12:34:05.415218"
  },
  {
    "epoch": 898,
    "batch_idx": 96,
    "loss": 0.33690184354782104,
    "learning_rate": 0.0005797683011992434,
    "timestep_mean": 491.28125,
    "epsilon_norm": 8.335911750793457,
    "prediction_norm": 7.244225978851318,
    "timestamp": "2025-07-14T12:34:05.420340"
  },
  {
    "epoch": 898,
    "batch_idx": 128,
    "loss": 0.19996586441993713,
    "learning_rate": 0.0005797683011992434,
    "timestep_mean": 566.65625,
    "epsilon_norm": 7.84872579574585,
    "prediction_norm": 7.201948642730713,
    "timestamp": "2025-07-14T12:34:05.425437"
  },
  {
    "epoch": 898,
    "batch_idx": 160,
    "loss": 0.3199332356452942,
    "learning_rate": 0.0005797683011992434,
    "timestep_mean": 482.25,
    "epsilon_norm": 8.528472900390625,
    "prediction_norm": 6.885989189147949,
    "timestamp": "2025-07-14T12:34:05.430393"
  },
  {
    "epoch": 898,
    "batch_idx": 192,
    "loss": 0.2943059802055359,
    "learning_rate": 0.0005797683011992434,
    "timestep_mean": 429.25,
    "epsilon_norm": 4.709380626678467,
    "prediction_norm": 3.827939033508301,
    "timestamp": "2025-07-14T12:34:05.435641"
  },
  {
    "epoch": 899,
    "batch_idx": 0,
    "loss": 0.3585413098335266,
    "learning_rate": 0.0005789928643131999,
    "timestep_mean": 592.3125,
    "epsilon_norm": 9.383967399597168,
    "prediction_norm": 7.936444282531738,
    "timestamp": "2025-07-14T12:34:05.440997"
  },
  {
    "epoch": 899,
    "batch_idx": 32,
    "loss": 0.3404414653778076,
    "learning_rate": 0.0005789928643131999,
    "timestep_mean": 513.28125,
    "epsilon_norm": 9.171537399291992,
    "prediction_norm": 7.769301891326904,
    "timestamp": "2025-07-14T12:34:05.446252"
  },
  {
    "epoch": 899,
    "batch_idx": 64,
    "loss": 0.41923636198043823,
    "learning_rate": 0.0005789928643131999,
    "timestep_mean": 450.90625,
    "epsilon_norm": 7.720014572143555,
    "prediction_norm": 6.841865062713623,
    "timestamp": "2025-07-14T12:34:05.451460"
  },
  {
    "epoch": 899,
    "batch_idx": 96,
    "loss": 0.3196158707141876,
    "learning_rate": 0.0005789928643131999,
    "timestep_mean": 527.03125,
    "epsilon_norm": 7.890378475189209,
    "prediction_norm": 6.856913089752197,
    "timestamp": "2025-07-14T12:34:05.456701"
  },
  {
    "epoch": 899,
    "batch_idx": 128,
    "loss": 0.29229068756103516,
    "learning_rate": 0.0005789928643131999,
    "timestep_mean": 508.3125,
    "epsilon_norm": 7.249527931213379,
    "prediction_norm": 5.945484638214111,
    "timestamp": "2025-07-14T12:34:05.461950"
  },
  {
    "epoch": 899,
    "batch_idx": 160,
    "loss": 0.2260103076696396,
    "learning_rate": 0.0005789928643131999,
    "timestep_mean": 521.0,
    "epsilon_norm": 7.999505519866943,
    "prediction_norm": 6.642727375030518,
    "timestamp": "2025-07-14T12:34:05.467198"
  },
  {
    "epoch": 899,
    "batch_idx": 192,
    "loss": 0.5584453344345093,
    "learning_rate": 0.0005789928643131999,
    "timestep_mean": 411.125,
    "epsilon_norm": 5.2543559074401855,
    "prediction_norm": 4.40954065322876,
    "timestamp": "2025-07-14T12:34:05.472345"
  },
  {
    "epoch": 900,
    "batch_idx": 0,
    "loss": 0.2204456776380539,
    "learning_rate": 0.0005782172325201159,
    "timestep_mean": 403.0,
    "epsilon_norm": 8.689565658569336,
    "prediction_norm": 7.646271705627441,
    "timestamp": "2025-07-14T12:34:05.477714"
  },
  {
    "epoch": 900,
    "batch_idx": 32,
    "loss": 0.18465448915958405,
    "learning_rate": 0.0005782172325201159,
    "timestep_mean": 521.59375,
    "epsilon_norm": 7.80086088180542,
    "prediction_norm": 7.00552225112915,
    "timestamp": "2025-07-14T12:34:05.483003"
  },
  {
    "epoch": 900,
    "batch_idx": 64,
    "loss": 0.26221585273742676,
    "learning_rate": 0.0005782172325201159,
    "timestep_mean": 496.625,
    "epsilon_norm": 9.161813735961914,
    "prediction_norm": 8.066452026367188,
    "timestamp": "2025-07-14T12:34:05.488214"
  },
  {
    "epoch": 900,
    "batch_idx": 96,
    "loss": 0.1857919692993164,
    "learning_rate": 0.0005782172325201159,
    "timestep_mean": 572.3125,
    "epsilon_norm": 7.89175271987915,
    "prediction_norm": 6.872438430786133,
    "timestamp": "2025-07-14T12:34:05.493349"
  },
  {
    "epoch": 900,
    "batch_idx": 128,
    "loss": 0.1979372501373291,
    "learning_rate": 0.0005782172325201159,
    "timestep_mean": 569.875,
    "epsilon_norm": 6.888350486755371,
    "prediction_norm": 5.924954414367676,
    "timestamp": "2025-07-14T12:34:05.498428"
  },
  {
    "epoch": 900,
    "batch_idx": 160,
    "loss": 0.1715373694896698,
    "learning_rate": 0.0005782172325201159,
    "timestep_mean": 569.46875,
    "epsilon_norm": 7.998134613037109,
    "prediction_norm": 7.3610358238220215,
    "timestamp": "2025-07-14T12:34:05.503466"
  },
  {
    "epoch": 900,
    "batch_idx": 192,
    "loss": 0.6232975721359253,
    "learning_rate": 0.0005782172325201159,
    "timestep_mean": 486.0,
    "epsilon_norm": 4.569525718688965,
    "prediction_norm": 3.7366182804107666,
    "timestamp": "2025-07-14T12:34:05.508552"
  },
  {
    "epoch": 901,
    "batch_idx": 0,
    "loss": 0.3634759783744812,
    "learning_rate": 0.000577441407733786,
    "timestep_mean": 482.15625,
    "epsilon_norm": 8.508953094482422,
    "prediction_norm": 6.8365607261657715,
    "timestamp": "2025-07-14T12:34:05.513821"
  },
  {
    "epoch": 901,
    "batch_idx": 32,
    "loss": 0.16921855509281158,
    "learning_rate": 0.000577441407733786,
    "timestep_mean": 413.96875,
    "epsilon_norm": 8.313529014587402,
    "prediction_norm": 7.407464504241943,
    "timestamp": "2025-07-14T12:34:05.518949"
  },
  {
    "epoch": 901,
    "batch_idx": 64,
    "loss": 0.21886229515075684,
    "learning_rate": 0.000577441407733786,
    "timestep_mean": 487.4375,
    "epsilon_norm": 7.493873596191406,
    "prediction_norm": 6.510461330413818,
    "timestamp": "2025-07-14T12:34:05.524139"
  },
  {
    "epoch": 901,
    "batch_idx": 96,
    "loss": 0.3128775656223297,
    "learning_rate": 0.000577441407733786,
    "timestep_mean": 527.0,
    "epsilon_norm": 7.769662857055664,
    "prediction_norm": 6.1989006996154785,
    "timestamp": "2025-07-14T12:34:05.529243"
  },
  {
    "epoch": 901,
    "batch_idx": 128,
    "loss": 0.22707322239875793,
    "learning_rate": 0.000577441407733786,
    "timestep_mean": 561.6875,
    "epsilon_norm": 8.119757652282715,
    "prediction_norm": 7.422922611236572,
    "timestamp": "2025-07-14T12:34:05.534399"
  },
  {
    "epoch": 901,
    "batch_idx": 160,
    "loss": 0.2965790033340454,
    "learning_rate": 0.000577441407733786,
    "timestep_mean": 485.4375,
    "epsilon_norm": 7.312162399291992,
    "prediction_norm": 5.669971942901611,
    "timestamp": "2025-07-14T12:34:05.539613"
  },
  {
    "epoch": 901,
    "batch_idx": 192,
    "loss": 0.12513962388038635,
    "learning_rate": 0.000577441407733786,
    "timestep_mean": 502.625,
    "epsilon_norm": 4.512378692626953,
    "prediction_norm": 4.393956184387207,
    "timestamp": "2025-07-14T12:34:05.544572"
  },
  {
    "epoch": 902,
    "batch_idx": 0,
    "loss": 0.19998528063297272,
    "learning_rate": 0.0005766653918684808,
    "timestep_mean": 526.5625,
    "epsilon_norm": 7.455514430999756,
    "prediction_norm": 6.211235523223877,
    "timestamp": "2025-07-14T12:34:05.552148"
  },
  {
    "epoch": 902,
    "batch_idx": 32,
    "loss": 0.4848693907260895,
    "learning_rate": 0.0005766653918684808,
    "timestep_mean": 431.9375,
    "epsilon_norm": 9.248067855834961,
    "prediction_norm": 7.119600296020508,
    "timestamp": "2025-07-14T12:34:05.557330"
  },
  {
    "epoch": 902,
    "batch_idx": 64,
    "loss": 0.4036484360694885,
    "learning_rate": 0.0005766653918684808,
    "timestep_mean": 506.25,
    "epsilon_norm": 8.301523208618164,
    "prediction_norm": 6.410300254821777,
    "timestamp": "2025-07-14T12:34:05.562480"
  },
  {
    "epoch": 902,
    "batch_idx": 96,
    "loss": 0.35207027196884155,
    "learning_rate": 0.0005766653918684808,
    "timestep_mean": 445.21875,
    "epsilon_norm": 7.731992721557617,
    "prediction_norm": 6.008092403411865,
    "timestamp": "2025-07-14T12:34:05.567570"
  },
  {
    "epoch": 902,
    "batch_idx": 128,
    "loss": 0.5112016201019287,
    "learning_rate": 0.0005766653918684808,
    "timestep_mean": 532.21875,
    "epsilon_norm": 8.57797622680664,
    "prediction_norm": 6.438175678253174,
    "timestamp": "2025-07-14T12:34:05.572792"
  },
  {
    "epoch": 902,
    "batch_idx": 160,
    "loss": 0.18186408281326294,
    "learning_rate": 0.0005766653918684808,
    "timestep_mean": 549.75,
    "epsilon_norm": 8.163498878479004,
    "prediction_norm": 7.228975296020508,
    "timestamp": "2025-07-14T12:34:05.577911"
  },
  {
    "epoch": 902,
    "batch_idx": 192,
    "loss": 0.5133023262023926,
    "learning_rate": 0.0005766653918684808,
    "timestep_mean": 413.625,
    "epsilon_norm": 3.299837589263916,
    "prediction_norm": 2.1605474948883057,
    "timestamp": "2025-07-14T12:34:05.583062"
  },
  {
    "epoch": 903,
    "batch_idx": 0,
    "loss": 0.20384275913238525,
    "learning_rate": 0.0005758891868389422,
    "timestep_mean": 509.875,
    "epsilon_norm": 8.081986427307129,
    "prediction_norm": 6.933840274810791,
    "timestamp": "2025-07-14T12:34:05.588240"
  },
  {
    "epoch": 903,
    "batch_idx": 32,
    "loss": 0.6265738010406494,
    "learning_rate": 0.0005758891868389422,
    "timestep_mean": 470.84375,
    "epsilon_norm": 7.796024799346924,
    "prediction_norm": 5.730630397796631,
    "timestamp": "2025-07-14T12:34:05.593139"
  },
  {
    "epoch": 903,
    "batch_idx": 64,
    "loss": 0.2866322994232178,
    "learning_rate": 0.0005758891868389422,
    "timestep_mean": 520.0,
    "epsilon_norm": 8.000539779663086,
    "prediction_norm": 6.618401050567627,
    "timestamp": "2025-07-14T12:34:05.598373"
  },
  {
    "epoch": 903,
    "batch_idx": 96,
    "loss": 0.2848799228668213,
    "learning_rate": 0.0005758891868389422,
    "timestep_mean": 491.6875,
    "epsilon_norm": 8.055574417114258,
    "prediction_norm": 6.866065979003906,
    "timestamp": "2025-07-14T12:34:05.603515"
  },
  {
    "epoch": 903,
    "batch_idx": 128,
    "loss": 0.34933456778526306,
    "learning_rate": 0.0005758891868389422,
    "timestep_mean": 485.0625,
    "epsilon_norm": 8.544803619384766,
    "prediction_norm": 7.358559608459473,
    "timestamp": "2025-07-14T12:34:05.608651"
  },
  {
    "epoch": 903,
    "batch_idx": 160,
    "loss": 0.3877277970314026,
    "learning_rate": 0.0005758891868389422,
    "timestep_mean": 553.90625,
    "epsilon_norm": 8.187565803527832,
    "prediction_norm": 6.286342620849609,
    "timestamp": "2025-07-14T12:34:05.613794"
  },
  {
    "epoch": 903,
    "batch_idx": 192,
    "loss": 0.06996425986289978,
    "learning_rate": 0.0005758891868389422,
    "timestep_mean": 596.75,
    "epsilon_norm": 3.4472157955169678,
    "prediction_norm": 3.5255253314971924,
    "timestamp": "2025-07-14T12:34:05.618970"
  },
  {
    "epoch": 904,
    "batch_idx": 0,
    "loss": 0.1409594714641571,
    "learning_rate": 0.0005751127945603791,
    "timestep_mean": 461.90625,
    "epsilon_norm": 7.128576755523682,
    "prediction_norm": 6.513596534729004,
    "timestamp": "2025-07-14T12:34:05.624139"
  },
  {
    "epoch": 904,
    "batch_idx": 32,
    "loss": 0.2104521095752716,
    "learning_rate": 0.0005751127945603791,
    "timestep_mean": 537.03125,
    "epsilon_norm": 8.380943298339844,
    "prediction_norm": 7.729640007019043,
    "timestamp": "2025-07-14T12:34:05.629392"
  },
  {
    "epoch": 904,
    "batch_idx": 64,
    "loss": 0.40034201741218567,
    "learning_rate": 0.0005751127945603791,
    "timestep_mean": 423.0,
    "epsilon_norm": 8.131003379821777,
    "prediction_norm": 6.395102024078369,
    "timestamp": "2025-07-14T12:34:05.634457"
  },
  {
    "epoch": 904,
    "batch_idx": 96,
    "loss": 0.28083014488220215,
    "learning_rate": 0.0005751127945603791,
    "timestep_mean": 560.40625,
    "epsilon_norm": 8.42556095123291,
    "prediction_norm": 7.387195110321045,
    "timestamp": "2025-07-14T12:34:05.639571"
  },
  {
    "epoch": 904,
    "batch_idx": 128,
    "loss": 0.22270408272743225,
    "learning_rate": 0.0005751127945603791,
    "timestep_mean": 562.65625,
    "epsilon_norm": 7.7578125,
    "prediction_norm": 6.7315778732299805,
    "timestamp": "2025-07-14T12:34:05.644709"
  },
  {
    "epoch": 904,
    "batch_idx": 160,
    "loss": 0.21401995420455933,
    "learning_rate": 0.0005751127945603791,
    "timestep_mean": 543.625,
    "epsilon_norm": 8.444463729858398,
    "prediction_norm": 7.513916492462158,
    "timestamp": "2025-07-14T12:34:05.649746"
  },
  {
    "epoch": 904,
    "batch_idx": 192,
    "loss": 0.05657602846622467,
    "learning_rate": 0.0005751127945603791,
    "timestep_mean": 498.75,
    "epsilon_norm": 2.940906524658203,
    "prediction_norm": 2.762194871902466,
    "timestamp": "2025-07-14T12:34:05.655015"
  },
  {
    "epoch": 905,
    "batch_idx": 0,
    "loss": 0.25946956872940063,
    "learning_rate": 0.0005743362169484622,
    "timestep_mean": 562.125,
    "epsilon_norm": 8.314918518066406,
    "prediction_norm": 7.265654563903809,
    "timestamp": "2025-07-14T12:34:05.660097"
  },
  {
    "epoch": 905,
    "batch_idx": 32,
    "loss": 0.2956189811229706,
    "learning_rate": 0.0005743362169484622,
    "timestep_mean": 536.03125,
    "epsilon_norm": 7.2144598960876465,
    "prediction_norm": 6.33744478225708,
    "timestamp": "2025-07-14T12:34:05.665218"
  },
  {
    "epoch": 905,
    "batch_idx": 64,
    "loss": 0.33645182847976685,
    "learning_rate": 0.0005743362169484622,
    "timestep_mean": 489.4375,
    "epsilon_norm": 8.738652229309082,
    "prediction_norm": 7.802997589111328,
    "timestamp": "2025-07-14T12:34:05.670358"
  },
  {
    "epoch": 905,
    "batch_idx": 96,
    "loss": 0.5582842826843262,
    "learning_rate": 0.0005743362169484622,
    "timestep_mean": 448.3125,
    "epsilon_norm": 8.836379051208496,
    "prediction_norm": 7.048369884490967,
    "timestamp": "2025-07-14T12:34:05.675505"
  },
  {
    "epoch": 905,
    "batch_idx": 128,
    "loss": 0.39753061532974243,
    "learning_rate": 0.0005743362169484622,
    "timestep_mean": 422.25,
    "epsilon_norm": 8.703082084655762,
    "prediction_norm": 7.136409759521484,
    "timestamp": "2025-07-14T12:34:05.680497"
  },
  {
    "epoch": 905,
    "batch_idx": 160,
    "loss": 0.25147950649261475,
    "learning_rate": 0.0005743362169484622,
    "timestep_mean": 571.25,
    "epsilon_norm": 9.258674621582031,
    "prediction_norm": 7.695965766906738,
    "timestamp": "2025-07-14T12:34:05.685649"
  },
  {
    "epoch": 905,
    "batch_idx": 192,
    "loss": 0.32815656065940857,
    "learning_rate": 0.0005743362169484622,
    "timestep_mean": 483.625,
    "epsilon_norm": 5.568080902099609,
    "prediction_norm": 4.957209587097168,
    "timestamp": "2025-07-14T12:34:05.690621"
  },
  {
    "epoch": 906,
    "batch_idx": 0,
    "loss": 0.25230684876441956,
    "learning_rate": 0.0005735594559193193,
    "timestep_mean": 521.71875,
    "epsilon_norm": 8.980864524841309,
    "prediction_norm": 7.518287181854248,
    "timestamp": "2025-07-14T12:34:05.695797"
  },
  {
    "epoch": 906,
    "batch_idx": 32,
    "loss": 0.287427693605423,
    "learning_rate": 0.0005735594559193193,
    "timestep_mean": 529.46875,
    "epsilon_norm": 7.671960830688477,
    "prediction_norm": 5.8840250968933105,
    "timestamp": "2025-07-14T12:34:05.700905"
  },
  {
    "epoch": 906,
    "batch_idx": 64,
    "loss": 0.17096754908561707,
    "learning_rate": 0.0005735594559193193,
    "timestep_mean": 464.875,
    "epsilon_norm": 7.74791145324707,
    "prediction_norm": 6.960709571838379,
    "timestamp": "2025-07-14T12:34:05.706106"
  },
  {
    "epoch": 906,
    "batch_idx": 96,
    "loss": 0.2107934057712555,
    "learning_rate": 0.0005735594559193193,
    "timestep_mean": 435.6875,
    "epsilon_norm": 7.053631782531738,
    "prediction_norm": 6.40226936340332,
    "timestamp": "2025-07-14T12:34:05.713548"
  },
  {
    "epoch": 906,
    "batch_idx": 128,
    "loss": 0.1887325644493103,
    "learning_rate": 0.0005735594559193193,
    "timestep_mean": 579.21875,
    "epsilon_norm": 7.567866325378418,
    "prediction_norm": 6.068504810333252,
    "timestamp": "2025-07-14T12:34:05.718409"
  },
  {
    "epoch": 906,
    "batch_idx": 160,
    "loss": 0.20535951852798462,
    "learning_rate": 0.0005735594559193193,
    "timestep_mean": 504.5,
    "epsilon_norm": 8.732232093811035,
    "prediction_norm": 7.673934459686279,
    "timestamp": "2025-07-14T12:34:05.723243"
  },
  {
    "epoch": 906,
    "batch_idx": 192,
    "loss": 0.3975923955440521,
    "learning_rate": 0.0005735594559193193,
    "timestep_mean": 608.875,
    "epsilon_norm": 4.184351444244385,
    "prediction_norm": 3.587711811065674,
    "timestamp": "2025-07-14T12:34:05.728561"
  },
  {
    "epoch": 907,
    "batch_idx": 0,
    "loss": 0.26293328404426575,
    "learning_rate": 0.0005727825133895314,
    "timestep_mean": 517.3125,
    "epsilon_norm": 7.821669101715088,
    "prediction_norm": 6.8236823081970215,
    "timestamp": "2025-07-14T12:34:05.734249"
  },
  {
    "epoch": 907,
    "batch_idx": 32,
    "loss": 0.33015716075897217,
    "learning_rate": 0.0005727825133895314,
    "timestep_mean": 412.6875,
    "epsilon_norm": 8.382784843444824,
    "prediction_norm": 6.808706283569336,
    "timestamp": "2025-07-14T12:34:05.739425"
  },
  {
    "epoch": 907,
    "batch_idx": 64,
    "loss": 0.2695388197898865,
    "learning_rate": 0.0005727825133895314,
    "timestep_mean": 486.6875,
    "epsilon_norm": 7.835634708404541,
    "prediction_norm": 6.75328254699707,
    "timestamp": "2025-07-14T12:34:05.744640"
  },
  {
    "epoch": 907,
    "batch_idx": 96,
    "loss": 0.2794477343559265,
    "learning_rate": 0.0005727825133895314,
    "timestep_mean": 547.34375,
    "epsilon_norm": 7.831078052520752,
    "prediction_norm": 6.7573370933532715,
    "timestamp": "2025-07-14T12:34:05.749834"
  },
  {
    "epoch": 907,
    "batch_idx": 128,
    "loss": 0.19483114778995514,
    "learning_rate": 0.0005727825133895314,
    "timestep_mean": 582.21875,
    "epsilon_norm": 8.534419059753418,
    "prediction_norm": 7.941013336181641,
    "timestamp": "2025-07-14T12:34:05.754926"
  },
  {
    "epoch": 907,
    "batch_idx": 160,
    "loss": 0.12562614679336548,
    "learning_rate": 0.0005727825133895314,
    "timestep_mean": 455.875,
    "epsilon_norm": 6.319965839385986,
    "prediction_norm": 5.736464500427246,
    "timestamp": "2025-07-14T12:34:05.759963"
  },
  {
    "epoch": 907,
    "batch_idx": 192,
    "loss": 0.484813928604126,
    "learning_rate": 0.0005727825133895314,
    "timestep_mean": 484.25,
    "epsilon_norm": 3.607962131500244,
    "prediction_norm": 2.7894322872161865,
    "timestamp": "2025-07-14T12:34:05.765030"
  },
  {
    "epoch": 908,
    "batch_idx": 0,
    "loss": 0.28505849838256836,
    "learning_rate": 0.0005720053912761267,
    "timestep_mean": 425.09375,
    "epsilon_norm": 7.845763206481934,
    "prediction_norm": 6.771965980529785,
    "timestamp": "2025-07-14T12:34:05.770037"
  },
  {
    "epoch": 908,
    "batch_idx": 32,
    "loss": 0.27613383531570435,
    "learning_rate": 0.0005720053912761267,
    "timestep_mean": 457.125,
    "epsilon_norm": 7.564994812011719,
    "prediction_norm": 6.039112091064453,
    "timestamp": "2025-07-14T12:34:05.775002"
  },
  {
    "epoch": 908,
    "batch_idx": 64,
    "loss": 0.2939907908439636,
    "learning_rate": 0.0005720053912761267,
    "timestep_mean": 488.5625,
    "epsilon_norm": 8.113941192626953,
    "prediction_norm": 6.4568772315979,
    "timestamp": "2025-07-14T12:34:05.780062"
  },
  {
    "epoch": 908,
    "batch_idx": 96,
    "loss": 0.24381370842456818,
    "learning_rate": 0.0005720053912761267,
    "timestep_mean": 547.3125,
    "epsilon_norm": 7.518481254577637,
    "prediction_norm": 7.270513534545898,
    "timestamp": "2025-07-14T12:34:05.785035"
  },
  {
    "epoch": 908,
    "batch_idx": 128,
    "loss": 0.3874853849411011,
    "learning_rate": 0.0005720053912761267,
    "timestep_mean": 472.96875,
    "epsilon_norm": 9.26313304901123,
    "prediction_norm": 8.064424514770508,
    "timestamp": "2025-07-14T12:34:05.790040"
  },
  {
    "epoch": 908,
    "batch_idx": 160,
    "loss": 0.1754293441772461,
    "learning_rate": 0.0005720053912761267,
    "timestep_mean": 518.1875,
    "epsilon_norm": 8.243363380432129,
    "prediction_norm": 7.754563331604004,
    "timestamp": "2025-07-14T12:34:05.795084"
  },
  {
    "epoch": 908,
    "batch_idx": 192,
    "loss": 0.40772321820259094,
    "learning_rate": 0.0005720053912761267,
    "timestep_mean": 655.0,
    "epsilon_norm": 4.304078578948975,
    "prediction_norm": 3.233877182006836,
    "timestamp": "2025-07-14T12:34:05.800057"
  },
  {
    "epoch": 909,
    "batch_idx": 0,
    "loss": 0.24924078583717346,
    "learning_rate": 0.0005712280914965769,
    "timestep_mean": 507.46875,
    "epsilon_norm": 8.1273832321167,
    "prediction_norm": 6.929134845733643,
    "timestamp": "2025-07-14T12:34:05.804887"
  },
  {
    "epoch": 909,
    "batch_idx": 32,
    "loss": 0.21436908841133118,
    "learning_rate": 0.0005712280914965769,
    "timestep_mean": 537.28125,
    "epsilon_norm": 7.793774127960205,
    "prediction_norm": 6.598304271697998,
    "timestamp": "2025-07-14T12:34:05.809655"
  },
  {
    "epoch": 909,
    "batch_idx": 64,
    "loss": 0.39157167077064514,
    "learning_rate": 0.0005712280914965769,
    "timestep_mean": 448.25,
    "epsilon_norm": 8.071457862854004,
    "prediction_norm": 6.042034149169922,
    "timestamp": "2025-07-14T12:34:05.814432"
  },
  {
    "epoch": 909,
    "batch_idx": 96,
    "loss": 0.10875679552555084,
    "learning_rate": 0.0005712280914965769,
    "timestep_mean": 606.21875,
    "epsilon_norm": 7.292412281036377,
    "prediction_norm": 6.4561448097229,
    "timestamp": "2025-07-14T12:34:05.819218"
  },
  {
    "epoch": 909,
    "batch_idx": 128,
    "loss": 0.5271050333976746,
    "learning_rate": 0.0005712280914965769,
    "timestep_mean": 476.625,
    "epsilon_norm": 8.716034889221191,
    "prediction_norm": 6.700857162475586,
    "timestamp": "2025-07-14T12:34:05.824156"
  },
  {
    "epoch": 909,
    "batch_idx": 160,
    "loss": 0.2759948968887329,
    "learning_rate": 0.0005712280914965769,
    "timestep_mean": 546.59375,
    "epsilon_norm": 9.081547737121582,
    "prediction_norm": 7.525717258453369,
    "timestamp": "2025-07-14T12:34:05.829226"
  },
  {
    "epoch": 909,
    "batch_idx": 192,
    "loss": 0.17934449017047882,
    "learning_rate": 0.0005712280914965769,
    "timestep_mean": 509.5,
    "epsilon_norm": 3.160714626312256,
    "prediction_norm": 2.571213722229004,
    "timestamp": "2025-07-14T12:34:05.834229"
  },
  {
    "epoch": 910,
    "batch_idx": 0,
    "loss": 0.25314831733703613,
    "learning_rate": 0.0005704506159687919,
    "timestep_mean": 512.5625,
    "epsilon_norm": 8.34986400604248,
    "prediction_norm": 7.147871494293213,
    "timestamp": "2025-07-14T12:34:05.839382"
  },
  {
    "epoch": 910,
    "batch_idx": 32,
    "loss": 0.3507423996925354,
    "learning_rate": 0.0005704506159687919,
    "timestep_mean": 517.5625,
    "epsilon_norm": 8.42574691772461,
    "prediction_norm": 6.9948811531066895,
    "timestamp": "2025-07-14T12:34:05.844431"
  },
  {
    "epoch": 910,
    "batch_idx": 64,
    "loss": 0.31255829334259033,
    "learning_rate": 0.0005704506159687919,
    "timestep_mean": 535.03125,
    "epsilon_norm": 7.952116966247559,
    "prediction_norm": 6.538572788238525,
    "timestamp": "2025-07-14T12:34:05.849515"
  },
  {
    "epoch": 910,
    "batch_idx": 96,
    "loss": 0.1626448631286621,
    "learning_rate": 0.0005704506159687919,
    "timestep_mean": 557.46875,
    "epsilon_norm": 8.304874420166016,
    "prediction_norm": 8.211124420166016,
    "timestamp": "2025-07-14T12:34:05.854601"
  },
  {
    "epoch": 910,
    "batch_idx": 128,
    "loss": 0.47037258744239807,
    "learning_rate": 0.0005704506159687919,
    "timestep_mean": 548.125,
    "epsilon_norm": 8.394200325012207,
    "prediction_norm": 6.780433654785156,
    "timestamp": "2025-07-14T12:34:05.859677"
  },
  {
    "epoch": 910,
    "batch_idx": 160,
    "loss": 0.4508857727050781,
    "learning_rate": 0.0005704506159687919,
    "timestep_mean": 462.28125,
    "epsilon_norm": 8.755654335021973,
    "prediction_norm": 7.011559009552002,
    "timestamp": "2025-07-14T12:34:05.864785"
  },
  {
    "epoch": 910,
    "batch_idx": 192,
    "loss": 0.11518629640340805,
    "learning_rate": 0.0005704506159687919,
    "timestep_mean": 580.125,
    "epsilon_norm": 3.368004560470581,
    "prediction_norm": 3.1739182472229004,
    "timestamp": "2025-07-14T12:34:05.869697"
  },
  {
    "epoch": 911,
    "batch_idx": 0,
    "loss": 0.42894676327705383,
    "learning_rate": 0.0005696729666111154,
    "timestep_mean": 477.40625,
    "epsilon_norm": 8.697294235229492,
    "prediction_norm": 7.770179271697998,
    "timestamp": "2025-07-14T12:34:05.874658"
  },
  {
    "epoch": 911,
    "batch_idx": 32,
    "loss": 0.20545965433120728,
    "learning_rate": 0.0005696729666111154,
    "timestep_mean": 524.375,
    "epsilon_norm": 8.134483337402344,
    "prediction_norm": 7.077098369598389,
    "timestamp": "2025-07-14T12:34:05.879704"
  },
  {
    "epoch": 911,
    "batch_idx": 64,
    "loss": 0.15953098237514496,
    "learning_rate": 0.0005696729666111154,
    "timestep_mean": 497.21875,
    "epsilon_norm": 6.889383316040039,
    "prediction_norm": 6.658326625823975,
    "timestamp": "2025-07-14T12:34:05.884755"
  },
  {
    "epoch": 911,
    "batch_idx": 96,
    "loss": 0.3232046365737915,
    "learning_rate": 0.0005696729666111154,
    "timestep_mean": 528.3125,
    "epsilon_norm": 7.9240007400512695,
    "prediction_norm": 7.348745822906494,
    "timestamp": "2025-07-14T12:34:05.889902"
  },
  {
    "epoch": 911,
    "batch_idx": 128,
    "loss": 0.3620493710041046,
    "learning_rate": 0.0005696729666111154,
    "timestep_mean": 451.40625,
    "epsilon_norm": 7.825727462768555,
    "prediction_norm": 6.920400142669678,
    "timestamp": "2025-07-14T12:34:05.894950"
  },
  {
    "epoch": 911,
    "batch_idx": 160,
    "loss": 0.233059361577034,
    "learning_rate": 0.0005696729666111154,
    "timestep_mean": 491.5,
    "epsilon_norm": 7.493762493133545,
    "prediction_norm": 6.9619832038879395,
    "timestamp": "2025-07-14T12:34:05.900233"
  },
  {
    "epoch": 911,
    "batch_idx": 192,
    "loss": 0.3727489709854126,
    "learning_rate": 0.0005696729666111154,
    "timestep_mean": 450.0,
    "epsilon_norm": 3.264446496963501,
    "prediction_norm": 1.8720020055770874,
    "timestamp": "2025-07-14T12:34:05.905288"
  },
  {
    "epoch": 912,
    "batch_idx": 0,
    "loss": 0.1894022822380066,
    "learning_rate": 0.0005688951453423197,
    "timestep_mean": 511.15625,
    "epsilon_norm": 8.214667320251465,
    "prediction_norm": 7.567323207855225,
    "timestamp": "2025-07-14T12:34:05.910274"
  },
  {
    "epoch": 912,
    "batch_idx": 32,
    "loss": 0.32186657190322876,
    "learning_rate": 0.0005688951453423197,
    "timestep_mean": 433.21875,
    "epsilon_norm": 8.326351165771484,
    "prediction_norm": 7.3744330406188965,
    "timestamp": "2025-07-14T12:34:05.915197"
  },
  {
    "epoch": 912,
    "batch_idx": 64,
    "loss": 0.19458048045635223,
    "learning_rate": 0.0005688951453423197,
    "timestep_mean": 500.375,
    "epsilon_norm": 7.744546890258789,
    "prediction_norm": 6.843395709991455,
    "timestamp": "2025-07-14T12:34:05.920093"
  },
  {
    "epoch": 912,
    "batch_idx": 96,
    "loss": 0.235472172498703,
    "learning_rate": 0.0005688951453423197,
    "timestep_mean": 548.75,
    "epsilon_norm": 7.54712438583374,
    "prediction_norm": 6.2291259765625,
    "timestamp": "2025-07-14T12:34:05.924976"
  },
  {
    "epoch": 912,
    "batch_idx": 128,
    "loss": 0.14057233929634094,
    "learning_rate": 0.0005688951453423197,
    "timestep_mean": 605.09375,
    "epsilon_norm": 7.390634059906006,
    "prediction_norm": 6.491754055023193,
    "timestamp": "2025-07-14T12:34:05.930226"
  },
  {
    "epoch": 912,
    "batch_idx": 160,
    "loss": 0.26643064618110657,
    "learning_rate": 0.0005688951453423197,
    "timestep_mean": 471.9375,
    "epsilon_norm": 7.904229640960693,
    "prediction_norm": 6.502196788787842,
    "timestamp": "2025-07-14T12:34:05.935502"
  },
  {
    "epoch": 912,
    "batch_idx": 192,
    "loss": 0.3077910542488098,
    "learning_rate": 0.0005688951453423197,
    "timestep_mean": 534.625,
    "epsilon_norm": 3.0409255027770996,
    "prediction_norm": 2.7325146198272705,
    "timestamp": "2025-07-14T12:34:05.942860"
  },
  {
    "epoch": 913,
    "batch_idx": 0,
    "loss": 0.3224911093711853,
    "learning_rate": 0.0005681171540816014,
    "timestep_mean": 468.90625,
    "epsilon_norm": 8.1630859375,
    "prediction_norm": 6.02455472946167,
    "timestamp": "2025-07-14T12:34:05.952847"
  },
  {
    "epoch": 913,
    "batch_idx": 32,
    "loss": 0.4148227572441101,
    "learning_rate": 0.0005681171540816014,
    "timestep_mean": 368.84375,
    "epsilon_norm": 7.606114387512207,
    "prediction_norm": 4.813307762145996,
    "timestamp": "2025-07-14T12:34:05.957886"
  },
  {
    "epoch": 913,
    "batch_idx": 64,
    "loss": 0.30481255054473877,
    "learning_rate": 0.0005681171540816014,
    "timestep_mean": 464.25,
    "epsilon_norm": 7.292048454284668,
    "prediction_norm": 5.479320526123047,
    "timestamp": "2025-07-14T12:34:05.963051"
  },
  {
    "epoch": 913,
    "batch_idx": 96,
    "loss": 0.23252607882022858,
    "learning_rate": 0.0005681171540816014,
    "timestep_mean": 537.96875,
    "epsilon_norm": 7.752995014190674,
    "prediction_norm": 6.418111324310303,
    "timestamp": "2025-07-14T12:34:05.968086"
  },
  {
    "epoch": 913,
    "batch_idx": 128,
    "loss": 0.3646955192089081,
    "learning_rate": 0.0005681171540816014,
    "timestep_mean": 434.34375,
    "epsilon_norm": 7.6918230056762695,
    "prediction_norm": 5.380128383636475,
    "timestamp": "2025-07-14T12:34:05.972967"
  },
  {
    "epoch": 913,
    "batch_idx": 160,
    "loss": 0.36609870195388794,
    "learning_rate": 0.0005681171540816014,
    "timestep_mean": 442.875,
    "epsilon_norm": 8.863515853881836,
    "prediction_norm": 6.766391754150391,
    "timestamp": "2025-07-14T12:34:05.977956"
  },
  {
    "epoch": 913,
    "batch_idx": 192,
    "loss": 0.7296950221061707,
    "learning_rate": 0.0005681171540816014,
    "timestep_mean": 579.625,
    "epsilon_norm": 5.150876522064209,
    "prediction_norm": 4.31273078918457,
    "timestamp": "2025-07-14T12:34:05.982928"
  },
  {
    "epoch": 914,
    "batch_idx": 0,
    "loss": 0.11631465703248978,
    "learning_rate": 0.0005673389947485768,
    "timestep_mean": 591.9375,
    "epsilon_norm": 7.202476978302002,
    "prediction_norm": 6.701446533203125,
    "timestamp": "2025-07-14T12:34:05.988019"
  },
  {
    "epoch": 914,
    "batch_idx": 32,
    "loss": 0.2690355181694031,
    "learning_rate": 0.0005673389947485768,
    "timestep_mean": 518.125,
    "epsilon_norm": 6.936899185180664,
    "prediction_norm": 5.470364570617676,
    "timestamp": "2025-07-14T12:34:05.992960"
  },
  {
    "epoch": 914,
    "batch_idx": 64,
    "loss": 0.27118679881095886,
    "learning_rate": 0.0005673389947485768,
    "timestep_mean": 561.59375,
    "epsilon_norm": 7.820128917694092,
    "prediction_norm": 5.856356143951416,
    "timestamp": "2025-07-14T12:34:05.997893"
  },
  {
    "epoch": 914,
    "batch_idx": 96,
    "loss": 0.44200149178504944,
    "learning_rate": 0.0005673389947485768,
    "timestep_mean": 472.5625,
    "epsilon_norm": 7.6177144050598145,
    "prediction_norm": 6.027248859405518,
    "timestamp": "2025-07-14T12:34:06.002840"
  },
  {
    "epoch": 914,
    "batch_idx": 128,
    "loss": 0.2307654619216919,
    "learning_rate": 0.0005673389947485768,
    "timestep_mean": 595.59375,
    "epsilon_norm": 7.905410289764404,
    "prediction_norm": 7.0809102058410645,
    "timestamp": "2025-07-14T12:34:06.007792"
  },
  {
    "epoch": 914,
    "batch_idx": 160,
    "loss": 0.30889180302619934,
    "learning_rate": 0.0005673389947485768,
    "timestep_mean": 454.46875,
    "epsilon_norm": 8.22439956665039,
    "prediction_norm": 6.996433258056641,
    "timestamp": "2025-07-14T12:34:06.012853"
  },
  {
    "epoch": 914,
    "batch_idx": 192,
    "loss": 0.3909858167171478,
    "learning_rate": 0.0005673389947485768,
    "timestep_mean": 386.375,
    "epsilon_norm": 5.697645664215088,
    "prediction_norm": 4.90760612487793,
    "timestamp": "2025-07-14T12:34:06.017807"
  },
  {
    "epoch": 915,
    "batch_idx": 0,
    "loss": 0.5265810489654541,
    "learning_rate": 0.0005665606692632769,
    "timestep_mean": 520.34375,
    "epsilon_norm": 8.543952941894531,
    "prediction_norm": 6.358277797698975,
    "timestamp": "2025-07-14T12:34:06.022846"
  },
  {
    "epoch": 915,
    "batch_idx": 32,
    "loss": 0.4205463230609894,
    "learning_rate": 0.0005665606692632769,
    "timestep_mean": 465.9375,
    "epsilon_norm": 8.746060371398926,
    "prediction_norm": 6.95917272567749,
    "timestamp": "2025-07-14T12:34:06.027836"
  },
  {
    "epoch": 915,
    "batch_idx": 64,
    "loss": 0.32640889286994934,
    "learning_rate": 0.0005665606692632769,
    "timestep_mean": 496.03125,
    "epsilon_norm": 7.90618371963501,
    "prediction_norm": 6.484377861022949,
    "timestamp": "2025-07-14T12:34:06.032862"
  },
  {
    "epoch": 915,
    "batch_idx": 96,
    "loss": 0.49738043546676636,
    "learning_rate": 0.0005665606692632769,
    "timestep_mean": 499.5,
    "epsilon_norm": 9.899617195129395,
    "prediction_norm": 8.247093200683594,
    "timestamp": "2025-07-14T12:34:06.037919"
  },
  {
    "epoch": 915,
    "batch_idx": 128,
    "loss": 0.3086029887199402,
    "learning_rate": 0.0005665606692632769,
    "timestep_mean": 557.6875,
    "epsilon_norm": 7.670957565307617,
    "prediction_norm": 6.542494773864746,
    "timestamp": "2025-07-14T12:34:06.042939"
  },
  {
    "epoch": 915,
    "batch_idx": 160,
    "loss": 0.17573405802249908,
    "learning_rate": 0.0005665606692632769,
    "timestep_mean": 521.5,
    "epsilon_norm": 7.148893356323242,
    "prediction_norm": 6.269059658050537,
    "timestamp": "2025-07-14T12:34:06.047955"
  },
  {
    "epoch": 915,
    "batch_idx": 192,
    "loss": 0.07971354573965073,
    "learning_rate": 0.0005665606692632769,
    "timestep_mean": 672.5,
    "epsilon_norm": 4.223128795623779,
    "prediction_norm": 4.175045490264893,
    "timestamp": "2025-07-14T12:34:06.052921"
  },
  {
    "epoch": 916,
    "batch_idx": 0,
    "loss": 0.3475343585014343,
    "learning_rate": 0.0005657821795461419,
    "timestep_mean": 396.96875,
    "epsilon_norm": 7.770991325378418,
    "prediction_norm": 6.391047477722168,
    "timestamp": "2025-07-14T12:34:06.058001"
  },
  {
    "epoch": 916,
    "batch_idx": 32,
    "loss": 0.2242167890071869,
    "learning_rate": 0.0005657821795461419,
    "timestep_mean": 511.625,
    "epsilon_norm": 7.479574680328369,
    "prediction_norm": 6.955977916717529,
    "timestamp": "2025-07-14T12:34:06.062916"
  },
  {
    "epoch": 916,
    "batch_idx": 64,
    "loss": 0.18332919478416443,
    "learning_rate": 0.0005657821795461419,
    "timestep_mean": 514.09375,
    "epsilon_norm": 7.938116073608398,
    "prediction_norm": 7.12703275680542,
    "timestamp": "2025-07-14T12:34:06.067944"
  },
  {
    "epoch": 916,
    "batch_idx": 96,
    "loss": 0.4474262595176697,
    "learning_rate": 0.0005657821795461419,
    "timestep_mean": 418.1875,
    "epsilon_norm": 8.552886962890625,
    "prediction_norm": 6.692952632904053,
    "timestamp": "2025-07-14T12:34:06.072938"
  },
  {
    "epoch": 916,
    "batch_idx": 128,
    "loss": 0.3922017812728882,
    "learning_rate": 0.0005657821795461419,
    "timestep_mean": 456.53125,
    "epsilon_norm": 8.117395401000977,
    "prediction_norm": 6.681220054626465,
    "timestamp": "2025-07-14T12:34:06.077900"
  },
  {
    "epoch": 916,
    "batch_idx": 160,
    "loss": 0.3266749680042267,
    "learning_rate": 0.0005657821795461419,
    "timestep_mean": 515.46875,
    "epsilon_norm": 7.2353715896606445,
    "prediction_norm": 5.576890468597412,
    "timestamp": "2025-07-14T12:34:06.083213"
  },
  {
    "epoch": 916,
    "batch_idx": 192,
    "loss": 0.326469361782074,
    "learning_rate": 0.0005657821795461419,
    "timestep_mean": 381.875,
    "epsilon_norm": 4.234651565551758,
    "prediction_norm": 3.618915319442749,
    "timestamp": "2025-07-14T12:34:06.088416"
  },
  {
    "epoch": 917,
    "batch_idx": 0,
    "loss": 0.21153977513313293,
    "learning_rate": 0.0005650035275180181,
    "timestep_mean": 463.78125,
    "epsilon_norm": 8.004284858703613,
    "prediction_norm": 7.335478782653809,
    "timestamp": "2025-07-14T12:34:06.093489"
  },
  {
    "epoch": 917,
    "batch_idx": 32,
    "loss": 0.43813979625701904,
    "learning_rate": 0.0005650035275180181,
    "timestep_mean": 492.90625,
    "epsilon_norm": 8.922406196594238,
    "prediction_norm": 7.002470016479492,
    "timestamp": "2025-07-14T12:34:06.098503"
  },
  {
    "epoch": 917,
    "batch_idx": 64,
    "loss": 0.31651952862739563,
    "learning_rate": 0.0005650035275180181,
    "timestep_mean": 473.65625,
    "epsilon_norm": 7.7084550857543945,
    "prediction_norm": 6.440719127655029,
    "timestamp": "2025-07-14T12:34:06.103479"
  },
  {
    "epoch": 917,
    "batch_idx": 96,
    "loss": 0.189138263463974,
    "learning_rate": 0.0005650035275180181,
    "timestep_mean": 512.3125,
    "epsilon_norm": 7.314778804779053,
    "prediction_norm": 6.725612163543701,
    "timestamp": "2025-07-14T12:34:06.108367"
  },
  {
    "epoch": 917,
    "batch_idx": 128,
    "loss": 0.23276901245117188,
    "learning_rate": 0.0005650035275180181,
    "timestep_mean": 406.90625,
    "epsilon_norm": 7.227367877960205,
    "prediction_norm": 5.87640380859375,
    "timestamp": "2025-07-14T12:34:06.113229"
  },
  {
    "epoch": 917,
    "batch_idx": 160,
    "loss": 0.24109384417533875,
    "learning_rate": 0.0005650035275180181,
    "timestep_mean": 462.875,
    "epsilon_norm": 7.6305952072143555,
    "prediction_norm": 6.05255126953125,
    "timestamp": "2025-07-14T12:34:06.118207"
  },
  {
    "epoch": 917,
    "batch_idx": 192,
    "loss": 0.21140021085739136,
    "learning_rate": 0.0005650035275180181,
    "timestep_mean": 579.875,
    "epsilon_norm": 2.910682201385498,
    "prediction_norm": 1.8295245170593262,
    "timestamp": "2025-07-14T12:34:06.123204"
  },
  {
    "epoch": 918,
    "batch_idx": 0,
    "loss": 0.32805371284484863,
    "learning_rate": 0.000564224715100152,
    "timestep_mean": 479.8125,
    "epsilon_norm": 8.307394027709961,
    "prediction_norm": 6.265420913696289,
    "timestamp": "2025-07-14T12:34:06.128169"
  },
  {
    "epoch": 918,
    "batch_idx": 32,
    "loss": 0.24323460459709167,
    "learning_rate": 0.000564224715100152,
    "timestep_mean": 460.0625,
    "epsilon_norm": 6.8393402099609375,
    "prediction_norm": 5.75590705871582,
    "timestamp": "2025-07-14T12:34:06.133199"
  },
  {
    "epoch": 918,
    "batch_idx": 64,
    "loss": 0.1952010989189148,
    "learning_rate": 0.000564224715100152,
    "timestep_mean": 523.34375,
    "epsilon_norm": 7.578185081481934,
    "prediction_norm": 6.008983135223389,
    "timestamp": "2025-07-14T12:34:06.138245"
  },
  {
    "epoch": 918,
    "batch_idx": 96,
    "loss": 0.23491401970386505,
    "learning_rate": 0.000564224715100152,
    "timestep_mean": 471.9375,
    "epsilon_norm": 7.5035505294799805,
    "prediction_norm": 6.007132530212402,
    "timestamp": "2025-07-14T12:34:06.143123"
  },
  {
    "epoch": 918,
    "batch_idx": 128,
    "loss": 0.4022371172904968,
    "learning_rate": 0.000564224715100152,
    "timestep_mean": 403.8125,
    "epsilon_norm": 7.837553977966309,
    "prediction_norm": 6.095694065093994,
    "timestamp": "2025-07-14T12:34:06.148099"
  },
  {
    "epoch": 918,
    "batch_idx": 160,
    "loss": 0.21044747531414032,
    "learning_rate": 0.000564224715100152,
    "timestep_mean": 582.25,
    "epsilon_norm": 7.658638954162598,
    "prediction_norm": 6.81099796295166,
    "timestamp": "2025-07-14T12:34:06.153067"
  },
  {
    "epoch": 918,
    "batch_idx": 192,
    "loss": 0.13879981637001038,
    "learning_rate": 0.000564224715100152,
    "timestep_mean": 689.875,
    "epsilon_norm": 4.086501598358154,
    "prediction_norm": 3.595043897628784,
    "timestamp": "2025-07-14T12:34:06.158046"
  },
  {
    "epoch": 919,
    "batch_idx": 0,
    "loss": 0.2530571222305298,
    "learning_rate": 0.0005634457442141858,
    "timestep_mean": 528.625,
    "epsilon_norm": 8.23285961151123,
    "prediction_norm": 7.270974636077881,
    "timestamp": "2025-07-14T12:34:06.163065"
  },
  {
    "epoch": 919,
    "batch_idx": 32,
    "loss": 0.31303560733795166,
    "learning_rate": 0.0005634457442141858,
    "timestep_mean": 481.71875,
    "epsilon_norm": 8.196310997009277,
    "prediction_norm": 7.090374946594238,
    "timestamp": "2025-07-14T12:34:06.168079"
  },
  {
    "epoch": 919,
    "batch_idx": 64,
    "loss": 0.45629996061325073,
    "learning_rate": 0.0005634457442141858,
    "timestep_mean": 482.0625,
    "epsilon_norm": 8.40721321105957,
    "prediction_norm": 6.498740196228027,
    "timestamp": "2025-07-14T12:34:06.173011"
  },
  {
    "epoch": 919,
    "batch_idx": 96,
    "loss": 0.13803532719612122,
    "learning_rate": 0.0005634457442141858,
    "timestep_mean": 565.90625,
    "epsilon_norm": 7.042270183563232,
    "prediction_norm": 6.50117301940918,
    "timestamp": "2025-07-14T12:34:06.177896"
  },
  {
    "epoch": 919,
    "batch_idx": 128,
    "loss": 0.36010706424713135,
    "learning_rate": 0.0005634457442141858,
    "timestep_mean": 336.84375,
    "epsilon_norm": 8.795001983642578,
    "prediction_norm": 7.2694621086120605,
    "timestamp": "2025-07-14T12:34:06.182840"
  },
  {
    "epoch": 919,
    "batch_idx": 160,
    "loss": 0.20335273444652557,
    "learning_rate": 0.0005634457442141858,
    "timestep_mean": 501.34375,
    "epsilon_norm": 9.933950424194336,
    "prediction_norm": 9.300243377685547,
    "timestamp": "2025-07-14T12:34:06.187883"
  },
  {
    "epoch": 919,
    "batch_idx": 192,
    "loss": 0.10273195803165436,
    "learning_rate": 0.0005634457442141858,
    "timestep_mean": 444.75,
    "epsilon_norm": 3.1528453826904297,
    "prediction_norm": 3.1843879222869873,
    "timestamp": "2025-07-14T12:34:06.192770"
  },
  {
    "epoch": 920,
    "batch_idx": 0,
    "loss": 0.2085884064435959,
    "learning_rate": 0.0005626666167821527,
    "timestep_mean": 496.6875,
    "epsilon_norm": 8.000940322875977,
    "prediction_norm": 7.163867950439453,
    "timestamp": "2025-07-14T12:34:06.197712"
  },
  {
    "epoch": 920,
    "batch_idx": 32,
    "loss": 0.49115997552871704,
    "learning_rate": 0.0005626666167821527,
    "timestep_mean": 438.28125,
    "epsilon_norm": 8.659753799438477,
    "prediction_norm": 6.56193733215332,
    "timestamp": "2025-07-14T12:34:06.202701"
  },
  {
    "epoch": 920,
    "batch_idx": 64,
    "loss": 0.3057960867881775,
    "learning_rate": 0.0005626666167821527,
    "timestep_mean": 468.15625,
    "epsilon_norm": 7.3553924560546875,
    "prediction_norm": 6.404535293579102,
    "timestamp": "2025-07-14T12:34:06.207614"
  },
  {
    "epoch": 920,
    "batch_idx": 96,
    "loss": 0.26136335730552673,
    "learning_rate": 0.0005626666167821527,
    "timestep_mean": 527.03125,
    "epsilon_norm": 7.56173849105835,
    "prediction_norm": 6.947135925292969,
    "timestamp": "2025-07-14T12:34:06.212591"
  },
  {
    "epoch": 920,
    "batch_idx": 128,
    "loss": 0.6212402582168579,
    "learning_rate": 0.0005626666167821527,
    "timestep_mean": 400.0625,
    "epsilon_norm": 8.715580940246582,
    "prediction_norm": 5.835832118988037,
    "timestamp": "2025-07-14T12:34:06.217594"
  },
  {
    "epoch": 920,
    "batch_idx": 160,
    "loss": 0.20492541790008545,
    "learning_rate": 0.0005626666167821527,
    "timestep_mean": 490.28125,
    "epsilon_norm": 8.040386199951172,
    "prediction_norm": 7.440839767456055,
    "timestamp": "2025-07-14T12:34:06.222522"
  },
  {
    "epoch": 920,
    "batch_idx": 192,
    "loss": 0.4922623336315155,
    "learning_rate": 0.0005626666167821527,
    "timestep_mean": 524.125,
    "epsilon_norm": 4.197424411773682,
    "prediction_norm": 3.276623487472534,
    "timestamp": "2025-07-14T12:34:06.227454"
  },
  {
    "epoch": 921,
    "batch_idx": 0,
    "loss": 0.10969637334346771,
    "learning_rate": 0.0005618873347264722,
    "timestep_mean": 565.1875,
    "epsilon_norm": 7.656530380249023,
    "prediction_norm": 7.587988376617432,
    "timestamp": "2025-07-14T12:34:06.232419"
  },
  {
    "epoch": 921,
    "batch_idx": 32,
    "loss": 0.3802688419818878,
    "learning_rate": 0.0005618873347264722,
    "timestep_mean": 459.65625,
    "epsilon_norm": 8.577310562133789,
    "prediction_norm": 7.373763084411621,
    "timestamp": "2025-07-14T12:34:06.237458"
  },
  {
    "epoch": 921,
    "batch_idx": 64,
    "loss": 0.30909332633018494,
    "learning_rate": 0.0005618873347264722,
    "timestep_mean": 481.3125,
    "epsilon_norm": 8.317861557006836,
    "prediction_norm": 7.653716087341309,
    "timestamp": "2025-07-14T12:34:06.242354"
  },
  {
    "epoch": 921,
    "batch_idx": 96,
    "loss": 0.2615140378475189,
    "learning_rate": 0.0005618873347264722,
    "timestep_mean": 482.03125,
    "epsilon_norm": 8.24715518951416,
    "prediction_norm": 7.247982025146484,
    "timestamp": "2025-07-14T12:34:06.247268"
  },
  {
    "epoch": 921,
    "batch_idx": 128,
    "loss": 0.2973495423793793,
    "learning_rate": 0.0005618873347264722,
    "timestep_mean": 477.3125,
    "epsilon_norm": 7.0681610107421875,
    "prediction_norm": 6.1481475830078125,
    "timestamp": "2025-07-14T12:34:06.252166"
  },
  {
    "epoch": 921,
    "batch_idx": 160,
    "loss": 0.1733119785785675,
    "learning_rate": 0.0005618873347264722,
    "timestep_mean": 606.15625,
    "epsilon_norm": 8.114370346069336,
    "prediction_norm": 7.080049514770508,
    "timestamp": "2025-07-14T12:34:06.257070"
  },
  {
    "epoch": 921,
    "batch_idx": 192,
    "loss": 0.21349553763866425,
    "learning_rate": 0.0005618873347264722,
    "timestep_mean": 485.5,
    "epsilon_norm": 3.901277780532837,
    "prediction_norm": 3.8064675331115723,
    "timestamp": "2025-07-14T12:34:06.262040"
  },
  {
    "epoch": 922,
    "batch_idx": 0,
    "loss": 0.26344430446624756,
    "learning_rate": 0.0005611078999699453,
    "timestep_mean": 489.25,
    "epsilon_norm": 8.534524917602539,
    "prediction_norm": 7.108030796051025,
    "timestamp": "2025-07-14T12:34:06.267114"
  },
  {
    "epoch": 922,
    "batch_idx": 32,
    "loss": 0.19734814763069153,
    "learning_rate": 0.0005611078999699453,
    "timestep_mean": 496.0,
    "epsilon_norm": 8.185067176818848,
    "prediction_norm": 6.635866641998291,
    "timestamp": "2025-07-14T12:34:06.271968"
  },
  {
    "epoch": 922,
    "batch_idx": 64,
    "loss": 0.25846511125564575,
    "learning_rate": 0.0005611078999699453,
    "timestep_mean": 414.59375,
    "epsilon_norm": 7.761101245880127,
    "prediction_norm": 6.173789978027344,
    "timestamp": "2025-07-14T12:34:06.276903"
  },
  {
    "epoch": 922,
    "batch_idx": 96,
    "loss": 0.3933957815170288,
    "learning_rate": 0.0005611078999699453,
    "timestep_mean": 550.34375,
    "epsilon_norm": 8.552760124206543,
    "prediction_norm": 6.00821590423584,
    "timestamp": "2025-07-14T12:34:06.281812"
  },
  {
    "epoch": 922,
    "batch_idx": 128,
    "loss": 0.4780978560447693,
    "learning_rate": 0.0005611078999699453,
    "timestep_mean": 415.625,
    "epsilon_norm": 8.843734741210938,
    "prediction_norm": 6.706816673278809,
    "timestamp": "2025-07-14T12:34:06.286721"
  },
  {
    "epoch": 922,
    "batch_idx": 160,
    "loss": 0.10341345518827438,
    "learning_rate": 0.0005611078999699453,
    "timestep_mean": 635.53125,
    "epsilon_norm": 8.22899341583252,
    "prediction_norm": 7.689426898956299,
    "timestamp": "2025-07-14T12:34:06.291639"
  },
  {
    "epoch": 922,
    "batch_idx": 192,
    "loss": 0.520839512348175,
    "learning_rate": 0.0005611078999699453,
    "timestep_mean": 351.75,
    "epsilon_norm": 4.493023872375488,
    "prediction_norm": 3.3433094024658203,
    "timestamp": "2025-07-14T12:34:06.296566"
  },
  {
    "epoch": 923,
    "batch_idx": 0,
    "loss": 0.26754030585289,
    "learning_rate": 0.0005603283144357498,
    "timestep_mean": 453.03125,
    "epsilon_norm": 6.8014140129089355,
    "prediction_norm": 5.736649513244629,
    "timestamp": "2025-07-14T12:34:06.301564"
  },
  {
    "epoch": 923,
    "batch_idx": 32,
    "loss": 0.16476351022720337,
    "learning_rate": 0.0005603283144357498,
    "timestep_mean": 520.90625,
    "epsilon_norm": 7.242525100708008,
    "prediction_norm": 6.224818706512451,
    "timestamp": "2025-07-14T12:34:06.306509"
  },
  {
    "epoch": 923,
    "batch_idx": 64,
    "loss": 0.327181339263916,
    "learning_rate": 0.0005603283144357498,
    "timestep_mean": 421.1875,
    "epsilon_norm": 8.148697853088379,
    "prediction_norm": 6.315329074859619,
    "timestamp": "2025-07-14T12:34:06.311483"
  },
  {
    "epoch": 923,
    "batch_idx": 96,
    "loss": 0.3805800676345825,
    "learning_rate": 0.0005603283144357498,
    "timestep_mean": 551.71875,
    "epsilon_norm": 8.08251667022705,
    "prediction_norm": 6.6611647605896,
    "timestamp": "2025-07-14T12:34:06.316427"
  },
  {
    "epoch": 923,
    "batch_idx": 128,
    "loss": 0.4079879820346832,
    "learning_rate": 0.0005603283144357498,
    "timestep_mean": 456.65625,
    "epsilon_norm": 8.11474895477295,
    "prediction_norm": 6.967841625213623,
    "timestamp": "2025-07-14T12:34:06.321402"
  },
  {
    "epoch": 923,
    "batch_idx": 160,
    "loss": 0.1829899400472641,
    "learning_rate": 0.0005603283144357498,
    "timestep_mean": 601.40625,
    "epsilon_norm": 7.85394811630249,
    "prediction_norm": 6.905040264129639,
    "timestamp": "2025-07-14T12:34:06.326312"
  },
  {
    "epoch": 923,
    "batch_idx": 192,
    "loss": 0.4767742156982422,
    "learning_rate": 0.0005603283144357498,
    "timestep_mean": 330.5,
    "epsilon_norm": 4.741067886352539,
    "prediction_norm": 4.292838096618652,
    "timestamp": "2025-07-14T12:34:06.331185"
  },
  {
    "epoch": 924,
    "batch_idx": 0,
    "loss": 0.5766798257827759,
    "learning_rate": 0.0005595485800474354,
    "timestep_mean": 420.28125,
    "epsilon_norm": 8.073383331298828,
    "prediction_norm": 6.379136562347412,
    "timestamp": "2025-07-14T12:34:06.336262"
  },
  {
    "epoch": 924,
    "batch_idx": 32,
    "loss": 0.2530938982963562,
    "learning_rate": 0.0005595485800474354,
    "timestep_mean": 501.375,
    "epsilon_norm": 8.875685691833496,
    "prediction_norm": 7.8755412101745605,
    "timestamp": "2025-07-14T12:34:06.341429"
  },
  {
    "epoch": 924,
    "batch_idx": 64,
    "loss": 0.2561493217945099,
    "learning_rate": 0.0005595485800474354,
    "timestep_mean": 478.4375,
    "epsilon_norm": 7.462113857269287,
    "prediction_norm": 6.317968845367432,
    "timestamp": "2025-07-14T12:34:06.346710"
  },
  {
    "epoch": 924,
    "batch_idx": 96,
    "loss": 0.21492788195610046,
    "learning_rate": 0.0005595485800474354,
    "timestep_mean": 445.9375,
    "epsilon_norm": 7.218756198883057,
    "prediction_norm": 6.163668155670166,
    "timestamp": "2025-07-14T12:34:06.351866"
  },
  {
    "epoch": 924,
    "batch_idx": 128,
    "loss": 0.16329523921012878,
    "learning_rate": 0.0005595485800474354,
    "timestep_mean": 525.21875,
    "epsilon_norm": 7.577317714691162,
    "prediction_norm": 7.2537841796875,
    "timestamp": "2025-07-14T12:34:06.357100"
  },
  {
    "epoch": 924,
    "batch_idx": 160,
    "loss": 0.20149610936641693,
    "learning_rate": 0.0005595485800474354,
    "timestep_mean": 473.9375,
    "epsilon_norm": 8.776824951171875,
    "prediction_norm": 8.320298194885254,
    "timestamp": "2025-07-14T12:34:06.362241"
  },
  {
    "epoch": 924,
    "batch_idx": 192,
    "loss": 0.08853177726268768,
    "learning_rate": 0.0005595485800474354,
    "timestep_mean": 577.625,
    "epsilon_norm": 4.152514934539795,
    "prediction_norm": 4.107102394104004,
    "timestamp": "2025-07-14T12:34:06.367378"
  },
  {
    "epoch": 925,
    "batch_idx": 0,
    "loss": 0.1882886439561844,
    "learning_rate": 0.0005587686987289194,
    "timestep_mean": 536.6875,
    "epsilon_norm": 8.431295394897461,
    "prediction_norm": 7.392269134521484,
    "timestamp": "2025-07-14T12:34:06.372674"
  },
  {
    "epoch": 925,
    "batch_idx": 32,
    "loss": 0.41964784264564514,
    "learning_rate": 0.0005587686987289194,
    "timestep_mean": 398.59375,
    "epsilon_norm": 9.146117210388184,
    "prediction_norm": 7.056975364685059,
    "timestamp": "2025-07-14T12:34:06.377734"
  },
  {
    "epoch": 925,
    "batch_idx": 64,
    "loss": 0.451067179441452,
    "learning_rate": 0.0005587686987289194,
    "timestep_mean": 470.25,
    "epsilon_norm": 7.881845474243164,
    "prediction_norm": 5.758467674255371,
    "timestamp": "2025-07-14T12:34:06.382590"
  },
  {
    "epoch": 925,
    "batch_idx": 96,
    "loss": 0.21380090713500977,
    "learning_rate": 0.0005587686987289194,
    "timestep_mean": 564.78125,
    "epsilon_norm": 8.590270042419434,
    "prediction_norm": 7.275254726409912,
    "timestamp": "2025-07-14T12:34:06.387701"
  },
  {
    "epoch": 925,
    "batch_idx": 128,
    "loss": 0.2919752895832062,
    "learning_rate": 0.0005587686987289194,
    "timestep_mean": 506.5,
    "epsilon_norm": 8.527487754821777,
    "prediction_norm": 7.295653343200684,
    "timestamp": "2025-07-14T12:34:06.392861"
  },
  {
    "epoch": 925,
    "batch_idx": 160,
    "loss": 0.20962952077388763,
    "learning_rate": 0.0005587686987289194,
    "timestep_mean": 530.90625,
    "epsilon_norm": 6.987278461456299,
    "prediction_norm": 6.25809383392334,
    "timestamp": "2025-07-14T12:34:06.397985"
  },
  {
    "epoch": 925,
    "batch_idx": 192,
    "loss": 0.7012566328048706,
    "learning_rate": 0.0005587686987289194,
    "timestep_mean": 217.5,
    "epsilon_norm": 3.711886405944824,
    "prediction_norm": 1.418651819229126,
    "timestamp": "2025-07-14T12:34:06.403123"
  },
  {
    "epoch": 926,
    "batch_idx": 0,
    "loss": 0.2235291302204132,
    "learning_rate": 0.0005579886724044813,
    "timestep_mean": 556.40625,
    "epsilon_norm": 7.911296844482422,
    "prediction_norm": 6.873495578765869,
    "timestamp": "2025-07-14T12:34:06.408068"
  },
  {
    "epoch": 926,
    "batch_idx": 32,
    "loss": 0.2541562020778656,
    "learning_rate": 0.0005579886724044813,
    "timestep_mean": 496.125,
    "epsilon_norm": 6.582666873931885,
    "prediction_norm": 5.350739002227783,
    "timestamp": "2025-07-14T12:34:06.413038"
  },
  {
    "epoch": 926,
    "batch_idx": 64,
    "loss": 0.2997712790966034,
    "learning_rate": 0.0005579886724044813,
    "timestep_mean": 513.71875,
    "epsilon_norm": 8.391502380371094,
    "prediction_norm": 7.272501468658447,
    "timestamp": "2025-07-14T12:34:06.418658"
  },
  {
    "epoch": 926,
    "batch_idx": 96,
    "loss": 0.2829219102859497,
    "learning_rate": 0.0005579886724044813,
    "timestep_mean": 497.90625,
    "epsilon_norm": 7.720352649688721,
    "prediction_norm": 5.933667182922363,
    "timestamp": "2025-07-14T12:34:06.423903"
  },
  {
    "epoch": 926,
    "batch_idx": 128,
    "loss": 0.29558783769607544,
    "learning_rate": 0.0005579886724044813,
    "timestep_mean": 552.9375,
    "epsilon_norm": 7.837716579437256,
    "prediction_norm": 6.888314247131348,
    "timestamp": "2025-07-14T12:34:06.429385"
  },
  {
    "epoch": 926,
    "batch_idx": 160,
    "loss": 0.38731539249420166,
    "learning_rate": 0.0005579886724044813,
    "timestep_mean": 530.0625,
    "epsilon_norm": 7.9898552894592285,
    "prediction_norm": 6.491220951080322,
    "timestamp": "2025-07-14T12:34:06.434473"
  },
  {
    "epoch": 926,
    "batch_idx": 192,
    "loss": 0.26341134309768677,
    "learning_rate": 0.0005579886724044813,
    "timestep_mean": 505.125,
    "epsilon_norm": 4.375344276428223,
    "prediction_norm": 3.2636966705322266,
    "timestamp": "2025-07-14T12:34:06.439646"
  },
  {
    "epoch": 927,
    "batch_idx": 0,
    "loss": 0.19208168983459473,
    "learning_rate": 0.0005572085029987585,
    "timestep_mean": 487.34375,
    "epsilon_norm": 7.7237629890441895,
    "prediction_norm": 6.502339839935303,
    "timestamp": "2025-07-14T12:34:06.444969"
  },
  {
    "epoch": 927,
    "batch_idx": 32,
    "loss": 0.351557195186615,
    "learning_rate": 0.0005572085029987585,
    "timestep_mean": 407.46875,
    "epsilon_norm": 8.308525085449219,
    "prediction_norm": 6.88414192199707,
    "timestamp": "2025-07-14T12:34:06.450131"
  },
  {
    "epoch": 927,
    "batch_idx": 64,
    "loss": 0.2858472466468811,
    "learning_rate": 0.0005572085029987585,
    "timestep_mean": 457.5625,
    "epsilon_norm": 7.563523292541504,
    "prediction_norm": 6.908519744873047,
    "timestamp": "2025-07-14T12:34:06.455325"
  },
  {
    "epoch": 927,
    "batch_idx": 96,
    "loss": 0.14078709483146667,
    "learning_rate": 0.0005572085029987585,
    "timestep_mean": 601.28125,
    "epsilon_norm": 8.163349151611328,
    "prediction_norm": 7.481476306915283,
    "timestamp": "2025-07-14T12:34:06.460485"
  },
  {
    "epoch": 927,
    "batch_idx": 128,
    "loss": 0.23908382654190063,
    "learning_rate": 0.0005572085029987585,
    "timestep_mean": 511.40625,
    "epsilon_norm": 7.848256587982178,
    "prediction_norm": 6.893388748168945,
    "timestamp": "2025-07-14T12:34:06.465686"
  },
  {
    "epoch": 927,
    "batch_idx": 160,
    "loss": 0.1526823490858078,
    "learning_rate": 0.0005572085029987585,
    "timestep_mean": 508.8125,
    "epsilon_norm": 7.64989709854126,
    "prediction_norm": 7.417102336883545,
    "timestamp": "2025-07-14T12:34:06.470998"
  },
  {
    "epoch": 927,
    "batch_idx": 192,
    "loss": 0.5063849687576294,
    "learning_rate": 0.0005572085029987585,
    "timestep_mean": 387.75,
    "epsilon_norm": 4.596587657928467,
    "prediction_norm": 4.164130210876465,
    "timestamp": "2025-07-14T12:34:06.476320"
  },
  {
    "epoch": 928,
    "batch_idx": 0,
    "loss": 0.17362523078918457,
    "learning_rate": 0.0005564281924367413,
    "timestep_mean": 513.9375,
    "epsilon_norm": 7.362436771392822,
    "prediction_norm": 6.030444622039795,
    "timestamp": "2025-07-14T12:34:06.481488"
  },
  {
    "epoch": 928,
    "batch_idx": 32,
    "loss": 0.23197132349014282,
    "learning_rate": 0.0005564281924367413,
    "timestep_mean": 497.5,
    "epsilon_norm": 8.157916069030762,
    "prediction_norm": 7.120606899261475,
    "timestamp": "2025-07-14T12:34:06.486494"
  },
  {
    "epoch": 928,
    "batch_idx": 64,
    "loss": 0.39274948835372925,
    "learning_rate": 0.0005564281924367413,
    "timestep_mean": 443.03125,
    "epsilon_norm": 8.644282341003418,
    "prediction_norm": 7.015894412994385,
    "timestamp": "2025-07-14T12:34:06.491825"
  },
  {
    "epoch": 928,
    "batch_idx": 96,
    "loss": 0.3814659118652344,
    "learning_rate": 0.0005564281924367413,
    "timestep_mean": 484.9375,
    "epsilon_norm": 9.334322929382324,
    "prediction_norm": 8.359295845031738,
    "timestamp": "2025-07-14T12:34:06.496939"
  },
  {
    "epoch": 928,
    "batch_idx": 128,
    "loss": 0.4533590078353882,
    "learning_rate": 0.0005564281924367413,
    "timestep_mean": 474.0625,
    "epsilon_norm": 8.825699806213379,
    "prediction_norm": 6.760988712310791,
    "timestamp": "2025-07-14T12:34:06.502081"
  },
  {
    "epoch": 928,
    "batch_idx": 160,
    "loss": 0.3343575596809387,
    "learning_rate": 0.0005564281924367413,
    "timestep_mean": 567.4375,
    "epsilon_norm": 7.985288143157959,
    "prediction_norm": 6.789099216461182,
    "timestamp": "2025-07-14T12:34:06.507255"
  },
  {
    "epoch": 928,
    "batch_idx": 192,
    "loss": 0.7806365489959717,
    "learning_rate": 0.0005564281924367413,
    "timestep_mean": 432.0,
    "epsilon_norm": 4.971715927124023,
    "prediction_norm": 3.7111589908599854,
    "timestamp": "2025-07-14T12:34:06.512536"
  },
  {
    "epoch": 929,
    "batch_idx": 0,
    "loss": 0.22943729162216187,
    "learning_rate": 0.0005556477426437689,
    "timestep_mean": 452.5,
    "epsilon_norm": 7.3958964347839355,
    "prediction_norm": 6.3491668701171875,
    "timestamp": "2025-07-14T12:34:06.517767"
  },
  {
    "epoch": 929,
    "batch_idx": 32,
    "loss": 0.41942766308784485,
    "learning_rate": 0.0005556477426437689,
    "timestep_mean": 464.4375,
    "epsilon_norm": 8.392951965332031,
    "prediction_norm": 7.077339172363281,
    "timestamp": "2025-07-14T12:34:06.522907"
  },
  {
    "epoch": 929,
    "batch_idx": 64,
    "loss": 0.14882946014404297,
    "learning_rate": 0.0005556477426437689,
    "timestep_mean": 499.75,
    "epsilon_norm": 6.840330123901367,
    "prediction_norm": 6.703910827636719,
    "timestamp": "2025-07-14T12:34:06.528212"
  },
  {
    "epoch": 929,
    "batch_idx": 96,
    "loss": 0.18792875111103058,
    "learning_rate": 0.0005556477426437689,
    "timestep_mean": 501.71875,
    "epsilon_norm": 7.675678730010986,
    "prediction_norm": 7.163428783416748,
    "timestamp": "2025-07-14T12:34:06.533425"
  },
  {
    "epoch": 929,
    "batch_idx": 128,
    "loss": 0.4372832477092743,
    "learning_rate": 0.0005556477426437689,
    "timestep_mean": 494.0,
    "epsilon_norm": 9.355088233947754,
    "prediction_norm": 7.492390155792236,
    "timestamp": "2025-07-14T12:34:06.538587"
  },
  {
    "epoch": 929,
    "batch_idx": 160,
    "loss": 0.5131868124008179,
    "learning_rate": 0.0005556477426437689,
    "timestep_mean": 457.4375,
    "epsilon_norm": 9.235595703125,
    "prediction_norm": 6.789853096008301,
    "timestamp": "2025-07-14T12:34:06.543763"
  },
  {
    "epoch": 929,
    "batch_idx": 192,
    "loss": 0.24217665195465088,
    "learning_rate": 0.0005556477426437689,
    "timestep_mean": 365.375,
    "epsilon_norm": 4.089996814727783,
    "prediction_norm": 3.485501527786255,
    "timestamp": "2025-07-14T12:34:06.548751"
  },
  {
    "epoch": 930,
    "batch_idx": 0,
    "loss": 0.254591166973114,
    "learning_rate": 0.0005548671555455232,
    "timestep_mean": 478.75,
    "epsilon_norm": 8.180747985839844,
    "prediction_norm": 6.519726753234863,
    "timestamp": "2025-07-14T12:34:06.553965"
  },
  {
    "epoch": 930,
    "batch_idx": 32,
    "loss": 0.26768696308135986,
    "learning_rate": 0.0005548671555455232,
    "timestep_mean": 452.125,
    "epsilon_norm": 7.61171293258667,
    "prediction_norm": 6.4554595947265625,
    "timestamp": "2025-07-14T12:34:06.559092"
  },
  {
    "epoch": 930,
    "batch_idx": 64,
    "loss": 0.37408292293548584,
    "learning_rate": 0.0005548671555455232,
    "timestep_mean": 480.125,
    "epsilon_norm": 8.185362815856934,
    "prediction_norm": 6.59744119644165,
    "timestamp": "2025-07-14T12:34:06.564341"
  },
  {
    "epoch": 930,
    "batch_idx": 96,
    "loss": 0.3192940950393677,
    "learning_rate": 0.0005548671555455232,
    "timestep_mean": 490.4375,
    "epsilon_norm": 8.26539134979248,
    "prediction_norm": 6.541348934173584,
    "timestamp": "2025-07-14T12:34:06.569506"
  },
  {
    "epoch": 930,
    "batch_idx": 128,
    "loss": 0.28056275844573975,
    "learning_rate": 0.0005548671555455232,
    "timestep_mean": 518.8125,
    "epsilon_norm": 8.024489402770996,
    "prediction_norm": 6.407917022705078,
    "timestamp": "2025-07-14T12:34:06.574668"
  },
  {
    "epoch": 930,
    "batch_idx": 160,
    "loss": 0.3278577923774719,
    "learning_rate": 0.0005548671555455232,
    "timestep_mean": 531.65625,
    "epsilon_norm": 8.519620895385742,
    "prediction_norm": 6.496387004852295,
    "timestamp": "2025-07-14T12:34:06.580035"
  },
  {
    "epoch": 930,
    "batch_idx": 192,
    "loss": 0.10569285601377487,
    "learning_rate": 0.0005548671555455232,
    "timestep_mean": 575.125,
    "epsilon_norm": 4.638465404510498,
    "prediction_norm": 4.3163371086120605,
    "timestamp": "2025-07-14T12:34:06.585222"
  },
  {
    "epoch": 931,
    "batch_idx": 0,
    "loss": 0.16282515227794647,
    "learning_rate": 0.0005540864330680254,
    "timestep_mean": 511.15625,
    "epsilon_norm": 7.86868953704834,
    "prediction_norm": 7.0313262939453125,
    "timestamp": "2025-07-14T12:34:06.590361"
  },
  {
    "epoch": 931,
    "batch_idx": 32,
    "loss": 0.22264054417610168,
    "learning_rate": 0.0005540864330680254,
    "timestep_mean": 597.8125,
    "epsilon_norm": 7.7512383460998535,
    "prediction_norm": 6.860194206237793,
    "timestamp": "2025-07-14T12:34:06.595627"
  },
  {
    "epoch": 931,
    "batch_idx": 64,
    "loss": 0.3526955842971802,
    "learning_rate": 0.0005540864330680254,
    "timestep_mean": 502.65625,
    "epsilon_norm": 8.308086395263672,
    "prediction_norm": 6.909103870391846,
    "timestamp": "2025-07-14T12:34:06.600832"
  },
  {
    "epoch": 931,
    "batch_idx": 96,
    "loss": 0.24890261888504028,
    "learning_rate": 0.0005540864330680254,
    "timestep_mean": 487.09375,
    "epsilon_norm": 7.767204284667969,
    "prediction_norm": 7.068057537078857,
    "timestamp": "2025-07-14T12:34:06.606011"
  },
  {
    "epoch": 931,
    "batch_idx": 128,
    "loss": 0.3392798602581024,
    "learning_rate": 0.0005540864330680254,
    "timestep_mean": 498.53125,
    "epsilon_norm": 8.230217933654785,
    "prediction_norm": 6.4642133712768555,
    "timestamp": "2025-07-14T12:34:06.611204"
  },
  {
    "epoch": 931,
    "batch_idx": 160,
    "loss": 0.22588811814785004,
    "learning_rate": 0.0005540864330680254,
    "timestep_mean": 552.71875,
    "epsilon_norm": 9.49576187133789,
    "prediction_norm": 8.201865196228027,
    "timestamp": "2025-07-14T12:34:06.616301"
  },
  {
    "epoch": 931,
    "batch_idx": 192,
    "loss": 0.17130708694458008,
    "learning_rate": 0.0005540864330680254,
    "timestep_mean": 607.25,
    "epsilon_norm": 3.819836139678955,
    "prediction_norm": 4.093475818634033,
    "timestamp": "2025-07-14T12:34:06.621489"
  },
  {
    "epoch": 932,
    "batch_idx": 0,
    "loss": 0.14272311329841614,
    "learning_rate": 0.0005533055771376305,
    "timestep_mean": 540.3125,
    "epsilon_norm": 7.723576545715332,
    "prediction_norm": 7.545607089996338,
    "timestamp": "2025-07-14T12:34:06.626653"
  },
  {
    "epoch": 932,
    "batch_idx": 32,
    "loss": 0.2300669252872467,
    "learning_rate": 0.0005533055771376305,
    "timestep_mean": 475.71875,
    "epsilon_norm": 7.508030891418457,
    "prediction_norm": 5.932998180389404,
    "timestamp": "2025-07-14T12:34:06.631768"
  },
  {
    "epoch": 932,
    "batch_idx": 64,
    "loss": 0.1544569730758667,
    "learning_rate": 0.0005533055771376305,
    "timestep_mean": 630.5625,
    "epsilon_norm": 8.808700561523438,
    "prediction_norm": 8.204480171203613,
    "timestamp": "2025-07-14T12:34:06.637035"
  },
  {
    "epoch": 932,
    "batch_idx": 96,
    "loss": 0.2353072613477707,
    "learning_rate": 0.0005533055771376305,
    "timestep_mean": 492.8125,
    "epsilon_norm": 7.9560089111328125,
    "prediction_norm": 7.1492919921875,
    "timestamp": "2025-07-14T12:34:06.642322"
  },
  {
    "epoch": 932,
    "batch_idx": 128,
    "loss": 0.26752227544784546,
    "learning_rate": 0.0005533055771376305,
    "timestep_mean": 442.78125,
    "epsilon_norm": 7.690432071685791,
    "prediction_norm": 6.261268615722656,
    "timestamp": "2025-07-14T12:34:06.647608"
  },
  {
    "epoch": 932,
    "batch_idx": 160,
    "loss": 0.1892213225364685,
    "learning_rate": 0.0005533055771376305,
    "timestep_mean": 555.40625,
    "epsilon_norm": 8.464704513549805,
    "prediction_norm": 7.725417613983154,
    "timestamp": "2025-07-14T12:34:06.652865"
  },
  {
    "epoch": 932,
    "batch_idx": 192,
    "loss": 0.237616628408432,
    "learning_rate": 0.0005533055771376305,
    "timestep_mean": 614.5,
    "epsilon_norm": 4.054375648498535,
    "prediction_norm": 2.8630549907684326,
    "timestamp": "2025-07-14T12:34:06.658083"
  },
  {
    "epoch": 933,
    "batch_idx": 0,
    "loss": 0.1771848201751709,
    "learning_rate": 0.0005525245896810231,
    "timestep_mean": 487.25,
    "epsilon_norm": 7.46608829498291,
    "prediction_norm": 7.184635639190674,
    "timestamp": "2025-07-14T12:34:06.663995"
  },
  {
    "epoch": 933,
    "batch_idx": 32,
    "loss": 0.3814842104911804,
    "learning_rate": 0.0005525245896810231,
    "timestep_mean": 449.0625,
    "epsilon_norm": 8.403706550598145,
    "prediction_norm": 6.749958515167236,
    "timestamp": "2025-07-14T12:34:06.669650"
  },
  {
    "epoch": 933,
    "batch_idx": 64,
    "loss": 0.30257460474967957,
    "learning_rate": 0.0005525245896810231,
    "timestep_mean": 513.875,
    "epsilon_norm": 8.415634155273438,
    "prediction_norm": 6.704404354095459,
    "timestamp": "2025-07-14T12:34:06.675085"
  },
  {
    "epoch": 933,
    "batch_idx": 96,
    "loss": 0.4006531238555908,
    "learning_rate": 0.0005525245896810231,
    "timestep_mean": 450.1875,
    "epsilon_norm": 8.28933334350586,
    "prediction_norm": 6.0009236335754395,
    "timestamp": "2025-07-14T12:34:06.680317"
  },
  {
    "epoch": 933,
    "batch_idx": 128,
    "loss": 0.397869735956192,
    "learning_rate": 0.0005525245896810231,
    "timestep_mean": 434.96875,
    "epsilon_norm": 7.431906700134277,
    "prediction_norm": 5.582817554473877,
    "timestamp": "2025-07-14T12:34:06.685633"
  },
  {
    "epoch": 933,
    "batch_idx": 160,
    "loss": 0.26825815439224243,
    "learning_rate": 0.0005525245896810231,
    "timestep_mean": 469.28125,
    "epsilon_norm": 9.155051231384277,
    "prediction_norm": 7.530548095703125,
    "timestamp": "2025-07-14T12:34:06.690819"
  },
  {
    "epoch": 933,
    "batch_idx": 192,
    "loss": 0.1114637553691864,
    "learning_rate": 0.0005525245896810231,
    "timestep_mean": 443.0,
    "epsilon_norm": 2.9184317588806152,
    "prediction_norm": 2.9424569606781006,
    "timestamp": "2025-07-14T12:34:06.696050"
  },
  {
    "epoch": 934,
    "batch_idx": 0,
    "loss": 0.3160049617290497,
    "learning_rate": 0.0005517434726252118,
    "timestep_mean": 457.40625,
    "epsilon_norm": 7.9764933586120605,
    "prediction_norm": 7.112143039703369,
    "timestamp": "2025-07-14T12:34:06.701372"
  },
  {
    "epoch": 934,
    "batch_idx": 32,
    "loss": 0.444975346326828,
    "learning_rate": 0.0005517434726252118,
    "timestep_mean": 479.03125,
    "epsilon_norm": 9.221399307250977,
    "prediction_norm": 7.50640869140625,
    "timestamp": "2025-07-14T12:34:06.706562"
  },
  {
    "epoch": 934,
    "batch_idx": 64,
    "loss": 0.28734827041625977,
    "learning_rate": 0.0005517434726252118,
    "timestep_mean": 565.21875,
    "epsilon_norm": 7.411677837371826,
    "prediction_norm": 6.9472761154174805,
    "timestamp": "2025-07-14T12:34:06.711805"
  },
  {
    "epoch": 934,
    "batch_idx": 96,
    "loss": 0.3093435764312744,
    "learning_rate": 0.0005517434726252118,
    "timestep_mean": 527.4375,
    "epsilon_norm": 7.94158935546875,
    "prediction_norm": 7.03007698059082,
    "timestamp": "2025-07-14T12:34:06.717153"
  },
  {
    "epoch": 934,
    "batch_idx": 128,
    "loss": 0.19851060211658478,
    "learning_rate": 0.0005517434726252118,
    "timestep_mean": 551.4375,
    "epsilon_norm": 7.224325656890869,
    "prediction_norm": 6.656678199768066,
    "timestamp": "2025-07-14T12:34:06.722280"
  },
  {
    "epoch": 934,
    "batch_idx": 160,
    "loss": 0.14161089062690735,
    "learning_rate": 0.0005517434726252118,
    "timestep_mean": 505.53125,
    "epsilon_norm": 7.1626434326171875,
    "prediction_norm": 7.171888828277588,
    "timestamp": "2025-07-14T12:34:06.727476"
  },
  {
    "epoch": 934,
    "batch_idx": 192,
    "loss": 0.055798597633838654,
    "learning_rate": 0.0005517434726252118,
    "timestep_mean": 615.875,
    "epsilon_norm": 4.50560188293457,
    "prediction_norm": 4.2831830978393555,
    "timestamp": "2025-07-14T12:34:06.732677"
  },
  {
    "epoch": 935,
    "batch_idx": 0,
    "loss": 0.29641100764274597,
    "learning_rate": 0.0005509622278975255,
    "timestep_mean": 527.1875,
    "epsilon_norm": 7.547155380249023,
    "prediction_norm": 5.8919997215271,
    "timestamp": "2025-07-14T12:34:06.737955"
  },
  {
    "epoch": 935,
    "batch_idx": 32,
    "loss": 0.14364179968833923,
    "learning_rate": 0.0005509622278975255,
    "timestep_mean": 462.21875,
    "epsilon_norm": 7.954208850860596,
    "prediction_norm": 7.952247619628906,
    "timestamp": "2025-07-14T12:34:06.743042"
  },
  {
    "epoch": 935,
    "batch_idx": 64,
    "loss": 0.18090663850307465,
    "learning_rate": 0.0005509622278975255,
    "timestep_mean": 454.15625,
    "epsilon_norm": 7.118263244628906,
    "prediction_norm": 6.372190475463867,
    "timestamp": "2025-07-14T12:34:06.748247"
  },
  {
    "epoch": 935,
    "batch_idx": 96,
    "loss": 0.2491312325000763,
    "learning_rate": 0.0005509622278975255,
    "timestep_mean": 458.46875,
    "epsilon_norm": 6.6133832931518555,
    "prediction_norm": 5.018425941467285,
    "timestamp": "2025-07-14T12:34:06.753528"
  },
  {
    "epoch": 935,
    "batch_idx": 128,
    "loss": 0.14203155040740967,
    "learning_rate": 0.0005509622278975255,
    "timestep_mean": 547.21875,
    "epsilon_norm": 8.202860832214355,
    "prediction_norm": 6.996027946472168,
    "timestamp": "2025-07-14T12:34:06.758794"
  },
  {
    "epoch": 935,
    "batch_idx": 160,
    "loss": 0.1897706687450409,
    "learning_rate": 0.0005509622278975255,
    "timestep_mean": 487.15625,
    "epsilon_norm": 7.027998447418213,
    "prediction_norm": 5.547072887420654,
    "timestamp": "2025-07-14T12:34:06.763936"
  },
  {
    "epoch": 935,
    "batch_idx": 192,
    "loss": 0.15939314663410187,
    "learning_rate": 0.0005509622278975255,
    "timestep_mean": 477.125,
    "epsilon_norm": 3.677402973175049,
    "prediction_norm": 3.5913219451904297,
    "timestamp": "2025-07-14T12:34:06.769069"
  },
  {
    "epoch": 936,
    "batch_idx": 0,
    "loss": 0.22242294251918793,
    "learning_rate": 0.0005501808574256079,
    "timestep_mean": 507.65625,
    "epsilon_norm": 9.26809024810791,
    "prediction_norm": 7.862082004547119,
    "timestamp": "2025-07-14T12:34:06.774250"
  },
  {
    "epoch": 936,
    "batch_idx": 32,
    "loss": 0.5519223213195801,
    "learning_rate": 0.0005501808574256079,
    "timestep_mean": 489.09375,
    "epsilon_norm": 9.184576034545898,
    "prediction_norm": 7.223507881164551,
    "timestamp": "2025-07-14T12:34:06.779881"
  },
  {
    "epoch": 936,
    "batch_idx": 64,
    "loss": 0.2717008590698242,
    "learning_rate": 0.0005501808574256079,
    "timestep_mean": 537.125,
    "epsilon_norm": 7.925511837005615,
    "prediction_norm": 7.27554178237915,
    "timestamp": "2025-07-14T12:34:06.785094"
  },
  {
    "epoch": 936,
    "batch_idx": 96,
    "loss": 0.28386154770851135,
    "learning_rate": 0.0005501808574256079,
    "timestep_mean": 438.40625,
    "epsilon_norm": 7.622442245483398,
    "prediction_norm": 5.906017303466797,
    "timestamp": "2025-07-14T12:34:06.790480"
  },
  {
    "epoch": 936,
    "batch_idx": 128,
    "loss": 0.3833518922328949,
    "learning_rate": 0.0005501808574256079,
    "timestep_mean": 484.8125,
    "epsilon_norm": 9.032638549804688,
    "prediction_norm": 7.634745121002197,
    "timestamp": "2025-07-14T12:34:06.795783"
  },
  {
    "epoch": 936,
    "batch_idx": 160,
    "loss": 0.16602128744125366,
    "learning_rate": 0.0005501808574256079,
    "timestep_mean": 465.4375,
    "epsilon_norm": 7.780439853668213,
    "prediction_norm": 7.430562973022461,
    "timestamp": "2025-07-14T12:34:06.801032"
  },
  {
    "epoch": 936,
    "batch_idx": 192,
    "loss": 0.489158034324646,
    "learning_rate": 0.0005501808574256079,
    "timestep_mean": 568.625,
    "epsilon_norm": 3.1202523708343506,
    "prediction_norm": 1.9126780033111572,
    "timestamp": "2025-07-14T12:34:06.806145"
  },
  {
    "epoch": 937,
    "batch_idx": 0,
    "loss": 0.26372578740119934,
    "learning_rate": 0.000549399363137413,
    "timestep_mean": 489.21875,
    "epsilon_norm": 7.2444748878479,
    "prediction_norm": 6.41613245010376,
    "timestamp": "2025-07-14T12:34:06.811221"
  },
  {
    "epoch": 937,
    "batch_idx": 32,
    "loss": 0.1964660882949829,
    "learning_rate": 0.000549399363137413,
    "timestep_mean": 537.75,
    "epsilon_norm": 8.344132423400879,
    "prediction_norm": 7.020240783691406,
    "timestamp": "2025-07-14T12:34:06.816327"
  },
  {
    "epoch": 937,
    "batch_idx": 64,
    "loss": 0.19462893903255463,
    "learning_rate": 0.000549399363137413,
    "timestep_mean": 537.09375,
    "epsilon_norm": 6.98563814163208,
    "prediction_norm": 6.334835529327393,
    "timestamp": "2025-07-14T12:34:06.821668"
  },
  {
    "epoch": 937,
    "batch_idx": 96,
    "loss": 0.1638013869524002,
    "learning_rate": 0.000549399363137413,
    "timestep_mean": 510.8125,
    "epsilon_norm": 8.282445907592773,
    "prediction_norm": 7.566920757293701,
    "timestamp": "2025-07-14T12:34:06.826819"
  },
  {
    "epoch": 937,
    "batch_idx": 128,
    "loss": 0.43943268060684204,
    "learning_rate": 0.000549399363137413,
    "timestep_mean": 457.53125,
    "epsilon_norm": 8.73629093170166,
    "prediction_norm": 6.901071071624756,
    "timestamp": "2025-07-14T12:34:06.831994"
  },
  {
    "epoch": 937,
    "batch_idx": 160,
    "loss": 0.19205817580223083,
    "learning_rate": 0.000549399363137413,
    "timestep_mean": 551.03125,
    "epsilon_norm": 7.828515529632568,
    "prediction_norm": 6.61733865737915,
    "timestamp": "2025-07-14T12:34:06.837230"
  },
  {
    "epoch": 937,
    "batch_idx": 192,
    "loss": 0.15357695519924164,
    "learning_rate": 0.000549399363137413,
    "timestep_mean": 375.75,
    "epsilon_norm": 3.641202449798584,
    "prediction_norm": 2.9396164417266846,
    "timestamp": "2025-07-14T12:34:06.842444"
  },
  {
    "epoch": 938,
    "batch_idx": 0,
    "loss": 0.2819403111934662,
    "learning_rate": 0.0005486177469612003,
    "timestep_mean": 542.03125,
    "epsilon_norm": 7.292029857635498,
    "prediction_norm": 5.499185562133789,
    "timestamp": "2025-07-14T12:34:06.847628"
  },
  {
    "epoch": 938,
    "batch_idx": 32,
    "loss": 0.38027533888816833,
    "learning_rate": 0.0005486177469612003,
    "timestep_mean": 575.1875,
    "epsilon_norm": 8.329355239868164,
    "prediction_norm": 6.488970756530762,
    "timestamp": "2025-07-14T12:34:06.852796"
  },
  {
    "epoch": 938,
    "batch_idx": 64,
    "loss": 0.11690520495176315,
    "learning_rate": 0.0005486177469612003,
    "timestep_mean": 608.78125,
    "epsilon_norm": 8.525936126708984,
    "prediction_norm": 7.929879188537598,
    "timestamp": "2025-07-14T12:34:06.857988"
  },
  {
    "epoch": 938,
    "batch_idx": 96,
    "loss": 0.19276034832000732,
    "learning_rate": 0.0005486177469612003,
    "timestep_mean": 502.625,
    "epsilon_norm": 7.489091396331787,
    "prediction_norm": 6.320095062255859,
    "timestamp": "2025-07-14T12:34:06.863198"
  },
  {
    "epoch": 938,
    "batch_idx": 128,
    "loss": 0.3517270088195801,
    "learning_rate": 0.0005486177469612003,
    "timestep_mean": 539.0,
    "epsilon_norm": 7.818202972412109,
    "prediction_norm": 6.256689548492432,
    "timestamp": "2025-07-14T12:34:06.868421"
  },
  {
    "epoch": 938,
    "batch_idx": 160,
    "loss": 0.31719934940338135,
    "learning_rate": 0.0005486177469612003,
    "timestep_mean": 454.46875,
    "epsilon_norm": 8.979413986206055,
    "prediction_norm": 7.588314533233643,
    "timestamp": "2025-07-14T12:34:06.873545"
  },
  {
    "epoch": 938,
    "batch_idx": 192,
    "loss": 0.201608806848526,
    "learning_rate": 0.0005486177469612003,
    "timestep_mean": 537.875,
    "epsilon_norm": 3.514740228652954,
    "prediction_norm": 2.5947306156158447,
    "timestamp": "2025-07-14T12:34:06.878555"
  },
  {
    "epoch": 939,
    "batch_idx": 0,
    "loss": 0.3310447931289673,
    "learning_rate": 0.0005478360108255297,
    "timestep_mean": 516.75,
    "epsilon_norm": 9.662092208862305,
    "prediction_norm": 8.072076797485352,
    "timestamp": "2025-07-14T12:34:06.883732"
  },
  {
    "epoch": 939,
    "batch_idx": 32,
    "loss": 0.19236208498477936,
    "learning_rate": 0.0005478360108255297,
    "timestep_mean": 540.6875,
    "epsilon_norm": 7.553493499755859,
    "prediction_norm": 7.018889904022217,
    "timestamp": "2025-07-14T12:34:06.888712"
  },
  {
    "epoch": 939,
    "batch_idx": 64,
    "loss": 0.5586957931518555,
    "learning_rate": 0.0005478360108255297,
    "timestep_mean": 430.65625,
    "epsilon_norm": 9.855294227600098,
    "prediction_norm": 8.11665153503418,
    "timestamp": "2025-07-14T12:34:06.894020"
  },
  {
    "epoch": 939,
    "batch_idx": 96,
    "loss": 0.19777721166610718,
    "learning_rate": 0.0005478360108255297,
    "timestep_mean": 418.5,
    "epsilon_norm": 6.207852840423584,
    "prediction_norm": 6.052048206329346,
    "timestamp": "2025-07-14T12:34:06.899185"
  },
  {
    "epoch": 939,
    "batch_idx": 128,
    "loss": 0.29602980613708496,
    "learning_rate": 0.0005478360108255297,
    "timestep_mean": 464.65625,
    "epsilon_norm": 7.1158647537231445,
    "prediction_norm": 6.2190752029418945,
    "timestamp": "2025-07-14T12:34:06.904420"
  },
  {
    "epoch": 939,
    "batch_idx": 160,
    "loss": 0.46248459815979004,
    "learning_rate": 0.0005478360108255297,
    "timestep_mean": 464.78125,
    "epsilon_norm": 8.263456344604492,
    "prediction_norm": 7.128347873687744,
    "timestamp": "2025-07-14T12:34:06.909530"
  },
  {
    "epoch": 939,
    "batch_idx": 192,
    "loss": 0.12252603471279144,
    "learning_rate": 0.0005478360108255297,
    "timestep_mean": 479.25,
    "epsilon_norm": 5.20839262008667,
    "prediction_norm": 5.390942573547363,
    "timestamp": "2025-07-14T12:34:06.914392"
  },
  {
    "epoch": 940,
    "batch_idx": 0,
    "loss": 0.3031066656112671,
    "learning_rate": 0.0005470541566592577,
    "timestep_mean": 501.15625,
    "epsilon_norm": 8.596632957458496,
    "prediction_norm": 8.402116775512695,
    "timestamp": "2025-07-14T12:34:06.919272"
  },
  {
    "epoch": 940,
    "batch_idx": 32,
    "loss": 0.13187313079833984,
    "learning_rate": 0.0005470541566592577,
    "timestep_mean": 490.625,
    "epsilon_norm": 8.16006851196289,
    "prediction_norm": 7.813781261444092,
    "timestamp": "2025-07-14T12:34:06.924300"
  },
  {
    "epoch": 940,
    "batch_idx": 64,
    "loss": 0.36483991146087646,
    "learning_rate": 0.0005470541566592577,
    "timestep_mean": 517.90625,
    "epsilon_norm": 9.01109504699707,
    "prediction_norm": 7.827602386474609,
    "timestamp": "2025-07-14T12:34:06.929712"
  },
  {
    "epoch": 940,
    "batch_idx": 96,
    "loss": 0.22244691848754883,
    "learning_rate": 0.0005470541566592577,
    "timestep_mean": 620.40625,
    "epsilon_norm": 9.020735740661621,
    "prediction_norm": 7.485477447509766,
    "timestamp": "2025-07-14T12:34:06.934877"
  },
  {
    "epoch": 940,
    "batch_idx": 128,
    "loss": 0.3123026490211487,
    "learning_rate": 0.0005470541566592577,
    "timestep_mean": 471.15625,
    "epsilon_norm": 9.119843482971191,
    "prediction_norm": 7.94367790222168,
    "timestamp": "2025-07-14T12:34:06.940161"
  },
  {
    "epoch": 940,
    "batch_idx": 160,
    "loss": 0.19777712225914001,
    "learning_rate": 0.0005470541566592577,
    "timestep_mean": 526.46875,
    "epsilon_norm": 6.586458206176758,
    "prediction_norm": 5.528924465179443,
    "timestamp": "2025-07-14T12:34:06.945172"
  },
  {
    "epoch": 940,
    "batch_idx": 192,
    "loss": 0.291659414768219,
    "learning_rate": 0.0005470541566592577,
    "timestep_mean": 419.625,
    "epsilon_norm": 3.9972784519195557,
    "prediction_norm": 3.2312426567077637,
    "timestamp": "2025-07-14T12:34:06.950212"
  },
  {
    "epoch": 941,
    "batch_idx": 0,
    "loss": 0.3403697609901428,
    "learning_rate": 0.0005462721863915316,
    "timestep_mean": 602.9375,
    "epsilon_norm": 7.347141265869141,
    "prediction_norm": 5.934781074523926,
    "timestamp": "2025-07-14T12:34:06.955088"
  },
  {
    "epoch": 941,
    "batch_idx": 32,
    "loss": 0.2870853543281555,
    "learning_rate": 0.0005462721863915316,
    "timestep_mean": 520.34375,
    "epsilon_norm": 7.236018180847168,
    "prediction_norm": 5.592166423797607,
    "timestamp": "2025-07-14T12:34:06.959918"
  },
  {
    "epoch": 941,
    "batch_idx": 64,
    "loss": 0.17894187569618225,
    "learning_rate": 0.0005462721863915316,
    "timestep_mean": 502.75,
    "epsilon_norm": 6.716403007507324,
    "prediction_norm": 5.65366792678833,
    "timestamp": "2025-07-14T12:34:06.964756"
  },
  {
    "epoch": 941,
    "batch_idx": 96,
    "loss": 0.32308945059776306,
    "learning_rate": 0.0005462721863915316,
    "timestep_mean": 502.1875,
    "epsilon_norm": 6.943204402923584,
    "prediction_norm": 5.472358226776123,
    "timestamp": "2025-07-14T12:34:06.969642"
  },
  {
    "epoch": 941,
    "batch_idx": 128,
    "loss": 0.5137834548950195,
    "learning_rate": 0.0005462721863915316,
    "timestep_mean": 472.09375,
    "epsilon_norm": 8.097888946533203,
    "prediction_norm": 5.562666893005371,
    "timestamp": "2025-07-14T12:34:06.974514"
  },
  {
    "epoch": 941,
    "batch_idx": 160,
    "loss": 0.19338002800941467,
    "learning_rate": 0.0005462721863915316,
    "timestep_mean": 485.8125,
    "epsilon_norm": 7.753472805023193,
    "prediction_norm": 6.885149955749512,
    "timestamp": "2025-07-14T12:34:06.979496"
  },
  {
    "epoch": 941,
    "batch_idx": 192,
    "loss": 0.30431807041168213,
    "learning_rate": 0.0005462721863915316,
    "timestep_mean": 514.25,
    "epsilon_norm": 3.7410852909088135,
    "prediction_norm": 2.5845766067504883,
    "timestamp": "2025-07-14T12:34:06.984363"
  },
  {
    "epoch": 942,
    "batch_idx": 0,
    "loss": 0.3277369737625122,
    "learning_rate": 0.0005454901019517854,
    "timestep_mean": 521.1875,
    "epsilon_norm": 8.713204383850098,
    "prediction_norm": 6.906008243560791,
    "timestamp": "2025-07-14T12:34:06.989215"
  },
  {
    "epoch": 942,
    "batch_idx": 32,
    "loss": 0.39382925629615784,
    "learning_rate": 0.0005454901019517854,
    "timestep_mean": 561.53125,
    "epsilon_norm": 8.77821159362793,
    "prediction_norm": 6.98342752456665,
    "timestamp": "2025-07-14T12:34:06.994058"
  },
  {
    "epoch": 942,
    "batch_idx": 64,
    "loss": 0.48955869674682617,
    "learning_rate": 0.0005454901019517854,
    "timestep_mean": 439.875,
    "epsilon_norm": 9.213505744934082,
    "prediction_norm": 7.753682613372803,
    "timestamp": "2025-07-14T12:34:06.998943"
  },
  {
    "epoch": 942,
    "batch_idx": 96,
    "loss": 0.32247674465179443,
    "learning_rate": 0.0005454901019517854,
    "timestep_mean": 540.28125,
    "epsilon_norm": 8.840620040893555,
    "prediction_norm": 6.610130310058594,
    "timestamp": "2025-07-14T12:34:07.003787"
  },
  {
    "epoch": 942,
    "batch_idx": 128,
    "loss": 0.24921104311943054,
    "learning_rate": 0.0005454901019517854,
    "timestep_mean": 492.5625,
    "epsilon_norm": 7.778662204742432,
    "prediction_norm": 6.532114505767822,
    "timestamp": "2025-07-14T12:34:07.008700"
  },
  {
    "epoch": 942,
    "batch_idx": 160,
    "loss": 0.24285191297531128,
    "learning_rate": 0.0005454901019517854,
    "timestep_mean": 448.5625,
    "epsilon_norm": 6.627583026885986,
    "prediction_norm": 5.256874084472656,
    "timestamp": "2025-07-14T12:34:07.013504"
  },
  {
    "epoch": 942,
    "batch_idx": 192,
    "loss": 0.13350754976272583,
    "learning_rate": 0.0005454901019517854,
    "timestep_mean": 489.0,
    "epsilon_norm": 3.6786599159240723,
    "prediction_norm": 3.5854735374450684,
    "timestamp": "2025-07-14T12:34:07.018510"
  },
  {
    "epoch": 943,
    "batch_idx": 0,
    "loss": 0.2080834060907364,
    "learning_rate": 0.0005447079052697347,
    "timestep_mean": 504.8125,
    "epsilon_norm": 7.9685587882995605,
    "prediction_norm": 7.420351982116699,
    "timestamp": "2025-07-14T12:34:07.023360"
  },
  {
    "epoch": 943,
    "batch_idx": 32,
    "loss": 0.4264081120491028,
    "learning_rate": 0.0005447079052697347,
    "timestep_mean": 475.875,
    "epsilon_norm": 7.49036979675293,
    "prediction_norm": 5.472720146179199,
    "timestamp": "2025-07-14T12:34:07.028334"
  },
  {
    "epoch": 943,
    "batch_idx": 64,
    "loss": 0.2421567738056183,
    "learning_rate": 0.0005447079052697347,
    "timestep_mean": 443.84375,
    "epsilon_norm": 7.856415748596191,
    "prediction_norm": 6.44645881652832,
    "timestamp": "2025-07-14T12:34:07.033205"
  },
  {
    "epoch": 943,
    "batch_idx": 96,
    "loss": 0.1661367267370224,
    "learning_rate": 0.0005447079052697347,
    "timestep_mean": 571.25,
    "epsilon_norm": 7.849491596221924,
    "prediction_norm": 6.8979878425598145,
    "timestamp": "2025-07-14T12:34:07.038210"
  },
  {
    "epoch": 943,
    "batch_idx": 128,
    "loss": 0.3215758502483368,
    "learning_rate": 0.0005447079052697347,
    "timestep_mean": 528.375,
    "epsilon_norm": 7.998060703277588,
    "prediction_norm": 6.591707706451416,
    "timestamp": "2025-07-14T12:34:07.043081"
  },
  {
    "epoch": 943,
    "batch_idx": 160,
    "loss": 0.2471356838941574,
    "learning_rate": 0.0005447079052697347,
    "timestep_mean": 569.03125,
    "epsilon_norm": 7.147533416748047,
    "prediction_norm": 6.795282363891602,
    "timestamp": "2025-07-14T12:34:07.047899"
  },
  {
    "epoch": 943,
    "batch_idx": 192,
    "loss": 0.30674344301223755,
    "learning_rate": 0.0005447079052697347,
    "timestep_mean": 436.25,
    "epsilon_norm": 3.0104329586029053,
    "prediction_norm": 2.312671184539795,
    "timestamp": "2025-07-14T12:34:07.052710"
  },
  {
    "epoch": 944,
    "batch_idx": 0,
    "loss": 0.33743754029273987,
    "learning_rate": 0.0005439255982753721,
    "timestep_mean": 515.375,
    "epsilon_norm": 8.208280563354492,
    "prediction_norm": 7.101860523223877,
    "timestamp": "2025-07-14T12:34:07.057631"
  },
  {
    "epoch": 944,
    "batch_idx": 32,
    "loss": 0.09675433486700058,
    "learning_rate": 0.0005439255982753721,
    "timestep_mean": 564.78125,
    "epsilon_norm": 6.79606294631958,
    "prediction_norm": 7.070586681365967,
    "timestamp": "2025-07-14T12:34:07.062618"
  },
  {
    "epoch": 944,
    "batch_idx": 64,
    "loss": 0.2891613841056824,
    "learning_rate": 0.0005439255982753721,
    "timestep_mean": 461.28125,
    "epsilon_norm": 8.102724075317383,
    "prediction_norm": 7.878774166107178,
    "timestamp": "2025-07-14T12:34:07.067433"
  },
  {
    "epoch": 944,
    "batch_idx": 96,
    "loss": 0.2408798635005951,
    "learning_rate": 0.0005439255982753721,
    "timestep_mean": 516.1875,
    "epsilon_norm": 8.101273536682129,
    "prediction_norm": 7.701127052307129,
    "timestamp": "2025-07-14T12:34:07.072260"
  },
  {
    "epoch": 944,
    "batch_idx": 128,
    "loss": 0.14458194375038147,
    "learning_rate": 0.0005439255982753721,
    "timestep_mean": 485.9375,
    "epsilon_norm": 7.873233318328857,
    "prediction_norm": 7.663607120513916,
    "timestamp": "2025-07-14T12:34:07.077119"
  },
  {
    "epoch": 944,
    "batch_idx": 160,
    "loss": 0.1910499930381775,
    "learning_rate": 0.0005439255982753721,
    "timestep_mean": 505.84375,
    "epsilon_norm": 9.761819839477539,
    "prediction_norm": 8.870530128479004,
    "timestamp": "2025-07-14T12:34:07.081917"
  },
  {
    "epoch": 944,
    "batch_idx": 192,
    "loss": 0.6608337163925171,
    "learning_rate": 0.0005439255982753721,
    "timestep_mean": 486.625,
    "epsilon_norm": 3.6224560737609863,
    "prediction_norm": 2.906132221221924,
    "timestamp": "2025-07-14T12:34:07.086847"
  },
  {
    "epoch": 945,
    "batch_idx": 0,
    "loss": 0.13462942838668823,
    "learning_rate": 0.0005431431828989622,
    "timestep_mean": 497.96875,
    "epsilon_norm": 8.20456600189209,
    "prediction_norm": 8.104490280151367,
    "timestamp": "2025-07-14T12:34:07.091777"
  },
  {
    "epoch": 945,
    "batch_idx": 32,
    "loss": 0.224931538105011,
    "learning_rate": 0.0005431431828989622,
    "timestep_mean": 525.1875,
    "epsilon_norm": 8.39923095703125,
    "prediction_norm": 7.492537498474121,
    "timestamp": "2025-07-14T12:34:07.096622"
  },
  {
    "epoch": 945,
    "batch_idx": 64,
    "loss": 0.23584617674350739,
    "learning_rate": 0.0005431431828989622,
    "timestep_mean": 448.78125,
    "epsilon_norm": 7.271334648132324,
    "prediction_norm": 6.031550407409668,
    "timestamp": "2025-07-14T12:34:07.101470"
  },
  {
    "epoch": 945,
    "batch_idx": 96,
    "loss": 0.2291092574596405,
    "learning_rate": 0.0005431431828989622,
    "timestep_mean": 528.46875,
    "epsilon_norm": 7.197718620300293,
    "prediction_norm": 5.644815444946289,
    "timestamp": "2025-07-14T12:34:07.106413"
  },
  {
    "epoch": 945,
    "batch_idx": 128,
    "loss": 0.17061910033226013,
    "learning_rate": 0.0005431431828989622,
    "timestep_mean": 562.4375,
    "epsilon_norm": 7.9545063972473145,
    "prediction_norm": 6.664092063903809,
    "timestamp": "2025-07-14T12:34:07.111253"
  },
  {
    "epoch": 945,
    "batch_idx": 160,
    "loss": 0.31865572929382324,
    "learning_rate": 0.0005431431828989622,
    "timestep_mean": 502.40625,
    "epsilon_norm": 8.865102767944336,
    "prediction_norm": 7.595986843109131,
    "timestamp": "2025-07-14T12:34:07.116076"
  },
  {
    "epoch": 945,
    "batch_idx": 192,
    "loss": 0.3102315664291382,
    "learning_rate": 0.0005431431828989622,
    "timestep_mean": 585.875,
    "epsilon_norm": 4.912391185760498,
    "prediction_norm": 3.8390603065490723,
    "timestamp": "2025-07-14T12:34:07.120993"
  },
  {
    "epoch": 946,
    "batch_idx": 0,
    "loss": 0.13572612404823303,
    "learning_rate": 0.000542360661071037,
    "timestep_mean": 502.8125,
    "epsilon_norm": 7.795505523681641,
    "prediction_norm": 7.1098809242248535,
    "timestamp": "2025-07-14T12:34:07.125850"
  },
  {
    "epoch": 946,
    "batch_idx": 32,
    "loss": 0.2900533080101013,
    "learning_rate": 0.000542360661071037,
    "timestep_mean": 467.4375,
    "epsilon_norm": 7.461601257324219,
    "prediction_norm": 6.189446449279785,
    "timestamp": "2025-07-14T12:34:07.130744"
  },
  {
    "epoch": 946,
    "batch_idx": 64,
    "loss": 0.3442752957344055,
    "learning_rate": 0.000542360661071037,
    "timestep_mean": 517.1875,
    "epsilon_norm": 8.737285614013672,
    "prediction_norm": 6.582883358001709,
    "timestamp": "2025-07-14T12:34:07.135781"
  },
  {
    "epoch": 946,
    "batch_idx": 96,
    "loss": 0.33033108711242676,
    "learning_rate": 0.000542360661071037,
    "timestep_mean": 487.09375,
    "epsilon_norm": 8.449345588684082,
    "prediction_norm": 6.9182820320129395,
    "timestamp": "2025-07-14T12:34:07.140911"
  },
  {
    "epoch": 946,
    "batch_idx": 128,
    "loss": 0.2485661804676056,
    "learning_rate": 0.000542360661071037,
    "timestep_mean": 524.625,
    "epsilon_norm": 7.891965866088867,
    "prediction_norm": 6.862525939941406,
    "timestamp": "2025-07-14T12:34:07.146394"
  },
  {
    "epoch": 946,
    "batch_idx": 160,
    "loss": 0.2255316823720932,
    "learning_rate": 0.000542360661071037,
    "timestep_mean": 520.90625,
    "epsilon_norm": 6.953328609466553,
    "prediction_norm": 5.7093610763549805,
    "timestamp": "2025-07-14T12:34:07.152046"
  },
  {
    "epoch": 946,
    "batch_idx": 192,
    "loss": 0.17152319848537445,
    "learning_rate": 0.000542360661071037,
    "timestep_mean": 564.375,
    "epsilon_norm": 3.706622838973999,
    "prediction_norm": 3.4694950580596924,
    "timestamp": "2025-07-14T12:34:07.157131"
  },
  {
    "epoch": 947,
    "batch_idx": 0,
    "loss": 0.1461954414844513,
    "learning_rate": 0.0005415780347223919,
    "timestep_mean": 573.75,
    "epsilon_norm": 7.994128227233887,
    "prediction_norm": 7.559635639190674,
    "timestamp": "2025-07-14T12:34:07.162245"
  },
  {
    "epoch": 947,
    "batch_idx": 32,
    "loss": 0.29668575525283813,
    "learning_rate": 0.0005415780347223919,
    "timestep_mean": 441.875,
    "epsilon_norm": 7.624240398406982,
    "prediction_norm": 6.655593395233154,
    "timestamp": "2025-07-14T12:34:07.167871"
  },
  {
    "epoch": 947,
    "batch_idx": 64,
    "loss": 0.19246360659599304,
    "learning_rate": 0.0005415780347223919,
    "timestep_mean": 520.96875,
    "epsilon_norm": 7.997564792633057,
    "prediction_norm": 6.8283209800720215,
    "timestamp": "2025-07-14T12:34:07.172801"
  },
  {
    "epoch": 947,
    "batch_idx": 96,
    "loss": 0.13958656787872314,
    "learning_rate": 0.0005415780347223919,
    "timestep_mean": 569.59375,
    "epsilon_norm": 6.622248649597168,
    "prediction_norm": 5.854562282562256,
    "timestamp": "2025-07-14T12:34:07.178417"
  },
  {
    "epoch": 947,
    "batch_idx": 128,
    "loss": 0.33322256803512573,
    "learning_rate": 0.0005415780347223919,
    "timestep_mean": 425.875,
    "epsilon_norm": 8.513853073120117,
    "prediction_norm": 7.047414302825928,
    "timestamp": "2025-07-14T12:34:07.183735"
  },
  {
    "epoch": 947,
    "batch_idx": 160,
    "loss": 0.30077046155929565,
    "learning_rate": 0.0005415780347223919,
    "timestep_mean": 485.1875,
    "epsilon_norm": 8.443229675292969,
    "prediction_norm": 7.258393287658691,
    "timestamp": "2025-07-14T12:34:07.188984"
  },
  {
    "epoch": 947,
    "batch_idx": 192,
    "loss": 0.18567678332328796,
    "learning_rate": 0.0005415780347223919,
    "timestep_mean": 593.625,
    "epsilon_norm": 3.8950355052948,
    "prediction_norm": 3.548743963241577,
    "timestamp": "2025-07-14T12:34:07.194483"
  },
  {
    "epoch": 948,
    "batch_idx": 0,
    "loss": 0.2716471552848816,
    "learning_rate": 0.0005407953057840792,
    "timestep_mean": 525.4375,
    "epsilon_norm": 7.411128044128418,
    "prediction_norm": 6.113187789916992,
    "timestamp": "2025-07-14T12:34:07.199843"
  },
  {
    "epoch": 948,
    "batch_idx": 32,
    "loss": 0.15108950436115265,
    "learning_rate": 0.0005407953057840792,
    "timestep_mean": 532.625,
    "epsilon_norm": 8.343036651611328,
    "prediction_norm": 7.215036392211914,
    "timestamp": "2025-07-14T12:34:07.205227"
  },
  {
    "epoch": 948,
    "batch_idx": 64,
    "loss": 0.17575183510780334,
    "learning_rate": 0.0005407953057840792,
    "timestep_mean": 509.0,
    "epsilon_norm": 7.246243953704834,
    "prediction_norm": 6.65705680847168,
    "timestamp": "2025-07-14T12:34:07.210672"
  },
  {
    "epoch": 948,
    "batch_idx": 96,
    "loss": 0.161224365234375,
    "learning_rate": 0.0005407953057840792,
    "timestep_mean": 467.0,
    "epsilon_norm": 8.022314071655273,
    "prediction_norm": 7.620362758636475,
    "timestamp": "2025-07-14T12:34:07.215780"
  },
  {
    "epoch": 948,
    "batch_idx": 128,
    "loss": 0.4213965833187103,
    "learning_rate": 0.0005407953057840792,
    "timestep_mean": 460.78125,
    "epsilon_norm": 8.659411430358887,
    "prediction_norm": 7.085179805755615,
    "timestamp": "2025-07-14T12:34:07.220976"
  },
  {
    "epoch": 948,
    "batch_idx": 160,
    "loss": 0.4544864892959595,
    "learning_rate": 0.0005407953057840792,
    "timestep_mean": 451.5625,
    "epsilon_norm": 8.886228561401367,
    "prediction_norm": 6.895008087158203,
    "timestamp": "2025-07-14T12:34:07.226236"
  },
  {
    "epoch": 948,
    "batch_idx": 192,
    "loss": 0.48177507519721985,
    "learning_rate": 0.0005407953057840792,
    "timestep_mean": 454.5,
    "epsilon_norm": 4.193826675415039,
    "prediction_norm": 3.565934181213379,
    "timestamp": "2025-07-14T12:34:07.231600"
  },
  {
    "epoch": 949,
    "batch_idx": 0,
    "loss": 0.1998276710510254,
    "learning_rate": 0.0005400124761874048,
    "timestep_mean": 523.9375,
    "epsilon_norm": 8.69127082824707,
    "prediction_norm": 7.574283123016357,
    "timestamp": "2025-07-14T12:34:07.236925"
  },
  {
    "epoch": 949,
    "batch_idx": 32,
    "loss": 0.21638235449790955,
    "learning_rate": 0.0005400124761874048,
    "timestep_mean": 498.53125,
    "epsilon_norm": 8.934892654418945,
    "prediction_norm": 7.859467029571533,
    "timestamp": "2025-07-14T12:34:07.242245"
  },
  {
    "epoch": 949,
    "batch_idx": 64,
    "loss": 0.37034374475479126,
    "learning_rate": 0.0005400124761874048,
    "timestep_mean": 449.1875,
    "epsilon_norm": 7.7750067710876465,
    "prediction_norm": 6.359381198883057,
    "timestamp": "2025-07-14T12:34:07.247506"
  },
  {
    "epoch": 949,
    "batch_idx": 96,
    "loss": 0.28637251257896423,
    "learning_rate": 0.0005400124761874048,
    "timestep_mean": 543.53125,
    "epsilon_norm": 8.238168716430664,
    "prediction_norm": 7.244174480438232,
    "timestamp": "2025-07-14T12:34:07.252812"
  },
  {
    "epoch": 949,
    "batch_idx": 128,
    "loss": 0.6244111061096191,
    "learning_rate": 0.0005400124761874048,
    "timestep_mean": 375.3125,
    "epsilon_norm": 8.247812271118164,
    "prediction_norm": 5.649768829345703,
    "timestamp": "2025-07-14T12:34:07.257985"
  },
  {
    "epoch": 949,
    "batch_idx": 160,
    "loss": 0.430527001619339,
    "learning_rate": 0.0005400124761874048,
    "timestep_mean": 431.09375,
    "epsilon_norm": 7.888261795043945,
    "prediction_norm": 6.854316234588623,
    "timestamp": "2025-07-14T12:34:07.263429"
  },
  {
    "epoch": 949,
    "batch_idx": 192,
    "loss": 0.226370170712471,
    "learning_rate": 0.0005400124761874048,
    "timestep_mean": 455.0,
    "epsilon_norm": 4.148588180541992,
    "prediction_norm": 3.504805326461792,
    "timestamp": "2025-07-14T12:34:07.268758"
  },
  {
    "epoch": 950,
    "batch_idx": 0,
    "loss": 0.2565370500087738,
    "learning_rate": 0.0005392295478639229,
    "timestep_mean": 476.6875,
    "epsilon_norm": 7.872062683105469,
    "prediction_norm": 7.205108642578125,
    "timestamp": "2025-07-14T12:34:07.273938"
  },
  {
    "epoch": 950,
    "batch_idx": 32,
    "loss": 0.241816908121109,
    "learning_rate": 0.0005392295478639229,
    "timestep_mean": 519.03125,
    "epsilon_norm": 8.433588981628418,
    "prediction_norm": 7.512390613555908,
    "timestamp": "2025-07-14T12:34:07.279201"
  },
  {
    "epoch": 950,
    "batch_idx": 64,
    "loss": 0.4412909746170044,
    "learning_rate": 0.0005392295478639229,
    "timestep_mean": 556.84375,
    "epsilon_norm": 9.249794960021973,
    "prediction_norm": 8.14934253692627,
    "timestamp": "2025-07-14T12:34:07.284333"
  },
  {
    "epoch": 950,
    "batch_idx": 96,
    "loss": 0.2585129737854004,
    "learning_rate": 0.0005392295478639229,
    "timestep_mean": 489.6875,
    "epsilon_norm": 7.7551398277282715,
    "prediction_norm": 6.520829200744629,
    "timestamp": "2025-07-14T12:34:07.289495"
  },
  {
    "epoch": 950,
    "batch_idx": 128,
    "loss": 0.19367168843746185,
    "learning_rate": 0.0005392295478639229,
    "timestep_mean": 480.65625,
    "epsilon_norm": 7.978764533996582,
    "prediction_norm": 7.056180953979492,
    "timestamp": "2025-07-14T12:34:07.294704"
  },
  {
    "epoch": 950,
    "batch_idx": 160,
    "loss": 0.4277765154838562,
    "learning_rate": 0.0005392295478639229,
    "timestep_mean": 466.40625,
    "epsilon_norm": 8.000971794128418,
    "prediction_norm": 5.947851657867432,
    "timestamp": "2025-07-14T12:34:07.300038"
  },
  {
    "epoch": 950,
    "batch_idx": 192,
    "loss": 0.07197108864784241,
    "learning_rate": 0.0005392295478639229,
    "timestep_mean": 625.375,
    "epsilon_norm": 4.426655292510986,
    "prediction_norm": 4.268955230712891,
    "timestamp": "2025-07-14T12:34:07.305297"
  },
  {
    "epoch": 951,
    "batch_idx": 0,
    "loss": 0.4404660761356354,
    "learning_rate": 0.0005384465227454314,
    "timestep_mean": 437.1875,
    "epsilon_norm": 8.274117469787598,
    "prediction_norm": 5.59482479095459,
    "timestamp": "2025-07-14T12:34:07.310531"
  },
  {
    "epoch": 951,
    "batch_idx": 32,
    "loss": 0.3722180426120758,
    "learning_rate": 0.0005384465227454314,
    "timestep_mean": 496.09375,
    "epsilon_norm": 8.222428321838379,
    "prediction_norm": 7.338136672973633,
    "timestamp": "2025-07-14T12:34:07.315831"
  },
  {
    "epoch": 951,
    "batch_idx": 64,
    "loss": 0.22422346472740173,
    "learning_rate": 0.0005384465227454314,
    "timestep_mean": 503.75,
    "epsilon_norm": 7.792008399963379,
    "prediction_norm": 6.268704891204834,
    "timestamp": "2025-07-14T12:34:07.321071"
  },
  {
    "epoch": 951,
    "batch_idx": 96,
    "loss": 0.3049522042274475,
    "learning_rate": 0.0005384465227454314,
    "timestep_mean": 419.09375,
    "epsilon_norm": 7.747270584106445,
    "prediction_norm": 5.630945205688477,
    "timestamp": "2025-07-14T12:34:07.326147"
  },
  {
    "epoch": 951,
    "batch_idx": 128,
    "loss": 0.24588146805763245,
    "learning_rate": 0.0005384465227454314,
    "timestep_mean": 516.5,
    "epsilon_norm": 8.266504287719727,
    "prediction_norm": 7.28057861328125,
    "timestamp": "2025-07-14T12:34:07.331414"
  },
  {
    "epoch": 951,
    "batch_idx": 160,
    "loss": 0.3102725148200989,
    "learning_rate": 0.0005384465227454314,
    "timestep_mean": 482.71875,
    "epsilon_norm": 7.637173652648926,
    "prediction_norm": 6.302331447601318,
    "timestamp": "2025-07-14T12:34:07.336664"
  },
  {
    "epoch": 951,
    "batch_idx": 192,
    "loss": 0.8157529234886169,
    "learning_rate": 0.0005384465227454314,
    "timestep_mean": 565.125,
    "epsilon_norm": 4.706468105316162,
    "prediction_norm": 3.202451705932617,
    "timestamp": "2025-07-14T12:34:07.341751"
  },
  {
    "epoch": 952,
    "batch_idx": 0,
    "loss": 0.2557995319366455,
    "learning_rate": 0.0005376634027639668,
    "timestep_mean": 490.78125,
    "epsilon_norm": 7.4721808433532715,
    "prediction_norm": 6.627679347991943,
    "timestamp": "2025-07-14T12:34:07.346870"
  },
  {
    "epoch": 952,
    "batch_idx": 32,
    "loss": 0.2310079038143158,
    "learning_rate": 0.0005376634027639668,
    "timestep_mean": 536.15625,
    "epsilon_norm": 7.596088409423828,
    "prediction_norm": 6.349080562591553,
    "timestamp": "2025-07-14T12:34:07.352009"
  },
  {
    "epoch": 952,
    "batch_idx": 64,
    "loss": 0.19961977005004883,
    "learning_rate": 0.0005376634027639668,
    "timestep_mean": 547.5,
    "epsilon_norm": 8.345340728759766,
    "prediction_norm": 7.335912227630615,
    "timestamp": "2025-07-14T12:34:07.357236"
  },
  {
    "epoch": 952,
    "batch_idx": 96,
    "loss": 0.2724340558052063,
    "learning_rate": 0.0005376634027639668,
    "timestep_mean": 471.0,
    "epsilon_norm": 7.897302150726318,
    "prediction_norm": 6.833352088928223,
    "timestamp": "2025-07-14T12:34:07.362546"
  },
  {
    "epoch": 952,
    "batch_idx": 128,
    "loss": 0.13759419322013855,
    "learning_rate": 0.0005376634027639668,
    "timestep_mean": 546.53125,
    "epsilon_norm": 7.338756561279297,
    "prediction_norm": 6.384002685546875,
    "timestamp": "2025-07-14T12:34:07.367791"
  },
  {
    "epoch": 952,
    "batch_idx": 160,
    "loss": 0.21269260346889496,
    "learning_rate": 0.0005376634027639668,
    "timestep_mean": 416.6875,
    "epsilon_norm": 7.3314948081970215,
    "prediction_norm": 6.054944038391113,
    "timestamp": "2025-07-14T12:34:07.372946"
  },
  {
    "epoch": 952,
    "batch_idx": 192,
    "loss": 0.13515469431877136,
    "learning_rate": 0.0005376634027639668,
    "timestep_mean": 689.125,
    "epsilon_norm": 3.551886558532715,
    "prediction_norm": 3.091341972351074,
    "timestamp": "2025-07-14T12:34:07.378073"
  },
  {
    "epoch": 953,
    "batch_idx": 0,
    "loss": 0.4478682577610016,
    "learning_rate": 0.0005368801898517998,
    "timestep_mean": 382.75,
    "epsilon_norm": 8.740984916687012,
    "prediction_norm": 6.699855327606201,
    "timestamp": "2025-07-14T12:34:07.383300"
  },
  {
    "epoch": 953,
    "batch_idx": 32,
    "loss": 0.28843849897384644,
    "learning_rate": 0.0005368801898517998,
    "timestep_mean": 517.46875,
    "epsilon_norm": 6.895841598510742,
    "prediction_norm": 6.335658550262451,
    "timestamp": "2025-07-14T12:34:07.388483"
  },
  {
    "epoch": 953,
    "batch_idx": 64,
    "loss": 0.23923513293266296,
    "learning_rate": 0.0005368801898517998,
    "timestep_mean": 593.71875,
    "epsilon_norm": 9.262008666992188,
    "prediction_norm": 8.632214546203613,
    "timestamp": "2025-07-14T12:34:07.393407"
  },
  {
    "epoch": 953,
    "batch_idx": 96,
    "loss": 0.3115062415599823,
    "learning_rate": 0.0005368801898517998,
    "timestep_mean": 457.28125,
    "epsilon_norm": 8.817873001098633,
    "prediction_norm": 6.841522693634033,
    "timestamp": "2025-07-14T12:34:07.398411"
  },
  {
    "epoch": 953,
    "batch_idx": 128,
    "loss": 0.24357029795646667,
    "learning_rate": 0.0005368801898517998,
    "timestep_mean": 558.28125,
    "epsilon_norm": 7.563105583190918,
    "prediction_norm": 7.037644863128662,
    "timestamp": "2025-07-14T12:34:07.403606"
  },
  {
    "epoch": 953,
    "batch_idx": 160,
    "loss": 0.26356586813926697,
    "learning_rate": 0.0005368801898517998,
    "timestep_mean": 504.5625,
    "epsilon_norm": 7.3754096031188965,
    "prediction_norm": 6.202459335327148,
    "timestamp": "2025-07-14T12:34:07.408596"
  },
  {
    "epoch": 953,
    "batch_idx": 192,
    "loss": 1.0725352764129639,
    "learning_rate": 0.0005368801898517998,
    "timestep_mean": 284.0,
    "epsilon_norm": 4.148933410644531,
    "prediction_norm": 2.0649898052215576,
    "timestamp": "2025-07-14T12:34:07.413507"
  },
  {
    "epoch": 954,
    "batch_idx": 0,
    "loss": 0.31633391976356506,
    "learning_rate": 0.0005360968859414307,
    "timestep_mean": 551.3125,
    "epsilon_norm": 9.434661865234375,
    "prediction_norm": 8.606386184692383,
    "timestamp": "2025-07-14T12:34:07.418843"
  },
  {
    "epoch": 954,
    "batch_idx": 32,
    "loss": 0.30007439851760864,
    "learning_rate": 0.0005360968859414307,
    "timestep_mean": 500.71875,
    "epsilon_norm": 7.838673114776611,
    "prediction_norm": 6.854640483856201,
    "timestamp": "2025-07-14T12:34:07.424059"
  },
  {
    "epoch": 954,
    "batch_idx": 64,
    "loss": 0.20451979339122772,
    "learning_rate": 0.0005360968859414307,
    "timestep_mean": 589.6875,
    "epsilon_norm": 7.019104480743408,
    "prediction_norm": 6.096471786499023,
    "timestamp": "2025-07-14T12:34:07.429366"
  },
  {
    "epoch": 954,
    "batch_idx": 96,
    "loss": 0.42755094170570374,
    "learning_rate": 0.0005360968859414307,
    "timestep_mean": 435.46875,
    "epsilon_norm": 7.878380298614502,
    "prediction_norm": 6.429387092590332,
    "timestamp": "2025-07-14T12:34:07.434549"
  },
  {
    "epoch": 954,
    "batch_idx": 128,
    "loss": 0.40478092432022095,
    "learning_rate": 0.0005360968859414307,
    "timestep_mean": 475.65625,
    "epsilon_norm": 8.267065048217773,
    "prediction_norm": 6.551429748535156,
    "timestamp": "2025-07-14T12:34:07.439813"
  },
  {
    "epoch": 954,
    "batch_idx": 160,
    "loss": 0.35509350895881653,
    "learning_rate": 0.0005360968859414307,
    "timestep_mean": 542.8125,
    "epsilon_norm": 9.034769058227539,
    "prediction_norm": 7.160080432891846,
    "timestamp": "2025-07-14T12:34:07.444928"
  },
  {
    "epoch": 954,
    "batch_idx": 192,
    "loss": 0.16996973752975464,
    "learning_rate": 0.0005360968859414307,
    "timestep_mean": 281.375,
    "epsilon_norm": 3.2735304832458496,
    "prediction_norm": 2.2766921520233154,
    "timestamp": "2025-07-14T12:34:07.450171"
  },
  {
    "epoch": 955,
    "batch_idx": 0,
    "loss": 0.1178765669465065,
    "learning_rate": 0.0005353134929655838,
    "timestep_mean": 578.21875,
    "epsilon_norm": 7.86626672744751,
    "prediction_norm": 6.860283374786377,
    "timestamp": "2025-07-14T12:34:07.455349"
  },
  {
    "epoch": 955,
    "batch_idx": 32,
    "loss": 0.22898158431053162,
    "learning_rate": 0.0005353134929655838,
    "timestep_mean": 461.8125,
    "epsilon_norm": 7.315769672393799,
    "prediction_norm": 6.34361457824707,
    "timestamp": "2025-07-14T12:34:07.460613"
  },
  {
    "epoch": 955,
    "batch_idx": 64,
    "loss": 0.20524504780769348,
    "learning_rate": 0.0005353134929655838,
    "timestep_mean": 543.09375,
    "epsilon_norm": 8.215699195861816,
    "prediction_norm": 7.72813606262207,
    "timestamp": "2025-07-14T12:34:07.465796"
  },
  {
    "epoch": 955,
    "batch_idx": 96,
    "loss": 0.18063512444496155,
    "learning_rate": 0.0005353134929655838,
    "timestep_mean": 500.59375,
    "epsilon_norm": 7.634552478790283,
    "prediction_norm": 6.790855884552002,
    "timestamp": "2025-07-14T12:34:07.471112"
  },
  {
    "epoch": 955,
    "batch_idx": 128,
    "loss": 0.4800630807876587,
    "learning_rate": 0.0005353134929655838,
    "timestep_mean": 459.28125,
    "epsilon_norm": 8.828092575073242,
    "prediction_norm": 7.366050720214844,
    "timestamp": "2025-07-14T12:34:07.476697"
  },
  {
    "epoch": 955,
    "batch_idx": 160,
    "loss": 0.32456153631210327,
    "learning_rate": 0.0005353134929655838,
    "timestep_mean": 463.34375,
    "epsilon_norm": 8.351682662963867,
    "prediction_norm": 7.367738246917725,
    "timestamp": "2025-07-14T12:34:07.481881"
  },
  {
    "epoch": 955,
    "batch_idx": 192,
    "loss": 0.3476663827896118,
    "learning_rate": 0.0005353134929655838,
    "timestep_mean": 383.875,
    "epsilon_norm": 4.344024658203125,
    "prediction_norm": 3.000122308731079,
    "timestamp": "2025-07-14T12:34:07.487227"
  },
  {
    "epoch": 956,
    "batch_idx": 0,
    "loss": 0.16941548883914948,
    "learning_rate": 0.0005345300128572034,
    "timestep_mean": 492.21875,
    "epsilon_norm": 8.59716796875,
    "prediction_norm": 7.515681743621826,
    "timestamp": "2025-07-14T12:34:07.492573"
  },
  {
    "epoch": 956,
    "batch_idx": 32,
    "loss": 0.4122549891471863,
    "learning_rate": 0.0005345300128572034,
    "timestep_mean": 486.75,
    "epsilon_norm": 7.7000651359558105,
    "prediction_norm": 5.853328227996826,
    "timestamp": "2025-07-14T12:34:07.497707"
  },
  {
    "epoch": 956,
    "batch_idx": 64,
    "loss": 0.21040403842926025,
    "learning_rate": 0.0005345300128572034,
    "timestep_mean": 481.25,
    "epsilon_norm": 8.006025314331055,
    "prediction_norm": 7.107237339019775,
    "timestamp": "2025-07-14T12:34:07.502703"
  },
  {
    "epoch": 956,
    "batch_idx": 96,
    "loss": 0.24273750185966492,
    "learning_rate": 0.0005345300128572034,
    "timestep_mean": 564.25,
    "epsilon_norm": 7.437185287475586,
    "prediction_norm": 6.238079071044922,
    "timestamp": "2025-07-14T12:34:07.507993"
  },
  {
    "epoch": 956,
    "batch_idx": 128,
    "loss": 0.3181834816932678,
    "learning_rate": 0.0005345300128572034,
    "timestep_mean": 500.4375,
    "epsilon_norm": 8.271125793457031,
    "prediction_norm": 6.49592924118042,
    "timestamp": "2025-07-14T12:34:07.513405"
  },
  {
    "epoch": 956,
    "batch_idx": 160,
    "loss": 0.3282510042190552,
    "learning_rate": 0.0005345300128572034,
    "timestep_mean": 615.96875,
    "epsilon_norm": 9.03096866607666,
    "prediction_norm": 7.991118907928467,
    "timestamp": "2025-07-14T12:34:07.518740"
  },
  {
    "epoch": 956,
    "batch_idx": 192,
    "loss": 0.0976639837026596,
    "learning_rate": 0.0005345300128572034,
    "timestep_mean": 550.625,
    "epsilon_norm": 4.041299343109131,
    "prediction_norm": 3.643411636352539,
    "timestamp": "2025-07-14T12:34:07.523938"
  },
  {
    "epoch": 957,
    "batch_idx": 0,
    "loss": 0.24424904584884644,
    "learning_rate": 0.0005337464475494488,
    "timestep_mean": 472.71875,
    "epsilon_norm": 8.114394187927246,
    "prediction_norm": 7.3555588722229,
    "timestamp": "2025-07-14T12:34:07.529171"
  },
  {
    "epoch": 957,
    "batch_idx": 32,
    "loss": 0.2788997292518616,
    "learning_rate": 0.0005337464475494488,
    "timestep_mean": 471.875,
    "epsilon_norm": 8.231194496154785,
    "prediction_norm": 6.5741047859191895,
    "timestamp": "2025-07-14T12:34:07.534051"
  },
  {
    "epoch": 957,
    "batch_idx": 64,
    "loss": 0.278781920671463,
    "learning_rate": 0.0005337464475494488,
    "timestep_mean": 477.40625,
    "epsilon_norm": 7.5697784423828125,
    "prediction_norm": 6.002895355224609,
    "timestamp": "2025-07-14T12:34:07.538911"
  },
  {
    "epoch": 957,
    "batch_idx": 96,
    "loss": 0.17417576909065247,
    "learning_rate": 0.0005337464475494488,
    "timestep_mean": 531.0,
    "epsilon_norm": 8.954910278320312,
    "prediction_norm": 8.211840629577637,
    "timestamp": "2025-07-14T12:34:07.543766"
  },
  {
    "epoch": 957,
    "batch_idx": 128,
    "loss": 0.2383519411087036,
    "learning_rate": 0.0005337464475494488,
    "timestep_mean": 537.59375,
    "epsilon_norm": 7.87843656539917,
    "prediction_norm": 6.850082874298096,
    "timestamp": "2025-07-14T12:34:07.548819"
  },
  {
    "epoch": 957,
    "batch_idx": 160,
    "loss": 0.47601571679115295,
    "learning_rate": 0.0005337464475494488,
    "timestep_mean": 522.40625,
    "epsilon_norm": 8.256589889526367,
    "prediction_norm": 6.653649806976318,
    "timestamp": "2025-07-14T12:34:07.553896"
  },
  {
    "epoch": 957,
    "batch_idx": 192,
    "loss": 0.047826625406742096,
    "learning_rate": 0.0005337464475494488,
    "timestep_mean": 602.75,
    "epsilon_norm": 3.266258955001831,
    "prediction_norm": 3.220308780670166,
    "timestamp": "2025-07-14T12:34:07.558989"
  },
  {
    "epoch": 958,
    "batch_idx": 0,
    "loss": 0.2335902750492096,
    "learning_rate": 0.0005329627989756893,
    "timestep_mean": 373.875,
    "epsilon_norm": 8.863709449768066,
    "prediction_norm": 7.809141635894775,
    "timestamp": "2025-07-14T12:34:07.564281"
  },
  {
    "epoch": 958,
    "batch_idx": 32,
    "loss": 0.3715584874153137,
    "learning_rate": 0.0005329627989756893,
    "timestep_mean": 462.78125,
    "epsilon_norm": 8.625689506530762,
    "prediction_norm": 7.3076653480529785,
    "timestamp": "2025-07-14T12:34:07.569408"
  },
  {
    "epoch": 958,
    "batch_idx": 64,
    "loss": 0.3071483373641968,
    "learning_rate": 0.0005329627989756893,
    "timestep_mean": 540.125,
    "epsilon_norm": 9.6146240234375,
    "prediction_norm": 8.571456909179688,
    "timestamp": "2025-07-14T12:34:07.574693"
  },
  {
    "epoch": 958,
    "batch_idx": 96,
    "loss": 0.5046218037605286,
    "learning_rate": 0.0005329627989756893,
    "timestep_mean": 427.46875,
    "epsilon_norm": 8.336050987243652,
    "prediction_norm": 6.575595378875732,
    "timestamp": "2025-07-14T12:34:07.579958"
  },
  {
    "epoch": 958,
    "batch_idx": 128,
    "loss": 0.16126686334609985,
    "learning_rate": 0.0005329627989756893,
    "timestep_mean": 558.96875,
    "epsilon_norm": 7.596094131469727,
    "prediction_norm": 7.078442096710205,
    "timestamp": "2025-07-14T12:34:07.585139"
  },
  {
    "epoch": 958,
    "batch_idx": 160,
    "loss": 0.27909523248672485,
    "learning_rate": 0.0005329627989756893,
    "timestep_mean": 386.59375,
    "epsilon_norm": 7.455523490905762,
    "prediction_norm": 5.975871562957764,
    "timestamp": "2025-07-14T12:34:07.590225"
  },
  {
    "epoch": 958,
    "batch_idx": 192,
    "loss": 0.25014084577560425,
    "learning_rate": 0.0005329627989756893,
    "timestep_mean": 575.625,
    "epsilon_norm": 4.796738624572754,
    "prediction_norm": 4.0928215980529785,
    "timestamp": "2025-07-14T12:34:07.595338"
  },
  {
    "epoch": 959,
    "batch_idx": 0,
    "loss": 0.1911938637495041,
    "learning_rate": 0.0005321790690695003,
    "timestep_mean": 429.21875,
    "epsilon_norm": 7.4177327156066895,
    "prediction_norm": 7.053503513336182,
    "timestamp": "2025-07-14T12:34:07.600660"
  },
  {
    "epoch": 959,
    "batch_idx": 32,
    "loss": 0.3530922830104828,
    "learning_rate": 0.0005321790690695003,
    "timestep_mean": 528.65625,
    "epsilon_norm": 8.019012451171875,
    "prediction_norm": 6.693824291229248,
    "timestamp": "2025-07-14T12:34:07.605742"
  },
  {
    "epoch": 959,
    "batch_idx": 64,
    "loss": 0.3908137083053589,
    "learning_rate": 0.0005321790690695003,
    "timestep_mean": 540.78125,
    "epsilon_norm": 9.194783210754395,
    "prediction_norm": 7.407172203063965,
    "timestamp": "2025-07-14T12:34:07.610975"
  },
  {
    "epoch": 959,
    "batch_idx": 96,
    "loss": 0.32652807235717773,
    "learning_rate": 0.0005321790690695003,
    "timestep_mean": 500.0625,
    "epsilon_norm": 7.971209526062012,
    "prediction_norm": 6.353429317474365,
    "timestamp": "2025-07-14T12:34:07.616123"
  },
  {
    "epoch": 959,
    "batch_idx": 128,
    "loss": 0.41402947902679443,
    "learning_rate": 0.0005321790690695003,
    "timestep_mean": 466.34375,
    "epsilon_norm": 7.599549770355225,
    "prediction_norm": 6.134311199188232,
    "timestamp": "2025-07-14T12:34:07.621470"
  },
  {
    "epoch": 959,
    "batch_idx": 160,
    "loss": 0.2445160150527954,
    "learning_rate": 0.0005321790690695003,
    "timestep_mean": 457.78125,
    "epsilon_norm": 7.799923419952393,
    "prediction_norm": 6.469468116760254,
    "timestamp": "2025-07-14T12:34:07.626579"
  },
  {
    "epoch": 959,
    "batch_idx": 192,
    "loss": 0.4909161627292633,
    "learning_rate": 0.0005321790690695003,
    "timestep_mean": 442.125,
    "epsilon_norm": 4.366450786590576,
    "prediction_norm": 2.612220287322998,
    "timestamp": "2025-07-14T12:34:07.631656"
  },
  {
    "epoch": 960,
    "batch_idx": 0,
    "loss": 0.25140610337257385,
    "learning_rate": 0.0005313952597646572,
    "timestep_mean": 535.1875,
    "epsilon_norm": 7.117455005645752,
    "prediction_norm": 6.368276596069336,
    "timestamp": "2025-07-14T12:34:07.636823"
  },
  {
    "epoch": 960,
    "batch_idx": 32,
    "loss": 0.26349949836730957,
    "learning_rate": 0.0005313952597646572,
    "timestep_mean": 467.875,
    "epsilon_norm": 9.05402946472168,
    "prediction_norm": 8.466144561767578,
    "timestamp": "2025-07-14T12:34:07.641926"
  },
  {
    "epoch": 960,
    "batch_idx": 64,
    "loss": 0.21950393915176392,
    "learning_rate": 0.0005313952597646572,
    "timestep_mean": 525.0625,
    "epsilon_norm": 8.156731605529785,
    "prediction_norm": 6.870720863342285,
    "timestamp": "2025-07-14T12:34:07.646922"
  },
  {
    "epoch": 960,
    "batch_idx": 96,
    "loss": 0.2930360436439514,
    "learning_rate": 0.0005313952597646572,
    "timestep_mean": 550.3125,
    "epsilon_norm": 8.118236541748047,
    "prediction_norm": 7.11224889755249,
    "timestamp": "2025-07-14T12:34:07.652127"
  },
  {
    "epoch": 960,
    "batch_idx": 128,
    "loss": 0.35184264183044434,
    "learning_rate": 0.0005313952597646572,
    "timestep_mean": 439.875,
    "epsilon_norm": 7.477726459503174,
    "prediction_norm": 6.317290306091309,
    "timestamp": "2025-07-14T12:34:07.657217"
  },
  {
    "epoch": 960,
    "batch_idx": 160,
    "loss": 0.32623976469039917,
    "learning_rate": 0.0005313952597646572,
    "timestep_mean": 558.3125,
    "epsilon_norm": 7.423233985900879,
    "prediction_norm": 5.92009973526001,
    "timestamp": "2025-07-14T12:34:07.662530"
  },
  {
    "epoch": 960,
    "batch_idx": 192,
    "loss": 0.3191797733306885,
    "learning_rate": 0.0005313952597646572,
    "timestep_mean": 525.625,
    "epsilon_norm": 3.83537220954895,
    "prediction_norm": 3.2096164226531982,
    "timestamp": "2025-07-14T12:34:07.667782"
  },
  {
    "epoch": 961,
    "batch_idx": 0,
    "loss": 0.2810899019241333,
    "learning_rate": 0.0005306113729951314,
    "timestep_mean": 512.71875,
    "epsilon_norm": 7.797047138214111,
    "prediction_norm": 6.420706272125244,
    "timestamp": "2025-07-14T12:34:07.673085"
  },
  {
    "epoch": 961,
    "batch_idx": 32,
    "loss": 0.4220605492591858,
    "learning_rate": 0.0005306113729951314,
    "timestep_mean": 514.21875,
    "epsilon_norm": 7.512358665466309,
    "prediction_norm": 6.33981466293335,
    "timestamp": "2025-07-14T12:34:07.678371"
  },
  {
    "epoch": 961,
    "batch_idx": 64,
    "loss": 0.27550336718559265,
    "learning_rate": 0.0005306113729951314,
    "timestep_mean": 459.5,
    "epsilon_norm": 7.31903076171875,
    "prediction_norm": 6.230954170227051,
    "timestamp": "2025-07-14T12:34:07.683605"
  },
  {
    "epoch": 961,
    "batch_idx": 96,
    "loss": 0.3151310682296753,
    "learning_rate": 0.0005306113729951314,
    "timestep_mean": 424.3125,
    "epsilon_norm": 6.589241981506348,
    "prediction_norm": 4.849521636962891,
    "timestamp": "2025-07-14T12:34:07.688876"
  },
  {
    "epoch": 961,
    "batch_idx": 128,
    "loss": 0.2986624240875244,
    "learning_rate": 0.0005306113729951314,
    "timestep_mean": 494.40625,
    "epsilon_norm": 8.741477966308594,
    "prediction_norm": 7.359458923339844,
    "timestamp": "2025-07-14T12:34:07.694143"
  },
  {
    "epoch": 961,
    "batch_idx": 160,
    "loss": 0.4108327031135559,
    "learning_rate": 0.0005306113729951314,
    "timestep_mean": 477.375,
    "epsilon_norm": 7.991332530975342,
    "prediction_norm": 6.828563213348389,
    "timestamp": "2025-07-14T12:34:07.699466"
  },
  {
    "epoch": 961,
    "batch_idx": 192,
    "loss": 0.12671208381652832,
    "learning_rate": 0.0005306113729951314,
    "timestep_mean": 516.0,
    "epsilon_norm": 3.6520416736602783,
    "prediction_norm": 3.6271581649780273,
    "timestamp": "2025-07-14T12:34:07.704585"
  },
  {
    "epoch": 962,
    "batch_idx": 0,
    "loss": 0.3339031934738159,
    "learning_rate": 0.0005298274106950859,
    "timestep_mean": 503.1875,
    "epsilon_norm": 7.234450340270996,
    "prediction_norm": 5.817144870758057,
    "timestamp": "2025-07-14T12:34:07.709717"
  },
  {
    "epoch": 962,
    "batch_idx": 32,
    "loss": 0.24419936537742615,
    "learning_rate": 0.0005298274106950859,
    "timestep_mean": 438.375,
    "epsilon_norm": 7.054358959197998,
    "prediction_norm": 5.898529529571533,
    "timestamp": "2025-07-14T12:34:07.714918"
  },
  {
    "epoch": 962,
    "batch_idx": 64,
    "loss": 0.31722068786621094,
    "learning_rate": 0.0005298274106950859,
    "timestep_mean": 499.59375,
    "epsilon_norm": 8.370701789855957,
    "prediction_norm": 7.391661643981934,
    "timestamp": "2025-07-14T12:34:07.720108"
  },
  {
    "epoch": 962,
    "batch_idx": 96,
    "loss": 0.3687799870967865,
    "learning_rate": 0.0005298274106950859,
    "timestep_mean": 453.75,
    "epsilon_norm": 7.505398273468018,
    "prediction_norm": 5.914584159851074,
    "timestamp": "2025-07-14T12:34:07.725318"
  },
  {
    "epoch": 962,
    "batch_idx": 128,
    "loss": 0.23226995766162872,
    "learning_rate": 0.0005298274106950859,
    "timestep_mean": 484.375,
    "epsilon_norm": 7.153121471405029,
    "prediction_norm": 6.170682430267334,
    "timestamp": "2025-07-14T12:34:07.730420"
  },
  {
    "epoch": 962,
    "batch_idx": 160,
    "loss": 0.25347569584846497,
    "learning_rate": 0.0005298274106950859,
    "timestep_mean": 647.03125,
    "epsilon_norm": 8.290655136108398,
    "prediction_norm": 7.595010757446289,
    "timestamp": "2025-07-14T12:34:07.735713"
  },
  {
    "epoch": 962,
    "batch_idx": 192,
    "loss": 0.26874208450317383,
    "learning_rate": 0.0005298274106950859,
    "timestep_mean": 617.125,
    "epsilon_norm": 4.342182636260986,
    "prediction_norm": 3.0256500244140625,
    "timestamp": "2025-07-14T12:34:07.740884"
  },
  {
    "epoch": 963,
    "batch_idx": 0,
    "loss": 0.2235163301229477,
    "learning_rate": 0.0005290433747988694,
    "timestep_mean": 485.4375,
    "epsilon_norm": 6.544647693634033,
    "prediction_norm": 6.040735721588135,
    "timestamp": "2025-07-14T12:34:07.746236"
  },
  {
    "epoch": 963,
    "batch_idx": 32,
    "loss": 0.21690595149993896,
    "learning_rate": 0.0005290433747988694,
    "timestep_mean": 539.1875,
    "epsilon_norm": 7.760814666748047,
    "prediction_norm": 6.486544132232666,
    "timestamp": "2025-07-14T12:34:07.751530"
  },
  {
    "epoch": 963,
    "batch_idx": 64,
    "loss": 0.19122420251369476,
    "learning_rate": 0.0005290433747988694,
    "timestep_mean": 548.34375,
    "epsilon_norm": 8.1087646484375,
    "prediction_norm": 6.942504405975342,
    "timestamp": "2025-07-14T12:34:07.757116"
  },
  {
    "epoch": 963,
    "batch_idx": 96,
    "loss": 0.2748742997646332,
    "learning_rate": 0.0005290433747988694,
    "timestep_mean": 518.84375,
    "epsilon_norm": 6.9317426681518555,
    "prediction_norm": 6.043486595153809,
    "timestamp": "2025-07-14T12:34:07.762486"
  },
  {
    "epoch": 963,
    "batch_idx": 128,
    "loss": 0.38382038474082947,
    "learning_rate": 0.0005290433747988694,
    "timestep_mean": 505.40625,
    "epsilon_norm": 8.565190315246582,
    "prediction_norm": 6.523165702819824,
    "timestamp": "2025-07-14T12:34:07.767846"
  },
  {
    "epoch": 963,
    "batch_idx": 160,
    "loss": 0.27627402544021606,
    "learning_rate": 0.0005290433747988694,
    "timestep_mean": 499.875,
    "epsilon_norm": 8.811034202575684,
    "prediction_norm": 7.08048152923584,
    "timestamp": "2025-07-14T12:34:07.773102"
  },
  {
    "epoch": 963,
    "batch_idx": 192,
    "loss": 0.1501350700855255,
    "learning_rate": 0.0005290433747988694,
    "timestep_mean": 503.25,
    "epsilon_norm": 2.8067009449005127,
    "prediction_norm": 1.9750592708587646,
    "timestamp": "2025-07-14T12:34:07.778326"
  },
  {
    "epoch": 964,
    "batch_idx": 0,
    "loss": 0.23321330547332764,
    "learning_rate": 0.0005282592672410127,
    "timestep_mean": 545.6875,
    "epsilon_norm": 7.337825775146484,
    "prediction_norm": 6.806781768798828,
    "timestamp": "2025-07-14T12:34:07.783598"
  },
  {
    "epoch": 964,
    "batch_idx": 32,
    "loss": 0.2810966968536377,
    "learning_rate": 0.0005282592672410127,
    "timestep_mean": 475.5,
    "epsilon_norm": 7.828223705291748,
    "prediction_norm": 6.366962432861328,
    "timestamp": "2025-07-14T12:34:07.788727"
  },
  {
    "epoch": 964,
    "batch_idx": 64,
    "loss": 0.16683965921401978,
    "learning_rate": 0.0005282592672410127,
    "timestep_mean": 534.59375,
    "epsilon_norm": 8.368027687072754,
    "prediction_norm": 7.804920673370361,
    "timestamp": "2025-07-14T12:34:07.793886"
  },
  {
    "epoch": 964,
    "batch_idx": 96,
    "loss": 0.2408684343099594,
    "learning_rate": 0.0005282592672410127,
    "timestep_mean": 441.8125,
    "epsilon_norm": 7.247613430023193,
    "prediction_norm": 6.113141059875488,
    "timestamp": "2025-07-14T12:34:07.799204"
  },
  {
    "epoch": 964,
    "batch_idx": 128,
    "loss": 0.2556811571121216,
    "learning_rate": 0.0005282592672410127,
    "timestep_mean": 439.71875,
    "epsilon_norm": 7.596138000488281,
    "prediction_norm": 6.879775047302246,
    "timestamp": "2025-07-14T12:34:07.805009"
  },
  {
    "epoch": 964,
    "batch_idx": 160,
    "loss": 0.2863212823867798,
    "learning_rate": 0.0005282592672410127,
    "timestep_mean": 433.4375,
    "epsilon_norm": 7.788568019866943,
    "prediction_norm": 6.791958808898926,
    "timestamp": "2025-07-14T12:34:07.810233"
  },
  {
    "epoch": 964,
    "batch_idx": 192,
    "loss": 0.32403072714805603,
    "learning_rate": 0.0005282592672410127,
    "timestep_mean": 459.75,
    "epsilon_norm": 3.001673936843872,
    "prediction_norm": 2.8024144172668457,
    "timestamp": "2025-07-14T12:34:07.815630"
  },
  {
    "epoch": 965,
    "batch_idx": 0,
    "loss": 0.22400906682014465,
    "learning_rate": 0.0005274750899562234,
    "timestep_mean": 486.4375,
    "epsilon_norm": 8.466493606567383,
    "prediction_norm": 7.6177191734313965,
    "timestamp": "2025-07-14T12:34:07.820988"
  },
  {
    "epoch": 965,
    "batch_idx": 32,
    "loss": 0.49817657470703125,
    "learning_rate": 0.0005274750899562234,
    "timestep_mean": 462.84375,
    "epsilon_norm": 9.011012077331543,
    "prediction_norm": 7.309965133666992,
    "timestamp": "2025-07-14T12:34:07.826278"
  },
  {
    "epoch": 965,
    "batch_idx": 64,
    "loss": 0.366131067276001,
    "learning_rate": 0.0005274750899562234,
    "timestep_mean": 502.5,
    "epsilon_norm": 7.961157321929932,
    "prediction_norm": 5.937507152557373,
    "timestamp": "2025-07-14T12:34:07.831509"
  },
  {
    "epoch": 965,
    "batch_idx": 96,
    "loss": 0.24785621464252472,
    "learning_rate": 0.0005274750899562234,
    "timestep_mean": 565.46875,
    "epsilon_norm": 7.371257781982422,
    "prediction_norm": 6.131665229797363,
    "timestamp": "2025-07-14T12:34:07.836748"
  },
  {
    "epoch": 965,
    "batch_idx": 128,
    "loss": 0.3798622488975525,
    "learning_rate": 0.0005274750899562234,
    "timestep_mean": 406.8125,
    "epsilon_norm": 7.979516983032227,
    "prediction_norm": 5.930929660797119,
    "timestamp": "2025-07-14T12:34:07.842124"
  },
  {
    "epoch": 965,
    "batch_idx": 160,
    "loss": 0.4342581033706665,
    "learning_rate": 0.0005274750899562234,
    "timestep_mean": 408.875,
    "epsilon_norm": 7.645096302032471,
    "prediction_norm": 5.960108757019043,
    "timestamp": "2025-07-14T12:34:07.847523"
  },
  {
    "epoch": 965,
    "batch_idx": 192,
    "loss": 0.060915883630514145,
    "learning_rate": 0.0005274750899562234,
    "timestep_mean": 564.75,
    "epsilon_norm": 2.8932876586914062,
    "prediction_norm": 2.3402457237243652,
    "timestamp": "2025-07-14T12:34:07.852719"
  },
  {
    "epoch": 966,
    "batch_idx": 0,
    "loss": 0.16225291788578033,
    "learning_rate": 0.0005266908448793808,
    "timestep_mean": 538.46875,
    "epsilon_norm": 7.794586181640625,
    "prediction_norm": 7.140225410461426,
    "timestamp": "2025-07-14T12:34:07.858012"
  },
  {
    "epoch": 966,
    "batch_idx": 32,
    "loss": 0.10237205028533936,
    "learning_rate": 0.0005266908448793808,
    "timestep_mean": 609.75,
    "epsilon_norm": 7.2873854637146,
    "prediction_norm": 7.127594470977783,
    "timestamp": "2025-07-14T12:34:07.863161"
  },
  {
    "epoch": 966,
    "batch_idx": 64,
    "loss": 0.28783613443374634,
    "learning_rate": 0.0005266908448793808,
    "timestep_mean": 471.03125,
    "epsilon_norm": 8.009771347045898,
    "prediction_norm": 6.344047546386719,
    "timestamp": "2025-07-14T12:34:07.868376"
  },
  {
    "epoch": 966,
    "batch_idx": 96,
    "loss": 0.35096079111099243,
    "learning_rate": 0.0005266908448793808,
    "timestep_mean": 446.65625,
    "epsilon_norm": 7.9803619384765625,
    "prediction_norm": 6.316361427307129,
    "timestamp": "2025-07-14T12:34:07.873503"
  },
  {
    "epoch": 966,
    "batch_idx": 128,
    "loss": 0.30863243341445923,
    "learning_rate": 0.0005266908448793808,
    "timestep_mean": 449.09375,
    "epsilon_norm": 7.045931339263916,
    "prediction_norm": 5.408228874206543,
    "timestamp": "2025-07-14T12:34:07.878814"
  },
  {
    "epoch": 966,
    "batch_idx": 160,
    "loss": 0.29460209608078003,
    "learning_rate": 0.0005266908448793808,
    "timestep_mean": 460.21875,
    "epsilon_norm": 8.910004615783691,
    "prediction_norm": 7.213700771331787,
    "timestamp": "2025-07-14T12:34:07.884006"
  },
  {
    "epoch": 966,
    "batch_idx": 192,
    "loss": 0.09094463288784027,
    "learning_rate": 0.0005266908448793808,
    "timestep_mean": 483.5,
    "epsilon_norm": 3.3301565647125244,
    "prediction_norm": 3.2081592082977295,
    "timestamp": "2025-07-14T12:34:07.889178"
  },
  {
    "epoch": 967,
    "batch_idx": 0,
    "loss": 0.2115945816040039,
    "learning_rate": 0.0005259065339455317,
    "timestep_mean": 538.5,
    "epsilon_norm": 7.220952987670898,
    "prediction_norm": 6.453756809234619,
    "timestamp": "2025-07-14T12:34:07.894498"
  },
  {
    "epoch": 967,
    "batch_idx": 32,
    "loss": 0.2107330709695816,
    "learning_rate": 0.0005259065339455317,
    "timestep_mean": 538.3125,
    "epsilon_norm": 8.602884292602539,
    "prediction_norm": 8.081405639648438,
    "timestamp": "2025-07-14T12:34:07.899468"
  },
  {
    "epoch": 967,
    "batch_idx": 64,
    "loss": 0.14019818603992462,
    "learning_rate": 0.0005259065339455317,
    "timestep_mean": 596.25,
    "epsilon_norm": 8.080570220947266,
    "prediction_norm": 7.061788082122803,
    "timestamp": "2025-07-14T12:34:07.904613"
  },
  {
    "epoch": 967,
    "batch_idx": 96,
    "loss": 0.24406222999095917,
    "learning_rate": 0.0005259065339455317,
    "timestep_mean": 510.15625,
    "epsilon_norm": 7.349552631378174,
    "prediction_norm": 6.190814971923828,
    "timestamp": "2025-07-14T12:34:07.909696"
  },
  {
    "epoch": 967,
    "batch_idx": 128,
    "loss": 0.336627721786499,
    "learning_rate": 0.0005259065339455317,
    "timestep_mean": 481.84375,
    "epsilon_norm": 8.087459564208984,
    "prediction_norm": 6.6730241775512695,
    "timestamp": "2025-07-14T12:34:07.914914"
  },
  {
    "epoch": 967,
    "batch_idx": 160,
    "loss": 0.27512115240097046,
    "learning_rate": 0.0005259065339455317,
    "timestep_mean": 493.46875,
    "epsilon_norm": 8.9270658493042,
    "prediction_norm": 6.6590986251831055,
    "timestamp": "2025-07-14T12:34:07.920255"
  },
  {
    "epoch": 967,
    "batch_idx": 192,
    "loss": 0.3360443115234375,
    "learning_rate": 0.0005259065339455317,
    "timestep_mean": 594.75,
    "epsilon_norm": 4.0997443199157715,
    "prediction_norm": 3.461768627166748,
    "timestamp": "2025-07-14T12:34:07.925439"
  },
  {
    "epoch": 968,
    "batch_idx": 0,
    "loss": 0.34325963258743286,
    "learning_rate": 0.0005251221590898853,
    "timestep_mean": 455.3125,
    "epsilon_norm": 7.996901988983154,
    "prediction_norm": 6.039830684661865,
    "timestamp": "2025-07-14T12:34:07.930618"
  },
  {
    "epoch": 968,
    "batch_idx": 32,
    "loss": 0.24196869134902954,
    "learning_rate": 0.0005251221590898853,
    "timestep_mean": 530.125,
    "epsilon_norm": 8.305929183959961,
    "prediction_norm": 7.206416130065918,
    "timestamp": "2025-07-14T12:34:07.935954"
  },
  {
    "epoch": 968,
    "batch_idx": 64,
    "loss": 0.3018367886543274,
    "learning_rate": 0.0005251221590898853,
    "timestep_mean": 494.0625,
    "epsilon_norm": 7.987077236175537,
    "prediction_norm": 6.236448287963867,
    "timestamp": "2025-07-14T12:34:07.941174"
  },
  {
    "epoch": 968,
    "batch_idx": 96,
    "loss": 0.2173323631286621,
    "learning_rate": 0.0005251221590898853,
    "timestep_mean": 399.0625,
    "epsilon_norm": 7.700977325439453,
    "prediction_norm": 6.870692253112793,
    "timestamp": "2025-07-14T12:34:07.946531"
  },
  {
    "epoch": 968,
    "batch_idx": 128,
    "loss": 0.21420390903949738,
    "learning_rate": 0.0005251221590898853,
    "timestep_mean": 435.5,
    "epsilon_norm": 7.43075704574585,
    "prediction_norm": 6.355838298797607,
    "timestamp": "2025-07-14T12:34:07.951814"
  },
  {
    "epoch": 968,
    "batch_idx": 160,
    "loss": 0.3383530378341675,
    "learning_rate": 0.0005251221590898853,
    "timestep_mean": 442.9375,
    "epsilon_norm": 8.044355392456055,
    "prediction_norm": 6.8139801025390625,
    "timestamp": "2025-07-14T12:34:07.957045"
  },
  {
    "epoch": 968,
    "batch_idx": 192,
    "loss": 0.09031633287668228,
    "learning_rate": 0.0005251221590898853,
    "timestep_mean": 638.75,
    "epsilon_norm": 3.3254587650299072,
    "prediction_norm": 3.2584118843078613,
    "timestamp": "2025-07-14T12:34:07.962462"
  },
  {
    "epoch": 969,
    "batch_idx": 0,
    "loss": 0.29049500823020935,
    "learning_rate": 0.0005243377222478088,
    "timestep_mean": 466.3125,
    "epsilon_norm": 8.730913162231445,
    "prediction_norm": 7.554446220397949,
    "timestamp": "2025-07-14T12:34:07.967646"
  },
  {
    "epoch": 969,
    "batch_idx": 32,
    "loss": 0.2961740791797638,
    "learning_rate": 0.0005243377222478088,
    "timestep_mean": 504.75,
    "epsilon_norm": 8.47828483581543,
    "prediction_norm": 7.134770393371582,
    "timestamp": "2025-07-14T12:34:07.972723"
  },
  {
    "epoch": 969,
    "batch_idx": 64,
    "loss": 0.2829100489616394,
    "learning_rate": 0.0005243377222478088,
    "timestep_mean": 496.96875,
    "epsilon_norm": 6.081307888031006,
    "prediction_norm": 4.832281589508057,
    "timestamp": "2025-07-14T12:34:07.977810"
  },
  {
    "epoch": 969,
    "batch_idx": 96,
    "loss": 0.30408233404159546,
    "learning_rate": 0.0005243377222478088,
    "timestep_mean": 449.84375,
    "epsilon_norm": 7.561385631561279,
    "prediction_norm": 6.266715049743652,
    "timestamp": "2025-07-14T12:34:07.983024"
  },
  {
    "epoch": 969,
    "batch_idx": 128,
    "loss": 0.45921409130096436,
    "learning_rate": 0.0005243377222478088,
    "timestep_mean": 536.0625,
    "epsilon_norm": 8.858516693115234,
    "prediction_norm": 6.237822532653809,
    "timestamp": "2025-07-14T12:34:07.988404"
  },
  {
    "epoch": 969,
    "batch_idx": 160,
    "loss": 0.29517996311187744,
    "learning_rate": 0.0005243377222478088,
    "timestep_mean": 494.9375,
    "epsilon_norm": 7.192579746246338,
    "prediction_norm": 5.739612579345703,
    "timestamp": "2025-07-14T12:34:07.993639"
  },
  {
    "epoch": 969,
    "batch_idx": 192,
    "loss": 0.17969241738319397,
    "learning_rate": 0.0005243377222478088,
    "timestep_mean": 475.25,
    "epsilon_norm": 5.408271312713623,
    "prediction_norm": 4.5570387840271,
    "timestamp": "2025-07-14T12:34:07.998823"
  },
  {
    "epoch": 970,
    "batch_idx": 0,
    "loss": 0.3788366913795471,
    "learning_rate": 0.0005235532253548218,
    "timestep_mean": 418.09375,
    "epsilon_norm": 7.265214920043945,
    "prediction_norm": 5.951691627502441,
    "timestamp": "2025-07-14T12:34:08.004198"
  },
  {
    "epoch": 970,
    "batch_idx": 32,
    "loss": 0.15229558944702148,
    "learning_rate": 0.0005235532253548218,
    "timestep_mean": 557.90625,
    "epsilon_norm": 7.908036231994629,
    "prediction_norm": 8.326204299926758,
    "timestamp": "2025-07-14T12:34:08.009288"
  },
  {
    "epoch": 970,
    "batch_idx": 64,
    "loss": 0.2215839922428131,
    "learning_rate": 0.0005235532253548218,
    "timestep_mean": 475.40625,
    "epsilon_norm": 7.339798450469971,
    "prediction_norm": 6.664439678192139,
    "timestamp": "2025-07-14T12:34:08.014460"
  },
  {
    "epoch": 970,
    "batch_idx": 96,
    "loss": 0.17931796610355377,
    "learning_rate": 0.0005235532253548218,
    "timestep_mean": 591.84375,
    "epsilon_norm": 8.494284629821777,
    "prediction_norm": 8.603564262390137,
    "timestamp": "2025-07-14T12:34:08.019666"
  },
  {
    "epoch": 970,
    "batch_idx": 128,
    "loss": 0.40943771600723267,
    "learning_rate": 0.0005235532253548218,
    "timestep_mean": 502.4375,
    "epsilon_norm": 7.867713928222656,
    "prediction_norm": 7.343405246734619,
    "timestamp": "2025-07-14T12:34:08.024895"
  },
  {
    "epoch": 970,
    "batch_idx": 160,
    "loss": 0.29135650396347046,
    "learning_rate": 0.0005235532253548218,
    "timestep_mean": 627.59375,
    "epsilon_norm": 9.954931259155273,
    "prediction_norm": 9.273341178894043,
    "timestamp": "2025-07-14T12:34:08.030007"
  },
  {
    "epoch": 970,
    "batch_idx": 192,
    "loss": 0.17429307103157043,
    "learning_rate": 0.0005235532253548218,
    "timestep_mean": 314.75,
    "epsilon_norm": 2.9442150592803955,
    "prediction_norm": 2.5870070457458496,
    "timestamp": "2025-07-14T12:34:08.034881"
  },
  {
    "epoch": 971,
    "batch_idx": 0,
    "loss": 0.24627739191055298,
    "learning_rate": 0.0005227686703465927,
    "timestep_mean": 548.375,
    "epsilon_norm": 9.808475494384766,
    "prediction_norm": 8.720734596252441,
    "timestamp": "2025-07-14T12:34:08.039894"
  },
  {
    "epoch": 971,
    "batch_idx": 32,
    "loss": 0.3005967140197754,
    "learning_rate": 0.0005227686703465927,
    "timestep_mean": 514.46875,
    "epsilon_norm": 8.261225700378418,
    "prediction_norm": 7.893529891967773,
    "timestamp": "2025-07-14T12:34:08.045546"
  },
  {
    "epoch": 971,
    "batch_idx": 64,
    "loss": 0.26440730690956116,
    "learning_rate": 0.0005227686703465927,
    "timestep_mean": 476.09375,
    "epsilon_norm": 8.407320022583008,
    "prediction_norm": 7.760444641113281,
    "timestamp": "2025-07-14T12:34:08.052640"
  },
  {
    "epoch": 971,
    "batch_idx": 96,
    "loss": 0.19871513545513153,
    "learning_rate": 0.0005227686703465927,
    "timestep_mean": 526.65625,
    "epsilon_norm": 7.43920373916626,
    "prediction_norm": 6.859098434448242,
    "timestamp": "2025-07-14T12:34:08.057974"
  },
  {
    "epoch": 971,
    "batch_idx": 128,
    "loss": 0.31587618589401245,
    "learning_rate": 0.0005227686703465927,
    "timestep_mean": 380.625,
    "epsilon_norm": 7.526426315307617,
    "prediction_norm": 6.636988162994385,
    "timestamp": "2025-07-14T12:34:08.063222"
  },
  {
    "epoch": 971,
    "batch_idx": 160,
    "loss": 0.3953043222427368,
    "learning_rate": 0.0005227686703465927,
    "timestep_mean": 490.03125,
    "epsilon_norm": 7.621191501617432,
    "prediction_norm": 5.2214765548706055,
    "timestamp": "2025-07-14T12:34:08.068553"
  },
  {
    "epoch": 971,
    "batch_idx": 192,
    "loss": 0.06708074361085892,
    "learning_rate": 0.0005227686703465927,
    "timestep_mean": 700.75,
    "epsilon_norm": 4.114471912384033,
    "prediction_norm": 4.023030757904053,
    "timestamp": "2025-07-14T12:34:08.073855"
  },
  {
    "epoch": 972,
    "batch_idx": 0,
    "loss": 0.22308745980262756,
    "learning_rate": 0.0005219840591589329,
    "timestep_mean": 487.0,
    "epsilon_norm": 7.491372108459473,
    "prediction_norm": 6.6258769035339355,
    "timestamp": "2025-07-14T12:34:08.079359"
  },
  {
    "epoch": 972,
    "batch_idx": 32,
    "loss": 0.3670216500759125,
    "learning_rate": 0.0005219840591589329,
    "timestep_mean": 388.1875,
    "epsilon_norm": 7.754861354827881,
    "prediction_norm": 5.980881214141846,
    "timestamp": "2025-07-14T12:34:08.084688"
  },
  {
    "epoch": 972,
    "batch_idx": 64,
    "loss": 0.43367302417755127,
    "learning_rate": 0.0005219840591589329,
    "timestep_mean": 445.125,
    "epsilon_norm": 8.992389678955078,
    "prediction_norm": 6.391114234924316,
    "timestamp": "2025-07-14T12:34:08.089779"
  },
  {
    "epoch": 972,
    "batch_idx": 96,
    "loss": 0.2787376642227173,
    "learning_rate": 0.0005219840591589329,
    "timestep_mean": 506.3125,
    "epsilon_norm": 8.210866928100586,
    "prediction_norm": 6.883781433105469,
    "timestamp": "2025-07-14T12:34:08.094847"
  },
  {
    "epoch": 972,
    "batch_idx": 128,
    "loss": 0.3798999786376953,
    "learning_rate": 0.0005219840591589329,
    "timestep_mean": 505.625,
    "epsilon_norm": 7.33876895904541,
    "prediction_norm": 5.411375522613525,
    "timestamp": "2025-07-14T12:34:08.100012"
  },
  {
    "epoch": 972,
    "batch_idx": 160,
    "loss": 0.32188650965690613,
    "learning_rate": 0.0005219840591589329,
    "timestep_mean": 479.09375,
    "epsilon_norm": 7.1116862297058105,
    "prediction_norm": 5.563775062561035,
    "timestamp": "2025-07-14T12:34:08.105187"
  },
  {
    "epoch": 972,
    "batch_idx": 192,
    "loss": 0.13581852614879608,
    "learning_rate": 0.0005219840591589329,
    "timestep_mean": 580.25,
    "epsilon_norm": 4.010136127471924,
    "prediction_norm": 3.7141759395599365,
    "timestamp": "2025-07-14T12:34:08.110254"
  },
  {
    "epoch": 973,
    "batch_idx": 0,
    "loss": 0.3095419406890869,
    "learning_rate": 0.0005211993937277926,
    "timestep_mean": 515.09375,
    "epsilon_norm": 10.043975830078125,
    "prediction_norm": 8.46823787689209,
    "timestamp": "2025-07-14T12:34:08.115510"
  },
  {
    "epoch": 973,
    "batch_idx": 32,
    "loss": 0.41043543815612793,
    "learning_rate": 0.0005211993937277926,
    "timestep_mean": 549.25,
    "epsilon_norm": 9.452828407287598,
    "prediction_norm": 7.358465671539307,
    "timestamp": "2025-07-14T12:34:08.121041"
  },
  {
    "epoch": 973,
    "batch_idx": 64,
    "loss": 0.2855394780635834,
    "learning_rate": 0.0005211993937277926,
    "timestep_mean": 549.65625,
    "epsilon_norm": 7.613279819488525,
    "prediction_norm": 5.592346668243408,
    "timestamp": "2025-07-14T12:34:08.126192"
  },
  {
    "epoch": 973,
    "batch_idx": 96,
    "loss": 0.3227907717227936,
    "learning_rate": 0.0005211993937277926,
    "timestep_mean": 468.3125,
    "epsilon_norm": 8.134058952331543,
    "prediction_norm": 7.2988739013671875,
    "timestamp": "2025-07-14T12:34:08.131457"
  },
  {
    "epoch": 973,
    "batch_idx": 128,
    "loss": 0.17450489103794098,
    "learning_rate": 0.0005211993937277926,
    "timestep_mean": 543.0,
    "epsilon_norm": 7.246286392211914,
    "prediction_norm": 6.433797836303711,
    "timestamp": "2025-07-14T12:34:08.136709"
  },
  {
    "epoch": 973,
    "batch_idx": 160,
    "loss": 0.3301578164100647,
    "learning_rate": 0.0005211993937277926,
    "timestep_mean": 539.875,
    "epsilon_norm": 7.909909248352051,
    "prediction_norm": 6.32703161239624,
    "timestamp": "2025-07-14T12:34:08.141904"
  },
  {
    "epoch": 973,
    "batch_idx": 192,
    "loss": 0.6450033187866211,
    "learning_rate": 0.0005211993937277926,
    "timestep_mean": 507.25,
    "epsilon_norm": 5.160254001617432,
    "prediction_norm": 4.011487007141113,
    "timestamp": "2025-07-14T12:34:08.146976"
  },
  {
    "epoch": 974,
    "batch_idx": 0,
    "loss": 0.2547895014286041,
    "learning_rate": 0.0005204146759892555,
    "timestep_mean": 402.75,
    "epsilon_norm": 8.481168746948242,
    "prediction_norm": 7.644158840179443,
    "timestamp": "2025-07-14T12:34:08.152216"
  },
  {
    "epoch": 974,
    "batch_idx": 32,
    "loss": 0.18692699074745178,
    "learning_rate": 0.0005204146759892555,
    "timestep_mean": 504.71875,
    "epsilon_norm": 7.214434623718262,
    "prediction_norm": 6.734559535980225,
    "timestamp": "2025-07-14T12:34:08.157445"
  },
  {
    "epoch": 974,
    "batch_idx": 64,
    "loss": 0.2731057405471802,
    "learning_rate": 0.0005204146759892555,
    "timestep_mean": 468.6875,
    "epsilon_norm": 7.729478359222412,
    "prediction_norm": 7.517419815063477,
    "timestamp": "2025-07-14T12:34:08.162748"
  },
  {
    "epoch": 974,
    "batch_idx": 96,
    "loss": 0.29383155703544617,
    "learning_rate": 0.0005204146759892555,
    "timestep_mean": 545.28125,
    "epsilon_norm": 8.426494598388672,
    "prediction_norm": 7.548886775970459,
    "timestamp": "2025-07-14T12:34:08.168029"
  },
  {
    "epoch": 974,
    "batch_idx": 128,
    "loss": 0.4000657796859741,
    "learning_rate": 0.0005204146759892555,
    "timestep_mean": 466.03125,
    "epsilon_norm": 6.933090686798096,
    "prediction_norm": 5.517980098724365,
    "timestamp": "2025-07-14T12:34:08.173297"
  },
  {
    "epoch": 974,
    "batch_idx": 160,
    "loss": 0.27363502979278564,
    "learning_rate": 0.0005204146759892555,
    "timestep_mean": 544.15625,
    "epsilon_norm": 8.699056625366211,
    "prediction_norm": 7.303049087524414,
    "timestamp": "2025-07-14T12:34:08.178441"
  },
  {
    "epoch": 974,
    "batch_idx": 192,
    "loss": 0.30779340863227844,
    "learning_rate": 0.0005204146759892555,
    "timestep_mean": 412.5,
    "epsilon_norm": 4.678773880004883,
    "prediction_norm": 3.9510838985443115,
    "timestamp": "2025-07-14T12:34:08.183827"
  },
  {
    "epoch": 975,
    "batch_idx": 0,
    "loss": 0.16312119364738464,
    "learning_rate": 0.0005196299078795348,
    "timestep_mean": 513.40625,
    "epsilon_norm": 7.695909023284912,
    "prediction_norm": 7.539241313934326,
    "timestamp": "2025-07-14T12:34:08.188982"
  },
  {
    "epoch": 975,
    "batch_idx": 32,
    "loss": 0.13204902410507202,
    "learning_rate": 0.0005196299078795348,
    "timestep_mean": 590.40625,
    "epsilon_norm": 7.501337051391602,
    "prediction_norm": 6.945034980773926,
    "timestamp": "2025-07-14T12:34:08.194227"
  },
  {
    "epoch": 975,
    "batch_idx": 64,
    "loss": 0.27194201946258545,
    "learning_rate": 0.0005196299078795348,
    "timestep_mean": 446.90625,
    "epsilon_norm": 6.9748854637146,
    "prediction_norm": 5.723906517028809,
    "timestamp": "2025-07-14T12:34:08.199388"
  },
  {
    "epoch": 975,
    "batch_idx": 96,
    "loss": 0.16315360367298126,
    "learning_rate": 0.0005196299078795348,
    "timestep_mean": 481.78125,
    "epsilon_norm": 7.502099990844727,
    "prediction_norm": 6.923092365264893,
    "timestamp": "2025-07-14T12:34:08.204694"
  },
  {
    "epoch": 975,
    "batch_idx": 128,
    "loss": 0.2479504942893982,
    "learning_rate": 0.0005196299078795348,
    "timestep_mean": 594.25,
    "epsilon_norm": 8.164031982421875,
    "prediction_norm": 6.694563865661621,
    "timestamp": "2025-07-14T12:34:08.209903"
  },
  {
    "epoch": 975,
    "batch_idx": 160,
    "loss": 0.3685336410999298,
    "learning_rate": 0.0005196299078795348,
    "timestep_mean": 482.0,
    "epsilon_norm": 8.597989082336426,
    "prediction_norm": 6.683302879333496,
    "timestamp": "2025-07-14T12:34:08.215138"
  },
  {
    "epoch": 975,
    "batch_idx": 192,
    "loss": 0.05903955176472664,
    "learning_rate": 0.0005196299078795348,
    "timestep_mean": 522.375,
    "epsilon_norm": 3.2312417030334473,
    "prediction_norm": 3.1502859592437744,
    "timestamp": "2025-07-14T12:34:08.220233"
  },
  {
    "epoch": 976,
    "batch_idx": 0,
    "loss": 0.26858511567115784,
    "learning_rate": 0.0005188450913349677,
    "timestep_mean": 477.125,
    "epsilon_norm": 7.235387802124023,
    "prediction_norm": 5.578215599060059,
    "timestamp": "2025-07-14T12:34:08.225618"
  },
  {
    "epoch": 976,
    "batch_idx": 32,
    "loss": 0.20726323127746582,
    "learning_rate": 0.0005188450913349677,
    "timestep_mean": 531.53125,
    "epsilon_norm": 6.7111968994140625,
    "prediction_norm": 5.314711570739746,
    "timestamp": "2025-07-14T12:34:08.230766"
  },
  {
    "epoch": 976,
    "batch_idx": 64,
    "loss": 0.24797937273979187,
    "learning_rate": 0.0005188450913349677,
    "timestep_mean": 494.84375,
    "epsilon_norm": 7.343705177307129,
    "prediction_norm": 6.350042343139648,
    "timestamp": "2025-07-14T12:34:08.235918"
  },
  {
    "epoch": 976,
    "batch_idx": 96,
    "loss": 0.45977672934532166,
    "learning_rate": 0.0005188450913349677,
    "timestep_mean": 516.34375,
    "epsilon_norm": 9.278789520263672,
    "prediction_norm": 7.184093952178955,
    "timestamp": "2025-07-14T12:34:08.240858"
  },
  {
    "epoch": 976,
    "batch_idx": 128,
    "loss": 0.19168980419635773,
    "learning_rate": 0.0005188450913349677,
    "timestep_mean": 569.0625,
    "epsilon_norm": 8.294707298278809,
    "prediction_norm": 7.198211193084717,
    "timestamp": "2025-07-14T12:34:08.245957"
  },
  {
    "epoch": 976,
    "batch_idx": 160,
    "loss": 0.12623628973960876,
    "learning_rate": 0.0005188450913349677,
    "timestep_mean": 500.59375,
    "epsilon_norm": 7.967897891998291,
    "prediction_norm": 7.018529891967773,
    "timestamp": "2025-07-14T12:34:08.250887"
  },
  {
    "epoch": 976,
    "batch_idx": 192,
    "loss": 0.44149884581565857,
    "learning_rate": 0.0005188450913349677,
    "timestep_mean": 554.125,
    "epsilon_norm": 3.5479254722595215,
    "prediction_norm": 2.452461004257202,
    "timestamp": "2025-07-14T12:34:08.256391"
  },
  {
    "epoch": 977,
    "batch_idx": 0,
    "loss": 0.29855114221572876,
    "learning_rate": 0.0005180602282920111,
    "timestep_mean": 564.34375,
    "epsilon_norm": 9.444681167602539,
    "prediction_norm": 8.610406875610352,
    "timestamp": "2025-07-14T12:34:08.261780"
  },
  {
    "epoch": 977,
    "batch_idx": 32,
    "loss": 0.1112491637468338,
    "learning_rate": 0.0005180602282920111,
    "timestep_mean": 553.65625,
    "epsilon_norm": 7.927365303039551,
    "prediction_norm": 7.2054924964904785,
    "timestamp": "2025-07-14T12:34:08.268899"
  },
  {
    "epoch": 977,
    "batch_idx": 64,
    "loss": 0.31561434268951416,
    "learning_rate": 0.0005180602282920111,
    "timestep_mean": 495.9375,
    "epsilon_norm": 8.745856285095215,
    "prediction_norm": 6.220358371734619,
    "timestamp": "2025-07-14T12:34:08.274251"
  },
  {
    "epoch": 977,
    "batch_idx": 96,
    "loss": 0.3421492576599121,
    "learning_rate": 0.0005180602282920111,
    "timestep_mean": 410.90625,
    "epsilon_norm": 8.38075065612793,
    "prediction_norm": 6.642214775085449,
    "timestamp": "2025-07-14T12:34:08.279556"
  },
  {
    "epoch": 977,
    "batch_idx": 128,
    "loss": 0.5369582772254944,
    "learning_rate": 0.0005180602282920111,
    "timestep_mean": 477.40625,
    "epsilon_norm": 9.342991828918457,
    "prediction_norm": 6.447159290313721,
    "timestamp": "2025-07-14T12:34:08.284704"
  },
  {
    "epoch": 977,
    "batch_idx": 160,
    "loss": 0.3229891061782837,
    "learning_rate": 0.0005180602282920111,
    "timestep_mean": 526.90625,
    "epsilon_norm": 6.868094444274902,
    "prediction_norm": 5.504391193389893,
    "timestamp": "2025-07-14T12:34:08.290033"
  },
  {
    "epoch": 977,
    "batch_idx": 192,
    "loss": 0.23314082622528076,
    "learning_rate": 0.0005180602282920111,
    "timestep_mean": 450.875,
    "epsilon_norm": 3.192229747772217,
    "prediction_norm": 2.216620445251465,
    "timestamp": "2025-07-14T12:34:08.295313"
  },
  {
    "epoch": 978,
    "batch_idx": 0,
    "loss": 0.13457441329956055,
    "learning_rate": 0.0005172753206872367,
    "timestep_mean": 494.9375,
    "epsilon_norm": 8.025233268737793,
    "prediction_norm": 7.2575364112854,
    "timestamp": "2025-07-14T12:34:08.300782"
  },
  {
    "epoch": 978,
    "batch_idx": 32,
    "loss": 0.25413358211517334,
    "learning_rate": 0.0005172753206872367,
    "timestep_mean": 490.25,
    "epsilon_norm": 8.566641807556152,
    "prediction_norm": 7.300342559814453,
    "timestamp": "2025-07-14T12:34:08.306021"
  },
  {
    "epoch": 978,
    "batch_idx": 64,
    "loss": 0.3917666971683502,
    "learning_rate": 0.0005172753206872367,
    "timestep_mean": 465.1875,
    "epsilon_norm": 8.039916038513184,
    "prediction_norm": 6.82476282119751,
    "timestamp": "2025-07-14T12:34:08.311113"
  },
  {
    "epoch": 978,
    "batch_idx": 96,
    "loss": 0.2880597710609436,
    "learning_rate": 0.0005172753206872367,
    "timestep_mean": 448.25,
    "epsilon_norm": 7.588765621185303,
    "prediction_norm": 6.713895320892334,
    "timestamp": "2025-07-14T12:34:08.316302"
  },
  {
    "epoch": 978,
    "batch_idx": 128,
    "loss": 0.39454272389411926,
    "learning_rate": 0.0005172753206872367,
    "timestep_mean": 421.6875,
    "epsilon_norm": 9.249309539794922,
    "prediction_norm": 8.286426544189453,
    "timestamp": "2025-07-14T12:34:08.321548"
  },
  {
    "epoch": 978,
    "batch_idx": 160,
    "loss": 0.290455162525177,
    "learning_rate": 0.0005172753206872367,
    "timestep_mean": 453.0,
    "epsilon_norm": 9.123011589050293,
    "prediction_norm": 8.184815406799316,
    "timestamp": "2025-07-14T12:34:08.326694"
  },
  {
    "epoch": 978,
    "batch_idx": 192,
    "loss": 0.20655646920204163,
    "learning_rate": 0.0005172753206872367,
    "timestep_mean": 384.375,
    "epsilon_norm": 3.356376886367798,
    "prediction_norm": 3.0566153526306152,
    "timestamp": "2025-07-14T12:34:08.331946"
  },
  {
    "epoch": 979,
    "batch_idx": 0,
    "loss": 0.4033324122428894,
    "learning_rate": 0.0005164903704573255,
    "timestep_mean": 467.0625,
    "epsilon_norm": 8.347545623779297,
    "prediction_norm": 6.8455729484558105,
    "timestamp": "2025-07-14T12:34:08.337304"
  },
  {
    "epoch": 979,
    "batch_idx": 32,
    "loss": 0.2372727245092392,
    "learning_rate": 0.0005164903704573255,
    "timestep_mean": 473.34375,
    "epsilon_norm": 8.125080108642578,
    "prediction_norm": 7.378492832183838,
    "timestamp": "2025-07-14T12:34:08.342622"
  },
  {
    "epoch": 979,
    "batch_idx": 64,
    "loss": 0.1937122941017151,
    "learning_rate": 0.0005164903704573255,
    "timestep_mean": 491.21875,
    "epsilon_norm": 7.513091087341309,
    "prediction_norm": 6.611926078796387,
    "timestamp": "2025-07-14T12:34:08.347830"
  },
  {
    "epoch": 979,
    "batch_idx": 96,
    "loss": 0.36540308594703674,
    "learning_rate": 0.0005164903704573255,
    "timestep_mean": 477.625,
    "epsilon_norm": 8.68910026550293,
    "prediction_norm": 7.420283317565918,
    "timestamp": "2025-07-14T12:34:08.353139"
  },
  {
    "epoch": 979,
    "batch_idx": 128,
    "loss": 0.11715364456176758,
    "learning_rate": 0.0005164903704573255,
    "timestep_mean": 495.09375,
    "epsilon_norm": 6.897665500640869,
    "prediction_norm": 6.577270030975342,
    "timestamp": "2025-07-14T12:34:08.358386"
  },
  {
    "epoch": 979,
    "batch_idx": 160,
    "loss": 0.23849274218082428,
    "learning_rate": 0.0005164903704573255,
    "timestep_mean": 495.75,
    "epsilon_norm": 8.782958030700684,
    "prediction_norm": 7.943387508392334,
    "timestamp": "2025-07-14T12:34:08.363600"
  },
  {
    "epoch": 979,
    "batch_idx": 192,
    "loss": 0.601776123046875,
    "learning_rate": 0.0005164903704573255,
    "timestep_mean": 434.25,
    "epsilon_norm": 4.995127201080322,
    "prediction_norm": 3.2569377422332764,
    "timestamp": "2025-07-14T12:34:08.368812"
  },
  {
    "epoch": 980,
    "batch_idx": 0,
    "loss": 0.2505185902118683,
    "learning_rate": 0.0005157053795390645,
    "timestep_mean": 489.03125,
    "epsilon_norm": 7.777430057525635,
    "prediction_norm": 6.95212984085083,
    "timestamp": "2025-07-14T12:34:08.374112"
  },
  {
    "epoch": 980,
    "batch_idx": 32,
    "loss": 0.30709072947502136,
    "learning_rate": 0.0005157053795390645,
    "timestep_mean": 430.40625,
    "epsilon_norm": 7.5518999099731445,
    "prediction_norm": 6.230405807495117,
    "timestamp": "2025-07-14T12:34:08.379397"
  },
  {
    "epoch": 980,
    "batch_idx": 64,
    "loss": 0.3310435116291046,
    "learning_rate": 0.0005157053795390645,
    "timestep_mean": 478.75,
    "epsilon_norm": 7.680339336395264,
    "prediction_norm": 6.315695285797119,
    "timestamp": "2025-07-14T12:34:08.384577"
  },
  {
    "epoch": 980,
    "batch_idx": 96,
    "loss": 0.37809789180755615,
    "learning_rate": 0.0005157053795390645,
    "timestep_mean": 440.75,
    "epsilon_norm": 8.80791187286377,
    "prediction_norm": 6.830314636230469,
    "timestamp": "2025-07-14T12:34:08.389823"
  },
  {
    "epoch": 980,
    "batch_idx": 128,
    "loss": 0.445006787776947,
    "learning_rate": 0.0005157053795390645,
    "timestep_mean": 454.6875,
    "epsilon_norm": 7.617585182189941,
    "prediction_norm": 6.1203083992004395,
    "timestamp": "2025-07-14T12:34:08.395048"
  },
  {
    "epoch": 980,
    "batch_idx": 160,
    "loss": 0.42437002062797546,
    "learning_rate": 0.0005157053795390645,
    "timestep_mean": 428.09375,
    "epsilon_norm": 7.299773693084717,
    "prediction_norm": 5.567032337188721,
    "timestamp": "2025-07-14T12:34:08.400353"
  },
  {
    "epoch": 980,
    "batch_idx": 192,
    "loss": 0.2113415002822876,
    "learning_rate": 0.0005157053795390645,
    "timestep_mean": 438.875,
    "epsilon_norm": 4.251589298248291,
    "prediction_norm": 3.894615888595581,
    "timestamp": "2025-07-14T12:34:08.405235"
  },
  {
    "epoch": 981,
    "batch_idx": 0,
    "loss": 0.18316149711608887,
    "learning_rate": 0.000514920349869341,
    "timestep_mean": 520.625,
    "epsilon_norm": 7.5307207107543945,
    "prediction_norm": 7.010834693908691,
    "timestamp": "2025-07-14T12:34:08.410186"
  },
  {
    "epoch": 981,
    "batch_idx": 32,
    "loss": 0.37439560890197754,
    "learning_rate": 0.000514920349869341,
    "timestep_mean": 547.375,
    "epsilon_norm": 8.621500968933105,
    "prediction_norm": 7.183631896972656,
    "timestamp": "2025-07-14T12:34:08.416105"
  },
  {
    "epoch": 981,
    "batch_idx": 64,
    "loss": 0.3604070246219635,
    "learning_rate": 0.000514920349869341,
    "timestep_mean": 431.1875,
    "epsilon_norm": 7.705262660980225,
    "prediction_norm": 5.841609001159668,
    "timestamp": "2025-07-14T12:34:08.421538"
  },
  {
    "epoch": 981,
    "batch_idx": 96,
    "loss": 0.24206095933914185,
    "learning_rate": 0.000514920349869341,
    "timestep_mean": 397.5,
    "epsilon_norm": 7.12542724609375,
    "prediction_norm": 6.010859489440918,
    "timestamp": "2025-07-14T12:34:08.426645"
  },
  {
    "epoch": 981,
    "batch_idx": 128,
    "loss": 0.1566997617483139,
    "learning_rate": 0.000514920349869341,
    "timestep_mean": 567.9375,
    "epsilon_norm": 8.12781047821045,
    "prediction_norm": 7.390621662139893,
    "timestamp": "2025-07-14T12:34:08.434798"
  },
  {
    "epoch": 981,
    "batch_idx": 160,
    "loss": 0.42848941683769226,
    "learning_rate": 0.000514920349869341,
    "timestep_mean": 502.59375,
    "epsilon_norm": 7.770195484161377,
    "prediction_norm": 5.674474239349365,
    "timestamp": "2025-07-14T12:34:08.440472"
  },
  {
    "epoch": 981,
    "batch_idx": 192,
    "loss": 0.39178526401519775,
    "learning_rate": 0.000514920349869341,
    "timestep_mean": 683.375,
    "epsilon_norm": 3.2442405223846436,
    "prediction_norm": 2.74863600730896,
    "timestamp": "2025-07-14T12:34:08.446205"
  },
  {
    "epoch": 982,
    "batch_idx": 0,
    "loss": 0.3268119692802429,
    "learning_rate": 0.0005141352833851371,
    "timestep_mean": 422.25,
    "epsilon_norm": 8.822392463684082,
    "prediction_norm": 6.519696235656738,
    "timestamp": "2025-07-14T12:34:08.451422"
  },
  {
    "epoch": 982,
    "batch_idx": 32,
    "loss": 0.15211576223373413,
    "learning_rate": 0.0005141352833851371,
    "timestep_mean": 634.8125,
    "epsilon_norm": 7.917142391204834,
    "prediction_norm": 7.147680282592773,
    "timestamp": "2025-07-14T12:34:08.456791"
  },
  {
    "epoch": 982,
    "batch_idx": 64,
    "loss": 0.44589272141456604,
    "learning_rate": 0.0005141352833851371,
    "timestep_mean": 376.4375,
    "epsilon_norm": 8.922607421875,
    "prediction_norm": 6.876376628875732,
    "timestamp": "2025-07-14T12:34:08.461991"
  },
  {
    "epoch": 982,
    "batch_idx": 96,
    "loss": 0.26452258229255676,
    "learning_rate": 0.0005141352833851371,
    "timestep_mean": 490.875,
    "epsilon_norm": 7.404710292816162,
    "prediction_norm": 6.275874614715576,
    "timestamp": "2025-07-14T12:34:08.467342"
  },
  {
    "epoch": 982,
    "batch_idx": 128,
    "loss": 0.24067169427871704,
    "learning_rate": 0.0005141352833851371,
    "timestep_mean": 529.3125,
    "epsilon_norm": 7.253108024597168,
    "prediction_norm": 6.406254768371582,
    "timestamp": "2025-07-14T12:34:08.472617"
  },
  {
    "epoch": 982,
    "batch_idx": 160,
    "loss": 0.18568149209022522,
    "learning_rate": 0.0005141352833851371,
    "timestep_mean": 545.03125,
    "epsilon_norm": 7.055182456970215,
    "prediction_norm": 6.635506629943848,
    "timestamp": "2025-07-14T12:34:08.477965"
  },
  {
    "epoch": 982,
    "batch_idx": 192,
    "loss": 0.04957905411720276,
    "learning_rate": 0.0005141352833851371,
    "timestep_mean": 621.25,
    "epsilon_norm": 3.8432414531707764,
    "prediction_norm": 3.9057071208953857,
    "timestamp": "2025-07-14T12:34:08.483253"
  },
  {
    "epoch": 983,
    "batch_idx": 0,
    "loss": 0.4212867021560669,
    "learning_rate": 0.0005133501820235267,
    "timestep_mean": 436.375,
    "epsilon_norm": 8.106471061706543,
    "prediction_norm": 6.303081035614014,
    "timestamp": "2025-07-14T12:34:08.488400"
  },
  {
    "epoch": 983,
    "batch_idx": 32,
    "loss": 0.19186997413635254,
    "learning_rate": 0.0005133501820235267,
    "timestep_mean": 519.3125,
    "epsilon_norm": 7.461019992828369,
    "prediction_norm": 6.335024833679199,
    "timestamp": "2025-07-14T12:34:08.493724"
  },
  {
    "epoch": 983,
    "batch_idx": 64,
    "loss": 0.3569895327091217,
    "learning_rate": 0.0005133501820235267,
    "timestep_mean": 543.9375,
    "epsilon_norm": 9.660181999206543,
    "prediction_norm": 8.114755630493164,
    "timestamp": "2025-07-14T12:34:08.498981"
  },
  {
    "epoch": 983,
    "batch_idx": 96,
    "loss": 0.2679872512817383,
    "learning_rate": 0.0005133501820235267,
    "timestep_mean": 485.03125,
    "epsilon_norm": 8.553582191467285,
    "prediction_norm": 7.482783794403076,
    "timestamp": "2025-07-14T12:34:08.504176"
  },
  {
    "epoch": 983,
    "batch_idx": 128,
    "loss": 0.2218511551618576,
    "learning_rate": 0.0005133501820235267,
    "timestep_mean": 435.0625,
    "epsilon_norm": 7.587302207946777,
    "prediction_norm": 6.05272102355957,
    "timestamp": "2025-07-14T12:34:08.509246"
  },
  {
    "epoch": 983,
    "batch_idx": 160,
    "loss": 0.24335791170597076,
    "learning_rate": 0.0005133501820235267,
    "timestep_mean": 478.9375,
    "epsilon_norm": 7.550804138183594,
    "prediction_norm": 6.017685413360596,
    "timestamp": "2025-07-14T12:34:08.514493"
  },
  {
    "epoch": 983,
    "batch_idx": 192,
    "loss": 0.4925535321235657,
    "learning_rate": 0.0005133501820235267,
    "timestep_mean": 467.625,
    "epsilon_norm": 3.832723617553711,
    "prediction_norm": 2.888627290725708,
    "timestamp": "2025-07-14T12:34:08.519679"
  },
  {
    "epoch": 984,
    "batch_idx": 0,
    "loss": 0.3614387810230255,
    "learning_rate": 0.0005125650477216693,
    "timestep_mean": 415.9375,
    "epsilon_norm": 7.9258341789245605,
    "prediction_norm": 6.289642810821533,
    "timestamp": "2025-07-14T12:34:08.524933"
  },
  {
    "epoch": 984,
    "batch_idx": 32,
    "loss": 0.27584367990493774,
    "learning_rate": 0.0005125650477216693,
    "timestep_mean": 437.03125,
    "epsilon_norm": 7.6807169914245605,
    "prediction_norm": 5.825899600982666,
    "timestamp": "2025-07-14T12:34:08.530029"
  },
  {
    "epoch": 984,
    "batch_idx": 64,
    "loss": 0.4732704162597656,
    "learning_rate": 0.0005125650477216693,
    "timestep_mean": 484.4375,
    "epsilon_norm": 8.869827270507812,
    "prediction_norm": 7.064789295196533,
    "timestamp": "2025-07-14T12:34:08.535068"
  },
  {
    "epoch": 984,
    "batch_idx": 96,
    "loss": 0.34539079666137695,
    "learning_rate": 0.0005125650477216693,
    "timestep_mean": 500.0625,
    "epsilon_norm": 7.791895866394043,
    "prediction_norm": 6.247109413146973,
    "timestamp": "2025-07-14T12:34:08.540251"
  },
  {
    "epoch": 984,
    "batch_idx": 128,
    "loss": 0.27660757303237915,
    "learning_rate": 0.0005125650477216693,
    "timestep_mean": 501.96875,
    "epsilon_norm": 9.348095893859863,
    "prediction_norm": 8.271668434143066,
    "timestamp": "2025-07-14T12:34:08.545819"
  },
  {
    "epoch": 984,
    "batch_idx": 160,
    "loss": 0.12619854509830475,
    "learning_rate": 0.0005125650477216693,
    "timestep_mean": 577.90625,
    "epsilon_norm": 8.101799964904785,
    "prediction_norm": 7.060805797576904,
    "timestamp": "2025-07-14T12:34:08.553117"
  },
  {
    "epoch": 984,
    "batch_idx": 192,
    "loss": 0.12663030624389648,
    "learning_rate": 0.0005125650477216693,
    "timestep_mean": 594.25,
    "epsilon_norm": 3.4463815689086914,
    "prediction_norm": 3.279780387878418,
    "timestamp": "2025-07-14T12:34:08.558476"
  },
  {
    "epoch": 985,
    "batch_idx": 0,
    "loss": 0.10009048879146576,
    "learning_rate": 0.0005117798824168056,
    "timestep_mean": 619.71875,
    "epsilon_norm": 6.6057658195495605,
    "prediction_norm": 6.586609363555908,
    "timestamp": "2025-07-14T12:34:08.563663"
  },
  {
    "epoch": 985,
    "batch_idx": 32,
    "loss": 0.3430708944797516,
    "learning_rate": 0.0005117798824168056,
    "timestep_mean": 520.8125,
    "epsilon_norm": 9.087118148803711,
    "prediction_norm": 8.126355171203613,
    "timestamp": "2025-07-14T12:34:08.568905"
  },
  {
    "epoch": 985,
    "batch_idx": 64,
    "loss": 0.14904329180717468,
    "learning_rate": 0.0005117798824168056,
    "timestep_mean": 563.1875,
    "epsilon_norm": 9.09625244140625,
    "prediction_norm": 8.621895790100098,
    "timestamp": "2025-07-14T12:34:08.574202"
  },
  {
    "epoch": 985,
    "batch_idx": 96,
    "loss": 0.2535445988178253,
    "learning_rate": 0.0005117798824168056,
    "timestep_mean": 477.9375,
    "epsilon_norm": 8.102489471435547,
    "prediction_norm": 7.142340660095215,
    "timestamp": "2025-07-14T12:34:08.579477"
  },
  {
    "epoch": 985,
    "batch_idx": 128,
    "loss": 0.20956197381019592,
    "learning_rate": 0.0005117798824168056,
    "timestep_mean": 591.5,
    "epsilon_norm": 8.02782917022705,
    "prediction_norm": 6.669100761413574,
    "timestamp": "2025-07-14T12:34:08.584689"
  },
  {
    "epoch": 985,
    "batch_idx": 160,
    "loss": 0.20752879977226257,
    "learning_rate": 0.0005117798824168056,
    "timestep_mean": 541.65625,
    "epsilon_norm": 7.653443336486816,
    "prediction_norm": 6.7535080909729,
    "timestamp": "2025-07-14T12:34:08.589857"
  },
  {
    "epoch": 985,
    "batch_idx": 192,
    "loss": 0.46238642930984497,
    "learning_rate": 0.0005117798824168056,
    "timestep_mean": 407.875,
    "epsilon_norm": 3.627054452896118,
    "prediction_norm": 2.5621933937072754,
    "timestamp": "2025-07-14T12:34:08.595262"
  },
  {
    "epoch": 986,
    "batch_idx": 0,
    "loss": 0.22564207017421722,
    "learning_rate": 0.0005109946880462531,
    "timestep_mean": 549.125,
    "epsilon_norm": 7.492011547088623,
    "prediction_norm": 6.464896202087402,
    "timestamp": "2025-07-14T12:34:08.600512"
  },
  {
    "epoch": 986,
    "batch_idx": 32,
    "loss": 0.326171338558197,
    "learning_rate": 0.0005109946880462531,
    "timestep_mean": 494.3125,
    "epsilon_norm": 8.158092498779297,
    "prediction_norm": 6.300261974334717,
    "timestamp": "2025-07-14T12:34:08.605664"
  },
  {
    "epoch": 986,
    "batch_idx": 64,
    "loss": 0.4160703420639038,
    "learning_rate": 0.0005109946880462531,
    "timestep_mean": 489.9375,
    "epsilon_norm": 7.5939507484436035,
    "prediction_norm": 5.784802436828613,
    "timestamp": "2025-07-14T12:34:08.610709"
  },
  {
    "epoch": 986,
    "batch_idx": 96,
    "loss": 0.4234972596168518,
    "learning_rate": 0.0005109946880462531,
    "timestep_mean": 430.21875,
    "epsilon_norm": 8.30246353149414,
    "prediction_norm": 6.463381767272949,
    "timestamp": "2025-07-14T12:34:08.615911"
  },
  {
    "epoch": 986,
    "batch_idx": 128,
    "loss": 0.21161682903766632,
    "learning_rate": 0.0005109946880462531,
    "timestep_mean": 478.25,
    "epsilon_norm": 7.6214094161987305,
    "prediction_norm": 6.654824733734131,
    "timestamp": "2025-07-14T12:34:08.621316"
  },
  {
    "epoch": 986,
    "batch_idx": 160,
    "loss": 0.13025251030921936,
    "learning_rate": 0.0005109946880462531,
    "timestep_mean": 549.75,
    "epsilon_norm": 7.9275665283203125,
    "prediction_norm": 7.85845422744751,
    "timestamp": "2025-07-14T12:34:08.626459"
  },
  {
    "epoch": 986,
    "batch_idx": 192,
    "loss": 0.08150994777679443,
    "learning_rate": 0.0005109946880462531,
    "timestep_mean": 682.375,
    "epsilon_norm": 4.308132171630859,
    "prediction_norm": 4.105673789978027,
    "timestamp": "2025-07-14T12:34:08.631706"
  },
  {
    "epoch": 987,
    "batch_idx": 0,
    "loss": 0.23295769095420837,
    "learning_rate": 0.0005102094665474008,
    "timestep_mean": 487.34375,
    "epsilon_norm": 7.3161797523498535,
    "prediction_norm": 6.546666622161865,
    "timestamp": "2025-07-14T12:34:08.637019"
  },
  {
    "epoch": 987,
    "batch_idx": 32,
    "loss": 0.2127029150724411,
    "learning_rate": 0.0005102094665474008,
    "timestep_mean": 513.5625,
    "epsilon_norm": 6.688041687011719,
    "prediction_norm": 6.347897052764893,
    "timestamp": "2025-07-14T12:34:08.642297"
  },
  {
    "epoch": 987,
    "batch_idx": 64,
    "loss": 0.40076345205307007,
    "learning_rate": 0.0005102094665474008,
    "timestep_mean": 433.1875,
    "epsilon_norm": 8.782269477844238,
    "prediction_norm": 7.353084564208984,
    "timestamp": "2025-07-14T12:34:08.647369"
  },
  {
    "epoch": 987,
    "batch_idx": 96,
    "loss": 0.2934827208518982,
    "learning_rate": 0.0005102094665474008,
    "timestep_mean": 527.46875,
    "epsilon_norm": 7.804906368255615,
    "prediction_norm": 6.477499008178711,
    "timestamp": "2025-07-14T12:34:08.652394"
  },
  {
    "epoch": 987,
    "batch_idx": 128,
    "loss": 0.1965692937374115,
    "learning_rate": 0.0005102094665474008,
    "timestep_mean": 545.25,
    "epsilon_norm": 8.74089241027832,
    "prediction_norm": 7.665616512298584,
    "timestamp": "2025-07-14T12:34:08.657425"
  },
  {
    "epoch": 987,
    "batch_idx": 160,
    "loss": 0.5197513699531555,
    "learning_rate": 0.0005102094665474008,
    "timestep_mean": 487.40625,
    "epsilon_norm": 9.08150863647461,
    "prediction_norm": 7.03122615814209,
    "timestamp": "2025-07-14T12:34:08.662740"
  },
  {
    "epoch": 987,
    "batch_idx": 192,
    "loss": 0.14957989752292633,
    "learning_rate": 0.0005102094665474008,
    "timestep_mean": 615.375,
    "epsilon_norm": 3.3341596126556396,
    "prediction_norm": 2.9565277099609375,
    "timestamp": "2025-07-14T12:34:08.668114"
  },
  {
    "epoch": 988,
    "batch_idx": 0,
    "loss": 0.327228844165802,
    "learning_rate": 0.0005094242198577047,
    "timestep_mean": 518.625,
    "epsilon_norm": 9.966859817504883,
    "prediction_norm": 8.14384937286377,
    "timestamp": "2025-07-14T12:34:08.673370"
  },
  {
    "epoch": 988,
    "batch_idx": 32,
    "loss": 0.3395264446735382,
    "learning_rate": 0.0005094242198577047,
    "timestep_mean": 438.28125,
    "epsilon_norm": 7.552860736846924,
    "prediction_norm": 6.183077812194824,
    "timestamp": "2025-07-14T12:34:08.679595"
  },
  {
    "epoch": 988,
    "batch_idx": 64,
    "loss": 0.24773909151554108,
    "learning_rate": 0.0005094242198577047,
    "timestep_mean": 527.1875,
    "epsilon_norm": 9.473685264587402,
    "prediction_norm": 8.432320594787598,
    "timestamp": "2025-07-14T12:34:08.684921"
  },
  {
    "epoch": 988,
    "batch_idx": 96,
    "loss": 0.34497717022895813,
    "learning_rate": 0.0005094242198577047,
    "timestep_mean": 442.09375,
    "epsilon_norm": 8.019335746765137,
    "prediction_norm": 6.125747203826904,
    "timestamp": "2025-07-14T12:34:08.690203"
  },
  {
    "epoch": 988,
    "batch_idx": 128,
    "loss": 0.3900705575942993,
    "learning_rate": 0.0005094242198577047,
    "timestep_mean": 475.96875,
    "epsilon_norm": 8.054953575134277,
    "prediction_norm": 6.184767723083496,
    "timestamp": "2025-07-14T12:34:08.695443"
  },
  {
    "epoch": 988,
    "batch_idx": 160,
    "loss": 0.373468279838562,
    "learning_rate": 0.0005094242198577047,
    "timestep_mean": 548.375,
    "epsilon_norm": 6.673395156860352,
    "prediction_norm": 6.046946048736572,
    "timestamp": "2025-07-14T12:34:08.700686"
  },
  {
    "epoch": 988,
    "batch_idx": 192,
    "loss": 0.5020684003829956,
    "learning_rate": 0.0005094242198577047,
    "timestep_mean": 465.75,
    "epsilon_norm": 5.10400915145874,
    "prediction_norm": 3.613011598587036,
    "timestamp": "2025-07-14T12:34:08.705925"
  },
  {
    "epoch": 989,
    "batch_idx": 0,
    "loss": 0.24438461661338806,
    "learning_rate": 0.0005086389499146828,
    "timestep_mean": 536.65625,
    "epsilon_norm": 7.701233863830566,
    "prediction_norm": 5.887441158294678,
    "timestamp": "2025-07-14T12:34:08.711218"
  },
  {
    "epoch": 989,
    "batch_idx": 32,
    "loss": 0.3014480769634247,
    "learning_rate": 0.0005086389499146828,
    "timestep_mean": 476.78125,
    "epsilon_norm": 7.699462890625,
    "prediction_norm": 6.766271114349365,
    "timestamp": "2025-07-14T12:34:08.716436"
  },
  {
    "epoch": 989,
    "batch_idx": 64,
    "loss": 0.1796686351299286,
    "learning_rate": 0.0005086389499146828,
    "timestep_mean": 485.15625,
    "epsilon_norm": 6.4537434577941895,
    "prediction_norm": 5.685027122497559,
    "timestamp": "2025-07-14T12:34:08.721706"
  },
  {
    "epoch": 989,
    "batch_idx": 96,
    "loss": 0.12257395684719086,
    "learning_rate": 0.0005086389499146828,
    "timestep_mean": 570.46875,
    "epsilon_norm": 7.1362690925598145,
    "prediction_norm": 6.423306941986084,
    "timestamp": "2025-07-14T12:34:08.726726"
  },
  {
    "epoch": 989,
    "batch_idx": 128,
    "loss": 0.18798375129699707,
    "learning_rate": 0.0005086389499146828,
    "timestep_mean": 538.25,
    "epsilon_norm": 6.8206095695495605,
    "prediction_norm": 5.8405656814575195,
    "timestamp": "2025-07-14T12:34:08.731649"
  },
  {
    "epoch": 989,
    "batch_idx": 160,
    "loss": 0.4550365209579468,
    "learning_rate": 0.0005086389499146828,
    "timestep_mean": 431.5625,
    "epsilon_norm": 8.59144115447998,
    "prediction_norm": 7.670924186706543,
    "timestamp": "2025-07-14T12:34:08.736578"
  },
  {
    "epoch": 989,
    "batch_idx": 192,
    "loss": 0.25086134672164917,
    "learning_rate": 0.0005086389499146828,
    "timestep_mean": 412.0,
    "epsilon_norm": 3.0284063816070557,
    "prediction_norm": 2.7913730144500732,
    "timestamp": "2025-07-14T12:34:08.741478"
  },
  {
    "epoch": 990,
    "batch_idx": 0,
    "loss": 0.15411332249641418,
    "learning_rate": 0.0005078536586559109,
    "timestep_mean": 462.96875,
    "epsilon_norm": 6.327365398406982,
    "prediction_norm": 6.018529415130615,
    "timestamp": "2025-07-14T12:34:08.746381"
  },
  {
    "epoch": 990,
    "batch_idx": 32,
    "loss": 0.3590201139450073,
    "learning_rate": 0.0005078536586559109,
    "timestep_mean": 527.71875,
    "epsilon_norm": 9.247220039367676,
    "prediction_norm": 8.01737117767334,
    "timestamp": "2025-07-14T12:34:08.751369"
  },
  {
    "epoch": 990,
    "batch_idx": 64,
    "loss": 0.2893112897872925,
    "learning_rate": 0.0005078536586559109,
    "timestep_mean": 450.9375,
    "epsilon_norm": 7.438799858093262,
    "prediction_norm": 5.47005033493042,
    "timestamp": "2025-07-14T12:34:08.756426"
  },
  {
    "epoch": 990,
    "batch_idx": 96,
    "loss": 0.22621816396713257,
    "learning_rate": 0.0005078536586559109,
    "timestep_mean": 458.5625,
    "epsilon_norm": 6.764466285705566,
    "prediction_norm": 5.526893615722656,
    "timestamp": "2025-07-14T12:34:08.762144"
  },
  {
    "epoch": 990,
    "batch_idx": 128,
    "loss": 0.14700639247894287,
    "learning_rate": 0.0005078536586559109,
    "timestep_mean": 584.15625,
    "epsilon_norm": 9.056535720825195,
    "prediction_norm": 8.283075332641602,
    "timestamp": "2025-07-14T12:34:08.769225"
  },
  {
    "epoch": 990,
    "batch_idx": 160,
    "loss": 0.20482954382896423,
    "learning_rate": 0.0005078536586559109,
    "timestep_mean": 544.5625,
    "epsilon_norm": 6.510163307189941,
    "prediction_norm": 5.662240505218506,
    "timestamp": "2025-07-14T12:34:08.774492"
  },
  {
    "epoch": 990,
    "batch_idx": 192,
    "loss": 0.004717142321169376,
    "learning_rate": 0.0005078536586559109,
    "timestep_mean": 874.875,
    "epsilon_norm": 4.1620683670043945,
    "prediction_norm": 4.148344039916992,
    "timestamp": "2025-07-14T12:34:08.779715"
  },
  {
    "epoch": 991,
    "batch_idx": 0,
    "loss": 0.13562613725662231,
    "learning_rate": 0.000507068348019017,
    "timestep_mean": 615.34375,
    "epsilon_norm": 9.349161148071289,
    "prediction_norm": 8.681177139282227,
    "timestamp": "2025-07-14T12:34:08.784856"
  },
  {
    "epoch": 991,
    "batch_idx": 32,
    "loss": 0.3217710256576538,
    "learning_rate": 0.000507068348019017,
    "timestep_mean": 482.59375,
    "epsilon_norm": 7.966233730316162,
    "prediction_norm": 5.977798938751221,
    "timestamp": "2025-07-14T12:34:08.789775"
  },
  {
    "epoch": 991,
    "batch_idx": 64,
    "loss": 0.2765645980834961,
    "learning_rate": 0.000507068348019017,
    "timestep_mean": 427.34375,
    "epsilon_norm": 7.946545124053955,
    "prediction_norm": 6.729957103729248,
    "timestamp": "2025-07-14T12:34:08.794688"
  },
  {
    "epoch": 991,
    "batch_idx": 96,
    "loss": 0.2870970666408539,
    "learning_rate": 0.000507068348019017,
    "timestep_mean": 525.0625,
    "epsilon_norm": 8.162999153137207,
    "prediction_norm": 6.852023601531982,
    "timestamp": "2025-07-14T12:34:08.799605"
  },
  {
    "epoch": 991,
    "batch_idx": 128,
    "loss": 0.32756853103637695,
    "learning_rate": 0.000507068348019017,
    "timestep_mean": 547.96875,
    "epsilon_norm": 9.010519981384277,
    "prediction_norm": 7.77376127243042,
    "timestamp": "2025-07-14T12:34:08.804531"
  },
  {
    "epoch": 991,
    "batch_idx": 160,
    "loss": 0.3753098249435425,
    "learning_rate": 0.000507068348019017,
    "timestep_mean": 452.3125,
    "epsilon_norm": 7.450099945068359,
    "prediction_norm": 5.781729698181152,
    "timestamp": "2025-07-14T12:34:08.809528"
  },
  {
    "epoch": 991,
    "batch_idx": 192,
    "loss": 0.1585153192281723,
    "learning_rate": 0.000507068348019017,
    "timestep_mean": 442.5,
    "epsilon_norm": 3.3875250816345215,
    "prediction_norm": 2.857001304626465,
    "timestamp": "2025-07-14T12:34:08.814516"
  },
  {
    "epoch": 992,
    "batch_idx": 0,
    "loss": 0.2618464529514313,
    "learning_rate": 0.0005062830199416768,
    "timestep_mean": 517.0625,
    "epsilon_norm": 8.70167064666748,
    "prediction_norm": 6.805221080780029,
    "timestamp": "2025-07-14T12:34:08.819894"
  },
  {
    "epoch": 992,
    "batch_idx": 32,
    "loss": 0.28772643208503723,
    "learning_rate": 0.0005062830199416768,
    "timestep_mean": 406.9375,
    "epsilon_norm": 8.29403305053711,
    "prediction_norm": 6.903675079345703,
    "timestamp": "2025-07-14T12:34:08.824970"
  },
  {
    "epoch": 992,
    "batch_idx": 64,
    "loss": 0.21455463767051697,
    "learning_rate": 0.0005062830199416768,
    "timestep_mean": 431.84375,
    "epsilon_norm": 7.932247161865234,
    "prediction_norm": 6.577383995056152,
    "timestamp": "2025-07-14T12:34:08.830016"
  },
  {
    "epoch": 992,
    "batch_idx": 96,
    "loss": 0.1421075463294983,
    "learning_rate": 0.0005062830199416768,
    "timestep_mean": 494.5,
    "epsilon_norm": 7.9498291015625,
    "prediction_norm": 7.1060872077941895,
    "timestamp": "2025-07-14T12:34:08.835031"
  },
  {
    "epoch": 992,
    "batch_idx": 128,
    "loss": 0.4851098656654358,
    "learning_rate": 0.0005062830199416768,
    "timestep_mean": 514.1875,
    "epsilon_norm": 8.93415641784668,
    "prediction_norm": 6.828304767608643,
    "timestamp": "2025-07-14T12:34:08.839914"
  },
  {
    "epoch": 992,
    "batch_idx": 160,
    "loss": 0.2447405606508255,
    "learning_rate": 0.0005062830199416768,
    "timestep_mean": 552.21875,
    "epsilon_norm": 7.8492841720581055,
    "prediction_norm": 6.500453472137451,
    "timestamp": "2025-07-14T12:34:08.844898"
  },
  {
    "epoch": 992,
    "batch_idx": 192,
    "loss": 0.18334299325942993,
    "learning_rate": 0.0005062830199416768,
    "timestep_mean": 490.375,
    "epsilon_norm": 2.692800283432007,
    "prediction_norm": 2.37007212638855,
    "timestamp": "2025-07-14T12:34:08.849924"
  },
  {
    "epoch": 993,
    "batch_idx": 0,
    "loss": 0.30989208817481995,
    "learning_rate": 0.0005054976763616097,
    "timestep_mean": 539.0,
    "epsilon_norm": 7.8516106605529785,
    "prediction_norm": 6.276637554168701,
    "timestamp": "2025-07-14T12:34:08.854976"
  },
  {
    "epoch": 993,
    "batch_idx": 32,
    "loss": 0.20178142189979553,
    "learning_rate": 0.0005054976763616097,
    "timestep_mean": 492.1875,
    "epsilon_norm": 7.107688903808594,
    "prediction_norm": 6.129885196685791,
    "timestamp": "2025-07-14T12:34:08.860235"
  },
  {
    "epoch": 993,
    "batch_idx": 64,
    "loss": 0.378969669342041,
    "learning_rate": 0.0005054976763616097,
    "timestep_mean": 461.90625,
    "epsilon_norm": 7.20943546295166,
    "prediction_norm": 5.4561567306518555,
    "timestamp": "2025-07-14T12:34:08.865298"
  },
  {
    "epoch": 993,
    "batch_idx": 96,
    "loss": 0.199817955493927,
    "learning_rate": 0.0005054976763616097,
    "timestep_mean": 550.34375,
    "epsilon_norm": 8.795448303222656,
    "prediction_norm": 8.348353385925293,
    "timestamp": "2025-07-14T12:34:08.870252"
  },
  {
    "epoch": 993,
    "batch_idx": 128,
    "loss": 0.1444138139486313,
    "learning_rate": 0.0005054976763616097,
    "timestep_mean": 581.5625,
    "epsilon_norm": 8.261334419250488,
    "prediction_norm": 8.091167449951172,
    "timestamp": "2025-07-14T12:34:08.875282"
  },
  {
    "epoch": 993,
    "batch_idx": 160,
    "loss": 0.33019715547561646,
    "learning_rate": 0.0005054976763616097,
    "timestep_mean": 448.4375,
    "epsilon_norm": 8.79478645324707,
    "prediction_norm": 7.555272579193115,
    "timestamp": "2025-07-14T12:34:08.880293"
  },
  {
    "epoch": 993,
    "batch_idx": 192,
    "loss": 0.4301145672798157,
    "learning_rate": 0.0005054976763616097,
    "timestep_mean": 384.375,
    "epsilon_norm": 3.584698438644409,
    "prediction_norm": 2.772557020187378,
    "timestamp": "2025-07-14T12:34:08.885318"
  },
  {
    "epoch": 994,
    "batch_idx": 0,
    "loss": 0.2945740222930908,
    "learning_rate": 0.0005047123192165725,
    "timestep_mean": 426.90625,
    "epsilon_norm": 8.834136962890625,
    "prediction_norm": 7.400809288024902,
    "timestamp": "2025-07-14T12:34:08.890889"
  },
  {
    "epoch": 994,
    "batch_idx": 32,
    "loss": 0.17199558019638062,
    "learning_rate": 0.0005047123192165725,
    "timestep_mean": 543.375,
    "epsilon_norm": 7.253450393676758,
    "prediction_norm": 6.827958583831787,
    "timestamp": "2025-07-14T12:34:08.896223"
  },
  {
    "epoch": 994,
    "batch_idx": 64,
    "loss": 0.21680398285388947,
    "learning_rate": 0.0005047123192165725,
    "timestep_mean": 527.21875,
    "epsilon_norm": 8.248997688293457,
    "prediction_norm": 6.854979038238525,
    "timestamp": "2025-07-14T12:34:08.902984"
  },
  {
    "epoch": 994,
    "batch_idx": 96,
    "loss": 0.20062974095344543,
    "learning_rate": 0.0005047123192165725,
    "timestep_mean": 460.46875,
    "epsilon_norm": 7.786802291870117,
    "prediction_norm": 7.122000217437744,
    "timestamp": "2025-07-14T12:34:08.908209"
  },
  {
    "epoch": 994,
    "batch_idx": 128,
    "loss": 0.4350805878639221,
    "learning_rate": 0.0005047123192165725,
    "timestep_mean": 399.6875,
    "epsilon_norm": 9.834376335144043,
    "prediction_norm": 7.9480390548706055,
    "timestamp": "2025-07-14T12:34:08.913434"
  },
  {
    "epoch": 994,
    "batch_idx": 160,
    "loss": 0.23987321555614471,
    "learning_rate": 0.0005047123192165725,
    "timestep_mean": 501.0,
    "epsilon_norm": 8.012277603149414,
    "prediction_norm": 7.441967487335205,
    "timestamp": "2025-07-14T12:34:08.918528"
  },
  {
    "epoch": 994,
    "batch_idx": 192,
    "loss": 0.2132611721754074,
    "learning_rate": 0.0005047123192165725,
    "timestep_mean": 276.125,
    "epsilon_norm": 2.4351422786712646,
    "prediction_norm": 2.3220489025115967,
    "timestamp": "2025-07-14T12:34:08.923652"
  },
  {
    "epoch": 995,
    "batch_idx": 0,
    "loss": 0.22388850152492523,
    "learning_rate": 0.0005039269504443562,
    "timestep_mean": 520.875,
    "epsilon_norm": 8.641749382019043,
    "prediction_norm": 7.85150146484375,
    "timestamp": "2025-07-14T12:34:08.928821"
  },
  {
    "epoch": 995,
    "batch_idx": 32,
    "loss": 0.27573588490486145,
    "learning_rate": 0.0005039269504443562,
    "timestep_mean": 490.21875,
    "epsilon_norm": 7.741577625274658,
    "prediction_norm": 6.405861854553223,
    "timestamp": "2025-07-14T12:34:08.933835"
  },
  {
    "epoch": 995,
    "batch_idx": 64,
    "loss": 0.257942795753479,
    "learning_rate": 0.0005039269504443562,
    "timestep_mean": 546.09375,
    "epsilon_norm": 7.371294975280762,
    "prediction_norm": 6.831535816192627,
    "timestamp": "2025-07-14T12:34:08.938760"
  },
  {
    "epoch": 995,
    "batch_idx": 96,
    "loss": 0.3043168783187866,
    "learning_rate": 0.0005039269504443562,
    "timestep_mean": 409.34375,
    "epsilon_norm": 7.165371894836426,
    "prediction_norm": 5.361634254455566,
    "timestamp": "2025-07-14T12:34:08.943781"
  },
  {
    "epoch": 995,
    "batch_idx": 128,
    "loss": 0.7011533975601196,
    "learning_rate": 0.0005039269504443562,
    "timestep_mean": 445.375,
    "epsilon_norm": 9.497420310974121,
    "prediction_norm": 7.0709757804870605,
    "timestamp": "2025-07-14T12:34:08.948835"
  },
  {
    "epoch": 995,
    "batch_idx": 160,
    "loss": 0.2666521668434143,
    "learning_rate": 0.0005039269504443562,
    "timestep_mean": 454.4375,
    "epsilon_norm": 6.078821659088135,
    "prediction_norm": 4.905789375305176,
    "timestamp": "2025-07-14T12:34:08.953864"
  },
  {
    "epoch": 995,
    "batch_idx": 192,
    "loss": 0.374972939491272,
    "learning_rate": 0.0005039269504443562,
    "timestep_mean": 614.0,
    "epsilon_norm": 3.3685100078582764,
    "prediction_norm": 2.3274295330047607,
    "timestamp": "2025-07-14T12:34:08.958992"
  },
  {
    "epoch": 996,
    "batch_idx": 0,
    "loss": 0.3026806116104126,
    "learning_rate": 0.0005031415719827801,
    "timestep_mean": 506.15625,
    "epsilon_norm": 8.67258071899414,
    "prediction_norm": 6.580273151397705,
    "timestamp": "2025-07-14T12:34:08.963979"
  },
  {
    "epoch": 996,
    "batch_idx": 32,
    "loss": 0.2949336767196655,
    "learning_rate": 0.0005031415719827801,
    "timestep_mean": 515.1875,
    "epsilon_norm": 8.617523193359375,
    "prediction_norm": 7.30479621887207,
    "timestamp": "2025-07-14T12:34:08.969023"
  },
  {
    "epoch": 996,
    "batch_idx": 64,
    "loss": 0.3950374126434326,
    "learning_rate": 0.0005031415719827801,
    "timestep_mean": 437.28125,
    "epsilon_norm": 8.79888916015625,
    "prediction_norm": 6.850120544433594,
    "timestamp": "2025-07-14T12:34:08.974276"
  },
  {
    "epoch": 996,
    "batch_idx": 96,
    "loss": 0.3970535695552826,
    "learning_rate": 0.0005031415719827801,
    "timestep_mean": 492.71875,
    "epsilon_norm": 9.57044792175293,
    "prediction_norm": 7.636159420013428,
    "timestamp": "2025-07-14T12:34:08.979712"
  },
  {
    "epoch": 996,
    "batch_idx": 128,
    "loss": 0.36955446004867554,
    "learning_rate": 0.0005031415719827801,
    "timestep_mean": 544.40625,
    "epsilon_norm": 7.942063808441162,
    "prediction_norm": 6.255014896392822,
    "timestamp": "2025-07-14T12:34:08.986349"
  },
  {
    "epoch": 996,
    "batch_idx": 160,
    "loss": 0.19380313158035278,
    "learning_rate": 0.0005031415719827801,
    "timestep_mean": 433.5625,
    "epsilon_norm": 7.680959701538086,
    "prediction_norm": 6.798226833343506,
    "timestamp": "2025-07-14T12:34:08.991708"
  },
  {
    "epoch": 996,
    "batch_idx": 192,
    "loss": 0.2094811499118805,
    "learning_rate": 0.0005031415719827801,
    "timestep_mean": 358.0,
    "epsilon_norm": 3.792358160018921,
    "prediction_norm": 3.174838066101074,
    "timestamp": "2025-07-14T12:34:08.996840"
  },
  {
    "epoch": 997,
    "batch_idx": 0,
    "loss": 0.47146737575531006,
    "learning_rate": 0.0005023561857696872,
    "timestep_mean": 431.5625,
    "epsilon_norm": 8.551862716674805,
    "prediction_norm": 6.391992092132568,
    "timestamp": "2025-07-14T12:34:09.001876"
  },
  {
    "epoch": 997,
    "batch_idx": 32,
    "loss": 0.2858407199382782,
    "learning_rate": 0.0005023561857696872,
    "timestep_mean": 577.21875,
    "epsilon_norm": 9.066543579101562,
    "prediction_norm": 7.3894548416137695,
    "timestamp": "2025-07-14T12:34:09.006911"
  },
  {
    "epoch": 997,
    "batch_idx": 64,
    "loss": 0.2599683701992035,
    "learning_rate": 0.0005023561857696872,
    "timestep_mean": 532.875,
    "epsilon_norm": 7.787042140960693,
    "prediction_norm": 6.861663341522217,
    "timestamp": "2025-07-14T12:34:09.011922"
  },
  {
    "epoch": 997,
    "batch_idx": 96,
    "loss": 0.3264565169811249,
    "learning_rate": 0.0005023561857696872,
    "timestep_mean": 480.90625,
    "epsilon_norm": 7.666697025299072,
    "prediction_norm": 5.978368759155273,
    "timestamp": "2025-07-14T12:34:09.017021"
  },
  {
    "epoch": 997,
    "batch_idx": 128,
    "loss": 0.3461913466453552,
    "learning_rate": 0.0005023561857696872,
    "timestep_mean": 512.0625,
    "epsilon_norm": 8.601120948791504,
    "prediction_norm": 7.031064987182617,
    "timestamp": "2025-07-14T12:34:09.021883"
  },
  {
    "epoch": 997,
    "batch_idx": 160,
    "loss": 0.3286977708339691,
    "learning_rate": 0.0005023561857696872,
    "timestep_mean": 473.71875,
    "epsilon_norm": 7.346218109130859,
    "prediction_norm": 6.222425937652588,
    "timestamp": "2025-07-14T12:34:09.026867"
  },
  {
    "epoch": 997,
    "batch_idx": 192,
    "loss": 0.4451115131378174,
    "learning_rate": 0.0005023561857696872,
    "timestep_mean": 424.25,
    "epsilon_norm": 3.8520846366882324,
    "prediction_norm": 4.173112869262695,
    "timestamp": "2025-07-14T12:34:09.031854"
  },
  {
    "epoch": 998,
    "batch_idx": 0,
    "loss": 0.18994489312171936,
    "learning_rate": 0.0005015707937429403,
    "timestep_mean": 496.875,
    "epsilon_norm": 7.067521572113037,
    "prediction_norm": 6.20922327041626,
    "timestamp": "2025-07-14T12:34:09.036802"
  },
  {
    "epoch": 998,
    "batch_idx": 32,
    "loss": 0.21356657147407532,
    "learning_rate": 0.0005015707937429403,
    "timestep_mean": 521.71875,
    "epsilon_norm": 7.416333198547363,
    "prediction_norm": 6.7920708656311035,
    "timestamp": "2025-07-14T12:34:09.041754"
  },
  {
    "epoch": 998,
    "batch_idx": 64,
    "loss": 0.4805539548397064,
    "learning_rate": 0.0005015707937429403,
    "timestep_mean": 399.09375,
    "epsilon_norm": 8.172086715698242,
    "prediction_norm": 6.752694129943848,
    "timestamp": "2025-07-14T12:34:09.046792"
  },
  {
    "epoch": 998,
    "batch_idx": 96,
    "loss": 0.3144568204879761,
    "learning_rate": 0.0005015707937429403,
    "timestep_mean": 404.90625,
    "epsilon_norm": 8.216005325317383,
    "prediction_norm": 6.198246955871582,
    "timestamp": "2025-07-14T12:34:09.051712"
  },
  {
    "epoch": 998,
    "batch_idx": 128,
    "loss": 0.26881444454193115,
    "learning_rate": 0.0005015707937429403,
    "timestep_mean": 479.96875,
    "epsilon_norm": 7.369835376739502,
    "prediction_norm": 6.939087390899658,
    "timestamp": "2025-07-14T12:34:09.056775"
  },
  {
    "epoch": 998,
    "batch_idx": 160,
    "loss": 0.16339419782161713,
    "learning_rate": 0.0005015707937429403,
    "timestep_mean": 527.15625,
    "epsilon_norm": 7.381047248840332,
    "prediction_norm": 6.378174304962158,
    "timestamp": "2025-07-14T12:34:09.061715"
  },
  {
    "epoch": 998,
    "batch_idx": 192,
    "loss": 0.14653730392456055,
    "learning_rate": 0.0005015707937429403,
    "timestep_mean": 529.75,
    "epsilon_norm": 4.135908126831055,
    "prediction_norm": 3.8827121257781982,
    "timestamp": "2025-07-14T12:34:09.066693"
  },
  {
    "epoch": 999,
    "batch_idx": 0,
    "loss": 0.2357967048883438,
    "learning_rate": 0.0005007853978404161,
    "timestep_mean": 508.25,
    "epsilon_norm": 7.978253364562988,
    "prediction_norm": 6.807233810424805,
    "timestamp": "2025-07-14T12:34:09.071806"
  },
  {
    "epoch": 999,
    "batch_idx": 32,
    "loss": 0.18962261080741882,
    "learning_rate": 0.0005007853978404161,
    "timestep_mean": 480.84375,
    "epsilon_norm": 7.398075103759766,
    "prediction_norm": 6.941407203674316,
    "timestamp": "2025-07-14T12:34:09.076804"
  },
  {
    "epoch": 999,
    "batch_idx": 64,
    "loss": 0.11424051970243454,
    "learning_rate": 0.0005007853978404161,
    "timestep_mean": 566.625,
    "epsilon_norm": 6.124901294708252,
    "prediction_norm": 5.3232340812683105,
    "timestamp": "2025-07-14T12:34:09.081679"
  },
  {
    "epoch": 999,
    "batch_idx": 96,
    "loss": 0.4022882580757141,
    "learning_rate": 0.0005007853978404161,
    "timestep_mean": 503.625,
    "epsilon_norm": 9.467591285705566,
    "prediction_norm": 7.944077491760254,
    "timestamp": "2025-07-14T12:34:09.086577"
  },
  {
    "epoch": 999,
    "batch_idx": 128,
    "loss": 0.23110663890838623,
    "learning_rate": 0.0005007853978404161,
    "timestep_mean": 490.0,
    "epsilon_norm": 6.460345268249512,
    "prediction_norm": 5.7406721115112305,
    "timestamp": "2025-07-14T12:34:09.091452"
  },
  {
    "epoch": 999,
    "batch_idx": 160,
    "loss": 0.11127592623233795,
    "learning_rate": 0.0005007853978404161,
    "timestep_mean": 644.75,
    "epsilon_norm": 8.792834281921387,
    "prediction_norm": 8.406075477600098,
    "timestamp": "2025-07-14T12:34:09.096399"
  },
  {
    "epoch": 999,
    "batch_idx": 192,
    "loss": 0.08865228295326233,
    "learning_rate": 0.0005007853978404161,
    "timestep_mean": 469.75,
    "epsilon_norm": 2.961226463317871,
    "prediction_norm": 2.559051990509033,
    "timestamp": "2025-07-14T12:34:09.101344"
  },
  {
    "epoch": 1000,
    "batch_idx": 0,
    "loss": 0.34070897102355957,
    "learning_rate": 0.0005000000000000006,
    "timestep_mean": 513.28125,
    "epsilon_norm": 7.975164413452148,
    "prediction_norm": 6.195044040679932,
    "timestamp": "2025-07-14T12:34:09.106248"
  },
  {
    "epoch": 1000,
    "batch_idx": 32,
    "loss": 0.13926546275615692,
    "learning_rate": 0.0005000000000000006,
    "timestep_mean": 459.9375,
    "epsilon_norm": 5.737939357757568,
    "prediction_norm": 4.823246002197266,
    "timestamp": "2025-07-14T12:34:09.111184"
  },
  {
    "epoch": 1000,
    "batch_idx": 64,
    "loss": 0.1995117962360382,
    "learning_rate": 0.0005000000000000006,
    "timestep_mean": 503.90625,
    "epsilon_norm": 8.092768669128418,
    "prediction_norm": 6.879179954528809,
    "timestamp": "2025-07-14T12:34:09.116102"
  },
  {
    "epoch": 1000,
    "batch_idx": 96,
    "loss": 0.142293319106102,
    "learning_rate": 0.0005000000000000006,
    "timestep_mean": 543.1875,
    "epsilon_norm": 7.412689685821533,
    "prediction_norm": 6.152246952056885,
    "timestamp": "2025-07-14T12:34:09.121005"
  },
  {
    "epoch": 1000,
    "batch_idx": 128,
    "loss": 0.20865461230278015,
    "learning_rate": 0.0005000000000000006,
    "timestep_mean": 484.25,
    "epsilon_norm": 8.07042121887207,
    "prediction_norm": 7.269585132598877,
    "timestamp": "2025-07-14T12:34:09.125905"
  },
  {
    "epoch": 1000,
    "batch_idx": 160,
    "loss": 0.22235149145126343,
    "learning_rate": 0.0005000000000000006,
    "timestep_mean": 489.84375,
    "epsilon_norm": 8.059030532836914,
    "prediction_norm": 6.416799068450928,
    "timestamp": "2025-07-14T12:34:09.130777"
  },
  {
    "epoch": 1000,
    "batch_idx": 192,
    "loss": 0.03696538507938385,
    "learning_rate": 0.0005000000000000006,
    "timestep_mean": 561.75,
    "epsilon_norm": 3.247382164001465,
    "prediction_norm": 3.0995922088623047,
    "timestamp": "2025-07-14T12:34:09.135656"
  },
  {
    "epoch": 1001,
    "batch_idx": 0,
    "loss": 0.5571684241294861,
    "learning_rate": 0.0004992146021595851,
    "timestep_mean": 555.1875,
    "epsilon_norm": 8.493961334228516,
    "prediction_norm": 6.376267433166504,
    "timestamp": "2025-07-14T12:34:09.140574"
  },
  {
    "epoch": 1001,
    "batch_idx": 32,
    "loss": 0.24836194515228271,
    "learning_rate": 0.0004992146021595851,
    "timestep_mean": 392.1875,
    "epsilon_norm": 7.996040344238281,
    "prediction_norm": 6.751774787902832,
    "timestamp": "2025-07-14T12:34:09.145481"
  },
  {
    "epoch": 1001,
    "batch_idx": 64,
    "loss": 0.47331252694129944,
    "learning_rate": 0.0004992146021595851,
    "timestep_mean": 456.65625,
    "epsilon_norm": 8.670300483703613,
    "prediction_norm": 6.588657379150391,
    "timestamp": "2025-07-14T12:34:09.150391"
  },
  {
    "epoch": 1001,
    "batch_idx": 96,
    "loss": 0.23175597190856934,
    "learning_rate": 0.0004992146021595851,
    "timestep_mean": 430.96875,
    "epsilon_norm": 6.104032039642334,
    "prediction_norm": 5.335453510284424,
    "timestamp": "2025-07-14T12:34:09.155233"
  },
  {
    "epoch": 1001,
    "batch_idx": 128,
    "loss": 0.27290406823158264,
    "learning_rate": 0.0004992146021595851,
    "timestep_mean": 551.5,
    "epsilon_norm": 7.933612823486328,
    "prediction_norm": 6.786494255065918,
    "timestamp": "2025-07-14T12:34:09.160213"
  },
  {
    "epoch": 1001,
    "batch_idx": 160,
    "loss": 0.23454725742340088,
    "learning_rate": 0.0004992146021595851,
    "timestep_mean": 535.90625,
    "epsilon_norm": 8.676234245300293,
    "prediction_norm": 7.429290771484375,
    "timestamp": "2025-07-14T12:34:09.165108"
  },
  {
    "epoch": 1001,
    "batch_idx": 192,
    "loss": 0.4495873749256134,
    "learning_rate": 0.0004992146021595851,
    "timestep_mean": 473.25,
    "epsilon_norm": 4.2599196434021,
    "prediction_norm": 3.062201976776123,
    "timestamp": "2025-07-14T12:34:09.169985"
  },
  {
    "epoch": 1002,
    "batch_idx": 0,
    "loss": 0.3973698914051056,
    "learning_rate": 0.0004984292062570608,
    "timestep_mean": 511.875,
    "epsilon_norm": 8.655305862426758,
    "prediction_norm": 7.704174995422363,
    "timestamp": "2025-07-14T12:34:09.174882"
  },
  {
    "epoch": 1002,
    "batch_idx": 32,
    "loss": 0.2850528359413147,
    "learning_rate": 0.0004984292062570608,
    "timestep_mean": 547.71875,
    "epsilon_norm": 8.34329891204834,
    "prediction_norm": 6.862350940704346,
    "timestamp": "2025-07-14T12:34:09.179885"
  },
  {
    "epoch": 1002,
    "batch_idx": 64,
    "loss": 0.19365328550338745,
    "learning_rate": 0.0004984292062570608,
    "timestep_mean": 496.28125,
    "epsilon_norm": 7.0753374099731445,
    "prediction_norm": 6.70710563659668,
    "timestamp": "2025-07-14T12:34:09.185020"
  },
  {
    "epoch": 1002,
    "batch_idx": 96,
    "loss": 0.42274796962738037,
    "learning_rate": 0.0004984292062570608,
    "timestep_mean": 453.3125,
    "epsilon_norm": 9.196700096130371,
    "prediction_norm": 8.137399673461914,
    "timestamp": "2025-07-14T12:34:09.190125"
  },
  {
    "epoch": 1002,
    "batch_idx": 128,
    "loss": 0.10986761748790741,
    "learning_rate": 0.0004984292062570608,
    "timestep_mean": 619.96875,
    "epsilon_norm": 6.465230464935303,
    "prediction_norm": 6.426758766174316,
    "timestamp": "2025-07-14T12:34:09.195064"
  },
  {
    "epoch": 1002,
    "batch_idx": 160,
    "loss": 0.3055700361728668,
    "learning_rate": 0.0004984292062570608,
    "timestep_mean": 562.21875,
    "epsilon_norm": 7.551322937011719,
    "prediction_norm": 7.559574604034424,
    "timestamp": "2025-07-14T12:34:09.200001"
  },
  {
    "epoch": 1002,
    "batch_idx": 192,
    "loss": 0.43358537554740906,
    "learning_rate": 0.0004984292062570608,
    "timestep_mean": 567.375,
    "epsilon_norm": 4.166239261627197,
    "prediction_norm": 2.924489736557007,
    "timestamp": "2025-07-14T12:34:09.205003"
  },
  {
    "epoch": 1003,
    "batch_idx": 0,
    "loss": 0.30352962017059326,
    "learning_rate": 0.0004976438142303139,
    "timestep_mean": 469.59375,
    "epsilon_norm": 8.06710433959961,
    "prediction_norm": 7.383428573608398,
    "timestamp": "2025-07-14T12:34:09.210145"
  },
  {
    "epoch": 1003,
    "batch_idx": 32,
    "loss": 0.26973721385002136,
    "learning_rate": 0.0004976438142303139,
    "timestep_mean": 451.15625,
    "epsilon_norm": 7.817060470581055,
    "prediction_norm": 6.956577301025391,
    "timestamp": "2025-07-14T12:34:09.218473"
  },
  {
    "epoch": 1003,
    "batch_idx": 64,
    "loss": 0.22998930513858795,
    "learning_rate": 0.0004976438142303139,
    "timestep_mean": 521.90625,
    "epsilon_norm": 7.218781471252441,
    "prediction_norm": 6.888730525970459,
    "timestamp": "2025-07-14T12:34:09.223886"
  },
  {
    "epoch": 1003,
    "batch_idx": 96,
    "loss": 0.11405386030673981,
    "learning_rate": 0.0004976438142303139,
    "timestep_mean": 499.21875,
    "epsilon_norm": 7.293488502502441,
    "prediction_norm": 6.743636608123779,
    "timestamp": "2025-07-14T12:34:09.229394"
  },
  {
    "epoch": 1003,
    "batch_idx": 128,
    "loss": 0.37963154911994934,
    "learning_rate": 0.0004976438142303139,
    "timestep_mean": 421.34375,
    "epsilon_norm": 7.741004943847656,
    "prediction_norm": 6.023322582244873,
    "timestamp": "2025-07-14T12:34:09.234590"
  },
  {
    "epoch": 1003,
    "batch_idx": 160,
    "loss": 0.20987163484096527,
    "learning_rate": 0.0004976438142303139,
    "timestep_mean": 495.53125,
    "epsilon_norm": 8.226725578308105,
    "prediction_norm": 7.271170616149902,
    "timestamp": "2025-07-14T12:34:09.239803"
  },
  {
    "epoch": 1003,
    "batch_idx": 192,
    "loss": 0.06945903599262238,
    "learning_rate": 0.0004976438142303139,
    "timestep_mean": 537.5,
    "epsilon_norm": 3.5592405796051025,
    "prediction_norm": 3.274728775024414,
    "timestamp": "2025-07-14T12:34:09.245265"
  },
  {
    "epoch": 1004,
    "batch_idx": 0,
    "loss": 0.18029683828353882,
    "learning_rate": 0.000496858428017221,
    "timestep_mean": 498.6875,
    "epsilon_norm": 7.722169876098633,
    "prediction_norm": 6.744683265686035,
    "timestamp": "2025-07-14T12:34:09.250493"
  },
  {
    "epoch": 1004,
    "batch_idx": 32,
    "loss": 0.24970191717147827,
    "learning_rate": 0.000496858428017221,
    "timestep_mean": 426.0,
    "epsilon_norm": 7.749207973480225,
    "prediction_norm": 6.2350873947143555,
    "timestamp": "2025-07-14T12:34:09.255601"
  },
  {
    "epoch": 1004,
    "batch_idx": 64,
    "loss": 0.15581710636615753,
    "learning_rate": 0.000496858428017221,
    "timestep_mean": 545.875,
    "epsilon_norm": 6.6987433433532715,
    "prediction_norm": 5.571768283843994,
    "timestamp": "2025-07-14T12:34:09.260877"
  },
  {
    "epoch": 1004,
    "batch_idx": 96,
    "loss": 0.11178894340991974,
    "learning_rate": 0.000496858428017221,
    "timestep_mean": 531.46875,
    "epsilon_norm": 7.775022506713867,
    "prediction_norm": 7.211120128631592,
    "timestamp": "2025-07-14T12:34:09.266043"
  },
  {
    "epoch": 1004,
    "batch_idx": 128,
    "loss": 0.3033350706100464,
    "learning_rate": 0.000496858428017221,
    "timestep_mean": 425.5625,
    "epsilon_norm": 9.145233154296875,
    "prediction_norm": 7.090452671051025,
    "timestamp": "2025-07-14T12:34:09.271273"
  },
  {
    "epoch": 1004,
    "batch_idx": 160,
    "loss": 0.2993263304233551,
    "learning_rate": 0.000496858428017221,
    "timestep_mean": 528.25,
    "epsilon_norm": 8.460471153259277,
    "prediction_norm": 6.642924785614014,
    "timestamp": "2025-07-14T12:34:09.276450"
  },
  {
    "epoch": 1004,
    "batch_idx": 192,
    "loss": 0.34395521879196167,
    "learning_rate": 0.000496858428017221,
    "timestep_mean": 596.5,
    "epsilon_norm": 3.6623260974884033,
    "prediction_norm": 2.304655075073242,
    "timestamp": "2025-07-14T12:34:09.281707"
  },
  {
    "epoch": 1005,
    "batch_idx": 0,
    "loss": 0.13041657209396362,
    "learning_rate": 0.0004960730495556449,
    "timestep_mean": 535.21875,
    "epsilon_norm": 7.25289249420166,
    "prediction_norm": 6.692511558532715,
    "timestamp": "2025-07-14T12:34:09.286913"
  },
  {
    "epoch": 1005,
    "batch_idx": 32,
    "loss": 0.22492563724517822,
    "learning_rate": 0.0004960730495556449,
    "timestep_mean": 501.875,
    "epsilon_norm": 8.58500862121582,
    "prediction_norm": 6.617368698120117,
    "timestamp": "2025-07-14T12:34:09.292274"
  },
  {
    "epoch": 1005,
    "batch_idx": 64,
    "loss": 0.2650188207626343,
    "learning_rate": 0.0004960730495556449,
    "timestep_mean": 469.8125,
    "epsilon_norm": 8.257107734680176,
    "prediction_norm": 6.958622932434082,
    "timestamp": "2025-07-14T12:34:09.297434"
  },
  {
    "epoch": 1005,
    "batch_idx": 96,
    "loss": 0.22368517518043518,
    "learning_rate": 0.0004960730495556449,
    "timestep_mean": 484.28125,
    "epsilon_norm": 7.074783802032471,
    "prediction_norm": 5.570549488067627,
    "timestamp": "2025-07-14T12:34:09.302808"
  },
  {
    "epoch": 1005,
    "batch_idx": 128,
    "loss": 0.27358710765838623,
    "learning_rate": 0.0004960730495556449,
    "timestep_mean": 479.125,
    "epsilon_norm": 7.1041388511657715,
    "prediction_norm": 5.853543281555176,
    "timestamp": "2025-07-14T12:34:09.308017"
  },
  {
    "epoch": 1005,
    "batch_idx": 160,
    "loss": 0.354226678609848,
    "learning_rate": 0.0004960730495556449,
    "timestep_mean": 536.4375,
    "epsilon_norm": 7.13282585144043,
    "prediction_norm": 5.8740973472595215,
    "timestamp": "2025-07-14T12:34:09.313351"
  },
  {
    "epoch": 1005,
    "batch_idx": 192,
    "loss": 0.06578329205513,
    "learning_rate": 0.0004960730495556449,
    "timestep_mean": 610.5,
    "epsilon_norm": 3.9836478233337402,
    "prediction_norm": 3.7252962589263916,
    "timestamp": "2025-07-14T12:34:09.318608"
  },
  {
    "epoch": 1006,
    "batch_idx": 0,
    "loss": 0.4892646372318268,
    "learning_rate": 0.0004952876807834285,
    "timestep_mean": 452.71875,
    "epsilon_norm": 9.197772026062012,
    "prediction_norm": 7.24863338470459,
    "timestamp": "2025-07-14T12:34:09.323898"
  },
  {
    "epoch": 1006,
    "batch_idx": 32,
    "loss": 0.8048006296157837,
    "learning_rate": 0.0004952876807834285,
    "timestep_mean": 427.03125,
    "epsilon_norm": 8.714608192443848,
    "prediction_norm": 6.276034355163574,
    "timestamp": "2025-07-14T12:34:09.329366"
  },
  {
    "epoch": 1006,
    "batch_idx": 64,
    "loss": 0.2760063707828522,
    "learning_rate": 0.0004952876807834285,
    "timestep_mean": 537.15625,
    "epsilon_norm": 6.642239093780518,
    "prediction_norm": 6.288654804229736,
    "timestamp": "2025-07-14T12:34:09.334590"
  },
  {
    "epoch": 1006,
    "batch_idx": 96,
    "loss": 0.30654141306877136,
    "learning_rate": 0.0004952876807834285,
    "timestep_mean": 525.03125,
    "epsilon_norm": 8.277240753173828,
    "prediction_norm": 7.468632698059082,
    "timestamp": "2025-07-14T12:34:09.339884"
  },
  {
    "epoch": 1006,
    "batch_idx": 128,
    "loss": 0.2208246886730194,
    "learning_rate": 0.0004952876807834285,
    "timestep_mean": 551.0625,
    "epsilon_norm": 9.136643409729004,
    "prediction_norm": 8.58998966217041,
    "timestamp": "2025-07-14T12:34:09.345023"
  },
  {
    "epoch": 1006,
    "batch_idx": 160,
    "loss": 0.30240851640701294,
    "learning_rate": 0.0004952876807834285,
    "timestep_mean": 522.46875,
    "epsilon_norm": 8.805990219116211,
    "prediction_norm": 8.028152465820312,
    "timestamp": "2025-07-14T12:34:09.350314"
  },
  {
    "epoch": 1006,
    "batch_idx": 192,
    "loss": 0.34939301013946533,
    "learning_rate": 0.0004952876807834285,
    "timestep_mean": 491.875,
    "epsilon_norm": 4.176042079925537,
    "prediction_norm": 3.5569567680358887,
    "timestamp": "2025-07-14T12:34:09.355632"
  },
  {
    "epoch": 1007,
    "batch_idx": 0,
    "loss": 0.3389570116996765,
    "learning_rate": 0.0004945023236383914,
    "timestep_mean": 474.78125,
    "epsilon_norm": 8.356610298156738,
    "prediction_norm": 6.513424873352051,
    "timestamp": "2025-07-14T12:34:09.360690"
  },
  {
    "epoch": 1007,
    "batch_idx": 32,
    "loss": 0.31379982829093933,
    "learning_rate": 0.0004945023236383914,
    "timestep_mean": 451.6875,
    "epsilon_norm": 8.254341125488281,
    "prediction_norm": 6.362967491149902,
    "timestamp": "2025-07-14T12:34:09.366008"
  },
  {
    "epoch": 1007,
    "batch_idx": 64,
    "loss": 0.264448881149292,
    "learning_rate": 0.0004945023236383914,
    "timestep_mean": 503.25,
    "epsilon_norm": 7.304507255554199,
    "prediction_norm": 6.095826625823975,
    "timestamp": "2025-07-14T12:34:09.371386"
  },
  {
    "epoch": 1007,
    "batch_idx": 96,
    "loss": 0.31583917140960693,
    "learning_rate": 0.0004945023236383914,
    "timestep_mean": 379.0,
    "epsilon_norm": 6.78782320022583,
    "prediction_norm": 5.408547878265381,
    "timestamp": "2025-07-14T12:34:09.376533"
  },
  {
    "epoch": 1007,
    "batch_idx": 128,
    "loss": 0.31825077533721924,
    "learning_rate": 0.0004945023236383914,
    "timestep_mean": 548.25,
    "epsilon_norm": 8.664446830749512,
    "prediction_norm": 7.486731052398682,
    "timestamp": "2025-07-14T12:34:09.381732"
  },
  {
    "epoch": 1007,
    "batch_idx": 160,
    "loss": 0.2594299912452698,
    "learning_rate": 0.0004945023236383914,
    "timestep_mean": 457.4375,
    "epsilon_norm": 7.89633321762085,
    "prediction_norm": 7.495591640472412,
    "timestamp": "2025-07-14T12:34:09.387045"
  },
  {
    "epoch": 1007,
    "batch_idx": 192,
    "loss": 0.04707915335893631,
    "learning_rate": 0.0004945023236383914,
    "timestep_mean": 600.25,
    "epsilon_norm": 4.0515360832214355,
    "prediction_norm": 4.462092399597168,
    "timestamp": "2025-07-14T12:34:09.392371"
  },
  {
    "epoch": 1008,
    "batch_idx": 0,
    "loss": 0.2255294919013977,
    "learning_rate": 0.0004937169800583242,
    "timestep_mean": 464.53125,
    "epsilon_norm": 7.478662490844727,
    "prediction_norm": 6.204349040985107,
    "timestamp": "2025-07-14T12:34:09.397597"
  },
  {
    "epoch": 1008,
    "batch_idx": 32,
    "loss": 0.31279733777046204,
    "learning_rate": 0.0004937169800583242,
    "timestep_mean": 486.78125,
    "epsilon_norm": 8.128851890563965,
    "prediction_norm": 7.161454677581787,
    "timestamp": "2025-07-14T12:34:09.402921"
  },
  {
    "epoch": 1008,
    "batch_idx": 64,
    "loss": 0.1825917363166809,
    "learning_rate": 0.0004937169800583242,
    "timestep_mean": 617.4375,
    "epsilon_norm": 7.032932281494141,
    "prediction_norm": 6.4418768882751465,
    "timestamp": "2025-07-14T12:34:09.407876"
  },
  {
    "epoch": 1008,
    "batch_idx": 96,
    "loss": 0.21344232559204102,
    "learning_rate": 0.0004937169800583242,
    "timestep_mean": 478.65625,
    "epsilon_norm": 8.226694107055664,
    "prediction_norm": 7.494097709655762,
    "timestamp": "2025-07-14T12:34:09.412891"
  },
  {
    "epoch": 1008,
    "batch_idx": 128,
    "loss": 0.3400377035140991,
    "learning_rate": 0.0004937169800583242,
    "timestep_mean": 433.5625,
    "epsilon_norm": 7.253603935241699,
    "prediction_norm": 5.919793128967285,
    "timestamp": "2025-07-14T12:34:09.418171"
  },
  {
    "epoch": 1008,
    "batch_idx": 160,
    "loss": 0.31874874234199524,
    "learning_rate": 0.0004937169800583242,
    "timestep_mean": 570.125,
    "epsilon_norm": 7.86571741104126,
    "prediction_norm": 6.371806621551514,
    "timestamp": "2025-07-14T12:34:09.423510"
  },
  {
    "epoch": 1008,
    "batch_idx": 192,
    "loss": 0.2421802431344986,
    "learning_rate": 0.0004937169800583242,
    "timestep_mean": 531.5,
    "epsilon_norm": 3.7863993644714355,
    "prediction_norm": 4.183802604675293,
    "timestamp": "2025-07-14T12:34:09.428798"
  },
  {
    "epoch": 1009,
    "batch_idx": 0,
    "loss": 0.2631628215312958,
    "learning_rate": 0.0004929316519809841,
    "timestep_mean": 525.40625,
    "epsilon_norm": 8.154593467712402,
    "prediction_norm": 6.8052077293396,
    "timestamp": "2025-07-14T12:34:09.434041"
  },
  {
    "epoch": 1009,
    "batch_idx": 32,
    "loss": 0.2804557681083679,
    "learning_rate": 0.0004929316519809841,
    "timestep_mean": 462.59375,
    "epsilon_norm": 7.754388809204102,
    "prediction_norm": 5.765021800994873,
    "timestamp": "2025-07-14T12:34:09.439323"
  },
  {
    "epoch": 1009,
    "batch_idx": 64,
    "loss": 0.27499255537986755,
    "learning_rate": 0.0004929316519809841,
    "timestep_mean": 420.84375,
    "epsilon_norm": 8.174078941345215,
    "prediction_norm": 6.482569217681885,
    "timestamp": "2025-07-14T12:34:09.444588"
  },
  {
    "epoch": 1009,
    "batch_idx": 96,
    "loss": 0.14545390009880066,
    "learning_rate": 0.0004929316519809841,
    "timestep_mean": 492.3125,
    "epsilon_norm": 7.671119689941406,
    "prediction_norm": 6.572676181793213,
    "timestamp": "2025-07-14T12:34:09.449801"
  },
  {
    "epoch": 1009,
    "batch_idx": 128,
    "loss": 0.45713281631469727,
    "learning_rate": 0.0004929316519809841,
    "timestep_mean": 405.71875,
    "epsilon_norm": 7.918223857879639,
    "prediction_norm": 5.819874286651611,
    "timestamp": "2025-07-14T12:34:09.455105"
  },
  {
    "epoch": 1009,
    "batch_idx": 160,
    "loss": 0.2008483111858368,
    "learning_rate": 0.0004929316519809841,
    "timestep_mean": 616.34375,
    "epsilon_norm": 7.3694682121276855,
    "prediction_norm": 6.230659008026123,
    "timestamp": "2025-07-14T12:34:09.460262"
  },
  {
    "epoch": 1009,
    "batch_idx": 192,
    "loss": 0.5846048593521118,
    "learning_rate": 0.0004929316519809841,
    "timestep_mean": 421.875,
    "epsilon_norm": 4.642309188842773,
    "prediction_norm": 3.0394647121429443,
    "timestamp": "2025-07-14T12:34:09.465213"
  },
  {
    "epoch": 1010,
    "batch_idx": 0,
    "loss": 0.2384239137172699,
    "learning_rate": 0.0004921463413440902,
    "timestep_mean": 541.4375,
    "epsilon_norm": 7.439499855041504,
    "prediction_norm": 5.799991607666016,
    "timestamp": "2025-07-14T12:34:09.470216"
  },
  {
    "epoch": 1010,
    "batch_idx": 32,
    "loss": 0.37536728382110596,
    "learning_rate": 0.0004921463413440902,
    "timestep_mean": 507.1875,
    "epsilon_norm": 9.382946968078613,
    "prediction_norm": 7.412134170532227,
    "timestamp": "2025-07-14T12:34:09.475679"
  },
  {
    "epoch": 1010,
    "batch_idx": 64,
    "loss": 0.49112915992736816,
    "learning_rate": 0.0004921463413440902,
    "timestep_mean": 489.375,
    "epsilon_norm": 8.896899223327637,
    "prediction_norm": 6.2393293380737305,
    "timestamp": "2025-07-14T12:34:09.480959"
  },
  {
    "epoch": 1010,
    "batch_idx": 96,
    "loss": 0.2798331379890442,
    "learning_rate": 0.0004921463413440902,
    "timestep_mean": 516.46875,
    "epsilon_norm": 7.7297797203063965,
    "prediction_norm": 6.32990837097168,
    "timestamp": "2025-07-14T12:34:09.486861"
  },
  {
    "epoch": 1010,
    "batch_idx": 128,
    "loss": 0.4376629889011383,
    "learning_rate": 0.0004921463413440902,
    "timestep_mean": 468.46875,
    "epsilon_norm": 8.279902458190918,
    "prediction_norm": 6.689513206481934,
    "timestamp": "2025-07-14T12:34:09.492395"
  },
  {
    "epoch": 1010,
    "batch_idx": 160,
    "loss": 0.40668541193008423,
    "learning_rate": 0.0004921463413440902,
    "timestep_mean": 442.96875,
    "epsilon_norm": 7.24735689163208,
    "prediction_norm": 5.196885108947754,
    "timestamp": "2025-07-14T12:34:09.497709"
  },
  {
    "epoch": 1010,
    "batch_idx": 192,
    "loss": 0.3398919701576233,
    "learning_rate": 0.0004921463413440902,
    "timestep_mean": 654.0,
    "epsilon_norm": 4.231522083282471,
    "prediction_norm": 3.5572152137756348,
    "timestamp": "2025-07-14T12:34:09.503097"
  },
  {
    "epoch": 1011,
    "batch_idx": 0,
    "loss": 0.3874787390232086,
    "learning_rate": 0.0004913610500853182,
    "timestep_mean": 483.0625,
    "epsilon_norm": 8.388358116149902,
    "prediction_norm": 7.531017780303955,
    "timestamp": "2025-07-14T12:34:09.508410"
  },
  {
    "epoch": 1011,
    "batch_idx": 32,
    "loss": 0.29233866930007935,
    "learning_rate": 0.0004913610500853182,
    "timestep_mean": 428.0,
    "epsilon_norm": 7.905367851257324,
    "prediction_norm": 6.778067111968994,
    "timestamp": "2025-07-14T12:34:09.513683"
  },
  {
    "epoch": 1011,
    "batch_idx": 64,
    "loss": 0.41636961698532104,
    "learning_rate": 0.0004913610500853182,
    "timestep_mean": 416.9375,
    "epsilon_norm": 9.07175350189209,
    "prediction_norm": 6.398475170135498,
    "timestamp": "2025-07-14T12:34:09.518981"
  },
  {
    "epoch": 1011,
    "batch_idx": 96,
    "loss": 0.27223411202430725,
    "learning_rate": 0.0004913610500853182,
    "timestep_mean": 489.90625,
    "epsilon_norm": 9.106111526489258,
    "prediction_norm": 8.597831726074219,
    "timestamp": "2025-07-14T12:34:09.524207"
  },
  {
    "epoch": 1011,
    "batch_idx": 128,
    "loss": 0.2230633646249771,
    "learning_rate": 0.0004913610500853182,
    "timestep_mean": 512.0625,
    "epsilon_norm": 7.4391255378723145,
    "prediction_norm": 6.2990546226501465,
    "timestamp": "2025-07-14T12:34:09.529538"
  },
  {
    "epoch": 1011,
    "batch_idx": 160,
    "loss": 0.2134547233581543,
    "learning_rate": 0.0004913610500853182,
    "timestep_mean": 561.625,
    "epsilon_norm": 8.098689079284668,
    "prediction_norm": 8.192745208740234,
    "timestamp": "2025-07-14T12:34:09.534720"
  },
  {
    "epoch": 1011,
    "batch_idx": 192,
    "loss": 0.20584538578987122,
    "learning_rate": 0.0004913610500853182,
    "timestep_mean": 413.625,
    "epsilon_norm": 3.0421600341796875,
    "prediction_norm": 2.855534076690674,
    "timestamp": "2025-07-14T12:34:09.539936"
  },
  {
    "epoch": 1012,
    "batch_idx": 0,
    "loss": 0.36716902256011963,
    "learning_rate": 0.0004905757801422962,
    "timestep_mean": 406.6875,
    "epsilon_norm": 8.420456886291504,
    "prediction_norm": 6.755157947540283,
    "timestamp": "2025-07-14T12:34:09.545322"
  },
  {
    "epoch": 1012,
    "batch_idx": 32,
    "loss": 0.5077204704284668,
    "learning_rate": 0.0004905757801422962,
    "timestep_mean": 566.90625,
    "epsilon_norm": 10.780828475952148,
    "prediction_norm": 9.213831901550293,
    "timestamp": "2025-07-14T12:34:09.550735"
  },
  {
    "epoch": 1012,
    "batch_idx": 64,
    "loss": 0.41145873069763184,
    "learning_rate": 0.0004905757801422962,
    "timestep_mean": 392.90625,
    "epsilon_norm": 7.868095874786377,
    "prediction_norm": 6.835571765899658,
    "timestamp": "2025-07-14T12:34:09.556044"
  },
  {
    "epoch": 1012,
    "batch_idx": 96,
    "loss": 0.3278908133506775,
    "learning_rate": 0.0004905757801422962,
    "timestep_mean": 460.8125,
    "epsilon_norm": 7.544615268707275,
    "prediction_norm": 6.717188358306885,
    "timestamp": "2025-07-14T12:34:09.561336"
  },
  {
    "epoch": 1012,
    "batch_idx": 128,
    "loss": 0.36050960421562195,
    "learning_rate": 0.0004905757801422962,
    "timestep_mean": 502.875,
    "epsilon_norm": 7.565622329711914,
    "prediction_norm": 7.305429458618164,
    "timestamp": "2025-07-14T12:34:09.566553"
  },
  {
    "epoch": 1012,
    "batch_idx": 160,
    "loss": 0.33580678701400757,
    "learning_rate": 0.0004905757801422962,
    "timestep_mean": 522.375,
    "epsilon_norm": 8.470162391662598,
    "prediction_norm": 7.823571681976318,
    "timestamp": "2025-07-14T12:34:09.571789"
  },
  {
    "epoch": 1012,
    "batch_idx": 192,
    "loss": 0.3259579539299011,
    "learning_rate": 0.0004905757801422962,
    "timestep_mean": 427.5,
    "epsilon_norm": 3.660156726837158,
    "prediction_norm": 3.5474472045898438,
    "timestamp": "2025-07-14T12:34:09.576621"
  },
  {
    "epoch": 1013,
    "batch_idx": 0,
    "loss": 0.3462015986442566,
    "learning_rate": 0.0004897905334526001,
    "timestep_mean": 509.46875,
    "epsilon_norm": 7.606841087341309,
    "prediction_norm": 6.297231674194336,
    "timestamp": "2025-07-14T12:34:09.581515"
  },
  {
    "epoch": 1013,
    "batch_idx": 32,
    "loss": 0.27327871322631836,
    "learning_rate": 0.0004897905334526001,
    "timestep_mean": 457.71875,
    "epsilon_norm": 8.238814353942871,
    "prediction_norm": 7.935944557189941,
    "timestamp": "2025-07-14T12:34:09.586493"
  },
  {
    "epoch": 1013,
    "batch_idx": 64,
    "loss": 0.37406057119369507,
    "learning_rate": 0.0004897905334526001,
    "timestep_mean": 420.03125,
    "epsilon_norm": 6.812168598175049,
    "prediction_norm": 5.226396560668945,
    "timestamp": "2025-07-14T12:34:09.591554"
  },
  {
    "epoch": 1013,
    "batch_idx": 96,
    "loss": 0.30184075236320496,
    "learning_rate": 0.0004897905334526001,
    "timestep_mean": 532.03125,
    "epsilon_norm": 8.011805534362793,
    "prediction_norm": 6.879220485687256,
    "timestamp": "2025-07-14T12:34:09.596707"
  },
  {
    "epoch": 1013,
    "batch_idx": 128,
    "loss": 0.3541736304759979,
    "learning_rate": 0.0004897905334526001,
    "timestep_mean": 419.0625,
    "epsilon_norm": 8.108189582824707,
    "prediction_norm": 6.540946960449219,
    "timestamp": "2025-07-14T12:34:09.601859"
  },
  {
    "epoch": 1013,
    "batch_idx": 160,
    "loss": 0.21542951464653015,
    "learning_rate": 0.0004897905334526001,
    "timestep_mean": 476.8125,
    "epsilon_norm": 7.053020000457764,
    "prediction_norm": 6.3205718994140625,
    "timestamp": "2025-07-14T12:34:09.607124"
  },
  {
    "epoch": 1013,
    "batch_idx": 192,
    "loss": 0.344893217086792,
    "learning_rate": 0.0004897905334526001,
    "timestep_mean": 589.0,
    "epsilon_norm": 3.8272135257720947,
    "prediction_norm": 2.9318296909332275,
    "timestamp": "2025-07-14T12:34:09.612246"
  },
  {
    "epoch": 1014,
    "batch_idx": 0,
    "loss": 0.3162594437599182,
    "learning_rate": 0.0004890053119537478,
    "timestep_mean": 522.9375,
    "epsilon_norm": 7.177901744842529,
    "prediction_norm": 5.539275169372559,
    "timestamp": "2025-07-14T12:34:09.617537"
  },
  {
    "epoch": 1014,
    "batch_idx": 32,
    "loss": 0.30184459686279297,
    "learning_rate": 0.0004890053119537478,
    "timestep_mean": 443.6875,
    "epsilon_norm": 7.637097358703613,
    "prediction_norm": 5.496585845947266,
    "timestamp": "2025-07-14T12:34:09.622815"
  },
  {
    "epoch": 1014,
    "batch_idx": 64,
    "loss": 0.08728368580341339,
    "learning_rate": 0.0004890053119537478,
    "timestep_mean": 660.875,
    "epsilon_norm": 7.108969688415527,
    "prediction_norm": 6.582038402557373,
    "timestamp": "2025-07-14T12:34:09.628096"
  },
  {
    "epoch": 1014,
    "batch_idx": 96,
    "loss": 0.3976729214191437,
    "learning_rate": 0.0004890053119537478,
    "timestep_mean": 493.125,
    "epsilon_norm": 7.627128601074219,
    "prediction_norm": 6.002805709838867,
    "timestamp": "2025-07-14T12:34:09.633354"
  },
  {
    "epoch": 1014,
    "batch_idx": 128,
    "loss": 0.22939258813858032,
    "learning_rate": 0.0004890053119537478,
    "timestep_mean": 467.71875,
    "epsilon_norm": 8.093658447265625,
    "prediction_norm": 6.486783504486084,
    "timestamp": "2025-07-14T12:34:09.638672"
  },
  {
    "epoch": 1014,
    "batch_idx": 160,
    "loss": 0.585282027721405,
    "learning_rate": 0.0004890053119537478,
    "timestep_mean": 417.40625,
    "epsilon_norm": 8.903794288635254,
    "prediction_norm": 6.354089736938477,
    "timestamp": "2025-07-14T12:34:09.643775"
  },
  {
    "epoch": 1014,
    "batch_idx": 192,
    "loss": 0.1351436823606491,
    "learning_rate": 0.0004890053119537478,
    "timestep_mean": 547.25,
    "epsilon_norm": 3.7658863067626953,
    "prediction_norm": 3.279103994369507,
    "timestamp": "2025-07-14T12:34:09.648666"
  },
  {
    "epoch": 1015,
    "batch_idx": 0,
    "loss": 0.20716556906700134,
    "learning_rate": 0.0004882201175831952,
    "timestep_mean": 477.0,
    "epsilon_norm": 7.96086311340332,
    "prediction_norm": 6.65634822845459,
    "timestamp": "2025-07-14T12:34:09.653563"
  },
  {
    "epoch": 1015,
    "batch_idx": 32,
    "loss": 0.31640708446502686,
    "learning_rate": 0.0004882201175831952,
    "timestep_mean": 463.5,
    "epsilon_norm": 9.015371322631836,
    "prediction_norm": 7.534616947174072,
    "timestamp": "2025-07-14T12:34:09.658524"
  },
  {
    "epoch": 1015,
    "batch_idx": 64,
    "loss": 0.29765528440475464,
    "learning_rate": 0.0004882201175831952,
    "timestep_mean": 408.125,
    "epsilon_norm": 7.8517913818359375,
    "prediction_norm": 5.872200965881348,
    "timestamp": "2025-07-14T12:34:09.663381"
  },
  {
    "epoch": 1015,
    "batch_idx": 96,
    "loss": 0.18711969256401062,
    "learning_rate": 0.0004882201175831952,
    "timestep_mean": 538.4375,
    "epsilon_norm": 8.737387657165527,
    "prediction_norm": 7.995490074157715,
    "timestamp": "2025-07-14T12:34:09.668328"
  },
  {
    "epoch": 1015,
    "batch_idx": 128,
    "loss": 0.2981921136379242,
    "learning_rate": 0.0004882201175831952,
    "timestep_mean": 454.09375,
    "epsilon_norm": 7.667262077331543,
    "prediction_norm": 6.307905673980713,
    "timestamp": "2025-07-14T12:34:09.673498"
  },
  {
    "epoch": 1015,
    "batch_idx": 160,
    "loss": 0.2061741203069687,
    "learning_rate": 0.0004882201175831952,
    "timestep_mean": 546.0625,
    "epsilon_norm": 7.365678787231445,
    "prediction_norm": 6.178271770477295,
    "timestamp": "2025-07-14T12:34:09.678700"
  },
  {
    "epoch": 1015,
    "batch_idx": 192,
    "loss": 0.2510972023010254,
    "learning_rate": 0.0004882201175831952,
    "timestep_mean": 499.625,
    "epsilon_norm": 4.335760116577148,
    "prediction_norm": 3.665821075439453,
    "timestamp": "2025-07-14T12:34:09.683960"
  },
  {
    "epoch": 1016,
    "batch_idx": 0,
    "loss": 0.29395201802253723,
    "learning_rate": 0.00048743495227833157,
    "timestep_mean": 550.5,
    "epsilon_norm": 8.248143196105957,
    "prediction_norm": 6.785037517547607,
    "timestamp": "2025-07-14T12:34:09.689197"
  },
  {
    "epoch": 1016,
    "batch_idx": 32,
    "loss": 0.3688243627548218,
    "learning_rate": 0.00048743495227833157,
    "timestep_mean": 545.25,
    "epsilon_norm": 9.928228378295898,
    "prediction_norm": 8.075021743774414,
    "timestamp": "2025-07-14T12:34:09.694378"
  },
  {
    "epoch": 1016,
    "batch_idx": 64,
    "loss": 0.2190563678741455,
    "learning_rate": 0.00048743495227833157,
    "timestep_mean": 453.4375,
    "epsilon_norm": 7.5507707595825195,
    "prediction_norm": 6.1833038330078125,
    "timestamp": "2025-07-14T12:34:09.699402"
  },
  {
    "epoch": 1016,
    "batch_idx": 96,
    "loss": 0.2646809220314026,
    "learning_rate": 0.00048743495227833157,
    "timestep_mean": 452.15625,
    "epsilon_norm": 8.145333290100098,
    "prediction_norm": 6.87711238861084,
    "timestamp": "2025-07-14T12:34:09.704641"
  },
  {
    "epoch": 1016,
    "batch_idx": 128,
    "loss": 0.41888976097106934,
    "learning_rate": 0.00048743495227833157,
    "timestep_mean": 531.15625,
    "epsilon_norm": 8.2532320022583,
    "prediction_norm": 6.269198417663574,
    "timestamp": "2025-07-14T12:34:09.709669"
  },
  {
    "epoch": 1016,
    "batch_idx": 160,
    "loss": 0.19122660160064697,
    "learning_rate": 0.00048743495227833157,
    "timestep_mean": 515.78125,
    "epsilon_norm": 7.2537713050842285,
    "prediction_norm": 6.482111930847168,
    "timestamp": "2025-07-14T12:34:09.714574"
  },
  {
    "epoch": 1016,
    "batch_idx": 192,
    "loss": 0.17674311995506287,
    "learning_rate": 0.00048743495227833157,
    "timestep_mean": 517.25,
    "epsilon_norm": 3.2860543727874756,
    "prediction_norm": 3.0556070804595947,
    "timestamp": "2025-07-14T12:34:09.719443"
  },
  {
    "epoch": 1017,
    "batch_idx": 0,
    "loss": 0.37096115946769714,
    "learning_rate": 0.0004866498179764742,
    "timestep_mean": 490.625,
    "epsilon_norm": 8.579629898071289,
    "prediction_norm": 6.7409539222717285,
    "timestamp": "2025-07-14T12:34:09.724493"
  },
  {
    "epoch": 1017,
    "batch_idx": 32,
    "loss": 0.23688343167304993,
    "learning_rate": 0.0004866498179764742,
    "timestep_mean": 490.03125,
    "epsilon_norm": 8.359092712402344,
    "prediction_norm": 7.368642330169678,
    "timestamp": "2025-07-14T12:34:09.729992"
  },
  {
    "epoch": 1017,
    "batch_idx": 64,
    "loss": 0.1922333538532257,
    "learning_rate": 0.0004866498179764742,
    "timestep_mean": 491.53125,
    "epsilon_norm": 7.499746322631836,
    "prediction_norm": 7.069841384887695,
    "timestamp": "2025-07-14T12:34:09.736095"
  },
  {
    "epoch": 1017,
    "batch_idx": 96,
    "loss": 0.1800759732723236,
    "learning_rate": 0.0004866498179764742,
    "timestep_mean": 624.375,
    "epsilon_norm": 7.868940353393555,
    "prediction_norm": 7.370391368865967,
    "timestamp": "2025-07-14T12:34:09.741321"
  },
  {
    "epoch": 1017,
    "batch_idx": 128,
    "loss": 0.26178330183029175,
    "learning_rate": 0.0004866498179764742,
    "timestep_mean": 502.5625,
    "epsilon_norm": 8.586448669433594,
    "prediction_norm": 7.781676292419434,
    "timestamp": "2025-07-14T12:34:09.746649"
  },
  {
    "epoch": 1017,
    "batch_idx": 160,
    "loss": 0.19438371062278748,
    "learning_rate": 0.0004866498179764742,
    "timestep_mean": 525.09375,
    "epsilon_norm": 8.183398246765137,
    "prediction_norm": 7.27890157699585,
    "timestamp": "2025-07-14T12:34:09.752007"
  },
  {
    "epoch": 1017,
    "batch_idx": 192,
    "loss": 0.20166748762130737,
    "learning_rate": 0.0004866498179764742,
    "timestep_mean": 655.0,
    "epsilon_norm": 4.386126518249512,
    "prediction_norm": 3.7856969833374023,
    "timestamp": "2025-07-14T12:34:09.757471"
  },
  {
    "epoch": 1018,
    "batch_idx": 0,
    "loss": 0.21307501196861267,
    "learning_rate": 0.00048586471661486366,
    "timestep_mean": 577.8125,
    "epsilon_norm": 7.088959217071533,
    "prediction_norm": 6.448684215545654,
    "timestamp": "2025-07-14T12:34:09.762761"
  },
  {
    "epoch": 1018,
    "batch_idx": 32,
    "loss": 0.20872074365615845,
    "learning_rate": 0.00048586471661486366,
    "timestep_mean": 535.125,
    "epsilon_norm": 7.158534526824951,
    "prediction_norm": 6.080458641052246,
    "timestamp": "2025-07-14T12:34:09.768059"
  },
  {
    "epoch": 1018,
    "batch_idx": 64,
    "loss": 0.24055495858192444,
    "learning_rate": 0.00048586471661486366,
    "timestep_mean": 470.8125,
    "epsilon_norm": 7.7318854331970215,
    "prediction_norm": 7.33359432220459,
    "timestamp": "2025-07-14T12:34:09.773362"
  },
  {
    "epoch": 1018,
    "batch_idx": 96,
    "loss": 0.3198901414871216,
    "learning_rate": 0.00048586471661486366,
    "timestep_mean": 546.125,
    "epsilon_norm": 8.612330436706543,
    "prediction_norm": 7.085917949676514,
    "timestamp": "2025-07-14T12:34:09.778607"
  },
  {
    "epoch": 1018,
    "batch_idx": 128,
    "loss": 0.18346503376960754,
    "learning_rate": 0.00048586471661486366,
    "timestep_mean": 536.75,
    "epsilon_norm": 7.664749622344971,
    "prediction_norm": 6.505699157714844,
    "timestamp": "2025-07-14T12:34:09.784007"
  },
  {
    "epoch": 1018,
    "batch_idx": 160,
    "loss": 0.22137562930583954,
    "learning_rate": 0.00048586471661486366,
    "timestep_mean": 538.71875,
    "epsilon_norm": 7.724813938140869,
    "prediction_norm": 6.711068630218506,
    "timestamp": "2025-07-14T12:34:09.789312"
  },
  {
    "epoch": 1018,
    "batch_idx": 192,
    "loss": 0.44602063298225403,
    "learning_rate": 0.00048586471661486366,
    "timestep_mean": 599.125,
    "epsilon_norm": 3.764615774154663,
    "prediction_norm": 3.184936046600342,
    "timestamp": "2025-07-14T12:34:09.794460"
  },
  {
    "epoch": 1019,
    "batch_idx": 0,
    "loss": 0.4033083915710449,
    "learning_rate": 0.0004850796501306598,
    "timestep_mean": 449.96875,
    "epsilon_norm": 7.016172409057617,
    "prediction_norm": 5.759740352630615,
    "timestamp": "2025-07-14T12:34:09.799685"
  },
  {
    "epoch": 1019,
    "batch_idx": 32,
    "loss": 0.15862885117530823,
    "learning_rate": 0.0004850796501306598,
    "timestep_mean": 627.375,
    "epsilon_norm": 7.9212236404418945,
    "prediction_norm": 7.198439598083496,
    "timestamp": "2025-07-14T12:34:09.804744"
  },
  {
    "epoch": 1019,
    "batch_idx": 64,
    "loss": 0.16444233059883118,
    "learning_rate": 0.0004850796501306598,
    "timestep_mean": 527.34375,
    "epsilon_norm": 7.011448860168457,
    "prediction_norm": 5.671220302581787,
    "timestamp": "2025-07-14T12:34:09.809771"
  },
  {
    "epoch": 1019,
    "batch_idx": 96,
    "loss": 0.4712148904800415,
    "learning_rate": 0.0004850796501306598,
    "timestep_mean": 440.125,
    "epsilon_norm": 7.858023166656494,
    "prediction_norm": 5.942634582519531,
    "timestamp": "2025-07-14T12:34:09.815374"
  },
  {
    "epoch": 1019,
    "batch_idx": 128,
    "loss": 0.34226900339126587,
    "learning_rate": 0.0004850796501306598,
    "timestep_mean": 438.75,
    "epsilon_norm": 8.427258491516113,
    "prediction_norm": 7.181158065795898,
    "timestamp": "2025-07-14T12:34:09.821550"
  },
  {
    "epoch": 1019,
    "batch_idx": 160,
    "loss": 0.33609098196029663,
    "learning_rate": 0.0004850796501306598,
    "timestep_mean": 405.75,
    "epsilon_norm": 8.058362007141113,
    "prediction_norm": 6.244008541107178,
    "timestamp": "2025-07-14T12:34:09.827007"
  },
  {
    "epoch": 1019,
    "batch_idx": 192,
    "loss": 0.258594810962677,
    "learning_rate": 0.0004850796501306598,
    "timestep_mean": 498.125,
    "epsilon_norm": 3.877742052078247,
    "prediction_norm": 3.1289708614349365,
    "timestamp": "2025-07-14T12:34:09.832527"
  },
  {
    "epoch": 1020,
    "batch_idx": 0,
    "loss": 0.5055770874023438,
    "learning_rate": 0.0004842946204609361,
    "timestep_mean": 476.09375,
    "epsilon_norm": 8.094853401184082,
    "prediction_norm": 5.673689842224121,
    "timestamp": "2025-07-14T12:34:09.838014"
  },
  {
    "epoch": 1020,
    "batch_idx": 32,
    "loss": 0.19451767206192017,
    "learning_rate": 0.0004842946204609361,
    "timestep_mean": 534.03125,
    "epsilon_norm": 6.926743984222412,
    "prediction_norm": 6.008344650268555,
    "timestamp": "2025-07-14T12:34:09.843202"
  },
  {
    "epoch": 1020,
    "batch_idx": 64,
    "loss": 0.2546335458755493,
    "learning_rate": 0.0004842946204609361,
    "timestep_mean": 489.84375,
    "epsilon_norm": 7.811529636383057,
    "prediction_norm": 6.477985858917236,
    "timestamp": "2025-07-14T12:34:09.848391"
  },
  {
    "epoch": 1020,
    "batch_idx": 96,
    "loss": 0.31145596504211426,
    "learning_rate": 0.0004842946204609361,
    "timestep_mean": 583.09375,
    "epsilon_norm": 8.527752876281738,
    "prediction_norm": 7.222492694854736,
    "timestamp": "2025-07-14T12:34:09.853560"
  },
  {
    "epoch": 1020,
    "batch_idx": 128,
    "loss": 0.28868529200553894,
    "learning_rate": 0.0004842946204609361,
    "timestep_mean": 416.5,
    "epsilon_norm": 7.433229446411133,
    "prediction_norm": 5.8642988204956055,
    "timestamp": "2025-07-14T12:34:09.859080"
  },
  {
    "epoch": 1020,
    "batch_idx": 160,
    "loss": 0.1715705394744873,
    "learning_rate": 0.0004842946204609361,
    "timestep_mean": 461.40625,
    "epsilon_norm": 8.133806228637695,
    "prediction_norm": 6.559539794921875,
    "timestamp": "2025-07-14T12:34:09.864366"
  },
  {
    "epoch": 1020,
    "batch_idx": 192,
    "loss": 0.14561142027378082,
    "learning_rate": 0.0004842946204609361,
    "timestep_mean": 571.125,
    "epsilon_norm": 3.038572072982788,
    "prediction_norm": 2.657273292541504,
    "timestamp": "2025-07-14T12:34:09.869601"
  },
  {
    "epoch": 1021,
    "batch_idx": 0,
    "loss": 0.2731940746307373,
    "learning_rate": 0.0004835096295426753,
    "timestep_mean": 580.84375,
    "epsilon_norm": 7.3731536865234375,
    "prediction_norm": 6.014637470245361,
    "timestamp": "2025-07-14T12:34:09.875032"
  },
  {
    "epoch": 1021,
    "batch_idx": 32,
    "loss": 0.2450416386127472,
    "learning_rate": 0.0004835096295426753,
    "timestep_mean": 491.25,
    "epsilon_norm": 7.15546178817749,
    "prediction_norm": 6.188674449920654,
    "timestamp": "2025-07-14T12:34:09.880278"
  },
  {
    "epoch": 1021,
    "batch_idx": 64,
    "loss": 0.3826073706150055,
    "learning_rate": 0.0004835096295426753,
    "timestep_mean": 467.0,
    "epsilon_norm": 8.48083209991455,
    "prediction_norm": 6.596771717071533,
    "timestamp": "2025-07-14T12:34:09.885646"
  },
  {
    "epoch": 1021,
    "batch_idx": 96,
    "loss": 0.13207848370075226,
    "learning_rate": 0.0004835096295426753,
    "timestep_mean": 552.71875,
    "epsilon_norm": 7.675408840179443,
    "prediction_norm": 7.157966613769531,
    "timestamp": "2025-07-14T12:34:09.890880"
  },
  {
    "epoch": 1021,
    "batch_idx": 128,
    "loss": 0.1534060537815094,
    "learning_rate": 0.0004835096295426753,
    "timestep_mean": 496.6875,
    "epsilon_norm": 7.776266574859619,
    "prediction_norm": 7.359207630157471,
    "timestamp": "2025-07-14T12:34:09.896319"
  },
  {
    "epoch": 1021,
    "batch_idx": 160,
    "loss": 0.37162572145462036,
    "learning_rate": 0.0004835096295426753,
    "timestep_mean": 554.0625,
    "epsilon_norm": 9.886115074157715,
    "prediction_norm": 8.284153938293457,
    "timestamp": "2025-07-14T12:34:09.901598"
  },
  {
    "epoch": 1021,
    "batch_idx": 192,
    "loss": 0.0789327621459961,
    "learning_rate": 0.0004835096295426753,
    "timestep_mean": 683.375,
    "epsilon_norm": 3.0452678203582764,
    "prediction_norm": 3.013411045074463,
    "timestamp": "2025-07-14T12:34:09.906740"
  },
  {
    "epoch": 1022,
    "batch_idx": 0,
    "loss": 0.32898208498954773,
    "learning_rate": 0.00048272467931276427,
    "timestep_mean": 386.0625,
    "epsilon_norm": 7.5060248374938965,
    "prediction_norm": 6.385915279388428,
    "timestamp": "2025-07-14T12:34:09.911689"
  },
  {
    "epoch": 1022,
    "batch_idx": 32,
    "loss": 0.15870419144630432,
    "learning_rate": 0.00048272467931276427,
    "timestep_mean": 532.21875,
    "epsilon_norm": 7.466512203216553,
    "prediction_norm": 7.059253692626953,
    "timestamp": "2025-07-14T12:34:09.916586"
  },
  {
    "epoch": 1022,
    "batch_idx": 64,
    "loss": 0.4306907057762146,
    "learning_rate": 0.00048272467931276427,
    "timestep_mean": 510.46875,
    "epsilon_norm": 8.999250411987305,
    "prediction_norm": 7.275003910064697,
    "timestamp": "2025-07-14T12:34:09.921617"
  },
  {
    "epoch": 1022,
    "batch_idx": 96,
    "loss": 0.25975301861763,
    "learning_rate": 0.00048272467931276427,
    "timestep_mean": 519.46875,
    "epsilon_norm": 7.80121374130249,
    "prediction_norm": 7.475409984588623,
    "timestamp": "2025-07-14T12:34:09.926535"
  },
  {
    "epoch": 1022,
    "batch_idx": 128,
    "loss": 0.23869886994361877,
    "learning_rate": 0.00048272467931276427,
    "timestep_mean": 466.6875,
    "epsilon_norm": 7.778722763061523,
    "prediction_norm": 7.817351341247559,
    "timestamp": "2025-07-14T12:34:09.931548"
  },
  {
    "epoch": 1022,
    "batch_idx": 160,
    "loss": 0.2744983434677124,
    "learning_rate": 0.00048272467931276427,
    "timestep_mean": 537.75,
    "epsilon_norm": 7.223662376403809,
    "prediction_norm": 6.463008403778076,
    "timestamp": "2025-07-14T12:34:09.936508"
  },
  {
    "epoch": 1022,
    "batch_idx": 192,
    "loss": 0.07042022049427032,
    "learning_rate": 0.00048272467931276427,
    "timestep_mean": 461.125,
    "epsilon_norm": 3.0364646911621094,
    "prediction_norm": 3.1089344024658203,
    "timestamp": "2025-07-14T12:34:09.941548"
  },
  {
    "epoch": 1023,
    "batch_idx": 0,
    "loss": 0.39336690306663513,
    "learning_rate": 0.00048193977170798963,
    "timestep_mean": 453.71875,
    "epsilon_norm": 6.900760650634766,
    "prediction_norm": 5.898860931396484,
    "timestamp": "2025-07-14T12:34:09.946614"
  },
  {
    "epoch": 1023,
    "batch_idx": 32,
    "loss": 0.2343946099281311,
    "learning_rate": 0.00048193977170798963,
    "timestep_mean": 556.59375,
    "epsilon_norm": 7.445122718811035,
    "prediction_norm": 6.459164142608643,
    "timestamp": "2025-07-14T12:34:09.951510"
  },
  {
    "epoch": 1023,
    "batch_idx": 64,
    "loss": 0.34479522705078125,
    "learning_rate": 0.00048193977170798963,
    "timestep_mean": 478.71875,
    "epsilon_norm": 8.243632316589355,
    "prediction_norm": 6.499980926513672,
    "timestamp": "2025-07-14T12:34:09.956454"
  },
  {
    "epoch": 1023,
    "batch_idx": 96,
    "loss": 0.1063709557056427,
    "learning_rate": 0.00048193977170798963,
    "timestep_mean": 617.78125,
    "epsilon_norm": 8.220714569091797,
    "prediction_norm": 8.015941619873047,
    "timestamp": "2025-07-14T12:34:09.961585"
  },
  {
    "epoch": 1023,
    "batch_idx": 128,
    "loss": 0.5050910711288452,
    "learning_rate": 0.00048193977170798963,
    "timestep_mean": 450.28125,
    "epsilon_norm": 9.6226167678833,
    "prediction_norm": 7.688926696777344,
    "timestamp": "2025-07-14T12:34:09.968558"
  },
  {
    "epoch": 1023,
    "batch_idx": 160,
    "loss": 0.1149158924818039,
    "learning_rate": 0.00048193977170798963,
    "timestep_mean": 532.75,
    "epsilon_norm": 7.841259479522705,
    "prediction_norm": 7.27822732925415,
    "timestamp": "2025-07-14T12:34:09.973823"
  },
  {
    "epoch": 1023,
    "batch_idx": 192,
    "loss": 0.16418537497520447,
    "learning_rate": 0.00048193977170798963,
    "timestep_mean": 731.375,
    "epsilon_norm": 4.01832389831543,
    "prediction_norm": 3.1418793201446533,
    "timestamp": "2025-07-14T12:34:09.979130"
  },
  {
    "epoch": 1024,
    "batch_idx": 0,
    "loss": 0.41979724168777466,
    "learning_rate": 0.00048115490866503294,
    "timestep_mean": 568.6875,
    "epsilon_norm": 8.015355110168457,
    "prediction_norm": 6.422318935394287,
    "timestamp": "2025-07-14T12:34:09.984201"
  },
  {
    "epoch": 1024,
    "batch_idx": 32,
    "loss": 0.21268445253372192,
    "learning_rate": 0.00048115490866503294,
    "timestep_mean": 510.0625,
    "epsilon_norm": 8.745681762695312,
    "prediction_norm": 7.492002487182617,
    "timestamp": "2025-07-14T12:34:09.989149"
  },
  {
    "epoch": 1024,
    "batch_idx": 64,
    "loss": 0.2156877964735031,
    "learning_rate": 0.00048115490866503294,
    "timestep_mean": 523.8125,
    "epsilon_norm": 7.888046741485596,
    "prediction_norm": 6.490708827972412,
    "timestamp": "2025-07-14T12:34:09.993984"
  },
  {
    "epoch": 1024,
    "batch_idx": 96,
    "loss": 0.2090207040309906,
    "learning_rate": 0.00048115490866503294,
    "timestep_mean": 452.46875,
    "epsilon_norm": 7.594915390014648,
    "prediction_norm": 6.875402450561523,
    "timestamp": "2025-07-14T12:34:09.998901"
  },
  {
    "epoch": 1024,
    "batch_idx": 128,
    "loss": 0.273086279630661,
    "learning_rate": 0.00048115490866503294,
    "timestep_mean": 546.375,
    "epsilon_norm": 8.580002784729004,
    "prediction_norm": 7.410991668701172,
    "timestamp": "2025-07-14T12:34:10.003829"
  },
  {
    "epoch": 1024,
    "batch_idx": 160,
    "loss": 0.26113951206207275,
    "learning_rate": 0.00048115490866503294,
    "timestep_mean": 598.1875,
    "epsilon_norm": 7.102725505828857,
    "prediction_norm": 5.804717540740967,
    "timestamp": "2025-07-14T12:34:10.009498"
  },
  {
    "epoch": 1024,
    "batch_idx": 192,
    "loss": 0.36655810475349426,
    "learning_rate": 0.00048115490866503294,
    "timestep_mean": 428.5,
    "epsilon_norm": 3.3998632431030273,
    "prediction_norm": 2.1541554927825928,
    "timestamp": "2025-07-14T12:34:10.014781"
  },
  {
    "epoch": 1025,
    "batch_idx": 0,
    "loss": 0.3353568911552429,
    "learning_rate": 0.00048037009212046597,
    "timestep_mean": 406.75,
    "epsilon_norm": 8.371870040893555,
    "prediction_norm": 6.039027690887451,
    "timestamp": "2025-07-14T12:34:10.027020"
  },
  {
    "epoch": 1025,
    "batch_idx": 32,
    "loss": 0.25716087222099304,
    "learning_rate": 0.00048037009212046597,
    "timestep_mean": 534.96875,
    "epsilon_norm": 8.611078262329102,
    "prediction_norm": 6.793000221252441,
    "timestamp": "2025-07-14T12:34:10.032677"
  },
  {
    "epoch": 1025,
    "batch_idx": 64,
    "loss": 0.31389519572257996,
    "learning_rate": 0.00048037009212046597,
    "timestep_mean": 463.125,
    "epsilon_norm": 7.464300155639648,
    "prediction_norm": 5.943293571472168,
    "timestamp": "2025-07-14T12:34:10.037834"
  },
  {
    "epoch": 1025,
    "batch_idx": 96,
    "loss": 0.12055277824401855,
    "learning_rate": 0.00048037009212046597,
    "timestep_mean": 572.96875,
    "epsilon_norm": 6.53890323638916,
    "prediction_norm": 6.334884166717529,
    "timestamp": "2025-07-14T12:34:10.042726"
  },
  {
    "epoch": 1025,
    "batch_idx": 128,
    "loss": 0.2069709300994873,
    "learning_rate": 0.00048037009212046597,
    "timestep_mean": 430.0,
    "epsilon_norm": 7.642580986022949,
    "prediction_norm": 6.915725231170654,
    "timestamp": "2025-07-14T12:34:10.047794"
  },
  {
    "epoch": 1025,
    "batch_idx": 160,
    "loss": 0.11915893107652664,
    "learning_rate": 0.00048037009212046597,
    "timestep_mean": 564.09375,
    "epsilon_norm": 7.952482223510742,
    "prediction_norm": 7.598212718963623,
    "timestamp": "2025-07-14T12:34:10.052664"
  },
  {
    "epoch": 1025,
    "batch_idx": 192,
    "loss": 0.09909114986658096,
    "learning_rate": 0.00048037009212046597,
    "timestep_mean": 548.375,
    "epsilon_norm": 2.0412983894348145,
    "prediction_norm": 1.700218915939331,
    "timestamp": "2025-07-14T12:34:10.057744"
  },
  {
    "epoch": 1026,
    "batch_idx": 0,
    "loss": 0.33075594902038574,
    "learning_rate": 0.00047958532401074525,
    "timestep_mean": 548.25,
    "epsilon_norm": 8.184621810913086,
    "prediction_norm": 7.1310014724731445,
    "timestamp": "2025-07-14T12:34:10.062860"
  },
  {
    "epoch": 1026,
    "batch_idx": 32,
    "loss": 0.36983680725097656,
    "learning_rate": 0.00047958532401074525,
    "timestep_mean": 406.125,
    "epsilon_norm": 7.537572860717773,
    "prediction_norm": 5.2014241218566895,
    "timestamp": "2025-07-14T12:34:10.067746"
  },
  {
    "epoch": 1026,
    "batch_idx": 64,
    "loss": 0.4730314612388611,
    "learning_rate": 0.00047958532401074525,
    "timestep_mean": 390.84375,
    "epsilon_norm": 7.7315850257873535,
    "prediction_norm": 6.05501651763916,
    "timestamp": "2025-07-14T12:34:10.072686"
  },
  {
    "epoch": 1026,
    "batch_idx": 96,
    "loss": 0.3114309310913086,
    "learning_rate": 0.00047958532401074525,
    "timestep_mean": 473.5625,
    "epsilon_norm": 7.167901515960693,
    "prediction_norm": 5.727126598358154,
    "timestamp": "2025-07-14T12:34:10.077784"
  },
  {
    "epoch": 1026,
    "batch_idx": 128,
    "loss": 0.22434596717357635,
    "learning_rate": 0.00047958532401074525,
    "timestep_mean": 512.875,
    "epsilon_norm": 7.331230640411377,
    "prediction_norm": 6.038510799407959,
    "timestamp": "2025-07-14T12:34:10.083014"
  },
  {
    "epoch": 1026,
    "batch_idx": 160,
    "loss": 0.4004991352558136,
    "learning_rate": 0.00047958532401074525,
    "timestep_mean": 581.4375,
    "epsilon_norm": 8.014989852905273,
    "prediction_norm": 5.964529037475586,
    "timestamp": "2025-07-14T12:34:10.088399"
  },
  {
    "epoch": 1026,
    "batch_idx": 192,
    "loss": 0.6553140878677368,
    "learning_rate": 0.00047958532401074525,
    "timestep_mean": 315.625,
    "epsilon_norm": 4.050954341888428,
    "prediction_norm": 2.3996965885162354,
    "timestamp": "2025-07-14T12:34:10.093629"
  },
  {
    "epoch": 1027,
    "batch_idx": 0,
    "loss": 0.2473561018705368,
    "learning_rate": 0.0004788006062722082,
    "timestep_mean": 566.0,
    "epsilon_norm": 7.9254961013793945,
    "prediction_norm": 6.828620910644531,
    "timestamp": "2025-07-14T12:34:10.098843"
  },
  {
    "epoch": 1027,
    "batch_idx": 32,
    "loss": 0.29925012588500977,
    "learning_rate": 0.0004788006062722082,
    "timestep_mean": 495.4375,
    "epsilon_norm": 8.816325187683105,
    "prediction_norm": 7.348243713378906,
    "timestamp": "2025-07-14T12:34:10.103843"
  },
  {
    "epoch": 1027,
    "batch_idx": 64,
    "loss": 0.22629401087760925,
    "learning_rate": 0.0004788006062722082,
    "timestep_mean": 511.59375,
    "epsilon_norm": 7.596735954284668,
    "prediction_norm": 6.050281524658203,
    "timestamp": "2025-07-14T12:34:10.109049"
  },
  {
    "epoch": 1027,
    "batch_idx": 96,
    "loss": 0.20765112340450287,
    "learning_rate": 0.0004788006062722082,
    "timestep_mean": 501.625,
    "epsilon_norm": 8.52243709564209,
    "prediction_norm": 7.693053722381592,
    "timestamp": "2025-07-14T12:34:10.114190"
  },
  {
    "epoch": 1027,
    "batch_idx": 128,
    "loss": 0.4323887228965759,
    "learning_rate": 0.0004788006062722082,
    "timestep_mean": 404.96875,
    "epsilon_norm": 8.37382698059082,
    "prediction_norm": 6.135937690734863,
    "timestamp": "2025-07-14T12:34:10.119433"
  },
  {
    "epoch": 1027,
    "batch_idx": 160,
    "loss": 0.4490775465965271,
    "learning_rate": 0.0004788006062722082,
    "timestep_mean": 400.28125,
    "epsilon_norm": 8.735067367553711,
    "prediction_norm": 6.761065483093262,
    "timestamp": "2025-07-14T12:34:10.124628"
  },
  {
    "epoch": 1027,
    "batch_idx": 192,
    "loss": 0.05861834064126015,
    "learning_rate": 0.0004788006062722082,
    "timestep_mean": 624.375,
    "epsilon_norm": 3.29036808013916,
    "prediction_norm": 2.87226939201355,
    "timestamp": "2025-07-14T12:34:10.130113"
  },
  {
    "epoch": 1028,
    "batch_idx": 0,
    "loss": 0.2580247223377228,
    "learning_rate": 0.0004780159408410679,
    "timestep_mean": 507.84375,
    "epsilon_norm": 6.756796836853027,
    "prediction_norm": 5.362112522125244,
    "timestamp": "2025-07-14T12:34:10.136673"
  },
  {
    "epoch": 1028,
    "batch_idx": 32,
    "loss": 0.21433371305465698,
    "learning_rate": 0.0004780159408410679,
    "timestep_mean": 506.4375,
    "epsilon_norm": 8.421540260314941,
    "prediction_norm": 7.626045227050781,
    "timestamp": "2025-07-14T12:34:10.142075"
  },
  {
    "epoch": 1028,
    "batch_idx": 64,
    "loss": 0.3177400231361389,
    "learning_rate": 0.0004780159408410679,
    "timestep_mean": 455.15625,
    "epsilon_norm": 8.247633934020996,
    "prediction_norm": 7.457345485687256,
    "timestamp": "2025-07-14T12:34:10.147156"
  },
  {
    "epoch": 1028,
    "batch_idx": 96,
    "loss": 0.37182825803756714,
    "learning_rate": 0.0004780159408410679,
    "timestep_mean": 484.53125,
    "epsilon_norm": 8.622915267944336,
    "prediction_norm": 7.697105884552002,
    "timestamp": "2025-07-14T12:34:10.152479"
  },
  {
    "epoch": 1028,
    "batch_idx": 128,
    "loss": 0.25408118963241577,
    "learning_rate": 0.0004780159408410679,
    "timestep_mean": 434.75,
    "epsilon_norm": 9.200549125671387,
    "prediction_norm": 7.633738040924072,
    "timestamp": "2025-07-14T12:34:10.157720"
  },
  {
    "epoch": 1028,
    "batch_idx": 160,
    "loss": 0.4394250512123108,
    "learning_rate": 0.0004780159408410679,
    "timestep_mean": 406.96875,
    "epsilon_norm": 9.324000358581543,
    "prediction_norm": 7.428369522094727,
    "timestamp": "2025-07-14T12:34:10.162995"
  },
  {
    "epoch": 1028,
    "batch_idx": 192,
    "loss": 0.03995309770107269,
    "learning_rate": 0.0004780159408410679,
    "timestep_mean": 714.25,
    "epsilon_norm": 5.229879856109619,
    "prediction_norm": 5.1782612800598145,
    "timestamp": "2025-07-14T12:34:10.168258"
  },
  {
    "epoch": 1029,
    "batch_idx": 0,
    "loss": 0.21974602341651917,
    "learning_rate": 0.00047723132965340814,
    "timestep_mean": 509.125,
    "epsilon_norm": 9.774194717407227,
    "prediction_norm": 9.351446151733398,
    "timestamp": "2025-07-14T12:34:10.173487"
  },
  {
    "epoch": 1029,
    "batch_idx": 32,
    "loss": 0.2108537256717682,
    "learning_rate": 0.00047723132965340814,
    "timestep_mean": 437.53125,
    "epsilon_norm": 9.098711967468262,
    "prediction_norm": 8.455828666687012,
    "timestamp": "2025-07-14T12:34:10.178744"
  },
  {
    "epoch": 1029,
    "batch_idx": 64,
    "loss": 0.2614525556564331,
    "learning_rate": 0.00047723132965340814,
    "timestep_mean": 507.6875,
    "epsilon_norm": 8.168014526367188,
    "prediction_norm": 7.070721626281738,
    "timestamp": "2025-07-14T12:34:10.184166"
  },
  {
    "epoch": 1029,
    "batch_idx": 96,
    "loss": 0.3437057137489319,
    "learning_rate": 0.00047723132965340814,
    "timestep_mean": 459.25,
    "epsilon_norm": 7.830313682556152,
    "prediction_norm": 6.676407337188721,
    "timestamp": "2025-07-14T12:34:10.189392"
  },
  {
    "epoch": 1029,
    "batch_idx": 128,
    "loss": 0.17612788081169128,
    "learning_rate": 0.00047723132965340814,
    "timestep_mean": 529.9375,
    "epsilon_norm": 7.834199905395508,
    "prediction_norm": 7.615913391113281,
    "timestamp": "2025-07-14T12:34:10.194456"
  },
  {
    "epoch": 1029,
    "batch_idx": 160,
    "loss": 0.174329936504364,
    "learning_rate": 0.00047723132965340814,
    "timestep_mean": 571.21875,
    "epsilon_norm": 6.878073215484619,
    "prediction_norm": 6.406455039978027,
    "timestamp": "2025-07-14T12:34:10.199600"
  },
  {
    "epoch": 1029,
    "batch_idx": 192,
    "loss": 0.46415501832962036,
    "learning_rate": 0.00047723132965340814,
    "timestep_mean": 527.75,
    "epsilon_norm": 5.08112907409668,
    "prediction_norm": 4.3279595375061035,
    "timestamp": "2025-07-14T12:34:10.204701"
  },
  {
    "epoch": 1030,
    "batch_idx": 0,
    "loss": 0.18380922079086304,
    "learning_rate": 0.00047644677464517884,
    "timestep_mean": 601.25,
    "epsilon_norm": 7.316874027252197,
    "prediction_norm": 5.944971561431885,
    "timestamp": "2025-07-14T12:34:10.209758"
  },
  {
    "epoch": 1030,
    "batch_idx": 32,
    "loss": 0.20764708518981934,
    "learning_rate": 0.00047644677464517884,
    "timestep_mean": 441.6875,
    "epsilon_norm": 6.786076068878174,
    "prediction_norm": 5.43186616897583,
    "timestamp": "2025-07-14T12:34:10.214754"
  },
  {
    "epoch": 1030,
    "batch_idx": 64,
    "loss": 0.1135726273059845,
    "learning_rate": 0.00047644677464517884,
    "timestep_mean": 557.9375,
    "epsilon_norm": 9.281375885009766,
    "prediction_norm": 8.39189624786377,
    "timestamp": "2025-07-14T12:34:10.219908"
  },
  {
    "epoch": 1030,
    "batch_idx": 96,
    "loss": 0.1566125899553299,
    "learning_rate": 0.00047644677464517884,
    "timestep_mean": 611.9375,
    "epsilon_norm": 7.813224792480469,
    "prediction_norm": 6.438268184661865,
    "timestamp": "2025-07-14T12:34:10.224866"
  },
  {
    "epoch": 1030,
    "batch_idx": 128,
    "loss": 0.22345933318138123,
    "learning_rate": 0.00047644677464517884,
    "timestep_mean": 504.03125,
    "epsilon_norm": 7.3901238441467285,
    "prediction_norm": 6.553089141845703,
    "timestamp": "2025-07-14T12:34:10.229892"
  },
  {
    "epoch": 1030,
    "batch_idx": 160,
    "loss": 0.2513796091079712,
    "learning_rate": 0.00047644677464517884,
    "timestep_mean": 610.75,
    "epsilon_norm": 9.190534591674805,
    "prediction_norm": 8.05018138885498,
    "timestamp": "2025-07-14T12:34:10.235029"
  },
  {
    "epoch": 1030,
    "batch_idx": 192,
    "loss": 0.09021040797233582,
    "learning_rate": 0.00047644677464517884,
    "timestep_mean": 564.375,
    "epsilon_norm": 3.805846691131592,
    "prediction_norm": 3.7077853679656982,
    "timestamp": "2025-07-14T12:34:10.240173"
  },
  {
    "epoch": 1031,
    "batch_idx": 0,
    "loss": 0.20222042500972748,
    "learning_rate": 0.00047566227775219203,
    "timestep_mean": 504.71875,
    "epsilon_norm": 7.931602954864502,
    "prediction_norm": 6.783301830291748,
    "timestamp": "2025-07-14T12:34:10.245415"
  },
  {
    "epoch": 1031,
    "batch_idx": 32,
    "loss": 0.23865020275115967,
    "learning_rate": 0.00047566227775219203,
    "timestep_mean": 471.25,
    "epsilon_norm": 7.5450215339660645,
    "prediction_norm": 6.530848026275635,
    "timestamp": "2025-07-14T12:34:10.250663"
  },
  {
    "epoch": 1031,
    "batch_idx": 64,
    "loss": 0.23799902200698853,
    "learning_rate": 0.00047566227775219203,
    "timestep_mean": 581.28125,
    "epsilon_norm": 7.414296627044678,
    "prediction_norm": 6.690535068511963,
    "timestamp": "2025-07-14T12:34:10.255903"
  },
  {
    "epoch": 1031,
    "batch_idx": 96,
    "loss": 0.2971858084201813,
    "learning_rate": 0.00047566227775219203,
    "timestep_mean": 447.65625,
    "epsilon_norm": 7.665524482727051,
    "prediction_norm": 5.717197895050049,
    "timestamp": "2025-07-14T12:34:10.261165"
  },
  {
    "epoch": 1031,
    "batch_idx": 128,
    "loss": 0.38519763946533203,
    "learning_rate": 0.00047566227775219203,
    "timestep_mean": 464.15625,
    "epsilon_norm": 8.934535026550293,
    "prediction_norm": 7.080615043640137,
    "timestamp": "2025-07-14T12:34:10.266370"
  },
  {
    "epoch": 1031,
    "batch_idx": 160,
    "loss": 0.38189584016799927,
    "learning_rate": 0.00047566227775219203,
    "timestep_mean": 481.4375,
    "epsilon_norm": 7.930834770202637,
    "prediction_norm": 5.637673377990723,
    "timestamp": "2025-07-14T12:34:10.271636"
  },
  {
    "epoch": 1031,
    "batch_idx": 192,
    "loss": 0.06918667256832123,
    "learning_rate": 0.00047566227775219203,
    "timestep_mean": 727.625,
    "epsilon_norm": 4.391293525695801,
    "prediction_norm": 4.3998565673828125,
    "timestamp": "2025-07-14T12:34:10.276831"
  },
  {
    "epoch": 1032,
    "batch_idx": 0,
    "loss": 0.41458770632743835,
    "learning_rate": 0.0004748778409101155,
    "timestep_mean": 381.59375,
    "epsilon_norm": 8.030590057373047,
    "prediction_norm": 5.787827014923096,
    "timestamp": "2025-07-14T12:34:10.282119"
  },
  {
    "epoch": 1032,
    "batch_idx": 32,
    "loss": 0.1854129582643509,
    "learning_rate": 0.0004748778409101155,
    "timestep_mean": 500.0,
    "epsilon_norm": 7.260784149169922,
    "prediction_norm": 7.02962064743042,
    "timestamp": "2025-07-14T12:34:10.287715"
  },
  {
    "epoch": 1032,
    "batch_idx": 64,
    "loss": 0.22968612611293793,
    "learning_rate": 0.0004748778409101155,
    "timestep_mean": 630.46875,
    "epsilon_norm": 8.404470443725586,
    "prediction_norm": 7.442687511444092,
    "timestamp": "2025-07-14T12:34:10.292873"
  },
  {
    "epoch": 1032,
    "batch_idx": 96,
    "loss": 0.22854948043823242,
    "learning_rate": 0.0004748778409101155,
    "timestep_mean": 490.96875,
    "epsilon_norm": 7.906173229217529,
    "prediction_norm": 7.203996658325195,
    "timestamp": "2025-07-14T12:34:10.298009"
  },
  {
    "epoch": 1032,
    "batch_idx": 128,
    "loss": 0.33106565475463867,
    "learning_rate": 0.0004748778409101155,
    "timestep_mean": 487.71875,
    "epsilon_norm": 7.055789470672607,
    "prediction_norm": 5.751306056976318,
    "timestamp": "2025-07-14T12:34:10.303610"
  },
  {
    "epoch": 1032,
    "batch_idx": 160,
    "loss": 0.5400806069374084,
    "learning_rate": 0.0004748778409101155,
    "timestep_mean": 425.96875,
    "epsilon_norm": 8.296212196350098,
    "prediction_norm": 6.0534138679504395,
    "timestamp": "2025-07-14T12:34:10.310493"
  },
  {
    "epoch": 1032,
    "batch_idx": 192,
    "loss": 0.5141645073890686,
    "learning_rate": 0.0004748778409101155,
    "timestep_mean": 351.625,
    "epsilon_norm": 4.132016181945801,
    "prediction_norm": 2.643956422805786,
    "timestamp": "2025-07-14T12:34:10.319112"
  },
  {
    "epoch": 1033,
    "batch_idx": 0,
    "loss": 0.2854403257369995,
    "learning_rate": 0.00047409346605446914,
    "timestep_mean": 465.65625,
    "epsilon_norm": 7.803406238555908,
    "prediction_norm": 6.698180675506592,
    "timestamp": "2025-07-14T12:34:10.331666"
  },
  {
    "epoch": 1033,
    "batch_idx": 32,
    "loss": 0.14581730961799622,
    "learning_rate": 0.00047409346605446914,
    "timestep_mean": 542.71875,
    "epsilon_norm": 6.309967994689941,
    "prediction_norm": 6.344832420349121,
    "timestamp": "2025-07-14T12:34:10.338597"
  },
  {
    "epoch": 1033,
    "batch_idx": 64,
    "loss": 0.330138623714447,
    "learning_rate": 0.00047409346605446914,
    "timestep_mean": 530.75,
    "epsilon_norm": 8.139998435974121,
    "prediction_norm": 6.833916187286377,
    "timestamp": "2025-07-14T12:34:10.349793"
  },
  {
    "epoch": 1033,
    "batch_idx": 96,
    "loss": 0.23468798398971558,
    "learning_rate": 0.00047409346605446914,
    "timestep_mean": 493.78125,
    "epsilon_norm": 7.887133598327637,
    "prediction_norm": 6.781219005584717,
    "timestamp": "2025-07-14T12:34:10.356117"
  },
  {
    "epoch": 1033,
    "batch_idx": 128,
    "loss": 0.20499631762504578,
    "learning_rate": 0.00047409346605446914,
    "timestep_mean": 459.8125,
    "epsilon_norm": 7.786798477172852,
    "prediction_norm": 6.73187255859375,
    "timestamp": "2025-07-14T12:34:10.361627"
  },
  {
    "epoch": 1033,
    "batch_idx": 160,
    "loss": 0.1546463966369629,
    "learning_rate": 0.00047409346605446914,
    "timestep_mean": 602.53125,
    "epsilon_norm": 7.893558025360107,
    "prediction_norm": 7.497856616973877,
    "timestamp": "2025-07-14T12:34:10.367027"
  },
  {
    "epoch": 1033,
    "batch_idx": 192,
    "loss": 0.20425285398960114,
    "learning_rate": 0.00047409346605446914,
    "timestep_mean": 454.75,
    "epsilon_norm": 4.2540788650512695,
    "prediction_norm": 3.5539932250976562,
    "timestamp": "2025-07-14T12:34:10.372610"
  },
  {
    "epoch": 1034,
    "batch_idx": 0,
    "loss": 0.2536720633506775,
    "learning_rate": 0.00047330915512061995,
    "timestep_mean": 444.5625,
    "epsilon_norm": 6.882107257843018,
    "prediction_norm": 5.752206325531006,
    "timestamp": "2025-07-14T12:34:10.377901"
  },
  {
    "epoch": 1034,
    "batch_idx": 32,
    "loss": 0.2920513153076172,
    "learning_rate": 0.00047330915512061995,
    "timestep_mean": 480.21875,
    "epsilon_norm": 8.7804536819458,
    "prediction_norm": 7.229514122009277,
    "timestamp": "2025-07-14T12:34:10.383042"
  },
  {
    "epoch": 1034,
    "batch_idx": 64,
    "loss": 0.23666705191135406,
    "learning_rate": 0.00047330915512061995,
    "timestep_mean": 501.78125,
    "epsilon_norm": 7.320496082305908,
    "prediction_norm": 5.616369247436523,
    "timestamp": "2025-07-14T12:34:10.388396"
  },
  {
    "epoch": 1034,
    "batch_idx": 96,
    "loss": 0.28280240297317505,
    "learning_rate": 0.00047330915512061995,
    "timestep_mean": 533.90625,
    "epsilon_norm": 8.3128662109375,
    "prediction_norm": 6.889628887176514,
    "timestamp": "2025-07-14T12:34:10.393722"
  },
  {
    "epoch": 1034,
    "batch_idx": 128,
    "loss": 0.4272058308124542,
    "learning_rate": 0.00047330915512061995,
    "timestep_mean": 475.09375,
    "epsilon_norm": 8.923237800598145,
    "prediction_norm": 7.350949287414551,
    "timestamp": "2025-07-14T12:34:10.399176"
  },
  {
    "epoch": 1034,
    "batch_idx": 160,
    "loss": 0.4201843738555908,
    "learning_rate": 0.00047330915512061995,
    "timestep_mean": 515.40625,
    "epsilon_norm": 9.975580215454102,
    "prediction_norm": 8.702022552490234,
    "timestamp": "2025-07-14T12:34:10.404631"
  },
  {
    "epoch": 1034,
    "batch_idx": 192,
    "loss": 0.49306175112724304,
    "learning_rate": 0.00047330915512061995,
    "timestep_mean": 520.5,
    "epsilon_norm": 4.474735260009766,
    "prediction_norm": 4.146812915802002,
    "timestamp": "2025-07-14T12:34:10.410246"
  },
  {
    "epoch": 1035,
    "batch_idx": 0,
    "loss": 0.27273377776145935,
    "learning_rate": 0.0004725249100437774,
    "timestep_mean": 530.125,
    "epsilon_norm": 8.396310806274414,
    "prediction_norm": 6.489657878875732,
    "timestamp": "2025-07-14T12:34:10.416036"
  },
  {
    "epoch": 1035,
    "batch_idx": 32,
    "loss": 0.16064003109931946,
    "learning_rate": 0.0004725249100437774,
    "timestep_mean": 571.1875,
    "epsilon_norm": 7.372316837310791,
    "prediction_norm": 7.114142894744873,
    "timestamp": "2025-07-14T12:34:10.421637"
  },
  {
    "epoch": 1035,
    "batch_idx": 64,
    "loss": 0.310115247964859,
    "learning_rate": 0.0004725249100437774,
    "timestep_mean": 473.78125,
    "epsilon_norm": 8.404250144958496,
    "prediction_norm": 6.99122953414917,
    "timestamp": "2025-07-14T12:34:10.427309"
  },
  {
    "epoch": 1035,
    "batch_idx": 96,
    "loss": 0.30737608671188354,
    "learning_rate": 0.0004725249100437774,
    "timestep_mean": 525.59375,
    "epsilon_norm": 8.354394912719727,
    "prediction_norm": 7.345548152923584,
    "timestamp": "2025-07-14T12:34:10.432814"
  },
  {
    "epoch": 1035,
    "batch_idx": 128,
    "loss": 0.38097184896469116,
    "learning_rate": 0.0004725249100437774,
    "timestep_mean": 460.09375,
    "epsilon_norm": 7.999151706695557,
    "prediction_norm": 6.455720901489258,
    "timestamp": "2025-07-14T12:34:10.438397"
  },
  {
    "epoch": 1035,
    "batch_idx": 160,
    "loss": 0.346630334854126,
    "learning_rate": 0.0004725249100437774,
    "timestep_mean": 461.0,
    "epsilon_norm": 9.060673713684082,
    "prediction_norm": 6.942296028137207,
    "timestamp": "2025-07-14T12:34:10.443715"
  },
  {
    "epoch": 1035,
    "batch_idx": 192,
    "loss": 0.1487862765789032,
    "learning_rate": 0.0004725249100437774,
    "timestep_mean": 457.875,
    "epsilon_norm": 3.137678623199463,
    "prediction_norm": 2.479081869125366,
    "timestamp": "2025-07-14T12:34:10.449108"
  },
  {
    "epoch": 1036,
    "batch_idx": 0,
    "loss": 0.23997285962104797,
    "learning_rate": 0.00047174073275898793,
    "timestep_mean": 558.3125,
    "epsilon_norm": 9.090974807739258,
    "prediction_norm": 8.015878677368164,
    "timestamp": "2025-07-14T12:34:10.454621"
  },
  {
    "epoch": 1036,
    "batch_idx": 32,
    "loss": 0.3072739243507385,
    "learning_rate": 0.00047174073275898793,
    "timestep_mean": 454.0,
    "epsilon_norm": 7.096617221832275,
    "prediction_norm": 5.9827470779418945,
    "timestamp": "2025-07-14T12:34:10.460037"
  },
  {
    "epoch": 1036,
    "batch_idx": 64,
    "loss": 0.526341438293457,
    "learning_rate": 0.00047174073275898793,
    "timestep_mean": 369.25,
    "epsilon_norm": 9.31405258178711,
    "prediction_norm": 6.877439498901367,
    "timestamp": "2025-07-14T12:34:10.465338"
  },
  {
    "epoch": 1036,
    "batch_idx": 96,
    "loss": 0.5030121803283691,
    "learning_rate": 0.00047174073275898793,
    "timestep_mean": 522.375,
    "epsilon_norm": 8.288113594055176,
    "prediction_norm": 6.45114803314209,
    "timestamp": "2025-07-14T12:34:10.470605"
  },
  {
    "epoch": 1036,
    "batch_idx": 128,
    "loss": 0.32505008578300476,
    "learning_rate": 0.00047174073275898793,
    "timestep_mean": 476.5,
    "epsilon_norm": 7.121033191680908,
    "prediction_norm": 5.937490463256836,
    "timestamp": "2025-07-14T12:34:10.475970"
  },
  {
    "epoch": 1036,
    "batch_idx": 160,
    "loss": 0.16915488243103027,
    "learning_rate": 0.00047174073275898793,
    "timestep_mean": 604.625,
    "epsilon_norm": 6.9984025955200195,
    "prediction_norm": 6.48789644241333,
    "timestamp": "2025-07-14T12:34:10.481229"
  },
  {
    "epoch": 1036,
    "batch_idx": 192,
    "loss": 0.19928522408008575,
    "learning_rate": 0.00047174073275898793,
    "timestep_mean": 339.25,
    "epsilon_norm": 4.3729166984558105,
    "prediction_norm": 4.317724704742432,
    "timestamp": "2025-07-14T12:34:10.486460"
  },
  {
    "epoch": 1037,
    "batch_idx": 0,
    "loss": 0.18992403149604797,
    "learning_rate": 0.0004709566252011312,
    "timestep_mean": 467.34375,
    "epsilon_norm": 7.224102020263672,
    "prediction_norm": 6.204415798187256,
    "timestamp": "2025-07-14T12:34:10.491721"
  },
  {
    "epoch": 1037,
    "batch_idx": 32,
    "loss": 0.3092695474624634,
    "learning_rate": 0.0004709566252011312,
    "timestep_mean": 454.09375,
    "epsilon_norm": 7.505395889282227,
    "prediction_norm": 6.528497219085693,
    "timestamp": "2025-07-14T12:34:10.496932"
  },
  {
    "epoch": 1037,
    "batch_idx": 64,
    "loss": 0.23750843107700348,
    "learning_rate": 0.0004709566252011312,
    "timestep_mean": 435.0625,
    "epsilon_norm": 7.833956718444824,
    "prediction_norm": 6.768405914306641,
    "timestamp": "2025-07-14T12:34:10.502380"
  },
  {
    "epoch": 1037,
    "batch_idx": 96,
    "loss": 0.3521963655948639,
    "learning_rate": 0.0004709566252011312,
    "timestep_mean": 546.40625,
    "epsilon_norm": 7.676385402679443,
    "prediction_norm": 6.039975643157959,
    "timestamp": "2025-07-14T12:34:10.507541"
  },
  {
    "epoch": 1037,
    "batch_idx": 128,
    "loss": 0.32317590713500977,
    "learning_rate": 0.0004709566252011312,
    "timestep_mean": 509.84375,
    "epsilon_norm": 7.970900058746338,
    "prediction_norm": 6.0545759201049805,
    "timestamp": "2025-07-14T12:34:10.513036"
  },
  {
    "epoch": 1037,
    "batch_idx": 160,
    "loss": 0.28958022594451904,
    "learning_rate": 0.0004709566252011312,
    "timestep_mean": 528.25,
    "epsilon_norm": 7.382652759552002,
    "prediction_norm": 6.821691513061523,
    "timestamp": "2025-07-14T12:34:10.518265"
  },
  {
    "epoch": 1037,
    "batch_idx": 192,
    "loss": 0.12647494673728943,
    "learning_rate": 0.0004709566252011312,
    "timestep_mean": 590.625,
    "epsilon_norm": 3.160440683364868,
    "prediction_norm": 2.2324161529541016,
    "timestamp": "2025-07-14T12:34:10.530083"
  },
  {
    "epoch": 1038,
    "batch_idx": 0,
    "loss": 0.2840549647808075,
    "learning_rate": 0.0004701725893049148,
    "timestep_mean": 570.28125,
    "epsilon_norm": 7.0224928855896,
    "prediction_norm": 6.044939994812012,
    "timestamp": "2025-07-14T12:34:10.535283"
  },
  {
    "epoch": 1038,
    "batch_idx": 32,
    "loss": 0.30100515484809875,
    "learning_rate": 0.0004701725893049148,
    "timestep_mean": 437.53125,
    "epsilon_norm": 7.8470258712768555,
    "prediction_norm": 6.7327704429626465,
    "timestamp": "2025-07-14T12:34:10.540365"
  },
  {
    "epoch": 1038,
    "batch_idx": 64,
    "loss": 0.28270837664604187,
    "learning_rate": 0.0004701725893049148,
    "timestep_mean": 445.90625,
    "epsilon_norm": 7.914741516113281,
    "prediction_norm": 7.285547733306885,
    "timestamp": "2025-07-14T12:34:10.545423"
  },
  {
    "epoch": 1038,
    "batch_idx": 96,
    "loss": 0.2751234173774719,
    "learning_rate": 0.0004701725893049148,
    "timestep_mean": 498.84375,
    "epsilon_norm": 9.095730781555176,
    "prediction_norm": 8.105725288391113,
    "timestamp": "2025-07-14T12:34:10.550581"
  },
  {
    "epoch": 1038,
    "batch_idx": 128,
    "loss": 0.346200168132782,
    "learning_rate": 0.0004701725893049148,
    "timestep_mean": 477.15625,
    "epsilon_norm": 8.630537033081055,
    "prediction_norm": 7.621845722198486,
    "timestamp": "2025-07-14T12:34:10.555734"
  },
  {
    "epoch": 1038,
    "batch_idx": 160,
    "loss": 0.5388516187667847,
    "learning_rate": 0.0004701725893049148,
    "timestep_mean": 451.15625,
    "epsilon_norm": 8.270692825317383,
    "prediction_norm": 5.706432819366455,
    "timestamp": "2025-07-14T12:34:10.560721"
  },
  {
    "epoch": 1038,
    "batch_idx": 192,
    "loss": 0.2733820974826813,
    "learning_rate": 0.0004701725893049148,
    "timestep_mean": 499.125,
    "epsilon_norm": 4.104096412658691,
    "prediction_norm": 3.7534186840057373,
    "timestamp": "2025-07-14T12:34:10.565470"
  },
  {
    "epoch": 1039,
    "batch_idx": 0,
    "loss": 0.3242674469947815,
    "learning_rate": 0.00046938862700486925,
    "timestep_mean": 402.21875,
    "epsilon_norm": 7.407638072967529,
    "prediction_norm": 5.861660957336426,
    "timestamp": "2025-07-14T12:34:10.570211"
  },
  {
    "epoch": 1039,
    "batch_idx": 32,
    "loss": 0.2560707926750183,
    "learning_rate": 0.00046938862700486925,
    "timestep_mean": 482.4375,
    "epsilon_norm": 7.695980072021484,
    "prediction_norm": 6.459797382354736,
    "timestamp": "2025-07-14T12:34:10.575045"
  },
  {
    "epoch": 1039,
    "batch_idx": 64,
    "loss": 0.35102999210357666,
    "learning_rate": 0.00046938862700486925,
    "timestep_mean": 480.125,
    "epsilon_norm": 7.958700656890869,
    "prediction_norm": 6.9619951248168945,
    "timestamp": "2025-07-14T12:34:10.579815"
  },
  {
    "epoch": 1039,
    "batch_idx": 96,
    "loss": 0.2951047420501709,
    "learning_rate": 0.00046938862700486925,
    "timestep_mean": 458.15625,
    "epsilon_norm": 8.19029426574707,
    "prediction_norm": 6.780386447906494,
    "timestamp": "2025-07-14T12:34:10.584517"
  },
  {
    "epoch": 1039,
    "batch_idx": 128,
    "loss": 0.26042211055755615,
    "learning_rate": 0.00046938862700486925,
    "timestep_mean": 554.15625,
    "epsilon_norm": 7.8155198097229,
    "prediction_norm": 7.2160162925720215,
    "timestamp": "2025-07-14T12:34:10.589309"
  },
  {
    "epoch": 1039,
    "batch_idx": 160,
    "loss": 0.2947160005569458,
    "learning_rate": 0.00046938862700486925,
    "timestep_mean": 423.34375,
    "epsilon_norm": 6.874283790588379,
    "prediction_norm": 5.6341352462768555,
    "timestamp": "2025-07-14T12:34:10.594018"
  },
  {
    "epoch": 1039,
    "batch_idx": 192,
    "loss": 0.3027111887931824,
    "learning_rate": 0.00046938862700486925,
    "timestep_mean": 429.75,
    "epsilon_norm": 2.8809914588928223,
    "prediction_norm": 2.537407159805298,
    "timestamp": "2025-07-14T12:34:10.598794"
  },
  {
    "epoch": 1040,
    "batch_idx": 0,
    "loss": 0.5314146876335144,
    "learning_rate": 0.0004686047402353436,
    "timestep_mean": 427.53125,
    "epsilon_norm": 8.073526382446289,
    "prediction_norm": 5.23887300491333,
    "timestamp": "2025-07-14T12:34:10.603532"
  },
  {
    "epoch": 1040,
    "batch_idx": 32,
    "loss": 0.3250550925731659,
    "learning_rate": 0.0004686047402353436,
    "timestep_mean": 504.28125,
    "epsilon_norm": 8.15651798248291,
    "prediction_norm": 6.4346022605896,
    "timestamp": "2025-07-14T12:34:10.608395"
  },
  {
    "epoch": 1040,
    "batch_idx": 64,
    "loss": 0.3511391878128052,
    "learning_rate": 0.0004686047402353436,
    "timestep_mean": 536.6875,
    "epsilon_norm": 8.380047798156738,
    "prediction_norm": 6.233269691467285,
    "timestamp": "2025-07-14T12:34:10.613166"
  },
  {
    "epoch": 1040,
    "batch_idx": 96,
    "loss": 0.2441416084766388,
    "learning_rate": 0.0004686047402353436,
    "timestep_mean": 568.65625,
    "epsilon_norm": 8.184212684631348,
    "prediction_norm": 6.498047828674316,
    "timestamp": "2025-07-14T12:34:10.617900"
  },
  {
    "epoch": 1040,
    "batch_idx": 128,
    "loss": 0.323494017124176,
    "learning_rate": 0.0004686047402353436,
    "timestep_mean": 485.3125,
    "epsilon_norm": 7.47862434387207,
    "prediction_norm": 5.8606858253479,
    "timestamp": "2025-07-14T12:34:10.622716"
  },
  {
    "epoch": 1040,
    "batch_idx": 160,
    "loss": 0.43660205602645874,
    "learning_rate": 0.0004686047402353436,
    "timestep_mean": 536.875,
    "epsilon_norm": 7.979630947113037,
    "prediction_norm": 5.658843994140625,
    "timestamp": "2025-07-14T12:34:10.627490"
  },
  {
    "epoch": 1040,
    "batch_idx": 192,
    "loss": 0.17489418387413025,
    "learning_rate": 0.0004686047402353436,
    "timestep_mean": 523.625,
    "epsilon_norm": 3.8153512477874756,
    "prediction_norm": 2.9958834648132324,
    "timestamp": "2025-07-14T12:34:10.632234"
  },
  {
    "epoch": 1041,
    "batch_idx": 0,
    "loss": 0.21757487952709198,
    "learning_rate": 0.00046782093093050036,
    "timestep_mean": 552.125,
    "epsilon_norm": 8.251734733581543,
    "prediction_norm": 7.199759006500244,
    "timestamp": "2025-07-14T12:34:10.637036"
  },
  {
    "epoch": 1041,
    "batch_idx": 32,
    "loss": 0.4287061393260956,
    "learning_rate": 0.00046782093093050036,
    "timestep_mean": 421.5,
    "epsilon_norm": 8.469320297241211,
    "prediction_norm": 6.290249347686768,
    "timestamp": "2025-07-14T12:34:10.641913"
  },
  {
    "epoch": 1041,
    "batch_idx": 64,
    "loss": 0.16000083088874817,
    "learning_rate": 0.00046782093093050036,
    "timestep_mean": 551.53125,
    "epsilon_norm": 8.759151458740234,
    "prediction_norm": 7.565838813781738,
    "timestamp": "2025-07-14T12:34:10.646898"
  },
  {
    "epoch": 1041,
    "batch_idx": 96,
    "loss": 0.2312690019607544,
    "learning_rate": 0.00046782093093050036,
    "timestep_mean": 487.78125,
    "epsilon_norm": 7.212127685546875,
    "prediction_norm": 5.8914475440979,
    "timestamp": "2025-07-14T12:34:10.651698"
  },
  {
    "epoch": 1041,
    "batch_idx": 128,
    "loss": 0.5726473927497864,
    "learning_rate": 0.00046782093093050036,
    "timestep_mean": 430.03125,
    "epsilon_norm": 9.062542915344238,
    "prediction_norm": 6.778321743011475,
    "timestamp": "2025-07-14T12:34:10.656594"
  },
  {
    "epoch": 1041,
    "batch_idx": 160,
    "loss": 0.21088138222694397,
    "learning_rate": 0.00046782093093050036,
    "timestep_mean": 547.21875,
    "epsilon_norm": 7.418735027313232,
    "prediction_norm": 6.356805801391602,
    "timestamp": "2025-07-14T12:34:10.661422"
  },
  {
    "epoch": 1041,
    "batch_idx": 192,
    "loss": 1.1785061359405518,
    "learning_rate": 0.00046782093093050036,
    "timestep_mean": 394.25,
    "epsilon_norm": 4.417919158935547,
    "prediction_norm": 2.8301491737365723,
    "timestamp": "2025-07-14T12:34:10.666215"
  },
  {
    "epoch": 1042,
    "batch_idx": 0,
    "loss": 0.34248679876327515,
    "learning_rate": 0.00046703720102431125,
    "timestep_mean": 476.15625,
    "epsilon_norm": 8.132159233093262,
    "prediction_norm": 7.056849002838135,
    "timestamp": "2025-07-14T12:34:10.671050"
  },
  {
    "epoch": 1042,
    "batch_idx": 32,
    "loss": 0.24593405425548553,
    "learning_rate": 0.00046703720102431125,
    "timestep_mean": 523.03125,
    "epsilon_norm": 7.398386478424072,
    "prediction_norm": 6.5636982917785645,
    "timestamp": "2025-07-14T12:34:10.675933"
  },
  {
    "epoch": 1042,
    "batch_idx": 64,
    "loss": 0.1872921884059906,
    "learning_rate": 0.00046703720102431125,
    "timestep_mean": 534.71875,
    "epsilon_norm": 7.785243034362793,
    "prediction_norm": 7.107758522033691,
    "timestamp": "2025-07-14T12:34:10.680827"
  },
  {
    "epoch": 1042,
    "batch_idx": 96,
    "loss": 0.13745702803134918,
    "learning_rate": 0.00046703720102431125,
    "timestep_mean": 571.625,
    "epsilon_norm": 8.170662879943848,
    "prediction_norm": 7.66238260269165,
    "timestamp": "2025-07-14T12:34:10.685789"
  },
  {
    "epoch": 1042,
    "batch_idx": 128,
    "loss": 0.25784578919410706,
    "learning_rate": 0.00046703720102431125,
    "timestep_mean": 512.0,
    "epsilon_norm": 8.069525718688965,
    "prediction_norm": 7.294059753417969,
    "timestamp": "2025-07-14T12:34:10.690760"
  },
  {
    "epoch": 1042,
    "batch_idx": 160,
    "loss": 0.29947730898857117,
    "learning_rate": 0.00046703720102431125,
    "timestep_mean": 541.53125,
    "epsilon_norm": 7.447535037994385,
    "prediction_norm": 6.418285369873047,
    "timestamp": "2025-07-14T12:34:10.695992"
  },
  {
    "epoch": 1042,
    "batch_idx": 192,
    "loss": 0.05539827421307564,
    "learning_rate": 0.00046703720102431125,
    "timestep_mean": 589.375,
    "epsilon_norm": 4.261523723602295,
    "prediction_norm": 4.3664093017578125,
    "timestamp": "2025-07-14T12:34:10.701060"
  },
  {
    "epoch": 1043,
    "batch_idx": 0,
    "loss": 0.32200631499290466,
    "learning_rate": 0.00046625355245055196,
    "timestep_mean": 514.84375,
    "epsilon_norm": 7.215121269226074,
    "prediction_norm": 5.909908771514893,
    "timestamp": "2025-07-14T12:34:10.706300"
  },
  {
    "epoch": 1043,
    "batch_idx": 32,
    "loss": 0.5029313564300537,
    "learning_rate": 0.00046625355245055196,
    "timestep_mean": 397.375,
    "epsilon_norm": 9.001852035522461,
    "prediction_norm": 7.497855186462402,
    "timestamp": "2025-07-14T12:34:10.711431"
  },
  {
    "epoch": 1043,
    "batch_idx": 64,
    "loss": 0.5047547221183777,
    "learning_rate": 0.00046625355245055196,
    "timestep_mean": 515.15625,
    "epsilon_norm": 8.592899322509766,
    "prediction_norm": 6.896749973297119,
    "timestamp": "2025-07-14T12:34:10.716489"
  },
  {
    "epoch": 1043,
    "batch_idx": 96,
    "loss": 0.23591738939285278,
    "learning_rate": 0.00046625355245055196,
    "timestep_mean": 488.03125,
    "epsilon_norm": 7.544733047485352,
    "prediction_norm": 6.505821704864502,
    "timestamp": "2025-07-14T12:34:10.721557"
  },
  {
    "epoch": 1043,
    "batch_idx": 128,
    "loss": 0.4774245023727417,
    "learning_rate": 0.00046625355245055196,
    "timestep_mean": 526.75,
    "epsilon_norm": 9.842483520507812,
    "prediction_norm": 8.164039611816406,
    "timestamp": "2025-07-14T12:34:10.726622"
  },
  {
    "epoch": 1043,
    "batch_idx": 160,
    "loss": 0.3609582781791687,
    "learning_rate": 0.00046625355245055196,
    "timestep_mean": 526.40625,
    "epsilon_norm": 8.198443412780762,
    "prediction_norm": 7.133170127868652,
    "timestamp": "2025-07-14T12:34:10.731712"
  },
  {
    "epoch": 1043,
    "batch_idx": 192,
    "loss": 0.10492189228534698,
    "learning_rate": 0.00046625355245055196,
    "timestep_mean": 628.5,
    "epsilon_norm": 3.6848764419555664,
    "prediction_norm": 3.5782177448272705,
    "timestamp": "2025-07-14T12:34:10.736970"
  },
  {
    "epoch": 1044,
    "batch_idx": 0,
    "loss": 0.28629302978515625,
    "learning_rate": 0.00046546998714279734,
    "timestep_mean": 494.21875,
    "epsilon_norm": 8.687333106994629,
    "prediction_norm": 7.243326663970947,
    "timestamp": "2025-07-14T12:34:10.742555"
  },
  {
    "epoch": 1044,
    "batch_idx": 32,
    "loss": 0.3369036316871643,
    "learning_rate": 0.00046546998714279734,
    "timestep_mean": 435.84375,
    "epsilon_norm": 7.819943904876709,
    "prediction_norm": 5.762834072113037,
    "timestamp": "2025-07-14T12:34:10.748162"
  },
  {
    "epoch": 1044,
    "batch_idx": 64,
    "loss": 0.351001501083374,
    "learning_rate": 0.00046546998714279734,
    "timestep_mean": 481.46875,
    "epsilon_norm": 8.061244010925293,
    "prediction_norm": 6.695559978485107,
    "timestamp": "2025-07-14T12:34:10.754075"
  },
  {
    "epoch": 1044,
    "batch_idx": 96,
    "loss": 0.21281975507736206,
    "learning_rate": 0.00046546998714279734,
    "timestep_mean": 519.21875,
    "epsilon_norm": 6.856556415557861,
    "prediction_norm": 6.322159290313721,
    "timestamp": "2025-07-14T12:34:10.759654"
  },
  {
    "epoch": 1044,
    "batch_idx": 128,
    "loss": 0.24061110615730286,
    "learning_rate": 0.00046546998714279734,
    "timestep_mean": 508.625,
    "epsilon_norm": 8.418601036071777,
    "prediction_norm": 7.700185775756836,
    "timestamp": "2025-07-14T12:34:10.765211"
  },
  {
    "epoch": 1044,
    "batch_idx": 160,
    "loss": 0.2766832709312439,
    "learning_rate": 0.00046546998714279734,
    "timestep_mean": 530.4375,
    "epsilon_norm": 8.758310317993164,
    "prediction_norm": 7.2418718338012695,
    "timestamp": "2025-07-14T12:34:10.770521"
  },
  {
    "epoch": 1044,
    "batch_idx": 192,
    "loss": 0.24463722109794617,
    "learning_rate": 0.00046546998714279734,
    "timestep_mean": 290.5,
    "epsilon_norm": 3.818347930908203,
    "prediction_norm": 3.0314180850982666,
    "timestamp": "2025-07-14T12:34:10.776008"
  },
  {
    "epoch": 1045,
    "batch_idx": 0,
    "loss": 0.26457449793815613,
    "learning_rate": 0.0004646865070344168,
    "timestep_mean": 526.5625,
    "epsilon_norm": 7.106706142425537,
    "prediction_norm": 7.016477584838867,
    "timestamp": "2025-07-14T12:34:10.782080"
  },
  {
    "epoch": 1045,
    "batch_idx": 32,
    "loss": 0.40433111786842346,
    "learning_rate": 0.0004646865070344168,
    "timestep_mean": 508.625,
    "epsilon_norm": 8.201197624206543,
    "prediction_norm": 6.972102642059326,
    "timestamp": "2025-07-14T12:34:10.787828"
  },
  {
    "epoch": 1045,
    "batch_idx": 64,
    "loss": 0.1985567808151245,
    "learning_rate": 0.0004646865070344168,
    "timestep_mean": 510.125,
    "epsilon_norm": 8.886920928955078,
    "prediction_norm": 7.584763526916504,
    "timestamp": "2025-07-14T12:34:10.801880"
  },
  {
    "epoch": 1045,
    "batch_idx": 96,
    "loss": 0.31098538637161255,
    "learning_rate": 0.0004646865070344168,
    "timestep_mean": 464.90625,
    "epsilon_norm": 7.253594875335693,
    "prediction_norm": 5.592276573181152,
    "timestamp": "2025-07-14T12:34:10.807893"
  },
  {
    "epoch": 1045,
    "batch_idx": 128,
    "loss": 0.1557522863149643,
    "learning_rate": 0.0004646865070344168,
    "timestep_mean": 437.65625,
    "epsilon_norm": 6.116876125335693,
    "prediction_norm": 5.321874141693115,
    "timestamp": "2025-07-14T12:34:10.813269"
  },
  {
    "epoch": 1045,
    "batch_idx": 160,
    "loss": 0.1954585760831833,
    "learning_rate": 0.0004646865070344168,
    "timestep_mean": 544.09375,
    "epsilon_norm": 6.854852199554443,
    "prediction_norm": 6.52970552444458,
    "timestamp": "2025-07-14T12:34:10.822034"
  },
  {
    "epoch": 1045,
    "batch_idx": 192,
    "loss": 0.29724225401878357,
    "learning_rate": 0.0004646865070344168,
    "timestep_mean": 417.25,
    "epsilon_norm": 4.110811233520508,
    "prediction_norm": 3.2486355304718018,
    "timestamp": "2025-07-14T12:34:10.830252"
  },
  {
    "epoch": 1046,
    "batch_idx": 0,
    "loss": 0.34802982211112976,
    "learning_rate": 0.00046390311405856993,
    "timestep_mean": 614.5625,
    "epsilon_norm": 8.697373390197754,
    "prediction_norm": 7.396523475646973,
    "timestamp": "2025-07-14T12:34:10.835885"
  },
  {
    "epoch": 1046,
    "batch_idx": 32,
    "loss": 0.23822976648807526,
    "learning_rate": 0.00046390311405856993,
    "timestep_mean": 424.6875,
    "epsilon_norm": 6.371493339538574,
    "prediction_norm": 5.384692192077637,
    "timestamp": "2025-07-14T12:34:10.841635"
  },
  {
    "epoch": 1046,
    "batch_idx": 64,
    "loss": 0.26967692375183105,
    "learning_rate": 0.00046390311405856993,
    "timestep_mean": 537.65625,
    "epsilon_norm": 8.067221641540527,
    "prediction_norm": 6.8790364265441895,
    "timestamp": "2025-07-14T12:34:10.848009"
  },
  {
    "epoch": 1046,
    "batch_idx": 96,
    "loss": 0.2990703880786896,
    "learning_rate": 0.00046390311405856993,
    "timestep_mean": 499.375,
    "epsilon_norm": 8.099040985107422,
    "prediction_norm": 6.223953723907471,
    "timestamp": "2025-07-14T12:34:10.855702"
  },
  {
    "epoch": 1046,
    "batch_idx": 128,
    "loss": 0.38912785053253174,
    "learning_rate": 0.00046390311405856993,
    "timestep_mean": 493.8125,
    "epsilon_norm": 7.872477054595947,
    "prediction_norm": 6.121696472167969,
    "timestamp": "2025-07-14T12:34:10.861145"
  },
  {
    "epoch": 1046,
    "batch_idx": 160,
    "loss": 0.27772200107574463,
    "learning_rate": 0.00046390311405856993,
    "timestep_mean": 512.78125,
    "epsilon_norm": 7.403220176696777,
    "prediction_norm": 5.782516956329346,
    "timestamp": "2025-07-14T12:34:10.866548"
  },
  {
    "epoch": 1046,
    "batch_idx": 192,
    "loss": 0.32701271772384644,
    "learning_rate": 0.00046390311405856993,
    "timestep_mean": 375.125,
    "epsilon_norm": 3.705174446105957,
    "prediction_norm": 2.3728723526000977,
    "timestamp": "2025-07-14T12:34:10.871766"
  },
  {
    "epoch": 1047,
    "batch_idx": 0,
    "loss": 0.4620376229286194,
    "learning_rate": 0.0004631198101482008,
    "timestep_mean": 443.1875,
    "epsilon_norm": 9.023472785949707,
    "prediction_norm": 6.3799920082092285,
    "timestamp": "2025-07-14T12:34:10.879874"
  },
  {
    "epoch": 1047,
    "batch_idx": 32,
    "loss": 0.30413317680358887,
    "learning_rate": 0.0004631198101482008,
    "timestep_mean": 530.09375,
    "epsilon_norm": 7.84976053237915,
    "prediction_norm": 6.220687389373779,
    "timestamp": "2025-07-14T12:34:10.888167"
  },
  {
    "epoch": 1047,
    "batch_idx": 64,
    "loss": 0.2853744328022003,
    "learning_rate": 0.0004631198101482008,
    "timestep_mean": 429.625,
    "epsilon_norm": 8.260834693908691,
    "prediction_norm": 6.84507942199707,
    "timestamp": "2025-07-14T12:34:10.893540"
  },
  {
    "epoch": 1047,
    "batch_idx": 96,
    "loss": 0.39223259687423706,
    "learning_rate": 0.0004631198101482008,
    "timestep_mean": 390.6875,
    "epsilon_norm": 9.242356300354004,
    "prediction_norm": 6.78080940246582,
    "timestamp": "2025-07-14T12:34:10.898905"
  },
  {
    "epoch": 1047,
    "batch_idx": 128,
    "loss": 0.3399551212787628,
    "learning_rate": 0.0004631198101482008,
    "timestep_mean": 407.0,
    "epsilon_norm": 7.804338455200195,
    "prediction_norm": 6.04118013381958,
    "timestamp": "2025-07-14T12:34:10.903967"
  },
  {
    "epoch": 1047,
    "batch_idx": 160,
    "loss": 0.33689141273498535,
    "learning_rate": 0.0004631198101482008,
    "timestep_mean": 505.0625,
    "epsilon_norm": 7.9099507331848145,
    "prediction_norm": 6.450992584228516,
    "timestamp": "2025-07-14T12:34:10.909385"
  },
  {
    "epoch": 1047,
    "batch_idx": 192,
    "loss": 0.39564231038093567,
    "learning_rate": 0.0004631198101482008,
    "timestep_mean": 634.75,
    "epsilon_norm": 5.368505477905273,
    "prediction_norm": 4.27020788192749,
    "timestamp": "2025-07-14T12:34:10.915580"
  },
  {
    "epoch": 1048,
    "batch_idx": 0,
    "loss": 0.24608829617500305,
    "learning_rate": 0.0004623365972360338,
    "timestep_mean": 547.125,
    "epsilon_norm": 7.5631585121154785,
    "prediction_norm": 6.547680377960205,
    "timestamp": "2025-07-14T12:34:10.922674"
  },
  {
    "epoch": 1048,
    "batch_idx": 32,
    "loss": 0.1749894767999649,
    "learning_rate": 0.0004623365972360338,
    "timestep_mean": 543.46875,
    "epsilon_norm": 7.391040802001953,
    "prediction_norm": 6.852245807647705,
    "timestamp": "2025-07-14T12:34:10.930521"
  },
  {
    "epoch": 1048,
    "batch_idx": 64,
    "loss": 0.3445817232131958,
    "learning_rate": 0.0004623365972360338,
    "timestep_mean": 514.09375,
    "epsilon_norm": 7.926149368286133,
    "prediction_norm": 6.124652862548828,
    "timestamp": "2025-07-14T12:34:10.936670"
  },
  {
    "epoch": 1048,
    "batch_idx": 96,
    "loss": 0.18632090091705322,
    "learning_rate": 0.0004623365972360338,
    "timestep_mean": 557.28125,
    "epsilon_norm": 8.143756866455078,
    "prediction_norm": 7.47832727432251,
    "timestamp": "2025-07-14T12:34:10.942473"
  },
  {
    "epoch": 1048,
    "batch_idx": 128,
    "loss": 0.13649050891399384,
    "learning_rate": 0.0004623365972360338,
    "timestep_mean": 570.75,
    "epsilon_norm": 7.333536148071289,
    "prediction_norm": 6.7577104568481445,
    "timestamp": "2025-07-14T12:34:10.949382"
  },
  {
    "epoch": 1048,
    "batch_idx": 160,
    "loss": 0.1426965892314911,
    "learning_rate": 0.0004623365972360338,
    "timestep_mean": 546.1875,
    "epsilon_norm": 7.643362522125244,
    "prediction_norm": 7.143954277038574,
    "timestamp": "2025-07-14T12:34:10.956623"
  },
  {
    "epoch": 1048,
    "batch_idx": 192,
    "loss": 0.1661725789308548,
    "learning_rate": 0.0004623365972360338,
    "timestep_mean": 589.125,
    "epsilon_norm": 3.5069234371185303,
    "prediction_norm": 3.0977494716644287,
    "timestamp": "2025-07-14T12:34:10.965621"
  },
  {
    "epoch": 1049,
    "batch_idx": 0,
    "loss": 0.14777663350105286,
    "learning_rate": 0.0004615534772545692,
    "timestep_mean": 494.75,
    "epsilon_norm": 7.149258613586426,
    "prediction_norm": 6.927305221557617,
    "timestamp": "2025-07-14T12:34:10.971092"
  },
  {
    "epoch": 1049,
    "batch_idx": 32,
    "loss": 0.30988264083862305,
    "learning_rate": 0.0004615534772545692,
    "timestep_mean": 400.0,
    "epsilon_norm": 7.790422439575195,
    "prediction_norm": 6.6062331199646,
    "timestamp": "2025-07-14T12:34:10.976360"
  },
  {
    "epoch": 1049,
    "batch_idx": 64,
    "loss": 0.37115901708602905,
    "learning_rate": 0.0004615534772545692,
    "timestep_mean": 434.875,
    "epsilon_norm": 9.214960098266602,
    "prediction_norm": 7.936040878295898,
    "timestamp": "2025-07-14T12:34:10.981634"
  },
  {
    "epoch": 1049,
    "batch_idx": 96,
    "loss": 0.13206817209720612,
    "learning_rate": 0.0004615534772545692,
    "timestep_mean": 556.3125,
    "epsilon_norm": 7.548745632171631,
    "prediction_norm": 6.962252616882324,
    "timestamp": "2025-07-14T12:34:10.986824"
  },
  {
    "epoch": 1049,
    "batch_idx": 128,
    "loss": 0.5895965099334717,
    "learning_rate": 0.0004615534772545692,
    "timestep_mean": 499.75,
    "epsilon_norm": 8.895584106445312,
    "prediction_norm": 6.797721862792969,
    "timestamp": "2025-07-14T12:34:10.992006"
  },
  {
    "epoch": 1049,
    "batch_idx": 160,
    "loss": 0.36421966552734375,
    "learning_rate": 0.0004615534772545692,
    "timestep_mean": 553.6875,
    "epsilon_norm": 8.817215919494629,
    "prediction_norm": 7.702432632446289,
    "timestamp": "2025-07-14T12:34:10.997167"
  },
  {
    "epoch": 1049,
    "batch_idx": 192,
    "loss": 0.3368886113166809,
    "learning_rate": 0.0004615534772545692,
    "timestep_mean": 291.875,
    "epsilon_norm": 3.3062150478363037,
    "prediction_norm": 2.929250717163086,
    "timestamp": "2025-07-14T12:34:11.002469"
  },
  {
    "epoch": 1050,
    "batch_idx": 0,
    "loss": 0.1848512589931488,
    "learning_rate": 0.0004607704521360777,
    "timestep_mean": 567.09375,
    "epsilon_norm": 8.284475326538086,
    "prediction_norm": 7.599966526031494,
    "timestamp": "2025-07-14T12:34:11.007744"
  },
  {
    "epoch": 1050,
    "batch_idx": 32,
    "loss": 0.2504149377346039,
    "learning_rate": 0.0004607704521360777,
    "timestep_mean": 481.78125,
    "epsilon_norm": 7.295321941375732,
    "prediction_norm": 6.6077728271484375,
    "timestamp": "2025-07-14T12:34:11.013553"
  },
  {
    "epoch": 1050,
    "batch_idx": 64,
    "loss": 0.34990474581718445,
    "learning_rate": 0.0004607704521360777,
    "timestep_mean": 561.03125,
    "epsilon_norm": 8.010810852050781,
    "prediction_norm": 6.413452625274658,
    "timestamp": "2025-07-14T12:34:11.018963"
  },
  {
    "epoch": 1050,
    "batch_idx": 96,
    "loss": 0.31259241700172424,
    "learning_rate": 0.0004607704521360777,
    "timestep_mean": 491.46875,
    "epsilon_norm": 7.4291911125183105,
    "prediction_norm": 6.717952251434326,
    "timestamp": "2025-07-14T12:34:11.024331"
  },
  {
    "epoch": 1050,
    "batch_idx": 128,
    "loss": 0.361653208732605,
    "learning_rate": 0.0004607704521360777,
    "timestep_mean": 490.84375,
    "epsilon_norm": 8.64865493774414,
    "prediction_norm": 7.119830131530762,
    "timestamp": "2025-07-14T12:34:11.029628"
  },
  {
    "epoch": 1050,
    "batch_idx": 160,
    "loss": 0.16488304734230042,
    "learning_rate": 0.0004607704521360777,
    "timestep_mean": 561.75,
    "epsilon_norm": 7.169870376586914,
    "prediction_norm": 6.118069648742676,
    "timestamp": "2025-07-14T12:34:11.034811"
  },
  {
    "epoch": 1050,
    "batch_idx": 192,
    "loss": 0.14435359835624695,
    "learning_rate": 0.0004607704521360777,
    "timestep_mean": 619.75,
    "epsilon_norm": 3.6067535877227783,
    "prediction_norm": 3.440094232559204,
    "timestamp": "2025-07-14T12:34:11.040154"
  },
  {
    "epoch": 1051,
    "batch_idx": 0,
    "loss": 0.3416212797164917,
    "learning_rate": 0.00045998752381259586,
    "timestep_mean": 480.3125,
    "epsilon_norm": 7.976532936096191,
    "prediction_norm": 6.323265075683594,
    "timestamp": "2025-07-14T12:34:11.045324"
  },
  {
    "epoch": 1051,
    "batch_idx": 32,
    "loss": 0.19024883210659027,
    "learning_rate": 0.00045998752381259586,
    "timestep_mean": 504.53125,
    "epsilon_norm": 7.081006050109863,
    "prediction_norm": 6.310678482055664,
    "timestamp": "2025-07-14T12:34:11.050540"
  },
  {
    "epoch": 1051,
    "batch_idx": 64,
    "loss": 0.3419814109802246,
    "learning_rate": 0.00045998752381259586,
    "timestep_mean": 535.625,
    "epsilon_norm": 7.955381393432617,
    "prediction_norm": 6.716825485229492,
    "timestamp": "2025-07-14T12:34:11.055634"
  },
  {
    "epoch": 1051,
    "batch_idx": 96,
    "loss": 0.3743443489074707,
    "learning_rate": 0.00045998752381259586,
    "timestep_mean": 525.625,
    "epsilon_norm": 7.3601789474487305,
    "prediction_norm": 5.794383525848389,
    "timestamp": "2025-07-14T12:34:11.060673"
  },
  {
    "epoch": 1051,
    "batch_idx": 128,
    "loss": 0.435879111289978,
    "learning_rate": 0.00045998752381259586,
    "timestep_mean": 455.03125,
    "epsilon_norm": 8.979207038879395,
    "prediction_norm": 7.251453399658203,
    "timestamp": "2025-07-14T12:34:11.065880"
  },
  {
    "epoch": 1051,
    "batch_idx": 160,
    "loss": 0.35695549845695496,
    "learning_rate": 0.00045998752381259586,
    "timestep_mean": 446.5625,
    "epsilon_norm": 7.5901312828063965,
    "prediction_norm": 5.794309139251709,
    "timestamp": "2025-07-14T12:34:11.070963"
  },
  {
    "epoch": 1051,
    "batch_idx": 192,
    "loss": 0.2960279583930969,
    "learning_rate": 0.00045998752381259586,
    "timestep_mean": 530.375,
    "epsilon_norm": 2.9670794010162354,
    "prediction_norm": 2.2050981521606445,
    "timestamp": "2025-07-14T12:34:11.076085"
  },
  {
    "epoch": 1052,
    "batch_idx": 0,
    "loss": 0.2763943374156952,
    "learning_rate": 0.0004592046942159214,
    "timestep_mean": 511.65625,
    "epsilon_norm": 8.16166877746582,
    "prediction_norm": 6.766264915466309,
    "timestamp": "2025-07-14T12:34:11.081162"
  },
  {
    "epoch": 1052,
    "batch_idx": 32,
    "loss": 0.6785826683044434,
    "learning_rate": 0.0004592046942159214,
    "timestep_mean": 470.71875,
    "epsilon_norm": 9.007662773132324,
    "prediction_norm": 6.675688743591309,
    "timestamp": "2025-07-14T12:34:11.086136"
  },
  {
    "epoch": 1052,
    "batch_idx": 64,
    "loss": 0.3985583186149597,
    "learning_rate": 0.0004592046942159214,
    "timestep_mean": 419.3125,
    "epsilon_norm": 7.877401828765869,
    "prediction_norm": 5.948892116546631,
    "timestamp": "2025-07-14T12:34:11.091159"
  },
  {
    "epoch": 1052,
    "batch_idx": 96,
    "loss": 0.3519374132156372,
    "learning_rate": 0.0004592046942159214,
    "timestep_mean": 560.53125,
    "epsilon_norm": 9.43333911895752,
    "prediction_norm": 7.691292762756348,
    "timestamp": "2025-07-14T12:34:11.096198"
  },
  {
    "epoch": 1052,
    "batch_idx": 128,
    "loss": 0.2046794891357422,
    "learning_rate": 0.0004592046942159214,
    "timestep_mean": 515.28125,
    "epsilon_norm": 7.138452529907227,
    "prediction_norm": 6.235728740692139,
    "timestamp": "2025-07-14T12:34:11.101432"
  },
  {
    "epoch": 1052,
    "batch_idx": 160,
    "loss": 0.28071722388267517,
    "learning_rate": 0.0004592046942159214,
    "timestep_mean": 488.5,
    "epsilon_norm": 6.809678077697754,
    "prediction_norm": 5.907108783721924,
    "timestamp": "2025-07-14T12:34:11.106670"
  },
  {
    "epoch": 1052,
    "batch_idx": 192,
    "loss": 0.5181404948234558,
    "learning_rate": 0.0004592046942159214,
    "timestep_mean": 549.5,
    "epsilon_norm": 3.4351234436035156,
    "prediction_norm": 2.6287670135498047,
    "timestamp": "2025-07-14T12:34:11.111844"
  },
  {
    "epoch": 1053,
    "batch_idx": 0,
    "loss": 0.18713857233524323,
    "learning_rate": 0.00045842196527760864,
    "timestep_mean": 482.5625,
    "epsilon_norm": 7.389591217041016,
    "prediction_norm": 6.561622142791748,
    "timestamp": "2025-07-14T12:34:11.117260"
  },
  {
    "epoch": 1053,
    "batch_idx": 32,
    "loss": 0.21513302624225616,
    "learning_rate": 0.00045842196527760864,
    "timestep_mean": 545.65625,
    "epsilon_norm": 7.2766265869140625,
    "prediction_norm": 6.034167766571045,
    "timestamp": "2025-07-14T12:34:11.122607"
  },
  {
    "epoch": 1053,
    "batch_idx": 64,
    "loss": 0.11747787147760391,
    "learning_rate": 0.00045842196527760864,
    "timestep_mean": 528.21875,
    "epsilon_norm": 9.236394882202148,
    "prediction_norm": 8.448579788208008,
    "timestamp": "2025-07-14T12:34:11.127826"
  },
  {
    "epoch": 1053,
    "batch_idx": 96,
    "loss": 0.4412481486797333,
    "learning_rate": 0.00045842196527760864,
    "timestep_mean": 426.75,
    "epsilon_norm": 8.803133010864258,
    "prediction_norm": 6.786768436431885,
    "timestamp": "2025-07-14T12:34:11.133098"
  },
  {
    "epoch": 1053,
    "batch_idx": 128,
    "loss": 0.4030721187591553,
    "learning_rate": 0.00045842196527760864,
    "timestep_mean": 427.34375,
    "epsilon_norm": 8.626608848571777,
    "prediction_norm": 7.426890850067139,
    "timestamp": "2025-07-14T12:34:11.138394"
  },
  {
    "epoch": 1053,
    "batch_idx": 160,
    "loss": 0.23667559027671814,
    "learning_rate": 0.00045842196527760864,
    "timestep_mean": 549.125,
    "epsilon_norm": 8.258062362670898,
    "prediction_norm": 7.462240695953369,
    "timestamp": "2025-07-14T12:34:11.143623"
  },
  {
    "epoch": 1053,
    "batch_idx": 192,
    "loss": 0.1208389550447464,
    "learning_rate": 0.00045842196527760864,
    "timestep_mean": 584.375,
    "epsilon_norm": 2.979076623916626,
    "prediction_norm": 2.639058828353882,
    "timestamp": "2025-07-14T12:34:11.148904"
  },
  {
    "epoch": 1054,
    "batch_idx": 0,
    "loss": 0.3560944199562073,
    "learning_rate": 0.00045763933892896343,
    "timestep_mean": 476.15625,
    "epsilon_norm": 7.1696319580078125,
    "prediction_norm": 6.04597806930542,
    "timestamp": "2025-07-14T12:34:11.154365"
  },
  {
    "epoch": 1054,
    "batch_idx": 32,
    "loss": 0.3924802541732788,
    "learning_rate": 0.00045763933892896343,
    "timestep_mean": 477.8125,
    "epsilon_norm": 8.998679161071777,
    "prediction_norm": 7.109054088592529,
    "timestamp": "2025-07-14T12:34:11.159553"
  },
  {
    "epoch": 1054,
    "batch_idx": 64,
    "loss": 0.20486080646514893,
    "learning_rate": 0.00045763933892896343,
    "timestep_mean": 468.65625,
    "epsilon_norm": 7.725351810455322,
    "prediction_norm": 6.669633865356445,
    "timestamp": "2025-07-14T12:34:11.164800"
  },
  {
    "epoch": 1054,
    "batch_idx": 96,
    "loss": 0.3388027548789978,
    "learning_rate": 0.00045763933892896343,
    "timestep_mean": 571.5,
    "epsilon_norm": 7.779606819152832,
    "prediction_norm": 6.085766315460205,
    "timestamp": "2025-07-14T12:34:11.170028"
  },
  {
    "epoch": 1054,
    "batch_idx": 128,
    "loss": 0.27825692296028137,
    "learning_rate": 0.00045763933892896343,
    "timestep_mean": 482.75,
    "epsilon_norm": 6.922011375427246,
    "prediction_norm": 5.896580219268799,
    "timestamp": "2025-07-14T12:34:11.175260"
  },
  {
    "epoch": 1054,
    "batch_idx": 160,
    "loss": 0.1842276155948639,
    "learning_rate": 0.00045763933892896343,
    "timestep_mean": 593.0625,
    "epsilon_norm": 5.935830593109131,
    "prediction_norm": 5.670078754425049,
    "timestamp": "2025-07-14T12:34:11.180458"
  },
  {
    "epoch": 1054,
    "batch_idx": 192,
    "loss": 0.1899421364068985,
    "learning_rate": 0.00045763933892896343,
    "timestep_mean": 474.5,
    "epsilon_norm": 4.585820198059082,
    "prediction_norm": 3.9498467445373535,
    "timestamp": "2025-07-14T12:34:11.185592"
  },
  {
    "epoch": 1055,
    "batch_idx": 0,
    "loss": 0.30520135164260864,
    "learning_rate": 0.00045685681710103853,
    "timestep_mean": 545.875,
    "epsilon_norm": 8.6860990524292,
    "prediction_norm": 6.873100280761719,
    "timestamp": "2025-07-14T12:34:11.190863"
  },
  {
    "epoch": 1055,
    "batch_idx": 32,
    "loss": 0.293722927570343,
    "learning_rate": 0.00045685681710103853,
    "timestep_mean": 591.59375,
    "epsilon_norm": 7.784968852996826,
    "prediction_norm": 6.6094207763671875,
    "timestamp": "2025-07-14T12:34:11.196230"
  },
  {
    "epoch": 1055,
    "batch_idx": 64,
    "loss": 0.18615517020225525,
    "learning_rate": 0.00045685681710103853,
    "timestep_mean": 478.4375,
    "epsilon_norm": 8.091532707214355,
    "prediction_norm": 6.764536380767822,
    "timestamp": "2025-07-14T12:34:11.201534"
  },
  {
    "epoch": 1055,
    "batch_idx": 96,
    "loss": 0.3405325412750244,
    "learning_rate": 0.00045685681710103853,
    "timestep_mean": 520.1875,
    "epsilon_norm": 9.37714672088623,
    "prediction_norm": 8.188591957092285,
    "timestamp": "2025-07-14T12:34:11.206754"
  },
  {
    "epoch": 1055,
    "batch_idx": 128,
    "loss": 0.24937036633491516,
    "learning_rate": 0.00045685681710103853,
    "timestep_mean": 539.0625,
    "epsilon_norm": 7.364676475524902,
    "prediction_norm": 6.728726863861084,
    "timestamp": "2025-07-14T12:34:11.211900"
  },
  {
    "epoch": 1055,
    "batch_idx": 160,
    "loss": 0.4538799524307251,
    "learning_rate": 0.00045685681710103853,
    "timestep_mean": 475.15625,
    "epsilon_norm": 8.5432710647583,
    "prediction_norm": 6.513296127319336,
    "timestamp": "2025-07-14T12:34:11.217101"
  },
  {
    "epoch": 1055,
    "batch_idx": 192,
    "loss": 0.33484137058258057,
    "learning_rate": 0.00045685681710103853,
    "timestep_mean": 296.5,
    "epsilon_norm": 3.5288918018341064,
    "prediction_norm": 2.389261245727539,
    "timestamp": "2025-07-14T12:34:11.222190"
  },
  {
    "epoch": 1056,
    "batch_idx": 0,
    "loss": 0.2725679278373718,
    "learning_rate": 0.0004560744017246286,
    "timestep_mean": 506.65625,
    "epsilon_norm": 8.606595993041992,
    "prediction_norm": 7.634373188018799,
    "timestamp": "2025-07-14T12:34:11.227474"
  },
  {
    "epoch": 1056,
    "batch_idx": 32,
    "loss": 0.40131473541259766,
    "learning_rate": 0.0004560744017246286,
    "timestep_mean": 357.90625,
    "epsilon_norm": 7.655168056488037,
    "prediction_norm": 6.3124542236328125,
    "timestamp": "2025-07-14T12:34:11.232713"
  },
  {
    "epoch": 1056,
    "batch_idx": 64,
    "loss": 0.24404014647006989,
    "learning_rate": 0.0004560744017246286,
    "timestep_mean": 565.125,
    "epsilon_norm": 9.330309867858887,
    "prediction_norm": 8.131892204284668,
    "timestamp": "2025-07-14T12:34:11.238198"
  },
  {
    "epoch": 1056,
    "batch_idx": 96,
    "loss": 0.22587642073631287,
    "learning_rate": 0.0004560744017246286,
    "timestep_mean": 496.125,
    "epsilon_norm": 8.311100006103516,
    "prediction_norm": 6.998843669891357,
    "timestamp": "2025-07-14T12:34:11.243487"
  },
  {
    "epoch": 1056,
    "batch_idx": 128,
    "loss": 0.11596590280532837,
    "learning_rate": 0.0004560744017246286,
    "timestep_mean": 563.625,
    "epsilon_norm": 7.5829949378967285,
    "prediction_norm": 6.977413177490234,
    "timestamp": "2025-07-14T12:34:11.248739"
  },
  {
    "epoch": 1056,
    "batch_idx": 160,
    "loss": 0.1891162097454071,
    "learning_rate": 0.0004560744017246286,
    "timestep_mean": 613.78125,
    "epsilon_norm": 7.383746147155762,
    "prediction_norm": 6.287872791290283,
    "timestamp": "2025-07-14T12:34:11.254145"
  },
  {
    "epoch": 1056,
    "batch_idx": 192,
    "loss": 0.31724226474761963,
    "learning_rate": 0.0004560744017246286,
    "timestep_mean": 439.5,
    "epsilon_norm": 3.6124250888824463,
    "prediction_norm": 3.1693296432495117,
    "timestamp": "2025-07-14T12:34:11.259317"
  },
  {
    "epoch": 1057,
    "batch_idx": 0,
    "loss": 0.41383442282676697,
    "learning_rate": 0.00045529209473026594,
    "timestep_mean": 356.96875,
    "epsilon_norm": 8.710537910461426,
    "prediction_norm": 6.553122520446777,
    "timestamp": "2025-07-14T12:34:11.264502"
  },
  {
    "epoch": 1057,
    "batch_idx": 32,
    "loss": 0.3312452435493469,
    "learning_rate": 0.00045529209473026594,
    "timestep_mean": 466.625,
    "epsilon_norm": 7.795215606689453,
    "prediction_norm": 5.906805515289307,
    "timestamp": "2025-07-14T12:34:11.269798"
  },
  {
    "epoch": 1057,
    "batch_idx": 64,
    "loss": 0.20843708515167236,
    "learning_rate": 0.00045529209473026594,
    "timestep_mean": 543.3125,
    "epsilon_norm": 7.231258869171143,
    "prediction_norm": 6.2511491775512695,
    "timestamp": "2025-07-14T12:34:11.275085"
  },
  {
    "epoch": 1057,
    "batch_idx": 96,
    "loss": 0.19124062359333038,
    "learning_rate": 0.00045529209473026594,
    "timestep_mean": 490.65625,
    "epsilon_norm": 7.7560014724731445,
    "prediction_norm": 6.679684638977051,
    "timestamp": "2025-07-14T12:34:11.280366"
  },
  {
    "epoch": 1057,
    "batch_idx": 128,
    "loss": 0.21742214262485504,
    "learning_rate": 0.00045529209473026594,
    "timestep_mean": 540.625,
    "epsilon_norm": 7.160921573638916,
    "prediction_norm": 6.443963050842285,
    "timestamp": "2025-07-14T12:34:11.285689"
  },
  {
    "epoch": 1057,
    "batch_idx": 160,
    "loss": 0.15904182195663452,
    "learning_rate": 0.00045529209473026594,
    "timestep_mean": 529.0,
    "epsilon_norm": 8.223304748535156,
    "prediction_norm": 7.360630512237549,
    "timestamp": "2025-07-14T12:34:11.290937"
  },
  {
    "epoch": 1057,
    "batch_idx": 192,
    "loss": 0.10738344490528107,
    "learning_rate": 0.00045529209473026594,
    "timestep_mean": 613.625,
    "epsilon_norm": 3.735912799835205,
    "prediction_norm": 3.273918628692627,
    "timestamp": "2025-07-14T12:34:11.296186"
  },
  {
    "epoch": 1058,
    "batch_idx": 0,
    "loss": 0.3190475404262543,
    "learning_rate": 0.0004545098980482153,
    "timestep_mean": 489.65625,
    "epsilon_norm": 8.817014694213867,
    "prediction_norm": 7.672043323516846,
    "timestamp": "2025-07-14T12:34:11.301332"
  },
  {
    "epoch": 1058,
    "batch_idx": 32,
    "loss": 0.3418121337890625,
    "learning_rate": 0.0004545098980482153,
    "timestep_mean": 475.65625,
    "epsilon_norm": 8.697888374328613,
    "prediction_norm": 7.625519752502441,
    "timestamp": "2025-07-14T12:34:11.306497"
  },
  {
    "epoch": 1058,
    "batch_idx": 64,
    "loss": 0.3106401860713959,
    "learning_rate": 0.0004545098980482153,
    "timestep_mean": 525.40625,
    "epsilon_norm": 8.003244400024414,
    "prediction_norm": 6.167271137237549,
    "timestamp": "2025-07-14T12:34:11.311741"
  },
  {
    "epoch": 1058,
    "batch_idx": 96,
    "loss": 0.2783234119415283,
    "learning_rate": 0.0004545098980482153,
    "timestep_mean": 567.25,
    "epsilon_norm": 8.369998931884766,
    "prediction_norm": 7.306336402893066,
    "timestamp": "2025-07-14T12:34:11.316883"
  },
  {
    "epoch": 1058,
    "batch_idx": 128,
    "loss": 0.11878323554992676,
    "learning_rate": 0.0004545098980482153,
    "timestep_mean": 561.40625,
    "epsilon_norm": 7.991013526916504,
    "prediction_norm": 6.977239608764648,
    "timestamp": "2025-07-14T12:34:11.321878"
  },
  {
    "epoch": 1058,
    "batch_idx": 160,
    "loss": 0.20921477675437927,
    "learning_rate": 0.0004545098980482153,
    "timestep_mean": 415.0625,
    "epsilon_norm": 7.317505359649658,
    "prediction_norm": 6.486541271209717,
    "timestamp": "2025-07-14T12:34:11.326941"
  },
  {
    "epoch": 1058,
    "batch_idx": 192,
    "loss": 0.22356697916984558,
    "learning_rate": 0.0004545098980482153,
    "timestep_mean": 524.0,
    "epsilon_norm": 3.8427345752716064,
    "prediction_norm": 3.4603829383850098,
    "timestamp": "2025-07-14T12:34:11.331940"
  },
  {
    "epoch": 1059,
    "batch_idx": 0,
    "loss": 0.39913758635520935,
    "learning_rate": 0.0004537278136084691,
    "timestep_mean": 454.53125,
    "epsilon_norm": 8.270318984985352,
    "prediction_norm": 7.213993072509766,
    "timestamp": "2025-07-14T12:34:11.336980"
  },
  {
    "epoch": 1059,
    "batch_idx": 32,
    "loss": 0.3202556073665619,
    "learning_rate": 0.0004537278136084691,
    "timestep_mean": 458.5625,
    "epsilon_norm": 9.144821166992188,
    "prediction_norm": 7.4415388107299805,
    "timestamp": "2025-07-14T12:34:11.342231"
  },
  {
    "epoch": 1059,
    "batch_idx": 64,
    "loss": 0.2729513943195343,
    "learning_rate": 0.0004537278136084691,
    "timestep_mean": 425.15625,
    "epsilon_norm": 7.744725704193115,
    "prediction_norm": 6.424118995666504,
    "timestamp": "2025-07-14T12:34:11.347318"
  },
  {
    "epoch": 1059,
    "batch_idx": 96,
    "loss": 0.14629288017749786,
    "learning_rate": 0.0004537278136084691,
    "timestep_mean": 563.1875,
    "epsilon_norm": 7.81649112701416,
    "prediction_norm": 6.907139778137207,
    "timestamp": "2025-07-14T12:34:11.352425"
  },
  {
    "epoch": 1059,
    "batch_idx": 128,
    "loss": 0.36635851860046387,
    "learning_rate": 0.0004537278136084691,
    "timestep_mean": 515.0,
    "epsilon_norm": 7.755545616149902,
    "prediction_norm": 6.994575023651123,
    "timestamp": "2025-07-14T12:34:11.357609"
  },
  {
    "epoch": 1059,
    "batch_idx": 160,
    "loss": 0.26850205659866333,
    "learning_rate": 0.0004537278136084691,
    "timestep_mean": 461.34375,
    "epsilon_norm": 7.119095325469971,
    "prediction_norm": 6.279748439788818,
    "timestamp": "2025-07-14T12:34:11.362667"
  },
  {
    "epoch": 1059,
    "batch_idx": 192,
    "loss": 0.1271451860666275,
    "learning_rate": 0.0004537278136084691,
    "timestep_mean": 547.25,
    "epsilon_norm": 4.443727016448975,
    "prediction_norm": 3.98125958442688,
    "timestamp": "2025-07-14T12:34:11.367660"
  },
  {
    "epoch": 1060,
    "batch_idx": 0,
    "loss": 0.17111802101135254,
    "learning_rate": 0.000452945843340743,
    "timestep_mean": 450.75,
    "epsilon_norm": 5.606265068054199,
    "prediction_norm": 4.657720565795898,
    "timestamp": "2025-07-14T12:34:11.372821"
  },
  {
    "epoch": 1060,
    "batch_idx": 32,
    "loss": 0.2938520014286041,
    "learning_rate": 0.000452945843340743,
    "timestep_mean": 422.1875,
    "epsilon_norm": 7.7688374519348145,
    "prediction_norm": 6.941956996917725,
    "timestamp": "2025-07-14T12:34:11.377942"
  },
  {
    "epoch": 1060,
    "batch_idx": 64,
    "loss": 0.327907919883728,
    "learning_rate": 0.000452945843340743,
    "timestep_mean": 447.125,
    "epsilon_norm": 7.001784324645996,
    "prediction_norm": 5.961150169372559,
    "timestamp": "2025-07-14T12:34:11.383028"
  },
  {
    "epoch": 1060,
    "batch_idx": 96,
    "loss": 0.440289705991745,
    "learning_rate": 0.000452945843340743,
    "timestep_mean": 503.34375,
    "epsilon_norm": 8.583322525024414,
    "prediction_norm": 7.252654075622559,
    "timestamp": "2025-07-14T12:34:11.388172"
  },
  {
    "epoch": 1060,
    "batch_idx": 128,
    "loss": 0.1402587890625,
    "learning_rate": 0.000452945843340743,
    "timestep_mean": 574.96875,
    "epsilon_norm": 8.011487007141113,
    "prediction_norm": 7.459848880767822,
    "timestamp": "2025-07-14T12:34:11.393312"
  },
  {
    "epoch": 1060,
    "batch_idx": 160,
    "loss": 0.26270318031311035,
    "learning_rate": 0.000452945843340743,
    "timestep_mean": 496.5625,
    "epsilon_norm": 7.913666248321533,
    "prediction_norm": 7.45654821395874,
    "timestamp": "2025-07-14T12:34:11.398479"
  },
  {
    "epoch": 1060,
    "batch_idx": 192,
    "loss": 0.1467193067073822,
    "learning_rate": 0.000452945843340743,
    "timestep_mean": 580.375,
    "epsilon_norm": 2.51546049118042,
    "prediction_norm": 1.8910080194473267,
    "timestamp": "2025-07-14T12:34:11.403597"
  },
  {
    "epoch": 1061,
    "batch_idx": 0,
    "loss": 0.24416403472423553,
    "learning_rate": 0.00045216398917447107,
    "timestep_mean": 583.125,
    "epsilon_norm": 8.164057731628418,
    "prediction_norm": 6.662962436676025,
    "timestamp": "2025-07-14T12:34:11.408555"
  },
  {
    "epoch": 1061,
    "batch_idx": 32,
    "loss": 0.26149195432662964,
    "learning_rate": 0.00045216398917447107,
    "timestep_mean": 513.21875,
    "epsilon_norm": 8.930498123168945,
    "prediction_norm": 7.887190341949463,
    "timestamp": "2025-07-14T12:34:11.413434"
  },
  {
    "epoch": 1061,
    "batch_idx": 64,
    "loss": 0.14718247950077057,
    "learning_rate": 0.00045216398917447107,
    "timestep_mean": 499.75,
    "epsilon_norm": 6.8989996910095215,
    "prediction_norm": 6.186368465423584,
    "timestamp": "2025-07-14T12:34:11.419053"
  },
  {
    "epoch": 1061,
    "batch_idx": 96,
    "loss": 0.22058019042015076,
    "learning_rate": 0.00045216398917447107,
    "timestep_mean": 541.0,
    "epsilon_norm": 8.509129524230957,
    "prediction_norm": 7.642284870147705,
    "timestamp": "2025-07-14T12:34:11.424470"
  },
  {
    "epoch": 1061,
    "batch_idx": 128,
    "loss": 0.22116833925247192,
    "learning_rate": 0.00045216398917447107,
    "timestep_mean": 420.6875,
    "epsilon_norm": 6.795677661895752,
    "prediction_norm": 6.226828575134277,
    "timestamp": "2025-07-14T12:34:11.429782"
  },
  {
    "epoch": 1061,
    "batch_idx": 160,
    "loss": 0.40439894795417786,
    "learning_rate": 0.00045216398917447107,
    "timestep_mean": 485.03125,
    "epsilon_norm": 8.340575218200684,
    "prediction_norm": 7.206915378570557,
    "timestamp": "2025-07-14T12:34:11.435399"
  },
  {
    "epoch": 1061,
    "batch_idx": 192,
    "loss": 0.07563653588294983,
    "learning_rate": 0.00045216398917447107,
    "timestep_mean": 551.25,
    "epsilon_norm": 2.9461729526519775,
    "prediction_norm": 2.256243944168091,
    "timestamp": "2025-07-14T12:34:11.440948"
  },
  {
    "epoch": 1062,
    "batch_idx": 0,
    "loss": 0.2440578192472458,
    "learning_rate": 0.00045138225303880064,
    "timestep_mean": 424.53125,
    "epsilon_norm": 6.689766883850098,
    "prediction_norm": 5.518708229064941,
    "timestamp": "2025-07-14T12:34:11.446295"
  },
  {
    "epoch": 1062,
    "batch_idx": 32,
    "loss": 0.21220746636390686,
    "learning_rate": 0.00045138225303880064,
    "timestep_mean": 433.625,
    "epsilon_norm": 7.399900436401367,
    "prediction_norm": 6.4436469078063965,
    "timestamp": "2025-07-14T12:34:11.451422"
  },
  {
    "epoch": 1062,
    "batch_idx": 64,
    "loss": 0.24388666450977325,
    "learning_rate": 0.00045138225303880064,
    "timestep_mean": 485.8125,
    "epsilon_norm": 7.367941856384277,
    "prediction_norm": 6.300729751586914,
    "timestamp": "2025-07-14T12:34:11.456560"
  },
  {
    "epoch": 1062,
    "batch_idx": 96,
    "loss": 0.2509598731994629,
    "learning_rate": 0.00045138225303880064,
    "timestep_mean": 494.25,
    "epsilon_norm": 8.074057579040527,
    "prediction_norm": 6.7514142990112305,
    "timestamp": "2025-07-14T12:34:11.461834"
  },
  {
    "epoch": 1062,
    "batch_idx": 128,
    "loss": 0.12904810905456543,
    "learning_rate": 0.00045138225303880064,
    "timestep_mean": 614.84375,
    "epsilon_norm": 7.297913074493408,
    "prediction_norm": 6.989306926727295,
    "timestamp": "2025-07-14T12:34:11.467093"
  },
  {
    "epoch": 1062,
    "batch_idx": 160,
    "loss": 0.2685456871986389,
    "learning_rate": 0.00045138225303880064,
    "timestep_mean": 479.15625,
    "epsilon_norm": 7.520612716674805,
    "prediction_norm": 6.4870381355285645,
    "timestamp": "2025-07-14T12:34:11.472221"
  },
  {
    "epoch": 1062,
    "batch_idx": 192,
    "loss": 0.2734002470970154,
    "learning_rate": 0.00045138225303880064,
    "timestep_mean": 460.5,
    "epsilon_norm": 3.4298675060272217,
    "prediction_norm": 2.878044605255127,
    "timestamp": "2025-07-14T12:34:11.477462"
  },
  {
    "epoch": 1063,
    "batch_idx": 0,
    "loss": 0.364460289478302,
    "learning_rate": 0.00045060063686258773,
    "timestep_mean": 450.71875,
    "epsilon_norm": 7.707871437072754,
    "prediction_norm": 6.4374918937683105,
    "timestamp": "2025-07-14T12:34:11.482779"
  },
  {
    "epoch": 1063,
    "batch_idx": 32,
    "loss": 0.2743872106075287,
    "learning_rate": 0.00045060063686258773,
    "timestep_mean": 444.625,
    "epsilon_norm": 8.185185432434082,
    "prediction_norm": 6.598750591278076,
    "timestamp": "2025-07-14T12:34:11.488029"
  },
  {
    "epoch": 1063,
    "batch_idx": 64,
    "loss": 0.19450195133686066,
    "learning_rate": 0.00045060063686258773,
    "timestep_mean": 499.15625,
    "epsilon_norm": 7.707194805145264,
    "prediction_norm": 7.112475395202637,
    "timestamp": "2025-07-14T12:34:11.493290"
  },
  {
    "epoch": 1063,
    "batch_idx": 96,
    "loss": 0.301327109336853,
    "learning_rate": 0.00045060063686258773,
    "timestep_mean": 436.34375,
    "epsilon_norm": 9.47514533996582,
    "prediction_norm": 7.717503070831299,
    "timestamp": "2025-07-14T12:34:11.498422"
  },
  {
    "epoch": 1063,
    "batch_idx": 128,
    "loss": 0.27456048130989075,
    "learning_rate": 0.00045060063686258773,
    "timestep_mean": 423.46875,
    "epsilon_norm": 6.696253299713135,
    "prediction_norm": 5.403693199157715,
    "timestamp": "2025-07-14T12:34:11.503575"
  },
  {
    "epoch": 1063,
    "batch_idx": 160,
    "loss": 0.280303955078125,
    "learning_rate": 0.00045060063686258773,
    "timestep_mean": 473.125,
    "epsilon_norm": 8.00562858581543,
    "prediction_norm": 6.122283935546875,
    "timestamp": "2025-07-14T12:34:11.508816"
  },
  {
    "epoch": 1063,
    "batch_idx": 192,
    "loss": 0.20791062712669373,
    "learning_rate": 0.00045060063686258773,
    "timestep_mean": 527.375,
    "epsilon_norm": 3.13500714302063,
    "prediction_norm": 2.414659023284912,
    "timestamp": "2025-07-14T12:34:11.513998"
  },
  {
    "epoch": 1064,
    "batch_idx": 0,
    "loss": 0.271543025970459,
    "learning_rate": 0.0004498191425743928,
    "timestep_mean": 462.4375,
    "epsilon_norm": 6.549365997314453,
    "prediction_norm": 5.2874064445495605,
    "timestamp": "2025-07-14T12:34:11.519254"
  },
  {
    "epoch": 1064,
    "batch_idx": 32,
    "loss": 0.22391581535339355,
    "learning_rate": 0.0004498191425743928,
    "timestep_mean": 547.4375,
    "epsilon_norm": 7.325736999511719,
    "prediction_norm": 6.293299674987793,
    "timestamp": "2025-07-14T12:34:11.524532"
  },
  {
    "epoch": 1064,
    "batch_idx": 64,
    "loss": 0.242942214012146,
    "learning_rate": 0.0004498191425743928,
    "timestep_mean": 482.6875,
    "epsilon_norm": 7.5629658699035645,
    "prediction_norm": 6.260890960693359,
    "timestamp": "2025-07-14T12:34:11.529714"
  },
  {
    "epoch": 1064,
    "batch_idx": 96,
    "loss": 0.2650523781776428,
    "learning_rate": 0.0004498191425743928,
    "timestep_mean": 507.6875,
    "epsilon_norm": 8.070149421691895,
    "prediction_norm": 7.356270790100098,
    "timestamp": "2025-07-14T12:34:11.534941"
  },
  {
    "epoch": 1064,
    "batch_idx": 128,
    "loss": 0.34409573674201965,
    "learning_rate": 0.0004498191425743928,
    "timestep_mean": 472.25,
    "epsilon_norm": 8.884610176086426,
    "prediction_norm": 8.0473051071167,
    "timestamp": "2025-07-14T12:34:11.540172"
  },
  {
    "epoch": 1064,
    "batch_idx": 160,
    "loss": 0.29026854038238525,
    "learning_rate": 0.0004498191425743928,
    "timestep_mean": 484.0,
    "epsilon_norm": 8.151833534240723,
    "prediction_norm": 6.752419948577881,
    "timestamp": "2025-07-14T12:34:11.545329"
  },
  {
    "epoch": 1064,
    "batch_idx": 192,
    "loss": 0.19483275711536407,
    "learning_rate": 0.0004498191425743928,
    "timestep_mean": 551.0,
    "epsilon_norm": 4.467559337615967,
    "prediction_norm": 3.9913086891174316,
    "timestamp": "2025-07-14T12:34:11.550648"
  },
  {
    "epoch": 1065,
    "batch_idx": 0,
    "loss": 0.12583503127098083,
    "learning_rate": 0.0004490377721024753,
    "timestep_mean": 625.09375,
    "epsilon_norm": 7.421847820281982,
    "prediction_norm": 7.159780025482178,
    "timestamp": "2025-07-14T12:34:11.555880"
  },
  {
    "epoch": 1065,
    "batch_idx": 32,
    "loss": 0.16071951389312744,
    "learning_rate": 0.0004490377721024753,
    "timestep_mean": 478.65625,
    "epsilon_norm": 7.113239288330078,
    "prediction_norm": 6.333374977111816,
    "timestamp": "2025-07-14T12:34:11.560924"
  },
  {
    "epoch": 1065,
    "batch_idx": 64,
    "loss": 0.2519223093986511,
    "learning_rate": 0.0004490377721024753,
    "timestep_mean": 527.25,
    "epsilon_norm": 7.655439853668213,
    "prediction_norm": 7.1794562339782715,
    "timestamp": "2025-07-14T12:34:11.565803"
  },
  {
    "epoch": 1065,
    "batch_idx": 96,
    "loss": 0.24686774611473083,
    "learning_rate": 0.0004490377721024753,
    "timestep_mean": 533.21875,
    "epsilon_norm": 7.808172225952148,
    "prediction_norm": 6.513310432434082,
    "timestamp": "2025-07-14T12:34:11.570692"
  },
  {
    "epoch": 1065,
    "batch_idx": 128,
    "loss": 0.2146710902452469,
    "learning_rate": 0.0004490377721024753,
    "timestep_mean": 492.1875,
    "epsilon_norm": 7.478462219238281,
    "prediction_norm": 6.823924541473389,
    "timestamp": "2025-07-14T12:34:11.575739"
  },
  {
    "epoch": 1065,
    "batch_idx": 160,
    "loss": 0.3214794397354126,
    "learning_rate": 0.0004490377721024753,
    "timestep_mean": 490.0625,
    "epsilon_norm": 7.837348461151123,
    "prediction_norm": 6.993500232696533,
    "timestamp": "2025-07-14T12:34:11.580626"
  },
  {
    "epoch": 1065,
    "batch_idx": 192,
    "loss": 0.155431866645813,
    "learning_rate": 0.0004490377721024753,
    "timestep_mean": 629.125,
    "epsilon_norm": 3.9611053466796875,
    "prediction_norm": 3.917658567428589,
    "timestamp": "2025-07-14T12:34:11.585542"
  },
  {
    "epoch": 1066,
    "batch_idx": 0,
    "loss": 0.20742134749889374,
    "learning_rate": 0.000448256527374789,
    "timestep_mean": 506.84375,
    "epsilon_norm": 8.043243408203125,
    "prediction_norm": 7.276275157928467,
    "timestamp": "2025-07-14T12:34:11.590416"
  },
  {
    "epoch": 1066,
    "batch_idx": 32,
    "loss": 0.2170839011669159,
    "learning_rate": 0.000448256527374789,
    "timestep_mean": 534.53125,
    "epsilon_norm": 7.761425971984863,
    "prediction_norm": 6.717581272125244,
    "timestamp": "2025-07-14T12:34:11.595284"
  },
  {
    "epoch": 1066,
    "batch_idx": 64,
    "loss": 0.29733026027679443,
    "learning_rate": 0.000448256527374789,
    "timestep_mean": 447.9375,
    "epsilon_norm": 7.214990615844727,
    "prediction_norm": 6.0374627113342285,
    "timestamp": "2025-07-14T12:34:11.600226"
  },
  {
    "epoch": 1066,
    "batch_idx": 96,
    "loss": 0.18399378657341003,
    "learning_rate": 0.000448256527374789,
    "timestep_mean": 615.71875,
    "epsilon_norm": 7.040813446044922,
    "prediction_norm": 5.885660648345947,
    "timestamp": "2025-07-14T12:34:11.605132"
  },
  {
    "epoch": 1066,
    "batch_idx": 128,
    "loss": 0.411946177482605,
    "learning_rate": 0.000448256527374789,
    "timestep_mean": 407.59375,
    "epsilon_norm": 7.9052863121032715,
    "prediction_norm": 6.2262701988220215,
    "timestamp": "2025-07-14T12:34:11.610042"
  },
  {
    "epoch": 1066,
    "batch_idx": 160,
    "loss": 0.3736134469509125,
    "learning_rate": 0.000448256527374789,
    "timestep_mean": 517.59375,
    "epsilon_norm": 8.063974380493164,
    "prediction_norm": 6.321355819702148,
    "timestamp": "2025-07-14T12:34:11.614899"
  },
  {
    "epoch": 1066,
    "batch_idx": 192,
    "loss": 0.17185445129871368,
    "learning_rate": 0.000448256527374789,
    "timestep_mean": 552.875,
    "epsilon_norm": 4.166488170623779,
    "prediction_norm": 3.6153507232666016,
    "timestamp": "2025-07-14T12:34:11.619881"
  },
  {
    "epoch": 1067,
    "batch_idx": 0,
    "loss": 0.23411013185977936,
    "learning_rate": 0.0004474754103189777,
    "timestep_mean": 482.53125,
    "epsilon_norm": 7.307156085968018,
    "prediction_norm": 6.369582176208496,
    "timestamp": "2025-07-14T12:34:11.624992"
  },
  {
    "epoch": 1067,
    "batch_idx": 32,
    "loss": 0.17352764308452606,
    "learning_rate": 0.0004474754103189777,
    "timestep_mean": 523.125,
    "epsilon_norm": 6.889873027801514,
    "prediction_norm": 6.405750274658203,
    "timestamp": "2025-07-14T12:34:11.630003"
  },
  {
    "epoch": 1067,
    "batch_idx": 64,
    "loss": 0.19044789671897888,
    "learning_rate": 0.0004474754103189777,
    "timestep_mean": 490.78125,
    "epsilon_norm": 7.485281467437744,
    "prediction_norm": 6.645657539367676,
    "timestamp": "2025-07-14T12:34:11.634998"
  },
  {
    "epoch": 1067,
    "batch_idx": 96,
    "loss": 0.33446004986763,
    "learning_rate": 0.0004474754103189777,
    "timestep_mean": 455.75,
    "epsilon_norm": 8.31358528137207,
    "prediction_norm": 6.657124996185303,
    "timestamp": "2025-07-14T12:34:11.639880"
  },
  {
    "epoch": 1067,
    "batch_idx": 128,
    "loss": 0.3508942425251007,
    "learning_rate": 0.0004474754103189777,
    "timestep_mean": 567.0625,
    "epsilon_norm": 7.894062519073486,
    "prediction_norm": 6.399531364440918,
    "timestamp": "2025-07-14T12:34:11.644788"
  },
  {
    "epoch": 1067,
    "batch_idx": 160,
    "loss": 0.3928861618041992,
    "learning_rate": 0.0004474754103189777,
    "timestep_mean": 547.15625,
    "epsilon_norm": 8.097450256347656,
    "prediction_norm": 6.059136867523193,
    "timestamp": "2025-07-14T12:34:11.649948"
  },
  {
    "epoch": 1067,
    "batch_idx": 192,
    "loss": 0.4237925410270691,
    "learning_rate": 0.0004474754103189777,
    "timestep_mean": 390.625,
    "epsilon_norm": 4.512389183044434,
    "prediction_norm": 3.4394350051879883,
    "timestamp": "2025-07-14T12:34:11.654850"
  },
  {
    "epoch": 1068,
    "batch_idx": 0,
    "loss": 0.3048429787158966,
    "learning_rate": 0.0004466944228623702,
    "timestep_mean": 566.0625,
    "epsilon_norm": 6.458933353424072,
    "prediction_norm": 5.063896656036377,
    "timestamp": "2025-07-14T12:34:11.659705"
  },
  {
    "epoch": 1068,
    "batch_idx": 32,
    "loss": 0.4324263036251068,
    "learning_rate": 0.0004466944228623702,
    "timestep_mean": 497.15625,
    "epsilon_norm": 7.152637481689453,
    "prediction_norm": 4.7825164794921875,
    "timestamp": "2025-07-14T12:34:11.664647"
  },
  {
    "epoch": 1068,
    "batch_idx": 64,
    "loss": 0.24208253622055054,
    "learning_rate": 0.0004466944228623702,
    "timestep_mean": 512.3125,
    "epsilon_norm": 8.088546752929688,
    "prediction_norm": 6.1087422370910645,
    "timestamp": "2025-07-14T12:34:11.669672"
  },
  {
    "epoch": 1068,
    "batch_idx": 96,
    "loss": 0.34417226910591125,
    "learning_rate": 0.0004466944228623702,
    "timestep_mean": 410.875,
    "epsilon_norm": 8.076972007751465,
    "prediction_norm": 6.234113693237305,
    "timestamp": "2025-07-14T12:34:11.674671"
  },
  {
    "epoch": 1068,
    "batch_idx": 128,
    "loss": 0.3491774797439575,
    "learning_rate": 0.0004466944228623702,
    "timestep_mean": 466.9375,
    "epsilon_norm": 8.040070533752441,
    "prediction_norm": 6.324603080749512,
    "timestamp": "2025-07-14T12:34:11.679579"
  },
  {
    "epoch": 1068,
    "batch_idx": 160,
    "loss": 0.2334076166152954,
    "learning_rate": 0.0004466944228623702,
    "timestep_mean": 383.34375,
    "epsilon_norm": 8.162811279296875,
    "prediction_norm": 6.574246406555176,
    "timestamp": "2025-07-14T12:34:11.684405"
  },
  {
    "epoch": 1068,
    "batch_idx": 192,
    "loss": 0.22792203724384308,
    "learning_rate": 0.0004466944228623702,
    "timestep_mean": 385.625,
    "epsilon_norm": 3.3699305057525635,
    "prediction_norm": 2.6650710105895996,
    "timestamp": "2025-07-14T12:34:11.689425"
  },
  {
    "epoch": 1069,
    "batch_idx": 0,
    "loss": 0.16521970927715302,
    "learning_rate": 0.0004459135669319754,
    "timestep_mean": 570.28125,
    "epsilon_norm": 7.74006986618042,
    "prediction_norm": 6.856236934661865,
    "timestamp": "2025-07-14T12:34:11.694369"
  },
  {
    "epoch": 1069,
    "batch_idx": 32,
    "loss": 0.3979390859603882,
    "learning_rate": 0.0004459135669319754,
    "timestep_mean": 589.90625,
    "epsilon_norm": 8.650806427001953,
    "prediction_norm": 7.359030246734619,
    "timestamp": "2025-07-14T12:34:11.699291"
  },
  {
    "epoch": 1069,
    "batch_idx": 64,
    "loss": 0.17185832560062408,
    "learning_rate": 0.0004459135669319754,
    "timestep_mean": 556.84375,
    "epsilon_norm": 7.942984104156494,
    "prediction_norm": 7.232264995574951,
    "timestamp": "2025-07-14T12:34:11.704190"
  },
  {
    "epoch": 1069,
    "batch_idx": 96,
    "loss": 0.29673436284065247,
    "learning_rate": 0.0004459135669319754,
    "timestep_mean": 531.71875,
    "epsilon_norm": 7.45667839050293,
    "prediction_norm": 6.224522113800049,
    "timestamp": "2025-07-14T12:34:11.709108"
  },
  {
    "epoch": 1069,
    "batch_idx": 128,
    "loss": 0.31146439909935,
    "learning_rate": 0.0004459135669319754,
    "timestep_mean": 499.75,
    "epsilon_norm": 8.61441421508789,
    "prediction_norm": 7.763672828674316,
    "timestamp": "2025-07-14T12:34:11.714005"
  },
  {
    "epoch": 1069,
    "batch_idx": 160,
    "loss": 0.4951539635658264,
    "learning_rate": 0.0004459135669319754,
    "timestep_mean": 489.25,
    "epsilon_norm": 9.54153823852539,
    "prediction_norm": 6.981064796447754,
    "timestamp": "2025-07-14T12:34:11.718889"
  },
  {
    "epoch": 1069,
    "batch_idx": 192,
    "loss": 0.18106496334075928,
    "learning_rate": 0.0004459135669319754,
    "timestep_mean": 530.25,
    "epsilon_norm": 3.8230905532836914,
    "prediction_norm": 3.6348657608032227,
    "timestamp": "2025-07-14T12:34:11.723792"
  },
  {
    "epoch": 1070,
    "batch_idx": 0,
    "loss": 0.12112490832805634,
    "learning_rate": 0.00044513284445447753,
    "timestep_mean": 488.03125,
    "epsilon_norm": 7.56942892074585,
    "prediction_norm": 7.801550388336182,
    "timestamp": "2025-07-14T12:34:11.728703"
  },
  {
    "epoch": 1070,
    "batch_idx": 32,
    "loss": 0.29800453782081604,
    "learning_rate": 0.00044513284445447753,
    "timestep_mean": 531.78125,
    "epsilon_norm": 7.7191853523254395,
    "prediction_norm": 6.5467047691345215,
    "timestamp": "2025-07-14T12:34:11.733553"
  },
  {
    "epoch": 1070,
    "batch_idx": 64,
    "loss": 0.4348953366279602,
    "learning_rate": 0.00044513284445447753,
    "timestep_mean": 447.375,
    "epsilon_norm": 8.460871696472168,
    "prediction_norm": 7.00685453414917,
    "timestamp": "2025-07-14T12:34:11.738498"
  },
  {
    "epoch": 1070,
    "batch_idx": 96,
    "loss": 0.22059421241283417,
    "learning_rate": 0.00044513284445447753,
    "timestep_mean": 464.46875,
    "epsilon_norm": 7.370913505554199,
    "prediction_norm": 6.047545433044434,
    "timestamp": "2025-07-14T12:34:11.743312"
  },
  {
    "epoch": 1070,
    "batch_idx": 128,
    "loss": 0.33334121108055115,
    "learning_rate": 0.00044513284445447753,
    "timestep_mean": 563.21875,
    "epsilon_norm": 8.041983604431152,
    "prediction_norm": 7.246110439300537,
    "timestamp": "2025-07-14T12:34:11.748201"
  },
  {
    "epoch": 1070,
    "batch_idx": 160,
    "loss": 0.2859248220920563,
    "learning_rate": 0.00044513284445447753,
    "timestep_mean": 485.34375,
    "epsilon_norm": 6.853714466094971,
    "prediction_norm": 6.072449207305908,
    "timestamp": "2025-07-14T12:34:11.753086"
  },
  {
    "epoch": 1070,
    "batch_idx": 192,
    "loss": 0.5992397665977478,
    "learning_rate": 0.00044513284445447753,
    "timestep_mean": 500.5,
    "epsilon_norm": 3.5763704776763916,
    "prediction_norm": 2.2716548442840576,
    "timestamp": "2025-07-14T12:34:11.757933"
  },
  {
    "epoch": 1071,
    "batch_idx": 0,
    "loss": 0.2775539457798004,
    "learning_rate": 0.0004443522573562319,
    "timestep_mean": 513.34375,
    "epsilon_norm": 7.4550628662109375,
    "prediction_norm": 6.557209491729736,
    "timestamp": "2025-07-14T12:34:11.762954"
  },
  {
    "epoch": 1071,
    "batch_idx": 32,
    "loss": 0.387568861246109,
    "learning_rate": 0.0004443522573562319,
    "timestep_mean": 432.28125,
    "epsilon_norm": 7.48595666885376,
    "prediction_norm": 6.021399021148682,
    "timestamp": "2025-07-14T12:34:11.767770"
  },
  {
    "epoch": 1071,
    "batch_idx": 64,
    "loss": 0.244550421833992,
    "learning_rate": 0.0004443522573562319,
    "timestep_mean": 654.78125,
    "epsilon_norm": 8.090840339660645,
    "prediction_norm": 6.769266605377197,
    "timestamp": "2025-07-14T12:34:11.772643"
  },
  {
    "epoch": 1071,
    "batch_idx": 96,
    "loss": 0.22675687074661255,
    "learning_rate": 0.0004443522573562319,
    "timestep_mean": 492.5,
    "epsilon_norm": 8.161450386047363,
    "prediction_norm": 6.899937152862549,
    "timestamp": "2025-07-14T12:34:11.777485"
  },
  {
    "epoch": 1071,
    "batch_idx": 128,
    "loss": 0.2043967843055725,
    "learning_rate": 0.0004443522573562319,
    "timestep_mean": 454.46875,
    "epsilon_norm": 6.909400463104248,
    "prediction_norm": 5.982435703277588,
    "timestamp": "2025-07-14T12:34:11.782340"
  },
  {
    "epoch": 1071,
    "batch_idx": 160,
    "loss": 0.2844613194465637,
    "learning_rate": 0.0004443522573562319,
    "timestep_mean": 566.28125,
    "epsilon_norm": 6.711546897888184,
    "prediction_norm": 5.3087592124938965,
    "timestamp": "2025-07-14T12:34:11.787543"
  },
  {
    "epoch": 1071,
    "batch_idx": 192,
    "loss": 0.47397857904434204,
    "learning_rate": 0.0004443522573562319,
    "timestep_mean": 498.0,
    "epsilon_norm": 3.972337245941162,
    "prediction_norm": 2.7933802604675293,
    "timestamp": "2025-07-14T12:34:11.792769"
  },
  {
    "epoch": 1072,
    "batch_idx": 0,
    "loss": 0.4138126075267792,
    "learning_rate": 0.0004435718075632593,
    "timestep_mean": 340.875,
    "epsilon_norm": 7.317634105682373,
    "prediction_norm": 5.49388313293457,
    "timestamp": "2025-07-14T12:34:11.797825"
  },
  {
    "epoch": 1072,
    "batch_idx": 32,
    "loss": 0.2391985058784485,
    "learning_rate": 0.0004435718075632593,
    "timestep_mean": 498.875,
    "epsilon_norm": 7.057756423950195,
    "prediction_norm": 6.2120256423950195,
    "timestamp": "2025-07-14T12:34:11.803026"
  },
  {
    "epoch": 1072,
    "batch_idx": 64,
    "loss": 0.23447871208190918,
    "learning_rate": 0.0004435718075632593,
    "timestep_mean": 529.71875,
    "epsilon_norm": 9.307023048400879,
    "prediction_norm": 7.63173770904541,
    "timestamp": "2025-07-14T12:34:11.808173"
  },
  {
    "epoch": 1072,
    "batch_idx": 96,
    "loss": 0.3041221499443054,
    "learning_rate": 0.0004435718075632593,
    "timestep_mean": 561.65625,
    "epsilon_norm": 7.592284679412842,
    "prediction_norm": 6.203609466552734,
    "timestamp": "2025-07-14T12:34:11.813366"
  },
  {
    "epoch": 1072,
    "batch_idx": 128,
    "loss": 0.3318626284599304,
    "learning_rate": 0.0004435718075632593,
    "timestep_mean": 515.875,
    "epsilon_norm": 8.649752616882324,
    "prediction_norm": 7.256746292114258,
    "timestamp": "2025-07-14T12:34:11.818554"
  },
  {
    "epoch": 1072,
    "batch_idx": 160,
    "loss": 0.31486794352531433,
    "learning_rate": 0.0004435718075632593,
    "timestep_mean": 492.96875,
    "epsilon_norm": 7.495996475219727,
    "prediction_norm": 6.142335414886475,
    "timestamp": "2025-07-14T12:34:11.823654"
  },
  {
    "epoch": 1072,
    "batch_idx": 192,
    "loss": 0.27169063687324524,
    "learning_rate": 0.0004435718075632593,
    "timestep_mean": 517.5,
    "epsilon_norm": 4.109930992126465,
    "prediction_norm": 2.7685704231262207,
    "timestamp": "2025-07-14T12:34:11.828827"
  },
  {
    "epoch": 1073,
    "batch_idx": 0,
    "loss": 0.21032869815826416,
    "learning_rate": 0.0004427914970012423,
    "timestep_mean": 507.21875,
    "epsilon_norm": 8.944903373718262,
    "prediction_norm": 7.782167911529541,
    "timestamp": "2025-07-14T12:34:11.834110"
  },
  {
    "epoch": 1073,
    "batch_idx": 32,
    "loss": 0.14109140634536743,
    "learning_rate": 0.0004427914970012423,
    "timestep_mean": 533.875,
    "epsilon_norm": 6.783449649810791,
    "prediction_norm": 5.705150604248047,
    "timestamp": "2025-07-14T12:34:11.839381"
  },
  {
    "epoch": 1073,
    "batch_idx": 64,
    "loss": 0.22176194190979004,
    "learning_rate": 0.0004427914970012423,
    "timestep_mean": 454.625,
    "epsilon_norm": 8.101516723632812,
    "prediction_norm": 6.730685234069824,
    "timestamp": "2025-07-14T12:34:11.844605"
  },
  {
    "epoch": 1073,
    "batch_idx": 96,
    "loss": 0.14258192479610443,
    "learning_rate": 0.0004427914970012423,
    "timestep_mean": 522.34375,
    "epsilon_norm": 7.640186786651611,
    "prediction_norm": 7.221160888671875,
    "timestamp": "2025-07-14T12:34:11.849781"
  },
  {
    "epoch": 1073,
    "batch_idx": 128,
    "loss": 0.2146833837032318,
    "learning_rate": 0.0004427914970012423,
    "timestep_mean": 516.5,
    "epsilon_norm": 7.635779857635498,
    "prediction_norm": 7.002543926239014,
    "timestamp": "2025-07-14T12:34:11.855136"
  },
  {
    "epoch": 1073,
    "batch_idx": 160,
    "loss": 0.15717598795890808,
    "learning_rate": 0.0004427914970012423,
    "timestep_mean": 521.6875,
    "epsilon_norm": 7.486786365509033,
    "prediction_norm": 6.686873435974121,
    "timestamp": "2025-07-14T12:34:11.860391"
  },
  {
    "epoch": 1073,
    "batch_idx": 192,
    "loss": 0.4140309691429138,
    "learning_rate": 0.0004427914970012423,
    "timestep_mean": 370.25,
    "epsilon_norm": 4.062103271484375,
    "prediction_norm": 2.850667953491211,
    "timestamp": "2025-07-14T12:34:11.865605"
  },
  {
    "epoch": 1074,
    "batch_idx": 0,
    "loss": 0.1844579130411148,
    "learning_rate": 0.00044201132759551947,
    "timestep_mean": 476.09375,
    "epsilon_norm": 7.621636390686035,
    "prediction_norm": 6.81373405456543,
    "timestamp": "2025-07-14T12:34:11.870748"
  },
  {
    "epoch": 1074,
    "batch_idx": 32,
    "loss": 0.3278493881225586,
    "learning_rate": 0.00044201132759551947,
    "timestep_mean": 449.0625,
    "epsilon_norm": 7.866459846496582,
    "prediction_norm": 6.712477684020996,
    "timestamp": "2025-07-14T12:34:11.875982"
  },
  {
    "epoch": 1074,
    "batch_idx": 64,
    "loss": 0.275969535112381,
    "learning_rate": 0.00044201132759551947,
    "timestep_mean": 581.65625,
    "epsilon_norm": 8.746013641357422,
    "prediction_norm": 7.082254409790039,
    "timestamp": "2025-07-14T12:34:11.881256"
  },
  {
    "epoch": 1074,
    "batch_idx": 96,
    "loss": 0.2435435652732849,
    "learning_rate": 0.00044201132759551947,
    "timestep_mean": 549.21875,
    "epsilon_norm": 7.2688164710998535,
    "prediction_norm": 6.331334590911865,
    "timestamp": "2025-07-14T12:34:11.886437"
  },
  {
    "epoch": 1074,
    "batch_idx": 128,
    "loss": 0.31863683462142944,
    "learning_rate": 0.00044201132759551947,
    "timestep_mean": 547.75,
    "epsilon_norm": 8.026041030883789,
    "prediction_norm": 6.24516487121582,
    "timestamp": "2025-07-14T12:34:11.891877"
  },
  {
    "epoch": 1074,
    "batch_idx": 160,
    "loss": 0.3951917886734009,
    "learning_rate": 0.00044201132759551947,
    "timestep_mean": 461.875,
    "epsilon_norm": 7.558990001678467,
    "prediction_norm": 6.283503532409668,
    "timestamp": "2025-07-14T12:34:11.897060"
  },
  {
    "epoch": 1074,
    "batch_idx": 192,
    "loss": 0.1809057891368866,
    "learning_rate": 0.00044201132759551947,
    "timestep_mean": 516.375,
    "epsilon_norm": 3.5133140087127686,
    "prediction_norm": 2.6714890003204346,
    "timestamp": "2025-07-14T12:34:11.902246"
  },
  {
    "epoch": 1075,
    "batch_idx": 0,
    "loss": 0.355721116065979,
    "learning_rate": 0.0004412313012710814,
    "timestep_mean": 407.1875,
    "epsilon_norm": 9.163841247558594,
    "prediction_norm": 7.038895130157471,
    "timestamp": "2025-07-14T12:34:11.907429"
  },
  {
    "epoch": 1075,
    "batch_idx": 32,
    "loss": 0.3275314271450043,
    "learning_rate": 0.0004412313012710814,
    "timestep_mean": 491.0625,
    "epsilon_norm": 7.961211204528809,
    "prediction_norm": 6.3634033203125,
    "timestamp": "2025-07-14T12:34:11.912642"
  },
  {
    "epoch": 1075,
    "batch_idx": 64,
    "loss": 0.21727673709392548,
    "learning_rate": 0.0004412313012710814,
    "timestep_mean": 514.875,
    "epsilon_norm": 8.10058307647705,
    "prediction_norm": 7.422519207000732,
    "timestamp": "2025-07-14T12:34:11.917857"
  },
  {
    "epoch": 1075,
    "batch_idx": 96,
    "loss": 0.15797770023345947,
    "learning_rate": 0.0004412313012710814,
    "timestep_mean": 541.65625,
    "epsilon_norm": 7.413576126098633,
    "prediction_norm": 7.132377624511719,
    "timestamp": "2025-07-14T12:34:11.923102"
  },
  {
    "epoch": 1075,
    "batch_idx": 128,
    "loss": 0.15401601791381836,
    "learning_rate": 0.0004412313012710814,
    "timestep_mean": 614.28125,
    "epsilon_norm": 7.521587371826172,
    "prediction_norm": 6.772606372833252,
    "timestamp": "2025-07-14T12:34:11.928268"
  },
  {
    "epoch": 1075,
    "batch_idx": 160,
    "loss": 0.21892082691192627,
    "learning_rate": 0.0004412313012710814,
    "timestep_mean": 463.375,
    "epsilon_norm": 7.369764804840088,
    "prediction_norm": 6.377734184265137,
    "timestamp": "2025-07-14T12:34:11.933302"
  },
  {
    "epoch": 1075,
    "batch_idx": 192,
    "loss": 0.7894338369369507,
    "learning_rate": 0.0004412313012710814,
    "timestep_mean": 446.375,
    "epsilon_norm": 4.86049747467041,
    "prediction_norm": 3.490490674972534,
    "timestamp": "2025-07-14T12:34:11.938420"
  },
  {
    "epoch": 1076,
    "batch_idx": 0,
    "loss": 0.2625974118709564,
    "learning_rate": 0.0004404514199525654,
    "timestep_mean": 454.75,
    "epsilon_norm": 7.21293830871582,
    "prediction_norm": 5.9962897300720215,
    "timestamp": "2025-07-14T12:34:11.943578"
  },
  {
    "epoch": 1076,
    "batch_idx": 32,
    "loss": 0.4398789405822754,
    "learning_rate": 0.0004404514199525654,
    "timestep_mean": 586.71875,
    "epsilon_norm": 9.128055572509766,
    "prediction_norm": 7.873600482940674,
    "timestamp": "2025-07-14T12:34:11.948806"
  },
  {
    "epoch": 1076,
    "batch_idx": 64,
    "loss": 0.3424196243286133,
    "learning_rate": 0.0004404514199525654,
    "timestep_mean": 451.5,
    "epsilon_norm": 7.219794750213623,
    "prediction_norm": 6.209558010101318,
    "timestamp": "2025-07-14T12:34:11.953989"
  },
  {
    "epoch": 1076,
    "batch_idx": 96,
    "loss": 0.5944223403930664,
    "learning_rate": 0.0004404514199525654,
    "timestep_mean": 483.65625,
    "epsilon_norm": 9.03372573852539,
    "prediction_norm": 7.174986362457275,
    "timestamp": "2025-07-14T12:34:11.959275"
  },
  {
    "epoch": 1076,
    "batch_idx": 128,
    "loss": 0.30184128880500793,
    "learning_rate": 0.0004404514199525654,
    "timestep_mean": 546.5625,
    "epsilon_norm": 8.781570434570312,
    "prediction_norm": 8.335969924926758,
    "timestamp": "2025-07-14T12:34:11.964503"
  },
  {
    "epoch": 1076,
    "batch_idx": 160,
    "loss": 0.12075286358594894,
    "learning_rate": 0.0004404514199525654,
    "timestep_mean": 538.0,
    "epsilon_norm": 7.851326942443848,
    "prediction_norm": 7.857067108154297,
    "timestamp": "2025-07-14T12:34:11.969775"
  },
  {
    "epoch": 1076,
    "batch_idx": 192,
    "loss": 0.9151589870452881,
    "learning_rate": 0.0004404514199525654,
    "timestep_mean": 280.0,
    "epsilon_norm": 4.314215660095215,
    "prediction_norm": 2.281002998352051,
    "timestamp": "2025-07-14T12:34:11.975049"
  },
  {
    "epoch": 1077,
    "batch_idx": 0,
    "loss": 0.4275314211845398,
    "learning_rate": 0.0004396716855642512,
    "timestep_mean": 522.46875,
    "epsilon_norm": 9.179560661315918,
    "prediction_norm": 7.594773292541504,
    "timestamp": "2025-07-14T12:34:11.980258"
  },
  {
    "epoch": 1077,
    "batch_idx": 32,
    "loss": 0.2747527062892914,
    "learning_rate": 0.0004396716855642512,
    "timestep_mean": 463.6875,
    "epsilon_norm": 7.728724002838135,
    "prediction_norm": 6.505413055419922,
    "timestamp": "2025-07-14T12:34:11.985518"
  },
  {
    "epoch": 1077,
    "batch_idx": 64,
    "loss": 0.2934230864048004,
    "learning_rate": 0.0004396716855642512,
    "timestep_mean": 431.8125,
    "epsilon_norm": 7.930939197540283,
    "prediction_norm": 6.604544162750244,
    "timestamp": "2025-07-14T12:34:11.990750"
  },
  {
    "epoch": 1077,
    "batch_idx": 96,
    "loss": 0.1988886147737503,
    "learning_rate": 0.0004396716855642512,
    "timestep_mean": 470.84375,
    "epsilon_norm": 6.460967540740967,
    "prediction_norm": 5.472397327423096,
    "timestamp": "2025-07-14T12:34:11.995915"
  },
  {
    "epoch": 1077,
    "batch_idx": 128,
    "loss": 0.29340416193008423,
    "learning_rate": 0.0004396716855642512,
    "timestep_mean": 475.84375,
    "epsilon_norm": 7.780401229858398,
    "prediction_norm": 6.867611408233643,
    "timestamp": "2025-07-14T12:34:12.001115"
  },
  {
    "epoch": 1077,
    "batch_idx": 160,
    "loss": 0.5398613214492798,
    "learning_rate": 0.0004396716855642512,
    "timestep_mean": 332.0,
    "epsilon_norm": 7.45318603515625,
    "prediction_norm": 5.5697832107543945,
    "timestamp": "2025-07-14T12:34:12.006352"
  },
  {
    "epoch": 1077,
    "batch_idx": 192,
    "loss": 0.18829819560050964,
    "learning_rate": 0.0004396716855642512,
    "timestep_mean": 481.0,
    "epsilon_norm": 3.960773468017578,
    "prediction_norm": 3.1092910766601562,
    "timestamp": "2025-07-14T12:34:12.011523"
  },
  {
    "epoch": 1078,
    "batch_idx": 0,
    "loss": 0.21299749612808228,
    "learning_rate": 0.00043889210003005544,
    "timestep_mean": 519.1875,
    "epsilon_norm": 7.8318891525268555,
    "prediction_norm": 7.161229133605957,
    "timestamp": "2025-07-14T12:34:12.016866"
  },
  {
    "epoch": 1078,
    "batch_idx": 32,
    "loss": 0.42430388927459717,
    "learning_rate": 0.00043889210003005544,
    "timestep_mean": 578.9375,
    "epsilon_norm": 9.754532814025879,
    "prediction_norm": 8.060038566589355,
    "timestamp": "2025-07-14T12:34:12.022029"
  },
  {
    "epoch": 1078,
    "batch_idx": 64,
    "loss": 0.18049375712871552,
    "learning_rate": 0.00043889210003005544,
    "timestep_mean": 491.375,
    "epsilon_norm": 8.551932334899902,
    "prediction_norm": 7.492531776428223,
    "timestamp": "2025-07-14T12:34:12.026949"
  },
  {
    "epoch": 1078,
    "batch_idx": 96,
    "loss": 0.42503154277801514,
    "learning_rate": 0.00043889210003005544,
    "timestep_mean": 426.875,
    "epsilon_norm": 7.941972255706787,
    "prediction_norm": 6.166097640991211,
    "timestamp": "2025-07-14T12:34:12.032053"
  },
  {
    "epoch": 1078,
    "batch_idx": 128,
    "loss": 0.22188769280910492,
    "learning_rate": 0.00043889210003005544,
    "timestep_mean": 559.125,
    "epsilon_norm": 7.328324794769287,
    "prediction_norm": 5.866427898406982,
    "timestamp": "2025-07-14T12:34:12.037251"
  },
  {
    "epoch": 1078,
    "batch_idx": 160,
    "loss": 0.3517443537712097,
    "learning_rate": 0.00043889210003005544,
    "timestep_mean": 466.90625,
    "epsilon_norm": 8.148962020874023,
    "prediction_norm": 6.138433456420898,
    "timestamp": "2025-07-14T12:34:12.042570"
  },
  {
    "epoch": 1078,
    "batch_idx": 192,
    "loss": 0.5980402231216431,
    "learning_rate": 0.00043889210003005544,
    "timestep_mean": 508.875,
    "epsilon_norm": 3.7633349895477295,
    "prediction_norm": 2.507647752761841,
    "timestamp": "2025-07-14T12:34:12.047569"
  },
  {
    "epoch": 1079,
    "batch_idx": 0,
    "loss": 0.23835858702659607,
    "learning_rate": 0.00043811266527352854,
    "timestep_mean": 469.40625,
    "epsilon_norm": 7.693603515625,
    "prediction_norm": 6.249441623687744,
    "timestamp": "2025-07-14T12:34:12.052429"
  },
  {
    "epoch": 1079,
    "batch_idx": 32,
    "loss": 0.16581806540489197,
    "learning_rate": 0.00043811266527352854,
    "timestep_mean": 508.9375,
    "epsilon_norm": 7.226141452789307,
    "prediction_norm": 6.347757816314697,
    "timestamp": "2025-07-14T12:34:12.057235"
  },
  {
    "epoch": 1079,
    "batch_idx": 64,
    "loss": 0.34502923488616943,
    "learning_rate": 0.00043811266527352854,
    "timestep_mean": 531.375,
    "epsilon_norm": 8.545438766479492,
    "prediction_norm": 7.248028755187988,
    "timestamp": "2025-07-14T12:34:12.062051"
  },
  {
    "epoch": 1079,
    "batch_idx": 96,
    "loss": 0.22664952278137207,
    "learning_rate": 0.00043811266527352854,
    "timestep_mean": 521.6875,
    "epsilon_norm": 7.456523895263672,
    "prediction_norm": 7.060952663421631,
    "timestamp": "2025-07-14T12:34:12.067045"
  },
  {
    "epoch": 1079,
    "batch_idx": 128,
    "loss": 0.2451513707637787,
    "learning_rate": 0.00043811266527352854,
    "timestep_mean": 526.8125,
    "epsilon_norm": 8.846613883972168,
    "prediction_norm": 8.547533988952637,
    "timestamp": "2025-07-14T12:34:12.071907"
  },
  {
    "epoch": 1079,
    "batch_idx": 160,
    "loss": 0.24779871106147766,
    "learning_rate": 0.00043811266527352854,
    "timestep_mean": 496.4375,
    "epsilon_norm": 8.272894859313965,
    "prediction_norm": 7.150203704833984,
    "timestamp": "2025-07-14T12:34:12.076726"
  },
  {
    "epoch": 1079,
    "batch_idx": 192,
    "loss": 0.47692424058914185,
    "learning_rate": 0.00043811266527352854,
    "timestep_mean": 478.875,
    "epsilon_norm": 4.94517183303833,
    "prediction_norm": 4.527525901794434,
    "timestamp": "2025-07-14T12:34:12.081538"
  },
  {
    "epoch": 1080,
    "batch_idx": 0,
    "loss": 0.2905440926551819,
    "learning_rate": 0.0004373333832178481,
    "timestep_mean": 506.9375,
    "epsilon_norm": 8.610819816589355,
    "prediction_norm": 7.8230390548706055,
    "timestamp": "2025-07-14T12:34:12.086400"
  },
  {
    "epoch": 1080,
    "batch_idx": 32,
    "loss": 0.24748387932777405,
    "learning_rate": 0.0004373333832178481,
    "timestep_mean": 524.84375,
    "epsilon_norm": 8.263764381408691,
    "prediction_norm": 7.17728328704834,
    "timestamp": "2025-07-14T12:34:12.091268"
  },
  {
    "epoch": 1080,
    "batch_idx": 64,
    "loss": 0.4498748183250427,
    "learning_rate": 0.0004373333832178481,
    "timestep_mean": 511.21875,
    "epsilon_norm": 9.187427520751953,
    "prediction_norm": 7.8777899742126465,
    "timestamp": "2025-07-14T12:34:12.096224"
  },
  {
    "epoch": 1080,
    "batch_idx": 96,
    "loss": 0.3325941264629364,
    "learning_rate": 0.0004373333832178481,
    "timestep_mean": 438.3125,
    "epsilon_norm": 7.365023136138916,
    "prediction_norm": 5.914534091949463,
    "timestamp": "2025-07-14T12:34:12.101009"
  },
  {
    "epoch": 1080,
    "batch_idx": 128,
    "loss": 0.2759852409362793,
    "learning_rate": 0.0004373333832178481,
    "timestep_mean": 498.3125,
    "epsilon_norm": 8.011417388916016,
    "prediction_norm": 7.066521644592285,
    "timestamp": "2025-07-14T12:34:12.106073"
  },
  {
    "epoch": 1080,
    "batch_idx": 160,
    "loss": 0.3096678555011749,
    "learning_rate": 0.0004373333832178481,
    "timestep_mean": 476.8125,
    "epsilon_norm": 7.895356178283691,
    "prediction_norm": 6.310328483581543,
    "timestamp": "2025-07-14T12:34:12.111005"
  },
  {
    "epoch": 1080,
    "batch_idx": 192,
    "loss": 0.07191526889801025,
    "learning_rate": 0.0004373333832178481,
    "timestep_mean": 675.75,
    "epsilon_norm": 3.5678975582122803,
    "prediction_norm": 3.3172335624694824,
    "timestamp": "2025-07-14T12:34:12.115915"
  },
  {
    "epoch": 1081,
    "batch_idx": 0,
    "loss": 0.2894933819770813,
    "learning_rate": 0.0004365542557858149,
    "timestep_mean": 488.75,
    "epsilon_norm": 7.540174961090088,
    "prediction_norm": 5.686611175537109,
    "timestamp": "2025-07-14T12:34:12.121268"
  },
  {
    "epoch": 1081,
    "batch_idx": 32,
    "loss": 0.46931004524230957,
    "learning_rate": 0.0004365542557858149,
    "timestep_mean": 527.5,
    "epsilon_norm": 8.461565017700195,
    "prediction_norm": 6.314438819885254,
    "timestamp": "2025-07-14T12:34:12.126064"
  },
  {
    "epoch": 1081,
    "batch_idx": 64,
    "loss": 0.11062648147344589,
    "learning_rate": 0.0004365542557858149,
    "timestep_mean": 519.28125,
    "epsilon_norm": 6.768957138061523,
    "prediction_norm": 6.166131973266602,
    "timestamp": "2025-07-14T12:34:12.131019"
  },
  {
    "epoch": 1081,
    "batch_idx": 96,
    "loss": 0.14130540192127228,
    "learning_rate": 0.0004365542557858149,
    "timestep_mean": 401.78125,
    "epsilon_norm": 8.144023895263672,
    "prediction_norm": 7.396359443664551,
    "timestamp": "2025-07-14T12:34:12.135971"
  },
  {
    "epoch": 1081,
    "batch_idx": 128,
    "loss": 0.17089125514030457,
    "learning_rate": 0.0004365542557858149,
    "timestep_mean": 565.28125,
    "epsilon_norm": 7.144979000091553,
    "prediction_norm": 6.033428192138672,
    "timestamp": "2025-07-14T12:34:12.140860"
  },
  {
    "epoch": 1081,
    "batch_idx": 160,
    "loss": 0.20537298917770386,
    "learning_rate": 0.0004365542557858149,
    "timestep_mean": 511.75,
    "epsilon_norm": 7.130883693695068,
    "prediction_norm": 6.083465099334717,
    "timestamp": "2025-07-14T12:34:12.145771"
  },
  {
    "epoch": 1081,
    "batch_idx": 192,
    "loss": 0.0997302234172821,
    "learning_rate": 0.0004365542557858149,
    "timestep_mean": 693.875,
    "epsilon_norm": 3.845831871032715,
    "prediction_norm": 3.389899492263794,
    "timestamp": "2025-07-14T12:34:12.150759"
  },
  {
    "epoch": 1082,
    "batch_idx": 0,
    "loss": 0.197831392288208,
    "learning_rate": 0.0004357752848998488,
    "timestep_mean": 579.78125,
    "epsilon_norm": 10.371163368225098,
    "prediction_norm": 8.899038314819336,
    "timestamp": "2025-07-14T12:34:12.155614"
  },
  {
    "epoch": 1082,
    "batch_idx": 32,
    "loss": 0.2915307581424713,
    "learning_rate": 0.0004357752848998488,
    "timestep_mean": 414.21875,
    "epsilon_norm": 8.094286918640137,
    "prediction_norm": 6.365604400634766,
    "timestamp": "2025-07-14T12:34:12.160516"
  },
  {
    "epoch": 1082,
    "batch_idx": 64,
    "loss": 0.18268559873104095,
    "learning_rate": 0.0004357752848998488,
    "timestep_mean": 513.96875,
    "epsilon_norm": 6.876554489135742,
    "prediction_norm": 5.72585391998291,
    "timestamp": "2025-07-14T12:34:12.165370"
  },
  {
    "epoch": 1082,
    "batch_idx": 96,
    "loss": 0.23176977038383484,
    "learning_rate": 0.0004357752848998488,
    "timestep_mean": 492.5625,
    "epsilon_norm": 7.994462490081787,
    "prediction_norm": 7.149345397949219,
    "timestamp": "2025-07-14T12:34:12.170213"
  },
  {
    "epoch": 1082,
    "batch_idx": 128,
    "loss": 0.2690068185329437,
    "learning_rate": 0.0004357752848998488,
    "timestep_mean": 459.28125,
    "epsilon_norm": 7.813543796539307,
    "prediction_norm": 6.554896354675293,
    "timestamp": "2025-07-14T12:34:12.175119"
  },
  {
    "epoch": 1082,
    "batch_idx": 160,
    "loss": 0.40115004777908325,
    "learning_rate": 0.0004357752848998488,
    "timestep_mean": 465.15625,
    "epsilon_norm": 9.491811752319336,
    "prediction_norm": 7.881278038024902,
    "timestamp": "2025-07-14T12:34:12.179954"
  },
  {
    "epoch": 1082,
    "batch_idx": 192,
    "loss": 0.24081359803676605,
    "learning_rate": 0.0004357752848998488,
    "timestep_mean": 509.125,
    "epsilon_norm": 3.8038740158081055,
    "prediction_norm": 2.8464951515197754,
    "timestamp": "2025-07-14T12:34:12.184834"
  },
  {
    "epoch": 1083,
    "batch_idx": 0,
    "loss": 0.1522161066532135,
    "learning_rate": 0.00043499647248198273,
    "timestep_mean": 520.53125,
    "epsilon_norm": 7.825911045074463,
    "prediction_norm": 7.0148725509643555,
    "timestamp": "2025-07-14T12:34:12.189703"
  },
  {
    "epoch": 1083,
    "batch_idx": 32,
    "loss": 0.31918227672576904,
    "learning_rate": 0.00043499647248198273,
    "timestep_mean": 495.9375,
    "epsilon_norm": 8.436038970947266,
    "prediction_norm": 7.327448844909668,
    "timestamp": "2025-07-14T12:34:12.194583"
  },
  {
    "epoch": 1083,
    "batch_idx": 64,
    "loss": 0.3115289509296417,
    "learning_rate": 0.00043499647248198273,
    "timestep_mean": 466.75,
    "epsilon_norm": 7.604684829711914,
    "prediction_norm": 5.828944206237793,
    "timestamp": "2025-07-14T12:34:12.199443"
  },
  {
    "epoch": 1083,
    "batch_idx": 96,
    "loss": 0.34801506996154785,
    "learning_rate": 0.00043499647248198273,
    "timestep_mean": 502.0,
    "epsilon_norm": 8.314529418945312,
    "prediction_norm": 6.980203628540039,
    "timestamp": "2025-07-14T12:34:12.204412"
  },
  {
    "epoch": 1083,
    "batch_idx": 128,
    "loss": 0.6226751804351807,
    "learning_rate": 0.00043499647248198273,
    "timestep_mean": 501.6875,
    "epsilon_norm": 7.774436950683594,
    "prediction_norm": 5.543066024780273,
    "timestamp": "2025-07-14T12:34:12.209278"
  },
  {
    "epoch": 1083,
    "batch_idx": 160,
    "loss": 0.375052809715271,
    "learning_rate": 0.00043499647248198273,
    "timestep_mean": 501.96875,
    "epsilon_norm": 7.76709508895874,
    "prediction_norm": 6.361296653747559,
    "timestamp": "2025-07-14T12:34:12.214160"
  },
  {
    "epoch": 1083,
    "batch_idx": 192,
    "loss": 0.381611168384552,
    "learning_rate": 0.00043499647248198273,
    "timestep_mean": 356.375,
    "epsilon_norm": 4.9020771980285645,
    "prediction_norm": 4.357779026031494,
    "timestamp": "2025-07-14T12:34:12.218989"
  },
  {
    "epoch": 1084,
    "batch_idx": 0,
    "loss": 0.31801149249076843,
    "learning_rate": 0.0004342178204538589,
    "timestep_mean": 489.15625,
    "epsilon_norm": 7.991066932678223,
    "prediction_norm": 6.257585048675537,
    "timestamp": "2025-07-14T12:34:12.223856"
  },
  {
    "epoch": 1084,
    "batch_idx": 32,
    "loss": 0.430463045835495,
    "learning_rate": 0.0004342178204538589,
    "timestep_mean": 482.03125,
    "epsilon_norm": 7.611769199371338,
    "prediction_norm": 5.6700119972229,
    "timestamp": "2025-07-14T12:34:12.228771"
  },
  {
    "epoch": 1084,
    "batch_idx": 64,
    "loss": 0.43390214443206787,
    "learning_rate": 0.0004342178204538589,
    "timestep_mean": 507.65625,
    "epsilon_norm": 8.9779634475708,
    "prediction_norm": 8.047420501708984,
    "timestamp": "2025-07-14T12:34:12.233673"
  },
  {
    "epoch": 1084,
    "batch_idx": 96,
    "loss": 0.4464193880558014,
    "learning_rate": 0.0004342178204538589,
    "timestep_mean": 467.75,
    "epsilon_norm": 7.8052287101745605,
    "prediction_norm": 6.390194416046143,
    "timestamp": "2025-07-14T12:34:12.238530"
  },
  {
    "epoch": 1084,
    "batch_idx": 128,
    "loss": 0.24045751988887787,
    "learning_rate": 0.0004342178204538589,
    "timestep_mean": 478.5,
    "epsilon_norm": 7.2798590660095215,
    "prediction_norm": 6.684741020202637,
    "timestamp": "2025-07-14T12:34:12.243346"
  },
  {
    "epoch": 1084,
    "batch_idx": 160,
    "loss": 0.21653610467910767,
    "learning_rate": 0.0004342178204538589,
    "timestep_mean": 478.5,
    "epsilon_norm": 8.29968547821045,
    "prediction_norm": 7.902658939361572,
    "timestamp": "2025-07-14T12:34:12.248193"
  },
  {
    "epoch": 1084,
    "batch_idx": 192,
    "loss": 0.3526425361633301,
    "learning_rate": 0.0004342178204538589,
    "timestep_mean": 524.125,
    "epsilon_norm": 4.8604254722595215,
    "prediction_norm": 3.865255832672119,
    "timestamp": "2025-07-14T12:34:12.253028"
  },
  {
    "epoch": 1085,
    "batch_idx": 0,
    "loss": 0.1199522614479065,
    "learning_rate": 0.00043343933073672396,
    "timestep_mean": 514.3125,
    "epsilon_norm": 8.226125717163086,
    "prediction_norm": 8.360688209533691,
    "timestamp": "2025-07-14T12:34:12.257901"
  },
  {
    "epoch": 1085,
    "batch_idx": 32,
    "loss": 0.3011479377746582,
    "learning_rate": 0.00043343933073672396,
    "timestep_mean": 616.03125,
    "epsilon_norm": 7.518039703369141,
    "prediction_norm": 6.568016529083252,
    "timestamp": "2025-07-14T12:34:12.262906"
  },
  {
    "epoch": 1085,
    "batch_idx": 64,
    "loss": 0.27975475788116455,
    "learning_rate": 0.00043343933073672396,
    "timestep_mean": 520.96875,
    "epsilon_norm": 9.219154357910156,
    "prediction_norm": 8.37225341796875,
    "timestamp": "2025-07-14T12:34:12.267795"
  },
  {
    "epoch": 1085,
    "batch_idx": 96,
    "loss": 0.30458036065101624,
    "learning_rate": 0.00043343933073672396,
    "timestep_mean": 452.8125,
    "epsilon_norm": 8.97657299041748,
    "prediction_norm": 7.827511787414551,
    "timestamp": "2025-07-14T12:34:12.272671"
  },
  {
    "epoch": 1085,
    "batch_idx": 128,
    "loss": 0.13746508955955505,
    "learning_rate": 0.00043343933073672396,
    "timestep_mean": 530.1875,
    "epsilon_norm": 6.766396999359131,
    "prediction_norm": 6.644425392150879,
    "timestamp": "2025-07-14T12:34:12.277604"
  },
  {
    "epoch": 1085,
    "batch_idx": 160,
    "loss": 0.2845255136489868,
    "learning_rate": 0.00043343933073672396,
    "timestep_mean": 481.09375,
    "epsilon_norm": 8.93363094329834,
    "prediction_norm": 7.4674577713012695,
    "timestamp": "2025-07-14T12:34:12.282730"
  },
  {
    "epoch": 1085,
    "batch_idx": 192,
    "loss": 0.37730443477630615,
    "learning_rate": 0.00043343933073672396,
    "timestep_mean": 374.25,
    "epsilon_norm": 3.7665724754333496,
    "prediction_norm": 3.2078781127929688,
    "timestamp": "2025-07-14T12:34:12.287865"
  },
  {
    "epoch": 1086,
    "batch_idx": 0,
    "loss": 0.2439565658569336,
    "learning_rate": 0.00043266100525142384,
    "timestep_mean": 512.90625,
    "epsilon_norm": 7.017597675323486,
    "prediction_norm": 5.6821160316467285,
    "timestamp": "2025-07-14T12:34:12.292796"
  },
  {
    "epoch": 1086,
    "batch_idx": 32,
    "loss": 0.3293225169181824,
    "learning_rate": 0.00043266100525142384,
    "timestep_mean": 433.46875,
    "epsilon_norm": 8.799288749694824,
    "prediction_norm": 6.706507682800293,
    "timestamp": "2025-07-14T12:34:12.297706"
  },
  {
    "epoch": 1086,
    "batch_idx": 64,
    "loss": 0.22050194442272186,
    "learning_rate": 0.00043266100525142384,
    "timestep_mean": 560.3125,
    "epsilon_norm": 7.997879505157471,
    "prediction_norm": 6.998102188110352,
    "timestamp": "2025-07-14T12:34:12.302630"
  },
  {
    "epoch": 1086,
    "batch_idx": 96,
    "loss": 0.3074291944503784,
    "learning_rate": 0.00043266100525142384,
    "timestep_mean": 537.8125,
    "epsilon_norm": 9.0713529586792,
    "prediction_norm": 7.672098636627197,
    "timestamp": "2025-07-14T12:34:12.307664"
  },
  {
    "epoch": 1086,
    "batch_idx": 128,
    "loss": 0.16911658644676208,
    "learning_rate": 0.00043266100525142384,
    "timestep_mean": 499.46875,
    "epsilon_norm": 7.433382034301758,
    "prediction_norm": 6.623575210571289,
    "timestamp": "2025-07-14T12:34:12.312548"
  },
  {
    "epoch": 1086,
    "batch_idx": 160,
    "loss": 0.20571470260620117,
    "learning_rate": 0.00043266100525142384,
    "timestep_mean": 539.0625,
    "epsilon_norm": 8.454788208007812,
    "prediction_norm": 7.331025123596191,
    "timestamp": "2025-07-14T12:34:12.317556"
  },
  {
    "epoch": 1086,
    "batch_idx": 192,
    "loss": 0.5917710065841675,
    "learning_rate": 0.00043266100525142384,
    "timestep_mean": 426.125,
    "epsilon_norm": 4.802191257476807,
    "prediction_norm": 3.4952235221862793,
    "timestamp": "2025-07-14T12:34:12.322391"
  },
  {
    "epoch": 1087,
    "batch_idx": 0,
    "loss": 0.28169798851013184,
    "learning_rate": 0.0004318828459183993,
    "timestep_mean": 491.6875,
    "epsilon_norm": 7.989114761352539,
    "prediction_norm": 6.585597038269043,
    "timestamp": "2025-07-14T12:34:12.327307"
  },
  {
    "epoch": 1087,
    "batch_idx": 32,
    "loss": 0.21848052740097046,
    "learning_rate": 0.0004318828459183993,
    "timestep_mean": 494.15625,
    "epsilon_norm": 6.883847236633301,
    "prediction_norm": 6.2987260818481445,
    "timestamp": "2025-07-14T12:34:12.332199"
  },
  {
    "epoch": 1087,
    "batch_idx": 64,
    "loss": 0.2797505855560303,
    "learning_rate": 0.0004318828459183993,
    "timestep_mean": 521.875,
    "epsilon_norm": 9.119832992553711,
    "prediction_norm": 7.336536884307861,
    "timestamp": "2025-07-14T12:34:12.337056"
  },
  {
    "epoch": 1087,
    "batch_idx": 96,
    "loss": 0.23835405707359314,
    "learning_rate": 0.0004318828459183993,
    "timestep_mean": 466.96875,
    "epsilon_norm": 8.178171157836914,
    "prediction_norm": 6.669267177581787,
    "timestamp": "2025-07-14T12:34:12.342105"
  },
  {
    "epoch": 1087,
    "batch_idx": 128,
    "loss": 0.17577385902404785,
    "learning_rate": 0.0004318828459183993,
    "timestep_mean": 564.8125,
    "epsilon_norm": 7.539836406707764,
    "prediction_norm": 6.6759514808654785,
    "timestamp": "2025-07-14T12:34:12.347065"
  },
  {
    "epoch": 1087,
    "batch_idx": 160,
    "loss": 0.25647038221359253,
    "learning_rate": 0.0004318828459183993,
    "timestep_mean": 501.84375,
    "epsilon_norm": 6.813526153564453,
    "prediction_norm": 5.548298358917236,
    "timestamp": "2025-07-14T12:34:12.351955"
  },
  {
    "epoch": 1087,
    "batch_idx": 192,
    "loss": 0.23925277590751648,
    "learning_rate": 0.0004318828459183993,
    "timestep_mean": 650.5,
    "epsilon_norm": 3.8208673000335693,
    "prediction_norm": 2.961930751800537,
    "timestamp": "2025-07-14T12:34:12.356889"
  },
  {
    "epoch": 1088,
    "batch_idx": 0,
    "loss": 0.17149347066879272,
    "learning_rate": 0.0004311048546576811,
    "timestep_mean": 503.4375,
    "epsilon_norm": 6.411287307739258,
    "prediction_norm": 5.72965669631958,
    "timestamp": "2025-07-14T12:34:12.361794"
  },
  {
    "epoch": 1088,
    "batch_idx": 32,
    "loss": 0.2195379137992859,
    "learning_rate": 0.0004311048546576811,
    "timestep_mean": 534.03125,
    "epsilon_norm": 7.624348163604736,
    "prediction_norm": 6.580442428588867,
    "timestamp": "2025-07-14T12:34:12.366836"
  },
  {
    "epoch": 1088,
    "batch_idx": 64,
    "loss": 0.22406429052352905,
    "learning_rate": 0.0004311048546576811,
    "timestep_mean": 474.25,
    "epsilon_norm": 7.008880615234375,
    "prediction_norm": 5.898531436920166,
    "timestamp": "2025-07-14T12:34:12.371803"
  },
  {
    "epoch": 1088,
    "batch_idx": 96,
    "loss": 0.26757168769836426,
    "learning_rate": 0.0004311048546576811,
    "timestep_mean": 492.21875,
    "epsilon_norm": 8.43182373046875,
    "prediction_norm": 7.321303367614746,
    "timestamp": "2025-07-14T12:34:12.376718"
  },
  {
    "epoch": 1088,
    "batch_idx": 128,
    "loss": 0.2709110379219055,
    "learning_rate": 0.0004311048546576811,
    "timestep_mean": 402.71875,
    "epsilon_norm": 6.461386203765869,
    "prediction_norm": 4.9422760009765625,
    "timestamp": "2025-07-14T12:34:12.381709"
  },
  {
    "epoch": 1088,
    "batch_idx": 160,
    "loss": 0.3042471408843994,
    "learning_rate": 0.0004311048546576811,
    "timestep_mean": 601.34375,
    "epsilon_norm": 8.80921745300293,
    "prediction_norm": 7.526735782623291,
    "timestamp": "2025-07-14T12:34:12.386633"
  },
  {
    "epoch": 1088,
    "batch_idx": 192,
    "loss": 0.20548105239868164,
    "learning_rate": 0.0004311048546576811,
    "timestep_mean": 489.5,
    "epsilon_norm": 4.201176166534424,
    "prediction_norm": 3.4916625022888184,
    "timestamp": "2025-07-14T12:34:12.391473"
  },
  {
    "epoch": 1089,
    "batch_idx": 0,
    "loss": 0.28713569045066833,
    "learning_rate": 0.0004303270333888854,
    "timestep_mean": 570.15625,
    "epsilon_norm": 7.757658004760742,
    "prediction_norm": 7.081162929534912,
    "timestamp": "2025-07-14T12:34:12.396477"
  },
  {
    "epoch": 1089,
    "batch_idx": 32,
    "loss": 0.24435998499393463,
    "learning_rate": 0.0004303270333888854,
    "timestep_mean": 404.59375,
    "epsilon_norm": 7.985776424407959,
    "prediction_norm": 6.731292247772217,
    "timestamp": "2025-07-14T12:34:12.401373"
  },
  {
    "epoch": 1089,
    "batch_idx": 64,
    "loss": 0.2998565137386322,
    "learning_rate": 0.0004303270333888854,
    "timestep_mean": 514.875,
    "epsilon_norm": 6.993924140930176,
    "prediction_norm": 6.039185523986816,
    "timestamp": "2025-07-14T12:34:12.406302"
  },
  {
    "epoch": 1089,
    "batch_idx": 96,
    "loss": 0.44389474391937256,
    "learning_rate": 0.0004303270333888854,
    "timestep_mean": 493.625,
    "epsilon_norm": 9.043617248535156,
    "prediction_norm": 7.111050605773926,
    "timestamp": "2025-07-14T12:34:12.411147"
  },
  {
    "epoch": 1089,
    "batch_idx": 128,
    "loss": 0.17776831984519958,
    "learning_rate": 0.0004303270333888854,
    "timestep_mean": 461.6875,
    "epsilon_norm": 8.65402603149414,
    "prediction_norm": 8.208271026611328,
    "timestamp": "2025-07-14T12:34:12.416409"
  },
  {
    "epoch": 1089,
    "batch_idx": 160,
    "loss": 0.18648478388786316,
    "learning_rate": 0.0004303270333888854,
    "timestep_mean": 498.4375,
    "epsilon_norm": 7.453408718109131,
    "prediction_norm": 6.780092239379883,
    "timestamp": "2025-07-14T12:34:12.421664"
  },
  {
    "epoch": 1089,
    "batch_idx": 192,
    "loss": 0.20941711962223053,
    "learning_rate": 0.0004303270333888854,
    "timestep_mean": 416.375,
    "epsilon_norm": 3.6838741302490234,
    "prediction_norm": 2.57989764213562,
    "timestamp": "2025-07-14T12:34:12.426741"
  },
  {
    "epoch": 1090,
    "batch_idx": 0,
    "loss": 0.28345006704330444,
    "learning_rate": 0.00042954938403120873,
    "timestep_mean": 400.78125,
    "epsilon_norm": 8.317681312561035,
    "prediction_norm": 7.280088424682617,
    "timestamp": "2025-07-14T12:34:12.432417"
  },
  {
    "epoch": 1090,
    "batch_idx": 32,
    "loss": 0.6187887787818909,
    "learning_rate": 0.00042954938403120873,
    "timestep_mean": 527.09375,
    "epsilon_norm": 9.22141170501709,
    "prediction_norm": 6.968283653259277,
    "timestamp": "2025-07-14T12:34:12.437756"
  },
  {
    "epoch": 1090,
    "batch_idx": 64,
    "loss": 0.2895963490009308,
    "learning_rate": 0.00042954938403120873,
    "timestep_mean": 506.71875,
    "epsilon_norm": 8.287626266479492,
    "prediction_norm": 6.789325714111328,
    "timestamp": "2025-07-14T12:34:12.442706"
  },
  {
    "epoch": 1090,
    "batch_idx": 96,
    "loss": 0.20049631595611572,
    "learning_rate": 0.00042954938403120873,
    "timestep_mean": 553.6875,
    "epsilon_norm": 8.107177734375,
    "prediction_norm": 7.232448101043701,
    "timestamp": "2025-07-14T12:34:12.447840"
  },
  {
    "epoch": 1090,
    "batch_idx": 128,
    "loss": 0.1531514823436737,
    "learning_rate": 0.00042954938403120873,
    "timestep_mean": 547.375,
    "epsilon_norm": 7.256232261657715,
    "prediction_norm": 6.898074626922607,
    "timestamp": "2025-07-14T12:34:12.452786"
  },
  {
    "epoch": 1090,
    "batch_idx": 160,
    "loss": 0.1334046721458435,
    "learning_rate": 0.00042954938403120873,
    "timestep_mean": 527.09375,
    "epsilon_norm": 9.737860679626465,
    "prediction_norm": 9.23892593383789,
    "timestamp": "2025-07-14T12:34:12.457819"
  },
  {
    "epoch": 1090,
    "batch_idx": 192,
    "loss": 0.08888622373342514,
    "learning_rate": 0.00042954938403120873,
    "timestep_mean": 586.25,
    "epsilon_norm": 3.7576777935028076,
    "prediction_norm": 3.743455171585083,
    "timestamp": "2025-07-14T12:34:12.462974"
  },
  {
    "epoch": 1091,
    "batch_idx": 0,
    "loss": 0.1789613962173462,
    "learning_rate": 0.00042877190850342374,
    "timestep_mean": 475.65625,
    "epsilon_norm": 7.707596778869629,
    "prediction_norm": 7.523307800292969,
    "timestamp": "2025-07-14T12:34:12.467931"
  },
  {
    "epoch": 1091,
    "batch_idx": 32,
    "loss": 0.2919413149356842,
    "learning_rate": 0.00042877190850342374,
    "timestep_mean": 489.46875,
    "epsilon_norm": 8.863186836242676,
    "prediction_norm": 7.834228992462158,
    "timestamp": "2025-07-14T12:34:12.472888"
  },
  {
    "epoch": 1091,
    "batch_idx": 64,
    "loss": 0.13484497368335724,
    "learning_rate": 0.00042877190850342374,
    "timestep_mean": 583.40625,
    "epsilon_norm": 8.123793601989746,
    "prediction_norm": 7.879429340362549,
    "timestamp": "2025-07-14T12:34:12.477867"
  },
  {
    "epoch": 1091,
    "batch_idx": 96,
    "loss": 0.2309601902961731,
    "learning_rate": 0.00042877190850342374,
    "timestep_mean": 574.09375,
    "epsilon_norm": 7.555606365203857,
    "prediction_norm": 6.146485805511475,
    "timestamp": "2025-07-14T12:34:12.482916"
  },
  {
    "epoch": 1091,
    "batch_idx": 128,
    "loss": 0.3792949318885803,
    "learning_rate": 0.00042877190850342374,
    "timestep_mean": 510.5,
    "epsilon_norm": 7.238297462463379,
    "prediction_norm": 6.175392150878906,
    "timestamp": "2025-07-14T12:34:12.488183"
  },
  {
    "epoch": 1091,
    "batch_idx": 160,
    "loss": 0.31514817476272583,
    "learning_rate": 0.00042877190850342374,
    "timestep_mean": 456.25,
    "epsilon_norm": 8.608053207397461,
    "prediction_norm": 6.898605823516846,
    "timestamp": "2025-07-14T12:34:12.493067"
  },
  {
    "epoch": 1091,
    "batch_idx": 192,
    "loss": 0.061239853501319885,
    "learning_rate": 0.00042877190850342374,
    "timestep_mean": 605.5,
    "epsilon_norm": 3.7091660499572754,
    "prediction_norm": 3.616826057434082,
    "timestamp": "2025-07-14T12:34:12.498020"
  },
  {
    "epoch": 1092,
    "batch_idx": 0,
    "loss": 0.22329695522785187,
    "learning_rate": 0.00042799460872387394,
    "timestep_mean": 489.75,
    "epsilon_norm": 8.175793647766113,
    "prediction_norm": 6.497687816619873,
    "timestamp": "2025-07-14T12:34:12.503034"
  },
  {
    "epoch": 1092,
    "batch_idx": 32,
    "loss": 0.3427320718765259,
    "learning_rate": 0.00042799460872387394,
    "timestep_mean": 497.8125,
    "epsilon_norm": 8.781318664550781,
    "prediction_norm": 7.168643951416016,
    "timestamp": "2025-07-14T12:34:12.508133"
  },
  {
    "epoch": 1092,
    "batch_idx": 64,
    "loss": 0.4907287657260895,
    "learning_rate": 0.00042799460872387394,
    "timestep_mean": 490.6875,
    "epsilon_norm": 9.13298225402832,
    "prediction_norm": 7.727997303009033,
    "timestamp": "2025-07-14T12:34:12.513557"
  },
  {
    "epoch": 1092,
    "batch_idx": 96,
    "loss": 0.33987152576446533,
    "learning_rate": 0.00042799460872387394,
    "timestep_mean": 509.5625,
    "epsilon_norm": 8.598655700683594,
    "prediction_norm": 7.180562496185303,
    "timestamp": "2025-07-14T12:34:12.518857"
  },
  {
    "epoch": 1092,
    "batch_idx": 128,
    "loss": 0.3458213210105896,
    "learning_rate": 0.00042799460872387394,
    "timestep_mean": 499.5,
    "epsilon_norm": 8.622183799743652,
    "prediction_norm": 7.251500606536865,
    "timestamp": "2025-07-14T12:34:12.524297"
  },
  {
    "epoch": 1092,
    "batch_idx": 160,
    "loss": 0.35518690943717957,
    "learning_rate": 0.00042799460872387394,
    "timestep_mean": 391.25,
    "epsilon_norm": 7.2618584632873535,
    "prediction_norm": 5.135640621185303,
    "timestamp": "2025-07-14T12:34:12.529685"
  },
  {
    "epoch": 1092,
    "batch_idx": 192,
    "loss": 0.10380832850933075,
    "learning_rate": 0.00042799460872387394,
    "timestep_mean": 595.875,
    "epsilon_norm": 3.5876963138580322,
    "prediction_norm": 3.573049545288086,
    "timestamp": "2025-07-14T12:34:12.535110"
  },
  {
    "epoch": 1093,
    "batch_idx": 0,
    "loss": 0.23800019919872284,
    "learning_rate": 0.0004272174866104694,
    "timestep_mean": 422.59375,
    "epsilon_norm": 6.735644340515137,
    "prediction_norm": 5.778263092041016,
    "timestamp": "2025-07-14T12:34:12.540321"
  },
  {
    "epoch": 1093,
    "batch_idx": 32,
    "loss": 0.2625506520271301,
    "learning_rate": 0.0004272174866104694,
    "timestep_mean": 523.34375,
    "epsilon_norm": 7.382158279418945,
    "prediction_norm": 6.854049205780029,
    "timestamp": "2025-07-14T12:34:12.545606"
  },
  {
    "epoch": 1093,
    "batch_idx": 64,
    "loss": 0.35597550868988037,
    "learning_rate": 0.0004272174866104694,
    "timestep_mean": 546.375,
    "epsilon_norm": 8.32087516784668,
    "prediction_norm": 7.2011260986328125,
    "timestamp": "2025-07-14T12:34:12.550918"
  },
  {
    "epoch": 1093,
    "batch_idx": 96,
    "loss": 0.252663254737854,
    "learning_rate": 0.0004272174866104694,
    "timestep_mean": 499.25,
    "epsilon_norm": 8.753924369812012,
    "prediction_norm": 7.740164756774902,
    "timestamp": "2025-07-14T12:34:12.556222"
  },
  {
    "epoch": 1093,
    "batch_idx": 128,
    "loss": 0.3348395526409149,
    "learning_rate": 0.0004272174866104694,
    "timestep_mean": 499.46875,
    "epsilon_norm": 8.366047859191895,
    "prediction_norm": 6.597012519836426,
    "timestamp": "2025-07-14T12:34:12.561432"
  },
  {
    "epoch": 1093,
    "batch_idx": 160,
    "loss": 0.32337456941604614,
    "learning_rate": 0.0004272174866104694,
    "timestep_mean": 617.65625,
    "epsilon_norm": 8.907754898071289,
    "prediction_norm": 7.167025566101074,
    "timestamp": "2025-07-14T12:34:12.566737"
  },
  {
    "epoch": 1093,
    "batch_idx": 192,
    "loss": 0.059445783495903015,
    "learning_rate": 0.0004272174866104694,
    "timestep_mean": 515.625,
    "epsilon_norm": 2.6629960536956787,
    "prediction_norm": 2.452845335006714,
    "timestamp": "2025-07-14T12:34:12.571938"
  },
  {
    "epoch": 1094,
    "batch_idx": 0,
    "loss": 0.22904320061206818,
    "learning_rate": 0.00042644054408068143,
    "timestep_mean": 594.84375,
    "epsilon_norm": 6.898685455322266,
    "prediction_norm": 6.138251304626465,
    "timestamp": "2025-07-14T12:34:12.577210"
  },
  {
    "epoch": 1094,
    "batch_idx": 32,
    "loss": 0.32427477836608887,
    "learning_rate": 0.00042644054408068143,
    "timestep_mean": 468.53125,
    "epsilon_norm": 7.724976062774658,
    "prediction_norm": 6.145637035369873,
    "timestamp": "2025-07-14T12:34:12.582385"
  },
  {
    "epoch": 1094,
    "batch_idx": 64,
    "loss": 0.28597360849380493,
    "learning_rate": 0.00042644054408068143,
    "timestep_mean": 383.3125,
    "epsilon_norm": 7.512221813201904,
    "prediction_norm": 6.208839416503906,
    "timestamp": "2025-07-14T12:34:12.587671"
  },
  {
    "epoch": 1094,
    "batch_idx": 96,
    "loss": 0.1933876872062683,
    "learning_rate": 0.00042644054408068143,
    "timestep_mean": 512.84375,
    "epsilon_norm": 7.304871559143066,
    "prediction_norm": 5.926654815673828,
    "timestamp": "2025-07-14T12:34:12.592833"
  },
  {
    "epoch": 1094,
    "batch_idx": 128,
    "loss": 0.27653059363365173,
    "learning_rate": 0.00042644054408068143,
    "timestep_mean": 480.28125,
    "epsilon_norm": 7.497504234313965,
    "prediction_norm": 6.056478023529053,
    "timestamp": "2025-07-14T12:34:12.598004"
  },
  {
    "epoch": 1094,
    "batch_idx": 160,
    "loss": 0.33337926864624023,
    "learning_rate": 0.00042644054408068143,
    "timestep_mean": 433.625,
    "epsilon_norm": 8.632488250732422,
    "prediction_norm": 7.599960803985596,
    "timestamp": "2025-07-14T12:34:12.603217"
  },
  {
    "epoch": 1094,
    "batch_idx": 192,
    "loss": 0.17179732024669647,
    "learning_rate": 0.00042644054408068143,
    "timestep_mean": 500.25,
    "epsilon_norm": 4.078885555267334,
    "prediction_norm": 3.4137725830078125,
    "timestamp": "2025-07-14T12:34:12.608394"
  },
  {
    "epoch": 1095,
    "batch_idx": 0,
    "loss": 0.20013108849525452,
    "learning_rate": 0.00042566378305153866,
    "timestep_mean": 513.09375,
    "epsilon_norm": 7.128103733062744,
    "prediction_norm": 6.341475009918213,
    "timestamp": "2025-07-14T12:34:12.613621"
  },
  {
    "epoch": 1095,
    "batch_idx": 32,
    "loss": 0.28563252091407776,
    "learning_rate": 0.00042566378305153866,
    "timestep_mean": 452.8125,
    "epsilon_norm": 7.342031002044678,
    "prediction_norm": 6.489666938781738,
    "timestamp": "2025-07-14T12:34:12.618941"
  },
  {
    "epoch": 1095,
    "batch_idx": 64,
    "loss": 0.3214484453201294,
    "learning_rate": 0.00042566378305153866,
    "timestep_mean": 447.5625,
    "epsilon_norm": 7.527287483215332,
    "prediction_norm": 6.567008018493652,
    "timestamp": "2025-07-14T12:34:12.624105"
  },
  {
    "epoch": 1095,
    "batch_idx": 96,
    "loss": 0.39071816205978394,
    "learning_rate": 0.00042566378305153866,
    "timestep_mean": 429.53125,
    "epsilon_norm": 8.907455444335938,
    "prediction_norm": 7.236823081970215,
    "timestamp": "2025-07-14T12:34:12.629302"
  },
  {
    "epoch": 1095,
    "batch_idx": 128,
    "loss": 0.21363070607185364,
    "learning_rate": 0.00042566378305153866,
    "timestep_mean": 525.375,
    "epsilon_norm": 7.215704917907715,
    "prediction_norm": 6.454107284545898,
    "timestamp": "2025-07-14T12:34:12.634531"
  },
  {
    "epoch": 1095,
    "batch_idx": 160,
    "loss": 0.19486519694328308,
    "learning_rate": 0.00042566378305153866,
    "timestep_mean": 492.6875,
    "epsilon_norm": 7.081259250640869,
    "prediction_norm": 5.759165287017822,
    "timestamp": "2025-07-14T12:34:12.639680"
  },
  {
    "epoch": 1095,
    "batch_idx": 192,
    "loss": 0.15371567010879517,
    "learning_rate": 0.00042566378305153866,
    "timestep_mean": 378.875,
    "epsilon_norm": 3.3121683597564697,
    "prediction_norm": 3.2113192081451416,
    "timestamp": "2025-07-14T12:34:12.644842"
  },
  {
    "epoch": 1096,
    "batch_idx": 0,
    "loss": 0.2897893786430359,
    "learning_rate": 0.00042488720543962146,
    "timestep_mean": 491.96875,
    "epsilon_norm": 7.394034385681152,
    "prediction_norm": 5.584429740905762,
    "timestamp": "2025-07-14T12:34:12.650213"
  },
  {
    "epoch": 1096,
    "batch_idx": 32,
    "loss": 0.2381410300731659,
    "learning_rate": 0.00042488720543962146,
    "timestep_mean": 490.34375,
    "epsilon_norm": 6.411226272583008,
    "prediction_norm": 5.608631610870361,
    "timestamp": "2025-07-14T12:34:12.655412"
  },
  {
    "epoch": 1096,
    "batch_idx": 64,
    "loss": 0.26396968960762024,
    "learning_rate": 0.00042488720543962146,
    "timestep_mean": 588.75,
    "epsilon_norm": 7.63314151763916,
    "prediction_norm": 7.076653480529785,
    "timestamp": "2025-07-14T12:34:12.660647"
  },
  {
    "epoch": 1096,
    "batch_idx": 96,
    "loss": 0.37117043137550354,
    "learning_rate": 0.00042488720543962146,
    "timestep_mean": 444.71875,
    "epsilon_norm": 9.342737197875977,
    "prediction_norm": 7.844005584716797,
    "timestamp": "2025-07-14T12:34:12.665892"
  },
  {
    "epoch": 1096,
    "batch_idx": 128,
    "loss": 0.19522982835769653,
    "learning_rate": 0.00042488720543962146,
    "timestep_mean": 402.90625,
    "epsilon_norm": 8.278518676757812,
    "prediction_norm": 7.304484844207764,
    "timestamp": "2025-07-14T12:34:12.671036"
  },
  {
    "epoch": 1096,
    "batch_idx": 160,
    "loss": 0.3083093762397766,
    "learning_rate": 0.00042488720543962146,
    "timestep_mean": 467.96875,
    "epsilon_norm": 8.619730949401855,
    "prediction_norm": 7.466369152069092,
    "timestamp": "2025-07-14T12:34:12.676193"
  },
  {
    "epoch": 1096,
    "batch_idx": 192,
    "loss": 0.3880234956741333,
    "learning_rate": 0.00042488720543962146,
    "timestep_mean": 411.625,
    "epsilon_norm": 3.6869328022003174,
    "prediction_norm": 3.1195695400238037,
    "timestamp": "2025-07-14T12:34:12.681356"
  },
  {
    "epoch": 1097,
    "batch_idx": 0,
    "loss": 0.18805596232414246,
    "learning_rate": 0.00042411081316105824,
    "timestep_mean": 526.28125,
    "epsilon_norm": 7.755661487579346,
    "prediction_norm": 7.147000312805176,
    "timestamp": "2025-07-14T12:34:12.686473"
  },
  {
    "epoch": 1097,
    "batch_idx": 32,
    "loss": 0.27962544560432434,
    "learning_rate": 0.00042411081316105824,
    "timestep_mean": 504.5,
    "epsilon_norm": 7.728320121765137,
    "prediction_norm": 6.805627346038818,
    "timestamp": "2025-07-14T12:34:12.691734"
  },
  {
    "epoch": 1097,
    "batch_idx": 64,
    "loss": 0.22118589282035828,
    "learning_rate": 0.00042411081316105824,
    "timestep_mean": 486.71875,
    "epsilon_norm": 7.189478397369385,
    "prediction_norm": 6.120944976806641,
    "timestamp": "2025-07-14T12:34:12.696978"
  },
  {
    "epoch": 1097,
    "batch_idx": 96,
    "loss": 0.35107338428497314,
    "learning_rate": 0.00042411081316105824,
    "timestep_mean": 472.875,
    "epsilon_norm": 8.8336820602417,
    "prediction_norm": 7.621413707733154,
    "timestamp": "2025-07-14T12:34:12.702135"
  },
  {
    "epoch": 1097,
    "batch_idx": 128,
    "loss": 0.21770153939723969,
    "learning_rate": 0.00042411081316105824,
    "timestep_mean": 484.25,
    "epsilon_norm": 6.221441745758057,
    "prediction_norm": 5.39824104309082,
    "timestamp": "2025-07-14T12:34:12.707349"
  },
  {
    "epoch": 1097,
    "batch_idx": 160,
    "loss": 0.24360178411006927,
    "learning_rate": 0.00042411081316105824,
    "timestep_mean": 495.78125,
    "epsilon_norm": 8.574090957641602,
    "prediction_norm": 7.718141555786133,
    "timestamp": "2025-07-14T12:34:12.712579"
  },
  {
    "epoch": 1097,
    "batch_idx": 192,
    "loss": 0.053765375167131424,
    "learning_rate": 0.00042411081316105824,
    "timestep_mean": 572.5,
    "epsilon_norm": 3.404974937438965,
    "prediction_norm": 3.2022440433502197,
    "timestamp": "2025-07-14T12:34:12.717770"
  },
  {
    "epoch": 1098,
    "batch_idx": 0,
    "loss": 0.3807847797870636,
    "learning_rate": 0.00042333460813151973,
    "timestep_mean": 526.46875,
    "epsilon_norm": 7.8528828620910645,
    "prediction_norm": 6.511068820953369,
    "timestamp": "2025-07-14T12:34:12.722984"
  },
  {
    "epoch": 1098,
    "batch_idx": 32,
    "loss": 0.34329625964164734,
    "learning_rate": 0.00042333460813151973,
    "timestep_mean": 449.4375,
    "epsilon_norm": 8.654242515563965,
    "prediction_norm": 7.325904846191406,
    "timestamp": "2025-07-14T12:34:12.728202"
  },
  {
    "epoch": 1098,
    "batch_idx": 64,
    "loss": 0.13688918948173523,
    "learning_rate": 0.00042333460813151973,
    "timestep_mean": 554.3125,
    "epsilon_norm": 8.19226360321045,
    "prediction_norm": 6.916388988494873,
    "timestamp": "2025-07-14T12:34:12.733519"
  },
  {
    "epoch": 1098,
    "batch_idx": 96,
    "loss": 0.21083566546440125,
    "learning_rate": 0.00042333460813151973,
    "timestep_mean": 516.125,
    "epsilon_norm": 7.363750457763672,
    "prediction_norm": 6.288851261138916,
    "timestamp": "2025-07-14T12:34:12.738737"
  },
  {
    "epoch": 1098,
    "batch_idx": 128,
    "loss": 0.18802425265312195,
    "learning_rate": 0.00042333460813151973,
    "timestep_mean": 467.0,
    "epsilon_norm": 8.30911922454834,
    "prediction_norm": 6.887455940246582,
    "timestamp": "2025-07-14T12:34:12.743991"
  },
  {
    "epoch": 1098,
    "batch_idx": 160,
    "loss": 0.1779174655675888,
    "learning_rate": 0.00042333460813151973,
    "timestep_mean": 497.8125,
    "epsilon_norm": 8.15275764465332,
    "prediction_norm": 7.075674057006836,
    "timestamp": "2025-07-14T12:34:12.749136"
  },
  {
    "epoch": 1098,
    "batch_idx": 192,
    "loss": 0.1083020269870758,
    "learning_rate": 0.00042333460813151973,
    "timestep_mean": 443.5,
    "epsilon_norm": 2.9767138957977295,
    "prediction_norm": 2.6970114707946777,
    "timestamp": "2025-07-14T12:34:12.754293"
  },
  {
    "epoch": 1099,
    "batch_idx": 0,
    "loss": 0.2435605376958847,
    "learning_rate": 0.00042255859226621444,
    "timestep_mean": 533.75,
    "epsilon_norm": 6.983161449432373,
    "prediction_norm": 5.767615795135498,
    "timestamp": "2025-07-14T12:34:12.759468"
  },
  {
    "epoch": 1099,
    "batch_idx": 32,
    "loss": 0.2515246570110321,
    "learning_rate": 0.00042255859226621444,
    "timestep_mean": 532.96875,
    "epsilon_norm": 8.593570709228516,
    "prediction_norm": 6.646726608276367,
    "timestamp": "2025-07-14T12:34:12.764677"
  },
  {
    "epoch": 1099,
    "batch_idx": 64,
    "loss": 0.3724727928638458,
    "learning_rate": 0.00042255859226621444,
    "timestep_mean": 444.84375,
    "epsilon_norm": 8.575911521911621,
    "prediction_norm": 6.746446132659912,
    "timestamp": "2025-07-14T12:34:12.769943"
  },
  {
    "epoch": 1099,
    "batch_idx": 96,
    "loss": 0.19801197946071625,
    "learning_rate": 0.00042255859226621444,
    "timestep_mean": 521.625,
    "epsilon_norm": 7.353024482727051,
    "prediction_norm": 6.349050998687744,
    "timestamp": "2025-07-14T12:34:12.775290"
  },
  {
    "epoch": 1099,
    "batch_idx": 128,
    "loss": 0.2064514011144638,
    "learning_rate": 0.00042255859226621444,
    "timestep_mean": 517.3125,
    "epsilon_norm": 7.811820983886719,
    "prediction_norm": 6.948294639587402,
    "timestamp": "2025-07-14T12:34:12.780539"
  },
  {
    "epoch": 1099,
    "batch_idx": 160,
    "loss": 0.4104122519493103,
    "learning_rate": 0.00042255859226621444,
    "timestep_mean": 462.84375,
    "epsilon_norm": 7.736828327178955,
    "prediction_norm": 6.307469367980957,
    "timestamp": "2025-07-14T12:34:12.785692"
  },
  {
    "epoch": 1099,
    "batch_idx": 192,
    "loss": 0.1553100347518921,
    "learning_rate": 0.00042255859226621444,
    "timestep_mean": 638.25,
    "epsilon_norm": 3.7794628143310547,
    "prediction_norm": 3.3254857063293457,
    "timestamp": "2025-07-14T12:34:12.790796"
  },
  {
    "epoch": 1100,
    "batch_idx": 0,
    "loss": 0.45978179574012756,
    "learning_rate": 0.0004217827674798846,
    "timestep_mean": 469.5625,
    "epsilon_norm": 8.671622276306152,
    "prediction_norm": 6.868691444396973,
    "timestamp": "2025-07-14T12:34:12.796018"
  },
  {
    "epoch": 1100,
    "batch_idx": 32,
    "loss": 0.29777470231056213,
    "learning_rate": 0.0004217827674798846,
    "timestep_mean": 448.375,
    "epsilon_norm": 8.724682807922363,
    "prediction_norm": 7.28213357925415,
    "timestamp": "2025-07-14T12:34:12.801256"
  },
  {
    "epoch": 1100,
    "batch_idx": 64,
    "loss": 0.3414262533187866,
    "learning_rate": 0.0004217827674798846,
    "timestep_mean": 457.15625,
    "epsilon_norm": 8.727299690246582,
    "prediction_norm": 7.263301849365234,
    "timestamp": "2025-07-14T12:34:12.806440"
  },
  {
    "epoch": 1100,
    "batch_idx": 96,
    "loss": 0.40147989988327026,
    "learning_rate": 0.0004217827674798846,
    "timestep_mean": 509.9375,
    "epsilon_norm": 7.010186672210693,
    "prediction_norm": 5.948097229003906,
    "timestamp": "2025-07-14T12:34:12.811654"
  },
  {
    "epoch": 1100,
    "batch_idx": 128,
    "loss": 0.19149234890937805,
    "learning_rate": 0.0004217827674798846,
    "timestep_mean": 545.15625,
    "epsilon_norm": 8.632673263549805,
    "prediction_norm": 7.732392311096191,
    "timestamp": "2025-07-14T12:34:12.816950"
  },
  {
    "epoch": 1100,
    "batch_idx": 160,
    "loss": 0.26088255643844604,
    "learning_rate": 0.0004217827674798846,
    "timestep_mean": 451.8125,
    "epsilon_norm": 8.411225318908691,
    "prediction_norm": 7.725308418273926,
    "timestamp": "2025-07-14T12:34:12.822215"
  },
  {
    "epoch": 1100,
    "batch_idx": 192,
    "loss": 0.5130875110626221,
    "learning_rate": 0.0004217827674798846,
    "timestep_mean": 585.625,
    "epsilon_norm": 4.5878119468688965,
    "prediction_norm": 3.5355043411254883,
    "timestamp": "2025-07-14T12:34:12.827357"
  },
  {
    "epoch": 1101,
    "batch_idx": 0,
    "loss": 0.41751736402511597,
    "learning_rate": 0.00042100713568680065,
    "timestep_mean": 432.6875,
    "epsilon_norm": 8.875537872314453,
    "prediction_norm": 6.641766548156738,
    "timestamp": "2025-07-14T12:34:12.832541"
  },
  {
    "epoch": 1101,
    "batch_idx": 32,
    "loss": 0.42227664589881897,
    "learning_rate": 0.00042100713568680065,
    "timestep_mean": 477.5625,
    "epsilon_norm": 7.536775588989258,
    "prediction_norm": 5.930356025695801,
    "timestamp": "2025-07-14T12:34:12.837768"
  },
  {
    "epoch": 1101,
    "batch_idx": 64,
    "loss": 0.2896365821361542,
    "learning_rate": 0.00042100713568680065,
    "timestep_mean": 483.59375,
    "epsilon_norm": 8.281627655029297,
    "prediction_norm": 6.48971700668335,
    "timestamp": "2025-07-14T12:34:12.842958"
  },
  {
    "epoch": 1101,
    "batch_idx": 96,
    "loss": 0.14911818504333496,
    "learning_rate": 0.00042100713568680065,
    "timestep_mean": 603.375,
    "epsilon_norm": 8.193492889404297,
    "prediction_norm": 7.614428997039795,
    "timestamp": "2025-07-14T12:34:12.848029"
  },
  {
    "epoch": 1101,
    "batch_idx": 128,
    "loss": 0.2921677529811859,
    "learning_rate": 0.00042100713568680065,
    "timestep_mean": 494.71875,
    "epsilon_norm": 7.958463191986084,
    "prediction_norm": 7.118075370788574,
    "timestamp": "2025-07-14T12:34:12.853111"
  },
  {
    "epoch": 1101,
    "batch_idx": 160,
    "loss": 0.29217445850372314,
    "learning_rate": 0.00042100713568680065,
    "timestep_mean": 489.25,
    "epsilon_norm": 9.504109382629395,
    "prediction_norm": 7.909474849700928,
    "timestamp": "2025-07-14T12:34:12.858394"
  },
  {
    "epoch": 1101,
    "batch_idx": 192,
    "loss": 0.8351941704750061,
    "learning_rate": 0.00042100713568680065,
    "timestep_mean": 429.625,
    "epsilon_norm": 6.621934413909912,
    "prediction_norm": 4.989312171936035,
    "timestamp": "2025-07-14T12:34:12.863568"
  },
  {
    "epoch": 1102,
    "batch_idx": 0,
    "loss": 0.26911062002182007,
    "learning_rate": 0.0004202316988007568,
    "timestep_mean": 494.34375,
    "epsilon_norm": 8.362285614013672,
    "prediction_norm": 7.239447593688965,
    "timestamp": "2025-07-14T12:34:12.868782"
  },
  {
    "epoch": 1102,
    "batch_idx": 32,
    "loss": 0.3812451958656311,
    "learning_rate": 0.0004202316988007568,
    "timestep_mean": 566.84375,
    "epsilon_norm": 7.767218112945557,
    "prediction_norm": 6.64382791519165,
    "timestamp": "2025-07-14T12:34:12.873957"
  },
  {
    "epoch": 1102,
    "batch_idx": 64,
    "loss": 0.26418089866638184,
    "learning_rate": 0.0004202316988007568,
    "timestep_mean": 471.53125,
    "epsilon_norm": 7.839022159576416,
    "prediction_norm": 7.1421051025390625,
    "timestamp": "2025-07-14T12:34:12.879169"
  },
  {
    "epoch": 1102,
    "batch_idx": 96,
    "loss": 0.3392384648323059,
    "learning_rate": 0.0004202316988007568,
    "timestep_mean": 477.59375,
    "epsilon_norm": 8.225102424621582,
    "prediction_norm": 6.932151794433594,
    "timestamp": "2025-07-14T12:34:12.884290"
  },
  {
    "epoch": 1102,
    "batch_idx": 128,
    "loss": 0.2761973440647125,
    "learning_rate": 0.0004202316988007568,
    "timestep_mean": 457.09375,
    "epsilon_norm": 8.528701782226562,
    "prediction_norm": 7.113518238067627,
    "timestamp": "2025-07-14T12:34:12.889285"
  },
  {
    "epoch": 1102,
    "batch_idx": 160,
    "loss": 0.35601723194122314,
    "learning_rate": 0.0004202316988007568,
    "timestep_mean": 468.28125,
    "epsilon_norm": 7.046733856201172,
    "prediction_norm": 5.772299289703369,
    "timestamp": "2025-07-14T12:34:12.894572"
  },
  {
    "epoch": 1102,
    "batch_idx": 192,
    "loss": 0.2214057594537735,
    "learning_rate": 0.0004202316988007568,
    "timestep_mean": 625.75,
    "epsilon_norm": 5.4542107582092285,
    "prediction_norm": 5.632060527801514,
    "timestamp": "2025-07-14T12:34:12.899798"
  },
  {
    "epoch": 1103,
    "batch_idx": 0,
    "loss": 0.09454210102558136,
    "learning_rate": 0.0004194564587350667,
    "timestep_mean": 705.03125,
    "epsilon_norm": 7.984974384307861,
    "prediction_norm": 7.747901439666748,
    "timestamp": "2025-07-14T12:34:12.904964"
  },
  {
    "epoch": 1103,
    "batch_idx": 32,
    "loss": 0.31886664032936096,
    "learning_rate": 0.0004194564587350667,
    "timestep_mean": 489.4375,
    "epsilon_norm": 7.832902908325195,
    "prediction_norm": 6.530397415161133,
    "timestamp": "2025-07-14T12:34:12.910182"
  },
  {
    "epoch": 1103,
    "batch_idx": 64,
    "loss": 0.2756029963493347,
    "learning_rate": 0.0004194564587350667,
    "timestep_mean": 473.96875,
    "epsilon_norm": 7.326766490936279,
    "prediction_norm": 5.8466644287109375,
    "timestamp": "2025-07-14T12:34:12.915380"
  },
  {
    "epoch": 1103,
    "batch_idx": 96,
    "loss": 0.2601168155670166,
    "learning_rate": 0.0004194564587350667,
    "timestep_mean": 338.25,
    "epsilon_norm": 7.548364639282227,
    "prediction_norm": 6.662510395050049,
    "timestamp": "2025-07-14T12:34:12.920534"
  },
  {
    "epoch": 1103,
    "batch_idx": 128,
    "loss": 0.26512619853019714,
    "learning_rate": 0.0004194564587350667,
    "timestep_mean": 566.625,
    "epsilon_norm": 8.903473854064941,
    "prediction_norm": 7.917328357696533,
    "timestamp": "2025-07-14T12:34:12.925697"
  },
  {
    "epoch": 1103,
    "batch_idx": 160,
    "loss": 0.31946054100990295,
    "learning_rate": 0.0004194564587350667,
    "timestep_mean": 578.875,
    "epsilon_norm": 7.720527648925781,
    "prediction_norm": 6.234258651733398,
    "timestamp": "2025-07-14T12:34:12.930611"
  },
  {
    "epoch": 1103,
    "batch_idx": 192,
    "loss": 0.16360919177532196,
    "learning_rate": 0.0004194564587350667,
    "timestep_mean": 612.125,
    "epsilon_norm": 3.385059118270874,
    "prediction_norm": 2.712005853652954,
    "timestamp": "2025-07-14T12:34:12.935726"
  },
  {
    "epoch": 1104,
    "batch_idx": 0,
    "loss": 0.4185274839401245,
    "learning_rate": 0.00041868141740255817,
    "timestep_mean": 415.03125,
    "epsilon_norm": 8.944880485534668,
    "prediction_norm": 6.827291965484619,
    "timestamp": "2025-07-14T12:34:12.940932"
  },
  {
    "epoch": 1104,
    "batch_idx": 32,
    "loss": 0.29030218720436096,
    "learning_rate": 0.00041868141740255817,
    "timestep_mean": 527.8125,
    "epsilon_norm": 9.563597679138184,
    "prediction_norm": 8.19785213470459,
    "timestamp": "2025-07-14T12:34:12.945864"
  },
  {
    "epoch": 1104,
    "batch_idx": 64,
    "loss": 0.5917381048202515,
    "learning_rate": 0.00041868141740255817,
    "timestep_mean": 425.40625,
    "epsilon_norm": 8.523445129394531,
    "prediction_norm": 6.306738376617432,
    "timestamp": "2025-07-14T12:34:12.950939"
  },
  {
    "epoch": 1104,
    "batch_idx": 96,
    "loss": 0.3612096309661865,
    "learning_rate": 0.00041868141740255817,
    "timestep_mean": 427.15625,
    "epsilon_norm": 8.532411575317383,
    "prediction_norm": 6.781691551208496,
    "timestamp": "2025-07-14T12:34:12.956129"
  },
  {
    "epoch": 1104,
    "batch_idx": 128,
    "loss": 0.3104317784309387,
    "learning_rate": 0.00041868141740255817,
    "timestep_mean": 441.59375,
    "epsilon_norm": 8.98377513885498,
    "prediction_norm": 7.759984016418457,
    "timestamp": "2025-07-14T12:34:12.961298"
  },
  {
    "epoch": 1104,
    "batch_idx": 160,
    "loss": 0.30999189615249634,
    "learning_rate": 0.00041868141740255817,
    "timestep_mean": 520.84375,
    "epsilon_norm": 8.215021133422852,
    "prediction_norm": 6.731810092926025,
    "timestamp": "2025-07-14T12:34:12.966426"
  },
  {
    "epoch": 1104,
    "batch_idx": 192,
    "loss": 0.07333371788263321,
    "learning_rate": 0.00041868141740255817,
    "timestep_mean": 502.375,
    "epsilon_norm": 3.1811423301696777,
    "prediction_norm": 3.4681155681610107,
    "timestamp": "2025-07-14T12:34:12.971624"
  },
  {
    "epoch": 1105,
    "batch_idx": 0,
    "loss": 0.21663165092468262,
    "learning_rate": 0.00041790657671556854,
    "timestep_mean": 512.28125,
    "epsilon_norm": 7.422912120819092,
    "prediction_norm": 5.936280727386475,
    "timestamp": "2025-07-14T12:34:12.976834"
  },
  {
    "epoch": 1105,
    "batch_idx": 32,
    "loss": 0.15147536993026733,
    "learning_rate": 0.00041790657671556854,
    "timestep_mean": 637.09375,
    "epsilon_norm": 8.584388732910156,
    "prediction_norm": 7.7836456298828125,
    "timestamp": "2025-07-14T12:34:12.981935"
  },
  {
    "epoch": 1105,
    "batch_idx": 64,
    "loss": 0.19920028746128082,
    "learning_rate": 0.00041790657671556854,
    "timestep_mean": 475.90625,
    "epsilon_norm": 8.000585556030273,
    "prediction_norm": 6.83557653427124,
    "timestamp": "2025-07-14T12:34:12.987026"
  },
  {
    "epoch": 1105,
    "batch_idx": 96,
    "loss": 0.4129202365875244,
    "learning_rate": 0.00041790657671556854,
    "timestep_mean": 387.0625,
    "epsilon_norm": 9.207054138183594,
    "prediction_norm": 7.386285781860352,
    "timestamp": "2025-07-14T12:34:12.992452"
  },
  {
    "epoch": 1105,
    "batch_idx": 128,
    "loss": 0.30773046612739563,
    "learning_rate": 0.00041790657671556854,
    "timestep_mean": 418.03125,
    "epsilon_norm": 7.157337665557861,
    "prediction_norm": 5.228623390197754,
    "timestamp": "2025-07-14T12:34:12.997631"
  },
  {
    "epoch": 1105,
    "batch_idx": 160,
    "loss": 0.22420695424079895,
    "learning_rate": 0.00041790657671556854,
    "timestep_mean": 530.9375,
    "epsilon_norm": 7.585333347320557,
    "prediction_norm": 6.516860008239746,
    "timestamp": "2025-07-14T12:34:13.002739"
  },
  {
    "epoch": 1105,
    "batch_idx": 192,
    "loss": 0.22453786432743073,
    "learning_rate": 0.00041790657671556854,
    "timestep_mean": 465.125,
    "epsilon_norm": 3.4822139739990234,
    "prediction_norm": 2.670828104019165,
    "timestamp": "2025-07-14T12:34:13.007920"
  },
  {
    "epoch": 1106,
    "batch_idx": 0,
    "loss": 0.3040386438369751,
    "learning_rate": 0.0004171319385859402,
    "timestep_mean": 462.25,
    "epsilon_norm": 7.52085542678833,
    "prediction_norm": 5.940090656280518,
    "timestamp": "2025-07-14T12:34:13.013189"
  },
  {
    "epoch": 1106,
    "batch_idx": 32,
    "loss": 0.34501686692237854,
    "learning_rate": 0.0004171319385859402,
    "timestep_mean": 488.71875,
    "epsilon_norm": 6.885254383087158,
    "prediction_norm": 5.430467128753662,
    "timestamp": "2025-07-14T12:34:13.018321"
  },
  {
    "epoch": 1106,
    "batch_idx": 64,
    "loss": 0.1646616905927658,
    "learning_rate": 0.0004171319385859402,
    "timestep_mean": 519.4375,
    "epsilon_norm": 6.744274616241455,
    "prediction_norm": 5.608330726623535,
    "timestamp": "2025-07-14T12:34:13.023429"
  },
  {
    "epoch": 1106,
    "batch_idx": 96,
    "loss": 0.09109827131032944,
    "learning_rate": 0.0004171319385859402,
    "timestep_mean": 586.0625,
    "epsilon_norm": 6.887803077697754,
    "prediction_norm": 6.764151096343994,
    "timestamp": "2025-07-14T12:34:13.028660"
  },
  {
    "epoch": 1106,
    "batch_idx": 128,
    "loss": 0.14535534381866455,
    "learning_rate": 0.0004171319385859402,
    "timestep_mean": 500.1875,
    "epsilon_norm": 6.89616060256958,
    "prediction_norm": 6.676918983459473,
    "timestamp": "2025-07-14T12:34:13.033862"
  },
  {
    "epoch": 1106,
    "batch_idx": 160,
    "loss": 0.1830650269985199,
    "learning_rate": 0.0004171319385859402,
    "timestep_mean": 449.0625,
    "epsilon_norm": 8.022714614868164,
    "prediction_norm": 7.293893337249756,
    "timestamp": "2025-07-14T12:34:13.038962"
  },
  {
    "epoch": 1106,
    "batch_idx": 192,
    "loss": 0.22153586149215698,
    "learning_rate": 0.0004171319385859402,
    "timestep_mean": 298.25,
    "epsilon_norm": 3.857285261154175,
    "prediction_norm": 2.658804416656494,
    "timestamp": "2025-07-14T12:34:13.043924"
  },
  {
    "epoch": 1107,
    "batch_idx": 0,
    "loss": 0.3837151527404785,
    "learning_rate": 0.00041635750492501566,
    "timestep_mean": 486.15625,
    "epsilon_norm": 7.983582973480225,
    "prediction_norm": 6.284395217895508,
    "timestamp": "2025-07-14T12:34:13.049128"
  },
  {
    "epoch": 1107,
    "batch_idx": 32,
    "loss": 0.282602995634079,
    "learning_rate": 0.00041635750492501566,
    "timestep_mean": 559.96875,
    "epsilon_norm": 8.389876365661621,
    "prediction_norm": 7.479345321655273,
    "timestamp": "2025-07-14T12:34:13.054326"
  },
  {
    "epoch": 1107,
    "batch_idx": 64,
    "loss": 0.24966830015182495,
    "learning_rate": 0.00041635750492501566,
    "timestep_mean": 623.375,
    "epsilon_norm": 8.167182922363281,
    "prediction_norm": 7.030229568481445,
    "timestamp": "2025-07-14T12:34:13.059437"
  },
  {
    "epoch": 1107,
    "batch_idx": 96,
    "loss": 0.3209369480609894,
    "learning_rate": 0.00041635750492501566,
    "timestep_mean": 526.0625,
    "epsilon_norm": 7.159633159637451,
    "prediction_norm": 6.016406536102295,
    "timestamp": "2025-07-14T12:34:13.064675"
  },
  {
    "epoch": 1107,
    "batch_idx": 128,
    "loss": 0.14559902250766754,
    "learning_rate": 0.00041635750492501566,
    "timestep_mean": 592.84375,
    "epsilon_norm": 8.519428253173828,
    "prediction_norm": 8.07542896270752,
    "timestamp": "2025-07-14T12:34:13.069826"
  },
  {
    "epoch": 1107,
    "batch_idx": 160,
    "loss": 0.4210081696510315,
    "learning_rate": 0.00041635750492501566,
    "timestep_mean": 431.96875,
    "epsilon_norm": 6.80134391784668,
    "prediction_norm": 5.187157154083252,
    "timestamp": "2025-07-14T12:34:13.074981"
  },
  {
    "epoch": 1107,
    "batch_idx": 192,
    "loss": 0.7276126146316528,
    "learning_rate": 0.00041635750492501566,
    "timestep_mean": 325.5,
    "epsilon_norm": 4.586757659912109,
    "prediction_norm": 3.2508959770202637,
    "timestamp": "2025-07-14T12:34:13.079824"
  },
  {
    "epoch": 1108,
    "batch_idx": 0,
    "loss": 0.39893996715545654,
    "learning_rate": 0.00041558327764363304,
    "timestep_mean": 425.3125,
    "epsilon_norm": 9.255447387695312,
    "prediction_norm": 7.975919723510742,
    "timestamp": "2025-07-14T12:34:13.084602"
  },
  {
    "epoch": 1108,
    "batch_idx": 32,
    "loss": 0.32191309332847595,
    "learning_rate": 0.00041558327764363304,
    "timestep_mean": 473.375,
    "epsilon_norm": 7.583720684051514,
    "prediction_norm": 5.71597957611084,
    "timestamp": "2025-07-14T12:34:13.089413"
  },
  {
    "epoch": 1108,
    "batch_idx": 64,
    "loss": 0.16988325119018555,
    "learning_rate": 0.00041558327764363304,
    "timestep_mean": 509.5625,
    "epsilon_norm": 8.222175598144531,
    "prediction_norm": 8.392678260803223,
    "timestamp": "2025-07-14T12:34:13.094279"
  },
  {
    "epoch": 1108,
    "batch_idx": 96,
    "loss": 0.3260018825531006,
    "learning_rate": 0.00041558327764363304,
    "timestep_mean": 373.9375,
    "epsilon_norm": 8.293462753295898,
    "prediction_norm": 6.197990417480469,
    "timestamp": "2025-07-14T12:34:13.099168"
  },
  {
    "epoch": 1108,
    "batch_idx": 128,
    "loss": 0.4554564654827118,
    "learning_rate": 0.00041558327764363304,
    "timestep_mean": 452.15625,
    "epsilon_norm": 8.628096580505371,
    "prediction_norm": 6.554165363311768,
    "timestamp": "2025-07-14T12:34:13.104007"
  },
  {
    "epoch": 1108,
    "batch_idx": 160,
    "loss": 0.3118698000907898,
    "learning_rate": 0.00041558327764363304,
    "timestep_mean": 486.40625,
    "epsilon_norm": 8.293135643005371,
    "prediction_norm": 7.255490303039551,
    "timestamp": "2025-07-14T12:34:13.108942"
  },
  {
    "epoch": 1108,
    "batch_idx": 192,
    "loss": 0.17649805545806885,
    "learning_rate": 0.00041558327764363304,
    "timestep_mean": 363.75,
    "epsilon_norm": 4.563381671905518,
    "prediction_norm": 4.1278252601623535,
    "timestamp": "2025-07-14T12:34:13.113918"
  },
  {
    "epoch": 1109,
    "batch_idx": 0,
    "loss": 0.2635350823402405,
    "learning_rate": 0.0004148092586521213,
    "timestep_mean": 546.71875,
    "epsilon_norm": 7.706454277038574,
    "prediction_norm": 6.5229597091674805,
    "timestamp": "2025-07-14T12:34:13.118799"
  },
  {
    "epoch": 1109,
    "batch_idx": 32,
    "loss": 0.2636868357658386,
    "learning_rate": 0.0004148092586521213,
    "timestep_mean": 475.625,
    "epsilon_norm": 7.92846155166626,
    "prediction_norm": 6.393747806549072,
    "timestamp": "2025-07-14T12:34:13.123769"
  },
  {
    "epoch": 1109,
    "batch_idx": 64,
    "loss": 0.37711620330810547,
    "learning_rate": 0.0004148092586521213,
    "timestep_mean": 451.875,
    "epsilon_norm": 8.628235816955566,
    "prediction_norm": 6.989344596862793,
    "timestamp": "2025-07-14T12:34:13.128627"
  },
  {
    "epoch": 1109,
    "batch_idx": 96,
    "loss": 0.2584848999977112,
    "learning_rate": 0.0004148092586521213,
    "timestep_mean": 535.40625,
    "epsilon_norm": 8.214001655578613,
    "prediction_norm": 7.123476505279541,
    "timestamp": "2025-07-14T12:34:13.133541"
  },
  {
    "epoch": 1109,
    "batch_idx": 128,
    "loss": 0.23094698786735535,
    "learning_rate": 0.0004148092586521213,
    "timestep_mean": 519.71875,
    "epsilon_norm": 6.986611843109131,
    "prediction_norm": 6.21336030960083,
    "timestamp": "2025-07-14T12:34:13.138316"
  },
  {
    "epoch": 1109,
    "batch_idx": 160,
    "loss": 0.22969049215316772,
    "learning_rate": 0.0004148092586521213,
    "timestep_mean": 515.84375,
    "epsilon_norm": 9.018087387084961,
    "prediction_norm": 8.012335777282715,
    "timestamp": "2025-07-14T12:34:13.143159"
  },
  {
    "epoch": 1109,
    "batch_idx": 192,
    "loss": 0.04953343793749809,
    "learning_rate": 0.0004148092586521213,
    "timestep_mean": 575.625,
    "epsilon_norm": 3.0103039741516113,
    "prediction_norm": 3.2260265350341797,
    "timestamp": "2025-07-14T12:34:13.147975"
  },
  {
    "epoch": 1110,
    "batch_idx": 0,
    "loss": 0.22467297315597534,
    "learning_rate": 0.00041403544986029516,
    "timestep_mean": 453.71875,
    "epsilon_norm": 8.513228416442871,
    "prediction_norm": 7.161489486694336,
    "timestamp": "2025-07-14T12:34:13.152801"
  },
  {
    "epoch": 1110,
    "batch_idx": 32,
    "loss": 0.2728850245475769,
    "learning_rate": 0.00041403544986029516,
    "timestep_mean": 609.4375,
    "epsilon_norm": 7.300539970397949,
    "prediction_norm": 6.5389485359191895,
    "timestamp": "2025-07-14T12:34:13.157708"
  },
  {
    "epoch": 1110,
    "batch_idx": 64,
    "loss": 0.2688480317592621,
    "learning_rate": 0.00041403544986029516,
    "timestep_mean": 530.8125,
    "epsilon_norm": 10.598891258239746,
    "prediction_norm": 9.561171531677246,
    "timestamp": "2025-07-14T12:34:13.162648"
  },
  {
    "epoch": 1110,
    "batch_idx": 96,
    "loss": 0.16772012412548065,
    "learning_rate": 0.00041403544986029516,
    "timestep_mean": 413.625,
    "epsilon_norm": 8.217677116394043,
    "prediction_norm": 7.6575469970703125,
    "timestamp": "2025-07-14T12:34:13.167461"
  },
  {
    "epoch": 1110,
    "batch_idx": 128,
    "loss": 0.5746026635169983,
    "learning_rate": 0.00041403544986029516,
    "timestep_mean": 437.4375,
    "epsilon_norm": 9.789358139038086,
    "prediction_norm": 7.124271869659424,
    "timestamp": "2025-07-14T12:34:13.172357"
  },
  {
    "epoch": 1110,
    "batch_idx": 160,
    "loss": 0.20763897895812988,
    "learning_rate": 0.00041403544986029516,
    "timestep_mean": 554.25,
    "epsilon_norm": 8.458714485168457,
    "prediction_norm": 7.9556884765625,
    "timestamp": "2025-07-14T12:34:13.177170"
  },
  {
    "epoch": 1110,
    "batch_idx": 192,
    "loss": 0.5107272267341614,
    "learning_rate": 0.00041403544986029516,
    "timestep_mean": 373.875,
    "epsilon_norm": 4.2796478271484375,
    "prediction_norm": 3.269852638244629,
    "timestamp": "2025-07-14T12:34:13.181987"
  },
  {
    "epoch": 1111,
    "batch_idx": 0,
    "loss": 0.11597806215286255,
    "learning_rate": 0.00041326185317745116,
    "timestep_mean": 559.03125,
    "epsilon_norm": 8.085991859436035,
    "prediction_norm": 7.456305027008057,
    "timestamp": "2025-07-14T12:34:13.186823"
  },
  {
    "epoch": 1111,
    "batch_idx": 32,
    "loss": 0.6372205018997192,
    "learning_rate": 0.00041326185317745116,
    "timestep_mean": 520.75,
    "epsilon_norm": 8.35302448272705,
    "prediction_norm": 5.9713454246521,
    "timestamp": "2025-07-14T12:34:13.191829"
  },
  {
    "epoch": 1111,
    "batch_idx": 64,
    "loss": 0.2152741402387619,
    "learning_rate": 0.00041326185317745116,
    "timestep_mean": 592.0625,
    "epsilon_norm": 7.27547025680542,
    "prediction_norm": 6.002944469451904,
    "timestamp": "2025-07-14T12:34:13.196667"
  },
  {
    "epoch": 1111,
    "batch_idx": 96,
    "loss": 0.30899930000305176,
    "learning_rate": 0.00041326185317745116,
    "timestep_mean": 523.8125,
    "epsilon_norm": 8.348444938659668,
    "prediction_norm": 7.440029144287109,
    "timestamp": "2025-07-14T12:34:13.201577"
  },
  {
    "epoch": 1111,
    "batch_idx": 128,
    "loss": 0.19393329322338104,
    "learning_rate": 0.00041326185317745116,
    "timestep_mean": 567.25,
    "epsilon_norm": 7.819697856903076,
    "prediction_norm": 7.227095603942871,
    "timestamp": "2025-07-14T12:34:13.206396"
  },
  {
    "epoch": 1111,
    "batch_idx": 160,
    "loss": 0.26205331087112427,
    "learning_rate": 0.00041326185317745116,
    "timestep_mean": 487.375,
    "epsilon_norm": 7.219526290893555,
    "prediction_norm": 5.850332736968994,
    "timestamp": "2025-07-14T12:34:13.211251"
  },
  {
    "epoch": 1111,
    "batch_idx": 192,
    "loss": 0.24976034462451935,
    "learning_rate": 0.00041326185317745116,
    "timestep_mean": 457.875,
    "epsilon_norm": 5.195124626159668,
    "prediction_norm": 4.321709156036377,
    "timestamp": "2025-07-14T12:34:13.216043"
  },
  {
    "epoch": 1112,
    "batch_idx": 0,
    "loss": 0.2784643769264221,
    "learning_rate": 0.000412488470512362,
    "timestep_mean": 466.0,
    "epsilon_norm": 5.959760665893555,
    "prediction_norm": 5.551410675048828,
    "timestamp": "2025-07-14T12:34:13.220989"
  },
  {
    "epoch": 1112,
    "batch_idx": 32,
    "loss": 0.26663124561309814,
    "learning_rate": 0.000412488470512362,
    "timestep_mean": 532.09375,
    "epsilon_norm": 7.878640651702881,
    "prediction_norm": 6.390469074249268,
    "timestamp": "2025-07-14T12:34:13.225797"
  },
  {
    "epoch": 1112,
    "batch_idx": 64,
    "loss": 0.46921706199645996,
    "learning_rate": 0.000412488470512362,
    "timestep_mean": 475.65625,
    "epsilon_norm": 7.37638521194458,
    "prediction_norm": 5.696900844573975,
    "timestamp": "2025-07-14T12:34:13.230669"
  },
  {
    "epoch": 1112,
    "batch_idx": 96,
    "loss": 0.24195584654808044,
    "learning_rate": 0.000412488470512362,
    "timestep_mean": 511.625,
    "epsilon_norm": 8.077019691467285,
    "prediction_norm": 6.962554931640625,
    "timestamp": "2025-07-14T12:34:13.235483"
  },
  {
    "epoch": 1112,
    "batch_idx": 128,
    "loss": 0.27144166827201843,
    "learning_rate": 0.000412488470512362,
    "timestep_mean": 481.25,
    "epsilon_norm": 7.725868225097656,
    "prediction_norm": 6.571324348449707,
    "timestamp": "2025-07-14T12:34:13.240344"
  },
  {
    "epoch": 1112,
    "batch_idx": 160,
    "loss": 0.25380340218544006,
    "learning_rate": 0.000412488470512362,
    "timestep_mean": 411.34375,
    "epsilon_norm": 7.243445873260498,
    "prediction_norm": 5.792608737945557,
    "timestamp": "2025-07-14T12:34:13.245232"
  },
  {
    "epoch": 1112,
    "batch_idx": 192,
    "loss": 0.060289740562438965,
    "learning_rate": 0.000412488470512362,
    "timestep_mean": 595.375,
    "epsilon_norm": 2.7386481761932373,
    "prediction_norm": 2.0346012115478516,
    "timestamp": "2025-07-14T12:34:13.250121"
  },
  {
    "epoch": 1113,
    "batch_idx": 0,
    "loss": 0.21737557649612427,
    "learning_rate": 0.00041171530377327265,
    "timestep_mean": 481.53125,
    "epsilon_norm": 7.965470790863037,
    "prediction_norm": 7.076609134674072,
    "timestamp": "2025-07-14T12:34:13.254966"
  },
  {
    "epoch": 1113,
    "batch_idx": 32,
    "loss": 0.3340449929237366,
    "learning_rate": 0.00041171530377327265,
    "timestep_mean": 477.9375,
    "epsilon_norm": 7.730219841003418,
    "prediction_norm": 6.005492687225342,
    "timestamp": "2025-07-14T12:34:13.259770"
  },
  {
    "epoch": 1113,
    "batch_idx": 64,
    "loss": 0.5201087594032288,
    "learning_rate": 0.00041171530377327265,
    "timestep_mean": 409.9375,
    "epsilon_norm": 7.822632312774658,
    "prediction_norm": 6.114642143249512,
    "timestamp": "2025-07-14T12:34:13.264617"
  },
  {
    "epoch": 1113,
    "batch_idx": 96,
    "loss": 0.2614782452583313,
    "learning_rate": 0.00041171530377327265,
    "timestep_mean": 437.09375,
    "epsilon_norm": 7.857209205627441,
    "prediction_norm": 6.769489765167236,
    "timestamp": "2025-07-14T12:34:13.269512"
  },
  {
    "epoch": 1113,
    "batch_idx": 128,
    "loss": 0.2886369228363037,
    "learning_rate": 0.00041171530377327265,
    "timestep_mean": 454.21875,
    "epsilon_norm": 7.602715969085693,
    "prediction_norm": 6.296361923217773,
    "timestamp": "2025-07-14T12:34:13.274378"
  },
  {
    "epoch": 1113,
    "batch_idx": 160,
    "loss": 0.12942886352539062,
    "learning_rate": 0.00041171530377327265,
    "timestep_mean": 576.09375,
    "epsilon_norm": 8.242621421813965,
    "prediction_norm": 7.635998725891113,
    "timestamp": "2025-07-14T12:34:13.279300"
  },
  {
    "epoch": 1113,
    "batch_idx": 192,
    "loss": 0.27294567227363586,
    "learning_rate": 0.00041171530377327265,
    "timestep_mean": 539.875,
    "epsilon_norm": 3.2068538665771484,
    "prediction_norm": 2.9595370292663574,
    "timestamp": "2025-07-14T12:34:13.284105"
  },
  {
    "epoch": 1114,
    "batch_idx": 0,
    "loss": 0.26258522272109985,
    "learning_rate": 0.00041094235486789486,
    "timestep_mean": 532.15625,
    "epsilon_norm": 7.202775001525879,
    "prediction_norm": 6.029356956481934,
    "timestamp": "2025-07-14T12:34:13.288977"
  },
  {
    "epoch": 1114,
    "batch_idx": 32,
    "loss": 0.13668276369571686,
    "learning_rate": 0.00041094235486789486,
    "timestep_mean": 524.28125,
    "epsilon_norm": 7.843742847442627,
    "prediction_norm": 7.137645721435547,
    "timestamp": "2025-07-14T12:34:13.293895"
  },
  {
    "epoch": 1114,
    "batch_idx": 64,
    "loss": 0.17749661207199097,
    "learning_rate": 0.00041094235486789486,
    "timestep_mean": 478.625,
    "epsilon_norm": 7.255731582641602,
    "prediction_norm": 7.230351448059082,
    "timestamp": "2025-07-14T12:34:13.298739"
  },
  {
    "epoch": 1114,
    "batch_idx": 96,
    "loss": 0.1242060661315918,
    "learning_rate": 0.00041094235486789486,
    "timestep_mean": 590.0625,
    "epsilon_norm": 7.25427770614624,
    "prediction_norm": 6.684548377990723,
    "timestamp": "2025-07-14T12:34:13.303703"
  },
  {
    "epoch": 1114,
    "batch_idx": 128,
    "loss": 0.35883861780166626,
    "learning_rate": 0.00041094235486789486,
    "timestep_mean": 401.6875,
    "epsilon_norm": 7.995978355407715,
    "prediction_norm": 6.396709442138672,
    "timestamp": "2025-07-14T12:34:13.308892"
  },
  {
    "epoch": 1114,
    "batch_idx": 160,
    "loss": 0.30536308884620667,
    "learning_rate": 0.00041094235486789486,
    "timestep_mean": 541.9375,
    "epsilon_norm": 8.117396354675293,
    "prediction_norm": 7.377575397491455,
    "timestamp": "2025-07-14T12:34:13.314052"
  },
  {
    "epoch": 1114,
    "batch_idx": 192,
    "loss": 0.15657824277877808,
    "learning_rate": 0.00041094235486789486,
    "timestep_mean": 456.75,
    "epsilon_norm": 3.067445993423462,
    "prediction_norm": 2.6136085987091064,
    "timestamp": "2025-07-14T12:34:13.319176"
  },
  {
    "epoch": 1115,
    "batch_idx": 0,
    "loss": 0.4929831624031067,
    "learning_rate": 0.0004101696257034037,
    "timestep_mean": 462.1875,
    "epsilon_norm": 8.364434242248535,
    "prediction_norm": 5.765648365020752,
    "timestamp": "2025-07-14T12:34:13.324270"
  },
  {
    "epoch": 1115,
    "batch_idx": 32,
    "loss": 0.33605721592903137,
    "learning_rate": 0.0004101696257034037,
    "timestep_mean": 545.84375,
    "epsilon_norm": 6.928840637207031,
    "prediction_norm": 5.966553211212158,
    "timestamp": "2025-07-14T12:34:13.329309"
  },
  {
    "epoch": 1115,
    "batch_idx": 64,
    "loss": 0.2883074879646301,
    "learning_rate": 0.0004101696257034037,
    "timestep_mean": 419.71875,
    "epsilon_norm": 8.294028282165527,
    "prediction_norm": 6.474546909332275,
    "timestamp": "2025-07-14T12:34:13.334438"
  },
  {
    "epoch": 1115,
    "batch_idx": 96,
    "loss": 0.17571017146110535,
    "learning_rate": 0.0004101696257034037,
    "timestep_mean": 488.1875,
    "epsilon_norm": 7.835040092468262,
    "prediction_norm": 7.033279895782471,
    "timestamp": "2025-07-14T12:34:13.339504"
  },
  {
    "epoch": 1115,
    "batch_idx": 128,
    "loss": 0.3546069860458374,
    "learning_rate": 0.0004101696257034037,
    "timestep_mean": 500.21875,
    "epsilon_norm": 7.675952911376953,
    "prediction_norm": 6.006043434143066,
    "timestamp": "2025-07-14T12:34:13.344595"
  },
  {
    "epoch": 1115,
    "batch_idx": 160,
    "loss": 0.2336253672838211,
    "learning_rate": 0.0004101696257034037,
    "timestep_mean": 431.03125,
    "epsilon_norm": 8.405651092529297,
    "prediction_norm": 7.141352653503418,
    "timestamp": "2025-07-14T12:34:13.350018"
  },
  {
    "epoch": 1115,
    "batch_idx": 192,
    "loss": 0.359785258769989,
    "learning_rate": 0.0004101696257034037,
    "timestep_mean": 338.875,
    "epsilon_norm": 4.018863201141357,
    "prediction_norm": 2.7721362113952637,
    "timestamp": "2025-07-14T12:34:13.355145"
  },
  {
    "epoch": 1116,
    "batch_idx": 0,
    "loss": 0.1688581109046936,
    "learning_rate": 0.00040939711818643133,
    "timestep_mean": 639.1875,
    "epsilon_norm": 7.92933988571167,
    "prediction_norm": 7.214949607849121,
    "timestamp": "2025-07-14T12:34:13.360253"
  },
  {
    "epoch": 1116,
    "batch_idx": 32,
    "loss": 0.2915736436843872,
    "learning_rate": 0.00040939711818643133,
    "timestep_mean": 430.53125,
    "epsilon_norm": 7.5657196044921875,
    "prediction_norm": 6.188723564147949,
    "timestamp": "2025-07-14T12:34:13.365446"
  },
  {
    "epoch": 1116,
    "batch_idx": 64,
    "loss": 0.3778429627418518,
    "learning_rate": 0.00040939711818643133,
    "timestep_mean": 448.65625,
    "epsilon_norm": 8.490432739257812,
    "prediction_norm": 6.765361309051514,
    "timestamp": "2025-07-14T12:34:13.370455"
  },
  {
    "epoch": 1116,
    "batch_idx": 96,
    "loss": 0.1915510892868042,
    "learning_rate": 0.00040939711818643133,
    "timestep_mean": 593.375,
    "epsilon_norm": 8.094888687133789,
    "prediction_norm": 7.483776569366455,
    "timestamp": "2025-07-14T12:34:13.375729"
  },
  {
    "epoch": 1116,
    "batch_idx": 128,
    "loss": 0.2167922854423523,
    "learning_rate": 0.00040939711818643133,
    "timestep_mean": 510.34375,
    "epsilon_norm": 6.590308666229248,
    "prediction_norm": 6.263187408447266,
    "timestamp": "2025-07-14T12:34:13.380582"
  },
  {
    "epoch": 1116,
    "batch_idx": 160,
    "loss": 0.3852282166481018,
    "learning_rate": 0.00040939711818643133,
    "timestep_mean": 349.59375,
    "epsilon_norm": 7.494239330291748,
    "prediction_norm": 5.704971790313721,
    "timestamp": "2025-07-14T12:34:13.385448"
  },
  {
    "epoch": 1116,
    "batch_idx": 192,
    "loss": 0.4336911141872406,
    "learning_rate": 0.00040939711818643133,
    "timestep_mean": 333.75,
    "epsilon_norm": 4.235003471374512,
    "prediction_norm": 2.9440276622772217,
    "timestamp": "2025-07-14T12:34:13.390270"
  },
  {
    "epoch": 1117,
    "batch_idx": 0,
    "loss": 0.2857459783554077,
    "learning_rate": 0.0004086248342230632,
    "timestep_mean": 510.75,
    "epsilon_norm": 8.750680923461914,
    "prediction_norm": 7.372492790222168,
    "timestamp": "2025-07-14T12:34:13.395116"
  },
  {
    "epoch": 1117,
    "batch_idx": 32,
    "loss": 0.26560530066490173,
    "learning_rate": 0.0004086248342230632,
    "timestep_mean": 523.9375,
    "epsilon_norm": 8.909106254577637,
    "prediction_norm": 7.560873031616211,
    "timestamp": "2025-07-14T12:34:13.400026"
  },
  {
    "epoch": 1117,
    "batch_idx": 64,
    "loss": 0.353834867477417,
    "learning_rate": 0.0004086248342230632,
    "timestep_mean": 399.96875,
    "epsilon_norm": 8.42225456237793,
    "prediction_norm": 6.538743495941162,
    "timestamp": "2025-07-14T12:34:13.404833"
  },
  {
    "epoch": 1117,
    "batch_idx": 96,
    "loss": 0.40837758779525757,
    "learning_rate": 0.0004086248342230632,
    "timestep_mean": 471.375,
    "epsilon_norm": 8.482177734375,
    "prediction_norm": 7.208946228027344,
    "timestamp": "2025-07-14T12:34:13.409767"
  },
  {
    "epoch": 1117,
    "batch_idx": 128,
    "loss": 0.2079959213733673,
    "learning_rate": 0.0004086248342230632,
    "timestep_mean": 512.96875,
    "epsilon_norm": 7.281966209411621,
    "prediction_norm": 6.330340385437012,
    "timestamp": "2025-07-14T12:34:13.414711"
  },
  {
    "epoch": 1117,
    "batch_idx": 160,
    "loss": 0.2936862111091614,
    "learning_rate": 0.0004086248342230632,
    "timestep_mean": 463.46875,
    "epsilon_norm": 6.817759990692139,
    "prediction_norm": 5.999587059020996,
    "timestamp": "2025-07-14T12:34:13.419777"
  },
  {
    "epoch": 1117,
    "batch_idx": 192,
    "loss": 0.47845658659935,
    "learning_rate": 0.0004086248342230632,
    "timestep_mean": 512.125,
    "epsilon_norm": 3.9214296340942383,
    "prediction_norm": 3.4809396266937256,
    "timestamp": "2025-07-14T12:34:13.424782"
  },
  {
    "epoch": 1118,
    "batch_idx": 0,
    "loss": 0.31474602222442627,
    "learning_rate": 0.0004078527757188333,
    "timestep_mean": 478.15625,
    "epsilon_norm": 7.902145862579346,
    "prediction_norm": 6.660304546356201,
    "timestamp": "2025-07-14T12:34:13.429872"
  },
  {
    "epoch": 1118,
    "batch_idx": 32,
    "loss": 0.19535723328590393,
    "learning_rate": 0.0004078527757188333,
    "timestep_mean": 526.28125,
    "epsilon_norm": 8.602797508239746,
    "prediction_norm": 7.771389007568359,
    "timestamp": "2025-07-14T12:34:13.434681"
  },
  {
    "epoch": 1118,
    "batch_idx": 64,
    "loss": 0.4331321716308594,
    "learning_rate": 0.0004078527757188333,
    "timestep_mean": 426.65625,
    "epsilon_norm": 8.363748550415039,
    "prediction_norm": 6.509155750274658,
    "timestamp": "2025-07-14T12:34:13.439745"
  },
  {
    "epoch": 1118,
    "batch_idx": 96,
    "loss": 0.21374492347240448,
    "learning_rate": 0.0004078527757188333,
    "timestep_mean": 531.5,
    "epsilon_norm": 6.719318866729736,
    "prediction_norm": 5.826870441436768,
    "timestamp": "2025-07-14T12:34:13.445015"
  },
  {
    "epoch": 1118,
    "batch_idx": 128,
    "loss": 0.24252428114414215,
    "learning_rate": 0.0004078527757188333,
    "timestep_mean": 476.40625,
    "epsilon_norm": 7.845480442047119,
    "prediction_norm": 6.531739234924316,
    "timestamp": "2025-07-14T12:34:13.450384"
  },
  {
    "epoch": 1118,
    "batch_idx": 160,
    "loss": 0.17065906524658203,
    "learning_rate": 0.0004078527757188333,
    "timestep_mean": 459.125,
    "epsilon_norm": 7.112170219421387,
    "prediction_norm": 6.539687156677246,
    "timestamp": "2025-07-14T12:34:13.455549"
  },
  {
    "epoch": 1118,
    "batch_idx": 192,
    "loss": 0.30831050872802734,
    "learning_rate": 0.0004078527757188333,
    "timestep_mean": 437.75,
    "epsilon_norm": 4.152237892150879,
    "prediction_norm": 2.63957142829895,
    "timestamp": "2025-07-14T12:34:13.460558"
  },
  {
    "epoch": 1119,
    "batch_idx": 0,
    "loss": 0.4363883435726166,
    "learning_rate": 0.00040708094457871926,
    "timestep_mean": 479.53125,
    "epsilon_norm": 6.801757335662842,
    "prediction_norm": 5.480189800262451,
    "timestamp": "2025-07-14T12:34:13.465485"
  },
  {
    "epoch": 1119,
    "batch_idx": 32,
    "loss": 0.1522429883480072,
    "learning_rate": 0.00040708094457871926,
    "timestep_mean": 485.96875,
    "epsilon_norm": 8.713079452514648,
    "prediction_norm": 7.891262054443359,
    "timestamp": "2025-07-14T12:34:13.470352"
  },
  {
    "epoch": 1119,
    "batch_idx": 64,
    "loss": 0.2974875867366791,
    "learning_rate": 0.00040708094457871926,
    "timestep_mean": 549.71875,
    "epsilon_norm": 7.4965715408325195,
    "prediction_norm": 6.218648910522461,
    "timestamp": "2025-07-14T12:34:13.475404"
  },
  {
    "epoch": 1119,
    "batch_idx": 96,
    "loss": 0.23150986433029175,
    "learning_rate": 0.00040708094457871926,
    "timestep_mean": 513.71875,
    "epsilon_norm": 7.242194175720215,
    "prediction_norm": 6.156367301940918,
    "timestamp": "2025-07-14T12:34:13.480307"
  },
  {
    "epoch": 1119,
    "batch_idx": 128,
    "loss": 0.22666268050670624,
    "learning_rate": 0.00040708094457871926,
    "timestep_mean": 551.90625,
    "epsilon_norm": 7.491913318634033,
    "prediction_norm": 6.42128849029541,
    "timestamp": "2025-07-14T12:34:13.485201"
  },
  {
    "epoch": 1119,
    "batch_idx": 160,
    "loss": 0.26158422231674194,
    "learning_rate": 0.00040708094457871926,
    "timestep_mean": 503.0625,
    "epsilon_norm": 7.150562763214111,
    "prediction_norm": 6.043100833892822,
    "timestamp": "2025-07-14T12:34:13.490284"
  },
  {
    "epoch": 1119,
    "batch_idx": 192,
    "loss": 0.044105980545282364,
    "learning_rate": 0.00040708094457871926,
    "timestep_mean": 579.125,
    "epsilon_norm": 3.907428026199341,
    "prediction_norm": 3.8974721431732178,
    "timestamp": "2025-07-14T12:34:13.495244"
  },
  {
    "epoch": 1120,
    "batch_idx": 0,
    "loss": 0.1732870638370514,
    "learning_rate": 0.0004063093427071376,
    "timestep_mean": 490.71875,
    "epsilon_norm": 8.055974006652832,
    "prediction_norm": 7.388424396514893,
    "timestamp": "2025-07-14T12:34:13.500222"
  },
  {
    "epoch": 1120,
    "batch_idx": 32,
    "loss": 0.1868181824684143,
    "learning_rate": 0.0004063093427071376,
    "timestep_mean": 495.59375,
    "epsilon_norm": 7.905566692352295,
    "prediction_norm": 6.635464668273926,
    "timestamp": "2025-07-14T12:34:13.505268"
  },
  {
    "epoch": 1120,
    "batch_idx": 64,
    "loss": 0.26435577869415283,
    "learning_rate": 0.0004063093427071376,
    "timestep_mean": 464.40625,
    "epsilon_norm": 7.487742900848389,
    "prediction_norm": 5.713794231414795,
    "timestamp": "2025-07-14T12:34:13.510235"
  },
  {
    "epoch": 1120,
    "batch_idx": 96,
    "loss": 0.260948121547699,
    "learning_rate": 0.0004063093427071376,
    "timestep_mean": 536.875,
    "epsilon_norm": 8.895604133605957,
    "prediction_norm": 7.582793235778809,
    "timestamp": "2025-07-14T12:34:13.515145"
  },
  {
    "epoch": 1120,
    "batch_idx": 128,
    "loss": 0.31752148270606995,
    "learning_rate": 0.0004063093427071376,
    "timestep_mean": 404.53125,
    "epsilon_norm": 7.930152893066406,
    "prediction_norm": 6.147251129150391,
    "timestamp": "2025-07-14T12:34:13.520203"
  },
  {
    "epoch": 1120,
    "batch_idx": 160,
    "loss": 0.29714471101760864,
    "learning_rate": 0.0004063093427071376,
    "timestep_mean": 421.4375,
    "epsilon_norm": 9.029735565185547,
    "prediction_norm": 7.5122222900390625,
    "timestamp": "2025-07-14T12:34:13.525227"
  },
  {
    "epoch": 1120,
    "batch_idx": 192,
    "loss": 0.20387530326843262,
    "learning_rate": 0.0004063093427071376,
    "timestep_mean": 233.625,
    "epsilon_norm": 4.194405555725098,
    "prediction_norm": 2.9826080799102783,
    "timestamp": "2025-07-14T12:34:13.530115"
  },
  {
    "epoch": 1121,
    "batch_idx": 0,
    "loss": 0.20966540277004242,
    "learning_rate": 0.00040553797200793934,
    "timestep_mean": 555.375,
    "epsilon_norm": 8.62295150756836,
    "prediction_norm": 7.772849082946777,
    "timestamp": "2025-07-14T12:34:13.535026"
  },
  {
    "epoch": 1121,
    "batch_idx": 32,
    "loss": 0.2600848078727722,
    "learning_rate": 0.00040553797200793934,
    "timestep_mean": 537.96875,
    "epsilon_norm": 7.589902400970459,
    "prediction_norm": 6.822779178619385,
    "timestamp": "2025-07-14T12:34:13.539961"
  },
  {
    "epoch": 1121,
    "batch_idx": 64,
    "loss": 0.08675511926412582,
    "learning_rate": 0.00040553797200793934,
    "timestep_mean": 520.8125,
    "epsilon_norm": 7.89076042175293,
    "prediction_norm": 7.507059574127197,
    "timestamp": "2025-07-14T12:34:13.544908"
  },
  {
    "epoch": 1121,
    "batch_idx": 96,
    "loss": 0.3726158142089844,
    "learning_rate": 0.00040553797200793934,
    "timestep_mean": 463.65625,
    "epsilon_norm": 7.7060112953186035,
    "prediction_norm": 6.181562423706055,
    "timestamp": "2025-07-14T12:34:13.549809"
  },
  {
    "epoch": 1121,
    "batch_idx": 128,
    "loss": 0.25642654299736023,
    "learning_rate": 0.00040553797200793934,
    "timestep_mean": 516.1875,
    "epsilon_norm": 7.0931782722473145,
    "prediction_norm": 5.869495391845703,
    "timestamp": "2025-07-14T12:34:13.554703"
  },
  {
    "epoch": 1121,
    "batch_idx": 160,
    "loss": 0.23050548136234283,
    "learning_rate": 0.00040553797200793934,
    "timestep_mean": 500.875,
    "epsilon_norm": 7.767263412475586,
    "prediction_norm": 6.505862236022949,
    "timestamp": "2025-07-14T12:34:13.559592"
  },
  {
    "epoch": 1121,
    "batch_idx": 192,
    "loss": 0.170598566532135,
    "learning_rate": 0.00040553797200793934,
    "timestep_mean": 541.0,
    "epsilon_norm": 3.6549675464630127,
    "prediction_norm": 3.2717437744140625,
    "timestamp": "2025-07-14T12:34:13.564622"
  },
  {
    "epoch": 1122,
    "batch_idx": 0,
    "loss": 0.3382716774940491,
    "learning_rate": 0.00040476683438440503,
    "timestep_mean": 465.3125,
    "epsilon_norm": 8.801438331604004,
    "prediction_norm": 7.3797502517700195,
    "timestamp": "2025-07-14T12:34:13.569502"
  },
  {
    "epoch": 1122,
    "batch_idx": 32,
    "loss": 0.29258137941360474,
    "learning_rate": 0.00040476683438440503,
    "timestep_mean": 493.09375,
    "epsilon_norm": 7.21938943862915,
    "prediction_norm": 5.927923679351807,
    "timestamp": "2025-07-14T12:34:13.574363"
  },
  {
    "epoch": 1122,
    "batch_idx": 64,
    "loss": 0.08896258473396301,
    "learning_rate": 0.00040476683438440503,
    "timestep_mean": 510.53125,
    "epsilon_norm": 7.1945037841796875,
    "prediction_norm": 6.795019149780273,
    "timestamp": "2025-07-14T12:34:13.579288"
  },
  {
    "epoch": 1122,
    "batch_idx": 96,
    "loss": 0.2776571810245514,
    "learning_rate": 0.00040476683438440503,
    "timestep_mean": 494.90625,
    "epsilon_norm": 7.275333881378174,
    "prediction_norm": 6.061589241027832,
    "timestamp": "2025-07-14T12:34:13.584200"
  },
  {
    "epoch": 1122,
    "batch_idx": 128,
    "loss": 0.4587884545326233,
    "learning_rate": 0.00040476683438440503,
    "timestep_mean": 493.46875,
    "epsilon_norm": 8.211321830749512,
    "prediction_norm": 6.732812404632568,
    "timestamp": "2025-07-14T12:34:13.589253"
  },
  {
    "epoch": 1122,
    "batch_idx": 160,
    "loss": 0.3080080449581146,
    "learning_rate": 0.00040476683438440503,
    "timestep_mean": 502.9375,
    "epsilon_norm": 8.21476936340332,
    "prediction_norm": 6.830102920532227,
    "timestamp": "2025-07-14T12:34:13.594143"
  },
  {
    "epoch": 1122,
    "batch_idx": 192,
    "loss": 0.0951809361577034,
    "learning_rate": 0.00040476683438440503,
    "timestep_mean": 679.5,
    "epsilon_norm": 3.945202112197876,
    "prediction_norm": 3.796196937561035,
    "timestamp": "2025-07-14T12:34:13.599019"
  },
  {
    "epoch": 1123,
    "batch_idx": 0,
    "loss": 0.0908181369304657,
    "learning_rate": 0.00040399593173924,
    "timestep_mean": 596.8125,
    "epsilon_norm": 6.454103469848633,
    "prediction_norm": 6.431032657623291,
    "timestamp": "2025-07-14T12:34:13.603927"
  },
  {
    "epoch": 1123,
    "batch_idx": 32,
    "loss": 0.4174755811691284,
    "learning_rate": 0.00040399593173924,
    "timestep_mean": 513.875,
    "epsilon_norm": 9.419476509094238,
    "prediction_norm": 7.516204833984375,
    "timestamp": "2025-07-14T12:34:13.608934"
  },
  {
    "epoch": 1123,
    "batch_idx": 64,
    "loss": 0.15658006072044373,
    "learning_rate": 0.00040399593173924,
    "timestep_mean": 512.40625,
    "epsilon_norm": 8.061711311340332,
    "prediction_norm": 7.959182262420654,
    "timestamp": "2025-07-14T12:34:13.613827"
  },
  {
    "epoch": 1123,
    "batch_idx": 96,
    "loss": 0.24736106395721436,
    "learning_rate": 0.00040399593173924,
    "timestep_mean": 511.375,
    "epsilon_norm": 8.159173965454102,
    "prediction_norm": 7.161170959472656,
    "timestamp": "2025-07-14T12:34:13.618790"
  },
  {
    "epoch": 1123,
    "batch_idx": 128,
    "loss": 0.31743910908699036,
    "learning_rate": 0.00040399593173924,
    "timestep_mean": 478.0625,
    "epsilon_norm": 7.235400676727295,
    "prediction_norm": 6.417458534240723,
    "timestamp": "2025-07-14T12:34:13.623834"
  },
  {
    "epoch": 1123,
    "batch_idx": 160,
    "loss": 0.17224353551864624,
    "learning_rate": 0.00040399593173924,
    "timestep_mean": 552.28125,
    "epsilon_norm": 6.439460754394531,
    "prediction_norm": 5.871502876281738,
    "timestamp": "2025-07-14T12:34:13.628749"
  },
  {
    "epoch": 1123,
    "batch_idx": 192,
    "loss": 0.28035444021224976,
    "learning_rate": 0.00040399593173924,
    "timestep_mean": 472.5,
    "epsilon_norm": 3.52986741065979,
    "prediction_norm": 2.41256046295166,
    "timestamp": "2025-07-14T12:34:13.633778"
  },
  {
    "epoch": 1124,
    "batch_idx": 0,
    "loss": 0.21798135340213776,
    "learning_rate": 0.0004032252659745699,
    "timestep_mean": 485.53125,
    "epsilon_norm": 6.6049957275390625,
    "prediction_norm": 5.585123538970947,
    "timestamp": "2025-07-14T12:34:13.638743"
  },
  {
    "epoch": 1124,
    "batch_idx": 32,
    "loss": 0.17661067843437195,
    "learning_rate": 0.0004032252659745699,
    "timestep_mean": 497.125,
    "epsilon_norm": 8.502934455871582,
    "prediction_norm": 7.655555248260498,
    "timestamp": "2025-07-14T12:34:13.643597"
  },
  {
    "epoch": 1124,
    "batch_idx": 64,
    "loss": 0.2177671194076538,
    "learning_rate": 0.0004032252659745699,
    "timestep_mean": 504.0625,
    "epsilon_norm": 8.410421371459961,
    "prediction_norm": 7.3727288246154785,
    "timestamp": "2025-07-14T12:34:13.648577"
  },
  {
    "epoch": 1124,
    "batch_idx": 96,
    "loss": 0.13564491271972656,
    "learning_rate": 0.0004032252659745699,
    "timestep_mean": 506.375,
    "epsilon_norm": 7.995684623718262,
    "prediction_norm": 7.393670082092285,
    "timestamp": "2025-07-14T12:34:13.653862"
  },
  {
    "epoch": 1124,
    "batch_idx": 128,
    "loss": 0.3266885280609131,
    "learning_rate": 0.0004032252659745699,
    "timestep_mean": 480.9375,
    "epsilon_norm": 7.382072448730469,
    "prediction_norm": 5.711306095123291,
    "timestamp": "2025-07-14T12:34:13.659400"
  },
  {
    "epoch": 1124,
    "batch_idx": 160,
    "loss": 0.29742199182510376,
    "learning_rate": 0.0004032252659745699,
    "timestep_mean": 502.8125,
    "epsilon_norm": 8.271197319030762,
    "prediction_norm": 6.653807163238525,
    "timestamp": "2025-07-14T12:34:13.664703"
  },
  {
    "epoch": 1124,
    "batch_idx": 192,
    "loss": 0.7069421410560608,
    "learning_rate": 0.0004032252659745699,
    "timestep_mean": 473.25,
    "epsilon_norm": 4.851113319396973,
    "prediction_norm": 3.5127387046813965,
    "timestamp": "2025-07-14T12:34:13.670012"
  },
  {
    "epoch": 1125,
    "batch_idx": 0,
    "loss": 0.19493132829666138,
    "learning_rate": 0.00040245483899193583,
    "timestep_mean": 517.125,
    "epsilon_norm": 6.397425651550293,
    "prediction_norm": 5.303859710693359,
    "timestamp": "2025-07-14T12:34:13.675168"
  },
  {
    "epoch": 1125,
    "batch_idx": 32,
    "loss": 0.3385113477706909,
    "learning_rate": 0.00040245483899193583,
    "timestep_mean": 408.1875,
    "epsilon_norm": 7.980759143829346,
    "prediction_norm": 6.531503677368164,
    "timestamp": "2025-07-14T12:34:13.680067"
  },
  {
    "epoch": 1125,
    "batch_idx": 64,
    "loss": 0.28768008947372437,
    "learning_rate": 0.00040245483899193583,
    "timestep_mean": 531.875,
    "epsilon_norm": 7.636200428009033,
    "prediction_norm": 6.043865203857422,
    "timestamp": "2025-07-14T12:34:13.685018"
  },
  {
    "epoch": 1125,
    "batch_idx": 96,
    "loss": 0.23633122444152832,
    "learning_rate": 0.00040245483899193583,
    "timestep_mean": 490.75,
    "epsilon_norm": 7.955045700073242,
    "prediction_norm": 7.194482326507568,
    "timestamp": "2025-07-14T12:34:13.690026"
  },
  {
    "epoch": 1125,
    "batch_idx": 128,
    "loss": 0.47177883982658386,
    "learning_rate": 0.00040245483899193583,
    "timestep_mean": 404.59375,
    "epsilon_norm": 8.38509750366211,
    "prediction_norm": 6.298304557800293,
    "timestamp": "2025-07-14T12:34:13.694884"
  },
  {
    "epoch": 1125,
    "batch_idx": 160,
    "loss": 0.31445106863975525,
    "learning_rate": 0.00040245483899193583,
    "timestep_mean": 424.25,
    "epsilon_norm": 8.542695999145508,
    "prediction_norm": 7.457083225250244,
    "timestamp": "2025-07-14T12:34:13.699771"
  },
  {
    "epoch": 1125,
    "batch_idx": 192,
    "loss": 0.8876187801361084,
    "learning_rate": 0.00040245483899193583,
    "timestep_mean": 480.75,
    "epsilon_norm": 3.9446921348571777,
    "prediction_norm": 2.0005366802215576,
    "timestamp": "2025-07-14T12:34:13.704771"
  },
  {
    "epoch": 1126,
    "batch_idx": 0,
    "loss": 0.28954461216926575,
    "learning_rate": 0.00040168465269228983,
    "timestep_mean": 470.5,
    "epsilon_norm": 8.756759643554688,
    "prediction_norm": 7.603485584259033,
    "timestamp": "2025-07-14T12:34:13.709653"
  },
  {
    "epoch": 1126,
    "batch_idx": 32,
    "loss": 0.3905990719795227,
    "learning_rate": 0.00040168465269228983,
    "timestep_mean": 395.9375,
    "epsilon_norm": 7.943403244018555,
    "prediction_norm": 6.313092231750488,
    "timestamp": "2025-07-14T12:34:13.714594"
  },
  {
    "epoch": 1126,
    "batch_idx": 64,
    "loss": 0.2469567060470581,
    "learning_rate": 0.00040168465269228983,
    "timestep_mean": 481.0625,
    "epsilon_norm": 5.892322540283203,
    "prediction_norm": 4.882059097290039,
    "timestamp": "2025-07-14T12:34:13.719539"
  },
  {
    "epoch": 1126,
    "batch_idx": 96,
    "loss": 0.15341678261756897,
    "learning_rate": 0.00040168465269228983,
    "timestep_mean": 481.5,
    "epsilon_norm": 8.379881858825684,
    "prediction_norm": 7.0146684646606445,
    "timestamp": "2025-07-14T12:34:13.724551"
  },
  {
    "epoch": 1126,
    "batch_idx": 128,
    "loss": 0.33628612756729126,
    "learning_rate": 0.00040168465269228983,
    "timestep_mean": 483.25,
    "epsilon_norm": 8.133074760437012,
    "prediction_norm": 6.396392345428467,
    "timestamp": "2025-07-14T12:34:13.729572"
  },
  {
    "epoch": 1126,
    "batch_idx": 160,
    "loss": 0.26910150051116943,
    "learning_rate": 0.00040168465269228983,
    "timestep_mean": 513.1875,
    "epsilon_norm": 7.53068733215332,
    "prediction_norm": 6.008725643157959,
    "timestamp": "2025-07-14T12:34:13.734397"
  },
  {
    "epoch": 1126,
    "batch_idx": 192,
    "loss": 0.2527707815170288,
    "learning_rate": 0.00040168465269228983,
    "timestep_mean": 627.5,
    "epsilon_norm": 3.566917896270752,
    "prediction_norm": 2.9431426525115967,
    "timestamp": "2025-07-14T12:34:13.739250"
  },
  {
    "epoch": 1127,
    "batch_idx": 0,
    "loss": 0.2984456419944763,
    "learning_rate": 0.00040091470897599027,
    "timestep_mean": 497.78125,
    "epsilon_norm": 6.939300537109375,
    "prediction_norm": 5.819174766540527,
    "timestamp": "2025-07-14T12:34:13.744249"
  },
  {
    "epoch": 1127,
    "batch_idx": 32,
    "loss": 0.3323746919631958,
    "learning_rate": 0.00040091470897599027,
    "timestep_mean": 528.3125,
    "epsilon_norm": 7.430350303649902,
    "prediction_norm": 6.221197128295898,
    "timestamp": "2025-07-14T12:34:13.749169"
  },
  {
    "epoch": 1127,
    "batch_idx": 64,
    "loss": 0.16708417236804962,
    "learning_rate": 0.00040091470897599027,
    "timestep_mean": 587.34375,
    "epsilon_norm": 6.9410271644592285,
    "prediction_norm": 6.486867904663086,
    "timestamp": "2025-07-14T12:34:13.754058"
  },
  {
    "epoch": 1127,
    "batch_idx": 96,
    "loss": 0.34091007709503174,
    "learning_rate": 0.00040091470897599027,
    "timestep_mean": 556.15625,
    "epsilon_norm": 6.531424522399902,
    "prediction_norm": 5.546199321746826,
    "timestamp": "2025-07-14T12:34:13.759097"
  },
  {
    "epoch": 1127,
    "batch_idx": 128,
    "loss": 0.21619443595409393,
    "learning_rate": 0.00040091470897599027,
    "timestep_mean": 454.375,
    "epsilon_norm": 7.645638942718506,
    "prediction_norm": 6.233405113220215,
    "timestamp": "2025-07-14T12:34:13.764052"
  },
  {
    "epoch": 1127,
    "batch_idx": 160,
    "loss": 0.24573665857315063,
    "learning_rate": 0.00040091470897599027,
    "timestep_mean": 553.65625,
    "epsilon_norm": 8.896137237548828,
    "prediction_norm": 8.122309684753418,
    "timestamp": "2025-07-14T12:34:13.769016"
  },
  {
    "epoch": 1127,
    "batch_idx": 192,
    "loss": 0.042088672518730164,
    "learning_rate": 0.00040091470897599027,
    "timestep_mean": 509.625,
    "epsilon_norm": 2.6980886459350586,
    "prediction_norm": 2.7835988998413086,
    "timestamp": "2025-07-14T12:34:13.773898"
  },
  {
    "epoch": 1128,
    "batch_idx": 0,
    "loss": 0.15793320536613464,
    "learning_rate": 0.00040014500974279644,
    "timestep_mean": 526.1875,
    "epsilon_norm": 7.626742839813232,
    "prediction_norm": 6.525051116943359,
    "timestamp": "2025-07-14T12:34:13.778817"
  },
  {
    "epoch": 1128,
    "batch_idx": 32,
    "loss": 0.3896011710166931,
    "learning_rate": 0.00040014500974279644,
    "timestep_mean": 467.4375,
    "epsilon_norm": 7.99256706237793,
    "prediction_norm": 6.50429105758667,
    "timestamp": "2025-07-14T12:34:13.783763"
  },
  {
    "epoch": 1128,
    "batch_idx": 64,
    "loss": 0.20031718909740448,
    "learning_rate": 0.00040014500974279644,
    "timestep_mean": 482.46875,
    "epsilon_norm": 8.377517700195312,
    "prediction_norm": 6.946864128112793,
    "timestamp": "2025-07-14T12:34:13.789553"
  },
  {
    "epoch": 1128,
    "batch_idx": 96,
    "loss": 0.1898140013217926,
    "learning_rate": 0.00040014500974279644,
    "timestep_mean": 427.96875,
    "epsilon_norm": 8.083984375,
    "prediction_norm": 6.600069522857666,
    "timestamp": "2025-07-14T12:34:13.799381"
  },
  {
    "epoch": 1128,
    "batch_idx": 128,
    "loss": 0.45379573106765747,
    "learning_rate": 0.00040014500974279644,
    "timestep_mean": 437.0625,
    "epsilon_norm": 7.79269552230835,
    "prediction_norm": 5.688292026519775,
    "timestamp": "2025-07-14T12:34:13.804621"
  },
  {
    "epoch": 1128,
    "batch_idx": 160,
    "loss": 0.18032866716384888,
    "learning_rate": 0.00040014500974279644,
    "timestep_mean": 507.6875,
    "epsilon_norm": 7.46503210067749,
    "prediction_norm": 6.090394973754883,
    "timestamp": "2025-07-14T12:34:13.809348"
  },
  {
    "epoch": 1128,
    "batch_idx": 192,
    "loss": 0.15137754380702972,
    "learning_rate": 0.00040014500974279644,
    "timestep_mean": 617.375,
    "epsilon_norm": 3.6881775856018066,
    "prediction_norm": 3.402834415435791,
    "timestamp": "2025-07-14T12:34:13.814250"
  },
  {
    "epoch": 1129,
    "batch_idx": 0,
    "loss": 0.33931446075439453,
    "learning_rate": 0.00039937555689186475,
    "timestep_mean": 468.28125,
    "epsilon_norm": 8.25335693359375,
    "prediction_norm": 5.901125431060791,
    "timestamp": "2025-07-14T12:34:13.819042"
  },
  {
    "epoch": 1129,
    "batch_idx": 32,
    "loss": 0.2492799609899521,
    "learning_rate": 0.00039937555689186475,
    "timestep_mean": 387.125,
    "epsilon_norm": 7.5515828132629395,
    "prediction_norm": 6.224985122680664,
    "timestamp": "2025-07-14T12:34:13.823765"
  },
  {
    "epoch": 1129,
    "batch_idx": 64,
    "loss": 0.258566677570343,
    "learning_rate": 0.00039937555689186475,
    "timestep_mean": 446.71875,
    "epsilon_norm": 8.091852188110352,
    "prediction_norm": 6.728281021118164,
    "timestamp": "2025-07-14T12:34:13.828496"
  },
  {
    "epoch": 1129,
    "batch_idx": 96,
    "loss": 0.3212851583957672,
    "learning_rate": 0.00039937555689186475,
    "timestep_mean": 487.59375,
    "epsilon_norm": 7.555502414703369,
    "prediction_norm": 5.3677263259887695,
    "timestamp": "2025-07-14T12:34:13.833179"
  },
  {
    "epoch": 1129,
    "batch_idx": 128,
    "loss": 0.264660120010376,
    "learning_rate": 0.00039937555689186475,
    "timestep_mean": 457.78125,
    "epsilon_norm": 7.923306941986084,
    "prediction_norm": 6.501439571380615,
    "timestamp": "2025-07-14T12:34:13.838069"
  },
  {
    "epoch": 1129,
    "batch_idx": 160,
    "loss": 0.2465066909790039,
    "learning_rate": 0.00039937555689186475,
    "timestep_mean": 478.75,
    "epsilon_norm": 6.919655799865723,
    "prediction_norm": 5.684032440185547,
    "timestamp": "2025-07-14T12:34:13.842741"
  },
  {
    "epoch": 1129,
    "batch_idx": 192,
    "loss": 0.31584250926971436,
    "learning_rate": 0.00039937555689186475,
    "timestep_mean": 641.25,
    "epsilon_norm": 4.411988258361816,
    "prediction_norm": 3.9247140884399414,
    "timestamp": "2025-07-14T12:34:13.847553"
  },
  {
    "epoch": 1130,
    "batch_idx": 0,
    "loss": 0.292397677898407,
    "learning_rate": 0.00039860635232174375,
    "timestep_mean": 525.75,
    "epsilon_norm": 9.55142879486084,
    "prediction_norm": 7.6978559494018555,
    "timestamp": "2025-07-14T12:34:13.852370"
  },
  {
    "epoch": 1130,
    "batch_idx": 32,
    "loss": 0.22709082067012787,
    "learning_rate": 0.00039860635232174375,
    "timestep_mean": 456.375,
    "epsilon_norm": 7.572969436645508,
    "prediction_norm": 6.4199347496032715,
    "timestamp": "2025-07-14T12:34:13.857103"
  },
  {
    "epoch": 1130,
    "batch_idx": 64,
    "loss": 0.21001583337783813,
    "learning_rate": 0.00039860635232174375,
    "timestep_mean": 487.53125,
    "epsilon_norm": 8.32964038848877,
    "prediction_norm": 8.067191123962402,
    "timestamp": "2025-07-14T12:34:13.861945"
  },
  {
    "epoch": 1130,
    "batch_idx": 96,
    "loss": 0.08735823631286621,
    "learning_rate": 0.00039860635232174375,
    "timestep_mean": 609.96875,
    "epsilon_norm": 7.644184589385986,
    "prediction_norm": 7.500755786895752,
    "timestamp": "2025-07-14T12:34:13.866974"
  },
  {
    "epoch": 1130,
    "batch_idx": 128,
    "loss": 0.18709607422351837,
    "learning_rate": 0.00039860635232174375,
    "timestep_mean": 503.40625,
    "epsilon_norm": 8.505167007446289,
    "prediction_norm": 7.589408874511719,
    "timestamp": "2025-07-14T12:34:13.872021"
  },
  {
    "epoch": 1130,
    "batch_idx": 160,
    "loss": 0.17287248373031616,
    "learning_rate": 0.00039860635232174375,
    "timestep_mean": 577.625,
    "epsilon_norm": 7.811465740203857,
    "prediction_norm": 7.500733375549316,
    "timestamp": "2025-07-14T12:34:13.876928"
  },
  {
    "epoch": 1130,
    "batch_idx": 192,
    "loss": 0.08206555992364883,
    "learning_rate": 0.00039860635232174375,
    "timestep_mean": 529.625,
    "epsilon_norm": 4.53671932220459,
    "prediction_norm": 4.016183853149414,
    "timestamp": "2025-07-14T12:34:13.881833"
  },
  {
    "epoch": 1131,
    "batch_idx": 0,
    "loss": 0.34657368063926697,
    "learning_rate": 0.00039783739793036916,
    "timestep_mean": 496.75,
    "epsilon_norm": 8.715574264526367,
    "prediction_norm": 7.756824016571045,
    "timestamp": "2025-07-14T12:34:13.887052"
  },
  {
    "epoch": 1131,
    "batch_idx": 32,
    "loss": 0.3286558985710144,
    "learning_rate": 0.00039783739793036916,
    "timestep_mean": 447.1875,
    "epsilon_norm": 7.985670566558838,
    "prediction_norm": 6.76584529876709,
    "timestamp": "2025-07-14T12:34:13.892180"
  },
  {
    "epoch": 1131,
    "batch_idx": 64,
    "loss": 0.4044463038444519,
    "learning_rate": 0.00039783739793036916,
    "timestep_mean": 459.5,
    "epsilon_norm": 8.206525802612305,
    "prediction_norm": 6.453160285949707,
    "timestamp": "2025-07-14T12:34:13.897304"
  },
  {
    "epoch": 1131,
    "batch_idx": 96,
    "loss": 0.2944830358028412,
    "learning_rate": 0.00039783739793036916,
    "timestep_mean": 477.8125,
    "epsilon_norm": 7.335728645324707,
    "prediction_norm": 7.155403137207031,
    "timestamp": "2025-07-14T12:34:13.902400"
  },
  {
    "epoch": 1131,
    "batch_idx": 128,
    "loss": 0.3399941027164459,
    "learning_rate": 0.00039783739793036916,
    "timestep_mean": 508.5,
    "epsilon_norm": 6.860202312469482,
    "prediction_norm": 6.130370140075684,
    "timestamp": "2025-07-14T12:34:13.907365"
  },
  {
    "epoch": 1131,
    "batch_idx": 160,
    "loss": 0.2544548511505127,
    "learning_rate": 0.00039783739793036916,
    "timestep_mean": 501.53125,
    "epsilon_norm": 8.619146347045898,
    "prediction_norm": 7.701533794403076,
    "timestamp": "2025-07-14T12:34:13.912491"
  },
  {
    "epoch": 1131,
    "batch_idx": 192,
    "loss": 0.40910935401916504,
    "learning_rate": 0.00039783739793036916,
    "timestep_mean": 428.75,
    "epsilon_norm": 5.081083297729492,
    "prediction_norm": 4.401875972747803,
    "timestamp": "2025-07-14T12:34:13.917633"
  },
  {
    "epoch": 1132,
    "batch_idx": 0,
    "loss": 0.1833876371383667,
    "learning_rate": 0.0003970686956150592,
    "timestep_mean": 466.6875,
    "epsilon_norm": 7.956714153289795,
    "prediction_norm": 7.071269512176514,
    "timestamp": "2025-07-14T12:34:13.922715"
  },
  {
    "epoch": 1132,
    "batch_idx": 32,
    "loss": 0.1524185836315155,
    "learning_rate": 0.0003970686956150592,
    "timestep_mean": 493.65625,
    "epsilon_norm": 5.857868194580078,
    "prediction_norm": 5.4401774406433105,
    "timestamp": "2025-07-14T12:34:13.927892"
  },
  {
    "epoch": 1132,
    "batch_idx": 64,
    "loss": 0.3060462176799774,
    "learning_rate": 0.0003970686956150592,
    "timestep_mean": 424.0625,
    "epsilon_norm": 7.013152599334717,
    "prediction_norm": 5.454968452453613,
    "timestamp": "2025-07-14T12:34:13.932894"
  },
  {
    "epoch": 1132,
    "batch_idx": 96,
    "loss": 0.27297765016555786,
    "learning_rate": 0.0003970686956150592,
    "timestep_mean": 594.5625,
    "epsilon_norm": 7.894796371459961,
    "prediction_norm": 6.477025985717773,
    "timestamp": "2025-07-14T12:34:13.938144"
  },
  {
    "epoch": 1132,
    "batch_idx": 128,
    "loss": 0.35479143261909485,
    "learning_rate": 0.0003970686956150592,
    "timestep_mean": 460.25,
    "epsilon_norm": 8.559444427490234,
    "prediction_norm": 7.308176040649414,
    "timestamp": "2025-07-14T12:34:13.943432"
  },
  {
    "epoch": 1132,
    "batch_idx": 160,
    "loss": 0.20533087849617004,
    "learning_rate": 0.0003970686956150592,
    "timestep_mean": 560.09375,
    "epsilon_norm": 7.658478736877441,
    "prediction_norm": 6.560242652893066,
    "timestamp": "2025-07-14T12:34:13.948589"
  },
  {
    "epoch": 1132,
    "batch_idx": 192,
    "loss": 0.18291717767715454,
    "learning_rate": 0.0003970686956150592,
    "timestep_mean": 635.25,
    "epsilon_norm": 3.1436820030212402,
    "prediction_norm": 2.8842015266418457,
    "timestamp": "2025-07-14T12:34:13.953769"
  },
  {
    "epoch": 1133,
    "batch_idx": 0,
    "loss": 0.39172467589378357,
    "learning_rate": 0.00039630024727251094,
    "timestep_mean": 485.78125,
    "epsilon_norm": 7.537563800811768,
    "prediction_norm": 6.247788906097412,
    "timestamp": "2025-07-14T12:34:13.958844"
  },
  {
    "epoch": 1133,
    "batch_idx": 32,
    "loss": 0.21612221002578735,
    "learning_rate": 0.00039630024727251094,
    "timestep_mean": 493.125,
    "epsilon_norm": 7.145933151245117,
    "prediction_norm": 6.308669567108154,
    "timestamp": "2025-07-14T12:34:13.963896"
  },
  {
    "epoch": 1133,
    "batch_idx": 64,
    "loss": 0.20182302594184875,
    "learning_rate": 0.00039630024727251094,
    "timestep_mean": 469.40625,
    "epsilon_norm": 7.668149471282959,
    "prediction_norm": 6.812182426452637,
    "timestamp": "2025-07-14T12:34:13.969115"
  },
  {
    "epoch": 1133,
    "batch_idx": 96,
    "loss": 0.26382067799568176,
    "learning_rate": 0.00039630024727251094,
    "timestep_mean": 509.25,
    "epsilon_norm": 7.23725700378418,
    "prediction_norm": 5.925331115722656,
    "timestamp": "2025-07-14T12:34:13.974392"
  },
  {
    "epoch": 1133,
    "batch_idx": 128,
    "loss": 0.19771608710289001,
    "learning_rate": 0.00039630024727251094,
    "timestep_mean": 584.5625,
    "epsilon_norm": 6.648147106170654,
    "prediction_norm": 5.598642826080322,
    "timestamp": "2025-07-14T12:34:13.979591"
  },
  {
    "epoch": 1133,
    "batch_idx": 160,
    "loss": 0.1824539303779602,
    "learning_rate": 0.00039630024727251094,
    "timestep_mean": 550.375,
    "epsilon_norm": 9.32208251953125,
    "prediction_norm": 8.147103309631348,
    "timestamp": "2025-07-14T12:34:13.984803"
  },
  {
    "epoch": 1133,
    "batch_idx": 192,
    "loss": 0.46316444873809814,
    "learning_rate": 0.00039630024727251094,
    "timestep_mean": 452.625,
    "epsilon_norm": 4.439464569091797,
    "prediction_norm": 3.110142230987549,
    "timestamp": "2025-07-14T12:34:13.989975"
  },
  {
    "epoch": 1134,
    "batch_idx": 0,
    "loss": 0.2213047295808792,
    "learning_rate": 0.00039553205479879406,
    "timestep_mean": 432.90625,
    "epsilon_norm": 8.041778564453125,
    "prediction_norm": 6.480164527893066,
    "timestamp": "2025-07-14T12:34:13.995061"
  },
  {
    "epoch": 1134,
    "batch_idx": 32,
    "loss": 0.28270694613456726,
    "learning_rate": 0.00039553205479879406,
    "timestep_mean": 439.5,
    "epsilon_norm": 6.9986891746521,
    "prediction_norm": 5.133944034576416,
    "timestamp": "2025-07-14T12:34:14.000475"
  },
  {
    "epoch": 1134,
    "batch_idx": 64,
    "loss": 0.4949076175689697,
    "learning_rate": 0.00039553205479879406,
    "timestep_mean": 429.71875,
    "epsilon_norm": 8.465314865112305,
    "prediction_norm": 6.83397102355957,
    "timestamp": "2025-07-14T12:34:14.005752"
  },
  {
    "epoch": 1134,
    "batch_idx": 96,
    "loss": 0.29061391949653625,
    "learning_rate": 0.00039553205479879406,
    "timestep_mean": 493.09375,
    "epsilon_norm": 8.310099601745605,
    "prediction_norm": 6.873495101928711,
    "timestamp": "2025-07-14T12:34:14.010953"
  },
  {
    "epoch": 1134,
    "batch_idx": 128,
    "loss": 0.344396710395813,
    "learning_rate": 0.00039553205479879406,
    "timestep_mean": 432.0625,
    "epsilon_norm": 8.48207950592041,
    "prediction_norm": 7.183844566345215,
    "timestamp": "2025-07-14T12:34:14.016222"
  },
  {
    "epoch": 1134,
    "batch_idx": 160,
    "loss": 0.48242950439453125,
    "learning_rate": 0.00039553205479879406,
    "timestep_mean": 442.78125,
    "epsilon_norm": 7.715450286865234,
    "prediction_norm": 5.5716166496276855,
    "timestamp": "2025-07-14T12:34:14.021561"
  },
  {
    "epoch": 1134,
    "batch_idx": 192,
    "loss": 0.11684928089380264,
    "learning_rate": 0.00039553205479879406,
    "timestep_mean": 669.0,
    "epsilon_norm": 3.794022798538208,
    "prediction_norm": 3.209289073944092,
    "timestamp": "2025-07-14T12:34:14.026696"
  },
  {
    "epoch": 1135,
    "batch_idx": 0,
    "loss": 0.22908075153827667,
    "learning_rate": 0.00039476412008934713,
    "timestep_mean": 488.4375,
    "epsilon_norm": 7.819116592407227,
    "prediction_norm": 6.348023891448975,
    "timestamp": "2025-07-14T12:34:14.031905"
  },
  {
    "epoch": 1135,
    "batch_idx": 32,
    "loss": 0.15365393459796906,
    "learning_rate": 0.00039476412008934713,
    "timestep_mean": 489.53125,
    "epsilon_norm": 7.089206695556641,
    "prediction_norm": 7.044166564941406,
    "timestamp": "2025-07-14T12:34:14.037114"
  },
  {
    "epoch": 1135,
    "batch_idx": 64,
    "loss": 0.14947834610939026,
    "learning_rate": 0.00039476412008934713,
    "timestep_mean": 500.0625,
    "epsilon_norm": 8.196650505065918,
    "prediction_norm": 7.30072546005249,
    "timestamp": "2025-07-14T12:34:14.042397"
  },
  {
    "epoch": 1135,
    "batch_idx": 96,
    "loss": 0.277430921792984,
    "learning_rate": 0.00039476412008934713,
    "timestep_mean": 497.0625,
    "epsilon_norm": 7.413262844085693,
    "prediction_norm": 6.224102020263672,
    "timestamp": "2025-07-14T12:34:14.047589"
  },
  {
    "epoch": 1135,
    "batch_idx": 128,
    "loss": 0.2751106023788452,
    "learning_rate": 0.00039476412008934713,
    "timestep_mean": 534.65625,
    "epsilon_norm": 7.669816970825195,
    "prediction_norm": 6.834609031677246,
    "timestamp": "2025-07-14T12:34:14.052667"
  },
  {
    "epoch": 1135,
    "batch_idx": 160,
    "loss": 0.22504404187202454,
    "learning_rate": 0.00039476412008934713,
    "timestep_mean": 513.8125,
    "epsilon_norm": 7.454267501831055,
    "prediction_norm": 7.265714168548584,
    "timestamp": "2025-07-14T12:34:14.057850"
  },
  {
    "epoch": 1135,
    "batch_idx": 192,
    "loss": 0.33261674642562866,
    "learning_rate": 0.00039476412008934713,
    "timestep_mean": 390.625,
    "epsilon_norm": 3.5578901767730713,
    "prediction_norm": 2.746659994125366,
    "timestamp": "2025-07-14T12:34:14.063297"
  },
  {
    "epoch": 1136,
    "batch_idx": 0,
    "loss": 0.19108088314533234,
    "learning_rate": 0.00039399644503897255,
    "timestep_mean": 598.65625,
    "epsilon_norm": 8.356751441955566,
    "prediction_norm": 7.606780052185059,
    "timestamp": "2025-07-14T12:34:14.068468"
  },
  {
    "epoch": 1136,
    "batch_idx": 32,
    "loss": 0.3350665271282196,
    "learning_rate": 0.00039399644503897255,
    "timestep_mean": 451.5,
    "epsilon_norm": 7.872028827667236,
    "prediction_norm": 6.352400779724121,
    "timestamp": "2025-07-14T12:34:14.073678"
  },
  {
    "epoch": 1136,
    "batch_idx": 64,
    "loss": 0.30434805154800415,
    "learning_rate": 0.00039399644503897255,
    "timestep_mean": 436.09375,
    "epsilon_norm": 7.527281761169434,
    "prediction_norm": 6.352514743804932,
    "timestamp": "2025-07-14T12:34:14.078931"
  },
  {
    "epoch": 1136,
    "batch_idx": 96,
    "loss": 0.10632670670747757,
    "learning_rate": 0.00039399644503897255,
    "timestep_mean": 482.90625,
    "epsilon_norm": 7.792798042297363,
    "prediction_norm": 7.444812774658203,
    "timestamp": "2025-07-14T12:34:14.084168"
  },
  {
    "epoch": 1136,
    "batch_idx": 128,
    "loss": 0.3340321183204651,
    "learning_rate": 0.00039399644503897255,
    "timestep_mean": 480.96875,
    "epsilon_norm": 8.08875560760498,
    "prediction_norm": 6.997134685516357,
    "timestamp": "2025-07-14T12:34:14.089474"
  },
  {
    "epoch": 1136,
    "batch_idx": 160,
    "loss": 0.31961220502853394,
    "learning_rate": 0.00039399644503897255,
    "timestep_mean": 381.8125,
    "epsilon_norm": 7.617100715637207,
    "prediction_norm": 6.011585235595703,
    "timestamp": "2025-07-14T12:34:14.094691"
  },
  {
    "epoch": 1136,
    "batch_idx": 192,
    "loss": 0.11809836328029633,
    "learning_rate": 0.00039399644503897255,
    "timestep_mean": 451.75,
    "epsilon_norm": 3.8024277687072754,
    "prediction_norm": 3.826406955718994,
    "timestamp": "2025-07-14T12:34:14.099790"
  },
  {
    "epoch": 1137,
    "batch_idx": 0,
    "loss": 0.14930081367492676,
    "learning_rate": 0.0003932290315418325,
    "timestep_mean": 591.34375,
    "epsilon_norm": 7.968245029449463,
    "prediction_norm": 7.254802227020264,
    "timestamp": "2025-07-14T12:34:14.104929"
  },
  {
    "epoch": 1137,
    "batch_idx": 32,
    "loss": 0.43271487951278687,
    "learning_rate": 0.0003932290315418325,
    "timestep_mean": 501.53125,
    "epsilon_norm": 8.103544235229492,
    "prediction_norm": 6.215404987335205,
    "timestamp": "2025-07-14T12:34:14.110157"
  },
  {
    "epoch": 1137,
    "batch_idx": 64,
    "loss": 0.24855081737041473,
    "learning_rate": 0.0003932290315418325,
    "timestep_mean": 456.0,
    "epsilon_norm": 6.750946044921875,
    "prediction_norm": 5.642316818237305,
    "timestamp": "2025-07-14T12:34:14.115416"
  },
  {
    "epoch": 1137,
    "batch_idx": 96,
    "loss": 0.16871419548988342,
    "learning_rate": 0.0003932290315418325,
    "timestep_mean": 458.03125,
    "epsilon_norm": 7.051204681396484,
    "prediction_norm": 5.681126117706299,
    "timestamp": "2025-07-14T12:34:14.120414"
  },
  {
    "epoch": 1137,
    "batch_idx": 128,
    "loss": 0.20063556730747223,
    "learning_rate": 0.0003932290315418325,
    "timestep_mean": 534.6875,
    "epsilon_norm": 7.3973846435546875,
    "prediction_norm": 5.807164669036865,
    "timestamp": "2025-07-14T12:34:14.125523"
  },
  {
    "epoch": 1137,
    "batch_idx": 160,
    "loss": 0.2231079638004303,
    "learning_rate": 0.0003932290315418325,
    "timestep_mean": 496.21875,
    "epsilon_norm": 7.182586669921875,
    "prediction_norm": 6.245485782623291,
    "timestamp": "2025-07-14T12:34:14.130664"
  },
  {
    "epoch": 1137,
    "batch_idx": 192,
    "loss": 0.33909255266189575,
    "learning_rate": 0.0003932290315418325,
    "timestep_mean": 439.5,
    "epsilon_norm": 2.782550573348999,
    "prediction_norm": 2.9012343883514404,
    "timestamp": "2025-07-14T12:34:14.135540"
  },
  {
    "epoch": 1138,
    "batch_idx": 0,
    "loss": 0.19957810640335083,
    "learning_rate": 0.0003924618814914432,
    "timestep_mean": 564.75,
    "epsilon_norm": 8.119312286376953,
    "prediction_norm": 7.176005840301514,
    "timestamp": "2025-07-14T12:34:14.140711"
  },
  {
    "epoch": 1138,
    "batch_idx": 32,
    "loss": 0.5018202662467957,
    "learning_rate": 0.0003924618814914432,
    "timestep_mean": 398.1875,
    "epsilon_norm": 8.998549461364746,
    "prediction_norm": 6.471611022949219,
    "timestamp": "2025-07-14T12:34:14.145923"
  },
  {
    "epoch": 1138,
    "batch_idx": 64,
    "loss": 0.17611092329025269,
    "learning_rate": 0.0003924618814914432,
    "timestep_mean": 542.25,
    "epsilon_norm": 8.257763862609863,
    "prediction_norm": 7.139106273651123,
    "timestamp": "2025-07-14T12:34:14.151113"
  },
  {
    "epoch": 1138,
    "batch_idx": 96,
    "loss": 0.20009148120880127,
    "learning_rate": 0.0003924618814914432,
    "timestep_mean": 500.8125,
    "epsilon_norm": 7.624078273773193,
    "prediction_norm": 6.3830766677856445,
    "timestamp": "2025-07-14T12:34:14.156307"
  },
  {
    "epoch": 1138,
    "batch_idx": 128,
    "loss": 0.4592140316963196,
    "learning_rate": 0.0003924618814914432,
    "timestep_mean": 469.4375,
    "epsilon_norm": 8.839962005615234,
    "prediction_norm": 6.495432376861572,
    "timestamp": "2025-07-14T12:34:14.161505"
  },
  {
    "epoch": 1138,
    "batch_idx": 160,
    "loss": 0.18226082623004913,
    "learning_rate": 0.0003924618814914432,
    "timestep_mean": 610.03125,
    "epsilon_norm": 8.607027053833008,
    "prediction_norm": 7.198851585388184,
    "timestamp": "2025-07-14T12:34:14.166747"
  },
  {
    "epoch": 1138,
    "batch_idx": 192,
    "loss": 0.30464988946914673,
    "learning_rate": 0.0003924618814914432,
    "timestep_mean": 507.625,
    "epsilon_norm": 3.717515468597412,
    "prediction_norm": 3.1668927669525146,
    "timestamp": "2025-07-14T12:34:14.171974"
  },
  {
    "epoch": 1139,
    "batch_idx": 0,
    "loss": 0.19348400831222534,
    "learning_rate": 0.00039169499678067134,
    "timestep_mean": 517.59375,
    "epsilon_norm": 8.737764358520508,
    "prediction_norm": 7.6366496086120605,
    "timestamp": "2025-07-14T12:34:14.177166"
  },
  {
    "epoch": 1139,
    "batch_idx": 32,
    "loss": 0.2822074294090271,
    "learning_rate": 0.00039169499678067134,
    "timestep_mean": 496.46875,
    "epsilon_norm": 8.164822578430176,
    "prediction_norm": 6.878658294677734,
    "timestamp": "2025-07-14T12:34:14.182303"
  },
  {
    "epoch": 1139,
    "batch_idx": 64,
    "loss": 0.2202628254890442,
    "learning_rate": 0.00039169499678067134,
    "timestep_mean": 560.25,
    "epsilon_norm": 7.121800899505615,
    "prediction_norm": 5.505734443664551,
    "timestamp": "2025-07-14T12:34:14.187507"
  },
  {
    "epoch": 1139,
    "batch_idx": 96,
    "loss": 0.3449004292488098,
    "learning_rate": 0.00039169499678067134,
    "timestep_mean": 454.8125,
    "epsilon_norm": 8.090002059936523,
    "prediction_norm": 6.244154453277588,
    "timestamp": "2025-07-14T12:34:14.192716"
  },
  {
    "epoch": 1139,
    "batch_idx": 128,
    "loss": 0.33619576692581177,
    "learning_rate": 0.00039169499678067134,
    "timestep_mean": 511.5625,
    "epsilon_norm": 8.089825630187988,
    "prediction_norm": 6.764580249786377,
    "timestamp": "2025-07-14T12:34:14.197841"
  },
  {
    "epoch": 1139,
    "batch_idx": 160,
    "loss": 0.25318607687950134,
    "learning_rate": 0.00039169499678067134,
    "timestep_mean": 548.71875,
    "epsilon_norm": 7.177679061889648,
    "prediction_norm": 5.6608357429504395,
    "timestamp": "2025-07-14T12:34:14.202912"
  },
  {
    "epoch": 1139,
    "batch_idx": 192,
    "loss": 0.22220779955387115,
    "learning_rate": 0.00039169499678067134,
    "timestep_mean": 699.25,
    "epsilon_norm": 4.878464698791504,
    "prediction_norm": 4.425751209259033,
    "timestamp": "2025-07-14T12:34:14.207779"
  },
  {
    "epoch": 1140,
    "batch_idx": 0,
    "loss": 0.11324609816074371,
    "learning_rate": 0.0003909283793017286,
    "timestep_mean": 556.46875,
    "epsilon_norm": 7.542847156524658,
    "prediction_norm": 7.286678314208984,
    "timestamp": "2025-07-14T12:34:14.212833"
  },
  {
    "epoch": 1140,
    "batch_idx": 32,
    "loss": 0.23372234404087067,
    "learning_rate": 0.0003909283793017286,
    "timestep_mean": 502.65625,
    "epsilon_norm": 7.3261542320251465,
    "prediction_norm": 6.473207473754883,
    "timestamp": "2025-07-14T12:34:14.217687"
  },
  {
    "epoch": 1140,
    "batch_idx": 64,
    "loss": 0.195149227976799,
    "learning_rate": 0.0003909283793017286,
    "timestep_mean": 499.3125,
    "epsilon_norm": 8.069124221801758,
    "prediction_norm": 7.270500183105469,
    "timestamp": "2025-07-14T12:34:14.222703"
  },
  {
    "epoch": 1140,
    "batch_idx": 96,
    "loss": 0.2261735200881958,
    "learning_rate": 0.0003909283793017286,
    "timestep_mean": 523.34375,
    "epsilon_norm": 7.571822643280029,
    "prediction_norm": 6.716710090637207,
    "timestamp": "2025-07-14T12:34:14.227868"
  },
  {
    "epoch": 1140,
    "batch_idx": 128,
    "loss": 0.28089848160743713,
    "learning_rate": 0.0003909283793017286,
    "timestep_mean": 453.15625,
    "epsilon_norm": 7.462567329406738,
    "prediction_norm": 5.749210357666016,
    "timestamp": "2025-07-14T12:34:14.233006"
  },
  {
    "epoch": 1140,
    "batch_idx": 160,
    "loss": 0.10311321169137955,
    "learning_rate": 0.0003909283793017286,
    "timestep_mean": 514.96875,
    "epsilon_norm": 8.380294799804688,
    "prediction_norm": 7.586203575134277,
    "timestamp": "2025-07-14T12:34:14.238323"
  },
  {
    "epoch": 1140,
    "batch_idx": 192,
    "loss": 0.24948924779891968,
    "learning_rate": 0.0003909283793017286,
    "timestep_mean": 371.125,
    "epsilon_norm": 4.334558010101318,
    "prediction_norm": 4.082257270812988,
    "timestamp": "2025-07-14T12:34:14.243558"
  },
  {
    "epoch": 1141,
    "batch_idx": 0,
    "loss": 0.18869829177856445,
    "learning_rate": 0.0003901620309461675,
    "timestep_mean": 545.875,
    "epsilon_norm": 8.046277046203613,
    "prediction_norm": 7.181256294250488,
    "timestamp": "2025-07-14T12:34:14.248830"
  },
  {
    "epoch": 1141,
    "batch_idx": 32,
    "loss": 0.19399990141391754,
    "learning_rate": 0.0003901620309461675,
    "timestep_mean": 525.4375,
    "epsilon_norm": 8.914575576782227,
    "prediction_norm": 8.016498565673828,
    "timestamp": "2025-07-14T12:34:14.253982"
  },
  {
    "epoch": 1141,
    "batch_idx": 64,
    "loss": 0.26308473944664,
    "learning_rate": 0.0003901620309461675,
    "timestep_mean": 533.0,
    "epsilon_norm": 8.132157325744629,
    "prediction_norm": 6.9078593254089355,
    "timestamp": "2025-07-14T12:34:14.259275"
  },
  {
    "epoch": 1141,
    "batch_idx": 96,
    "loss": 0.3992539346218109,
    "learning_rate": 0.0003901620309461675,
    "timestep_mean": 478.5625,
    "epsilon_norm": 8.355016708374023,
    "prediction_norm": 7.060328483581543,
    "timestamp": "2025-07-14T12:34:14.264463"
  },
  {
    "epoch": 1141,
    "batch_idx": 128,
    "loss": 0.26425451040267944,
    "learning_rate": 0.0003901620309461675,
    "timestep_mean": 439.0,
    "epsilon_norm": 8.475561141967773,
    "prediction_norm": 7.529911041259766,
    "timestamp": "2025-07-14T12:34:14.269666"
  },
  {
    "epoch": 1141,
    "batch_idx": 160,
    "loss": 0.3168759346008301,
    "learning_rate": 0.0003901620309461675,
    "timestep_mean": 577.09375,
    "epsilon_norm": 8.041716575622559,
    "prediction_norm": 6.647580623626709,
    "timestamp": "2025-07-14T12:34:14.274847"
  },
  {
    "epoch": 1141,
    "batch_idx": 192,
    "loss": 0.09322121739387512,
    "learning_rate": 0.0003901620309461675,
    "timestep_mean": 552.875,
    "epsilon_norm": 2.528040885925293,
    "prediction_norm": 2.536400318145752,
    "timestamp": "2025-07-14T12:34:14.279966"
  },
  {
    "epoch": 1142,
    "batch_idx": 0,
    "loss": 0.24845640361309052,
    "learning_rate": 0.00038939595360487644,
    "timestep_mean": 490.84375,
    "epsilon_norm": 8.06585693359375,
    "prediction_norm": 7.574098110198975,
    "timestamp": "2025-07-14T12:34:14.285194"
  },
  {
    "epoch": 1142,
    "batch_idx": 32,
    "loss": 0.24710695445537567,
    "learning_rate": 0.00038939595360487644,
    "timestep_mean": 522.03125,
    "epsilon_norm": 9.54500961303711,
    "prediction_norm": 8.437070846557617,
    "timestamp": "2025-07-14T12:34:14.290490"
  },
  {
    "epoch": 1142,
    "batch_idx": 64,
    "loss": 0.22760778665542603,
    "learning_rate": 0.00038939595360487644,
    "timestep_mean": 561.46875,
    "epsilon_norm": 7.867049217224121,
    "prediction_norm": 7.53903865814209,
    "timestamp": "2025-07-14T12:34:14.295648"
  },
  {
    "epoch": 1142,
    "batch_idx": 96,
    "loss": 0.15810713171958923,
    "learning_rate": 0.00038939595360487644,
    "timestep_mean": 503.125,
    "epsilon_norm": 7.410882472991943,
    "prediction_norm": 7.40782356262207,
    "timestamp": "2025-07-14T12:34:14.300909"
  },
  {
    "epoch": 1142,
    "batch_idx": 128,
    "loss": 0.2141558974981308,
    "learning_rate": 0.00038939595360487644,
    "timestep_mean": 538.4375,
    "epsilon_norm": 7.209786415100098,
    "prediction_norm": 6.776214599609375,
    "timestamp": "2025-07-14T12:34:14.305998"
  },
  {
    "epoch": 1142,
    "batch_idx": 160,
    "loss": 0.1892009824514389,
    "learning_rate": 0.00038939595360487644,
    "timestep_mean": 445.0,
    "epsilon_norm": 8.186197280883789,
    "prediction_norm": 7.7199907302856445,
    "timestamp": "2025-07-14T12:34:14.311009"
  },
  {
    "epoch": 1142,
    "batch_idx": 192,
    "loss": 0.08169180154800415,
    "learning_rate": 0.00038939595360487644,
    "timestep_mean": 586.75,
    "epsilon_norm": 3.8381640911102295,
    "prediction_norm": 4.168459892272949,
    "timestamp": "2025-07-14T12:34:14.316105"
  },
  {
    "epoch": 1143,
    "batch_idx": 0,
    "loss": 0.232985720038414,
    "learning_rate": 0.00038863014916807486,
    "timestep_mean": 442.9375,
    "epsilon_norm": 7.3144941329956055,
    "prediction_norm": 6.526333332061768,
    "timestamp": "2025-07-14T12:34:14.321200"
  },
  {
    "epoch": 1143,
    "batch_idx": 32,
    "loss": 0.4374391734600067,
    "learning_rate": 0.00038863014916807486,
    "timestep_mean": 478.15625,
    "epsilon_norm": 7.861785888671875,
    "prediction_norm": 6.4033522605896,
    "timestamp": "2025-07-14T12:34:14.326282"
  },
  {
    "epoch": 1143,
    "batch_idx": 64,
    "loss": 0.1744692623615265,
    "learning_rate": 0.00038863014916807486,
    "timestep_mean": 490.59375,
    "epsilon_norm": 7.9345479011535645,
    "prediction_norm": 7.250197410583496,
    "timestamp": "2025-07-14T12:34:14.331430"
  },
  {
    "epoch": 1143,
    "batch_idx": 96,
    "loss": 0.2365788221359253,
    "learning_rate": 0.00038863014916807486,
    "timestep_mean": 521.5,
    "epsilon_norm": 7.419851303100586,
    "prediction_norm": 6.516036033630371,
    "timestamp": "2025-07-14T12:34:14.336502"
  },
  {
    "epoch": 1143,
    "batch_idx": 128,
    "loss": 0.21178996562957764,
    "learning_rate": 0.00038863014916807486,
    "timestep_mean": 529.59375,
    "epsilon_norm": 7.661927223205566,
    "prediction_norm": 7.135593414306641,
    "timestamp": "2025-07-14T12:34:14.341904"
  },
  {
    "epoch": 1143,
    "batch_idx": 160,
    "loss": 0.17030610144138336,
    "learning_rate": 0.00038863014916807486,
    "timestep_mean": 461.8125,
    "epsilon_norm": 7.530954837799072,
    "prediction_norm": 6.443047523498535,
    "timestamp": "2025-07-14T12:34:14.347193"
  },
  {
    "epoch": 1143,
    "batch_idx": 192,
    "loss": 0.08188464492559433,
    "learning_rate": 0.00038863014916807486,
    "timestep_mean": 724.125,
    "epsilon_norm": 3.597212553024292,
    "prediction_norm": 2.993896722793579,
    "timestamp": "2025-07-14T12:34:14.352270"
  },
  {
    "epoch": 1144,
    "batch_idx": 0,
    "loss": 0.40201854705810547,
    "learning_rate": 0.0003878646195253092,
    "timestep_mean": 493.84375,
    "epsilon_norm": 8.975985527038574,
    "prediction_norm": 6.690441608428955,
    "timestamp": "2025-07-14T12:34:14.357518"
  },
  {
    "epoch": 1144,
    "batch_idx": 32,
    "loss": 0.2951720952987671,
    "learning_rate": 0.0003878646195253092,
    "timestep_mean": 559.90625,
    "epsilon_norm": 7.411529541015625,
    "prediction_norm": 6.128789901733398,
    "timestamp": "2025-07-14T12:34:14.362662"
  },
  {
    "epoch": 1144,
    "batch_idx": 64,
    "loss": 0.23332113027572632,
    "learning_rate": 0.0003878646195253092,
    "timestep_mean": 413.4375,
    "epsilon_norm": 6.627084255218506,
    "prediction_norm": 5.223755836486816,
    "timestamp": "2025-07-14T12:34:14.367872"
  },
  {
    "epoch": 1144,
    "batch_idx": 96,
    "loss": 0.25809377431869507,
    "learning_rate": 0.0003878646195253092,
    "timestep_mean": 525.78125,
    "epsilon_norm": 7.736645698547363,
    "prediction_norm": 6.38984489440918,
    "timestamp": "2025-07-14T12:34:14.373014"
  },
  {
    "epoch": 1144,
    "batch_idx": 128,
    "loss": 0.27608051896095276,
    "learning_rate": 0.0003878646195253092,
    "timestep_mean": 441.875,
    "epsilon_norm": 8.280862808227539,
    "prediction_norm": 6.902055740356445,
    "timestamp": "2025-07-14T12:34:14.378195"
  },
  {
    "epoch": 1144,
    "batch_idx": 160,
    "loss": 0.28083786368370056,
    "learning_rate": 0.0003878646195253092,
    "timestep_mean": 500.71875,
    "epsilon_norm": 8.397833824157715,
    "prediction_norm": 6.832579135894775,
    "timestamp": "2025-07-14T12:34:14.383382"
  },
  {
    "epoch": 1144,
    "batch_idx": 192,
    "loss": 0.42811331152915955,
    "learning_rate": 0.0003878646195253092,
    "timestep_mean": 532.0,
    "epsilon_norm": 3.9073002338409424,
    "prediction_norm": 2.7337486743927,
    "timestamp": "2025-07-14T12:34:14.388555"
  },
  {
    "epoch": 1145,
    "batch_idx": 0,
    "loss": 0.21367381513118744,
    "learning_rate": 0.00038709936656544804,
    "timestep_mean": 406.4375,
    "epsilon_norm": 7.956206321716309,
    "prediction_norm": 6.271746635437012,
    "timestamp": "2025-07-14T12:34:14.393682"
  },
  {
    "epoch": 1145,
    "batch_idx": 32,
    "loss": 0.29745301604270935,
    "learning_rate": 0.00038709936656544804,
    "timestep_mean": 544.1875,
    "epsilon_norm": 8.382450103759766,
    "prediction_norm": 7.018435955047607,
    "timestamp": "2025-07-14T12:34:14.398908"
  },
  {
    "epoch": 1145,
    "batch_idx": 64,
    "loss": 0.4132973551750183,
    "learning_rate": 0.00038709936656544804,
    "timestep_mean": 494.71875,
    "epsilon_norm": 7.97046422958374,
    "prediction_norm": 6.663322925567627,
    "timestamp": "2025-07-14T12:34:14.404080"
  },
  {
    "epoch": 1145,
    "batch_idx": 96,
    "loss": 0.3050330579280853,
    "learning_rate": 0.00038709936656544804,
    "timestep_mean": 471.3125,
    "epsilon_norm": 8.11947250366211,
    "prediction_norm": 7.0681257247924805,
    "timestamp": "2025-07-14T12:34:14.409115"
  },
  {
    "epoch": 1145,
    "batch_idx": 128,
    "loss": 0.13329169154167175,
    "learning_rate": 0.00038709936656544804,
    "timestep_mean": 561.34375,
    "epsilon_norm": 6.530550956726074,
    "prediction_norm": 5.898603439331055,
    "timestamp": "2025-07-14T12:34:14.413938"
  },
  {
    "epoch": 1145,
    "batch_idx": 160,
    "loss": 0.5124356746673584,
    "learning_rate": 0.00038709936656544804,
    "timestep_mean": 490.0625,
    "epsilon_norm": 9.738018035888672,
    "prediction_norm": 7.397817611694336,
    "timestamp": "2025-07-14T12:34:14.419225"
  },
  {
    "epoch": 1145,
    "batch_idx": 192,
    "loss": 0.050976675003767014,
    "learning_rate": 0.00038709936656544804,
    "timestep_mean": 466.5,
    "epsilon_norm": 4.853281021118164,
    "prediction_norm": 4.901108264923096,
    "timestamp": "2025-07-14T12:34:14.424441"
  },
  {
    "epoch": 1146,
    "batch_idx": 0,
    "loss": 0.22654405236244202,
    "learning_rate": 0.0003863343921766767,
    "timestep_mean": 475.8125,
    "epsilon_norm": 7.910593032836914,
    "prediction_norm": 6.424256801605225,
    "timestamp": "2025-07-14T12:34:14.429766"
  },
  {
    "epoch": 1146,
    "batch_idx": 32,
    "loss": 0.34810033440589905,
    "learning_rate": 0.0003863343921766767,
    "timestep_mean": 482.71875,
    "epsilon_norm": 7.100918292999268,
    "prediction_norm": 5.453718662261963,
    "timestamp": "2025-07-14T12:34:14.435235"
  },
  {
    "epoch": 1146,
    "batch_idx": 64,
    "loss": 0.27775174379348755,
    "learning_rate": 0.0003863343921766767,
    "timestep_mean": 515.4375,
    "epsilon_norm": 8.49486255645752,
    "prediction_norm": 7.536690711975098,
    "timestamp": "2025-07-14T12:34:14.440444"
  },
  {
    "epoch": 1146,
    "batch_idx": 96,
    "loss": 0.3222845792770386,
    "learning_rate": 0.0003863343921766767,
    "timestep_mean": 365.1875,
    "epsilon_norm": 7.093249320983887,
    "prediction_norm": 5.304594039916992,
    "timestamp": "2025-07-14T12:34:14.445590"
  },
  {
    "epoch": 1146,
    "batch_idx": 128,
    "loss": 0.22015400230884552,
    "learning_rate": 0.0003863343921766767,
    "timestep_mean": 499.53125,
    "epsilon_norm": 7.263319969177246,
    "prediction_norm": 6.208366394042969,
    "timestamp": "2025-07-14T12:34:14.450990"
  },
  {
    "epoch": 1146,
    "batch_idx": 160,
    "loss": 0.4386720657348633,
    "learning_rate": 0.0003863343921766767,
    "timestep_mean": 463.15625,
    "epsilon_norm": 8.587862014770508,
    "prediction_norm": 6.524398326873779,
    "timestamp": "2025-07-14T12:34:14.456133"
  },
  {
    "epoch": 1146,
    "batch_idx": 192,
    "loss": 0.11682041734457016,
    "learning_rate": 0.0003863343921766767,
    "timestep_mean": 583.875,
    "epsilon_norm": 3.5680954456329346,
    "prediction_norm": 3.163735866546631,
    "timestamp": "2025-07-14T12:34:14.461141"
  },
  {
    "epoch": 1147,
    "batch_idx": 0,
    "loss": 0.2616308927536011,
    "learning_rate": 0.00038556969824649344,
    "timestep_mean": 522.5625,
    "epsilon_norm": 7.653915882110596,
    "prediction_norm": 7.032606601715088,
    "timestamp": "2025-07-14T12:34:14.466194"
  },
  {
    "epoch": 1147,
    "batch_idx": 32,
    "loss": 0.2183685600757599,
    "learning_rate": 0.00038556969824649344,
    "timestep_mean": 448.6875,
    "epsilon_norm": 7.7225871086120605,
    "prediction_norm": 6.881021976470947,
    "timestamp": "2025-07-14T12:34:14.471707"
  },
  {
    "epoch": 1147,
    "batch_idx": 64,
    "loss": 0.19022870063781738,
    "learning_rate": 0.00038556969824649344,
    "timestep_mean": 479.5625,
    "epsilon_norm": 7.1367645263671875,
    "prediction_norm": 6.535770893096924,
    "timestamp": "2025-07-14T12:34:14.476802"
  },
  {
    "epoch": 1147,
    "batch_idx": 96,
    "loss": 0.31287768483161926,
    "learning_rate": 0.00038556969824649344,
    "timestep_mean": 572.25,
    "epsilon_norm": 7.783660411834717,
    "prediction_norm": 6.1686811447143555,
    "timestamp": "2025-07-14T12:34:14.481683"
  },
  {
    "epoch": 1147,
    "batch_idx": 128,
    "loss": 0.16141179203987122,
    "learning_rate": 0.00038556969824649344,
    "timestep_mean": 529.15625,
    "epsilon_norm": 8.00202465057373,
    "prediction_norm": 7.998108863830566,
    "timestamp": "2025-07-14T12:34:14.486609"
  },
  {
    "epoch": 1147,
    "batch_idx": 160,
    "loss": 0.327404648065567,
    "learning_rate": 0.00038556969824649344,
    "timestep_mean": 432.375,
    "epsilon_norm": 8.670903205871582,
    "prediction_norm": 7.289681434631348,
    "timestamp": "2025-07-14T12:34:14.491478"
  },
  {
    "epoch": 1147,
    "batch_idx": 192,
    "loss": 0.28483253717422485,
    "learning_rate": 0.00038556969824649344,
    "timestep_mean": 672.875,
    "epsilon_norm": 5.344793319702148,
    "prediction_norm": 5.226555347442627,
    "timestamp": "2025-07-14T12:34:14.496386"
  },
  {
    "epoch": 1148,
    "batch_idx": 0,
    "loss": 0.18807922303676605,
    "learning_rate": 0.00038480528666170463,
    "timestep_mean": 538.21875,
    "epsilon_norm": 6.759009838104248,
    "prediction_norm": 6.397653579711914,
    "timestamp": "2025-07-14T12:34:14.501299"
  },
  {
    "epoch": 1148,
    "batch_idx": 32,
    "loss": 0.2236679196357727,
    "learning_rate": 0.00038480528666170463,
    "timestep_mean": 494.875,
    "epsilon_norm": 6.994873523712158,
    "prediction_norm": 6.188251495361328,
    "timestamp": "2025-07-14T12:34:14.506222"
  },
  {
    "epoch": 1148,
    "batch_idx": 64,
    "loss": 0.6385575532913208,
    "learning_rate": 0.00038480528666170463,
    "timestep_mean": 435.25,
    "epsilon_norm": 8.515135765075684,
    "prediction_norm": 5.923656463623047,
    "timestamp": "2025-07-14T12:34:14.511145"
  },
  {
    "epoch": 1148,
    "batch_idx": 96,
    "loss": 0.19531238079071045,
    "learning_rate": 0.00038480528666170463,
    "timestep_mean": 570.75,
    "epsilon_norm": 7.677871227264404,
    "prediction_norm": 6.671882629394531,
    "timestamp": "2025-07-14T12:34:14.516010"
  },
  {
    "epoch": 1148,
    "batch_idx": 128,
    "loss": 0.3773643970489502,
    "learning_rate": 0.00038480528666170463,
    "timestep_mean": 562.3125,
    "epsilon_norm": 8.075915336608887,
    "prediction_norm": 6.758361339569092,
    "timestamp": "2025-07-14T12:34:14.520865"
  },
  {
    "epoch": 1148,
    "batch_idx": 160,
    "loss": 0.3336881697177887,
    "learning_rate": 0.00038480528666170463,
    "timestep_mean": 456.125,
    "epsilon_norm": 9.235779762268066,
    "prediction_norm": 7.34541654586792,
    "timestamp": "2025-07-14T12:34:14.525809"
  },
  {
    "epoch": 1148,
    "batch_idx": 192,
    "loss": 0.7540422677993774,
    "learning_rate": 0.00038480528666170463,
    "timestep_mean": 441.625,
    "epsilon_norm": 4.097268104553223,
    "prediction_norm": 3.3684980869293213,
    "timestamp": "2025-07-14T12:34:14.530697"
  },
  {
    "epoch": 1149,
    "batch_idx": 0,
    "loss": 0.17254400253295898,
    "learning_rate": 0.0003840411593084198,
    "timestep_mean": 518.46875,
    "epsilon_norm": 7.4440083503723145,
    "prediction_norm": 6.554245948791504,
    "timestamp": "2025-07-14T12:34:14.535593"
  },
  {
    "epoch": 1149,
    "batch_idx": 32,
    "loss": 0.4424597918987274,
    "learning_rate": 0.0003840411593084198,
    "timestep_mean": 541.5625,
    "epsilon_norm": 8.284396171569824,
    "prediction_norm": 6.347774028778076,
    "timestamp": "2025-07-14T12:34:14.540451"
  },
  {
    "epoch": 1149,
    "batch_idx": 64,
    "loss": 0.20879754424095154,
    "learning_rate": 0.0003840411593084198,
    "timestep_mean": 559.875,
    "epsilon_norm": 8.358979225158691,
    "prediction_norm": 7.199143886566162,
    "timestamp": "2025-07-14T12:34:14.545339"
  },
  {
    "epoch": 1149,
    "batch_idx": 96,
    "loss": 0.3420718312263489,
    "learning_rate": 0.0003840411593084198,
    "timestep_mean": 589.125,
    "epsilon_norm": 9.25540542602539,
    "prediction_norm": 8.431989669799805,
    "timestamp": "2025-07-14T12:34:14.550431"
  },
  {
    "epoch": 1149,
    "batch_idx": 128,
    "loss": 0.16834288835525513,
    "learning_rate": 0.0003840411593084198,
    "timestep_mean": 581.15625,
    "epsilon_norm": 6.747078895568848,
    "prediction_norm": 6.029935359954834,
    "timestamp": "2025-07-14T12:34:14.555278"
  },
  {
    "epoch": 1149,
    "batch_idx": 160,
    "loss": 0.16410675644874573,
    "learning_rate": 0.0003840411593084198,
    "timestep_mean": 527.375,
    "epsilon_norm": 6.90925931930542,
    "prediction_norm": 5.856637477874756,
    "timestamp": "2025-07-14T12:34:14.560215"
  },
  {
    "epoch": 1149,
    "batch_idx": 192,
    "loss": 0.2086797058582306,
    "learning_rate": 0.0003840411593084198,
    "timestep_mean": 582.625,
    "epsilon_norm": 4.032726764678955,
    "prediction_norm": 3.2738356590270996,
    "timestamp": "2025-07-14T12:34:14.565077"
  },
  {
    "epoch": 1150,
    "batch_idx": 0,
    "loss": 0.5729899406433105,
    "learning_rate": 0.00038327731807204717,
    "timestep_mean": 423.0,
    "epsilon_norm": 8.698664665222168,
    "prediction_norm": 6.142075538635254,
    "timestamp": "2025-07-14T12:34:14.569954"
  },
  {
    "epoch": 1150,
    "batch_idx": 32,
    "loss": 0.16962534189224243,
    "learning_rate": 0.00038327731807204717,
    "timestep_mean": 574.1875,
    "epsilon_norm": 7.442669868469238,
    "prediction_norm": 6.86381196975708,
    "timestamp": "2025-07-14T12:34:14.574790"
  },
  {
    "epoch": 1150,
    "batch_idx": 64,
    "loss": 0.31980615854263306,
    "learning_rate": 0.00038327731807204717,
    "timestep_mean": 521.65625,
    "epsilon_norm": 8.363436698913574,
    "prediction_norm": 6.66786003112793,
    "timestamp": "2025-07-14T12:34:14.579780"
  },
  {
    "epoch": 1150,
    "batch_idx": 96,
    "loss": 0.27563387155532837,
    "learning_rate": 0.00038327731807204717,
    "timestep_mean": 477.25,
    "epsilon_norm": 7.940129280090332,
    "prediction_norm": 6.388766288757324,
    "timestamp": "2025-07-14T12:34:14.584632"
  },
  {
    "epoch": 1150,
    "batch_idx": 128,
    "loss": 0.36016321182250977,
    "learning_rate": 0.00038327731807204717,
    "timestep_mean": 474.34375,
    "epsilon_norm": 8.375988006591797,
    "prediction_norm": 7.49703311920166,
    "timestamp": "2025-07-14T12:34:14.589563"
  },
  {
    "epoch": 1150,
    "batch_idx": 160,
    "loss": 0.2103731781244278,
    "learning_rate": 0.00038327731807204717,
    "timestep_mean": 448.5625,
    "epsilon_norm": 6.737645149230957,
    "prediction_norm": 5.089021682739258,
    "timestamp": "2025-07-14T12:34:14.594486"
  },
  {
    "epoch": 1150,
    "batch_idx": 192,
    "loss": 0.14047051966190338,
    "learning_rate": 0.00038327731807204717,
    "timestep_mean": 403.5,
    "epsilon_norm": 4.496184349060059,
    "prediction_norm": 4.465927600860596,
    "timestamp": "2025-07-14T12:34:14.599280"
  },
  {
    "epoch": 1151,
    "batch_idx": 0,
    "loss": 0.4060859680175781,
    "learning_rate": 0.0003825137648372891,
    "timestep_mean": 400.78125,
    "epsilon_norm": 7.383345127105713,
    "prediction_norm": 5.642061233520508,
    "timestamp": "2025-07-14T12:34:14.604199"
  },
  {
    "epoch": 1151,
    "batch_idx": 32,
    "loss": 0.3035600185394287,
    "learning_rate": 0.0003825137648372891,
    "timestep_mean": 411.78125,
    "epsilon_norm": 8.534214973449707,
    "prediction_norm": 7.237525463104248,
    "timestamp": "2025-07-14T12:34:14.609111"
  },
  {
    "epoch": 1151,
    "batch_idx": 64,
    "loss": 0.4030056297779083,
    "learning_rate": 0.0003825137648372891,
    "timestep_mean": 451.21875,
    "epsilon_norm": 7.593627452850342,
    "prediction_norm": 5.814691543579102,
    "timestamp": "2025-07-14T12:34:14.614105"
  },
  {
    "epoch": 1151,
    "batch_idx": 96,
    "loss": 0.1957854926586151,
    "learning_rate": 0.0003825137648372891,
    "timestep_mean": 562.9375,
    "epsilon_norm": 9.609109878540039,
    "prediction_norm": 8.82343578338623,
    "timestamp": "2025-07-14T12:34:14.619005"
  },
  {
    "epoch": 1151,
    "batch_idx": 128,
    "loss": 0.3066522479057312,
    "learning_rate": 0.0003825137648372891,
    "timestep_mean": 470.15625,
    "epsilon_norm": 7.143108367919922,
    "prediction_norm": 5.256006717681885,
    "timestamp": "2025-07-14T12:34:14.623804"
  },
  {
    "epoch": 1151,
    "batch_idx": 160,
    "loss": 0.24340690672397614,
    "learning_rate": 0.0003825137648372891,
    "timestep_mean": 534.9375,
    "epsilon_norm": 7.847781181335449,
    "prediction_norm": 6.223819732666016,
    "timestamp": "2025-07-14T12:34:14.628692"
  },
  {
    "epoch": 1151,
    "batch_idx": 192,
    "loss": 0.11130045354366302,
    "learning_rate": 0.0003825137648372891,
    "timestep_mean": 465.625,
    "epsilon_norm": 3.401416063308716,
    "prediction_norm": 2.960383892059326,
    "timestamp": "2025-07-14T12:34:14.633561"
  },
  {
    "epoch": 1152,
    "batch_idx": 0,
    "loss": 0.38765817880630493,
    "learning_rate": 0.0003817505014881375,
    "timestep_mean": 501.84375,
    "epsilon_norm": 8.102508544921875,
    "prediction_norm": 6.520224094390869,
    "timestamp": "2025-07-14T12:34:14.638441"
  },
  {
    "epoch": 1152,
    "batch_idx": 32,
    "loss": 0.21342581510543823,
    "learning_rate": 0.0003817505014881375,
    "timestep_mean": 573.25,
    "epsilon_norm": 7.331770420074463,
    "prediction_norm": 6.631172180175781,
    "timestamp": "2025-07-14T12:34:14.643296"
  },
  {
    "epoch": 1152,
    "batch_idx": 64,
    "loss": 0.2360561192035675,
    "learning_rate": 0.0003817505014881375,
    "timestep_mean": 522.25,
    "epsilon_norm": 8.240038871765137,
    "prediction_norm": 7.108082294464111,
    "timestamp": "2025-07-14T12:34:14.648204"
  },
  {
    "epoch": 1152,
    "batch_idx": 96,
    "loss": 0.155246302485466,
    "learning_rate": 0.0003817505014881375,
    "timestep_mean": 493.15625,
    "epsilon_norm": 7.865317344665527,
    "prediction_norm": 7.127927303314209,
    "timestamp": "2025-07-14T12:34:14.653147"
  },
  {
    "epoch": 1152,
    "batch_idx": 128,
    "loss": 0.32506752014160156,
    "learning_rate": 0.0003817505014881375,
    "timestep_mean": 486.84375,
    "epsilon_norm": 7.567896842956543,
    "prediction_norm": 6.408529758453369,
    "timestamp": "2025-07-14T12:34:14.658036"
  },
  {
    "epoch": 1152,
    "batch_idx": 160,
    "loss": 0.2418176680803299,
    "learning_rate": 0.0003817505014881375,
    "timestep_mean": 477.3125,
    "epsilon_norm": 8.410438537597656,
    "prediction_norm": 7.369794845581055,
    "timestamp": "2025-07-14T12:34:14.663021"
  },
  {
    "epoch": 1152,
    "batch_idx": 192,
    "loss": 0.2665313482284546,
    "learning_rate": 0.0003817505014881375,
    "timestep_mean": 394.875,
    "epsilon_norm": 4.355729103088379,
    "prediction_norm": 3.5333476066589355,
    "timestamp": "2025-07-14T12:34:14.667823"
  },
  {
    "epoch": 1153,
    "batch_idx": 0,
    "loss": 0.3440323770046234,
    "learning_rate": 0.0003809875299078686,
    "timestep_mean": 476.75,
    "epsilon_norm": 8.80195140838623,
    "prediction_norm": 6.9502458572387695,
    "timestamp": "2025-07-14T12:34:14.672707"
  },
  {
    "epoch": 1153,
    "batch_idx": 32,
    "loss": 0.3988019824028015,
    "learning_rate": 0.0003809875299078686,
    "timestep_mean": 422.0,
    "epsilon_norm": 9.403347969055176,
    "prediction_norm": 7.401435375213623,
    "timestamp": "2025-07-14T12:34:14.677625"
  },
  {
    "epoch": 1153,
    "batch_idx": 64,
    "loss": 0.302129328250885,
    "learning_rate": 0.0003809875299078686,
    "timestep_mean": 623.78125,
    "epsilon_norm": 7.922424793243408,
    "prediction_norm": 6.990270137786865,
    "timestamp": "2025-07-14T12:34:14.682509"
  },
  {
    "epoch": 1153,
    "batch_idx": 96,
    "loss": 0.14823958277702332,
    "learning_rate": 0.0003809875299078686,
    "timestep_mean": 486.9375,
    "epsilon_norm": 6.821291923522949,
    "prediction_norm": 6.115530490875244,
    "timestamp": "2025-07-14T12:34:14.687366"
  },
  {
    "epoch": 1153,
    "batch_idx": 128,
    "loss": 0.35763728618621826,
    "learning_rate": 0.0003809875299078686,
    "timestep_mean": 567.9375,
    "epsilon_norm": 9.18839168548584,
    "prediction_norm": 7.727714538574219,
    "timestamp": "2025-07-14T12:34:14.692283"
  },
  {
    "epoch": 1153,
    "batch_idx": 160,
    "loss": 0.30078256130218506,
    "learning_rate": 0.0003809875299078686,
    "timestep_mean": 498.09375,
    "epsilon_norm": 8.068182945251465,
    "prediction_norm": 6.990318298339844,
    "timestamp": "2025-07-14T12:34:14.697125"
  },
  {
    "epoch": 1153,
    "batch_idx": 192,
    "loss": 0.19068704545497894,
    "learning_rate": 0.0003809875299078686,
    "timestep_mean": 531.375,
    "epsilon_norm": 3.835707187652588,
    "prediction_norm": 3.0708870887756348,
    "timestamp": "2025-07-14T12:34:14.701955"
  },
  {
    "epoch": 1154,
    "batch_idx": 0,
    "loss": 0.3216322064399719,
    "learning_rate": 0.000380224851979039,
    "timestep_mean": 469.15625,
    "epsilon_norm": 9.371986389160156,
    "prediction_norm": 8.284706115722656,
    "timestamp": "2025-07-14T12:34:14.707033"
  },
  {
    "epoch": 1154,
    "batch_idx": 32,
    "loss": 0.18764951825141907,
    "learning_rate": 0.000380224851979039,
    "timestep_mean": 562.34375,
    "epsilon_norm": 6.607174873352051,
    "prediction_norm": 6.1893486976623535,
    "timestamp": "2025-07-14T12:34:14.712162"
  },
  {
    "epoch": 1154,
    "batch_idx": 64,
    "loss": 0.41554737091064453,
    "learning_rate": 0.000380224851979039,
    "timestep_mean": 450.1875,
    "epsilon_norm": 9.457783699035645,
    "prediction_norm": 8.040592193603516,
    "timestamp": "2025-07-14T12:34:14.717311"
  },
  {
    "epoch": 1154,
    "batch_idx": 96,
    "loss": 0.20345866680145264,
    "learning_rate": 0.000380224851979039,
    "timestep_mean": 516.8125,
    "epsilon_norm": 7.76812219619751,
    "prediction_norm": 7.614262104034424,
    "timestamp": "2025-07-14T12:34:14.722591"
  },
  {
    "epoch": 1154,
    "batch_idx": 128,
    "loss": 0.2039438784122467,
    "learning_rate": 0.000380224851979039,
    "timestep_mean": 574.15625,
    "epsilon_norm": 7.261051654815674,
    "prediction_norm": 6.79917049407959,
    "timestamp": "2025-07-14T12:34:14.727704"
  },
  {
    "epoch": 1154,
    "batch_idx": 160,
    "loss": 0.2173319011926651,
    "learning_rate": 0.000380224851979039,
    "timestep_mean": 499.09375,
    "epsilon_norm": 8.068732261657715,
    "prediction_norm": 7.596053600311279,
    "timestamp": "2025-07-14T12:34:14.732853"
  },
  {
    "epoch": 1154,
    "batch_idx": 192,
    "loss": 0.4562463164329529,
    "learning_rate": 0.000380224851979039,
    "timestep_mean": 314.0,
    "epsilon_norm": 3.4983720779418945,
    "prediction_norm": 3.3647043704986572,
    "timestamp": "2025-07-14T12:34:14.737980"
  },
  {
    "epoch": 1155,
    "batch_idx": 0,
    "loss": 0.3590245842933655,
    "learning_rate": 0.0003794624695834806,
    "timestep_mean": 390.0,
    "epsilon_norm": 8.30148983001709,
    "prediction_norm": 6.296774864196777,
    "timestamp": "2025-07-14T12:34:14.743200"
  },
  {
    "epoch": 1155,
    "batch_idx": 32,
    "loss": 0.15711212158203125,
    "learning_rate": 0.0003794624695834806,
    "timestep_mean": 550.6875,
    "epsilon_norm": 7.312657833099365,
    "prediction_norm": 6.617916584014893,
    "timestamp": "2025-07-14T12:34:14.748398"
  },
  {
    "epoch": 1155,
    "batch_idx": 64,
    "loss": 0.09508886933326721,
    "learning_rate": 0.0003794624695834806,
    "timestep_mean": 577.59375,
    "epsilon_norm": 8.388264656066895,
    "prediction_norm": 8.060280799865723,
    "timestamp": "2025-07-14T12:34:14.753566"
  },
  {
    "epoch": 1155,
    "batch_idx": 96,
    "loss": 0.3663327097892761,
    "learning_rate": 0.0003794624695834806,
    "timestep_mean": 400.5,
    "epsilon_norm": 8.005738258361816,
    "prediction_norm": 7.023094177246094,
    "timestamp": "2025-07-14T12:34:14.758860"
  },
  {
    "epoch": 1155,
    "batch_idx": 128,
    "loss": 0.2918069362640381,
    "learning_rate": 0.0003794624695834806,
    "timestep_mean": 454.40625,
    "epsilon_norm": 8.17298412322998,
    "prediction_norm": 6.936602592468262,
    "timestamp": "2025-07-14T12:34:14.764083"
  },
  {
    "epoch": 1155,
    "batch_idx": 160,
    "loss": 0.2092195451259613,
    "learning_rate": 0.0003794624695834806,
    "timestep_mean": 559.78125,
    "epsilon_norm": 7.770327568054199,
    "prediction_norm": 6.601092338562012,
    "timestamp": "2025-07-14T12:34:14.769258"
  },
  {
    "epoch": 1155,
    "batch_idx": 192,
    "loss": 0.46014177799224854,
    "learning_rate": 0.0003794624695834806,
    "timestep_mean": 420.75,
    "epsilon_norm": 3.9376018047332764,
    "prediction_norm": 3.3658335208892822,
    "timestamp": "2025-07-14T12:34:14.774369"
  },
  {
    "epoch": 1156,
    "batch_idx": 0,
    "loss": 0.26745593547821045,
    "learning_rate": 0.0003787003846022962,
    "timestep_mean": 490.03125,
    "epsilon_norm": 7.477935791015625,
    "prediction_norm": 5.985622882843018,
    "timestamp": "2025-07-14T12:34:14.779801"
  },
  {
    "epoch": 1156,
    "batch_idx": 32,
    "loss": 0.313441663980484,
    "learning_rate": 0.0003787003846022962,
    "timestep_mean": 587.75,
    "epsilon_norm": 7.730723857879639,
    "prediction_norm": 6.694818019866943,
    "timestamp": "2025-07-14T12:34:14.784987"
  },
  {
    "epoch": 1156,
    "batch_idx": 64,
    "loss": 0.3207705616950989,
    "learning_rate": 0.0003787003846022962,
    "timestep_mean": 510.5,
    "epsilon_norm": 8.123260498046875,
    "prediction_norm": 6.192878723144531,
    "timestamp": "2025-07-14T12:34:14.790179"
  },
  {
    "epoch": 1156,
    "batch_idx": 96,
    "loss": 0.18814261257648468,
    "learning_rate": 0.0003787003846022962,
    "timestep_mean": 522.34375,
    "epsilon_norm": 8.00999641418457,
    "prediction_norm": 7.357501983642578,
    "timestamp": "2025-07-14T12:34:14.795421"
  },
  {
    "epoch": 1156,
    "batch_idx": 128,
    "loss": 0.3478522002696991,
    "learning_rate": 0.0003787003846022962,
    "timestep_mean": 494.4375,
    "epsilon_norm": 7.899311542510986,
    "prediction_norm": 6.278475761413574,
    "timestamp": "2025-07-14T12:34:14.800810"
  },
  {
    "epoch": 1156,
    "batch_idx": 160,
    "loss": 0.15001024305820465,
    "learning_rate": 0.0003787003846022962,
    "timestep_mean": 530.03125,
    "epsilon_norm": 7.5016632080078125,
    "prediction_norm": 7.01080322265625,
    "timestamp": "2025-07-14T12:34:14.805991"
  },
  {
    "epoch": 1156,
    "batch_idx": 192,
    "loss": 0.18037086725234985,
    "learning_rate": 0.0003787003846022962,
    "timestep_mean": 427.25,
    "epsilon_norm": 4.7528076171875,
    "prediction_norm": 4.316948890686035,
    "timestamp": "2025-07-14T12:34:14.811133"
  },
  {
    "epoch": 1157,
    "batch_idx": 0,
    "loss": 0.1795876920223236,
    "learning_rate": 0.0003779385989158548,
    "timestep_mean": 494.25,
    "epsilon_norm": 7.818473815917969,
    "prediction_norm": 7.202820777893066,
    "timestamp": "2025-07-14T12:34:14.816403"
  },
  {
    "epoch": 1157,
    "batch_idx": 32,
    "loss": 0.2423555850982666,
    "learning_rate": 0.0003779385989158548,
    "timestep_mean": 436.21875,
    "epsilon_norm": 6.760908603668213,
    "prediction_norm": 5.063647270202637,
    "timestamp": "2025-07-14T12:34:14.821763"
  },
  {
    "epoch": 1157,
    "batch_idx": 64,
    "loss": 0.1504235863685608,
    "learning_rate": 0.0003779385989158548,
    "timestep_mean": 534.59375,
    "epsilon_norm": 7.303770542144775,
    "prediction_norm": 6.653841495513916,
    "timestamp": "2025-07-14T12:34:14.826970"
  },
  {
    "epoch": 1157,
    "batch_idx": 96,
    "loss": 0.45462971925735474,
    "learning_rate": 0.0003779385989158548,
    "timestep_mean": 437.0625,
    "epsilon_norm": 8.643369674682617,
    "prediction_norm": 6.318875312805176,
    "timestamp": "2025-07-14T12:34:14.832135"
  },
  {
    "epoch": 1157,
    "batch_idx": 128,
    "loss": 0.3286336362361908,
    "learning_rate": 0.0003779385989158548,
    "timestep_mean": 449.8125,
    "epsilon_norm": 7.507889270782471,
    "prediction_norm": 5.518478870391846,
    "timestamp": "2025-07-14T12:34:14.837400"
  },
  {
    "epoch": 1157,
    "batch_idx": 160,
    "loss": 0.26466166973114014,
    "learning_rate": 0.0003779385989158548,
    "timestep_mean": 498.0,
    "epsilon_norm": 7.6722331047058105,
    "prediction_norm": 6.169321060180664,
    "timestamp": "2025-07-14T12:34:14.842581"
  },
  {
    "epoch": 1157,
    "batch_idx": 192,
    "loss": 0.06877017021179199,
    "learning_rate": 0.0003779385989158548,
    "timestep_mean": 628.125,
    "epsilon_norm": 3.4133150577545166,
    "prediction_norm": 3.6210544109344482,
    "timestamp": "2025-07-14T12:34:14.847654"
  },
  {
    "epoch": 1158,
    "batch_idx": 0,
    "loss": 0.2783866226673126,
    "learning_rate": 0.0003771771144037867,
    "timestep_mean": 462.40625,
    "epsilon_norm": 7.869954586029053,
    "prediction_norm": 6.370975017547607,
    "timestamp": "2025-07-14T12:34:14.852902"
  },
  {
    "epoch": 1158,
    "batch_idx": 32,
    "loss": 0.21827831864356995,
    "learning_rate": 0.0003771771144037867,
    "timestep_mean": 542.53125,
    "epsilon_norm": 7.686077117919922,
    "prediction_norm": 6.636558532714844,
    "timestamp": "2025-07-14T12:34:14.858007"
  },
  {
    "epoch": 1158,
    "batch_idx": 64,
    "loss": 0.2518496513366699,
    "learning_rate": 0.0003771771144037867,
    "timestep_mean": 464.4375,
    "epsilon_norm": 9.053055763244629,
    "prediction_norm": 6.989895343780518,
    "timestamp": "2025-07-14T12:34:14.863281"
  },
  {
    "epoch": 1158,
    "batch_idx": 96,
    "loss": 0.4058186411857605,
    "learning_rate": 0.0003771771144037867,
    "timestep_mean": 468.90625,
    "epsilon_norm": 8.571215629577637,
    "prediction_norm": 6.42489767074585,
    "timestamp": "2025-07-14T12:34:14.868583"
  },
  {
    "epoch": 1158,
    "batch_idx": 128,
    "loss": 0.31645041704177856,
    "learning_rate": 0.0003771771144037867,
    "timestep_mean": 482.125,
    "epsilon_norm": 7.759493827819824,
    "prediction_norm": 6.2921295166015625,
    "timestamp": "2025-07-14T12:34:14.873887"
  },
  {
    "epoch": 1158,
    "batch_idx": 160,
    "loss": 0.20120897889137268,
    "learning_rate": 0.0003771771144037867,
    "timestep_mean": 495.9375,
    "epsilon_norm": 8.522573471069336,
    "prediction_norm": 7.414970874786377,
    "timestamp": "2025-07-14T12:34:14.879003"
  },
  {
    "epoch": 1158,
    "batch_idx": 192,
    "loss": 0.08453258872032166,
    "learning_rate": 0.0003771771144037867,
    "timestep_mean": 576.75,
    "epsilon_norm": 2.718513250350952,
    "prediction_norm": 2.7689619064331055,
    "timestamp": "2025-07-14T12:34:14.884309"
  },
  {
    "epoch": 1159,
    "batch_idx": 0,
    "loss": 0.523608922958374,
    "learning_rate": 0.00037641593294497947,
    "timestep_mean": 429.8125,
    "epsilon_norm": 7.806474685668945,
    "prediction_norm": 5.840147972106934,
    "timestamp": "2025-07-14T12:34:14.889535"
  },
  {
    "epoch": 1159,
    "batch_idx": 32,
    "loss": 0.3797464668750763,
    "learning_rate": 0.00037641593294497947,
    "timestep_mean": 462.71875,
    "epsilon_norm": 7.422323703765869,
    "prediction_norm": 4.968911647796631,
    "timestamp": "2025-07-14T12:34:14.894701"
  },
  {
    "epoch": 1159,
    "batch_idx": 64,
    "loss": 0.26662248373031616,
    "learning_rate": 0.00037641593294497947,
    "timestep_mean": 488.59375,
    "epsilon_norm": 7.889953136444092,
    "prediction_norm": 6.547708034515381,
    "timestamp": "2025-07-14T12:34:14.900025"
  },
  {
    "epoch": 1159,
    "batch_idx": 96,
    "loss": 0.15691126883029938,
    "learning_rate": 0.00037641593294497947,
    "timestep_mean": 537.84375,
    "epsilon_norm": 7.62263298034668,
    "prediction_norm": 6.769969463348389,
    "timestamp": "2025-07-14T12:34:14.905177"
  },
  {
    "epoch": 1159,
    "batch_idx": 128,
    "loss": 0.26388269662857056,
    "learning_rate": 0.00037641593294497947,
    "timestep_mean": 488.53125,
    "epsilon_norm": 7.477722644805908,
    "prediction_norm": 6.207688808441162,
    "timestamp": "2025-07-14T12:34:14.910414"
  },
  {
    "epoch": 1159,
    "batch_idx": 160,
    "loss": 0.31632596254348755,
    "learning_rate": 0.00037641593294497947,
    "timestep_mean": 460.59375,
    "epsilon_norm": 7.897005558013916,
    "prediction_norm": 6.747742652893066,
    "timestamp": "2025-07-14T12:34:14.915610"
  },
  {
    "epoch": 1159,
    "batch_idx": 192,
    "loss": 0.24631713330745697,
    "learning_rate": 0.00037641593294497947,
    "timestep_mean": 500.75,
    "epsilon_norm": 3.2153680324554443,
    "prediction_norm": 3.3836140632629395,
    "timestamp": "2025-07-14T12:34:14.920727"
  },
  {
    "epoch": 1160,
    "batch_idx": 0,
    "loss": 0.10707458853721619,
    "learning_rate": 0.00037565505641757257,
    "timestep_mean": 543.5,
    "epsilon_norm": 6.5943450927734375,
    "prediction_norm": 6.227200984954834,
    "timestamp": "2025-07-14T12:34:14.925928"
  },
  {
    "epoch": 1160,
    "batch_idx": 32,
    "loss": 0.2833768129348755,
    "learning_rate": 0.00037565505641757257,
    "timestep_mean": 451.84375,
    "epsilon_norm": 8.099137306213379,
    "prediction_norm": 6.731204032897949,
    "timestamp": "2025-07-14T12:34:14.931132"
  },
  {
    "epoch": 1160,
    "batch_idx": 64,
    "loss": 0.47496891021728516,
    "learning_rate": 0.00037565505641757257,
    "timestep_mean": 457.375,
    "epsilon_norm": 8.540112495422363,
    "prediction_norm": 6.424280166625977,
    "timestamp": "2025-07-14T12:34:14.936159"
  },
  {
    "epoch": 1160,
    "batch_idx": 96,
    "loss": 0.2219245880842209,
    "learning_rate": 0.00037565505641757257,
    "timestep_mean": 496.875,
    "epsilon_norm": 7.955824375152588,
    "prediction_norm": 7.063666820526123,
    "timestamp": "2025-07-14T12:34:14.941272"
  },
  {
    "epoch": 1160,
    "batch_idx": 128,
    "loss": 0.26806214451789856,
    "learning_rate": 0.00037565505641757257,
    "timestep_mean": 406.96875,
    "epsilon_norm": 6.693153381347656,
    "prediction_norm": 5.352443695068359,
    "timestamp": "2025-07-14T12:34:14.946573"
  },
  {
    "epoch": 1160,
    "batch_idx": 160,
    "loss": 0.3096066117286682,
    "learning_rate": 0.00037565505641757257,
    "timestep_mean": 446.625,
    "epsilon_norm": 8.092846870422363,
    "prediction_norm": 7.4165239334106445,
    "timestamp": "2025-07-14T12:34:14.951815"
  },
  {
    "epoch": 1160,
    "batch_idx": 192,
    "loss": 0.7639074325561523,
    "learning_rate": 0.00037565505641757257,
    "timestep_mean": 291.75,
    "epsilon_norm": 4.921382427215576,
    "prediction_norm": 3.6604621410369873,
    "timestamp": "2025-07-14T12:34:14.956873"
  },
  {
    "epoch": 1161,
    "batch_idx": 0,
    "loss": 0.19255989789962769,
    "learning_rate": 0.0003748944866989532,
    "timestep_mean": 570.0,
    "epsilon_norm": 8.529291152954102,
    "prediction_norm": 7.900486946105957,
    "timestamp": "2025-07-14T12:34:14.962086"
  },
  {
    "epoch": 1161,
    "batch_idx": 32,
    "loss": 0.3884194791316986,
    "learning_rate": 0.0003748944866989532,
    "timestep_mean": 530.28125,
    "epsilon_norm": 8.486254692077637,
    "prediction_norm": 7.376425266265869,
    "timestamp": "2025-07-14T12:34:14.967318"
  },
  {
    "epoch": 1161,
    "batch_idx": 64,
    "loss": 0.13050533831119537,
    "learning_rate": 0.0003748944866989532,
    "timestep_mean": 567.46875,
    "epsilon_norm": 7.682697296142578,
    "prediction_norm": 7.104706287384033,
    "timestamp": "2025-07-14T12:34:14.972223"
  },
  {
    "epoch": 1161,
    "batch_idx": 96,
    "loss": 0.28144571185112,
    "learning_rate": 0.0003748944866989532,
    "timestep_mean": 523.4375,
    "epsilon_norm": 7.712867736816406,
    "prediction_norm": 6.477614879608154,
    "timestamp": "2025-07-14T12:34:14.977161"
  },
  {
    "epoch": 1161,
    "batch_idx": 128,
    "loss": 0.254993736743927,
    "learning_rate": 0.0003748944866989532,
    "timestep_mean": 495.3125,
    "epsilon_norm": 8.33382511138916,
    "prediction_norm": 7.385308265686035,
    "timestamp": "2025-07-14T12:34:14.982345"
  },
  {
    "epoch": 1161,
    "batch_idx": 160,
    "loss": 0.17352184653282166,
    "learning_rate": 0.0003748944866989532,
    "timestep_mean": 488.78125,
    "epsilon_norm": 7.627002239227295,
    "prediction_norm": 6.477113723754883,
    "timestamp": "2025-07-14T12:34:14.987404"
  },
  {
    "epoch": 1161,
    "batch_idx": 192,
    "loss": 0.25706803798675537,
    "learning_rate": 0.0003748944866989532,
    "timestep_mean": 393.25,
    "epsilon_norm": 4.517874717712402,
    "prediction_norm": 3.856553077697754,
    "timestamp": "2025-07-14T12:34:14.992492"
  },
  {
    "epoch": 1162,
    "batch_idx": 0,
    "loss": 0.2566543221473694,
    "learning_rate": 0.00037413422566575126,
    "timestep_mean": 529.96875,
    "epsilon_norm": 7.752859115600586,
    "prediction_norm": 6.5198588371276855,
    "timestamp": "2025-07-14T12:34:14.997359"
  },
  {
    "epoch": 1162,
    "batch_idx": 32,
    "loss": 0.260018527507782,
    "learning_rate": 0.00037413422566575126,
    "timestep_mean": 460.46875,
    "epsilon_norm": 6.650444984436035,
    "prediction_norm": 5.444871425628662,
    "timestamp": "2025-07-14T12:34:15.002211"
  },
  {
    "epoch": 1162,
    "batch_idx": 64,
    "loss": 0.25454092025756836,
    "learning_rate": 0.00037413422566575126,
    "timestep_mean": 567.96875,
    "epsilon_norm": 7.3632121086120605,
    "prediction_norm": 6.255521774291992,
    "timestamp": "2025-07-14T12:34:15.007134"
  },
  {
    "epoch": 1162,
    "batch_idx": 96,
    "loss": 0.4205375015735626,
    "learning_rate": 0.00037413422566575126,
    "timestep_mean": 446.875,
    "epsilon_norm": 8.647560119628906,
    "prediction_norm": 6.669998645782471,
    "timestamp": "2025-07-14T12:34:15.012059"
  },
  {
    "epoch": 1162,
    "batch_idx": 128,
    "loss": 0.3196253776550293,
    "learning_rate": 0.00037413422566575126,
    "timestep_mean": 448.96875,
    "epsilon_norm": 7.041217803955078,
    "prediction_norm": 5.405387878417969,
    "timestamp": "2025-07-14T12:34:15.016874"
  },
  {
    "epoch": 1162,
    "batch_idx": 160,
    "loss": 0.4278240501880646,
    "learning_rate": 0.00037413422566575126,
    "timestep_mean": 465.5625,
    "epsilon_norm": 8.875619888305664,
    "prediction_norm": 7.0638604164123535,
    "timestamp": "2025-07-14T12:34:15.021719"
  },
  {
    "epoch": 1162,
    "batch_idx": 192,
    "loss": 0.2276744395494461,
    "learning_rate": 0.00037413422566575126,
    "timestep_mean": 585.375,
    "epsilon_norm": 3.5477640628814697,
    "prediction_norm": 3.073437213897705,
    "timestamp": "2025-07-14T12:34:15.026646"
  },
  {
    "epoch": 1163,
    "batch_idx": 0,
    "loss": 0.24177426099777222,
    "learning_rate": 0.0003733742751938358,
    "timestep_mean": 505.5,
    "epsilon_norm": 8.068416595458984,
    "prediction_norm": 7.303555011749268,
    "timestamp": "2025-07-14T12:34:15.031493"
  },
  {
    "epoch": 1163,
    "batch_idx": 32,
    "loss": 0.28868696093559265,
    "learning_rate": 0.0003733742751938358,
    "timestep_mean": 509.34375,
    "epsilon_norm": 7.762164115905762,
    "prediction_norm": 5.9594550132751465,
    "timestamp": "2025-07-14T12:34:15.036375"
  },
  {
    "epoch": 1163,
    "batch_idx": 64,
    "loss": 0.4121153950691223,
    "learning_rate": 0.0003733742751938358,
    "timestep_mean": 510.0,
    "epsilon_norm": 7.766068458557129,
    "prediction_norm": 5.874405860900879,
    "timestamp": "2025-07-14T12:34:15.041502"
  },
  {
    "epoch": 1163,
    "batch_idx": 96,
    "loss": 0.20993967354297638,
    "learning_rate": 0.0003733742751938358,
    "timestep_mean": 615.0,
    "epsilon_norm": 7.260091781616211,
    "prediction_norm": 6.455121040344238,
    "timestamp": "2025-07-14T12:34:15.046464"
  },
  {
    "epoch": 1163,
    "batch_idx": 128,
    "loss": 0.11332565546035767,
    "learning_rate": 0.0003733742751938358,
    "timestep_mean": 642.9375,
    "epsilon_norm": 7.1343817710876465,
    "prediction_norm": 6.85920524597168,
    "timestamp": "2025-07-14T12:34:15.051264"
  },
  {
    "epoch": 1163,
    "batch_idx": 160,
    "loss": 0.22035744786262512,
    "learning_rate": 0.0003733742751938358,
    "timestep_mean": 504.40625,
    "epsilon_norm": 7.801674842834473,
    "prediction_norm": 6.494076728820801,
    "timestamp": "2025-07-14T12:34:15.056215"
  },
  {
    "epoch": 1163,
    "batch_idx": 192,
    "loss": 0.07452236860990524,
    "learning_rate": 0.0003733742751938358,
    "timestep_mean": 513.75,
    "epsilon_norm": 2.951511859893799,
    "prediction_norm": 3.008646249771118,
    "timestamp": "2025-07-14T12:34:15.061101"
  },
  {
    "epoch": 1164,
    "batch_idx": 0,
    "loss": 0.41031426191329956,
    "learning_rate": 0.0003726146371583089,
    "timestep_mean": 481.4375,
    "epsilon_norm": 7.501769542694092,
    "prediction_norm": 6.108876705169678,
    "timestamp": "2025-07-14T12:34:15.065993"
  },
  {
    "epoch": 1164,
    "batch_idx": 32,
    "loss": 0.2814125418663025,
    "learning_rate": 0.0003726146371583089,
    "timestep_mean": 476.21875,
    "epsilon_norm": 7.428208827972412,
    "prediction_norm": 5.376571178436279,
    "timestamp": "2025-07-14T12:34:15.070852"
  },
  {
    "epoch": 1164,
    "batch_idx": 64,
    "loss": 0.36239922046661377,
    "learning_rate": 0.0003726146371583089,
    "timestep_mean": 502.25,
    "epsilon_norm": 8.057524681091309,
    "prediction_norm": 5.99418306350708,
    "timestamp": "2025-07-14T12:34:15.075801"
  },
  {
    "epoch": 1164,
    "batch_idx": 96,
    "loss": 0.1719822883605957,
    "learning_rate": 0.0003726146371583089,
    "timestep_mean": 462.125,
    "epsilon_norm": 8.193902969360352,
    "prediction_norm": 6.989330291748047,
    "timestamp": "2025-07-14T12:34:15.080615"
  },
  {
    "epoch": 1164,
    "batch_idx": 128,
    "loss": 0.1258084625005722,
    "learning_rate": 0.0003726146371583089,
    "timestep_mean": 568.03125,
    "epsilon_norm": 6.30218505859375,
    "prediction_norm": 6.091116428375244,
    "timestamp": "2025-07-14T12:34:15.085434"
  },
  {
    "epoch": 1164,
    "batch_idx": 160,
    "loss": 0.1121244728565216,
    "learning_rate": 0.0003726146371583089,
    "timestep_mean": 663.0,
    "epsilon_norm": 7.312097072601318,
    "prediction_norm": 6.957888603210449,
    "timestamp": "2025-07-14T12:34:15.090302"
  },
  {
    "epoch": 1164,
    "batch_idx": 192,
    "loss": 0.7181318402290344,
    "learning_rate": 0.0003726146371583089,
    "timestep_mean": 445.625,
    "epsilon_norm": 4.765989780426025,
    "prediction_norm": 3.0352113246917725,
    "timestamp": "2025-07-14T12:34:15.095162"
  },
  {
    "epoch": 1165,
    "batch_idx": 0,
    "loss": 0.2338804453611374,
    "learning_rate": 0.0003718553134335016,
    "timestep_mean": 454.6875,
    "epsilon_norm": 8.079606056213379,
    "prediction_norm": 6.708800315856934,
    "timestamp": "2025-07-14T12:34:15.100112"
  },
  {
    "epoch": 1165,
    "batch_idx": 32,
    "loss": 0.3824872374534607,
    "learning_rate": 0.0003718553134335016,
    "timestep_mean": 448.4375,
    "epsilon_norm": 9.07978343963623,
    "prediction_norm": 6.790061950683594,
    "timestamp": "2025-07-14T12:34:15.104997"
  },
  {
    "epoch": 1165,
    "batch_idx": 64,
    "loss": 0.5791258215904236,
    "learning_rate": 0.0003718553134335016,
    "timestep_mean": 453.9375,
    "epsilon_norm": 8.63414478302002,
    "prediction_norm": 5.863049030303955,
    "timestamp": "2025-07-14T12:34:15.109940"
  },
  {
    "epoch": 1165,
    "batch_idx": 96,
    "loss": 0.42032742500305176,
    "learning_rate": 0.0003718553134335016,
    "timestep_mean": 555.0,
    "epsilon_norm": 8.106568336486816,
    "prediction_norm": 6.788999557495117,
    "timestamp": "2025-07-14T12:34:15.114843"
  },
  {
    "epoch": 1165,
    "batch_idx": 128,
    "loss": 0.34994232654571533,
    "learning_rate": 0.0003718553134335016,
    "timestep_mean": 590.28125,
    "epsilon_norm": 8.466361045837402,
    "prediction_norm": 7.029148578643799,
    "timestamp": "2025-07-14T12:34:15.119749"
  },
  {
    "epoch": 1165,
    "batch_idx": 160,
    "loss": 0.5001243352890015,
    "learning_rate": 0.0003718553134335016,
    "timestep_mean": 433.25,
    "epsilon_norm": 8.703786849975586,
    "prediction_norm": 7.445461273193359,
    "timestamp": "2025-07-14T12:34:15.124642"
  },
  {
    "epoch": 1165,
    "batch_idx": 192,
    "loss": 0.3809875249862671,
    "learning_rate": 0.0003718553134335016,
    "timestep_mean": 421.5,
    "epsilon_norm": 2.771277666091919,
    "prediction_norm": 2.237064838409424,
    "timestamp": "2025-07-14T12:34:15.129606"
  },
  {
    "epoch": 1166,
    "batch_idx": 0,
    "loss": 0.0964217334985733,
    "learning_rate": 0.00037109630589297004,
    "timestep_mean": 603.125,
    "epsilon_norm": 7.364906311035156,
    "prediction_norm": 7.127715110778809,
    "timestamp": "2025-07-14T12:34:15.134485"
  },
  {
    "epoch": 1166,
    "batch_idx": 32,
    "loss": 0.3039824366569519,
    "learning_rate": 0.00037109630589297004,
    "timestep_mean": 483.4375,
    "epsilon_norm": 7.692474842071533,
    "prediction_norm": 6.221745014190674,
    "timestamp": "2025-07-14T12:34:15.139360"
  },
  {
    "epoch": 1166,
    "batch_idx": 64,
    "loss": 0.22448453307151794,
    "learning_rate": 0.00037109630589297004,
    "timestep_mean": 477.09375,
    "epsilon_norm": 8.407580375671387,
    "prediction_norm": 7.336426258087158,
    "timestamp": "2025-07-14T12:34:15.144244"
  },
  {
    "epoch": 1166,
    "batch_idx": 96,
    "loss": 0.4019760191440582,
    "learning_rate": 0.00037109630589297004,
    "timestep_mean": 480.9375,
    "epsilon_norm": 8.710745811462402,
    "prediction_norm": 7.331255912780762,
    "timestamp": "2025-07-14T12:34:15.149131"
  },
  {
    "epoch": 1166,
    "batch_idx": 128,
    "loss": 0.260976105928421,
    "learning_rate": 0.00037109630589297004,
    "timestep_mean": 577.65625,
    "epsilon_norm": 8.296223640441895,
    "prediction_norm": 6.800680160522461,
    "timestamp": "2025-07-14T12:34:15.153976"
  },
  {
    "epoch": 1166,
    "batch_idx": 160,
    "loss": 0.3384503722190857,
    "learning_rate": 0.00037109630589297004,
    "timestep_mean": 471.1875,
    "epsilon_norm": 9.573800086975098,
    "prediction_norm": 8.523879051208496,
    "timestamp": "2025-07-14T12:34:15.158931"
  },
  {
    "epoch": 1166,
    "batch_idx": 192,
    "loss": 0.2358601987361908,
    "learning_rate": 0.00037109630589297004,
    "timestep_mean": 530.125,
    "epsilon_norm": 4.770864963531494,
    "prediction_norm": 4.40138578414917,
    "timestamp": "2025-07-14T12:34:15.163800"
  },
  {
    "epoch": 1167,
    "batch_idx": 0,
    "loss": 0.19311711192131042,
    "learning_rate": 0.00037033761640948974,
    "timestep_mean": 474.5625,
    "epsilon_norm": 7.572341442108154,
    "prediction_norm": 6.55137825012207,
    "timestamp": "2025-07-14T12:34:15.168730"
  },
  {
    "epoch": 1167,
    "batch_idx": 32,
    "loss": 0.19539088010787964,
    "learning_rate": 0.00037033761640948974,
    "timestep_mean": 467.75,
    "epsilon_norm": 7.070761203765869,
    "prediction_norm": 6.150470733642578,
    "timestamp": "2025-07-14T12:34:15.173595"
  },
  {
    "epoch": 1167,
    "batch_idx": 64,
    "loss": 0.47482359409332275,
    "learning_rate": 0.00037033761640948974,
    "timestep_mean": 376.03125,
    "epsilon_norm": 8.223560333251953,
    "prediction_norm": 6.665106773376465,
    "timestamp": "2025-07-14T12:34:15.178461"
  },
  {
    "epoch": 1167,
    "batch_idx": 96,
    "loss": 0.30229049921035767,
    "learning_rate": 0.00037033761640948974,
    "timestep_mean": 546.125,
    "epsilon_norm": 7.683088779449463,
    "prediction_norm": 5.989511966705322,
    "timestamp": "2025-07-14T12:34:15.183364"
  },
  {
    "epoch": 1167,
    "batch_idx": 128,
    "loss": 0.40398919582366943,
    "learning_rate": 0.00037033761640948974,
    "timestep_mean": 519.96875,
    "epsilon_norm": 9.082893371582031,
    "prediction_norm": 6.852935791015625,
    "timestamp": "2025-07-14T12:34:15.188230"
  },
  {
    "epoch": 1167,
    "batch_idx": 160,
    "loss": 0.25666916370391846,
    "learning_rate": 0.00037033761640948974,
    "timestep_mean": 367.21875,
    "epsilon_norm": 8.243302345275879,
    "prediction_norm": 6.931650638580322,
    "timestamp": "2025-07-14T12:34:15.193159"
  },
  {
    "epoch": 1167,
    "batch_idx": 192,
    "loss": 0.06307744234800339,
    "learning_rate": 0.00037033761640948974,
    "timestep_mean": 603.25,
    "epsilon_norm": 3.780977249145508,
    "prediction_norm": 3.4991965293884277,
    "timestamp": "2025-07-14T12:34:15.198123"
  },
  {
    "epoch": 1168,
    "batch_idx": 0,
    "loss": 0.35658544301986694,
    "learning_rate": 0.0003695792468550515,
    "timestep_mean": 446.53125,
    "epsilon_norm": 7.5175461769104,
    "prediction_norm": 6.093222141265869,
    "timestamp": "2025-07-14T12:34:15.203061"
  },
  {
    "epoch": 1168,
    "batch_idx": 32,
    "loss": 0.41357484459877014,
    "learning_rate": 0.0003695792468550515,
    "timestep_mean": 516.625,
    "epsilon_norm": 9.289859771728516,
    "prediction_norm": 7.9161834716796875,
    "timestamp": "2025-07-14T12:34:15.207955"
  },
  {
    "epoch": 1168,
    "batch_idx": 64,
    "loss": 0.2902439534664154,
    "learning_rate": 0.0003695792468550515,
    "timestep_mean": 548.03125,
    "epsilon_norm": 8.226714134216309,
    "prediction_norm": 7.01651668548584,
    "timestamp": "2025-07-14T12:34:15.212961"
  },
  {
    "epoch": 1168,
    "batch_idx": 96,
    "loss": 0.2544638514518738,
    "learning_rate": 0.0003695792468550515,
    "timestep_mean": 486.5,
    "epsilon_norm": 7.057353496551514,
    "prediction_norm": 5.452330112457275,
    "timestamp": "2025-07-14T12:34:15.217863"
  },
  {
    "epoch": 1168,
    "batch_idx": 128,
    "loss": 0.43062785267829895,
    "learning_rate": 0.0003695792468550515,
    "timestep_mean": 482.1875,
    "epsilon_norm": 7.622367858886719,
    "prediction_norm": 5.808813571929932,
    "timestamp": "2025-07-14T12:34:15.222725"
  },
  {
    "epoch": 1168,
    "batch_idx": 160,
    "loss": 0.38741248846054077,
    "learning_rate": 0.0003695792468550515,
    "timestep_mean": 510.0625,
    "epsilon_norm": 7.966275215148926,
    "prediction_norm": 6.562026500701904,
    "timestamp": "2025-07-14T12:34:15.227626"
  },
  {
    "epoch": 1168,
    "batch_idx": 192,
    "loss": 0.7732131481170654,
    "learning_rate": 0.0003695792468550515,
    "timestep_mean": 316.875,
    "epsilon_norm": 4.156354904174805,
    "prediction_norm": 2.782808780670166,
    "timestamp": "2025-07-14T12:34:15.232521"
  },
  {
    "epoch": 1169,
    "batch_idx": 0,
    "loss": 0.4066832661628723,
    "learning_rate": 0.0003688211991008569,
    "timestep_mean": 511.0625,
    "epsilon_norm": 9.28073501586914,
    "prediction_norm": 7.45973539352417,
    "timestamp": "2025-07-14T12:34:15.237405"
  },
  {
    "epoch": 1169,
    "batch_idx": 32,
    "loss": 0.23970812559127808,
    "learning_rate": 0.0003688211991008569,
    "timestep_mean": 437.875,
    "epsilon_norm": 8.177486419677734,
    "prediction_norm": 6.987411022186279,
    "timestamp": "2025-07-14T12:34:15.242414"
  },
  {
    "epoch": 1169,
    "batch_idx": 64,
    "loss": 0.18845383822917938,
    "learning_rate": 0.0003688211991008569,
    "timestep_mean": 511.9375,
    "epsilon_norm": 7.022987365722656,
    "prediction_norm": 6.935873031616211,
    "timestamp": "2025-07-14T12:34:15.247510"
  },
  {
    "epoch": 1169,
    "batch_idx": 96,
    "loss": 0.2834903299808502,
    "learning_rate": 0.0003688211991008569,
    "timestep_mean": 388.625,
    "epsilon_norm": 6.512775421142578,
    "prediction_norm": 5.310985565185547,
    "timestamp": "2025-07-14T12:34:15.252480"
  },
  {
    "epoch": 1169,
    "batch_idx": 128,
    "loss": 0.20906615257263184,
    "learning_rate": 0.0003688211991008569,
    "timestep_mean": 512.90625,
    "epsilon_norm": 6.8717041015625,
    "prediction_norm": 6.117454528808594,
    "timestamp": "2025-07-14T12:34:15.257351"
  },
  {
    "epoch": 1169,
    "batch_idx": 160,
    "loss": 0.30201905965805054,
    "learning_rate": 0.0003688211991008569,
    "timestep_mean": 469.0625,
    "epsilon_norm": 8.041202545166016,
    "prediction_norm": 6.946983337402344,
    "timestamp": "2025-07-14T12:34:15.262233"
  },
  {
    "epoch": 1169,
    "batch_idx": 192,
    "loss": 0.47029751539230347,
    "learning_rate": 0.0003688211991008569,
    "timestep_mean": 502.75,
    "epsilon_norm": 4.362151622772217,
    "prediction_norm": 3.773360013961792,
    "timestamp": "2025-07-14T12:34:15.267283"
  },
  {
    "epoch": 1170,
    "batch_idx": 0,
    "loss": 0.3065817058086395,
    "learning_rate": 0.00036806347501731353,
    "timestep_mean": 473.1875,
    "epsilon_norm": 7.46571683883667,
    "prediction_norm": 6.398195266723633,
    "timestamp": "2025-07-14T12:34:15.272127"
  },
  {
    "epoch": 1170,
    "batch_idx": 32,
    "loss": 0.3273472785949707,
    "learning_rate": 0.00036806347501731353,
    "timestep_mean": 489.3125,
    "epsilon_norm": 6.9105143547058105,
    "prediction_norm": 6.20780086517334,
    "timestamp": "2025-07-14T12:34:15.277051"
  },
  {
    "epoch": 1170,
    "batch_idx": 64,
    "loss": 0.14423143863677979,
    "learning_rate": 0.00036806347501731353,
    "timestep_mean": 502.28125,
    "epsilon_norm": 8.512277603149414,
    "prediction_norm": 8.1990385055542,
    "timestamp": "2025-07-14T12:34:15.282029"
  },
  {
    "epoch": 1170,
    "batch_idx": 96,
    "loss": 0.24475137889385223,
    "learning_rate": 0.00036806347501731353,
    "timestep_mean": 542.25,
    "epsilon_norm": 6.991201400756836,
    "prediction_norm": 6.305687427520752,
    "timestamp": "2025-07-14T12:34:15.286951"
  },
  {
    "epoch": 1170,
    "batch_idx": 128,
    "loss": 0.3742305636405945,
    "learning_rate": 0.00036806347501731353,
    "timestep_mean": 475.1875,
    "epsilon_norm": 8.711308479309082,
    "prediction_norm": 6.95195198059082,
    "timestamp": "2025-07-14T12:34:15.291864"
  },
  {
    "epoch": 1170,
    "batch_idx": 160,
    "loss": 0.39988645911216736,
    "learning_rate": 0.00036806347501731353,
    "timestep_mean": 480.5,
    "epsilon_norm": 8.276619911193848,
    "prediction_norm": 7.1330389976501465,
    "timestamp": "2025-07-14T12:34:15.296790"
  },
  {
    "epoch": 1170,
    "batch_idx": 192,
    "loss": 0.30153733491897583,
    "learning_rate": 0.00036806347501731353,
    "timestep_mean": 556.625,
    "epsilon_norm": 4.774974346160889,
    "prediction_norm": 4.02556037902832,
    "timestamp": "2025-07-14T12:34:15.301675"
  },
  {
    "epoch": 1171,
    "batch_idx": 0,
    "loss": 0.38803398609161377,
    "learning_rate": 0.0003673060764740301,
    "timestep_mean": 549.28125,
    "epsilon_norm": 7.996409893035889,
    "prediction_norm": 6.530611038208008,
    "timestamp": "2025-07-14T12:34:15.306627"
  },
  {
    "epoch": 1171,
    "batch_idx": 32,
    "loss": 0.20375028252601624,
    "learning_rate": 0.0003673060764740301,
    "timestep_mean": 490.5625,
    "epsilon_norm": 7.389925479888916,
    "prediction_norm": 6.093510627746582,
    "timestamp": "2025-07-14T12:34:15.311536"
  },
  {
    "epoch": 1171,
    "batch_idx": 64,
    "loss": 0.25755131244659424,
    "learning_rate": 0.0003673060764740301,
    "timestep_mean": 516.21875,
    "epsilon_norm": 7.217907428741455,
    "prediction_norm": 6.656999588012695,
    "timestamp": "2025-07-14T12:34:15.316631"
  },
  {
    "epoch": 1171,
    "batch_idx": 96,
    "loss": 0.5819158554077148,
    "learning_rate": 0.0003673060764740301,
    "timestep_mean": 487.5625,
    "epsilon_norm": 8.905075073242188,
    "prediction_norm": 6.175559043884277,
    "timestamp": "2025-07-14T12:34:15.321536"
  },
  {
    "epoch": 1171,
    "batch_idx": 128,
    "loss": 0.1655724197626114,
    "learning_rate": 0.0003673060764740301,
    "timestep_mean": 638.03125,
    "epsilon_norm": 7.924004554748535,
    "prediction_norm": 7.217597007751465,
    "timestamp": "2025-07-14T12:34:15.326358"
  },
  {
    "epoch": 1171,
    "batch_idx": 160,
    "loss": 0.22865398228168488,
    "learning_rate": 0.0003673060764740301,
    "timestep_mean": 528.59375,
    "epsilon_norm": 7.399076461791992,
    "prediction_norm": 6.114001274108887,
    "timestamp": "2025-07-14T12:34:15.331274"
  },
  {
    "epoch": 1171,
    "batch_idx": 192,
    "loss": 0.4996805191040039,
    "learning_rate": 0.0003673060764740301,
    "timestep_mean": 340.375,
    "epsilon_norm": 4.663942337036133,
    "prediction_norm": 3.535956621170044,
    "timestamp": "2025-07-14T12:34:15.336200"
  },
  {
    "epoch": 1172,
    "batch_idx": 0,
    "loss": 0.30942779779434204,
    "learning_rate": 0.0003665490053398123,
    "timestep_mean": 544.25,
    "epsilon_norm": 7.91353702545166,
    "prediction_norm": 6.27493143081665,
    "timestamp": "2025-07-14T12:34:15.341126"
  },
  {
    "epoch": 1172,
    "batch_idx": 32,
    "loss": 0.4130093455314636,
    "learning_rate": 0.0003665490053398123,
    "timestep_mean": 414.6875,
    "epsilon_norm": 8.584759712219238,
    "prediction_norm": 6.468516826629639,
    "timestamp": "2025-07-14T12:34:15.346122"
  },
  {
    "epoch": 1172,
    "batch_idx": 64,
    "loss": 0.29663729667663574,
    "learning_rate": 0.0003665490053398123,
    "timestep_mean": 461.625,
    "epsilon_norm": 7.361730575561523,
    "prediction_norm": 5.37470817565918,
    "timestamp": "2025-07-14T12:34:15.350976"
  },
  {
    "epoch": 1172,
    "batch_idx": 96,
    "loss": 0.25385117530822754,
    "learning_rate": 0.0003665490053398123,
    "timestep_mean": 425.6875,
    "epsilon_norm": 8.152271270751953,
    "prediction_norm": 6.107365131378174,
    "timestamp": "2025-07-14T12:34:15.355892"
  },
  {
    "epoch": 1172,
    "batch_idx": 128,
    "loss": 0.29758042097091675,
    "learning_rate": 0.0003665490053398123,
    "timestep_mean": 460.59375,
    "epsilon_norm": 7.338027000427246,
    "prediction_norm": 5.832038879394531,
    "timestamp": "2025-07-14T12:34:15.360800"
  },
  {
    "epoch": 1172,
    "batch_idx": 160,
    "loss": 0.27127495408058167,
    "learning_rate": 0.0003665490053398123,
    "timestep_mean": 493.28125,
    "epsilon_norm": 8.603439331054688,
    "prediction_norm": 7.16029167175293,
    "timestamp": "2025-07-14T12:34:15.365680"
  },
  {
    "epoch": 1172,
    "batch_idx": 192,
    "loss": 0.12136966735124588,
    "learning_rate": 0.0003665490053398123,
    "timestep_mean": 475.375,
    "epsilon_norm": 2.881831169128418,
    "prediction_norm": 2.688385009765625,
    "timestamp": "2025-07-14T12:34:15.370579"
  },
  {
    "epoch": 1173,
    "batch_idx": 0,
    "loss": 0.23700189590454102,
    "learning_rate": 0.0003657922634826577,
    "timestep_mean": 552.8125,
    "epsilon_norm": 8.217239379882812,
    "prediction_norm": 6.913431167602539,
    "timestamp": "2025-07-14T12:34:15.375455"
  },
  {
    "epoch": 1173,
    "batch_idx": 32,
    "loss": 0.3427950143814087,
    "learning_rate": 0.0003657922634826577,
    "timestep_mean": 495.21875,
    "epsilon_norm": 8.273687362670898,
    "prediction_norm": 6.946374893188477,
    "timestamp": "2025-07-14T12:34:15.380354"
  },
  {
    "epoch": 1173,
    "batch_idx": 64,
    "loss": 0.12512625753879547,
    "learning_rate": 0.0003657922634826577,
    "timestep_mean": 598.78125,
    "epsilon_norm": 7.290151119232178,
    "prediction_norm": 6.619612216949463,
    "timestamp": "2025-07-14T12:34:15.385231"
  },
  {
    "epoch": 1173,
    "batch_idx": 96,
    "loss": 0.29165083169937134,
    "learning_rate": 0.0003657922634826577,
    "timestep_mean": 449.75,
    "epsilon_norm": 7.867552280426025,
    "prediction_norm": 6.638175010681152,
    "timestamp": "2025-07-14T12:34:15.390132"
  },
  {
    "epoch": 1173,
    "batch_idx": 128,
    "loss": 0.2164328396320343,
    "learning_rate": 0.0003657922634826577,
    "timestep_mean": 563.34375,
    "epsilon_norm": 7.413206100463867,
    "prediction_norm": 6.417499542236328,
    "timestamp": "2025-07-14T12:34:15.394999"
  },
  {
    "epoch": 1173,
    "batch_idx": 160,
    "loss": 0.1857723593711853,
    "learning_rate": 0.0003657922634826577,
    "timestep_mean": 525.96875,
    "epsilon_norm": 7.503938674926758,
    "prediction_norm": 6.429243087768555,
    "timestamp": "2025-07-14T12:34:15.400017"
  },
  {
    "epoch": 1173,
    "batch_idx": 192,
    "loss": 0.1318226158618927,
    "learning_rate": 0.0003657922634826577,
    "timestep_mean": 389.125,
    "epsilon_norm": 3.6514055728912354,
    "prediction_norm": 3.0141241550445557,
    "timestamp": "2025-07-14T12:34:15.404827"
  },
  {
    "epoch": 1174,
    "batch_idx": 0,
    "loss": 0.31354063749313354,
    "learning_rate": 0.0003650358527697518,
    "timestep_mean": 438.5625,
    "epsilon_norm": 7.631017684936523,
    "prediction_norm": 6.1847991943359375,
    "timestamp": "2025-07-14T12:34:15.409682"
  },
  {
    "epoch": 1174,
    "batch_idx": 32,
    "loss": 0.26398542523384094,
    "learning_rate": 0.0003650358527697518,
    "timestep_mean": 482.78125,
    "epsilon_norm": 7.3705525398254395,
    "prediction_norm": 6.080944061279297,
    "timestamp": "2025-07-14T12:34:15.414615"
  },
  {
    "epoch": 1174,
    "batch_idx": 64,
    "loss": 0.225762277841568,
    "learning_rate": 0.0003650358527697518,
    "timestep_mean": 512.65625,
    "epsilon_norm": 7.212937355041504,
    "prediction_norm": 5.8356242179870605,
    "timestamp": "2025-07-14T12:34:15.419627"
  },
  {
    "epoch": 1174,
    "batch_idx": 96,
    "loss": 0.2594248056411743,
    "learning_rate": 0.0003650358527697518,
    "timestep_mean": 600.15625,
    "epsilon_norm": 8.732860565185547,
    "prediction_norm": 7.524751663208008,
    "timestamp": "2025-07-14T12:34:15.424639"
  },
  {
    "epoch": 1174,
    "batch_idx": 128,
    "loss": 0.352225661277771,
    "learning_rate": 0.0003650358527697518,
    "timestep_mean": 570.84375,
    "epsilon_norm": 7.521166801452637,
    "prediction_norm": 6.427122116088867,
    "timestamp": "2025-07-14T12:34:15.429624"
  },
  {
    "epoch": 1174,
    "batch_idx": 160,
    "loss": 0.26495060324668884,
    "learning_rate": 0.0003650358527697518,
    "timestep_mean": 466.875,
    "epsilon_norm": 7.903219223022461,
    "prediction_norm": 6.720790386199951,
    "timestamp": "2025-07-14T12:34:15.434482"
  },
  {
    "epoch": 1174,
    "batch_idx": 192,
    "loss": 0.18302962183952332,
    "learning_rate": 0.0003650358527697518,
    "timestep_mean": 595.0,
    "epsilon_norm": 3.6738154888153076,
    "prediction_norm": 3.1513984203338623,
    "timestamp": "2025-07-14T12:34:15.439459"
  },
  {
    "epoch": 1175,
    "batch_idx": 0,
    "loss": 0.3208571970462799,
    "learning_rate": 0.00036427977506746286,
    "timestep_mean": 541.84375,
    "epsilon_norm": 6.799456596374512,
    "prediction_norm": 5.6251397132873535,
    "timestamp": "2025-07-14T12:34:15.444574"
  },
  {
    "epoch": 1175,
    "batch_idx": 32,
    "loss": 0.1781989336013794,
    "learning_rate": 0.00036427977506746286,
    "timestep_mean": 533.65625,
    "epsilon_norm": 8.137776374816895,
    "prediction_norm": 7.3536376953125,
    "timestamp": "2025-07-14T12:34:15.449790"
  },
  {
    "epoch": 1175,
    "batch_idx": 64,
    "loss": 0.24164170026779175,
    "learning_rate": 0.00036427977506746286,
    "timestep_mean": 537.28125,
    "epsilon_norm": 7.497895240783691,
    "prediction_norm": 6.35562801361084,
    "timestamp": "2025-07-14T12:34:15.455099"
  },
  {
    "epoch": 1175,
    "batch_idx": 96,
    "loss": 0.1473064422607422,
    "learning_rate": 0.00036427977506746286,
    "timestep_mean": 553.90625,
    "epsilon_norm": 6.123415470123291,
    "prediction_norm": 5.471944332122803,
    "timestamp": "2025-07-14T12:34:15.460244"
  },
  {
    "epoch": 1175,
    "batch_idx": 128,
    "loss": 0.4334864020347595,
    "learning_rate": 0.00036427977506746286,
    "timestep_mean": 431.21875,
    "epsilon_norm": 8.162331581115723,
    "prediction_norm": 6.146798610687256,
    "timestamp": "2025-07-14T12:34:15.465490"
  },
  {
    "epoch": 1175,
    "batch_idx": 160,
    "loss": 0.13460296392440796,
    "learning_rate": 0.00036427977506746286,
    "timestep_mean": 561.84375,
    "epsilon_norm": 8.222442626953125,
    "prediction_norm": 7.444718360900879,
    "timestamp": "2025-07-14T12:34:15.470617"
  },
  {
    "epoch": 1175,
    "batch_idx": 192,
    "loss": 0.020664134994149208,
    "learning_rate": 0.00036427977506746286,
    "timestep_mean": 508.875,
    "epsilon_norm": 4.619518280029297,
    "prediction_norm": 4.41661262512207,
    "timestamp": "2025-07-14T12:34:15.475583"
  },
  {
    "epoch": 1176,
    "batch_idx": 0,
    "loss": 0.459690660238266,
    "learning_rate": 0.0003635240322413373,
    "timestep_mean": 394.90625,
    "epsilon_norm": 8.061210632324219,
    "prediction_norm": 6.379417419433594,
    "timestamp": "2025-07-14T12:34:15.480532"
  },
  {
    "epoch": 1176,
    "batch_idx": 32,
    "loss": 0.21485194563865662,
    "learning_rate": 0.0003635240322413373,
    "timestep_mean": 490.90625,
    "epsilon_norm": 7.806380271911621,
    "prediction_norm": 6.729038715362549,
    "timestamp": "2025-07-14T12:34:15.485662"
  },
  {
    "epoch": 1176,
    "batch_idx": 64,
    "loss": 0.46140581369400024,
    "learning_rate": 0.0003635240322413373,
    "timestep_mean": 470.625,
    "epsilon_norm": 8.519007682800293,
    "prediction_norm": 6.506134986877441,
    "timestamp": "2025-07-14T12:34:15.490898"
  },
  {
    "epoch": 1176,
    "batch_idx": 96,
    "loss": 0.25676870346069336,
    "learning_rate": 0.0003635240322413373,
    "timestep_mean": 485.96875,
    "epsilon_norm": 6.978253364562988,
    "prediction_norm": 5.978945732116699,
    "timestamp": "2025-07-14T12:34:15.496001"
  },
  {
    "epoch": 1176,
    "batch_idx": 128,
    "loss": 0.26113662123680115,
    "learning_rate": 0.0003635240322413373,
    "timestep_mean": 503.71875,
    "epsilon_norm": 7.850125312805176,
    "prediction_norm": 6.473211765289307,
    "timestamp": "2025-07-14T12:34:15.501077"
  },
  {
    "epoch": 1176,
    "batch_idx": 160,
    "loss": 0.13837385177612305,
    "learning_rate": 0.0003635240322413373,
    "timestep_mean": 492.59375,
    "epsilon_norm": 8.057185173034668,
    "prediction_norm": 7.288026809692383,
    "timestamp": "2025-07-14T12:34:15.506267"
  },
  {
    "epoch": 1176,
    "batch_idx": 192,
    "loss": 0.34617888927459717,
    "learning_rate": 0.0003635240322413373,
    "timestep_mean": 434.25,
    "epsilon_norm": 3.4685614109039307,
    "prediction_norm": 2.937897205352783,
    "timestamp": "2025-07-14T12:34:15.511418"
  },
  {
    "epoch": 1177,
    "batch_idx": 0,
    "loss": 0.30291521549224854,
    "learning_rate": 0.0003627686261560957,
    "timestep_mean": 481.25,
    "epsilon_norm": 8.133402824401855,
    "prediction_norm": 6.613412380218506,
    "timestamp": "2025-07-14T12:34:15.516748"
  },
  {
    "epoch": 1177,
    "batch_idx": 32,
    "loss": 0.19470123946666718,
    "learning_rate": 0.0003627686261560957,
    "timestep_mean": 589.3125,
    "epsilon_norm": 7.520481109619141,
    "prediction_norm": 6.725641250610352,
    "timestamp": "2025-07-14T12:34:15.521822"
  },
  {
    "epoch": 1177,
    "batch_idx": 64,
    "loss": 0.1691751629114151,
    "learning_rate": 0.0003627686261560957,
    "timestep_mean": 549.03125,
    "epsilon_norm": 7.210637092590332,
    "prediction_norm": 6.378321170806885,
    "timestamp": "2025-07-14T12:34:15.526967"
  },
  {
    "epoch": 1177,
    "batch_idx": 96,
    "loss": 0.46076393127441406,
    "learning_rate": 0.0003627686261560957,
    "timestep_mean": 425.65625,
    "epsilon_norm": 8.916184425354004,
    "prediction_norm": 6.858921527862549,
    "timestamp": "2025-07-14T12:34:15.532309"
  },
  {
    "epoch": 1177,
    "batch_idx": 128,
    "loss": 0.44248828291893005,
    "learning_rate": 0.0003627686261560957,
    "timestep_mean": 500.28125,
    "epsilon_norm": 8.693214416503906,
    "prediction_norm": 7.401769161224365,
    "timestamp": "2025-07-14T12:34:15.537474"
  },
  {
    "epoch": 1177,
    "batch_idx": 160,
    "loss": 0.24987950921058655,
    "learning_rate": 0.0003627686261560957,
    "timestep_mean": 453.0625,
    "epsilon_norm": 7.174893379211426,
    "prediction_norm": 6.200425148010254,
    "timestamp": "2025-07-14T12:34:15.542677"
  },
  {
    "epoch": 1177,
    "batch_idx": 192,
    "loss": 0.25082314014434814,
    "learning_rate": 0.0003627686261560957,
    "timestep_mean": 343.375,
    "epsilon_norm": 1.9132442474365234,
    "prediction_norm": 1.7921062707901,
    "timestamp": "2025-07-14T12:34:15.547891"
  },
  {
    "epoch": 1178,
    "batch_idx": 0,
    "loss": 0.19893696904182434,
    "learning_rate": 0.00036201355867562725,
    "timestep_mean": 472.53125,
    "epsilon_norm": 6.386943817138672,
    "prediction_norm": 5.232665061950684,
    "timestamp": "2025-07-14T12:34:15.553036"
  },
  {
    "epoch": 1178,
    "batch_idx": 32,
    "loss": 0.25171154737472534,
    "learning_rate": 0.00036201355867562725,
    "timestep_mean": 487.5,
    "epsilon_norm": 8.427906036376953,
    "prediction_norm": 6.875681400299072,
    "timestamp": "2025-07-14T12:34:15.558415"
  },
  {
    "epoch": 1178,
    "batch_idx": 64,
    "loss": 0.2954990863800049,
    "learning_rate": 0.00036201355867562725,
    "timestep_mean": 459.90625,
    "epsilon_norm": 8.381016731262207,
    "prediction_norm": 6.505651473999023,
    "timestamp": "2025-07-14T12:34:15.563602"
  },
  {
    "epoch": 1178,
    "batch_idx": 96,
    "loss": 0.23957456648349762,
    "learning_rate": 0.00036201355867562725,
    "timestep_mean": 491.1875,
    "epsilon_norm": 7.619044303894043,
    "prediction_norm": 6.563478946685791,
    "timestamp": "2025-07-14T12:34:15.568714"
  },
  {
    "epoch": 1178,
    "batch_idx": 128,
    "loss": 0.18830937147140503,
    "learning_rate": 0.00036201355867562725,
    "timestep_mean": 606.5625,
    "epsilon_norm": 8.078731536865234,
    "prediction_norm": 7.223454475402832,
    "timestamp": "2025-07-14T12:34:15.573877"
  },
  {
    "epoch": 1178,
    "batch_idx": 160,
    "loss": 0.2658129930496216,
    "learning_rate": 0.00036201355867562725,
    "timestep_mean": 508.53125,
    "epsilon_norm": 7.131824016571045,
    "prediction_norm": 6.115312099456787,
    "timestamp": "2025-07-14T12:34:15.579073"
  },
  {
    "epoch": 1178,
    "batch_idx": 192,
    "loss": 0.35243648290634155,
    "learning_rate": 0.00036201355867562725,
    "timestep_mean": 332.5,
    "epsilon_norm": 3.5123941898345947,
    "prediction_norm": 2.1971797943115234,
    "timestamp": "2025-07-14T12:34:15.584362"
  },
  {
    "epoch": 1179,
    "batch_idx": 0,
    "loss": 0.44159233570098877,
    "learning_rate": 0.00036125883166298586,
    "timestep_mean": 471.09375,
    "epsilon_norm": 7.464484214782715,
    "prediction_norm": 5.682823657989502,
    "timestamp": "2025-07-14T12:34:15.589492"
  },
  {
    "epoch": 1179,
    "batch_idx": 32,
    "loss": 0.1712890863418579,
    "learning_rate": 0.00036125883166298586,
    "timestep_mean": 530.28125,
    "epsilon_norm": 7.154696941375732,
    "prediction_norm": 6.606939792633057,
    "timestamp": "2025-07-14T12:34:15.594605"
  },
  {
    "epoch": 1179,
    "batch_idx": 64,
    "loss": 0.0789453461766243,
    "learning_rate": 0.00036125883166298586,
    "timestep_mean": 647.40625,
    "epsilon_norm": 7.839488983154297,
    "prediction_norm": 7.719263553619385,
    "timestamp": "2025-07-14T12:34:15.600017"
  },
  {
    "epoch": 1179,
    "batch_idx": 96,
    "loss": 0.10958515107631683,
    "learning_rate": 0.00036125883166298586,
    "timestep_mean": 521.28125,
    "epsilon_norm": 6.649819374084473,
    "prediction_norm": 6.180475234985352,
    "timestamp": "2025-07-14T12:34:15.605304"
  },
  {
    "epoch": 1179,
    "batch_idx": 128,
    "loss": 0.2989153265953064,
    "learning_rate": 0.00036125883166298586,
    "timestep_mean": 499.03125,
    "epsilon_norm": 8.179405212402344,
    "prediction_norm": 7.1670451164245605,
    "timestamp": "2025-07-14T12:34:15.610458"
  },
  {
    "epoch": 1179,
    "batch_idx": 160,
    "loss": 0.3479979336261749,
    "learning_rate": 0.00036125883166298586,
    "timestep_mean": 559.875,
    "epsilon_norm": 8.742907524108887,
    "prediction_norm": 6.954400539398193,
    "timestamp": "2025-07-14T12:34:15.615821"
  },
  {
    "epoch": 1179,
    "batch_idx": 192,
    "loss": 0.30408042669296265,
    "learning_rate": 0.00036125883166298586,
    "timestep_mean": 422.125,
    "epsilon_norm": 3.8934693336486816,
    "prediction_norm": 2.5743918418884277,
    "timestamp": "2025-07-14T12:34:15.621026"
  },
  {
    "epoch": 1180,
    "batch_idx": 0,
    "loss": 0.19807232916355133,
    "learning_rate": 0.00036050444698038526,
    "timestep_mean": 445.5,
    "epsilon_norm": 6.529854774475098,
    "prediction_norm": 5.685409069061279,
    "timestamp": "2025-07-14T12:34:15.626103"
  },
  {
    "epoch": 1180,
    "batch_idx": 32,
    "loss": 0.25339776277542114,
    "learning_rate": 0.00036050444698038526,
    "timestep_mean": 434.9375,
    "epsilon_norm": 8.516393661499023,
    "prediction_norm": 7.546154975891113,
    "timestamp": "2025-07-14T12:34:15.631353"
  },
  {
    "epoch": 1180,
    "batch_idx": 64,
    "loss": 0.2245979607105255,
    "learning_rate": 0.00036050444698038526,
    "timestep_mean": 453.1875,
    "epsilon_norm": 7.729759693145752,
    "prediction_norm": 7.225508213043213,
    "timestamp": "2025-07-14T12:34:15.636530"
  },
  {
    "epoch": 1180,
    "batch_idx": 96,
    "loss": 0.3912833631038666,
    "learning_rate": 0.00036050444698038526,
    "timestep_mean": 409.25,
    "epsilon_norm": 7.750895023345947,
    "prediction_norm": 5.9598259925842285,
    "timestamp": "2025-07-14T12:34:15.641839"
  },
  {
    "epoch": 1180,
    "batch_idx": 128,
    "loss": 0.3426677882671356,
    "learning_rate": 0.00036050444698038526,
    "timestep_mean": 404.3125,
    "epsilon_norm": 6.341436862945557,
    "prediction_norm": 4.935882568359375,
    "timestamp": "2025-07-14T12:34:15.647006"
  },
  {
    "epoch": 1180,
    "batch_idx": 160,
    "loss": 0.33987414836883545,
    "learning_rate": 0.00036050444698038526,
    "timestep_mean": 472.0,
    "epsilon_norm": 8.595949172973633,
    "prediction_norm": 6.435004234313965,
    "timestamp": "2025-07-14T12:34:15.652286"
  },
  {
    "epoch": 1180,
    "batch_idx": 192,
    "loss": 1.0076966285705566,
    "learning_rate": 0.00036050444698038526,
    "timestep_mean": 402.125,
    "epsilon_norm": 4.755548000335693,
    "prediction_norm": 3.435889482498169,
    "timestamp": "2025-07-14T12:34:15.657446"
  },
  {
    "epoch": 1181,
    "batch_idx": 0,
    "loss": 0.10333116352558136,
    "learning_rate": 0.00035975040648919516,
    "timestep_mean": 567.46875,
    "epsilon_norm": 7.233031749725342,
    "prediction_norm": 7.030955791473389,
    "timestamp": "2025-07-14T12:34:15.662597"
  },
  {
    "epoch": 1181,
    "batch_idx": 32,
    "loss": 0.1809222549200058,
    "learning_rate": 0.00035975040648919516,
    "timestep_mean": 585.46875,
    "epsilon_norm": 7.644955635070801,
    "prediction_norm": 7.0247883796691895,
    "timestamp": "2025-07-14T12:34:15.667751"
  },
  {
    "epoch": 1181,
    "batch_idx": 64,
    "loss": 0.2755280137062073,
    "learning_rate": 0.00035975040648919516,
    "timestep_mean": 444.21875,
    "epsilon_norm": 8.138009071350098,
    "prediction_norm": 6.786985874176025,
    "timestamp": "2025-07-14T12:34:15.672938"
  },
  {
    "epoch": 1181,
    "batch_idx": 96,
    "loss": 0.38323286175727844,
    "learning_rate": 0.00035975040648919516,
    "timestep_mean": 411.09375,
    "epsilon_norm": 8.3347806930542,
    "prediction_norm": 6.954845905303955,
    "timestamp": "2025-07-14T12:34:15.678130"
  },
  {
    "epoch": 1181,
    "batch_idx": 128,
    "loss": 0.25638777017593384,
    "learning_rate": 0.00035975040648919516,
    "timestep_mean": 482.46875,
    "epsilon_norm": 8.566028594970703,
    "prediction_norm": 7.5181169509887695,
    "timestamp": "2025-07-14T12:34:15.683430"
  },
  {
    "epoch": 1181,
    "batch_idx": 160,
    "loss": 0.2625959515571594,
    "learning_rate": 0.00035975040648919516,
    "timestep_mean": 525.78125,
    "epsilon_norm": 7.6417012214660645,
    "prediction_norm": 6.902070045471191,
    "timestamp": "2025-07-14T12:34:15.688702"
  },
  {
    "epoch": 1181,
    "batch_idx": 192,
    "loss": 0.26646000146865845,
    "learning_rate": 0.00035975040648919516,
    "timestep_mean": 390.5,
    "epsilon_norm": 4.707700729370117,
    "prediction_norm": 3.7376773357391357,
    "timestamp": "2025-07-14T12:34:15.693932"
  },
  {
    "epoch": 1182,
    "batch_idx": 0,
    "loss": 0.3628047704696655,
    "learning_rate": 0.0003589967120499353,
    "timestep_mean": 411.34375,
    "epsilon_norm": 8.58039379119873,
    "prediction_norm": 6.497225761413574,
    "timestamp": "2025-07-14T12:34:15.699007"
  },
  {
    "epoch": 1182,
    "batch_idx": 32,
    "loss": 0.17790120840072632,
    "learning_rate": 0.0003589967120499353,
    "timestep_mean": 470.28125,
    "epsilon_norm": 7.147416114807129,
    "prediction_norm": 6.324285507202148,
    "timestamp": "2025-07-14T12:34:15.704244"
  },
  {
    "epoch": 1182,
    "batch_idx": 64,
    "loss": 0.3212125897407532,
    "learning_rate": 0.0003589967120499353,
    "timestep_mean": 436.125,
    "epsilon_norm": 9.886059761047363,
    "prediction_norm": 7.940434455871582,
    "timestamp": "2025-07-14T12:34:15.709352"
  },
  {
    "epoch": 1182,
    "batch_idx": 96,
    "loss": 0.20638932287693024,
    "learning_rate": 0.0003589967120499353,
    "timestep_mean": 448.0625,
    "epsilon_norm": 8.735284805297852,
    "prediction_norm": 8.699197769165039,
    "timestamp": "2025-07-14T12:34:15.714574"
  },
  {
    "epoch": 1182,
    "batch_idx": 128,
    "loss": 0.08452323079109192,
    "learning_rate": 0.0003589967120499353,
    "timestep_mean": 532.59375,
    "epsilon_norm": 7.846978187561035,
    "prediction_norm": 7.315117835998535,
    "timestamp": "2025-07-14T12:34:15.719684"
  },
  {
    "epoch": 1182,
    "batch_idx": 160,
    "loss": 0.21708372235298157,
    "learning_rate": 0.0003589967120499353,
    "timestep_mean": 543.3125,
    "epsilon_norm": 8.778016090393066,
    "prediction_norm": 8.089606285095215,
    "timestamp": "2025-07-14T12:34:15.724807"
  },
  {
    "epoch": 1182,
    "batch_idx": 192,
    "loss": 0.12484552711248398,
    "learning_rate": 0.0003589967120499353,
    "timestep_mean": 650.75,
    "epsilon_norm": 4.321878433227539,
    "prediction_norm": 3.5229039192199707,
    "timestamp": "2025-07-14T12:34:15.730085"
  },
  {
    "epoch": 1183,
    "batch_idx": 0,
    "loss": 0.29538118839263916,
    "learning_rate": 0.0003582433655222717,
    "timestep_mean": 500.34375,
    "epsilon_norm": 7.7885308265686035,
    "prediction_norm": 6.799500942230225,
    "timestamp": "2025-07-14T12:34:15.735363"
  },
  {
    "epoch": 1183,
    "batch_idx": 32,
    "loss": 0.3385058641433716,
    "learning_rate": 0.0003582433655222717,
    "timestep_mean": 515.71875,
    "epsilon_norm": 8.011534690856934,
    "prediction_norm": 6.242049217224121,
    "timestamp": "2025-07-14T12:34:15.740435"
  },
  {
    "epoch": 1183,
    "batch_idx": 64,
    "loss": 0.19698473811149597,
    "learning_rate": 0.0003582433655222717,
    "timestep_mean": 521.3125,
    "epsilon_norm": 7.409279823303223,
    "prediction_norm": 6.287320613861084,
    "timestamp": "2025-07-14T12:34:15.745628"
  },
  {
    "epoch": 1183,
    "batch_idx": 96,
    "loss": 0.20898520946502686,
    "learning_rate": 0.0003582433655222717,
    "timestep_mean": 485.65625,
    "epsilon_norm": 7.135379314422607,
    "prediction_norm": 5.855440616607666,
    "timestamp": "2025-07-14T12:34:15.750926"
  },
  {
    "epoch": 1183,
    "batch_idx": 128,
    "loss": 0.1341322958469391,
    "learning_rate": 0.0003582433655222717,
    "timestep_mean": 503.90625,
    "epsilon_norm": 7.2364301681518555,
    "prediction_norm": 6.927794456481934,
    "timestamp": "2025-07-14T12:34:15.756061"
  },
  {
    "epoch": 1183,
    "batch_idx": 160,
    "loss": 0.30061569809913635,
    "learning_rate": 0.0003582433655222717,
    "timestep_mean": 522.71875,
    "epsilon_norm": 8.522115707397461,
    "prediction_norm": 6.9713921546936035,
    "timestamp": "2025-07-14T12:34:15.761318"
  },
  {
    "epoch": 1183,
    "batch_idx": 192,
    "loss": 0.17997035384178162,
    "learning_rate": 0.0003582433655222717,
    "timestep_mean": 428.75,
    "epsilon_norm": 3.734675884246826,
    "prediction_norm": 3.0462422370910645,
    "timestamp": "2025-07-14T12:34:15.766525"
  },
  {
    "epoch": 1184,
    "batch_idx": 0,
    "loss": 0.2725642919540405,
    "learning_rate": 0.00035749036876501193,
    "timestep_mean": 434.1875,
    "epsilon_norm": 8.465812683105469,
    "prediction_norm": 7.023247241973877,
    "timestamp": "2025-07-14T12:34:15.771679"
  },
  {
    "epoch": 1184,
    "batch_idx": 32,
    "loss": 0.21121451258659363,
    "learning_rate": 0.00035749036876501193,
    "timestep_mean": 466.3125,
    "epsilon_norm": 7.277395725250244,
    "prediction_norm": 7.053367614746094,
    "timestamp": "2025-07-14T12:34:15.776866"
  },
  {
    "epoch": 1184,
    "batch_idx": 64,
    "loss": 0.31993234157562256,
    "learning_rate": 0.00035749036876501193,
    "timestep_mean": 509.8125,
    "epsilon_norm": 6.7866950035095215,
    "prediction_norm": 5.873301029205322,
    "timestamp": "2025-07-14T12:34:15.782015"
  },
  {
    "epoch": 1184,
    "batch_idx": 96,
    "loss": 0.30453163385391235,
    "learning_rate": 0.00035749036876501193,
    "timestep_mean": 551.71875,
    "epsilon_norm": 8.066330909729004,
    "prediction_norm": 6.325438976287842,
    "timestamp": "2025-07-14T12:34:15.787173"
  },
  {
    "epoch": 1184,
    "batch_idx": 128,
    "loss": 0.5527340173721313,
    "learning_rate": 0.00035749036876501193,
    "timestep_mean": 448.53125,
    "epsilon_norm": 7.616246223449707,
    "prediction_norm": 5.662799835205078,
    "timestamp": "2025-07-14T12:34:15.792656"
  },
  {
    "epoch": 1184,
    "batch_idx": 160,
    "loss": 0.32286375761032104,
    "learning_rate": 0.00035749036876501193,
    "timestep_mean": 376.21875,
    "epsilon_norm": 8.175020217895508,
    "prediction_norm": 6.469337463378906,
    "timestamp": "2025-07-14T12:34:15.797702"
  },
  {
    "epoch": 1184,
    "batch_idx": 192,
    "loss": 0.08268748223781586,
    "learning_rate": 0.00035749036876501193,
    "timestep_mean": 495.75,
    "epsilon_norm": 2.9194064140319824,
    "prediction_norm": 2.7283921241760254,
    "timestamp": "2025-07-14T12:34:15.802924"
  },
  {
    "epoch": 1185,
    "batch_idx": 0,
    "loss": 0.1522403359413147,
    "learning_rate": 0.0003567377236361009,
    "timestep_mean": 499.21875,
    "epsilon_norm": 7.699370861053467,
    "prediction_norm": 6.958256244659424,
    "timestamp": "2025-07-14T12:34:15.807969"
  },
  {
    "epoch": 1185,
    "batch_idx": 32,
    "loss": 0.26626500487327576,
    "learning_rate": 0.0003567377236361009,
    "timestep_mean": 568.4375,
    "epsilon_norm": 8.682147026062012,
    "prediction_norm": 7.16727876663208,
    "timestamp": "2025-07-14T12:34:15.813105"
  },
  {
    "epoch": 1185,
    "batch_idx": 64,
    "loss": 0.29610949754714966,
    "learning_rate": 0.0003567377236361009,
    "timestep_mean": 423.75,
    "epsilon_norm": 7.059660911560059,
    "prediction_norm": 5.58003044128418,
    "timestamp": "2025-07-14T12:34:15.818339"
  },
  {
    "epoch": 1185,
    "batch_idx": 96,
    "loss": 0.3030385971069336,
    "learning_rate": 0.0003567377236361009,
    "timestep_mean": 397.6875,
    "epsilon_norm": 6.942263603210449,
    "prediction_norm": 5.305019378662109,
    "timestamp": "2025-07-14T12:34:15.823387"
  },
  {
    "epoch": 1185,
    "batch_idx": 128,
    "loss": 0.34297195076942444,
    "learning_rate": 0.0003567377236361009,
    "timestep_mean": 436.15625,
    "epsilon_norm": 6.208421230316162,
    "prediction_norm": 5.015868186950684,
    "timestamp": "2025-07-14T12:34:15.828542"
  },
  {
    "epoch": 1185,
    "batch_idx": 160,
    "loss": 0.31220245361328125,
    "learning_rate": 0.0003567377236361009,
    "timestep_mean": 453.9375,
    "epsilon_norm": 6.834563255310059,
    "prediction_norm": 5.289247035980225,
    "timestamp": "2025-07-14T12:34:15.833781"
  },
  {
    "epoch": 1185,
    "batch_idx": 192,
    "loss": 0.452705979347229,
    "learning_rate": 0.0003567377236361009,
    "timestep_mean": 435.125,
    "epsilon_norm": 3.931983709335327,
    "prediction_norm": 2.5940234661102295,
    "timestamp": "2025-07-14T12:34:15.838922"
  },
  {
    "epoch": 1186,
    "batch_idx": 0,
    "loss": 0.28336483240127563,
    "learning_rate": 0.00035598543199261547,
    "timestep_mean": 462.90625,
    "epsilon_norm": 7.549399375915527,
    "prediction_norm": 6.405792236328125,
    "timestamp": "2025-07-14T12:34:15.844187"
  },
  {
    "epoch": 1186,
    "batch_idx": 32,
    "loss": 0.2623859643936157,
    "learning_rate": 0.00035598543199261547,
    "timestep_mean": 450.75,
    "epsilon_norm": 8.13310432434082,
    "prediction_norm": 7.016626834869385,
    "timestamp": "2025-07-14T12:34:15.849470"
  },
  {
    "epoch": 1186,
    "batch_idx": 64,
    "loss": 0.22943942248821259,
    "learning_rate": 0.00035598543199261547,
    "timestep_mean": 463.1875,
    "epsilon_norm": 7.652708530426025,
    "prediction_norm": 6.251662731170654,
    "timestamp": "2025-07-14T12:34:15.854694"
  },
  {
    "epoch": 1186,
    "batch_idx": 96,
    "loss": 0.23566511273384094,
    "learning_rate": 0.00035598543199261547,
    "timestep_mean": 494.28125,
    "epsilon_norm": 8.968049049377441,
    "prediction_norm": 7.962143421173096,
    "timestamp": "2025-07-14T12:34:15.859848"
  },
  {
    "epoch": 1186,
    "batch_idx": 128,
    "loss": 0.32234668731689453,
    "learning_rate": 0.00035598543199261547,
    "timestep_mean": 484.0625,
    "epsilon_norm": 7.807890892028809,
    "prediction_norm": 5.997128963470459,
    "timestamp": "2025-07-14T12:34:15.865016"
  },
  {
    "epoch": 1186,
    "batch_idx": 160,
    "loss": 0.24576710164546967,
    "learning_rate": 0.00035598543199261547,
    "timestep_mean": 569.5625,
    "epsilon_norm": 7.23258638381958,
    "prediction_norm": 6.722074508666992,
    "timestamp": "2025-07-14T12:34:15.870223"
  },
  {
    "epoch": 1186,
    "batch_idx": 192,
    "loss": 0.2774256467819214,
    "learning_rate": 0.00035598543199261547,
    "timestep_mean": 535.375,
    "epsilon_norm": 4.633397102355957,
    "prediction_norm": 3.594233274459839,
    "timestamp": "2025-07-14T12:34:15.875221"
  },
  {
    "epoch": 1187,
    "batch_idx": 0,
    "loss": 0.21437300741672516,
    "learning_rate": 0.0003552334956907604,
    "timestep_mean": 585.59375,
    "epsilon_norm": 7.986095905303955,
    "prediction_norm": 6.830775260925293,
    "timestamp": "2025-07-14T12:34:15.880387"
  },
  {
    "epoch": 1187,
    "batch_idx": 32,
    "loss": 0.1852286458015442,
    "learning_rate": 0.0003552334956907604,
    "timestep_mean": 579.8125,
    "epsilon_norm": 7.580905437469482,
    "prediction_norm": 6.6672234535217285,
    "timestamp": "2025-07-14T12:34:15.885609"
  },
  {
    "epoch": 1187,
    "batch_idx": 64,
    "loss": 0.23594330251216888,
    "learning_rate": 0.0003552334956907604,
    "timestep_mean": 541.15625,
    "epsilon_norm": 8.939801216125488,
    "prediction_norm": 8.283823013305664,
    "timestamp": "2025-07-14T12:34:15.890868"
  },
  {
    "epoch": 1187,
    "batch_idx": 96,
    "loss": 0.4438285827636719,
    "learning_rate": 0.0003552334956907604,
    "timestep_mean": 356.21875,
    "epsilon_norm": 9.179534912109375,
    "prediction_norm": 6.914984703063965,
    "timestamp": "2025-07-14T12:34:15.895985"
  },
  {
    "epoch": 1187,
    "batch_idx": 128,
    "loss": 0.17693930864334106,
    "learning_rate": 0.0003552334956907604,
    "timestep_mean": 619.09375,
    "epsilon_norm": 8.00247573852539,
    "prediction_norm": 7.512065887451172,
    "timestamp": "2025-07-14T12:34:15.901103"
  },
  {
    "epoch": 1187,
    "batch_idx": 160,
    "loss": 0.27120932936668396,
    "learning_rate": 0.0003552334956907604,
    "timestep_mean": 527.25,
    "epsilon_norm": 8.551826477050781,
    "prediction_norm": 7.3860554695129395,
    "timestamp": "2025-07-14T12:34:15.906309"
  },
  {
    "epoch": 1187,
    "batch_idx": 192,
    "loss": 0.24968385696411133,
    "learning_rate": 0.0003552334956907604,
    "timestep_mean": 381.75,
    "epsilon_norm": 4.3024773597717285,
    "prediction_norm": 3.5025668144226074,
    "timestamp": "2025-07-14T12:34:15.911392"
  },
  {
    "epoch": 1188,
    "batch_idx": 0,
    "loss": 0.28514549136161804,
    "learning_rate": 0.0003544819165858641,
    "timestep_mean": 512.125,
    "epsilon_norm": 6.654361248016357,
    "prediction_norm": 5.692826271057129,
    "timestamp": "2025-07-14T12:34:15.916510"
  },
  {
    "epoch": 1188,
    "batch_idx": 32,
    "loss": 0.19823166728019714,
    "learning_rate": 0.0003544819165858641,
    "timestep_mean": 523.8125,
    "epsilon_norm": 7.8500657081604,
    "prediction_norm": 7.283463954925537,
    "timestamp": "2025-07-14T12:34:15.921799"
  },
  {
    "epoch": 1188,
    "batch_idx": 64,
    "loss": 0.3368406593799591,
    "learning_rate": 0.0003544819165858641,
    "timestep_mean": 396.625,
    "epsilon_norm": 7.828498363494873,
    "prediction_norm": 7.2451348304748535,
    "timestamp": "2025-07-14T12:34:15.926984"
  },
  {
    "epoch": 1188,
    "batch_idx": 96,
    "loss": 0.4005107283592224,
    "learning_rate": 0.0003544819165858641,
    "timestep_mean": 471.40625,
    "epsilon_norm": 8.476791381835938,
    "prediction_norm": 6.976410865783691,
    "timestamp": "2025-07-14T12:34:15.932117"
  },
  {
    "epoch": 1188,
    "batch_idx": 128,
    "loss": 0.23020221292972565,
    "learning_rate": 0.0003544819165858641,
    "timestep_mean": 490.4375,
    "epsilon_norm": 7.242090702056885,
    "prediction_norm": 6.885230541229248,
    "timestamp": "2025-07-14T12:34:15.937239"
  },
  {
    "epoch": 1188,
    "batch_idx": 160,
    "loss": 0.23484262824058533,
    "learning_rate": 0.0003544819165858641,
    "timestep_mean": 507.0,
    "epsilon_norm": 7.959923267364502,
    "prediction_norm": 7.16463565826416,
    "timestamp": "2025-07-14T12:34:15.942412"
  },
  {
    "epoch": 1188,
    "batch_idx": 192,
    "loss": 0.30348077416419983,
    "learning_rate": 0.0003544819165858641,
    "timestep_mean": 516.5,
    "epsilon_norm": 4.267904758453369,
    "prediction_norm": 3.596341848373413,
    "timestamp": "2025-07-14T12:34:15.947517"
  },
  {
    "epoch": 1189,
    "batch_idx": 0,
    "loss": 0.2780744731426239,
    "learning_rate": 0.0003537306965323731,
    "timestep_mean": 468.03125,
    "epsilon_norm": 9.137495040893555,
    "prediction_norm": 8.261375427246094,
    "timestamp": "2025-07-14T12:34:15.952698"
  },
  {
    "epoch": 1189,
    "batch_idx": 32,
    "loss": 0.3567764163017273,
    "learning_rate": 0.0003537306965323731,
    "timestep_mean": 468.59375,
    "epsilon_norm": 8.691620826721191,
    "prediction_norm": 7.288072109222412,
    "timestamp": "2025-07-14T12:34:15.957936"
  },
  {
    "epoch": 1189,
    "batch_idx": 64,
    "loss": 0.1571895182132721,
    "learning_rate": 0.0003537306965323731,
    "timestep_mean": 483.78125,
    "epsilon_norm": 8.473520278930664,
    "prediction_norm": 7.979373455047607,
    "timestamp": "2025-07-14T12:34:15.963268"
  },
  {
    "epoch": 1189,
    "batch_idx": 96,
    "loss": 0.16280090808868408,
    "learning_rate": 0.0003537306965323731,
    "timestep_mean": 534.8125,
    "epsilon_norm": 8.6964693069458,
    "prediction_norm": 8.038087844848633,
    "timestamp": "2025-07-14T12:34:15.968395"
  },
  {
    "epoch": 1189,
    "batch_idx": 128,
    "loss": 0.2290543019771576,
    "learning_rate": 0.0003537306965323731,
    "timestep_mean": 443.9375,
    "epsilon_norm": 7.326386451721191,
    "prediction_norm": 6.140668869018555,
    "timestamp": "2025-07-14T12:34:15.973561"
  },
  {
    "epoch": 1189,
    "batch_idx": 160,
    "loss": 0.30399277806282043,
    "learning_rate": 0.0003537306965323731,
    "timestep_mean": 478.21875,
    "epsilon_norm": 7.989330768585205,
    "prediction_norm": 6.8010430335998535,
    "timestamp": "2025-07-14T12:34:15.978820"
  },
  {
    "epoch": 1189,
    "batch_idx": 192,
    "loss": 0.11749182641506195,
    "learning_rate": 0.0003537306965323731,
    "timestep_mean": 612.75,
    "epsilon_norm": 3.851360559463501,
    "prediction_norm": 3.2111446857452393,
    "timestamp": "2025-07-14T12:34:15.983847"
  },
  {
    "epoch": 1190,
    "batch_idx": 0,
    "loss": 0.1258390247821808,
    "learning_rate": 0.00035297983738384814,
    "timestep_mean": 510.5625,
    "epsilon_norm": 6.061405658721924,
    "prediction_norm": 5.338322639465332,
    "timestamp": "2025-07-14T12:34:15.988956"
  },
  {
    "epoch": 1190,
    "batch_idx": 32,
    "loss": 0.24974358081817627,
    "learning_rate": 0.00035297983738384814,
    "timestep_mean": 566.71875,
    "epsilon_norm": 7.68543004989624,
    "prediction_norm": 6.399723529815674,
    "timestamp": "2025-07-14T12:34:15.994137"
  },
  {
    "epoch": 1190,
    "batch_idx": 64,
    "loss": 0.22106188535690308,
    "learning_rate": 0.00035297983738384814,
    "timestep_mean": 452.625,
    "epsilon_norm": 7.428075313568115,
    "prediction_norm": 6.024657249450684,
    "timestamp": "2025-07-14T12:34:15.999379"
  },
  {
    "epoch": 1190,
    "batch_idx": 96,
    "loss": 0.33875319361686707,
    "learning_rate": 0.00035297983738384814,
    "timestep_mean": 452.40625,
    "epsilon_norm": 8.270766258239746,
    "prediction_norm": 6.99353551864624,
    "timestamp": "2025-07-14T12:34:16.004669"
  },
  {
    "epoch": 1190,
    "batch_idx": 128,
    "loss": 0.32790079712867737,
    "learning_rate": 0.00035297983738384814,
    "timestep_mean": 461.59375,
    "epsilon_norm": 8.206168174743652,
    "prediction_norm": 7.1046037673950195,
    "timestamp": "2025-07-14T12:34:16.009639"
  },
  {
    "epoch": 1190,
    "batch_idx": 160,
    "loss": 0.2302725911140442,
    "learning_rate": 0.00035297983738384814,
    "timestep_mean": 546.8125,
    "epsilon_norm": 7.685558319091797,
    "prediction_norm": 6.41654109954834,
    "timestamp": "2025-07-14T12:34:16.014532"
  },
  {
    "epoch": 1190,
    "batch_idx": 192,
    "loss": 0.10272238403558731,
    "learning_rate": 0.00035297983738384814,
    "timestep_mean": 651.625,
    "epsilon_norm": 5.542036533355713,
    "prediction_norm": 5.212698936462402,
    "timestamp": "2025-07-14T12:34:16.019342"
  },
  {
    "epoch": 1191,
    "batch_idx": 0,
    "loss": 0.3421098589897156,
    "learning_rate": 0.0003522293409929596,
    "timestep_mean": 519.46875,
    "epsilon_norm": 7.2421135902404785,
    "prediction_norm": 6.231999397277832,
    "timestamp": "2025-07-14T12:34:16.024222"
  },
  {
    "epoch": 1191,
    "batch_idx": 32,
    "loss": 0.3486846685409546,
    "learning_rate": 0.0003522293409929596,
    "timestep_mean": 618.375,
    "epsilon_norm": 9.131879806518555,
    "prediction_norm": 7.744691371917725,
    "timestamp": "2025-07-14T12:34:16.029161"
  },
  {
    "epoch": 1191,
    "batch_idx": 64,
    "loss": 0.3183842897415161,
    "learning_rate": 0.0003522293409929596,
    "timestep_mean": 497.21875,
    "epsilon_norm": 8.041301727294922,
    "prediction_norm": 6.496700286865234,
    "timestamp": "2025-07-14T12:34:16.034109"
  },
  {
    "epoch": 1191,
    "batch_idx": 96,
    "loss": 0.14694394171237946,
    "learning_rate": 0.0003522293409929596,
    "timestep_mean": 566.25,
    "epsilon_norm": 7.534566879272461,
    "prediction_norm": 6.223735809326172,
    "timestamp": "2025-07-14T12:34:16.039055"
  },
  {
    "epoch": 1191,
    "batch_idx": 128,
    "loss": 0.4513983428478241,
    "learning_rate": 0.0003522293409929596,
    "timestep_mean": 470.0625,
    "epsilon_norm": 8.657470703125,
    "prediction_norm": 6.648598670959473,
    "timestamp": "2025-07-14T12:34:16.043881"
  },
  {
    "epoch": 1191,
    "batch_idx": 160,
    "loss": 0.31108248233795166,
    "learning_rate": 0.0003522293409929596,
    "timestep_mean": 479.0625,
    "epsilon_norm": 9.919942855834961,
    "prediction_norm": 8.238197326660156,
    "timestamp": "2025-07-14T12:34:16.048719"
  },
  {
    "epoch": 1191,
    "batch_idx": 192,
    "loss": 0.39888375997543335,
    "learning_rate": 0.0003522293409929596,
    "timestep_mean": 350.25,
    "epsilon_norm": 4.171975612640381,
    "prediction_norm": 3.622957468032837,
    "timestamp": "2025-07-14T12:34:16.053578"
  },
  {
    "epoch": 1192,
    "batch_idx": 0,
    "loss": 0.19544154405593872,
    "learning_rate": 0.0003514792092114826,
    "timestep_mean": 447.78125,
    "epsilon_norm": 7.067167282104492,
    "prediction_norm": 6.276608943939209,
    "timestamp": "2025-07-14T12:34:16.058529"
  },
  {
    "epoch": 1192,
    "batch_idx": 32,
    "loss": 0.2467646598815918,
    "learning_rate": 0.0003514792092114826,
    "timestep_mean": 459.6875,
    "epsilon_norm": 8.448650360107422,
    "prediction_norm": 7.2782440185546875,
    "timestamp": "2025-07-14T12:34:16.063401"
  },
  {
    "epoch": 1192,
    "batch_idx": 64,
    "loss": 0.4316079616546631,
    "learning_rate": 0.0003514792092114826,
    "timestep_mean": 454.65625,
    "epsilon_norm": 8.783044815063477,
    "prediction_norm": 7.137266159057617,
    "timestamp": "2025-07-14T12:34:16.068286"
  },
  {
    "epoch": 1192,
    "batch_idx": 96,
    "loss": 0.07845597714185715,
    "learning_rate": 0.0003514792092114826,
    "timestep_mean": 596.65625,
    "epsilon_norm": 7.237146377563477,
    "prediction_norm": 6.855554580688477,
    "timestamp": "2025-07-14T12:34:16.073100"
  },
  {
    "epoch": 1192,
    "batch_idx": 128,
    "loss": 0.15432409942150116,
    "learning_rate": 0.0003514792092114826,
    "timestep_mean": 463.125,
    "epsilon_norm": 7.8759446144104,
    "prediction_norm": 7.6647539138793945,
    "timestamp": "2025-07-14T12:34:16.077929"
  },
  {
    "epoch": 1192,
    "batch_idx": 160,
    "loss": 0.18554770946502686,
    "learning_rate": 0.0003514792092114826,
    "timestep_mean": 595.0625,
    "epsilon_norm": 7.561648368835449,
    "prediction_norm": 7.422065258026123,
    "timestamp": "2025-07-14T12:34:16.082732"
  },
  {
    "epoch": 1192,
    "batch_idx": 192,
    "loss": 0.4392682909965515,
    "learning_rate": 0.0003514792092114826,
    "timestep_mean": 563.75,
    "epsilon_norm": 4.570335388183594,
    "prediction_norm": 2.971907138824463,
    "timestamp": "2025-07-14T12:34:16.087642"
  },
  {
    "epoch": 1193,
    "batch_idx": 0,
    "loss": 0.18600918352603912,
    "learning_rate": 0.000350729443890293,
    "timestep_mean": 524.0,
    "epsilon_norm": 7.705532550811768,
    "prediction_norm": 7.49146842956543,
    "timestamp": "2025-07-14T12:34:16.092536"
  },
  {
    "epoch": 1193,
    "batch_idx": 32,
    "loss": 0.23290923237800598,
    "learning_rate": 0.000350729443890293,
    "timestep_mean": 498.3125,
    "epsilon_norm": 8.213894844055176,
    "prediction_norm": 6.79178524017334,
    "timestamp": "2025-07-14T12:34:16.097354"
  },
  {
    "epoch": 1193,
    "batch_idx": 64,
    "loss": 0.19550101459026337,
    "learning_rate": 0.000350729443890293,
    "timestep_mean": 504.96875,
    "epsilon_norm": 7.939987659454346,
    "prediction_norm": 7.151980400085449,
    "timestamp": "2025-07-14T12:34:16.102215"
  },
  {
    "epoch": 1193,
    "batch_idx": 96,
    "loss": 0.3796963095664978,
    "learning_rate": 0.000350729443890293,
    "timestep_mean": 517.875,
    "epsilon_norm": 8.74980640411377,
    "prediction_norm": 6.663187026977539,
    "timestamp": "2025-07-14T12:34:16.107084"
  },
  {
    "epoch": 1193,
    "batch_idx": 128,
    "loss": 0.23181281983852386,
    "learning_rate": 0.000350729443890293,
    "timestep_mean": 622.28125,
    "epsilon_norm": 7.869236469268799,
    "prediction_norm": 7.614656925201416,
    "timestamp": "2025-07-14T12:34:16.111936"
  },
  {
    "epoch": 1193,
    "batch_idx": 160,
    "loss": 0.26577869057655334,
    "learning_rate": 0.000350729443890293,
    "timestep_mean": 537.96875,
    "epsilon_norm": 7.6718010902404785,
    "prediction_norm": 6.8881449699401855,
    "timestamp": "2025-07-14T12:34:16.116897"
  },
  {
    "epoch": 1193,
    "batch_idx": 192,
    "loss": 0.31732961535453796,
    "learning_rate": 0.000350729443890293,
    "timestep_mean": 419.875,
    "epsilon_norm": 4.4910759925842285,
    "prediction_norm": 3.6942684650421143,
    "timestamp": "2025-07-14T12:34:16.121772"
  },
  {
    "epoch": 1194,
    "batch_idx": 0,
    "loss": 0.18353918194770813,
    "learning_rate": 0.000349980046879362,
    "timestep_mean": 551.625,
    "epsilon_norm": 7.4115447998046875,
    "prediction_norm": 6.5778021812438965,
    "timestamp": "2025-07-14T12:34:16.126653"
  },
  {
    "epoch": 1194,
    "batch_idx": 32,
    "loss": 0.3970922827720642,
    "learning_rate": 0.000349980046879362,
    "timestep_mean": 495.34375,
    "epsilon_norm": 7.199082374572754,
    "prediction_norm": 6.444259166717529,
    "timestamp": "2025-07-14T12:34:16.131515"
  },
  {
    "epoch": 1194,
    "batch_idx": 64,
    "loss": 0.22188617289066315,
    "learning_rate": 0.000349980046879362,
    "timestep_mean": 519.875,
    "epsilon_norm": 7.448816776275635,
    "prediction_norm": 6.9258928298950195,
    "timestamp": "2025-07-14T12:34:16.136363"
  },
  {
    "epoch": 1194,
    "batch_idx": 96,
    "loss": 0.28676778078079224,
    "learning_rate": 0.000349980046879362,
    "timestep_mean": 429.59375,
    "epsilon_norm": 7.083000659942627,
    "prediction_norm": 5.279517650604248,
    "timestamp": "2025-07-14T12:34:16.141202"
  },
  {
    "epoch": 1194,
    "batch_idx": 128,
    "loss": 0.3196079730987549,
    "learning_rate": 0.000349980046879362,
    "timestep_mean": 489.78125,
    "epsilon_norm": 8.672101974487305,
    "prediction_norm": 7.011771202087402,
    "timestamp": "2025-07-14T12:34:16.146171"
  },
  {
    "epoch": 1194,
    "batch_idx": 160,
    "loss": 0.2658776640892029,
    "learning_rate": 0.000349980046879362,
    "timestep_mean": 538.15625,
    "epsilon_norm": 7.984393119812012,
    "prediction_norm": 6.476343631744385,
    "timestamp": "2025-07-14T12:34:16.151076"
  },
  {
    "epoch": 1194,
    "batch_idx": 192,
    "loss": 0.24243980646133423,
    "learning_rate": 0.000349980046879362,
    "timestep_mean": 305.125,
    "epsilon_norm": 3.950463056564331,
    "prediction_norm": 2.9259941577911377,
    "timestamp": "2025-07-14T12:34:16.155925"
  },
  {
    "epoch": 1195,
    "batch_idx": 0,
    "loss": 0.24640914797782898,
    "learning_rate": 0.0003492310200277522,
    "timestep_mean": 468.625,
    "epsilon_norm": 6.580659866333008,
    "prediction_norm": 5.504063129425049,
    "timestamp": "2025-07-14T12:34:16.160829"
  },
  {
    "epoch": 1195,
    "batch_idx": 32,
    "loss": 0.4250969886779785,
    "learning_rate": 0.0003492310200277522,
    "timestep_mean": 513.59375,
    "epsilon_norm": 8.10125732421875,
    "prediction_norm": 5.95025634765625,
    "timestamp": "2025-07-14T12:34:16.165769"
  },
  {
    "epoch": 1195,
    "batch_idx": 64,
    "loss": 0.289138525724411,
    "learning_rate": 0.0003492310200277522,
    "timestep_mean": 545.46875,
    "epsilon_norm": 8.662185668945312,
    "prediction_norm": 6.885486125946045,
    "timestamp": "2025-07-14T12:34:16.170632"
  },
  {
    "epoch": 1195,
    "batch_idx": 96,
    "loss": 0.4413752555847168,
    "learning_rate": 0.0003492310200277522,
    "timestep_mean": 458.28125,
    "epsilon_norm": 7.464919090270996,
    "prediction_norm": 5.2892045974731445,
    "timestamp": "2025-07-14T12:34:16.175584"
  },
  {
    "epoch": 1195,
    "batch_idx": 128,
    "loss": 0.17761938273906708,
    "learning_rate": 0.0003492310200277522,
    "timestep_mean": 499.15625,
    "epsilon_norm": 6.768833160400391,
    "prediction_norm": 5.862586498260498,
    "timestamp": "2025-07-14T12:34:16.180461"
  },
  {
    "epoch": 1195,
    "batch_idx": 160,
    "loss": 0.31671300530433655,
    "learning_rate": 0.0003492310200277522,
    "timestep_mean": 387.96875,
    "epsilon_norm": 6.9150776863098145,
    "prediction_norm": 5.075418472290039,
    "timestamp": "2025-07-14T12:34:16.185305"
  },
  {
    "epoch": 1195,
    "batch_idx": 192,
    "loss": 0.10085413604974747,
    "learning_rate": 0.0003492310200277522,
    "timestep_mean": 584.625,
    "epsilon_norm": 3.3466086387634277,
    "prediction_norm": 3.1868841648101807,
    "timestamp": "2025-07-14T12:34:16.190114"
  },
  {
    "epoch": 1196,
    "batch_idx": 0,
    "loss": 0.30993151664733887,
    "learning_rate": 0.0003484823651836131,
    "timestep_mean": 493.8125,
    "epsilon_norm": 7.348749160766602,
    "prediction_norm": 5.970036029815674,
    "timestamp": "2025-07-14T12:34:16.194979"
  },
  {
    "epoch": 1196,
    "batch_idx": 32,
    "loss": 0.4239809811115265,
    "learning_rate": 0.0003484823651836131,
    "timestep_mean": 487.59375,
    "epsilon_norm": 7.253026962280273,
    "prediction_norm": 6.13204288482666,
    "timestamp": "2025-07-14T12:34:16.199817"
  },
  {
    "epoch": 1196,
    "batch_idx": 64,
    "loss": 0.2681211233139038,
    "learning_rate": 0.0003484823651836131,
    "timestep_mean": 472.53125,
    "epsilon_norm": 8.932719230651855,
    "prediction_norm": 7.919509410858154,
    "timestamp": "2025-07-14T12:34:16.204759"
  },
  {
    "epoch": 1196,
    "batch_idx": 96,
    "loss": 0.37352442741394043,
    "learning_rate": 0.0003484823651836131,
    "timestep_mean": 546.21875,
    "epsilon_norm": 7.1067280769348145,
    "prediction_norm": 6.179075241088867,
    "timestamp": "2025-07-14T12:34:16.209527"
  },
  {
    "epoch": 1196,
    "batch_idx": 128,
    "loss": 0.4421032965183258,
    "learning_rate": 0.0003484823651836131,
    "timestep_mean": 542.96875,
    "epsilon_norm": 7.89966344833374,
    "prediction_norm": 5.850106239318848,
    "timestamp": "2025-07-14T12:34:16.214368"
  },
  {
    "epoch": 1196,
    "batch_idx": 160,
    "loss": 0.16454565525054932,
    "learning_rate": 0.0003484823651836131,
    "timestep_mean": 564.34375,
    "epsilon_norm": 7.748650074005127,
    "prediction_norm": 6.839023113250732,
    "timestamp": "2025-07-14T12:34:16.219249"
  },
  {
    "epoch": 1196,
    "batch_idx": 192,
    "loss": 0.4404311180114746,
    "learning_rate": 0.0003484823651836131,
    "timestep_mean": 567.125,
    "epsilon_norm": 3.907543420791626,
    "prediction_norm": 3.3146777153015137,
    "timestamp": "2025-07-14T12:34:16.224117"
  },
  {
    "epoch": 1197,
    "batch_idx": 0,
    "loss": 0.24030953645706177,
    "learning_rate": 0.000347734084194176,
    "timestep_mean": 456.4375,
    "epsilon_norm": 7.9774017333984375,
    "prediction_norm": 7.1804890632629395,
    "timestamp": "2025-07-14T12:34:16.229030"
  },
  {
    "epoch": 1197,
    "batch_idx": 32,
    "loss": 0.18763497471809387,
    "learning_rate": 0.000347734084194176,
    "timestep_mean": 545.8125,
    "epsilon_norm": 8.473051071166992,
    "prediction_norm": 7.821730613708496,
    "timestamp": "2025-07-14T12:34:16.234117"
  },
  {
    "epoch": 1197,
    "batch_idx": 64,
    "loss": 0.2679775357246399,
    "learning_rate": 0.000347734084194176,
    "timestep_mean": 514.78125,
    "epsilon_norm": 7.793696880340576,
    "prediction_norm": 6.28111457824707,
    "timestamp": "2025-07-14T12:34:16.239253"
  },
  {
    "epoch": 1197,
    "batch_idx": 96,
    "loss": 0.47210896015167236,
    "learning_rate": 0.000347734084194176,
    "timestep_mean": 386.84375,
    "epsilon_norm": 7.813519477844238,
    "prediction_norm": 6.51442289352417,
    "timestamp": "2025-07-14T12:34:16.244422"
  },
  {
    "epoch": 1197,
    "batch_idx": 128,
    "loss": 0.20458847284317017,
    "learning_rate": 0.000347734084194176,
    "timestep_mean": 520.625,
    "epsilon_norm": 7.823854923248291,
    "prediction_norm": 7.124120712280273,
    "timestamp": "2025-07-14T12:34:16.249647"
  },
  {
    "epoch": 1197,
    "batch_idx": 160,
    "loss": 0.29793912172317505,
    "learning_rate": 0.000347734084194176,
    "timestep_mean": 505.96875,
    "epsilon_norm": 8.246341705322266,
    "prediction_norm": 7.120811939239502,
    "timestamp": "2025-07-14T12:34:16.254869"
  },
  {
    "epoch": 1197,
    "batch_idx": 192,
    "loss": 0.4488029181957245,
    "learning_rate": 0.000347734084194176,
    "timestep_mean": 459.5,
    "epsilon_norm": 4.722413063049316,
    "prediction_norm": 3.5213451385498047,
    "timestamp": "2025-07-14T12:34:16.259919"
  },
  {
    "epoch": 1198,
    "batch_idx": 0,
    "loss": 0.15730589628219604,
    "learning_rate": 0.0003469861789057496,
    "timestep_mean": 497.84375,
    "epsilon_norm": 7.316253662109375,
    "prediction_norm": 6.986134052276611,
    "timestamp": "2025-07-14T12:34:16.265039"
  },
  {
    "epoch": 1198,
    "batch_idx": 32,
    "loss": 0.5358105897903442,
    "learning_rate": 0.0003469861789057496,
    "timestep_mean": 476.6875,
    "epsilon_norm": 9.118823051452637,
    "prediction_norm": 7.7186713218688965,
    "timestamp": "2025-07-14T12:34:16.270210"
  },
  {
    "epoch": 1198,
    "batch_idx": 64,
    "loss": 0.20374950766563416,
    "learning_rate": 0.0003469861789057496,
    "timestep_mean": 505.4375,
    "epsilon_norm": 7.300445556640625,
    "prediction_norm": 7.010687828063965,
    "timestamp": "2025-07-14T12:34:16.275465"
  },
  {
    "epoch": 1198,
    "batch_idx": 96,
    "loss": 0.39975088834762573,
    "learning_rate": 0.0003469861789057496,
    "timestep_mean": 453.53125,
    "epsilon_norm": 8.690351486206055,
    "prediction_norm": 6.740938663482666,
    "timestamp": "2025-07-14T12:34:16.280697"
  },
  {
    "epoch": 1198,
    "batch_idx": 128,
    "loss": 0.41192176938056946,
    "learning_rate": 0.0003469861789057496,
    "timestep_mean": 410.25,
    "epsilon_norm": 7.1141438484191895,
    "prediction_norm": 5.844163417816162,
    "timestamp": "2025-07-14T12:34:16.285888"
  },
  {
    "epoch": 1198,
    "batch_idx": 160,
    "loss": 0.45610225200653076,
    "learning_rate": 0.0003469861789057496,
    "timestep_mean": 520.28125,
    "epsilon_norm": 8.499944686889648,
    "prediction_norm": 6.579807758331299,
    "timestamp": "2025-07-14T12:34:16.291082"
  },
  {
    "epoch": 1198,
    "batch_idx": 192,
    "loss": 0.5342240929603577,
    "learning_rate": 0.0003469861789057496,
    "timestep_mean": 494.5,
    "epsilon_norm": 4.912374973297119,
    "prediction_norm": 3.5672929286956787,
    "timestamp": "2025-07-14T12:34:16.296260"
  },
  {
    "epoch": 1199,
    "batch_idx": 0,
    "loss": 0.369758278131485,
    "learning_rate": 0.0003462386511637164,
    "timestep_mean": 486.09375,
    "epsilon_norm": 8.597858428955078,
    "prediction_norm": 6.6890869140625,
    "timestamp": "2025-07-14T12:34:16.301488"
  },
  {
    "epoch": 1199,
    "batch_idx": 32,
    "loss": 0.24228525161743164,
    "learning_rate": 0.0003462386511637164,
    "timestep_mean": 551.25,
    "epsilon_norm": 8.748276710510254,
    "prediction_norm": 7.302587032318115,
    "timestamp": "2025-07-14T12:34:16.306661"
  },
  {
    "epoch": 1199,
    "batch_idx": 64,
    "loss": 0.28264299035072327,
    "learning_rate": 0.0003462386511637164,
    "timestep_mean": 480.0,
    "epsilon_norm": 8.453716278076172,
    "prediction_norm": 7.084771156311035,
    "timestamp": "2025-07-14T12:34:16.311716"
  },
  {
    "epoch": 1199,
    "batch_idx": 96,
    "loss": 0.3830856680870056,
    "learning_rate": 0.0003462386511637164,
    "timestep_mean": 438.9375,
    "epsilon_norm": 7.030449390411377,
    "prediction_norm": 5.320405006408691,
    "timestamp": "2025-07-14T12:34:16.316918"
  },
  {
    "epoch": 1199,
    "batch_idx": 128,
    "loss": 0.22003428637981415,
    "learning_rate": 0.0003462386511637164,
    "timestep_mean": 550.375,
    "epsilon_norm": 8.917251586914062,
    "prediction_norm": 7.740634441375732,
    "timestamp": "2025-07-14T12:34:16.321858"
  },
  {
    "epoch": 1199,
    "batch_idx": 160,
    "loss": 0.22913046181201935,
    "learning_rate": 0.0003462386511637164,
    "timestep_mean": 585.28125,
    "epsilon_norm": 8.32653522491455,
    "prediction_norm": 7.393132209777832,
    "timestamp": "2025-07-14T12:34:16.327035"
  },
  {
    "epoch": 1199,
    "batch_idx": 192,
    "loss": 0.10899153351783752,
    "learning_rate": 0.0003462386511637164,
    "timestep_mean": 626.375,
    "epsilon_norm": 3.5130808353424072,
    "prediction_norm": 2.9914703369140625,
    "timestamp": "2025-07-14T12:34:16.332230"
  },
  {
    "epoch": 1200,
    "batch_idx": 0,
    "loss": 0.2227228879928589,
    "learning_rate": 0.0003454915028125264,
    "timestep_mean": 518.4375,
    "epsilon_norm": 8.295605659484863,
    "prediction_norm": 6.6731367111206055,
    "timestamp": "2025-07-14T12:34:16.337181"
  },
  {
    "epoch": 1200,
    "batch_idx": 32,
    "loss": 0.2710829973220825,
    "learning_rate": 0.0003454915028125264,
    "timestep_mean": 540.5,
    "epsilon_norm": 6.853489398956299,
    "prediction_norm": 5.736329555511475,
    "timestamp": "2025-07-14T12:34:16.342399"
  },
  {
    "epoch": 1200,
    "batch_idx": 64,
    "loss": 0.13812117278575897,
    "learning_rate": 0.0003454915028125264,
    "timestep_mean": 582.25,
    "epsilon_norm": 7.419060230255127,
    "prediction_norm": 6.724219799041748,
    "timestamp": "2025-07-14T12:34:16.347595"
  },
  {
    "epoch": 1200,
    "batch_idx": 96,
    "loss": 0.09518453478813171,
    "learning_rate": 0.0003454915028125264,
    "timestep_mean": 484.46875,
    "epsilon_norm": 7.819483757019043,
    "prediction_norm": 6.875470161437988,
    "timestamp": "2025-07-14T12:34:16.352687"
  },
  {
    "epoch": 1200,
    "batch_idx": 128,
    "loss": 0.2276647686958313,
    "learning_rate": 0.0003454915028125264,
    "timestep_mean": 462.4375,
    "epsilon_norm": 7.773518085479736,
    "prediction_norm": 6.7660980224609375,
    "timestamp": "2025-07-14T12:34:16.357872"
  },
  {
    "epoch": 1200,
    "batch_idx": 160,
    "loss": 0.20920684933662415,
    "learning_rate": 0.0003454915028125264,
    "timestep_mean": 498.53125,
    "epsilon_norm": 7.649120330810547,
    "prediction_norm": 6.2092814445495605,
    "timestamp": "2025-07-14T12:34:16.363124"
  },
  {
    "epoch": 1200,
    "batch_idx": 192,
    "loss": 0.07606575638055801,
    "learning_rate": 0.0003454915028125264,
    "timestep_mean": 545.625,
    "epsilon_norm": 4.449346542358398,
    "prediction_norm": 4.594609260559082,
    "timestamp": "2025-07-14T12:34:16.368235"
  },
  {
    "epoch": 1201,
    "batch_idx": 0,
    "loss": 0.4527302384376526,
    "learning_rate": 0.000344744735695694,
    "timestep_mean": 454.75,
    "epsilon_norm": 8.384572982788086,
    "prediction_norm": 7.030679225921631,
    "timestamp": "2025-07-14T12:34:16.373483"
  },
  {
    "epoch": 1201,
    "batch_idx": 32,
    "loss": 0.25486308336257935,
    "learning_rate": 0.000344744735695694,
    "timestep_mean": 498.9375,
    "epsilon_norm": 7.8301682472229,
    "prediction_norm": 6.578183650970459,
    "timestamp": "2025-07-14T12:34:16.378711"
  },
  {
    "epoch": 1201,
    "batch_idx": 64,
    "loss": 0.2945128381252289,
    "learning_rate": 0.000344744735695694,
    "timestep_mean": 475.4375,
    "epsilon_norm": 8.320446968078613,
    "prediction_norm": 7.242528915405273,
    "timestamp": "2025-07-14T12:34:16.383889"
  },
  {
    "epoch": 1201,
    "batch_idx": 96,
    "loss": 0.30643242597579956,
    "learning_rate": 0.000344744735695694,
    "timestep_mean": 564.75,
    "epsilon_norm": 7.554869174957275,
    "prediction_norm": 6.713352203369141,
    "timestamp": "2025-07-14T12:34:16.388986"
  },
  {
    "epoch": 1201,
    "batch_idx": 128,
    "loss": 0.20738281309604645,
    "learning_rate": 0.000344744735695694,
    "timestep_mean": 427.84375,
    "epsilon_norm": 7.117237567901611,
    "prediction_norm": 6.353435516357422,
    "timestamp": "2025-07-14T12:34:16.394153"
  },
  {
    "epoch": 1201,
    "batch_idx": 160,
    "loss": 0.13738808035850525,
    "learning_rate": 0.000344744735695694,
    "timestep_mean": 572.625,
    "epsilon_norm": 8.262838363647461,
    "prediction_norm": 7.423511981964111,
    "timestamp": "2025-07-14T12:34:16.399336"
  },
  {
    "epoch": 1201,
    "batch_idx": 192,
    "loss": 0.061206310987472534,
    "learning_rate": 0.000344744735695694,
    "timestep_mean": 696.375,
    "epsilon_norm": 3.0465612411499023,
    "prediction_norm": 2.6836960315704346,
    "timestamp": "2025-07-14T12:34:16.404307"
  },
  {
    "epoch": 1202,
    "batch_idx": 0,
    "loss": 0.3298133611679077,
    "learning_rate": 0.0003439983516557927,
    "timestep_mean": 538.9375,
    "epsilon_norm": 8.39404582977295,
    "prediction_norm": 6.520343780517578,
    "timestamp": "2025-07-14T12:34:16.409201"
  },
  {
    "epoch": 1202,
    "batch_idx": 32,
    "loss": 0.1796949803829193,
    "learning_rate": 0.0003439983516557927,
    "timestep_mean": 548.96875,
    "epsilon_norm": 8.08929443359375,
    "prediction_norm": 6.980473041534424,
    "timestamp": "2025-07-14T12:34:16.414029"
  },
  {
    "epoch": 1202,
    "batch_idx": 64,
    "loss": 0.2721925377845764,
    "learning_rate": 0.0003439983516557927,
    "timestep_mean": 493.0625,
    "epsilon_norm": 9.599421501159668,
    "prediction_norm": 8.330219268798828,
    "timestamp": "2025-07-14T12:34:16.419352"
  },
  {
    "epoch": 1202,
    "batch_idx": 96,
    "loss": 0.33982154726982117,
    "learning_rate": 0.0003439983516557927,
    "timestep_mean": 434.875,
    "epsilon_norm": 7.850663661956787,
    "prediction_norm": 6.475150108337402,
    "timestamp": "2025-07-14T12:34:16.424442"
  },
  {
    "epoch": 1202,
    "batch_idx": 128,
    "loss": 0.3754598796367645,
    "learning_rate": 0.0003439983516557927,
    "timestep_mean": 424.125,
    "epsilon_norm": 7.709615230560303,
    "prediction_norm": 5.834156513214111,
    "timestamp": "2025-07-14T12:34:16.429767"
  },
  {
    "epoch": 1202,
    "batch_idx": 160,
    "loss": 0.26848703622817993,
    "learning_rate": 0.0003439983516557927,
    "timestep_mean": 470.96875,
    "epsilon_norm": 7.480291843414307,
    "prediction_norm": 5.894364833831787,
    "timestamp": "2025-07-14T12:34:16.434934"
  },
  {
    "epoch": 1202,
    "batch_idx": 192,
    "loss": 0.28604376316070557,
    "learning_rate": 0.0003439983516557927,
    "timestep_mean": 396.375,
    "epsilon_norm": 4.122359752655029,
    "prediction_norm": 3.8921115398406982,
    "timestamp": "2025-07-14T12:34:16.439947"
  },
  {
    "epoch": 1203,
    "batch_idx": 0,
    "loss": 0.13898609578609467,
    "learning_rate": 0.000343252352534451,
    "timestep_mean": 567.0,
    "epsilon_norm": 8.708638191223145,
    "prediction_norm": 8.194523811340332,
    "timestamp": "2025-07-14T12:34:16.445195"
  },
  {
    "epoch": 1203,
    "batch_idx": 32,
    "loss": 0.1755116879940033,
    "learning_rate": 0.000343252352534451,
    "timestep_mean": 463.0625,
    "epsilon_norm": 7.788713455200195,
    "prediction_norm": 7.148530960083008,
    "timestamp": "2025-07-14T12:34:16.450731"
  },
  {
    "epoch": 1203,
    "batch_idx": 64,
    "loss": 0.21052059531211853,
    "learning_rate": 0.000343252352534451,
    "timestep_mean": 593.4375,
    "epsilon_norm": 7.820517539978027,
    "prediction_norm": 7.3623456954956055,
    "timestamp": "2025-07-14T12:34:16.455879"
  },
  {
    "epoch": 1203,
    "batch_idx": 96,
    "loss": 0.24017208814620972,
    "learning_rate": 0.000343252352534451,
    "timestep_mean": 538.625,
    "epsilon_norm": 7.763771057128906,
    "prediction_norm": 6.746481418609619,
    "timestamp": "2025-07-14T12:34:16.460995"
  },
  {
    "epoch": 1203,
    "batch_idx": 128,
    "loss": 0.30368533730506897,
    "learning_rate": 0.000343252352534451,
    "timestep_mean": 445.96875,
    "epsilon_norm": 8.056798934936523,
    "prediction_norm": 6.71865701675415,
    "timestamp": "2025-07-14T12:34:16.466128"
  },
  {
    "epoch": 1203,
    "batch_idx": 160,
    "loss": 0.2800726592540741,
    "learning_rate": 0.000343252352534451,
    "timestep_mean": 476.625,
    "epsilon_norm": 7.785982608795166,
    "prediction_norm": 6.1764936447143555,
    "timestamp": "2025-07-14T12:34:16.471464"
  },
  {
    "epoch": 1203,
    "batch_idx": 192,
    "loss": 0.11041224747896194,
    "learning_rate": 0.000343252352534451,
    "timestep_mean": 375.25,
    "epsilon_norm": 3.9647164344787598,
    "prediction_norm": 3.2342045307159424,
    "timestamp": "2025-07-14T12:34:16.476593"
  },
  {
    "epoch": 1204,
    "batch_idx": 0,
    "loss": 0.27990955114364624,
    "learning_rate": 0.00034250674017234764,
    "timestep_mean": 526.53125,
    "epsilon_norm": 8.238755226135254,
    "prediction_norm": 7.382061004638672,
    "timestamp": "2025-07-14T12:34:16.481679"
  },
  {
    "epoch": 1204,
    "batch_idx": 32,
    "loss": 0.3510640561580658,
    "learning_rate": 0.00034250674017234764,
    "timestep_mean": 458.5625,
    "epsilon_norm": 9.016948699951172,
    "prediction_norm": 8.042623519897461,
    "timestamp": "2025-07-14T12:34:16.486660"
  },
  {
    "epoch": 1204,
    "batch_idx": 64,
    "loss": 0.3591616153717041,
    "learning_rate": 0.00034250674017234764,
    "timestep_mean": 522.34375,
    "epsilon_norm": 8.786330223083496,
    "prediction_norm": 7.458332538604736,
    "timestamp": "2025-07-14T12:34:16.491849"
  },
  {
    "epoch": 1204,
    "batch_idx": 96,
    "loss": 0.15295258164405823,
    "learning_rate": 0.00034250674017234764,
    "timestep_mean": 625.5625,
    "epsilon_norm": 7.3995232582092285,
    "prediction_norm": 7.561422824859619,
    "timestamp": "2025-07-14T12:34:16.497122"
  },
  {
    "epoch": 1204,
    "batch_idx": 128,
    "loss": 0.31209132075309753,
    "learning_rate": 0.00034250674017234764,
    "timestep_mean": 509.21875,
    "epsilon_norm": 8.273476600646973,
    "prediction_norm": 7.13020133972168,
    "timestamp": "2025-07-14T12:34:16.502318"
  },
  {
    "epoch": 1204,
    "batch_idx": 160,
    "loss": 0.2512529492378235,
    "learning_rate": 0.00034250674017234764,
    "timestep_mean": 447.71875,
    "epsilon_norm": 7.881161689758301,
    "prediction_norm": 7.381912708282471,
    "timestamp": "2025-07-14T12:34:16.507481"
  },
  {
    "epoch": 1204,
    "batch_idx": 192,
    "loss": 0.4642132520675659,
    "learning_rate": 0.00034250674017234764,
    "timestep_mean": 457.0,
    "epsilon_norm": 4.126650333404541,
    "prediction_norm": 3.129838228225708,
    "timestamp": "2025-07-14T12:34:16.512647"
  },
  {
    "epoch": 1205,
    "batch_idx": 0,
    "loss": 0.1974530816078186,
    "learning_rate": 0.000341761516409207,
    "timestep_mean": 505.5,
    "epsilon_norm": 7.135026931762695,
    "prediction_norm": 6.70829963684082,
    "timestamp": "2025-07-14T12:34:16.517811"
  },
  {
    "epoch": 1205,
    "batch_idx": 32,
    "loss": 0.2574254870414734,
    "learning_rate": 0.000341761516409207,
    "timestep_mean": 486.28125,
    "epsilon_norm": 7.791018962860107,
    "prediction_norm": 6.899228096008301,
    "timestamp": "2025-07-14T12:34:16.522979"
  },
  {
    "epoch": 1205,
    "batch_idx": 64,
    "loss": 0.26252108812332153,
    "learning_rate": 0.000341761516409207,
    "timestep_mean": 543.625,
    "epsilon_norm": 8.301206588745117,
    "prediction_norm": 6.967191696166992,
    "timestamp": "2025-07-14T12:34:16.528237"
  },
  {
    "epoch": 1205,
    "batch_idx": 96,
    "loss": 0.45591357350349426,
    "learning_rate": 0.000341761516409207,
    "timestep_mean": 485.0,
    "epsilon_norm": 8.29145622253418,
    "prediction_norm": 7.314120292663574,
    "timestamp": "2025-07-14T12:34:16.533178"
  },
  {
    "epoch": 1205,
    "batch_idx": 128,
    "loss": 0.3433438539505005,
    "learning_rate": 0.000341761516409207,
    "timestep_mean": 481.03125,
    "epsilon_norm": 7.545677185058594,
    "prediction_norm": 6.289478778839111,
    "timestamp": "2025-07-14T12:34:16.538144"
  },
  {
    "epoch": 1205,
    "batch_idx": 160,
    "loss": 0.3360675871372223,
    "learning_rate": 0.000341761516409207,
    "timestep_mean": 513.3125,
    "epsilon_norm": 8.929084777832031,
    "prediction_norm": 7.84092378616333,
    "timestamp": "2025-07-14T12:34:16.543028"
  },
  {
    "epoch": 1205,
    "batch_idx": 192,
    "loss": 0.462125301361084,
    "learning_rate": 0.000341761516409207,
    "timestep_mean": 525.75,
    "epsilon_norm": 3.7023251056671143,
    "prediction_norm": 2.8911805152893066,
    "timestamp": "2025-07-14T12:34:16.548035"
  },
  {
    "epoch": 1206,
    "batch_idx": 0,
    "loss": 0.4488319456577301,
    "learning_rate": 0.00034101668308379453,
    "timestep_mean": 494.3125,
    "epsilon_norm": 7.992768287658691,
    "prediction_norm": 5.703406810760498,
    "timestamp": "2025-07-14T12:34:16.553014"
  },
  {
    "epoch": 1206,
    "batch_idx": 32,
    "loss": 0.1528247892856598,
    "learning_rate": 0.00034101668308379453,
    "timestep_mean": 503.625,
    "epsilon_norm": 7.319407939910889,
    "prediction_norm": 6.822298526763916,
    "timestamp": "2025-07-14T12:34:16.557915"
  },
  {
    "epoch": 1206,
    "batch_idx": 64,
    "loss": 0.10775846242904663,
    "learning_rate": 0.00034101668308379453,
    "timestep_mean": 560.09375,
    "epsilon_norm": 6.906860828399658,
    "prediction_norm": 6.10200834274292,
    "timestamp": "2025-07-14T12:34:16.563301"
  },
  {
    "epoch": 1206,
    "batch_idx": 96,
    "loss": 0.3417361378669739,
    "learning_rate": 0.00034101668308379453,
    "timestep_mean": 462.125,
    "epsilon_norm": 8.16614818572998,
    "prediction_norm": 6.544188022613525,
    "timestamp": "2025-07-14T12:34:16.568460"
  },
  {
    "epoch": 1206,
    "batch_idx": 128,
    "loss": 0.23861953616142273,
    "learning_rate": 0.00034101668308379453,
    "timestep_mean": 434.59375,
    "epsilon_norm": 7.859043121337891,
    "prediction_norm": 6.486084938049316,
    "timestamp": "2025-07-14T12:34:16.573581"
  },
  {
    "epoch": 1206,
    "batch_idx": 160,
    "loss": 0.2849203646183014,
    "learning_rate": 0.00034101668308379453,
    "timestep_mean": 498.25,
    "epsilon_norm": 7.160813808441162,
    "prediction_norm": 5.885405540466309,
    "timestamp": "2025-07-14T12:34:16.578510"
  },
  {
    "epoch": 1206,
    "batch_idx": 192,
    "loss": 0.06706841289997101,
    "learning_rate": 0.00034101668308379453,
    "timestep_mean": 552.625,
    "epsilon_norm": 2.4216690063476562,
    "prediction_norm": 2.482435703277588,
    "timestamp": "2025-07-14T12:34:16.583792"
  },
  {
    "epoch": 1207,
    "batch_idx": 0,
    "loss": 0.5235894322395325,
    "learning_rate": 0.00034027224203391263,
    "timestep_mean": 433.71875,
    "epsilon_norm": 8.53249454498291,
    "prediction_norm": 5.7886881828308105,
    "timestamp": "2025-07-14T12:34:16.589268"
  },
  {
    "epoch": 1207,
    "batch_idx": 32,
    "loss": 0.2620237171649933,
    "learning_rate": 0.00034027224203391263,
    "timestep_mean": 560.125,
    "epsilon_norm": 8.98184871673584,
    "prediction_norm": 7.422063827514648,
    "timestamp": "2025-07-14T12:34:16.594299"
  },
  {
    "epoch": 1207,
    "batch_idx": 64,
    "loss": 0.22829416394233704,
    "learning_rate": 0.00034027224203391263,
    "timestep_mean": 615.9375,
    "epsilon_norm": 8.955080032348633,
    "prediction_norm": 7.04059362411499,
    "timestamp": "2025-07-14T12:34:16.599356"
  },
  {
    "epoch": 1207,
    "batch_idx": 96,
    "loss": 0.13506466150283813,
    "learning_rate": 0.00034027224203391263,
    "timestep_mean": 565.28125,
    "epsilon_norm": 7.092371940612793,
    "prediction_norm": 5.968917369842529,
    "timestamp": "2025-07-14T12:34:16.606864"
  },
  {
    "epoch": 1207,
    "batch_idx": 128,
    "loss": 0.39952248334884644,
    "learning_rate": 0.00034027224203391263,
    "timestep_mean": 572.65625,
    "epsilon_norm": 8.928078651428223,
    "prediction_norm": 6.81919002532959,
    "timestamp": "2025-07-14T12:34:16.615373"
  },
  {
    "epoch": 1207,
    "batch_idx": 160,
    "loss": 0.32868385314941406,
    "learning_rate": 0.00034027224203391263,
    "timestep_mean": 542.96875,
    "epsilon_norm": 6.656491756439209,
    "prediction_norm": 5.31638240814209,
    "timestamp": "2025-07-14T12:34:16.620793"
  },
  {
    "epoch": 1207,
    "batch_idx": 192,
    "loss": 0.7467851042747498,
    "learning_rate": 0.00034027224203391263,
    "timestep_mean": 482.5,
    "epsilon_norm": 4.00988245010376,
    "prediction_norm": 2.551314115524292,
    "timestamp": "2025-07-14T12:34:16.625974"
  },
  {
    "epoch": 1208,
    "batch_idx": 0,
    "loss": 0.20687036216259003,
    "learning_rate": 0.00033952819509639533,
    "timestep_mean": 447.40625,
    "epsilon_norm": 7.507627964019775,
    "prediction_norm": 6.552706241607666,
    "timestamp": "2025-07-14T12:34:16.630974"
  },
  {
    "epoch": 1208,
    "batch_idx": 32,
    "loss": 0.1803458034992218,
    "learning_rate": 0.00033952819509639533,
    "timestep_mean": 513.375,
    "epsilon_norm": 7.092203140258789,
    "prediction_norm": 5.852474689483643,
    "timestamp": "2025-07-14T12:34:16.636059"
  },
  {
    "epoch": 1208,
    "batch_idx": 64,
    "loss": 0.23823192715644836,
    "learning_rate": 0.00033952819509639533,
    "timestep_mean": 566.09375,
    "epsilon_norm": 6.844013214111328,
    "prediction_norm": 6.080811500549316,
    "timestamp": "2025-07-14T12:34:16.641325"
  },
  {
    "epoch": 1208,
    "batch_idx": 96,
    "loss": 0.18765124678611755,
    "learning_rate": 0.00033952819509639533,
    "timestep_mean": 503.125,
    "epsilon_norm": 6.888838291168213,
    "prediction_norm": 5.72062349319458,
    "timestamp": "2025-07-14T12:34:16.646464"
  },
  {
    "epoch": 1208,
    "batch_idx": 128,
    "loss": 0.5734554529190063,
    "learning_rate": 0.00033952819509639533,
    "timestep_mean": 438.96875,
    "epsilon_norm": 8.20152473449707,
    "prediction_norm": 5.814647197723389,
    "timestamp": "2025-07-14T12:34:16.651738"
  },
  {
    "epoch": 1208,
    "batch_idx": 160,
    "loss": 0.3572468161582947,
    "learning_rate": 0.00033952819509639533,
    "timestep_mean": 499.875,
    "epsilon_norm": 7.659084320068359,
    "prediction_norm": 6.921098232269287,
    "timestamp": "2025-07-14T12:34:16.657122"
  },
  {
    "epoch": 1208,
    "batch_idx": 192,
    "loss": 0.1514434814453125,
    "learning_rate": 0.00033952819509639533,
    "timestep_mean": 457.5,
    "epsilon_norm": 3.2094311714172363,
    "prediction_norm": 3.0662710666656494,
    "timestamp": "2025-07-14T12:34:16.662309"
  },
  {
    "epoch": 1209,
    "batch_idx": 0,
    "loss": 0.20190607011318207,
    "learning_rate": 0.0003387845441071046,
    "timestep_mean": 525.75,
    "epsilon_norm": 7.6270751953125,
    "prediction_norm": 6.72280740737915,
    "timestamp": "2025-07-14T12:34:16.667487"
  },
  {
    "epoch": 1209,
    "batch_idx": 32,
    "loss": 0.3884504437446594,
    "learning_rate": 0.0003387845441071046,
    "timestep_mean": 518.5625,
    "epsilon_norm": 8.710674285888672,
    "prediction_norm": 6.895409107208252,
    "timestamp": "2025-07-14T12:34:16.672696"
  },
  {
    "epoch": 1209,
    "batch_idx": 64,
    "loss": 0.3045033812522888,
    "learning_rate": 0.0003387845441071046,
    "timestep_mean": 505.78125,
    "epsilon_norm": 8.57031536102295,
    "prediction_norm": 6.758596897125244,
    "timestamp": "2025-07-14T12:34:16.677849"
  },
  {
    "epoch": 1209,
    "batch_idx": 96,
    "loss": 0.1485580950975418,
    "learning_rate": 0.0003387845441071046,
    "timestep_mean": 602.9375,
    "epsilon_norm": 7.507158279418945,
    "prediction_norm": 6.74908971786499,
    "timestamp": "2025-07-14T12:34:16.683013"
  },
  {
    "epoch": 1209,
    "batch_idx": 128,
    "loss": 0.18013958632946014,
    "learning_rate": 0.0003387845441071046,
    "timestep_mean": 437.53125,
    "epsilon_norm": 7.338495254516602,
    "prediction_norm": 6.566173076629639,
    "timestamp": "2025-07-14T12:34:16.688345"
  },
  {
    "epoch": 1209,
    "batch_idx": 160,
    "loss": 0.3434847593307495,
    "learning_rate": 0.0003387845441071046,
    "timestep_mean": 459.125,
    "epsilon_norm": 7.123905181884766,
    "prediction_norm": 6.159276962280273,
    "timestamp": "2025-07-14T12:34:16.693518"
  },
  {
    "epoch": 1209,
    "batch_idx": 192,
    "loss": 0.31796687841415405,
    "learning_rate": 0.0003387845441071046,
    "timestep_mean": 652.0,
    "epsilon_norm": 5.550998687744141,
    "prediction_norm": 5.6667304039001465,
    "timestamp": "2025-07-14T12:34:16.698705"
  },
  {
    "epoch": 1210,
    "batch_idx": 0,
    "loss": 0.16222608089447021,
    "learning_rate": 0.0003380412909009253,
    "timestep_mean": 516.375,
    "epsilon_norm": 8.132586479187012,
    "prediction_norm": 6.935673236846924,
    "timestamp": "2025-07-14T12:34:16.703973"
  },
  {
    "epoch": 1210,
    "batch_idx": 32,
    "loss": 0.42496463656425476,
    "learning_rate": 0.0003380412909009253,
    "timestep_mean": 500.0,
    "epsilon_norm": 8.569146156311035,
    "prediction_norm": 6.117410659790039,
    "timestamp": "2025-07-14T12:34:16.709255"
  },
  {
    "epoch": 1210,
    "batch_idx": 64,
    "loss": 0.1530407816171646,
    "learning_rate": 0.0003380412909009253,
    "timestep_mean": 524.21875,
    "epsilon_norm": 8.662239074707031,
    "prediction_norm": 8.548713684082031,
    "timestamp": "2025-07-14T12:34:16.714380"
  },
  {
    "epoch": 1210,
    "batch_idx": 96,
    "loss": 0.2990047335624695,
    "learning_rate": 0.0003380412909009253,
    "timestep_mean": 480.125,
    "epsilon_norm": 8.372847557067871,
    "prediction_norm": 7.138639450073242,
    "timestamp": "2025-07-14T12:34:16.719565"
  },
  {
    "epoch": 1210,
    "batch_idx": 128,
    "loss": 0.2743586599826813,
    "learning_rate": 0.0003380412909009253,
    "timestep_mean": 592.125,
    "epsilon_norm": 9.504270553588867,
    "prediction_norm": 8.26864242553711,
    "timestamp": "2025-07-14T12:34:16.724766"
  },
  {
    "epoch": 1210,
    "batch_idx": 160,
    "loss": 0.3470434248447418,
    "learning_rate": 0.0003380412909009253,
    "timestep_mean": 477.65625,
    "epsilon_norm": 7.035521507263184,
    "prediction_norm": 6.269134521484375,
    "timestamp": "2025-07-14T12:34:16.730062"
  },
  {
    "epoch": 1210,
    "batch_idx": 192,
    "loss": 0.1783600002527237,
    "learning_rate": 0.0003380412909009253,
    "timestep_mean": 470.625,
    "epsilon_norm": 3.602884531021118,
    "prediction_norm": 3.2090516090393066,
    "timestamp": "2025-07-14T12:34:16.735133"
  },
  {
    "epoch": 1211,
    "batch_idx": 0,
    "loss": 0.27704495191574097,
    "learning_rate": 0.0003372984373117609,
    "timestep_mean": 533.34375,
    "epsilon_norm": 8.24768352508545,
    "prediction_norm": 7.387140274047852,
    "timestamp": "2025-07-14T12:34:16.740423"
  },
  {
    "epoch": 1211,
    "batch_idx": 32,
    "loss": 0.3561263680458069,
    "learning_rate": 0.0003372984373117609,
    "timestep_mean": 518.1875,
    "epsilon_norm": 8.125365257263184,
    "prediction_norm": 6.771113872528076,
    "timestamp": "2025-07-14T12:34:16.745591"
  },
  {
    "epoch": 1211,
    "batch_idx": 64,
    "loss": 0.2855749726295471,
    "learning_rate": 0.0003372984373117609,
    "timestep_mean": 506.0625,
    "epsilon_norm": 9.787514686584473,
    "prediction_norm": 8.715435981750488,
    "timestamp": "2025-07-14T12:34:16.750933"
  },
  {
    "epoch": 1211,
    "batch_idx": 96,
    "loss": 0.2843879461288452,
    "learning_rate": 0.0003372984373117609,
    "timestep_mean": 502.8125,
    "epsilon_norm": 7.805853843688965,
    "prediction_norm": 7.3417816162109375,
    "timestamp": "2025-07-14T12:34:16.756116"
  },
  {
    "epoch": 1211,
    "batch_idx": 128,
    "loss": 0.22810617089271545,
    "learning_rate": 0.0003372984373117609,
    "timestep_mean": 459.4375,
    "epsilon_norm": 7.290616989135742,
    "prediction_norm": 5.833288192749023,
    "timestamp": "2025-07-14T12:34:16.761276"
  },
  {
    "epoch": 1211,
    "batch_idx": 160,
    "loss": 0.15872985124588013,
    "learning_rate": 0.0003372984373117609,
    "timestep_mean": 523.625,
    "epsilon_norm": 7.093080043792725,
    "prediction_norm": 6.540704250335693,
    "timestamp": "2025-07-14T12:34:16.766476"
  },
  {
    "epoch": 1211,
    "batch_idx": 192,
    "loss": 0.468885600566864,
    "learning_rate": 0.0003372984373117609,
    "timestep_mean": 586.375,
    "epsilon_norm": 3.874891757965088,
    "prediction_norm": 3.0362024307250977,
    "timestamp": "2025-07-14T12:34:16.771713"
  },
  {
    "epoch": 1212,
    "batch_idx": 0,
    "loss": 0.40056878328323364,
    "learning_rate": 0.00033655598517252883,
    "timestep_mean": 515.40625,
    "epsilon_norm": 8.927836418151855,
    "prediction_norm": 7.657991886138916,
    "timestamp": "2025-07-14T12:34:16.776957"
  },
  {
    "epoch": 1212,
    "batch_idx": 32,
    "loss": 0.3264879584312439,
    "learning_rate": 0.00033655598517252883,
    "timestep_mean": 451.9375,
    "epsilon_norm": 8.229032516479492,
    "prediction_norm": 6.9759840965271,
    "timestamp": "2025-07-14T12:34:16.782192"
  },
  {
    "epoch": 1212,
    "batch_idx": 64,
    "loss": 0.15728473663330078,
    "learning_rate": 0.00033655598517252883,
    "timestep_mean": 527.6875,
    "epsilon_norm": 8.434798240661621,
    "prediction_norm": 7.941483020782471,
    "timestamp": "2025-07-14T12:34:16.787344"
  },
  {
    "epoch": 1212,
    "batch_idx": 96,
    "loss": 0.4798824191093445,
    "learning_rate": 0.00033655598517252883,
    "timestep_mean": 392.6875,
    "epsilon_norm": 8.731948852539062,
    "prediction_norm": 6.683881759643555,
    "timestamp": "2025-07-14T12:34:16.792589"
  },
  {
    "epoch": 1212,
    "batch_idx": 128,
    "loss": 0.4294036030769348,
    "learning_rate": 0.00033655598517252883,
    "timestep_mean": 360.0,
    "epsilon_norm": 7.401409149169922,
    "prediction_norm": 6.022478103637695,
    "timestamp": "2025-07-14T12:34:16.797880"
  },
  {
    "epoch": 1212,
    "batch_idx": 160,
    "loss": 0.25032511353492737,
    "learning_rate": 0.00033655598517252883,
    "timestep_mean": 503.65625,
    "epsilon_norm": 7.706852912902832,
    "prediction_norm": 7.425032138824463,
    "timestamp": "2025-07-14T12:34:16.803031"
  },
  {
    "epoch": 1212,
    "batch_idx": 192,
    "loss": 0.15536516904830933,
    "learning_rate": 0.00033655598517252883,
    "timestep_mean": 384.75,
    "epsilon_norm": 3.9943642616271973,
    "prediction_norm": 3.791382312774658,
    "timestamp": "2025-07-14T12:34:16.808042"
  },
  {
    "epoch": 1213,
    "batch_idx": 0,
    "loss": 0.2303040325641632,
    "learning_rate": 0.00033581393631515574,
    "timestep_mean": 505.875,
    "epsilon_norm": 7.624687194824219,
    "prediction_norm": 6.405378341674805,
    "timestamp": "2025-07-14T12:34:16.813444"
  },
  {
    "epoch": 1213,
    "batch_idx": 32,
    "loss": 0.4789403975009918,
    "learning_rate": 0.00033581393631515574,
    "timestep_mean": 449.34375,
    "epsilon_norm": 9.300904273986816,
    "prediction_norm": 7.369921684265137,
    "timestamp": "2025-07-14T12:34:16.818629"
  },
  {
    "epoch": 1213,
    "batch_idx": 64,
    "loss": 0.28474047780036926,
    "learning_rate": 0.00033581393631515574,
    "timestep_mean": 457.78125,
    "epsilon_norm": 7.721477508544922,
    "prediction_norm": 6.2292304039001465,
    "timestamp": "2025-07-14T12:34:16.823783"
  },
  {
    "epoch": 1213,
    "batch_idx": 96,
    "loss": 0.14224708080291748,
    "learning_rate": 0.00033581393631515574,
    "timestep_mean": 474.84375,
    "epsilon_norm": 6.330554008483887,
    "prediction_norm": 6.012536525726318,
    "timestamp": "2025-07-14T12:34:16.828985"
  },
  {
    "epoch": 1213,
    "batch_idx": 128,
    "loss": 0.21738401055335999,
    "learning_rate": 0.00033581393631515574,
    "timestep_mean": 535.125,
    "epsilon_norm": 7.58629846572876,
    "prediction_norm": 6.5152812004089355,
    "timestamp": "2025-07-14T12:34:16.834252"
  },
  {
    "epoch": 1213,
    "batch_idx": 160,
    "loss": 0.18589961528778076,
    "learning_rate": 0.00033581393631515574,
    "timestep_mean": 527.71875,
    "epsilon_norm": 9.16568660736084,
    "prediction_norm": 7.692722797393799,
    "timestamp": "2025-07-14T12:34:16.839408"
  },
  {
    "epoch": 1213,
    "batch_idx": 192,
    "loss": 0.1277124583721161,
    "learning_rate": 0.00033581393631515574,
    "timestep_mean": 431.0,
    "epsilon_norm": 3.128453016281128,
    "prediction_norm": 2.483441114425659,
    "timestamp": "2025-07-14T12:34:16.844495"
  },
  {
    "epoch": 1214,
    "batch_idx": 0,
    "loss": 0.33002156019210815,
    "learning_rate": 0.0003350722925705736,
    "timestep_mean": 498.75,
    "epsilon_norm": 7.127905368804932,
    "prediction_norm": 5.39598274230957,
    "timestamp": "2025-07-14T12:34:16.849733"
  },
  {
    "epoch": 1214,
    "batch_idx": 32,
    "loss": 0.2203235626220703,
    "learning_rate": 0.0003350722925705736,
    "timestep_mean": 512.21875,
    "epsilon_norm": 9.426146507263184,
    "prediction_norm": 7.5002946853637695,
    "timestamp": "2025-07-14T12:34:16.855116"
  },
  {
    "epoch": 1214,
    "batch_idx": 64,
    "loss": 0.23667767643928528,
    "learning_rate": 0.0003350722925705736,
    "timestep_mean": 570.8125,
    "epsilon_norm": 7.3215413093566895,
    "prediction_norm": 5.831605911254883,
    "timestamp": "2025-07-14T12:34:16.860206"
  },
  {
    "epoch": 1214,
    "batch_idx": 96,
    "loss": 0.4152138829231262,
    "learning_rate": 0.0003350722925705736,
    "timestep_mean": 460.5625,
    "epsilon_norm": 8.692234992980957,
    "prediction_norm": 6.46440315246582,
    "timestamp": "2025-07-14T12:34:16.865080"
  },
  {
    "epoch": 1214,
    "batch_idx": 128,
    "loss": 0.2180623710155487,
    "learning_rate": 0.0003350722925705736,
    "timestep_mean": 579.5,
    "epsilon_norm": 8.765580177307129,
    "prediction_norm": 7.151217937469482,
    "timestamp": "2025-07-14T12:34:16.869942"
  },
  {
    "epoch": 1214,
    "batch_idx": 160,
    "loss": 0.44919517636299133,
    "learning_rate": 0.0003350722925705736,
    "timestep_mean": 489.28125,
    "epsilon_norm": 8.848043441772461,
    "prediction_norm": 6.733968734741211,
    "timestamp": "2025-07-14T12:34:16.874964"
  },
  {
    "epoch": 1214,
    "batch_idx": 192,
    "loss": 0.4294830560684204,
    "learning_rate": 0.0003350722925705736,
    "timestep_mean": 483.375,
    "epsilon_norm": 4.17331075668335,
    "prediction_norm": 3.2330753803253174,
    "timestamp": "2025-07-14T12:34:16.879891"
  },
  {
    "epoch": 1215,
    "batch_idx": 0,
    "loss": 0.24579697847366333,
    "learning_rate": 0.0003343310557687146,
    "timestep_mean": 452.96875,
    "epsilon_norm": 7.599996089935303,
    "prediction_norm": 6.0173563957214355,
    "timestamp": "2025-07-14T12:34:16.884705"
  },
  {
    "epoch": 1215,
    "batch_idx": 32,
    "loss": 0.5174038410186768,
    "learning_rate": 0.0003343310557687146,
    "timestep_mean": 508.625,
    "epsilon_norm": 9.44052505493164,
    "prediction_norm": 7.262918472290039,
    "timestamp": "2025-07-14T12:34:16.889597"
  },
  {
    "epoch": 1215,
    "batch_idx": 64,
    "loss": 0.24892401695251465,
    "learning_rate": 0.0003343310557687146,
    "timestep_mean": 530.8125,
    "epsilon_norm": 7.750713348388672,
    "prediction_norm": 6.541903972625732,
    "timestamp": "2025-07-14T12:34:16.894497"
  },
  {
    "epoch": 1215,
    "batch_idx": 96,
    "loss": 0.3178728222846985,
    "learning_rate": 0.0003343310557687146,
    "timestep_mean": 527.125,
    "epsilon_norm": 8.166523933410645,
    "prediction_norm": 6.836610794067383,
    "timestamp": "2025-07-14T12:34:16.899342"
  },
  {
    "epoch": 1215,
    "batch_idx": 128,
    "loss": 0.17014959454536438,
    "learning_rate": 0.0003343310557687146,
    "timestep_mean": 549.875,
    "epsilon_norm": 8.068890571594238,
    "prediction_norm": 7.129855632781982,
    "timestamp": "2025-07-14T12:34:16.904241"
  },
  {
    "epoch": 1215,
    "batch_idx": 160,
    "loss": 0.2857831120491028,
    "learning_rate": 0.0003343310557687146,
    "timestep_mean": 443.71875,
    "epsilon_norm": 6.937150478363037,
    "prediction_norm": 5.867516994476318,
    "timestamp": "2025-07-14T12:34:16.909097"
  },
  {
    "epoch": 1215,
    "batch_idx": 192,
    "loss": 0.13041411340236664,
    "learning_rate": 0.0003343310557687146,
    "timestep_mean": 526.875,
    "epsilon_norm": 2.7366604804992676,
    "prediction_norm": 2.5013303756713867,
    "timestamp": "2025-07-14T12:34:16.914044"
  },
  {
    "epoch": 1216,
    "batch_idx": 0,
    "loss": 0.3111231029033661,
    "learning_rate": 0.00033359022773850675,
    "timestep_mean": 506.5625,
    "epsilon_norm": 6.734628677368164,
    "prediction_norm": 5.775811195373535,
    "timestamp": "2025-07-14T12:34:16.918978"
  },
  {
    "epoch": 1216,
    "batch_idx": 32,
    "loss": 0.4170498549938202,
    "learning_rate": 0.00033359022773850675,
    "timestep_mean": 443.0625,
    "epsilon_norm": 8.562260627746582,
    "prediction_norm": 6.772972106933594,
    "timestamp": "2025-07-14T12:34:16.923826"
  },
  {
    "epoch": 1216,
    "batch_idx": 64,
    "loss": 0.27832484245300293,
    "learning_rate": 0.00033359022773850675,
    "timestep_mean": 461.53125,
    "epsilon_norm": 7.411451816558838,
    "prediction_norm": 6.489435195922852,
    "timestamp": "2025-07-14T12:34:16.928702"
  },
  {
    "epoch": 1216,
    "batch_idx": 96,
    "loss": 0.3775423765182495,
    "learning_rate": 0.00033359022773850675,
    "timestep_mean": 547.0,
    "epsilon_norm": 8.867327690124512,
    "prediction_norm": 7.628440856933594,
    "timestamp": "2025-07-14T12:34:16.933654"
  },
  {
    "epoch": 1216,
    "batch_idx": 128,
    "loss": 0.1756974458694458,
    "learning_rate": 0.00033359022773850675,
    "timestep_mean": 546.6875,
    "epsilon_norm": 7.281457424163818,
    "prediction_norm": 6.3151960372924805,
    "timestamp": "2025-07-14T12:34:16.938561"
  },
  {
    "epoch": 1216,
    "batch_idx": 160,
    "loss": 0.19038715958595276,
    "learning_rate": 0.00033359022773850675,
    "timestep_mean": 500.65625,
    "epsilon_norm": 9.061206817626953,
    "prediction_norm": 8.253697395324707,
    "timestamp": "2025-07-14T12:34:16.943398"
  },
  {
    "epoch": 1216,
    "batch_idx": 192,
    "loss": 0.37426307797431946,
    "learning_rate": 0.00033359022773850675,
    "timestep_mean": 511.625,
    "epsilon_norm": 5.0623860359191895,
    "prediction_norm": 4.471501350402832,
    "timestamp": "2025-07-14T12:34:16.948310"
  },
  {
    "epoch": 1217,
    "batch_idx": 0,
    "loss": 0.43153801560401917,
    "learning_rate": 0.00033284981030786966,
    "timestep_mean": 459.15625,
    "epsilon_norm": 8.910226821899414,
    "prediction_norm": 7.175556182861328,
    "timestamp": "2025-07-14T12:34:16.953183"
  },
  {
    "epoch": 1217,
    "batch_idx": 32,
    "loss": 0.31797799468040466,
    "learning_rate": 0.00033284981030786966,
    "timestep_mean": 506.09375,
    "epsilon_norm": 7.947300910949707,
    "prediction_norm": 7.1427412033081055,
    "timestamp": "2025-07-14T12:34:16.958045"
  },
  {
    "epoch": 1217,
    "batch_idx": 64,
    "loss": 0.4221891164779663,
    "learning_rate": 0.00033284981030786966,
    "timestep_mean": 488.84375,
    "epsilon_norm": 9.099239349365234,
    "prediction_norm": 7.472267150878906,
    "timestamp": "2025-07-14T12:34:16.963065"
  },
  {
    "epoch": 1217,
    "batch_idx": 96,
    "loss": 0.27715516090393066,
    "learning_rate": 0.00033284981030786966,
    "timestep_mean": 533.59375,
    "epsilon_norm": 7.259466648101807,
    "prediction_norm": 6.10917854309082,
    "timestamp": "2025-07-14T12:34:16.967991"
  },
  {
    "epoch": 1217,
    "batch_idx": 128,
    "loss": 0.34015873074531555,
    "learning_rate": 0.00033284981030786966,
    "timestep_mean": 478.75,
    "epsilon_norm": 8.950196266174316,
    "prediction_norm": 6.899170875549316,
    "timestamp": "2025-07-14T12:34:16.972876"
  },
  {
    "epoch": 1217,
    "batch_idx": 160,
    "loss": 0.3309085965156555,
    "learning_rate": 0.00033284981030786966,
    "timestep_mean": 386.09375,
    "epsilon_norm": 8.420785903930664,
    "prediction_norm": 6.960652828216553,
    "timestamp": "2025-07-14T12:34:16.977765"
  },
  {
    "epoch": 1217,
    "batch_idx": 192,
    "loss": 0.16241607069969177,
    "learning_rate": 0.00033284981030786966,
    "timestep_mean": 376.0,
    "epsilon_norm": 3.9839982986450195,
    "prediction_norm": 3.554569959640503,
    "timestamp": "2025-07-14T12:34:16.982612"
  },
  {
    "epoch": 1218,
    "batch_idx": 0,
    "loss": 0.40215903520584106,
    "learning_rate": 0.00033210980530370976,
    "timestep_mean": 525.03125,
    "epsilon_norm": 7.306520462036133,
    "prediction_norm": 6.092496871948242,
    "timestamp": "2025-07-14T12:34:16.987532"
  },
  {
    "epoch": 1218,
    "batch_idx": 32,
    "loss": 0.365366667509079,
    "learning_rate": 0.00033210980530370976,
    "timestep_mean": 439.90625,
    "epsilon_norm": 8.372971534729004,
    "prediction_norm": 6.853135108947754,
    "timestamp": "2025-07-14T12:34:16.992459"
  },
  {
    "epoch": 1218,
    "batch_idx": 64,
    "loss": 0.33764415979385376,
    "learning_rate": 0.00033210980530370976,
    "timestep_mean": 456.5625,
    "epsilon_norm": 8.366759300231934,
    "prediction_norm": 6.542193412780762,
    "timestamp": "2025-07-14T12:34:16.997345"
  },
  {
    "epoch": 1218,
    "batch_idx": 96,
    "loss": 0.2582358121871948,
    "learning_rate": 0.00033210980530370976,
    "timestep_mean": 559.875,
    "epsilon_norm": 7.316371440887451,
    "prediction_norm": 6.813235759735107,
    "timestamp": "2025-07-14T12:34:17.002176"
  },
  {
    "epoch": 1218,
    "batch_idx": 128,
    "loss": 0.27628153562545776,
    "learning_rate": 0.00033210980530370976,
    "timestep_mean": 458.84375,
    "epsilon_norm": 7.453302383422852,
    "prediction_norm": 6.074145793914795,
    "timestamp": "2025-07-14T12:34:17.007030"
  },
  {
    "epoch": 1218,
    "batch_idx": 160,
    "loss": 0.2635638415813446,
    "learning_rate": 0.00033210980530370976,
    "timestep_mean": 554.34375,
    "epsilon_norm": 8.290614128112793,
    "prediction_norm": 7.708354949951172,
    "timestamp": "2025-07-14T12:34:17.011911"
  },
  {
    "epoch": 1218,
    "batch_idx": 192,
    "loss": 0.2487744390964508,
    "learning_rate": 0.00033210980530370976,
    "timestep_mean": 386.5,
    "epsilon_norm": 2.709085702896118,
    "prediction_norm": 2.6032679080963135,
    "timestamp": "2025-07-14T12:34:17.016899"
  },
  {
    "epoch": 1219,
    "batch_idx": 0,
    "loss": 0.2866542637348175,
    "learning_rate": 0.0003313702145519158,
    "timestep_mean": 549.125,
    "epsilon_norm": 7.8883843421936035,
    "prediction_norm": 6.463654041290283,
    "timestamp": "2025-07-14T12:34:17.021782"
  },
  {
    "epoch": 1219,
    "batch_idx": 32,
    "loss": 0.3758509159088135,
    "learning_rate": 0.0003313702145519158,
    "timestep_mean": 503.875,
    "epsilon_norm": 8.52047061920166,
    "prediction_norm": 6.580878257751465,
    "timestamp": "2025-07-14T12:34:17.026654"
  },
  {
    "epoch": 1219,
    "batch_idx": 64,
    "loss": 0.2651197910308838,
    "learning_rate": 0.0003313702145519158,
    "timestep_mean": 458.0625,
    "epsilon_norm": 8.132660865783691,
    "prediction_norm": 6.751719951629639,
    "timestamp": "2025-07-14T12:34:17.031481"
  },
  {
    "epoch": 1219,
    "batch_idx": 96,
    "loss": 0.17141848802566528,
    "learning_rate": 0.0003313702145519158,
    "timestep_mean": 508.25,
    "epsilon_norm": 7.931975841522217,
    "prediction_norm": 7.111989498138428,
    "timestamp": "2025-07-14T12:34:17.036371"
  },
  {
    "epoch": 1219,
    "batch_idx": 128,
    "loss": 0.28077054023742676,
    "learning_rate": 0.0003313702145519158,
    "timestep_mean": 438.71875,
    "epsilon_norm": 8.457749366760254,
    "prediction_norm": 6.791680812835693,
    "timestamp": "2025-07-14T12:34:17.041247"
  },
  {
    "epoch": 1219,
    "batch_idx": 160,
    "loss": 0.22627772390842438,
    "learning_rate": 0.0003313702145519158,
    "timestep_mean": 551.9375,
    "epsilon_norm": 7.2688307762146,
    "prediction_norm": 5.789867877960205,
    "timestamp": "2025-07-14T12:34:17.046197"
  },
  {
    "epoch": 1219,
    "batch_idx": 192,
    "loss": 0.27796730399131775,
    "learning_rate": 0.0003313702145519158,
    "timestep_mean": 489.75,
    "epsilon_norm": 3.9146957397460938,
    "prediction_norm": 3.435551166534424,
    "timestamp": "2025-07-14T12:34:17.051018"
  },
  {
    "epoch": 1220,
    "batch_idx": 0,
    "loss": 0.1334252655506134,
    "learning_rate": 0.00033063103987735436,
    "timestep_mean": 567.78125,
    "epsilon_norm": 7.131261825561523,
    "prediction_norm": 6.305147171020508,
    "timestamp": "2025-07-14T12:34:17.055939"
  },
  {
    "epoch": 1220,
    "batch_idx": 32,
    "loss": 0.37104883790016174,
    "learning_rate": 0.00033063103987735436,
    "timestep_mean": 438.65625,
    "epsilon_norm": 8.960556030273438,
    "prediction_norm": 7.34591817855835,
    "timestamp": "2025-07-14T12:34:17.060802"
  },
  {
    "epoch": 1220,
    "batch_idx": 64,
    "loss": 0.23985356092453003,
    "learning_rate": 0.00033063103987735436,
    "timestep_mean": 425.9375,
    "epsilon_norm": 7.145652770996094,
    "prediction_norm": 6.150652885437012,
    "timestamp": "2025-07-14T12:34:17.065741"
  },
  {
    "epoch": 1220,
    "batch_idx": 96,
    "loss": 0.16978999972343445,
    "learning_rate": 0.00033063103987735436,
    "timestep_mean": 521.75,
    "epsilon_norm": 7.967190742492676,
    "prediction_norm": 7.462746620178223,
    "timestamp": "2025-07-14T12:34:17.070640"
  },
  {
    "epoch": 1220,
    "batch_idx": 128,
    "loss": 0.10188878327608109,
    "learning_rate": 0.00033063103987735436,
    "timestep_mean": 635.34375,
    "epsilon_norm": 7.361055850982666,
    "prediction_norm": 7.076393127441406,
    "timestamp": "2025-07-14T12:34:17.075599"
  },
  {
    "epoch": 1220,
    "batch_idx": 160,
    "loss": 0.13001082837581635,
    "learning_rate": 0.00033063103987735436,
    "timestep_mean": 529.28125,
    "epsilon_norm": 7.258124351501465,
    "prediction_norm": 6.6748456954956055,
    "timestamp": "2025-07-14T12:34:17.080564"
  },
  {
    "epoch": 1220,
    "batch_idx": 192,
    "loss": 0.5608938932418823,
    "learning_rate": 0.00033063103987735436,
    "timestep_mean": 417.25,
    "epsilon_norm": 5.470409870147705,
    "prediction_norm": 3.9380078315734863,
    "timestamp": "2025-07-14T12:34:17.085427"
  },
  {
    "epoch": 1221,
    "batch_idx": 0,
    "loss": 0.23930124938488007,
    "learning_rate": 0.0003298922831038657,
    "timestep_mean": 500.25,
    "epsilon_norm": 8.486132621765137,
    "prediction_norm": 7.2993998527526855,
    "timestamp": "2025-07-14T12:34:17.090360"
  },
  {
    "epoch": 1221,
    "batch_idx": 32,
    "loss": 0.42767229676246643,
    "learning_rate": 0.0003298922831038657,
    "timestep_mean": 466.84375,
    "epsilon_norm": 8.567946434020996,
    "prediction_norm": 6.906121730804443,
    "timestamp": "2025-07-14T12:34:17.095265"
  },
  {
    "epoch": 1221,
    "batch_idx": 64,
    "loss": 0.21563762426376343,
    "learning_rate": 0.0003298922831038657,
    "timestep_mean": 483.15625,
    "epsilon_norm": 7.349595069885254,
    "prediction_norm": 6.531790733337402,
    "timestamp": "2025-07-14T12:34:17.100231"
  },
  {
    "epoch": 1221,
    "batch_idx": 96,
    "loss": 0.35343486070632935,
    "learning_rate": 0.0003298922831038657,
    "timestep_mean": 484.125,
    "epsilon_norm": 7.603636741638184,
    "prediction_norm": 5.351241111755371,
    "timestamp": "2025-07-14T12:34:17.105054"
  },
  {
    "epoch": 1221,
    "batch_idx": 128,
    "loss": 0.3107144832611084,
    "learning_rate": 0.0003298922831038657,
    "timestep_mean": 518.34375,
    "epsilon_norm": 8.116569519042969,
    "prediction_norm": 6.75811767578125,
    "timestamp": "2025-07-14T12:34:17.110046"
  },
  {
    "epoch": 1221,
    "batch_idx": 160,
    "loss": 0.3223833739757538,
    "learning_rate": 0.0003298922831038657,
    "timestep_mean": 519.1875,
    "epsilon_norm": 8.213099479675293,
    "prediction_norm": 6.3767900466918945,
    "timestamp": "2025-07-14T12:34:17.114919"
  },
  {
    "epoch": 1221,
    "batch_idx": 192,
    "loss": 0.29324236512184143,
    "learning_rate": 0.0003298922831038657,
    "timestep_mean": 328.625,
    "epsilon_norm": 3.36959171295166,
    "prediction_norm": 3.1560564041137695,
    "timestamp": "2025-07-14T12:34:17.119798"
  },
  {
    "epoch": 1222,
    "batch_idx": 0,
    "loss": 0.15589995682239532,
    "learning_rate": 0.0003291539460542584,
    "timestep_mean": 604.40625,
    "epsilon_norm": 6.343914031982422,
    "prediction_norm": 5.674526691436768,
    "timestamp": "2025-07-14T12:34:17.124743"
  },
  {
    "epoch": 1222,
    "batch_idx": 32,
    "loss": 0.33664852380752563,
    "learning_rate": 0.0003291539460542584,
    "timestep_mean": 473.15625,
    "epsilon_norm": 7.489006996154785,
    "prediction_norm": 6.1601176261901855,
    "timestamp": "2025-07-14T12:34:17.129715"
  },
  {
    "epoch": 1222,
    "batch_idx": 64,
    "loss": 0.5363771915435791,
    "learning_rate": 0.0003291539460542584,
    "timestep_mean": 478.5,
    "epsilon_norm": 9.272345542907715,
    "prediction_norm": 7.280582427978516,
    "timestamp": "2025-07-14T12:34:17.134545"
  },
  {
    "epoch": 1222,
    "batch_idx": 96,
    "loss": 0.4933638274669647,
    "learning_rate": 0.0003291539460542584,
    "timestep_mean": 421.84375,
    "epsilon_norm": 9.362651824951172,
    "prediction_norm": 7.428077220916748,
    "timestamp": "2025-07-14T12:34:17.139400"
  },
  {
    "epoch": 1222,
    "batch_idx": 128,
    "loss": 0.22584228217601776,
    "learning_rate": 0.0003291539460542584,
    "timestep_mean": 531.0,
    "epsilon_norm": 7.571969509124756,
    "prediction_norm": 6.581223964691162,
    "timestamp": "2025-07-14T12:34:17.144356"
  },
  {
    "epoch": 1222,
    "batch_idx": 160,
    "loss": 0.3427492380142212,
    "learning_rate": 0.0003291539460542584,
    "timestep_mean": 466.875,
    "epsilon_norm": 9.332432746887207,
    "prediction_norm": 7.033434867858887,
    "timestamp": "2025-07-14T12:34:17.149233"
  },
  {
    "epoch": 1222,
    "batch_idx": 192,
    "loss": 0.10599155724048615,
    "learning_rate": 0.0003291539460542584,
    "timestep_mean": 543.625,
    "epsilon_norm": 3.5304174423217773,
    "prediction_norm": 3.527400493621826,
    "timestamp": "2025-07-14T12:34:17.154147"
  },
  {
    "epoch": 1223,
    "batch_idx": 0,
    "loss": 0.2570802867412567,
    "learning_rate": 0.00032841603055030604,
    "timestep_mean": 503.09375,
    "epsilon_norm": 7.73049783706665,
    "prediction_norm": 6.516453742980957,
    "timestamp": "2025-07-14T12:34:17.159173"
  },
  {
    "epoch": 1223,
    "batch_idx": 32,
    "loss": 0.20885367691516876,
    "learning_rate": 0.00032841603055030604,
    "timestep_mean": 454.875,
    "epsilon_norm": 7.439298152923584,
    "prediction_norm": 5.896289825439453,
    "timestamp": "2025-07-14T12:34:17.164072"
  },
  {
    "epoch": 1223,
    "batch_idx": 64,
    "loss": 0.21837490797042847,
    "learning_rate": 0.00032841603055030604,
    "timestep_mean": 508.40625,
    "epsilon_norm": 6.989291191101074,
    "prediction_norm": 6.5196404457092285,
    "timestamp": "2025-07-14T12:34:17.169015"
  },
  {
    "epoch": 1223,
    "batch_idx": 96,
    "loss": 0.2509083151817322,
    "learning_rate": 0.00032841603055030604,
    "timestep_mean": 476.46875,
    "epsilon_norm": 7.1091814041137695,
    "prediction_norm": 5.770143985748291,
    "timestamp": "2025-07-14T12:34:17.173949"
  },
  {
    "epoch": 1223,
    "batch_idx": 128,
    "loss": 0.26585620641708374,
    "learning_rate": 0.00032841603055030604,
    "timestep_mean": 597.4375,
    "epsilon_norm": 8.462367057800293,
    "prediction_norm": 8.024239540100098,
    "timestamp": "2025-07-14T12:34:17.178849"
  },
  {
    "epoch": 1223,
    "batch_idx": 160,
    "loss": 0.14696133136749268,
    "learning_rate": 0.00032841603055030604,
    "timestep_mean": 486.59375,
    "epsilon_norm": 8.138280868530273,
    "prediction_norm": 7.529675006866455,
    "timestamp": "2025-07-14T12:34:17.183781"
  },
  {
    "epoch": 1223,
    "batch_idx": 192,
    "loss": 0.7502322793006897,
    "learning_rate": 0.00032841603055030604,
    "timestep_mean": 331.25,
    "epsilon_norm": 4.781891345977783,
    "prediction_norm": 3.863581895828247,
    "timestamp": "2025-07-14T12:34:17.188657"
  },
  {
    "epoch": 1224,
    "batch_idx": 0,
    "loss": 0.3283655047416687,
    "learning_rate": 0.0003276785384127415,
    "timestep_mean": 469.21875,
    "epsilon_norm": 6.69482421875,
    "prediction_norm": 5.677106857299805,
    "timestamp": "2025-07-14T12:34:17.193511"
  },
  {
    "epoch": 1224,
    "batch_idx": 32,
    "loss": 0.19347445666790009,
    "learning_rate": 0.0003276785384127415,
    "timestep_mean": 429.59375,
    "epsilon_norm": 7.1591796875,
    "prediction_norm": 6.594249248504639,
    "timestamp": "2025-07-14T12:34:17.198368"
  },
  {
    "epoch": 1224,
    "batch_idx": 64,
    "loss": 0.20810383558273315,
    "learning_rate": 0.0003276785384127415,
    "timestep_mean": 525.84375,
    "epsilon_norm": 7.919041633605957,
    "prediction_norm": 7.116307735443115,
    "timestamp": "2025-07-14T12:34:17.203260"
  },
  {
    "epoch": 1224,
    "batch_idx": 96,
    "loss": 0.1995895355939865,
    "learning_rate": 0.0003276785384127415,
    "timestep_mean": 524.34375,
    "epsilon_norm": 7.451113700866699,
    "prediction_norm": 7.1008782386779785,
    "timestamp": "2025-07-14T12:34:17.208144"
  },
  {
    "epoch": 1224,
    "batch_idx": 128,
    "loss": 0.11729387193918228,
    "learning_rate": 0.0003276785384127415,
    "timestep_mean": 530.5625,
    "epsilon_norm": 8.25651741027832,
    "prediction_norm": 7.8724141120910645,
    "timestamp": "2025-07-14T12:34:17.213124"
  },
  {
    "epoch": 1224,
    "batch_idx": 160,
    "loss": 0.4334374666213989,
    "learning_rate": 0.0003276785384127415,
    "timestep_mean": 525.3125,
    "epsilon_norm": 7.9860076904296875,
    "prediction_norm": 5.784709930419922,
    "timestamp": "2025-07-14T12:34:17.217960"
  },
  {
    "epoch": 1224,
    "batch_idx": 192,
    "loss": 0.08435913920402527,
    "learning_rate": 0.0003276785384127415,
    "timestep_mean": 578.625,
    "epsilon_norm": 4.037629127502441,
    "prediction_norm": 4.024862766265869,
    "timestamp": "2025-07-14T12:34:17.222893"
  },
  {
    "epoch": 1225,
    "batch_idx": 0,
    "loss": 0.42915937304496765,
    "learning_rate": 0.00032694147146125353,
    "timestep_mean": 485.84375,
    "epsilon_norm": 8.815865516662598,
    "prediction_norm": 7.157081127166748,
    "timestamp": "2025-07-14T12:34:17.227816"
  },
  {
    "epoch": 1225,
    "batch_idx": 32,
    "loss": 0.181954026222229,
    "learning_rate": 0.00032694147146125353,
    "timestep_mean": 468.34375,
    "epsilon_norm": 8.4375,
    "prediction_norm": 7.19595193862915,
    "timestamp": "2025-07-14T12:34:17.232675"
  },
  {
    "epoch": 1225,
    "batch_idx": 64,
    "loss": 0.26319289207458496,
    "learning_rate": 0.00032694147146125353,
    "timestep_mean": 552.28125,
    "epsilon_norm": 7.723732948303223,
    "prediction_norm": 6.037527084350586,
    "timestamp": "2025-07-14T12:34:17.237604"
  },
  {
    "epoch": 1225,
    "batch_idx": 96,
    "loss": 0.10769955068826675,
    "learning_rate": 0.00032694147146125353,
    "timestep_mean": 604.5625,
    "epsilon_norm": 7.466616630554199,
    "prediction_norm": 7.355497360229492,
    "timestamp": "2025-07-14T12:34:17.242546"
  },
  {
    "epoch": 1225,
    "batch_idx": 128,
    "loss": 0.5914851427078247,
    "learning_rate": 0.00032694147146125353,
    "timestep_mean": 381.9375,
    "epsilon_norm": 8.879618644714355,
    "prediction_norm": 7.042933940887451,
    "timestamp": "2025-07-14T12:34:17.247439"
  },
  {
    "epoch": 1225,
    "batch_idx": 160,
    "loss": 0.17420463263988495,
    "learning_rate": 0.00032694147146125353,
    "timestep_mean": 470.40625,
    "epsilon_norm": 7.704118728637695,
    "prediction_norm": 6.42453145980835,
    "timestamp": "2025-07-14T12:34:17.252325"
  },
  {
    "epoch": 1225,
    "batch_idx": 192,
    "loss": 0.02930620312690735,
    "learning_rate": 0.00032694147146125353,
    "timestep_mean": 583.75,
    "epsilon_norm": 3.9489054679870605,
    "prediction_norm": 4.036506175994873,
    "timestamp": "2025-07-14T12:34:17.257176"
  },
  {
    "epoch": 1226,
    "batch_idx": 0,
    "loss": 0.2910535931587219,
    "learning_rate": 0.00032620483151448153,
    "timestep_mean": 491.53125,
    "epsilon_norm": 7.918042182922363,
    "prediction_norm": 6.216476917266846,
    "timestamp": "2025-07-14T12:34:17.262161"
  },
  {
    "epoch": 1226,
    "batch_idx": 32,
    "loss": 0.20293153822422028,
    "learning_rate": 0.00032620483151448153,
    "timestep_mean": 477.34375,
    "epsilon_norm": 6.389255046844482,
    "prediction_norm": 6.271122932434082,
    "timestamp": "2025-07-14T12:34:17.267127"
  },
  {
    "epoch": 1226,
    "batch_idx": 64,
    "loss": 0.31043529510498047,
    "learning_rate": 0.00032620483151448153,
    "timestep_mean": 504.96875,
    "epsilon_norm": 8.186921119689941,
    "prediction_norm": 6.795693397521973,
    "timestamp": "2025-07-14T12:34:17.272091"
  },
  {
    "epoch": 1226,
    "batch_idx": 96,
    "loss": 0.4249352812767029,
    "learning_rate": 0.00032620483151448153,
    "timestep_mean": 494.8125,
    "epsilon_norm": 8.646759033203125,
    "prediction_norm": 6.523667812347412,
    "timestamp": "2025-07-14T12:34:17.276966"
  },
  {
    "epoch": 1226,
    "batch_idx": 128,
    "loss": 0.19169190526008606,
    "learning_rate": 0.00032620483151448153,
    "timestep_mean": 425.0625,
    "epsilon_norm": 7.512332916259766,
    "prediction_norm": 6.411480903625488,
    "timestamp": "2025-07-14T12:34:17.281876"
  },
  {
    "epoch": 1226,
    "batch_idx": 160,
    "loss": 0.22693437337875366,
    "learning_rate": 0.00032620483151448153,
    "timestep_mean": 559.25,
    "epsilon_norm": 7.501599311828613,
    "prediction_norm": 6.555245399475098,
    "timestamp": "2025-07-14T12:34:17.286806"
  },
  {
    "epoch": 1226,
    "batch_idx": 192,
    "loss": 0.11408182978630066,
    "learning_rate": 0.00032620483151448153,
    "timestep_mean": 625.75,
    "epsilon_norm": 2.5119082927703857,
    "prediction_norm": 2.13547420501709,
    "timestamp": "2025-07-14T12:34:17.291697"
  },
  {
    "epoch": 1227,
    "batch_idx": 0,
    "loss": 0.2909848093986511,
    "learning_rate": 0.0003254686203900112,
    "timestep_mean": 486.0,
    "epsilon_norm": 7.13205099105835,
    "prediction_norm": 5.691183090209961,
    "timestamp": "2025-07-14T12:34:17.296595"
  },
  {
    "epoch": 1227,
    "batch_idx": 32,
    "loss": 0.3615718483924866,
    "learning_rate": 0.0003254686203900112,
    "timestep_mean": 453.875,
    "epsilon_norm": 8.569610595703125,
    "prediction_norm": 7.10839319229126,
    "timestamp": "2025-07-14T12:34:17.301509"
  },
  {
    "epoch": 1227,
    "batch_idx": 64,
    "loss": 0.32415127754211426,
    "learning_rate": 0.0003254686203900112,
    "timestep_mean": 456.09375,
    "epsilon_norm": 8.472380638122559,
    "prediction_norm": 6.435815334320068,
    "timestamp": "2025-07-14T12:34:17.306607"
  },
  {
    "epoch": 1227,
    "batch_idx": 96,
    "loss": 0.1794777810573578,
    "learning_rate": 0.0003254686203900112,
    "timestep_mean": 484.34375,
    "epsilon_norm": 7.146356105804443,
    "prediction_norm": 6.082352161407471,
    "timestamp": "2025-07-14T12:34:17.311757"
  },
  {
    "epoch": 1227,
    "batch_idx": 128,
    "loss": 0.2831970751285553,
    "learning_rate": 0.0003254686203900112,
    "timestep_mean": 499.0625,
    "epsilon_norm": 9.075581550598145,
    "prediction_norm": 7.832849979400635,
    "timestamp": "2025-07-14T12:34:17.316938"
  },
  {
    "epoch": 1227,
    "batch_idx": 160,
    "loss": 0.1488146036863327,
    "learning_rate": 0.0003254686203900112,
    "timestep_mean": 609.90625,
    "epsilon_norm": 7.6269049644470215,
    "prediction_norm": 6.598812580108643,
    "timestamp": "2025-07-14T12:34:17.322025"
  },
  {
    "epoch": 1227,
    "batch_idx": 192,
    "loss": 0.2396572381258011,
    "learning_rate": 0.0003254686203900112,
    "timestep_mean": 438.0,
    "epsilon_norm": 4.531918525695801,
    "prediction_norm": 3.38301682472229,
    "timestamp": "2025-07-14T12:34:17.327111"
  },
  {
    "epoch": 1228,
    "batch_idx": 0,
    "loss": 0.3356649577617645,
    "learning_rate": 0.00032473283990437044,
    "timestep_mean": 398.5,
    "epsilon_norm": 8.66558837890625,
    "prediction_norm": 7.0778374671936035,
    "timestamp": "2025-07-14T12:34:17.332177"
  },
  {
    "epoch": 1228,
    "batch_idx": 32,
    "loss": 0.23818756639957428,
    "learning_rate": 0.00032473283990437044,
    "timestep_mean": 464.78125,
    "epsilon_norm": 7.60601806640625,
    "prediction_norm": 6.412683963775635,
    "timestamp": "2025-07-14T12:34:17.337382"
  },
  {
    "epoch": 1228,
    "batch_idx": 64,
    "loss": 0.27853086590766907,
    "learning_rate": 0.00032473283990437044,
    "timestep_mean": 491.21875,
    "epsilon_norm": 7.7261962890625,
    "prediction_norm": 6.615458011627197,
    "timestamp": "2025-07-14T12:34:17.342587"
  },
  {
    "epoch": 1228,
    "batch_idx": 96,
    "loss": 0.30272233486175537,
    "learning_rate": 0.00032473283990437044,
    "timestep_mean": 518.9375,
    "epsilon_norm": 7.59579610824585,
    "prediction_norm": 6.618674278259277,
    "timestamp": "2025-07-14T12:34:17.347602"
  },
  {
    "epoch": 1228,
    "batch_idx": 128,
    "loss": 0.14560845494270325,
    "learning_rate": 0.00032473283990437044,
    "timestep_mean": 504.8125,
    "epsilon_norm": 7.939084053039551,
    "prediction_norm": 7.230240345001221,
    "timestamp": "2025-07-14T12:34:17.352674"
  },
  {
    "epoch": 1228,
    "batch_idx": 160,
    "loss": 0.2665705382823944,
    "learning_rate": 0.00032473283990437044,
    "timestep_mean": 464.75,
    "epsilon_norm": 7.41662073135376,
    "prediction_norm": 6.871918678283691,
    "timestamp": "2025-07-14T12:34:17.357883"
  },
  {
    "epoch": 1228,
    "batch_idx": 192,
    "loss": 0.2096703201532364,
    "learning_rate": 0.00032473283990437044,
    "timestep_mean": 557.375,
    "epsilon_norm": 4.297879695892334,
    "prediction_norm": 4.076748847961426,
    "timestamp": "2025-07-14T12:34:17.363097"
  },
  {
    "epoch": 1229,
    "batch_idx": 0,
    "loss": 0.216373011469841,
    "learning_rate": 0.0003239974918730245,
    "timestep_mean": 506.0625,
    "epsilon_norm": 7.386116981506348,
    "prediction_norm": 6.159748554229736,
    "timestamp": "2025-07-14T12:34:17.368186"
  },
  {
    "epoch": 1229,
    "batch_idx": 32,
    "loss": 0.2113177627325058,
    "learning_rate": 0.0003239974918730245,
    "timestep_mean": 531.5,
    "epsilon_norm": 8.216021537780762,
    "prediction_norm": 7.401994705200195,
    "timestamp": "2025-07-14T12:34:17.373222"
  },
  {
    "epoch": 1229,
    "batch_idx": 64,
    "loss": 0.1658644676208496,
    "learning_rate": 0.0003239974918730245,
    "timestep_mean": 608.375,
    "epsilon_norm": 7.799187660217285,
    "prediction_norm": 7.095669746398926,
    "timestamp": "2025-07-14T12:34:17.378358"
  },
  {
    "epoch": 1229,
    "batch_idx": 96,
    "loss": 0.34112679958343506,
    "learning_rate": 0.0003239974918730245,
    "timestep_mean": 456.53125,
    "epsilon_norm": 8.188952445983887,
    "prediction_norm": 6.481450080871582,
    "timestamp": "2025-07-14T12:34:17.383507"
  },
  {
    "epoch": 1229,
    "batch_idx": 128,
    "loss": 0.3446826636791229,
    "learning_rate": 0.0003239974918730245,
    "timestep_mean": 533.625,
    "epsilon_norm": 8.17104434967041,
    "prediction_norm": 7.228039741516113,
    "timestamp": "2025-07-14T12:34:17.388549"
  },
  {
    "epoch": 1229,
    "batch_idx": 160,
    "loss": 0.23506224155426025,
    "learning_rate": 0.0003239974918730245,
    "timestep_mean": 475.65625,
    "epsilon_norm": 6.031513214111328,
    "prediction_norm": 5.324934482574463,
    "timestamp": "2025-07-14T12:34:17.393701"
  },
  {
    "epoch": 1229,
    "batch_idx": 192,
    "loss": 0.4714163541793823,
    "learning_rate": 0.0003239974918730245,
    "timestep_mean": 573.375,
    "epsilon_norm": 4.479625225067139,
    "prediction_norm": 3.8236703872680664,
    "timestamp": "2025-07-14T12:34:17.398866"
  },
  {
    "epoch": 1230,
    "batch_idx": 0,
    "loss": 0.226166769862175,
    "learning_rate": 0.0003232625781103715,
    "timestep_mean": 603.6875,
    "epsilon_norm": 8.378329277038574,
    "prediction_norm": 7.296262264251709,
    "timestamp": "2025-07-14T12:34:17.403956"
  },
  {
    "epoch": 1230,
    "batch_idx": 32,
    "loss": 0.33752870559692383,
    "learning_rate": 0.0003232625781103715,
    "timestep_mean": 578.34375,
    "epsilon_norm": 7.878110408782959,
    "prediction_norm": 6.448927402496338,
    "timestamp": "2025-07-14T12:34:17.408942"
  },
  {
    "epoch": 1230,
    "batch_idx": 64,
    "loss": 0.17744186520576477,
    "learning_rate": 0.0003232625781103715,
    "timestep_mean": 556.40625,
    "epsilon_norm": 7.896050930023193,
    "prediction_norm": 7.162714004516602,
    "timestamp": "2025-07-14T12:34:17.413847"
  },
  {
    "epoch": 1230,
    "batch_idx": 96,
    "loss": 0.5210340619087219,
    "learning_rate": 0.0003232625781103715,
    "timestep_mean": 471.75,
    "epsilon_norm": 9.721034049987793,
    "prediction_norm": 7.620772838592529,
    "timestamp": "2025-07-14T12:34:17.419199"
  },
  {
    "epoch": 1230,
    "batch_idx": 128,
    "loss": 0.2339664250612259,
    "learning_rate": 0.0003232625781103715,
    "timestep_mean": 521.5,
    "epsilon_norm": 7.4133782386779785,
    "prediction_norm": 6.5534539222717285,
    "timestamp": "2025-07-14T12:34:17.424384"
  },
  {
    "epoch": 1230,
    "batch_idx": 160,
    "loss": 0.41853150725364685,
    "learning_rate": 0.0003232625781103715,
    "timestep_mean": 448.15625,
    "epsilon_norm": 7.6640238761901855,
    "prediction_norm": 6.104137420654297,
    "timestamp": "2025-07-14T12:34:17.429663"
  },
  {
    "epoch": 1230,
    "batch_idx": 192,
    "loss": 0.13555890321731567,
    "learning_rate": 0.0003232625781103715,
    "timestep_mean": 517.0,
    "epsilon_norm": 2.9830212593078613,
    "prediction_norm": 2.774137258529663,
    "timestamp": "2025-07-14T12:34:17.434787"
  },
  {
    "epoch": 1231,
    "batch_idx": 0,
    "loss": 0.17190703749656677,
    "learning_rate": 0.00032252810042973786,
    "timestep_mean": 564.8125,
    "epsilon_norm": 6.440562725067139,
    "prediction_norm": 6.181487083435059,
    "timestamp": "2025-07-14T12:34:17.440107"
  },
  {
    "epoch": 1231,
    "batch_idx": 32,
    "loss": 0.39119255542755127,
    "learning_rate": 0.00032252810042973786,
    "timestep_mean": 432.21875,
    "epsilon_norm": 8.08643913269043,
    "prediction_norm": 5.891567707061768,
    "timestamp": "2025-07-14T12:34:17.445458"
  },
  {
    "epoch": 1231,
    "batch_idx": 64,
    "loss": 0.29085010290145874,
    "learning_rate": 0.00032252810042973786,
    "timestep_mean": 442.15625,
    "epsilon_norm": 8.44872760772705,
    "prediction_norm": 6.615917205810547,
    "timestamp": "2025-07-14T12:34:17.450699"
  },
  {
    "epoch": 1231,
    "batch_idx": 96,
    "loss": 0.2880289554595947,
    "learning_rate": 0.00032252810042973786,
    "timestep_mean": 481.53125,
    "epsilon_norm": 9.286293029785156,
    "prediction_norm": 7.5713067054748535,
    "timestamp": "2025-07-14T12:34:17.455874"
  },
  {
    "epoch": 1231,
    "batch_idx": 128,
    "loss": 0.41578564047813416,
    "learning_rate": 0.00032252810042973786,
    "timestep_mean": 494.28125,
    "epsilon_norm": 7.874970436096191,
    "prediction_norm": 5.828432083129883,
    "timestamp": "2025-07-14T12:34:17.460837"
  },
  {
    "epoch": 1231,
    "batch_idx": 160,
    "loss": 0.14856404066085815,
    "learning_rate": 0.00032252810042973786,
    "timestep_mean": 591.75,
    "epsilon_norm": 8.394039154052734,
    "prediction_norm": 7.6615118980407715,
    "timestamp": "2025-07-14T12:34:17.465768"
  },
  {
    "epoch": 1231,
    "batch_idx": 192,
    "loss": 0.22213415801525116,
    "learning_rate": 0.00032252810042973786,
    "timestep_mean": 392.375,
    "epsilon_norm": 2.6160216331481934,
    "prediction_norm": 2.2229578495025635,
    "timestamp": "2025-07-14T12:34:17.470630"
  },
  {
    "epoch": 1232,
    "batch_idx": 0,
    "loss": 0.3573019206523895,
    "learning_rate": 0.00032179406064337463,
    "timestep_mean": 530.71875,
    "epsilon_norm": 8.920158386230469,
    "prediction_norm": 7.100107192993164,
    "timestamp": "2025-07-14T12:34:17.475611"
  },
  {
    "epoch": 1232,
    "batch_idx": 32,
    "loss": 0.19153062999248505,
    "learning_rate": 0.00032179406064337463,
    "timestep_mean": 571.5625,
    "epsilon_norm": 7.467126846313477,
    "prediction_norm": 6.978845596313477,
    "timestamp": "2025-07-14T12:34:17.480807"
  },
  {
    "epoch": 1232,
    "batch_idx": 64,
    "loss": 0.16845020651817322,
    "learning_rate": 0.00032179406064337463,
    "timestep_mean": 472.15625,
    "epsilon_norm": 7.333803176879883,
    "prediction_norm": 6.923312187194824,
    "timestamp": "2025-07-14T12:34:17.485985"
  },
  {
    "epoch": 1232,
    "batch_idx": 96,
    "loss": 0.3063580393791199,
    "learning_rate": 0.00032179406064337463,
    "timestep_mean": 486.53125,
    "epsilon_norm": 8.69960880279541,
    "prediction_norm": 7.10075569152832,
    "timestamp": "2025-07-14T12:34:17.491090"
  },
  {
    "epoch": 1232,
    "batch_idx": 128,
    "loss": 0.29431769251823425,
    "learning_rate": 0.00032179406064337463,
    "timestep_mean": 589.8125,
    "epsilon_norm": 9.035759925842285,
    "prediction_norm": 7.613832473754883,
    "timestamp": "2025-07-14T12:34:17.496228"
  },
  {
    "epoch": 1232,
    "batch_idx": 160,
    "loss": 0.21711163222789764,
    "learning_rate": 0.00032179406064337463,
    "timestep_mean": 528.46875,
    "epsilon_norm": 7.151636600494385,
    "prediction_norm": 6.330673694610596,
    "timestamp": "2025-07-14T12:34:17.501213"
  },
  {
    "epoch": 1232,
    "batch_idx": 192,
    "loss": 0.14464989304542542,
    "learning_rate": 0.00032179406064337463,
    "timestep_mean": 569.75,
    "epsilon_norm": 2.6427574157714844,
    "prediction_norm": 2.1010797023773193,
    "timestamp": "2025-07-14T12:34:17.506360"
  },
  {
    "epoch": 1233,
    "batch_idx": 0,
    "loss": 0.35284292697906494,
    "learning_rate": 0.0003210604605624518,
    "timestep_mean": 471.375,
    "epsilon_norm": 8.576273918151855,
    "prediction_norm": 7.173134803771973,
    "timestamp": "2025-07-14T12:34:17.511478"
  },
  {
    "epoch": 1233,
    "batch_idx": 32,
    "loss": 0.1283857226371765,
    "learning_rate": 0.0003210604605624518,
    "timestep_mean": 552.53125,
    "epsilon_norm": 7.87757682800293,
    "prediction_norm": 6.831735610961914,
    "timestamp": "2025-07-14T12:34:17.516738"
  },
  {
    "epoch": 1233,
    "batch_idx": 64,
    "loss": 0.41949957609176636,
    "learning_rate": 0.0003210604605624518,
    "timestep_mean": 498.5,
    "epsilon_norm": 7.777579307556152,
    "prediction_norm": 5.6377973556518555,
    "timestamp": "2025-07-14T12:34:17.521919"
  },
  {
    "epoch": 1233,
    "batch_idx": 96,
    "loss": 0.18146778643131256,
    "learning_rate": 0.0003210604605624518,
    "timestep_mean": 585.21875,
    "epsilon_norm": 7.853188991546631,
    "prediction_norm": 6.477553844451904,
    "timestamp": "2025-07-14T12:34:17.526996"
  },
  {
    "epoch": 1233,
    "batch_idx": 128,
    "loss": 0.3949894607067108,
    "learning_rate": 0.0003210604605624518,
    "timestep_mean": 552.34375,
    "epsilon_norm": 8.687967300415039,
    "prediction_norm": 6.865710258483887,
    "timestamp": "2025-07-14T12:34:17.532096"
  },
  {
    "epoch": 1233,
    "batch_idx": 160,
    "loss": 0.38985157012939453,
    "learning_rate": 0.0003210604605624518,
    "timestep_mean": 468.53125,
    "epsilon_norm": 9.81656551361084,
    "prediction_norm": 7.620082378387451,
    "timestamp": "2025-07-14T12:34:17.537314"
  },
  {
    "epoch": 1233,
    "batch_idx": 192,
    "loss": 0.05352116376161575,
    "learning_rate": 0.0003210604605624518,
    "timestep_mean": 717.75,
    "epsilon_norm": 3.6977622509002686,
    "prediction_norm": 3.7582521438598633,
    "timestamp": "2025-07-14T12:34:17.542619"
  },
  {
    "epoch": 1234,
    "batch_idx": 0,
    "loss": 0.3517235517501831,
    "learning_rate": 0.00032032730199705467,
    "timestep_mean": 443.0625,
    "epsilon_norm": 9.357665061950684,
    "prediction_norm": 8.20121955871582,
    "timestamp": "2025-07-14T12:34:17.547699"
  },
  {
    "epoch": 1234,
    "batch_idx": 32,
    "loss": 0.1725405901670456,
    "learning_rate": 0.00032032730199705467,
    "timestep_mean": 509.25,
    "epsilon_norm": 7.588502883911133,
    "prediction_norm": 6.397815704345703,
    "timestamp": "2025-07-14T12:34:17.552986"
  },
  {
    "epoch": 1234,
    "batch_idx": 64,
    "loss": 0.31665241718292236,
    "learning_rate": 0.00032032730199705467,
    "timestep_mean": 474.09375,
    "epsilon_norm": 8.367952346801758,
    "prediction_norm": 7.722506523132324,
    "timestamp": "2025-07-14T12:34:17.558165"
  },
  {
    "epoch": 1234,
    "batch_idx": 96,
    "loss": 0.24802887439727783,
    "learning_rate": 0.00032032730199705467,
    "timestep_mean": 493.75,
    "epsilon_norm": 7.952134609222412,
    "prediction_norm": 6.747077941894531,
    "timestamp": "2025-07-14T12:34:17.563384"
  },
  {
    "epoch": 1234,
    "batch_idx": 128,
    "loss": 0.1641666442155838,
    "learning_rate": 0.00032032730199705467,
    "timestep_mean": 555.375,
    "epsilon_norm": 6.656894207000732,
    "prediction_norm": 5.845461368560791,
    "timestamp": "2025-07-14T12:34:17.568578"
  },
  {
    "epoch": 1234,
    "batch_idx": 160,
    "loss": 0.15423841774463654,
    "learning_rate": 0.00032032730199705467,
    "timestep_mean": 560.84375,
    "epsilon_norm": 8.198365211486816,
    "prediction_norm": 7.896002292633057,
    "timestamp": "2025-07-14T12:34:17.573759"
  },
  {
    "epoch": 1234,
    "batch_idx": 192,
    "loss": 0.8146225810050964,
    "learning_rate": 0.00032032730199705467,
    "timestep_mean": 572.5,
    "epsilon_norm": 6.01567268371582,
    "prediction_norm": 4.321235656738281,
    "timestamp": "2025-07-14T12:34:17.578823"
  },
  {
    "epoch": 1235,
    "batch_idx": 0,
    "loss": 0.3993750810623169,
    "learning_rate": 0.000319594586756179,
    "timestep_mean": 535.90625,
    "epsilon_norm": 8.574580192565918,
    "prediction_norm": 6.985868453979492,
    "timestamp": "2025-07-14T12:34:17.583874"
  },
  {
    "epoch": 1235,
    "batch_idx": 32,
    "loss": 0.20177645981311798,
    "learning_rate": 0.000319594586756179,
    "timestep_mean": 510.59375,
    "epsilon_norm": 8.398245811462402,
    "prediction_norm": 7.703114986419678,
    "timestamp": "2025-07-14T12:34:17.589091"
  },
  {
    "epoch": 1235,
    "batch_idx": 64,
    "loss": 0.4521695375442505,
    "learning_rate": 0.000319594586756179,
    "timestep_mean": 474.84375,
    "epsilon_norm": 8.309985160827637,
    "prediction_norm": 6.131427764892578,
    "timestamp": "2025-07-14T12:34:17.594318"
  },
  {
    "epoch": 1235,
    "batch_idx": 96,
    "loss": 0.34419912099838257,
    "learning_rate": 0.000319594586756179,
    "timestep_mean": 421.90625,
    "epsilon_norm": 8.879127502441406,
    "prediction_norm": 7.726588726043701,
    "timestamp": "2025-07-14T12:34:17.599555"
  },
  {
    "epoch": 1235,
    "batch_idx": 128,
    "loss": 0.20080780982971191,
    "learning_rate": 0.000319594586756179,
    "timestep_mean": 520.9375,
    "epsilon_norm": 8.245695114135742,
    "prediction_norm": 7.4620866775512695,
    "timestamp": "2025-07-14T12:34:17.604847"
  },
  {
    "epoch": 1235,
    "batch_idx": 160,
    "loss": 0.26517894864082336,
    "learning_rate": 0.000319594586756179,
    "timestep_mean": 528.59375,
    "epsilon_norm": 8.202168464660645,
    "prediction_norm": 7.092074394226074,
    "timestamp": "2025-07-14T12:34:17.610122"
  },
  {
    "epoch": 1235,
    "batch_idx": 192,
    "loss": 0.2695154547691345,
    "learning_rate": 0.000319594586756179,
    "timestep_mean": 627.625,
    "epsilon_norm": 3.259856700897217,
    "prediction_norm": 2.5598742961883545,
    "timestamp": "2025-07-14T12:34:17.615222"
  },
  {
    "epoch": 1236,
    "batch_idx": 0,
    "loss": 0.39218568801879883,
    "learning_rate": 0.0003188623166477271,
    "timestep_mean": 477.625,
    "epsilon_norm": 7.839212894439697,
    "prediction_norm": 6.665761470794678,
    "timestamp": "2025-07-14T12:34:17.620290"
  },
  {
    "epoch": 1236,
    "batch_idx": 32,
    "loss": 0.2897762060165405,
    "learning_rate": 0.0003188623166477271,
    "timestep_mean": 492.46875,
    "epsilon_norm": 7.955533504486084,
    "prediction_norm": 6.796880722045898,
    "timestamp": "2025-07-14T12:34:17.625562"
  },
  {
    "epoch": 1236,
    "batch_idx": 64,
    "loss": 0.23331649601459503,
    "learning_rate": 0.0003188623166477271,
    "timestep_mean": 554.53125,
    "epsilon_norm": 8.201333045959473,
    "prediction_norm": 7.0099053382873535,
    "timestamp": "2025-07-14T12:34:17.630755"
  },
  {
    "epoch": 1236,
    "batch_idx": 96,
    "loss": 0.3570186495780945,
    "learning_rate": 0.0003188623166477271,
    "timestep_mean": 391.1875,
    "epsilon_norm": 7.866693496704102,
    "prediction_norm": 5.789994239807129,
    "timestamp": "2025-07-14T12:34:17.636034"
  },
  {
    "epoch": 1236,
    "batch_idx": 128,
    "loss": 0.22579069435596466,
    "learning_rate": 0.0003188623166477271,
    "timestep_mean": 506.875,
    "epsilon_norm": 8.15361213684082,
    "prediction_norm": 7.57845401763916,
    "timestamp": "2025-07-14T12:34:17.641180"
  },
  {
    "epoch": 1236,
    "batch_idx": 160,
    "loss": 0.36650550365448,
    "learning_rate": 0.0003188623166477271,
    "timestep_mean": 487.25,
    "epsilon_norm": 8.502545356750488,
    "prediction_norm": 7.409036159515381,
    "timestamp": "2025-07-14T12:34:17.646310"
  },
  {
    "epoch": 1236,
    "batch_idx": 192,
    "loss": 0.16810345649719238,
    "learning_rate": 0.0003188623166477271,
    "timestep_mean": 278.625,
    "epsilon_norm": 3.127582311630249,
    "prediction_norm": 2.6268362998962402,
    "timestamp": "2025-07-14T12:34:17.651051"
  },
  {
    "epoch": 1237,
    "batch_idx": 0,
    "loss": 0.41771942377090454,
    "learning_rate": 0.00031813049347850244,
    "timestep_mean": 517.5,
    "epsilon_norm": 9.487327575683594,
    "prediction_norm": 7.356405258178711,
    "timestamp": "2025-07-14T12:34:17.655912"
  },
  {
    "epoch": 1237,
    "batch_idx": 32,
    "loss": 0.35125109553337097,
    "learning_rate": 0.00031813049347850244,
    "timestep_mean": 437.15625,
    "epsilon_norm": 7.909659385681152,
    "prediction_norm": 6.355588912963867,
    "timestamp": "2025-07-14T12:34:17.660774"
  },
  {
    "epoch": 1237,
    "batch_idx": 64,
    "loss": 0.2503144443035126,
    "learning_rate": 0.00031813049347850244,
    "timestep_mean": 579.65625,
    "epsilon_norm": 8.619107246398926,
    "prediction_norm": 7.500656604766846,
    "timestamp": "2025-07-14T12:34:17.665742"
  },
  {
    "epoch": 1237,
    "batch_idx": 96,
    "loss": 0.2728826403617859,
    "learning_rate": 0.00031813049347850244,
    "timestep_mean": 525.53125,
    "epsilon_norm": 9.279471397399902,
    "prediction_norm": 8.279544830322266,
    "timestamp": "2025-07-14T12:34:17.670613"
  },
  {
    "epoch": 1237,
    "batch_idx": 128,
    "loss": 0.2786875069141388,
    "learning_rate": 0.00031813049347850244,
    "timestep_mean": 479.0,
    "epsilon_norm": 8.663064956665039,
    "prediction_norm": 6.60484504699707,
    "timestamp": "2025-07-14T12:34:17.675423"
  },
  {
    "epoch": 1237,
    "batch_idx": 160,
    "loss": 0.33271461725234985,
    "learning_rate": 0.00031813049347850244,
    "timestep_mean": 486.375,
    "epsilon_norm": 7.818798542022705,
    "prediction_norm": 6.334770679473877,
    "timestamp": "2025-07-14T12:34:17.680286"
  },
  {
    "epoch": 1237,
    "batch_idx": 192,
    "loss": 0.19818875193595886,
    "learning_rate": 0.00031813049347850244,
    "timestep_mean": 425.5,
    "epsilon_norm": 3.3123538494110107,
    "prediction_norm": 2.488363027572632,
    "timestamp": "2025-07-14T12:34:17.685128"
  },
  {
    "epoch": 1238,
    "batch_idx": 0,
    "loss": 0.24684810638427734,
    "learning_rate": 0.000317399119054206,
    "timestep_mean": 493.59375,
    "epsilon_norm": 6.876364707946777,
    "prediction_norm": 5.190441131591797,
    "timestamp": "2025-07-14T12:34:17.690025"
  },
  {
    "epoch": 1238,
    "batch_idx": 32,
    "loss": 0.4014143943786621,
    "learning_rate": 0.000317399119054206,
    "timestep_mean": 460.71875,
    "epsilon_norm": 8.382076263427734,
    "prediction_norm": 6.861968517303467,
    "timestamp": "2025-07-14T12:34:17.694993"
  },
  {
    "epoch": 1238,
    "batch_idx": 64,
    "loss": 0.24562829732894897,
    "learning_rate": 0.000317399119054206,
    "timestep_mean": 476.28125,
    "epsilon_norm": 7.426967620849609,
    "prediction_norm": 6.084003448486328,
    "timestamp": "2025-07-14T12:34:17.699998"
  },
  {
    "epoch": 1238,
    "batch_idx": 96,
    "loss": 0.4012352228164673,
    "learning_rate": 0.000317399119054206,
    "timestep_mean": 467.71875,
    "epsilon_norm": 8.786195755004883,
    "prediction_norm": 7.023281574249268,
    "timestamp": "2025-07-14T12:34:17.704890"
  },
  {
    "epoch": 1238,
    "batch_idx": 128,
    "loss": 0.326253741979599,
    "learning_rate": 0.000317399119054206,
    "timestep_mean": 496.21875,
    "epsilon_norm": 8.425538063049316,
    "prediction_norm": 7.526935577392578,
    "timestamp": "2025-07-14T12:34:17.709733"
  },
  {
    "epoch": 1238,
    "batch_idx": 160,
    "loss": 0.29840826988220215,
    "learning_rate": 0.000317399119054206,
    "timestep_mean": 484.84375,
    "epsilon_norm": 7.070646286010742,
    "prediction_norm": 6.478342533111572,
    "timestamp": "2025-07-14T12:34:17.714564"
  },
  {
    "epoch": 1238,
    "batch_idx": 192,
    "loss": 0.6589577198028564,
    "learning_rate": 0.000317399119054206,
    "timestep_mean": 563.375,
    "epsilon_norm": 4.756715774536133,
    "prediction_norm": 3.2229340076446533,
    "timestamp": "2025-07-14T12:34:17.719382"
  },
  {
    "epoch": 1239,
    "batch_idx": 0,
    "loss": 0.19360029697418213,
    "learning_rate": 0.0003166681951794315,
    "timestep_mean": 453.0625,
    "epsilon_norm": 7.445789813995361,
    "prediction_norm": 6.7528862953186035,
    "timestamp": "2025-07-14T12:34:17.724231"
  },
  {
    "epoch": 1239,
    "batch_idx": 32,
    "loss": 0.29829007387161255,
    "learning_rate": 0.0003166681951794315,
    "timestep_mean": 481.40625,
    "epsilon_norm": 9.19101619720459,
    "prediction_norm": 7.482882976531982,
    "timestamp": "2025-07-14T12:34:17.729069"
  },
  {
    "epoch": 1239,
    "batch_idx": 64,
    "loss": 0.19799873232841492,
    "learning_rate": 0.0003166681951794315,
    "timestep_mean": 478.34375,
    "epsilon_norm": 8.382890701293945,
    "prediction_norm": 7.698713779449463,
    "timestamp": "2025-07-14T12:34:17.733951"
  },
  {
    "epoch": 1239,
    "batch_idx": 96,
    "loss": 0.2860807180404663,
    "learning_rate": 0.0003166681951794315,
    "timestep_mean": 555.4375,
    "epsilon_norm": 7.979081153869629,
    "prediction_norm": 7.1092963218688965,
    "timestamp": "2025-07-14T12:34:17.738837"
  },
  {
    "epoch": 1239,
    "batch_idx": 128,
    "loss": 0.21415621042251587,
    "learning_rate": 0.0003166681951794315,
    "timestep_mean": 577.125,
    "epsilon_norm": 8.302144050598145,
    "prediction_norm": 6.944469928741455,
    "timestamp": "2025-07-14T12:34:17.743699"
  },
  {
    "epoch": 1239,
    "batch_idx": 160,
    "loss": 0.2574898600578308,
    "learning_rate": 0.0003166681951794315,
    "timestep_mean": 551.6875,
    "epsilon_norm": 8.641162872314453,
    "prediction_norm": 7.615512371063232,
    "timestamp": "2025-07-14T12:34:17.748589"
  },
  {
    "epoch": 1239,
    "batch_idx": 192,
    "loss": 0.8024648427963257,
    "learning_rate": 0.0003166681951794315,
    "timestep_mean": 418.625,
    "epsilon_norm": 4.546426296234131,
    "prediction_norm": 3.37089467048645,
    "timestamp": "2025-07-14T12:34:17.753413"
  },
  {
    "epoch": 1240,
    "batch_idx": 0,
    "loss": 0.13507217168807983,
    "learning_rate": 0.0003159377236576609,
    "timestep_mean": 432.875,
    "epsilon_norm": 6.838680744171143,
    "prediction_norm": 6.3821868896484375,
    "timestamp": "2025-07-14T12:34:17.758392"
  },
  {
    "epoch": 1240,
    "batch_idx": 32,
    "loss": 0.32075005769729614,
    "learning_rate": 0.0003159377236576609,
    "timestep_mean": 457.90625,
    "epsilon_norm": 8.44849967956543,
    "prediction_norm": 6.253507614135742,
    "timestamp": "2025-07-14T12:34:17.763243"
  },
  {
    "epoch": 1240,
    "batch_idx": 64,
    "loss": 0.37391334772109985,
    "learning_rate": 0.0003159377236576609,
    "timestep_mean": 507.65625,
    "epsilon_norm": 8.003583908081055,
    "prediction_norm": 6.528297424316406,
    "timestamp": "2025-07-14T12:34:17.768058"
  },
  {
    "epoch": 1240,
    "batch_idx": 96,
    "loss": 0.2849777638912201,
    "learning_rate": 0.0003159377236576609,
    "timestep_mean": 540.75,
    "epsilon_norm": 8.437237739562988,
    "prediction_norm": 7.350388526916504,
    "timestamp": "2025-07-14T12:34:17.772993"
  },
  {
    "epoch": 1240,
    "batch_idx": 128,
    "loss": 0.3596903681755066,
    "learning_rate": 0.0003159377236576609,
    "timestep_mean": 413.75,
    "epsilon_norm": 7.75851583480835,
    "prediction_norm": 6.016142845153809,
    "timestamp": "2025-07-14T12:34:17.777805"
  },
  {
    "epoch": 1240,
    "batch_idx": 160,
    "loss": 0.1544368416070938,
    "learning_rate": 0.0003159377236576609,
    "timestep_mean": 470.96875,
    "epsilon_norm": 7.489729404449463,
    "prediction_norm": 6.259459018707275,
    "timestamp": "2025-07-14T12:34:17.782680"
  },
  {
    "epoch": 1240,
    "batch_idx": 192,
    "loss": 0.3304593861103058,
    "learning_rate": 0.0003159377236576609,
    "timestep_mean": 429.875,
    "epsilon_norm": 3.825131893157959,
    "prediction_norm": 2.7158477306365967,
    "timestamp": "2025-07-14T12:34:17.787568"
  },
  {
    "epoch": 1241,
    "batch_idx": 0,
    "loss": 0.38422858715057373,
    "learning_rate": 0.00031520770629126014,
    "timestep_mean": 481.0,
    "epsilon_norm": 7.786405563354492,
    "prediction_norm": 5.992982387542725,
    "timestamp": "2025-07-14T12:34:17.792458"
  },
  {
    "epoch": 1241,
    "batch_idx": 32,
    "loss": 0.24802182614803314,
    "learning_rate": 0.00031520770629126014,
    "timestep_mean": 480.6875,
    "epsilon_norm": 7.7765116691589355,
    "prediction_norm": 6.350379467010498,
    "timestamp": "2025-07-14T12:34:17.797357"
  },
  {
    "epoch": 1241,
    "batch_idx": 64,
    "loss": 0.3223642408847809,
    "learning_rate": 0.00031520770629126014,
    "timestep_mean": 531.1875,
    "epsilon_norm": 7.509573936462402,
    "prediction_norm": 6.029363632202148,
    "timestamp": "2025-07-14T12:34:17.802224"
  },
  {
    "epoch": 1241,
    "batch_idx": 96,
    "loss": 0.21847517788410187,
    "learning_rate": 0.00031520770629126014,
    "timestep_mean": 420.875,
    "epsilon_norm": 6.656695365905762,
    "prediction_norm": 5.809453964233398,
    "timestamp": "2025-07-14T12:34:17.807152"
  },
  {
    "epoch": 1241,
    "batch_idx": 128,
    "loss": 0.34949228167533875,
    "learning_rate": 0.00031520770629126014,
    "timestep_mean": 460.40625,
    "epsilon_norm": 8.75559139251709,
    "prediction_norm": 7.169830322265625,
    "timestamp": "2025-07-14T12:34:17.811982"
  },
  {
    "epoch": 1241,
    "batch_idx": 160,
    "loss": 0.14954908192157745,
    "learning_rate": 0.00031520770629126014,
    "timestep_mean": 571.75,
    "epsilon_norm": 8.000091552734375,
    "prediction_norm": 7.223500728607178,
    "timestamp": "2025-07-14T12:34:17.816917"
  },
  {
    "epoch": 1241,
    "batch_idx": 192,
    "loss": 0.506405770778656,
    "learning_rate": 0.00031520770629126014,
    "timestep_mean": 486.625,
    "epsilon_norm": 3.6799850463867188,
    "prediction_norm": 2.7838923931121826,
    "timestamp": "2025-07-14T12:34:17.821717"
  },
  {
    "epoch": 1242,
    "batch_idx": 0,
    "loss": 0.3014851212501526,
    "learning_rate": 0.0003144781448814743,
    "timestep_mean": 441.21875,
    "epsilon_norm": 7.075207710266113,
    "prediction_norm": 5.899442195892334,
    "timestamp": "2025-07-14T12:34:17.826657"
  },
  {
    "epoch": 1242,
    "batch_idx": 32,
    "loss": 0.24997738003730774,
    "learning_rate": 0.0003144781448814743,
    "timestep_mean": 518.15625,
    "epsilon_norm": 8.168191909790039,
    "prediction_norm": 7.258775234222412,
    "timestamp": "2025-07-14T12:34:17.831481"
  },
  {
    "epoch": 1242,
    "batch_idx": 64,
    "loss": 0.1866125762462616,
    "learning_rate": 0.0003144781448814743,
    "timestep_mean": 444.25,
    "epsilon_norm": 6.497001647949219,
    "prediction_norm": 5.788259506225586,
    "timestamp": "2025-07-14T12:34:17.836307"
  },
  {
    "epoch": 1242,
    "batch_idx": 96,
    "loss": 0.2821913957595825,
    "learning_rate": 0.0003144781448814743,
    "timestep_mean": 516.28125,
    "epsilon_norm": 6.990662574768066,
    "prediction_norm": 5.761516094207764,
    "timestamp": "2025-07-14T12:34:17.841197"
  },
  {
    "epoch": 1242,
    "batch_idx": 128,
    "loss": 0.16854393482208252,
    "learning_rate": 0.0003144781448814743,
    "timestep_mean": 553.96875,
    "epsilon_norm": 6.283061981201172,
    "prediction_norm": 5.735340118408203,
    "timestamp": "2025-07-14T12:34:17.846138"
  },
  {
    "epoch": 1242,
    "batch_idx": 160,
    "loss": 0.26120465993881226,
    "learning_rate": 0.0003144781448814743,
    "timestep_mean": 475.625,
    "epsilon_norm": 8.574484825134277,
    "prediction_norm": 7.6094207763671875,
    "timestamp": "2025-07-14T12:34:17.850993"
  },
  {
    "epoch": 1242,
    "batch_idx": 192,
    "loss": 0.293606698513031,
    "learning_rate": 0.0003144781448814743,
    "timestep_mean": 539.0,
    "epsilon_norm": 3.5993001461029053,
    "prediction_norm": 3.238133192062378,
    "timestamp": "2025-07-14T12:34:17.855955"
  },
  {
    "epoch": 1243,
    "batch_idx": 0,
    "loss": 0.306614488363266,
    "learning_rate": 0.000313749041228424,
    "timestep_mean": 502.03125,
    "epsilon_norm": 7.680032730102539,
    "prediction_norm": 6.699897289276123,
    "timestamp": "2025-07-14T12:34:17.860832"
  },
  {
    "epoch": 1243,
    "batch_idx": 32,
    "loss": 0.21634799242019653,
    "learning_rate": 0.000313749041228424,
    "timestep_mean": 557.28125,
    "epsilon_norm": 7.731928825378418,
    "prediction_norm": 6.308610916137695,
    "timestamp": "2025-07-14T12:34:17.865683"
  },
  {
    "epoch": 1243,
    "batch_idx": 64,
    "loss": 0.23026807606220245,
    "learning_rate": 0.000313749041228424,
    "timestep_mean": 510.59375,
    "epsilon_norm": 7.069430351257324,
    "prediction_norm": 6.188552379608154,
    "timestamp": "2025-07-14T12:34:17.870532"
  },
  {
    "epoch": 1243,
    "batch_idx": 96,
    "loss": 0.17056122422218323,
    "learning_rate": 0.000313749041228424,
    "timestep_mean": 481.3125,
    "epsilon_norm": 7.405365467071533,
    "prediction_norm": 6.799188137054443,
    "timestamp": "2025-07-14T12:34:17.875791"
  },
  {
    "epoch": 1243,
    "batch_idx": 128,
    "loss": 0.2189459651708603,
    "learning_rate": 0.000313749041228424,
    "timestep_mean": 470.65625,
    "epsilon_norm": 6.970623016357422,
    "prediction_norm": 6.27694845199585,
    "timestamp": "2025-07-14T12:34:17.880849"
  },
  {
    "epoch": 1243,
    "batch_idx": 160,
    "loss": 0.22301310300827026,
    "learning_rate": 0.000313749041228424,
    "timestep_mean": 592.53125,
    "epsilon_norm": 6.846338272094727,
    "prediction_norm": 6.252871513366699,
    "timestamp": "2025-07-14T12:34:17.886061"
  },
  {
    "epoch": 1243,
    "batch_idx": 192,
    "loss": 0.20613087713718414,
    "learning_rate": 0.000313749041228424,
    "timestep_mean": 606.125,
    "epsilon_norm": 4.27694845199585,
    "prediction_norm": 3.7962493896484375,
    "timestamp": "2025-07-14T12:34:17.891166"
  },
  {
    "epoch": 1244,
    "batch_idx": 0,
    "loss": 0.2907080054283142,
    "learning_rate": 0.00031302039713109977,
    "timestep_mean": 572.8125,
    "epsilon_norm": 6.937745094299316,
    "prediction_norm": 5.347888469696045,
    "timestamp": "2025-07-14T12:34:17.896288"
  },
  {
    "epoch": 1244,
    "batch_idx": 32,
    "loss": 0.22500291466712952,
    "learning_rate": 0.00031302039713109977,
    "timestep_mean": 459.0625,
    "epsilon_norm": 9.154960632324219,
    "prediction_norm": 7.8460917472839355,
    "timestamp": "2025-07-14T12:34:17.901508"
  },
  {
    "epoch": 1244,
    "batch_idx": 64,
    "loss": 0.21094995737075806,
    "learning_rate": 0.00031302039713109977,
    "timestep_mean": 491.40625,
    "epsilon_norm": 7.73069953918457,
    "prediction_norm": 6.987210273742676,
    "timestamp": "2025-07-14T12:34:17.906632"
  },
  {
    "epoch": 1244,
    "batch_idx": 96,
    "loss": 0.3591628074645996,
    "learning_rate": 0.00031302039713109977,
    "timestep_mean": 523.3125,
    "epsilon_norm": 7.154292106628418,
    "prediction_norm": 6.43143892288208,
    "timestamp": "2025-07-14T12:34:17.911847"
  },
  {
    "epoch": 1244,
    "batch_idx": 128,
    "loss": 0.30822598934173584,
    "learning_rate": 0.00031302039713109977,
    "timestep_mean": 450.625,
    "epsilon_norm": 7.212244987487793,
    "prediction_norm": 5.466008186340332,
    "timestamp": "2025-07-14T12:34:17.916931"
  },
  {
    "epoch": 1244,
    "batch_idx": 160,
    "loss": 0.3475562632083893,
    "learning_rate": 0.00031302039713109977,
    "timestep_mean": 528.9375,
    "epsilon_norm": 7.47600793838501,
    "prediction_norm": 6.230053424835205,
    "timestamp": "2025-07-14T12:34:17.921824"
  },
  {
    "epoch": 1244,
    "batch_idx": 192,
    "loss": 0.23298384249210358,
    "learning_rate": 0.00031302039713109977,
    "timestep_mean": 443.125,
    "epsilon_norm": 4.691258430480957,
    "prediction_norm": 3.467104434967041,
    "timestamp": "2025-07-14T12:34:17.927172"
  },
  {
    "epoch": 1245,
    "batch_idx": 0,
    "loss": 0.47579219937324524,
    "learning_rate": 0.00031229221438735844,
    "timestep_mean": 435.375,
    "epsilon_norm": 8.497091293334961,
    "prediction_norm": 5.9489426612854,
    "timestamp": "2025-07-14T12:34:17.935391"
  },
  {
    "epoch": 1245,
    "batch_idx": 32,
    "loss": 0.31017300486564636,
    "learning_rate": 0.00031229221438735844,
    "timestep_mean": 497.84375,
    "epsilon_norm": 8.040070533752441,
    "prediction_norm": 6.494877815246582,
    "timestamp": "2025-07-14T12:34:17.940617"
  },
  {
    "epoch": 1245,
    "batch_idx": 64,
    "loss": 0.2699342966079712,
    "learning_rate": 0.00031229221438735844,
    "timestep_mean": 529.15625,
    "epsilon_norm": 8.531895637512207,
    "prediction_norm": 7.486936092376709,
    "timestamp": "2025-07-14T12:34:17.945677"
  },
  {
    "epoch": 1245,
    "batch_idx": 96,
    "loss": 0.4266001582145691,
    "learning_rate": 0.00031229221438735844,
    "timestep_mean": 438.125,
    "epsilon_norm": 7.559070587158203,
    "prediction_norm": 5.543013095855713,
    "timestamp": "2025-07-14T12:34:17.950827"
  },
  {
    "epoch": 1245,
    "batch_idx": 128,
    "loss": 0.4502754807472229,
    "learning_rate": 0.00031229221438735844,
    "timestep_mean": 473.375,
    "epsilon_norm": 8.300188064575195,
    "prediction_norm": 6.75413703918457,
    "timestamp": "2025-07-14T12:34:17.955852"
  },
  {
    "epoch": 1245,
    "batch_idx": 160,
    "loss": 0.29768484830856323,
    "learning_rate": 0.00031229221438735844,
    "timestep_mean": 573.71875,
    "epsilon_norm": 8.663378715515137,
    "prediction_norm": 7.541886329650879,
    "timestamp": "2025-07-14T12:34:17.960803"
  },
  {
    "epoch": 1245,
    "batch_idx": 192,
    "loss": 0.3853176236152649,
    "learning_rate": 0.00031229221438735844,
    "timestep_mean": 229.875,
    "epsilon_norm": 3.177252769470215,
    "prediction_norm": 1.305385708808899,
    "timestamp": "2025-07-14T12:34:17.965835"
  },
  {
    "epoch": 1246,
    "batch_idx": 0,
    "loss": 0.28175467252731323,
    "learning_rate": 0.0003115644947939185,
    "timestep_mean": 537.21875,
    "epsilon_norm": 8.17822265625,
    "prediction_norm": 6.735193252563477,
    "timestamp": "2025-07-14T12:34:17.970757"
  },
  {
    "epoch": 1246,
    "batch_idx": 32,
    "loss": 0.2833396792411804,
    "learning_rate": 0.0003115644947939185,
    "timestep_mean": 403.34375,
    "epsilon_norm": 7.636572360992432,
    "prediction_norm": 6.098073959350586,
    "timestamp": "2025-07-14T12:34:17.975779"
  },
  {
    "epoch": 1246,
    "batch_idx": 64,
    "loss": 0.39394912123680115,
    "learning_rate": 0.0003115644947939185,
    "timestep_mean": 404.59375,
    "epsilon_norm": 8.038508415222168,
    "prediction_norm": 5.805972099304199,
    "timestamp": "2025-07-14T12:34:17.980787"
  },
  {
    "epoch": 1246,
    "batch_idx": 96,
    "loss": 0.3505879342556,
    "learning_rate": 0.0003115644947939185,
    "timestep_mean": 510.15625,
    "epsilon_norm": 7.538712978363037,
    "prediction_norm": 5.852587699890137,
    "timestamp": "2025-07-14T12:34:17.985752"
  },
  {
    "epoch": 1246,
    "batch_idx": 128,
    "loss": 0.26799994707107544,
    "learning_rate": 0.0003115644947939185,
    "timestep_mean": 477.34375,
    "epsilon_norm": 7.375877380371094,
    "prediction_norm": 6.01702880859375,
    "timestamp": "2025-07-14T12:34:17.990767"
  },
  {
    "epoch": 1246,
    "batch_idx": 160,
    "loss": 0.1821919083595276,
    "learning_rate": 0.0003115644947939185,
    "timestep_mean": 505.09375,
    "epsilon_norm": 9.178301811218262,
    "prediction_norm": 8.213220596313477,
    "timestamp": "2025-07-14T12:34:17.995736"
  },
  {
    "epoch": 1246,
    "batch_idx": 192,
    "loss": 0.41486304998397827,
    "learning_rate": 0.0003115644947939185,
    "timestep_mean": 416.75,
    "epsilon_norm": 4.7161126136779785,
    "prediction_norm": 3.60370135307312,
    "timestamp": "2025-07-14T12:34:18.000855"
  },
  {
    "epoch": 1247,
    "batch_idx": 0,
    "loss": 0.2487487941980362,
    "learning_rate": 0.0003108372401463561,
    "timestep_mean": 507.5,
    "epsilon_norm": 8.647704124450684,
    "prediction_norm": 7.993026256561279,
    "timestamp": "2025-07-14T12:34:18.005637"
  },
  {
    "epoch": 1247,
    "batch_idx": 32,
    "loss": 0.12984970211982727,
    "learning_rate": 0.0003108372401463561,
    "timestep_mean": 561.15625,
    "epsilon_norm": 6.367963790893555,
    "prediction_norm": 5.8384690284729,
    "timestamp": "2025-07-14T12:34:18.010626"
  },
  {
    "epoch": 1247,
    "batch_idx": 64,
    "loss": 0.6198256611824036,
    "learning_rate": 0.0003108372401463561,
    "timestep_mean": 399.90625,
    "epsilon_norm": 8.987736701965332,
    "prediction_norm": 6.799208164215088,
    "timestamp": "2025-07-14T12:34:18.015548"
  },
  {
    "epoch": 1247,
    "batch_idx": 96,
    "loss": 0.30990296602249146,
    "learning_rate": 0.0003108372401463561,
    "timestep_mean": 496.6875,
    "epsilon_norm": 7.58168888092041,
    "prediction_norm": 5.813287258148193,
    "timestamp": "2025-07-14T12:34:18.020625"
  },
  {
    "epoch": 1247,
    "batch_idx": 128,
    "loss": 0.3981930613517761,
    "learning_rate": 0.0003108372401463561,
    "timestep_mean": 523.96875,
    "epsilon_norm": 9.439349174499512,
    "prediction_norm": 7.612544536590576,
    "timestamp": "2025-07-14T12:34:18.025711"
  },
  {
    "epoch": 1247,
    "batch_idx": 160,
    "loss": 0.3149719834327698,
    "learning_rate": 0.0003108372401463561,
    "timestep_mean": 445.96875,
    "epsilon_norm": 8.001924514770508,
    "prediction_norm": 6.563431262969971,
    "timestamp": "2025-07-14T12:34:18.030736"
  },
  {
    "epoch": 1247,
    "batch_idx": 192,
    "loss": 0.3789352476596832,
    "learning_rate": 0.0003108372401463561,
    "timestep_mean": 572.0,
    "epsilon_norm": 4.091699600219727,
    "prediction_norm": 3.4588818550109863,
    "timestamp": "2025-07-14T12:34:18.035703"
  },
  {
    "epoch": 1248,
    "batch_idx": 0,
    "loss": 0.2877880334854126,
    "learning_rate": 0.00031011045223909946,
    "timestep_mean": 470.5625,
    "epsilon_norm": 7.93064022064209,
    "prediction_norm": 7.021667957305908,
    "timestamp": "2025-07-14T12:34:18.040662"
  },
  {
    "epoch": 1248,
    "batch_idx": 32,
    "loss": 0.18170003592967987,
    "learning_rate": 0.00031011045223909946,
    "timestep_mean": 528.5625,
    "epsilon_norm": 7.128489971160889,
    "prediction_norm": 6.246222496032715,
    "timestamp": "2025-07-14T12:34:18.045601"
  },
  {
    "epoch": 1248,
    "batch_idx": 64,
    "loss": 0.4087413549423218,
    "learning_rate": 0.00031011045223909946,
    "timestep_mean": 511.4375,
    "epsilon_norm": 8.654328346252441,
    "prediction_norm": 7.202238082885742,
    "timestamp": "2025-07-14T12:34:18.050758"
  },
  {
    "epoch": 1248,
    "batch_idx": 96,
    "loss": 0.15674367547035217,
    "learning_rate": 0.00031011045223909946,
    "timestep_mean": 561.3125,
    "epsilon_norm": 7.352303504943848,
    "prediction_norm": 6.674729824066162,
    "timestamp": "2025-07-14T12:34:18.055728"
  },
  {
    "epoch": 1248,
    "batch_idx": 128,
    "loss": 0.19809406995773315,
    "learning_rate": 0.00031011045223909946,
    "timestep_mean": 471.53125,
    "epsilon_norm": 7.025467395782471,
    "prediction_norm": 5.726291656494141,
    "timestamp": "2025-07-14T12:34:18.060796"
  },
  {
    "epoch": 1248,
    "batch_idx": 160,
    "loss": 0.17124590277671814,
    "learning_rate": 0.00031011045223909946,
    "timestep_mean": 534.59375,
    "epsilon_norm": 6.9952473640441895,
    "prediction_norm": 6.352377891540527,
    "timestamp": "2025-07-14T12:34:18.065780"
  },
  {
    "epoch": 1248,
    "batch_idx": 192,
    "loss": 0.1347401738166809,
    "learning_rate": 0.00031011045223909946,
    "timestep_mean": 637.125,
    "epsilon_norm": 4.509451389312744,
    "prediction_norm": 3.84091854095459,
    "timestamp": "2025-07-14T12:34:18.070853"
  },
  {
    "epoch": 1249,
    "batch_idx": 0,
    "loss": 0.19690188765525818,
    "learning_rate": 0.0003093841328654254,
    "timestep_mean": 491.375,
    "epsilon_norm": 8.039959907531738,
    "prediction_norm": 7.910251617431641,
    "timestamp": "2025-07-14T12:34:18.076059"
  },
  {
    "epoch": 1249,
    "batch_idx": 32,
    "loss": 0.44442272186279297,
    "learning_rate": 0.0003093841328654254,
    "timestep_mean": 535.4375,
    "epsilon_norm": 9.11345386505127,
    "prediction_norm": 7.278591632843018,
    "timestamp": "2025-07-14T12:34:18.081216"
  },
  {
    "epoch": 1249,
    "batch_idx": 64,
    "loss": 0.20606586337089539,
    "learning_rate": 0.0003093841328654254,
    "timestep_mean": 502.34375,
    "epsilon_norm": 8.064266204833984,
    "prediction_norm": 7.563905715942383,
    "timestamp": "2025-07-14T12:34:18.086408"
  },
  {
    "epoch": 1249,
    "batch_idx": 96,
    "loss": 0.33446526527404785,
    "learning_rate": 0.0003093841328654254,
    "timestep_mean": 482.65625,
    "epsilon_norm": 8.73592758178711,
    "prediction_norm": 7.298148155212402,
    "timestamp": "2025-07-14T12:34:18.091742"
  },
  {
    "epoch": 1249,
    "batch_idx": 128,
    "loss": 0.31451430916786194,
    "learning_rate": 0.0003093841328654254,
    "timestep_mean": 445.875,
    "epsilon_norm": 8.873701095581055,
    "prediction_norm": 6.800873279571533,
    "timestamp": "2025-07-14T12:34:18.096881"
  },
  {
    "epoch": 1249,
    "batch_idx": 160,
    "loss": 0.2686225175857544,
    "learning_rate": 0.0003093841328654254,
    "timestep_mean": 578.1875,
    "epsilon_norm": 7.3710808753967285,
    "prediction_norm": 6.381011962890625,
    "timestamp": "2025-07-14T12:34:18.102050"
  },
  {
    "epoch": 1249,
    "batch_idx": 192,
    "loss": 0.03259772062301636,
    "learning_rate": 0.0003093841328654254,
    "timestep_mean": 707.0,
    "epsilon_norm": 3.5410349369049072,
    "prediction_norm": 3.40791654586792,
    "timestamp": "2025-07-14T12:34:18.107252"
  },
  {
    "epoch": 1250,
    "batch_idx": 0,
    "loss": 0.3478740453720093,
    "learning_rate": 0.00030865828381745506,
    "timestep_mean": 443.5,
    "epsilon_norm": 7.114410400390625,
    "prediction_norm": 6.456310272216797,
    "timestamp": "2025-07-14T12:34:18.112328"
  },
  {
    "epoch": 1250,
    "batch_idx": 32,
    "loss": 0.2342161387205124,
    "learning_rate": 0.00030865828381745506,
    "timestep_mean": 476.59375,
    "epsilon_norm": 7.398506164550781,
    "prediction_norm": 6.4988274574279785,
    "timestamp": "2025-07-14T12:34:18.117584"
  },
  {
    "epoch": 1250,
    "batch_idx": 64,
    "loss": 0.25722622871398926,
    "learning_rate": 0.00030865828381745506,
    "timestep_mean": 499.09375,
    "epsilon_norm": 7.173085689544678,
    "prediction_norm": 6.198173999786377,
    "timestamp": "2025-07-14T12:34:18.122730"
  },
  {
    "epoch": 1250,
    "batch_idx": 96,
    "loss": 0.2417721450328827,
    "learning_rate": 0.00030865828381745506,
    "timestep_mean": 528.90625,
    "epsilon_norm": 8.170313835144043,
    "prediction_norm": 6.870327949523926,
    "timestamp": "2025-07-14T12:34:18.127768"
  },
  {
    "epoch": 1250,
    "batch_idx": 128,
    "loss": 0.15006454288959503,
    "learning_rate": 0.00030865828381745506,
    "timestep_mean": 445.40625,
    "epsilon_norm": 8.883627891540527,
    "prediction_norm": 8.61655330657959,
    "timestamp": "2025-07-14T12:34:18.132886"
  },
  {
    "epoch": 1250,
    "batch_idx": 160,
    "loss": 0.25156745314598083,
    "learning_rate": 0.00030865828381745506,
    "timestep_mean": 514.78125,
    "epsilon_norm": 7.671146392822266,
    "prediction_norm": 7.1556477546691895,
    "timestamp": "2025-07-14T12:34:18.138155"
  },
  {
    "epoch": 1250,
    "batch_idx": 192,
    "loss": 0.36202818155288696,
    "learning_rate": 0.00030865828381745506,
    "timestep_mean": 542.0,
    "epsilon_norm": 3.360370397567749,
    "prediction_norm": 2.021108627319336,
    "timestamp": "2025-07-14T12:34:18.143288"
  },
  {
    "epoch": 1251,
    "batch_idx": 0,
    "loss": 0.14355029165744781,
    "learning_rate": 0.0003079329068861487,
    "timestep_mean": 528.0625,
    "epsilon_norm": 7.014524459838867,
    "prediction_norm": 6.662047863006592,
    "timestamp": "2025-07-14T12:34:18.148421"
  },
  {
    "epoch": 1251,
    "batch_idx": 32,
    "loss": 0.44844645261764526,
    "learning_rate": 0.0003079329068861487,
    "timestep_mean": 562.28125,
    "epsilon_norm": 7.781403064727783,
    "prediction_norm": 5.340043067932129,
    "timestamp": "2025-07-14T12:34:18.153563"
  },
  {
    "epoch": 1251,
    "batch_idx": 64,
    "loss": 0.1654677540063858,
    "learning_rate": 0.0003079329068861487,
    "timestep_mean": 425.03125,
    "epsilon_norm": 7.944005489349365,
    "prediction_norm": 7.251418113708496,
    "timestamp": "2025-07-14T12:34:18.158856"
  },
  {
    "epoch": 1251,
    "batch_idx": 96,
    "loss": 0.3333407938480377,
    "learning_rate": 0.0003079329068861487,
    "timestep_mean": 391.5,
    "epsilon_norm": 8.107399940490723,
    "prediction_norm": 6.3438873291015625,
    "timestamp": "2025-07-14T12:34:18.164013"
  },
  {
    "epoch": 1251,
    "batch_idx": 128,
    "loss": 0.319938063621521,
    "learning_rate": 0.0003079329068861487,
    "timestep_mean": 502.75,
    "epsilon_norm": 8.078733444213867,
    "prediction_norm": 6.760210037231445,
    "timestamp": "2025-07-14T12:34:18.169107"
  },
  {
    "epoch": 1251,
    "batch_idx": 160,
    "loss": 0.3492799401283264,
    "learning_rate": 0.0003079329068861487,
    "timestep_mean": 496.5625,
    "epsilon_norm": 8.996415138244629,
    "prediction_norm": 7.084474563598633,
    "timestamp": "2025-07-14T12:34:18.174357"
  },
  {
    "epoch": 1251,
    "batch_idx": 192,
    "loss": 0.5837909579277039,
    "learning_rate": 0.0003079329068861487,
    "timestep_mean": 549.625,
    "epsilon_norm": 3.8731932640075684,
    "prediction_norm": 2.7756102085113525,
    "timestamp": "2025-07-14T12:34:18.179610"
  },
  {
    "epoch": 1252,
    "batch_idx": 0,
    "loss": 0.18191058933734894,
    "learning_rate": 0.0003072080038613017,
    "timestep_mean": 489.5625,
    "epsilon_norm": 6.8909687995910645,
    "prediction_norm": 6.19073486328125,
    "timestamp": "2025-07-14T12:34:18.184801"
  },
  {
    "epoch": 1252,
    "batch_idx": 32,
    "loss": 0.3394016623497009,
    "learning_rate": 0.0003072080038613017,
    "timestep_mean": 573.125,
    "epsilon_norm": 7.538494110107422,
    "prediction_norm": 6.214970111846924,
    "timestamp": "2025-07-14T12:34:18.190014"
  },
  {
    "epoch": 1252,
    "batch_idx": 64,
    "loss": 0.3145183324813843,
    "learning_rate": 0.0003072080038613017,
    "timestep_mean": 467.375,
    "epsilon_norm": 8.185160636901855,
    "prediction_norm": 6.772232532501221,
    "timestamp": "2025-07-14T12:34:18.195129"
  },
  {
    "epoch": 1252,
    "batch_idx": 96,
    "loss": 0.17845214903354645,
    "learning_rate": 0.0003072080038613017,
    "timestep_mean": 524.90625,
    "epsilon_norm": 8.227995872497559,
    "prediction_norm": 7.413891315460205,
    "timestamp": "2025-07-14T12:34:18.200477"
  },
  {
    "epoch": 1252,
    "batch_idx": 128,
    "loss": 0.16039209067821503,
    "learning_rate": 0.0003072080038613017,
    "timestep_mean": 473.65625,
    "epsilon_norm": 7.33025598526001,
    "prediction_norm": 6.722696304321289,
    "timestamp": "2025-07-14T12:34:18.205518"
  },
  {
    "epoch": 1252,
    "batch_idx": 160,
    "loss": 0.3724357485771179,
    "learning_rate": 0.0003072080038613017,
    "timestep_mean": 474.34375,
    "epsilon_norm": 10.060131072998047,
    "prediction_norm": 9.011163711547852,
    "timestamp": "2025-07-14T12:34:18.210410"
  },
  {
    "epoch": 1252,
    "batch_idx": 192,
    "loss": 0.0934678465127945,
    "learning_rate": 0.0003072080038613017,
    "timestep_mean": 488.875,
    "epsilon_norm": 3.3799829483032227,
    "prediction_norm": 3.253448247909546,
    "timestamp": "2025-07-14T12:34:18.215307"
  },
  {
    "epoch": 1253,
    "batch_idx": 0,
    "loss": 0.20175495743751526,
    "learning_rate": 0.0003064835765315402,
    "timestep_mean": 550.4375,
    "epsilon_norm": 9.413362503051758,
    "prediction_norm": 8.9544677734375,
    "timestamp": "2025-07-14T12:34:18.220265"
  },
  {
    "epoch": 1253,
    "batch_idx": 32,
    "loss": 0.3529975712299347,
    "learning_rate": 0.0003064835765315402,
    "timestep_mean": 451.21875,
    "epsilon_norm": 8.257203102111816,
    "prediction_norm": 7.096402168273926,
    "timestamp": "2025-07-14T12:34:18.225261"
  },
  {
    "epoch": 1253,
    "batch_idx": 64,
    "loss": 0.284839928150177,
    "learning_rate": 0.0003064835765315402,
    "timestep_mean": 526.03125,
    "epsilon_norm": 8.15244197845459,
    "prediction_norm": 7.408936023712158,
    "timestamp": "2025-07-14T12:34:18.230236"
  },
  {
    "epoch": 1253,
    "batch_idx": 96,
    "loss": 0.23929128050804138,
    "learning_rate": 0.0003064835765315402,
    "timestep_mean": 484.9375,
    "epsilon_norm": 7.935974597930908,
    "prediction_norm": 6.506621360778809,
    "timestamp": "2025-07-14T12:34:18.235248"
  },
  {
    "epoch": 1253,
    "batch_idx": 128,
    "loss": 0.20403935015201569,
    "learning_rate": 0.0003064835765315402,
    "timestep_mean": 464.46875,
    "epsilon_norm": 7.384093284606934,
    "prediction_norm": 6.442927360534668,
    "timestamp": "2025-07-14T12:34:18.240379"
  },
  {
    "epoch": 1253,
    "batch_idx": 160,
    "loss": 0.2782862186431885,
    "learning_rate": 0.0003064835765315402,
    "timestep_mean": 563.28125,
    "epsilon_norm": 8.526677131652832,
    "prediction_norm": 7.213582992553711,
    "timestamp": "2025-07-14T12:34:18.245513"
  },
  {
    "epoch": 1253,
    "batch_idx": 192,
    "loss": 0.27830636501312256,
    "learning_rate": 0.0003064835765315402,
    "timestep_mean": 273.375,
    "epsilon_norm": 3.1552951335906982,
    "prediction_norm": 2.631016969680786,
    "timestamp": "2025-07-14T12:34:18.250510"
  },
  {
    "epoch": 1254,
    "batch_idx": 0,
    "loss": 0.20628602802753448,
    "learning_rate": 0.0003057596266843169,
    "timestep_mean": 518.03125,
    "epsilon_norm": 7.285139560699463,
    "prediction_norm": 5.919042110443115,
    "timestamp": "2025-07-14T12:34:18.255423"
  },
  {
    "epoch": 1254,
    "batch_idx": 32,
    "loss": 0.405414879322052,
    "learning_rate": 0.0003057596266843169,
    "timestep_mean": 454.4375,
    "epsilon_norm": 8.25178337097168,
    "prediction_norm": 6.225010871887207,
    "timestamp": "2025-07-14T12:34:18.260463"
  },
  {
    "epoch": 1254,
    "batch_idx": 64,
    "loss": 0.16711080074310303,
    "learning_rate": 0.0003057596266843169,
    "timestep_mean": 459.03125,
    "epsilon_norm": 7.160395622253418,
    "prediction_norm": 6.165362358093262,
    "timestamp": "2025-07-14T12:34:18.265637"
  },
  {
    "epoch": 1254,
    "batch_idx": 96,
    "loss": 0.18257677555084229,
    "learning_rate": 0.0003057596266843169,
    "timestep_mean": 483.71875,
    "epsilon_norm": 7.996988773345947,
    "prediction_norm": 6.8475446701049805,
    "timestamp": "2025-07-14T12:34:18.270791"
  },
  {
    "epoch": 1254,
    "batch_idx": 128,
    "loss": 0.5666205883026123,
    "learning_rate": 0.0003057596266843169,
    "timestep_mean": 569.0,
    "epsilon_norm": 8.36417293548584,
    "prediction_norm": 6.525221347808838,
    "timestamp": "2025-07-14T12:34:18.275868"
  },
  {
    "epoch": 1254,
    "batch_idx": 160,
    "loss": 0.4189133048057556,
    "learning_rate": 0.0003057596266843169,
    "timestep_mean": 498.03125,
    "epsilon_norm": 10.363677024841309,
    "prediction_norm": 8.666638374328613,
    "timestamp": "2025-07-14T12:34:18.281020"
  },
  {
    "epoch": 1254,
    "batch_idx": 192,
    "loss": 0.3288835883140564,
    "learning_rate": 0.0003057596266843169,
    "timestep_mean": 415.875,
    "epsilon_norm": 4.128335475921631,
    "prediction_norm": 3.180969715118408,
    "timestamp": "2025-07-14T12:34:18.286094"
  },
  {
    "epoch": 1255,
    "batch_idx": 0,
    "loss": 0.24018484354019165,
    "learning_rate": 0.0003050361561059059,
    "timestep_mean": 508.53125,
    "epsilon_norm": 8.108409881591797,
    "prediction_norm": 7.404749393463135,
    "timestamp": "2025-07-14T12:34:18.291206"
  },
  {
    "epoch": 1255,
    "batch_idx": 32,
    "loss": 0.2947703003883362,
    "learning_rate": 0.0003050361561059059,
    "timestep_mean": 481.375,
    "epsilon_norm": 8.813961029052734,
    "prediction_norm": 7.51411771774292,
    "timestamp": "2025-07-14T12:34:18.296419"
  },
  {
    "epoch": 1255,
    "batch_idx": 64,
    "loss": 0.4185526371002197,
    "learning_rate": 0.0003050361561059059,
    "timestep_mean": 443.5,
    "epsilon_norm": 7.688479900360107,
    "prediction_norm": 5.813345909118652,
    "timestamp": "2025-07-14T12:34:18.301524"
  },
  {
    "epoch": 1255,
    "batch_idx": 96,
    "loss": 0.34024137258529663,
    "learning_rate": 0.0003050361561059059,
    "timestep_mean": 436.53125,
    "epsilon_norm": 7.220992565155029,
    "prediction_norm": 5.784149646759033,
    "timestamp": "2025-07-14T12:34:18.306694"
  },
  {
    "epoch": 1255,
    "batch_idx": 128,
    "loss": 0.35269856452941895,
    "learning_rate": 0.0003050361561059059,
    "timestep_mean": 398.4375,
    "epsilon_norm": 7.787662029266357,
    "prediction_norm": 6.1135454177856445,
    "timestamp": "2025-07-14T12:34:18.311819"
  },
  {
    "epoch": 1255,
    "batch_idx": 160,
    "loss": 0.35957854986190796,
    "learning_rate": 0.0003050361561059059,
    "timestep_mean": 476.09375,
    "epsilon_norm": 8.731934547424316,
    "prediction_norm": 7.101536273956299,
    "timestamp": "2025-07-14T12:34:18.316913"
  },
  {
    "epoch": 1255,
    "batch_idx": 192,
    "loss": 0.08373822271823883,
    "learning_rate": 0.0003050361561059059,
    "timestep_mean": 666.25,
    "epsilon_norm": 2.62445330619812,
    "prediction_norm": 2.282608985900879,
    "timestamp": "2025-07-14T12:34:18.321893"
  },
  {
    "epoch": 1256,
    "batch_idx": 0,
    "loss": 0.3879595696926117,
    "learning_rate": 0.0003043131665813988,
    "timestep_mean": 481.6875,
    "epsilon_norm": 8.628189086914062,
    "prediction_norm": 7.008699893951416,
    "timestamp": "2025-07-14T12:34:18.327004"
  },
  {
    "epoch": 1256,
    "batch_idx": 32,
    "loss": 0.2549671530723572,
    "learning_rate": 0.0003043131665813988,
    "timestep_mean": 511.375,
    "epsilon_norm": 8.777511596679688,
    "prediction_norm": 7.371028900146484,
    "timestamp": "2025-07-14T12:34:18.332099"
  },
  {
    "epoch": 1256,
    "batch_idx": 64,
    "loss": 0.41652798652648926,
    "learning_rate": 0.0003043131665813988,
    "timestep_mean": 492.53125,
    "epsilon_norm": 7.69263219833374,
    "prediction_norm": 5.535777568817139,
    "timestamp": "2025-07-14T12:34:18.337271"
  },
  {
    "epoch": 1256,
    "batch_idx": 96,
    "loss": 0.2789461612701416,
    "learning_rate": 0.0003043131665813988,
    "timestep_mean": 557.0625,
    "epsilon_norm": 7.890662670135498,
    "prediction_norm": 7.140722274780273,
    "timestamp": "2025-07-14T12:34:18.342425"
  },
  {
    "epoch": 1256,
    "batch_idx": 128,
    "loss": 0.634619414806366,
    "learning_rate": 0.0003043131665813988,
    "timestep_mean": 456.84375,
    "epsilon_norm": 8.610688209533691,
    "prediction_norm": 6.360940933227539,
    "timestamp": "2025-07-14T12:34:18.347526"
  },
  {
    "epoch": 1256,
    "batch_idx": 160,
    "loss": 0.25404202938079834,
    "learning_rate": 0.0003043131665813988,
    "timestep_mean": 478.75,
    "epsilon_norm": 7.820746421813965,
    "prediction_norm": 6.931480884552002,
    "timestamp": "2025-07-14T12:34:18.352721"
  },
  {
    "epoch": 1256,
    "batch_idx": 192,
    "loss": 0.15316733717918396,
    "learning_rate": 0.0003043131665813988,
    "timestep_mean": 611.25,
    "epsilon_norm": 3.601789951324463,
    "prediction_norm": 2.9522790908813477,
    "timestamp": "2025-07-14T12:34:18.357886"
  },
  {
    "epoch": 1257,
    "batch_idx": 0,
    "loss": 0.16484391689300537,
    "learning_rate": 0.0003035906598947005,
    "timestep_mean": 585.0,
    "epsilon_norm": 7.668223857879639,
    "prediction_norm": 7.206003189086914,
    "timestamp": "2025-07-14T12:34:18.363264"
  },
  {
    "epoch": 1257,
    "batch_idx": 32,
    "loss": 0.19123181700706482,
    "learning_rate": 0.0003035906598947005,
    "timestep_mean": 577.4375,
    "epsilon_norm": 8.73077392578125,
    "prediction_norm": 7.826912879943848,
    "timestamp": "2025-07-14T12:34:18.368421"
  },
  {
    "epoch": 1257,
    "batch_idx": 64,
    "loss": 0.20900101959705353,
    "learning_rate": 0.0003035906598947005,
    "timestep_mean": 548.4375,
    "epsilon_norm": 7.575683116912842,
    "prediction_norm": 6.023997783660889,
    "timestamp": "2025-07-14T12:34:18.373628"
  },
  {
    "epoch": 1257,
    "batch_idx": 96,
    "loss": 0.21355512738227844,
    "learning_rate": 0.0003035906598947005,
    "timestep_mean": 500.15625,
    "epsilon_norm": 7.393413066864014,
    "prediction_norm": 6.843120574951172,
    "timestamp": "2025-07-14T12:34:18.378829"
  },
  {
    "epoch": 1257,
    "batch_idx": 128,
    "loss": 0.3219340443611145,
    "learning_rate": 0.0003035906598947005,
    "timestep_mean": 445.71875,
    "epsilon_norm": 7.474444389343262,
    "prediction_norm": 5.932698726654053,
    "timestamp": "2025-07-14T12:34:18.384085"
  },
  {
    "epoch": 1257,
    "batch_idx": 160,
    "loss": 0.2572648823261261,
    "learning_rate": 0.0003035906598947005,
    "timestep_mean": 556.125,
    "epsilon_norm": 8.798084259033203,
    "prediction_norm": 7.467185974121094,
    "timestamp": "2025-07-14T12:34:18.389160"
  },
  {
    "epoch": 1257,
    "batch_idx": 192,
    "loss": 0.7580119371414185,
    "learning_rate": 0.0003035906598947005,
    "timestep_mean": 295.375,
    "epsilon_norm": 4.193007469177246,
    "prediction_norm": 2.392155170440674,
    "timestamp": "2025-07-14T12:34:18.394285"
  },
  {
    "epoch": 1258,
    "batch_idx": 0,
    "loss": 0.2553981840610504,
    "learning_rate": 0.00030286863782852435,
    "timestep_mean": 537.09375,
    "epsilon_norm": 7.438764572143555,
    "prediction_norm": 5.857356548309326,
    "timestamp": "2025-07-14T12:34:18.399529"
  },
  {
    "epoch": 1258,
    "batch_idx": 32,
    "loss": 0.18390582501888275,
    "learning_rate": 0.00030286863782852435,
    "timestep_mean": 521.875,
    "epsilon_norm": 7.015946865081787,
    "prediction_norm": 6.224156856536865,
    "timestamp": "2025-07-14T12:34:18.404703"
  },
  {
    "epoch": 1258,
    "batch_idx": 64,
    "loss": 0.2344912886619568,
    "learning_rate": 0.00030286863782852435,
    "timestep_mean": 479.25,
    "epsilon_norm": 8.285308837890625,
    "prediction_norm": 7.997961521148682,
    "timestamp": "2025-07-14T12:34:18.409522"
  },
  {
    "epoch": 1258,
    "batch_idx": 96,
    "loss": 0.25827687978744507,
    "learning_rate": 0.00030286863782852435,
    "timestep_mean": 489.0,
    "epsilon_norm": 8.389617919921875,
    "prediction_norm": 7.1826090812683105,
    "timestamp": "2025-07-14T12:34:18.414443"
  },
  {
    "epoch": 1258,
    "batch_idx": 128,
    "loss": 0.2771419882774353,
    "learning_rate": 0.00030286863782852435,
    "timestep_mean": 454.3125,
    "epsilon_norm": 7.029078483581543,
    "prediction_norm": 6.465963840484619,
    "timestamp": "2025-07-14T12:34:18.419516"
  },
  {
    "epoch": 1258,
    "batch_idx": 160,
    "loss": 0.25322502851486206,
    "learning_rate": 0.00030286863782852435,
    "timestep_mean": 513.4375,
    "epsilon_norm": 7.3185811042785645,
    "prediction_norm": 6.56151819229126,
    "timestamp": "2025-07-14T12:34:18.424356"
  },
  {
    "epoch": 1258,
    "batch_idx": 192,
    "loss": 0.3068844676017761,
    "learning_rate": 0.00030286863782852435,
    "timestep_mean": 414.5,
    "epsilon_norm": 3.4932477474212646,
    "prediction_norm": 3.376331090927124,
    "timestamp": "2025-07-14T12:34:18.429315"
  },
  {
    "epoch": 1259,
    "batch_idx": 0,
    "loss": 0.34727272391319275,
    "learning_rate": 0.00030214710216438834,
    "timestep_mean": 531.71875,
    "epsilon_norm": 8.70047664642334,
    "prediction_norm": 7.274091720581055,
    "timestamp": "2025-07-14T12:34:18.434419"
  },
  {
    "epoch": 1259,
    "batch_idx": 32,
    "loss": 0.15919463336467743,
    "learning_rate": 0.00030214710216438834,
    "timestep_mean": 564.5625,
    "epsilon_norm": 8.122121810913086,
    "prediction_norm": 7.348393440246582,
    "timestamp": "2025-07-14T12:34:18.439499"
  },
  {
    "epoch": 1259,
    "batch_idx": 64,
    "loss": 0.4383562207221985,
    "learning_rate": 0.00030214710216438834,
    "timestep_mean": 460.25,
    "epsilon_norm": 8.477785110473633,
    "prediction_norm": 6.148044586181641,
    "timestamp": "2025-07-14T12:34:18.444601"
  },
  {
    "epoch": 1259,
    "batch_idx": 96,
    "loss": 0.14875710010528564,
    "learning_rate": 0.00030214710216438834,
    "timestep_mean": 541.53125,
    "epsilon_norm": 6.671962261199951,
    "prediction_norm": 6.162930965423584,
    "timestamp": "2025-07-14T12:34:18.449732"
  },
  {
    "epoch": 1259,
    "batch_idx": 128,
    "loss": 0.3275899887084961,
    "learning_rate": 0.00030214710216438834,
    "timestep_mean": 420.46875,
    "epsilon_norm": 8.72679615020752,
    "prediction_norm": 7.701027870178223,
    "timestamp": "2025-07-14T12:34:18.454960"
  },
  {
    "epoch": 1259,
    "batch_idx": 160,
    "loss": 0.1985265016555786,
    "learning_rate": 0.00030214710216438834,
    "timestep_mean": 488.4375,
    "epsilon_norm": 8.130667686462402,
    "prediction_norm": 7.216605186462402,
    "timestamp": "2025-07-14T12:34:18.460150"
  },
  {
    "epoch": 1259,
    "batch_idx": 192,
    "loss": 0.02516566962003708,
    "learning_rate": 0.00030214710216438834,
    "timestep_mean": 711.625,
    "epsilon_norm": 3.898376941680908,
    "prediction_norm": 4.078678131103516,
    "timestamp": "2025-07-14T12:34:18.465196"
  },
  {
    "epoch": 1260,
    "batch_idx": 0,
    "loss": 0.3666093945503235,
    "learning_rate": 0.00030142605468260955,
    "timestep_mean": 544.28125,
    "epsilon_norm": 8.426023483276367,
    "prediction_norm": 7.010125637054443,
    "timestamp": "2025-07-14T12:34:18.470280"
  },
  {
    "epoch": 1260,
    "batch_idx": 32,
    "loss": 0.16424983739852905,
    "learning_rate": 0.00030142605468260955,
    "timestep_mean": 473.40625,
    "epsilon_norm": 8.711917877197266,
    "prediction_norm": 8.261313438415527,
    "timestamp": "2025-07-14T12:34:18.475555"
  },
  {
    "epoch": 1260,
    "batch_idx": 64,
    "loss": 0.41411176323890686,
    "learning_rate": 0.00030142605468260955,
    "timestep_mean": 482.875,
    "epsilon_norm": 8.614131927490234,
    "prediction_norm": 7.033583641052246,
    "timestamp": "2025-07-14T12:34:18.480549"
  },
  {
    "epoch": 1260,
    "batch_idx": 96,
    "loss": 0.35650137066841125,
    "learning_rate": 0.00030142605468260955,
    "timestep_mean": 555.25,
    "epsilon_norm": 8.416739463806152,
    "prediction_norm": 7.4604058265686035,
    "timestamp": "2025-07-14T12:34:18.485587"
  },
  {
    "epoch": 1260,
    "batch_idx": 128,
    "loss": 0.27195316553115845,
    "learning_rate": 0.00030142605468260955,
    "timestep_mean": 424.59375,
    "epsilon_norm": 7.4716691970825195,
    "prediction_norm": 6.70793342590332,
    "timestamp": "2025-07-14T12:34:18.490597"
  },
  {
    "epoch": 1260,
    "batch_idx": 160,
    "loss": 0.2289198786020279,
    "learning_rate": 0.00030142605468260955,
    "timestep_mean": 505.125,
    "epsilon_norm": 8.41299819946289,
    "prediction_norm": 7.051879405975342,
    "timestamp": "2025-07-14T12:34:18.495707"
  },
  {
    "epoch": 1260,
    "batch_idx": 192,
    "loss": 0.08614220470190048,
    "learning_rate": 0.00030142605468260955,
    "timestep_mean": 363.75,
    "epsilon_norm": 3.161696195602417,
    "prediction_norm": 2.9564507007598877,
    "timestamp": "2025-07-14T12:34:18.500624"
  },
  {
    "epoch": 1261,
    "batch_idx": 0,
    "loss": 0.15676090121269226,
    "learning_rate": 0.00030070549716230136,
    "timestep_mean": 567.46875,
    "epsilon_norm": 9.017245292663574,
    "prediction_norm": 8.579991340637207,
    "timestamp": "2025-07-14T12:34:18.505487"
  },
  {
    "epoch": 1261,
    "batch_idx": 32,
    "loss": 0.26372000575065613,
    "learning_rate": 0.00030070549716230136,
    "timestep_mean": 541.59375,
    "epsilon_norm": 9.301301002502441,
    "prediction_norm": 8.118752479553223,
    "timestamp": "2025-07-14T12:34:18.510348"
  },
  {
    "epoch": 1261,
    "batch_idx": 64,
    "loss": 0.29691046476364136,
    "learning_rate": 0.00030070549716230136,
    "timestep_mean": 509.78125,
    "epsilon_norm": 8.876291275024414,
    "prediction_norm": 7.904155254364014,
    "timestamp": "2025-07-14T12:34:18.515224"
  },
  {
    "epoch": 1261,
    "batch_idx": 96,
    "loss": 0.18540018796920776,
    "learning_rate": 0.00030070549716230136,
    "timestep_mean": 511.09375,
    "epsilon_norm": 7.522892951965332,
    "prediction_norm": 6.620875835418701,
    "timestamp": "2025-07-14T12:34:18.520097"
  },
  {
    "epoch": 1261,
    "batch_idx": 128,
    "loss": 0.4219830632209778,
    "learning_rate": 0.00030070549716230136,
    "timestep_mean": 368.28125,
    "epsilon_norm": 7.334645748138428,
    "prediction_norm": 6.242992877960205,
    "timestamp": "2025-07-14T12:34:18.525039"
  },
  {
    "epoch": 1261,
    "batch_idx": 160,
    "loss": 0.2442367672920227,
    "learning_rate": 0.00030070549716230136,
    "timestep_mean": 492.40625,
    "epsilon_norm": 6.76661491394043,
    "prediction_norm": 5.207362174987793,
    "timestamp": "2025-07-14T12:34:18.529888"
  },
  {
    "epoch": 1261,
    "batch_idx": 192,
    "loss": 0.2702520489692688,
    "learning_rate": 0.00030070549716230136,
    "timestep_mean": 485.0,
    "epsilon_norm": 3.4995572566986084,
    "prediction_norm": 2.0541951656341553,
    "timestamp": "2025-07-14T12:34:18.534711"
  },
  {
    "epoch": 1262,
    "batch_idx": 0,
    "loss": 0.35888516902923584,
    "learning_rate": 0.0002999854313813676,
    "timestep_mean": 457.28125,
    "epsilon_norm": 8.284329414367676,
    "prediction_norm": 6.719235897064209,
    "timestamp": "2025-07-14T12:34:18.539545"
  },
  {
    "epoch": 1262,
    "batch_idx": 32,
    "loss": 0.19988420605659485,
    "learning_rate": 0.0002999854313813676,
    "timestep_mean": 530.53125,
    "epsilon_norm": 7.899425506591797,
    "prediction_norm": 6.774311065673828,
    "timestamp": "2025-07-14T12:34:18.544388"
  },
  {
    "epoch": 1262,
    "batch_idx": 64,
    "loss": 0.48509976267814636,
    "learning_rate": 0.0002999854313813676,
    "timestep_mean": 469.09375,
    "epsilon_norm": 8.84882926940918,
    "prediction_norm": 6.399479389190674,
    "timestamp": "2025-07-14T12:34:18.549197"
  },
  {
    "epoch": 1262,
    "batch_idx": 96,
    "loss": 0.17126943171024323,
    "learning_rate": 0.0002999854313813676,
    "timestep_mean": 576.78125,
    "epsilon_norm": 7.274096488952637,
    "prediction_norm": 6.762337684631348,
    "timestamp": "2025-07-14T12:34:18.554053"
  },
  {
    "epoch": 1262,
    "batch_idx": 128,
    "loss": 0.4009335935115814,
    "learning_rate": 0.0002999854313813676,
    "timestep_mean": 448.0625,
    "epsilon_norm": 7.191206455230713,
    "prediction_norm": 4.995389461517334,
    "timestamp": "2025-07-14T12:34:18.558943"
  },
  {
    "epoch": 1262,
    "batch_idx": 160,
    "loss": 0.30396944284439087,
    "learning_rate": 0.0002999854313813676,
    "timestep_mean": 491.03125,
    "epsilon_norm": 9.268962860107422,
    "prediction_norm": 7.324592590332031,
    "timestamp": "2025-07-14T12:34:18.563807"
  },
  {
    "epoch": 1262,
    "batch_idx": 192,
    "loss": 0.15743176639080048,
    "learning_rate": 0.0002999854313813676,
    "timestep_mean": 435.25,
    "epsilon_norm": 2.4207262992858887,
    "prediction_norm": 2.811682939529419,
    "timestamp": "2025-07-14T12:34:18.568645"
  },
  {
    "epoch": 1263,
    "batch_idx": 0,
    "loss": 0.25442689657211304,
    "learning_rate": 0.0002992658591164991,
    "timestep_mean": 534.1875,
    "epsilon_norm": 7.679069995880127,
    "prediction_norm": 6.57090950012207,
    "timestamp": "2025-07-14T12:34:18.573475"
  },
  {
    "epoch": 1263,
    "batch_idx": 32,
    "loss": 0.3168896436691284,
    "learning_rate": 0.0002992658591164991,
    "timestep_mean": 553.1875,
    "epsilon_norm": 7.728521347045898,
    "prediction_norm": 6.49120569229126,
    "timestamp": "2025-07-14T12:34:18.578313"
  },
  {
    "epoch": 1263,
    "batch_idx": 64,
    "loss": 0.24162960052490234,
    "learning_rate": 0.0002992658591164991,
    "timestep_mean": 516.21875,
    "epsilon_norm": 7.887925624847412,
    "prediction_norm": 6.72155237197876,
    "timestamp": "2025-07-14T12:34:18.583184"
  },
  {
    "epoch": 1263,
    "batch_idx": 96,
    "loss": 0.4985823631286621,
    "learning_rate": 0.0002992658591164991,
    "timestep_mean": 396.90625,
    "epsilon_norm": 9.160552978515625,
    "prediction_norm": 7.154801368713379,
    "timestamp": "2025-07-14T12:34:18.588021"
  },
  {
    "epoch": 1263,
    "batch_idx": 128,
    "loss": 0.19504314661026,
    "learning_rate": 0.0002992658591164991,
    "timestep_mean": 529.40625,
    "epsilon_norm": 7.632264137268066,
    "prediction_norm": 6.3716721534729,
    "timestamp": "2025-07-14T12:34:18.592840"
  },
  {
    "epoch": 1263,
    "batch_idx": 160,
    "loss": 0.37783342599868774,
    "learning_rate": 0.0002992658591164991,
    "timestep_mean": 514.5625,
    "epsilon_norm": 8.69166374206543,
    "prediction_norm": 7.04713773727417,
    "timestamp": "2025-07-14T12:34:18.597644"
  },
  {
    "epoch": 1263,
    "batch_idx": 192,
    "loss": 0.32937973737716675,
    "learning_rate": 0.0002992658591164991,
    "timestep_mean": 475.625,
    "epsilon_norm": 4.438014030456543,
    "prediction_norm": 3.945228099822998,
    "timestamp": "2025-07-14T12:34:18.605757"
  },
  {
    "epoch": 1264,
    "batch_idx": 0,
    "loss": 0.1433517336845398,
    "learning_rate": 0.00029854678214316855,
    "timestep_mean": 583.25,
    "epsilon_norm": 9.240903854370117,
    "prediction_norm": 8.526691436767578,
    "timestamp": "2025-07-14T12:34:18.612392"
  },
  {
    "epoch": 1264,
    "batch_idx": 32,
    "loss": 0.18495631217956543,
    "learning_rate": 0.00029854678214316855,
    "timestep_mean": 587.9375,
    "epsilon_norm": 7.026402473449707,
    "prediction_norm": 6.465798377990723,
    "timestamp": "2025-07-14T12:34:18.618502"
  },
  {
    "epoch": 1264,
    "batch_idx": 64,
    "loss": 0.29201042652130127,
    "learning_rate": 0.00029854678214316855,
    "timestep_mean": 467.65625,
    "epsilon_norm": 7.463932037353516,
    "prediction_norm": 5.785263538360596,
    "timestamp": "2025-07-14T12:34:18.624830"
  },
  {
    "epoch": 1264,
    "batch_idx": 96,
    "loss": 0.1963358223438263,
    "learning_rate": 0.00029854678214316855,
    "timestep_mean": 487.875,
    "epsilon_norm": 8.625746726989746,
    "prediction_norm": 7.826070785522461,
    "timestamp": "2025-07-14T12:34:18.633343"
  },
  {
    "epoch": 1264,
    "batch_idx": 128,
    "loss": 0.3706199526786804,
    "learning_rate": 0.00029854678214316855,
    "timestep_mean": 501.125,
    "epsilon_norm": 8.100234031677246,
    "prediction_norm": 6.309030055999756,
    "timestamp": "2025-07-14T12:34:18.639828"
  },
  {
    "epoch": 1264,
    "batch_idx": 160,
    "loss": 0.6808339357376099,
    "learning_rate": 0.00029854678214316855,
    "timestep_mean": 392.09375,
    "epsilon_norm": 7.472353458404541,
    "prediction_norm": 4.197124481201172,
    "timestamp": "2025-07-14T12:34:18.646053"
  },
  {
    "epoch": 1264,
    "batch_idx": 192,
    "loss": 0.6764290928840637,
    "learning_rate": 0.00029854678214316855,
    "timestep_mean": 389.625,
    "epsilon_norm": 4.878741264343262,
    "prediction_norm": 3.2259302139282227,
    "timestamp": "2025-07-14T12:34:18.651585"
  },
  {
    "epoch": 1265,
    "batch_idx": 0,
    "loss": 0.16666211187839508,
    "learning_rate": 0.00029782820223562737,
    "timestep_mean": 563.03125,
    "epsilon_norm": 6.8252763748168945,
    "prediction_norm": 5.863768100738525,
    "timestamp": "2025-07-14T12:34:18.657224"
  },
  {
    "epoch": 1265,
    "batch_idx": 32,
    "loss": 0.23666426539421082,
    "learning_rate": 0.00029782820223562737,
    "timestep_mean": 484.59375,
    "epsilon_norm": 6.879814624786377,
    "prediction_norm": 5.501803874969482,
    "timestamp": "2025-07-14T12:34:18.662394"
  },
  {
    "epoch": 1265,
    "batch_idx": 64,
    "loss": 0.4169732332229614,
    "learning_rate": 0.00029782820223562737,
    "timestep_mean": 465.78125,
    "epsilon_norm": 8.182083129882812,
    "prediction_norm": 6.208041667938232,
    "timestamp": "2025-07-14T12:34:18.667469"
  },
  {
    "epoch": 1265,
    "batch_idx": 96,
    "loss": 0.22312584519386292,
    "learning_rate": 0.00029782820223562737,
    "timestep_mean": 443.59375,
    "epsilon_norm": 7.016638278961182,
    "prediction_norm": 6.436595439910889,
    "timestamp": "2025-07-14T12:34:18.672555"
  },
  {
    "epoch": 1265,
    "batch_idx": 128,
    "loss": 0.21194693446159363,
    "learning_rate": 0.00029782820223562737,
    "timestep_mean": 472.40625,
    "epsilon_norm": 7.482365131378174,
    "prediction_norm": 6.716541767120361,
    "timestamp": "2025-07-14T12:34:18.677724"
  },
  {
    "epoch": 1265,
    "batch_idx": 160,
    "loss": 0.455360472202301,
    "learning_rate": 0.00029782820223562737,
    "timestep_mean": 487.75,
    "epsilon_norm": 8.030208587646484,
    "prediction_norm": 6.405221462249756,
    "timestamp": "2025-07-14T12:34:18.682779"
  },
  {
    "epoch": 1265,
    "batch_idx": 192,
    "loss": 0.7731294631958008,
    "learning_rate": 0.00029782820223562737,
    "timestep_mean": 538.0,
    "epsilon_norm": 4.934579372406006,
    "prediction_norm": 4.14302396774292,
    "timestamp": "2025-07-14T12:34:18.688097"
  },
  {
    "epoch": 1266,
    "batch_idx": 0,
    "loss": 0.3557892143726349,
    "learning_rate": 0.0002971101211668999,
    "timestep_mean": 474.65625,
    "epsilon_norm": 8.481352806091309,
    "prediction_norm": 7.406203746795654,
    "timestamp": "2025-07-14T12:34:18.693137"
  },
  {
    "epoch": 1266,
    "batch_idx": 32,
    "loss": 0.3150584399700165,
    "learning_rate": 0.0002971101211668999,
    "timestep_mean": 477.125,
    "epsilon_norm": 8.22730827331543,
    "prediction_norm": 6.719740867614746,
    "timestamp": "2025-07-14T12:34:18.698250"
  },
  {
    "epoch": 1266,
    "batch_idx": 64,
    "loss": 0.22309395670890808,
    "learning_rate": 0.0002971101211668999,
    "timestep_mean": 414.625,
    "epsilon_norm": 7.259192943572998,
    "prediction_norm": 5.80712890625,
    "timestamp": "2025-07-14T12:34:18.703293"
  },
  {
    "epoch": 1266,
    "batch_idx": 96,
    "loss": 0.18787947297096252,
    "learning_rate": 0.0002971101211668999,
    "timestep_mean": 464.875,
    "epsilon_norm": 6.891254425048828,
    "prediction_norm": 6.005162715911865,
    "timestamp": "2025-07-14T12:34:18.708360"
  },
  {
    "epoch": 1266,
    "batch_idx": 128,
    "loss": 0.18284085392951965,
    "learning_rate": 0.0002971101211668999,
    "timestep_mean": 470.75,
    "epsilon_norm": 8.109538078308105,
    "prediction_norm": 7.628711700439453,
    "timestamp": "2025-07-14T12:34:18.713364"
  },
  {
    "epoch": 1266,
    "batch_idx": 160,
    "loss": 0.3433187007904053,
    "learning_rate": 0.0002971101211668999,
    "timestep_mean": 547.90625,
    "epsilon_norm": 7.88515567779541,
    "prediction_norm": 6.776761054992676,
    "timestamp": "2025-07-14T12:34:18.718759"
  },
  {
    "epoch": 1266,
    "batch_idx": 192,
    "loss": 0.09706270694732666,
    "learning_rate": 0.0002971101211668999,
    "timestep_mean": 249.875,
    "epsilon_norm": 3.0355112552642822,
    "prediction_norm": 2.957648515701294,
    "timestamp": "2025-07-14T12:34:18.724080"
  },
  {
    "epoch": 1267,
    "batch_idx": 0,
    "loss": 0.2176702916622162,
    "learning_rate": 0.0002963925407087798,
    "timestep_mean": 602.40625,
    "epsilon_norm": 6.715611457824707,
    "prediction_norm": 5.781129360198975,
    "timestamp": "2025-07-14T12:34:18.729278"
  },
  {
    "epoch": 1267,
    "batch_idx": 32,
    "loss": 0.13943301141262054,
    "learning_rate": 0.0002963925407087798,
    "timestep_mean": 529.4375,
    "epsilon_norm": 7.001271724700928,
    "prediction_norm": 6.408466815948486,
    "timestamp": "2025-07-14T12:34:18.734371"
  },
  {
    "epoch": 1267,
    "batch_idx": 64,
    "loss": 0.3539777994155884,
    "learning_rate": 0.0002963925407087798,
    "timestep_mean": 550.40625,
    "epsilon_norm": 8.382122039794922,
    "prediction_norm": 6.302443981170654,
    "timestamp": "2025-07-14T12:34:18.739414"
  },
  {
    "epoch": 1267,
    "batch_idx": 96,
    "loss": 0.24669209122657776,
    "learning_rate": 0.0002963925407087798,
    "timestep_mean": 498.0,
    "epsilon_norm": 8.447770118713379,
    "prediction_norm": 7.003413677215576,
    "timestamp": "2025-07-14T12:34:18.744445"
  },
  {
    "epoch": 1267,
    "batch_idx": 128,
    "loss": 0.45089370012283325,
    "learning_rate": 0.0002963925407087798,
    "timestep_mean": 495.375,
    "epsilon_norm": 8.618553161621094,
    "prediction_norm": 6.979134559631348,
    "timestamp": "2025-07-14T12:34:18.749532"
  },
  {
    "epoch": 1267,
    "batch_idx": 160,
    "loss": 0.22157281637191772,
    "learning_rate": 0.0002963925407087798,
    "timestep_mean": 557.25,
    "epsilon_norm": 7.640505313873291,
    "prediction_norm": 6.679880619049072,
    "timestamp": "2025-07-14T12:34:18.754645"
  },
  {
    "epoch": 1267,
    "batch_idx": 192,
    "loss": 0.3469851613044739,
    "learning_rate": 0.0002963925407087798,
    "timestep_mean": 465.75,
    "epsilon_norm": 4.803621292114258,
    "prediction_norm": 3.4564671516418457,
    "timestamp": "2025-07-14T12:34:18.759769"
  },
  {
    "epoch": 1268,
    "batch_idx": 0,
    "loss": 0.1897244155406952,
    "learning_rate": 0.0002956754626318254,
    "timestep_mean": 505.1875,
    "epsilon_norm": 8.690967559814453,
    "prediction_norm": 8.077936172485352,
    "timestamp": "2025-07-14T12:34:18.764845"
  },
  {
    "epoch": 1268,
    "batch_idx": 32,
    "loss": 0.19036585092544556,
    "learning_rate": 0.0002956754626318254,
    "timestep_mean": 527.84375,
    "epsilon_norm": 6.619566440582275,
    "prediction_norm": 5.219174385070801,
    "timestamp": "2025-07-14T12:34:18.769821"
  },
  {
    "epoch": 1268,
    "batch_idx": 64,
    "loss": 0.13209515810012817,
    "learning_rate": 0.0002956754626318254,
    "timestep_mean": 586.40625,
    "epsilon_norm": 7.723222732543945,
    "prediction_norm": 6.601388931274414,
    "timestamp": "2025-07-14T12:34:18.775390"
  },
  {
    "epoch": 1268,
    "batch_idx": 96,
    "loss": 0.24767157435417175,
    "learning_rate": 0.0002956754626318254,
    "timestep_mean": 465.28125,
    "epsilon_norm": 8.0166015625,
    "prediction_norm": 6.872462272644043,
    "timestamp": "2025-07-14T12:34:18.780782"
  },
  {
    "epoch": 1268,
    "batch_idx": 128,
    "loss": 0.209907665848732,
    "learning_rate": 0.0002956754626318254,
    "timestep_mean": 499.15625,
    "epsilon_norm": 7.966141700744629,
    "prediction_norm": 6.869665145874023,
    "timestamp": "2025-07-14T12:34:18.786054"
  },
  {
    "epoch": 1268,
    "batch_idx": 160,
    "loss": 0.1675334870815277,
    "learning_rate": 0.0002956754626318254,
    "timestep_mean": 489.3125,
    "epsilon_norm": 8.222505569458008,
    "prediction_norm": 7.369697570800781,
    "timestamp": "2025-07-14T12:34:18.791497"
  },
  {
    "epoch": 1268,
    "batch_idx": 192,
    "loss": 0.5252764225006104,
    "learning_rate": 0.0002956754626318254,
    "timestep_mean": 420.5,
    "epsilon_norm": 4.715063095092773,
    "prediction_norm": 4.28627872467041,
    "timestamp": "2025-07-14T12:34:18.796649"
  },
  {
    "epoch": 1269,
    "batch_idx": 0,
    "loss": 0.21887639164924622,
    "learning_rate": 0.0002949588887053557,
    "timestep_mean": 496.53125,
    "epsilon_norm": 8.118783950805664,
    "prediction_norm": 7.242367267608643,
    "timestamp": "2025-07-14T12:34:18.801994"
  },
  {
    "epoch": 1269,
    "batch_idx": 32,
    "loss": 0.3119717836380005,
    "learning_rate": 0.0002949588887053557,
    "timestep_mean": 462.90625,
    "epsilon_norm": 8.19544506072998,
    "prediction_norm": 6.840205669403076,
    "timestamp": "2025-07-14T12:34:18.807557"
  },
  {
    "epoch": 1269,
    "batch_idx": 64,
    "loss": 0.23932573199272156,
    "learning_rate": 0.0002949588887053557,
    "timestep_mean": 598.5,
    "epsilon_norm": 7.048594951629639,
    "prediction_norm": 5.862248420715332,
    "timestamp": "2025-07-14T12:34:18.812711"
  },
  {
    "epoch": 1269,
    "batch_idx": 96,
    "loss": 0.3551331162452698,
    "learning_rate": 0.0002949588887053557,
    "timestep_mean": 492.34375,
    "epsilon_norm": 8.578858375549316,
    "prediction_norm": 6.223119258880615,
    "timestamp": "2025-07-14T12:34:18.817703"
  },
  {
    "epoch": 1269,
    "batch_idx": 128,
    "loss": 0.2752014398574829,
    "learning_rate": 0.0002949588887053557,
    "timestep_mean": 496.40625,
    "epsilon_norm": 7.7007856369018555,
    "prediction_norm": 6.371722221374512,
    "timestamp": "2025-07-14T12:34:18.822785"
  },
  {
    "epoch": 1269,
    "batch_idx": 160,
    "loss": 0.48643994331359863,
    "learning_rate": 0.0002949588887053557,
    "timestep_mean": 464.9375,
    "epsilon_norm": 8.275425910949707,
    "prediction_norm": 6.188831329345703,
    "timestamp": "2025-07-14T12:34:18.827800"
  },
  {
    "epoch": 1269,
    "batch_idx": 192,
    "loss": 0.06165879964828491,
    "learning_rate": 0.0002949588887053557,
    "timestep_mean": 733.25,
    "epsilon_norm": 3.2980847358703613,
    "prediction_norm": 3.4270355701446533,
    "timestamp": "2025-07-14T12:34:18.832995"
  },
  {
    "epoch": 1270,
    "batch_idx": 0,
    "loss": 0.2745244801044464,
    "learning_rate": 0.0002942428206974455,
    "timestep_mean": 550.5625,
    "epsilon_norm": 7.468367576599121,
    "prediction_norm": 6.079509258270264,
    "timestamp": "2025-07-14T12:34:18.838093"
  },
  {
    "epoch": 1270,
    "batch_idx": 32,
    "loss": 0.15357549488544464,
    "learning_rate": 0.0002942428206974455,
    "timestep_mean": 493.78125,
    "epsilon_norm": 9.063675880432129,
    "prediction_norm": 7.946784973144531,
    "timestamp": "2025-07-14T12:34:18.843062"
  },
  {
    "epoch": 1270,
    "batch_idx": 64,
    "loss": 0.2078828066587448,
    "learning_rate": 0.0002942428206974455,
    "timestep_mean": 535.1875,
    "epsilon_norm": 7.461977481842041,
    "prediction_norm": 6.418546676635742,
    "timestamp": "2025-07-14T12:34:18.848026"
  },
  {
    "epoch": 1270,
    "batch_idx": 96,
    "loss": 0.3443886637687683,
    "learning_rate": 0.0002942428206974455,
    "timestep_mean": 459.28125,
    "epsilon_norm": 7.416867256164551,
    "prediction_norm": 5.817552089691162,
    "timestamp": "2025-07-14T12:34:18.853041"
  },
  {
    "epoch": 1270,
    "batch_idx": 128,
    "loss": 0.15713973343372345,
    "learning_rate": 0.0002942428206974455,
    "timestep_mean": 470.90625,
    "epsilon_norm": 7.915502548217773,
    "prediction_norm": 7.0413713455200195,
    "timestamp": "2025-07-14T12:34:18.858038"
  },
  {
    "epoch": 1270,
    "batch_idx": 160,
    "loss": 0.15755686163902283,
    "learning_rate": 0.0002942428206974455,
    "timestep_mean": 534.59375,
    "epsilon_norm": 6.532055854797363,
    "prediction_norm": 5.208934307098389,
    "timestamp": "2025-07-14T12:34:18.863110"
  },
  {
    "epoch": 1270,
    "batch_idx": 192,
    "loss": 0.23194476962089539,
    "learning_rate": 0.0002942428206974455,
    "timestep_mean": 489.0,
    "epsilon_norm": 4.4429931640625,
    "prediction_norm": 3.883061647415161,
    "timestamp": "2025-07-14T12:34:18.867978"
  },
  {
    "epoch": 1271,
    "batch_idx": 0,
    "loss": 0.1703229695558548,
    "learning_rate": 0.0002935272603749215,
    "timestep_mean": 505.46875,
    "epsilon_norm": 8.57400131225586,
    "prediction_norm": 8.07717514038086,
    "timestamp": "2025-07-14T12:34:18.872854"
  },
  {
    "epoch": 1271,
    "batch_idx": 32,
    "loss": 0.16856497526168823,
    "learning_rate": 0.0002935272603749215,
    "timestep_mean": 527.5,
    "epsilon_norm": 7.7484917640686035,
    "prediction_norm": 6.756720542907715,
    "timestamp": "2025-07-14T12:34:18.877846"
  },
  {
    "epoch": 1271,
    "batch_idx": 64,
    "loss": 0.5066014528274536,
    "learning_rate": 0.0002935272603749215,
    "timestep_mean": 435.53125,
    "epsilon_norm": 10.224296569824219,
    "prediction_norm": 7.821399211883545,
    "timestamp": "2025-07-14T12:34:18.882763"
  },
  {
    "epoch": 1271,
    "batch_idx": 96,
    "loss": 0.16428512334823608,
    "learning_rate": 0.0002935272603749215,
    "timestep_mean": 550.1875,
    "epsilon_norm": 6.942742347717285,
    "prediction_norm": 6.294353008270264,
    "timestamp": "2025-07-14T12:34:18.887796"
  },
  {
    "epoch": 1271,
    "batch_idx": 128,
    "loss": 0.22957032918930054,
    "learning_rate": 0.0002935272603749215,
    "timestep_mean": 450.25,
    "epsilon_norm": 7.208493232727051,
    "prediction_norm": 5.937664985656738,
    "timestamp": "2025-07-14T12:34:18.892784"
  },
  {
    "epoch": 1271,
    "batch_idx": 160,
    "loss": 0.24331055581569672,
    "learning_rate": 0.0002935272603749215,
    "timestep_mean": 552.28125,
    "epsilon_norm": 8.806407928466797,
    "prediction_norm": 7.999915599822998,
    "timestamp": "2025-07-14T12:34:18.897836"
  },
  {
    "epoch": 1271,
    "batch_idx": 192,
    "loss": 0.3845447301864624,
    "learning_rate": 0.0002935272603749215,
    "timestep_mean": 592.75,
    "epsilon_norm": 3.372802972793579,
    "prediction_norm": 2.5633015632629395,
    "timestamp": "2025-07-14T12:34:18.902787"
  },
  {
    "epoch": 1272,
    "batch_idx": 0,
    "loss": 0.1177276074886322,
    "learning_rate": 0.0002928122095033579,
    "timestep_mean": 574.34375,
    "epsilon_norm": 8.139161109924316,
    "prediction_norm": 7.4623565673828125,
    "timestamp": "2025-07-14T12:34:18.907702"
  },
  {
    "epoch": 1272,
    "batch_idx": 32,
    "loss": 0.3205912411212921,
    "learning_rate": 0.0002928122095033579,
    "timestep_mean": 508.65625,
    "epsilon_norm": 8.99940013885498,
    "prediction_norm": 6.436734199523926,
    "timestamp": "2025-07-14T12:34:18.912647"
  },
  {
    "epoch": 1272,
    "batch_idx": 64,
    "loss": 0.2783339321613312,
    "learning_rate": 0.0002928122095033579,
    "timestep_mean": 505.5,
    "epsilon_norm": 8.07907772064209,
    "prediction_norm": 7.5048508644104,
    "timestamp": "2025-07-14T12:34:18.917559"
  },
  {
    "epoch": 1272,
    "batch_idx": 96,
    "loss": 0.2086217999458313,
    "learning_rate": 0.0002928122095033579,
    "timestep_mean": 523.25,
    "epsilon_norm": 7.7899017333984375,
    "prediction_norm": 6.650425434112549,
    "timestamp": "2025-07-14T12:34:18.922545"
  },
  {
    "epoch": 1272,
    "batch_idx": 128,
    "loss": 0.4048001170158386,
    "learning_rate": 0.0002928122095033579,
    "timestep_mean": 532.46875,
    "epsilon_norm": 9.388823509216309,
    "prediction_norm": 8.02091121673584,
    "timestamp": "2025-07-14T12:34:18.927416"
  },
  {
    "epoch": 1272,
    "batch_idx": 160,
    "loss": 0.33835744857788086,
    "learning_rate": 0.0002928122095033579,
    "timestep_mean": 474.75,
    "epsilon_norm": 8.630791664123535,
    "prediction_norm": 7.002167701721191,
    "timestamp": "2025-07-14T12:34:18.932277"
  },
  {
    "epoch": 1272,
    "batch_idx": 192,
    "loss": 0.5076325535774231,
    "learning_rate": 0.0002928122095033579,
    "timestep_mean": 399.25,
    "epsilon_norm": 3.7236766815185547,
    "prediction_norm": 2.0446665287017822,
    "timestamp": "2025-07-14T12:34:18.937179"
  },
  {
    "epoch": 1273,
    "batch_idx": 0,
    "loss": 0.3050882816314697,
    "learning_rate": 0.00029209766984707126,
    "timestep_mean": 536.34375,
    "epsilon_norm": 8.921585083007812,
    "prediction_norm": 8.183221817016602,
    "timestamp": "2025-07-14T12:34:18.942119"
  },
  {
    "epoch": 1273,
    "batch_idx": 32,
    "loss": 0.3181157410144806,
    "learning_rate": 0.00029209766984707126,
    "timestep_mean": 497.71875,
    "epsilon_norm": 9.17885684967041,
    "prediction_norm": 8.173872947692871,
    "timestamp": "2025-07-14T12:34:18.947023"
  },
  {
    "epoch": 1273,
    "batch_idx": 64,
    "loss": 0.33494946360588074,
    "learning_rate": 0.00029209766984707126,
    "timestep_mean": 500.9375,
    "epsilon_norm": 7.420801639556885,
    "prediction_norm": 5.903400897979736,
    "timestamp": "2025-07-14T12:34:18.951980"
  },
  {
    "epoch": 1273,
    "batch_idx": 96,
    "loss": 0.2142530381679535,
    "learning_rate": 0.00029209766984707126,
    "timestep_mean": 545.6875,
    "epsilon_norm": 8.941784858703613,
    "prediction_norm": 8.321219444274902,
    "timestamp": "2025-07-14T12:34:18.956859"
  },
  {
    "epoch": 1273,
    "batch_idx": 128,
    "loss": 0.2882145047187805,
    "learning_rate": 0.00029209766984707126,
    "timestep_mean": 541.375,
    "epsilon_norm": 8.656902313232422,
    "prediction_norm": 7.099612712860107,
    "timestamp": "2025-07-14T12:34:18.961769"
  },
  {
    "epoch": 1273,
    "batch_idx": 160,
    "loss": 0.37436288595199585,
    "learning_rate": 0.00029209766984707126,
    "timestep_mean": 511.25,
    "epsilon_norm": 7.065009593963623,
    "prediction_norm": 5.228158473968506,
    "timestamp": "2025-07-14T12:34:18.966724"
  },
  {
    "epoch": 1273,
    "batch_idx": 192,
    "loss": 0.3173511028289795,
    "learning_rate": 0.00029209766984707126,
    "timestep_mean": 557.25,
    "epsilon_norm": 2.9780333042144775,
    "prediction_norm": 2.097041606903076,
    "timestamp": "2025-07-14T12:34:18.971662"
  },
  {
    "epoch": 1274,
    "batch_idx": 0,
    "loss": 0.44381821155548096,
    "learning_rate": 0.00029138364316911745,
    "timestep_mean": 454.03125,
    "epsilon_norm": 8.413992881774902,
    "prediction_norm": 6.3520941734313965,
    "timestamp": "2025-07-14T12:34:18.976536"
  },
  {
    "epoch": 1274,
    "batch_idx": 32,
    "loss": 0.44184815883636475,
    "learning_rate": 0.00029138364316911745,
    "timestep_mean": 502.84375,
    "epsilon_norm": 7.637929916381836,
    "prediction_norm": 6.282630443572998,
    "timestamp": "2025-07-14T12:34:18.981438"
  },
  {
    "epoch": 1274,
    "batch_idx": 64,
    "loss": 0.20343175530433655,
    "learning_rate": 0.00029138364316911745,
    "timestep_mean": 559.53125,
    "epsilon_norm": 8.304794311523438,
    "prediction_norm": 7.342691898345947,
    "timestamp": "2025-07-14T12:34:18.986527"
  },
  {
    "epoch": 1274,
    "batch_idx": 96,
    "loss": 0.21802301704883575,
    "learning_rate": 0.00029138364316911745,
    "timestep_mean": 429.53125,
    "epsilon_norm": 6.785830497741699,
    "prediction_norm": 6.177567005157471,
    "timestamp": "2025-07-14T12:34:18.991761"
  },
  {
    "epoch": 1274,
    "batch_idx": 128,
    "loss": 0.27034205198287964,
    "learning_rate": 0.00029138364316911745,
    "timestep_mean": 482.40625,
    "epsilon_norm": 7.502599716186523,
    "prediction_norm": 6.777787685394287,
    "timestamp": "2025-07-14T12:34:18.996894"
  },
  {
    "epoch": 1274,
    "batch_idx": 160,
    "loss": 0.19844508171081543,
    "learning_rate": 0.00029138364316911745,
    "timestep_mean": 577.375,
    "epsilon_norm": 7.6009521484375,
    "prediction_norm": 7.2680768966674805,
    "timestamp": "2025-07-14T12:34:19.001954"
  },
  {
    "epoch": 1274,
    "batch_idx": 192,
    "loss": 0.24312248826026917,
    "learning_rate": 0.00029138364316911745,
    "timestep_mean": 475.625,
    "epsilon_norm": 3.5771470069885254,
    "prediction_norm": 3.310041904449463,
    "timestamp": "2025-07-14T12:34:19.006928"
  },
  {
    "epoch": 1275,
    "batch_idx": 0,
    "loss": 0.26081493496894836,
    "learning_rate": 0.00029067013123128586,
    "timestep_mean": 551.46875,
    "epsilon_norm": 7.062150478363037,
    "prediction_norm": 6.164296627044678,
    "timestamp": "2025-07-14T12:34:19.012209"
  },
  {
    "epoch": 1275,
    "batch_idx": 32,
    "loss": 0.2570449113845825,
    "learning_rate": 0.00029067013123128586,
    "timestep_mean": 508.71875,
    "epsilon_norm": 7.3036041259765625,
    "prediction_norm": 6.611819267272949,
    "timestamp": "2025-07-14T12:34:19.017123"
  },
  {
    "epoch": 1275,
    "batch_idx": 64,
    "loss": 0.15296123921871185,
    "learning_rate": 0.00029067013123128586,
    "timestep_mean": 575.9375,
    "epsilon_norm": 9.058550834655762,
    "prediction_norm": 8.395014762878418,
    "timestamp": "2025-07-14T12:34:19.022117"
  },
  {
    "epoch": 1275,
    "batch_idx": 96,
    "loss": 0.40059858560562134,
    "learning_rate": 0.00029067013123128586,
    "timestep_mean": 489.28125,
    "epsilon_norm": 9.272655487060547,
    "prediction_norm": 7.064544677734375,
    "timestamp": "2025-07-14T12:34:19.027192"
  },
  {
    "epoch": 1275,
    "batch_idx": 128,
    "loss": 0.26911529898643494,
    "learning_rate": 0.00029067013123128586,
    "timestep_mean": 469.90625,
    "epsilon_norm": 7.887272357940674,
    "prediction_norm": 6.290907382965088,
    "timestamp": "2025-07-14T12:34:19.032359"
  },
  {
    "epoch": 1275,
    "batch_idx": 160,
    "loss": 0.14553551375865936,
    "learning_rate": 0.00029067013123128586,
    "timestep_mean": 564.28125,
    "epsilon_norm": 7.679121017456055,
    "prediction_norm": 7.2183451652526855,
    "timestamp": "2025-07-14T12:34:19.037477"
  },
  {
    "epoch": 1275,
    "batch_idx": 192,
    "loss": 0.6024901866912842,
    "learning_rate": 0.00029067013123128586,
    "timestep_mean": 381.5,
    "epsilon_norm": 4.709517002105713,
    "prediction_norm": 2.5589187145233154,
    "timestamp": "2025-07-14T12:34:19.042554"
  },
  {
    "epoch": 1276,
    "batch_idx": 0,
    "loss": 0.17614971101284027,
    "learning_rate": 0.00028995713579409686,
    "timestep_mean": 588.21875,
    "epsilon_norm": 7.408395290374756,
    "prediction_norm": 6.316246032714844,
    "timestamp": "2025-07-14T12:34:19.047681"
  },
  {
    "epoch": 1276,
    "batch_idx": 32,
    "loss": 0.47804075479507446,
    "learning_rate": 0.00028995713579409686,
    "timestep_mean": 445.59375,
    "epsilon_norm": 7.497607231140137,
    "prediction_norm": 6.48621129989624,
    "timestamp": "2025-07-14T12:34:19.052831"
  },
  {
    "epoch": 1276,
    "batch_idx": 64,
    "loss": 0.2539520859718323,
    "learning_rate": 0.00028995713579409686,
    "timestep_mean": 544.90625,
    "epsilon_norm": 7.7199883460998535,
    "prediction_norm": 6.4460530281066895,
    "timestamp": "2025-07-14T12:34:19.057944"
  },
  {
    "epoch": 1276,
    "batch_idx": 96,
    "loss": 0.2210545837879181,
    "learning_rate": 0.00028995713579409686,
    "timestep_mean": 538.15625,
    "epsilon_norm": 8.539203643798828,
    "prediction_norm": 7.2024078369140625,
    "timestamp": "2025-07-14T12:34:19.063205"
  },
  {
    "epoch": 1276,
    "batch_idx": 128,
    "loss": 0.3003939390182495,
    "learning_rate": 0.00028995713579409686,
    "timestep_mean": 526.84375,
    "epsilon_norm": 8.392221450805664,
    "prediction_norm": 7.327481269836426,
    "timestamp": "2025-07-14T12:34:19.068458"
  },
  {
    "epoch": 1276,
    "batch_idx": 160,
    "loss": 0.20829066634178162,
    "learning_rate": 0.00028995713579409686,
    "timestep_mean": 537.09375,
    "epsilon_norm": 7.9412522315979,
    "prediction_norm": 6.598842620849609,
    "timestamp": "2025-07-14T12:34:19.073566"
  },
  {
    "epoch": 1276,
    "batch_idx": 192,
    "loss": 0.1284378468990326,
    "learning_rate": 0.00028995713579409686,
    "timestep_mean": 455.75,
    "epsilon_norm": 3.452928304672241,
    "prediction_norm": 3.3187813758850098,
    "timestamp": "2025-07-14T12:34:19.078519"
  },
  {
    "epoch": 1277,
    "batch_idx": 0,
    "loss": 0.49252942204475403,
    "learning_rate": 0.00028924465861679546,
    "timestep_mean": 408.03125,
    "epsilon_norm": 8.014825820922852,
    "prediction_norm": 6.310433864593506,
    "timestamp": "2025-07-14T12:34:19.083814"
  },
  {
    "epoch": 1277,
    "batch_idx": 32,
    "loss": 0.27980220317840576,
    "learning_rate": 0.00028924465861679546,
    "timestep_mean": 471.5,
    "epsilon_norm": 7.740678787231445,
    "prediction_norm": 6.067254543304443,
    "timestamp": "2025-07-14T12:34:19.088724"
  },
  {
    "epoch": 1277,
    "batch_idx": 64,
    "loss": 0.37768447399139404,
    "learning_rate": 0.00028924465861679546,
    "timestep_mean": 400.5625,
    "epsilon_norm": 7.2546186447143555,
    "prediction_norm": 5.221668720245361,
    "timestamp": "2025-07-14T12:34:19.093556"
  },
  {
    "epoch": 1277,
    "batch_idx": 96,
    "loss": 0.16523712873458862,
    "learning_rate": 0.00028924465861679546,
    "timestep_mean": 555.59375,
    "epsilon_norm": 8.052706718444824,
    "prediction_norm": 6.79252815246582,
    "timestamp": "2025-07-14T12:34:19.098441"
  },
  {
    "epoch": 1277,
    "batch_idx": 128,
    "loss": 0.288380891084671,
    "learning_rate": 0.00028924465861679546,
    "timestep_mean": 488.5625,
    "epsilon_norm": 7.635833740234375,
    "prediction_norm": 6.17796516418457,
    "timestamp": "2025-07-14T12:34:19.103311"
  },
  {
    "epoch": 1277,
    "batch_idx": 160,
    "loss": 0.2758561372756958,
    "learning_rate": 0.00028924465861679546,
    "timestep_mean": 502.8125,
    "epsilon_norm": 7.628814697265625,
    "prediction_norm": 6.176946640014648,
    "timestamp": "2025-07-14T12:34:19.108664"
  },
  {
    "epoch": 1277,
    "batch_idx": 192,
    "loss": 0.45715096592903137,
    "learning_rate": 0.00028924465861679546,
    "timestep_mean": 648.375,
    "epsilon_norm": 3.6275570392608643,
    "prediction_norm": 3.143916130065918,
    "timestamp": "2025-07-14T12:34:19.113778"
  },
  {
    "epoch": 1278,
    "batch_idx": 0,
    "loss": 0.3053880035877228,
    "learning_rate": 0.0002885327014573484,
    "timestep_mean": 563.03125,
    "epsilon_norm": 8.442095756530762,
    "prediction_norm": 7.2846999168396,
    "timestamp": "2025-07-14T12:34:19.118829"
  },
  {
    "epoch": 1278,
    "batch_idx": 32,
    "loss": 0.3861932158470154,
    "learning_rate": 0.0002885327014573484,
    "timestep_mean": 434.125,
    "epsilon_norm": 8.628954887390137,
    "prediction_norm": 6.947742938995361,
    "timestamp": "2025-07-14T12:34:19.123979"
  },
  {
    "epoch": 1278,
    "batch_idx": 64,
    "loss": 0.22846180200576782,
    "learning_rate": 0.0002885327014573484,
    "timestep_mean": 554.1875,
    "epsilon_norm": 8.545883178710938,
    "prediction_norm": 7.345152854919434,
    "timestamp": "2025-07-14T12:34:19.129063"
  },
  {
    "epoch": 1278,
    "batch_idx": 96,
    "loss": 0.3508099913597107,
    "learning_rate": 0.0002885327014573484,
    "timestep_mean": 497.84375,
    "epsilon_norm": 7.878657817840576,
    "prediction_norm": 6.816282749176025,
    "timestamp": "2025-07-14T12:34:19.134157"
  },
  {
    "epoch": 1278,
    "batch_idx": 128,
    "loss": 0.19937455654144287,
    "learning_rate": 0.0002885327014573484,
    "timestep_mean": 526.3125,
    "epsilon_norm": 7.07110071182251,
    "prediction_norm": 6.107511520385742,
    "timestamp": "2025-07-14T12:34:19.139347"
  },
  {
    "epoch": 1278,
    "batch_idx": 160,
    "loss": 0.6646177768707275,
    "learning_rate": 0.0002885327014573484,
    "timestep_mean": 504.28125,
    "epsilon_norm": 8.258909225463867,
    "prediction_norm": 5.787789821624756,
    "timestamp": "2025-07-14T12:34:19.144581"
  },
  {
    "epoch": 1278,
    "batch_idx": 192,
    "loss": 0.268104612827301,
    "learning_rate": 0.0002885327014573484,
    "timestep_mean": 413.75,
    "epsilon_norm": 3.5887680053710938,
    "prediction_norm": 2.6142640113830566,
    "timestamp": "2025-07-14T12:34:19.149728"
  },
  {
    "epoch": 1279,
    "batch_idx": 0,
    "loss": 0.19785888493061066,
    "learning_rate": 0.00028782126607243894,
    "timestep_mean": 461.21875,
    "epsilon_norm": 7.6516594886779785,
    "prediction_norm": 6.711873531341553,
    "timestamp": "2025-07-14T12:34:19.155043"
  },
  {
    "epoch": 1279,
    "batch_idx": 32,
    "loss": 0.24072080850601196,
    "learning_rate": 0.00028782126607243894,
    "timestep_mean": 553.8125,
    "epsilon_norm": 7.430846691131592,
    "prediction_norm": 6.703164100646973,
    "timestamp": "2025-07-14T12:34:19.160181"
  },
  {
    "epoch": 1279,
    "batch_idx": 64,
    "loss": 0.19059911370277405,
    "learning_rate": 0.00028782126607243894,
    "timestep_mean": 562.21875,
    "epsilon_norm": 7.565258502960205,
    "prediction_norm": 6.687037467956543,
    "timestamp": "2025-07-14T12:34:19.165389"
  },
  {
    "epoch": 1279,
    "batch_idx": 96,
    "loss": 0.19148652255535126,
    "learning_rate": 0.00028782126607243894,
    "timestep_mean": 508.625,
    "epsilon_norm": 7.810782432556152,
    "prediction_norm": 7.351254940032959,
    "timestamp": "2025-07-14T12:34:19.170532"
  },
  {
    "epoch": 1279,
    "batch_idx": 128,
    "loss": 0.2714540660381317,
    "learning_rate": 0.00028782126607243894,
    "timestep_mean": 468.125,
    "epsilon_norm": 7.654580593109131,
    "prediction_norm": 6.074922561645508,
    "timestamp": "2025-07-14T12:34:19.175685"
  },
  {
    "epoch": 1279,
    "batch_idx": 160,
    "loss": 0.31327393651008606,
    "learning_rate": 0.00028782126607243894,
    "timestep_mean": 476.09375,
    "epsilon_norm": 8.838730812072754,
    "prediction_norm": 7.24644136428833,
    "timestamp": "2025-07-14T12:34:19.180551"
  },
  {
    "epoch": 1279,
    "batch_idx": 192,
    "loss": 0.20237354934215546,
    "learning_rate": 0.00028782126607243894,
    "timestep_mean": 454.625,
    "epsilon_norm": 4.247626304626465,
    "prediction_norm": 3.455451726913452,
    "timestamp": "2025-07-14T12:34:19.185372"
  },
  {
    "epoch": 1280,
    "batch_idx": 0,
    "loss": 0.46944501996040344,
    "learning_rate": 0.00028711035421746366,
    "timestep_mean": 501.34375,
    "epsilon_norm": 9.053744316101074,
    "prediction_norm": 6.7856879234313965,
    "timestamp": "2025-07-14T12:34:19.190308"
  },
  {
    "epoch": 1280,
    "batch_idx": 32,
    "loss": 0.1941695511341095,
    "learning_rate": 0.00028711035421746366,
    "timestep_mean": 482.28125,
    "epsilon_norm": 7.120289325714111,
    "prediction_norm": 6.023530006408691,
    "timestamp": "2025-07-14T12:34:19.195195"
  },
  {
    "epoch": 1280,
    "batch_idx": 64,
    "loss": 0.23195409774780273,
    "learning_rate": 0.00028711035421746366,
    "timestep_mean": 530.28125,
    "epsilon_norm": 7.026757717132568,
    "prediction_norm": 5.673133373260498,
    "timestamp": "2025-07-14T12:34:19.200168"
  },
  {
    "epoch": 1280,
    "batch_idx": 96,
    "loss": 0.2815564274787903,
    "learning_rate": 0.00028711035421746366,
    "timestep_mean": 589.71875,
    "epsilon_norm": 9.115077018737793,
    "prediction_norm": 8.182705879211426,
    "timestamp": "2025-07-14T12:34:19.205079"
  },
  {
    "epoch": 1280,
    "batch_idx": 128,
    "loss": 0.2089829444885254,
    "learning_rate": 0.00028711035421746366,
    "timestep_mean": 483.8125,
    "epsilon_norm": 7.869243621826172,
    "prediction_norm": 6.870650291442871,
    "timestamp": "2025-07-14T12:34:19.209998"
  },
  {
    "epoch": 1280,
    "batch_idx": 160,
    "loss": 0.3649201989173889,
    "learning_rate": 0.00028711035421746366,
    "timestep_mean": 492.46875,
    "epsilon_norm": 8.28178882598877,
    "prediction_norm": 6.849498271942139,
    "timestamp": "2025-07-14T12:34:19.214860"
  },
  {
    "epoch": 1280,
    "batch_idx": 192,
    "loss": 0.3258865773677826,
    "learning_rate": 0.00028711035421746366,
    "timestep_mean": 606.625,
    "epsilon_norm": 3.6772372722625732,
    "prediction_norm": 3.024775981903076,
    "timestamp": "2025-07-14T12:34:19.219737"
  },
  {
    "epoch": 1281,
    "batch_idx": 0,
    "loss": 0.3692297339439392,
    "learning_rate": 0.0002863999676465265,
    "timestep_mean": 557.9375,
    "epsilon_norm": 8.63298225402832,
    "prediction_norm": 6.702762126922607,
    "timestamp": "2025-07-14T12:34:19.224658"
  },
  {
    "epoch": 1281,
    "batch_idx": 32,
    "loss": 0.1707252860069275,
    "learning_rate": 0.0002863999676465265,
    "timestep_mean": 525.21875,
    "epsilon_norm": 7.523462295532227,
    "prediction_norm": 6.480748176574707,
    "timestamp": "2025-07-14T12:34:19.229603"
  },
  {
    "epoch": 1281,
    "batch_idx": 64,
    "loss": 0.10871811211109161,
    "learning_rate": 0.0002863999676465265,
    "timestep_mean": 640.53125,
    "epsilon_norm": 7.607901573181152,
    "prediction_norm": 7.559514045715332,
    "timestamp": "2025-07-14T12:34:19.234538"
  },
  {
    "epoch": 1281,
    "batch_idx": 96,
    "loss": 0.2530510723590851,
    "learning_rate": 0.0002863999676465265,
    "timestep_mean": 539.375,
    "epsilon_norm": 8.591423034667969,
    "prediction_norm": 7.968430995941162,
    "timestamp": "2025-07-14T12:34:19.239397"
  },
  {
    "epoch": 1281,
    "batch_idx": 128,
    "loss": 0.2641822397708893,
    "learning_rate": 0.0002863999676465265,
    "timestep_mean": 503.78125,
    "epsilon_norm": 7.713953018188477,
    "prediction_norm": 6.469976425170898,
    "timestamp": "2025-07-14T12:34:19.244273"
  },
  {
    "epoch": 1281,
    "batch_idx": 160,
    "loss": 0.15362003445625305,
    "learning_rate": 0.0002863999676465265,
    "timestep_mean": 514.15625,
    "epsilon_norm": 6.874802112579346,
    "prediction_norm": 6.07924222946167,
    "timestamp": "2025-07-14T12:34:19.249114"
  },
  {
    "epoch": 1281,
    "batch_idx": 192,
    "loss": 0.187677800655365,
    "learning_rate": 0.0002863999676465265,
    "timestep_mean": 477.75,
    "epsilon_norm": 3.75533390045166,
    "prediction_norm": 3.48113751411438,
    "timestamp": "2025-07-14T12:34:19.253952"
  },
  {
    "epoch": 1282,
    "batch_idx": 0,
    "loss": 0.181494802236557,
    "learning_rate": 0.00028569010811243584,
    "timestep_mean": 536.53125,
    "epsilon_norm": 7.97404146194458,
    "prediction_norm": 7.151331424713135,
    "timestamp": "2025-07-14T12:34:19.258941"
  },
  {
    "epoch": 1282,
    "batch_idx": 32,
    "loss": 0.2255793958902359,
    "learning_rate": 0.00028569010811243584,
    "timestep_mean": 432.40625,
    "epsilon_norm": 8.616623878479004,
    "prediction_norm": 7.810323238372803,
    "timestamp": "2025-07-14T12:34:19.263854"
  },
  {
    "epoch": 1282,
    "batch_idx": 64,
    "loss": 0.23406288027763367,
    "learning_rate": 0.00028569010811243584,
    "timestep_mean": 501.875,
    "epsilon_norm": 7.472503662109375,
    "prediction_norm": 6.522593021392822,
    "timestamp": "2025-07-14T12:34:19.268697"
  },
  {
    "epoch": 1282,
    "batch_idx": 96,
    "loss": 0.10652099549770355,
    "learning_rate": 0.00028569010811243584,
    "timestep_mean": 459.15625,
    "epsilon_norm": 7.41271448135376,
    "prediction_norm": 7.0341105461120605,
    "timestamp": "2025-07-14T12:34:19.273611"
  },
  {
    "epoch": 1282,
    "batch_idx": 128,
    "loss": 0.29058265686035156,
    "learning_rate": 0.00028569010811243584,
    "timestep_mean": 494.8125,
    "epsilon_norm": 8.54189682006836,
    "prediction_norm": 6.832901477813721,
    "timestamp": "2025-07-14T12:34:19.278509"
  },
  {
    "epoch": 1282,
    "batch_idx": 160,
    "loss": 0.25706160068511963,
    "learning_rate": 0.00028569010811243584,
    "timestep_mean": 488.15625,
    "epsilon_norm": 7.360894680023193,
    "prediction_norm": 6.52323579788208,
    "timestamp": "2025-07-14T12:34:19.283496"
  },
  {
    "epoch": 1282,
    "batch_idx": 192,
    "loss": 0.1694379597902298,
    "learning_rate": 0.00028569010811243584,
    "timestep_mean": 499.125,
    "epsilon_norm": 4.333583831787109,
    "prediction_norm": 3.910297393798828,
    "timestamp": "2025-07-14T12:34:19.288313"
  },
  {
    "epoch": 1283,
    "batch_idx": 0,
    "loss": 0.19851939380168915,
    "learning_rate": 0.00028498077736669955,
    "timestep_mean": 550.25,
    "epsilon_norm": 7.871639728546143,
    "prediction_norm": 6.9344916343688965,
    "timestamp": "2025-07-14T12:34:19.293148"
  },
  {
    "epoch": 1283,
    "batch_idx": 32,
    "loss": 0.10679707676172256,
    "learning_rate": 0.00028498077736669955,
    "timestep_mean": 573.53125,
    "epsilon_norm": 5.612505912780762,
    "prediction_norm": 5.306960105895996,
    "timestamp": "2025-07-14T12:34:19.298041"
  },
  {
    "epoch": 1283,
    "batch_idx": 64,
    "loss": 0.3018285930156708,
    "learning_rate": 0.00028498077736669955,
    "timestep_mean": 472.8125,
    "epsilon_norm": 7.906705856323242,
    "prediction_norm": 6.373226165771484,
    "timestamp": "2025-07-14T12:34:19.302924"
  },
  {
    "epoch": 1283,
    "batch_idx": 96,
    "loss": 0.41292309761047363,
    "learning_rate": 0.00028498077736669955,
    "timestep_mean": 486.3125,
    "epsilon_norm": 7.956089973449707,
    "prediction_norm": 5.90482234954834,
    "timestamp": "2025-07-14T12:34:19.307821"
  },
  {
    "epoch": 1283,
    "batch_idx": 128,
    "loss": 0.49377644062042236,
    "learning_rate": 0.00028498077736669955,
    "timestep_mean": 502.96875,
    "epsilon_norm": 8.173219680786133,
    "prediction_norm": 6.2171430587768555,
    "timestamp": "2025-07-14T12:34:19.312793"
  },
  {
    "epoch": 1283,
    "batch_idx": 160,
    "loss": 0.2704591453075409,
    "learning_rate": 0.00028498077736669955,
    "timestep_mean": 440.53125,
    "epsilon_norm": 8.815648078918457,
    "prediction_norm": 7.273303985595703,
    "timestamp": "2025-07-14T12:34:19.317713"
  },
  {
    "epoch": 1283,
    "batch_idx": 192,
    "loss": 0.43942415714263916,
    "learning_rate": 0.00028498077736669955,
    "timestep_mean": 417.625,
    "epsilon_norm": 3.302845001220703,
    "prediction_norm": 2.50274395942688,
    "timestamp": "2025-07-14T12:34:19.322536"
  },
  {
    "epoch": 1284,
    "batch_idx": 0,
    "loss": 0.2417571097612381,
    "learning_rate": 0.0002842719771595206,
    "timestep_mean": 537.65625,
    "epsilon_norm": 8.023392677307129,
    "prediction_norm": 7.132122993469238,
    "timestamp": "2025-07-14T12:34:19.327416"
  },
  {
    "epoch": 1284,
    "batch_idx": 32,
    "loss": 0.3980027139186859,
    "learning_rate": 0.0002842719771595206,
    "timestep_mean": 449.3125,
    "epsilon_norm": 7.835251808166504,
    "prediction_norm": 6.11273717880249,
    "timestamp": "2025-07-14T12:34:19.332237"
  },
  {
    "epoch": 1284,
    "batch_idx": 64,
    "loss": 0.29438841342926025,
    "learning_rate": 0.0002842719771595206,
    "timestep_mean": 456.40625,
    "epsilon_norm": 7.0724196434021,
    "prediction_norm": 5.9715895652771,
    "timestamp": "2025-07-14T12:34:19.337096"
  },
  {
    "epoch": 1284,
    "batch_idx": 96,
    "loss": 0.29936641454696655,
    "learning_rate": 0.0002842719771595206,
    "timestep_mean": 544.4375,
    "epsilon_norm": 8.295987129211426,
    "prediction_norm": 7.6075029373168945,
    "timestamp": "2025-07-14T12:34:19.342082"
  },
  {
    "epoch": 1284,
    "batch_idx": 128,
    "loss": 0.3266058564186096,
    "learning_rate": 0.0002842719771595206,
    "timestep_mean": 529.8125,
    "epsilon_norm": 8.761042594909668,
    "prediction_norm": 7.6280975341796875,
    "timestamp": "2025-07-14T12:34:19.346941"
  },
  {
    "epoch": 1284,
    "batch_idx": 160,
    "loss": 0.43073248863220215,
    "learning_rate": 0.0002842719771595206,
    "timestep_mean": 483.15625,
    "epsilon_norm": 8.11876106262207,
    "prediction_norm": 6.422387599945068,
    "timestamp": "2025-07-14T12:34:19.351857"
  },
  {
    "epoch": 1284,
    "batch_idx": 192,
    "loss": 0.255320280790329,
    "learning_rate": 0.0002842719771595206,
    "timestep_mean": 586.25,
    "epsilon_norm": 3.9994585514068604,
    "prediction_norm": 3.2340011596679688,
    "timestamp": "2025-07-14T12:34:19.356703"
  },
  {
    "epoch": 1285,
    "batch_idx": 0,
    "loss": 0.3393864333629608,
    "learning_rate": 0.000283563709239793,
    "timestep_mean": 493.5,
    "epsilon_norm": 7.527224063873291,
    "prediction_norm": 6.704360008239746,
    "timestamp": "2025-07-14T12:34:19.361563"
  },
  {
    "epoch": 1285,
    "batch_idx": 32,
    "loss": 0.22320735454559326,
    "learning_rate": 0.000283563709239793,
    "timestep_mean": 498.0,
    "epsilon_norm": 8.406652450561523,
    "prediction_norm": 7.873098373413086,
    "timestamp": "2025-07-14T12:34:19.366408"
  },
  {
    "epoch": 1285,
    "batch_idx": 64,
    "loss": 0.2902385890483856,
    "learning_rate": 0.000283563709239793,
    "timestep_mean": 511.78125,
    "epsilon_norm": 8.64253044128418,
    "prediction_norm": 7.564460754394531,
    "timestamp": "2025-07-14T12:34:19.371338"
  },
  {
    "epoch": 1285,
    "batch_idx": 96,
    "loss": 0.44684037566185,
    "learning_rate": 0.000283563709239793,
    "timestep_mean": 593.65625,
    "epsilon_norm": 8.371471405029297,
    "prediction_norm": 6.628785133361816,
    "timestamp": "2025-07-14T12:34:19.376196"
  },
  {
    "epoch": 1285,
    "batch_idx": 128,
    "loss": 0.24876391887664795,
    "learning_rate": 0.000283563709239793,
    "timestep_mean": 489.5625,
    "epsilon_norm": 7.856587886810303,
    "prediction_norm": 6.776003360748291,
    "timestamp": "2025-07-14T12:34:19.381039"
  },
  {
    "epoch": 1285,
    "batch_idx": 160,
    "loss": 0.22507578134536743,
    "learning_rate": 0.000283563709239793,
    "timestep_mean": 475.0625,
    "epsilon_norm": 7.507155895233154,
    "prediction_norm": 6.642935752868652,
    "timestamp": "2025-07-14T12:34:19.385940"
  },
  {
    "epoch": 1285,
    "batch_idx": 192,
    "loss": 0.25526463985443115,
    "learning_rate": 0.000283563709239793,
    "timestep_mean": 555.25,
    "epsilon_norm": 3.335221767425537,
    "prediction_norm": 3.2671408653259277,
    "timestamp": "2025-07-14T12:34:19.390771"
  },
  {
    "epoch": 1286,
    "batch_idx": 0,
    "loss": 0.3357922434806824,
    "learning_rate": 0.00028285597535509773,
    "timestep_mean": 466.78125,
    "epsilon_norm": 8.544135093688965,
    "prediction_norm": 7.1208624839782715,
    "timestamp": "2025-07-14T12:34:19.395714"
  },
  {
    "epoch": 1286,
    "batch_idx": 32,
    "loss": 0.4590722918510437,
    "learning_rate": 0.00028285597535509773,
    "timestep_mean": 419.5,
    "epsilon_norm": 8.787473678588867,
    "prediction_norm": 6.279932498931885,
    "timestamp": "2025-07-14T12:34:19.400673"
  },
  {
    "epoch": 1286,
    "batch_idx": 64,
    "loss": 0.3788302540779114,
    "learning_rate": 0.00028285597535509773,
    "timestep_mean": 447.90625,
    "epsilon_norm": 8.173456192016602,
    "prediction_norm": 6.211910724639893,
    "timestamp": "2025-07-14T12:34:19.405716"
  },
  {
    "epoch": 1286,
    "batch_idx": 96,
    "loss": 0.34294265508651733,
    "learning_rate": 0.00028285597535509773,
    "timestep_mean": 577.875,
    "epsilon_norm": 8.752870559692383,
    "prediction_norm": 7.69523811340332,
    "timestamp": "2025-07-14T12:34:19.410658"
  },
  {
    "epoch": 1286,
    "batch_idx": 128,
    "loss": 0.3020526170730591,
    "learning_rate": 0.00028285597535509773,
    "timestep_mean": 432.59375,
    "epsilon_norm": 7.716426849365234,
    "prediction_norm": 6.245588302612305,
    "timestamp": "2025-07-14T12:34:19.415750"
  },
  {
    "epoch": 1286,
    "batch_idx": 160,
    "loss": 0.30176377296447754,
    "learning_rate": 0.00028285597535509773,
    "timestep_mean": 604.15625,
    "epsilon_norm": 8.101886749267578,
    "prediction_norm": 6.489195823669434,
    "timestamp": "2025-07-14T12:34:19.421095"
  },
  {
    "epoch": 1286,
    "batch_idx": 192,
    "loss": 0.7440018653869629,
    "learning_rate": 0.00028285597535509773,
    "timestep_mean": 276.625,
    "epsilon_norm": 4.71759033203125,
    "prediction_norm": 2.760657548904419,
    "timestamp": "2025-07-14T12:34:19.426143"
  },
  {
    "epoch": 1287,
    "batch_idx": 0,
    "loss": 0.3342944383621216,
    "learning_rate": 0.00028214877725169765,
    "timestep_mean": 502.96875,
    "epsilon_norm": 8.081815719604492,
    "prediction_norm": 5.87146520614624,
    "timestamp": "2025-07-14T12:34:19.431334"
  },
  {
    "epoch": 1287,
    "batch_idx": 32,
    "loss": 0.16596759855747223,
    "learning_rate": 0.00028214877725169765,
    "timestep_mean": 498.09375,
    "epsilon_norm": 6.670979022979736,
    "prediction_norm": 6.055863857269287,
    "timestamp": "2025-07-14T12:34:19.436440"
  },
  {
    "epoch": 1287,
    "batch_idx": 64,
    "loss": 0.455323189496994,
    "learning_rate": 0.00028214877725169765,
    "timestep_mean": 502.15625,
    "epsilon_norm": 7.752786636352539,
    "prediction_norm": 5.741289138793945,
    "timestamp": "2025-07-14T12:34:19.441769"
  },
  {
    "epoch": 1287,
    "batch_idx": 96,
    "loss": 0.31466323137283325,
    "learning_rate": 0.00028214877725169765,
    "timestep_mean": 457.90625,
    "epsilon_norm": 9.6138277053833,
    "prediction_norm": 8.393765449523926,
    "timestamp": "2025-07-14T12:34:19.446979"
  },
  {
    "epoch": 1287,
    "batch_idx": 128,
    "loss": 0.4335262179374695,
    "learning_rate": 0.00028214877725169765,
    "timestep_mean": 468.0,
    "epsilon_norm": 7.725468635559082,
    "prediction_norm": 6.159172058105469,
    "timestamp": "2025-07-14T12:34:19.452169"
  },
  {
    "epoch": 1287,
    "batch_idx": 160,
    "loss": 0.5319803953170776,
    "learning_rate": 0.00028214877725169765,
    "timestep_mean": 398.84375,
    "epsilon_norm": 8.359282493591309,
    "prediction_norm": 6.0031352043151855,
    "timestamp": "2025-07-14T12:34:19.457415"
  },
  {
    "epoch": 1287,
    "batch_idx": 192,
    "loss": 0.5281604528427124,
    "learning_rate": 0.00028214877725169765,
    "timestep_mean": 403.5,
    "epsilon_norm": 4.187746047973633,
    "prediction_norm": 2.4502933025360107,
    "timestamp": "2025-07-14T12:34:19.462695"
  },
  {
    "epoch": 1288,
    "batch_idx": 0,
    "loss": 0.38496002554893494,
    "learning_rate": 0.0002814421166745337,
    "timestep_mean": 438.53125,
    "epsilon_norm": 8.062400817871094,
    "prediction_norm": 5.970454692840576,
    "timestamp": "2025-07-14T12:34:19.467905"
  },
  {
    "epoch": 1288,
    "batch_idx": 32,
    "loss": 0.3283082842826843,
    "learning_rate": 0.0002814421166745337,
    "timestep_mean": 481.15625,
    "epsilon_norm": 6.737992286682129,
    "prediction_norm": 5.482985973358154,
    "timestamp": "2025-07-14T12:34:19.473175"
  },
  {
    "epoch": 1288,
    "batch_idx": 64,
    "loss": 0.5490962266921997,
    "learning_rate": 0.0002814421166745337,
    "timestep_mean": 414.03125,
    "epsilon_norm": 8.949014663696289,
    "prediction_norm": 6.080632209777832,
    "timestamp": "2025-07-14T12:34:19.478478"
  },
  {
    "epoch": 1288,
    "batch_idx": 96,
    "loss": 0.32375335693359375,
    "learning_rate": 0.0002814421166745337,
    "timestep_mean": 484.28125,
    "epsilon_norm": 8.894011497497559,
    "prediction_norm": 6.873264312744141,
    "timestamp": "2025-07-14T12:34:19.483551"
  },
  {
    "epoch": 1288,
    "batch_idx": 128,
    "loss": 0.32668882608413696,
    "learning_rate": 0.0002814421166745337,
    "timestep_mean": 544.96875,
    "epsilon_norm": 8.676288604736328,
    "prediction_norm": 7.934695720672607,
    "timestamp": "2025-07-14T12:34:19.488691"
  },
  {
    "epoch": 1288,
    "batch_idx": 160,
    "loss": 0.2568897008895874,
    "learning_rate": 0.0002814421166745337,
    "timestep_mean": 556.34375,
    "epsilon_norm": 8.454257011413574,
    "prediction_norm": 7.217291831970215,
    "timestamp": "2025-07-14T12:34:19.493926"
  },
  {
    "epoch": 1288,
    "batch_idx": 192,
    "loss": 0.06345058232545853,
    "learning_rate": 0.0002814421166745337,
    "timestep_mean": 570.75,
    "epsilon_norm": 3.2180817127227783,
    "prediction_norm": 3.5562992095947266,
    "timestamp": "2025-07-14T12:34:19.498961"
  },
  {
    "epoch": 1289,
    "batch_idx": 0,
    "loss": 0.27131152153015137,
    "learning_rate": 0.00028073599536722067,
    "timestep_mean": 487.71875,
    "epsilon_norm": 7.96094274520874,
    "prediction_norm": 6.841509819030762,
    "timestamp": "2025-07-14T12:34:19.504045"
  },
  {
    "epoch": 1289,
    "batch_idx": 32,
    "loss": 0.20880499482154846,
    "learning_rate": 0.00028073599536722067,
    "timestep_mean": 460.96875,
    "epsilon_norm": 6.60709285736084,
    "prediction_norm": 5.460785388946533,
    "timestamp": "2025-07-14T12:34:19.509233"
  },
  {
    "epoch": 1289,
    "batch_idx": 64,
    "loss": 0.1802206188440323,
    "learning_rate": 0.00028073599536722067,
    "timestep_mean": 591.90625,
    "epsilon_norm": 9.767333984375,
    "prediction_norm": 8.967225074768066,
    "timestamp": "2025-07-14T12:34:19.514369"
  },
  {
    "epoch": 1289,
    "batch_idx": 96,
    "loss": 0.3428652584552765,
    "learning_rate": 0.00028073599536722067,
    "timestep_mean": 490.15625,
    "epsilon_norm": 8.523721694946289,
    "prediction_norm": 7.233412742614746,
    "timestamp": "2025-07-14T12:34:19.519592"
  },
  {
    "epoch": 1289,
    "batch_idx": 128,
    "loss": 0.49189749360084534,
    "learning_rate": 0.00028073599536722067,
    "timestep_mean": 478.28125,
    "epsilon_norm": 8.598089218139648,
    "prediction_norm": 6.892482280731201,
    "timestamp": "2025-07-14T12:34:19.524510"
  },
  {
    "epoch": 1289,
    "batch_idx": 160,
    "loss": 0.18796998262405396,
    "learning_rate": 0.00028073599536722067,
    "timestep_mean": 579.90625,
    "epsilon_norm": 8.472548484802246,
    "prediction_norm": 7.288296222686768,
    "timestamp": "2025-07-14T12:34:19.529673"
  },
  {
    "epoch": 1289,
    "batch_idx": 192,
    "loss": 0.41104137897491455,
    "learning_rate": 0.00028073599536722067,
    "timestep_mean": 351.125,
    "epsilon_norm": 3.881700277328491,
    "prediction_norm": 2.864755153656006,
    "timestamp": "2025-07-14T12:34:19.534759"
  },
  {
    "epoch": 1290,
    "batch_idx": 0,
    "loss": 0.1283099353313446,
    "learning_rate": 0.0002800304150720425,
    "timestep_mean": 521.71875,
    "epsilon_norm": 6.860265731811523,
    "prediction_norm": 6.468762397766113,
    "timestamp": "2025-07-14T12:34:19.539942"
  },
  {
    "epoch": 1290,
    "batch_idx": 32,
    "loss": 0.1846667230129242,
    "learning_rate": 0.0002800304150720425,
    "timestep_mean": 455.21875,
    "epsilon_norm": 7.603492259979248,
    "prediction_norm": 6.118441581726074,
    "timestamp": "2025-07-14T12:34:19.545189"
  },
  {
    "epoch": 1290,
    "batch_idx": 64,
    "loss": 0.2089243233203888,
    "learning_rate": 0.0002800304150720425,
    "timestep_mean": 468.3125,
    "epsilon_norm": 7.841578006744385,
    "prediction_norm": 6.721997261047363,
    "timestamp": "2025-07-14T12:34:19.550535"
  },
  {
    "epoch": 1290,
    "batch_idx": 96,
    "loss": 0.46604594588279724,
    "learning_rate": 0.0002800304150720425,
    "timestep_mean": 429.0,
    "epsilon_norm": 7.803775310516357,
    "prediction_norm": 6.131618022918701,
    "timestamp": "2025-07-14T12:34:19.555697"
  },
  {
    "epoch": 1290,
    "batch_idx": 128,
    "loss": 0.1819499135017395,
    "learning_rate": 0.0002800304150720425,
    "timestep_mean": 530.375,
    "epsilon_norm": 8.408917427062988,
    "prediction_norm": 7.073057174682617,
    "timestamp": "2025-07-14T12:34:19.560916"
  },
  {
    "epoch": 1290,
    "batch_idx": 160,
    "loss": 0.27770933508872986,
    "learning_rate": 0.0002800304150720425,
    "timestep_mean": 495.34375,
    "epsilon_norm": 7.8927903175354,
    "prediction_norm": 6.692587375640869,
    "timestamp": "2025-07-14T12:34:19.566126"
  },
  {
    "epoch": 1290,
    "batch_idx": 192,
    "loss": 0.11759309470653534,
    "learning_rate": 0.0002800304150720425,
    "timestep_mean": 467.0,
    "epsilon_norm": 3.0895919799804688,
    "prediction_norm": 2.686615467071533,
    "timestamp": "2025-07-14T12:34:19.571303"
  },
  {
    "epoch": 1291,
    "batch_idx": 0,
    "loss": 0.24242404103279114,
    "learning_rate": 0.0002793253775299488,
    "timestep_mean": 561.09375,
    "epsilon_norm": 8.282708168029785,
    "prediction_norm": 6.792538166046143,
    "timestamp": "2025-07-14T12:34:19.576402"
  },
  {
    "epoch": 1291,
    "batch_idx": 32,
    "loss": 0.4154197573661804,
    "learning_rate": 0.0002793253775299488,
    "timestep_mean": 376.21875,
    "epsilon_norm": 8.26220703125,
    "prediction_norm": 6.100390911102295,
    "timestamp": "2025-07-14T12:34:19.581589"
  },
  {
    "epoch": 1291,
    "batch_idx": 64,
    "loss": 0.24897401034832,
    "learning_rate": 0.0002793253775299488,
    "timestep_mean": 461.90625,
    "epsilon_norm": 8.959626197814941,
    "prediction_norm": 7.487552642822266,
    "timestamp": "2025-07-14T12:34:19.586794"
  },
  {
    "epoch": 1291,
    "batch_idx": 96,
    "loss": 0.5168029069900513,
    "learning_rate": 0.0002793253775299488,
    "timestep_mean": 485.9375,
    "epsilon_norm": 9.514430046081543,
    "prediction_norm": 7.057271957397461,
    "timestamp": "2025-07-14T12:34:19.592080"
  },
  {
    "epoch": 1291,
    "batch_idx": 128,
    "loss": 0.19277402758598328,
    "learning_rate": 0.0002793253775299488,
    "timestep_mean": 567.4375,
    "epsilon_norm": 7.003057479858398,
    "prediction_norm": 6.45371150970459,
    "timestamp": "2025-07-14T12:34:19.597237"
  },
  {
    "epoch": 1291,
    "batch_idx": 160,
    "loss": 0.5708463191986084,
    "learning_rate": 0.0002793253775299488,
    "timestep_mean": 466.125,
    "epsilon_norm": 9.476037979125977,
    "prediction_norm": 7.771968364715576,
    "timestamp": "2025-07-14T12:34:19.602405"
  },
  {
    "epoch": 1291,
    "batch_idx": 192,
    "loss": 0.23620493710041046,
    "learning_rate": 0.0002793253775299488,
    "timestep_mean": 548.875,
    "epsilon_norm": 4.062311172485352,
    "prediction_norm": 3.5971519947052,
    "timestamp": "2025-07-14T12:34:19.607584"
  },
  {
    "epoch": 1292,
    "batch_idx": 0,
    "loss": 0.7272096276283264,
    "learning_rate": 0.00027862088448054944,
    "timestep_mean": 499.8125,
    "epsilon_norm": 9.620537757873535,
    "prediction_norm": 7.606151580810547,
    "timestamp": "2025-07-14T12:34:19.612810"
  },
  {
    "epoch": 1292,
    "batch_idx": 32,
    "loss": 0.4280468225479126,
    "learning_rate": 0.00027862088448054944,
    "timestep_mean": 419.875,
    "epsilon_norm": 7.8045783042907715,
    "prediction_norm": 6.106979846954346,
    "timestamp": "2025-07-14T12:34:19.618018"
  },
  {
    "epoch": 1292,
    "batch_idx": 64,
    "loss": 0.20690768957138062,
    "learning_rate": 0.00027862088448054944,
    "timestep_mean": 503.21875,
    "epsilon_norm": 8.34716796875,
    "prediction_norm": 8.06663703918457,
    "timestamp": "2025-07-14T12:34:19.623180"
  },
  {
    "epoch": 1292,
    "batch_idx": 96,
    "loss": 0.15432360768318176,
    "learning_rate": 0.00027862088448054944,
    "timestep_mean": 498.09375,
    "epsilon_norm": 7.372366905212402,
    "prediction_norm": 6.944571018218994,
    "timestamp": "2025-07-14T12:34:19.628401"
  },
  {
    "epoch": 1292,
    "batch_idx": 128,
    "loss": 0.15632127225399017,
    "learning_rate": 0.00027862088448054944,
    "timestep_mean": 522.3125,
    "epsilon_norm": 8.086870193481445,
    "prediction_norm": 7.730438709259033,
    "timestamp": "2025-07-14T12:34:19.633610"
  },
  {
    "epoch": 1292,
    "batch_idx": 160,
    "loss": 0.18785201013088226,
    "learning_rate": 0.00027862088448054944,
    "timestep_mean": 585.78125,
    "epsilon_norm": 7.797857284545898,
    "prediction_norm": 7.383577823638916,
    "timestamp": "2025-07-14T12:34:19.638813"
  },
  {
    "epoch": 1292,
    "batch_idx": 192,
    "loss": 0.22483041882514954,
    "learning_rate": 0.00027862088448054944,
    "timestep_mean": 560.375,
    "epsilon_norm": 5.073925971984863,
    "prediction_norm": 4.360846042633057,
    "timestamp": "2025-07-14T12:34:19.644039"
  },
  {
    "epoch": 1293,
    "batch_idx": 0,
    "loss": 0.10952159762382507,
    "learning_rate": 0.00027791693766211073,
    "timestep_mean": 562.53125,
    "epsilon_norm": 7.696695327758789,
    "prediction_norm": 7.6406731605529785,
    "timestamp": "2025-07-14T12:34:19.649210"
  },
  {
    "epoch": 1293,
    "batch_idx": 32,
    "loss": 0.6077982783317566,
    "learning_rate": 0.00027791693766211073,
    "timestep_mean": 433.375,
    "epsilon_norm": 7.331579208374023,
    "prediction_norm": 5.17529296875,
    "timestamp": "2025-07-14T12:34:19.654451"
  },
  {
    "epoch": 1293,
    "batch_idx": 64,
    "loss": 0.442198246717453,
    "learning_rate": 0.00027791693766211073,
    "timestep_mean": 447.03125,
    "epsilon_norm": 8.15780258178711,
    "prediction_norm": 6.859897613525391,
    "timestamp": "2025-07-14T12:34:19.659570"
  },
  {
    "epoch": 1293,
    "batch_idx": 96,
    "loss": 0.19190767407417297,
    "learning_rate": 0.00027791693766211073,
    "timestep_mean": 464.0,
    "epsilon_norm": 7.404526710510254,
    "prediction_norm": 6.610100269317627,
    "timestamp": "2025-07-14T12:34:19.664811"
  },
  {
    "epoch": 1293,
    "batch_idx": 128,
    "loss": 0.31299078464508057,
    "learning_rate": 0.00027791693766211073,
    "timestep_mean": 524.3125,
    "epsilon_norm": 8.414796829223633,
    "prediction_norm": 7.134308338165283,
    "timestamp": "2025-07-14T12:34:19.670042"
  },
  {
    "epoch": 1293,
    "batch_idx": 160,
    "loss": 0.34305375814437866,
    "learning_rate": 0.00027791693766211073,
    "timestep_mean": 534.53125,
    "epsilon_norm": 8.030874252319336,
    "prediction_norm": 5.997251510620117,
    "timestamp": "2025-07-14T12:34:19.675291"
  },
  {
    "epoch": 1293,
    "batch_idx": 192,
    "loss": 0.4092913269996643,
    "learning_rate": 0.00027791693766211073,
    "timestep_mean": 372.25,
    "epsilon_norm": 3.9005839824676514,
    "prediction_norm": 2.710834503173828,
    "timestamp": "2025-07-14T12:34:19.680501"
  },
  {
    "epoch": 1294,
    "batch_idx": 0,
    "loss": 0.19172939658164978,
    "learning_rate": 0.00027721353881155197,
    "timestep_mean": 515.21875,
    "epsilon_norm": 6.876873016357422,
    "prediction_norm": 6.05034065246582,
    "timestamp": "2025-07-14T12:34:19.685713"
  },
  {
    "epoch": 1294,
    "batch_idx": 32,
    "loss": 0.2290688157081604,
    "learning_rate": 0.00027721353881155197,
    "timestep_mean": 451.65625,
    "epsilon_norm": 8.39002513885498,
    "prediction_norm": 6.776942253112793,
    "timestamp": "2025-07-14T12:34:19.690894"
  },
  {
    "epoch": 1294,
    "batch_idx": 64,
    "loss": 0.30435189604759216,
    "learning_rate": 0.00027721353881155197,
    "timestep_mean": 469.6875,
    "epsilon_norm": 8.102242469787598,
    "prediction_norm": 6.799222946166992,
    "timestamp": "2025-07-14T12:34:19.696054"
  },
  {
    "epoch": 1294,
    "batch_idx": 96,
    "loss": 0.2746937870979309,
    "learning_rate": 0.00027721353881155197,
    "timestep_mean": 505.5625,
    "epsilon_norm": 7.476888179779053,
    "prediction_norm": 6.360791206359863,
    "timestamp": "2025-07-14T12:34:19.701188"
  },
  {
    "epoch": 1294,
    "batch_idx": 128,
    "loss": 0.1383441835641861,
    "learning_rate": 0.00027721353881155197,
    "timestep_mean": 544.71875,
    "epsilon_norm": 7.593631744384766,
    "prediction_norm": 6.982031345367432,
    "timestamp": "2025-07-14T12:34:19.706361"
  },
  {
    "epoch": 1294,
    "batch_idx": 160,
    "loss": 0.39741063117980957,
    "learning_rate": 0.00027721353881155197,
    "timestep_mean": 430.9375,
    "epsilon_norm": 8.443465232849121,
    "prediction_norm": 6.897931098937988,
    "timestamp": "2025-07-14T12:34:19.711643"
  },
  {
    "epoch": 1294,
    "batch_idx": 192,
    "loss": 0.06715549528598785,
    "learning_rate": 0.00027721353881155197,
    "timestep_mean": 574.375,
    "epsilon_norm": 4.338771343231201,
    "prediction_norm": 4.105537414550781,
    "timestamp": "2025-07-14T12:34:19.716931"
  },
  {
    "epoch": 1295,
    "batch_idx": 0,
    "loss": 0.5020134449005127,
    "learning_rate": 0.0002765106896644396,
    "timestep_mean": 432.28125,
    "epsilon_norm": 8.967463493347168,
    "prediction_norm": 6.458564281463623,
    "timestamp": "2025-07-14T12:34:19.722159"
  },
  {
    "epoch": 1295,
    "batch_idx": 32,
    "loss": 0.257138729095459,
    "learning_rate": 0.0002765106896644396,
    "timestep_mean": 499.84375,
    "epsilon_norm": 6.918747425079346,
    "prediction_norm": 5.832515716552734,
    "timestamp": "2025-07-14T12:34:19.727363"
  },
  {
    "epoch": 1295,
    "batch_idx": 64,
    "loss": 0.2330911010503769,
    "learning_rate": 0.0002765106896644396,
    "timestep_mean": 531.90625,
    "epsilon_norm": 8.861138343811035,
    "prediction_norm": 7.862828254699707,
    "timestamp": "2025-07-14T12:34:19.732538"
  },
  {
    "epoch": 1295,
    "batch_idx": 96,
    "loss": 0.2469523847103119,
    "learning_rate": 0.0002765106896644396,
    "timestep_mean": 489.71875,
    "epsilon_norm": 7.471437454223633,
    "prediction_norm": 6.2906670570373535,
    "timestamp": "2025-07-14T12:34:19.737801"
  },
  {
    "epoch": 1295,
    "batch_idx": 128,
    "loss": 0.36143165826797485,
    "learning_rate": 0.0002765106896644396,
    "timestep_mean": 483.4375,
    "epsilon_norm": 8.088468551635742,
    "prediction_norm": 6.391297817230225,
    "timestamp": "2025-07-14T12:34:19.742886"
  },
  {
    "epoch": 1295,
    "batch_idx": 160,
    "loss": 0.25657176971435547,
    "learning_rate": 0.0002765106896644396,
    "timestep_mean": 484.75,
    "epsilon_norm": 6.73724889755249,
    "prediction_norm": 5.593013763427734,
    "timestamp": "2025-07-14T12:34:19.748143"
  },
  {
    "epoch": 1295,
    "batch_idx": 192,
    "loss": 0.16997259855270386,
    "learning_rate": 0.0002765106896644396,
    "timestep_mean": 429.625,
    "epsilon_norm": 3.170330762863159,
    "prediction_norm": 2.395406723022461,
    "timestamp": "2025-07-14T12:34:19.753315"
  },
  {
    "epoch": 1296,
    "batch_idx": 0,
    "loss": 0.39723435044288635,
    "learning_rate": 0.0002758083919549841,
    "timestep_mean": 463.5,
    "epsilon_norm": 9.231074333190918,
    "prediction_norm": 7.409163475036621,
    "timestamp": "2025-07-14T12:34:19.758633"
  },
  {
    "epoch": 1296,
    "batch_idx": 32,
    "loss": 0.3189772665500641,
    "learning_rate": 0.0002758083919549841,
    "timestep_mean": 589.1875,
    "epsilon_norm": 7.869465351104736,
    "prediction_norm": 6.796083927154541,
    "timestamp": "2025-07-14T12:34:19.763825"
  },
  {
    "epoch": 1296,
    "batch_idx": 64,
    "loss": 0.283236026763916,
    "learning_rate": 0.0002758083919549841,
    "timestep_mean": 396.5,
    "epsilon_norm": 8.042193412780762,
    "prediction_norm": 6.9048967361450195,
    "timestamp": "2025-07-14T12:34:19.768831"
  },
  {
    "epoch": 1296,
    "batch_idx": 96,
    "loss": 0.29594627022743225,
    "learning_rate": 0.0002758083919549841,
    "timestep_mean": 457.9375,
    "epsilon_norm": 7.240059852600098,
    "prediction_norm": 6.3068952560424805,
    "timestamp": "2025-07-14T12:34:19.774079"
  },
  {
    "epoch": 1296,
    "batch_idx": 128,
    "loss": 0.20599417388439178,
    "learning_rate": 0.0002758083919549841,
    "timestep_mean": 442.5,
    "epsilon_norm": 8.444392204284668,
    "prediction_norm": 7.656663417816162,
    "timestamp": "2025-07-14T12:34:19.779305"
  },
  {
    "epoch": 1296,
    "batch_idx": 160,
    "loss": 0.2542042136192322,
    "learning_rate": 0.0002758083919549841,
    "timestep_mean": 534.40625,
    "epsilon_norm": 7.48549747467041,
    "prediction_norm": 6.467850208282471,
    "timestamp": "2025-07-14T12:34:19.784261"
  },
  {
    "epoch": 1296,
    "batch_idx": 192,
    "loss": 0.40501201152801514,
    "learning_rate": 0.0002758083919549841,
    "timestep_mean": 586.0,
    "epsilon_norm": 4.314596176147461,
    "prediction_norm": 3.7629456520080566,
    "timestamp": "2025-07-14T12:34:19.789111"
  },
  {
    "epoch": 1297,
    "batch_idx": 0,
    "loss": 0.3834916949272156,
    "learning_rate": 0.000275106647416035,
    "timestep_mean": 456.9375,
    "epsilon_norm": 7.9530816078186035,
    "prediction_norm": 6.1613616943359375,
    "timestamp": "2025-07-14T12:34:19.794008"
  },
  {
    "epoch": 1297,
    "batch_idx": 32,
    "loss": 0.16630089282989502,
    "learning_rate": 0.000275106647416035,
    "timestep_mean": 508.28125,
    "epsilon_norm": 6.894997596740723,
    "prediction_norm": 6.382419109344482,
    "timestamp": "2025-07-14T12:34:19.798849"
  },
  {
    "epoch": 1297,
    "batch_idx": 64,
    "loss": 0.2735781669616699,
    "learning_rate": 0.000275106647416035,
    "timestep_mean": 451.625,
    "epsilon_norm": 7.410854339599609,
    "prediction_norm": 6.080483913421631,
    "timestamp": "2025-07-14T12:34:19.803700"
  },
  {
    "epoch": 1297,
    "batch_idx": 96,
    "loss": 0.35317787528038025,
    "learning_rate": 0.000275106647416035,
    "timestep_mean": 495.125,
    "epsilon_norm": 8.522276878356934,
    "prediction_norm": 6.700597286224365,
    "timestamp": "2025-07-14T12:34:19.808703"
  },
  {
    "epoch": 1297,
    "batch_idx": 128,
    "loss": 0.32334452867507935,
    "learning_rate": 0.000275106647416035,
    "timestep_mean": 546.78125,
    "epsilon_norm": 8.034444808959961,
    "prediction_norm": 6.459802150726318,
    "timestamp": "2025-07-14T12:34:19.813549"
  },
  {
    "epoch": 1297,
    "batch_idx": 160,
    "loss": 0.30627477169036865,
    "learning_rate": 0.000275106647416035,
    "timestep_mean": 386.3125,
    "epsilon_norm": 7.945338726043701,
    "prediction_norm": 5.614927768707275,
    "timestamp": "2025-07-14T12:34:19.818628"
  },
  {
    "epoch": 1297,
    "batch_idx": 192,
    "loss": 0.05109003931283951,
    "learning_rate": 0.000275106647416035,
    "timestep_mean": 645.875,
    "epsilon_norm": 4.368518352508545,
    "prediction_norm": 3.9473793506622314,
    "timestamp": "2025-07-14T12:34:19.823717"
  },
  {
    "epoch": 1298,
    "batch_idx": 0,
    "loss": 0.1554420441389084,
    "learning_rate": 0.0002744054577790776,
    "timestep_mean": 570.75,
    "epsilon_norm": 6.400768280029297,
    "prediction_norm": 5.761695384979248,
    "timestamp": "2025-07-14T12:34:19.828918"
  },
  {
    "epoch": 1298,
    "batch_idx": 32,
    "loss": 0.22727014124393463,
    "learning_rate": 0.0002744054577790776,
    "timestep_mean": 456.40625,
    "epsilon_norm": 7.761518478393555,
    "prediction_norm": 6.851110458374023,
    "timestamp": "2025-07-14T12:34:19.833999"
  },
  {
    "epoch": 1298,
    "batch_idx": 64,
    "loss": 0.38808339834213257,
    "learning_rate": 0.0002744054577790776,
    "timestep_mean": 552.09375,
    "epsilon_norm": 7.865307331085205,
    "prediction_norm": 6.3960981369018555,
    "timestamp": "2025-07-14T12:34:19.839159"
  },
  {
    "epoch": 1298,
    "batch_idx": 96,
    "loss": 0.36217203736305237,
    "learning_rate": 0.0002744054577790776,
    "timestep_mean": 468.9375,
    "epsilon_norm": 8.269099235534668,
    "prediction_norm": 6.051119804382324,
    "timestamp": "2025-07-14T12:34:19.844348"
  },
  {
    "epoch": 1298,
    "batch_idx": 128,
    "loss": 0.23355379700660706,
    "learning_rate": 0.0002744054577790776,
    "timestep_mean": 528.9375,
    "epsilon_norm": 8.077615737915039,
    "prediction_norm": 7.654361248016357,
    "timestamp": "2025-07-14T12:34:19.849515"
  },
  {
    "epoch": 1298,
    "batch_idx": 160,
    "loss": 0.19912943243980408,
    "learning_rate": 0.0002744054577790776,
    "timestep_mean": 632.1875,
    "epsilon_norm": 7.3809099197387695,
    "prediction_norm": 6.706768989562988,
    "timestamp": "2025-07-14T12:34:19.854613"
  },
  {
    "epoch": 1298,
    "batch_idx": 192,
    "loss": 0.05840844660997391,
    "learning_rate": 0.0002744054577790776,
    "timestep_mean": 601.0,
    "epsilon_norm": 3.7158560752868652,
    "prediction_norm": 3.997976541519165,
    "timestamp": "2025-07-14T12:34:19.859657"
  },
  {
    "epoch": 1299,
    "batch_idx": 0,
    "loss": 0.3209775686264038,
    "learning_rate": 0.0002737048247742274,
    "timestep_mean": 570.5625,
    "epsilon_norm": 8.952799797058105,
    "prediction_norm": 7.641357421875,
    "timestamp": "2025-07-14T12:34:19.864775"
  },
  {
    "epoch": 1299,
    "batch_idx": 32,
    "loss": 0.21633639931678772,
    "learning_rate": 0.0002737048247742274,
    "timestep_mean": 417.09375,
    "epsilon_norm": 8.253569602966309,
    "prediction_norm": 7.08641242980957,
    "timestamp": "2025-07-14T12:34:19.869987"
  },
  {
    "epoch": 1299,
    "batch_idx": 64,
    "loss": 0.4069772958755493,
    "learning_rate": 0.0002737048247742274,
    "timestep_mean": 501.25,
    "epsilon_norm": 9.591198921203613,
    "prediction_norm": 7.5857038497924805,
    "timestamp": "2025-07-14T12:34:19.875271"
  },
  {
    "epoch": 1299,
    "batch_idx": 96,
    "loss": 0.2960548996925354,
    "learning_rate": 0.0002737048247742274,
    "timestep_mean": 547.28125,
    "epsilon_norm": 7.801754474639893,
    "prediction_norm": 6.699033260345459,
    "timestamp": "2025-07-14T12:34:19.880421"
  },
  {
    "epoch": 1299,
    "batch_idx": 128,
    "loss": 0.28981655836105347,
    "learning_rate": 0.0002737048247742274,
    "timestep_mean": 577.25,
    "epsilon_norm": 8.40701675415039,
    "prediction_norm": 6.800857067108154,
    "timestamp": "2025-07-14T12:34:19.885632"
  },
  {
    "epoch": 1299,
    "batch_idx": 160,
    "loss": 0.2706877291202545,
    "learning_rate": 0.0002737048247742274,
    "timestep_mean": 512.0625,
    "epsilon_norm": 8.2406005859375,
    "prediction_norm": 6.957120418548584,
    "timestamp": "2025-07-14T12:34:19.890898"
  },
  {
    "epoch": 1299,
    "batch_idx": 192,
    "loss": 0.20988817512989044,
    "learning_rate": 0.0002737048247742274,
    "timestep_mean": 462.5,
    "epsilon_norm": 3.0717508792877197,
    "prediction_norm": 2.377511739730835,
    "timestamp": "2025-07-14T12:34:19.896056"
  },
  {
    "epoch": 1300,
    "batch_idx": 0,
    "loss": 0.15524521470069885,
    "learning_rate": 0.00027300475013022674,
    "timestep_mean": 520.84375,
    "epsilon_norm": 7.738590717315674,
    "prediction_norm": 7.18744421005249,
    "timestamp": "2025-07-14T12:34:19.901236"
  },
  {
    "epoch": 1300,
    "batch_idx": 32,
    "loss": 0.3185858726501465,
    "learning_rate": 0.00027300475013022674,
    "timestep_mean": 575.25,
    "epsilon_norm": 7.8481855392456055,
    "prediction_norm": 6.881155490875244,
    "timestamp": "2025-07-14T12:34:19.906403"
  },
  {
    "epoch": 1300,
    "batch_idx": 64,
    "loss": 0.3822975754737854,
    "learning_rate": 0.00027300475013022674,
    "timestep_mean": 529.21875,
    "epsilon_norm": 9.346939086914062,
    "prediction_norm": 7.424226760864258,
    "timestamp": "2025-07-14T12:34:19.911678"
  },
  {
    "epoch": 1300,
    "batch_idx": 96,
    "loss": 0.285228431224823,
    "learning_rate": 0.00027300475013022674,
    "timestep_mean": 521.375,
    "epsilon_norm": 8.306909561157227,
    "prediction_norm": 7.257727146148682,
    "timestamp": "2025-07-14T12:34:19.917007"
  },
  {
    "epoch": 1300,
    "batch_idx": 128,
    "loss": 0.18534569442272186,
    "learning_rate": 0.00027300475013022674,
    "timestep_mean": 541.125,
    "epsilon_norm": 7.768795490264893,
    "prediction_norm": 6.444599628448486,
    "timestamp": "2025-07-14T12:34:19.922148"
  },
  {
    "epoch": 1300,
    "batch_idx": 160,
    "loss": 0.24324889481067657,
    "learning_rate": 0.00027300475013022674,
    "timestep_mean": 505.15625,
    "epsilon_norm": 8.074418067932129,
    "prediction_norm": 7.011368274688721,
    "timestamp": "2025-07-14T12:34:19.927307"
  },
  {
    "epoch": 1300,
    "batch_idx": 192,
    "loss": 0.2383800894021988,
    "learning_rate": 0.00027300475013022674,
    "timestep_mean": 370.5,
    "epsilon_norm": 3.5345675945281982,
    "prediction_norm": 3.171257734298706,
    "timestamp": "2025-07-14T12:34:19.932242"
  },
  {
    "epoch": 1301,
    "batch_idx": 0,
    "loss": 0.4714195430278778,
    "learning_rate": 0.00027230523557444025,
    "timestep_mean": 550.8125,
    "epsilon_norm": 8.374383926391602,
    "prediction_norm": 6.85506534576416,
    "timestamp": "2025-07-14T12:34:19.937120"
  },
  {
    "epoch": 1301,
    "batch_idx": 32,
    "loss": 0.40409427881240845,
    "learning_rate": 0.00027230523557444025,
    "timestep_mean": 517.5625,
    "epsilon_norm": 8.514924049377441,
    "prediction_norm": 6.990505218505859,
    "timestamp": "2025-07-14T12:34:19.942202"
  },
  {
    "epoch": 1301,
    "batch_idx": 64,
    "loss": 0.5757935047149658,
    "learning_rate": 0.00027230523557444025,
    "timestep_mean": 446.8125,
    "epsilon_norm": 8.788994789123535,
    "prediction_norm": 6.302082061767578,
    "timestamp": "2025-07-14T12:34:19.947294"
  },
  {
    "epoch": 1301,
    "batch_idx": 96,
    "loss": 0.24759399890899658,
    "learning_rate": 0.00027230523557444025,
    "timestep_mean": 543.6875,
    "epsilon_norm": 7.61575174331665,
    "prediction_norm": 6.167485237121582,
    "timestamp": "2025-07-14T12:34:19.952322"
  },
  {
    "epoch": 1301,
    "batch_idx": 128,
    "loss": 0.20558574795722961,
    "learning_rate": 0.00027230523557444025,
    "timestep_mean": 482.59375,
    "epsilon_norm": 7.494981288909912,
    "prediction_norm": 6.753150939941406,
    "timestamp": "2025-07-14T12:34:19.957482"
  },
  {
    "epoch": 1301,
    "batch_idx": 160,
    "loss": 0.11830106377601624,
    "learning_rate": 0.00027230523557444025,
    "timestep_mean": 541.9375,
    "epsilon_norm": 6.6211700439453125,
    "prediction_norm": 6.508511543273926,
    "timestamp": "2025-07-14T12:34:19.962640"
  },
  {
    "epoch": 1301,
    "batch_idx": 192,
    "loss": 0.10788273811340332,
    "learning_rate": 0.00027230523557444025,
    "timestep_mean": 400.375,
    "epsilon_norm": 3.7999420166015625,
    "prediction_norm": 3.5916459560394287,
    "timestamp": "2025-07-14T12:34:19.967732"
  },
  {
    "epoch": 1302,
    "batch_idx": 0,
    "loss": 0.2538379430770874,
    "learning_rate": 0.0002716062828328505,
    "timestep_mean": 531.875,
    "epsilon_norm": 7.28935432434082,
    "prediction_norm": 5.922627925872803,
    "timestamp": "2025-07-14T12:34:19.972898"
  },
  {
    "epoch": 1302,
    "batch_idx": 32,
    "loss": 0.3240188658237457,
    "learning_rate": 0.0002716062828328505,
    "timestep_mean": 465.78125,
    "epsilon_norm": 8.219582557678223,
    "prediction_norm": 6.732224464416504,
    "timestamp": "2025-07-14T12:34:19.978077"
  },
  {
    "epoch": 1302,
    "batch_idx": 64,
    "loss": 0.39057600498199463,
    "learning_rate": 0.0002716062828328505,
    "timestep_mean": 407.875,
    "epsilon_norm": 7.96645975112915,
    "prediction_norm": 5.920021057128906,
    "timestamp": "2025-07-14T12:34:19.983373"
  },
  {
    "epoch": 1302,
    "batch_idx": 96,
    "loss": 0.5044758915901184,
    "learning_rate": 0.0002716062828328505,
    "timestep_mean": 522.3125,
    "epsilon_norm": 8.354888916015625,
    "prediction_norm": 7.211969375610352,
    "timestamp": "2025-07-14T12:34:19.988512"
  },
  {
    "epoch": 1302,
    "batch_idx": 128,
    "loss": 0.3076002597808838,
    "learning_rate": 0.0002716062828328505,
    "timestep_mean": 460.5,
    "epsilon_norm": 8.025721549987793,
    "prediction_norm": 6.20849084854126,
    "timestamp": "2025-07-14T12:34:19.993690"
  },
  {
    "epoch": 1302,
    "batch_idx": 160,
    "loss": 0.1300419121980667,
    "learning_rate": 0.0002716062828328505,
    "timestep_mean": 603.28125,
    "epsilon_norm": 8.405031204223633,
    "prediction_norm": 7.94853401184082,
    "timestamp": "2025-07-14T12:34:19.998603"
  },
  {
    "epoch": 1302,
    "batch_idx": 192,
    "loss": 0.32043880224227905,
    "learning_rate": 0.0002716062828328505,
    "timestep_mean": 410.375,
    "epsilon_norm": 4.210258960723877,
    "prediction_norm": 3.123508930206299,
    "timestamp": "2025-07-14T12:34:20.003436"
  },
  {
    "epoch": 1303,
    "batch_idx": 0,
    "loss": 0.34565120935440063,
    "learning_rate": 0.0002709078936300537,
    "timestep_mean": 509.4375,
    "epsilon_norm": 9.776256561279297,
    "prediction_norm": 7.9103593826293945,
    "timestamp": "2025-07-14T12:34:20.008378"
  },
  {
    "epoch": 1303,
    "batch_idx": 32,
    "loss": 0.31213539838790894,
    "learning_rate": 0.0002709078936300537,
    "timestep_mean": 499.5625,
    "epsilon_norm": 8.026834487915039,
    "prediction_norm": 6.918800354003906,
    "timestamp": "2025-07-14T12:34:20.013211"
  },
  {
    "epoch": 1303,
    "batch_idx": 64,
    "loss": 0.33963772654533386,
    "learning_rate": 0.0002709078936300537,
    "timestep_mean": 644.375,
    "epsilon_norm": 7.780735015869141,
    "prediction_norm": 6.361093521118164,
    "timestamp": "2025-07-14T12:34:20.018052"
  },
  {
    "epoch": 1303,
    "batch_idx": 96,
    "loss": 0.2647596001625061,
    "learning_rate": 0.0002709078936300537,
    "timestep_mean": 530.875,
    "epsilon_norm": 8.395774841308594,
    "prediction_norm": 7.2627034187316895,
    "timestamp": "2025-07-14T12:34:20.022863"
  },
  {
    "epoch": 1303,
    "batch_idx": 128,
    "loss": 0.18344157934188843,
    "learning_rate": 0.0002709078936300537,
    "timestep_mean": 477.90625,
    "epsilon_norm": 6.9640421867370605,
    "prediction_norm": 6.228901386260986,
    "timestamp": "2025-07-14T12:34:20.027751"
  },
  {
    "epoch": 1303,
    "batch_idx": 160,
    "loss": 0.337178111076355,
    "learning_rate": 0.0002709078936300537,
    "timestep_mean": 507.03125,
    "epsilon_norm": 7.617554664611816,
    "prediction_norm": 5.449497222900391,
    "timestamp": "2025-07-14T12:34:20.032604"
  },
  {
    "epoch": 1303,
    "batch_idx": 192,
    "loss": 0.23219211399555206,
    "learning_rate": 0.0002709078936300537,
    "timestep_mean": 537.625,
    "epsilon_norm": 5.35388708114624,
    "prediction_norm": 4.806941509246826,
    "timestamp": "2025-07-14T12:34:20.037567"
  },
  {
    "epoch": 1304,
    "batch_idx": 0,
    "loss": 0.3209211230278015,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 490.09375,
    "epsilon_norm": 7.981008052825928,
    "prediction_norm": 7.264483451843262,
    "timestamp": "2025-07-14T12:34:20.042499"
  },
  {
    "epoch": 1304,
    "batch_idx": 32,
    "loss": 0.3008514940738678,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 523.625,
    "epsilon_norm": 8.198346138000488,
    "prediction_norm": 7.218488693237305,
    "timestamp": "2025-07-14T12:34:20.047346"
  },
  {
    "epoch": 1304,
    "batch_idx": 64,
    "loss": 0.2741759121417999,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 591.1875,
    "epsilon_norm": 8.074929237365723,
    "prediction_norm": 7.411504745483398,
    "timestamp": "2025-07-14T12:34:20.052233"
  },
  {
    "epoch": 1304,
    "batch_idx": 96,
    "loss": 0.39470189809799194,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 410.96875,
    "epsilon_norm": 8.803352355957031,
    "prediction_norm": 6.845947265625,
    "timestamp": "2025-07-14T12:34:20.057153"
  },
  {
    "epoch": 1304,
    "batch_idx": 128,
    "loss": 0.20030902326107025,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 461.78125,
    "epsilon_norm": 8.008054733276367,
    "prediction_norm": 7.239896297454834,
    "timestamp": "2025-07-14T12:34:20.061969"
  },
  {
    "epoch": 1304,
    "batch_idx": 160,
    "loss": 0.3056011199951172,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 501.40625,
    "epsilon_norm": 7.221263408660889,
    "prediction_norm": 5.9534525871276855,
    "timestamp": "2025-07-14T12:34:20.066891"
  },
  {
    "epoch": 1304,
    "batch_idx": 192,
    "loss": 0.1808602213859558,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 371.75,
    "epsilon_norm": 2.678971290588379,
    "prediction_norm": 2.340357780456543,
    "timestamp": "2025-07-14T12:34:20.071683"
  },
  {
    "epoch": 1305,
    "batch_idx": 0,
    "loss": 0.2521201968193054,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 507.84375,
    "epsilon_norm": 7.699542045593262,
    "prediction_norm": 6.769864082336426,
    "timestamp": "2025-07-14T12:34:20.076622"
  },
  {
    "epoch": 1305,
    "batch_idx": 32,
    "loss": 0.5121431350708008,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 392.71875,
    "epsilon_norm": 7.378565311431885,
    "prediction_norm": 6.062551021575928,
    "timestamp": "2025-07-14T12:34:20.081447"
  },
  {
    "epoch": 1305,
    "batch_idx": 64,
    "loss": 0.21830147504806519,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 489.40625,
    "epsilon_norm": 6.72113037109375,
    "prediction_norm": 5.74629545211792,
    "timestamp": "2025-07-14T12:34:20.086385"
  },
  {
    "epoch": 1305,
    "batch_idx": 96,
    "loss": 0.30279433727264404,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 483.3125,
    "epsilon_norm": 7.9315619468688965,
    "prediction_norm": 5.908528804779053,
    "timestamp": "2025-07-14T12:34:20.091208"
  },
  {
    "epoch": 1305,
    "batch_idx": 128,
    "loss": 0.29067230224609375,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 464.90625,
    "epsilon_norm": 8.477105140686035,
    "prediction_norm": 7.130499839782715,
    "timestamp": "2025-07-14T12:34:20.096081"
  },
  {
    "epoch": 1305,
    "batch_idx": 160,
    "loss": 0.2805114686489105,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 484.09375,
    "epsilon_norm": 7.786065578460693,
    "prediction_norm": 6.65680456161499,
    "timestamp": "2025-07-14T12:34:20.100937"
  },
  {
    "epoch": 1305,
    "batch_idx": 192,
    "loss": 0.3174954950809479,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 603.875,
    "epsilon_norm": 4.276848793029785,
    "prediction_norm": 3.1962969303131104,
    "timestamp": "2025-07-14T12:34:20.105812"
  },
  {
    "epoch": 1306,
    "batch_idx": 0,
    "loss": 0.34373241662979126,
    "learning_rate": 0.0002688161244795042,
    "timestep_mean": 491.40625,
    "epsilon_norm": 8.370532035827637,
    "prediction_norm": 6.513157844543457,
    "timestamp": "2025-07-14T12:34:20.110689"
  },
  {
    "epoch": 1306,
    "batch_idx": 32,
    "loss": 0.23520609736442566,
    "learning_rate": 0.0002688161244795042,
    "timestep_mean": 487.75,
    "epsilon_norm": 7.13939094543457,
    "prediction_norm": 6.789875507354736,
    "timestamp": "2025-07-14T12:34:20.115562"
  },
  {
    "epoch": 1306,
    "batch_idx": 64,
    "loss": 0.47278326749801636,
    "learning_rate": 0.0002688161244795042,
    "timestep_mean": 535.40625,
    "epsilon_norm": 8.927895545959473,
    "prediction_norm": 6.600623607635498,
    "timestamp": "2025-07-14T12:34:20.120425"
  },
  {
    "epoch": 1306,
    "batch_idx": 96,
    "loss": 0.19087593257427216,
    "learning_rate": 0.0002688161244795042,
    "timestep_mean": 523.0,
    "epsilon_norm": 8.291035652160645,
    "prediction_norm": 7.408330917358398,
    "timestamp": "2025-07-14T12:34:20.125361"
  },
  {
    "epoch": 1306,
    "batch_idx": 128,
    "loss": 0.15797671675682068,
    "learning_rate": 0.0002688161244795042,
    "timestep_mean": 522.96875,
    "epsilon_norm": 8.747064590454102,
    "prediction_norm": 8.447188377380371,
    "timestamp": "2025-07-14T12:34:20.130230"
  },
  {
    "epoch": 1306,
    "batch_idx": 160,
    "loss": 0.09346846491098404,
    "learning_rate": 0.0002688161244795042,
    "timestep_mean": 583.34375,
    "epsilon_norm": 8.080196380615234,
    "prediction_norm": 7.530154228210449,
    "timestamp": "2025-07-14T12:34:20.135147"
  },
  {
    "epoch": 1306,
    "batch_idx": 192,
    "loss": 0.04612646624445915,
    "learning_rate": 0.0002688161244795042,
    "timestep_mean": 519.25,
    "epsilon_norm": 1.783307671546936,
    "prediction_norm": 1.8451753854751587,
    "timestamp": "2025-07-14T12:34:20.139963"
  },
  {
    "epoch": 1307,
    "batch_idx": 0,
    "loss": 0.25703418254852295,
    "learning_rate": 0.00026812000664997126,
    "timestep_mean": 444.75,
    "epsilon_norm": 6.5857415199279785,
    "prediction_norm": 4.874833106994629,
    "timestamp": "2025-07-14T12:34:20.144841"
  },
  {
    "epoch": 1307,
    "batch_idx": 32,
    "loss": 0.2544611692428589,
    "learning_rate": 0.00026812000664997126,
    "timestep_mean": 575.125,
    "epsilon_norm": 8.367788314819336,
    "prediction_norm": 7.506925106048584,
    "timestamp": "2025-07-14T12:34:20.149668"
  },
  {
    "epoch": 1307,
    "batch_idx": 64,
    "loss": 0.23534275591373444,
    "learning_rate": 0.00026812000664997126,
    "timestep_mean": 408.0625,
    "epsilon_norm": 7.295114994049072,
    "prediction_norm": 6.316963195800781,
    "timestamp": "2025-07-14T12:34:20.154629"
  },
  {
    "epoch": 1307,
    "batch_idx": 96,
    "loss": 0.41494280099868774,
    "learning_rate": 0.00026812000664997126,
    "timestep_mean": 453.65625,
    "epsilon_norm": 9.145188331604004,
    "prediction_norm": 6.971514701843262,
    "timestamp": "2025-07-14T12:34:20.159453"
  },
  {
    "epoch": 1307,
    "batch_idx": 128,
    "loss": 0.4754044711589813,
    "learning_rate": 0.00026812000664997126,
    "timestep_mean": 553.53125,
    "epsilon_norm": 8.912930488586426,
    "prediction_norm": 6.639981746673584,
    "timestamp": "2025-07-14T12:34:20.164342"
  },
  {
    "epoch": 1307,
    "batch_idx": 160,
    "loss": 0.24986329674720764,
    "learning_rate": 0.00026812000664997126,
    "timestep_mean": 448.96875,
    "epsilon_norm": 8.728529930114746,
    "prediction_norm": 7.028165817260742,
    "timestamp": "2025-07-14T12:34:20.169215"
  },
  {
    "epoch": 1307,
    "batch_idx": 192,
    "loss": 0.08567866683006287,
    "learning_rate": 0.00026812000664997126,
    "timestep_mean": 502.25,
    "epsilon_norm": 2.535154104232788,
    "prediction_norm": 2.5020856857299805,
    "timestamp": "2025-07-14T12:34:20.174111"
  },
  {
    "epoch": 1308,
    "batch_idx": 0,
    "loss": 0.27262672781944275,
    "learning_rate": 0.00026742446096127105,
    "timestep_mean": 478.25,
    "epsilon_norm": 7.882228374481201,
    "prediction_norm": 6.59382963180542,
    "timestamp": "2025-07-14T12:34:20.178951"
  },
  {
    "epoch": 1308,
    "batch_idx": 32,
    "loss": 0.2565406858921051,
    "learning_rate": 0.00026742446096127105,
    "timestep_mean": 558.8125,
    "epsilon_norm": 7.820015907287598,
    "prediction_norm": 6.651301383972168,
    "timestamp": "2025-07-14T12:34:20.183864"
  },
  {
    "epoch": 1308,
    "batch_idx": 64,
    "loss": 0.14839349687099457,
    "learning_rate": 0.00026742446096127105,
    "timestep_mean": 601.03125,
    "epsilon_norm": 8.608203887939453,
    "prediction_norm": 7.611730098724365,
    "timestamp": "2025-07-14T12:34:20.188719"
  },
  {
    "epoch": 1308,
    "batch_idx": 96,
    "loss": 0.3592103123664856,
    "learning_rate": 0.00026742446096127105,
    "timestep_mean": 493.03125,
    "epsilon_norm": 9.353798866271973,
    "prediction_norm": 8.155510902404785,
    "timestamp": "2025-07-14T12:34:20.193586"
  },
  {
    "epoch": 1308,
    "batch_idx": 128,
    "loss": 0.16670861840248108,
    "learning_rate": 0.00026742446096127105,
    "timestep_mean": 465.21875,
    "epsilon_norm": 7.167678356170654,
    "prediction_norm": 6.504199981689453,
    "timestamp": "2025-07-14T12:34:20.198423"
  },
  {
    "epoch": 1308,
    "batch_idx": 160,
    "loss": 0.20652779936790466,
    "learning_rate": 0.00026742446096127105,
    "timestep_mean": 576.34375,
    "epsilon_norm": 8.23838996887207,
    "prediction_norm": 7.2651848793029785,
    "timestamp": "2025-07-14T12:34:20.203271"
  },
  {
    "epoch": 1308,
    "batch_idx": 192,
    "loss": 0.7598875761032104,
    "learning_rate": 0.00026742446096127105,
    "timestep_mean": 578.625,
    "epsilon_norm": 6.0433502197265625,
    "prediction_norm": 3.75527024269104,
    "timestamp": "2025-07-14T12:34:20.208156"
  },
  {
    "epoch": 1309,
    "batch_idx": 0,
    "loss": 0.342579185962677,
    "learning_rate": 0.0002667294891295939,
    "timestep_mean": 518.71875,
    "epsilon_norm": 10.57336139678955,
    "prediction_norm": 9.93045425415039,
    "timestamp": "2025-07-14T12:34:20.213164"
  },
  {
    "epoch": 1309,
    "batch_idx": 32,
    "loss": 0.24310818314552307,
    "learning_rate": 0.0002667294891295939,
    "timestep_mean": 487.8125,
    "epsilon_norm": 8.022330284118652,
    "prediction_norm": 6.448574066162109,
    "timestamp": "2025-07-14T12:34:20.218112"
  },
  {
    "epoch": 1309,
    "batch_idx": 64,
    "loss": 0.16177599132061005,
    "learning_rate": 0.0002667294891295939,
    "timestep_mean": 542.78125,
    "epsilon_norm": 8.222188949584961,
    "prediction_norm": 7.464408874511719,
    "timestamp": "2025-07-14T12:34:20.222962"
  },
  {
    "epoch": 1309,
    "batch_idx": 96,
    "loss": 0.4169166684150696,
    "learning_rate": 0.0002667294891295939,
    "timestep_mean": 460.75,
    "epsilon_norm": 9.14953899383545,
    "prediction_norm": 7.2540717124938965,
    "timestamp": "2025-07-14T12:34:20.227855"
  },
  {
    "epoch": 1309,
    "batch_idx": 128,
    "loss": 0.29507097601890564,
    "learning_rate": 0.0002667294891295939,
    "timestep_mean": 547.8125,
    "epsilon_norm": 8.545952796936035,
    "prediction_norm": 7.3857879638671875,
    "timestamp": "2025-07-14T12:34:20.232730"
  },
  {
    "epoch": 1309,
    "batch_idx": 160,
    "loss": 0.2732793688774109,
    "learning_rate": 0.0002667294891295939,
    "timestep_mean": 548.96875,
    "epsilon_norm": 8.670025825500488,
    "prediction_norm": 7.875837326049805,
    "timestamp": "2025-07-14T12:34:20.237694"
  },
  {
    "epoch": 1309,
    "batch_idx": 192,
    "loss": 0.18705399334430695,
    "learning_rate": 0.0002667294891295939,
    "timestep_mean": 482.5,
    "epsilon_norm": 3.713266372680664,
    "prediction_norm": 3.2511098384857178,
    "timestamp": "2025-07-14T12:34:20.242576"
  },
  {
    "epoch": 1310,
    "batch_idx": 0,
    "loss": 0.29033058881759644,
    "learning_rate": 0.0002660350928697136,
    "timestep_mean": 476.96875,
    "epsilon_norm": 7.58104944229126,
    "prediction_norm": 6.0436015129089355,
    "timestamp": "2025-07-14T12:34:20.247479"
  },
  {
    "epoch": 1310,
    "batch_idx": 32,
    "loss": 0.3429754078388214,
    "learning_rate": 0.0002660350928697136,
    "timestep_mean": 481.03125,
    "epsilon_norm": 7.10517692565918,
    "prediction_norm": 6.014122009277344,
    "timestamp": "2025-07-14T12:34:20.252357"
  },
  {
    "epoch": 1310,
    "batch_idx": 64,
    "loss": 0.3841647505760193,
    "learning_rate": 0.0002660350928697136,
    "timestep_mean": 466.25,
    "epsilon_norm": 9.615314483642578,
    "prediction_norm": 7.865270137786865,
    "timestamp": "2025-07-14T12:34:20.257208"
  },
  {
    "epoch": 1310,
    "batch_idx": 96,
    "loss": 0.28598570823669434,
    "learning_rate": 0.0002660350928697136,
    "timestep_mean": 494.625,
    "epsilon_norm": 8.059191703796387,
    "prediction_norm": 6.436307430267334,
    "timestamp": "2025-07-14T12:34:20.262111"
  },
  {
    "epoch": 1310,
    "batch_idx": 128,
    "loss": 0.17418840527534485,
    "learning_rate": 0.0002660350928697136,
    "timestep_mean": 513.15625,
    "epsilon_norm": 7.099476337432861,
    "prediction_norm": 6.3954291343688965,
    "timestamp": "2025-07-14T12:34:20.267071"
  },
  {
    "epoch": 1310,
    "batch_idx": 160,
    "loss": 0.17083880305290222,
    "learning_rate": 0.0002660350928697136,
    "timestep_mean": 509.28125,
    "epsilon_norm": 7.133691310882568,
    "prediction_norm": 6.524470806121826,
    "timestamp": "2025-07-14T12:34:20.271986"
  },
  {
    "epoch": 1310,
    "batch_idx": 192,
    "loss": 0.13006232678890228,
    "learning_rate": 0.0002660350928697136,
    "timestep_mean": 460.5,
    "epsilon_norm": 3.2423148155212402,
    "prediction_norm": 3.0969741344451904,
    "timestamp": "2025-07-14T12:34:20.276834"
  },
  {
    "epoch": 1311,
    "batch_idx": 0,
    "loss": 0.18376758694648743,
    "learning_rate": 0.0002653412738949837,
    "timestep_mean": 558.71875,
    "epsilon_norm": 7.453969478607178,
    "prediction_norm": 6.823393821716309,
    "timestamp": "2025-07-14T12:34:20.281754"
  },
  {
    "epoch": 1311,
    "batch_idx": 32,
    "loss": 0.42758727073669434,
    "learning_rate": 0.0002653412738949837,
    "timestep_mean": 420.5625,
    "epsilon_norm": 7.692399978637695,
    "prediction_norm": 5.9360504150390625,
    "timestamp": "2025-07-14T12:34:20.286707"
  },
  {
    "epoch": 1311,
    "batch_idx": 64,
    "loss": 0.2244131863117218,
    "learning_rate": 0.0002653412738949837,
    "timestep_mean": 582.625,
    "epsilon_norm": 8.4312162399292,
    "prediction_norm": 7.057281017303467,
    "timestamp": "2025-07-14T12:34:20.291583"
  },
  {
    "epoch": 1311,
    "batch_idx": 96,
    "loss": 0.41872653365135193,
    "learning_rate": 0.0002653412738949837,
    "timestep_mean": 449.5625,
    "epsilon_norm": 7.872386932373047,
    "prediction_norm": 5.839361667633057,
    "timestamp": "2025-07-14T12:34:20.296480"
  },
  {
    "epoch": 1311,
    "batch_idx": 128,
    "loss": 0.3738124966621399,
    "learning_rate": 0.0002653412738949837,
    "timestep_mean": 509.5,
    "epsilon_norm": 8.244146347045898,
    "prediction_norm": 6.75272274017334,
    "timestamp": "2025-07-14T12:34:20.301335"
  },
  {
    "epoch": 1311,
    "batch_idx": 160,
    "loss": 0.28360578417778015,
    "learning_rate": 0.0002653412738949837,
    "timestep_mean": 478.09375,
    "epsilon_norm": 7.501491069793701,
    "prediction_norm": 6.1725592613220215,
    "timestamp": "2025-07-14T12:34:20.306244"
  },
  {
    "epoch": 1311,
    "batch_idx": 192,
    "loss": 0.5589559674263,
    "learning_rate": 0.0002653412738949837,
    "timestep_mean": 365.0,
    "epsilon_norm": 5.296983242034912,
    "prediction_norm": 4.431001663208008,
    "timestamp": "2025-07-14T12:34:20.311088"
  },
  {
    "epoch": 1312,
    "batch_idx": 0,
    "loss": 0.10949137806892395,
    "learning_rate": 0.00026464803391733414,
    "timestep_mean": 587.5,
    "epsilon_norm": 7.5744829177856445,
    "prediction_norm": 6.555978775024414,
    "timestamp": "2025-07-14T12:34:20.315934"
  },
  {
    "epoch": 1312,
    "batch_idx": 32,
    "loss": 0.26785770058631897,
    "learning_rate": 0.00026464803391733414,
    "timestep_mean": 524.28125,
    "epsilon_norm": 8.040384292602539,
    "prediction_norm": 6.984381198883057,
    "timestamp": "2025-07-14T12:34:20.320894"
  },
  {
    "epoch": 1312,
    "batch_idx": 64,
    "loss": 0.21140873432159424,
    "learning_rate": 0.00026464803391733414,
    "timestep_mean": 511.03125,
    "epsilon_norm": 7.677743434906006,
    "prediction_norm": 7.5143842697143555,
    "timestamp": "2025-07-14T12:34:20.325875"
  },
  {
    "epoch": 1312,
    "batch_idx": 96,
    "loss": 0.1785280406475067,
    "learning_rate": 0.00026464803391733414,
    "timestep_mean": 544.96875,
    "epsilon_norm": 7.112606048583984,
    "prediction_norm": 6.29475736618042,
    "timestamp": "2025-07-14T12:34:20.330831"
  },
  {
    "epoch": 1312,
    "batch_idx": 128,
    "loss": 0.4847927689552307,
    "learning_rate": 0.00026464803391733414,
    "timestep_mean": 486.90625,
    "epsilon_norm": 7.712899208068848,
    "prediction_norm": 5.598557949066162,
    "timestamp": "2025-07-14T12:34:20.335735"
  },
  {
    "epoch": 1312,
    "batch_idx": 160,
    "loss": 0.2836659848690033,
    "learning_rate": 0.00026464803391733414,
    "timestep_mean": 514.5,
    "epsilon_norm": 7.325232028961182,
    "prediction_norm": 6.2922186851501465,
    "timestamp": "2025-07-14T12:34:20.340621"
  },
  {
    "epoch": 1312,
    "batch_idx": 192,
    "loss": 0.16676482558250427,
    "learning_rate": 0.00026464803391733414,
    "timestep_mean": 433.875,
    "epsilon_norm": 3.0090227127075195,
    "prediction_norm": 2.7725231647491455,
    "timestamp": "2025-07-14T12:34:20.345542"
  },
  {
    "epoch": 1313,
    "batch_idx": 0,
    "loss": 0.14073319733142853,
    "learning_rate": 0.00026395537464726483,
    "timestep_mean": 502.5625,
    "epsilon_norm": 8.808587074279785,
    "prediction_norm": 8.210599899291992,
    "timestamp": "2025-07-14T12:34:20.350428"
  },
  {
    "epoch": 1313,
    "batch_idx": 32,
    "loss": 0.30245262384414673,
    "learning_rate": 0.00026395537464726483,
    "timestep_mean": 494.65625,
    "epsilon_norm": 8.002082824707031,
    "prediction_norm": 6.686670303344727,
    "timestamp": "2025-07-14T12:34:20.355252"
  },
  {
    "epoch": 1313,
    "batch_idx": 64,
    "loss": 0.2114245891571045,
    "learning_rate": 0.00026395537464726483,
    "timestep_mean": 568.65625,
    "epsilon_norm": 9.341632843017578,
    "prediction_norm": 7.957239151000977,
    "timestamp": "2025-07-14T12:34:20.360159"
  },
  {
    "epoch": 1313,
    "batch_idx": 96,
    "loss": 0.3705236315727234,
    "learning_rate": 0.00026395537464726483,
    "timestep_mean": 499.78125,
    "epsilon_norm": 7.8078484535217285,
    "prediction_norm": 5.696454048156738,
    "timestamp": "2025-07-14T12:34:20.365075"
  },
  {
    "epoch": 1313,
    "batch_idx": 128,
    "loss": 0.32666444778442383,
    "learning_rate": 0.00026395537464726483,
    "timestep_mean": 591.75,
    "epsilon_norm": 8.707889556884766,
    "prediction_norm": 6.6869635581970215,
    "timestamp": "2025-07-14T12:34:20.369954"
  },
  {
    "epoch": 1313,
    "batch_idx": 160,
    "loss": 0.2008880078792572,
    "learning_rate": 0.00026395537464726483,
    "timestep_mean": 470.71875,
    "epsilon_norm": 7.365461826324463,
    "prediction_norm": 6.337399005889893,
    "timestamp": "2025-07-14T12:34:20.374838"
  },
  {
    "epoch": 1313,
    "batch_idx": 192,
    "loss": 0.2675713896751404,
    "learning_rate": 0.00026395537464726483,
    "timestep_mean": 573.5,
    "epsilon_norm": 4.797218322753906,
    "prediction_norm": 4.301706314086914,
    "timestamp": "2025-07-14T12:34:20.379806"
  },
  {
    "epoch": 1314,
    "batch_idx": 0,
    "loss": 0.4036717712879181,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 460.40625,
    "epsilon_norm": 7.811339855194092,
    "prediction_norm": 6.554746150970459,
    "timestamp": "2025-07-14T12:34:20.384639"
  },
  {
    "epoch": 1314,
    "batch_idx": 32,
    "loss": 0.24317720532417297,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 546.625,
    "epsilon_norm": 7.633969306945801,
    "prediction_norm": 6.340046405792236,
    "timestamp": "2025-07-14T12:34:20.389514"
  },
  {
    "epoch": 1314,
    "batch_idx": 64,
    "loss": 0.08798389881849289,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 545.6875,
    "epsilon_norm": 7.423811912536621,
    "prediction_norm": 6.901721477508545,
    "timestamp": "2025-07-14T12:34:20.394396"
  },
  {
    "epoch": 1314,
    "batch_idx": 96,
    "loss": 0.30271977186203003,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 480.875,
    "epsilon_norm": 7.5292253494262695,
    "prediction_norm": 6.322835922241211,
    "timestamp": "2025-07-14T12:34:20.399232"
  },
  {
    "epoch": 1314,
    "batch_idx": 128,
    "loss": 0.4516303241252899,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 529.875,
    "epsilon_norm": 8.306794166564941,
    "prediction_norm": 6.544371128082275,
    "timestamp": "2025-07-14T12:34:20.404198"
  },
  {
    "epoch": 1314,
    "batch_idx": 160,
    "loss": 0.34626707434654236,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 483.125,
    "epsilon_norm": 7.838226795196533,
    "prediction_norm": 6.52826452255249,
    "timestamp": "2025-07-14T12:34:20.409034"
  },
  {
    "epoch": 1314,
    "batch_idx": 192,
    "loss": 0.10967982560396194,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 490.0,
    "epsilon_norm": 3.370100975036621,
    "prediction_norm": 2.7843761444091797,
    "timestamp": "2025-07-14T12:34:20.413894"
  },
  {
    "epoch": 1315,
    "batch_idx": 0,
    "loss": 0.2697136402130127,
    "learning_rate": 0.00026257180506470307,
    "timestep_mean": 516.5625,
    "epsilon_norm": 8.361430168151855,
    "prediction_norm": 7.079683780670166,
    "timestamp": "2025-07-14T12:34:20.418942"
  },
  {
    "epoch": 1315,
    "batch_idx": 32,
    "loss": 0.2783088684082031,
    "learning_rate": 0.00026257180506470307,
    "timestep_mean": 469.0625,
    "epsilon_norm": 6.927597999572754,
    "prediction_norm": 6.009786128997803,
    "timestamp": "2025-07-14T12:34:20.423925"
  },
  {
    "epoch": 1315,
    "batch_idx": 64,
    "loss": 0.2505839467048645,
    "learning_rate": 0.00026257180506470307,
    "timestep_mean": 497.65625,
    "epsilon_norm": 8.391031265258789,
    "prediction_norm": 7.264503479003906,
    "timestamp": "2025-07-14T12:34:20.428841"
  },
  {
    "epoch": 1315,
    "batch_idx": 96,
    "loss": 0.1915920376777649,
    "learning_rate": 0.00026257180506470307,
    "timestep_mean": 602.78125,
    "epsilon_norm": 7.787078857421875,
    "prediction_norm": 7.155134677886963,
    "timestamp": "2025-07-14T12:34:20.433874"
  },
  {
    "epoch": 1315,
    "batch_idx": 128,
    "loss": 0.3740936815738678,
    "learning_rate": 0.00026257180506470307,
    "timestep_mean": 426.96875,
    "epsilon_norm": 7.436573028564453,
    "prediction_norm": 5.608097553253174,
    "timestamp": "2025-07-14T12:34:20.438815"
  },
  {
    "epoch": 1315,
    "batch_idx": 160,
    "loss": 0.3672466278076172,
    "learning_rate": 0.00026257180506470307,
    "timestep_mean": 514.0,
    "epsilon_norm": 8.646860122680664,
    "prediction_norm": 7.316461563110352,
    "timestamp": "2025-07-14T12:34:20.443703"
  },
  {
    "epoch": 1315,
    "batch_idx": 192,
    "loss": 0.1476118117570877,
    "learning_rate": 0.00026257180506470307,
    "timestep_mean": 540.25,
    "epsilon_norm": 3.523483991622925,
    "prediction_norm": 2.97689151763916,
    "timestamp": "2025-07-14T12:34:20.448719"
  },
  {
    "epoch": 1316,
    "batch_idx": 0,
    "loss": 0.1914052814245224,
    "learning_rate": 0.0002618808981660307,
    "timestep_mean": 450.375,
    "epsilon_norm": 7.25091552734375,
    "prediction_norm": 6.9076828956604,
    "timestamp": "2025-07-14T12:34:20.453630"
  },
  {
    "epoch": 1316,
    "batch_idx": 32,
    "loss": 0.15869277715682983,
    "learning_rate": 0.0002618808981660307,
    "timestep_mean": 524.0,
    "epsilon_norm": 6.221493721008301,
    "prediction_norm": 6.140661239624023,
    "timestamp": "2025-07-14T12:34:20.458643"
  },
  {
    "epoch": 1316,
    "batch_idx": 64,
    "loss": 0.15002784132957458,
    "learning_rate": 0.0002618808981660307,
    "timestep_mean": 533.53125,
    "epsilon_norm": 8.225872039794922,
    "prediction_norm": 8.50913143157959,
    "timestamp": "2025-07-14T12:34:20.463717"
  },
  {
    "epoch": 1316,
    "batch_idx": 96,
    "loss": 0.17050039768218994,
    "learning_rate": 0.0002618808981660307,
    "timestep_mean": 515.9375,
    "epsilon_norm": 7.620117664337158,
    "prediction_norm": 7.18862771987915,
    "timestamp": "2025-07-14T12:34:20.468954"
  },
  {
    "epoch": 1316,
    "batch_idx": 128,
    "loss": 0.3322056531906128,
    "learning_rate": 0.0002618808981660307,
    "timestep_mean": 513.15625,
    "epsilon_norm": 7.771108627319336,
    "prediction_norm": 6.677763938903809,
    "timestamp": "2025-07-14T12:34:20.474158"
  },
  {
    "epoch": 1316,
    "batch_idx": 160,
    "loss": 0.23559480905532837,
    "learning_rate": 0.0002618808981660307,
    "timestep_mean": 545.40625,
    "epsilon_norm": 8.101512908935547,
    "prediction_norm": 7.6181769371032715,
    "timestamp": "2025-07-14T12:34:20.479458"
  },
  {
    "epoch": 1316,
    "batch_idx": 192,
    "loss": 0.3581603467464447,
    "learning_rate": 0.0002618808981660307,
    "timestep_mean": 220.25,
    "epsilon_norm": 2.9734318256378174,
    "prediction_norm": 1.750922679901123,
    "timestamp": "2025-07-14T12:34:20.484613"
  },
  {
    "epoch": 1317,
    "batch_idx": 0,
    "loss": 0.18603171408176422,
    "learning_rate": 0.00026119057880257153,
    "timestep_mean": 456.625,
    "epsilon_norm": 6.309511661529541,
    "prediction_norm": 5.799274921417236,
    "timestamp": "2025-07-14T12:34:20.489820"
  },
  {
    "epoch": 1317,
    "batch_idx": 32,
    "loss": 0.267438143491745,
    "learning_rate": 0.00026119057880257153,
    "timestep_mean": 504.96875,
    "epsilon_norm": 7.66397762298584,
    "prediction_norm": 6.380868434906006,
    "timestamp": "2025-07-14T12:34:20.495058"
  },
  {
    "epoch": 1317,
    "batch_idx": 64,
    "loss": 0.4918363094329834,
    "learning_rate": 0.00026119057880257153,
    "timestep_mean": 521.15625,
    "epsilon_norm": 9.022440910339355,
    "prediction_norm": 7.103128433227539,
    "timestamp": "2025-07-14T12:34:20.500253"
  },
  {
    "epoch": 1317,
    "batch_idx": 96,
    "loss": 0.27029645442962646,
    "learning_rate": 0.00026119057880257153,
    "timestep_mean": 547.125,
    "epsilon_norm": 7.825032711029053,
    "prediction_norm": 7.235706806182861,
    "timestamp": "2025-07-14T12:34:20.505443"
  },
  {
    "epoch": 1317,
    "batch_idx": 128,
    "loss": 0.3634735941886902,
    "learning_rate": 0.00026119057880257153,
    "timestep_mean": 490.34375,
    "epsilon_norm": 8.189469337463379,
    "prediction_norm": 6.531590938568115,
    "timestamp": "2025-07-14T12:34:20.510640"
  },
  {
    "epoch": 1317,
    "batch_idx": 160,
    "loss": 0.23567163944244385,
    "learning_rate": 0.00026119057880257153,
    "timestep_mean": 513.0625,
    "epsilon_norm": 7.600718975067139,
    "prediction_norm": 6.68934440612793,
    "timestamp": "2025-07-14T12:34:20.515949"
  },
  {
    "epoch": 1317,
    "batch_idx": 192,
    "loss": 0.34516817331314087,
    "learning_rate": 0.00026119057880257153,
    "timestep_mean": 460.75,
    "epsilon_norm": 4.156609058380127,
    "prediction_norm": 2.8329267501831055,
    "timestamp": "2025-07-14T12:34:20.521108"
  },
  {
    "epoch": 1318,
    "batch_idx": 0,
    "loss": 0.2575807571411133,
    "learning_rate": 0.0002605008486776198,
    "timestep_mean": 486.375,
    "epsilon_norm": 8.373329162597656,
    "prediction_norm": 7.485301494598389,
    "timestamp": "2025-07-14T12:34:20.526293"
  },
  {
    "epoch": 1318,
    "batch_idx": 32,
    "loss": 0.3431813716888428,
    "learning_rate": 0.0002605008486776198,
    "timestep_mean": 454.78125,
    "epsilon_norm": 6.916362285614014,
    "prediction_norm": 4.984824180603027,
    "timestamp": "2025-07-14T12:34:20.531401"
  },
  {
    "epoch": 1318,
    "batch_idx": 64,
    "loss": 0.17200012505054474,
    "learning_rate": 0.0002605008486776198,
    "timestep_mean": 647.875,
    "epsilon_norm": 6.408519744873047,
    "prediction_norm": 5.722326278686523,
    "timestamp": "2025-07-14T12:34:20.536299"
  },
  {
    "epoch": 1318,
    "batch_idx": 96,
    "loss": 0.3989693224430084,
    "learning_rate": 0.0002605008486776198,
    "timestep_mean": 492.78125,
    "epsilon_norm": 8.127129554748535,
    "prediction_norm": 5.928872108459473,
    "timestamp": "2025-07-14T12:34:20.541431"
  },
  {
    "epoch": 1318,
    "batch_idx": 128,
    "loss": 0.23926430940628052,
    "learning_rate": 0.0002605008486776198,
    "timestep_mean": 522.78125,
    "epsilon_norm": 9.284695625305176,
    "prediction_norm": 8.050673484802246,
    "timestamp": "2025-07-14T12:34:20.546609"
  },
  {
    "epoch": 1318,
    "batch_idx": 160,
    "loss": 0.297940194606781,
    "learning_rate": 0.0002605008486776198,
    "timestep_mean": 488.21875,
    "epsilon_norm": 7.038440704345703,
    "prediction_norm": 5.902083396911621,
    "timestamp": "2025-07-14T12:34:20.551799"
  },
  {
    "epoch": 1318,
    "batch_idx": 192,
    "loss": 0.07541264593601227,
    "learning_rate": 0.0002605008486776198,
    "timestep_mean": 545.5,
    "epsilon_norm": 4.102794170379639,
    "prediction_norm": 4.210380554199219,
    "timestamp": "2025-07-14T12:34:20.556991"
  },
  {
    "epoch": 1319,
    "batch_idx": 0,
    "loss": 0.25834912061691284,
    "learning_rate": 0.0002598117094930161,
    "timestep_mean": 454.75,
    "epsilon_norm": 7.170166969299316,
    "prediction_norm": 5.6079912185668945,
    "timestamp": "2025-07-14T12:34:20.562163"
  },
  {
    "epoch": 1319,
    "batch_idx": 32,
    "loss": 0.16073483228683472,
    "learning_rate": 0.0002598117094930161,
    "timestep_mean": 507.53125,
    "epsilon_norm": 7.14863395690918,
    "prediction_norm": 6.167912006378174,
    "timestamp": "2025-07-14T12:34:20.567452"
  },
  {
    "epoch": 1319,
    "batch_idx": 64,
    "loss": 0.4016653597354889,
    "learning_rate": 0.0002598117094930161,
    "timestep_mean": 460.28125,
    "epsilon_norm": 8.707262992858887,
    "prediction_norm": 6.27679443359375,
    "timestamp": "2025-07-14T12:34:20.572537"
  },
  {
    "epoch": 1319,
    "batch_idx": 96,
    "loss": 0.12621472775936127,
    "learning_rate": 0.0002598117094930161,
    "timestep_mean": 658.28125,
    "epsilon_norm": 8.108513832092285,
    "prediction_norm": 7.243099212646484,
    "timestamp": "2025-07-14T12:34:20.577669"
  },
  {
    "epoch": 1319,
    "batch_idx": 128,
    "loss": 0.16038888692855835,
    "learning_rate": 0.0002598117094930161,
    "timestep_mean": 589.4375,
    "epsilon_norm": 7.530125141143799,
    "prediction_norm": 6.597311496734619,
    "timestamp": "2025-07-14T12:34:20.582859"
  },
  {
    "epoch": 1319,
    "batch_idx": 160,
    "loss": 0.21097798645496368,
    "learning_rate": 0.0002598117094930161,
    "timestep_mean": 452.5625,
    "epsilon_norm": 7.203988552093506,
    "prediction_norm": 6.352427959442139,
    "timestamp": "2025-07-14T12:34:20.588049"
  },
  {
    "epoch": 1319,
    "batch_idx": 192,
    "loss": 0.33590394258499146,
    "learning_rate": 0.0002598117094930161,
    "timestep_mean": 550.375,
    "epsilon_norm": 6.384980201721191,
    "prediction_norm": 5.758316516876221,
    "timestamp": "2025-07-14T12:34:20.593220"
  },
  {
    "epoch": 1320,
    "batch_idx": 0,
    "loss": 0.30541694164276123,
    "learning_rate": 0.0002591231629491428,
    "timestep_mean": 584.25,
    "epsilon_norm": 7.9403276443481445,
    "prediction_norm": 5.9623517990112305,
    "timestamp": "2025-07-14T12:34:20.598405"
  },
  {
    "epoch": 1320,
    "batch_idx": 32,
    "loss": 0.39380133152008057,
    "learning_rate": 0.0002591231629491428,
    "timestep_mean": 530.3125,
    "epsilon_norm": 8.704306602478027,
    "prediction_norm": 6.903676509857178,
    "timestamp": "2025-07-14T12:34:20.603673"
  },
  {
    "epoch": 1320,
    "batch_idx": 64,
    "loss": 0.29790377616882324,
    "learning_rate": 0.0002591231629491428,
    "timestep_mean": 487.46875,
    "epsilon_norm": 9.415085792541504,
    "prediction_norm": 7.54196834564209,
    "timestamp": "2025-07-14T12:34:20.608883"
  },
  {
    "epoch": 1320,
    "batch_idx": 96,
    "loss": 0.18887409567832947,
    "learning_rate": 0.0002591231629491428,
    "timestep_mean": 455.0625,
    "epsilon_norm": 7.136123180389404,
    "prediction_norm": 6.27907133102417,
    "timestamp": "2025-07-14T12:34:20.614048"
  },
  {
    "epoch": 1320,
    "batch_idx": 128,
    "loss": 0.348052442073822,
    "learning_rate": 0.0002591231629491428,
    "timestep_mean": 437.8125,
    "epsilon_norm": 7.86665153503418,
    "prediction_norm": 6.5570268630981445,
    "timestamp": "2025-07-14T12:34:20.619270"
  },
  {
    "epoch": 1320,
    "batch_idx": 160,
    "loss": 0.27137064933776855,
    "learning_rate": 0.0002591231629491428,
    "timestep_mean": 491.6875,
    "epsilon_norm": 8.66220474243164,
    "prediction_norm": 6.960577487945557,
    "timestamp": "2025-07-14T12:34:20.624476"
  },
  {
    "epoch": 1320,
    "batch_idx": 192,
    "loss": 0.11786650866270065,
    "learning_rate": 0.0002591231629491428,
    "timestep_mean": 433.25,
    "epsilon_norm": 2.885924816131592,
    "prediction_norm": 3.1794369220733643,
    "timestamp": "2025-07-14T12:34:20.629764"
  },
  {
    "epoch": 1321,
    "batch_idx": 0,
    "loss": 0.3163604438304901,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 493.4375,
    "epsilon_norm": 7.431827068328857,
    "prediction_norm": 6.110389709472656,
    "timestamp": "2025-07-14T12:34:20.634888"
  },
  {
    "epoch": 1321,
    "batch_idx": 32,
    "loss": 0.27805954217910767,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 501.25,
    "epsilon_norm": 8.958490371704102,
    "prediction_norm": 7.891871929168701,
    "timestamp": "2025-07-14T12:34:20.640113"
  },
  {
    "epoch": 1321,
    "batch_idx": 64,
    "loss": 0.2891656458377838,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 458.96875,
    "epsilon_norm": 7.66810941696167,
    "prediction_norm": 6.048529624938965,
    "timestamp": "2025-07-14T12:34:20.645303"
  },
  {
    "epoch": 1321,
    "batch_idx": 96,
    "loss": 0.23776641488075256,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 587.6875,
    "epsilon_norm": 7.923181533813477,
    "prediction_norm": 7.060213565826416,
    "timestamp": "2025-07-14T12:34:20.650560"
  },
  {
    "epoch": 1321,
    "batch_idx": 128,
    "loss": 0.3703421354293823,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 438.96875,
    "epsilon_norm": 6.950669765472412,
    "prediction_norm": 5.676699161529541,
    "timestamp": "2025-07-14T12:34:20.655757"
  },
  {
    "epoch": 1321,
    "batch_idx": 160,
    "loss": 0.26239240169525146,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 548.375,
    "epsilon_norm": 8.796846389770508,
    "prediction_norm": 7.307015419006348,
    "timestamp": "2025-07-14T12:34:20.660961"
  },
  {
    "epoch": 1321,
    "batch_idx": 192,
    "loss": 0.1309734731912613,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 460.375,
    "epsilon_norm": 3.132370710372925,
    "prediction_norm": 2.9079110622406006,
    "timestamp": "2025-07-14T12:34:20.666138"
  },
  {
    "epoch": 1322,
    "batch_idx": 0,
    "loss": 0.17667701840400696,
    "learning_rate": 0.00025774785457780133,
    "timestep_mean": 604.5625,
    "epsilon_norm": 7.932619094848633,
    "prediction_norm": 7.102458953857422,
    "timestamp": "2025-07-14T12:34:20.671226"
  },
  {
    "epoch": 1322,
    "batch_idx": 32,
    "loss": 0.22562327980995178,
    "learning_rate": 0.00025774785457780133,
    "timestep_mean": 493.25,
    "epsilon_norm": 8.019915580749512,
    "prediction_norm": 6.397639274597168,
    "timestamp": "2025-07-14T12:34:20.676378"
  },
  {
    "epoch": 1322,
    "batch_idx": 64,
    "loss": 0.3042469024658203,
    "learning_rate": 0.00025774785457780133,
    "timestep_mean": 508.75,
    "epsilon_norm": 8.820176124572754,
    "prediction_norm": 7.606995105743408,
    "timestamp": "2025-07-14T12:34:20.681469"
  },
  {
    "epoch": 1322,
    "batch_idx": 96,
    "loss": 0.21133489906787872,
    "learning_rate": 0.00025774785457780133,
    "timestep_mean": 536.25,
    "epsilon_norm": 8.43329906463623,
    "prediction_norm": 7.503407001495361,
    "timestamp": "2025-07-14T12:34:20.686524"
  },
  {
    "epoch": 1322,
    "batch_idx": 128,
    "loss": 0.18226730823516846,
    "learning_rate": 0.00025774785457780133,
    "timestep_mean": 504.21875,
    "epsilon_norm": 8.588445663452148,
    "prediction_norm": 7.675279140472412,
    "timestamp": "2025-07-14T12:34:20.691620"
  },
  {
    "epoch": 1322,
    "batch_idx": 160,
    "loss": 0.2076818346977234,
    "learning_rate": 0.00025774785457780133,
    "timestep_mean": 523.84375,
    "epsilon_norm": 7.194603443145752,
    "prediction_norm": 6.209934234619141,
    "timestamp": "2025-07-14T12:34:20.696780"
  },
  {
    "epoch": 1322,
    "batch_idx": 192,
    "loss": 0.08238089084625244,
    "learning_rate": 0.00025774785457780133,
    "timestep_mean": 648.0,
    "epsilon_norm": 4.119548320770264,
    "prediction_norm": 4.218550205230713,
    "timestamp": "2025-07-14T12:34:20.701873"
  },
  {
    "epoch": 1323,
    "batch_idx": 0,
    "loss": 0.12022868543863297,
    "learning_rate": 0.0002570610961437701,
    "timestep_mean": 486.15625,
    "epsilon_norm": 7.668467044830322,
    "prediction_norm": 7.225843906402588,
    "timestamp": "2025-07-14T12:34:20.706992"
  },
  {
    "epoch": 1323,
    "batch_idx": 32,
    "loss": 0.22546076774597168,
    "learning_rate": 0.0002570610961437701,
    "timestep_mean": 490.15625,
    "epsilon_norm": 7.525576114654541,
    "prediction_norm": 6.513032913208008,
    "timestamp": "2025-07-14T12:34:20.712066"
  },
  {
    "epoch": 1323,
    "batch_idx": 64,
    "loss": 0.16746333241462708,
    "learning_rate": 0.0002570610961437701,
    "timestep_mean": 563.1875,
    "epsilon_norm": 8.384352684020996,
    "prediction_norm": 7.650187015533447,
    "timestamp": "2025-07-14T12:34:20.717334"
  },
  {
    "epoch": 1323,
    "batch_idx": 96,
    "loss": 0.19976961612701416,
    "learning_rate": 0.0002570610961437701,
    "timestep_mean": 477.65625,
    "epsilon_norm": 8.903881072998047,
    "prediction_norm": 8.22379207611084,
    "timestamp": "2025-07-14T12:34:20.722343"
  },
  {
    "epoch": 1323,
    "batch_idx": 128,
    "loss": 0.3762126564979553,
    "learning_rate": 0.0002570610961437701,
    "timestep_mean": 435.6875,
    "epsilon_norm": 7.548003673553467,
    "prediction_norm": 5.948712348937988,
    "timestamp": "2025-07-14T12:34:20.727477"
  },
  {
    "epoch": 1323,
    "batch_idx": 160,
    "loss": 0.37428516149520874,
    "learning_rate": 0.0002570610961437701,
    "timestep_mean": 442.3125,
    "epsilon_norm": 8.142833709716797,
    "prediction_norm": 6.435837745666504,
    "timestamp": "2025-07-14T12:34:20.732549"
  },
  {
    "epoch": 1323,
    "batch_idx": 192,
    "loss": 0.23530125617980957,
    "learning_rate": 0.0002570610961437701,
    "timestep_mean": 458.0,
    "epsilon_norm": 6.641362190246582,
    "prediction_norm": 6.673640727996826,
    "timestamp": "2025-07-14T12:34:20.737817"
  },
  {
    "epoch": 1324,
    "batch_idx": 0,
    "loss": 0.37409359216690063,
    "learning_rate": 0.0002563749371373343,
    "timestep_mean": 486.96875,
    "epsilon_norm": 7.6101765632629395,
    "prediction_norm": 6.342494487762451,
    "timestamp": "2025-07-14T12:34:20.742789"
  },
  {
    "epoch": 1324,
    "batch_idx": 32,
    "loss": 0.15525829792022705,
    "learning_rate": 0.0002563749371373343,
    "timestep_mean": 503.15625,
    "epsilon_norm": 7.025813102722168,
    "prediction_norm": 6.40433931350708,
    "timestamp": "2025-07-14T12:34:20.747866"
  },
  {
    "epoch": 1324,
    "batch_idx": 64,
    "loss": 0.1702166646718979,
    "learning_rate": 0.0002563749371373343,
    "timestep_mean": 559.8125,
    "epsilon_norm": 7.70793342590332,
    "prediction_norm": 6.760373592376709,
    "timestamp": "2025-07-14T12:34:20.752996"
  },
  {
    "epoch": 1324,
    "batch_idx": 96,
    "loss": 0.1591966152191162,
    "learning_rate": 0.0002563749371373343,
    "timestep_mean": 552.375,
    "epsilon_norm": 6.905396938323975,
    "prediction_norm": 6.131814956665039,
    "timestamp": "2025-07-14T12:34:20.758137"
  },
  {
    "epoch": 1324,
    "batch_idx": 128,
    "loss": 0.15465685725212097,
    "learning_rate": 0.0002563749371373343,
    "timestep_mean": 517.8125,
    "epsilon_norm": 7.82425594329834,
    "prediction_norm": 7.0023064613342285,
    "timestamp": "2025-07-14T12:34:20.763476"
  },
  {
    "epoch": 1324,
    "batch_idx": 160,
    "loss": 0.16635897755622864,
    "learning_rate": 0.0002563749371373343,
    "timestep_mean": 549.1875,
    "epsilon_norm": 7.6313347816467285,
    "prediction_norm": 6.437272071838379,
    "timestamp": "2025-07-14T12:34:20.768644"
  },
  {
    "epoch": 1324,
    "batch_idx": 192,
    "loss": 0.15919426083564758,
    "learning_rate": 0.0002563749371373343,
    "timestep_mean": 551.5,
    "epsilon_norm": 4.552279472351074,
    "prediction_norm": 3.831873655319214,
    "timestamp": "2025-07-14T12:34:20.773776"
  },
  {
    "epoch": 1325,
    "batch_idx": 0,
    "loss": 0.3303036391735077,
    "learning_rate": 0.0002556893792515228,
    "timestep_mean": 490.90625,
    "epsilon_norm": 8.209674835205078,
    "prediction_norm": 6.941542625427246,
    "timestamp": "2025-07-14T12:34:20.779020"
  },
  {
    "epoch": 1325,
    "batch_idx": 32,
    "loss": 0.17723774909973145,
    "learning_rate": 0.0002556893792515228,
    "timestep_mean": 578.1875,
    "epsilon_norm": 7.770141124725342,
    "prediction_norm": 7.040255546569824,
    "timestamp": "2025-07-14T12:34:20.784326"
  },
  {
    "epoch": 1325,
    "batch_idx": 64,
    "loss": 0.44371703267097473,
    "learning_rate": 0.0002556893792515228,
    "timestep_mean": 492.78125,
    "epsilon_norm": 7.533681392669678,
    "prediction_norm": 5.164848327636719,
    "timestamp": "2025-07-14T12:34:20.789539"
  },
  {
    "epoch": 1325,
    "batch_idx": 96,
    "loss": 0.44312596321105957,
    "learning_rate": 0.0002556893792515228,
    "timestep_mean": 480.875,
    "epsilon_norm": 8.093914985656738,
    "prediction_norm": 6.229034423828125,
    "timestamp": "2025-07-14T12:34:20.794755"
  },
  {
    "epoch": 1325,
    "batch_idx": 128,
    "loss": 0.4268583357334137,
    "learning_rate": 0.0002556893792515228,
    "timestep_mean": 473.28125,
    "epsilon_norm": 8.054685592651367,
    "prediction_norm": 5.906301498413086,
    "timestamp": "2025-07-14T12:34:20.799957"
  },
  {
    "epoch": 1325,
    "batch_idx": 160,
    "loss": 0.2738487124443054,
    "learning_rate": 0.0002556893792515228,
    "timestep_mean": 519.46875,
    "epsilon_norm": 8.082249641418457,
    "prediction_norm": 6.154101848602295,
    "timestamp": "2025-07-14T12:34:20.805367"
  },
  {
    "epoch": 1325,
    "batch_idx": 192,
    "loss": 0.43182113766670227,
    "learning_rate": 0.0002556893792515228,
    "timestep_mean": 355.125,
    "epsilon_norm": 3.678840398788452,
    "prediction_norm": 2.5888991355895996,
    "timestamp": "2025-07-14T12:34:20.810529"
  },
  {
    "epoch": 1326,
    "batch_idx": 0,
    "loss": 0.3750045895576477,
    "learning_rate": 0.0002550044241778822,
    "timestep_mean": 421.6875,
    "epsilon_norm": 8.910884857177734,
    "prediction_norm": 6.8080549240112305,
    "timestamp": "2025-07-14T12:34:20.815760"
  },
  {
    "epoch": 1326,
    "batch_idx": 32,
    "loss": 0.19373473525047302,
    "learning_rate": 0.0002550044241778822,
    "timestep_mean": 441.21875,
    "epsilon_norm": 7.334135055541992,
    "prediction_norm": 6.6194939613342285,
    "timestamp": "2025-07-14T12:34:20.821083"
  },
  {
    "epoch": 1326,
    "batch_idx": 64,
    "loss": 0.22603195905685425,
    "learning_rate": 0.0002550044241778822,
    "timestep_mean": 506.34375,
    "epsilon_norm": 7.724322319030762,
    "prediction_norm": 6.779774188995361,
    "timestamp": "2025-07-14T12:34:20.826195"
  },
  {
    "epoch": 1326,
    "batch_idx": 96,
    "loss": 0.2770542502403259,
    "learning_rate": 0.0002550044241778822,
    "timestep_mean": 562.875,
    "epsilon_norm": 7.159807205200195,
    "prediction_norm": 6.722158432006836,
    "timestamp": "2025-07-14T12:34:20.831257"
  },
  {
    "epoch": 1326,
    "batch_idx": 128,
    "loss": 0.30315399169921875,
    "learning_rate": 0.0002550044241778822,
    "timestep_mean": 446.71875,
    "epsilon_norm": 6.922240257263184,
    "prediction_norm": 5.529853820800781,
    "timestamp": "2025-07-14T12:34:20.836496"
  },
  {
    "epoch": 1326,
    "batch_idx": 160,
    "loss": 0.2448594570159912,
    "learning_rate": 0.0002550044241778822,
    "timestep_mean": 504.9375,
    "epsilon_norm": 7.4799485206604,
    "prediction_norm": 6.678357124328613,
    "timestamp": "2025-07-14T12:34:20.841777"
  },
  {
    "epoch": 1326,
    "batch_idx": 192,
    "loss": 0.38586825132369995,
    "learning_rate": 0.0002550044241778822,
    "timestep_mean": 464.875,
    "epsilon_norm": 4.408824443817139,
    "prediction_norm": 3.9515902996063232,
    "timestamp": "2025-07-14T12:34:20.846638"
  },
  {
    "epoch": 1327,
    "batch_idx": 0,
    "loss": 0.26270049810409546,
    "learning_rate": 0.0002543200736064703,
    "timestep_mean": 470.53125,
    "epsilon_norm": 7.233119010925293,
    "prediction_norm": 5.924134731292725,
    "timestamp": "2025-07-14T12:34:20.851544"
  },
  {
    "epoch": 1327,
    "batch_idx": 32,
    "loss": 0.218522846698761,
    "learning_rate": 0.0002543200736064703,
    "timestep_mean": 563.8125,
    "epsilon_norm": 8.124008178710938,
    "prediction_norm": 7.31053352355957,
    "timestamp": "2025-07-14T12:34:20.856351"
  },
  {
    "epoch": 1327,
    "batch_idx": 64,
    "loss": 0.30799901485443115,
    "learning_rate": 0.0002543200736064703,
    "timestep_mean": 436.75,
    "epsilon_norm": 7.692209720611572,
    "prediction_norm": 6.292633533477783,
    "timestamp": "2025-07-14T12:34:20.861263"
  },
  {
    "epoch": 1327,
    "batch_idx": 96,
    "loss": 0.1489105224609375,
    "learning_rate": 0.0002543200736064703,
    "timestep_mean": 472.0625,
    "epsilon_norm": 7.743783473968506,
    "prediction_norm": 7.208296298980713,
    "timestamp": "2025-07-14T12:34:20.866174"
  },
  {
    "epoch": 1327,
    "batch_idx": 128,
    "loss": 0.22335614264011383,
    "learning_rate": 0.0002543200736064703,
    "timestep_mean": 582.875,
    "epsilon_norm": 6.874399185180664,
    "prediction_norm": 5.838772773742676,
    "timestamp": "2025-07-14T12:34:20.871220"
  },
  {
    "epoch": 1327,
    "batch_idx": 160,
    "loss": 0.25328686833381653,
    "learning_rate": 0.0002543200736064703,
    "timestep_mean": 466.21875,
    "epsilon_norm": 7.820743560791016,
    "prediction_norm": 6.505722999572754,
    "timestamp": "2025-07-14T12:34:20.876069"
  },
  {
    "epoch": 1327,
    "batch_idx": 192,
    "loss": 0.28298476338386536,
    "learning_rate": 0.0002543200736064703,
    "timestep_mean": 528.25,
    "epsilon_norm": 4.349351406097412,
    "prediction_norm": 3.2015790939331055,
    "timestamp": "2025-07-14T12:34:20.880927"
  },
  {
    "epoch": 1328,
    "batch_idx": 0,
    "loss": 0.2897418141365051,
    "learning_rate": 0.00025363632922585443,
    "timestep_mean": 491.125,
    "epsilon_norm": 8.447150230407715,
    "prediction_norm": 7.637136936187744,
    "timestamp": "2025-07-14T12:34:20.885757"
  },
  {
    "epoch": 1328,
    "batch_idx": 32,
    "loss": 0.17307864129543304,
    "learning_rate": 0.00025363632922585443,
    "timestep_mean": 568.0,
    "epsilon_norm": 7.635401248931885,
    "prediction_norm": 6.849579811096191,
    "timestamp": "2025-07-14T12:34:20.890724"
  },
  {
    "epoch": 1328,
    "batch_idx": 64,
    "loss": 0.251106321811676,
    "learning_rate": 0.00025363632922585443,
    "timestep_mean": 591.5,
    "epsilon_norm": 8.165420532226562,
    "prediction_norm": 7.156839370727539,
    "timestamp": "2025-07-14T12:34:20.895625"
  },
  {
    "epoch": 1328,
    "batch_idx": 96,
    "loss": 0.29281678795814514,
    "learning_rate": 0.00025363632922585443,
    "timestep_mean": 516.75,
    "epsilon_norm": 8.735920906066895,
    "prediction_norm": 7.562055587768555,
    "timestamp": "2025-07-14T12:34:20.900549"
  },
  {
    "epoch": 1328,
    "batch_idx": 128,
    "loss": 0.2765409052371979,
    "learning_rate": 0.00025363632922585443,
    "timestep_mean": 440.28125,
    "epsilon_norm": 6.517659664154053,
    "prediction_norm": 5.401861190795898,
    "timestamp": "2025-07-14T12:34:20.905337"
  },
  {
    "epoch": 1328,
    "batch_idx": 160,
    "loss": 0.14926698803901672,
    "learning_rate": 0.00025363632922585443,
    "timestep_mean": 496.40625,
    "epsilon_norm": 6.841100215911865,
    "prediction_norm": 6.470329761505127,
    "timestamp": "2025-07-14T12:34:20.910210"
  },
  {
    "epoch": 1328,
    "batch_idx": 192,
    "loss": 0.16921812295913696,
    "learning_rate": 0.00025363632922585443,
    "timestep_mean": 495.75,
    "epsilon_norm": 3.8967199325561523,
    "prediction_norm": 3.908640146255493,
    "timestamp": "2025-07-14T12:34:20.915017"
  },
  {
    "epoch": 1329,
    "batch_idx": 0,
    "loss": 0.3302425146102905,
    "learning_rate": 0.0002529531927231063,
    "timestep_mean": 524.59375,
    "epsilon_norm": 7.770218372344971,
    "prediction_norm": 6.3816351890563965,
    "timestamp": "2025-07-14T12:34:20.919931"
  },
  {
    "epoch": 1329,
    "batch_idx": 32,
    "loss": 0.31212273240089417,
    "learning_rate": 0.0002529531927231063,
    "timestep_mean": 630.625,
    "epsilon_norm": 9.135150909423828,
    "prediction_norm": 7.5362091064453125,
    "timestamp": "2025-07-14T12:34:20.924770"
  },
  {
    "epoch": 1329,
    "batch_idx": 64,
    "loss": 0.4495700001716614,
    "learning_rate": 0.0002529531927231063,
    "timestep_mean": 373.21875,
    "epsilon_norm": 8.60777473449707,
    "prediction_norm": 6.2173614501953125,
    "timestamp": "2025-07-14T12:34:20.929659"
  },
  {
    "epoch": 1329,
    "batch_idx": 96,
    "loss": 0.11018076539039612,
    "learning_rate": 0.0002529531927231063,
    "timestep_mean": 551.34375,
    "epsilon_norm": 7.548236846923828,
    "prediction_norm": 7.18745756149292,
    "timestamp": "2025-07-14T12:34:20.934489"
  },
  {
    "epoch": 1329,
    "batch_idx": 128,
    "loss": 0.20668309926986694,
    "learning_rate": 0.0002529531927231063,
    "timestep_mean": 523.53125,
    "epsilon_norm": 8.851973533630371,
    "prediction_norm": 7.810619354248047,
    "timestamp": "2025-07-14T12:34:20.939324"
  },
  {
    "epoch": 1329,
    "batch_idx": 160,
    "loss": 0.24948960542678833,
    "learning_rate": 0.0002529531927231063,
    "timestep_mean": 421.8125,
    "epsilon_norm": 7.464482307434082,
    "prediction_norm": 6.103604793548584,
    "timestamp": "2025-07-14T12:34:20.944171"
  },
  {
    "epoch": 1329,
    "batch_idx": 192,
    "loss": 0.6434783935546875,
    "learning_rate": 0.0002529531927231063,
    "timestep_mean": 391.875,
    "epsilon_norm": 4.672936916351318,
    "prediction_norm": 3.1259913444519043,
    "timestamp": "2025-07-14T12:34:20.948997"
  },
  {
    "epoch": 1330,
    "batch_idx": 0,
    "loss": 0.09868312627077103,
    "learning_rate": 0.0002522706657837965,
    "timestep_mean": 549.6875,
    "epsilon_norm": 5.895328998565674,
    "prediction_norm": 5.415592670440674,
    "timestamp": "2025-07-14T12:34:20.953904"
  },
  {
    "epoch": 1330,
    "batch_idx": 32,
    "loss": 0.45488986372947693,
    "learning_rate": 0.0002522706657837965,
    "timestep_mean": 457.1875,
    "epsilon_norm": 8.202400207519531,
    "prediction_norm": 6.228180408477783,
    "timestamp": "2025-07-14T12:34:20.958792"
  },
  {
    "epoch": 1330,
    "batch_idx": 64,
    "loss": 0.45103347301483154,
    "learning_rate": 0.0002522706657837965,
    "timestep_mean": 456.875,
    "epsilon_norm": 8.447046279907227,
    "prediction_norm": 6.340634822845459,
    "timestamp": "2025-07-14T12:34:20.963595"
  },
  {
    "epoch": 1330,
    "batch_idx": 96,
    "loss": 0.18272367119789124,
    "learning_rate": 0.0002522706657837965,
    "timestep_mean": 523.125,
    "epsilon_norm": 8.363551139831543,
    "prediction_norm": 7.227571964263916,
    "timestamp": "2025-07-14T12:34:20.968440"
  },
  {
    "epoch": 1330,
    "batch_idx": 128,
    "loss": 0.17772896587848663,
    "learning_rate": 0.0002522706657837965,
    "timestep_mean": 542.8125,
    "epsilon_norm": 9.243490219116211,
    "prediction_norm": 8.323843955993652,
    "timestamp": "2025-07-14T12:34:20.973286"
  },
  {
    "epoch": 1330,
    "batch_idx": 160,
    "loss": 0.20889155566692352,
    "learning_rate": 0.0002522706657837965,
    "timestep_mean": 523.3125,
    "epsilon_norm": 8.335796356201172,
    "prediction_norm": 7.149926662445068,
    "timestamp": "2025-07-14T12:34:20.978184"
  },
  {
    "epoch": 1330,
    "batch_idx": 192,
    "loss": 0.16631314158439636,
    "learning_rate": 0.0002522706657837965,
    "timestep_mean": 735.625,
    "epsilon_norm": 4.495627403259277,
    "prediction_norm": 3.9615302085876465,
    "timestamp": "2025-07-14T12:34:20.982997"
  },
  {
    "epoch": 1331,
    "batch_idx": 0,
    "loss": 0.20772603154182434,
    "learning_rate": 0.00025158875009199326,
    "timestep_mean": 542.15625,
    "epsilon_norm": 8.046370506286621,
    "prediction_norm": 7.1857829093933105,
    "timestamp": "2025-07-14T12:34:20.988025"
  },
  {
    "epoch": 1331,
    "batch_idx": 32,
    "loss": 0.13958127796649933,
    "learning_rate": 0.00025158875009199326,
    "timestep_mean": 592.375,
    "epsilon_norm": 7.582056999206543,
    "prediction_norm": 6.6411309242248535,
    "timestamp": "2025-07-14T12:34:20.992870"
  },
  {
    "epoch": 1331,
    "batch_idx": 64,
    "loss": 0.26578783988952637,
    "learning_rate": 0.00025158875009199326,
    "timestep_mean": 514.53125,
    "epsilon_norm": 6.801445007324219,
    "prediction_norm": 6.063007354736328,
    "timestamp": "2025-07-14T12:34:20.997721"
  },
  {
    "epoch": 1331,
    "batch_idx": 96,
    "loss": 0.15593627095222473,
    "learning_rate": 0.00025158875009199326,
    "timestep_mean": 417.21875,
    "epsilon_norm": 6.9044294357299805,
    "prediction_norm": 6.370059490203857,
    "timestamp": "2025-07-14T12:34:21.002568"
  },
  {
    "epoch": 1331,
    "batch_idx": 128,
    "loss": 0.28676360845565796,
    "learning_rate": 0.00025158875009199326,
    "timestep_mean": 462.03125,
    "epsilon_norm": 7.7595624923706055,
    "prediction_norm": 6.806296348571777,
    "timestamp": "2025-07-14T12:34:21.007487"
  },
  {
    "epoch": 1331,
    "batch_idx": 160,
    "loss": 0.23486238718032837,
    "learning_rate": 0.00025158875009199326,
    "timestep_mean": 485.71875,
    "epsilon_norm": 7.500783443450928,
    "prediction_norm": 6.207302570343018,
    "timestamp": "2025-07-14T12:34:21.012507"
  },
  {
    "epoch": 1331,
    "batch_idx": 192,
    "loss": 0.0407930351793766,
    "learning_rate": 0.00025158875009199326,
    "timestep_mean": 571.25,
    "epsilon_norm": 2.787611961364746,
    "prediction_norm": 2.500361680984497,
    "timestamp": "2025-07-14T12:34:21.017330"
  },
  {
    "epoch": 1332,
    "batch_idx": 0,
    "loss": 0.23026971518993378,
    "learning_rate": 0.00025090744733025495,
    "timestep_mean": 586.375,
    "epsilon_norm": 8.496256828308105,
    "prediction_norm": 7.427279472351074,
    "timestamp": "2025-07-14T12:34:21.022224"
  },
  {
    "epoch": 1332,
    "batch_idx": 32,
    "loss": 0.4758022427558899,
    "learning_rate": 0.00025090744733025495,
    "timestep_mean": 507.4375,
    "epsilon_norm": 8.511412620544434,
    "prediction_norm": 7.204631328582764,
    "timestamp": "2025-07-14T12:34:21.027130"
  },
  {
    "epoch": 1332,
    "batch_idx": 64,
    "loss": 0.2394753098487854,
    "learning_rate": 0.00025090744733025495,
    "timestep_mean": 470.09375,
    "epsilon_norm": 8.856656074523926,
    "prediction_norm": 7.770715713500977,
    "timestamp": "2025-07-14T12:34:21.031965"
  },
  {
    "epoch": 1332,
    "batch_idx": 96,
    "loss": 0.27905428409576416,
    "learning_rate": 0.00025090744733025495,
    "timestep_mean": 536.84375,
    "epsilon_norm": 8.607402801513672,
    "prediction_norm": 7.5165181159973145,
    "timestamp": "2025-07-14T12:34:21.036924"
  },
  {
    "epoch": 1332,
    "batch_idx": 128,
    "loss": 0.20711609721183777,
    "learning_rate": 0.00025090744733025495,
    "timestep_mean": 488.78125,
    "epsilon_norm": 7.03306245803833,
    "prediction_norm": 6.100500583648682,
    "timestamp": "2025-07-14T12:34:21.041821"
  },
  {
    "epoch": 1332,
    "batch_idx": 160,
    "loss": 0.3066926598548889,
    "learning_rate": 0.00025090744733025495,
    "timestep_mean": 490.53125,
    "epsilon_norm": 8.349666595458984,
    "prediction_norm": 7.3404693603515625,
    "timestamp": "2025-07-14T12:34:21.046687"
  },
  {
    "epoch": 1332,
    "batch_idx": 192,
    "loss": 0.423904150724411,
    "learning_rate": 0.00025090744733025495,
    "timestep_mean": 359.25,
    "epsilon_norm": 4.634381294250488,
    "prediction_norm": 3.8896729946136475,
    "timestamp": "2025-07-14T12:34:21.051542"
  },
  {
    "epoch": 1333,
    "batch_idx": 0,
    "loss": 0.22894489765167236,
    "learning_rate": 0.00025022675917962876,
    "timestep_mean": 487.21875,
    "epsilon_norm": 7.922689914703369,
    "prediction_norm": 6.565747261047363,
    "timestamp": "2025-07-14T12:34:21.056426"
  },
  {
    "epoch": 1333,
    "batch_idx": 32,
    "loss": 0.23428188264369965,
    "learning_rate": 0.00025022675917962876,
    "timestep_mean": 511.4375,
    "epsilon_norm": 7.04290246963501,
    "prediction_norm": 6.334160804748535,
    "timestamp": "2025-07-14T12:34:21.061288"
  },
  {
    "epoch": 1333,
    "batch_idx": 64,
    "loss": 0.20950648188591003,
    "learning_rate": 0.00025022675917962876,
    "timestep_mean": 533.25,
    "epsilon_norm": 7.506192207336426,
    "prediction_norm": 6.474370002746582,
    "timestamp": "2025-07-14T12:34:21.066132"
  },
  {
    "epoch": 1333,
    "batch_idx": 96,
    "loss": 0.3600408732891083,
    "learning_rate": 0.00025022675917962876,
    "timestep_mean": 514.15625,
    "epsilon_norm": 7.800334930419922,
    "prediction_norm": 6.614818096160889,
    "timestamp": "2025-07-14T12:34:21.071134"
  },
  {
    "epoch": 1333,
    "batch_idx": 128,
    "loss": 0.460685133934021,
    "learning_rate": 0.00025022675917962876,
    "timestep_mean": 398.0625,
    "epsilon_norm": 8.476255416870117,
    "prediction_norm": 6.314887046813965,
    "timestamp": "2025-07-14T12:34:21.076285"
  },
  {
    "epoch": 1333,
    "batch_idx": 160,
    "loss": 0.22373750805854797,
    "learning_rate": 0.00025022675917962876,
    "timestep_mean": 485.71875,
    "epsilon_norm": 7.943822383880615,
    "prediction_norm": 7.581867694854736,
    "timestamp": "2025-07-14T12:34:21.081515"
  },
  {
    "epoch": 1333,
    "batch_idx": 192,
    "loss": 0.10522574931383133,
    "learning_rate": 0.00025022675917962876,
    "timestep_mean": 527.5,
    "epsilon_norm": 3.780247688293457,
    "prediction_norm": 3.1909189224243164,
    "timestamp": "2025-07-14T12:34:21.086753"
  },
  {
    "epoch": 1334,
    "batch_idx": 0,
    "loss": 0.2090359628200531,
    "learning_rate": 0.0002495466873196455,
    "timestep_mean": 578.03125,
    "epsilon_norm": 8.507509231567383,
    "prediction_norm": 7.158686637878418,
    "timestamp": "2025-07-14T12:34:21.091963"
  },
  {
    "epoch": 1334,
    "batch_idx": 32,
    "loss": 0.2743823230266571,
    "learning_rate": 0.0002495466873196455,
    "timestep_mean": 506.03125,
    "epsilon_norm": 8.484773635864258,
    "prediction_norm": 7.349581718444824,
    "timestamp": "2025-07-14T12:34:21.097149"
  },
  {
    "epoch": 1334,
    "batch_idx": 64,
    "loss": 0.21883463859558105,
    "learning_rate": 0.0002495466873196455,
    "timestep_mean": 509.0625,
    "epsilon_norm": 8.101088523864746,
    "prediction_norm": 7.05609130859375,
    "timestamp": "2025-07-14T12:34:21.102314"
  },
  {
    "epoch": 1334,
    "batch_idx": 96,
    "loss": 0.23686322569847107,
    "learning_rate": 0.0002495466873196455,
    "timestep_mean": 513.4375,
    "epsilon_norm": 7.988983154296875,
    "prediction_norm": 6.571600437164307,
    "timestamp": "2025-07-14T12:34:21.107497"
  },
  {
    "epoch": 1334,
    "batch_idx": 128,
    "loss": 0.3466304838657379,
    "learning_rate": 0.0002495466873196455,
    "timestep_mean": 470.9375,
    "epsilon_norm": 8.74404239654541,
    "prediction_norm": 6.988877773284912,
    "timestamp": "2025-07-14T12:34:21.112674"
  },
  {
    "epoch": 1334,
    "batch_idx": 160,
    "loss": 0.22662559151649475,
    "learning_rate": 0.0002495466873196455,
    "timestep_mean": 484.5,
    "epsilon_norm": 7.5685224533081055,
    "prediction_norm": 6.541111946105957,
    "timestamp": "2025-07-14T12:34:21.117791"
  },
  {
    "epoch": 1334,
    "batch_idx": 192,
    "loss": 0.5033172965049744,
    "learning_rate": 0.0002495466873196455,
    "timestep_mean": 327.25,
    "epsilon_norm": 3.53731632232666,
    "prediction_norm": 2.553220272064209,
    "timestamp": "2025-07-14T12:34:21.122936"
  },
  {
    "epoch": 1335,
    "batch_idx": 0,
    "loss": 0.2752133905887604,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 440.59375,
    "epsilon_norm": 9.503006935119629,
    "prediction_norm": 7.9085493087768555,
    "timestamp": "2025-07-14T12:34:21.128132"
  },
  {
    "epoch": 1335,
    "batch_idx": 32,
    "loss": 0.16037406027317047,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 592.84375,
    "epsilon_norm": 8.422380447387695,
    "prediction_norm": 7.485702991485596,
    "timestamp": "2025-07-14T12:34:21.133424"
  },
  {
    "epoch": 1335,
    "batch_idx": 64,
    "loss": 0.10801366716623306,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 593.59375,
    "epsilon_norm": 7.0997114181518555,
    "prediction_norm": 6.826859951019287,
    "timestamp": "2025-07-14T12:34:21.138582"
  },
  {
    "epoch": 1335,
    "batch_idx": 96,
    "loss": 0.17931310832500458,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 487.84375,
    "epsilon_norm": 7.115450859069824,
    "prediction_norm": 6.395959377288818,
    "timestamp": "2025-07-14T12:34:21.143774"
  },
  {
    "epoch": 1335,
    "batch_idx": 128,
    "loss": 0.16610132157802582,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 569.125,
    "epsilon_norm": 7.587230682373047,
    "prediction_norm": 6.961729526519775,
    "timestamp": "2025-07-14T12:34:21.148986"
  },
  {
    "epoch": 1335,
    "batch_idx": 160,
    "loss": 0.25312697887420654,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 534.78125,
    "epsilon_norm": 8.591206550598145,
    "prediction_norm": 7.712909698486328,
    "timestamp": "2025-07-14T12:34:21.154157"
  },
  {
    "epoch": 1335,
    "batch_idx": 192,
    "loss": 0.5648048520088196,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 394.125,
    "epsilon_norm": 4.602065563201904,
    "prediction_norm": 3.1222803592681885,
    "timestamp": "2025-07-14T12:34:21.159282"
  },
  {
    "epoch": 1336,
    "batch_idx": 0,
    "loss": 0.11551280319690704,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 476.09375,
    "epsilon_norm": 6.612367630004883,
    "prediction_norm": 6.518158435821533,
    "timestamp": "2025-07-14T12:34:21.164480"
  },
  {
    "epoch": 1336,
    "batch_idx": 32,
    "loss": 0.29653459787368774,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 487.25,
    "epsilon_norm": 7.843574047088623,
    "prediction_norm": 6.338270664215088,
    "timestamp": "2025-07-14T12:34:21.169652"
  },
  {
    "epoch": 1336,
    "batch_idx": 64,
    "loss": 0.23424488306045532,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 530.75,
    "epsilon_norm": 7.413295745849609,
    "prediction_norm": 6.06720495223999,
    "timestamp": "2025-07-14T12:34:21.174838"
  },
  {
    "epoch": 1336,
    "batch_idx": 96,
    "loss": 0.2716754078865051,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 445.03125,
    "epsilon_norm": 8.086448669433594,
    "prediction_norm": 7.468106269836426,
    "timestamp": "2025-07-14T12:34:21.180155"
  },
  {
    "epoch": 1336,
    "batch_idx": 128,
    "loss": 0.25956881046295166,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 483.40625,
    "epsilon_norm": 8.867417335510254,
    "prediction_norm": 7.339348316192627,
    "timestamp": "2025-07-14T12:34:21.185297"
  },
  {
    "epoch": 1336,
    "batch_idx": 160,
    "loss": 0.48769447207450867,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 449.96875,
    "epsilon_norm": 8.711423873901367,
    "prediction_norm": 6.928902626037598,
    "timestamp": "2025-07-14T12:34:21.190526"
  },
  {
    "epoch": 1336,
    "batch_idx": 192,
    "loss": 0.42596927285194397,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 346.0,
    "epsilon_norm": 4.09998893737793,
    "prediction_norm": 3.8723549842834473,
    "timestamp": "2025-07-14T12:34:21.195684"
  },
  {
    "epoch": 1337,
    "batch_idx": 0,
    "loss": 0.3216753602027893,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 494.75,
    "epsilon_norm": 8.459827423095703,
    "prediction_norm": 7.3363566398620605,
    "timestamp": "2025-07-14T12:34:21.200921"
  },
  {
    "epoch": 1337,
    "batch_idx": 32,
    "loss": 0.30810868740081787,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 463.375,
    "epsilon_norm": 8.544804573059082,
    "prediction_norm": 7.243601322174072,
    "timestamp": "2025-07-14T12:34:21.206152"
  },
  {
    "epoch": 1337,
    "batch_idx": 64,
    "loss": 0.21740135550498962,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 507.09375,
    "epsilon_norm": 8.40872573852539,
    "prediction_norm": 7.567336082458496,
    "timestamp": "2025-07-14T12:34:21.211345"
  },
  {
    "epoch": 1337,
    "batch_idx": 96,
    "loss": 0.242072194814682,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 519.5,
    "epsilon_norm": 6.833646297454834,
    "prediction_norm": 5.5873003005981445,
    "timestamp": "2025-07-14T12:34:21.216783"
  },
  {
    "epoch": 1337,
    "batch_idx": 128,
    "loss": 0.39980727434158325,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 490.3125,
    "epsilon_norm": 8.013008117675781,
    "prediction_norm": 6.440279006958008,
    "timestamp": "2025-07-14T12:34:21.221950"
  },
  {
    "epoch": 1337,
    "batch_idx": 160,
    "loss": 0.19353356957435608,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 477.15625,
    "epsilon_norm": 6.890319347381592,
    "prediction_norm": 5.982751369476318,
    "timestamp": "2025-07-14T12:34:21.227172"
  },
  {
    "epoch": 1337,
    "batch_idx": 192,
    "loss": 0.18715961277484894,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 522.5,
    "epsilon_norm": 4.40272855758667,
    "prediction_norm": 3.924988031387329,
    "timestamp": "2025-07-14T12:34:21.232268"
  },
  {
    "epoch": 1338,
    "batch_idx": 0,
    "loss": 0.3034094572067261,
    "learning_rate": 0.0002468325963234342,
    "timestep_mean": 500.375,
    "epsilon_norm": 7.562463760375977,
    "prediction_norm": 6.346436977386475,
    "timestamp": "2025-07-14T12:34:21.237511"
  },
  {
    "epoch": 1338,
    "batch_idx": 32,
    "loss": 0.18018190562725067,
    "learning_rate": 0.0002468325963234342,
    "timestep_mean": 505.375,
    "epsilon_norm": 8.327707290649414,
    "prediction_norm": 7.424315452575684,
    "timestamp": "2025-07-14T12:34:21.242742"
  },
  {
    "epoch": 1338,
    "batch_idx": 64,
    "loss": 0.27063214778900146,
    "learning_rate": 0.0002468325963234342,
    "timestep_mean": 422.4375,
    "epsilon_norm": 7.6177978515625,
    "prediction_norm": 6.394880294799805,
    "timestamp": "2025-07-14T12:34:21.247952"
  },
  {
    "epoch": 1338,
    "batch_idx": 96,
    "loss": 0.20867952704429626,
    "learning_rate": 0.0002468325963234342,
    "timestep_mean": 436.71875,
    "epsilon_norm": 7.674128532409668,
    "prediction_norm": 6.382267951965332,
    "timestamp": "2025-07-14T12:34:21.253160"
  },
  {
    "epoch": 1338,
    "batch_idx": 128,
    "loss": 0.176611989736557,
    "learning_rate": 0.0002468325963234342,
    "timestep_mean": 547.125,
    "epsilon_norm": 8.553462028503418,
    "prediction_norm": 7.735283374786377,
    "timestamp": "2025-07-14T12:34:21.258433"
  },
  {
    "epoch": 1338,
    "batch_idx": 160,
    "loss": 0.32157012820243835,
    "learning_rate": 0.0002468325963234342,
    "timestep_mean": 612.0,
    "epsilon_norm": 9.043251037597656,
    "prediction_norm": 8.07258129119873,
    "timestamp": "2025-07-14T12:34:21.263600"
  },
  {
    "epoch": 1338,
    "batch_idx": 192,
    "loss": 0.19807130098342896,
    "learning_rate": 0.0002468325963234342,
    "timestep_mean": 592.375,
    "epsilon_norm": 3.600998640060425,
    "prediction_norm": 2.6728262901306152,
    "timestamp": "2025-07-14T12:34:21.268738"
  },
  {
    "epoch": 1339,
    "batch_idx": 0,
    "loss": 0.37817561626434326,
    "learning_rate": 0.0002461556310562511,
    "timestep_mean": 490.75,
    "epsilon_norm": 8.899903297424316,
    "prediction_norm": 7.073536396026611,
    "timestamp": "2025-07-14T12:34:21.274032"
  },
  {
    "epoch": 1339,
    "batch_idx": 32,
    "loss": 0.17681753635406494,
    "learning_rate": 0.0002461556310562511,
    "timestep_mean": 531.46875,
    "epsilon_norm": 8.021659851074219,
    "prediction_norm": 7.812809944152832,
    "timestamp": "2025-07-14T12:34:21.279212"
  },
  {
    "epoch": 1339,
    "batch_idx": 64,
    "loss": 0.604714572429657,
    "learning_rate": 0.0002461556310562511,
    "timestep_mean": 461.53125,
    "epsilon_norm": 8.896953582763672,
    "prediction_norm": 5.995298862457275,
    "timestamp": "2025-07-14T12:34:21.284408"
  },
  {
    "epoch": 1339,
    "batch_idx": 96,
    "loss": 0.2444809228181839,
    "learning_rate": 0.0002461556310562511,
    "timestep_mean": 516.53125,
    "epsilon_norm": 8.295781135559082,
    "prediction_norm": 7.486260890960693,
    "timestamp": "2025-07-14T12:34:21.289603"
  },
  {
    "epoch": 1339,
    "batch_idx": 128,
    "loss": 0.19369962811470032,
    "learning_rate": 0.0002461556310562511,
    "timestep_mean": 499.65625,
    "epsilon_norm": 8.314884185791016,
    "prediction_norm": 7.425632953643799,
    "timestamp": "2025-07-14T12:34:21.294823"
  },
  {
    "epoch": 1339,
    "batch_idx": 160,
    "loss": 0.3426116406917572,
    "learning_rate": 0.0002461556310562511,
    "timestep_mean": 394.15625,
    "epsilon_norm": 6.958126068115234,
    "prediction_norm": 5.583066463470459,
    "timestamp": "2025-07-14T12:34:21.300103"
  },
  {
    "epoch": 1339,
    "batch_idx": 192,
    "loss": 0.2473086416721344,
    "learning_rate": 0.0002461556310562511,
    "timestep_mean": 479.625,
    "epsilon_norm": 3.0604782104492188,
    "prediction_norm": 3.073350429534912,
    "timestamp": "2025-07-14T12:34:21.305224"
  },
  {
    "epoch": 1340,
    "batch_idx": 0,
    "loss": 0.46364524960517883,
    "learning_rate": 0.0002454792921248147,
    "timestep_mean": 411.65625,
    "epsilon_norm": 8.684408187866211,
    "prediction_norm": 7.100180625915527,
    "timestamp": "2025-07-14T12:34:21.310433"
  },
  {
    "epoch": 1340,
    "batch_idx": 32,
    "loss": 0.34821802377700806,
    "learning_rate": 0.0002454792921248147,
    "timestep_mean": 436.15625,
    "epsilon_norm": 8.633220672607422,
    "prediction_norm": 7.821234226226807,
    "timestamp": "2025-07-14T12:34:21.315697"
  },
  {
    "epoch": 1340,
    "batch_idx": 64,
    "loss": 0.32340821623802185,
    "learning_rate": 0.0002454792921248147,
    "timestep_mean": 616.1875,
    "epsilon_norm": 9.290934562683105,
    "prediction_norm": 8.248775482177734,
    "timestamp": "2025-07-14T12:34:21.320865"
  },
  {
    "epoch": 1340,
    "batch_idx": 96,
    "loss": 0.285366415977478,
    "learning_rate": 0.0002454792921248147,
    "timestep_mean": 545.78125,
    "epsilon_norm": 9.085285186767578,
    "prediction_norm": 7.806457042694092,
    "timestamp": "2025-07-14T12:34:21.325892"
  },
  {
    "epoch": 1340,
    "batch_idx": 128,
    "loss": 0.42013752460479736,
    "learning_rate": 0.0002454792921248147,
    "timestep_mean": 391.625,
    "epsilon_norm": 7.196383476257324,
    "prediction_norm": 5.221867084503174,
    "timestamp": "2025-07-14T12:34:21.331018"
  },
  {
    "epoch": 1340,
    "batch_idx": 160,
    "loss": 0.4629930853843689,
    "learning_rate": 0.0002454792921248147,
    "timestep_mean": 476.9375,
    "epsilon_norm": 8.984583854675293,
    "prediction_norm": 7.161506175994873,
    "timestamp": "2025-07-14T12:34:21.336209"
  },
  {
    "epoch": 1340,
    "batch_idx": 192,
    "loss": 0.3096410036087036,
    "learning_rate": 0.0002454792921248147,
    "timestep_mean": 411.0,
    "epsilon_norm": 3.322737455368042,
    "prediction_norm": 2.5145344734191895,
    "timestamp": "2025-07-14T12:34:21.341438"
  },
  {
    "epoch": 1341,
    "batch_idx": 0,
    "loss": 0.25503379106521606,
    "learning_rate": 0.00024480358119792375,
    "timestep_mean": 552.8125,
    "epsilon_norm": 8.93375301361084,
    "prediction_norm": 7.212430000305176,
    "timestamp": "2025-07-14T12:34:21.346807"
  },
  {
    "epoch": 1341,
    "batch_idx": 32,
    "loss": 0.34756797552108765,
    "learning_rate": 0.00024480358119792375,
    "timestep_mean": 535.46875,
    "epsilon_norm": 7.925502777099609,
    "prediction_norm": 6.648284912109375,
    "timestamp": "2025-07-14T12:34:21.351969"
  },
  {
    "epoch": 1341,
    "batch_idx": 64,
    "loss": 0.4912295937538147,
    "learning_rate": 0.00024480358119792375,
    "timestep_mean": 448.96875,
    "epsilon_norm": 9.955471992492676,
    "prediction_norm": 7.883037090301514,
    "timestamp": "2025-07-14T12:34:21.357166"
  },
  {
    "epoch": 1341,
    "batch_idx": 96,
    "loss": 0.3719657063484192,
    "learning_rate": 0.00024480358119792375,
    "timestep_mean": 480.8125,
    "epsilon_norm": 6.633453369140625,
    "prediction_norm": 5.125400066375732,
    "timestamp": "2025-07-14T12:34:21.362323"
  },
  {
    "epoch": 1341,
    "batch_idx": 128,
    "loss": 0.1738692820072174,
    "learning_rate": 0.00024480358119792375,
    "timestep_mean": 510.28125,
    "epsilon_norm": 6.542875289916992,
    "prediction_norm": 5.875683784484863,
    "timestamp": "2025-07-14T12:34:21.367601"
  },
  {
    "epoch": 1341,
    "batch_idx": 160,
    "loss": 0.31034988164901733,
    "learning_rate": 0.00024480358119792375,
    "timestep_mean": 449.28125,
    "epsilon_norm": 7.5642313957214355,
    "prediction_norm": 6.172010898590088,
    "timestamp": "2025-07-14T12:34:21.372831"
  },
  {
    "epoch": 1341,
    "batch_idx": 192,
    "loss": 0.241184264421463,
    "learning_rate": 0.00024480358119792375,
    "timestep_mean": 513.25,
    "epsilon_norm": 3.29109787940979,
    "prediction_norm": 2.201927661895752,
    "timestamp": "2025-07-14T12:34:21.378028"
  },
  {
    "epoch": 1342,
    "batch_idx": 0,
    "loss": 0.18835848569869995,
    "learning_rate": 0.0002441284999428279,
    "timestep_mean": 466.0,
    "epsilon_norm": 7.609839916229248,
    "prediction_norm": 6.612412452697754,
    "timestamp": "2025-07-14T12:34:21.383403"
  },
  {
    "epoch": 1342,
    "batch_idx": 32,
    "loss": 0.2797510027885437,
    "learning_rate": 0.0002441284999428279,
    "timestep_mean": 445.90625,
    "epsilon_norm": 7.591912746429443,
    "prediction_norm": 6.453458786010742,
    "timestamp": "2025-07-14T12:34:21.388604"
  },
  {
    "epoch": 1342,
    "batch_idx": 64,
    "loss": 0.3807489275932312,
    "learning_rate": 0.0002441284999428279,
    "timestep_mean": 446.625,
    "epsilon_norm": 8.043539047241211,
    "prediction_norm": 5.946889877319336,
    "timestamp": "2025-07-14T12:34:21.393795"
  },
  {
    "epoch": 1342,
    "batch_idx": 96,
    "loss": 0.35499459505081177,
    "learning_rate": 0.0002441284999428279,
    "timestep_mean": 480.53125,
    "epsilon_norm": 7.882680892944336,
    "prediction_norm": 6.341928482055664,
    "timestamp": "2025-07-14T12:34:21.399099"
  },
  {
    "epoch": 1342,
    "batch_idx": 128,
    "loss": 0.45675572752952576,
    "learning_rate": 0.0002441284999428279,
    "timestep_mean": 490.03125,
    "epsilon_norm": 7.772331714630127,
    "prediction_norm": 5.64321756362915,
    "timestamp": "2025-07-14T12:34:21.404157"
  },
  {
    "epoch": 1342,
    "batch_idx": 160,
    "loss": 0.31452876329421997,
    "learning_rate": 0.0002441284999428279,
    "timestep_mean": 519.59375,
    "epsilon_norm": 7.935083866119385,
    "prediction_norm": 6.564136028289795,
    "timestamp": "2025-07-14T12:34:21.409144"
  },
  {
    "epoch": 1342,
    "batch_idx": 192,
    "loss": 0.22876876592636108,
    "learning_rate": 0.0002441284999428279,
    "timestep_mean": 514.0,
    "epsilon_norm": 4.577445030212402,
    "prediction_norm": 3.465040922164917,
    "timestamp": "2025-07-14T12:34:21.414121"
  },
  {
    "epoch": 1343,
    "batch_idx": 0,
    "loss": 0.20342102646827698,
    "learning_rate": 0.00024345405002522302,
    "timestep_mean": 423.03125,
    "epsilon_norm": 8.931449890136719,
    "prediction_norm": 8.6078462600708,
    "timestamp": "2025-07-14T12:34:21.419215"
  },
  {
    "epoch": 1343,
    "batch_idx": 32,
    "loss": 0.3214465379714966,
    "learning_rate": 0.00024345405002522302,
    "timestep_mean": 531.03125,
    "epsilon_norm": 9.15533447265625,
    "prediction_norm": 7.325591087341309,
    "timestamp": "2025-07-14T12:34:21.424456"
  },
  {
    "epoch": 1343,
    "batch_idx": 64,
    "loss": 0.2324739545583725,
    "learning_rate": 0.00024345405002522302,
    "timestep_mean": 470.1875,
    "epsilon_norm": 9.29951000213623,
    "prediction_norm": 7.952080249786377,
    "timestamp": "2025-07-14T12:34:21.429804"
  },
  {
    "epoch": 1343,
    "batch_idx": 96,
    "loss": 0.2586011290550232,
    "learning_rate": 0.00024345405002522302,
    "timestep_mean": 482.65625,
    "epsilon_norm": 7.54583215713501,
    "prediction_norm": 6.104974269866943,
    "timestamp": "2025-07-14T12:34:21.434988"
  },
  {
    "epoch": 1343,
    "batch_idx": 128,
    "loss": 0.23641332983970642,
    "learning_rate": 0.00024345405002522302,
    "timestep_mean": 468.40625,
    "epsilon_norm": 8.024191856384277,
    "prediction_norm": 6.55158805847168,
    "timestamp": "2025-07-14T12:34:21.440184"
  },
  {
    "epoch": 1343,
    "batch_idx": 160,
    "loss": 0.43470075726509094,
    "learning_rate": 0.00024345405002522302,
    "timestep_mean": 426.5,
    "epsilon_norm": 7.553986549377441,
    "prediction_norm": 5.784981727600098,
    "timestamp": "2025-07-14T12:34:21.445378"
  },
  {
    "epoch": 1343,
    "batch_idx": 192,
    "loss": 0.5221231579780579,
    "learning_rate": 0.00024345405002522302,
    "timestep_mean": 480.375,
    "epsilon_norm": 3.4771482944488525,
    "prediction_norm": 2.650540828704834,
    "timestamp": "2025-07-14T12:34:21.450771"
  },
  {
    "epoch": 1344,
    "batch_idx": 0,
    "loss": 0.16731971502304077,
    "learning_rate": 0.000242780233109247,
    "timestep_mean": 537.25,
    "epsilon_norm": 7.960211277008057,
    "prediction_norm": 7.45387601852417,
    "timestamp": "2025-07-14T12:34:21.455926"
  },
  {
    "epoch": 1344,
    "batch_idx": 32,
    "loss": 0.2766995131969452,
    "learning_rate": 0.000242780233109247,
    "timestep_mean": 505.4375,
    "epsilon_norm": 8.968311309814453,
    "prediction_norm": 7.69765043258667,
    "timestamp": "2025-07-14T12:34:21.460958"
  },
  {
    "epoch": 1344,
    "batch_idx": 64,
    "loss": 0.22446785867214203,
    "learning_rate": 0.000242780233109247,
    "timestep_mean": 524.375,
    "epsilon_norm": 8.5399808883667,
    "prediction_norm": 7.822550296783447,
    "timestamp": "2025-07-14T12:34:21.465906"
  },
  {
    "epoch": 1344,
    "batch_idx": 96,
    "loss": 0.41760528087615967,
    "learning_rate": 0.000242780233109247,
    "timestep_mean": 438.6875,
    "epsilon_norm": 8.846511840820312,
    "prediction_norm": 6.5942769050598145,
    "timestamp": "2025-07-14T12:34:21.470738"
  },
  {
    "epoch": 1344,
    "batch_idx": 128,
    "loss": 0.3619763255119324,
    "learning_rate": 0.000242780233109247,
    "timestep_mean": 499.34375,
    "epsilon_norm": 8.575129508972168,
    "prediction_norm": 6.793651580810547,
    "timestamp": "2025-07-14T12:34:21.475625"
  },
  {
    "epoch": 1344,
    "batch_idx": 160,
    "loss": 0.1384846568107605,
    "learning_rate": 0.000242780233109247,
    "timestep_mean": 554.28125,
    "epsilon_norm": 8.54416561126709,
    "prediction_norm": 7.6165876388549805,
    "timestamp": "2025-07-14T12:34:21.480466"
  },
  {
    "epoch": 1344,
    "batch_idx": 192,
    "loss": 0.21240103244781494,
    "learning_rate": 0.000242780233109247,
    "timestep_mean": 540.875,
    "epsilon_norm": 3.395224094390869,
    "prediction_norm": 2.250945568084717,
    "timestamp": "2025-07-14T12:34:21.485307"
  },
  {
    "epoch": 1345,
    "batch_idx": 0,
    "loss": 0.47357362508773804,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 439.40625,
    "epsilon_norm": 7.641485214233398,
    "prediction_norm": 5.608308792114258,
    "timestamp": "2025-07-14T12:34:21.490328"
  },
  {
    "epoch": 1345,
    "batch_idx": 32,
    "loss": 0.10446441173553467,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 622.9375,
    "epsilon_norm": 7.538137435913086,
    "prediction_norm": 6.861055374145508,
    "timestamp": "2025-07-14T12:34:21.495541"
  },
  {
    "epoch": 1345,
    "batch_idx": 64,
    "loss": 0.3454602360725403,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 500.8125,
    "epsilon_norm": 7.378894805908203,
    "prediction_norm": 5.68184757232666,
    "timestamp": "2025-07-14T12:34:21.500739"
  },
  {
    "epoch": 1345,
    "batch_idx": 96,
    "loss": 0.5195760130882263,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 489.78125,
    "epsilon_norm": 9.44566535949707,
    "prediction_norm": 7.359731674194336,
    "timestamp": "2025-07-14T12:34:21.505782"
  },
  {
    "epoch": 1345,
    "batch_idx": 128,
    "loss": 0.2846546769142151,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 500.84375,
    "epsilon_norm": 7.423825263977051,
    "prediction_norm": 6.357394218444824,
    "timestamp": "2025-07-14T12:34:21.510908"
  },
  {
    "epoch": 1345,
    "batch_idx": 160,
    "loss": 0.31936115026474,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 569.125,
    "epsilon_norm": 8.425464630126953,
    "prediction_norm": 7.2646074295043945,
    "timestamp": "2025-07-14T12:34:21.516091"
  },
  {
    "epoch": 1345,
    "batch_idx": 192,
    "loss": 0.07533273100852966,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 471.75,
    "epsilon_norm": 3.949286460876465,
    "prediction_norm": 3.8869752883911133,
    "timestamp": "2025-07-14T12:34:21.521359"
  },
  {
    "epoch": 1346,
    "batch_idx": 0,
    "loss": 0.32207050919532776,
    "learning_rate": 0.0002414345049309217,
    "timestep_mean": 523.59375,
    "epsilon_norm": 7.123641490936279,
    "prediction_norm": 5.989088535308838,
    "timestamp": "2025-07-14T12:34:21.526380"
  },
  {
    "epoch": 1346,
    "batch_idx": 32,
    "loss": 0.24296195805072784,
    "learning_rate": 0.0002414345049309217,
    "timestep_mean": 502.125,
    "epsilon_norm": 9.324498176574707,
    "prediction_norm": 8.423493385314941,
    "timestamp": "2025-07-14T12:34:21.531417"
  },
  {
    "epoch": 1346,
    "batch_idx": 64,
    "loss": 0.18924295902252197,
    "learning_rate": 0.0002414345049309217,
    "timestep_mean": 548.8125,
    "epsilon_norm": 7.297212600708008,
    "prediction_norm": 6.2538161277771,
    "timestamp": "2025-07-14T12:34:21.536645"
  },
  {
    "epoch": 1346,
    "batch_idx": 96,
    "loss": 0.4586501121520996,
    "learning_rate": 0.0002414345049309217,
    "timestep_mean": 451.21875,
    "epsilon_norm": 8.743024826049805,
    "prediction_norm": 6.914772987365723,
    "timestamp": "2025-07-14T12:34:21.541934"
  },
  {
    "epoch": 1346,
    "batch_idx": 128,
    "loss": 0.3846944570541382,
    "learning_rate": 0.0002414345049309217,
    "timestep_mean": 496.40625,
    "epsilon_norm": 7.980383396148682,
    "prediction_norm": 6.314620018005371,
    "timestamp": "2025-07-14T12:34:21.546863"
  },
  {
    "epoch": 1346,
    "batch_idx": 160,
    "loss": 0.24178850650787354,
    "learning_rate": 0.0002414345049309217,
    "timestep_mean": 501.3125,
    "epsilon_norm": 8.105250358581543,
    "prediction_norm": 6.541369438171387,
    "timestamp": "2025-07-14T12:34:21.551927"
  },
  {
    "epoch": 1346,
    "batch_idx": 192,
    "loss": 0.39374127984046936,
    "learning_rate": 0.0002414345049309217,
    "timestep_mean": 465.125,
    "epsilon_norm": 4.333470344543457,
    "prediction_norm": 3.3506627082824707,
    "timestamp": "2025-07-14T12:34:21.557054"
  },
  {
    "epoch": 1347,
    "batch_idx": 0,
    "loss": 0.31510111689567566,
    "learning_rate": 0.00024076259698902258,
    "timestep_mean": 484.34375,
    "epsilon_norm": 7.557987689971924,
    "prediction_norm": 6.403543949127197,
    "timestamp": "2025-07-14T12:34:21.562351"
  },
  {
    "epoch": 1347,
    "batch_idx": 32,
    "loss": 0.18676148355007172,
    "learning_rate": 0.00024076259698902258,
    "timestep_mean": 493.6875,
    "epsilon_norm": 7.326845169067383,
    "prediction_norm": 6.351301193237305,
    "timestamp": "2025-07-14T12:34:21.567618"
  },
  {
    "epoch": 1347,
    "batch_idx": 64,
    "loss": 0.27022674679756165,
    "learning_rate": 0.00024076259698902258,
    "timestep_mean": 448.59375,
    "epsilon_norm": 8.487253189086914,
    "prediction_norm": 6.878105640411377,
    "timestamp": "2025-07-14T12:34:21.572806"
  },
  {
    "epoch": 1347,
    "batch_idx": 96,
    "loss": 0.2689220905303955,
    "learning_rate": 0.00024076259698902258,
    "timestep_mean": 428.3125,
    "epsilon_norm": 8.844293594360352,
    "prediction_norm": 7.52427339553833,
    "timestamp": "2025-07-14T12:34:21.578000"
  },
  {
    "epoch": 1347,
    "batch_idx": 128,
    "loss": 0.18377386033535004,
    "learning_rate": 0.00024076259698902258,
    "timestep_mean": 521.0625,
    "epsilon_norm": 8.75360107421875,
    "prediction_norm": 8.099960327148438,
    "timestamp": "2025-07-14T12:34:21.583379"
  },
  {
    "epoch": 1347,
    "batch_idx": 160,
    "loss": 0.19231101870536804,
    "learning_rate": 0.00024076259698902258,
    "timestep_mean": 537.125,
    "epsilon_norm": 8.430273056030273,
    "prediction_norm": 7.942395210266113,
    "timestamp": "2025-07-14T12:34:21.588553"
  },
  {
    "epoch": 1347,
    "batch_idx": 192,
    "loss": 0.06724131852388382,
    "learning_rate": 0.00024076259698902258,
    "timestep_mean": 691.25,
    "epsilon_norm": 4.288055896759033,
    "prediction_norm": 3.8620827198028564,
    "timestamp": "2025-07-14T12:34:21.593593"
  },
  {
    "epoch": 1348,
    "batch_idx": 0,
    "loss": 0.2509686052799225,
    "learning_rate": 0.00024009132868964568,
    "timestep_mean": 534.71875,
    "epsilon_norm": 8.21360969543457,
    "prediction_norm": 6.855930328369141,
    "timestamp": "2025-07-14T12:34:21.598720"
  },
  {
    "epoch": 1348,
    "batch_idx": 32,
    "loss": 0.3876628279685974,
    "learning_rate": 0.00024009132868964568,
    "timestep_mean": 498.375,
    "epsilon_norm": 7.25234317779541,
    "prediction_norm": 5.805969715118408,
    "timestamp": "2025-07-14T12:34:21.603848"
  },
  {
    "epoch": 1348,
    "batch_idx": 64,
    "loss": 0.18783092498779297,
    "learning_rate": 0.00024009132868964568,
    "timestep_mean": 523.96875,
    "epsilon_norm": 7.786746025085449,
    "prediction_norm": 6.530949592590332,
    "timestamp": "2025-07-14T12:34:21.609056"
  },
  {
    "epoch": 1348,
    "batch_idx": 96,
    "loss": 0.2512732744216919,
    "learning_rate": 0.00024009132868964568,
    "timestep_mean": 450.65625,
    "epsilon_norm": 7.650362968444824,
    "prediction_norm": 6.730532169342041,
    "timestamp": "2025-07-14T12:34:21.614184"
  },
  {
    "epoch": 1348,
    "batch_idx": 128,
    "loss": 0.23119044303894043,
    "learning_rate": 0.00024009132868964568,
    "timestep_mean": 482.40625,
    "epsilon_norm": 6.716156959533691,
    "prediction_norm": 5.840714454650879,
    "timestamp": "2025-07-14T12:34:21.619469"
  },
  {
    "epoch": 1348,
    "batch_idx": 160,
    "loss": 0.36027950048446655,
    "learning_rate": 0.00024009132868964568,
    "timestep_mean": 437.375,
    "epsilon_norm": 7.725244045257568,
    "prediction_norm": 6.0896687507629395,
    "timestamp": "2025-07-14T12:34:21.624734"
  },
  {
    "epoch": 1348,
    "batch_idx": 192,
    "loss": 0.5363362431526184,
    "learning_rate": 0.00024009132868964568,
    "timestep_mean": 499.75,
    "epsilon_norm": 4.8416900634765625,
    "prediction_norm": 3.7175650596618652,
    "timestamp": "2025-07-14T12:34:21.629883"
  },
  {
    "epoch": 1349,
    "batch_idx": 0,
    "loss": 0.2427811622619629,
    "learning_rate": 0.0002394207016890781,
    "timestep_mean": 539.0625,
    "epsilon_norm": 8.304567337036133,
    "prediction_norm": 7.256284713745117,
    "timestamp": "2025-07-14T12:34:21.635004"
  },
  {
    "epoch": 1349,
    "batch_idx": 32,
    "loss": 0.1418556421995163,
    "learning_rate": 0.0002394207016890781,
    "timestep_mean": 535.8125,
    "epsilon_norm": 8.62033748626709,
    "prediction_norm": 7.97828483581543,
    "timestamp": "2025-07-14T12:34:21.640150"
  },
  {
    "epoch": 1349,
    "batch_idx": 64,
    "loss": 0.21037574112415314,
    "learning_rate": 0.0002394207016890781,
    "timestep_mean": 512.375,
    "epsilon_norm": 7.2718186378479,
    "prediction_norm": 6.622316360473633,
    "timestamp": "2025-07-14T12:34:21.645292"
  },
  {
    "epoch": 1349,
    "batch_idx": 96,
    "loss": 0.3431120216846466,
    "learning_rate": 0.0002394207016890781,
    "timestep_mean": 455.3125,
    "epsilon_norm": 8.517915725708008,
    "prediction_norm": 6.556156158447266,
    "timestamp": "2025-07-14T12:34:21.650489"
  },
  {
    "epoch": 1349,
    "batch_idx": 128,
    "loss": 0.12641748785972595,
    "learning_rate": 0.0002394207016890781,
    "timestep_mean": 518.21875,
    "epsilon_norm": 6.17622709274292,
    "prediction_norm": 5.501677513122559,
    "timestamp": "2025-07-14T12:34:21.655678"
  },
  {
    "epoch": 1349,
    "batch_idx": 160,
    "loss": 0.31581491231918335,
    "learning_rate": 0.0002394207016890781,
    "timestep_mean": 493.28125,
    "epsilon_norm": 7.337238788604736,
    "prediction_norm": 5.94793176651001,
    "timestamp": "2025-07-14T12:34:21.660866"
  },
  {
    "epoch": 1349,
    "batch_idx": 192,
    "loss": 0.10469328612089157,
    "learning_rate": 0.0002394207016890781,
    "timestep_mean": 512.75,
    "epsilon_norm": 4.298933506011963,
    "prediction_norm": 3.8413026332855225,
    "timestamp": "2025-07-14T12:34:21.665959"
  },
  {
    "epoch": 1350,
    "batch_idx": 0,
    "loss": 0.15076108276844025,
    "learning_rate": 0.0002387507176420259,
    "timestep_mean": 534.03125,
    "epsilon_norm": 8.133759498596191,
    "prediction_norm": 7.310326099395752,
    "timestamp": "2025-07-14T12:34:21.671186"
  },
  {
    "epoch": 1350,
    "batch_idx": 32,
    "loss": 0.18239617347717285,
    "learning_rate": 0.0002387507176420259,
    "timestep_mean": 583.0,
    "epsilon_norm": 7.4243292808532715,
    "prediction_norm": 6.705044746398926,
    "timestamp": "2025-07-14T12:34:21.676263"
  },
  {
    "epoch": 1350,
    "batch_idx": 64,
    "loss": 0.245811328291893,
    "learning_rate": 0.0002387507176420259,
    "timestep_mean": 468.34375,
    "epsilon_norm": 7.162786960601807,
    "prediction_norm": 6.649322509765625,
    "timestamp": "2025-07-14T12:34:21.681453"
  },
  {
    "epoch": 1350,
    "batch_idx": 96,
    "loss": 0.19657349586486816,
    "learning_rate": 0.0002387507176420259,
    "timestep_mean": 466.09375,
    "epsilon_norm": 7.661855697631836,
    "prediction_norm": 7.597207069396973,
    "timestamp": "2025-07-14T12:34:21.686639"
  },
  {
    "epoch": 1350,
    "batch_idx": 128,
    "loss": 0.33439356088638306,
    "learning_rate": 0.0002387507176420259,
    "timestep_mean": 550.0,
    "epsilon_norm": 7.314298152923584,
    "prediction_norm": 6.328273773193359,
    "timestamp": "2025-07-14T12:34:21.692031"
  },
  {
    "epoch": 1350,
    "batch_idx": 160,
    "loss": 0.3000987768173218,
    "learning_rate": 0.0002387507176420259,
    "timestep_mean": 497.40625,
    "epsilon_norm": 7.775201320648193,
    "prediction_norm": 6.184471607208252,
    "timestamp": "2025-07-14T12:34:21.697208"
  },
  {
    "epoch": 1350,
    "batch_idx": 192,
    "loss": 0.8502779603004456,
    "learning_rate": 0.0002387507176420259,
    "timestep_mean": 366.75,
    "epsilon_norm": 5.7443647384643555,
    "prediction_norm": 4.485824108123779,
    "timestamp": "2025-07-14T12:34:21.702360"
  },
  {
    "epoch": 1351,
    "batch_idx": 0,
    "loss": 0.3038427233695984,
    "learning_rate": 0.00023808137820160788,
    "timestep_mean": 487.15625,
    "epsilon_norm": 8.50556755065918,
    "prediction_norm": 7.5026535987854,
    "timestamp": "2025-07-14T12:34:21.707593"
  },
  {
    "epoch": 1351,
    "batch_idx": 32,
    "loss": 0.42434316873550415,
    "learning_rate": 0.00023808137820160788,
    "timestep_mean": 481.65625,
    "epsilon_norm": 9.168618202209473,
    "prediction_norm": 7.573697090148926,
    "timestamp": "2025-07-14T12:34:21.712806"
  },
  {
    "epoch": 1351,
    "batch_idx": 64,
    "loss": 0.24928522109985352,
    "learning_rate": 0.00023808137820160788,
    "timestep_mean": 553.09375,
    "epsilon_norm": 8.114984512329102,
    "prediction_norm": 7.434746265411377,
    "timestamp": "2025-07-14T12:34:21.717976"
  },
  {
    "epoch": 1351,
    "batch_idx": 96,
    "loss": 0.1082131564617157,
    "learning_rate": 0.00023808137820160788,
    "timestep_mean": 557.34375,
    "epsilon_norm": 6.769825458526611,
    "prediction_norm": 5.822876453399658,
    "timestamp": "2025-07-14T12:34:21.723246"
  },
  {
    "epoch": 1351,
    "batch_idx": 128,
    "loss": 0.42891180515289307,
    "learning_rate": 0.00023808137820160788,
    "timestep_mean": 489.25,
    "epsilon_norm": 8.687163352966309,
    "prediction_norm": 6.665730953216553,
    "timestamp": "2025-07-14T12:34:21.728372"
  },
  {
    "epoch": 1351,
    "batch_idx": 160,
    "loss": 0.24449463188648224,
    "learning_rate": 0.00023808137820160788,
    "timestep_mean": 500.625,
    "epsilon_norm": 7.731782913208008,
    "prediction_norm": 5.980678081512451,
    "timestamp": "2025-07-14T12:34:21.733648"
  },
  {
    "epoch": 1351,
    "batch_idx": 192,
    "loss": 0.1898638904094696,
    "learning_rate": 0.00023808137820160788,
    "timestep_mean": 518.125,
    "epsilon_norm": 2.507878303527832,
    "prediction_norm": 1.6647214889526367,
    "timestamp": "2025-07-14T12:34:21.738793"
  },
  {
    "epoch": 1352,
    "batch_idx": 0,
    "loss": 0.5341764688491821,
    "learning_rate": 0.0002374126850193524,
    "timestep_mean": 499.28125,
    "epsilon_norm": 8.318326950073242,
    "prediction_norm": 6.346395492553711,
    "timestamp": "2025-07-14T12:34:21.744010"
  },
  {
    "epoch": 1352,
    "batch_idx": 32,
    "loss": 0.4008468687534332,
    "learning_rate": 0.0002374126850193524,
    "timestep_mean": 461.34375,
    "epsilon_norm": 7.871438503265381,
    "prediction_norm": 6.44758415222168,
    "timestamp": "2025-07-14T12:34:21.749220"
  },
  {
    "epoch": 1352,
    "batch_idx": 64,
    "loss": 0.26081597805023193,
    "learning_rate": 0.0002374126850193524,
    "timestep_mean": 493.21875,
    "epsilon_norm": 7.444338798522949,
    "prediction_norm": 5.880584239959717,
    "timestamp": "2025-07-14T12:34:21.754412"
  },
  {
    "epoch": 1352,
    "batch_idx": 96,
    "loss": 0.3359408378601074,
    "learning_rate": 0.0002374126850193524,
    "timestep_mean": 500.5,
    "epsilon_norm": 6.856078147888184,
    "prediction_norm": 5.677597999572754,
    "timestamp": "2025-07-14T12:34:21.759568"
  },
  {
    "epoch": 1352,
    "batch_idx": 128,
    "loss": 0.46520116925239563,
    "learning_rate": 0.0002374126850193524,
    "timestep_mean": 497.9375,
    "epsilon_norm": 8.373480796813965,
    "prediction_norm": 6.483420372009277,
    "timestamp": "2025-07-14T12:34:21.764777"
  },
  {
    "epoch": 1352,
    "batch_idx": 160,
    "loss": 0.21882671117782593,
    "learning_rate": 0.0002374126850193524,
    "timestep_mean": 531.28125,
    "epsilon_norm": 7.917629241943359,
    "prediction_norm": 6.623180389404297,
    "timestamp": "2025-07-14T12:34:21.770012"
  },
  {
    "epoch": 1352,
    "batch_idx": 192,
    "loss": 0.46212223172187805,
    "learning_rate": 0.0002374126850193524,
    "timestep_mean": 387.5,
    "epsilon_norm": 3.9314863681793213,
    "prediction_norm": 3.843294858932495,
    "timestamp": "2025-07-14T12:34:21.775376"
  },
  {
    "epoch": 1353,
    "batch_idx": 0,
    "loss": 0.4571649432182312,
    "learning_rate": 0.000236744639745194,
    "timestep_mean": 463.65625,
    "epsilon_norm": 8.420441627502441,
    "prediction_norm": 6.480355262756348,
    "timestamp": "2025-07-14T12:34:21.780554"
  },
  {
    "epoch": 1353,
    "batch_idx": 32,
    "loss": 0.2549710273742676,
    "learning_rate": 0.000236744639745194,
    "timestep_mean": 443.3125,
    "epsilon_norm": 7.634417533874512,
    "prediction_norm": 7.071836948394775,
    "timestamp": "2025-07-14T12:34:21.785727"
  },
  {
    "epoch": 1353,
    "batch_idx": 64,
    "loss": 0.21638676524162292,
    "learning_rate": 0.000236744639745194,
    "timestep_mean": 455.3125,
    "epsilon_norm": 7.062628746032715,
    "prediction_norm": 5.952145099639893,
    "timestamp": "2025-07-14T12:34:21.790988"
  },
  {
    "epoch": 1353,
    "batch_idx": 96,
    "loss": 0.27839213609695435,
    "learning_rate": 0.000236744639745194,
    "timestep_mean": 509.5,
    "epsilon_norm": 7.569972991943359,
    "prediction_norm": 6.5060038566589355,
    "timestamp": "2025-07-14T12:34:21.796164"
  },
  {
    "epoch": 1353,
    "batch_idx": 128,
    "loss": 0.07067304849624634,
    "learning_rate": 0.000236744639745194,
    "timestep_mean": 624.53125,
    "epsilon_norm": 7.703387260437012,
    "prediction_norm": 7.173898220062256,
    "timestamp": "2025-07-14T12:34:21.801411"
  },
  {
    "epoch": 1353,
    "batch_idx": 160,
    "loss": 0.2780216336250305,
    "learning_rate": 0.000236744639745194,
    "timestep_mean": 455.15625,
    "epsilon_norm": 6.782446384429932,
    "prediction_norm": 5.51399564743042,
    "timestamp": "2025-07-14T12:34:21.806519"
  },
  {
    "epoch": 1353,
    "batch_idx": 192,
    "loss": 0.058649517595767975,
    "learning_rate": 0.000236744639745194,
    "timestep_mean": 463.25,
    "epsilon_norm": 3.5416295528411865,
    "prediction_norm": 3.5941638946533203,
    "timestamp": "2025-07-14T12:34:21.811622"
  },
  {
    "epoch": 1354,
    "batch_idx": 0,
    "loss": 0.2746536135673523,
    "learning_rate": 0.00023607724402746716,
    "timestep_mean": 470.84375,
    "epsilon_norm": 7.524539947509766,
    "prediction_norm": 5.843266487121582,
    "timestamp": "2025-07-14T12:34:21.817022"
  },
  {
    "epoch": 1354,
    "batch_idx": 32,
    "loss": 0.14914213120937347,
    "learning_rate": 0.00023607724402746716,
    "timestep_mean": 518.75,
    "epsilon_norm": 8.39852523803711,
    "prediction_norm": 8.474847793579102,
    "timestamp": "2025-07-14T12:34:21.822125"
  },
  {
    "epoch": 1354,
    "batch_idx": 64,
    "loss": 0.23359867930412292,
    "learning_rate": 0.00023607724402746716,
    "timestep_mean": 487.0625,
    "epsilon_norm": 7.981902599334717,
    "prediction_norm": 6.575765132904053,
    "timestamp": "2025-07-14T12:34:21.827323"
  },
  {
    "epoch": 1354,
    "batch_idx": 96,
    "loss": 0.259130597114563,
    "learning_rate": 0.00023607724402746716,
    "timestep_mean": 508.125,
    "epsilon_norm": 8.482029914855957,
    "prediction_norm": 7.103673458099365,
    "timestamp": "2025-07-14T12:34:21.832530"
  },
  {
    "epoch": 1354,
    "batch_idx": 128,
    "loss": 0.2902684211730957,
    "learning_rate": 0.00023607724402746716,
    "timestep_mean": 516.625,
    "epsilon_norm": 8.331855773925781,
    "prediction_norm": 6.349330902099609,
    "timestamp": "2025-07-14T12:34:21.837765"
  },
  {
    "epoch": 1354,
    "batch_idx": 160,
    "loss": 0.2625077962875366,
    "learning_rate": 0.00023607724402746716,
    "timestep_mean": 463.96875,
    "epsilon_norm": 8.62807559967041,
    "prediction_norm": 7.43684720993042,
    "timestamp": "2025-07-14T12:34:21.842985"
  },
  {
    "epoch": 1354,
    "batch_idx": 192,
    "loss": 0.10884202271699905,
    "learning_rate": 0.00023607724402746716,
    "timestep_mean": 569.125,
    "epsilon_norm": 3.73215651512146,
    "prediction_norm": 3.309159994125366,
    "timestamp": "2025-07-14T12:34:21.848106"
  },
  {
    "epoch": 1355,
    "batch_idx": 0,
    "loss": 0.20811817049980164,
    "learning_rate": 0.00023541049951290505,
    "timestep_mean": 514.5,
    "epsilon_norm": 7.345460891723633,
    "prediction_norm": 6.063339710235596,
    "timestamp": "2025-07-14T12:34:21.853337"
  },
  {
    "epoch": 1355,
    "batch_idx": 32,
    "loss": 0.24464847147464752,
    "learning_rate": 0.00023541049951290505,
    "timestep_mean": 484.28125,
    "epsilon_norm": 7.517321586608887,
    "prediction_norm": 6.493549346923828,
    "timestamp": "2025-07-14T12:34:21.858634"
  },
  {
    "epoch": 1355,
    "batch_idx": 64,
    "loss": 0.17638295888900757,
    "learning_rate": 0.00023541049951290505,
    "timestep_mean": 547.96875,
    "epsilon_norm": 7.532465934753418,
    "prediction_norm": 6.565358638763428,
    "timestamp": "2025-07-14T12:34:21.863876"
  },
  {
    "epoch": 1355,
    "batch_idx": 96,
    "loss": 0.3338044583797455,
    "learning_rate": 0.00023541049951290505,
    "timestep_mean": 492.6875,
    "epsilon_norm": 8.10293197631836,
    "prediction_norm": 6.6817731857299805,
    "timestamp": "2025-07-14T12:34:21.868986"
  },
  {
    "epoch": 1355,
    "batch_idx": 128,
    "loss": 0.27544674277305603,
    "learning_rate": 0.00023541049951290505,
    "timestep_mean": 531.125,
    "epsilon_norm": 7.555838108062744,
    "prediction_norm": 6.186949253082275,
    "timestamp": "2025-07-14T12:34:21.874255"
  },
  {
    "epoch": 1355,
    "batch_idx": 160,
    "loss": 0.20839044451713562,
    "learning_rate": 0.00023541049951290505,
    "timestep_mean": 501.34375,
    "epsilon_norm": 7.051196098327637,
    "prediction_norm": 6.013532638549805,
    "timestamp": "2025-07-14T12:34:21.879414"
  },
  {
    "epoch": 1355,
    "batch_idx": 192,
    "loss": 0.4163704514503479,
    "learning_rate": 0.00023541049951290505,
    "timestep_mean": 295.75,
    "epsilon_norm": 3.9503986835479736,
    "prediction_norm": 2.846341133117676,
    "timestamp": "2025-07-14T12:34:21.884585"
  },
  {
    "epoch": 1356,
    "batch_idx": 0,
    "loss": 0.4783972203731537,
    "learning_rate": 0.00023474440784663342,
    "timestep_mean": 462.53125,
    "epsilon_norm": 8.314181327819824,
    "prediction_norm": 5.9820876121521,
    "timestamp": "2025-07-14T12:34:21.889717"
  },
  {
    "epoch": 1356,
    "batch_idx": 32,
    "loss": 0.18763448297977448,
    "learning_rate": 0.00023474440784663342,
    "timestep_mean": 545.4375,
    "epsilon_norm": 7.892191410064697,
    "prediction_norm": 7.00820779800415,
    "timestamp": "2025-07-14T12:34:21.894968"
  },
  {
    "epoch": 1356,
    "batch_idx": 64,
    "loss": 0.12986832857131958,
    "learning_rate": 0.00023474440784663342,
    "timestep_mean": 565.8125,
    "epsilon_norm": 7.782989978790283,
    "prediction_norm": 7.187723636627197,
    "timestamp": "2025-07-14T12:34:21.900307"
  },
  {
    "epoch": 1356,
    "batch_idx": 96,
    "loss": 0.22272250056266785,
    "learning_rate": 0.00023474440784663342,
    "timestep_mean": 537.4375,
    "epsilon_norm": 6.487773418426514,
    "prediction_norm": 5.846972942352295,
    "timestamp": "2025-07-14T12:34:21.905518"
  },
  {
    "epoch": 1356,
    "batch_idx": 128,
    "loss": 0.22861850261688232,
    "learning_rate": 0.00023474440784663342,
    "timestep_mean": 521.40625,
    "epsilon_norm": 8.264032363891602,
    "prediction_norm": 7.415506839752197,
    "timestamp": "2025-07-14T12:34:21.910755"
  },
  {
    "epoch": 1356,
    "batch_idx": 160,
    "loss": 0.3464387059211731,
    "learning_rate": 0.00023474440784663342,
    "timestep_mean": 495.53125,
    "epsilon_norm": 7.393874645233154,
    "prediction_norm": 5.429675579071045,
    "timestamp": "2025-07-14T12:34:21.915952"
  },
  {
    "epoch": 1356,
    "batch_idx": 192,
    "loss": 0.5742285847663879,
    "learning_rate": 0.00023474440784663342,
    "timestep_mean": 313.875,
    "epsilon_norm": 4.426181316375732,
    "prediction_norm": 2.9240238666534424,
    "timestamp": "2025-07-14T12:34:21.921104"
  },
  {
    "epoch": 1357,
    "batch_idx": 0,
    "loss": 0.2643921375274658,
    "learning_rate": 0.00023407897067216686,
    "timestep_mean": 523.96875,
    "epsilon_norm": 9.144006729125977,
    "prediction_norm": 8.21937084197998,
    "timestamp": "2025-07-14T12:34:21.926263"
  },
  {
    "epoch": 1357,
    "batch_idx": 32,
    "loss": 0.09768244624137878,
    "learning_rate": 0.00023407897067216686,
    "timestep_mean": 568.53125,
    "epsilon_norm": 7.060906887054443,
    "prediction_norm": 7.169325351715088,
    "timestamp": "2025-07-14T12:34:21.931368"
  },
  {
    "epoch": 1357,
    "batch_idx": 64,
    "loss": 0.3466866612434387,
    "learning_rate": 0.00023407897067216686,
    "timestep_mean": 526.71875,
    "epsilon_norm": 7.48155403137207,
    "prediction_norm": 6.35461950302124,
    "timestamp": "2025-07-14T12:34:21.936623"
  },
  {
    "epoch": 1357,
    "batch_idx": 96,
    "loss": 0.3824118375778198,
    "learning_rate": 0.00023407897067216686,
    "timestep_mean": 476.9375,
    "epsilon_norm": 8.400395393371582,
    "prediction_norm": 6.779766082763672,
    "timestamp": "2025-07-14T12:34:21.941956"
  },
  {
    "epoch": 1357,
    "batch_idx": 128,
    "loss": 0.24456974864006042,
    "learning_rate": 0.00023407897067216686,
    "timestep_mean": 527.375,
    "epsilon_norm": 7.92636251449585,
    "prediction_norm": 7.0537848472595215,
    "timestamp": "2025-07-14T12:34:21.947107"
  },
  {
    "epoch": 1357,
    "batch_idx": 160,
    "loss": 0.16796982288360596,
    "learning_rate": 0.00023407897067216686,
    "timestep_mean": 547.34375,
    "epsilon_norm": 8.355560302734375,
    "prediction_norm": 7.497315406799316,
    "timestamp": "2025-07-14T12:34:21.952251"
  },
  {
    "epoch": 1357,
    "batch_idx": 192,
    "loss": 0.2014351189136505,
    "learning_rate": 0.00023407897067216686,
    "timestep_mean": 382.125,
    "epsilon_norm": 3.450040578842163,
    "prediction_norm": 2.9146363735198975,
    "timestamp": "2025-07-14T12:34:21.957207"
  },
  {
    "epoch": 1358,
    "batch_idx": 0,
    "loss": 0.4417375922203064,
    "learning_rate": 0.00023341418963140597,
    "timestep_mean": 487.9375,
    "epsilon_norm": 8.482215881347656,
    "prediction_norm": 6.398225784301758,
    "timestamp": "2025-07-14T12:34:21.962068"
  },
  {
    "epoch": 1358,
    "batch_idx": 32,
    "loss": 0.3787651062011719,
    "learning_rate": 0.00023341418963140597,
    "timestep_mean": 459.5,
    "epsilon_norm": 8.494102478027344,
    "prediction_norm": 6.657556056976318,
    "timestamp": "2025-07-14T12:34:21.966953"
  },
  {
    "epoch": 1358,
    "batch_idx": 64,
    "loss": 0.24353398382663727,
    "learning_rate": 0.00023341418963140597,
    "timestep_mean": 473.1875,
    "epsilon_norm": 7.137552738189697,
    "prediction_norm": 5.959168910980225,
    "timestamp": "2025-07-14T12:34:21.971828"
  },
  {
    "epoch": 1358,
    "batch_idx": 96,
    "loss": 0.28501439094543457,
    "learning_rate": 0.00023341418963140597,
    "timestep_mean": 482.1875,
    "epsilon_norm": 7.448666095733643,
    "prediction_norm": 6.179925441741943,
    "timestamp": "2025-07-14T12:34:21.976731"
  },
  {
    "epoch": 1358,
    "batch_idx": 128,
    "loss": 0.21743977069854736,
    "learning_rate": 0.00023341418963140597,
    "timestep_mean": 521.1875,
    "epsilon_norm": 7.7675395011901855,
    "prediction_norm": 6.631967067718506,
    "timestamp": "2025-07-14T12:34:21.981617"
  },
  {
    "epoch": 1358,
    "batch_idx": 160,
    "loss": 0.2005913406610489,
    "learning_rate": 0.00023341418963140597,
    "timestep_mean": 459.0625,
    "epsilon_norm": 7.618746280670166,
    "prediction_norm": 7.06223201751709,
    "timestamp": "2025-07-14T12:34:21.986434"
  },
  {
    "epoch": 1358,
    "batch_idx": 192,
    "loss": 0.2607845962047577,
    "learning_rate": 0.00023341418963140597,
    "timestep_mean": 572.875,
    "epsilon_norm": 3.942605495452881,
    "prediction_norm": 3.4418466091156006,
    "timestamp": "2025-07-14T12:34:21.991198"
  },
  {
    "epoch": 1359,
    "batch_idx": 0,
    "loss": 0.22559130191802979,
    "learning_rate": 0.0002327500663646317,
    "timestep_mean": 562.6875,
    "epsilon_norm": 9.701576232910156,
    "prediction_norm": 8.494173049926758,
    "timestamp": "2025-07-14T12:34:21.996166"
  },
  {
    "epoch": 1359,
    "batch_idx": 32,
    "loss": 0.26789021492004395,
    "learning_rate": 0.0002327500663646317,
    "timestep_mean": 571.40625,
    "epsilon_norm": 7.8372626304626465,
    "prediction_norm": 6.975531578063965,
    "timestamp": "2025-07-14T12:34:22.001026"
  },
  {
    "epoch": 1359,
    "batch_idx": 64,
    "loss": 0.14908108115196228,
    "learning_rate": 0.0002327500663646317,
    "timestep_mean": 551.59375,
    "epsilon_norm": 7.239714622497559,
    "prediction_norm": 6.183965682983398,
    "timestamp": "2025-07-14T12:34:22.005883"
  },
  {
    "epoch": 1359,
    "batch_idx": 96,
    "loss": 0.23279568552970886,
    "learning_rate": 0.0002327500663646317,
    "timestep_mean": 489.59375,
    "epsilon_norm": 6.386432647705078,
    "prediction_norm": 5.618404865264893,
    "timestamp": "2025-07-14T12:34:22.010728"
  },
  {
    "epoch": 1359,
    "batch_idx": 128,
    "loss": 0.5809515714645386,
    "learning_rate": 0.0002327500663646317,
    "timestep_mean": 479.25,
    "epsilon_norm": 9.472684860229492,
    "prediction_norm": 6.732922077178955,
    "timestamp": "2025-07-14T12:34:22.015601"
  },
  {
    "epoch": 1359,
    "batch_idx": 160,
    "loss": 0.21597371995449066,
    "learning_rate": 0.0002327500663646317,
    "timestep_mean": 487.625,
    "epsilon_norm": 8.054732322692871,
    "prediction_norm": 7.645496368408203,
    "timestamp": "2025-07-14T12:34:22.020443"
  },
  {
    "epoch": 1359,
    "batch_idx": 192,
    "loss": 0.2217423915863037,
    "learning_rate": 0.0002327500663646317,
    "timestep_mean": 524.625,
    "epsilon_norm": 3.446845293045044,
    "prediction_norm": 2.9420711994171143,
    "timestamp": "2025-07-14T12:34:22.025406"
  },
  {
    "epoch": 1360,
    "batch_idx": 0,
    "loss": 0.16236039996147156,
    "learning_rate": 0.00023208660251050213,
    "timestep_mean": 522.71875,
    "epsilon_norm": 7.5930633544921875,
    "prediction_norm": 7.023626327514648,
    "timestamp": "2025-07-14T12:34:22.030256"
  },
  {
    "epoch": 1360,
    "batch_idx": 32,
    "loss": 0.35178494453430176,
    "learning_rate": 0.00023208660251050213,
    "timestep_mean": 490.625,
    "epsilon_norm": 8.737092971801758,
    "prediction_norm": 7.084750175476074,
    "timestamp": "2025-07-14T12:34:22.035083"
  },
  {
    "epoch": 1360,
    "batch_idx": 64,
    "loss": 0.24463534355163574,
    "learning_rate": 0.00023208660251050213,
    "timestep_mean": 452.1875,
    "epsilon_norm": 7.750209808349609,
    "prediction_norm": 6.779157638549805,
    "timestamp": "2025-07-14T12:34:22.039971"
  },
  {
    "epoch": 1360,
    "batch_idx": 96,
    "loss": 0.2163153886795044,
    "learning_rate": 0.00023208660251050213,
    "timestep_mean": 545.53125,
    "epsilon_norm": 7.625965118408203,
    "prediction_norm": 6.693592071533203,
    "timestamp": "2025-07-14T12:34:22.044878"
  },
  {
    "epoch": 1360,
    "batch_idx": 128,
    "loss": 0.14628687500953674,
    "learning_rate": 0.00023208660251050213,
    "timestep_mean": 521.65625,
    "epsilon_norm": 7.617526054382324,
    "prediction_norm": 6.870880126953125,
    "timestamp": "2025-07-14T12:34:22.049730"
  },
  {
    "epoch": 1360,
    "batch_idx": 160,
    "loss": 0.2562539875507355,
    "learning_rate": 0.00023208660251050213,
    "timestep_mean": 565.125,
    "epsilon_norm": 7.387783050537109,
    "prediction_norm": 6.038455963134766,
    "timestamp": "2025-07-14T12:34:22.054617"
  },
  {
    "epoch": 1360,
    "batch_idx": 192,
    "loss": 0.3680906891822815,
    "learning_rate": 0.00023208660251050213,
    "timestep_mean": 661.0,
    "epsilon_norm": 3.696424722671509,
    "prediction_norm": 2.641810655593872,
    "timestamp": "2025-07-14T12:34:22.059407"
  },
  {
    "epoch": 1361,
    "batch_idx": 0,
    "loss": 0.4356139302253723,
    "learning_rate": 0.00023142379970604818,
    "timestep_mean": 466.65625,
    "epsilon_norm": 8.349581718444824,
    "prediction_norm": 6.571294784545898,
    "timestamp": "2025-07-14T12:34:22.064235"
  },
  {
    "epoch": 1361,
    "batch_idx": 32,
    "loss": 0.42034152150154114,
    "learning_rate": 0.00023142379970604818,
    "timestep_mean": 528.96875,
    "epsilon_norm": 7.386790752410889,
    "prediction_norm": 5.932496547698975,
    "timestamp": "2025-07-14T12:34:22.069050"
  },
  {
    "epoch": 1361,
    "batch_idx": 64,
    "loss": 0.2113790065050125,
    "learning_rate": 0.00023142379970604818,
    "timestep_mean": 527.25,
    "epsilon_norm": 6.450361728668213,
    "prediction_norm": 5.249795436859131,
    "timestamp": "2025-07-14T12:34:22.073877"
  },
  {
    "epoch": 1361,
    "batch_idx": 96,
    "loss": 0.21476317942142487,
    "learning_rate": 0.00023142379970604818,
    "timestep_mean": 500.5625,
    "epsilon_norm": 6.9866156578063965,
    "prediction_norm": 6.012514591217041,
    "timestamp": "2025-07-14T12:34:22.078766"
  },
  {
    "epoch": 1361,
    "batch_idx": 128,
    "loss": 0.29145485162734985,
    "learning_rate": 0.00023142379970604818,
    "timestep_mean": 503.75,
    "epsilon_norm": 9.855323791503906,
    "prediction_norm": 8.066184043884277,
    "timestamp": "2025-07-14T12:34:22.083795"
  },
  {
    "epoch": 1361,
    "batch_idx": 160,
    "loss": 0.06625973433256149,
    "learning_rate": 0.00023142379970604818,
    "timestep_mean": 651.5625,
    "epsilon_norm": 7.84437894821167,
    "prediction_norm": 7.684101581573486,
    "timestamp": "2025-07-14T12:34:22.088667"
  },
  {
    "epoch": 1361,
    "batch_idx": 192,
    "loss": 0.07750634104013443,
    "learning_rate": 0.00023142379970604818,
    "timestep_mean": 601.125,
    "epsilon_norm": 3.767254590988159,
    "prediction_norm": 3.421006441116333,
    "timestamp": "2025-07-14T12:34:22.093504"
  },
  {
    "epoch": 1362,
    "batch_idx": 0,
    "loss": 0.22880913317203522,
    "learning_rate": 0.00023076165958667042,
    "timestep_mean": 463.0,
    "epsilon_norm": 6.781782627105713,
    "prediction_norm": 5.586264610290527,
    "timestamp": "2025-07-14T12:34:22.098432"
  },
  {
    "epoch": 1362,
    "batch_idx": 32,
    "loss": 0.1110672727227211,
    "learning_rate": 0.00023076165958667042,
    "timestep_mean": 555.59375,
    "epsilon_norm": 6.975872039794922,
    "prediction_norm": 6.740324020385742,
    "timestamp": "2025-07-14T12:34:22.103291"
  },
  {
    "epoch": 1362,
    "batch_idx": 64,
    "loss": 0.5114051103591919,
    "learning_rate": 0.00023076165958667042,
    "timestep_mean": 491.15625,
    "epsilon_norm": 8.240439414978027,
    "prediction_norm": 6.4048285484313965,
    "timestamp": "2025-07-14T12:34:22.108162"
  },
  {
    "epoch": 1362,
    "batch_idx": 96,
    "loss": 0.32438603043556213,
    "learning_rate": 0.00023076165958667042,
    "timestep_mean": 486.09375,
    "epsilon_norm": 8.292579650878906,
    "prediction_norm": 6.9360857009887695,
    "timestamp": "2025-07-14T12:34:22.113138"
  },
  {
    "epoch": 1362,
    "batch_idx": 128,
    "loss": 0.42562538385391235,
    "learning_rate": 0.00023076165958667042,
    "timestep_mean": 462.15625,
    "epsilon_norm": 8.568373680114746,
    "prediction_norm": 6.325397968292236,
    "timestamp": "2025-07-14T12:34:22.117983"
  },
  {
    "epoch": 1362,
    "batch_idx": 160,
    "loss": 0.14824552834033966,
    "learning_rate": 0.00023076165958667042,
    "timestep_mean": 597.3125,
    "epsilon_norm": 8.77531623840332,
    "prediction_norm": 8.278857231140137,
    "timestamp": "2025-07-14T12:34:22.122839"
  },
  {
    "epoch": 1362,
    "batch_idx": 192,
    "loss": 0.3142845928668976,
    "learning_rate": 0.00023076165958667042,
    "timestep_mean": 583.375,
    "epsilon_norm": 3.151376247406006,
    "prediction_norm": 2.5143494606018066,
    "timestamp": "2025-07-14T12:34:22.127645"
  },
  {
    "epoch": 1363,
    "batch_idx": 0,
    "loss": 0.22801974415779114,
    "learning_rate": 0.0002301001837861331,
    "timestep_mean": 602.6875,
    "epsilon_norm": 8.507804870605469,
    "prediction_norm": 8.09249496459961,
    "timestamp": "2025-07-14T12:34:22.132524"
  },
  {
    "epoch": 1363,
    "batch_idx": 32,
    "loss": 0.1524161845445633,
    "learning_rate": 0.0002301001837861331,
    "timestep_mean": 485.5,
    "epsilon_norm": 6.513197898864746,
    "prediction_norm": 6.143299579620361,
    "timestamp": "2025-07-14T12:34:22.137322"
  },
  {
    "epoch": 1363,
    "batch_idx": 64,
    "loss": 0.38675493001937866,
    "learning_rate": 0.0002301001837861331,
    "timestep_mean": 442.34375,
    "epsilon_norm": 8.123309135437012,
    "prediction_norm": 5.4703874588012695,
    "timestamp": "2025-07-14T12:34:22.142394"
  },
  {
    "epoch": 1363,
    "batch_idx": 96,
    "loss": 0.1867174506187439,
    "learning_rate": 0.0002301001837861331,
    "timestep_mean": 509.6875,
    "epsilon_norm": 7.826582431793213,
    "prediction_norm": 7.414066314697266,
    "timestamp": "2025-07-14T12:34:22.147155"
  },
  {
    "epoch": 1363,
    "batch_idx": 128,
    "loss": 0.22385528683662415,
    "learning_rate": 0.0002301001837861331,
    "timestep_mean": 614.75,
    "epsilon_norm": 8.40029525756836,
    "prediction_norm": 7.479025840759277,
    "timestamp": "2025-07-14T12:34:22.151985"
  },
  {
    "epoch": 1363,
    "batch_idx": 160,
    "loss": 0.1352662742137909,
    "learning_rate": 0.0002301001837861331,
    "timestep_mean": 574.25,
    "epsilon_norm": 7.28619384765625,
    "prediction_norm": 6.801814556121826,
    "timestamp": "2025-07-14T12:34:22.156804"
  },
  {
    "epoch": 1363,
    "batch_idx": 192,
    "loss": 0.11024855077266693,
    "learning_rate": 0.0002301001837861331,
    "timestep_mean": 411.75,
    "epsilon_norm": 2.919154167175293,
    "prediction_norm": 2.728010654449463,
    "timestamp": "2025-07-14T12:34:22.161606"
  },
  {
    "epoch": 1364,
    "batch_idx": 0,
    "loss": 0.4226805865764618,
    "learning_rate": 0.00022943937393656247,
    "timestep_mean": 501.59375,
    "epsilon_norm": 7.414878845214844,
    "prediction_norm": 5.5206732749938965,
    "timestamp": "2025-07-14T12:34:22.166507"
  },
  {
    "epoch": 1364,
    "batch_idx": 32,
    "loss": 0.17983636260032654,
    "learning_rate": 0.00022943937393656247,
    "timestep_mean": 513.34375,
    "epsilon_norm": 9.08870792388916,
    "prediction_norm": 7.697910785675049,
    "timestamp": "2025-07-14T12:34:22.171500"
  },
  {
    "epoch": 1364,
    "batch_idx": 64,
    "loss": 0.21759629249572754,
    "learning_rate": 0.00022943937393656247,
    "timestep_mean": 479.90625,
    "epsilon_norm": 8.431927680969238,
    "prediction_norm": 7.0967607498168945,
    "timestamp": "2025-07-14T12:34:22.176372"
  },
  {
    "epoch": 1364,
    "batch_idx": 96,
    "loss": 0.14627400040626526,
    "learning_rate": 0.00022943937393656247,
    "timestep_mean": 535.21875,
    "epsilon_norm": 7.750422477722168,
    "prediction_norm": 6.471345901489258,
    "timestamp": "2025-07-14T12:34:22.181257"
  },
  {
    "epoch": 1364,
    "batch_idx": 128,
    "loss": 0.12023968249559402,
    "learning_rate": 0.00022943937393656247,
    "timestep_mean": 568.875,
    "epsilon_norm": 6.545622825622559,
    "prediction_norm": 6.031383991241455,
    "timestamp": "2025-07-14T12:34:22.186125"
  },
  {
    "epoch": 1364,
    "batch_idx": 160,
    "loss": 0.2061384618282318,
    "learning_rate": 0.00022943937393656247,
    "timestep_mean": 472.875,
    "epsilon_norm": 8.204922676086426,
    "prediction_norm": 7.296317100524902,
    "timestamp": "2025-07-14T12:34:22.191014"
  },
  {
    "epoch": 1364,
    "batch_idx": 192,
    "loss": 0.5150524377822876,
    "learning_rate": 0.00022943937393656247,
    "timestep_mean": 361.75,
    "epsilon_norm": 4.881948947906494,
    "prediction_norm": 3.294766902923584,
    "timestamp": "2025-07-14T12:34:22.196032"
  },
  {
    "epoch": 1365,
    "batch_idx": 0,
    "loss": 0.20245568454265594,
    "learning_rate": 0.00022877923166844108,
    "timestep_mean": 437.28125,
    "epsilon_norm": 8.450974464416504,
    "prediction_norm": 7.330272674560547,
    "timestamp": "2025-07-14T12:34:22.200951"
  },
  {
    "epoch": 1365,
    "batch_idx": 32,
    "loss": 0.2254001796245575,
    "learning_rate": 0.00022877923166844108,
    "timestep_mean": 410.1875,
    "epsilon_norm": 7.505687236785889,
    "prediction_norm": 6.497118949890137,
    "timestamp": "2025-07-14T12:34:22.205908"
  },
  {
    "epoch": 1365,
    "batch_idx": 64,
    "loss": 0.3670662045478821,
    "learning_rate": 0.00022877923166844108,
    "timestep_mean": 480.34375,
    "epsilon_norm": 8.045424461364746,
    "prediction_norm": 6.549676418304443,
    "timestamp": "2025-07-14T12:34:22.210824"
  },
  {
    "epoch": 1365,
    "batch_idx": 96,
    "loss": 0.23315311968326569,
    "learning_rate": 0.00022877923166844108,
    "timestep_mean": 525.25,
    "epsilon_norm": 8.592923164367676,
    "prediction_norm": 7.6522321701049805,
    "timestamp": "2025-07-14T12:34:22.215733"
  },
  {
    "epoch": 1365,
    "batch_idx": 128,
    "loss": 0.2079721838235855,
    "learning_rate": 0.00022877923166844108,
    "timestep_mean": 428.15625,
    "epsilon_norm": 7.448627948760986,
    "prediction_norm": 6.743892192840576,
    "timestamp": "2025-07-14T12:34:22.220632"
  },
  {
    "epoch": 1365,
    "batch_idx": 160,
    "loss": 0.39904865622520447,
    "learning_rate": 0.00022877923166844108,
    "timestep_mean": 484.6875,
    "epsilon_norm": 7.6148362159729,
    "prediction_norm": 6.081595420837402,
    "timestamp": "2025-07-14T12:34:22.225602"
  },
  {
    "epoch": 1365,
    "batch_idx": 192,
    "loss": 0.20132628083229065,
    "learning_rate": 0.00022877923166844108,
    "timestep_mean": 529.875,
    "epsilon_norm": 3.4616713523864746,
    "prediction_norm": 3.446852684020996,
    "timestamp": "2025-07-14T12:34:22.230407"
  },
  {
    "epoch": 1366,
    "batch_idx": 0,
    "loss": 0.18281930685043335,
    "learning_rate": 0.00022811975861060404,
    "timestep_mean": 521.6875,
    "epsilon_norm": 8.337158203125,
    "prediction_norm": 7.618961811065674,
    "timestamp": "2025-07-14T12:34:22.235458"
  },
  {
    "epoch": 1366,
    "batch_idx": 32,
    "loss": 0.2807534337043762,
    "learning_rate": 0.00022811975861060404,
    "timestep_mean": 508.625,
    "epsilon_norm": 9.157357215881348,
    "prediction_norm": 7.660867214202881,
    "timestamp": "2025-07-14T12:34:22.240415"
  },
  {
    "epoch": 1366,
    "batch_idx": 64,
    "loss": 0.4682310223579407,
    "learning_rate": 0.00022811975861060404,
    "timestep_mean": 492.21875,
    "epsilon_norm": 8.6377592086792,
    "prediction_norm": 6.399163246154785,
    "timestamp": "2025-07-14T12:34:22.245372"
  },
  {
    "epoch": 1366,
    "batch_idx": 96,
    "loss": 0.3656057119369507,
    "learning_rate": 0.00022811975861060404,
    "timestep_mean": 475.53125,
    "epsilon_norm": 7.0979180335998535,
    "prediction_norm": 5.293431758880615,
    "timestamp": "2025-07-14T12:34:22.250410"
  },
  {
    "epoch": 1366,
    "batch_idx": 128,
    "loss": 0.27605393528938293,
    "learning_rate": 0.00022811975861060404,
    "timestep_mean": 519.15625,
    "epsilon_norm": 8.138925552368164,
    "prediction_norm": 6.698696613311768,
    "timestamp": "2025-07-14T12:34:22.255370"
  },
  {
    "epoch": 1366,
    "batch_idx": 160,
    "loss": 0.2221297323703766,
    "learning_rate": 0.00022811975861060404,
    "timestep_mean": 601.21875,
    "epsilon_norm": 9.050554275512695,
    "prediction_norm": 7.906280517578125,
    "timestamp": "2025-07-14T12:34:22.260265"
  },
  {
    "epoch": 1366,
    "batch_idx": 192,
    "loss": 0.39381060004234314,
    "learning_rate": 0.00022811975861060404,
    "timestep_mean": 409.25,
    "epsilon_norm": 3.925166606903076,
    "prediction_norm": 2.817466974258423,
    "timestamp": "2025-07-14T12:34:22.265146"
  },
  {
    "epoch": 1367,
    "batch_idx": 0,
    "loss": 0.14751556515693665,
    "learning_rate": 0.00022746095639023623,
    "timestep_mean": 507.3125,
    "epsilon_norm": 6.1031880378723145,
    "prediction_norm": 5.951179027557373,
    "timestamp": "2025-07-14T12:34:22.269999"
  },
  {
    "epoch": 1367,
    "batch_idx": 32,
    "loss": 0.32805368304252625,
    "learning_rate": 0.00022746095639023623,
    "timestep_mean": 451.15625,
    "epsilon_norm": 8.904973983764648,
    "prediction_norm": 7.274399280548096,
    "timestamp": "2025-07-14T12:34:22.274886"
  },
  {
    "epoch": 1367,
    "batch_idx": 64,
    "loss": 0.3816506266593933,
    "learning_rate": 0.00022746095639023623,
    "timestep_mean": 481.625,
    "epsilon_norm": 8.102520942687988,
    "prediction_norm": 6.212068557739258,
    "timestamp": "2025-07-14T12:34:22.279843"
  },
  {
    "epoch": 1367,
    "batch_idx": 96,
    "loss": 0.0657087042927742,
    "learning_rate": 0.00022746095639023623,
    "timestep_mean": 609.3125,
    "epsilon_norm": 7.636314868927002,
    "prediction_norm": 7.397524833679199,
    "timestamp": "2025-07-14T12:34:22.284717"
  },
  {
    "epoch": 1367,
    "batch_idx": 128,
    "loss": 0.2106224000453949,
    "learning_rate": 0.00022746095639023623,
    "timestep_mean": 591.0,
    "epsilon_norm": 7.932310581207275,
    "prediction_norm": 7.196112632751465,
    "timestamp": "2025-07-14T12:34:22.289648"
  },
  {
    "epoch": 1367,
    "batch_idx": 160,
    "loss": 0.23575565218925476,
    "learning_rate": 0.00022746095639023623,
    "timestep_mean": 557.90625,
    "epsilon_norm": 8.575555801391602,
    "prediction_norm": 7.468506336212158,
    "timestamp": "2025-07-14T12:34:22.294581"
  },
  {
    "epoch": 1367,
    "batch_idx": 192,
    "loss": 0.15671643614768982,
    "learning_rate": 0.00022746095639023623,
    "timestep_mean": 537.75,
    "epsilon_norm": 3.270415782928467,
    "prediction_norm": 3.096024990081787,
    "timestamp": "2025-07-14T12:34:22.299430"
  },
  {
    "epoch": 1368,
    "batch_idx": 0,
    "loss": 0.36661210656166077,
    "learning_rate": 0.00022680282663286582,
    "timestep_mean": 511.875,
    "epsilon_norm": 6.984766006469727,
    "prediction_norm": 6.4068074226379395,
    "timestamp": "2025-07-14T12:34:22.304388"
  },
  {
    "epoch": 1368,
    "batch_idx": 32,
    "loss": 0.20038770139217377,
    "learning_rate": 0.00022680282663286582,
    "timestep_mean": 518.28125,
    "epsilon_norm": 7.9754838943481445,
    "prediction_norm": 7.633808612823486,
    "timestamp": "2025-07-14T12:34:22.309311"
  },
  {
    "epoch": 1368,
    "batch_idx": 64,
    "loss": 0.23086878657341003,
    "learning_rate": 0.00022680282663286582,
    "timestep_mean": 424.71875,
    "epsilon_norm": 7.781862258911133,
    "prediction_norm": 6.812332630157471,
    "timestamp": "2025-07-14T12:34:22.314242"
  },
  {
    "epoch": 1368,
    "batch_idx": 96,
    "loss": 0.23067380487918854,
    "learning_rate": 0.00022680282663286582,
    "timestep_mean": 506.53125,
    "epsilon_norm": 9.026396751403809,
    "prediction_norm": 8.644082069396973,
    "timestamp": "2025-07-14T12:34:22.319145"
  },
  {
    "epoch": 1368,
    "batch_idx": 128,
    "loss": 0.24619512259960175,
    "learning_rate": 0.00022680282663286582,
    "timestep_mean": 571.96875,
    "epsilon_norm": 8.083136558532715,
    "prediction_norm": 7.411966323852539,
    "timestamp": "2025-07-14T12:34:22.324011"
  },
  {
    "epoch": 1368,
    "batch_idx": 160,
    "loss": 0.14905691146850586,
    "learning_rate": 0.00022680282663286582,
    "timestep_mean": 511.65625,
    "epsilon_norm": 7.314618110656738,
    "prediction_norm": 6.7036614418029785,
    "timestamp": "2025-07-14T12:34:22.328895"
  },
  {
    "epoch": 1368,
    "batch_idx": 192,
    "loss": 0.11296108365058899,
    "learning_rate": 0.00022680282663286582,
    "timestep_mean": 500.125,
    "epsilon_norm": 3.8729212284088135,
    "prediction_norm": 3.2997875213623047,
    "timestamp": "2025-07-14T12:34:22.333800"
  },
  {
    "epoch": 1369,
    "batch_idx": 0,
    "loss": 0.2975676357746124,
    "learning_rate": 0.00022614537096236296,
    "timestep_mean": 477.625,
    "epsilon_norm": 8.799421310424805,
    "prediction_norm": 7.475062847137451,
    "timestamp": "2025-07-14T12:34:22.338638"
  },
  {
    "epoch": 1369,
    "batch_idx": 32,
    "loss": 0.276384562253952,
    "learning_rate": 0.00022614537096236296,
    "timestep_mean": 483.15625,
    "epsilon_norm": 7.12183952331543,
    "prediction_norm": 5.451501846313477,
    "timestamp": "2025-07-14T12:34:22.343514"
  },
  {
    "epoch": 1369,
    "batch_idx": 64,
    "loss": 0.3445068597793579,
    "learning_rate": 0.00022614537096236296,
    "timestep_mean": 493.125,
    "epsilon_norm": 7.405396938323975,
    "prediction_norm": 5.600953578948975,
    "timestamp": "2025-07-14T12:34:22.348395"
  },
  {
    "epoch": 1369,
    "batch_idx": 96,
    "loss": 0.24858419597148895,
    "learning_rate": 0.00022614537096236296,
    "timestep_mean": 478.90625,
    "epsilon_norm": 8.666913032531738,
    "prediction_norm": 7.0709052085876465,
    "timestamp": "2025-07-14T12:34:22.353283"
  },
  {
    "epoch": 1369,
    "batch_idx": 128,
    "loss": 0.3644983768463135,
    "learning_rate": 0.00022614537096236296,
    "timestep_mean": 396.0625,
    "epsilon_norm": 7.75966739654541,
    "prediction_norm": 6.11177921295166,
    "timestamp": "2025-07-14T12:34:22.358178"
  },
  {
    "epoch": 1369,
    "batch_idx": 160,
    "loss": 0.39346885681152344,
    "learning_rate": 0.00022614537096236296,
    "timestep_mean": 554.09375,
    "epsilon_norm": 7.899567127227783,
    "prediction_norm": 6.203464508056641,
    "timestamp": "2025-07-14T12:34:22.363199"
  },
  {
    "epoch": 1369,
    "batch_idx": 192,
    "loss": 0.23198390007019043,
    "learning_rate": 0.00022614537096236296,
    "timestep_mean": 337.625,
    "epsilon_norm": 3.450206756591797,
    "prediction_norm": 2.209601402282715,
    "timestamp": "2025-07-14T12:34:22.368077"
  },
  {
    "epoch": 1370,
    "batch_idx": 0,
    "loss": 0.2439875602722168,
    "learning_rate": 0.0002254885910009345,
    "timestep_mean": 476.1875,
    "epsilon_norm": 7.5631842613220215,
    "prediction_norm": 6.240729808807373,
    "timestamp": "2025-07-14T12:34:22.373042"
  },
  {
    "epoch": 1370,
    "batch_idx": 32,
    "loss": 0.15792204439640045,
    "learning_rate": 0.0002254885910009345,
    "timestep_mean": 537.3125,
    "epsilon_norm": 7.387871742248535,
    "prediction_norm": 6.41024923324585,
    "timestamp": "2025-07-14T12:34:22.377874"
  },
  {
    "epoch": 1370,
    "batch_idx": 64,
    "loss": 0.12675166130065918,
    "learning_rate": 0.0002254885910009345,
    "timestep_mean": 567.15625,
    "epsilon_norm": 7.940542221069336,
    "prediction_norm": 7.407157897949219,
    "timestamp": "2025-07-14T12:34:22.382734"
  },
  {
    "epoch": 1370,
    "batch_idx": 96,
    "loss": 0.34943723678588867,
    "learning_rate": 0.0002254885910009345,
    "timestep_mean": 544.96875,
    "epsilon_norm": 8.967591285705566,
    "prediction_norm": 7.73040771484375,
    "timestamp": "2025-07-14T12:34:22.387698"
  },
  {
    "epoch": 1370,
    "batch_idx": 128,
    "loss": 0.332230806350708,
    "learning_rate": 0.0002254885910009345,
    "timestep_mean": 516.375,
    "epsilon_norm": 8.573511123657227,
    "prediction_norm": 6.9743218421936035,
    "timestamp": "2025-07-14T12:34:22.392544"
  },
  {
    "epoch": 1370,
    "batch_idx": 160,
    "loss": 0.31244876980781555,
    "learning_rate": 0.0002254885910009345,
    "timestep_mean": 495.625,
    "epsilon_norm": 8.476802825927734,
    "prediction_norm": 6.8449320793151855,
    "timestamp": "2025-07-14T12:34:22.397386"
  },
  {
    "epoch": 1370,
    "batch_idx": 192,
    "loss": 0.22665400803089142,
    "learning_rate": 0.0002254885910009345,
    "timestep_mean": 537.375,
    "epsilon_norm": 4.320423126220703,
    "prediction_norm": 4.235927581787109,
    "timestamp": "2025-07-14T12:34:22.402224"
  },
  {
    "epoch": 1371,
    "batch_idx": 0,
    "loss": 0.25586652755737305,
    "learning_rate": 0.00022483248836911908,
    "timestep_mean": 446.90625,
    "epsilon_norm": 7.587096691131592,
    "prediction_norm": 6.271796226501465,
    "timestamp": "2025-07-14T12:34:22.407093"
  },
  {
    "epoch": 1371,
    "batch_idx": 32,
    "loss": 0.45798957347869873,
    "learning_rate": 0.00022483248836911908,
    "timestep_mean": 452.25,
    "epsilon_norm": 8.813321113586426,
    "prediction_norm": 7.020182132720947,
    "timestamp": "2025-07-14T12:34:22.412004"
  },
  {
    "epoch": 1371,
    "batch_idx": 64,
    "loss": 0.4183744788169861,
    "learning_rate": 0.00022483248836911908,
    "timestep_mean": 426.25,
    "epsilon_norm": 7.556672096252441,
    "prediction_norm": 6.029470920562744,
    "timestamp": "2025-07-14T12:34:22.417198"
  },
  {
    "epoch": 1371,
    "batch_idx": 96,
    "loss": 0.12860698997974396,
    "learning_rate": 0.00022483248836911908,
    "timestep_mean": 583.75,
    "epsilon_norm": 7.3000383377075195,
    "prediction_norm": 6.71185302734375,
    "timestamp": "2025-07-14T12:34:22.422158"
  },
  {
    "epoch": 1371,
    "batch_idx": 128,
    "loss": 0.31974899768829346,
    "learning_rate": 0.00022483248836911908,
    "timestep_mean": 557.25,
    "epsilon_norm": 8.203989028930664,
    "prediction_norm": 6.775200366973877,
    "timestamp": "2025-07-14T12:34:22.427057"
  },
  {
    "epoch": 1371,
    "batch_idx": 160,
    "loss": 0.2732250392436981,
    "learning_rate": 0.00022483248836911908,
    "timestep_mean": 509.625,
    "epsilon_norm": 7.345285415649414,
    "prediction_norm": 6.309700965881348,
    "timestamp": "2025-07-14T12:34:22.431985"
  },
  {
    "epoch": 1371,
    "batch_idx": 192,
    "loss": 0.3658634424209595,
    "learning_rate": 0.00022483248836911908,
    "timestep_mean": 406.875,
    "epsilon_norm": 3.47986102104187,
    "prediction_norm": 2.65684175491333,
    "timestamp": "2025-07-14T12:34:22.436988"
  },
  {
    "epoch": 1372,
    "batch_idx": 0,
    "loss": 0.3636953830718994,
    "learning_rate": 0.0002241770646857852,
    "timestep_mean": 500.0,
    "epsilon_norm": 8.632917404174805,
    "prediction_norm": 6.5189361572265625,
    "timestamp": "2025-07-14T12:34:22.441990"
  },
  {
    "epoch": 1372,
    "batch_idx": 32,
    "loss": 0.6168262958526611,
    "learning_rate": 0.0002241770646857852,
    "timestep_mean": 400.34375,
    "epsilon_norm": 8.526104927062988,
    "prediction_norm": 5.875787734985352,
    "timestamp": "2025-07-14T12:34:22.446982"
  },
  {
    "epoch": 1372,
    "batch_idx": 64,
    "loss": 0.26948773860931396,
    "learning_rate": 0.0002241770646857852,
    "timestep_mean": 429.4375,
    "epsilon_norm": 8.313523292541504,
    "prediction_norm": 6.655914783477783,
    "timestamp": "2025-07-14T12:34:22.451966"
  },
  {
    "epoch": 1372,
    "batch_idx": 96,
    "loss": 0.6370784640312195,
    "learning_rate": 0.0002241770646857852,
    "timestep_mean": 401.65625,
    "epsilon_norm": 9.143844604492188,
    "prediction_norm": 6.38381290435791,
    "timestamp": "2025-07-14T12:34:22.456991"
  },
  {
    "epoch": 1372,
    "batch_idx": 128,
    "loss": 0.20355209708213806,
    "learning_rate": 0.0002241770646857852,
    "timestep_mean": 551.46875,
    "epsilon_norm": 6.006299018859863,
    "prediction_norm": 5.349639892578125,
    "timestamp": "2025-07-14T12:34:22.461866"
  },
  {
    "epoch": 1372,
    "batch_idx": 160,
    "loss": 0.25692665576934814,
    "learning_rate": 0.0002241770646857852,
    "timestep_mean": 466.375,
    "epsilon_norm": 7.509155750274658,
    "prediction_norm": 6.1115498542785645,
    "timestamp": "2025-07-14T12:34:22.466718"
  },
  {
    "epoch": 1372,
    "batch_idx": 192,
    "loss": 0.11745084822177887,
    "learning_rate": 0.0002241770646857852,
    "timestep_mean": 570.875,
    "epsilon_norm": 2.5555026531219482,
    "prediction_norm": 2.210965394973755,
    "timestamp": "2025-07-14T12:34:22.471667"
  },
  {
    "epoch": 1373,
    "batch_idx": 0,
    "loss": 0.20922234654426575,
    "learning_rate": 0.00022352232156812543,
    "timestep_mean": 506.34375,
    "epsilon_norm": 7.864912986755371,
    "prediction_norm": 6.8403639793396,
    "timestamp": "2025-07-14T12:34:22.476555"
  },
  {
    "epoch": 1373,
    "batch_idx": 32,
    "loss": 0.1478257030248642,
    "learning_rate": 0.00022352232156812543,
    "timestep_mean": 513.40625,
    "epsilon_norm": 8.467971801757812,
    "prediction_norm": 7.436028480529785,
    "timestamp": "2025-07-14T12:34:22.481406"
  },
  {
    "epoch": 1373,
    "batch_idx": 64,
    "loss": 0.16820061206817627,
    "learning_rate": 0.00022352232156812543,
    "timestep_mean": 568.40625,
    "epsilon_norm": 7.5686726570129395,
    "prediction_norm": 6.556433200836182,
    "timestamp": "2025-07-14T12:34:22.486310"
  },
  {
    "epoch": 1373,
    "batch_idx": 96,
    "loss": 0.20390135049819946,
    "learning_rate": 0.00022352232156812543,
    "timestep_mean": 472.9375,
    "epsilon_norm": 8.258302688598633,
    "prediction_norm": 7.286149978637695,
    "timestamp": "2025-07-14T12:34:22.491157"
  },
  {
    "epoch": 1373,
    "batch_idx": 128,
    "loss": 0.3565590977668762,
    "learning_rate": 0.00022352232156812543,
    "timestep_mean": 433.125,
    "epsilon_norm": 7.6975297927856445,
    "prediction_norm": 5.471229076385498,
    "timestamp": "2025-07-14T12:34:22.496098"
  },
  {
    "epoch": 1373,
    "batch_idx": 160,
    "loss": 0.14215238392353058,
    "learning_rate": 0.00022352232156812543,
    "timestep_mean": 450.90625,
    "epsilon_norm": 6.775656700134277,
    "prediction_norm": 5.839321136474609,
    "timestamp": "2025-07-14T12:34:22.500910"
  },
  {
    "epoch": 1373,
    "batch_idx": 192,
    "loss": 0.8674290776252747,
    "learning_rate": 0.00022352232156812543,
    "timestep_mean": 327.75,
    "epsilon_norm": 4.791385173797607,
    "prediction_norm": 3.5295748710632324,
    "timestamp": "2025-07-14T12:34:22.505801"
  },
  {
    "epoch": 1374,
    "batch_idx": 0,
    "loss": 0.22543787956237793,
    "learning_rate": 0.00022286826063165337,
    "timestep_mean": 537.21875,
    "epsilon_norm": 8.20634937286377,
    "prediction_norm": 7.325326442718506,
    "timestamp": "2025-07-14T12:34:22.510717"
  },
  {
    "epoch": 1374,
    "batch_idx": 32,
    "loss": 0.32706165313720703,
    "learning_rate": 0.00022286826063165337,
    "timestep_mean": 474.1875,
    "epsilon_norm": 9.208304405212402,
    "prediction_norm": 7.254612445831299,
    "timestamp": "2025-07-14T12:34:22.515601"
  },
  {
    "epoch": 1374,
    "batch_idx": 64,
    "loss": 0.24626508355140686,
    "learning_rate": 0.00022286826063165337,
    "timestep_mean": 494.3125,
    "epsilon_norm": 8.242502212524414,
    "prediction_norm": 6.643182754516602,
    "timestamp": "2025-07-14T12:34:22.520439"
  },
  {
    "epoch": 1374,
    "batch_idx": 96,
    "loss": 0.21010923385620117,
    "learning_rate": 0.00022286826063165337,
    "timestep_mean": 485.84375,
    "epsilon_norm": 6.747541427612305,
    "prediction_norm": 5.515254020690918,
    "timestamp": "2025-07-14T12:34:22.525412"
  },
  {
    "epoch": 1374,
    "batch_idx": 128,
    "loss": 0.3280770778656006,
    "learning_rate": 0.00022286826063165337,
    "timestep_mean": 529.6875,
    "epsilon_norm": 8.190434455871582,
    "prediction_norm": 6.7099127769470215,
    "timestamp": "2025-07-14T12:34:22.530243"
  },
  {
    "epoch": 1374,
    "batch_idx": 160,
    "loss": 0.33797726035118103,
    "learning_rate": 0.00022286826063165337,
    "timestep_mean": 551.78125,
    "epsilon_norm": 7.568994522094727,
    "prediction_norm": 5.8629350662231445,
    "timestamp": "2025-07-14T12:34:22.535044"
  },
  {
    "epoch": 1374,
    "batch_idx": 192,
    "loss": 0.2815346419811249,
    "learning_rate": 0.00022286826063165337,
    "timestep_mean": 529.875,
    "epsilon_norm": 4.0265913009643555,
    "prediction_norm": 3.590965747833252,
    "timestamp": "2025-07-14T12:34:22.539886"
  },
  {
    "epoch": 1375,
    "batch_idx": 0,
    "loss": 0.2178625911474228,
    "learning_rate": 0.00022221488349019933,
    "timestep_mean": 556.40625,
    "epsilon_norm": 7.9429826736450195,
    "prediction_norm": 7.0262556076049805,
    "timestamp": "2025-07-14T12:34:22.544741"
  },
  {
    "epoch": 1375,
    "batch_idx": 32,
    "loss": 0.38280731439590454,
    "learning_rate": 0.00022221488349019933,
    "timestep_mean": 472.75,
    "epsilon_norm": 8.596881866455078,
    "prediction_norm": 7.158090114593506,
    "timestamp": "2025-07-14T12:34:22.549545"
  },
  {
    "epoch": 1375,
    "batch_idx": 64,
    "loss": 0.31981194019317627,
    "learning_rate": 0.00022221488349019933,
    "timestep_mean": 477.875,
    "epsilon_norm": 7.666619777679443,
    "prediction_norm": 6.021090507507324,
    "timestamp": "2025-07-14T12:34:22.554463"
  },
  {
    "epoch": 1375,
    "batch_idx": 96,
    "loss": 0.29015052318573,
    "learning_rate": 0.00022221488349019933,
    "timestep_mean": 455.46875,
    "epsilon_norm": 8.33924388885498,
    "prediction_norm": 7.237002372741699,
    "timestamp": "2025-07-14T12:34:22.559276"
  },
  {
    "epoch": 1375,
    "batch_idx": 128,
    "loss": 0.3189963102340698,
    "learning_rate": 0.00022221488349019933,
    "timestep_mean": 430.75,
    "epsilon_norm": 7.604457855224609,
    "prediction_norm": 5.800506591796875,
    "timestamp": "2025-07-14T12:34:22.564179"
  },
  {
    "epoch": 1375,
    "batch_idx": 160,
    "loss": 0.1777631789445877,
    "learning_rate": 0.00022221488349019933,
    "timestep_mean": 445.59375,
    "epsilon_norm": 7.4366960525512695,
    "prediction_norm": 6.671616554260254,
    "timestamp": "2025-07-14T12:34:22.569076"
  },
  {
    "epoch": 1375,
    "batch_idx": 192,
    "loss": 0.4383208155632019,
    "learning_rate": 0.00022221488349019933,
    "timestep_mean": 445.375,
    "epsilon_norm": 4.652022361755371,
    "prediction_norm": 3.0688343048095703,
    "timestamp": "2025-07-14T12:34:22.573928"
  },
  {
    "epoch": 1376,
    "batch_idx": 0,
    "loss": 0.1773867905139923,
    "learning_rate": 0.00022156219175590628,
    "timestep_mean": 571.5625,
    "epsilon_norm": 8.23483943939209,
    "prediction_norm": 7.357808589935303,
    "timestamp": "2025-07-14T12:34:22.578768"
  },
  {
    "epoch": 1376,
    "batch_idx": 32,
    "loss": 0.13915449380874634,
    "learning_rate": 0.00022156219175590628,
    "timestep_mean": 507.3125,
    "epsilon_norm": 7.319477558135986,
    "prediction_norm": 6.1757588386535645,
    "timestamp": "2025-07-14T12:34:22.583690"
  },
  {
    "epoch": 1376,
    "batch_idx": 64,
    "loss": 0.2410130500793457,
    "learning_rate": 0.00022156219175590628,
    "timestep_mean": 510.75,
    "epsilon_norm": 7.75009298324585,
    "prediction_norm": 7.009931564331055,
    "timestamp": "2025-07-14T12:34:22.588544"
  },
  {
    "epoch": 1376,
    "batch_idx": 96,
    "loss": 0.2368486523628235,
    "learning_rate": 0.00022156219175590628,
    "timestep_mean": 488.6875,
    "epsilon_norm": 7.148520469665527,
    "prediction_norm": 6.2024149894714355,
    "timestamp": "2025-07-14T12:34:22.593378"
  },
  {
    "epoch": 1376,
    "batch_idx": 128,
    "loss": 0.3038454055786133,
    "learning_rate": 0.00022156219175590628,
    "timestep_mean": 477.15625,
    "epsilon_norm": 8.943049430847168,
    "prediction_norm": 8.066023826599121,
    "timestamp": "2025-07-14T12:34:22.598275"
  },
  {
    "epoch": 1376,
    "batch_idx": 160,
    "loss": 0.5126990079879761,
    "learning_rate": 0.00022156219175590628,
    "timestep_mean": 479.625,
    "epsilon_norm": 7.977274417877197,
    "prediction_norm": 6.564539909362793,
    "timestamp": "2025-07-14T12:34:22.603188"
  },
  {
    "epoch": 1376,
    "batch_idx": 192,
    "loss": 0.09129224717617035,
    "learning_rate": 0.00022156219175590628,
    "timestep_mean": 642.125,
    "epsilon_norm": 4.055602550506592,
    "prediction_norm": 3.8748764991760254,
    "timestamp": "2025-07-14T12:34:22.608020"
  },
  {
    "epoch": 1377,
    "batch_idx": 0,
    "loss": 0.19595158100128174,
    "learning_rate": 0.00022091018703922658,
    "timestep_mean": 543.5625,
    "epsilon_norm": 7.088500499725342,
    "prediction_norm": 6.5681257247924805,
    "timestamp": "2025-07-14T12:34:22.612930"
  },
  {
    "epoch": 1377,
    "batch_idx": 32,
    "loss": 0.41872209310531616,
    "learning_rate": 0.00022091018703922658,
    "timestep_mean": 452.25,
    "epsilon_norm": 9.182933807373047,
    "prediction_norm": 7.287199020385742,
    "timestamp": "2025-07-14T12:34:22.617775"
  },
  {
    "epoch": 1377,
    "batch_idx": 64,
    "loss": 0.3165425658226013,
    "learning_rate": 0.00022091018703922658,
    "timestep_mean": 585.59375,
    "epsilon_norm": 7.685686111450195,
    "prediction_norm": 6.439182758331299,
    "timestamp": "2025-07-14T12:34:22.622598"
  },
  {
    "epoch": 1377,
    "batch_idx": 96,
    "loss": 0.3684738278388977,
    "learning_rate": 0.00022091018703922658,
    "timestep_mean": 513.59375,
    "epsilon_norm": 7.657540321350098,
    "prediction_norm": 6.119470596313477,
    "timestamp": "2025-07-14T12:34:22.627481"
  },
  {
    "epoch": 1377,
    "batch_idx": 128,
    "loss": 0.3709743618965149,
    "learning_rate": 0.00022091018703922658,
    "timestep_mean": 503.4375,
    "epsilon_norm": 8.513693809509277,
    "prediction_norm": 7.349175453186035,
    "timestamp": "2025-07-14T12:34:22.632284"
  },
  {
    "epoch": 1377,
    "batch_idx": 160,
    "loss": 0.24907946586608887,
    "learning_rate": 0.00022091018703922658,
    "timestep_mean": 540.46875,
    "epsilon_norm": 7.669400215148926,
    "prediction_norm": 6.90675687789917,
    "timestamp": "2025-07-14T12:34:22.637118"
  },
  {
    "epoch": 1377,
    "batch_idx": 192,
    "loss": 0.333769828081131,
    "learning_rate": 0.00022091018703922658,
    "timestep_mean": 399.625,
    "epsilon_norm": 4.3515944480896,
    "prediction_norm": 3.3625078201293945,
    "timestamp": "2025-07-14T12:34:22.642068"
  },
  {
    "epoch": 1378,
    "batch_idx": 0,
    "loss": 0.24324539303779602,
    "learning_rate": 0.0002202588709489169,
    "timestep_mean": 509.90625,
    "epsilon_norm": 9.441564559936523,
    "prediction_norm": 7.638606071472168,
    "timestamp": "2025-07-14T12:34:22.646918"
  },
  {
    "epoch": 1378,
    "batch_idx": 32,
    "loss": 0.34917721152305603,
    "learning_rate": 0.0002202588709489169,
    "timestep_mean": 457.25,
    "epsilon_norm": 7.794052600860596,
    "prediction_norm": 6.192030429840088,
    "timestamp": "2025-07-14T12:34:22.651767"
  },
  {
    "epoch": 1378,
    "batch_idx": 64,
    "loss": 0.2953387498855591,
    "learning_rate": 0.0002202588709489169,
    "timestep_mean": 551.5,
    "epsilon_norm": 9.261813163757324,
    "prediction_norm": 7.693196773529053,
    "timestamp": "2025-07-14T12:34:22.656641"
  },
  {
    "epoch": 1378,
    "batch_idx": 96,
    "loss": 0.5186969041824341,
    "learning_rate": 0.0002202588709489169,
    "timestep_mean": 425.65625,
    "epsilon_norm": 8.700176239013672,
    "prediction_norm": 7.078152179718018,
    "timestamp": "2025-07-14T12:34:22.661696"
  },
  {
    "epoch": 1378,
    "batch_idx": 128,
    "loss": 0.21115374565124512,
    "learning_rate": 0.0002202588709489169,
    "timestep_mean": 460.8125,
    "epsilon_norm": 6.903530597686768,
    "prediction_norm": 6.307975769042969,
    "timestamp": "2025-07-14T12:34:22.666852"
  },
  {
    "epoch": 1378,
    "batch_idx": 160,
    "loss": 0.3980758786201477,
    "learning_rate": 0.0002202588709489169,
    "timestep_mean": 529.96875,
    "epsilon_norm": 8.387200355529785,
    "prediction_norm": 6.8516526222229,
    "timestamp": "2025-07-14T12:34:22.672001"
  },
  {
    "epoch": 1378,
    "batch_idx": 192,
    "loss": 0.14156827330589294,
    "learning_rate": 0.0002202588709489169,
    "timestep_mean": 550.625,
    "epsilon_norm": 3.5952467918395996,
    "prediction_norm": 3.5743467807769775,
    "timestamp": "2025-07-14T12:34:22.677203"
  },
  {
    "epoch": 1379,
    "batch_idx": 0,
    "loss": 0.3984122574329376,
    "learning_rate": 0.00021960824509203484,
    "timestep_mean": 384.375,
    "epsilon_norm": 9.942245483398438,
    "prediction_norm": 8.306474685668945,
    "timestamp": "2025-07-14T12:34:22.682354"
  },
  {
    "epoch": 1379,
    "batch_idx": 32,
    "loss": 0.37952175736427307,
    "learning_rate": 0.00021960824509203484,
    "timestep_mean": 442.0,
    "epsilon_norm": 7.9775390625,
    "prediction_norm": 6.540373802185059,
    "timestamp": "2025-07-14T12:34:22.687630"
  },
  {
    "epoch": 1379,
    "batch_idx": 64,
    "loss": 0.4565923511981964,
    "learning_rate": 0.00021960824509203484,
    "timestep_mean": 460.3125,
    "epsilon_norm": 7.051102161407471,
    "prediction_norm": 5.155526638031006,
    "timestamp": "2025-07-14T12:34:22.692767"
  },
  {
    "epoch": 1379,
    "batch_idx": 96,
    "loss": 0.3893783986568451,
    "learning_rate": 0.00021960824509203484,
    "timestep_mean": 505.78125,
    "epsilon_norm": 8.155656814575195,
    "prediction_norm": 6.609664440155029,
    "timestamp": "2025-07-14T12:34:22.697913"
  },
  {
    "epoch": 1379,
    "batch_idx": 128,
    "loss": 0.45932984352111816,
    "learning_rate": 0.00021960824509203484,
    "timestep_mean": 463.25,
    "epsilon_norm": 8.540578842163086,
    "prediction_norm": 7.425184726715088,
    "timestamp": "2025-07-14T12:34:22.703049"
  },
  {
    "epoch": 1379,
    "batch_idx": 160,
    "loss": 0.3466235101222992,
    "learning_rate": 0.00021960824509203484,
    "timestep_mean": 501.0625,
    "epsilon_norm": 8.662644386291504,
    "prediction_norm": 7.183195114135742,
    "timestamp": "2025-07-14T12:34:22.708157"
  },
  {
    "epoch": 1379,
    "batch_idx": 192,
    "loss": 0.09633480757474899,
    "learning_rate": 0.00021960824509203484,
    "timestep_mean": 624.0,
    "epsilon_norm": 3.692484140396118,
    "prediction_norm": 3.580430030822754,
    "timestamp": "2025-07-14T12:34:22.713264"
  },
  {
    "epoch": 1380,
    "batch_idx": 0,
    "loss": 0.32287168502807617,
    "learning_rate": 0.00021895831107393498,
    "timestep_mean": 453.875,
    "epsilon_norm": 8.348767280578613,
    "prediction_norm": 6.529053688049316,
    "timestamp": "2025-07-14T12:34:22.718409"
  },
  {
    "epoch": 1380,
    "batch_idx": 32,
    "loss": 0.2674817442893982,
    "learning_rate": 0.00021895831107393498,
    "timestep_mean": 498.78125,
    "epsilon_norm": 7.901676654815674,
    "prediction_norm": 6.917994499206543,
    "timestamp": "2025-07-14T12:34:22.723569"
  },
  {
    "epoch": 1380,
    "batch_idx": 64,
    "loss": 0.41039663553237915,
    "learning_rate": 0.00021895831107393498,
    "timestep_mean": 413.59375,
    "epsilon_norm": 8.946735382080078,
    "prediction_norm": 6.997963905334473,
    "timestamp": "2025-07-14T12:34:22.728716"
  },
  {
    "epoch": 1380,
    "batch_idx": 96,
    "loss": 0.12491998821496964,
    "learning_rate": 0.00021895831107393498,
    "timestep_mean": 555.28125,
    "epsilon_norm": 8.330672264099121,
    "prediction_norm": 7.498826026916504,
    "timestamp": "2025-07-14T12:34:22.733944"
  },
  {
    "epoch": 1380,
    "batch_idx": 128,
    "loss": 0.2130281627178192,
    "learning_rate": 0.00021895831107393498,
    "timestep_mean": 517.34375,
    "epsilon_norm": 7.2256975173950195,
    "prediction_norm": 6.3728437423706055,
    "timestamp": "2025-07-14T12:34:22.739057"
  },
  {
    "epoch": 1380,
    "batch_idx": 160,
    "loss": 0.25200217962265015,
    "learning_rate": 0.00021895831107393498,
    "timestep_mean": 518.4375,
    "epsilon_norm": 7.642361640930176,
    "prediction_norm": 6.1397857666015625,
    "timestamp": "2025-07-14T12:34:22.744262"
  },
  {
    "epoch": 1380,
    "batch_idx": 192,
    "loss": 0.5907005071640015,
    "learning_rate": 0.00021895831107393498,
    "timestep_mean": 446.0,
    "epsilon_norm": 4.069798946380615,
    "prediction_norm": 2.569103717803955,
    "timestamp": "2025-07-14T12:34:22.749407"
  },
  {
    "epoch": 1381,
    "batch_idx": 0,
    "loss": 0.2322320193052292,
    "learning_rate": 0.0002183090704982652,
    "timestep_mean": 486.75,
    "epsilon_norm": 8.327247619628906,
    "prediction_norm": 7.709958553314209,
    "timestamp": "2025-07-14T12:34:22.754672"
  },
  {
    "epoch": 1381,
    "batch_idx": 32,
    "loss": 0.19736947119235992,
    "learning_rate": 0.0002183090704982652,
    "timestep_mean": 544.09375,
    "epsilon_norm": 6.718824863433838,
    "prediction_norm": 5.916110992431641,
    "timestamp": "2025-07-14T12:34:22.759921"
  },
  {
    "epoch": 1381,
    "batch_idx": 64,
    "loss": 0.19020500779151917,
    "learning_rate": 0.0002183090704982652,
    "timestep_mean": 530.90625,
    "epsilon_norm": 8.948451042175293,
    "prediction_norm": 7.68053674697876,
    "timestamp": "2025-07-14T12:34:22.765075"
  },
  {
    "epoch": 1381,
    "batch_idx": 96,
    "loss": 0.18556375801563263,
    "learning_rate": 0.0002183090704982652,
    "timestep_mean": 492.5625,
    "epsilon_norm": 7.340921878814697,
    "prediction_norm": 6.729722023010254,
    "timestamp": "2025-07-14T12:34:22.770269"
  },
  {
    "epoch": 1381,
    "batch_idx": 128,
    "loss": 0.24129101634025574,
    "learning_rate": 0.0002183090704982652,
    "timestep_mean": 506.25,
    "epsilon_norm": 8.392494201660156,
    "prediction_norm": 7.291504383087158,
    "timestamp": "2025-07-14T12:34:22.775589"
  },
  {
    "epoch": 1381,
    "batch_idx": 160,
    "loss": 0.20833703875541687,
    "learning_rate": 0.0002183090704982652,
    "timestep_mean": 536.4375,
    "epsilon_norm": 7.649201393127441,
    "prediction_norm": 6.843194007873535,
    "timestamp": "2025-07-14T12:34:22.780778"
  },
  {
    "epoch": 1381,
    "batch_idx": 192,
    "loss": 0.1623765528202057,
    "learning_rate": 0.0002183090704982652,
    "timestep_mean": 491.5,
    "epsilon_norm": 4.639302730560303,
    "prediction_norm": 4.294477462768555,
    "timestamp": "2025-07-14T12:34:22.785897"
  },
  {
    "epoch": 1382,
    "batch_idx": 0,
    "loss": 0.1570536196231842,
    "learning_rate": 0.00021766052496696186,
    "timestep_mean": 493.9375,
    "epsilon_norm": 7.407312870025635,
    "prediction_norm": 7.14000940322876,
    "timestamp": "2025-07-14T12:34:22.791148"
  },
  {
    "epoch": 1382,
    "batch_idx": 32,
    "loss": 0.18052034080028534,
    "learning_rate": 0.00021766052496696186,
    "timestep_mean": 496.78125,
    "epsilon_norm": 6.416271686553955,
    "prediction_norm": 5.852114677429199,
    "timestamp": "2025-07-14T12:34:22.796454"
  },
  {
    "epoch": 1382,
    "batch_idx": 64,
    "loss": 0.24569477140903473,
    "learning_rate": 0.00021766052496696186,
    "timestep_mean": 508.1875,
    "epsilon_norm": 6.184099197387695,
    "prediction_norm": 5.230798244476318,
    "timestamp": "2025-07-14T12:34:22.801584"
  },
  {
    "epoch": 1382,
    "batch_idx": 96,
    "loss": 0.15699952840805054,
    "learning_rate": 0.00021766052496696186,
    "timestep_mean": 597.5,
    "epsilon_norm": 8.307886123657227,
    "prediction_norm": 7.770998477935791,
    "timestamp": "2025-07-14T12:34:22.806728"
  },
  {
    "epoch": 1382,
    "batch_idx": 128,
    "loss": 0.3475651741027832,
    "learning_rate": 0.00021766052496696186,
    "timestep_mean": 504.65625,
    "epsilon_norm": 9.827255249023438,
    "prediction_norm": 8.316787719726562,
    "timestamp": "2025-07-14T12:34:22.811925"
  },
  {
    "epoch": 1382,
    "batch_idx": 160,
    "loss": 0.2711351215839386,
    "learning_rate": 0.00021766052496696186,
    "timestep_mean": 494.875,
    "epsilon_norm": 8.886016845703125,
    "prediction_norm": 7.972984790802002,
    "timestamp": "2025-07-14T12:34:22.816989"
  },
  {
    "epoch": 1382,
    "batch_idx": 192,
    "loss": 0.2090853750705719,
    "learning_rate": 0.00021766052496696186,
    "timestep_mean": 578.125,
    "epsilon_norm": 3.1701197624206543,
    "prediction_norm": 2.813162088394165,
    "timestamp": "2025-07-14T12:34:22.822052"
  },
  {
    "epoch": 1383,
    "batch_idx": 0,
    "loss": 0.24355953931808472,
    "learning_rate": 0.0002170126760802464,
    "timestep_mean": 495.53125,
    "epsilon_norm": 8.323139190673828,
    "prediction_norm": 6.546029567718506,
    "timestamp": "2025-07-14T12:34:22.827213"
  },
  {
    "epoch": 1383,
    "batch_idx": 32,
    "loss": 0.2432527393102646,
    "learning_rate": 0.0002170126760802464,
    "timestep_mean": 507.96875,
    "epsilon_norm": 9.343588829040527,
    "prediction_norm": 7.77946662902832,
    "timestamp": "2025-07-14T12:34:22.832346"
  },
  {
    "epoch": 1383,
    "batch_idx": 64,
    "loss": 0.1838088482618332,
    "learning_rate": 0.0002170126760802464,
    "timestep_mean": 601.5,
    "epsilon_norm": 8.04449462890625,
    "prediction_norm": 7.021806716918945,
    "timestamp": "2025-07-14T12:34:22.837591"
  },
  {
    "epoch": 1383,
    "batch_idx": 96,
    "loss": 0.4798602759838104,
    "learning_rate": 0.0002170126760802464,
    "timestep_mean": 460.59375,
    "epsilon_norm": 9.048150062561035,
    "prediction_norm": 7.0459885597229,
    "timestamp": "2025-07-14T12:34:22.842771"
  },
  {
    "epoch": 1383,
    "batch_idx": 128,
    "loss": 0.2768360376358032,
    "learning_rate": 0.0002170126760802464,
    "timestep_mean": 476.0625,
    "epsilon_norm": 8.131444931030273,
    "prediction_norm": 6.731552600860596,
    "timestamp": "2025-07-14T12:34:22.847934"
  },
  {
    "epoch": 1383,
    "batch_idx": 160,
    "loss": 0.17194543778896332,
    "learning_rate": 0.0002170126760802464,
    "timestep_mean": 595.15625,
    "epsilon_norm": 7.424167156219482,
    "prediction_norm": 6.349350929260254,
    "timestamp": "2025-07-14T12:34:22.853082"
  },
  {
    "epoch": 1383,
    "batch_idx": 192,
    "loss": 0.09156478941440582,
    "learning_rate": 0.0002170126760802464,
    "timestep_mean": 466.375,
    "epsilon_norm": 3.9287338256835938,
    "prediction_norm": 3.892606258392334,
    "timestamp": "2025-07-14T12:34:22.858221"
  },
  {
    "epoch": 1384,
    "batch_idx": 0,
    "loss": 0.2314312756061554,
    "learning_rate": 0.00021636552543662216,
    "timestep_mean": 401.59375,
    "epsilon_norm": 7.675328254699707,
    "prediction_norm": 6.771769046783447,
    "timestamp": "2025-07-14T12:34:22.863533"
  },
  {
    "epoch": 1384,
    "batch_idx": 32,
    "loss": 0.17233547568321228,
    "learning_rate": 0.00021636552543662216,
    "timestep_mean": 474.40625,
    "epsilon_norm": 8.52630615234375,
    "prediction_norm": 7.54966402053833,
    "timestamp": "2025-07-14T12:34:22.868524"
  },
  {
    "epoch": 1384,
    "batch_idx": 64,
    "loss": 0.2877886891365051,
    "learning_rate": 0.00021636552543662216,
    "timestep_mean": 533.8125,
    "epsilon_norm": 8.392923355102539,
    "prediction_norm": 6.8674116134643555,
    "timestamp": "2025-07-14T12:34:22.873628"
  },
  {
    "epoch": 1384,
    "batch_idx": 96,
    "loss": 0.40688031911849976,
    "learning_rate": 0.00021636552543662216,
    "timestep_mean": 479.0625,
    "epsilon_norm": 8.36510181427002,
    "prediction_norm": 7.242895603179932,
    "timestamp": "2025-07-14T12:34:22.878672"
  },
  {
    "epoch": 1384,
    "batch_idx": 128,
    "loss": 0.4177091419696808,
    "learning_rate": 0.00021636552543662216,
    "timestep_mean": 500.53125,
    "epsilon_norm": 8.480953216552734,
    "prediction_norm": 6.2881999015808105,
    "timestamp": "2025-07-14T12:34:22.883977"
  },
  {
    "epoch": 1384,
    "batch_idx": 160,
    "loss": 0.2151048332452774,
    "learning_rate": 0.00021636552543662216,
    "timestep_mean": 419.3125,
    "epsilon_norm": 7.210049152374268,
    "prediction_norm": 6.647083759307861,
    "timestamp": "2025-07-14T12:34:22.889041"
  },
  {
    "epoch": 1384,
    "batch_idx": 192,
    "loss": 0.15607896447181702,
    "learning_rate": 0.00021636552543662216,
    "timestep_mean": 612.5,
    "epsilon_norm": 3.758758068084717,
    "prediction_norm": 3.1695823669433594,
    "timestamp": "2025-07-14T12:34:22.894278"
  },
  {
    "epoch": 1385,
    "batch_idx": 0,
    "loss": 0.11921902000904083,
    "learning_rate": 0.0002157190746328683,
    "timestep_mean": 459.125,
    "epsilon_norm": 6.689938068389893,
    "prediction_norm": 6.190888404846191,
    "timestamp": "2025-07-14T12:34:22.899440"
  },
  {
    "epoch": 1385,
    "batch_idx": 32,
    "loss": 0.34661537408828735,
    "learning_rate": 0.0002157190746328683,
    "timestep_mean": 495.125,
    "epsilon_norm": 7.756432056427002,
    "prediction_norm": 6.133106231689453,
    "timestamp": "2025-07-14T12:34:22.904700"
  },
  {
    "epoch": 1385,
    "batch_idx": 64,
    "loss": 0.16363567113876343,
    "learning_rate": 0.0002157190746328683,
    "timestep_mean": 459.21875,
    "epsilon_norm": 8.348394393920898,
    "prediction_norm": 7.577393531799316,
    "timestamp": "2025-07-14T12:34:22.909849"
  },
  {
    "epoch": 1385,
    "batch_idx": 96,
    "loss": 0.1373518705368042,
    "learning_rate": 0.0002157190746328683,
    "timestep_mean": 532.59375,
    "epsilon_norm": 8.24182415008545,
    "prediction_norm": 7.6751909255981445,
    "timestamp": "2025-07-14T12:34:22.915047"
  },
  {
    "epoch": 1385,
    "batch_idx": 128,
    "loss": 0.25055038928985596,
    "learning_rate": 0.0002157190746328683,
    "timestep_mean": 485.96875,
    "epsilon_norm": 8.120063781738281,
    "prediction_norm": 7.562340259552002,
    "timestamp": "2025-07-14T12:34:22.920181"
  },
  {
    "epoch": 1385,
    "batch_idx": 160,
    "loss": 0.3106592893600464,
    "learning_rate": 0.0002157190746328683,
    "timestep_mean": 480.4375,
    "epsilon_norm": 8.720345497131348,
    "prediction_norm": 7.293038845062256,
    "timestamp": "2025-07-14T12:34:22.925421"
  },
  {
    "epoch": 1385,
    "batch_idx": 192,
    "loss": 0.3277331292629242,
    "learning_rate": 0.0002157190746328683,
    "timestep_mean": 533.125,
    "epsilon_norm": 4.880782604217529,
    "prediction_norm": 4.1016926765441895,
    "timestamp": "2025-07-14T12:34:22.930495"
  },
  {
    "epoch": 1386,
    "batch_idx": 0,
    "loss": 0.6445626020431519,
    "learning_rate": 0.00021507332526403864,
    "timestep_mean": 457.34375,
    "epsilon_norm": 7.860403060913086,
    "prediction_norm": 5.55648136138916,
    "timestamp": "2025-07-14T12:34:22.935633"
  },
  {
    "epoch": 1386,
    "batch_idx": 32,
    "loss": 0.3847111761569977,
    "learning_rate": 0.00021507332526403864,
    "timestep_mean": 481.75,
    "epsilon_norm": 8.009003639221191,
    "prediction_norm": 6.457920551300049,
    "timestamp": "2025-07-14T12:34:22.940799"
  },
  {
    "epoch": 1386,
    "batch_idx": 64,
    "loss": 0.4512304663658142,
    "learning_rate": 0.00021507332526403864,
    "timestep_mean": 490.0,
    "epsilon_norm": 8.075273513793945,
    "prediction_norm": 6.601572513580322,
    "timestamp": "2025-07-14T12:34:22.946033"
  },
  {
    "epoch": 1386,
    "batch_idx": 96,
    "loss": 0.21464115381240845,
    "learning_rate": 0.00021507332526403864,
    "timestep_mean": 474.28125,
    "epsilon_norm": 8.730072021484375,
    "prediction_norm": 7.268514156341553,
    "timestamp": "2025-07-14T12:34:22.951185"
  },
  {
    "epoch": 1386,
    "batch_idx": 128,
    "loss": 0.3890911340713501,
    "learning_rate": 0.00021507332526403864,
    "timestep_mean": 403.90625,
    "epsilon_norm": 7.647920608520508,
    "prediction_norm": 5.281683444976807,
    "timestamp": "2025-07-14T12:34:22.956208"
  },
  {
    "epoch": 1386,
    "batch_idx": 160,
    "loss": 0.30105409026145935,
    "learning_rate": 0.00021507332526403864,
    "timestep_mean": 506.3125,
    "epsilon_norm": 7.844358921051025,
    "prediction_norm": 6.20904541015625,
    "timestamp": "2025-07-14T12:34:22.961200"
  },
  {
    "epoch": 1386,
    "batch_idx": 192,
    "loss": 0.14420776069164276,
    "learning_rate": 0.00021507332526403864,
    "timestep_mean": 431.75,
    "epsilon_norm": 4.089840888977051,
    "prediction_norm": 3.571668863296509,
    "timestamp": "2025-07-14T12:34:22.965989"
  },
  {
    "epoch": 1387,
    "batch_idx": 0,
    "loss": 0.19136233627796173,
    "learning_rate": 0.00021442827892345474,
    "timestep_mean": 541.0,
    "epsilon_norm": 8.59586238861084,
    "prediction_norm": 7.779796600341797,
    "timestamp": "2025-07-14T12:34:22.970923"
  },
  {
    "epoch": 1387,
    "batch_idx": 32,
    "loss": 0.6544888019561768,
    "learning_rate": 0.00021442827892345474,
    "timestep_mean": 478.0,
    "epsilon_norm": 8.678282737731934,
    "prediction_norm": 6.507531642913818,
    "timestamp": "2025-07-14T12:34:22.975808"
  },
  {
    "epoch": 1387,
    "batch_idx": 64,
    "loss": 0.21039998531341553,
    "learning_rate": 0.00021442827892345474,
    "timestep_mean": 476.59375,
    "epsilon_norm": 7.098112106323242,
    "prediction_norm": 6.172278881072998,
    "timestamp": "2025-07-14T12:34:22.980602"
  },
  {
    "epoch": 1387,
    "batch_idx": 96,
    "loss": 0.277954638004303,
    "learning_rate": 0.00021442827892345474,
    "timestep_mean": 504.875,
    "epsilon_norm": 8.29123306274414,
    "prediction_norm": 6.976380348205566,
    "timestamp": "2025-07-14T12:34:22.985634"
  },
  {
    "epoch": 1387,
    "batch_idx": 128,
    "loss": 0.5225429534912109,
    "learning_rate": 0.00021442827892345474,
    "timestep_mean": 432.1875,
    "epsilon_norm": 7.499222755432129,
    "prediction_norm": 5.473207473754883,
    "timestamp": "2025-07-14T12:34:22.990800"
  },
  {
    "epoch": 1387,
    "batch_idx": 160,
    "loss": 0.24436785280704498,
    "learning_rate": 0.00021442827892345474,
    "timestep_mean": 550.90625,
    "epsilon_norm": 8.159144401550293,
    "prediction_norm": 6.962527751922607,
    "timestamp": "2025-07-14T12:34:22.996065"
  },
  {
    "epoch": 1387,
    "batch_idx": 192,
    "loss": 0.43231621384620667,
    "learning_rate": 0.00021442827892345474,
    "timestep_mean": 526.875,
    "epsilon_norm": 4.203949928283691,
    "prediction_norm": 3.466937303543091,
    "timestamp": "2025-07-14T12:34:23.001124"
  },
  {
    "epoch": 1388,
    "batch_idx": 0,
    "loss": 0.516366183757782,
    "learning_rate": 0.0002137839372027048,
    "timestep_mean": 439.0,
    "epsilon_norm": 8.662797927856445,
    "prediction_norm": 6.584692478179932,
    "timestamp": "2025-07-14T12:34:23.006220"
  },
  {
    "epoch": 1388,
    "batch_idx": 32,
    "loss": 0.42917996644973755,
    "learning_rate": 0.0002137839372027048,
    "timestep_mean": 524.28125,
    "epsilon_norm": 7.549713134765625,
    "prediction_norm": 5.314077854156494,
    "timestamp": "2025-07-14T12:34:23.011309"
  },
  {
    "epoch": 1388,
    "batch_idx": 64,
    "loss": 0.2347518801689148,
    "learning_rate": 0.0002137839372027048,
    "timestep_mean": 550.8125,
    "epsilon_norm": 7.540851593017578,
    "prediction_norm": 6.441255569458008,
    "timestamp": "2025-07-14T12:34:23.016306"
  },
  {
    "epoch": 1388,
    "batch_idx": 96,
    "loss": 0.4304679036140442,
    "learning_rate": 0.0002137839372027048,
    "timestep_mean": 533.03125,
    "epsilon_norm": 8.678970336914062,
    "prediction_norm": 6.373507499694824,
    "timestamp": "2025-07-14T12:34:23.021391"
  },
  {
    "epoch": 1388,
    "batch_idx": 128,
    "loss": 0.20468290150165558,
    "learning_rate": 0.0002137839372027048,
    "timestep_mean": 497.3125,
    "epsilon_norm": 7.962904453277588,
    "prediction_norm": 6.654420852661133,
    "timestamp": "2025-07-14T12:34:23.026588"
  },
  {
    "epoch": 1388,
    "batch_idx": 160,
    "loss": 0.3723784387111664,
    "learning_rate": 0.0002137839372027048,
    "timestep_mean": 505.59375,
    "epsilon_norm": 8.21926212310791,
    "prediction_norm": 6.834105968475342,
    "timestamp": "2025-07-14T12:34:23.031681"
  },
  {
    "epoch": 1388,
    "batch_idx": 192,
    "loss": 0.2982790470123291,
    "learning_rate": 0.0002137839372027048,
    "timestep_mean": 527.0,
    "epsilon_norm": 3.818563222885132,
    "prediction_norm": 2.9874861240386963,
    "timestamp": "2025-07-14T12:34:23.036908"
  },
  {
    "epoch": 1389,
    "batch_idx": 0,
    "loss": 0.28532299399375916,
    "learning_rate": 0.00021314030169163823,
    "timestep_mean": 509.53125,
    "epsilon_norm": 8.693860054016113,
    "prediction_norm": 7.2367753982543945,
    "timestamp": "2025-07-14T12:34:23.042052"
  },
  {
    "epoch": 1389,
    "batch_idx": 32,
    "loss": 0.1647697240114212,
    "learning_rate": 0.00021314030169163823,
    "timestep_mean": 481.4375,
    "epsilon_norm": 5.952835559844971,
    "prediction_norm": 5.602060794830322,
    "timestamp": "2025-07-14T12:34:23.047123"
  },
  {
    "epoch": 1389,
    "batch_idx": 64,
    "loss": 0.17625126242637634,
    "learning_rate": 0.00021314030169163823,
    "timestep_mean": 505.84375,
    "epsilon_norm": 8.737009048461914,
    "prediction_norm": 7.85382604598999,
    "timestamp": "2025-07-14T12:34:23.052354"
  },
  {
    "epoch": 1389,
    "batch_idx": 96,
    "loss": 0.3078107535839081,
    "learning_rate": 0.00021314030169163823,
    "timestep_mean": 429.46875,
    "epsilon_norm": 8.249110221862793,
    "prediction_norm": 6.99766731262207,
    "timestamp": "2025-07-14T12:34:23.057521"
  },
  {
    "epoch": 1389,
    "batch_idx": 128,
    "loss": 0.21966564655303955,
    "learning_rate": 0.00021314030169163823,
    "timestep_mean": 560.40625,
    "epsilon_norm": 7.5599493980407715,
    "prediction_norm": 7.069016456604004,
    "timestamp": "2025-07-14T12:34:23.062684"
  },
  {
    "epoch": 1389,
    "batch_idx": 160,
    "loss": 0.4153248965740204,
    "learning_rate": 0.00021314030169163823,
    "timestep_mean": 411.375,
    "epsilon_norm": 9.14807415008545,
    "prediction_norm": 6.369542598724365,
    "timestamp": "2025-07-14T12:34:23.067675"
  },
  {
    "epoch": 1389,
    "batch_idx": 192,
    "loss": 0.2676176428794861,
    "learning_rate": 0.00021314030169163823,
    "timestep_mean": 532.5,
    "epsilon_norm": 4.7766618728637695,
    "prediction_norm": 3.7345199584960938,
    "timestamp": "2025-07-14T12:34:23.072632"
  },
  {
    "epoch": 1390,
    "batch_idx": 0,
    "loss": 0.37249863147735596,
    "learning_rate": 0.000212497373978361,
    "timestep_mean": 465.5,
    "epsilon_norm": 7.261246681213379,
    "prediction_norm": 5.719228267669678,
    "timestamp": "2025-07-14T12:34:23.077713"
  },
  {
    "epoch": 1390,
    "batch_idx": 32,
    "loss": 0.44514936208724976,
    "learning_rate": 0.000212497373978361,
    "timestep_mean": 513.59375,
    "epsilon_norm": 8.515127182006836,
    "prediction_norm": 6.517690658569336,
    "timestamp": "2025-07-14T12:34:23.082840"
  },
  {
    "epoch": 1390,
    "batch_idx": 64,
    "loss": 0.3434889316558838,
    "learning_rate": 0.000212497373978361,
    "timestep_mean": 438.75,
    "epsilon_norm": 7.2508063316345215,
    "prediction_norm": 5.777422904968262,
    "timestamp": "2025-07-14T12:34:23.088073"
  },
  {
    "epoch": 1390,
    "batch_idx": 96,
    "loss": 0.3511512577533722,
    "learning_rate": 0.000212497373978361,
    "timestep_mean": 459.5,
    "epsilon_norm": 8.106976509094238,
    "prediction_norm": 7.315843105316162,
    "timestamp": "2025-07-14T12:34:23.093201"
  },
  {
    "epoch": 1390,
    "batch_idx": 128,
    "loss": 0.3352762460708618,
    "learning_rate": 0.000212497373978361,
    "timestep_mean": 476.9375,
    "epsilon_norm": 7.498504161834717,
    "prediction_norm": 6.070170879364014,
    "timestamp": "2025-07-14T12:34:23.098332"
  },
  {
    "epoch": 1390,
    "batch_idx": 160,
    "loss": 0.36417970061302185,
    "learning_rate": 0.000212497373978361,
    "timestep_mean": 485.84375,
    "epsilon_norm": 8.956311225891113,
    "prediction_norm": 7.4274468421936035,
    "timestamp": "2025-07-14T12:34:23.103475"
  },
  {
    "epoch": 1390,
    "batch_idx": 192,
    "loss": 0.17260953783988953,
    "learning_rate": 0.000212497373978361,
    "timestep_mean": 598.875,
    "epsilon_norm": 6.330978870391846,
    "prediction_norm": 6.140945911407471,
    "timestamp": "2025-07-14T12:34:23.108560"
  },
  {
    "epoch": 1391,
    "batch_idx": 0,
    "loss": 0.32170942425727844,
    "learning_rate": 0.00021185515564923384,
    "timestep_mean": 445.0625,
    "epsilon_norm": 8.685233116149902,
    "prediction_norm": 7.770844459533691,
    "timestamp": "2025-07-14T12:34:23.113717"
  },
  {
    "epoch": 1391,
    "batch_idx": 32,
    "loss": 0.29696810245513916,
    "learning_rate": 0.00021185515564923384,
    "timestep_mean": 503.5625,
    "epsilon_norm": 8.396336555480957,
    "prediction_norm": 7.4486775398254395,
    "timestamp": "2025-07-14T12:34:23.118797"
  },
  {
    "epoch": 1391,
    "batch_idx": 64,
    "loss": 0.17270943522453308,
    "learning_rate": 0.00021185515564923384,
    "timestep_mean": 473.84375,
    "epsilon_norm": 7.445080757141113,
    "prediction_norm": 7.603781223297119,
    "timestamp": "2025-07-14T12:34:23.123660"
  },
  {
    "epoch": 1391,
    "batch_idx": 96,
    "loss": 0.251955509185791,
    "learning_rate": 0.00021185515564923384,
    "timestep_mean": 500.84375,
    "epsilon_norm": 7.299159049987793,
    "prediction_norm": 6.269255638122559,
    "timestamp": "2025-07-14T12:34:23.128552"
  },
  {
    "epoch": 1391,
    "batch_idx": 128,
    "loss": 0.21081241965293884,
    "learning_rate": 0.00021185515564923384,
    "timestep_mean": 534.1875,
    "epsilon_norm": 7.4923224449157715,
    "prediction_norm": 7.208329677581787,
    "timestamp": "2025-07-14T12:34:23.133389"
  },
  {
    "epoch": 1391,
    "batch_idx": 160,
    "loss": 0.2901439070701599,
    "learning_rate": 0.00021185515564923384,
    "timestep_mean": 443.625,
    "epsilon_norm": 8.032235145568848,
    "prediction_norm": 6.94043493270874,
    "timestamp": "2025-07-14T12:34:23.138365"
  },
  {
    "epoch": 1391,
    "batch_idx": 192,
    "loss": 0.16736140847206116,
    "learning_rate": 0.00021185515564923384,
    "timestep_mean": 506.875,
    "epsilon_norm": 3.6288301944732666,
    "prediction_norm": 3.370255708694458,
    "timestamp": "2025-07-14T12:34:23.143179"
  },
  {
    "epoch": 1392,
    "batch_idx": 0,
    "loss": 0.24466189742088318,
    "learning_rate": 0.00021121364828886657,
    "timestep_mean": 520.59375,
    "epsilon_norm": 7.289355754852295,
    "prediction_norm": 6.405293941497803,
    "timestamp": "2025-07-14T12:34:23.148069"
  },
  {
    "epoch": 1392,
    "batch_idx": 32,
    "loss": 0.24436138570308685,
    "learning_rate": 0.00021121364828886657,
    "timestep_mean": 592.3125,
    "epsilon_norm": 7.106176376342773,
    "prediction_norm": 6.4500627517700195,
    "timestamp": "2025-07-14T12:34:23.152906"
  },
  {
    "epoch": 1392,
    "batch_idx": 64,
    "loss": 0.2589870095252991,
    "learning_rate": 0.00021121364828886657,
    "timestep_mean": 415.0625,
    "epsilon_norm": 8.082712173461914,
    "prediction_norm": 7.672851085662842,
    "timestamp": "2025-07-14T12:34:23.157979"
  },
  {
    "epoch": 1392,
    "batch_idx": 96,
    "loss": 0.20985881984233856,
    "learning_rate": 0.00021121364828886657,
    "timestep_mean": 513.96875,
    "epsilon_norm": 8.728982925415039,
    "prediction_norm": 7.299034595489502,
    "timestamp": "2025-07-14T12:34:23.163326"
  },
  {
    "epoch": 1392,
    "batch_idx": 128,
    "loss": 0.1482146978378296,
    "learning_rate": 0.00021121364828886657,
    "timestep_mean": 503.8125,
    "epsilon_norm": 8.43734073638916,
    "prediction_norm": 8.043289184570312,
    "timestamp": "2025-07-14T12:34:23.168570"
  },
  {
    "epoch": 1392,
    "batch_idx": 160,
    "loss": 0.2813912630081177,
    "learning_rate": 0.00021121364828886657,
    "timestep_mean": 484.75,
    "epsilon_norm": 7.506921768188477,
    "prediction_norm": 6.642776012420654,
    "timestamp": "2025-07-14T12:34:23.173728"
  },
  {
    "epoch": 1392,
    "batch_idx": 192,
    "loss": 0.2308812290430069,
    "learning_rate": 0.00021121364828886657,
    "timestep_mean": 569.875,
    "epsilon_norm": 3.8842437267303467,
    "prediction_norm": 2.9412291049957275,
    "timestamp": "2025-07-14T12:34:23.178920"
  },
  {
    "epoch": 1393,
    "batch_idx": 0,
    "loss": 0.1580462008714676,
    "learning_rate": 0.0002105728534801145,
    "timestep_mean": 545.84375,
    "epsilon_norm": 8.03504753112793,
    "prediction_norm": 6.9252495765686035,
    "timestamp": "2025-07-14T12:34:23.184124"
  },
  {
    "epoch": 1393,
    "batch_idx": 32,
    "loss": 0.16804881393909454,
    "learning_rate": 0.0002105728534801145,
    "timestep_mean": 479.65625,
    "epsilon_norm": 8.493040084838867,
    "prediction_norm": 7.233750820159912,
    "timestamp": "2025-07-14T12:34:23.189292"
  },
  {
    "epoch": 1393,
    "batch_idx": 64,
    "loss": 0.378936767578125,
    "learning_rate": 0.0002105728534801145,
    "timestep_mean": 513.65625,
    "epsilon_norm": 8.303582191467285,
    "prediction_norm": 7.3624348640441895,
    "timestamp": "2025-07-14T12:34:23.194545"
  },
  {
    "epoch": 1393,
    "batch_idx": 96,
    "loss": 0.3430873453617096,
    "learning_rate": 0.0002105728534801145,
    "timestep_mean": 415.625,
    "epsilon_norm": 8.006858825683594,
    "prediction_norm": 6.453363418579102,
    "timestamp": "2025-07-14T12:34:23.199707"
  },
  {
    "epoch": 1393,
    "batch_idx": 128,
    "loss": 0.10333073884248734,
    "learning_rate": 0.0002105728534801145,
    "timestep_mean": 580.84375,
    "epsilon_norm": 8.19144058227539,
    "prediction_norm": 7.708563804626465,
    "timestamp": "2025-07-14T12:34:23.204937"
  },
  {
    "epoch": 1393,
    "batch_idx": 160,
    "loss": 0.19970500469207764,
    "learning_rate": 0.0002105728534801145,
    "timestep_mean": 554.15625,
    "epsilon_norm": 7.452357769012451,
    "prediction_norm": 6.190404891967773,
    "timestamp": "2025-07-14T12:34:23.210191"
  },
  {
    "epoch": 1393,
    "batch_idx": 192,
    "loss": 0.6455387473106384,
    "learning_rate": 0.0002105728534801145,
    "timestep_mean": 445.75,
    "epsilon_norm": 4.8598103523254395,
    "prediction_norm": 3.1943743228912354,
    "timestamp": "2025-07-14T12:34:23.215314"
  },
  {
    "epoch": 1394,
    "batch_idx": 0,
    "loss": 0.3562124967575073,
    "learning_rate": 0.00020993277280407556,
    "timestep_mean": 517.0,
    "epsilon_norm": 8.712509155273438,
    "prediction_norm": 7.024849891662598,
    "timestamp": "2025-07-14T12:34:23.220557"
  },
  {
    "epoch": 1394,
    "batch_idx": 32,
    "loss": 0.2401483952999115,
    "learning_rate": 0.00020993277280407556,
    "timestep_mean": 528.59375,
    "epsilon_norm": 7.399321556091309,
    "prediction_norm": 6.238668441772461,
    "timestamp": "2025-07-14T12:34:23.225724"
  },
  {
    "epoch": 1394,
    "batch_idx": 64,
    "loss": 0.23975303769111633,
    "learning_rate": 0.00020993277280407556,
    "timestep_mean": 524.625,
    "epsilon_norm": 7.404843330383301,
    "prediction_norm": 6.32307243347168,
    "timestamp": "2025-07-14T12:34:23.230684"
  },
  {
    "epoch": 1394,
    "batch_idx": 96,
    "loss": 0.33804693818092346,
    "learning_rate": 0.00020993277280407556,
    "timestep_mean": 504.5,
    "epsilon_norm": 7.747217178344727,
    "prediction_norm": 6.706002235412598,
    "timestamp": "2025-07-14T12:34:23.235500"
  },
  {
    "epoch": 1394,
    "batch_idx": 128,
    "loss": 0.14031222462654114,
    "learning_rate": 0.00020993277280407556,
    "timestep_mean": 476.375,
    "epsilon_norm": 6.931643009185791,
    "prediction_norm": 6.114249229431152,
    "timestamp": "2025-07-14T12:34:23.240352"
  },
  {
    "epoch": 1394,
    "batch_idx": 160,
    "loss": 0.4841785430908203,
    "learning_rate": 0.00020993277280407556,
    "timestep_mean": 453.625,
    "epsilon_norm": 8.174657821655273,
    "prediction_norm": 5.556310653686523,
    "timestamp": "2025-07-14T12:34:23.245154"
  },
  {
    "epoch": 1394,
    "batch_idx": 192,
    "loss": 0.18418774008750916,
    "learning_rate": 0.00020993277280407556,
    "timestep_mean": 650.0,
    "epsilon_norm": 2.8377668857574463,
    "prediction_norm": 2.3848092555999756,
    "timestamp": "2025-07-14T12:34:23.250084"
  },
  {
    "epoch": 1395,
    "batch_idx": 0,
    "loss": 0.2758864462375641,
    "learning_rate": 0.00020929340784008496,
    "timestep_mean": 581.46875,
    "epsilon_norm": 6.620378017425537,
    "prediction_norm": 5.552105903625488,
    "timestamp": "2025-07-14T12:34:23.254964"
  },
  {
    "epoch": 1395,
    "batch_idx": 32,
    "loss": 0.3056696057319641,
    "learning_rate": 0.00020929340784008496,
    "timestep_mean": 527.0,
    "epsilon_norm": 8.451894760131836,
    "prediction_norm": 6.228229999542236,
    "timestamp": "2025-07-14T12:34:23.259890"
  },
  {
    "epoch": 1395,
    "batch_idx": 64,
    "loss": 0.4068342447280884,
    "learning_rate": 0.00020929340784008496,
    "timestep_mean": 409.625,
    "epsilon_norm": 8.621108055114746,
    "prediction_norm": 6.230692386627197,
    "timestamp": "2025-07-14T12:34:23.264953"
  },
  {
    "epoch": 1395,
    "batch_idx": 96,
    "loss": 0.09428320825099945,
    "learning_rate": 0.00020929340784008496,
    "timestep_mean": 611.40625,
    "epsilon_norm": 7.425004959106445,
    "prediction_norm": 7.016586780548096,
    "timestamp": "2025-07-14T12:34:23.270044"
  },
  {
    "epoch": 1395,
    "batch_idx": 128,
    "loss": 0.2823668122291565,
    "learning_rate": 0.00020929340784008496,
    "timestep_mean": 565.4375,
    "epsilon_norm": 7.2105817794799805,
    "prediction_norm": 6.216802597045898,
    "timestamp": "2025-07-14T12:34:23.275256"
  },
  {
    "epoch": 1395,
    "batch_idx": 160,
    "loss": 0.29496800899505615,
    "learning_rate": 0.00020929340784008496,
    "timestep_mean": 487.96875,
    "epsilon_norm": 8.223663330078125,
    "prediction_norm": 5.916095733642578,
    "timestamp": "2025-07-14T12:34:23.280314"
  },
  {
    "epoch": 1395,
    "batch_idx": 192,
    "loss": 0.3599317669868469,
    "learning_rate": 0.00020929340784008496,
    "timestep_mean": 546.625,
    "epsilon_norm": 4.211015224456787,
    "prediction_norm": 3.7050058841705322,
    "timestamp": "2025-07-14T12:34:23.285349"
  },
  {
    "epoch": 1396,
    "batch_idx": 0,
    "loss": 0.15193329751491547,
    "learning_rate": 0.00020865476016571224,
    "timestep_mean": 541.0,
    "epsilon_norm": 7.754063606262207,
    "prediction_norm": 6.8798980712890625,
    "timestamp": "2025-07-14T12:34:23.290463"
  },
  {
    "epoch": 1396,
    "batch_idx": 32,
    "loss": 0.15361705422401428,
    "learning_rate": 0.00020865476016571224,
    "timestep_mean": 514.15625,
    "epsilon_norm": 6.980907440185547,
    "prediction_norm": 6.627655506134033,
    "timestamp": "2025-07-14T12:34:23.295593"
  },
  {
    "epoch": 1396,
    "batch_idx": 64,
    "loss": 0.3555413484573364,
    "learning_rate": 0.00020865476016571224,
    "timestep_mean": 504.53125,
    "epsilon_norm": 8.258709907531738,
    "prediction_norm": 6.0832366943359375,
    "timestamp": "2025-07-14T12:34:23.300754"
  },
  {
    "epoch": 1396,
    "batch_idx": 96,
    "loss": 0.1733459085226059,
    "learning_rate": 0.00020865476016571224,
    "timestep_mean": 552.25,
    "epsilon_norm": 7.654341220855713,
    "prediction_norm": 6.995357036590576,
    "timestamp": "2025-07-14T12:34:23.305883"
  },
  {
    "epoch": 1396,
    "batch_idx": 128,
    "loss": 0.2774151563644409,
    "learning_rate": 0.00020865476016571224,
    "timestep_mean": 500.5625,
    "epsilon_norm": 8.853090286254883,
    "prediction_norm": 6.981263160705566,
    "timestamp": "2025-07-14T12:34:23.310929"
  },
  {
    "epoch": 1396,
    "batch_idx": 160,
    "loss": 0.2803809642791748,
    "learning_rate": 0.00020865476016571224,
    "timestep_mean": 437.15625,
    "epsilon_norm": 8.007394790649414,
    "prediction_norm": 6.483603477478027,
    "timestamp": "2025-07-14T12:34:23.315960"
  },
  {
    "epoch": 1396,
    "batch_idx": 192,
    "loss": 0.12476831674575806,
    "learning_rate": 0.00020865476016571224,
    "timestep_mean": 455.25,
    "epsilon_norm": 3.924175262451172,
    "prediction_norm": 3.8250691890716553,
    "timestamp": "2025-07-14T12:34:23.320772"
  },
  {
    "epoch": 1397,
    "batch_idx": 0,
    "loss": 0.17769332230091095,
    "learning_rate": 0.00020801683135675709,
    "timestep_mean": 536.625,
    "epsilon_norm": 7.545292377471924,
    "prediction_norm": 6.585282325744629,
    "timestamp": "2025-07-14T12:34:23.325653"
  },
  {
    "epoch": 1397,
    "batch_idx": 32,
    "loss": 0.0937688872218132,
    "learning_rate": 0.00020801683135675709,
    "timestep_mean": 501.84375,
    "epsilon_norm": 7.604411602020264,
    "prediction_norm": 7.1616740226745605,
    "timestamp": "2025-07-14T12:34:23.330500"
  },
  {
    "epoch": 1397,
    "batch_idx": 64,
    "loss": 0.5372022390365601,
    "learning_rate": 0.00020801683135675709,
    "timestep_mean": 412.84375,
    "epsilon_norm": 8.398048400878906,
    "prediction_norm": 5.6508965492248535,
    "timestamp": "2025-07-14T12:34:23.335433"
  },
  {
    "epoch": 1397,
    "batch_idx": 96,
    "loss": 0.356365442276001,
    "learning_rate": 0.00020801683135675709,
    "timestep_mean": 415.46875,
    "epsilon_norm": 6.623077869415283,
    "prediction_norm": 5.1283087730407715,
    "timestamp": "2025-07-14T12:34:23.340265"
  },
  {
    "epoch": 1397,
    "batch_idx": 128,
    "loss": 0.24639317393302917,
    "learning_rate": 0.00020801683135675709,
    "timestep_mean": 519.9375,
    "epsilon_norm": 7.601032257080078,
    "prediction_norm": 6.403017044067383,
    "timestamp": "2025-07-14T12:34:23.345115"
  },
  {
    "epoch": 1397,
    "batch_idx": 160,
    "loss": 0.2547152638435364,
    "learning_rate": 0.00020801683135675709,
    "timestep_mean": 498.34375,
    "epsilon_norm": 5.869355201721191,
    "prediction_norm": 4.460048675537109,
    "timestamp": "2025-07-14T12:34:23.350074"
  },
  {
    "epoch": 1397,
    "batch_idx": 192,
    "loss": 0.14858105778694153,
    "learning_rate": 0.00020801683135675709,
    "timestep_mean": 634.75,
    "epsilon_norm": 4.502167224884033,
    "prediction_norm": 4.0242204666137695,
    "timestamp": "2025-07-14T12:34:23.354952"
  },
  {
    "epoch": 1398,
    "batch_idx": 0,
    "loss": 0.3536168038845062,
    "learning_rate": 0.00020737962298724534,
    "timestep_mean": 540.75,
    "epsilon_norm": 7.81830358505249,
    "prediction_norm": 6.074450492858887,
    "timestamp": "2025-07-14T12:34:23.359760"
  },
  {
    "epoch": 1398,
    "batch_idx": 32,
    "loss": 0.3153451085090637,
    "learning_rate": 0.00020737962298724534,
    "timestep_mean": 541.625,
    "epsilon_norm": 9.08205795288086,
    "prediction_norm": 8.133034706115723,
    "timestamp": "2025-07-14T12:34:23.364600"
  },
  {
    "epoch": 1398,
    "batch_idx": 64,
    "loss": 0.3687394857406616,
    "learning_rate": 0.00020737962298724534,
    "timestep_mean": 449.53125,
    "epsilon_norm": 8.396461486816406,
    "prediction_norm": 6.79381799697876,
    "timestamp": "2025-07-14T12:34:23.369443"
  },
  {
    "epoch": 1398,
    "batch_idx": 96,
    "loss": 0.23212307691574097,
    "learning_rate": 0.00020737962298724534,
    "timestep_mean": 503.65625,
    "epsilon_norm": 7.211635112762451,
    "prediction_norm": 6.467434406280518,
    "timestamp": "2025-07-14T12:34:23.374267"
  },
  {
    "epoch": 1398,
    "batch_idx": 128,
    "loss": 0.4803861975669861,
    "learning_rate": 0.00020737962298724534,
    "timestep_mean": 426.6875,
    "epsilon_norm": 8.071273803710938,
    "prediction_norm": 6.611340522766113,
    "timestamp": "2025-07-14T12:34:23.379106"
  },
  {
    "epoch": 1398,
    "batch_idx": 160,
    "loss": 0.31096965074539185,
    "learning_rate": 0.00020737962298724534,
    "timestep_mean": 520.75,
    "epsilon_norm": 8.248597145080566,
    "prediction_norm": 6.8733391761779785,
    "timestamp": "2025-07-14T12:34:23.383956"
  },
  {
    "epoch": 1398,
    "batch_idx": 192,
    "loss": 0.23647207021713257,
    "learning_rate": 0.00020737962298724534,
    "timestep_mean": 606.375,
    "epsilon_norm": 4.328579902648926,
    "prediction_norm": 3.1948063373565674,
    "timestamp": "2025-07-14T12:34:23.388791"
  },
  {
    "epoch": 1399,
    "batch_idx": 0,
    "loss": 0.29227596521377563,
    "learning_rate": 0.00020674313662942515,
    "timestep_mean": 533.59375,
    "epsilon_norm": 7.863425254821777,
    "prediction_norm": 6.296780586242676,
    "timestamp": "2025-07-14T12:34:23.393706"
  },
  {
    "epoch": 1399,
    "batch_idx": 32,
    "loss": 0.2100927233695984,
    "learning_rate": 0.00020674313662942515,
    "timestep_mean": 471.5625,
    "epsilon_norm": 8.238673210144043,
    "prediction_norm": 7.412198543548584,
    "timestamp": "2025-07-14T12:34:23.398541"
  },
  {
    "epoch": 1399,
    "batch_idx": 64,
    "loss": 0.3672134280204773,
    "learning_rate": 0.00020674313662942515,
    "timestep_mean": 463.3125,
    "epsilon_norm": 7.716738700866699,
    "prediction_norm": 6.094611167907715,
    "timestamp": "2025-07-14T12:34:23.403445"
  },
  {
    "epoch": 1399,
    "batch_idx": 96,
    "loss": 0.34449559450149536,
    "learning_rate": 0.00020674313662942515,
    "timestep_mean": 476.625,
    "epsilon_norm": 7.897403717041016,
    "prediction_norm": 6.346578121185303,
    "timestamp": "2025-07-14T12:34:23.408418"
  },
  {
    "epoch": 1399,
    "batch_idx": 128,
    "loss": 0.18678490817546844,
    "learning_rate": 0.00020674313662942515,
    "timestep_mean": 578.75,
    "epsilon_norm": 7.235611915588379,
    "prediction_norm": 6.383122444152832,
    "timestamp": "2025-07-14T12:34:23.413264"
  },
  {
    "epoch": 1399,
    "batch_idx": 160,
    "loss": 0.1558288037776947,
    "learning_rate": 0.00020674313662942515,
    "timestep_mean": 460.84375,
    "epsilon_norm": 6.636023044586182,
    "prediction_norm": 5.9583001136779785,
    "timestamp": "2025-07-14T12:34:23.418399"
  },
  {
    "epoch": 1399,
    "batch_idx": 192,
    "loss": 0.38048025965690613,
    "learning_rate": 0.00020674313662942515,
    "timestep_mean": 340.625,
    "epsilon_norm": 3.693176031112671,
    "prediction_norm": 2.6400816440582275,
    "timestamp": "2025-07-14T12:34:23.423565"
  },
  {
    "epoch": 1400,
    "batch_idx": 0,
    "loss": 0.6471366882324219,
    "learning_rate": 0.00020610737385376375,
    "timestep_mean": 361.09375,
    "epsilon_norm": 9.186918258666992,
    "prediction_norm": 6.4163079261779785,
    "timestamp": "2025-07-14T12:34:23.428804"
  },
  {
    "epoch": 1400,
    "batch_idx": 32,
    "loss": 0.18407167494297028,
    "learning_rate": 0.00020610737385376375,
    "timestep_mean": 502.375,
    "epsilon_norm": 7.653639793395996,
    "prediction_norm": 6.37503719329834,
    "timestamp": "2025-07-14T12:34:23.433802"
  },
  {
    "epoch": 1400,
    "batch_idx": 64,
    "loss": 0.16175132989883423,
    "learning_rate": 0.00020610737385376375,
    "timestep_mean": 563.1875,
    "epsilon_norm": 8.264022827148438,
    "prediction_norm": 7.707156181335449,
    "timestamp": "2025-07-14T12:34:23.438809"
  },
  {
    "epoch": 1400,
    "batch_idx": 96,
    "loss": 0.48147881031036377,
    "learning_rate": 0.00020610737385376375,
    "timestep_mean": 496.40625,
    "epsilon_norm": 7.985796928405762,
    "prediction_norm": 5.891627788543701,
    "timestamp": "2025-07-14T12:34:23.444467"
  },
  {
    "epoch": 1400,
    "batch_idx": 128,
    "loss": 0.2022264301776886,
    "learning_rate": 0.00020610737385376375,
    "timestep_mean": 571.5625,
    "epsilon_norm": 7.744884490966797,
    "prediction_norm": 7.05210542678833,
    "timestamp": "2025-07-14T12:34:23.449761"
  },
  {
    "epoch": 1400,
    "batch_idx": 160,
    "loss": 0.41553735733032227,
    "learning_rate": 0.00020610737385376375,
    "timestep_mean": 460.6875,
    "epsilon_norm": 7.295699596405029,
    "prediction_norm": 4.9658403396606445,
    "timestamp": "2025-07-14T12:34:23.454882"
  },
  {
    "epoch": 1400,
    "batch_idx": 192,
    "loss": 0.06303300708532333,
    "learning_rate": 0.00020610737385376375,
    "timestep_mean": 479.75,
    "epsilon_norm": 3.2548203468322754,
    "prediction_norm": 3.5069308280944824,
    "timestamp": "2025-07-14T12:34:23.460088"
  },
  {
    "epoch": 1401,
    "batch_idx": 0,
    "loss": 0.44480210542678833,
    "learning_rate": 0.00020547233622894234,
    "timestep_mean": 424.125,
    "epsilon_norm": 8.699063301086426,
    "prediction_norm": 6.863938331604004,
    "timestamp": "2025-07-14T12:34:23.465041"
  },
  {
    "epoch": 1401,
    "batch_idx": 32,
    "loss": 0.2624276280403137,
    "learning_rate": 0.00020547233622894234,
    "timestep_mean": 537.125,
    "epsilon_norm": 8.837023735046387,
    "prediction_norm": 8.139970779418945,
    "timestamp": "2025-07-14T12:34:23.470015"
  },
  {
    "epoch": 1401,
    "batch_idx": 64,
    "loss": 0.21302883327007294,
    "learning_rate": 0.00020547233622894234,
    "timestep_mean": 510.03125,
    "epsilon_norm": 8.205720901489258,
    "prediction_norm": 7.189503192901611,
    "timestamp": "2025-07-14T12:34:23.475020"
  },
  {
    "epoch": 1401,
    "batch_idx": 96,
    "loss": 0.24490514397621155,
    "learning_rate": 0.00020547233622894234,
    "timestep_mean": 563.3125,
    "epsilon_norm": 8.024823188781738,
    "prediction_norm": 7.002919673919678,
    "timestamp": "2025-07-14T12:34:23.479957"
  },
  {
    "epoch": 1401,
    "batch_idx": 128,
    "loss": 0.24765419960021973,
    "learning_rate": 0.00020547233622894234,
    "timestep_mean": 398.5625,
    "epsilon_norm": 6.98296594619751,
    "prediction_norm": 5.86244010925293,
    "timestamp": "2025-07-14T12:34:23.484833"
  },
  {
    "epoch": 1401,
    "batch_idx": 160,
    "loss": 0.34054720401763916,
    "learning_rate": 0.00020547233622894234,
    "timestep_mean": 526.125,
    "epsilon_norm": 7.986883640289307,
    "prediction_norm": 6.118912220001221,
    "timestamp": "2025-07-14T12:34:23.489785"
  },
  {
    "epoch": 1401,
    "batch_idx": 192,
    "loss": 0.49303945899009705,
    "learning_rate": 0.00020547233622894234,
    "timestep_mean": 424.75,
    "epsilon_norm": 4.046136856079102,
    "prediction_norm": 3.271130323410034,
    "timestamp": "2025-07-14T12:34:23.494742"
  },
  {
    "epoch": 1402,
    "batch_idx": 0,
    "loss": 0.3105090260505676,
    "learning_rate": 0.0002048380253218529,
    "timestep_mean": 480.5,
    "epsilon_norm": 8.619736671447754,
    "prediction_norm": 7.222568988800049,
    "timestamp": "2025-07-14T12:34:23.499632"
  },
  {
    "epoch": 1402,
    "batch_idx": 32,
    "loss": 0.3075034022331238,
    "learning_rate": 0.0002048380253218529,
    "timestep_mean": 557.25,
    "epsilon_norm": 7.716922283172607,
    "prediction_norm": 6.189360618591309,
    "timestamp": "2025-07-14T12:34:23.504647"
  },
  {
    "epoch": 1402,
    "batch_idx": 64,
    "loss": 0.2610960304737091,
    "learning_rate": 0.0002048380253218529,
    "timestep_mean": 514.28125,
    "epsilon_norm": 7.138556480407715,
    "prediction_norm": 6.504031658172607,
    "timestamp": "2025-07-14T12:34:23.509599"
  },
  {
    "epoch": 1402,
    "batch_idx": 96,
    "loss": 0.24651341140270233,
    "learning_rate": 0.0002048380253218529,
    "timestep_mean": 417.84375,
    "epsilon_norm": 7.722693920135498,
    "prediction_norm": 7.549032688140869,
    "timestamp": "2025-07-14T12:34:23.514538"
  },
  {
    "epoch": 1402,
    "batch_idx": 128,
    "loss": 0.25329816341400146,
    "learning_rate": 0.0002048380253218529,
    "timestep_mean": 530.5,
    "epsilon_norm": 8.86535358428955,
    "prediction_norm": 7.6360249519348145,
    "timestamp": "2025-07-14T12:34:23.519420"
  },
  {
    "epoch": 1402,
    "batch_idx": 160,
    "loss": 0.2495613545179367,
    "learning_rate": 0.0002048380253218529,
    "timestep_mean": 476.625,
    "epsilon_norm": 6.786930084228516,
    "prediction_norm": 5.728259086608887,
    "timestamp": "2025-07-14T12:34:23.524330"
  },
  {
    "epoch": 1402,
    "batch_idx": 192,
    "loss": 0.04011355713009834,
    "learning_rate": 0.0002048380253218529,
    "timestep_mean": 553.875,
    "epsilon_norm": 4.223268032073975,
    "prediction_norm": 4.513242721557617,
    "timestamp": "2025-07-14T12:34:23.529192"
  },
  {
    "epoch": 1403,
    "batch_idx": 0,
    "loss": 0.3147018551826477,
    "learning_rate": 0.00020420444269759523,
    "timestep_mean": 520.15625,
    "epsilon_norm": 8.32307243347168,
    "prediction_norm": 7.363570213317871,
    "timestamp": "2025-07-14T12:34:23.534109"
  },
  {
    "epoch": 1403,
    "batch_idx": 32,
    "loss": 0.20988139510154724,
    "learning_rate": 0.00020420444269759523,
    "timestep_mean": 529.65625,
    "epsilon_norm": 7.627564907073975,
    "prediction_norm": 6.8353776931762695,
    "timestamp": "2025-07-14T12:34:23.538960"
  },
  {
    "epoch": 1403,
    "batch_idx": 64,
    "loss": 0.32394739985466003,
    "learning_rate": 0.00020420444269759523,
    "timestep_mean": 515.125,
    "epsilon_norm": 7.834706783294678,
    "prediction_norm": 6.896629333496094,
    "timestamp": "2025-07-14T12:34:23.543806"
  },
  {
    "epoch": 1403,
    "batch_idx": 96,
    "loss": 0.25641292333602905,
    "learning_rate": 0.00020420444269759523,
    "timestep_mean": 553.53125,
    "epsilon_norm": 6.959366321563721,
    "prediction_norm": 6.669122695922852,
    "timestamp": "2025-07-14T12:34:23.548701"
  },
  {
    "epoch": 1403,
    "batch_idx": 128,
    "loss": 0.26151037216186523,
    "learning_rate": 0.00020420444269759523,
    "timestep_mean": 441.75,
    "epsilon_norm": 7.4588093757629395,
    "prediction_norm": 7.229885578155518,
    "timestamp": "2025-07-14T12:34:23.553570"
  },
  {
    "epoch": 1403,
    "batch_idx": 160,
    "loss": 0.15601040422916412,
    "learning_rate": 0.00020420444269759523,
    "timestep_mean": 593.3125,
    "epsilon_norm": 7.498407363891602,
    "prediction_norm": 6.6501007080078125,
    "timestamp": "2025-07-14T12:34:23.558532"
  },
  {
    "epoch": 1403,
    "batch_idx": 192,
    "loss": 0.05833711102604866,
    "learning_rate": 0.00020420444269759523,
    "timestep_mean": 537.5,
    "epsilon_norm": 3.626209259033203,
    "prediction_norm": 3.3447670936584473,
    "timestamp": "2025-07-14T12:34:23.563438"
  },
  {
    "epoch": 1404,
    "batch_idx": 0,
    "loss": 0.525762140750885,
    "learning_rate": 0.00020357158991947063,
    "timestep_mean": 381.9375,
    "epsilon_norm": 8.49856948852539,
    "prediction_norm": 6.106054306030273,
    "timestamp": "2025-07-14T12:34:23.568396"
  },
  {
    "epoch": 1404,
    "batch_idx": 32,
    "loss": 0.28111541271209717,
    "learning_rate": 0.00020357158991947063,
    "timestep_mean": 427.6875,
    "epsilon_norm": 7.758358001708984,
    "prediction_norm": 6.450188159942627,
    "timestamp": "2025-07-14T12:34:23.573291"
  },
  {
    "epoch": 1404,
    "batch_idx": 64,
    "loss": 0.49314504861831665,
    "learning_rate": 0.00020357158991947063,
    "timestep_mean": 458.09375,
    "epsilon_norm": 8.7827730178833,
    "prediction_norm": 6.98813009262085,
    "timestamp": "2025-07-14T12:34:23.578166"
  },
  {
    "epoch": 1404,
    "batch_idx": 96,
    "loss": 0.16050858795642853,
    "learning_rate": 0.00020357158991947063,
    "timestep_mean": 507.625,
    "epsilon_norm": 8.139589309692383,
    "prediction_norm": 7.266491889953613,
    "timestamp": "2025-07-14T12:34:23.583050"
  },
  {
    "epoch": 1404,
    "batch_idx": 128,
    "loss": 0.1340702772140503,
    "learning_rate": 0.00020357158991947063,
    "timestep_mean": 500.40625,
    "epsilon_norm": 8.03564167022705,
    "prediction_norm": 7.277109622955322,
    "timestamp": "2025-07-14T12:34:23.587980"
  },
  {
    "epoch": 1404,
    "batch_idx": 160,
    "loss": 0.28955408930778503,
    "learning_rate": 0.00020357158991947063,
    "timestep_mean": 482.4375,
    "epsilon_norm": 8.498716354370117,
    "prediction_norm": 7.443081378936768,
    "timestamp": "2025-07-14T12:34:23.592899"
  },
  {
    "epoch": 1404,
    "batch_idx": 192,
    "loss": 0.9545350670814514,
    "learning_rate": 0.00020357158991947063,
    "timestep_mean": 340.0,
    "epsilon_norm": 5.493602275848389,
    "prediction_norm": 3.325000047683716,
    "timestamp": "2025-07-14T12:34:23.598043"
  },
  {
    "epoch": 1405,
    "batch_idx": 0,
    "loss": 0.3432263135910034,
    "learning_rate": 0.00020293946854898085,
    "timestep_mean": 510.0625,
    "epsilon_norm": 9.257651329040527,
    "prediction_norm": 7.177622318267822,
    "timestamp": "2025-07-14T12:34:23.603145"
  },
  {
    "epoch": 1405,
    "batch_idx": 32,
    "loss": 0.2511773705482483,
    "learning_rate": 0.00020293946854898085,
    "timestep_mean": 579.1875,
    "epsilon_norm": 9.784406661987305,
    "prediction_norm": 8.625824928283691,
    "timestamp": "2025-07-14T12:34:23.608225"
  },
  {
    "epoch": 1405,
    "batch_idx": 64,
    "loss": 0.15330198407173157,
    "learning_rate": 0.00020293946854898085,
    "timestep_mean": 539.1875,
    "epsilon_norm": 6.656955242156982,
    "prediction_norm": 6.3261332511901855,
    "timestamp": "2025-07-14T12:34:23.613373"
  },
  {
    "epoch": 1405,
    "batch_idx": 96,
    "loss": 0.3193800449371338,
    "learning_rate": 0.00020293946854898085,
    "timestep_mean": 532.34375,
    "epsilon_norm": 8.580732345581055,
    "prediction_norm": 7.946920871734619,
    "timestamp": "2025-07-14T12:34:23.618353"
  },
  {
    "epoch": 1405,
    "batch_idx": 128,
    "loss": 0.3393329977989197,
    "learning_rate": 0.00020293946854898085,
    "timestep_mean": 526.15625,
    "epsilon_norm": 7.791905879974365,
    "prediction_norm": 6.471884250640869,
    "timestamp": "2025-07-14T12:34:23.623299"
  },
  {
    "epoch": 1405,
    "batch_idx": 160,
    "loss": 0.29355692863464355,
    "learning_rate": 0.00020293946854898085,
    "timestep_mean": 504.25,
    "epsilon_norm": 7.706523418426514,
    "prediction_norm": 6.271576404571533,
    "timestamp": "2025-07-14T12:34:23.628194"
  },
  {
    "epoch": 1405,
    "batch_idx": 192,
    "loss": 0.19174353778362274,
    "learning_rate": 0.00020293946854898085,
    "timestep_mean": 632.125,
    "epsilon_norm": 5.537870407104492,
    "prediction_norm": 4.474209308624268,
    "timestamp": "2025-07-14T12:34:23.633230"
  },
  {
    "epoch": 1406,
    "batch_idx": 0,
    "loss": 0.5862858295440674,
    "learning_rate": 0.0002023080801458229,
    "timestep_mean": 457.84375,
    "epsilon_norm": 7.11714506149292,
    "prediction_norm": 5.070316314697266,
    "timestamp": "2025-07-14T12:34:23.638252"
  },
  {
    "epoch": 1406,
    "batch_idx": 32,
    "loss": 0.1599523276090622,
    "learning_rate": 0.0002023080801458229,
    "timestep_mean": 635.53125,
    "epsilon_norm": 7.305866718292236,
    "prediction_norm": 6.580515384674072,
    "timestamp": "2025-07-14T12:34:23.643166"
  },
  {
    "epoch": 1406,
    "batch_idx": 64,
    "loss": 0.3674381375312805,
    "learning_rate": 0.0002023080801458229,
    "timestep_mean": 504.0625,
    "epsilon_norm": 8.315811157226562,
    "prediction_norm": 6.804566860198975,
    "timestamp": "2025-07-14T12:34:23.648082"
  },
  {
    "epoch": 1406,
    "batch_idx": 96,
    "loss": 0.28239086270332336,
    "learning_rate": 0.0002023080801458229,
    "timestep_mean": 481.3125,
    "epsilon_norm": 8.084039688110352,
    "prediction_norm": 7.393318176269531,
    "timestamp": "2025-07-14T12:34:23.652931"
  },
  {
    "epoch": 1406,
    "batch_idx": 128,
    "loss": 0.2956085801124573,
    "learning_rate": 0.0002023080801458229,
    "timestep_mean": 521.0625,
    "epsilon_norm": 7.777235984802246,
    "prediction_norm": 6.510044574737549,
    "timestamp": "2025-07-14T12:34:23.657815"
  },
  {
    "epoch": 1406,
    "batch_idx": 160,
    "loss": 0.423821359872818,
    "learning_rate": 0.0002023080801458229,
    "timestep_mean": 409.15625,
    "epsilon_norm": 7.875639915466309,
    "prediction_norm": 6.290434837341309,
    "timestamp": "2025-07-14T12:34:23.662773"
  },
  {
    "epoch": 1406,
    "batch_idx": 192,
    "loss": 0.23951424658298492,
    "learning_rate": 0.0002023080801458229,
    "timestep_mean": 398.0,
    "epsilon_norm": 3.3589186668395996,
    "prediction_norm": 2.9132330417633057,
    "timestamp": "2025-07-14T12:34:23.667690"
  },
  {
    "epoch": 1407,
    "batch_idx": 0,
    "loss": 0.17768125236034393,
    "learning_rate": 0.00020167742626788407,
    "timestep_mean": 493.21875,
    "epsilon_norm": 6.835495471954346,
    "prediction_norm": 5.956589221954346,
    "timestamp": "2025-07-14T12:34:23.672599"
  },
  {
    "epoch": 1407,
    "batch_idx": 32,
    "loss": 0.2612559497356415,
    "learning_rate": 0.00020167742626788407,
    "timestep_mean": 473.59375,
    "epsilon_norm": 7.883469104766846,
    "prediction_norm": 6.323709011077881,
    "timestamp": "2025-07-14T12:34:23.677463"
  },
  {
    "epoch": 1407,
    "batch_idx": 64,
    "loss": 0.31467515230178833,
    "learning_rate": 0.00020167742626788407,
    "timestep_mean": 453.375,
    "epsilon_norm": 8.749868392944336,
    "prediction_norm": 7.090771198272705,
    "timestamp": "2025-07-14T12:34:23.682327"
  },
  {
    "epoch": 1407,
    "batch_idx": 96,
    "loss": 0.31870079040527344,
    "learning_rate": 0.00020167742626788407,
    "timestep_mean": 427.65625,
    "epsilon_norm": 6.55091667175293,
    "prediction_norm": 5.0791239738464355,
    "timestamp": "2025-07-14T12:34:23.687224"
  },
  {
    "epoch": 1407,
    "batch_idx": 128,
    "loss": 0.2272762954235077,
    "learning_rate": 0.00020167742626788407,
    "timestep_mean": 604.53125,
    "epsilon_norm": 8.35313606262207,
    "prediction_norm": 7.46876335144043,
    "timestamp": "2025-07-14T12:34:23.692189"
  },
  {
    "epoch": 1407,
    "batch_idx": 160,
    "loss": 0.331615686416626,
    "learning_rate": 0.00020167742626788407,
    "timestep_mean": 455.28125,
    "epsilon_norm": 8.464409828186035,
    "prediction_norm": 7.0920281410217285,
    "timestamp": "2025-07-14T12:34:23.697085"
  },
  {
    "epoch": 1407,
    "batch_idx": 192,
    "loss": 0.32670313119888306,
    "learning_rate": 0.00020167742626788407,
    "timestep_mean": 568.625,
    "epsilon_norm": 3.958916664123535,
    "prediction_norm": 3.199373483657837,
    "timestamp": "2025-07-14T12:34:23.701940"
  },
  {
    "epoch": 1408,
    "batch_idx": 0,
    "loss": 0.33905574679374695,
    "learning_rate": 0.00020104750847124098,
    "timestep_mean": 513.34375,
    "epsilon_norm": 7.660887718200684,
    "prediction_norm": 7.146249294281006,
    "timestamp": "2025-07-14T12:34:23.706827"
  },
  {
    "epoch": 1408,
    "batch_idx": 32,
    "loss": 0.2360478639602661,
    "learning_rate": 0.00020104750847124098,
    "timestep_mean": 543.5,
    "epsilon_norm": 8.119553565979004,
    "prediction_norm": 6.686443328857422,
    "timestamp": "2025-07-14T12:34:23.711689"
  },
  {
    "epoch": 1408,
    "batch_idx": 64,
    "loss": 0.3146326541900635,
    "learning_rate": 0.00020104750847124098,
    "timestep_mean": 493.71875,
    "epsilon_norm": 8.364529609680176,
    "prediction_norm": 7.22517728805542,
    "timestamp": "2025-07-14T12:34:23.716697"
  },
  {
    "epoch": 1408,
    "batch_idx": 96,
    "loss": 0.1548299789428711,
    "learning_rate": 0.00020104750847124098,
    "timestep_mean": 538.625,
    "epsilon_norm": 6.329456806182861,
    "prediction_norm": 5.693953990936279,
    "timestamp": "2025-07-14T12:34:23.721577"
  },
  {
    "epoch": 1408,
    "batch_idx": 128,
    "loss": 0.154626727104187,
    "learning_rate": 0.00020104750847124098,
    "timestep_mean": 517.75,
    "epsilon_norm": 7.868821144104004,
    "prediction_norm": 7.299771785736084,
    "timestamp": "2025-07-14T12:34:23.726427"
  },
  {
    "epoch": 1408,
    "batch_idx": 160,
    "loss": 0.13347776234149933,
    "learning_rate": 0.00020104750847124098,
    "timestep_mean": 565.5,
    "epsilon_norm": 8.024948120117188,
    "prediction_norm": 7.221297740936279,
    "timestamp": "2025-07-14T12:34:23.731334"
  },
  {
    "epoch": 1408,
    "batch_idx": 192,
    "loss": 0.11400839686393738,
    "learning_rate": 0.00020104750847124098,
    "timestep_mean": 491.75,
    "epsilon_norm": 2.777311086654663,
    "prediction_norm": 2.1218273639678955,
    "timestamp": "2025-07-14T12:34:23.736495"
  },
  {
    "epoch": 1409,
    "batch_idx": 0,
    "loss": 0.13151293992996216,
    "learning_rate": 0.00020041832831015234,
    "timestep_mean": 495.375,
    "epsilon_norm": 8.061702728271484,
    "prediction_norm": 7.296111106872559,
    "timestamp": "2025-07-14T12:34:23.741549"
  },
  {
    "epoch": 1409,
    "batch_idx": 32,
    "loss": 0.24202793836593628,
    "learning_rate": 0.00020041832831015234,
    "timestep_mean": 621.96875,
    "epsilon_norm": 8.713959693908691,
    "prediction_norm": 7.201879978179932,
    "timestamp": "2025-07-14T12:34:23.746874"
  },
  {
    "epoch": 1409,
    "batch_idx": 64,
    "loss": 0.2172679305076599,
    "learning_rate": 0.00020041832831015234,
    "timestep_mean": 433.6875,
    "epsilon_norm": 8.737092018127441,
    "prediction_norm": 7.82363224029541,
    "timestamp": "2025-07-14T12:34:23.751871"
  },
  {
    "epoch": 1409,
    "batch_idx": 96,
    "loss": 0.27927136421203613,
    "learning_rate": 0.00020041832831015234,
    "timestep_mean": 481.96875,
    "epsilon_norm": 8.56714153289795,
    "prediction_norm": 7.253695964813232,
    "timestamp": "2025-07-14T12:34:23.756772"
  },
  {
    "epoch": 1409,
    "batch_idx": 128,
    "loss": 0.3498510718345642,
    "learning_rate": 0.00020041832831015234,
    "timestep_mean": 494.9375,
    "epsilon_norm": 8.075818061828613,
    "prediction_norm": 6.392916679382324,
    "timestamp": "2025-07-14T12:34:23.761736"
  },
  {
    "epoch": 1409,
    "batch_idx": 160,
    "loss": 0.25533348321914673,
    "learning_rate": 0.00020041832831015234,
    "timestep_mean": 511.03125,
    "epsilon_norm": 9.443601608276367,
    "prediction_norm": 8.314356803894043,
    "timestamp": "2025-07-14T12:34:23.766743"
  },
  {
    "epoch": 1409,
    "batch_idx": 192,
    "loss": 0.569744348526001,
    "learning_rate": 0.00020041832831015234,
    "timestep_mean": 542.75,
    "epsilon_norm": 5.758398056030273,
    "prediction_norm": 4.409073352813721,
    "timestamp": "2025-07-14T12:34:23.771768"
  },
  {
    "epoch": 1410,
    "batch_idx": 0,
    "loss": 0.2815486490726471,
    "learning_rate": 0.0001997898873370583,
    "timestep_mean": 490.875,
    "epsilon_norm": 8.774846076965332,
    "prediction_norm": 7.357916831970215,
    "timestamp": "2025-07-14T12:34:23.776747"
  },
  {
    "epoch": 1410,
    "batch_idx": 32,
    "loss": 0.2386932075023651,
    "learning_rate": 0.0001997898873370583,
    "timestep_mean": 558.46875,
    "epsilon_norm": 8.15782356262207,
    "prediction_norm": 7.359730243682861,
    "timestamp": "2025-07-14T12:34:23.781605"
  },
  {
    "epoch": 1410,
    "batch_idx": 64,
    "loss": 0.31220290064811707,
    "learning_rate": 0.0001997898873370583,
    "timestep_mean": 509.09375,
    "epsilon_norm": 7.357908725738525,
    "prediction_norm": 5.232586860656738,
    "timestamp": "2025-07-14T12:34:23.786547"
  },
  {
    "epoch": 1410,
    "batch_idx": 96,
    "loss": 0.28325170278549194,
    "learning_rate": 0.0001997898873370583,
    "timestep_mean": 456.84375,
    "epsilon_norm": 7.4838690757751465,
    "prediction_norm": 6.157962322235107,
    "timestamp": "2025-07-14T12:34:23.791430"
  },
  {
    "epoch": 1410,
    "batch_idx": 128,
    "loss": 0.4491319954395294,
    "learning_rate": 0.0001997898873370583,
    "timestep_mean": 476.21875,
    "epsilon_norm": 8.511025428771973,
    "prediction_norm": 6.968875885009766,
    "timestamp": "2025-07-14T12:34:23.796455"
  },
  {
    "epoch": 1410,
    "batch_idx": 160,
    "loss": 0.26937663555145264,
    "learning_rate": 0.0001997898873370583,
    "timestep_mean": 468.96875,
    "epsilon_norm": 7.992917060852051,
    "prediction_norm": 6.540907859802246,
    "timestamp": "2025-07-14T12:34:23.801351"
  },
  {
    "epoch": 1410,
    "batch_idx": 192,
    "loss": 0.0960293710231781,
    "learning_rate": 0.0001997898873370583,
    "timestep_mean": 628.875,
    "epsilon_norm": 4.5029778480529785,
    "prediction_norm": 4.195678234100342,
    "timestamp": "2025-07-14T12:34:23.806260"
  },
  {
    "epoch": 1411,
    "batch_idx": 0,
    "loss": 0.23588252067565918,
    "learning_rate": 0.0001991621871025742,
    "timestep_mean": 498.40625,
    "epsilon_norm": 8.24416732788086,
    "prediction_norm": 7.0008673667907715,
    "timestamp": "2025-07-14T12:34:23.811154"
  },
  {
    "epoch": 1411,
    "batch_idx": 32,
    "loss": 0.29556024074554443,
    "learning_rate": 0.0001991621871025742,
    "timestep_mean": 524.28125,
    "epsilon_norm": 7.820163249969482,
    "prediction_norm": 7.267807483673096,
    "timestamp": "2025-07-14T12:34:23.816090"
  },
  {
    "epoch": 1411,
    "batch_idx": 64,
    "loss": 0.33881744742393494,
    "learning_rate": 0.0001991621871025742,
    "timestep_mean": 523.09375,
    "epsilon_norm": 7.660709857940674,
    "prediction_norm": 6.083721160888672,
    "timestamp": "2025-07-14T12:34:23.820939"
  },
  {
    "epoch": 1411,
    "batch_idx": 96,
    "loss": 0.24587005376815796,
    "learning_rate": 0.0001991621871025742,
    "timestep_mean": 525.125,
    "epsilon_norm": 7.209700584411621,
    "prediction_norm": 5.820552825927734,
    "timestamp": "2025-07-14T12:34:23.825882"
  },
  {
    "epoch": 1411,
    "batch_idx": 128,
    "loss": 0.14280639588832855,
    "learning_rate": 0.0001991621871025742,
    "timestep_mean": 552.0,
    "epsilon_norm": 7.014842987060547,
    "prediction_norm": 6.23522424697876,
    "timestamp": "2025-07-14T12:34:23.831060"
  },
  {
    "epoch": 1411,
    "batch_idx": 160,
    "loss": 0.28010624647140503,
    "learning_rate": 0.0001991621871025742,
    "timestep_mean": 474.46875,
    "epsilon_norm": 7.649811267852783,
    "prediction_norm": 6.446218490600586,
    "timestamp": "2025-07-14T12:34:23.836199"
  },
  {
    "epoch": 1411,
    "batch_idx": 192,
    "loss": 0.021834850311279297,
    "learning_rate": 0.0001991621871025742,
    "timestep_mean": 485.5,
    "epsilon_norm": 3.61921763420105,
    "prediction_norm": 3.584308385848999,
    "timestamp": "2025-07-14T12:34:23.841462"
  },
  {
    "epoch": 1412,
    "batch_idx": 0,
    "loss": 0.16522730886936188,
    "learning_rate": 0.00019853522915548786,
    "timestep_mean": 587.53125,
    "epsilon_norm": 7.190845012664795,
    "prediction_norm": 6.877511501312256,
    "timestamp": "2025-07-14T12:34:23.846795"
  },
  {
    "epoch": 1412,
    "batch_idx": 32,
    "loss": 0.21387530863285065,
    "learning_rate": 0.00019853522915548786,
    "timestep_mean": 621.9375,
    "epsilon_norm": 9.019804000854492,
    "prediction_norm": 7.979282855987549,
    "timestamp": "2025-07-14T12:34:23.851955"
  },
  {
    "epoch": 1412,
    "batch_idx": 64,
    "loss": 0.22045499086380005,
    "learning_rate": 0.00019853522915548786,
    "timestep_mean": 570.59375,
    "epsilon_norm": 8.213634490966797,
    "prediction_norm": 6.975428581237793,
    "timestamp": "2025-07-14T12:34:23.857164"
  },
  {
    "epoch": 1412,
    "batch_idx": 96,
    "loss": 0.1539095640182495,
    "learning_rate": 0.00019853522915548786,
    "timestep_mean": 557.40625,
    "epsilon_norm": 6.823179721832275,
    "prediction_norm": 6.45540714263916,
    "timestamp": "2025-07-14T12:34:23.862383"
  },
  {
    "epoch": 1412,
    "batch_idx": 128,
    "loss": 0.4141533076763153,
    "learning_rate": 0.00019853522915548786,
    "timestep_mean": 531.0625,
    "epsilon_norm": 7.50714635848999,
    "prediction_norm": 6.0238728523254395,
    "timestamp": "2025-07-14T12:34:23.867598"
  },
  {
    "epoch": 1412,
    "batch_idx": 160,
    "loss": 0.2757352590560913,
    "learning_rate": 0.00019853522915548786,
    "timestep_mean": 479.375,
    "epsilon_norm": 8.265314102172852,
    "prediction_norm": 7.034643650054932,
    "timestamp": "2025-07-14T12:34:23.872825"
  },
  {
    "epoch": 1412,
    "batch_idx": 192,
    "loss": 0.09819239377975464,
    "learning_rate": 0.00019853522915548786,
    "timestep_mean": 625.875,
    "epsilon_norm": 3.921908140182495,
    "prediction_norm": 3.8189241886138916,
    "timestamp": "2025-07-14T12:34:23.878041"
  },
  {
    "epoch": 1413,
    "batch_idx": 0,
    "loss": 0.280150830745697,
    "learning_rate": 0.00019790901504275601,
    "timestep_mean": 438.34375,
    "epsilon_norm": 7.652149200439453,
    "prediction_norm": 6.501628875732422,
    "timestamp": "2025-07-14T12:34:23.883210"
  },
  {
    "epoch": 1413,
    "batch_idx": 32,
    "loss": 0.19027341902256012,
    "learning_rate": 0.00019790901504275601,
    "timestep_mean": 560.3125,
    "epsilon_norm": 7.727947235107422,
    "prediction_norm": 6.83071231842041,
    "timestamp": "2025-07-14T12:34:23.888654"
  },
  {
    "epoch": 1413,
    "batch_idx": 64,
    "loss": 0.39038413763046265,
    "learning_rate": 0.00019790901504275601,
    "timestep_mean": 451.4375,
    "epsilon_norm": 8.331281661987305,
    "prediction_norm": 7.111809730529785,
    "timestamp": "2025-07-14T12:34:23.893940"
  },
  {
    "epoch": 1413,
    "batch_idx": 96,
    "loss": 0.2903655767440796,
    "learning_rate": 0.00019790901504275601,
    "timestep_mean": 534.65625,
    "epsilon_norm": 6.307775974273682,
    "prediction_norm": 5.547804355621338,
    "timestamp": "2025-07-14T12:34:23.899188"
  },
  {
    "epoch": 1413,
    "batch_idx": 128,
    "loss": 0.34875863790512085,
    "learning_rate": 0.00019790901504275601,
    "timestep_mean": 459.65625,
    "epsilon_norm": 8.174101829528809,
    "prediction_norm": 6.682210445404053,
    "timestamp": "2025-07-14T12:34:23.904340"
  },
  {
    "epoch": 1413,
    "batch_idx": 160,
    "loss": 0.34415972232818604,
    "learning_rate": 0.00019790901504275601,
    "timestep_mean": 507.9375,
    "epsilon_norm": 9.225471496582031,
    "prediction_norm": 7.6528778076171875,
    "timestamp": "2025-07-14T12:34:23.909512"
  },
  {
    "epoch": 1413,
    "batch_idx": 192,
    "loss": 0.12900416553020477,
    "learning_rate": 0.00019790901504275601,
    "timestep_mean": 417.625,
    "epsilon_norm": 3.309617757797241,
    "prediction_norm": 2.72072696685791,
    "timestamp": "2025-07-14T12:34:23.914690"
  },
  {
    "epoch": 1414,
    "batch_idx": 0,
    "loss": 0.12237662076950073,
    "learning_rate": 0.0001972835463094996,
    "timestep_mean": 518.46875,
    "epsilon_norm": 6.800380229949951,
    "prediction_norm": 6.498743057250977,
    "timestamp": "2025-07-14T12:34:23.919922"
  },
  {
    "epoch": 1414,
    "batch_idx": 32,
    "loss": 0.43124064803123474,
    "learning_rate": 0.0001972835463094996,
    "timestep_mean": 431.4375,
    "epsilon_norm": 7.700784206390381,
    "prediction_norm": 5.449223518371582,
    "timestamp": "2025-07-14T12:34:23.925230"
  },
  {
    "epoch": 1414,
    "batch_idx": 64,
    "loss": 0.1906755119562149,
    "learning_rate": 0.0001972835463094996,
    "timestep_mean": 533.34375,
    "epsilon_norm": 6.849647045135498,
    "prediction_norm": 5.88530158996582,
    "timestamp": "2025-07-14T12:34:23.930441"
  },
  {
    "epoch": 1414,
    "batch_idx": 96,
    "loss": 0.25112631916999817,
    "learning_rate": 0.0001972835463094996,
    "timestep_mean": 559.5,
    "epsilon_norm": 8.191370964050293,
    "prediction_norm": 7.337681293487549,
    "timestamp": "2025-07-14T12:34:23.935495"
  },
  {
    "epoch": 1414,
    "batch_idx": 128,
    "loss": 0.2439250648021698,
    "learning_rate": 0.0001972835463094996,
    "timestep_mean": 464.0,
    "epsilon_norm": 7.2101545333862305,
    "prediction_norm": 5.92364501953125,
    "timestamp": "2025-07-14T12:34:23.940786"
  },
  {
    "epoch": 1414,
    "batch_idx": 160,
    "loss": 0.5337440967559814,
    "learning_rate": 0.0001972835463094996,
    "timestep_mean": 309.6875,
    "epsilon_norm": 9.22117805480957,
    "prediction_norm": 6.051656723022461,
    "timestamp": "2025-07-14T12:34:23.945919"
  },
  {
    "epoch": 1414,
    "batch_idx": 192,
    "loss": 0.21598190069198608,
    "learning_rate": 0.0001972835463094996,
    "timestep_mean": 458.5,
    "epsilon_norm": 3.7448766231536865,
    "prediction_norm": 3.519747495651245,
    "timestamp": "2025-07-14T12:34:23.951055"
  },
  {
    "epoch": 1415,
    "batch_idx": 0,
    "loss": 0.2265227735042572,
    "learning_rate": 0.00019665882449900049,
    "timestep_mean": 449.4375,
    "epsilon_norm": 7.0838117599487305,
    "prediction_norm": 5.853739261627197,
    "timestamp": "2025-07-14T12:34:23.956209"
  },
  {
    "epoch": 1415,
    "batch_idx": 32,
    "loss": 0.196808859705925,
    "learning_rate": 0.00019665882449900049,
    "timestep_mean": 593.96875,
    "epsilon_norm": 7.552307605743408,
    "prediction_norm": 6.628597736358643,
    "timestamp": "2025-07-14T12:34:23.961358"
  },
  {
    "epoch": 1415,
    "batch_idx": 64,
    "loss": 0.1971207559108734,
    "learning_rate": 0.00019665882449900049,
    "timestep_mean": 475.46875,
    "epsilon_norm": 7.70824670791626,
    "prediction_norm": 6.353775501251221,
    "timestamp": "2025-07-14T12:34:23.966780"
  },
  {
    "epoch": 1415,
    "batch_idx": 96,
    "loss": 0.35610276460647583,
    "learning_rate": 0.00019665882449900049,
    "timestep_mean": 469.3125,
    "epsilon_norm": 7.576453685760498,
    "prediction_norm": 6.087670803070068,
    "timestamp": "2025-07-14T12:34:23.971988"
  },
  {
    "epoch": 1415,
    "batch_idx": 128,
    "loss": 0.26696687936782837,
    "learning_rate": 0.00019665882449900049,
    "timestep_mean": 509.9375,
    "epsilon_norm": 8.05173397064209,
    "prediction_norm": 6.89649772644043,
    "timestamp": "2025-07-14T12:34:23.977164"
  },
  {
    "epoch": 1415,
    "batch_idx": 160,
    "loss": 0.23423394560813904,
    "learning_rate": 0.00019665882449900049,
    "timestep_mean": 423.96875,
    "epsilon_norm": 8.130845069885254,
    "prediction_norm": 7.109274387359619,
    "timestamp": "2025-07-14T12:34:23.982315"
  },
  {
    "epoch": 1415,
    "batch_idx": 192,
    "loss": 0.11320846527814865,
    "learning_rate": 0.00019665882449900049,
    "timestep_mean": 639.25,
    "epsilon_norm": 3.756187915802002,
    "prediction_norm": 3.383030652999878,
    "timestamp": "2025-07-14T12:34:23.987445"
  },
  {
    "epoch": 1416,
    "batch_idx": 0,
    "loss": 0.16001732647418976,
    "learning_rate": 0.00019603485115269752,
    "timestep_mean": 576.21875,
    "epsilon_norm": 8.24850082397461,
    "prediction_norm": 7.696839332580566,
    "timestamp": "2025-07-14T12:34:23.992603"
  },
  {
    "epoch": 1416,
    "batch_idx": 32,
    "loss": 0.24254830181598663,
    "learning_rate": 0.00019603485115269752,
    "timestep_mean": 550.90625,
    "epsilon_norm": 7.392310619354248,
    "prediction_norm": 6.481982707977295,
    "timestamp": "2025-07-14T12:34:23.997806"
  },
  {
    "epoch": 1416,
    "batch_idx": 64,
    "loss": 0.20560936629772186,
    "learning_rate": 0.00019603485115269752,
    "timestep_mean": 543.03125,
    "epsilon_norm": 8.154701232910156,
    "prediction_norm": 7.572175025939941,
    "timestamp": "2025-07-14T12:34:24.003004"
  },
  {
    "epoch": 1416,
    "batch_idx": 96,
    "loss": 0.2846493721008301,
    "learning_rate": 0.00019603485115269752,
    "timestep_mean": 488.65625,
    "epsilon_norm": 8.062566757202148,
    "prediction_norm": 6.337714672088623,
    "timestamp": "2025-07-14T12:34:24.008079"
  },
  {
    "epoch": 1416,
    "batch_idx": 128,
    "loss": 0.6888656616210938,
    "learning_rate": 0.00019603485115269752,
    "timestep_mean": 425.1875,
    "epsilon_norm": 8.976661682128906,
    "prediction_norm": 6.376196384429932,
    "timestamp": "2025-07-14T12:34:24.013229"
  },
  {
    "epoch": 1416,
    "batch_idx": 160,
    "loss": 0.25713270902633667,
    "learning_rate": 0.00019603485115269752,
    "timestep_mean": 533.9375,
    "epsilon_norm": 7.171917915344238,
    "prediction_norm": 6.29251766204834,
    "timestamp": "2025-07-14T12:34:24.018369"
  },
  {
    "epoch": 1416,
    "batch_idx": 192,
    "loss": 0.7589192390441895,
    "learning_rate": 0.00019603485115269752,
    "timestep_mean": 567.625,
    "epsilon_norm": 5.82738733291626,
    "prediction_norm": 4.6917266845703125,
    "timestamp": "2025-07-14T12:34:24.023484"
  },
  {
    "epoch": 1417,
    "batch_idx": 0,
    "loss": 0.42834848165512085,
    "learning_rate": 0.00019541162781018322,
    "timestep_mean": 382.09375,
    "epsilon_norm": 8.505534172058105,
    "prediction_norm": 6.617222785949707,
    "timestamp": "2025-07-14T12:34:24.028679"
  },
  {
    "epoch": 1417,
    "batch_idx": 32,
    "loss": 0.2780020833015442,
    "learning_rate": 0.00019541162781018322,
    "timestep_mean": 454.3125,
    "epsilon_norm": 7.491226673126221,
    "prediction_norm": 6.110232353210449,
    "timestamp": "2025-07-14T12:34:24.034106"
  },
  {
    "epoch": 1417,
    "batch_idx": 64,
    "loss": 0.1612125188112259,
    "learning_rate": 0.00019541162781018322,
    "timestep_mean": 434.875,
    "epsilon_norm": 9.258295059204102,
    "prediction_norm": 8.620676040649414,
    "timestamp": "2025-07-14T12:34:24.039172"
  },
  {
    "epoch": 1417,
    "batch_idx": 96,
    "loss": 0.4709433913230896,
    "learning_rate": 0.00019541162781018322,
    "timestep_mean": 465.96875,
    "epsilon_norm": 7.56224250793457,
    "prediction_norm": 5.872608184814453,
    "timestamp": "2025-07-14T12:34:24.044406"
  },
  {
    "epoch": 1417,
    "batch_idx": 128,
    "loss": 0.5319983959197998,
    "learning_rate": 0.00019541162781018322,
    "timestep_mean": 439.34375,
    "epsilon_norm": 8.206007957458496,
    "prediction_norm": 5.650888442993164,
    "timestamp": "2025-07-14T12:34:24.049498"
  },
  {
    "epoch": 1417,
    "batch_idx": 160,
    "loss": 0.24912357330322266,
    "learning_rate": 0.00019541162781018322,
    "timestep_mean": 440.15625,
    "epsilon_norm": 8.434541702270508,
    "prediction_norm": 7.938688278198242,
    "timestamp": "2025-07-14T12:34:24.054656"
  },
  {
    "epoch": 1417,
    "batch_idx": 192,
    "loss": 0.5735758543014526,
    "learning_rate": 0.00019541162781018322,
    "timestep_mean": 427.25,
    "epsilon_norm": 4.6262006759643555,
    "prediction_norm": 3.8105533123016357,
    "timestamp": "2025-07-14T12:34:24.059911"
  },
  {
    "epoch": 1418,
    "batch_idx": 0,
    "loss": 0.3986680209636688,
    "learning_rate": 0.00019478915600919903,
    "timestep_mean": 476.90625,
    "epsilon_norm": 7.359646797180176,
    "prediction_norm": 5.696626663208008,
    "timestamp": "2025-07-14T12:34:24.065103"
  },
  {
    "epoch": 1418,
    "batch_idx": 32,
    "loss": 0.46515753865242004,
    "learning_rate": 0.00019478915600919903,
    "timestep_mean": 487.09375,
    "epsilon_norm": 8.049793243408203,
    "prediction_norm": 6.151160717010498,
    "timestamp": "2025-07-14T12:34:24.070169"
  },
  {
    "epoch": 1418,
    "batch_idx": 64,
    "loss": 0.2449016571044922,
    "learning_rate": 0.00019478915600919903,
    "timestep_mean": 482.8125,
    "epsilon_norm": 8.232359886169434,
    "prediction_norm": 7.18375301361084,
    "timestamp": "2025-07-14T12:34:24.075287"
  },
  {
    "epoch": 1418,
    "batch_idx": 96,
    "loss": 0.12865981459617615,
    "learning_rate": 0.00019478915600919903,
    "timestep_mean": 533.84375,
    "epsilon_norm": 7.617547035217285,
    "prediction_norm": 7.0301384925842285,
    "timestamp": "2025-07-14T12:34:24.080394"
  },
  {
    "epoch": 1418,
    "batch_idx": 128,
    "loss": 0.18376599252223969,
    "learning_rate": 0.00019478915600919903,
    "timestep_mean": 581.34375,
    "epsilon_norm": 7.763378143310547,
    "prediction_norm": 7.251241683959961,
    "timestamp": "2025-07-14T12:34:24.085480"
  },
  {
    "epoch": 1418,
    "batch_idx": 160,
    "loss": 0.16738227009773254,
    "learning_rate": 0.00019478915600919903,
    "timestep_mean": 499.28125,
    "epsilon_norm": 7.703174591064453,
    "prediction_norm": 7.045634746551514,
    "timestamp": "2025-07-14T12:34:24.090748"
  },
  {
    "epoch": 1418,
    "batch_idx": 192,
    "loss": 0.3352179229259491,
    "learning_rate": 0.00019478915600919903,
    "timestep_mean": 412.0,
    "epsilon_norm": 4.365116596221924,
    "prediction_norm": 4.035494327545166,
    "timestamp": "2025-07-14T12:34:24.095899"
  },
  {
    "epoch": 1419,
    "batch_idx": 0,
    "loss": 0.23500479757785797,
    "learning_rate": 0.0001941674372856321,
    "timestep_mean": 449.6875,
    "epsilon_norm": 8.072787284851074,
    "prediction_norm": 7.214293956756592,
    "timestamp": "2025-07-14T12:34:24.101033"
  },
  {
    "epoch": 1419,
    "batch_idx": 32,
    "loss": 0.30995458364486694,
    "learning_rate": 0.0001941674372856321,
    "timestep_mean": 430.4375,
    "epsilon_norm": 8.69508171081543,
    "prediction_norm": 7.483078479766846,
    "timestamp": "2025-07-14T12:34:24.106295"
  },
  {
    "epoch": 1419,
    "batch_idx": 64,
    "loss": 0.32617777585983276,
    "learning_rate": 0.0001941674372856321,
    "timestep_mean": 525.34375,
    "epsilon_norm": 8.900985717773438,
    "prediction_norm": 6.988035678863525,
    "timestamp": "2025-07-14T12:34:24.111490"
  },
  {
    "epoch": 1419,
    "batch_idx": 96,
    "loss": 0.22748811542987823,
    "learning_rate": 0.0001941674372856321,
    "timestep_mean": 511.59375,
    "epsilon_norm": 7.659580230712891,
    "prediction_norm": 6.303641319274902,
    "timestamp": "2025-07-14T12:34:24.116785"
  },
  {
    "epoch": 1419,
    "batch_idx": 128,
    "loss": 0.17623993754386902,
    "learning_rate": 0.0001941674372856321,
    "timestep_mean": 539.78125,
    "epsilon_norm": 7.116482257843018,
    "prediction_norm": 5.938924789428711,
    "timestamp": "2025-07-14T12:34:24.121989"
  },
  {
    "epoch": 1419,
    "batch_idx": 160,
    "loss": 0.44367533922195435,
    "learning_rate": 0.0001941674372856321,
    "timestep_mean": 456.40625,
    "epsilon_norm": 7.312570095062256,
    "prediction_norm": 5.677903652191162,
    "timestamp": "2025-07-14T12:34:24.127158"
  },
  {
    "epoch": 1419,
    "batch_idx": 192,
    "loss": 0.11949079483747482,
    "learning_rate": 0.0001941674372856321,
    "timestep_mean": 570.875,
    "epsilon_norm": 3.660933256149292,
    "prediction_norm": 3.52193546295166,
    "timestamp": "2025-07-14T12:34:24.132219"
  },
  {
    "epoch": 1420,
    "batch_idx": 0,
    "loss": 0.4067274332046509,
    "learning_rate": 0.00019354647317351212,
    "timestep_mean": 497.84375,
    "epsilon_norm": 8.462886810302734,
    "prediction_norm": 6.157504081726074,
    "timestamp": "2025-07-14T12:34:24.137406"
  },
  {
    "epoch": 1420,
    "batch_idx": 32,
    "loss": 0.1688777506351471,
    "learning_rate": 0.00019354647317351212,
    "timestep_mean": 454.28125,
    "epsilon_norm": 7.9837517738342285,
    "prediction_norm": 7.201552867889404,
    "timestamp": "2025-07-14T12:34:24.142637"
  },
  {
    "epoch": 1420,
    "batch_idx": 64,
    "loss": 0.23904143273830414,
    "learning_rate": 0.00019354647317351212,
    "timestep_mean": 489.4375,
    "epsilon_norm": 8.116272926330566,
    "prediction_norm": 6.988423824310303,
    "timestamp": "2025-07-14T12:34:24.147781"
  },
  {
    "epoch": 1420,
    "batch_idx": 96,
    "loss": 0.43530982732772827,
    "learning_rate": 0.00019354647317351212,
    "timestep_mean": 450.625,
    "epsilon_norm": 8.743927955627441,
    "prediction_norm": 6.369452953338623,
    "timestamp": "2025-07-14T12:34:24.153065"
  },
  {
    "epoch": 1420,
    "batch_idx": 128,
    "loss": 0.26658734679222107,
    "learning_rate": 0.00019354647317351212,
    "timestep_mean": 561.96875,
    "epsilon_norm": 8.292943000793457,
    "prediction_norm": 6.71231746673584,
    "timestamp": "2025-07-14T12:34:24.158120"
  },
  {
    "epoch": 1420,
    "batch_idx": 160,
    "loss": 0.17495840787887573,
    "learning_rate": 0.00019354647317351212,
    "timestep_mean": 582.15625,
    "epsilon_norm": 7.672940731048584,
    "prediction_norm": 6.665618896484375,
    "timestamp": "2025-07-14T12:34:24.163184"
  },
  {
    "epoch": 1420,
    "batch_idx": 192,
    "loss": 0.21884137392044067,
    "learning_rate": 0.00019354647317351212,
    "timestep_mean": 396.375,
    "epsilon_norm": 3.7799887657165527,
    "prediction_norm": 3.590143918991089,
    "timestamp": "2025-07-14T12:34:24.168323"
  },
  {
    "epoch": 1421,
    "batch_idx": 0,
    "loss": 0.24935929477214813,
    "learning_rate": 0.00019292626520500558,
    "timestep_mean": 480.1875,
    "epsilon_norm": 8.190113067626953,
    "prediction_norm": 7.462032318115234,
    "timestamp": "2025-07-14T12:34:24.173561"
  },
  {
    "epoch": 1421,
    "batch_idx": 32,
    "loss": 0.4592670202255249,
    "learning_rate": 0.00019292626520500558,
    "timestep_mean": 465.25,
    "epsilon_norm": 9.436216354370117,
    "prediction_norm": 6.854977130889893,
    "timestamp": "2025-07-14T12:34:24.178658"
  },
  {
    "epoch": 1421,
    "batch_idx": 64,
    "loss": 0.2037079930305481,
    "learning_rate": 0.00019292626520500558,
    "timestep_mean": 553.71875,
    "epsilon_norm": 7.721737861633301,
    "prediction_norm": 6.95495080947876,
    "timestamp": "2025-07-14T12:34:24.183936"
  },
  {
    "epoch": 1421,
    "batch_idx": 96,
    "loss": 0.35160326957702637,
    "learning_rate": 0.00019292626520500558,
    "timestep_mean": 509.53125,
    "epsilon_norm": 7.735157012939453,
    "prediction_norm": 5.727904796600342,
    "timestamp": "2025-07-14T12:34:24.189122"
  },
  {
    "epoch": 1421,
    "batch_idx": 128,
    "loss": 0.22232772409915924,
    "learning_rate": 0.00019292626520500558,
    "timestep_mean": 508.09375,
    "epsilon_norm": 7.219882488250732,
    "prediction_norm": 6.649508953094482,
    "timestamp": "2025-07-14T12:34:24.194246"
  },
  {
    "epoch": 1421,
    "batch_idx": 160,
    "loss": 0.17038795351982117,
    "learning_rate": 0.00019292626520500558,
    "timestep_mean": 491.40625,
    "epsilon_norm": 6.240375995635986,
    "prediction_norm": 5.205631256103516,
    "timestamp": "2025-07-14T12:34:24.199441"
  },
  {
    "epoch": 1421,
    "batch_idx": 192,
    "loss": 0.07085151225328445,
    "learning_rate": 0.00019292626520500558,
    "timestep_mean": 695.375,
    "epsilon_norm": 3.658188581466675,
    "prediction_norm": 3.56608247756958,
    "timestamp": "2025-07-14T12:34:24.204696"
  },
  {
    "epoch": 1422,
    "batch_idx": 0,
    "loss": 0.26839369535446167,
    "learning_rate": 0.00019230681491041473,
    "timestep_mean": 553.65625,
    "epsilon_norm": 8.36231803894043,
    "prediction_norm": 6.925311088562012,
    "timestamp": "2025-07-14T12:34:24.209805"
  },
  {
    "epoch": 1422,
    "batch_idx": 32,
    "loss": 0.29570630192756653,
    "learning_rate": 0.00019230681491041473,
    "timestep_mean": 432.46875,
    "epsilon_norm": 7.9139723777771,
    "prediction_norm": 6.9148783683776855,
    "timestamp": "2025-07-14T12:34:24.215046"
  },
  {
    "epoch": 1422,
    "batch_idx": 64,
    "loss": 0.17119792103767395,
    "learning_rate": 0.00019230681491041473,
    "timestep_mean": 521.59375,
    "epsilon_norm": 7.344344615936279,
    "prediction_norm": 6.173393249511719,
    "timestamp": "2025-07-14T12:34:24.220237"
  },
  {
    "epoch": 1422,
    "batch_idx": 96,
    "loss": 0.14630979299545288,
    "learning_rate": 0.00019230681491041473,
    "timestep_mean": 592.90625,
    "epsilon_norm": 8.093670845031738,
    "prediction_norm": 7.899786472320557,
    "timestamp": "2025-07-14T12:34:24.225565"
  },
  {
    "epoch": 1422,
    "batch_idx": 128,
    "loss": 0.23070281744003296,
    "learning_rate": 0.00019230681491041473,
    "timestep_mean": 505.3125,
    "epsilon_norm": 8.509833335876465,
    "prediction_norm": 7.620431423187256,
    "timestamp": "2025-07-14T12:34:24.230787"
  },
  {
    "epoch": 1422,
    "batch_idx": 160,
    "loss": 0.3811376094818115,
    "learning_rate": 0.00019230681491041473,
    "timestep_mean": 491.625,
    "epsilon_norm": 8.154638290405273,
    "prediction_norm": 6.01522970199585,
    "timestamp": "2025-07-14T12:34:24.235955"
  },
  {
    "epoch": 1422,
    "batch_idx": 192,
    "loss": 0.17471584677696228,
    "learning_rate": 0.00019230681491041473,
    "timestep_mean": 368.0,
    "epsilon_norm": 3.942160129547119,
    "prediction_norm": 3.1824939250946045,
    "timestamp": "2025-07-14T12:34:24.240888"
  },
  {
    "epoch": 1423,
    "batch_idx": 0,
    "loss": 0.36015745997428894,
    "learning_rate": 0.0001916881238181708,
    "timestep_mean": 447.90625,
    "epsilon_norm": 7.980620861053467,
    "prediction_norm": 6.624266147613525,
    "timestamp": "2025-07-14T12:34:24.245776"
  },
  {
    "epoch": 1423,
    "batch_idx": 32,
    "loss": 0.4316186308860779,
    "learning_rate": 0.0001916881238181708,
    "timestep_mean": 482.625,
    "epsilon_norm": 8.819644927978516,
    "prediction_norm": 6.866671562194824,
    "timestamp": "2025-07-14T12:34:24.250704"
  },
  {
    "epoch": 1423,
    "batch_idx": 64,
    "loss": 0.18565846979618073,
    "learning_rate": 0.0001916881238181708,
    "timestep_mean": 526.9375,
    "epsilon_norm": 7.256767749786377,
    "prediction_norm": 6.228835105895996,
    "timestamp": "2025-07-14T12:34:24.255590"
  },
  {
    "epoch": 1423,
    "batch_idx": 96,
    "loss": 0.37452423572540283,
    "learning_rate": 0.0001916881238181708,
    "timestep_mean": 443.71875,
    "epsilon_norm": 9.222509384155273,
    "prediction_norm": 7.173813343048096,
    "timestamp": "2025-07-14T12:34:24.260481"
  },
  {
    "epoch": 1423,
    "batch_idx": 128,
    "loss": 0.199586421251297,
    "learning_rate": 0.0001916881238181708,
    "timestep_mean": 565.4375,
    "epsilon_norm": 8.15068244934082,
    "prediction_norm": 6.7180657386779785,
    "timestamp": "2025-07-14T12:34:24.265361"
  },
  {
    "epoch": 1423,
    "batch_idx": 160,
    "loss": 0.14681684970855713,
    "learning_rate": 0.0001916881238181708,
    "timestep_mean": 498.84375,
    "epsilon_norm": 7.239517688751221,
    "prediction_norm": 6.4697957038879395,
    "timestamp": "2025-07-14T12:34:24.270223"
  },
  {
    "epoch": 1423,
    "batch_idx": 192,
    "loss": 0.05527954176068306,
    "learning_rate": 0.0001916881238181708,
    "timestep_mean": 588.25,
    "epsilon_norm": 4.637238502502441,
    "prediction_norm": 4.388296604156494,
    "timestamp": "2025-07-14T12:34:24.275241"
  },
  {
    "epoch": 1424,
    "batch_idx": 0,
    "loss": 0.20184051990509033,
    "learning_rate": 0.00019107019345483302,
    "timestep_mean": 551.65625,
    "epsilon_norm": 8.635627746582031,
    "prediction_norm": 7.602962970733643,
    "timestamp": "2025-07-14T12:34:24.280146"
  },
  {
    "epoch": 1424,
    "batch_idx": 32,
    "loss": 0.3431926369667053,
    "learning_rate": 0.00019107019345483302,
    "timestep_mean": 473.375,
    "epsilon_norm": 7.675092697143555,
    "prediction_norm": 6.164343357086182,
    "timestamp": "2025-07-14T12:34:24.285082"
  },
  {
    "epoch": 1424,
    "batch_idx": 64,
    "loss": 0.23070135712623596,
    "learning_rate": 0.00019107019345483302,
    "timestep_mean": 469.75,
    "epsilon_norm": 6.826258182525635,
    "prediction_norm": 6.0384297370910645,
    "timestamp": "2025-07-14T12:34:24.290018"
  },
  {
    "epoch": 1424,
    "batch_idx": 96,
    "loss": 0.3738402724266052,
    "learning_rate": 0.00019107019345483302,
    "timestep_mean": 481.375,
    "epsilon_norm": 9.26201343536377,
    "prediction_norm": 7.369844913482666,
    "timestamp": "2025-07-14T12:34:24.295003"
  },
  {
    "epoch": 1424,
    "batch_idx": 128,
    "loss": 0.09325219690799713,
    "learning_rate": 0.00019107019345483302,
    "timestep_mean": 546.34375,
    "epsilon_norm": 8.013482093811035,
    "prediction_norm": 7.653589248657227,
    "timestamp": "2025-07-14T12:34:24.300319"
  },
  {
    "epoch": 1424,
    "batch_idx": 160,
    "loss": 0.3392406702041626,
    "learning_rate": 0.00019107019345483302,
    "timestep_mean": 467.90625,
    "epsilon_norm": 7.391044616699219,
    "prediction_norm": 5.805500030517578,
    "timestamp": "2025-07-14T12:34:24.305483"
  },
  {
    "epoch": 1424,
    "batch_idx": 192,
    "loss": 0.05768810212612152,
    "learning_rate": 0.00019107019345483302,
    "timestep_mean": 663.5,
    "epsilon_norm": 5.406254768371582,
    "prediction_norm": 5.027266025543213,
    "timestamp": "2025-07-14T12:34:24.310609"
  },
  {
    "epoch": 1425,
    "batch_idx": 0,
    "loss": 0.3938937783241272,
    "learning_rate": 0.00019045302534508341,
    "timestep_mean": 505.875,
    "epsilon_norm": 9.2874755859375,
    "prediction_norm": 7.369711399078369,
    "timestamp": "2025-07-14T12:34:24.315841"
  },
  {
    "epoch": 1425,
    "batch_idx": 32,
    "loss": 0.26008498668670654,
    "learning_rate": 0.00019045302534508341,
    "timestep_mean": 491.25,
    "epsilon_norm": 7.779517650604248,
    "prediction_norm": 6.400146484375,
    "timestamp": "2025-07-14T12:34:24.320936"
  },
  {
    "epoch": 1425,
    "batch_idx": 64,
    "loss": 0.3071914613246918,
    "learning_rate": 0.00019045302534508341,
    "timestep_mean": 437.375,
    "epsilon_norm": 7.842608451843262,
    "prediction_norm": 6.274656772613525,
    "timestamp": "2025-07-14T12:34:24.325977"
  },
  {
    "epoch": 1425,
    "batch_idx": 96,
    "loss": 0.24154706299304962,
    "learning_rate": 0.00019045302534508341,
    "timestep_mean": 481.71875,
    "epsilon_norm": 8.758673667907715,
    "prediction_norm": 7.522193431854248,
    "timestamp": "2025-07-14T12:34:24.331108"
  },
  {
    "epoch": 1425,
    "batch_idx": 128,
    "loss": 0.22361476719379425,
    "learning_rate": 0.00019045302534508341,
    "timestep_mean": 565.84375,
    "epsilon_norm": 8.08672046661377,
    "prediction_norm": 7.331048488616943,
    "timestamp": "2025-07-14T12:34:24.336142"
  },
  {
    "epoch": 1425,
    "batch_idx": 160,
    "loss": 0.21292313933372498,
    "learning_rate": 0.00019045302534508341,
    "timestep_mean": 519.9375,
    "epsilon_norm": 8.196961402893066,
    "prediction_norm": 7.417400360107422,
    "timestamp": "2025-07-14T12:34:24.341328"
  },
  {
    "epoch": 1425,
    "batch_idx": 192,
    "loss": 0.09873466193675995,
    "learning_rate": 0.00019045302534508341,
    "timestep_mean": 494.0,
    "epsilon_norm": 3.1639909744262695,
    "prediction_norm": 3.2667853832244873,
    "timestamp": "2025-07-14T12:34:24.346646"
  },
  {
    "epoch": 1426,
    "batch_idx": 0,
    "loss": 0.2418990582227707,
    "learning_rate": 0.00018983662101172222,
    "timestep_mean": 522.46875,
    "epsilon_norm": 8.349709510803223,
    "prediction_norm": 7.193845272064209,
    "timestamp": "2025-07-14T12:34:24.351817"
  },
  {
    "epoch": 1426,
    "batch_idx": 32,
    "loss": 0.2417675405740738,
    "learning_rate": 0.00018983662101172222,
    "timestep_mean": 506.125,
    "epsilon_norm": 6.430114269256592,
    "prediction_norm": 4.703917980194092,
    "timestamp": "2025-07-14T12:34:24.356956"
  },
  {
    "epoch": 1426,
    "batch_idx": 64,
    "loss": 0.2590841054916382,
    "learning_rate": 0.00018983662101172222,
    "timestep_mean": 457.59375,
    "epsilon_norm": 6.9423604011535645,
    "prediction_norm": 5.7539448738098145,
    "timestamp": "2025-07-14T12:34:24.361971"
  },
  {
    "epoch": 1426,
    "batch_idx": 96,
    "loss": 0.4688747227191925,
    "learning_rate": 0.00018983662101172222,
    "timestep_mean": 500.75,
    "epsilon_norm": 7.711598873138428,
    "prediction_norm": 5.745298385620117,
    "timestamp": "2025-07-14T12:34:24.366936"
  },
  {
    "epoch": 1426,
    "batch_idx": 128,
    "loss": 0.16118484735488892,
    "learning_rate": 0.00018983662101172222,
    "timestep_mean": 518.6875,
    "epsilon_norm": 7.5688676834106445,
    "prediction_norm": 6.991633415222168,
    "timestamp": "2025-07-14T12:34:24.371721"
  },
  {
    "epoch": 1426,
    "batch_idx": 160,
    "loss": 0.17214922606945038,
    "learning_rate": 0.00018983662101172222,
    "timestep_mean": 452.625,
    "epsilon_norm": 7.704574108123779,
    "prediction_norm": 7.121267795562744,
    "timestamp": "2025-07-14T12:34:24.376658"
  },
  {
    "epoch": 1426,
    "batch_idx": 192,
    "loss": 0.5576065182685852,
    "learning_rate": 0.00018983662101172222,
    "timestep_mean": 572.25,
    "epsilon_norm": 4.5305495262146,
    "prediction_norm": 2.9972774982452393,
    "timestamp": "2025-07-14T12:34:24.381517"
  },
  {
    "epoch": 1427,
    "batch_idx": 0,
    "loss": 0.262772798538208,
    "learning_rate": 0.0001892209819756665,
    "timestep_mean": 422.40625,
    "epsilon_norm": 6.945295333862305,
    "prediction_norm": 5.7259111404418945,
    "timestamp": "2025-07-14T12:34:24.386428"
  },
  {
    "epoch": 1427,
    "batch_idx": 32,
    "loss": 0.27979031205177307,
    "learning_rate": 0.0001892209819756665,
    "timestep_mean": 590.90625,
    "epsilon_norm": 8.423625946044922,
    "prediction_norm": 7.3549370765686035,
    "timestamp": "2025-07-14T12:34:24.391295"
  },
  {
    "epoch": 1427,
    "batch_idx": 64,
    "loss": 0.20906206965446472,
    "learning_rate": 0.0001892209819756665,
    "timestep_mean": 619.96875,
    "epsilon_norm": 8.887434005737305,
    "prediction_norm": 7.814574241638184,
    "timestamp": "2025-07-14T12:34:24.396232"
  },
  {
    "epoch": 1427,
    "batch_idx": 96,
    "loss": 0.35982075333595276,
    "learning_rate": 0.0001892209819756665,
    "timestep_mean": 541.46875,
    "epsilon_norm": 7.9889302253723145,
    "prediction_norm": 7.087898254394531,
    "timestamp": "2025-07-14T12:34:24.401131"
  },
  {
    "epoch": 1427,
    "batch_idx": 128,
    "loss": 0.2409130334854126,
    "learning_rate": 0.0001892209819756665,
    "timestep_mean": 568.71875,
    "epsilon_norm": 7.314550876617432,
    "prediction_norm": 6.616686820983887,
    "timestamp": "2025-07-14T12:34:24.406026"
  },
  {
    "epoch": 1427,
    "batch_idx": 160,
    "loss": 0.27387091517448425,
    "learning_rate": 0.0001892209819756665,
    "timestep_mean": 493.0625,
    "epsilon_norm": 6.929989337921143,
    "prediction_norm": 5.6918721199035645,
    "timestamp": "2025-07-14T12:34:24.410863"
  },
  {
    "epoch": 1427,
    "batch_idx": 192,
    "loss": 0.048357002437114716,
    "learning_rate": 0.0001892209819756665,
    "timestep_mean": 675.625,
    "epsilon_norm": 3.4403178691864014,
    "prediction_norm": 3.3050429821014404,
    "timestamp": "2025-07-14T12:34:24.415703"
  },
  {
    "epoch": 1428,
    "batch_idx": 0,
    "loss": 0.48456287384033203,
    "learning_rate": 0.00018860610975594412,
    "timestep_mean": 473.375,
    "epsilon_norm": 7.938223838806152,
    "prediction_norm": 5.95481014251709,
    "timestamp": "2025-07-14T12:34:24.420836"
  },
  {
    "epoch": 1428,
    "batch_idx": 32,
    "loss": 0.1574600636959076,
    "learning_rate": 0.00018860610975594412,
    "timestep_mean": 540.5625,
    "epsilon_norm": 7.2515974044799805,
    "prediction_norm": 6.688679218292236,
    "timestamp": "2025-07-14T12:34:24.425805"
  },
  {
    "epoch": 1428,
    "batch_idx": 64,
    "loss": 0.2564346194267273,
    "learning_rate": 0.00018860610975594412,
    "timestep_mean": 448.90625,
    "epsilon_norm": 7.773104667663574,
    "prediction_norm": 6.194318771362305,
    "timestamp": "2025-07-14T12:34:24.430696"
  },
  {
    "epoch": 1428,
    "batch_idx": 96,
    "loss": 0.19792407751083374,
    "learning_rate": 0.00018860610975594412,
    "timestep_mean": 445.875,
    "epsilon_norm": 7.370016098022461,
    "prediction_norm": 6.327486515045166,
    "timestamp": "2025-07-14T12:34:24.435557"
  },
  {
    "epoch": 1428,
    "batch_idx": 128,
    "loss": 0.3312510848045349,
    "learning_rate": 0.00018860610975594412,
    "timestep_mean": 525.9375,
    "epsilon_norm": 8.708207130432129,
    "prediction_norm": 7.420932769775391,
    "timestamp": "2025-07-14T12:34:24.440418"
  },
  {
    "epoch": 1428,
    "batch_idx": 160,
    "loss": 0.2812204957008362,
    "learning_rate": 0.00018860610975594412,
    "timestep_mean": 466.6875,
    "epsilon_norm": 11.024894714355469,
    "prediction_norm": 9.847768783569336,
    "timestamp": "2025-07-14T12:34:24.445502"
  },
  {
    "epoch": 1428,
    "batch_idx": 192,
    "loss": 0.10114125162363052,
    "learning_rate": 0.00018860610975594412,
    "timestep_mean": 513.375,
    "epsilon_norm": 4.380766868591309,
    "prediction_norm": 3.9311327934265137,
    "timestamp": "2025-07-14T12:34:24.450519"
  },
  {
    "epoch": 1429,
    "batch_idx": 0,
    "loss": 0.2678118348121643,
    "learning_rate": 0.00018799200586969096,
    "timestep_mean": 519.625,
    "epsilon_norm": 7.270426273345947,
    "prediction_norm": 6.129528045654297,
    "timestamp": "2025-07-14T12:34:24.455436"
  },
  {
    "epoch": 1429,
    "batch_idx": 32,
    "loss": 0.37070518732070923,
    "learning_rate": 0.00018799200586969096,
    "timestep_mean": 425.625,
    "epsilon_norm": 8.689543724060059,
    "prediction_norm": 7.355753421783447,
    "timestamp": "2025-07-14T12:34:24.460272"
  },
  {
    "epoch": 1429,
    "batch_idx": 64,
    "loss": 0.24523010849952698,
    "learning_rate": 0.00018799200586969096,
    "timestep_mean": 503.1875,
    "epsilon_norm": 7.880980968475342,
    "prediction_norm": 6.637599945068359,
    "timestamp": "2025-07-14T12:34:24.465107"
  },
  {
    "epoch": 1429,
    "batch_idx": 96,
    "loss": 0.25708702206611633,
    "learning_rate": 0.00018799200586969096,
    "timestep_mean": 488.34375,
    "epsilon_norm": 7.991409778594971,
    "prediction_norm": 6.387197971343994,
    "timestamp": "2025-07-14T12:34:24.470082"
  },
  {
    "epoch": 1429,
    "batch_idx": 128,
    "loss": 0.4444252848625183,
    "learning_rate": 0.00018799200586969096,
    "timestep_mean": 464.84375,
    "epsilon_norm": 8.312468528747559,
    "prediction_norm": 6.39815616607666,
    "timestamp": "2025-07-14T12:34:24.475082"
  },
  {
    "epoch": 1429,
    "batch_idx": 160,
    "loss": 0.22988161444664001,
    "learning_rate": 0.00018799200586969096,
    "timestep_mean": 563.4375,
    "epsilon_norm": 8.745992660522461,
    "prediction_norm": 7.842879772186279,
    "timestamp": "2025-07-14T12:34:24.479938"
  },
  {
    "epoch": 1429,
    "batch_idx": 192,
    "loss": 0.43792659044265747,
    "learning_rate": 0.00018799200586969096,
    "timestep_mean": 421.5,
    "epsilon_norm": 4.262162208557129,
    "prediction_norm": 3.4580187797546387,
    "timestamp": "2025-07-14T12:34:24.484736"
  },
  {
    "epoch": 1430,
    "batch_idx": 0,
    "loss": 0.1869937926530838,
    "learning_rate": 0.00018737867183214765,
    "timestep_mean": 497.65625,
    "epsilon_norm": 6.498116970062256,
    "prediction_norm": 5.868168354034424,
    "timestamp": "2025-07-14T12:34:24.489624"
  },
  {
    "epoch": 1430,
    "batch_idx": 32,
    "loss": 0.25753647089004517,
    "learning_rate": 0.00018737867183214765,
    "timestep_mean": 504.1875,
    "epsilon_norm": 7.299435615539551,
    "prediction_norm": 6.1454315185546875,
    "timestamp": "2025-07-14T12:34:24.494453"
  },
  {
    "epoch": 1430,
    "batch_idx": 64,
    "loss": 0.12107907980680466,
    "learning_rate": 0.00018737867183214765,
    "timestep_mean": 494.1875,
    "epsilon_norm": 6.972869396209717,
    "prediction_norm": 6.102388858795166,
    "timestamp": "2025-07-14T12:34:24.499318"
  },
  {
    "epoch": 1430,
    "batch_idx": 96,
    "loss": 0.21172569692134857,
    "learning_rate": 0.00018737867183214765,
    "timestep_mean": 470.125,
    "epsilon_norm": 7.486837863922119,
    "prediction_norm": 6.003885269165039,
    "timestamp": "2025-07-14T12:34:24.504219"
  },
  {
    "epoch": 1430,
    "batch_idx": 128,
    "loss": 0.22078050673007965,
    "learning_rate": 0.00018737867183214765,
    "timestep_mean": 475.65625,
    "epsilon_norm": 7.153716564178467,
    "prediction_norm": 6.398906230926514,
    "timestamp": "2025-07-14T12:34:24.509198"
  },
  {
    "epoch": 1430,
    "batch_idx": 160,
    "loss": 0.22581064701080322,
    "learning_rate": 0.00018737867183214765,
    "timestep_mean": 336.6875,
    "epsilon_norm": 6.7046637535095215,
    "prediction_norm": 6.0416646003723145,
    "timestamp": "2025-07-14T12:34:24.514031"
  },
  {
    "epoch": 1430,
    "batch_idx": 192,
    "loss": 0.15693186223506927,
    "learning_rate": 0.00018737867183214765,
    "timestep_mean": 467.75,
    "epsilon_norm": 4.449438571929932,
    "prediction_norm": 4.093638896942139,
    "timestamp": "2025-07-14T12:34:24.519264"
  },
  {
    "epoch": 1431,
    "batch_idx": 0,
    "loss": 0.28379374742507935,
    "learning_rate": 0.0001867661091566549,
    "timestep_mean": 524.96875,
    "epsilon_norm": 9.368875503540039,
    "prediction_norm": 7.800169944763184,
    "timestamp": "2025-07-14T12:34:24.524123"
  },
  {
    "epoch": 1431,
    "batch_idx": 32,
    "loss": 0.17620524764060974,
    "learning_rate": 0.0001867661091566549,
    "timestep_mean": 477.0,
    "epsilon_norm": 7.437414646148682,
    "prediction_norm": 7.030901908874512,
    "timestamp": "2025-07-14T12:34:24.529046"
  },
  {
    "epoch": 1431,
    "batch_idx": 64,
    "loss": 0.18168511986732483,
    "learning_rate": 0.0001867661091566549,
    "timestep_mean": 547.5625,
    "epsilon_norm": 7.912121772766113,
    "prediction_norm": 7.260282516479492,
    "timestamp": "2025-07-14T12:34:24.534189"
  },
  {
    "epoch": 1431,
    "batch_idx": 96,
    "loss": 0.3517373502254486,
    "learning_rate": 0.0001867661091566549,
    "timestep_mean": 483.46875,
    "epsilon_norm": 8.423775672912598,
    "prediction_norm": 6.378019332885742,
    "timestamp": "2025-07-14T12:34:24.539040"
  },
  {
    "epoch": 1431,
    "batch_idx": 128,
    "loss": 0.2888813614845276,
    "learning_rate": 0.0001867661091566549,
    "timestep_mean": 461.46875,
    "epsilon_norm": 8.61186408996582,
    "prediction_norm": 6.712178707122803,
    "timestamp": "2025-07-14T12:34:24.543883"
  },
  {
    "epoch": 1431,
    "batch_idx": 160,
    "loss": 0.3830166459083557,
    "learning_rate": 0.0001867661091566549,
    "timestep_mean": 441.34375,
    "epsilon_norm": 8.095240592956543,
    "prediction_norm": 6.253894329071045,
    "timestamp": "2025-07-14T12:34:24.548968"
  },
  {
    "epoch": 1431,
    "batch_idx": 192,
    "loss": 0.06320834904909134,
    "learning_rate": 0.0001867661091566549,
    "timestep_mean": 691.625,
    "epsilon_norm": 3.4075911045074463,
    "prediction_norm": 3.3539226055145264,
    "timestamp": "2025-07-14T12:34:24.553781"
  },
  {
    "epoch": 1432,
    "batch_idx": 0,
    "loss": 0.2019139528274536,
    "learning_rate": 0.00018615431935465007,
    "timestep_mean": 483.59375,
    "epsilon_norm": 7.132203102111816,
    "prediction_norm": 5.81193733215332,
    "timestamp": "2025-07-14T12:34:24.558732"
  },
  {
    "epoch": 1432,
    "batch_idx": 32,
    "loss": 0.2611658573150635,
    "learning_rate": 0.00018615431935465007,
    "timestep_mean": 463.03125,
    "epsilon_norm": 8.249526023864746,
    "prediction_norm": 6.903152942657471,
    "timestamp": "2025-07-14T12:34:24.563587"
  },
  {
    "epoch": 1432,
    "batch_idx": 64,
    "loss": 0.20380280911922455,
    "learning_rate": 0.00018615431935465007,
    "timestep_mean": 560.03125,
    "epsilon_norm": 6.527854919433594,
    "prediction_norm": 5.419606685638428,
    "timestamp": "2025-07-14T12:34:24.568528"
  },
  {
    "epoch": 1432,
    "batch_idx": 96,
    "loss": 0.5437337160110474,
    "learning_rate": 0.00018615431935465007,
    "timestep_mean": 504.5,
    "epsilon_norm": 9.337146759033203,
    "prediction_norm": 8.303922653198242,
    "timestamp": "2025-07-14T12:34:24.573384"
  },
  {
    "epoch": 1432,
    "batch_idx": 128,
    "loss": 0.2894538938999176,
    "learning_rate": 0.00018615431935465007,
    "timestep_mean": 451.0,
    "epsilon_norm": 7.8799729347229,
    "prediction_norm": 7.075071334838867,
    "timestamp": "2025-07-14T12:34:24.578397"
  },
  {
    "epoch": 1432,
    "batch_idx": 160,
    "loss": 0.19530382752418518,
    "learning_rate": 0.00018615431935465007,
    "timestep_mean": 522.96875,
    "epsilon_norm": 8.550443649291992,
    "prediction_norm": 8.337906837463379,
    "timestamp": "2025-07-14T12:34:24.583392"
  },
  {
    "epoch": 1432,
    "batch_idx": 192,
    "loss": 0.0900309830904007,
    "learning_rate": 0.00018615431935465007,
    "timestep_mean": 652.875,
    "epsilon_norm": 4.3786234855651855,
    "prediction_norm": 3.814540147781372,
    "timestamp": "2025-07-14T12:34:24.588346"
  },
  {
    "epoch": 1433,
    "batch_idx": 0,
    "loss": 0.27417847514152527,
    "learning_rate": 0.00018554330393566384,
    "timestep_mean": 444.125,
    "epsilon_norm": 8.079756736755371,
    "prediction_norm": 6.334732532501221,
    "timestamp": "2025-07-14T12:34:24.593188"
  },
  {
    "epoch": 1433,
    "batch_idx": 32,
    "loss": 0.13450540602207184,
    "learning_rate": 0.00018554330393566384,
    "timestep_mean": 514.9375,
    "epsilon_norm": 7.398966312408447,
    "prediction_norm": 6.727758407592773,
    "timestamp": "2025-07-14T12:34:24.598113"
  },
  {
    "epoch": 1433,
    "batch_idx": 64,
    "loss": 0.14588101208209991,
    "learning_rate": 0.00018554330393566384,
    "timestep_mean": 544.71875,
    "epsilon_norm": 7.375880241394043,
    "prediction_norm": 7.360760688781738,
    "timestamp": "2025-07-14T12:34:24.603048"
  },
  {
    "epoch": 1433,
    "batch_idx": 96,
    "loss": 0.20503655076026917,
    "learning_rate": 0.00018554330393566384,
    "timestep_mean": 554.53125,
    "epsilon_norm": 8.767208099365234,
    "prediction_norm": 8.115346908569336,
    "timestamp": "2025-07-14T12:34:24.608090"
  },
  {
    "epoch": 1433,
    "batch_idx": 128,
    "loss": 0.23222683370113373,
    "learning_rate": 0.00018554330393566384,
    "timestep_mean": 515.125,
    "epsilon_norm": 7.102356433868408,
    "prediction_norm": 6.3200507164001465,
    "timestamp": "2025-07-14T12:34:24.613150"
  },
  {
    "epoch": 1433,
    "batch_idx": 160,
    "loss": 0.316591739654541,
    "learning_rate": 0.00018554330393566384,
    "timestep_mean": 518.71875,
    "epsilon_norm": 7.428786754608154,
    "prediction_norm": 5.972918510437012,
    "timestamp": "2025-07-14T12:34:24.618107"
  },
  {
    "epoch": 1433,
    "batch_idx": 192,
    "loss": 0.36338889598846436,
    "learning_rate": 0.00018554330393566384,
    "timestep_mean": 375.75,
    "epsilon_norm": 4.505373001098633,
    "prediction_norm": 3.7351479530334473,
    "timestamp": "2025-07-14T12:34:24.623023"
  },
  {
    "epoch": 1434,
    "batch_idx": 0,
    "loss": 0.28879621624946594,
    "learning_rate": 0.00018493306440731585,
    "timestep_mean": 530.28125,
    "epsilon_norm": 8.212912559509277,
    "prediction_norm": 7.184516429901123,
    "timestamp": "2025-07-14T12:34:24.628083"
  },
  {
    "epoch": 1434,
    "batch_idx": 32,
    "loss": 0.43615642189979553,
    "learning_rate": 0.00018493306440731585,
    "timestep_mean": 409.0625,
    "epsilon_norm": 7.942254543304443,
    "prediction_norm": 5.802672386169434,
    "timestamp": "2025-07-14T12:34:24.632971"
  },
  {
    "epoch": 1434,
    "batch_idx": 64,
    "loss": 0.1546781361103058,
    "learning_rate": 0.00018493306440731585,
    "timestep_mean": 543.65625,
    "epsilon_norm": 7.910573482513428,
    "prediction_norm": 7.449491500854492,
    "timestamp": "2025-07-14T12:34:24.638118"
  },
  {
    "epoch": 1434,
    "batch_idx": 96,
    "loss": 0.19926688075065613,
    "learning_rate": 0.00018493306440731585,
    "timestep_mean": 474.34375,
    "epsilon_norm": 8.105863571166992,
    "prediction_norm": 7.003291130065918,
    "timestamp": "2025-07-14T12:34:24.643001"
  },
  {
    "epoch": 1434,
    "batch_idx": 128,
    "loss": 0.2740404009819031,
    "learning_rate": 0.00018493306440731585,
    "timestep_mean": 580.90625,
    "epsilon_norm": 9.098438262939453,
    "prediction_norm": 7.697511672973633,
    "timestamp": "2025-07-14T12:34:24.647889"
  },
  {
    "epoch": 1434,
    "batch_idx": 160,
    "loss": 0.3223097324371338,
    "learning_rate": 0.00018493306440731585,
    "timestep_mean": 479.25,
    "epsilon_norm": 7.938066005706787,
    "prediction_norm": 6.6757588386535645,
    "timestamp": "2025-07-14T12:34:24.652754"
  },
  {
    "epoch": 1434,
    "batch_idx": 192,
    "loss": 0.28882071375846863,
    "learning_rate": 0.00018493306440731585,
    "timestep_mean": 440.0,
    "epsilon_norm": 4.805360317230225,
    "prediction_norm": 4.221378803253174,
    "timestamp": "2025-07-14T12:34:24.657604"
  },
  {
    "epoch": 1435,
    "batch_idx": 0,
    "loss": 0.2235082983970642,
    "learning_rate": 0.00018432360227531145,
    "timestep_mean": 466.1875,
    "epsilon_norm": 8.004372596740723,
    "prediction_norm": 6.854931831359863,
    "timestamp": "2025-07-14T12:34:24.662552"
  },
  {
    "epoch": 1435,
    "batch_idx": 32,
    "loss": 0.21939577162265778,
    "learning_rate": 0.00018432360227531145,
    "timestep_mean": 464.0,
    "epsilon_norm": 6.851395130157471,
    "prediction_norm": 5.890904903411865,
    "timestamp": "2025-07-14T12:34:24.667548"
  },
  {
    "epoch": 1435,
    "batch_idx": 64,
    "loss": 0.27543574571609497,
    "learning_rate": 0.00018432360227531145,
    "timestep_mean": 471.875,
    "epsilon_norm": 7.16637659072876,
    "prediction_norm": 5.407825469970703,
    "timestamp": "2025-07-14T12:34:24.672379"
  },
  {
    "epoch": 1435,
    "batch_idx": 96,
    "loss": 0.5075920820236206,
    "learning_rate": 0.00018432360227531145,
    "timestep_mean": 470.3125,
    "epsilon_norm": 10.054308891296387,
    "prediction_norm": 7.7769880294799805,
    "timestamp": "2025-07-14T12:34:24.677253"
  },
  {
    "epoch": 1435,
    "batch_idx": 128,
    "loss": 0.2049599140882492,
    "learning_rate": 0.00018432360227531145,
    "timestep_mean": 514.125,
    "epsilon_norm": 8.757975578308105,
    "prediction_norm": 7.8326873779296875,
    "timestamp": "2025-07-14T12:34:24.682111"
  },
  {
    "epoch": 1435,
    "batch_idx": 160,
    "loss": 0.31525760889053345,
    "learning_rate": 0.00018432360227531145,
    "timestep_mean": 493.6875,
    "epsilon_norm": 8.575121879577637,
    "prediction_norm": 7.689351558685303,
    "timestamp": "2025-07-14T12:34:24.687014"
  },
  {
    "epoch": 1435,
    "batch_idx": 192,
    "loss": 0.08310258388519287,
    "learning_rate": 0.00018432360227531145,
    "timestep_mean": 383.75,
    "epsilon_norm": 3.5192034244537354,
    "prediction_norm": 3.573118209838867,
    "timestamp": "2025-07-14T12:34:24.691951"
  },
  {
    "epoch": 1436,
    "batch_idx": 0,
    "loss": 0.3317202925682068,
    "learning_rate": 0.0001837149190434381,
    "timestep_mean": 476.6875,
    "epsilon_norm": 8.277050018310547,
    "prediction_norm": 6.811760425567627,
    "timestamp": "2025-07-14T12:34:24.696834"
  },
  {
    "epoch": 1436,
    "batch_idx": 32,
    "loss": 0.23821403086185455,
    "learning_rate": 0.0001837149190434381,
    "timestep_mean": 518.09375,
    "epsilon_norm": 7.761085033416748,
    "prediction_norm": 7.20115327835083,
    "timestamp": "2025-07-14T12:34:24.701770"
  },
  {
    "epoch": 1436,
    "batch_idx": 64,
    "loss": 0.2734659016132355,
    "learning_rate": 0.0001837149190434381,
    "timestep_mean": 502.5625,
    "epsilon_norm": 7.18071985244751,
    "prediction_norm": 5.5327982902526855,
    "timestamp": "2025-07-14T12:34:24.706630"
  },
  {
    "epoch": 1436,
    "batch_idx": 96,
    "loss": 0.350508451461792,
    "learning_rate": 0.0001837149190434381,
    "timestep_mean": 581.125,
    "epsilon_norm": 9.63939094543457,
    "prediction_norm": 8.19411563873291,
    "timestamp": "2025-07-14T12:34:24.711473"
  },
  {
    "epoch": 1436,
    "batch_idx": 128,
    "loss": 0.26194101572036743,
    "learning_rate": 0.0001837149190434381,
    "timestep_mean": 550.625,
    "epsilon_norm": 8.918825149536133,
    "prediction_norm": 8.053762435913086,
    "timestamp": "2025-07-14T12:34:24.716389"
  },
  {
    "epoch": 1436,
    "batch_idx": 160,
    "loss": 0.4356614947319031,
    "learning_rate": 0.0001837149190434381,
    "timestep_mean": 431.03125,
    "epsilon_norm": 8.498054504394531,
    "prediction_norm": 7.233198165893555,
    "timestamp": "2025-07-14T12:34:24.721253"
  },
  {
    "epoch": 1436,
    "batch_idx": 192,
    "loss": 0.618337869644165,
    "learning_rate": 0.0001837149190434381,
    "timestep_mean": 430.0,
    "epsilon_norm": 3.875361442565918,
    "prediction_norm": 2.2299530506134033,
    "timestamp": "2025-07-14T12:34:24.726105"
  },
  {
    "epoch": 1437,
    "batch_idx": 0,
    "loss": 0.2528397738933563,
    "learning_rate": 0.00018310701621356105,
    "timestep_mean": 464.75,
    "epsilon_norm": 7.694448947906494,
    "prediction_norm": 7.328802108764648,
    "timestamp": "2025-07-14T12:34:24.730960"
  },
  {
    "epoch": 1437,
    "batch_idx": 32,
    "loss": 0.42443910241127014,
    "learning_rate": 0.00018310701621356105,
    "timestep_mean": 425.0,
    "epsilon_norm": 8.795433044433594,
    "prediction_norm": 7.362683296203613,
    "timestamp": "2025-07-14T12:34:24.735883"
  },
  {
    "epoch": 1437,
    "batch_idx": 64,
    "loss": 0.18086373805999756,
    "learning_rate": 0.00018310701621356105,
    "timestep_mean": 504.46875,
    "epsilon_norm": 6.80230188369751,
    "prediction_norm": 5.916916370391846,
    "timestamp": "2025-07-14T12:34:24.740720"
  },
  {
    "epoch": 1437,
    "batch_idx": 96,
    "loss": 0.4819345474243164,
    "learning_rate": 0.00018310701621356105,
    "timestep_mean": 359.34375,
    "epsilon_norm": 8.425444602966309,
    "prediction_norm": 7.116919040679932,
    "timestamp": "2025-07-14T12:34:24.745675"
  },
  {
    "epoch": 1437,
    "batch_idx": 128,
    "loss": 0.25624364614486694,
    "learning_rate": 0.00018310701621356105,
    "timestep_mean": 420.34375,
    "epsilon_norm": 8.273198127746582,
    "prediction_norm": 7.8515305519104,
    "timestamp": "2025-07-14T12:34:24.750662"
  },
  {
    "epoch": 1437,
    "batch_idx": 160,
    "loss": 0.13485082983970642,
    "learning_rate": 0.00018310701621356105,
    "timestep_mean": 557.09375,
    "epsilon_norm": 6.5878729820251465,
    "prediction_norm": 6.58478307723999,
    "timestamp": "2025-07-14T12:34:24.755576"
  },
  {
    "epoch": 1437,
    "batch_idx": 192,
    "loss": 0.7095810174942017,
    "learning_rate": 0.00018310701621356105,
    "timestep_mean": 384.625,
    "epsilon_norm": 4.468555927276611,
    "prediction_norm": 3.38625431060791,
    "timestamp": "2025-07-14T12:34:24.760425"
  },
  {
    "epoch": 1438,
    "batch_idx": 0,
    "loss": 0.41846731305122375,
    "learning_rate": 0.00018249989528561992,
    "timestep_mean": 549.65625,
    "epsilon_norm": 10.076094627380371,
    "prediction_norm": 7.791065692901611,
    "timestamp": "2025-07-14T12:34:24.765389"
  },
  {
    "epoch": 1438,
    "batch_idx": 32,
    "loss": 0.3451237082481384,
    "learning_rate": 0.00018249989528561992,
    "timestep_mean": 534.4375,
    "epsilon_norm": 8.492623329162598,
    "prediction_norm": 6.972444534301758,
    "timestamp": "2025-07-14T12:34:24.770276"
  },
  {
    "epoch": 1438,
    "batch_idx": 64,
    "loss": 0.10052041709423065,
    "learning_rate": 0.00018249989528561992,
    "timestep_mean": 583.84375,
    "epsilon_norm": 8.284223556518555,
    "prediction_norm": 8.109000205993652,
    "timestamp": "2025-07-14T12:34:24.775274"
  },
  {
    "epoch": 1438,
    "batch_idx": 96,
    "loss": 0.3899076581001282,
    "learning_rate": 0.00018249989528561992,
    "timestep_mean": 458.625,
    "epsilon_norm": 8.571769714355469,
    "prediction_norm": 6.881420612335205,
    "timestamp": "2025-07-14T12:34:24.780101"
  },
  {
    "epoch": 1438,
    "batch_idx": 128,
    "loss": 0.3454240560531616,
    "learning_rate": 0.00018249989528561992,
    "timestep_mean": 413.6875,
    "epsilon_norm": 8.5682373046875,
    "prediction_norm": 6.7607645988464355,
    "timestamp": "2025-07-14T12:34:24.785019"
  },
  {
    "epoch": 1438,
    "batch_idx": 160,
    "loss": 0.5322392582893372,
    "learning_rate": 0.00018249989528561992,
    "timestep_mean": 433.6875,
    "epsilon_norm": 9.052774429321289,
    "prediction_norm": 7.361989974975586,
    "timestamp": "2025-07-14T12:34:24.789909"
  },
  {
    "epoch": 1438,
    "batch_idx": 192,
    "loss": 0.16577261686325073,
    "learning_rate": 0.00018249989528561992,
    "timestep_mean": 520.125,
    "epsilon_norm": 3.8023667335510254,
    "prediction_norm": 3.0072779655456543,
    "timestamp": "2025-07-14T12:34:24.794784"
  },
  {
    "epoch": 1439,
    "batch_idx": 0,
    "loss": 0.2874693274497986,
    "learning_rate": 0.00018189355775762576,
    "timestep_mean": 413.78125,
    "epsilon_norm": 7.931785583496094,
    "prediction_norm": 6.447669506072998,
    "timestamp": "2025-07-14T12:34:24.799751"
  },
  {
    "epoch": 1439,
    "batch_idx": 32,
    "loss": 0.1961624026298523,
    "learning_rate": 0.00018189355775762576,
    "timestep_mean": 523.65625,
    "epsilon_norm": 8.386017799377441,
    "prediction_norm": 7.853651523590088,
    "timestamp": "2025-07-14T12:34:24.804600"
  },
  {
    "epoch": 1439,
    "batch_idx": 64,
    "loss": 0.3464910089969635,
    "learning_rate": 0.00018189355775762576,
    "timestep_mean": 519.03125,
    "epsilon_norm": 7.961947917938232,
    "prediction_norm": 5.515257358551025,
    "timestamp": "2025-07-14T12:34:24.809404"
  },
  {
    "epoch": 1439,
    "batch_idx": 96,
    "loss": 0.21510115265846252,
    "learning_rate": 0.00018189355775762576,
    "timestep_mean": 414.78125,
    "epsilon_norm": 8.175006866455078,
    "prediction_norm": 7.589753150939941,
    "timestamp": "2025-07-14T12:34:24.814247"
  },
  {
    "epoch": 1439,
    "batch_idx": 128,
    "loss": 0.262234091758728,
    "learning_rate": 0.00018189355775762576,
    "timestep_mean": 561.875,
    "epsilon_norm": 7.803542137145996,
    "prediction_norm": 6.5713396072387695,
    "timestamp": "2025-07-14T12:34:24.819156"
  },
  {
    "epoch": 1439,
    "batch_idx": 160,
    "loss": 0.41464024782180786,
    "learning_rate": 0.00018189355775762576,
    "timestep_mean": 417.90625,
    "epsilon_norm": 7.539570331573486,
    "prediction_norm": 6.3089494705200195,
    "timestamp": "2025-07-14T12:34:24.824030"
  },
  {
    "epoch": 1439,
    "batch_idx": 192,
    "loss": 0.2372148483991623,
    "learning_rate": 0.00018189355775762576,
    "timestep_mean": 467.75,
    "epsilon_norm": 4.604852676391602,
    "prediction_norm": 4.625219345092773,
    "timestamp": "2025-07-14T12:34:24.828847"
  },
  {
    "epoch": 1440,
    "batch_idx": 0,
    "loss": 0.16698786616325378,
    "learning_rate": 0.00018128800512565542,
    "timestep_mean": 423.34375,
    "epsilon_norm": 7.507872581481934,
    "prediction_norm": 7.357694625854492,
    "timestamp": "2025-07-14T12:34:24.833830"
  },
  {
    "epoch": 1440,
    "batch_idx": 32,
    "loss": 0.16483061015605927,
    "learning_rate": 0.00018128800512565542,
    "timestep_mean": 517.84375,
    "epsilon_norm": 6.460778713226318,
    "prediction_norm": 6.083678245544434,
    "timestamp": "2025-07-14T12:34:24.838684"
  },
  {
    "epoch": 1440,
    "batch_idx": 64,
    "loss": 0.4277929663658142,
    "learning_rate": 0.00018128800512565542,
    "timestep_mean": 449.53125,
    "epsilon_norm": 7.864844799041748,
    "prediction_norm": 5.849039077758789,
    "timestamp": "2025-07-14T12:34:24.843618"
  },
  {
    "epoch": 1440,
    "batch_idx": 96,
    "loss": 0.1682872474193573,
    "learning_rate": 0.00018128800512565542,
    "timestep_mean": 587.40625,
    "epsilon_norm": 8.044635772705078,
    "prediction_norm": 7.498630046844482,
    "timestamp": "2025-07-14T12:34:24.848540"
  },
  {
    "epoch": 1440,
    "batch_idx": 128,
    "loss": 0.5559219121932983,
    "learning_rate": 0.00018128800512565542,
    "timestep_mean": 439.5,
    "epsilon_norm": 7.493685245513916,
    "prediction_norm": 5.706573963165283,
    "timestamp": "2025-07-14T12:34:24.853508"
  },
  {
    "epoch": 1440,
    "batch_idx": 160,
    "loss": 0.3300195336341858,
    "learning_rate": 0.00018128800512565542,
    "timestep_mean": 497.75,
    "epsilon_norm": 7.6007466316223145,
    "prediction_norm": 6.782817363739014,
    "timestamp": "2025-07-14T12:34:24.858574"
  },
  {
    "epoch": 1440,
    "batch_idx": 192,
    "loss": 0.5050622820854187,
    "learning_rate": 0.00018128800512565542,
    "timestep_mean": 441.0,
    "epsilon_norm": 4.022254467010498,
    "prediction_norm": 2.4681036472320557,
    "timestamp": "2025-07-14T12:34:24.863429"
  },
  {
    "epoch": 1441,
    "batch_idx": 0,
    "loss": 0.36791521310806274,
    "learning_rate": 0.00018068323888385026,
    "timestep_mean": 483.875,
    "epsilon_norm": 9.447298049926758,
    "prediction_norm": 7.569186687469482,
    "timestamp": "2025-07-14T12:34:24.868350"
  },
  {
    "epoch": 1441,
    "batch_idx": 32,
    "loss": 0.17579960823059082,
    "learning_rate": 0.00018068323888385026,
    "timestep_mean": 503.28125,
    "epsilon_norm": 8.614255905151367,
    "prediction_norm": 7.173122406005859,
    "timestamp": "2025-07-14T12:34:24.873250"
  },
  {
    "epoch": 1441,
    "batch_idx": 64,
    "loss": 0.12718507647514343,
    "learning_rate": 0.00018068323888385026,
    "timestep_mean": 541.65625,
    "epsilon_norm": 6.955624103546143,
    "prediction_norm": 6.360011100769043,
    "timestamp": "2025-07-14T12:34:24.878135"
  },
  {
    "epoch": 1441,
    "batch_idx": 96,
    "loss": 0.4214118719100952,
    "learning_rate": 0.00018068323888385026,
    "timestep_mean": 524.59375,
    "epsilon_norm": 9.829096794128418,
    "prediction_norm": 7.88726806640625,
    "timestamp": "2025-07-14T12:34:24.883049"
  },
  {
    "epoch": 1441,
    "batch_idx": 128,
    "loss": 0.21482080221176147,
    "learning_rate": 0.00018068323888385026,
    "timestep_mean": 512.0625,
    "epsilon_norm": 8.675130844116211,
    "prediction_norm": 7.174029350280762,
    "timestamp": "2025-07-14T12:34:24.887969"
  },
  {
    "epoch": 1441,
    "batch_idx": 160,
    "loss": 0.23651129007339478,
    "learning_rate": 0.00018068323888385026,
    "timestep_mean": 512.59375,
    "epsilon_norm": 7.12557315826416,
    "prediction_norm": 5.469288349151611,
    "timestamp": "2025-07-14T12:34:24.892936"
  },
  {
    "epoch": 1441,
    "batch_idx": 192,
    "loss": 0.06427977979183197,
    "learning_rate": 0.00018068323888385026,
    "timestep_mean": 532.375,
    "epsilon_norm": 3.8222403526306152,
    "prediction_norm": 3.6270339488983154,
    "timestamp": "2025-07-14T12:34:24.897833"
  },
  {
    "epoch": 1442,
    "batch_idx": 0,
    "loss": 0.3194313645362854,
    "learning_rate": 0.00018007926052441116,
    "timestep_mean": 496.71875,
    "epsilon_norm": 7.635438919067383,
    "prediction_norm": 6.2399725914001465,
    "timestamp": "2025-07-14T12:34:24.902728"
  },
  {
    "epoch": 1442,
    "batch_idx": 32,
    "loss": 0.17151768505573273,
    "learning_rate": 0.00018007926052441116,
    "timestep_mean": 514.6875,
    "epsilon_norm": 7.796504974365234,
    "prediction_norm": 7.0720906257629395,
    "timestamp": "2025-07-14T12:34:24.907699"
  },
  {
    "epoch": 1442,
    "batch_idx": 64,
    "loss": 0.16047701239585876,
    "learning_rate": 0.00018007926052441116,
    "timestep_mean": 473.71875,
    "epsilon_norm": 6.763369083404541,
    "prediction_norm": 5.894476890563965,
    "timestamp": "2025-07-14T12:34:24.912640"
  },
  {
    "epoch": 1442,
    "batch_idx": 96,
    "loss": 0.5023742318153381,
    "learning_rate": 0.00018007926052441116,
    "timestep_mean": 498.78125,
    "epsilon_norm": 8.852346420288086,
    "prediction_norm": 7.277969837188721,
    "timestamp": "2025-07-14T12:34:24.917616"
  },
  {
    "epoch": 1442,
    "batch_idx": 128,
    "loss": 0.16459733247756958,
    "learning_rate": 0.00018007926052441116,
    "timestep_mean": 545.71875,
    "epsilon_norm": 6.283699989318848,
    "prediction_norm": 6.069761753082275,
    "timestamp": "2025-07-14T12:34:24.922465"
  },
  {
    "epoch": 1442,
    "batch_idx": 160,
    "loss": 0.2675117254257202,
    "learning_rate": 0.00018007926052441116,
    "timestep_mean": 613.8125,
    "epsilon_norm": 8.187456130981445,
    "prediction_norm": 6.967276573181152,
    "timestamp": "2025-07-14T12:34:24.927386"
  },
  {
    "epoch": 1442,
    "batch_idx": 192,
    "loss": 0.1646335870027542,
    "learning_rate": 0.00018007926052441116,
    "timestep_mean": 696.375,
    "epsilon_norm": 4.762986183166504,
    "prediction_norm": 4.36180305480957,
    "timestamp": "2025-07-14T12:34:24.932336"
  },
  {
    "epoch": 1443,
    "batch_idx": 0,
    "loss": 0.27860280871391296,
    "learning_rate": 0.00017947607153759394,
    "timestep_mean": 539.46875,
    "epsilon_norm": 7.547464847564697,
    "prediction_norm": 7.171473026275635,
    "timestamp": "2025-07-14T12:34:24.937147"
  },
  {
    "epoch": 1443,
    "batch_idx": 32,
    "loss": 0.2179393768310547,
    "learning_rate": 0.00017947607153759394,
    "timestep_mean": 542.3125,
    "epsilon_norm": 8.167362213134766,
    "prediction_norm": 7.463416576385498,
    "timestamp": "2025-07-14T12:34:24.942142"
  },
  {
    "epoch": 1443,
    "batch_idx": 64,
    "loss": 0.2026975452899933,
    "learning_rate": 0.00017947607153759394,
    "timestep_mean": 634.09375,
    "epsilon_norm": 7.303557395935059,
    "prediction_norm": 6.503452777862549,
    "timestamp": "2025-07-14T12:34:24.947002"
  },
  {
    "epoch": 1443,
    "batch_idx": 96,
    "loss": 0.30810609459877014,
    "learning_rate": 0.00017947607153759394,
    "timestep_mean": 424.5,
    "epsilon_norm": 8.385650634765625,
    "prediction_norm": 6.9218974113464355,
    "timestamp": "2025-07-14T12:34:24.951882"
  },
  {
    "epoch": 1443,
    "batch_idx": 128,
    "loss": 0.47254806756973267,
    "learning_rate": 0.00017947607153759394,
    "timestep_mean": 480.375,
    "epsilon_norm": 8.595782279968262,
    "prediction_norm": 6.76179838180542,
    "timestamp": "2025-07-14T12:34:24.956816"
  },
  {
    "epoch": 1443,
    "batch_idx": 160,
    "loss": 0.3345828354358673,
    "learning_rate": 0.00017947607153759394,
    "timestep_mean": 517.8125,
    "epsilon_norm": 8.806283950805664,
    "prediction_norm": 7.542790412902832,
    "timestamp": "2025-07-14T12:34:24.961738"
  },
  {
    "epoch": 1443,
    "batch_idx": 192,
    "loss": 0.20075097680091858,
    "learning_rate": 0.00017947607153759394,
    "timestep_mean": 493.5,
    "epsilon_norm": 3.2066500186920166,
    "prediction_norm": 2.7073371410369873,
    "timestamp": "2025-07-14T12:34:24.966724"
  },
  {
    "epoch": 1444,
    "batch_idx": 0,
    "loss": 0.20125515758991241,
    "learning_rate": 0.00017887367341170825,
    "timestep_mean": 618.4375,
    "epsilon_norm": 8.237939834594727,
    "prediction_norm": 7.325150489807129,
    "timestamp": "2025-07-14T12:34:24.971649"
  },
  {
    "epoch": 1444,
    "batch_idx": 32,
    "loss": 0.43748900294303894,
    "learning_rate": 0.00017887367341170825,
    "timestep_mean": 472.53125,
    "epsilon_norm": 9.174118041992188,
    "prediction_norm": 7.674837112426758,
    "timestamp": "2025-07-14T12:34:24.976492"
  },
  {
    "epoch": 1444,
    "batch_idx": 64,
    "loss": 0.2727697193622589,
    "learning_rate": 0.00017887367341170825,
    "timestep_mean": 455.625,
    "epsilon_norm": 8.503141403198242,
    "prediction_norm": 7.678977966308594,
    "timestamp": "2025-07-14T12:34:24.981304"
  },
  {
    "epoch": 1444,
    "batch_idx": 96,
    "loss": 0.1993282437324524,
    "learning_rate": 0.00017887367341170825,
    "timestep_mean": 543.8125,
    "epsilon_norm": 8.4078950881958,
    "prediction_norm": 7.599846363067627,
    "timestamp": "2025-07-14T12:34:24.986176"
  },
  {
    "epoch": 1444,
    "batch_idx": 128,
    "loss": 0.32405659556388855,
    "learning_rate": 0.00017887367341170825,
    "timestep_mean": 507.34375,
    "epsilon_norm": 8.472271919250488,
    "prediction_norm": 7.211360931396484,
    "timestamp": "2025-07-14T12:34:24.991106"
  },
  {
    "epoch": 1444,
    "batch_idx": 160,
    "loss": 0.3383544087409973,
    "learning_rate": 0.00017887367341170825,
    "timestep_mean": 551.65625,
    "epsilon_norm": 8.73678970336914,
    "prediction_norm": 7.62448787689209,
    "timestamp": "2025-07-14T12:34:24.995997"
  },
  {
    "epoch": 1444,
    "batch_idx": 192,
    "loss": 0.5710804462432861,
    "learning_rate": 0.00017887367341170825,
    "timestep_mean": 510.125,
    "epsilon_norm": 4.192597389221191,
    "prediction_norm": 2.5916619300842285,
    "timestamp": "2025-07-14T12:34:25.000861"
  },
  {
    "epoch": 1445,
    "batch_idx": 0,
    "loss": 0.14869365096092224,
    "learning_rate": 0.00017827206763311085,
    "timestep_mean": 520.9375,
    "epsilon_norm": 8.115476608276367,
    "prediction_norm": 7.401539325714111,
    "timestamp": "2025-07-14T12:34:25.005821"
  },
  {
    "epoch": 1445,
    "batch_idx": 32,
    "loss": 0.19916291534900665,
    "learning_rate": 0.00017827206763311085,
    "timestep_mean": 556.875,
    "epsilon_norm": 7.125384330749512,
    "prediction_norm": 6.717249393463135,
    "timestamp": "2025-07-14T12:34:25.010687"
  },
  {
    "epoch": 1445,
    "batch_idx": 64,
    "loss": 0.26676392555236816,
    "learning_rate": 0.00017827206763311085,
    "timestep_mean": 484.46875,
    "epsilon_norm": 7.947615146636963,
    "prediction_norm": 6.586587905883789,
    "timestamp": "2025-07-14T12:34:25.015582"
  },
  {
    "epoch": 1445,
    "batch_idx": 96,
    "loss": 0.29507315158843994,
    "learning_rate": 0.00017827206763311085,
    "timestep_mean": 475.53125,
    "epsilon_norm": 8.557926177978516,
    "prediction_norm": 7.015503883361816,
    "timestamp": "2025-07-14T12:34:25.020480"
  },
  {
    "epoch": 1445,
    "batch_idx": 128,
    "loss": 0.5412824153900146,
    "learning_rate": 0.00017827206763311085,
    "timestep_mean": 388.59375,
    "epsilon_norm": 7.778161525726318,
    "prediction_norm": 5.494264602661133,
    "timestamp": "2025-07-14T12:34:25.025496"
  },
  {
    "epoch": 1445,
    "batch_idx": 160,
    "loss": 0.30579185485839844,
    "learning_rate": 0.00017827206763311085,
    "timestep_mean": 537.1875,
    "epsilon_norm": 7.066330909729004,
    "prediction_norm": 6.0337419509887695,
    "timestamp": "2025-07-14T12:34:25.030434"
  },
  {
    "epoch": 1445,
    "batch_idx": 192,
    "loss": 0.0784587562084198,
    "learning_rate": 0.00017827206763311085,
    "timestep_mean": 667.75,
    "epsilon_norm": 2.866462469100952,
    "prediction_norm": 2.6869728565216064,
    "timestamp": "2025-07-14T12:34:25.035245"
  },
  {
    "epoch": 1446,
    "batch_idx": 0,
    "loss": 0.26643744111061096,
    "learning_rate": 0.00017767125568620464,
    "timestep_mean": 536.90625,
    "epsilon_norm": 7.951401233673096,
    "prediction_norm": 6.8237385749816895,
    "timestamp": "2025-07-14T12:34:25.040131"
  },
  {
    "epoch": 1446,
    "batch_idx": 32,
    "loss": 0.3422149419784546,
    "learning_rate": 0.00017767125568620464,
    "timestep_mean": 480.75,
    "epsilon_norm": 8.10626220703125,
    "prediction_norm": 6.55902624130249,
    "timestamp": "2025-07-14T12:34:25.044992"
  },
  {
    "epoch": 1446,
    "batch_idx": 64,
    "loss": 0.2250525802373886,
    "learning_rate": 0.00017767125568620464,
    "timestep_mean": 575.875,
    "epsilon_norm": 8.355961799621582,
    "prediction_norm": 7.723623275756836,
    "timestamp": "2025-07-14T12:34:25.049885"
  },
  {
    "epoch": 1446,
    "batch_idx": 96,
    "loss": 0.2378610372543335,
    "learning_rate": 0.00017767125568620464,
    "timestep_mean": 501.59375,
    "epsilon_norm": 8.001976013183594,
    "prediction_norm": 6.491603374481201,
    "timestamp": "2025-07-14T12:34:25.054824"
  },
  {
    "epoch": 1446,
    "batch_idx": 128,
    "loss": 0.24982252717018127,
    "learning_rate": 0.00017767125568620464,
    "timestep_mean": 518.53125,
    "epsilon_norm": 8.573663711547852,
    "prediction_norm": 7.307530879974365,
    "timestamp": "2025-07-14T12:34:25.059666"
  },
  {
    "epoch": 1446,
    "batch_idx": 160,
    "loss": 0.26395928859710693,
    "learning_rate": 0.00017767125568620464,
    "timestep_mean": 503.90625,
    "epsilon_norm": 8.499321937561035,
    "prediction_norm": 7.258020401000977,
    "timestamp": "2025-07-14T12:34:25.064521"
  },
  {
    "epoch": 1446,
    "batch_idx": 192,
    "loss": 0.14997771382331848,
    "learning_rate": 0.00017767125568620464,
    "timestep_mean": 509.0,
    "epsilon_norm": 3.063962697982788,
    "prediction_norm": 3.1181883811950684,
    "timestamp": "2025-07-14T12:34:25.069390"
  },
  {
    "epoch": 1447,
    "batch_idx": 0,
    "loss": 0.19927138090133667,
    "learning_rate": 0.00017707123905343332,
    "timestep_mean": 468.6875,
    "epsilon_norm": 8.201647758483887,
    "prediction_norm": 6.995530128479004,
    "timestamp": "2025-07-14T12:34:25.074271"
  },
  {
    "epoch": 1447,
    "batch_idx": 32,
    "loss": 0.3353479206562042,
    "learning_rate": 0.00017707123905343332,
    "timestep_mean": 504.03125,
    "epsilon_norm": 8.360551834106445,
    "prediction_norm": 6.312559604644775,
    "timestamp": "2025-07-14T12:34:25.079158"
  },
  {
    "epoch": 1447,
    "batch_idx": 64,
    "loss": 0.15890106558799744,
    "learning_rate": 0.00017707123905343332,
    "timestep_mean": 475.53125,
    "epsilon_norm": 7.729474067687988,
    "prediction_norm": 6.34691858291626,
    "timestamp": "2025-07-14T12:34:25.084206"
  },
  {
    "epoch": 1447,
    "batch_idx": 96,
    "loss": 0.16095082461833954,
    "learning_rate": 0.00017707123905343332,
    "timestep_mean": 549.28125,
    "epsilon_norm": 7.664830684661865,
    "prediction_norm": 6.603568077087402,
    "timestamp": "2025-07-14T12:34:25.089066"
  },
  {
    "epoch": 1447,
    "batch_idx": 128,
    "loss": 0.32476645708084106,
    "learning_rate": 0.00017707123905343332,
    "timestep_mean": 515.15625,
    "epsilon_norm": 9.198760986328125,
    "prediction_norm": 7.509400844573975,
    "timestamp": "2025-07-14T12:34:25.094035"
  },
  {
    "epoch": 1447,
    "batch_idx": 160,
    "loss": 0.3804071843624115,
    "learning_rate": 0.00017707123905343332,
    "timestep_mean": 502.21875,
    "epsilon_norm": 8.282404899597168,
    "prediction_norm": 6.815474510192871,
    "timestamp": "2025-07-14T12:34:25.098932"
  },
  {
    "epoch": 1447,
    "batch_idx": 192,
    "loss": 0.2547721862792969,
    "learning_rate": 0.00017707123905343332,
    "timestep_mean": 578.0,
    "epsilon_norm": 3.341479778289795,
    "prediction_norm": 2.8883557319641113,
    "timestamp": "2025-07-14T12:34:25.103862"
  },
  {
    "epoch": 1448,
    "batch_idx": 0,
    "loss": 0.2138112187385559,
    "learning_rate": 0.00017647201921527808,
    "timestep_mean": 438.3125,
    "epsilon_norm": 6.41572904586792,
    "prediction_norm": 5.336174488067627,
    "timestamp": "2025-07-14T12:34:25.108850"
  },
  {
    "epoch": 1448,
    "batch_idx": 32,
    "loss": 0.2899646759033203,
    "learning_rate": 0.00017647201921527808,
    "timestep_mean": 466.125,
    "epsilon_norm": 8.367164611816406,
    "prediction_norm": 6.627919673919678,
    "timestamp": "2025-07-14T12:34:25.113730"
  },
  {
    "epoch": 1448,
    "batch_idx": 64,
    "loss": 0.1972699761390686,
    "learning_rate": 0.00017647201921527808,
    "timestep_mean": 526.46875,
    "epsilon_norm": 7.57794189453125,
    "prediction_norm": 6.784317493438721,
    "timestamp": "2025-07-14T12:34:25.118591"
  },
  {
    "epoch": 1448,
    "batch_idx": 96,
    "loss": 0.2867446541786194,
    "learning_rate": 0.00017647201921527808,
    "timestep_mean": 491.78125,
    "epsilon_norm": 8.14782428741455,
    "prediction_norm": 7.362462043762207,
    "timestamp": "2025-07-14T12:34:25.123450"
  },
  {
    "epoch": 1448,
    "batch_idx": 128,
    "loss": 0.6350730061531067,
    "learning_rate": 0.00017647201921527808,
    "timestep_mean": 446.71875,
    "epsilon_norm": 7.918659687042236,
    "prediction_norm": 6.6446123123168945,
    "timestamp": "2025-07-14T12:34:25.128304"
  },
  {
    "epoch": 1448,
    "batch_idx": 160,
    "loss": 0.24619387090206146,
    "learning_rate": 0.00017647201921527808,
    "timestep_mean": 500.21875,
    "epsilon_norm": 7.634615898132324,
    "prediction_norm": 6.98299503326416,
    "timestamp": "2025-07-14T12:34:25.133474"
  },
  {
    "epoch": 1448,
    "batch_idx": 192,
    "loss": 0.3382555842399597,
    "learning_rate": 0.00017647201921527808,
    "timestep_mean": 401.0,
    "epsilon_norm": 3.242064952850342,
    "prediction_norm": 2.7676424980163574,
    "timestamp": "2025-07-14T12:34:25.138670"
  },
  {
    "epoch": 1449,
    "batch_idx": 0,
    "loss": 0.21024523675441742,
    "learning_rate": 0.00017587359765025465,
    "timestep_mean": 501.25,
    "epsilon_norm": 6.960334777832031,
    "prediction_norm": 6.201913833618164,
    "timestamp": "2025-07-14T12:34:25.143746"
  },
  {
    "epoch": 1449,
    "batch_idx": 32,
    "loss": 0.2206190526485443,
    "learning_rate": 0.00017587359765025465,
    "timestep_mean": 436.5,
    "epsilon_norm": 7.7419328689575195,
    "prediction_norm": 6.89655065536499,
    "timestamp": "2025-07-14T12:34:25.148914"
  },
  {
    "epoch": 1449,
    "batch_idx": 64,
    "loss": 0.31982237100601196,
    "learning_rate": 0.00017587359765025465,
    "timestep_mean": 485.65625,
    "epsilon_norm": 7.784402847290039,
    "prediction_norm": 6.259206295013428,
    "timestamp": "2025-07-14T12:34:25.154171"
  },
  {
    "epoch": 1449,
    "batch_idx": 96,
    "loss": 0.17587634921073914,
    "learning_rate": 0.00017587359765025465,
    "timestep_mean": 520.03125,
    "epsilon_norm": 8.090935707092285,
    "prediction_norm": 7.283713340759277,
    "timestamp": "2025-07-14T12:34:25.159290"
  },
  {
    "epoch": 1449,
    "batch_idx": 128,
    "loss": 0.3163864016532898,
    "learning_rate": 0.00017587359765025465,
    "timestep_mean": 448.59375,
    "epsilon_norm": 8.215290069580078,
    "prediction_norm": 6.85416316986084,
    "timestamp": "2025-07-14T12:34:25.164464"
  },
  {
    "epoch": 1449,
    "batch_idx": 160,
    "loss": 0.3751971423625946,
    "learning_rate": 0.00017587359765025465,
    "timestep_mean": 519.40625,
    "epsilon_norm": 6.925344467163086,
    "prediction_norm": 5.348341941833496,
    "timestamp": "2025-07-14T12:34:25.169666"
  },
  {
    "epoch": 1449,
    "batch_idx": 192,
    "loss": 0.3703705370426178,
    "learning_rate": 0.00017587359765025465,
    "timestep_mean": 339.875,
    "epsilon_norm": 3.4805092811584473,
    "prediction_norm": 2.5929083824157715,
    "timestamp": "2025-07-14T12:34:25.174764"
  },
  {
    "epoch": 1450,
    "batch_idx": 0,
    "loss": 0.3258017301559448,
    "learning_rate": 0.00017527597583490853,
    "timestep_mean": 439.21875,
    "epsilon_norm": 7.652246952056885,
    "prediction_norm": 5.913812160491943,
    "timestamp": "2025-07-14T12:34:25.180108"
  },
  {
    "epoch": 1450,
    "batch_idx": 32,
    "loss": 0.2173219919204712,
    "learning_rate": 0.00017527597583490853,
    "timestep_mean": 516.84375,
    "epsilon_norm": 7.61539363861084,
    "prediction_norm": 6.595381736755371,
    "timestamp": "2025-07-14T12:34:25.185199"
  },
  {
    "epoch": 1450,
    "batch_idx": 64,
    "loss": 0.1609877049922943,
    "learning_rate": 0.00017527597583490853,
    "timestep_mean": 499.53125,
    "epsilon_norm": 7.27288293838501,
    "prediction_norm": 6.742843151092529,
    "timestamp": "2025-07-14T12:34:25.190522"
  },
  {
    "epoch": 1450,
    "batch_idx": 96,
    "loss": 0.5509241819381714,
    "learning_rate": 0.00017527597583490853,
    "timestep_mean": 428.65625,
    "epsilon_norm": 8.620471954345703,
    "prediction_norm": 5.735384941101074,
    "timestamp": "2025-07-14T12:34:25.195731"
  },
  {
    "epoch": 1450,
    "batch_idx": 128,
    "loss": 0.2944192886352539,
    "learning_rate": 0.00017527597583490853,
    "timestep_mean": 473.375,
    "epsilon_norm": 6.512514114379883,
    "prediction_norm": 6.177820682525635,
    "timestamp": "2025-07-14T12:34:25.200897"
  },
  {
    "epoch": 1450,
    "batch_idx": 160,
    "loss": 0.3023748993873596,
    "learning_rate": 0.00017527597583490853,
    "timestep_mean": 471.875,
    "epsilon_norm": 7.742345333099365,
    "prediction_norm": 6.357326984405518,
    "timestamp": "2025-07-14T12:34:25.206044"
  },
  {
    "epoch": 1450,
    "batch_idx": 192,
    "loss": 0.046791285276412964,
    "learning_rate": 0.00017527597583490853,
    "timestep_mean": 595.5,
    "epsilon_norm": 4.277257442474365,
    "prediction_norm": 4.032283306121826,
    "timestamp": "2025-07-14T12:34:25.211189"
  },
  {
    "epoch": 1451,
    "batch_idx": 0,
    "loss": 0.3850589096546173,
    "learning_rate": 0.00017467915524381212,
    "timestep_mean": 545.40625,
    "epsilon_norm": 8.4419527053833,
    "prediction_norm": 7.041897773742676,
    "timestamp": "2025-07-14T12:34:25.216399"
  },
  {
    "epoch": 1451,
    "batch_idx": 32,
    "loss": 0.2338089942932129,
    "learning_rate": 0.00017467915524381212,
    "timestep_mean": 423.375,
    "epsilon_norm": 7.936583042144775,
    "prediction_norm": 6.494045257568359,
    "timestamp": "2025-07-14T12:34:25.221594"
  },
  {
    "epoch": 1451,
    "batch_idx": 64,
    "loss": 0.2557632625102997,
    "learning_rate": 0.00017467915524381212,
    "timestep_mean": 541.0625,
    "epsilon_norm": 8.924577713012695,
    "prediction_norm": 7.111883640289307,
    "timestamp": "2025-07-14T12:34:25.226739"
  },
  {
    "epoch": 1451,
    "batch_idx": 96,
    "loss": 0.2543509006500244,
    "learning_rate": 0.00017467915524381212,
    "timestep_mean": 496.46875,
    "epsilon_norm": 7.443253517150879,
    "prediction_norm": 6.1493239402771,
    "timestamp": "2025-07-14T12:34:25.231890"
  },
  {
    "epoch": 1451,
    "batch_idx": 128,
    "loss": 0.29730185866355896,
    "learning_rate": 0.00017467915524381212,
    "timestep_mean": 552.40625,
    "epsilon_norm": 8.22433090209961,
    "prediction_norm": 6.09235143661499,
    "timestamp": "2025-07-14T12:34:25.237107"
  },
  {
    "epoch": 1451,
    "batch_idx": 160,
    "loss": 0.3889238238334656,
    "learning_rate": 0.00017467915524381212,
    "timestep_mean": 503.84375,
    "epsilon_norm": 8.482176780700684,
    "prediction_norm": 7.0085344314575195,
    "timestamp": "2025-07-14T12:34:25.242497"
  },
  {
    "epoch": 1451,
    "batch_idx": 192,
    "loss": 0.2000320553779602,
    "learning_rate": 0.00017467915524381212,
    "timestep_mean": 631.125,
    "epsilon_norm": 4.494511604309082,
    "prediction_norm": 3.9880852699279785,
    "timestamp": "2025-07-14T12:34:25.247541"
  },
  {
    "epoch": 1452,
    "batch_idx": 0,
    "loss": 0.33761560916900635,
    "learning_rate": 0.00017408313734956085,
    "timestep_mean": 434.1875,
    "epsilon_norm": 8.520113945007324,
    "prediction_norm": 6.721047878265381,
    "timestamp": "2025-07-14T12:34:25.252769"
  },
  {
    "epoch": 1452,
    "batch_idx": 32,
    "loss": 0.21196478605270386,
    "learning_rate": 0.00017408313734956085,
    "timestep_mean": 599.34375,
    "epsilon_norm": 7.895480155944824,
    "prediction_norm": 6.7603230476379395,
    "timestamp": "2025-07-14T12:34:25.257932"
  },
  {
    "epoch": 1452,
    "batch_idx": 64,
    "loss": 0.2030160129070282,
    "learning_rate": 0.00017408313734956085,
    "timestep_mean": 534.65625,
    "epsilon_norm": 6.774379730224609,
    "prediction_norm": 6.34716272354126,
    "timestamp": "2025-07-14T12:34:25.263189"
  },
  {
    "epoch": 1452,
    "batch_idx": 96,
    "loss": 0.2064094841480255,
    "learning_rate": 0.00017408313734956085,
    "timestep_mean": 517.6875,
    "epsilon_norm": 7.690552234649658,
    "prediction_norm": 6.798238277435303,
    "timestamp": "2025-07-14T12:34:25.268366"
  },
  {
    "epoch": 1452,
    "batch_idx": 128,
    "loss": 0.18435345590114594,
    "learning_rate": 0.00017408313734956085,
    "timestep_mean": 475.9375,
    "epsilon_norm": 8.393060684204102,
    "prediction_norm": 7.228959560394287,
    "timestamp": "2025-07-14T12:34:25.273486"
  },
  {
    "epoch": 1452,
    "batch_idx": 160,
    "loss": 0.3443523049354553,
    "learning_rate": 0.00017408313734956085,
    "timestep_mean": 433.8125,
    "epsilon_norm": 8.895931243896484,
    "prediction_norm": 7.685484409332275,
    "timestamp": "2025-07-14T12:34:25.278537"
  },
  {
    "epoch": 1452,
    "batch_idx": 192,
    "loss": 0.24413836002349854,
    "learning_rate": 0.00017408313734956085,
    "timestep_mean": 370.375,
    "epsilon_norm": 2.406534433364868,
    "prediction_norm": 2.2416419982910156,
    "timestamp": "2025-07-14T12:34:25.283877"
  },
  {
    "epoch": 1453,
    "batch_idx": 0,
    "loss": 0.38885682821273804,
    "learning_rate": 0.0001734879236227699,
    "timestep_mean": 455.28125,
    "epsilon_norm": 7.333725452423096,
    "prediction_norm": 5.867558479309082,
    "timestamp": "2025-07-14T12:34:25.289087"
  },
  {
    "epoch": 1453,
    "batch_idx": 32,
    "loss": 0.21147547662258148,
    "learning_rate": 0.0001734879236227699,
    "timestep_mean": 484.0625,
    "epsilon_norm": 7.124466896057129,
    "prediction_norm": 6.627152442932129,
    "timestamp": "2025-07-14T12:34:25.294233"
  },
  {
    "epoch": 1453,
    "batch_idx": 64,
    "loss": 0.4160469174385071,
    "learning_rate": 0.0001734879236227699,
    "timestep_mean": 396.25,
    "epsilon_norm": 8.465209007263184,
    "prediction_norm": 7.131287097930908,
    "timestamp": "2025-07-14T12:34:25.299455"
  },
  {
    "epoch": 1453,
    "batch_idx": 96,
    "loss": 0.22266337275505066,
    "learning_rate": 0.0001734879236227699,
    "timestep_mean": 495.4375,
    "epsilon_norm": 8.973358154296875,
    "prediction_norm": 7.310347080230713,
    "timestamp": "2025-07-14T12:34:25.304789"
  },
  {
    "epoch": 1453,
    "batch_idx": 128,
    "loss": 0.1989191770553589,
    "learning_rate": 0.0001734879236227699,
    "timestep_mean": 490.78125,
    "epsilon_norm": 8.524526596069336,
    "prediction_norm": 7.782834529876709,
    "timestamp": "2025-07-14T12:34:25.309734"
  },
  {
    "epoch": 1453,
    "batch_idx": 160,
    "loss": 0.1629733443260193,
    "learning_rate": 0.0001734879236227699,
    "timestep_mean": 476.65625,
    "epsilon_norm": 7.488554000854492,
    "prediction_norm": 6.543300628662109,
    "timestamp": "2025-07-14T12:34:25.314645"
  },
  {
    "epoch": 1453,
    "batch_idx": 192,
    "loss": 0.2407960295677185,
    "learning_rate": 0.0001734879236227699,
    "timestep_mean": 580.5,
    "epsilon_norm": 3.81660795211792,
    "prediction_norm": 2.941066026687622,
    "timestamp": "2025-07-14T12:34:25.319639"
  },
  {
    "epoch": 1454,
    "batch_idx": 0,
    "loss": 0.24511653184890747,
    "learning_rate": 0.00017289351553206975,
    "timestep_mean": 541.25,
    "epsilon_norm": 7.989871978759766,
    "prediction_norm": 6.8226637840271,
    "timestamp": "2025-07-14T12:34:25.324600"
  },
  {
    "epoch": 1454,
    "batch_idx": 32,
    "loss": 0.3659922480583191,
    "learning_rate": 0.00017289351553206975,
    "timestep_mean": 408.78125,
    "epsilon_norm": 7.888742923736572,
    "prediction_norm": 6.396838188171387,
    "timestamp": "2025-07-14T12:34:25.329862"
  },
  {
    "epoch": 1454,
    "batch_idx": 64,
    "loss": 0.11821290850639343,
    "learning_rate": 0.00017289351553206975,
    "timestep_mean": 542.1875,
    "epsilon_norm": 7.643381595611572,
    "prediction_norm": 6.7433977127075195,
    "timestamp": "2025-07-14T12:34:25.334848"
  },
  {
    "epoch": 1454,
    "batch_idx": 96,
    "loss": 0.38962090015411377,
    "learning_rate": 0.00017289351553206975,
    "timestep_mean": 427.625,
    "epsilon_norm": 7.47415828704834,
    "prediction_norm": 5.923102855682373,
    "timestamp": "2025-07-14T12:34:25.339903"
  },
  {
    "epoch": 1454,
    "batch_idx": 128,
    "loss": 0.6052130460739136,
    "learning_rate": 0.00017289351553206975,
    "timestep_mean": 468.71875,
    "epsilon_norm": 9.256617546081543,
    "prediction_norm": 6.996615409851074,
    "timestamp": "2025-07-14T12:34:25.345008"
  },
  {
    "epoch": 1454,
    "batch_idx": 160,
    "loss": 0.3548215627670288,
    "learning_rate": 0.00017289351553206975,
    "timestep_mean": 460.09375,
    "epsilon_norm": 7.494563579559326,
    "prediction_norm": 6.086157321929932,
    "timestamp": "2025-07-14T12:34:25.350276"
  },
  {
    "epoch": 1454,
    "batch_idx": 192,
    "loss": 0.08216530084609985,
    "learning_rate": 0.00017289351553206975,
    "timestep_mean": 587.75,
    "epsilon_norm": 2.6626405715942383,
    "prediction_norm": 2.595360517501831,
    "timestamp": "2025-07-14T12:34:25.355307"
  },
  {
    "epoch": 1455,
    "batch_idx": 0,
    "loss": 0.12789735198020935,
    "learning_rate": 0.0001722999145441034,
    "timestep_mean": 543.59375,
    "epsilon_norm": 8.14521312713623,
    "prediction_norm": 7.540624141693115,
    "timestamp": "2025-07-14T12:34:25.360385"
  },
  {
    "epoch": 1455,
    "batch_idx": 32,
    "loss": 0.3128359019756317,
    "learning_rate": 0.0001722999145441034,
    "timestep_mean": 534.6875,
    "epsilon_norm": 8.905343055725098,
    "prediction_norm": 7.276373863220215,
    "timestamp": "2025-07-14T12:34:25.365555"
  },
  {
    "epoch": 1455,
    "batch_idx": 64,
    "loss": 0.5469359159469604,
    "learning_rate": 0.0001722999145441034,
    "timestep_mean": 575.46875,
    "epsilon_norm": 9.154167175292969,
    "prediction_norm": 7.382422924041748,
    "timestamp": "2025-07-14T12:34:25.370676"
  },
  {
    "epoch": 1455,
    "batch_idx": 96,
    "loss": 0.22288846969604492,
    "learning_rate": 0.0001722999145441034,
    "timestep_mean": 545.34375,
    "epsilon_norm": 8.195882797241211,
    "prediction_norm": 6.9574761390686035,
    "timestamp": "2025-07-14T12:34:25.376023"
  },
  {
    "epoch": 1455,
    "batch_idx": 128,
    "loss": 0.29307109117507935,
    "learning_rate": 0.0001722999145441034,
    "timestep_mean": 412.28125,
    "epsilon_norm": 8.262533187866211,
    "prediction_norm": 6.991377353668213,
    "timestamp": "2025-07-14T12:34:25.381212"
  },
  {
    "epoch": 1455,
    "batch_idx": 160,
    "loss": 0.32263126969337463,
    "learning_rate": 0.0001722999145441034,
    "timestep_mean": 513.03125,
    "epsilon_norm": 7.515787601470947,
    "prediction_norm": 6.327818393707275,
    "timestamp": "2025-07-14T12:34:25.386449"
  },
  {
    "epoch": 1455,
    "batch_idx": 192,
    "loss": 0.29832690954208374,
    "learning_rate": 0.0001722999145441034,
    "timestep_mean": 411.625,
    "epsilon_norm": 3.929725408554077,
    "prediction_norm": 3.0527961254119873,
    "timestamp": "2025-07-14T12:34:25.391536"
  },
  {
    "epoch": 1456,
    "batch_idx": 0,
    "loss": 0.16626445949077606,
    "learning_rate": 0.00017170712212352215,
    "timestep_mean": 473.6875,
    "epsilon_norm": 7.846739768981934,
    "prediction_norm": 7.211182594299316,
    "timestamp": "2025-07-14T12:34:25.396612"
  },
  {
    "epoch": 1456,
    "batch_idx": 32,
    "loss": 0.24479952454566956,
    "learning_rate": 0.00017170712212352215,
    "timestep_mean": 477.09375,
    "epsilon_norm": 6.9291181564331055,
    "prediction_norm": 6.313954830169678,
    "timestamp": "2025-07-14T12:34:25.401711"
  },
  {
    "epoch": 1456,
    "batch_idx": 64,
    "loss": 0.2817496061325073,
    "learning_rate": 0.00017170712212352215,
    "timestep_mean": 556.21875,
    "epsilon_norm": 9.125593185424805,
    "prediction_norm": 8.0348539352417,
    "timestamp": "2025-07-14T12:34:25.406765"
  },
  {
    "epoch": 1456,
    "batch_idx": 96,
    "loss": 0.3298989534378052,
    "learning_rate": 0.00017170712212352215,
    "timestep_mean": 475.71875,
    "epsilon_norm": 8.070818901062012,
    "prediction_norm": 6.602224826812744,
    "timestamp": "2025-07-14T12:34:25.411614"
  },
  {
    "epoch": 1456,
    "batch_idx": 128,
    "loss": 0.22170954942703247,
    "learning_rate": 0.00017170712212352215,
    "timestep_mean": 510.28125,
    "epsilon_norm": 8.045952796936035,
    "prediction_norm": 7.200446605682373,
    "timestamp": "2025-07-14T12:34:25.417023"
  },
  {
    "epoch": 1456,
    "batch_idx": 160,
    "loss": 0.16006909310817719,
    "learning_rate": 0.00017170712212352215,
    "timestep_mean": 518.53125,
    "epsilon_norm": 6.200199127197266,
    "prediction_norm": 5.821207523345947,
    "timestamp": "2025-07-14T12:34:25.422084"
  },
  {
    "epoch": 1456,
    "batch_idx": 192,
    "loss": 0.23275041580200195,
    "learning_rate": 0.00017170712212352215,
    "timestep_mean": 602.125,
    "epsilon_norm": 4.2645158767700195,
    "prediction_norm": 3.9091336727142334,
    "timestamp": "2025-07-14T12:34:25.427316"
  },
  {
    "epoch": 1457,
    "batch_idx": 0,
    "loss": 0.18267302215099335,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 560.25,
    "epsilon_norm": 7.9859724044799805,
    "prediction_norm": 6.98153018951416,
    "timestamp": "2025-07-14T12:34:25.432577"
  },
  {
    "epoch": 1457,
    "batch_idx": 32,
    "loss": 0.3623737394809723,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 471.4375,
    "epsilon_norm": 7.976277828216553,
    "prediction_norm": 6.461740970611572,
    "timestamp": "2025-07-14T12:34:25.437727"
  },
  {
    "epoch": 1457,
    "batch_idx": 64,
    "loss": 0.2859811782836914,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 498.875,
    "epsilon_norm": 7.991397857666016,
    "prediction_norm": 6.541152000427246,
    "timestamp": "2025-07-14T12:34:25.442721"
  },
  {
    "epoch": 1457,
    "batch_idx": 96,
    "loss": 0.46285200119018555,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 455.0625,
    "epsilon_norm": 7.830168724060059,
    "prediction_norm": 6.326055526733398,
    "timestamp": "2025-07-14T12:34:25.447937"
  },
  {
    "epoch": 1457,
    "batch_idx": 128,
    "loss": 0.38725724816322327,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 538.09375,
    "epsilon_norm": 8.163421630859375,
    "prediction_norm": 7.4857497215271,
    "timestamp": "2025-07-14T12:34:25.453134"
  },
  {
    "epoch": 1457,
    "batch_idx": 160,
    "loss": 0.3710899353027344,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 423.6875,
    "epsilon_norm": 8.145146369934082,
    "prediction_norm": 6.611009120941162,
    "timestamp": "2025-07-14T12:34:25.458405"
  },
  {
    "epoch": 1457,
    "batch_idx": 192,
    "loss": 0.09301469475030899,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 567.75,
    "epsilon_norm": 3.2862186431884766,
    "prediction_norm": 3.3290977478027344,
    "timestamp": "2025-07-14T12:34:25.463641"
  },
  {
    "epoch": 1458,
    "batch_idx": 0,
    "loss": 0.2814664840698242,
    "learning_rate": 0.00017052396883314197,
    "timestep_mean": 514.65625,
    "epsilon_norm": 8.895472526550293,
    "prediction_norm": 7.536461353302002,
    "timestamp": "2025-07-14T12:34:25.468855"
  },
  {
    "epoch": 1458,
    "batch_idx": 32,
    "loss": 0.24026188254356384,
    "learning_rate": 0.00017052396883314197,
    "timestep_mean": 437.34375,
    "epsilon_norm": 7.38199520111084,
    "prediction_norm": 6.7233757972717285,
    "timestamp": "2025-07-14T12:34:25.474061"
  },
  {
    "epoch": 1458,
    "batch_idx": 64,
    "loss": 0.4527985155582428,
    "learning_rate": 0.00017052396883314197,
    "timestep_mean": 438.78125,
    "epsilon_norm": 9.176421165466309,
    "prediction_norm": 7.680559158325195,
    "timestamp": "2025-07-14T12:34:25.479252"
  },
  {
    "epoch": 1458,
    "batch_idx": 96,
    "loss": 0.24951985478401184,
    "learning_rate": 0.00017052396883314197,
    "timestep_mean": 472.03125,
    "epsilon_norm": 8.627910614013672,
    "prediction_norm": 7.452329158782959,
    "timestamp": "2025-07-14T12:34:25.484476"
  },
  {
    "epoch": 1458,
    "batch_idx": 128,
    "loss": 0.45138978958129883,
    "learning_rate": 0.00017052396883314197,
    "timestep_mean": 468.375,
    "epsilon_norm": 9.625733375549316,
    "prediction_norm": 7.331610679626465,
    "timestamp": "2025-07-14T12:34:25.489649"
  },
  {
    "epoch": 1458,
    "batch_idx": 160,
    "loss": 0.3947375416755676,
    "learning_rate": 0.00017052396883314197,
    "timestep_mean": 472.59375,
    "epsilon_norm": 7.876626491546631,
    "prediction_norm": 6.805659294128418,
    "timestamp": "2025-07-14T12:34:25.494820"
  },
  {
    "epoch": 1458,
    "batch_idx": 192,
    "loss": 0.1858157217502594,
    "learning_rate": 0.00017052396883314197,
    "timestep_mean": 687.125,
    "epsilon_norm": 2.8108513355255127,
    "prediction_norm": 2.7806026935577393,
    "timestamp": "2025-07-14T12:34:25.500061"
  },
  {
    "epoch": 1459,
    "batch_idx": 0,
    "loss": 0.29605239629745483,
    "learning_rate": 0.00016993361088265603,
    "timestep_mean": 483.65625,
    "epsilon_norm": 8.616455078125,
    "prediction_norm": 6.865511894226074,
    "timestamp": "2025-07-14T12:34:25.505270"
  },
  {
    "epoch": 1459,
    "batch_idx": 32,
    "loss": 0.23328988254070282,
    "learning_rate": 0.00016993361088265603,
    "timestep_mean": 482.6875,
    "epsilon_norm": 6.629000186920166,
    "prediction_norm": 6.029916763305664,
    "timestamp": "2025-07-14T12:34:25.510528"
  },
  {
    "epoch": 1459,
    "batch_idx": 64,
    "loss": 0.22278685867786407,
    "learning_rate": 0.00016993361088265603,
    "timestep_mean": 518.625,
    "epsilon_norm": 8.969610214233398,
    "prediction_norm": 7.8828935623168945,
    "timestamp": "2025-07-14T12:34:25.515761"
  },
  {
    "epoch": 1459,
    "batch_idx": 96,
    "loss": 0.2336072325706482,
    "learning_rate": 0.00016993361088265603,
    "timestep_mean": 507.21875,
    "epsilon_norm": 8.042346954345703,
    "prediction_norm": 6.957798480987549,
    "timestamp": "2025-07-14T12:34:25.520845"
  },
  {
    "epoch": 1459,
    "batch_idx": 128,
    "loss": 0.376885324716568,
    "learning_rate": 0.00016993361088265603,
    "timestep_mean": 447.03125,
    "epsilon_norm": 8.447526931762695,
    "prediction_norm": 6.953397274017334,
    "timestamp": "2025-07-14T12:34:25.526000"
  },
  {
    "epoch": 1459,
    "batch_idx": 160,
    "loss": 0.20145899057388306,
    "learning_rate": 0.00016993361088265603,
    "timestep_mean": 482.0625,
    "epsilon_norm": 7.472924709320068,
    "prediction_norm": 6.7288498878479,
    "timestamp": "2025-07-14T12:34:25.531117"
  },
  {
    "epoch": 1459,
    "batch_idx": 192,
    "loss": 0.16515758633613586,
    "learning_rate": 0.00016993361088265603,
    "timestep_mean": 575.125,
    "epsilon_norm": 3.6383819580078125,
    "prediction_norm": 2.938295841217041,
    "timestamp": "2025-07-14T12:34:25.536306"
  },
  {
    "epoch": 1460,
    "batch_idx": 0,
    "loss": 0.24906253814697266,
    "learning_rate": 0.00016934406733817427,
    "timestep_mean": 457.25,
    "epsilon_norm": 8.304948806762695,
    "prediction_norm": 6.814962387084961,
    "timestamp": "2025-07-14T12:34:25.541487"
  },
  {
    "epoch": 1460,
    "batch_idx": 32,
    "loss": 0.5533660054206848,
    "learning_rate": 0.00016934406733817427,
    "timestep_mean": 374.03125,
    "epsilon_norm": 9.308541297912598,
    "prediction_norm": 7.331763744354248,
    "timestamp": "2025-07-14T12:34:25.546709"
  },
  {
    "epoch": 1460,
    "batch_idx": 64,
    "loss": 0.20871511101722717,
    "learning_rate": 0.00016934406733817427,
    "timestep_mean": 542.0625,
    "epsilon_norm": 8.192106246948242,
    "prediction_norm": 7.164608955383301,
    "timestamp": "2025-07-14T12:34:25.551912"
  },
  {
    "epoch": 1460,
    "batch_idx": 96,
    "loss": 0.21042034029960632,
    "learning_rate": 0.00016934406733817427,
    "timestep_mean": 558.8125,
    "epsilon_norm": 8.181500434875488,
    "prediction_norm": 7.209050178527832,
    "timestamp": "2025-07-14T12:34:25.556891"
  },
  {
    "epoch": 1460,
    "batch_idx": 128,
    "loss": 0.31796714663505554,
    "learning_rate": 0.00016934406733817427,
    "timestep_mean": 413.75,
    "epsilon_norm": 8.274057388305664,
    "prediction_norm": 6.603517532348633,
    "timestamp": "2025-07-14T12:34:25.561995"
  },
  {
    "epoch": 1460,
    "batch_idx": 160,
    "loss": 0.24110549688339233,
    "learning_rate": 0.00016934406733817427,
    "timestep_mean": 511.09375,
    "epsilon_norm": 8.811834335327148,
    "prediction_norm": 7.14042329788208,
    "timestamp": "2025-07-14T12:34:25.567261"
  },
  {
    "epoch": 1460,
    "batch_idx": 192,
    "loss": 0.13355548679828644,
    "learning_rate": 0.00016934406733817427,
    "timestep_mean": 636.125,
    "epsilon_norm": 3.804690361022949,
    "prediction_norm": 3.2025349140167236,
    "timestamp": "2025-07-14T12:34:25.572280"
  },
  {
    "epoch": 1461,
    "batch_idx": 0,
    "loss": 0.3056792914867401,
    "learning_rate": 0.0001687553396543371,
    "timestep_mean": 515.53125,
    "epsilon_norm": 7.802237510681152,
    "prediction_norm": 6.184779167175293,
    "timestamp": "2025-07-14T12:34:25.577454"
  },
  {
    "epoch": 1461,
    "batch_idx": 32,
    "loss": 0.3204303979873657,
    "learning_rate": 0.0001687553396543371,
    "timestep_mean": 470.90625,
    "epsilon_norm": 7.585256099700928,
    "prediction_norm": 6.014937400817871,
    "timestamp": "2025-07-14T12:34:25.582632"
  },
  {
    "epoch": 1461,
    "batch_idx": 64,
    "loss": 0.21624426543712616,
    "learning_rate": 0.0001687553396543371,
    "timestep_mean": 525.84375,
    "epsilon_norm": 7.606109142303467,
    "prediction_norm": 6.0852227210998535,
    "timestamp": "2025-07-14T12:34:25.587788"
  },
  {
    "epoch": 1461,
    "batch_idx": 96,
    "loss": 0.15622927248477936,
    "learning_rate": 0.0001687553396543371,
    "timestep_mean": 524.75,
    "epsilon_norm": 7.5951924324035645,
    "prediction_norm": 7.565272331237793,
    "timestamp": "2025-07-14T12:34:25.592754"
  },
  {
    "epoch": 1461,
    "batch_idx": 128,
    "loss": 0.15917649865150452,
    "learning_rate": 0.0001687553396543371,
    "timestep_mean": 527.8125,
    "epsilon_norm": 7.4715256690979,
    "prediction_norm": 6.591642379760742,
    "timestamp": "2025-07-14T12:34:25.597757"
  },
  {
    "epoch": 1461,
    "batch_idx": 160,
    "loss": 0.23893174529075623,
    "learning_rate": 0.0001687553396543371,
    "timestep_mean": 502.53125,
    "epsilon_norm": 7.775184631347656,
    "prediction_norm": 6.531463623046875,
    "timestamp": "2025-07-14T12:34:25.602993"
  },
  {
    "epoch": 1461,
    "batch_idx": 192,
    "loss": 0.11725044250488281,
    "learning_rate": 0.0001687553396543371,
    "timestep_mean": 386.875,
    "epsilon_norm": 4.052021026611328,
    "prediction_norm": 3.5866706371307373,
    "timestamp": "2025-07-14T12:34:25.608059"
  },
  {
    "epoch": 1462,
    "batch_idx": 0,
    "loss": 0.31035107374191284,
    "learning_rate": 0.000168167429283771,
    "timestep_mean": 477.9375,
    "epsilon_norm": 7.819002628326416,
    "prediction_norm": 6.122165679931641,
    "timestamp": "2025-07-14T12:34:25.613129"
  },
  {
    "epoch": 1462,
    "batch_idx": 32,
    "loss": 0.25801724195480347,
    "learning_rate": 0.000168167429283771,
    "timestep_mean": 520.90625,
    "epsilon_norm": 8.905126571655273,
    "prediction_norm": 7.767683506011963,
    "timestamp": "2025-07-14T12:34:25.618250"
  },
  {
    "epoch": 1462,
    "batch_idx": 64,
    "loss": 0.2617042064666748,
    "learning_rate": 0.000168167429283771,
    "timestep_mean": 544.75,
    "epsilon_norm": 8.403467178344727,
    "prediction_norm": 6.642166614532471,
    "timestamp": "2025-07-14T12:34:25.623398"
  },
  {
    "epoch": 1462,
    "batch_idx": 96,
    "loss": 0.22849474847316742,
    "learning_rate": 0.000168167429283771,
    "timestep_mean": 507.3125,
    "epsilon_norm": 7.989498615264893,
    "prediction_norm": 6.680331230163574,
    "timestamp": "2025-07-14T12:34:25.628632"
  },
  {
    "epoch": 1462,
    "batch_idx": 128,
    "loss": 0.23995116353034973,
    "learning_rate": 0.000168167429283771,
    "timestep_mean": 601.40625,
    "epsilon_norm": 7.379052639007568,
    "prediction_norm": 6.743276119232178,
    "timestamp": "2025-07-14T12:34:25.633885"
  },
  {
    "epoch": 1462,
    "batch_idx": 160,
    "loss": 0.2363346517086029,
    "learning_rate": 0.000168167429283771,
    "timestep_mean": 530.28125,
    "epsilon_norm": 7.140133857727051,
    "prediction_norm": 6.871919631958008,
    "timestamp": "2025-07-14T12:34:25.638967"
  },
  {
    "epoch": 1462,
    "batch_idx": 192,
    "loss": 0.6590574383735657,
    "learning_rate": 0.000168167429283771,
    "timestep_mean": 332.25,
    "epsilon_norm": 3.865100383758545,
    "prediction_norm": 2.9764249324798584,
    "timestamp": "2025-07-14T12:34:25.644075"
  },
  {
    "epoch": 1463,
    "batch_idx": 0,
    "loss": 0.19151705503463745,
    "learning_rate": 0.0001675803376770868,
    "timestep_mean": 492.34375,
    "epsilon_norm": 7.762818336486816,
    "prediction_norm": 6.668369293212891,
    "timestamp": "2025-07-14T12:34:25.649204"
  },
  {
    "epoch": 1463,
    "batch_idx": 32,
    "loss": 0.1785935014486313,
    "learning_rate": 0.0001675803376770868,
    "timestep_mean": 543.34375,
    "epsilon_norm": 8.10947322845459,
    "prediction_norm": 8.1768217086792,
    "timestamp": "2025-07-14T12:34:25.654351"
  },
  {
    "epoch": 1463,
    "batch_idx": 64,
    "loss": 0.21001428365707397,
    "learning_rate": 0.0001675803376770868,
    "timestep_mean": 535.90625,
    "epsilon_norm": 7.834733486175537,
    "prediction_norm": 7.2231316566467285,
    "timestamp": "2025-07-14T12:34:25.659543"
  },
  {
    "epoch": 1463,
    "batch_idx": 96,
    "loss": 0.24202056229114532,
    "learning_rate": 0.0001675803376770868,
    "timestep_mean": 489.6875,
    "epsilon_norm": 8.619607925415039,
    "prediction_norm": 8.09853458404541,
    "timestamp": "2025-07-14T12:34:25.664532"
  },
  {
    "epoch": 1463,
    "batch_idx": 128,
    "loss": 0.28841882944107056,
    "learning_rate": 0.0001675803376770868,
    "timestep_mean": 489.03125,
    "epsilon_norm": 7.820581912994385,
    "prediction_norm": 6.363523006439209,
    "timestamp": "2025-07-14T12:34:25.669395"
  },
  {
    "epoch": 1463,
    "batch_idx": 160,
    "loss": 0.3213961720466614,
    "learning_rate": 0.0001675803376770868,
    "timestep_mean": 453.96875,
    "epsilon_norm": 7.645534992218018,
    "prediction_norm": 5.938955783843994,
    "timestamp": "2025-07-14T12:34:25.674193"
  },
  {
    "epoch": 1463,
    "batch_idx": 192,
    "loss": 0.6417359113693237,
    "learning_rate": 0.0001675803376770868,
    "timestep_mean": 481.25,
    "epsilon_norm": 5.104644298553467,
    "prediction_norm": 3.6164863109588623,
    "timestamp": "2025-07-14T12:34:25.679010"
  },
  {
    "epoch": 1464,
    "batch_idx": 0,
    "loss": 0.2711459994316101,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 563.78125,
    "epsilon_norm": 7.268749237060547,
    "prediction_norm": 6.550018310546875,
    "timestamp": "2025-07-14T12:34:25.684017"
  },
  {
    "epoch": 1464,
    "batch_idx": 32,
    "loss": 0.2654957175254822,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 509.75,
    "epsilon_norm": 9.039334297180176,
    "prediction_norm": 7.909826755523682,
    "timestamp": "2025-07-14T12:34:25.688963"
  },
  {
    "epoch": 1464,
    "batch_idx": 64,
    "loss": 0.2852368950843811,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 484.3125,
    "epsilon_norm": 6.170072555541992,
    "prediction_norm": 4.843475341796875,
    "timestamp": "2025-07-14T12:34:25.693869"
  },
  {
    "epoch": 1464,
    "batch_idx": 96,
    "loss": 0.40154358744621277,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 372.375,
    "epsilon_norm": 7.733377456665039,
    "prediction_norm": 6.379084587097168,
    "timestamp": "2025-07-14T12:34:25.699202"
  },
  {
    "epoch": 1464,
    "batch_idx": 128,
    "loss": 0.34842389822006226,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 485.4375,
    "epsilon_norm": 8.034183502197266,
    "prediction_norm": 6.521531105041504,
    "timestamp": "2025-07-14T12:34:25.704563"
  },
  {
    "epoch": 1464,
    "batch_idx": 160,
    "loss": 0.26965755224227905,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 526.59375,
    "epsilon_norm": 8.285799026489258,
    "prediction_norm": 6.991461277008057,
    "timestamp": "2025-07-14T12:34:25.709675"
  },
  {
    "epoch": 1464,
    "batch_idx": 192,
    "loss": 0.06211717426776886,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 539.5,
    "epsilon_norm": 3.5885908603668213,
    "prediction_norm": 3.2222115993499756,
    "timestamp": "2025-07-14T12:34:25.714800"
  },
  {
    "epoch": 1465,
    "batch_idx": 0,
    "loss": 0.15590113401412964,
    "learning_rate": 0.00016640861654770037,
    "timestep_mean": 556.125,
    "epsilon_norm": 6.685939311981201,
    "prediction_norm": 6.290310859680176,
    "timestamp": "2025-07-14T12:34:25.719844"
  },
  {
    "epoch": 1465,
    "batch_idx": 32,
    "loss": 0.16161136329174042,
    "learning_rate": 0.00016640861654770037,
    "timestep_mean": 566.5625,
    "epsilon_norm": 8.205520629882812,
    "prediction_norm": 7.360753059387207,
    "timestamp": "2025-07-14T12:34:25.724761"
  },
  {
    "epoch": 1465,
    "batch_idx": 64,
    "loss": 0.21379967033863068,
    "learning_rate": 0.00016640861654770037,
    "timestep_mean": 532.0625,
    "epsilon_norm": 8.188272476196289,
    "prediction_norm": 6.739383220672607,
    "timestamp": "2025-07-14T12:34:25.729638"
  },
  {
    "epoch": 1465,
    "batch_idx": 96,
    "loss": 0.26693835854530334,
    "learning_rate": 0.00016640861654770037,
    "timestep_mean": 473.96875,
    "epsilon_norm": 8.979735374450684,
    "prediction_norm": 7.515336036682129,
    "timestamp": "2025-07-14T12:34:25.734485"
  },
  {
    "epoch": 1465,
    "batch_idx": 128,
    "loss": 0.2675950527191162,
    "learning_rate": 0.00016640861654770037,
    "timestep_mean": 509.0,
    "epsilon_norm": 7.459859848022461,
    "prediction_norm": 5.929751396179199,
    "timestamp": "2025-07-14T12:34:25.739299"
  },
  {
    "epoch": 1465,
    "batch_idx": 160,
    "loss": 0.3972095549106598,
    "learning_rate": 0.00016640861654770037,
    "timestep_mean": 442.5625,
    "epsilon_norm": 7.428234100341797,
    "prediction_norm": 5.872091293334961,
    "timestamp": "2025-07-14T12:34:25.744253"
  },
  {
    "epoch": 1465,
    "batch_idx": 192,
    "loss": 0.1346609890460968,
    "learning_rate": 0.00016640861654770037,
    "timestep_mean": 494.125,
    "epsilon_norm": 3.653294563293457,
    "prediction_norm": 3.1021018028259277,
    "timestamp": "2025-07-14T12:34:25.749088"
  },
  {
    "epoch": 1466,
    "batch_idx": 0,
    "loss": 0.35629451274871826,
    "learning_rate": 0.0001658239899161039,
    "timestep_mean": 610.15625,
    "epsilon_norm": 7.21938943862915,
    "prediction_norm": 6.133548259735107,
    "timestamp": "2025-07-14T12:34:25.753872"
  },
  {
    "epoch": 1466,
    "batch_idx": 32,
    "loss": 0.22685083746910095,
    "learning_rate": 0.0001658239899161039,
    "timestep_mean": 532.71875,
    "epsilon_norm": 8.205138206481934,
    "prediction_norm": 6.560584545135498,
    "timestamp": "2025-07-14T12:34:25.758838"
  },
  {
    "epoch": 1466,
    "batch_idx": 64,
    "loss": 0.2079140543937683,
    "learning_rate": 0.0001658239899161039,
    "timestep_mean": 507.28125,
    "epsilon_norm": 8.19677448272705,
    "prediction_norm": 7.423080921173096,
    "timestamp": "2025-07-14T12:34:25.763666"
  },
  {
    "epoch": 1466,
    "batch_idx": 96,
    "loss": 0.19938458502292633,
    "learning_rate": 0.0001658239899161039,
    "timestep_mean": 479.125,
    "epsilon_norm": 6.165131568908691,
    "prediction_norm": 5.190910816192627,
    "timestamp": "2025-07-14T12:34:25.768494"
  },
  {
    "epoch": 1466,
    "batch_idx": 128,
    "loss": 0.35767367482185364,
    "learning_rate": 0.0001658239899161039,
    "timestep_mean": 483.96875,
    "epsilon_norm": 8.095216751098633,
    "prediction_norm": 6.461062908172607,
    "timestamp": "2025-07-14T12:34:25.773400"
  },
  {
    "epoch": 1466,
    "batch_idx": 160,
    "loss": 0.3782320022583008,
    "learning_rate": 0.0001658239899161039,
    "timestep_mean": 487.21875,
    "epsilon_norm": 8.314068794250488,
    "prediction_norm": 7.056155204772949,
    "timestamp": "2025-07-14T12:34:25.778293"
  },
  {
    "epoch": 1466,
    "batch_idx": 192,
    "loss": 0.6388376355171204,
    "learning_rate": 0.0001658239899161039,
    "timestep_mean": 467.625,
    "epsilon_norm": 4.063014507293701,
    "prediction_norm": 3.175330400466919,
    "timestamp": "2025-07-14T12:34:25.783250"
  },
  {
    "epoch": 1467,
    "batch_idx": 0,
    "loss": 0.3133091330528259,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 501.03125,
    "epsilon_norm": 8.55986213684082,
    "prediction_norm": 7.082128524780273,
    "timestamp": "2025-07-14T12:34:25.788409"
  },
  {
    "epoch": 1467,
    "batch_idx": 32,
    "loss": 0.4209060072898865,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 440.46875,
    "epsilon_norm": 8.00894546508789,
    "prediction_norm": 5.832962512969971,
    "timestamp": "2025-07-14T12:34:25.793257"
  },
  {
    "epoch": 1467,
    "batch_idx": 64,
    "loss": 0.431935578584671,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 374.90625,
    "epsilon_norm": 7.7640838623046875,
    "prediction_norm": 6.010318279266357,
    "timestamp": "2025-07-14T12:34:25.798144"
  },
  {
    "epoch": 1467,
    "batch_idx": 96,
    "loss": 0.32833659648895264,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 480.5625,
    "epsilon_norm": 7.378631114959717,
    "prediction_norm": 6.419288158416748,
    "timestamp": "2025-07-14T12:34:25.803075"
  },
  {
    "epoch": 1467,
    "batch_idx": 128,
    "loss": 0.3804171681404114,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 447.25,
    "epsilon_norm": 8.856620788574219,
    "prediction_norm": 6.757726669311523,
    "timestamp": "2025-07-14T12:34:25.808168"
  },
  {
    "epoch": 1467,
    "batch_idx": 160,
    "loss": 0.2453879714012146,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 553.34375,
    "epsilon_norm": 6.812159061431885,
    "prediction_norm": 6.275782585144043,
    "timestamp": "2025-07-14T12:34:25.813210"
  },
  {
    "epoch": 1467,
    "batch_idx": 192,
    "loss": 0.09278009086847305,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 516.25,
    "epsilon_norm": 2.443871259689331,
    "prediction_norm": 2.4935455322265625,
    "timestamp": "2025-07-14T12:34:25.818007"
  },
  {
    "epoch": 1468,
    "batch_idx": 0,
    "loss": 0.2469015121459961,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 468.1875,
    "epsilon_norm": 7.462454319000244,
    "prediction_norm": 6.459052562713623,
    "timestamp": "2025-07-14T12:34:25.822893"
  },
  {
    "epoch": 1468,
    "batch_idx": 32,
    "loss": 0.19102896749973297,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 465.0625,
    "epsilon_norm": 7.457984924316406,
    "prediction_norm": 6.441164016723633,
    "timestamp": "2025-07-14T12:34:25.827691"
  },
  {
    "epoch": 1468,
    "batch_idx": 64,
    "loss": 0.28677791357040405,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 487.96875,
    "epsilon_norm": 7.73419713973999,
    "prediction_norm": 6.028667449951172,
    "timestamp": "2025-07-14T12:34:25.832656"
  },
  {
    "epoch": 1468,
    "batch_idx": 96,
    "loss": 0.2938570976257324,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 407.5625,
    "epsilon_norm": 7.6617817878723145,
    "prediction_norm": 6.040935516357422,
    "timestamp": "2025-07-14T12:34:25.837501"
  },
  {
    "epoch": 1468,
    "batch_idx": 128,
    "loss": 0.23884239792823792,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 496.71875,
    "epsilon_norm": 7.685605525970459,
    "prediction_norm": 6.614805221557617,
    "timestamp": "2025-07-14T12:34:25.842339"
  },
  {
    "epoch": 1468,
    "batch_idx": 160,
    "loss": 0.3272849917411804,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 480.15625,
    "epsilon_norm": 7.570793628692627,
    "prediction_norm": 6.691824913024902,
    "timestamp": "2025-07-14T12:34:25.847174"
  },
  {
    "epoch": 1468,
    "batch_idx": 192,
    "loss": 0.20399439334869385,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 591.875,
    "epsilon_norm": 4.183128833770752,
    "prediction_norm": 3.6253700256347656,
    "timestamp": "2025-07-14T12:34:25.852010"
  },
  {
    "epoch": 1469,
    "batch_idx": 0,
    "loss": 0.22566862404346466,
    "learning_rate": 0.00016407506305768276,
    "timestep_mean": 403.71875,
    "epsilon_norm": 7.690411567687988,
    "prediction_norm": 6.432894229888916,
    "timestamp": "2025-07-14T12:34:25.856842"
  },
  {
    "epoch": 1469,
    "batch_idx": 32,
    "loss": 0.26945051550865173,
    "learning_rate": 0.00016407506305768276,
    "timestep_mean": 526.71875,
    "epsilon_norm": 6.549227237701416,
    "prediction_norm": 5.860096454620361,
    "timestamp": "2025-07-14T12:34:25.861675"
  },
  {
    "epoch": 1469,
    "batch_idx": 64,
    "loss": 0.3531898856163025,
    "learning_rate": 0.00016407506305768276,
    "timestep_mean": 464.5625,
    "epsilon_norm": 6.969118595123291,
    "prediction_norm": 6.112877368927002,
    "timestamp": "2025-07-14T12:34:25.866489"
  },
  {
    "epoch": 1469,
    "batch_idx": 96,
    "loss": 0.20801526308059692,
    "learning_rate": 0.00016407506305768276,
    "timestep_mean": 542.0625,
    "epsilon_norm": 6.927857875823975,
    "prediction_norm": 5.938989639282227,
    "timestamp": "2025-07-14T12:34:25.871474"
  },
  {
    "epoch": 1469,
    "batch_idx": 128,
    "loss": 0.2503533959388733,
    "learning_rate": 0.00016407506305768276,
    "timestep_mean": 517.0625,
    "epsilon_norm": 7.590367794036865,
    "prediction_norm": 6.174474716186523,
    "timestamp": "2025-07-14T12:34:25.876245"
  },
  {
    "epoch": 1469,
    "batch_idx": 160,
    "loss": 0.23982354998588562,
    "learning_rate": 0.00016407506305768276,
    "timestep_mean": 556.8125,
    "epsilon_norm": 6.755923271179199,
    "prediction_norm": 5.180154323577881,
    "timestamp": "2025-07-14T12:34:25.881058"
  },
  {
    "epoch": 1469,
    "batch_idx": 192,
    "loss": 0.24045026302337646,
    "learning_rate": 0.00016407506305768276,
    "timestep_mean": 576.125,
    "epsilon_norm": 4.315356731414795,
    "prediction_norm": 3.4132261276245117,
    "timestamp": "2025-07-14T12:34:25.885907"
  },
  {
    "epoch": 1470,
    "batch_idx": 0,
    "loss": 0.12758009135723114,
    "learning_rate": 0.00016349374324511377,
    "timestep_mean": 479.0,
    "epsilon_norm": 6.921685218811035,
    "prediction_norm": 6.014142036437988,
    "timestamp": "2025-07-14T12:34:25.890754"
  },
  {
    "epoch": 1470,
    "batch_idx": 32,
    "loss": 0.26613113284111023,
    "learning_rate": 0.00016349374324511377,
    "timestep_mean": 531.78125,
    "epsilon_norm": 7.105068206787109,
    "prediction_norm": 5.871885299682617,
    "timestamp": "2025-07-14T12:34:25.895631"
  },
  {
    "epoch": 1470,
    "batch_idx": 64,
    "loss": 0.35692280530929565,
    "learning_rate": 0.00016349374324511377,
    "timestep_mean": 456.875,
    "epsilon_norm": 8.695857048034668,
    "prediction_norm": 6.316778659820557,
    "timestamp": "2025-07-14T12:34:25.900901"
  },
  {
    "epoch": 1470,
    "batch_idx": 96,
    "loss": 0.14416863024234772,
    "learning_rate": 0.00016349374324511377,
    "timestep_mean": 534.5625,
    "epsilon_norm": 7.785964012145996,
    "prediction_norm": 7.059303283691406,
    "timestamp": "2025-07-14T12:34:25.905892"
  },
  {
    "epoch": 1470,
    "batch_idx": 128,
    "loss": 0.21530380845069885,
    "learning_rate": 0.00016349374324511377,
    "timestep_mean": 559.09375,
    "epsilon_norm": 8.520825386047363,
    "prediction_norm": 7.005462646484375,
    "timestamp": "2025-07-14T12:34:25.911041"
  },
  {
    "epoch": 1470,
    "batch_idx": 160,
    "loss": 0.1844644844532013,
    "learning_rate": 0.00016349374324511377,
    "timestep_mean": 575.65625,
    "epsilon_norm": 7.643184661865234,
    "prediction_norm": 6.859059810638428,
    "timestamp": "2025-07-14T12:34:25.916240"
  },
  {
    "epoch": 1470,
    "batch_idx": 192,
    "loss": 0.25322428345680237,
    "learning_rate": 0.00016349374324511377,
    "timestep_mean": 564.375,
    "epsilon_norm": 4.970132350921631,
    "prediction_norm": 3.892584800720215,
    "timestamp": "2025-07-14T12:34:25.921382"
  },
  {
    "epoch": 1471,
    "batch_idx": 0,
    "loss": 0.14569436013698578,
    "learning_rate": 0.000162913253728282,
    "timestep_mean": 511.5,
    "epsilon_norm": 7.3452653884887695,
    "prediction_norm": 6.361258029937744,
    "timestamp": "2025-07-14T12:34:25.926513"
  },
  {
    "epoch": 1471,
    "batch_idx": 32,
    "loss": 0.25390321016311646,
    "learning_rate": 0.000162913253728282,
    "timestep_mean": 478.15625,
    "epsilon_norm": 8.120796203613281,
    "prediction_norm": 6.813797950744629,
    "timestamp": "2025-07-14T12:34:25.931566"
  },
  {
    "epoch": 1471,
    "batch_idx": 64,
    "loss": 0.2777990996837616,
    "learning_rate": 0.000162913253728282,
    "timestep_mean": 576.78125,
    "epsilon_norm": 8.153582572937012,
    "prediction_norm": 6.765452861785889,
    "timestamp": "2025-07-14T12:34:25.936619"
  },
  {
    "epoch": 1471,
    "batch_idx": 96,
    "loss": 0.3563039004802704,
    "learning_rate": 0.000162913253728282,
    "timestep_mean": 465.9375,
    "epsilon_norm": 7.684445381164551,
    "prediction_norm": 6.294765949249268,
    "timestamp": "2025-07-14T12:34:25.941830"
  },
  {
    "epoch": 1471,
    "batch_idx": 128,
    "loss": 0.14492399990558624,
    "learning_rate": 0.000162913253728282,
    "timestep_mean": 436.78125,
    "epsilon_norm": 8.61296272277832,
    "prediction_norm": 8.010758399963379,
    "timestamp": "2025-07-14T12:34:25.947012"
  },
  {
    "epoch": 1471,
    "batch_idx": 160,
    "loss": 0.18573233485221863,
    "learning_rate": 0.000162913253728282,
    "timestep_mean": 531.15625,
    "epsilon_norm": 8.939953804016113,
    "prediction_norm": 8.016988754272461,
    "timestamp": "2025-07-14T12:34:25.952094"
  },
  {
    "epoch": 1471,
    "batch_idx": 192,
    "loss": 0.8283277153968811,
    "learning_rate": 0.000162913253728282,
    "timestep_mean": 355.125,
    "epsilon_norm": 4.72060489654541,
    "prediction_norm": 3.591507911682129,
    "timestamp": "2025-07-14T12:34:25.957222"
  },
  {
    "epoch": 1472,
    "batch_idx": 0,
    "loss": 0.1849195659160614,
    "learning_rate": 0.00016233359593948807,
    "timestep_mean": 562.40625,
    "epsilon_norm": 7.147404670715332,
    "prediction_norm": 5.8603410720825195,
    "timestamp": "2025-07-14T12:34:25.962383"
  },
  {
    "epoch": 1472,
    "batch_idx": 32,
    "loss": 0.4404071271419525,
    "learning_rate": 0.00016233359593948807,
    "timestep_mean": 524.0625,
    "epsilon_norm": 6.610705375671387,
    "prediction_norm": 5.533920764923096,
    "timestamp": "2025-07-14T12:34:25.967701"
  },
  {
    "epoch": 1472,
    "batch_idx": 64,
    "loss": 0.32656151056289673,
    "learning_rate": 0.00016233359593948807,
    "timestep_mean": 402.15625,
    "epsilon_norm": 8.144922256469727,
    "prediction_norm": 6.687058925628662,
    "timestamp": "2025-07-14T12:34:25.972910"
  },
  {
    "epoch": 1472,
    "batch_idx": 96,
    "loss": 0.23074987530708313,
    "learning_rate": 0.00016233359593948807,
    "timestep_mean": 498.8125,
    "epsilon_norm": 8.102521896362305,
    "prediction_norm": 7.380465507507324,
    "timestamp": "2025-07-14T12:34:25.978174"
  },
  {
    "epoch": 1472,
    "batch_idx": 128,
    "loss": 0.28063201904296875,
    "learning_rate": 0.00016233359593948807,
    "timestep_mean": 560.875,
    "epsilon_norm": 8.682973861694336,
    "prediction_norm": 7.452782154083252,
    "timestamp": "2025-07-14T12:34:25.983457"
  },
  {
    "epoch": 1472,
    "batch_idx": 160,
    "loss": 0.28775346279144287,
    "learning_rate": 0.00016233359593948807,
    "timestep_mean": 474.15625,
    "epsilon_norm": 7.96812105178833,
    "prediction_norm": 7.028682708740234,
    "timestamp": "2025-07-14T12:34:25.988568"
  },
  {
    "epoch": 1472,
    "batch_idx": 192,
    "loss": 0.1697159707546234,
    "learning_rate": 0.00016233359593948807,
    "timestep_mean": 528.125,
    "epsilon_norm": 3.3452978134155273,
    "prediction_norm": 2.9562935829162598,
    "timestamp": "2025-07-14T12:34:25.993547"
  },
  {
    "epoch": 1473,
    "batch_idx": 0,
    "loss": 0.1778288632631302,
    "learning_rate": 0.00016175477130897972,
    "timestep_mean": 515.8125,
    "epsilon_norm": 7.942503929138184,
    "prediction_norm": 7.137179374694824,
    "timestamp": "2025-07-14T12:34:25.998586"
  },
  {
    "epoch": 1473,
    "batch_idx": 32,
    "loss": 0.24166086316108704,
    "learning_rate": 0.00016175477130897972,
    "timestep_mean": 459.21875,
    "epsilon_norm": 8.588372230529785,
    "prediction_norm": 7.034877300262451,
    "timestamp": "2025-07-14T12:34:26.003552"
  },
  {
    "epoch": 1473,
    "batch_idx": 64,
    "loss": 0.25455424189567566,
    "learning_rate": 0.00016175477130897972,
    "timestep_mean": 414.0,
    "epsilon_norm": 6.984495162963867,
    "prediction_norm": 5.947882175445557,
    "timestamp": "2025-07-14T12:34:26.008427"
  },
  {
    "epoch": 1473,
    "batch_idx": 96,
    "loss": 0.29772818088531494,
    "learning_rate": 0.00016175477130897972,
    "timestep_mean": 485.4375,
    "epsilon_norm": 8.58774185180664,
    "prediction_norm": 7.063937187194824,
    "timestamp": "2025-07-14T12:34:26.013212"
  },
  {
    "epoch": 1473,
    "batch_idx": 128,
    "loss": 0.23704078793525696,
    "learning_rate": 0.00016175477130897972,
    "timestep_mean": 548.3125,
    "epsilon_norm": 8.103742599487305,
    "prediction_norm": 7.411022186279297,
    "timestamp": "2025-07-14T12:34:26.017983"
  },
  {
    "epoch": 1473,
    "batch_idx": 160,
    "loss": 0.15480847656726837,
    "learning_rate": 0.00016175477130897972,
    "timestep_mean": 604.71875,
    "epsilon_norm": 8.692745208740234,
    "prediction_norm": 8.201288223266602,
    "timestamp": "2025-07-14T12:34:26.022772"
  },
  {
    "epoch": 1473,
    "batch_idx": 192,
    "loss": 0.4238361120223999,
    "learning_rate": 0.00016175477130897972,
    "timestep_mean": 400.375,
    "epsilon_norm": 4.811722278594971,
    "prediction_norm": 4.203495502471924,
    "timestamp": "2025-07-14T12:34:26.027606"
  },
  {
    "epoch": 1474,
    "batch_idx": 0,
    "loss": 0.2591477334499359,
    "learning_rate": 0.00016117678126494907,
    "timestep_mean": 508.0,
    "epsilon_norm": 7.900538921356201,
    "prediction_norm": 6.91710090637207,
    "timestamp": "2025-07-14T12:34:26.032454"
  },
  {
    "epoch": 1474,
    "batch_idx": 32,
    "loss": 0.28615283966064453,
    "learning_rate": 0.00016117678126494907,
    "timestep_mean": 499.03125,
    "epsilon_norm": 8.585648536682129,
    "prediction_norm": 8.212021827697754,
    "timestamp": "2025-07-14T12:34:26.037352"
  },
  {
    "epoch": 1474,
    "batch_idx": 64,
    "loss": 0.3068743944168091,
    "learning_rate": 0.00016117678126494907,
    "timestep_mean": 467.4375,
    "epsilon_norm": 7.710670471191406,
    "prediction_norm": 6.266761779785156,
    "timestamp": "2025-07-14T12:34:26.042352"
  },
  {
    "epoch": 1474,
    "batch_idx": 96,
    "loss": 0.3468988537788391,
    "learning_rate": 0.00016117678126494907,
    "timestep_mean": 504.65625,
    "epsilon_norm": 9.023944854736328,
    "prediction_norm": 7.712069988250732,
    "timestamp": "2025-07-14T12:34:26.047134"
  },
  {
    "epoch": 1474,
    "batch_idx": 128,
    "loss": 0.19626358151435852,
    "learning_rate": 0.00016117678126494907,
    "timestep_mean": 470.34375,
    "epsilon_norm": 9.100968360900879,
    "prediction_norm": 8.356439590454102,
    "timestamp": "2025-07-14T12:34:26.051951"
  },
  {
    "epoch": 1474,
    "batch_idx": 160,
    "loss": 0.1657380759716034,
    "learning_rate": 0.00016117678126494907,
    "timestep_mean": 512.5,
    "epsilon_norm": 7.56462287902832,
    "prediction_norm": 6.473022937774658,
    "timestamp": "2025-07-14T12:34:26.056765"
  },
  {
    "epoch": 1474,
    "batch_idx": 192,
    "loss": 0.07193915545940399,
    "learning_rate": 0.00016117678126494907,
    "timestep_mean": 335.5,
    "epsilon_norm": 2.7024006843566895,
    "prediction_norm": 2.725848436355591,
    "timestamp": "2025-07-14T12:34:26.061611"
  },
  {
    "epoch": 1475,
    "batch_idx": 0,
    "loss": 0.18047890067100525,
    "learning_rate": 0.0001605996272335295,
    "timestep_mean": 419.25,
    "epsilon_norm": 7.132315635681152,
    "prediction_norm": 6.358941555023193,
    "timestamp": "2025-07-14T12:34:26.066437"
  },
  {
    "epoch": 1475,
    "batch_idx": 32,
    "loss": 0.48599255084991455,
    "learning_rate": 0.0001605996272335295,
    "timestep_mean": 504.21875,
    "epsilon_norm": 7.76505708694458,
    "prediction_norm": 6.3342976570129395,
    "timestamp": "2025-07-14T12:34:26.071397"
  },
  {
    "epoch": 1475,
    "batch_idx": 64,
    "loss": 0.3365456461906433,
    "learning_rate": 0.0001605996272335295,
    "timestep_mean": 482.53125,
    "epsilon_norm": 7.629669189453125,
    "prediction_norm": 6.963161468505859,
    "timestamp": "2025-07-14T12:34:26.076213"
  },
  {
    "epoch": 1475,
    "batch_idx": 96,
    "loss": 0.4884607493877411,
    "learning_rate": 0.0001605996272335295,
    "timestep_mean": 443.40625,
    "epsilon_norm": 8.305450439453125,
    "prediction_norm": 6.095219612121582,
    "timestamp": "2025-07-14T12:34:26.081052"
  },
  {
    "epoch": 1475,
    "batch_idx": 128,
    "loss": 0.22881273925304413,
    "learning_rate": 0.0001605996272335295,
    "timestep_mean": 497.40625,
    "epsilon_norm": 8.74793529510498,
    "prediction_norm": 7.7722320556640625,
    "timestamp": "2025-07-14T12:34:26.085936"
  },
  {
    "epoch": 1475,
    "batch_idx": 160,
    "loss": 0.30808305740356445,
    "learning_rate": 0.0001605996272335295,
    "timestep_mean": 450.3125,
    "epsilon_norm": 7.014312267303467,
    "prediction_norm": 6.5877251625061035,
    "timestamp": "2025-07-14T12:34:26.090760"
  },
  {
    "epoch": 1475,
    "batch_idx": 192,
    "loss": 0.16554279625415802,
    "learning_rate": 0.0001605996272335295,
    "timestep_mean": 392.375,
    "epsilon_norm": 3.422028064727783,
    "prediction_norm": 3.3453996181488037,
    "timestamp": "2025-07-14T12:34:26.095576"
  },
  {
    "epoch": 1476,
    "batch_idx": 0,
    "loss": 0.24161314964294434,
    "learning_rate": 0.00016002331063879067,
    "timestep_mean": 493.84375,
    "epsilon_norm": 8.009570121765137,
    "prediction_norm": 6.876946926116943,
    "timestamp": "2025-07-14T12:34:26.100489"
  },
  {
    "epoch": 1476,
    "batch_idx": 32,
    "loss": 0.2296898066997528,
    "learning_rate": 0.00016002331063879067,
    "timestep_mean": 472.96875,
    "epsilon_norm": 6.801448822021484,
    "prediction_norm": 6.147471904754639,
    "timestamp": "2025-07-14T12:34:26.105347"
  },
  {
    "epoch": 1476,
    "batch_idx": 64,
    "loss": 0.3126681447029114,
    "learning_rate": 0.00016002331063879067,
    "timestep_mean": 483.53125,
    "epsilon_norm": 8.069100379943848,
    "prediction_norm": 7.012711048126221,
    "timestamp": "2025-07-14T12:34:26.110123"
  },
  {
    "epoch": 1476,
    "batch_idx": 96,
    "loss": 0.3897116184234619,
    "learning_rate": 0.00016002331063879067,
    "timestep_mean": 462.125,
    "epsilon_norm": 7.117684364318848,
    "prediction_norm": 5.336182117462158,
    "timestamp": "2025-07-14T12:34:26.114975"
  },
  {
    "epoch": 1476,
    "batch_idx": 128,
    "loss": 0.4542253911495209,
    "learning_rate": 0.00016002331063879067,
    "timestep_mean": 459.03125,
    "epsilon_norm": 8.182829856872559,
    "prediction_norm": 6.806489944458008,
    "timestamp": "2025-07-14T12:34:26.119801"
  },
  {
    "epoch": 1476,
    "batch_idx": 160,
    "loss": 0.26057857275009155,
    "learning_rate": 0.00016002331063879067,
    "timestep_mean": 462.96875,
    "epsilon_norm": 6.815927982330322,
    "prediction_norm": 5.0707197189331055,
    "timestamp": "2025-07-14T12:34:26.124622"
  },
  {
    "epoch": 1476,
    "batch_idx": 192,
    "loss": 0.059403207153081894,
    "learning_rate": 0.00016002331063879067,
    "timestep_mean": 617.75,
    "epsilon_norm": 3.087890625,
    "prediction_norm": 3.2488770484924316,
    "timestamp": "2025-07-14T12:34:26.129465"
  },
  {
    "epoch": 1477,
    "batch_idx": 0,
    "loss": 0.16680818796157837,
    "learning_rate": 0.00015944783290273708,
    "timestep_mean": 432.59375,
    "epsilon_norm": 6.474283695220947,
    "prediction_norm": 5.834794044494629,
    "timestamp": "2025-07-14T12:34:26.134320"
  },
  {
    "epoch": 1477,
    "batch_idx": 32,
    "loss": 0.3801082968711853,
    "learning_rate": 0.00015944783290273708,
    "timestep_mean": 531.90625,
    "epsilon_norm": 8.250083923339844,
    "prediction_norm": 5.542740821838379,
    "timestamp": "2025-07-14T12:34:26.139227"
  },
  {
    "epoch": 1477,
    "batch_idx": 64,
    "loss": 0.22532802820205688,
    "learning_rate": 0.00015944783290273708,
    "timestep_mean": 527.3125,
    "epsilon_norm": 8.583152770996094,
    "prediction_norm": 7.4537129402160645,
    "timestamp": "2025-07-14T12:34:26.144100"
  },
  {
    "epoch": 1477,
    "batch_idx": 96,
    "loss": 0.15837673842906952,
    "learning_rate": 0.00015944783290273708,
    "timestep_mean": 511.5625,
    "epsilon_norm": 7.363089561462402,
    "prediction_norm": 6.6463942527771,
    "timestamp": "2025-07-14T12:34:26.148942"
  },
  {
    "epoch": 1477,
    "batch_idx": 128,
    "loss": 0.2290584146976471,
    "learning_rate": 0.00015944783290273708,
    "timestep_mean": 423.1875,
    "epsilon_norm": 7.2824296951293945,
    "prediction_norm": 6.223005771636963,
    "timestamp": "2025-07-14T12:34:26.153796"
  },
  {
    "epoch": 1477,
    "batch_idx": 160,
    "loss": 0.20962408185005188,
    "learning_rate": 0.00015944783290273708,
    "timestep_mean": 534.6875,
    "epsilon_norm": 7.395326614379883,
    "prediction_norm": 6.057620525360107,
    "timestamp": "2025-07-14T12:34:26.158771"
  },
  {
    "epoch": 1477,
    "batch_idx": 192,
    "loss": 0.5543925762176514,
    "learning_rate": 0.00015944783290273708,
    "timestep_mean": 390.25,
    "epsilon_norm": 3.7211153507232666,
    "prediction_norm": 2.3358633518218994,
    "timestamp": "2025-07-14T12:34:26.163664"
  },
  {
    "epoch": 1478,
    "batch_idx": 0,
    "loss": 0.23220527172088623,
    "learning_rate": 0.0001588731954453021,
    "timestep_mean": 483.5625,
    "epsilon_norm": 7.155587673187256,
    "prediction_norm": 6.143520355224609,
    "timestamp": "2025-07-14T12:34:26.168494"
  },
  {
    "epoch": 1478,
    "batch_idx": 32,
    "loss": 0.29892489314079285,
    "learning_rate": 0.0001588731954453021,
    "timestep_mean": 525.65625,
    "epsilon_norm": 8.725622177124023,
    "prediction_norm": 7.303693771362305,
    "timestamp": "2025-07-14T12:34:26.173400"
  },
  {
    "epoch": 1478,
    "batch_idx": 64,
    "loss": 0.4767219126224518,
    "learning_rate": 0.0001588731954453021,
    "timestep_mean": 521.1875,
    "epsilon_norm": 9.238921165466309,
    "prediction_norm": 7.0846662521362305,
    "timestamp": "2025-07-14T12:34:26.178194"
  },
  {
    "epoch": 1478,
    "batch_idx": 96,
    "loss": 0.3580077886581421,
    "learning_rate": 0.0001588731954453021,
    "timestep_mean": 476.96875,
    "epsilon_norm": 7.683373928070068,
    "prediction_norm": 5.815418243408203,
    "timestamp": "2025-07-14T12:34:26.183011"
  },
  {
    "epoch": 1478,
    "batch_idx": 128,
    "loss": 0.4457520842552185,
    "learning_rate": 0.0001588731954453021,
    "timestep_mean": 445.875,
    "epsilon_norm": 7.553093433380127,
    "prediction_norm": 5.619913578033447,
    "timestamp": "2025-07-14T12:34:26.187981"
  },
  {
    "epoch": 1478,
    "batch_idx": 160,
    "loss": 0.2400486171245575,
    "learning_rate": 0.0001588731954453021,
    "timestep_mean": 507.25,
    "epsilon_norm": 8.708691596984863,
    "prediction_norm": 7.110196590423584,
    "timestamp": "2025-07-14T12:34:26.192824"
  },
  {
    "epoch": 1478,
    "batch_idx": 192,
    "loss": 0.09347153455018997,
    "learning_rate": 0.0001588731954453021,
    "timestep_mean": 652.75,
    "epsilon_norm": 3.4349544048309326,
    "prediction_norm": 3.29082989692688,
    "timestamp": "2025-07-14T12:34:26.197622"
  },
  {
    "epoch": 1479,
    "batch_idx": 0,
    "loss": 0.29099392890930176,
    "learning_rate": 0.00015829939968434704,
    "timestep_mean": 511.0625,
    "epsilon_norm": 7.751119613647461,
    "prediction_norm": 6.484065055847168,
    "timestamp": "2025-07-14T12:34:26.202490"
  },
  {
    "epoch": 1479,
    "batch_idx": 32,
    "loss": 0.41666269302368164,
    "learning_rate": 0.00015829939968434704,
    "timestep_mean": 369.0,
    "epsilon_norm": 8.433196067810059,
    "prediction_norm": 6.183428764343262,
    "timestamp": "2025-07-14T12:34:26.207286"
  },
  {
    "epoch": 1479,
    "batch_idx": 64,
    "loss": 0.24942588806152344,
    "learning_rate": 0.00015829939968434704,
    "timestep_mean": 530.09375,
    "epsilon_norm": 6.881004333496094,
    "prediction_norm": 5.0279459953308105,
    "timestamp": "2025-07-14T12:34:26.212126"
  },
  {
    "epoch": 1479,
    "batch_idx": 96,
    "loss": 0.363575279712677,
    "learning_rate": 0.00015829939968434704,
    "timestep_mean": 505.5,
    "epsilon_norm": 8.902918815612793,
    "prediction_norm": 7.620335102081299,
    "timestamp": "2025-07-14T12:34:26.217069"
  },
  {
    "epoch": 1479,
    "batch_idx": 128,
    "loss": 0.36359113454818726,
    "learning_rate": 0.00015829939968434704,
    "timestep_mean": 590.53125,
    "epsilon_norm": 8.950469017028809,
    "prediction_norm": 6.909289360046387,
    "timestamp": "2025-07-14T12:34:26.221923"
  },
  {
    "epoch": 1479,
    "batch_idx": 160,
    "loss": 0.1511095017194748,
    "learning_rate": 0.00015829939968434704,
    "timestep_mean": 518.1875,
    "epsilon_norm": 7.366805076599121,
    "prediction_norm": 6.5784592628479,
    "timestamp": "2025-07-14T12:34:26.226822"
  },
  {
    "epoch": 1479,
    "batch_idx": 192,
    "loss": 0.16593265533447266,
    "learning_rate": 0.00015829939968434704,
    "timestep_mean": 524.0,
    "epsilon_norm": 4.368930339813232,
    "prediction_norm": 3.6384167671203613,
    "timestamp": "2025-07-14T12:34:26.231688"
  },
  {
    "epoch": 1480,
    "batch_idx": 0,
    "loss": 0.18074649572372437,
    "learning_rate": 0.0001577264470356561,
    "timestep_mean": 563.09375,
    "epsilon_norm": 7.35833215713501,
    "prediction_norm": 6.5650434494018555,
    "timestamp": "2025-07-14T12:34:26.236547"
  },
  {
    "epoch": 1480,
    "batch_idx": 32,
    "loss": 0.3528941869735718,
    "learning_rate": 0.0001577264470356561,
    "timestep_mean": 551.96875,
    "epsilon_norm": 8.334433555603027,
    "prediction_norm": 6.38906192779541,
    "timestamp": "2025-07-14T12:34:26.241422"
  },
  {
    "epoch": 1480,
    "batch_idx": 64,
    "loss": 0.2930159270763397,
    "learning_rate": 0.0001577264470356561,
    "timestep_mean": 446.5625,
    "epsilon_norm": 6.634389400482178,
    "prediction_norm": 5.161781311035156,
    "timestamp": "2025-07-14T12:34:26.246362"
  },
  {
    "epoch": 1480,
    "batch_idx": 96,
    "loss": 0.29734253883361816,
    "learning_rate": 0.0001577264470356561,
    "timestep_mean": 456.46875,
    "epsilon_norm": 7.304963111877441,
    "prediction_norm": 6.0340256690979,
    "timestamp": "2025-07-14T12:34:26.251257"
  },
  {
    "epoch": 1480,
    "batch_idx": 128,
    "loss": 0.2319049835205078,
    "learning_rate": 0.0001577264470356561,
    "timestep_mean": 465.90625,
    "epsilon_norm": 7.645761489868164,
    "prediction_norm": 6.120350360870361,
    "timestamp": "2025-07-14T12:34:26.256110"
  },
  {
    "epoch": 1480,
    "batch_idx": 160,
    "loss": 0.4966244697570801,
    "learning_rate": 0.0001577264470356561,
    "timestep_mean": 534.65625,
    "epsilon_norm": 8.561403274536133,
    "prediction_norm": 7.2006354331970215,
    "timestamp": "2025-07-14T12:34:26.261096"
  },
  {
    "epoch": 1480,
    "batch_idx": 192,
    "loss": 0.1436154842376709,
    "learning_rate": 0.0001577264470356561,
    "timestep_mean": 512.875,
    "epsilon_norm": 4.619359493255615,
    "prediction_norm": 3.8191447257995605,
    "timestamp": "2025-07-14T12:34:26.265959"
  },
  {
    "epoch": 1481,
    "batch_idx": 0,
    "loss": 0.44970011711120605,
    "learning_rate": 0.00015715433891293237,
    "timestep_mean": 445.59375,
    "epsilon_norm": 7.855423450469971,
    "prediction_norm": 6.038942337036133,
    "timestamp": "2025-07-14T12:34:26.270971"
  },
  {
    "epoch": 1481,
    "batch_idx": 32,
    "loss": 0.4611450135707855,
    "learning_rate": 0.00015715433891293237,
    "timestep_mean": 453.71875,
    "epsilon_norm": 9.387547492980957,
    "prediction_norm": 6.672068119049072,
    "timestamp": "2025-07-14T12:34:26.275886"
  },
  {
    "epoch": 1481,
    "batch_idx": 64,
    "loss": 0.22406630218029022,
    "learning_rate": 0.00015715433891293237,
    "timestep_mean": 473.40625,
    "epsilon_norm": 6.190107345581055,
    "prediction_norm": 5.9211297035217285,
    "timestamp": "2025-07-14T12:34:26.280765"
  },
  {
    "epoch": 1481,
    "batch_idx": 96,
    "loss": 0.457077294588089,
    "learning_rate": 0.00015715433891293237,
    "timestep_mean": 466.15625,
    "epsilon_norm": 8.301669120788574,
    "prediction_norm": 6.532475471496582,
    "timestamp": "2025-07-14T12:34:26.285564"
  },
  {
    "epoch": 1481,
    "batch_idx": 128,
    "loss": 0.25092828273773193,
    "learning_rate": 0.00015715433891293237,
    "timestep_mean": 521.5,
    "epsilon_norm": 8.387940406799316,
    "prediction_norm": 7.833171367645264,
    "timestamp": "2025-07-14T12:34:26.290369"
  },
  {
    "epoch": 1481,
    "batch_idx": 160,
    "loss": 0.4174375832080841,
    "learning_rate": 0.00015715433891293237,
    "timestep_mean": 473.59375,
    "epsilon_norm": 7.877415657043457,
    "prediction_norm": 6.135091304779053,
    "timestamp": "2025-07-14T12:34:26.295201"
  },
  {
    "epoch": 1481,
    "batch_idx": 192,
    "loss": 0.19918277859687805,
    "learning_rate": 0.00015715433891293237,
    "timestep_mean": 546.25,
    "epsilon_norm": 3.3361904621124268,
    "prediction_norm": 2.7965543270111084,
    "timestamp": "2025-07-14T12:34:26.299993"
  },
  {
    "epoch": 1482,
    "batch_idx": 0,
    "loss": 0.4525962769985199,
    "learning_rate": 0.00015658307672779622,
    "timestep_mean": 472.75,
    "epsilon_norm": 8.0689697265625,
    "prediction_norm": 6.488657474517822,
    "timestamp": "2025-07-14T12:34:26.304858"
  },
  {
    "epoch": 1482,
    "batch_idx": 32,
    "loss": 0.23027780652046204,
    "learning_rate": 0.00015658307672779622,
    "timestep_mean": 569.90625,
    "epsilon_norm": 7.590609550476074,
    "prediction_norm": 6.863560199737549,
    "timestamp": "2025-07-14T12:34:26.309683"
  },
  {
    "epoch": 1482,
    "batch_idx": 64,
    "loss": 0.37529098987579346,
    "learning_rate": 0.00015658307672779622,
    "timestep_mean": 420.09375,
    "epsilon_norm": 7.985445022583008,
    "prediction_norm": 6.428459644317627,
    "timestamp": "2025-07-14T12:34:26.314547"
  },
  {
    "epoch": 1482,
    "batch_idx": 96,
    "loss": 0.3290402889251709,
    "learning_rate": 0.00015658307672779622,
    "timestep_mean": 488.21875,
    "epsilon_norm": 8.590728759765625,
    "prediction_norm": 6.925854682922363,
    "timestamp": "2025-07-14T12:34:26.319403"
  },
  {
    "epoch": 1482,
    "batch_idx": 128,
    "loss": 0.21338161826133728,
    "learning_rate": 0.00015658307672779622,
    "timestep_mean": 597.15625,
    "epsilon_norm": 7.006597518920898,
    "prediction_norm": 5.936255931854248,
    "timestamp": "2025-07-14T12:34:26.324243"
  },
  {
    "epoch": 1482,
    "batch_idx": 160,
    "loss": 0.510581374168396,
    "learning_rate": 0.00015658307672779622,
    "timestep_mean": 482.75,
    "epsilon_norm": 8.62621021270752,
    "prediction_norm": 6.7197418212890625,
    "timestamp": "2025-07-14T12:34:26.329094"
  },
  {
    "epoch": 1482,
    "batch_idx": 192,
    "loss": 0.7588646411895752,
    "learning_rate": 0.00015658307672779622,
    "timestep_mean": 624.25,
    "epsilon_norm": 4.071109771728516,
    "prediction_norm": 2.4077911376953125,
    "timestamp": "2025-07-14T12:34:26.333948"
  },
  {
    "epoch": 1483,
    "batch_idx": 0,
    "loss": 0.2376173883676529,
    "learning_rate": 0.00015601266188978015,
    "timestep_mean": 431.78125,
    "epsilon_norm": 8.13750171661377,
    "prediction_norm": 7.091450214385986,
    "timestamp": "2025-07-14T12:34:26.338746"
  },
  {
    "epoch": 1483,
    "batch_idx": 32,
    "loss": 0.3462037444114685,
    "learning_rate": 0.00015601266188978015,
    "timestep_mean": 533.78125,
    "epsilon_norm": 8.994804382324219,
    "prediction_norm": 7.709386825561523,
    "timestamp": "2025-07-14T12:34:26.343631"
  },
  {
    "epoch": 1483,
    "batch_idx": 64,
    "loss": 0.2659146189689636,
    "learning_rate": 0.00015601266188978015,
    "timestep_mean": 533.375,
    "epsilon_norm": 8.973467826843262,
    "prediction_norm": 7.831374645233154,
    "timestamp": "2025-07-14T12:34:26.348497"
  },
  {
    "epoch": 1483,
    "batch_idx": 96,
    "loss": 0.2867124378681183,
    "learning_rate": 0.00015601266188978015,
    "timestep_mean": 421.3125,
    "epsilon_norm": 7.493392467498779,
    "prediction_norm": 6.689394474029541,
    "timestamp": "2025-07-14T12:34:26.353411"
  },
  {
    "epoch": 1483,
    "batch_idx": 128,
    "loss": 0.15978676080703735,
    "learning_rate": 0.00015601266188978015,
    "timestep_mean": 522.9375,
    "epsilon_norm": 8.113210678100586,
    "prediction_norm": 7.510124683380127,
    "timestamp": "2025-07-14T12:34:26.358269"
  },
  {
    "epoch": 1483,
    "batch_idx": 160,
    "loss": 0.21753734350204468,
    "learning_rate": 0.00015601266188978015,
    "timestep_mean": 533.1875,
    "epsilon_norm": 6.952460765838623,
    "prediction_norm": 6.3911542892456055,
    "timestamp": "2025-07-14T12:34:26.363189"
  },
  {
    "epoch": 1483,
    "batch_idx": 192,
    "loss": 0.3097352087497711,
    "learning_rate": 0.00015601266188978015,
    "timestep_mean": 514.375,
    "epsilon_norm": 4.396899223327637,
    "prediction_norm": 3.642498254776001,
    "timestamp": "2025-07-14T12:34:26.368009"
  },
  {
    "epoch": 1484,
    "batch_idx": 0,
    "loss": 0.3585057258605957,
    "learning_rate": 0.000155443095806326,
    "timestep_mean": 456.53125,
    "epsilon_norm": 7.382333278656006,
    "prediction_norm": 5.753702163696289,
    "timestamp": "2025-07-14T12:34:26.372896"
  },
  {
    "epoch": 1484,
    "batch_idx": 32,
    "loss": 0.20054100453853607,
    "learning_rate": 0.000155443095806326,
    "timestep_mean": 550.5625,
    "epsilon_norm": 7.311356544494629,
    "prediction_norm": 7.210643291473389,
    "timestamp": "2025-07-14T12:34:26.377732"
  },
  {
    "epoch": 1484,
    "batch_idx": 64,
    "loss": 0.20093417167663574,
    "learning_rate": 0.000155443095806326,
    "timestep_mean": 471.375,
    "epsilon_norm": 7.305144309997559,
    "prediction_norm": 6.287032604217529,
    "timestamp": "2025-07-14T12:34:26.382583"
  },
  {
    "epoch": 1484,
    "batch_idx": 96,
    "loss": 0.1638982892036438,
    "learning_rate": 0.000155443095806326,
    "timestep_mean": 536.6875,
    "epsilon_norm": 7.348345756530762,
    "prediction_norm": 6.306121349334717,
    "timestamp": "2025-07-14T12:34:26.387549"
  },
  {
    "epoch": 1484,
    "batch_idx": 128,
    "loss": 0.32424336671829224,
    "learning_rate": 0.000155443095806326,
    "timestep_mean": 510.75,
    "epsilon_norm": 7.751571178436279,
    "prediction_norm": 6.898329734802246,
    "timestamp": "2025-07-14T12:34:26.392464"
  },
  {
    "epoch": 1484,
    "batch_idx": 160,
    "loss": 0.13667495548725128,
    "learning_rate": 0.000155443095806326,
    "timestep_mean": 513.375,
    "epsilon_norm": 7.024573802947998,
    "prediction_norm": 5.7627387046813965,
    "timestamp": "2025-07-14T12:34:26.397239"
  },
  {
    "epoch": 1484,
    "batch_idx": 192,
    "loss": 0.5423576831817627,
    "learning_rate": 0.000155443095806326,
    "timestep_mean": 345.375,
    "epsilon_norm": 4.276615142822266,
    "prediction_norm": 2.9759812355041504,
    "timestamp": "2025-07-14T12:34:26.402029"
  },
  {
    "epoch": 1485,
    "batch_idx": 0,
    "loss": 0.39087527990341187,
    "learning_rate": 0.0001548743798827817,
    "timestep_mean": 367.96875,
    "epsilon_norm": 8.028313636779785,
    "prediction_norm": 6.450324058532715,
    "timestamp": "2025-07-14T12:34:26.406890"
  },
  {
    "epoch": 1485,
    "batch_idx": 32,
    "loss": 0.20268382132053375,
    "learning_rate": 0.0001548743798827817,
    "timestep_mean": 456.09375,
    "epsilon_norm": 7.911500930786133,
    "prediction_norm": 7.156284332275391,
    "timestamp": "2025-07-14T12:34:26.411761"
  },
  {
    "epoch": 1485,
    "batch_idx": 64,
    "loss": 0.11810516566038132,
    "learning_rate": 0.0001548743798827817,
    "timestep_mean": 555.4375,
    "epsilon_norm": 7.038870811462402,
    "prediction_norm": 6.589445114135742,
    "timestamp": "2025-07-14T12:34:26.416715"
  },
  {
    "epoch": 1485,
    "batch_idx": 96,
    "loss": 0.20600315928459167,
    "learning_rate": 0.0001548743798827817,
    "timestep_mean": 471.75,
    "epsilon_norm": 7.538063049316406,
    "prediction_norm": 7.005572319030762,
    "timestamp": "2025-07-14T12:34:26.421739"
  },
  {
    "epoch": 1485,
    "batch_idx": 128,
    "loss": 0.2686687707901001,
    "learning_rate": 0.0001548743798827817,
    "timestep_mean": 536.8125,
    "epsilon_norm": 8.114243507385254,
    "prediction_norm": 7.114001274108887,
    "timestamp": "2025-07-14T12:34:26.426582"
  },
  {
    "epoch": 1485,
    "batch_idx": 160,
    "loss": 0.22803831100463867,
    "learning_rate": 0.0001548743798827817,
    "timestep_mean": 513.0,
    "epsilon_norm": 7.884645938873291,
    "prediction_norm": 6.572947978973389,
    "timestamp": "2025-07-14T12:34:26.431500"
  },
  {
    "epoch": 1485,
    "batch_idx": 192,
    "loss": 0.8086562752723694,
    "learning_rate": 0.0001548743798827817,
    "timestep_mean": 398.75,
    "epsilon_norm": 4.163924217224121,
    "prediction_norm": 2.5912599563598633,
    "timestamp": "2025-07-14T12:34:26.436683"
  },
  {
    "epoch": 1486,
    "batch_idx": 0,
    "loss": 0.3074934482574463,
    "learning_rate": 0.0001543065155223971,
    "timestep_mean": 467.21875,
    "epsilon_norm": 7.860308647155762,
    "prediction_norm": 6.46044921875,
    "timestamp": "2025-07-14T12:34:26.441624"
  },
  {
    "epoch": 1486,
    "batch_idx": 32,
    "loss": 0.41752421855926514,
    "learning_rate": 0.0001543065155223971,
    "timestep_mean": 367.53125,
    "epsilon_norm": 7.649679183959961,
    "prediction_norm": 6.041290760040283,
    "timestamp": "2025-07-14T12:34:26.446544"
  },
  {
    "epoch": 1486,
    "batch_idx": 64,
    "loss": 0.2063141167163849,
    "learning_rate": 0.0001543065155223971,
    "timestep_mean": 493.25,
    "epsilon_norm": 6.528337001800537,
    "prediction_norm": 5.536760330200195,
    "timestamp": "2025-07-14T12:34:26.451413"
  },
  {
    "epoch": 1486,
    "batch_idx": 96,
    "loss": 0.3797808885574341,
    "learning_rate": 0.0001543065155223971,
    "timestep_mean": 437.625,
    "epsilon_norm": 8.675824165344238,
    "prediction_norm": 6.688408851623535,
    "timestamp": "2025-07-14T12:34:26.456301"
  },
  {
    "epoch": 1486,
    "batch_idx": 128,
    "loss": 0.21486994624137878,
    "learning_rate": 0.0001543065155223971,
    "timestep_mean": 417.40625,
    "epsilon_norm": 7.601919174194336,
    "prediction_norm": 6.464085578918457,
    "timestamp": "2025-07-14T12:34:26.461210"
  },
  {
    "epoch": 1486,
    "batch_idx": 160,
    "loss": 0.25703415274620056,
    "learning_rate": 0.0001543065155223971,
    "timestep_mean": 475.59375,
    "epsilon_norm": 7.988515853881836,
    "prediction_norm": 6.335638046264648,
    "timestamp": "2025-07-14T12:34:26.466067"
  },
  {
    "epoch": 1486,
    "batch_idx": 192,
    "loss": 0.14795108139514923,
    "learning_rate": 0.0001543065155223971,
    "timestep_mean": 456.375,
    "epsilon_norm": 4.568940162658691,
    "prediction_norm": 4.4896416664123535,
    "timestamp": "2025-07-14T12:34:26.471057"
  },
  {
    "epoch": 1487,
    "batch_idx": 0,
    "loss": 0.24012218415737152,
    "learning_rate": 0.00015373950412632102,
    "timestep_mean": 517.65625,
    "epsilon_norm": 7.804856300354004,
    "prediction_norm": 6.953855991363525,
    "timestamp": "2025-07-14T12:34:26.476021"
  },
  {
    "epoch": 1487,
    "batch_idx": 32,
    "loss": 0.46451207995414734,
    "learning_rate": 0.00015373950412632102,
    "timestep_mean": 467.125,
    "epsilon_norm": 7.968626499176025,
    "prediction_norm": 6.659521579742432,
    "timestamp": "2025-07-14T12:34:26.480895"
  },
  {
    "epoch": 1487,
    "batch_idx": 64,
    "loss": 0.16751965880393982,
    "learning_rate": 0.00015373950412632102,
    "timestep_mean": 524.71875,
    "epsilon_norm": 7.330039024353027,
    "prediction_norm": 6.515712738037109,
    "timestamp": "2025-07-14T12:34:26.485756"
  },
  {
    "epoch": 1487,
    "batch_idx": 96,
    "loss": 0.35458672046661377,
    "learning_rate": 0.00015373950412632102,
    "timestep_mean": 458.0,
    "epsilon_norm": 8.882356643676758,
    "prediction_norm": 6.797726154327393,
    "timestamp": "2025-07-14T12:34:26.490949"
  },
  {
    "epoch": 1487,
    "batch_idx": 128,
    "loss": 0.2120479941368103,
    "learning_rate": 0.00015373950412632102,
    "timestep_mean": 490.59375,
    "epsilon_norm": 7.2581400871276855,
    "prediction_norm": 6.3714704513549805,
    "timestamp": "2025-07-14T12:34:26.496145"
  },
  {
    "epoch": 1487,
    "batch_idx": 160,
    "loss": 0.30460333824157715,
    "learning_rate": 0.00015373950412632102,
    "timestep_mean": 515.53125,
    "epsilon_norm": 8.66396427154541,
    "prediction_norm": 7.304894924163818,
    "timestamp": "2025-07-14T12:34:26.501245"
  },
  {
    "epoch": 1487,
    "batch_idx": 192,
    "loss": 0.15779916942119598,
    "learning_rate": 0.00015373950412632102,
    "timestep_mean": 601.25,
    "epsilon_norm": 3.492898941040039,
    "prediction_norm": 2.7411530017852783,
    "timestamp": "2025-07-14T12:34:26.506506"
  },
  {
    "epoch": 1488,
    "batch_idx": 0,
    "loss": 0.21784177422523499,
    "learning_rate": 0.00015317334709359786,
    "timestep_mean": 426.90625,
    "epsilon_norm": 7.878545761108398,
    "prediction_norm": 6.50444221496582,
    "timestamp": "2025-07-14T12:34:26.511786"
  },
  {
    "epoch": 1488,
    "batch_idx": 32,
    "loss": 0.14929740130901337,
    "learning_rate": 0.00015317334709359786,
    "timestep_mean": 519.75,
    "epsilon_norm": 7.198026657104492,
    "prediction_norm": 6.098320007324219,
    "timestamp": "2025-07-14T12:34:26.517022"
  },
  {
    "epoch": 1488,
    "batch_idx": 64,
    "loss": 0.14350195229053497,
    "learning_rate": 0.00015317334709359786,
    "timestep_mean": 474.28125,
    "epsilon_norm": 7.3101911544799805,
    "prediction_norm": 6.15464448928833,
    "timestamp": "2025-07-14T12:34:26.522026"
  },
  {
    "epoch": 1488,
    "batch_idx": 96,
    "loss": 0.34376487135887146,
    "learning_rate": 0.00015317334709359786,
    "timestep_mean": 505.78125,
    "epsilon_norm": 8.210672378540039,
    "prediction_norm": 6.453434944152832,
    "timestamp": "2025-07-14T12:34:26.527094"
  },
  {
    "epoch": 1488,
    "batch_idx": 128,
    "loss": 0.18491367995738983,
    "learning_rate": 0.00015317334709359786,
    "timestep_mean": 474.4375,
    "epsilon_norm": 7.7103800773620605,
    "prediction_norm": 6.163224697113037,
    "timestamp": "2025-07-14T12:34:26.532252"
  },
  {
    "epoch": 1488,
    "batch_idx": 160,
    "loss": 0.27776631712913513,
    "learning_rate": 0.00015317334709359786,
    "timestep_mean": 470.9375,
    "epsilon_norm": 7.616562843322754,
    "prediction_norm": 6.318559646606445,
    "timestamp": "2025-07-14T12:34:26.537408"
  },
  {
    "epoch": 1488,
    "batch_idx": 192,
    "loss": 0.4830377995967865,
    "learning_rate": 0.00015317334709359786,
    "timestep_mean": 596.375,
    "epsilon_norm": 4.403945446014404,
    "prediction_norm": 3.280324935913086,
    "timestamp": "2025-07-14T12:34:26.542487"
  },
  {
    "epoch": 1489,
    "batch_idx": 0,
    "loss": 0.1979156732559204,
    "learning_rate": 0.00015260804582116355,
    "timestep_mean": 516.78125,
    "epsilon_norm": 8.137203216552734,
    "prediction_norm": 7.2451677322387695,
    "timestamp": "2025-07-14T12:34:26.547696"
  },
  {
    "epoch": 1489,
    "batch_idx": 32,
    "loss": 0.12811681628227234,
    "learning_rate": 0.00015260804582116355,
    "timestep_mean": 554.0625,
    "epsilon_norm": 7.341822624206543,
    "prediction_norm": 6.441883087158203,
    "timestamp": "2025-07-14T12:34:26.552909"
  },
  {
    "epoch": 1489,
    "batch_idx": 64,
    "loss": 0.28120875358581543,
    "learning_rate": 0.00015260804582116355,
    "timestep_mean": 515.5,
    "epsilon_norm": 8.261563301086426,
    "prediction_norm": 7.257723808288574,
    "timestamp": "2025-07-14T12:34:26.558014"
  },
  {
    "epoch": 1489,
    "batch_idx": 96,
    "loss": 0.17314200103282928,
    "learning_rate": 0.00015260804582116355,
    "timestep_mean": 574.5,
    "epsilon_norm": 7.567968368530273,
    "prediction_norm": 6.676754951477051,
    "timestamp": "2025-07-14T12:34:26.563249"
  },
  {
    "epoch": 1489,
    "batch_idx": 128,
    "loss": 0.16099447011947632,
    "learning_rate": 0.00015260804582116355,
    "timestep_mean": 457.875,
    "epsilon_norm": 7.906875133514404,
    "prediction_norm": 6.797409534454346,
    "timestamp": "2025-07-14T12:34:26.568334"
  },
  {
    "epoch": 1489,
    "batch_idx": 160,
    "loss": 0.19422271847724915,
    "learning_rate": 0.00015260804582116355,
    "timestep_mean": 499.0,
    "epsilon_norm": 6.792737007141113,
    "prediction_norm": 6.0376176834106445,
    "timestamp": "2025-07-14T12:34:26.573390"
  },
  {
    "epoch": 1489,
    "batch_idx": 192,
    "loss": 0.23748627305030823,
    "learning_rate": 0.00015260804582116355,
    "timestep_mean": 506.125,
    "epsilon_norm": 3.7444941997528076,
    "prediction_norm": 2.870126962661743,
    "timestamp": "2025-07-14T12:34:26.578618"
  },
  {
    "epoch": 1490,
    "batch_idx": 0,
    "loss": 0.14895352721214294,
    "learning_rate": 0.00015204360170384312,
    "timestep_mean": 473.40625,
    "epsilon_norm": 7.676016330718994,
    "prediction_norm": 6.562191486358643,
    "timestamp": "2025-07-14T12:34:26.583836"
  },
  {
    "epoch": 1490,
    "batch_idx": 32,
    "loss": 0.38251256942749023,
    "learning_rate": 0.00015204360170384312,
    "timestep_mean": 454.25,
    "epsilon_norm": 7.500818729400635,
    "prediction_norm": 6.527028560638428,
    "timestamp": "2025-07-14T12:34:26.588811"
  },
  {
    "epoch": 1490,
    "batch_idx": 64,
    "loss": 0.183025062084198,
    "learning_rate": 0.00015204360170384312,
    "timestep_mean": 528.875,
    "epsilon_norm": 7.978550910949707,
    "prediction_norm": 6.604328155517578,
    "timestamp": "2025-07-14T12:34:26.594018"
  },
  {
    "epoch": 1490,
    "batch_idx": 96,
    "loss": 0.3035888969898224,
    "learning_rate": 0.00015204360170384312,
    "timestep_mean": 538.8125,
    "epsilon_norm": 7.454800128936768,
    "prediction_norm": 5.632410526275635,
    "timestamp": "2025-07-14T12:34:26.599272"
  },
  {
    "epoch": 1490,
    "batch_idx": 128,
    "loss": 0.2762829065322876,
    "learning_rate": 0.00015204360170384312,
    "timestep_mean": 507.4375,
    "epsilon_norm": 7.5788469314575195,
    "prediction_norm": 7.229347229003906,
    "timestamp": "2025-07-14T12:34:26.604461"
  },
  {
    "epoch": 1490,
    "batch_idx": 160,
    "loss": 0.49564164876937866,
    "learning_rate": 0.00015204360170384312,
    "timestep_mean": 521.0625,
    "epsilon_norm": 7.656303882598877,
    "prediction_norm": 6.522058010101318,
    "timestamp": "2025-07-14T12:34:26.609410"
  },
  {
    "epoch": 1490,
    "batch_idx": 192,
    "loss": 0.03813950717449188,
    "learning_rate": 0.00015204360170384312,
    "timestep_mean": 708.125,
    "epsilon_norm": 4.399134635925293,
    "prediction_norm": 4.581679821014404,
    "timestamp": "2025-07-14T12:34:26.614444"
  },
  {
    "epoch": 1491,
    "batch_idx": 0,
    "loss": 0.2866280674934387,
    "learning_rate": 0.00015148001613434612,
    "timestep_mean": 543.15625,
    "epsilon_norm": 8.948677062988281,
    "prediction_norm": 7.78601598739624,
    "timestamp": "2025-07-14T12:34:26.619950"
  },
  {
    "epoch": 1491,
    "batch_idx": 32,
    "loss": 0.16473238170146942,
    "learning_rate": 0.00015148001613434612,
    "timestep_mean": 518.09375,
    "epsilon_norm": 8.513707160949707,
    "prediction_norm": 8.402362823486328,
    "timestamp": "2025-07-14T12:34:26.625117"
  },
  {
    "epoch": 1491,
    "batch_idx": 64,
    "loss": 0.27563706040382385,
    "learning_rate": 0.00015148001613434612,
    "timestep_mean": 488.3125,
    "epsilon_norm": 7.99456787109375,
    "prediction_norm": 7.019600868225098,
    "timestamp": "2025-07-14T12:34:26.630220"
  },
  {
    "epoch": 1491,
    "batch_idx": 96,
    "loss": 0.24212363362312317,
    "learning_rate": 0.00015148001613434612,
    "timestep_mean": 478.53125,
    "epsilon_norm": 7.6243577003479,
    "prediction_norm": 6.637940883636475,
    "timestamp": "2025-07-14T12:34:26.635398"
  },
  {
    "epoch": 1491,
    "batch_idx": 128,
    "loss": 0.20884305238723755,
    "learning_rate": 0.00015148001613434612,
    "timestep_mean": 520.5625,
    "epsilon_norm": 8.62856674194336,
    "prediction_norm": 7.99557638168335,
    "timestamp": "2025-07-14T12:34:26.640733"
  },
  {
    "epoch": 1491,
    "batch_idx": 160,
    "loss": 0.33172863721847534,
    "learning_rate": 0.00015148001613434612,
    "timestep_mean": 458.46875,
    "epsilon_norm": 7.722878932952881,
    "prediction_norm": 6.267447471618652,
    "timestamp": "2025-07-14T12:34:26.646110"
  },
  {
    "epoch": 1491,
    "batch_idx": 192,
    "loss": 0.03582824021577835,
    "learning_rate": 0.00015148001613434612,
    "timestep_mean": 598.625,
    "epsilon_norm": 2.4375009536743164,
    "prediction_norm": 2.337341070175171,
    "timestamp": "2025-07-14T12:34:26.651355"
  },
  {
    "epoch": 1492,
    "batch_idx": 0,
    "loss": 0.2835714519023895,
    "learning_rate": 0.000150917290503264,
    "timestep_mean": 483.625,
    "epsilon_norm": 7.117825984954834,
    "prediction_norm": 6.135290622711182,
    "timestamp": "2025-07-14T12:34:26.656572"
  },
  {
    "epoch": 1492,
    "batch_idx": 32,
    "loss": 0.1767261028289795,
    "learning_rate": 0.000150917290503264,
    "timestep_mean": 511.375,
    "epsilon_norm": 9.0193510055542,
    "prediction_norm": 8.037521362304688,
    "timestamp": "2025-07-14T12:34:26.661754"
  },
  {
    "epoch": 1492,
    "batch_idx": 64,
    "loss": 0.35348212718963623,
    "learning_rate": 0.000150917290503264,
    "timestep_mean": 501.15625,
    "epsilon_norm": 8.453152656555176,
    "prediction_norm": 7.172731399536133,
    "timestamp": "2025-07-14T12:34:26.666902"
  },
  {
    "epoch": 1492,
    "batch_idx": 96,
    "loss": 0.22292889654636383,
    "learning_rate": 0.000150917290503264,
    "timestep_mean": 466.34375,
    "epsilon_norm": 7.457921504974365,
    "prediction_norm": 6.258082389831543,
    "timestamp": "2025-07-14T12:34:26.672034"
  },
  {
    "epoch": 1492,
    "batch_idx": 128,
    "loss": 0.46746891736984253,
    "learning_rate": 0.000150917290503264,
    "timestep_mean": 434.84375,
    "epsilon_norm": 8.597241401672363,
    "prediction_norm": 6.695247173309326,
    "timestamp": "2025-07-14T12:34:26.677098"
  },
  {
    "epoch": 1492,
    "batch_idx": 160,
    "loss": 0.3506377935409546,
    "learning_rate": 0.000150917290503264,
    "timestep_mean": 466.84375,
    "epsilon_norm": 7.634433746337891,
    "prediction_norm": 6.5204997062683105,
    "timestamp": "2025-07-14T12:34:26.682361"
  },
  {
    "epoch": 1492,
    "batch_idx": 192,
    "loss": 1.6877108812332153,
    "learning_rate": 0.000150917290503264,
    "timestep_mean": 316.125,
    "epsilon_norm": 5.486788272857666,
    "prediction_norm": 2.9446048736572266,
    "timestamp": "2025-07-14T12:34:26.687610"
  },
  {
    "epoch": 1493,
    "batch_idx": 0,
    "loss": 0.17944782972335815,
    "learning_rate": 0.00015035542619906608,
    "timestep_mean": 558.75,
    "epsilon_norm": 7.12487268447876,
    "prediction_norm": 5.843646049499512,
    "timestamp": "2025-07-14T12:34:26.692787"
  },
  {
    "epoch": 1493,
    "batch_idx": 32,
    "loss": 0.13206034898757935,
    "learning_rate": 0.00015035542619906608,
    "timestep_mean": 506.5,
    "epsilon_norm": 7.073125839233398,
    "prediction_norm": 5.954268932342529,
    "timestamp": "2025-07-14T12:34:26.697894"
  },
  {
    "epoch": 1493,
    "batch_idx": 64,
    "loss": 0.1806875467300415,
    "learning_rate": 0.00015035542619906608,
    "timestep_mean": 577.34375,
    "epsilon_norm": 8.6619291305542,
    "prediction_norm": 7.672141075134277,
    "timestamp": "2025-07-14T12:34:26.703060"
  },
  {
    "epoch": 1493,
    "batch_idx": 96,
    "loss": 0.3445906639099121,
    "learning_rate": 0.00015035542619906608,
    "timestep_mean": 472.5,
    "epsilon_norm": 8.126585006713867,
    "prediction_norm": 6.423988342285156,
    "timestamp": "2025-07-14T12:34:26.708271"
  },
  {
    "epoch": 1493,
    "batch_idx": 128,
    "loss": 0.28325071930885315,
    "learning_rate": 0.00015035542619906608,
    "timestep_mean": 440.625,
    "epsilon_norm": 7.403159141540527,
    "prediction_norm": 6.24997615814209,
    "timestamp": "2025-07-14T12:34:26.713465"
  },
  {
    "epoch": 1493,
    "batch_idx": 160,
    "loss": 0.2179892361164093,
    "learning_rate": 0.00015035542619906608,
    "timestep_mean": 565.0,
    "epsilon_norm": 8.529603958129883,
    "prediction_norm": 7.133471965789795,
    "timestamp": "2025-07-14T12:34:26.718675"
  },
  {
    "epoch": 1493,
    "batch_idx": 192,
    "loss": 0.24815280735492706,
    "learning_rate": 0.00015035542619906608,
    "timestep_mean": 466.875,
    "epsilon_norm": 5.0987653732299805,
    "prediction_norm": 4.585080623626709,
    "timestamp": "2025-07-14T12:34:26.723831"
  },
  {
    "epoch": 1494,
    "batch_idx": 0,
    "loss": 0.3689730167388916,
    "learning_rate": 0.0001497944246080972,
    "timestep_mean": 465.21875,
    "epsilon_norm": 8.296135902404785,
    "prediction_norm": 6.638923168182373,
    "timestamp": "2025-07-14T12:34:26.729061"
  },
  {
    "epoch": 1494,
    "batch_idx": 32,
    "loss": 0.17957234382629395,
    "learning_rate": 0.0001497944246080972,
    "timestep_mean": 541.1875,
    "epsilon_norm": 7.434850215911865,
    "prediction_norm": 6.667184829711914,
    "timestamp": "2025-07-14T12:34:26.734154"
  },
  {
    "epoch": 1494,
    "batch_idx": 64,
    "loss": 0.461516797542572,
    "learning_rate": 0.0001497944246080972,
    "timestep_mean": 370.8125,
    "epsilon_norm": 8.519734382629395,
    "prediction_norm": 6.263016700744629,
    "timestamp": "2025-07-14T12:34:26.739344"
  },
  {
    "epoch": 1494,
    "batch_idx": 96,
    "loss": 0.1928807497024536,
    "learning_rate": 0.0001497944246080972,
    "timestep_mean": 541.9375,
    "epsilon_norm": 7.58963680267334,
    "prediction_norm": 6.751527786254883,
    "timestamp": "2025-07-14T12:34:26.744336"
  },
  {
    "epoch": 1494,
    "batch_idx": 128,
    "loss": 0.21508486568927765,
    "learning_rate": 0.0001497944246080972,
    "timestep_mean": 463.78125,
    "epsilon_norm": 7.110667705535889,
    "prediction_norm": 6.53901481628418,
    "timestamp": "2025-07-14T12:34:26.749577"
  },
  {
    "epoch": 1494,
    "batch_idx": 160,
    "loss": 0.2755270004272461,
    "learning_rate": 0.0001497944246080972,
    "timestep_mean": 438.65625,
    "epsilon_norm": 7.521365642547607,
    "prediction_norm": 6.764977931976318,
    "timestamp": "2025-07-14T12:34:26.754761"
  },
  {
    "epoch": 1494,
    "batch_idx": 192,
    "loss": 0.3860909044742584,
    "learning_rate": 0.0001497944246080972,
    "timestep_mean": 355.75,
    "epsilon_norm": 3.8690028190612793,
    "prediction_norm": 3.870274305343628,
    "timestamp": "2025-07-14T12:34:26.759801"
  },
  {
    "epoch": 1495,
    "batch_idx": 0,
    "loss": 0.22161775827407837,
    "learning_rate": 0.0001492342871145724,
    "timestep_mean": 502.21875,
    "epsilon_norm": 7.408548831939697,
    "prediction_norm": 6.096272945404053,
    "timestamp": "2025-07-14T12:34:26.764941"
  },
  {
    "epoch": 1495,
    "batch_idx": 32,
    "loss": 0.2758740186691284,
    "learning_rate": 0.0001492342871145724,
    "timestep_mean": 601.0625,
    "epsilon_norm": 7.917109966278076,
    "prediction_norm": 6.774819850921631,
    "timestamp": "2025-07-14T12:34:26.770093"
  },
  {
    "epoch": 1495,
    "batch_idx": 64,
    "loss": 0.3533359169960022,
    "learning_rate": 0.0001492342871145724,
    "timestep_mean": 489.96875,
    "epsilon_norm": 6.850264549255371,
    "prediction_norm": 5.478992938995361,
    "timestamp": "2025-07-14T12:34:26.775206"
  },
  {
    "epoch": 1495,
    "batch_idx": 96,
    "loss": 0.2127259224653244,
    "learning_rate": 0.0001492342871145724,
    "timestep_mean": 478.28125,
    "epsilon_norm": 7.753929138183594,
    "prediction_norm": 6.437540054321289,
    "timestamp": "2025-07-14T12:34:26.780262"
  },
  {
    "epoch": 1495,
    "batch_idx": 128,
    "loss": 0.3377626836299896,
    "learning_rate": 0.0001492342871145724,
    "timestep_mean": 498.46875,
    "epsilon_norm": 9.544052124023438,
    "prediction_norm": 8.569486618041992,
    "timestamp": "2025-07-14T12:34:26.785333"
  },
  {
    "epoch": 1495,
    "batch_idx": 160,
    "loss": 0.4456164538860321,
    "learning_rate": 0.0001492342871145724,
    "timestep_mean": 423.03125,
    "epsilon_norm": 9.703302383422852,
    "prediction_norm": 7.314303398132324,
    "timestamp": "2025-07-14T12:34:26.790526"
  },
  {
    "epoch": 1495,
    "batch_idx": 192,
    "loss": 0.36134085059165955,
    "learning_rate": 0.0001492342871145724,
    "timestep_mean": 388.625,
    "epsilon_norm": 3.4342989921569824,
    "prediction_norm": 2.5847973823547363,
    "timestamp": "2025-07-14T12:34:26.795753"
  },
  {
    "epoch": 1496,
    "batch_idx": 0,
    "loss": 0.2392670065164566,
    "learning_rate": 0.00014867501510057554,
    "timestep_mean": 548.5625,
    "epsilon_norm": 7.348231315612793,
    "prediction_norm": 6.1583571434021,
    "timestamp": "2025-07-14T12:34:26.800972"
  },
  {
    "epoch": 1496,
    "batch_idx": 32,
    "loss": 0.2899610698223114,
    "learning_rate": 0.00014867501510057554,
    "timestep_mean": 487.0,
    "epsilon_norm": 7.780249118804932,
    "prediction_norm": 6.411178112030029,
    "timestamp": "2025-07-14T12:34:26.806126"
  },
  {
    "epoch": 1496,
    "batch_idx": 64,
    "loss": 0.3410288691520691,
    "learning_rate": 0.00014867501510057554,
    "timestep_mean": 419.125,
    "epsilon_norm": 7.736590385437012,
    "prediction_norm": 6.343600749969482,
    "timestamp": "2025-07-14T12:34:26.811253"
  },
  {
    "epoch": 1496,
    "batch_idx": 96,
    "loss": 0.27524334192276,
    "learning_rate": 0.00014867501510057554,
    "timestep_mean": 433.21875,
    "epsilon_norm": 7.592881202697754,
    "prediction_norm": 6.41566276550293,
    "timestamp": "2025-07-14T12:34:26.816522"
  },
  {
    "epoch": 1496,
    "batch_idx": 128,
    "loss": 0.2442474663257599,
    "learning_rate": 0.00014867501510057554,
    "timestep_mean": 575.84375,
    "epsilon_norm": 8.520734786987305,
    "prediction_norm": 7.08062219619751,
    "timestamp": "2025-07-14T12:34:26.821614"
  },
  {
    "epoch": 1496,
    "batch_idx": 160,
    "loss": 0.3224853575229645,
    "learning_rate": 0.00014867501510057554,
    "timestep_mean": 403.34375,
    "epsilon_norm": 7.418421745300293,
    "prediction_norm": 5.628566265106201,
    "timestamp": "2025-07-14T12:34:26.826622"
  },
  {
    "epoch": 1496,
    "batch_idx": 192,
    "loss": 0.4186212122440338,
    "learning_rate": 0.00014867501510057554,
    "timestep_mean": 398.0,
    "epsilon_norm": 4.424323081970215,
    "prediction_norm": 3.463097095489502,
    "timestamp": "2025-07-14T12:34:26.831699"
  },
  {
    "epoch": 1497,
    "batch_idx": 0,
    "loss": 0.3354477882385254,
    "learning_rate": 0.00014811660994605504,
    "timestep_mean": 472.78125,
    "epsilon_norm": 7.612988471984863,
    "prediction_norm": 6.600619792938232,
    "timestamp": "2025-07-14T12:34:26.836827"
  },
  {
    "epoch": 1497,
    "batch_idx": 32,
    "loss": 0.5784486532211304,
    "learning_rate": 0.00014811660994605504,
    "timestep_mean": 384.625,
    "epsilon_norm": 8.458465576171875,
    "prediction_norm": 6.075416564941406,
    "timestamp": "2025-07-14T12:34:26.842053"
  },
  {
    "epoch": 1497,
    "batch_idx": 64,
    "loss": 0.28921711444854736,
    "learning_rate": 0.00014811660994605504,
    "timestep_mean": 471.09375,
    "epsilon_norm": 8.346844673156738,
    "prediction_norm": 6.887002944946289,
    "timestamp": "2025-07-14T12:34:26.847076"
  },
  {
    "epoch": 1497,
    "batch_idx": 96,
    "loss": 0.23042258620262146,
    "learning_rate": 0.00014811660994605504,
    "timestep_mean": 519.0,
    "epsilon_norm": 7.529400825500488,
    "prediction_norm": 6.749044895172119,
    "timestamp": "2025-07-14T12:34:26.852215"
  },
  {
    "epoch": 1497,
    "batch_idx": 128,
    "loss": 0.16065962612628937,
    "learning_rate": 0.00014811660994605504,
    "timestep_mean": 555.3125,
    "epsilon_norm": 8.725621223449707,
    "prediction_norm": 8.21546745300293,
    "timestamp": "2025-07-14T12:34:26.857465"
  },
  {
    "epoch": 1497,
    "batch_idx": 160,
    "loss": 0.24129220843315125,
    "learning_rate": 0.00014811660994605504,
    "timestep_mean": 464.0625,
    "epsilon_norm": 7.968922138214111,
    "prediction_norm": 7.220156192779541,
    "timestamp": "2025-07-14T12:34:26.862769"
  },
  {
    "epoch": 1497,
    "batch_idx": 192,
    "loss": 0.16276663541793823,
    "learning_rate": 0.00014811660994605504,
    "timestep_mean": 447.625,
    "epsilon_norm": 4.470061302185059,
    "prediction_norm": 4.281826972961426,
    "timestamp": "2025-07-14T12:34:26.867878"
  },
  {
    "epoch": 1498,
    "batch_idx": 0,
    "loss": 0.15063418447971344,
    "learning_rate": 0.0001475590730288195,
    "timestep_mean": 449.6875,
    "epsilon_norm": 8.098750114440918,
    "prediction_norm": 7.627941131591797,
    "timestamp": "2025-07-14T12:34:26.873134"
  },
  {
    "epoch": 1498,
    "batch_idx": 32,
    "loss": 0.341445654630661,
    "learning_rate": 0.0001475590730288195,
    "timestep_mean": 474.5625,
    "epsilon_norm": 8.197368621826172,
    "prediction_norm": 6.677665710449219,
    "timestamp": "2025-07-14T12:34:26.878277"
  },
  {
    "epoch": 1498,
    "batch_idx": 64,
    "loss": 0.13486340641975403,
    "learning_rate": 0.0001475590730288195,
    "timestep_mean": 587.53125,
    "epsilon_norm": 8.388200759887695,
    "prediction_norm": 8.12509822845459,
    "timestamp": "2025-07-14T12:34:26.883467"
  },
  {
    "epoch": 1498,
    "batch_idx": 96,
    "loss": 0.18630115687847137,
    "learning_rate": 0.0001475590730288195,
    "timestep_mean": 525.6875,
    "epsilon_norm": 9.550641059875488,
    "prediction_norm": 8.297500610351562,
    "timestamp": "2025-07-14T12:34:26.888567"
  },
  {
    "epoch": 1498,
    "batch_idx": 128,
    "loss": 0.18456213176250458,
    "learning_rate": 0.0001475590730288195,
    "timestep_mean": 549.53125,
    "epsilon_norm": 8.252816200256348,
    "prediction_norm": 7.685605525970459,
    "timestamp": "2025-07-14T12:34:26.893522"
  },
  {
    "epoch": 1498,
    "batch_idx": 160,
    "loss": 0.26108062267303467,
    "learning_rate": 0.0001475590730288195,
    "timestep_mean": 454.6875,
    "epsilon_norm": 8.257312774658203,
    "prediction_norm": 6.807577133178711,
    "timestamp": "2025-07-14T12:34:26.898898"
  },
  {
    "epoch": 1498,
    "batch_idx": 192,
    "loss": 0.4585091471672058,
    "learning_rate": 0.0001475590730288195,
    "timestep_mean": 416.625,
    "epsilon_norm": 3.8733091354370117,
    "prediction_norm": 3.010066270828247,
    "timestamp": "2025-07-14T12:34:26.903993"
  },
  {
    "epoch": 1499,
    "batch_idx": 0,
    "loss": 0.2357696145772934,
    "learning_rate": 0.00014700240572453643,
    "timestep_mean": 509.21875,
    "epsilon_norm": 8.5767822265625,
    "prediction_norm": 7.525699138641357,
    "timestamp": "2025-07-14T12:34:26.909209"
  },
  {
    "epoch": 1499,
    "batch_idx": 32,
    "loss": 0.295674592256546,
    "learning_rate": 0.00014700240572453643,
    "timestep_mean": 483.84375,
    "epsilon_norm": 7.565877914428711,
    "prediction_norm": 6.407968997955322,
    "timestamp": "2025-07-14T12:34:26.914467"
  },
  {
    "epoch": 1499,
    "batch_idx": 64,
    "loss": 0.5244356989860535,
    "learning_rate": 0.00014700240572453643,
    "timestep_mean": 407.0,
    "epsilon_norm": 8.310248374938965,
    "prediction_norm": 5.954448699951172,
    "timestamp": "2025-07-14T12:34:26.919612"
  },
  {
    "epoch": 1499,
    "batch_idx": 96,
    "loss": 0.19021518528461456,
    "learning_rate": 0.00014700240572453643,
    "timestep_mean": 526.03125,
    "epsilon_norm": 8.013113021850586,
    "prediction_norm": 7.383687973022461,
    "timestamp": "2025-07-14T12:34:26.924890"
  },
  {
    "epoch": 1499,
    "batch_idx": 128,
    "loss": 0.22340308129787445,
    "learning_rate": 0.00014700240572453643,
    "timestep_mean": 503.96875,
    "epsilon_norm": 8.827637672424316,
    "prediction_norm": 7.415349960327148,
    "timestamp": "2025-07-14T12:34:26.930127"
  },
  {
    "epoch": 1499,
    "batch_idx": 160,
    "loss": 0.2141539305448532,
    "learning_rate": 0.00014700240572453643,
    "timestep_mean": 496.75,
    "epsilon_norm": 6.649877071380615,
    "prediction_norm": 5.477367877960205,
    "timestamp": "2025-07-14T12:34:26.935222"
  },
  {
    "epoch": 1499,
    "batch_idx": 192,
    "loss": 0.20475056767463684,
    "learning_rate": 0.00014700240572453643,
    "timestep_mean": 394.125,
    "epsilon_norm": 3.565214157104492,
    "prediction_norm": 3.123173236846924,
    "timestamp": "2025-07-14T12:34:26.940476"
  },
  {
    "epoch": 1500,
    "batch_idx": 0,
    "loss": 0.22314350306987762,
    "learning_rate": 0.00014644660940672647,
    "timestep_mean": 507.53125,
    "epsilon_norm": 7.78355073928833,
    "prediction_norm": 6.842938423156738,
    "timestamp": "2025-07-14T12:34:26.945651"
  },
  {
    "epoch": 1500,
    "batch_idx": 32,
    "loss": 0.11385466158390045,
    "learning_rate": 0.00014644660940672647,
    "timestep_mean": 554.96875,
    "epsilon_norm": 8.809121131896973,
    "prediction_norm": 8.160531997680664,
    "timestamp": "2025-07-14T12:34:26.950915"
  },
  {
    "epoch": 1500,
    "batch_idx": 64,
    "loss": 0.29257330298423767,
    "learning_rate": 0.00014644660940672647,
    "timestep_mean": 488.28125,
    "epsilon_norm": 8.666738510131836,
    "prediction_norm": 7.6548967361450195,
    "timestamp": "2025-07-14T12:34:26.956087"
  },
  {
    "epoch": 1500,
    "batch_idx": 96,
    "loss": 0.2815745770931244,
    "learning_rate": 0.00014644660940672647,
    "timestep_mean": 573.75,
    "epsilon_norm": 8.479904174804688,
    "prediction_norm": 7.319019317626953,
    "timestamp": "2025-07-14T12:34:26.961187"
  },
  {
    "epoch": 1500,
    "batch_idx": 128,
    "loss": 0.4625374376773834,
    "learning_rate": 0.00014644660940672647,
    "timestep_mean": 474.03125,
    "epsilon_norm": 9.42732048034668,
    "prediction_norm": 7.911520957946777,
    "timestamp": "2025-07-14T12:34:26.966379"
  },
  {
    "epoch": 1500,
    "batch_idx": 160,
    "loss": 0.19991567730903625,
    "learning_rate": 0.00014644660940672647,
    "timestep_mean": 518.65625,
    "epsilon_norm": 7.543951988220215,
    "prediction_norm": 6.923558235168457,
    "timestamp": "2025-07-14T12:34:26.971633"
  },
  {
    "epoch": 1500,
    "batch_idx": 192,
    "loss": 0.8916890621185303,
    "learning_rate": 0.00014644660940672647,
    "timestep_mean": 328.0,
    "epsilon_norm": 4.70428466796875,
    "prediction_norm": 2.322643280029297,
    "timestamp": "2025-07-14T12:34:26.976661"
  },
  {
    "epoch": 1501,
    "batch_idx": 0,
    "loss": 0.17669112980365753,
    "learning_rate": 0.00014589168544676208,
    "timestep_mean": 441.25,
    "epsilon_norm": 9.000993728637695,
    "prediction_norm": 8.197705268859863,
    "timestamp": "2025-07-14T12:34:26.981786"
  },
  {
    "epoch": 1501,
    "batch_idx": 32,
    "loss": 0.24673806130886078,
    "learning_rate": 0.00014589168544676208,
    "timestep_mean": 477.09375,
    "epsilon_norm": 10.045381546020508,
    "prediction_norm": 8.853665351867676,
    "timestamp": "2025-07-14T12:34:26.987071"
  },
  {
    "epoch": 1501,
    "batch_idx": 64,
    "loss": 0.41199803352355957,
    "learning_rate": 0.00014589168544676208,
    "timestep_mean": 466.6875,
    "epsilon_norm": 7.828338146209717,
    "prediction_norm": 6.249042510986328,
    "timestamp": "2025-07-14T12:34:26.992256"
  },
  {
    "epoch": 1501,
    "batch_idx": 96,
    "loss": 0.31956005096435547,
    "learning_rate": 0.00014589168544676208,
    "timestep_mean": 463.90625,
    "epsilon_norm": 7.407713890075684,
    "prediction_norm": 5.970884323120117,
    "timestamp": "2025-07-14T12:34:26.997416"
  },
  {
    "epoch": 1501,
    "batch_idx": 128,
    "loss": 0.20798353850841522,
    "learning_rate": 0.00014589168544676208,
    "timestep_mean": 475.9375,
    "epsilon_norm": 6.5981926918029785,
    "prediction_norm": 5.211274147033691,
    "timestamp": "2025-07-14T12:34:27.002559"
  },
  {
    "epoch": 1501,
    "batch_idx": 160,
    "loss": 0.17873990535736084,
    "learning_rate": 0.00014589168544676208,
    "timestep_mean": 479.78125,
    "epsilon_norm": 8.148359298706055,
    "prediction_norm": 7.135525226593018,
    "timestamp": "2025-07-14T12:34:27.007759"
  },
  {
    "epoch": 1501,
    "batch_idx": 192,
    "loss": 0.23217053711414337,
    "learning_rate": 0.00014589168544676208,
    "timestep_mean": 599.25,
    "epsilon_norm": 3.3202829360961914,
    "prediction_norm": 2.7836456298828125,
    "timestamp": "2025-07-14T12:34:27.012795"
  },
  {
    "epoch": 1502,
    "batch_idx": 0,
    "loss": 0.22386296093463898,
    "learning_rate": 0.00014533763521386336,
    "timestep_mean": 503.59375,
    "epsilon_norm": 8.807249069213867,
    "prediction_norm": 8.251286506652832,
    "timestamp": "2025-07-14T12:34:27.017646"
  },
  {
    "epoch": 1502,
    "batch_idx": 32,
    "loss": 0.28275445103645325,
    "learning_rate": 0.00014533763521386336,
    "timestep_mean": 469.65625,
    "epsilon_norm": 7.704739093780518,
    "prediction_norm": 6.865292072296143,
    "timestamp": "2025-07-14T12:34:27.022537"
  },
  {
    "epoch": 1502,
    "batch_idx": 64,
    "loss": 0.22007882595062256,
    "learning_rate": 0.00014533763521386336,
    "timestep_mean": 502.5625,
    "epsilon_norm": 7.16933012008667,
    "prediction_norm": 6.011160373687744,
    "timestamp": "2025-07-14T12:34:27.027491"
  },
  {
    "epoch": 1502,
    "batch_idx": 96,
    "loss": 0.23985141515731812,
    "learning_rate": 0.00014533763521386336,
    "timestep_mean": 527.34375,
    "epsilon_norm": 7.750632286071777,
    "prediction_norm": 6.801502704620361,
    "timestamp": "2025-07-14T12:34:27.032423"
  },
  {
    "epoch": 1502,
    "batch_idx": 128,
    "loss": 0.5312918424606323,
    "learning_rate": 0.00014533763521386336,
    "timestep_mean": 439.125,
    "epsilon_norm": 7.765377044677734,
    "prediction_norm": 5.495090961456299,
    "timestamp": "2025-07-14T12:34:27.037326"
  },
  {
    "epoch": 1502,
    "batch_idx": 160,
    "loss": 0.33584946393966675,
    "learning_rate": 0.00014533763521386336,
    "timestep_mean": 495.65625,
    "epsilon_norm": 9.042189598083496,
    "prediction_norm": 7.5837249755859375,
    "timestamp": "2025-07-14T12:34:27.042336"
  },
  {
    "epoch": 1502,
    "batch_idx": 192,
    "loss": 0.4538094401359558,
    "learning_rate": 0.00014533763521386336,
    "timestep_mean": 362.75,
    "epsilon_norm": 3.8107235431671143,
    "prediction_norm": 2.3382375240325928,
    "timestamp": "2025-07-14T12:34:27.047205"
  },
  {
    "epoch": 1503,
    "batch_idx": 0,
    "loss": 0.35020968317985535,
    "learning_rate": 0.00014478446007509346,
    "timestep_mean": 502.1875,
    "epsilon_norm": 8.243478775024414,
    "prediction_norm": 6.675475597381592,
    "timestamp": "2025-07-14T12:34:27.052090"
  },
  {
    "epoch": 1503,
    "batch_idx": 32,
    "loss": 0.6513991355895996,
    "learning_rate": 0.00014478446007509346,
    "timestep_mean": 495.6875,
    "epsilon_norm": 9.332056999206543,
    "prediction_norm": 7.12011194229126,
    "timestamp": "2025-07-14T12:34:27.057014"
  },
  {
    "epoch": 1503,
    "batch_idx": 64,
    "loss": 0.17866560816764832,
    "learning_rate": 0.00014478446007509346,
    "timestep_mean": 593.5625,
    "epsilon_norm": 8.042576789855957,
    "prediction_norm": 7.0008673667907715,
    "timestamp": "2025-07-14T12:34:27.061956"
  },
  {
    "epoch": 1503,
    "batch_idx": 96,
    "loss": 0.22663486003875732,
    "learning_rate": 0.00014478446007509346,
    "timestep_mean": 488.46875,
    "epsilon_norm": 7.606988430023193,
    "prediction_norm": 6.778780937194824,
    "timestamp": "2025-07-14T12:34:27.066759"
  },
  {
    "epoch": 1503,
    "batch_idx": 128,
    "loss": 0.33593568205833435,
    "learning_rate": 0.00014478446007509346,
    "timestep_mean": 537.1875,
    "epsilon_norm": 8.241480827331543,
    "prediction_norm": 6.672206878662109,
    "timestamp": "2025-07-14T12:34:27.071668"
  },
  {
    "epoch": 1503,
    "batch_idx": 160,
    "loss": 0.41484951972961426,
    "learning_rate": 0.00014478446007509346,
    "timestep_mean": 521.0625,
    "epsilon_norm": 8.432408332824707,
    "prediction_norm": 6.491523265838623,
    "timestamp": "2025-07-14T12:34:27.076478"
  },
  {
    "epoch": 1503,
    "batch_idx": 192,
    "loss": 0.2639455199241638,
    "learning_rate": 0.00014478446007509346,
    "timestep_mean": 492.125,
    "epsilon_norm": 2.9431281089782715,
    "prediction_norm": 2.1409692764282227,
    "timestamp": "2025-07-14T12:34:27.081376"
  },
  {
    "epoch": 1504,
    "batch_idx": 0,
    "loss": 0.520215630531311,
    "learning_rate": 0.00014423216139535753,
    "timestep_mean": 409.28125,
    "epsilon_norm": 9.125398635864258,
    "prediction_norm": 6.623836994171143,
    "timestamp": "2025-07-14T12:34:27.086224"
  },
  {
    "epoch": 1504,
    "batch_idx": 32,
    "loss": 0.27768251299858093,
    "learning_rate": 0.00014423216139535753,
    "timestep_mean": 434.1875,
    "epsilon_norm": 7.613230228424072,
    "prediction_norm": 5.650190353393555,
    "timestamp": "2025-07-14T12:34:27.091087"
  },
  {
    "epoch": 1504,
    "batch_idx": 64,
    "loss": 0.21508212387561798,
    "learning_rate": 0.00014423216139535753,
    "timestep_mean": 487.71875,
    "epsilon_norm": 6.61776876449585,
    "prediction_norm": 6.487459659576416,
    "timestamp": "2025-07-14T12:34:27.095911"
  },
  {
    "epoch": 1504,
    "batch_idx": 96,
    "loss": 0.27696123719215393,
    "learning_rate": 0.00014423216139535753,
    "timestep_mean": 370.34375,
    "epsilon_norm": 7.650177478790283,
    "prediction_norm": 6.1677422523498535,
    "timestamp": "2025-07-14T12:34:27.100842"
  },
  {
    "epoch": 1504,
    "batch_idx": 128,
    "loss": 0.35378551483154297,
    "learning_rate": 0.00014423216139535753,
    "timestep_mean": 460.53125,
    "epsilon_norm": 8.01500415802002,
    "prediction_norm": 6.792067527770996,
    "timestamp": "2025-07-14T12:34:27.105691"
  },
  {
    "epoch": 1504,
    "batch_idx": 160,
    "loss": 0.2678676247596741,
    "learning_rate": 0.00014423216139535753,
    "timestep_mean": 446.15625,
    "epsilon_norm": 7.491172790527344,
    "prediction_norm": 6.732769012451172,
    "timestamp": "2025-07-14T12:34:27.110553"
  },
  {
    "epoch": 1504,
    "batch_idx": 192,
    "loss": 0.36759623885154724,
    "learning_rate": 0.00014423216139535753,
    "timestep_mean": 462.125,
    "epsilon_norm": 4.545816421508789,
    "prediction_norm": 3.7974841594696045,
    "timestamp": "2025-07-14T12:34:27.115386"
  },
  {
    "epoch": 1505,
    "batch_idx": 0,
    "loss": 0.5041103959083557,
    "learning_rate": 0.00014368074053739752,
    "timestep_mean": 504.96875,
    "epsilon_norm": 7.740848064422607,
    "prediction_norm": 5.959938049316406,
    "timestamp": "2025-07-14T12:34:27.120224"
  },
  {
    "epoch": 1505,
    "batch_idx": 32,
    "loss": 0.23321057856082916,
    "learning_rate": 0.00014368074053739752,
    "timestep_mean": 485.125,
    "epsilon_norm": 8.407917022705078,
    "prediction_norm": 7.268428325653076,
    "timestamp": "2025-07-14T12:34:27.125109"
  },
  {
    "epoch": 1505,
    "batch_idx": 64,
    "loss": 0.1770535111427307,
    "learning_rate": 0.00014368074053739752,
    "timestep_mean": 520.375,
    "epsilon_norm": 6.069403648376465,
    "prediction_norm": 5.362661361694336,
    "timestamp": "2025-07-14T12:34:27.129995"
  },
  {
    "epoch": 1505,
    "batch_idx": 96,
    "loss": 0.2869747281074524,
    "learning_rate": 0.00014368074053739752,
    "timestep_mean": 483.4375,
    "epsilon_norm": 7.465878009796143,
    "prediction_norm": 6.020035266876221,
    "timestamp": "2025-07-14T12:34:27.134730"
  },
  {
    "epoch": 1505,
    "batch_idx": 128,
    "loss": 0.4836607575416565,
    "learning_rate": 0.00014368074053739752,
    "timestep_mean": 425.90625,
    "epsilon_norm": 8.070865631103516,
    "prediction_norm": 6.454196453094482,
    "timestamp": "2025-07-14T12:34:27.139583"
  },
  {
    "epoch": 1505,
    "batch_idx": 160,
    "loss": 0.7860289812088013,
    "learning_rate": 0.00014368074053739752,
    "timestep_mean": 501.25,
    "epsilon_norm": 9.373808860778809,
    "prediction_norm": 6.698678493499756,
    "timestamp": "2025-07-14T12:34:27.144413"
  },
  {
    "epoch": 1505,
    "batch_idx": 192,
    "loss": 0.5879435539245605,
    "learning_rate": 0.00014368074053739752,
    "timestep_mean": 584.625,
    "epsilon_norm": 5.92584753036499,
    "prediction_norm": 5.424526691436768,
    "timestamp": "2025-07-14T12:34:27.149233"
  },
  {
    "epoch": 1506,
    "batch_idx": 0,
    "loss": 0.08080092817544937,
    "learning_rate": 0.00014313019886178966,
    "timestep_mean": 644.46875,
    "epsilon_norm": 7.0986127853393555,
    "prediction_norm": 6.630572319030762,
    "timestamp": "2025-07-14T12:34:27.154064"
  },
  {
    "epoch": 1506,
    "batch_idx": 32,
    "loss": 0.27228158712387085,
    "learning_rate": 0.00014313019886178966,
    "timestep_mean": 478.3125,
    "epsilon_norm": 7.40598726272583,
    "prediction_norm": 6.15627384185791,
    "timestamp": "2025-07-14T12:34:27.159023"
  },
  {
    "epoch": 1506,
    "batch_idx": 64,
    "loss": 0.35386019945144653,
    "learning_rate": 0.00014313019886178966,
    "timestep_mean": 481.09375,
    "epsilon_norm": 7.887515544891357,
    "prediction_norm": 6.564510822296143,
    "timestamp": "2025-07-14T12:34:27.163854"
  },
  {
    "epoch": 1506,
    "batch_idx": 96,
    "loss": 0.2813870906829834,
    "learning_rate": 0.00014313019886178966,
    "timestep_mean": 486.75,
    "epsilon_norm": 8.062335968017578,
    "prediction_norm": 7.140642166137695,
    "timestamp": "2025-07-14T12:34:27.168719"
  },
  {
    "epoch": 1506,
    "batch_idx": 128,
    "loss": 0.31025007367134094,
    "learning_rate": 0.00014313019886178966,
    "timestep_mean": 437.0625,
    "epsilon_norm": 8.733869552612305,
    "prediction_norm": 7.433377265930176,
    "timestamp": "2025-07-14T12:34:27.173564"
  },
  {
    "epoch": 1506,
    "batch_idx": 160,
    "loss": 0.13421641290187836,
    "learning_rate": 0.00014313019886178966,
    "timestep_mean": 546.25,
    "epsilon_norm": 8.286842346191406,
    "prediction_norm": 7.99972677230835,
    "timestamp": "2025-07-14T12:34:27.178448"
  },
  {
    "epoch": 1506,
    "batch_idx": 192,
    "loss": 0.09879232943058014,
    "learning_rate": 0.00014313019886178966,
    "timestep_mean": 589.5,
    "epsilon_norm": 4.65008020401001,
    "prediction_norm": 4.525362491607666,
    "timestamp": "2025-07-14T12:34:27.183383"
  },
  {
    "epoch": 1507,
    "batch_idx": 0,
    "loss": 0.13043756783008575,
    "learning_rate": 0.00014258053772694048,
    "timestep_mean": 464.625,
    "epsilon_norm": 7.055624961853027,
    "prediction_norm": 6.314696311950684,
    "timestamp": "2025-07-14T12:34:27.188212"
  },
  {
    "epoch": 1507,
    "batch_idx": 32,
    "loss": 0.34700506925582886,
    "learning_rate": 0.00014258053772694048,
    "timestep_mean": 357.25,
    "epsilon_norm": 7.648820877075195,
    "prediction_norm": 6.7015790939331055,
    "timestamp": "2025-07-14T12:34:27.193622"
  },
  {
    "epoch": 1507,
    "batch_idx": 64,
    "loss": 0.14755979180335999,
    "learning_rate": 0.00014258053772694048,
    "timestep_mean": 503.5625,
    "epsilon_norm": 7.340090274810791,
    "prediction_norm": 7.217635631561279,
    "timestamp": "2025-07-14T12:34:27.198827"
  },
  {
    "epoch": 1507,
    "batch_idx": 96,
    "loss": 0.21949198842048645,
    "learning_rate": 0.00014258053772694048,
    "timestep_mean": 459.71875,
    "epsilon_norm": 7.181140422821045,
    "prediction_norm": 6.171482563018799,
    "timestamp": "2025-07-14T12:34:27.203892"
  },
  {
    "epoch": 1507,
    "batch_idx": 128,
    "loss": 0.611067533493042,
    "learning_rate": 0.00014258053772694048,
    "timestep_mean": 586.0625,
    "epsilon_norm": 10.085357666015625,
    "prediction_norm": 8.169727325439453,
    "timestamp": "2025-07-14T12:34:27.208845"
  },
  {
    "epoch": 1507,
    "batch_idx": 160,
    "loss": 0.3237835168838501,
    "learning_rate": 0.00014258053772694048,
    "timestep_mean": 525.25,
    "epsilon_norm": 7.760812759399414,
    "prediction_norm": 6.4169769287109375,
    "timestamp": "2025-07-14T12:34:27.214796"
  },
  {
    "epoch": 1507,
    "batch_idx": 192,
    "loss": 0.23935288190841675,
    "learning_rate": 0.00014258053772694048,
    "timestep_mean": 423.5,
    "epsilon_norm": 4.405044078826904,
    "prediction_norm": 3.3727517127990723,
    "timestamp": "2025-07-14T12:34:27.224292"
  },
  {
    "epoch": 1508,
    "batch_idx": 0,
    "loss": 0.2929297387599945,
    "learning_rate": 0.0001420317584890846,
    "timestep_mean": 531.03125,
    "epsilon_norm": 7.758559703826904,
    "prediction_norm": 6.211061477661133,
    "timestamp": "2025-07-14T12:34:27.229630"
  },
  {
    "epoch": 1508,
    "batch_idx": 32,
    "loss": 0.38739579916000366,
    "learning_rate": 0.0001420317584890846,
    "timestep_mean": 446.96875,
    "epsilon_norm": 8.515706062316895,
    "prediction_norm": 6.807076454162598,
    "timestamp": "2025-07-14T12:34:27.234550"
  },
  {
    "epoch": 1508,
    "batch_idx": 64,
    "loss": 0.2511710822582245,
    "learning_rate": 0.0001420317584890846,
    "timestep_mean": 473.5625,
    "epsilon_norm": 7.659426212310791,
    "prediction_norm": 6.021501064300537,
    "timestamp": "2025-07-14T12:34:27.239451"
  },
  {
    "epoch": 1508,
    "batch_idx": 96,
    "loss": 0.34024274349212646,
    "learning_rate": 0.0001420317584890846,
    "timestep_mean": 422.96875,
    "epsilon_norm": 7.39661169052124,
    "prediction_norm": 6.552670001983643,
    "timestamp": "2025-07-14T12:34:27.244259"
  },
  {
    "epoch": 1508,
    "batch_idx": 128,
    "loss": 0.17464782297611237,
    "learning_rate": 0.0001420317584890846,
    "timestep_mean": 490.75,
    "epsilon_norm": 7.564778804779053,
    "prediction_norm": 6.808134078979492,
    "timestamp": "2025-07-14T12:34:27.249179"
  },
  {
    "epoch": 1508,
    "batch_idx": 160,
    "loss": 0.35587435960769653,
    "learning_rate": 0.0001420317584890846,
    "timestep_mean": 504.53125,
    "epsilon_norm": 9.084885597229004,
    "prediction_norm": 7.401549339294434,
    "timestamp": "2025-07-14T12:34:27.254376"
  },
  {
    "epoch": 1508,
    "batch_idx": 192,
    "loss": 0.1033487319946289,
    "learning_rate": 0.0001420317584890846,
    "timestep_mean": 441.375,
    "epsilon_norm": 3.9569790363311768,
    "prediction_norm": 3.9257569313049316,
    "timestamp": "2025-07-14T12:34:27.259702"
  },
  {
    "epoch": 1509,
    "batch_idx": 0,
    "loss": 0.20536133646965027,
    "learning_rate": 0.0001414838625022801,
    "timestep_mean": 574.625,
    "epsilon_norm": 7.330162525177002,
    "prediction_norm": 6.265369892120361,
    "timestamp": "2025-07-14T12:34:27.264946"
  },
  {
    "epoch": 1509,
    "batch_idx": 32,
    "loss": 0.1598346084356308,
    "learning_rate": 0.0001414838625022801,
    "timestep_mean": 495.96875,
    "epsilon_norm": 8.65621280670166,
    "prediction_norm": 8.070067405700684,
    "timestamp": "2025-07-14T12:34:27.270134"
  },
  {
    "epoch": 1509,
    "batch_idx": 64,
    "loss": 0.34898602962493896,
    "learning_rate": 0.0001414838625022801,
    "timestep_mean": 523.8125,
    "epsilon_norm": 8.9025239944458,
    "prediction_norm": 8.228323936462402,
    "timestamp": "2025-07-14T12:34:27.275515"
  },
  {
    "epoch": 1509,
    "batch_idx": 96,
    "loss": 0.2127939760684967,
    "learning_rate": 0.0001414838625022801,
    "timestep_mean": 530.4375,
    "epsilon_norm": 7.076028823852539,
    "prediction_norm": 5.946902751922607,
    "timestamp": "2025-07-14T12:34:27.280569"
  },
  {
    "epoch": 1509,
    "batch_idx": 128,
    "loss": 0.1589004248380661,
    "learning_rate": 0.0001414838625022801,
    "timestep_mean": 530.46875,
    "epsilon_norm": 8.489656448364258,
    "prediction_norm": 7.505453109741211,
    "timestamp": "2025-07-14T12:34:27.285689"
  },
  {
    "epoch": 1509,
    "batch_idx": 160,
    "loss": 0.1306159645318985,
    "learning_rate": 0.0001414838625022801,
    "timestep_mean": 475.96875,
    "epsilon_norm": 6.529629707336426,
    "prediction_norm": 5.916841983795166,
    "timestamp": "2025-07-14T12:34:27.290875"
  },
  {
    "epoch": 1509,
    "batch_idx": 192,
    "loss": 0.13055306673049927,
    "learning_rate": 0.0001414838625022801,
    "timestep_mean": 369.125,
    "epsilon_norm": 3.3920552730560303,
    "prediction_norm": 3.3930697441101074,
    "timestamp": "2025-07-14T12:34:27.296037"
  },
  {
    "epoch": 1510,
    "batch_idx": 0,
    "loss": 0.23044124245643616,
    "learning_rate": 0.00014093685111840585,
    "timestep_mean": 539.0,
    "epsilon_norm": 6.637691020965576,
    "prediction_norm": 5.9362382888793945,
    "timestamp": "2025-07-14T12:34:27.301179"
  },
  {
    "epoch": 1510,
    "batch_idx": 32,
    "loss": 0.23410290479660034,
    "learning_rate": 0.00014093685111840585,
    "timestep_mean": 525.8125,
    "epsilon_norm": 7.646151065826416,
    "prediction_norm": 6.64569091796875,
    "timestamp": "2025-07-14T12:34:27.306492"
  },
  {
    "epoch": 1510,
    "batch_idx": 64,
    "loss": 0.5839003324508667,
    "learning_rate": 0.00014093685111840585,
    "timestep_mean": 451.21875,
    "epsilon_norm": 8.135199546813965,
    "prediction_norm": 5.003599643707275,
    "timestamp": "2025-07-14T12:34:27.311708"
  },
  {
    "epoch": 1510,
    "batch_idx": 96,
    "loss": 0.1540956199169159,
    "learning_rate": 0.00014093685111840585,
    "timestep_mean": 539.5625,
    "epsilon_norm": 7.429139614105225,
    "prediction_norm": 6.692654132843018,
    "timestamp": "2025-07-14T12:34:27.316996"
  },
  {
    "epoch": 1510,
    "batch_idx": 128,
    "loss": 0.2114349752664566,
    "learning_rate": 0.00014093685111840585,
    "timestep_mean": 461.8125,
    "epsilon_norm": 6.92752742767334,
    "prediction_norm": 5.904603004455566,
    "timestamp": "2025-07-14T12:34:27.322008"
  },
  {
    "epoch": 1510,
    "batch_idx": 160,
    "loss": 0.38639938831329346,
    "learning_rate": 0.00014093685111840585,
    "timestep_mean": 501.0,
    "epsilon_norm": 8.677825927734375,
    "prediction_norm": 6.519109725952148,
    "timestamp": "2025-07-14T12:34:27.327112"
  },
  {
    "epoch": 1510,
    "batch_idx": 192,
    "loss": 0.15279823541641235,
    "learning_rate": 0.00014093685111840585,
    "timestep_mean": 538.125,
    "epsilon_norm": 4.440991401672363,
    "prediction_norm": 3.6669998168945312,
    "timestamp": "2025-07-14T12:34:27.332341"
  },
  {
    "epoch": 1511,
    "batch_idx": 0,
    "loss": 0.13535936176776886,
    "learning_rate": 0.00014039072568715802,
    "timestep_mean": 530.0,
    "epsilon_norm": 7.52351188659668,
    "prediction_norm": 6.3870978355407715,
    "timestamp": "2025-07-14T12:34:27.337346"
  },
  {
    "epoch": 1511,
    "batch_idx": 32,
    "loss": 0.3534112274646759,
    "learning_rate": 0.00014039072568715802,
    "timestep_mean": 542.53125,
    "epsilon_norm": 9.095527648925781,
    "prediction_norm": 8.141545295715332,
    "timestamp": "2025-07-14T12:34:27.342333"
  },
  {
    "epoch": 1511,
    "batch_idx": 64,
    "loss": 0.23949883878231049,
    "learning_rate": 0.00014039072568715802,
    "timestep_mean": 524.125,
    "epsilon_norm": 8.580540657043457,
    "prediction_norm": 7.200521469116211,
    "timestamp": "2025-07-14T12:34:27.347427"
  },
  {
    "epoch": 1511,
    "batch_idx": 96,
    "loss": 0.20577000081539154,
    "learning_rate": 0.00014039072568715802,
    "timestep_mean": 527.375,
    "epsilon_norm": 7.882666110992432,
    "prediction_norm": 6.7770538330078125,
    "timestamp": "2025-07-14T12:34:27.352616"
  },
  {
    "epoch": 1511,
    "batch_idx": 128,
    "loss": 0.38806870579719543,
    "learning_rate": 0.00014039072568715802,
    "timestep_mean": 475.5,
    "epsilon_norm": 8.681788444519043,
    "prediction_norm": 6.989880561828613,
    "timestamp": "2025-07-14T12:34:27.357812"
  },
  {
    "epoch": 1511,
    "batch_idx": 160,
    "loss": 0.33360129594802856,
    "learning_rate": 0.00014039072568715802,
    "timestep_mean": 437.125,
    "epsilon_norm": 7.89486026763916,
    "prediction_norm": 6.248081207275391,
    "timestamp": "2025-07-14T12:34:27.363059"
  },
  {
    "epoch": 1511,
    "batch_idx": 192,
    "loss": 0.08079014718532562,
    "learning_rate": 0.00014039072568715802,
    "timestep_mean": 447.0,
    "epsilon_norm": 3.7415215969085693,
    "prediction_norm": 3.433823823928833,
    "timestamp": "2025-07-14T12:34:27.368028"
  },
  {
    "epoch": 1512,
    "batch_idx": 0,
    "loss": 0.1528037190437317,
    "learning_rate": 0.00013984548755604674,
    "timestep_mean": 572.65625,
    "epsilon_norm": 7.633583068847656,
    "prediction_norm": 6.657185077667236,
    "timestamp": "2025-07-14T12:34:27.373264"
  },
  {
    "epoch": 1512,
    "batch_idx": 32,
    "loss": 0.2479424923658371,
    "learning_rate": 0.00013984548755604674,
    "timestep_mean": 468.0625,
    "epsilon_norm": 8.051753997802734,
    "prediction_norm": 7.0829057693481445,
    "timestamp": "2025-07-14T12:34:27.378366"
  },
  {
    "epoch": 1512,
    "batch_idx": 64,
    "loss": 0.4594513177871704,
    "learning_rate": 0.00013984548755604674,
    "timestep_mean": 445.78125,
    "epsilon_norm": 9.259527206420898,
    "prediction_norm": 7.136366844177246,
    "timestamp": "2025-07-14T12:34:27.383623"
  },
  {
    "epoch": 1512,
    "batch_idx": 96,
    "loss": 0.24811838567256927,
    "learning_rate": 0.00013984548755604674,
    "timestep_mean": 504.375,
    "epsilon_norm": 7.460493564605713,
    "prediction_norm": 6.531722545623779,
    "timestamp": "2025-07-14T12:34:27.388739"
  },
  {
    "epoch": 1512,
    "batch_idx": 128,
    "loss": 0.28498053550720215,
    "learning_rate": 0.00013984548755604674,
    "timestep_mean": 488.75,
    "epsilon_norm": 7.2523956298828125,
    "prediction_norm": 5.996647357940674,
    "timestamp": "2025-07-14T12:34:27.393755"
  },
  {
    "epoch": 1512,
    "batch_idx": 160,
    "loss": 0.22209498286247253,
    "learning_rate": 0.00013984548755604674,
    "timestep_mean": 504.03125,
    "epsilon_norm": 8.167841911315918,
    "prediction_norm": 7.2182769775390625,
    "timestamp": "2025-07-14T12:34:27.398985"
  },
  {
    "epoch": 1512,
    "batch_idx": 192,
    "loss": 0.31654223799705505,
    "learning_rate": 0.00013984548755604674,
    "timestep_mean": 356.75,
    "epsilon_norm": 3.4789490699768066,
    "prediction_norm": 2.254802703857422,
    "timestamp": "2025-07-14T12:34:27.404130"
  },
  {
    "epoch": 1513,
    "batch_idx": 0,
    "loss": 0.12576326727867126,
    "learning_rate": 0.00013930113807039327,
    "timestep_mean": 515.0,
    "epsilon_norm": 6.288127422332764,
    "prediction_norm": 5.843965530395508,
    "timestamp": "2025-07-14T12:34:27.409037"
  },
  {
    "epoch": 1513,
    "batch_idx": 32,
    "loss": 0.5321088433265686,
    "learning_rate": 0.00013930113807039327,
    "timestep_mean": 447.78125,
    "epsilon_norm": 7.715875625610352,
    "prediction_norm": 5.290998458862305,
    "timestamp": "2025-07-14T12:34:27.413934"
  },
  {
    "epoch": 1513,
    "batch_idx": 64,
    "loss": 0.34945785999298096,
    "learning_rate": 0.00013930113807039327,
    "timestep_mean": 494.8125,
    "epsilon_norm": 7.95165491104126,
    "prediction_norm": 6.43593692779541,
    "timestamp": "2025-07-14T12:34:27.419126"
  },
  {
    "epoch": 1513,
    "batch_idx": 96,
    "loss": 0.21202248334884644,
    "learning_rate": 0.00013930113807039327,
    "timestep_mean": 499.09375,
    "epsilon_norm": 8.064066886901855,
    "prediction_norm": 7.220168113708496,
    "timestamp": "2025-07-14T12:34:27.424299"
  },
  {
    "epoch": 1513,
    "batch_idx": 128,
    "loss": 0.20754870772361755,
    "learning_rate": 0.00013930113807039327,
    "timestep_mean": 560.25,
    "epsilon_norm": 7.883674621582031,
    "prediction_norm": 6.575930118560791,
    "timestamp": "2025-07-14T12:34:27.429391"
  },
  {
    "epoch": 1513,
    "batch_idx": 160,
    "loss": 0.31362414360046387,
    "learning_rate": 0.00013930113807039327,
    "timestep_mean": 508.5625,
    "epsilon_norm": 8.06590461730957,
    "prediction_norm": 6.888626575469971,
    "timestamp": "2025-07-14T12:34:27.434376"
  },
  {
    "epoch": 1513,
    "batch_idx": 192,
    "loss": 0.11295740306377411,
    "learning_rate": 0.00013930113807039327,
    "timestep_mean": 573.625,
    "epsilon_norm": 5.6461405754089355,
    "prediction_norm": 5.174218654632568,
    "timestamp": "2025-07-14T12:34:27.439449"
  },
  {
    "epoch": 1514,
    "batch_idx": 0,
    "loss": 0.3292079567909241,
    "learning_rate": 0.0001387576785733253,
    "timestep_mean": 467.1875,
    "epsilon_norm": 7.455857753753662,
    "prediction_norm": 5.679219722747803,
    "timestamp": "2025-07-14T12:34:27.444464"
  },
  {
    "epoch": 1514,
    "batch_idx": 32,
    "loss": 0.212492436170578,
    "learning_rate": 0.0001387576785733253,
    "timestep_mean": 562.21875,
    "epsilon_norm": 8.742144584655762,
    "prediction_norm": 7.655605792999268,
    "timestamp": "2025-07-14T12:34:27.449339"
  },
  {
    "epoch": 1514,
    "batch_idx": 64,
    "loss": 0.1753644347190857,
    "learning_rate": 0.0001387576785733253,
    "timestep_mean": 594.09375,
    "epsilon_norm": 7.370971202850342,
    "prediction_norm": 7.0403642654418945,
    "timestamp": "2025-07-14T12:34:27.454233"
  },
  {
    "epoch": 1514,
    "batch_idx": 96,
    "loss": 0.2668038606643677,
    "learning_rate": 0.0001387576785733253,
    "timestep_mean": 537.71875,
    "epsilon_norm": 9.20019817352295,
    "prediction_norm": 8.088737487792969,
    "timestamp": "2025-07-14T12:34:27.459198"
  },
  {
    "epoch": 1514,
    "batch_idx": 128,
    "loss": 0.38171446323394775,
    "learning_rate": 0.0001387576785733253,
    "timestep_mean": 464.40625,
    "epsilon_norm": 7.688750743865967,
    "prediction_norm": 6.0597004890441895,
    "timestamp": "2025-07-14T12:34:27.464042"
  },
  {
    "epoch": 1514,
    "batch_idx": 160,
    "loss": 0.1664298176765442,
    "learning_rate": 0.0001387576785733253,
    "timestep_mean": 484.53125,
    "epsilon_norm": 6.802198886871338,
    "prediction_norm": 6.000297546386719,
    "timestamp": "2025-07-14T12:34:27.468913"
  },
  {
    "epoch": 1514,
    "batch_idx": 192,
    "loss": 0.08193013072013855,
    "learning_rate": 0.0001387576785733253,
    "timestep_mean": 617.75,
    "epsilon_norm": 2.6057164669036865,
    "prediction_norm": 2.552297353744507,
    "timestamp": "2025-07-14T12:34:27.473772"
  },
  {
    "epoch": 1515,
    "batch_idx": 0,
    "loss": 0.23670703172683716,
    "learning_rate": 0.00013821511040577547,
    "timestep_mean": 583.15625,
    "epsilon_norm": 9.348785400390625,
    "prediction_norm": 8.390782356262207,
    "timestamp": "2025-07-14T12:34:27.478678"
  },
  {
    "epoch": 1515,
    "batch_idx": 32,
    "loss": 0.10786738991737366,
    "learning_rate": 0.00013821511040577547,
    "timestep_mean": 540.15625,
    "epsilon_norm": 7.97020959854126,
    "prediction_norm": 7.415741443634033,
    "timestamp": "2025-07-14T12:34:27.483636"
  },
  {
    "epoch": 1515,
    "batch_idx": 64,
    "loss": 0.1427946388721466,
    "learning_rate": 0.00013821511040577547,
    "timestep_mean": 539.9375,
    "epsilon_norm": 6.681293487548828,
    "prediction_norm": 6.031439304351807,
    "timestamp": "2025-07-14T12:34:27.488501"
  },
  {
    "epoch": 1515,
    "batch_idx": 96,
    "loss": 0.3225046992301941,
    "learning_rate": 0.00013821511040577547,
    "timestep_mean": 451.59375,
    "epsilon_norm": 7.201893329620361,
    "prediction_norm": 6.189916133880615,
    "timestamp": "2025-07-14T12:34:27.493454"
  },
  {
    "epoch": 1515,
    "batch_idx": 128,
    "loss": 0.3796287477016449,
    "learning_rate": 0.00013821511040577547,
    "timestep_mean": 450.03125,
    "epsilon_norm": 8.009305000305176,
    "prediction_norm": 6.5593414306640625,
    "timestamp": "2025-07-14T12:34:27.498314"
  },
  {
    "epoch": 1515,
    "batch_idx": 160,
    "loss": 0.15865713357925415,
    "learning_rate": 0.00013821511040577547,
    "timestep_mean": 479.0625,
    "epsilon_norm": 6.868008613586426,
    "prediction_norm": 6.337432861328125,
    "timestamp": "2025-07-14T12:34:27.503166"
  },
  {
    "epoch": 1515,
    "batch_idx": 192,
    "loss": 0.4958122670650482,
    "learning_rate": 0.00013821511040577547,
    "timestep_mean": 560.375,
    "epsilon_norm": 3.9425148963928223,
    "prediction_norm": 2.971304416656494,
    "timestamp": "2025-07-14T12:34:27.508034"
  },
  {
    "epoch": 1516,
    "batch_idx": 0,
    "loss": 0.27075597643852234,
    "learning_rate": 0.00013767343490647696,
    "timestep_mean": 506.34375,
    "epsilon_norm": 8.83152961730957,
    "prediction_norm": 7.329629898071289,
    "timestamp": "2025-07-14T12:34:27.513003"
  },
  {
    "epoch": 1516,
    "batch_idx": 32,
    "loss": 0.29315680265426636,
    "learning_rate": 0.00013767343490647696,
    "timestep_mean": 433.125,
    "epsilon_norm": 7.338226318359375,
    "prediction_norm": 5.9584269523620605,
    "timestamp": "2025-07-14T12:34:27.517823"
  },
  {
    "epoch": 1516,
    "batch_idx": 64,
    "loss": 0.18201151490211487,
    "learning_rate": 0.00013767343490647696,
    "timestep_mean": 461.9375,
    "epsilon_norm": 7.204431056976318,
    "prediction_norm": 6.362078666687012,
    "timestamp": "2025-07-14T12:34:27.522703"
  },
  {
    "epoch": 1516,
    "batch_idx": 96,
    "loss": 0.21475917100906372,
    "learning_rate": 0.00013767343490647696,
    "timestep_mean": 562.125,
    "epsilon_norm": 7.49621057510376,
    "prediction_norm": 6.456684589385986,
    "timestamp": "2025-07-14T12:34:27.527561"
  },
  {
    "epoch": 1516,
    "batch_idx": 128,
    "loss": 0.1409062296152115,
    "learning_rate": 0.00013767343490647696,
    "timestep_mean": 507.8125,
    "epsilon_norm": 7.581649303436279,
    "prediction_norm": 7.239383697509766,
    "timestamp": "2025-07-14T12:34:27.532398"
  },
  {
    "epoch": 1516,
    "batch_idx": 160,
    "loss": 0.2054748386144638,
    "learning_rate": 0.00013767343490647696,
    "timestep_mean": 501.78125,
    "epsilon_norm": 7.476093292236328,
    "prediction_norm": 6.63771390914917,
    "timestamp": "2025-07-14T12:34:27.537275"
  },
  {
    "epoch": 1516,
    "batch_idx": 192,
    "loss": 0.14026254415512085,
    "learning_rate": 0.00013767343490647696,
    "timestep_mean": 530.875,
    "epsilon_norm": 3.6117289066314697,
    "prediction_norm": 3.194061517715454,
    "timestamp": "2025-07-14T12:34:27.542177"
  },
  {
    "epoch": 1517,
    "batch_idx": 0,
    "loss": 0.2487199306488037,
    "learning_rate": 0.00013713265341195975,
    "timestep_mean": 522.0625,
    "epsilon_norm": 7.985857009887695,
    "prediction_norm": 6.682767391204834,
    "timestamp": "2025-07-14T12:34:27.547075"
  },
  {
    "epoch": 1517,
    "batch_idx": 32,
    "loss": 0.17272117733955383,
    "learning_rate": 0.00013713265341195975,
    "timestep_mean": 456.15625,
    "epsilon_norm": 7.708151817321777,
    "prediction_norm": 6.938162326812744,
    "timestamp": "2025-07-14T12:34:27.551904"
  },
  {
    "epoch": 1517,
    "batch_idx": 64,
    "loss": 0.5901700258255005,
    "learning_rate": 0.00013713265341195975,
    "timestep_mean": 519.71875,
    "epsilon_norm": 8.734862327575684,
    "prediction_norm": 6.585782527923584,
    "timestamp": "2025-07-14T12:34:27.556750"
  },
  {
    "epoch": 1517,
    "batch_idx": 96,
    "loss": 0.18075495958328247,
    "learning_rate": 0.00013713265341195975,
    "timestep_mean": 544.28125,
    "epsilon_norm": 7.556034088134766,
    "prediction_norm": 6.672931671142578,
    "timestamp": "2025-07-14T12:34:27.561643"
  },
  {
    "epoch": 1517,
    "batch_idx": 128,
    "loss": 0.28424516320228577,
    "learning_rate": 0.00013713265341195975,
    "timestep_mean": 516.125,
    "epsilon_norm": 7.706867694854736,
    "prediction_norm": 6.869451522827148,
    "timestamp": "2025-07-14T12:34:27.566516"
  },
  {
    "epoch": 1517,
    "batch_idx": 160,
    "loss": 0.19567465782165527,
    "learning_rate": 0.00013713265341195975,
    "timestep_mean": 425.46875,
    "epsilon_norm": 7.209308624267578,
    "prediction_norm": 6.473451137542725,
    "timestamp": "2025-07-14T12:34:27.571508"
  },
  {
    "epoch": 1517,
    "batch_idx": 192,
    "loss": 0.15565937757492065,
    "learning_rate": 0.00013713265341195975,
    "timestep_mean": 437.375,
    "epsilon_norm": 4.012519836425781,
    "prediction_norm": 4.196213722229004,
    "timestamp": "2025-07-14T12:34:27.576389"
  },
  {
    "epoch": 1518,
    "batch_idx": 0,
    "loss": 0.1904793679714203,
    "learning_rate": 0.0001365927672565488,
    "timestep_mean": 473.03125,
    "epsilon_norm": 8.762445449829102,
    "prediction_norm": 7.07110595703125,
    "timestamp": "2025-07-14T12:34:27.581256"
  },
  {
    "epoch": 1518,
    "batch_idx": 32,
    "loss": 0.34837719798088074,
    "learning_rate": 0.0001365927672565488,
    "timestep_mean": 461.75,
    "epsilon_norm": 8.833869934082031,
    "prediction_norm": 6.888210773468018,
    "timestamp": "2025-07-14T12:34:27.586131"
  },
  {
    "epoch": 1518,
    "batch_idx": 64,
    "loss": 0.2642032504081726,
    "learning_rate": 0.0001365927672565488,
    "timestep_mean": 437.125,
    "epsilon_norm": 7.573843002319336,
    "prediction_norm": 6.717829704284668,
    "timestamp": "2025-07-14T12:34:27.591017"
  },
  {
    "epoch": 1518,
    "batch_idx": 96,
    "loss": 0.37071090936660767,
    "learning_rate": 0.0001365927672565488,
    "timestep_mean": 515.375,
    "epsilon_norm": 8.256378173828125,
    "prediction_norm": 7.328622341156006,
    "timestamp": "2025-07-14T12:34:27.595875"
  },
  {
    "epoch": 1518,
    "batch_idx": 128,
    "loss": 0.23809805512428284,
    "learning_rate": 0.0001365927672565488,
    "timestep_mean": 481.65625,
    "epsilon_norm": 7.720423221588135,
    "prediction_norm": 6.567480564117432,
    "timestamp": "2025-07-14T12:34:27.600662"
  },
  {
    "epoch": 1518,
    "batch_idx": 160,
    "loss": 0.3041088581085205,
    "learning_rate": 0.0001365927672565488,
    "timestep_mean": 515.28125,
    "epsilon_norm": 9.454204559326172,
    "prediction_norm": 8.51308822631836,
    "timestamp": "2025-07-14T12:34:27.605521"
  },
  {
    "epoch": 1518,
    "batch_idx": 192,
    "loss": 0.17164529860019684,
    "learning_rate": 0.0001365927672565488,
    "timestep_mean": 383.875,
    "epsilon_norm": 3.0416948795318604,
    "prediction_norm": 3.0115787982940674,
    "timestamp": "2025-07-14T12:34:27.610391"
  },
  {
    "epoch": 1519,
    "batch_idx": 0,
    "loss": 0.36919957399368286,
    "learning_rate": 0.0001360537777723594,
    "timestep_mean": 519.875,
    "epsilon_norm": 9.826308250427246,
    "prediction_norm": 8.693004608154297,
    "timestamp": "2025-07-14T12:34:27.615260"
  },
  {
    "epoch": 1519,
    "batch_idx": 32,
    "loss": 0.24443355202674866,
    "learning_rate": 0.0001360537777723594,
    "timestep_mean": 610.9375,
    "epsilon_norm": 6.759275436401367,
    "prediction_norm": 5.872901439666748,
    "timestamp": "2025-07-14T12:34:27.620177"
  },
  {
    "epoch": 1519,
    "batch_idx": 64,
    "loss": 0.3394039571285248,
    "learning_rate": 0.0001360537777723594,
    "timestep_mean": 456.15625,
    "epsilon_norm": 8.310348510742188,
    "prediction_norm": 6.7585344314575195,
    "timestamp": "2025-07-14T12:34:27.625107"
  },
  {
    "epoch": 1519,
    "batch_idx": 96,
    "loss": 0.34354931116104126,
    "learning_rate": 0.0001360537777723594,
    "timestep_mean": 461.0,
    "epsilon_norm": 8.739680290222168,
    "prediction_norm": 7.404051303863525,
    "timestamp": "2025-07-14T12:34:27.630004"
  },
  {
    "epoch": 1519,
    "batch_idx": 128,
    "loss": 0.7279820442199707,
    "learning_rate": 0.0001360537777723594,
    "timestep_mean": 505.5,
    "epsilon_norm": 9.05798053741455,
    "prediction_norm": 7.06745719909668,
    "timestamp": "2025-07-14T12:34:27.634885"
  },
  {
    "epoch": 1519,
    "batch_idx": 160,
    "loss": 0.2631917893886566,
    "learning_rate": 0.0001360537777723594,
    "timestep_mean": 466.125,
    "epsilon_norm": 7.671646595001221,
    "prediction_norm": 6.845001220703125,
    "timestamp": "2025-07-14T12:34:27.639912"
  },
  {
    "epoch": 1519,
    "batch_idx": 192,
    "loss": 0.4619169235229492,
    "learning_rate": 0.0001360537777723594,
    "timestep_mean": 462.75,
    "epsilon_norm": 4.268555164337158,
    "prediction_norm": 3.2185628414154053,
    "timestamp": "2025-07-14T12:34:27.644793"
  },
  {
    "epoch": 1520,
    "batch_idx": 0,
    "loss": 0.2532358169555664,
    "learning_rate": 0.00013551568628929433,
    "timestep_mean": 534.90625,
    "epsilon_norm": 7.946979999542236,
    "prediction_norm": 7.04054594039917,
    "timestamp": "2025-07-14T12:34:27.649685"
  },
  {
    "epoch": 1520,
    "batch_idx": 32,
    "loss": 0.1669866144657135,
    "learning_rate": 0.00013551568628929433,
    "timestep_mean": 541.59375,
    "epsilon_norm": 7.8833441734313965,
    "prediction_norm": 7.270853042602539,
    "timestamp": "2025-07-14T12:34:27.654653"
  },
  {
    "epoch": 1520,
    "batch_idx": 64,
    "loss": 0.37578946352005005,
    "learning_rate": 0.00013551568628929433,
    "timestep_mean": 497.09375,
    "epsilon_norm": 9.478635787963867,
    "prediction_norm": 7.807488441467285,
    "timestamp": "2025-07-14T12:34:27.659624"
  },
  {
    "epoch": 1520,
    "batch_idx": 96,
    "loss": 0.2833646237850189,
    "learning_rate": 0.00013551568628929433,
    "timestep_mean": 540.25,
    "epsilon_norm": 8.086468696594238,
    "prediction_norm": 6.842128753662109,
    "timestamp": "2025-07-14T12:34:27.664576"
  },
  {
    "epoch": 1520,
    "batch_idx": 128,
    "loss": 0.15327990055084229,
    "learning_rate": 0.00013551568628929433,
    "timestep_mean": 547.6875,
    "epsilon_norm": 7.063522815704346,
    "prediction_norm": 6.550174236297607,
    "timestamp": "2025-07-14T12:34:27.669458"
  },
  {
    "epoch": 1520,
    "batch_idx": 160,
    "loss": 0.283296674489975,
    "learning_rate": 0.00013551568628929433,
    "timestep_mean": 561.25,
    "epsilon_norm": 7.88718318939209,
    "prediction_norm": 6.816246032714844,
    "timestamp": "2025-07-14T12:34:27.674380"
  },
  {
    "epoch": 1520,
    "batch_idx": 192,
    "loss": 0.41104769706726074,
    "learning_rate": 0.00013551568628929433,
    "timestep_mean": 682.875,
    "epsilon_norm": 6.004556655883789,
    "prediction_norm": 5.471379280090332,
    "timestamp": "2025-07-14T12:34:27.679252"
  },
  {
    "epoch": 1521,
    "batch_idx": 0,
    "loss": 0.19062452018260956,
    "learning_rate": 0.00013497849413504135,
    "timestep_mean": 549.1875,
    "epsilon_norm": 8.35137939453125,
    "prediction_norm": 7.4366278648376465,
    "timestamp": "2025-07-14T12:34:27.684186"
  },
  {
    "epoch": 1521,
    "batch_idx": 32,
    "loss": 0.2687867283821106,
    "learning_rate": 0.00013497849413504135,
    "timestep_mean": 471.5,
    "epsilon_norm": 8.087530136108398,
    "prediction_norm": 6.977025985717773,
    "timestamp": "2025-07-14T12:34:27.689120"
  },
  {
    "epoch": 1521,
    "batch_idx": 64,
    "loss": 0.1978607475757599,
    "learning_rate": 0.00013497849413504135,
    "timestep_mean": 551.15625,
    "epsilon_norm": 7.890030860900879,
    "prediction_norm": 6.9410319328308105,
    "timestamp": "2025-07-14T12:34:27.693999"
  },
  {
    "epoch": 1521,
    "batch_idx": 96,
    "loss": 0.11374281346797943,
    "learning_rate": 0.00013497849413504135,
    "timestep_mean": 657.71875,
    "epsilon_norm": 7.974979877471924,
    "prediction_norm": 7.916418552398682,
    "timestamp": "2025-07-14T12:34:27.698922"
  },
  {
    "epoch": 1521,
    "batch_idx": 128,
    "loss": 0.33979612588882446,
    "learning_rate": 0.00013497849413504135,
    "timestep_mean": 435.25,
    "epsilon_norm": 7.860365867614746,
    "prediction_norm": 6.422762393951416,
    "timestamp": "2025-07-14T12:34:27.703793"
  },
  {
    "epoch": 1521,
    "batch_idx": 160,
    "loss": 0.2835228145122528,
    "learning_rate": 0.00013497849413504135,
    "timestep_mean": 516.4375,
    "epsilon_norm": 8.035423278808594,
    "prediction_norm": 6.848872184753418,
    "timestamp": "2025-07-14T12:34:27.708888"
  },
  {
    "epoch": 1521,
    "batch_idx": 192,
    "loss": 0.1697062999010086,
    "learning_rate": 0.00013497849413504135,
    "timestep_mean": 660.625,
    "epsilon_norm": 3.6680145263671875,
    "prediction_norm": 3.206664562225342,
    "timestamp": "2025-07-14T12:34:27.713770"
  },
  {
    "epoch": 1522,
    "batch_idx": 0,
    "loss": 0.4148574471473694,
    "learning_rate": 0.0001344422026350681,
    "timestep_mean": 486.28125,
    "epsilon_norm": 9.098217964172363,
    "prediction_norm": 7.321436882019043,
    "timestamp": "2025-07-14T12:34:27.718711"
  },
  {
    "epoch": 1522,
    "batch_idx": 32,
    "loss": 0.32339999079704285,
    "learning_rate": 0.0001344422026350681,
    "timestep_mean": 474.25,
    "epsilon_norm": 8.464094161987305,
    "prediction_norm": 6.643275737762451,
    "timestamp": "2025-07-14T12:34:27.723652"
  },
  {
    "epoch": 1522,
    "batch_idx": 64,
    "loss": 0.17467910051345825,
    "learning_rate": 0.0001344422026350681,
    "timestep_mean": 453.75,
    "epsilon_norm": 6.805246829986572,
    "prediction_norm": 5.6472344398498535,
    "timestamp": "2025-07-14T12:34:27.728544"
  },
  {
    "epoch": 1522,
    "batch_idx": 96,
    "loss": 0.3308815658092499,
    "learning_rate": 0.0001344422026350681,
    "timestep_mean": 561.71875,
    "epsilon_norm": 7.882604598999023,
    "prediction_norm": 6.366264343261719,
    "timestamp": "2025-07-14T12:34:27.733567"
  },
  {
    "epoch": 1522,
    "batch_idx": 128,
    "loss": 0.24131235480308533,
    "learning_rate": 0.0001344422026350681,
    "timestep_mean": 513.78125,
    "epsilon_norm": 7.648437976837158,
    "prediction_norm": 6.5114874839782715,
    "timestamp": "2025-07-14T12:34:27.738408"
  },
  {
    "epoch": 1522,
    "batch_idx": 160,
    "loss": 0.31754249334335327,
    "learning_rate": 0.0001344422026350681,
    "timestep_mean": 422.15625,
    "epsilon_norm": 8.505257606506348,
    "prediction_norm": 6.918403148651123,
    "timestamp": "2025-07-14T12:34:27.743356"
  },
  {
    "epoch": 1522,
    "batch_idx": 192,
    "loss": 0.5866819620132446,
    "learning_rate": 0.0001344422026350681,
    "timestep_mean": 515.25,
    "epsilon_norm": 4.228294372558594,
    "prediction_norm": 3.2926011085510254,
    "timestamp": "2025-07-14T12:34:27.748370"
  },
  {
    "epoch": 1523,
    "batch_idx": 0,
    "loss": 0.37445443868637085,
    "learning_rate": 0.00013390681311262095,
    "timestep_mean": 524.875,
    "epsilon_norm": 9.325980186462402,
    "prediction_norm": 7.376796722412109,
    "timestamp": "2025-07-14T12:34:27.753452"
  },
  {
    "epoch": 1523,
    "batch_idx": 32,
    "loss": 0.32571324706077576,
    "learning_rate": 0.00013390681311262095,
    "timestep_mean": 372.125,
    "epsilon_norm": 8.015650749206543,
    "prediction_norm": 6.585284233093262,
    "timestamp": "2025-07-14T12:34:27.758605"
  },
  {
    "epoch": 1523,
    "batch_idx": 64,
    "loss": 0.39441537857055664,
    "learning_rate": 0.00013390681311262095,
    "timestep_mean": 487.40625,
    "epsilon_norm": 8.582529067993164,
    "prediction_norm": 7.1921796798706055,
    "timestamp": "2025-07-14T12:34:27.763500"
  },
  {
    "epoch": 1523,
    "batch_idx": 96,
    "loss": 0.15566010773181915,
    "learning_rate": 0.00013390681311262095,
    "timestep_mean": 502.9375,
    "epsilon_norm": 7.334636688232422,
    "prediction_norm": 7.27756929397583,
    "timestamp": "2025-07-14T12:34:27.768471"
  },
  {
    "epoch": 1523,
    "batch_idx": 128,
    "loss": 0.40581777691841125,
    "learning_rate": 0.00013390681311262095,
    "timestep_mean": 534.0625,
    "epsilon_norm": 8.995955467224121,
    "prediction_norm": 7.63800573348999,
    "timestamp": "2025-07-14T12:34:27.773380"
  },
  {
    "epoch": 1523,
    "batch_idx": 160,
    "loss": 0.5396337509155273,
    "learning_rate": 0.00013390681311262095,
    "timestep_mean": 452.65625,
    "epsilon_norm": 8.378754615783691,
    "prediction_norm": 6.349533557891846,
    "timestamp": "2025-07-14T12:34:27.778277"
  },
  {
    "epoch": 1523,
    "batch_idx": 192,
    "loss": 0.0788811445236206,
    "learning_rate": 0.00013390681311262095,
    "timestep_mean": 625.5,
    "epsilon_norm": 5.216712951660156,
    "prediction_norm": 4.93297815322876,
    "timestamp": "2025-07-14T12:34:27.783386"
  },
  {
    "epoch": 1524,
    "batch_idx": 0,
    "loss": 0.3235721290111542,
    "learning_rate": 0.0001333723268887202,
    "timestep_mean": 537.84375,
    "epsilon_norm": 8.879762649536133,
    "prediction_norm": 7.156373977661133,
    "timestamp": "2025-07-14T12:34:27.788307"
  },
  {
    "epoch": 1524,
    "batch_idx": 32,
    "loss": 0.5488226413726807,
    "learning_rate": 0.0001333723268887202,
    "timestep_mean": 373.375,
    "epsilon_norm": 9.426250457763672,
    "prediction_norm": 7.004351615905762,
    "timestamp": "2025-07-14T12:34:27.793169"
  },
  {
    "epoch": 1524,
    "batch_idx": 64,
    "loss": 0.19209472835063934,
    "learning_rate": 0.0001333723268887202,
    "timestep_mean": 543.0,
    "epsilon_norm": 7.409666061401367,
    "prediction_norm": 6.5557355880737305,
    "timestamp": "2025-07-14T12:34:27.798068"
  },
  {
    "epoch": 1524,
    "batch_idx": 96,
    "loss": 0.21829208731651306,
    "learning_rate": 0.0001333723268887202,
    "timestep_mean": 525.5625,
    "epsilon_norm": 8.344338417053223,
    "prediction_norm": 7.463715076446533,
    "timestamp": "2025-07-14T12:34:27.803064"
  },
  {
    "epoch": 1524,
    "batch_idx": 128,
    "loss": 0.1620384007692337,
    "learning_rate": 0.0001333723268887202,
    "timestep_mean": 511.28125,
    "epsilon_norm": 7.779648780822754,
    "prediction_norm": 7.691587448120117,
    "timestamp": "2025-07-14T12:34:27.807961"
  },
  {
    "epoch": 1524,
    "batch_idx": 160,
    "loss": 0.1919083297252655,
    "learning_rate": 0.0001333723268887202,
    "timestep_mean": 499.4375,
    "epsilon_norm": 7.737863540649414,
    "prediction_norm": 7.077805995941162,
    "timestamp": "2025-07-14T12:34:27.812957"
  },
  {
    "epoch": 1524,
    "batch_idx": 192,
    "loss": 0.21818234026432037,
    "learning_rate": 0.0001333723268887202,
    "timestep_mean": 591.5,
    "epsilon_norm": 3.793712854385376,
    "prediction_norm": 3.650280714035034,
    "timestamp": "2025-07-14T12:34:27.817830"
  },
  {
    "epoch": 1525,
    "batch_idx": 0,
    "loss": 0.2455415576696396,
    "learning_rate": 0.00013283874528215734,
    "timestep_mean": 539.4375,
    "epsilon_norm": 7.242295742034912,
    "prediction_norm": 5.780175685882568,
    "timestamp": "2025-07-14T12:34:27.822769"
  },
  {
    "epoch": 1525,
    "batch_idx": 32,
    "loss": 0.24117252230644226,
    "learning_rate": 0.00013283874528215734,
    "timestep_mean": 557.96875,
    "epsilon_norm": 7.891053199768066,
    "prediction_norm": 6.498070240020752,
    "timestamp": "2025-07-14T12:34:27.827727"
  },
  {
    "epoch": 1525,
    "batch_idx": 64,
    "loss": 0.12456769496202469,
    "learning_rate": 0.00013283874528215734,
    "timestep_mean": 533.53125,
    "epsilon_norm": 6.858123302459717,
    "prediction_norm": 6.221510887145996,
    "timestamp": "2025-07-14T12:34:27.832697"
  },
  {
    "epoch": 1525,
    "batch_idx": 96,
    "loss": 0.3162466287612915,
    "learning_rate": 0.00013283874528215734,
    "timestep_mean": 448.90625,
    "epsilon_norm": 9.2128267288208,
    "prediction_norm": 7.597525596618652,
    "timestamp": "2025-07-14T12:34:27.837559"
  },
  {
    "epoch": 1525,
    "batch_idx": 128,
    "loss": 0.22478467226028442,
    "learning_rate": 0.00013283874528215734,
    "timestep_mean": 537.5625,
    "epsilon_norm": 8.693399429321289,
    "prediction_norm": 7.676320552825928,
    "timestamp": "2025-07-14T12:34:27.842502"
  },
  {
    "epoch": 1525,
    "batch_idx": 160,
    "loss": 0.19368886947631836,
    "learning_rate": 0.00013283874528215734,
    "timestep_mean": 518.21875,
    "epsilon_norm": 6.740583896636963,
    "prediction_norm": 6.05408239364624,
    "timestamp": "2025-07-14T12:34:27.847422"
  },
  {
    "epoch": 1525,
    "batch_idx": 192,
    "loss": 0.29566705226898193,
    "learning_rate": 0.00013283874528215734,
    "timestep_mean": 601.25,
    "epsilon_norm": 3.494579792022705,
    "prediction_norm": 3.302380323410034,
    "timestamp": "2025-07-14T12:34:27.852310"
  },
  {
    "epoch": 1526,
    "batch_idx": 0,
    "loss": 0.5029658675193787,
    "learning_rate": 0.00013230606960949214,
    "timestep_mean": 442.125,
    "epsilon_norm": 8.704044342041016,
    "prediction_norm": 6.401862144470215,
    "timestamp": "2025-07-14T12:34:27.857265"
  },
  {
    "epoch": 1526,
    "batch_idx": 32,
    "loss": 0.28728875517845154,
    "learning_rate": 0.00013230606960949214,
    "timestep_mean": 496.40625,
    "epsilon_norm": 7.198093414306641,
    "prediction_norm": 5.810468673706055,
    "timestamp": "2025-07-14T12:34:27.862107"
  },
  {
    "epoch": 1526,
    "batch_idx": 64,
    "loss": 0.2004881501197815,
    "learning_rate": 0.00013230606960949214,
    "timestep_mean": 504.09375,
    "epsilon_norm": 7.217067241668701,
    "prediction_norm": 6.415804862976074,
    "timestamp": "2025-07-14T12:34:27.867138"
  },
  {
    "epoch": 1526,
    "batch_idx": 96,
    "loss": 0.20959237217903137,
    "learning_rate": 0.00013230606960949214,
    "timestep_mean": 538.78125,
    "epsilon_norm": 8.937432289123535,
    "prediction_norm": 8.208712577819824,
    "timestamp": "2025-07-14T12:34:27.872020"
  },
  {
    "epoch": 1526,
    "batch_idx": 128,
    "loss": 0.3790821135044098,
    "learning_rate": 0.00013230606960949214,
    "timestep_mean": 506.25,
    "epsilon_norm": 7.871486663818359,
    "prediction_norm": 6.745445251464844,
    "timestamp": "2025-07-14T12:34:27.876888"
  },
  {
    "epoch": 1526,
    "batch_idx": 160,
    "loss": 0.30764061212539673,
    "learning_rate": 0.00013230606960949214,
    "timestep_mean": 439.65625,
    "epsilon_norm": 8.174428939819336,
    "prediction_norm": 6.771853923797607,
    "timestamp": "2025-07-14T12:34:27.881777"
  },
  {
    "epoch": 1526,
    "batch_idx": 192,
    "loss": 0.38062694668769836,
    "learning_rate": 0.00013230606960949214,
    "timestep_mean": 457.625,
    "epsilon_norm": 4.398887634277344,
    "prediction_norm": 3.6156582832336426,
    "timestamp": "2025-07-14T12:34:27.886620"
  },
  {
    "epoch": 1527,
    "batch_idx": 0,
    "loss": 0.3519380986690521,
    "learning_rate": 0.00013177430118504888,
    "timestep_mean": 461.09375,
    "epsilon_norm": 8.48904037475586,
    "prediction_norm": 7.176451206207275,
    "timestamp": "2025-07-14T12:34:27.891539"
  },
  {
    "epoch": 1527,
    "batch_idx": 32,
    "loss": 0.29133695363998413,
    "learning_rate": 0.00013177430118504888,
    "timestep_mean": 466.78125,
    "epsilon_norm": 7.45109748840332,
    "prediction_norm": 6.259483337402344,
    "timestamp": "2025-07-14T12:34:27.896508"
  },
  {
    "epoch": 1527,
    "batch_idx": 64,
    "loss": 0.2555624544620514,
    "learning_rate": 0.00013177430118504888,
    "timestep_mean": 436.125,
    "epsilon_norm": 6.856218338012695,
    "prediction_norm": 5.414549827575684,
    "timestamp": "2025-07-14T12:34:27.901333"
  },
  {
    "epoch": 1527,
    "batch_idx": 96,
    "loss": 0.15883779525756836,
    "learning_rate": 0.00013177430118504888,
    "timestep_mean": 484.9375,
    "epsilon_norm": 6.223476886749268,
    "prediction_norm": 5.010496139526367,
    "timestamp": "2025-07-14T12:34:27.906270"
  },
  {
    "epoch": 1527,
    "batch_idx": 128,
    "loss": 0.16229695081710815,
    "learning_rate": 0.00013177430118504888,
    "timestep_mean": 611.6875,
    "epsilon_norm": 7.693541526794434,
    "prediction_norm": 7.085536003112793,
    "timestamp": "2025-07-14T12:34:27.911161"
  },
  {
    "epoch": 1527,
    "batch_idx": 160,
    "loss": 0.21877488493919373,
    "learning_rate": 0.00013177430118504888,
    "timestep_mean": 552.53125,
    "epsilon_norm": 8.2196044921875,
    "prediction_norm": 7.10804557800293,
    "timestamp": "2025-07-14T12:34:27.916067"
  },
  {
    "epoch": 1527,
    "batch_idx": 192,
    "loss": 0.0871143490076065,
    "learning_rate": 0.00013177430118504888,
    "timestep_mean": 478.0,
    "epsilon_norm": 3.6820998191833496,
    "prediction_norm": 3.5551583766937256,
    "timestamp": "2025-07-14T12:34:27.921094"
  },
  {
    "epoch": 1528,
    "batch_idx": 0,
    "loss": 0.14836707711219788,
    "learning_rate": 0.00013124344132091318,
    "timestep_mean": 563.6875,
    "epsilon_norm": 7.902097702026367,
    "prediction_norm": 6.742469310760498,
    "timestamp": "2025-07-14T12:34:27.926079"
  },
  {
    "epoch": 1528,
    "batch_idx": 32,
    "loss": 0.05456000193953514,
    "learning_rate": 0.00013124344132091318,
    "timestep_mean": 640.53125,
    "epsilon_norm": 6.861838340759277,
    "prediction_norm": 6.389740943908691,
    "timestamp": "2025-07-14T12:34:27.930953"
  },
  {
    "epoch": 1528,
    "batch_idx": 64,
    "loss": 0.45232704281806946,
    "learning_rate": 0.00013124344132091318,
    "timestep_mean": 473.53125,
    "epsilon_norm": 7.685108184814453,
    "prediction_norm": 6.10452938079834,
    "timestamp": "2025-07-14T12:34:27.938653"
  },
  {
    "epoch": 1528,
    "batch_idx": 96,
    "loss": 0.2007269263267517,
    "learning_rate": 0.00013124344132091318,
    "timestep_mean": 487.3125,
    "epsilon_norm": 8.197185516357422,
    "prediction_norm": 7.3494696617126465,
    "timestamp": "2025-07-14T12:34:27.948538"
  },
  {
    "epoch": 1528,
    "batch_idx": 128,
    "loss": 0.46083134412765503,
    "learning_rate": 0.00013124344132091318,
    "timestep_mean": 391.0625,
    "epsilon_norm": 7.118530750274658,
    "prediction_norm": 4.999138832092285,
    "timestamp": "2025-07-14T12:34:27.960997"
  },
  {
    "epoch": 1528,
    "batch_idx": 160,
    "loss": 0.435494065284729,
    "learning_rate": 0.00013124344132091318,
    "timestep_mean": 445.1875,
    "epsilon_norm": 8.358881950378418,
    "prediction_norm": 6.820000171661377,
    "timestamp": "2025-07-14T12:34:27.968706"
  },
  {
    "epoch": 1528,
    "batch_idx": 192,
    "loss": 0.3608643412590027,
    "learning_rate": 0.00013124344132091318,
    "timestep_mean": 311.75,
    "epsilon_norm": 3.0443286895751953,
    "prediction_norm": 2.362349510192871,
    "timestamp": "2025-07-14T12:34:27.976284"
  },
  {
    "epoch": 1529,
    "batch_idx": 0,
    "loss": 0.2686007618904114,
    "learning_rate": 0.00013071349132692892,
    "timestep_mean": 502.03125,
    "epsilon_norm": 7.418971538543701,
    "prediction_norm": 6.589045524597168,
    "timestamp": "2025-07-14T12:34:27.984432"
  },
  {
    "epoch": 1529,
    "batch_idx": 32,
    "loss": 0.2268758863210678,
    "learning_rate": 0.00013071349132692892,
    "timestep_mean": 483.4375,
    "epsilon_norm": 8.752381324768066,
    "prediction_norm": 7.497647762298584,
    "timestamp": "2025-07-14T12:34:27.991950"
  },
  {
    "epoch": 1529,
    "batch_idx": 64,
    "loss": 0.2845425009727478,
    "learning_rate": 0.00013071349132692892,
    "timestep_mean": 496.4375,
    "epsilon_norm": 8.550003051757812,
    "prediction_norm": 6.907153606414795,
    "timestamp": "2025-07-14T12:34:27.998493"
  },
  {
    "epoch": 1529,
    "batch_idx": 96,
    "loss": 0.33136633038520813,
    "learning_rate": 0.00013071349132692892,
    "timestep_mean": 486.25,
    "epsilon_norm": 9.634223937988281,
    "prediction_norm": 8.400293350219727,
    "timestamp": "2025-07-14T12:34:28.007472"
  },
  {
    "epoch": 1529,
    "batch_idx": 128,
    "loss": 0.17432105541229248,
    "learning_rate": 0.00013071349132692892,
    "timestep_mean": 513.75,
    "epsilon_norm": 7.780345916748047,
    "prediction_norm": 6.746540069580078,
    "timestamp": "2025-07-14T12:34:28.017113"
  },
  {
    "epoch": 1529,
    "batch_idx": 160,
    "loss": 0.33444374799728394,
    "learning_rate": 0.00013071349132692892,
    "timestep_mean": 549.40625,
    "epsilon_norm": 8.151601791381836,
    "prediction_norm": 6.295714378356934,
    "timestamp": "2025-07-14T12:34:28.023846"
  },
  {
    "epoch": 1529,
    "batch_idx": 192,
    "loss": 0.759905219078064,
    "learning_rate": 0.00013071349132692892,
    "timestep_mean": 324.25,
    "epsilon_norm": 4.280314922332764,
    "prediction_norm": 2.3200764656066895,
    "timestamp": "2025-07-14T12:34:28.033966"
  },
  {
    "epoch": 1530,
    "batch_idx": 0,
    "loss": 0.21427875757217407,
    "learning_rate": 0.00013018445251069535,
    "timestep_mean": 479.03125,
    "epsilon_norm": 7.202629566192627,
    "prediction_norm": 6.134235382080078,
    "timestamp": "2025-07-14T12:34:28.040165"
  },
  {
    "epoch": 1530,
    "batch_idx": 32,
    "loss": 0.46961814165115356,
    "learning_rate": 0.00013018445251069535,
    "timestep_mean": 479.78125,
    "epsilon_norm": 8.871574401855469,
    "prediction_norm": 7.051857948303223,
    "timestamp": "2025-07-14T12:34:28.050224"
  },
  {
    "epoch": 1530,
    "batch_idx": 64,
    "loss": 0.22531680762767792,
    "learning_rate": 0.00013018445251069535,
    "timestep_mean": 480.84375,
    "epsilon_norm": 8.313817024230957,
    "prediction_norm": 6.945290565490723,
    "timestamp": "2025-07-14T12:34:28.056551"
  },
  {
    "epoch": 1530,
    "batch_idx": 96,
    "loss": 0.2888627052307129,
    "learning_rate": 0.00013018445251069535,
    "timestep_mean": 438.65625,
    "epsilon_norm": 7.446263313293457,
    "prediction_norm": 5.700400352478027,
    "timestamp": "2025-07-14T12:34:28.063019"
  },
  {
    "epoch": 1530,
    "batch_idx": 128,
    "loss": 0.14669615030288696,
    "learning_rate": 0.00013018445251069535,
    "timestep_mean": 565.3125,
    "epsilon_norm": 8.786492347717285,
    "prediction_norm": 8.04998779296875,
    "timestamp": "2025-07-14T12:34:28.068312"
  },
  {
    "epoch": 1530,
    "batch_idx": 160,
    "loss": 0.3364222049713135,
    "learning_rate": 0.00013018445251069535,
    "timestep_mean": 466.21875,
    "epsilon_norm": 7.562573432922363,
    "prediction_norm": 5.892489910125732,
    "timestamp": "2025-07-14T12:34:28.074063"
  },
  {
    "epoch": 1530,
    "batch_idx": 192,
    "loss": 0.9866774082183838,
    "learning_rate": 0.00013018445251069535,
    "timestep_mean": 590.125,
    "epsilon_norm": 6.5087480545043945,
    "prediction_norm": 4.465449333190918,
    "timestamp": "2025-07-14T12:34:28.079392"
  },
  {
    "epoch": 1531,
    "batch_idx": 0,
    "loss": 0.252871572971344,
    "learning_rate": 0.00012965632617756268,
    "timestep_mean": 468.15625,
    "epsilon_norm": 8.271750450134277,
    "prediction_norm": 7.7041850090026855,
    "timestamp": "2025-07-14T12:34:28.084638"
  },
  {
    "epoch": 1531,
    "batch_idx": 32,
    "loss": 0.34845229983329773,
    "learning_rate": 0.00012965632617756268,
    "timestep_mean": 532.625,
    "epsilon_norm": 7.8894429206848145,
    "prediction_norm": 5.7536444664001465,
    "timestamp": "2025-07-14T12:34:28.089763"
  },
  {
    "epoch": 1531,
    "batch_idx": 64,
    "loss": 0.4602307677268982,
    "learning_rate": 0.00012965632617756268,
    "timestep_mean": 459.3125,
    "epsilon_norm": 8.775951385498047,
    "prediction_norm": 6.311539173126221,
    "timestamp": "2025-07-14T12:34:28.094887"
  },
  {
    "epoch": 1531,
    "batch_idx": 96,
    "loss": 0.2527869641780853,
    "learning_rate": 0.00012965632617756268,
    "timestep_mean": 577.125,
    "epsilon_norm": 8.295236587524414,
    "prediction_norm": 7.300498962402344,
    "timestamp": "2025-07-14T12:34:28.100222"
  },
  {
    "epoch": 1531,
    "batch_idx": 128,
    "loss": 0.2097446322441101,
    "learning_rate": 0.00012965632617756268,
    "timestep_mean": 436.28125,
    "epsilon_norm": 7.385629653930664,
    "prediction_norm": 6.411520004272461,
    "timestamp": "2025-07-14T12:34:28.105290"
  },
  {
    "epoch": 1531,
    "batch_idx": 160,
    "loss": 0.09737730026245117,
    "learning_rate": 0.00012965632617756268,
    "timestep_mean": 584.0625,
    "epsilon_norm": 7.054662227630615,
    "prediction_norm": 6.756645679473877,
    "timestamp": "2025-07-14T12:34:28.110480"
  },
  {
    "epoch": 1531,
    "batch_idx": 192,
    "loss": 0.06542223691940308,
    "learning_rate": 0.00012965632617756268,
    "timestep_mean": 528.75,
    "epsilon_norm": 4.165789604187012,
    "prediction_norm": 4.017140865325928,
    "timestamp": "2025-07-14T12:34:28.115558"
  },
  {
    "epoch": 1532,
    "batch_idx": 0,
    "loss": 0.19925056397914886,
    "learning_rate": 0.00012912911363063053,
    "timestep_mean": 564.53125,
    "epsilon_norm": 8.081196784973145,
    "prediction_norm": 7.107070446014404,
    "timestamp": "2025-07-14T12:34:28.120752"
  },
  {
    "epoch": 1532,
    "batch_idx": 32,
    "loss": 0.2689446806907654,
    "learning_rate": 0.00012912911363063053,
    "timestep_mean": 471.375,
    "epsilon_norm": 7.224891662597656,
    "prediction_norm": 6.067286968231201,
    "timestamp": "2025-07-14T12:34:28.125662"
  },
  {
    "epoch": 1532,
    "batch_idx": 64,
    "loss": 0.23844578862190247,
    "learning_rate": 0.00012912911363063053,
    "timestep_mean": 564.34375,
    "epsilon_norm": 7.618492126464844,
    "prediction_norm": 7.200027942657471,
    "timestamp": "2025-07-14T12:34:28.130760"
  },
  {
    "epoch": 1532,
    "batch_idx": 96,
    "loss": 0.4174678325653076,
    "learning_rate": 0.00012912911363063053,
    "timestep_mean": 546.40625,
    "epsilon_norm": 8.247513771057129,
    "prediction_norm": 6.4793901443481445,
    "timestamp": "2025-07-14T12:34:28.135973"
  },
  {
    "epoch": 1532,
    "batch_idx": 128,
    "loss": 0.3233199715614319,
    "learning_rate": 0.00012912911363063053,
    "timestep_mean": 467.3125,
    "epsilon_norm": 7.541964530944824,
    "prediction_norm": 6.095196723937988,
    "timestamp": "2025-07-14T12:34:28.141271"
  },
  {
    "epoch": 1532,
    "batch_idx": 160,
    "loss": 0.2151867151260376,
    "learning_rate": 0.00012912911363063053,
    "timestep_mean": 470.46875,
    "epsilon_norm": 9.038594245910645,
    "prediction_norm": 8.459615707397461,
    "timestamp": "2025-07-14T12:34:28.146660"
  },
  {
    "epoch": 1532,
    "batch_idx": 192,
    "loss": 0.11805133521556854,
    "learning_rate": 0.00012912911363063053,
    "timestep_mean": 629.125,
    "epsilon_norm": 4.308159351348877,
    "prediction_norm": 3.973970413208008,
    "timestamp": "2025-07-14T12:34:28.152027"
  },
  {
    "epoch": 1533,
    "batch_idx": 0,
    "loss": 0.1678120493888855,
    "learning_rate": 0.00012860281617074328,
    "timestep_mean": 506.75,
    "epsilon_norm": 6.65061616897583,
    "prediction_norm": 5.93914270401001,
    "timestamp": "2025-07-14T12:34:28.157353"
  },
  {
    "epoch": 1533,
    "batch_idx": 32,
    "loss": 0.19538727402687073,
    "learning_rate": 0.00012860281617074328,
    "timestep_mean": 544.125,
    "epsilon_norm": 7.314480304718018,
    "prediction_norm": 6.5984787940979,
    "timestamp": "2025-07-14T12:34:28.162795"
  },
  {
    "epoch": 1533,
    "batch_idx": 64,
    "loss": 0.1473233699798584,
    "learning_rate": 0.00012860281617074328,
    "timestep_mean": 538.3125,
    "epsilon_norm": 8.287389755249023,
    "prediction_norm": 7.646319389343262,
    "timestamp": "2025-07-14T12:34:28.168177"
  },
  {
    "epoch": 1533,
    "batch_idx": 96,
    "loss": 0.25366467237472534,
    "learning_rate": 0.00012860281617074328,
    "timestep_mean": 477.21875,
    "epsilon_norm": 8.210975646972656,
    "prediction_norm": 6.416740417480469,
    "timestamp": "2025-07-14T12:34:28.173464"
  },
  {
    "epoch": 1533,
    "batch_idx": 128,
    "loss": 0.3631719946861267,
    "learning_rate": 0.00012860281617074328,
    "timestep_mean": 464.78125,
    "epsilon_norm": 8.23147201538086,
    "prediction_norm": 6.404553413391113,
    "timestamp": "2025-07-14T12:34:28.178595"
  },
  {
    "epoch": 1533,
    "batch_idx": 160,
    "loss": 0.18970715999603271,
    "learning_rate": 0.00012860281617074328,
    "timestep_mean": 612.25,
    "epsilon_norm": 8.29219913482666,
    "prediction_norm": 7.235048770904541,
    "timestamp": "2025-07-14T12:34:28.184029"
  },
  {
    "epoch": 1533,
    "batch_idx": 192,
    "loss": 0.33252617716789246,
    "learning_rate": 0.00012860281617074328,
    "timestep_mean": 364.5,
    "epsilon_norm": 3.411282777786255,
    "prediction_norm": 3.18391489982605,
    "timestamp": "2025-07-14T12:34:28.189451"
  },
  {
    "epoch": 1534,
    "batch_idx": 0,
    "loss": 0.19250139594078064,
    "learning_rate": 0.00012807743509648753,
    "timestep_mean": 522.75,
    "epsilon_norm": 8.876453399658203,
    "prediction_norm": 7.977744102478027,
    "timestamp": "2025-07-14T12:34:28.194620"
  },
  {
    "epoch": 1534,
    "batch_idx": 32,
    "loss": 0.27477702498435974,
    "learning_rate": 0.00012807743509648753,
    "timestep_mean": 468.6875,
    "epsilon_norm": 8.330995559692383,
    "prediction_norm": 7.5994873046875,
    "timestamp": "2025-07-14T12:34:28.199933"
  },
  {
    "epoch": 1534,
    "batch_idx": 64,
    "loss": 0.24089089035987854,
    "learning_rate": 0.00012807743509648753,
    "timestep_mean": 555.3125,
    "epsilon_norm": 8.984901428222656,
    "prediction_norm": 7.710941314697266,
    "timestamp": "2025-07-14T12:34:28.205195"
  },
  {
    "epoch": 1534,
    "batch_idx": 96,
    "loss": 0.40964698791503906,
    "learning_rate": 0.00012807743509648753,
    "timestep_mean": 493.96875,
    "epsilon_norm": 7.860356330871582,
    "prediction_norm": 5.699678421020508,
    "timestamp": "2025-07-14T12:34:28.210379"
  },
  {
    "epoch": 1534,
    "batch_idx": 128,
    "loss": 0.126054584980011,
    "learning_rate": 0.00012807743509648753,
    "timestep_mean": 512.4375,
    "epsilon_norm": 7.710468292236328,
    "prediction_norm": 7.2431488037109375,
    "timestamp": "2025-07-14T12:34:28.215632"
  },
  {
    "epoch": 1534,
    "batch_idx": 160,
    "loss": 0.17681771516799927,
    "learning_rate": 0.00012807743509648753,
    "timestep_mean": 546.4375,
    "epsilon_norm": 7.37052059173584,
    "prediction_norm": 7.0450520515441895,
    "timestamp": "2025-07-14T12:34:28.220727"
  },
  {
    "epoch": 1534,
    "batch_idx": 192,
    "loss": 0.28201502561569214,
    "learning_rate": 0.00012807743509648753,
    "timestep_mean": 344.875,
    "epsilon_norm": 3.9251809120178223,
    "prediction_norm": 2.831894636154175,
    "timestamp": "2025-07-14T12:34:28.225969"
  },
  {
    "epoch": 1535,
    "batch_idx": 0,
    "loss": 0.17504136264324188,
    "learning_rate": 0.00012755297170418919,
    "timestep_mean": 505.03125,
    "epsilon_norm": 7.846040725708008,
    "prediction_norm": 7.304781436920166,
    "timestamp": "2025-07-14T12:34:28.231107"
  },
  {
    "epoch": 1535,
    "batch_idx": 32,
    "loss": 0.18006464838981628,
    "learning_rate": 0.00012755297170418919,
    "timestep_mean": 555.84375,
    "epsilon_norm": 9.138014793395996,
    "prediction_norm": 8.120437622070312,
    "timestamp": "2025-07-14T12:34:28.236403"
  },
  {
    "epoch": 1535,
    "batch_idx": 64,
    "loss": 0.2867963910102844,
    "learning_rate": 0.00012755297170418919,
    "timestep_mean": 534.28125,
    "epsilon_norm": 8.405473709106445,
    "prediction_norm": 7.350924491882324,
    "timestamp": "2025-07-14T12:34:28.241762"
  },
  {
    "epoch": 1535,
    "batch_idx": 96,
    "loss": 0.335345596075058,
    "learning_rate": 0.00012755297170418919,
    "timestep_mean": 398.0,
    "epsilon_norm": 7.631810188293457,
    "prediction_norm": 5.776979446411133,
    "timestamp": "2025-07-14T12:34:28.246912"
  },
  {
    "epoch": 1535,
    "batch_idx": 128,
    "loss": 0.3810160756111145,
    "learning_rate": 0.00012755297170418919,
    "timestep_mean": 523.59375,
    "epsilon_norm": 7.866860866546631,
    "prediction_norm": 6.700288772583008,
    "timestamp": "2025-07-14T12:34:28.252127"
  },
  {
    "epoch": 1535,
    "batch_idx": 160,
    "loss": 0.13326647877693176,
    "learning_rate": 0.00012755297170418919,
    "timestep_mean": 450.5625,
    "epsilon_norm": 6.411807537078857,
    "prediction_norm": 5.9702630043029785,
    "timestamp": "2025-07-14T12:34:28.257427"
  },
  {
    "epoch": 1535,
    "batch_idx": 192,
    "loss": 0.0922805666923523,
    "learning_rate": 0.00012755297170418919,
    "timestep_mean": 613.75,
    "epsilon_norm": 3.2965595722198486,
    "prediction_norm": 3.1798627376556396,
    "timestamp": "2025-07-14T12:34:28.262592"
  },
  {
    "epoch": 1536,
    "batch_idx": 0,
    "loss": 0.33115676045417786,
    "learning_rate": 0.00012702942728790901,
    "timestep_mean": 452.34375,
    "epsilon_norm": 8.402836799621582,
    "prediction_norm": 7.079946994781494,
    "timestamp": "2025-07-14T12:34:28.267722"
  },
  {
    "epoch": 1536,
    "batch_idx": 32,
    "loss": 0.12757736444473267,
    "learning_rate": 0.00012702942728790901,
    "timestep_mean": 526.5,
    "epsilon_norm": 7.295313358306885,
    "prediction_norm": 7.722931861877441,
    "timestamp": "2025-07-14T12:34:28.272923"
  },
  {
    "epoch": 1536,
    "batch_idx": 64,
    "loss": 0.6051571369171143,
    "learning_rate": 0.00012702942728790901,
    "timestep_mean": 485.6875,
    "epsilon_norm": 9.152353286743164,
    "prediction_norm": 7.230616569519043,
    "timestamp": "2025-07-14T12:34:28.278165"
  },
  {
    "epoch": 1536,
    "batch_idx": 96,
    "loss": 0.26075291633605957,
    "learning_rate": 0.00012702942728790901,
    "timestep_mean": 523.0,
    "epsilon_norm": 7.662198066711426,
    "prediction_norm": 6.156056880950928,
    "timestamp": "2025-07-14T12:34:28.283478"
  },
  {
    "epoch": 1536,
    "batch_idx": 128,
    "loss": 0.42150622606277466,
    "learning_rate": 0.00012702942728790901,
    "timestep_mean": 507.59375,
    "epsilon_norm": 8.723959922790527,
    "prediction_norm": 7.010214805603027,
    "timestamp": "2025-07-14T12:34:28.295144"
  },
  {
    "epoch": 1536,
    "batch_idx": 160,
    "loss": 0.15021242201328278,
    "learning_rate": 0.00012702942728790901,
    "timestep_mean": 551.28125,
    "epsilon_norm": 7.237964630126953,
    "prediction_norm": 6.332912445068359,
    "timestamp": "2025-07-14T12:34:28.301372"
  },
  {
    "epoch": 1536,
    "batch_idx": 192,
    "loss": 0.12759296596050262,
    "learning_rate": 0.00012702942728790901,
    "timestep_mean": 576.5,
    "epsilon_norm": 3.4925646781921387,
    "prediction_norm": 2.975782871246338,
    "timestamp": "2025-07-14T12:34:28.311124"
  },
  {
    "epoch": 1537,
    "batch_idx": 0,
    "loss": 0.12327045947313309,
    "learning_rate": 0.00012650680313944107,
    "timestep_mean": 643.21875,
    "epsilon_norm": 8.194510459899902,
    "prediction_norm": 7.73135232925415,
    "timestamp": "2025-07-14T12:34:28.318421"
  },
  {
    "epoch": 1537,
    "batch_idx": 32,
    "loss": 0.2553709149360657,
    "learning_rate": 0.00012650680313944107,
    "timestep_mean": 524.84375,
    "epsilon_norm": 7.935324192047119,
    "prediction_norm": 6.7426276206970215,
    "timestamp": "2025-07-14T12:34:28.326265"
  },
  {
    "epoch": 1537,
    "batch_idx": 64,
    "loss": 0.1516304910182953,
    "learning_rate": 0.00012650680313944107,
    "timestep_mean": 458.8125,
    "epsilon_norm": 8.04891586303711,
    "prediction_norm": 7.587291240692139,
    "timestamp": "2025-07-14T12:34:28.334606"
  },
  {
    "epoch": 1537,
    "batch_idx": 96,
    "loss": 0.2689768671989441,
    "learning_rate": 0.00012650680313944107,
    "timestep_mean": 422.6875,
    "epsilon_norm": 8.10849666595459,
    "prediction_norm": 6.603485107421875,
    "timestamp": "2025-07-14T12:34:28.340918"
  },
  {
    "epoch": 1537,
    "batch_idx": 128,
    "loss": 0.38814324140548706,
    "learning_rate": 0.00012650680313944107,
    "timestep_mean": 455.96875,
    "epsilon_norm": 7.320616245269775,
    "prediction_norm": 6.106287956237793,
    "timestamp": "2025-07-14T12:34:28.346918"
  },
  {
    "epoch": 1537,
    "batch_idx": 160,
    "loss": 0.23672199249267578,
    "learning_rate": 0.00012650680313944107,
    "timestep_mean": 445.71875,
    "epsilon_norm": 8.566682815551758,
    "prediction_norm": 7.604249000549316,
    "timestamp": "2025-07-14T12:34:28.352919"
  },
  {
    "epoch": 1537,
    "batch_idx": 192,
    "loss": 0.09898334741592407,
    "learning_rate": 0.00012650680313944107,
    "timestep_mean": 684.625,
    "epsilon_norm": 3.6076436042785645,
    "prediction_norm": 3.7945163249969482,
    "timestamp": "2025-07-14T12:34:28.359754"
  },
  {
    "epoch": 1538,
    "batch_idx": 0,
    "loss": 0.1355552077293396,
    "learning_rate": 0.00012598510054830886,
    "timestep_mean": 513.4375,
    "epsilon_norm": 7.527454376220703,
    "prediction_norm": 7.02353572845459,
    "timestamp": "2025-07-14T12:34:28.368424"
  },
  {
    "epoch": 1538,
    "batch_idx": 32,
    "loss": 0.16735798120498657,
    "learning_rate": 0.00012598510054830886,
    "timestep_mean": 522.78125,
    "epsilon_norm": 7.53215217590332,
    "prediction_norm": 7.0070013999938965,
    "timestamp": "2025-07-14T12:34:28.374570"
  },
  {
    "epoch": 1538,
    "batch_idx": 64,
    "loss": 0.15656176209449768,
    "learning_rate": 0.00012598510054830886,
    "timestep_mean": 458.0625,
    "epsilon_norm": 7.990286827087402,
    "prediction_norm": 7.150308132171631,
    "timestamp": "2025-07-14T12:34:28.380594"
  },
  {
    "epoch": 1538,
    "batch_idx": 96,
    "loss": 0.3399917483329773,
    "learning_rate": 0.00012598510054830886,
    "timestep_mean": 466.625,
    "epsilon_norm": 7.992981910705566,
    "prediction_norm": 7.131186008453369,
    "timestamp": "2025-07-14T12:34:28.386906"
  },
  {
    "epoch": 1538,
    "batch_idx": 128,
    "loss": 0.2998723089694977,
    "learning_rate": 0.00012598510054830886,
    "timestep_mean": 503.53125,
    "epsilon_norm": 7.677610874176025,
    "prediction_norm": 6.686192512512207,
    "timestamp": "2025-07-14T12:34:28.393091"
  },
  {
    "epoch": 1538,
    "batch_idx": 160,
    "loss": 0.21333742141723633,
    "learning_rate": 0.00012598510054830886,
    "timestep_mean": 497.75,
    "epsilon_norm": 9.454890251159668,
    "prediction_norm": 8.412891387939453,
    "timestamp": "2025-07-14T12:34:28.400815"
  },
  {
    "epoch": 1538,
    "batch_idx": 192,
    "loss": 0.24039241671562195,
    "learning_rate": 0.00012598510054830886,
    "timestep_mean": 438.125,
    "epsilon_norm": 4.233438491821289,
    "prediction_norm": 3.060809373855591,
    "timestamp": "2025-07-14T12:34:28.410082"
  },
  {
    "epoch": 1539,
    "batch_idx": 0,
    "loss": 0.18296825885772705,
    "learning_rate": 0.00012546432080176107,
    "timestep_mean": 473.5625,
    "epsilon_norm": 7.83233118057251,
    "prediction_norm": 6.740060329437256,
    "timestamp": "2025-07-14T12:34:28.416110"
  },
  {
    "epoch": 1539,
    "batch_idx": 32,
    "loss": 0.16164827346801758,
    "learning_rate": 0.00012546432080176107,
    "timestep_mean": 497.25,
    "epsilon_norm": 7.548952102661133,
    "prediction_norm": 6.468408107757568,
    "timestamp": "2025-07-14T12:34:28.422754"
  },
  {
    "epoch": 1539,
    "batch_idx": 64,
    "loss": 0.2476610243320465,
    "learning_rate": 0.00012546432080176107,
    "timestep_mean": 450.09375,
    "epsilon_norm": 7.995404243469238,
    "prediction_norm": 6.599869251251221,
    "timestamp": "2025-07-14T12:34:28.431143"
  },
  {
    "epoch": 1539,
    "batch_idx": 96,
    "loss": 0.32957401871681213,
    "learning_rate": 0.00012546432080176107,
    "timestep_mean": 461.46875,
    "epsilon_norm": 7.144876956939697,
    "prediction_norm": 6.1096296310424805,
    "timestamp": "2025-07-14T12:34:28.438981"
  },
  {
    "epoch": 1539,
    "batch_idx": 128,
    "loss": 0.32716992497444153,
    "learning_rate": 0.00012546432080176107,
    "timestep_mean": 544.0625,
    "epsilon_norm": 7.960572719573975,
    "prediction_norm": 6.636517524719238,
    "timestamp": "2025-07-14T12:34:28.445006"
  },
  {
    "epoch": 1539,
    "batch_idx": 160,
    "loss": 0.20285411179065704,
    "learning_rate": 0.00012546432080176107,
    "timestep_mean": 529.34375,
    "epsilon_norm": 7.511318683624268,
    "prediction_norm": 6.928480625152588,
    "timestamp": "2025-07-14T12:34:28.452189"
  },
  {
    "epoch": 1539,
    "batch_idx": 192,
    "loss": 0.7099578976631165,
    "learning_rate": 0.00012546432080176107,
    "timestep_mean": 323.375,
    "epsilon_norm": 3.8196048736572266,
    "prediction_norm": 2.6858487129211426,
    "timestamp": "2025-07-14T12:34:28.461869"
  },
  {
    "epoch": 1540,
    "batch_idx": 0,
    "loss": 0.20213884115219116,
    "learning_rate": 0.00012494446518477025,
    "timestep_mean": 453.59375,
    "epsilon_norm": 7.49979829788208,
    "prediction_norm": 6.189301013946533,
    "timestamp": "2025-07-14T12:34:28.471234"
  },
  {
    "epoch": 1540,
    "batch_idx": 32,
    "loss": 0.3178739547729492,
    "learning_rate": 0.00012494446518477025,
    "timestep_mean": 519.28125,
    "epsilon_norm": 7.714409828186035,
    "prediction_norm": 6.086348056793213,
    "timestamp": "2025-07-14T12:34:28.480894"
  },
  {
    "epoch": 1540,
    "batch_idx": 64,
    "loss": 0.22996889054775238,
    "learning_rate": 0.00012494446518477025,
    "timestep_mean": 507.5,
    "epsilon_norm": 8.699938774108887,
    "prediction_norm": 7.776904106140137,
    "timestamp": "2025-07-14T12:34:28.488992"
  },
  {
    "epoch": 1540,
    "batch_idx": 96,
    "loss": 0.2087005078792572,
    "learning_rate": 0.00012494446518477025,
    "timestep_mean": 509.375,
    "epsilon_norm": 6.737428188323975,
    "prediction_norm": 5.443756580352783,
    "timestamp": "2025-07-14T12:34:28.497617"
  },
  {
    "epoch": 1540,
    "batch_idx": 128,
    "loss": 0.2617607116699219,
    "learning_rate": 0.00012494446518477025,
    "timestep_mean": 480.3125,
    "epsilon_norm": 8.488720893859863,
    "prediction_norm": 7.45081901550293,
    "timestamp": "2025-07-14T12:34:28.503258"
  },
  {
    "epoch": 1540,
    "batch_idx": 160,
    "loss": 0.22372546792030334,
    "learning_rate": 0.00012494446518477025,
    "timestep_mean": 499.125,
    "epsilon_norm": 7.567446231842041,
    "prediction_norm": 6.7113776206970215,
    "timestamp": "2025-07-14T12:34:28.513253"
  },
  {
    "epoch": 1540,
    "batch_idx": 192,
    "loss": 0.0880969911813736,
    "learning_rate": 0.00012494446518477025,
    "timestep_mean": 501.0,
    "epsilon_norm": 3.571136236190796,
    "prediction_norm": 3.187622308731079,
    "timestamp": "2025-07-14T12:34:28.518594"
  },
  {
    "epoch": 1541,
    "batch_idx": 0,
    "loss": 0.3280288875102997,
    "learning_rate": 0.0001244255349800285,
    "timestep_mean": 425.6875,
    "epsilon_norm": 8.874813079833984,
    "prediction_norm": 7.142716884613037,
    "timestamp": "2025-07-14T12:34:28.523824"
  },
  {
    "epoch": 1541,
    "batch_idx": 32,
    "loss": 0.30949336290359497,
    "learning_rate": 0.0001244255349800285,
    "timestep_mean": 542.03125,
    "epsilon_norm": 8.208065032958984,
    "prediction_norm": 6.914806365966797,
    "timestamp": "2025-07-14T12:34:28.529086"
  },
  {
    "epoch": 1541,
    "batch_idx": 64,
    "loss": 0.38664156198501587,
    "learning_rate": 0.0001244255349800285,
    "timestep_mean": 529.4375,
    "epsilon_norm": 7.704944133758545,
    "prediction_norm": 6.945914268493652,
    "timestamp": "2025-07-14T12:34:28.534044"
  },
  {
    "epoch": 1541,
    "batch_idx": 96,
    "loss": 0.2215687483549118,
    "learning_rate": 0.0001244255349800285,
    "timestep_mean": 488.8125,
    "epsilon_norm": 7.192571640014648,
    "prediction_norm": 6.498908519744873,
    "timestamp": "2025-07-14T12:34:28.539123"
  },
  {
    "epoch": 1541,
    "batch_idx": 128,
    "loss": 0.3804353177547455,
    "learning_rate": 0.0001244255349800285,
    "timestep_mean": 548.53125,
    "epsilon_norm": 9.784097671508789,
    "prediction_norm": 7.879220008850098,
    "timestamp": "2025-07-14T12:34:28.544138"
  },
  {
    "epoch": 1541,
    "batch_idx": 160,
    "loss": 0.23879121243953705,
    "learning_rate": 0.0001244255349800285,
    "timestep_mean": 486.03125,
    "epsilon_norm": 7.897452354431152,
    "prediction_norm": 7.014170169830322,
    "timestamp": "2025-07-14T12:34:28.549233"
  },
  {
    "epoch": 1541,
    "batch_idx": 192,
    "loss": 0.07381652295589447,
    "learning_rate": 0.0001244255349800285,
    "timestep_mean": 686.625,
    "epsilon_norm": 3.2583842277526855,
    "prediction_norm": 3.2112393379211426,
    "timestamp": "2025-07-14T12:34:28.554289"
  },
  {
    "epoch": 1542,
    "batch_idx": 0,
    "loss": 0.22725442051887512,
    "learning_rate": 0.0001239075314679444,
    "timestep_mean": 478.3125,
    "epsilon_norm": 8.86876106262207,
    "prediction_norm": 8.098167419433594,
    "timestamp": "2025-07-14T12:34:28.559361"
  },
  {
    "epoch": 1542,
    "batch_idx": 32,
    "loss": 0.21200409531593323,
    "learning_rate": 0.0001239075314679444,
    "timestep_mean": 475.3125,
    "epsilon_norm": 7.188710689544678,
    "prediction_norm": 6.4997124671936035,
    "timestamp": "2025-07-14T12:34:28.564247"
  },
  {
    "epoch": 1542,
    "batch_idx": 64,
    "loss": 0.2017848789691925,
    "learning_rate": 0.0001239075314679444,
    "timestep_mean": 509.6875,
    "epsilon_norm": 8.446097373962402,
    "prediction_norm": 7.5409626960754395,
    "timestamp": "2025-07-14T12:34:28.569234"
  },
  {
    "epoch": 1542,
    "batch_idx": 96,
    "loss": 0.2672833800315857,
    "learning_rate": 0.0001239075314679444,
    "timestep_mean": 461.59375,
    "epsilon_norm": 7.990682125091553,
    "prediction_norm": 6.648350715637207,
    "timestamp": "2025-07-14T12:34:28.574155"
  },
  {
    "epoch": 1542,
    "batch_idx": 128,
    "loss": 0.16316772997379303,
    "learning_rate": 0.0001239075314679444,
    "timestep_mean": 537.03125,
    "epsilon_norm": 7.364019870758057,
    "prediction_norm": 6.609344005584717,
    "timestamp": "2025-07-14T12:34:28.579151"
  },
  {
    "epoch": 1542,
    "batch_idx": 160,
    "loss": 0.26883840560913086,
    "learning_rate": 0.0001239075314679444,
    "timestep_mean": 609.6875,
    "epsilon_norm": 7.797378063201904,
    "prediction_norm": 6.460679531097412,
    "timestamp": "2025-07-14T12:34:28.584252"
  },
  {
    "epoch": 1542,
    "batch_idx": 192,
    "loss": 0.15180039405822754,
    "learning_rate": 0.0001239075314679444,
    "timestep_mean": 513.5,
    "epsilon_norm": 5.25827169418335,
    "prediction_norm": 4.541801452636719,
    "timestamp": "2025-07-14T12:34:28.589184"
  },
  {
    "epoch": 1543,
    "batch_idx": 0,
    "loss": 0.2686058580875397,
    "learning_rate": 0.00012339045592664013,
    "timestep_mean": 474.71875,
    "epsilon_norm": 6.840328216552734,
    "prediction_norm": 5.573856830596924,
    "timestamp": "2025-07-14T12:34:28.594176"
  },
  {
    "epoch": 1543,
    "batch_idx": 32,
    "loss": 0.16993555426597595,
    "learning_rate": 0.00012339045592664013,
    "timestep_mean": 475.09375,
    "epsilon_norm": 7.955926895141602,
    "prediction_norm": 7.756269931793213,
    "timestamp": "2025-07-14T12:34:28.599077"
  },
  {
    "epoch": 1543,
    "batch_idx": 64,
    "loss": 0.20417949557304382,
    "learning_rate": 0.00012339045592664013,
    "timestep_mean": 555.15625,
    "epsilon_norm": 7.465754985809326,
    "prediction_norm": 6.390139579772949,
    "timestamp": "2025-07-14T12:34:28.604142"
  },
  {
    "epoch": 1543,
    "batch_idx": 96,
    "loss": 0.3189442455768585,
    "learning_rate": 0.00012339045592664013,
    "timestep_mean": 455.28125,
    "epsilon_norm": 8.87177562713623,
    "prediction_norm": 7.271071434020996,
    "timestamp": "2025-07-14T12:34:28.609070"
  },
  {
    "epoch": 1543,
    "batch_idx": 128,
    "loss": 0.2191636562347412,
    "learning_rate": 0.00012339045592664013,
    "timestep_mean": 633.4375,
    "epsilon_norm": 7.614262580871582,
    "prediction_norm": 6.361598491668701,
    "timestamp": "2025-07-14T12:34:28.614074"
  },
  {
    "epoch": 1543,
    "batch_idx": 160,
    "loss": 0.39771074056625366,
    "learning_rate": 0.00012339045592664013,
    "timestep_mean": 459.03125,
    "epsilon_norm": 8.72858715057373,
    "prediction_norm": 7.1496381759643555,
    "timestamp": "2025-07-14T12:34:28.618973"
  },
  {
    "epoch": 1543,
    "batch_idx": 192,
    "loss": 0.19255831837654114,
    "learning_rate": 0.00012339045592664013,
    "timestep_mean": 489.0,
    "epsilon_norm": 3.3788046836853027,
    "prediction_norm": 3.180110216140747,
    "timestamp": "2025-07-14T12:34:28.623935"
  },
  {
    "epoch": 1544,
    "batch_idx": 0,
    "loss": 0.17228202521800995,
    "learning_rate": 0.0001228743096319481,
    "timestep_mean": 548.0625,
    "epsilon_norm": 8.089491844177246,
    "prediction_norm": 7.1714959144592285,
    "timestamp": "2025-07-14T12:34:28.628826"
  },
  {
    "epoch": 1544,
    "batch_idx": 32,
    "loss": 0.29381412267684937,
    "learning_rate": 0.0001228743096319481,
    "timestep_mean": 449.625,
    "epsilon_norm": 8.18871784210205,
    "prediction_norm": 6.238500595092773,
    "timestamp": "2025-07-14T12:34:28.633899"
  },
  {
    "epoch": 1544,
    "batch_idx": 64,
    "loss": 0.1793404519557953,
    "learning_rate": 0.0001228743096319481,
    "timestep_mean": 512.8125,
    "epsilon_norm": 7.809980392456055,
    "prediction_norm": 7.212238311767578,
    "timestamp": "2025-07-14T12:34:28.638803"
  },
  {
    "epoch": 1544,
    "batch_idx": 96,
    "loss": 0.20936459302902222,
    "learning_rate": 0.0001228743096319481,
    "timestep_mean": 487.8125,
    "epsilon_norm": 7.56132698059082,
    "prediction_norm": 6.802156448364258,
    "timestamp": "2025-07-14T12:34:28.643827"
  },
  {
    "epoch": 1544,
    "batch_idx": 128,
    "loss": 0.34232163429260254,
    "learning_rate": 0.0001228743096319481,
    "timestep_mean": 467.9375,
    "epsilon_norm": 8.586808204650879,
    "prediction_norm": 6.895170211791992,
    "timestamp": "2025-07-14T12:34:28.648760"
  },
  {
    "epoch": 1544,
    "batch_idx": 160,
    "loss": 0.36504262685775757,
    "learning_rate": 0.0001228743096319481,
    "timestep_mean": 381.09375,
    "epsilon_norm": 8.621939659118652,
    "prediction_norm": 6.572061061859131,
    "timestamp": "2025-07-14T12:34:28.653684"
  },
  {
    "epoch": 1544,
    "batch_idx": 192,
    "loss": 0.418607234954834,
    "learning_rate": 0.0001228743096319481,
    "timestep_mean": 397.25,
    "epsilon_norm": 3.7966558933258057,
    "prediction_norm": 3.044337749481201,
    "timestamp": "2025-07-14T12:34:28.658681"
  },
  {
    "epoch": 1545,
    "batch_idx": 0,
    "loss": 0.24701496958732605,
    "learning_rate": 0.00012235909385740825,
    "timestep_mean": 534.84375,
    "epsilon_norm": 8.302011489868164,
    "prediction_norm": 7.224536895751953,
    "timestamp": "2025-07-14T12:34:28.663620"
  },
  {
    "epoch": 1545,
    "batch_idx": 32,
    "loss": 0.2575424909591675,
    "learning_rate": 0.00012235909385740825,
    "timestep_mean": 513.6875,
    "epsilon_norm": 7.603647232055664,
    "prediction_norm": 6.158050537109375,
    "timestamp": "2025-07-14T12:34:28.668463"
  },
  {
    "epoch": 1545,
    "batch_idx": 64,
    "loss": 0.21163567900657654,
    "learning_rate": 0.00012235909385740825,
    "timestep_mean": 523.59375,
    "epsilon_norm": 7.970613956451416,
    "prediction_norm": 7.300998210906982,
    "timestamp": "2025-07-14T12:34:28.673317"
  },
  {
    "epoch": 1545,
    "batch_idx": 96,
    "loss": 0.3804948329925537,
    "learning_rate": 0.00012235909385740825,
    "timestep_mean": 556.53125,
    "epsilon_norm": 8.270203590393066,
    "prediction_norm": 7.660417556762695,
    "timestamp": "2025-07-14T12:34:28.678216"
  },
  {
    "epoch": 1545,
    "batch_idx": 128,
    "loss": 0.19431045651435852,
    "learning_rate": 0.00012235909385740825,
    "timestep_mean": 472.5625,
    "epsilon_norm": 7.341069221496582,
    "prediction_norm": 5.794191360473633,
    "timestamp": "2025-07-14T12:34:28.683236"
  },
  {
    "epoch": 1545,
    "batch_idx": 160,
    "loss": 0.24840997159481049,
    "learning_rate": 0.00012235909385740825,
    "timestep_mean": 436.0625,
    "epsilon_norm": 6.836280822753906,
    "prediction_norm": 5.577474117279053,
    "timestamp": "2025-07-14T12:34:28.688486"
  },
  {
    "epoch": 1545,
    "batch_idx": 192,
    "loss": 0.18130487203598022,
    "learning_rate": 0.00012235909385740825,
    "timestep_mean": 436.0,
    "epsilon_norm": 3.6946089267730713,
    "prediction_norm": 3.202120542526245,
    "timestamp": "2025-07-14T12:34:28.693499"
  },
  {
    "epoch": 1546,
    "batch_idx": 0,
    "loss": 0.26465603709220886,
    "learning_rate": 0.00012184480987426407,
    "timestep_mean": 547.75,
    "epsilon_norm": 8.246806144714355,
    "prediction_norm": 6.951747417449951,
    "timestamp": "2025-07-14T12:34:28.698876"
  },
  {
    "epoch": 1546,
    "batch_idx": 32,
    "loss": 0.45295074582099915,
    "learning_rate": 0.00012184480987426407,
    "timestep_mean": 449.1875,
    "epsilon_norm": 8.083361625671387,
    "prediction_norm": 6.5803985595703125,
    "timestamp": "2025-07-14T12:34:28.705190"
  },
  {
    "epoch": 1546,
    "batch_idx": 64,
    "loss": 0.38524606823921204,
    "learning_rate": 0.00012184480987426407,
    "timestep_mean": 407.9375,
    "epsilon_norm": 8.456375122070312,
    "prediction_norm": 5.780344486236572,
    "timestamp": "2025-07-14T12:34:28.718440"
  },
  {
    "epoch": 1546,
    "batch_idx": 96,
    "loss": 0.4715338945388794,
    "learning_rate": 0.00012184480987426407,
    "timestep_mean": 516.71875,
    "epsilon_norm": 8.82791805267334,
    "prediction_norm": 6.887967586517334,
    "timestamp": "2025-07-14T12:34:28.726873"
  },
  {
    "epoch": 1546,
    "batch_idx": 128,
    "loss": 0.26927506923675537,
    "learning_rate": 0.00012184480987426407,
    "timestep_mean": 474.03125,
    "epsilon_norm": 6.778100967407227,
    "prediction_norm": 5.25168514251709,
    "timestamp": "2025-07-14T12:34:28.734717"
  },
  {
    "epoch": 1546,
    "batch_idx": 160,
    "loss": 0.16463981568813324,
    "learning_rate": 0.00012184480987426407,
    "timestep_mean": 536.5,
    "epsilon_norm": 8.60296630859375,
    "prediction_norm": 8.255319595336914,
    "timestamp": "2025-07-14T12:34:28.743123"
  },
  {
    "epoch": 1546,
    "batch_idx": 192,
    "loss": 0.2814408838748932,
    "learning_rate": 0.00012184480987426407,
    "timestep_mean": 439.875,
    "epsilon_norm": 3.9042763710021973,
    "prediction_norm": 3.1202642917633057,
    "timestamp": "2025-07-14T12:34:28.751089"
  },
  {
    "epoch": 1547,
    "batch_idx": 0,
    "loss": 0.3066742420196533,
    "learning_rate": 0.00012133145895146033,
    "timestep_mean": 379.4375,
    "epsilon_norm": 8.399940490722656,
    "prediction_norm": 6.688532829284668,
    "timestamp": "2025-07-14T12:34:28.757240"
  },
  {
    "epoch": 1547,
    "batch_idx": 32,
    "loss": 0.25094208121299744,
    "learning_rate": 0.00012133145895146033,
    "timestep_mean": 487.6875,
    "epsilon_norm": 8.18700122833252,
    "prediction_norm": 7.326484680175781,
    "timestamp": "2025-07-14T12:34:28.763333"
  },
  {
    "epoch": 1547,
    "batch_idx": 64,
    "loss": 0.22110748291015625,
    "learning_rate": 0.00012133145895146033,
    "timestep_mean": 610.375,
    "epsilon_norm": 7.784700870513916,
    "prediction_norm": 6.539761543273926,
    "timestamp": "2025-07-14T12:34:28.769161"
  },
  {
    "epoch": 1547,
    "batch_idx": 96,
    "loss": 0.2680956721305847,
    "learning_rate": 0.00012133145895146033,
    "timestep_mean": 462.125,
    "epsilon_norm": 6.893956184387207,
    "prediction_norm": 6.060095310211182,
    "timestamp": "2025-07-14T12:34:28.775402"
  },
  {
    "epoch": 1547,
    "batch_idx": 128,
    "loss": 0.11892985552549362,
    "learning_rate": 0.00012133145895146033,
    "timestep_mean": 607.40625,
    "epsilon_norm": 8.174234390258789,
    "prediction_norm": 7.137964725494385,
    "timestamp": "2025-07-14T12:34:28.781637"
  },
  {
    "epoch": 1547,
    "batch_idx": 160,
    "loss": 0.22276534140110016,
    "learning_rate": 0.00012133145895146033,
    "timestep_mean": 544.25,
    "epsilon_norm": 8.151617050170898,
    "prediction_norm": 7.039633750915527,
    "timestamp": "2025-07-14T12:34:28.788026"
  },
  {
    "epoch": 1547,
    "batch_idx": 192,
    "loss": 0.2419530600309372,
    "learning_rate": 0.00012133145895146033,
    "timestep_mean": 594.0,
    "epsilon_norm": 5.035034656524658,
    "prediction_norm": 4.099338054656982,
    "timestamp": "2025-07-14T12:34:28.794159"
  },
  {
    "epoch": 1548,
    "batch_idx": 0,
    "loss": 0.10672503709793091,
    "learning_rate": 0.00012081904235563908,
    "timestep_mean": 566.09375,
    "epsilon_norm": 6.811578273773193,
    "prediction_norm": 6.815649509429932,
    "timestamp": "2025-07-14T12:34:28.800929"
  },
  {
    "epoch": 1548,
    "batch_idx": 32,
    "loss": 0.19962267577648163,
    "learning_rate": 0.00012081904235563908,
    "timestep_mean": 485.34375,
    "epsilon_norm": 7.907980442047119,
    "prediction_norm": 6.909053325653076,
    "timestamp": "2025-07-14T12:34:28.807246"
  },
  {
    "epoch": 1548,
    "batch_idx": 64,
    "loss": 0.20615226030349731,
    "learning_rate": 0.00012081904235563908,
    "timestep_mean": 470.9375,
    "epsilon_norm": 8.98005199432373,
    "prediction_norm": 8.068819046020508,
    "timestamp": "2025-07-14T12:34:28.814023"
  },
  {
    "epoch": 1548,
    "batch_idx": 96,
    "loss": 0.2946748435497284,
    "learning_rate": 0.00012081904235563908,
    "timestep_mean": 496.90625,
    "epsilon_norm": 8.374429702758789,
    "prediction_norm": 7.418859958648682,
    "timestamp": "2025-07-14T12:34:28.819369"
  },
  {
    "epoch": 1548,
    "batch_idx": 128,
    "loss": 0.3045903444290161,
    "learning_rate": 0.00012081904235563908,
    "timestep_mean": 455.84375,
    "epsilon_norm": 8.035442352294922,
    "prediction_norm": 6.207601547241211,
    "timestamp": "2025-07-14T12:34:28.825209"
  },
  {
    "epoch": 1548,
    "batch_idx": 160,
    "loss": 0.1953446865081787,
    "learning_rate": 0.00012081904235563908,
    "timestep_mean": 528.65625,
    "epsilon_norm": 6.679985523223877,
    "prediction_norm": 6.037069797515869,
    "timestamp": "2025-07-14T12:34:28.830520"
  },
  {
    "epoch": 1548,
    "batch_idx": 192,
    "loss": 0.18912068009376526,
    "learning_rate": 0.00012081904235563908,
    "timestep_mean": 477.875,
    "epsilon_norm": 3.263779640197754,
    "prediction_norm": 2.745253086090088,
    "timestamp": "2025-07-14T12:34:28.835874"
  },
  {
    "epoch": 1549,
    "batch_idx": 0,
    "loss": 0.3335452377796173,
    "learning_rate": 0.00012030756135113787,
    "timestep_mean": 478.125,
    "epsilon_norm": 7.502713203430176,
    "prediction_norm": 6.661500930786133,
    "timestamp": "2025-07-14T12:34:28.841156"
  },
  {
    "epoch": 1549,
    "batch_idx": 32,
    "loss": 0.1905679553747177,
    "learning_rate": 0.00012030756135113787,
    "timestep_mean": 517.78125,
    "epsilon_norm": 8.9089994430542,
    "prediction_norm": 8.419711112976074,
    "timestamp": "2025-07-14T12:34:28.846485"
  },
  {
    "epoch": 1549,
    "batch_idx": 64,
    "loss": 0.3843454420566559,
    "learning_rate": 0.00012030756135113787,
    "timestep_mean": 395.40625,
    "epsilon_norm": 7.199802398681641,
    "prediction_norm": 5.918967247009277,
    "timestamp": "2025-07-14T12:34:28.851647"
  },
  {
    "epoch": 1549,
    "batch_idx": 96,
    "loss": 0.3772493004798889,
    "learning_rate": 0.00012030756135113787,
    "timestep_mean": 474.3125,
    "epsilon_norm": 9.18984603881836,
    "prediction_norm": 7.110138893127441,
    "timestamp": "2025-07-14T12:34:28.856770"
  },
  {
    "epoch": 1549,
    "batch_idx": 128,
    "loss": 0.30408263206481934,
    "learning_rate": 0.00012030756135113787,
    "timestep_mean": 545.5,
    "epsilon_norm": 7.23581600189209,
    "prediction_norm": 5.48970365524292,
    "timestamp": "2025-07-14T12:34:28.861919"
  },
  {
    "epoch": 1549,
    "batch_idx": 160,
    "loss": 0.2921561002731323,
    "learning_rate": 0.00012030756135113787,
    "timestep_mean": 512.09375,
    "epsilon_norm": 8.116110801696777,
    "prediction_norm": 6.609989166259766,
    "timestamp": "2025-07-14T12:34:28.867274"
  },
  {
    "epoch": 1549,
    "batch_idx": 192,
    "loss": 0.15652874112129211,
    "learning_rate": 0.00012030756135113787,
    "timestep_mean": 509.125,
    "epsilon_norm": 3.909907341003418,
    "prediction_norm": 3.295205593109131,
    "timestamp": "2025-07-14T12:34:28.872456"
  },
  {
    "epoch": 1550,
    "batch_idx": 0,
    "loss": 0.22761471569538116,
    "learning_rate": 0.0001197970171999846,
    "timestep_mean": 580.625,
    "epsilon_norm": 7.590976238250732,
    "prediction_norm": 6.775193691253662,
    "timestamp": "2025-07-14T12:34:28.877529"
  },
  {
    "epoch": 1550,
    "batch_idx": 32,
    "loss": 0.23397739231586456,
    "learning_rate": 0.0001197970171999846,
    "timestep_mean": 505.9375,
    "epsilon_norm": 7.873081207275391,
    "prediction_norm": 7.41123104095459,
    "timestamp": "2025-07-14T12:34:28.882688"
  },
  {
    "epoch": 1550,
    "batch_idx": 64,
    "loss": 0.2536712884902954,
    "learning_rate": 0.0001197970171999846,
    "timestep_mean": 456.53125,
    "epsilon_norm": 6.031587600708008,
    "prediction_norm": 5.327724933624268,
    "timestamp": "2025-07-14T12:34:28.887784"
  },
  {
    "epoch": 1550,
    "batch_idx": 96,
    "loss": 0.4553256630897522,
    "learning_rate": 0.0001197970171999846,
    "timestep_mean": 459.25,
    "epsilon_norm": 8.202164649963379,
    "prediction_norm": 6.01953125,
    "timestamp": "2025-07-14T12:34:28.892898"
  },
  {
    "epoch": 1550,
    "batch_idx": 128,
    "loss": 0.14650657773017883,
    "learning_rate": 0.0001197970171999846,
    "timestep_mean": 544.0625,
    "epsilon_norm": 7.61138916015625,
    "prediction_norm": 6.822376251220703,
    "timestamp": "2025-07-14T12:34:28.898261"
  },
  {
    "epoch": 1550,
    "batch_idx": 160,
    "loss": 0.2595828175544739,
    "learning_rate": 0.0001197970171999846,
    "timestep_mean": 560.15625,
    "epsilon_norm": 8.104313850402832,
    "prediction_norm": 6.774871349334717,
    "timestamp": "2025-07-14T12:34:28.903456"
  },
  {
    "epoch": 1550,
    "batch_idx": 192,
    "loss": 0.32589051127433777,
    "learning_rate": 0.0001197970171999846,
    "timestep_mean": 304.875,
    "epsilon_norm": 4.884848594665527,
    "prediction_norm": 3.6038010120391846,
    "timestamp": "2025-07-14T12:34:28.908722"
  },
  {
    "epoch": 1551,
    "batch_idx": 0,
    "loss": 0.14857885241508484,
    "learning_rate": 0.00011928741116189652,
    "timestep_mean": 633.1875,
    "epsilon_norm": 8.030562400817871,
    "prediction_norm": 7.701486587524414,
    "timestamp": "2025-07-14T12:34:28.913885"
  },
  {
    "epoch": 1551,
    "batch_idx": 32,
    "loss": 0.12499919533729553,
    "learning_rate": 0.00011928741116189652,
    "timestep_mean": 471.15625,
    "epsilon_norm": 7.962827682495117,
    "prediction_norm": 7.307060718536377,
    "timestamp": "2025-07-14T12:34:28.919107"
  },
  {
    "epoch": 1551,
    "batch_idx": 64,
    "loss": 0.18284371495246887,
    "learning_rate": 0.00011928741116189652,
    "timestep_mean": 515.59375,
    "epsilon_norm": 7.480447292327881,
    "prediction_norm": 6.7337327003479,
    "timestamp": "2025-07-14T12:34:28.924813"
  },
  {
    "epoch": 1551,
    "batch_idx": 96,
    "loss": 0.09554257243871689,
    "learning_rate": 0.00011928741116189652,
    "timestep_mean": 520.9375,
    "epsilon_norm": 8.873395919799805,
    "prediction_norm": 8.18798542022705,
    "timestamp": "2025-07-14T12:34:28.929961"
  },
  {
    "epoch": 1551,
    "batch_idx": 128,
    "loss": 0.18107175827026367,
    "learning_rate": 0.00011928741116189652,
    "timestep_mean": 512.625,
    "epsilon_norm": 7.659554481506348,
    "prediction_norm": 6.546360015869141,
    "timestamp": "2025-07-14T12:34:28.937332"
  },
  {
    "epoch": 1551,
    "batch_idx": 160,
    "loss": 0.24254052340984344,
    "learning_rate": 0.00011928741116189652,
    "timestep_mean": 512.0,
    "epsilon_norm": 7.4723076820373535,
    "prediction_norm": 5.743863582611084,
    "timestamp": "2025-07-14T12:34:28.942470"
  },
  {
    "epoch": 1551,
    "batch_idx": 192,
    "loss": 0.38314080238342285,
    "learning_rate": 0.00011928741116189652,
    "timestep_mean": 381.125,
    "epsilon_norm": 4.384318828582764,
    "prediction_norm": 3.7504823207855225,
    "timestamp": "2025-07-14T12:34:28.947771"
  },
  {
    "epoch": 1552,
    "batch_idx": 0,
    "loss": 0.26982152462005615,
    "learning_rate": 0.00011877874449427619,
    "timestep_mean": 555.59375,
    "epsilon_norm": 7.657711982727051,
    "prediction_norm": 6.440352439880371,
    "timestamp": "2025-07-14T12:34:28.952762"
  },
  {
    "epoch": 1552,
    "batch_idx": 32,
    "loss": 0.34794488549232483,
    "learning_rate": 0.00011877874449427619,
    "timestep_mean": 494.125,
    "epsilon_norm": 9.129481315612793,
    "prediction_norm": 6.883243083953857,
    "timestamp": "2025-07-14T12:34:28.957742"
  },
  {
    "epoch": 1552,
    "batch_idx": 64,
    "loss": 0.3290538489818573,
    "learning_rate": 0.00011877874449427619,
    "timestep_mean": 433.8125,
    "epsilon_norm": 8.038146018981934,
    "prediction_norm": 6.601088523864746,
    "timestamp": "2025-07-14T12:34:28.962939"
  },
  {
    "epoch": 1552,
    "batch_idx": 96,
    "loss": 0.31596848368644714,
    "learning_rate": 0.00011877874449427619,
    "timestep_mean": 476.84375,
    "epsilon_norm": 7.766106128692627,
    "prediction_norm": 5.921665668487549,
    "timestamp": "2025-07-14T12:34:28.967897"
  },
  {
    "epoch": 1552,
    "batch_idx": 128,
    "loss": 0.29102498292922974,
    "learning_rate": 0.00011877874449427619,
    "timestep_mean": 432.53125,
    "epsilon_norm": 7.201282978057861,
    "prediction_norm": 5.390805244445801,
    "timestamp": "2025-07-14T12:34:28.972875"
  },
  {
    "epoch": 1552,
    "batch_idx": 160,
    "loss": 0.4825246334075928,
    "learning_rate": 0.00011877874449427619,
    "timestep_mean": 435.28125,
    "epsilon_norm": 9.620818138122559,
    "prediction_norm": 7.85044527053833,
    "timestamp": "2025-07-14T12:34:28.977807"
  },
  {
    "epoch": 1552,
    "batch_idx": 192,
    "loss": 0.1562139391899109,
    "learning_rate": 0.00011877874449427619,
    "timestep_mean": 650.0,
    "epsilon_norm": 4.388118267059326,
    "prediction_norm": 3.4238290786743164,
    "timestamp": "2025-07-14T12:34:28.982739"
  },
  {
    "epoch": 1553,
    "batch_idx": 0,
    "loss": 0.18038274347782135,
    "learning_rate": 0.0001182710184522074,
    "timestep_mean": 521.28125,
    "epsilon_norm": 8.68689250946045,
    "prediction_norm": 7.349080562591553,
    "timestamp": "2025-07-14T12:34:28.987798"
  },
  {
    "epoch": 1553,
    "batch_idx": 32,
    "loss": 0.11232049763202667,
    "learning_rate": 0.0001182710184522074,
    "timestep_mean": 571.53125,
    "epsilon_norm": 7.76604700088501,
    "prediction_norm": 6.920966148376465,
    "timestamp": "2025-07-14T12:34:28.992647"
  },
  {
    "epoch": 1553,
    "batch_idx": 64,
    "loss": 0.18268655240535736,
    "learning_rate": 0.0001182710184522074,
    "timestep_mean": 507.78125,
    "epsilon_norm": 8.030323028564453,
    "prediction_norm": 7.401682376861572,
    "timestamp": "2025-07-14T12:34:28.998144"
  },
  {
    "epoch": 1553,
    "batch_idx": 96,
    "loss": 0.25956839323043823,
    "learning_rate": 0.0001182710184522074,
    "timestep_mean": 467.0,
    "epsilon_norm": 7.699434757232666,
    "prediction_norm": 6.030850410461426,
    "timestamp": "2025-07-14T12:34:29.003373"
  },
  {
    "epoch": 1553,
    "batch_idx": 128,
    "loss": 0.28730878233909607,
    "learning_rate": 0.0001182710184522074,
    "timestep_mean": 455.84375,
    "epsilon_norm": 9.18294906616211,
    "prediction_norm": 7.0237250328063965,
    "timestamp": "2025-07-14T12:34:29.008820"
  },
  {
    "epoch": 1553,
    "batch_idx": 160,
    "loss": 0.31765878200531006,
    "learning_rate": 0.0001182710184522074,
    "timestep_mean": 477.0,
    "epsilon_norm": 7.164159297943115,
    "prediction_norm": 5.462913513183594,
    "timestamp": "2025-07-14T12:34:29.016020"
  },
  {
    "epoch": 1553,
    "batch_idx": 192,
    "loss": 0.37756484746932983,
    "learning_rate": 0.0001182710184522074,
    "timestep_mean": 434.375,
    "epsilon_norm": 3.761824607849121,
    "prediction_norm": 2.650024890899658,
    "timestamp": "2025-07-14T12:34:29.021695"
  },
  {
    "epoch": 1554,
    "batch_idx": 0,
    "loss": 0.19883546233177185,
    "learning_rate": 0.00011776423428845437,
    "timestep_mean": 515.375,
    "epsilon_norm": 7.426846027374268,
    "prediction_norm": 6.736161708831787,
    "timestamp": "2025-07-14T12:34:29.027844"
  },
  {
    "epoch": 1554,
    "batch_idx": 32,
    "loss": 0.3483794629573822,
    "learning_rate": 0.00011776423428845437,
    "timestep_mean": 585.5625,
    "epsilon_norm": 8.407052040100098,
    "prediction_norm": 6.924979209899902,
    "timestamp": "2025-07-14T12:34:29.033687"
  },
  {
    "epoch": 1554,
    "batch_idx": 64,
    "loss": 0.2881764769554138,
    "learning_rate": 0.00011776423428845437,
    "timestep_mean": 442.28125,
    "epsilon_norm": 6.569432735443115,
    "prediction_norm": 6.027997016906738,
    "timestamp": "2025-07-14T12:34:29.039004"
  },
  {
    "epoch": 1554,
    "batch_idx": 96,
    "loss": 0.169904887676239,
    "learning_rate": 0.00011776423428845437,
    "timestep_mean": 482.21875,
    "epsilon_norm": 7.502880096435547,
    "prediction_norm": 7.054375171661377,
    "timestamp": "2025-07-14T12:34:29.044895"
  },
  {
    "epoch": 1554,
    "batch_idx": 128,
    "loss": 0.10038910806179047,
    "learning_rate": 0.00011776423428845437,
    "timestep_mean": 559.53125,
    "epsilon_norm": 6.870532989501953,
    "prediction_norm": 6.5493669509887695,
    "timestamp": "2025-07-14T12:34:29.050304"
  },
  {
    "epoch": 1554,
    "batch_idx": 160,
    "loss": 0.250568687915802,
    "learning_rate": 0.00011776423428845437,
    "timestep_mean": 573.4375,
    "epsilon_norm": 8.37181568145752,
    "prediction_norm": 7.439795017242432,
    "timestamp": "2025-07-14T12:34:29.055935"
  },
  {
    "epoch": 1554,
    "batch_idx": 192,
    "loss": 0.7815865278244019,
    "learning_rate": 0.00011776423428845437,
    "timestep_mean": 426.25,
    "epsilon_norm": 4.732212543487549,
    "prediction_norm": 3.1321563720703125,
    "timestamp": "2025-07-14T12:34:29.063778"
  },
  {
    "epoch": 1555,
    "batch_idx": 0,
    "loss": 0.2068464159965515,
    "learning_rate": 0.00011725839325345601,
    "timestep_mean": 528.03125,
    "epsilon_norm": 7.8822021484375,
    "prediction_norm": 6.941980361938477,
    "timestamp": "2025-07-14T12:34:29.069269"
  },
  {
    "epoch": 1555,
    "batch_idx": 32,
    "loss": 0.47912368178367615,
    "learning_rate": 0.00011725839325345601,
    "timestep_mean": 517.625,
    "epsilon_norm": 7.540550708770752,
    "prediction_norm": 6.395135402679443,
    "timestamp": "2025-07-14T12:34:29.074251"
  },
  {
    "epoch": 1555,
    "batch_idx": 64,
    "loss": 0.3669411540031433,
    "learning_rate": 0.00011725839325345601,
    "timestep_mean": 513.40625,
    "epsilon_norm": 8.29816722869873,
    "prediction_norm": 6.737311363220215,
    "timestamp": "2025-07-14T12:34:29.079375"
  },
  {
    "epoch": 1555,
    "batch_idx": 96,
    "loss": 0.3916977643966675,
    "learning_rate": 0.00011725839325345601,
    "timestep_mean": 463.40625,
    "epsilon_norm": 8.291325569152832,
    "prediction_norm": 6.9354023933410645,
    "timestamp": "2025-07-14T12:34:29.084503"
  },
  {
    "epoch": 1555,
    "batch_idx": 128,
    "loss": 0.2056104689836502,
    "learning_rate": 0.00011725839325345601,
    "timestep_mean": 481.8125,
    "epsilon_norm": 7.409645080566406,
    "prediction_norm": 6.34107780456543,
    "timestamp": "2025-07-14T12:34:29.090368"
  },
  {
    "epoch": 1555,
    "batch_idx": 160,
    "loss": 0.15373949706554413,
    "learning_rate": 0.00011725839325345601,
    "timestep_mean": 609.625,
    "epsilon_norm": 7.37736177444458,
    "prediction_norm": 6.707315444946289,
    "timestamp": "2025-07-14T12:34:29.095718"
  },
  {
    "epoch": 1555,
    "batch_idx": 192,
    "loss": 0.2988511025905609,
    "learning_rate": 0.00011725839325345601,
    "timestep_mean": 606.875,
    "epsilon_norm": 3.24723482131958,
    "prediction_norm": 2.6972715854644775,
    "timestamp": "2025-07-14T12:34:29.107015"
  },
  {
    "epoch": 1556,
    "batch_idx": 0,
    "loss": 0.32664987444877625,
    "learning_rate": 0.00011675349659532504,
    "timestep_mean": 469.15625,
    "epsilon_norm": 8.06629753112793,
    "prediction_norm": 5.993716239929199,
    "timestamp": "2025-07-14T12:34:29.112626"
  },
  {
    "epoch": 1556,
    "batch_idx": 32,
    "loss": 0.40023425221443176,
    "learning_rate": 0.00011675349659532504,
    "timestep_mean": 518.75,
    "epsilon_norm": 8.239107131958008,
    "prediction_norm": 7.089168548583984,
    "timestamp": "2025-07-14T12:34:29.117910"
  },
  {
    "epoch": 1556,
    "batch_idx": 64,
    "loss": 0.53337162733078,
    "learning_rate": 0.00011675349659532504,
    "timestep_mean": 483.1875,
    "epsilon_norm": 9.522533416748047,
    "prediction_norm": 7.440408706665039,
    "timestamp": "2025-07-14T12:34:29.122998"
  },
  {
    "epoch": 1556,
    "batch_idx": 96,
    "loss": 0.41818636655807495,
    "learning_rate": 0.00011675349659532504,
    "timestep_mean": 500.9375,
    "epsilon_norm": 10.256940841674805,
    "prediction_norm": 8.498334884643555,
    "timestamp": "2025-07-14T12:34:29.128108"
  },
  {
    "epoch": 1556,
    "batch_idx": 128,
    "loss": 0.11702940613031387,
    "learning_rate": 0.00011675349659532504,
    "timestep_mean": 514.0625,
    "epsilon_norm": 7.658366680145264,
    "prediction_norm": 7.283148765563965,
    "timestamp": "2025-07-14T12:34:29.133222"
  },
  {
    "epoch": 1556,
    "batch_idx": 160,
    "loss": 0.2543659806251526,
    "learning_rate": 0.00011675349659532504,
    "timestep_mean": 498.125,
    "epsilon_norm": 6.555442810058594,
    "prediction_norm": 5.7664875984191895,
    "timestamp": "2025-07-14T12:34:29.138183"
  },
  {
    "epoch": 1556,
    "batch_idx": 192,
    "loss": 0.16021905839443207,
    "learning_rate": 0.00011675349659532504,
    "timestep_mean": 425.5,
    "epsilon_norm": 3.2706704139709473,
    "prediction_norm": 3.279315710067749,
    "timestamp": "2025-07-14T12:34:29.143209"
  },
  {
    "epoch": 1557,
    "batch_idx": 0,
    "loss": 0.5499494075775146,
    "learning_rate": 0.00011624954555984408,
    "timestep_mean": 427.65625,
    "epsilon_norm": 9.373435020446777,
    "prediction_norm": 7.370858192443848,
    "timestamp": "2025-07-14T12:34:29.148231"
  },
  {
    "epoch": 1557,
    "batch_idx": 32,
    "loss": 0.21180418133735657,
    "learning_rate": 0.00011624954555984408,
    "timestep_mean": 524.71875,
    "epsilon_norm": 8.639263153076172,
    "prediction_norm": 7.744636535644531,
    "timestamp": "2025-07-14T12:34:29.153200"
  },
  {
    "epoch": 1557,
    "batch_idx": 64,
    "loss": 0.32042816281318665,
    "learning_rate": 0.00011624954555984408,
    "timestep_mean": 508.625,
    "epsilon_norm": 8.555290222167969,
    "prediction_norm": 7.379486083984375,
    "timestamp": "2025-07-14T12:34:29.158257"
  },
  {
    "epoch": 1557,
    "batch_idx": 96,
    "loss": 0.3041932284832001,
    "learning_rate": 0.00011624954555984408,
    "timestep_mean": 460.625,
    "epsilon_norm": 8.45627212524414,
    "prediction_norm": 6.950922966003418,
    "timestamp": "2025-07-14T12:34:29.163365"
  },
  {
    "epoch": 1557,
    "batch_idx": 128,
    "loss": 0.17020300030708313,
    "learning_rate": 0.00011624954555984408,
    "timestep_mean": 535.4375,
    "epsilon_norm": 7.15415096282959,
    "prediction_norm": 6.328962802886963,
    "timestamp": "2025-07-14T12:34:29.168263"
  },
  {
    "epoch": 1557,
    "batch_idx": 160,
    "loss": 0.1747620552778244,
    "learning_rate": 0.00011624954555984408,
    "timestep_mean": 492.90625,
    "epsilon_norm": 7.0861406326293945,
    "prediction_norm": 6.362488746643066,
    "timestamp": "2025-07-14T12:34:29.173175"
  },
  {
    "epoch": 1557,
    "batch_idx": 192,
    "loss": 0.9285054206848145,
    "learning_rate": 0.00011624954555984408,
    "timestep_mean": 486.125,
    "epsilon_norm": 4.0185651779174805,
    "prediction_norm": 2.8535773754119873,
    "timestamp": "2025-07-14T12:34:29.178118"
  },
  {
    "epoch": 1558,
    "batch_idx": 0,
    "loss": 0.3407399356365204,
    "learning_rate": 0.0001157465413904617,
    "timestep_mean": 486.78125,
    "epsilon_norm": 7.134843826293945,
    "prediction_norm": 5.640990734100342,
    "timestamp": "2025-07-14T12:34:29.183099"
  },
  {
    "epoch": 1558,
    "batch_idx": 32,
    "loss": 0.1435086727142334,
    "learning_rate": 0.0001157465413904617,
    "timestep_mean": 522.71875,
    "epsilon_norm": 6.769964694976807,
    "prediction_norm": 6.068792819976807,
    "timestamp": "2025-07-14T12:34:29.188202"
  },
  {
    "epoch": 1558,
    "batch_idx": 64,
    "loss": 0.3219476640224457,
    "learning_rate": 0.0001157465413904617,
    "timestep_mean": 435.28125,
    "epsilon_norm": 8.511785507202148,
    "prediction_norm": 7.1150288581848145,
    "timestamp": "2025-07-14T12:34:29.193121"
  },
  {
    "epoch": 1558,
    "batch_idx": 96,
    "loss": 0.2812064588069916,
    "learning_rate": 0.0001157465413904617,
    "timestep_mean": 468.25,
    "epsilon_norm": 7.754648208618164,
    "prediction_norm": 6.624623775482178,
    "timestamp": "2025-07-14T12:34:29.198044"
  },
  {
    "epoch": 1558,
    "batch_idx": 128,
    "loss": 0.26087141036987305,
    "learning_rate": 0.0001157465413904617,
    "timestep_mean": 569.5,
    "epsilon_norm": 8.989269256591797,
    "prediction_norm": 7.457152366638184,
    "timestamp": "2025-07-14T12:34:29.202971"
  },
  {
    "epoch": 1558,
    "batch_idx": 160,
    "loss": 0.39674806594848633,
    "learning_rate": 0.0001157465413904617,
    "timestep_mean": 430.53125,
    "epsilon_norm": 9.451309204101562,
    "prediction_norm": 8.03102970123291,
    "timestamp": "2025-07-14T12:34:29.207986"
  },
  {
    "epoch": 1558,
    "batch_idx": 192,
    "loss": 0.2850993871688843,
    "learning_rate": 0.0001157465413904617,
    "timestep_mean": 519.5,
    "epsilon_norm": 3.8733255863189697,
    "prediction_norm": 3.0811855792999268,
    "timestamp": "2025-07-14T12:34:29.213051"
  },
  {
    "epoch": 1559,
    "batch_idx": 0,
    "loss": 0.14074856042861938,
    "learning_rate": 0.00011524448532829107,
    "timestep_mean": 563.75,
    "epsilon_norm": 7.640213966369629,
    "prediction_norm": 7.33305549621582,
    "timestamp": "2025-07-14T12:34:29.218032"
  },
  {
    "epoch": 1559,
    "batch_idx": 32,
    "loss": 0.3463681638240814,
    "learning_rate": 0.00011524448532829107,
    "timestep_mean": 472.78125,
    "epsilon_norm": 8.549283027648926,
    "prediction_norm": 6.564448833465576,
    "timestamp": "2025-07-14T12:34:29.222997"
  },
  {
    "epoch": 1559,
    "batch_idx": 64,
    "loss": 0.1775875687599182,
    "learning_rate": 0.00011524448532829107,
    "timestep_mean": 553.9375,
    "epsilon_norm": 8.423417091369629,
    "prediction_norm": 7.638424873352051,
    "timestamp": "2025-07-14T12:34:29.227920"
  },
  {
    "epoch": 1559,
    "batch_idx": 96,
    "loss": 0.2794835567474365,
    "learning_rate": 0.00011524448532829107,
    "timestep_mean": 572.21875,
    "epsilon_norm": 7.996119976043701,
    "prediction_norm": 6.61292839050293,
    "timestamp": "2025-07-14T12:34:29.232896"
  },
  {
    "epoch": 1559,
    "batch_idx": 128,
    "loss": 0.360599547624588,
    "learning_rate": 0.00011524448532829107,
    "timestep_mean": 430.875,
    "epsilon_norm": 8.650283813476562,
    "prediction_norm": 7.170693874359131,
    "timestamp": "2025-07-14T12:34:29.237986"
  },
  {
    "epoch": 1559,
    "batch_idx": 160,
    "loss": 0.18221110105514526,
    "learning_rate": 0.00011524448532829107,
    "timestep_mean": 556.3125,
    "epsilon_norm": 7.508141040802002,
    "prediction_norm": 6.392994403839111,
    "timestamp": "2025-07-14T12:34:29.242922"
  },
  {
    "epoch": 1559,
    "batch_idx": 192,
    "loss": 0.3823663890361786,
    "learning_rate": 0.00011524448532829107,
    "timestep_mean": 459.0,
    "epsilon_norm": 4.595066547393799,
    "prediction_norm": 3.568298816680908,
    "timestamp": "2025-07-14T12:34:29.247898"
  },
  {
    "epoch": 1560,
    "batch_idx": 0,
    "loss": 0.2244860827922821,
    "learning_rate": 0.00011474337861210544,
    "timestep_mean": 427.28125,
    "epsilon_norm": 7.76560115814209,
    "prediction_norm": 6.742228984832764,
    "timestamp": "2025-07-14T12:34:29.252855"
  },
  {
    "epoch": 1560,
    "batch_idx": 32,
    "loss": 0.3232840895652771,
    "learning_rate": 0.00011474337861210544,
    "timestep_mean": 434.0625,
    "epsilon_norm": 8.150032043457031,
    "prediction_norm": 6.676571369171143,
    "timestamp": "2025-07-14T12:34:29.257789"
  },
  {
    "epoch": 1560,
    "batch_idx": 64,
    "loss": 0.2847914397716522,
    "learning_rate": 0.00011474337861210544,
    "timestep_mean": 497.875,
    "epsilon_norm": 7.592591762542725,
    "prediction_norm": 6.386541843414307,
    "timestamp": "2025-07-14T12:34:29.262822"
  },
  {
    "epoch": 1560,
    "batch_idx": 96,
    "loss": 0.36867380142211914,
    "learning_rate": 0.00011474337861210544,
    "timestep_mean": 455.84375,
    "epsilon_norm": 8.139723777770996,
    "prediction_norm": 6.642356872558594,
    "timestamp": "2025-07-14T12:34:29.267777"
  },
  {
    "epoch": 1560,
    "batch_idx": 128,
    "loss": 0.22422118484973907,
    "learning_rate": 0.00011474337861210544,
    "timestep_mean": 459.8125,
    "epsilon_norm": 8.024548530578613,
    "prediction_norm": 6.708344459533691,
    "timestamp": "2025-07-14T12:34:29.272730"
  },
  {
    "epoch": 1560,
    "batch_idx": 160,
    "loss": 0.446090966463089,
    "learning_rate": 0.00011474337861210544,
    "timestep_mean": 476.78125,
    "epsilon_norm": 8.40359878540039,
    "prediction_norm": 5.9201250076293945,
    "timestamp": "2025-07-14T12:34:29.277686"
  },
  {
    "epoch": 1560,
    "batch_idx": 192,
    "loss": 0.451494961977005,
    "learning_rate": 0.00011474337861210544,
    "timestep_mean": 381.125,
    "epsilon_norm": 4.992728233337402,
    "prediction_norm": 3.8968894481658936,
    "timestamp": "2025-07-14T12:34:29.282550"
  },
  {
    "epoch": 1561,
    "batch_idx": 0,
    "loss": 0.05952121689915657,
    "learning_rate": 0.00011424322247833575,
    "timestep_mean": 562.3125,
    "epsilon_norm": 8.105997085571289,
    "prediction_norm": 7.88488245010376,
    "timestamp": "2025-07-14T12:34:29.287649"
  },
  {
    "epoch": 1561,
    "batch_idx": 32,
    "loss": 0.5430339574813843,
    "learning_rate": 0.00011424322247833575,
    "timestep_mean": 410.15625,
    "epsilon_norm": 9.355225563049316,
    "prediction_norm": 6.9450531005859375,
    "timestamp": "2025-07-14T12:34:29.292592"
  },
  {
    "epoch": 1561,
    "batch_idx": 64,
    "loss": 0.2751063406467438,
    "learning_rate": 0.00011424322247833575,
    "timestep_mean": 519.125,
    "epsilon_norm": 8.128151893615723,
    "prediction_norm": 6.872978210449219,
    "timestamp": "2025-07-14T12:34:29.297585"
  },
  {
    "epoch": 1561,
    "batch_idx": 96,
    "loss": 0.24862615764141083,
    "learning_rate": 0.00011424322247833575,
    "timestep_mean": 460.96875,
    "epsilon_norm": 7.894665718078613,
    "prediction_norm": 6.839297771453857,
    "timestamp": "2025-07-14T12:34:29.302516"
  },
  {
    "epoch": 1561,
    "batch_idx": 128,
    "loss": 0.3452754318714142,
    "learning_rate": 0.00011424322247833575,
    "timestep_mean": 341.6875,
    "epsilon_norm": 7.664702415466309,
    "prediction_norm": 5.962646484375,
    "timestamp": "2025-07-14T12:34:29.307463"
  },
  {
    "epoch": 1561,
    "batch_idx": 160,
    "loss": 0.29159390926361084,
    "learning_rate": 0.00011424322247833575,
    "timestep_mean": 546.40625,
    "epsilon_norm": 8.10799503326416,
    "prediction_norm": 7.491747856140137,
    "timestamp": "2025-07-14T12:34:29.312561"
  },
  {
    "epoch": 1561,
    "batch_idx": 192,
    "loss": 0.05822303891181946,
    "learning_rate": 0.00011424322247833575,
    "timestep_mean": 574.25,
    "epsilon_norm": 2.7262587547302246,
    "prediction_norm": 3.014826536178589,
    "timestamp": "2025-07-14T12:34:29.317549"
  },
  {
    "epoch": 1562,
    "batch_idx": 0,
    "loss": 0.2951146960258484,
    "learning_rate": 0.00011374401816106778,
    "timestep_mean": 519.0625,
    "epsilon_norm": 8.120096206665039,
    "prediction_norm": 6.6083879470825195,
    "timestamp": "2025-07-14T12:34:29.322498"
  },
  {
    "epoch": 1562,
    "batch_idx": 32,
    "loss": 0.3966241776943207,
    "learning_rate": 0.00011374401816106778,
    "timestep_mean": 403.78125,
    "epsilon_norm": 7.675623416900635,
    "prediction_norm": 6.372151851654053,
    "timestamp": "2025-07-14T12:34:29.327482"
  },
  {
    "epoch": 1562,
    "batch_idx": 64,
    "loss": 0.3669610619544983,
    "learning_rate": 0.00011374401816106778,
    "timestep_mean": 432.75,
    "epsilon_norm": 8.328632354736328,
    "prediction_norm": 6.463488578796387,
    "timestamp": "2025-07-14T12:34:29.332373"
  },
  {
    "epoch": 1562,
    "batch_idx": 96,
    "loss": 0.20837388932704926,
    "learning_rate": 0.00011374401816106778,
    "timestep_mean": 546.125,
    "epsilon_norm": 7.165358543395996,
    "prediction_norm": 6.258955955505371,
    "timestamp": "2025-07-14T12:34:29.337340"
  },
  {
    "epoch": 1562,
    "batch_idx": 128,
    "loss": 0.319497674703598,
    "learning_rate": 0.00011374401816106778,
    "timestep_mean": 532.46875,
    "epsilon_norm": 8.642983436584473,
    "prediction_norm": 7.118407249450684,
    "timestamp": "2025-07-14T12:34:29.343124"
  },
  {
    "epoch": 1562,
    "batch_idx": 160,
    "loss": 0.161283940076828,
    "learning_rate": 0.00011374401816106778,
    "timestep_mean": 442.28125,
    "epsilon_norm": 7.190748691558838,
    "prediction_norm": 6.779935836791992,
    "timestamp": "2025-07-14T12:34:29.349570"
  },
  {
    "epoch": 1562,
    "batch_idx": 192,
    "loss": 0.24386417865753174,
    "learning_rate": 0.00011374401816106778,
    "timestep_mean": 289.875,
    "epsilon_norm": 3.7349061965942383,
    "prediction_norm": 3.2686097621917725,
    "timestamp": "2025-07-14T12:34:29.365022"
  },
  {
    "epoch": 1563,
    "batch_idx": 0,
    "loss": 0.11515732109546661,
    "learning_rate": 0.00011324576689203842,
    "timestep_mean": 486.875,
    "epsilon_norm": 6.607074737548828,
    "prediction_norm": 6.355302333831787,
    "timestamp": "2025-07-14T12:34:29.372087"
  },
  {
    "epoch": 1563,
    "batch_idx": 32,
    "loss": 0.3047817349433899,
    "learning_rate": 0.00011324576689203842,
    "timestep_mean": 433.5625,
    "epsilon_norm": 7.6321539878845215,
    "prediction_norm": 6.126655101776123,
    "timestamp": "2025-07-14T12:34:29.380299"
  },
  {
    "epoch": 1563,
    "batch_idx": 64,
    "loss": 0.18380343914031982,
    "learning_rate": 0.00011324576689203842,
    "timestep_mean": 517.71875,
    "epsilon_norm": 8.2788724899292,
    "prediction_norm": 7.616716384887695,
    "timestamp": "2025-07-14T12:34:29.388921"
  },
  {
    "epoch": 1563,
    "batch_idx": 96,
    "loss": 0.24799196422100067,
    "learning_rate": 0.00011324576689203842,
    "timestep_mean": 544.34375,
    "epsilon_norm": 7.350503921508789,
    "prediction_norm": 7.031247138977051,
    "timestamp": "2025-07-14T12:34:29.397108"
  },
  {
    "epoch": 1563,
    "batch_idx": 128,
    "loss": 0.3456709384918213,
    "learning_rate": 0.00011324576689203842,
    "timestep_mean": 391.625,
    "epsilon_norm": 7.983992099761963,
    "prediction_norm": 6.690125942230225,
    "timestamp": "2025-07-14T12:34:29.405405"
  },
  {
    "epoch": 1563,
    "batch_idx": 160,
    "loss": 0.314983606338501,
    "learning_rate": 0.00011324576689203842,
    "timestep_mean": 515.1875,
    "epsilon_norm": 8.589315414428711,
    "prediction_norm": 6.783679008483887,
    "timestamp": "2025-07-14T12:34:29.411581"
  },
  {
    "epoch": 1563,
    "batch_idx": 192,
    "loss": 0.3814127445220947,
    "learning_rate": 0.00011324576689203842,
    "timestep_mean": 421.25,
    "epsilon_norm": 3.9500792026519775,
    "prediction_norm": 2.677129030227661,
    "timestamp": "2025-07-14T12:34:29.417799"
  },
  {
    "epoch": 1564,
    "batch_idx": 0,
    "loss": 0.1949245184659958,
    "learning_rate": 0.00011274846990063315,
    "timestep_mean": 555.25,
    "epsilon_norm": 6.928218841552734,
    "prediction_norm": 6.064349174499512,
    "timestamp": "2025-07-14T12:34:29.425896"
  },
  {
    "epoch": 1564,
    "batch_idx": 32,
    "loss": 0.2598764896392822,
    "learning_rate": 0.00011274846990063315,
    "timestep_mean": 533.6875,
    "epsilon_norm": 7.722479343414307,
    "prediction_norm": 6.002089023590088,
    "timestamp": "2025-07-14T12:34:29.431657"
  },
  {
    "epoch": 1564,
    "batch_idx": 64,
    "loss": 0.12368202209472656,
    "learning_rate": 0.00011274846990063315,
    "timestep_mean": 635.15625,
    "epsilon_norm": 7.864852428436279,
    "prediction_norm": 7.3110456466674805,
    "timestamp": "2025-07-14T12:34:29.439472"
  },
  {
    "epoch": 1564,
    "batch_idx": 96,
    "loss": 0.14667931199073792,
    "learning_rate": 0.00011274846990063315,
    "timestep_mean": 493.78125,
    "epsilon_norm": 7.920163154602051,
    "prediction_norm": 7.026605606079102,
    "timestamp": "2025-07-14T12:34:29.445590"
  },
  {
    "epoch": 1564,
    "batch_idx": 128,
    "loss": 0.25732558965682983,
    "learning_rate": 0.00011274846990063315,
    "timestep_mean": 480.8125,
    "epsilon_norm": 8.368003845214844,
    "prediction_norm": 7.009675025939941,
    "timestamp": "2025-07-14T12:34:29.451300"
  },
  {
    "epoch": 1564,
    "batch_idx": 160,
    "loss": 0.3249593675136566,
    "learning_rate": 0.00011274846990063315,
    "timestep_mean": 431.34375,
    "epsilon_norm": 7.892227649688721,
    "prediction_norm": 6.894811630249023,
    "timestamp": "2025-07-14T12:34:29.457073"
  },
  {
    "epoch": 1564,
    "batch_idx": 192,
    "loss": 0.05014362931251526,
    "learning_rate": 0.00011274846990063315,
    "timestep_mean": 601.5,
    "epsilon_norm": 3.6364989280700684,
    "prediction_norm": 3.3722360134124756,
    "timestamp": "2025-07-14T12:34:29.462308"
  },
  {
    "epoch": 1565,
    "batch_idx": 0,
    "loss": 0.3626653254032135,
    "learning_rate": 0.00011225212841388282,
    "timestep_mean": 572.03125,
    "epsilon_norm": 7.912062644958496,
    "prediction_norm": 6.957863807678223,
    "timestamp": "2025-07-14T12:34:29.467705"
  },
  {
    "epoch": 1565,
    "batch_idx": 32,
    "loss": 0.23492223024368286,
    "learning_rate": 0.00011225212841388282,
    "timestep_mean": 448.71875,
    "epsilon_norm": 7.976684093475342,
    "prediction_norm": 6.505499839782715,
    "timestamp": "2025-07-14T12:34:29.472781"
  },
  {
    "epoch": 1565,
    "batch_idx": 64,
    "loss": 0.24939599633216858,
    "learning_rate": 0.00011225212841388282,
    "timestep_mean": 456.6875,
    "epsilon_norm": 7.1559367179870605,
    "prediction_norm": 5.845524787902832,
    "timestamp": "2025-07-14T12:34:29.477929"
  },
  {
    "epoch": 1565,
    "batch_idx": 96,
    "loss": 0.5651829242706299,
    "learning_rate": 0.00011225212841388282,
    "timestep_mean": 508.84375,
    "epsilon_norm": 7.780148029327393,
    "prediction_norm": 5.577662467956543,
    "timestamp": "2025-07-14T12:34:29.483235"
  },
  {
    "epoch": 1565,
    "batch_idx": 128,
    "loss": 0.3237285017967224,
    "learning_rate": 0.00011225212841388282,
    "timestep_mean": 568.1875,
    "epsilon_norm": 8.545806884765625,
    "prediction_norm": 7.359452724456787,
    "timestamp": "2025-07-14T12:34:29.488311"
  },
  {
    "epoch": 1565,
    "batch_idx": 160,
    "loss": 0.14298966526985168,
    "learning_rate": 0.00011225212841388282,
    "timestep_mean": 549.71875,
    "epsilon_norm": 8.04818344116211,
    "prediction_norm": 7.569197177886963,
    "timestamp": "2025-07-14T12:34:29.493358"
  },
  {
    "epoch": 1565,
    "batch_idx": 192,
    "loss": 0.3210376799106598,
    "learning_rate": 0.00011225212841388282,
    "timestep_mean": 518.25,
    "epsilon_norm": 4.749475479125977,
    "prediction_norm": 4.458433628082275,
    "timestamp": "2025-07-14T12:34:29.498612"
  },
  {
    "epoch": 1566,
    "batch_idx": 0,
    "loss": 0.25458142161369324,
    "learning_rate": 0.00011175674365646083,
    "timestep_mean": 528.0,
    "epsilon_norm": 7.446470260620117,
    "prediction_norm": 5.917273998260498,
    "timestamp": "2025-07-14T12:34:29.503600"
  },
  {
    "epoch": 1566,
    "batch_idx": 32,
    "loss": 0.15523982048034668,
    "learning_rate": 0.00011175674365646083,
    "timestep_mean": 518.09375,
    "epsilon_norm": 7.999553680419922,
    "prediction_norm": 6.819956302642822,
    "timestamp": "2025-07-14T12:34:29.508652"
  },
  {
    "epoch": 1566,
    "batch_idx": 64,
    "loss": 0.40280526876449585,
    "learning_rate": 0.00011175674365646083,
    "timestep_mean": 489.5,
    "epsilon_norm": 8.543734550476074,
    "prediction_norm": 6.427739143371582,
    "timestamp": "2025-07-14T12:34:29.513645"
  },
  {
    "epoch": 1566,
    "batch_idx": 96,
    "loss": 0.13172243535518646,
    "learning_rate": 0.00011175674365646083,
    "timestep_mean": 498.96875,
    "epsilon_norm": 7.62324333190918,
    "prediction_norm": 7.130527973175049,
    "timestamp": "2025-07-14T12:34:29.518560"
  },
  {
    "epoch": 1566,
    "batch_idx": 128,
    "loss": 0.25334858894348145,
    "learning_rate": 0.00011175674365646083,
    "timestep_mean": 524.53125,
    "epsilon_norm": 7.626412391662598,
    "prediction_norm": 6.633161544799805,
    "timestamp": "2025-07-14T12:34:29.523921"
  },
  {
    "epoch": 1566,
    "batch_idx": 160,
    "loss": 0.1854093074798584,
    "learning_rate": 0.00011175674365646083,
    "timestep_mean": 513.3125,
    "epsilon_norm": 7.658825874328613,
    "prediction_norm": 6.61692476272583,
    "timestamp": "2025-07-14T12:34:29.531137"
  },
  {
    "epoch": 1566,
    "batch_idx": 192,
    "loss": 0.12877757847309113,
    "learning_rate": 0.00011175674365646083,
    "timestep_mean": 462.75,
    "epsilon_norm": 4.137506484985352,
    "prediction_norm": 4.108834266662598,
    "timestamp": "2025-07-14T12:34:29.543873"
  },
  {
    "epoch": 1567,
    "batch_idx": 0,
    "loss": 0.34707963466644287,
    "learning_rate": 0.00011126231685067956,
    "timestep_mean": 442.28125,
    "epsilon_norm": 7.96074914932251,
    "prediction_norm": 5.9326491355896,
    "timestamp": "2025-07-14T12:34:29.550227"
  },
  {
    "epoch": 1567,
    "batch_idx": 32,
    "loss": 0.2669028341770172,
    "learning_rate": 0.00011126231685067956,
    "timestep_mean": 519.625,
    "epsilon_norm": 8.589256286621094,
    "prediction_norm": 7.503971099853516,
    "timestamp": "2025-07-14T12:34:29.555861"
  },
  {
    "epoch": 1567,
    "batch_idx": 64,
    "loss": 0.28833237290382385,
    "learning_rate": 0.00011126231685067956,
    "timestep_mean": 516.25,
    "epsilon_norm": 7.904598236083984,
    "prediction_norm": 6.494758605957031,
    "timestamp": "2025-07-14T12:34:29.566200"
  },
  {
    "epoch": 1567,
    "batch_idx": 96,
    "loss": 0.3079507052898407,
    "learning_rate": 0.00011126231685067956,
    "timestep_mean": 560.75,
    "epsilon_norm": 8.151902198791504,
    "prediction_norm": 7.142376899719238,
    "timestamp": "2025-07-14T12:34:29.575305"
  },
  {
    "epoch": 1567,
    "batch_idx": 128,
    "loss": 0.36423712968826294,
    "learning_rate": 0.00011126231685067956,
    "timestep_mean": 466.59375,
    "epsilon_norm": 8.762761116027832,
    "prediction_norm": 6.7473626136779785,
    "timestamp": "2025-07-14T12:34:29.585616"
  },
  {
    "epoch": 1567,
    "batch_idx": 160,
    "loss": 0.22566020488739014,
    "learning_rate": 0.00011126231685067956,
    "timestep_mean": 505.0625,
    "epsilon_norm": 8.300553321838379,
    "prediction_norm": 6.72796630859375,
    "timestamp": "2025-07-14T12:34:29.594812"
  },
  {
    "epoch": 1567,
    "batch_idx": 192,
    "loss": 0.36996787786483765,
    "learning_rate": 0.00011126231685067956,
    "timestep_mean": 402.125,
    "epsilon_norm": 4.585994243621826,
    "prediction_norm": 3.574172258377075,
    "timestamp": "2025-07-14T12:34:29.602024"
  },
  {
    "epoch": 1568,
    "batch_idx": 0,
    "loss": 0.6001790761947632,
    "learning_rate": 0.00011076884921648833,
    "timestep_mean": 443.875,
    "epsilon_norm": 9.45070743560791,
    "prediction_norm": 6.76131534576416,
    "timestamp": "2025-07-14T12:34:29.608052"
  },
  {
    "epoch": 1568,
    "batch_idx": 32,
    "loss": 0.2000260353088379,
    "learning_rate": 0.00011076884921648833,
    "timestep_mean": 489.34375,
    "epsilon_norm": 7.31836462020874,
    "prediction_norm": 6.834695816040039,
    "timestamp": "2025-07-14T12:34:29.614209"
  },
  {
    "epoch": 1568,
    "batch_idx": 64,
    "loss": 0.23597531020641327,
    "learning_rate": 0.00011076884921648833,
    "timestep_mean": 500.21875,
    "epsilon_norm": 8.111541748046875,
    "prediction_norm": 6.348677158355713,
    "timestamp": "2025-07-14T12:34:29.622289"
  },
  {
    "epoch": 1568,
    "batch_idx": 96,
    "loss": 0.3905051648616791,
    "learning_rate": 0.00011076884921648833,
    "timestep_mean": 473.4375,
    "epsilon_norm": 8.82714557647705,
    "prediction_norm": 7.279679775238037,
    "timestamp": "2025-07-14T12:34:29.628072"
  },
  {
    "epoch": 1568,
    "batch_idx": 128,
    "loss": 0.36793050169944763,
    "learning_rate": 0.00011076884921648833,
    "timestep_mean": 435.40625,
    "epsilon_norm": 7.833518028259277,
    "prediction_norm": 6.443249702453613,
    "timestamp": "2025-07-14T12:34:29.633235"
  },
  {
    "epoch": 1568,
    "batch_idx": 160,
    "loss": 0.22714346647262573,
    "learning_rate": 0.00011076884921648833,
    "timestep_mean": 512.0625,
    "epsilon_norm": 8.07400894165039,
    "prediction_norm": 7.193734169006348,
    "timestamp": "2025-07-14T12:34:29.639030"
  },
  {
    "epoch": 1568,
    "batch_idx": 192,
    "loss": 0.44199520349502563,
    "learning_rate": 0.00011076884921648833,
    "timestep_mean": 627.625,
    "epsilon_norm": 4.756828784942627,
    "prediction_norm": 4.473270416259766,
    "timestamp": "2025-07-14T12:34:29.644279"
  },
  {
    "epoch": 1569,
    "batch_idx": 0,
    "loss": 0.13789141178131104,
    "learning_rate": 0.00011027634197146929,
    "timestep_mean": 515.4375,
    "epsilon_norm": 6.652789115905762,
    "prediction_norm": 5.966996669769287,
    "timestamp": "2025-07-14T12:34:29.649623"
  },
  {
    "epoch": 1569,
    "batch_idx": 32,
    "loss": 0.22463372349739075,
    "learning_rate": 0.00011027634197146929,
    "timestep_mean": 391.65625,
    "epsilon_norm": 7.972354888916016,
    "prediction_norm": 6.531433582305908,
    "timestamp": "2025-07-14T12:34:29.656097"
  },
  {
    "epoch": 1569,
    "batch_idx": 64,
    "loss": 0.21580788493156433,
    "learning_rate": 0.00011027634197146929,
    "timestep_mean": 521.0625,
    "epsilon_norm": 8.642111778259277,
    "prediction_norm": 7.845979690551758,
    "timestamp": "2025-07-14T12:34:29.661855"
  },
  {
    "epoch": 1569,
    "batch_idx": 96,
    "loss": 0.25083115696907043,
    "learning_rate": 0.00011027634197146929,
    "timestep_mean": 412.21875,
    "epsilon_norm": 6.778383255004883,
    "prediction_norm": 5.495146751403809,
    "timestamp": "2025-07-14T12:34:29.668701"
  },
  {
    "epoch": 1569,
    "batch_idx": 128,
    "loss": 0.2046995311975479,
    "learning_rate": 0.00011027634197146929,
    "timestep_mean": 529.40625,
    "epsilon_norm": 7.689508438110352,
    "prediction_norm": 7.2361907958984375,
    "timestamp": "2025-07-14T12:34:29.674246"
  },
  {
    "epoch": 1569,
    "batch_idx": 160,
    "loss": 0.18775424361228943,
    "learning_rate": 0.00011027634197146929,
    "timestep_mean": 582.125,
    "epsilon_norm": 8.29208755493164,
    "prediction_norm": 7.53196382522583,
    "timestamp": "2025-07-14T12:34:29.680783"
  },
  {
    "epoch": 1569,
    "batch_idx": 192,
    "loss": 0.840717613697052,
    "learning_rate": 0.00011027634197146929,
    "timestep_mean": 495.5,
    "epsilon_norm": 5.012826919555664,
    "prediction_norm": 3.3812575340270996,
    "timestamp": "2025-07-14T12:34:29.686554"
  },
  {
    "epoch": 1570,
    "batch_idx": 0,
    "loss": 0.27397871017456055,
    "learning_rate": 0.0001097847963308351,
    "timestep_mean": 581.03125,
    "epsilon_norm": 8.281070709228516,
    "prediction_norm": 6.687669277191162,
    "timestamp": "2025-07-14T12:34:29.694521"
  },
  {
    "epoch": 1570,
    "batch_idx": 32,
    "loss": 0.25342193245887756,
    "learning_rate": 0.0001097847963308351,
    "timestep_mean": 513.09375,
    "epsilon_norm": 7.166887283325195,
    "prediction_norm": 6.644359111785889,
    "timestamp": "2025-07-14T12:34:29.701200"
  },
  {
    "epoch": 1570,
    "batch_idx": 64,
    "loss": 0.2835776209831238,
    "learning_rate": 0.0001097847963308351,
    "timestep_mean": 514.9375,
    "epsilon_norm": 7.932623863220215,
    "prediction_norm": 6.550640106201172,
    "timestamp": "2025-07-14T12:34:29.707111"
  },
  {
    "epoch": 1570,
    "batch_idx": 96,
    "loss": 0.15967699885368347,
    "learning_rate": 0.0001097847963308351,
    "timestep_mean": 578.6875,
    "epsilon_norm": 7.862384796142578,
    "prediction_norm": 7.532113552093506,
    "timestamp": "2025-07-14T12:34:29.717723"
  },
  {
    "epoch": 1570,
    "batch_idx": 128,
    "loss": 0.32473188638687134,
    "learning_rate": 0.0001097847963308351,
    "timestep_mean": 522.3125,
    "epsilon_norm": 7.837809085845947,
    "prediction_norm": 7.140661239624023,
    "timestamp": "2025-07-14T12:34:29.727447"
  },
  {
    "epoch": 1570,
    "batch_idx": 160,
    "loss": 0.29696863889694214,
    "learning_rate": 0.0001097847963308351,
    "timestep_mean": 478.6875,
    "epsilon_norm": 6.880119323730469,
    "prediction_norm": 6.0224289894104,
    "timestamp": "2025-07-14T12:34:29.735984"
  },
  {
    "epoch": 1570,
    "batch_idx": 192,
    "loss": 0.5237424373626709,
    "learning_rate": 0.0001097847963308351,
    "timestep_mean": 589.75,
    "epsilon_norm": 3.7994208335876465,
    "prediction_norm": 2.5516345500946045,
    "timestamp": "2025-07-14T12:34:29.742500"
  },
  {
    "epoch": 1571,
    "batch_idx": 0,
    "loss": 0.3287964463233948,
    "learning_rate": 0.00010929421350742601,
    "timestep_mean": 412.9375,
    "epsilon_norm": 8.320685386657715,
    "prediction_norm": 7.4929938316345215,
    "timestamp": "2025-07-14T12:34:29.752447"
  },
  {
    "epoch": 1571,
    "batch_idx": 32,
    "loss": 0.2279188632965088,
    "learning_rate": 0.00010929421350742601,
    "timestep_mean": 507.78125,
    "epsilon_norm": 8.635009765625,
    "prediction_norm": 7.649865627288818,
    "timestamp": "2025-07-14T12:34:29.758991"
  },
  {
    "epoch": 1571,
    "batch_idx": 64,
    "loss": 0.21886023879051208,
    "learning_rate": 0.00010929421350742601,
    "timestep_mean": 441.28125,
    "epsilon_norm": 8.451838493347168,
    "prediction_norm": 7.221818923950195,
    "timestamp": "2025-07-14T12:34:29.767385"
  },
  {
    "epoch": 1571,
    "batch_idx": 96,
    "loss": 0.20246775448322296,
    "learning_rate": 0.00010929421350742601,
    "timestep_mean": 534.28125,
    "epsilon_norm": 7.7925238609313965,
    "prediction_norm": 7.0385212898254395,
    "timestamp": "2025-07-14T12:34:29.780079"
  },
  {
    "epoch": 1571,
    "batch_idx": 128,
    "loss": 0.24940107762813568,
    "learning_rate": 0.00010929421350742601,
    "timestep_mean": 413.15625,
    "epsilon_norm": 8.152665138244629,
    "prediction_norm": 7.531879425048828,
    "timestamp": "2025-07-14T12:34:29.787697"
  },
  {
    "epoch": 1571,
    "batch_idx": 160,
    "loss": 0.48360583186149597,
    "learning_rate": 0.00010929421350742601,
    "timestep_mean": 550.0625,
    "epsilon_norm": 8.358461380004883,
    "prediction_norm": 7.50698709487915,
    "timestamp": "2025-07-14T12:34:29.796946"
  },
  {
    "epoch": 1571,
    "batch_idx": 192,
    "loss": 0.2619602382183075,
    "learning_rate": 0.00010929421350742601,
    "timestep_mean": 451.625,
    "epsilon_norm": 4.46924352645874,
    "prediction_norm": 4.1443772315979,
    "timestamp": "2025-07-14T12:34:29.806091"
  },
  {
    "epoch": 1572,
    "batch_idx": 0,
    "loss": 0.144734188914299,
    "learning_rate": 0.00010880459471170594,
    "timestep_mean": 567.40625,
    "epsilon_norm": 7.068885803222656,
    "prediction_norm": 6.520621299743652,
    "timestamp": "2025-07-14T12:34:29.813918"
  },
  {
    "epoch": 1572,
    "batch_idx": 32,
    "loss": 0.23107066750526428,
    "learning_rate": 0.00010880459471170594,
    "timestep_mean": 493.65625,
    "epsilon_norm": 7.944366455078125,
    "prediction_norm": 7.150104522705078,
    "timestamp": "2025-07-14T12:34:29.820187"
  },
  {
    "epoch": 1572,
    "batch_idx": 64,
    "loss": 0.23149938881397247,
    "learning_rate": 0.00010880459471170594,
    "timestep_mean": 456.21875,
    "epsilon_norm": 7.332962989807129,
    "prediction_norm": 6.863330364227295,
    "timestamp": "2025-07-14T12:34:29.826329"
  },
  {
    "epoch": 1572,
    "batch_idx": 96,
    "loss": 0.6464698314666748,
    "learning_rate": 0.00010880459471170594,
    "timestep_mean": 499.34375,
    "epsilon_norm": 8.686545372009277,
    "prediction_norm": 5.88995361328125,
    "timestamp": "2025-07-14T12:34:29.832057"
  },
  {
    "epoch": 1572,
    "batch_idx": 128,
    "loss": 0.43380439281463623,
    "learning_rate": 0.00010880459471170594,
    "timestep_mean": 483.5,
    "epsilon_norm": 8.319099426269531,
    "prediction_norm": 6.693519115447998,
    "timestamp": "2025-07-14T12:34:29.837873"
  },
  {
    "epoch": 1572,
    "batch_idx": 160,
    "loss": 0.25650060176849365,
    "learning_rate": 0.00010880459471170594,
    "timestep_mean": 531.25,
    "epsilon_norm": 7.82680606842041,
    "prediction_norm": 6.619521141052246,
    "timestamp": "2025-07-14T12:34:29.845477"
  },
  {
    "epoch": 1572,
    "batch_idx": 192,
    "loss": 0.027326054871082306,
    "learning_rate": 0.00010880459471170594,
    "timestep_mean": 557.5,
    "epsilon_norm": 3.261359930038452,
    "prediction_norm": 3.34816837310791,
    "timestamp": "2025-07-14T12:34:29.853714"
  },
  {
    "epoch": 1573,
    "batch_idx": 0,
    "loss": 0.1965392678976059,
    "learning_rate": 0.00010831594115176085,
    "timestep_mean": 511.4375,
    "epsilon_norm": 6.426974773406982,
    "prediction_norm": 5.900690078735352,
    "timestamp": "2025-07-14T12:34:29.860594"
  },
  {
    "epoch": 1573,
    "batch_idx": 32,
    "loss": 0.37061017751693726,
    "learning_rate": 0.00010831594115176085,
    "timestep_mean": 475.90625,
    "epsilon_norm": 8.58107852935791,
    "prediction_norm": 6.898112773895264,
    "timestamp": "2025-07-14T12:34:29.866658"
  },
  {
    "epoch": 1573,
    "batch_idx": 64,
    "loss": 0.32657262682914734,
    "learning_rate": 0.00010831594115176085,
    "timestep_mean": 512.03125,
    "epsilon_norm": 8.925621032714844,
    "prediction_norm": 7.791799068450928,
    "timestamp": "2025-07-14T12:34:29.872507"
  },
  {
    "epoch": 1573,
    "batch_idx": 96,
    "loss": 0.3781220316886902,
    "learning_rate": 0.00010831594115176085,
    "timestep_mean": 414.09375,
    "epsilon_norm": 8.360479354858398,
    "prediction_norm": 6.794602870941162,
    "timestamp": "2025-07-14T12:34:29.878924"
  },
  {
    "epoch": 1573,
    "batch_idx": 128,
    "loss": 0.25834938883781433,
    "learning_rate": 0.00010831594115176085,
    "timestep_mean": 481.59375,
    "epsilon_norm": 8.129734992980957,
    "prediction_norm": 7.6085052490234375,
    "timestamp": "2025-07-14T12:34:29.887018"
  },
  {
    "epoch": 1573,
    "batch_idx": 160,
    "loss": 0.3496602475643158,
    "learning_rate": 0.00010831594115176085,
    "timestep_mean": 523.96875,
    "epsilon_norm": 7.630840301513672,
    "prediction_norm": 5.911010265350342,
    "timestamp": "2025-07-14T12:34:29.893563"
  },
  {
    "epoch": 1573,
    "batch_idx": 192,
    "loss": 0.429851233959198,
    "learning_rate": 0.00010831594115176085,
    "timestep_mean": 507.25,
    "epsilon_norm": 4.038400650024414,
    "prediction_norm": 3.309624433517456,
    "timestamp": "2025-07-14T12:34:29.899342"
  },
  {
    "epoch": 1574,
    "batch_idx": 0,
    "loss": 0.21072907745838165,
    "learning_rate": 0.00010782825403329504,
    "timestep_mean": 477.75,
    "epsilon_norm": 7.245016098022461,
    "prediction_norm": 6.970213413238525,
    "timestamp": "2025-07-14T12:34:29.904903"
  },
  {
    "epoch": 1574,
    "batch_idx": 32,
    "loss": 0.22540327906608582,
    "learning_rate": 0.00010782825403329504,
    "timestep_mean": 454.28125,
    "epsilon_norm": 7.0738844871521,
    "prediction_norm": 6.061661243438721,
    "timestamp": "2025-07-14T12:34:29.909967"
  },
  {
    "epoch": 1574,
    "batch_idx": 64,
    "loss": 0.17139479517936707,
    "learning_rate": 0.00010782825403329504,
    "timestep_mean": 569.5625,
    "epsilon_norm": 8.904210090637207,
    "prediction_norm": 7.888211250305176,
    "timestamp": "2025-07-14T12:34:29.915046"
  },
  {
    "epoch": 1574,
    "batch_idx": 96,
    "loss": 0.2529679536819458,
    "learning_rate": 0.00010782825403329504,
    "timestep_mean": 489.53125,
    "epsilon_norm": 7.396631240844727,
    "prediction_norm": 5.607349872589111,
    "timestamp": "2025-07-14T12:34:29.920083"
  },
  {
    "epoch": 1574,
    "batch_idx": 128,
    "loss": 0.2894781827926636,
    "learning_rate": 0.00010782825403329504,
    "timestep_mean": 624.53125,
    "epsilon_norm": 9.189996719360352,
    "prediction_norm": 8.004903793334961,
    "timestamp": "2025-07-14T12:34:29.925123"
  },
  {
    "epoch": 1574,
    "batch_idx": 160,
    "loss": 0.36024972796440125,
    "learning_rate": 0.00010782825403329504,
    "timestep_mean": 473.96875,
    "epsilon_norm": 8.686341285705566,
    "prediction_norm": 7.044755458831787,
    "timestamp": "2025-07-14T12:34:29.930068"
  },
  {
    "epoch": 1574,
    "batch_idx": 192,
    "loss": 0.19838522374629974,
    "learning_rate": 0.00010782825403329504,
    "timestep_mean": 597.25,
    "epsilon_norm": 3.1963894367218018,
    "prediction_norm": 2.484748125076294,
    "timestamp": "2025-07-14T12:34:29.935050"
  },
  {
    "epoch": 1575,
    "batch_idx": 0,
    "loss": 0.2051803171634674,
    "learning_rate": 0.00010734153455962745,
    "timestep_mean": 463.78125,
    "epsilon_norm": 7.3419294357299805,
    "prediction_norm": 6.544029712677002,
    "timestamp": "2025-07-14T12:34:29.939973"
  },
  {
    "epoch": 1575,
    "batch_idx": 32,
    "loss": 0.17446543276309967,
    "learning_rate": 0.00010734153455962745,
    "timestep_mean": 554.875,
    "epsilon_norm": 9.081202507019043,
    "prediction_norm": 7.734866619110107,
    "timestamp": "2025-07-14T12:34:29.944748"
  },
  {
    "epoch": 1575,
    "batch_idx": 64,
    "loss": 0.16727761924266815,
    "learning_rate": 0.00010734153455962745,
    "timestep_mean": 493.40625,
    "epsilon_norm": 8.261914253234863,
    "prediction_norm": 7.43677282333374,
    "timestamp": "2025-07-14T12:34:29.949485"
  },
  {
    "epoch": 1575,
    "batch_idx": 96,
    "loss": 0.2583927810192108,
    "learning_rate": 0.00010734153455962745,
    "timestep_mean": 474.125,
    "epsilon_norm": 7.862672805786133,
    "prediction_norm": 6.6120734214782715,
    "timestamp": "2025-07-14T12:34:29.954375"
  },
  {
    "epoch": 1575,
    "batch_idx": 128,
    "loss": 0.0791245847940445,
    "learning_rate": 0.00010734153455962745,
    "timestep_mean": 565.78125,
    "epsilon_norm": 7.372291564941406,
    "prediction_norm": 7.381539344787598,
    "timestamp": "2025-07-14T12:34:29.959184"
  },
  {
    "epoch": 1575,
    "batch_idx": 160,
    "loss": 0.210723876953125,
    "learning_rate": 0.00010734153455962745,
    "timestep_mean": 561.34375,
    "epsilon_norm": 7.757397651672363,
    "prediction_norm": 7.177745342254639,
    "timestamp": "2025-07-14T12:34:29.963927"
  },
  {
    "epoch": 1575,
    "batch_idx": 192,
    "loss": 0.25559520721435547,
    "learning_rate": 0.00010734153455962745,
    "timestep_mean": 689.0,
    "epsilon_norm": 2.8391783237457275,
    "prediction_norm": 2.437992811203003,
    "timestamp": "2025-07-14T12:34:29.968884"
  },
  {
    "epoch": 1576,
    "batch_idx": 0,
    "loss": 0.24093259871006012,
    "learning_rate": 0.00010685578393169069,
    "timestep_mean": 562.84375,
    "epsilon_norm": 8.027363777160645,
    "prediction_norm": 6.828562259674072,
    "timestamp": "2025-07-14T12:34:29.973645"
  },
  {
    "epoch": 1576,
    "batch_idx": 32,
    "loss": 0.21440137922763824,
    "learning_rate": 0.00010685578393169069,
    "timestep_mean": 552.1875,
    "epsilon_norm": 7.975935459136963,
    "prediction_norm": 6.988868713378906,
    "timestamp": "2025-07-14T12:34:29.978430"
  },
  {
    "epoch": 1576,
    "batch_idx": 64,
    "loss": 0.3381480574607849,
    "learning_rate": 0.00010685578393169069,
    "timestep_mean": 495.625,
    "epsilon_norm": 9.122363090515137,
    "prediction_norm": 8.215439796447754,
    "timestamp": "2025-07-14T12:34:29.983144"
  },
  {
    "epoch": 1576,
    "batch_idx": 96,
    "loss": 0.2037261724472046,
    "learning_rate": 0.00010685578393169069,
    "timestep_mean": 486.375,
    "epsilon_norm": 8.490313529968262,
    "prediction_norm": 7.700342655181885,
    "timestamp": "2025-07-14T12:34:29.988021"
  },
  {
    "epoch": 1576,
    "batch_idx": 128,
    "loss": 0.2391907423734665,
    "learning_rate": 0.00010685578393169069,
    "timestep_mean": 511.53125,
    "epsilon_norm": 7.266188621520996,
    "prediction_norm": 6.250715255737305,
    "timestamp": "2025-07-14T12:34:29.992811"
  },
  {
    "epoch": 1576,
    "batch_idx": 160,
    "loss": 0.2825385332107544,
    "learning_rate": 0.00010685578393169069,
    "timestep_mean": 475.9375,
    "epsilon_norm": 8.198785781860352,
    "prediction_norm": 7.001073360443115,
    "timestamp": "2025-07-14T12:34:29.998034"
  },
  {
    "epoch": 1576,
    "batch_idx": 192,
    "loss": 0.07583140581846237,
    "learning_rate": 0.00010685578393169069,
    "timestep_mean": 641.875,
    "epsilon_norm": 3.011657238006592,
    "prediction_norm": 2.7489781379699707,
    "timestamp": "2025-07-14T12:34:30.003155"
  },
  {
    "epoch": 1577,
    "batch_idx": 0,
    "loss": 0.1549525260925293,
    "learning_rate": 0.00010637100334802539,
    "timestep_mean": 547.03125,
    "epsilon_norm": 9.045284271240234,
    "prediction_norm": 7.701938152313232,
    "timestamp": "2025-07-14T12:34:30.008015"
  },
  {
    "epoch": 1577,
    "batch_idx": 32,
    "loss": 0.29681944847106934,
    "learning_rate": 0.00010637100334802539,
    "timestep_mean": 466.34375,
    "epsilon_norm": 7.275196075439453,
    "prediction_norm": 6.03593111038208,
    "timestamp": "2025-07-14T12:34:30.013015"
  },
  {
    "epoch": 1577,
    "batch_idx": 64,
    "loss": 0.5169349908828735,
    "learning_rate": 0.00010637100334802539,
    "timestep_mean": 390.09375,
    "epsilon_norm": 8.558162689208984,
    "prediction_norm": 5.586620807647705,
    "timestamp": "2025-07-14T12:34:30.017868"
  },
  {
    "epoch": 1577,
    "batch_idx": 96,
    "loss": 0.2536221146583557,
    "learning_rate": 0.00010637100334802539,
    "timestep_mean": 415.375,
    "epsilon_norm": 7.384922027587891,
    "prediction_norm": 6.011805534362793,
    "timestamp": "2025-07-14T12:34:30.022610"
  },
  {
    "epoch": 1577,
    "batch_idx": 128,
    "loss": 0.474229097366333,
    "learning_rate": 0.00010637100334802539,
    "timestep_mean": 481.375,
    "epsilon_norm": 9.27880573272705,
    "prediction_norm": 7.767898082733154,
    "timestamp": "2025-07-14T12:34:30.027370"
  },
  {
    "epoch": 1577,
    "batch_idx": 160,
    "loss": 0.29320406913757324,
    "learning_rate": 0.00010637100334802539,
    "timestep_mean": 485.90625,
    "epsilon_norm": 7.462148666381836,
    "prediction_norm": 5.854889392852783,
    "timestamp": "2025-07-14T12:34:30.032114"
  },
  {
    "epoch": 1577,
    "batch_idx": 192,
    "loss": 0.36316564679145813,
    "learning_rate": 0.00010637100334802539,
    "timestep_mean": 354.75,
    "epsilon_norm": 3.3134090900421143,
    "prediction_norm": 2.386615514755249,
    "timestamp": "2025-07-14T12:34:30.036932"
  },
  {
    "epoch": 1578,
    "batch_idx": 0,
    "loss": 0.38851282000541687,
    "learning_rate": 0.00010588719400478001,
    "timestep_mean": 391.46875,
    "epsilon_norm": 8.037755966186523,
    "prediction_norm": 5.848872184753418,
    "timestamp": "2025-07-14T12:34:30.041801"
  },
  {
    "epoch": 1578,
    "batch_idx": 32,
    "loss": 0.2355508953332901,
    "learning_rate": 0.00010588719400478001,
    "timestep_mean": 526.53125,
    "epsilon_norm": 8.293147087097168,
    "prediction_norm": 6.726813793182373,
    "timestamp": "2025-07-14T12:34:30.046632"
  },
  {
    "epoch": 1578,
    "batch_idx": 64,
    "loss": 0.2702833116054535,
    "learning_rate": 0.00010588719400478001,
    "timestep_mean": 495.125,
    "epsilon_norm": 6.662450790405273,
    "prediction_norm": 5.931442737579346,
    "timestamp": "2025-07-14T12:34:30.051360"
  },
  {
    "epoch": 1578,
    "batch_idx": 96,
    "loss": 0.22543585300445557,
    "learning_rate": 0.00010588719400478001,
    "timestep_mean": 505.5,
    "epsilon_norm": 7.55393648147583,
    "prediction_norm": 6.807605266571045,
    "timestamp": "2025-07-14T12:34:30.056265"
  },
  {
    "epoch": 1578,
    "batch_idx": 128,
    "loss": 0.3520019054412842,
    "learning_rate": 0.00010588719400478001,
    "timestep_mean": 507.40625,
    "epsilon_norm": 9.336483001708984,
    "prediction_norm": 7.601484298706055,
    "timestamp": "2025-07-14T12:34:30.061050"
  },
  {
    "epoch": 1578,
    "batch_idx": 160,
    "loss": 0.437593936920166,
    "learning_rate": 0.00010588719400478001,
    "timestep_mean": 469.40625,
    "epsilon_norm": 8.798795700073242,
    "prediction_norm": 6.95728874206543,
    "timestamp": "2025-07-14T12:34:30.065878"
  },
  {
    "epoch": 1578,
    "batch_idx": 192,
    "loss": 0.19614240527153015,
    "learning_rate": 0.00010588719400478001,
    "timestep_mean": 561.375,
    "epsilon_norm": 3.4093737602233887,
    "prediction_norm": 3.0873992443084717,
    "timestamp": "2025-07-14T12:34:30.070636"
  },
  {
    "epoch": 1579,
    "batch_idx": 0,
    "loss": 0.295969694852829,
    "learning_rate": 0.00010540435709570582,
    "timestep_mean": 589.4375,
    "epsilon_norm": 8.812966346740723,
    "prediction_norm": 7.417081832885742,
    "timestamp": "2025-07-14T12:34:30.075452"
  },
  {
    "epoch": 1579,
    "batch_idx": 32,
    "loss": 0.3803956210613251,
    "learning_rate": 0.00010540435709570582,
    "timestep_mean": 431.09375,
    "epsilon_norm": 8.188084602355957,
    "prediction_norm": 6.656446933746338,
    "timestamp": "2025-07-14T12:34:30.080339"
  },
  {
    "epoch": 1579,
    "batch_idx": 64,
    "loss": 0.4722568988800049,
    "learning_rate": 0.00010540435709570582,
    "timestep_mean": 438.46875,
    "epsilon_norm": 8.187244415283203,
    "prediction_norm": 6.848501682281494,
    "timestamp": "2025-07-14T12:34:30.085124"
  },
  {
    "epoch": 1579,
    "batch_idx": 96,
    "loss": 0.14031928777694702,
    "learning_rate": 0.00010540435709570582,
    "timestep_mean": 470.6875,
    "epsilon_norm": 7.294956207275391,
    "prediction_norm": 6.633695125579834,
    "timestamp": "2025-07-14T12:34:30.090002"
  },
  {
    "epoch": 1579,
    "batch_idx": 128,
    "loss": 0.26729869842529297,
    "learning_rate": 0.00010540435709570582,
    "timestep_mean": 531.125,
    "epsilon_norm": 7.528897285461426,
    "prediction_norm": 6.326754570007324,
    "timestamp": "2025-07-14T12:34:30.094857"
  },
  {
    "epoch": 1579,
    "batch_idx": 160,
    "loss": 0.21951454877853394,
    "learning_rate": 0.00010540435709570582,
    "timestep_mean": 548.0,
    "epsilon_norm": 8.543169021606445,
    "prediction_norm": 7.085776329040527,
    "timestamp": "2025-07-14T12:34:30.099779"
  },
  {
    "epoch": 1579,
    "batch_idx": 192,
    "loss": 0.7387605905532837,
    "learning_rate": 0.00010540435709570582,
    "timestep_mean": 305.375,
    "epsilon_norm": 4.324221611022949,
    "prediction_norm": 3.0441055297851562,
    "timestamp": "2025-07-14T12:34:30.104762"
  },
  {
    "epoch": 1580,
    "batch_idx": 0,
    "loss": 0.2815955877304077,
    "learning_rate": 0.00010492249381215478,
    "timestep_mean": 489.125,
    "epsilon_norm": 7.997186183929443,
    "prediction_norm": 6.446052074432373,
    "timestamp": "2025-07-14T12:34:30.109651"
  },
  {
    "epoch": 1580,
    "batch_idx": 32,
    "loss": 0.34187835454940796,
    "learning_rate": 0.00010492249381215478,
    "timestep_mean": 485.03125,
    "epsilon_norm": 7.628741264343262,
    "prediction_norm": 5.9527268409729,
    "timestamp": "2025-07-14T12:34:30.114750"
  },
  {
    "epoch": 1580,
    "batch_idx": 64,
    "loss": 0.21582841873168945,
    "learning_rate": 0.00010492249381215478,
    "timestep_mean": 507.59375,
    "epsilon_norm": 8.40201473236084,
    "prediction_norm": 7.307766914367676,
    "timestamp": "2025-07-14T12:34:30.119828"
  },
  {
    "epoch": 1580,
    "batch_idx": 96,
    "loss": 0.622205376625061,
    "learning_rate": 0.00010492249381215478,
    "timestep_mean": 466.53125,
    "epsilon_norm": 8.956745147705078,
    "prediction_norm": 6.593619346618652,
    "timestamp": "2025-07-14T12:34:30.124893"
  },
  {
    "epoch": 1580,
    "batch_idx": 128,
    "loss": 0.23561863601207733,
    "learning_rate": 0.00010492249381215478,
    "timestep_mean": 540.28125,
    "epsilon_norm": 7.50980806350708,
    "prediction_norm": 6.511740684509277,
    "timestamp": "2025-07-14T12:34:30.130209"
  },
  {
    "epoch": 1580,
    "batch_idx": 160,
    "loss": 0.2528175115585327,
    "learning_rate": 0.00010492249381215478,
    "timestep_mean": 474.4375,
    "epsilon_norm": 7.44687032699585,
    "prediction_norm": 6.193431377410889,
    "timestamp": "2025-07-14T12:34:30.135201"
  },
  {
    "epoch": 1580,
    "batch_idx": 192,
    "loss": 0.21489384770393372,
    "learning_rate": 0.00010492249381215478,
    "timestep_mean": 393.75,
    "epsilon_norm": 4.080293655395508,
    "prediction_norm": 4.270230770111084,
    "timestamp": "2025-07-14T12:34:30.140203"
  },
  {
    "epoch": 1581,
    "batch_idx": 0,
    "loss": 0.27378690242767334,
    "learning_rate": 0.00010444160534307688,
    "timestep_mean": 505.40625,
    "epsilon_norm": 7.839380741119385,
    "prediction_norm": 6.932971954345703,
    "timestamp": "2025-07-14T12:34:30.145412"
  },
  {
    "epoch": 1581,
    "batch_idx": 32,
    "loss": 0.2666659951210022,
    "learning_rate": 0.00010444160534307688,
    "timestep_mean": 531.28125,
    "epsilon_norm": 7.638272285461426,
    "prediction_norm": 6.808403968811035,
    "timestamp": "2025-07-14T12:34:30.150708"
  },
  {
    "epoch": 1581,
    "batch_idx": 64,
    "loss": 0.19368226826190948,
    "learning_rate": 0.00010444160534307688,
    "timestep_mean": 562.78125,
    "epsilon_norm": 8.289929389953613,
    "prediction_norm": 7.182590007781982,
    "timestamp": "2025-07-14T12:34:30.155825"
  },
  {
    "epoch": 1581,
    "batch_idx": 96,
    "loss": 0.33538907766342163,
    "learning_rate": 0.00010444160534307688,
    "timestep_mean": 465.21875,
    "epsilon_norm": 8.495017051696777,
    "prediction_norm": 6.361772060394287,
    "timestamp": "2025-07-14T12:34:30.160896"
  },
  {
    "epoch": 1581,
    "batch_idx": 128,
    "loss": 0.2748844623565674,
    "learning_rate": 0.00010444160534307688,
    "timestep_mean": 502.3125,
    "epsilon_norm": 8.096685409545898,
    "prediction_norm": 6.160703659057617,
    "timestamp": "2025-07-14T12:34:30.165924"
  },
  {
    "epoch": 1581,
    "batch_idx": 160,
    "loss": 0.2661733627319336,
    "learning_rate": 0.00010444160534307688,
    "timestep_mean": 497.8125,
    "epsilon_norm": 7.6629509925842285,
    "prediction_norm": 6.093275547027588,
    "timestamp": "2025-07-14T12:34:30.171274"
  },
  {
    "epoch": 1581,
    "batch_idx": 192,
    "loss": 0.14645427465438843,
    "learning_rate": 0.00010444160534307688,
    "timestep_mean": 637.0,
    "epsilon_norm": 4.088371276855469,
    "prediction_norm": 3.581571102142334,
    "timestamp": "2025-07-14T12:34:30.176668"
  },
  {
    "epoch": 1582,
    "batch_idx": 0,
    "loss": 0.2688525319099426,
    "learning_rate": 0.00010396169287501648,
    "timestep_mean": 485.09375,
    "epsilon_norm": 7.477025032043457,
    "prediction_norm": 6.323188304901123,
    "timestamp": "2025-07-14T12:34:30.182009"
  },
  {
    "epoch": 1582,
    "batch_idx": 32,
    "loss": 0.21340316534042358,
    "learning_rate": 0.00010396169287501648,
    "timestep_mean": 430.46875,
    "epsilon_norm": 7.2983808517456055,
    "prediction_norm": 6.136902809143066,
    "timestamp": "2025-07-14T12:34:30.187573"
  },
  {
    "epoch": 1582,
    "batch_idx": 64,
    "loss": 0.34360021352767944,
    "learning_rate": 0.00010396169287501648,
    "timestep_mean": 408.9375,
    "epsilon_norm": 8.249303817749023,
    "prediction_norm": 6.463435173034668,
    "timestamp": "2025-07-14T12:34:30.192766"
  },
  {
    "epoch": 1582,
    "batch_idx": 96,
    "loss": 0.16401006281375885,
    "learning_rate": 0.00010396169287501648,
    "timestep_mean": 548.53125,
    "epsilon_norm": 7.700774669647217,
    "prediction_norm": 7.233951568603516,
    "timestamp": "2025-07-14T12:34:30.198013"
  },
  {
    "epoch": 1582,
    "batch_idx": 128,
    "loss": 0.12553545832633972,
    "learning_rate": 0.00010396169287501648,
    "timestep_mean": 577.9375,
    "epsilon_norm": 8.352926254272461,
    "prediction_norm": 8.123409271240234,
    "timestamp": "2025-07-14T12:34:30.203279"
  },
  {
    "epoch": 1582,
    "batch_idx": 160,
    "loss": 0.278944194316864,
    "learning_rate": 0.00010396169287501648,
    "timestep_mean": 499.375,
    "epsilon_norm": 7.208281517028809,
    "prediction_norm": 6.097188949584961,
    "timestamp": "2025-07-14T12:34:30.208377"
  },
  {
    "epoch": 1582,
    "batch_idx": 192,
    "loss": 0.9621139764785767,
    "learning_rate": 0.00010396169287501648,
    "timestep_mean": 308.25,
    "epsilon_norm": 4.490245819091797,
    "prediction_norm": 2.683912515640259,
    "timestamp": "2025-07-14T12:34:30.213652"
  },
  {
    "epoch": 1583,
    "batch_idx": 0,
    "loss": 0.2299378365278244,
    "learning_rate": 0.00010348275759210993,
    "timestep_mean": 488.3125,
    "epsilon_norm": 7.748589515686035,
    "prediction_norm": 6.560176849365234,
    "timestamp": "2025-07-14T12:34:30.219033"
  },
  {
    "epoch": 1583,
    "batch_idx": 32,
    "loss": 0.1076168566942215,
    "learning_rate": 0.00010348275759210993,
    "timestep_mean": 654.3125,
    "epsilon_norm": 6.5936994552612305,
    "prediction_norm": 6.130700588226318,
    "timestamp": "2025-07-14T12:34:30.224241"
  },
  {
    "epoch": 1583,
    "batch_idx": 64,
    "loss": 0.2468569129705429,
    "learning_rate": 0.00010348275759210993,
    "timestep_mean": 545.8125,
    "epsilon_norm": 8.923182487487793,
    "prediction_norm": 7.88500452041626,
    "timestamp": "2025-07-14T12:34:30.229795"
  },
  {
    "epoch": 1583,
    "batch_idx": 96,
    "loss": 0.3309857249259949,
    "learning_rate": 0.00010348275759210993,
    "timestep_mean": 490.65625,
    "epsilon_norm": 8.217890739440918,
    "prediction_norm": 7.427607536315918,
    "timestamp": "2025-07-14T12:34:30.235210"
  },
  {
    "epoch": 1583,
    "batch_idx": 128,
    "loss": 0.17005230486392975,
    "learning_rate": 0.00010348275759210993,
    "timestep_mean": 513.5625,
    "epsilon_norm": 8.092580795288086,
    "prediction_norm": 7.222377777099609,
    "timestamp": "2025-07-14T12:34:30.240526"
  },
  {
    "epoch": 1583,
    "batch_idx": 160,
    "loss": 0.3246561884880066,
    "learning_rate": 0.00010348275759210993,
    "timestep_mean": 503.46875,
    "epsilon_norm": 7.868429660797119,
    "prediction_norm": 6.588009834289551,
    "timestamp": "2025-07-14T12:34:30.245918"
  },
  {
    "epoch": 1583,
    "batch_idx": 192,
    "loss": 0.07832947373390198,
    "learning_rate": 0.00010348275759210993,
    "timestep_mean": 585.25,
    "epsilon_norm": 3.8810482025146484,
    "prediction_norm": 3.689678192138672,
    "timestamp": "2025-07-14T12:34:30.251357"
  },
  {
    "epoch": 1584,
    "batch_idx": 0,
    "loss": 0.12995274364948273,
    "learning_rate": 0.00010300480067608228,
    "timestep_mean": 516.3125,
    "epsilon_norm": 7.7632317543029785,
    "prediction_norm": 6.989648342132568,
    "timestamp": "2025-07-14T12:34:30.256564"
  },
  {
    "epoch": 1584,
    "batch_idx": 32,
    "loss": 0.32725441455841064,
    "learning_rate": 0.00010300480067608228,
    "timestep_mean": 445.15625,
    "epsilon_norm": 7.16769552230835,
    "prediction_norm": 5.6439690589904785,
    "timestamp": "2025-07-14T12:34:30.261878"
  },
  {
    "epoch": 1584,
    "batch_idx": 64,
    "loss": 0.26113361120224,
    "learning_rate": 0.00010300480067608228,
    "timestep_mean": 556.40625,
    "epsilon_norm": 8.745866775512695,
    "prediction_norm": 7.500193119049072,
    "timestamp": "2025-07-14T12:34:30.267154"
  },
  {
    "epoch": 1584,
    "batch_idx": 96,
    "loss": 0.2620105743408203,
    "learning_rate": 0.00010300480067608228,
    "timestep_mean": 510.03125,
    "epsilon_norm": 7.758127212524414,
    "prediction_norm": 6.807224750518799,
    "timestamp": "2025-07-14T12:34:30.272330"
  },
  {
    "epoch": 1584,
    "batch_idx": 128,
    "loss": 0.1842377781867981,
    "learning_rate": 0.00010300480067608228,
    "timestep_mean": 571.3125,
    "epsilon_norm": 8.408918380737305,
    "prediction_norm": 7.76188850402832,
    "timestamp": "2025-07-14T12:34:30.277701"
  },
  {
    "epoch": 1584,
    "batch_idx": 160,
    "loss": 0.15622064471244812,
    "learning_rate": 0.00010300480067608228,
    "timestep_mean": 586.46875,
    "epsilon_norm": 8.740291595458984,
    "prediction_norm": 8.342117309570312,
    "timestamp": "2025-07-14T12:34:30.283131"
  },
  {
    "epoch": 1584,
    "batch_idx": 192,
    "loss": 0.7131417989730835,
    "learning_rate": 0.00010300480067608228,
    "timestep_mean": 339.25,
    "epsilon_norm": 4.427000045776367,
    "prediction_norm": 2.6916725635528564,
    "timestamp": "2025-07-14T12:34:30.288452"
  },
  {
    "epoch": 1585,
    "batch_idx": 0,
    "loss": 0.15254108607769012,
    "learning_rate": 0.00010252782330624507,
    "timestep_mean": 496.5625,
    "epsilon_norm": 7.068444728851318,
    "prediction_norm": 6.439239978790283,
    "timestamp": "2025-07-14T12:34:30.294015"
  },
  {
    "epoch": 1585,
    "batch_idx": 32,
    "loss": 0.2465566098690033,
    "learning_rate": 0.00010252782330624507,
    "timestep_mean": 538.25,
    "epsilon_norm": 8.570427894592285,
    "prediction_norm": 7.417644023895264,
    "timestamp": "2025-07-14T12:34:30.299323"
  },
  {
    "epoch": 1585,
    "batch_idx": 64,
    "loss": 0.13826006650924683,
    "learning_rate": 0.00010252782330624507,
    "timestep_mean": 557.09375,
    "epsilon_norm": 7.047816276550293,
    "prediction_norm": 6.330308437347412,
    "timestamp": "2025-07-14T12:34:30.304485"
  },
  {
    "epoch": 1585,
    "batch_idx": 96,
    "loss": 0.24785125255584717,
    "learning_rate": 0.00010252782330624507,
    "timestep_mean": 483.8125,
    "epsilon_norm": 6.692006587982178,
    "prediction_norm": 5.446567058563232,
    "timestamp": "2025-07-14T12:34:30.309651"
  },
  {
    "epoch": 1585,
    "batch_idx": 128,
    "loss": 0.20885904133319855,
    "learning_rate": 0.00010252782330624507,
    "timestep_mean": 555.09375,
    "epsilon_norm": 8.630014419555664,
    "prediction_norm": 7.746399402618408,
    "timestamp": "2025-07-14T12:34:30.314943"
  },
  {
    "epoch": 1585,
    "batch_idx": 160,
    "loss": 0.23445077240467072,
    "learning_rate": 0.00010252782330624507,
    "timestep_mean": 537.65625,
    "epsilon_norm": 7.3759942054748535,
    "prediction_norm": 6.65660285949707,
    "timestamp": "2025-07-14T12:34:30.320273"
  },
  {
    "epoch": 1585,
    "batch_idx": 192,
    "loss": 0.3325083255767822,
    "learning_rate": 0.00010252782330624507,
    "timestep_mean": 567.625,
    "epsilon_norm": 4.396931171417236,
    "prediction_norm": 4.013654708862305,
    "timestamp": "2025-07-14T12:34:30.325580"
  },
  {
    "epoch": 1586,
    "batch_idx": 0,
    "loss": 0.13656602799892426,
    "learning_rate": 0.00010205182665949206,
    "timestep_mean": 591.9375,
    "epsilon_norm": 8.13277816772461,
    "prediction_norm": 7.276416301727295,
    "timestamp": "2025-07-14T12:34:30.330760"
  },
  {
    "epoch": 1586,
    "batch_idx": 32,
    "loss": 0.2900284230709076,
    "learning_rate": 0.00010205182665949206,
    "timestep_mean": 462.21875,
    "epsilon_norm": 8.42331600189209,
    "prediction_norm": 7.124423980712891,
    "timestamp": "2025-07-14T12:34:30.336486"
  },
  {
    "epoch": 1586,
    "batch_idx": 64,
    "loss": 0.3067670464515686,
    "learning_rate": 0.00010205182665949206,
    "timestep_mean": 529.21875,
    "epsilon_norm": 8.64120864868164,
    "prediction_norm": 7.438283920288086,
    "timestamp": "2025-07-14T12:34:30.341806"
  },
  {
    "epoch": 1586,
    "batch_idx": 96,
    "loss": 0.4417765140533447,
    "learning_rate": 0.00010205182665949206,
    "timestep_mean": 385.28125,
    "epsilon_norm": 8.580362319946289,
    "prediction_norm": 6.456559658050537,
    "timestamp": "2025-07-14T12:34:30.346941"
  },
  {
    "epoch": 1586,
    "batch_idx": 128,
    "loss": 0.29105544090270996,
    "learning_rate": 0.00010205182665949206,
    "timestep_mean": 471.0,
    "epsilon_norm": 9.133306503295898,
    "prediction_norm": 7.694543838500977,
    "timestamp": "2025-07-14T12:34:30.352169"
  },
  {
    "epoch": 1586,
    "batch_idx": 160,
    "loss": 0.29634031653404236,
    "learning_rate": 0.00010205182665949206,
    "timestep_mean": 551.34375,
    "epsilon_norm": 8.678995132446289,
    "prediction_norm": 7.400663375854492,
    "timestamp": "2025-07-14T12:34:30.357420"
  },
  {
    "epoch": 1586,
    "batch_idx": 192,
    "loss": 0.058609090745449066,
    "learning_rate": 0.00010205182665949206,
    "timestep_mean": 627.5,
    "epsilon_norm": 2.698122978210449,
    "prediction_norm": 2.8772082328796387,
    "timestamp": "2025-07-14T12:34:30.362914"
  },
  {
    "epoch": 1587,
    "batch_idx": 0,
    "loss": 0.24895623326301575,
    "learning_rate": 0.00010157681191029802,
    "timestep_mean": 440.0,
    "epsilon_norm": 6.5835676193237305,
    "prediction_norm": 5.4695916175842285,
    "timestamp": "2025-07-14T12:34:30.368340"
  },
  {
    "epoch": 1587,
    "batch_idx": 32,
    "loss": 0.18612879514694214,
    "learning_rate": 0.00010157681191029802,
    "timestep_mean": 509.875,
    "epsilon_norm": 7.643773555755615,
    "prediction_norm": 6.418459415435791,
    "timestamp": "2025-07-14T12:34:30.373543"
  },
  {
    "epoch": 1587,
    "batch_idx": 64,
    "loss": 0.3260948359966278,
    "learning_rate": 0.00010157681191029802,
    "timestep_mean": 552.125,
    "epsilon_norm": 8.219435691833496,
    "prediction_norm": 6.899010181427002,
    "timestamp": "2025-07-14T12:34:30.378993"
  },
  {
    "epoch": 1587,
    "batch_idx": 96,
    "loss": 0.3314792513847351,
    "learning_rate": 0.00010157681191029802,
    "timestep_mean": 521.71875,
    "epsilon_norm": 8.6160306930542,
    "prediction_norm": 7.165603160858154,
    "timestamp": "2025-07-14T12:34:30.384457"
  },
  {
    "epoch": 1587,
    "batch_idx": 128,
    "loss": 0.2887064516544342,
    "learning_rate": 0.00010157681191029802,
    "timestep_mean": 504.9375,
    "epsilon_norm": 8.537382125854492,
    "prediction_norm": 6.926114082336426,
    "timestamp": "2025-07-14T12:34:30.389764"
  },
  {
    "epoch": 1587,
    "batch_idx": 160,
    "loss": 0.2277626395225525,
    "learning_rate": 0.00010157681191029802,
    "timestep_mean": 544.65625,
    "epsilon_norm": 8.504213333129883,
    "prediction_norm": 7.41097354888916,
    "timestamp": "2025-07-14T12:34:30.395111"
  },
  {
    "epoch": 1587,
    "batch_idx": 192,
    "loss": 0.15952551364898682,
    "learning_rate": 0.00010157681191029802,
    "timestep_mean": 473.25,
    "epsilon_norm": 4.971020221710205,
    "prediction_norm": 4.822575092315674,
    "timestamp": "2025-07-14T12:34:30.400448"
  },
  {
    "epoch": 1588,
    "batch_idx": 0,
    "loss": 0.1591702401638031,
    "learning_rate": 0.00010110278023071453,
    "timestep_mean": 534.78125,
    "epsilon_norm": 8.495912551879883,
    "prediction_norm": 7.641912937164307,
    "timestamp": "2025-07-14T12:34:30.405592"
  },
  {
    "epoch": 1588,
    "batch_idx": 32,
    "loss": 0.2237314134836197,
    "learning_rate": 0.00010110278023071453,
    "timestep_mean": 465.90625,
    "epsilon_norm": 8.241998672485352,
    "prediction_norm": 7.078372955322266,
    "timestamp": "2025-07-14T12:34:30.410576"
  },
  {
    "epoch": 1588,
    "batch_idx": 64,
    "loss": 0.3233976662158966,
    "learning_rate": 0.00010110278023071453,
    "timestep_mean": 514.0625,
    "epsilon_norm": 9.588772773742676,
    "prediction_norm": 7.341683387756348,
    "timestamp": "2025-07-14T12:34:30.415885"
  },
  {
    "epoch": 1588,
    "batch_idx": 96,
    "loss": 0.2488279491662979,
    "learning_rate": 0.00010110278023071453,
    "timestep_mean": 495.125,
    "epsilon_norm": 6.845458984375,
    "prediction_norm": 5.290665149688721,
    "timestamp": "2025-07-14T12:34:30.421321"
  },
  {
    "epoch": 1588,
    "batch_idx": 128,
    "loss": 0.2348659187555313,
    "learning_rate": 0.00010110278023071453,
    "timestep_mean": 557.15625,
    "epsilon_norm": 7.376533031463623,
    "prediction_norm": 6.798353672027588,
    "timestamp": "2025-07-14T12:34:30.426377"
  },
  {
    "epoch": 1588,
    "batch_idx": 160,
    "loss": 0.2084055244922638,
    "learning_rate": 0.00010110278023071453,
    "timestep_mean": 540.59375,
    "epsilon_norm": 7.134897232055664,
    "prediction_norm": 6.244055271148682,
    "timestamp": "2025-07-14T12:34:30.431406"
  },
  {
    "epoch": 1588,
    "batch_idx": 192,
    "loss": 0.34949302673339844,
    "learning_rate": 0.00010110278023071453,
    "timestep_mean": 455.375,
    "epsilon_norm": 4.645226955413818,
    "prediction_norm": 4.073399543762207,
    "timestamp": "2025-07-14T12:34:30.436481"
  },
  {
    "epoch": 1589,
    "batch_idx": 0,
    "loss": 0.15673816204071045,
    "learning_rate": 0.00010062973279036749,
    "timestep_mean": 556.0,
    "epsilon_norm": 8.098479270935059,
    "prediction_norm": 7.568258762359619,
    "timestamp": "2025-07-14T12:34:30.441789"
  },
  {
    "epoch": 1589,
    "batch_idx": 32,
    "loss": 0.34198421239852905,
    "learning_rate": 0.00010062973279036749,
    "timestep_mean": 478.5,
    "epsilon_norm": 7.726077079772949,
    "prediction_norm": 6.241794109344482,
    "timestamp": "2025-07-14T12:34:30.446924"
  },
  {
    "epoch": 1589,
    "batch_idx": 64,
    "loss": 0.180082768201828,
    "learning_rate": 0.00010062973279036749,
    "timestep_mean": 517.15625,
    "epsilon_norm": 8.006647109985352,
    "prediction_norm": 7.927257061004639,
    "timestamp": "2025-07-14T12:34:30.451971"
  },
  {
    "epoch": 1589,
    "batch_idx": 96,
    "loss": 0.19524382054805756,
    "learning_rate": 0.00010062973279036749,
    "timestep_mean": 494.09375,
    "epsilon_norm": 8.274006843566895,
    "prediction_norm": 7.497666358947754,
    "timestamp": "2025-07-14T12:34:30.456928"
  },
  {
    "epoch": 1589,
    "batch_idx": 128,
    "loss": 0.21009942889213562,
    "learning_rate": 0.00010062973279036749,
    "timestep_mean": 489.59375,
    "epsilon_norm": 7.620820999145508,
    "prediction_norm": 6.508971691131592,
    "timestamp": "2025-07-14T12:34:30.461979"
  },
  {
    "epoch": 1589,
    "batch_idx": 160,
    "loss": 0.27978515625,
    "learning_rate": 0.00010062973279036749,
    "timestep_mean": 546.375,
    "epsilon_norm": 8.341979026794434,
    "prediction_norm": 7.410219192504883,
    "timestamp": "2025-07-14T12:34:30.467092"
  },
  {
    "epoch": 1589,
    "batch_idx": 192,
    "loss": 0.33032527565956116,
    "learning_rate": 0.00010062973279036749,
    "timestep_mean": 366.875,
    "epsilon_norm": 3.213505268096924,
    "prediction_norm": 2.513408660888672,
    "timestamp": "2025-07-14T12:34:30.472078"
  },
  {
    "epoch": 1590,
    "batch_idx": 0,
    "loss": 0.136166051030159,
    "learning_rate": 0.00010015767075645485,
    "timestep_mean": 531.96875,
    "epsilon_norm": 7.885730266571045,
    "prediction_norm": 7.2847900390625,
    "timestamp": "2025-07-14T12:34:30.477238"
  },
  {
    "epoch": 1590,
    "batch_idx": 32,
    "loss": 0.3990749418735504,
    "learning_rate": 0.00010015767075645485,
    "timestep_mean": 454.90625,
    "epsilon_norm": 7.818434715270996,
    "prediction_norm": 5.849135398864746,
    "timestamp": "2025-07-14T12:34:30.482187"
  },
  {
    "epoch": 1590,
    "batch_idx": 64,
    "loss": 0.3399975597858429,
    "learning_rate": 0.00010015767075645485,
    "timestep_mean": 496.3125,
    "epsilon_norm": 8.413538932800293,
    "prediction_norm": 6.66342830657959,
    "timestamp": "2025-07-14T12:34:30.487172"
  },
  {
    "epoch": 1590,
    "batch_idx": 96,
    "loss": 0.25746673345565796,
    "learning_rate": 0.00010015767075645485,
    "timestep_mean": 468.375,
    "epsilon_norm": 7.117221355438232,
    "prediction_norm": 6.169222354888916,
    "timestamp": "2025-07-14T12:34:30.492327"
  },
  {
    "epoch": 1590,
    "batch_idx": 128,
    "loss": 0.15588463842868805,
    "learning_rate": 0.00010015767075645485,
    "timestep_mean": 539.90625,
    "epsilon_norm": 7.132470607757568,
    "prediction_norm": 6.313508987426758,
    "timestamp": "2025-07-14T12:34:30.497307"
  },
  {
    "epoch": 1590,
    "batch_idx": 160,
    "loss": 0.11228485405445099,
    "learning_rate": 0.00010015767075645485,
    "timestep_mean": 591.21875,
    "epsilon_norm": 7.943477630615234,
    "prediction_norm": 7.487477779388428,
    "timestamp": "2025-07-14T12:34:30.502221"
  },
  {
    "epoch": 1590,
    "batch_idx": 192,
    "loss": 0.03811581805348396,
    "learning_rate": 0.00010015767075645485,
    "timestep_mean": 507.75,
    "epsilon_norm": 4.068193435668945,
    "prediction_norm": 3.958226442337036,
    "timestamp": "2025-07-14T12:34:30.507222"
  },
  {
    "epoch": 1591,
    "batch_idx": 0,
    "loss": 0.1610913872718811,
    "learning_rate": 9.968659529374216e-05,
    "timestep_mean": 490.4375,
    "epsilon_norm": 7.659003257751465,
    "prediction_norm": 6.932651996612549,
    "timestamp": "2025-07-14T12:34:30.512227"
  },
  {
    "epoch": 1591,
    "batch_idx": 32,
    "loss": 0.22874413430690765,
    "learning_rate": 9.968659529374216e-05,
    "timestep_mean": 462.21875,
    "epsilon_norm": 6.636128902435303,
    "prediction_norm": 6.056674480438232,
    "timestamp": "2025-07-14T12:34:30.517274"
  },
  {
    "epoch": 1591,
    "batch_idx": 64,
    "loss": 0.3323083519935608,
    "learning_rate": 9.968659529374216e-05,
    "timestep_mean": 508.3125,
    "epsilon_norm": 8.257285118103027,
    "prediction_norm": 6.6289873123168945,
    "timestamp": "2025-07-14T12:34:30.522238"
  },
  {
    "epoch": 1591,
    "batch_idx": 96,
    "loss": 0.20670434832572937,
    "learning_rate": 9.968659529374216e-05,
    "timestep_mean": 542.09375,
    "epsilon_norm": 8.359848976135254,
    "prediction_norm": 7.329441070556641,
    "timestamp": "2025-07-14T12:34:30.527165"
  },
  {
    "epoch": 1591,
    "batch_idx": 128,
    "loss": 0.33354002237319946,
    "learning_rate": 9.968659529374216e-05,
    "timestep_mean": 472.9375,
    "epsilon_norm": 8.724211692810059,
    "prediction_norm": 7.206012725830078,
    "timestamp": "2025-07-14T12:34:30.532095"
  },
  {
    "epoch": 1591,
    "batch_idx": 160,
    "loss": 0.2465486228466034,
    "learning_rate": 9.968659529374216e-05,
    "timestep_mean": 514.5,
    "epsilon_norm": 8.004231452941895,
    "prediction_norm": 7.260164737701416,
    "timestamp": "2025-07-14T12:34:30.537016"
  },
  {
    "epoch": 1591,
    "batch_idx": 192,
    "loss": 0.22928917407989502,
    "learning_rate": 9.968659529374216e-05,
    "timestep_mean": 591.5,
    "epsilon_norm": 3.587132215499878,
    "prediction_norm": 3.2275357246398926,
    "timestamp": "2025-07-14T12:34:30.542125"
  },
  {
    "epoch": 1592,
    "batch_idx": 0,
    "loss": 0.2655852437019348,
    "learning_rate": 9.921650756456161e-05,
    "timestep_mean": 486.1875,
    "epsilon_norm": 8.084613800048828,
    "prediction_norm": 6.5243916511535645,
    "timestamp": "2025-07-14T12:34:30.547073"
  },
  {
    "epoch": 1592,
    "batch_idx": 32,
    "loss": 0.2863750457763672,
    "learning_rate": 9.921650756456161e-05,
    "timestep_mean": 456.0625,
    "epsilon_norm": 7.402961254119873,
    "prediction_norm": 5.920513153076172,
    "timestamp": "2025-07-14T12:34:30.552020"
  },
  {
    "epoch": 1592,
    "batch_idx": 64,
    "loss": 0.2724490761756897,
    "learning_rate": 9.921650756456161e-05,
    "timestep_mean": 480.375,
    "epsilon_norm": 8.080435752868652,
    "prediction_norm": 6.813820838928223,
    "timestamp": "2025-07-14T12:34:30.556915"
  },
  {
    "epoch": 1592,
    "batch_idx": 96,
    "loss": 0.32222679257392883,
    "learning_rate": 9.921650756456161e-05,
    "timestep_mean": 514.40625,
    "epsilon_norm": 7.953101634979248,
    "prediction_norm": 6.848811626434326,
    "timestamp": "2025-07-14T12:34:30.561892"
  },
  {
    "epoch": 1592,
    "batch_idx": 128,
    "loss": 0.2620559334754944,
    "learning_rate": 9.921650756456161e-05,
    "timestep_mean": 382.15625,
    "epsilon_norm": 7.51122522354126,
    "prediction_norm": 6.61642599105835,
    "timestamp": "2025-07-14T12:34:30.566991"
  },
  {
    "epoch": 1592,
    "batch_idx": 160,
    "loss": 0.3845175802707672,
    "learning_rate": 9.921650756456161e-05,
    "timestep_mean": 411.0,
    "epsilon_norm": 7.530187129974365,
    "prediction_norm": 5.2707695960998535,
    "timestamp": "2025-07-14T12:34:30.571942"
  },
  {
    "epoch": 1592,
    "batch_idx": 192,
    "loss": 0.19428043067455292,
    "learning_rate": 9.921650756456161e-05,
    "timestep_mean": 475.75,
    "epsilon_norm": 3.746506452560425,
    "prediction_norm": 3.0993492603302,
    "timestamp": "2025-07-14T12:34:30.576976"
  },
  {
    "epoch": 1593,
    "batch_idx": 0,
    "loss": 0.18011915683746338,
    "learning_rate": 9.87474087288082e-05,
    "timestep_mean": 545.28125,
    "epsilon_norm": 6.947481155395508,
    "prediction_norm": 5.683330059051514,
    "timestamp": "2025-07-14T12:34:30.582033"
  },
  {
    "epoch": 1593,
    "batch_idx": 32,
    "loss": 0.25983861088752747,
    "learning_rate": 9.87474087288082e-05,
    "timestep_mean": 479.5,
    "epsilon_norm": 7.636862754821777,
    "prediction_norm": 6.299045085906982,
    "timestamp": "2025-07-14T12:34:30.587106"
  },
  {
    "epoch": 1593,
    "batch_idx": 64,
    "loss": 0.1834721714258194,
    "learning_rate": 9.87474087288082e-05,
    "timestep_mean": 559.625,
    "epsilon_norm": 8.104763984680176,
    "prediction_norm": 7.604894638061523,
    "timestamp": "2025-07-14T12:34:30.592109"
  },
  {
    "epoch": 1593,
    "batch_idx": 96,
    "loss": 0.17780835926532745,
    "learning_rate": 9.87474087288082e-05,
    "timestep_mean": 527.25,
    "epsilon_norm": 7.977485179901123,
    "prediction_norm": 7.124844551086426,
    "timestamp": "2025-07-14T12:34:30.597134"
  },
  {
    "epoch": 1593,
    "batch_idx": 128,
    "loss": 0.22089345753192902,
    "learning_rate": 9.87474087288082e-05,
    "timestep_mean": 510.8125,
    "epsilon_norm": 8.257174491882324,
    "prediction_norm": 7.156461715698242,
    "timestamp": "2025-07-14T12:34:30.602137"
  },
  {
    "epoch": 1593,
    "batch_idx": 160,
    "loss": 0.2828361392021179,
    "learning_rate": 9.87474087288082e-05,
    "timestep_mean": 449.0,
    "epsilon_norm": 7.620790958404541,
    "prediction_norm": 6.582470417022705,
    "timestamp": "2025-07-14T12:34:30.607199"
  },
  {
    "epoch": 1593,
    "batch_idx": 192,
    "loss": 0.16201171278953552,
    "learning_rate": 9.87474087288082e-05,
    "timestep_mean": 505.625,
    "epsilon_norm": 5.426341533660889,
    "prediction_norm": 5.049317359924316,
    "timestamp": "2025-07-14T12:34:30.612191"
  },
  {
    "epoch": 1594,
    "batch_idx": 0,
    "loss": 0.22679468989372253,
    "learning_rate": 9.827929994393617e-05,
    "timestep_mean": 477.53125,
    "epsilon_norm": 7.063409328460693,
    "prediction_norm": 6.252468585968018,
    "timestamp": "2025-07-14T12:34:30.617294"
  },
  {
    "epoch": 1594,
    "batch_idx": 32,
    "loss": 0.23494163155555725,
    "learning_rate": 9.827929994393617e-05,
    "timestep_mean": 499.125,
    "epsilon_norm": 7.472464084625244,
    "prediction_norm": 5.835034370422363,
    "timestamp": "2025-07-14T12:34:30.622362"
  },
  {
    "epoch": 1594,
    "batch_idx": 64,
    "loss": 0.21779942512512207,
    "learning_rate": 9.827929994393617e-05,
    "timestep_mean": 595.59375,
    "epsilon_norm": 6.73370361328125,
    "prediction_norm": 6.0943284034729,
    "timestamp": "2025-07-14T12:34:30.627580"
  },
  {
    "epoch": 1594,
    "batch_idx": 96,
    "loss": 0.1371414065361023,
    "learning_rate": 9.827929994393617e-05,
    "timestep_mean": 537.59375,
    "epsilon_norm": 7.773827075958252,
    "prediction_norm": 7.616247653961182,
    "timestamp": "2025-07-14T12:34:30.632765"
  },
  {
    "epoch": 1594,
    "batch_idx": 128,
    "loss": 0.39917027950286865,
    "learning_rate": 9.827929994393617e-05,
    "timestep_mean": 433.21875,
    "epsilon_norm": 7.790289878845215,
    "prediction_norm": 6.37545919418335,
    "timestamp": "2025-07-14T12:34:30.637802"
  },
  {
    "epoch": 1594,
    "batch_idx": 160,
    "loss": 0.13035258650779724,
    "learning_rate": 9.827929994393617e-05,
    "timestep_mean": 546.8125,
    "epsilon_norm": 8.867729187011719,
    "prediction_norm": 8.010820388793945,
    "timestamp": "2025-07-14T12:34:30.642766"
  },
  {
    "epoch": 1594,
    "batch_idx": 192,
    "loss": 0.3558306097984314,
    "learning_rate": 9.827929994393617e-05,
    "timestep_mean": 327.125,
    "epsilon_norm": 3.0220446586608887,
    "prediction_norm": 2.064758539199829,
    "timestamp": "2025-07-14T12:34:30.647927"
  },
  {
    "epoch": 1595,
    "batch_idx": 0,
    "loss": 0.3603764772415161,
    "learning_rate": 9.781218236495773e-05,
    "timestep_mean": 521.96875,
    "epsilon_norm": 8.048054695129395,
    "prediction_norm": 6.9389872550964355,
    "timestamp": "2025-07-14T12:34:30.652959"
  },
  {
    "epoch": 1595,
    "batch_idx": 32,
    "loss": 0.3290151357650757,
    "learning_rate": 9.781218236495773e-05,
    "timestep_mean": 481.96875,
    "epsilon_norm": 7.860074043273926,
    "prediction_norm": 6.452149391174316,
    "timestamp": "2025-07-14T12:34:30.658012"
  },
  {
    "epoch": 1595,
    "batch_idx": 64,
    "loss": 0.25648409128189087,
    "learning_rate": 9.781218236495773e-05,
    "timestep_mean": 536.71875,
    "epsilon_norm": 7.169713497161865,
    "prediction_norm": 6.059906005859375,
    "timestamp": "2025-07-14T12:34:30.663215"
  },
  {
    "epoch": 1595,
    "batch_idx": 96,
    "loss": 0.16634777188301086,
    "learning_rate": 9.781218236495773e-05,
    "timestep_mean": 609.3125,
    "epsilon_norm": 6.904179096221924,
    "prediction_norm": 6.0647077560424805,
    "timestamp": "2025-07-14T12:34:30.668147"
  },
  {
    "epoch": 1595,
    "batch_idx": 128,
    "loss": 0.30336764454841614,
    "learning_rate": 9.781218236495773e-05,
    "timestep_mean": 544.1875,
    "epsilon_norm": 8.036155700683594,
    "prediction_norm": 7.066164493560791,
    "timestamp": "2025-07-14T12:34:30.673209"
  },
  {
    "epoch": 1595,
    "batch_idx": 160,
    "loss": 0.15600454807281494,
    "learning_rate": 9.781218236495773e-05,
    "timestep_mean": 499.875,
    "epsilon_norm": 7.331184387207031,
    "prediction_norm": 6.451560020446777,
    "timestamp": "2025-07-14T12:34:30.678167"
  },
  {
    "epoch": 1595,
    "batch_idx": 192,
    "loss": 0.15707743167877197,
    "learning_rate": 9.781218236495773e-05,
    "timestep_mean": 610.0,
    "epsilon_norm": 4.365102767944336,
    "prediction_norm": 3.8839547634124756,
    "timestamp": "2025-07-14T12:34:30.683163"
  },
  {
    "epoch": 1596,
    "batch_idx": 0,
    "loss": 0.2432713359594345,
    "learning_rate": 9.734605714443904e-05,
    "timestep_mean": 464.8125,
    "epsilon_norm": 8.232738494873047,
    "prediction_norm": 6.4458231925964355,
    "timestamp": "2025-07-14T12:34:30.688198"
  },
  {
    "epoch": 1596,
    "batch_idx": 32,
    "loss": 0.31143099069595337,
    "learning_rate": 9.734605714443904e-05,
    "timestep_mean": 458.65625,
    "epsilon_norm": 8.170022964477539,
    "prediction_norm": 6.703139781951904,
    "timestamp": "2025-07-14T12:34:30.693197"
  },
  {
    "epoch": 1596,
    "batch_idx": 64,
    "loss": 0.34542950987815857,
    "learning_rate": 9.734605714443904e-05,
    "timestep_mean": 465.90625,
    "epsilon_norm": 7.461779594421387,
    "prediction_norm": 5.502213954925537,
    "timestamp": "2025-07-14T12:34:30.698285"
  },
  {
    "epoch": 1596,
    "batch_idx": 96,
    "loss": 0.5501794219017029,
    "learning_rate": 9.734605714443904e-05,
    "timestep_mean": 484.6875,
    "epsilon_norm": 7.939319610595703,
    "prediction_norm": 5.728431224822998,
    "timestamp": "2025-07-14T12:34:30.703267"
  },
  {
    "epoch": 1596,
    "batch_idx": 128,
    "loss": 0.20173271000385284,
    "learning_rate": 9.734605714443904e-05,
    "timestep_mean": 488.21875,
    "epsilon_norm": 8.41444206237793,
    "prediction_norm": 7.550755977630615,
    "timestamp": "2025-07-14T12:34:30.708229"
  },
  {
    "epoch": 1596,
    "batch_idx": 160,
    "loss": 0.19772177934646606,
    "learning_rate": 9.734605714443904e-05,
    "timestep_mean": 539.28125,
    "epsilon_norm": 7.177724361419678,
    "prediction_norm": 6.212911605834961,
    "timestamp": "2025-07-14T12:34:30.713211"
  },
  {
    "epoch": 1596,
    "batch_idx": 192,
    "loss": 0.15021009743213654,
    "learning_rate": 9.734605714443904e-05,
    "timestep_mean": 535.125,
    "epsilon_norm": 3.417346715927124,
    "prediction_norm": 2.891519069671631,
    "timestamp": "2025-07-14T12:34:30.718186"
  },
  {
    "epoch": 1597,
    "batch_idx": 0,
    "loss": 0.2205040454864502,
    "learning_rate": 9.688092543249758e-05,
    "timestep_mean": 458.03125,
    "epsilon_norm": 8.114745140075684,
    "prediction_norm": 7.046867847442627,
    "timestamp": "2025-07-14T12:34:30.723297"
  },
  {
    "epoch": 1597,
    "batch_idx": 32,
    "loss": 0.46804922819137573,
    "learning_rate": 9.688092543249758e-05,
    "timestep_mean": 458.65625,
    "epsilon_norm": 7.893893718719482,
    "prediction_norm": 6.495285511016846,
    "timestamp": "2025-07-14T12:34:30.728427"
  },
  {
    "epoch": 1597,
    "batch_idx": 64,
    "loss": 0.17514611780643463,
    "learning_rate": 9.688092543249758e-05,
    "timestep_mean": 536.71875,
    "epsilon_norm": 7.427659511566162,
    "prediction_norm": 6.576311111450195,
    "timestamp": "2025-07-14T12:34:30.733417"
  },
  {
    "epoch": 1597,
    "batch_idx": 96,
    "loss": 0.21711108088493347,
    "learning_rate": 9.688092543249758e-05,
    "timestep_mean": 483.9375,
    "epsilon_norm": 7.638843536376953,
    "prediction_norm": 7.137660503387451,
    "timestamp": "2025-07-14T12:34:30.738460"
  },
  {
    "epoch": 1597,
    "batch_idx": 128,
    "loss": 0.26462745666503906,
    "learning_rate": 9.688092543249758e-05,
    "timestep_mean": 500.28125,
    "epsilon_norm": 7.5800018310546875,
    "prediction_norm": 6.859657287597656,
    "timestamp": "2025-07-14T12:34:30.743471"
  },
  {
    "epoch": 1597,
    "batch_idx": 160,
    "loss": 0.15132969617843628,
    "learning_rate": 9.688092543249758e-05,
    "timestep_mean": 541.03125,
    "epsilon_norm": 8.075704574584961,
    "prediction_norm": 7.229010581970215,
    "timestamp": "2025-07-14T12:34:30.748447"
  },
  {
    "epoch": 1597,
    "batch_idx": 192,
    "loss": 0.06837840378284454,
    "learning_rate": 9.688092543249758e-05,
    "timestep_mean": 572.625,
    "epsilon_norm": 3.6662814617156982,
    "prediction_norm": 3.8099796772003174,
    "timestamp": "2025-07-14T12:34:30.753531"
  },
  {
    "epoch": 1598,
    "batch_idx": 0,
    "loss": 0.45877015590667725,
    "learning_rate": 9.641678837679994e-05,
    "timestep_mean": 466.875,
    "epsilon_norm": 9.501448631286621,
    "prediction_norm": 7.454692840576172,
    "timestamp": "2025-07-14T12:34:30.758502"
  },
  {
    "epoch": 1598,
    "batch_idx": 32,
    "loss": 0.2288879156112671,
    "learning_rate": 9.641678837679994e-05,
    "timestep_mean": 508.4375,
    "epsilon_norm": 7.635885715484619,
    "prediction_norm": 6.952564239501953,
    "timestamp": "2025-07-14T12:34:30.763568"
  },
  {
    "epoch": 1598,
    "batch_idx": 64,
    "loss": 0.2537636458873749,
    "learning_rate": 9.641678837679994e-05,
    "timestep_mean": 472.84375,
    "epsilon_norm": 6.820833683013916,
    "prediction_norm": 6.008309841156006,
    "timestamp": "2025-07-14T12:34:30.768659"
  },
  {
    "epoch": 1598,
    "batch_idx": 96,
    "loss": 0.31961965560913086,
    "learning_rate": 9.641678837679994e-05,
    "timestep_mean": 486.78125,
    "epsilon_norm": 7.457187175750732,
    "prediction_norm": 5.384036064147949,
    "timestamp": "2025-07-14T12:34:30.773604"
  },
  {
    "epoch": 1598,
    "batch_idx": 128,
    "loss": 0.316892147064209,
    "learning_rate": 9.641678837679994e-05,
    "timestep_mean": 428.1875,
    "epsilon_norm": 7.598320484161377,
    "prediction_norm": 5.986454963684082,
    "timestamp": "2025-07-14T12:34:30.778668"
  },
  {
    "epoch": 1598,
    "batch_idx": 160,
    "loss": 0.2662302255630493,
    "learning_rate": 9.641678837679994e-05,
    "timestep_mean": 440.0,
    "epsilon_norm": 7.729323863983154,
    "prediction_norm": 6.692962169647217,
    "timestamp": "2025-07-14T12:34:30.783690"
  },
  {
    "epoch": 1598,
    "batch_idx": 192,
    "loss": 0.09919144213199615,
    "learning_rate": 9.641678837679994e-05,
    "timestep_mean": 570.5,
    "epsilon_norm": 3.127455234527588,
    "prediction_norm": 2.5353970527648926,
    "timestamp": "2025-07-14T12:34:30.788704"
  },
  {
    "epoch": 1599,
    "batch_idx": 0,
    "loss": 0.23431673645973206,
    "learning_rate": 9.595364712255771e-05,
    "timestep_mean": 510.65625,
    "epsilon_norm": 7.20519495010376,
    "prediction_norm": 6.180440425872803,
    "timestamp": "2025-07-14T12:34:30.793771"
  },
  {
    "epoch": 1599,
    "batch_idx": 32,
    "loss": 0.5324582457542419,
    "learning_rate": 9.595364712255771e-05,
    "timestep_mean": 450.21875,
    "epsilon_norm": 8.336162567138672,
    "prediction_norm": 6.303905963897705,
    "timestamp": "2025-07-14T12:34:30.798751"
  },
  {
    "epoch": 1599,
    "batch_idx": 64,
    "loss": 0.31606629490852356,
    "learning_rate": 9.595364712255771e-05,
    "timestep_mean": 449.625,
    "epsilon_norm": 7.757850170135498,
    "prediction_norm": 5.983731269836426,
    "timestamp": "2025-07-14T12:34:30.803695"
  },
  {
    "epoch": 1599,
    "batch_idx": 96,
    "loss": 0.28478336334228516,
    "learning_rate": 9.595364712255771e-05,
    "timestep_mean": 425.1875,
    "epsilon_norm": 7.405614376068115,
    "prediction_norm": 6.407025337219238,
    "timestamp": "2025-07-14T12:34:30.808739"
  },
  {
    "epoch": 1599,
    "batch_idx": 128,
    "loss": 0.2559497058391571,
    "learning_rate": 9.595364712255771e-05,
    "timestep_mean": 433.09375,
    "epsilon_norm": 8.60399341583252,
    "prediction_norm": 7.104994773864746,
    "timestamp": "2025-07-14T12:34:30.813687"
  },
  {
    "epoch": 1599,
    "batch_idx": 160,
    "loss": 0.32724106311798096,
    "learning_rate": 9.595364712255771e-05,
    "timestep_mean": 480.21875,
    "epsilon_norm": 8.212215423583984,
    "prediction_norm": 6.676442623138428,
    "timestamp": "2025-07-14T12:34:30.818700"
  },
  {
    "epoch": 1599,
    "batch_idx": 192,
    "loss": 0.5645890831947327,
    "learning_rate": 9.595364712255771e-05,
    "timestep_mean": 650.25,
    "epsilon_norm": 4.961700439453125,
    "prediction_norm": 3.7885076999664307,
    "timestamp": "2025-07-14T12:34:30.823694"
  },
  {
    "epoch": 1600,
    "batch_idx": 0,
    "loss": 0.1574067622423172,
    "learning_rate": 9.549150281252633e-05,
    "timestep_mean": 481.3125,
    "epsilon_norm": 7.780520915985107,
    "prediction_norm": 6.550908088684082,
    "timestamp": "2025-07-14T12:34:30.828690"
  },
  {
    "epoch": 1600,
    "batch_idx": 32,
    "loss": 0.2138034850358963,
    "learning_rate": 9.549150281252633e-05,
    "timestep_mean": 531.84375,
    "epsilon_norm": 7.085043430328369,
    "prediction_norm": 6.27816915512085,
    "timestamp": "2025-07-14T12:34:30.833762"
  },
  {
    "epoch": 1600,
    "batch_idx": 64,
    "loss": 0.4553736448287964,
    "learning_rate": 9.549150281252633e-05,
    "timestep_mean": 461.53125,
    "epsilon_norm": 8.654354095458984,
    "prediction_norm": 7.3939409255981445,
    "timestamp": "2025-07-14T12:34:30.838661"
  },
  {
    "epoch": 1600,
    "batch_idx": 96,
    "loss": 0.1230524554848671,
    "learning_rate": 9.549150281252633e-05,
    "timestep_mean": 531.28125,
    "epsilon_norm": 8.320843696594238,
    "prediction_norm": 7.51249361038208,
    "timestamp": "2025-07-14T12:34:30.843622"
  },
  {
    "epoch": 1600,
    "batch_idx": 128,
    "loss": 0.21754911541938782,
    "learning_rate": 9.549150281252633e-05,
    "timestep_mean": 606.09375,
    "epsilon_norm": 7.628001689910889,
    "prediction_norm": 6.498502254486084,
    "timestamp": "2025-07-14T12:34:30.848553"
  },
  {
    "epoch": 1600,
    "batch_idx": 160,
    "loss": 0.19575628638267517,
    "learning_rate": 9.549150281252633e-05,
    "timestep_mean": 487.9375,
    "epsilon_norm": 7.029559135437012,
    "prediction_norm": 6.253422737121582,
    "timestamp": "2025-07-14T12:34:30.853546"
  },
  {
    "epoch": 1600,
    "batch_idx": 192,
    "loss": 0.11559906601905823,
    "learning_rate": 9.549150281252633e-05,
    "timestep_mean": 487.0,
    "epsilon_norm": 3.5318634510040283,
    "prediction_norm": 3.1791181564331055,
    "timestamp": "2025-07-14T12:34:30.858485"
  },
  {
    "epoch": 1601,
    "batch_idx": 0,
    "loss": 0.29144981503486633,
    "learning_rate": 9.503035658700071e-05,
    "timestep_mean": 384.09375,
    "epsilon_norm": 6.603885173797607,
    "prediction_norm": 5.635254859924316,
    "timestamp": "2025-07-14T12:34:30.863448"
  },
  {
    "epoch": 1601,
    "batch_idx": 32,
    "loss": 0.4430517554283142,
    "learning_rate": 9.503035658700071e-05,
    "timestep_mean": 499.0625,
    "epsilon_norm": 8.810474395751953,
    "prediction_norm": 7.3861870765686035,
    "timestamp": "2025-07-14T12:34:30.868414"
  },
  {
    "epoch": 1601,
    "batch_idx": 64,
    "loss": 0.3511841595172882,
    "learning_rate": 9.503035658700071e-05,
    "timestep_mean": 554.65625,
    "epsilon_norm": 7.717482089996338,
    "prediction_norm": 5.802402496337891,
    "timestamp": "2025-07-14T12:34:30.873423"
  },
  {
    "epoch": 1601,
    "batch_idx": 96,
    "loss": 0.2404872179031372,
    "learning_rate": 9.503035658700071e-05,
    "timestep_mean": 378.21875,
    "epsilon_norm": 8.729752540588379,
    "prediction_norm": 7.6755690574646,
    "timestamp": "2025-07-14T12:34:30.878395"
  },
  {
    "epoch": 1601,
    "batch_idx": 128,
    "loss": 0.2637755274772644,
    "learning_rate": 9.503035658700071e-05,
    "timestep_mean": 619.21875,
    "epsilon_norm": 7.563835144042969,
    "prediction_norm": 6.3036980628967285,
    "timestamp": "2025-07-14T12:34:30.883271"
  },
  {
    "epoch": 1601,
    "batch_idx": 160,
    "loss": 0.3207564651966095,
    "learning_rate": 9.503035658700071e-05,
    "timestep_mean": 495.34375,
    "epsilon_norm": 7.551797866821289,
    "prediction_norm": 5.803242206573486,
    "timestamp": "2025-07-14T12:34:30.888332"
  },
  {
    "epoch": 1601,
    "batch_idx": 192,
    "loss": 0.17196768522262573,
    "learning_rate": 9.503035658700071e-05,
    "timestep_mean": 446.875,
    "epsilon_norm": 3.748394727706909,
    "prediction_norm": 3.18796706199646,
    "timestamp": "2025-07-14T12:34:30.893263"
  },
  {
    "epoch": 1602,
    "batch_idx": 0,
    "loss": 0.3229614198207855,
    "learning_rate": 9.457020958381334e-05,
    "timestep_mean": 615.8125,
    "epsilon_norm": 8.489006996154785,
    "prediction_norm": 6.826700687408447,
    "timestamp": "2025-07-14T12:34:30.898286"
  },
  {
    "epoch": 1602,
    "batch_idx": 32,
    "loss": 0.314281702041626,
    "learning_rate": 9.457020958381334e-05,
    "timestep_mean": 566.78125,
    "epsilon_norm": 8.036689758300781,
    "prediction_norm": 6.569421291351318,
    "timestamp": "2025-07-14T12:34:30.903233"
  },
  {
    "epoch": 1602,
    "batch_idx": 64,
    "loss": 0.27521538734436035,
    "learning_rate": 9.457020958381334e-05,
    "timestep_mean": 488.78125,
    "epsilon_norm": 8.009065628051758,
    "prediction_norm": 6.485709190368652,
    "timestamp": "2025-07-14T12:34:30.908137"
  },
  {
    "epoch": 1602,
    "batch_idx": 96,
    "loss": 0.22136659920215607,
    "learning_rate": 9.457020958381334e-05,
    "timestep_mean": 510.4375,
    "epsilon_norm": 7.162324905395508,
    "prediction_norm": 5.916132926940918,
    "timestamp": "2025-07-14T12:34:30.913211"
  },
  {
    "epoch": 1602,
    "batch_idx": 128,
    "loss": 0.2733209431171417,
    "learning_rate": 9.457020958381334e-05,
    "timestep_mean": 529.5,
    "epsilon_norm": 7.965820789337158,
    "prediction_norm": 6.321600914001465,
    "timestamp": "2025-07-14T12:34:30.918431"
  },
  {
    "epoch": 1602,
    "batch_idx": 160,
    "loss": 0.17664429545402527,
    "learning_rate": 9.457020958381334e-05,
    "timestep_mean": 479.34375,
    "epsilon_norm": 7.042276382446289,
    "prediction_norm": 6.415557384490967,
    "timestamp": "2025-07-14T12:34:30.923728"
  },
  {
    "epoch": 1602,
    "batch_idx": 192,
    "loss": 0.31555840373039246,
    "learning_rate": 9.457020958381334e-05,
    "timestep_mean": 481.25,
    "epsilon_norm": 3.511802911758423,
    "prediction_norm": 3.0049140453338623,
    "timestamp": "2025-07-14T12:34:30.928936"
  },
  {
    "epoch": 1603,
    "batch_idx": 0,
    "loss": 0.2054082751274109,
    "learning_rate": 9.411106293833115e-05,
    "timestep_mean": 534.875,
    "epsilon_norm": 7.493946552276611,
    "prediction_norm": 6.723252773284912,
    "timestamp": "2025-07-14T12:34:30.934151"
  },
  {
    "epoch": 1603,
    "batch_idx": 32,
    "loss": 0.30124974250793457,
    "learning_rate": 9.411106293833115e-05,
    "timestep_mean": 466.6875,
    "epsilon_norm": 7.528142929077148,
    "prediction_norm": 6.516303062438965,
    "timestamp": "2025-07-14T12:34:30.939558"
  },
  {
    "epoch": 1603,
    "batch_idx": 64,
    "loss": 0.18254642188549042,
    "learning_rate": 9.411106293833115e-05,
    "timestep_mean": 482.46875,
    "epsilon_norm": 7.4235520362854,
    "prediction_norm": 5.8706865310668945,
    "timestamp": "2025-07-14T12:34:30.944802"
  },
  {
    "epoch": 1603,
    "batch_idx": 96,
    "loss": 0.15148289501667023,
    "learning_rate": 9.411106293833115e-05,
    "timestep_mean": 559.625,
    "epsilon_norm": 6.224728107452393,
    "prediction_norm": 5.830883979797363,
    "timestamp": "2025-07-14T12:34:30.950074"
  },
  {
    "epoch": 1603,
    "batch_idx": 128,
    "loss": 0.14652103185653687,
    "learning_rate": 9.411106293833115e-05,
    "timestep_mean": 516.71875,
    "epsilon_norm": 7.569262981414795,
    "prediction_norm": 6.377439975738525,
    "timestamp": "2025-07-14T12:34:30.955027"
  },
  {
    "epoch": 1603,
    "batch_idx": 160,
    "loss": 0.16163155436515808,
    "learning_rate": 9.411106293833115e-05,
    "timestep_mean": 552.84375,
    "epsilon_norm": 7.987846851348877,
    "prediction_norm": 7.299023628234863,
    "timestamp": "2025-07-14T12:34:30.960313"
  },
  {
    "epoch": 1603,
    "batch_idx": 192,
    "loss": 0.705069363117218,
    "learning_rate": 9.411106293833115e-05,
    "timestep_mean": 468.125,
    "epsilon_norm": 3.7690699100494385,
    "prediction_norm": 2.860156774520874,
    "timestamp": "2025-07-14T12:34:30.965541"
  },
  {
    "epoch": 1604,
    "batch_idx": 0,
    "loss": 0.15713798999786377,
    "learning_rate": 9.365291778345303e-05,
    "timestep_mean": 593.78125,
    "epsilon_norm": 7.02039098739624,
    "prediction_norm": 6.030584812164307,
    "timestamp": "2025-07-14T12:34:30.970698"
  },
  {
    "epoch": 1604,
    "batch_idx": 32,
    "loss": 0.3130354881286621,
    "learning_rate": 9.365291778345303e-05,
    "timestep_mean": 467.6875,
    "epsilon_norm": 8.33342170715332,
    "prediction_norm": 6.457119464874268,
    "timestamp": "2025-07-14T12:34:30.976075"
  },
  {
    "epoch": 1604,
    "batch_idx": 64,
    "loss": 0.2830703556537628,
    "learning_rate": 9.365291778345303e-05,
    "timestep_mean": 573.65625,
    "epsilon_norm": 8.759608268737793,
    "prediction_norm": 7.677468776702881,
    "timestamp": "2025-07-14T12:34:30.981328"
  },
  {
    "epoch": 1604,
    "batch_idx": 96,
    "loss": 0.1759481132030487,
    "learning_rate": 9.365291778345303e-05,
    "timestep_mean": 526.28125,
    "epsilon_norm": 6.513105392456055,
    "prediction_norm": 5.650363445281982,
    "timestamp": "2025-07-14T12:34:30.986551"
  },
  {
    "epoch": 1604,
    "batch_idx": 128,
    "loss": 0.176622211933136,
    "learning_rate": 9.365291778345303e-05,
    "timestep_mean": 579.78125,
    "epsilon_norm": 7.147513389587402,
    "prediction_norm": 5.976682662963867,
    "timestamp": "2025-07-14T12:34:30.991846"
  },
  {
    "epoch": 1604,
    "batch_idx": 160,
    "loss": 0.22125136852264404,
    "learning_rate": 9.365291778345303e-05,
    "timestep_mean": 532.75,
    "epsilon_norm": 7.522520065307617,
    "prediction_norm": 7.090582370758057,
    "timestamp": "2025-07-14T12:34:30.997093"
  },
  {
    "epoch": 1604,
    "batch_idx": 192,
    "loss": 0.35250693559646606,
    "learning_rate": 9.365291778345303e-05,
    "timestep_mean": 438.625,
    "epsilon_norm": 4.96433687210083,
    "prediction_norm": 3.8020358085632324,
    "timestamp": "2025-07-14T12:34:31.002219"
  },
  {
    "epoch": 1605,
    "batch_idx": 0,
    "loss": 0.3382904529571533,
    "learning_rate": 9.319577524960654e-05,
    "timestep_mean": 385.59375,
    "epsilon_norm": 7.987601280212402,
    "prediction_norm": 6.206678867340088,
    "timestamp": "2025-07-14T12:34:31.007443"
  },
  {
    "epoch": 1605,
    "batch_idx": 32,
    "loss": 0.2532239556312561,
    "learning_rate": 9.319577524960654e-05,
    "timestep_mean": 477.09375,
    "epsilon_norm": 7.4292449951171875,
    "prediction_norm": 6.195138931274414,
    "timestamp": "2025-07-14T12:34:31.012898"
  },
  {
    "epoch": 1605,
    "batch_idx": 64,
    "loss": 0.378775030374527,
    "learning_rate": 9.319577524960654e-05,
    "timestep_mean": 504.65625,
    "epsilon_norm": 8.751324653625488,
    "prediction_norm": 6.468631267547607,
    "timestamp": "2025-07-14T12:34:31.018006"
  },
  {
    "epoch": 1605,
    "batch_idx": 96,
    "loss": 0.3454638719558716,
    "learning_rate": 9.319577524960654e-05,
    "timestep_mean": 448.84375,
    "epsilon_norm": 8.370095252990723,
    "prediction_norm": 6.908486843109131,
    "timestamp": "2025-07-14T12:34:31.029886"
  },
  {
    "epoch": 1605,
    "batch_idx": 128,
    "loss": 0.5049172639846802,
    "learning_rate": 9.319577524960654e-05,
    "timestep_mean": 587.9375,
    "epsilon_norm": 8.887069702148438,
    "prediction_norm": 6.947171688079834,
    "timestamp": "2025-07-14T12:34:31.034964"
  },
  {
    "epoch": 1605,
    "batch_idx": 160,
    "loss": 0.181848406791687,
    "learning_rate": 9.319577524960654e-05,
    "timestep_mean": 503.15625,
    "epsilon_norm": 6.922755241394043,
    "prediction_norm": 6.320268154144287,
    "timestamp": "2025-07-14T12:34:31.039991"
  },
  {
    "epoch": 1605,
    "batch_idx": 192,
    "loss": 0.3950175642967224,
    "learning_rate": 9.319577524960654e-05,
    "timestep_mean": 383.5,
    "epsilon_norm": 3.9430108070373535,
    "prediction_norm": 2.998563289642334,
    "timestamp": "2025-07-14T12:34:31.045046"
  },
  {
    "epoch": 1606,
    "batch_idx": 0,
    "loss": 0.23907244205474854,
    "learning_rate": 9.273963646474525e-05,
    "timestep_mean": 385.25,
    "epsilon_norm": 7.526427268981934,
    "prediction_norm": 6.268181324005127,
    "timestamp": "2025-07-14T12:34:31.050088"
  },
  {
    "epoch": 1606,
    "batch_idx": 32,
    "loss": 0.523485004901886,
    "learning_rate": 9.273963646474525e-05,
    "timestep_mean": 435.90625,
    "epsilon_norm": 7.857822895050049,
    "prediction_norm": 5.9825758934021,
    "timestamp": "2025-07-14T12:34:31.055082"
  },
  {
    "epoch": 1606,
    "batch_idx": 64,
    "loss": 0.22966814041137695,
    "learning_rate": 9.273963646474525e-05,
    "timestep_mean": 462.125,
    "epsilon_norm": 7.98931884765625,
    "prediction_norm": 7.240030765533447,
    "timestamp": "2025-07-14T12:34:31.060162"
  },
  {
    "epoch": 1606,
    "batch_idx": 96,
    "loss": 0.3360399305820465,
    "learning_rate": 9.273963646474525e-05,
    "timestep_mean": 447.25,
    "epsilon_norm": 9.219365119934082,
    "prediction_norm": 7.57946252822876,
    "timestamp": "2025-07-14T12:34:31.065107"
  },
  {
    "epoch": 1606,
    "batch_idx": 128,
    "loss": 0.2008811980485916,
    "learning_rate": 9.273963646474525e-05,
    "timestep_mean": 593.65625,
    "epsilon_norm": 8.04570198059082,
    "prediction_norm": 7.53640079498291,
    "timestamp": "2025-07-14T12:34:31.070114"
  },
  {
    "epoch": 1606,
    "batch_idx": 160,
    "loss": 0.30368080735206604,
    "learning_rate": 9.273963646474525e-05,
    "timestep_mean": 451.0625,
    "epsilon_norm": 7.68071985244751,
    "prediction_norm": 5.774580001831055,
    "timestamp": "2025-07-14T12:34:31.075053"
  },
  {
    "epoch": 1606,
    "batch_idx": 192,
    "loss": 0.16804978251457214,
    "learning_rate": 9.273963646474525e-05,
    "timestep_mean": 388.125,
    "epsilon_norm": 3.3894972801208496,
    "prediction_norm": 2.662461757659912,
    "timestamp": "2025-07-14T12:34:31.080079"
  },
  {
    "epoch": 1607,
    "batch_idx": 0,
    "loss": 0.14659589529037476,
    "learning_rate": 9.228450255434666e-05,
    "timestep_mean": 561.0625,
    "epsilon_norm": 6.517457962036133,
    "prediction_norm": 5.63501501083374,
    "timestamp": "2025-07-14T12:34:31.085104"
  },
  {
    "epoch": 1607,
    "batch_idx": 32,
    "loss": 0.35552483797073364,
    "learning_rate": 9.228450255434666e-05,
    "timestep_mean": 455.21875,
    "epsilon_norm": 8.25238037109375,
    "prediction_norm": 6.240220069885254,
    "timestamp": "2025-07-14T12:34:31.090183"
  },
  {
    "epoch": 1607,
    "batch_idx": 64,
    "loss": 0.31408625841140747,
    "learning_rate": 9.228450255434666e-05,
    "timestep_mean": 543.5625,
    "epsilon_norm": 8.950237274169922,
    "prediction_norm": 8.147526741027832,
    "timestamp": "2025-07-14T12:34:31.095188"
  },
  {
    "epoch": 1607,
    "batch_idx": 96,
    "loss": 0.14290329813957214,
    "learning_rate": 9.228450255434666e-05,
    "timestep_mean": 503.3125,
    "epsilon_norm": 8.436400413513184,
    "prediction_norm": 7.450334548950195,
    "timestamp": "2025-07-14T12:34:31.100282"
  },
  {
    "epoch": 1607,
    "batch_idx": 128,
    "loss": 0.34556329250335693,
    "learning_rate": 9.228450255434666e-05,
    "timestep_mean": 528.21875,
    "epsilon_norm": 8.080450057983398,
    "prediction_norm": 6.044041156768799,
    "timestamp": "2025-07-14T12:34:31.105396"
  },
  {
    "epoch": 1607,
    "batch_idx": 160,
    "loss": 0.4623643457889557,
    "learning_rate": 9.228450255434666e-05,
    "timestep_mean": 444.28125,
    "epsilon_norm": 8.202569007873535,
    "prediction_norm": 6.784224987030029,
    "timestamp": "2025-07-14T12:34:31.110473"
  },
  {
    "epoch": 1607,
    "batch_idx": 192,
    "loss": 0.19135574996471405,
    "learning_rate": 9.228450255434666e-05,
    "timestep_mean": 605.25,
    "epsilon_norm": 3.389253616333008,
    "prediction_norm": 3.124436855316162,
    "timestamp": "2025-07-14T12:34:31.115479"
  },
  {
    "epoch": 1608,
    "batch_idx": 0,
    "loss": 0.17896884679794312,
    "learning_rate": 9.183037464140802e-05,
    "timestep_mean": 571.71875,
    "epsilon_norm": 7.665429592132568,
    "prediction_norm": 7.153817653656006,
    "timestamp": "2025-07-14T12:34:31.120498"
  },
  {
    "epoch": 1608,
    "batch_idx": 32,
    "loss": 0.282947301864624,
    "learning_rate": 9.183037464140802e-05,
    "timestep_mean": 553.59375,
    "epsilon_norm": 8.13263988494873,
    "prediction_norm": 6.945123672485352,
    "timestamp": "2025-07-14T12:34:31.125763"
  },
  {
    "epoch": 1608,
    "batch_idx": 64,
    "loss": 0.2426479458808899,
    "learning_rate": 9.183037464140802e-05,
    "timestep_mean": 540.90625,
    "epsilon_norm": 7.478871822357178,
    "prediction_norm": 6.5213446617126465,
    "timestamp": "2025-07-14T12:34:31.130766"
  },
  {
    "epoch": 1608,
    "batch_idx": 96,
    "loss": 0.4583255648612976,
    "learning_rate": 9.183037464140802e-05,
    "timestep_mean": 357.125,
    "epsilon_norm": 7.882180213928223,
    "prediction_norm": 5.589532375335693,
    "timestamp": "2025-07-14T12:34:31.135840"
  },
  {
    "epoch": 1608,
    "batch_idx": 128,
    "loss": 0.5951879024505615,
    "learning_rate": 9.183037464140802e-05,
    "timestep_mean": 473.96875,
    "epsilon_norm": 8.443631172180176,
    "prediction_norm": 6.591275215148926,
    "timestamp": "2025-07-14T12:34:31.140886"
  },
  {
    "epoch": 1608,
    "batch_idx": 160,
    "loss": 0.27967116236686707,
    "learning_rate": 9.183037464140802e-05,
    "timestep_mean": 457.1875,
    "epsilon_norm": 7.6972126960754395,
    "prediction_norm": 6.291311264038086,
    "timestamp": "2025-07-14T12:34:31.145956"
  },
  {
    "epoch": 1608,
    "batch_idx": 192,
    "loss": 0.898321807384491,
    "learning_rate": 9.183037464140802e-05,
    "timestep_mean": 285.125,
    "epsilon_norm": 4.438712120056152,
    "prediction_norm": 2.7878992557525635,
    "timestamp": "2025-07-14T12:34:31.150904"
  },
  {
    "epoch": 1609,
    "batch_idx": 0,
    "loss": 0.3597673773765564,
    "learning_rate": 9.137725384644522e-05,
    "timestep_mean": 508.96875,
    "epsilon_norm": 8.31312084197998,
    "prediction_norm": 6.943141937255859,
    "timestamp": "2025-07-14T12:34:31.156088"
  },
  {
    "epoch": 1609,
    "batch_idx": 32,
    "loss": 0.17687223851680756,
    "learning_rate": 9.137725384644522e-05,
    "timestep_mean": 517.59375,
    "epsilon_norm": 6.255077838897705,
    "prediction_norm": 5.266556739807129,
    "timestamp": "2025-07-14T12:34:31.160959"
  },
  {
    "epoch": 1609,
    "batch_idx": 64,
    "loss": 0.2216934859752655,
    "learning_rate": 9.137725384644522e-05,
    "timestep_mean": 448.65625,
    "epsilon_norm": 7.7805681228637695,
    "prediction_norm": 6.493919372558594,
    "timestamp": "2025-07-14T12:34:31.166040"
  },
  {
    "epoch": 1609,
    "batch_idx": 96,
    "loss": 0.19906361401081085,
    "learning_rate": 9.137725384644522e-05,
    "timestep_mean": 583.09375,
    "epsilon_norm": 7.254563331604004,
    "prediction_norm": 6.3124494552612305,
    "timestamp": "2025-07-14T12:34:31.171277"
  },
  {
    "epoch": 1609,
    "batch_idx": 128,
    "loss": 0.3036262094974518,
    "learning_rate": 9.137725384644522e-05,
    "timestep_mean": 466.5,
    "epsilon_norm": 7.137725830078125,
    "prediction_norm": 6.030949115753174,
    "timestamp": "2025-07-14T12:34:31.176478"
  },
  {
    "epoch": 1609,
    "batch_idx": 160,
    "loss": 0.5288398265838623,
    "learning_rate": 9.137725384644522e-05,
    "timestep_mean": 459.46875,
    "epsilon_norm": 8.863893508911133,
    "prediction_norm": 6.112179279327393,
    "timestamp": "2025-07-14T12:34:31.181779"
  },
  {
    "epoch": 1609,
    "batch_idx": 192,
    "loss": 0.26661375164985657,
    "learning_rate": 9.137725384644522e-05,
    "timestep_mean": 535.5,
    "epsilon_norm": 5.288573741912842,
    "prediction_norm": 4.303282260894775,
    "timestamp": "2025-07-14T12:34:31.187008"
  },
  {
    "epoch": 1610,
    "batch_idx": 0,
    "loss": 0.18452395498752594,
    "learning_rate": 9.092514128748834e-05,
    "timestep_mean": 473.96875,
    "epsilon_norm": 6.941155910491943,
    "prediction_norm": 6.218162536621094,
    "timestamp": "2025-07-14T12:34:31.192290"
  },
  {
    "epoch": 1610,
    "batch_idx": 32,
    "loss": 0.2281583696603775,
    "learning_rate": 9.092514128748834e-05,
    "timestep_mean": 442.46875,
    "epsilon_norm": 8.36672306060791,
    "prediction_norm": 7.051294326782227,
    "timestamp": "2025-07-14T12:34:31.197411"
  },
  {
    "epoch": 1610,
    "batch_idx": 64,
    "loss": 0.11640992760658264,
    "learning_rate": 9.092514128748834e-05,
    "timestep_mean": 611.3125,
    "epsilon_norm": 8.849963188171387,
    "prediction_norm": 7.809792518615723,
    "timestamp": "2025-07-14T12:34:31.202486"
  },
  {
    "epoch": 1610,
    "batch_idx": 96,
    "loss": 0.16086623072624207,
    "learning_rate": 9.092514128748834e-05,
    "timestep_mean": 474.75,
    "epsilon_norm": 8.062966346740723,
    "prediction_norm": 6.857853412628174,
    "timestamp": "2025-07-14T12:34:31.207520"
  },
  {
    "epoch": 1610,
    "batch_idx": 128,
    "loss": 0.37256544828414917,
    "learning_rate": 9.092514128748834e-05,
    "timestep_mean": 521.78125,
    "epsilon_norm": 8.163330078125,
    "prediction_norm": 6.157271862030029,
    "timestamp": "2025-07-14T12:34:31.212859"
  },
  {
    "epoch": 1610,
    "batch_idx": 160,
    "loss": 0.3838152289390564,
    "learning_rate": 9.092514128748834e-05,
    "timestep_mean": 513.46875,
    "epsilon_norm": 8.137150764465332,
    "prediction_norm": 6.374107837677002,
    "timestamp": "2025-07-14T12:34:31.218081"
  },
  {
    "epoch": 1610,
    "batch_idx": 192,
    "loss": 0.20692801475524902,
    "learning_rate": 9.092514128748834e-05,
    "timestep_mean": 480.875,
    "epsilon_norm": 3.1979992389678955,
    "prediction_norm": 2.264106273651123,
    "timestamp": "2025-07-14T12:34:31.223250"
  },
  {
    "epoch": 1611,
    "batch_idx": 0,
    "loss": 0.3280313014984131,
    "learning_rate": 9.047403808008035e-05,
    "timestep_mean": 453.375,
    "epsilon_norm": 8.267175674438477,
    "prediction_norm": 6.422457218170166,
    "timestamp": "2025-07-14T12:34:31.228565"
  },
  {
    "epoch": 1611,
    "batch_idx": 32,
    "loss": 0.16652807593345642,
    "learning_rate": 9.047403808008035e-05,
    "timestep_mean": 509.625,
    "epsilon_norm": 7.909407138824463,
    "prediction_norm": 6.837213039398193,
    "timestamp": "2025-07-14T12:34:31.233868"
  },
  {
    "epoch": 1611,
    "batch_idx": 64,
    "loss": 0.25801199674606323,
    "learning_rate": 9.047403808008035e-05,
    "timestep_mean": 577.46875,
    "epsilon_norm": 8.711028099060059,
    "prediction_norm": 7.870569229125977,
    "timestamp": "2025-07-14T12:34:31.239127"
  },
  {
    "epoch": 1611,
    "batch_idx": 96,
    "loss": 0.17917227745056152,
    "learning_rate": 9.047403808008035e-05,
    "timestep_mean": 514.9375,
    "epsilon_norm": 7.737374305725098,
    "prediction_norm": 7.434373378753662,
    "timestamp": "2025-07-14T12:34:31.244388"
  },
  {
    "epoch": 1611,
    "batch_idx": 128,
    "loss": 0.3239642083644867,
    "learning_rate": 9.047403808008035e-05,
    "timestep_mean": 555.0,
    "epsilon_norm": 7.5575432777404785,
    "prediction_norm": 6.288090229034424,
    "timestamp": "2025-07-14T12:34:31.249602"
  },
  {
    "epoch": 1611,
    "batch_idx": 160,
    "loss": 0.2752770781517029,
    "learning_rate": 9.047403808008035e-05,
    "timestep_mean": 451.1875,
    "epsilon_norm": 8.149287223815918,
    "prediction_norm": 7.101871013641357,
    "timestamp": "2025-07-14T12:34:31.255028"
  },
  {
    "epoch": 1611,
    "batch_idx": 192,
    "loss": 0.2393171191215515,
    "learning_rate": 9.047403808008035e-05,
    "timestep_mean": 524.375,
    "epsilon_norm": 3.837555408477783,
    "prediction_norm": 3.237138271331787,
    "timestamp": "2025-07-14T12:34:31.260201"
  },
  {
    "epoch": 1612,
    "batch_idx": 0,
    "loss": 0.2556964159011841,
    "learning_rate": 9.00239453372739e-05,
    "timestep_mean": 590.21875,
    "epsilon_norm": 9.61277961730957,
    "prediction_norm": 8.410711288452148,
    "timestamp": "2025-07-14T12:34:31.265426"
  },
  {
    "epoch": 1612,
    "batch_idx": 32,
    "loss": 0.3095378279685974,
    "learning_rate": 9.00239453372739e-05,
    "timestep_mean": 497.09375,
    "epsilon_norm": 7.919538497924805,
    "prediction_norm": 6.362170696258545,
    "timestamp": "2025-07-14T12:34:31.270700"
  },
  {
    "epoch": 1612,
    "batch_idx": 64,
    "loss": 0.331891804933548,
    "learning_rate": 9.00239453372739e-05,
    "timestep_mean": 536.40625,
    "epsilon_norm": 7.626880645751953,
    "prediction_norm": 5.687938213348389,
    "timestamp": "2025-07-14T12:34:31.275947"
  },
  {
    "epoch": 1612,
    "batch_idx": 96,
    "loss": 0.25181856751441956,
    "learning_rate": 9.00239453372739e-05,
    "timestep_mean": 567.8125,
    "epsilon_norm": 6.612349987030029,
    "prediction_norm": 6.233132839202881,
    "timestamp": "2025-07-14T12:34:31.281227"
  },
  {
    "epoch": 1612,
    "batch_idx": 128,
    "loss": 0.329256534576416,
    "learning_rate": 9.00239453372739e-05,
    "timestep_mean": 459.53125,
    "epsilon_norm": 8.268064498901367,
    "prediction_norm": 7.001733303070068,
    "timestamp": "2025-07-14T12:34:31.286409"
  },
  {
    "epoch": 1612,
    "batch_idx": 160,
    "loss": 0.2601587176322937,
    "learning_rate": 9.00239453372739e-05,
    "timestep_mean": 528.125,
    "epsilon_norm": 8.019680976867676,
    "prediction_norm": 7.238584518432617,
    "timestamp": "2025-07-14T12:34:31.291642"
  },
  {
    "epoch": 1612,
    "batch_idx": 192,
    "loss": 0.4061858355998993,
    "learning_rate": 9.00239453372739e-05,
    "timestep_mean": 494.0,
    "epsilon_norm": 4.525386810302734,
    "prediction_norm": 3.6774165630340576,
    "timestamp": "2025-07-14T12:34:31.296877"
  },
  {
    "epoch": 1613,
    "batch_idx": 0,
    "loss": 0.375497043132782,
    "learning_rate": 8.95748641696276e-05,
    "timestep_mean": 418.84375,
    "epsilon_norm": 8.15070629119873,
    "prediction_norm": 5.539547920227051,
    "timestamp": "2025-07-14T12:34:31.302155"
  },
  {
    "epoch": 1613,
    "batch_idx": 32,
    "loss": 0.2763773202896118,
    "learning_rate": 8.95748641696276e-05,
    "timestep_mean": 565.875,
    "epsilon_norm": 8.577674865722656,
    "prediction_norm": 7.189599990844727,
    "timestamp": "2025-07-14T12:34:31.307384"
  },
  {
    "epoch": 1613,
    "batch_idx": 64,
    "loss": 0.12921613454818726,
    "learning_rate": 8.95748641696276e-05,
    "timestep_mean": 606.46875,
    "epsilon_norm": 7.118432998657227,
    "prediction_norm": 6.944876194000244,
    "timestamp": "2025-07-14T12:34:31.312565"
  },
  {
    "epoch": 1613,
    "batch_idx": 96,
    "loss": 0.16982042789459229,
    "learning_rate": 8.95748641696276e-05,
    "timestep_mean": 531.6875,
    "epsilon_norm": 7.660632133483887,
    "prediction_norm": 7.224610328674316,
    "timestamp": "2025-07-14T12:34:31.317620"
  },
  {
    "epoch": 1613,
    "batch_idx": 128,
    "loss": 0.2792019844055176,
    "learning_rate": 8.95748641696276e-05,
    "timestep_mean": 604.0625,
    "epsilon_norm": 7.690595626831055,
    "prediction_norm": 6.9305419921875,
    "timestamp": "2025-07-14T12:34:31.322721"
  },
  {
    "epoch": 1613,
    "batch_idx": 160,
    "loss": 0.256534218788147,
    "learning_rate": 8.95748641696276e-05,
    "timestep_mean": 529.8125,
    "epsilon_norm": 7.079247951507568,
    "prediction_norm": 6.179296970367432,
    "timestamp": "2025-07-14T12:34:31.328013"
  },
  {
    "epoch": 1613,
    "batch_idx": 192,
    "loss": 0.2931506931781769,
    "learning_rate": 8.95748641696276e-05,
    "timestep_mean": 505.375,
    "epsilon_norm": 3.4028892517089844,
    "prediction_norm": 2.5889158248901367,
    "timestamp": "2025-07-14T12:34:31.333192"
  },
  {
    "epoch": 1614,
    "batch_idx": 0,
    "loss": 0.14166073501110077,
    "learning_rate": 8.912679568520491e-05,
    "timestep_mean": 407.15625,
    "epsilon_norm": 7.7175445556640625,
    "prediction_norm": 6.726978778839111,
    "timestamp": "2025-07-14T12:34:31.338613"
  },
  {
    "epoch": 1614,
    "batch_idx": 32,
    "loss": 0.17523279786109924,
    "learning_rate": 8.912679568520491e-05,
    "timestep_mean": 498.65625,
    "epsilon_norm": 7.791312217712402,
    "prediction_norm": 6.964685440063477,
    "timestamp": "2025-07-14T12:34:31.343772"
  },
  {
    "epoch": 1614,
    "batch_idx": 64,
    "loss": 0.2743483781814575,
    "learning_rate": 8.912679568520491e-05,
    "timestep_mean": 555.90625,
    "epsilon_norm": 9.213183403015137,
    "prediction_norm": 7.944064140319824,
    "timestamp": "2025-07-14T12:34:31.349019"
  },
  {
    "epoch": 1614,
    "batch_idx": 96,
    "loss": 0.32207995653152466,
    "learning_rate": 8.912679568520491e-05,
    "timestep_mean": 445.28125,
    "epsilon_norm": 8.174729347229004,
    "prediction_norm": 7.168702125549316,
    "timestamp": "2025-07-14T12:34:31.354365"
  },
  {
    "epoch": 1614,
    "batch_idx": 128,
    "loss": 0.3428881764411926,
    "learning_rate": 8.912679568520491e-05,
    "timestep_mean": 525.625,
    "epsilon_norm": 8.332999229431152,
    "prediction_norm": 6.767307758331299,
    "timestamp": "2025-07-14T12:34:31.359537"
  },
  {
    "epoch": 1614,
    "batch_idx": 160,
    "loss": 0.5059342384338379,
    "learning_rate": 8.912679568520491e-05,
    "timestep_mean": 475.3125,
    "epsilon_norm": 8.097783088684082,
    "prediction_norm": 6.4761128425598145,
    "timestamp": "2025-07-14T12:34:31.364797"
  },
  {
    "epoch": 1614,
    "batch_idx": 192,
    "loss": 0.5154378414154053,
    "learning_rate": 8.912679568520491e-05,
    "timestep_mean": 426.875,
    "epsilon_norm": 3.4232826232910156,
    "prediction_norm": 2.753451108932495,
    "timestamp": "2025-07-14T12:34:31.370007"
  },
  {
    "epoch": 1615,
    "batch_idx": 0,
    "loss": 0.403175413608551,
    "learning_rate": 8.867974098957014e-05,
    "timestep_mean": 449.96875,
    "epsilon_norm": 8.862948417663574,
    "prediction_norm": 6.62391471862793,
    "timestamp": "2025-07-14T12:34:31.375223"
  },
  {
    "epoch": 1615,
    "batch_idx": 32,
    "loss": 0.1740654855966568,
    "learning_rate": 8.867974098957014e-05,
    "timestep_mean": 570.625,
    "epsilon_norm": 8.103278160095215,
    "prediction_norm": 7.741751670837402,
    "timestamp": "2025-07-14T12:34:31.380428"
  },
  {
    "epoch": 1615,
    "batch_idx": 64,
    "loss": 0.2863706648349762,
    "learning_rate": 8.867974098957014e-05,
    "timestep_mean": 505.625,
    "epsilon_norm": 7.775548458099365,
    "prediction_norm": 6.636582374572754,
    "timestamp": "2025-07-14T12:34:31.385634"
  },
  {
    "epoch": 1615,
    "batch_idx": 96,
    "loss": 0.35784244537353516,
    "learning_rate": 8.867974098957014e-05,
    "timestep_mean": 553.09375,
    "epsilon_norm": 7.875508785247803,
    "prediction_norm": 6.012172222137451,
    "timestamp": "2025-07-14T12:34:31.390866"
  },
  {
    "epoch": 1615,
    "batch_idx": 128,
    "loss": 0.18542833626270294,
    "learning_rate": 8.867974098957014e-05,
    "timestep_mean": 522.1875,
    "epsilon_norm": 8.226766586303711,
    "prediction_norm": 7.331573009490967,
    "timestamp": "2025-07-14T12:34:31.396183"
  },
  {
    "epoch": 1615,
    "batch_idx": 160,
    "loss": 0.28184714913368225,
    "learning_rate": 8.867974098957014e-05,
    "timestep_mean": 508.34375,
    "epsilon_norm": 7.326406002044678,
    "prediction_norm": 5.3959736824035645,
    "timestamp": "2025-07-14T12:34:31.401405"
  },
  {
    "epoch": 1615,
    "batch_idx": 192,
    "loss": 0.08928389102220535,
    "learning_rate": 8.867974098957014e-05,
    "timestep_mean": 608.5,
    "epsilon_norm": 2.761300802230835,
    "prediction_norm": 2.728727340698242,
    "timestamp": "2025-07-14T12:34:31.406340"
  },
  {
    "epoch": 1616,
    "batch_idx": 0,
    "loss": 0.33890098333358765,
    "learning_rate": 8.823370118578625e-05,
    "timestep_mean": 532.96875,
    "epsilon_norm": 8.846465110778809,
    "prediction_norm": 7.546883583068848,
    "timestamp": "2025-07-14T12:34:31.411221"
  },
  {
    "epoch": 1616,
    "batch_idx": 32,
    "loss": 0.3165607750415802,
    "learning_rate": 8.823370118578625e-05,
    "timestep_mean": 452.75,
    "epsilon_norm": 8.776612281799316,
    "prediction_norm": 7.7894134521484375,
    "timestamp": "2025-07-14T12:34:31.416437"
  },
  {
    "epoch": 1616,
    "batch_idx": 64,
    "loss": 0.2608875632286072,
    "learning_rate": 8.823370118578625e-05,
    "timestep_mean": 533.40625,
    "epsilon_norm": 7.658405303955078,
    "prediction_norm": 6.6294145584106445,
    "timestamp": "2025-07-14T12:34:31.421969"
  },
  {
    "epoch": 1616,
    "batch_idx": 96,
    "loss": 0.23077601194381714,
    "learning_rate": 8.823370118578625e-05,
    "timestep_mean": 476.34375,
    "epsilon_norm": 7.64253568649292,
    "prediction_norm": 7.005607604980469,
    "timestamp": "2025-07-14T12:34:31.427158"
  },
  {
    "epoch": 1616,
    "batch_idx": 128,
    "loss": 0.242820143699646,
    "learning_rate": 8.823370118578625e-05,
    "timestep_mean": 480.75,
    "epsilon_norm": 7.932525634765625,
    "prediction_norm": 7.128940582275391,
    "timestamp": "2025-07-14T12:34:31.432028"
  },
  {
    "epoch": 1616,
    "batch_idx": 160,
    "loss": 0.25601476430892944,
    "learning_rate": 8.823370118578625e-05,
    "timestep_mean": 489.46875,
    "epsilon_norm": 7.919058322906494,
    "prediction_norm": 7.222007751464844,
    "timestamp": "2025-07-14T12:34:31.437008"
  },
  {
    "epoch": 1616,
    "batch_idx": 192,
    "loss": 0.5339506268501282,
    "learning_rate": 8.823370118578625e-05,
    "timestep_mean": 533.875,
    "epsilon_norm": 3.8987438678741455,
    "prediction_norm": 2.4087793827056885,
    "timestamp": "2025-07-14T12:34:31.442180"
  },
  {
    "epoch": 1617,
    "batch_idx": 0,
    "loss": 0.33222082257270813,
    "learning_rate": 8.778867737441229e-05,
    "timestep_mean": 446.03125,
    "epsilon_norm": 7.620940208435059,
    "prediction_norm": 6.296175479888916,
    "timestamp": "2025-07-14T12:34:31.447235"
  },
  {
    "epoch": 1617,
    "batch_idx": 32,
    "loss": 0.1657385528087616,
    "learning_rate": 8.778867737441229e-05,
    "timestep_mean": 486.96875,
    "epsilon_norm": 8.362639427185059,
    "prediction_norm": 8.30892276763916,
    "timestamp": "2025-07-14T12:34:31.452177"
  },
  {
    "epoch": 1617,
    "batch_idx": 64,
    "loss": 0.2733939290046692,
    "learning_rate": 8.778867737441229e-05,
    "timestep_mean": 472.4375,
    "epsilon_norm": 7.8459320068359375,
    "prediction_norm": 6.149105548858643,
    "timestamp": "2025-07-14T12:34:31.457076"
  },
  {
    "epoch": 1617,
    "batch_idx": 96,
    "loss": 0.33038145303726196,
    "learning_rate": 8.778867737441229e-05,
    "timestep_mean": 471.1875,
    "epsilon_norm": 9.241466522216797,
    "prediction_norm": 7.650134563446045,
    "timestamp": "2025-07-14T12:34:31.461966"
  },
  {
    "epoch": 1617,
    "batch_idx": 128,
    "loss": 0.3668597340583801,
    "learning_rate": 8.778867737441229e-05,
    "timestep_mean": 574.5,
    "epsilon_norm": 7.626540660858154,
    "prediction_norm": 6.130906581878662,
    "timestamp": "2025-07-14T12:34:31.467453"
  },
  {
    "epoch": 1617,
    "batch_idx": 160,
    "loss": 0.34816810488700867,
    "learning_rate": 8.778867737441229e-05,
    "timestep_mean": 518.65625,
    "epsilon_norm": 7.58648157119751,
    "prediction_norm": 6.455069541931152,
    "timestamp": "2025-07-14T12:34:31.472733"
  },
  {
    "epoch": 1617,
    "batch_idx": 192,
    "loss": 0.7874974012374878,
    "learning_rate": 8.778867737441229e-05,
    "timestep_mean": 376.75,
    "epsilon_norm": 5.281050205230713,
    "prediction_norm": 3.081319570541382,
    "timestamp": "2025-07-14T12:34:31.477734"
  },
  {
    "epoch": 1618,
    "batch_idx": 0,
    "loss": 0.25970834493637085,
    "learning_rate": 8.734467065350018e-05,
    "timestep_mean": 496.8125,
    "epsilon_norm": 8.051762580871582,
    "prediction_norm": 6.618923187255859,
    "timestamp": "2025-07-14T12:34:31.482845"
  },
  {
    "epoch": 1618,
    "batch_idx": 32,
    "loss": 0.25374048948287964,
    "learning_rate": 8.734467065350018e-05,
    "timestep_mean": 508.0625,
    "epsilon_norm": 7.498690605163574,
    "prediction_norm": 6.448763847351074,
    "timestamp": "2025-07-14T12:34:31.487830"
  },
  {
    "epoch": 1618,
    "batch_idx": 64,
    "loss": 0.1883557140827179,
    "learning_rate": 8.734467065350018e-05,
    "timestep_mean": 527.625,
    "epsilon_norm": 7.472657203674316,
    "prediction_norm": 6.5970234870910645,
    "timestamp": "2025-07-14T12:34:31.492880"
  },
  {
    "epoch": 1618,
    "batch_idx": 96,
    "loss": 0.4552736282348633,
    "learning_rate": 8.734467065350018e-05,
    "timestep_mean": 465.96875,
    "epsilon_norm": 9.189583778381348,
    "prediction_norm": 7.324896812438965,
    "timestamp": "2025-07-14T12:34:31.497810"
  },
  {
    "epoch": 1618,
    "batch_idx": 128,
    "loss": 0.19545188546180725,
    "learning_rate": 8.734467065350018e-05,
    "timestep_mean": 516.5,
    "epsilon_norm": 8.048665046691895,
    "prediction_norm": 7.149972915649414,
    "timestamp": "2025-07-14T12:34:31.502696"
  },
  {
    "epoch": 1618,
    "batch_idx": 160,
    "loss": 0.2157018482685089,
    "learning_rate": 8.734467065350018e-05,
    "timestep_mean": 466.84375,
    "epsilon_norm": 8.312984466552734,
    "prediction_norm": 7.278542995452881,
    "timestamp": "2025-07-14T12:34:31.507558"
  },
  {
    "epoch": 1618,
    "batch_idx": 192,
    "loss": 0.2585991621017456,
    "learning_rate": 8.734467065350018e-05,
    "timestep_mean": 520.25,
    "epsilon_norm": 3.3805441856384277,
    "prediction_norm": 3.0698368549346924,
    "timestamp": "2025-07-14T12:34:31.512592"
  },
  {
    "epoch": 1619,
    "batch_idx": 0,
    "loss": 0.20834505558013916,
    "learning_rate": 8.690168211859242e-05,
    "timestep_mean": 496.9375,
    "epsilon_norm": 8.277740478515625,
    "prediction_norm": 7.739686012268066,
    "timestamp": "2025-07-14T12:34:31.517488"
  },
  {
    "epoch": 1619,
    "batch_idx": 32,
    "loss": 0.26842302083969116,
    "learning_rate": 8.690168211859242e-05,
    "timestep_mean": 537.90625,
    "epsilon_norm": 8.370823860168457,
    "prediction_norm": 6.585498809814453,
    "timestamp": "2025-07-14T12:34:31.522266"
  },
  {
    "epoch": 1619,
    "batch_idx": 64,
    "loss": 0.4472881853580475,
    "learning_rate": 8.690168211859242e-05,
    "timestep_mean": 490.4375,
    "epsilon_norm": 8.699114799499512,
    "prediction_norm": 7.338632583618164,
    "timestamp": "2025-07-14T12:34:31.527155"
  },
  {
    "epoch": 1619,
    "batch_idx": 96,
    "loss": 0.33580687642097473,
    "learning_rate": 8.690168211859242e-05,
    "timestep_mean": 483.5625,
    "epsilon_norm": 8.387104034423828,
    "prediction_norm": 7.293158054351807,
    "timestamp": "2025-07-14T12:34:31.532073"
  },
  {
    "epoch": 1619,
    "batch_idx": 128,
    "loss": 0.3122672438621521,
    "learning_rate": 8.690168211859242e-05,
    "timestep_mean": 478.40625,
    "epsilon_norm": 8.185745239257812,
    "prediction_norm": 7.060456275939941,
    "timestamp": "2025-07-14T12:34:31.536967"
  },
  {
    "epoch": 1619,
    "batch_idx": 160,
    "loss": 0.4825582802295685,
    "learning_rate": 8.690168211859242e-05,
    "timestep_mean": 432.96875,
    "epsilon_norm": 9.429977416992188,
    "prediction_norm": 7.4286041259765625,
    "timestamp": "2025-07-14T12:34:31.541894"
  },
  {
    "epoch": 1619,
    "batch_idx": 192,
    "loss": 0.715203583240509,
    "learning_rate": 8.690168211859242e-05,
    "timestep_mean": 590.75,
    "epsilon_norm": 4.138432025909424,
    "prediction_norm": 2.4448421001434326,
    "timestamp": "2025-07-14T12:34:31.546832"
  },
  {
    "epoch": 1620,
    "batch_idx": 0,
    "loss": 0.31599509716033936,
    "learning_rate": 8.645971286271911e-05,
    "timestep_mean": 446.6875,
    "epsilon_norm": 7.602613925933838,
    "prediction_norm": 7.0372700691223145,
    "timestamp": "2025-07-14T12:34:31.552046"
  },
  {
    "epoch": 1620,
    "batch_idx": 32,
    "loss": 0.4372699558734894,
    "learning_rate": 8.645971286271911e-05,
    "timestep_mean": 495.15625,
    "epsilon_norm": 8.499323844909668,
    "prediction_norm": 6.377724647521973,
    "timestamp": "2025-07-14T12:34:31.556968"
  },
  {
    "epoch": 1620,
    "batch_idx": 64,
    "loss": 0.23430012166500092,
    "learning_rate": 8.645971286271911e-05,
    "timestep_mean": 474.09375,
    "epsilon_norm": 6.593066215515137,
    "prediction_norm": 5.520786762237549,
    "timestamp": "2025-07-14T12:34:31.561803"
  },
  {
    "epoch": 1620,
    "batch_idx": 96,
    "loss": 0.18577904999256134,
    "learning_rate": 8.645971286271911e-05,
    "timestep_mean": 575.03125,
    "epsilon_norm": 7.581329822540283,
    "prediction_norm": 6.581060409545898,
    "timestamp": "2025-07-14T12:34:31.566670"
  },
  {
    "epoch": 1620,
    "batch_idx": 128,
    "loss": 0.1323649287223816,
    "learning_rate": 8.645971286271911e-05,
    "timestep_mean": 610.375,
    "epsilon_norm": 7.037990570068359,
    "prediction_norm": 6.54928731918335,
    "timestamp": "2025-07-14T12:34:31.571541"
  },
  {
    "epoch": 1620,
    "batch_idx": 160,
    "loss": 0.24153290688991547,
    "learning_rate": 8.645971286271911e-05,
    "timestep_mean": 509.15625,
    "epsilon_norm": 7.92554235458374,
    "prediction_norm": 6.004128932952881,
    "timestamp": "2025-07-14T12:34:31.576663"
  },
  {
    "epoch": 1620,
    "batch_idx": 192,
    "loss": 0.14489293098449707,
    "learning_rate": 8.645971286271911e-05,
    "timestep_mean": 532.5,
    "epsilon_norm": 2.323636770248413,
    "prediction_norm": 1.8158806562423706,
    "timestamp": "2025-07-14T12:34:31.581486"
  },
  {
    "epoch": 1621,
    "batch_idx": 0,
    "loss": 0.28696101903915405,
    "learning_rate": 8.601876397639555e-05,
    "timestep_mean": 486.59375,
    "epsilon_norm": 8.071396827697754,
    "prediction_norm": 7.042867660522461,
    "timestamp": "2025-07-14T12:34:31.586355"
  },
  {
    "epoch": 1621,
    "batch_idx": 32,
    "loss": 0.25550299882888794,
    "learning_rate": 8.601876397639555e-05,
    "timestep_mean": 479.3125,
    "epsilon_norm": 7.621682643890381,
    "prediction_norm": 6.6440110206604,
    "timestamp": "2025-07-14T12:34:31.591223"
  },
  {
    "epoch": 1621,
    "batch_idx": 64,
    "loss": 0.47779321670532227,
    "learning_rate": 8.601876397639555e-05,
    "timestep_mean": 411.75,
    "epsilon_norm": 8.342216491699219,
    "prediction_norm": 6.871536731719971,
    "timestamp": "2025-07-14T12:34:31.596189"
  },
  {
    "epoch": 1621,
    "batch_idx": 96,
    "loss": 0.1928570568561554,
    "learning_rate": 8.601876397639555e-05,
    "timestep_mean": 489.875,
    "epsilon_norm": 8.063713073730469,
    "prediction_norm": 6.709600448608398,
    "timestamp": "2025-07-14T12:34:31.601120"
  },
  {
    "epoch": 1621,
    "batch_idx": 128,
    "loss": 0.07927808165550232,
    "learning_rate": 8.601876397639555e-05,
    "timestep_mean": 590.875,
    "epsilon_norm": 6.7631940841674805,
    "prediction_norm": 6.632936954498291,
    "timestamp": "2025-07-14T12:34:31.607049"
  },
  {
    "epoch": 1621,
    "batch_idx": 160,
    "loss": 0.25551939010620117,
    "learning_rate": 8.601876397639555e-05,
    "timestep_mean": 501.40625,
    "epsilon_norm": 8.002384185791016,
    "prediction_norm": 6.7860870361328125,
    "timestamp": "2025-07-14T12:34:31.612713"
  },
  {
    "epoch": 1621,
    "batch_idx": 192,
    "loss": 0.17850635945796967,
    "learning_rate": 8.601876397639555e-05,
    "timestep_mean": 532.875,
    "epsilon_norm": 3.8972418308258057,
    "prediction_norm": 3.7395780086517334,
    "timestamp": "2025-07-14T12:34:31.617862"
  },
  {
    "epoch": 1622,
    "batch_idx": 0,
    "loss": 0.3411296308040619,
    "learning_rate": 8.557883654761903e-05,
    "timestep_mean": 518.6875,
    "epsilon_norm": 7.042698383331299,
    "prediction_norm": 5.305986404418945,
    "timestamp": "2025-07-14T12:34:31.622976"
  },
  {
    "epoch": 1622,
    "batch_idx": 32,
    "loss": 0.42992663383483887,
    "learning_rate": 8.557883654761903e-05,
    "timestep_mean": 453.90625,
    "epsilon_norm": 8.957279205322266,
    "prediction_norm": 7.274131774902344,
    "timestamp": "2025-07-14T12:34:31.627931"
  },
  {
    "epoch": 1622,
    "batch_idx": 64,
    "loss": 0.28831061720848083,
    "learning_rate": 8.557883654761903e-05,
    "timestep_mean": 515.375,
    "epsilon_norm": 7.973896026611328,
    "prediction_norm": 6.757901668548584,
    "timestamp": "2025-07-14T12:34:31.632966"
  },
  {
    "epoch": 1622,
    "batch_idx": 96,
    "loss": 0.4203856289386749,
    "learning_rate": 8.557883654761903e-05,
    "timestep_mean": 405.1875,
    "epsilon_norm": 8.19559097290039,
    "prediction_norm": 6.33122444152832,
    "timestamp": "2025-07-14T12:34:31.638067"
  },
  {
    "epoch": 1622,
    "batch_idx": 128,
    "loss": 0.3019123673439026,
    "learning_rate": 8.557883654761903e-05,
    "timestep_mean": 442.59375,
    "epsilon_norm": 8.800774574279785,
    "prediction_norm": 7.406667232513428,
    "timestamp": "2025-07-14T12:34:31.642934"
  },
  {
    "epoch": 1622,
    "batch_idx": 160,
    "loss": 0.36208558082580566,
    "learning_rate": 8.557883654761903e-05,
    "timestep_mean": 566.28125,
    "epsilon_norm": 9.239391326904297,
    "prediction_norm": 7.715817928314209,
    "timestamp": "2025-07-14T12:34:31.647884"
  },
  {
    "epoch": 1622,
    "batch_idx": 192,
    "loss": 0.3747226595878601,
    "learning_rate": 8.557883654761903e-05,
    "timestep_mean": 477.0,
    "epsilon_norm": 4.448892593383789,
    "prediction_norm": 3.0569660663604736,
    "timestamp": "2025-07-14T12:34:31.653105"
  },
  {
    "epoch": 1623,
    "batch_idx": 0,
    "loss": 0.22881969809532166,
    "learning_rate": 8.513993166186709e-05,
    "timestep_mean": 532.40625,
    "epsilon_norm": 6.487072467803955,
    "prediction_norm": 5.384054660797119,
    "timestamp": "2025-07-14T12:34:31.658306"
  },
  {
    "epoch": 1623,
    "batch_idx": 32,
    "loss": 0.3609938621520996,
    "learning_rate": 8.513993166186709e-05,
    "timestep_mean": 413.09375,
    "epsilon_norm": 7.238444805145264,
    "prediction_norm": 5.348824977874756,
    "timestamp": "2025-07-14T12:34:31.663571"
  },
  {
    "epoch": 1623,
    "batch_idx": 64,
    "loss": 0.3511277437210083,
    "learning_rate": 8.513993166186709e-05,
    "timestep_mean": 493.84375,
    "epsilon_norm": 8.463644027709961,
    "prediction_norm": 6.825091361999512,
    "timestamp": "2025-07-14T12:34:31.668601"
  },
  {
    "epoch": 1623,
    "batch_idx": 96,
    "loss": 0.19954589009284973,
    "learning_rate": 8.513993166186709e-05,
    "timestep_mean": 528.28125,
    "epsilon_norm": 7.831445217132568,
    "prediction_norm": 7.025717258453369,
    "timestamp": "2025-07-14T12:34:31.673716"
  },
  {
    "epoch": 1623,
    "batch_idx": 128,
    "loss": 0.33444201946258545,
    "learning_rate": 8.513993166186709e-05,
    "timestep_mean": 484.125,
    "epsilon_norm": 8.397576332092285,
    "prediction_norm": 6.630729675292969,
    "timestamp": "2025-07-14T12:34:31.678828"
  },
  {
    "epoch": 1623,
    "batch_idx": 160,
    "loss": 0.23838943243026733,
    "learning_rate": 8.513993166186709e-05,
    "timestep_mean": 420.21875,
    "epsilon_norm": 7.784902572631836,
    "prediction_norm": 6.473596572875977,
    "timestamp": "2025-07-14T12:34:31.684212"
  },
  {
    "epoch": 1623,
    "batch_idx": 192,
    "loss": 0.21891295909881592,
    "learning_rate": 8.513993166186709e-05,
    "timestep_mean": 416.375,
    "epsilon_norm": 2.8146564960479736,
    "prediction_norm": 2.072148084640503,
    "timestamp": "2025-07-14T12:34:31.689493"
  },
  {
    "epoch": 1624,
    "batch_idx": 0,
    "loss": 0.27916866540908813,
    "learning_rate": 8.470205040209369e-05,
    "timestep_mean": 487.625,
    "epsilon_norm": 8.0944242477417,
    "prediction_norm": 6.54170036315918,
    "timestamp": "2025-07-14T12:34:31.694765"
  },
  {
    "epoch": 1624,
    "batch_idx": 32,
    "loss": 0.2644020915031433,
    "learning_rate": 8.470205040209369e-05,
    "timestep_mean": 464.90625,
    "epsilon_norm": 6.5502028465271,
    "prediction_norm": 5.4697465896606445,
    "timestamp": "2025-07-14T12:34:31.700002"
  },
  {
    "epoch": 1624,
    "batch_idx": 64,
    "loss": 0.1626281887292862,
    "learning_rate": 8.470205040209369e-05,
    "timestep_mean": 520.1875,
    "epsilon_norm": 7.727405548095703,
    "prediction_norm": 6.81113338470459,
    "timestamp": "2025-07-14T12:34:31.705260"
  },
  {
    "epoch": 1624,
    "batch_idx": 96,
    "loss": 0.3180009126663208,
    "learning_rate": 8.470205040209369e-05,
    "timestep_mean": 450.9375,
    "epsilon_norm": 8.947221755981445,
    "prediction_norm": 7.541695594787598,
    "timestamp": "2025-07-14T12:34:31.711581"
  },
  {
    "epoch": 1624,
    "batch_idx": 128,
    "loss": 0.2068743109703064,
    "learning_rate": 8.470205040209369e-05,
    "timestep_mean": 487.6875,
    "epsilon_norm": 7.877452850341797,
    "prediction_norm": 6.9834723472595215,
    "timestamp": "2025-07-14T12:34:31.717139"
  },
  {
    "epoch": 1624,
    "batch_idx": 160,
    "loss": 0.38370993733406067,
    "learning_rate": 8.470205040209369e-05,
    "timestep_mean": 469.28125,
    "epsilon_norm": 8.071927070617676,
    "prediction_norm": 6.40049409866333,
    "timestamp": "2025-07-14T12:34:31.722439"
  },
  {
    "epoch": 1624,
    "batch_idx": 192,
    "loss": 0.3028327524662018,
    "learning_rate": 8.470205040209369e-05,
    "timestep_mean": 548.125,
    "epsilon_norm": 3.466944932937622,
    "prediction_norm": 2.6640191078186035,
    "timestamp": "2025-07-14T12:34:31.727807"
  },
  {
    "epoch": 1625,
    "batch_idx": 0,
    "loss": 0.2447306215763092,
    "learning_rate": 8.426519384872729e-05,
    "timestep_mean": 482.96875,
    "epsilon_norm": 8.117457389831543,
    "prediction_norm": 6.7528181076049805,
    "timestamp": "2025-07-14T12:34:31.733120"
  },
  {
    "epoch": 1625,
    "batch_idx": 32,
    "loss": 0.29649871587753296,
    "learning_rate": 8.426519384872729e-05,
    "timestep_mean": 448.78125,
    "epsilon_norm": 8.348957061767578,
    "prediction_norm": 6.800312519073486,
    "timestamp": "2025-07-14T12:34:31.738604"
  },
  {
    "epoch": 1625,
    "batch_idx": 64,
    "loss": 0.41468337178230286,
    "learning_rate": 8.426519384872729e-05,
    "timestep_mean": 541.0,
    "epsilon_norm": 7.992591857910156,
    "prediction_norm": 6.028945446014404,
    "timestamp": "2025-07-14T12:34:31.743822"
  },
  {
    "epoch": 1625,
    "batch_idx": 96,
    "loss": 0.3803357481956482,
    "learning_rate": 8.426519384872729e-05,
    "timestep_mean": 445.09375,
    "epsilon_norm": 8.45483112335205,
    "prediction_norm": 7.336349010467529,
    "timestamp": "2025-07-14T12:34:31.749075"
  },
  {
    "epoch": 1625,
    "batch_idx": 128,
    "loss": 0.36072230339050293,
    "learning_rate": 8.426519384872729e-05,
    "timestep_mean": 376.8125,
    "epsilon_norm": 8.659619331359863,
    "prediction_norm": 7.38897705078125,
    "timestamp": "2025-07-14T12:34:31.754334"
  },
  {
    "epoch": 1625,
    "batch_idx": 160,
    "loss": 0.22092553973197937,
    "learning_rate": 8.426519384872729e-05,
    "timestep_mean": 481.5625,
    "epsilon_norm": 8.17082405090332,
    "prediction_norm": 7.129693508148193,
    "timestamp": "2025-07-14T12:34:31.759617"
  },
  {
    "epoch": 1625,
    "batch_idx": 192,
    "loss": 0.2685616910457611,
    "learning_rate": 8.426519384872729e-05,
    "timestep_mean": 465.375,
    "epsilon_norm": 4.284919261932373,
    "prediction_norm": 3.3449060916900635,
    "timestamp": "2025-07-14T12:34:31.764639"
  },
  {
    "epoch": 1626,
    "batch_idx": 0,
    "loss": 0.1273849755525589,
    "learning_rate": 8.38293630796684e-05,
    "timestep_mean": 605.5,
    "epsilon_norm": 7.907330513000488,
    "prediction_norm": 7.103289604187012,
    "timestamp": "2025-07-14T12:34:31.769647"
  },
  {
    "epoch": 1626,
    "batch_idx": 32,
    "loss": 0.31157997250556946,
    "learning_rate": 8.38293630796684e-05,
    "timestep_mean": 437.6875,
    "epsilon_norm": 7.900885581970215,
    "prediction_norm": 6.480996131896973,
    "timestamp": "2025-07-14T12:34:31.774507"
  },
  {
    "epoch": 1626,
    "batch_idx": 64,
    "loss": 0.21166864037513733,
    "learning_rate": 8.38293630796684e-05,
    "timestep_mean": 507.40625,
    "epsilon_norm": 7.234981060028076,
    "prediction_norm": 7.118942737579346,
    "timestamp": "2025-07-14T12:34:31.779462"
  },
  {
    "epoch": 1626,
    "batch_idx": 96,
    "loss": 0.15126517415046692,
    "learning_rate": 8.38293630796684e-05,
    "timestep_mean": 562.84375,
    "epsilon_norm": 8.472898483276367,
    "prediction_norm": 7.822205543518066,
    "timestamp": "2025-07-14T12:34:31.784300"
  },
  {
    "epoch": 1626,
    "batch_idx": 128,
    "loss": 0.18617703020572662,
    "learning_rate": 8.38293630796684e-05,
    "timestep_mean": 501.125,
    "epsilon_norm": 6.833458423614502,
    "prediction_norm": 6.145918846130371,
    "timestamp": "2025-07-14T12:34:31.789257"
  },
  {
    "epoch": 1626,
    "batch_idx": 160,
    "loss": 0.19541656970977783,
    "learning_rate": 8.38293630796684e-05,
    "timestep_mean": 437.5625,
    "epsilon_norm": 6.422487258911133,
    "prediction_norm": 5.981387615203857,
    "timestamp": "2025-07-14T12:34:31.794152"
  },
  {
    "epoch": 1626,
    "batch_idx": 192,
    "loss": 0.14453479647636414,
    "learning_rate": 8.38293630796684e-05,
    "timestep_mean": 428.875,
    "epsilon_norm": 3.239351511001587,
    "prediction_norm": 2.9356582164764404,
    "timestamp": "2025-07-14T12:34:31.799015"
  },
  {
    "epoch": 1627,
    "batch_idx": 0,
    "loss": 0.5746554732322693,
    "learning_rate": 8.339455917028545e-05,
    "timestep_mean": 530.34375,
    "epsilon_norm": 8.960559844970703,
    "prediction_norm": 6.414918899536133,
    "timestamp": "2025-07-14T12:34:31.803881"
  },
  {
    "epoch": 1627,
    "batch_idx": 32,
    "loss": 0.1984291970729828,
    "learning_rate": 8.339455917028545e-05,
    "timestep_mean": 506.375,
    "epsilon_norm": 6.680975914001465,
    "prediction_norm": 6.0136637687683105,
    "timestamp": "2025-07-14T12:34:31.808830"
  },
  {
    "epoch": 1627,
    "batch_idx": 64,
    "loss": 0.1733415722846985,
    "learning_rate": 8.339455917028545e-05,
    "timestep_mean": 441.6875,
    "epsilon_norm": 8.560630798339844,
    "prediction_norm": 7.673375129699707,
    "timestamp": "2025-07-14T12:34:31.814072"
  },
  {
    "epoch": 1627,
    "batch_idx": 96,
    "loss": 0.23356381058692932,
    "learning_rate": 8.339455917028545e-05,
    "timestep_mean": 563.21875,
    "epsilon_norm": 8.418618202209473,
    "prediction_norm": 7.281980514526367,
    "timestamp": "2025-07-14T12:34:31.818980"
  },
  {
    "epoch": 1627,
    "batch_idx": 128,
    "loss": 0.14959466457366943,
    "learning_rate": 8.339455917028545e-05,
    "timestep_mean": 559.375,
    "epsilon_norm": 8.629133224487305,
    "prediction_norm": 7.7465105056762695,
    "timestamp": "2025-07-14T12:34:31.823882"
  },
  {
    "epoch": 1627,
    "batch_idx": 160,
    "loss": 0.23347006738185883,
    "learning_rate": 8.339455917028545e-05,
    "timestep_mean": 500.875,
    "epsilon_norm": 7.096465587615967,
    "prediction_norm": 6.41141414642334,
    "timestamp": "2025-07-14T12:34:31.828762"
  },
  {
    "epoch": 1627,
    "batch_idx": 192,
    "loss": 0.23765486478805542,
    "learning_rate": 8.339455917028545e-05,
    "timestep_mean": 298.75,
    "epsilon_norm": 3.2020156383514404,
    "prediction_norm": 2.227935552597046,
    "timestamp": "2025-07-14T12:34:31.833877"
  },
  {
    "epoch": 1628,
    "batch_idx": 0,
    "loss": 0.18835769593715668,
    "learning_rate": 8.296078319341429e-05,
    "timestep_mean": 542.78125,
    "epsilon_norm": 8.823775291442871,
    "prediction_norm": 8.002196311950684,
    "timestamp": "2025-07-14T12:34:31.838740"
  },
  {
    "epoch": 1628,
    "batch_idx": 32,
    "loss": 0.2128324955701828,
    "learning_rate": 8.296078319341429e-05,
    "timestep_mean": 515.3125,
    "epsilon_norm": 8.772116661071777,
    "prediction_norm": 7.776388645172119,
    "timestamp": "2025-07-14T12:34:31.843726"
  },
  {
    "epoch": 1628,
    "batch_idx": 64,
    "loss": 0.43313586711883545,
    "learning_rate": 8.296078319341429e-05,
    "timestep_mean": 452.5625,
    "epsilon_norm": 8.49423599243164,
    "prediction_norm": 6.7578582763671875,
    "timestamp": "2025-07-14T12:34:31.848694"
  },
  {
    "epoch": 1628,
    "batch_idx": 96,
    "loss": 0.21112476289272308,
    "learning_rate": 8.296078319341429e-05,
    "timestep_mean": 578.5625,
    "epsilon_norm": 8.441749572753906,
    "prediction_norm": 7.665709972381592,
    "timestamp": "2025-07-14T12:34:31.853610"
  },
  {
    "epoch": 1628,
    "batch_idx": 128,
    "loss": 0.13464555144309998,
    "learning_rate": 8.296078319341429e-05,
    "timestep_mean": 549.875,
    "epsilon_norm": 6.972061634063721,
    "prediction_norm": 6.480435371398926,
    "timestamp": "2025-07-14T12:34:31.858917"
  },
  {
    "epoch": 1628,
    "batch_idx": 160,
    "loss": 0.1573936939239502,
    "learning_rate": 8.296078319341429e-05,
    "timestep_mean": 468.15625,
    "epsilon_norm": 7.535615921020508,
    "prediction_norm": 6.538793087005615,
    "timestamp": "2025-07-14T12:34:31.864323"
  },
  {
    "epoch": 1628,
    "batch_idx": 192,
    "loss": 0.5380675196647644,
    "learning_rate": 8.296078319341429e-05,
    "timestep_mean": 360.5,
    "epsilon_norm": 4.006632328033447,
    "prediction_norm": 2.2323951721191406,
    "timestamp": "2025-07-14T12:34:31.869474"
  },
  {
    "epoch": 1629,
    "batch_idx": 0,
    "loss": 0.10157997906208038,
    "learning_rate": 8.25280362193542e-05,
    "timestep_mean": 641.6875,
    "epsilon_norm": 7.153375148773193,
    "prediction_norm": 6.673933029174805,
    "timestamp": "2025-07-14T12:34:31.874773"
  },
  {
    "epoch": 1629,
    "batch_idx": 32,
    "loss": 0.3019685745239258,
    "learning_rate": 8.25280362193542e-05,
    "timestep_mean": 497.4375,
    "epsilon_norm": 7.500834941864014,
    "prediction_norm": 6.226099014282227,
    "timestamp": "2025-07-14T12:34:31.879893"
  },
  {
    "epoch": 1629,
    "batch_idx": 64,
    "loss": 0.22217847406864166,
    "learning_rate": 8.25280362193542e-05,
    "timestep_mean": 480.125,
    "epsilon_norm": 8.116772651672363,
    "prediction_norm": 6.903735637664795,
    "timestamp": "2025-07-14T12:34:31.884987"
  },
  {
    "epoch": 1629,
    "batch_idx": 96,
    "loss": 0.13015572726726532,
    "learning_rate": 8.25280362193542e-05,
    "timestep_mean": 484.90625,
    "epsilon_norm": 7.1389336585998535,
    "prediction_norm": 6.2391743659973145,
    "timestamp": "2025-07-14T12:34:31.890064"
  },
  {
    "epoch": 1629,
    "batch_idx": 128,
    "loss": 0.29557275772094727,
    "learning_rate": 8.25280362193542e-05,
    "timestep_mean": 402.625,
    "epsilon_norm": 8.124063491821289,
    "prediction_norm": 6.092148303985596,
    "timestamp": "2025-07-14T12:34:31.895183"
  },
  {
    "epoch": 1629,
    "batch_idx": 160,
    "loss": 0.21560704708099365,
    "learning_rate": 8.25280362193542e-05,
    "timestep_mean": 438.34375,
    "epsilon_norm": 7.678967475891113,
    "prediction_norm": 6.359321594238281,
    "timestamp": "2025-07-14T12:34:31.900309"
  },
  {
    "epoch": 1629,
    "batch_idx": 192,
    "loss": 0.45775124430656433,
    "learning_rate": 8.25280362193542e-05,
    "timestep_mean": 380.875,
    "epsilon_norm": 5.280048847198486,
    "prediction_norm": 4.331646919250488,
    "timestamp": "2025-07-14T12:34:31.905413"
  },
  {
    "epoch": 1630,
    "batch_idx": 0,
    "loss": 0.27895838022232056,
    "learning_rate": 8.209631931586483e-05,
    "timestep_mean": 539.125,
    "epsilon_norm": 8.93542194366455,
    "prediction_norm": 7.566412448883057,
    "timestamp": "2025-07-14T12:34:31.910491"
  },
  {
    "epoch": 1630,
    "batch_idx": 32,
    "loss": 0.18975308537483215,
    "learning_rate": 8.209631931586483e-05,
    "timestep_mean": 580.96875,
    "epsilon_norm": 7.754727840423584,
    "prediction_norm": 7.042322635650635,
    "timestamp": "2025-07-14T12:34:31.915746"
  },
  {
    "epoch": 1630,
    "batch_idx": 64,
    "loss": 0.19719111919403076,
    "learning_rate": 8.209631931586483e-05,
    "timestep_mean": 505.84375,
    "epsilon_norm": 7.987723350524902,
    "prediction_norm": 6.839412212371826,
    "timestamp": "2025-07-14T12:34:31.920906"
  },
  {
    "epoch": 1630,
    "batch_idx": 96,
    "loss": 0.340844064950943,
    "learning_rate": 8.209631931586483e-05,
    "timestep_mean": 435.71875,
    "epsilon_norm": 9.523794174194336,
    "prediction_norm": 7.844087600708008,
    "timestamp": "2025-07-14T12:34:31.926216"
  },
  {
    "epoch": 1630,
    "batch_idx": 128,
    "loss": 0.13167759776115417,
    "learning_rate": 8.209631931586483e-05,
    "timestep_mean": 550.84375,
    "epsilon_norm": 7.3393754959106445,
    "prediction_norm": 6.792529106140137,
    "timestamp": "2025-07-14T12:34:31.931937"
  },
  {
    "epoch": 1630,
    "batch_idx": 160,
    "loss": 0.2685055732727051,
    "learning_rate": 8.209631931586483e-05,
    "timestep_mean": 455.15625,
    "epsilon_norm": 7.900353908538818,
    "prediction_norm": 6.335636615753174,
    "timestamp": "2025-07-14T12:34:31.946191"
  },
  {
    "epoch": 1630,
    "batch_idx": 192,
    "loss": 0.45733845233917236,
    "learning_rate": 8.209631931586483e-05,
    "timestep_mean": 511.25,
    "epsilon_norm": 4.207928657531738,
    "prediction_norm": 2.955606698989868,
    "timestamp": "2025-07-14T12:34:31.954113"
  },
  {
    "epoch": 1631,
    "batch_idx": 0,
    "loss": 0.23625579476356506,
    "learning_rate": 8.166563354816515e-05,
    "timestep_mean": 465.6875,
    "epsilon_norm": 7.8736572265625,
    "prediction_norm": 6.217837810516357,
    "timestamp": "2025-07-14T12:34:31.961606"
  },
  {
    "epoch": 1631,
    "batch_idx": 32,
    "loss": 0.29816144704818726,
    "learning_rate": 8.166563354816515e-05,
    "timestep_mean": 510.09375,
    "epsilon_norm": 7.319101810455322,
    "prediction_norm": 5.8150315284729,
    "timestamp": "2025-07-14T12:34:31.970404"
  },
  {
    "epoch": 1631,
    "batch_idx": 64,
    "loss": 0.4816865026950836,
    "learning_rate": 8.166563354816515e-05,
    "timestep_mean": 411.1875,
    "epsilon_norm": 9.809370994567871,
    "prediction_norm": 7.39178466796875,
    "timestamp": "2025-07-14T12:34:31.978341"
  },
  {
    "epoch": 1631,
    "batch_idx": 96,
    "loss": 0.21432417631149292,
    "learning_rate": 8.166563354816515e-05,
    "timestep_mean": 537.90625,
    "epsilon_norm": 8.35268497467041,
    "prediction_norm": 7.2221245765686035,
    "timestamp": "2025-07-14T12:34:31.985545"
  },
  {
    "epoch": 1631,
    "batch_idx": 128,
    "loss": 0.2394728660583496,
    "learning_rate": 8.166563354816515e-05,
    "timestep_mean": 487.75,
    "epsilon_norm": 6.824319839477539,
    "prediction_norm": 6.264561176300049,
    "timestamp": "2025-07-14T12:34:31.993415"
  },
  {
    "epoch": 1631,
    "batch_idx": 160,
    "loss": 0.3226912319660187,
    "learning_rate": 8.166563354816515e-05,
    "timestep_mean": 484.5625,
    "epsilon_norm": 8.188275337219238,
    "prediction_norm": 7.041000843048096,
    "timestamp": "2025-07-14T12:34:31.999545"
  },
  {
    "epoch": 1631,
    "batch_idx": 192,
    "loss": 1.302802562713623,
    "learning_rate": 8.166563354816515e-05,
    "timestep_mean": 399.75,
    "epsilon_norm": 4.9182305335998535,
    "prediction_norm": 3.293123483657837,
    "timestamp": "2025-07-14T12:34:32.005816"
  },
  {
    "epoch": 1632,
    "batch_idx": 0,
    "loss": 0.3430353105068207,
    "learning_rate": 8.123597997892913e-05,
    "timestep_mean": 492.59375,
    "epsilon_norm": 8.615357398986816,
    "prediction_norm": 7.108027935028076,
    "timestamp": "2025-07-14T12:34:32.013842"
  },
  {
    "epoch": 1632,
    "batch_idx": 32,
    "loss": 0.350283145904541,
    "learning_rate": 8.123597997892913e-05,
    "timestep_mean": 483.53125,
    "epsilon_norm": 7.719755172729492,
    "prediction_norm": 6.010467529296875,
    "timestamp": "2025-07-14T12:34:32.019699"
  },
  {
    "epoch": 1632,
    "batch_idx": 64,
    "loss": 0.11910946667194366,
    "learning_rate": 8.123597997892913e-05,
    "timestep_mean": 587.28125,
    "epsilon_norm": 6.825793743133545,
    "prediction_norm": 6.103509902954102,
    "timestamp": "2025-07-14T12:34:32.026884"
  },
  {
    "epoch": 1632,
    "batch_idx": 96,
    "loss": 0.15596255660057068,
    "learning_rate": 8.123597997892913e-05,
    "timestep_mean": 498.09375,
    "epsilon_norm": 7.8339619636535645,
    "prediction_norm": 6.937857627868652,
    "timestamp": "2025-07-14T12:34:32.035114"
  },
  {
    "epoch": 1632,
    "batch_idx": 128,
    "loss": 0.21543851494789124,
    "learning_rate": 8.123597997892913e-05,
    "timestep_mean": 521.28125,
    "epsilon_norm": 7.610630989074707,
    "prediction_norm": 6.850463390350342,
    "timestamp": "2025-07-14T12:34:32.041056"
  },
  {
    "epoch": 1632,
    "batch_idx": 160,
    "loss": 0.3635159730911255,
    "learning_rate": 8.123597997892913e-05,
    "timestep_mean": 431.03125,
    "epsilon_norm": 8.159989356994629,
    "prediction_norm": 6.942677021026611,
    "timestamp": "2025-07-14T12:34:32.046255"
  },
  {
    "epoch": 1632,
    "batch_idx": 192,
    "loss": 0.2613002061843872,
    "learning_rate": 8.123597997892913e-05,
    "timestep_mean": 716.125,
    "epsilon_norm": 3.985053539276123,
    "prediction_norm": 2.9722323417663574,
    "timestamp": "2025-07-14T12:34:32.052171"
  },
  {
    "epoch": 1633,
    "batch_idx": 0,
    "loss": 0.40575918555259705,
    "learning_rate": 8.080735966828441e-05,
    "timestep_mean": 459.59375,
    "epsilon_norm": 7.501701831817627,
    "prediction_norm": 5.736498832702637,
    "timestamp": "2025-07-14T12:34:32.057250"
  },
  {
    "epoch": 1633,
    "batch_idx": 32,
    "loss": 0.15383481979370117,
    "learning_rate": 8.080735966828441e-05,
    "timestep_mean": 538.125,
    "epsilon_norm": 8.253057479858398,
    "prediction_norm": 7.610269069671631,
    "timestamp": "2025-07-14T12:34:32.062203"
  },
  {
    "epoch": 1633,
    "batch_idx": 64,
    "loss": 0.4181392192840576,
    "learning_rate": 8.080735966828441e-05,
    "timestep_mean": 413.4375,
    "epsilon_norm": 7.448907852172852,
    "prediction_norm": 5.863057613372803,
    "timestamp": "2025-07-14T12:34:32.067452"
  },
  {
    "epoch": 1633,
    "batch_idx": 96,
    "loss": 0.2176159918308258,
    "learning_rate": 8.080735966828441e-05,
    "timestep_mean": 465.0,
    "epsilon_norm": 5.998535633087158,
    "prediction_norm": 5.081624507904053,
    "timestamp": "2025-07-14T12:34:32.072474"
  },
  {
    "epoch": 1633,
    "batch_idx": 128,
    "loss": 0.5203790664672852,
    "learning_rate": 8.080735966828441e-05,
    "timestep_mean": 505.09375,
    "epsilon_norm": 8.836482048034668,
    "prediction_norm": 7.299820899963379,
    "timestamp": "2025-07-14T12:34:32.077311"
  },
  {
    "epoch": 1633,
    "batch_idx": 160,
    "loss": 0.39134880900382996,
    "learning_rate": 8.080735966828441e-05,
    "timestep_mean": 544.03125,
    "epsilon_norm": 9.217341423034668,
    "prediction_norm": 7.069234848022461,
    "timestamp": "2025-07-14T12:34:32.082199"
  },
  {
    "epoch": 1633,
    "batch_idx": 192,
    "loss": 0.2247580736875534,
    "learning_rate": 8.080735966828441e-05,
    "timestep_mean": 506.375,
    "epsilon_norm": 3.4900784492492676,
    "prediction_norm": 3.081016778945923,
    "timestamp": "2025-07-14T12:34:32.087085"
  },
  {
    "epoch": 1634,
    "batch_idx": 0,
    "loss": 0.47801095247268677,
    "learning_rate": 8.037977367380918e-05,
    "timestep_mean": 506.59375,
    "epsilon_norm": 7.425300121307373,
    "prediction_norm": 5.705782413482666,
    "timestamp": "2025-07-14T12:34:32.092078"
  },
  {
    "epoch": 1634,
    "batch_idx": 32,
    "loss": 0.1607227623462677,
    "learning_rate": 8.037977367380918e-05,
    "timestep_mean": 601.96875,
    "epsilon_norm": 7.650842189788818,
    "prediction_norm": 7.50037145614624,
    "timestamp": "2025-07-14T12:34:32.097063"
  },
  {
    "epoch": 1634,
    "batch_idx": 64,
    "loss": 0.34881073236465454,
    "learning_rate": 8.037977367380918e-05,
    "timestep_mean": 568.9375,
    "epsilon_norm": 8.549166679382324,
    "prediction_norm": 7.514462471008301,
    "timestamp": "2025-07-14T12:34:32.101949"
  },
  {
    "epoch": 1634,
    "batch_idx": 96,
    "loss": 0.1812785267829895,
    "learning_rate": 8.037977367380918e-05,
    "timestep_mean": 532.59375,
    "epsilon_norm": 7.032680511474609,
    "prediction_norm": 6.2768049240112305,
    "timestamp": "2025-07-14T12:34:32.106958"
  },
  {
    "epoch": 1634,
    "batch_idx": 128,
    "loss": 0.2689978778362274,
    "learning_rate": 8.037977367380918e-05,
    "timestep_mean": 567.9375,
    "epsilon_norm": 7.561062812805176,
    "prediction_norm": 6.16898250579834,
    "timestamp": "2025-07-14T12:34:32.111842"
  },
  {
    "epoch": 1634,
    "batch_idx": 160,
    "loss": 0.24822787940502167,
    "learning_rate": 8.037977367380918e-05,
    "timestep_mean": 483.90625,
    "epsilon_norm": 7.588837146759033,
    "prediction_norm": 6.634934902191162,
    "timestamp": "2025-07-14T12:34:32.116773"
  },
  {
    "epoch": 1634,
    "batch_idx": 192,
    "loss": 0.12602074444293976,
    "learning_rate": 8.037977367380918e-05,
    "timestep_mean": 312.125,
    "epsilon_norm": 3.3734428882598877,
    "prediction_norm": 3.027696132659912,
    "timestamp": "2025-07-14T12:34:32.121809"
  },
  {
    "epoch": 1635,
    "batch_idx": 0,
    "loss": 0.1482686996459961,
    "learning_rate": 7.995322305052901e-05,
    "timestep_mean": 532.84375,
    "epsilon_norm": 7.59337854385376,
    "prediction_norm": 7.420129776000977,
    "timestamp": "2025-07-14T12:34:32.126832"
  },
  {
    "epoch": 1635,
    "batch_idx": 32,
    "loss": 0.20973005890846252,
    "learning_rate": 7.995322305052901e-05,
    "timestep_mean": 482.65625,
    "epsilon_norm": 7.9905219078063965,
    "prediction_norm": 7.147459983825684,
    "timestamp": "2025-07-14T12:34:32.131806"
  },
  {
    "epoch": 1635,
    "batch_idx": 64,
    "loss": 0.316755473613739,
    "learning_rate": 7.995322305052901e-05,
    "timestep_mean": 509.34375,
    "epsilon_norm": 9.42094612121582,
    "prediction_norm": 7.9240922927856445,
    "timestamp": "2025-07-14T12:34:32.136765"
  },
  {
    "epoch": 1635,
    "batch_idx": 96,
    "loss": 0.40433192253112793,
    "learning_rate": 7.995322305052901e-05,
    "timestep_mean": 464.625,
    "epsilon_norm": 8.744695663452148,
    "prediction_norm": 7.13225793838501,
    "timestamp": "2025-07-14T12:34:32.141714"
  },
  {
    "epoch": 1635,
    "batch_idx": 128,
    "loss": 0.22751224040985107,
    "learning_rate": 7.995322305052901e-05,
    "timestep_mean": 490.1875,
    "epsilon_norm": 7.670916557312012,
    "prediction_norm": 6.188337326049805,
    "timestamp": "2025-07-14T12:34:32.146736"
  },
  {
    "epoch": 1635,
    "batch_idx": 160,
    "loss": 0.2933582067489624,
    "learning_rate": 7.995322305052901e-05,
    "timestep_mean": 398.25,
    "epsilon_norm": 8.189815521240234,
    "prediction_norm": 6.283876895904541,
    "timestamp": "2025-07-14T12:34:32.151711"
  },
  {
    "epoch": 1635,
    "batch_idx": 192,
    "loss": 0.18628016114234924,
    "learning_rate": 7.995322305052901e-05,
    "timestep_mean": 611.0,
    "epsilon_norm": 3.0615837574005127,
    "prediction_norm": 2.721531391143799,
    "timestamp": "2025-07-14T12:34:32.156640"
  },
  {
    "epoch": 1636,
    "batch_idx": 0,
    "loss": 0.11956140398979187,
    "learning_rate": 7.952770885091545e-05,
    "timestep_mean": 512.5625,
    "epsilon_norm": 7.500540733337402,
    "prediction_norm": 7.048319339752197,
    "timestamp": "2025-07-14T12:34:32.161665"
  },
  {
    "epoch": 1636,
    "batch_idx": 32,
    "loss": 0.3218861520290375,
    "learning_rate": 7.952770885091545e-05,
    "timestep_mean": 487.75,
    "epsilon_norm": 8.101296424865723,
    "prediction_norm": 6.973961353302002,
    "timestamp": "2025-07-14T12:34:32.166628"
  },
  {
    "epoch": 1636,
    "batch_idx": 64,
    "loss": 0.34787747263908386,
    "learning_rate": 7.952770885091545e-05,
    "timestep_mean": 584.21875,
    "epsilon_norm": 7.966146945953369,
    "prediction_norm": 6.815479278564453,
    "timestamp": "2025-07-14T12:34:32.171604"
  },
  {
    "epoch": 1636,
    "batch_idx": 96,
    "loss": 0.19430822134017944,
    "learning_rate": 7.952770885091545e-05,
    "timestep_mean": 593.9375,
    "epsilon_norm": 7.12400484085083,
    "prediction_norm": 6.268069744110107,
    "timestamp": "2025-07-14T12:34:32.176522"
  },
  {
    "epoch": 1636,
    "batch_idx": 128,
    "loss": 0.37362486124038696,
    "learning_rate": 7.952770885091545e-05,
    "timestep_mean": 547.875,
    "epsilon_norm": 7.705471992492676,
    "prediction_norm": 6.208401203155518,
    "timestamp": "2025-07-14T12:34:32.181461"
  },
  {
    "epoch": 1636,
    "batch_idx": 160,
    "loss": 0.3732106685638428,
    "learning_rate": 7.952770885091545e-05,
    "timestep_mean": 462.21875,
    "epsilon_norm": 8.380736351013184,
    "prediction_norm": 6.456357002258301,
    "timestamp": "2025-07-14T12:34:32.186459"
  },
  {
    "epoch": 1636,
    "batch_idx": 192,
    "loss": 0.1551431268453598,
    "learning_rate": 7.952770885091545e-05,
    "timestep_mean": 431.375,
    "epsilon_norm": 4.087854385375977,
    "prediction_norm": 3.3508718013763428,
    "timestamp": "2025-07-14T12:34:32.191418"
  },
  {
    "epoch": 1637,
    "batch_idx": 0,
    "loss": 0.17158398032188416,
    "learning_rate": 7.910323212488238e-05,
    "timestep_mean": 490.21875,
    "epsilon_norm": 7.027459621429443,
    "prediction_norm": 6.350001335144043,
    "timestamp": "2025-07-14T12:34:32.196396"
  },
  {
    "epoch": 1637,
    "batch_idx": 32,
    "loss": 0.3116123080253601,
    "learning_rate": 7.910323212488238e-05,
    "timestep_mean": 515.5625,
    "epsilon_norm": 8.225255012512207,
    "prediction_norm": 7.108585357666016,
    "timestamp": "2025-07-14T12:34:32.201348"
  },
  {
    "epoch": 1637,
    "batch_idx": 64,
    "loss": 0.11971033364534378,
    "learning_rate": 7.910323212488238e-05,
    "timestep_mean": 560.125,
    "epsilon_norm": 8.24726390838623,
    "prediction_norm": 7.901159763336182,
    "timestamp": "2025-07-14T12:34:32.206289"
  },
  {
    "epoch": 1637,
    "batch_idx": 96,
    "loss": 0.39619266986846924,
    "learning_rate": 7.910323212488238e-05,
    "timestep_mean": 502.0,
    "epsilon_norm": 8.779507637023926,
    "prediction_norm": 6.599456787109375,
    "timestamp": "2025-07-14T12:34:32.211235"
  },
  {
    "epoch": 1637,
    "batch_idx": 128,
    "loss": 0.4330528974533081,
    "learning_rate": 7.910323212488238e-05,
    "timestep_mean": 384.03125,
    "epsilon_norm": 8.711712837219238,
    "prediction_norm": 7.321136474609375,
    "timestamp": "2025-07-14T12:34:32.216180"
  },
  {
    "epoch": 1637,
    "batch_idx": 160,
    "loss": 0.13696354627609253,
    "learning_rate": 7.910323212488238e-05,
    "timestep_mean": 607.59375,
    "epsilon_norm": 8.235363960266113,
    "prediction_norm": 7.1696672439575195,
    "timestamp": "2025-07-14T12:34:32.221181"
  },
  {
    "epoch": 1637,
    "batch_idx": 192,
    "loss": 0.3746500015258789,
    "learning_rate": 7.910323212488238e-05,
    "timestep_mean": 441.875,
    "epsilon_norm": 4.23333740234375,
    "prediction_norm": 3.4854323863983154,
    "timestamp": "2025-07-14T12:34:32.226112"
  },
  {
    "epoch": 1638,
    "batch_idx": 0,
    "loss": 0.29258254170417786,
    "learning_rate": 7.867979391978394e-05,
    "timestep_mean": 456.6875,
    "epsilon_norm": 7.270709037780762,
    "prediction_norm": 6.399190425872803,
    "timestamp": "2025-07-14T12:34:32.231107"
  },
  {
    "epoch": 1638,
    "batch_idx": 32,
    "loss": 0.28271594643592834,
    "learning_rate": 7.867979391978394e-05,
    "timestep_mean": 428.15625,
    "epsilon_norm": 8.570326805114746,
    "prediction_norm": 7.728513240814209,
    "timestamp": "2025-07-14T12:34:32.236050"
  },
  {
    "epoch": 1638,
    "batch_idx": 64,
    "loss": 0.25269997119903564,
    "learning_rate": 7.867979391978394e-05,
    "timestep_mean": 575.3125,
    "epsilon_norm": 8.665204048156738,
    "prediction_norm": 7.901689529418945,
    "timestamp": "2025-07-14T12:34:32.240994"
  },
  {
    "epoch": 1638,
    "batch_idx": 96,
    "loss": 0.247138112783432,
    "learning_rate": 7.867979391978394e-05,
    "timestep_mean": 563.21875,
    "epsilon_norm": 7.32360315322876,
    "prediction_norm": 6.489481449127197,
    "timestamp": "2025-07-14T12:34:32.245982"
  },
  {
    "epoch": 1638,
    "batch_idx": 128,
    "loss": 0.26487302780151367,
    "learning_rate": 7.867979391978394e-05,
    "timestep_mean": 453.875,
    "epsilon_norm": 6.371744632720947,
    "prediction_norm": 5.411334991455078,
    "timestamp": "2025-07-14T12:34:32.251054"
  },
  {
    "epoch": 1638,
    "batch_idx": 160,
    "loss": 0.3816566467285156,
    "learning_rate": 7.867979391978394e-05,
    "timestep_mean": 452.34375,
    "epsilon_norm": 7.359189033508301,
    "prediction_norm": 6.2920308113098145,
    "timestamp": "2025-07-14T12:34:32.256147"
  },
  {
    "epoch": 1638,
    "batch_idx": 192,
    "loss": 0.03401702642440796,
    "learning_rate": 7.867979391978394e-05,
    "timestep_mean": 684.875,
    "epsilon_norm": 3.5427865982055664,
    "prediction_norm": 3.6451072692871094,
    "timestamp": "2025-07-14T12:34:32.261127"
  },
  {
    "epoch": 1639,
    "batch_idx": 0,
    "loss": 0.4663681983947754,
    "learning_rate": 7.825739528041167e-05,
    "timestep_mean": 527.125,
    "epsilon_norm": 8.449686050415039,
    "prediction_norm": 6.2921648025512695,
    "timestamp": "2025-07-14T12:34:32.266872"
  },
  {
    "epoch": 1639,
    "batch_idx": 32,
    "loss": 0.315884530544281,
    "learning_rate": 7.825739528041167e-05,
    "timestep_mean": 504.46875,
    "epsilon_norm": 8.574906349182129,
    "prediction_norm": 7.275031089782715,
    "timestamp": "2025-07-14T12:34:32.272246"
  },
  {
    "epoch": 1639,
    "batch_idx": 64,
    "loss": 0.4215947985649109,
    "learning_rate": 7.825739528041167e-05,
    "timestep_mean": 454.25,
    "epsilon_norm": 8.52751636505127,
    "prediction_norm": 7.068850040435791,
    "timestamp": "2025-07-14T12:34:32.283038"
  },
  {
    "epoch": 1639,
    "batch_idx": 96,
    "loss": 0.3524826765060425,
    "learning_rate": 7.825739528041167e-05,
    "timestep_mean": 501.53125,
    "epsilon_norm": 8.215657234191895,
    "prediction_norm": 7.016096591949463,
    "timestamp": "2025-07-14T12:34:32.290399"
  },
  {
    "epoch": 1639,
    "batch_idx": 128,
    "loss": 0.246529221534729,
    "learning_rate": 7.825739528041167e-05,
    "timestep_mean": 442.9375,
    "epsilon_norm": 8.271480560302734,
    "prediction_norm": 6.946450233459473,
    "timestamp": "2025-07-14T12:34:32.296678"
  },
  {
    "epoch": 1639,
    "batch_idx": 160,
    "loss": 0.24991698563098907,
    "learning_rate": 7.825739528041167e-05,
    "timestep_mean": 503.28125,
    "epsilon_norm": 7.420044422149658,
    "prediction_norm": 6.337549686431885,
    "timestamp": "2025-07-14T12:34:32.309154"
  },
  {
    "epoch": 1639,
    "batch_idx": 192,
    "loss": 0.11080388724803925,
    "learning_rate": 7.825739528041167e-05,
    "timestep_mean": 480.5,
    "epsilon_norm": 2.864622116088867,
    "prediction_norm": 2.7649996280670166,
    "timestamp": "2025-07-14T12:34:32.315465"
  },
  {
    "epoch": 1640,
    "batch_idx": 0,
    "loss": 0.3456663489341736,
    "learning_rate": 7.783603724899243e-05,
    "timestep_mean": 429.90625,
    "epsilon_norm": 9.262055397033691,
    "prediction_norm": 8.300321578979492,
    "timestamp": "2025-07-14T12:34:32.326182"
  },
  {
    "epoch": 1640,
    "batch_idx": 32,
    "loss": 0.18589870631694794,
    "learning_rate": 7.783603724899243e-05,
    "timestep_mean": 555.03125,
    "epsilon_norm": 6.870731830596924,
    "prediction_norm": 6.17136287689209,
    "timestamp": "2025-07-14T12:34:32.337905"
  },
  {
    "epoch": 1640,
    "batch_idx": 64,
    "loss": 0.23362889885902405,
    "learning_rate": 7.783603724899243e-05,
    "timestep_mean": 572.40625,
    "epsilon_norm": 10.020267486572266,
    "prediction_norm": 8.819204330444336,
    "timestamp": "2025-07-14T12:34:32.346155"
  },
  {
    "epoch": 1640,
    "batch_idx": 96,
    "loss": 0.21294081211090088,
    "learning_rate": 7.783603724899243e-05,
    "timestep_mean": 539.90625,
    "epsilon_norm": 8.976569175720215,
    "prediction_norm": 7.737376689910889,
    "timestamp": "2025-07-14T12:34:32.351816"
  },
  {
    "epoch": 1640,
    "batch_idx": 128,
    "loss": 0.10584630072116852,
    "learning_rate": 7.783603724899243e-05,
    "timestep_mean": 619.625,
    "epsilon_norm": 8.398726463317871,
    "prediction_norm": 8.080851554870605,
    "timestamp": "2025-07-14T12:34:32.357367"
  },
  {
    "epoch": 1640,
    "batch_idx": 160,
    "loss": 0.3895701766014099,
    "learning_rate": 7.783603724899243e-05,
    "timestep_mean": 494.71875,
    "epsilon_norm": 8.40449047088623,
    "prediction_norm": 6.897134304046631,
    "timestamp": "2025-07-14T12:34:32.362879"
  },
  {
    "epoch": 1640,
    "batch_idx": 192,
    "loss": 0.5879381895065308,
    "learning_rate": 7.783603724899243e-05,
    "timestep_mean": 588.875,
    "epsilon_norm": 4.326840400695801,
    "prediction_norm": 3.51955246925354,
    "timestamp": "2025-07-14T12:34:32.368266"
  },
  {
    "epoch": 1641,
    "batch_idx": 0,
    "loss": 0.44709479808807373,
    "learning_rate": 7.741572086518524e-05,
    "timestep_mean": 411.9375,
    "epsilon_norm": 8.041714668273926,
    "prediction_norm": 6.225198268890381,
    "timestamp": "2025-07-14T12:34:32.373608"
  },
  {
    "epoch": 1641,
    "batch_idx": 32,
    "loss": 0.393976092338562,
    "learning_rate": 7.741572086518524e-05,
    "timestep_mean": 449.59375,
    "epsilon_norm": 7.863927364349365,
    "prediction_norm": 6.598960876464844,
    "timestamp": "2025-07-14T12:34:32.378874"
  },
  {
    "epoch": 1641,
    "batch_idx": 64,
    "loss": 0.27229055762290955,
    "learning_rate": 7.741572086518524e-05,
    "timestep_mean": 502.0,
    "epsilon_norm": 6.966714382171631,
    "prediction_norm": 5.494593143463135,
    "timestamp": "2025-07-14T12:34:32.384383"
  },
  {
    "epoch": 1641,
    "batch_idx": 96,
    "loss": 0.2115926295518875,
    "learning_rate": 7.741572086518524e-05,
    "timestep_mean": 472.4375,
    "epsilon_norm": 6.796059608459473,
    "prediction_norm": 5.619232654571533,
    "timestamp": "2025-07-14T12:34:32.390357"
  },
  {
    "epoch": 1641,
    "batch_idx": 128,
    "loss": 0.3561018407344818,
    "learning_rate": 7.741572086518524e-05,
    "timestep_mean": 534.6875,
    "epsilon_norm": 7.082871913909912,
    "prediction_norm": 6.297478199005127,
    "timestamp": "2025-07-14T12:34:32.395958"
  },
  {
    "epoch": 1641,
    "batch_idx": 160,
    "loss": 0.3493698239326477,
    "learning_rate": 7.741572086518524e-05,
    "timestep_mean": 460.3125,
    "epsilon_norm": 9.496118545532227,
    "prediction_norm": 8.02756404876709,
    "timestamp": "2025-07-14T12:34:32.402164"
  },
  {
    "epoch": 1641,
    "batch_idx": 192,
    "loss": 0.36060023307800293,
    "learning_rate": 7.741572086518524e-05,
    "timestep_mean": 485.5,
    "epsilon_norm": 4.482157230377197,
    "prediction_norm": 4.096503734588623,
    "timestamp": "2025-07-14T12:34:32.407498"
  },
  {
    "epoch": 1642,
    "batch_idx": 0,
    "loss": 0.3161351680755615,
    "learning_rate": 7.699644716607892e-05,
    "timestep_mean": 515.75,
    "epsilon_norm": 8.7705078125,
    "prediction_norm": 7.2047882080078125,
    "timestamp": "2025-07-14T12:34:32.412628"
  },
  {
    "epoch": 1642,
    "batch_idx": 32,
    "loss": 0.16508617997169495,
    "learning_rate": 7.699644716607892e-05,
    "timestep_mean": 567.65625,
    "epsilon_norm": 7.629784107208252,
    "prediction_norm": 7.367612361907959,
    "timestamp": "2025-07-14T12:34:32.418243"
  },
  {
    "epoch": 1642,
    "batch_idx": 64,
    "loss": 0.28038153052330017,
    "learning_rate": 7.699644716607892e-05,
    "timestep_mean": 480.9375,
    "epsilon_norm": 8.303683280944824,
    "prediction_norm": 6.69153356552124,
    "timestamp": "2025-07-14T12:34:32.423966"
  },
  {
    "epoch": 1642,
    "batch_idx": 96,
    "loss": 0.25463587045669556,
    "learning_rate": 7.699644716607892e-05,
    "timestep_mean": 548.65625,
    "epsilon_norm": 7.511183738708496,
    "prediction_norm": 6.083690643310547,
    "timestamp": "2025-07-14T12:34:32.429498"
  },
  {
    "epoch": 1642,
    "batch_idx": 128,
    "loss": 0.3958072364330292,
    "learning_rate": 7.699644716607892e-05,
    "timestep_mean": 469.90625,
    "epsilon_norm": 7.495680332183838,
    "prediction_norm": 5.703604221343994,
    "timestamp": "2025-07-14T12:34:32.434809"
  },
  {
    "epoch": 1642,
    "batch_idx": 160,
    "loss": 0.3825720548629761,
    "learning_rate": 7.699644716607892e-05,
    "timestep_mean": 517.28125,
    "epsilon_norm": 7.850981712341309,
    "prediction_norm": 6.1275200843811035,
    "timestamp": "2025-07-14T12:34:32.440127"
  },
  {
    "epoch": 1642,
    "batch_idx": 192,
    "loss": 0.3360980749130249,
    "learning_rate": 7.699644716607892e-05,
    "timestep_mean": 219.5,
    "epsilon_norm": 3.2131519317626953,
    "prediction_norm": 2.637970209121704,
    "timestamp": "2025-07-14T12:34:32.445496"
  },
  {
    "epoch": 1643,
    "batch_idx": 0,
    "loss": 0.22641518712043762,
    "learning_rate": 7.657821718618971e-05,
    "timestep_mean": 488.6875,
    "epsilon_norm": 7.578366756439209,
    "prediction_norm": 6.092158317565918,
    "timestamp": "2025-07-14T12:34:32.451027"
  },
  {
    "epoch": 1643,
    "batch_idx": 32,
    "loss": 0.18385940790176392,
    "learning_rate": 7.657821718618971e-05,
    "timestep_mean": 555.09375,
    "epsilon_norm": 7.2892842292785645,
    "prediction_norm": 6.618662357330322,
    "timestamp": "2025-07-14T12:34:32.456453"
  },
  {
    "epoch": 1643,
    "batch_idx": 64,
    "loss": 0.11580118536949158,
    "learning_rate": 7.657821718618971e-05,
    "timestep_mean": 568.6875,
    "epsilon_norm": 7.092007637023926,
    "prediction_norm": 6.658729553222656,
    "timestamp": "2025-07-14T12:34:32.461848"
  },
  {
    "epoch": 1643,
    "batch_idx": 96,
    "loss": 0.258279412984848,
    "learning_rate": 7.657821718618971e-05,
    "timestep_mean": 516.65625,
    "epsilon_norm": 7.375441074371338,
    "prediction_norm": 6.030046463012695,
    "timestamp": "2025-07-14T12:34:32.467205"
  },
  {
    "epoch": 1643,
    "batch_idx": 128,
    "loss": 0.40995174646377563,
    "learning_rate": 7.657821718618971e-05,
    "timestep_mean": 455.1875,
    "epsilon_norm": 8.35411548614502,
    "prediction_norm": 6.37273645401001,
    "timestamp": "2025-07-14T12:34:32.472494"
  },
  {
    "epoch": 1643,
    "batch_idx": 160,
    "loss": 0.22442224621772766,
    "learning_rate": 7.657821718618971e-05,
    "timestep_mean": 496.53125,
    "epsilon_norm": 7.287723541259766,
    "prediction_norm": 6.526642322540283,
    "timestamp": "2025-07-14T12:34:32.477861"
  },
  {
    "epoch": 1643,
    "batch_idx": 192,
    "loss": 0.21271590888500214,
    "learning_rate": 7.657821718618971e-05,
    "timestep_mean": 614.125,
    "epsilon_norm": 6.052361965179443,
    "prediction_norm": 5.0161261558532715,
    "timestamp": "2025-07-14T12:34:32.483321"
  },
  {
    "epoch": 1644,
    "batch_idx": 0,
    "loss": 0.31106048822402954,
    "learning_rate": 7.616103195745836e-05,
    "timestep_mean": 546.78125,
    "epsilon_norm": 7.888735771179199,
    "prediction_norm": 6.388722896575928,
    "timestamp": "2025-07-14T12:34:32.489488"
  },
  {
    "epoch": 1644,
    "batch_idx": 32,
    "loss": 0.15860101580619812,
    "learning_rate": 7.616103195745836e-05,
    "timestep_mean": 532.46875,
    "epsilon_norm": 6.892941474914551,
    "prediction_norm": 6.022369861602783,
    "timestamp": "2025-07-14T12:34:32.495258"
  },
  {
    "epoch": 1644,
    "batch_idx": 64,
    "loss": 0.15717828273773193,
    "learning_rate": 7.616103195745836e-05,
    "timestep_mean": 501.1875,
    "epsilon_norm": 7.510120391845703,
    "prediction_norm": 6.8913726806640625,
    "timestamp": "2025-07-14T12:34:32.507196"
  },
  {
    "epoch": 1644,
    "batch_idx": 96,
    "loss": 0.3077751398086548,
    "learning_rate": 7.616103195745836e-05,
    "timestep_mean": 567.6875,
    "epsilon_norm": 6.696477890014648,
    "prediction_norm": 5.23543643951416,
    "timestamp": "2025-07-14T12:34:32.512972"
  },
  {
    "epoch": 1644,
    "batch_idx": 128,
    "loss": 0.20718315243721008,
    "learning_rate": 7.616103195745836e-05,
    "timestep_mean": 560.46875,
    "epsilon_norm": 7.695473670959473,
    "prediction_norm": 7.385583877563477,
    "timestamp": "2025-07-14T12:34:32.518293"
  },
  {
    "epoch": 1644,
    "batch_idx": 160,
    "loss": 0.10411199927330017,
    "learning_rate": 7.616103195745836e-05,
    "timestep_mean": 581.84375,
    "epsilon_norm": 8.569981575012207,
    "prediction_norm": 7.7586798667907715,
    "timestamp": "2025-07-14T12:34:32.524373"
  },
  {
    "epoch": 1644,
    "batch_idx": 192,
    "loss": 0.2360132783651352,
    "learning_rate": 7.616103195745836e-05,
    "timestep_mean": 507.75,
    "epsilon_norm": 4.925083637237549,
    "prediction_norm": 3.979116678237915,
    "timestamp": "2025-07-14T12:34:32.530175"
  },
  {
    "epoch": 1645,
    "batch_idx": 0,
    "loss": 0.260057270526886,
    "learning_rate": 7.57448925092483e-05,
    "timestep_mean": 510.71875,
    "epsilon_norm": 7.669595241546631,
    "prediction_norm": 6.643360614776611,
    "timestamp": "2025-07-14T12:34:32.540803"
  },
  {
    "epoch": 1645,
    "batch_idx": 32,
    "loss": 0.3738304674625397,
    "learning_rate": 7.57448925092483e-05,
    "timestep_mean": 471.03125,
    "epsilon_norm": 8.666824340820312,
    "prediction_norm": 6.666581630706787,
    "timestamp": "2025-07-14T12:34:32.546854"
  },
  {
    "epoch": 1645,
    "batch_idx": 64,
    "loss": 0.3885345757007599,
    "learning_rate": 7.57448925092483e-05,
    "timestep_mean": 464.1875,
    "epsilon_norm": 8.065848350524902,
    "prediction_norm": 5.590952396392822,
    "timestamp": "2025-07-14T12:34:32.552180"
  },
  {
    "epoch": 1645,
    "batch_idx": 96,
    "loss": 0.3872181177139282,
    "learning_rate": 7.57448925092483e-05,
    "timestep_mean": 545.03125,
    "epsilon_norm": 7.983953952789307,
    "prediction_norm": 6.766354084014893,
    "timestamp": "2025-07-14T12:34:32.558541"
  },
  {
    "epoch": 1645,
    "batch_idx": 128,
    "loss": 0.19036151468753815,
    "learning_rate": 7.57448925092483e-05,
    "timestep_mean": 538.34375,
    "epsilon_norm": 7.63304328918457,
    "prediction_norm": 7.242898941040039,
    "timestamp": "2025-07-14T12:34:32.564076"
  },
  {
    "epoch": 1645,
    "batch_idx": 160,
    "loss": 0.12827511131763458,
    "learning_rate": 7.57448925092483e-05,
    "timestep_mean": 620.75,
    "epsilon_norm": 7.358423709869385,
    "prediction_norm": 6.5901618003845215,
    "timestamp": "2025-07-14T12:34:32.574975"
  },
  {
    "epoch": 1645,
    "batch_idx": 192,
    "loss": 0.04368523508310318,
    "learning_rate": 7.57448925092483e-05,
    "timestep_mean": 627.5,
    "epsilon_norm": 2.5687849521636963,
    "prediction_norm": 2.4530181884765625,
    "timestamp": "2025-07-14T12:34:32.580873"
  },
  {
    "epoch": 1646,
    "batch_idx": 0,
    "loss": 0.2310517430305481,
    "learning_rate": 7.532979986834174e-05,
    "timestep_mean": 535.5,
    "epsilon_norm": 9.015945434570312,
    "prediction_norm": 8.120841026306152,
    "timestamp": "2025-07-14T12:34:32.586435"
  },
  {
    "epoch": 1646,
    "batch_idx": 32,
    "loss": 0.20237183570861816,
    "learning_rate": 7.532979986834174e-05,
    "timestep_mean": 491.84375,
    "epsilon_norm": 6.859829425811768,
    "prediction_norm": 5.5610785484313965,
    "timestamp": "2025-07-14T12:34:32.591863"
  },
  {
    "epoch": 1646,
    "batch_idx": 64,
    "loss": 0.45712271332740784,
    "learning_rate": 7.532979986834174e-05,
    "timestep_mean": 464.15625,
    "epsilon_norm": 9.381043434143066,
    "prediction_norm": 7.411005020141602,
    "timestamp": "2025-07-14T12:34:32.598234"
  },
  {
    "epoch": 1646,
    "batch_idx": 96,
    "loss": 0.22004477679729462,
    "learning_rate": 7.532979986834174e-05,
    "timestep_mean": 466.90625,
    "epsilon_norm": 6.762755393981934,
    "prediction_norm": 5.081509590148926,
    "timestamp": "2025-07-14T12:34:32.604105"
  },
  {
    "epoch": 1646,
    "batch_idx": 128,
    "loss": 0.3641323745250702,
    "learning_rate": 7.532979986834174e-05,
    "timestep_mean": 529.9375,
    "epsilon_norm": 7.551634788513184,
    "prediction_norm": 6.561857223510742,
    "timestamp": "2025-07-14T12:34:32.615591"
  },
  {
    "epoch": 1646,
    "batch_idx": 160,
    "loss": 0.3268347382545471,
    "learning_rate": 7.532979986834174e-05,
    "timestep_mean": 488.65625,
    "epsilon_norm": 8.494534492492676,
    "prediction_norm": 7.20453405380249,
    "timestamp": "2025-07-14T12:34:32.621385"
  },
  {
    "epoch": 1646,
    "batch_idx": 192,
    "loss": 0.10128343105316162,
    "learning_rate": 7.532979986834174e-05,
    "timestep_mean": 574.75,
    "epsilon_norm": 2.693101167678833,
    "prediction_norm": 2.9900107383728027,
    "timestamp": "2025-07-14T12:34:32.626652"
  },
  {
    "epoch": 1647,
    "batch_idx": 0,
    "loss": 0.25973716378211975,
    "learning_rate": 7.49157550589388e-05,
    "timestep_mean": 565.1875,
    "epsilon_norm": 8.868110656738281,
    "prediction_norm": 7.3717570304870605,
    "timestamp": "2025-07-14T12:34:32.631900"
  },
  {
    "epoch": 1647,
    "batch_idx": 32,
    "loss": 0.3988149166107178,
    "learning_rate": 7.49157550589388e-05,
    "timestep_mean": 498.03125,
    "epsilon_norm": 8.561746597290039,
    "prediction_norm": 6.942702770233154,
    "timestamp": "2025-07-14T12:34:32.637178"
  },
  {
    "epoch": 1647,
    "batch_idx": 64,
    "loss": 0.28415346145629883,
    "learning_rate": 7.49157550589388e-05,
    "timestep_mean": 564.4375,
    "epsilon_norm": 9.164247512817383,
    "prediction_norm": 7.7122063636779785,
    "timestamp": "2025-07-14T12:34:32.642249"
  },
  {
    "epoch": 1647,
    "batch_idx": 96,
    "loss": 0.18535500764846802,
    "learning_rate": 7.49157550589388e-05,
    "timestep_mean": 478.4375,
    "epsilon_norm": 7.449334621429443,
    "prediction_norm": 6.6483659744262695,
    "timestamp": "2025-07-14T12:34:32.647354"
  },
  {
    "epoch": 1647,
    "batch_idx": 128,
    "loss": 0.12507162988185883,
    "learning_rate": 7.49157550589388e-05,
    "timestep_mean": 525.09375,
    "epsilon_norm": 8.040863037109375,
    "prediction_norm": 7.433267593383789,
    "timestamp": "2025-07-14T12:34:32.652484"
  },
  {
    "epoch": 1647,
    "batch_idx": 160,
    "loss": 0.1629800796508789,
    "learning_rate": 7.49157550589388e-05,
    "timestep_mean": 576.5,
    "epsilon_norm": 7.055954933166504,
    "prediction_norm": 6.360845565795898,
    "timestamp": "2025-07-14T12:34:32.657606"
  },
  {
    "epoch": 1647,
    "batch_idx": 192,
    "loss": 0.3198896646499634,
    "learning_rate": 7.49157550589388e-05,
    "timestep_mean": 414.25,
    "epsilon_norm": 3.668996572494507,
    "prediction_norm": 2.918713331222534,
    "timestamp": "2025-07-14T12:34:32.663076"
  },
  {
    "epoch": 1648,
    "batch_idx": 0,
    "loss": 0.39393681287765503,
    "learning_rate": 7.450275910265417e-05,
    "timestep_mean": 559.34375,
    "epsilon_norm": 8.968391418457031,
    "prediction_norm": 7.456043243408203,
    "timestamp": "2025-07-14T12:34:32.668365"
  },
  {
    "epoch": 1648,
    "batch_idx": 32,
    "loss": 0.1924750804901123,
    "learning_rate": 7.450275910265417e-05,
    "timestep_mean": 475.1875,
    "epsilon_norm": 8.250795364379883,
    "prediction_norm": 6.831096172332764,
    "timestamp": "2025-07-14T12:34:32.673883"
  },
  {
    "epoch": 1648,
    "batch_idx": 64,
    "loss": 0.36278223991394043,
    "learning_rate": 7.450275910265417e-05,
    "timestep_mean": 457.5,
    "epsilon_norm": 7.541567325592041,
    "prediction_norm": 6.128804683685303,
    "timestamp": "2025-07-14T12:34:32.679417"
  },
  {
    "epoch": 1648,
    "batch_idx": 96,
    "loss": 0.43454575538635254,
    "learning_rate": 7.450275910265417e-05,
    "timestep_mean": 545.96875,
    "epsilon_norm": 9.852860450744629,
    "prediction_norm": 7.922823905944824,
    "timestamp": "2025-07-14T12:34:32.684608"
  },
  {
    "epoch": 1648,
    "batch_idx": 128,
    "loss": 0.14679814875125885,
    "learning_rate": 7.450275910265417e-05,
    "timestep_mean": 596.21875,
    "epsilon_norm": 7.831194877624512,
    "prediction_norm": 6.728182792663574,
    "timestamp": "2025-07-14T12:34:32.689981"
  },
  {
    "epoch": 1648,
    "batch_idx": 160,
    "loss": 0.24716705083847046,
    "learning_rate": 7.450275910265417e-05,
    "timestep_mean": 510.28125,
    "epsilon_norm": 7.867097854614258,
    "prediction_norm": 6.777659893035889,
    "timestamp": "2025-07-14T12:34:32.695628"
  },
  {
    "epoch": 1648,
    "batch_idx": 192,
    "loss": 0.6022005677223206,
    "learning_rate": 7.450275910265417e-05,
    "timestep_mean": 438.625,
    "epsilon_norm": 4.682680130004883,
    "prediction_norm": 3.49819016456604,
    "timestamp": "2025-07-14T12:34:32.701023"
  },
  {
    "epoch": 1649,
    "batch_idx": 0,
    "loss": 0.23697666823863983,
    "learning_rate": 7.409081301851371e-05,
    "timestep_mean": 410.53125,
    "epsilon_norm": 7.063823699951172,
    "prediction_norm": 5.98218297958374,
    "timestamp": "2025-07-14T12:34:32.706346"
  },
  {
    "epoch": 1649,
    "batch_idx": 32,
    "loss": 0.2843126058578491,
    "learning_rate": 7.409081301851371e-05,
    "timestep_mean": 493.09375,
    "epsilon_norm": 9.055377006530762,
    "prediction_norm": 8.401737213134766,
    "timestamp": "2025-07-14T12:34:32.711835"
  },
  {
    "epoch": 1649,
    "batch_idx": 64,
    "loss": 0.17528805136680603,
    "learning_rate": 7.409081301851371e-05,
    "timestep_mean": 479.34375,
    "epsilon_norm": 8.1277437210083,
    "prediction_norm": 7.526021957397461,
    "timestamp": "2025-07-14T12:34:32.717334"
  },
  {
    "epoch": 1649,
    "batch_idx": 96,
    "loss": 0.2949497699737549,
    "learning_rate": 7.409081301851371e-05,
    "timestep_mean": 467.90625,
    "epsilon_norm": 8.994942665100098,
    "prediction_norm": 7.215368270874023,
    "timestamp": "2025-07-14T12:34:32.722664"
  },
  {
    "epoch": 1649,
    "batch_idx": 128,
    "loss": 0.2279130518436432,
    "learning_rate": 7.409081301851371e-05,
    "timestep_mean": 503.625,
    "epsilon_norm": 6.85687780380249,
    "prediction_norm": 5.679042816162109,
    "timestamp": "2025-07-14T12:34:32.728239"
  },
  {
    "epoch": 1649,
    "batch_idx": 160,
    "loss": 0.34698688983917236,
    "learning_rate": 7.409081301851371e-05,
    "timestep_mean": 515.03125,
    "epsilon_norm": 7.872559070587158,
    "prediction_norm": 5.559693813323975,
    "timestamp": "2025-07-14T12:34:32.733612"
  },
  {
    "epoch": 1649,
    "batch_idx": 192,
    "loss": 0.09684402495622635,
    "learning_rate": 7.409081301851371e-05,
    "timestep_mean": 479.5,
    "epsilon_norm": 3.9532108306884766,
    "prediction_norm": 4.093323707580566,
    "timestamp": "2025-07-14T12:34:32.739055"
  },
  {
    "epoch": 1650,
    "batch_idx": 0,
    "loss": 0.33393198251724243,
    "learning_rate": 7.36799178229539e-05,
    "timestep_mean": 407.875,
    "epsilon_norm": 8.059693336486816,
    "prediction_norm": 6.478282451629639,
    "timestamp": "2025-07-14T12:34:32.744545"
  },
  {
    "epoch": 1650,
    "batch_idx": 32,
    "loss": 0.1321605145931244,
    "learning_rate": 7.36799178229539e-05,
    "timestep_mean": 636.375,
    "epsilon_norm": 8.284987449645996,
    "prediction_norm": 7.795348167419434,
    "timestamp": "2025-07-14T12:34:32.749937"
  },
  {
    "epoch": 1650,
    "batch_idx": 64,
    "loss": 0.18306782841682434,
    "learning_rate": 7.36799178229539e-05,
    "timestep_mean": 492.71875,
    "epsilon_norm": 6.601309776306152,
    "prediction_norm": 5.700414657592773,
    "timestamp": "2025-07-14T12:34:32.755255"
  },
  {
    "epoch": 1650,
    "batch_idx": 96,
    "loss": 0.4587559103965759,
    "learning_rate": 7.36799178229539e-05,
    "timestep_mean": 407.1875,
    "epsilon_norm": 7.553365230560303,
    "prediction_norm": 5.673720359802246,
    "timestamp": "2025-07-14T12:34:32.760450"
  },
  {
    "epoch": 1650,
    "batch_idx": 128,
    "loss": 0.27808868885040283,
    "learning_rate": 7.36799178229539e-05,
    "timestep_mean": 432.0,
    "epsilon_norm": 7.051328182220459,
    "prediction_norm": 5.161756992340088,
    "timestamp": "2025-07-14T12:34:32.766137"
  },
  {
    "epoch": 1650,
    "batch_idx": 160,
    "loss": 0.2680877149105072,
    "learning_rate": 7.36799178229539e-05,
    "timestep_mean": 470.21875,
    "epsilon_norm": 7.947004795074463,
    "prediction_norm": 7.313596248626709,
    "timestamp": "2025-07-14T12:34:32.771378"
  },
  {
    "epoch": 1650,
    "batch_idx": 192,
    "loss": 0.009908339008688927,
    "learning_rate": 7.36799178229539e-05,
    "timestep_mean": 745.375,
    "epsilon_norm": 4.854234218597412,
    "prediction_norm": 4.779374599456787,
    "timestamp": "2025-07-14T12:34:32.782833"
  },
  {
    "epoch": 1651,
    "batch_idx": 0,
    "loss": 0.1650950014591217,
    "learning_rate": 7.327007452981764e-05,
    "timestep_mean": 489.375,
    "epsilon_norm": 7.2557830810546875,
    "prediction_norm": 6.4983978271484375,
    "timestamp": "2025-07-14T12:34:32.788402"
  },
  {
    "epoch": 1651,
    "batch_idx": 32,
    "loss": 0.22457490861415863,
    "learning_rate": 7.327007452981764e-05,
    "timestep_mean": 534.375,
    "epsilon_norm": 8.677346229553223,
    "prediction_norm": 7.468553066253662,
    "timestamp": "2025-07-14T12:34:32.793393"
  },
  {
    "epoch": 1651,
    "batch_idx": 64,
    "loss": 0.2655629813671112,
    "learning_rate": 7.327007452981764e-05,
    "timestep_mean": 473.40625,
    "epsilon_norm": 7.947567462921143,
    "prediction_norm": 7.28035306930542,
    "timestamp": "2025-07-14T12:34:32.798485"
  },
  {
    "epoch": 1651,
    "batch_idx": 96,
    "loss": 0.24031507968902588,
    "learning_rate": 7.327007452981764e-05,
    "timestep_mean": 628.625,
    "epsilon_norm": 8.017831802368164,
    "prediction_norm": 6.592957496643066,
    "timestamp": "2025-07-14T12:34:32.803512"
  },
  {
    "epoch": 1651,
    "batch_idx": 128,
    "loss": 0.3753168284893036,
    "learning_rate": 7.327007452981764e-05,
    "timestep_mean": 558.71875,
    "epsilon_norm": 7.4122209548950195,
    "prediction_norm": 5.464015007019043,
    "timestamp": "2025-07-14T12:34:32.808551"
  },
  {
    "epoch": 1651,
    "batch_idx": 160,
    "loss": 0.29061296582221985,
    "learning_rate": 7.327007452981764e-05,
    "timestep_mean": 519.625,
    "epsilon_norm": 9.622645378112793,
    "prediction_norm": 8.01658821105957,
    "timestamp": "2025-07-14T12:34:32.813652"
  },
  {
    "epoch": 1651,
    "batch_idx": 192,
    "loss": 0.43738698959350586,
    "learning_rate": 7.327007452981764e-05,
    "timestep_mean": 442.0,
    "epsilon_norm": 3.1732113361358643,
    "prediction_norm": 2.3253068923950195,
    "timestamp": "2025-07-14T12:34:32.818896"
  },
  {
    "epoch": 1652,
    "batch_idx": 0,
    "loss": 0.16549444198608398,
    "learning_rate": 7.286128415035236e-05,
    "timestep_mean": 533.53125,
    "epsilon_norm": 8.023719787597656,
    "prediction_norm": 7.124589443206787,
    "timestamp": "2025-07-14T12:34:32.824008"
  },
  {
    "epoch": 1652,
    "batch_idx": 32,
    "loss": 0.17886969447135925,
    "learning_rate": 7.286128415035236e-05,
    "timestep_mean": 509.6875,
    "epsilon_norm": 8.821694374084473,
    "prediction_norm": 7.811086654663086,
    "timestamp": "2025-07-14T12:34:32.828999"
  },
  {
    "epoch": 1652,
    "batch_idx": 64,
    "loss": 0.24538038671016693,
    "learning_rate": 7.286128415035236e-05,
    "timestep_mean": 523.09375,
    "epsilon_norm": 7.401220798492432,
    "prediction_norm": 6.5331268310546875,
    "timestamp": "2025-07-14T12:34:32.834048"
  },
  {
    "epoch": 1652,
    "batch_idx": 96,
    "loss": 0.19569982588291168,
    "learning_rate": 7.286128415035236e-05,
    "timestep_mean": 495.34375,
    "epsilon_norm": 6.995209217071533,
    "prediction_norm": 5.766253471374512,
    "timestamp": "2025-07-14T12:34:32.838978"
  },
  {
    "epoch": 1652,
    "batch_idx": 128,
    "loss": 0.258207768201828,
    "learning_rate": 7.286128415035236e-05,
    "timestep_mean": 548.84375,
    "epsilon_norm": 8.216984748840332,
    "prediction_norm": 6.790168285369873,
    "timestamp": "2025-07-14T12:34:32.843942"
  },
  {
    "epoch": 1652,
    "batch_idx": 160,
    "loss": 0.10543449223041534,
    "learning_rate": 7.286128415035236e-05,
    "timestep_mean": 590.625,
    "epsilon_norm": 7.799850940704346,
    "prediction_norm": 7.218898296356201,
    "timestamp": "2025-07-14T12:34:32.848954"
  },
  {
    "epoch": 1652,
    "batch_idx": 192,
    "loss": 0.15921524167060852,
    "learning_rate": 7.286128415035236e-05,
    "timestep_mean": 675.5,
    "epsilon_norm": 3.8573358058929443,
    "prediction_norm": 3.655998468399048,
    "timestamp": "2025-07-14T12:34:32.854038"
  },
  {
    "epoch": 1653,
    "batch_idx": 0,
    "loss": 0.2296975702047348,
    "learning_rate": 7.24535476932081e-05,
    "timestep_mean": 496.96875,
    "epsilon_norm": 8.104586601257324,
    "prediction_norm": 6.917591094970703,
    "timestamp": "2025-07-14T12:34:32.859134"
  },
  {
    "epoch": 1653,
    "batch_idx": 32,
    "loss": 0.22779083251953125,
    "learning_rate": 7.24535476932081e-05,
    "timestep_mean": 525.15625,
    "epsilon_norm": 7.220455646514893,
    "prediction_norm": 6.312407493591309,
    "timestamp": "2025-07-14T12:34:32.864159"
  },
  {
    "epoch": 1653,
    "batch_idx": 64,
    "loss": 0.18887796998023987,
    "learning_rate": 7.24535476932081e-05,
    "timestep_mean": 510.125,
    "epsilon_norm": 8.669732093811035,
    "prediction_norm": 7.328612804412842,
    "timestamp": "2025-07-14T12:34:32.869129"
  },
  {
    "epoch": 1653,
    "batch_idx": 96,
    "loss": 0.14098092913627625,
    "learning_rate": 7.24535476932081e-05,
    "timestep_mean": 550.4375,
    "epsilon_norm": 8.588837623596191,
    "prediction_norm": 8.019469261169434,
    "timestamp": "2025-07-14T12:34:32.874144"
  },
  {
    "epoch": 1653,
    "batch_idx": 128,
    "loss": 0.22072163224220276,
    "learning_rate": 7.24535476932081e-05,
    "timestep_mean": 565.34375,
    "epsilon_norm": 7.917092800140381,
    "prediction_norm": 7.135761260986328,
    "timestamp": "2025-07-14T12:34:32.879291"
  },
  {
    "epoch": 1653,
    "batch_idx": 160,
    "loss": 0.21746081113815308,
    "learning_rate": 7.24535476932081e-05,
    "timestep_mean": 481.21875,
    "epsilon_norm": 7.634548187255859,
    "prediction_norm": 7.1799235343933105,
    "timestamp": "2025-07-14T12:34:32.884211"
  },
  {
    "epoch": 1653,
    "batch_idx": 192,
    "loss": 0.35209858417510986,
    "learning_rate": 7.24535476932081e-05,
    "timestep_mean": 382.375,
    "epsilon_norm": 4.083579063415527,
    "prediction_norm": 3.486937999725342,
    "timestamp": "2025-07-14T12:34:32.889251"
  },
  {
    "epoch": 1654,
    "batch_idx": 0,
    "loss": 0.6976968050003052,
    "learning_rate": 7.204686616443349e-05,
    "timestep_mean": 390.9375,
    "epsilon_norm": 9.052505493164062,
    "prediction_norm": 6.183928966522217,
    "timestamp": "2025-07-14T12:34:32.894253"
  },
  {
    "epoch": 1654,
    "batch_idx": 32,
    "loss": 0.36564260721206665,
    "learning_rate": 7.204686616443349e-05,
    "timestep_mean": 464.40625,
    "epsilon_norm": 8.219178199768066,
    "prediction_norm": 6.956445217132568,
    "timestamp": "2025-07-14T12:34:32.899182"
  },
  {
    "epoch": 1654,
    "batch_idx": 64,
    "loss": 0.34344756603240967,
    "learning_rate": 7.204686616443349e-05,
    "timestep_mean": 456.25,
    "epsilon_norm": 8.62169075012207,
    "prediction_norm": 6.842269420623779,
    "timestamp": "2025-07-14T12:34:32.904112"
  },
  {
    "epoch": 1654,
    "batch_idx": 96,
    "loss": 0.2550155222415924,
    "learning_rate": 7.204686616443349e-05,
    "timestep_mean": 529.34375,
    "epsilon_norm": 8.986798286437988,
    "prediction_norm": 7.819922924041748,
    "timestamp": "2025-07-14T12:34:32.909102"
  },
  {
    "epoch": 1654,
    "batch_idx": 128,
    "loss": 0.7730040550231934,
    "learning_rate": 7.204686616443349e-05,
    "timestep_mean": 531.21875,
    "epsilon_norm": 9.841204643249512,
    "prediction_norm": 7.105103969573975,
    "timestamp": "2025-07-14T12:34:32.914026"
  },
  {
    "epoch": 1654,
    "batch_idx": 160,
    "loss": 0.290437251329422,
    "learning_rate": 7.204686616443349e-05,
    "timestep_mean": 466.65625,
    "epsilon_norm": 7.6044921875,
    "prediction_norm": 6.103777885437012,
    "timestamp": "2025-07-14T12:34:32.918947"
  },
  {
    "epoch": 1654,
    "batch_idx": 192,
    "loss": 0.23746377229690552,
    "learning_rate": 7.204686616443349e-05,
    "timestep_mean": 540.0,
    "epsilon_norm": 3.016455888748169,
    "prediction_norm": 1.734209418296814,
    "timestamp": "2025-07-14T12:34:32.923865"
  },
  {
    "epoch": 1655,
    "batch_idx": 0,
    "loss": 0.20210400223731995,
    "learning_rate": 7.164124056747522e-05,
    "timestep_mean": 525.1875,
    "epsilon_norm": 7.215697765350342,
    "prediction_norm": 6.212738513946533,
    "timestamp": "2025-07-14T12:34:32.928740"
  },
  {
    "epoch": 1655,
    "batch_idx": 32,
    "loss": 0.1497078835964203,
    "learning_rate": 7.164124056747522e-05,
    "timestep_mean": 575.03125,
    "epsilon_norm": 7.635174751281738,
    "prediction_norm": 7.337935924530029,
    "timestamp": "2025-07-14T12:34:32.933762"
  },
  {
    "epoch": 1655,
    "batch_idx": 64,
    "loss": 0.2702102065086365,
    "learning_rate": 7.164124056747522e-05,
    "timestep_mean": 485.65625,
    "epsilon_norm": 6.776998519897461,
    "prediction_norm": 5.320992469787598,
    "timestamp": "2025-07-14T12:34:32.938693"
  },
  {
    "epoch": 1655,
    "batch_idx": 96,
    "loss": 0.4584365785121918,
    "learning_rate": 7.164124056747522e-05,
    "timestep_mean": 452.40625,
    "epsilon_norm": 8.878410339355469,
    "prediction_norm": 7.322678565979004,
    "timestamp": "2025-07-14T12:34:32.943604"
  },
  {
    "epoch": 1655,
    "batch_idx": 128,
    "loss": 0.19520467519760132,
    "learning_rate": 7.164124056747522e-05,
    "timestep_mean": 551.0625,
    "epsilon_norm": 8.544991493225098,
    "prediction_norm": 7.678299427032471,
    "timestamp": "2025-07-14T12:34:32.948512"
  },
  {
    "epoch": 1655,
    "batch_idx": 160,
    "loss": 0.24391281604766846,
    "learning_rate": 7.164124056747522e-05,
    "timestep_mean": 511.6875,
    "epsilon_norm": 7.307965278625488,
    "prediction_norm": 6.467956066131592,
    "timestamp": "2025-07-14T12:34:32.953439"
  },
  {
    "epoch": 1655,
    "batch_idx": 192,
    "loss": 0.2723015546798706,
    "learning_rate": 7.164124056747522e-05,
    "timestep_mean": 379.125,
    "epsilon_norm": 4.52529239654541,
    "prediction_norm": 4.527650356292725,
    "timestamp": "2025-07-14T12:34:32.958437"
  },
  {
    "epoch": 1656,
    "batch_idx": 0,
    "loss": 0.3489305377006531,
    "learning_rate": 7.123667190317395e-05,
    "timestep_mean": 461.625,
    "epsilon_norm": 8.687994956970215,
    "prediction_norm": 7.281832695007324,
    "timestamp": "2025-07-14T12:34:32.963447"
  },
  {
    "epoch": 1656,
    "batch_idx": 32,
    "loss": 0.36289650201797485,
    "learning_rate": 7.123667190317395e-05,
    "timestep_mean": 553.75,
    "epsilon_norm": 9.339107513427734,
    "prediction_norm": 7.549360275268555,
    "timestamp": "2025-07-14T12:34:32.968352"
  },
  {
    "epoch": 1656,
    "batch_idx": 64,
    "loss": 0.21155434846878052,
    "learning_rate": 7.123667190317395e-05,
    "timestep_mean": 500.6875,
    "epsilon_norm": 7.922731876373291,
    "prediction_norm": 6.874508380889893,
    "timestamp": "2025-07-14T12:34:32.973265"
  },
  {
    "epoch": 1656,
    "batch_idx": 96,
    "loss": 0.25388580560684204,
    "learning_rate": 7.123667190317395e-05,
    "timestep_mean": 482.59375,
    "epsilon_norm": 8.987640380859375,
    "prediction_norm": 7.627365589141846,
    "timestamp": "2025-07-14T12:34:32.978209"
  },
  {
    "epoch": 1656,
    "batch_idx": 128,
    "loss": 0.1917102187871933,
    "learning_rate": 7.123667190317395e-05,
    "timestep_mean": 479.625,
    "epsilon_norm": 8.104584693908691,
    "prediction_norm": 7.089295387268066,
    "timestamp": "2025-07-14T12:34:32.983088"
  },
  {
    "epoch": 1656,
    "batch_idx": 160,
    "loss": 0.455785870552063,
    "learning_rate": 7.123667190317395e-05,
    "timestep_mean": 516.375,
    "epsilon_norm": 7.598635196685791,
    "prediction_norm": 5.7867608070373535,
    "timestamp": "2025-07-14T12:34:32.988060"
  },
  {
    "epoch": 1656,
    "batch_idx": 192,
    "loss": 0.4349512457847595,
    "learning_rate": 7.123667190317395e-05,
    "timestep_mean": 788.375,
    "epsilon_norm": 4.815364837646484,
    "prediction_norm": 4.245023727416992,
    "timestamp": "2025-07-14T12:34:32.993035"
  },
  {
    "epoch": 1657,
    "batch_idx": 0,
    "loss": 0.33399707078933716,
    "learning_rate": 7.083316116976263e-05,
    "timestep_mean": 490.1875,
    "epsilon_norm": 6.735256671905518,
    "prediction_norm": 5.569699287414551,
    "timestamp": "2025-07-14T12:34:32.998038"
  },
  {
    "epoch": 1657,
    "batch_idx": 32,
    "loss": 0.17510853707790375,
    "learning_rate": 7.083316116976263e-05,
    "timestep_mean": 522.75,
    "epsilon_norm": 8.554699897766113,
    "prediction_norm": 7.623482704162598,
    "timestamp": "2025-07-14T12:34:33.003313"
  },
  {
    "epoch": 1657,
    "batch_idx": 64,
    "loss": 0.29703325033187866,
    "learning_rate": 7.083316116976263e-05,
    "timestep_mean": 554.5,
    "epsilon_norm": 6.75319242477417,
    "prediction_norm": 5.047116279602051,
    "timestamp": "2025-07-14T12:34:33.008376"
  },
  {
    "epoch": 1657,
    "batch_idx": 96,
    "loss": 0.20047594606876373,
    "learning_rate": 7.083316116976263e-05,
    "timestep_mean": 518.1875,
    "epsilon_norm": 7.888321399688721,
    "prediction_norm": 7.099070072174072,
    "timestamp": "2025-07-14T12:34:33.013719"
  },
  {
    "epoch": 1657,
    "batch_idx": 128,
    "loss": 0.32993578910827637,
    "learning_rate": 7.083316116976263e-05,
    "timestep_mean": 460.71875,
    "epsilon_norm": 7.011643886566162,
    "prediction_norm": 5.209002494812012,
    "timestamp": "2025-07-14T12:34:33.018939"
  },
  {
    "epoch": 1657,
    "batch_idx": 160,
    "loss": 0.315837562084198,
    "learning_rate": 7.083316116976263e-05,
    "timestep_mean": 512.84375,
    "epsilon_norm": 8.036646842956543,
    "prediction_norm": 6.770134925842285,
    "timestamp": "2025-07-14T12:34:33.024031"
  },
  {
    "epoch": 1657,
    "batch_idx": 192,
    "loss": 0.4716566205024719,
    "learning_rate": 7.083316116976263e-05,
    "timestep_mean": 389.25,
    "epsilon_norm": 4.1576433181762695,
    "prediction_norm": 2.702803134918213,
    "timestamp": "2025-07-14T12:34:33.029364"
  },
  {
    "epoch": 1658,
    "batch_idx": 0,
    "loss": 0.30980318784713745,
    "learning_rate": 7.043070936286383e-05,
    "timestep_mean": 485.28125,
    "epsilon_norm": 8.687295913696289,
    "prediction_norm": 7.366894721984863,
    "timestamp": "2025-07-14T12:34:33.034645"
  },
  {
    "epoch": 1658,
    "batch_idx": 32,
    "loss": 0.25995075702667236,
    "learning_rate": 7.043070936286383e-05,
    "timestep_mean": 444.5,
    "epsilon_norm": 8.732340812683105,
    "prediction_norm": 7.055001258850098,
    "timestamp": "2025-07-14T12:34:33.039889"
  },
  {
    "epoch": 1658,
    "batch_idx": 64,
    "loss": 0.2927916347980499,
    "learning_rate": 7.043070936286383e-05,
    "timestep_mean": 467.09375,
    "epsilon_norm": 7.785238265991211,
    "prediction_norm": 6.272132396697998,
    "timestamp": "2025-07-14T12:34:33.045167"
  },
  {
    "epoch": 1658,
    "batch_idx": 96,
    "loss": 0.27845633029937744,
    "learning_rate": 7.043070936286383e-05,
    "timestep_mean": 497.4375,
    "epsilon_norm": 8.566781997680664,
    "prediction_norm": 7.366323471069336,
    "timestamp": "2025-07-14T12:34:33.050231"
  },
  {
    "epoch": 1658,
    "batch_idx": 128,
    "loss": 0.2200045883655548,
    "learning_rate": 7.043070936286383e-05,
    "timestep_mean": 426.875,
    "epsilon_norm": 7.230232238769531,
    "prediction_norm": 6.303529262542725,
    "timestamp": "2025-07-14T12:34:33.055417"
  },
  {
    "epoch": 1658,
    "batch_idx": 160,
    "loss": 0.3944471776485443,
    "learning_rate": 7.043070936286383e-05,
    "timestep_mean": 500.625,
    "epsilon_norm": 8.163946151733398,
    "prediction_norm": 6.473019123077393,
    "timestamp": "2025-07-14T12:34:33.060689"
  },
  {
    "epoch": 1658,
    "batch_idx": 192,
    "loss": 0.490680068731308,
    "learning_rate": 7.043070936286383e-05,
    "timestep_mean": 339.875,
    "epsilon_norm": 4.024555683135986,
    "prediction_norm": 2.789825201034546,
    "timestamp": "2025-07-14T12:34:33.065822"
  },
  {
    "epoch": 1659,
    "batch_idx": 0,
    "loss": 0.2225000113248825,
    "learning_rate": 7.002931747548751e-05,
    "timestep_mean": 548.9375,
    "epsilon_norm": 7.626125335693359,
    "prediction_norm": 6.659561634063721,
    "timestamp": "2025-07-14T12:34:33.071005"
  },
  {
    "epoch": 1659,
    "batch_idx": 32,
    "loss": 0.3182002305984497,
    "learning_rate": 7.002931747548751e-05,
    "timestep_mean": 446.71875,
    "epsilon_norm": 7.791040897369385,
    "prediction_norm": 6.02387809753418,
    "timestamp": "2025-07-14T12:34:33.076345"
  },
  {
    "epoch": 1659,
    "batch_idx": 64,
    "loss": 0.22023287415504456,
    "learning_rate": 7.002931747548751e-05,
    "timestep_mean": 480.34375,
    "epsilon_norm": 7.296463489532471,
    "prediction_norm": 6.565305709838867,
    "timestamp": "2025-07-14T12:34:33.082340"
  },
  {
    "epoch": 1659,
    "batch_idx": 96,
    "loss": 0.27640464901924133,
    "learning_rate": 7.002931747548751e-05,
    "timestep_mean": 494.59375,
    "epsilon_norm": 7.019195079803467,
    "prediction_norm": 5.55485725402832,
    "timestamp": "2025-07-14T12:34:33.087886"
  },
  {
    "epoch": 1659,
    "batch_idx": 128,
    "loss": 0.19759970903396606,
    "learning_rate": 7.002931747548751e-05,
    "timestep_mean": 573.28125,
    "epsilon_norm": 7.098705768585205,
    "prediction_norm": 6.619468688964844,
    "timestamp": "2025-07-14T12:34:33.097901"
  },
  {
    "epoch": 1659,
    "batch_idx": 160,
    "loss": 0.20365077257156372,
    "learning_rate": 7.002931747548751e-05,
    "timestep_mean": 511.8125,
    "epsilon_norm": 8.575565338134766,
    "prediction_norm": 8.020380973815918,
    "timestamp": "2025-07-14T12:34:33.103367"
  },
  {
    "epoch": 1659,
    "batch_idx": 192,
    "loss": 0.29732969403266907,
    "learning_rate": 7.002931747548751e-05,
    "timestep_mean": 626.75,
    "epsilon_norm": 3.7264280319213867,
    "prediction_norm": 2.882145881652832,
    "timestamp": "2025-07-14T12:34:33.108844"
  },
  {
    "epoch": 1660,
    "batch_idx": 0,
    "loss": 0.4741787314414978,
    "learning_rate": 6.962898649802822e-05,
    "timestep_mean": 512.96875,
    "epsilon_norm": 8.552268981933594,
    "prediction_norm": 7.020585060119629,
    "timestamp": "2025-07-14T12:34:33.114041"
  },
  {
    "epoch": 1660,
    "batch_idx": 32,
    "loss": 0.07773898541927338,
    "learning_rate": 6.962898649802822e-05,
    "timestep_mean": 652.5625,
    "epsilon_norm": 8.001300811767578,
    "prediction_norm": 7.645890712738037,
    "timestamp": "2025-07-14T12:34:33.118977"
  },
  {
    "epoch": 1660,
    "batch_idx": 64,
    "loss": 0.2640175223350525,
    "learning_rate": 6.962898649802822e-05,
    "timestep_mean": 556.21875,
    "epsilon_norm": 8.34060001373291,
    "prediction_norm": 7.3806047439575195,
    "timestamp": "2025-07-14T12:34:33.124088"
  },
  {
    "epoch": 1660,
    "batch_idx": 96,
    "loss": 0.33060646057128906,
    "learning_rate": 6.962898649802822e-05,
    "timestep_mean": 481.125,
    "epsilon_norm": 7.505343437194824,
    "prediction_norm": 6.377002716064453,
    "timestamp": "2025-07-14T12:34:33.128936"
  },
  {
    "epoch": 1660,
    "batch_idx": 128,
    "loss": 0.21012499928474426,
    "learning_rate": 6.962898649802822e-05,
    "timestep_mean": 386.40625,
    "epsilon_norm": 8.302959442138672,
    "prediction_norm": 7.382440567016602,
    "timestamp": "2025-07-14T12:34:33.133876"
  },
  {
    "epoch": 1660,
    "batch_idx": 160,
    "loss": 0.270301878452301,
    "learning_rate": 6.962898649802822e-05,
    "timestep_mean": 379.8125,
    "epsilon_norm": 8.122049331665039,
    "prediction_norm": 7.105417251586914,
    "timestamp": "2025-07-14T12:34:33.138815"
  },
  {
    "epoch": 1660,
    "batch_idx": 192,
    "loss": 0.2623966932296753,
    "learning_rate": 6.962898649802822e-05,
    "timestep_mean": 462.25,
    "epsilon_norm": 3.9835281372070312,
    "prediction_norm": 3.516313076019287,
    "timestamp": "2025-07-14T12:34:33.143799"
  },
  {
    "epoch": 1661,
    "batch_idx": 0,
    "loss": 0.17642809450626373,
    "learning_rate": 6.92297174182627e-05,
    "timestep_mean": 502.0,
    "epsilon_norm": 8.272126197814941,
    "prediction_norm": 7.343151092529297,
    "timestamp": "2025-07-14T12:34:33.149453"
  },
  {
    "epoch": 1661,
    "batch_idx": 32,
    "loss": 0.18942560255527496,
    "learning_rate": 6.92297174182627e-05,
    "timestep_mean": 561.0,
    "epsilon_norm": 7.80119514465332,
    "prediction_norm": 7.222166538238525,
    "timestamp": "2025-07-14T12:34:33.154730"
  },
  {
    "epoch": 1661,
    "batch_idx": 64,
    "loss": 0.22449399530887604,
    "learning_rate": 6.92297174182627e-05,
    "timestep_mean": 511.78125,
    "epsilon_norm": 8.613892555236816,
    "prediction_norm": 7.125110626220703,
    "timestamp": "2025-07-14T12:34:33.160479"
  },
  {
    "epoch": 1661,
    "batch_idx": 96,
    "loss": 0.3323739767074585,
    "learning_rate": 6.92297174182627e-05,
    "timestep_mean": 504.125,
    "epsilon_norm": 9.179266929626465,
    "prediction_norm": 8.637807846069336,
    "timestamp": "2025-07-14T12:34:33.165691"
  },
  {
    "epoch": 1661,
    "batch_idx": 128,
    "loss": 0.3517514169216156,
    "learning_rate": 6.92297174182627e-05,
    "timestep_mean": 467.25,
    "epsilon_norm": 9.220273971557617,
    "prediction_norm": 7.704793453216553,
    "timestamp": "2025-07-14T12:34:33.170651"
  },
  {
    "epoch": 1661,
    "batch_idx": 160,
    "loss": 0.2932528257369995,
    "learning_rate": 6.92297174182627e-05,
    "timestep_mean": 467.75,
    "epsilon_norm": 7.785913467407227,
    "prediction_norm": 6.8370041847229,
    "timestamp": "2025-07-14T12:34:33.175690"
  },
  {
    "epoch": 1661,
    "batch_idx": 192,
    "loss": 0.735753059387207,
    "learning_rate": 6.92297174182627e-05,
    "timestep_mean": 456.375,
    "epsilon_norm": 4.47534704208374,
    "prediction_norm": 3.334192991256714,
    "timestamp": "2025-07-14T12:34:33.180809"
  },
  {
    "epoch": 1662,
    "batch_idx": 0,
    "loss": 0.22916004061698914,
    "learning_rate": 6.88315112213481e-05,
    "timestep_mean": 509.75,
    "epsilon_norm": 8.250140190124512,
    "prediction_norm": 7.202491283416748,
    "timestamp": "2025-07-14T12:34:33.185791"
  },
  {
    "epoch": 1662,
    "batch_idx": 32,
    "loss": 0.347622811794281,
    "learning_rate": 6.88315112213481e-05,
    "timestep_mean": 511.65625,
    "epsilon_norm": 8.578267097473145,
    "prediction_norm": 7.0416412353515625,
    "timestamp": "2025-07-14T12:34:33.190902"
  },
  {
    "epoch": 1662,
    "batch_idx": 64,
    "loss": 0.2160864621400833,
    "learning_rate": 6.88315112213481e-05,
    "timestep_mean": 508.40625,
    "epsilon_norm": 7.042394161224365,
    "prediction_norm": 6.219626426696777,
    "timestamp": "2025-07-14T12:34:33.196020"
  },
  {
    "epoch": 1662,
    "batch_idx": 96,
    "loss": 0.429385781288147,
    "learning_rate": 6.88315112213481e-05,
    "timestep_mean": 451.21875,
    "epsilon_norm": 7.290452480316162,
    "prediction_norm": 5.202734470367432,
    "timestamp": "2025-07-14T12:34:33.201055"
  },
  {
    "epoch": 1662,
    "batch_idx": 128,
    "loss": 0.27780258655548096,
    "learning_rate": 6.88315112213481e-05,
    "timestep_mean": 469.5,
    "epsilon_norm": 7.454246520996094,
    "prediction_norm": 6.078510284423828,
    "timestamp": "2025-07-14T12:34:33.205999"
  },
  {
    "epoch": 1662,
    "batch_idx": 160,
    "loss": 0.22387966513633728,
    "learning_rate": 6.88315112213481e-05,
    "timestep_mean": 551.15625,
    "epsilon_norm": 7.57030725479126,
    "prediction_norm": 7.235973834991455,
    "timestamp": "2025-07-14T12:34:33.210895"
  },
  {
    "epoch": 1662,
    "batch_idx": 192,
    "loss": 0.14372678101062775,
    "learning_rate": 6.88315112213481e-05,
    "timestep_mean": 505.5,
    "epsilon_norm": 4.649659633636475,
    "prediction_norm": 4.491508483886719,
    "timestamp": "2025-07-14T12:34:33.215940"
  },
  {
    "epoch": 1663,
    "batch_idx": 0,
    "loss": 0.3701348304748535,
    "learning_rate": 6.843436888981811e-05,
    "timestep_mean": 355.0625,
    "epsilon_norm": 8.519542694091797,
    "prediction_norm": 6.862849712371826,
    "timestamp": "2025-07-14T12:34:33.221218"
  },
  {
    "epoch": 1663,
    "batch_idx": 32,
    "loss": 0.1316489577293396,
    "learning_rate": 6.843436888981811e-05,
    "timestep_mean": 536.46875,
    "epsilon_norm": 7.624098777770996,
    "prediction_norm": 6.995503902435303,
    "timestamp": "2025-07-14T12:34:33.226243"
  },
  {
    "epoch": 1663,
    "batch_idx": 64,
    "loss": 0.28253674507141113,
    "learning_rate": 6.843436888981811e-05,
    "timestep_mean": 516.4375,
    "epsilon_norm": 8.033958435058594,
    "prediction_norm": 6.593962669372559,
    "timestamp": "2025-07-14T12:34:33.231325"
  },
  {
    "epoch": 1663,
    "batch_idx": 96,
    "loss": 0.2671816051006317,
    "learning_rate": 6.843436888981811e-05,
    "timestep_mean": 470.15625,
    "epsilon_norm": 7.98897123336792,
    "prediction_norm": 6.498297691345215,
    "timestamp": "2025-07-14T12:34:33.236429"
  },
  {
    "epoch": 1663,
    "batch_idx": 128,
    "loss": 0.3353668451309204,
    "learning_rate": 6.843436888981811e-05,
    "timestep_mean": 473.09375,
    "epsilon_norm": 7.041045188903809,
    "prediction_norm": 5.931090354919434,
    "timestamp": "2025-07-14T12:34:33.241373"
  },
  {
    "epoch": 1663,
    "batch_idx": 160,
    "loss": 0.2867951989173889,
    "learning_rate": 6.843436888981811e-05,
    "timestep_mean": 563.375,
    "epsilon_norm": 6.559449195861816,
    "prediction_norm": 5.642814636230469,
    "timestamp": "2025-07-14T12:34:33.246674"
  },
  {
    "epoch": 1663,
    "batch_idx": 192,
    "loss": 0.3850111663341522,
    "learning_rate": 6.843436888981811e-05,
    "timestep_mean": 414.125,
    "epsilon_norm": 4.393068313598633,
    "prediction_norm": 3.5628392696380615,
    "timestamp": "2025-07-14T12:34:33.251857"
  },
  {
    "epoch": 1664,
    "batch_idx": 0,
    "loss": 0.16978704929351807,
    "learning_rate": 6.803829140358223e-05,
    "timestep_mean": 617.6875,
    "epsilon_norm": 7.6171674728393555,
    "prediction_norm": 7.200351238250732,
    "timestamp": "2025-07-14T12:34:33.256941"
  },
  {
    "epoch": 1664,
    "batch_idx": 32,
    "loss": 0.20983976125717163,
    "learning_rate": 6.803829140358223e-05,
    "timestep_mean": 448.5,
    "epsilon_norm": 7.644336223602295,
    "prediction_norm": 6.434728145599365,
    "timestamp": "2025-07-14T12:34:33.262033"
  },
  {
    "epoch": 1664,
    "batch_idx": 64,
    "loss": 0.42785510420799255,
    "learning_rate": 6.803829140358223e-05,
    "timestep_mean": 443.78125,
    "epsilon_norm": 7.159147262573242,
    "prediction_norm": 5.2526021003723145,
    "timestamp": "2025-07-14T12:34:33.267079"
  },
  {
    "epoch": 1664,
    "batch_idx": 96,
    "loss": 0.4235926866531372,
    "learning_rate": 6.803829140358223e-05,
    "timestep_mean": 500.375,
    "epsilon_norm": 8.797567367553711,
    "prediction_norm": 7.314146041870117,
    "timestamp": "2025-07-14T12:34:33.272048"
  },
  {
    "epoch": 1664,
    "batch_idx": 128,
    "loss": 0.2069069743156433,
    "learning_rate": 6.803829140358223e-05,
    "timestep_mean": 541.3125,
    "epsilon_norm": 8.181514739990234,
    "prediction_norm": 7.073099613189697,
    "timestamp": "2025-07-14T12:34:33.277117"
  },
  {
    "epoch": 1664,
    "batch_idx": 160,
    "loss": 0.36966729164123535,
    "learning_rate": 6.803829140358223e-05,
    "timestep_mean": 447.84375,
    "epsilon_norm": 7.913815975189209,
    "prediction_norm": 6.662118434906006,
    "timestamp": "2025-07-14T12:34:33.282085"
  },
  {
    "epoch": 1664,
    "batch_idx": 192,
    "loss": 0.5221202373504639,
    "learning_rate": 6.803829140358223e-05,
    "timestep_mean": 459.625,
    "epsilon_norm": 4.896331310272217,
    "prediction_norm": 3.7055044174194336,
    "timestamp": "2025-07-14T12:34:33.287085"
  },
  {
    "epoch": 1665,
    "batch_idx": 0,
    "loss": 0.4424750804901123,
    "learning_rate": 6.764327973992249e-05,
    "timestep_mean": 441.65625,
    "epsilon_norm": 8.791583061218262,
    "prediction_norm": 6.88145637512207,
    "timestamp": "2025-07-14T12:34:33.292074"
  },
  {
    "epoch": 1665,
    "batch_idx": 32,
    "loss": 0.30786317586898804,
    "learning_rate": 6.764327973992249e-05,
    "timestep_mean": 528.4375,
    "epsilon_norm": 7.5073466300964355,
    "prediction_norm": 6.1474385261535645,
    "timestamp": "2025-07-14T12:34:33.297097"
  },
  {
    "epoch": 1665,
    "batch_idx": 64,
    "loss": 0.25512373447418213,
    "learning_rate": 6.764327973992249e-05,
    "timestep_mean": 490.8125,
    "epsilon_norm": 7.958560943603516,
    "prediction_norm": 6.697247505187988,
    "timestamp": "2025-07-14T12:34:33.302122"
  },
  {
    "epoch": 1665,
    "batch_idx": 96,
    "loss": 0.20626220107078552,
    "learning_rate": 6.764327973992249e-05,
    "timestep_mean": 532.1875,
    "epsilon_norm": 7.743746757507324,
    "prediction_norm": 6.520779132843018,
    "timestamp": "2025-07-14T12:34:33.307143"
  },
  {
    "epoch": 1665,
    "batch_idx": 128,
    "loss": 0.361841082572937,
    "learning_rate": 6.764327973992249e-05,
    "timestep_mean": 497.125,
    "epsilon_norm": 8.534421920776367,
    "prediction_norm": 6.553415298461914,
    "timestamp": "2025-07-14T12:34:33.312124"
  },
  {
    "epoch": 1665,
    "batch_idx": 160,
    "loss": 0.18576455116271973,
    "learning_rate": 6.764327973992249e-05,
    "timestep_mean": 494.6875,
    "epsilon_norm": 7.551653861999512,
    "prediction_norm": 7.0463547706604,
    "timestamp": "2025-07-14T12:34:33.317112"
  },
  {
    "epoch": 1665,
    "batch_idx": 192,
    "loss": 0.09552126377820969,
    "learning_rate": 6.764327973992249e-05,
    "timestep_mean": 637.375,
    "epsilon_norm": 3.9899179935455322,
    "prediction_norm": 3.4270126819610596,
    "timestamp": "2025-07-14T12:34:33.322169"
  },
  {
    "epoch": 1666,
    "batch_idx": 0,
    "loss": 0.31561610102653503,
    "learning_rate": 6.724933487349048e-05,
    "timestep_mean": 507.0,
    "epsilon_norm": 7.381740570068359,
    "prediction_norm": 6.220977783203125,
    "timestamp": "2025-07-14T12:34:33.327209"
  },
  {
    "epoch": 1666,
    "batch_idx": 32,
    "loss": 0.23739080131053925,
    "learning_rate": 6.724933487349048e-05,
    "timestep_mean": 553.90625,
    "epsilon_norm": 7.4254231452941895,
    "prediction_norm": 6.45342493057251,
    "timestamp": "2025-07-14T12:34:33.332186"
  },
  {
    "epoch": 1666,
    "batch_idx": 64,
    "loss": 0.41705796122550964,
    "learning_rate": 6.724933487349048e-05,
    "timestep_mean": 524.46875,
    "epsilon_norm": 9.598798751831055,
    "prediction_norm": 7.551205635070801,
    "timestamp": "2025-07-14T12:34:33.338255"
  },
  {
    "epoch": 1666,
    "batch_idx": 96,
    "loss": 0.12960536777973175,
    "learning_rate": 6.724933487349048e-05,
    "timestep_mean": 498.96875,
    "epsilon_norm": 7.339227676391602,
    "prediction_norm": 6.815333366394043,
    "timestamp": "2025-07-14T12:34:33.343656"
  },
  {
    "epoch": 1666,
    "batch_idx": 128,
    "loss": 0.3160291314125061,
    "learning_rate": 6.724933487349048e-05,
    "timestep_mean": 549.625,
    "epsilon_norm": 8.101208686828613,
    "prediction_norm": 6.356828689575195,
    "timestamp": "2025-07-14T12:34:33.354817"
  },
  {
    "epoch": 1666,
    "batch_idx": 160,
    "loss": 0.1623614877462387,
    "learning_rate": 6.724933487349048e-05,
    "timestep_mean": 517.4375,
    "epsilon_norm": 7.3848042488098145,
    "prediction_norm": 6.557524681091309,
    "timestamp": "2025-07-14T12:34:33.359998"
  },
  {
    "epoch": 1666,
    "batch_idx": 192,
    "loss": 0.15804988145828247,
    "learning_rate": 6.724933487349048e-05,
    "timestep_mean": 474.0,
    "epsilon_norm": 3.667950391769409,
    "prediction_norm": 3.39556622505188,
    "timestamp": "2025-07-14T12:34:33.364978"
  },
  {
    "epoch": 1667,
    "batch_idx": 0,
    "loss": 0.21074873208999634,
    "learning_rate": 6.685645777630638e-05,
    "timestep_mean": 493.125,
    "epsilon_norm": 8.613173484802246,
    "prediction_norm": 7.433160305023193,
    "timestamp": "2025-07-14T12:34:33.369910"
  },
  {
    "epoch": 1667,
    "batch_idx": 32,
    "loss": 0.17901210486888885,
    "learning_rate": 6.685645777630638e-05,
    "timestep_mean": 551.90625,
    "epsilon_norm": 8.567002296447754,
    "prediction_norm": 7.36835241317749,
    "timestamp": "2025-07-14T12:34:33.374941"
  },
  {
    "epoch": 1667,
    "batch_idx": 64,
    "loss": 0.23143887519836426,
    "learning_rate": 6.685645777630638e-05,
    "timestep_mean": 471.40625,
    "epsilon_norm": 8.99086856842041,
    "prediction_norm": 7.679495811462402,
    "timestamp": "2025-07-14T12:34:33.379952"
  },
  {
    "epoch": 1667,
    "batch_idx": 96,
    "loss": 0.16420702636241913,
    "learning_rate": 6.685645777630638e-05,
    "timestep_mean": 515.84375,
    "epsilon_norm": 7.787734031677246,
    "prediction_norm": 7.20841646194458,
    "timestamp": "2025-07-14T12:34:33.384871"
  },
  {
    "epoch": 1667,
    "batch_idx": 128,
    "loss": 0.20348551869392395,
    "learning_rate": 6.685645777630638e-05,
    "timestep_mean": 516.5625,
    "epsilon_norm": 7.625165939331055,
    "prediction_norm": 6.483119964599609,
    "timestamp": "2025-07-14T12:34:33.389856"
  },
  {
    "epoch": 1667,
    "batch_idx": 160,
    "loss": 0.2161996066570282,
    "learning_rate": 6.685645777630638e-05,
    "timestep_mean": 536.65625,
    "epsilon_norm": 8.794132232666016,
    "prediction_norm": 7.26490592956543,
    "timestamp": "2025-07-14T12:34:33.394649"
  },
  {
    "epoch": 1667,
    "batch_idx": 192,
    "loss": 0.5848981142044067,
    "learning_rate": 6.685645777630638e-05,
    "timestep_mean": 281.875,
    "epsilon_norm": 4.3479814529418945,
    "prediction_norm": 3.1192545890808105,
    "timestamp": "2025-07-14T12:34:33.399607"
  },
  {
    "epoch": 1668,
    "batch_idx": 0,
    "loss": 0.5366053581237793,
    "learning_rate": 6.646464941775497e-05,
    "timestep_mean": 424.21875,
    "epsilon_norm": 8.759283065795898,
    "prediction_norm": 6.525879383087158,
    "timestamp": "2025-07-14T12:34:33.404524"
  },
  {
    "epoch": 1668,
    "batch_idx": 32,
    "loss": 0.5443058013916016,
    "learning_rate": 6.646464941775497e-05,
    "timestep_mean": 439.5625,
    "epsilon_norm": 8.694753646850586,
    "prediction_norm": 6.636408805847168,
    "timestamp": "2025-07-14T12:34:33.409406"
  },
  {
    "epoch": 1668,
    "batch_idx": 64,
    "loss": 0.5009576082229614,
    "learning_rate": 6.646464941775497e-05,
    "timestep_mean": 552.65625,
    "epsilon_norm": 8.833430290222168,
    "prediction_norm": 6.602755546569824,
    "timestamp": "2025-07-14T12:34:33.414608"
  },
  {
    "epoch": 1668,
    "batch_idx": 96,
    "loss": 0.20176178216934204,
    "learning_rate": 6.646464941775497e-05,
    "timestep_mean": 521.96875,
    "epsilon_norm": 7.917933940887451,
    "prediction_norm": 6.935471057891846,
    "timestamp": "2025-07-14T12:34:33.419732"
  },
  {
    "epoch": 1668,
    "batch_idx": 128,
    "loss": 0.24899627268314362,
    "learning_rate": 6.646464941775497e-05,
    "timestep_mean": 576.0,
    "epsilon_norm": 7.3305559158325195,
    "prediction_norm": 6.21090841293335,
    "timestamp": "2025-07-14T12:34:33.424874"
  },
  {
    "epoch": 1668,
    "batch_idx": 160,
    "loss": 0.27967798709869385,
    "learning_rate": 6.646464941775497e-05,
    "timestep_mean": 529.375,
    "epsilon_norm": 8.666938781738281,
    "prediction_norm": 7.215117931365967,
    "timestamp": "2025-07-14T12:34:33.429963"
  },
  {
    "epoch": 1668,
    "batch_idx": 192,
    "loss": 0.39059996604919434,
    "learning_rate": 6.646464941775497e-05,
    "timestep_mean": 412.25,
    "epsilon_norm": 5.0537872314453125,
    "prediction_norm": 3.581763505935669,
    "timestamp": "2025-07-14T12:34:33.435139"
  },
  {
    "epoch": 1669,
    "batch_idx": 0,
    "loss": 0.11470082402229309,
    "learning_rate": 6.607391076458467e-05,
    "timestep_mean": 457.15625,
    "epsilon_norm": 7.185370922088623,
    "prediction_norm": 6.496506690979004,
    "timestamp": "2025-07-14T12:34:33.440193"
  },
  {
    "epoch": 1669,
    "batch_idx": 32,
    "loss": 0.3391004800796509,
    "learning_rate": 6.607391076458467e-05,
    "timestep_mean": 505.75,
    "epsilon_norm": 7.916075706481934,
    "prediction_norm": 6.125274181365967,
    "timestamp": "2025-07-14T12:34:33.445115"
  },
  {
    "epoch": 1669,
    "batch_idx": 64,
    "loss": 0.3820277452468872,
    "learning_rate": 6.607391076458467e-05,
    "timestep_mean": 530.0,
    "epsilon_norm": 7.915197849273682,
    "prediction_norm": 5.310861587524414,
    "timestamp": "2025-07-14T12:34:33.450215"
  },
  {
    "epoch": 1669,
    "batch_idx": 96,
    "loss": 0.2722623348236084,
    "learning_rate": 6.607391076458467e-05,
    "timestep_mean": 471.96875,
    "epsilon_norm": 8.424308776855469,
    "prediction_norm": 6.880939483642578,
    "timestamp": "2025-07-14T12:34:33.455272"
  },
  {
    "epoch": 1669,
    "batch_idx": 128,
    "loss": 0.18073420226573944,
    "learning_rate": 6.607391076458467e-05,
    "timestep_mean": 560.9375,
    "epsilon_norm": 7.072003364562988,
    "prediction_norm": 6.192349433898926,
    "timestamp": "2025-07-14T12:34:33.460163"
  },
  {
    "epoch": 1669,
    "batch_idx": 160,
    "loss": 0.5075676441192627,
    "learning_rate": 6.607391076458467e-05,
    "timestep_mean": 394.75,
    "epsilon_norm": 8.38913345336914,
    "prediction_norm": 5.869660377502441,
    "timestamp": "2025-07-14T12:34:33.465202"
  },
  {
    "epoch": 1669,
    "batch_idx": 192,
    "loss": 0.31872493028640747,
    "learning_rate": 6.607391076458467e-05,
    "timestep_mean": 484.875,
    "epsilon_norm": 4.113202095031738,
    "prediction_norm": 2.9467649459838867,
    "timestamp": "2025-07-14T12:34:33.470202"
  },
  {
    "epoch": 1670,
    "batch_idx": 0,
    "loss": 0.39473757147789,
    "learning_rate": 6.568424278090443e-05,
    "timestep_mean": 452.4375,
    "epsilon_norm": 7.414050102233887,
    "prediction_norm": 5.838531494140625,
    "timestamp": "2025-07-14T12:34:33.475169"
  },
  {
    "epoch": 1670,
    "batch_idx": 32,
    "loss": 0.6318193674087524,
    "learning_rate": 6.568424278090443e-05,
    "timestep_mean": 387.28125,
    "epsilon_norm": 8.87647533416748,
    "prediction_norm": 5.949347972869873,
    "timestamp": "2025-07-14T12:34:33.480186"
  },
  {
    "epoch": 1670,
    "batch_idx": 64,
    "loss": 0.2404700517654419,
    "learning_rate": 6.568424278090443e-05,
    "timestep_mean": 500.0,
    "epsilon_norm": 7.280755043029785,
    "prediction_norm": 6.392853736877441,
    "timestamp": "2025-07-14T12:34:33.485110"
  },
  {
    "epoch": 1670,
    "batch_idx": 96,
    "loss": 0.23974454402923584,
    "learning_rate": 6.568424278090443e-05,
    "timestep_mean": 434.9375,
    "epsilon_norm": 6.598456859588623,
    "prediction_norm": 6.3269124031066895,
    "timestamp": "2025-07-14T12:34:33.490147"
  },
  {
    "epoch": 1670,
    "batch_idx": 128,
    "loss": 0.30675679445266724,
    "learning_rate": 6.568424278090443e-05,
    "timestep_mean": 517.15625,
    "epsilon_norm": 7.572808742523193,
    "prediction_norm": 5.94122314453125,
    "timestamp": "2025-07-14T12:34:33.495133"
  },
  {
    "epoch": 1670,
    "batch_idx": 160,
    "loss": 0.10374721884727478,
    "learning_rate": 6.568424278090443e-05,
    "timestep_mean": 550.84375,
    "epsilon_norm": 9.747730255126953,
    "prediction_norm": 8.605731010437012,
    "timestamp": "2025-07-14T12:34:33.500178"
  },
  {
    "epoch": 1670,
    "batch_idx": 192,
    "loss": 0.009895524010062218,
    "learning_rate": 6.568424278090443e-05,
    "timestep_mean": 567.875,
    "epsilon_norm": 3.2373485565185547,
    "prediction_norm": 3.2881526947021484,
    "timestamp": "2025-07-14T12:34:33.505350"
  },
  {
    "epoch": 1671,
    "batch_idx": 0,
    "loss": 0.23164057731628418,
    "learning_rate": 6.529564642818084e-05,
    "timestep_mean": 630.03125,
    "epsilon_norm": 6.8020853996276855,
    "prediction_norm": 5.977475166320801,
    "timestamp": "2025-07-14T12:34:33.510441"
  },
  {
    "epoch": 1671,
    "batch_idx": 32,
    "loss": 0.2214551866054535,
    "learning_rate": 6.529564642818084e-05,
    "timestep_mean": 580.0625,
    "epsilon_norm": 8.794480323791504,
    "prediction_norm": 7.9127607345581055,
    "timestamp": "2025-07-14T12:34:33.515476"
  },
  {
    "epoch": 1671,
    "batch_idx": 64,
    "loss": 0.19638046622276306,
    "learning_rate": 6.529564642818084e-05,
    "timestep_mean": 492.71875,
    "epsilon_norm": 8.100608825683594,
    "prediction_norm": 6.692943572998047,
    "timestamp": "2025-07-14T12:34:33.520483"
  },
  {
    "epoch": 1671,
    "batch_idx": 96,
    "loss": 0.1283644735813141,
    "learning_rate": 6.529564642818084e-05,
    "timestep_mean": 571.0625,
    "epsilon_norm": 7.120407581329346,
    "prediction_norm": 6.499295234680176,
    "timestamp": "2025-07-14T12:34:33.526606"
  },
  {
    "epoch": 1671,
    "batch_idx": 128,
    "loss": 0.42098182439804077,
    "learning_rate": 6.529564642818084e-05,
    "timestep_mean": 427.15625,
    "epsilon_norm": 7.698545932769775,
    "prediction_norm": 5.612706184387207,
    "timestamp": "2025-07-14T12:34:33.537362"
  },
  {
    "epoch": 1671,
    "batch_idx": 160,
    "loss": 0.1230861246585846,
    "learning_rate": 6.529564642818084e-05,
    "timestep_mean": 518.34375,
    "epsilon_norm": 8.138400077819824,
    "prediction_norm": 7.793831825256348,
    "timestamp": "2025-07-14T12:34:33.542609"
  },
  {
    "epoch": 1671,
    "batch_idx": 192,
    "loss": 0.42056041955947876,
    "learning_rate": 6.529564642818084e-05,
    "timestep_mean": 475.625,
    "epsilon_norm": 5.417728900909424,
    "prediction_norm": 4.680122375488281,
    "timestamp": "2025-07-14T12:34:33.547776"
  },
  {
    "epoch": 1672,
    "batch_idx": 0,
    "loss": 0.31549930572509766,
    "learning_rate": 6.490812266523714e-05,
    "timestep_mean": 435.875,
    "epsilon_norm": 8.570695877075195,
    "prediction_norm": 7.104854106903076,
    "timestamp": "2025-07-14T12:34:33.552834"
  },
  {
    "epoch": 1672,
    "batch_idx": 32,
    "loss": 0.3154332637786865,
    "learning_rate": 6.490812266523714e-05,
    "timestep_mean": 482.46875,
    "epsilon_norm": 8.734322547912598,
    "prediction_norm": 7.067333698272705,
    "timestamp": "2025-07-14T12:34:33.557809"
  },
  {
    "epoch": 1672,
    "batch_idx": 64,
    "loss": 0.17761287093162537,
    "learning_rate": 6.490812266523714e-05,
    "timestep_mean": 616.125,
    "epsilon_norm": 7.355989933013916,
    "prediction_norm": 6.792176723480225,
    "timestamp": "2025-07-14T12:34:33.562943"
  },
  {
    "epoch": 1672,
    "batch_idx": 96,
    "loss": 0.23999324440956116,
    "learning_rate": 6.490812266523714e-05,
    "timestep_mean": 454.75,
    "epsilon_norm": 8.452045440673828,
    "prediction_norm": 7.586987018585205,
    "timestamp": "2025-07-14T12:34:33.567825"
  },
  {
    "epoch": 1672,
    "batch_idx": 128,
    "loss": 0.2933835983276367,
    "learning_rate": 6.490812266523714e-05,
    "timestep_mean": 406.75,
    "epsilon_norm": 7.064996719360352,
    "prediction_norm": 5.692018985748291,
    "timestamp": "2025-07-14T12:34:33.572716"
  },
  {
    "epoch": 1672,
    "batch_idx": 160,
    "loss": 0.29215532541275024,
    "learning_rate": 6.490812266523714e-05,
    "timestep_mean": 382.375,
    "epsilon_norm": 7.280239105224609,
    "prediction_norm": 5.910948276519775,
    "timestamp": "2025-07-14T12:34:33.577710"
  },
  {
    "epoch": 1672,
    "batch_idx": 192,
    "loss": 0.2119937241077423,
    "learning_rate": 6.490812266523714e-05,
    "timestep_mean": 407.0,
    "epsilon_norm": 3.2105154991149902,
    "prediction_norm": 3.3127236366271973,
    "timestamp": "2025-07-14T12:34:33.582681"
  },
  {
    "epoch": 1673,
    "batch_idx": 0,
    "loss": 0.1497526466846466,
    "learning_rate": 6.452167244824956e-05,
    "timestep_mean": 492.46875,
    "epsilon_norm": 7.872888565063477,
    "prediction_norm": 7.26810884475708,
    "timestamp": "2025-07-14T12:34:33.587716"
  },
  {
    "epoch": 1673,
    "batch_idx": 32,
    "loss": 0.07148368656635284,
    "learning_rate": 6.452167244824956e-05,
    "timestep_mean": 625.375,
    "epsilon_norm": 6.2568135261535645,
    "prediction_norm": 6.578382968902588,
    "timestamp": "2025-07-14T12:34:33.592642"
  },
  {
    "epoch": 1673,
    "batch_idx": 64,
    "loss": 0.17637194693088531,
    "learning_rate": 6.452167244824956e-05,
    "timestep_mean": 512.09375,
    "epsilon_norm": 7.853590488433838,
    "prediction_norm": 7.160550117492676,
    "timestamp": "2025-07-14T12:34:33.597591"
  },
  {
    "epoch": 1673,
    "batch_idx": 96,
    "loss": 0.26795658469200134,
    "learning_rate": 6.452167244824956e-05,
    "timestep_mean": 540.9375,
    "epsilon_norm": 8.773777961730957,
    "prediction_norm": 7.7940168380737305,
    "timestamp": "2025-07-14T12:34:33.602515"
  },
  {
    "epoch": 1673,
    "batch_idx": 128,
    "loss": 0.4095730781555176,
    "learning_rate": 6.452167244824956e-05,
    "timestep_mean": 508.5625,
    "epsilon_norm": 8.26508903503418,
    "prediction_norm": 6.40529727935791,
    "timestamp": "2025-07-14T12:34:33.607458"
  },
  {
    "epoch": 1673,
    "batch_idx": 160,
    "loss": 0.3069458603858948,
    "learning_rate": 6.452167244824956e-05,
    "timestep_mean": 418.8125,
    "epsilon_norm": 8.358034133911133,
    "prediction_norm": 6.724406719207764,
    "timestamp": "2025-07-14T12:34:33.612701"
  },
  {
    "epoch": 1673,
    "batch_idx": 192,
    "loss": 0.08446142077445984,
    "learning_rate": 6.452167244824956e-05,
    "timestep_mean": 482.875,
    "epsilon_norm": 3.479565143585205,
    "prediction_norm": 3.008902072906494,
    "timestamp": "2025-07-14T12:34:33.617731"
  },
  {
    "epoch": 1674,
    "batch_idx": 0,
    "loss": 0.3192582130432129,
    "learning_rate": 6.41362967307456e-05,
    "timestep_mean": 491.15625,
    "epsilon_norm": 7.2069411277771,
    "prediction_norm": 5.582055568695068,
    "timestamp": "2025-07-14T12:34:33.622694"
  },
  {
    "epoch": 1674,
    "batch_idx": 32,
    "loss": 0.25495821237564087,
    "learning_rate": 6.41362967307456e-05,
    "timestep_mean": 529.09375,
    "epsilon_norm": 8.332474708557129,
    "prediction_norm": 6.921219348907471,
    "timestamp": "2025-07-14T12:34:33.627665"
  },
  {
    "epoch": 1674,
    "batch_idx": 64,
    "loss": 0.21989859640598297,
    "learning_rate": 6.41362967307456e-05,
    "timestep_mean": 501.8125,
    "epsilon_norm": 7.41157341003418,
    "prediction_norm": 6.162525177001953,
    "timestamp": "2025-07-14T12:34:33.632522"
  },
  {
    "epoch": 1674,
    "batch_idx": 96,
    "loss": 0.3214491307735443,
    "learning_rate": 6.41362967307456e-05,
    "timestep_mean": 591.46875,
    "epsilon_norm": 9.074396133422852,
    "prediction_norm": 7.551313400268555,
    "timestamp": "2025-07-14T12:34:33.637344"
  },
  {
    "epoch": 1674,
    "batch_idx": 128,
    "loss": 0.22825655341148376,
    "learning_rate": 6.41362967307456e-05,
    "timestep_mean": 540.90625,
    "epsilon_norm": 9.033846855163574,
    "prediction_norm": 7.908079147338867,
    "timestamp": "2025-07-14T12:34:33.642336"
  },
  {
    "epoch": 1674,
    "batch_idx": 160,
    "loss": 0.31349289417266846,
    "learning_rate": 6.41362967307456e-05,
    "timestep_mean": 477.90625,
    "epsilon_norm": 8.675496101379395,
    "prediction_norm": 7.024158954620361,
    "timestamp": "2025-07-14T12:34:33.647228"
  },
  {
    "epoch": 1674,
    "batch_idx": 192,
    "loss": 0.15659883618354797,
    "learning_rate": 6.41362967307456e-05,
    "timestep_mean": 358.625,
    "epsilon_norm": 4.160973072052002,
    "prediction_norm": 3.7831194400787354,
    "timestamp": "2025-07-14T12:34:33.652179"
  },
  {
    "epoch": 1675,
    "batch_idx": 0,
    "loss": 0.20864605903625488,
    "learning_rate": 6.37519964636014e-05,
    "timestep_mean": 503.625,
    "epsilon_norm": 8.097201347351074,
    "prediction_norm": 7.037233352661133,
    "timestamp": "2025-07-14T12:34:33.657303"
  },
  {
    "epoch": 1675,
    "batch_idx": 32,
    "loss": 0.1662549376487732,
    "learning_rate": 6.37519964636014e-05,
    "timestep_mean": 585.03125,
    "epsilon_norm": 7.577181816101074,
    "prediction_norm": 6.410287857055664,
    "timestamp": "2025-07-14T12:34:33.662402"
  },
  {
    "epoch": 1675,
    "batch_idx": 64,
    "loss": 0.24988852441310883,
    "learning_rate": 6.37519964636014e-05,
    "timestep_mean": 518.65625,
    "epsilon_norm": 8.420415878295898,
    "prediction_norm": 7.075893402099609,
    "timestamp": "2025-07-14T12:34:33.667799"
  },
  {
    "epoch": 1675,
    "batch_idx": 96,
    "loss": 0.2907499670982361,
    "learning_rate": 6.37519964636014e-05,
    "timestep_mean": 514.15625,
    "epsilon_norm": 8.050085067749023,
    "prediction_norm": 7.461112022399902,
    "timestamp": "2025-07-14T12:34:33.672907"
  },
  {
    "epoch": 1675,
    "batch_idx": 128,
    "loss": 0.348777174949646,
    "learning_rate": 6.37519964636014e-05,
    "timestep_mean": 530.46875,
    "epsilon_norm": 8.583264350891113,
    "prediction_norm": 7.285419464111328,
    "timestamp": "2025-07-14T12:34:33.678168"
  },
  {
    "epoch": 1675,
    "batch_idx": 160,
    "loss": 0.24704815447330475,
    "learning_rate": 6.37519964636014e-05,
    "timestep_mean": 531.53125,
    "epsilon_norm": 7.923713684082031,
    "prediction_norm": 6.678765773773193,
    "timestamp": "2025-07-14T12:34:33.683440"
  },
  {
    "epoch": 1675,
    "batch_idx": 192,
    "loss": 0.14007480442523956,
    "learning_rate": 6.37519964636014e-05,
    "timestep_mean": 535.25,
    "epsilon_norm": 3.9207205772399902,
    "prediction_norm": 3.338244676589966,
    "timestamp": "2025-07-14T12:34:33.688643"
  },
  {
    "epoch": 1676,
    "batch_idx": 0,
    "loss": 0.13202494382858276,
    "learning_rate": 6.336877259503991e-05,
    "timestep_mean": 550.3125,
    "epsilon_norm": 8.429303169250488,
    "prediction_norm": 7.428474426269531,
    "timestamp": "2025-07-14T12:34:33.693752"
  },
  {
    "epoch": 1676,
    "batch_idx": 32,
    "loss": 0.20985743403434753,
    "learning_rate": 6.336877259503991e-05,
    "timestep_mean": 505.21875,
    "epsilon_norm": 9.028048515319824,
    "prediction_norm": 7.766030788421631,
    "timestamp": "2025-07-14T12:34:33.698794"
  },
  {
    "epoch": 1676,
    "batch_idx": 64,
    "loss": 0.2890119254589081,
    "learning_rate": 6.336877259503991e-05,
    "timestep_mean": 462.125,
    "epsilon_norm": 8.103870391845703,
    "prediction_norm": 6.309809684753418,
    "timestamp": "2025-07-14T12:34:33.704069"
  },
  {
    "epoch": 1676,
    "batch_idx": 96,
    "loss": 0.5246315002441406,
    "learning_rate": 6.336877259503991e-05,
    "timestep_mean": 452.5625,
    "epsilon_norm": 9.016693115234375,
    "prediction_norm": 6.964821815490723,
    "timestamp": "2025-07-14T12:34:33.709433"
  },
  {
    "epoch": 1676,
    "batch_idx": 128,
    "loss": 0.22845852375030518,
    "learning_rate": 6.336877259503991e-05,
    "timestep_mean": 501.5,
    "epsilon_norm": 7.46867036819458,
    "prediction_norm": 6.449618816375732,
    "timestamp": "2025-07-14T12:34:33.714526"
  },
  {
    "epoch": 1676,
    "batch_idx": 160,
    "loss": 0.22695016860961914,
    "learning_rate": 6.336877259503991e-05,
    "timestep_mean": 613.28125,
    "epsilon_norm": 8.019723892211914,
    "prediction_norm": 6.7355170249938965,
    "timestamp": "2025-07-14T12:34:33.719799"
  },
  {
    "epoch": 1676,
    "batch_idx": 192,
    "loss": 0.5765525698661804,
    "learning_rate": 6.336877259503991e-05,
    "timestep_mean": 525.875,
    "epsilon_norm": 4.4880781173706055,
    "prediction_norm": 3.49918270111084,
    "timestamp": "2025-07-14T12:34:33.725111"
  },
  {
    "epoch": 1677,
    "batch_idx": 0,
    "loss": 0.4457489848136902,
    "learning_rate": 6.298662607062782e-05,
    "timestep_mean": 463.65625,
    "epsilon_norm": 8.455559730529785,
    "prediction_norm": 6.07554292678833,
    "timestamp": "2025-07-14T12:34:33.730338"
  },
  {
    "epoch": 1677,
    "batch_idx": 32,
    "loss": 0.2544406056404114,
    "learning_rate": 6.298662607062782e-05,
    "timestep_mean": 521.5,
    "epsilon_norm": 7.932656764984131,
    "prediction_norm": 6.900246620178223,
    "timestamp": "2025-07-14T12:34:33.735574"
  },
  {
    "epoch": 1677,
    "batch_idx": 64,
    "loss": 0.20382297039031982,
    "learning_rate": 6.298662607062782e-05,
    "timestep_mean": 538.125,
    "epsilon_norm": 8.310628890991211,
    "prediction_norm": 7.844675064086914,
    "timestamp": "2025-07-14T12:34:33.740972"
  },
  {
    "epoch": 1677,
    "batch_idx": 96,
    "loss": 0.2963966727256775,
    "learning_rate": 6.298662607062782e-05,
    "timestep_mean": 525.78125,
    "epsilon_norm": 8.676225662231445,
    "prediction_norm": 7.6949143409729,
    "timestamp": "2025-07-14T12:34:33.746199"
  },
  {
    "epoch": 1677,
    "batch_idx": 128,
    "loss": 0.30266353487968445,
    "learning_rate": 6.298662607062782e-05,
    "timestep_mean": 530.75,
    "epsilon_norm": 9.410118103027344,
    "prediction_norm": 8.305704116821289,
    "timestamp": "2025-07-14T12:34:33.751346"
  },
  {
    "epoch": 1677,
    "batch_idx": 160,
    "loss": 0.22814670205116272,
    "learning_rate": 6.298662607062782e-05,
    "timestep_mean": 584.9375,
    "epsilon_norm": 7.822544574737549,
    "prediction_norm": 6.993424415588379,
    "timestamp": "2025-07-14T12:34:33.756557"
  },
  {
    "epoch": 1677,
    "batch_idx": 192,
    "loss": 0.08845747262239456,
    "learning_rate": 6.298662607062782e-05,
    "timestep_mean": 616.625,
    "epsilon_norm": 3.3214433193206787,
    "prediction_norm": 3.3892886638641357,
    "timestamp": "2025-07-14T12:34:33.761908"
  },
  {
    "epoch": 1678,
    "batch_idx": 0,
    "loss": 0.4838656783103943,
    "learning_rate": 6.260555783327365e-05,
    "timestep_mean": 518.625,
    "epsilon_norm": 9.084855079650879,
    "prediction_norm": 7.1786065101623535,
    "timestamp": "2025-07-14T12:34:33.767265"
  },
  {
    "epoch": 1678,
    "batch_idx": 32,
    "loss": 0.33429983258247375,
    "learning_rate": 6.260555783327365e-05,
    "timestep_mean": 472.65625,
    "epsilon_norm": 6.938720703125,
    "prediction_norm": 5.831104755401611,
    "timestamp": "2025-07-14T12:34:33.772443"
  },
  {
    "epoch": 1678,
    "batch_idx": 64,
    "loss": 0.2976113557815552,
    "learning_rate": 6.260555783327365e-05,
    "timestep_mean": 452.25,
    "epsilon_norm": 7.639703273773193,
    "prediction_norm": 5.778117656707764,
    "timestamp": "2025-07-14T12:34:33.777772"
  },
  {
    "epoch": 1678,
    "batch_idx": 96,
    "loss": 0.17363683879375458,
    "learning_rate": 6.260555783327365e-05,
    "timestep_mean": 450.65625,
    "epsilon_norm": 9.010503768920898,
    "prediction_norm": 7.793458461761475,
    "timestamp": "2025-07-14T12:34:33.783016"
  },
  {
    "epoch": 1678,
    "batch_idx": 128,
    "loss": 0.1583459973335266,
    "learning_rate": 6.260555783327365e-05,
    "timestep_mean": 485.46875,
    "epsilon_norm": 7.683431625366211,
    "prediction_norm": 6.474550724029541,
    "timestamp": "2025-07-14T12:34:33.788483"
  },
  {
    "epoch": 1678,
    "batch_idx": 160,
    "loss": 0.34267309308052063,
    "learning_rate": 6.260555783327365e-05,
    "timestep_mean": 495.40625,
    "epsilon_norm": 8.819772720336914,
    "prediction_norm": 7.184428691864014,
    "timestamp": "2025-07-14T12:34:33.794055"
  },
  {
    "epoch": 1678,
    "batch_idx": 192,
    "loss": 0.15902617573738098,
    "learning_rate": 6.260555783327365e-05,
    "timestep_mean": 516.75,
    "epsilon_norm": 3.7467589378356934,
    "prediction_norm": 3.608004570007324,
    "timestamp": "2025-07-14T12:34:33.804718"
  },
  {
    "epoch": 1679,
    "batch_idx": 0,
    "loss": 0.3014511466026306,
    "learning_rate": 6.222556882322544e-05,
    "timestep_mean": 497.0625,
    "epsilon_norm": 7.733861923217773,
    "prediction_norm": 6.394537448883057,
    "timestamp": "2025-07-14T12:34:33.814502"
  },
  {
    "epoch": 1679,
    "batch_idx": 32,
    "loss": 0.5395656824111938,
    "learning_rate": 6.222556882322544e-05,
    "timestep_mean": 480.125,
    "epsilon_norm": 8.272452354431152,
    "prediction_norm": 6.208225727081299,
    "timestamp": "2025-07-14T12:34:33.821023"
  },
  {
    "epoch": 1679,
    "batch_idx": 64,
    "loss": 0.21199217438697815,
    "learning_rate": 6.222556882322544e-05,
    "timestep_mean": 548.0,
    "epsilon_norm": 8.883309364318848,
    "prediction_norm": 7.397481441497803,
    "timestamp": "2025-07-14T12:34:33.827442"
  },
  {
    "epoch": 1679,
    "batch_idx": 96,
    "loss": 0.3126213252544403,
    "learning_rate": 6.222556882322544e-05,
    "timestep_mean": 466.125,
    "epsilon_norm": 8.436112403869629,
    "prediction_norm": 7.8566460609436035,
    "timestamp": "2025-07-14T12:34:33.833314"
  },
  {
    "epoch": 1679,
    "batch_idx": 128,
    "loss": 0.24744334816932678,
    "learning_rate": 6.222556882322544e-05,
    "timestep_mean": 545.03125,
    "epsilon_norm": 7.892584800720215,
    "prediction_norm": 6.9639973640441895,
    "timestamp": "2025-07-14T12:34:33.838659"
  },
  {
    "epoch": 1679,
    "batch_idx": 160,
    "loss": 0.23086321353912354,
    "learning_rate": 6.222556882322544e-05,
    "timestep_mean": 547.9375,
    "epsilon_norm": 7.4355244636535645,
    "prediction_norm": 6.070332050323486,
    "timestamp": "2025-07-14T12:34:33.843593"
  },
  {
    "epoch": 1679,
    "batch_idx": 192,
    "loss": 0.15824265778064728,
    "learning_rate": 6.222556882322544e-05,
    "timestep_mean": 410.625,
    "epsilon_norm": 3.294628143310547,
    "prediction_norm": 3.3538711071014404,
    "timestamp": "2025-07-14T12:34:33.848903"
  },
  {
    "epoch": 1680,
    "batch_idx": 0,
    "loss": 0.45772847533226013,
    "learning_rate": 6.184665997806821e-05,
    "timestep_mean": 592.28125,
    "epsilon_norm": 8.624995231628418,
    "prediction_norm": 6.984454154968262,
    "timestamp": "2025-07-14T12:34:33.854344"
  },
  {
    "epoch": 1680,
    "batch_idx": 32,
    "loss": 0.2786291837692261,
    "learning_rate": 6.184665997806821e-05,
    "timestep_mean": 544.21875,
    "epsilon_norm": 7.605703353881836,
    "prediction_norm": 6.471324920654297,
    "timestamp": "2025-07-14T12:34:33.859461"
  },
  {
    "epoch": 1680,
    "batch_idx": 64,
    "loss": 0.33991795778274536,
    "learning_rate": 6.184665997806821e-05,
    "timestep_mean": 532.1875,
    "epsilon_norm": 9.328024864196777,
    "prediction_norm": 7.84664249420166,
    "timestamp": "2025-07-14T12:34:33.864439"
  },
  {
    "epoch": 1680,
    "batch_idx": 96,
    "loss": 0.19201114773750305,
    "learning_rate": 6.184665997806821e-05,
    "timestep_mean": 582.34375,
    "epsilon_norm": 8.008206367492676,
    "prediction_norm": 7.555445194244385,
    "timestamp": "2025-07-14T12:34:33.869553"
  },
  {
    "epoch": 1680,
    "batch_idx": 128,
    "loss": 0.2817339301109314,
    "learning_rate": 6.184665997806821e-05,
    "timestep_mean": 521.46875,
    "epsilon_norm": 8.205583572387695,
    "prediction_norm": 7.569834232330322,
    "timestamp": "2025-07-14T12:34:33.874638"
  },
  {
    "epoch": 1680,
    "batch_idx": 160,
    "loss": 0.2019689977169037,
    "learning_rate": 6.184665997806821e-05,
    "timestep_mean": 498.84375,
    "epsilon_norm": 8.636995315551758,
    "prediction_norm": 8.051210403442383,
    "timestamp": "2025-07-14T12:34:33.879786"
  },
  {
    "epoch": 1680,
    "batch_idx": 192,
    "loss": 0.20086126029491425,
    "learning_rate": 6.184665997806821e-05,
    "timestep_mean": 636.375,
    "epsilon_norm": 3.3882007598876953,
    "prediction_norm": 2.871387243270874,
    "timestamp": "2025-07-14T12:34:33.884997"
  },
  {
    "epoch": 1681,
    "batch_idx": 0,
    "loss": 0.1743985265493393,
    "learning_rate": 6.14688322327221e-05,
    "timestep_mean": 498.59375,
    "epsilon_norm": 8.513144493103027,
    "prediction_norm": 8.04592227935791,
    "timestamp": "2025-07-14T12:34:33.892189"
  },
  {
    "epoch": 1681,
    "batch_idx": 32,
    "loss": 0.31149619817733765,
    "learning_rate": 6.14688322327221e-05,
    "timestep_mean": 581.03125,
    "epsilon_norm": 9.208090782165527,
    "prediction_norm": 8.409476280212402,
    "timestamp": "2025-07-14T12:34:33.900164"
  },
  {
    "epoch": 1681,
    "batch_idx": 64,
    "loss": 0.4113713502883911,
    "learning_rate": 6.14688322327221e-05,
    "timestep_mean": 467.1875,
    "epsilon_norm": 8.278716087341309,
    "prediction_norm": 6.683662414550781,
    "timestamp": "2025-07-14T12:34:33.906061"
  },
  {
    "epoch": 1681,
    "batch_idx": 96,
    "loss": 0.2990957498550415,
    "learning_rate": 6.14688322327221e-05,
    "timestep_mean": 536.6875,
    "epsilon_norm": 6.922953128814697,
    "prediction_norm": 5.915886878967285,
    "timestamp": "2025-07-14T12:34:33.913978"
  },
  {
    "epoch": 1681,
    "batch_idx": 128,
    "loss": 0.1818835437297821,
    "learning_rate": 6.14688322327221e-05,
    "timestep_mean": 524.21875,
    "epsilon_norm": 8.266491889953613,
    "prediction_norm": 7.443331718444824,
    "timestamp": "2025-07-14T12:34:33.919596"
  },
  {
    "epoch": 1681,
    "batch_idx": 160,
    "loss": 0.29412999749183655,
    "learning_rate": 6.14688322327221e-05,
    "timestep_mean": 480.375,
    "epsilon_norm": 7.627429485321045,
    "prediction_norm": 6.416402816772461,
    "timestamp": "2025-07-14T12:34:33.924996"
  },
  {
    "epoch": 1681,
    "batch_idx": 192,
    "loss": 0.24387119710445404,
    "learning_rate": 6.14688322327221e-05,
    "timestep_mean": 601.0,
    "epsilon_norm": 4.075296878814697,
    "prediction_norm": 3.04028058052063,
    "timestamp": "2025-07-14T12:34:33.930877"
  },
  {
    "epoch": 1682,
    "batch_idx": 0,
    "loss": 0.24298474192619324,
    "learning_rate": 6.10920865194392e-05,
    "timestep_mean": 506.46875,
    "epsilon_norm": 7.131247520446777,
    "prediction_norm": 6.085718154907227,
    "timestamp": "2025-07-14T12:34:33.938145"
  },
  {
    "epoch": 1682,
    "batch_idx": 32,
    "loss": 0.2468687742948532,
    "learning_rate": 6.10920865194392e-05,
    "timestep_mean": 474.34375,
    "epsilon_norm": 7.898687839508057,
    "prediction_norm": 7.086264610290527,
    "timestamp": "2025-07-14T12:34:33.945222"
  },
  {
    "epoch": 1682,
    "batch_idx": 64,
    "loss": 0.23136673867702484,
    "learning_rate": 6.10920865194392e-05,
    "timestep_mean": 551.0625,
    "epsilon_norm": 7.53472900390625,
    "prediction_norm": 6.659863471984863,
    "timestamp": "2025-07-14T12:34:33.951217"
  },
  {
    "epoch": 1682,
    "batch_idx": 96,
    "loss": 0.30702659487724304,
    "learning_rate": 6.10920865194392e-05,
    "timestep_mean": 377.84375,
    "epsilon_norm": 7.7888641357421875,
    "prediction_norm": 6.092513084411621,
    "timestamp": "2025-07-14T12:34:33.957289"
  },
  {
    "epoch": 1682,
    "batch_idx": 128,
    "loss": 0.195596382021904,
    "learning_rate": 6.10920865194392e-05,
    "timestep_mean": 566.90625,
    "epsilon_norm": 7.933001518249512,
    "prediction_norm": 7.158688545227051,
    "timestamp": "2025-07-14T12:34:33.963315"
  },
  {
    "epoch": 1682,
    "batch_idx": 160,
    "loss": 0.28815722465515137,
    "learning_rate": 6.10920865194392e-05,
    "timestep_mean": 399.1875,
    "epsilon_norm": 7.951541900634766,
    "prediction_norm": 6.602968692779541,
    "timestamp": "2025-07-14T12:34:33.971492"
  },
  {
    "epoch": 1682,
    "batch_idx": 192,
    "loss": 0.21402235329151154,
    "learning_rate": 6.10920865194392e-05,
    "timestep_mean": 417.0,
    "epsilon_norm": 5.271878242492676,
    "prediction_norm": 4.343291282653809,
    "timestamp": "2025-07-14T12:34:33.977205"
  },
  {
    "epoch": 1683,
    "batch_idx": 0,
    "loss": 0.3091175854206085,
    "learning_rate": 6.071642376780223e-05,
    "timestep_mean": 517.09375,
    "epsilon_norm": 7.294922351837158,
    "prediction_norm": 6.048361301422119,
    "timestamp": "2025-07-14T12:34:33.982959"
  },
  {
    "epoch": 1683,
    "batch_idx": 32,
    "loss": 0.319286048412323,
    "learning_rate": 6.071642376780223e-05,
    "timestep_mean": 549.75,
    "epsilon_norm": 8.057219505310059,
    "prediction_norm": 7.133908748626709,
    "timestamp": "2025-07-14T12:34:33.989223"
  },
  {
    "epoch": 1683,
    "batch_idx": 64,
    "loss": 0.3716038167476654,
    "learning_rate": 6.071642376780223e-05,
    "timestep_mean": 574.40625,
    "epsilon_norm": 8.073163032531738,
    "prediction_norm": 6.271936416625977,
    "timestamp": "2025-07-14T12:34:33.994798"
  },
  {
    "epoch": 1683,
    "batch_idx": 96,
    "loss": 0.2418207973241806,
    "learning_rate": 6.071642376780223e-05,
    "timestep_mean": 489.96875,
    "epsilon_norm": 6.713079929351807,
    "prediction_norm": 5.651784420013428,
    "timestamp": "2025-07-14T12:34:34.000134"
  },
  {
    "epoch": 1683,
    "batch_idx": 128,
    "loss": 0.3929431438446045,
    "learning_rate": 6.071642376780223e-05,
    "timestep_mean": 427.90625,
    "epsilon_norm": 7.92037296295166,
    "prediction_norm": 5.760651111602783,
    "timestamp": "2025-07-14T12:34:34.007731"
  },
  {
    "epoch": 1683,
    "batch_idx": 160,
    "loss": 0.23110990226268768,
    "learning_rate": 6.071642376780223e-05,
    "timestep_mean": 461.03125,
    "epsilon_norm": 8.015702247619629,
    "prediction_norm": 6.745476722717285,
    "timestamp": "2025-07-14T12:34:34.013295"
  },
  {
    "epoch": 1683,
    "batch_idx": 192,
    "loss": 0.1330903023481369,
    "learning_rate": 6.071642376780223e-05,
    "timestep_mean": 547.375,
    "epsilon_norm": 4.150808334350586,
    "prediction_norm": 4.0607805252075195,
    "timestamp": "2025-07-14T12:34:34.019686"
  },
  {
    "epoch": 1684,
    "batch_idx": 0,
    "loss": 0.18613052368164062,
    "learning_rate": 6.034184490472194e-05,
    "timestep_mean": 493.59375,
    "epsilon_norm": 7.3298139572143555,
    "prediction_norm": 6.528134822845459,
    "timestamp": "2025-07-14T12:34:34.024979"
  },
  {
    "epoch": 1684,
    "batch_idx": 32,
    "loss": 0.24477505683898926,
    "learning_rate": 6.034184490472194e-05,
    "timestep_mean": 591.53125,
    "epsilon_norm": 8.231724739074707,
    "prediction_norm": 6.671614646911621,
    "timestamp": "2025-07-14T12:34:34.030076"
  },
  {
    "epoch": 1684,
    "batch_idx": 64,
    "loss": 0.23644906282424927,
    "learning_rate": 6.034184490472194e-05,
    "timestep_mean": 491.3125,
    "epsilon_norm": 8.329586029052734,
    "prediction_norm": 6.844758033752441,
    "timestamp": "2025-07-14T12:34:34.035067"
  },
  {
    "epoch": 1684,
    "batch_idx": 96,
    "loss": 0.2768930196762085,
    "learning_rate": 6.034184490472194e-05,
    "timestep_mean": 515.21875,
    "epsilon_norm": 7.498322486877441,
    "prediction_norm": 6.388613224029541,
    "timestamp": "2025-07-14T12:34:34.040147"
  },
  {
    "epoch": 1684,
    "batch_idx": 128,
    "loss": 0.18657922744750977,
    "learning_rate": 6.034184490472194e-05,
    "timestep_mean": 491.40625,
    "epsilon_norm": 7.202009677886963,
    "prediction_norm": 6.082438945770264,
    "timestamp": "2025-07-14T12:34:34.045120"
  },
  {
    "epoch": 1684,
    "batch_idx": 160,
    "loss": 0.2692358195781708,
    "learning_rate": 6.034184490472194e-05,
    "timestep_mean": 445.625,
    "epsilon_norm": 7.828068733215332,
    "prediction_norm": 6.376218795776367,
    "timestamp": "2025-07-14T12:34:34.050272"
  },
  {
    "epoch": 1684,
    "batch_idx": 192,
    "loss": 0.07411007583141327,
    "learning_rate": 6.034184490472194e-05,
    "timestep_mean": 502.5,
    "epsilon_norm": 4.521744251251221,
    "prediction_norm": 4.197866916656494,
    "timestamp": "2025-07-14T12:34:34.055349"
  },
  {
    "epoch": 1685,
    "batch_idx": 0,
    "loss": 0.29547804594039917,
    "learning_rate": 5.9968350854434026e-05,
    "timestep_mean": 533.21875,
    "epsilon_norm": 7.589957237243652,
    "prediction_norm": 6.014931678771973,
    "timestamp": "2025-07-14T12:34:34.060302"
  },
  {
    "epoch": 1685,
    "batch_idx": 32,
    "loss": 0.10173141956329346,
    "learning_rate": 5.9968350854434026e-05,
    "timestep_mean": 609.9375,
    "epsilon_norm": 9.27573299407959,
    "prediction_norm": 8.740272521972656,
    "timestamp": "2025-07-14T12:34:34.065369"
  },
  {
    "epoch": 1685,
    "batch_idx": 64,
    "loss": 0.22694523632526398,
    "learning_rate": 5.9968350854434026e-05,
    "timestep_mean": 563.71875,
    "epsilon_norm": 9.466617584228516,
    "prediction_norm": 8.63907241821289,
    "timestamp": "2025-07-14T12:34:34.070421"
  },
  {
    "epoch": 1685,
    "batch_idx": 96,
    "loss": 0.31243935227394104,
    "learning_rate": 5.9968350854434026e-05,
    "timestep_mean": 360.65625,
    "epsilon_norm": 7.908476829528809,
    "prediction_norm": 6.173588752746582,
    "timestamp": "2025-07-14T12:34:34.075496"
  },
  {
    "epoch": 1685,
    "batch_idx": 128,
    "loss": 0.324688196182251,
    "learning_rate": 5.9968350854434026e-05,
    "timestep_mean": 445.6875,
    "epsilon_norm": 6.908755779266357,
    "prediction_norm": 5.757012844085693,
    "timestamp": "2025-07-14T12:34:34.080629"
  },
  {
    "epoch": 1685,
    "batch_idx": 160,
    "loss": 0.1815001368522644,
    "learning_rate": 5.9968350854434026e-05,
    "timestep_mean": 445.625,
    "epsilon_norm": 7.633688926696777,
    "prediction_norm": 7.262975692749023,
    "timestamp": "2025-07-14T12:34:34.085665"
  },
  {
    "epoch": 1685,
    "batch_idx": 192,
    "loss": 0.26637688279151917,
    "learning_rate": 5.9968350854434026e-05,
    "timestep_mean": 585.75,
    "epsilon_norm": 4.6626667976379395,
    "prediction_norm": 4.4148335456848145,
    "timestamp": "2025-07-14T12:34:34.090840"
  },
  {
    "epoch": 1686,
    "batch_idx": 0,
    "loss": 0.41944077610969543,
    "learning_rate": 5.95959425384982e-05,
    "timestep_mean": 548.3125,
    "epsilon_norm": 8.853487014770508,
    "prediction_norm": 6.9390869140625,
    "timestamp": "2025-07-14T12:34:34.096068"
  },
  {
    "epoch": 1686,
    "batch_idx": 32,
    "loss": 0.25079524517059326,
    "learning_rate": 5.95959425384982e-05,
    "timestep_mean": 501.71875,
    "epsilon_norm": 9.008720397949219,
    "prediction_norm": 8.171229362487793,
    "timestamp": "2025-07-14T12:34:34.101144"
  },
  {
    "epoch": 1686,
    "batch_idx": 64,
    "loss": 0.42116957902908325,
    "learning_rate": 5.95959425384982e-05,
    "timestep_mean": 438.03125,
    "epsilon_norm": 7.75667667388916,
    "prediction_norm": 5.620769500732422,
    "timestamp": "2025-07-14T12:34:34.106176"
  },
  {
    "epoch": 1686,
    "batch_idx": 96,
    "loss": 0.15624989569187164,
    "learning_rate": 5.95959425384982e-05,
    "timestep_mean": 377.25,
    "epsilon_norm": 6.089404106140137,
    "prediction_norm": 5.4931721687316895,
    "timestamp": "2025-07-14T12:34:34.111063"
  },
  {
    "epoch": 1686,
    "batch_idx": 128,
    "loss": 0.29664164781570435,
    "learning_rate": 5.95959425384982e-05,
    "timestep_mean": 527.78125,
    "epsilon_norm": 7.1094255447387695,
    "prediction_norm": 6.005395889282227,
    "timestamp": "2025-07-14T12:34:34.115934"
  },
  {
    "epoch": 1686,
    "batch_idx": 160,
    "loss": 0.24824075400829315,
    "learning_rate": 5.95959425384982e-05,
    "timestep_mean": 479.71875,
    "epsilon_norm": 7.862639904022217,
    "prediction_norm": 6.50850248336792,
    "timestamp": "2025-07-14T12:34:34.121010"
  },
  {
    "epoch": 1686,
    "batch_idx": 192,
    "loss": 0.06788698583841324,
    "learning_rate": 5.95959425384982e-05,
    "timestep_mean": 756.125,
    "epsilon_norm": 2.271848678588867,
    "prediction_norm": 2.152724266052246,
    "timestamp": "2025-07-14T12:34:34.126053"
  },
  {
    "epoch": 1687,
    "batch_idx": 0,
    "loss": 0.3431275486946106,
    "learning_rate": 5.922462087579494e-05,
    "timestep_mean": 486.28125,
    "epsilon_norm": 7.993370056152344,
    "prediction_norm": 5.752061367034912,
    "timestamp": "2025-07-14T12:34:34.130993"
  },
  {
    "epoch": 1687,
    "batch_idx": 32,
    "loss": 0.148602694272995,
    "learning_rate": 5.922462087579494e-05,
    "timestep_mean": 501.5,
    "epsilon_norm": 7.457648754119873,
    "prediction_norm": 6.191609859466553,
    "timestamp": "2025-07-14T12:34:34.135984"
  },
  {
    "epoch": 1687,
    "batch_idx": 64,
    "loss": 0.2322913110256195,
    "learning_rate": 5.922462087579494e-05,
    "timestep_mean": 481.0625,
    "epsilon_norm": 9.07597827911377,
    "prediction_norm": 7.243405342102051,
    "timestamp": "2025-07-14T12:34:34.140969"
  },
  {
    "epoch": 1687,
    "batch_idx": 96,
    "loss": 0.07730956375598907,
    "learning_rate": 5.922462087579494e-05,
    "timestep_mean": 595.40625,
    "epsilon_norm": 6.4054107666015625,
    "prediction_norm": 6.416905403137207,
    "timestamp": "2025-07-14T12:34:34.145969"
  },
  {
    "epoch": 1687,
    "batch_idx": 128,
    "loss": 0.3238224685192108,
    "learning_rate": 5.922462087579494e-05,
    "timestep_mean": 424.84375,
    "epsilon_norm": 7.621048927307129,
    "prediction_norm": 6.289383411407471,
    "timestamp": "2025-07-14T12:34:34.150867"
  },
  {
    "epoch": 1687,
    "batch_idx": 160,
    "loss": 0.3238074779510498,
    "learning_rate": 5.922462087579494e-05,
    "timestep_mean": 475.46875,
    "epsilon_norm": 8.172199249267578,
    "prediction_norm": 7.56110954284668,
    "timestamp": "2025-07-14T12:34:34.155754"
  },
  {
    "epoch": 1687,
    "batch_idx": 192,
    "loss": 0.7529961466789246,
    "learning_rate": 5.922462087579494e-05,
    "timestep_mean": 361.75,
    "epsilon_norm": 3.7465736865997314,
    "prediction_norm": 2.330935001373291,
    "timestamp": "2025-07-14T12:34:34.160647"
  },
  {
    "epoch": 1688,
    "batch_idx": 0,
    "loss": 0.2901196777820587,
    "learning_rate": 5.8854386782523296e-05,
    "timestep_mean": 486.125,
    "epsilon_norm": 7.07345724105835,
    "prediction_norm": 5.358834266662598,
    "timestamp": "2025-07-14T12:34:34.165544"
  },
  {
    "epoch": 1688,
    "batch_idx": 32,
    "loss": 0.27580997347831726,
    "learning_rate": 5.8854386782523296e-05,
    "timestep_mean": 521.375,
    "epsilon_norm": 8.838677406311035,
    "prediction_norm": 7.8352227210998535,
    "timestamp": "2025-07-14T12:34:34.170538"
  },
  {
    "epoch": 1688,
    "batch_idx": 64,
    "loss": 0.22074630856513977,
    "learning_rate": 5.8854386782523296e-05,
    "timestep_mean": 526.75,
    "epsilon_norm": 7.494152545928955,
    "prediction_norm": 6.634880542755127,
    "timestamp": "2025-07-14T12:34:34.175523"
  },
  {
    "epoch": 1688,
    "batch_idx": 96,
    "loss": 0.26465359330177307,
    "learning_rate": 5.8854386782523296e-05,
    "timestep_mean": 573.90625,
    "epsilon_norm": 8.088628768920898,
    "prediction_norm": 6.951460361480713,
    "timestamp": "2025-07-14T12:34:34.180413"
  },
  {
    "epoch": 1688,
    "batch_idx": 128,
    "loss": 0.1723213642835617,
    "learning_rate": 5.8854386782523296e-05,
    "timestep_mean": 483.375,
    "epsilon_norm": 8.19864559173584,
    "prediction_norm": 7.109896659851074,
    "timestamp": "2025-07-14T12:34:34.185301"
  },
  {
    "epoch": 1688,
    "batch_idx": 160,
    "loss": 0.5541409850120544,
    "learning_rate": 5.8854386782523296e-05,
    "timestep_mean": 401.09375,
    "epsilon_norm": 8.26706314086914,
    "prediction_norm": 5.696103572845459,
    "timestamp": "2025-07-14T12:34:34.190169"
  },
  {
    "epoch": 1688,
    "batch_idx": 192,
    "loss": 0.17453524470329285,
    "learning_rate": 5.8854386782523296e-05,
    "timestep_mean": 567.875,
    "epsilon_norm": 2.132760524749756,
    "prediction_norm": 2.392996311187744,
    "timestamp": "2025-07-14T12:34:34.195063"
  },
  {
    "epoch": 1689,
    "batch_idx": 0,
    "loss": 0.2127230167388916,
    "learning_rate": 5.848524117219954e-05,
    "timestep_mean": 528.0,
    "epsilon_norm": 7.758471965789795,
    "prediction_norm": 6.437929153442383,
    "timestamp": "2025-07-14T12:34:34.200001"
  },
  {
    "epoch": 1689,
    "batch_idx": 32,
    "loss": 0.25807368755340576,
    "learning_rate": 5.848524117219954e-05,
    "timestep_mean": 475.40625,
    "epsilon_norm": 7.276589393615723,
    "prediction_norm": 5.808974742889404,
    "timestamp": "2025-07-14T12:34:34.204916"
  },
  {
    "epoch": 1689,
    "batch_idx": 64,
    "loss": 0.28929629921913147,
    "learning_rate": 5.848524117219954e-05,
    "timestep_mean": 474.3125,
    "epsilon_norm": 8.448923110961914,
    "prediction_norm": 7.038607120513916,
    "timestamp": "2025-07-14T12:34:34.209840"
  },
  {
    "epoch": 1689,
    "batch_idx": 96,
    "loss": 0.17431554198265076,
    "learning_rate": 5.848524117219954e-05,
    "timestep_mean": 506.21875,
    "epsilon_norm": 7.691338539123535,
    "prediction_norm": 6.8137383460998535,
    "timestamp": "2025-07-14T12:34:34.214782"
  },
  {
    "epoch": 1689,
    "batch_idx": 128,
    "loss": 0.2776016294956207,
    "learning_rate": 5.848524117219954e-05,
    "timestep_mean": 449.375,
    "epsilon_norm": 7.293591022491455,
    "prediction_norm": 5.817744255065918,
    "timestamp": "2025-07-14T12:34:34.219618"
  },
  {
    "epoch": 1689,
    "batch_idx": 160,
    "loss": 0.4179999828338623,
    "learning_rate": 5.848524117219954e-05,
    "timestep_mean": 431.875,
    "epsilon_norm": 6.7825398445129395,
    "prediction_norm": 5.508364677429199,
    "timestamp": "2025-07-14T12:34:34.224548"
  },
  {
    "epoch": 1689,
    "batch_idx": 192,
    "loss": 0.049570485949516296,
    "learning_rate": 5.848524117219954e-05,
    "timestep_mean": 599.875,
    "epsilon_norm": 4.024259090423584,
    "prediction_norm": 4.203526020050049,
    "timestamp": "2025-07-14T12:34:34.229549"
  },
  {
    "epoch": 1690,
    "batch_idx": 0,
    "loss": 0.46323806047439575,
    "learning_rate": 5.811718495565326e-05,
    "timestep_mean": 479.90625,
    "epsilon_norm": 8.279415130615234,
    "prediction_norm": 6.370841026306152,
    "timestamp": "2025-07-14T12:34:34.234445"
  },
  {
    "epoch": 1690,
    "batch_idx": 32,
    "loss": 0.2831953465938568,
    "learning_rate": 5.811718495565326e-05,
    "timestep_mean": 422.125,
    "epsilon_norm": 9.214186668395996,
    "prediction_norm": 8.44417667388916,
    "timestamp": "2025-07-14T12:34:34.239344"
  },
  {
    "epoch": 1690,
    "batch_idx": 64,
    "loss": 0.33815598487854004,
    "learning_rate": 5.811718495565326e-05,
    "timestep_mean": 464.65625,
    "epsilon_norm": 8.033591270446777,
    "prediction_norm": 6.634708881378174,
    "timestamp": "2025-07-14T12:34:34.244248"
  },
  {
    "epoch": 1690,
    "batch_idx": 96,
    "loss": 0.22629278898239136,
    "learning_rate": 5.811718495565326e-05,
    "timestep_mean": 439.75,
    "epsilon_norm": 6.211501121520996,
    "prediction_norm": 5.24908971786499,
    "timestamp": "2025-07-14T12:34:34.249153"
  },
  {
    "epoch": 1690,
    "batch_idx": 128,
    "loss": 0.1980942189693451,
    "learning_rate": 5.811718495565326e-05,
    "timestep_mean": 550.0625,
    "epsilon_norm": 7.7013983726501465,
    "prediction_norm": 7.104258060455322,
    "timestamp": "2025-07-14T12:34:34.254004"
  },
  {
    "epoch": 1690,
    "batch_idx": 160,
    "loss": 0.21174447238445282,
    "learning_rate": 5.811718495565326e-05,
    "timestep_mean": 593.65625,
    "epsilon_norm": 8.75684928894043,
    "prediction_norm": 7.437873840332031,
    "timestamp": "2025-07-14T12:34:34.258976"
  },
  {
    "epoch": 1690,
    "batch_idx": 192,
    "loss": 0.07298021763563156,
    "learning_rate": 5.811718495565326e-05,
    "timestep_mean": 535.375,
    "epsilon_norm": 4.132397174835205,
    "prediction_norm": 3.63586688041687,
    "timestamp": "2025-07-14T12:34:34.263835"
  },
  {
    "epoch": 1691,
    "batch_idx": 0,
    "loss": 0.3006100654602051,
    "learning_rate": 5.7750219041026996e-05,
    "timestep_mean": 538.71875,
    "epsilon_norm": 6.964282512664795,
    "prediction_norm": 5.4305877685546875,
    "timestamp": "2025-07-14T12:34:34.268747"
  },
  {
    "epoch": 1691,
    "batch_idx": 32,
    "loss": 0.16090944409370422,
    "learning_rate": 5.7750219041026996e-05,
    "timestep_mean": 543.0625,
    "epsilon_norm": 9.130047798156738,
    "prediction_norm": 8.5488862991333,
    "timestamp": "2025-07-14T12:34:34.273631"
  },
  {
    "epoch": 1691,
    "batch_idx": 64,
    "loss": 0.2746940553188324,
    "learning_rate": 5.7750219041026996e-05,
    "timestep_mean": 503.875,
    "epsilon_norm": 8.81929874420166,
    "prediction_norm": 7.932037353515625,
    "timestamp": "2025-07-14T12:34:34.278520"
  },
  {
    "epoch": 1691,
    "batch_idx": 96,
    "loss": 0.22586220502853394,
    "learning_rate": 5.7750219041026996e-05,
    "timestep_mean": 529.9375,
    "epsilon_norm": 8.318390846252441,
    "prediction_norm": 6.807394027709961,
    "timestamp": "2025-07-14T12:34:34.283404"
  },
  {
    "epoch": 1691,
    "batch_idx": 128,
    "loss": 0.29860877990722656,
    "learning_rate": 5.7750219041026996e-05,
    "timestep_mean": 536.375,
    "epsilon_norm": 8.296422958374023,
    "prediction_norm": 6.917303085327148,
    "timestamp": "2025-07-14T12:34:34.288264"
  },
  {
    "epoch": 1691,
    "batch_idx": 160,
    "loss": 0.33337995409965515,
    "learning_rate": 5.7750219041026996e-05,
    "timestep_mean": 458.25,
    "epsilon_norm": 7.58186674118042,
    "prediction_norm": 6.613717079162598,
    "timestamp": "2025-07-14T12:34:34.293133"
  },
  {
    "epoch": 1691,
    "batch_idx": 192,
    "loss": 0.27146658301353455,
    "learning_rate": 5.7750219041026996e-05,
    "timestep_mean": 542.125,
    "epsilon_norm": 4.428854942321777,
    "prediction_norm": 3.5073065757751465,
    "timestamp": "2025-07-14T12:34:34.297976"
  },
  {
    "epoch": 1692,
    "batch_idx": 0,
    "loss": 0.2938654124736786,
    "learning_rate": 5.738434433377242e-05,
    "timestep_mean": 469.15625,
    "epsilon_norm": 7.412148952484131,
    "prediction_norm": 5.909191131591797,
    "timestamp": "2025-07-14T12:34:34.302933"
  },
  {
    "epoch": 1692,
    "batch_idx": 32,
    "loss": 0.37046724557876587,
    "learning_rate": 5.738434433377242e-05,
    "timestep_mean": 408.75,
    "epsilon_norm": 7.6860246658325195,
    "prediction_norm": 6.294156551361084,
    "timestamp": "2025-07-14T12:34:34.307777"
  },
  {
    "epoch": 1692,
    "batch_idx": 64,
    "loss": 0.34593552350997925,
    "learning_rate": 5.738434433377242e-05,
    "timestep_mean": 476.625,
    "epsilon_norm": 7.958805084228516,
    "prediction_norm": 5.814887046813965,
    "timestamp": "2025-07-14T12:34:34.312811"
  },
  {
    "epoch": 1692,
    "batch_idx": 96,
    "loss": 0.14109215140342712,
    "learning_rate": 5.738434433377242e-05,
    "timestep_mean": 482.625,
    "epsilon_norm": 7.438295841217041,
    "prediction_norm": 6.854227542877197,
    "timestamp": "2025-07-14T12:34:34.317689"
  },
  {
    "epoch": 1692,
    "batch_idx": 128,
    "loss": 0.3643573224544525,
    "learning_rate": 5.738434433377242e-05,
    "timestep_mean": 495.75,
    "epsilon_norm": 9.02006721496582,
    "prediction_norm": 7.323824405670166,
    "timestamp": "2025-07-14T12:34:34.322553"
  },
  {
    "epoch": 1692,
    "batch_idx": 160,
    "loss": 0.2461581826210022,
    "learning_rate": 5.738434433377242e-05,
    "timestep_mean": 506.65625,
    "epsilon_norm": 7.103752136230469,
    "prediction_norm": 6.329465389251709,
    "timestamp": "2025-07-14T12:34:34.327392"
  },
  {
    "epoch": 1692,
    "batch_idx": 192,
    "loss": 0.18372350931167603,
    "learning_rate": 5.738434433377242e-05,
    "timestep_mean": 589.125,
    "epsilon_norm": 5.522569179534912,
    "prediction_norm": 4.4564528465271,
    "timestamp": "2025-07-14T12:34:34.332299"
  },
  {
    "epoch": 1693,
    "batch_idx": 0,
    "loss": 0.45786505937576294,
    "learning_rate": 5.7019561736649117e-05,
    "timestep_mean": 462.78125,
    "epsilon_norm": 9.67135238647461,
    "prediction_norm": 8.742203712463379,
    "timestamp": "2025-07-14T12:34:34.337478"
  },
  {
    "epoch": 1693,
    "batch_idx": 32,
    "loss": 0.3447306156158447,
    "learning_rate": 5.7019561736649117e-05,
    "timestep_mean": 429.21875,
    "epsilon_norm": 8.602241516113281,
    "prediction_norm": 6.946088790893555,
    "timestamp": "2025-07-14T12:34:34.342686"
  },
  {
    "epoch": 1693,
    "batch_idx": 64,
    "loss": 0.20088854432106018,
    "learning_rate": 5.7019561736649117e-05,
    "timestep_mean": 447.03125,
    "epsilon_norm": 7.771493434906006,
    "prediction_norm": 6.350367546081543,
    "timestamp": "2025-07-14T12:34:34.347813"
  },
  {
    "epoch": 1693,
    "batch_idx": 96,
    "loss": 0.33884257078170776,
    "learning_rate": 5.7019561736649117e-05,
    "timestep_mean": 625.34375,
    "epsilon_norm": 8.005213737487793,
    "prediction_norm": 6.674264430999756,
    "timestamp": "2025-07-14T12:34:34.353039"
  },
  {
    "epoch": 1693,
    "batch_idx": 128,
    "loss": 0.49942106008529663,
    "learning_rate": 5.7019561736649117e-05,
    "timestep_mean": 395.09375,
    "epsilon_norm": 8.10489559173584,
    "prediction_norm": 5.867867946624756,
    "timestamp": "2025-07-14T12:34:34.358227"
  },
  {
    "epoch": 1693,
    "batch_idx": 160,
    "loss": 0.4185406565666199,
    "learning_rate": 5.7019561736649117e-05,
    "timestep_mean": 527.90625,
    "epsilon_norm": 9.153873443603516,
    "prediction_norm": 7.124556064605713,
    "timestamp": "2025-07-14T12:34:34.363516"
  },
  {
    "epoch": 1693,
    "batch_idx": 192,
    "loss": 0.2087610960006714,
    "learning_rate": 5.7019561736649117e-05,
    "timestep_mean": 558.875,
    "epsilon_norm": 3.1996312141418457,
    "prediction_norm": 3.527280569076538,
    "timestamp": "2025-07-14T12:34:34.368605"
  },
  {
    "epoch": 1694,
    "batch_idx": 0,
    "loss": 0.21404245495796204,
    "learning_rate": 5.6655872149721724e-05,
    "timestep_mean": 533.46875,
    "epsilon_norm": 7.728983402252197,
    "prediction_norm": 7.305121421813965,
    "timestamp": "2025-07-14T12:34:34.373835"
  },
  {
    "epoch": 1694,
    "batch_idx": 32,
    "loss": 0.14492417871952057,
    "learning_rate": 5.6655872149721724e-05,
    "timestep_mean": 527.15625,
    "epsilon_norm": 7.229728698730469,
    "prediction_norm": 7.296705722808838,
    "timestamp": "2025-07-14T12:34:34.379167"
  },
  {
    "epoch": 1694,
    "batch_idx": 64,
    "loss": 0.21499797701835632,
    "learning_rate": 5.6655872149721724e-05,
    "timestep_mean": 474.28125,
    "epsilon_norm": 5.879525184631348,
    "prediction_norm": 5.109984874725342,
    "timestamp": "2025-07-14T12:34:34.384380"
  },
  {
    "epoch": 1694,
    "batch_idx": 96,
    "loss": 0.33788424730300903,
    "learning_rate": 5.6655872149721724e-05,
    "timestep_mean": 525.375,
    "epsilon_norm": 8.486760139465332,
    "prediction_norm": 7.446864604949951,
    "timestamp": "2025-07-14T12:34:34.389516"
  },
  {
    "epoch": 1694,
    "batch_idx": 128,
    "loss": 0.15471641719341278,
    "learning_rate": 5.6655872149721724e-05,
    "timestep_mean": 554.96875,
    "epsilon_norm": 8.358158111572266,
    "prediction_norm": 7.507647514343262,
    "timestamp": "2025-07-14T12:34:34.394774"
  },
  {
    "epoch": 1694,
    "batch_idx": 160,
    "loss": 0.11879906058311462,
    "learning_rate": 5.6655872149721724e-05,
    "timestep_mean": 545.28125,
    "epsilon_norm": 7.2269463539123535,
    "prediction_norm": 6.977737903594971,
    "timestamp": "2025-07-14T12:34:34.399926"
  },
  {
    "epoch": 1694,
    "batch_idx": 192,
    "loss": 0.8163737654685974,
    "learning_rate": 5.6655872149721724e-05,
    "timestep_mean": 394.625,
    "epsilon_norm": 5.421781063079834,
    "prediction_norm": 3.6836347579956055,
    "timestamp": "2025-07-14T12:34:34.405159"
  },
  {
    "epoch": 1695,
    "batch_idx": 0,
    "loss": 0.18176791071891785,
    "learning_rate": 5.629327647035842e-05,
    "timestep_mean": 547.78125,
    "epsilon_norm": 8.135946273803711,
    "prediction_norm": 6.70660400390625,
    "timestamp": "2025-07-14T12:34:34.410523"
  },
  {
    "epoch": 1695,
    "batch_idx": 32,
    "loss": 0.3391689360141754,
    "learning_rate": 5.629327647035842e-05,
    "timestep_mean": 478.90625,
    "epsilon_norm": 8.256464958190918,
    "prediction_norm": 6.985215663909912,
    "timestamp": "2025-07-14T12:34:34.415726"
  },
  {
    "epoch": 1695,
    "batch_idx": 64,
    "loss": 0.2622395157814026,
    "learning_rate": 5.629327647035842e-05,
    "timestep_mean": 462.625,
    "epsilon_norm": 7.907699108123779,
    "prediction_norm": 6.79264497756958,
    "timestamp": "2025-07-14T12:34:34.421298"
  },
  {
    "epoch": 1695,
    "batch_idx": 96,
    "loss": 0.12511509656906128,
    "learning_rate": 5.629327647035842e-05,
    "timestep_mean": 528.375,
    "epsilon_norm": 7.7216572761535645,
    "prediction_norm": 7.2343950271606445,
    "timestamp": "2025-07-14T12:34:34.426526"
  },
  {
    "epoch": 1695,
    "batch_idx": 128,
    "loss": 0.1922062784433365,
    "learning_rate": 5.629327647035842e-05,
    "timestep_mean": 472.28125,
    "epsilon_norm": 6.657783031463623,
    "prediction_norm": 6.030892848968506,
    "timestamp": "2025-07-14T12:34:34.432090"
  },
  {
    "epoch": 1695,
    "batch_idx": 160,
    "loss": 0.3542274534702301,
    "learning_rate": 5.629327647035842e-05,
    "timestep_mean": 409.6875,
    "epsilon_norm": 8.838666915893555,
    "prediction_norm": 7.304830551147461,
    "timestamp": "2025-07-14T12:34:34.437330"
  },
  {
    "epoch": 1695,
    "batch_idx": 192,
    "loss": 0.19713476300239563,
    "learning_rate": 5.629327647035842e-05,
    "timestep_mean": 351.5,
    "epsilon_norm": 3.5083162784576416,
    "prediction_norm": 3.2136926651000977,
    "timestamp": "2025-07-14T12:34:34.442963"
  },
  {
    "epoch": 1696,
    "batch_idx": 0,
    "loss": 0.27551648020744324,
    "learning_rate": 5.5931775593227755e-05,
    "timestep_mean": 465.34375,
    "epsilon_norm": 8.000826835632324,
    "prediction_norm": 7.023859977722168,
    "timestamp": "2025-07-14T12:34:34.448119"
  },
  {
    "epoch": 1696,
    "batch_idx": 32,
    "loss": 0.26791054010391235,
    "learning_rate": 5.5931775593227755e-05,
    "timestep_mean": 575.4375,
    "epsilon_norm": 8.029082298278809,
    "prediction_norm": 6.911130905151367,
    "timestamp": "2025-07-14T12:34:34.453779"
  },
  {
    "epoch": 1696,
    "batch_idx": 64,
    "loss": 0.3911985754966736,
    "learning_rate": 5.5931775593227755e-05,
    "timestep_mean": 482.46875,
    "epsilon_norm": 8.725194931030273,
    "prediction_norm": 6.5261688232421875,
    "timestamp": "2025-07-14T12:34:34.459161"
  },
  {
    "epoch": 1696,
    "batch_idx": 96,
    "loss": 0.25138986110687256,
    "learning_rate": 5.5931775593227755e-05,
    "timestep_mean": 508.21875,
    "epsilon_norm": 7.924955368041992,
    "prediction_norm": 6.615202903747559,
    "timestamp": "2025-07-14T12:34:34.464377"
  },
  {
    "epoch": 1696,
    "batch_idx": 128,
    "loss": 0.16179321706295013,
    "learning_rate": 5.5931775593227755e-05,
    "timestep_mean": 447.9375,
    "epsilon_norm": 7.008477687835693,
    "prediction_norm": 5.929111003875732,
    "timestamp": "2025-07-14T12:34:34.469619"
  },
  {
    "epoch": 1696,
    "batch_idx": 160,
    "loss": 0.1962086260318756,
    "learning_rate": 5.5931775593227755e-05,
    "timestep_mean": 470.90625,
    "epsilon_norm": 6.428943634033203,
    "prediction_norm": 5.791206359863281,
    "timestamp": "2025-07-14T12:34:34.474854"
  },
  {
    "epoch": 1696,
    "batch_idx": 192,
    "loss": 0.20014865696430206,
    "learning_rate": 5.5931775593227755e-05,
    "timestep_mean": 497.5,
    "epsilon_norm": 3.9989326000213623,
    "prediction_norm": 3.964686870574951,
    "timestamp": "2025-07-14T12:34:34.480322"
  },
  {
    "epoch": 1697,
    "batch_idx": 0,
    "loss": 0.15388627350330353,
    "learning_rate": 5.557137041029724e-05,
    "timestep_mean": 484.78125,
    "epsilon_norm": 7.632378101348877,
    "prediction_norm": 7.3888726234436035,
    "timestamp": "2025-07-14T12:34:34.485315"
  },
  {
    "epoch": 1697,
    "batch_idx": 32,
    "loss": 0.24087762832641602,
    "learning_rate": 5.557137041029724e-05,
    "timestep_mean": 505.625,
    "epsilon_norm": 8.21652889251709,
    "prediction_norm": 7.258911609649658,
    "timestamp": "2025-07-14T12:34:34.491252"
  },
  {
    "epoch": 1697,
    "batch_idx": 64,
    "loss": 0.2466868758201599,
    "learning_rate": 5.557137041029724e-05,
    "timestep_mean": 516.15625,
    "epsilon_norm": 8.52877426147461,
    "prediction_norm": 7.627328872680664,
    "timestamp": "2025-07-14T12:34:34.496555"
  },
  {
    "epoch": 1697,
    "batch_idx": 96,
    "loss": 0.22437353432178497,
    "learning_rate": 5.557137041029724e-05,
    "timestep_mean": 481.40625,
    "epsilon_norm": 7.538817405700684,
    "prediction_norm": 6.288840293884277,
    "timestamp": "2025-07-14T12:34:34.501603"
  },
  {
    "epoch": 1697,
    "batch_idx": 128,
    "loss": 0.301384299993515,
    "learning_rate": 5.557137041029724e-05,
    "timestep_mean": 435.9375,
    "epsilon_norm": 8.280722618103027,
    "prediction_norm": 6.597390651702881,
    "timestamp": "2025-07-14T12:34:34.506762"
  },
  {
    "epoch": 1697,
    "batch_idx": 160,
    "loss": 0.13893987238407135,
    "learning_rate": 5.557137041029724e-05,
    "timestep_mean": 522.71875,
    "epsilon_norm": 6.813808917999268,
    "prediction_norm": 6.490762233734131,
    "timestamp": "2025-07-14T12:34:34.511877"
  },
  {
    "epoch": 1697,
    "batch_idx": 192,
    "loss": 0.14668089151382446,
    "learning_rate": 5.557137041029724e-05,
    "timestep_mean": 551.0,
    "epsilon_norm": 3.3057193756103516,
    "prediction_norm": 3.252652883529663,
    "timestamp": "2025-07-14T12:34:34.516994"
  },
  {
    "epoch": 1698,
    "batch_idx": 0,
    "loss": 0.31143927574157715,
    "learning_rate": 5.52120618108311e-05,
    "timestep_mean": 542.375,
    "epsilon_norm": 7.244444847106934,
    "prediction_norm": 6.07805061340332,
    "timestamp": "2025-07-14T12:34:34.522016"
  },
  {
    "epoch": 1698,
    "batch_idx": 32,
    "loss": 0.24411329627037048,
    "learning_rate": 5.52120618108311e-05,
    "timestep_mean": 486.59375,
    "epsilon_norm": 7.0638747215271,
    "prediction_norm": 5.790858745574951,
    "timestamp": "2025-07-14T12:34:34.527049"
  },
  {
    "epoch": 1698,
    "batch_idx": 64,
    "loss": 0.3322776257991791,
    "learning_rate": 5.52120618108311e-05,
    "timestep_mean": 415.96875,
    "epsilon_norm": 7.808465480804443,
    "prediction_norm": 6.505375385284424,
    "timestamp": "2025-07-14T12:34:34.532191"
  },
  {
    "epoch": 1698,
    "batch_idx": 96,
    "loss": 0.21421369910240173,
    "learning_rate": 5.52120618108311e-05,
    "timestep_mean": 491.09375,
    "epsilon_norm": 8.554654121398926,
    "prediction_norm": 7.312175750732422,
    "timestamp": "2025-07-14T12:34:34.537149"
  },
  {
    "epoch": 1698,
    "batch_idx": 128,
    "loss": 0.30784231424331665,
    "learning_rate": 5.52120618108311e-05,
    "timestep_mean": 547.5,
    "epsilon_norm": 8.113873481750488,
    "prediction_norm": 7.233148574829102,
    "timestamp": "2025-07-14T12:34:34.542334"
  },
  {
    "epoch": 1698,
    "batch_idx": 160,
    "loss": 0.2247934192419052,
    "learning_rate": 5.52120618108311e-05,
    "timestep_mean": 553.125,
    "epsilon_norm": 7.383498668670654,
    "prediction_norm": 6.632040023803711,
    "timestamp": "2025-07-14T12:34:34.547460"
  },
  {
    "epoch": 1698,
    "batch_idx": 192,
    "loss": 0.17854049801826477,
    "learning_rate": 5.52120618108311e-05,
    "timestep_mean": 484.75,
    "epsilon_norm": 2.516195297241211,
    "prediction_norm": 2.1986711025238037,
    "timestamp": "2025-07-14T12:34:34.552420"
  },
  {
    "epoch": 1699,
    "batch_idx": 0,
    "loss": 0.16535025835037231,
    "learning_rate": 5.485385068138715e-05,
    "timestep_mean": 532.15625,
    "epsilon_norm": 7.38047981262207,
    "prediction_norm": 6.737320899963379,
    "timestamp": "2025-07-14T12:34:34.557466"
  },
  {
    "epoch": 1699,
    "batch_idx": 32,
    "loss": 0.2518506646156311,
    "learning_rate": 5.485385068138715e-05,
    "timestep_mean": 473.59375,
    "epsilon_norm": 8.539491653442383,
    "prediction_norm": 6.867936134338379,
    "timestamp": "2025-07-14T12:34:34.562527"
  },
  {
    "epoch": 1699,
    "batch_idx": 64,
    "loss": 0.1959659457206726,
    "learning_rate": 5.485385068138715e-05,
    "timestep_mean": 500.6875,
    "epsilon_norm": 8.22282600402832,
    "prediction_norm": 7.326959133148193,
    "timestamp": "2025-07-14T12:34:34.567663"
  },
  {
    "epoch": 1699,
    "batch_idx": 96,
    "loss": 0.2260131984949112,
    "learning_rate": 5.485385068138715e-05,
    "timestep_mean": 588.8125,
    "epsilon_norm": 7.962432384490967,
    "prediction_norm": 7.349236011505127,
    "timestamp": "2025-07-14T12:34:34.572648"
  },
  {
    "epoch": 1699,
    "batch_idx": 128,
    "loss": 0.38787853717803955,
    "learning_rate": 5.485385068138715e-05,
    "timestep_mean": 465.59375,
    "epsilon_norm": 8.84656047821045,
    "prediction_norm": 7.267709255218506,
    "timestamp": "2025-07-14T12:34:34.577679"
  },
  {
    "epoch": 1699,
    "batch_idx": 160,
    "loss": 0.4854888319969177,
    "learning_rate": 5.485385068138715e-05,
    "timestep_mean": 442.71875,
    "epsilon_norm": 7.211904525756836,
    "prediction_norm": 5.474437236785889,
    "timestamp": "2025-07-14T12:34:34.582738"
  },
  {
    "epoch": 1699,
    "batch_idx": 192,
    "loss": 0.20141616463661194,
    "learning_rate": 5.485385068138715e-05,
    "timestep_mean": 492.0,
    "epsilon_norm": 3.5128700733184814,
    "prediction_norm": 3.16816782951355,
    "timestamp": "2025-07-14T12:34:34.587833"
  },
  {
    "epoch": 1700,
    "batch_idx": 0,
    "loss": 0.17452861368656158,
    "learning_rate": 5.44967379058161e-05,
    "timestep_mean": 536.53125,
    "epsilon_norm": 8.325286865234375,
    "prediction_norm": 7.2156853675842285,
    "timestamp": "2025-07-14T12:34:34.593163"
  },
  {
    "epoch": 1700,
    "batch_idx": 32,
    "loss": 0.27262067794799805,
    "learning_rate": 5.44967379058161e-05,
    "timestep_mean": 407.1875,
    "epsilon_norm": 7.5401611328125,
    "prediction_norm": 6.061221122741699,
    "timestamp": "2025-07-14T12:34:34.598322"
  },
  {
    "epoch": 1700,
    "batch_idx": 64,
    "loss": 0.29986217617988586,
    "learning_rate": 5.44967379058161e-05,
    "timestep_mean": 458.25,
    "epsilon_norm": 8.456762313842773,
    "prediction_norm": 7.023675918579102,
    "timestamp": "2025-07-14T12:34:34.603465"
  },
  {
    "epoch": 1700,
    "batch_idx": 96,
    "loss": 0.4351827800273895,
    "learning_rate": 5.44967379058161e-05,
    "timestep_mean": 389.96875,
    "epsilon_norm": 7.739195346832275,
    "prediction_norm": 5.7906880378723145,
    "timestamp": "2025-07-14T12:34:34.608551"
  },
  {
    "epoch": 1700,
    "batch_idx": 128,
    "loss": 0.2754545211791992,
    "learning_rate": 5.44967379058161e-05,
    "timestep_mean": 522.96875,
    "epsilon_norm": 7.624602317810059,
    "prediction_norm": 6.201323509216309,
    "timestamp": "2025-07-14T12:34:34.613915"
  },
  {
    "epoch": 1700,
    "batch_idx": 160,
    "loss": 0.31587114930152893,
    "learning_rate": 5.44967379058161e-05,
    "timestep_mean": 421.40625,
    "epsilon_norm": 8.228364944458008,
    "prediction_norm": 7.130825996398926,
    "timestamp": "2025-07-14T12:34:34.619246"
  },
  {
    "epoch": 1700,
    "batch_idx": 192,
    "loss": 0.581998348236084,
    "learning_rate": 5.44967379058161e-05,
    "timestep_mean": 530.125,
    "epsilon_norm": 4.060754776000977,
    "prediction_norm": 2.358818531036377,
    "timestamp": "2025-07-14T12:34:34.624301"
  },
  {
    "epoch": 1701,
    "batch_idx": 0,
    "loss": 0.23536550998687744,
    "learning_rate": 5.414072436525807e-05,
    "timestep_mean": 442.65625,
    "epsilon_norm": 8.449423789978027,
    "prediction_norm": 6.863460063934326,
    "timestamp": "2025-07-14T12:34:34.629549"
  },
  {
    "epoch": 1701,
    "batch_idx": 32,
    "loss": 0.25471723079681396,
    "learning_rate": 5.414072436525807e-05,
    "timestep_mean": 504.4375,
    "epsilon_norm": 8.560432434082031,
    "prediction_norm": 7.26264762878418,
    "timestamp": "2025-07-14T12:34:34.634722"
  },
  {
    "epoch": 1701,
    "batch_idx": 64,
    "loss": 0.2823568880558014,
    "learning_rate": 5.414072436525807e-05,
    "timestep_mean": 370.53125,
    "epsilon_norm": 7.483457565307617,
    "prediction_norm": 6.585414886474609,
    "timestamp": "2025-07-14T12:34:34.640158"
  },
  {
    "epoch": 1701,
    "batch_idx": 96,
    "loss": 0.2565556764602661,
    "learning_rate": 5.414072436525807e-05,
    "timestep_mean": 484.65625,
    "epsilon_norm": 6.476783752441406,
    "prediction_norm": 5.026244163513184,
    "timestamp": "2025-07-14T12:34:34.645389"
  },
  {
    "epoch": 1701,
    "batch_idx": 128,
    "loss": 0.2812066078186035,
    "learning_rate": 5.414072436525807e-05,
    "timestep_mean": 587.625,
    "epsilon_norm": 7.702801704406738,
    "prediction_norm": 5.883687973022461,
    "timestamp": "2025-07-14T12:34:34.650627"
  },
  {
    "epoch": 1701,
    "batch_idx": 160,
    "loss": 0.38005322217941284,
    "learning_rate": 5.414072436525807e-05,
    "timestep_mean": 469.8125,
    "epsilon_norm": 7.0681047439575195,
    "prediction_norm": 5.580968379974365,
    "timestamp": "2025-07-14T12:34:34.655947"
  },
  {
    "epoch": 1701,
    "batch_idx": 192,
    "loss": 0.16762249171733856,
    "learning_rate": 5.414072436525807e-05,
    "timestep_mean": 706.375,
    "epsilon_norm": 3.4559805393218994,
    "prediction_norm": 3.131817579269409,
    "timestamp": "2025-07-14T12:34:34.661151"
  },
  {
    "epoch": 1702,
    "batch_idx": 0,
    "loss": 0.1241971105337143,
    "learning_rate": 5.3785810938140993e-05,
    "timestep_mean": 519.875,
    "epsilon_norm": 8.243197441101074,
    "prediction_norm": 7.566556453704834,
    "timestamp": "2025-07-14T12:34:34.666446"
  },
  {
    "epoch": 1702,
    "batch_idx": 32,
    "loss": 0.39245426654815674,
    "learning_rate": 5.3785810938140993e-05,
    "timestep_mean": 493.53125,
    "epsilon_norm": 8.666876792907715,
    "prediction_norm": 6.911865711212158,
    "timestamp": "2025-07-14T12:34:34.671810"
  },
  {
    "epoch": 1702,
    "batch_idx": 64,
    "loss": 0.2473081350326538,
    "learning_rate": 5.3785810938140993e-05,
    "timestep_mean": 496.125,
    "epsilon_norm": 8.97824764251709,
    "prediction_norm": 7.404888153076172,
    "timestamp": "2025-07-14T12:34:34.676941"
  },
  {
    "epoch": 1702,
    "batch_idx": 96,
    "loss": 0.457209974527359,
    "learning_rate": 5.3785810938140993e-05,
    "timestep_mean": 550.25,
    "epsilon_norm": 8.021011352539062,
    "prediction_norm": 6.655487537384033,
    "timestamp": "2025-07-14T12:34:34.682313"
  },
  {
    "epoch": 1702,
    "batch_idx": 128,
    "loss": 0.22435951232910156,
    "learning_rate": 5.3785810938140993e-05,
    "timestep_mean": 512.34375,
    "epsilon_norm": 8.250301361083984,
    "prediction_norm": 7.193064212799072,
    "timestamp": "2025-07-14T12:34:34.687643"
  },
  {
    "epoch": 1702,
    "batch_idx": 160,
    "loss": 0.2790088951587677,
    "learning_rate": 5.3785810938140993e-05,
    "timestep_mean": 472.3125,
    "epsilon_norm": 8.391961097717285,
    "prediction_norm": 7.334070682525635,
    "timestamp": "2025-07-14T12:34:34.692856"
  },
  {
    "epoch": 1702,
    "batch_idx": 192,
    "loss": 0.9612218141555786,
    "learning_rate": 5.3785810938140993e-05,
    "timestep_mean": 511.625,
    "epsilon_norm": 4.406765937805176,
    "prediction_norm": 3.405914783477783,
    "timestamp": "2025-07-14T12:34:34.698157"
  },
  {
    "epoch": 1703,
    "batch_idx": 0,
    "loss": 0.21853293478488922,
    "learning_rate": 5.343199850017882e-05,
    "timestep_mean": 517.09375,
    "epsilon_norm": 7.393517971038818,
    "prediction_norm": 6.5498199462890625,
    "timestamp": "2025-07-14T12:34:34.703540"
  },
  {
    "epoch": 1703,
    "batch_idx": 32,
    "loss": 0.4423553943634033,
    "learning_rate": 5.343199850017882e-05,
    "timestep_mean": 452.84375,
    "epsilon_norm": 8.611893653869629,
    "prediction_norm": 6.2945556640625,
    "timestamp": "2025-07-14T12:34:34.708647"
  },
  {
    "epoch": 1703,
    "batch_idx": 64,
    "loss": 0.49059513211250305,
    "learning_rate": 5.343199850017882e-05,
    "timestep_mean": 533.96875,
    "epsilon_norm": 8.378029823303223,
    "prediction_norm": 6.3188557624816895,
    "timestamp": "2025-07-14T12:34:34.713963"
  },
  {
    "epoch": 1703,
    "batch_idx": 96,
    "loss": 0.20861545205116272,
    "learning_rate": 5.343199850017882e-05,
    "timestep_mean": 561.375,
    "epsilon_norm": 7.750617027282715,
    "prediction_norm": 6.940836429595947,
    "timestamp": "2025-07-14T12:34:34.719275"
  },
  {
    "epoch": 1703,
    "batch_idx": 128,
    "loss": 0.2595530152320862,
    "learning_rate": 5.343199850017882e-05,
    "timestep_mean": 520.8125,
    "epsilon_norm": 6.700557231903076,
    "prediction_norm": 5.536803245544434,
    "timestamp": "2025-07-14T12:34:34.724316"
  },
  {
    "epoch": 1703,
    "batch_idx": 160,
    "loss": 0.3984869718551636,
    "learning_rate": 5.343199850017882e-05,
    "timestep_mean": 450.59375,
    "epsilon_norm": 8.047289848327637,
    "prediction_norm": 5.88732385635376,
    "timestamp": "2025-07-14T12:34:34.729321"
  },
  {
    "epoch": 1703,
    "batch_idx": 192,
    "loss": 0.24664965271949768,
    "learning_rate": 5.343199850017882e-05,
    "timestep_mean": 424.0,
    "epsilon_norm": 4.8328962326049805,
    "prediction_norm": 3.6325738430023193,
    "timestamp": "2025-07-14T12:34:34.734235"
  },
  {
    "epoch": 1704,
    "batch_idx": 0,
    "loss": 0.25563594698905945,
    "learning_rate": 5.30792879243681e-05,
    "timestep_mean": 619.21875,
    "epsilon_norm": 8.073756217956543,
    "prediction_norm": 6.760913848876953,
    "timestamp": "2025-07-14T12:34:34.739172"
  },
  {
    "epoch": 1704,
    "batch_idx": 32,
    "loss": 0.17786414921283722,
    "learning_rate": 5.30792879243681e-05,
    "timestep_mean": 570.9375,
    "epsilon_norm": 8.123300552368164,
    "prediction_norm": 7.020901679992676,
    "timestamp": "2025-07-14T12:34:34.744145"
  },
  {
    "epoch": 1704,
    "batch_idx": 64,
    "loss": 0.24970847368240356,
    "learning_rate": 5.30792879243681e-05,
    "timestep_mean": 484.59375,
    "epsilon_norm": 9.825446128845215,
    "prediction_norm": 8.116985321044922,
    "timestamp": "2025-07-14T12:34:34.749263"
  },
  {
    "epoch": 1704,
    "batch_idx": 96,
    "loss": 0.17715473473072052,
    "learning_rate": 5.30792879243681e-05,
    "timestep_mean": 481.46875,
    "epsilon_norm": 9.013704299926758,
    "prediction_norm": 7.744818210601807,
    "timestamp": "2025-07-14T12:34:34.754428"
  },
  {
    "epoch": 1704,
    "batch_idx": 128,
    "loss": 0.33773601055145264,
    "learning_rate": 5.30792879243681e-05,
    "timestep_mean": 496.65625,
    "epsilon_norm": 7.3342461585998535,
    "prediction_norm": 6.115479469299316,
    "timestamp": "2025-07-14T12:34:34.759559"
  },
  {
    "epoch": 1704,
    "batch_idx": 160,
    "loss": 0.1821746677160263,
    "learning_rate": 5.30792879243681e-05,
    "timestep_mean": 541.4375,
    "epsilon_norm": 7.264690399169922,
    "prediction_norm": 5.341212272644043,
    "timestamp": "2025-07-14T12:34:34.764685"
  },
  {
    "epoch": 1704,
    "batch_idx": 192,
    "loss": 0.29049551486968994,
    "learning_rate": 5.30792879243681e-05,
    "timestep_mean": 559.5,
    "epsilon_norm": 4.389125823974609,
    "prediction_norm": 3.8207311630249023,
    "timestamp": "2025-07-14T12:34:34.769899"
  },
  {
    "epoch": 1705,
    "batch_idx": 0,
    "loss": 0.39910590648651123,
    "learning_rate": 5.272768008098737e-05,
    "timestep_mean": 438.28125,
    "epsilon_norm": 8.890257835388184,
    "prediction_norm": 6.95905876159668,
    "timestamp": "2025-07-14T12:34:34.774985"
  },
  {
    "epoch": 1705,
    "batch_idx": 32,
    "loss": 0.12736503779888153,
    "learning_rate": 5.272768008098737e-05,
    "timestep_mean": 506.90625,
    "epsilon_norm": 6.178747653961182,
    "prediction_norm": 5.685396671295166,
    "timestamp": "2025-07-14T12:34:34.780201"
  },
  {
    "epoch": 1705,
    "batch_idx": 64,
    "loss": 0.37452152371406555,
    "learning_rate": 5.272768008098737e-05,
    "timestep_mean": 405.8125,
    "epsilon_norm": 9.312240600585938,
    "prediction_norm": 7.397666931152344,
    "timestamp": "2025-07-14T12:34:34.785478"
  },
  {
    "epoch": 1705,
    "batch_idx": 96,
    "loss": 0.34283706545829773,
    "learning_rate": 5.272768008098737e-05,
    "timestep_mean": 455.25,
    "epsilon_norm": 7.532054424285889,
    "prediction_norm": 6.038305282592773,
    "timestamp": "2025-07-14T12:34:34.790685"
  },
  {
    "epoch": 1705,
    "batch_idx": 128,
    "loss": 0.2156643271446228,
    "learning_rate": 5.272768008098737e-05,
    "timestep_mean": 522.34375,
    "epsilon_norm": 8.707962036132812,
    "prediction_norm": 7.3030171394348145,
    "timestamp": "2025-07-14T12:34:34.795810"
  },
  {
    "epoch": 1705,
    "batch_idx": 160,
    "loss": 0.3697512149810791,
    "learning_rate": 5.272768008098737e-05,
    "timestep_mean": 466.0,
    "epsilon_norm": 7.175731658935547,
    "prediction_norm": 5.448281288146973,
    "timestamp": "2025-07-14T12:34:34.800757"
  },
  {
    "epoch": 1705,
    "batch_idx": 192,
    "loss": 0.42834311723709106,
    "learning_rate": 5.272768008098737e-05,
    "timestep_mean": 539.25,
    "epsilon_norm": 4.287639617919922,
    "prediction_norm": 3.425393581390381,
    "timestamp": "2025-07-14T12:34:34.805775"
  },
  {
    "epoch": 1706,
    "batch_idx": 0,
    "loss": 0.1696777194738388,
    "learning_rate": 5.2377175837594204e-05,
    "timestep_mean": 517.71875,
    "epsilon_norm": 6.909950256347656,
    "prediction_norm": 6.6247758865356445,
    "timestamp": "2025-07-14T12:34:34.810690"
  },
  {
    "epoch": 1706,
    "batch_idx": 32,
    "loss": 0.21231740713119507,
    "learning_rate": 5.2377175837594204e-05,
    "timestep_mean": 491.375,
    "epsilon_norm": 7.645308017730713,
    "prediction_norm": 6.907154560089111,
    "timestamp": "2025-07-14T12:34:34.815598"
  },
  {
    "epoch": 1706,
    "batch_idx": 64,
    "loss": 0.17035943269729614,
    "learning_rate": 5.2377175837594204e-05,
    "timestep_mean": 557.875,
    "epsilon_norm": 8.125213623046875,
    "prediction_norm": 7.542076110839844,
    "timestamp": "2025-07-14T12:34:34.820482"
  },
  {
    "epoch": 1706,
    "batch_idx": 96,
    "loss": 0.1875612586736679,
    "learning_rate": 5.2377175837594204e-05,
    "timestep_mean": 583.375,
    "epsilon_norm": 8.800830841064453,
    "prediction_norm": 7.431656360626221,
    "timestamp": "2025-07-14T12:34:34.825538"
  },
  {
    "epoch": 1706,
    "batch_idx": 128,
    "loss": 0.31554365158081055,
    "learning_rate": 5.2377175837594204e-05,
    "timestep_mean": 487.15625,
    "epsilon_norm": 8.265816688537598,
    "prediction_norm": 6.753908157348633,
    "timestamp": "2025-07-14T12:34:34.830456"
  },
  {
    "epoch": 1706,
    "batch_idx": 160,
    "loss": 0.3217107653617859,
    "learning_rate": 5.2377175837594204e-05,
    "timestep_mean": 468.875,
    "epsilon_norm": 8.319133758544922,
    "prediction_norm": 7.186213493347168,
    "timestamp": "2025-07-14T12:34:34.835331"
  },
  {
    "epoch": 1706,
    "batch_idx": 192,
    "loss": 0.2508115768432617,
    "learning_rate": 5.2377175837594204e-05,
    "timestep_mean": 566.0,
    "epsilon_norm": 3.5654728412628174,
    "prediction_norm": 2.995555877685547,
    "timestamp": "2025-07-14T12:34:34.840183"
  },
  {
    "epoch": 1707,
    "batch_idx": 0,
    "loss": 0.12695901095867157,
    "learning_rate": 5.202777605902258e-05,
    "timestep_mean": 532.71875,
    "epsilon_norm": 8.354898452758789,
    "prediction_norm": 7.962545871734619,
    "timestamp": "2025-07-14T12:34:34.845094"
  },
  {
    "epoch": 1707,
    "batch_idx": 32,
    "loss": 0.26104485988616943,
    "learning_rate": 5.202777605902258e-05,
    "timestep_mean": 444.125,
    "epsilon_norm": 8.337084770202637,
    "prediction_norm": 6.697347640991211,
    "timestamp": "2025-07-14T12:34:34.850162"
  },
  {
    "epoch": 1707,
    "batch_idx": 64,
    "loss": 0.2982574999332428,
    "learning_rate": 5.202777605902258e-05,
    "timestep_mean": 442.03125,
    "epsilon_norm": 6.801693916320801,
    "prediction_norm": 4.941768646240234,
    "timestamp": "2025-07-14T12:34:34.855117"
  },
  {
    "epoch": 1707,
    "batch_idx": 96,
    "loss": 0.22889533638954163,
    "learning_rate": 5.202777605902258e-05,
    "timestep_mean": 517.25,
    "epsilon_norm": 8.611112594604492,
    "prediction_norm": 7.28765869140625,
    "timestamp": "2025-07-14T12:34:34.860144"
  },
  {
    "epoch": 1707,
    "batch_idx": 128,
    "loss": 0.41768234968185425,
    "learning_rate": 5.202777605902258e-05,
    "timestep_mean": 403.1875,
    "epsilon_norm": 8.144430160522461,
    "prediction_norm": 6.237497806549072,
    "timestamp": "2025-07-14T12:34:34.865042"
  },
  {
    "epoch": 1707,
    "batch_idx": 160,
    "loss": 0.27678075432777405,
    "learning_rate": 5.202777605902258e-05,
    "timestep_mean": 522.71875,
    "epsilon_norm": 8.390968322753906,
    "prediction_norm": 7.149380207061768,
    "timestamp": "2025-07-14T12:34:34.870002"
  },
  {
    "epoch": 1707,
    "batch_idx": 192,
    "loss": 0.7140523791313171,
    "learning_rate": 5.202777605902258e-05,
    "timestep_mean": 461.75,
    "epsilon_norm": 4.17073917388916,
    "prediction_norm": 3.1691830158233643,
    "timestamp": "2025-07-14T12:34:34.875102"
  },
  {
    "epoch": 1708,
    "batch_idx": 0,
    "loss": 0.3754822313785553,
    "learning_rate": 5.167948160738205e-05,
    "timestep_mean": 538.4375,
    "epsilon_norm": 7.869372844696045,
    "prediction_norm": 7.1999382972717285,
    "timestamp": "2025-07-14T12:34:34.880142"
  },
  {
    "epoch": 1708,
    "batch_idx": 32,
    "loss": 0.16898475587368011,
    "learning_rate": 5.167948160738205e-05,
    "timestep_mean": 526.1875,
    "epsilon_norm": 7.63482141494751,
    "prediction_norm": 7.145946502685547,
    "timestamp": "2025-07-14T12:34:34.885103"
  },
  {
    "epoch": 1708,
    "batch_idx": 64,
    "loss": 0.1835007667541504,
    "learning_rate": 5.167948160738205e-05,
    "timestep_mean": 550.40625,
    "epsilon_norm": 7.536239147186279,
    "prediction_norm": 6.923702239990234,
    "timestamp": "2025-07-14T12:34:34.890105"
  },
  {
    "epoch": 1708,
    "batch_idx": 96,
    "loss": 0.2967309355735779,
    "learning_rate": 5.167948160738205e-05,
    "timestep_mean": 367.34375,
    "epsilon_norm": 7.622281074523926,
    "prediction_norm": 6.064131736755371,
    "timestamp": "2025-07-14T12:34:34.895085"
  },
  {
    "epoch": 1708,
    "batch_idx": 128,
    "loss": 0.2932683229446411,
    "learning_rate": 5.167948160738205e-05,
    "timestep_mean": 515.125,
    "epsilon_norm": 7.712658882141113,
    "prediction_norm": 5.813871383666992,
    "timestamp": "2025-07-14T12:34:34.900086"
  },
  {
    "epoch": 1708,
    "batch_idx": 160,
    "loss": 0.26766636967658997,
    "learning_rate": 5.167948160738205e-05,
    "timestep_mean": 461.65625,
    "epsilon_norm": 7.739587783813477,
    "prediction_norm": 6.7814130783081055,
    "timestamp": "2025-07-14T12:34:34.905042"
  },
  {
    "epoch": 1708,
    "batch_idx": 192,
    "loss": 0.4848182797431946,
    "learning_rate": 5.167948160738205e-05,
    "timestep_mean": 461.75,
    "epsilon_norm": 4.260507583618164,
    "prediction_norm": 2.7297468185424805,
    "timestamp": "2025-07-14T12:34:34.909983"
  },
  {
    "epoch": 1709,
    "batch_idx": 0,
    "loss": 0.23676680028438568,
    "learning_rate": 5.1332293342054405e-05,
    "timestep_mean": 512.8125,
    "epsilon_norm": 8.595194816589355,
    "prediction_norm": 7.7217698097229,
    "timestamp": "2025-07-14T12:34:34.915036"
  },
  {
    "epoch": 1709,
    "batch_idx": 32,
    "loss": 0.5385484099388123,
    "learning_rate": 5.1332293342054405e-05,
    "timestep_mean": 488.375,
    "epsilon_norm": 8.77367115020752,
    "prediction_norm": 7.202664852142334,
    "timestamp": "2025-07-14T12:34:34.919993"
  },
  {
    "epoch": 1709,
    "batch_idx": 64,
    "loss": 0.15039464831352234,
    "learning_rate": 5.1332293342054405e-05,
    "timestep_mean": 435.3125,
    "epsilon_norm": 7.214082717895508,
    "prediction_norm": 5.651230335235596,
    "timestamp": "2025-07-14T12:34:34.925080"
  },
  {
    "epoch": 1709,
    "batch_idx": 96,
    "loss": 0.3527708649635315,
    "learning_rate": 5.1332293342054405e-05,
    "timestep_mean": 477.21875,
    "epsilon_norm": 6.611961841583252,
    "prediction_norm": 5.93828010559082,
    "timestamp": "2025-07-14T12:34:34.930233"
  },
  {
    "epoch": 1709,
    "batch_idx": 128,
    "loss": 0.274593710899353,
    "learning_rate": 5.1332293342054405e-05,
    "timestep_mean": 513.75,
    "epsilon_norm": 8.11125373840332,
    "prediction_norm": 6.678578853607178,
    "timestamp": "2025-07-14T12:34:34.935246"
  },
  {
    "epoch": 1709,
    "batch_idx": 160,
    "loss": 0.31679612398147583,
    "learning_rate": 5.1332293342054405e-05,
    "timestep_mean": 468.71875,
    "epsilon_norm": 7.534699440002441,
    "prediction_norm": 5.990240573883057,
    "timestamp": "2025-07-14T12:34:34.940298"
  },
  {
    "epoch": 1709,
    "batch_idx": 192,
    "loss": 0.19039997458457947,
    "learning_rate": 5.1332293342054405e-05,
    "timestep_mean": 556.875,
    "epsilon_norm": 3.128004312515259,
    "prediction_norm": 2.81040358543396,
    "timestamp": "2025-07-14T12:34:34.945504"
  },
  {
    "epoch": 1710,
    "batch_idx": 0,
    "loss": 0.34453561902046204,
    "learning_rate": 5.098621211969222e-05,
    "timestep_mean": 473.0625,
    "epsilon_norm": 7.3675336837768555,
    "prediction_norm": 5.693247318267822,
    "timestamp": "2025-07-14T12:34:34.950671"
  },
  {
    "epoch": 1710,
    "batch_idx": 32,
    "loss": 0.36735039949417114,
    "learning_rate": 5.098621211969222e-05,
    "timestep_mean": 561.28125,
    "epsilon_norm": 8.796588897705078,
    "prediction_norm": 7.506319046020508,
    "timestamp": "2025-07-14T12:34:34.955776"
  },
  {
    "epoch": 1710,
    "batch_idx": 64,
    "loss": 0.3104826807975769,
    "learning_rate": 5.098621211969222e-05,
    "timestep_mean": 456.25,
    "epsilon_norm": 6.361660480499268,
    "prediction_norm": 5.2566938400268555,
    "timestamp": "2025-07-14T12:34:34.960828"
  },
  {
    "epoch": 1710,
    "batch_idx": 96,
    "loss": 0.20700421929359436,
    "learning_rate": 5.098621211969222e-05,
    "timestep_mean": 488.21875,
    "epsilon_norm": 7.061888694763184,
    "prediction_norm": 6.007204055786133,
    "timestamp": "2025-07-14T12:34:34.965753"
  },
  {
    "epoch": 1710,
    "batch_idx": 128,
    "loss": 0.21423926949501038,
    "learning_rate": 5.098621211969222e-05,
    "timestep_mean": 551.1875,
    "epsilon_norm": 8.385811805725098,
    "prediction_norm": 7.236749172210693,
    "timestamp": "2025-07-14T12:34:34.970723"
  },
  {
    "epoch": 1710,
    "batch_idx": 160,
    "loss": 0.1645447164773941,
    "learning_rate": 5.098621211969222e-05,
    "timestep_mean": 453.875,
    "epsilon_norm": 6.922788619995117,
    "prediction_norm": 6.306894779205322,
    "timestamp": "2025-07-14T12:34:34.975727"
  },
  {
    "epoch": 1710,
    "batch_idx": 192,
    "loss": 0.8602712154388428,
    "learning_rate": 5.098621211969222e-05,
    "timestep_mean": 491.375,
    "epsilon_norm": 5.52431583404541,
    "prediction_norm": 5.0028767585754395,
    "timestamp": "2025-07-14T12:34:34.980713"
  },
  {
    "epoch": 1711,
    "batch_idx": 0,
    "loss": 0.17562855780124664,
    "learning_rate": 5.0641238794216495e-05,
    "timestep_mean": 541.53125,
    "epsilon_norm": 8.422429084777832,
    "prediction_norm": 7.989644527435303,
    "timestamp": "2025-07-14T12:34:34.985614"
  },
  {
    "epoch": 1711,
    "batch_idx": 32,
    "loss": 0.12247514724731445,
    "learning_rate": 5.0641238794216495e-05,
    "timestep_mean": 510.5625,
    "epsilon_norm": 6.856907367706299,
    "prediction_norm": 6.468127250671387,
    "timestamp": "2025-07-14T12:34:34.990560"
  },
  {
    "epoch": 1711,
    "batch_idx": 64,
    "loss": 0.18773260712623596,
    "learning_rate": 5.0641238794216495e-05,
    "timestep_mean": 487.6875,
    "epsilon_norm": 8.233590126037598,
    "prediction_norm": 7.769548416137695,
    "timestamp": "2025-07-14T12:34:34.995456"
  },
  {
    "epoch": 1711,
    "batch_idx": 96,
    "loss": 0.2506044805049896,
    "learning_rate": 5.0641238794216495e-05,
    "timestep_mean": 443.6875,
    "epsilon_norm": 7.317996501922607,
    "prediction_norm": 6.40873908996582,
    "timestamp": "2025-07-14T12:34:35.000454"
  },
  {
    "epoch": 1711,
    "batch_idx": 128,
    "loss": 0.16338367760181427,
    "learning_rate": 5.0641238794216495e-05,
    "timestep_mean": 528.34375,
    "epsilon_norm": 7.424388408660889,
    "prediction_norm": 6.98037576675415,
    "timestamp": "2025-07-14T12:34:35.005355"
  },
  {
    "epoch": 1711,
    "batch_idx": 160,
    "loss": 0.28973686695098877,
    "learning_rate": 5.0641238794216495e-05,
    "timestep_mean": 463.375,
    "epsilon_norm": 8.23510456085205,
    "prediction_norm": 6.821484565734863,
    "timestamp": "2025-07-14T12:34:35.010466"
  },
  {
    "epoch": 1711,
    "batch_idx": 192,
    "loss": 0.17078125476837158,
    "learning_rate": 5.0641238794216495e-05,
    "timestep_mean": 464.25,
    "epsilon_norm": 3.6385421752929688,
    "prediction_norm": 3.2018675804138184,
    "timestamp": "2025-07-14T12:34:35.015412"
  },
  {
    "epoch": 1712,
    "batch_idx": 0,
    "loss": 0.2453940510749817,
    "learning_rate": 5.0297374216814454e-05,
    "timestep_mean": 435.5,
    "epsilon_norm": 7.488610744476318,
    "prediction_norm": 6.389838218688965,
    "timestamp": "2025-07-14T12:34:35.020365"
  },
  {
    "epoch": 1712,
    "batch_idx": 32,
    "loss": 0.5093024373054504,
    "learning_rate": 5.0297374216814454e-05,
    "timestep_mean": 484.28125,
    "epsilon_norm": 9.408064842224121,
    "prediction_norm": 7.508689880371094,
    "timestamp": "2025-07-14T12:34:35.025361"
  },
  {
    "epoch": 1712,
    "batch_idx": 64,
    "loss": 0.23008403182029724,
    "learning_rate": 5.0297374216814454e-05,
    "timestep_mean": 485.0,
    "epsilon_norm": 7.56774377822876,
    "prediction_norm": 6.407665729522705,
    "timestamp": "2025-07-14T12:34:35.030341"
  },
  {
    "epoch": 1712,
    "batch_idx": 96,
    "loss": 0.20777475833892822,
    "learning_rate": 5.0297374216814454e-05,
    "timestep_mean": 478.46875,
    "epsilon_norm": 7.459657669067383,
    "prediction_norm": 6.598966121673584,
    "timestamp": "2025-07-14T12:34:35.035297"
  },
  {
    "epoch": 1712,
    "batch_idx": 128,
    "loss": 0.19237126410007477,
    "learning_rate": 5.0297374216814454e-05,
    "timestep_mean": 486.8125,
    "epsilon_norm": 7.173879623413086,
    "prediction_norm": 6.53273344039917,
    "timestamp": "2025-07-14T12:34:35.040227"
  },
  {
    "epoch": 1712,
    "batch_idx": 160,
    "loss": 0.14816871285438538,
    "learning_rate": 5.0297374216814454e-05,
    "timestep_mean": 607.875,
    "epsilon_norm": 7.912775993347168,
    "prediction_norm": 7.120392322540283,
    "timestamp": "2025-07-14T12:34:35.045183"
  },
  {
    "epoch": 1712,
    "batch_idx": 192,
    "loss": 0.3270109295845032,
    "learning_rate": 5.0297374216814454e-05,
    "timestep_mean": 492.25,
    "epsilon_norm": 2.8950746059417725,
    "prediction_norm": 2.8627586364746094,
    "timestamp": "2025-07-14T12:34:35.050315"
  },
  {
    "epoch": 1713,
    "batch_idx": 0,
    "loss": 0.24851159751415253,
    "learning_rate": 4.9954619235938076e-05,
    "timestep_mean": 490.625,
    "epsilon_norm": 7.643744468688965,
    "prediction_norm": 6.465060710906982,
    "timestamp": "2025-07-14T12:34:35.055238"
  },
  {
    "epoch": 1713,
    "batch_idx": 32,
    "loss": 0.348314493894577,
    "learning_rate": 4.9954619235938076e-05,
    "timestep_mean": 438.53125,
    "epsilon_norm": 8.578144073486328,
    "prediction_norm": 6.72969913482666,
    "timestamp": "2025-07-14T12:34:35.060267"
  },
  {
    "epoch": 1713,
    "batch_idx": 64,
    "loss": 0.3498942255973816,
    "learning_rate": 4.9954619235938076e-05,
    "timestep_mean": 497.53125,
    "epsilon_norm": 9.127311706542969,
    "prediction_norm": 7.227290630340576,
    "timestamp": "2025-07-14T12:34:35.065420"
  },
  {
    "epoch": 1713,
    "batch_idx": 96,
    "loss": 0.292949914932251,
    "learning_rate": 4.9954619235938076e-05,
    "timestep_mean": 499.8125,
    "epsilon_norm": 7.709773063659668,
    "prediction_norm": 6.992424964904785,
    "timestamp": "2025-07-14T12:34:35.070358"
  },
  {
    "epoch": 1713,
    "batch_idx": 128,
    "loss": 0.2811751663684845,
    "learning_rate": 4.9954619235938076e-05,
    "timestep_mean": 472.40625,
    "epsilon_norm": 8.317437171936035,
    "prediction_norm": 7.193823337554932,
    "timestamp": "2025-07-14T12:34:35.075357"
  },
  {
    "epoch": 1713,
    "batch_idx": 160,
    "loss": 0.23138752579689026,
    "learning_rate": 4.9954619235938076e-05,
    "timestep_mean": 578.59375,
    "epsilon_norm": 7.226123332977295,
    "prediction_norm": 6.1985602378845215,
    "timestamp": "2025-07-14T12:34:35.080282"
  },
  {
    "epoch": 1713,
    "batch_idx": 192,
    "loss": 0.3283461332321167,
    "learning_rate": 4.9954619235938076e-05,
    "timestep_mean": 443.125,
    "epsilon_norm": 4.597768783569336,
    "prediction_norm": 3.471506118774414,
    "timestamp": "2025-07-14T12:34:35.085239"
  },
  {
    "epoch": 1714,
    "batch_idx": 0,
    "loss": 0.3367258906364441,
    "learning_rate": 4.9612974697300964e-05,
    "timestep_mean": 476.28125,
    "epsilon_norm": 8.76913070678711,
    "prediction_norm": 7.334556579589844,
    "timestamp": "2025-07-14T12:34:35.090694"
  },
  {
    "epoch": 1714,
    "batch_idx": 32,
    "loss": 0.20106494426727295,
    "learning_rate": 4.9612974697300964e-05,
    "timestep_mean": 458.125,
    "epsilon_norm": 8.00777816772461,
    "prediction_norm": 7.032614231109619,
    "timestamp": "2025-07-14T12:34:35.100919"
  },
  {
    "epoch": 1714,
    "batch_idx": 64,
    "loss": 0.3956323266029358,
    "learning_rate": 4.9612974697300964e-05,
    "timestep_mean": 467.1875,
    "epsilon_norm": 8.169913291931152,
    "prediction_norm": 6.4522552490234375,
    "timestamp": "2025-07-14T12:34:35.105894"
  },
  {
    "epoch": 1714,
    "batch_idx": 96,
    "loss": 0.18586713075637817,
    "learning_rate": 4.9612974697300964e-05,
    "timestep_mean": 465.96875,
    "epsilon_norm": 5.897568702697754,
    "prediction_norm": 5.270359039306641,
    "timestamp": "2025-07-14T12:34:35.110680"
  },
  {
    "epoch": 1714,
    "batch_idx": 128,
    "loss": 0.20790991187095642,
    "learning_rate": 4.9612974697300964e-05,
    "timestep_mean": 542.59375,
    "epsilon_norm": 7.508737564086914,
    "prediction_norm": 6.62873649597168,
    "timestamp": "2025-07-14T12:34:35.115479"
  },
  {
    "epoch": 1714,
    "batch_idx": 160,
    "loss": 0.20862187445163727,
    "learning_rate": 4.9612974697300964e-05,
    "timestep_mean": 487.15625,
    "epsilon_norm": 7.882663726806641,
    "prediction_norm": 6.901674270629883,
    "timestamp": "2025-07-14T12:34:35.120210"
  },
  {
    "epoch": 1714,
    "batch_idx": 192,
    "loss": 0.09866271913051605,
    "learning_rate": 4.9612974697300964e-05,
    "timestep_mean": 517.75,
    "epsilon_norm": 3.644460916519165,
    "prediction_norm": 3.362241744995117,
    "timestamp": "2025-07-14T12:34:35.125092"
  },
  {
    "epoch": 1715,
    "batch_idx": 0,
    "loss": 0.5271000862121582,
    "learning_rate": 4.9272441443877206e-05,
    "timestep_mean": 470.65625,
    "epsilon_norm": 9.515438079833984,
    "prediction_norm": 7.449066162109375,
    "timestamp": "2025-07-14T12:34:35.129804"
  },
  {
    "epoch": 1715,
    "batch_idx": 32,
    "loss": 0.24770545959472656,
    "learning_rate": 4.9272441443877206e-05,
    "timestep_mean": 486.125,
    "epsilon_norm": 8.385658264160156,
    "prediction_norm": 7.448695182800293,
    "timestamp": "2025-07-14T12:34:35.134684"
  },
  {
    "epoch": 1715,
    "batch_idx": 64,
    "loss": 0.21583771705627441,
    "learning_rate": 4.9272441443877206e-05,
    "timestep_mean": 523.59375,
    "epsilon_norm": 7.718349456787109,
    "prediction_norm": 6.598504543304443,
    "timestamp": "2025-07-14T12:34:35.139548"
  },
  {
    "epoch": 1715,
    "batch_idx": 96,
    "loss": 0.2571518123149872,
    "learning_rate": 4.9272441443877206e-05,
    "timestep_mean": 511.78125,
    "epsilon_norm": 7.356627941131592,
    "prediction_norm": 5.715036869049072,
    "timestamp": "2025-07-14T12:34:35.144364"
  },
  {
    "epoch": 1715,
    "batch_idx": 128,
    "loss": 0.19237342476844788,
    "learning_rate": 4.9272441443877206e-05,
    "timestep_mean": 512.1875,
    "epsilon_norm": 7.972198486328125,
    "prediction_norm": 7.121445178985596,
    "timestamp": "2025-07-14T12:34:35.149105"
  },
  {
    "epoch": 1715,
    "batch_idx": 160,
    "loss": 0.41292282938957214,
    "learning_rate": 4.9272441443877206e-05,
    "timestep_mean": 391.875,
    "epsilon_norm": 8.8649263381958,
    "prediction_norm": 7.557140827178955,
    "timestamp": "2025-07-14T12:34:35.153858"
  },
  {
    "epoch": 1715,
    "batch_idx": 192,
    "loss": 0.09463192522525787,
    "learning_rate": 4.9272441443877206e-05,
    "timestep_mean": 509.0,
    "epsilon_norm": 3.4887490272521973,
    "prediction_norm": 3.201223611831665,
    "timestamp": "2025-07-14T12:34:35.158641"
  },
  {
    "epoch": 1716,
    "batch_idx": 0,
    "loss": 0.19544151425361633,
    "learning_rate": 4.8933020315898575e-05,
    "timestep_mean": 434.09375,
    "epsilon_norm": 8.05868148803711,
    "prediction_norm": 6.972146987915039,
    "timestamp": "2025-07-14T12:34:35.163395"
  },
  {
    "epoch": 1716,
    "batch_idx": 32,
    "loss": 0.524339497089386,
    "learning_rate": 4.8933020315898575e-05,
    "timestep_mean": 421.53125,
    "epsilon_norm": 9.002635955810547,
    "prediction_norm": 6.933599472045898,
    "timestamp": "2025-07-14T12:34:35.168111"
  },
  {
    "epoch": 1716,
    "batch_idx": 64,
    "loss": 0.13139639794826508,
    "learning_rate": 4.8933020315898575e-05,
    "timestep_mean": 535.03125,
    "epsilon_norm": 7.9984002113342285,
    "prediction_norm": 7.0021209716796875,
    "timestamp": "2025-07-14T12:34:35.173016"
  },
  {
    "epoch": 1716,
    "batch_idx": 96,
    "loss": 0.38471701741218567,
    "learning_rate": 4.8933020315898575e-05,
    "timestep_mean": 428.78125,
    "epsilon_norm": 8.938529014587402,
    "prediction_norm": 7.204469680786133,
    "timestamp": "2025-07-14T12:34:35.177901"
  },
  {
    "epoch": 1716,
    "batch_idx": 128,
    "loss": 0.23948416113853455,
    "learning_rate": 4.8933020315898575e-05,
    "timestep_mean": 537.0,
    "epsilon_norm": 8.654775619506836,
    "prediction_norm": 7.605612277984619,
    "timestamp": "2025-07-14T12:34:35.182752"
  },
  {
    "epoch": 1716,
    "batch_idx": 160,
    "loss": 0.18488863110542297,
    "learning_rate": 4.8933020315898575e-05,
    "timestep_mean": 564.75,
    "epsilon_norm": 8.343541145324707,
    "prediction_norm": 6.961553573608398,
    "timestamp": "2025-07-14T12:34:35.187541"
  },
  {
    "epoch": 1716,
    "batch_idx": 192,
    "loss": 0.09486192464828491,
    "learning_rate": 4.8933020315898575e-05,
    "timestep_mean": 666.625,
    "epsilon_norm": 3.817833185195923,
    "prediction_norm": 3.8520314693450928,
    "timestamp": "2025-07-14T12:34:35.192528"
  },
  {
    "epoch": 1717,
    "batch_idx": 0,
    "loss": 0.20710106194019318,
    "learning_rate": 4.8594712150853204e-05,
    "timestep_mean": 570.96875,
    "epsilon_norm": 8.563858985900879,
    "prediction_norm": 7.502902507781982,
    "timestamp": "2025-07-14T12:34:35.197433"
  },
  {
    "epoch": 1717,
    "batch_idx": 32,
    "loss": 0.16183029115200043,
    "learning_rate": 4.8594712150853204e-05,
    "timestep_mean": 539.1875,
    "epsilon_norm": 8.392176628112793,
    "prediction_norm": 7.318800926208496,
    "timestamp": "2025-07-14T12:34:35.202296"
  },
  {
    "epoch": 1717,
    "batch_idx": 64,
    "loss": 0.3005956709384918,
    "learning_rate": 4.8594712150853204e-05,
    "timestep_mean": 465.28125,
    "epsilon_norm": 7.377315044403076,
    "prediction_norm": 6.145567417144775,
    "timestamp": "2025-07-14T12:34:35.207122"
  },
  {
    "epoch": 1717,
    "batch_idx": 96,
    "loss": 0.569766640663147,
    "learning_rate": 4.8594712150853204e-05,
    "timestep_mean": 449.34375,
    "epsilon_norm": 8.152924537658691,
    "prediction_norm": 6.26374626159668,
    "timestamp": "2025-07-14T12:34:35.211938"
  },
  {
    "epoch": 1717,
    "batch_idx": 128,
    "loss": 0.2798584997653961,
    "learning_rate": 4.8594712150853204e-05,
    "timestep_mean": 468.1875,
    "epsilon_norm": 7.984710693359375,
    "prediction_norm": 6.635979175567627,
    "timestamp": "2025-07-14T12:34:35.216873"
  },
  {
    "epoch": 1717,
    "batch_idx": 160,
    "loss": 0.12406962364912033,
    "learning_rate": 4.8594712150853204e-05,
    "timestep_mean": 457.09375,
    "epsilon_norm": 6.6811203956604,
    "prediction_norm": 6.420548439025879,
    "timestamp": "2025-07-14T12:34:35.221715"
  },
  {
    "epoch": 1717,
    "batch_idx": 192,
    "loss": 0.06244182586669922,
    "learning_rate": 4.8594712150853204e-05,
    "timestep_mean": 589.375,
    "epsilon_norm": 3.069685935974121,
    "prediction_norm": 2.8079521656036377,
    "timestamp": "2025-07-14T12:34:35.226865"
  },
  {
    "epoch": 1718,
    "batch_idx": 0,
    "loss": 0.21278753876686096,
    "learning_rate": 4.825751778348258e-05,
    "timestep_mean": 467.8125,
    "epsilon_norm": 6.404202938079834,
    "prediction_norm": 5.261199951171875,
    "timestamp": "2025-07-14T12:34:35.231895"
  },
  {
    "epoch": 1718,
    "batch_idx": 32,
    "loss": 0.610164225101471,
    "learning_rate": 4.825751778348258e-05,
    "timestep_mean": 464.75,
    "epsilon_norm": 9.204085350036621,
    "prediction_norm": 6.985787868499756,
    "timestamp": "2025-07-14T12:34:35.236799"
  },
  {
    "epoch": 1718,
    "batch_idx": 64,
    "loss": 0.214439257979393,
    "learning_rate": 4.825751778348258e-05,
    "timestep_mean": 503.40625,
    "epsilon_norm": 7.775599002838135,
    "prediction_norm": 7.330544948577881,
    "timestamp": "2025-07-14T12:34:35.241557"
  },
  {
    "epoch": 1718,
    "batch_idx": 96,
    "loss": 0.32793158292770386,
    "learning_rate": 4.825751778348258e-05,
    "timestep_mean": 505.8125,
    "epsilon_norm": 7.6466264724731445,
    "prediction_norm": 6.197833061218262,
    "timestamp": "2025-07-14T12:34:35.246685"
  },
  {
    "epoch": 1718,
    "batch_idx": 128,
    "loss": 0.13493356108665466,
    "learning_rate": 4.825751778348258e-05,
    "timestep_mean": 576.625,
    "epsilon_norm": 8.531451225280762,
    "prediction_norm": 7.6764912605285645,
    "timestamp": "2025-07-14T12:34:35.251560"
  },
  {
    "epoch": 1718,
    "batch_idx": 160,
    "loss": 0.1681526005268097,
    "learning_rate": 4.825751778348258e-05,
    "timestep_mean": 530.96875,
    "epsilon_norm": 8.42512321472168,
    "prediction_norm": 8.049112319946289,
    "timestamp": "2025-07-14T12:34:35.256534"
  },
  {
    "epoch": 1718,
    "batch_idx": 192,
    "loss": 0.44425851106643677,
    "learning_rate": 4.825751778348258e-05,
    "timestep_mean": 553.875,
    "epsilon_norm": 4.623603820800781,
    "prediction_norm": 3.9684746265411377,
    "timestamp": "2025-07-14T12:34:35.261616"
  },
  {
    "epoch": 1719,
    "batch_idx": 0,
    "loss": 0.17036454379558563,
    "learning_rate": 4.792143804578042e-05,
    "timestep_mean": 549.0,
    "epsilon_norm": 8.467503547668457,
    "prediction_norm": 7.841394901275635,
    "timestamp": "2025-07-14T12:34:35.266603"
  },
  {
    "epoch": 1719,
    "batch_idx": 32,
    "loss": 0.331123411655426,
    "learning_rate": 4.792143804578042e-05,
    "timestep_mean": 470.28125,
    "epsilon_norm": 7.07814884185791,
    "prediction_norm": 5.573996067047119,
    "timestamp": "2025-07-14T12:34:35.271665"
  },
  {
    "epoch": 1719,
    "batch_idx": 64,
    "loss": 0.250506728887558,
    "learning_rate": 4.792143804578042e-05,
    "timestep_mean": 629.0625,
    "epsilon_norm": 7.346162796020508,
    "prediction_norm": 6.639707088470459,
    "timestamp": "2025-07-14T12:34:35.276522"
  },
  {
    "epoch": 1719,
    "batch_idx": 96,
    "loss": 0.3184155821800232,
    "learning_rate": 4.792143804578042e-05,
    "timestep_mean": 494.28125,
    "epsilon_norm": 7.510471820831299,
    "prediction_norm": 6.290432453155518,
    "timestamp": "2025-07-14T12:34:35.281453"
  },
  {
    "epoch": 1719,
    "batch_idx": 128,
    "loss": 0.18375620245933533,
    "learning_rate": 4.792143804578042e-05,
    "timestep_mean": 489.65625,
    "epsilon_norm": 7.820928573608398,
    "prediction_norm": 7.249596118927002,
    "timestamp": "2025-07-14T12:34:35.286410"
  },
  {
    "epoch": 1719,
    "batch_idx": 160,
    "loss": 0.19391894340515137,
    "learning_rate": 4.792143804578042e-05,
    "timestep_mean": 553.875,
    "epsilon_norm": 7.907863140106201,
    "prediction_norm": 7.119624614715576,
    "timestamp": "2025-07-14T12:34:35.291430"
  },
  {
    "epoch": 1719,
    "batch_idx": 192,
    "loss": 0.10218801349401474,
    "learning_rate": 4.792143804578042e-05,
    "timestep_mean": 598.625,
    "epsilon_norm": 4.251313209533691,
    "prediction_norm": 4.225580215454102,
    "timestamp": "2025-07-14T12:34:35.296465"
  },
  {
    "epoch": 1720,
    "batch_idx": 0,
    "loss": 0.1730133444070816,
    "learning_rate": 4.758647376699033e-05,
    "timestep_mean": 515.75,
    "epsilon_norm": 7.410964488983154,
    "prediction_norm": 6.778427600860596,
    "timestamp": "2025-07-14T12:34:35.301349"
  },
  {
    "epoch": 1720,
    "batch_idx": 32,
    "loss": 0.22218172252178192,
    "learning_rate": 4.758647376699033e-05,
    "timestep_mean": 533.625,
    "epsilon_norm": 8.071538925170898,
    "prediction_norm": 7.035384178161621,
    "timestamp": "2025-07-14T12:34:35.306310"
  },
  {
    "epoch": 1720,
    "batch_idx": 64,
    "loss": 0.1778024137020111,
    "learning_rate": 4.758647376699033e-05,
    "timestep_mean": 535.375,
    "epsilon_norm": 6.419294834136963,
    "prediction_norm": 5.874788761138916,
    "timestamp": "2025-07-14T12:34:35.311277"
  },
  {
    "epoch": 1720,
    "batch_idx": 96,
    "loss": 0.3338990807533264,
    "learning_rate": 4.758647376699033e-05,
    "timestep_mean": 494.5625,
    "epsilon_norm": 7.849956035614014,
    "prediction_norm": 6.04940938949585,
    "timestamp": "2025-07-14T12:34:35.316218"
  },
  {
    "epoch": 1720,
    "batch_idx": 128,
    "loss": 0.44368770718574524,
    "learning_rate": 4.758647376699033e-05,
    "timestep_mean": 493.6875,
    "epsilon_norm": 7.286344051361084,
    "prediction_norm": 5.478808879852295,
    "timestamp": "2025-07-14T12:34:35.321235"
  },
  {
    "epoch": 1720,
    "batch_idx": 160,
    "loss": 0.30455493927001953,
    "learning_rate": 4.758647376699033e-05,
    "timestep_mean": 525.90625,
    "epsilon_norm": 7.977169990539551,
    "prediction_norm": 7.451815605163574,
    "timestamp": "2025-07-14T12:34:35.326220"
  },
  {
    "epoch": 1720,
    "batch_idx": 192,
    "loss": 0.26300275325775146,
    "learning_rate": 4.758647376699033e-05,
    "timestep_mean": 457.0,
    "epsilon_norm": 4.216249465942383,
    "prediction_norm": 3.9026482105255127,
    "timestamp": "2025-07-14T12:34:35.331439"
  },
  {
    "epoch": 1721,
    "batch_idx": 0,
    "loss": 0.2531101107597351,
    "learning_rate": 4.7252625773602845e-05,
    "timestep_mean": 448.15625,
    "epsilon_norm": 7.5840935707092285,
    "prediction_norm": 6.737870693206787,
    "timestamp": "2025-07-14T12:34:35.336714"
  },
  {
    "epoch": 1721,
    "batch_idx": 32,
    "loss": 0.4981409013271332,
    "learning_rate": 4.7252625773602845e-05,
    "timestep_mean": 472.71875,
    "epsilon_norm": 8.235648155212402,
    "prediction_norm": 6.673821449279785,
    "timestamp": "2025-07-14T12:34:35.341825"
  },
  {
    "epoch": 1721,
    "batch_idx": 64,
    "loss": 0.32224392890930176,
    "learning_rate": 4.7252625773602845e-05,
    "timestep_mean": 481.21875,
    "epsilon_norm": 8.006006240844727,
    "prediction_norm": 6.229897975921631,
    "timestamp": "2025-07-14T12:34:35.346934"
  },
  {
    "epoch": 1721,
    "batch_idx": 96,
    "loss": 0.2283109724521637,
    "learning_rate": 4.7252625773602845e-05,
    "timestep_mean": 569.0,
    "epsilon_norm": 8.090984344482422,
    "prediction_norm": 6.784979343414307,
    "timestamp": "2025-07-14T12:34:35.352283"
  },
  {
    "epoch": 1721,
    "batch_idx": 128,
    "loss": 0.1831500381231308,
    "learning_rate": 4.7252625773602845e-05,
    "timestep_mean": 522.6875,
    "epsilon_norm": 7.220944881439209,
    "prediction_norm": 5.926555633544922,
    "timestamp": "2025-07-14T12:34:35.357478"
  },
  {
    "epoch": 1721,
    "batch_idx": 160,
    "loss": 0.1848594844341278,
    "learning_rate": 4.7252625773602845e-05,
    "timestep_mean": 510.15625,
    "epsilon_norm": 7.217758655548096,
    "prediction_norm": 6.570580005645752,
    "timestamp": "2025-07-14T12:34:35.362721"
  },
  {
    "epoch": 1721,
    "batch_idx": 192,
    "loss": 0.39388060569763184,
    "learning_rate": 4.7252625773602845e-05,
    "timestep_mean": 444.5,
    "epsilon_norm": 3.7732093334198,
    "prediction_norm": 2.6292715072631836,
    "timestamp": "2025-07-14T12:34:35.367858"
  },
  {
    "epoch": 1722,
    "batch_idx": 0,
    "loss": 0.19101104140281677,
    "learning_rate": 4.691989488935517e-05,
    "timestep_mean": 506.46875,
    "epsilon_norm": 7.642571449279785,
    "prediction_norm": 7.450702667236328,
    "timestamp": "2025-07-14T12:34:35.373031"
  },
  {
    "epoch": 1722,
    "batch_idx": 32,
    "loss": 0.11397206783294678,
    "learning_rate": 4.691989488935517e-05,
    "timestep_mean": 441.46875,
    "epsilon_norm": 7.430725574493408,
    "prediction_norm": 7.1021342277526855,
    "timestamp": "2025-07-14T12:34:35.378246"
  },
  {
    "epoch": 1722,
    "batch_idx": 64,
    "loss": 0.6748888492584229,
    "learning_rate": 4.691989488935517e-05,
    "timestep_mean": 469.875,
    "epsilon_norm": 9.407689094543457,
    "prediction_norm": 7.032626152038574,
    "timestamp": "2025-07-14T12:34:35.383722"
  },
  {
    "epoch": 1722,
    "batch_idx": 96,
    "loss": 0.1162635087966919,
    "learning_rate": 4.691989488935517e-05,
    "timestep_mean": 500.90625,
    "epsilon_norm": 7.652526378631592,
    "prediction_norm": 6.946058750152588,
    "timestamp": "2025-07-14T12:34:35.392790"
  },
  {
    "epoch": 1722,
    "batch_idx": 128,
    "loss": 0.21620279550552368,
    "learning_rate": 4.691989488935517e-05,
    "timestep_mean": 612.875,
    "epsilon_norm": 8.215438842773438,
    "prediction_norm": 6.998149394989014,
    "timestamp": "2025-07-14T12:34:35.402055"
  },
  {
    "epoch": 1722,
    "batch_idx": 160,
    "loss": 0.39174890518188477,
    "learning_rate": 4.691989488935517e-05,
    "timestep_mean": 528.4375,
    "epsilon_norm": 7.590825080871582,
    "prediction_norm": 6.940133571624756,
    "timestamp": "2025-07-14T12:34:35.409361"
  },
  {
    "epoch": 1722,
    "batch_idx": 192,
    "loss": 0.5412415862083435,
    "learning_rate": 4.691989488935517e-05,
    "timestep_mean": 329.75,
    "epsilon_norm": 3.7652862071990967,
    "prediction_norm": 2.363906145095825,
    "timestamp": "2025-07-14T12:34:35.417279"
  },
  {
    "epoch": 1723,
    "batch_idx": 0,
    "loss": 0.12449803948402405,
    "learning_rate": 4.658828193522735e-05,
    "timestep_mean": 560.5625,
    "epsilon_norm": 7.711906909942627,
    "prediction_norm": 7.37283182144165,
    "timestamp": "2025-07-14T12:34:35.425351"
  },
  {
    "epoch": 1723,
    "batch_idx": 32,
    "loss": 0.24515891075134277,
    "learning_rate": 4.658828193522735e-05,
    "timestep_mean": 549.03125,
    "epsilon_norm": 9.333888053894043,
    "prediction_norm": 8.41636848449707,
    "timestamp": "2025-07-14T12:34:35.432123"
  },
  {
    "epoch": 1723,
    "batch_idx": 64,
    "loss": 0.5238928198814392,
    "learning_rate": 4.658828193522735e-05,
    "timestep_mean": 480.84375,
    "epsilon_norm": 8.225379943847656,
    "prediction_norm": 5.852619647979736,
    "timestamp": "2025-07-14T12:34:35.439138"
  },
  {
    "epoch": 1723,
    "batch_idx": 96,
    "loss": 0.35769185423851013,
    "learning_rate": 4.658828193522735e-05,
    "timestep_mean": 479.4375,
    "epsilon_norm": 7.50342321395874,
    "prediction_norm": 6.49493408203125,
    "timestamp": "2025-07-14T12:34:35.446734"
  },
  {
    "epoch": 1723,
    "batch_idx": 128,
    "loss": 0.2889588475227356,
    "learning_rate": 4.658828193522735e-05,
    "timestep_mean": 492.0625,
    "epsilon_norm": 7.337113380432129,
    "prediction_norm": 5.738515853881836,
    "timestamp": "2025-07-14T12:34:35.452776"
  },
  {
    "epoch": 1723,
    "batch_idx": 160,
    "loss": 0.09716162085533142,
    "learning_rate": 4.658828193522735e-05,
    "timestep_mean": 544.0,
    "epsilon_norm": 7.654533386230469,
    "prediction_norm": 7.300685405731201,
    "timestamp": "2025-07-14T12:34:35.460536"
  },
  {
    "epoch": 1723,
    "batch_idx": 192,
    "loss": 0.11875561624765396,
    "learning_rate": 4.658828193522735e-05,
    "timestep_mean": 532.875,
    "epsilon_norm": 3.66690993309021,
    "prediction_norm": 3.2838213443756104,
    "timestamp": "2025-07-14T12:34:35.472004"
  },
  {
    "epoch": 1724,
    "batch_idx": 0,
    "loss": 0.4524902105331421,
    "learning_rate": 4.6257787729441505e-05,
    "timestep_mean": 464.53125,
    "epsilon_norm": 8.901110649108887,
    "prediction_norm": 7.546508312225342,
    "timestamp": "2025-07-14T12:34:35.480624"
  },
  {
    "epoch": 1724,
    "batch_idx": 32,
    "loss": 0.28724241256713867,
    "learning_rate": 4.6257787729441505e-05,
    "timestep_mean": 447.15625,
    "epsilon_norm": 9.787552833557129,
    "prediction_norm": 8.598731994628906,
    "timestamp": "2025-07-14T12:34:35.487945"
  },
  {
    "epoch": 1724,
    "batch_idx": 64,
    "loss": 0.11616604775190353,
    "learning_rate": 4.6257787729441505e-05,
    "timestep_mean": 505.1875,
    "epsilon_norm": 7.207663536071777,
    "prediction_norm": 6.6772027015686035,
    "timestamp": "2025-07-14T12:34:35.493228"
  },
  {
    "epoch": 1724,
    "batch_idx": 96,
    "loss": 0.45701247453689575,
    "learning_rate": 4.6257787729441505e-05,
    "timestep_mean": 432.375,
    "epsilon_norm": 9.338542938232422,
    "prediction_norm": 7.050382137298584,
    "timestamp": "2025-07-14T12:34:35.498253"
  },
  {
    "epoch": 1724,
    "batch_idx": 128,
    "loss": 0.29465094208717346,
    "learning_rate": 4.6257787729441505e-05,
    "timestep_mean": 469.71875,
    "epsilon_norm": 7.525106430053711,
    "prediction_norm": 6.2176833152771,
    "timestamp": "2025-07-14T12:34:35.503248"
  },
  {
    "epoch": 1724,
    "batch_idx": 160,
    "loss": 0.14688831567764282,
    "learning_rate": 4.6257787729441505e-05,
    "timestep_mean": 511.9375,
    "epsilon_norm": 7.066948890686035,
    "prediction_norm": 6.520021915435791,
    "timestamp": "2025-07-14T12:34:35.508496"
  },
  {
    "epoch": 1724,
    "batch_idx": 192,
    "loss": 0.08814139664173126,
    "learning_rate": 4.6257787729441505e-05,
    "timestep_mean": 450.875,
    "epsilon_norm": 3.0886831283569336,
    "prediction_norm": 2.9132111072540283,
    "timestamp": "2025-07-14T12:34:35.513801"
  },
  {
    "epoch": 1725,
    "batch_idx": 0,
    "loss": 0.19269782304763794,
    "learning_rate": 4.592841308745943e-05,
    "timestep_mean": 496.53125,
    "epsilon_norm": 7.436045169830322,
    "prediction_norm": 6.374675750732422,
    "timestamp": "2025-07-14T12:34:35.518855"
  },
  {
    "epoch": 1725,
    "batch_idx": 32,
    "loss": 0.40385809540748596,
    "learning_rate": 4.592841308745943e-05,
    "timestep_mean": 448.65625,
    "epsilon_norm": 8.2611083984375,
    "prediction_norm": 6.452445030212402,
    "timestamp": "2025-07-14T12:34:35.524016"
  },
  {
    "epoch": 1725,
    "batch_idx": 64,
    "loss": 0.28729307651519775,
    "learning_rate": 4.592841308745943e-05,
    "timestep_mean": 471.5,
    "epsilon_norm": 8.50204086303711,
    "prediction_norm": 7.248133659362793,
    "timestamp": "2025-07-14T12:34:35.529121"
  },
  {
    "epoch": 1725,
    "batch_idx": 96,
    "loss": 0.18641497194766998,
    "learning_rate": 4.592841308745943e-05,
    "timestep_mean": 582.8125,
    "epsilon_norm": 7.371140480041504,
    "prediction_norm": 6.019619941711426,
    "timestamp": "2025-07-14T12:34:35.534395"
  },
  {
    "epoch": 1725,
    "batch_idx": 128,
    "loss": 0.3661677837371826,
    "learning_rate": 4.592841308745943e-05,
    "timestep_mean": 493.84375,
    "epsilon_norm": 7.845788955688477,
    "prediction_norm": 6.607600212097168,
    "timestamp": "2025-07-14T12:34:35.539490"
  },
  {
    "epoch": 1725,
    "batch_idx": 160,
    "loss": 0.1701478809118271,
    "learning_rate": 4.592841308745943e-05,
    "timestep_mean": 568.03125,
    "epsilon_norm": 8.574630737304688,
    "prediction_norm": 8.034126281738281,
    "timestamp": "2025-07-14T12:34:35.544645"
  },
  {
    "epoch": 1725,
    "batch_idx": 192,
    "loss": 0.12313657253980637,
    "learning_rate": 4.592841308745943e-05,
    "timestep_mean": 608.125,
    "epsilon_norm": 3.3262789249420166,
    "prediction_norm": 3.3920631408691406,
    "timestamp": "2025-07-14T12:34:35.549758"
  },
  {
    "epoch": 1726,
    "batch_idx": 0,
    "loss": 0.2961888313293457,
    "learning_rate": 4.560015882197993e-05,
    "timestep_mean": 515.5,
    "epsilon_norm": 8.024420738220215,
    "prediction_norm": 7.126372337341309,
    "timestamp": "2025-07-14T12:34:35.555032"
  },
  {
    "epoch": 1726,
    "batch_idx": 32,
    "loss": 0.23425042629241943,
    "learning_rate": 4.560015882197993e-05,
    "timestep_mean": 476.125,
    "epsilon_norm": 7.478793621063232,
    "prediction_norm": 5.671377182006836,
    "timestamp": "2025-07-14T12:34:35.560099"
  },
  {
    "epoch": 1726,
    "batch_idx": 64,
    "loss": 0.3167194724082947,
    "learning_rate": 4.560015882197993e-05,
    "timestep_mean": 439.25,
    "epsilon_norm": 7.868887424468994,
    "prediction_norm": 6.164743900299072,
    "timestamp": "2025-07-14T12:34:35.565130"
  },
  {
    "epoch": 1726,
    "batch_idx": 96,
    "loss": 0.23774895071983337,
    "learning_rate": 4.560015882197993e-05,
    "timestep_mean": 531.75,
    "epsilon_norm": 6.930499076843262,
    "prediction_norm": 6.304894924163818,
    "timestamp": "2025-07-14T12:34:35.570315"
  },
  {
    "epoch": 1726,
    "batch_idx": 128,
    "loss": 0.1300584077835083,
    "learning_rate": 4.560015882197993e-05,
    "timestep_mean": 589.4375,
    "epsilon_norm": 7.6272125244140625,
    "prediction_norm": 6.761503219604492,
    "timestamp": "2025-07-14T12:34:35.575568"
  },
  {
    "epoch": 1726,
    "batch_idx": 160,
    "loss": 0.18291078507900238,
    "learning_rate": 4.560015882197993e-05,
    "timestep_mean": 598.125,
    "epsilon_norm": 7.84639310836792,
    "prediction_norm": 6.873179912567139,
    "timestamp": "2025-07-14T12:34:35.580801"
  },
  {
    "epoch": 1726,
    "batch_idx": 192,
    "loss": 0.21236902475357056,
    "learning_rate": 4.560015882197993e-05,
    "timestep_mean": 587.5,
    "epsilon_norm": 4.46524715423584,
    "prediction_norm": 4.060328483581543,
    "timestamp": "2025-07-14T12:34:35.585967"
  },
  {
    "epoch": 1727,
    "batch_idx": 0,
    "loss": 0.334722638130188,
    "learning_rate": 4.527302574293813e-05,
    "timestep_mean": 493.78125,
    "epsilon_norm": 8.020983695983887,
    "prediction_norm": 6.624623775482178,
    "timestamp": "2025-07-14T12:34:35.591275"
  },
  {
    "epoch": 1727,
    "batch_idx": 32,
    "loss": 0.18183234333992004,
    "learning_rate": 4.527302574293813e-05,
    "timestep_mean": 511.5,
    "epsilon_norm": 7.565996170043945,
    "prediction_norm": 6.815685749053955,
    "timestamp": "2025-07-14T12:34:35.596497"
  },
  {
    "epoch": 1727,
    "batch_idx": 64,
    "loss": 0.42751166224479675,
    "learning_rate": 4.527302574293813e-05,
    "timestep_mean": 440.625,
    "epsilon_norm": 8.125202178955078,
    "prediction_norm": 6.837878704071045,
    "timestamp": "2025-07-14T12:34:35.601612"
  },
  {
    "epoch": 1727,
    "batch_idx": 96,
    "loss": 0.3245552182197571,
    "learning_rate": 4.527302574293813e-05,
    "timestep_mean": 459.125,
    "epsilon_norm": 8.191259384155273,
    "prediction_norm": 6.71997594833374,
    "timestamp": "2025-07-14T12:34:35.606782"
  },
  {
    "epoch": 1727,
    "batch_idx": 128,
    "loss": 0.33908313512802124,
    "learning_rate": 4.527302574293813e-05,
    "timestep_mean": 420.625,
    "epsilon_norm": 6.802016258239746,
    "prediction_norm": 6.17564058303833,
    "timestamp": "2025-07-14T12:34:35.612006"
  },
  {
    "epoch": 1727,
    "batch_idx": 160,
    "loss": 0.4318946301937103,
    "learning_rate": 4.527302574293813e-05,
    "timestep_mean": 432.03125,
    "epsilon_norm": 7.890121936798096,
    "prediction_norm": 5.708915710449219,
    "timestamp": "2025-07-14T12:34:35.617316"
  },
  {
    "epoch": 1727,
    "batch_idx": 192,
    "loss": 0.13221371173858643,
    "learning_rate": 4.527302574293813e-05,
    "timestep_mean": 496.625,
    "epsilon_norm": 4.067806243896484,
    "prediction_norm": 3.8686935901641846,
    "timestamp": "2025-07-14T12:34:35.622450"
  },
  {
    "epoch": 1728,
    "batch_idx": 0,
    "loss": 0.24690672755241394,
    "learning_rate": 4.4947014657502176e-05,
    "timestep_mean": 562.84375,
    "epsilon_norm": 7.634829521179199,
    "prediction_norm": 7.2161946296691895,
    "timestamp": "2025-07-14T12:34:35.627553"
  },
  {
    "epoch": 1728,
    "batch_idx": 32,
    "loss": 0.20430435240268707,
    "learning_rate": 4.4947014657502176e-05,
    "timestep_mean": 575.71875,
    "epsilon_norm": 8.50597858428955,
    "prediction_norm": 7.043398857116699,
    "timestamp": "2025-07-14T12:34:35.632768"
  },
  {
    "epoch": 1728,
    "batch_idx": 64,
    "loss": 0.36737746000289917,
    "learning_rate": 4.4947014657502176e-05,
    "timestep_mean": 462.1875,
    "epsilon_norm": 8.872563362121582,
    "prediction_norm": 7.801595211029053,
    "timestamp": "2025-07-14T12:34:35.638002"
  },
  {
    "epoch": 1728,
    "batch_idx": 96,
    "loss": 0.19388936460018158,
    "learning_rate": 4.4947014657502176e-05,
    "timestep_mean": 484.09375,
    "epsilon_norm": 7.945619106292725,
    "prediction_norm": 6.93372917175293,
    "timestamp": "2025-07-14T12:34:35.643088"
  },
  {
    "epoch": 1728,
    "batch_idx": 128,
    "loss": 0.22479671239852905,
    "learning_rate": 4.4947014657502176e-05,
    "timestep_mean": 479.65625,
    "epsilon_norm": 9.197114944458008,
    "prediction_norm": 7.842693328857422,
    "timestamp": "2025-07-14T12:34:35.648347"
  },
  {
    "epoch": 1728,
    "batch_idx": 160,
    "loss": 0.2364020198583603,
    "learning_rate": 4.4947014657502176e-05,
    "timestep_mean": 478.25,
    "epsilon_norm": 8.755463600158691,
    "prediction_norm": 7.02946662902832,
    "timestamp": "2025-07-14T12:34:35.653598"
  },
  {
    "epoch": 1728,
    "batch_idx": 192,
    "loss": 0.0576997846364975,
    "learning_rate": 4.4947014657502176e-05,
    "timestep_mean": 672.125,
    "epsilon_norm": 3.7505781650543213,
    "prediction_norm": 3.442643880844116,
    "timestamp": "2025-07-14T12:34:35.658870"
  },
  {
    "epoch": 1729,
    "batch_idx": 0,
    "loss": 0.1911398321390152,
    "learning_rate": 4.462212637007213e-05,
    "timestep_mean": 495.84375,
    "epsilon_norm": 8.098982810974121,
    "prediction_norm": 7.552119255065918,
    "timestamp": "2025-07-14T12:34:35.664067"
  },
  {
    "epoch": 1729,
    "batch_idx": 32,
    "loss": 0.45694851875305176,
    "learning_rate": 4.462212637007213e-05,
    "timestep_mean": 511.59375,
    "epsilon_norm": 8.655961036682129,
    "prediction_norm": 7.103768348693848,
    "timestamp": "2025-07-14T12:34:35.669182"
  },
  {
    "epoch": 1729,
    "batch_idx": 64,
    "loss": 0.12410838156938553,
    "learning_rate": 4.462212637007213e-05,
    "timestep_mean": 568.6875,
    "epsilon_norm": 8.781411170959473,
    "prediction_norm": 7.995504379272461,
    "timestamp": "2025-07-14T12:34:35.674340"
  },
  {
    "epoch": 1729,
    "batch_idx": 96,
    "loss": 0.21983850002288818,
    "learning_rate": 4.462212637007213e-05,
    "timestep_mean": 538.9375,
    "epsilon_norm": 8.030088424682617,
    "prediction_norm": 6.6548004150390625,
    "timestamp": "2025-07-14T12:34:35.679567"
  },
  {
    "epoch": 1729,
    "batch_idx": 128,
    "loss": 0.45076191425323486,
    "learning_rate": 4.462212637007213e-05,
    "timestep_mean": 436.4375,
    "epsilon_norm": 7.822466850280762,
    "prediction_norm": 6.301429748535156,
    "timestamp": "2025-07-14T12:34:35.684769"
  },
  {
    "epoch": 1729,
    "batch_idx": 160,
    "loss": 0.3876245617866516,
    "learning_rate": 4.462212637007213e-05,
    "timestep_mean": 465.75,
    "epsilon_norm": 8.18907356262207,
    "prediction_norm": 6.625848293304443,
    "timestamp": "2025-07-14T12:34:35.689885"
  },
  {
    "epoch": 1729,
    "batch_idx": 192,
    "loss": 0.08819177746772766,
    "learning_rate": 4.462212637007213e-05,
    "timestep_mean": 697.875,
    "epsilon_norm": 2.379910469055176,
    "prediction_norm": 2.436511993408203,
    "timestamp": "2025-07-14T12:34:35.694967"
  },
  {
    "epoch": 1730,
    "batch_idx": 0,
    "loss": 0.23081821203231812,
    "learning_rate": 4.4298361682277364e-05,
    "timestep_mean": 486.375,
    "epsilon_norm": 8.07812786102295,
    "prediction_norm": 6.994115352630615,
    "timestamp": "2025-07-14T12:34:35.700350"
  },
  {
    "epoch": 1730,
    "batch_idx": 32,
    "loss": 0.14498576521873474,
    "learning_rate": 4.4298361682277364e-05,
    "timestep_mean": 489.65625,
    "epsilon_norm": 7.395905017852783,
    "prediction_norm": 6.277708530426025,
    "timestamp": "2025-07-14T12:34:35.705507"
  },
  {
    "epoch": 1730,
    "batch_idx": 64,
    "loss": 0.1529703140258789,
    "learning_rate": 4.4298361682277364e-05,
    "timestep_mean": 610.53125,
    "epsilon_norm": 9.075603485107422,
    "prediction_norm": 8.367584228515625,
    "timestamp": "2025-07-14T12:34:35.710736"
  },
  {
    "epoch": 1730,
    "batch_idx": 96,
    "loss": 0.14251437783241272,
    "learning_rate": 4.4298361682277364e-05,
    "timestep_mean": 601.96875,
    "epsilon_norm": 8.26242446899414,
    "prediction_norm": 7.984019756317139,
    "timestamp": "2025-07-14T12:34:35.715864"
  },
  {
    "epoch": 1730,
    "batch_idx": 128,
    "loss": 0.1501472145318985,
    "learning_rate": 4.4298361682277364e-05,
    "timestep_mean": 525.15625,
    "epsilon_norm": 7.3206682205200195,
    "prediction_norm": 7.012246131896973,
    "timestamp": "2025-07-14T12:34:35.720933"
  },
  {
    "epoch": 1730,
    "batch_idx": 160,
    "loss": 0.17192184925079346,
    "learning_rate": 4.4298361682277364e-05,
    "timestep_mean": 525.75,
    "epsilon_norm": 7.296710968017578,
    "prediction_norm": 6.5806193351745605,
    "timestamp": "2025-07-14T12:34:35.728789"
  },
  {
    "epoch": 1730,
    "batch_idx": 192,
    "loss": 0.03353121876716614,
    "learning_rate": 4.4298361682277364e-05,
    "timestep_mean": 666.375,
    "epsilon_norm": 4.424522876739502,
    "prediction_norm": 4.412160873413086,
    "timestamp": "2025-07-14T12:34:35.735765"
  },
  {
    "epoch": 1731,
    "batch_idx": 0,
    "loss": 0.30781275033950806,
    "learning_rate": 4.3975721392975295e-05,
    "timestep_mean": 519.625,
    "epsilon_norm": 7.813992977142334,
    "prediction_norm": 6.829104900360107,
    "timestamp": "2025-07-14T12:34:35.742541"
  },
  {
    "epoch": 1731,
    "batch_idx": 32,
    "loss": 0.27780425548553467,
    "learning_rate": 4.3975721392975295e-05,
    "timestep_mean": 514.71875,
    "epsilon_norm": 8.84876537322998,
    "prediction_norm": 7.929931163787842,
    "timestamp": "2025-07-14T12:34:35.750681"
  },
  {
    "epoch": 1731,
    "batch_idx": 64,
    "loss": 0.22013403475284576,
    "learning_rate": 4.3975721392975295e-05,
    "timestep_mean": 486.0,
    "epsilon_norm": 7.505146026611328,
    "prediction_norm": 6.327320575714111,
    "timestamp": "2025-07-14T12:34:35.757117"
  },
  {
    "epoch": 1731,
    "batch_idx": 96,
    "loss": 0.5625324249267578,
    "learning_rate": 4.3975721392975295e-05,
    "timestep_mean": 490.0625,
    "epsilon_norm": 8.287498474121094,
    "prediction_norm": 6.342773914337158,
    "timestamp": "2025-07-14T12:34:35.763712"
  },
  {
    "epoch": 1731,
    "batch_idx": 128,
    "loss": 0.1844072788953781,
    "learning_rate": 4.3975721392975295e-05,
    "timestep_mean": 569.125,
    "epsilon_norm": 8.603975296020508,
    "prediction_norm": 7.025274276733398,
    "timestamp": "2025-07-14T12:34:35.770138"
  },
  {
    "epoch": 1731,
    "batch_idx": 160,
    "loss": 0.36880797147750854,
    "learning_rate": 4.3975721392975295e-05,
    "timestep_mean": 428.15625,
    "epsilon_norm": 9.16140365600586,
    "prediction_norm": 7.871310710906982,
    "timestamp": "2025-07-14T12:34:35.776658"
  },
  {
    "epoch": 1731,
    "batch_idx": 192,
    "loss": 0.5347887873649597,
    "learning_rate": 4.3975721392975295e-05,
    "timestep_mean": 541.375,
    "epsilon_norm": 3.5761771202087402,
    "prediction_norm": 2.7866084575653076,
    "timestamp": "2025-07-14T12:34:35.782551"
  },
  {
    "epoch": 1732,
    "batch_idx": 0,
    "loss": 0.3861720561981201,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 406.25,
    "epsilon_norm": 8.433547973632812,
    "prediction_norm": 6.290695667266846,
    "timestamp": "2025-07-14T12:34:35.788126"
  },
  {
    "epoch": 1732,
    "batch_idx": 32,
    "loss": 0.17106367647647858,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 509.375,
    "epsilon_norm": 7.23956823348999,
    "prediction_norm": 6.374200820922852,
    "timestamp": "2025-07-14T12:34:35.793322"
  },
  {
    "epoch": 1732,
    "batch_idx": 64,
    "loss": 0.25503453612327576,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 498.15625,
    "epsilon_norm": 7.863293170928955,
    "prediction_norm": 6.113106727600098,
    "timestamp": "2025-07-14T12:34:35.798515"
  },
  {
    "epoch": 1732,
    "batch_idx": 96,
    "loss": 0.3405953645706177,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 462.25,
    "epsilon_norm": 8.458855628967285,
    "prediction_norm": 7.2278289794921875,
    "timestamp": "2025-07-14T12:34:35.803754"
  },
  {
    "epoch": 1732,
    "batch_idx": 128,
    "loss": 0.11558607220649719,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 571.25,
    "epsilon_norm": 7.657642841339111,
    "prediction_norm": 6.934174537658691,
    "timestamp": "2025-07-14T12:34:35.809094"
  },
  {
    "epoch": 1732,
    "batch_idx": 160,
    "loss": 0.14208726584911346,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 494.3125,
    "epsilon_norm": 7.862140655517578,
    "prediction_norm": 7.427834987640381,
    "timestamp": "2025-07-14T12:34:35.814177"
  },
  {
    "epoch": 1732,
    "batch_idx": 192,
    "loss": 0.062302038073539734,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 578.5,
    "epsilon_norm": 2.854846715927124,
    "prediction_norm": 2.640617847442627,
    "timestamp": "2025-07-14T12:34:35.819248"
  },
  {
    "epoch": 1733,
    "batch_idx": 0,
    "loss": 0.2618462145328522,
    "learning_rate": 4.3333817191403955e-05,
    "timestep_mean": 522.75,
    "epsilon_norm": 7.742112636566162,
    "prediction_norm": 7.561262607574463,
    "timestamp": "2025-07-14T12:34:35.824418"
  },
  {
    "epoch": 1733,
    "batch_idx": 32,
    "loss": 0.5586762428283691,
    "learning_rate": 4.3333817191403955e-05,
    "timestep_mean": 522.75,
    "epsilon_norm": 9.63101863861084,
    "prediction_norm": 7.635307788848877,
    "timestamp": "2025-07-14T12:34:35.829453"
  },
  {
    "epoch": 1733,
    "batch_idx": 64,
    "loss": 0.2518436312675476,
    "learning_rate": 4.3333817191403955e-05,
    "timestep_mean": 561.4375,
    "epsilon_norm": 7.111771583557129,
    "prediction_norm": 6.450737953186035,
    "timestamp": "2025-07-14T12:34:35.834619"
  },
  {
    "epoch": 1733,
    "batch_idx": 96,
    "loss": 0.2996288537979126,
    "learning_rate": 4.3333817191403955e-05,
    "timestep_mean": 511.0625,
    "epsilon_norm": 7.467658042907715,
    "prediction_norm": 6.19032621383667,
    "timestamp": "2025-07-14T12:34:35.839738"
  },
  {
    "epoch": 1733,
    "batch_idx": 128,
    "loss": 0.34260550141334534,
    "learning_rate": 4.3333817191403955e-05,
    "timestep_mean": 490.84375,
    "epsilon_norm": 8.13707160949707,
    "prediction_norm": 6.437326431274414,
    "timestamp": "2025-07-14T12:34:35.844731"
  },
  {
    "epoch": 1733,
    "batch_idx": 160,
    "loss": 0.14476758241653442,
    "learning_rate": 4.3333817191403955e-05,
    "timestep_mean": 513.75,
    "epsilon_norm": 7.642129421234131,
    "prediction_norm": 7.0525407791137695,
    "timestamp": "2025-07-14T12:34:35.849665"
  },
  {
    "epoch": 1733,
    "batch_idx": 192,
    "loss": 0.2531943917274475,
    "learning_rate": 4.3333817191403955e-05,
    "timestep_mean": 394.75,
    "epsilon_norm": 4.6647443771362305,
    "prediction_norm": 3.835047721862793,
    "timestamp": "2025-07-14T12:34:35.854929"
  },
  {
    "epoch": 1734,
    "batch_idx": 0,
    "loss": 0.17424404621124268,
    "learning_rate": 4.3014554862969466e-05,
    "timestep_mean": 500.28125,
    "epsilon_norm": 8.076348304748535,
    "prediction_norm": 6.792356491088867,
    "timestamp": "2025-07-14T12:34:35.860073"
  },
  {
    "epoch": 1734,
    "batch_idx": 32,
    "loss": 0.3694438934326172,
    "learning_rate": 4.3014554862969466e-05,
    "timestep_mean": 488.15625,
    "epsilon_norm": 8.123035430908203,
    "prediction_norm": 6.643538475036621,
    "timestamp": "2025-07-14T12:34:35.865201"
  },
  {
    "epoch": 1734,
    "batch_idx": 64,
    "loss": 0.30861514806747437,
    "learning_rate": 4.3014554862969466e-05,
    "timestep_mean": 570.8125,
    "epsilon_norm": 8.611542701721191,
    "prediction_norm": 6.492210865020752,
    "timestamp": "2025-07-14T12:34:35.870193"
  },
  {
    "epoch": 1734,
    "batch_idx": 96,
    "loss": 0.38590866327285767,
    "learning_rate": 4.3014554862969466e-05,
    "timestep_mean": 430.25,
    "epsilon_norm": 8.277168273925781,
    "prediction_norm": 6.454843521118164,
    "timestamp": "2025-07-14T12:34:35.875339"
  },
  {
    "epoch": 1734,
    "batch_idx": 128,
    "loss": 0.13541258871555328,
    "learning_rate": 4.3014554862969466e-05,
    "timestep_mean": 508.375,
    "epsilon_norm": 8.887762069702148,
    "prediction_norm": 8.433100700378418,
    "timestamp": "2025-07-14T12:34:35.880735"
  },
  {
    "epoch": 1734,
    "batch_idx": 160,
    "loss": 0.47104954719543457,
    "learning_rate": 4.3014554862969466e-05,
    "timestep_mean": 503.375,
    "epsilon_norm": 8.807586669921875,
    "prediction_norm": 7.304436206817627,
    "timestamp": "2025-07-14T12:34:35.886997"
  },
  {
    "epoch": 1734,
    "batch_idx": 192,
    "loss": 0.3621634244918823,
    "learning_rate": 4.3014554862969466e-05,
    "timestep_mean": 511.25,
    "epsilon_norm": 3.468377113342285,
    "prediction_norm": 2.8194189071655273,
    "timestamp": "2025-07-14T12:34:35.892890"
  },
  {
    "epoch": 1735,
    "batch_idx": 0,
    "loss": 0.1162150576710701,
    "learning_rate": 4.26964201006932e-05,
    "timestep_mean": 566.71875,
    "epsilon_norm": 6.817238807678223,
    "prediction_norm": 6.087253570556641,
    "timestamp": "2025-07-14T12:34:35.901255"
  },
  {
    "epoch": 1735,
    "batch_idx": 32,
    "loss": 0.329179048538208,
    "learning_rate": 4.26964201006932e-05,
    "timestep_mean": 476.59375,
    "epsilon_norm": 7.864506244659424,
    "prediction_norm": 6.103143692016602,
    "timestamp": "2025-07-14T12:34:35.908319"
  },
  {
    "epoch": 1735,
    "batch_idx": 64,
    "loss": 0.21300339698791504,
    "learning_rate": 4.26964201006932e-05,
    "timestep_mean": 503.40625,
    "epsilon_norm": 6.649864673614502,
    "prediction_norm": 5.7871809005737305,
    "timestamp": "2025-07-14T12:34:35.914910"
  },
  {
    "epoch": 1735,
    "batch_idx": 96,
    "loss": 0.20779499411582947,
    "learning_rate": 4.26964201006932e-05,
    "timestep_mean": 481.625,
    "epsilon_norm": 8.591280937194824,
    "prediction_norm": 7.679171562194824,
    "timestamp": "2025-07-14T12:34:35.921921"
  },
  {
    "epoch": 1735,
    "batch_idx": 128,
    "loss": 0.27698343992233276,
    "learning_rate": 4.26964201006932e-05,
    "timestep_mean": 568.84375,
    "epsilon_norm": 7.121518135070801,
    "prediction_norm": 5.659750461578369,
    "timestamp": "2025-07-14T12:34:35.929762"
  },
  {
    "epoch": 1735,
    "batch_idx": 160,
    "loss": 0.19544504582881927,
    "learning_rate": 4.26964201006932e-05,
    "timestep_mean": 446.65625,
    "epsilon_norm": 7.261921405792236,
    "prediction_norm": 6.494230270385742,
    "timestamp": "2025-07-14T12:34:35.936340"
  },
  {
    "epoch": 1735,
    "batch_idx": 192,
    "loss": 0.09267829358577728,
    "learning_rate": 4.26964201006932e-05,
    "timestep_mean": 617.375,
    "epsilon_norm": 4.806249141693115,
    "prediction_norm": 4.569312572479248,
    "timestamp": "2025-07-14T12:34:35.943141"
  },
  {
    "epoch": 1736,
    "batch_idx": 0,
    "loss": 0.09244579076766968,
    "learning_rate": 4.237941368954124e-05,
    "timestep_mean": 474.1875,
    "epsilon_norm": 7.602378845214844,
    "prediction_norm": 7.37260103225708,
    "timestamp": "2025-07-14T12:34:35.950612"
  },
  {
    "epoch": 1736,
    "batch_idx": 32,
    "loss": 0.23508217930793762,
    "learning_rate": 4.237941368954124e-05,
    "timestep_mean": 472.09375,
    "epsilon_norm": 8.000123977661133,
    "prediction_norm": 7.095512866973877,
    "timestamp": "2025-07-14T12:34:35.959013"
  },
  {
    "epoch": 1736,
    "batch_idx": 64,
    "loss": 0.35676681995391846,
    "learning_rate": 4.237941368954124e-05,
    "timestep_mean": 497.625,
    "epsilon_norm": 7.641360282897949,
    "prediction_norm": 6.398634910583496,
    "timestamp": "2025-07-14T12:34:35.967337"
  },
  {
    "epoch": 1736,
    "batch_idx": 96,
    "loss": 0.27511656284332275,
    "learning_rate": 4.237941368954124e-05,
    "timestep_mean": 486.40625,
    "epsilon_norm": 8.859304428100586,
    "prediction_norm": 7.661071300506592,
    "timestamp": "2025-07-14T12:34:35.975693"
  },
  {
    "epoch": 1736,
    "batch_idx": 128,
    "loss": 0.41997450590133667,
    "learning_rate": 4.237941368954124e-05,
    "timestep_mean": 421.34375,
    "epsilon_norm": 7.916631698608398,
    "prediction_norm": 5.843443393707275,
    "timestamp": "2025-07-14T12:34:35.983890"
  },
  {
    "epoch": 1736,
    "batch_idx": 160,
    "loss": 0.14874954521656036,
    "learning_rate": 4.237941368954124e-05,
    "timestep_mean": 580.53125,
    "epsilon_norm": 7.037569522857666,
    "prediction_norm": 6.294425010681152,
    "timestamp": "2025-07-14T12:34:35.991144"
  },
  {
    "epoch": 1736,
    "batch_idx": 192,
    "loss": 0.5509231090545654,
    "learning_rate": 4.237941368954124e-05,
    "timestep_mean": 344.0,
    "epsilon_norm": 3.471853494644165,
    "prediction_norm": 2.511990785598755,
    "timestamp": "2025-07-14T12:34:35.998476"
  },
  {
    "epoch": 1737,
    "batch_idx": 0,
    "loss": 0.43440407514572144,
    "learning_rate": 4.2063536411695255e-05,
    "timestep_mean": 442.21875,
    "epsilon_norm": 8.412138938903809,
    "prediction_norm": 6.456554889678955,
    "timestamp": "2025-07-14T12:34:36.005287"
  },
  {
    "epoch": 1737,
    "batch_idx": 32,
    "loss": 0.11981672048568726,
    "learning_rate": 4.2063536411695255e-05,
    "timestep_mean": 532.125,
    "epsilon_norm": 8.909358978271484,
    "prediction_norm": 8.1430025100708,
    "timestamp": "2025-07-14T12:34:36.010873"
  },
  {
    "epoch": 1737,
    "batch_idx": 64,
    "loss": 0.2033003270626068,
    "learning_rate": 4.2063536411695255e-05,
    "timestep_mean": 556.625,
    "epsilon_norm": 7.493655204772949,
    "prediction_norm": 6.602916717529297,
    "timestamp": "2025-07-14T12:34:36.015847"
  },
  {
    "epoch": 1737,
    "batch_idx": 96,
    "loss": 0.22054758667945862,
    "learning_rate": 4.2063536411695255e-05,
    "timestep_mean": 440.96875,
    "epsilon_norm": 7.729514122009277,
    "prediction_norm": 6.622692584991455,
    "timestamp": "2025-07-14T12:34:36.020666"
  },
  {
    "epoch": 1737,
    "batch_idx": 128,
    "loss": 0.2897473871707916,
    "learning_rate": 4.2063536411695255e-05,
    "timestep_mean": 518.75,
    "epsilon_norm": 9.712506294250488,
    "prediction_norm": 8.220014572143555,
    "timestamp": "2025-07-14T12:34:36.025555"
  },
  {
    "epoch": 1737,
    "batch_idx": 160,
    "loss": 0.320182204246521,
    "learning_rate": 4.2063536411695255e-05,
    "timestep_mean": 518.40625,
    "epsilon_norm": 8.407326698303223,
    "prediction_norm": 7.61093282699585,
    "timestamp": "2025-07-14T12:34:36.030498"
  },
  {
    "epoch": 1737,
    "batch_idx": 192,
    "loss": 0.8198460340499878,
    "learning_rate": 4.2063536411695255e-05,
    "timestep_mean": 376.125,
    "epsilon_norm": 5.130366325378418,
    "prediction_norm": 3.216917037963867,
    "timestamp": "2025-07-14T12:34:36.035493"
  },
  {
    "epoch": 1738,
    "batch_idx": 0,
    "loss": 0.24757829308509827,
    "learning_rate": 4.1748789046551005e-05,
    "timestep_mean": 467.75,
    "epsilon_norm": 7.842550277709961,
    "prediction_norm": 6.666985034942627,
    "timestamp": "2025-07-14T12:34:36.040359"
  },
  {
    "epoch": 1738,
    "batch_idx": 32,
    "loss": 0.4708254933357239,
    "learning_rate": 4.1748789046551005e-05,
    "timestep_mean": 482.34375,
    "epsilon_norm": 8.493673324584961,
    "prediction_norm": 6.271403789520264,
    "timestamp": "2025-07-14T12:34:36.045325"
  },
  {
    "epoch": 1738,
    "batch_idx": 64,
    "loss": 0.14171382784843445,
    "learning_rate": 4.1748789046551005e-05,
    "timestep_mean": 557.40625,
    "epsilon_norm": 7.5873613357543945,
    "prediction_norm": 7.426051139831543,
    "timestamp": "2025-07-14T12:34:36.050430"
  },
  {
    "epoch": 1738,
    "batch_idx": 96,
    "loss": 0.4013822674751282,
    "learning_rate": 4.1748789046551005e-05,
    "timestep_mean": 477.0,
    "epsilon_norm": 8.709169387817383,
    "prediction_norm": 7.147683620452881,
    "timestamp": "2025-07-14T12:34:36.055574"
  },
  {
    "epoch": 1738,
    "batch_idx": 128,
    "loss": 0.23862594366073608,
    "learning_rate": 4.1748789046551005e-05,
    "timestep_mean": 490.71875,
    "epsilon_norm": 6.997503280639648,
    "prediction_norm": 5.745115756988525,
    "timestamp": "2025-07-14T12:34:36.060459"
  },
  {
    "epoch": 1738,
    "batch_idx": 160,
    "loss": 0.11356622725725174,
    "learning_rate": 4.1748789046551005e-05,
    "timestep_mean": 504.6875,
    "epsilon_norm": 7.376588344573975,
    "prediction_norm": 6.73663330078125,
    "timestamp": "2025-07-14T12:34:36.065433"
  },
  {
    "epoch": 1738,
    "batch_idx": 192,
    "loss": 0.607624888420105,
    "learning_rate": 4.1748789046551005e-05,
    "timestep_mean": 425.625,
    "epsilon_norm": 4.7846150398254395,
    "prediction_norm": 3.825000762939453,
    "timestamp": "2025-07-14T12:34:36.070544"
  },
  {
    "epoch": 1739,
    "batch_idx": 0,
    "loss": 0.3159525394439697,
    "learning_rate": 4.143517237071649e-05,
    "timestep_mean": 469.96875,
    "epsilon_norm": 7.847082138061523,
    "prediction_norm": 6.844104766845703,
    "timestamp": "2025-07-14T12:34:36.075680"
  },
  {
    "epoch": 1739,
    "batch_idx": 32,
    "loss": 0.27557647228240967,
    "learning_rate": 4.143517237071649e-05,
    "timestep_mean": 499.21875,
    "epsilon_norm": 9.019991874694824,
    "prediction_norm": 7.794964790344238,
    "timestamp": "2025-07-14T12:34:36.080625"
  },
  {
    "epoch": 1739,
    "batch_idx": 64,
    "loss": 0.21727019548416138,
    "learning_rate": 4.143517237071649e-05,
    "timestep_mean": 509.90625,
    "epsilon_norm": 8.791467666625977,
    "prediction_norm": 7.501043319702148,
    "timestamp": "2025-07-14T12:34:36.085548"
  },
  {
    "epoch": 1739,
    "batch_idx": 96,
    "loss": 0.1750405728816986,
    "learning_rate": 4.143517237071649e-05,
    "timestep_mean": 492.59375,
    "epsilon_norm": 8.113622665405273,
    "prediction_norm": 7.012931823730469,
    "timestamp": "2025-07-14T12:34:36.090435"
  },
  {
    "epoch": 1739,
    "batch_idx": 128,
    "loss": 0.24343132972717285,
    "learning_rate": 4.143517237071649e-05,
    "timestep_mean": 480.53125,
    "epsilon_norm": 8.397018432617188,
    "prediction_norm": 7.19803524017334,
    "timestamp": "2025-07-14T12:34:36.095617"
  },
  {
    "epoch": 1739,
    "batch_idx": 160,
    "loss": 0.1823621690273285,
    "learning_rate": 4.143517237071649e-05,
    "timestep_mean": 563.71875,
    "epsilon_norm": 8.211335182189941,
    "prediction_norm": 7.545744895935059,
    "timestamp": "2025-07-14T12:34:36.100896"
  },
  {
    "epoch": 1739,
    "batch_idx": 192,
    "loss": 0.6808013916015625,
    "learning_rate": 4.143517237071649e-05,
    "timestep_mean": 370.875,
    "epsilon_norm": 4.236270427703857,
    "prediction_norm": 2.673661708831787,
    "timestamp": "2025-07-14T12:34:36.105923"
  },
  {
    "epoch": 1740,
    "batch_idx": 0,
    "loss": 0.3098989427089691,
    "learning_rate": 4.112268715800944e-05,
    "timestep_mean": 418.0625,
    "epsilon_norm": 7.953652858734131,
    "prediction_norm": 6.065967082977295,
    "timestamp": "2025-07-14T12:34:36.110997"
  },
  {
    "epoch": 1740,
    "batch_idx": 32,
    "loss": 0.19712218642234802,
    "learning_rate": 4.112268715800944e-05,
    "timestep_mean": 555.9375,
    "epsilon_norm": 8.651291847229004,
    "prediction_norm": 8.01871395111084,
    "timestamp": "2025-07-14T12:34:36.116120"
  },
  {
    "epoch": 1740,
    "batch_idx": 64,
    "loss": 0.15543313324451447,
    "learning_rate": 4.112268715800944e-05,
    "timestep_mean": 513.0625,
    "epsilon_norm": 7.798561096191406,
    "prediction_norm": 7.149786472320557,
    "timestamp": "2025-07-14T12:34:36.121226"
  },
  {
    "epoch": 1740,
    "batch_idx": 96,
    "loss": 0.23782551288604736,
    "learning_rate": 4.112268715800944e-05,
    "timestep_mean": 558.4375,
    "epsilon_norm": 9.020938873291016,
    "prediction_norm": 8.10628890991211,
    "timestamp": "2025-07-14T12:34:36.126302"
  },
  {
    "epoch": 1740,
    "batch_idx": 128,
    "loss": 0.17361092567443848,
    "learning_rate": 4.112268715800944e-05,
    "timestep_mean": 487.875,
    "epsilon_norm": 8.192466735839844,
    "prediction_norm": 7.59099006652832,
    "timestamp": "2025-07-14T12:34:36.131437"
  },
  {
    "epoch": 1740,
    "batch_idx": 160,
    "loss": 0.15886065363883972,
    "learning_rate": 4.112268715800944e-05,
    "timestep_mean": 511.21875,
    "epsilon_norm": 7.3068671226501465,
    "prediction_norm": 7.260122776031494,
    "timestamp": "2025-07-14T12:34:36.136547"
  },
  {
    "epoch": 1740,
    "batch_idx": 192,
    "loss": 0.2427995800971985,
    "learning_rate": 4.112268715800944e-05,
    "timestep_mean": 596.25,
    "epsilon_norm": 6.269313812255859,
    "prediction_norm": 5.436017990112305,
    "timestamp": "2025-07-14T12:34:36.141884"
  },
  {
    "epoch": 1741,
    "batch_idx": 0,
    "loss": 0.25923585891723633,
    "learning_rate": 4.0811334179456154e-05,
    "timestep_mean": 445.90625,
    "epsilon_norm": 7.888919830322266,
    "prediction_norm": 6.605472564697266,
    "timestamp": "2025-07-14T12:34:36.146978"
  },
  {
    "epoch": 1741,
    "batch_idx": 32,
    "loss": 0.40844792127609253,
    "learning_rate": 4.0811334179456154e-05,
    "timestep_mean": 453.5,
    "epsilon_norm": 8.531128883361816,
    "prediction_norm": 6.479736804962158,
    "timestamp": "2025-07-14T12:34:36.152040"
  },
  {
    "epoch": 1741,
    "batch_idx": 64,
    "loss": 0.2560824155807495,
    "learning_rate": 4.0811334179456154e-05,
    "timestep_mean": 490.21875,
    "epsilon_norm": 7.60538911819458,
    "prediction_norm": 6.313133716583252,
    "timestamp": "2025-07-14T12:34:36.157165"
  },
  {
    "epoch": 1741,
    "batch_idx": 96,
    "loss": 0.360637366771698,
    "learning_rate": 4.0811334179456154e-05,
    "timestep_mean": 514.09375,
    "epsilon_norm": 7.665158271789551,
    "prediction_norm": 6.5112223625183105,
    "timestamp": "2025-07-14T12:34:36.162288"
  },
  {
    "epoch": 1741,
    "batch_idx": 128,
    "loss": 0.20761986076831818,
    "learning_rate": 4.0811334179456154e-05,
    "timestep_mean": 421.0,
    "epsilon_norm": 7.404284954071045,
    "prediction_norm": 6.414688587188721,
    "timestamp": "2025-07-14T12:34:36.167478"
  },
  {
    "epoch": 1741,
    "batch_idx": 160,
    "loss": 0.35340723395347595,
    "learning_rate": 4.0811334179456154e-05,
    "timestep_mean": 475.75,
    "epsilon_norm": 8.796268463134766,
    "prediction_norm": 7.580039978027344,
    "timestamp": "2025-07-14T12:34:36.172587"
  },
  {
    "epoch": 1741,
    "batch_idx": 192,
    "loss": 0.4080711901187897,
    "learning_rate": 4.0811334179456154e-05,
    "timestep_mean": 612.375,
    "epsilon_norm": 3.4693427085876465,
    "prediction_norm": 2.66813588142395,
    "timestamp": "2025-07-14T12:34:36.177784"
  },
  {
    "epoch": 1742,
    "batch_idx": 0,
    "loss": 0.29087018966674805,
    "learning_rate": 4.0501114203289394e-05,
    "timestep_mean": 511.90625,
    "epsilon_norm": 8.144213676452637,
    "prediction_norm": 6.805539608001709,
    "timestamp": "2025-07-14T12:34:36.182923"
  },
  {
    "epoch": 1742,
    "batch_idx": 32,
    "loss": 0.12994170188903809,
    "learning_rate": 4.0501114203289394e-05,
    "timestep_mean": 608.65625,
    "epsilon_norm": 7.398673057556152,
    "prediction_norm": 6.906300067901611,
    "timestamp": "2025-07-14T12:34:36.188342"
  },
  {
    "epoch": 1742,
    "batch_idx": 64,
    "loss": 0.31536245346069336,
    "learning_rate": 4.0501114203289394e-05,
    "timestep_mean": 387.96875,
    "epsilon_norm": 7.214971542358398,
    "prediction_norm": 5.400646209716797,
    "timestamp": "2025-07-14T12:34:36.193500"
  },
  {
    "epoch": 1742,
    "batch_idx": 96,
    "loss": 0.2374582141637802,
    "learning_rate": 4.0501114203289394e-05,
    "timestep_mean": 428.3125,
    "epsilon_norm": 9.129858016967773,
    "prediction_norm": 7.549840450286865,
    "timestamp": "2025-07-14T12:34:36.198724"
  },
  {
    "epoch": 1742,
    "batch_idx": 128,
    "loss": 0.43120601773262024,
    "learning_rate": 4.0501114203289394e-05,
    "timestep_mean": 466.5625,
    "epsilon_norm": 7.899110794067383,
    "prediction_norm": 5.89054012298584,
    "timestamp": "2025-07-14T12:34:36.203963"
  },
  {
    "epoch": 1742,
    "batch_idx": 160,
    "loss": 0.15892446041107178,
    "learning_rate": 4.0501114203289394e-05,
    "timestep_mean": 571.96875,
    "epsilon_norm": 6.803386688232422,
    "prediction_norm": 6.208774566650391,
    "timestamp": "2025-07-14T12:34:36.209352"
  },
  {
    "epoch": 1742,
    "batch_idx": 192,
    "loss": 0.17799967527389526,
    "learning_rate": 4.0501114203289394e-05,
    "timestep_mean": 592.125,
    "epsilon_norm": 4.80810022354126,
    "prediction_norm": 4.07865571975708,
    "timestamp": "2025-07-14T12:34:36.214539"
  },
  {
    "epoch": 1743,
    "batch_idx": 0,
    "loss": 0.24603575468063354,
    "learning_rate": 4.019202799494565e-05,
    "timestep_mean": 564.40625,
    "epsilon_norm": 7.8786301612854,
    "prediction_norm": 6.754730224609375,
    "timestamp": "2025-07-14T12:34:36.219888"
  },
  {
    "epoch": 1743,
    "batch_idx": 32,
    "loss": 0.4377727806568146,
    "learning_rate": 4.019202799494565e-05,
    "timestep_mean": 458.375,
    "epsilon_norm": 8.878363609313965,
    "prediction_norm": 7.668748378753662,
    "timestamp": "2025-07-14T12:34:36.225176"
  },
  {
    "epoch": 1743,
    "batch_idx": 64,
    "loss": 0.140466570854187,
    "learning_rate": 4.019202799494565e-05,
    "timestep_mean": 569.28125,
    "epsilon_norm": 7.609360694885254,
    "prediction_norm": 7.096158027648926,
    "timestamp": "2025-07-14T12:34:36.230324"
  },
  {
    "epoch": 1743,
    "batch_idx": 96,
    "loss": 0.11595724523067474,
    "learning_rate": 4.019202799494565e-05,
    "timestep_mean": 619.21875,
    "epsilon_norm": 7.33414363861084,
    "prediction_norm": 6.937330722808838,
    "timestamp": "2025-07-14T12:34:36.235560"
  },
  {
    "epoch": 1743,
    "batch_idx": 128,
    "loss": 0.22044697403907776,
    "learning_rate": 4.019202799494565e-05,
    "timestep_mean": 508.0625,
    "epsilon_norm": 8.369799613952637,
    "prediction_norm": 7.211918354034424,
    "timestamp": "2025-07-14T12:34:36.240623"
  },
  {
    "epoch": 1743,
    "batch_idx": 160,
    "loss": 0.22777046263217926,
    "learning_rate": 4.019202799494565e-05,
    "timestep_mean": 587.0625,
    "epsilon_norm": 7.8316802978515625,
    "prediction_norm": 6.7272796630859375,
    "timestamp": "2025-07-14T12:34:36.247912"
  },
  {
    "epoch": 1743,
    "batch_idx": 192,
    "loss": 0.07086227089166641,
    "learning_rate": 4.019202799494565e-05,
    "timestep_mean": 589.875,
    "epsilon_norm": 4.178202152252197,
    "prediction_norm": 4.418039321899414,
    "timestamp": "2025-07-14T12:34:36.252984"
  },
  {
    "epoch": 1744,
    "batch_idx": 0,
    "loss": 0.2213279902935028,
    "learning_rate": 3.9884076317064924e-05,
    "timestep_mean": 508.84375,
    "epsilon_norm": 7.089231491088867,
    "prediction_norm": 5.885406017303467,
    "timestamp": "2025-07-14T12:34:36.261214"
  },
  {
    "epoch": 1744,
    "batch_idx": 32,
    "loss": 0.4976235330104828,
    "learning_rate": 3.9884076317064924e-05,
    "timestep_mean": 383.59375,
    "epsilon_norm": 7.57756233215332,
    "prediction_norm": 5.3850250244140625,
    "timestamp": "2025-07-14T12:34:36.266356"
  },
  {
    "epoch": 1744,
    "batch_idx": 64,
    "loss": 0.13458575308322906,
    "learning_rate": 3.9884076317064924e-05,
    "timestep_mean": 488.03125,
    "epsilon_norm": 8.31224250793457,
    "prediction_norm": 7.406922817230225,
    "timestamp": "2025-07-14T12:34:36.271334"
  },
  {
    "epoch": 1744,
    "batch_idx": 96,
    "loss": 0.363288551568985,
    "learning_rate": 3.9884076317064924e-05,
    "timestep_mean": 503.875,
    "epsilon_norm": 8.06213092803955,
    "prediction_norm": 6.877790451049805,
    "timestamp": "2025-07-14T12:34:36.276194"
  },
  {
    "epoch": 1744,
    "batch_idx": 128,
    "loss": 0.2725541889667511,
    "learning_rate": 3.9884076317064924e-05,
    "timestep_mean": 430.9375,
    "epsilon_norm": 8.135703086853027,
    "prediction_norm": 6.827019214630127,
    "timestamp": "2025-07-14T12:34:36.281050"
  },
  {
    "epoch": 1744,
    "batch_idx": 160,
    "loss": 0.19429035484790802,
    "learning_rate": 3.9884076317064924e-05,
    "timestep_mean": 505.375,
    "epsilon_norm": 7.54082727432251,
    "prediction_norm": 6.523605823516846,
    "timestamp": "2025-07-14T12:34:36.285965"
  },
  {
    "epoch": 1744,
    "batch_idx": 192,
    "loss": 0.4540364742279053,
    "learning_rate": 3.9884076317064924e-05,
    "timestep_mean": 538.75,
    "epsilon_norm": 4.160030841827393,
    "prediction_norm": 3.2945544719696045,
    "timestamp": "2025-07-14T12:34:36.290925"
  },
  {
    "epoch": 1745,
    "batch_idx": 0,
    "loss": 0.46668535470962524,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 432.125,
    "epsilon_norm": 7.5609540939331055,
    "prediction_norm": 5.941099166870117,
    "timestamp": "2025-07-14T12:34:36.295802"
  },
  {
    "epoch": 1745,
    "batch_idx": 32,
    "loss": 0.2247687131166458,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 547.625,
    "epsilon_norm": 8.492920875549316,
    "prediction_norm": 7.630351543426514,
    "timestamp": "2025-07-14T12:34:36.300654"
  },
  {
    "epoch": 1745,
    "batch_idx": 64,
    "loss": 0.22924219071865082,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 425.0625,
    "epsilon_norm": 7.984180927276611,
    "prediction_norm": 7.133887767791748,
    "timestamp": "2025-07-14T12:34:36.305394"
  },
  {
    "epoch": 1745,
    "batch_idx": 96,
    "loss": 0.3086867928504944,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 428.96875,
    "epsilon_norm": 8.561322212219238,
    "prediction_norm": 6.836387634277344,
    "timestamp": "2025-07-14T12:34:36.310159"
  },
  {
    "epoch": 1745,
    "batch_idx": 128,
    "loss": 0.205396369099617,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 509.6875,
    "epsilon_norm": 7.764385223388672,
    "prediction_norm": 6.47534704208374,
    "timestamp": "2025-07-14T12:34:36.314953"
  },
  {
    "epoch": 1745,
    "batch_idx": 160,
    "loss": 0.16670072078704834,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 562.9375,
    "epsilon_norm": 8.01542854309082,
    "prediction_norm": 7.406566143035889,
    "timestamp": "2025-07-14T12:34:36.319675"
  },
  {
    "epoch": 1745,
    "batch_idx": 192,
    "loss": 0.22683122754096985,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 373.25,
    "epsilon_norm": 3.3382065296173096,
    "prediction_norm": 2.9403233528137207,
    "timestamp": "2025-07-14T12:34:36.324594"
  },
  {
    "epoch": 1746,
    "batch_idx": 0,
    "loss": 0.2553534209728241,
    "learning_rate": 3.927157958925082e-05,
    "timestep_mean": 499.8125,
    "epsilon_norm": 7.585415840148926,
    "prediction_norm": 6.774186134338379,
    "timestamp": "2025-07-14T12:34:36.329632"
  },
  {
    "epoch": 1746,
    "batch_idx": 32,
    "loss": 0.14020174741744995,
    "learning_rate": 3.927157958925082e-05,
    "timestep_mean": 504.53125,
    "epsilon_norm": 8.381205558776855,
    "prediction_norm": 7.719594955444336,
    "timestamp": "2025-07-14T12:34:36.334501"
  },
  {
    "epoch": 1746,
    "batch_idx": 64,
    "loss": 0.17230017483234406,
    "learning_rate": 3.927157958925082e-05,
    "timestep_mean": 563.125,
    "epsilon_norm": 7.91767692565918,
    "prediction_norm": 7.612529754638672,
    "timestamp": "2025-07-14T12:34:36.339327"
  },
  {
    "epoch": 1746,
    "batch_idx": 96,
    "loss": 0.6787406802177429,
    "learning_rate": 3.927157958925082e-05,
    "timestep_mean": 412.1875,
    "epsilon_norm": 8.815613746643066,
    "prediction_norm": 5.418493747711182,
    "timestamp": "2025-07-14T12:34:36.344116"
  },
  {
    "epoch": 1746,
    "batch_idx": 128,
    "loss": 0.21149355173110962,
    "learning_rate": 3.927157958925082e-05,
    "timestep_mean": 430.84375,
    "epsilon_norm": 7.659791946411133,
    "prediction_norm": 6.691714286804199,
    "timestamp": "2025-07-14T12:34:36.348964"
  },
  {
    "epoch": 1746,
    "batch_idx": 160,
    "loss": 0.19877761602401733,
    "learning_rate": 3.927157958925082e-05,
    "timestep_mean": 534.375,
    "epsilon_norm": 7.282374382019043,
    "prediction_norm": 5.919534683227539,
    "timestamp": "2025-07-14T12:34:36.353817"
  },
  {
    "epoch": 1746,
    "batch_idx": 192,
    "loss": 0.2526325583457947,
    "learning_rate": 3.927157958925082e-05,
    "timestep_mean": 521.25,
    "epsilon_norm": 3.7150232791900635,
    "prediction_norm": 3.1401970386505127,
    "timestamp": "2025-07-14T12:34:36.358905"
  },
  {
    "epoch": 1747,
    "batch_idx": 0,
    "loss": 0.31438204646110535,
    "learning_rate": 3.896703605059243e-05,
    "timestep_mean": 552.59375,
    "epsilon_norm": 8.11612606048584,
    "prediction_norm": 6.705033302307129,
    "timestamp": "2025-07-14T12:34:36.363757"
  },
  {
    "epoch": 1747,
    "batch_idx": 32,
    "loss": 0.45510396361351013,
    "learning_rate": 3.896703605059243e-05,
    "timestep_mean": 428.875,
    "epsilon_norm": 7.0362162590026855,
    "prediction_norm": 4.976455211639404,
    "timestamp": "2025-07-14T12:34:36.368907"
  },
  {
    "epoch": 1747,
    "batch_idx": 64,
    "loss": 0.22619566321372986,
    "learning_rate": 3.896703605059243e-05,
    "timestep_mean": 492.375,
    "epsilon_norm": 7.741839408874512,
    "prediction_norm": 7.445036888122559,
    "timestamp": "2025-07-14T12:34:36.373853"
  },
  {
    "epoch": 1747,
    "batch_idx": 96,
    "loss": 0.17388635873794556,
    "learning_rate": 3.896703605059243e-05,
    "timestep_mean": 572.5625,
    "epsilon_norm": 8.539939880371094,
    "prediction_norm": 7.214362621307373,
    "timestamp": "2025-07-14T12:34:36.378787"
  },
  {
    "epoch": 1747,
    "batch_idx": 128,
    "loss": 0.3386545777320862,
    "learning_rate": 3.896703605059243e-05,
    "timestep_mean": 521.125,
    "epsilon_norm": 7.321957111358643,
    "prediction_norm": 6.002790927886963,
    "timestamp": "2025-07-14T12:34:36.383736"
  },
  {
    "epoch": 1747,
    "batch_idx": 160,
    "loss": 0.20545841753482819,
    "learning_rate": 3.896703605059243e-05,
    "timestep_mean": 549.25,
    "epsilon_norm": 7.17296028137207,
    "prediction_norm": 5.894704341888428,
    "timestamp": "2025-07-14T12:34:36.388815"
  },
  {
    "epoch": 1747,
    "batch_idx": 192,
    "loss": 0.5231768488883972,
    "learning_rate": 3.896703605059243e-05,
    "timestep_mean": 515.125,
    "epsilon_norm": 4.6348958015441895,
    "prediction_norm": 3.439816951751709,
    "timestamp": "2025-07-14T12:34:36.393918"
  },
  {
    "epoch": 1748,
    "batch_idx": 0,
    "loss": 0.37338340282440186,
    "learning_rate": 3.866363006494256e-05,
    "timestep_mean": 493.3125,
    "epsilon_norm": 9.540465354919434,
    "prediction_norm": 8.016310691833496,
    "timestamp": "2025-07-14T12:34:36.399058"
  },
  {
    "epoch": 1748,
    "batch_idx": 32,
    "loss": 0.23652958869934082,
    "learning_rate": 3.866363006494256e-05,
    "timestep_mean": 484.03125,
    "epsilon_norm": 7.580511093139648,
    "prediction_norm": 5.99355936050415,
    "timestamp": "2025-07-14T12:34:36.404035"
  },
  {
    "epoch": 1748,
    "batch_idx": 64,
    "loss": 0.31975528597831726,
    "learning_rate": 3.866363006494256e-05,
    "timestep_mean": 498.96875,
    "epsilon_norm": 8.219684600830078,
    "prediction_norm": 6.572962284088135,
    "timestamp": "2025-07-14T12:34:36.409038"
  },
  {
    "epoch": 1748,
    "batch_idx": 96,
    "loss": 0.2588086724281311,
    "learning_rate": 3.866363006494256e-05,
    "timestep_mean": 520.3125,
    "epsilon_norm": 8.62855339050293,
    "prediction_norm": 7.486323356628418,
    "timestamp": "2025-07-14T12:34:36.414213"
  },
  {
    "epoch": 1748,
    "batch_idx": 128,
    "loss": 0.28110942244529724,
    "learning_rate": 3.866363006494256e-05,
    "timestep_mean": 478.84375,
    "epsilon_norm": 7.843087673187256,
    "prediction_norm": 6.265727519989014,
    "timestamp": "2025-07-14T12:34:36.419346"
  },
  {
    "epoch": 1748,
    "batch_idx": 160,
    "loss": 0.2593795359134674,
    "learning_rate": 3.866363006494256e-05,
    "timestep_mean": 477.59375,
    "epsilon_norm": 7.681864261627197,
    "prediction_norm": 6.314335823059082,
    "timestamp": "2025-07-14T12:34:36.424438"
  },
  {
    "epoch": 1748,
    "batch_idx": 192,
    "loss": 0.13562244176864624,
    "learning_rate": 3.866363006494256e-05,
    "timestep_mean": 473.375,
    "epsilon_norm": 4.280487060546875,
    "prediction_norm": 3.9122581481933594,
    "timestamp": "2025-07-14T12:34:36.429667"
  },
  {
    "epoch": 1749,
    "batch_idx": 0,
    "loss": 0.31323087215423584,
    "learning_rate": 3.836136238092553e-05,
    "timestep_mean": 459.375,
    "epsilon_norm": 8.268709182739258,
    "prediction_norm": 6.793542861938477,
    "timestamp": "2025-07-14T12:34:36.434652"
  },
  {
    "epoch": 1749,
    "batch_idx": 32,
    "loss": 0.30576249957084656,
    "learning_rate": 3.836136238092553e-05,
    "timestep_mean": 510.625,
    "epsilon_norm": 8.756340980529785,
    "prediction_norm": 7.257730484008789,
    "timestamp": "2025-07-14T12:34:36.439854"
  },
  {
    "epoch": 1749,
    "batch_idx": 64,
    "loss": 0.34456372261047363,
    "learning_rate": 3.836136238092553e-05,
    "timestep_mean": 522.71875,
    "epsilon_norm": 8.607121467590332,
    "prediction_norm": 6.6698408126831055,
    "timestamp": "2025-07-14T12:34:36.444842"
  },
  {
    "epoch": 1749,
    "batch_idx": 96,
    "loss": 0.23150615394115448,
    "learning_rate": 3.836136238092553e-05,
    "timestep_mean": 468.53125,
    "epsilon_norm": 8.174878120422363,
    "prediction_norm": 7.282837867736816,
    "timestamp": "2025-07-14T12:34:36.449835"
  },
  {
    "epoch": 1749,
    "batch_idx": 128,
    "loss": 0.3051796853542328,
    "learning_rate": 3.836136238092553e-05,
    "timestep_mean": 557.65625,
    "epsilon_norm": 7.7584228515625,
    "prediction_norm": 6.665988922119141,
    "timestamp": "2025-07-14T12:34:36.455054"
  },
  {
    "epoch": 1749,
    "batch_idx": 160,
    "loss": 0.39364397525787354,
    "learning_rate": 3.836136238092553e-05,
    "timestep_mean": 506.84375,
    "epsilon_norm": 9.125136375427246,
    "prediction_norm": 7.2488508224487305,
    "timestamp": "2025-07-14T12:34:36.459985"
  },
  {
    "epoch": 1749,
    "batch_idx": 192,
    "loss": 1.0683313608169556,
    "learning_rate": 3.836136238092553e-05,
    "timestep_mean": 484.25,
    "epsilon_norm": 5.0398478507995605,
    "prediction_norm": 3.2462570667266846,
    "timestamp": "2025-07-14T12:34:36.465107"
  },
  {
    "epoch": 1750,
    "batch_idx": 0,
    "loss": 0.19394215941429138,
    "learning_rate": 3.806023374435663e-05,
    "timestep_mean": 483.96875,
    "epsilon_norm": 8.370598793029785,
    "prediction_norm": 7.743000507354736,
    "timestamp": "2025-07-14T12:34:36.470122"
  },
  {
    "epoch": 1750,
    "batch_idx": 32,
    "loss": 0.28704899549484253,
    "learning_rate": 3.806023374435663e-05,
    "timestep_mean": 516.5625,
    "epsilon_norm": 8.674125671386719,
    "prediction_norm": 7.331549167633057,
    "timestamp": "2025-07-14T12:34:36.475168"
  },
  {
    "epoch": 1750,
    "batch_idx": 64,
    "loss": 0.2439044564962387,
    "learning_rate": 3.806023374435663e-05,
    "timestep_mean": 499.09375,
    "epsilon_norm": 7.032036781311035,
    "prediction_norm": 5.803523063659668,
    "timestamp": "2025-07-14T12:34:36.480162"
  },
  {
    "epoch": 1750,
    "batch_idx": 96,
    "loss": 0.27542757987976074,
    "learning_rate": 3.806023374435663e-05,
    "timestep_mean": 455.6875,
    "epsilon_norm": 8.531797409057617,
    "prediction_norm": 7.37990140914917,
    "timestamp": "2025-07-14T12:34:36.485232"
  },
  {
    "epoch": 1750,
    "batch_idx": 128,
    "loss": 0.13466781377792358,
    "learning_rate": 3.806023374435663e-05,
    "timestep_mean": 549.53125,
    "epsilon_norm": 8.165301322937012,
    "prediction_norm": 7.793449878692627,
    "timestamp": "2025-07-14T12:34:36.490270"
  },
  {
    "epoch": 1750,
    "batch_idx": 160,
    "loss": 0.3585491478443146,
    "learning_rate": 3.806023374435663e-05,
    "timestep_mean": 432.71875,
    "epsilon_norm": 9.2013521194458,
    "prediction_norm": 7.242832183837891,
    "timestamp": "2025-07-14T12:34:36.495293"
  },
  {
    "epoch": 1750,
    "batch_idx": 192,
    "loss": 0.24754536151885986,
    "learning_rate": 3.806023374435663e-05,
    "timestep_mean": 549.625,
    "epsilon_norm": 4.542046070098877,
    "prediction_norm": 4.104033470153809,
    "timestamp": "2025-07-14T12:34:36.500359"
  },
  {
    "epoch": 1751,
    "batch_idx": 0,
    "loss": 0.2910357117652893,
    "learning_rate": 3.7760244898240926e-05,
    "timestep_mean": 439.46875,
    "epsilon_norm": 7.468164443969727,
    "prediction_norm": 5.9599289894104,
    "timestamp": "2025-07-14T12:34:36.505332"
  },
  {
    "epoch": 1751,
    "batch_idx": 32,
    "loss": 0.278423547744751,
    "learning_rate": 3.7760244898240926e-05,
    "timestep_mean": 490.4375,
    "epsilon_norm": 8.541092872619629,
    "prediction_norm": 7.711104393005371,
    "timestamp": "2025-07-14T12:34:36.510236"
  },
  {
    "epoch": 1751,
    "batch_idx": 64,
    "loss": 0.32994943857192993,
    "learning_rate": 3.7760244898240926e-05,
    "timestep_mean": 464.3125,
    "epsilon_norm": 8.589223861694336,
    "prediction_norm": 6.682584762573242,
    "timestamp": "2025-07-14T12:34:36.515176"
  },
  {
    "epoch": 1751,
    "batch_idx": 96,
    "loss": 0.1784004420042038,
    "learning_rate": 3.7760244898240926e-05,
    "timestep_mean": 502.4375,
    "epsilon_norm": 7.116170883178711,
    "prediction_norm": 6.560335159301758,
    "timestamp": "2025-07-14T12:34:36.520183"
  },
  {
    "epoch": 1751,
    "batch_idx": 128,
    "loss": 0.31411653757095337,
    "learning_rate": 3.7760244898240926e-05,
    "timestep_mean": 474.78125,
    "epsilon_norm": 8.131047248840332,
    "prediction_norm": 6.156680583953857,
    "timestamp": "2025-07-14T12:34:36.525293"
  },
  {
    "epoch": 1751,
    "batch_idx": 160,
    "loss": 0.4177554249763489,
    "learning_rate": 3.7760244898240926e-05,
    "timestep_mean": 470.28125,
    "epsilon_norm": 8.157119750976562,
    "prediction_norm": 6.118200302124023,
    "timestamp": "2025-07-14T12:34:36.530637"
  },
  {
    "epoch": 1751,
    "batch_idx": 192,
    "loss": 0.1770850419998169,
    "learning_rate": 3.7760244898240926e-05,
    "timestep_mean": 470.625,
    "epsilon_norm": 2.8848795890808105,
    "prediction_norm": 2.322078227996826,
    "timestamp": "2025-07-14T12:34:36.537687"
  },
  {
    "epoch": 1752,
    "batch_idx": 0,
    "loss": 0.2547376751899719,
    "learning_rate": 3.746139658277092e-05,
    "timestep_mean": 491.875,
    "epsilon_norm": 6.996767520904541,
    "prediction_norm": 6.174887180328369,
    "timestamp": "2025-07-14T12:34:36.553812"
  },
  {
    "epoch": 1752,
    "batch_idx": 32,
    "loss": 0.33488526940345764,
    "learning_rate": 3.746139658277092e-05,
    "timestep_mean": 513.375,
    "epsilon_norm": 7.675729751586914,
    "prediction_norm": 6.4539713859558105,
    "timestamp": "2025-07-14T12:34:36.563666"
  },
  {
    "epoch": 1752,
    "batch_idx": 64,
    "loss": 0.37379133701324463,
    "learning_rate": 3.746139658277092e-05,
    "timestep_mean": 476.21875,
    "epsilon_norm": 8.258955001831055,
    "prediction_norm": 7.468109130859375,
    "timestamp": "2025-07-14T12:34:36.571204"
  },
  {
    "epoch": 1752,
    "batch_idx": 96,
    "loss": 0.33588069677352905,
    "learning_rate": 3.746139658277092e-05,
    "timestep_mean": 448.09375,
    "epsilon_norm": 7.8108601570129395,
    "prediction_norm": 6.052634239196777,
    "timestamp": "2025-07-14T12:34:36.577929"
  },
  {
    "epoch": 1752,
    "batch_idx": 128,
    "loss": 0.17256388068199158,
    "learning_rate": 3.746139658277092e-05,
    "timestep_mean": 596.34375,
    "epsilon_norm": 8.371485710144043,
    "prediction_norm": 7.46902322769165,
    "timestamp": "2025-07-14T12:34:36.584801"
  },
  {
    "epoch": 1752,
    "batch_idx": 160,
    "loss": 0.3570650815963745,
    "learning_rate": 3.746139658277092e-05,
    "timestep_mean": 535.75,
    "epsilon_norm": 8.413481712341309,
    "prediction_norm": 6.935297012329102,
    "timestamp": "2025-07-14T12:34:36.591031"
  },
  {
    "epoch": 1752,
    "batch_idx": 192,
    "loss": 0.34692198038101196,
    "learning_rate": 3.746139658277092e-05,
    "timestep_mean": 349.375,
    "epsilon_norm": 3.3578546047210693,
    "prediction_norm": 2.6913037300109863,
    "timestamp": "2025-07-14T12:34:36.597276"
  },
  {
    "epoch": 1753,
    "batch_idx": 0,
    "loss": 0.25889861583709717,
    "learning_rate": 3.716368953532545e-05,
    "timestep_mean": 492.125,
    "epsilon_norm": 8.321725845336914,
    "prediction_norm": 6.893013954162598,
    "timestamp": "2025-07-14T12:34:36.605567"
  },
  {
    "epoch": 1753,
    "batch_idx": 32,
    "loss": 0.5108560919761658,
    "learning_rate": 3.716368953532545e-05,
    "timestep_mean": 412.34375,
    "epsilon_norm": 8.239251136779785,
    "prediction_norm": 6.5660624504089355,
    "timestamp": "2025-07-14T12:34:36.613732"
  },
  {
    "epoch": 1753,
    "batch_idx": 64,
    "loss": 0.225579172372818,
    "learning_rate": 3.716368953532545e-05,
    "timestep_mean": 520.4375,
    "epsilon_norm": 7.181107521057129,
    "prediction_norm": 6.407868385314941,
    "timestamp": "2025-07-14T12:34:36.622576"
  },
  {
    "epoch": 1753,
    "batch_idx": 96,
    "loss": 0.3814300298690796,
    "learning_rate": 3.716368953532545e-05,
    "timestep_mean": 526.1875,
    "epsilon_norm": 8.602486610412598,
    "prediction_norm": 6.574130535125732,
    "timestamp": "2025-07-14T12:34:36.631873"
  },
  {
    "epoch": 1753,
    "batch_idx": 128,
    "loss": 0.3208983242511749,
    "learning_rate": 3.716368953532545e-05,
    "timestep_mean": 541.8125,
    "epsilon_norm": 8.585550308227539,
    "prediction_norm": 7.874267101287842,
    "timestamp": "2025-07-14T12:34:36.640503"
  },
  {
    "epoch": 1753,
    "batch_idx": 160,
    "loss": 0.5051826238632202,
    "learning_rate": 3.716368953532545e-05,
    "timestep_mean": 509.46875,
    "epsilon_norm": 8.563672065734863,
    "prediction_norm": 6.46130895614624,
    "timestamp": "2025-07-14T12:34:36.648342"
  },
  {
    "epoch": 1753,
    "batch_idx": 192,
    "loss": 0.2651295065879822,
    "learning_rate": 3.716368953532545e-05,
    "timestep_mean": 460.5,
    "epsilon_norm": 4.099698066711426,
    "prediction_norm": 3.218627452850342,
    "timestamp": "2025-07-14T12:34:36.655128"
  },
  {
    "epoch": 1754,
    "batch_idx": 0,
    "loss": 0.420227587223053,
    "learning_rate": 3.6867124490466693e-05,
    "timestep_mean": 532.25,
    "epsilon_norm": 8.659931182861328,
    "prediction_norm": 6.136250019073486,
    "timestamp": "2025-07-14T12:34:36.661004"
  },
  {
    "epoch": 1754,
    "batch_idx": 32,
    "loss": 0.30805861949920654,
    "learning_rate": 3.6867124490466693e-05,
    "timestep_mean": 482.4375,
    "epsilon_norm": 7.3231706619262695,
    "prediction_norm": 6.115573406219482,
    "timestamp": "2025-07-14T12:34:36.666347"
  },
  {
    "epoch": 1754,
    "batch_idx": 64,
    "loss": 0.5305909514427185,
    "learning_rate": 3.6867124490466693e-05,
    "timestep_mean": 470.6875,
    "epsilon_norm": 8.552870750427246,
    "prediction_norm": 6.326535701751709,
    "timestamp": "2025-07-14T12:34:36.671440"
  },
  {
    "epoch": 1754,
    "batch_idx": 96,
    "loss": 0.2370031625032425,
    "learning_rate": 3.6867124490466693e-05,
    "timestep_mean": 442.5625,
    "epsilon_norm": 8.018872261047363,
    "prediction_norm": 6.7748918533325195,
    "timestamp": "2025-07-14T12:34:36.676350"
  },
  {
    "epoch": 1754,
    "batch_idx": 128,
    "loss": 0.2029963731765747,
    "learning_rate": 3.6867124490466693e-05,
    "timestep_mean": 436.625,
    "epsilon_norm": 8.719449043273926,
    "prediction_norm": 8.186479568481445,
    "timestamp": "2025-07-14T12:34:36.681279"
  },
  {
    "epoch": 1754,
    "batch_idx": 160,
    "loss": 0.20523923635482788,
    "learning_rate": 3.6867124490466693e-05,
    "timestep_mean": 450.625,
    "epsilon_norm": 7.568885803222656,
    "prediction_norm": 6.985219478607178,
    "timestamp": "2025-07-14T12:34:36.686171"
  },
  {
    "epoch": 1754,
    "batch_idx": 192,
    "loss": 0.12633979320526123,
    "learning_rate": 3.6867124490466693e-05,
    "timestep_mean": 400.5,
    "epsilon_norm": 3.9023940563201904,
    "prediction_norm": 3.4709386825561523,
    "timestamp": "2025-07-14T12:34:36.691157"
  },
  {
    "epoch": 1755,
    "batch_idx": 0,
    "loss": 0.5043731331825256,
    "learning_rate": 3.65717021799396e-05,
    "timestep_mean": 438.9375,
    "epsilon_norm": 9.155130386352539,
    "prediction_norm": 7.007297515869141,
    "timestamp": "2025-07-14T12:34:36.696187"
  },
  {
    "epoch": 1755,
    "batch_idx": 32,
    "loss": 0.25137490034103394,
    "learning_rate": 3.65717021799396e-05,
    "timestep_mean": 536.15625,
    "epsilon_norm": 8.480462074279785,
    "prediction_norm": 7.168551445007324,
    "timestamp": "2025-07-14T12:34:36.701001"
  },
  {
    "epoch": 1755,
    "batch_idx": 64,
    "loss": 0.41678497195243835,
    "learning_rate": 3.65717021799396e-05,
    "timestep_mean": 551.0625,
    "epsilon_norm": 9.379801750183105,
    "prediction_norm": 7.8080244064331055,
    "timestamp": "2025-07-14T12:34:36.705865"
  },
  {
    "epoch": 1755,
    "batch_idx": 96,
    "loss": 0.5020660161972046,
    "learning_rate": 3.65717021799396e-05,
    "timestep_mean": 417.40625,
    "epsilon_norm": 9.077640533447266,
    "prediction_norm": 6.439040660858154,
    "timestamp": "2025-07-14T12:34:36.710726"
  },
  {
    "epoch": 1755,
    "batch_idx": 128,
    "loss": 0.4419962465763092,
    "learning_rate": 3.65717021799396e-05,
    "timestep_mean": 475.0625,
    "epsilon_norm": 7.812449932098389,
    "prediction_norm": 5.9983367919921875,
    "timestamp": "2025-07-14T12:34:36.715610"
  },
  {
    "epoch": 1755,
    "batch_idx": 160,
    "loss": 0.2950383126735687,
    "learning_rate": 3.65717021799396e-05,
    "timestep_mean": 495.53125,
    "epsilon_norm": 7.785198211669922,
    "prediction_norm": 6.205532073974609,
    "timestamp": "2025-07-14T12:34:36.720421"
  },
  {
    "epoch": 1755,
    "batch_idx": 192,
    "loss": 0.2617875039577484,
    "learning_rate": 3.65717021799396e-05,
    "timestep_mean": 563.75,
    "epsilon_norm": 3.256197929382324,
    "prediction_norm": 2.726428270339966,
    "timestamp": "2025-07-14T12:34:36.725485"
  },
  {
    "epoch": 1756,
    "batch_idx": 0,
    "loss": 0.26483818888664246,
    "learning_rate": 3.627742333266936e-05,
    "timestep_mean": 537.875,
    "epsilon_norm": 7.869286060333252,
    "prediction_norm": 6.815886497497559,
    "timestamp": "2025-07-14T12:34:36.730343"
  },
  {
    "epoch": 1756,
    "batch_idx": 32,
    "loss": 0.5911838412284851,
    "learning_rate": 3.627742333266936e-05,
    "timestep_mean": 396.1875,
    "epsilon_norm": 8.645001411437988,
    "prediction_norm": 6.90907096862793,
    "timestamp": "2025-07-14T12:34:36.735254"
  },
  {
    "epoch": 1756,
    "batch_idx": 64,
    "loss": 0.3970925211906433,
    "learning_rate": 3.627742333266936e-05,
    "timestep_mean": 425.71875,
    "epsilon_norm": 8.173296928405762,
    "prediction_norm": 7.0167951583862305,
    "timestamp": "2025-07-14T12:34:36.740375"
  },
  {
    "epoch": 1756,
    "batch_idx": 96,
    "loss": 0.2887113690376282,
    "learning_rate": 3.627742333266936e-05,
    "timestep_mean": 535.625,
    "epsilon_norm": 9.138023376464844,
    "prediction_norm": 7.270429611206055,
    "timestamp": "2025-07-14T12:34:36.745305"
  },
  {
    "epoch": 1756,
    "batch_idx": 128,
    "loss": 0.2642424404621124,
    "learning_rate": 3.627742333266936e-05,
    "timestep_mean": 477.8125,
    "epsilon_norm": 7.893918991088867,
    "prediction_norm": 6.310330390930176,
    "timestamp": "2025-07-14T12:34:36.750377"
  },
  {
    "epoch": 1756,
    "batch_idx": 160,
    "loss": 0.24711012840270996,
    "learning_rate": 3.627742333266936e-05,
    "timestep_mean": 524.84375,
    "epsilon_norm": 8.404765129089355,
    "prediction_norm": 7.386873722076416,
    "timestamp": "2025-07-14T12:34:36.755143"
  },
  {
    "epoch": 1756,
    "batch_idx": 192,
    "loss": 0.11543167382478714,
    "learning_rate": 3.627742333266936e-05,
    "timestep_mean": 482.875,
    "epsilon_norm": 3.735269784927368,
    "prediction_norm": 3.2399849891662598,
    "timestamp": "2025-07-14T12:34:36.760384"
  },
  {
    "epoch": 1757,
    "batch_idx": 0,
    "loss": 0.2729620337486267,
    "learning_rate": 3.598428867475966e-05,
    "timestep_mean": 498.3125,
    "epsilon_norm": 7.709567546844482,
    "prediction_norm": 6.171436786651611,
    "timestamp": "2025-07-14T12:34:36.765272"
  },
  {
    "epoch": 1757,
    "batch_idx": 32,
    "loss": 0.4001728296279907,
    "learning_rate": 3.598428867475966e-05,
    "timestep_mean": 487.15625,
    "epsilon_norm": 8.255407333374023,
    "prediction_norm": 6.483757019042969,
    "timestamp": "2025-07-14T12:34:36.770195"
  },
  {
    "epoch": 1757,
    "batch_idx": 64,
    "loss": 0.2197059541940689,
    "learning_rate": 3.598428867475966e-05,
    "timestep_mean": 502.3125,
    "epsilon_norm": 7.338207244873047,
    "prediction_norm": 6.1633076667785645,
    "timestamp": "2025-07-14T12:34:36.775082"
  },
  {
    "epoch": 1757,
    "batch_idx": 96,
    "loss": 0.2732512056827545,
    "learning_rate": 3.598428867475966e-05,
    "timestep_mean": 525.6875,
    "epsilon_norm": 7.745190620422363,
    "prediction_norm": 6.746213436126709,
    "timestamp": "2025-07-14T12:34:36.780064"
  },
  {
    "epoch": 1757,
    "batch_idx": 128,
    "loss": 0.2854238450527191,
    "learning_rate": 3.598428867475966e-05,
    "timestep_mean": 486.0,
    "epsilon_norm": 8.020365715026855,
    "prediction_norm": 6.431707859039307,
    "timestamp": "2025-07-14T12:34:36.784949"
  },
  {
    "epoch": 1757,
    "batch_idx": 160,
    "loss": 0.162051260471344,
    "learning_rate": 3.598428867475966e-05,
    "timestep_mean": 559.03125,
    "epsilon_norm": 7.378610134124756,
    "prediction_norm": 6.626049995422363,
    "timestamp": "2025-07-14T12:34:36.789908"
  },
  {
    "epoch": 1757,
    "batch_idx": 192,
    "loss": 0.5169018507003784,
    "learning_rate": 3.598428867475966e-05,
    "timestep_mean": 547.625,
    "epsilon_norm": 4.143904209136963,
    "prediction_norm": 3.839268922805786,
    "timestamp": "2025-07-14T12:34:36.794914"
  },
  {
    "epoch": 1758,
    "batch_idx": 0,
    "loss": 0.22365736961364746,
    "learning_rate": 3.5692298929491434e-05,
    "timestep_mean": 516.40625,
    "epsilon_norm": 8.302231788635254,
    "prediction_norm": 7.374373912811279,
    "timestamp": "2025-07-14T12:34:36.799863"
  },
  {
    "epoch": 1758,
    "batch_idx": 32,
    "loss": 0.31942760944366455,
    "learning_rate": 3.5692298929491434e-05,
    "timestep_mean": 500.53125,
    "epsilon_norm": 7.353109836578369,
    "prediction_norm": 5.897228717803955,
    "timestamp": "2025-07-14T12:34:36.804951"
  },
  {
    "epoch": 1758,
    "batch_idx": 64,
    "loss": 0.21455520391464233,
    "learning_rate": 3.5692298929491434e-05,
    "timestep_mean": 590.8125,
    "epsilon_norm": 8.60565185546875,
    "prediction_norm": 7.751835823059082,
    "timestamp": "2025-07-14T12:34:36.809831"
  },
  {
    "epoch": 1758,
    "batch_idx": 96,
    "loss": 0.3288726806640625,
    "learning_rate": 3.5692298929491434e-05,
    "timestep_mean": 458.59375,
    "epsilon_norm": 7.524993896484375,
    "prediction_norm": 6.119228363037109,
    "timestamp": "2025-07-14T12:34:36.814814"
  },
  {
    "epoch": 1758,
    "batch_idx": 128,
    "loss": 0.25577765703201294,
    "learning_rate": 3.5692298929491434e-05,
    "timestep_mean": 497.125,
    "epsilon_norm": 9.214479446411133,
    "prediction_norm": 7.970667839050293,
    "timestamp": "2025-07-14T12:34:36.819694"
  },
  {
    "epoch": 1758,
    "batch_idx": 160,
    "loss": 0.17011788487434387,
    "learning_rate": 3.5692298929491434e-05,
    "timestep_mean": 579.40625,
    "epsilon_norm": 8.381614685058594,
    "prediction_norm": 7.993354797363281,
    "timestamp": "2025-07-14T12:34:36.824845"
  },
  {
    "epoch": 1758,
    "batch_idx": 192,
    "loss": 0.18776656687259674,
    "learning_rate": 3.5692298929491434e-05,
    "timestep_mean": 406.625,
    "epsilon_norm": 3.2053303718566895,
    "prediction_norm": 2.5219264030456543,
    "timestamp": "2025-07-14T12:34:36.830033"
  },
  {
    "epoch": 1759,
    "batch_idx": 0,
    "loss": 0.24931183457374573,
    "learning_rate": 3.540145481731987e-05,
    "timestep_mean": 494.4375,
    "epsilon_norm": 8.501450538635254,
    "prediction_norm": 7.145049571990967,
    "timestamp": "2025-07-14T12:34:36.835185"
  },
  {
    "epoch": 1759,
    "batch_idx": 32,
    "loss": 0.17099952697753906,
    "learning_rate": 3.540145481731987e-05,
    "timestep_mean": 580.9375,
    "epsilon_norm": 8.600062370300293,
    "prediction_norm": 7.530751705169678,
    "timestamp": "2025-07-14T12:34:36.840416"
  },
  {
    "epoch": 1759,
    "batch_idx": 64,
    "loss": 0.334482878446579,
    "learning_rate": 3.540145481731987e-05,
    "timestep_mean": 439.875,
    "epsilon_norm": 8.251986503601074,
    "prediction_norm": 6.845098972320557,
    "timestamp": "2025-07-14T12:34:36.845680"
  },
  {
    "epoch": 1759,
    "batch_idx": 96,
    "loss": 0.21257424354553223,
    "learning_rate": 3.540145481731987e-05,
    "timestep_mean": 438.125,
    "epsilon_norm": 7.591641902923584,
    "prediction_norm": 6.965728282928467,
    "timestamp": "2025-07-14T12:34:36.850961"
  },
  {
    "epoch": 1759,
    "batch_idx": 128,
    "loss": 0.4098302125930786,
    "learning_rate": 3.540145481731987e-05,
    "timestep_mean": 533.03125,
    "epsilon_norm": 8.550164222717285,
    "prediction_norm": 7.390316963195801,
    "timestamp": "2025-07-14T12:34:36.856102"
  },
  {
    "epoch": 1759,
    "batch_idx": 160,
    "loss": 0.1007121354341507,
    "learning_rate": 3.540145481731987e-05,
    "timestep_mean": 561.6875,
    "epsilon_norm": 8.449094772338867,
    "prediction_norm": 8.100987434387207,
    "timestamp": "2025-07-14T12:34:36.861255"
  },
  {
    "epoch": 1759,
    "batch_idx": 192,
    "loss": 0.16948197782039642,
    "learning_rate": 3.540145481731987e-05,
    "timestep_mean": 527.75,
    "epsilon_norm": 6.222019672393799,
    "prediction_norm": 6.181405067443848,
    "timestamp": "2025-07-14T12:34:36.866469"
  },
  {
    "epoch": 1760,
    "batch_idx": 0,
    "loss": 0.32524174451828003,
    "learning_rate": 3.5111757055874265e-05,
    "timestep_mean": 441.46875,
    "epsilon_norm": 7.440000534057617,
    "prediction_norm": 5.463015079498291,
    "timestamp": "2025-07-14T12:34:36.871662"
  },
  {
    "epoch": 1760,
    "batch_idx": 32,
    "loss": 0.17131341993808746,
    "learning_rate": 3.5111757055874265e-05,
    "timestep_mean": 525.71875,
    "epsilon_norm": 7.737762928009033,
    "prediction_norm": 6.901607036590576,
    "timestamp": "2025-07-14T12:34:36.876725"
  },
  {
    "epoch": 1760,
    "batch_idx": 64,
    "loss": 0.18756958842277527,
    "learning_rate": 3.5111757055874265e-05,
    "timestep_mean": 475.5625,
    "epsilon_norm": 6.621193885803223,
    "prediction_norm": 5.776263236999512,
    "timestamp": "2025-07-14T12:34:36.881942"
  },
  {
    "epoch": 1760,
    "batch_idx": 96,
    "loss": 0.2914935052394867,
    "learning_rate": 3.5111757055874265e-05,
    "timestep_mean": 418.59375,
    "epsilon_norm": 8.380755424499512,
    "prediction_norm": 7.588067054748535,
    "timestamp": "2025-07-14T12:34:36.887231"
  },
  {
    "epoch": 1760,
    "batch_idx": 128,
    "loss": 0.28178006410598755,
    "learning_rate": 3.5111757055874265e-05,
    "timestep_mean": 496.28125,
    "epsilon_norm": 8.860573768615723,
    "prediction_norm": 7.53676176071167,
    "timestamp": "2025-07-14T12:34:36.892457"
  },
  {
    "epoch": 1760,
    "batch_idx": 160,
    "loss": 0.2225390523672104,
    "learning_rate": 3.5111757055874265e-05,
    "timestep_mean": 464.4375,
    "epsilon_norm": 7.918849945068359,
    "prediction_norm": 6.950882911682129,
    "timestamp": "2025-07-14T12:34:36.897710"
  },
  {
    "epoch": 1760,
    "batch_idx": 192,
    "loss": 0.17805393040180206,
    "learning_rate": 3.5111757055874265e-05,
    "timestep_mean": 551.375,
    "epsilon_norm": 4.134041786193848,
    "prediction_norm": 3.724169969558716,
    "timestamp": "2025-07-14T12:34:36.902964"
  },
  {
    "epoch": 1761,
    "batch_idx": 0,
    "loss": 0.4494851231575012,
    "learning_rate": 3.482320635995505e-05,
    "timestep_mean": 503.1875,
    "epsilon_norm": 8.438898086547852,
    "prediction_norm": 6.557950973510742,
    "timestamp": "2025-07-14T12:34:36.908214"
  },
  {
    "epoch": 1761,
    "batch_idx": 32,
    "loss": 0.3585512340068817,
    "learning_rate": 3.482320635995505e-05,
    "timestep_mean": 468.96875,
    "epsilon_norm": 8.590524673461914,
    "prediction_norm": 6.827851295471191,
    "timestamp": "2025-07-14T12:34:36.913492"
  },
  {
    "epoch": 1761,
    "batch_idx": 64,
    "loss": 0.2421494722366333,
    "learning_rate": 3.482320635995505e-05,
    "timestep_mean": 448.40625,
    "epsilon_norm": 7.487706661224365,
    "prediction_norm": 6.285052299499512,
    "timestamp": "2025-07-14T12:34:36.918793"
  },
  {
    "epoch": 1761,
    "batch_idx": 96,
    "loss": 0.18316581845283508,
    "learning_rate": 3.482320635995505e-05,
    "timestep_mean": 500.34375,
    "epsilon_norm": 7.096976280212402,
    "prediction_norm": 5.984915256500244,
    "timestamp": "2025-07-14T12:34:36.924030"
  },
  {
    "epoch": 1761,
    "batch_idx": 128,
    "loss": 0.35326993465423584,
    "learning_rate": 3.482320635995505e-05,
    "timestep_mean": 527.96875,
    "epsilon_norm": 9.004244804382324,
    "prediction_norm": 7.8199462890625,
    "timestamp": "2025-07-14T12:34:36.929190"
  },
  {
    "epoch": 1761,
    "batch_idx": 160,
    "loss": 0.20109891891479492,
    "learning_rate": 3.482320635995505e-05,
    "timestep_mean": 420.5,
    "epsilon_norm": 7.177905082702637,
    "prediction_norm": 6.395864486694336,
    "timestamp": "2025-07-14T12:34:36.934282"
  },
  {
    "epoch": 1761,
    "batch_idx": 192,
    "loss": 0.20080435276031494,
    "learning_rate": 3.482320635995505e-05,
    "timestep_mean": 491.125,
    "epsilon_norm": 3.188001871109009,
    "prediction_norm": 3.32055926322937,
    "timestamp": "2025-07-14T12:34:36.939515"
  },
  {
    "epoch": 1762,
    "batch_idx": 0,
    "loss": 0.30939555168151855,
    "learning_rate": 3.4535803441532116e-05,
    "timestep_mean": 494.875,
    "epsilon_norm": 7.38129186630249,
    "prediction_norm": 6.36422061920166,
    "timestamp": "2025-07-14T12:34:36.944867"
  },
  {
    "epoch": 1762,
    "batch_idx": 32,
    "loss": 0.21412956714630127,
    "learning_rate": 3.4535803441532116e-05,
    "timestep_mean": 533.46875,
    "epsilon_norm": 8.196578025817871,
    "prediction_norm": 7.161698818206787,
    "timestamp": "2025-07-14T12:34:36.950137"
  },
  {
    "epoch": 1762,
    "batch_idx": 64,
    "loss": 0.1446940302848816,
    "learning_rate": 3.4535803441532116e-05,
    "timestep_mean": 566.9375,
    "epsilon_norm": 8.766861915588379,
    "prediction_norm": 7.8477606773376465,
    "timestamp": "2025-07-14T12:34:36.955306"
  },
  {
    "epoch": 1762,
    "batch_idx": 96,
    "loss": 0.3029857873916626,
    "learning_rate": 3.4535803441532116e-05,
    "timestep_mean": 500.21875,
    "epsilon_norm": 7.5979743003845215,
    "prediction_norm": 6.551084041595459,
    "timestamp": "2025-07-14T12:34:36.960455"
  },
  {
    "epoch": 1762,
    "batch_idx": 128,
    "loss": 0.36034491658210754,
    "learning_rate": 3.4535803441532116e-05,
    "timestep_mean": 510.78125,
    "epsilon_norm": 8.293307304382324,
    "prediction_norm": 6.2310686111450195,
    "timestamp": "2025-07-14T12:34:36.965481"
  },
  {
    "epoch": 1762,
    "batch_idx": 160,
    "loss": 0.1980261653661728,
    "learning_rate": 3.4535803441532116e-05,
    "timestep_mean": 558.40625,
    "epsilon_norm": 8.33372974395752,
    "prediction_norm": 7.656882286071777,
    "timestamp": "2025-07-14T12:34:36.970658"
  },
  {
    "epoch": 1762,
    "batch_idx": 192,
    "loss": 0.2116926908493042,
    "learning_rate": 3.4535803441532116e-05,
    "timestep_mean": 386.0,
    "epsilon_norm": 3.541767120361328,
    "prediction_norm": 3.1966118812561035,
    "timestamp": "2025-07-14T12:34:36.975841"
  },
  {
    "epoch": 1763,
    "batch_idx": 0,
    "loss": 0.1738564670085907,
    "learning_rate": 3.424954900974386e-05,
    "timestep_mean": 540.625,
    "epsilon_norm": 7.470494747161865,
    "prediction_norm": 7.129488945007324,
    "timestamp": "2025-07-14T12:34:36.981079"
  },
  {
    "epoch": 1763,
    "batch_idx": 32,
    "loss": 0.1978897601366043,
    "learning_rate": 3.424954900974386e-05,
    "timestep_mean": 475.1875,
    "epsilon_norm": 7.638734340667725,
    "prediction_norm": 6.978257179260254,
    "timestamp": "2025-07-14T12:34:36.986282"
  },
  {
    "epoch": 1763,
    "batch_idx": 64,
    "loss": 0.11419204622507095,
    "learning_rate": 3.424954900974386e-05,
    "timestep_mean": 541.21875,
    "epsilon_norm": 8.043445587158203,
    "prediction_norm": 7.482170104980469,
    "timestamp": "2025-07-14T12:34:36.991434"
  },
  {
    "epoch": 1763,
    "batch_idx": 96,
    "loss": 0.37629151344299316,
    "learning_rate": 3.424954900974386e-05,
    "timestep_mean": 517.5625,
    "epsilon_norm": 9.223995208740234,
    "prediction_norm": 7.553350925445557,
    "timestamp": "2025-07-14T12:34:36.996655"
  },
  {
    "epoch": 1763,
    "batch_idx": 128,
    "loss": 0.1500355452299118,
    "learning_rate": 3.424954900974386e-05,
    "timestep_mean": 546.5,
    "epsilon_norm": 7.451694011688232,
    "prediction_norm": 7.153910160064697,
    "timestamp": "2025-07-14T12:34:37.002085"
  },
  {
    "epoch": 1763,
    "batch_idx": 160,
    "loss": 0.17182433605194092,
    "learning_rate": 3.424954900974386e-05,
    "timestep_mean": 516.21875,
    "epsilon_norm": 8.120803833007812,
    "prediction_norm": 7.238250255584717,
    "timestamp": "2025-07-14T12:34:37.007426"
  },
  {
    "epoch": 1763,
    "batch_idx": 192,
    "loss": 0.46557995676994324,
    "learning_rate": 3.424954900974386e-05,
    "timestep_mean": 403.625,
    "epsilon_norm": 4.526210308074951,
    "prediction_norm": 3.259946823120117,
    "timestamp": "2025-07-14T12:34:37.012646"
  },
  {
    "epoch": 1764,
    "batch_idx": 0,
    "loss": 0.3153875470161438,
    "learning_rate": 3.3964443770894525e-05,
    "timestep_mean": 477.90625,
    "epsilon_norm": 8.625170707702637,
    "prediction_norm": 7.142354488372803,
    "timestamp": "2025-07-14T12:34:37.017826"
  },
  {
    "epoch": 1764,
    "batch_idx": 32,
    "loss": 0.24850553274154663,
    "learning_rate": 3.3964443770894525e-05,
    "timestep_mean": 453.25,
    "epsilon_norm": 7.395632266998291,
    "prediction_norm": 6.216306686401367,
    "timestamp": "2025-07-14T12:34:37.023225"
  },
  {
    "epoch": 1764,
    "batch_idx": 64,
    "loss": 0.4365931749343872,
    "learning_rate": 3.3964443770894525e-05,
    "timestep_mean": 498.03125,
    "epsilon_norm": 7.719545364379883,
    "prediction_norm": 6.233405590057373,
    "timestamp": "2025-07-14T12:34:37.028624"
  },
  {
    "epoch": 1764,
    "batch_idx": 96,
    "loss": 0.5171900987625122,
    "learning_rate": 3.3964443770894525e-05,
    "timestep_mean": 398.125,
    "epsilon_norm": 7.635699272155762,
    "prediction_norm": 5.28469705581665,
    "timestamp": "2025-07-14T12:34:37.033930"
  },
  {
    "epoch": 1764,
    "batch_idx": 128,
    "loss": 0.41217556595802307,
    "learning_rate": 3.3964443770894525e-05,
    "timestep_mean": 479.8125,
    "epsilon_norm": 8.400313377380371,
    "prediction_norm": 6.577183723449707,
    "timestamp": "2025-07-14T12:34:37.039353"
  },
  {
    "epoch": 1764,
    "batch_idx": 160,
    "loss": 0.20817719399929047,
    "learning_rate": 3.3964443770894525e-05,
    "timestep_mean": 508.9375,
    "epsilon_norm": 8.6697416305542,
    "prediction_norm": 7.56987190246582,
    "timestamp": "2025-07-14T12:34:37.044747"
  },
  {
    "epoch": 1764,
    "batch_idx": 192,
    "loss": 0.1863163262605667,
    "learning_rate": 3.3964443770894525e-05,
    "timestep_mean": 355.75,
    "epsilon_norm": 1.9968444108963013,
    "prediction_norm": 1.4734153747558594,
    "timestamp": "2025-07-14T12:34:37.050172"
  },
  {
    "epoch": 1765,
    "batch_idx": 0,
    "loss": 0.20354154706001282,
    "learning_rate": 3.3680488428453e-05,
    "timestep_mean": 545.59375,
    "epsilon_norm": 8.164095878601074,
    "prediction_norm": 7.128992557525635,
    "timestamp": "2025-07-14T12:34:37.055724"
  },
  {
    "epoch": 1765,
    "batch_idx": 32,
    "loss": 0.3688875436782837,
    "learning_rate": 3.3680488428453e-05,
    "timestep_mean": 533.84375,
    "epsilon_norm": 8.785569190979004,
    "prediction_norm": 7.342879295349121,
    "timestamp": "2025-07-14T12:34:37.060979"
  },
  {
    "epoch": 1765,
    "batch_idx": 64,
    "loss": 0.2433559000492096,
    "learning_rate": 3.3680488428453e-05,
    "timestep_mean": 554.625,
    "epsilon_norm": 7.172569751739502,
    "prediction_norm": 6.293486595153809,
    "timestamp": "2025-07-14T12:34:37.066178"
  },
  {
    "epoch": 1765,
    "batch_idx": 96,
    "loss": 0.33761775493621826,
    "learning_rate": 3.3680488428453e-05,
    "timestep_mean": 584.4375,
    "epsilon_norm": 7.806382656097412,
    "prediction_norm": 6.804224967956543,
    "timestamp": "2025-07-14T12:34:37.071613"
  },
  {
    "epoch": 1765,
    "batch_idx": 128,
    "loss": 0.22145885229110718,
    "learning_rate": 3.3680488428453e-05,
    "timestep_mean": 552.125,
    "epsilon_norm": 7.607716083526611,
    "prediction_norm": 6.32958459854126,
    "timestamp": "2025-07-14T12:34:37.076795"
  },
  {
    "epoch": 1765,
    "batch_idx": 160,
    "loss": 0.38741323351860046,
    "learning_rate": 3.3680488428453e-05,
    "timestep_mean": 503.65625,
    "epsilon_norm": 8.430693626403809,
    "prediction_norm": 6.308229446411133,
    "timestamp": "2025-07-14T12:34:37.082102"
  },
  {
    "epoch": 1765,
    "batch_idx": 192,
    "loss": 0.6273569464683533,
    "learning_rate": 3.3680488428453e-05,
    "timestep_mean": 299.25,
    "epsilon_norm": 3.9690940380096436,
    "prediction_norm": 2.6742823123931885,
    "timestamp": "2025-07-14T12:34:37.087322"
  },
  {
    "epoch": 1766,
    "batch_idx": 0,
    "loss": 0.17228001356124878,
    "learning_rate": 3.339768368305079e-05,
    "timestep_mean": 529.21875,
    "epsilon_norm": 7.931487083435059,
    "prediction_norm": 7.211906433105469,
    "timestamp": "2025-07-14T12:34:37.092535"
  },
  {
    "epoch": 1766,
    "batch_idx": 32,
    "loss": 0.3370126485824585,
    "learning_rate": 3.339768368305079e-05,
    "timestep_mean": 550.6875,
    "epsilon_norm": 7.538177013397217,
    "prediction_norm": 5.910253047943115,
    "timestamp": "2025-07-14T12:34:37.097963"
  },
  {
    "epoch": 1766,
    "batch_idx": 64,
    "loss": 0.14823681116104126,
    "learning_rate": 3.339768368305079e-05,
    "timestep_mean": 550.75,
    "epsilon_norm": 7.3588972091674805,
    "prediction_norm": 6.753454208374023,
    "timestamp": "2025-07-14T12:34:37.103450"
  },
  {
    "epoch": 1766,
    "batch_idx": 96,
    "loss": 0.25697964429855347,
    "learning_rate": 3.339768368305079e-05,
    "timestep_mean": 554.4375,
    "epsilon_norm": 8.139796257019043,
    "prediction_norm": 7.317577838897705,
    "timestamp": "2025-07-14T12:34:37.108721"
  },
  {
    "epoch": 1766,
    "batch_idx": 128,
    "loss": 0.2713719606399536,
    "learning_rate": 3.339768368305079e-05,
    "timestep_mean": 475.78125,
    "epsilon_norm": 7.31890344619751,
    "prediction_norm": 6.371048450469971,
    "timestamp": "2025-07-14T12:34:37.113947"
  },
  {
    "epoch": 1766,
    "batch_idx": 160,
    "loss": 0.23737114667892456,
    "learning_rate": 3.339768368305079e-05,
    "timestep_mean": 599.125,
    "epsilon_norm": 7.960928916931152,
    "prediction_norm": 6.634997844696045,
    "timestamp": "2025-07-14T12:34:37.119005"
  },
  {
    "epoch": 1766,
    "batch_idx": 192,
    "loss": 0.4274592399597168,
    "learning_rate": 3.339768368305079e-05,
    "timestep_mean": 513.125,
    "epsilon_norm": 3.3716628551483154,
    "prediction_norm": 2.265693187713623,
    "timestamp": "2025-07-14T12:34:37.126463"
  },
  {
    "epoch": 1767,
    "batch_idx": 0,
    "loss": 0.325093537569046,
    "learning_rate": 3.311603023248044e-05,
    "timestep_mean": 503.875,
    "epsilon_norm": 7.0363335609436035,
    "prediction_norm": 5.6612043380737305,
    "timestamp": "2025-07-14T12:34:37.138645"
  },
  {
    "epoch": 1767,
    "batch_idx": 32,
    "loss": 0.21278658509254456,
    "learning_rate": 3.311603023248044e-05,
    "timestep_mean": 541.53125,
    "epsilon_norm": 6.853972434997559,
    "prediction_norm": 5.995785236358643,
    "timestamp": "2025-07-14T12:34:37.147377"
  },
  {
    "epoch": 1767,
    "batch_idx": 64,
    "loss": 0.4149266183376312,
    "learning_rate": 3.311603023248044e-05,
    "timestep_mean": 554.28125,
    "epsilon_norm": 8.803925514221191,
    "prediction_norm": 7.769152641296387,
    "timestamp": "2025-07-14T12:34:37.155789"
  },
  {
    "epoch": 1767,
    "batch_idx": 96,
    "loss": 0.3073052763938904,
    "learning_rate": 3.311603023248044e-05,
    "timestep_mean": 487.1875,
    "epsilon_norm": 8.700148582458496,
    "prediction_norm": 7.33489465713501,
    "timestamp": "2025-07-14T12:34:37.161571"
  },
  {
    "epoch": 1767,
    "batch_idx": 128,
    "loss": 0.2615109980106354,
    "learning_rate": 3.311603023248044e-05,
    "timestep_mean": 547.34375,
    "epsilon_norm": 6.997219562530518,
    "prediction_norm": 5.719174861907959,
    "timestamp": "2025-07-14T12:34:37.168152"
  },
  {
    "epoch": 1767,
    "batch_idx": 160,
    "loss": 0.3053300380706787,
    "learning_rate": 3.311603023248044e-05,
    "timestep_mean": 414.28125,
    "epsilon_norm": 8.40672492980957,
    "prediction_norm": 6.71470308303833,
    "timestamp": "2025-07-14T12:34:37.176495"
  },
  {
    "epoch": 1767,
    "batch_idx": 192,
    "loss": 0.32160845398902893,
    "learning_rate": 3.311603023248044e-05,
    "timestep_mean": 348.625,
    "epsilon_norm": 3.5812435150146484,
    "prediction_norm": 3.5341105461120605,
    "timestamp": "2025-07-14T12:34:37.184718"
  },
  {
    "epoch": 1768,
    "batch_idx": 0,
    "loss": 0.24856770038604736,
    "learning_rate": 3.2835528771693986e-05,
    "timestep_mean": 513.875,
    "epsilon_norm": 8.102952003479004,
    "prediction_norm": 7.2167768478393555,
    "timestamp": "2025-07-14T12:34:37.193167"
  },
  {
    "epoch": 1768,
    "batch_idx": 32,
    "loss": 0.14859849214553833,
    "learning_rate": 3.2835528771693986e-05,
    "timestep_mean": 492.8125,
    "epsilon_norm": 8.205936431884766,
    "prediction_norm": 7.503242492675781,
    "timestamp": "2025-07-14T12:34:37.201355"
  },
  {
    "epoch": 1768,
    "batch_idx": 64,
    "loss": 0.43287941813468933,
    "learning_rate": 3.2835528771693986e-05,
    "timestep_mean": 426.3125,
    "epsilon_norm": 7.9778618812561035,
    "prediction_norm": 4.9890851974487305,
    "timestamp": "2025-07-14T12:34:37.209900"
  },
  {
    "epoch": 1768,
    "batch_idx": 96,
    "loss": 0.2278919816017151,
    "learning_rate": 3.2835528771693986e-05,
    "timestep_mean": 481.0,
    "epsilon_norm": 7.754626750946045,
    "prediction_norm": 6.654979705810547,
    "timestamp": "2025-07-14T12:34:37.218232"
  },
  {
    "epoch": 1768,
    "batch_idx": 128,
    "loss": 0.322164386510849,
    "learning_rate": 3.2835528771693986e-05,
    "timestep_mean": 509.25,
    "epsilon_norm": 8.357353210449219,
    "prediction_norm": 6.707430839538574,
    "timestamp": "2025-07-14T12:34:37.225701"
  },
  {
    "epoch": 1768,
    "batch_idx": 160,
    "loss": 0.3264131247997284,
    "learning_rate": 3.2835528771693986e-05,
    "timestep_mean": 502.15625,
    "epsilon_norm": 8.76228141784668,
    "prediction_norm": 7.475141525268555,
    "timestamp": "2025-07-14T12:34:37.231292"
  },
  {
    "epoch": 1768,
    "batch_idx": 192,
    "loss": 0.23266565799713135,
    "learning_rate": 3.2835528771693986e-05,
    "timestep_mean": 500.375,
    "epsilon_norm": 3.539044141769409,
    "prediction_norm": 3.533775806427002,
    "timestamp": "2025-07-14T12:34:37.236468"
  },
  {
    "epoch": 1769,
    "batch_idx": 0,
    "loss": 0.5384427309036255,
    "learning_rate": 3.2556179992800854e-05,
    "timestep_mean": 478.3125,
    "epsilon_norm": 8.570352554321289,
    "prediction_norm": 6.115082263946533,
    "timestamp": "2025-07-14T12:34:37.241609"
  },
  {
    "epoch": 1769,
    "batch_idx": 32,
    "loss": 0.24636700749397278,
    "learning_rate": 3.2556179992800854e-05,
    "timestep_mean": 528.5,
    "epsilon_norm": 8.468448638916016,
    "prediction_norm": 7.753368854522705,
    "timestamp": "2025-07-14T12:34:37.246546"
  },
  {
    "epoch": 1769,
    "batch_idx": 64,
    "loss": 0.29642432928085327,
    "learning_rate": 3.2556179992800854e-05,
    "timestep_mean": 554.40625,
    "epsilon_norm": 9.099562644958496,
    "prediction_norm": 8.085851669311523,
    "timestamp": "2025-07-14T12:34:37.251328"
  },
  {
    "epoch": 1769,
    "batch_idx": 96,
    "loss": 0.3865915834903717,
    "learning_rate": 3.2556179992800854e-05,
    "timestep_mean": 489.28125,
    "epsilon_norm": 8.893671035766602,
    "prediction_norm": 6.728567600250244,
    "timestamp": "2025-07-14T12:34:37.256195"
  },
  {
    "epoch": 1769,
    "batch_idx": 128,
    "loss": 0.25621649622917175,
    "learning_rate": 3.2556179992800854e-05,
    "timestep_mean": 462.25,
    "epsilon_norm": 8.309836387634277,
    "prediction_norm": 7.177751541137695,
    "timestamp": "2025-07-14T12:34:37.260958"
  },
  {
    "epoch": 1769,
    "batch_idx": 160,
    "loss": 0.15511208772659302,
    "learning_rate": 3.2556179992800854e-05,
    "timestep_mean": 526.5,
    "epsilon_norm": 6.913573741912842,
    "prediction_norm": 6.3853936195373535,
    "timestamp": "2025-07-14T12:34:37.265804"
  },
  {
    "epoch": 1769,
    "batch_idx": 192,
    "loss": 0.4289610981941223,
    "learning_rate": 3.2556179992800854e-05,
    "timestep_mean": 511.25,
    "epsilon_norm": 3.6978812217712402,
    "prediction_norm": 2.5579185485839844,
    "timestamp": "2025-07-14T12:34:37.270719"
  },
  {
    "epoch": 1770,
    "batch_idx": 0,
    "loss": 0.1722450852394104,
    "learning_rate": 3.227798458506636e-05,
    "timestep_mean": 495.1875,
    "epsilon_norm": 7.080245018005371,
    "prediction_norm": 6.664515018463135,
    "timestamp": "2025-07-14T12:34:37.275726"
  },
  {
    "epoch": 1770,
    "batch_idx": 32,
    "loss": 0.20304718613624573,
    "learning_rate": 3.227798458506636e-05,
    "timestep_mean": 522.75,
    "epsilon_norm": 7.965148448944092,
    "prediction_norm": 7.091268539428711,
    "timestamp": "2025-07-14T12:34:37.280521"
  },
  {
    "epoch": 1770,
    "batch_idx": 64,
    "loss": 0.25314509868621826,
    "learning_rate": 3.227798458506636e-05,
    "timestep_mean": 462.375,
    "epsilon_norm": 8.120680809020996,
    "prediction_norm": 6.944748401641846,
    "timestamp": "2025-07-14T12:34:37.285472"
  },
  {
    "epoch": 1770,
    "batch_idx": 96,
    "loss": 0.3790052533149719,
    "learning_rate": 3.227798458506636e-05,
    "timestep_mean": 595.125,
    "epsilon_norm": 8.059356689453125,
    "prediction_norm": 6.863595485687256,
    "timestamp": "2025-07-14T12:34:37.290334"
  },
  {
    "epoch": 1770,
    "batch_idx": 128,
    "loss": 0.2895303964614868,
    "learning_rate": 3.227798458506636e-05,
    "timestep_mean": 443.96875,
    "epsilon_norm": 8.150993347167969,
    "prediction_norm": 7.441263198852539,
    "timestamp": "2025-07-14T12:34:37.295260"
  },
  {
    "epoch": 1770,
    "batch_idx": 160,
    "loss": 0.30970293283462524,
    "learning_rate": 3.227798458506636e-05,
    "timestep_mean": 436.875,
    "epsilon_norm": 8.348505020141602,
    "prediction_norm": 7.663867950439453,
    "timestamp": "2025-07-14T12:34:37.300229"
  },
  {
    "epoch": 1770,
    "batch_idx": 192,
    "loss": 0.1296282708644867,
    "learning_rate": 3.227798458506636e-05,
    "timestep_mean": 559.5,
    "epsilon_norm": 3.265828847885132,
    "prediction_norm": 2.7454349994659424,
    "timestamp": "2025-07-14T12:34:37.305179"
  },
  {
    "epoch": 1771,
    "batch_idx": 0,
    "loss": 0.27167779207229614,
    "learning_rate": 3.2000943234909976e-05,
    "timestep_mean": 573.3125,
    "epsilon_norm": 8.00503158569336,
    "prediction_norm": 6.897132396697998,
    "timestamp": "2025-07-14T12:34:37.310102"
  },
  {
    "epoch": 1771,
    "batch_idx": 32,
    "loss": 0.2823181450366974,
    "learning_rate": 3.2000943234909976e-05,
    "timestep_mean": 466.53125,
    "epsilon_norm": 8.19022274017334,
    "prediction_norm": 6.703300476074219,
    "timestamp": "2025-07-14T12:34:37.315030"
  },
  {
    "epoch": 1771,
    "batch_idx": 64,
    "loss": 0.25777167081832886,
    "learning_rate": 3.2000943234909976e-05,
    "timestep_mean": 466.6875,
    "epsilon_norm": 7.444586753845215,
    "prediction_norm": 6.279054641723633,
    "timestamp": "2025-07-14T12:34:37.320129"
  },
  {
    "epoch": 1771,
    "batch_idx": 96,
    "loss": 0.15944620966911316,
    "learning_rate": 3.2000943234909976e-05,
    "timestep_mean": 583.625,
    "epsilon_norm": 8.595853805541992,
    "prediction_norm": 7.895602226257324,
    "timestamp": "2025-07-14T12:34:37.325319"
  },
  {
    "epoch": 1771,
    "batch_idx": 128,
    "loss": 0.3508099913597107,
    "learning_rate": 3.2000943234909976e-05,
    "timestep_mean": 523.65625,
    "epsilon_norm": 9.341681480407715,
    "prediction_norm": 7.681992530822754,
    "timestamp": "2025-07-14T12:34:37.330421"
  },
  {
    "epoch": 1771,
    "batch_idx": 160,
    "loss": 0.30836164951324463,
    "learning_rate": 3.2000943234909976e-05,
    "timestep_mean": 479.15625,
    "epsilon_norm": 7.484918594360352,
    "prediction_norm": 5.984127998352051,
    "timestamp": "2025-07-14T12:34:37.335538"
  },
  {
    "epoch": 1771,
    "batch_idx": 192,
    "loss": 0.03825681284070015,
    "learning_rate": 3.2000943234909976e-05,
    "timestep_mean": 532.0,
    "epsilon_norm": 1.9830193519592285,
    "prediction_norm": 2.1791586875915527,
    "timestamp": "2025-07-14T12:34:37.340862"
  },
  {
    "epoch": 1772,
    "batch_idx": 0,
    "loss": 0.2394658774137497,
    "learning_rate": 3.172505662590385e-05,
    "timestep_mean": 508.09375,
    "epsilon_norm": 7.870150566101074,
    "prediction_norm": 6.937553405761719,
    "timestamp": "2025-07-14T12:34:37.345920"
  },
  {
    "epoch": 1772,
    "batch_idx": 32,
    "loss": 0.43159765005111694,
    "learning_rate": 3.172505662590385e-05,
    "timestep_mean": 470.0625,
    "epsilon_norm": 7.645304203033447,
    "prediction_norm": 6.293778419494629,
    "timestamp": "2025-07-14T12:34:37.351004"
  },
  {
    "epoch": 1772,
    "batch_idx": 64,
    "loss": 0.22587831318378448,
    "learning_rate": 3.172505662590385e-05,
    "timestep_mean": 464.53125,
    "epsilon_norm": 7.554229736328125,
    "prediction_norm": 6.702284812927246,
    "timestamp": "2025-07-14T12:34:37.356014"
  },
  {
    "epoch": 1772,
    "batch_idx": 96,
    "loss": 0.3029364347457886,
    "learning_rate": 3.172505662590385e-05,
    "timestep_mean": 525.15625,
    "epsilon_norm": 8.406270027160645,
    "prediction_norm": 7.25408411026001,
    "timestamp": "2025-07-14T12:34:37.360930"
  },
  {
    "epoch": 1772,
    "batch_idx": 128,
    "loss": 0.6039702892303467,
    "learning_rate": 3.172505662590385e-05,
    "timestep_mean": 455.375,
    "epsilon_norm": 10.06291389465332,
    "prediction_norm": 8.309982299804688,
    "timestamp": "2025-07-14T12:34:37.368000"
  },
  {
    "epoch": 1772,
    "batch_idx": 160,
    "loss": 0.15090006589889526,
    "learning_rate": 3.172505662590385e-05,
    "timestep_mean": 574.625,
    "epsilon_norm": 8.295726776123047,
    "prediction_norm": 7.317162036895752,
    "timestamp": "2025-07-14T12:34:37.373967"
  },
  {
    "epoch": 1772,
    "batch_idx": 192,
    "loss": 0.3490774631500244,
    "learning_rate": 3.172505662590385e-05,
    "timestep_mean": 372.125,
    "epsilon_norm": 4.085144996643066,
    "prediction_norm": 2.911466598510742,
    "timestamp": "2025-07-14T12:34:37.379229"
  },
  {
    "epoch": 1773,
    "batch_idx": 0,
    "loss": 0.2863895297050476,
    "learning_rate": 3.145032543877062e-05,
    "timestep_mean": 513.40625,
    "epsilon_norm": 8.191280364990234,
    "prediction_norm": 6.7568817138671875,
    "timestamp": "2025-07-14T12:34:37.389475"
  },
  {
    "epoch": 1773,
    "batch_idx": 32,
    "loss": 0.3437198996543884,
    "learning_rate": 3.145032543877062e-05,
    "timestep_mean": 390.34375,
    "epsilon_norm": 8.105696678161621,
    "prediction_norm": 6.68747091293335,
    "timestamp": "2025-07-14T12:34:37.394568"
  },
  {
    "epoch": 1773,
    "batch_idx": 64,
    "loss": 0.3670070171356201,
    "learning_rate": 3.145032543877062e-05,
    "timestep_mean": 493.8125,
    "epsilon_norm": 8.442874908447266,
    "prediction_norm": 7.051529407501221,
    "timestamp": "2025-07-14T12:34:37.399647"
  },
  {
    "epoch": 1773,
    "batch_idx": 96,
    "loss": 0.3149302303791046,
    "learning_rate": 3.145032543877062e-05,
    "timestep_mean": 582.5,
    "epsilon_norm": 8.230669975280762,
    "prediction_norm": 7.21956729888916,
    "timestamp": "2025-07-14T12:34:37.404689"
  },
  {
    "epoch": 1773,
    "batch_idx": 128,
    "loss": 0.4301586449146271,
    "learning_rate": 3.145032543877062e-05,
    "timestep_mean": 444.09375,
    "epsilon_norm": 8.62681770324707,
    "prediction_norm": 6.349608898162842,
    "timestamp": "2025-07-14T12:34:37.409716"
  },
  {
    "epoch": 1773,
    "batch_idx": 160,
    "loss": 0.3296664357185364,
    "learning_rate": 3.145032543877062e-05,
    "timestep_mean": 459.25,
    "epsilon_norm": 7.426830291748047,
    "prediction_norm": 6.671950817108154,
    "timestamp": "2025-07-14T12:34:37.415079"
  },
  {
    "epoch": 1773,
    "batch_idx": 192,
    "loss": 0.2108481079339981,
    "learning_rate": 3.145032543877062e-05,
    "timestep_mean": 469.25,
    "epsilon_norm": 4.735213279724121,
    "prediction_norm": 3.942664861679077,
    "timestamp": "2025-07-14T12:34:37.420615"
  },
  {
    "epoch": 1774,
    "batch_idx": 0,
    "loss": 0.3469948172569275,
    "learning_rate": 3.117675035138211e-05,
    "timestep_mean": 472.375,
    "epsilon_norm": 8.893298149108887,
    "prediction_norm": 7.2433576583862305,
    "timestamp": "2025-07-14T12:34:37.426289"
  },
  {
    "epoch": 1774,
    "batch_idx": 32,
    "loss": 0.3149828016757965,
    "learning_rate": 3.117675035138211e-05,
    "timestep_mean": 556.9375,
    "epsilon_norm": 8.653600692749023,
    "prediction_norm": 7.344676494598389,
    "timestamp": "2025-07-14T12:34:37.431819"
  },
  {
    "epoch": 1774,
    "batch_idx": 64,
    "loss": 0.3709774911403656,
    "learning_rate": 3.117675035138211e-05,
    "timestep_mean": 441.3125,
    "epsilon_norm": 7.808054447174072,
    "prediction_norm": 6.242593288421631,
    "timestamp": "2025-07-14T12:34:37.437210"
  },
  {
    "epoch": 1774,
    "batch_idx": 96,
    "loss": 0.2927088141441345,
    "learning_rate": 3.117675035138211e-05,
    "timestep_mean": 497.46875,
    "epsilon_norm": 6.692497253417969,
    "prediction_norm": 5.212823867797852,
    "timestamp": "2025-07-14T12:34:37.442524"
  },
  {
    "epoch": 1774,
    "batch_idx": 128,
    "loss": 0.22307491302490234,
    "learning_rate": 3.117675035138211e-05,
    "timestep_mean": 505.75,
    "epsilon_norm": 8.09986686706543,
    "prediction_norm": 7.022485733032227,
    "timestamp": "2025-07-14T12:34:37.447798"
  },
  {
    "epoch": 1774,
    "batch_idx": 160,
    "loss": 0.2111627757549286,
    "learning_rate": 3.117675035138211e-05,
    "timestep_mean": 506.21875,
    "epsilon_norm": 7.544376373291016,
    "prediction_norm": 6.75730037689209,
    "timestamp": "2025-07-14T12:34:37.452937"
  },
  {
    "epoch": 1774,
    "batch_idx": 192,
    "loss": 0.06517663598060608,
    "learning_rate": 3.117675035138211e-05,
    "timestep_mean": 517.5,
    "epsilon_norm": 3.179403781890869,
    "prediction_norm": 3.2675533294677734,
    "timestamp": "2025-07-14T12:34:37.458509"
  },
  {
    "epoch": 1775,
    "batch_idx": 0,
    "loss": 0.3128877282142639,
    "learning_rate": 3.090433203875797e-05,
    "timestep_mean": 460.46875,
    "epsilon_norm": 7.9981231689453125,
    "prediction_norm": 6.4410505294799805,
    "timestamp": "2025-07-14T12:34:37.463829"
  },
  {
    "epoch": 1775,
    "batch_idx": 32,
    "loss": 0.4135262370109558,
    "learning_rate": 3.090433203875797e-05,
    "timestep_mean": 448.5,
    "epsilon_norm": 10.695096015930176,
    "prediction_norm": 8.82981014251709,
    "timestamp": "2025-07-14T12:34:37.469044"
  },
  {
    "epoch": 1775,
    "batch_idx": 64,
    "loss": 0.29784199595451355,
    "learning_rate": 3.090433203875797e-05,
    "timestep_mean": 420.6875,
    "epsilon_norm": 7.5467705726623535,
    "prediction_norm": 6.220262050628662,
    "timestamp": "2025-07-14T12:34:37.474443"
  },
  {
    "epoch": 1775,
    "batch_idx": 96,
    "loss": 0.4386520981788635,
    "learning_rate": 3.090433203875797e-05,
    "timestep_mean": 466.65625,
    "epsilon_norm": 9.185237884521484,
    "prediction_norm": 7.303715229034424,
    "timestamp": "2025-07-14T12:34:37.479810"
  },
  {
    "epoch": 1775,
    "batch_idx": 128,
    "loss": 0.23227718472480774,
    "learning_rate": 3.090433203875797e-05,
    "timestep_mean": 517.03125,
    "epsilon_norm": 7.850208282470703,
    "prediction_norm": 6.746005058288574,
    "timestamp": "2025-07-14T12:34:37.484949"
  },
  {
    "epoch": 1775,
    "batch_idx": 160,
    "loss": 0.16467174887657166,
    "learning_rate": 3.090433203875797e-05,
    "timestep_mean": 530.0,
    "epsilon_norm": 7.076819896697998,
    "prediction_norm": 6.129961013793945,
    "timestamp": "2025-07-14T12:34:37.490130"
  },
  {
    "epoch": 1775,
    "batch_idx": 192,
    "loss": 0.6436286568641663,
    "learning_rate": 3.090433203875797e-05,
    "timestep_mean": 384.625,
    "epsilon_norm": 3.8453121185302734,
    "prediction_norm": 2.513127088546753,
    "timestamp": "2025-07-14T12:34:37.495372"
  },
  {
    "epoch": 1776,
    "batch_idx": 0,
    "loss": 0.4287436604499817,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 545.34375,
    "epsilon_norm": 8.95063304901123,
    "prediction_norm": 6.7170915603637695,
    "timestamp": "2025-07-14T12:34:37.500625"
  },
  {
    "epoch": 1776,
    "batch_idx": 32,
    "loss": 0.18933159112930298,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 538.8125,
    "epsilon_norm": 6.883492946624756,
    "prediction_norm": 6.527523517608643,
    "timestamp": "2025-07-14T12:34:37.505893"
  },
  {
    "epoch": 1776,
    "batch_idx": 64,
    "loss": 0.3871568441390991,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 461.09375,
    "epsilon_norm": 8.55063533782959,
    "prediction_norm": 6.576913356781006,
    "timestamp": "2025-07-14T12:34:37.511082"
  },
  {
    "epoch": 1776,
    "batch_idx": 96,
    "loss": 0.5260165929794312,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 449.34375,
    "epsilon_norm": 8.536474227905273,
    "prediction_norm": 6.245954513549805,
    "timestamp": "2025-07-14T12:34:37.516374"
  },
  {
    "epoch": 1776,
    "batch_idx": 128,
    "loss": 0.13807767629623413,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 530.59375,
    "epsilon_norm": 8.114047050476074,
    "prediction_norm": 7.491327285766602,
    "timestamp": "2025-07-14T12:34:37.521745"
  },
  {
    "epoch": 1776,
    "batch_idx": 160,
    "loss": 0.20427051186561584,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 554.125,
    "epsilon_norm": 7.799087047576904,
    "prediction_norm": 7.092652797698975,
    "timestamp": "2025-07-14T12:34:37.527014"
  },
  {
    "epoch": 1776,
    "batch_idx": 192,
    "loss": 0.2893484830856323,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 536.625,
    "epsilon_norm": 4.175809860229492,
    "prediction_norm": 3.500171184539795,
    "timestamp": "2025-07-14T12:34:37.532172"
  },
  {
    "epoch": 1777,
    "batch_idx": 0,
    "loss": 0.23549699783325195,
    "learning_rate": 3.0362968423606577e-05,
    "timestep_mean": 463.3125,
    "epsilon_norm": 9.082730293273926,
    "prediction_norm": 7.871975898742676,
    "timestamp": "2025-07-14T12:34:37.537512"
  },
  {
    "epoch": 1777,
    "batch_idx": 32,
    "loss": 0.333579421043396,
    "learning_rate": 3.0362968423606577e-05,
    "timestep_mean": 514.71875,
    "epsilon_norm": 7.953340530395508,
    "prediction_norm": 6.296391487121582,
    "timestamp": "2025-07-14T12:34:37.542743"
  },
  {
    "epoch": 1777,
    "batch_idx": 64,
    "loss": 0.31845182180404663,
    "learning_rate": 3.0362968423606577e-05,
    "timestep_mean": 506.5,
    "epsilon_norm": 8.842511177062988,
    "prediction_norm": 7.653966426849365,
    "timestamp": "2025-07-14T12:34:37.547933"
  },
  {
    "epoch": 1777,
    "batch_idx": 96,
    "loss": 0.2796539068222046,
    "learning_rate": 3.0362968423606577e-05,
    "timestep_mean": 450.65625,
    "epsilon_norm": 8.077537536621094,
    "prediction_norm": 6.477383136749268,
    "timestamp": "2025-07-14T12:34:37.553244"
  },
  {
    "epoch": 1777,
    "batch_idx": 128,
    "loss": 0.11328072100877762,
    "learning_rate": 3.0362968423606577e-05,
    "timestep_mean": 483.0,
    "epsilon_norm": 7.24813175201416,
    "prediction_norm": 6.649479866027832,
    "timestamp": "2025-07-14T12:34:37.558532"
  },
  {
    "epoch": 1777,
    "batch_idx": 160,
    "loss": 0.22458970546722412,
    "learning_rate": 3.0362968423606577e-05,
    "timestep_mean": 492.84375,
    "epsilon_norm": 9.46458625793457,
    "prediction_norm": 8.43986701965332,
    "timestamp": "2025-07-14T12:34:37.563798"
  },
  {
    "epoch": 1777,
    "batch_idx": 192,
    "loss": 0.051696449518203735,
    "learning_rate": 3.0362968423606577e-05,
    "timestep_mean": 532.875,
    "epsilon_norm": 4.206711769104004,
    "prediction_norm": 4.225245475769043,
    "timestamp": "2025-07-14T12:34:37.568816"
  },
  {
    "epoch": 1778,
    "batch_idx": 0,
    "loss": 0.3933919072151184,
    "learning_rate": 3.009402445684017e-05,
    "timestep_mean": 446.9375,
    "epsilon_norm": 7.836798191070557,
    "prediction_norm": 6.707091331481934,
    "timestamp": "2025-07-14T12:34:37.573816"
  },
  {
    "epoch": 1778,
    "batch_idx": 32,
    "loss": 0.3844641149044037,
    "learning_rate": 3.009402445684017e-05,
    "timestep_mean": 406.0625,
    "epsilon_norm": 7.538562774658203,
    "prediction_norm": 5.08350944519043,
    "timestamp": "2025-07-14T12:34:37.578959"
  },
  {
    "epoch": 1778,
    "batch_idx": 64,
    "loss": 0.4536215662956238,
    "learning_rate": 3.009402445684017e-05,
    "timestep_mean": 457.03125,
    "epsilon_norm": 7.497616291046143,
    "prediction_norm": 5.7800750732421875,
    "timestamp": "2025-07-14T12:34:37.584185"
  },
  {
    "epoch": 1778,
    "batch_idx": 96,
    "loss": 0.19503843784332275,
    "learning_rate": 3.009402445684017e-05,
    "timestep_mean": 508.3125,
    "epsilon_norm": 8.880202293395996,
    "prediction_norm": 7.668119430541992,
    "timestamp": "2025-07-14T12:34:37.589361"
  },
  {
    "epoch": 1778,
    "batch_idx": 128,
    "loss": 0.18910396099090576,
    "learning_rate": 3.009402445684017e-05,
    "timestep_mean": 496.75,
    "epsilon_norm": 8.14913272857666,
    "prediction_norm": 7.56789493560791,
    "timestamp": "2025-07-14T12:34:37.594663"
  },
  {
    "epoch": 1778,
    "batch_idx": 160,
    "loss": 0.42458873987197876,
    "learning_rate": 3.009402445684017e-05,
    "timestep_mean": 508.15625,
    "epsilon_norm": 8.883184432983398,
    "prediction_norm": 6.507022857666016,
    "timestamp": "2025-07-14T12:34:37.599898"
  },
  {
    "epoch": 1778,
    "batch_idx": 192,
    "loss": 0.27646732330322266,
    "learning_rate": 3.009402445684017e-05,
    "timestep_mean": 565.375,
    "epsilon_norm": 3.345856189727783,
    "prediction_norm": 2.5350184440612793,
    "timestamp": "2025-07-14T12:34:37.605258"
  },
  {
    "epoch": 1779,
    "batch_idx": 0,
    "loss": 0.22943270206451416,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 502.0625,
    "epsilon_norm": 7.3008294105529785,
    "prediction_norm": 5.771762847900391,
    "timestamp": "2025-07-14T12:34:37.610406"
  },
  {
    "epoch": 1779,
    "batch_idx": 32,
    "loss": 0.1238013282418251,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 601.15625,
    "epsilon_norm": 8.468749046325684,
    "prediction_norm": 8.041177749633789,
    "timestamp": "2025-07-14T12:34:37.615584"
  },
  {
    "epoch": 1779,
    "batch_idx": 64,
    "loss": 0.3242622911930084,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 472.3125,
    "epsilon_norm": 8.968583106994629,
    "prediction_norm": 7.205071449279785,
    "timestamp": "2025-07-14T12:34:37.620801"
  },
  {
    "epoch": 1779,
    "batch_idx": 96,
    "loss": 0.1788264811038971,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 505.21875,
    "epsilon_norm": 7.272587299346924,
    "prediction_norm": 6.246540069580078,
    "timestamp": "2025-07-14T12:34:37.626135"
  },
  {
    "epoch": 1779,
    "batch_idx": 128,
    "loss": 0.49719440937042236,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 479.40625,
    "epsilon_norm": 8.447595596313477,
    "prediction_norm": 5.838160037994385,
    "timestamp": "2025-07-14T12:34:37.631316"
  },
  {
    "epoch": 1779,
    "batch_idx": 160,
    "loss": 0.3989759683609009,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 511.8125,
    "epsilon_norm": 9.203469276428223,
    "prediction_norm": 7.506675720214844,
    "timestamp": "2025-07-14T12:34:37.636492"
  },
  {
    "epoch": 1779,
    "batch_idx": 192,
    "loss": 0.9806464314460754,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 385.25,
    "epsilon_norm": 4.999290943145752,
    "prediction_norm": 3.2110183238983154,
    "timestamp": "2025-07-14T12:34:37.641659"
  },
  {
    "epoch": 1780,
    "batch_idx": 0,
    "loss": 0.19132566452026367,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 555.84375,
    "epsilon_norm": 8.4574556350708,
    "prediction_norm": 7.633062362670898,
    "timestamp": "2025-07-14T12:34:37.646925"
  },
  {
    "epoch": 1780,
    "batch_idx": 32,
    "loss": 0.32854196429252625,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 518.46875,
    "epsilon_norm": 7.578429698944092,
    "prediction_norm": 6.3696088790893555,
    "timestamp": "2025-07-14T12:34:37.652120"
  },
  {
    "epoch": 1780,
    "batch_idx": 64,
    "loss": 0.2533420920372009,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 363.625,
    "epsilon_norm": 8.036388397216797,
    "prediction_norm": 6.9158034324646,
    "timestamp": "2025-07-14T12:34:37.657255"
  },
  {
    "epoch": 1780,
    "batch_idx": 96,
    "loss": 0.2948949635028839,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 510.78125,
    "epsilon_norm": 7.120691776275635,
    "prediction_norm": 5.983279228210449,
    "timestamp": "2025-07-14T12:34:37.662518"
  },
  {
    "epoch": 1780,
    "batch_idx": 128,
    "loss": 0.22128817439079285,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 482.03125,
    "epsilon_norm": 7.356836795806885,
    "prediction_norm": 6.143622398376465,
    "timestamp": "2025-07-14T12:34:37.667696"
  },
  {
    "epoch": 1780,
    "batch_idx": 160,
    "loss": 0.47443369030952454,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 459.65625,
    "epsilon_norm": 8.763652801513672,
    "prediction_norm": 6.625335216522217,
    "timestamp": "2025-07-14T12:34:37.672819"
  },
  {
    "epoch": 1780,
    "batch_idx": 192,
    "loss": 0.21800640225410461,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 533.75,
    "epsilon_norm": 3.6737897396087646,
    "prediction_norm": 2.7779884338378906,
    "timestamp": "2025-07-14T12:34:37.677948"
  },
  {
    "epoch": 1781,
    "batch_idx": 0,
    "loss": 0.38893163204193115,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 393.09375,
    "epsilon_norm": 7.832700252532959,
    "prediction_norm": 6.505660533905029,
    "timestamp": "2025-07-14T12:34:37.683038"
  },
  {
    "epoch": 1781,
    "batch_idx": 32,
    "loss": 0.2058342695236206,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 509.1875,
    "epsilon_norm": 6.641979217529297,
    "prediction_norm": 6.035150527954102,
    "timestamp": "2025-07-14T12:34:37.688094"
  },
  {
    "epoch": 1781,
    "batch_idx": 64,
    "loss": 0.18861021101474762,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 511.46875,
    "epsilon_norm": 8.757580757141113,
    "prediction_norm": 7.621706962585449,
    "timestamp": "2025-07-14T12:34:37.692966"
  },
  {
    "epoch": 1781,
    "batch_idx": 96,
    "loss": 0.415011465549469,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 529.8125,
    "epsilon_norm": 8.959002494812012,
    "prediction_norm": 7.343754291534424,
    "timestamp": "2025-07-14T12:34:37.697942"
  },
  {
    "epoch": 1781,
    "batch_idx": 128,
    "loss": 0.4761529862880707,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 470.375,
    "epsilon_norm": 9.353041648864746,
    "prediction_norm": 7.3770294189453125,
    "timestamp": "2025-07-14T12:34:37.702864"
  },
  {
    "epoch": 1781,
    "batch_idx": 160,
    "loss": 0.17408382892608643,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 432.78125,
    "epsilon_norm": 7.893192768096924,
    "prediction_norm": 7.181208610534668,
    "timestamp": "2025-07-14T12:34:37.707891"
  },
  {
    "epoch": 1781,
    "batch_idx": 192,
    "loss": 0.12822076678276062,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 479.375,
    "epsilon_norm": 3.462404489517212,
    "prediction_norm": 2.9810664653778076,
    "timestamp": "2025-07-14T12:34:37.712938"
  },
  {
    "epoch": 1782,
    "batch_idx": 0,
    "loss": 0.286691278219223,
    "learning_rate": 2.9029849645604726e-05,
    "timestep_mean": 430.28125,
    "epsilon_norm": 7.375050067901611,
    "prediction_norm": 5.235068321228027,
    "timestamp": "2025-07-14T12:34:37.717882"
  },
  {
    "epoch": 1782,
    "batch_idx": 32,
    "loss": 0.08120296895503998,
    "learning_rate": 2.9029849645604726e-05,
    "timestep_mean": 542.28125,
    "epsilon_norm": 7.6504716873168945,
    "prediction_norm": 7.524112224578857,
    "timestamp": "2025-07-14T12:34:37.722906"
  },
  {
    "epoch": 1782,
    "batch_idx": 64,
    "loss": 0.1678207516670227,
    "learning_rate": 2.9029849645604726e-05,
    "timestep_mean": 513.46875,
    "epsilon_norm": 6.895781517028809,
    "prediction_norm": 6.56861686706543,
    "timestamp": "2025-07-14T12:34:37.727782"
  },
  {
    "epoch": 1782,
    "batch_idx": 96,
    "loss": 0.1503591388463974,
    "learning_rate": 2.9029849645604726e-05,
    "timestep_mean": 499.0,
    "epsilon_norm": 7.742587566375732,
    "prediction_norm": 7.267272472381592,
    "timestamp": "2025-07-14T12:34:37.732750"
  },
  {
    "epoch": 1782,
    "batch_idx": 128,
    "loss": 0.4622405469417572,
    "learning_rate": 2.9029849645604726e-05,
    "timestep_mean": 433.375,
    "epsilon_norm": 9.316573143005371,
    "prediction_norm": 7.022581577301025,
    "timestamp": "2025-07-14T12:34:37.737786"
  },
  {
    "epoch": 1782,
    "batch_idx": 160,
    "loss": 0.37304461002349854,
    "learning_rate": 2.9029849645604726e-05,
    "timestep_mean": 481.3125,
    "epsilon_norm": 9.019501686096191,
    "prediction_norm": 7.005948543548584,
    "timestamp": "2025-07-14T12:34:37.742716"
  },
  {
    "epoch": 1782,
    "batch_idx": 192,
    "loss": 0.2827185094356537,
    "learning_rate": 2.9029849645604726e-05,
    "timestep_mean": 491.875,
    "epsilon_norm": 3.7486412525177,
    "prediction_norm": 3.2819180488586426,
    "timestamp": "2025-07-14T12:34:37.747754"
  },
  {
    "epoch": 1783,
    "batch_idx": 0,
    "loss": 0.19366928935050964,
    "learning_rate": 2.8766709488936052e-05,
    "timestep_mean": 570.78125,
    "epsilon_norm": 7.052483081817627,
    "prediction_norm": 6.747621059417725,
    "timestamp": "2025-07-14T12:34:37.752636"
  },
  {
    "epoch": 1783,
    "batch_idx": 32,
    "loss": 0.24127310514450073,
    "learning_rate": 2.8766709488936052e-05,
    "timestep_mean": 484.75,
    "epsilon_norm": 7.6184234619140625,
    "prediction_norm": 6.258087635040283,
    "timestamp": "2025-07-14T12:34:37.757823"
  },
  {
    "epoch": 1783,
    "batch_idx": 64,
    "loss": 0.18260912597179413,
    "learning_rate": 2.8766709488936052e-05,
    "timestep_mean": 567.625,
    "epsilon_norm": 8.120909690856934,
    "prediction_norm": 7.116147518157959,
    "timestamp": "2025-07-14T12:34:37.762726"
  },
  {
    "epoch": 1783,
    "batch_idx": 96,
    "loss": 0.3147313594818115,
    "learning_rate": 2.8766709488936052e-05,
    "timestep_mean": 477.21875,
    "epsilon_norm": 7.928868770599365,
    "prediction_norm": 6.592639923095703,
    "timestamp": "2025-07-14T12:34:37.767681"
  },
  {
    "epoch": 1783,
    "batch_idx": 128,
    "loss": 0.23364652693271637,
    "learning_rate": 2.8766709488936052e-05,
    "timestep_mean": 476.375,
    "epsilon_norm": 7.111799716949463,
    "prediction_norm": 6.2963762283325195,
    "timestamp": "2025-07-14T12:34:37.772599"
  },
  {
    "epoch": 1783,
    "batch_idx": 160,
    "loss": 0.32115352153778076,
    "learning_rate": 2.8766709488936052e-05,
    "timestep_mean": 470.3125,
    "epsilon_norm": 7.721158504486084,
    "prediction_norm": 6.230206489562988,
    "timestamp": "2025-07-14T12:34:37.777573"
  },
  {
    "epoch": 1783,
    "batch_idx": 192,
    "loss": 0.10906572639942169,
    "learning_rate": 2.8766709488936052e-05,
    "timestep_mean": 574.125,
    "epsilon_norm": 3.0827720165252686,
    "prediction_norm": 3.184169292449951,
    "timestamp": "2025-07-14T12:34:37.782518"
  },
  {
    "epoch": 1784,
    "batch_idx": 0,
    "loss": 0.2362011969089508,
    "learning_rate": 2.850473205356773e-05,
    "timestep_mean": 447.8125,
    "epsilon_norm": 7.377941131591797,
    "prediction_norm": 6.521335601806641,
    "timestamp": "2025-07-14T12:34:37.787452"
  },
  {
    "epoch": 1784,
    "batch_idx": 32,
    "loss": 0.24186265468597412,
    "learning_rate": 2.850473205356773e-05,
    "timestep_mean": 422.96875,
    "epsilon_norm": 8.044903755187988,
    "prediction_norm": 7.116392135620117,
    "timestamp": "2025-07-14T12:34:37.792412"
  },
  {
    "epoch": 1784,
    "batch_idx": 64,
    "loss": 0.228313609957695,
    "learning_rate": 2.850473205356773e-05,
    "timestep_mean": 465.15625,
    "epsilon_norm": 8.068323135375977,
    "prediction_norm": 7.216361999511719,
    "timestamp": "2025-07-14T12:34:37.797358"
  },
  {
    "epoch": 1784,
    "batch_idx": 96,
    "loss": 0.3087195158004761,
    "learning_rate": 2.850473205356773e-05,
    "timestep_mean": 480.0,
    "epsilon_norm": 7.618701457977295,
    "prediction_norm": 6.613099575042725,
    "timestamp": "2025-07-14T12:34:37.802348"
  },
  {
    "epoch": 1784,
    "batch_idx": 128,
    "loss": 0.2563760280609131,
    "learning_rate": 2.850473205356773e-05,
    "timestep_mean": 494.6875,
    "epsilon_norm": 6.7876057624816895,
    "prediction_norm": 5.973718166351318,
    "timestamp": "2025-07-14T12:34:37.807266"
  },
  {
    "epoch": 1784,
    "batch_idx": 160,
    "loss": 0.2960614562034607,
    "learning_rate": 2.850473205356773e-05,
    "timestep_mean": 411.6875,
    "epsilon_norm": 7.651739120483398,
    "prediction_norm": 5.930181980133057,
    "timestamp": "2025-07-14T12:34:37.812206"
  },
  {
    "epoch": 1784,
    "batch_idx": 192,
    "loss": 0.2807508409023285,
    "learning_rate": 2.850473205356773e-05,
    "timestep_mean": 316.625,
    "epsilon_norm": 3.2079219818115234,
    "prediction_norm": 2.082953929901123,
    "timestamp": "2025-07-14T12:34:37.817237"
  },
  {
    "epoch": 1785,
    "batch_idx": 0,
    "loss": 0.41351568698883057,
    "learning_rate": 2.8243917985903195e-05,
    "timestep_mean": 565.6875,
    "epsilon_norm": 8.418543815612793,
    "prediction_norm": 6.622955799102783,
    "timestamp": "2025-07-14T12:34:37.822142"
  },
  {
    "epoch": 1785,
    "batch_idx": 32,
    "loss": 0.2899232506752014,
    "learning_rate": 2.8243917985903195e-05,
    "timestep_mean": 498.78125,
    "epsilon_norm": 7.775018215179443,
    "prediction_norm": 6.374151706695557,
    "timestamp": "2025-07-14T12:34:37.827131"
  },
  {
    "epoch": 1785,
    "batch_idx": 64,
    "loss": 0.1086149513721466,
    "learning_rate": 2.8243917985903195e-05,
    "timestep_mean": 536.1875,
    "epsilon_norm": 6.40004301071167,
    "prediction_norm": 5.4596052169799805,
    "timestamp": "2025-07-14T12:34:37.832083"
  },
  {
    "epoch": 1785,
    "batch_idx": 96,
    "loss": 0.4059429168701172,
    "learning_rate": 2.8243917985903195e-05,
    "timestep_mean": 458.90625,
    "epsilon_norm": 8.672172546386719,
    "prediction_norm": 6.428191661834717,
    "timestamp": "2025-07-14T12:34:37.837075"
  },
  {
    "epoch": 1785,
    "batch_idx": 128,
    "loss": 0.3716470003128052,
    "learning_rate": 2.8243917985903195e-05,
    "timestep_mean": 456.90625,
    "epsilon_norm": 7.320944786071777,
    "prediction_norm": 5.998327255249023,
    "timestamp": "2025-07-14T12:34:37.842008"
  },
  {
    "epoch": 1785,
    "batch_idx": 160,
    "loss": 0.20874366164207458,
    "learning_rate": 2.8243917985903195e-05,
    "timestep_mean": 530.1875,
    "epsilon_norm": 7.313018798828125,
    "prediction_norm": 6.7363972663879395,
    "timestamp": "2025-07-14T12:34:37.846929"
  },
  {
    "epoch": 1785,
    "batch_idx": 192,
    "loss": 0.09423409402370453,
    "learning_rate": 2.8243917985903195e-05,
    "timestep_mean": 698.0,
    "epsilon_norm": 2.5341296195983887,
    "prediction_norm": 2.0562052726745605,
    "timestamp": "2025-07-14T12:34:37.851908"
  },
  {
    "epoch": 1786,
    "batch_idx": 0,
    "loss": 0.5125741958618164,
    "learning_rate": 2.7984267929475168e-05,
    "timestep_mean": 420.9375,
    "epsilon_norm": 8.306767463684082,
    "prediction_norm": 5.943068504333496,
    "timestamp": "2025-07-14T12:34:37.857026"
  },
  {
    "epoch": 1786,
    "batch_idx": 32,
    "loss": 0.3015793263912201,
    "learning_rate": 2.7984267929475168e-05,
    "timestep_mean": 471.71875,
    "epsilon_norm": 8.33176040649414,
    "prediction_norm": 7.220782279968262,
    "timestamp": "2025-07-14T12:34:37.861950"
  },
  {
    "epoch": 1786,
    "batch_idx": 64,
    "loss": 0.33591020107269287,
    "learning_rate": 2.7984267929475168e-05,
    "timestep_mean": 454.125,
    "epsilon_norm": 6.491716384887695,
    "prediction_norm": 4.878607749938965,
    "timestamp": "2025-07-14T12:34:37.867001"
  },
  {
    "epoch": 1786,
    "batch_idx": 96,
    "loss": 0.25429651141166687,
    "learning_rate": 2.7984267929475168e-05,
    "timestep_mean": 478.125,
    "epsilon_norm": 8.967146873474121,
    "prediction_norm": 7.4280171394348145,
    "timestamp": "2025-07-14T12:34:37.871902"
  },
  {
    "epoch": 1786,
    "batch_idx": 128,
    "loss": 0.2831433415412903,
    "learning_rate": 2.7984267929475168e-05,
    "timestep_mean": 517.875,
    "epsilon_norm": 8.487059593200684,
    "prediction_norm": 7.43887996673584,
    "timestamp": "2025-07-14T12:34:37.876850"
  },
  {
    "epoch": 1786,
    "batch_idx": 160,
    "loss": 0.2460610270500183,
    "learning_rate": 2.7984267929475168e-05,
    "timestep_mean": 520.5,
    "epsilon_norm": 8.67137336730957,
    "prediction_norm": 7.525548934936523,
    "timestamp": "2025-07-14T12:34:37.881787"
  },
  {
    "epoch": 1786,
    "batch_idx": 192,
    "loss": 0.360870361328125,
    "learning_rate": 2.7984267929475168e-05,
    "timestep_mean": 577.5,
    "epsilon_norm": 4.1020426750183105,
    "prediction_norm": 3.2446439266204834,
    "timestamp": "2025-07-14T12:34:37.886874"
  },
  {
    "epoch": 1787,
    "batch_idx": 0,
    "loss": 0.30450522899627686,
    "learning_rate": 2.7725782524944284e-05,
    "timestep_mean": 489.0625,
    "epsilon_norm": 8.380627632141113,
    "prediction_norm": 7.080058574676514,
    "timestamp": "2025-07-14T12:34:37.891985"
  },
  {
    "epoch": 1787,
    "batch_idx": 32,
    "loss": 0.16413474082946777,
    "learning_rate": 2.7725782524944284e-05,
    "timestep_mean": 532.4375,
    "epsilon_norm": 8.525716781616211,
    "prediction_norm": 7.610189437866211,
    "timestamp": "2025-07-14T12:34:37.896946"
  },
  {
    "epoch": 1787,
    "batch_idx": 64,
    "loss": 0.22065575420856476,
    "learning_rate": 2.7725782524944284e-05,
    "timestep_mean": 473.09375,
    "epsilon_norm": 7.839869976043701,
    "prediction_norm": 7.108238697052002,
    "timestamp": "2025-07-14T12:34:37.901920"
  },
  {
    "epoch": 1787,
    "batch_idx": 96,
    "loss": 0.3919432759284973,
    "learning_rate": 2.7725782524944284e-05,
    "timestep_mean": 466.25,
    "epsilon_norm": 8.962137222290039,
    "prediction_norm": 6.8057122230529785,
    "timestamp": "2025-07-14T12:34:37.906853"
  },
  {
    "epoch": 1787,
    "batch_idx": 128,
    "loss": 0.11299224197864532,
    "learning_rate": 2.7725782524944284e-05,
    "timestep_mean": 557.40625,
    "epsilon_norm": 6.784770488739014,
    "prediction_norm": 6.369997501373291,
    "timestamp": "2025-07-14T12:34:37.911844"
  },
  {
    "epoch": 1787,
    "batch_idx": 160,
    "loss": 0.5236650109291077,
    "learning_rate": 2.7725782524944284e-05,
    "timestep_mean": 427.34375,
    "epsilon_norm": 9.041790008544922,
    "prediction_norm": 6.6872663497924805,
    "timestamp": "2025-07-14T12:34:37.916936"
  },
  {
    "epoch": 1787,
    "batch_idx": 192,
    "loss": 0.14262855052947998,
    "learning_rate": 2.7725782524944284e-05,
    "timestep_mean": 512.875,
    "epsilon_norm": 4.201486587524414,
    "prediction_norm": 4.033478736877441,
    "timestamp": "2025-07-14T12:34:37.921912"
  },
  {
    "epoch": 1788,
    "batch_idx": 0,
    "loss": 0.22765682637691498,
    "learning_rate": 2.7468462410097645e-05,
    "timestep_mean": 504.84375,
    "epsilon_norm": 8.073187828063965,
    "prediction_norm": 7.079939842224121,
    "timestamp": "2025-07-14T12:34:37.926922"
  },
  {
    "epoch": 1788,
    "batch_idx": 32,
    "loss": 0.14634785056114197,
    "learning_rate": 2.7468462410097645e-05,
    "timestep_mean": 538.03125,
    "epsilon_norm": 7.4757304191589355,
    "prediction_norm": 7.193458080291748,
    "timestamp": "2025-07-14T12:34:37.931905"
  },
  {
    "epoch": 1788,
    "batch_idx": 64,
    "loss": 0.4168110191822052,
    "learning_rate": 2.7468462410097645e-05,
    "timestep_mean": 487.40625,
    "epsilon_norm": 8.198725700378418,
    "prediction_norm": 6.114658355712891,
    "timestamp": "2025-07-14T12:34:37.936847"
  },
  {
    "epoch": 1788,
    "batch_idx": 96,
    "loss": 0.45631274580955505,
    "learning_rate": 2.7468462410097645e-05,
    "timestep_mean": 454.40625,
    "epsilon_norm": 8.535301208496094,
    "prediction_norm": 6.5010151863098145,
    "timestamp": "2025-07-14T12:34:37.941953"
  },
  {
    "epoch": 1788,
    "batch_idx": 128,
    "loss": 0.24922913312911987,
    "learning_rate": 2.7468462410097645e-05,
    "timestep_mean": 503.6875,
    "epsilon_norm": 7.841059684753418,
    "prediction_norm": 6.8895721435546875,
    "timestamp": "2025-07-14T12:34:37.946851"
  },
  {
    "epoch": 1788,
    "batch_idx": 160,
    "loss": 0.31545206904411316,
    "learning_rate": 2.7468462410097645e-05,
    "timestep_mean": 478.3125,
    "epsilon_norm": 7.563092231750488,
    "prediction_norm": 6.977950572967529,
    "timestamp": "2025-07-14T12:34:37.951879"
  },
  {
    "epoch": 1788,
    "batch_idx": 192,
    "loss": 0.17438024282455444,
    "learning_rate": 2.7468462410097645e-05,
    "timestep_mean": 606.875,
    "epsilon_norm": 4.067350387573242,
    "prediction_norm": 3.519392490386963,
    "timestamp": "2025-07-14T12:34:37.956936"
  },
  {
    "epoch": 1789,
    "batch_idx": 0,
    "loss": 0.14359626173973083,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 502.1875,
    "epsilon_norm": 8.168169021606445,
    "prediction_norm": 7.9333319664001465,
    "timestamp": "2025-07-14T12:34:37.961990"
  },
  {
    "epoch": 1789,
    "batch_idx": 32,
    "loss": 0.2376033514738083,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 517.65625,
    "epsilon_norm": 8.333049774169922,
    "prediction_norm": 7.1889190673828125,
    "timestamp": "2025-07-14T12:34:37.967393"
  },
  {
    "epoch": 1789,
    "batch_idx": 64,
    "loss": 0.3079829216003418,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 386.125,
    "epsilon_norm": 7.590925216674805,
    "prediction_norm": 6.39229679107666,
    "timestamp": "2025-07-14T12:34:37.973250"
  },
  {
    "epoch": 1789,
    "batch_idx": 96,
    "loss": 0.4101242125034332,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 554.4375,
    "epsilon_norm": 7.93937873840332,
    "prediction_norm": 6.567712783813477,
    "timestamp": "2025-07-14T12:34:37.988369"
  },
  {
    "epoch": 1789,
    "batch_idx": 128,
    "loss": 0.2622341513633728,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 480.0,
    "epsilon_norm": 7.849256992340088,
    "prediction_norm": 6.785423755645752,
    "timestamp": "2025-07-14T12:34:37.997529"
  },
  {
    "epoch": 1789,
    "batch_idx": 160,
    "loss": 0.18677619099617004,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 462.75,
    "epsilon_norm": 7.826496124267578,
    "prediction_norm": 6.7723212242126465,
    "timestamp": "2025-07-14T12:34:38.005734"
  },
  {
    "epoch": 1789,
    "batch_idx": 192,
    "loss": 0.12264799326658249,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 628.75,
    "epsilon_norm": 3.5861716270446777,
    "prediction_norm": 3.0739712715148926,
    "timestamp": "2025-07-14T12:34:38.014172"
  },
  {
    "epoch": 1790,
    "batch_idx": 0,
    "loss": 0.2153235226869583,
    "learning_rate": 2.6957320586227344e-05,
    "timestep_mean": 479.28125,
    "epsilon_norm": 7.583052158355713,
    "prediction_norm": 6.331881046295166,
    "timestamp": "2025-07-14T12:34:38.020097"
  },
  {
    "epoch": 1790,
    "batch_idx": 32,
    "loss": 0.2846941649913788,
    "learning_rate": 2.6957320586227344e-05,
    "timestep_mean": 487.25,
    "epsilon_norm": 8.444188117980957,
    "prediction_norm": 7.5202765464782715,
    "timestamp": "2025-07-14T12:34:38.026491"
  },
  {
    "epoch": 1790,
    "batch_idx": 64,
    "loss": 0.14411327242851257,
    "learning_rate": 2.6957320586227344e-05,
    "timestep_mean": 573.375,
    "epsilon_norm": 7.376040458679199,
    "prediction_norm": 6.75659704208374,
    "timestamp": "2025-07-14T12:34:38.034752"
  },
  {
    "epoch": 1790,
    "batch_idx": 96,
    "loss": 0.28371676802635193,
    "learning_rate": 2.6957320586227344e-05,
    "timestep_mean": 580.8125,
    "epsilon_norm": 8.456938743591309,
    "prediction_norm": 6.980163097381592,
    "timestamp": "2025-07-14T12:34:38.043477"
  },
  {
    "epoch": 1790,
    "batch_idx": 128,
    "loss": 0.41672617197036743,
    "learning_rate": 2.6957320586227344e-05,
    "timestep_mean": 519.875,
    "epsilon_norm": 8.216862678527832,
    "prediction_norm": 7.287907123565674,
    "timestamp": "2025-07-14T12:34:38.051791"
  },
  {
    "epoch": 1790,
    "batch_idx": 160,
    "loss": 0.3277706503868103,
    "learning_rate": 2.6957320586227344e-05,
    "timestep_mean": 442.0625,
    "epsilon_norm": 7.870982646942139,
    "prediction_norm": 6.226482391357422,
    "timestamp": "2025-07-14T12:34:38.059581"
  },
  {
    "epoch": 1790,
    "batch_idx": 192,
    "loss": 0.4196816384792328,
    "learning_rate": 2.6957320586227344e-05,
    "timestep_mean": 401.5,
    "epsilon_norm": 3.551936626434326,
    "prediction_norm": 1.9179933071136475,
    "timestamp": "2025-07-14T12:34:38.067409"
  },
  {
    "epoch": 1791,
    "batch_idx": 0,
    "loss": 0.4159858524799347,
    "learning_rate": 2.670350013839539e-05,
    "timestep_mean": 427.75,
    "epsilon_norm": 8.16834831237793,
    "prediction_norm": 6.009946823120117,
    "timestamp": "2025-07-14T12:34:38.075746"
  },
  {
    "epoch": 1791,
    "batch_idx": 32,
    "loss": 0.2363642007112503,
    "learning_rate": 2.670350013839539e-05,
    "timestep_mean": 504.9375,
    "epsilon_norm": 8.054388046264648,
    "prediction_norm": 6.988346099853516,
    "timestamp": "2025-07-14T12:34:38.082211"
  },
  {
    "epoch": 1791,
    "batch_idx": 64,
    "loss": 0.23081645369529724,
    "learning_rate": 2.670350013839539e-05,
    "timestep_mean": 476.15625,
    "epsilon_norm": 7.513378620147705,
    "prediction_norm": 5.872309684753418,
    "timestamp": "2025-07-14T12:34:38.087343"
  },
  {
    "epoch": 1791,
    "batch_idx": 96,
    "loss": 0.24665632843971252,
    "learning_rate": 2.670350013839539e-05,
    "timestep_mean": 395.0625,
    "epsilon_norm": 8.022485733032227,
    "prediction_norm": 6.880151748657227,
    "timestamp": "2025-07-14T12:34:38.092464"
  },
  {
    "epoch": 1791,
    "batch_idx": 128,
    "loss": 0.18251198530197144,
    "learning_rate": 2.670350013839539e-05,
    "timestep_mean": 493.5,
    "epsilon_norm": 7.51914119720459,
    "prediction_norm": 6.707268714904785,
    "timestamp": "2025-07-14T12:34:38.097268"
  },
  {
    "epoch": 1791,
    "batch_idx": 160,
    "loss": 0.31890735030174255,
    "learning_rate": 2.670350013839539e-05,
    "timestep_mean": 503.6875,
    "epsilon_norm": 7.687934875488281,
    "prediction_norm": 6.653094291687012,
    "timestamp": "2025-07-14T12:34:38.102017"
  },
  {
    "epoch": 1791,
    "batch_idx": 192,
    "loss": 0.2034299373626709,
    "learning_rate": 2.670350013839539e-05,
    "timestep_mean": 493.25,
    "epsilon_norm": 4.363493919372559,
    "prediction_norm": 3.513362407684326,
    "timestamp": "2025-07-14T12:34:38.106949"
  },
  {
    "epoch": 1792,
    "batch_idx": 0,
    "loss": 0.18407265841960907,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 521.46875,
    "epsilon_norm": 7.901496410369873,
    "prediction_norm": 7.333026885986328,
    "timestamp": "2025-07-14T12:34:38.111809"
  },
  {
    "epoch": 1792,
    "batch_idx": 32,
    "loss": 0.4167410731315613,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 391.6875,
    "epsilon_norm": 8.813910484313965,
    "prediction_norm": 6.864697456359863,
    "timestamp": "2025-07-14T12:34:38.116588"
  },
  {
    "epoch": 1792,
    "batch_idx": 64,
    "loss": 0.24457213282585144,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 563.75,
    "epsilon_norm": 7.41188383102417,
    "prediction_norm": 6.161371231079102,
    "timestamp": "2025-07-14T12:34:38.121479"
  },
  {
    "epoch": 1792,
    "batch_idx": 96,
    "loss": 0.19938114285469055,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 509.71875,
    "epsilon_norm": 8.336095809936523,
    "prediction_norm": 7.459589958190918,
    "timestamp": "2025-07-14T12:34:38.126252"
  },
  {
    "epoch": 1792,
    "batch_idx": 128,
    "loss": 0.3470751643180847,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 510.375,
    "epsilon_norm": 8.542451858520508,
    "prediction_norm": 7.48238468170166,
    "timestamp": "2025-07-14T12:34:38.131016"
  },
  {
    "epoch": 1792,
    "batch_idx": 160,
    "loss": 0.4167836606502533,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 463.625,
    "epsilon_norm": 7.903693199157715,
    "prediction_norm": 6.029970169067383,
    "timestamp": "2025-07-14T12:34:38.135827"
  },
  {
    "epoch": 1792,
    "batch_idx": 192,
    "loss": 0.03306538984179497,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 656.25,
    "epsilon_norm": 3.46500825881958,
    "prediction_norm": 3.146545886993408,
    "timestamp": "2025-07-14T12:34:38.140623"
  },
  {
    "epoch": 1793,
    "batch_idx": 0,
    "loss": 0.41192758083343506,
    "learning_rate": 2.619936330231992e-05,
    "timestep_mean": 443.5625,
    "epsilon_norm": 7.806797981262207,
    "prediction_norm": 6.308578014373779,
    "timestamp": "2025-07-14T12:34:38.145418"
  },
  {
    "epoch": 1793,
    "batch_idx": 32,
    "loss": 0.27806705236434937,
    "learning_rate": 2.619936330231992e-05,
    "timestep_mean": 467.71875,
    "epsilon_norm": 8.535820007324219,
    "prediction_norm": 6.999633312225342,
    "timestamp": "2025-07-14T12:34:38.150293"
  },
  {
    "epoch": 1793,
    "batch_idx": 64,
    "loss": 0.14120444655418396,
    "learning_rate": 2.619936330231992e-05,
    "timestep_mean": 588.78125,
    "epsilon_norm": 7.366476058959961,
    "prediction_norm": 6.9840288162231445,
    "timestamp": "2025-07-14T12:34:38.155004"
  },
  {
    "epoch": 1793,
    "batch_idx": 96,
    "loss": 0.38745975494384766,
    "learning_rate": 2.619936330231992e-05,
    "timestep_mean": 524.25,
    "epsilon_norm": 7.760720729827881,
    "prediction_norm": 5.987327575683594,
    "timestamp": "2025-07-14T12:34:38.159809"
  },
  {
    "epoch": 1793,
    "batch_idx": 128,
    "loss": 0.2429710477590561,
    "learning_rate": 2.619936330231992e-05,
    "timestep_mean": 546.75,
    "epsilon_norm": 7.788259506225586,
    "prediction_norm": 6.30734395980835,
    "timestamp": "2025-07-14T12:34:38.164527"
  },
  {
    "epoch": 1793,
    "batch_idx": 160,
    "loss": 0.3054138720035553,
    "learning_rate": 2.619936330231992e-05,
    "timestep_mean": 482.53125,
    "epsilon_norm": 8.016709327697754,
    "prediction_norm": 6.651575088500977,
    "timestamp": "2025-07-14T12:34:38.169367"
  },
  {
    "epoch": 1793,
    "batch_idx": 192,
    "loss": 0.0332280732691288,
    "learning_rate": 2.619936330231992e-05,
    "timestep_mean": 776.375,
    "epsilon_norm": 4.072497367858887,
    "prediction_norm": 3.9760403633117676,
    "timestamp": "2025-07-14T12:34:38.174232"
  },
  {
    "epoch": 1794,
    "batch_idx": 0,
    "loss": 0.4251956045627594,
    "learning_rate": 2.594904815798404e-05,
    "timestep_mean": 442.71875,
    "epsilon_norm": 8.402664184570312,
    "prediction_norm": 6.192327499389648,
    "timestamp": "2025-07-14T12:34:38.179074"
  },
  {
    "epoch": 1794,
    "batch_idx": 32,
    "loss": 0.15293331444263458,
    "learning_rate": 2.594904815798404e-05,
    "timestep_mean": 497.65625,
    "epsilon_norm": 7.085779190063477,
    "prediction_norm": 6.546080112457275,
    "timestamp": "2025-07-14T12:34:38.183972"
  },
  {
    "epoch": 1794,
    "batch_idx": 64,
    "loss": 0.1937597393989563,
    "learning_rate": 2.594904815798404e-05,
    "timestep_mean": 520.1875,
    "epsilon_norm": 6.62313175201416,
    "prediction_norm": 5.928366184234619,
    "timestamp": "2025-07-14T12:34:38.188800"
  },
  {
    "epoch": 1794,
    "batch_idx": 96,
    "loss": 0.08707365393638611,
    "learning_rate": 2.594904815798404e-05,
    "timestep_mean": 542.15625,
    "epsilon_norm": 7.356138229370117,
    "prediction_norm": 7.342288017272949,
    "timestamp": "2025-07-14T12:34:38.193540"
  },
  {
    "epoch": 1794,
    "batch_idx": 128,
    "loss": 0.33935636281967163,
    "learning_rate": 2.594904815798404e-05,
    "timestep_mean": 474.90625,
    "epsilon_norm": 7.7849202156066895,
    "prediction_norm": 6.366122245788574,
    "timestamp": "2025-07-14T12:34:38.198267"
  },
  {
    "epoch": 1794,
    "batch_idx": 160,
    "loss": 0.3359145522117615,
    "learning_rate": 2.594904815798404e-05,
    "timestep_mean": 377.65625,
    "epsilon_norm": 8.808064460754395,
    "prediction_norm": 6.947775840759277,
    "timestamp": "2025-07-14T12:34:38.203016"
  },
  {
    "epoch": 1794,
    "batch_idx": 192,
    "loss": 0.12275728583335876,
    "learning_rate": 2.594904815798404e-05,
    "timestep_mean": 600.5,
    "epsilon_norm": 3.859797716140747,
    "prediction_norm": 3.4064955711364746,
    "timestamp": "2025-07-14T12:34:38.207750"
  },
  {
    "epoch": 1795,
    "batch_idx": 0,
    "loss": 0.2850363850593567,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 553.96875,
    "epsilon_norm": 7.270008563995361,
    "prediction_norm": 6.2687859535217285,
    "timestamp": "2025-07-14T12:34:38.212511"
  },
  {
    "epoch": 1795,
    "batch_idx": 32,
    "loss": 0.21187351644039154,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 487.53125,
    "epsilon_norm": 8.900275230407715,
    "prediction_norm": 7.58427095413208,
    "timestamp": "2025-07-14T12:34:38.217458"
  },
  {
    "epoch": 1795,
    "batch_idx": 64,
    "loss": 0.19333337247371674,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 497.125,
    "epsilon_norm": 7.690526962280273,
    "prediction_norm": 7.034600257873535,
    "timestamp": "2025-07-14T12:34:38.222255"
  },
  {
    "epoch": 1795,
    "batch_idx": 96,
    "loss": 0.2024618685245514,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 517.0,
    "epsilon_norm": 7.011618137359619,
    "prediction_norm": 6.0229082107543945,
    "timestamp": "2025-07-14T12:34:38.227147"
  },
  {
    "epoch": 1795,
    "batch_idx": 128,
    "loss": 0.2373761087656021,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 511.53125,
    "epsilon_norm": 8.202095031738281,
    "prediction_norm": 6.895940780639648,
    "timestamp": "2025-07-14T12:34:38.231976"
  },
  {
    "epoch": 1795,
    "batch_idx": 160,
    "loss": 0.48946595191955566,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 435.09375,
    "epsilon_norm": 8.471062660217285,
    "prediction_norm": 6.352138996124268,
    "timestamp": "2025-07-14T12:34:38.237020"
  },
  {
    "epoch": 1795,
    "batch_idx": 192,
    "loss": 0.7138685584068298,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 511.5,
    "epsilon_norm": 5.0589141845703125,
    "prediction_norm": 3.062623977661133,
    "timestamp": "2025-07-14T12:34:38.242323"
  },
  {
    "epoch": 1796,
    "batch_idx": 0,
    "loss": 0.25024744868278503,
    "learning_rate": 2.545192750485264e-05,
    "timestep_mean": 449.96875,
    "epsilon_norm": 8.211792945861816,
    "prediction_norm": 7.039214611053467,
    "timestamp": "2025-07-14T12:34:38.247268"
  },
  {
    "epoch": 1796,
    "batch_idx": 32,
    "loss": 0.34913650155067444,
    "learning_rate": 2.545192750485264e-05,
    "timestep_mean": 464.90625,
    "epsilon_norm": 8.807056427001953,
    "prediction_norm": 6.820852279663086,
    "timestamp": "2025-07-14T12:34:38.252097"
  },
  {
    "epoch": 1796,
    "batch_idx": 64,
    "loss": 0.17694330215454102,
    "learning_rate": 2.545192750485264e-05,
    "timestep_mean": 503.53125,
    "epsilon_norm": 7.2569780349731445,
    "prediction_norm": 6.583176136016846,
    "timestamp": "2025-07-14T12:34:38.257125"
  },
  {
    "epoch": 1796,
    "batch_idx": 96,
    "loss": 0.29393309354782104,
    "learning_rate": 2.545192750485264e-05,
    "timestep_mean": 528.1875,
    "epsilon_norm": 7.981931209564209,
    "prediction_norm": 6.610421180725098,
    "timestamp": "2025-07-14T12:34:38.262489"
  },
  {
    "epoch": 1796,
    "batch_idx": 128,
    "loss": 0.15992599725723267,
    "learning_rate": 2.545192750485264e-05,
    "timestep_mean": 523.28125,
    "epsilon_norm": 6.689155578613281,
    "prediction_norm": 5.74694299697876,
    "timestamp": "2025-07-14T12:34:38.267733"
  },
  {
    "epoch": 1796,
    "batch_idx": 160,
    "loss": 0.31951165199279785,
    "learning_rate": 2.545192750485264e-05,
    "timestep_mean": 444.375,
    "epsilon_norm": 7.450154781341553,
    "prediction_norm": 6.0656352043151855,
    "timestamp": "2025-07-14T12:34:38.272744"
  },
  {
    "epoch": 1796,
    "batch_idx": 192,
    "loss": 0.4144093692302704,
    "learning_rate": 2.545192750485264e-05,
    "timestep_mean": 306.25,
    "epsilon_norm": 3.7517993450164795,
    "prediction_norm": 3.147035598754883,
    "timestamp": "2025-07-14T12:34:38.277809"
  },
  {
    "epoch": 1797,
    "batch_idx": 0,
    "loss": 0.2534739375114441,
    "learning_rate": 2.520512322265322e-05,
    "timestep_mean": 489.90625,
    "epsilon_norm": 8.304337501525879,
    "prediction_norm": 6.983057022094727,
    "timestamp": "2025-07-14T12:34:38.282947"
  },
  {
    "epoch": 1797,
    "batch_idx": 32,
    "loss": 0.16525667905807495,
    "learning_rate": 2.520512322265322e-05,
    "timestep_mean": 526.34375,
    "epsilon_norm": 8.4678373336792,
    "prediction_norm": 7.568310737609863,
    "timestamp": "2025-07-14T12:34:38.288044"
  },
  {
    "epoch": 1797,
    "batch_idx": 64,
    "loss": 0.2918301522731781,
    "learning_rate": 2.520512322265322e-05,
    "timestep_mean": 458.46875,
    "epsilon_norm": 8.052959442138672,
    "prediction_norm": 7.160463333129883,
    "timestamp": "2025-07-14T12:34:38.293018"
  },
  {
    "epoch": 1797,
    "batch_idx": 96,
    "loss": 0.19505125284194946,
    "learning_rate": 2.520512322265322e-05,
    "timestep_mean": 445.875,
    "epsilon_norm": 6.802730560302734,
    "prediction_norm": 5.5345845222473145,
    "timestamp": "2025-07-14T12:34:38.297930"
  },
  {
    "epoch": 1797,
    "batch_idx": 128,
    "loss": 0.3242833614349365,
    "learning_rate": 2.520512322265322e-05,
    "timestep_mean": 414.65625,
    "epsilon_norm": 6.770722389221191,
    "prediction_norm": 5.450096130371094,
    "timestamp": "2025-07-14T12:34:38.302922"
  },
  {
    "epoch": 1797,
    "batch_idx": 160,
    "loss": 0.32756876945495605,
    "learning_rate": 2.520512322265322e-05,
    "timestep_mean": 385.25,
    "epsilon_norm": 8.316618919372559,
    "prediction_norm": 7.104003429412842,
    "timestamp": "2025-07-14T12:34:38.308041"
  },
  {
    "epoch": 1797,
    "batch_idx": 192,
    "loss": 0.0593777559697628,
    "learning_rate": 2.520512322265322e-05,
    "timestep_mean": 695.25,
    "epsilon_norm": 2.822657585144043,
    "prediction_norm": 2.970097541809082,
    "timestamp": "2025-07-14T12:34:38.313173"
  },
  {
    "epoch": 1798,
    "batch_idx": 0,
    "loss": 0.28481319546699524,
    "learning_rate": 2.4959490449614094e-05,
    "timestep_mean": 524.21875,
    "epsilon_norm": 7.571482181549072,
    "prediction_norm": 6.473414897918701,
    "timestamp": "2025-07-14T12:34:38.319020"
  },
  {
    "epoch": 1798,
    "batch_idx": 32,
    "loss": 0.2561872601509094,
    "learning_rate": 2.4959490449614094e-05,
    "timestep_mean": 525.375,
    "epsilon_norm": 8.474130630493164,
    "prediction_norm": 7.178828716278076,
    "timestamp": "2025-07-14T12:34:38.324359"
  },
  {
    "epoch": 1798,
    "batch_idx": 64,
    "loss": 0.25406762957572937,
    "learning_rate": 2.4959490449614094e-05,
    "timestep_mean": 486.1875,
    "epsilon_norm": 8.531850814819336,
    "prediction_norm": 6.930442810058594,
    "timestamp": "2025-07-14T12:34:38.329290"
  },
  {
    "epoch": 1798,
    "batch_idx": 96,
    "loss": 0.1929844319820404,
    "learning_rate": 2.4959490449614094e-05,
    "timestep_mean": 532.5625,
    "epsilon_norm": 8.253806114196777,
    "prediction_norm": 7.07281494140625,
    "timestamp": "2025-07-14T12:34:38.335071"
  },
  {
    "epoch": 1798,
    "batch_idx": 128,
    "loss": 0.3805311918258667,
    "learning_rate": 2.4959490449614094e-05,
    "timestep_mean": 539.0,
    "epsilon_norm": 9.565351486206055,
    "prediction_norm": 7.897339344024658,
    "timestamp": "2025-07-14T12:34:38.340086"
  },
  {
    "epoch": 1798,
    "batch_idx": 160,
    "loss": 0.12851549685001373,
    "learning_rate": 2.4959490449614094e-05,
    "timestep_mean": 514.84375,
    "epsilon_norm": 6.895992755889893,
    "prediction_norm": 6.822768211364746,
    "timestamp": "2025-07-14T12:34:38.345175"
  },
  {
    "epoch": 1798,
    "batch_idx": 192,
    "loss": 0.1083107441663742,
    "learning_rate": 2.4959490449614094e-05,
    "timestep_mean": 673.625,
    "epsilon_norm": 3.7977991104125977,
    "prediction_norm": 3.5382447242736816,
    "timestamp": "2025-07-14T12:34:38.350501"
  },
  {
    "epoch": 1799,
    "batch_idx": 0,
    "loss": 0.32413366436958313,
    "learning_rate": 2.4715029791810115e-05,
    "timestep_mean": 469.96875,
    "epsilon_norm": 8.364400863647461,
    "prediction_norm": 7.056640148162842,
    "timestamp": "2025-07-14T12:34:38.355608"
  },
  {
    "epoch": 1799,
    "batch_idx": 32,
    "loss": 0.2203233540058136,
    "learning_rate": 2.4715029791810115e-05,
    "timestep_mean": 512.65625,
    "epsilon_norm": 7.372090816497803,
    "prediction_norm": 6.173243999481201,
    "timestamp": "2025-07-14T12:34:38.360716"
  },
  {
    "epoch": 1799,
    "batch_idx": 64,
    "loss": 0.15199054777622223,
    "learning_rate": 2.4715029791810115e-05,
    "timestep_mean": 481.59375,
    "epsilon_norm": 6.127737045288086,
    "prediction_norm": 5.500913143157959,
    "timestamp": "2025-07-14T12:34:38.365788"
  },
  {
    "epoch": 1799,
    "batch_idx": 96,
    "loss": 0.19935806095600128,
    "learning_rate": 2.4715029791810115e-05,
    "timestep_mean": 538.5625,
    "epsilon_norm": 6.535862445831299,
    "prediction_norm": 5.388153076171875,
    "timestamp": "2025-07-14T12:34:38.371002"
  },
  {
    "epoch": 1799,
    "batch_idx": 128,
    "loss": 0.33714038133621216,
    "learning_rate": 2.4715029791810115e-05,
    "timestep_mean": 454.8125,
    "epsilon_norm": 6.994317054748535,
    "prediction_norm": 5.619806289672852,
    "timestamp": "2025-07-14T12:34:38.376258"
  },
  {
    "epoch": 1799,
    "batch_idx": 160,
    "loss": 0.24035637080669403,
    "learning_rate": 2.4715029791810115e-05,
    "timestep_mean": 479.25,
    "epsilon_norm": 7.432061195373535,
    "prediction_norm": 6.521159648895264,
    "timestamp": "2025-07-14T12:34:38.381459"
  },
  {
    "epoch": 1799,
    "batch_idx": 192,
    "loss": 0.18075965344905853,
    "learning_rate": 2.4715029791810115e-05,
    "timestep_mean": 496.375,
    "epsilon_norm": 3.800239324569702,
    "prediction_norm": 3.26814341545105,
    "timestamp": "2025-07-14T12:34:38.386755"
  },
  {
    "epoch": 1800,
    "batch_idx": 0,
    "loss": 0.23152796924114227,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 511.3125,
    "epsilon_norm": 8.599597930908203,
    "prediction_norm": 7.792696475982666,
    "timestamp": "2025-07-14T12:34:38.392080"
  },
  {
    "epoch": 1800,
    "batch_idx": 32,
    "loss": 0.45460033416748047,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 375.40625,
    "epsilon_norm": 7.504889488220215,
    "prediction_norm": 5.092380046844482,
    "timestamp": "2025-07-14T12:34:38.397366"
  },
  {
    "epoch": 1800,
    "batch_idx": 64,
    "loss": 0.1938823014497757,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 495.9375,
    "epsilon_norm": 8.105646133422852,
    "prediction_norm": 6.655834674835205,
    "timestamp": "2025-07-14T12:34:38.402412"
  },
  {
    "epoch": 1800,
    "batch_idx": 96,
    "loss": 0.13962145149707794,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 432.34375,
    "epsilon_norm": 7.321278095245361,
    "prediction_norm": 6.723734378814697,
    "timestamp": "2025-07-14T12:34:38.407498"
  },
  {
    "epoch": 1800,
    "batch_idx": 128,
    "loss": 0.13925784826278687,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 560.4375,
    "epsilon_norm": 8.32018756866455,
    "prediction_norm": 7.463046550750732,
    "timestamp": "2025-07-14T12:34:38.412461"
  },
  {
    "epoch": 1800,
    "batch_idx": 160,
    "loss": 0.4999490976333618,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 425.875,
    "epsilon_norm": 9.082937240600586,
    "prediction_norm": 8.025105476379395,
    "timestamp": "2025-07-14T12:34:38.417838"
  },
  {
    "epoch": 1800,
    "batch_idx": 192,
    "loss": 0.21934087574481964,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 496.25,
    "epsilon_norm": 3.385092258453369,
    "prediction_norm": 3.082660675048828,
    "timestamp": "2025-07-14T12:34:38.423031"
  },
  {
    "epoch": 1801,
    "batch_idx": 0,
    "loss": 0.39557528495788574,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 517.78125,
    "epsilon_norm": 7.538271903991699,
    "prediction_norm": 5.590206623077393,
    "timestamp": "2025-07-14T12:34:38.428394"
  },
  {
    "epoch": 1801,
    "batch_idx": 32,
    "loss": 0.2225649207830429,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 459.125,
    "epsilon_norm": 8.198880195617676,
    "prediction_norm": 6.611962795257568,
    "timestamp": "2025-07-14T12:34:38.433663"
  },
  {
    "epoch": 1801,
    "batch_idx": 64,
    "loss": 0.28072720766067505,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 512.84375,
    "epsilon_norm": 7.8245697021484375,
    "prediction_norm": 6.580019950866699,
    "timestamp": "2025-07-14T12:34:38.438888"
  },
  {
    "epoch": 1801,
    "batch_idx": 96,
    "loss": 0.23729673027992249,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 464.46875,
    "epsilon_norm": 6.938814163208008,
    "prediction_norm": 6.213169574737549,
    "timestamp": "2025-07-14T12:34:38.444352"
  },
  {
    "epoch": 1801,
    "batch_idx": 128,
    "loss": 0.12083688378334045,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 478.4375,
    "epsilon_norm": 7.126679420471191,
    "prediction_norm": 6.140660285949707,
    "timestamp": "2025-07-14T12:34:38.449518"
  },
  {
    "epoch": 1801,
    "batch_idx": 160,
    "loss": 0.5822588801383972,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 415.1875,
    "epsilon_norm": 8.09443187713623,
    "prediction_norm": 5.53695011138916,
    "timestamp": "2025-07-14T12:34:38.454803"
  },
  {
    "epoch": 1801,
    "batch_idx": 192,
    "loss": 0.11281630396842957,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 510.25,
    "epsilon_norm": 4.276076793670654,
    "prediction_norm": 3.9719557762145996,
    "timestamp": "2025-07-14T12:34:38.460025"
  },
  {
    "epoch": 1802,
    "batch_idx": 0,
    "loss": 0.23191872239112854,
    "learning_rate": 2.3988686527161737e-05,
    "timestep_mean": 512.09375,
    "epsilon_norm": 7.795871734619141,
    "prediction_norm": 6.950641632080078,
    "timestamp": "2025-07-14T12:34:38.465506"
  },
  {
    "epoch": 1802,
    "batch_idx": 32,
    "loss": 0.40213295817375183,
    "learning_rate": 2.3988686527161737e-05,
    "timestep_mean": 425.78125,
    "epsilon_norm": 7.01851749420166,
    "prediction_norm": 5.928839683532715,
    "timestamp": "2025-07-14T12:34:38.470767"
  },
  {
    "epoch": 1802,
    "batch_idx": 64,
    "loss": 0.25644630193710327,
    "learning_rate": 2.3988686527161737e-05,
    "timestep_mean": 432.21875,
    "epsilon_norm": 8.211423873901367,
    "prediction_norm": 6.981071472167969,
    "timestamp": "2025-07-14T12:34:38.476237"
  },
  {
    "epoch": 1802,
    "batch_idx": 96,
    "loss": 0.37421518564224243,
    "learning_rate": 2.3988686527161737e-05,
    "timestep_mean": 511.5625,
    "epsilon_norm": 8.172592163085938,
    "prediction_norm": 6.508415222167969,
    "timestamp": "2025-07-14T12:34:38.481532"
  },
  {
    "epoch": 1802,
    "batch_idx": 128,
    "loss": 0.1033131405711174,
    "learning_rate": 2.3988686527161737e-05,
    "timestep_mean": 577.0,
    "epsilon_norm": 8.359692573547363,
    "prediction_norm": 7.712627410888672,
    "timestamp": "2025-07-14T12:34:38.486741"
  },
  {
    "epoch": 1802,
    "batch_idx": 160,
    "loss": 0.15786191821098328,
    "learning_rate": 2.3988686527161737e-05,
    "timestep_mean": 590.34375,
    "epsilon_norm": 8.025493621826172,
    "prediction_norm": 7.105672836303711,
    "timestamp": "2025-07-14T12:34:38.492017"
  },
  {
    "epoch": 1802,
    "batch_idx": 192,
    "loss": 0.16615891456604004,
    "learning_rate": 2.3988686527161737e-05,
    "timestep_mean": 550.5,
    "epsilon_norm": 4.736248970031738,
    "prediction_norm": 3.7891480922698975,
    "timestamp": "2025-07-14T12:34:38.497270"
  },
  {
    "epoch": 1803,
    "batch_idx": 0,
    "loss": 0.17966940999031067,
    "learning_rate": 2.3748920333177933e-05,
    "timestep_mean": 426.5625,
    "epsilon_norm": 7.5458984375,
    "prediction_norm": 6.5697808265686035,
    "timestamp": "2025-07-14T12:34:38.502620"
  },
  {
    "epoch": 1803,
    "batch_idx": 32,
    "loss": 0.2467660903930664,
    "learning_rate": 2.3748920333177933e-05,
    "timestep_mean": 485.21875,
    "epsilon_norm": 7.642349720001221,
    "prediction_norm": 6.6975226402282715,
    "timestamp": "2025-07-14T12:34:38.507720"
  },
  {
    "epoch": 1803,
    "batch_idx": 64,
    "loss": 0.24340781569480896,
    "learning_rate": 2.3748920333177933e-05,
    "timestep_mean": 553.5625,
    "epsilon_norm": 7.427708625793457,
    "prediction_norm": 6.7673258781433105,
    "timestamp": "2025-07-14T12:34:38.512951"
  },
  {
    "epoch": 1803,
    "batch_idx": 96,
    "loss": 0.3507940471172333,
    "learning_rate": 2.3748920333177933e-05,
    "timestep_mean": 485.8125,
    "epsilon_norm": 7.8238630294799805,
    "prediction_norm": 5.968371868133545,
    "timestamp": "2025-07-14T12:34:38.518098"
  },
  {
    "epoch": 1803,
    "batch_idx": 128,
    "loss": 0.21242448687553406,
    "learning_rate": 2.3748920333177933e-05,
    "timestep_mean": 486.96875,
    "epsilon_norm": 6.694103240966797,
    "prediction_norm": 5.623579502105713,
    "timestamp": "2025-07-14T12:34:38.523184"
  },
  {
    "epoch": 1803,
    "batch_idx": 160,
    "loss": 0.23842409253120422,
    "learning_rate": 2.3748920333177933e-05,
    "timestep_mean": 612.6875,
    "epsilon_norm": 8.028202056884766,
    "prediction_norm": 7.167464256286621,
    "timestamp": "2025-07-14T12:34:38.528393"
  },
  {
    "epoch": 1803,
    "batch_idx": 192,
    "loss": 0.2243349254131317,
    "learning_rate": 2.3748920333177933e-05,
    "timestep_mean": 555.0,
    "epsilon_norm": 4.764701843261719,
    "prediction_norm": 4.423209190368652,
    "timestamp": "2025-07-14T12:34:38.533748"
  },
  {
    "epoch": 1804,
    "batch_idx": 0,
    "loss": 0.25174546241760254,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 490.4375,
    "epsilon_norm": 7.7154364585876465,
    "prediction_norm": 6.787032127380371,
    "timestamp": "2025-07-14T12:34:38.538960"
  },
  {
    "epoch": 1804,
    "batch_idx": 32,
    "loss": 0.20274318754673004,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 554.25,
    "epsilon_norm": 8.019206047058105,
    "prediction_norm": 7.080992698669434,
    "timestamp": "2025-07-14T12:34:38.544255"
  },
  {
    "epoch": 1804,
    "batch_idx": 64,
    "loss": 0.13216550648212433,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 625.3125,
    "epsilon_norm": 6.2987470626831055,
    "prediction_norm": 5.7443366050720215,
    "timestamp": "2025-07-14T12:34:38.549496"
  },
  {
    "epoch": 1804,
    "batch_idx": 96,
    "loss": 0.31334352493286133,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 416.65625,
    "epsilon_norm": 7.718252182006836,
    "prediction_norm": 6.160415172576904,
    "timestamp": "2025-07-14T12:34:38.554956"
  },
  {
    "epoch": 1804,
    "batch_idx": 128,
    "loss": 0.4270237386226654,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 453.75,
    "epsilon_norm": 8.195571899414062,
    "prediction_norm": 6.0408172607421875,
    "timestamp": "2025-07-14T12:34:38.560248"
  },
  {
    "epoch": 1804,
    "batch_idx": 160,
    "loss": 0.4744538366794586,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 408.75,
    "epsilon_norm": 7.861235618591309,
    "prediction_norm": 5.4498209953308105,
    "timestamp": "2025-07-14T12:34:38.565424"
  },
  {
    "epoch": 1804,
    "batch_idx": 192,
    "loss": 0.5832297205924988,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 281.0,
    "epsilon_norm": 4.932743072509766,
    "prediction_norm": 4.343008518218994,
    "timestamp": "2025-07-14T12:34:38.570741"
  },
  {
    "epoch": 1805,
    "batch_idx": 0,
    "loss": 0.12872479856014252,
    "learning_rate": 2.3272913840499397e-05,
    "timestep_mean": 581.59375,
    "epsilon_norm": 7.560622692108154,
    "prediction_norm": 7.347736835479736,
    "timestamp": "2025-07-14T12:34:38.575726"
  },
  {
    "epoch": 1805,
    "batch_idx": 32,
    "loss": 0.16539674997329712,
    "learning_rate": 2.3272913840499397e-05,
    "timestep_mean": 425.21875,
    "epsilon_norm": 7.967799186706543,
    "prediction_norm": 7.470433712005615,
    "timestamp": "2025-07-14T12:34:38.580810"
  },
  {
    "epoch": 1805,
    "batch_idx": 64,
    "loss": 0.505047082901001,
    "learning_rate": 2.3272913840499397e-05,
    "timestep_mean": 447.375,
    "epsilon_norm": 8.470637321472168,
    "prediction_norm": 5.47432804107666,
    "timestamp": "2025-07-14T12:34:38.585984"
  },
  {
    "epoch": 1805,
    "batch_idx": 96,
    "loss": 0.21612806618213654,
    "learning_rate": 2.3272913840499397e-05,
    "timestep_mean": 498.5625,
    "epsilon_norm": 8.393021583557129,
    "prediction_norm": 7.427011966705322,
    "timestamp": "2025-07-14T12:34:38.591078"
  },
  {
    "epoch": 1805,
    "batch_idx": 128,
    "loss": 0.1550540030002594,
    "learning_rate": 2.3272913840499397e-05,
    "timestep_mean": 536.03125,
    "epsilon_norm": 7.36903715133667,
    "prediction_norm": 6.512551307678223,
    "timestamp": "2025-07-14T12:34:38.596294"
  },
  {
    "epoch": 1805,
    "batch_idx": 160,
    "loss": 0.3842424154281616,
    "learning_rate": 2.3272913840499397e-05,
    "timestep_mean": 438.15625,
    "epsilon_norm": 8.08232593536377,
    "prediction_norm": 6.528172492980957,
    "timestamp": "2025-07-14T12:34:38.601412"
  },
  {
    "epoch": 1805,
    "batch_idx": 192,
    "loss": 0.0648152083158493,
    "learning_rate": 2.3272913840499397e-05,
    "timestep_mean": 583.875,
    "epsilon_norm": 3.122894287109375,
    "prediction_norm": 2.7542593479156494,
    "timestamp": "2025-07-14T12:34:38.606510"
  },
  {
    "epoch": 1806,
    "batch_idx": 0,
    "loss": 0.3431991934776306,
    "learning_rate": 2.3036674716303273e-05,
    "timestep_mean": 513.0625,
    "epsilon_norm": 8.77431869506836,
    "prediction_norm": 6.5961833000183105,
    "timestamp": "2025-07-14T12:34:38.611800"
  },
  {
    "epoch": 1806,
    "batch_idx": 32,
    "loss": 0.22986724972724915,
    "learning_rate": 2.3036674716303273e-05,
    "timestep_mean": 527.1875,
    "epsilon_norm": 8.39969253540039,
    "prediction_norm": 7.407804489135742,
    "timestamp": "2025-07-14T12:34:38.617087"
  },
  {
    "epoch": 1806,
    "batch_idx": 64,
    "loss": 0.16621346771717072,
    "learning_rate": 2.3036674716303273e-05,
    "timestep_mean": 607.59375,
    "epsilon_norm": 8.576335906982422,
    "prediction_norm": 7.6741461753845215,
    "timestamp": "2025-07-14T12:34:38.622183"
  },
  {
    "epoch": 1806,
    "batch_idx": 96,
    "loss": 0.3708146810531616,
    "learning_rate": 2.3036674716303273e-05,
    "timestep_mean": 525.40625,
    "epsilon_norm": 8.8607177734375,
    "prediction_norm": 7.663943767547607,
    "timestamp": "2025-07-14T12:34:38.627363"
  },
  {
    "epoch": 1806,
    "batch_idx": 128,
    "loss": 0.3095918893814087,
    "learning_rate": 2.3036674716303273e-05,
    "timestep_mean": 466.90625,
    "epsilon_norm": 7.6728949546813965,
    "prediction_norm": 5.995759010314941,
    "timestamp": "2025-07-14T12:34:38.632490"
  },
  {
    "epoch": 1806,
    "batch_idx": 160,
    "loss": 0.15124304592609406,
    "learning_rate": 2.3036674716303273e-05,
    "timestep_mean": 589.375,
    "epsilon_norm": 7.844101428985596,
    "prediction_norm": 7.197356224060059,
    "timestamp": "2025-07-14T12:34:38.637729"
  },
  {
    "epoch": 1806,
    "batch_idx": 192,
    "loss": 0.10859768837690353,
    "learning_rate": 2.3036674716303273e-05,
    "timestep_mean": 512.5,
    "epsilon_norm": 3.1678032875061035,
    "prediction_norm": 3.042513847351074,
    "timestamp": "2025-07-14T12:34:38.643222"
  },
  {
    "epoch": 1807,
    "batch_idx": 0,
    "loss": 0.2911943197250366,
    "learning_rate": 2.2801612451698647e-05,
    "timestep_mean": 394.375,
    "epsilon_norm": 6.604055404663086,
    "prediction_norm": 5.11854887008667,
    "timestamp": "2025-07-14T12:34:38.652320"
  },
  {
    "epoch": 1807,
    "batch_idx": 32,
    "loss": 0.44733870029449463,
    "learning_rate": 2.2801612451698647e-05,
    "timestep_mean": 452.40625,
    "epsilon_norm": 8.13290023803711,
    "prediction_norm": 5.837488651275635,
    "timestamp": "2025-07-14T12:34:38.657847"
  },
  {
    "epoch": 1807,
    "batch_idx": 64,
    "loss": 0.3397977650165558,
    "learning_rate": 2.2801612451698647e-05,
    "timestep_mean": 491.65625,
    "epsilon_norm": 8.385284423828125,
    "prediction_norm": 7.025366306304932,
    "timestamp": "2025-07-14T12:34:38.662899"
  },
  {
    "epoch": 1807,
    "batch_idx": 96,
    "loss": 0.2793230414390564,
    "learning_rate": 2.2801612451698647e-05,
    "timestep_mean": 539.5625,
    "epsilon_norm": 7.995020389556885,
    "prediction_norm": 6.86359977722168,
    "timestamp": "2025-07-14T12:34:38.667722"
  },
  {
    "epoch": 1807,
    "batch_idx": 128,
    "loss": 0.26248499751091003,
    "learning_rate": 2.2801612451698647e-05,
    "timestep_mean": 433.09375,
    "epsilon_norm": 7.535594940185547,
    "prediction_norm": 6.608705520629883,
    "timestamp": "2025-07-14T12:34:38.672658"
  },
  {
    "epoch": 1807,
    "batch_idx": 160,
    "loss": 0.13365645706653595,
    "learning_rate": 2.2801612451698647e-05,
    "timestep_mean": 495.46875,
    "epsilon_norm": 7.243614196777344,
    "prediction_norm": 6.585599422454834,
    "timestamp": "2025-07-14T12:34:38.677496"
  },
  {
    "epoch": 1807,
    "batch_idx": 192,
    "loss": 0.25708818435668945,
    "learning_rate": 2.2801612451698647e-05,
    "timestep_mean": 350.0,
    "epsilon_norm": 3.297207832336426,
    "prediction_norm": 2.9309351444244385,
    "timestamp": "2025-07-14T12:34:38.682484"
  },
  {
    "epoch": 1808,
    "batch_idx": 0,
    "loss": 0.21690240502357483,
    "learning_rate": 2.256772762667852e-05,
    "timestep_mean": 415.78125,
    "epsilon_norm": 7.013298511505127,
    "prediction_norm": 6.061220169067383,
    "timestamp": "2025-07-14T12:34:38.687478"
  },
  {
    "epoch": 1808,
    "batch_idx": 32,
    "loss": 0.3295883536338806,
    "learning_rate": 2.256772762667852e-05,
    "timestep_mean": 419.59375,
    "epsilon_norm": 8.71325397491455,
    "prediction_norm": 7.53420352935791,
    "timestamp": "2025-07-14T12:34:38.692429"
  },
  {
    "epoch": 1808,
    "batch_idx": 64,
    "loss": 0.2253609001636505,
    "learning_rate": 2.256772762667852e-05,
    "timestep_mean": 538.0625,
    "epsilon_norm": 7.742242336273193,
    "prediction_norm": 7.02878999710083,
    "timestamp": "2025-07-14T12:34:38.697431"
  },
  {
    "epoch": 1808,
    "batch_idx": 96,
    "loss": 0.2950315475463867,
    "learning_rate": 2.256772762667852e-05,
    "timestep_mean": 514.6875,
    "epsilon_norm": 8.126605033874512,
    "prediction_norm": 6.909494876861572,
    "timestamp": "2025-07-14T12:34:38.702292"
  },
  {
    "epoch": 1808,
    "batch_idx": 128,
    "loss": 0.23504434525966644,
    "learning_rate": 2.256772762667852e-05,
    "timestep_mean": 500.75,
    "epsilon_norm": 7.471641540527344,
    "prediction_norm": 6.459652423858643,
    "timestamp": "2025-07-14T12:34:38.707196"
  },
  {
    "epoch": 1808,
    "batch_idx": 160,
    "loss": 0.14028355479240417,
    "learning_rate": 2.256772762667852e-05,
    "timestep_mean": 522.90625,
    "epsilon_norm": 7.309074401855469,
    "prediction_norm": 6.6993231773376465,
    "timestamp": "2025-07-14T12:34:38.711995"
  },
  {
    "epoch": 1808,
    "batch_idx": 192,
    "loss": 0.1893545389175415,
    "learning_rate": 2.256772762667852e-05,
    "timestep_mean": 611.625,
    "epsilon_norm": 4.131089687347412,
    "prediction_norm": 3.497445583343506,
    "timestamp": "2025-07-14T12:34:38.717058"
  },
  {
    "epoch": 1809,
    "batch_idx": 0,
    "loss": 0.24144232273101807,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 492.6875,
    "epsilon_norm": 8.231436729431152,
    "prediction_norm": 6.699894428253174,
    "timestamp": "2025-07-14T12:34:38.721957"
  },
  {
    "epoch": 1809,
    "batch_idx": 32,
    "loss": 0.3858456015586853,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 540.46875,
    "epsilon_norm": 7.555486679077148,
    "prediction_norm": 5.809206962585449,
    "timestamp": "2025-07-14T12:34:38.726793"
  },
  {
    "epoch": 1809,
    "batch_idx": 64,
    "loss": 0.5597416758537292,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 522.75,
    "epsilon_norm": 7.8554182052612305,
    "prediction_norm": 5.488838195800781,
    "timestamp": "2025-07-14T12:34:38.731588"
  },
  {
    "epoch": 1809,
    "batch_idx": 96,
    "loss": 0.2471681833267212,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 476.15625,
    "epsilon_norm": 8.379122734069824,
    "prediction_norm": 6.776609897613525,
    "timestamp": "2025-07-14T12:34:38.736371"
  },
  {
    "epoch": 1809,
    "batch_idx": 128,
    "loss": 0.2775843143463135,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 540.03125,
    "epsilon_norm": 8.684138298034668,
    "prediction_norm": 7.8029632568359375,
    "timestamp": "2025-07-14T12:34:38.741162"
  },
  {
    "epoch": 1809,
    "batch_idx": 160,
    "loss": 0.25939926505088806,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 533.1875,
    "epsilon_norm": 7.825323104858398,
    "prediction_norm": 6.934963226318359,
    "timestamp": "2025-07-14T12:34:38.746075"
  },
  {
    "epoch": 1809,
    "batch_idx": 192,
    "loss": 0.23012666404247284,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 544.75,
    "epsilon_norm": 3.1377243995666504,
    "prediction_norm": 2.527461051940918,
    "timestamp": "2025-07-14T12:34:38.750983"
  },
  {
    "epoch": 1810,
    "batch_idx": 0,
    "loss": 0.42921367287635803,
    "learning_rate": 2.2103492600834934e-05,
    "timestep_mean": 476.53125,
    "epsilon_norm": 9.0011625289917,
    "prediction_norm": 7.397736072540283,
    "timestamp": "2025-07-14T12:34:38.755834"
  },
  {
    "epoch": 1810,
    "batch_idx": 32,
    "loss": 0.13234573602676392,
    "learning_rate": 2.2103492600834934e-05,
    "timestep_mean": 593.8125,
    "epsilon_norm": 7.233206748962402,
    "prediction_norm": 6.691351890563965,
    "timestamp": "2025-07-14T12:34:38.760817"
  },
  {
    "epoch": 1810,
    "batch_idx": 64,
    "loss": 0.2476431429386139,
    "learning_rate": 2.2103492600834934e-05,
    "timestep_mean": 537.46875,
    "epsilon_norm": 9.468478202819824,
    "prediction_norm": 7.997335433959961,
    "timestamp": "2025-07-14T12:34:38.765651"
  },
  {
    "epoch": 1810,
    "batch_idx": 96,
    "loss": 0.2491745948791504,
    "learning_rate": 2.2103492600834934e-05,
    "timestep_mean": 494.4375,
    "epsilon_norm": 6.6966986656188965,
    "prediction_norm": 5.323245525360107,
    "timestamp": "2025-07-14T12:34:38.770634"
  },
  {
    "epoch": 1810,
    "batch_idx": 128,
    "loss": 0.1738366037607193,
    "learning_rate": 2.2103492600834934e-05,
    "timestep_mean": 530.90625,
    "epsilon_norm": 8.44931411743164,
    "prediction_norm": 7.322288990020752,
    "timestamp": "2025-07-14T12:34:38.775559"
  },
  {
    "epoch": 1810,
    "batch_idx": 160,
    "loss": 0.3381360173225403,
    "learning_rate": 2.2103492600834934e-05,
    "timestep_mean": 585.5,
    "epsilon_norm": 7.957237243652344,
    "prediction_norm": 6.88348388671875,
    "timestamp": "2025-07-14T12:34:38.780498"
  },
  {
    "epoch": 1810,
    "batch_idx": 192,
    "loss": 0.483146607875824,
    "learning_rate": 2.2103492600834934e-05,
    "timestep_mean": 503.875,
    "epsilon_norm": 3.858232021331787,
    "prediction_norm": 3.059417247772217,
    "timestamp": "2025-07-14T12:34:38.785477"
  },
  {
    "epoch": 1811,
    "batch_idx": 0,
    "loss": 0.18023107945919037,
    "learning_rate": 2.1873143545465318e-05,
    "timestep_mean": 509.53125,
    "epsilon_norm": 7.1279120445251465,
    "prediction_norm": 6.5795416831970215,
    "timestamp": "2025-07-14T12:34:38.790521"
  },
  {
    "epoch": 1811,
    "batch_idx": 32,
    "loss": 0.4541941285133362,
    "learning_rate": 2.1873143545465318e-05,
    "timestep_mean": 412.90625,
    "epsilon_norm": 7.837994575500488,
    "prediction_norm": 5.705084323883057,
    "timestamp": "2025-07-14T12:34:38.795537"
  },
  {
    "epoch": 1811,
    "batch_idx": 64,
    "loss": 0.27248549461364746,
    "learning_rate": 2.1873143545465318e-05,
    "timestep_mean": 419.59375,
    "epsilon_norm": 7.770712852478027,
    "prediction_norm": 6.634641170501709,
    "timestamp": "2025-07-14T12:34:38.800591"
  },
  {
    "epoch": 1811,
    "batch_idx": 96,
    "loss": 0.13375595211982727,
    "learning_rate": 2.1873143545465318e-05,
    "timestep_mean": 560.3125,
    "epsilon_norm": 7.62957763671875,
    "prediction_norm": 6.894817352294922,
    "timestamp": "2025-07-14T12:34:38.805537"
  },
  {
    "epoch": 1811,
    "batch_idx": 128,
    "loss": 0.38173723220825195,
    "learning_rate": 2.1873143545465318e-05,
    "timestep_mean": 574.53125,
    "epsilon_norm": 8.21660041809082,
    "prediction_norm": 7.30571985244751,
    "timestamp": "2025-07-14T12:34:38.810487"
  },
  {
    "epoch": 1811,
    "batch_idx": 160,
    "loss": 0.31940901279449463,
    "learning_rate": 2.1873143545465318e-05,
    "timestep_mean": 487.09375,
    "epsilon_norm": 7.71082067489624,
    "prediction_norm": 5.857255458831787,
    "timestamp": "2025-07-14T12:34:38.815474"
  },
  {
    "epoch": 1811,
    "batch_idx": 192,
    "loss": 1.0859241485595703,
    "learning_rate": 2.1873143545465318e-05,
    "timestep_mean": 500.625,
    "epsilon_norm": 5.4502034187316895,
    "prediction_norm": 2.8320727348327637,
    "timestamp": "2025-07-14T12:34:38.820443"
  },
  {
    "epoch": 1812,
    "batch_idx": 0,
    "loss": 0.23399454355239868,
    "learning_rate": 2.164397422058472e-05,
    "timestep_mean": 456.53125,
    "epsilon_norm": 7.5680928230285645,
    "prediction_norm": 6.053579330444336,
    "timestamp": "2025-07-14T12:34:38.825552"
  },
  {
    "epoch": 1812,
    "batch_idx": 32,
    "loss": 0.3087134063243866,
    "learning_rate": 2.164397422058472e-05,
    "timestep_mean": 413.40625,
    "epsilon_norm": 7.863522529602051,
    "prediction_norm": 6.449141025543213,
    "timestamp": "2025-07-14T12:34:38.830487"
  },
  {
    "epoch": 1812,
    "batch_idx": 64,
    "loss": 0.588733434677124,
    "learning_rate": 2.164397422058472e-05,
    "timestep_mean": 409.53125,
    "epsilon_norm": 8.224540710449219,
    "prediction_norm": 5.579957008361816,
    "timestamp": "2025-07-14T12:34:38.835422"
  },
  {
    "epoch": 1812,
    "batch_idx": 96,
    "loss": 0.35479551553726196,
    "learning_rate": 2.164397422058472e-05,
    "timestep_mean": 475.125,
    "epsilon_norm": 9.166147232055664,
    "prediction_norm": 8.074275970458984,
    "timestamp": "2025-07-14T12:34:38.840357"
  },
  {
    "epoch": 1812,
    "batch_idx": 128,
    "loss": 0.2506232261657715,
    "learning_rate": 2.164397422058472e-05,
    "timestep_mean": 552.75,
    "epsilon_norm": 7.765619277954102,
    "prediction_norm": 6.7231550216674805,
    "timestamp": "2025-07-14T12:34:38.845330"
  },
  {
    "epoch": 1812,
    "batch_idx": 160,
    "loss": 0.26736515760421753,
    "learning_rate": 2.164397422058472e-05,
    "timestep_mean": 596.78125,
    "epsilon_norm": 10.213706970214844,
    "prediction_norm": 9.234926223754883,
    "timestamp": "2025-07-14T12:34:38.850350"
  },
  {
    "epoch": 1812,
    "batch_idx": 192,
    "loss": 0.07917323708534241,
    "learning_rate": 2.164397422058472e-05,
    "timestep_mean": 634.125,
    "epsilon_norm": 3.341881036758423,
    "prediction_norm": 2.8692729473114014,
    "timestamp": "2025-07-14T12:34:38.855262"
  },
  {
    "epoch": 1813,
    "batch_idx": 0,
    "loss": 0.2674003839492798,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 493.4375,
    "epsilon_norm": 7.96251916885376,
    "prediction_norm": 6.787330150604248,
    "timestamp": "2025-07-14T12:34:38.860253"
  },
  {
    "epoch": 1813,
    "batch_idx": 32,
    "loss": 0.24916014075279236,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 403.65625,
    "epsilon_norm": 7.118114471435547,
    "prediction_norm": 5.961066722869873,
    "timestamp": "2025-07-14T12:34:38.865185"
  },
  {
    "epoch": 1813,
    "batch_idx": 64,
    "loss": 0.29294413328170776,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 447.90625,
    "epsilon_norm": 9.169739723205566,
    "prediction_norm": 7.9604010581970215,
    "timestamp": "2025-07-14T12:34:38.870144"
  },
  {
    "epoch": 1813,
    "batch_idx": 96,
    "loss": 0.09916660189628601,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 565.28125,
    "epsilon_norm": 6.89730978012085,
    "prediction_norm": 6.782677173614502,
    "timestamp": "2025-07-14T12:34:38.875299"
  },
  {
    "epoch": 1813,
    "batch_idx": 128,
    "loss": 0.278273344039917,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 480.25,
    "epsilon_norm": 8.552480697631836,
    "prediction_norm": 7.291932106018066,
    "timestamp": "2025-07-14T12:34:38.880223"
  },
  {
    "epoch": 1813,
    "batch_idx": 160,
    "loss": 0.4150583744049072,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 499.4375,
    "epsilon_norm": 8.817980766296387,
    "prediction_norm": 7.602086067199707,
    "timestamp": "2025-07-14T12:34:38.886001"
  },
  {
    "epoch": 1813,
    "batch_idx": 192,
    "loss": 0.14944426715373993,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 442.5,
    "epsilon_norm": 3.311392307281494,
    "prediction_norm": 3.314363718032837,
    "timestamp": "2025-07-14T12:34:38.891381"
  },
  {
    "epoch": 1814,
    "batch_idx": 0,
    "loss": 0.751496434211731,
    "learning_rate": 2.118917702118888e-05,
    "timestep_mean": 408.75,
    "epsilon_norm": 9.25196647644043,
    "prediction_norm": 6.191483974456787,
    "timestamp": "2025-07-14T12:34:38.902091"
  },
  {
    "epoch": 1814,
    "batch_idx": 32,
    "loss": 0.3426138758659363,
    "learning_rate": 2.118917702118888e-05,
    "timestep_mean": 426.1875,
    "epsilon_norm": 9.078374862670898,
    "prediction_norm": 7.614892959594727,
    "timestamp": "2025-07-14T12:34:38.908033"
  },
  {
    "epoch": 1814,
    "batch_idx": 64,
    "loss": 0.17353487014770508,
    "learning_rate": 2.118917702118888e-05,
    "timestep_mean": 470.125,
    "epsilon_norm": 7.493182182312012,
    "prediction_norm": 6.71596097946167,
    "timestamp": "2025-07-14T12:34:38.913407"
  },
  {
    "epoch": 1814,
    "batch_idx": 96,
    "loss": 0.2956029176712036,
    "learning_rate": 2.118917702118888e-05,
    "timestep_mean": 465.4375,
    "epsilon_norm": 7.573495864868164,
    "prediction_norm": 6.132569789886475,
    "timestamp": "2025-07-14T12:34:38.918423"
  },
  {
    "epoch": 1814,
    "batch_idx": 128,
    "loss": 0.5475393533706665,
    "learning_rate": 2.118917702118888e-05,
    "timestep_mean": 541.875,
    "epsilon_norm": 9.029759407043457,
    "prediction_norm": 6.339387893676758,
    "timestamp": "2025-07-14T12:34:38.923799"
  },
  {
    "epoch": 1814,
    "batch_idx": 160,
    "loss": 0.1795942783355713,
    "learning_rate": 2.118917702118888e-05,
    "timestep_mean": 463.21875,
    "epsilon_norm": 7.889244556427002,
    "prediction_norm": 7.162042140960693,
    "timestamp": "2025-07-14T12:34:38.929287"
  },
  {
    "epoch": 1814,
    "batch_idx": 192,
    "loss": 0.43836596608161926,
    "learning_rate": 2.118917702118888e-05,
    "timestep_mean": 476.0,
    "epsilon_norm": 4.362172603607178,
    "prediction_norm": 2.9198780059814453,
    "timestamp": "2025-07-14T12:34:38.934654"
  },
  {
    "epoch": 1815,
    "batch_idx": 0,
    "loss": 0.2561489939689636,
    "learning_rate": 2.0963550268840384e-05,
    "timestep_mean": 524.09375,
    "epsilon_norm": 8.531335830688477,
    "prediction_norm": 7.642307758331299,
    "timestamp": "2025-07-14T12:34:38.940063"
  },
  {
    "epoch": 1815,
    "batch_idx": 32,
    "loss": 0.13507196307182312,
    "learning_rate": 2.0963550268840384e-05,
    "timestep_mean": 503.875,
    "epsilon_norm": 7.708739280700684,
    "prediction_norm": 6.649896621704102,
    "timestamp": "2025-07-14T12:34:38.945171"
  },
  {
    "epoch": 1815,
    "batch_idx": 64,
    "loss": 0.2998921573162079,
    "learning_rate": 2.0963550268840384e-05,
    "timestep_mean": 486.46875,
    "epsilon_norm": 8.054855346679688,
    "prediction_norm": 6.046576499938965,
    "timestamp": "2025-07-14T12:34:38.950628"
  },
  {
    "epoch": 1815,
    "batch_idx": 96,
    "loss": 0.15218275785446167,
    "learning_rate": 2.0963550268840384e-05,
    "timestep_mean": 542.03125,
    "epsilon_norm": 8.67124080657959,
    "prediction_norm": 7.745551109313965,
    "timestamp": "2025-07-14T12:34:38.955676"
  },
  {
    "epoch": 1815,
    "batch_idx": 128,
    "loss": 0.266448438167572,
    "learning_rate": 2.0963550268840384e-05,
    "timestep_mean": 503.1875,
    "epsilon_norm": 8.89395809173584,
    "prediction_norm": 7.748972415924072,
    "timestamp": "2025-07-14T12:34:38.960770"
  },
  {
    "epoch": 1815,
    "batch_idx": 160,
    "loss": 0.3246069848537445,
    "learning_rate": 2.0963550268840384e-05,
    "timestep_mean": 359.84375,
    "epsilon_norm": 7.7411065101623535,
    "prediction_norm": 6.278129577636719,
    "timestamp": "2025-07-14T12:34:38.966048"
  },
  {
    "epoch": 1815,
    "batch_idx": 192,
    "loss": 0.1016821563243866,
    "learning_rate": 2.0963550268840384e-05,
    "timestep_mean": 559.25,
    "epsilon_norm": 2.4776885509490967,
    "prediction_norm": 2.02130126953125,
    "timestamp": "2025-07-14T12:34:38.971388"
  },
  {
    "epoch": 1816,
    "batch_idx": 0,
    "loss": 0.19237110018730164,
    "learning_rate": 2.0739105491312072e-05,
    "timestep_mean": 510.3125,
    "epsilon_norm": 7.011259078979492,
    "prediction_norm": 6.274628639221191,
    "timestamp": "2025-07-14T12:34:38.976478"
  },
  {
    "epoch": 1816,
    "batch_idx": 32,
    "loss": 0.1577976644039154,
    "learning_rate": 2.0739105491312072e-05,
    "timestep_mean": 563.625,
    "epsilon_norm": 8.148509979248047,
    "prediction_norm": 7.6555399894714355,
    "timestamp": "2025-07-14T12:34:38.981596"
  },
  {
    "epoch": 1816,
    "batch_idx": 64,
    "loss": 0.18870961666107178,
    "learning_rate": 2.0739105491312072e-05,
    "timestep_mean": 545.21875,
    "epsilon_norm": 8.493026733398438,
    "prediction_norm": 7.748113632202148,
    "timestamp": "2025-07-14T12:34:38.986681"
  },
  {
    "epoch": 1816,
    "batch_idx": 96,
    "loss": 0.273174911737442,
    "learning_rate": 2.0739105491312072e-05,
    "timestep_mean": 441.65625,
    "epsilon_norm": 7.820765972137451,
    "prediction_norm": 6.296527862548828,
    "timestamp": "2025-07-14T12:34:38.992130"
  },
  {
    "epoch": 1816,
    "batch_idx": 128,
    "loss": 0.1310935765504837,
    "learning_rate": 2.0739105491312072e-05,
    "timestep_mean": 535.09375,
    "epsilon_norm": 8.015130996704102,
    "prediction_norm": 7.673855781555176,
    "timestamp": "2025-07-14T12:34:38.997316"
  },
  {
    "epoch": 1816,
    "batch_idx": 160,
    "loss": 0.17281334102153778,
    "learning_rate": 2.0739105491312072e-05,
    "timestep_mean": 566.28125,
    "epsilon_norm": 7.143126487731934,
    "prediction_norm": 6.047945499420166,
    "timestamp": "2025-07-14T12:34:39.002444"
  },
  {
    "epoch": 1816,
    "batch_idx": 192,
    "loss": 0.3300066292285919,
    "learning_rate": 2.0739105491312072e-05,
    "timestep_mean": 472.75,
    "epsilon_norm": 4.324892044067383,
    "prediction_norm": 3.4738245010375977,
    "timestamp": "2025-07-14T12:34:39.007808"
  },
  {
    "epoch": 1817,
    "batch_idx": 0,
    "loss": 0.18876200914382935,
    "learning_rate": 2.0515843242398958e-05,
    "timestep_mean": 438.21875,
    "epsilon_norm": 7.311282157897949,
    "prediction_norm": 5.7083563804626465,
    "timestamp": "2025-07-14T12:34:39.013347"
  },
  {
    "epoch": 1817,
    "batch_idx": 32,
    "loss": 0.18150119483470917,
    "learning_rate": 2.0515843242398958e-05,
    "timestep_mean": 539.5625,
    "epsilon_norm": 8.708464622497559,
    "prediction_norm": 7.640850067138672,
    "timestamp": "2025-07-14T12:34:39.018706"
  },
  {
    "epoch": 1817,
    "batch_idx": 64,
    "loss": 0.29895925521850586,
    "learning_rate": 2.0515843242398958e-05,
    "timestep_mean": 531.15625,
    "epsilon_norm": 9.617524147033691,
    "prediction_norm": 7.87853479385376,
    "timestamp": "2025-07-14T12:34:39.023937"
  },
  {
    "epoch": 1817,
    "batch_idx": 96,
    "loss": 0.35411107540130615,
    "learning_rate": 2.0515843242398958e-05,
    "timestep_mean": 490.15625,
    "epsilon_norm": 8.23452091217041,
    "prediction_norm": 6.983088493347168,
    "timestamp": "2025-07-14T12:34:39.029282"
  },
  {
    "epoch": 1817,
    "batch_idx": 128,
    "loss": 0.15849994122982025,
    "learning_rate": 2.0515843242398958e-05,
    "timestep_mean": 604.34375,
    "epsilon_norm": 8.115215301513672,
    "prediction_norm": 7.384558200836182,
    "timestamp": "2025-07-14T12:34:39.034661"
  },
  {
    "epoch": 1817,
    "batch_idx": 160,
    "loss": 0.12753361463546753,
    "learning_rate": 2.0515843242398958e-05,
    "timestep_mean": 560.9375,
    "epsilon_norm": 7.364589691162109,
    "prediction_norm": 6.88265323638916,
    "timestamp": "2025-07-14T12:34:39.039999"
  },
  {
    "epoch": 1817,
    "batch_idx": 192,
    "loss": 0.11064700782299042,
    "learning_rate": 2.0515843242398958e-05,
    "timestep_mean": 493.125,
    "epsilon_norm": 4.026546478271484,
    "prediction_norm": 3.8227014541625977,
    "timestamp": "2025-07-14T12:34:39.045394"
  },
  {
    "epoch": 1818,
    "batch_idx": 0,
    "loss": 0.29341065883636475,
    "learning_rate": 2.0293764072978607e-05,
    "timestep_mean": 484.125,
    "epsilon_norm": 8.573638916015625,
    "prediction_norm": 7.127934455871582,
    "timestamp": "2025-07-14T12:34:39.050893"
  },
  {
    "epoch": 1818,
    "batch_idx": 32,
    "loss": 0.2528306841850281,
    "learning_rate": 2.0293764072978607e-05,
    "timestep_mean": 439.0,
    "epsilon_norm": 7.651772975921631,
    "prediction_norm": 5.974305629730225,
    "timestamp": "2025-07-14T12:34:39.056109"
  },
  {
    "epoch": 1818,
    "batch_idx": 64,
    "loss": 0.18573763966560364,
    "learning_rate": 2.0293764072978607e-05,
    "timestep_mean": 540.90625,
    "epsilon_norm": 8.411172866821289,
    "prediction_norm": 7.272221565246582,
    "timestamp": "2025-07-14T12:34:39.061336"
  },
  {
    "epoch": 1818,
    "batch_idx": 96,
    "loss": 0.529323399066925,
    "learning_rate": 2.0293764072978607e-05,
    "timestep_mean": 458.46875,
    "epsilon_norm": 8.055183410644531,
    "prediction_norm": 6.157559871673584,
    "timestamp": "2025-07-14T12:34:39.066816"
  },
  {
    "epoch": 1818,
    "batch_idx": 128,
    "loss": 0.2691502571105957,
    "learning_rate": 2.0293764072978607e-05,
    "timestep_mean": 396.0625,
    "epsilon_norm": 7.463814735412598,
    "prediction_norm": 6.023116588592529,
    "timestamp": "2025-07-14T12:34:39.072173"
  },
  {
    "epoch": 1818,
    "batch_idx": 160,
    "loss": 0.43034082651138306,
    "learning_rate": 2.0293764072978607e-05,
    "timestep_mean": 414.75,
    "epsilon_norm": 8.482441902160645,
    "prediction_norm": 6.85374641418457,
    "timestamp": "2025-07-14T12:34:39.077437"
  },
  {
    "epoch": 1818,
    "batch_idx": 192,
    "loss": 0.22326241433620453,
    "learning_rate": 2.0293764072978607e-05,
    "timestep_mean": 537.375,
    "epsilon_norm": 3.627506732940674,
    "prediction_norm": 2.8615691661834717,
    "timestamp": "2025-07-14T12:34:39.082678"
  },
  {
    "epoch": 1819,
    "batch_idx": 0,
    "loss": 0.2335422784090042,
    "learning_rate": 2.007286853100914e-05,
    "timestep_mean": 496.28125,
    "epsilon_norm": 6.967986106872559,
    "prediction_norm": 6.5183563232421875,
    "timestamp": "2025-07-14T12:34:39.087949"
  },
  {
    "epoch": 1819,
    "batch_idx": 32,
    "loss": 0.4320812523365021,
    "learning_rate": 2.007286853100914e-05,
    "timestep_mean": 422.25,
    "epsilon_norm": 8.769485473632812,
    "prediction_norm": 7.042762279510498,
    "timestamp": "2025-07-14T12:34:39.093149"
  },
  {
    "epoch": 1819,
    "batch_idx": 64,
    "loss": 0.3078579306602478,
    "learning_rate": 2.007286853100914e-05,
    "timestep_mean": 554.28125,
    "epsilon_norm": 8.298284530639648,
    "prediction_norm": 7.189974784851074,
    "timestamp": "2025-07-14T12:34:39.098305"
  },
  {
    "epoch": 1819,
    "batch_idx": 96,
    "loss": 0.28197818994522095,
    "learning_rate": 2.007286853100914e-05,
    "timestep_mean": 506.0,
    "epsilon_norm": 7.189395427703857,
    "prediction_norm": 6.107336044311523,
    "timestamp": "2025-07-14T12:34:39.103638"
  },
  {
    "epoch": 1819,
    "batch_idx": 128,
    "loss": 0.24073663353919983,
    "learning_rate": 2.007286853100914e-05,
    "timestep_mean": 532.8125,
    "epsilon_norm": 8.217839241027832,
    "prediction_norm": 6.827615261077881,
    "timestamp": "2025-07-14T12:34:39.108996"
  },
  {
    "epoch": 1819,
    "batch_idx": 160,
    "loss": 0.25361067056655884,
    "learning_rate": 2.007286853100914e-05,
    "timestep_mean": 542.21875,
    "epsilon_norm": 8.341596603393555,
    "prediction_norm": 6.8142218589782715,
    "timestamp": "2025-07-14T12:34:39.114259"
  },
  {
    "epoch": 1819,
    "batch_idx": 192,
    "loss": 0.2901933193206787,
    "learning_rate": 2.007286853100914e-05,
    "timestep_mean": 551.0,
    "epsilon_norm": 3.5037143230438232,
    "prediction_norm": 3.1233887672424316,
    "timestamp": "2025-07-14T12:34:39.119710"
  },
  {
    "epoch": 1820,
    "batch_idx": 0,
    "loss": 0.2684206962585449,
    "learning_rate": 1.9853157161528462e-05,
    "timestep_mean": 448.625,
    "epsilon_norm": 6.350780963897705,
    "prediction_norm": 5.249520778656006,
    "timestamp": "2025-07-14T12:34:39.125195"
  },
  {
    "epoch": 1820,
    "batch_idx": 32,
    "loss": 0.3314707279205322,
    "learning_rate": 1.9853157161528462e-05,
    "timestep_mean": 399.5,
    "epsilon_norm": 7.777401447296143,
    "prediction_norm": 6.07456111907959,
    "timestamp": "2025-07-14T12:34:39.130653"
  },
  {
    "epoch": 1820,
    "batch_idx": 64,
    "loss": 0.3286355137825012,
    "learning_rate": 1.9853157161528462e-05,
    "timestep_mean": 486.8125,
    "epsilon_norm": 8.17829418182373,
    "prediction_norm": 6.3216071128845215,
    "timestamp": "2025-07-14T12:34:39.136209"
  },
  {
    "epoch": 1820,
    "batch_idx": 96,
    "loss": 0.2941078245639801,
    "learning_rate": 1.9853157161528462e-05,
    "timestep_mean": 423.4375,
    "epsilon_norm": 7.635787010192871,
    "prediction_norm": 6.64583158493042,
    "timestamp": "2025-07-14T12:34:39.143342"
  },
  {
    "epoch": 1820,
    "batch_idx": 128,
    "loss": 0.4223216772079468,
    "learning_rate": 1.9853157161528462e-05,
    "timestep_mean": 384.6875,
    "epsilon_norm": 7.414462089538574,
    "prediction_norm": 5.520713806152344,
    "timestamp": "2025-07-14T12:34:39.149308"
  },
  {
    "epoch": 1820,
    "batch_idx": 160,
    "loss": 0.4122430682182312,
    "learning_rate": 1.9853157161528462e-05,
    "timestep_mean": 493.03125,
    "epsilon_norm": 8.465007781982422,
    "prediction_norm": 7.068929195404053,
    "timestamp": "2025-07-14T12:34:39.160289"
  },
  {
    "epoch": 1820,
    "batch_idx": 192,
    "loss": 0.10169777274131775,
    "learning_rate": 1.9853157161528462e-05,
    "timestep_mean": 477.0,
    "epsilon_norm": 3.3148574829101562,
    "prediction_norm": 3.208430290222168,
    "timestamp": "2025-07-14T12:34:39.165984"
  },
  {
    "epoch": 1821,
    "batch_idx": 0,
    "loss": 0.3837938904762268,
    "learning_rate": 1.9634630506652533e-05,
    "timestep_mean": 415.375,
    "epsilon_norm": 7.618461608886719,
    "prediction_norm": 6.8364715576171875,
    "timestamp": "2025-07-14T12:34:39.171416"
  },
  {
    "epoch": 1821,
    "batch_idx": 32,
    "loss": 0.2536769509315491,
    "learning_rate": 1.9634630506652533e-05,
    "timestep_mean": 470.8125,
    "epsilon_norm": 7.06250524520874,
    "prediction_norm": 6.141976833343506,
    "timestamp": "2025-07-14T12:34:39.176586"
  },
  {
    "epoch": 1821,
    "batch_idx": 64,
    "loss": 0.31812065839767456,
    "learning_rate": 1.9634630506652533e-05,
    "timestep_mean": 482.03125,
    "epsilon_norm": 7.365202903747559,
    "prediction_norm": 5.838576793670654,
    "timestamp": "2025-07-14T12:34:39.181720"
  },
  {
    "epoch": 1821,
    "batch_idx": 96,
    "loss": 0.22383633255958557,
    "learning_rate": 1.9634630506652533e-05,
    "timestep_mean": 572.40625,
    "epsilon_norm": 8.653423309326172,
    "prediction_norm": 7.155018329620361,
    "timestamp": "2025-07-14T12:34:39.186847"
  },
  {
    "epoch": 1821,
    "batch_idx": 128,
    "loss": 0.5073660016059875,
    "learning_rate": 1.9634630506652533e-05,
    "timestep_mean": 436.28125,
    "epsilon_norm": 8.353303909301758,
    "prediction_norm": 6.477667808532715,
    "timestamp": "2025-07-14T12:34:39.192105"
  },
  {
    "epoch": 1821,
    "batch_idx": 160,
    "loss": 0.1748201549053192,
    "learning_rate": 1.9634630506652533e-05,
    "timestep_mean": 468.53125,
    "epsilon_norm": 9.284629821777344,
    "prediction_norm": 8.462657928466797,
    "timestamp": "2025-07-14T12:34:39.196993"
  },
  {
    "epoch": 1821,
    "batch_idx": 192,
    "loss": 0.10377795994281769,
    "learning_rate": 1.9634630506652533e-05,
    "timestep_mean": 506.875,
    "epsilon_norm": 3.8418357372283936,
    "prediction_norm": 3.3654212951660156,
    "timestamp": "2025-07-14T12:34:39.202066"
  },
  {
    "epoch": 1822,
    "batch_idx": 0,
    "loss": 0.1703779101371765,
    "learning_rate": 1.941728910557405e-05,
    "timestep_mean": 508.625,
    "epsilon_norm": 7.176180839538574,
    "prediction_norm": 6.462286949157715,
    "timestamp": "2025-07-14T12:34:39.207006"
  },
  {
    "epoch": 1822,
    "batch_idx": 32,
    "loss": 0.3747009038925171,
    "learning_rate": 1.941728910557405e-05,
    "timestep_mean": 494.8125,
    "epsilon_norm": 9.990748405456543,
    "prediction_norm": 8.06125545501709,
    "timestamp": "2025-07-14T12:34:39.211867"
  },
  {
    "epoch": 1822,
    "batch_idx": 64,
    "loss": 0.26621514558792114,
    "learning_rate": 1.941728910557405e-05,
    "timestep_mean": 472.0625,
    "epsilon_norm": 7.694098949432373,
    "prediction_norm": 6.915750503540039,
    "timestamp": "2025-07-14T12:34:39.216941"
  },
  {
    "epoch": 1822,
    "batch_idx": 96,
    "loss": 0.2136131227016449,
    "learning_rate": 1.941728910557405e-05,
    "timestep_mean": 563.8125,
    "epsilon_norm": 7.951185703277588,
    "prediction_norm": 6.822546005249023,
    "timestamp": "2025-07-14T12:34:39.221771"
  },
  {
    "epoch": 1822,
    "batch_idx": 128,
    "loss": 0.17302948236465454,
    "learning_rate": 1.941728910557405e-05,
    "timestep_mean": 520.53125,
    "epsilon_norm": 8.255956649780273,
    "prediction_norm": 7.430737018585205,
    "timestamp": "2025-07-14T12:34:39.226750"
  },
  {
    "epoch": 1822,
    "batch_idx": 160,
    "loss": 0.26068729162216187,
    "learning_rate": 1.941728910557405e-05,
    "timestep_mean": 516.09375,
    "epsilon_norm": 8.490107536315918,
    "prediction_norm": 7.2283172607421875,
    "timestamp": "2025-07-14T12:34:39.231713"
  },
  {
    "epoch": 1822,
    "batch_idx": 192,
    "loss": 0.11146807670593262,
    "learning_rate": 1.941728910557405e-05,
    "timestep_mean": 458.75,
    "epsilon_norm": 2.7481274604797363,
    "prediction_norm": 2.5264453887939453,
    "timestamp": "2025-07-14T12:34:39.236683"
  },
  {
    "epoch": 1823,
    "batch_idx": 0,
    "loss": 0.16489636898040771,
    "learning_rate": 1.9201133494561425e-05,
    "timestep_mean": 480.0625,
    "epsilon_norm": 6.966078758239746,
    "prediction_norm": 6.091337203979492,
    "timestamp": "2025-07-14T12:34:39.241802"
  },
  {
    "epoch": 1823,
    "batch_idx": 32,
    "loss": 0.2452210783958435,
    "learning_rate": 1.9201133494561425e-05,
    "timestep_mean": 506.59375,
    "epsilon_norm": 8.428702354431152,
    "prediction_norm": 7.623110771179199,
    "timestamp": "2025-07-14T12:34:39.246686"
  },
  {
    "epoch": 1823,
    "batch_idx": 64,
    "loss": 0.1445346474647522,
    "learning_rate": 1.9201133494561425e-05,
    "timestep_mean": 570.40625,
    "epsilon_norm": 8.737090110778809,
    "prediction_norm": 8.133950233459473,
    "timestamp": "2025-07-14T12:34:39.251488"
  },
  {
    "epoch": 1823,
    "batch_idx": 96,
    "loss": 0.34188634157180786,
    "learning_rate": 1.9201133494561425e-05,
    "timestep_mean": 493.40625,
    "epsilon_norm": 8.416547775268555,
    "prediction_norm": 6.70913028717041,
    "timestamp": "2025-07-14T12:34:39.256350"
  },
  {
    "epoch": 1823,
    "batch_idx": 128,
    "loss": 0.1811094880104065,
    "learning_rate": 1.9201133494561425e-05,
    "timestep_mean": 534.875,
    "epsilon_norm": 8.625341415405273,
    "prediction_norm": 8.32046127319336,
    "timestamp": "2025-07-14T12:34:39.261220"
  },
  {
    "epoch": 1823,
    "batch_idx": 160,
    "loss": 0.3483903408050537,
    "learning_rate": 1.9201133494561425e-05,
    "timestep_mean": 417.53125,
    "epsilon_norm": 8.422479629516602,
    "prediction_norm": 7.085999488830566,
    "timestamp": "2025-07-14T12:34:39.266075"
  },
  {
    "epoch": 1823,
    "batch_idx": 192,
    "loss": 0.1810636818408966,
    "learning_rate": 1.9201133494561425e-05,
    "timestep_mean": 668.375,
    "epsilon_norm": 3.5986974239349365,
    "prediction_norm": 3.1178009510040283,
    "timestamp": "2025-07-14T12:34:39.270972"
  },
  {
    "epoch": 1824,
    "batch_idx": 0,
    "loss": 0.42259442806243896,
    "learning_rate": 1.8986164206957087e-05,
    "timestep_mean": 554.625,
    "epsilon_norm": 8.024571418762207,
    "prediction_norm": 6.364903926849365,
    "timestamp": "2025-07-14T12:34:39.276085"
  },
  {
    "epoch": 1824,
    "batch_idx": 32,
    "loss": 0.14105644822120667,
    "learning_rate": 1.8986164206957087e-05,
    "timestep_mean": 590.78125,
    "epsilon_norm": 9.188806533813477,
    "prediction_norm": 8.600834846496582,
    "timestamp": "2025-07-14T12:34:39.280967"
  },
  {
    "epoch": 1824,
    "batch_idx": 64,
    "loss": 0.300594300031662,
    "learning_rate": 1.8986164206957087e-05,
    "timestep_mean": 380.53125,
    "epsilon_norm": 8.007513046264648,
    "prediction_norm": 6.273004531860352,
    "timestamp": "2025-07-14T12:34:39.285889"
  },
  {
    "epoch": 1824,
    "batch_idx": 96,
    "loss": 0.1926669180393219,
    "learning_rate": 1.8986164206957087e-05,
    "timestep_mean": 452.4375,
    "epsilon_norm": 6.490400791168213,
    "prediction_norm": 5.603267669677734,
    "timestamp": "2025-07-14T12:34:39.290845"
  },
  {
    "epoch": 1824,
    "batch_idx": 128,
    "loss": 0.3421703279018402,
    "learning_rate": 1.8986164206957087e-05,
    "timestep_mean": 531.625,
    "epsilon_norm": 8.048133850097656,
    "prediction_norm": 7.070465564727783,
    "timestamp": "2025-07-14T12:34:39.295796"
  },
  {
    "epoch": 1824,
    "batch_idx": 160,
    "loss": 0.3531193733215332,
    "learning_rate": 1.8986164206957087e-05,
    "timestep_mean": 504.59375,
    "epsilon_norm": 7.376495361328125,
    "prediction_norm": 5.863356590270996,
    "timestamp": "2025-07-14T12:34:39.300879"
  },
  {
    "epoch": 1824,
    "batch_idx": 192,
    "loss": 0.09994445741176605,
    "learning_rate": 1.8986164206957087e-05,
    "timestep_mean": 639.375,
    "epsilon_norm": 3.1714136600494385,
    "prediction_norm": 3.312208652496338,
    "timestamp": "2025-07-14T12:34:39.305966"
  },
  {
    "epoch": 1825,
    "batch_idx": 0,
    "loss": 0.1469508707523346,
    "learning_rate": 1.8772381773176416e-05,
    "timestep_mean": 528.28125,
    "epsilon_norm": 7.404071807861328,
    "prediction_norm": 7.122040748596191,
    "timestamp": "2025-07-14T12:34:39.311004"
  },
  {
    "epoch": 1825,
    "batch_idx": 32,
    "loss": 0.24037188291549683,
    "learning_rate": 1.8772381773176416e-05,
    "timestep_mean": 485.5625,
    "epsilon_norm": 8.462553024291992,
    "prediction_norm": 7.547820568084717,
    "timestamp": "2025-07-14T12:34:39.316002"
  },
  {
    "epoch": 1825,
    "batch_idx": 64,
    "loss": 0.2008959949016571,
    "learning_rate": 1.8772381773176416e-05,
    "timestep_mean": 511.0,
    "epsilon_norm": 7.90755033493042,
    "prediction_norm": 7.059706211090088,
    "timestamp": "2025-07-14T12:34:39.320960"
  },
  {
    "epoch": 1825,
    "batch_idx": 96,
    "loss": 0.2645663619041443,
    "learning_rate": 1.8772381773176416e-05,
    "timestep_mean": 464.25,
    "epsilon_norm": 6.825926780700684,
    "prediction_norm": 5.578324317932129,
    "timestamp": "2025-07-14T12:34:39.325900"
  },
  {
    "epoch": 1825,
    "batch_idx": 128,
    "loss": 0.1862412989139557,
    "learning_rate": 1.8772381773176416e-05,
    "timestep_mean": 547.15625,
    "epsilon_norm": 8.1499662399292,
    "prediction_norm": 7.604833126068115,
    "timestamp": "2025-07-14T12:34:39.330846"
  },
  {
    "epoch": 1825,
    "batch_idx": 160,
    "loss": 0.26966217160224915,
    "learning_rate": 1.8772381773176416e-05,
    "timestep_mean": 479.5,
    "epsilon_norm": 7.669929504394531,
    "prediction_norm": 6.5291056632995605,
    "timestamp": "2025-07-14T12:34:39.335904"
  },
  {
    "epoch": 1825,
    "batch_idx": 192,
    "loss": 0.3476473093032837,
    "learning_rate": 1.8772381773176416e-05,
    "timestep_mean": 527.375,
    "epsilon_norm": 4.284974575042725,
    "prediction_norm": 3.8789846897125244,
    "timestamp": "2025-07-14T12:34:39.340996"
  },
  {
    "epoch": 1826,
    "batch_idx": 0,
    "loss": 0.24458461999893188,
    "learning_rate": 1.8559786720706182e-05,
    "timestep_mean": 487.5,
    "epsilon_norm": 7.7046661376953125,
    "prediction_norm": 6.80758810043335,
    "timestamp": "2025-07-14T12:34:39.346074"
  },
  {
    "epoch": 1826,
    "batch_idx": 32,
    "loss": 0.3695056736469269,
    "learning_rate": 1.8559786720706182e-05,
    "timestep_mean": 440.96875,
    "epsilon_norm": 7.179537296295166,
    "prediction_norm": 5.772292137145996,
    "timestamp": "2025-07-14T12:34:39.351062"
  },
  {
    "epoch": 1826,
    "batch_idx": 64,
    "loss": 0.2042607069015503,
    "learning_rate": 1.8559786720706182e-05,
    "timestep_mean": 491.84375,
    "epsilon_norm": 7.899508953094482,
    "prediction_norm": 6.878039360046387,
    "timestamp": "2025-07-14T12:34:39.356067"
  },
  {
    "epoch": 1826,
    "batch_idx": 96,
    "loss": 0.15491408109664917,
    "learning_rate": 1.8559786720706182e-05,
    "timestep_mean": 628.625,
    "epsilon_norm": 7.7585368156433105,
    "prediction_norm": 7.381602764129639,
    "timestamp": "2025-07-14T12:34:39.361055"
  },
  {
    "epoch": 1826,
    "batch_idx": 128,
    "loss": 0.2944343090057373,
    "learning_rate": 1.8559786720706182e-05,
    "timestep_mean": 527.40625,
    "epsilon_norm": 8.213166236877441,
    "prediction_norm": 7.2816925048828125,
    "timestamp": "2025-07-14T12:34:39.365969"
  },
  {
    "epoch": 1826,
    "batch_idx": 160,
    "loss": 0.274366557598114,
    "learning_rate": 1.8559786720706182e-05,
    "timestep_mean": 538.3125,
    "epsilon_norm": 8.09825611114502,
    "prediction_norm": 6.708084583282471,
    "timestamp": "2025-07-14T12:34:39.370890"
  },
  {
    "epoch": 1826,
    "batch_idx": 192,
    "loss": 0.4164940416812897,
    "learning_rate": 1.8559786720706182e-05,
    "timestep_mean": 504.375,
    "epsilon_norm": 4.45076322555542,
    "prediction_norm": 3.208649158477783,
    "timestamp": "2025-07-14T12:34:39.375983"
  },
  {
    "epoch": 1827,
    "batch_idx": 0,
    "loss": 0.21985848248004913,
    "learning_rate": 1.8348379574103788e-05,
    "timestep_mean": 512.875,
    "epsilon_norm": 8.343558311462402,
    "prediction_norm": 6.891213417053223,
    "timestamp": "2025-07-14T12:34:39.380904"
  },
  {
    "epoch": 1827,
    "batch_idx": 32,
    "loss": 0.1843462884426117,
    "learning_rate": 1.8348379574103788e-05,
    "timestep_mean": 476.1875,
    "epsilon_norm": 7.177880764007568,
    "prediction_norm": 6.540369510650635,
    "timestamp": "2025-07-14T12:34:39.385832"
  },
  {
    "epoch": 1827,
    "batch_idx": 64,
    "loss": 0.21418845653533936,
    "learning_rate": 1.8348379574103788e-05,
    "timestep_mean": 541.0625,
    "epsilon_norm": 7.345737457275391,
    "prediction_norm": 6.461557388305664,
    "timestamp": "2025-07-14T12:34:39.390827"
  },
  {
    "epoch": 1827,
    "batch_idx": 96,
    "loss": 0.23805630207061768,
    "learning_rate": 1.8348379574103788e-05,
    "timestep_mean": 535.9375,
    "epsilon_norm": 6.772752285003662,
    "prediction_norm": 5.838115692138672,
    "timestamp": "2025-07-14T12:34:39.395758"
  },
  {
    "epoch": 1827,
    "batch_idx": 128,
    "loss": 0.2575366199016571,
    "learning_rate": 1.8348379574103788e-05,
    "timestep_mean": 519.9375,
    "epsilon_norm": 8.129186630249023,
    "prediction_norm": 7.08751916885376,
    "timestamp": "2025-07-14T12:34:39.400749"
  },
  {
    "epoch": 1827,
    "batch_idx": 160,
    "loss": 0.1642812192440033,
    "learning_rate": 1.8348379574103788e-05,
    "timestep_mean": 442.09375,
    "epsilon_norm": 8.028803825378418,
    "prediction_norm": 7.0099029541015625,
    "timestamp": "2025-07-14T12:34:39.405721"
  },
  {
    "epoch": 1827,
    "batch_idx": 192,
    "loss": 0.7605284452438354,
    "learning_rate": 1.8348379574103788e-05,
    "timestep_mean": 437.25,
    "epsilon_norm": 5.250747203826904,
    "prediction_norm": 3.711794137954712,
    "timestamp": "2025-07-14T12:34:39.410705"
  },
  {
    "epoch": 1828,
    "batch_idx": 0,
    "loss": 0.15229684114456177,
    "learning_rate": 1.8138160854995144e-05,
    "timestep_mean": 510.25,
    "epsilon_norm": 8.009588241577148,
    "prediction_norm": 7.211767673492432,
    "timestamp": "2025-07-14T12:34:39.415933"
  },
  {
    "epoch": 1828,
    "batch_idx": 32,
    "loss": 0.32360348105430603,
    "learning_rate": 1.8138160854995144e-05,
    "timestep_mean": 476.84375,
    "epsilon_norm": 8.614252090454102,
    "prediction_norm": 6.76788330078125,
    "timestamp": "2025-07-14T12:34:39.421087"
  },
  {
    "epoch": 1828,
    "batch_idx": 64,
    "loss": 0.3293198347091675,
    "learning_rate": 1.8138160854995144e-05,
    "timestep_mean": 406.15625,
    "epsilon_norm": 8.55004596710205,
    "prediction_norm": 6.977243900299072,
    "timestamp": "2025-07-14T12:34:39.426226"
  },
  {
    "epoch": 1828,
    "batch_idx": 96,
    "loss": 0.21114042401313782,
    "learning_rate": 1.8138160854995144e-05,
    "timestep_mean": 490.84375,
    "epsilon_norm": 7.752170085906982,
    "prediction_norm": 6.513565540313721,
    "timestamp": "2025-07-14T12:34:39.431228"
  },
  {
    "epoch": 1828,
    "batch_idx": 128,
    "loss": 0.3964402973651886,
    "learning_rate": 1.8138160854995144e-05,
    "timestep_mean": 472.03125,
    "epsilon_norm": 7.62628173828125,
    "prediction_norm": 6.1465582847595215,
    "timestamp": "2025-07-14T12:34:39.436360"
  },
  {
    "epoch": 1828,
    "batch_idx": 160,
    "loss": 0.223394513130188,
    "learning_rate": 1.8138160854995144e-05,
    "timestep_mean": 580.375,
    "epsilon_norm": 8.219152450561523,
    "prediction_norm": 6.897030830383301,
    "timestamp": "2025-07-14T12:34:39.441319"
  },
  {
    "epoch": 1828,
    "batch_idx": 192,
    "loss": 0.09650741517543793,
    "learning_rate": 1.8138160854995144e-05,
    "timestep_mean": 498.5,
    "epsilon_norm": 3.253493309020996,
    "prediction_norm": 3.1954009532928467,
    "timestamp": "2025-07-14T12:34:39.446535"
  },
  {
    "epoch": 1829,
    "batch_idx": 0,
    "loss": 0.32936573028564453,
    "learning_rate": 1.792913108207411e-05,
    "timestep_mean": 444.5625,
    "epsilon_norm": 8.418537139892578,
    "prediction_norm": 7.549211025238037,
    "timestamp": "2025-07-14T12:34:39.451605"
  },
  {
    "epoch": 1829,
    "batch_idx": 32,
    "loss": 0.17092379927635193,
    "learning_rate": 1.792913108207411e-05,
    "timestep_mean": 593.5625,
    "epsilon_norm": 7.941655158996582,
    "prediction_norm": 7.860843658447266,
    "timestamp": "2025-07-14T12:34:39.456628"
  },
  {
    "epoch": 1829,
    "batch_idx": 64,
    "loss": 0.24194654822349548,
    "learning_rate": 1.792913108207411e-05,
    "timestep_mean": 505.09375,
    "epsilon_norm": 7.372491359710693,
    "prediction_norm": 6.509183883666992,
    "timestamp": "2025-07-14T12:34:39.461661"
  },
  {
    "epoch": 1829,
    "batch_idx": 96,
    "loss": 0.3752160370349884,
    "learning_rate": 1.792913108207411e-05,
    "timestep_mean": 406.21875,
    "epsilon_norm": 8.01357650756836,
    "prediction_norm": 6.7276530265808105,
    "timestamp": "2025-07-14T12:34:39.466616"
  },
  {
    "epoch": 1829,
    "batch_idx": 128,
    "loss": 0.19592474400997162,
    "learning_rate": 1.792913108207411e-05,
    "timestep_mean": 455.8125,
    "epsilon_norm": 7.570497989654541,
    "prediction_norm": 6.915703296661377,
    "timestamp": "2025-07-14T12:34:39.471599"
  },
  {
    "epoch": 1829,
    "batch_idx": 160,
    "loss": 0.26569175720214844,
    "learning_rate": 1.792913108207411e-05,
    "timestep_mean": 471.71875,
    "epsilon_norm": 7.562589645385742,
    "prediction_norm": 6.261807918548584,
    "timestamp": "2025-07-14T12:34:39.476506"
  },
  {
    "epoch": 1829,
    "batch_idx": 192,
    "loss": 0.21357184648513794,
    "learning_rate": 1.792913108207411e-05,
    "timestep_mean": 353.0,
    "epsilon_norm": 2.991243839263916,
    "prediction_norm": 2.2018613815307617,
    "timestamp": "2025-07-14T12:34:39.481424"
  },
  {
    "epoch": 1830,
    "batch_idx": 0,
    "loss": 0.2035159468650818,
    "learning_rate": 1.7721290771101016e-05,
    "timestep_mean": 572.40625,
    "epsilon_norm": 6.380299091339111,
    "prediction_norm": 6.024228572845459,
    "timestamp": "2025-07-14T12:34:39.486407"
  },
  {
    "epoch": 1830,
    "batch_idx": 32,
    "loss": 0.2991017699241638,
    "learning_rate": 1.7721290771101016e-05,
    "timestep_mean": 446.84375,
    "epsilon_norm": 8.112345695495605,
    "prediction_norm": 6.779957294464111,
    "timestamp": "2025-07-14T12:34:39.491434"
  },
  {
    "epoch": 1830,
    "batch_idx": 64,
    "loss": 0.17531557381153107,
    "learning_rate": 1.7721290771101016e-05,
    "timestep_mean": 534.125,
    "epsilon_norm": 8.312265396118164,
    "prediction_norm": 7.221719741821289,
    "timestamp": "2025-07-14T12:34:39.496473"
  },
  {
    "epoch": 1830,
    "batch_idx": 96,
    "loss": 0.17086923122406006,
    "learning_rate": 1.7721290771101016e-05,
    "timestep_mean": 506.78125,
    "epsilon_norm": 8.197149276733398,
    "prediction_norm": 7.46576452255249,
    "timestamp": "2025-07-14T12:34:39.501553"
  },
  {
    "epoch": 1830,
    "batch_idx": 128,
    "loss": 0.15242694318294525,
    "learning_rate": 1.7721290771101016e-05,
    "timestep_mean": 479.375,
    "epsilon_norm": 6.771670341491699,
    "prediction_norm": 6.168522357940674,
    "timestamp": "2025-07-14T12:34:39.506582"
  },
  {
    "epoch": 1830,
    "batch_idx": 160,
    "loss": 0.3641684353351593,
    "learning_rate": 1.7721290771101016e-05,
    "timestep_mean": 506.28125,
    "epsilon_norm": 9.115443229675293,
    "prediction_norm": 7.408121109008789,
    "timestamp": "2025-07-14T12:34:39.511544"
  },
  {
    "epoch": 1830,
    "batch_idx": 192,
    "loss": 0.05632850527763367,
    "learning_rate": 1.7721290771101016e-05,
    "timestep_mean": 653.125,
    "epsilon_norm": 4.550039291381836,
    "prediction_norm": 4.472305774688721,
    "timestamp": "2025-07-14T12:34:39.516632"
  },
  {
    "epoch": 1831,
    "batch_idx": 0,
    "loss": 0.20505833625793457,
    "learning_rate": 1.7514640434900863e-05,
    "timestep_mean": 559.71875,
    "epsilon_norm": 8.341700553894043,
    "prediction_norm": 7.590424060821533,
    "timestamp": "2025-07-14T12:34:39.521682"
  },
  {
    "epoch": 1831,
    "batch_idx": 32,
    "loss": 0.17445677518844604,
    "learning_rate": 1.7514640434900863e-05,
    "timestep_mean": 443.125,
    "epsilon_norm": 6.5729289054870605,
    "prediction_norm": 5.630584239959717,
    "timestamp": "2025-07-14T12:34:39.526616"
  },
  {
    "epoch": 1831,
    "batch_idx": 64,
    "loss": 0.31407275795936584,
    "learning_rate": 1.7514640434900863e-05,
    "timestep_mean": 469.0625,
    "epsilon_norm": 7.606504440307617,
    "prediction_norm": 6.475650787353516,
    "timestamp": "2025-07-14T12:34:39.531628"
  },
  {
    "epoch": 1831,
    "batch_idx": 96,
    "loss": 0.19983170926570892,
    "learning_rate": 1.7514640434900863e-05,
    "timestep_mean": 531.5,
    "epsilon_norm": 8.219741821289062,
    "prediction_norm": 6.9361066818237305,
    "timestamp": "2025-07-14T12:34:39.536652"
  },
  {
    "epoch": 1831,
    "batch_idx": 128,
    "loss": 0.34470486640930176,
    "learning_rate": 1.7514640434900863e-05,
    "timestep_mean": 476.6875,
    "epsilon_norm": 8.987920761108398,
    "prediction_norm": 7.568150520324707,
    "timestamp": "2025-07-14T12:34:39.541804"
  },
  {
    "epoch": 1831,
    "batch_idx": 160,
    "loss": 0.5202131271362305,
    "learning_rate": 1.7514640434900863e-05,
    "timestep_mean": 448.0,
    "epsilon_norm": 8.121892929077148,
    "prediction_norm": 5.7207207679748535,
    "timestamp": "2025-07-14T12:34:39.546719"
  },
  {
    "epoch": 1831,
    "batch_idx": 192,
    "loss": 0.06468065828084946,
    "learning_rate": 1.7514640434900863e-05,
    "timestep_mean": 534.125,
    "epsilon_norm": 4.196883678436279,
    "prediction_norm": 3.966447114944458,
    "timestamp": "2025-07-14T12:34:39.551723"
  },
  {
    "epoch": 1832,
    "batch_idx": 0,
    "loss": 0.28463274240493774,
    "learning_rate": 1.730918058336306e-05,
    "timestep_mean": 442.9375,
    "epsilon_norm": 8.021554946899414,
    "prediction_norm": 6.650393009185791,
    "timestamp": "2025-07-14T12:34:39.556759"
  },
  {
    "epoch": 1832,
    "batch_idx": 32,
    "loss": 0.4817796051502228,
    "learning_rate": 1.730918058336306e-05,
    "timestep_mean": 484.5625,
    "epsilon_norm": 8.483062744140625,
    "prediction_norm": 6.714438438415527,
    "timestamp": "2025-07-14T12:34:39.561900"
  },
  {
    "epoch": 1832,
    "batch_idx": 64,
    "loss": 0.3360946774482727,
    "learning_rate": 1.730918058336306e-05,
    "timestep_mean": 467.40625,
    "epsilon_norm": 8.732184410095215,
    "prediction_norm": 7.21277379989624,
    "timestamp": "2025-07-14T12:34:39.567057"
  },
  {
    "epoch": 1832,
    "batch_idx": 96,
    "loss": 0.17971104383468628,
    "learning_rate": 1.730918058336306e-05,
    "timestep_mean": 520.03125,
    "epsilon_norm": 7.930473804473877,
    "prediction_norm": 6.683879852294922,
    "timestamp": "2025-07-14T12:34:39.572179"
  },
  {
    "epoch": 1832,
    "batch_idx": 128,
    "loss": 0.2599819004535675,
    "learning_rate": 1.730918058336306e-05,
    "timestep_mean": 501.21875,
    "epsilon_norm": 8.547385215759277,
    "prediction_norm": 7.61903715133667,
    "timestamp": "2025-07-14T12:34:39.577164"
  },
  {
    "epoch": 1832,
    "batch_idx": 160,
    "loss": 0.2139964997768402,
    "learning_rate": 1.730918058336306e-05,
    "timestep_mean": 420.0625,
    "epsilon_norm": 7.946009635925293,
    "prediction_norm": 6.7041497230529785,
    "timestamp": "2025-07-14T12:34:39.582127"
  },
  {
    "epoch": 1832,
    "batch_idx": 192,
    "loss": 0.19201600551605225,
    "learning_rate": 1.730918058336306e-05,
    "timestep_mean": 442.0,
    "epsilon_norm": 3.9480793476104736,
    "prediction_norm": 3.653947353363037,
    "timestamp": "2025-07-14T12:34:39.587106"
  },
  {
    "epoch": 1833,
    "batch_idx": 0,
    "loss": 0.2970202565193176,
    "learning_rate": 1.7104911723439307e-05,
    "timestep_mean": 510.8125,
    "epsilon_norm": 8.095115661621094,
    "prediction_norm": 7.004295349121094,
    "timestamp": "2025-07-14T12:34:39.592235"
  },
  {
    "epoch": 1833,
    "batch_idx": 32,
    "loss": 0.2576460838317871,
    "learning_rate": 1.7104911723439307e-05,
    "timestep_mean": 446.9375,
    "epsilon_norm": 7.970350742340088,
    "prediction_norm": 6.654376983642578,
    "timestamp": "2025-07-14T12:34:39.597196"
  },
  {
    "epoch": 1833,
    "batch_idx": 64,
    "loss": 0.34126460552215576,
    "learning_rate": 1.7104911723439307e-05,
    "timestep_mean": 493.0,
    "epsilon_norm": 7.964629650115967,
    "prediction_norm": 6.72085428237915,
    "timestamp": "2025-07-14T12:34:39.602166"
  },
  {
    "epoch": 1833,
    "batch_idx": 96,
    "loss": 0.23703297972679138,
    "learning_rate": 1.7104911723439307e-05,
    "timestep_mean": 477.8125,
    "epsilon_norm": 7.594803333282471,
    "prediction_norm": 7.013558387756348,
    "timestamp": "2025-07-14T12:34:39.607209"
  },
  {
    "epoch": 1833,
    "batch_idx": 128,
    "loss": 0.42237114906311035,
    "learning_rate": 1.7104911723439307e-05,
    "timestep_mean": 454.65625,
    "epsilon_norm": 8.91055679321289,
    "prediction_norm": 6.763063430786133,
    "timestamp": "2025-07-14T12:34:39.612267"
  },
  {
    "epoch": 1833,
    "batch_idx": 160,
    "loss": 0.3749924898147583,
    "learning_rate": 1.7104911723439307e-05,
    "timestep_mean": 485.53125,
    "epsilon_norm": 8.01828384399414,
    "prediction_norm": 7.144037246704102,
    "timestamp": "2025-07-14T12:34:39.617426"
  },
  {
    "epoch": 1833,
    "batch_idx": 192,
    "loss": 0.266757607460022,
    "learning_rate": 1.7104911723439307e-05,
    "timestep_mean": 357.25,
    "epsilon_norm": 4.477518081665039,
    "prediction_norm": 4.018489360809326,
    "timestamp": "2025-07-14T12:34:39.622383"
  },
  {
    "epoch": 1834,
    "batch_idx": 0,
    "loss": 0.21386097371578217,
    "learning_rate": 1.6901834359142608e-05,
    "timestep_mean": 500.90625,
    "epsilon_norm": 7.168882846832275,
    "prediction_norm": 5.660480976104736,
    "timestamp": "2025-07-14T12:34:39.627324"
  },
  {
    "epoch": 1834,
    "batch_idx": 32,
    "loss": 0.27077415585517883,
    "learning_rate": 1.6901834359142608e-05,
    "timestep_mean": 496.78125,
    "epsilon_norm": 8.376890182495117,
    "prediction_norm": 7.497109413146973,
    "timestamp": "2025-07-14T12:34:39.632285"
  },
  {
    "epoch": 1834,
    "batch_idx": 64,
    "loss": 0.32085925340652466,
    "learning_rate": 1.6901834359142608e-05,
    "timestep_mean": 477.5,
    "epsilon_norm": 7.341317653656006,
    "prediction_norm": 6.04200553894043,
    "timestamp": "2025-07-14T12:34:39.637226"
  },
  {
    "epoch": 1834,
    "batch_idx": 96,
    "loss": 0.21475329995155334,
    "learning_rate": 1.6901834359142608e-05,
    "timestep_mean": 551.59375,
    "epsilon_norm": 8.226390838623047,
    "prediction_norm": 7.286470890045166,
    "timestamp": "2025-07-14T12:34:39.642272"
  },
  {
    "epoch": 1834,
    "batch_idx": 128,
    "loss": 0.28611332178115845,
    "learning_rate": 1.6901834359142608e-05,
    "timestep_mean": 452.40625,
    "epsilon_norm": 7.007177352905273,
    "prediction_norm": 5.856049060821533,
    "timestamp": "2025-07-14T12:34:39.647281"
  },
  {
    "epoch": 1834,
    "batch_idx": 160,
    "loss": 0.24297182261943817,
    "learning_rate": 1.6901834359142608e-05,
    "timestep_mean": 452.1875,
    "epsilon_norm": 8.229597091674805,
    "prediction_norm": 6.980365753173828,
    "timestamp": "2025-07-14T12:34:39.652196"
  },
  {
    "epoch": 1834,
    "batch_idx": 192,
    "loss": 0.22587065398693085,
    "learning_rate": 1.6901834359142608e-05,
    "timestep_mean": 646.125,
    "epsilon_norm": 4.184055328369141,
    "prediction_norm": 3.7025976181030273,
    "timestamp": "2025-07-14T12:34:39.658688"
  },
  {
    "epoch": 1835,
    "batch_idx": 0,
    "loss": 0.45717960596084595,
    "learning_rate": 1.6699948991546418e-05,
    "timestep_mean": 471.0,
    "epsilon_norm": 8.184937477111816,
    "prediction_norm": 6.3609843254089355,
    "timestamp": "2025-07-14T12:34:39.672600"
  },
  {
    "epoch": 1835,
    "batch_idx": 32,
    "loss": 0.35885846614837646,
    "learning_rate": 1.6699948991546418e-05,
    "timestep_mean": 470.84375,
    "epsilon_norm": 7.6941118240356445,
    "prediction_norm": 6.0983123779296875,
    "timestamp": "2025-07-14T12:34:39.680523"
  },
  {
    "epoch": 1835,
    "batch_idx": 64,
    "loss": 0.1768529713153839,
    "learning_rate": 1.6699948991546418e-05,
    "timestep_mean": 510.9375,
    "epsilon_norm": 7.651233673095703,
    "prediction_norm": 6.444150924682617,
    "timestamp": "2025-07-14T12:34:39.689141"
  },
  {
    "epoch": 1835,
    "batch_idx": 96,
    "loss": 0.36157476902008057,
    "learning_rate": 1.6699948991546418e-05,
    "timestep_mean": 417.8125,
    "epsilon_norm": 8.312541007995605,
    "prediction_norm": 6.5865583419799805,
    "timestamp": "2025-07-14T12:34:39.697609"
  },
  {
    "epoch": 1835,
    "batch_idx": 128,
    "loss": 0.251178503036499,
    "learning_rate": 1.6699948991546418e-05,
    "timestep_mean": 437.5625,
    "epsilon_norm": 8.375746726989746,
    "prediction_norm": 6.994948387145996,
    "timestamp": "2025-07-14T12:34:39.705729"
  },
  {
    "epoch": 1835,
    "batch_idx": 160,
    "loss": 0.3819189667701721,
    "learning_rate": 1.6699948991546418e-05,
    "timestep_mean": 584.71875,
    "epsilon_norm": 8.386293411254883,
    "prediction_norm": 7.322421073913574,
    "timestamp": "2025-07-14T12:34:39.714079"
  },
  {
    "epoch": 1835,
    "batch_idx": 192,
    "loss": 0.12590517103672028,
    "learning_rate": 1.6699948991546418e-05,
    "timestep_mean": 642.375,
    "epsilon_norm": 5.023776531219482,
    "prediction_norm": 4.341323375701904,
    "timestamp": "2025-07-14T12:34:39.722445"
  },
  {
    "epoch": 1836,
    "batch_idx": 0,
    "loss": 0.11323121935129166,
    "learning_rate": 1.64992561187825e-05,
    "timestep_mean": 551.28125,
    "epsilon_norm": 6.968255043029785,
    "prediction_norm": 6.668519020080566,
    "timestamp": "2025-07-14T12:34:39.728396"
  },
  {
    "epoch": 1836,
    "batch_idx": 32,
    "loss": 0.1947035789489746,
    "learning_rate": 1.64992561187825e-05,
    "timestep_mean": 537.25,
    "epsilon_norm": 8.094673156738281,
    "prediction_norm": 7.30072546005249,
    "timestamp": "2025-07-14T12:34:39.734694"
  },
  {
    "epoch": 1836,
    "batch_idx": 64,
    "loss": 0.1981881558895111,
    "learning_rate": 1.64992561187825e-05,
    "timestep_mean": 581.03125,
    "epsilon_norm": 7.67431640625,
    "prediction_norm": 6.8634562492370605,
    "timestamp": "2025-07-14T12:34:39.743125"
  },
  {
    "epoch": 1836,
    "batch_idx": 96,
    "loss": 0.3244176506996155,
    "learning_rate": 1.64992561187825e-05,
    "timestep_mean": 531.0,
    "epsilon_norm": 7.306253910064697,
    "prediction_norm": 5.509409427642822,
    "timestamp": "2025-07-14T12:34:39.751145"
  },
  {
    "epoch": 1836,
    "batch_idx": 128,
    "loss": 0.36500126123428345,
    "learning_rate": 1.64992561187825e-05,
    "timestep_mean": 486.28125,
    "epsilon_norm": 7.614962100982666,
    "prediction_norm": 6.394340991973877,
    "timestamp": "2025-07-14T12:34:39.758218"
  },
  {
    "epoch": 1836,
    "batch_idx": 160,
    "loss": 0.29025107622146606,
    "learning_rate": 1.64992561187825e-05,
    "timestep_mean": 563.40625,
    "epsilon_norm": 7.804328918457031,
    "prediction_norm": 6.3612060546875,
    "timestamp": "2025-07-14T12:34:39.764996"
  },
  {
    "epoch": 1836,
    "batch_idx": 192,
    "loss": 0.4489300549030304,
    "learning_rate": 1.64992561187825e-05,
    "timestep_mean": 470.75,
    "epsilon_norm": 4.9224700927734375,
    "prediction_norm": 4.046101093292236,
    "timestamp": "2025-07-14T12:34:39.771115"
  },
  {
    "epoch": 1837,
    "batch_idx": 0,
    "loss": 0.11595325917005539,
    "learning_rate": 1.629975623604074e-05,
    "timestep_mean": 554.9375,
    "epsilon_norm": 8.177967071533203,
    "prediction_norm": 7.3018646240234375,
    "timestamp": "2025-07-14T12:34:39.777472"
  },
  {
    "epoch": 1837,
    "batch_idx": 32,
    "loss": 0.14435122907161713,
    "learning_rate": 1.629975623604074e-05,
    "timestep_mean": 489.4375,
    "epsilon_norm": 6.696779727935791,
    "prediction_norm": 6.373706340789795,
    "timestamp": "2025-07-14T12:34:39.784636"
  },
  {
    "epoch": 1837,
    "batch_idx": 64,
    "loss": 0.2506486475467682,
    "learning_rate": 1.629975623604074e-05,
    "timestep_mean": 545.875,
    "epsilon_norm": 8.165486335754395,
    "prediction_norm": 6.579569339752197,
    "timestamp": "2025-07-14T12:34:39.790022"
  },
  {
    "epoch": 1837,
    "batch_idx": 96,
    "loss": 0.20243865251541138,
    "learning_rate": 1.629975623604074e-05,
    "timestep_mean": 440.0,
    "epsilon_norm": 8.274802207946777,
    "prediction_norm": 6.910260200500488,
    "timestamp": "2025-07-14T12:34:39.795272"
  },
  {
    "epoch": 1837,
    "batch_idx": 128,
    "loss": 0.22395209968090057,
    "learning_rate": 1.629975623604074e-05,
    "timestep_mean": 507.84375,
    "epsilon_norm": 7.687923431396484,
    "prediction_norm": 6.819143295288086,
    "timestamp": "2025-07-14T12:34:39.800657"
  },
  {
    "epoch": 1837,
    "batch_idx": 160,
    "loss": 0.23033612966537476,
    "learning_rate": 1.629975623604074e-05,
    "timestep_mean": 526.9375,
    "epsilon_norm": 7.829061508178711,
    "prediction_norm": 7.2635297775268555,
    "timestamp": "2025-07-14T12:34:39.805783"
  },
  {
    "epoch": 1837,
    "batch_idx": 192,
    "loss": 0.40752679109573364,
    "learning_rate": 1.629975623604074e-05,
    "timestep_mean": 392.5,
    "epsilon_norm": 4.125617027282715,
    "prediction_norm": 3.269503116607666,
    "timestamp": "2025-07-14T12:34:39.811024"
  },
  {
    "epoch": 1838,
    "batch_idx": 0,
    "loss": 0.1672746241092682,
    "learning_rate": 1.6101449835567325e-05,
    "timestep_mean": 469.75,
    "epsilon_norm": 7.279850959777832,
    "prediction_norm": 6.374152660369873,
    "timestamp": "2025-07-14T12:34:39.816177"
  },
  {
    "epoch": 1838,
    "batch_idx": 32,
    "loss": 0.2216479778289795,
    "learning_rate": 1.6101449835567325e-05,
    "timestep_mean": 539.59375,
    "epsilon_norm": 8.695913314819336,
    "prediction_norm": 7.807689666748047,
    "timestamp": "2025-07-14T12:34:39.821496"
  },
  {
    "epoch": 1838,
    "batch_idx": 64,
    "loss": 0.17832911014556885,
    "learning_rate": 1.6101449835567325e-05,
    "timestep_mean": 463.84375,
    "epsilon_norm": 6.839151382446289,
    "prediction_norm": 6.619949817657471,
    "timestamp": "2025-07-14T12:34:39.826661"
  },
  {
    "epoch": 1838,
    "batch_idx": 96,
    "loss": 0.37644779682159424,
    "learning_rate": 1.6101449835567325e-05,
    "timestep_mean": 428.03125,
    "epsilon_norm": 8.14133358001709,
    "prediction_norm": 6.224517822265625,
    "timestamp": "2025-07-14T12:34:39.831902"
  },
  {
    "epoch": 1838,
    "batch_idx": 128,
    "loss": 0.30299389362335205,
    "learning_rate": 1.6101449835567325e-05,
    "timestep_mean": 538.40625,
    "epsilon_norm": 8.743184089660645,
    "prediction_norm": 7.424258232116699,
    "timestamp": "2025-07-14T12:34:39.837182"
  },
  {
    "epoch": 1838,
    "batch_idx": 160,
    "loss": 0.19431401789188385,
    "learning_rate": 1.6101449835567325e-05,
    "timestep_mean": 517.59375,
    "epsilon_norm": 6.973336219787598,
    "prediction_norm": 6.399629592895508,
    "timestamp": "2025-07-14T12:34:39.842476"
  },
  {
    "epoch": 1838,
    "batch_idx": 192,
    "loss": 0.4016571640968323,
    "learning_rate": 1.6101449835567325e-05,
    "timestep_mean": 569.75,
    "epsilon_norm": 6.373058319091797,
    "prediction_norm": 5.6090521812438965,
    "timestamp": "2025-07-14T12:34:39.847713"
  },
  {
    "epoch": 1839,
    "batch_idx": 0,
    "loss": 0.22115281224250793,
    "learning_rate": 1.5904337406663358e-05,
    "timestep_mean": 539.53125,
    "epsilon_norm": 7.854257106781006,
    "prediction_norm": 6.717473983764648,
    "timestamp": "2025-07-14T12:34:39.852963"
  },
  {
    "epoch": 1839,
    "batch_idx": 32,
    "loss": 0.32762083411216736,
    "learning_rate": 1.5904337406663358e-05,
    "timestep_mean": 490.21875,
    "epsilon_norm": 6.57549524307251,
    "prediction_norm": 5.178950786590576,
    "timestamp": "2025-07-14T12:34:39.858078"
  },
  {
    "epoch": 1839,
    "batch_idx": 64,
    "loss": 0.2565055191516876,
    "learning_rate": 1.5904337406663358e-05,
    "timestep_mean": 538.40625,
    "epsilon_norm": 8.064011573791504,
    "prediction_norm": 6.562846660614014,
    "timestamp": "2025-07-14T12:34:39.863421"
  },
  {
    "epoch": 1839,
    "batch_idx": 96,
    "loss": 0.3164539039134979,
    "learning_rate": 1.5904337406663358e-05,
    "timestep_mean": 483.03125,
    "epsilon_norm": 9.786361694335938,
    "prediction_norm": 8.01335334777832,
    "timestamp": "2025-07-14T12:34:39.868636"
  },
  {
    "epoch": 1839,
    "batch_idx": 128,
    "loss": 0.32005298137664795,
    "learning_rate": 1.5904337406663358e-05,
    "timestep_mean": 518.46875,
    "epsilon_norm": 8.927485466003418,
    "prediction_norm": 7.214690685272217,
    "timestamp": "2025-07-14T12:34:39.873940"
  },
  {
    "epoch": 1839,
    "batch_idx": 160,
    "loss": 0.41837334632873535,
    "learning_rate": 1.5904337406663358e-05,
    "timestep_mean": 491.90625,
    "epsilon_norm": 8.689072608947754,
    "prediction_norm": 6.8638505935668945,
    "timestamp": "2025-07-14T12:34:39.879194"
  },
  {
    "epoch": 1839,
    "batch_idx": 192,
    "loss": 0.11999984085559845,
    "learning_rate": 1.5904337406663358e-05,
    "timestep_mean": 538.5,
    "epsilon_norm": 3.6984615325927734,
    "prediction_norm": 3.700100898742676,
    "timestamp": "2025-07-14T12:34:39.884512"
  },
  {
    "epoch": 1840,
    "batch_idx": 0,
    "loss": 0.1783616840839386,
    "learning_rate": 1.570841943568446e-05,
    "timestep_mean": 467.625,
    "epsilon_norm": 7.608295917510986,
    "prediction_norm": 6.918414115905762,
    "timestamp": "2025-07-14T12:34:39.889823"
  },
  {
    "epoch": 1840,
    "batch_idx": 32,
    "loss": 0.1512974500656128,
    "learning_rate": 1.570841943568446e-05,
    "timestep_mean": 582.78125,
    "epsilon_norm": 7.2115044593811035,
    "prediction_norm": 6.474770545959473,
    "timestamp": "2025-07-14T12:34:39.895137"
  },
  {
    "epoch": 1840,
    "batch_idx": 64,
    "loss": 0.3155997097492218,
    "learning_rate": 1.570841943568446e-05,
    "timestep_mean": 463.65625,
    "epsilon_norm": 8.314617156982422,
    "prediction_norm": 6.114720821380615,
    "timestamp": "2025-07-14T12:34:39.900398"
  },
  {
    "epoch": 1840,
    "batch_idx": 96,
    "loss": 0.274077445268631,
    "learning_rate": 1.570841943568446e-05,
    "timestep_mean": 482.15625,
    "epsilon_norm": 7.382990837097168,
    "prediction_norm": 6.700949192047119,
    "timestamp": "2025-07-14T12:34:39.905686"
  },
  {
    "epoch": 1840,
    "batch_idx": 128,
    "loss": 0.310859739780426,
    "learning_rate": 1.570841943568446e-05,
    "timestep_mean": 457.0,
    "epsilon_norm": 6.744482040405273,
    "prediction_norm": 5.7798752784729,
    "timestamp": "2025-07-14T12:34:39.910875"
  },
  {
    "epoch": 1840,
    "batch_idx": 160,
    "loss": 0.20558932423591614,
    "learning_rate": 1.570841943568446e-05,
    "timestep_mean": 547.90625,
    "epsilon_norm": 8.207622528076172,
    "prediction_norm": 6.949368476867676,
    "timestamp": "2025-07-14T12:34:39.916095"
  },
  {
    "epoch": 1840,
    "batch_idx": 192,
    "loss": 0.14500108361244202,
    "learning_rate": 1.570841943568446e-05,
    "timestep_mean": 456.875,
    "epsilon_norm": 3.2568697929382324,
    "prediction_norm": 3.0675735473632812,
    "timestamp": "2025-07-14T12:34:39.921320"
  },
  {
    "epoch": 1841,
    "batch_idx": 0,
    "loss": 0.2021985948085785,
    "learning_rate": 1.5513696406038555e-05,
    "timestep_mean": 583.375,
    "epsilon_norm": 7.26146125793457,
    "prediction_norm": 6.600501537322998,
    "timestamp": "2025-07-14T12:34:39.926665"
  },
  {
    "epoch": 1841,
    "batch_idx": 32,
    "loss": 0.15129169821739197,
    "learning_rate": 1.5513696406038555e-05,
    "timestep_mean": 564.6875,
    "epsilon_norm": 7.8199381828308105,
    "prediction_norm": 7.211521625518799,
    "timestamp": "2025-07-14T12:34:39.931968"
  },
  {
    "epoch": 1841,
    "batch_idx": 64,
    "loss": 0.1742081642150879,
    "learning_rate": 1.5513696406038555e-05,
    "timestep_mean": 492.0625,
    "epsilon_norm": 7.232324600219727,
    "prediction_norm": 7.094586372375488,
    "timestamp": "2025-07-14T12:34:39.937229"
  },
  {
    "epoch": 1841,
    "batch_idx": 96,
    "loss": 0.24226996302604675,
    "learning_rate": 1.5513696406038555e-05,
    "timestep_mean": 488.28125,
    "epsilon_norm": 8.801566123962402,
    "prediction_norm": 7.716628551483154,
    "timestamp": "2025-07-14T12:34:39.942796"
  },
  {
    "epoch": 1841,
    "batch_idx": 128,
    "loss": 0.5718216896057129,
    "learning_rate": 1.5513696406038555e-05,
    "timestep_mean": 321.1875,
    "epsilon_norm": 8.168228149414062,
    "prediction_norm": 5.662845611572266,
    "timestamp": "2025-07-14T12:34:39.948039"
  },
  {
    "epoch": 1841,
    "batch_idx": 160,
    "loss": 0.3734092712402344,
    "learning_rate": 1.5513696406038555e-05,
    "timestep_mean": 521.03125,
    "epsilon_norm": 7.918362140655518,
    "prediction_norm": 6.779611587524414,
    "timestamp": "2025-07-14T12:34:39.953439"
  },
  {
    "epoch": 1841,
    "batch_idx": 192,
    "loss": 0.4538104832172394,
    "learning_rate": 1.5513696406038555e-05,
    "timestep_mean": 491.375,
    "epsilon_norm": 4.323103904724121,
    "prediction_norm": 3.4015910625457764,
    "timestamp": "2025-07-14T12:34:39.958773"
  },
  {
    "epoch": 1842,
    "batch_idx": 0,
    "loss": 0.25028568506240845,
    "learning_rate": 1.5320168798185372e-05,
    "timestep_mean": 500.53125,
    "epsilon_norm": 8.591129302978516,
    "prediction_norm": 7.082089900970459,
    "timestamp": "2025-07-14T12:34:39.964044"
  },
  {
    "epoch": 1842,
    "batch_idx": 32,
    "loss": 0.26270753145217896,
    "learning_rate": 1.5320168798185372e-05,
    "timestep_mean": 454.75,
    "epsilon_norm": 6.807377815246582,
    "prediction_norm": 5.96852970123291,
    "timestamp": "2025-07-14T12:34:39.969395"
  },
  {
    "epoch": 1842,
    "batch_idx": 64,
    "loss": 0.35921335220336914,
    "learning_rate": 1.5320168798185372e-05,
    "timestep_mean": 451.9375,
    "epsilon_norm": 9.752140045166016,
    "prediction_norm": 8.187512397766113,
    "timestamp": "2025-07-14T12:34:39.974702"
  },
  {
    "epoch": 1842,
    "batch_idx": 96,
    "loss": 0.11910924315452576,
    "learning_rate": 1.5320168798185372e-05,
    "timestep_mean": 579.0,
    "epsilon_norm": 7.380774021148682,
    "prediction_norm": 6.805235862731934,
    "timestamp": "2025-07-14T12:34:39.980013"
  },
  {
    "epoch": 1842,
    "batch_idx": 128,
    "loss": 0.1581716239452362,
    "learning_rate": 1.5320168798185372e-05,
    "timestep_mean": 474.78125,
    "epsilon_norm": 8.54899787902832,
    "prediction_norm": 7.1415934562683105,
    "timestamp": "2025-07-14T12:34:39.985265"
  },
  {
    "epoch": 1842,
    "batch_idx": 160,
    "loss": 0.1658286303281784,
    "learning_rate": 1.5320168798185372e-05,
    "timestep_mean": 504.09375,
    "epsilon_norm": 7.463433265686035,
    "prediction_norm": 6.594121932983398,
    "timestamp": "2025-07-14T12:34:39.990523"
  },
  {
    "epoch": 1842,
    "batch_idx": 192,
    "loss": 0.34706735610961914,
    "learning_rate": 1.5320168798185372e-05,
    "timestep_mean": 473.25,
    "epsilon_norm": 2.4603822231292725,
    "prediction_norm": 1.7744417190551758,
    "timestamp": "2025-07-14T12:34:39.995836"
  },
  {
    "epoch": 1843,
    "batch_idx": 0,
    "loss": 0.48988088965415955,
    "learning_rate": 1.5127837089635109e-05,
    "timestep_mean": 407.0625,
    "epsilon_norm": 8.086349487304688,
    "prediction_norm": 5.9744486808776855,
    "timestamp": "2025-07-14T12:34:40.001073"
  },
  {
    "epoch": 1843,
    "batch_idx": 32,
    "loss": 0.5911660194396973,
    "learning_rate": 1.5127837089635109e-05,
    "timestep_mean": 438.5,
    "epsilon_norm": 9.820213317871094,
    "prediction_norm": 6.79633092880249,
    "timestamp": "2025-07-14T12:34:40.006294"
  },
  {
    "epoch": 1843,
    "batch_idx": 64,
    "loss": 0.4939981997013092,
    "learning_rate": 1.5127837089635109e-05,
    "timestep_mean": 465.71875,
    "epsilon_norm": 7.482486248016357,
    "prediction_norm": 5.060074329376221,
    "timestamp": "2025-07-14T12:34:40.011632"
  },
  {
    "epoch": 1843,
    "batch_idx": 96,
    "loss": 0.17594321072101593,
    "learning_rate": 1.5127837089635109e-05,
    "timestep_mean": 598.5,
    "epsilon_norm": 7.548342227935791,
    "prediction_norm": 7.148727893829346,
    "timestamp": "2025-07-14T12:34:40.017054"
  },
  {
    "epoch": 1843,
    "batch_idx": 128,
    "loss": 0.2616020441055298,
    "learning_rate": 1.5127837089635109e-05,
    "timestep_mean": 520.71875,
    "epsilon_norm": 7.8456926345825195,
    "prediction_norm": 6.385490417480469,
    "timestamp": "2025-07-14T12:34:40.022377"
  },
  {
    "epoch": 1843,
    "batch_idx": 160,
    "loss": 0.2817718982696533,
    "learning_rate": 1.5127837089635109e-05,
    "timestep_mean": 476.25,
    "epsilon_norm": 7.721299648284912,
    "prediction_norm": 7.065034866333008,
    "timestamp": "2025-07-14T12:34:40.027658"
  },
  {
    "epoch": 1843,
    "batch_idx": 192,
    "loss": 0.1588359922170639,
    "learning_rate": 1.5127837089635109e-05,
    "timestep_mean": 444.25,
    "epsilon_norm": 3.9197964668273926,
    "prediction_norm": 3.276669979095459,
    "timestamp": "2025-07-14T12:34:40.032946"
  },
  {
    "epoch": 1844,
    "batch_idx": 0,
    "loss": 0.23246127367019653,
    "learning_rate": 1.4936701754947101e-05,
    "timestep_mean": 498.0625,
    "epsilon_norm": 7.51112174987793,
    "prediction_norm": 6.8075127601623535,
    "timestamp": "2025-07-14T12:34:40.038244"
  },
  {
    "epoch": 1844,
    "batch_idx": 32,
    "loss": 0.11426734924316406,
    "learning_rate": 1.4936701754947101e-05,
    "timestep_mean": 587.15625,
    "epsilon_norm": 7.859978199005127,
    "prediction_norm": 7.592396259307861,
    "timestamp": "2025-07-14T12:34:40.043362"
  },
  {
    "epoch": 1844,
    "batch_idx": 64,
    "loss": 0.22152908146381378,
    "learning_rate": 1.4936701754947101e-05,
    "timestep_mean": 505.9375,
    "epsilon_norm": 8.699344635009766,
    "prediction_norm": 7.0764875411987305,
    "timestamp": "2025-07-14T12:34:40.048617"
  },
  {
    "epoch": 1844,
    "batch_idx": 96,
    "loss": 0.2302992194890976,
    "learning_rate": 1.4936701754947101e-05,
    "timestep_mean": 522.375,
    "epsilon_norm": 7.371583938598633,
    "prediction_norm": 6.1615447998046875,
    "timestamp": "2025-07-14T12:34:40.053840"
  },
  {
    "epoch": 1844,
    "batch_idx": 128,
    "loss": 0.2432398796081543,
    "learning_rate": 1.4936701754947101e-05,
    "timestep_mean": 566.96875,
    "epsilon_norm": 8.349261283874512,
    "prediction_norm": 6.924952983856201,
    "timestamp": "2025-07-14T12:34:40.059187"
  },
  {
    "epoch": 1844,
    "batch_idx": 160,
    "loss": 0.4252470135688782,
    "learning_rate": 1.4936701754947101e-05,
    "timestep_mean": 440.84375,
    "epsilon_norm": 8.311173439025879,
    "prediction_norm": 6.559817790985107,
    "timestamp": "2025-07-14T12:34:40.064446"
  },
  {
    "epoch": 1844,
    "batch_idx": 192,
    "loss": 0.10843193531036377,
    "learning_rate": 1.4936701754947101e-05,
    "timestep_mean": 493.125,
    "epsilon_norm": 3.225877523422241,
    "prediction_norm": 2.910444736480713,
    "timestamp": "2025-07-14T12:34:40.069739"
  },
  {
    "epoch": 1845,
    "batch_idx": 0,
    "loss": 0.16291889548301697,
    "learning_rate": 1.4746763265728769e-05,
    "timestep_mean": 583.5625,
    "epsilon_norm": 6.473995685577393,
    "prediction_norm": 5.847029209136963,
    "timestamp": "2025-07-14T12:34:40.075191"
  },
  {
    "epoch": 1845,
    "batch_idx": 32,
    "loss": 0.4191757142543793,
    "learning_rate": 1.4746763265728769e-05,
    "timestep_mean": 440.1875,
    "epsilon_norm": 9.725201606750488,
    "prediction_norm": 7.8308234214782715,
    "timestamp": "2025-07-14T12:34:40.080458"
  },
  {
    "epoch": 1845,
    "batch_idx": 64,
    "loss": 0.24987320601940155,
    "learning_rate": 1.4746763265728769e-05,
    "timestep_mean": 582.3125,
    "epsilon_norm": 8.277748107910156,
    "prediction_norm": 6.659970283508301,
    "timestamp": "2025-07-14T12:34:40.085777"
  },
  {
    "epoch": 1845,
    "batch_idx": 96,
    "loss": 0.20681661367416382,
    "learning_rate": 1.4746763265728769e-05,
    "timestep_mean": 477.4375,
    "epsilon_norm": 7.731176376342773,
    "prediction_norm": 7.447965621948242,
    "timestamp": "2025-07-14T12:34:40.090977"
  },
  {
    "epoch": 1845,
    "batch_idx": 128,
    "loss": 0.18707703053951263,
    "learning_rate": 1.4746763265728769e-05,
    "timestep_mean": 519.03125,
    "epsilon_norm": 6.92563009262085,
    "prediction_norm": 6.145568370819092,
    "timestamp": "2025-07-14T12:34:40.096350"
  },
  {
    "epoch": 1845,
    "batch_idx": 160,
    "loss": 0.14726224541664124,
    "learning_rate": 1.4746763265728769e-05,
    "timestep_mean": 578.125,
    "epsilon_norm": 7.49233865737915,
    "prediction_norm": 7.093035697937012,
    "timestamp": "2025-07-14T12:34:40.101495"
  },
  {
    "epoch": 1845,
    "batch_idx": 192,
    "loss": 0.4586954712867737,
    "learning_rate": 1.4746763265728769e-05,
    "timestep_mean": 559.875,
    "epsilon_norm": 3.8076953887939453,
    "prediction_norm": 2.5850908756256104,
    "timestamp": "2025-07-14T12:34:40.106500"
  },
  {
    "epoch": 1846,
    "batch_idx": 0,
    "loss": 0.48977407813072205,
    "learning_rate": 1.4558022090634503e-05,
    "timestep_mean": 484.46875,
    "epsilon_norm": 8.157393455505371,
    "prediction_norm": 6.383056163787842,
    "timestamp": "2025-07-14T12:34:40.111685"
  },
  {
    "epoch": 1846,
    "batch_idx": 32,
    "loss": 0.2874028980731964,
    "learning_rate": 1.4558022090634503e-05,
    "timestep_mean": 539.1875,
    "epsilon_norm": 8.449056625366211,
    "prediction_norm": 6.9465107917785645,
    "timestamp": "2025-07-14T12:34:40.117015"
  },
  {
    "epoch": 1846,
    "batch_idx": 64,
    "loss": 0.1757807731628418,
    "learning_rate": 1.4558022090634503e-05,
    "timestep_mean": 543.0625,
    "epsilon_norm": 7.568446159362793,
    "prediction_norm": 7.221147537231445,
    "timestamp": "2025-07-14T12:34:40.122236"
  },
  {
    "epoch": 1846,
    "batch_idx": 96,
    "loss": 0.2846840023994446,
    "learning_rate": 1.4558022090634503e-05,
    "timestep_mean": 503.8125,
    "epsilon_norm": 8.225943565368652,
    "prediction_norm": 6.781776428222656,
    "timestamp": "2025-07-14T12:34:40.127404"
  },
  {
    "epoch": 1846,
    "batch_idx": 128,
    "loss": 0.37413325905799866,
    "learning_rate": 1.4558022090634503e-05,
    "timestep_mean": 461.1875,
    "epsilon_norm": 8.492115020751953,
    "prediction_norm": 6.521975994110107,
    "timestamp": "2025-07-14T12:34:40.132698"
  },
  {
    "epoch": 1846,
    "batch_idx": 160,
    "loss": 0.34226399660110474,
    "learning_rate": 1.4558022090634503e-05,
    "timestep_mean": 479.15625,
    "epsilon_norm": 8.46487045288086,
    "prediction_norm": 7.228913307189941,
    "timestamp": "2025-07-14T12:34:40.137918"
  },
  {
    "epoch": 1846,
    "batch_idx": 192,
    "loss": 0.6766363978385925,
    "learning_rate": 1.4558022090634503e-05,
    "timestep_mean": 450.0,
    "epsilon_norm": 5.222317218780518,
    "prediction_norm": 4.45874547958374,
    "timestamp": "2025-07-14T12:34:40.142940"
  },
  {
    "epoch": 1847,
    "batch_idx": 0,
    "loss": 0.35409611463546753,
    "learning_rate": 1.4370478695364397e-05,
    "timestep_mean": 612.28125,
    "epsilon_norm": 8.509849548339844,
    "prediction_norm": 6.972259998321533,
    "timestamp": "2025-07-14T12:34:40.148321"
  },
  {
    "epoch": 1847,
    "batch_idx": 32,
    "loss": 0.20697632431983948,
    "learning_rate": 1.4370478695364397e-05,
    "timestep_mean": 543.0625,
    "epsilon_norm": 8.780282974243164,
    "prediction_norm": 7.795809268951416,
    "timestamp": "2025-07-14T12:34:40.153526"
  },
  {
    "epoch": 1847,
    "batch_idx": 64,
    "loss": 0.2749694287776947,
    "learning_rate": 1.4370478695364397e-05,
    "timestep_mean": 473.0,
    "epsilon_norm": 7.887296199798584,
    "prediction_norm": 6.401692867279053,
    "timestamp": "2025-07-14T12:34:40.158969"
  },
  {
    "epoch": 1847,
    "batch_idx": 96,
    "loss": 0.19837142527103424,
    "learning_rate": 1.4370478695364397e-05,
    "timestep_mean": 480.28125,
    "epsilon_norm": 8.283899307250977,
    "prediction_norm": 7.7280192375183105,
    "timestamp": "2025-07-14T12:34:40.164083"
  },
  {
    "epoch": 1847,
    "batch_idx": 128,
    "loss": 0.18903452157974243,
    "learning_rate": 1.4370478695364397e-05,
    "timestep_mean": 505.34375,
    "epsilon_norm": 7.913474082946777,
    "prediction_norm": 6.853883266448975,
    "timestamp": "2025-07-14T12:34:40.169413"
  },
  {
    "epoch": 1847,
    "batch_idx": 160,
    "loss": 0.2246004045009613,
    "learning_rate": 1.4370478695364397e-05,
    "timestep_mean": 520.6875,
    "epsilon_norm": 7.440948486328125,
    "prediction_norm": 6.407190799713135,
    "timestamp": "2025-07-14T12:34:40.174982"
  },
  {
    "epoch": 1847,
    "batch_idx": 192,
    "loss": 0.28584325313568115,
    "learning_rate": 1.4370478695364397e-05,
    "timestep_mean": 552.375,
    "epsilon_norm": 4.323765754699707,
    "prediction_norm": 3.8653225898742676,
    "timestamp": "2025-07-14T12:34:40.180162"
  },
  {
    "epoch": 1848,
    "batch_idx": 0,
    "loss": 0.17876771092414856,
    "learning_rate": 1.4184133542663014e-05,
    "timestep_mean": 611.125,
    "epsilon_norm": 7.24934720993042,
    "prediction_norm": 6.470330715179443,
    "timestamp": "2025-07-14T12:34:40.185179"
  },
  {
    "epoch": 1848,
    "batch_idx": 32,
    "loss": 0.43242329359054565,
    "learning_rate": 1.4184133542663014e-05,
    "timestep_mean": 433.40625,
    "epsilon_norm": 8.355815887451172,
    "prediction_norm": 6.318471431732178,
    "timestamp": "2025-07-14T12:34:40.190361"
  },
  {
    "epoch": 1848,
    "batch_idx": 64,
    "loss": 0.23762811720371246,
    "learning_rate": 1.4184133542663014e-05,
    "timestep_mean": 506.53125,
    "epsilon_norm": 8.980470657348633,
    "prediction_norm": 8.329950332641602,
    "timestamp": "2025-07-14T12:34:40.195326"
  },
  {
    "epoch": 1848,
    "batch_idx": 96,
    "loss": 0.23850885033607483,
    "learning_rate": 1.4184133542663014e-05,
    "timestep_mean": 486.46875,
    "epsilon_norm": 7.968020439147949,
    "prediction_norm": 6.779237747192383,
    "timestamp": "2025-07-14T12:34:40.200364"
  },
  {
    "epoch": 1848,
    "batch_idx": 128,
    "loss": 0.2804689407348633,
    "learning_rate": 1.4184133542663014e-05,
    "timestep_mean": 492.28125,
    "epsilon_norm": 7.361661434173584,
    "prediction_norm": 6.682168483734131,
    "timestamp": "2025-07-14T12:34:40.205380"
  },
  {
    "epoch": 1848,
    "batch_idx": 160,
    "loss": 0.30746763944625854,
    "learning_rate": 1.4184133542663014e-05,
    "timestep_mean": 541.6875,
    "epsilon_norm": 8.116388320922852,
    "prediction_norm": 6.453005313873291,
    "timestamp": "2025-07-14T12:34:40.210338"
  },
  {
    "epoch": 1848,
    "batch_idx": 192,
    "loss": 0.11419882625341415,
    "learning_rate": 1.4184133542663014e-05,
    "timestep_mean": 618.625,
    "epsilon_norm": 3.452004909515381,
    "prediction_norm": 3.204763174057007,
    "timestamp": "2025-07-14T12:34:40.215278"
  },
  {
    "epoch": 1849,
    "batch_idx": 0,
    "loss": 0.3547295928001404,
    "learning_rate": 1.3998987092318726e-05,
    "timestep_mean": 454.6875,
    "epsilon_norm": 8.109295845031738,
    "prediction_norm": 7.1687822341918945,
    "timestamp": "2025-07-14T12:34:40.220243"
  },
  {
    "epoch": 1849,
    "batch_idx": 32,
    "loss": 0.20306959748268127,
    "learning_rate": 1.3998987092318726e-05,
    "timestep_mean": 546.3125,
    "epsilon_norm": 7.829097270965576,
    "prediction_norm": 6.77142333984375,
    "timestamp": "2025-07-14T12:34:40.225278"
  },
  {
    "epoch": 1849,
    "batch_idx": 64,
    "loss": 0.3122509717941284,
    "learning_rate": 1.3998987092318726e-05,
    "timestep_mean": 510.21875,
    "epsilon_norm": 7.039126396179199,
    "prediction_norm": 5.979586601257324,
    "timestamp": "2025-07-14T12:34:40.230188"
  },
  {
    "epoch": 1849,
    "batch_idx": 96,
    "loss": 0.16788265109062195,
    "learning_rate": 1.3998987092318726e-05,
    "timestep_mean": 480.625,
    "epsilon_norm": 7.037276268005371,
    "prediction_norm": 6.194371700286865,
    "timestamp": "2025-07-14T12:34:40.235112"
  },
  {
    "epoch": 1849,
    "batch_idx": 128,
    "loss": 0.2558453679084778,
    "learning_rate": 1.3998987092318726e-05,
    "timestep_mean": 456.53125,
    "epsilon_norm": 6.645073890686035,
    "prediction_norm": 5.368658065795898,
    "timestamp": "2025-07-14T12:34:40.240129"
  },
  {
    "epoch": 1849,
    "batch_idx": 160,
    "loss": 0.2862309217453003,
    "learning_rate": 1.3998987092318726e-05,
    "timestep_mean": 540.5,
    "epsilon_norm": 8.240646362304688,
    "prediction_norm": 7.057536602020264,
    "timestamp": "2025-07-14T12:34:40.245069"
  },
  {
    "epoch": 1849,
    "batch_idx": 192,
    "loss": 0.11169988662004471,
    "learning_rate": 1.3998987092318726e-05,
    "timestep_mean": 574.75,
    "epsilon_norm": 3.236114263534546,
    "prediction_norm": 2.8552143573760986,
    "timestamp": "2025-07-14T12:34:40.249931"
  },
  {
    "epoch": 1850,
    "batch_idx": 0,
    "loss": 0.5142260193824768,
    "learning_rate": 1.3815039801161721e-05,
    "timestep_mean": 422.96875,
    "epsilon_norm": 9.489538192749023,
    "prediction_norm": 7.139902591705322,
    "timestamp": "2025-07-14T12:34:40.254952"
  },
  {
    "epoch": 1850,
    "batch_idx": 32,
    "loss": 0.3800816237926483,
    "learning_rate": 1.3815039801161721e-05,
    "timestep_mean": 528.75,
    "epsilon_norm": 9.55345344543457,
    "prediction_norm": 8.580822944641113,
    "timestamp": "2025-07-14T12:34:40.259793"
  },
  {
    "epoch": 1850,
    "batch_idx": 64,
    "loss": 0.43076545000076294,
    "learning_rate": 1.3815039801161721e-05,
    "timestep_mean": 471.21875,
    "epsilon_norm": 8.75127124786377,
    "prediction_norm": 7.090938568115234,
    "timestamp": "2025-07-14T12:34:40.264664"
  },
  {
    "epoch": 1850,
    "batch_idx": 96,
    "loss": 0.3682234287261963,
    "learning_rate": 1.3815039801161721e-05,
    "timestep_mean": 469.21875,
    "epsilon_norm": 8.80215072631836,
    "prediction_norm": 6.9868693351745605,
    "timestamp": "2025-07-14T12:34:40.269547"
  },
  {
    "epoch": 1850,
    "batch_idx": 128,
    "loss": 0.17560473084449768,
    "learning_rate": 1.3815039801161721e-05,
    "timestep_mean": 423.6875,
    "epsilon_norm": 7.049757480621338,
    "prediction_norm": 6.4115986824035645,
    "timestamp": "2025-07-14T12:34:40.274438"
  },
  {
    "epoch": 1850,
    "batch_idx": 160,
    "loss": 0.49736958742141724,
    "learning_rate": 1.3815039801161721e-05,
    "timestep_mean": 452.6875,
    "epsilon_norm": 8.787586212158203,
    "prediction_norm": 6.875404357910156,
    "timestamp": "2025-07-14T12:34:40.279435"
  },
  {
    "epoch": 1850,
    "batch_idx": 192,
    "loss": 0.030266664922237396,
    "learning_rate": 1.3815039801161721e-05,
    "timestep_mean": 677.625,
    "epsilon_norm": 3.805742025375366,
    "prediction_norm": 3.8248541355133057,
    "timestamp": "2025-07-14T12:34:40.284457"
  },
  {
    "epoch": 1851,
    "batch_idx": 0,
    "loss": 0.5923280119895935,
    "learning_rate": 1.3632292123063772e-05,
    "timestep_mean": 466.4375,
    "epsilon_norm": 9.235416412353516,
    "prediction_norm": 7.089642524719238,
    "timestamp": "2025-07-14T12:34:40.289395"
  },
  {
    "epoch": 1851,
    "batch_idx": 32,
    "loss": 0.34772735834121704,
    "learning_rate": 1.3632292123063772e-05,
    "timestep_mean": 383.5625,
    "epsilon_norm": 7.381259918212891,
    "prediction_norm": 5.863587379455566,
    "timestamp": "2025-07-14T12:34:40.294349"
  },
  {
    "epoch": 1851,
    "batch_idx": 64,
    "loss": 0.4214087724685669,
    "learning_rate": 1.3632292123063772e-05,
    "timestep_mean": 505.03125,
    "epsilon_norm": 8.259553909301758,
    "prediction_norm": 6.447025775909424,
    "timestamp": "2025-07-14T12:34:40.299229"
  },
  {
    "epoch": 1851,
    "batch_idx": 96,
    "loss": 0.13625553250312805,
    "learning_rate": 1.3632292123063772e-05,
    "timestep_mean": 479.4375,
    "epsilon_norm": 7.229930400848389,
    "prediction_norm": 6.954198360443115,
    "timestamp": "2025-07-14T12:34:40.304177"
  },
  {
    "epoch": 1851,
    "batch_idx": 128,
    "loss": 0.36687415838241577,
    "learning_rate": 1.3632292123063772e-05,
    "timestep_mean": 457.75,
    "epsilon_norm": 8.471661567687988,
    "prediction_norm": 6.730003833770752,
    "timestamp": "2025-07-14T12:34:40.309235"
  },
  {
    "epoch": 1851,
    "batch_idx": 160,
    "loss": 0.36917126178741455,
    "learning_rate": 1.3632292123063772e-05,
    "timestep_mean": 498.15625,
    "epsilon_norm": 9.28125,
    "prediction_norm": 7.861898899078369,
    "timestamp": "2025-07-14T12:34:40.314155"
  },
  {
    "epoch": 1851,
    "batch_idx": 192,
    "loss": 0.1977156102657318,
    "learning_rate": 1.3632292123063772e-05,
    "timestep_mean": 444.25,
    "epsilon_norm": 4.317575454711914,
    "prediction_norm": 3.559039831161499,
    "timestamp": "2025-07-14T12:34:40.319107"
  },
  {
    "epoch": 1852,
    "batch_idx": 0,
    "loss": 0.521318793296814,
    "learning_rate": 1.3450744508936741e-05,
    "timestep_mean": 561.46875,
    "epsilon_norm": 8.217877388000488,
    "prediction_norm": 6.62260627746582,
    "timestamp": "2025-07-14T12:34:40.324007"
  },
  {
    "epoch": 1852,
    "batch_idx": 32,
    "loss": 0.42363157868385315,
    "learning_rate": 1.3450744508936741e-05,
    "timestep_mean": 526.4375,
    "epsilon_norm": 7.833667755126953,
    "prediction_norm": 5.91969633102417,
    "timestamp": "2025-07-14T12:34:40.328946"
  },
  {
    "epoch": 1852,
    "batch_idx": 64,
    "loss": 0.33584511280059814,
    "learning_rate": 1.3450744508936741e-05,
    "timestep_mean": 453.75,
    "epsilon_norm": 7.252435684204102,
    "prediction_norm": 4.914702892303467,
    "timestamp": "2025-07-14T12:34:40.334073"
  },
  {
    "epoch": 1852,
    "batch_idx": 96,
    "loss": 0.3960335850715637,
    "learning_rate": 1.3450744508936741e-05,
    "timestep_mean": 436.28125,
    "epsilon_norm": 8.691957473754883,
    "prediction_norm": 7.042993068695068,
    "timestamp": "2025-07-14T12:34:40.339162"
  },
  {
    "epoch": 1852,
    "batch_idx": 128,
    "loss": 0.13482198119163513,
    "learning_rate": 1.3450744508936741e-05,
    "timestep_mean": 512.6875,
    "epsilon_norm": 6.872894287109375,
    "prediction_norm": 6.0427565574646,
    "timestamp": "2025-07-14T12:34:40.344226"
  },
  {
    "epoch": 1852,
    "batch_idx": 160,
    "loss": 0.19192245602607727,
    "learning_rate": 1.3450744508936741e-05,
    "timestep_mean": 529.84375,
    "epsilon_norm": 7.782124996185303,
    "prediction_norm": 6.7707061767578125,
    "timestamp": "2025-07-14T12:34:40.349252"
  },
  {
    "epoch": 1852,
    "batch_idx": 192,
    "loss": 0.35026815533638,
    "learning_rate": 1.3450744508936741e-05,
    "timestep_mean": 616.375,
    "epsilon_norm": 3.2758750915527344,
    "prediction_norm": 2.6454272270202637,
    "timestamp": "2025-07-14T12:34:40.354236"
  },
  {
    "epoch": 1853,
    "batch_idx": 0,
    "loss": 0.19387289881706238,
    "learning_rate": 1.3270397406731138e-05,
    "timestep_mean": 525.21875,
    "epsilon_norm": 7.014026641845703,
    "prediction_norm": 5.497457027435303,
    "timestamp": "2025-07-14T12:34:40.359246"
  },
  {
    "epoch": 1853,
    "batch_idx": 32,
    "loss": 0.2386474609375,
    "learning_rate": 1.3270397406731138e-05,
    "timestep_mean": 470.625,
    "epsilon_norm": 8.320005416870117,
    "prediction_norm": 7.246773719787598,
    "timestamp": "2025-07-14T12:34:40.364122"
  },
  {
    "epoch": 1853,
    "batch_idx": 64,
    "loss": 0.17060858011245728,
    "learning_rate": 1.3270397406731138e-05,
    "timestep_mean": 550.71875,
    "epsilon_norm": 8.11404800415039,
    "prediction_norm": 7.249831676483154,
    "timestamp": "2025-07-14T12:34:40.369117"
  },
  {
    "epoch": 1853,
    "batch_idx": 96,
    "loss": 0.19742995500564575,
    "learning_rate": 1.3270397406731138e-05,
    "timestep_mean": 514.9375,
    "epsilon_norm": 7.657423973083496,
    "prediction_norm": 6.954739093780518,
    "timestamp": "2025-07-14T12:34:40.374082"
  },
  {
    "epoch": 1853,
    "batch_idx": 128,
    "loss": 0.3115285038948059,
    "learning_rate": 1.3270397406731138e-05,
    "timestep_mean": 415.0,
    "epsilon_norm": 7.1525492668151855,
    "prediction_norm": 6.290431976318359,
    "timestamp": "2025-07-14T12:34:40.379039"
  },
  {
    "epoch": 1853,
    "batch_idx": 160,
    "loss": 0.40476372838020325,
    "learning_rate": 1.3270397406731138e-05,
    "timestep_mean": 463.90625,
    "epsilon_norm": 8.949134826660156,
    "prediction_norm": 7.416141033172607,
    "timestamp": "2025-07-14T12:34:40.384105"
  },
  {
    "epoch": 1853,
    "batch_idx": 192,
    "loss": 0.11058773100376129,
    "learning_rate": 1.3270397406731138e-05,
    "timestep_mean": 526.875,
    "epsilon_norm": 2.996636152267456,
    "prediction_norm": 2.539168357849121,
    "timestamp": "2025-07-14T12:34:40.389044"
  },
  {
    "epoch": 1854,
    "batch_idx": 0,
    "loss": 0.23972231149673462,
    "learning_rate": 1.309125126143562e-05,
    "timestep_mean": 574.84375,
    "epsilon_norm": 8.587748527526855,
    "prediction_norm": 8.084545135498047,
    "timestamp": "2025-07-14T12:34:40.394006"
  },
  {
    "epoch": 1854,
    "batch_idx": 32,
    "loss": 0.08003807067871094,
    "learning_rate": 1.309125126143562e-05,
    "timestep_mean": 567.125,
    "epsilon_norm": 7.5215373039245605,
    "prediction_norm": 6.981872081756592,
    "timestamp": "2025-07-14T12:34:40.398852"
  },
  {
    "epoch": 1854,
    "batch_idx": 64,
    "loss": 0.24457938969135284,
    "learning_rate": 1.309125126143562e-05,
    "timestep_mean": 551.71875,
    "epsilon_norm": 7.783276557922363,
    "prediction_norm": 6.253891468048096,
    "timestamp": "2025-07-14T12:34:40.403753"
  },
  {
    "epoch": 1854,
    "batch_idx": 96,
    "loss": 0.4717179536819458,
    "learning_rate": 1.309125126143562e-05,
    "timestep_mean": 528.84375,
    "epsilon_norm": 9.464293479919434,
    "prediction_norm": 7.359821319580078,
    "timestamp": "2025-07-14T12:34:40.408684"
  },
  {
    "epoch": 1854,
    "batch_idx": 128,
    "loss": 0.17080722749233246,
    "learning_rate": 1.309125126143562e-05,
    "timestep_mean": 513.875,
    "epsilon_norm": 7.98799991607666,
    "prediction_norm": 7.035180568695068,
    "timestamp": "2025-07-14T12:34:40.413626"
  },
  {
    "epoch": 1854,
    "batch_idx": 160,
    "loss": 0.33840152621269226,
    "learning_rate": 1.309125126143562e-05,
    "timestep_mean": 522.53125,
    "epsilon_norm": 7.9477620124816895,
    "prediction_norm": 6.4270219802856445,
    "timestamp": "2025-07-14T12:34:40.418971"
  },
  {
    "epoch": 1854,
    "batch_idx": 192,
    "loss": 0.06789811700582504,
    "learning_rate": 1.309125126143562e-05,
    "timestep_mean": 609.75,
    "epsilon_norm": 3.5524051189422607,
    "prediction_norm": 3.48007869720459,
    "timestamp": "2025-07-14T12:34:40.424582"
  },
  {
    "epoch": 1855,
    "batch_idx": 0,
    "loss": 0.35838305950164795,
    "learning_rate": 1.2913306515075328e-05,
    "timestep_mean": 557.5,
    "epsilon_norm": 9.183040618896484,
    "prediction_norm": 7.892613410949707,
    "timestamp": "2025-07-14T12:34:40.430328"
  },
  {
    "epoch": 1855,
    "batch_idx": 32,
    "loss": 0.2019961178302765,
    "learning_rate": 1.2913306515075328e-05,
    "timestep_mean": 470.84375,
    "epsilon_norm": 8.958819389343262,
    "prediction_norm": 7.865322589874268,
    "timestamp": "2025-07-14T12:34:40.435622"
  },
  {
    "epoch": 1855,
    "batch_idx": 64,
    "loss": 0.2132566273212433,
    "learning_rate": 1.2913306515075328e-05,
    "timestep_mean": 531.9375,
    "epsilon_norm": 9.114502906799316,
    "prediction_norm": 8.0271635055542,
    "timestamp": "2025-07-14T12:34:40.440896"
  },
  {
    "epoch": 1855,
    "batch_idx": 96,
    "loss": 0.23612675070762634,
    "learning_rate": 1.2913306515075328e-05,
    "timestep_mean": 570.71875,
    "epsilon_norm": 8.276302337646484,
    "prediction_norm": 7.613393306732178,
    "timestamp": "2025-07-14T12:34:40.446190"
  },
  {
    "epoch": 1855,
    "batch_idx": 128,
    "loss": 0.3218769431114197,
    "learning_rate": 1.2913306515075328e-05,
    "timestep_mean": 489.40625,
    "epsilon_norm": 7.945930004119873,
    "prediction_norm": 6.202626705169678,
    "timestamp": "2025-07-14T12:34:40.451480"
  },
  {
    "epoch": 1855,
    "batch_idx": 160,
    "loss": 0.1353870928287506,
    "learning_rate": 1.2913306515075328e-05,
    "timestep_mean": 535.4375,
    "epsilon_norm": 6.744456768035889,
    "prediction_norm": 6.057394504547119,
    "timestamp": "2025-07-14T12:34:40.456758"
  },
  {
    "epoch": 1855,
    "batch_idx": 192,
    "loss": 0.14265835285186768,
    "learning_rate": 1.2913306515075328e-05,
    "timestep_mean": 599.25,
    "epsilon_norm": 3.0843772888183594,
    "prediction_norm": 3.3463501930236816,
    "timestamp": "2025-07-14T12:34:40.461944"
  },
  {
    "epoch": 1856,
    "batch_idx": 0,
    "loss": 0.2656134068965912,
    "learning_rate": 1.2736563606711382e-05,
    "timestep_mean": 533.21875,
    "epsilon_norm": 7.656799793243408,
    "prediction_norm": 6.603724002838135,
    "timestamp": "2025-07-14T12:34:40.467372"
  },
  {
    "epoch": 1856,
    "batch_idx": 32,
    "loss": 0.16116583347320557,
    "learning_rate": 1.2736563606711382e-05,
    "timestep_mean": 513.4375,
    "epsilon_norm": 8.743884086608887,
    "prediction_norm": 8.099438667297363,
    "timestamp": "2025-07-14T12:34:40.472627"
  },
  {
    "epoch": 1856,
    "batch_idx": 64,
    "loss": 0.20623263716697693,
    "learning_rate": 1.2736563606711382e-05,
    "timestep_mean": 475.65625,
    "epsilon_norm": 7.852985858917236,
    "prediction_norm": 6.541726589202881,
    "timestamp": "2025-07-14T12:34:40.477859"
  },
  {
    "epoch": 1856,
    "batch_idx": 96,
    "loss": 0.12897701561450958,
    "learning_rate": 1.2736563606711382e-05,
    "timestep_mean": 517.4375,
    "epsilon_norm": 6.687308311462402,
    "prediction_norm": 6.503034591674805,
    "timestamp": "2025-07-14T12:34:40.483004"
  },
  {
    "epoch": 1856,
    "batch_idx": 128,
    "loss": 0.2584267854690552,
    "learning_rate": 1.2736563606711382e-05,
    "timestep_mean": 466.09375,
    "epsilon_norm": 6.942279815673828,
    "prediction_norm": 5.655381202697754,
    "timestamp": "2025-07-14T12:34:40.488320"
  },
  {
    "epoch": 1856,
    "batch_idx": 160,
    "loss": 0.3921971023082733,
    "learning_rate": 1.2736563606711382e-05,
    "timestep_mean": 451.25,
    "epsilon_norm": 7.865533828735352,
    "prediction_norm": 6.426995277404785,
    "timestamp": "2025-07-14T12:34:40.493532"
  },
  {
    "epoch": 1856,
    "batch_idx": 192,
    "loss": 0.10044384002685547,
    "learning_rate": 1.2736563606711382e-05,
    "timestep_mean": 555.625,
    "epsilon_norm": 4.607554912567139,
    "prediction_norm": 4.533657073974609,
    "timestamp": "2025-07-14T12:34:40.498750"
  },
  {
    "epoch": 1857,
    "batch_idx": 0,
    "loss": 0.27116942405700684,
    "learning_rate": 1.256102297243944e-05,
    "timestep_mean": 548.5,
    "epsilon_norm": 7.274384021759033,
    "prediction_norm": 6.186767101287842,
    "timestamp": "2025-07-14T12:34:40.503965"
  },
  {
    "epoch": 1857,
    "batch_idx": 32,
    "loss": 0.1769106388092041,
    "learning_rate": 1.256102297243944e-05,
    "timestep_mean": 512.25,
    "epsilon_norm": 6.796933174133301,
    "prediction_norm": 6.017138481140137,
    "timestamp": "2025-07-14T12:34:40.509178"
  },
  {
    "epoch": 1857,
    "batch_idx": 64,
    "loss": 0.1326829493045807,
    "learning_rate": 1.256102297243944e-05,
    "timestep_mean": 551.75,
    "epsilon_norm": 7.180593013763428,
    "prediction_norm": 6.342301368713379,
    "timestamp": "2025-07-14T12:34:40.514495"
  },
  {
    "epoch": 1857,
    "batch_idx": 96,
    "loss": 0.33176136016845703,
    "learning_rate": 1.256102297243944e-05,
    "timestep_mean": 462.90625,
    "epsilon_norm": 7.87839412689209,
    "prediction_norm": 6.413305282592773,
    "timestamp": "2025-07-14T12:34:40.519729"
  },
  {
    "epoch": 1857,
    "batch_idx": 128,
    "loss": 0.16246867179870605,
    "learning_rate": 1.256102297243944e-05,
    "timestep_mean": 566.125,
    "epsilon_norm": 7.37237548828125,
    "prediction_norm": 6.334380149841309,
    "timestamp": "2025-07-14T12:34:40.524924"
  },
  {
    "epoch": 1857,
    "batch_idx": 160,
    "loss": 0.23994717001914978,
    "learning_rate": 1.256102297243944e-05,
    "timestep_mean": 528.46875,
    "epsilon_norm": 7.709386825561523,
    "prediction_norm": 6.544231414794922,
    "timestamp": "2025-07-14T12:34:40.530044"
  },
  {
    "epoch": 1857,
    "batch_idx": 192,
    "loss": 0.0556909516453743,
    "learning_rate": 1.256102297243944e-05,
    "timestep_mean": 686.25,
    "epsilon_norm": 3.97715425491333,
    "prediction_norm": 3.972186803817749,
    "timestamp": "2025-07-14T12:34:40.535040"
  },
  {
    "epoch": 1858,
    "batch_idx": 0,
    "loss": 0.22523458302021027,
    "learning_rate": 1.2386685045388313e-05,
    "timestep_mean": 522.03125,
    "epsilon_norm": 7.361072063446045,
    "prediction_norm": 6.2014970779418945,
    "timestamp": "2025-07-14T12:34:40.539963"
  },
  {
    "epoch": 1858,
    "batch_idx": 32,
    "loss": 0.3843876123428345,
    "learning_rate": 1.2386685045388313e-05,
    "timestep_mean": 459.03125,
    "epsilon_norm": 8.653687477111816,
    "prediction_norm": 6.491726875305176,
    "timestamp": "2025-07-14T12:34:40.544889"
  },
  {
    "epoch": 1858,
    "batch_idx": 64,
    "loss": 0.20801445841789246,
    "learning_rate": 1.2386685045388313e-05,
    "timestep_mean": 541.03125,
    "epsilon_norm": 7.6117329597473145,
    "prediction_norm": 6.8641357421875,
    "timestamp": "2025-07-14T12:34:40.549952"
  },
  {
    "epoch": 1858,
    "batch_idx": 96,
    "loss": 0.28885987401008606,
    "learning_rate": 1.2386685045388313e-05,
    "timestep_mean": 441.40625,
    "epsilon_norm": 7.5256781578063965,
    "prediction_norm": 6.725809574127197,
    "timestamp": "2025-07-14T12:34:40.555113"
  },
  {
    "epoch": 1858,
    "batch_idx": 128,
    "loss": 0.2689962387084961,
    "learning_rate": 1.2386685045388313e-05,
    "timestep_mean": 500.34375,
    "epsilon_norm": 7.144624710083008,
    "prediction_norm": 5.3694047927856445,
    "timestamp": "2025-07-14T12:34:40.560364"
  },
  {
    "epoch": 1858,
    "batch_idx": 160,
    "loss": 0.21438166499137878,
    "learning_rate": 1.2386685045388313e-05,
    "timestep_mean": 508.84375,
    "epsilon_norm": 8.543875694274902,
    "prediction_norm": 7.331977367401123,
    "timestamp": "2025-07-14T12:34:40.565333"
  },
  {
    "epoch": 1858,
    "batch_idx": 192,
    "loss": 0.25594455003738403,
    "learning_rate": 1.2386685045388313e-05,
    "timestep_mean": 518.75,
    "epsilon_norm": 4.519239902496338,
    "prediction_norm": 4.011549472808838,
    "timestamp": "2025-07-14T12:34:40.570199"
  },
  {
    "epoch": 1859,
    "batch_idx": 0,
    "loss": 0.26654326915740967,
    "learning_rate": 1.2213550255719629e-05,
    "timestep_mean": 444.40625,
    "epsilon_norm": 7.36646032333374,
    "prediction_norm": 6.185093879699707,
    "timestamp": "2025-07-14T12:34:40.575269"
  },
  {
    "epoch": 1859,
    "batch_idx": 32,
    "loss": 0.25852298736572266,
    "learning_rate": 1.2213550255719629e-05,
    "timestep_mean": 519.25,
    "epsilon_norm": 9.171783447265625,
    "prediction_norm": 7.943967342376709,
    "timestamp": "2025-07-14T12:34:40.580152"
  },
  {
    "epoch": 1859,
    "batch_idx": 64,
    "loss": 0.23174303770065308,
    "learning_rate": 1.2213550255719629e-05,
    "timestep_mean": 508.46875,
    "epsilon_norm": 7.50065279006958,
    "prediction_norm": 6.1747050285339355,
    "timestamp": "2025-07-14T12:34:40.585226"
  },
  {
    "epoch": 1859,
    "batch_idx": 96,
    "loss": 0.2035713791847229,
    "learning_rate": 1.2213550255719629e-05,
    "timestep_mean": 551.28125,
    "epsilon_norm": 7.946378707885742,
    "prediction_norm": 7.179941654205322,
    "timestamp": "2025-07-14T12:34:40.590247"
  },
  {
    "epoch": 1859,
    "batch_idx": 128,
    "loss": 0.31798994541168213,
    "learning_rate": 1.2213550255719629e-05,
    "timestep_mean": 435.4375,
    "epsilon_norm": 7.869882106781006,
    "prediction_norm": 6.783312797546387,
    "timestamp": "2025-07-14T12:34:40.595099"
  },
  {
    "epoch": 1859,
    "batch_idx": 160,
    "loss": 0.1420731544494629,
    "learning_rate": 1.2213550255719629e-05,
    "timestep_mean": 595.125,
    "epsilon_norm": 7.065395355224609,
    "prediction_norm": 6.369874000549316,
    "timestamp": "2025-07-14T12:34:40.599947"
  },
  {
    "epoch": 1859,
    "batch_idx": 192,
    "loss": 0.1442175954580307,
    "learning_rate": 1.2213550255719629e-05,
    "timestep_mean": 633.125,
    "epsilon_norm": 3.4248673915863037,
    "prediction_norm": 3.4595322608947754,
    "timestamp": "2025-07-14T12:34:40.604851"
  },
  {
    "epoch": 1860,
    "batch_idx": 0,
    "loss": 0.4236912727355957,
    "learning_rate": 1.2041619030626337e-05,
    "timestep_mean": 518.125,
    "epsilon_norm": 7.421485424041748,
    "prediction_norm": 5.963351726531982,
    "timestamp": "2025-07-14T12:34:40.609712"
  },
  {
    "epoch": 1860,
    "batch_idx": 32,
    "loss": 0.14103253185749054,
    "learning_rate": 1.2041619030626337e-05,
    "timestep_mean": 488.9375,
    "epsilon_norm": 7.038665294647217,
    "prediction_norm": 6.047985076904297,
    "timestamp": "2025-07-14T12:34:40.614540"
  },
  {
    "epoch": 1860,
    "batch_idx": 64,
    "loss": 0.3030443489551544,
    "learning_rate": 1.2041619030626337e-05,
    "timestep_mean": 472.5625,
    "epsilon_norm": 7.718839645385742,
    "prediction_norm": 6.452757835388184,
    "timestamp": "2025-07-14T12:34:40.619436"
  },
  {
    "epoch": 1860,
    "batch_idx": 96,
    "loss": 0.16397182643413544,
    "learning_rate": 1.2041619030626337e-05,
    "timestep_mean": 604.5,
    "epsilon_norm": 7.19675874710083,
    "prediction_norm": 6.56081485748291,
    "timestamp": "2025-07-14T12:34:40.624337"
  },
  {
    "epoch": 1860,
    "batch_idx": 128,
    "loss": 0.14729470014572144,
    "learning_rate": 1.2041619030626337e-05,
    "timestep_mean": 547.84375,
    "epsilon_norm": 7.38375186920166,
    "prediction_norm": 6.2671098709106445,
    "timestamp": "2025-07-14T12:34:40.629276"
  },
  {
    "epoch": 1860,
    "batch_idx": 160,
    "loss": 0.31972888112068176,
    "learning_rate": 1.2041619030626337e-05,
    "timestep_mean": 441.5625,
    "epsilon_norm": 8.648311614990234,
    "prediction_norm": 7.0497283935546875,
    "timestamp": "2025-07-14T12:34:40.634237"
  },
  {
    "epoch": 1860,
    "batch_idx": 192,
    "loss": 0.2846878170967102,
    "learning_rate": 1.2041619030626337e-05,
    "timestep_mean": 524.0,
    "epsilon_norm": 3.640761375427246,
    "prediction_norm": 2.786231279373169,
    "timestamp": "2025-07-14T12:34:40.639149"
  },
  {
    "epoch": 1861,
    "batch_idx": 0,
    "loss": 0.17462803423404694,
    "learning_rate": 1.1870891794331537e-05,
    "timestep_mean": 598.875,
    "epsilon_norm": 8.94003677368164,
    "prediction_norm": 7.880522727966309,
    "timestamp": "2025-07-14T12:34:40.644070"
  },
  {
    "epoch": 1861,
    "batch_idx": 32,
    "loss": 0.23249423503875732,
    "learning_rate": 1.1870891794331537e-05,
    "timestep_mean": 455.90625,
    "epsilon_norm": 7.173667907714844,
    "prediction_norm": 6.177821159362793,
    "timestamp": "2025-07-14T12:34:40.648983"
  },
  {
    "epoch": 1861,
    "batch_idx": 64,
    "loss": 0.30383336544036865,
    "learning_rate": 1.1870891794331537e-05,
    "timestep_mean": 443.0,
    "epsilon_norm": 8.115577697753906,
    "prediction_norm": 7.1590895652771,
    "timestamp": "2025-07-14T12:34:40.653921"
  },
  {
    "epoch": 1861,
    "batch_idx": 96,
    "loss": 0.21243293583393097,
    "learning_rate": 1.1870891794331537e-05,
    "timestep_mean": 501.4375,
    "epsilon_norm": 8.397295951843262,
    "prediction_norm": 7.317292213439941,
    "timestamp": "2025-07-14T12:34:40.658884"
  },
  {
    "epoch": 1861,
    "batch_idx": 128,
    "loss": 0.21422961354255676,
    "learning_rate": 1.1870891794331537e-05,
    "timestep_mean": 638.53125,
    "epsilon_norm": 8.381450653076172,
    "prediction_norm": 7.307682991027832,
    "timestamp": "2025-07-14T12:34:40.663743"
  },
  {
    "epoch": 1861,
    "batch_idx": 160,
    "loss": 0.12908098101615906,
    "learning_rate": 1.1870891794331537e-05,
    "timestep_mean": 614.0625,
    "epsilon_norm": 7.605162143707275,
    "prediction_norm": 7.224534034729004,
    "timestamp": "2025-07-14T12:34:40.668622"
  },
  {
    "epoch": 1861,
    "batch_idx": 192,
    "loss": 0.2380591630935669,
    "learning_rate": 1.1870891794331537e-05,
    "timestep_mean": 586.625,
    "epsilon_norm": 3.961764335632324,
    "prediction_norm": 3.59578537940979,
    "timestamp": "2025-07-14T12:34:40.673479"
  },
  {
    "epoch": 1862,
    "batch_idx": 0,
    "loss": 0.2014136016368866,
    "learning_rate": 1.1701368968087708e-05,
    "timestep_mean": 427.375,
    "epsilon_norm": 6.942519664764404,
    "prediction_norm": 5.552695274353027,
    "timestamp": "2025-07-14T12:34:40.678430"
  },
  {
    "epoch": 1862,
    "batch_idx": 32,
    "loss": 0.32045912742614746,
    "learning_rate": 1.1701368968087708e-05,
    "timestep_mean": 526.5625,
    "epsilon_norm": 8.16799259185791,
    "prediction_norm": 7.588486194610596,
    "timestamp": "2025-07-14T12:34:40.683256"
  },
  {
    "epoch": 1862,
    "batch_idx": 64,
    "loss": 0.28854915499687195,
    "learning_rate": 1.1701368968087708e-05,
    "timestep_mean": 504.78125,
    "epsilon_norm": 7.470539093017578,
    "prediction_norm": 6.225177764892578,
    "timestamp": "2025-07-14T12:34:40.688264"
  },
  {
    "epoch": 1862,
    "batch_idx": 96,
    "loss": 0.21907666325569153,
    "learning_rate": 1.1701368968087708e-05,
    "timestep_mean": 539.5,
    "epsilon_norm": 7.684806823730469,
    "prediction_norm": 6.70326042175293,
    "timestamp": "2025-07-14T12:34:40.693135"
  },
  {
    "epoch": 1862,
    "batch_idx": 128,
    "loss": 0.24754878878593445,
    "learning_rate": 1.1701368968087708e-05,
    "timestep_mean": 553.59375,
    "epsilon_norm": 7.280516147613525,
    "prediction_norm": 6.244089126586914,
    "timestamp": "2025-07-14T12:34:40.698022"
  },
  {
    "epoch": 1862,
    "batch_idx": 160,
    "loss": 0.23405805230140686,
    "learning_rate": 1.1701368968087708e-05,
    "timestep_mean": 440.625,
    "epsilon_norm": 6.50161600112915,
    "prediction_norm": 5.7149658203125,
    "timestamp": "2025-07-14T12:34:40.702917"
  },
  {
    "epoch": 1862,
    "batch_idx": 192,
    "loss": 0.22205552458763123,
    "learning_rate": 1.1701368968087708e-05,
    "timestep_mean": 582.5,
    "epsilon_norm": 4.1933183670043945,
    "prediction_norm": 3.3633875846862793,
    "timestamp": "2025-07-14T12:34:40.707774"
  },
  {
    "epoch": 1863,
    "batch_idx": 0,
    "loss": 0.3452330231666565,
    "learning_rate": 1.1533050970175649e-05,
    "timestep_mean": 456.96875,
    "epsilon_norm": 8.45058536529541,
    "prediction_norm": 7.014277935028076,
    "timestamp": "2025-07-14T12:34:40.712636"
  },
  {
    "epoch": 1863,
    "batch_idx": 32,
    "loss": 0.33142775297164917,
    "learning_rate": 1.1533050970175649e-05,
    "timestep_mean": 439.25,
    "epsilon_norm": 8.026677131652832,
    "prediction_norm": 6.708740234375,
    "timestamp": "2025-07-14T12:34:40.717569"
  },
  {
    "epoch": 1863,
    "batch_idx": 64,
    "loss": 0.1477654129266739,
    "learning_rate": 1.1533050970175649e-05,
    "timestep_mean": 561.9375,
    "epsilon_norm": 6.954174041748047,
    "prediction_norm": 6.052996635437012,
    "timestamp": "2025-07-14T12:34:40.722469"
  },
  {
    "epoch": 1863,
    "batch_idx": 96,
    "loss": 0.4441472589969635,
    "learning_rate": 1.1533050970175649e-05,
    "timestep_mean": 522.125,
    "epsilon_norm": 9.197535514831543,
    "prediction_norm": 7.55369234085083,
    "timestamp": "2025-07-14T12:34:40.727367"
  },
  {
    "epoch": 1863,
    "batch_idx": 128,
    "loss": 0.2611703872680664,
    "learning_rate": 1.1533050970175649e-05,
    "timestep_mean": 482.15625,
    "epsilon_norm": 7.962719440460205,
    "prediction_norm": 7.080682754516602,
    "timestamp": "2025-07-14T12:34:40.732323"
  },
  {
    "epoch": 1863,
    "batch_idx": 160,
    "loss": 0.23141475021839142,
    "learning_rate": 1.1533050970175649e-05,
    "timestep_mean": 556.5,
    "epsilon_norm": 8.053382873535156,
    "prediction_norm": 6.747251033782959,
    "timestamp": "2025-07-14T12:34:40.737213"
  },
  {
    "epoch": 1863,
    "batch_idx": 192,
    "loss": 0.2534017264842987,
    "learning_rate": 1.1533050970175649e-05,
    "timestep_mean": 444.625,
    "epsilon_norm": 3.8076415061950684,
    "prediction_norm": 2.809666633605957,
    "timestamp": "2025-07-14T12:34:40.742172"
  },
  {
    "epoch": 1864,
    "batch_idx": 0,
    "loss": 0.2677120864391327,
    "learning_rate": 1.1365938215903257e-05,
    "timestep_mean": 506.0625,
    "epsilon_norm": 8.954028129577637,
    "prediction_norm": 7.91995906829834,
    "timestamp": "2025-07-14T12:34:40.747060"
  },
  {
    "epoch": 1864,
    "batch_idx": 32,
    "loss": 0.22365060448646545,
    "learning_rate": 1.1365938215903257e-05,
    "timestep_mean": 515.65625,
    "epsilon_norm": 7.872859477996826,
    "prediction_norm": 7.438235759735107,
    "timestamp": "2025-07-14T12:34:40.751934"
  },
  {
    "epoch": 1864,
    "batch_idx": 64,
    "loss": 0.20562826097011566,
    "learning_rate": 1.1365938215903257e-05,
    "timestep_mean": 490.03125,
    "epsilon_norm": 6.847231864929199,
    "prediction_norm": 5.340065956115723,
    "timestamp": "2025-07-14T12:34:40.756808"
  },
  {
    "epoch": 1864,
    "batch_idx": 96,
    "loss": 0.4330393671989441,
    "learning_rate": 1.1365938215903257e-05,
    "timestep_mean": 390.125,
    "epsilon_norm": 8.02737045288086,
    "prediction_norm": 6.176455020904541,
    "timestamp": "2025-07-14T12:34:40.761653"
  },
  {
    "epoch": 1864,
    "batch_idx": 128,
    "loss": 0.25045016407966614,
    "learning_rate": 1.1365938215903257e-05,
    "timestep_mean": 532.90625,
    "epsilon_norm": 6.917873859405518,
    "prediction_norm": 5.2603044509887695,
    "timestamp": "2025-07-14T12:34:40.766552"
  },
  {
    "epoch": 1864,
    "batch_idx": 160,
    "loss": 0.26349937915802,
    "learning_rate": 1.1365938215903257e-05,
    "timestep_mean": 450.78125,
    "epsilon_norm": 7.195577144622803,
    "prediction_norm": 5.7853875160217285,
    "timestamp": "2025-07-14T12:34:40.771480"
  },
  {
    "epoch": 1864,
    "batch_idx": 192,
    "loss": 0.33053839206695557,
    "learning_rate": 1.1365938215903257e-05,
    "timestep_mean": 610.125,
    "epsilon_norm": 5.147633075714111,
    "prediction_norm": 5.01871919631958,
    "timestamp": "2025-07-14T12:34:40.776300"
  },
  {
    "epoch": 1865,
    "batch_idx": 0,
    "loss": 0.22734656929969788,
    "learning_rate": 1.12000311176047e-05,
    "timestep_mean": 512.5625,
    "epsilon_norm": 8.2492036819458,
    "prediction_norm": 7.359662055969238,
    "timestamp": "2025-07-14T12:34:40.781175"
  },
  {
    "epoch": 1865,
    "batch_idx": 32,
    "loss": 0.2427680492401123,
    "learning_rate": 1.12000311176047e-05,
    "timestep_mean": 465.71875,
    "epsilon_norm": 6.799583435058594,
    "prediction_norm": 5.293118953704834,
    "timestamp": "2025-07-14T12:34:40.786064"
  },
  {
    "epoch": 1865,
    "batch_idx": 64,
    "loss": 0.28092843294143677,
    "learning_rate": 1.12000311176047e-05,
    "timestep_mean": 488.6875,
    "epsilon_norm": 8.501505851745605,
    "prediction_norm": 7.212518215179443,
    "timestamp": "2025-07-14T12:34:40.790954"
  },
  {
    "epoch": 1865,
    "batch_idx": 96,
    "loss": 0.5454766750335693,
    "learning_rate": 1.12000311176047e-05,
    "timestep_mean": 513.46875,
    "epsilon_norm": 7.452210426330566,
    "prediction_norm": 5.399300575256348,
    "timestamp": "2025-07-14T12:34:40.795809"
  },
  {
    "epoch": 1865,
    "batch_idx": 128,
    "loss": 0.1292148232460022,
    "learning_rate": 1.12000311176047e-05,
    "timestep_mean": 483.9375,
    "epsilon_norm": 8.303840637207031,
    "prediction_norm": 7.308459281921387,
    "timestamp": "2025-07-14T12:34:40.800737"
  },
  {
    "epoch": 1865,
    "batch_idx": 160,
    "loss": 0.19543369114398956,
    "learning_rate": 1.12000311176047e-05,
    "timestep_mean": 459.03125,
    "epsilon_norm": 7.083101272583008,
    "prediction_norm": 6.255768299102783,
    "timestamp": "2025-07-14T12:34:40.805601"
  },
  {
    "epoch": 1865,
    "batch_idx": 192,
    "loss": 0.10197288542985916,
    "learning_rate": 1.12000311176047e-05,
    "timestep_mean": 517.125,
    "epsilon_norm": 3.10841965675354,
    "prediction_norm": 2.620668649673462,
    "timestamp": "2025-07-14T12:34:40.810565"
  },
  {
    "epoch": 1866,
    "batch_idx": 0,
    "loss": 0.25022023916244507,
    "learning_rate": 1.1035330084639135e-05,
    "timestep_mean": 484.0625,
    "epsilon_norm": 7.469274044036865,
    "prediction_norm": 6.046911239624023,
    "timestamp": "2025-07-14T12:34:40.815502"
  },
  {
    "epoch": 1866,
    "batch_idx": 32,
    "loss": 0.559215247631073,
    "learning_rate": 1.1035330084639135e-05,
    "timestep_mean": 389.59375,
    "epsilon_norm": 8.690587997436523,
    "prediction_norm": 6.239748954772949,
    "timestamp": "2025-07-14T12:34:40.820389"
  },
  {
    "epoch": 1866,
    "batch_idx": 64,
    "loss": 0.2271331548690796,
    "learning_rate": 1.1035330084639135e-05,
    "timestep_mean": 587.84375,
    "epsilon_norm": 8.617584228515625,
    "prediction_norm": 7.119743824005127,
    "timestamp": "2025-07-14T12:34:40.825358"
  },
  {
    "epoch": 1866,
    "batch_idx": 96,
    "loss": 0.08808747678995132,
    "learning_rate": 1.1035330084639135e-05,
    "timestep_mean": 551.96875,
    "epsilon_norm": 7.936644077301025,
    "prediction_norm": 7.2423601150512695,
    "timestamp": "2025-07-14T12:34:40.830284"
  },
  {
    "epoch": 1866,
    "batch_idx": 128,
    "loss": 0.26201000809669495,
    "learning_rate": 1.1035330084639135e-05,
    "timestep_mean": 490.8125,
    "epsilon_norm": 8.853531837463379,
    "prediction_norm": 7.921635627746582,
    "timestamp": "2025-07-14T12:34:40.835152"
  },
  {
    "epoch": 1866,
    "batch_idx": 160,
    "loss": 0.33705541491508484,
    "learning_rate": 1.1035330084639135e-05,
    "timestep_mean": 539.71875,
    "epsilon_norm": 7.9377617835998535,
    "prediction_norm": 5.899838924407959,
    "timestamp": "2025-07-14T12:34:40.840068"
  },
  {
    "epoch": 1866,
    "batch_idx": 192,
    "loss": 0.11564011126756668,
    "learning_rate": 1.1035330084639135e-05,
    "timestep_mean": 546.875,
    "epsilon_norm": 4.097851276397705,
    "prediction_norm": 3.6862540245056152,
    "timestamp": "2025-07-14T12:34:40.844933"
  },
  {
    "epoch": 1867,
    "batch_idx": 0,
    "loss": 0.17418798804283142,
    "learning_rate": 1.087183552338999e-05,
    "timestep_mean": 429.0,
    "epsilon_norm": 7.90018367767334,
    "prediction_norm": 6.841359615325928,
    "timestamp": "2025-07-14T12:34:40.849880"
  },
  {
    "epoch": 1867,
    "batch_idx": 32,
    "loss": 0.20038364827632904,
    "learning_rate": 1.087183552338999e-05,
    "timestep_mean": 445.21875,
    "epsilon_norm": 7.520385265350342,
    "prediction_norm": 7.0575761795043945,
    "timestamp": "2025-07-14T12:34:40.854880"
  },
  {
    "epoch": 1867,
    "batch_idx": 64,
    "loss": 0.1660470962524414,
    "learning_rate": 1.087183552338999e-05,
    "timestep_mean": 538.5625,
    "epsilon_norm": 7.425967216491699,
    "prediction_norm": 6.953421592712402,
    "timestamp": "2025-07-14T12:34:40.859730"
  },
  {
    "epoch": 1867,
    "batch_idx": 96,
    "loss": 0.39437779784202576,
    "learning_rate": 1.087183552338999e-05,
    "timestep_mean": 478.84375,
    "epsilon_norm": 9.176860809326172,
    "prediction_norm": 7.857182502746582,
    "timestamp": "2025-07-14T12:34:40.864666"
  },
  {
    "epoch": 1867,
    "batch_idx": 128,
    "loss": 0.10262884199619293,
    "learning_rate": 1.087183552338999e-05,
    "timestep_mean": 576.84375,
    "epsilon_norm": 7.741661548614502,
    "prediction_norm": 7.014479160308838,
    "timestamp": "2025-07-14T12:34:40.869577"
  },
  {
    "epoch": 1867,
    "batch_idx": 160,
    "loss": 0.29734691977500916,
    "learning_rate": 1.087183552338999e-05,
    "timestep_mean": 494.0,
    "epsilon_norm": 8.515981674194336,
    "prediction_norm": 7.256627559661865,
    "timestamp": "2025-07-14T12:34:40.874401"
  },
  {
    "epoch": 1867,
    "batch_idx": 192,
    "loss": 0.37159544229507446,
    "learning_rate": 1.087183552338999e-05,
    "timestep_mean": 561.0,
    "epsilon_norm": 3.9245657920837402,
    "prediction_norm": 3.427501916885376,
    "timestamp": "2025-07-14T12:34:40.879226"
  },
  {
    "epoch": 1868,
    "batch_idx": 0,
    "loss": 0.46695226430892944,
    "learning_rate": 1.0709547837263962e-05,
    "timestep_mean": 527.0,
    "epsilon_norm": 8.320804595947266,
    "prediction_norm": 6.661975860595703,
    "timestamp": "2025-07-14T12:34:40.884094"
  },
  {
    "epoch": 1868,
    "batch_idx": 32,
    "loss": 0.16901445388793945,
    "learning_rate": 1.0709547837263962e-05,
    "timestep_mean": 548.46875,
    "epsilon_norm": 7.51478385925293,
    "prediction_norm": 7.360904216766357,
    "timestamp": "2025-07-14T12:34:40.888994"
  },
  {
    "epoch": 1868,
    "batch_idx": 64,
    "loss": 0.31937992572784424,
    "learning_rate": 1.0709547837263962e-05,
    "timestep_mean": 502.53125,
    "epsilon_norm": 8.66500473022461,
    "prediction_norm": 7.620684623718262,
    "timestamp": "2025-07-14T12:34:40.893867"
  },
  {
    "epoch": 1868,
    "batch_idx": 96,
    "loss": 0.3772351145744324,
    "learning_rate": 1.0709547837263962e-05,
    "timestep_mean": 504.90625,
    "epsilon_norm": 8.516154289245605,
    "prediction_norm": 6.566050052642822,
    "timestamp": "2025-07-14T12:34:40.898743"
  },
  {
    "epoch": 1868,
    "batch_idx": 128,
    "loss": 0.2638217508792877,
    "learning_rate": 1.0709547837263962e-05,
    "timestep_mean": 480.21875,
    "epsilon_norm": 7.684903144836426,
    "prediction_norm": 6.757511615753174,
    "timestamp": "2025-07-14T12:34:40.903645"
  },
  {
    "epoch": 1868,
    "batch_idx": 160,
    "loss": 0.3179768919944763,
    "learning_rate": 1.0709547837263962e-05,
    "timestep_mean": 498.34375,
    "epsilon_norm": 7.3149895668029785,
    "prediction_norm": 5.717563629150391,
    "timestamp": "2025-07-14T12:34:40.908649"
  },
  {
    "epoch": 1868,
    "batch_idx": 192,
    "loss": 0.1312796175479889,
    "learning_rate": 1.0709547837263962e-05,
    "timestep_mean": 334.75,
    "epsilon_norm": 2.362253427505493,
    "prediction_norm": 1.830896258354187,
    "timestamp": "2025-07-14T12:34:40.913529"
  },
  {
    "epoch": 1869,
    "batch_idx": 0,
    "loss": 0.32054686546325684,
    "learning_rate": 1.0548467426689685e-05,
    "timestep_mean": 456.875,
    "epsilon_norm": 7.694703578948975,
    "prediction_norm": 6.241257667541504,
    "timestamp": "2025-07-14T12:34:40.918467"
  },
  {
    "epoch": 1869,
    "batch_idx": 32,
    "loss": 0.39196130633354187,
    "learning_rate": 1.0548467426689685e-05,
    "timestep_mean": 540.4375,
    "epsilon_norm": 8.305362701416016,
    "prediction_norm": 6.424998760223389,
    "timestamp": "2025-07-14T12:34:40.923405"
  },
  {
    "epoch": 1869,
    "batch_idx": 64,
    "loss": 0.3865790069103241,
    "learning_rate": 1.0548467426689685e-05,
    "timestep_mean": 631.96875,
    "epsilon_norm": 8.45168685913086,
    "prediction_norm": 6.791138172149658,
    "timestamp": "2025-07-14T12:34:40.928284"
  },
  {
    "epoch": 1869,
    "batch_idx": 96,
    "loss": 0.28450727462768555,
    "learning_rate": 1.0548467426689685e-05,
    "timestep_mean": 416.75,
    "epsilon_norm": 6.974344253540039,
    "prediction_norm": 5.649078369140625,
    "timestamp": "2025-07-14T12:34:40.933168"
  },
  {
    "epoch": 1869,
    "batch_idx": 128,
    "loss": 0.2771137058734894,
    "learning_rate": 1.0548467426689685e-05,
    "timestep_mean": 527.03125,
    "epsilon_norm": 6.927164077758789,
    "prediction_norm": 5.72125768661499,
    "timestamp": "2025-07-14T12:34:40.938176"
  },
  {
    "epoch": 1869,
    "batch_idx": 160,
    "loss": 0.17851094901561737,
    "learning_rate": 1.0548467426689685e-05,
    "timestep_mean": 556.625,
    "epsilon_norm": 8.800379753112793,
    "prediction_norm": 8.236313819885254,
    "timestamp": "2025-07-14T12:34:40.943040"
  },
  {
    "epoch": 1869,
    "batch_idx": 192,
    "loss": 0.25538936257362366,
    "learning_rate": 1.0548467426689685e-05,
    "timestep_mean": 229.25,
    "epsilon_norm": 3.2416954040527344,
    "prediction_norm": 2.8066506385803223,
    "timestamp": "2025-07-14T12:34:40.947972"
  },
  {
    "epoch": 1870,
    "batch_idx": 0,
    "loss": 0.26773521304130554,
    "learning_rate": 1.0388594689117066e-05,
    "timestep_mean": 431.34375,
    "epsilon_norm": 8.628924369812012,
    "prediction_norm": 7.627359867095947,
    "timestamp": "2025-07-14T12:34:40.952882"
  },
  {
    "epoch": 1870,
    "batch_idx": 32,
    "loss": 0.3566284775733948,
    "learning_rate": 1.0388594689117066e-05,
    "timestep_mean": 409.40625,
    "epsilon_norm": 7.717319965362549,
    "prediction_norm": 5.624156951904297,
    "timestamp": "2025-07-14T12:34:40.957784"
  },
  {
    "epoch": 1870,
    "batch_idx": 64,
    "loss": 0.1311522126197815,
    "learning_rate": 1.0388594689117066e-05,
    "timestep_mean": 541.03125,
    "epsilon_norm": 8.67929744720459,
    "prediction_norm": 8.274316787719727,
    "timestamp": "2025-07-14T12:34:40.962642"
  },
  {
    "epoch": 1870,
    "batch_idx": 96,
    "loss": 0.38038021326065063,
    "learning_rate": 1.0388594689117066e-05,
    "timestep_mean": 519.1875,
    "epsilon_norm": 8.862189292907715,
    "prediction_norm": 7.214142799377441,
    "timestamp": "2025-07-14T12:34:40.967601"
  },
  {
    "epoch": 1870,
    "batch_idx": 128,
    "loss": 0.37231773138046265,
    "learning_rate": 1.0388594689117066e-05,
    "timestep_mean": 525.3125,
    "epsilon_norm": 8.274564743041992,
    "prediction_norm": 6.851367473602295,
    "timestamp": "2025-07-14T12:34:40.972503"
  },
  {
    "epoch": 1870,
    "batch_idx": 160,
    "loss": 0.20438691973686218,
    "learning_rate": 1.0388594689117066e-05,
    "timestep_mean": 544.09375,
    "epsilon_norm": 8.156880378723145,
    "prediction_norm": 7.314676284790039,
    "timestamp": "2025-07-14T12:34:40.977364"
  },
  {
    "epoch": 1870,
    "batch_idx": 192,
    "loss": 0.26951053738594055,
    "learning_rate": 1.0388594689117066e-05,
    "timestep_mean": 489.75,
    "epsilon_norm": 4.276788234710693,
    "prediction_norm": 3.5344462394714355,
    "timestamp": "2025-07-14T12:34:40.982191"
  },
  {
    "epoch": 1871,
    "batch_idx": 0,
    "loss": 0.18320856988430023,
    "learning_rate": 1.0229930019016337e-05,
    "timestep_mean": 588.3125,
    "epsilon_norm": 7.78603458404541,
    "prediction_norm": 7.049820423126221,
    "timestamp": "2025-07-14T12:34:40.987026"
  },
  {
    "epoch": 1871,
    "batch_idx": 32,
    "loss": 0.225095734000206,
    "learning_rate": 1.0229930019016337e-05,
    "timestep_mean": 525.875,
    "epsilon_norm": 8.814742088317871,
    "prediction_norm": 7.941289901733398,
    "timestamp": "2025-07-14T12:34:40.991950"
  },
  {
    "epoch": 1871,
    "batch_idx": 64,
    "loss": 0.22402329742908478,
    "learning_rate": 1.0229930019016337e-05,
    "timestep_mean": 439.625,
    "epsilon_norm": 8.033647537231445,
    "prediction_norm": 6.6310296058654785,
    "timestamp": "2025-07-14T12:34:40.996816"
  },
  {
    "epoch": 1871,
    "batch_idx": 96,
    "loss": 0.1806102842092514,
    "learning_rate": 1.0229930019016337e-05,
    "timestep_mean": 424.09375,
    "epsilon_norm": 6.630230903625488,
    "prediction_norm": 5.980093002319336,
    "timestamp": "2025-07-14T12:34:41.001954"
  },
  {
    "epoch": 1871,
    "batch_idx": 128,
    "loss": 0.2709020972251892,
    "learning_rate": 1.0229930019016337e-05,
    "timestep_mean": 500.34375,
    "epsilon_norm": 9.584783554077148,
    "prediction_norm": 8.543920516967773,
    "timestamp": "2025-07-14T12:34:41.006829"
  },
  {
    "epoch": 1871,
    "batch_idx": 160,
    "loss": 0.2798709571361542,
    "learning_rate": 1.0229930019016337e-05,
    "timestep_mean": 514.40625,
    "epsilon_norm": 8.506614685058594,
    "prediction_norm": 6.898285388946533,
    "timestamp": "2025-07-14T12:34:41.011713"
  },
  {
    "epoch": 1871,
    "batch_idx": 192,
    "loss": 0.41692396998405457,
    "learning_rate": 1.0229930019016337e-05,
    "timestep_mean": 387.375,
    "epsilon_norm": 4.330398082733154,
    "prediction_norm": 3.8897650241851807,
    "timestamp": "2025-07-14T12:34:41.016593"
  },
  {
    "epoch": 1872,
    "batch_idx": 0,
    "loss": 0.3307066857814789,
    "learning_rate": 1.0072473807876564e-05,
    "timestep_mean": 525.625,
    "epsilon_norm": 8.58317756652832,
    "prediction_norm": 6.799381732940674,
    "timestamp": "2025-07-14T12:34:41.021602"
  },
  {
    "epoch": 1872,
    "batch_idx": 32,
    "loss": 0.17458373308181763,
    "learning_rate": 1.0072473807876564e-05,
    "timestep_mean": 494.0,
    "epsilon_norm": 7.311537265777588,
    "prediction_norm": 6.62594747543335,
    "timestamp": "2025-07-14T12:34:41.026472"
  },
  {
    "epoch": 1872,
    "batch_idx": 64,
    "loss": 0.2955522835254669,
    "learning_rate": 1.0072473807876564e-05,
    "timestep_mean": 537.59375,
    "epsilon_norm": 6.759897232055664,
    "prediction_norm": 6.078857421875,
    "timestamp": "2025-07-14T12:34:41.031362"
  },
  {
    "epoch": 1872,
    "batch_idx": 96,
    "loss": 0.518727719783783,
    "learning_rate": 1.0072473807876564e-05,
    "timestep_mean": 449.28125,
    "epsilon_norm": 9.612006187438965,
    "prediction_norm": 7.049291610717773,
    "timestamp": "2025-07-14T12:34:41.036256"
  },
  {
    "epoch": 1872,
    "batch_idx": 128,
    "loss": 0.301494300365448,
    "learning_rate": 1.0072473807876564e-05,
    "timestep_mean": 389.875,
    "epsilon_norm": 7.368466377258301,
    "prediction_norm": 5.874858856201172,
    "timestamp": "2025-07-14T12:34:41.041160"
  },
  {
    "epoch": 1872,
    "batch_idx": 160,
    "loss": 0.21481052041053772,
    "learning_rate": 1.0072473807876564e-05,
    "timestep_mean": 523.875,
    "epsilon_norm": 6.847982883453369,
    "prediction_norm": 6.329275608062744,
    "timestamp": "2025-07-14T12:34:41.046055"
  },
  {
    "epoch": 1872,
    "batch_idx": 192,
    "loss": 0.33542731404304504,
    "learning_rate": 1.0072473807876564e-05,
    "timestep_mean": 519.5,
    "epsilon_norm": 4.933619976043701,
    "prediction_norm": 4.688916206359863,
    "timestamp": "2025-07-14T12:34:41.050965"
  },
  {
    "epoch": 1873,
    "batch_idx": 0,
    "loss": 0.14202065765857697,
    "learning_rate": 9.916226444205473e-06,
    "timestep_mean": 537.0,
    "epsilon_norm": 7.925220489501953,
    "prediction_norm": 7.472301959991455,
    "timestamp": "2025-07-14T12:34:41.055893"
  },
  {
    "epoch": 1873,
    "batch_idx": 32,
    "loss": 0.4370371997356415,
    "learning_rate": 9.916226444205473e-06,
    "timestep_mean": 415.03125,
    "epsilon_norm": 7.903379917144775,
    "prediction_norm": 5.6305251121521,
    "timestamp": "2025-07-14T12:34:41.060801"
  },
  {
    "epoch": 1873,
    "batch_idx": 64,
    "loss": 0.28133028745651245,
    "learning_rate": 9.916226444205473e-06,
    "timestep_mean": 499.0,
    "epsilon_norm": 7.606022357940674,
    "prediction_norm": 6.609457015991211,
    "timestamp": "2025-07-14T12:34:41.065701"
  },
  {
    "epoch": 1873,
    "batch_idx": 96,
    "loss": 0.3550977408885956,
    "learning_rate": 9.916226444205473e-06,
    "timestep_mean": 449.90625,
    "epsilon_norm": 9.154572486877441,
    "prediction_norm": 7.0896124839782715,
    "timestamp": "2025-07-14T12:34:41.070599"
  },
  {
    "epoch": 1873,
    "batch_idx": 128,
    "loss": 0.1961817890405655,
    "learning_rate": 9.916226444205473e-06,
    "timestep_mean": 521.53125,
    "epsilon_norm": 7.581039905548096,
    "prediction_norm": 6.479776382446289,
    "timestamp": "2025-07-14T12:34:41.075608"
  },
  {
    "epoch": 1873,
    "batch_idx": 160,
    "loss": 0.34739431738853455,
    "learning_rate": 9.916226444205473e-06,
    "timestep_mean": 493.75,
    "epsilon_norm": 7.923131942749023,
    "prediction_norm": 5.606618404388428,
    "timestamp": "2025-07-14T12:34:41.080480"
  },
  {
    "epoch": 1873,
    "batch_idx": 192,
    "loss": 0.1697070151567459,
    "learning_rate": 9.916226444205473e-06,
    "timestep_mean": 416.625,
    "epsilon_norm": 4.253108501434326,
    "prediction_norm": 3.852353572845459,
    "timestamp": "2025-07-14T12:34:41.085337"
  },
  {
    "epoch": 1874,
    "batch_idx": 0,
    "loss": 0.22439491748809814,
    "learning_rate": 9.76118831352784e-06,
    "timestep_mean": 461.34375,
    "epsilon_norm": 7.008084774017334,
    "prediction_norm": 6.156355857849121,
    "timestamp": "2025-07-14T12:34:41.090254"
  },
  {
    "epoch": 1874,
    "batch_idx": 32,
    "loss": 0.16387242078781128,
    "learning_rate": 9.76118831352784e-06,
    "timestep_mean": 491.09375,
    "epsilon_norm": 6.780799388885498,
    "prediction_norm": 5.56103515625,
    "timestamp": "2025-07-14T12:34:41.095118"
  },
  {
    "epoch": 1874,
    "batch_idx": 64,
    "loss": 0.4093232750892639,
    "learning_rate": 9.76118831352784e-06,
    "timestep_mean": 508.5625,
    "epsilon_norm": 8.899205207824707,
    "prediction_norm": 7.81575345993042,
    "timestamp": "2025-07-14T12:34:41.100152"
  },
  {
    "epoch": 1874,
    "batch_idx": 96,
    "loss": 0.21127842366695404,
    "learning_rate": 9.76118831352784e-06,
    "timestep_mean": 420.1875,
    "epsilon_norm": 7.789065361022949,
    "prediction_norm": 6.41710090637207,
    "timestamp": "2025-07-14T12:34:41.105008"
  },
  {
    "epoch": 1874,
    "batch_idx": 128,
    "loss": 0.24951419234275818,
    "learning_rate": 9.76118831352784e-06,
    "timestep_mean": 492.5,
    "epsilon_norm": 8.276586532592773,
    "prediction_norm": 7.26185417175293,
    "timestamp": "2025-07-14T12:34:41.109886"
  },
  {
    "epoch": 1874,
    "batch_idx": 160,
    "loss": 0.4174697995185852,
    "learning_rate": 9.76118831352784e-06,
    "timestep_mean": 454.5625,
    "epsilon_norm": 8.33186149597168,
    "prediction_norm": 6.372574806213379,
    "timestamp": "2025-07-14T12:34:41.115029"
  },
  {
    "epoch": 1874,
    "batch_idx": 192,
    "loss": 0.15309970080852509,
    "learning_rate": 9.76118831352784e-06,
    "timestep_mean": 501.25,
    "epsilon_norm": 3.9472815990448,
    "prediction_norm": 3.4398815631866455,
    "timestamp": "2025-07-14T12:34:41.120158"
  },
  {
    "epoch": 1875,
    "batch_idx": 0,
    "loss": 0.23363572359085083,
    "learning_rate": 9.607359798384779e-06,
    "timestep_mean": 556.0,
    "epsilon_norm": 7.786534786224365,
    "prediction_norm": 7.242462158203125,
    "timestamp": "2025-07-14T12:34:41.125436"
  },
  {
    "epoch": 1875,
    "batch_idx": 32,
    "loss": 0.34818023443222046,
    "learning_rate": 9.607359798384779e-06,
    "timestep_mean": 513.15625,
    "epsilon_norm": 8.736747741699219,
    "prediction_norm": 7.811239242553711,
    "timestamp": "2025-07-14T12:34:41.130623"
  },
  {
    "epoch": 1875,
    "batch_idx": 64,
    "loss": 0.26204949617385864,
    "learning_rate": 9.607359798384779e-06,
    "timestep_mean": 453.125,
    "epsilon_norm": 7.183382987976074,
    "prediction_norm": 6.259357929229736,
    "timestamp": "2025-07-14T12:34:41.135796"
  },
  {
    "epoch": 1875,
    "batch_idx": 96,
    "loss": 0.3099935054779053,
    "learning_rate": 9.607359798384779e-06,
    "timestep_mean": 463.09375,
    "epsilon_norm": 8.181663513183594,
    "prediction_norm": 6.68098783493042,
    "timestamp": "2025-07-14T12:34:41.140950"
  },
  {
    "epoch": 1875,
    "batch_idx": 128,
    "loss": 0.308864563703537,
    "learning_rate": 9.607359798384779e-06,
    "timestep_mean": 559.71875,
    "epsilon_norm": 8.550060272216797,
    "prediction_norm": 6.811758518218994,
    "timestamp": "2025-07-14T12:34:41.146172"
  },
  {
    "epoch": 1875,
    "batch_idx": 160,
    "loss": 0.34348264336586,
    "learning_rate": 9.607359798384779e-06,
    "timestep_mean": 422.59375,
    "epsilon_norm": 7.806307315826416,
    "prediction_norm": 6.90889310836792,
    "timestamp": "2025-07-14T12:34:41.151377"
  },
  {
    "epoch": 1875,
    "batch_idx": 192,
    "loss": 0.12184688448905945,
    "learning_rate": 9.607359798384779e-06,
    "timestep_mean": 470.25,
    "epsilon_norm": 5.780981540679932,
    "prediction_norm": 5.355534076690674,
    "timestamp": "2025-07-14T12:34:41.156596"
  },
  {
    "epoch": 1876,
    "batch_idx": 0,
    "loss": 0.2655874788761139,
    "learning_rate": 9.454741278333005e-06,
    "timestep_mean": 454.40625,
    "epsilon_norm": 8.807740211486816,
    "prediction_norm": 7.827357292175293,
    "timestamp": "2025-07-14T12:34:41.161768"
  },
  {
    "epoch": 1876,
    "batch_idx": 32,
    "loss": 0.15689323842525482,
    "learning_rate": 9.454741278333005e-06,
    "timestep_mean": 500.125,
    "epsilon_norm": 8.694343566894531,
    "prediction_norm": 7.517043590545654,
    "timestamp": "2025-07-14T12:34:41.167092"
  },
  {
    "epoch": 1876,
    "batch_idx": 64,
    "loss": 0.22784365713596344,
    "learning_rate": 9.454741278333005e-06,
    "timestep_mean": 502.84375,
    "epsilon_norm": 8.075364112854004,
    "prediction_norm": 6.599542140960693,
    "timestamp": "2025-07-14T12:34:41.172306"
  },
  {
    "epoch": 1876,
    "batch_idx": 96,
    "loss": 0.2934175729751587,
    "learning_rate": 9.454741278333005e-06,
    "timestep_mean": 468.0625,
    "epsilon_norm": 7.187741756439209,
    "prediction_norm": 5.819797992706299,
    "timestamp": "2025-07-14T12:34:41.177224"
  },
  {
    "epoch": 1876,
    "batch_idx": 128,
    "loss": 0.24189847707748413,
    "learning_rate": 9.454741278333005e-06,
    "timestep_mean": 517.90625,
    "epsilon_norm": 8.330747604370117,
    "prediction_norm": 6.2681050300598145,
    "timestamp": "2025-07-14T12:34:41.182351"
  },
  {
    "epoch": 1876,
    "batch_idx": 160,
    "loss": 0.24833999574184418,
    "learning_rate": 9.454741278333005e-06,
    "timestep_mean": 513.125,
    "epsilon_norm": 7.455054759979248,
    "prediction_norm": 6.418844223022461,
    "timestamp": "2025-07-14T12:34:41.187571"
  },
  {
    "epoch": 1876,
    "batch_idx": 192,
    "loss": 0.2133677899837494,
    "learning_rate": 9.454741278333005e-06,
    "timestep_mean": 400.875,
    "epsilon_norm": 2.55263352394104,
    "prediction_norm": 2.0947799682617188,
    "timestamp": "2025-07-14T12:34:41.192785"
  },
  {
    "epoch": 1877,
    "batch_idx": 0,
    "loss": 0.21611559391021729,
    "learning_rate": 9.303333129943351e-06,
    "timestep_mean": 475.9375,
    "epsilon_norm": 7.4378533363342285,
    "prediction_norm": 6.384414196014404,
    "timestamp": "2025-07-14T12:34:41.198036"
  },
  {
    "epoch": 1877,
    "batch_idx": 32,
    "loss": 0.35389813780784607,
    "learning_rate": 9.303333129943351e-06,
    "timestep_mean": 560.375,
    "epsilon_norm": 9.283853530883789,
    "prediction_norm": 7.564090251922607,
    "timestamp": "2025-07-14T12:34:41.203085"
  },
  {
    "epoch": 1877,
    "batch_idx": 64,
    "loss": 0.41199782490730286,
    "learning_rate": 9.303333129943351e-06,
    "timestep_mean": 419.4375,
    "epsilon_norm": 7.973120212554932,
    "prediction_norm": 6.081460952758789,
    "timestamp": "2025-07-14T12:34:41.207959"
  },
  {
    "epoch": 1877,
    "batch_idx": 96,
    "loss": 0.3076995313167572,
    "learning_rate": 9.303333129943351e-06,
    "timestep_mean": 532.59375,
    "epsilon_norm": 9.025375366210938,
    "prediction_norm": 7.708096981048584,
    "timestamp": "2025-07-14T12:34:41.212950"
  },
  {
    "epoch": 1877,
    "batch_idx": 128,
    "loss": 0.2755235731601715,
    "learning_rate": 9.303333129943351e-06,
    "timestep_mean": 526.71875,
    "epsilon_norm": 7.443440914154053,
    "prediction_norm": 6.073982238769531,
    "timestamp": "2025-07-14T12:34:41.217893"
  },
  {
    "epoch": 1877,
    "batch_idx": 160,
    "loss": 0.34374988079071045,
    "learning_rate": 9.303333129943351e-06,
    "timestep_mean": 453.84375,
    "epsilon_norm": 7.789735317230225,
    "prediction_norm": 5.904726982116699,
    "timestamp": "2025-07-14T12:34:41.223057"
  },
  {
    "epoch": 1877,
    "batch_idx": 192,
    "loss": 0.21577024459838867,
    "learning_rate": 9.303333129943351e-06,
    "timestep_mean": 307.25,
    "epsilon_norm": 3.003679037094116,
    "prediction_norm": 2.5961191654205322,
    "timestamp": "2025-07-14T12:34:41.228213"
  },
  {
    "epoch": 1878,
    "batch_idx": 0,
    "loss": 0.09617125988006592,
    "learning_rate": 9.153135726800593e-06,
    "timestep_mean": 526.03125,
    "epsilon_norm": 6.859668254852295,
    "prediction_norm": 6.622602462768555,
    "timestamp": "2025-07-14T12:34:41.233550"
  },
  {
    "epoch": 1878,
    "batch_idx": 32,
    "loss": 0.24637222290039062,
    "learning_rate": 9.153135726800593e-06,
    "timestep_mean": 504.625,
    "epsilon_norm": 7.819867134094238,
    "prediction_norm": 7.131396770477295,
    "timestamp": "2025-07-14T12:34:41.238789"
  },
  {
    "epoch": 1878,
    "batch_idx": 64,
    "loss": 0.23745042085647583,
    "learning_rate": 9.153135726800593e-06,
    "timestep_mean": 466.40625,
    "epsilon_norm": 8.260899543762207,
    "prediction_norm": 6.6737895011901855,
    "timestamp": "2025-07-14T12:34:41.243992"
  },
  {
    "epoch": 1878,
    "batch_idx": 96,
    "loss": 0.340387225151062,
    "learning_rate": 9.153135726800593e-06,
    "timestep_mean": 514.8125,
    "epsilon_norm": 7.930132865905762,
    "prediction_norm": 6.812249183654785,
    "timestamp": "2025-07-14T12:34:41.249220"
  },
  {
    "epoch": 1878,
    "batch_idx": 128,
    "loss": 0.15564671158790588,
    "learning_rate": 9.153135726800593e-06,
    "timestep_mean": 580.4375,
    "epsilon_norm": 7.389193534851074,
    "prediction_norm": 6.977835655212402,
    "timestamp": "2025-07-14T12:34:41.254404"
  },
  {
    "epoch": 1878,
    "batch_idx": 160,
    "loss": 0.558112382888794,
    "learning_rate": 9.153135726800593e-06,
    "timestep_mean": 548.8125,
    "epsilon_norm": 7.926302909851074,
    "prediction_norm": 5.590843200683594,
    "timestamp": "2025-07-14T12:34:41.259680"
  },
  {
    "epoch": 1878,
    "batch_idx": 192,
    "loss": 0.4091457724571228,
    "learning_rate": 9.153135726800593e-06,
    "timestep_mean": 445.25,
    "epsilon_norm": 3.7984399795532227,
    "prediction_norm": 2.9019787311553955,
    "timestamp": "2025-07-14T12:34:41.264851"
  },
  {
    "epoch": 1879,
    "batch_idx": 0,
    "loss": 0.24577313661575317,
    "learning_rate": 9.004149439501723e-06,
    "timestep_mean": 530.09375,
    "epsilon_norm": 7.091289043426514,
    "prediction_norm": 6.3216872215271,
    "timestamp": "2025-07-14T12:34:41.270032"
  },
  {
    "epoch": 1879,
    "batch_idx": 32,
    "loss": 0.33411097526550293,
    "learning_rate": 9.004149439501723e-06,
    "timestep_mean": 495.6875,
    "epsilon_norm": 8.54878044128418,
    "prediction_norm": 6.684263706207275,
    "timestamp": "2025-07-14T12:34:41.275331"
  },
  {
    "epoch": 1879,
    "batch_idx": 64,
    "loss": 0.47220832109451294,
    "learning_rate": 9.004149439501723e-06,
    "timestep_mean": 459.6875,
    "epsilon_norm": 9.56831169128418,
    "prediction_norm": 7.610804557800293,
    "timestamp": "2025-07-14T12:34:41.280515"
  },
  {
    "epoch": 1879,
    "batch_idx": 96,
    "loss": 0.3631744384765625,
    "learning_rate": 9.004149439501723e-06,
    "timestep_mean": 486.5625,
    "epsilon_norm": 7.364563465118408,
    "prediction_norm": 5.853063106536865,
    "timestamp": "2025-07-14T12:34:41.285765"
  },
  {
    "epoch": 1879,
    "batch_idx": 128,
    "loss": 0.08071450889110565,
    "learning_rate": 9.004149439501723e-06,
    "timestep_mean": 583.25,
    "epsilon_norm": 7.42107629776001,
    "prediction_norm": 6.910975456237793,
    "timestamp": "2025-07-14T12:34:41.290637"
  },
  {
    "epoch": 1879,
    "batch_idx": 160,
    "loss": 0.1769605576992035,
    "learning_rate": 9.004149439501723e-06,
    "timestep_mean": 457.59375,
    "epsilon_norm": 8.198896408081055,
    "prediction_norm": 7.68765115737915,
    "timestamp": "2025-07-14T12:34:41.295761"
  },
  {
    "epoch": 1879,
    "batch_idx": 192,
    "loss": 0.15478818118572235,
    "learning_rate": 9.004149439501723e-06,
    "timestep_mean": 519.5,
    "epsilon_norm": 3.175042152404785,
    "prediction_norm": 3.0235960483551025,
    "timestamp": "2025-07-14T12:34:41.301134"
  },
  {
    "epoch": 1880,
    "batch_idx": 0,
    "loss": 0.15708215534687042,
    "learning_rate": 8.856374635655634e-06,
    "timestep_mean": 567.5625,
    "epsilon_norm": 7.674584865570068,
    "prediction_norm": 7.086948871612549,
    "timestamp": "2025-07-14T12:34:41.306320"
  },
  {
    "epoch": 1880,
    "batch_idx": 32,
    "loss": 0.33803093433380127,
    "learning_rate": 8.856374635655634e-06,
    "timestep_mean": 566.125,
    "epsilon_norm": 7.49938440322876,
    "prediction_norm": 6.18116569519043,
    "timestamp": "2025-07-14T12:34:41.311552"
  },
  {
    "epoch": 1880,
    "batch_idx": 64,
    "loss": 0.3018350899219513,
    "learning_rate": 8.856374635655634e-06,
    "timestep_mean": 470.875,
    "epsilon_norm": 8.397826194763184,
    "prediction_norm": 7.338483810424805,
    "timestamp": "2025-07-14T12:34:41.316882"
  },
  {
    "epoch": 1880,
    "batch_idx": 96,
    "loss": 0.1159667819738388,
    "learning_rate": 8.856374635655634e-06,
    "timestep_mean": 634.96875,
    "epsilon_norm": 9.096776962280273,
    "prediction_norm": 9.008198738098145,
    "timestamp": "2025-07-14T12:34:41.321992"
  },
  {
    "epoch": 1880,
    "batch_idx": 128,
    "loss": 0.11553998291492462,
    "learning_rate": 8.856374635655634e-06,
    "timestep_mean": 500.15625,
    "epsilon_norm": 8.201224327087402,
    "prediction_norm": 7.717614650726318,
    "timestamp": "2025-07-14T12:34:41.327117"
  },
  {
    "epoch": 1880,
    "batch_idx": 160,
    "loss": 0.3374597132205963,
    "learning_rate": 8.856374635655634e-06,
    "timestep_mean": 459.21875,
    "epsilon_norm": 8.765495300292969,
    "prediction_norm": 6.690089225769043,
    "timestamp": "2025-07-14T12:34:41.332222"
  },
  {
    "epoch": 1880,
    "batch_idx": 192,
    "loss": 0.11545351147651672,
    "learning_rate": 8.856374635655634e-06,
    "timestep_mean": 588.5,
    "epsilon_norm": 3.2358744144439697,
    "prediction_norm": 2.8392109870910645,
    "timestamp": "2025-07-14T12:34:41.337362"
  },
  {
    "epoch": 1881,
    "batch_idx": 0,
    "loss": 0.19327668845653534,
    "learning_rate": 8.709811679882045e-06,
    "timestep_mean": 546.53125,
    "epsilon_norm": 7.130038261413574,
    "prediction_norm": 6.23459529876709,
    "timestamp": "2025-07-14T12:34:41.342731"
  },
  {
    "epoch": 1881,
    "batch_idx": 32,
    "loss": 0.25434258580207825,
    "learning_rate": 8.709811679882045e-06,
    "timestep_mean": 515.96875,
    "epsilon_norm": 8.747662544250488,
    "prediction_norm": 7.7771501541137695,
    "timestamp": "2025-07-14T12:34:41.347952"
  },
  {
    "epoch": 1881,
    "batch_idx": 64,
    "loss": 0.25077176094055176,
    "learning_rate": 8.709811679882045e-06,
    "timestep_mean": 610.15625,
    "epsilon_norm": 7.896939277648926,
    "prediction_norm": 6.801474094390869,
    "timestamp": "2025-07-14T12:34:41.353151"
  },
  {
    "epoch": 1881,
    "batch_idx": 96,
    "loss": 0.28845450282096863,
    "learning_rate": 8.709811679882045e-06,
    "timestep_mean": 482.1875,
    "epsilon_norm": 7.569952964782715,
    "prediction_norm": 6.07760763168335,
    "timestamp": "2025-07-14T12:34:41.358513"
  },
  {
    "epoch": 1881,
    "batch_idx": 128,
    "loss": 0.2600184381008148,
    "learning_rate": 8.709811679882045e-06,
    "timestep_mean": 549.6875,
    "epsilon_norm": 7.663631916046143,
    "prediction_norm": 6.568434715270996,
    "timestamp": "2025-07-14T12:34:41.363774"
  },
  {
    "epoch": 1881,
    "batch_idx": 160,
    "loss": 0.3591468930244446,
    "learning_rate": 8.709811679882045e-06,
    "timestep_mean": 506.96875,
    "epsilon_norm": 8.242319107055664,
    "prediction_norm": 6.535724639892578,
    "timestamp": "2025-07-14T12:34:41.368984"
  },
  {
    "epoch": 1881,
    "batch_idx": 192,
    "loss": 0.0982041284441948,
    "learning_rate": 8.709811679882045e-06,
    "timestep_mean": 436.75,
    "epsilon_norm": 3.068530559539795,
    "prediction_norm": 3.085775136947632,
    "timestamp": "2025-07-14T12:34:41.374130"
  },
  {
    "epoch": 1882,
    "batch_idx": 0,
    "loss": 0.29010865092277527,
    "learning_rate": 8.564460933810407e-06,
    "timestep_mean": 487.1875,
    "epsilon_norm": 8.104368209838867,
    "prediction_norm": 6.465688228607178,
    "timestamp": "2025-07-14T12:34:41.379673"
  },
  {
    "epoch": 1882,
    "batch_idx": 32,
    "loss": 0.41517162322998047,
    "learning_rate": 8.564460933810407e-06,
    "timestep_mean": 386.8125,
    "epsilon_norm": 8.293036460876465,
    "prediction_norm": 6.283744812011719,
    "timestamp": "2025-07-14T12:34:41.385287"
  },
  {
    "epoch": 1882,
    "batch_idx": 64,
    "loss": 0.29885926842689514,
    "learning_rate": 8.564460933810407e-06,
    "timestep_mean": 490.03125,
    "epsilon_norm": 8.673728942871094,
    "prediction_norm": 7.325442314147949,
    "timestamp": "2025-07-14T12:34:41.390218"
  },
  {
    "epoch": 1882,
    "batch_idx": 96,
    "loss": 0.474489688873291,
    "learning_rate": 8.564460933810407e-06,
    "timestep_mean": 502.90625,
    "epsilon_norm": 8.516277313232422,
    "prediction_norm": 6.660519123077393,
    "timestamp": "2025-07-14T12:34:41.395255"
  },
  {
    "epoch": 1882,
    "batch_idx": 128,
    "loss": 0.2981955409049988,
    "learning_rate": 8.564460933810407e-06,
    "timestep_mean": 548.0,
    "epsilon_norm": 7.231101989746094,
    "prediction_norm": 6.175290107727051,
    "timestamp": "2025-07-14T12:34:41.400646"
  },
  {
    "epoch": 1882,
    "batch_idx": 160,
    "loss": 0.2790576219558716,
    "learning_rate": 8.564460933810407e-06,
    "timestep_mean": 432.4375,
    "epsilon_norm": 7.9298553466796875,
    "prediction_norm": 6.964502811431885,
    "timestamp": "2025-07-14T12:34:41.405507"
  },
  {
    "epoch": 1882,
    "batch_idx": 192,
    "loss": 0.09812119603157043,
    "learning_rate": 8.564460933810407e-06,
    "timestep_mean": 625.5,
    "epsilon_norm": 4.741825580596924,
    "prediction_norm": 4.1367082595825195,
    "timestamp": "2025-07-14T12:34:41.414683"
  },
  {
    "epoch": 1883,
    "batch_idx": 0,
    "loss": 0.23413152992725372,
    "learning_rate": 8.420322756079229e-06,
    "timestep_mean": 472.4375,
    "epsilon_norm": 7.927826881408691,
    "prediction_norm": 6.585193157196045,
    "timestamp": "2025-07-14T12:34:41.420568"
  },
  {
    "epoch": 1883,
    "batch_idx": 32,
    "loss": 0.18022650480270386,
    "learning_rate": 8.420322756079229e-06,
    "timestep_mean": 558.875,
    "epsilon_norm": 7.691162586212158,
    "prediction_norm": 6.917625427246094,
    "timestamp": "2025-07-14T12:34:41.426139"
  },
  {
    "epoch": 1883,
    "batch_idx": 64,
    "loss": 0.2858421504497528,
    "learning_rate": 8.420322756079229e-06,
    "timestep_mean": 542.125,
    "epsilon_norm": 8.666976928710938,
    "prediction_norm": 7.279589653015137,
    "timestamp": "2025-07-14T12:34:41.431529"
  },
  {
    "epoch": 1883,
    "batch_idx": 96,
    "loss": 0.1766391396522522,
    "learning_rate": 8.420322756079229e-06,
    "timestep_mean": 569.28125,
    "epsilon_norm": 7.817995071411133,
    "prediction_norm": 6.750916481018066,
    "timestamp": "2025-07-14T12:34:41.436869"
  },
  {
    "epoch": 1883,
    "batch_idx": 128,
    "loss": 0.2763025164604187,
    "learning_rate": 8.420322756079229e-06,
    "timestep_mean": 469.46875,
    "epsilon_norm": 8.245375633239746,
    "prediction_norm": 6.488624095916748,
    "timestamp": "2025-07-14T12:34:41.442128"
  },
  {
    "epoch": 1883,
    "batch_idx": 160,
    "loss": 0.24460530281066895,
    "learning_rate": 8.420322756079229e-06,
    "timestep_mean": 493.59375,
    "epsilon_norm": 7.561553001403809,
    "prediction_norm": 6.250185489654541,
    "timestamp": "2025-07-14T12:34:41.447194"
  },
  {
    "epoch": 1883,
    "batch_idx": 192,
    "loss": 0.29897773265838623,
    "learning_rate": 8.420322756079229e-06,
    "timestep_mean": 639.375,
    "epsilon_norm": 3.9344706535339355,
    "prediction_norm": 3.537684679031372,
    "timestamp": "2025-07-14T12:34:41.452508"
  },
  {
    "epoch": 1884,
    "batch_idx": 0,
    "loss": 0.3480461537837982,
    "learning_rate": 8.277397502335132e-06,
    "timestep_mean": 461.875,
    "epsilon_norm": 8.503360748291016,
    "prediction_norm": 6.787059307098389,
    "timestamp": "2025-07-14T12:34:41.457893"
  },
  {
    "epoch": 1884,
    "batch_idx": 32,
    "loss": 0.1910267174243927,
    "learning_rate": 8.277397502335132e-06,
    "timestep_mean": 485.5625,
    "epsilon_norm": 7.089483261108398,
    "prediction_norm": 5.558108806610107,
    "timestamp": "2025-07-14T12:34:41.463231"
  },
  {
    "epoch": 1884,
    "batch_idx": 64,
    "loss": 0.30937460064888,
    "learning_rate": 8.277397502335132e-06,
    "timestep_mean": 466.34375,
    "epsilon_norm": 8.059786796569824,
    "prediction_norm": 7.172420501708984,
    "timestamp": "2025-07-14T12:34:41.468446"
  },
  {
    "epoch": 1884,
    "batch_idx": 96,
    "loss": 0.21970778703689575,
    "learning_rate": 8.277397502335132e-06,
    "timestep_mean": 472.0,
    "epsilon_norm": 7.690547943115234,
    "prediction_norm": 6.9432053565979,
    "timestamp": "2025-07-14T12:34:41.473836"
  },
  {
    "epoch": 1884,
    "batch_idx": 128,
    "loss": 0.22410055994987488,
    "learning_rate": 8.277397502335132e-06,
    "timestep_mean": 563.28125,
    "epsilon_norm": 8.800606727600098,
    "prediction_norm": 7.837434768676758,
    "timestamp": "2025-07-14T12:34:41.479045"
  },
  {
    "epoch": 1884,
    "batch_idx": 160,
    "loss": 0.31608983874320984,
    "learning_rate": 8.277397502335132e-06,
    "timestep_mean": 496.21875,
    "epsilon_norm": 7.829258441925049,
    "prediction_norm": 6.391182899475098,
    "timestamp": "2025-07-14T12:34:41.484436"
  },
  {
    "epoch": 1884,
    "batch_idx": 192,
    "loss": 0.3038250207901001,
    "learning_rate": 8.277397502335132e-06,
    "timestep_mean": 474.125,
    "epsilon_norm": 4.2735137939453125,
    "prediction_norm": 3.980088710784912,
    "timestamp": "2025-07-14T12:34:41.489687"
  },
  {
    "epoch": 1885,
    "batch_idx": 0,
    "loss": 0.14942008256912231,
    "learning_rate": 8.135685525232077e-06,
    "timestep_mean": 573.40625,
    "epsilon_norm": 8.44164752960205,
    "prediction_norm": 7.682248115539551,
    "timestamp": "2025-07-14T12:34:41.494837"
  },
  {
    "epoch": 1885,
    "batch_idx": 32,
    "loss": 0.32073259353637695,
    "learning_rate": 8.135685525232077e-06,
    "timestep_mean": 553.09375,
    "epsilon_norm": 7.173622131347656,
    "prediction_norm": 5.405455112457275,
    "timestamp": "2025-07-14T12:34:41.500245"
  },
  {
    "epoch": 1885,
    "batch_idx": 64,
    "loss": 0.23405878245830536,
    "learning_rate": 8.135685525232077e-06,
    "timestep_mean": 445.125,
    "epsilon_norm": 7.883899688720703,
    "prediction_norm": 6.72334623336792,
    "timestamp": "2025-07-14T12:34:41.505589"
  },
  {
    "epoch": 1885,
    "batch_idx": 96,
    "loss": 0.1883769929409027,
    "learning_rate": 8.135685525232077e-06,
    "timestep_mean": 459.75,
    "epsilon_norm": 7.822900295257568,
    "prediction_norm": 6.9883713722229,
    "timestamp": "2025-07-14T12:34:41.510727"
  },
  {
    "epoch": 1885,
    "batch_idx": 128,
    "loss": 0.3139989674091339,
    "learning_rate": 8.135685525232077e-06,
    "timestep_mean": 598.5625,
    "epsilon_norm": 9.897819519042969,
    "prediction_norm": 8.517202377319336,
    "timestamp": "2025-07-14T12:34:41.515927"
  },
  {
    "epoch": 1885,
    "batch_idx": 160,
    "loss": 0.23938670754432678,
    "learning_rate": 8.135685525232077e-06,
    "timestep_mean": 467.59375,
    "epsilon_norm": 7.115323543548584,
    "prediction_norm": 5.886315822601318,
    "timestamp": "2025-07-14T12:34:41.521279"
  },
  {
    "epoch": 1885,
    "batch_idx": 192,
    "loss": 0.09276285767555237,
    "learning_rate": 8.135685525232077e-06,
    "timestep_mean": 602.5,
    "epsilon_norm": 3.425410509109497,
    "prediction_norm": 2.5763299465179443,
    "timestamp": "2025-07-14T12:34:41.526385"
  },
  {
    "epoch": 1886,
    "batch_idx": 0,
    "loss": 0.21423183381557465,
    "learning_rate": 7.995187174430146e-06,
    "timestep_mean": 518.3125,
    "epsilon_norm": 8.863837242126465,
    "prediction_norm": 8.334479331970215,
    "timestamp": "2025-07-14T12:34:41.531339"
  },
  {
    "epoch": 1886,
    "batch_idx": 32,
    "loss": 0.323297381401062,
    "learning_rate": 7.995187174430146e-06,
    "timestep_mean": 496.625,
    "epsilon_norm": 9.359307289123535,
    "prediction_norm": 7.661328315734863,
    "timestamp": "2025-07-14T12:34:41.536384"
  },
  {
    "epoch": 1886,
    "batch_idx": 64,
    "loss": 0.20269788801670074,
    "learning_rate": 7.995187174430146e-06,
    "timestep_mean": 496.53125,
    "epsilon_norm": 7.739684581756592,
    "prediction_norm": 6.617312431335449,
    "timestamp": "2025-07-14T12:34:41.541517"
  },
  {
    "epoch": 1886,
    "batch_idx": 96,
    "loss": 0.3142006993293762,
    "learning_rate": 7.995187174430146e-06,
    "timestep_mean": 454.53125,
    "epsilon_norm": 7.9504923820495605,
    "prediction_norm": 6.165183067321777,
    "timestamp": "2025-07-14T12:34:41.546495"
  },
  {
    "epoch": 1886,
    "batch_idx": 128,
    "loss": 0.17944461107254028,
    "learning_rate": 7.995187174430146e-06,
    "timestep_mean": 521.28125,
    "epsilon_norm": 8.374292373657227,
    "prediction_norm": 7.479651927947998,
    "timestamp": "2025-07-14T12:34:41.551407"
  },
  {
    "epoch": 1886,
    "batch_idx": 160,
    "loss": 0.22727710008621216,
    "learning_rate": 7.995187174430146e-06,
    "timestep_mean": 531.28125,
    "epsilon_norm": 7.8925251960754395,
    "prediction_norm": 7.082076549530029,
    "timestamp": "2025-07-14T12:34:41.556367"
  },
  {
    "epoch": 1886,
    "batch_idx": 192,
    "loss": 0.32088690996170044,
    "learning_rate": 7.995187174430146e-06,
    "timestep_mean": 649.25,
    "epsilon_norm": 5.04754638671875,
    "prediction_norm": 4.470005989074707,
    "timestamp": "2025-07-14T12:34:41.561370"
  },
  {
    "epoch": 1887,
    "batch_idx": 0,
    "loss": 0.1705930083990097,
    "learning_rate": 7.855902796595087e-06,
    "timestep_mean": 556.65625,
    "epsilon_norm": 8.428451538085938,
    "prediction_norm": 7.451749324798584,
    "timestamp": "2025-07-14T12:34:41.566348"
  },
  {
    "epoch": 1887,
    "batch_idx": 32,
    "loss": 0.4373950660228729,
    "learning_rate": 7.855902796595087e-06,
    "timestep_mean": 437.53125,
    "epsilon_norm": 8.339391708374023,
    "prediction_norm": 6.968442440032959,
    "timestamp": "2025-07-14T12:34:41.571258"
  },
  {
    "epoch": 1887,
    "batch_idx": 64,
    "loss": 0.16843228042125702,
    "learning_rate": 7.855902796595087e-06,
    "timestep_mean": 505.1875,
    "epsilon_norm": 7.540419578552246,
    "prediction_norm": 6.8949809074401855,
    "timestamp": "2025-07-14T12:34:41.576175"
  },
  {
    "epoch": 1887,
    "batch_idx": 96,
    "loss": 0.2200789749622345,
    "learning_rate": 7.855902796595087e-06,
    "timestep_mean": 524.25,
    "epsilon_norm": 6.987268924713135,
    "prediction_norm": 5.987680435180664,
    "timestamp": "2025-07-14T12:34:41.581056"
  },
  {
    "epoch": 1887,
    "batch_idx": 128,
    "loss": 0.2248396873474121,
    "learning_rate": 7.855902796595087e-06,
    "timestep_mean": 461.0,
    "epsilon_norm": 6.695306301116943,
    "prediction_norm": 6.22911262512207,
    "timestamp": "2025-07-14T12:34:41.586177"
  },
  {
    "epoch": 1887,
    "batch_idx": 160,
    "loss": 0.4240841269493103,
    "learning_rate": 7.855902796595087e-06,
    "timestep_mean": 541.9375,
    "epsilon_norm": 9.353983879089355,
    "prediction_norm": 7.04159688949585,
    "timestamp": "2025-07-14T12:34:41.591044"
  },
  {
    "epoch": 1887,
    "batch_idx": 192,
    "loss": 0.059164948761463165,
    "learning_rate": 7.855902796595087e-06,
    "timestep_mean": 526.0,
    "epsilon_norm": 4.363019943237305,
    "prediction_norm": 4.1225433349609375,
    "timestamp": "2025-07-14T12:34:41.596012"
  },
  {
    "epoch": 1888,
    "batch_idx": 0,
    "loss": 0.2870473563671112,
    "learning_rate": 7.717832735397329e-06,
    "timestep_mean": 572.1875,
    "epsilon_norm": 8.231761932373047,
    "prediction_norm": 7.380682945251465,
    "timestamp": "2025-07-14T12:34:41.600914"
  },
  {
    "epoch": 1888,
    "batch_idx": 32,
    "loss": 0.3773260712623596,
    "learning_rate": 7.717832735397329e-06,
    "timestep_mean": 476.90625,
    "epsilon_norm": 8.4338960647583,
    "prediction_norm": 6.584447383880615,
    "timestamp": "2025-07-14T12:34:41.605801"
  },
  {
    "epoch": 1888,
    "batch_idx": 64,
    "loss": 0.43266284465789795,
    "learning_rate": 7.717832735397329e-06,
    "timestep_mean": 484.21875,
    "epsilon_norm": 9.630478858947754,
    "prediction_norm": 7.904688835144043,
    "timestamp": "2025-07-14T12:34:41.610697"
  },
  {
    "epoch": 1888,
    "batch_idx": 96,
    "loss": 0.23238058388233185,
    "learning_rate": 7.717832735397329e-06,
    "timestep_mean": 548.5625,
    "epsilon_norm": 8.314435005187988,
    "prediction_norm": 7.498410224914551,
    "timestamp": "2025-07-14T12:34:41.615572"
  },
  {
    "epoch": 1888,
    "batch_idx": 128,
    "loss": 0.2083040177822113,
    "learning_rate": 7.717832735397329e-06,
    "timestep_mean": 393.71875,
    "epsilon_norm": 7.576087474822998,
    "prediction_norm": 7.200063228607178,
    "timestamp": "2025-07-14T12:34:41.620824"
  },
  {
    "epoch": 1888,
    "batch_idx": 160,
    "loss": 0.2961732745170593,
    "learning_rate": 7.717832735397329e-06,
    "timestep_mean": 401.5625,
    "epsilon_norm": 7.299501419067383,
    "prediction_norm": 5.335681438446045,
    "timestamp": "2025-07-14T12:34:41.625768"
  },
  {
    "epoch": 1888,
    "batch_idx": 192,
    "loss": 0.41018781065940857,
    "learning_rate": 7.717832735397329e-06,
    "timestep_mean": 471.25,
    "epsilon_norm": 4.136322498321533,
    "prediction_norm": 2.6082515716552734,
    "timestamp": "2025-07-14T12:34:41.630730"
  },
  {
    "epoch": 1889,
    "batch_idx": 0,
    "loss": 0.4389914274215698,
    "learning_rate": 7.5809773315109124e-06,
    "timestep_mean": 502.65625,
    "epsilon_norm": 8.734020233154297,
    "prediction_norm": 7.415491580963135,
    "timestamp": "2025-07-14T12:34:41.635595"
  },
  {
    "epoch": 1889,
    "batch_idx": 32,
    "loss": 0.21191740036010742,
    "learning_rate": 7.5809773315109124e-06,
    "timestep_mean": 549.84375,
    "epsilon_norm": 7.996020793914795,
    "prediction_norm": 7.216774940490723,
    "timestamp": "2025-07-14T12:34:41.640959"
  },
  {
    "epoch": 1889,
    "batch_idx": 64,
    "loss": 0.19451621174812317,
    "learning_rate": 7.5809773315109124e-06,
    "timestep_mean": 501.65625,
    "epsilon_norm": 8.242616653442383,
    "prediction_norm": 7.170653820037842,
    "timestamp": "2025-07-14T12:34:41.648903"
  },
  {
    "epoch": 1889,
    "batch_idx": 96,
    "loss": 0.212215855717659,
    "learning_rate": 7.5809773315109124e-06,
    "timestep_mean": 448.46875,
    "epsilon_norm": 7.624638557434082,
    "prediction_norm": 6.5349860191345215,
    "timestamp": "2025-07-14T12:34:41.655890"
  },
  {
    "epoch": 1889,
    "batch_idx": 128,
    "loss": 0.35882246494293213,
    "learning_rate": 7.5809773315109124e-06,
    "timestep_mean": 520.96875,
    "epsilon_norm": 7.9368815422058105,
    "prediction_norm": 6.62467098236084,
    "timestamp": "2025-07-14T12:34:41.663041"
  },
  {
    "epoch": 1889,
    "batch_idx": 160,
    "loss": 0.28359246253967285,
    "learning_rate": 7.5809773315109124e-06,
    "timestep_mean": 519.0625,
    "epsilon_norm": 8.413521766662598,
    "prediction_norm": 7.481125354766846,
    "timestamp": "2025-07-14T12:34:41.668871"
  },
  {
    "epoch": 1889,
    "batch_idx": 192,
    "loss": 0.13033027946949005,
    "learning_rate": 7.5809773315109124e-06,
    "timestep_mean": 630.75,
    "epsilon_norm": 4.667896270751953,
    "prediction_norm": 4.542691707611084,
    "timestamp": "2025-07-14T12:34:41.676090"
  },
  {
    "epoch": 1890,
    "batch_idx": 0,
    "loss": 0.4185318946838379,
    "learning_rate": 7.44533692261306e-06,
    "timestep_mean": 415.5625,
    "epsilon_norm": 8.344769477844238,
    "prediction_norm": 6.306640148162842,
    "timestamp": "2025-07-14T12:34:41.684689"
  },
  {
    "epoch": 1890,
    "batch_idx": 32,
    "loss": 0.14990833401679993,
    "learning_rate": 7.44533692261306e-06,
    "timestep_mean": 529.75,
    "epsilon_norm": 6.851566314697266,
    "prediction_norm": 6.633533477783203,
    "timestamp": "2025-07-14T12:34:41.692843"
  },
  {
    "epoch": 1890,
    "batch_idx": 64,
    "loss": 0.39706820249557495,
    "learning_rate": 7.44533692261306e-06,
    "timestep_mean": 535.625,
    "epsilon_norm": 7.581671237945557,
    "prediction_norm": 6.2572102546691895,
    "timestamp": "2025-07-14T12:34:41.701122"
  },
  {
    "epoch": 1890,
    "batch_idx": 96,
    "loss": 0.2769193649291992,
    "learning_rate": 7.44533692261306e-06,
    "timestep_mean": 464.21875,
    "epsilon_norm": 8.158329963684082,
    "prediction_norm": 6.7659807205200195,
    "timestamp": "2025-07-14T12:34:41.709563"
  },
  {
    "epoch": 1890,
    "batch_idx": 128,
    "loss": 0.27425724267959595,
    "learning_rate": 7.44533692261306e-06,
    "timestep_mean": 417.21875,
    "epsilon_norm": 7.672962188720703,
    "prediction_norm": 6.343945026397705,
    "timestamp": "2025-07-14T12:34:41.718325"
  },
  {
    "epoch": 1890,
    "batch_idx": 160,
    "loss": 0.18152230978012085,
    "learning_rate": 7.44533692261306e-06,
    "timestep_mean": 526.1875,
    "epsilon_norm": 6.892884254455566,
    "prediction_norm": 6.048654556274414,
    "timestamp": "2025-07-14T12:34:41.726839"
  },
  {
    "epoch": 1890,
    "batch_idx": 192,
    "loss": 0.065707728266716,
    "learning_rate": 7.44533692261306e-06,
    "timestep_mean": 641.125,
    "epsilon_norm": 3.324619770050049,
    "prediction_norm": 3.3428142070770264,
    "timestamp": "2025-07-14T12:34:41.735107"
  },
  {
    "epoch": 1891,
    "batch_idx": 0,
    "loss": 0.24958613514900208,
    "learning_rate": 7.310911843382888e-06,
    "timestep_mean": 483.90625,
    "epsilon_norm": 7.329514026641846,
    "prediction_norm": 6.082942485809326,
    "timestamp": "2025-07-14T12:34:41.743404"
  },
  {
    "epoch": 1891,
    "batch_idx": 32,
    "loss": 0.22605209052562714,
    "learning_rate": 7.310911843382888e-06,
    "timestep_mean": 463.53125,
    "epsilon_norm": 7.189347267150879,
    "prediction_norm": 5.650577068328857,
    "timestamp": "2025-07-14T12:34:41.748754"
  },
  {
    "epoch": 1891,
    "batch_idx": 64,
    "loss": 0.24871915578842163,
    "learning_rate": 7.310911843382888e-06,
    "timestep_mean": 556.21875,
    "epsilon_norm": 9.091727256774902,
    "prediction_norm": 8.322728157043457,
    "timestamp": "2025-07-14T12:34:41.753757"
  },
  {
    "epoch": 1891,
    "batch_idx": 96,
    "loss": 0.5565429329872131,
    "learning_rate": 7.310911843382888e-06,
    "timestep_mean": 383.1875,
    "epsilon_norm": 9.38929271697998,
    "prediction_norm": 6.3275957107543945,
    "timestamp": "2025-07-14T12:34:41.758913"
  },
  {
    "epoch": 1891,
    "batch_idx": 128,
    "loss": 0.27194032073020935,
    "learning_rate": 7.310911843382888e-06,
    "timestep_mean": 577.875,
    "epsilon_norm": 6.7045488357543945,
    "prediction_norm": 5.56247615814209,
    "timestamp": "2025-07-14T12:34:41.763836"
  },
  {
    "epoch": 1891,
    "batch_idx": 160,
    "loss": 0.2577720880508423,
    "learning_rate": 7.310911843382888e-06,
    "timestep_mean": 526.0,
    "epsilon_norm": 8.43803596496582,
    "prediction_norm": 6.773374080657959,
    "timestamp": "2025-07-14T12:34:41.768716"
  },
  {
    "epoch": 1891,
    "batch_idx": 192,
    "loss": 0.23873601853847504,
    "learning_rate": 7.310911843382888e-06,
    "timestep_mean": 379.5,
    "epsilon_norm": 3.460481643676758,
    "prediction_norm": 3.1120235919952393,
    "timestamp": "2025-07-14T12:34:41.773668"
  },
  {
    "epoch": 1892,
    "batch_idx": 0,
    "loss": 0.18017686903476715,
    "learning_rate": 7.17770242550097e-06,
    "timestep_mean": 510.125,
    "epsilon_norm": 7.360949993133545,
    "prediction_norm": 6.07776403427124,
    "timestamp": "2025-07-14T12:34:41.778769"
  },
  {
    "epoch": 1892,
    "batch_idx": 32,
    "loss": 0.20621493458747864,
    "learning_rate": 7.17770242550097e-06,
    "timestep_mean": 471.0625,
    "epsilon_norm": 8.074825286865234,
    "prediction_norm": 6.949571132659912,
    "timestamp": "2025-07-14T12:34:41.783753"
  },
  {
    "epoch": 1892,
    "batch_idx": 64,
    "loss": 0.18758782744407654,
    "learning_rate": 7.17770242550097e-06,
    "timestep_mean": 474.375,
    "epsilon_norm": 7.677258014678955,
    "prediction_norm": 6.708279609680176,
    "timestamp": "2025-07-14T12:34:41.788631"
  },
  {
    "epoch": 1892,
    "batch_idx": 96,
    "loss": 0.270868182182312,
    "learning_rate": 7.17770242550097e-06,
    "timestep_mean": 430.3125,
    "epsilon_norm": 8.570021629333496,
    "prediction_norm": 7.053591728210449,
    "timestamp": "2025-07-14T12:34:41.793636"
  },
  {
    "epoch": 1892,
    "batch_idx": 128,
    "loss": 0.08890047669410706,
    "learning_rate": 7.17770242550097e-06,
    "timestep_mean": 531.40625,
    "epsilon_norm": 6.856542110443115,
    "prediction_norm": 7.39221715927124,
    "timestamp": "2025-07-14T12:34:41.798587"
  },
  {
    "epoch": 1892,
    "batch_idx": 160,
    "loss": 0.19485163688659668,
    "learning_rate": 7.17770242550097e-06,
    "timestep_mean": 586.84375,
    "epsilon_norm": 8.993791580200195,
    "prediction_norm": 7.899975299835205,
    "timestamp": "2025-07-14T12:34:41.803688"
  },
  {
    "epoch": 1892,
    "batch_idx": 192,
    "loss": 0.273555725812912,
    "learning_rate": 7.17770242550097e-06,
    "timestep_mean": 476.625,
    "epsilon_norm": 4.109697341918945,
    "prediction_norm": 3.128185749053955,
    "timestamp": "2025-07-14T12:34:41.809039"
  },
  {
    "epoch": 1893,
    "batch_idx": 0,
    "loss": 0.2362539917230606,
    "learning_rate": 7.0457089976483305e-06,
    "timestep_mean": 466.0,
    "epsilon_norm": 8.845440864562988,
    "prediction_norm": 7.311246395111084,
    "timestamp": "2025-07-14T12:34:41.814344"
  },
  {
    "epoch": 1893,
    "batch_idx": 32,
    "loss": 0.13301169872283936,
    "learning_rate": 7.0457089976483305e-06,
    "timestep_mean": 550.25,
    "epsilon_norm": 7.396791934967041,
    "prediction_norm": 7.022275447845459,
    "timestamp": "2025-07-14T12:34:41.819587"
  },
  {
    "epoch": 1893,
    "batch_idx": 64,
    "loss": 0.3458700180053711,
    "learning_rate": 7.0457089976483305e-06,
    "timestep_mean": 430.21875,
    "epsilon_norm": 7.050683498382568,
    "prediction_norm": 5.495182514190674,
    "timestamp": "2025-07-14T12:34:41.824635"
  },
  {
    "epoch": 1893,
    "batch_idx": 96,
    "loss": 0.29029756784439087,
    "learning_rate": 7.0457089976483305e-06,
    "timestep_mean": 475.21875,
    "epsilon_norm": 8.839693069458008,
    "prediction_norm": 7.458855628967285,
    "timestamp": "2025-07-14T12:34:41.829998"
  },
  {
    "epoch": 1893,
    "batch_idx": 128,
    "loss": 0.21087849140167236,
    "learning_rate": 7.0457089976483305e-06,
    "timestep_mean": 498.59375,
    "epsilon_norm": 7.454607963562012,
    "prediction_norm": 6.493418216705322,
    "timestamp": "2025-07-14T12:34:41.835100"
  },
  {
    "epoch": 1893,
    "batch_idx": 160,
    "loss": 0.19485829770565033,
    "learning_rate": 7.0457089976483305e-06,
    "timestep_mean": 536.1875,
    "epsilon_norm": 7.245543479919434,
    "prediction_norm": 6.6657586097717285,
    "timestamp": "2025-07-14T12:34:41.840189"
  },
  {
    "epoch": 1893,
    "batch_idx": 192,
    "loss": 0.30433595180511475,
    "learning_rate": 7.0457089976483305e-06,
    "timestep_mean": 400.875,
    "epsilon_norm": 3.948014259338379,
    "prediction_norm": 3.5137038230895996,
    "timestamp": "2025-07-14T12:34:41.845496"
  },
  {
    "epoch": 1894,
    "batch_idx": 0,
    "loss": 0.13578060269355774,
    "learning_rate": 6.914931885505565e-06,
    "timestep_mean": 531.28125,
    "epsilon_norm": 8.679097175598145,
    "prediction_norm": 8.18687915802002,
    "timestamp": "2025-07-14T12:34:41.850784"
  },
  {
    "epoch": 1894,
    "batch_idx": 32,
    "loss": 0.21182671189308167,
    "learning_rate": 6.914931885505565e-06,
    "timestep_mean": 565.125,
    "epsilon_norm": 8.39451789855957,
    "prediction_norm": 7.777559280395508,
    "timestamp": "2025-07-14T12:34:41.856030"
  },
  {
    "epoch": 1894,
    "batch_idx": 64,
    "loss": 0.37095165252685547,
    "learning_rate": 6.914931885505565e-06,
    "timestep_mean": 511.15625,
    "epsilon_norm": 7.374745845794678,
    "prediction_norm": 5.485284805297852,
    "timestamp": "2025-07-14T12:34:41.861187"
  },
  {
    "epoch": 1894,
    "batch_idx": 96,
    "loss": 0.3925844430923462,
    "learning_rate": 6.914931885505565e-06,
    "timestep_mean": 442.6875,
    "epsilon_norm": 8.104804992675781,
    "prediction_norm": 6.550858020782471,
    "timestamp": "2025-07-14T12:34:41.866407"
  },
  {
    "epoch": 1894,
    "batch_idx": 128,
    "loss": 0.33095771074295044,
    "learning_rate": 6.914931885505565e-06,
    "timestep_mean": 412.75,
    "epsilon_norm": 7.738679885864258,
    "prediction_norm": 5.50143575668335,
    "timestamp": "2025-07-14T12:34:41.871762"
  },
  {
    "epoch": 1894,
    "batch_idx": 160,
    "loss": 0.3212004601955414,
    "learning_rate": 6.914931885505565e-06,
    "timestep_mean": 500.03125,
    "epsilon_norm": 7.808022975921631,
    "prediction_norm": 6.481400489807129,
    "timestamp": "2025-07-14T12:34:41.876933"
  },
  {
    "epoch": 1894,
    "batch_idx": 192,
    "loss": 0.39674413204193115,
    "learning_rate": 6.914931885505565e-06,
    "timestep_mean": 402.625,
    "epsilon_norm": 3.6968820095062256,
    "prediction_norm": 2.595134973526001,
    "timestamp": "2025-07-14T12:34:41.882123"
  },
  {
    "epoch": 1895,
    "batch_idx": 0,
    "loss": 0.25795984268188477,
    "learning_rate": 6.785371411752278e-06,
    "timestep_mean": 463.21875,
    "epsilon_norm": 8.650259017944336,
    "prediction_norm": 7.430866718292236,
    "timestamp": "2025-07-14T12:34:41.887416"
  },
  {
    "epoch": 1895,
    "batch_idx": 32,
    "loss": 0.48771384358406067,
    "learning_rate": 6.785371411752278e-06,
    "timestep_mean": 433.65625,
    "epsilon_norm": 7.742679595947266,
    "prediction_norm": 5.972393035888672,
    "timestamp": "2025-07-14T12:34:41.892623"
  },
  {
    "epoch": 1895,
    "batch_idx": 64,
    "loss": 0.3533897399902344,
    "learning_rate": 6.785371411752278e-06,
    "timestep_mean": 525.6875,
    "epsilon_norm": 8.204599380493164,
    "prediction_norm": 6.814288139343262,
    "timestamp": "2025-07-14T12:34:41.897928"
  },
  {
    "epoch": 1895,
    "batch_idx": 96,
    "loss": 0.15258735418319702,
    "learning_rate": 6.785371411752278e-06,
    "timestep_mean": 500.8125,
    "epsilon_norm": 6.955279350280762,
    "prediction_norm": 5.882668495178223,
    "timestamp": "2025-07-14T12:34:41.903358"
  },
  {
    "epoch": 1895,
    "batch_idx": 128,
    "loss": 0.3334065079689026,
    "learning_rate": 6.785371411752278e-06,
    "timestep_mean": 451.6875,
    "epsilon_norm": 7.838384628295898,
    "prediction_norm": 6.6807098388671875,
    "timestamp": "2025-07-14T12:34:41.908665"
  },
  {
    "epoch": 1895,
    "batch_idx": 160,
    "loss": 0.33139199018478394,
    "learning_rate": 6.785371411752278e-06,
    "timestep_mean": 549.59375,
    "epsilon_norm": 9.577784538269043,
    "prediction_norm": 8.084513664245605,
    "timestamp": "2025-07-14T12:34:41.913923"
  },
  {
    "epoch": 1895,
    "batch_idx": 192,
    "loss": 0.35746535658836365,
    "learning_rate": 6.785371411752278e-06,
    "timestep_mean": 528.875,
    "epsilon_norm": 5.412156105041504,
    "prediction_norm": 4.411736488342285,
    "timestamp": "2025-07-14T12:34:41.919232"
  },
  {
    "epoch": 1896,
    "batch_idx": 0,
    "loss": 0.33902156352996826,
    "learning_rate": 6.657027896065976e-06,
    "timestep_mean": 459.03125,
    "epsilon_norm": 7.549813270568848,
    "prediction_norm": 6.101493835449219,
    "timestamp": "2025-07-14T12:34:41.924449"
  },
  {
    "epoch": 1896,
    "batch_idx": 32,
    "loss": 0.5303951501846313,
    "learning_rate": 6.657027896065976e-06,
    "timestep_mean": 484.78125,
    "epsilon_norm": 8.518925666809082,
    "prediction_norm": 6.3484673500061035,
    "timestamp": "2025-07-14T12:34:41.929705"
  },
  {
    "epoch": 1896,
    "batch_idx": 64,
    "loss": 0.2752467095851898,
    "learning_rate": 6.657027896065976e-06,
    "timestep_mean": 455.875,
    "epsilon_norm": 7.604826927185059,
    "prediction_norm": 6.3166375160217285,
    "timestamp": "2025-07-14T12:34:41.934870"
  },
  {
    "epoch": 1896,
    "batch_idx": 96,
    "loss": 0.2754451334476471,
    "learning_rate": 6.657027896065976e-06,
    "timestep_mean": 577.25,
    "epsilon_norm": 7.653967380523682,
    "prediction_norm": 6.2450666427612305,
    "timestamp": "2025-07-14T12:34:41.940195"
  },
  {
    "epoch": 1896,
    "batch_idx": 128,
    "loss": 0.2871997356414795,
    "learning_rate": 6.657027896065976e-06,
    "timestep_mean": 440.28125,
    "epsilon_norm": 7.858771800994873,
    "prediction_norm": 6.875267505645752,
    "timestamp": "2025-07-14T12:34:41.945603"
  },
  {
    "epoch": 1896,
    "batch_idx": 160,
    "loss": 0.39251065254211426,
    "learning_rate": 6.657027896065976e-06,
    "timestep_mean": 608.0625,
    "epsilon_norm": 8.566692352294922,
    "prediction_norm": 7.260283470153809,
    "timestamp": "2025-07-14T12:34:41.950920"
  },
  {
    "epoch": 1896,
    "batch_idx": 192,
    "loss": 0.47499099373817444,
    "learning_rate": 6.657027896065976e-06,
    "timestep_mean": 391.625,
    "epsilon_norm": 3.662409543991089,
    "prediction_norm": 2.76499605178833,
    "timestamp": "2025-07-14T12:34:41.956064"
  },
  {
    "epoch": 1897,
    "batch_idx": 0,
    "loss": 0.14895419776439667,
    "learning_rate": 6.5299016551215676e-06,
    "timestep_mean": 516.8125,
    "epsilon_norm": 7.997678756713867,
    "prediction_norm": 7.480092525482178,
    "timestamp": "2025-07-14T12:34:41.961394"
  },
  {
    "epoch": 1897,
    "batch_idx": 32,
    "loss": 0.19151188433170319,
    "learning_rate": 6.5299016551215676e-06,
    "timestep_mean": 483.65625,
    "epsilon_norm": 6.880548477172852,
    "prediction_norm": 6.209927082061768,
    "timestamp": "2025-07-14T12:34:41.966596"
  },
  {
    "epoch": 1897,
    "batch_idx": 64,
    "loss": 0.11203092336654663,
    "learning_rate": 6.5299016551215676e-06,
    "timestep_mean": 658.875,
    "epsilon_norm": 7.13536262512207,
    "prediction_norm": 6.621290683746338,
    "timestamp": "2025-07-14T12:34:41.972007"
  },
  {
    "epoch": 1897,
    "batch_idx": 96,
    "loss": 0.44073909521102905,
    "learning_rate": 6.5299016551215676e-06,
    "timestep_mean": 443.46875,
    "epsilon_norm": 8.302855491638184,
    "prediction_norm": 6.663424015045166,
    "timestamp": "2025-07-14T12:34:41.977328"
  },
  {
    "epoch": 1897,
    "batch_idx": 128,
    "loss": 0.29344642162323,
    "learning_rate": 6.5299016551215676e-06,
    "timestep_mean": 512.59375,
    "epsilon_norm": 7.474947452545166,
    "prediction_norm": 5.863072395324707,
    "timestamp": "2025-07-14T12:34:41.982608"
  },
  {
    "epoch": 1897,
    "batch_idx": 160,
    "loss": 0.34725669026374817,
    "learning_rate": 6.5299016551215676e-06,
    "timestep_mean": 484.75,
    "epsilon_norm": 7.366328716278076,
    "prediction_norm": 6.792011260986328,
    "timestamp": "2025-07-14T12:34:41.987841"
  },
  {
    "epoch": 1897,
    "batch_idx": 192,
    "loss": 0.08958449959754944,
    "learning_rate": 6.5299016551215676e-06,
    "timestep_mean": 620.5,
    "epsilon_norm": 3.8606746196746826,
    "prediction_norm": 3.418029546737671,
    "timestamp": "2025-07-14T12:34:41.993448"
  },
  {
    "epoch": 1898,
    "batch_idx": 0,
    "loss": 0.31090760231018066,
    "learning_rate": 6.40399300259042e-06,
    "timestep_mean": 523.34375,
    "epsilon_norm": 8.808966636657715,
    "prediction_norm": 7.647948741912842,
    "timestamp": "2025-07-14T12:34:41.999012"
  },
  {
    "epoch": 1898,
    "batch_idx": 32,
    "loss": 0.36092257499694824,
    "learning_rate": 6.40399300259042e-06,
    "timestep_mean": 490.46875,
    "epsilon_norm": 8.610419273376465,
    "prediction_norm": 6.948950290679932,
    "timestamp": "2025-07-14T12:34:42.004438"
  },
  {
    "epoch": 1898,
    "batch_idx": 64,
    "loss": 0.13280627131462097,
    "learning_rate": 6.40399300259042e-06,
    "timestep_mean": 530.71875,
    "epsilon_norm": 6.698441028594971,
    "prediction_norm": 6.234650135040283,
    "timestamp": "2025-07-14T12:34:42.009657"
  },
  {
    "epoch": 1898,
    "batch_idx": 96,
    "loss": 0.35191610455513,
    "learning_rate": 6.40399300259042e-06,
    "timestep_mean": 515.125,
    "epsilon_norm": 7.475752830505371,
    "prediction_norm": 6.161675453186035,
    "timestamp": "2025-07-14T12:34:42.014922"
  },
  {
    "epoch": 1898,
    "batch_idx": 128,
    "loss": 0.24215425550937653,
    "learning_rate": 6.40399300259042e-06,
    "timestep_mean": 425.65625,
    "epsilon_norm": 7.315523624420166,
    "prediction_norm": 6.3411078453063965,
    "timestamp": "2025-07-14T12:34:42.020223"
  },
  {
    "epoch": 1898,
    "batch_idx": 160,
    "loss": 0.38529255986213684,
    "learning_rate": 6.40399300259042e-06,
    "timestep_mean": 494.875,
    "epsilon_norm": 9.563043594360352,
    "prediction_norm": 8.20794677734375,
    "timestamp": "2025-07-14T12:34:42.025694"
  },
  {
    "epoch": 1898,
    "batch_idx": 192,
    "loss": 0.05988935008645058,
    "learning_rate": 6.40399300259042e-06,
    "timestep_mean": 662.5,
    "epsilon_norm": 3.6498844623565674,
    "prediction_norm": 3.91825532913208,
    "timestamp": "2025-07-14T12:34:42.030983"
  },
  {
    "epoch": 1899,
    "batch_idx": 0,
    "loss": 0.3141797184944153,
    "learning_rate": 6.27930224913958e-06,
    "timestep_mean": 523.6875,
    "epsilon_norm": 7.486595630645752,
    "prediction_norm": 6.797840595245361,
    "timestamp": "2025-07-14T12:34:42.036309"
  },
  {
    "epoch": 1899,
    "batch_idx": 32,
    "loss": 0.17960435152053833,
    "learning_rate": 6.27930224913958e-06,
    "timestep_mean": 550.84375,
    "epsilon_norm": 8.045489311218262,
    "prediction_norm": 7.825177192687988,
    "timestamp": "2025-07-14T12:34:42.041916"
  },
  {
    "epoch": 1899,
    "batch_idx": 64,
    "loss": 0.3213236629962921,
    "learning_rate": 6.27930224913958e-06,
    "timestep_mean": 486.75,
    "epsilon_norm": 7.79669189453125,
    "prediction_norm": 6.7921857833862305,
    "timestamp": "2025-07-14T12:34:42.047114"
  },
  {
    "epoch": 1899,
    "batch_idx": 96,
    "loss": 0.26812487840652466,
    "learning_rate": 6.27930224913958e-06,
    "timestep_mean": 553.5,
    "epsilon_norm": 9.286873817443848,
    "prediction_norm": 8.018320083618164,
    "timestamp": "2025-07-14T12:34:42.052347"
  },
  {
    "epoch": 1899,
    "batch_idx": 128,
    "loss": 0.2091996967792511,
    "learning_rate": 6.27930224913958e-06,
    "timestep_mean": 541.78125,
    "epsilon_norm": 7.601563930511475,
    "prediction_norm": 6.687368869781494,
    "timestamp": "2025-07-14T12:34:42.057698"
  },
  {
    "epoch": 1899,
    "batch_idx": 160,
    "loss": 0.2794865369796753,
    "learning_rate": 6.27930224913958e-06,
    "timestep_mean": 458.5,
    "epsilon_norm": 8.307843208312988,
    "prediction_norm": 7.020252227783203,
    "timestamp": "2025-07-14T12:34:42.062910"
  },
  {
    "epoch": 1899,
    "batch_idx": 192,
    "loss": 0.14691373705863953,
    "learning_rate": 6.27930224913958e-06,
    "timestep_mean": 666.625,
    "epsilon_norm": 3.7596793174743652,
    "prediction_norm": 3.2461764812469482,
    "timestamp": "2025-07-14T12:34:42.068131"
  },
  {
    "epoch": 1900,
    "batch_idx": 0,
    "loss": 0.16363824903964996,
    "learning_rate": 6.155829702431166e-06,
    "timestep_mean": 543.4375,
    "epsilon_norm": 7.801756858825684,
    "prediction_norm": 7.197507381439209,
    "timestamp": "2025-07-14T12:34:42.073447"
  },
  {
    "epoch": 1900,
    "batch_idx": 32,
    "loss": 0.3235752284526825,
    "learning_rate": 6.155829702431166e-06,
    "timestep_mean": 396.0625,
    "epsilon_norm": 8.345961570739746,
    "prediction_norm": 6.663557529449463,
    "timestamp": "2025-07-14T12:34:42.078777"
  },
  {
    "epoch": 1900,
    "batch_idx": 64,
    "loss": 0.30712711811065674,
    "learning_rate": 6.155829702431166e-06,
    "timestep_mean": 596.25,
    "epsilon_norm": 9.898770332336426,
    "prediction_norm": 8.722530364990234,
    "timestamp": "2025-07-14T12:34:42.084026"
  },
  {
    "epoch": 1900,
    "batch_idx": 96,
    "loss": 0.17648489773273468,
    "learning_rate": 6.155829702431166e-06,
    "timestep_mean": 541.0625,
    "epsilon_norm": 8.879485130310059,
    "prediction_norm": 8.297233581542969,
    "timestamp": "2025-07-14T12:34:42.089154"
  },
  {
    "epoch": 1900,
    "batch_idx": 128,
    "loss": 0.26919907331466675,
    "learning_rate": 6.155829702431166e-06,
    "timestep_mean": 440.15625,
    "epsilon_norm": 7.830406665802002,
    "prediction_norm": 6.52759313583374,
    "timestamp": "2025-07-14T12:34:42.094412"
  },
  {
    "epoch": 1900,
    "batch_idx": 160,
    "loss": 0.26542165875434875,
    "learning_rate": 6.155829702431166e-06,
    "timestep_mean": 543.21875,
    "epsilon_norm": 8.204425811767578,
    "prediction_norm": 7.433422088623047,
    "timestamp": "2025-07-14T12:34:42.099482"
  },
  {
    "epoch": 1900,
    "batch_idx": 192,
    "loss": 0.20986081659793854,
    "learning_rate": 6.155829702431166e-06,
    "timestep_mean": 614.5,
    "epsilon_norm": 4.52095890045166,
    "prediction_norm": 3.942950487136841,
    "timestamp": "2025-07-14T12:34:42.104458"
  },
  {
    "epoch": 1901,
    "batch_idx": 0,
    "loss": 0.25799012184143066,
    "learning_rate": 6.033575667121311e-06,
    "timestep_mean": 493.6875,
    "epsilon_norm": 7.049064636230469,
    "prediction_norm": 5.539182662963867,
    "timestamp": "2025-07-14T12:34:42.109421"
  },
  {
    "epoch": 1901,
    "batch_idx": 32,
    "loss": 0.36755257844924927,
    "learning_rate": 6.033575667121311e-06,
    "timestep_mean": 470.46875,
    "epsilon_norm": 9.133746147155762,
    "prediction_norm": 7.8147196769714355,
    "timestamp": "2025-07-14T12:34:42.114371"
  },
  {
    "epoch": 1901,
    "batch_idx": 64,
    "loss": 0.3006249666213989,
    "learning_rate": 6.033575667121311e-06,
    "timestep_mean": 519.25,
    "epsilon_norm": 7.726807594299316,
    "prediction_norm": 6.621826171875,
    "timestamp": "2025-07-14T12:34:42.119374"
  },
  {
    "epoch": 1901,
    "batch_idx": 96,
    "loss": 0.4575330913066864,
    "learning_rate": 6.033575667121311e-06,
    "timestep_mean": 429.78125,
    "epsilon_norm": 8.776663780212402,
    "prediction_norm": 6.668286323547363,
    "timestamp": "2025-07-14T12:34:42.124278"
  },
  {
    "epoch": 1901,
    "batch_idx": 128,
    "loss": 0.28661438822746277,
    "learning_rate": 6.033575667121311e-06,
    "timestep_mean": 460.4375,
    "epsilon_norm": 9.012290954589844,
    "prediction_norm": 7.802873134613037,
    "timestamp": "2025-07-14T12:34:42.129437"
  },
  {
    "epoch": 1901,
    "batch_idx": 160,
    "loss": 0.4898865818977356,
    "learning_rate": 6.033575667121311e-06,
    "timestep_mean": 464.5,
    "epsilon_norm": 9.104433059692383,
    "prediction_norm": 7.535834312438965,
    "timestamp": "2025-07-14T12:34:42.136129"
  },
  {
    "epoch": 1901,
    "batch_idx": 192,
    "loss": 0.121254563331604,
    "learning_rate": 6.033575667121311e-06,
    "timestep_mean": 433.125,
    "epsilon_norm": 3.2918894290924072,
    "prediction_norm": 3.34667706489563,
    "timestamp": "2025-07-14T12:34:42.145748"
  },
  {
    "epoch": 1902,
    "batch_idx": 0,
    "loss": 0.3779411315917969,
    "learning_rate": 5.912540444859778e-06,
    "timestep_mean": 470.375,
    "epsilon_norm": 8.322596549987793,
    "prediction_norm": 7.101415157318115,
    "timestamp": "2025-07-14T12:34:42.152441"
  },
  {
    "epoch": 1902,
    "batch_idx": 32,
    "loss": 0.22488215565681458,
    "learning_rate": 5.912540444859778e-06,
    "timestep_mean": 502.15625,
    "epsilon_norm": 8.454890251159668,
    "prediction_norm": 7.069794178009033,
    "timestamp": "2025-07-14T12:34:42.162823"
  },
  {
    "epoch": 1902,
    "batch_idx": 64,
    "loss": 0.3810589611530304,
    "learning_rate": 5.912540444859778e-06,
    "timestep_mean": 523.875,
    "epsilon_norm": 7.260622501373291,
    "prediction_norm": 6.127142906188965,
    "timestamp": "2025-07-14T12:34:42.168536"
  },
  {
    "epoch": 1902,
    "batch_idx": 96,
    "loss": 0.18130230903625488,
    "learning_rate": 5.912540444859778e-06,
    "timestep_mean": 568.34375,
    "epsilon_norm": 7.106733798980713,
    "prediction_norm": 6.021853923797607,
    "timestamp": "2025-07-14T12:34:42.176516"
  },
  {
    "epoch": 1902,
    "batch_idx": 128,
    "loss": 0.3319169878959656,
    "learning_rate": 5.912540444859778e-06,
    "timestep_mean": 489.28125,
    "epsilon_norm": 9.22216796875,
    "prediction_norm": 8.067996978759766,
    "timestamp": "2025-07-14T12:34:42.184841"
  },
  {
    "epoch": 1902,
    "batch_idx": 160,
    "loss": 0.21878227591514587,
    "learning_rate": 5.912540444859778e-06,
    "timestep_mean": 554.03125,
    "epsilon_norm": 8.046357154846191,
    "prediction_norm": 7.270139217376709,
    "timestamp": "2025-07-14T12:34:42.193031"
  },
  {
    "epoch": 1902,
    "batch_idx": 192,
    "loss": 0.5418988466262817,
    "learning_rate": 5.912540444859778e-06,
    "timestep_mean": 571.375,
    "epsilon_norm": 4.681345462799072,
    "prediction_norm": 3.228053569793701,
    "timestamp": "2025-07-14T12:34:42.201532"
  },
  {
    "epoch": 1903,
    "batch_idx": 0,
    "loss": 0.20593610405921936,
    "learning_rate": 5.792724334288842e-06,
    "timestep_mean": 507.125,
    "epsilon_norm": 7.5958075523376465,
    "prediction_norm": 6.964325428009033,
    "timestamp": "2025-07-14T12:34:42.209830"
  },
  {
    "epoch": 1903,
    "batch_idx": 32,
    "loss": 0.3459465503692627,
    "learning_rate": 5.792724334288842e-06,
    "timestep_mean": 479.0625,
    "epsilon_norm": 8.279266357421875,
    "prediction_norm": 6.717780113220215,
    "timestamp": "2025-07-14T12:34:42.218363"
  },
  {
    "epoch": 1903,
    "batch_idx": 64,
    "loss": 0.19006115198135376,
    "learning_rate": 5.792724334288842e-06,
    "timestep_mean": 539.40625,
    "epsilon_norm": 7.960151672363281,
    "prediction_norm": 7.130730628967285,
    "timestamp": "2025-07-14T12:34:42.226530"
  },
  {
    "epoch": 1903,
    "batch_idx": 96,
    "loss": 0.24651861190795898,
    "learning_rate": 5.792724334288842e-06,
    "timestep_mean": 484.5625,
    "epsilon_norm": 7.626413345336914,
    "prediction_norm": 6.321782112121582,
    "timestamp": "2025-07-14T12:34:42.235091"
  },
  {
    "epoch": 1903,
    "batch_idx": 128,
    "loss": 0.12492148578166962,
    "learning_rate": 5.792724334288842e-06,
    "timestep_mean": 534.125,
    "epsilon_norm": 7.377566814422607,
    "prediction_norm": 7.141049385070801,
    "timestamp": "2025-07-14T12:34:42.243229"
  },
  {
    "epoch": 1903,
    "batch_idx": 160,
    "loss": 0.2782934904098511,
    "learning_rate": 5.792724334288842e-06,
    "timestep_mean": 506.65625,
    "epsilon_norm": 9.106588363647461,
    "prediction_norm": 7.762170314788818,
    "timestamp": "2025-07-14T12:34:42.248445"
  },
  {
    "epoch": 1903,
    "batch_idx": 192,
    "loss": 0.14261724054813385,
    "learning_rate": 5.792724334288842e-06,
    "timestep_mean": 472.125,
    "epsilon_norm": 3.5963563919067383,
    "prediction_norm": 3.6006176471710205,
    "timestamp": "2025-07-14T12:34:42.253521"
  },
  {
    "epoch": 1904,
    "batch_idx": 0,
    "loss": 0.31385964155197144,
    "learning_rate": 5.67412763104302e-06,
    "timestep_mean": 508.65625,
    "epsilon_norm": 8.6102294921875,
    "prediction_norm": 7.518414497375488,
    "timestamp": "2025-07-14T12:34:42.258517"
  },
  {
    "epoch": 1904,
    "batch_idx": 32,
    "loss": 0.20801609754562378,
    "learning_rate": 5.67412763104302e-06,
    "timestep_mean": 611.9375,
    "epsilon_norm": 7.552566051483154,
    "prediction_norm": 7.124107360839844,
    "timestamp": "2025-07-14T12:34:42.263474"
  },
  {
    "epoch": 1904,
    "batch_idx": 64,
    "loss": 0.4374379813671112,
    "learning_rate": 5.67412763104302e-06,
    "timestep_mean": 450.9375,
    "epsilon_norm": 8.493718147277832,
    "prediction_norm": 6.787600517272949,
    "timestamp": "2025-07-14T12:34:42.268408"
  },
  {
    "epoch": 1904,
    "batch_idx": 96,
    "loss": 0.5219046473503113,
    "learning_rate": 5.67412763104302e-06,
    "timestep_mean": 455.21875,
    "epsilon_norm": 9.007513046264648,
    "prediction_norm": 7.205790042877197,
    "timestamp": "2025-07-14T12:34:42.273371"
  },
  {
    "epoch": 1904,
    "batch_idx": 128,
    "loss": 0.2121160477399826,
    "learning_rate": 5.67412763104302e-06,
    "timestep_mean": 557.34375,
    "epsilon_norm": 8.58008861541748,
    "prediction_norm": 7.444865703582764,
    "timestamp": "2025-07-14T12:34:42.278314"
  },
  {
    "epoch": 1904,
    "batch_idx": 160,
    "loss": 0.10953167825937271,
    "learning_rate": 5.67412763104302e-06,
    "timestep_mean": 552.375,
    "epsilon_norm": 7.9298906326293945,
    "prediction_norm": 6.976239204406738,
    "timestamp": "2025-07-14T12:34:42.283126"
  },
  {
    "epoch": 1904,
    "batch_idx": 192,
    "loss": 1.1695328950881958,
    "learning_rate": 5.67412763104302e-06,
    "timestep_mean": 455.75,
    "epsilon_norm": 5.018395900726318,
    "prediction_norm": 2.8098948001861572,
    "timestamp": "2025-07-14T12:34:42.288147"
  },
  {
    "epoch": 1905,
    "batch_idx": 0,
    "loss": 0.35832035541534424,
    "learning_rate": 5.556750627747738e-06,
    "timestep_mean": 476.25,
    "epsilon_norm": 7.805478572845459,
    "prediction_norm": 5.924813747406006,
    "timestamp": "2025-07-14T12:34:42.293158"
  },
  {
    "epoch": 1905,
    "batch_idx": 32,
    "loss": 0.17743754386901855,
    "learning_rate": 5.556750627747738e-06,
    "timestep_mean": 477.96875,
    "epsilon_norm": 7.204558372497559,
    "prediction_norm": 6.381577491760254,
    "timestamp": "2025-07-14T12:34:42.298020"
  },
  {
    "epoch": 1905,
    "batch_idx": 64,
    "loss": 0.2608630061149597,
    "learning_rate": 5.556750627747738e-06,
    "timestep_mean": 530.0625,
    "epsilon_norm": 6.577473163604736,
    "prediction_norm": 5.912991046905518,
    "timestamp": "2025-07-14T12:34:42.302943"
  },
  {
    "epoch": 1905,
    "batch_idx": 96,
    "loss": 0.5102514624595642,
    "learning_rate": 5.556750627747738e-06,
    "timestep_mean": 342.9375,
    "epsilon_norm": 7.727057456970215,
    "prediction_norm": 5.389781951904297,
    "timestamp": "2025-07-14T12:34:42.307758"
  },
  {
    "epoch": 1905,
    "batch_idx": 128,
    "loss": 0.3607932925224304,
    "learning_rate": 5.556750627747738e-06,
    "timestep_mean": 443.875,
    "epsilon_norm": 8.330510139465332,
    "prediction_norm": 6.268224716186523,
    "timestamp": "2025-07-14T12:34:42.312708"
  },
  {
    "epoch": 1905,
    "batch_idx": 160,
    "loss": 0.2171674370765686,
    "learning_rate": 5.556750627747738e-06,
    "timestep_mean": 488.71875,
    "epsilon_norm": 7.906017780303955,
    "prediction_norm": 7.377419471740723,
    "timestamp": "2025-07-14T12:34:42.317709"
  },
  {
    "epoch": 1905,
    "batch_idx": 192,
    "loss": 0.6962167024612427,
    "learning_rate": 5.556750627747738e-06,
    "timestep_mean": 328.5,
    "epsilon_norm": 4.683807849884033,
    "prediction_norm": 3.1899147033691406,
    "timestamp": "2025-07-14T12:34:42.322679"
  },
  {
    "epoch": 1906,
    "batch_idx": 0,
    "loss": 0.2643622159957886,
    "learning_rate": 5.440593614019103e-06,
    "timestep_mean": 515.53125,
    "epsilon_norm": 8.016637802124023,
    "prediction_norm": 6.975675582885742,
    "timestamp": "2025-07-14T12:34:42.327669"
  },
  {
    "epoch": 1906,
    "batch_idx": 32,
    "loss": 0.2804332971572876,
    "learning_rate": 5.440593614019103e-06,
    "timestep_mean": 497.375,
    "epsilon_norm": 7.831543922424316,
    "prediction_norm": 6.229954242706299,
    "timestamp": "2025-07-14T12:34:42.332652"
  },
  {
    "epoch": 1906,
    "batch_idx": 64,
    "loss": 0.3912310302257538,
    "learning_rate": 5.440593614019103e-06,
    "timestep_mean": 461.78125,
    "epsilon_norm": 7.980278968811035,
    "prediction_norm": 6.534953594207764,
    "timestamp": "2025-07-14T12:34:42.337705"
  },
  {
    "epoch": 1906,
    "batch_idx": 96,
    "loss": 0.30729085206985474,
    "learning_rate": 5.440593614019103e-06,
    "timestep_mean": 507.3125,
    "epsilon_norm": 8.128609657287598,
    "prediction_norm": 7.089573860168457,
    "timestamp": "2025-07-14T12:34:42.342761"
  },
  {
    "epoch": 1906,
    "batch_idx": 128,
    "loss": 0.14573076367378235,
    "learning_rate": 5.440593614019103e-06,
    "timestep_mean": 432.84375,
    "epsilon_norm": 7.010446071624756,
    "prediction_norm": 5.924368858337402,
    "timestamp": "2025-07-14T12:34:42.347802"
  },
  {
    "epoch": 1906,
    "batch_idx": 160,
    "loss": 0.32967060804367065,
    "learning_rate": 5.440593614019103e-06,
    "timestep_mean": 559.375,
    "epsilon_norm": 8.558460235595703,
    "prediction_norm": 6.612114906311035,
    "timestamp": "2025-07-14T12:34:42.352832"
  },
  {
    "epoch": 1906,
    "batch_idx": 192,
    "loss": 0.3255382776260376,
    "learning_rate": 5.440593614019103e-06,
    "timestep_mean": 365.5,
    "epsilon_norm": 5.29570198059082,
    "prediction_norm": 3.6831111907958984,
    "timestamp": "2025-07-14T12:34:42.357893"
  },
  {
    "epoch": 1907,
    "batch_idx": 0,
    "loss": 0.3006526827812195,
    "learning_rate": 5.325656876463077e-06,
    "timestep_mean": 488.84375,
    "epsilon_norm": 7.967274188995361,
    "prediction_norm": 6.43173885345459,
    "timestamp": "2025-07-14T12:34:42.362972"
  },
  {
    "epoch": 1907,
    "batch_idx": 32,
    "loss": 0.3793262243270874,
    "learning_rate": 5.325656876463077e-06,
    "timestep_mean": 553.28125,
    "epsilon_norm": 8.716242790222168,
    "prediction_norm": 7.065490245819092,
    "timestamp": "2025-07-14T12:34:42.367933"
  },
  {
    "epoch": 1907,
    "batch_idx": 64,
    "loss": 0.5704323053359985,
    "learning_rate": 5.325656876463077e-06,
    "timestep_mean": 448.03125,
    "epsilon_norm": 9.350358009338379,
    "prediction_norm": 6.851324558258057,
    "timestamp": "2025-07-14T12:34:42.372942"
  },
  {
    "epoch": 1907,
    "batch_idx": 96,
    "loss": 0.2117086797952652,
    "learning_rate": 5.325656876463077e-06,
    "timestep_mean": 491.25,
    "epsilon_norm": 7.531093597412109,
    "prediction_norm": 6.659614562988281,
    "timestamp": "2025-07-14T12:34:42.377962"
  },
  {
    "epoch": 1907,
    "batch_idx": 128,
    "loss": 0.22653429210186005,
    "learning_rate": 5.325656876463077e-06,
    "timestep_mean": 539.90625,
    "epsilon_norm": 7.875741004943848,
    "prediction_norm": 7.00611686706543,
    "timestamp": "2025-07-14T12:34:42.382964"
  },
  {
    "epoch": 1907,
    "batch_idx": 160,
    "loss": 0.4475456476211548,
    "learning_rate": 5.325656876463077e-06,
    "timestep_mean": 384.8125,
    "epsilon_norm": 7.991042613983154,
    "prediction_norm": 6.357922077178955,
    "timestamp": "2025-07-14T12:34:42.388105"
  },
  {
    "epoch": 1907,
    "batch_idx": 192,
    "loss": 0.8200858235359192,
    "learning_rate": 5.325656876463077e-06,
    "timestep_mean": 429.875,
    "epsilon_norm": 4.2205810546875,
    "prediction_norm": 3.1794919967651367,
    "timestamp": "2025-07-14T12:34:42.393049"
  },
  {
    "epoch": 1908,
    "batch_idx": 0,
    "loss": 0.26023000478744507,
    "learning_rate": 5.211940698674529e-06,
    "timestep_mean": 486.90625,
    "epsilon_norm": 8.295284271240234,
    "prediction_norm": 7.036609649658203,
    "timestamp": "2025-07-14T12:34:42.398076"
  },
  {
    "epoch": 1908,
    "batch_idx": 32,
    "loss": 0.27887195348739624,
    "learning_rate": 5.211940698674529e-06,
    "timestep_mean": 500.90625,
    "epsilon_norm": 7.719479084014893,
    "prediction_norm": 6.539682865142822,
    "timestamp": "2025-07-14T12:34:42.403101"
  },
  {
    "epoch": 1908,
    "batch_idx": 64,
    "loss": 0.2471752017736435,
    "learning_rate": 5.211940698674529e-06,
    "timestep_mean": 453.75,
    "epsilon_norm": 8.42479419708252,
    "prediction_norm": 7.772783279418945,
    "timestamp": "2025-07-14T12:34:42.408022"
  },
  {
    "epoch": 1908,
    "batch_idx": 96,
    "loss": 0.458894819021225,
    "learning_rate": 5.211940698674529e-06,
    "timestep_mean": 451.6875,
    "epsilon_norm": 7.957695484161377,
    "prediction_norm": 6.458933353424072,
    "timestamp": "2025-07-14T12:34:42.413088"
  },
  {
    "epoch": 1908,
    "batch_idx": 128,
    "loss": 0.23619569838047028,
    "learning_rate": 5.211940698674529e-06,
    "timestep_mean": 529.34375,
    "epsilon_norm": 7.924575328826904,
    "prediction_norm": 7.492354393005371,
    "timestamp": "2025-07-14T12:34:42.418213"
  },
  {
    "epoch": 1908,
    "batch_idx": 160,
    "loss": 0.2511286735534668,
    "learning_rate": 5.211940698674529e-06,
    "timestep_mean": 547.5,
    "epsilon_norm": 7.367002487182617,
    "prediction_norm": 7.110994338989258,
    "timestamp": "2025-07-14T12:34:42.423285"
  },
  {
    "epoch": 1908,
    "batch_idx": 192,
    "loss": 0.13704578578472137,
    "learning_rate": 5.211940698674529e-06,
    "timestep_mean": 649.0,
    "epsilon_norm": 4.401965618133545,
    "prediction_norm": 4.403068542480469,
    "timestamp": "2025-07-14T12:34:42.428332"
  },
  {
    "epoch": 1909,
    "batch_idx": 0,
    "loss": 0.1659294068813324,
    "learning_rate": 5.099445361236905e-06,
    "timestep_mean": 503.84375,
    "epsilon_norm": 7.703479290008545,
    "prediction_norm": 6.80750036239624,
    "timestamp": "2025-07-14T12:34:42.433308"
  },
  {
    "epoch": 1909,
    "batch_idx": 32,
    "loss": 0.18831057846546173,
    "learning_rate": 5.099445361236905e-06,
    "timestep_mean": 500.1875,
    "epsilon_norm": 8.686592102050781,
    "prediction_norm": 7.502437591552734,
    "timestamp": "2025-07-14T12:34:42.438528"
  },
  {
    "epoch": 1909,
    "batch_idx": 64,
    "loss": 0.28442007303237915,
    "learning_rate": 5.099445361236905e-06,
    "timestep_mean": 424.53125,
    "epsilon_norm": 7.70266580581665,
    "prediction_norm": 6.111457824707031,
    "timestamp": "2025-07-14T12:34:42.443555"
  },
  {
    "epoch": 1909,
    "batch_idx": 96,
    "loss": 0.32227402925491333,
    "learning_rate": 5.099445361236905e-06,
    "timestep_mean": 528.46875,
    "epsilon_norm": 8.974853515625,
    "prediction_norm": 7.433964729309082,
    "timestamp": "2025-07-14T12:34:42.448552"
  },
  {
    "epoch": 1909,
    "batch_idx": 128,
    "loss": 0.25938844680786133,
    "learning_rate": 5.099445361236905e-06,
    "timestep_mean": 463.4375,
    "epsilon_norm": 6.72471284866333,
    "prediction_norm": 5.364940166473389,
    "timestamp": "2025-07-14T12:34:42.453554"
  },
  {
    "epoch": 1909,
    "batch_idx": 160,
    "loss": 0.24828937649726868,
    "learning_rate": 5.099445361236905e-06,
    "timestep_mean": 478.25,
    "epsilon_norm": 8.100378036499023,
    "prediction_norm": 6.790594577789307,
    "timestamp": "2025-07-14T12:34:42.458760"
  },
  {
    "epoch": 1909,
    "batch_idx": 192,
    "loss": 0.35423704981803894,
    "learning_rate": 5.099445361236905e-06,
    "timestep_mean": 447.125,
    "epsilon_norm": 3.8752315044403076,
    "prediction_norm": 3.2398364543914795,
    "timestamp": "2025-07-14T12:34:42.463817"
  },
  {
    "epoch": 1910,
    "batch_idx": 0,
    "loss": 0.2252216339111328,
    "learning_rate": 4.9881711417212276e-06,
    "timestep_mean": 457.25,
    "epsilon_norm": 7.469618797302246,
    "prediction_norm": 6.081935882568359,
    "timestamp": "2025-07-14T12:34:42.468808"
  },
  {
    "epoch": 1910,
    "batch_idx": 32,
    "loss": 0.1697162389755249,
    "learning_rate": 4.9881711417212276e-06,
    "timestep_mean": 483.1875,
    "epsilon_norm": 8.007553100585938,
    "prediction_norm": 7.250198841094971,
    "timestamp": "2025-07-14T12:34:42.473800"
  },
  {
    "epoch": 1910,
    "batch_idx": 64,
    "loss": 0.639228343963623,
    "learning_rate": 4.9881711417212276e-06,
    "timestep_mean": 479.65625,
    "epsilon_norm": 8.359545707702637,
    "prediction_norm": 5.566743850708008,
    "timestamp": "2025-07-14T12:34:42.478782"
  },
  {
    "epoch": 1910,
    "batch_idx": 96,
    "loss": 0.3446817994117737,
    "learning_rate": 4.9881711417212276e-06,
    "timestep_mean": 500.125,
    "epsilon_norm": 7.576459884643555,
    "prediction_norm": 6.398875713348389,
    "timestamp": "2025-07-14T12:34:42.483752"
  },
  {
    "epoch": 1910,
    "batch_idx": 128,
    "loss": 0.46177709102630615,
    "learning_rate": 4.9881711417212276e-06,
    "timestep_mean": 514.09375,
    "epsilon_norm": 8.820595741271973,
    "prediction_norm": 6.230193138122559,
    "timestamp": "2025-07-14T12:34:42.488791"
  },
  {
    "epoch": 1910,
    "batch_idx": 160,
    "loss": 0.20509041845798492,
    "learning_rate": 4.9881711417212276e-06,
    "timestep_mean": 526.5625,
    "epsilon_norm": 7.794910907745361,
    "prediction_norm": 6.607020378112793,
    "timestamp": "2025-07-14T12:34:42.493676"
  },
  {
    "epoch": 1910,
    "batch_idx": 192,
    "loss": 0.39772504568099976,
    "learning_rate": 4.9881711417212276e-06,
    "timestep_mean": 282.0,
    "epsilon_norm": 3.31429386138916,
    "prediction_norm": 2.148998498916626,
    "timestamp": "2025-07-14T12:34:42.498643"
  },
  {
    "epoch": 1911,
    "batch_idx": 0,
    "loss": 0.28886091709136963,
    "learning_rate": 4.8781183146855945e-06,
    "timestep_mean": 512.3125,
    "epsilon_norm": 8.037446022033691,
    "prediction_norm": 6.487035751342773,
    "timestamp": "2025-07-14T12:34:42.503614"
  },
  {
    "epoch": 1911,
    "batch_idx": 32,
    "loss": 0.3631107211112976,
    "learning_rate": 4.8781183146855945e-06,
    "timestep_mean": 472.21875,
    "epsilon_norm": 8.469212532043457,
    "prediction_norm": 6.593048572540283,
    "timestamp": "2025-07-14T12:34:42.508648"
  },
  {
    "epoch": 1911,
    "batch_idx": 64,
    "loss": 0.17299243807792664,
    "learning_rate": 4.8781183146855945e-06,
    "timestep_mean": 520.59375,
    "epsilon_norm": 7.292473316192627,
    "prediction_norm": 6.250752925872803,
    "timestamp": "2025-07-14T12:34:42.513550"
  },
  {
    "epoch": 1911,
    "batch_idx": 96,
    "loss": 0.26884907484054565,
    "learning_rate": 4.8781183146855945e-06,
    "timestep_mean": 519.28125,
    "epsilon_norm": 8.398755073547363,
    "prediction_norm": 7.1470465660095215,
    "timestamp": "2025-07-14T12:34:42.518599"
  },
  {
    "epoch": 1911,
    "batch_idx": 128,
    "loss": 0.24316343665122986,
    "learning_rate": 4.8781183146855945e-06,
    "timestep_mean": 559.0625,
    "epsilon_norm": 8.230711936950684,
    "prediction_norm": 7.3031415939331055,
    "timestamp": "2025-07-14T12:34:42.523587"
  },
  {
    "epoch": 1911,
    "batch_idx": 160,
    "loss": 0.2246653139591217,
    "learning_rate": 4.8781183146855945e-06,
    "timestep_mean": 473.5,
    "epsilon_norm": 7.451703071594238,
    "prediction_norm": 6.48445463180542,
    "timestamp": "2025-07-14T12:34:42.528556"
  },
  {
    "epoch": 1911,
    "batch_idx": 192,
    "loss": 0.9058220982551575,
    "learning_rate": 4.8781183146855945e-06,
    "timestep_mean": 438.375,
    "epsilon_norm": 5.386992454528809,
    "prediction_norm": 3.166400671005249,
    "timestamp": "2025-07-14T12:34:42.533659"
  },
  {
    "epoch": 1912,
    "batch_idx": 0,
    "loss": 0.18868927657604218,
    "learning_rate": 4.769287151674403e-06,
    "timestep_mean": 485.3125,
    "epsilon_norm": 7.275522708892822,
    "prediction_norm": 6.230983734130859,
    "timestamp": "2025-07-14T12:34:42.538918"
  },
  {
    "epoch": 1912,
    "batch_idx": 32,
    "loss": 0.26452064514160156,
    "learning_rate": 4.769287151674403e-06,
    "timestep_mean": 477.53125,
    "epsilon_norm": 8.6912260055542,
    "prediction_norm": 6.9163594245910645,
    "timestamp": "2025-07-14T12:34:42.544017"
  },
  {
    "epoch": 1912,
    "batch_idx": 64,
    "loss": 0.13976018130779266,
    "learning_rate": 4.769287151674403e-06,
    "timestep_mean": 457.1875,
    "epsilon_norm": 8.788368225097656,
    "prediction_norm": 7.951175212860107,
    "timestamp": "2025-07-14T12:34:42.549022"
  },
  {
    "epoch": 1912,
    "batch_idx": 96,
    "loss": 0.5422068238258362,
    "learning_rate": 4.769287151674403e-06,
    "timestep_mean": 419.15625,
    "epsilon_norm": 8.76608657836914,
    "prediction_norm": 6.0892109870910645,
    "timestamp": "2025-07-14T12:34:42.554049"
  },
  {
    "epoch": 1912,
    "batch_idx": 128,
    "loss": 0.16508573293685913,
    "learning_rate": 4.769287151674403e-06,
    "timestep_mean": 455.125,
    "epsilon_norm": 7.305854797363281,
    "prediction_norm": 6.534869194030762,
    "timestamp": "2025-07-14T12:34:42.559113"
  },
  {
    "epoch": 1912,
    "batch_idx": 160,
    "loss": 0.1537906378507614,
    "learning_rate": 4.769287151674403e-06,
    "timestep_mean": 482.8125,
    "epsilon_norm": 7.368716239929199,
    "prediction_norm": 6.292943477630615,
    "timestamp": "2025-07-14T12:34:42.564100"
  },
  {
    "epoch": 1912,
    "batch_idx": 192,
    "loss": 0.24039553105831146,
    "learning_rate": 4.769287151674403e-06,
    "timestep_mean": 558.125,
    "epsilon_norm": 3.830193042755127,
    "prediction_norm": 3.3289661407470703,
    "timestamp": "2025-07-14T12:34:42.569213"
  },
  {
    "epoch": 1913,
    "batch_idx": 0,
    "loss": 0.24744319915771484,
    "learning_rate": 4.661677921217741e-06,
    "timestep_mean": 556.0625,
    "epsilon_norm": 8.508980751037598,
    "prediction_norm": 7.702327251434326,
    "timestamp": "2025-07-14T12:34:42.574260"
  },
  {
    "epoch": 1913,
    "batch_idx": 32,
    "loss": 0.15948647260665894,
    "learning_rate": 4.661677921217741e-06,
    "timestep_mean": 553.40625,
    "epsilon_norm": 8.100505828857422,
    "prediction_norm": 7.151559352874756,
    "timestamp": "2025-07-14T12:34:42.579281"
  },
  {
    "epoch": 1913,
    "batch_idx": 64,
    "loss": 0.2508876621723175,
    "learning_rate": 4.661677921217741e-06,
    "timestep_mean": 516.53125,
    "epsilon_norm": 7.152407646179199,
    "prediction_norm": 6.1089863777160645,
    "timestamp": "2025-07-14T12:34:42.584324"
  },
  {
    "epoch": 1913,
    "batch_idx": 96,
    "loss": 0.39271843433380127,
    "learning_rate": 4.661677921217741e-06,
    "timestep_mean": 446.09375,
    "epsilon_norm": 8.744247436523438,
    "prediction_norm": 6.442684650421143,
    "timestamp": "2025-07-14T12:34:42.589210"
  },
  {
    "epoch": 1913,
    "batch_idx": 128,
    "loss": 0.25198894739151,
    "learning_rate": 4.661677921217741e-06,
    "timestep_mean": 486.40625,
    "epsilon_norm": 7.393856048583984,
    "prediction_norm": 6.058929443359375,
    "timestamp": "2025-07-14T12:34:42.594210"
  },
  {
    "epoch": 1913,
    "batch_idx": 160,
    "loss": 0.4241970181465149,
    "learning_rate": 4.661677921217741e-06,
    "timestep_mean": 535.375,
    "epsilon_norm": 8.252153396606445,
    "prediction_norm": 6.676215171813965,
    "timestamp": "2025-07-14T12:34:42.599163"
  },
  {
    "epoch": 1913,
    "batch_idx": 192,
    "loss": 0.03443995118141174,
    "learning_rate": 4.661677921217741e-06,
    "timestep_mean": 514.5,
    "epsilon_norm": 3.016436815261841,
    "prediction_norm": 2.9960877895355225,
    "timestamp": "2025-07-14T12:34:42.604152"
  },
  {
    "epoch": 1914,
    "batch_idx": 0,
    "loss": 0.3119719922542572,
    "learning_rate": 4.5552908888306626e-06,
    "timestep_mean": 597.71875,
    "epsilon_norm": 7.702636241912842,
    "prediction_norm": 6.352717876434326,
    "timestamp": "2025-07-14T12:34:42.609130"
  },
  {
    "epoch": 1914,
    "batch_idx": 32,
    "loss": 0.3296612501144409,
    "learning_rate": 4.5552908888306626e-06,
    "timestep_mean": 531.1875,
    "epsilon_norm": 7.379698276519775,
    "prediction_norm": 6.08026647567749,
    "timestamp": "2025-07-14T12:34:42.614043"
  },
  {
    "epoch": 1914,
    "batch_idx": 64,
    "loss": 0.21459557116031647,
    "learning_rate": 4.5552908888306626e-06,
    "timestep_mean": 477.65625,
    "epsilon_norm": 7.470785617828369,
    "prediction_norm": 6.376856327056885,
    "timestamp": "2025-07-14T12:34:42.619074"
  },
  {
    "epoch": 1914,
    "batch_idx": 96,
    "loss": 0.3077200949192047,
    "learning_rate": 4.5552908888306626e-06,
    "timestep_mean": 436.21875,
    "epsilon_norm": 7.454492092132568,
    "prediction_norm": 6.07625150680542,
    "timestamp": "2025-07-14T12:34:42.624063"
  },
  {
    "epoch": 1914,
    "batch_idx": 128,
    "loss": 0.2053304761648178,
    "learning_rate": 4.5552908888306626e-06,
    "timestep_mean": 487.28125,
    "epsilon_norm": 7.906077861785889,
    "prediction_norm": 6.911592960357666,
    "timestamp": "2025-07-14T12:34:42.629222"
  },
  {
    "epoch": 1914,
    "batch_idx": 160,
    "loss": 0.2363073229789734,
    "learning_rate": 4.5552908888306626e-06,
    "timestep_mean": 469.375,
    "epsilon_norm": 7.879138946533203,
    "prediction_norm": 6.891229152679443,
    "timestamp": "2025-07-14T12:34:42.634540"
  },
  {
    "epoch": 1914,
    "batch_idx": 192,
    "loss": 0.0741250216960907,
    "learning_rate": 4.5552908888306626e-06,
    "timestep_mean": 597.0,
    "epsilon_norm": 3.669933319091797,
    "prediction_norm": 3.37349009513855,
    "timestamp": "2025-07-14T12:34:42.639752"
  },
  {
    "epoch": 1915,
    "batch_idx": 0,
    "loss": 0.16362135112285614,
    "learning_rate": 4.450126317012635e-06,
    "timestep_mean": 506.09375,
    "epsilon_norm": 7.795411586761475,
    "prediction_norm": 7.273276329040527,
    "timestamp": "2025-07-14T12:34:42.645088"
  },
  {
    "epoch": 1915,
    "batch_idx": 32,
    "loss": 0.3231056034564972,
    "learning_rate": 4.450126317012635e-06,
    "timestep_mean": 459.375,
    "epsilon_norm": 8.252425193786621,
    "prediction_norm": 6.8898024559021,
    "timestamp": "2025-07-14T12:34:42.650486"
  },
  {
    "epoch": 1915,
    "batch_idx": 64,
    "loss": 0.19797539710998535,
    "learning_rate": 4.450126317012635e-06,
    "timestep_mean": 436.5625,
    "epsilon_norm": 6.9858903884887695,
    "prediction_norm": 5.817180156707764,
    "timestamp": "2025-07-14T12:34:42.655718"
  },
  {
    "epoch": 1915,
    "batch_idx": 96,
    "loss": 0.30118080973625183,
    "learning_rate": 4.450126317012635e-06,
    "timestep_mean": 491.09375,
    "epsilon_norm": 7.437389850616455,
    "prediction_norm": 6.310254096984863,
    "timestamp": "2025-07-14T12:34:42.661037"
  },
  {
    "epoch": 1915,
    "batch_idx": 128,
    "loss": 0.3035612106323242,
    "learning_rate": 4.450126317012635e-06,
    "timestep_mean": 439.96875,
    "epsilon_norm": 7.542760848999023,
    "prediction_norm": 5.793447971343994,
    "timestamp": "2025-07-14T12:34:42.668426"
  },
  {
    "epoch": 1915,
    "batch_idx": 160,
    "loss": 0.12979596853256226,
    "learning_rate": 4.450126317012635e-06,
    "timestep_mean": 545.3125,
    "epsilon_norm": 7.569006443023682,
    "prediction_norm": 6.9643378257751465,
    "timestamp": "2025-07-14T12:34:42.681441"
  },
  {
    "epoch": 1915,
    "batch_idx": 192,
    "loss": 0.09495534002780914,
    "learning_rate": 4.450126317012635e-06,
    "timestep_mean": 462.625,
    "epsilon_norm": 3.8747386932373047,
    "prediction_norm": 3.5909056663513184,
    "timestamp": "2025-07-14T12:34:42.689776"
  },
  {
    "epoch": 1916,
    "batch_idx": 0,
    "loss": 0.2070103883743286,
    "learning_rate": 4.346184465246703e-06,
    "timestep_mean": 454.3125,
    "epsilon_norm": 8.05824089050293,
    "prediction_norm": 7.482696533203125,
    "timestamp": "2025-07-14T12:34:42.697260"
  },
  {
    "epoch": 1916,
    "batch_idx": 32,
    "loss": 0.1995326280593872,
    "learning_rate": 4.346184465246703e-06,
    "timestep_mean": 521.53125,
    "epsilon_norm": 7.8062052726745605,
    "prediction_norm": 6.763145446777344,
    "timestamp": "2025-07-14T12:34:42.705783"
  },
  {
    "epoch": 1916,
    "batch_idx": 64,
    "loss": 0.3018195629119873,
    "learning_rate": 4.346184465246703e-06,
    "timestep_mean": 458.34375,
    "epsilon_norm": 7.415359973907471,
    "prediction_norm": 6.070179462432861,
    "timestamp": "2025-07-14T12:34:42.713826"
  },
  {
    "epoch": 1916,
    "batch_idx": 96,
    "loss": 0.2800747752189636,
    "learning_rate": 4.346184465246703e-06,
    "timestep_mean": 468.5,
    "epsilon_norm": 8.177419662475586,
    "prediction_norm": 7.3451104164123535,
    "timestamp": "2025-07-14T12:34:42.722156"
  },
  {
    "epoch": 1916,
    "batch_idx": 128,
    "loss": 0.297235906124115,
    "learning_rate": 4.346184465246703e-06,
    "timestep_mean": 548.125,
    "epsilon_norm": 8.488121032714844,
    "prediction_norm": 7.296253204345703,
    "timestamp": "2025-07-14T12:34:42.730561"
  },
  {
    "epoch": 1916,
    "batch_idx": 160,
    "loss": 0.2270793616771698,
    "learning_rate": 4.346184465246703e-06,
    "timestep_mean": 502.6875,
    "epsilon_norm": 9.230774879455566,
    "prediction_norm": 7.645852565765381,
    "timestamp": "2025-07-14T12:34:42.738970"
  },
  {
    "epoch": 1916,
    "batch_idx": 192,
    "loss": 0.16186559200286865,
    "learning_rate": 4.346184465246703e-06,
    "timestep_mean": 286.0,
    "epsilon_norm": 2.1155903339385986,
    "prediction_norm": 2.607175827026367,
    "timestamp": "2025-07-14T12:34:42.744737"
  },
  {
    "epoch": 1917,
    "batch_idx": 0,
    "loss": 0.10369117558002472,
    "learning_rate": 4.243465589999161e-06,
    "timestep_mean": 500.25,
    "epsilon_norm": 7.899139404296875,
    "prediction_norm": 7.550563335418701,
    "timestamp": "2025-07-14T12:34:42.751720"
  },
  {
    "epoch": 1917,
    "batch_idx": 32,
    "loss": 0.33829450607299805,
    "learning_rate": 4.243465589999161e-06,
    "timestep_mean": 439.71875,
    "epsilon_norm": 7.9826765060424805,
    "prediction_norm": 6.310491561889648,
    "timestamp": "2025-07-14T12:34:42.760098"
  },
  {
    "epoch": 1917,
    "batch_idx": 64,
    "loss": 0.15737037360668182,
    "learning_rate": 4.243465589999161e-06,
    "timestep_mean": 510.25,
    "epsilon_norm": 7.130926132202148,
    "prediction_norm": 6.451333045959473,
    "timestamp": "2025-07-14T12:34:42.767376"
  },
  {
    "epoch": 1917,
    "batch_idx": 96,
    "loss": 0.23589298129081726,
    "learning_rate": 4.243465589999161e-06,
    "timestep_mean": 463.5625,
    "epsilon_norm": 7.4653472900390625,
    "prediction_norm": 6.086507320404053,
    "timestamp": "2025-07-14T12:34:42.772558"
  },
  {
    "epoch": 1917,
    "batch_idx": 128,
    "loss": 0.21399271488189697,
    "learning_rate": 4.243465589999161e-06,
    "timestep_mean": 554.6875,
    "epsilon_norm": 7.568730354309082,
    "prediction_norm": 7.082009792327881,
    "timestamp": "2025-07-14T12:34:42.777603"
  },
  {
    "epoch": 1917,
    "batch_idx": 160,
    "loss": 0.3333006203174591,
    "learning_rate": 4.243465589999161e-06,
    "timestep_mean": 459.9375,
    "epsilon_norm": 8.254761695861816,
    "prediction_norm": 6.656735420227051,
    "timestamp": "2025-07-14T12:34:42.782762"
  },
  {
    "epoch": 1917,
    "batch_idx": 192,
    "loss": 0.9240422248840332,
    "learning_rate": 4.243465589999161e-06,
    "timestep_mean": 443.375,
    "epsilon_norm": 4.6617231369018555,
    "prediction_norm": 2.693939685821533,
    "timestamp": "2025-07-14T12:34:42.788144"
  },
  {
    "epoch": 1918,
    "batch_idx": 0,
    "loss": 0.24948738515377045,
    "learning_rate": 4.141969944718602e-06,
    "timestep_mean": 466.28125,
    "epsilon_norm": 9.154248237609863,
    "prediction_norm": 8.532666206359863,
    "timestamp": "2025-07-14T12:34:42.793305"
  },
  {
    "epoch": 1918,
    "batch_idx": 32,
    "loss": 0.3044300973415375,
    "learning_rate": 4.141969944718602e-06,
    "timestep_mean": 481.34375,
    "epsilon_norm": 8.710041999816895,
    "prediction_norm": 7.390567302703857,
    "timestamp": "2025-07-14T12:34:42.798447"
  },
  {
    "epoch": 1918,
    "batch_idx": 64,
    "loss": 0.19323775172233582,
    "learning_rate": 4.141969944718602e-06,
    "timestep_mean": 414.5625,
    "epsilon_norm": 7.770254135131836,
    "prediction_norm": 7.302525043487549,
    "timestamp": "2025-07-14T12:34:42.803780"
  },
  {
    "epoch": 1918,
    "batch_idx": 96,
    "loss": 0.42598289251327515,
    "learning_rate": 4.141969944718602e-06,
    "timestep_mean": 486.28125,
    "epsilon_norm": 8.935993194580078,
    "prediction_norm": 7.531770706176758,
    "timestamp": "2025-07-14T12:34:42.808881"
  },
  {
    "epoch": 1918,
    "batch_idx": 128,
    "loss": 0.17700088024139404,
    "learning_rate": 4.141969944718602e-06,
    "timestep_mean": 544.6875,
    "epsilon_norm": 9.372901916503906,
    "prediction_norm": 8.614901542663574,
    "timestamp": "2025-07-14T12:34:42.813808"
  },
  {
    "epoch": 1918,
    "batch_idx": 160,
    "loss": 0.19320949912071228,
    "learning_rate": 4.141969944718602e-06,
    "timestep_mean": 562.75,
    "epsilon_norm": 8.283011436462402,
    "prediction_norm": 7.734041690826416,
    "timestamp": "2025-07-14T12:34:42.818862"
  },
  {
    "epoch": 1918,
    "batch_idx": 192,
    "loss": 0.11074677109718323,
    "learning_rate": 4.141969944718602e-06,
    "timestep_mean": 431.0,
    "epsilon_norm": 4.253142833709717,
    "prediction_norm": 3.4855167865753174,
    "timestamp": "2025-07-14T12:34:42.824005"
  },
  {
    "epoch": 1919,
    "batch_idx": 0,
    "loss": 0.20041900873184204,
    "learning_rate": 4.041697779835371e-06,
    "timestep_mean": 562.625,
    "epsilon_norm": 8.993093490600586,
    "prediction_norm": 7.974074840545654,
    "timestamp": "2025-07-14T12:34:42.829152"
  },
  {
    "epoch": 1919,
    "batch_idx": 32,
    "loss": 0.23544850945472717,
    "learning_rate": 4.041697779835371e-06,
    "timestep_mean": 515.53125,
    "epsilon_norm": 7.724827766418457,
    "prediction_norm": 6.8132643699646,
    "timestamp": "2025-07-14T12:34:42.834376"
  },
  {
    "epoch": 1919,
    "batch_idx": 64,
    "loss": 0.282371461391449,
    "learning_rate": 4.041697779835371e-06,
    "timestep_mean": 547.5625,
    "epsilon_norm": 8.084966659545898,
    "prediction_norm": 6.9037184715271,
    "timestamp": "2025-07-14T12:34:42.839566"
  },
  {
    "epoch": 1919,
    "batch_idx": 96,
    "loss": 0.22020024061203003,
    "learning_rate": 4.041697779835371e-06,
    "timestep_mean": 463.96875,
    "epsilon_norm": 7.481311798095703,
    "prediction_norm": 6.562958717346191,
    "timestamp": "2025-07-14T12:34:42.844711"
  },
  {
    "epoch": 1919,
    "batch_idx": 128,
    "loss": 0.44208598136901855,
    "learning_rate": 4.041697779835371e-06,
    "timestep_mean": 454.65625,
    "epsilon_norm": 8.24719524383545,
    "prediction_norm": 5.9301228523254395,
    "timestamp": "2025-07-14T12:34:42.849913"
  },
  {
    "epoch": 1919,
    "batch_idx": 160,
    "loss": 0.28380897641181946,
    "learning_rate": 4.041697779835371e-06,
    "timestep_mean": 514.625,
    "epsilon_norm": 8.71210765838623,
    "prediction_norm": 7.347795486450195,
    "timestamp": "2025-07-14T12:34:42.855289"
  },
  {
    "epoch": 1919,
    "batch_idx": 192,
    "loss": 0.13856935501098633,
    "learning_rate": 4.041697779835371e-06,
    "timestep_mean": 408.375,
    "epsilon_norm": 4.369637489318848,
    "prediction_norm": 4.152362823486328,
    "timestamp": "2025-07-14T12:34:42.860563"
  },
  {
    "epoch": 1920,
    "batch_idx": 0,
    "loss": 0.2600606679916382,
    "learning_rate": 3.942649342761115e-06,
    "timestep_mean": 483.5625,
    "epsilon_norm": 8.34726333618164,
    "prediction_norm": 7.332704544067383,
    "timestamp": "2025-07-14T12:34:42.865729"
  },
  {
    "epoch": 1920,
    "batch_idx": 32,
    "loss": 0.3300129175186157,
    "learning_rate": 3.942649342761115e-06,
    "timestep_mean": 611.0,
    "epsilon_norm": 8.661091804504395,
    "prediction_norm": 7.669384956359863,
    "timestamp": "2025-07-14T12:34:42.871241"
  },
  {
    "epoch": 1920,
    "batch_idx": 64,
    "loss": 0.2439180612564087,
    "learning_rate": 3.942649342761115e-06,
    "timestep_mean": 583.65625,
    "epsilon_norm": 8.139872550964355,
    "prediction_norm": 7.038259983062744,
    "timestamp": "2025-07-14T12:34:42.876599"
  },
  {
    "epoch": 1920,
    "batch_idx": 96,
    "loss": 0.27299967408180237,
    "learning_rate": 3.942649342761115e-06,
    "timestep_mean": 461.875,
    "epsilon_norm": 7.624603748321533,
    "prediction_norm": 6.208322525024414,
    "timestamp": "2025-07-14T12:34:42.881738"
  },
  {
    "epoch": 1920,
    "batch_idx": 128,
    "loss": 0.24483221769332886,
    "learning_rate": 3.942649342761115e-06,
    "timestep_mean": 603.375,
    "epsilon_norm": 8.304946899414062,
    "prediction_norm": 7.432772159576416,
    "timestamp": "2025-07-14T12:34:42.887037"
  },
  {
    "epoch": 1920,
    "batch_idx": 160,
    "loss": 0.421041876077652,
    "learning_rate": 3.942649342761115e-06,
    "timestep_mean": 425.0625,
    "epsilon_norm": 8.453968048095703,
    "prediction_norm": 7.070036888122559,
    "timestamp": "2025-07-14T12:34:42.892306"
  },
  {
    "epoch": 1920,
    "batch_idx": 192,
    "loss": 0.3507419526576996,
    "learning_rate": 3.942649342761115e-06,
    "timestep_mean": 533.5,
    "epsilon_norm": 3.5989797115325928,
    "prediction_norm": 2.060199737548828,
    "timestamp": "2025-07-14T12:34:42.897458"
  },
  {
    "epoch": 1921,
    "batch_idx": 0,
    "loss": 0.26921918988227844,
    "learning_rate": 3.844824877888006e-06,
    "timestep_mean": 471.65625,
    "epsilon_norm": 8.259227752685547,
    "prediction_norm": 7.057325839996338,
    "timestamp": "2025-07-14T12:34:42.902763"
  },
  {
    "epoch": 1921,
    "batch_idx": 32,
    "loss": 0.31267082691192627,
    "learning_rate": 3.844824877888006e-06,
    "timestep_mean": 398.75,
    "epsilon_norm": 7.781026840209961,
    "prediction_norm": 6.414089679718018,
    "timestamp": "2025-07-14T12:34:42.907941"
  },
  {
    "epoch": 1921,
    "batch_idx": 64,
    "loss": 0.20662225782871246,
    "learning_rate": 3.844824877888006e-06,
    "timestep_mean": 558.03125,
    "epsilon_norm": 8.200617790222168,
    "prediction_norm": 7.222513198852539,
    "timestamp": "2025-07-14T12:34:42.913222"
  },
  {
    "epoch": 1921,
    "batch_idx": 96,
    "loss": 0.304615318775177,
    "learning_rate": 3.844824877888006e-06,
    "timestep_mean": 465.8125,
    "epsilon_norm": 8.430489540100098,
    "prediction_norm": 6.462454795837402,
    "timestamp": "2025-07-14T12:34:42.918561"
  },
  {
    "epoch": 1921,
    "batch_idx": 128,
    "loss": 0.4078575372695923,
    "learning_rate": 3.844824877888006e-06,
    "timestep_mean": 493.5,
    "epsilon_norm": 8.511960983276367,
    "prediction_norm": 7.0834879875183105,
    "timestamp": "2025-07-14T12:34:42.923667"
  },
  {
    "epoch": 1921,
    "batch_idx": 160,
    "loss": 0.3334527015686035,
    "learning_rate": 3.844824877888006e-06,
    "timestep_mean": 476.78125,
    "epsilon_norm": 8.21883487701416,
    "prediction_norm": 6.23752498626709,
    "timestamp": "2025-07-14T12:34:42.928935"
  },
  {
    "epoch": 1921,
    "batch_idx": 192,
    "loss": 0.22958698868751526,
    "learning_rate": 3.844824877888006e-06,
    "timestep_mean": 472.625,
    "epsilon_norm": 4.037782669067383,
    "prediction_norm": 3.0701215267181396,
    "timestamp": "2025-07-14T12:34:42.934477"
  },
  {
    "epoch": 1922,
    "batch_idx": 0,
    "loss": 0.2408672571182251,
    "learning_rate": 3.748224626588134e-06,
    "timestep_mean": 599.09375,
    "epsilon_norm": 7.926501274108887,
    "prediction_norm": 6.9642863273620605,
    "timestamp": "2025-07-14T12:34:42.939878"
  },
  {
    "epoch": 1922,
    "batch_idx": 32,
    "loss": 0.4465595483779907,
    "learning_rate": 3.748224626588134e-06,
    "timestep_mean": 442.90625,
    "epsilon_norm": 6.897182464599609,
    "prediction_norm": 4.912117958068848,
    "timestamp": "2025-07-14T12:34:42.945986"
  },
  {
    "epoch": 1922,
    "batch_idx": 64,
    "loss": 0.32115638256073,
    "learning_rate": 3.748224626588134e-06,
    "timestep_mean": 445.25,
    "epsilon_norm": 7.4762701988220215,
    "prediction_norm": 5.787322521209717,
    "timestamp": "2025-07-14T12:34:42.951480"
  },
  {
    "epoch": 1922,
    "batch_idx": 96,
    "loss": 0.2880792021751404,
    "learning_rate": 3.748224626588134e-06,
    "timestep_mean": 525.09375,
    "epsilon_norm": 7.743659496307373,
    "prediction_norm": 6.701501846313477,
    "timestamp": "2025-07-14T12:34:42.962265"
  },
  {
    "epoch": 1922,
    "batch_idx": 128,
    "loss": 0.2866756319999695,
    "learning_rate": 3.748224626588134e-06,
    "timestep_mean": 511.25,
    "epsilon_norm": 8.434586524963379,
    "prediction_norm": 6.852987289428711,
    "timestamp": "2025-07-14T12:34:42.967854"
  },
  {
    "epoch": 1922,
    "batch_idx": 160,
    "loss": 0.229872465133667,
    "learning_rate": 3.748224626588134e-06,
    "timestep_mean": 409.40625,
    "epsilon_norm": 6.257273197174072,
    "prediction_norm": 5.64432430267334,
    "timestamp": "2025-07-14T12:34:42.972923"
  },
  {
    "epoch": 1922,
    "batch_idx": 192,
    "loss": 0.16298358142375946,
    "learning_rate": 3.748224626588134e-06,
    "timestep_mean": 353.625,
    "epsilon_norm": 3.647392988204956,
    "prediction_norm": 3.4148974418640137,
    "timestamp": "2025-07-14T12:34:42.978106"
  },
  {
    "epoch": 1923,
    "batch_idx": 0,
    "loss": 0.16371917724609375,
    "learning_rate": 3.652848827213115e-06,
    "timestep_mean": 523.53125,
    "epsilon_norm": 8.12726879119873,
    "prediction_norm": 7.703841209411621,
    "timestamp": "2025-07-14T12:34:42.983280"
  },
  {
    "epoch": 1923,
    "batch_idx": 32,
    "loss": 0.23207449913024902,
    "learning_rate": 3.652848827213115e-06,
    "timestep_mean": 532.28125,
    "epsilon_norm": 7.929203033447266,
    "prediction_norm": 6.767265319824219,
    "timestamp": "2025-07-14T12:34:42.988359"
  },
  {
    "epoch": 1923,
    "batch_idx": 64,
    "loss": 0.4140816628932953,
    "learning_rate": 3.652848827213115e-06,
    "timestep_mean": 475.71875,
    "epsilon_norm": 8.75509262084961,
    "prediction_norm": 7.035327911376953,
    "timestamp": "2025-07-14T12:34:42.993300"
  },
  {
    "epoch": 1923,
    "batch_idx": 96,
    "loss": 0.2256803959608078,
    "learning_rate": 3.652848827213115e-06,
    "timestep_mean": 580.5,
    "epsilon_norm": 7.886480331420898,
    "prediction_norm": 7.054955005645752,
    "timestamp": "2025-07-14T12:34:42.998292"
  },
  {
    "epoch": 1923,
    "batch_idx": 128,
    "loss": 0.5094879269599915,
    "learning_rate": 3.652848827213115e-06,
    "timestep_mean": 446.46875,
    "epsilon_norm": 8.595142364501953,
    "prediction_norm": 6.578042507171631,
    "timestamp": "2025-07-14T12:34:43.004050"
  },
  {
    "epoch": 1923,
    "batch_idx": 160,
    "loss": 0.11073487251996994,
    "learning_rate": 3.652848827213115e-06,
    "timestep_mean": 521.84375,
    "epsilon_norm": 6.689970970153809,
    "prediction_norm": 6.455731391906738,
    "timestamp": "2025-07-14T12:34:43.009657"
  },
  {
    "epoch": 1923,
    "batch_idx": 192,
    "loss": 0.11910152435302734,
    "learning_rate": 3.652848827213115e-06,
    "timestep_mean": 545.25,
    "epsilon_norm": 4.136855125427246,
    "prediction_norm": 4.121335983276367,
    "timestamp": "2025-07-14T12:34:43.014955"
  },
  {
    "epoch": 1924,
    "batch_idx": 0,
    "loss": 0.2785678207874298,
    "learning_rate": 3.5586977150932035e-06,
    "timestep_mean": 481.71875,
    "epsilon_norm": 8.860896110534668,
    "prediction_norm": 7.3212666511535645,
    "timestamp": "2025-07-14T12:34:43.020283"
  },
  {
    "epoch": 1924,
    "batch_idx": 32,
    "loss": 0.421058714389801,
    "learning_rate": 3.5586977150932035e-06,
    "timestep_mean": 481.90625,
    "epsilon_norm": 8.296944618225098,
    "prediction_norm": 6.330021858215332,
    "timestamp": "2025-07-14T12:34:43.025557"
  },
  {
    "epoch": 1924,
    "batch_idx": 64,
    "loss": 0.19730651378631592,
    "learning_rate": 3.5586977150932035e-06,
    "timestep_mean": 549.1875,
    "epsilon_norm": 6.881308555603027,
    "prediction_norm": 6.143082141876221,
    "timestamp": "2025-07-14T12:34:43.030655"
  },
  {
    "epoch": 1924,
    "batch_idx": 96,
    "loss": 0.28263381123542786,
    "learning_rate": 3.5586977150932035e-06,
    "timestep_mean": 505.03125,
    "epsilon_norm": 8.298737525939941,
    "prediction_norm": 7.48799991607666,
    "timestamp": "2025-07-14T12:34:43.035872"
  },
  {
    "epoch": 1924,
    "batch_idx": 128,
    "loss": 0.14486122131347656,
    "learning_rate": 3.5586977150932035e-06,
    "timestep_mean": 565.9375,
    "epsilon_norm": 7.460631370544434,
    "prediction_norm": 6.703495025634766,
    "timestamp": "2025-07-14T12:34:43.040988"
  },
  {
    "epoch": 1924,
    "batch_idx": 160,
    "loss": 0.20106641948223114,
    "learning_rate": 3.5586977150932035e-06,
    "timestep_mean": 459.625,
    "epsilon_norm": 7.4094390869140625,
    "prediction_norm": 6.854693412780762,
    "timestamp": "2025-07-14T12:34:43.046244"
  },
  {
    "epoch": 1924,
    "batch_idx": 192,
    "loss": 0.06370387226343155,
    "learning_rate": 3.5586977150932035e-06,
    "timestep_mean": 635.0,
    "epsilon_norm": 3.1978275775909424,
    "prediction_norm": 2.7125816345214844,
    "timestamp": "2025-07-14T12:34:43.051336"
  },
  {
    "epoch": 1925,
    "batch_idx": 0,
    "loss": 0.3499791622161865,
    "learning_rate": 3.465771522536851e-06,
    "timestep_mean": 487.71875,
    "epsilon_norm": 8.099724769592285,
    "prediction_norm": 6.5163893699646,
    "timestamp": "2025-07-14T12:34:43.056520"
  },
  {
    "epoch": 1925,
    "batch_idx": 32,
    "loss": 0.17705106735229492,
    "learning_rate": 3.465771522536851e-06,
    "timestep_mean": 502.5,
    "epsilon_norm": 6.3961663246154785,
    "prediction_norm": 5.759025573730469,
    "timestamp": "2025-07-14T12:34:43.061585"
  },
  {
    "epoch": 1925,
    "batch_idx": 64,
    "loss": 0.2715737819671631,
    "learning_rate": 3.465771522536851e-06,
    "timestep_mean": 451.84375,
    "epsilon_norm": 7.968505859375,
    "prediction_norm": 6.503571033477783,
    "timestamp": "2025-07-14T12:34:43.066876"
  },
  {
    "epoch": 1925,
    "batch_idx": 96,
    "loss": 0.4322253465652466,
    "learning_rate": 3.465771522536851e-06,
    "timestep_mean": 532.6875,
    "epsilon_norm": 8.614693641662598,
    "prediction_norm": 6.796127796173096,
    "timestamp": "2025-07-14T12:34:43.072209"
  },
  {
    "epoch": 1925,
    "batch_idx": 128,
    "loss": 0.16454942524433136,
    "learning_rate": 3.465771522536851e-06,
    "timestep_mean": 562.65625,
    "epsilon_norm": 6.394905090332031,
    "prediction_norm": 5.682424068450928,
    "timestamp": "2025-07-14T12:34:43.077120"
  },
  {
    "epoch": 1925,
    "batch_idx": 160,
    "loss": 0.32413366436958313,
    "learning_rate": 3.465771522536851e-06,
    "timestep_mean": 444.96875,
    "epsilon_norm": 8.13039779663086,
    "prediction_norm": 6.441720485687256,
    "timestamp": "2025-07-14T12:34:43.082083"
  },
  {
    "epoch": 1925,
    "batch_idx": 192,
    "loss": 0.08755079656839371,
    "learning_rate": 3.465771522536851e-06,
    "timestep_mean": 595.125,
    "epsilon_norm": 3.717215061187744,
    "prediction_norm": 3.753957986831665,
    "timestamp": "2025-07-14T12:34:43.087080"
  },
  {
    "epoch": 1926,
    "batch_idx": 0,
    "loss": 0.302435040473938,
    "learning_rate": 3.374070478830313e-06,
    "timestep_mean": 435.46875,
    "epsilon_norm": 7.570674896240234,
    "prediction_norm": 5.540698051452637,
    "timestamp": "2025-07-14T12:34:43.092372"
  },
  {
    "epoch": 1926,
    "batch_idx": 32,
    "loss": 0.4911027252674103,
    "learning_rate": 3.374070478830313e-06,
    "timestep_mean": 389.0625,
    "epsilon_norm": 7.920472145080566,
    "prediction_norm": 6.125850200653076,
    "timestamp": "2025-07-14T12:34:43.097559"
  },
  {
    "epoch": 1926,
    "batch_idx": 64,
    "loss": 0.24219539761543274,
    "learning_rate": 3.374070478830313e-06,
    "timestep_mean": 439.65625,
    "epsilon_norm": 6.992470741271973,
    "prediction_norm": 6.226254940032959,
    "timestamp": "2025-07-14T12:34:43.102580"
  },
  {
    "epoch": 1926,
    "batch_idx": 96,
    "loss": 0.3275638818740845,
    "learning_rate": 3.374070478830313e-06,
    "timestep_mean": 496.375,
    "epsilon_norm": 8.655085563659668,
    "prediction_norm": 7.055924892425537,
    "timestamp": "2025-07-14T12:34:43.107639"
  },
  {
    "epoch": 1926,
    "batch_idx": 128,
    "loss": 0.28694218397140503,
    "learning_rate": 3.374070478830313e-06,
    "timestep_mean": 555.84375,
    "epsilon_norm": 8.924985885620117,
    "prediction_norm": 7.789496898651123,
    "timestamp": "2025-07-14T12:34:43.112872"
  },
  {
    "epoch": 1926,
    "batch_idx": 160,
    "loss": 0.19426557421684265,
    "learning_rate": 3.374070478830313e-06,
    "timestep_mean": 498.125,
    "epsilon_norm": 7.434108257293701,
    "prediction_norm": 6.3615851402282715,
    "timestamp": "2025-07-14T12:34:43.117978"
  },
  {
    "epoch": 1926,
    "batch_idx": 192,
    "loss": 0.7427230477333069,
    "learning_rate": 3.374070478830313e-06,
    "timestep_mean": 524.25,
    "epsilon_norm": 4.215241432189941,
    "prediction_norm": 2.629737377166748,
    "timestamp": "2025-07-14T12:34:43.123031"
  },
  {
    "epoch": 1927,
    "batch_idx": 0,
    "loss": 0.18903037905693054,
    "learning_rate": 3.283594810236709e-06,
    "timestep_mean": 448.03125,
    "epsilon_norm": 6.824600696563721,
    "prediction_norm": 5.929187297821045,
    "timestamp": "2025-07-14T12:34:43.128129"
  },
  {
    "epoch": 1927,
    "batch_idx": 32,
    "loss": 0.10463708639144897,
    "learning_rate": 3.283594810236709e-06,
    "timestep_mean": 544.9375,
    "epsilon_norm": 7.888937473297119,
    "prediction_norm": 7.623376369476318,
    "timestamp": "2025-07-14T12:34:43.133225"
  },
  {
    "epoch": 1927,
    "batch_idx": 64,
    "loss": 0.3603309988975525,
    "learning_rate": 3.283594810236709e-06,
    "timestep_mean": 371.59375,
    "epsilon_norm": 7.418048858642578,
    "prediction_norm": 6.021815776824951,
    "timestamp": "2025-07-14T12:34:43.138311"
  },
  {
    "epoch": 1927,
    "batch_idx": 96,
    "loss": 0.16930365562438965,
    "learning_rate": 3.283594810236709e-06,
    "timestep_mean": 507.65625,
    "epsilon_norm": 8.496072769165039,
    "prediction_norm": 7.619630336761475,
    "timestamp": "2025-07-14T12:34:43.143368"
  },
  {
    "epoch": 1927,
    "batch_idx": 128,
    "loss": 0.2901284694671631,
    "learning_rate": 3.283594810236709e-06,
    "timestep_mean": 493.375,
    "epsilon_norm": 7.637350559234619,
    "prediction_norm": 6.287046432495117,
    "timestamp": "2025-07-14T12:34:43.148364"
  },
  {
    "epoch": 1927,
    "batch_idx": 160,
    "loss": 0.36817777156829834,
    "learning_rate": 3.283594810236709e-06,
    "timestep_mean": 453.40625,
    "epsilon_norm": 8.37360954284668,
    "prediction_norm": 6.938567161560059,
    "timestamp": "2025-07-14T12:34:43.153328"
  },
  {
    "epoch": 1927,
    "batch_idx": 192,
    "loss": 0.3403559625148773,
    "learning_rate": 3.283594810236709e-06,
    "timestep_mean": 464.25,
    "epsilon_norm": 4.458429336547852,
    "prediction_norm": 3.5068042278289795,
    "timestamp": "2025-07-14T12:34:43.158302"
  },
  {
    "epoch": 1928,
    "batch_idx": 0,
    "loss": 0.23583579063415527,
    "learning_rate": 3.1943447399958e-06,
    "timestep_mean": 394.875,
    "epsilon_norm": 8.515487670898438,
    "prediction_norm": 6.969609260559082,
    "timestamp": "2025-07-14T12:34:43.163271"
  },
  {
    "epoch": 1928,
    "batch_idx": 32,
    "loss": 0.24039186537265778,
    "learning_rate": 3.1943447399958e-06,
    "timestep_mean": 488.15625,
    "epsilon_norm": 7.46774435043335,
    "prediction_norm": 6.7170491218566895,
    "timestamp": "2025-07-14T12:34:43.168290"
  },
  {
    "epoch": 1928,
    "batch_idx": 64,
    "loss": 0.31454068422317505,
    "learning_rate": 3.1943447399958e-06,
    "timestep_mean": 549.0,
    "epsilon_norm": 7.739066123962402,
    "prediction_norm": 6.659094333648682,
    "timestamp": "2025-07-14T12:34:43.173336"
  },
  {
    "epoch": 1928,
    "batch_idx": 96,
    "loss": 0.5974258184432983,
    "learning_rate": 3.1943447399958e-06,
    "timestep_mean": 439.15625,
    "epsilon_norm": 8.21960163116455,
    "prediction_norm": 6.303737640380859,
    "timestamp": "2025-07-14T12:34:43.178408"
  },
  {
    "epoch": 1928,
    "batch_idx": 128,
    "loss": 0.28395718336105347,
    "learning_rate": 3.1943447399958e-06,
    "timestep_mean": 470.84375,
    "epsilon_norm": 8.756714820861816,
    "prediction_norm": 7.392154216766357,
    "timestamp": "2025-07-14T12:34:43.183408"
  },
  {
    "epoch": 1928,
    "batch_idx": 160,
    "loss": 0.255261093378067,
    "learning_rate": 3.1943447399958e-06,
    "timestep_mean": 515.84375,
    "epsilon_norm": 8.919108390808105,
    "prediction_norm": 7.476750373840332,
    "timestamp": "2025-07-14T12:34:43.188413"
  },
  {
    "epoch": 1928,
    "batch_idx": 192,
    "loss": 0.3441551625728607,
    "learning_rate": 3.1943447399958e-06,
    "timestep_mean": 527.5,
    "epsilon_norm": 3.712360143661499,
    "prediction_norm": 2.5550878047943115,
    "timestamp": "2025-07-14T12:34:43.193457"
  },
  {
    "epoch": 1929,
    "batch_idx": 0,
    "loss": 0.289847195148468,
    "learning_rate": 3.106320488323264e-06,
    "timestep_mean": 525.40625,
    "epsilon_norm": 7.0419487953186035,
    "prediction_norm": 6.032527446746826,
    "timestamp": "2025-07-14T12:34:43.198467"
  },
  {
    "epoch": 1929,
    "batch_idx": 32,
    "loss": 0.21078111231327057,
    "learning_rate": 3.106320488323264e-06,
    "timestep_mean": 490.25,
    "epsilon_norm": 7.623115539550781,
    "prediction_norm": 6.692200660705566,
    "timestamp": "2025-07-14T12:34:43.203563"
  },
  {
    "epoch": 1929,
    "batch_idx": 64,
    "loss": 0.1514778584241867,
    "learning_rate": 3.106320488323264e-06,
    "timestep_mean": 541.625,
    "epsilon_norm": 7.088830947875977,
    "prediction_norm": 6.347288608551025,
    "timestamp": "2025-07-14T12:34:43.208600"
  },
  {
    "epoch": 1929,
    "batch_idx": 96,
    "loss": 0.21980544924736023,
    "learning_rate": 3.106320488323264e-06,
    "timestep_mean": 520.4375,
    "epsilon_norm": 7.074112415313721,
    "prediction_norm": 6.14797306060791,
    "timestamp": "2025-07-14T12:34:43.213625"
  },
  {
    "epoch": 1929,
    "batch_idx": 128,
    "loss": 0.24095088243484497,
    "learning_rate": 3.106320488323264e-06,
    "timestep_mean": 543.375,
    "epsilon_norm": 7.916896343231201,
    "prediction_norm": 7.214760780334473,
    "timestamp": "2025-07-14T12:34:43.218619"
  },
  {
    "epoch": 1929,
    "batch_idx": 160,
    "loss": 0.43548890948295593,
    "learning_rate": 3.106320488323264e-06,
    "timestep_mean": 457.5,
    "epsilon_norm": 6.653336524963379,
    "prediction_norm": 4.689603328704834,
    "timestamp": "2025-07-14T12:34:43.223578"
  },
  {
    "epoch": 1929,
    "batch_idx": 192,
    "loss": 0.4126184582710266,
    "learning_rate": 3.106320488323264e-06,
    "timestep_mean": 575.875,
    "epsilon_norm": 4.270605087280273,
    "prediction_norm": 3.4593429565429688,
    "timestamp": "2025-07-14T12:34:43.228480"
  },
  {
    "epoch": 1930,
    "batch_idx": 0,
    "loss": 0.36997032165527344,
    "learning_rate": 3.0195222724101443e-06,
    "timestep_mean": 460.5,
    "epsilon_norm": 8.873008728027344,
    "prediction_norm": 7.019825458526611,
    "timestamp": "2025-07-14T12:34:43.233371"
  },
  {
    "epoch": 1930,
    "batch_idx": 32,
    "loss": 0.21908819675445557,
    "learning_rate": 3.0195222724101443e-06,
    "timestep_mean": 481.78125,
    "epsilon_norm": 7.754136085510254,
    "prediction_norm": 6.640416145324707,
    "timestamp": "2025-07-14T12:34:43.238432"
  },
  {
    "epoch": 1930,
    "batch_idx": 64,
    "loss": 0.1309167444705963,
    "learning_rate": 3.0195222724101443e-06,
    "timestep_mean": 481.0,
    "epsilon_norm": 7.4101243019104,
    "prediction_norm": 6.611842155456543,
    "timestamp": "2025-07-14T12:34:43.243461"
  },
  {
    "epoch": 1930,
    "batch_idx": 96,
    "loss": 0.2736007869243622,
    "learning_rate": 3.0195222724101443e-06,
    "timestep_mean": 514.0,
    "epsilon_norm": 9.73688793182373,
    "prediction_norm": 8.403900146484375,
    "timestamp": "2025-07-14T12:34:43.248338"
  },
  {
    "epoch": 1930,
    "batch_idx": 128,
    "loss": 0.2998713254928589,
    "learning_rate": 3.0195222724101443e-06,
    "timestep_mean": 511.84375,
    "epsilon_norm": 8.192283630371094,
    "prediction_norm": 6.711546421051025,
    "timestamp": "2025-07-14T12:34:43.253290"
  },
  {
    "epoch": 1930,
    "batch_idx": 160,
    "loss": 0.2218436896800995,
    "learning_rate": 3.0195222724101443e-06,
    "timestep_mean": 481.28125,
    "epsilon_norm": 8.226494789123535,
    "prediction_norm": 7.033278465270996,
    "timestamp": "2025-07-14T12:34:43.258291"
  },
  {
    "epoch": 1930,
    "batch_idx": 192,
    "loss": 0.38629356026649475,
    "learning_rate": 3.0195222724101443e-06,
    "timestep_mean": 454.25,
    "epsilon_norm": 3.9043242931365967,
    "prediction_norm": 3.020418882369995,
    "timestamp": "2025-07-14T12:34:43.263256"
  },
  {
    "epoch": 1931,
    "batch_idx": 0,
    "loss": 0.20075564086437225,
    "learning_rate": 2.933950306422459e-06,
    "timestep_mean": 456.4375,
    "epsilon_norm": 9.164802551269531,
    "prediction_norm": 7.833528995513916,
    "timestamp": "2025-07-14T12:34:43.268211"
  },
  {
    "epoch": 1931,
    "batch_idx": 32,
    "loss": 0.1509893536567688,
    "learning_rate": 2.933950306422459e-06,
    "timestep_mean": 563.09375,
    "epsilon_norm": 7.225627899169922,
    "prediction_norm": 6.873103618621826,
    "timestamp": "2025-07-14T12:34:43.273194"
  },
  {
    "epoch": 1931,
    "batch_idx": 64,
    "loss": 0.32048243284225464,
    "learning_rate": 2.933950306422459e-06,
    "timestep_mean": 504.125,
    "epsilon_norm": 8.423260688781738,
    "prediction_norm": 6.770745277404785,
    "timestamp": "2025-07-14T12:34:43.278402"
  },
  {
    "epoch": 1931,
    "batch_idx": 96,
    "loss": 0.35354575514793396,
    "learning_rate": 2.933950306422459e-06,
    "timestep_mean": 462.15625,
    "epsilon_norm": 8.088927268981934,
    "prediction_norm": 6.917429447174072,
    "timestamp": "2025-07-14T12:34:43.284488"
  },
  {
    "epoch": 1931,
    "batch_idx": 128,
    "loss": 0.2314118891954422,
    "learning_rate": 2.933950306422459e-06,
    "timestep_mean": 399.15625,
    "epsilon_norm": 7.54761266708374,
    "prediction_norm": 6.461234092712402,
    "timestamp": "2025-07-14T12:34:43.298042"
  },
  {
    "epoch": 1931,
    "batch_idx": 160,
    "loss": 0.1843537986278534,
    "learning_rate": 2.933950306422459e-06,
    "timestep_mean": 506.0,
    "epsilon_norm": 7.547084808349609,
    "prediction_norm": 7.281471252441406,
    "timestamp": "2025-07-14T12:34:43.305852"
  },
  {
    "epoch": 1931,
    "batch_idx": 192,
    "loss": 0.07375223934650421,
    "learning_rate": 2.933950306422459e-06,
    "timestep_mean": 452.125,
    "epsilon_norm": 3.2677149772644043,
    "prediction_norm": 2.8602166175842285,
    "timestamp": "2025-07-14T12:34:43.313860"
  },
  {
    "epoch": 1932,
    "batch_idx": 0,
    "loss": 0.306709349155426,
    "learning_rate": 2.8496048015005356e-06,
    "timestep_mean": 482.875,
    "epsilon_norm": 7.650557041168213,
    "prediction_norm": 6.3099446296691895,
    "timestamp": "2025-07-14T12:34:43.322213"
  },
  {
    "epoch": 1932,
    "batch_idx": 32,
    "loss": 0.27651748061180115,
    "learning_rate": 2.8496048015005356e-06,
    "timestep_mean": 469.0625,
    "epsilon_norm": 7.805248737335205,
    "prediction_norm": 6.4008469581604,
    "timestamp": "2025-07-14T12:34:43.330466"
  },
  {
    "epoch": 1932,
    "batch_idx": 64,
    "loss": 0.46465158462524414,
    "learning_rate": 2.8496048015005356e-06,
    "timestep_mean": 498.21875,
    "epsilon_norm": 7.261544704437256,
    "prediction_norm": 5.064167022705078,
    "timestamp": "2025-07-14T12:34:43.338801"
  },
  {
    "epoch": 1932,
    "batch_idx": 96,
    "loss": 0.309845894575119,
    "learning_rate": 2.8496048015005356e-06,
    "timestep_mean": 498.15625,
    "epsilon_norm": 8.863624572753906,
    "prediction_norm": 7.2514872550964355,
    "timestamp": "2025-07-14T12:34:43.344547"
  },
  {
    "epoch": 1932,
    "batch_idx": 128,
    "loss": 0.25676843523979187,
    "learning_rate": 2.8496048015005356e-06,
    "timestep_mean": 493.03125,
    "epsilon_norm": 8.389327049255371,
    "prediction_norm": 7.1371283531188965,
    "timestamp": "2025-07-14T12:34:43.350969"
  },
  {
    "epoch": 1932,
    "batch_idx": 160,
    "loss": 0.42107412219047546,
    "learning_rate": 2.8496048015005356e-06,
    "timestep_mean": 398.125,
    "epsilon_norm": 9.684782028198242,
    "prediction_norm": 7.719531536102295,
    "timestamp": "2025-07-14T12:34:43.359712"
  },
  {
    "epoch": 1932,
    "batch_idx": 192,
    "loss": 0.2627834677696228,
    "learning_rate": 2.8496048015005356e-06,
    "timestep_mean": 498.125,
    "epsilon_norm": 4.990528106689453,
    "prediction_norm": 4.34975528717041,
    "timestamp": "2025-07-14T12:34:43.367112"
  },
  {
    "epoch": 1933,
    "batch_idx": 0,
    "loss": 0.5045086145401001,
    "learning_rate": 2.7664859657585116e-06,
    "timestep_mean": 485.75,
    "epsilon_norm": 8.783567428588867,
    "prediction_norm": 7.120483875274658,
    "timestamp": "2025-07-14T12:34:43.374219"
  },
  {
    "epoch": 1933,
    "batch_idx": 32,
    "loss": 0.27173352241516113,
    "learning_rate": 2.7664859657585116e-06,
    "timestep_mean": 476.34375,
    "epsilon_norm": 8.572188377380371,
    "prediction_norm": 7.170197010040283,
    "timestamp": "2025-07-14T12:34:43.380994"
  },
  {
    "epoch": 1933,
    "batch_idx": 64,
    "loss": 0.2263699620962143,
    "learning_rate": 2.7664859657585116e-06,
    "timestep_mean": 570.96875,
    "epsilon_norm": 7.451868534088135,
    "prediction_norm": 6.399338245391846,
    "timestamp": "2025-07-14T12:34:43.386682"
  },
  {
    "epoch": 1933,
    "batch_idx": 96,
    "loss": 0.08698569238185883,
    "learning_rate": 2.7664859657585116e-06,
    "timestep_mean": 504.75,
    "epsilon_norm": 6.913922309875488,
    "prediction_norm": 6.294865131378174,
    "timestamp": "2025-07-14T12:34:43.391778"
  },
  {
    "epoch": 1933,
    "batch_idx": 128,
    "loss": 0.25985968112945557,
    "learning_rate": 2.7664859657585116e-06,
    "timestep_mean": 483.4375,
    "epsilon_norm": 7.644737243652344,
    "prediction_norm": 6.704005718231201,
    "timestamp": "2025-07-14T12:34:43.400285"
  },
  {
    "epoch": 1933,
    "batch_idx": 160,
    "loss": 0.3269602060317993,
    "learning_rate": 2.7664859657585116e-06,
    "timestep_mean": 589.75,
    "epsilon_norm": 8.079594612121582,
    "prediction_norm": 7.195066928863525,
    "timestamp": "2025-07-14T12:34:43.405250"
  },
  {
    "epoch": 1933,
    "batch_idx": 192,
    "loss": 0.3614361882209778,
    "learning_rate": 2.7664859657585116e-06,
    "timestep_mean": 538.25,
    "epsilon_norm": 4.05083703994751,
    "prediction_norm": 3.1333529949188232,
    "timestamp": "2025-07-14T12:34:43.410338"
  },
  {
    "epoch": 1934,
    "batch_idx": 0,
    "loss": 0.15335023403167725,
    "learning_rate": 2.6845940042838335e-06,
    "timestep_mean": 471.375,
    "epsilon_norm": 7.481767654418945,
    "prediction_norm": 6.822381973266602,
    "timestamp": "2025-07-14T12:34:43.415513"
  },
  {
    "epoch": 1934,
    "batch_idx": 32,
    "loss": 0.2986757457256317,
    "learning_rate": 2.6845940042838335e-06,
    "timestep_mean": 542.78125,
    "epsilon_norm": 7.868951797485352,
    "prediction_norm": 6.792896747589111,
    "timestamp": "2025-07-14T12:34:43.420637"
  },
  {
    "epoch": 1934,
    "batch_idx": 64,
    "loss": 0.35428163409233093,
    "learning_rate": 2.6845940042838335e-06,
    "timestep_mean": 421.375,
    "epsilon_norm": 7.546436309814453,
    "prediction_norm": 6.025462627410889,
    "timestamp": "2025-07-14T12:34:43.425647"
  },
  {
    "epoch": 1934,
    "batch_idx": 96,
    "loss": 0.3393512964248657,
    "learning_rate": 2.6845940042838335e-06,
    "timestep_mean": 534.5,
    "epsilon_norm": 7.344135284423828,
    "prediction_norm": 5.80745267868042,
    "timestamp": "2025-07-14T12:34:43.430671"
  },
  {
    "epoch": 1934,
    "batch_idx": 128,
    "loss": 0.32069727778434753,
    "learning_rate": 2.6845940042838335e-06,
    "timestep_mean": 534.5,
    "epsilon_norm": 8.209602355957031,
    "prediction_norm": 6.824604511260986,
    "timestamp": "2025-07-14T12:34:43.435669"
  },
  {
    "epoch": 1934,
    "batch_idx": 160,
    "loss": 0.13003462553024292,
    "learning_rate": 2.6845940042838335e-06,
    "timestep_mean": 521.40625,
    "epsilon_norm": 7.863703727722168,
    "prediction_norm": 7.458086013793945,
    "timestamp": "2025-07-14T12:34:43.440545"
  },
  {
    "epoch": 1934,
    "batch_idx": 192,
    "loss": 0.0678749829530716,
    "learning_rate": 2.6845940042838335e-06,
    "timestep_mean": 739.0,
    "epsilon_norm": 5.627000331878662,
    "prediction_norm": 5.758316516876221,
    "timestamp": "2025-07-14T12:34:43.445705"
  },
  {
    "epoch": 1935,
    "batch_idx": 0,
    "loss": 0.15150907635688782,
    "learning_rate": 2.6039291191367586e-06,
    "timestep_mean": 592.46875,
    "epsilon_norm": 8.31285572052002,
    "prediction_norm": 7.790322780609131,
    "timestamp": "2025-07-14T12:34:43.450836"
  },
  {
    "epoch": 1935,
    "batch_idx": 32,
    "loss": 0.18625056743621826,
    "learning_rate": 2.6039291191367586e-06,
    "timestep_mean": 468.71875,
    "epsilon_norm": 7.828684329986572,
    "prediction_norm": 6.974286079406738,
    "timestamp": "2025-07-14T12:34:43.455869"
  },
  {
    "epoch": 1935,
    "batch_idx": 64,
    "loss": 0.18050971627235413,
    "learning_rate": 2.6039291191367586e-06,
    "timestep_mean": 529.0,
    "epsilon_norm": 7.741806983947754,
    "prediction_norm": 6.907247066497803,
    "timestamp": "2025-07-14T12:34:43.460828"
  },
  {
    "epoch": 1935,
    "batch_idx": 96,
    "loss": 0.22741296887397766,
    "learning_rate": 2.6039291191367586e-06,
    "timestep_mean": 530.0625,
    "epsilon_norm": 7.996261119842529,
    "prediction_norm": 6.9396562576293945,
    "timestamp": "2025-07-14T12:34:43.465799"
  },
  {
    "epoch": 1935,
    "batch_idx": 128,
    "loss": 0.1668892651796341,
    "learning_rate": 2.6039291191367586e-06,
    "timestep_mean": 639.09375,
    "epsilon_norm": 8.583627700805664,
    "prediction_norm": 7.884223937988281,
    "timestamp": "2025-07-14T12:34:43.470985"
  },
  {
    "epoch": 1935,
    "batch_idx": 160,
    "loss": 0.21921151876449585,
    "learning_rate": 2.6039291191367586e-06,
    "timestep_mean": 480.0625,
    "epsilon_norm": 7.881478786468506,
    "prediction_norm": 6.499585151672363,
    "timestamp": "2025-07-14T12:34:43.476206"
  },
  {
    "epoch": 1935,
    "batch_idx": 192,
    "loss": 0.14615429937839508,
    "learning_rate": 2.6039291191367586e-06,
    "timestep_mean": 512.875,
    "epsilon_norm": 3.2742245197296143,
    "prediction_norm": 3.6834661960601807,
    "timestamp": "2025-07-14T12:34:43.481419"
  },
  {
    "epoch": 1936,
    "batch_idx": 0,
    "loss": 0.3137829601764679,
    "learning_rate": 2.5244915093499107e-06,
    "timestep_mean": 552.15625,
    "epsilon_norm": 9.7598295211792,
    "prediction_norm": 8.686281204223633,
    "timestamp": "2025-07-14T12:34:43.486467"
  },
  {
    "epoch": 1936,
    "batch_idx": 32,
    "loss": 0.1997409164905548,
    "learning_rate": 2.5244915093499107e-06,
    "timestep_mean": 490.46875,
    "epsilon_norm": 7.383327960968018,
    "prediction_norm": 6.529659748077393,
    "timestamp": "2025-07-14T12:34:43.491636"
  },
  {
    "epoch": 1936,
    "batch_idx": 64,
    "loss": 0.22339047491550446,
    "learning_rate": 2.5244915093499107e-06,
    "timestep_mean": 466.90625,
    "epsilon_norm": 7.7465009689331055,
    "prediction_norm": 6.6405348777771,
    "timestamp": "2025-07-14T12:34:43.496939"
  },
  {
    "epoch": 1936,
    "batch_idx": 96,
    "loss": 0.10772918909788132,
    "learning_rate": 2.5244915093499107e-06,
    "timestep_mean": 496.9375,
    "epsilon_norm": 9.204263687133789,
    "prediction_norm": 8.769091606140137,
    "timestamp": "2025-07-14T12:34:43.502093"
  },
  {
    "epoch": 1936,
    "batch_idx": 128,
    "loss": 0.45020800828933716,
    "learning_rate": 2.5244915093499107e-06,
    "timestep_mean": 483.53125,
    "epsilon_norm": 8.379439353942871,
    "prediction_norm": 6.145814418792725,
    "timestamp": "2025-07-14T12:34:43.507238"
  },
  {
    "epoch": 1936,
    "batch_idx": 160,
    "loss": 0.26226866245269775,
    "learning_rate": 2.5244915093499107e-06,
    "timestep_mean": 495.375,
    "epsilon_norm": 9.50916576385498,
    "prediction_norm": 7.812980651855469,
    "timestamp": "2025-07-14T12:34:43.512422"
  },
  {
    "epoch": 1936,
    "batch_idx": 192,
    "loss": 0.28479811549186707,
    "learning_rate": 2.5244915093499107e-06,
    "timestep_mean": 526.875,
    "epsilon_norm": 4.874352931976318,
    "prediction_norm": 3.871206521987915,
    "timestamp": "2025-07-14T12:34:43.517961"
  },
  {
    "epoch": 1937,
    "batch_idx": 0,
    "loss": 0.24223023653030396,
    "learning_rate": 2.4462813709277253e-06,
    "timestep_mean": 503.625,
    "epsilon_norm": 8.119767189025879,
    "prediction_norm": 6.454833984375,
    "timestamp": "2025-07-14T12:34:43.523188"
  },
  {
    "epoch": 1937,
    "batch_idx": 32,
    "loss": 0.1634833663702011,
    "learning_rate": 2.4462813709277253e-06,
    "timestep_mean": 490.28125,
    "epsilon_norm": 7.896353244781494,
    "prediction_norm": 7.190402507781982,
    "timestamp": "2025-07-14T12:34:43.528354"
  },
  {
    "epoch": 1937,
    "batch_idx": 64,
    "loss": 0.2823636531829834,
    "learning_rate": 2.4462813709277253e-06,
    "timestep_mean": 443.21875,
    "epsilon_norm": 7.349455833435059,
    "prediction_norm": 6.125333309173584,
    "timestamp": "2025-07-14T12:34:43.533388"
  },
  {
    "epoch": 1937,
    "batch_idx": 96,
    "loss": 0.5175620913505554,
    "learning_rate": 2.4462813709277253e-06,
    "timestep_mean": 442.84375,
    "epsilon_norm": 9.228715896606445,
    "prediction_norm": 7.296397686004639,
    "timestamp": "2025-07-14T12:34:43.538578"
  },
  {
    "epoch": 1937,
    "batch_idx": 128,
    "loss": 0.2495439499616623,
    "learning_rate": 2.4462813709277253e-06,
    "timestep_mean": 520.90625,
    "epsilon_norm": 7.982577800750732,
    "prediction_norm": 6.657191753387451,
    "timestamp": "2025-07-14T12:34:43.543781"
  },
  {
    "epoch": 1937,
    "batch_idx": 160,
    "loss": 0.18747377395629883,
    "learning_rate": 2.4462813709277253e-06,
    "timestep_mean": 587.84375,
    "epsilon_norm": 7.245790958404541,
    "prediction_norm": 6.40253210067749,
    "timestamp": "2025-07-14T12:34:43.548896"
  },
  {
    "epoch": 1937,
    "batch_idx": 192,
    "loss": 0.2977442741394043,
    "learning_rate": 2.4462813709277253e-06,
    "timestep_mean": 465.5,
    "epsilon_norm": 2.468250274658203,
    "prediction_norm": 1.554914116859436,
    "timestamp": "2025-07-14T12:34:43.554164"
  },
  {
    "epoch": 1938,
    "batch_idx": 0,
    "loss": 0.24699918925762177,
    "learning_rate": 2.369298896845837e-06,
    "timestep_mean": 468.125,
    "epsilon_norm": 8.078601837158203,
    "prediction_norm": 6.61605167388916,
    "timestamp": "2025-07-14T12:34:43.559454"
  },
  {
    "epoch": 1938,
    "batch_idx": 32,
    "loss": 0.21993617713451385,
    "learning_rate": 2.369298896845837e-06,
    "timestep_mean": 545.59375,
    "epsilon_norm": 8.723678588867188,
    "prediction_norm": 7.659697532653809,
    "timestamp": "2025-07-14T12:34:43.564576"
  },
  {
    "epoch": 1938,
    "batch_idx": 64,
    "loss": 0.2379572093486786,
    "learning_rate": 2.369298896845837e-06,
    "timestep_mean": 384.90625,
    "epsilon_norm": 7.860568523406982,
    "prediction_norm": 6.603919982910156,
    "timestamp": "2025-07-14T12:34:43.569849"
  },
  {
    "epoch": 1938,
    "batch_idx": 96,
    "loss": 0.2656196355819702,
    "learning_rate": 2.369298896845837e-06,
    "timestep_mean": 530.875,
    "epsilon_norm": 8.695548057556152,
    "prediction_norm": 7.978938102722168,
    "timestamp": "2025-07-14T12:34:43.575071"
  },
  {
    "epoch": 1938,
    "batch_idx": 128,
    "loss": 0.18209728598594666,
    "learning_rate": 2.369298896845837e-06,
    "timestep_mean": 506.875,
    "epsilon_norm": 6.887004375457764,
    "prediction_norm": 5.96036434173584,
    "timestamp": "2025-07-14T12:34:43.580343"
  },
  {
    "epoch": 1938,
    "batch_idx": 160,
    "loss": 0.15662646293640137,
    "learning_rate": 2.369298896845837e-06,
    "timestep_mean": 587.78125,
    "epsilon_norm": 7.654196739196777,
    "prediction_norm": 6.808858871459961,
    "timestamp": "2025-07-14T12:34:43.585501"
  },
  {
    "epoch": 1938,
    "batch_idx": 192,
    "loss": 0.48001688718795776,
    "learning_rate": 2.369298896845837e-06,
    "timestep_mean": 604.875,
    "epsilon_norm": 5.581332683563232,
    "prediction_norm": 4.447033882141113,
    "timestamp": "2025-07-14T12:34:43.590479"
  },
  {
    "epoch": 1939,
    "batch_idx": 0,
    "loss": 0.19178009033203125,
    "learning_rate": 2.293544277050916e-06,
    "timestep_mean": 567.0625,
    "epsilon_norm": 8.216951370239258,
    "prediction_norm": 7.281284809112549,
    "timestamp": "2025-07-14T12:34:43.595830"
  },
  {
    "epoch": 1939,
    "batch_idx": 32,
    "loss": 0.19704672694206238,
    "learning_rate": 2.293544277050916e-06,
    "timestep_mean": 463.625,
    "epsilon_norm": 6.8609161376953125,
    "prediction_norm": 6.241525650024414,
    "timestamp": "2025-07-14T12:34:43.601376"
  },
  {
    "epoch": 1939,
    "batch_idx": 64,
    "loss": 0.25190985202789307,
    "learning_rate": 2.293544277050916e-06,
    "timestep_mean": 556.0,
    "epsilon_norm": 8.462493896484375,
    "prediction_norm": 7.1912522315979,
    "timestamp": "2025-07-14T12:34:43.606626"
  },
  {
    "epoch": 1939,
    "batch_idx": 96,
    "loss": 0.17045575380325317,
    "learning_rate": 2.293544277050916e-06,
    "timestep_mean": 532.1875,
    "epsilon_norm": 8.818790435791016,
    "prediction_norm": 7.833080291748047,
    "timestamp": "2025-07-14T12:34:43.612041"
  },
  {
    "epoch": 1939,
    "batch_idx": 128,
    "loss": 0.37380334734916687,
    "learning_rate": 2.293544277050916e-06,
    "timestep_mean": 503.40625,
    "epsilon_norm": 8.329916000366211,
    "prediction_norm": 6.34648323059082,
    "timestamp": "2025-07-14T12:34:43.617459"
  },
  {
    "epoch": 1939,
    "batch_idx": 160,
    "loss": 0.24108296632766724,
    "learning_rate": 2.293544277050916e-06,
    "timestep_mean": 549.375,
    "epsilon_norm": 7.492363452911377,
    "prediction_norm": 6.849090099334717,
    "timestamp": "2025-07-14T12:34:43.622579"
  },
  {
    "epoch": 1939,
    "batch_idx": 192,
    "loss": 0.4955686330795288,
    "learning_rate": 2.293544277050916e-06,
    "timestep_mean": 440.625,
    "epsilon_norm": 3.9412899017333984,
    "prediction_norm": 2.564204216003418,
    "timestamp": "2025-07-14T12:34:43.627808"
  },
  {
    "epoch": 1940,
    "batch_idx": 0,
    "loss": 0.25181037187576294,
    "learning_rate": 2.2190176984599993e-06,
    "timestep_mean": 490.21875,
    "epsilon_norm": 7.328791618347168,
    "prediction_norm": 6.010749816894531,
    "timestamp": "2025-07-14T12:34:43.633073"
  },
  {
    "epoch": 1940,
    "batch_idx": 32,
    "loss": 0.2647343873977661,
    "learning_rate": 2.2190176984599993e-06,
    "timestep_mean": 527.0,
    "epsilon_norm": 6.7807512283325195,
    "prediction_norm": 5.577188491821289,
    "timestamp": "2025-07-14T12:34:43.638247"
  },
  {
    "epoch": 1940,
    "batch_idx": 64,
    "loss": 0.35863661766052246,
    "learning_rate": 2.2190176984599993e-06,
    "timestep_mean": 429.9375,
    "epsilon_norm": 8.915919303894043,
    "prediction_norm": 7.934936046600342,
    "timestamp": "2025-07-14T12:34:43.643176"
  },
  {
    "epoch": 1940,
    "batch_idx": 96,
    "loss": 0.18379530310630798,
    "learning_rate": 2.2190176984599993e-06,
    "timestep_mean": 555.625,
    "epsilon_norm": 8.465253829956055,
    "prediction_norm": 7.711615085601807,
    "timestamp": "2025-07-14T12:34:43.648108"
  },
  {
    "epoch": 1940,
    "batch_idx": 128,
    "loss": 0.13164730370044708,
    "learning_rate": 2.2190176984599993e-06,
    "timestep_mean": 701.875,
    "epsilon_norm": 8.964252471923828,
    "prediction_norm": 8.852531433105469,
    "timestamp": "2025-07-14T12:34:43.653071"
  },
  {
    "epoch": 1940,
    "batch_idx": 160,
    "loss": 0.30166980624198914,
    "learning_rate": 2.2190176984599993e-06,
    "timestep_mean": 423.8125,
    "epsilon_norm": 7.313640117645264,
    "prediction_norm": 5.7579216957092285,
    "timestamp": "2025-07-14T12:34:43.658042"
  },
  {
    "epoch": 1940,
    "batch_idx": 192,
    "loss": 0.14751014113426208,
    "learning_rate": 2.2190176984599993e-06,
    "timestep_mean": 606.5,
    "epsilon_norm": 2.6860930919647217,
    "prediction_norm": 2.2269463539123535,
    "timestamp": "2025-07-14T12:34:43.663075"
  },
  {
    "epoch": 1941,
    "batch_idx": 0,
    "loss": 0.3754807710647583,
    "learning_rate": 2.1457193449599373e-06,
    "timestep_mean": 493.59375,
    "epsilon_norm": 9.72089958190918,
    "prediction_norm": 8.448731422424316,
    "timestamp": "2025-07-14T12:34:43.668076"
  },
  {
    "epoch": 1941,
    "batch_idx": 32,
    "loss": 0.3028886020183563,
    "learning_rate": 2.1457193449599373e-06,
    "timestep_mean": 478.96875,
    "epsilon_norm": 7.763558387756348,
    "prediction_norm": 5.943352222442627,
    "timestamp": "2025-07-14T12:34:43.673009"
  },
  {
    "epoch": 1941,
    "batch_idx": 64,
    "loss": 0.23486152291297913,
    "learning_rate": 2.1457193449599373e-06,
    "timestep_mean": 491.28125,
    "epsilon_norm": 7.424421310424805,
    "prediction_norm": 6.4737091064453125,
    "timestamp": "2025-07-14T12:34:43.677954"
  },
  {
    "epoch": 1941,
    "batch_idx": 96,
    "loss": 0.312469482421875,
    "learning_rate": 2.1457193449599373e-06,
    "timestep_mean": 613.4375,
    "epsilon_norm": 8.850249290466309,
    "prediction_norm": 7.687884330749512,
    "timestamp": "2025-07-14T12:34:43.682851"
  },
  {
    "epoch": 1941,
    "batch_idx": 128,
    "loss": 0.26848727464675903,
    "learning_rate": 2.1457193449599373e-06,
    "timestep_mean": 482.9375,
    "epsilon_norm": 7.38837194442749,
    "prediction_norm": 6.767220497131348,
    "timestamp": "2025-07-14T12:34:43.687888"
  },
  {
    "epoch": 1941,
    "batch_idx": 160,
    "loss": 0.2877476215362549,
    "learning_rate": 2.1457193449599373e-06,
    "timestep_mean": 375.375,
    "epsilon_norm": 6.925494194030762,
    "prediction_norm": 5.329498291015625,
    "timestamp": "2025-07-14T12:34:43.692889"
  },
  {
    "epoch": 1941,
    "batch_idx": 192,
    "loss": 0.2995568811893463,
    "learning_rate": 2.1457193449599373e-06,
    "timestep_mean": 440.875,
    "epsilon_norm": 2.5299882888793945,
    "prediction_norm": 2.7570230960845947,
    "timestamp": "2025-07-14T12:34:43.697805"
  },
  {
    "epoch": 1942,
    "batch_idx": 0,
    "loss": 0.09960627555847168,
    "learning_rate": 2.073649397407171e-06,
    "timestep_mean": 509.21875,
    "epsilon_norm": 7.357067108154297,
    "prediction_norm": 6.865749835968018,
    "timestamp": "2025-07-14T12:34:43.702805"
  },
  {
    "epoch": 1942,
    "batch_idx": 32,
    "loss": 0.17116251587867737,
    "learning_rate": 2.073649397407171e-06,
    "timestep_mean": 482.90625,
    "epsilon_norm": 6.846629619598389,
    "prediction_norm": 5.9281415939331055,
    "timestamp": "2025-07-14T12:34:43.707768"
  },
  {
    "epoch": 1942,
    "batch_idx": 64,
    "loss": 0.4502634108066559,
    "learning_rate": 2.073649397407171e-06,
    "timestep_mean": 517.8125,
    "epsilon_norm": 10.239572525024414,
    "prediction_norm": 8.830300331115723,
    "timestamp": "2025-07-14T12:34:43.713195"
  },
  {
    "epoch": 1942,
    "batch_idx": 96,
    "loss": 0.29064348340034485,
    "learning_rate": 2.073649397407171e-06,
    "timestep_mean": 519.1875,
    "epsilon_norm": 6.64932107925415,
    "prediction_norm": 5.154138565063477,
    "timestamp": "2025-07-14T12:34:43.718344"
  },
  {
    "epoch": 1942,
    "batch_idx": 128,
    "loss": 0.14949139952659607,
    "learning_rate": 2.073649397407171e-06,
    "timestep_mean": 531.625,
    "epsilon_norm": 7.7213826179504395,
    "prediction_norm": 7.233809947967529,
    "timestamp": "2025-07-14T12:34:43.723386"
  },
  {
    "epoch": 1942,
    "batch_idx": 160,
    "loss": 0.48574650287628174,
    "learning_rate": 2.073649397407171e-06,
    "timestep_mean": 540.46875,
    "epsilon_norm": 7.2822585105896,
    "prediction_norm": 5.636077880859375,
    "timestamp": "2025-07-14T12:34:43.728669"
  },
  {
    "epoch": 1942,
    "batch_idx": 192,
    "loss": 0.09034021198749542,
    "learning_rate": 2.073649397407171e-06,
    "timestep_mean": 573.25,
    "epsilon_norm": 3.9637017250061035,
    "prediction_norm": 3.958577871322632,
    "timestamp": "2025-07-14T12:34:43.734064"
  },
  {
    "epoch": 1943,
    "batch_idx": 0,
    "loss": 0.16748051345348358,
    "learning_rate": 2.0028080336271215e-06,
    "timestep_mean": 515.34375,
    "epsilon_norm": 6.898932456970215,
    "prediction_norm": 6.102477073669434,
    "timestamp": "2025-07-14T12:34:43.739556"
  },
  {
    "epoch": 1943,
    "batch_idx": 32,
    "loss": 0.1574978530406952,
    "learning_rate": 2.0028080336271215e-06,
    "timestep_mean": 592.875,
    "epsilon_norm": 7.44333028793335,
    "prediction_norm": 6.6973137855529785,
    "timestamp": "2025-07-14T12:34:43.745018"
  },
  {
    "epoch": 1943,
    "batch_idx": 64,
    "loss": 0.3445732593536377,
    "learning_rate": 2.0028080336271215e-06,
    "timestep_mean": 457.4375,
    "epsilon_norm": 8.127769470214844,
    "prediction_norm": 6.764100074768066,
    "timestamp": "2025-07-14T12:34:43.751366"
  },
  {
    "epoch": 1943,
    "batch_idx": 96,
    "loss": 0.23037900030612946,
    "learning_rate": 2.0028080336271215e-06,
    "timestep_mean": 542.5,
    "epsilon_norm": 8.072620391845703,
    "prediction_norm": 6.876641273498535,
    "timestamp": "2025-07-14T12:34:43.767021"
  },
  {
    "epoch": 1943,
    "batch_idx": 128,
    "loss": 0.22310291230678558,
    "learning_rate": 2.0028080336271215e-06,
    "timestep_mean": 564.15625,
    "epsilon_norm": 8.064867973327637,
    "prediction_norm": 6.95123291015625,
    "timestamp": "2025-07-14T12:34:43.774697"
  },
  {
    "epoch": 1943,
    "batch_idx": 160,
    "loss": 0.2984071373939514,
    "learning_rate": 2.0028080336271215e-06,
    "timestep_mean": 581.15625,
    "epsilon_norm": 8.414765357971191,
    "prediction_norm": 7.154215335845947,
    "timestamp": "2025-07-14T12:34:43.780929"
  },
  {
    "epoch": 1943,
    "batch_idx": 192,
    "loss": 0.17592491209506989,
    "learning_rate": 2.0028080336271215e-06,
    "timestep_mean": 480.5,
    "epsilon_norm": 4.184047698974609,
    "prediction_norm": 4.102627754211426,
    "timestamp": "2025-07-14T12:34:43.789557"
  },
  {
    "epoch": 1944,
    "batch_idx": 0,
    "loss": 0.23877090215682983,
    "learning_rate": 1.9331954284137454e-06,
    "timestep_mean": 537.9375,
    "epsilon_norm": 7.6274027824401855,
    "prediction_norm": 6.653286933898926,
    "timestamp": "2025-07-14T12:34:43.797702"
  },
  {
    "epoch": 1944,
    "batch_idx": 32,
    "loss": 0.329456627368927,
    "learning_rate": 1.9331954284137454e-06,
    "timestep_mean": 537.15625,
    "epsilon_norm": 8.4340238571167,
    "prediction_norm": 7.412973403930664,
    "timestamp": "2025-07-14T12:34:43.806080"
  },
  {
    "epoch": 1944,
    "batch_idx": 64,
    "loss": 0.24044519662857056,
    "learning_rate": 1.9331954284137454e-06,
    "timestep_mean": 476.4375,
    "epsilon_norm": 6.951119422912598,
    "prediction_norm": 6.227389335632324,
    "timestamp": "2025-07-14T12:34:43.815445"
  },
  {
    "epoch": 1944,
    "batch_idx": 96,
    "loss": 0.23855379223823547,
    "learning_rate": 1.9331954284137454e-06,
    "timestep_mean": 452.25,
    "epsilon_norm": 8.34510612487793,
    "prediction_norm": 6.6497721672058105,
    "timestamp": "2025-07-14T12:34:43.826454"
  },
  {
    "epoch": 1944,
    "batch_idx": 128,
    "loss": 0.36362192034721375,
    "learning_rate": 1.9331954284137454e-06,
    "timestep_mean": 388.0,
    "epsilon_norm": 6.166829586029053,
    "prediction_norm": 4.643474578857422,
    "timestamp": "2025-07-14T12:34:43.835054"
  },
  {
    "epoch": 1944,
    "batch_idx": 160,
    "loss": 0.5178329944610596,
    "learning_rate": 1.9331954284137454e-06,
    "timestep_mean": 414.5625,
    "epsilon_norm": 8.923919677734375,
    "prediction_norm": 6.63773250579834,
    "timestamp": "2025-07-14T12:34:43.843294"
  },
  {
    "epoch": 1944,
    "batch_idx": 192,
    "loss": 0.3791238069534302,
    "learning_rate": 1.9331954284137454e-06,
    "timestep_mean": 431.375,
    "epsilon_norm": 4.0006561279296875,
    "prediction_norm": 3.177974224090576,
    "timestamp": "2025-07-14T12:34:43.852007"
  },
  {
    "epoch": 1945,
    "batch_idx": 0,
    "loss": 0.13523316383361816,
    "learning_rate": 1.8648117535293699e-06,
    "timestep_mean": 557.40625,
    "epsilon_norm": 7.257567405700684,
    "prediction_norm": 6.563821792602539,
    "timestamp": "2025-07-14T12:34:43.859217"
  },
  {
    "epoch": 1945,
    "batch_idx": 32,
    "loss": 0.16257449984550476,
    "learning_rate": 1.8648117535293699e-06,
    "timestep_mean": 555.375,
    "epsilon_norm": 7.851776599884033,
    "prediction_norm": 6.627207279205322,
    "timestamp": "2025-07-14T12:34:43.864364"
  },
  {
    "epoch": 1945,
    "batch_idx": 64,
    "loss": 0.11537272483110428,
    "learning_rate": 1.8648117535293699e-06,
    "timestep_mean": 478.96875,
    "epsilon_norm": 8.48021125793457,
    "prediction_norm": 7.657227039337158,
    "timestamp": "2025-07-14T12:34:43.869481"
  },
  {
    "epoch": 1945,
    "batch_idx": 96,
    "loss": 0.27850624918937683,
    "learning_rate": 1.8648117535293699e-06,
    "timestep_mean": 455.25,
    "epsilon_norm": 7.271633625030518,
    "prediction_norm": 5.962432861328125,
    "timestamp": "2025-07-14T12:34:43.874692"
  },
  {
    "epoch": 1945,
    "batch_idx": 128,
    "loss": 0.3837059736251831,
    "learning_rate": 1.8648117535293699e-06,
    "timestep_mean": 502.1875,
    "epsilon_norm": 9.299866676330566,
    "prediction_norm": 7.659584045410156,
    "timestamp": "2025-07-14T12:34:43.879999"
  },
  {
    "epoch": 1945,
    "batch_idx": 160,
    "loss": 0.19137747585773468,
    "learning_rate": 1.8648117535293699e-06,
    "timestep_mean": 490.21875,
    "epsilon_norm": 7.268041610717773,
    "prediction_norm": 6.800970077514648,
    "timestamp": "2025-07-14T12:34:43.885045"
  },
  {
    "epoch": 1945,
    "batch_idx": 192,
    "loss": 0.23953627049922943,
    "learning_rate": 1.8648117535293699e-06,
    "timestep_mean": 507.625,
    "epsilon_norm": 3.643221378326416,
    "prediction_norm": 3.0013906955718994,
    "timestamp": "2025-07-14T12:34:43.890142"
  },
  {
    "epoch": 1946,
    "batch_idx": 0,
    "loss": 0.202798992395401,
    "learning_rate": 1.7976571777038024e-06,
    "timestep_mean": 489.125,
    "epsilon_norm": 7.862371921539307,
    "prediction_norm": 7.081092834472656,
    "timestamp": "2025-07-14T12:34:43.895229"
  },
  {
    "epoch": 1946,
    "batch_idx": 32,
    "loss": 0.3378217816352844,
    "learning_rate": 1.7976571777038024e-06,
    "timestep_mean": 524.28125,
    "epsilon_norm": 8.193819999694824,
    "prediction_norm": 6.843235015869141,
    "timestamp": "2025-07-14T12:34:43.900514"
  },
  {
    "epoch": 1946,
    "batch_idx": 64,
    "loss": 0.23813705146312714,
    "learning_rate": 1.7976571777038024e-06,
    "timestep_mean": 563.625,
    "epsilon_norm": 8.174118995666504,
    "prediction_norm": 6.896158695220947,
    "timestamp": "2025-07-14T12:34:43.905694"
  },
  {
    "epoch": 1946,
    "batch_idx": 96,
    "loss": 0.21979156136512756,
    "learning_rate": 1.7976571777038024e-06,
    "timestep_mean": 508.09375,
    "epsilon_norm": 8.731632232666016,
    "prediction_norm": 7.460968971252441,
    "timestamp": "2025-07-14T12:34:43.910820"
  },
  {
    "epoch": 1946,
    "batch_idx": 128,
    "loss": 0.19331228733062744,
    "learning_rate": 1.7976571777038024e-06,
    "timestep_mean": 568.78125,
    "epsilon_norm": 7.280989170074463,
    "prediction_norm": 6.2882184982299805,
    "timestamp": "2025-07-14T12:34:43.915840"
  },
  {
    "epoch": 1946,
    "batch_idx": 160,
    "loss": 0.2667770981788635,
    "learning_rate": 1.7976571777038024e-06,
    "timestep_mean": 517.84375,
    "epsilon_norm": 8.833762168884277,
    "prediction_norm": 7.930381774902344,
    "timestamp": "2025-07-14T12:34:43.920747"
  },
  {
    "epoch": 1946,
    "batch_idx": 192,
    "loss": 0.27922558784484863,
    "learning_rate": 1.7976571777038024e-06,
    "timestep_mean": 418.25,
    "epsilon_norm": 3.755791187286377,
    "prediction_norm": 3.717308759689331,
    "timestamp": "2025-07-14T12:34:43.925783"
  },
  {
    "epoch": 1947,
    "batch_idx": 0,
    "loss": 0.15217725932598114,
    "learning_rate": 1.7317318666342767e-06,
    "timestep_mean": 579.4375,
    "epsilon_norm": 7.053605556488037,
    "prediction_norm": 6.375019550323486,
    "timestamp": "2025-07-14T12:34:43.930887"
  },
  {
    "epoch": 1947,
    "batch_idx": 32,
    "loss": 0.20925530791282654,
    "learning_rate": 1.7317318666342767e-06,
    "timestep_mean": 425.34375,
    "epsilon_norm": 7.043837070465088,
    "prediction_norm": 5.868374824523926,
    "timestamp": "2025-07-14T12:34:43.935755"
  },
  {
    "epoch": 1947,
    "batch_idx": 64,
    "loss": 0.35869380831718445,
    "learning_rate": 1.7317318666342767e-06,
    "timestep_mean": 463.15625,
    "epsilon_norm": 8.01011848449707,
    "prediction_norm": 6.804812908172607,
    "timestamp": "2025-07-14T12:34:43.940645"
  },
  {
    "epoch": 1947,
    "batch_idx": 96,
    "loss": 0.27253225445747375,
    "learning_rate": 1.7317318666342767e-06,
    "timestep_mean": 512.40625,
    "epsilon_norm": 8.654777526855469,
    "prediction_norm": 7.779811382293701,
    "timestamp": "2025-07-14T12:34:43.945718"
  },
  {
    "epoch": 1947,
    "batch_idx": 128,
    "loss": 0.15113750100135803,
    "learning_rate": 1.7317318666342767e-06,
    "timestep_mean": 532.59375,
    "epsilon_norm": 5.725440502166748,
    "prediction_norm": 4.990197658538818,
    "timestamp": "2025-07-14T12:34:43.950856"
  },
  {
    "epoch": 1947,
    "batch_idx": 160,
    "loss": 0.23997019231319427,
    "learning_rate": 1.7317318666342767e-06,
    "timestep_mean": 462.5625,
    "epsilon_norm": 6.919471740722656,
    "prediction_norm": 6.002978324890137,
    "timestamp": "2025-07-14T12:34:43.955905"
  },
  {
    "epoch": 1947,
    "batch_idx": 192,
    "loss": 0.12760579586029053,
    "learning_rate": 1.7317318666342767e-06,
    "timestep_mean": 353.875,
    "epsilon_norm": 3.0283122062683105,
    "prediction_norm": 2.5755510330200195,
    "timestamp": "2025-07-14T12:34:43.961152"
  },
  {
    "epoch": 1948,
    "batch_idx": 0,
    "loss": 0.26808589696884155,
    "learning_rate": 1.667035982985064e-06,
    "timestep_mean": 487.25,
    "epsilon_norm": 7.388730049133301,
    "prediction_norm": 6.017757415771484,
    "timestamp": "2025-07-14T12:34:43.966241"
  },
  {
    "epoch": 1948,
    "batch_idx": 32,
    "loss": 0.2619744539260864,
    "learning_rate": 1.667035982985064e-06,
    "timestep_mean": 468.84375,
    "epsilon_norm": 8.55772590637207,
    "prediction_norm": 7.512818336486816,
    "timestamp": "2025-07-14T12:34:43.971307"
  },
  {
    "epoch": 1948,
    "batch_idx": 64,
    "loss": 0.1833934485912323,
    "learning_rate": 1.667035982985064e-06,
    "timestep_mean": 577.71875,
    "epsilon_norm": 8.309924125671387,
    "prediction_norm": 7.639723777770996,
    "timestamp": "2025-07-14T12:34:43.976225"
  },
  {
    "epoch": 1948,
    "batch_idx": 96,
    "loss": 0.3267884850502014,
    "learning_rate": 1.667035982985064e-06,
    "timestep_mean": 435.6875,
    "epsilon_norm": 8.53022289276123,
    "prediction_norm": 6.983760833740234,
    "timestamp": "2025-07-14T12:34:43.981180"
  },
  {
    "epoch": 1948,
    "batch_idx": 128,
    "loss": 0.17440098524093628,
    "learning_rate": 1.667035982985064e-06,
    "timestep_mean": 439.8125,
    "epsilon_norm": 7.165139675140381,
    "prediction_norm": 7.06232213973999,
    "timestamp": "2025-07-14T12:34:43.986130"
  },
  {
    "epoch": 1948,
    "batch_idx": 160,
    "loss": 0.27289819717407227,
    "learning_rate": 1.667035982985064e-06,
    "timestep_mean": 506.84375,
    "epsilon_norm": 8.297074317932129,
    "prediction_norm": 7.431614398956299,
    "timestamp": "2025-07-14T12:34:43.991125"
  },
  {
    "epoch": 1948,
    "batch_idx": 192,
    "loss": 0.06372754275798798,
    "learning_rate": 1.667035982985064e-06,
    "timestep_mean": 599.0,
    "epsilon_norm": 3.690682888031006,
    "prediction_norm": 3.5637340545654297,
    "timestamp": "2025-07-14T12:34:43.996055"
  },
  {
    "epoch": 1949,
    "batch_idx": 0,
    "loss": 0.1597209870815277,
    "learning_rate": 1.603569686386695e-06,
    "timestep_mean": 557.78125,
    "epsilon_norm": 7.490427017211914,
    "prediction_norm": 6.760505199432373,
    "timestamp": "2025-07-14T12:34:44.001010"
  },
  {
    "epoch": 1949,
    "batch_idx": 32,
    "loss": 0.14860935509204865,
    "learning_rate": 1.603569686386695e-06,
    "timestep_mean": 469.3125,
    "epsilon_norm": 7.746807098388672,
    "prediction_norm": 7.126211166381836,
    "timestamp": "2025-07-14T12:34:44.005875"
  },
  {
    "epoch": 1949,
    "batch_idx": 64,
    "loss": 0.20433203876018524,
    "learning_rate": 1.603569686386695e-06,
    "timestep_mean": 531.15625,
    "epsilon_norm": 7.839620113372803,
    "prediction_norm": 7.165179252624512,
    "timestamp": "2025-07-14T12:34:44.010704"
  },
  {
    "epoch": 1949,
    "batch_idx": 96,
    "loss": 0.10389526188373566,
    "learning_rate": 1.603569686386695e-06,
    "timestep_mean": 574.0625,
    "epsilon_norm": 7.888815879821777,
    "prediction_norm": 7.293139457702637,
    "timestamp": "2025-07-14T12:34:44.015710"
  },
  {
    "epoch": 1949,
    "batch_idx": 128,
    "loss": 0.24258701503276825,
    "learning_rate": 1.603569686386695e-06,
    "timestep_mean": 576.03125,
    "epsilon_norm": 9.36144733428955,
    "prediction_norm": 8.697628021240234,
    "timestamp": "2025-07-14T12:34:44.020567"
  },
  {
    "epoch": 1949,
    "batch_idx": 160,
    "loss": 0.3106881082057953,
    "learning_rate": 1.603569686386695e-06,
    "timestep_mean": 477.34375,
    "epsilon_norm": 8.526634216308594,
    "prediction_norm": 7.0948333740234375,
    "timestamp": "2025-07-14T12:34:44.025626"
  },
  {
    "epoch": 1949,
    "batch_idx": 192,
    "loss": 0.1548345685005188,
    "learning_rate": 1.603569686386695e-06,
    "timestep_mean": 578.875,
    "epsilon_norm": 4.822792053222656,
    "prediction_norm": 4.568353652954102,
    "timestamp": "2025-07-14T12:34:44.030569"
  },
  {
    "epoch": 1950,
    "batch_idx": 0,
    "loss": 0.30195608735084534,
    "learning_rate": 1.5413331334360165e-06,
    "timestep_mean": 577.28125,
    "epsilon_norm": 8.05123519897461,
    "prediction_norm": 6.617612838745117,
    "timestamp": "2025-07-14T12:34:44.035507"
  },
  {
    "epoch": 1950,
    "batch_idx": 32,
    "loss": 0.3777703046798706,
    "learning_rate": 1.5413331334360165e-06,
    "timestep_mean": 498.9375,
    "epsilon_norm": 9.390048027038574,
    "prediction_norm": 6.958252906799316,
    "timestamp": "2025-07-14T12:34:44.040514"
  },
  {
    "epoch": 1950,
    "batch_idx": 64,
    "loss": 0.16483280062675476,
    "learning_rate": 1.5413331334360165e-06,
    "timestep_mean": 546.4375,
    "epsilon_norm": 7.845524787902832,
    "prediction_norm": 7.231486797332764,
    "timestamp": "2025-07-14T12:34:44.045821"
  },
  {
    "epoch": 1950,
    "batch_idx": 96,
    "loss": 0.22481787204742432,
    "learning_rate": 1.5413331334360165e-06,
    "timestep_mean": 470.25,
    "epsilon_norm": 8.34424114227295,
    "prediction_norm": 7.3105149269104,
    "timestamp": "2025-07-14T12:34:44.051088"
  },
  {
    "epoch": 1950,
    "batch_idx": 128,
    "loss": 0.28955844044685364,
    "learning_rate": 1.5413331334360165e-06,
    "timestep_mean": 542.65625,
    "epsilon_norm": 8.494855880737305,
    "prediction_norm": 7.148633003234863,
    "timestamp": "2025-07-14T12:34:44.056173"
  },
  {
    "epoch": 1950,
    "batch_idx": 160,
    "loss": 0.4103485941886902,
    "learning_rate": 1.5413331334360165e-06,
    "timestep_mean": 462.5,
    "epsilon_norm": 7.906599998474121,
    "prediction_norm": 6.618727684020996,
    "timestamp": "2025-07-14T12:34:44.061412"
  },
  {
    "epoch": 1950,
    "batch_idx": 192,
    "loss": 0.06692678481340408,
    "learning_rate": 1.5413331334360165e-06,
    "timestep_mean": 704.875,
    "epsilon_norm": 4.24833869934082,
    "prediction_norm": 4.007658004760742,
    "timestamp": "2025-07-14T12:34:44.066591"
  },
  {
    "epoch": 1951,
    "batch_idx": 0,
    "loss": 0.26128819584846497,
    "learning_rate": 1.4803264776955246e-06,
    "timestep_mean": 455.09375,
    "epsilon_norm": 7.5753889083862305,
    "prediction_norm": 6.147903919219971,
    "timestamp": "2025-07-14T12:34:44.072020"
  },
  {
    "epoch": 1951,
    "batch_idx": 32,
    "loss": 0.40809333324432373,
    "learning_rate": 1.4803264776955246e-06,
    "timestep_mean": 487.28125,
    "epsilon_norm": 8.750130653381348,
    "prediction_norm": 6.925313949584961,
    "timestamp": "2025-07-14T12:34:44.077188"
  },
  {
    "epoch": 1951,
    "batch_idx": 64,
    "loss": 0.24469350278377533,
    "learning_rate": 1.4803264776955246e-06,
    "timestep_mean": 426.78125,
    "epsilon_norm": 7.777502059936523,
    "prediction_norm": 6.362460613250732,
    "timestamp": "2025-07-14T12:34:44.082356"
  },
  {
    "epoch": 1951,
    "batch_idx": 96,
    "loss": 0.15799455344676971,
    "learning_rate": 1.4803264776955246e-06,
    "timestep_mean": 577.34375,
    "epsilon_norm": 7.228629112243652,
    "prediction_norm": 6.480494499206543,
    "timestamp": "2025-07-14T12:34:44.087603"
  },
  {
    "epoch": 1951,
    "batch_idx": 128,
    "loss": 0.3309710621833801,
    "learning_rate": 1.4803264776955246e-06,
    "timestep_mean": 436.75,
    "epsilon_norm": 8.48890495300293,
    "prediction_norm": 6.844480037689209,
    "timestamp": "2025-07-14T12:34:44.092851"
  },
  {
    "epoch": 1951,
    "batch_idx": 160,
    "loss": 0.12725913524627686,
    "learning_rate": 1.4803264776955246e-06,
    "timestep_mean": 594.28125,
    "epsilon_norm": 7.419437408447266,
    "prediction_norm": 6.544203281402588,
    "timestamp": "2025-07-14T12:34:44.098053"
  },
  {
    "epoch": 1951,
    "batch_idx": 192,
    "loss": 0.22229787707328796,
    "learning_rate": 1.4803264776955246e-06,
    "timestep_mean": 404.625,
    "epsilon_norm": 3.34950852394104,
    "prediction_norm": 3.2776739597320557,
    "timestamp": "2025-07-14T12:34:44.103251"
  },
  {
    "epoch": 1952,
    "batch_idx": 0,
    "loss": 0.3832462430000305,
    "learning_rate": 1.4205498696930317e-06,
    "timestep_mean": 459.3125,
    "epsilon_norm": 8.797197341918945,
    "prediction_norm": 7.0281662940979,
    "timestamp": "2025-07-14T12:34:44.108607"
  },
  {
    "epoch": 1952,
    "batch_idx": 32,
    "loss": 0.25107240676879883,
    "learning_rate": 1.4205498696930317e-06,
    "timestep_mean": 514.90625,
    "epsilon_norm": 7.247507095336914,
    "prediction_norm": 6.579438209533691,
    "timestamp": "2025-07-14T12:34:44.113703"
  },
  {
    "epoch": 1952,
    "batch_idx": 64,
    "loss": 0.3325618803501129,
    "learning_rate": 1.4205498696930317e-06,
    "timestep_mean": 496.65625,
    "epsilon_norm": 8.59769344329834,
    "prediction_norm": 7.184231281280518,
    "timestamp": "2025-07-14T12:34:44.118640"
  },
  {
    "epoch": 1952,
    "batch_idx": 96,
    "loss": 0.39291125535964966,
    "learning_rate": 1.4205498696930317e-06,
    "timestep_mean": 455.6875,
    "epsilon_norm": 7.706413745880127,
    "prediction_norm": 5.766182899475098,
    "timestamp": "2025-07-14T12:34:44.126608"
  },
  {
    "epoch": 1952,
    "batch_idx": 128,
    "loss": 0.2987755537033081,
    "learning_rate": 1.4205498696930317e-06,
    "timestep_mean": 542.0,
    "epsilon_norm": 7.250068187713623,
    "prediction_norm": 6.21492862701416,
    "timestamp": "2025-07-14T12:34:44.137344"
  },
  {
    "epoch": 1952,
    "batch_idx": 160,
    "loss": 0.1688399612903595,
    "learning_rate": 1.4205498696930317e-06,
    "timestep_mean": 526.0,
    "epsilon_norm": 6.394712448120117,
    "prediction_norm": 5.77426815032959,
    "timestamp": "2025-07-14T12:34:44.146336"
  },
  {
    "epoch": 1952,
    "batch_idx": 192,
    "loss": 0.3862382769584656,
    "learning_rate": 1.4205498696930317e-06,
    "timestep_mean": 475.375,
    "epsilon_norm": 4.4661712646484375,
    "prediction_norm": 3.7367777824401855,
    "timestamp": "2025-07-14T12:34:44.155102"
  },
  {
    "epoch": 1953,
    "batch_idx": 0,
    "loss": 0.4404450058937073,
    "learning_rate": 1.3620034569214445e-06,
    "timestep_mean": 457.96875,
    "epsilon_norm": 9.421457290649414,
    "prediction_norm": 7.65404748916626,
    "timestamp": "2025-07-14T12:34:44.161912"
  },
  {
    "epoch": 1953,
    "batch_idx": 32,
    "loss": 0.19835427403450012,
    "learning_rate": 1.3620034569214445e-06,
    "timestep_mean": 480.21875,
    "epsilon_norm": 8.275771141052246,
    "prediction_norm": 7.513795375823975,
    "timestamp": "2025-07-14T12:34:44.167951"
  },
  {
    "epoch": 1953,
    "batch_idx": 64,
    "loss": 0.27930450439453125,
    "learning_rate": 1.3620034569214445e-06,
    "timestep_mean": 552.40625,
    "epsilon_norm": 8.043991088867188,
    "prediction_norm": 6.657710075378418,
    "timestamp": "2025-07-14T12:34:44.173177"
  },
  {
    "epoch": 1953,
    "batch_idx": 96,
    "loss": 0.3116524815559387,
    "learning_rate": 1.3620034569214445e-06,
    "timestep_mean": 484.65625,
    "epsilon_norm": 7.718799114227295,
    "prediction_norm": 5.8571882247924805,
    "timestamp": "2025-07-14T12:34:44.178510"
  },
  {
    "epoch": 1953,
    "batch_idx": 128,
    "loss": 0.22047626972198486,
    "learning_rate": 1.3620034569214445e-06,
    "timestep_mean": 573.46875,
    "epsilon_norm": 7.505849838256836,
    "prediction_norm": 6.72905969619751,
    "timestamp": "2025-07-14T12:34:44.183626"
  },
  {
    "epoch": 1953,
    "batch_idx": 160,
    "loss": 0.26767435669898987,
    "learning_rate": 1.3620034569214445e-06,
    "timestep_mean": 479.71875,
    "epsilon_norm": 6.761505126953125,
    "prediction_norm": 5.999298572540283,
    "timestamp": "2025-07-14T12:34:44.188772"
  },
  {
    "epoch": 1953,
    "batch_idx": 192,
    "loss": 0.24385026097297668,
    "learning_rate": 1.3620034569214445e-06,
    "timestep_mean": 374.625,
    "epsilon_norm": 4.02559757232666,
    "prediction_norm": 3.4802074432373047,
    "timestamp": "2025-07-14T12:34:44.194204"
  },
  {
    "epoch": 1954,
    "batch_idx": 0,
    "loss": 0.16014613211154938,
    "learning_rate": 1.3046873838381533e-06,
    "timestep_mean": 596.9375,
    "epsilon_norm": 8.789350509643555,
    "prediction_norm": 8.187686920166016,
    "timestamp": "2025-07-14T12:34:44.199204"
  },
  {
    "epoch": 1954,
    "batch_idx": 32,
    "loss": 0.44670531153678894,
    "learning_rate": 1.3046873838381533e-06,
    "timestep_mean": 455.9375,
    "epsilon_norm": 7.595305919647217,
    "prediction_norm": 5.7717604637146,
    "timestamp": "2025-07-14T12:34:44.204261"
  },
  {
    "epoch": 1954,
    "batch_idx": 64,
    "loss": 0.1610351800918579,
    "learning_rate": 1.3046873838381533e-06,
    "timestep_mean": 561.125,
    "epsilon_norm": 7.513284683227539,
    "prediction_norm": 6.882268905639648,
    "timestamp": "2025-07-14T12:34:44.209424"
  },
  {
    "epoch": 1954,
    "batch_idx": 96,
    "loss": 0.3254132866859436,
    "learning_rate": 1.3046873838381533e-06,
    "timestep_mean": 509.25,
    "epsilon_norm": 7.46484375,
    "prediction_norm": 6.145943641662598,
    "timestamp": "2025-07-14T12:34:44.214338"
  },
  {
    "epoch": 1954,
    "batch_idx": 128,
    "loss": 0.1950196921825409,
    "learning_rate": 1.3046873838381533e-06,
    "timestep_mean": 514.9375,
    "epsilon_norm": 7.742887020111084,
    "prediction_norm": 7.192530632019043,
    "timestamp": "2025-07-14T12:34:44.219171"
  },
  {
    "epoch": 1954,
    "batch_idx": 160,
    "loss": 0.04761160537600517,
    "learning_rate": 1.3046873838381533e-06,
    "timestep_mean": 676.59375,
    "epsilon_norm": 7.585770130157471,
    "prediction_norm": 7.72011137008667,
    "timestamp": "2025-07-14T12:34:44.224048"
  },
  {
    "epoch": 1954,
    "batch_idx": 192,
    "loss": 0.26798319816589355,
    "learning_rate": 1.3046873838381533e-06,
    "timestep_mean": 399.5,
    "epsilon_norm": 3.4157590866088867,
    "prediction_norm": 2.699711322784424,
    "timestamp": "2025-07-14T12:34:44.229324"
  },
  {
    "epoch": 1955,
    "batch_idx": 0,
    "loss": 0.3564070761203766,
    "learning_rate": 1.2486017918649772e-06,
    "timestep_mean": 523.875,
    "epsilon_norm": 8.071276664733887,
    "prediction_norm": 6.8180975914001465,
    "timestamp": "2025-07-14T12:34:44.234276"
  },
  {
    "epoch": 1955,
    "batch_idx": 32,
    "loss": 0.14522354304790497,
    "learning_rate": 1.2486017918649772e-06,
    "timestep_mean": 503.125,
    "epsilon_norm": 7.0242156982421875,
    "prediction_norm": 6.386881351470947,
    "timestamp": "2025-07-14T12:34:44.239234"
  },
  {
    "epoch": 1955,
    "batch_idx": 64,
    "loss": 0.313209593296051,
    "learning_rate": 1.2486017918649772e-06,
    "timestep_mean": 535.8125,
    "epsilon_norm": 8.687053680419922,
    "prediction_norm": 7.78331184387207,
    "timestamp": "2025-07-14T12:34:44.244214"
  },
  {
    "epoch": 1955,
    "batch_idx": 96,
    "loss": 0.2491760402917862,
    "learning_rate": 1.2486017918649772e-06,
    "timestep_mean": 561.53125,
    "epsilon_norm": 7.575342655181885,
    "prediction_norm": 7.324240684509277,
    "timestamp": "2025-07-14T12:34:44.249338"
  },
  {
    "epoch": 1955,
    "batch_idx": 128,
    "loss": 0.2592296600341797,
    "learning_rate": 1.2486017918649772e-06,
    "timestep_mean": 482.625,
    "epsilon_norm": 9.489150047302246,
    "prediction_norm": 8.377401351928711,
    "timestamp": "2025-07-14T12:34:44.254392"
  },
  {
    "epoch": 1955,
    "batch_idx": 160,
    "loss": 0.28144630789756775,
    "learning_rate": 1.2486017918649772e-06,
    "timestep_mean": 458.40625,
    "epsilon_norm": 7.880785942077637,
    "prediction_norm": 6.412370681762695,
    "timestamp": "2025-07-14T12:34:44.259520"
  },
  {
    "epoch": 1955,
    "batch_idx": 192,
    "loss": 0.22499260306358337,
    "learning_rate": 1.2486017918649772e-06,
    "timestep_mean": 580.75,
    "epsilon_norm": 3.174199342727661,
    "prediction_norm": 2.655965566635132,
    "timestamp": "2025-07-14T12:34:44.264607"
  },
  {
    "epoch": 1956,
    "batch_idx": 0,
    "loss": 0.16464529931545258,
    "learning_rate": 1.1937468193873857e-06,
    "timestep_mean": 449.9375,
    "epsilon_norm": 7.456414222717285,
    "prediction_norm": 6.961733818054199,
    "timestamp": "2025-07-14T12:34:44.269998"
  },
  {
    "epoch": 1956,
    "batch_idx": 32,
    "loss": 0.29824045300483704,
    "learning_rate": 1.1937468193873857e-06,
    "timestep_mean": 568.125,
    "epsilon_norm": 8.860553741455078,
    "prediction_norm": 6.876697063446045,
    "timestamp": "2025-07-14T12:34:44.274980"
  },
  {
    "epoch": 1956,
    "batch_idx": 64,
    "loss": 0.15102896094322205,
    "learning_rate": 1.1937468193873857e-06,
    "timestep_mean": 516.8125,
    "epsilon_norm": 7.189404487609863,
    "prediction_norm": 6.580688953399658,
    "timestamp": "2025-07-14T12:34:44.280341"
  },
  {
    "epoch": 1956,
    "batch_idx": 96,
    "loss": 0.5654453635215759,
    "learning_rate": 1.1937468193873857e-06,
    "timestep_mean": 451.5625,
    "epsilon_norm": 9.056077003479004,
    "prediction_norm": 6.441305637359619,
    "timestamp": "2025-07-14T12:34:44.285474"
  },
  {
    "epoch": 1956,
    "batch_idx": 128,
    "loss": 0.22237838804721832,
    "learning_rate": 1.1937468193873857e-06,
    "timestep_mean": 479.90625,
    "epsilon_norm": 9.141672134399414,
    "prediction_norm": 8.050495147705078,
    "timestamp": "2025-07-14T12:34:44.290624"
  },
  {
    "epoch": 1956,
    "batch_idx": 160,
    "loss": 0.5224410891532898,
    "learning_rate": 1.1937468193873857e-06,
    "timestep_mean": 464.625,
    "epsilon_norm": 9.369439125061035,
    "prediction_norm": 7.66983699798584,
    "timestamp": "2025-07-14T12:34:44.295822"
  },
  {
    "epoch": 1956,
    "batch_idx": 192,
    "loss": 0.011255604214966297,
    "learning_rate": 1.1937468193873857e-06,
    "timestep_mean": 858.75,
    "epsilon_norm": 4.299199104309082,
    "prediction_norm": 4.4189958572387695,
    "timestamp": "2025-07-14T12:34:44.301005"
  },
  {
    "epoch": 1957,
    "batch_idx": 0,
    "loss": 0.2528706192970276,
    "learning_rate": 1.1401226017547218e-06,
    "timestep_mean": 488.03125,
    "epsilon_norm": 8.34118938446045,
    "prediction_norm": 7.477991580963135,
    "timestamp": "2025-07-14T12:34:44.306208"
  },
  {
    "epoch": 1957,
    "batch_idx": 32,
    "loss": 0.29670998454093933,
    "learning_rate": 1.1401226017547218e-06,
    "timestep_mean": 514.0625,
    "epsilon_norm": 7.990037441253662,
    "prediction_norm": 6.978307247161865,
    "timestamp": "2025-07-14T12:34:44.311407"
  },
  {
    "epoch": 1957,
    "batch_idx": 64,
    "loss": 0.1380656510591507,
    "learning_rate": 1.1401226017547218e-06,
    "timestep_mean": 483.28125,
    "epsilon_norm": 7.354637622833252,
    "prediction_norm": 6.93939733505249,
    "timestamp": "2025-07-14T12:34:44.316701"
  },
  {
    "epoch": 1957,
    "batch_idx": 96,
    "loss": 0.2833667993545532,
    "learning_rate": 1.1401226017547218e-06,
    "timestep_mean": 573.65625,
    "epsilon_norm": 8.834548950195312,
    "prediction_norm": 7.693986892700195,
    "timestamp": "2025-07-14T12:34:44.321832"
  },
  {
    "epoch": 1957,
    "batch_idx": 128,
    "loss": 0.20238731801509857,
    "learning_rate": 1.1401226017547218e-06,
    "timestep_mean": 599.15625,
    "epsilon_norm": 9.560819625854492,
    "prediction_norm": 8.744606971740723,
    "timestamp": "2025-07-14T12:34:44.326928"
  },
  {
    "epoch": 1957,
    "batch_idx": 160,
    "loss": 0.347479909658432,
    "learning_rate": 1.1401226017547218e-06,
    "timestep_mean": 439.4375,
    "epsilon_norm": 7.087766647338867,
    "prediction_norm": 6.157472610473633,
    "timestamp": "2025-07-14T12:34:44.331852"
  },
  {
    "epoch": 1957,
    "batch_idx": 192,
    "loss": 0.03305024653673172,
    "learning_rate": 1.1401226017547218e-06,
    "timestep_mean": 543.625,
    "epsilon_norm": 3.2262701988220215,
    "prediction_norm": 3.242264747619629,
    "timestamp": "2025-07-14T12:34:44.337118"
  },
  {
    "epoch": 1958,
    "batch_idx": 0,
    "loss": 0.4545547366142273,
    "learning_rate": 1.0877292712792576e-06,
    "timestep_mean": 534.5,
    "epsilon_norm": 8.73349380493164,
    "prediction_norm": 7.721838474273682,
    "timestamp": "2025-07-14T12:34:44.342303"
  },
  {
    "epoch": 1958,
    "batch_idx": 32,
    "loss": 0.3853659927845001,
    "learning_rate": 1.0877292712792576e-06,
    "timestep_mean": 548.21875,
    "epsilon_norm": 8.133084297180176,
    "prediction_norm": 7.06045389175415,
    "timestamp": "2025-07-14T12:34:44.347538"
  },
  {
    "epoch": 1958,
    "batch_idx": 64,
    "loss": 0.32747119665145874,
    "learning_rate": 1.0877292712792576e-06,
    "timestep_mean": 459.9375,
    "epsilon_norm": 7.359119892120361,
    "prediction_norm": 5.782454013824463,
    "timestamp": "2025-07-14T12:34:44.352649"
  },
  {
    "epoch": 1958,
    "batch_idx": 96,
    "loss": 0.3067971467971802,
    "learning_rate": 1.0877292712792576e-06,
    "timestep_mean": 422.875,
    "epsilon_norm": 7.447147846221924,
    "prediction_norm": 5.883206367492676,
    "timestamp": "2025-07-14T12:34:44.357640"
  },
  {
    "epoch": 1958,
    "batch_idx": 128,
    "loss": 0.19340862333774567,
    "learning_rate": 1.0877292712792576e-06,
    "timestep_mean": 430.40625,
    "epsilon_norm": 7.221312999725342,
    "prediction_norm": 5.721724510192871,
    "timestamp": "2025-07-14T12:34:44.362725"
  },
  {
    "epoch": 1958,
    "batch_idx": 160,
    "loss": 0.3068636357784271,
    "learning_rate": 1.0877292712792576e-06,
    "timestep_mean": 506.5625,
    "epsilon_norm": 7.79095458984375,
    "prediction_norm": 7.091954708099365,
    "timestamp": "2025-07-14T12:34:44.367857"
  },
  {
    "epoch": 1958,
    "batch_idx": 192,
    "loss": 0.6364165544509888,
    "learning_rate": 1.0877292712792576e-06,
    "timestep_mean": 570.125,
    "epsilon_norm": 3.9220547676086426,
    "prediction_norm": 2.848220109939575,
    "timestamp": "2025-07-14T12:34:44.372992"
  },
  {
    "epoch": 1959,
    "batch_idx": 0,
    "loss": 0.2929658591747284,
    "learning_rate": 1.0365669572364724e-06,
    "timestep_mean": 513.90625,
    "epsilon_norm": 7.783982753753662,
    "prediction_norm": 6.198765754699707,
    "timestamp": "2025-07-14T12:34:44.378217"
  },
  {
    "epoch": 1959,
    "batch_idx": 32,
    "loss": 0.18518617749214172,
    "learning_rate": 1.0365669572364724e-06,
    "timestep_mean": 527.46875,
    "epsilon_norm": 6.897392749786377,
    "prediction_norm": 6.165916919708252,
    "timestamp": "2025-07-14T12:34:44.383526"
  },
  {
    "epoch": 1959,
    "batch_idx": 64,
    "loss": 0.4770580232143402,
    "learning_rate": 1.0365669572364724e-06,
    "timestep_mean": 360.78125,
    "epsilon_norm": 8.517430305480957,
    "prediction_norm": 6.481043338775635,
    "timestamp": "2025-07-14T12:34:44.388984"
  },
  {
    "epoch": 1959,
    "batch_idx": 96,
    "loss": 0.3193626403808594,
    "learning_rate": 1.0365669572364724e-06,
    "timestep_mean": 420.03125,
    "epsilon_norm": 8.202617645263672,
    "prediction_norm": 6.879189968109131,
    "timestamp": "2025-07-14T12:34:44.394201"
  },
  {
    "epoch": 1959,
    "batch_idx": 128,
    "loss": 0.40108421444892883,
    "learning_rate": 1.0365669572364724e-06,
    "timestep_mean": 386.59375,
    "epsilon_norm": 8.388351440429688,
    "prediction_norm": 7.015149116516113,
    "timestamp": "2025-07-14T12:34:44.399522"
  },
  {
    "epoch": 1959,
    "batch_idx": 160,
    "loss": 0.1431419849395752,
    "learning_rate": 1.0365669572364724e-06,
    "timestep_mean": 487.8125,
    "epsilon_norm": 6.555261135101318,
    "prediction_norm": 5.879915237426758,
    "timestamp": "2025-07-14T12:34:44.404647"
  },
  {
    "epoch": 1959,
    "batch_idx": 192,
    "loss": 0.20165950059890747,
    "learning_rate": 1.0365669572364724e-06,
    "timestep_mean": 640.125,
    "epsilon_norm": 3.5120956897735596,
    "prediction_norm": 3.268768787384033,
    "timestamp": "2025-07-14T12:34:44.409602"
  },
  {
    "epoch": 1960,
    "batch_idx": 0,
    "loss": 0.27054232358932495,
    "learning_rate": 9.866357858642196e-07,
    "timestep_mean": 488.46875,
    "epsilon_norm": 8.19715690612793,
    "prediction_norm": 7.33282470703125,
    "timestamp": "2025-07-14T12:34:44.414732"
  },
  {
    "epoch": 1960,
    "batch_idx": 32,
    "loss": 0.2392888218164444,
    "learning_rate": 9.866357858642196e-07,
    "timestep_mean": 555.1875,
    "epsilon_norm": 8.051383018493652,
    "prediction_norm": 6.897912502288818,
    "timestamp": "2025-07-14T12:34:44.427024"
  },
  {
    "epoch": 1960,
    "batch_idx": 64,
    "loss": 0.14883071184158325,
    "learning_rate": 9.866357858642196e-07,
    "timestep_mean": 472.375,
    "epsilon_norm": 7.279541015625,
    "prediction_norm": 6.899186134338379,
    "timestamp": "2025-07-14T12:34:44.432497"
  },
  {
    "epoch": 1960,
    "batch_idx": 96,
    "loss": 0.2796095609664917,
    "learning_rate": 9.866357858642196e-07,
    "timestep_mean": 474.875,
    "epsilon_norm": 7.632841110229492,
    "prediction_norm": 6.59514045715332,
    "timestamp": "2025-07-14T12:34:44.437942"
  },
  {
    "epoch": 1960,
    "batch_idx": 128,
    "loss": 0.43325239419937134,
    "learning_rate": 9.866357858642196e-07,
    "timestep_mean": 456.6875,
    "epsilon_norm": 8.071667671203613,
    "prediction_norm": 5.613563060760498,
    "timestamp": "2025-07-14T12:34:44.443192"
  },
  {
    "epoch": 1960,
    "batch_idx": 160,
    "loss": 0.09020605683326721,
    "learning_rate": 9.866357858642196e-07,
    "timestep_mean": 526.90625,
    "epsilon_norm": 7.884422779083252,
    "prediction_norm": 7.561915874481201,
    "timestamp": "2025-07-14T12:34:44.448375"
  },
  {
    "epoch": 1960,
    "batch_idx": 192,
    "loss": 0.2771279215812683,
    "learning_rate": 9.866357858642196e-07,
    "timestep_mean": 485.75,
    "epsilon_norm": 3.651374578475952,
    "prediction_norm": 2.8232030868530273,
    "timestamp": "2025-07-14T12:34:44.453550"
  },
  {
    "epoch": 1961,
    "batch_idx": 0,
    "loss": 0.3223888576030731,
    "learning_rate": 9.379358803627271e-07,
    "timestep_mean": 494.46875,
    "epsilon_norm": 8.880637168884277,
    "prediction_norm": 8.053348541259766,
    "timestamp": "2025-07-14T12:34:44.458596"
  },
  {
    "epoch": 1961,
    "batch_idx": 32,
    "loss": 0.4746113419532776,
    "learning_rate": 9.379358803627271e-07,
    "timestep_mean": 401.1875,
    "epsilon_norm": 9.710383415222168,
    "prediction_norm": 7.652186393737793,
    "timestamp": "2025-07-14T12:34:44.463314"
  },
  {
    "epoch": 1961,
    "batch_idx": 64,
    "loss": 0.1861293613910675,
    "learning_rate": 9.379358803627271e-07,
    "timestep_mean": 491.3125,
    "epsilon_norm": 7.958714485168457,
    "prediction_norm": 7.250037670135498,
    "timestamp": "2025-07-14T12:34:44.468121"
  },
  {
    "epoch": 1961,
    "batch_idx": 96,
    "loss": 0.22407914698123932,
    "learning_rate": 9.379358803627271e-07,
    "timestep_mean": 547.15625,
    "epsilon_norm": 7.678514003753662,
    "prediction_norm": 7.0321431159973145,
    "timestamp": "2025-07-14T12:34:44.472897"
  },
  {
    "epoch": 1961,
    "batch_idx": 128,
    "loss": 0.3903495669364929,
    "learning_rate": 9.379358803627271e-07,
    "timestep_mean": 386.90625,
    "epsilon_norm": 6.794158458709717,
    "prediction_norm": 4.730700969696045,
    "timestamp": "2025-07-14T12:34:44.477687"
  },
  {
    "epoch": 1961,
    "batch_idx": 160,
    "loss": 0.19952377676963806,
    "learning_rate": 9.379358803627271e-07,
    "timestep_mean": 496.75,
    "epsilon_norm": 7.650511264801025,
    "prediction_norm": 6.6146674156188965,
    "timestamp": "2025-07-14T12:34:44.482661"
  },
  {
    "epoch": 1961,
    "batch_idx": 192,
    "loss": 0.36088332533836365,
    "learning_rate": 9.379358803627271e-07,
    "timestep_mean": 639.875,
    "epsilon_norm": 4.145040988922119,
    "prediction_norm": 3.4752206802368164,
    "timestamp": "2025-07-14T12:34:44.487741"
  },
  {
    "epoch": 1962,
    "batch_idx": 0,
    "loss": 0.13509228825569153,
    "learning_rate": 8.904673608940974e-07,
    "timestep_mean": 481.78125,
    "epsilon_norm": 7.797480583190918,
    "prediction_norm": 7.494233131408691,
    "timestamp": "2025-07-14T12:34:44.492767"
  },
  {
    "epoch": 1962,
    "batch_idx": 32,
    "loss": 0.2565348148345947,
    "learning_rate": 8.904673608940974e-07,
    "timestep_mean": 453.46875,
    "epsilon_norm": 7.913381576538086,
    "prediction_norm": 7.198975563049316,
    "timestamp": "2025-07-14T12:34:44.497753"
  },
  {
    "epoch": 1962,
    "batch_idx": 64,
    "loss": 0.11440395563840866,
    "learning_rate": 8.904673608940974e-07,
    "timestep_mean": 535.4375,
    "epsilon_norm": 7.238481044769287,
    "prediction_norm": 6.765378952026367,
    "timestamp": "2025-07-14T12:34:44.502803"
  },
  {
    "epoch": 1962,
    "batch_idx": 96,
    "loss": 0.1933038830757141,
    "learning_rate": 8.904673608940974e-07,
    "timestep_mean": 516.0,
    "epsilon_norm": 8.024347305297852,
    "prediction_norm": 6.874923229217529,
    "timestamp": "2025-07-14T12:34:44.507727"
  },
  {
    "epoch": 1962,
    "batch_idx": 128,
    "loss": 0.21343432366847992,
    "learning_rate": 8.904673608940974e-07,
    "timestep_mean": 568.34375,
    "epsilon_norm": 7.995640754699707,
    "prediction_norm": 7.024775981903076,
    "timestamp": "2025-07-14T12:34:44.512637"
  },
  {
    "epoch": 1962,
    "batch_idx": 160,
    "loss": 0.2215503752231598,
    "learning_rate": 8.904673608940974e-07,
    "timestep_mean": 604.71875,
    "epsilon_norm": 7.625108242034912,
    "prediction_norm": 6.999690055847168,
    "timestamp": "2025-07-14T12:34:44.517594"
  },
  {
    "epoch": 1962,
    "batch_idx": 192,
    "loss": 0.5529381632804871,
    "learning_rate": 8.904673608940974e-07,
    "timestep_mean": 717.125,
    "epsilon_norm": 4.775165557861328,
    "prediction_norm": 4.051061630249023,
    "timestamp": "2025-07-14T12:34:44.522400"
  },
  {
    "epoch": 1963,
    "batch_idx": 0,
    "loss": 0.22441497445106506,
    "learning_rate": 8.442303445823077e-07,
    "timestep_mean": 441.4375,
    "epsilon_norm": 7.115808963775635,
    "prediction_norm": 6.311416149139404,
    "timestamp": "2025-07-14T12:34:44.527182"
  },
  {
    "epoch": 1963,
    "batch_idx": 32,
    "loss": 0.22521989047527313,
    "learning_rate": 8.442303445823077e-07,
    "timestep_mean": 529.34375,
    "epsilon_norm": 8.324047088623047,
    "prediction_norm": 7.705618381500244,
    "timestamp": "2025-07-14T12:34:44.531945"
  },
  {
    "epoch": 1963,
    "batch_idx": 64,
    "loss": 0.18873745203018188,
    "learning_rate": 8.442303445823077e-07,
    "timestep_mean": 447.34375,
    "epsilon_norm": 7.51481819152832,
    "prediction_norm": 6.668729782104492,
    "timestamp": "2025-07-14T12:34:44.536687"
  },
  {
    "epoch": 1963,
    "batch_idx": 96,
    "loss": 0.37653782963752747,
    "learning_rate": 8.442303445823077e-07,
    "timestep_mean": 494.25,
    "epsilon_norm": 8.708200454711914,
    "prediction_norm": 6.883727073669434,
    "timestamp": "2025-07-14T12:34:44.541462"
  },
  {
    "epoch": 1963,
    "batch_idx": 128,
    "loss": 0.4383019804954529,
    "learning_rate": 8.442303445823077e-07,
    "timestep_mean": 451.75,
    "epsilon_norm": 7.57002592086792,
    "prediction_norm": 5.805328369140625,
    "timestamp": "2025-07-14T12:34:44.546430"
  },
  {
    "epoch": 1963,
    "batch_idx": 160,
    "loss": 0.44190776348114014,
    "learning_rate": 8.442303445823077e-07,
    "timestep_mean": 487.15625,
    "epsilon_norm": 8.72851276397705,
    "prediction_norm": 6.571135520935059,
    "timestamp": "2025-07-14T12:34:44.551250"
  },
  {
    "epoch": 1963,
    "batch_idx": 192,
    "loss": 0.18177896738052368,
    "learning_rate": 8.442303445823077e-07,
    "timestep_mean": 486.125,
    "epsilon_norm": 3.9409470558166504,
    "prediction_norm": 3.476452589035034,
    "timestamp": "2025-07-14T12:34:44.557488"
  },
  {
    "epoch": 1964,
    "batch_idx": 0,
    "loss": 0.3371002972126007,
    "learning_rate": 7.992249455124881e-07,
    "timestep_mean": 491.875,
    "epsilon_norm": 7.5707502365112305,
    "prediction_norm": 6.587314605712891,
    "timestamp": "2025-07-14T12:34:44.562449"
  },
  {
    "epoch": 1964,
    "batch_idx": 32,
    "loss": 0.41210684180259705,
    "learning_rate": 7.992249455124881e-07,
    "timestep_mean": 467.09375,
    "epsilon_norm": 7.939080238342285,
    "prediction_norm": 6.838838577270508,
    "timestamp": "2025-07-14T12:34:44.567343"
  },
  {
    "epoch": 1964,
    "batch_idx": 64,
    "loss": 0.2246461808681488,
    "learning_rate": 7.992249455124881e-07,
    "timestep_mean": 575.5625,
    "epsilon_norm": 7.727367401123047,
    "prediction_norm": 6.666476249694824,
    "timestamp": "2025-07-14T12:34:44.572205"
  },
  {
    "epoch": 1964,
    "batch_idx": 96,
    "loss": 0.2650560140609741,
    "learning_rate": 7.992249455124881e-07,
    "timestep_mean": 600.21875,
    "epsilon_norm": 6.966712951660156,
    "prediction_norm": 5.5289411544799805,
    "timestamp": "2025-07-14T12:34:44.577188"
  },
  {
    "epoch": 1964,
    "batch_idx": 128,
    "loss": 0.17356973886489868,
    "learning_rate": 7.992249455124881e-07,
    "timestep_mean": 493.21875,
    "epsilon_norm": 8.674141883850098,
    "prediction_norm": 7.861104488372803,
    "timestamp": "2025-07-14T12:34:44.582098"
  },
  {
    "epoch": 1964,
    "batch_idx": 160,
    "loss": 0.28024500608444214,
    "learning_rate": 7.992249455124881e-07,
    "timestep_mean": 455.9375,
    "epsilon_norm": 8.230478286743164,
    "prediction_norm": 7.320643424987793,
    "timestamp": "2025-07-14T12:34:44.586959"
  },
  {
    "epoch": 1964,
    "batch_idx": 192,
    "loss": 0.3455151319503784,
    "learning_rate": 7.992249455124881e-07,
    "timestep_mean": 322.625,
    "epsilon_norm": 3.9056713581085205,
    "prediction_norm": 2.613152265548706,
    "timestamp": "2025-07-14T12:34:44.591907"
  },
  {
    "epoch": 1965,
    "batch_idx": 0,
    "loss": 0.1998911201953888,
    "learning_rate": 7.55451274731033e-07,
    "timestep_mean": 514.625,
    "epsilon_norm": 8.797637939453125,
    "prediction_norm": 8.316100120544434,
    "timestamp": "2025-07-14T12:34:44.596757"
  },
  {
    "epoch": 1965,
    "batch_idx": 32,
    "loss": 0.11367958784103394,
    "learning_rate": 7.55451274731033e-07,
    "timestep_mean": 538.84375,
    "epsilon_norm": 6.526326656341553,
    "prediction_norm": 5.9306721687316895,
    "timestamp": "2025-07-14T12:34:44.602393"
  },
  {
    "epoch": 1965,
    "batch_idx": 64,
    "loss": 0.23080089688301086,
    "learning_rate": 7.55451274731033e-07,
    "timestep_mean": 411.15625,
    "epsilon_norm": 8.275885581970215,
    "prediction_norm": 7.098826885223389,
    "timestamp": "2025-07-14T12:34:44.614036"
  },
  {
    "epoch": 1965,
    "batch_idx": 96,
    "loss": 0.30637994408607483,
    "learning_rate": 7.55451274731033e-07,
    "timestep_mean": 503.84375,
    "epsilon_norm": 8.021330833435059,
    "prediction_norm": 7.071908950805664,
    "timestamp": "2025-07-14T12:34:44.619469"
  },
  {
    "epoch": 1965,
    "batch_idx": 128,
    "loss": 0.16924303770065308,
    "learning_rate": 7.55451274731033e-07,
    "timestep_mean": 541.3125,
    "epsilon_norm": 8.561613082885742,
    "prediction_norm": 7.921614646911621,
    "timestamp": "2025-07-14T12:34:44.624631"
  },
  {
    "epoch": 1965,
    "batch_idx": 160,
    "loss": 0.35038307309150696,
    "learning_rate": 7.55451274731033e-07,
    "timestep_mean": 441.6875,
    "epsilon_norm": 8.89365291595459,
    "prediction_norm": 6.601502895355225,
    "timestamp": "2025-07-14T12:34:44.629812"
  },
  {
    "epoch": 1965,
    "batch_idx": 192,
    "loss": 0.37165045738220215,
    "learning_rate": 7.55451274731033e-07,
    "timestep_mean": 411.75,
    "epsilon_norm": 3.079136848449707,
    "prediction_norm": 2.5170204639434814,
    "timestamp": "2025-07-14T12:34:44.635181"
  },
  {
    "epoch": 1966,
    "batch_idx": 0,
    "loss": 0.416729599237442,
    "learning_rate": 7.129094402451012e-07,
    "timestep_mean": 590.875,
    "epsilon_norm": 7.800025463104248,
    "prediction_norm": 6.793274879455566,
    "timestamp": "2025-07-14T12:34:44.640501"
  },
  {
    "epoch": 1966,
    "batch_idx": 32,
    "loss": 0.11850526928901672,
    "learning_rate": 7.129094402451012e-07,
    "timestep_mean": 631.28125,
    "epsilon_norm": 8.338406562805176,
    "prediction_norm": 7.522503852844238,
    "timestamp": "2025-07-14T12:34:44.645632"
  },
  {
    "epoch": 1966,
    "batch_idx": 64,
    "loss": 0.23688006401062012,
    "learning_rate": 7.129094402451012e-07,
    "timestep_mean": 582.0,
    "epsilon_norm": 7.93566370010376,
    "prediction_norm": 6.945899486541748,
    "timestamp": "2025-07-14T12:34:44.650860"
  },
  {
    "epoch": 1966,
    "batch_idx": 96,
    "loss": 0.2544028162956238,
    "learning_rate": 7.129094402451012e-07,
    "timestep_mean": 511.78125,
    "epsilon_norm": 8.014091491699219,
    "prediction_norm": 6.412685871124268,
    "timestamp": "2025-07-14T12:34:44.656244"
  },
  {
    "epoch": 1966,
    "batch_idx": 128,
    "loss": 0.14604493975639343,
    "learning_rate": 7.129094402451012e-07,
    "timestep_mean": 470.15625,
    "epsilon_norm": 6.953274726867676,
    "prediction_norm": 6.155836582183838,
    "timestamp": "2025-07-14T12:34:44.661722"
  },
  {
    "epoch": 1966,
    "batch_idx": 160,
    "loss": 0.13071145117282867,
    "learning_rate": 7.129094402451012e-07,
    "timestep_mean": 590.625,
    "epsilon_norm": 8.01986312866211,
    "prediction_norm": 7.6727471351623535,
    "timestamp": "2025-07-14T12:34:44.667135"
  },
  {
    "epoch": 1966,
    "batch_idx": 192,
    "loss": 0.20286373794078827,
    "learning_rate": 7.129094402451012e-07,
    "timestep_mean": 652.0,
    "epsilon_norm": 3.811985731124878,
    "prediction_norm": 3.485069513320923,
    "timestamp": "2025-07-14T12:34:44.672487"
  },
  {
    "epoch": 1967,
    "batch_idx": 0,
    "loss": 0.3095638155937195,
    "learning_rate": 6.715995470225048e-07,
    "timestep_mean": 547.53125,
    "epsilon_norm": 8.624309539794922,
    "prediction_norm": 7.249600410461426,
    "timestamp": "2025-07-14T12:34:44.677845"
  },
  {
    "epoch": 1967,
    "batch_idx": 32,
    "loss": 0.20257848501205444,
    "learning_rate": 6.715995470225048e-07,
    "timestep_mean": 490.84375,
    "epsilon_norm": 7.27910852432251,
    "prediction_norm": 6.792276382446289,
    "timestamp": "2025-07-14T12:34:44.683259"
  },
  {
    "epoch": 1967,
    "batch_idx": 64,
    "loss": 0.18517372012138367,
    "learning_rate": 6.715995470225048e-07,
    "timestep_mean": 451.78125,
    "epsilon_norm": 7.922749042510986,
    "prediction_norm": 6.856568813323975,
    "timestamp": "2025-07-14T12:34:44.688710"
  },
  {
    "epoch": 1967,
    "batch_idx": 96,
    "loss": 0.4738411605358124,
    "learning_rate": 6.715995470225048e-07,
    "timestep_mean": 453.15625,
    "epsilon_norm": 8.51430892944336,
    "prediction_norm": 6.776823043823242,
    "timestamp": "2025-07-14T12:34:44.694002"
  },
  {
    "epoch": 1967,
    "batch_idx": 128,
    "loss": 0.19371308386325836,
    "learning_rate": 6.715995470225048e-07,
    "timestep_mean": 508.21875,
    "epsilon_norm": 8.875221252441406,
    "prediction_norm": 7.693267345428467,
    "timestamp": "2025-07-14T12:34:44.699310"
  },
  {
    "epoch": 1967,
    "batch_idx": 160,
    "loss": 0.2188943475484848,
    "learning_rate": 6.715995470225048e-07,
    "timestep_mean": 548.9375,
    "epsilon_norm": 7.6278839111328125,
    "prediction_norm": 6.664848327636719,
    "timestamp": "2025-07-14T12:34:44.704798"
  },
  {
    "epoch": 1967,
    "batch_idx": 192,
    "loss": 0.2630595862865448,
    "learning_rate": 6.715995470225048e-07,
    "timestep_mean": 351.375,
    "epsilon_norm": 2.581796884536743,
    "prediction_norm": 2.4830009937286377,
    "timestamp": "2025-07-14T12:34:44.710097"
  },
  {
    "epoch": 1968,
    "batch_idx": 0,
    "loss": 0.32091057300567627,
    "learning_rate": 6.315216969912655e-07,
    "timestep_mean": 482.5,
    "epsilon_norm": 8.278763771057129,
    "prediction_norm": 6.180835723876953,
    "timestamp": "2025-07-14T12:34:44.715274"
  },
  {
    "epoch": 1968,
    "batch_idx": 32,
    "loss": 0.4947376251220703,
    "learning_rate": 6.315216969912655e-07,
    "timestep_mean": 515.125,
    "epsilon_norm": 9.641744613647461,
    "prediction_norm": 7.416517734527588,
    "timestamp": "2025-07-14T12:34:44.720579"
  },
  {
    "epoch": 1968,
    "batch_idx": 64,
    "loss": 0.20247893035411835,
    "learning_rate": 6.315216969912655e-07,
    "timestep_mean": 571.53125,
    "epsilon_norm": 7.573853969573975,
    "prediction_norm": 6.798064708709717,
    "timestamp": "2025-07-14T12:34:44.725646"
  },
  {
    "epoch": 1968,
    "batch_idx": 96,
    "loss": 0.21111950278282166,
    "learning_rate": 6.315216969912655e-07,
    "timestep_mean": 501.375,
    "epsilon_norm": 8.19756031036377,
    "prediction_norm": 7.014803409576416,
    "timestamp": "2025-07-14T12:34:44.730819"
  },
  {
    "epoch": 1968,
    "batch_idx": 128,
    "loss": 0.3172382414340973,
    "learning_rate": 6.315216969912655e-07,
    "timestep_mean": 423.3125,
    "epsilon_norm": 6.563408851623535,
    "prediction_norm": 5.611907005310059,
    "timestamp": "2025-07-14T12:34:44.735802"
  },
  {
    "epoch": 1968,
    "batch_idx": 160,
    "loss": 0.09558569639921188,
    "learning_rate": 6.315216969912655e-07,
    "timestep_mean": 579.4375,
    "epsilon_norm": 8.824065208435059,
    "prediction_norm": 8.591254234313965,
    "timestamp": "2025-07-14T12:34:44.740732"
  },
  {
    "epoch": 1968,
    "batch_idx": 192,
    "loss": 0.4362088441848755,
    "learning_rate": 6.315216969912655e-07,
    "timestep_mean": 391.875,
    "epsilon_norm": 4.199055194854736,
    "prediction_norm": 3.1617074012756348,
    "timestamp": "2025-07-14T12:34:44.745669"
  },
  {
    "epoch": 1969,
    "batch_idx": 0,
    "loss": 0.47921067476272583,
    "learning_rate": 5.926759890394478e-07,
    "timestep_mean": 486.34375,
    "epsilon_norm": 7.900770664215088,
    "prediction_norm": 5.6540117263793945,
    "timestamp": "2025-07-14T12:34:44.750669"
  },
  {
    "epoch": 1969,
    "batch_idx": 32,
    "loss": 0.4015812277793884,
    "learning_rate": 5.926759890394478e-07,
    "timestep_mean": 442.46875,
    "epsilon_norm": 8.254230499267578,
    "prediction_norm": 6.080819129943848,
    "timestamp": "2025-07-14T12:34:44.755774"
  },
  {
    "epoch": 1969,
    "batch_idx": 64,
    "loss": 0.38646233081817627,
    "learning_rate": 5.926759890394478e-07,
    "timestep_mean": 485.59375,
    "epsilon_norm": 8.711126327514648,
    "prediction_norm": 7.073509216308594,
    "timestamp": "2025-07-14T12:34:44.760910"
  },
  {
    "epoch": 1969,
    "batch_idx": 96,
    "loss": 0.11501093208789825,
    "learning_rate": 5.926759890394478e-07,
    "timestep_mean": 495.65625,
    "epsilon_norm": 7.241095542907715,
    "prediction_norm": 6.455469131469727,
    "timestamp": "2025-07-14T12:34:44.766000"
  },
  {
    "epoch": 1969,
    "batch_idx": 128,
    "loss": 0.28539711236953735,
    "learning_rate": 5.926759890394478e-07,
    "timestep_mean": 447.125,
    "epsilon_norm": 7.4116902351379395,
    "prediction_norm": 6.237458229064941,
    "timestamp": "2025-07-14T12:34:44.771151"
  },
  {
    "epoch": 1969,
    "batch_idx": 160,
    "loss": 0.1957760751247406,
    "learning_rate": 5.926759890394478e-07,
    "timestep_mean": 442.46875,
    "epsilon_norm": 8.209745407104492,
    "prediction_norm": 7.071519374847412,
    "timestamp": "2025-07-14T12:34:44.776318"
  },
  {
    "epoch": 1969,
    "batch_idx": 192,
    "loss": 0.32909396290779114,
    "learning_rate": 5.926759890394478e-07,
    "timestep_mean": 523.5,
    "epsilon_norm": 4.470299243927002,
    "prediction_norm": 3.3416898250579834,
    "timestamp": "2025-07-14T12:34:44.781423"
  },
  {
    "epoch": 1970,
    "batch_idx": 0,
    "loss": 0.31883013248443604,
    "learning_rate": 5.550625190150476e-07,
    "timestep_mean": 505.59375,
    "epsilon_norm": 8.544184684753418,
    "prediction_norm": 6.457281589508057,
    "timestamp": "2025-07-14T12:34:44.786680"
  },
  {
    "epoch": 1970,
    "batch_idx": 32,
    "loss": 0.11666461825370789,
    "learning_rate": 5.550625190150476e-07,
    "timestep_mean": 431.96875,
    "epsilon_norm": 7.564525604248047,
    "prediction_norm": 7.034733772277832,
    "timestamp": "2025-07-14T12:34:44.791921"
  },
  {
    "epoch": 1970,
    "batch_idx": 64,
    "loss": 0.29018956422805786,
    "learning_rate": 5.550625190150476e-07,
    "timestep_mean": 513.375,
    "epsilon_norm": 8.205697059631348,
    "prediction_norm": 7.300631999969482,
    "timestamp": "2025-07-14T12:34:44.797099"
  },
  {
    "epoch": 1970,
    "batch_idx": 96,
    "loss": 0.22544212639331818,
    "learning_rate": 5.550625190150476e-07,
    "timestep_mean": 581.46875,
    "epsilon_norm": 8.665019035339355,
    "prediction_norm": 7.37961483001709,
    "timestamp": "2025-07-14T12:34:44.802234"
  },
  {
    "epoch": 1970,
    "batch_idx": 128,
    "loss": 0.4566758871078491,
    "learning_rate": 5.550625190150476e-07,
    "timestep_mean": 361.34375,
    "epsilon_norm": 7.541846752166748,
    "prediction_norm": 5.846305847167969,
    "timestamp": "2025-07-14T12:34:44.807408"
  },
  {
    "epoch": 1970,
    "batch_idx": 160,
    "loss": 0.26440683007240295,
    "learning_rate": 5.550625190150476e-07,
    "timestep_mean": 532.34375,
    "epsilon_norm": 8.58226490020752,
    "prediction_norm": 7.3252787590026855,
    "timestamp": "2025-07-14T12:34:44.812676"
  },
  {
    "epoch": 1970,
    "batch_idx": 192,
    "loss": 0.0959198921918869,
    "learning_rate": 5.550625190150476e-07,
    "timestep_mean": 417.5,
    "epsilon_norm": 3.2531280517578125,
    "prediction_norm": 3.1369571685791016,
    "timestamp": "2025-07-14T12:34:44.817880"
  },
  {
    "epoch": 1971,
    "batch_idx": 0,
    "loss": 0.2891301214694977,
    "learning_rate": 5.186813797254381e-07,
    "timestep_mean": 554.28125,
    "epsilon_norm": 9.143799781799316,
    "prediction_norm": 7.836328029632568,
    "timestamp": "2025-07-14T12:34:44.822976"
  },
  {
    "epoch": 1971,
    "batch_idx": 32,
    "loss": 0.1872372329235077,
    "learning_rate": 5.186813797254381e-07,
    "timestep_mean": 527.25,
    "epsilon_norm": 7.508035659790039,
    "prediction_norm": 6.431091785430908,
    "timestamp": "2025-07-14T12:34:44.828277"
  },
  {
    "epoch": 1971,
    "batch_idx": 64,
    "loss": 0.10588166862726212,
    "learning_rate": 5.186813797254381e-07,
    "timestep_mean": 595.6875,
    "epsilon_norm": 7.184289932250977,
    "prediction_norm": 6.876782417297363,
    "timestamp": "2025-07-14T12:34:44.833568"
  },
  {
    "epoch": 1971,
    "batch_idx": 96,
    "loss": 0.2543449401855469,
    "learning_rate": 5.186813797254381e-07,
    "timestep_mean": 495.5625,
    "epsilon_norm": 8.587632179260254,
    "prediction_norm": 7.277245044708252,
    "timestamp": "2025-07-14T12:34:44.838561"
  },
  {
    "epoch": 1971,
    "batch_idx": 128,
    "loss": 0.26747098565101624,
    "learning_rate": 5.186813797254381e-07,
    "timestep_mean": 499.625,
    "epsilon_norm": 6.775057792663574,
    "prediction_norm": 5.646205425262451,
    "timestamp": "2025-07-14T12:34:44.843768"
  },
  {
    "epoch": 1971,
    "batch_idx": 160,
    "loss": 0.2595219016075134,
    "learning_rate": 5.186813797254381e-07,
    "timestep_mean": 490.09375,
    "epsilon_norm": 7.577627658843994,
    "prediction_norm": 6.554196357727051,
    "timestamp": "2025-07-14T12:34:44.848937"
  },
  {
    "epoch": 1971,
    "batch_idx": 192,
    "loss": 0.05860793590545654,
    "learning_rate": 5.186813797254381e-07,
    "timestep_mean": 613.75,
    "epsilon_norm": 3.3466570377349854,
    "prediction_norm": 3.0979182720184326,
    "timestamp": "2025-07-14T12:34:44.854159"
  },
  {
    "epoch": 1972,
    "batch_idx": 0,
    "loss": 0.2600661516189575,
    "learning_rate": 4.835326609376462e-07,
    "timestep_mean": 524.0625,
    "epsilon_norm": 8.692440032958984,
    "prediction_norm": 7.404092788696289,
    "timestamp": "2025-07-14T12:34:44.859434"
  },
  {
    "epoch": 1972,
    "batch_idx": 32,
    "loss": 0.27142414450645447,
    "learning_rate": 4.835326609376462e-07,
    "timestep_mean": 458.71875,
    "epsilon_norm": 7.452977180480957,
    "prediction_norm": 6.479616165161133,
    "timestamp": "2025-07-14T12:34:44.864565"
  },
  {
    "epoch": 1972,
    "batch_idx": 64,
    "loss": 0.2644779086112976,
    "learning_rate": 4.835326609376462e-07,
    "timestep_mean": 569.375,
    "epsilon_norm": 8.314547538757324,
    "prediction_norm": 7.030329704284668,
    "timestamp": "2025-07-14T12:34:44.869732"
  },
  {
    "epoch": 1972,
    "batch_idx": 96,
    "loss": 0.33536583185195923,
    "learning_rate": 4.835326609376462e-07,
    "timestep_mean": 424.34375,
    "epsilon_norm": 7.283456802368164,
    "prediction_norm": 6.135376930236816,
    "timestamp": "2025-07-14T12:34:44.874852"
  },
  {
    "epoch": 1972,
    "batch_idx": 128,
    "loss": 0.18897540867328644,
    "learning_rate": 4.835326609376462e-07,
    "timestep_mean": 602.09375,
    "epsilon_norm": 9.40078067779541,
    "prediction_norm": 8.5212984085083,
    "timestamp": "2025-07-14T12:34:44.880121"
  },
  {
    "epoch": 1972,
    "batch_idx": 160,
    "loss": 0.22975721955299377,
    "learning_rate": 4.835326609376462e-07,
    "timestep_mean": 478.1875,
    "epsilon_norm": 6.869845390319824,
    "prediction_norm": 5.83657169342041,
    "timestamp": "2025-07-14T12:34:44.885247"
  },
  {
    "epoch": 1972,
    "batch_idx": 192,
    "loss": 0.13978375494480133,
    "learning_rate": 4.835326609376462e-07,
    "timestep_mean": 566.5,
    "epsilon_norm": 5.281387805938721,
    "prediction_norm": 4.943111419677734,
    "timestamp": "2025-07-14T12:34:44.890387"
  },
  {
    "epoch": 1973,
    "batch_idx": 0,
    "loss": 0.31788140535354614,
    "learning_rate": 4.4961644937752074e-07,
    "timestep_mean": 461.15625,
    "epsilon_norm": 6.9991455078125,
    "prediction_norm": 5.996115207672119,
    "timestamp": "2025-07-14T12:34:44.895618"
  },
  {
    "epoch": 1973,
    "batch_idx": 32,
    "loss": 0.27237874269485474,
    "learning_rate": 4.4961644937752074e-07,
    "timestep_mean": 484.28125,
    "epsilon_norm": 8.599380493164062,
    "prediction_norm": 7.438248157501221,
    "timestamp": "2025-07-14T12:34:44.900875"
  },
  {
    "epoch": 1973,
    "batch_idx": 64,
    "loss": 0.3882778286933899,
    "learning_rate": 4.4961644937752074e-07,
    "timestep_mean": 456.6875,
    "epsilon_norm": 8.490903854370117,
    "prediction_norm": 7.236042499542236,
    "timestamp": "2025-07-14T12:34:44.906061"
  },
  {
    "epoch": 1973,
    "batch_idx": 96,
    "loss": 0.2979061007499695,
    "learning_rate": 4.4961644937752074e-07,
    "timestep_mean": 570.8125,
    "epsilon_norm": 8.984509468078613,
    "prediction_norm": 7.553010940551758,
    "timestamp": "2025-07-14T12:34:44.911282"
  },
  {
    "epoch": 1973,
    "batch_idx": 128,
    "loss": 0.3585243225097656,
    "learning_rate": 4.4961644937752074e-07,
    "timestep_mean": 463.625,
    "epsilon_norm": 7.75709342956543,
    "prediction_norm": 6.790096282958984,
    "timestamp": "2025-07-14T12:34:44.916490"
  },
  {
    "epoch": 1973,
    "batch_idx": 160,
    "loss": 0.22886386513710022,
    "learning_rate": 4.4961644937752074e-07,
    "timestep_mean": 527.59375,
    "epsilon_norm": 7.787026405334473,
    "prediction_norm": 6.280324935913086,
    "timestamp": "2025-07-14T12:34:44.921845"
  },
  {
    "epoch": 1973,
    "batch_idx": 192,
    "loss": 0.19990476965904236,
    "learning_rate": 4.4961644937752074e-07,
    "timestep_mean": 641.25,
    "epsilon_norm": 3.9280476570129395,
    "prediction_norm": 3.5226805210113525,
    "timestamp": "2025-07-14T12:34:44.926972"
  },
  {
    "epoch": 1974,
    "batch_idx": 0,
    "loss": 0.24042537808418274,
    "learning_rate": 4.169328287299541e-07,
    "timestep_mean": 519.90625,
    "epsilon_norm": 8.026875495910645,
    "prediction_norm": 7.119966983795166,
    "timestamp": "2025-07-14T12:34:44.932158"
  },
  {
    "epoch": 1974,
    "batch_idx": 32,
    "loss": 0.1623794436454773,
    "learning_rate": 4.169328287299541e-07,
    "timestep_mean": 542.8125,
    "epsilon_norm": 7.703945636749268,
    "prediction_norm": 6.8579325675964355,
    "timestamp": "2025-07-14T12:34:44.937393"
  },
  {
    "epoch": 1974,
    "batch_idx": 64,
    "loss": 0.18829578161239624,
    "learning_rate": 4.169328287299541e-07,
    "timestep_mean": 641.25,
    "epsilon_norm": 7.494311332702637,
    "prediction_norm": 6.748343467712402,
    "timestamp": "2025-07-14T12:34:44.942819"
  },
  {
    "epoch": 1974,
    "batch_idx": 96,
    "loss": 0.23447322845458984,
    "learning_rate": 4.169328287299541e-07,
    "timestep_mean": 521.375,
    "epsilon_norm": 7.868964195251465,
    "prediction_norm": 7.267219543457031,
    "timestamp": "2025-07-14T12:34:44.947950"
  },
  {
    "epoch": 1974,
    "batch_idx": 128,
    "loss": 0.2880434989929199,
    "learning_rate": 4.169328287299541e-07,
    "timestep_mean": 481.40625,
    "epsilon_norm": 7.9040703773498535,
    "prediction_norm": 6.972902297973633,
    "timestamp": "2025-07-14T12:34:44.953175"
  },
  {
    "epoch": 1974,
    "batch_idx": 160,
    "loss": 0.21388623118400574,
    "learning_rate": 4.169328287299541e-07,
    "timestep_mean": 490.40625,
    "epsilon_norm": 7.760453701019287,
    "prediction_norm": 6.382420063018799,
    "timestamp": "2025-07-14T12:34:44.958360"
  },
  {
    "epoch": 1974,
    "batch_idx": 192,
    "loss": 0.07946091890335083,
    "learning_rate": 4.169328287299541e-07,
    "timestep_mean": 545.875,
    "epsilon_norm": 3.218200206756592,
    "prediction_norm": 3.755671501159668,
    "timestamp": "2025-07-14T12:34:44.963541"
  },
  {
    "epoch": 1975,
    "batch_idx": 0,
    "loss": 0.3070555031299591,
    "learning_rate": 3.854818796385491e-07,
    "timestep_mean": 440.5,
    "epsilon_norm": 7.893798828125,
    "prediction_norm": 6.412811279296875,
    "timestamp": "2025-07-14T12:34:44.968664"
  },
  {
    "epoch": 1975,
    "batch_idx": 32,
    "loss": 0.18743129074573517,
    "learning_rate": 3.854818796385491e-07,
    "timestep_mean": 548.71875,
    "epsilon_norm": 8.001594543457031,
    "prediction_norm": 6.456892967224121,
    "timestamp": "2025-07-14T12:34:44.973826"
  },
  {
    "epoch": 1975,
    "batch_idx": 64,
    "loss": 0.23085835576057434,
    "learning_rate": 3.854818796385491e-07,
    "timestep_mean": 480.53125,
    "epsilon_norm": 8.696663856506348,
    "prediction_norm": 7.998325347900391,
    "timestamp": "2025-07-14T12:34:44.979084"
  },
  {
    "epoch": 1975,
    "batch_idx": 96,
    "loss": 0.2524884343147278,
    "learning_rate": 3.854818796385491e-07,
    "timestep_mean": 509.84375,
    "epsilon_norm": 8.3718900680542,
    "prediction_norm": 6.807347297668457,
    "timestamp": "2025-07-14T12:34:44.984380"
  },
  {
    "epoch": 1975,
    "batch_idx": 128,
    "loss": 0.21409466862678528,
    "learning_rate": 3.854818796385491e-07,
    "timestep_mean": 486.875,
    "epsilon_norm": 8.343727111816406,
    "prediction_norm": 7.302717208862305,
    "timestamp": "2025-07-14T12:34:44.989552"
  },
  {
    "epoch": 1975,
    "batch_idx": 160,
    "loss": 0.24758681654930115,
    "learning_rate": 3.854818796385491e-07,
    "timestep_mean": 535.71875,
    "epsilon_norm": 8.583587646484375,
    "prediction_norm": 7.197963237762451,
    "timestamp": "2025-07-14T12:34:44.994674"
  },
  {
    "epoch": 1975,
    "batch_idx": 192,
    "loss": 0.35336610674858093,
    "learning_rate": 3.854818796385491e-07,
    "timestep_mean": 409.5,
    "epsilon_norm": 4.6328816413879395,
    "prediction_norm": 4.147469520568848,
    "timestamp": "2025-07-14T12:34:44.999722"
  },
  {
    "epoch": 1976,
    "batch_idx": 0,
    "loss": 0.26376309990882874,
    "learning_rate": 3.5526367970539726e-07,
    "timestep_mean": 440.96875,
    "epsilon_norm": 7.348486423492432,
    "prediction_norm": 6.154544830322266,
    "timestamp": "2025-07-14T12:34:45.005054"
  },
  {
    "epoch": 1976,
    "batch_idx": 32,
    "loss": 0.22687239944934845,
    "learning_rate": 3.5526367970539726e-07,
    "timestep_mean": 542.3125,
    "epsilon_norm": 8.682991027832031,
    "prediction_norm": 7.655359268188477,
    "timestamp": "2025-07-14T12:34:45.010179"
  },
  {
    "epoch": 1976,
    "batch_idx": 64,
    "loss": 0.307498574256897,
    "learning_rate": 3.5526367970539726e-07,
    "timestep_mean": 497.9375,
    "epsilon_norm": 7.256778717041016,
    "prediction_norm": 6.44264554977417,
    "timestamp": "2025-07-14T12:34:45.015297"
  },
  {
    "epoch": 1976,
    "batch_idx": 96,
    "loss": 0.23820945620536804,
    "learning_rate": 3.5526367970539726e-07,
    "timestep_mean": 484.34375,
    "epsilon_norm": 7.512821197509766,
    "prediction_norm": 6.2607245445251465,
    "timestamp": "2025-07-14T12:34:45.020563"
  },
  {
    "epoch": 1976,
    "batch_idx": 128,
    "loss": 0.22325551509857178,
    "learning_rate": 3.5526367970539726e-07,
    "timestep_mean": 550.53125,
    "epsilon_norm": 8.222439765930176,
    "prediction_norm": 6.811587810516357,
    "timestamp": "2025-07-14T12:34:45.025837"
  },
  {
    "epoch": 1976,
    "batch_idx": 160,
    "loss": 0.24786873161792755,
    "learning_rate": 3.5526367970539726e-07,
    "timestep_mean": 464.375,
    "epsilon_norm": 8.995453834533691,
    "prediction_norm": 7.596970558166504,
    "timestamp": "2025-07-14T12:34:45.030994"
  },
  {
    "epoch": 1976,
    "batch_idx": 192,
    "loss": 0.17906352877616882,
    "learning_rate": 3.5526367970539726e-07,
    "timestep_mean": 605.75,
    "epsilon_norm": 4.0963521003723145,
    "prediction_norm": 3.724107027053833,
    "timestamp": "2025-07-14T12:34:45.036157"
  },
  {
    "epoch": 1977,
    "batch_idx": 0,
    "loss": 0.2464328110218048,
    "learning_rate": 3.262783034908564e-07,
    "timestep_mean": 581.4375,
    "epsilon_norm": 8.83813762664795,
    "prediction_norm": 7.857411861419678,
    "timestamp": "2025-07-14T12:34:45.041373"
  },
  {
    "epoch": 1977,
    "batch_idx": 32,
    "loss": 0.3168014883995056,
    "learning_rate": 3.262783034908564e-07,
    "timestep_mean": 528.84375,
    "epsilon_norm": 8.275979042053223,
    "prediction_norm": 7.343487739562988,
    "timestamp": "2025-07-14T12:34:45.046725"
  },
  {
    "epoch": 1977,
    "batch_idx": 64,
    "loss": 0.40050607919692993,
    "learning_rate": 3.262783034908564e-07,
    "timestep_mean": 459.03125,
    "epsilon_norm": 9.368485450744629,
    "prediction_norm": 7.658954620361328,
    "timestamp": "2025-07-14T12:34:45.051921"
  },
  {
    "epoch": 1977,
    "batch_idx": 96,
    "loss": 0.23146244883537292,
    "learning_rate": 3.262783034908564e-07,
    "timestep_mean": 509.0625,
    "epsilon_norm": 7.037581443786621,
    "prediction_norm": 6.427664279937744,
    "timestamp": "2025-07-14T12:34:45.057053"
  },
  {
    "epoch": 1977,
    "batch_idx": 128,
    "loss": 0.22526392340660095,
    "learning_rate": 3.262783034908564e-07,
    "timestep_mean": 525.8125,
    "epsilon_norm": 7.644104957580566,
    "prediction_norm": 6.890722274780273,
    "timestamp": "2025-07-14T12:34:45.062206"
  },
  {
    "epoch": 1977,
    "batch_idx": 160,
    "loss": 0.17796483635902405,
    "learning_rate": 3.262783034908564e-07,
    "timestep_mean": 597.21875,
    "epsilon_norm": 9.309478759765625,
    "prediction_norm": 8.730376243591309,
    "timestamp": "2025-07-14T12:34:45.067540"
  },
  {
    "epoch": 1977,
    "batch_idx": 192,
    "loss": 0.27047470211982727,
    "learning_rate": 3.262783034908564e-07,
    "timestep_mean": 561.25,
    "epsilon_norm": 4.534088611602783,
    "prediction_norm": 3.8510804176330566,
    "timestamp": "2025-07-14T12:34:45.072636"
  },
  {
    "epoch": 1978,
    "batch_idx": 0,
    "loss": 0.3460220694541931,
    "learning_rate": 2.985258225135509e-07,
    "timestep_mean": 544.71875,
    "epsilon_norm": 8.480791091918945,
    "prediction_norm": 7.176337242126465,
    "timestamp": "2025-07-14T12:34:45.077838"
  },
  {
    "epoch": 1978,
    "batch_idx": 32,
    "loss": 0.2985912263393402,
    "learning_rate": 2.985258225135509e-07,
    "timestep_mean": 426.375,
    "epsilon_norm": 8.538310050964355,
    "prediction_norm": 6.7977070808410645,
    "timestamp": "2025-07-14T12:34:45.082763"
  },
  {
    "epoch": 1978,
    "batch_idx": 64,
    "loss": 0.21581044793128967,
    "learning_rate": 2.985258225135509e-07,
    "timestep_mean": 575.90625,
    "epsilon_norm": 8.370981216430664,
    "prediction_norm": 7.320396423339844,
    "timestamp": "2025-07-14T12:34:45.087650"
  },
  {
    "epoch": 1978,
    "batch_idx": 96,
    "loss": 0.30976545810699463,
    "learning_rate": 2.985258225135509e-07,
    "timestep_mean": 503.125,
    "epsilon_norm": 7.63110876083374,
    "prediction_norm": 6.4903788566589355,
    "timestamp": "2025-07-14T12:34:45.092733"
  },
  {
    "epoch": 1978,
    "batch_idx": 128,
    "loss": 0.5238752961158752,
    "learning_rate": 2.985258225135509e-07,
    "timestep_mean": 540.65625,
    "epsilon_norm": 8.09204387664795,
    "prediction_norm": 6.074638366699219,
    "timestamp": "2025-07-14T12:34:45.097674"
  },
  {
    "epoch": 1978,
    "batch_idx": 160,
    "loss": 0.3730786442756653,
    "learning_rate": 2.985258225135509e-07,
    "timestep_mean": 461.0625,
    "epsilon_norm": 8.265247344970703,
    "prediction_norm": 7.097507953643799,
    "timestamp": "2025-07-14T12:34:45.102535"
  },
  {
    "epoch": 1978,
    "batch_idx": 192,
    "loss": 0.37991195917129517,
    "learning_rate": 2.985258225135509e-07,
    "timestep_mean": 409.5,
    "epsilon_norm": 3.428895950317383,
    "prediction_norm": 2.231505870819092,
    "timestamp": "2025-07-14T12:34:45.107379"
  },
  {
    "epoch": 1979,
    "batch_idx": 0,
    "loss": 0.170973002910614,
    "learning_rate": 2.7200630524992745e-07,
    "timestep_mean": 521.4375,
    "epsilon_norm": 8.769270896911621,
    "prediction_norm": 8.132695198059082,
    "timestamp": "2025-07-14T12:34:45.112279"
  },
  {
    "epoch": 1979,
    "batch_idx": 32,
    "loss": 0.2246909737586975,
    "learning_rate": 2.7200630524992745e-07,
    "timestep_mean": 538.4375,
    "epsilon_norm": 7.725288391113281,
    "prediction_norm": 6.644896984100342,
    "timestamp": "2025-07-14T12:34:45.117260"
  },
  {
    "epoch": 1979,
    "batch_idx": 64,
    "loss": 0.11703809350728989,
    "learning_rate": 2.7200630524992745e-07,
    "timestep_mean": 555.84375,
    "epsilon_norm": 7.275386810302734,
    "prediction_norm": 6.634763240814209,
    "timestamp": "2025-07-14T12:34:45.122137"
  },
  {
    "epoch": 1979,
    "batch_idx": 96,
    "loss": 0.35732316970825195,
    "learning_rate": 2.7200630524992745e-07,
    "timestep_mean": 494.96875,
    "epsilon_norm": 8.532506942749023,
    "prediction_norm": 7.41096305847168,
    "timestamp": "2025-07-14T12:34:45.127022"
  },
  {
    "epoch": 1979,
    "batch_idx": 128,
    "loss": 0.18592830002307892,
    "learning_rate": 2.7200630524992745e-07,
    "timestep_mean": 482.0625,
    "epsilon_norm": 8.729327201843262,
    "prediction_norm": 7.140318393707275,
    "timestamp": "2025-07-14T12:34:45.131876"
  },
  {
    "epoch": 1979,
    "batch_idx": 160,
    "loss": 0.2541287839412689,
    "learning_rate": 2.7200630524992745e-07,
    "timestep_mean": 427.1875,
    "epsilon_norm": 7.608599662780762,
    "prediction_norm": 6.343356609344482,
    "timestamp": "2025-07-14T12:34:45.136811"
  },
  {
    "epoch": 1979,
    "batch_idx": 192,
    "loss": 0.07171684503555298,
    "learning_rate": 2.7200630524992745e-07,
    "timestep_mean": 684.625,
    "epsilon_norm": 3.246137857437134,
    "prediction_norm": 3.241976499557495,
    "timestamp": "2025-07-14T12:34:45.141659"
  },
  {
    "epoch": 1980,
    "batch_idx": 0,
    "loss": 0.42634081840515137,
    "learning_rate": 2.467198171341997e-07,
    "timestep_mean": 558.75,
    "epsilon_norm": 8.766483306884766,
    "prediction_norm": 6.40263557434082,
    "timestamp": "2025-07-14T12:34:45.146664"
  },
  {
    "epoch": 1980,
    "batch_idx": 32,
    "loss": 0.20290037989616394,
    "learning_rate": 2.467198171341997e-07,
    "timestep_mean": 496.59375,
    "epsilon_norm": 6.553798198699951,
    "prediction_norm": 5.611343860626221,
    "timestamp": "2025-07-14T12:34:45.151507"
  },
  {
    "epoch": 1980,
    "batch_idx": 64,
    "loss": 0.12492910772562027,
    "learning_rate": 2.467198171341997e-07,
    "timestep_mean": 546.46875,
    "epsilon_norm": 6.557557106018066,
    "prediction_norm": 6.263979434967041,
    "timestamp": "2025-07-14T12:34:45.156503"
  },
  {
    "epoch": 1980,
    "batch_idx": 96,
    "loss": 0.37344473600387573,
    "learning_rate": 2.467198171341997e-07,
    "timestep_mean": 455.84375,
    "epsilon_norm": 8.892641067504883,
    "prediction_norm": 7.616144180297852,
    "timestamp": "2025-07-14T12:34:45.161518"
  },
  {
    "epoch": 1980,
    "batch_idx": 128,
    "loss": 0.2606395483016968,
    "learning_rate": 2.467198171341997e-07,
    "timestep_mean": 519.8125,
    "epsilon_norm": 6.956752777099609,
    "prediction_norm": 6.2145609855651855,
    "timestamp": "2025-07-14T12:34:45.166544"
  },
  {
    "epoch": 1980,
    "batch_idx": 160,
    "loss": 0.4152217507362366,
    "learning_rate": 2.467198171341997e-07,
    "timestep_mean": 463.15625,
    "epsilon_norm": 8.765035629272461,
    "prediction_norm": 6.641943454742432,
    "timestamp": "2025-07-14T12:34:45.171738"
  },
  {
    "epoch": 1980,
    "batch_idx": 192,
    "loss": 0.25625336170196533,
    "learning_rate": 2.467198171341997e-07,
    "timestep_mean": 468.25,
    "epsilon_norm": 4.799482822418213,
    "prediction_norm": 4.196409702301025,
    "timestamp": "2025-07-14T12:34:45.176820"
  },
  {
    "epoch": 1981,
    "batch_idx": 0,
    "loss": 0.2922101318836212,
    "learning_rate": 2.2266642055840367e-07,
    "timestep_mean": 404.28125,
    "epsilon_norm": 7.39887809753418,
    "prediction_norm": 6.07985782623291,
    "timestamp": "2025-07-14T12:34:45.181890"
  },
  {
    "epoch": 1981,
    "batch_idx": 32,
    "loss": 0.49606209993362427,
    "learning_rate": 2.2266642055840367e-07,
    "timestep_mean": 441.90625,
    "epsilon_norm": 8.066160202026367,
    "prediction_norm": 6.065613269805908,
    "timestamp": "2025-07-14T12:34:45.187052"
  },
  {
    "epoch": 1981,
    "batch_idx": 64,
    "loss": 0.29446277022361755,
    "learning_rate": 2.2266642055840367e-07,
    "timestep_mean": 483.5,
    "epsilon_norm": 7.761168479919434,
    "prediction_norm": 6.201569080352783,
    "timestamp": "2025-07-14T12:34:45.192330"
  },
  {
    "epoch": 1981,
    "batch_idx": 96,
    "loss": 0.14793919026851654,
    "learning_rate": 2.2266642055840367e-07,
    "timestep_mean": 497.875,
    "epsilon_norm": 7.987395286560059,
    "prediction_norm": 6.780283451080322,
    "timestamp": "2025-07-14T12:34:45.197419"
  },
  {
    "epoch": 1981,
    "batch_idx": 128,
    "loss": 0.363796204328537,
    "learning_rate": 2.2266642055840367e-07,
    "timestep_mean": 457.0625,
    "epsilon_norm": 8.889440536499023,
    "prediction_norm": 7.1531476974487305,
    "timestamp": "2025-07-14T12:34:45.202565"
  },
  {
    "epoch": 1981,
    "batch_idx": 160,
    "loss": 0.21953490376472473,
    "learning_rate": 2.2266642055840367e-07,
    "timestep_mean": 549.96875,
    "epsilon_norm": 9.474082946777344,
    "prediction_norm": 7.972390174865723,
    "timestamp": "2025-07-14T12:34:45.207721"
  },
  {
    "epoch": 1981,
    "batch_idx": 192,
    "loss": 0.39224374294281006,
    "learning_rate": 2.2266642055840367e-07,
    "timestep_mean": 494.5,
    "epsilon_norm": 3.7468016147613525,
    "prediction_norm": 3.008225202560425,
    "timestamp": "2025-07-14T12:34:45.212930"
  },
  {
    "epoch": 1982,
    "batch_idx": 0,
    "loss": 0.19082850217819214,
    "learning_rate": 1.998461748717315e-07,
    "timestep_mean": 443.9375,
    "epsilon_norm": 6.960975170135498,
    "prediction_norm": 6.397202491760254,
    "timestamp": "2025-07-14T12:34:45.218109"
  },
  {
    "epoch": 1982,
    "batch_idx": 32,
    "loss": 0.5099515914916992,
    "learning_rate": 1.998461748717315e-07,
    "timestep_mean": 419.375,
    "epsilon_norm": 7.908893585205078,
    "prediction_norm": 5.550872802734375,
    "timestamp": "2025-07-14T12:34:45.223253"
  },
  {
    "epoch": 1982,
    "batch_idx": 64,
    "loss": 0.18309715390205383,
    "learning_rate": 1.998461748717315e-07,
    "timestep_mean": 508.53125,
    "epsilon_norm": 7.284280300140381,
    "prediction_norm": 6.468391418457031,
    "timestamp": "2025-07-14T12:34:45.228536"
  },
  {
    "epoch": 1982,
    "batch_idx": 96,
    "loss": 0.21906907856464386,
    "learning_rate": 1.998461748717315e-07,
    "timestep_mean": 500.03125,
    "epsilon_norm": 7.257194519042969,
    "prediction_norm": 6.553398132324219,
    "timestamp": "2025-07-14T12:34:45.233683"
  },
  {
    "epoch": 1982,
    "batch_idx": 128,
    "loss": 0.32119062542915344,
    "learning_rate": 1.998461748717315e-07,
    "timestep_mean": 463.4375,
    "epsilon_norm": 6.844542980194092,
    "prediction_norm": 5.888505935668945,
    "timestamp": "2025-07-14T12:34:45.238805"
  },
  {
    "epoch": 1982,
    "batch_idx": 160,
    "loss": 0.3170006275177002,
    "learning_rate": 1.998461748717315e-07,
    "timestep_mean": 440.8125,
    "epsilon_norm": 8.056681632995605,
    "prediction_norm": 6.602311134338379,
    "timestamp": "2025-07-14T12:34:45.243960"
  },
  {
    "epoch": 1982,
    "batch_idx": 192,
    "loss": 1.5000861883163452,
    "learning_rate": 1.998461748717315e-07,
    "timestep_mean": 387.125,
    "epsilon_norm": 5.221567630767822,
    "prediction_norm": 2.706672191619873,
    "timestamp": "2025-07-14T12:34:45.249082"
  },
  {
    "epoch": 1983,
    "batch_idx": 0,
    "loss": 0.16197869181632996,
    "learning_rate": 1.7825913638097583e-07,
    "timestep_mean": 499.625,
    "epsilon_norm": 8.167936325073242,
    "prediction_norm": 7.297781944274902,
    "timestamp": "2025-07-14T12:34:45.254264"
  },
  {
    "epoch": 1983,
    "batch_idx": 32,
    "loss": 0.331137478351593,
    "learning_rate": 1.7825913638097583e-07,
    "timestep_mean": 417.53125,
    "epsilon_norm": 7.592419624328613,
    "prediction_norm": 5.73801326751709,
    "timestamp": "2025-07-14T12:34:45.259474"
  },
  {
    "epoch": 1983,
    "batch_idx": 64,
    "loss": 0.34576910734176636,
    "learning_rate": 1.7825913638097583e-07,
    "timestep_mean": 504.5625,
    "epsilon_norm": 7.777278423309326,
    "prediction_norm": 6.612274646759033,
    "timestamp": "2025-07-14T12:34:45.264613"
  },
  {
    "epoch": 1983,
    "batch_idx": 96,
    "loss": 0.2422402948141098,
    "learning_rate": 1.7825913638097583e-07,
    "timestep_mean": 439.53125,
    "epsilon_norm": 7.446900367736816,
    "prediction_norm": 6.384446620941162,
    "timestamp": "2025-07-14T12:34:45.269731"
  },
  {
    "epoch": 1983,
    "batch_idx": 128,
    "loss": 0.15110307931900024,
    "learning_rate": 1.7825913638097583e-07,
    "timestep_mean": 502.96875,
    "epsilon_norm": 8.364310264587402,
    "prediction_norm": 7.931623458862305,
    "timestamp": "2025-07-14T12:34:45.274951"
  },
  {
    "epoch": 1983,
    "batch_idx": 160,
    "loss": 0.31674450635910034,
    "learning_rate": 1.7825913638097583e-07,
    "timestep_mean": 523.5,
    "epsilon_norm": 8.006810188293457,
    "prediction_norm": 6.962766647338867,
    "timestamp": "2025-07-14T12:34:45.280286"
  },
  {
    "epoch": 1983,
    "batch_idx": 192,
    "loss": 0.14091002941131592,
    "learning_rate": 1.7825913638097583e-07,
    "timestep_mean": 528.5,
    "epsilon_norm": 3.92372989654541,
    "prediction_norm": 3.51092529296875,
    "timestamp": "2025-07-14T12:34:45.285424"
  },
  {
    "epoch": 1984,
    "batch_idx": 0,
    "loss": 0.1503886580467224,
    "learning_rate": 1.579053583500299e-07,
    "timestep_mean": 527.40625,
    "epsilon_norm": 8.732853889465332,
    "prediction_norm": 7.964510917663574,
    "timestamp": "2025-07-14T12:34:45.290604"
  },
  {
    "epoch": 1984,
    "batch_idx": 32,
    "loss": 0.29438096284866333,
    "learning_rate": 1.579053583500299e-07,
    "timestep_mean": 472.96875,
    "epsilon_norm": 8.52544116973877,
    "prediction_norm": 7.0571794509887695,
    "timestamp": "2025-07-14T12:34:45.295805"
  },
  {
    "epoch": 1984,
    "batch_idx": 64,
    "loss": 0.25685638189315796,
    "learning_rate": 1.579053583500299e-07,
    "timestep_mean": 458.21875,
    "epsilon_norm": 6.473594665527344,
    "prediction_norm": 5.922549247741699,
    "timestamp": "2025-07-14T12:34:45.301218"
  },
  {
    "epoch": 1984,
    "batch_idx": 96,
    "loss": 0.11876283586025238,
    "learning_rate": 1.579053583500299e-07,
    "timestep_mean": 521.8125,
    "epsilon_norm": 8.270727157592773,
    "prediction_norm": 7.227529048919678,
    "timestamp": "2025-07-14T12:34:45.306324"
  },
  {
    "epoch": 1984,
    "batch_idx": 128,
    "loss": 0.4711458086967468,
    "learning_rate": 1.579053583500299e-07,
    "timestep_mean": 446.15625,
    "epsilon_norm": 8.024596214294434,
    "prediction_norm": 6.125805377960205,
    "timestamp": "2025-07-14T12:34:45.311585"
  },
  {
    "epoch": 1984,
    "batch_idx": 160,
    "loss": 0.16919797658920288,
    "learning_rate": 1.579053583500299e-07,
    "timestep_mean": 630.5625,
    "epsilon_norm": 8.03829288482666,
    "prediction_norm": 7.4740166664123535,
    "timestamp": "2025-07-14T12:34:45.316909"
  },
  {
    "epoch": 1984,
    "batch_idx": 192,
    "loss": 0.24827882647514343,
    "learning_rate": 1.579053583500299e-07,
    "timestep_mean": 552.5,
    "epsilon_norm": 3.7591257095336914,
    "prediction_norm": 3.4593148231506348,
    "timestamp": "2025-07-14T12:34:45.321898"
  },
  {
    "epoch": 1985,
    "batch_idx": 0,
    "loss": 0.21247634291648865,
    "learning_rate": 1.387848909997212e-07,
    "timestep_mean": 536.34375,
    "epsilon_norm": 7.4518656730651855,
    "prediction_norm": 6.3856964111328125,
    "timestamp": "2025-07-14T12:34:45.326987"
  },
  {
    "epoch": 1985,
    "batch_idx": 32,
    "loss": 0.3769167363643646,
    "learning_rate": 1.387848909997212e-07,
    "timestep_mean": 499.6875,
    "epsilon_norm": 9.452345848083496,
    "prediction_norm": 7.575678825378418,
    "timestamp": "2025-07-14T12:34:45.332122"
  },
  {
    "epoch": 1985,
    "batch_idx": 64,
    "loss": 0.20469672977924347,
    "learning_rate": 1.387848909997212e-07,
    "timestep_mean": 569.71875,
    "epsilon_norm": 7.0148138999938965,
    "prediction_norm": 5.883691310882568,
    "timestamp": "2025-07-14T12:34:45.337089"
  },
  {
    "epoch": 1985,
    "batch_idx": 96,
    "loss": 0.27301928400993347,
    "learning_rate": 1.387848909997212e-07,
    "timestep_mean": 415.59375,
    "epsilon_norm": 7.376636981964111,
    "prediction_norm": 6.23836612701416,
    "timestamp": "2025-07-14T12:34:45.342029"
  },
  {
    "epoch": 1985,
    "batch_idx": 128,
    "loss": 0.22159649431705475,
    "learning_rate": 1.387848909997212e-07,
    "timestep_mean": 551.375,
    "epsilon_norm": 8.045788764953613,
    "prediction_norm": 7.525063991546631,
    "timestamp": "2025-07-14T12:34:45.346918"
  },
  {
    "epoch": 1985,
    "batch_idx": 160,
    "loss": 0.20301121473312378,
    "learning_rate": 1.387848909997212e-07,
    "timestep_mean": 453.125,
    "epsilon_norm": 8.686610221862793,
    "prediction_norm": 7.2736897468566895,
    "timestamp": "2025-07-14T12:34:45.351862"
  },
  {
    "epoch": 1985,
    "batch_idx": 192,
    "loss": 0.9010394811630249,
    "learning_rate": 1.387848909997212e-07,
    "timestep_mean": 573.375,
    "epsilon_norm": 4.853060245513916,
    "prediction_norm": 3.4752261638641357,
    "timestamp": "2025-07-14T12:34:45.356720"
  },
  {
    "epoch": 1986,
    "batch_idx": 0,
    "loss": 0.20522654056549072,
    "learning_rate": 1.20897781507978e-07,
    "timestep_mean": 534.46875,
    "epsilon_norm": 6.734370231628418,
    "prediction_norm": 5.92877197265625,
    "timestamp": "2025-07-14T12:34:45.361828"
  },
  {
    "epoch": 1986,
    "batch_idx": 32,
    "loss": 0.3068959414958954,
    "learning_rate": 1.20897781507978e-07,
    "timestep_mean": 478.9375,
    "epsilon_norm": 7.926941871643066,
    "prediction_norm": 6.696016788482666,
    "timestamp": "2025-07-14T12:34:45.367006"
  },
  {
    "epoch": 1986,
    "batch_idx": 64,
    "loss": 0.18202587962150574,
    "learning_rate": 1.20897781507978e-07,
    "timestep_mean": 548.6875,
    "epsilon_norm": 6.766497611999512,
    "prediction_norm": 5.854007720947266,
    "timestamp": "2025-07-14T12:34:45.372107"
  },
  {
    "epoch": 1986,
    "batch_idx": 96,
    "loss": 0.3383231461048126,
    "learning_rate": 1.20897781507978e-07,
    "timestep_mean": 538.34375,
    "epsilon_norm": 8.07784366607666,
    "prediction_norm": 6.622640609741211,
    "timestamp": "2025-07-14T12:34:45.377208"
  },
  {
    "epoch": 1986,
    "batch_idx": 128,
    "loss": 0.294126033782959,
    "learning_rate": 1.20897781507978e-07,
    "timestep_mean": 424.78125,
    "epsilon_norm": 8.203064918518066,
    "prediction_norm": 6.942135810852051,
    "timestamp": "2025-07-14T12:34:45.382340"
  },
  {
    "epoch": 1986,
    "batch_idx": 160,
    "loss": 0.15913835167884827,
    "learning_rate": 1.20897781507978e-07,
    "timestep_mean": 494.21875,
    "epsilon_norm": 7.548557758331299,
    "prediction_norm": 6.228058338165283,
    "timestamp": "2025-07-14T12:34:45.387498"
  },
  {
    "epoch": 1986,
    "batch_idx": 192,
    "loss": 0.07637571543455124,
    "learning_rate": 1.20897781507978e-07,
    "timestep_mean": 569.375,
    "epsilon_norm": 3.1647491455078125,
    "prediction_norm": 2.8149845600128174,
    "timestamp": "2025-07-14T12:34:45.392695"
  },
  {
    "epoch": 1987,
    "batch_idx": 0,
    "loss": 0.10931792110204697,
    "learning_rate": 1.0424407400944068e-07,
    "timestep_mean": 553.71875,
    "epsilon_norm": 6.8210930824279785,
    "prediction_norm": 6.249821662902832,
    "timestamp": "2025-07-14T12:34:45.397826"
  },
  {
    "epoch": 1987,
    "batch_idx": 32,
    "loss": 0.17841529846191406,
    "learning_rate": 1.0424407400944068e-07,
    "timestep_mean": 495.71875,
    "epsilon_norm": 8.553777694702148,
    "prediction_norm": 8.326436996459961,
    "timestamp": "2025-07-14T12:34:45.402724"
  },
  {
    "epoch": 1987,
    "batch_idx": 64,
    "loss": 0.10351064056158066,
    "learning_rate": 1.0424407400944068e-07,
    "timestep_mean": 568.125,
    "epsilon_norm": 7.368546962738037,
    "prediction_norm": 6.786435127258301,
    "timestamp": "2025-07-14T12:34:45.407612"
  },
  {
    "epoch": 1987,
    "batch_idx": 96,
    "loss": 0.3013046979904175,
    "learning_rate": 1.0424407400944068e-07,
    "timestep_mean": 486.9375,
    "epsilon_norm": 8.360803604125977,
    "prediction_norm": 7.049742221832275,
    "timestamp": "2025-07-14T12:34:45.412469"
  },
  {
    "epoch": 1987,
    "batch_idx": 128,
    "loss": 0.22481051087379456,
    "learning_rate": 1.0424407400944068e-07,
    "timestep_mean": 584.5625,
    "epsilon_norm": 8.358917236328125,
    "prediction_norm": 7.381228446960449,
    "timestamp": "2025-07-14T12:34:45.417753"
  },
  {
    "epoch": 1987,
    "batch_idx": 160,
    "loss": 0.3048064112663269,
    "learning_rate": 1.0424407400944068e-07,
    "timestep_mean": 466.15625,
    "epsilon_norm": 8.791650772094727,
    "prediction_norm": 7.776822566986084,
    "timestamp": "2025-07-14T12:34:45.423065"
  },
  {
    "epoch": 1987,
    "batch_idx": 192,
    "loss": 0.30987879633903503,
    "learning_rate": 1.0424407400944068e-07,
    "timestep_mean": 508.5,
    "epsilon_norm": 3.853116273880005,
    "prediction_norm": 3.2017288208007812,
    "timestamp": "2025-07-14T12:34:45.428395"
  },
  {
    "epoch": 1988,
    "batch_idx": 0,
    "loss": 0.18355673551559448,
    "learning_rate": 8.882380959551728e-08,
    "timestep_mean": 559.71875,
    "epsilon_norm": 7.446311950683594,
    "prediction_norm": 6.892214298248291,
    "timestamp": "2025-07-14T12:34:45.433768"
  },
  {
    "epoch": 1988,
    "batch_idx": 32,
    "loss": 0.47723907232284546,
    "learning_rate": 8.882380959551728e-08,
    "timestep_mean": 430.75,
    "epsilon_norm": 8.068632125854492,
    "prediction_norm": 6.274762153625488,
    "timestamp": "2025-07-14T12:34:45.438982"
  },
  {
    "epoch": 1988,
    "batch_idx": 64,
    "loss": 0.3372017741203308,
    "learning_rate": 8.882380959551728e-08,
    "timestep_mean": 469.59375,
    "epsilon_norm": 8.484949111938477,
    "prediction_norm": 6.972304344177246,
    "timestamp": "2025-07-14T12:34:45.444082"
  },
  {
    "epoch": 1988,
    "batch_idx": 96,
    "loss": 0.33640211820602417,
    "learning_rate": 8.882380959551728e-08,
    "timestep_mean": 523.28125,
    "epsilon_norm": 8.513955116271973,
    "prediction_norm": 6.686304092407227,
    "timestamp": "2025-07-14T12:34:45.449172"
  },
  {
    "epoch": 1988,
    "batch_idx": 128,
    "loss": 0.18769291043281555,
    "learning_rate": 8.882380959551728e-08,
    "timestep_mean": 533.375,
    "epsilon_norm": 7.022512912750244,
    "prediction_norm": 6.856949806213379,
    "timestamp": "2025-07-14T12:34:45.454338"
  },
  {
    "epoch": 1988,
    "batch_idx": 160,
    "loss": 0.31870219111442566,
    "learning_rate": 8.882380959551728e-08,
    "timestep_mean": 495.21875,
    "epsilon_norm": 9.18952751159668,
    "prediction_norm": 7.406091690063477,
    "timestamp": "2025-07-14T12:34:45.459455"
  },
  {
    "epoch": 1988,
    "batch_idx": 192,
    "loss": 0.02390475943684578,
    "learning_rate": 8.882380959551728e-08,
    "timestep_mean": 751.375,
    "epsilon_norm": 5.00400972366333,
    "prediction_norm": 4.853988170623779,
    "timestamp": "2025-07-14T12:34:45.464507"
  },
  {
    "epoch": 1989,
    "batch_idx": 0,
    "loss": 0.21656981110572815,
    "learning_rate": 7.463702631410598e-08,
    "timestep_mean": 563.8125,
    "epsilon_norm": 8.444339752197266,
    "prediction_norm": 7.654046535491943,
    "timestamp": "2025-07-14T12:34:45.469654"
  },
  {
    "epoch": 1989,
    "batch_idx": 32,
    "loss": 0.4259842336177826,
    "learning_rate": 7.463702631410598e-08,
    "timestep_mean": 479.75,
    "epsilon_norm": 8.704063415527344,
    "prediction_norm": 7.417688846588135,
    "timestamp": "2025-07-14T12:34:45.474849"
  },
  {
    "epoch": 1989,
    "batch_idx": 64,
    "loss": 0.31291908025741577,
    "learning_rate": 7.463702631410598e-08,
    "timestep_mean": 489.28125,
    "epsilon_norm": 8.106876373291016,
    "prediction_norm": 6.970376014709473,
    "timestamp": "2025-07-14T12:34:45.479993"
  },
  {
    "epoch": 1989,
    "batch_idx": 96,
    "loss": 0.12873150408267975,
    "learning_rate": 7.463702631410598e-08,
    "timestep_mean": 551.59375,
    "epsilon_norm": 6.066528797149658,
    "prediction_norm": 5.513573169708252,
    "timestamp": "2025-07-14T12:34:45.485114"
  },
  {
    "epoch": 1989,
    "batch_idx": 128,
    "loss": 0.3712995648384094,
    "learning_rate": 7.463702631410598e-08,
    "timestep_mean": 492.21875,
    "epsilon_norm": 8.682504653930664,
    "prediction_norm": 6.759275436401367,
    "timestamp": "2025-07-14T12:34:45.490215"
  },
  {
    "epoch": 1989,
    "batch_idx": 160,
    "loss": 0.43718221783638,
    "learning_rate": 7.463702631410598e-08,
    "timestep_mean": 429.9375,
    "epsilon_norm": 8.181906700134277,
    "prediction_norm": 5.739786624908447,
    "timestamp": "2025-07-14T12:34:45.495361"
  },
  {
    "epoch": 1989,
    "batch_idx": 192,
    "loss": 0.3386383652687073,
    "learning_rate": 7.463702631410598e-08,
    "timestep_mean": 500.375,
    "epsilon_norm": 3.6427016258239746,
    "prediction_norm": 2.548078775405884,
    "timestamp": "2025-07-14T12:34:45.500567"
  },
  {
    "epoch": 1990,
    "batch_idx": 0,
    "loss": 0.379803329706192,
    "learning_rate": 6.168375916970607e-08,
    "timestep_mean": 394.4375,
    "epsilon_norm": 7.464416980743408,
    "prediction_norm": 5.635301113128662,
    "timestamp": "2025-07-14T12:34:45.505705"
  },
  {
    "epoch": 1990,
    "batch_idx": 32,
    "loss": 0.20884636044502258,
    "learning_rate": 6.168375916970607e-08,
    "timestep_mean": 566.375,
    "epsilon_norm": 8.348252296447754,
    "prediction_norm": 7.406996250152588,
    "timestamp": "2025-07-14T12:34:45.510890"
  },
  {
    "epoch": 1990,
    "batch_idx": 64,
    "loss": 0.41860613226890564,
    "learning_rate": 6.168375916970607e-08,
    "timestep_mean": 475.9375,
    "epsilon_norm": 8.628820419311523,
    "prediction_norm": 6.371829986572266,
    "timestamp": "2025-07-14T12:34:45.516063"
  },
  {
    "epoch": 1990,
    "batch_idx": 96,
    "loss": 0.3127283751964569,
    "learning_rate": 6.168375916970607e-08,
    "timestep_mean": 512.6875,
    "epsilon_norm": 7.1129469871521,
    "prediction_norm": 5.730189323425293,
    "timestamp": "2025-07-14T12:34:45.521291"
  },
  {
    "epoch": 1990,
    "batch_idx": 128,
    "loss": 0.23760105669498444,
    "learning_rate": 6.168375916970607e-08,
    "timestep_mean": 544.71875,
    "epsilon_norm": 8.316822052001953,
    "prediction_norm": 7.1822381019592285,
    "timestamp": "2025-07-14T12:34:45.526302"
  },
  {
    "epoch": 1990,
    "batch_idx": 160,
    "loss": 0.30124199390411377,
    "learning_rate": 6.168375916970607e-08,
    "timestep_mean": 519.125,
    "epsilon_norm": 8.449379920959473,
    "prediction_norm": 7.240670680999756,
    "timestamp": "2025-07-14T12:34:45.531397"
  },
  {
    "epoch": 1990,
    "batch_idx": 192,
    "loss": 0.0838480219244957,
    "learning_rate": 6.168375916970607e-08,
    "timestep_mean": 527.125,
    "epsilon_norm": 4.454960823059082,
    "prediction_norm": 4.055154800415039,
    "timestamp": "2025-07-14T12:34:45.536497"
  },
  {
    "epoch": 1991,
    "batch_idx": 0,
    "loss": 0.19929391145706177,
    "learning_rate": 4.996404012325145e-08,
    "timestep_mean": 544.125,
    "epsilon_norm": 8.194746971130371,
    "prediction_norm": 6.881514549255371,
    "timestamp": "2025-07-14T12:34:45.541473"
  },
  {
    "epoch": 1991,
    "batch_idx": 32,
    "loss": 0.19263184070587158,
    "learning_rate": 4.996404012325145e-08,
    "timestep_mean": 449.5625,
    "epsilon_norm": 7.405101776123047,
    "prediction_norm": 6.375148773193359,
    "timestamp": "2025-07-14T12:34:45.546780"
  },
  {
    "epoch": 1991,
    "batch_idx": 64,
    "loss": 0.11048842966556549,
    "learning_rate": 4.996404012325145e-08,
    "timestep_mean": 562.8125,
    "epsilon_norm": 8.395757675170898,
    "prediction_norm": 7.723439693450928,
    "timestamp": "2025-07-14T12:34:45.551944"
  },
  {
    "epoch": 1991,
    "batch_idx": 96,
    "loss": 0.40045541524887085,
    "learning_rate": 4.996404012325145e-08,
    "timestep_mean": 511.90625,
    "epsilon_norm": 8.449835777282715,
    "prediction_norm": 7.177546977996826,
    "timestamp": "2025-07-14T12:34:45.557047"
  },
  {
    "epoch": 1991,
    "batch_idx": 128,
    "loss": 0.29279276728630066,
    "learning_rate": 4.996404012325145e-08,
    "timestep_mean": 457.1875,
    "epsilon_norm": 7.20535945892334,
    "prediction_norm": 5.858462333679199,
    "timestamp": "2025-07-14T12:34:45.562264"
  },
  {
    "epoch": 1991,
    "batch_idx": 160,
    "loss": 0.45910608768463135,
    "learning_rate": 4.996404012325145e-08,
    "timestep_mean": 425.5625,
    "epsilon_norm": 9.00769329071045,
    "prediction_norm": 7.255688667297363,
    "timestamp": "2025-07-14T12:34:45.567567"
  },
  {
    "epoch": 1991,
    "batch_idx": 192,
    "loss": 0.13967370986938477,
    "learning_rate": 4.996404012325145e-08,
    "timestep_mean": 607.625,
    "epsilon_norm": 3.0741124153137207,
    "prediction_norm": 2.5418038368225098,
    "timestamp": "2025-07-14T12:34:45.572660"
  },
  {
    "epoch": 1992,
    "batch_idx": 0,
    "loss": 0.5268644094467163,
    "learning_rate": 3.947789809194408e-08,
    "timestep_mean": 457.625,
    "epsilon_norm": 8.858708381652832,
    "prediction_norm": 6.456861972808838,
    "timestamp": "2025-07-14T12:34:45.577873"
  },
  {
    "epoch": 1992,
    "batch_idx": 32,
    "loss": 0.30532947182655334,
    "learning_rate": 3.947789809194408e-08,
    "timestep_mean": 465.0625,
    "epsilon_norm": 7.552252769470215,
    "prediction_norm": 5.783445835113525,
    "timestamp": "2025-07-14T12:34:45.583008"
  },
  {
    "epoch": 1992,
    "batch_idx": 64,
    "loss": 0.2541634142398834,
    "learning_rate": 3.947789809194408e-08,
    "timestep_mean": 520.3125,
    "epsilon_norm": 8.078271865844727,
    "prediction_norm": 6.630857467651367,
    "timestamp": "2025-07-14T12:34:45.588317"
  },
  {
    "epoch": 1992,
    "batch_idx": 96,
    "loss": 0.35552364587783813,
    "learning_rate": 3.947789809194408e-08,
    "timestep_mean": 442.90625,
    "epsilon_norm": 8.312975883483887,
    "prediction_norm": 7.242526531219482,
    "timestamp": "2025-07-14T12:34:45.593540"
  },
  {
    "epoch": 1992,
    "batch_idx": 128,
    "loss": 0.07903192192316055,
    "learning_rate": 3.947789809194408e-08,
    "timestep_mean": 615.625,
    "epsilon_norm": 8.401151657104492,
    "prediction_norm": 7.562169075012207,
    "timestamp": "2025-07-14T12:34:45.598552"
  },
  {
    "epoch": 1992,
    "batch_idx": 160,
    "loss": 0.22868391871452332,
    "learning_rate": 3.947789809194408e-08,
    "timestep_mean": 465.0625,
    "epsilon_norm": 7.9222412109375,
    "prediction_norm": 6.743739128112793,
    "timestamp": "2025-07-14T12:34:45.603765"
  },
  {
    "epoch": 1992,
    "batch_idx": 192,
    "loss": 0.5382524728775024,
    "learning_rate": 3.947789809194408e-08,
    "timestep_mean": 372.75,
    "epsilon_norm": 4.1910529136657715,
    "prediction_norm": 2.4132800102233887,
    "timestamp": "2025-07-14T12:34:45.609045"
  },
  {
    "epoch": 1993,
    "batch_idx": 0,
    "loss": 0.23343107104301453,
    "learning_rate": 3.022535894930952e-08,
    "timestep_mean": 506.96875,
    "epsilon_norm": 8.905097961425781,
    "prediction_norm": 7.516758918762207,
    "timestamp": "2025-07-14T12:34:45.614265"
  },
  {
    "epoch": 1993,
    "batch_idx": 32,
    "loss": 0.46619054675102234,
    "learning_rate": 3.022535894930952e-08,
    "timestep_mean": 548.40625,
    "epsilon_norm": 7.841280460357666,
    "prediction_norm": 6.296192169189453,
    "timestamp": "2025-07-14T12:34:45.619418"
  },
  {
    "epoch": 1993,
    "batch_idx": 64,
    "loss": 0.244109109044075,
    "learning_rate": 3.022535894930952e-08,
    "timestep_mean": 502.84375,
    "epsilon_norm": 6.786818027496338,
    "prediction_norm": 5.956867218017578,
    "timestamp": "2025-07-14T12:34:45.624602"
  },
  {
    "epoch": 1993,
    "batch_idx": 96,
    "loss": 0.37125125527381897,
    "learning_rate": 3.022535894930952e-08,
    "timestep_mean": 459.40625,
    "epsilon_norm": 7.148087024688721,
    "prediction_norm": 5.950343132019043,
    "timestamp": "2025-07-14T12:34:45.629938"
  },
  {
    "epoch": 1993,
    "batch_idx": 128,
    "loss": 0.2139514684677124,
    "learning_rate": 3.022535894930952e-08,
    "timestep_mean": 476.25,
    "epsilon_norm": 8.700762748718262,
    "prediction_norm": 7.981479644775391,
    "timestamp": "2025-07-14T12:34:45.635017"
  },
  {
    "epoch": 1993,
    "batch_idx": 160,
    "loss": 0.2509978711605072,
    "learning_rate": 3.022535894930952e-08,
    "timestep_mean": 509.21875,
    "epsilon_norm": 7.892393589019775,
    "prediction_norm": 6.70067834854126,
    "timestamp": "2025-07-14T12:34:45.640123"
  },
  {
    "epoch": 1993,
    "batch_idx": 192,
    "loss": 0.22712558507919312,
    "learning_rate": 3.022535894930952e-08,
    "timestep_mean": 497.25,
    "epsilon_norm": 4.092855453491211,
    "prediction_norm": 3.3463268280029297,
    "timestamp": "2025-07-14T12:34:45.645205"
  },
  {
    "epoch": 1994,
    "batch_idx": 0,
    "loss": 0.262448251247406,
    "learning_rate": 2.2206445525085855e-08,
    "timestep_mean": 492.84375,
    "epsilon_norm": 7.106161117553711,
    "prediction_norm": 6.224088191986084,
    "timestamp": "2025-07-14T12:34:45.650395"
  },
  {
    "epoch": 1994,
    "batch_idx": 32,
    "loss": 0.1743071973323822,
    "learning_rate": 2.2206445525085855e-08,
    "timestep_mean": 622.5,
    "epsilon_norm": 6.333511829376221,
    "prediction_norm": 5.506492614746094,
    "timestamp": "2025-07-14T12:34:45.655267"
  },
  {
    "epoch": 1994,
    "batch_idx": 64,
    "loss": 0.16659294068813324,
    "learning_rate": 2.2206445525085855e-08,
    "timestep_mean": 437.6875,
    "epsilon_norm": 6.263068675994873,
    "prediction_norm": 5.873767375946045,
    "timestamp": "2025-07-14T12:34:45.660178"
  },
  {
    "epoch": 1994,
    "batch_idx": 96,
    "loss": 0.30257320404052734,
    "learning_rate": 2.2206445525085855e-08,
    "timestep_mean": 523.84375,
    "epsilon_norm": 7.643557071685791,
    "prediction_norm": 6.880493640899658,
    "timestamp": "2025-07-14T12:34:45.665004"
  },
  {
    "epoch": 1994,
    "batch_idx": 128,
    "loss": 0.27271148562431335,
    "learning_rate": 2.2206445525085855e-08,
    "timestep_mean": 442.0625,
    "epsilon_norm": 8.262203216552734,
    "prediction_norm": 6.753617286682129,
    "timestamp": "2025-07-14T12:34:45.669825"
  },
  {
    "epoch": 1994,
    "batch_idx": 160,
    "loss": 0.2959825098514557,
    "learning_rate": 2.2206445525085855e-08,
    "timestep_mean": 534.25,
    "epsilon_norm": 7.31088399887085,
    "prediction_norm": 6.280914306640625,
    "timestamp": "2025-07-14T12:34:45.674666"
  },
  {
    "epoch": 1994,
    "batch_idx": 192,
    "loss": 0.0425332635641098,
    "learning_rate": 2.2206445525085855e-08,
    "timestep_mean": 488.25,
    "epsilon_norm": 3.98592472076416,
    "prediction_norm": 4.260024070739746,
    "timestamp": "2025-07-14T12:34:45.679563"
  },
  {
    "epoch": 1995,
    "batch_idx": 0,
    "loss": 0.37697505950927734,
    "learning_rate": 1.5421177605168238e-08,
    "timestep_mean": 511.90625,
    "epsilon_norm": 8.925588607788086,
    "prediction_norm": 7.868684768676758,
    "timestamp": "2025-07-14T12:34:45.684603"
  },
  {
    "epoch": 1995,
    "batch_idx": 32,
    "loss": 0.30666857957839966,
    "learning_rate": 1.5421177605168238e-08,
    "timestep_mean": 621.40625,
    "epsilon_norm": 8.905216217041016,
    "prediction_norm": 8.038509368896484,
    "timestamp": "2025-07-14T12:34:45.689655"
  },
  {
    "epoch": 1995,
    "batch_idx": 64,
    "loss": 0.16734907031059265,
    "learning_rate": 1.5421177605168238e-08,
    "timestep_mean": 507.625,
    "epsilon_norm": 6.271418571472168,
    "prediction_norm": 5.557671546936035,
    "timestamp": "2025-07-14T12:34:45.694750"
  },
  {
    "epoch": 1995,
    "batch_idx": 96,
    "loss": 0.12906748056411743,
    "learning_rate": 1.5421177605168238e-08,
    "timestep_mean": 501.5,
    "epsilon_norm": 6.210880756378174,
    "prediction_norm": 5.554234981536865,
    "timestamp": "2025-07-14T12:34:45.699865"
  },
  {
    "epoch": 1995,
    "batch_idx": 128,
    "loss": 0.2513670027256012,
    "learning_rate": 1.5421177605168238e-08,
    "timestep_mean": 420.59375,
    "epsilon_norm": 8.183783531188965,
    "prediction_norm": 7.458987236022949,
    "timestamp": "2025-07-14T12:34:45.705100"
  },
  {
    "epoch": 1995,
    "batch_idx": 160,
    "loss": 0.2173812985420227,
    "learning_rate": 1.5421177605168238e-08,
    "timestep_mean": 487.6875,
    "epsilon_norm": 6.744717597961426,
    "prediction_norm": 6.049226760864258,
    "timestamp": "2025-07-14T12:34:45.710203"
  },
  {
    "epoch": 1995,
    "batch_idx": 192,
    "loss": 0.4242018461227417,
    "learning_rate": 1.5421177605168238e-08,
    "timestep_mean": 332.5,
    "epsilon_norm": 4.660519123077393,
    "prediction_norm": 3.80383563041687,
    "timestamp": "2025-07-14T12:34:45.715278"
  },
  {
    "epoch": 1996,
    "batch_idx": 0,
    "loss": 0.26082760095596313,
    "learning_rate": 9.86957193144232e-09,
    "timestep_mean": 571.0,
    "epsilon_norm": 7.2441086769104,
    "prediction_norm": 5.448973178863525,
    "timestamp": "2025-07-14T12:34:45.720504"
  },
  {
    "epoch": 1996,
    "batch_idx": 32,
    "loss": 0.19131335616111755,
    "learning_rate": 9.86957193144232e-09,
    "timestep_mean": 565.21875,
    "epsilon_norm": 9.095464706420898,
    "prediction_norm": 8.543658256530762,
    "timestamp": "2025-07-14T12:34:45.725706"
  },
  {
    "epoch": 1996,
    "batch_idx": 64,
    "loss": 0.3257454037666321,
    "learning_rate": 9.86957193144232e-09,
    "timestep_mean": 497.03125,
    "epsilon_norm": 8.590946197509766,
    "prediction_norm": 7.1111578941345215,
    "timestamp": "2025-07-14T12:34:45.730868"
  },
  {
    "epoch": 1996,
    "batch_idx": 96,
    "loss": 0.2601814270019531,
    "learning_rate": 9.86957193144232e-09,
    "timestep_mean": 462.0625,
    "epsilon_norm": 6.6656365394592285,
    "prediction_norm": 5.439820289611816,
    "timestamp": "2025-07-14T12:34:45.736011"
  },
  {
    "epoch": 1996,
    "batch_idx": 128,
    "loss": 0.4905391335487366,
    "learning_rate": 9.86957193144232e-09,
    "timestep_mean": 407.21875,
    "epsilon_norm": 7.742193698883057,
    "prediction_norm": 5.7363176345825195,
    "timestamp": "2025-07-14T12:34:45.741167"
  },
  {
    "epoch": 1996,
    "batch_idx": 160,
    "loss": 0.35042452812194824,
    "learning_rate": 9.86957193144232e-09,
    "timestep_mean": 470.1875,
    "epsilon_norm": 8.400646209716797,
    "prediction_norm": 7.026433944702148,
    "timestamp": "2025-07-14T12:34:45.746495"
  },
  {
    "epoch": 1996,
    "batch_idx": 192,
    "loss": 0.8846811652183533,
    "learning_rate": 9.86957193144232e-09,
    "timestep_mean": 533.125,
    "epsilon_norm": 4.5467071533203125,
    "prediction_norm": 3.233229637145996,
    "timestamp": "2025-07-14T12:34:45.751598"
  },
  {
    "epoch": 1997,
    "batch_idx": 0,
    "loss": 0.3721756339073181,
    "learning_rate": 5.551642202006317e-09,
    "timestep_mean": 390.09375,
    "epsilon_norm": 8.199488639831543,
    "prediction_norm": 6.560333728790283,
    "timestamp": "2025-07-14T12:34:45.756711"
  },
  {
    "epoch": 1997,
    "batch_idx": 32,
    "loss": 0.19471481442451477,
    "learning_rate": 5.551642202006317e-09,
    "timestep_mean": 481.59375,
    "epsilon_norm": 7.8990936279296875,
    "prediction_norm": 6.821253776550293,
    "timestamp": "2025-07-14T12:34:45.761945"
  },
  {
    "epoch": 1997,
    "batch_idx": 64,
    "loss": 0.1933298110961914,
    "learning_rate": 5.551642202006317e-09,
    "timestep_mean": 533.6875,
    "epsilon_norm": 7.352922439575195,
    "prediction_norm": 6.898579120635986,
    "timestamp": "2025-07-14T12:34:45.767259"
  },
  {
    "epoch": 1997,
    "batch_idx": 96,
    "loss": 0.07477282732725143,
    "learning_rate": 5.551642202006317e-09,
    "timestep_mean": 619.875,
    "epsilon_norm": 7.352676868438721,
    "prediction_norm": 6.961706638336182,
    "timestamp": "2025-07-14T12:34:45.772458"
  },
  {
    "epoch": 1997,
    "batch_idx": 128,
    "loss": 0.2548794448375702,
    "learning_rate": 5.551642202006317e-09,
    "timestep_mean": 501.96875,
    "epsilon_norm": 6.949882507324219,
    "prediction_norm": 6.020275592803955,
    "timestamp": "2025-07-14T12:34:45.777624"
  },
  {
    "epoch": 1997,
    "batch_idx": 160,
    "loss": 0.2118874490261078,
    "learning_rate": 5.551642202006317e-09,
    "timestep_mean": 501.875,
    "epsilon_norm": 7.772165775299072,
    "prediction_norm": 7.0711493492126465,
    "timestamp": "2025-07-14T12:34:45.782691"
  },
  {
    "epoch": 1997,
    "batch_idx": 192,
    "loss": 0.18201479315757751,
    "learning_rate": 5.551642202006317e-09,
    "timestep_mean": 545.375,
    "epsilon_norm": 4.263891220092773,
    "prediction_norm": 3.83382248878479,
    "timestamp": "2025-07-14T12:34:45.787838"
  },
  {
    "epoch": 1998,
    "batch_idx": 0,
    "loss": 0.15670745074748993,
    "learning_rate": 2.467399070893436e-09,
    "timestep_mean": 555.875,
    "epsilon_norm": 8.170336723327637,
    "prediction_norm": 7.263626575469971,
    "timestamp": "2025-07-14T12:34:45.793045"
  },
  {
    "epoch": 1998,
    "batch_idx": 32,
    "loss": 0.09158742427825928,
    "learning_rate": 2.467399070893436e-09,
    "timestep_mean": 602.28125,
    "epsilon_norm": 8.969364166259766,
    "prediction_norm": 8.414731979370117,
    "timestamp": "2025-07-14T12:34:45.798101"
  },
  {
    "epoch": 1998,
    "batch_idx": 64,
    "loss": 0.5529276132583618,
    "learning_rate": 2.467399070893436e-09,
    "timestep_mean": 511.1875,
    "epsilon_norm": 8.752400398254395,
    "prediction_norm": 6.575862407684326,
    "timestamp": "2025-07-14T12:34:45.803260"
  },
  {
    "epoch": 1998,
    "batch_idx": 96,
    "loss": 0.3394143283367157,
    "learning_rate": 2.467399070893436e-09,
    "timestep_mean": 414.53125,
    "epsilon_norm": 8.148601531982422,
    "prediction_norm": 6.370409965515137,
    "timestamp": "2025-07-14T12:34:45.808575"
  },
  {
    "epoch": 1998,
    "batch_idx": 128,
    "loss": 0.40668123960494995,
    "learning_rate": 2.467399070893436e-09,
    "timestep_mean": 444.15625,
    "epsilon_norm": 8.160552024841309,
    "prediction_norm": 6.455117225646973,
    "timestamp": "2025-07-14T12:34:45.813799"
  },
  {
    "epoch": 1998,
    "batch_idx": 160,
    "loss": 0.30329692363739014,
    "learning_rate": 2.467399070893436e-09,
    "timestep_mean": 473.84375,
    "epsilon_norm": 7.01481294631958,
    "prediction_norm": 5.768001556396484,
    "timestamp": "2025-07-14T12:34:45.818920"
  },
  {
    "epoch": 1998,
    "batch_idx": 192,
    "loss": 0.20780164003372192,
    "learning_rate": 2.467399070893436e-09,
    "timestep_mean": 576.5,
    "epsilon_norm": 3.3539772033691406,
    "prediction_norm": 2.9046692848205566,
    "timestamp": "2025-07-14T12:34:45.824018"
  },
  {
    "epoch": 1999,
    "batch_idx": 0,
    "loss": 0.2084852159023285,
    "learning_rate": 6.168501482384229e-10,
    "timestep_mean": 512.84375,
    "epsilon_norm": 8.49306869506836,
    "prediction_norm": 7.14573860168457,
    "timestamp": "2025-07-14T12:34:45.829249"
  },
  {
    "epoch": 1999,
    "batch_idx": 32,
    "loss": 0.22842542827129364,
    "learning_rate": 6.168501482384229e-10,
    "timestep_mean": 514.09375,
    "epsilon_norm": 9.059733390808105,
    "prediction_norm": 7.993016242980957,
    "timestamp": "2025-07-14T12:34:45.834483"
  },
  {
    "epoch": 1999,
    "batch_idx": 64,
    "loss": 0.149652361869812,
    "learning_rate": 6.168501482384229e-10,
    "timestep_mean": 449.90625,
    "epsilon_norm": 7.534159183502197,
    "prediction_norm": 6.793392658233643,
    "timestamp": "2025-07-14T12:34:45.839601"
  },
  {
    "epoch": 1999,
    "batch_idx": 96,
    "loss": 0.18111446499824524,
    "learning_rate": 6.168501482384229e-10,
    "timestep_mean": 548.5625,
    "epsilon_norm": 7.894271373748779,
    "prediction_norm": 6.981801509857178,
    "timestamp": "2025-07-14T12:34:45.844789"
  },
  {
    "epoch": 1999,
    "batch_idx": 128,
    "loss": 0.3246767520904541,
    "learning_rate": 6.168501482384229e-10,
    "timestep_mean": 537.84375,
    "epsilon_norm": 8.740707397460938,
    "prediction_norm": 7.179518222808838,
    "timestamp": "2025-07-14T12:34:45.849639"
  },
  {
    "epoch": 1999,
    "batch_idx": 160,
    "loss": 0.2254250943660736,
    "learning_rate": 6.168501482384229e-10,
    "timestep_mean": 454.15625,
    "epsilon_norm": 7.445241451263428,
    "prediction_norm": 6.817237377166748,
    "timestamp": "2025-07-14T12:34:45.854899"
  },
  {
    "epoch": 1999,
    "batch_idx": 192,
    "loss": 0.09346063435077667,
    "learning_rate": 6.168501482384229e-10,
    "timestep_mean": 588.375,
    "epsilon_norm": 2.652785539627075,
    "prediction_norm": 2.4155433177948,
    "timestamp": "2025-07-14T12:34:45.859798"
  },
  {
    "epoch": 2000,
    "batch_idx": 0,
    "loss": 0.2754670977592468,
    "learning_rate": 0.0,
    "timestep_mean": 500.21875,
    "epsilon_norm": 7.6226701736450195,
    "prediction_norm": 5.86544132232666,
    "timestamp": "2025-07-14T12:34:45.864394"
  },
  {
    "epoch": 2000,
    "batch_idx": 32,
    "loss": 0.32409271597862244,
    "learning_rate": 0.0,
    "timestep_mean": 458.21875,
    "epsilon_norm": 7.750141143798828,
    "prediction_norm": 6.545510768890381,
    "timestamp": "2025-07-14T12:34:45.868922"
  },
  {
    "epoch": 2000,
    "batch_idx": 64,
    "loss": 0.4486657381057739,
    "learning_rate": 0.0,
    "timestep_mean": 544.5,
    "epsilon_norm": 8.969834327697754,
    "prediction_norm": 7.079329490661621,
    "timestamp": "2025-07-14T12:34:45.873531"
  },
  {
    "epoch": 2000,
    "batch_idx": 96,
    "loss": 0.24059313535690308,
    "learning_rate": 0.0,
    "timestep_mean": 455.3125,
    "epsilon_norm": 7.51449728012085,
    "prediction_norm": 6.351263523101807,
    "timestamp": "2025-07-14T12:34:45.878218"
  },
  {
    "epoch": 2000,
    "batch_idx": 128,
    "loss": 0.46138009428977966,
    "learning_rate": 0.0,
    "timestep_mean": 535.78125,
    "epsilon_norm": 8.089509963989258,
    "prediction_norm": 5.996744632720947,
    "timestamp": "2025-07-14T12:34:45.882959"
  },
  {
    "epoch": 2000,
    "batch_idx": 160,
    "loss": 0.0984613224864006,
    "learning_rate": 0.0,
    "timestep_mean": 640.5625,
    "epsilon_norm": 7.583261489868164,
    "prediction_norm": 6.922745227813721,
    "timestamp": "2025-07-14T12:34:45.887831"
  },
  {
    "epoch": 2000,
    "batch_idx": 192,
    "loss": 0.13269856572151184,
    "learning_rate": 0.0,
    "timestep_mean": 278.625,
    "epsilon_norm": 4.017440319061279,
    "prediction_norm": 3.7973902225494385,
    "timestamp": "2025-07-14T12:34:45.892750"
  },
  {
    "epoch": 2001,
    "batch_idx": 0,
    "loss": 0.20281821489334106,
    "learning_rate": 6.168501482384237e-10,
    "timestep_mean": 461.875,
    "epsilon_norm": 8.287820816040039,
    "prediction_norm": 7.313547611236572,
    "timestamp": "2025-07-14T12:34:45.897911"
  },
  {
    "epoch": 2001,
    "batch_idx": 32,
    "loss": 0.420829176902771,
    "learning_rate": 6.168501482384237e-10,
    "timestep_mean": 478.15625,
    "epsilon_norm": 7.6785664558410645,
    "prediction_norm": 6.694278717041016,
    "timestamp": "2025-07-14T12:34:45.902999"
  },
  {
    "epoch": 2001,
    "batch_idx": 64,
    "loss": 0.2016945481300354,
    "learning_rate": 6.168501482384237e-10,
    "timestep_mean": 527.9375,
    "epsilon_norm": 7.107685565948486,
    "prediction_norm": 6.1200385093688965,
    "timestamp": "2025-07-14T12:34:45.908179"
  },
  {
    "epoch": 2001,
    "batch_idx": 96,
    "loss": 0.26820340752601624,
    "learning_rate": 6.168501482384237e-10,
    "timestep_mean": 509.125,
    "epsilon_norm": 8.42262077331543,
    "prediction_norm": 7.429649829864502,
    "timestamp": "2025-07-14T12:34:45.913416"
  },
  {
    "epoch": 2001,
    "batch_idx": 128,
    "loss": 0.13643288612365723,
    "learning_rate": 6.168501482384237e-10,
    "timestep_mean": 546.78125,
    "epsilon_norm": 7.8850789070129395,
    "prediction_norm": 7.449822902679443,
    "timestamp": "2025-07-14T12:34:45.918452"
  },
  {
    "epoch": 2001,
    "batch_idx": 160,
    "loss": 0.24224911630153656,
    "learning_rate": 6.168501482384237e-10,
    "timestep_mean": 491.5625,
    "epsilon_norm": 7.323684215545654,
    "prediction_norm": 5.990538597106934,
    "timestamp": "2025-07-14T12:34:45.923508"
  },
  {
    "epoch": 2001,
    "batch_idx": 192,
    "loss": 0.2048703283071518,
    "learning_rate": 6.168501482384237e-10,
    "timestep_mean": 505.0,
    "epsilon_norm": 5.41762113571167,
    "prediction_norm": 5.0363688468933105,
    "timestamp": "2025-07-14T12:34:45.928698"
  },
  {
    "epoch": 2002,
    "batch_idx": 0,
    "loss": 0.25111138820648193,
    "learning_rate": 2.467399070893439e-09,
    "timestep_mean": 558.03125,
    "epsilon_norm": 8.531376838684082,
    "prediction_norm": 7.3762969970703125,
    "timestamp": "2025-07-14T12:34:45.933986"
  },
  {
    "epoch": 2002,
    "batch_idx": 32,
    "loss": 0.45113879442214966,
    "learning_rate": 2.467399070893439e-09,
    "timestep_mean": 531.96875,
    "epsilon_norm": 7.475042819976807,
    "prediction_norm": 6.38994836807251,
    "timestamp": "2025-07-14T12:34:45.939143"
  },
  {
    "epoch": 2002,
    "batch_idx": 64,
    "loss": 0.30274033546447754,
    "learning_rate": 2.467399070893439e-09,
    "timestep_mean": 445.9375,
    "epsilon_norm": 7.298751354217529,
    "prediction_norm": 6.286497592926025,
    "timestamp": "2025-07-14T12:34:45.944332"
  },
  {
    "epoch": 2002,
    "batch_idx": 96,
    "loss": 0.21856966614723206,
    "learning_rate": 2.467399070893439e-09,
    "timestep_mean": 495.71875,
    "epsilon_norm": 7.99371337890625,
    "prediction_norm": 7.193982124328613,
    "timestamp": "2025-07-14T12:34:45.949452"
  },
  {
    "epoch": 2002,
    "batch_idx": 128,
    "loss": 0.4859035015106201,
    "learning_rate": 2.467399070893439e-09,
    "timestep_mean": 514.71875,
    "epsilon_norm": 8.125632286071777,
    "prediction_norm": 6.191400527954102,
    "timestamp": "2025-07-14T12:34:45.954665"
  },
  {
    "epoch": 2002,
    "batch_idx": 160,
    "loss": 0.13448914885520935,
    "learning_rate": 2.467399070893439e-09,
    "timestep_mean": 478.59375,
    "epsilon_norm": 7.667906284332275,
    "prediction_norm": 7.289750099182129,
    "timestamp": "2025-07-14T12:34:45.959747"
  },
  {
    "epoch": 2002,
    "batch_idx": 192,
    "loss": 0.3190382122993469,
    "learning_rate": 2.467399070893439e-09,
    "timestep_mean": 171.125,
    "epsilon_norm": 4.397505760192871,
    "prediction_norm": 3.439589500427246,
    "timestamp": "2025-07-14T12:34:45.964790"
  },
  {
    "epoch": 2003,
    "batch_idx": 0,
    "loss": 0.3955863416194916,
    "learning_rate": 5.551642202006324e-09,
    "timestep_mean": 514.59375,
    "epsilon_norm": 8.257905006408691,
    "prediction_norm": 6.426395416259766,
    "timestamp": "2025-07-14T12:34:45.970050"
  },
  {
    "epoch": 2003,
    "batch_idx": 32,
    "loss": 0.302644282579422,
    "learning_rate": 5.551642202006324e-09,
    "timestep_mean": 505.53125,
    "epsilon_norm": 8.014405250549316,
    "prediction_norm": 6.539068698883057,
    "timestamp": "2025-07-14T12:34:45.975400"
  },
  {
    "epoch": 2003,
    "batch_idx": 64,
    "loss": 0.30225443840026855,
    "learning_rate": 5.551642202006324e-09,
    "timestep_mean": 516.5,
    "epsilon_norm": 8.652881622314453,
    "prediction_norm": 7.7597761154174805,
    "timestamp": "2025-07-14T12:34:45.980613"
  },
  {
    "epoch": 2003,
    "batch_idx": 96,
    "loss": 0.22699879109859467,
    "learning_rate": 5.551642202006324e-09,
    "timestep_mean": 512.96875,
    "epsilon_norm": 8.24551773071289,
    "prediction_norm": 6.8988471031188965,
    "timestamp": "2025-07-14T12:34:45.985830"
  },
  {
    "epoch": 2003,
    "batch_idx": 128,
    "loss": 0.19630900025367737,
    "learning_rate": 5.551642202006324e-09,
    "timestep_mean": 448.21875,
    "epsilon_norm": 8.331008911132812,
    "prediction_norm": 7.669151306152344,
    "timestamp": "2025-07-14T12:34:45.990918"
  },
  {
    "epoch": 2003,
    "batch_idx": 160,
    "loss": 0.3251505494117737,
    "learning_rate": 5.551642202006324e-09,
    "timestep_mean": 482.0625,
    "epsilon_norm": 7.343771934509277,
    "prediction_norm": 5.937617778778076,
    "timestamp": "2025-07-14T12:34:45.996115"
  },
  {
    "epoch": 2003,
    "batch_idx": 192,
    "loss": 0.2453218698501587,
    "learning_rate": 5.551642202006324e-09,
    "timestep_mean": 389.375,
    "epsilon_norm": 3.4226877689361572,
    "prediction_norm": 2.49222469329834,
    "timestamp": "2025-07-14T12:34:46.001259"
  },
  {
    "epoch": 2004,
    "batch_idx": 0,
    "loss": 0.31669026613235474,
    "learning_rate": 9.869571931442334e-09,
    "timestep_mean": 520.84375,
    "epsilon_norm": 8.788719177246094,
    "prediction_norm": 7.1822943687438965,
    "timestamp": "2025-07-14T12:34:46.006490"
  },
  {
    "epoch": 2004,
    "batch_idx": 32,
    "loss": 0.288042813539505,
    "learning_rate": 9.869571931442334e-09,
    "timestep_mean": 393.90625,
    "epsilon_norm": 8.23027515411377,
    "prediction_norm": 7.254272937774658,
    "timestamp": "2025-07-14T12:34:46.011699"
  },
  {
    "epoch": 2004,
    "batch_idx": 64,
    "loss": 0.20375841856002808,
    "learning_rate": 9.869571931442334e-09,
    "timestep_mean": 493.5,
    "epsilon_norm": 8.321813583374023,
    "prediction_norm": 6.947993278503418,
    "timestamp": "2025-07-14T12:34:46.016927"
  },
  {
    "epoch": 2004,
    "batch_idx": 96,
    "loss": 0.187126025557518,
    "learning_rate": 9.869571931442334e-09,
    "timestep_mean": 432.5625,
    "epsilon_norm": 7.342583656311035,
    "prediction_norm": 6.576550006866455,
    "timestamp": "2025-07-14T12:34:46.022152"
  },
  {
    "epoch": 2004,
    "batch_idx": 128,
    "loss": 0.4123821556568146,
    "learning_rate": 9.869571931442334e-09,
    "timestep_mean": 537.6875,
    "epsilon_norm": 8.088186264038086,
    "prediction_norm": 6.289390563964844,
    "timestamp": "2025-07-14T12:34:46.027394"
  },
  {
    "epoch": 2004,
    "batch_idx": 160,
    "loss": 0.18865028023719788,
    "learning_rate": 9.869571931442334e-09,
    "timestep_mean": 531.78125,
    "epsilon_norm": 7.873101711273193,
    "prediction_norm": 7.041999340057373,
    "timestamp": "2025-07-14T12:34:46.032474"
  },
  {
    "epoch": 2004,
    "batch_idx": 192,
    "loss": 0.3614193797111511,
    "learning_rate": 9.869571931442334e-09,
    "timestep_mean": 332.125,
    "epsilon_norm": 3.9120373725891113,
    "prediction_norm": 3.4264304637908936,
    "timestamp": "2025-07-14T12:34:46.037681"
  },
  {
    "epoch": 2005,
    "batch_idx": 0,
    "loss": 0.29926350712776184,
    "learning_rate": 1.5421177605168258e-08,
    "timestep_mean": 492.15625,
    "epsilon_norm": 7.608362197875977,
    "prediction_norm": 6.725562572479248,
    "timestamp": "2025-07-14T12:34:46.042864"
  },
  {
    "epoch": 2005,
    "batch_idx": 32,
    "loss": 0.19140729308128357,
    "learning_rate": 1.5421177605168258e-08,
    "timestep_mean": 520.5625,
    "epsilon_norm": 7.846795558929443,
    "prediction_norm": 7.110672473907471,
    "timestamp": "2025-07-14T12:34:46.048014"
  },
  {
    "epoch": 2005,
    "batch_idx": 64,
    "loss": 0.37192434072494507,
    "learning_rate": 1.5421177605168258e-08,
    "timestep_mean": 492.46875,
    "epsilon_norm": 7.131235599517822,
    "prediction_norm": 5.711572170257568,
    "timestamp": "2025-07-14T12:34:46.053281"
  },
  {
    "epoch": 2005,
    "batch_idx": 96,
    "loss": 0.4422270953655243,
    "learning_rate": 1.5421177605168258e-08,
    "timestep_mean": 431.09375,
    "epsilon_norm": 8.387682914733887,
    "prediction_norm": 6.654860496520996,
    "timestamp": "2025-07-14T12:34:46.058518"
  },
  {
    "epoch": 2005,
    "batch_idx": 128,
    "loss": 0.15398366749286652,
    "learning_rate": 1.5421177605168258e-08,
    "timestep_mean": 546.40625,
    "epsilon_norm": 7.328059673309326,
    "prediction_norm": 6.297025203704834,
    "timestamp": "2025-07-14T12:34:46.063783"
  },
  {
    "epoch": 2005,
    "batch_idx": 160,
    "loss": 0.26205307245254517,
    "learning_rate": 1.5421177605168258e-08,
    "timestep_mean": 514.65625,
    "epsilon_norm": 8.42595386505127,
    "prediction_norm": 7.492433547973633,
    "timestamp": "2025-07-14T12:34:46.069007"
  },
  {
    "epoch": 2005,
    "batch_idx": 192,
    "loss": 0.23848839104175568,
    "learning_rate": 1.5421177605168258e-08,
    "timestep_mean": 473.625,
    "epsilon_norm": 3.1651611328125,
    "prediction_norm": 2.785888433456421,
    "timestamp": "2025-07-14T12:34:46.074115"
  },
  {
    "epoch": 2006,
    "batch_idx": 0,
    "loss": 0.16644690930843353,
    "learning_rate": 2.2206445525085882e-08,
    "timestep_mean": 451.0,
    "epsilon_norm": 6.926459789276123,
    "prediction_norm": 6.107026100158691,
    "timestamp": "2025-07-14T12:34:46.079284"
  },
  {
    "epoch": 2006,
    "batch_idx": 32,
    "loss": 0.14714281260967255,
    "learning_rate": 2.2206445525085882e-08,
    "timestep_mean": 486.34375,
    "epsilon_norm": 7.110901355743408,
    "prediction_norm": 6.665228366851807,
    "timestamp": "2025-07-14T12:34:46.084461"
  },
  {
    "epoch": 2006,
    "batch_idx": 64,
    "loss": 0.40635913610458374,
    "learning_rate": 2.2206445525085882e-08,
    "timestep_mean": 360.5,
    "epsilon_norm": 8.962148666381836,
    "prediction_norm": 7.153553009033203,
    "timestamp": "2025-07-14T12:34:46.089573"
  },
  {
    "epoch": 2006,
    "batch_idx": 96,
    "loss": 0.22314511239528656,
    "learning_rate": 2.2206445525085882e-08,
    "timestep_mean": 436.25,
    "epsilon_norm": 6.896268367767334,
    "prediction_norm": 5.571208477020264,
    "timestamp": "2025-07-14T12:34:46.094794"
  },
  {
    "epoch": 2006,
    "batch_idx": 128,
    "loss": 0.24248939752578735,
    "learning_rate": 2.2206445525085882e-08,
    "timestep_mean": 499.90625,
    "epsilon_norm": 7.326702117919922,
    "prediction_norm": 6.1228413581848145,
    "timestamp": "2025-07-14T12:34:46.100016"
  },
  {
    "epoch": 2006,
    "batch_idx": 160,
    "loss": 0.2112959772348404,
    "learning_rate": 2.2206445525085882e-08,
    "timestep_mean": 543.78125,
    "epsilon_norm": 8.564626693725586,
    "prediction_norm": 7.2083587646484375,
    "timestamp": "2025-07-14T12:34:46.105302"
  },
  {
    "epoch": 2006,
    "batch_idx": 192,
    "loss": 0.08726753294467926,
    "learning_rate": 2.2206445525085882e-08,
    "timestep_mean": 492.375,
    "epsilon_norm": 3.7939579486846924,
    "prediction_norm": 3.1765925884246826,
    "timestamp": "2025-07-14T12:34:46.110428"
  },
  {
    "epoch": 2007,
    "batch_idx": 0,
    "loss": 0.28910207748413086,
    "learning_rate": 3.022535894930956e-08,
    "timestep_mean": 444.0,
    "epsilon_norm": 7.944705009460449,
    "prediction_norm": 6.796820163726807,
    "timestamp": "2025-07-14T12:34:46.115584"
  },
  {
    "epoch": 2007,
    "batch_idx": 32,
    "loss": 0.45702898502349854,
    "learning_rate": 3.022535894930956e-08,
    "timestep_mean": 468.875,
    "epsilon_norm": 7.868221282958984,
    "prediction_norm": 6.243106365203857,
    "timestamp": "2025-07-14T12:34:46.120778"
  },
  {
    "epoch": 2007,
    "batch_idx": 64,
    "loss": 0.22695405781269073,
    "learning_rate": 3.022535894930956e-08,
    "timestep_mean": 522.0,
    "epsilon_norm": 8.966978073120117,
    "prediction_norm": 8.31475830078125,
    "timestamp": "2025-07-14T12:34:46.125940"
  },
  {
    "epoch": 2007,
    "batch_idx": 96,
    "loss": 0.37194517254829407,
    "learning_rate": 3.022535894930956e-08,
    "timestep_mean": 475.5625,
    "epsilon_norm": 9.43476390838623,
    "prediction_norm": 8.187723159790039,
    "timestamp": "2025-07-14T12:34:46.131012"
  },
  {
    "epoch": 2007,
    "batch_idx": 128,
    "loss": 0.17261016368865967,
    "learning_rate": 3.022535894930956e-08,
    "timestep_mean": 591.3125,
    "epsilon_norm": 7.670202255249023,
    "prediction_norm": 7.290256023406982,
    "timestamp": "2025-07-14T12:34:46.136168"
  },
  {
    "epoch": 2007,
    "batch_idx": 160,
    "loss": 0.21441063284873962,
    "learning_rate": 3.022535894930956e-08,
    "timestep_mean": 514.6875,
    "epsilon_norm": 7.91060209274292,
    "prediction_norm": 7.324102401733398,
    "timestamp": "2025-07-14T12:34:46.141395"
  },
  {
    "epoch": 2007,
    "batch_idx": 192,
    "loss": 0.3416712284088135,
    "learning_rate": 3.022535894930956e-08,
    "timestep_mean": 360.875,
    "epsilon_norm": 3.404940605163574,
    "prediction_norm": 2.706603765487671,
    "timestamp": "2025-07-14T12:34:46.146666"
  },
  {
    "epoch": 2008,
    "batch_idx": 0,
    "loss": 0.42223504185676575,
    "learning_rate": 3.9477898091944134e-08,
    "timestep_mean": 503.96875,
    "epsilon_norm": 8.567855834960938,
    "prediction_norm": 7.148982524871826,
    "timestamp": "2025-07-14T12:34:46.151624"
  },
  {
    "epoch": 2008,
    "batch_idx": 32,
    "loss": 0.3304033875465393,
    "learning_rate": 3.9477898091944134e-08,
    "timestep_mean": 460.78125,
    "epsilon_norm": 8.607633590698242,
    "prediction_norm": 6.79217004776001,
    "timestamp": "2025-07-14T12:34:46.156606"
  },
  {
    "epoch": 2008,
    "batch_idx": 64,
    "loss": 0.1372540146112442,
    "learning_rate": 3.9477898091944134e-08,
    "timestep_mean": 531.3125,
    "epsilon_norm": 6.379785060882568,
    "prediction_norm": 5.52128267288208,
    "timestamp": "2025-07-14T12:34:46.161753"
  },
  {
    "epoch": 2008,
    "batch_idx": 96,
    "loss": 0.3450945019721985,
    "learning_rate": 3.9477898091944134e-08,
    "timestep_mean": 550.8125,
    "epsilon_norm": 7.824005603790283,
    "prediction_norm": 6.19787073135376,
    "timestamp": "2025-07-14T12:34:46.167079"
  },
  {
    "epoch": 2008,
    "batch_idx": 128,
    "loss": 0.34165728092193604,
    "learning_rate": 3.9477898091944134e-08,
    "timestep_mean": 556.59375,
    "epsilon_norm": 8.840331077575684,
    "prediction_norm": 7.064556121826172,
    "timestamp": "2025-07-14T12:34:46.172158"
  },
  {
    "epoch": 2008,
    "batch_idx": 160,
    "loss": 0.17246472835540771,
    "learning_rate": 3.9477898091944134e-08,
    "timestep_mean": 565.90625,
    "epsilon_norm": 7.426294326782227,
    "prediction_norm": 6.6786932945251465,
    "timestamp": "2025-07-14T12:34:46.177280"
  },
  {
    "epoch": 2008,
    "batch_idx": 192,
    "loss": 0.08041498064994812,
    "learning_rate": 3.9477898091944134e-08,
    "timestep_mean": 642.375,
    "epsilon_norm": 3.662400484085083,
    "prediction_norm": 3.744809627532959,
    "timestamp": "2025-07-14T12:34:46.182285"
  },
  {
    "epoch": 2009,
    "batch_idx": 0,
    "loss": 0.245059534907341,
    "learning_rate": 4.9964040123251515e-08,
    "timestep_mean": 499.90625,
    "epsilon_norm": 7.751409530639648,
    "prediction_norm": 6.809933185577393,
    "timestamp": "2025-07-14T12:34:46.187470"
  },
  {
    "epoch": 2009,
    "batch_idx": 32,
    "loss": 0.21919900178909302,
    "learning_rate": 4.9964040123251515e-08,
    "timestep_mean": 404.4375,
    "epsilon_norm": 8.320034980773926,
    "prediction_norm": 6.585004806518555,
    "timestamp": "2025-07-14T12:34:46.192818"
  },
  {
    "epoch": 2009,
    "batch_idx": 64,
    "loss": 0.19393326342105865,
    "learning_rate": 4.9964040123251515e-08,
    "timestep_mean": 498.71875,
    "epsilon_norm": 8.122233390808105,
    "prediction_norm": 6.862265110015869,
    "timestamp": "2025-07-14T12:34:46.197931"
  },
  {
    "epoch": 2009,
    "batch_idx": 96,
    "loss": 0.29845771193504333,
    "learning_rate": 4.9964040123251515e-08,
    "timestep_mean": 413.5625,
    "epsilon_norm": 7.998418807983398,
    "prediction_norm": 7.479151725769043,
    "timestamp": "2025-07-14T12:34:46.203066"
  },
  {
    "epoch": 2009,
    "batch_idx": 128,
    "loss": 0.13513919711112976,
    "learning_rate": 4.9964040123251515e-08,
    "timestep_mean": 572.40625,
    "epsilon_norm": 8.799324035644531,
    "prediction_norm": 7.808647155761719,
    "timestamp": "2025-07-14T12:34:46.208086"
  },
  {
    "epoch": 2009,
    "batch_idx": 160,
    "loss": 0.2153072953224182,
    "learning_rate": 4.9964040123251515e-08,
    "timestep_mean": 481.09375,
    "epsilon_norm": 8.460260391235352,
    "prediction_norm": 7.2959818840026855,
    "timestamp": "2025-07-14T12:34:46.213441"
  },
  {
    "epoch": 2009,
    "batch_idx": 192,
    "loss": 0.21897678077220917,
    "learning_rate": 4.9964040123251515e-08,
    "timestep_mean": 670.0,
    "epsilon_norm": 4.427073955535889,
    "prediction_norm": 4.2657976150512695,
    "timestamp": "2025-07-14T12:34:46.218563"
  },
  {
    "epoch": 2010,
    "batch_idx": 0,
    "loss": 0.5258599519729614,
    "learning_rate": 6.168375916970616e-08,
    "timestep_mean": 443.53125,
    "epsilon_norm": 7.738814830780029,
    "prediction_norm": 5.532378196716309,
    "timestamp": "2025-07-14T12:34:46.223741"
  },
  {
    "epoch": 2010,
    "batch_idx": 32,
    "loss": 0.22386379539966583,
    "learning_rate": 6.168375916970616e-08,
    "timestep_mean": 388.40625,
    "epsilon_norm": 7.843871116638184,
    "prediction_norm": 6.513467311859131,
    "timestamp": "2025-07-14T12:34:46.228902"
  },
  {
    "epoch": 2010,
    "batch_idx": 64,
    "loss": 0.4167560935020447,
    "learning_rate": 6.168375916970616e-08,
    "timestep_mean": 456.09375,
    "epsilon_norm": 7.802753448486328,
    "prediction_norm": 6.099620819091797,
    "timestamp": "2025-07-14T12:34:46.234119"
  },
  {
    "epoch": 2010,
    "batch_idx": 96,
    "loss": 0.12856179475784302,
    "learning_rate": 6.168375916970616e-08,
    "timestep_mean": 608.09375,
    "epsilon_norm": 8.034472465515137,
    "prediction_norm": 7.6627278327941895,
    "timestamp": "2025-07-14T12:34:46.239074"
  },
  {
    "epoch": 2010,
    "batch_idx": 128,
    "loss": 0.19204331934452057,
    "learning_rate": 6.168375916970616e-08,
    "timestep_mean": 544.3125,
    "epsilon_norm": 7.546696186065674,
    "prediction_norm": 6.7312188148498535,
    "timestamp": "2025-07-14T12:34:46.243947"
  },
  {
    "epoch": 2010,
    "batch_idx": 160,
    "loss": 0.2933531105518341,
    "learning_rate": 6.168375916970616e-08,
    "timestep_mean": 511.28125,
    "epsilon_norm": 7.298732757568359,
    "prediction_norm": 5.935588836669922,
    "timestamp": "2025-07-14T12:34:46.248793"
  },
  {
    "epoch": 2010,
    "batch_idx": 192,
    "loss": 0.32667797803878784,
    "learning_rate": 6.168375916970616e-08,
    "timestep_mean": 536.0,
    "epsilon_norm": 4.544233798980713,
    "prediction_norm": 3.598393201828003,
    "timestamp": "2025-07-14T12:34:46.253667"
  },
  {
    "epoch": 2011,
    "batch_idx": 0,
    "loss": 0.3172166645526886,
    "learning_rate": 7.463702631410609e-08,
    "timestep_mean": 514.875,
    "epsilon_norm": 8.522716522216797,
    "prediction_norm": 7.085228443145752,
    "timestamp": "2025-07-14T12:34:46.258658"
  },
  {
    "epoch": 2011,
    "batch_idx": 32,
    "loss": 0.26639407873153687,
    "learning_rate": 7.463702631410609e-08,
    "timestep_mean": 513.6875,
    "epsilon_norm": 7.522223949432373,
    "prediction_norm": 6.172688007354736,
    "timestamp": "2025-07-14T12:34:46.263525"
  },
  {
    "epoch": 2011,
    "batch_idx": 64,
    "loss": 0.195913165807724,
    "learning_rate": 7.463702631410609e-08,
    "timestep_mean": 409.8125,
    "epsilon_norm": 6.598372936248779,
    "prediction_norm": 5.189543724060059,
    "timestamp": "2025-07-14T12:34:46.268369"
  },
  {
    "epoch": 2011,
    "batch_idx": 96,
    "loss": 0.1642291247844696,
    "learning_rate": 7.463702631410609e-08,
    "timestep_mean": 462.90625,
    "epsilon_norm": 6.106305122375488,
    "prediction_norm": 5.423386573791504,
    "timestamp": "2025-07-14T12:34:46.273236"
  },
  {
    "epoch": 2011,
    "batch_idx": 128,
    "loss": 0.32029789686203003,
    "learning_rate": 7.463702631410609e-08,
    "timestep_mean": 562.78125,
    "epsilon_norm": 9.569716453552246,
    "prediction_norm": 7.838550567626953,
    "timestamp": "2025-07-14T12:34:46.278116"
  },
  {
    "epoch": 2011,
    "batch_idx": 160,
    "loss": 0.530409574508667,
    "learning_rate": 7.463702631410609e-08,
    "timestep_mean": 454.5,
    "epsilon_norm": 9.285228729248047,
    "prediction_norm": 6.568858623504639,
    "timestamp": "2025-07-14T12:34:46.282956"
  },
  {
    "epoch": 2011,
    "batch_idx": 192,
    "loss": 0.41833630204200745,
    "learning_rate": 7.463702631410609e-08,
    "timestep_mean": 489.75,
    "epsilon_norm": 4.923316478729248,
    "prediction_norm": 3.884350538253784,
    "timestamp": "2025-07-14T12:34:46.287912"
  },
  {
    "epoch": 2012,
    "batch_idx": 0,
    "loss": 0.1712651550769806,
    "learning_rate": 8.88238095955174e-08,
    "timestep_mean": 451.09375,
    "epsilon_norm": 8.646500587463379,
    "prediction_norm": 7.8320536613464355,
    "timestamp": "2025-07-14T12:34:46.292813"
  },
  {
    "epoch": 2012,
    "batch_idx": 32,
    "loss": 0.0936453565955162,
    "learning_rate": 8.88238095955174e-08,
    "timestep_mean": 551.5,
    "epsilon_norm": 7.027711391448975,
    "prediction_norm": 6.458600997924805,
    "timestamp": "2025-07-14T12:34:46.297738"
  },
  {
    "epoch": 2012,
    "batch_idx": 64,
    "loss": 0.1467953324317932,
    "learning_rate": 8.88238095955174e-08,
    "timestep_mean": 448.875,
    "epsilon_norm": 6.906147480010986,
    "prediction_norm": 6.025717735290527,
    "timestamp": "2025-07-14T12:34:46.302569"
  },
  {
    "epoch": 2012,
    "batch_idx": 96,
    "loss": 0.23916593194007874,
    "learning_rate": 8.88238095955174e-08,
    "timestep_mean": 525.46875,
    "epsilon_norm": 6.886447429656982,
    "prediction_norm": 5.416945457458496,
    "timestamp": "2025-07-14T12:34:46.307480"
  },
  {
    "epoch": 2012,
    "batch_idx": 128,
    "loss": 0.3362681269645691,
    "learning_rate": 8.88238095955174e-08,
    "timestep_mean": 502.84375,
    "epsilon_norm": 9.021805763244629,
    "prediction_norm": 7.917289733886719,
    "timestamp": "2025-07-14T12:34:46.312409"
  },
  {
    "epoch": 2012,
    "batch_idx": 160,
    "loss": 0.15021660923957825,
    "learning_rate": 8.88238095955174e-08,
    "timestep_mean": 453.3125,
    "epsilon_norm": 6.733951091766357,
    "prediction_norm": 6.491137504577637,
    "timestamp": "2025-07-14T12:34:46.317420"
  },
  {
    "epoch": 2012,
    "batch_idx": 192,
    "loss": 0.07463449984788895,
    "learning_rate": 8.88238095955174e-08,
    "timestep_mean": 589.125,
    "epsilon_norm": 4.171299934387207,
    "prediction_norm": 4.474576473236084,
    "timestamp": "2025-07-14T12:34:46.322171"
  },
  {
    "epoch": 2013,
    "batch_idx": 0,
    "loss": 0.3151649236679077,
    "learning_rate": 1.0424407400944081e-07,
    "timestep_mean": 516.40625,
    "epsilon_norm": 8.116289138793945,
    "prediction_norm": 6.704514503479004,
    "timestamp": "2025-07-14T12:34:46.326933"
  },
  {
    "epoch": 2013,
    "batch_idx": 32,
    "loss": 0.23868893086910248,
    "learning_rate": 1.0424407400944081e-07,
    "timestep_mean": 465.3125,
    "epsilon_norm": 8.667113304138184,
    "prediction_norm": 7.4235429763793945,
    "timestamp": "2025-07-14T12:34:46.331708"
  },
  {
    "epoch": 2013,
    "batch_idx": 64,
    "loss": 0.3343825936317444,
    "learning_rate": 1.0424407400944081e-07,
    "timestep_mean": 428.21875,
    "epsilon_norm": 7.424264907836914,
    "prediction_norm": 5.684444904327393,
    "timestamp": "2025-07-14T12:34:46.336532"
  },
  {
    "epoch": 2013,
    "batch_idx": 96,
    "loss": 0.20028352737426758,
    "learning_rate": 1.0424407400944081e-07,
    "timestep_mean": 509.40625,
    "epsilon_norm": 7.744535446166992,
    "prediction_norm": 7.354317665100098,
    "timestamp": "2025-07-14T12:34:46.341475"
  },
  {
    "epoch": 2013,
    "batch_idx": 128,
    "loss": 0.19194182753562927,
    "learning_rate": 1.0424407400944081e-07,
    "timestep_mean": 528.21875,
    "epsilon_norm": 8.259170532226562,
    "prediction_norm": 7.269300937652588,
    "timestamp": "2025-07-14T12:34:46.346453"
  },
  {
    "epoch": 2013,
    "batch_idx": 160,
    "loss": 0.2657327353954315,
    "learning_rate": 1.0424407400944081e-07,
    "timestep_mean": 533.9375,
    "epsilon_norm": 7.951103687286377,
    "prediction_norm": 6.775646686553955,
    "timestamp": "2025-07-14T12:34:46.351268"
  },
  {
    "epoch": 2013,
    "batch_idx": 192,
    "loss": 0.49770253896713257,
    "learning_rate": 1.0424407400944081e-07,
    "timestep_mean": 556.375,
    "epsilon_norm": 3.4521701335906982,
    "prediction_norm": 3.2789266109466553,
    "timestamp": "2025-07-14T12:34:46.356052"
  },
  {
    "epoch": 2014,
    "batch_idx": 0,
    "loss": 0.3944655656814575,
    "learning_rate": 1.2089778150797815e-07,
    "timestep_mean": 504.40625,
    "epsilon_norm": 8.90658187866211,
    "prediction_norm": 6.619839668273926,
    "timestamp": "2025-07-14T12:34:46.360903"
  },
  {
    "epoch": 2014,
    "batch_idx": 32,
    "loss": 0.3264024257659912,
    "learning_rate": 1.2089778150797815e-07,
    "timestep_mean": 442.25,
    "epsilon_norm": 7.492701530456543,
    "prediction_norm": 6.260842800140381,
    "timestamp": "2025-07-14T12:34:46.365816"
  },
  {
    "epoch": 2014,
    "batch_idx": 64,
    "loss": 0.27693265676498413,
    "learning_rate": 1.2089778150797815e-07,
    "timestep_mean": 554.75,
    "epsilon_norm": 6.860639572143555,
    "prediction_norm": 5.633775234222412,
    "timestamp": "2025-07-14T12:34:46.370659"
  },
  {
    "epoch": 2014,
    "batch_idx": 96,
    "loss": 0.3451782763004303,
    "learning_rate": 1.2089778150797815e-07,
    "timestep_mean": 446.84375,
    "epsilon_norm": 8.298158645629883,
    "prediction_norm": 6.541141510009766,
    "timestamp": "2025-07-14T12:34:46.375612"
  },
  {
    "epoch": 2014,
    "batch_idx": 128,
    "loss": 0.19289474189281464,
    "learning_rate": 1.2089778150797815e-07,
    "timestep_mean": 545.15625,
    "epsilon_norm": 7.322609901428223,
    "prediction_norm": 6.225387096405029,
    "timestamp": "2025-07-14T12:34:46.380481"
  },
  {
    "epoch": 2014,
    "batch_idx": 160,
    "loss": 0.2412550300359726,
    "learning_rate": 1.2089778150797815e-07,
    "timestep_mean": 591.5,
    "epsilon_norm": 8.20880126953125,
    "prediction_norm": 7.145985126495361,
    "timestamp": "2025-07-14T12:34:46.385321"
  },
  {
    "epoch": 2014,
    "batch_idx": 192,
    "loss": 0.08370842784643173,
    "learning_rate": 1.2089778150797815e-07,
    "timestep_mean": 565.875,
    "epsilon_norm": 3.103701591491699,
    "prediction_norm": 2.6562440395355225,
    "timestamp": "2025-07-14T12:34:46.390168"
  },
  {
    "epoch": 2015,
    "batch_idx": 0,
    "loss": 0.19362854957580566,
    "learning_rate": 1.3878489099972136e-07,
    "timestep_mean": 503.8125,
    "epsilon_norm": 7.8377156257629395,
    "prediction_norm": 6.581485271453857,
    "timestamp": "2025-07-14T12:34:46.395063"
  },
  {
    "epoch": 2015,
    "batch_idx": 32,
    "loss": 0.31318360567092896,
    "learning_rate": 1.3878489099972136e-07,
    "timestep_mean": 435.375,
    "epsilon_norm": 6.977900505065918,
    "prediction_norm": 5.737476825714111,
    "timestamp": "2025-07-14T12:34:46.399953"
  },
  {
    "epoch": 2015,
    "batch_idx": 64,
    "loss": 0.44707614183425903,
    "learning_rate": 1.3878489099972136e-07,
    "timestep_mean": 473.375,
    "epsilon_norm": 9.454689025878906,
    "prediction_norm": 7.568263053894043,
    "timestamp": "2025-07-14T12:34:46.404880"
  },
  {
    "epoch": 2015,
    "batch_idx": 96,
    "loss": 0.2634391486644745,
    "learning_rate": 1.3878489099972136e-07,
    "timestep_mean": 449.4375,
    "epsilon_norm": 7.469918727874756,
    "prediction_norm": 6.636319637298584,
    "timestamp": "2025-07-14T12:34:46.409735"
  },
  {
    "epoch": 2015,
    "batch_idx": 128,
    "loss": 0.24468320608139038,
    "learning_rate": 1.3878489099972136e-07,
    "timestep_mean": 517.125,
    "epsilon_norm": 8.903383255004883,
    "prediction_norm": 7.666725158691406,
    "timestamp": "2025-07-14T12:34:46.414656"
  },
  {
    "epoch": 2015,
    "batch_idx": 160,
    "loss": 0.4160512685775757,
    "learning_rate": 1.3878489099972136e-07,
    "timestep_mean": 449.53125,
    "epsilon_norm": 8.08244800567627,
    "prediction_norm": 6.322448253631592,
    "timestamp": "2025-07-14T12:34:46.419701"
  },
  {
    "epoch": 2015,
    "batch_idx": 192,
    "loss": 0.10531340539455414,
    "learning_rate": 1.3878489099972136e-07,
    "timestep_mean": 470.25,
    "epsilon_norm": 4.198437690734863,
    "prediction_norm": 3.611999988555908,
    "timestamp": "2025-07-14T12:34:46.424585"
  },
  {
    "epoch": 2016,
    "batch_idx": 0,
    "loss": 0.1684350073337555,
    "learning_rate": 1.5790535835003008e-07,
    "timestep_mean": 508.3125,
    "epsilon_norm": 8.199901580810547,
    "prediction_norm": 7.186168670654297,
    "timestamp": "2025-07-14T12:34:46.429589"
  },
  {
    "epoch": 2016,
    "batch_idx": 32,
    "loss": 0.2537592947483063,
    "learning_rate": 1.5790535835003008e-07,
    "timestep_mean": 530.8125,
    "epsilon_norm": 8.786694526672363,
    "prediction_norm": 7.857229232788086,
    "timestamp": "2025-07-14T12:34:46.434502"
  },
  {
    "epoch": 2016,
    "batch_idx": 64,
    "loss": 0.20883625745773315,
    "learning_rate": 1.5790535835003008e-07,
    "timestep_mean": 599.75,
    "epsilon_norm": 8.361339569091797,
    "prediction_norm": 7.650665760040283,
    "timestamp": "2025-07-14T12:34:46.439359"
  },
  {
    "epoch": 2016,
    "batch_idx": 96,
    "loss": 0.21508538722991943,
    "learning_rate": 1.5790535835003008e-07,
    "timestep_mean": 502.5,
    "epsilon_norm": 8.091614723205566,
    "prediction_norm": 7.315596580505371,
    "timestamp": "2025-07-14T12:34:46.444399"
  },
  {
    "epoch": 2016,
    "batch_idx": 128,
    "loss": 0.4534316956996918,
    "learning_rate": 1.5790535835003008e-07,
    "timestep_mean": 353.75,
    "epsilon_norm": 7.709116458892822,
    "prediction_norm": 5.6717729568481445,
    "timestamp": "2025-07-14T12:34:46.449381"
  },
  {
    "epoch": 2016,
    "batch_idx": 160,
    "loss": 0.23391816020011902,
    "learning_rate": 1.5790535835003008e-07,
    "timestep_mean": 415.5,
    "epsilon_norm": 7.689250469207764,
    "prediction_norm": 6.281158447265625,
    "timestamp": "2025-07-14T12:34:46.454253"
  },
  {
    "epoch": 2016,
    "batch_idx": 192,
    "loss": 0.03787042200565338,
    "learning_rate": 1.5790535835003008e-07,
    "timestep_mean": 571.75,
    "epsilon_norm": 3.1713366508483887,
    "prediction_norm": 3.0155515670776367,
    "timestamp": "2025-07-14T12:34:46.459085"
  },
  {
    "epoch": 2017,
    "batch_idx": 0,
    "loss": 0.5000044703483582,
    "learning_rate": 1.7825913638097604e-07,
    "timestep_mean": 460.875,
    "epsilon_norm": 8.29145622253418,
    "prediction_norm": 6.7175774574279785,
    "timestamp": "2025-07-14T12:34:46.463987"
  },
  {
    "epoch": 2017,
    "batch_idx": 32,
    "loss": 0.3221428692340851,
    "learning_rate": 1.7825913638097604e-07,
    "timestep_mean": 444.03125,
    "epsilon_norm": 7.594734191894531,
    "prediction_norm": 6.482178688049316,
    "timestamp": "2025-07-14T12:34:46.468844"
  },
  {
    "epoch": 2017,
    "batch_idx": 64,
    "loss": 0.26205798983573914,
    "learning_rate": 1.7825913638097604e-07,
    "timestep_mean": 543.75,
    "epsilon_norm": 9.270085334777832,
    "prediction_norm": 8.127236366271973,
    "timestamp": "2025-07-14T12:34:46.473761"
  },
  {
    "epoch": 2017,
    "batch_idx": 96,
    "loss": 0.2885284125804901,
    "learning_rate": 1.7825913638097604e-07,
    "timestep_mean": 441.59375,
    "epsilon_norm": 8.500472068786621,
    "prediction_norm": 7.0745158195495605,
    "timestamp": "2025-07-14T12:34:46.478688"
  },
  {
    "epoch": 2017,
    "batch_idx": 128,
    "loss": 0.2490006387233734,
    "learning_rate": 1.7825913638097604e-07,
    "timestep_mean": 510.40625,
    "epsilon_norm": 8.073403358459473,
    "prediction_norm": 7.274460792541504,
    "timestamp": "2025-07-14T12:34:46.483686"
  },
  {
    "epoch": 2017,
    "batch_idx": 160,
    "loss": 0.20792430639266968,
    "learning_rate": 1.7825913638097604e-07,
    "timestep_mean": 516.375,
    "epsilon_norm": 7.772258281707764,
    "prediction_norm": 6.8603997230529785,
    "timestamp": "2025-07-14T12:34:46.488536"
  },
  {
    "epoch": 2017,
    "batch_idx": 192,
    "loss": 0.4597451090812683,
    "learning_rate": 1.7825913638097604e-07,
    "timestep_mean": 527.0,
    "epsilon_norm": 3.7364156246185303,
    "prediction_norm": 2.599158763885498,
    "timestamp": "2025-07-14T12:34:46.493471"
  },
  {
    "epoch": 2018,
    "batch_idx": 0,
    "loss": 0.12341408431529999,
    "learning_rate": 1.9984617487173176e-07,
    "timestep_mean": 492.25,
    "epsilon_norm": 7.146207809448242,
    "prediction_norm": 6.78877592086792,
    "timestamp": "2025-07-14T12:34:46.498437"
  },
  {
    "epoch": 2018,
    "batch_idx": 32,
    "loss": 0.1501455008983612,
    "learning_rate": 1.9984617487173176e-07,
    "timestep_mean": 465.5,
    "epsilon_norm": 7.377610683441162,
    "prediction_norm": 6.952678680419922,
    "timestamp": "2025-07-14T12:34:46.503639"
  },
  {
    "epoch": 2018,
    "batch_idx": 64,
    "loss": 0.2807376980781555,
    "learning_rate": 1.9984617487173176e-07,
    "timestep_mean": 523.09375,
    "epsilon_norm": 7.390987873077393,
    "prediction_norm": 6.496817588806152,
    "timestamp": "2025-07-14T12:34:46.508537"
  },
  {
    "epoch": 2018,
    "batch_idx": 96,
    "loss": 0.29189443588256836,
    "learning_rate": 1.9984617487173176e-07,
    "timestep_mean": 505.375,
    "epsilon_norm": 8.312826156616211,
    "prediction_norm": 6.628590106964111,
    "timestamp": "2025-07-14T12:34:46.513440"
  },
  {
    "epoch": 2018,
    "batch_idx": 128,
    "loss": 0.27377036213874817,
    "learning_rate": 1.9984617487173176e-07,
    "timestep_mean": 475.21875,
    "epsilon_norm": 8.989858627319336,
    "prediction_norm": 7.915297031402588,
    "timestamp": "2025-07-14T12:34:46.518352"
  },
  {
    "epoch": 2018,
    "batch_idx": 160,
    "loss": 0.1879594922065735,
    "learning_rate": 1.9984617487173176e-07,
    "timestep_mean": 499.5625,
    "epsilon_norm": 7.188982009887695,
    "prediction_norm": 6.296782970428467,
    "timestamp": "2025-07-14T12:34:46.523289"
  },
  {
    "epoch": 2018,
    "batch_idx": 192,
    "loss": 0.7064070105552673,
    "learning_rate": 1.9984617487173176e-07,
    "timestep_mean": 363.875,
    "epsilon_norm": 3.8761017322540283,
    "prediction_norm": 2.854266405105591,
    "timestamp": "2025-07-14T12:34:46.528102"
  },
  {
    "epoch": 2019,
    "batch_idx": 0,
    "loss": 0.3482476472854614,
    "learning_rate": 2.2266642055834843e-07,
    "timestep_mean": 502.25,
    "epsilon_norm": 8.554475784301758,
    "prediction_norm": 6.552324295043945,
    "timestamp": "2025-07-14T12:34:46.532965"
  },
  {
    "epoch": 2019,
    "batch_idx": 32,
    "loss": 0.1596023440361023,
    "learning_rate": 2.2266642055834843e-07,
    "timestep_mean": 536.5625,
    "epsilon_norm": 7.446468830108643,
    "prediction_norm": 6.869968891143799,
    "timestamp": "2025-07-14T12:34:46.537873"
  },
  {
    "epoch": 2019,
    "batch_idx": 64,
    "loss": 0.14974765479564667,
    "learning_rate": 2.2266642055834843e-07,
    "timestep_mean": 541.21875,
    "epsilon_norm": 8.249391555786133,
    "prediction_norm": 8.050646781921387,
    "timestamp": "2025-07-14T12:34:46.542728"
  },
  {
    "epoch": 2019,
    "batch_idx": 96,
    "loss": 0.2274298071861267,
    "learning_rate": 2.2266642055834843e-07,
    "timestep_mean": 540.5625,
    "epsilon_norm": 7.515963077545166,
    "prediction_norm": 6.518361568450928,
    "timestamp": "2025-07-14T12:34:46.547582"
  },
  {
    "epoch": 2019,
    "batch_idx": 128,
    "loss": 0.4464236795902252,
    "learning_rate": 2.2266642055834843e-07,
    "timestep_mean": 492.53125,
    "epsilon_norm": 9.492081642150879,
    "prediction_norm": 7.5625505447387695,
    "timestamp": "2025-07-14T12:34:46.552429"
  },
  {
    "epoch": 2019,
    "batch_idx": 160,
    "loss": 0.11888018995523453,
    "learning_rate": 2.2266642055834843e-07,
    "timestep_mean": 533.5625,
    "epsilon_norm": 7.992084503173828,
    "prediction_norm": 7.372249126434326,
    "timestamp": "2025-07-14T12:34:46.557337"
  },
  {
    "epoch": 2019,
    "batch_idx": 192,
    "loss": 0.17699271440505981,
    "learning_rate": 2.2266642055834843e-07,
    "timestep_mean": 462.125,
    "epsilon_norm": 3.317779779434204,
    "prediction_norm": 2.6992392539978027,
    "timestamp": "2025-07-14T12:34:46.562183"
  },
  {
    "epoch": 2020,
    "batch_idx": 0,
    "loss": 0.25604259967803955,
    "learning_rate": 2.4671981713420004e-07,
    "timestep_mean": 435.75,
    "epsilon_norm": 8.331062316894531,
    "prediction_norm": 6.726197242736816,
    "timestamp": "2025-07-14T12:34:46.567095"
  },
  {
    "epoch": 2020,
    "batch_idx": 32,
    "loss": 0.2371629774570465,
    "learning_rate": 2.4671981713420004e-07,
    "timestep_mean": 526.34375,
    "epsilon_norm": 8.000751495361328,
    "prediction_norm": 6.910285472869873,
    "timestamp": "2025-07-14T12:34:46.572002"
  },
  {
    "epoch": 2020,
    "batch_idx": 64,
    "loss": 0.22114121913909912,
    "learning_rate": 2.4671981713420004e-07,
    "timestep_mean": 497.21875,
    "epsilon_norm": 6.770461559295654,
    "prediction_norm": 5.584373950958252,
    "timestamp": "2025-07-14T12:34:46.576829"
  },
  {
    "epoch": 2020,
    "batch_idx": 96,
    "loss": 0.3343868851661682,
    "learning_rate": 2.4671981713420004e-07,
    "timestep_mean": 515.15625,
    "epsilon_norm": 7.835562705993652,
    "prediction_norm": 6.312149524688721,
    "timestamp": "2025-07-14T12:34:46.581695"
  },
  {
    "epoch": 2020,
    "batch_idx": 128,
    "loss": 0.26190251111984253,
    "learning_rate": 2.4671981713420004e-07,
    "timestep_mean": 525.53125,
    "epsilon_norm": 8.278112411499023,
    "prediction_norm": 6.83789587020874,
    "timestamp": "2025-07-14T12:34:46.586541"
  },
  {
    "epoch": 2020,
    "batch_idx": 160,
    "loss": 0.3274488151073456,
    "learning_rate": 2.4671981713420004e-07,
    "timestep_mean": 444.96875,
    "epsilon_norm": 7.683647155761719,
    "prediction_norm": 6.3585028648376465,
    "timestamp": "2025-07-14T12:34:46.591428"
  },
  {
    "epoch": 2020,
    "batch_idx": 192,
    "loss": 0.06073343753814697,
    "learning_rate": 2.4671981713420004e-07,
    "timestep_mean": 574.375,
    "epsilon_norm": 3.130342483520508,
    "prediction_norm": 2.988842010498047,
    "timestamp": "2025-07-14T12:34:46.596561"
  },
  {
    "epoch": 2021,
    "batch_idx": 0,
    "loss": 0.18239912390708923,
    "learning_rate": 2.720063052499278e-07,
    "timestep_mean": 504.125,
    "epsilon_norm": 7.370554447174072,
    "prediction_norm": 6.585707664489746,
    "timestamp": "2025-07-14T12:34:46.601576"
  },
  {
    "epoch": 2021,
    "batch_idx": 32,
    "loss": 0.32862886786460876,
    "learning_rate": 2.720063052499278e-07,
    "timestep_mean": 480.875,
    "epsilon_norm": 8.737226486206055,
    "prediction_norm": 7.075932025909424,
    "timestamp": "2025-07-14T12:34:46.606641"
  },
  {
    "epoch": 2021,
    "batch_idx": 64,
    "loss": 0.20828193426132202,
    "learning_rate": 2.720063052499278e-07,
    "timestep_mean": 544.71875,
    "epsilon_norm": 7.731257438659668,
    "prediction_norm": 7.0662384033203125,
    "timestamp": "2025-07-14T12:34:46.611844"
  },
  {
    "epoch": 2021,
    "batch_idx": 96,
    "loss": 0.10631325840950012,
    "learning_rate": 2.720063052499278e-07,
    "timestep_mean": 568.5,
    "epsilon_norm": 8.148513793945312,
    "prediction_norm": 7.398956298828125,
    "timestamp": "2025-07-14T12:34:46.617022"
  },
  {
    "epoch": 2021,
    "batch_idx": 128,
    "loss": 0.2514996826648712,
    "learning_rate": 2.720063052499278e-07,
    "timestep_mean": 454.03125,
    "epsilon_norm": 7.271737098693848,
    "prediction_norm": 6.435746669769287,
    "timestamp": "2025-07-14T12:34:46.622219"
  },
  {
    "epoch": 2021,
    "batch_idx": 160,
    "loss": 0.1621536910533905,
    "learning_rate": 2.720063052499278e-07,
    "timestep_mean": 544.90625,
    "epsilon_norm": 7.429348945617676,
    "prediction_norm": 6.67933464050293,
    "timestamp": "2025-07-14T12:34:46.627404"
  },
  {
    "epoch": 2021,
    "batch_idx": 192,
    "loss": 0.48376303911209106,
    "learning_rate": 2.720063052499278e-07,
    "timestep_mean": 437.125,
    "epsilon_norm": 5.124724388122559,
    "prediction_norm": 3.7574267387390137,
    "timestamp": "2025-07-14T12:34:46.632522"
  },
  {
    "epoch": 2022,
    "batch_idx": 0,
    "loss": 0.34261152148246765,
    "learning_rate": 2.9852582251355126e-07,
    "timestep_mean": 441.96875,
    "epsilon_norm": 9.003351211547852,
    "prediction_norm": 7.43637228012085,
    "timestamp": "2025-07-14T12:34:46.637726"
  },
  {
    "epoch": 2022,
    "batch_idx": 32,
    "loss": 0.19209431111812592,
    "learning_rate": 2.9852582251355126e-07,
    "timestep_mean": 479.21875,
    "epsilon_norm": 7.299510478973389,
    "prediction_norm": 6.159248352050781,
    "timestamp": "2025-07-14T12:34:46.642944"
  },
  {
    "epoch": 2022,
    "batch_idx": 64,
    "loss": 0.2591272294521332,
    "learning_rate": 2.9852582251355126e-07,
    "timestep_mean": 541.75,
    "epsilon_norm": 8.672673225402832,
    "prediction_norm": 7.452365875244141,
    "timestamp": "2025-07-14T12:34:46.648055"
  },
  {
    "epoch": 2022,
    "batch_idx": 96,
    "loss": 0.2503887414932251,
    "learning_rate": 2.9852582251355126e-07,
    "timestep_mean": 366.03125,
    "epsilon_norm": 8.452648162841797,
    "prediction_norm": 7.169561862945557,
    "timestamp": "2025-07-14T12:34:46.653301"
  },
  {
    "epoch": 2022,
    "batch_idx": 128,
    "loss": 0.27135878801345825,
    "learning_rate": 2.9852582251355126e-07,
    "timestep_mean": 511.625,
    "epsilon_norm": 7.431793689727783,
    "prediction_norm": 5.9142327308654785,
    "timestamp": "2025-07-14T12:34:46.658563"
  },
  {
    "epoch": 2022,
    "batch_idx": 160,
    "loss": 0.2902430295944214,
    "learning_rate": 2.9852582251355126e-07,
    "timestep_mean": 437.84375,
    "epsilon_norm": 7.686089038848877,
    "prediction_norm": 7.053631782531738,
    "timestamp": "2025-07-14T12:34:46.663768"
  },
  {
    "epoch": 2022,
    "batch_idx": 192,
    "loss": 0.19928789138793945,
    "learning_rate": 2.9852582251355126e-07,
    "timestep_mean": 613.0,
    "epsilon_norm": 4.156808376312256,
    "prediction_norm": 3.776785135269165,
    "timestamp": "2025-07-14T12:34:46.668742"
  },
  {
    "epoch": 2023,
    "batch_idx": 0,
    "loss": 0.18928200006484985,
    "learning_rate": 3.262783034908568e-07,
    "timestep_mean": 506.09375,
    "epsilon_norm": 7.513065338134766,
    "prediction_norm": 6.966683387756348,
    "timestamp": "2025-07-14T12:34:46.673875"
  },
  {
    "epoch": 2023,
    "batch_idx": 32,
    "loss": 0.2512776255607605,
    "learning_rate": 3.262783034908568e-07,
    "timestep_mean": 479.84375,
    "epsilon_norm": 6.735900402069092,
    "prediction_norm": 5.305049896240234,
    "timestamp": "2025-07-14T12:34:46.679142"
  },
  {
    "epoch": 2023,
    "batch_idx": 64,
    "loss": 0.3518282473087311,
    "learning_rate": 3.262783034908568e-07,
    "timestep_mean": 561.0625,
    "epsilon_norm": 8.780591011047363,
    "prediction_norm": 7.233073711395264,
    "timestamp": "2025-07-14T12:34:46.684477"
  },
  {
    "epoch": 2023,
    "batch_idx": 96,
    "loss": 0.21715012192726135,
    "learning_rate": 3.262783034908568e-07,
    "timestep_mean": 372.65625,
    "epsilon_norm": 6.64548921585083,
    "prediction_norm": 5.658403396606445,
    "timestamp": "2025-07-14T12:34:46.689638"
  },
  {
    "epoch": 2023,
    "batch_idx": 128,
    "loss": 0.34121182560920715,
    "learning_rate": 3.262783034908568e-07,
    "timestep_mean": 531.84375,
    "epsilon_norm": 9.070810317993164,
    "prediction_norm": 7.171696662902832,
    "timestamp": "2025-07-14T12:34:46.694855"
  },
  {
    "epoch": 2023,
    "batch_idx": 160,
    "loss": 0.5141072273254395,
    "learning_rate": 3.262783034908568e-07,
    "timestep_mean": 427.78125,
    "epsilon_norm": 9.270284652709961,
    "prediction_norm": 7.573453426361084,
    "timestamp": "2025-07-14T12:34:46.700164"
  },
  {
    "epoch": 2023,
    "batch_idx": 192,
    "loss": 0.42899012565612793,
    "learning_rate": 3.262783034908568e-07,
    "timestep_mean": 515.375,
    "epsilon_norm": 3.6930007934570312,
    "prediction_norm": 3.031087875366211,
    "timestamp": "2025-07-14T12:34:46.705106"
  },
  {
    "epoch": 2024,
    "batch_idx": 0,
    "loss": 0.35542482137680054,
    "learning_rate": 3.552636797053977e-07,
    "timestep_mean": 453.15625,
    "epsilon_norm": 8.134380340576172,
    "prediction_norm": 6.6902666091918945,
    "timestamp": "2025-07-14T12:34:46.710197"
  },
  {
    "epoch": 2024,
    "batch_idx": 32,
    "loss": 0.14980462193489075,
    "learning_rate": 3.552636797053977e-07,
    "timestep_mean": 541.6875,
    "epsilon_norm": 7.449477195739746,
    "prediction_norm": 7.16795015335083,
    "timestamp": "2025-07-14T12:34:46.715264"
  },
  {
    "epoch": 2024,
    "batch_idx": 64,
    "loss": 0.28451642394065857,
    "learning_rate": 3.552636797053977e-07,
    "timestep_mean": 588.125,
    "epsilon_norm": 7.683496475219727,
    "prediction_norm": 6.315005779266357,
    "timestamp": "2025-07-14T12:34:46.720517"
  },
  {
    "epoch": 2024,
    "batch_idx": 96,
    "loss": 0.35567301511764526,
    "learning_rate": 3.552636797053977e-07,
    "timestep_mean": 423.3125,
    "epsilon_norm": 7.738696575164795,
    "prediction_norm": 6.179398536682129,
    "timestamp": "2025-07-14T12:34:46.725863"
  },
  {
    "epoch": 2024,
    "batch_idx": 128,
    "loss": 0.16252735257148743,
    "learning_rate": 3.552636797053977e-07,
    "timestep_mean": 458.59375,
    "epsilon_norm": 6.9983649253845215,
    "prediction_norm": 6.761768341064453,
    "timestamp": "2025-07-14T12:34:46.731002"
  },
  {
    "epoch": 2024,
    "batch_idx": 160,
    "loss": 0.18398280441761017,
    "learning_rate": 3.552636797053977e-07,
    "timestep_mean": 435.03125,
    "epsilon_norm": 6.058969020843506,
    "prediction_norm": 5.681056022644043,
    "timestamp": "2025-07-14T12:34:46.736177"
  },
  {
    "epoch": 2024,
    "batch_idx": 192,
    "loss": 0.08295989036560059,
    "learning_rate": 3.552636797053977e-07,
    "timestep_mean": 649.25,
    "epsilon_norm": 3.5880582332611084,
    "prediction_norm": 3.382781744003296,
    "timestamp": "2025-07-14T12:34:46.741285"
  },
  {
    "epoch": 2025,
    "batch_idx": 0,
    "loss": 0.19191274046897888,
    "learning_rate": 3.8548187963854957e-07,
    "timestep_mean": 565.84375,
    "epsilon_norm": 7.238293647766113,
    "prediction_norm": 6.265769958496094,
    "timestamp": "2025-07-14T12:34:46.746536"
  },
  {
    "epoch": 2025,
    "batch_idx": 32,
    "loss": 0.2549443244934082,
    "learning_rate": 3.8548187963854957e-07,
    "timestep_mean": 421.90625,
    "epsilon_norm": 7.291606426239014,
    "prediction_norm": 6.183224201202393,
    "timestamp": "2025-07-14T12:34:46.751640"
  },
  {
    "epoch": 2025,
    "batch_idx": 64,
    "loss": 0.17203393578529358,
    "learning_rate": 3.8548187963854957e-07,
    "timestep_mean": 486.4375,
    "epsilon_norm": 8.02564811706543,
    "prediction_norm": 6.801241397857666,
    "timestamp": "2025-07-14T12:34:46.756824"
  },
  {
    "epoch": 2025,
    "batch_idx": 96,
    "loss": 0.3179234266281128,
    "learning_rate": 3.8548187963854957e-07,
    "timestep_mean": 557.625,
    "epsilon_norm": 8.433012962341309,
    "prediction_norm": 7.102629661560059,
    "timestamp": "2025-07-14T12:34:46.762005"
  },
  {
    "epoch": 2025,
    "batch_idx": 128,
    "loss": 0.3389711380004883,
    "learning_rate": 3.8548187963854957e-07,
    "timestep_mean": 433.90625,
    "epsilon_norm": 7.380793571472168,
    "prediction_norm": 6.0327653884887695,
    "timestamp": "2025-07-14T12:34:46.767297"
  },
  {
    "epoch": 2025,
    "batch_idx": 160,
    "loss": 0.14151820540428162,
    "learning_rate": 3.8548187963854957e-07,
    "timestep_mean": 532.21875,
    "epsilon_norm": 7.453377723693848,
    "prediction_norm": 6.939386367797852,
    "timestamp": "2025-07-14T12:34:46.772482"
  },
  {
    "epoch": 2025,
    "batch_idx": 192,
    "loss": 0.052580077201128006,
    "learning_rate": 3.8548187963854957e-07,
    "timestep_mean": 733.375,
    "epsilon_norm": 5.098665714263916,
    "prediction_norm": 4.736508846282959,
    "timestamp": "2025-07-14T12:34:46.777609"
  },
  {
    "epoch": 2026,
    "batch_idx": 0,
    "loss": 0.46182331442832947,
    "learning_rate": 4.169328287299546e-07,
    "timestep_mean": 504.0,
    "epsilon_norm": 8.25676441192627,
    "prediction_norm": 6.701717853546143,
    "timestamp": "2025-07-14T12:34:46.782805"
  },
  {
    "epoch": 2026,
    "batch_idx": 32,
    "loss": 0.2527408003807068,
    "learning_rate": 4.169328287299546e-07,
    "timestep_mean": 452.71875,
    "epsilon_norm": 7.290299892425537,
    "prediction_norm": 6.615365982055664,
    "timestamp": "2025-07-14T12:34:46.788014"
  },
  {
    "epoch": 2026,
    "batch_idx": 64,
    "loss": 0.328747034072876,
    "learning_rate": 4.169328287299546e-07,
    "timestep_mean": 465.5625,
    "epsilon_norm": 8.757124900817871,
    "prediction_norm": 7.112703323364258,
    "timestamp": "2025-07-14T12:34:46.793124"
  },
  {
    "epoch": 2026,
    "batch_idx": 96,
    "loss": 0.23353378474712372,
    "learning_rate": 4.169328287299546e-07,
    "timestep_mean": 480.875,
    "epsilon_norm": 7.530850887298584,
    "prediction_norm": 7.0414018630981445,
    "timestamp": "2025-07-14T12:34:46.798296"
  },
  {
    "epoch": 2026,
    "batch_idx": 128,
    "loss": 0.09162890911102295,
    "learning_rate": 4.169328287299546e-07,
    "timestep_mean": 633.65625,
    "epsilon_norm": 7.767496585845947,
    "prediction_norm": 7.3329854011535645,
    "timestamp": "2025-07-14T12:34:46.803469"
  },
  {
    "epoch": 2026,
    "batch_idx": 160,
    "loss": 0.5240107774734497,
    "learning_rate": 4.169328287299546e-07,
    "timestep_mean": 423.28125,
    "epsilon_norm": 7.651665687561035,
    "prediction_norm": 5.245704174041748,
    "timestamp": "2025-07-14T12:34:46.808780"
  },
  {
    "epoch": 2026,
    "batch_idx": 192,
    "loss": 0.32187771797180176,
    "learning_rate": 4.169328287299546e-07,
    "timestep_mean": 459.25,
    "epsilon_norm": 4.093877792358398,
    "prediction_norm": 3.9513766765594482,
    "timestamp": "2025-07-14T12:34:46.813902"
  },
  {
    "epoch": 2027,
    "batch_idx": 0,
    "loss": 0.1954917311668396,
    "learning_rate": 4.4961644937752127e-07,
    "timestep_mean": 470.78125,
    "epsilon_norm": 8.035746574401855,
    "prediction_norm": 7.275590896606445,
    "timestamp": "2025-07-14T12:34:46.819099"
  },
  {
    "epoch": 2027,
    "batch_idx": 32,
    "loss": 0.06658850610256195,
    "learning_rate": 4.4961644937752127e-07,
    "timestep_mean": 615.65625,
    "epsilon_norm": 8.24778938293457,
    "prediction_norm": 8.194156646728516,
    "timestamp": "2025-07-14T12:34:46.824201"
  },
  {
    "epoch": 2027,
    "batch_idx": 64,
    "loss": 0.22134791314601898,
    "learning_rate": 4.4961644937752127e-07,
    "timestep_mean": 560.625,
    "epsilon_norm": 7.0613694190979,
    "prediction_norm": 6.137813091278076,
    "timestamp": "2025-07-14T12:34:46.829591"
  },
  {
    "epoch": 2027,
    "batch_idx": 96,
    "loss": 0.19923052191734314,
    "learning_rate": 4.4961644937752127e-07,
    "timestep_mean": 474.59375,
    "epsilon_norm": 7.491708278656006,
    "prediction_norm": 6.231307506561279,
    "timestamp": "2025-07-14T12:34:46.834743"
  },
  {
    "epoch": 2027,
    "batch_idx": 128,
    "loss": 0.17378130555152893,
    "learning_rate": 4.4961644937752127e-07,
    "timestep_mean": 510.25,
    "epsilon_norm": 7.612997055053711,
    "prediction_norm": 6.441800594329834,
    "timestamp": "2025-07-14T12:34:46.839845"
  },
  {
    "epoch": 2027,
    "batch_idx": 160,
    "loss": 0.2091430127620697,
    "learning_rate": 4.4961644937752127e-07,
    "timestep_mean": 492.3125,
    "epsilon_norm": 7.450118541717529,
    "prediction_norm": 7.216756343841553,
    "timestamp": "2025-07-14T12:34:46.845122"
  },
  {
    "epoch": 2027,
    "batch_idx": 192,
    "loss": 0.07794556021690369,
    "learning_rate": 4.4961644937752127e-07,
    "timestep_mean": 621.875,
    "epsilon_norm": 3.3763880729675293,
    "prediction_norm": 3.247251272201538,
    "timestamp": "2025-07-14T12:34:46.850420"
  },
  {
    "epoch": 2028,
    "batch_idx": 0,
    "loss": 0.10327087342739105,
    "learning_rate": 4.835326609376467e-07,
    "timestep_mean": 558.53125,
    "epsilon_norm": 7.785438060760498,
    "prediction_norm": 7.427182674407959,
    "timestamp": "2025-07-14T12:34:46.855611"
  },
  {
    "epoch": 2028,
    "batch_idx": 32,
    "loss": 0.24988774955272675,
    "learning_rate": 4.835326609376467e-07,
    "timestep_mean": 466.78125,
    "epsilon_norm": 8.12125015258789,
    "prediction_norm": 7.444438457489014,
    "timestamp": "2025-07-14T12:34:46.860519"
  },
  {
    "epoch": 2028,
    "batch_idx": 64,
    "loss": 0.47670459747314453,
    "learning_rate": 4.835326609376467e-07,
    "timestep_mean": 498.375,
    "epsilon_norm": 8.594585418701172,
    "prediction_norm": 6.964095115661621,
    "timestamp": "2025-07-14T12:34:46.865440"
  },
  {
    "epoch": 2028,
    "batch_idx": 96,
    "loss": 0.21200577914714813,
    "learning_rate": 4.835326609376467e-07,
    "timestep_mean": 506.4375,
    "epsilon_norm": 8.267641067504883,
    "prediction_norm": 7.20803689956665,
    "timestamp": "2025-07-14T12:34:46.870665"
  },
  {
    "epoch": 2028,
    "batch_idx": 128,
    "loss": 0.27382954955101013,
    "learning_rate": 4.835326609376467e-07,
    "timestep_mean": 491.75,
    "epsilon_norm": 8.514138221740723,
    "prediction_norm": 7.252002239227295,
    "timestamp": "2025-07-14T12:34:46.875921"
  },
  {
    "epoch": 2028,
    "batch_idx": 160,
    "loss": 0.2872355580329895,
    "learning_rate": 4.835326609376467e-07,
    "timestep_mean": 504.59375,
    "epsilon_norm": 7.585406303405762,
    "prediction_norm": 6.450750350952148,
    "timestamp": "2025-07-14T12:34:46.881053"
  },
  {
    "epoch": 2028,
    "batch_idx": 192,
    "loss": 0.2948237955570221,
    "learning_rate": 4.835326609376467e-07,
    "timestep_mean": 558.5,
    "epsilon_norm": 4.186769485473633,
    "prediction_norm": 3.2555551528930664,
    "timestamp": "2025-07-14T12:34:46.886185"
  },
  {
    "epoch": 2029,
    "batch_idx": 0,
    "loss": 0.21025046706199646,
    "learning_rate": 5.186813797254387e-07,
    "timestep_mean": 576.75,
    "epsilon_norm": 8.208024024963379,
    "prediction_norm": 7.31353235244751,
    "timestamp": "2025-07-14T12:34:46.891399"
  },
  {
    "epoch": 2029,
    "batch_idx": 32,
    "loss": 0.23268863558769226,
    "learning_rate": 5.186813797254387e-07,
    "timestep_mean": 490.375,
    "epsilon_norm": 7.362235069274902,
    "prediction_norm": 6.277947902679443,
    "timestamp": "2025-07-14T12:34:46.896725"
  },
  {
    "epoch": 2029,
    "batch_idx": 64,
    "loss": 0.3584861755371094,
    "learning_rate": 5.186813797254387e-07,
    "timestep_mean": 502.03125,
    "epsilon_norm": 8.344776153564453,
    "prediction_norm": 7.122801780700684,
    "timestamp": "2025-07-14T12:34:46.901863"
  },
  {
    "epoch": 2029,
    "batch_idx": 96,
    "loss": 0.14310061931610107,
    "learning_rate": 5.186813797254387e-07,
    "timestep_mean": 558.4375,
    "epsilon_norm": 8.530186653137207,
    "prediction_norm": 8.251081466674805,
    "timestamp": "2025-07-14T12:34:46.907002"
  },
  {
    "epoch": 2029,
    "batch_idx": 128,
    "loss": 0.3639858365058899,
    "learning_rate": 5.186813797254387e-07,
    "timestep_mean": 549.375,
    "epsilon_norm": 8.537800788879395,
    "prediction_norm": 6.311082363128662,
    "timestamp": "2025-07-14T12:34:46.912226"
  },
  {
    "epoch": 2029,
    "batch_idx": 160,
    "loss": 0.383339524269104,
    "learning_rate": 5.186813797254387e-07,
    "timestep_mean": 553.5,
    "epsilon_norm": 8.73289680480957,
    "prediction_norm": 6.5602803230285645,
    "timestamp": "2025-07-14T12:34:46.917616"
  },
  {
    "epoch": 2029,
    "batch_idx": 192,
    "loss": 0.20905640721321106,
    "learning_rate": 5.186813797254387e-07,
    "timestep_mean": 531.875,
    "epsilon_norm": 4.791909217834473,
    "prediction_norm": 4.520177364349365,
    "timestamp": "2025-07-14T12:34:46.922695"
  },
  {
    "epoch": 2030,
    "batch_idx": 0,
    "loss": 0.34400880336761475,
    "learning_rate": 5.550625190149929e-07,
    "timestep_mean": 479.125,
    "epsilon_norm": 7.313624858856201,
    "prediction_norm": 5.930300235748291,
    "timestamp": "2025-07-14T12:34:46.927901"
  },
  {
    "epoch": 2030,
    "batch_idx": 32,
    "loss": 0.15407973527908325,
    "learning_rate": 5.550625190149929e-07,
    "timestep_mean": 518.90625,
    "epsilon_norm": 6.772201061248779,
    "prediction_norm": 6.268994331359863,
    "timestamp": "2025-07-14T12:34:46.933077"
  },
  {
    "epoch": 2030,
    "batch_idx": 64,
    "loss": 0.21981385350227356,
    "learning_rate": 5.550625190149929e-07,
    "timestep_mean": 482.90625,
    "epsilon_norm": 7.175991058349609,
    "prediction_norm": 5.848615646362305,
    "timestamp": "2025-07-14T12:34:46.938451"
  },
  {
    "epoch": 2030,
    "batch_idx": 96,
    "loss": 0.2359427511692047,
    "learning_rate": 5.550625190149929e-07,
    "timestep_mean": 531.84375,
    "epsilon_norm": 8.447711944580078,
    "prediction_norm": 7.5499267578125,
    "timestamp": "2025-07-14T12:34:46.943576"
  },
  {
    "epoch": 2030,
    "batch_idx": 128,
    "loss": 0.4200315475463867,
    "learning_rate": 5.550625190149929e-07,
    "timestep_mean": 488.0,
    "epsilon_norm": 8.226883888244629,
    "prediction_norm": 6.057163238525391,
    "timestamp": "2025-07-14T12:34:46.948714"
  },
  {
    "epoch": 2030,
    "batch_idx": 160,
    "loss": 0.32902538776397705,
    "learning_rate": 5.550625190149929e-07,
    "timestep_mean": 447.28125,
    "epsilon_norm": 8.323657035827637,
    "prediction_norm": 6.934315204620361,
    "timestamp": "2025-07-14T12:34:46.953964"
  },
  {
    "epoch": 2030,
    "batch_idx": 192,
    "loss": 0.27813273668289185,
    "learning_rate": 5.550625190149929e-07,
    "timestep_mean": 669.375,
    "epsilon_norm": 3.3846988677978516,
    "prediction_norm": 2.353595495223999,
    "timestamp": "2025-07-14T12:34:46.959215"
  },
  {
    "epoch": 2031,
    "batch_idx": 0,
    "loss": 0.3165355920791626,
    "learning_rate": 5.926759890394486e-07,
    "timestep_mean": 572.25,
    "epsilon_norm": 9.254043579101562,
    "prediction_norm": 8.004460334777832,
    "timestamp": "2025-07-14T12:34:46.964396"
  },
  {
    "epoch": 2031,
    "batch_idx": 32,
    "loss": 0.15017780661582947,
    "learning_rate": 5.926759890394486e-07,
    "timestep_mean": 481.34375,
    "epsilon_norm": 7.682911396026611,
    "prediction_norm": 6.811295032501221,
    "timestamp": "2025-07-14T12:34:46.969616"
  },
  {
    "epoch": 2031,
    "batch_idx": 64,
    "loss": 0.3186652660369873,
    "learning_rate": 5.926759890394486e-07,
    "timestep_mean": 513.9375,
    "epsilon_norm": 8.957829475402832,
    "prediction_norm": 7.508617401123047,
    "timestamp": "2025-07-14T12:34:46.974749"
  },
  {
    "epoch": 2031,
    "batch_idx": 96,
    "loss": 0.1768314242362976,
    "learning_rate": 5.926759890394486e-07,
    "timestep_mean": 497.34375,
    "epsilon_norm": 7.632600784301758,
    "prediction_norm": 6.861003875732422,
    "timestamp": "2025-07-14T12:34:46.980128"
  },
  {
    "epoch": 2031,
    "batch_idx": 128,
    "loss": 0.3250422477722168,
    "learning_rate": 5.926759890394486e-07,
    "timestep_mean": 530.59375,
    "epsilon_norm": 7.207915306091309,
    "prediction_norm": 5.821249961853027,
    "timestamp": "2025-07-14T12:34:46.985253"
  },
  {
    "epoch": 2031,
    "batch_idx": 160,
    "loss": 0.29239508509635925,
    "learning_rate": 5.926759890394486e-07,
    "timestep_mean": 393.59375,
    "epsilon_norm": 7.59469747543335,
    "prediction_norm": 6.5160627365112305,
    "timestamp": "2025-07-14T12:34:46.990338"
  },
  {
    "epoch": 2031,
    "batch_idx": 192,
    "loss": 0.10883773863315582,
    "learning_rate": 5.926759890394486e-07,
    "timestep_mean": 519.125,
    "epsilon_norm": 3.427966594696045,
    "prediction_norm": 3.4435513019561768,
    "timestamp": "2025-07-14T12:34:46.995550"
  },
  {
    "epoch": 2032,
    "batch_idx": 0,
    "loss": 0.1892997920513153,
    "learning_rate": 6.315216969912666e-07,
    "timestep_mean": 437.4375,
    "epsilon_norm": 7.699696063995361,
    "prediction_norm": 6.264547824859619,
    "timestamp": "2025-07-14T12:34:47.000669"
  },
  {
    "epoch": 2032,
    "batch_idx": 32,
    "loss": 0.1449652463197708,
    "learning_rate": 6.315216969912666e-07,
    "timestep_mean": 571.875,
    "epsilon_norm": 7.556628227233887,
    "prediction_norm": 6.566015720367432,
    "timestamp": "2025-07-14T12:34:47.005622"
  },
  {
    "epoch": 2032,
    "batch_idx": 64,
    "loss": 0.3170095682144165,
    "learning_rate": 6.315216969912666e-07,
    "timestep_mean": 539.0625,
    "epsilon_norm": 7.980309009552002,
    "prediction_norm": 7.072690010070801,
    "timestamp": "2025-07-14T12:34:47.010524"
  },
  {
    "epoch": 2032,
    "batch_idx": 96,
    "loss": 0.2660834789276123,
    "learning_rate": 6.315216969912666e-07,
    "timestep_mean": 516.84375,
    "epsilon_norm": 7.719386577606201,
    "prediction_norm": 6.908769130706787,
    "timestamp": "2025-07-14T12:34:47.015404"
  },
  {
    "epoch": 2032,
    "batch_idx": 128,
    "loss": 0.22950692474842072,
    "learning_rate": 6.315216969912666e-07,
    "timestep_mean": 452.6875,
    "epsilon_norm": 8.622346878051758,
    "prediction_norm": 7.328032970428467,
    "timestamp": "2025-07-14T12:34:47.020273"
  },
  {
    "epoch": 2032,
    "batch_idx": 160,
    "loss": 0.23422889411449432,
    "learning_rate": 6.315216969912666e-07,
    "timestep_mean": 503.8125,
    "epsilon_norm": 7.803233623504639,
    "prediction_norm": 6.275025367736816,
    "timestamp": "2025-07-14T12:34:47.025227"
  },
  {
    "epoch": 2032,
    "batch_idx": 192,
    "loss": 0.13106657564640045,
    "learning_rate": 6.315216969912666e-07,
    "timestep_mean": 485.375,
    "epsilon_norm": 4.610132217407227,
    "prediction_norm": 4.288753032684326,
    "timestamp": "2025-07-14T12:34:47.029986"
  },
  {
    "epoch": 2033,
    "batch_idx": 0,
    "loss": 0.30370041728019714,
    "learning_rate": 6.71599547022506e-07,
    "timestep_mean": 524.375,
    "epsilon_norm": 7.058989524841309,
    "prediction_norm": 6.364662170410156,
    "timestamp": "2025-07-14T12:34:47.034760"
  },
  {
    "epoch": 2033,
    "batch_idx": 32,
    "loss": 0.3710334897041321,
    "learning_rate": 6.71599547022506e-07,
    "timestep_mean": 389.46875,
    "epsilon_norm": 8.022357940673828,
    "prediction_norm": 5.8990478515625,
    "timestamp": "2025-07-14T12:34:47.039637"
  },
  {
    "epoch": 2033,
    "batch_idx": 64,
    "loss": 0.22964546084403992,
    "learning_rate": 6.71599547022506e-07,
    "timestep_mean": 440.6875,
    "epsilon_norm": 7.618147373199463,
    "prediction_norm": 6.744268894195557,
    "timestamp": "2025-07-14T12:34:47.044530"
  },
  {
    "epoch": 2033,
    "batch_idx": 96,
    "loss": 0.16709744930267334,
    "learning_rate": 6.71599547022506e-07,
    "timestep_mean": 446.46875,
    "epsilon_norm": 8.610490798950195,
    "prediction_norm": 7.353789806365967,
    "timestamp": "2025-07-14T12:34:47.049374"
  },
  {
    "epoch": 2033,
    "batch_idx": 128,
    "loss": 0.24342189729213715,
    "learning_rate": 6.71599547022506e-07,
    "timestep_mean": 463.21875,
    "epsilon_norm": 7.103903293609619,
    "prediction_norm": 6.0342841148376465,
    "timestamp": "2025-07-14T12:34:47.054184"
  },
  {
    "epoch": 2033,
    "batch_idx": 160,
    "loss": 0.29509156942367554,
    "learning_rate": 6.71599547022506e-07,
    "timestep_mean": 461.65625,
    "epsilon_norm": 6.978394985198975,
    "prediction_norm": 5.2152299880981445,
    "timestamp": "2025-07-14T12:34:47.059133"
  },
  {
    "epoch": 2033,
    "batch_idx": 192,
    "loss": 0.19602251052856445,
    "learning_rate": 6.71599547022506e-07,
    "timestep_mean": 383.5,
    "epsilon_norm": 3.1653716564178467,
    "prediction_norm": 2.9633054733276367,
    "timestamp": "2025-07-14T12:34:47.063964"
  },
  {
    "epoch": 2034,
    "batch_idx": 0,
    "loss": 0.40752291679382324,
    "learning_rate": 7.129094402451025e-07,
    "timestep_mean": 458.28125,
    "epsilon_norm": 8.574443817138672,
    "prediction_norm": 6.352375507354736,
    "timestamp": "2025-07-14T12:34:47.068856"
  },
  {
    "epoch": 2034,
    "batch_idx": 32,
    "loss": 0.4821211099624634,
    "learning_rate": 7.129094402451025e-07,
    "timestep_mean": 494.1875,
    "epsilon_norm": 9.05276107788086,
    "prediction_norm": 7.0758490562438965,
    "timestamp": "2025-07-14T12:34:47.073664"
  },
  {
    "epoch": 2034,
    "batch_idx": 64,
    "loss": 0.19068175554275513,
    "learning_rate": 7.129094402451025e-07,
    "timestep_mean": 455.4375,
    "epsilon_norm": 6.115597724914551,
    "prediction_norm": 5.296938896179199,
    "timestamp": "2025-07-14T12:34:47.078472"
  },
  {
    "epoch": 2034,
    "batch_idx": 96,
    "loss": 0.20411686599254608,
    "learning_rate": 7.129094402451025e-07,
    "timestep_mean": 537.0,
    "epsilon_norm": 7.674581527709961,
    "prediction_norm": 6.754281044006348,
    "timestamp": "2025-07-14T12:34:47.083473"
  },
  {
    "epoch": 2034,
    "batch_idx": 128,
    "loss": 0.22769610583782196,
    "learning_rate": 7.129094402451025e-07,
    "timestep_mean": 492.71875,
    "epsilon_norm": 7.797037124633789,
    "prediction_norm": 6.638715744018555,
    "timestamp": "2025-07-14T12:34:47.088293"
  },
  {
    "epoch": 2034,
    "batch_idx": 160,
    "loss": 0.18891508877277374,
    "learning_rate": 7.129094402451025e-07,
    "timestep_mean": 505.59375,
    "epsilon_norm": 8.70477294921875,
    "prediction_norm": 8.339323043823242,
    "timestamp": "2025-07-14T12:34:47.093120"
  },
  {
    "epoch": 2034,
    "batch_idx": 192,
    "loss": 0.1032557263970375,
    "learning_rate": 7.129094402451025e-07,
    "timestep_mean": 577.375,
    "epsilon_norm": 2.797506093978882,
    "prediction_norm": 2.5156660079956055,
    "timestamp": "2025-07-14T12:34:47.097947"
  },
  {
    "epoch": 2035,
    "batch_idx": 0,
    "loss": 0.31940943002700806,
    "learning_rate": 7.554512747310344e-07,
    "timestep_mean": 418.96875,
    "epsilon_norm": 7.9566168785095215,
    "prediction_norm": 6.326665878295898,
    "timestamp": "2025-07-14T12:34:47.102816"
  },
  {
    "epoch": 2035,
    "batch_idx": 32,
    "loss": 0.20124375820159912,
    "learning_rate": 7.554512747310344e-07,
    "timestep_mean": 542.625,
    "epsilon_norm": 8.107256889343262,
    "prediction_norm": 7.101466655731201,
    "timestamp": "2025-07-14T12:34:47.107652"
  },
  {
    "epoch": 2035,
    "batch_idx": 64,
    "loss": 0.27764278650283813,
    "learning_rate": 7.554512747310344e-07,
    "timestep_mean": 430.1875,
    "epsilon_norm": 7.3555169105529785,
    "prediction_norm": 5.943902015686035,
    "timestamp": "2025-07-14T12:34:47.112485"
  },
  {
    "epoch": 2035,
    "batch_idx": 96,
    "loss": 0.1745433509349823,
    "learning_rate": 7.554512747310344e-07,
    "timestep_mean": 516.6875,
    "epsilon_norm": 7.668391227722168,
    "prediction_norm": 6.6464104652404785,
    "timestamp": "2025-07-14T12:34:47.117440"
  },
  {
    "epoch": 2035,
    "batch_idx": 128,
    "loss": 0.3296620845794678,
    "learning_rate": 7.554512747310344e-07,
    "timestep_mean": 483.15625,
    "epsilon_norm": 7.383674144744873,
    "prediction_norm": 6.184787273406982,
    "timestamp": "2025-07-14T12:34:47.122295"
  },
  {
    "epoch": 2035,
    "batch_idx": 160,
    "loss": 0.3118802309036255,
    "learning_rate": 7.554512747310344e-07,
    "timestep_mean": 502.65625,
    "epsilon_norm": 7.740551471710205,
    "prediction_norm": 6.559384346008301,
    "timestamp": "2025-07-14T12:34:47.127340"
  },
  {
    "epoch": 2035,
    "batch_idx": 192,
    "loss": 0.2344132363796234,
    "learning_rate": 7.554512747310344e-07,
    "timestep_mean": 472.75,
    "epsilon_norm": 4.486958980560303,
    "prediction_norm": 3.5607075691223145,
    "timestamp": "2025-07-14T12:34:47.132330"
  },
  {
    "epoch": 2036,
    "batch_idx": 0,
    "loss": 0.33799466490745544,
    "learning_rate": 7.992249455124896e-07,
    "timestep_mean": 460.40625,
    "epsilon_norm": 9.491134643554688,
    "prediction_norm": 7.243000507354736,
    "timestamp": "2025-07-14T12:34:47.137576"
  },
  {
    "epoch": 2036,
    "batch_idx": 32,
    "loss": 0.13385051488876343,
    "learning_rate": 7.992249455124896e-07,
    "timestep_mean": 573.5,
    "epsilon_norm": 7.258767127990723,
    "prediction_norm": 6.491025924682617,
    "timestamp": "2025-07-14T12:34:47.142738"
  },
  {
    "epoch": 2036,
    "batch_idx": 64,
    "loss": 0.3000296652317047,
    "learning_rate": 7.992249455124896e-07,
    "timestep_mean": 471.75,
    "epsilon_norm": 7.6861467361450195,
    "prediction_norm": 6.710277557373047,
    "timestamp": "2025-07-14T12:34:47.147842"
  },
  {
    "epoch": 2036,
    "batch_idx": 96,
    "loss": 0.22050254046916962,
    "learning_rate": 7.992249455124896e-07,
    "timestep_mean": 579.09375,
    "epsilon_norm": 8.624847412109375,
    "prediction_norm": 7.582820415496826,
    "timestamp": "2025-07-14T12:34:47.153004"
  },
  {
    "epoch": 2036,
    "batch_idx": 128,
    "loss": 0.163286030292511,
    "learning_rate": 7.992249455124896e-07,
    "timestep_mean": 544.34375,
    "epsilon_norm": 7.888206958770752,
    "prediction_norm": 7.489153861999512,
    "timestamp": "2025-07-14T12:34:47.158160"
  },
  {
    "epoch": 2036,
    "batch_idx": 160,
    "loss": 0.3260596990585327,
    "learning_rate": 7.992249455124896e-07,
    "timestep_mean": 527.84375,
    "epsilon_norm": 8.936323165893555,
    "prediction_norm": 7.467880725860596,
    "timestamp": "2025-07-14T12:34:47.163429"
  },
  {
    "epoch": 2036,
    "batch_idx": 192,
    "loss": 0.6048535108566284,
    "learning_rate": 7.992249455124896e-07,
    "timestep_mean": 250.25,
    "epsilon_norm": 4.516265869140625,
    "prediction_norm": 2.657716989517212,
    "timestamp": "2025-07-14T12:34:47.168515"
  },
  {
    "epoch": 2037,
    "batch_idx": 0,
    "loss": 0.3241926431655884,
    "learning_rate": 8.442303445823093e-07,
    "timestep_mean": 494.15625,
    "epsilon_norm": 8.311702728271484,
    "prediction_norm": 6.746626853942871,
    "timestamp": "2025-07-14T12:34:47.173595"
  },
  {
    "epoch": 2037,
    "batch_idx": 32,
    "loss": 0.1971430629491806,
    "learning_rate": 8.442303445823093e-07,
    "timestep_mean": 530.71875,
    "epsilon_norm": 7.3579230308532715,
    "prediction_norm": 6.368607044219971,
    "timestamp": "2025-07-14T12:34:47.178770"
  },
  {
    "epoch": 2037,
    "batch_idx": 64,
    "loss": 0.26768285036087036,
    "learning_rate": 8.442303445823093e-07,
    "timestep_mean": 434.5,
    "epsilon_norm": 6.358294486999512,
    "prediction_norm": 5.875423431396484,
    "timestamp": "2025-07-14T12:34:47.184074"
  },
  {
    "epoch": 2037,
    "batch_idx": 96,
    "loss": 0.3329637944698334,
    "learning_rate": 8.442303445823093e-07,
    "timestep_mean": 521.5625,
    "epsilon_norm": 7.952124118804932,
    "prediction_norm": 6.848420143127441,
    "timestamp": "2025-07-14T12:34:47.189225"
  },
  {
    "epoch": 2037,
    "batch_idx": 128,
    "loss": 0.21326856315135956,
    "learning_rate": 8.442303445823093e-07,
    "timestep_mean": 499.84375,
    "epsilon_norm": 7.727427959442139,
    "prediction_norm": 6.405509948730469,
    "timestamp": "2025-07-14T12:34:47.194400"
  },
  {
    "epoch": 2037,
    "batch_idx": 160,
    "loss": 0.268005907535553,
    "learning_rate": 8.442303445823093e-07,
    "timestep_mean": 496.84375,
    "epsilon_norm": 8.163276672363281,
    "prediction_norm": 7.20754337310791,
    "timestamp": "2025-07-14T12:34:47.199518"
  },
  {
    "epoch": 2037,
    "batch_idx": 192,
    "loss": 0.3561454117298126,
    "learning_rate": 8.442303445823093e-07,
    "timestep_mean": 408.875,
    "epsilon_norm": 3.048985719680786,
    "prediction_norm": 2.3545608520507812,
    "timestamp": "2025-07-14T12:34:47.204811"
  },
  {
    "epoch": 2038,
    "batch_idx": 0,
    "loss": 0.3753277063369751,
    "learning_rate": 8.904673608940991e-07,
    "timestep_mean": 470.34375,
    "epsilon_norm": 8.254744529724121,
    "prediction_norm": 6.695363998413086,
    "timestamp": "2025-07-14T12:34:47.209878"
  },
  {
    "epoch": 2038,
    "batch_idx": 32,
    "loss": 0.4377000629901886,
    "learning_rate": 8.904673608940991e-07,
    "timestep_mean": 471.5625,
    "epsilon_norm": 8.127866744995117,
    "prediction_norm": 6.66713285446167,
    "timestamp": "2025-07-14T12:34:47.214996"
  },
  {
    "epoch": 2038,
    "batch_idx": 64,
    "loss": 0.2419777810573578,
    "learning_rate": 8.904673608940991e-07,
    "timestep_mean": 512.84375,
    "epsilon_norm": 7.377875328063965,
    "prediction_norm": 6.5772552490234375,
    "timestamp": "2025-07-14T12:34:47.220238"
  },
  {
    "epoch": 2038,
    "batch_idx": 96,
    "loss": 0.4436298906803131,
    "learning_rate": 8.904673608940991e-07,
    "timestep_mean": 515.34375,
    "epsilon_norm": 8.646711349487305,
    "prediction_norm": 6.975150108337402,
    "timestamp": "2025-07-14T12:34:47.225469"
  },
  {
    "epoch": 2038,
    "batch_idx": 128,
    "loss": 0.6023558378219604,
    "learning_rate": 8.904673608940991e-07,
    "timestep_mean": 502.0,
    "epsilon_norm": 9.202987670898438,
    "prediction_norm": 7.639000415802002,
    "timestamp": "2025-07-14T12:34:47.230643"
  },
  {
    "epoch": 2038,
    "batch_idx": 160,
    "loss": 0.2717383801937103,
    "learning_rate": 8.904673608940991e-07,
    "timestep_mean": 579.09375,
    "epsilon_norm": 9.218409538269043,
    "prediction_norm": 7.329357624053955,
    "timestamp": "2025-07-14T12:34:47.235539"
  },
  {
    "epoch": 2038,
    "batch_idx": 192,
    "loss": 0.3602745532989502,
    "learning_rate": 8.904673608940991e-07,
    "timestep_mean": 294.375,
    "epsilon_norm": 4.2842793464660645,
    "prediction_norm": 3.3640899658203125,
    "timestamp": "2025-07-14T12:34:47.240448"
  },
  {
    "epoch": 2039,
    "batch_idx": 0,
    "loss": 0.3935101628303528,
    "learning_rate": 9.379358803626733e-07,
    "timestep_mean": 543.75,
    "epsilon_norm": 9.202988624572754,
    "prediction_norm": 7.2330827713012695,
    "timestamp": "2025-07-14T12:34:47.245625"
  },
  {
    "epoch": 2039,
    "batch_idx": 32,
    "loss": 0.16441065073013306,
    "learning_rate": 9.379358803626733e-07,
    "timestep_mean": 553.125,
    "epsilon_norm": 8.2878999710083,
    "prediction_norm": 7.5466837882995605,
    "timestamp": "2025-07-14T12:34:47.250953"
  },
  {
    "epoch": 2039,
    "batch_idx": 64,
    "loss": 0.30422016978263855,
    "learning_rate": 9.379358803626733e-07,
    "timestep_mean": 475.25,
    "epsilon_norm": 9.222321510314941,
    "prediction_norm": 7.85712194442749,
    "timestamp": "2025-07-14T12:34:47.256149"
  },
  {
    "epoch": 2039,
    "batch_idx": 96,
    "loss": 0.25160354375839233,
    "learning_rate": 9.379358803626733e-07,
    "timestep_mean": 529.59375,
    "epsilon_norm": 8.442150115966797,
    "prediction_norm": 7.0741047859191895,
    "timestamp": "2025-07-14T12:34:47.261340"
  },
  {
    "epoch": 2039,
    "batch_idx": 128,
    "loss": 0.37074708938598633,
    "learning_rate": 9.379358803626733e-07,
    "timestep_mean": 490.15625,
    "epsilon_norm": 8.207876205444336,
    "prediction_norm": 6.7100934982299805,
    "timestamp": "2025-07-14T12:34:47.266485"
  },
  {
    "epoch": 2039,
    "batch_idx": 160,
    "loss": 0.44954872131347656,
    "learning_rate": 9.379358803626733e-07,
    "timestep_mean": 560.46875,
    "epsilon_norm": 9.189571380615234,
    "prediction_norm": 7.740300178527832,
    "timestamp": "2025-07-14T12:34:47.271762"
  },
  {
    "epoch": 2039,
    "batch_idx": 192,
    "loss": 0.3417258858680725,
    "learning_rate": 9.379358803626733e-07,
    "timestep_mean": 500.625,
    "epsilon_norm": 5.6140313148498535,
    "prediction_norm": 4.249216556549072,
    "timestamp": "2025-07-14T12:34:47.276925"
  },
  {
    "epoch": 2040,
    "batch_idx": 0,
    "loss": 0.23934322595596313,
    "learning_rate": 9.866357858642213e-07,
    "timestep_mean": 582.1875,
    "epsilon_norm": 7.827943801879883,
    "prediction_norm": 7.232913017272949,
    "timestamp": "2025-07-14T12:34:47.281999"
  },
  {
    "epoch": 2040,
    "batch_idx": 32,
    "loss": 0.2775980830192566,
    "learning_rate": 9.866357858642213e-07,
    "timestep_mean": 491.46875,
    "epsilon_norm": 7.960445880889893,
    "prediction_norm": 7.1719512939453125,
    "timestamp": "2025-07-14T12:34:47.287235"
  },
  {
    "epoch": 2040,
    "batch_idx": 64,
    "loss": 0.2715488076210022,
    "learning_rate": 9.866357858642213e-07,
    "timestep_mean": 518.8125,
    "epsilon_norm": 8.051944732666016,
    "prediction_norm": 6.808566093444824,
    "timestamp": "2025-07-14T12:34:47.292372"
  },
  {
    "epoch": 2040,
    "batch_idx": 96,
    "loss": 0.1652948260307312,
    "learning_rate": 9.866357858642213e-07,
    "timestep_mean": 457.3125,
    "epsilon_norm": 7.369818210601807,
    "prediction_norm": 6.366577625274658,
    "timestamp": "2025-07-14T12:34:47.297522"
  },
  {
    "epoch": 2040,
    "batch_idx": 128,
    "loss": 0.23426979780197144,
    "learning_rate": 9.866357858642213e-07,
    "timestep_mean": 548.40625,
    "epsilon_norm": 8.63782787322998,
    "prediction_norm": 7.708540439605713,
    "timestamp": "2025-07-14T12:34:47.302639"
  },
  {
    "epoch": 2040,
    "batch_idx": 160,
    "loss": 0.13784947991371155,
    "learning_rate": 9.866357858642213e-07,
    "timestep_mean": 440.8125,
    "epsilon_norm": 8.108134269714355,
    "prediction_norm": 7.657190322875977,
    "timestamp": "2025-07-14T12:34:47.307771"
  },
  {
    "epoch": 2040,
    "batch_idx": 192,
    "loss": 0.30758339166641235,
    "learning_rate": 9.866357858642213e-07,
    "timestep_mean": 414.25,
    "epsilon_norm": 3.241882562637329,
    "prediction_norm": 2.347456455230713,
    "timestamp": "2025-07-14T12:34:47.312966"
  },
  {
    "epoch": 2041,
    "batch_idx": 0,
    "loss": 0.12969040870666504,
    "learning_rate": 1.036566957236474e-06,
    "timestep_mean": 556.46875,
    "epsilon_norm": 7.875244617462158,
    "prediction_norm": 7.228122711181641,
    "timestamp": "2025-07-14T12:34:47.318080"
  },
  {
    "epoch": 2041,
    "batch_idx": 32,
    "loss": 0.18530206382274628,
    "learning_rate": 1.036566957236474e-06,
    "timestep_mean": 516.84375,
    "epsilon_norm": 7.7854533195495605,
    "prediction_norm": 7.201870441436768,
    "timestamp": "2025-07-14T12:34:47.323079"
  },
  {
    "epoch": 2041,
    "batch_idx": 64,
    "loss": 0.2499852031469345,
    "learning_rate": 1.036566957236474e-06,
    "timestep_mean": 493.75,
    "epsilon_norm": 7.618873596191406,
    "prediction_norm": 6.258198261260986,
    "timestamp": "2025-07-14T12:34:47.328166"
  },
  {
    "epoch": 2041,
    "batch_idx": 96,
    "loss": 0.12824882566928864,
    "learning_rate": 1.036566957236474e-06,
    "timestep_mean": 532.15625,
    "epsilon_norm": 7.622883319854736,
    "prediction_norm": 6.766099452972412,
    "timestamp": "2025-07-14T12:34:47.333241"
  },
  {
    "epoch": 2041,
    "batch_idx": 128,
    "loss": 0.38069748878479004,
    "learning_rate": 1.036566957236474e-06,
    "timestep_mean": 470.6875,
    "epsilon_norm": 8.460826873779297,
    "prediction_norm": 6.405519485473633,
    "timestamp": "2025-07-14T12:34:47.338558"
  },
  {
    "epoch": 2041,
    "batch_idx": 160,
    "loss": 0.20083652436733246,
    "learning_rate": 1.036566957236474e-06,
    "timestep_mean": 495.59375,
    "epsilon_norm": 8.226398468017578,
    "prediction_norm": 7.155532360076904,
    "timestamp": "2025-07-14T12:34:47.343723"
  },
  {
    "epoch": 2041,
    "batch_idx": 192,
    "loss": 0.5186538696289062,
    "learning_rate": 1.036566957236474e-06,
    "timestep_mean": 600.875,
    "epsilon_norm": 4.33950662612915,
    "prediction_norm": 2.2978262901306152,
    "timestamp": "2025-07-14T12:34:47.348751"
  },
  {
    "epoch": 2042,
    "batch_idx": 0,
    "loss": 0.19042930006980896,
    "learning_rate": 1.0877292712792595e-06,
    "timestep_mean": 587.34375,
    "epsilon_norm": 7.948480606079102,
    "prediction_norm": 7.279391765594482,
    "timestamp": "2025-07-14T12:34:47.353961"
  },
  {
    "epoch": 2042,
    "batch_idx": 32,
    "loss": 0.22058914601802826,
    "learning_rate": 1.0877292712792595e-06,
    "timestep_mean": 492.1875,
    "epsilon_norm": 8.099129676818848,
    "prediction_norm": 7.290583610534668,
    "timestamp": "2025-07-14T12:34:47.359227"
  },
  {
    "epoch": 2042,
    "batch_idx": 64,
    "loss": 0.19299496710300446,
    "learning_rate": 1.0877292712792595e-06,
    "timestep_mean": 529.90625,
    "epsilon_norm": 7.513627052307129,
    "prediction_norm": 6.6211347579956055,
    "timestamp": "2025-07-14T12:34:47.364193"
  },
  {
    "epoch": 2042,
    "batch_idx": 96,
    "loss": 0.1883707344532013,
    "learning_rate": 1.0877292712792595e-06,
    "timestep_mean": 538.84375,
    "epsilon_norm": 7.803430080413818,
    "prediction_norm": 6.732841491699219,
    "timestamp": "2025-07-14T12:34:47.369184"
  },
  {
    "epoch": 2042,
    "batch_idx": 128,
    "loss": 0.4033026099205017,
    "learning_rate": 1.0877292712792595e-06,
    "timestep_mean": 439.9375,
    "epsilon_norm": 9.379149436950684,
    "prediction_norm": 7.301704406738281,
    "timestamp": "2025-07-14T12:34:47.374285"
  },
  {
    "epoch": 2042,
    "batch_idx": 160,
    "loss": 0.3936005234718323,
    "learning_rate": 1.0877292712792595e-06,
    "timestep_mean": 498.09375,
    "epsilon_norm": 7.693088531494141,
    "prediction_norm": 5.519759654998779,
    "timestamp": "2025-07-14T12:34:47.379411"
  },
  {
    "epoch": 2042,
    "batch_idx": 192,
    "loss": 0.6432991027832031,
    "learning_rate": 1.0877292712792595e-06,
    "timestep_mean": 498.75,
    "epsilon_norm": 4.103331565856934,
    "prediction_norm": 2.3194127082824707,
    "timestamp": "2025-07-14T12:34:47.384607"
  },
  {
    "epoch": 2043,
    "batch_idx": 0,
    "loss": 0.07741117477416992,
    "learning_rate": 1.1401226017547238e-06,
    "timestep_mean": 540.65625,
    "epsilon_norm": 8.250102996826172,
    "prediction_norm": 7.976229190826416,
    "timestamp": "2025-07-14T12:34:47.389729"
  },
  {
    "epoch": 2043,
    "batch_idx": 32,
    "loss": 0.2559661269187927,
    "learning_rate": 1.1401226017547238e-06,
    "timestep_mean": 421.28125,
    "epsilon_norm": 7.652652263641357,
    "prediction_norm": 6.806190013885498,
    "timestamp": "2025-07-14T12:34:47.394929"
  },
  {
    "epoch": 2043,
    "batch_idx": 64,
    "loss": 0.2760361135005951,
    "learning_rate": 1.1401226017547238e-06,
    "timestep_mean": 528.34375,
    "epsilon_norm": 7.445544242858887,
    "prediction_norm": 6.6307783126831055,
    "timestamp": "2025-07-14T12:34:47.400217"
  },
  {
    "epoch": 2043,
    "batch_idx": 96,
    "loss": 0.17656484246253967,
    "learning_rate": 1.1401226017547238e-06,
    "timestep_mean": 483.375,
    "epsilon_norm": 7.215847969055176,
    "prediction_norm": 6.272707462310791,
    "timestamp": "2025-07-14T12:34:47.405320"
  },
  {
    "epoch": 2043,
    "batch_idx": 128,
    "loss": 0.2713721990585327,
    "learning_rate": 1.1401226017547238e-06,
    "timestep_mean": 460.28125,
    "epsilon_norm": 7.734200477600098,
    "prediction_norm": 7.11170768737793,
    "timestamp": "2025-07-14T12:34:47.410221"
  },
  {
    "epoch": 2043,
    "batch_idx": 160,
    "loss": 0.34258973598480225,
    "learning_rate": 1.1401226017547238e-06,
    "timestep_mean": 465.40625,
    "epsilon_norm": 8.7178955078125,
    "prediction_norm": 7.383386135101318,
    "timestamp": "2025-07-14T12:34:47.415150"
  },
  {
    "epoch": 2043,
    "batch_idx": 192,
    "loss": 0.2371828407049179,
    "learning_rate": 1.1401226017547238e-06,
    "timestep_mean": 460.375,
    "epsilon_norm": 4.16818904876709,
    "prediction_norm": 3.6811633110046387,
    "timestamp": "2025-07-14T12:34:47.420421"
  },
  {
    "epoch": 2044,
    "batch_idx": 0,
    "loss": 0.21614600718021393,
    "learning_rate": 1.1937468193873878e-06,
    "timestep_mean": 490.1875,
    "epsilon_norm": 7.904793739318848,
    "prediction_norm": 6.919403076171875,
    "timestamp": "2025-07-14T12:34:47.425735"
  },
  {
    "epoch": 2044,
    "batch_idx": 32,
    "loss": 0.5068987607955933,
    "learning_rate": 1.1937468193873878e-06,
    "timestep_mean": 415.53125,
    "epsilon_norm": 8.639280319213867,
    "prediction_norm": 6.633112907409668,
    "timestamp": "2025-07-14T12:34:47.431047"
  },
  {
    "epoch": 2044,
    "batch_idx": 64,
    "loss": 0.22917892038822174,
    "learning_rate": 1.1937468193873878e-06,
    "timestep_mean": 449.34375,
    "epsilon_norm": 8.195000648498535,
    "prediction_norm": 7.32959508895874,
    "timestamp": "2025-07-14T12:34:47.436201"
  },
  {
    "epoch": 2044,
    "batch_idx": 96,
    "loss": 0.28327709436416626,
    "learning_rate": 1.1937468193873878e-06,
    "timestep_mean": 511.90625,
    "epsilon_norm": 7.882530212402344,
    "prediction_norm": 6.3463311195373535,
    "timestamp": "2025-07-14T12:34:47.441368"
  },
  {
    "epoch": 2044,
    "batch_idx": 128,
    "loss": 0.25409990549087524,
    "learning_rate": 1.1937468193873878e-06,
    "timestep_mean": 541.875,
    "epsilon_norm": 7.591244697570801,
    "prediction_norm": 6.270848751068115,
    "timestamp": "2025-07-14T12:34:47.446680"
  },
  {
    "epoch": 2044,
    "batch_idx": 160,
    "loss": 0.2926709055900574,
    "learning_rate": 1.1937468193873878e-06,
    "timestep_mean": 510.65625,
    "epsilon_norm": 7.787953853607178,
    "prediction_norm": 6.287336826324463,
    "timestamp": "2025-07-14T12:34:47.451838"
  },
  {
    "epoch": 2044,
    "batch_idx": 192,
    "loss": 0.41473081707954407,
    "learning_rate": 1.1937468193873878e-06,
    "timestep_mean": 595.625,
    "epsilon_norm": 4.809670448303223,
    "prediction_norm": 3.9779672622680664,
    "timestamp": "2025-07-14T12:34:47.456936"
  },
  {
    "epoch": 2045,
    "batch_idx": 0,
    "loss": 0.18218418955802917,
    "learning_rate": 1.248601791864924e-06,
    "timestep_mean": 496.9375,
    "epsilon_norm": 7.425693988800049,
    "prediction_norm": 6.822237491607666,
    "timestamp": "2025-07-14T12:34:47.462046"
  },
  {
    "epoch": 2045,
    "batch_idx": 32,
    "loss": 0.30452242493629456,
    "learning_rate": 1.248601791864924e-06,
    "timestep_mean": 524.71875,
    "epsilon_norm": 7.474750995635986,
    "prediction_norm": 5.940999984741211,
    "timestamp": "2025-07-14T12:34:47.467331"
  },
  {
    "epoch": 2045,
    "batch_idx": 64,
    "loss": 0.19619885087013245,
    "learning_rate": 1.248601791864924e-06,
    "timestep_mean": 489.875,
    "epsilon_norm": 8.317007064819336,
    "prediction_norm": 7.055522441864014,
    "timestamp": "2025-07-14T12:34:47.472482"
  },
  {
    "epoch": 2045,
    "batch_idx": 96,
    "loss": 0.28315120935440063,
    "learning_rate": 1.248601791864924e-06,
    "timestep_mean": 587.96875,
    "epsilon_norm": 8.413213729858398,
    "prediction_norm": 6.637636661529541,
    "timestamp": "2025-07-14T12:34:47.477649"
  },
  {
    "epoch": 2045,
    "batch_idx": 128,
    "loss": 0.2841547727584839,
    "learning_rate": 1.248601791864924e-06,
    "timestep_mean": 475.5625,
    "epsilon_norm": 6.861369609832764,
    "prediction_norm": 6.065699577331543,
    "timestamp": "2025-07-14T12:34:47.482770"
  },
  {
    "epoch": 2045,
    "batch_idx": 160,
    "loss": 0.5274609327316284,
    "learning_rate": 1.248601791864924e-06,
    "timestep_mean": 394.5,
    "epsilon_norm": 8.40410327911377,
    "prediction_norm": 5.9689249992370605,
    "timestamp": "2025-07-14T12:34:47.488026"
  },
  {
    "epoch": 2045,
    "batch_idx": 192,
    "loss": 0.23832577466964722,
    "learning_rate": 1.248601791864924e-06,
    "timestep_mean": 560.125,
    "epsilon_norm": 3.0775020122528076,
    "prediction_norm": 2.257389545440674,
    "timestamp": "2025-07-14T12:34:47.493151"
  },
  {
    "epoch": 2046,
    "batch_idx": 0,
    "loss": 0.2778986394405365,
    "learning_rate": 1.3046873838381559e-06,
    "timestep_mean": 468.53125,
    "epsilon_norm": 7.607690334320068,
    "prediction_norm": 6.462888240814209,
    "timestamp": "2025-07-14T12:34:47.498240"
  },
  {
    "epoch": 2046,
    "batch_idx": 32,
    "loss": 0.33621668815612793,
    "learning_rate": 1.3046873838381559e-06,
    "timestep_mean": 498.9375,
    "epsilon_norm": 8.200023651123047,
    "prediction_norm": 6.943227291107178,
    "timestamp": "2025-07-14T12:34:47.503171"
  },
  {
    "epoch": 2046,
    "batch_idx": 64,
    "loss": 0.20271781086921692,
    "learning_rate": 1.3046873838381559e-06,
    "timestep_mean": 559.6875,
    "epsilon_norm": 7.991138458251953,
    "prediction_norm": 7.087881565093994,
    "timestamp": "2025-07-14T12:34:47.507998"
  },
  {
    "epoch": 2046,
    "batch_idx": 96,
    "loss": 0.2971680760383606,
    "learning_rate": 1.3046873838381559e-06,
    "timestep_mean": 501.75,
    "epsilon_norm": 8.425092697143555,
    "prediction_norm": 6.374500751495361,
    "timestamp": "2025-07-14T12:34:47.512899"
  },
  {
    "epoch": 2046,
    "batch_idx": 128,
    "loss": 0.2633826732635498,
    "learning_rate": 1.3046873838381559e-06,
    "timestep_mean": 483.34375,
    "epsilon_norm": 7.054297924041748,
    "prediction_norm": 5.449850559234619,
    "timestamp": "2025-07-14T12:34:47.517844"
  },
  {
    "epoch": 2046,
    "batch_idx": 160,
    "loss": 0.21886713802814484,
    "learning_rate": 1.3046873838381559e-06,
    "timestep_mean": 568.21875,
    "epsilon_norm": 9.006211280822754,
    "prediction_norm": 8.467549324035645,
    "timestamp": "2025-07-14T12:34:47.522668"
  },
  {
    "epoch": 2046,
    "batch_idx": 192,
    "loss": 0.2637050151824951,
    "learning_rate": 1.3046873838381559e-06,
    "timestep_mean": 315.0,
    "epsilon_norm": 4.545052528381348,
    "prediction_norm": 3.782276153564453,
    "timestamp": "2025-07-14T12:34:47.527482"
  },
  {
    "epoch": 2047,
    "batch_idx": 0,
    "loss": 0.30362725257873535,
    "learning_rate": 1.3620034569214472e-06,
    "timestep_mean": 568.71875,
    "epsilon_norm": 7.985861301422119,
    "prediction_norm": 6.498906135559082,
    "timestamp": "2025-07-14T12:34:47.532299"
  },
  {
    "epoch": 2047,
    "batch_idx": 32,
    "loss": 0.1691381186246872,
    "learning_rate": 1.3620034569214472e-06,
    "timestep_mean": 565.75,
    "epsilon_norm": 8.078368186950684,
    "prediction_norm": 7.4822869300842285,
    "timestamp": "2025-07-14T12:34:47.537134"
  },
  {
    "epoch": 2047,
    "batch_idx": 64,
    "loss": 0.42258545756340027,
    "learning_rate": 1.3620034569214472e-06,
    "timestep_mean": 432.21875,
    "epsilon_norm": 7.276684761047363,
    "prediction_norm": 5.630873680114746,
    "timestamp": "2025-07-14T12:34:47.542044"
  },
  {
    "epoch": 2047,
    "batch_idx": 96,
    "loss": 0.11536731570959091,
    "learning_rate": 1.3620034569214472e-06,
    "timestep_mean": 573.40625,
    "epsilon_norm": 6.965897083282471,
    "prediction_norm": 6.605372428894043,
    "timestamp": "2025-07-14T12:34:47.546862"
  },
  {
    "epoch": 2047,
    "batch_idx": 128,
    "loss": 0.3520384132862091,
    "learning_rate": 1.3620034569214472e-06,
    "timestep_mean": 582.78125,
    "epsilon_norm": 8.091838836669922,
    "prediction_norm": 7.053319931030273,
    "timestamp": "2025-07-14T12:34:47.551660"
  },
  {
    "epoch": 2047,
    "batch_idx": 160,
    "loss": 0.20793698728084564,
    "learning_rate": 1.3620034569214472e-06,
    "timestep_mean": 461.90625,
    "epsilon_norm": 7.350335597991943,
    "prediction_norm": 6.220162391662598,
    "timestamp": "2025-07-14T12:34:47.556431"
  },
  {
    "epoch": 2047,
    "batch_idx": 192,
    "loss": 0.33156636357307434,
    "learning_rate": 1.3620034569214472e-06,
    "timestep_mean": 486.625,
    "epsilon_norm": 4.889434337615967,
    "prediction_norm": 3.5014829635620117,
    "timestamp": "2025-07-14T12:34:47.561241"
  },
  {
    "epoch": 2048,
    "batch_idx": 0,
    "loss": 0.1873515546321869,
    "learning_rate": 1.4205498696930347e-06,
    "timestep_mean": 538.71875,
    "epsilon_norm": 7.675119876861572,
    "prediction_norm": 6.766205310821533,
    "timestamp": "2025-07-14T12:34:47.566084"
  },
  {
    "epoch": 2048,
    "batch_idx": 32,
    "loss": 0.24120305478572845,
    "learning_rate": 1.4205498696930347e-06,
    "timestep_mean": 434.75,
    "epsilon_norm": 6.932443618774414,
    "prediction_norm": 5.872193813323975,
    "timestamp": "2025-07-14T12:34:47.570960"
  },
  {
    "epoch": 2048,
    "batch_idx": 64,
    "loss": 0.19441303610801697,
    "learning_rate": 1.4205498696930347e-06,
    "timestep_mean": 512.09375,
    "epsilon_norm": 8.297517776489258,
    "prediction_norm": 7.490823745727539,
    "timestamp": "2025-07-14T12:34:47.576012"
  },
  {
    "epoch": 2048,
    "batch_idx": 96,
    "loss": 0.4641045928001404,
    "learning_rate": 1.4205498696930347e-06,
    "timestep_mean": 507.4375,
    "epsilon_norm": 8.636001586914062,
    "prediction_norm": 6.824676990509033,
    "timestamp": "2025-07-14T12:34:47.581017"
  },
  {
    "epoch": 2048,
    "batch_idx": 128,
    "loss": 0.2422996163368225,
    "learning_rate": 1.4205498696930347e-06,
    "timestep_mean": 378.40625,
    "epsilon_norm": 7.110203266143799,
    "prediction_norm": 6.174746513366699,
    "timestamp": "2025-07-14T12:34:47.586024"
  },
  {
    "epoch": 2048,
    "batch_idx": 160,
    "loss": 0.5342414379119873,
    "learning_rate": 1.4205498696930347e-06,
    "timestep_mean": 399.8125,
    "epsilon_norm": 7.594616413116455,
    "prediction_norm": 5.319484233856201,
    "timestamp": "2025-07-14T12:34:47.591002"
  },
  {
    "epoch": 2048,
    "batch_idx": 192,
    "loss": 0.31054285168647766,
    "learning_rate": 1.4205498696930347e-06,
    "timestep_mean": 627.25,
    "epsilon_norm": 3.0698649883270264,
    "prediction_norm": 3.1022603511810303,
    "timestamp": "2025-07-14T12:34:47.596102"
  },
  {
    "epoch": 2049,
    "batch_idx": 0,
    "loss": 0.25838980078697205,
    "learning_rate": 1.4803264776955278e-06,
    "timestep_mean": 442.65625,
    "epsilon_norm": 9.024282455444336,
    "prediction_norm": 8.255765914916992,
    "timestamp": "2025-07-14T12:34:47.601217"
  },
  {
    "epoch": 2049,
    "batch_idx": 32,
    "loss": 0.2016637772321701,
    "learning_rate": 1.4803264776955278e-06,
    "timestep_mean": 483.09375,
    "epsilon_norm": 8.698124885559082,
    "prediction_norm": 7.653447151184082,
    "timestamp": "2025-07-14T12:34:47.606256"
  },
  {
    "epoch": 2049,
    "batch_idx": 64,
    "loss": 0.27446725964546204,
    "learning_rate": 1.4803264776955278e-06,
    "timestep_mean": 390.5625,
    "epsilon_norm": 7.959740161895752,
    "prediction_norm": 6.550168991088867,
    "timestamp": "2025-07-14T12:34:47.611433"
  },
  {
    "epoch": 2049,
    "batch_idx": 96,
    "loss": 0.2503127455711365,
    "learning_rate": 1.4803264776955278e-06,
    "timestep_mean": 473.0625,
    "epsilon_norm": 7.1749653816223145,
    "prediction_norm": 5.932173252105713,
    "timestamp": "2025-07-14T12:34:47.616593"
  },
  {
    "epoch": 2049,
    "batch_idx": 128,
    "loss": 0.2286776304244995,
    "learning_rate": 1.4803264776955278e-06,
    "timestep_mean": 450.53125,
    "epsilon_norm": 7.719157695770264,
    "prediction_norm": 6.740041732788086,
    "timestamp": "2025-07-14T12:34:47.621862"
  },
  {
    "epoch": 2049,
    "batch_idx": 160,
    "loss": 0.17061907052993774,
    "learning_rate": 1.4803264776955278e-06,
    "timestep_mean": 494.71875,
    "epsilon_norm": 6.835962772369385,
    "prediction_norm": 6.422665596008301,
    "timestamp": "2025-07-14T12:34:47.627003"
  },
  {
    "epoch": 2049,
    "batch_idx": 192,
    "loss": 0.3361895680427551,
    "learning_rate": 1.4803264776955278e-06,
    "timestep_mean": 455.875,
    "epsilon_norm": 4.650722980499268,
    "prediction_norm": 3.1486270427703857,
    "timestamp": "2025-07-14T12:34:47.632019"
  },
  {
    "epoch": 2050,
    "batch_idx": 0,
    "loss": 0.20589400827884674,
    "learning_rate": 1.5413331334360201e-06,
    "timestep_mean": 545.5625,
    "epsilon_norm": 8.628669738769531,
    "prediction_norm": 7.728201389312744,
    "timestamp": "2025-07-14T12:34:47.637178"
  },
  {
    "epoch": 2050,
    "batch_idx": 32,
    "loss": 0.13400134444236755,
    "learning_rate": 1.5413331334360201e-06,
    "timestep_mean": 541.875,
    "epsilon_norm": 7.840422630310059,
    "prediction_norm": 6.640491962432861,
    "timestamp": "2025-07-14T12:34:47.642395"
  },
  {
    "epoch": 2050,
    "batch_idx": 64,
    "loss": 0.2657386064529419,
    "learning_rate": 1.5413331334360201e-06,
    "timestep_mean": 480.25,
    "epsilon_norm": 6.71968936920166,
    "prediction_norm": 6.064350605010986,
    "timestamp": "2025-07-14T12:34:47.647509"
  },
  {
    "epoch": 2050,
    "batch_idx": 96,
    "loss": 0.1990167200565338,
    "learning_rate": 1.5413331334360201e-06,
    "timestep_mean": 472.5625,
    "epsilon_norm": 8.228225708007812,
    "prediction_norm": 7.381017208099365,
    "timestamp": "2025-07-14T12:34:47.652687"
  },
  {
    "epoch": 2050,
    "batch_idx": 128,
    "loss": 0.42550167441368103,
    "learning_rate": 1.5413331334360201e-06,
    "timestep_mean": 442.5,
    "epsilon_norm": 8.4921293258667,
    "prediction_norm": 6.673159122467041,
    "timestamp": "2025-07-14T12:34:47.657771"
  },
  {
    "epoch": 2050,
    "batch_idx": 160,
    "loss": 0.2805922329425812,
    "learning_rate": 1.5413331334360201e-06,
    "timestep_mean": 407.375,
    "epsilon_norm": 8.080670356750488,
    "prediction_norm": 6.547602653503418,
    "timestamp": "2025-07-14T12:34:47.662978"
  },
  {
    "epoch": 2050,
    "batch_idx": 192,
    "loss": 0.6073995232582092,
    "learning_rate": 1.5413331334360201e-06,
    "timestep_mean": 524.875,
    "epsilon_norm": 4.545392990112305,
    "prediction_norm": 3.1563355922698975,
    "timestamp": "2025-07-14T12:34:47.668058"
  },
  {
    "epoch": 2051,
    "batch_idx": 0,
    "loss": 0.16999737918376923,
    "learning_rate": 1.6035696863866988e-06,
    "timestep_mean": 472.90625,
    "epsilon_norm": 6.445538520812988,
    "prediction_norm": 6.223296642303467,
    "timestamp": "2025-07-14T12:34:47.673127"
  },
  {
    "epoch": 2051,
    "batch_idx": 32,
    "loss": 0.3545001745223999,
    "learning_rate": 1.6035696863866988e-06,
    "timestep_mean": 476.34375,
    "epsilon_norm": 8.126764297485352,
    "prediction_norm": 6.419554233551025,
    "timestamp": "2025-07-14T12:34:47.678404"
  },
  {
    "epoch": 2051,
    "batch_idx": 64,
    "loss": 0.1818736344575882,
    "learning_rate": 1.6035696863866988e-06,
    "timestep_mean": 511.5,
    "epsilon_norm": 7.460826396942139,
    "prediction_norm": 6.827337741851807,
    "timestamp": "2025-07-14T12:34:47.683720"
  },
  {
    "epoch": 2051,
    "batch_idx": 96,
    "loss": 0.14981582760810852,
    "learning_rate": 1.6035696863866988e-06,
    "timestep_mean": 467.9375,
    "epsilon_norm": 7.3465166091918945,
    "prediction_norm": 6.809882164001465,
    "timestamp": "2025-07-14T12:34:47.688916"
  },
  {
    "epoch": 2051,
    "batch_idx": 128,
    "loss": 0.4623875319957733,
    "learning_rate": 1.6035696863866988e-06,
    "timestep_mean": 483.4375,
    "epsilon_norm": 8.440622329711914,
    "prediction_norm": 6.9560394287109375,
    "timestamp": "2025-07-14T12:34:47.694057"
  },
  {
    "epoch": 2051,
    "batch_idx": 160,
    "loss": 0.11800302565097809,
    "learning_rate": 1.6035696863866988e-06,
    "timestep_mean": 517.1875,
    "epsilon_norm": 7.039365291595459,
    "prediction_norm": 6.370757579803467,
    "timestamp": "2025-07-14T12:34:47.699163"
  },
  {
    "epoch": 2051,
    "batch_idx": 192,
    "loss": 0.3019028902053833,
    "learning_rate": 1.6035696863866988e-06,
    "timestep_mean": 420.5,
    "epsilon_norm": 4.385997772216797,
    "prediction_norm": 3.3652961254119873,
    "timestamp": "2025-07-14T12:34:47.704347"
  },
  {
    "epoch": 2052,
    "batch_idx": 0,
    "loss": 0.2447178065776825,
    "learning_rate": 1.667035982985068e-06,
    "timestep_mean": 474.90625,
    "epsilon_norm": 7.157408714294434,
    "prediction_norm": 6.695542812347412,
    "timestamp": "2025-07-14T12:34:47.709495"
  },
  {
    "epoch": 2052,
    "batch_idx": 32,
    "loss": 0.3364834785461426,
    "learning_rate": 1.667035982985068e-06,
    "timestep_mean": 491.875,
    "epsilon_norm": 7.029076099395752,
    "prediction_norm": 5.715651035308838,
    "timestamp": "2025-07-14T12:34:47.714589"
  },
  {
    "epoch": 2052,
    "batch_idx": 64,
    "loss": 0.3005824089050293,
    "learning_rate": 1.667035982985068e-06,
    "timestep_mean": 446.40625,
    "epsilon_norm": 9.304227828979492,
    "prediction_norm": 8.449222564697266,
    "timestamp": "2025-07-14T12:34:47.719565"
  },
  {
    "epoch": 2052,
    "batch_idx": 96,
    "loss": 0.20120374858379364,
    "learning_rate": 1.667035982985068e-06,
    "timestep_mean": 561.875,
    "epsilon_norm": 7.635308742523193,
    "prediction_norm": 6.552736759185791,
    "timestamp": "2025-07-14T12:34:47.724517"
  },
  {
    "epoch": 2052,
    "batch_idx": 128,
    "loss": 0.3071368336677551,
    "learning_rate": 1.667035982985068e-06,
    "timestep_mean": 571.59375,
    "epsilon_norm": 8.87015438079834,
    "prediction_norm": 7.6144700050354,
    "timestamp": "2025-07-14T12:34:47.729820"
  },
  {
    "epoch": 2052,
    "batch_idx": 160,
    "loss": 0.3016144335269928,
    "learning_rate": 1.667035982985068e-06,
    "timestep_mean": 358.59375,
    "epsilon_norm": 8.21152400970459,
    "prediction_norm": 6.651425838470459,
    "timestamp": "2025-07-14T12:34:47.734940"
  },
  {
    "epoch": 2052,
    "batch_idx": 192,
    "loss": 0.1464683711528778,
    "learning_rate": 1.667035982985068e-06,
    "timestep_mean": 541.625,
    "epsilon_norm": 4.053425312042236,
    "prediction_norm": 4.013691425323486,
    "timestamp": "2025-07-14T12:34:47.740012"
  },
  {
    "epoch": 2053,
    "batch_idx": 0,
    "loss": 0.381281316280365,
    "learning_rate": 1.731731866634281e-06,
    "timestep_mean": 457.46875,
    "epsilon_norm": 7.856380462646484,
    "prediction_norm": 5.769469261169434,
    "timestamp": "2025-07-14T12:34:47.745257"
  },
  {
    "epoch": 2053,
    "batch_idx": 32,
    "loss": 0.3952862620353699,
    "learning_rate": 1.731731866634281e-06,
    "timestep_mean": 443.9375,
    "epsilon_norm": 7.5477705001831055,
    "prediction_norm": 6.195513725280762,
    "timestamp": "2025-07-14T12:34:47.750434"
  },
  {
    "epoch": 2053,
    "batch_idx": 64,
    "loss": 0.3742507994174957,
    "learning_rate": 1.731731866634281e-06,
    "timestep_mean": 499.53125,
    "epsilon_norm": 8.284126281738281,
    "prediction_norm": 6.101667404174805,
    "timestamp": "2025-07-14T12:34:47.755572"
  },
  {
    "epoch": 2053,
    "batch_idx": 96,
    "loss": 0.3366307318210602,
    "learning_rate": 1.731731866634281e-06,
    "timestep_mean": 470.40625,
    "epsilon_norm": 8.002901077270508,
    "prediction_norm": 6.8646111488342285,
    "timestamp": "2025-07-14T12:34:47.760707"
  },
  {
    "epoch": 2053,
    "batch_idx": 128,
    "loss": 0.31282490491867065,
    "learning_rate": 1.731731866634281e-06,
    "timestep_mean": 431.15625,
    "epsilon_norm": 8.096155166625977,
    "prediction_norm": 6.1401166915893555,
    "timestamp": "2025-07-14T12:34:47.765845"
  },
  {
    "epoch": 2053,
    "batch_idx": 160,
    "loss": 0.24492725729942322,
    "learning_rate": 1.731731866634281e-06,
    "timestep_mean": 535.90625,
    "epsilon_norm": 7.667312145233154,
    "prediction_norm": 6.4278693199157715,
    "timestamp": "2025-07-14T12:34:47.771070"
  },
  {
    "epoch": 2053,
    "batch_idx": 192,
    "loss": 0.30651652812957764,
    "learning_rate": 1.731731866634281e-06,
    "timestep_mean": 347.0,
    "epsilon_norm": 4.5723557472229,
    "prediction_norm": 3.5926530361175537,
    "timestamp": "2025-07-14T12:34:47.776203"
  },
  {
    "epoch": 2054,
    "batch_idx": 0,
    "loss": 0.1850559562444687,
    "learning_rate": 1.7976571777038066e-06,
    "timestep_mean": 502.28125,
    "epsilon_norm": 6.501579761505127,
    "prediction_norm": 5.9538187980651855,
    "timestamp": "2025-07-14T12:34:47.781297"
  },
  {
    "epoch": 2054,
    "batch_idx": 32,
    "loss": 0.27606451511383057,
    "learning_rate": 1.7976571777038066e-06,
    "timestep_mean": 509.78125,
    "epsilon_norm": 8.379559516906738,
    "prediction_norm": 6.751187801361084,
    "timestamp": "2025-07-14T12:34:47.786594"
  },
  {
    "epoch": 2054,
    "batch_idx": 64,
    "loss": 0.25548362731933594,
    "learning_rate": 1.7976571777038066e-06,
    "timestep_mean": 539.6875,
    "epsilon_norm": 7.603801727294922,
    "prediction_norm": 6.418551445007324,
    "timestamp": "2025-07-14T12:34:47.791859"
  },
  {
    "epoch": 2054,
    "batch_idx": 96,
    "loss": 0.24393978714942932,
    "learning_rate": 1.7976571777038066e-06,
    "timestep_mean": 526.75,
    "epsilon_norm": 7.092438697814941,
    "prediction_norm": 5.372322082519531,
    "timestamp": "2025-07-14T12:34:47.797078"
  },
  {
    "epoch": 2054,
    "batch_idx": 128,
    "loss": 0.23356321454048157,
    "learning_rate": 1.7976571777038066e-06,
    "timestep_mean": 553.90625,
    "epsilon_norm": 8.251564979553223,
    "prediction_norm": 6.847842693328857,
    "timestamp": "2025-07-14T12:34:47.802186"
  },
  {
    "epoch": 2054,
    "batch_idx": 160,
    "loss": 0.22139158844947815,
    "learning_rate": 1.7976571777038066e-06,
    "timestep_mean": 550.125,
    "epsilon_norm": 6.997199058532715,
    "prediction_norm": 5.949737071990967,
    "timestamp": "2025-07-14T12:34:47.807314"
  },
  {
    "epoch": 2054,
    "batch_idx": 192,
    "loss": 0.24079179763793945,
    "learning_rate": 1.7976571777038066e-06,
    "timestep_mean": 568.625,
    "epsilon_norm": 3.6522419452667236,
    "prediction_norm": 3.19411039352417,
    "timestamp": "2025-07-14T12:34:47.812436"
  },
  {
    "epoch": 2055,
    "batch_idx": 0,
    "loss": 0.22285613417625427,
    "learning_rate": 1.8648117535293745e-06,
    "timestep_mean": 568.0,
    "epsilon_norm": 7.695013999938965,
    "prediction_norm": 7.095308303833008,
    "timestamp": "2025-07-14T12:34:47.817761"
  },
  {
    "epoch": 2055,
    "batch_idx": 32,
    "loss": 0.4530991315841675,
    "learning_rate": 1.8648117535293745e-06,
    "timestep_mean": 488.75,
    "epsilon_norm": 9.11437702178955,
    "prediction_norm": 7.083316802978516,
    "timestamp": "2025-07-14T12:34:47.822886"
  },
  {
    "epoch": 2055,
    "batch_idx": 64,
    "loss": 0.44525355100631714,
    "learning_rate": 1.8648117535293745e-06,
    "timestep_mean": 541.71875,
    "epsilon_norm": 7.837055683135986,
    "prediction_norm": 6.336212158203125,
    "timestamp": "2025-07-14T12:34:47.828047"
  },
  {
    "epoch": 2055,
    "batch_idx": 96,
    "loss": 0.24670153856277466,
    "learning_rate": 1.8648117535293745e-06,
    "timestep_mean": 508.875,
    "epsilon_norm": 8.719672203063965,
    "prediction_norm": 7.757266044616699,
    "timestamp": "2025-07-14T12:34:47.833227"
  },
  {
    "epoch": 2055,
    "batch_idx": 128,
    "loss": 0.2837291359901428,
    "learning_rate": 1.8648117535293745e-06,
    "timestep_mean": 581.34375,
    "epsilon_norm": 7.539180755615234,
    "prediction_norm": 6.504469394683838,
    "timestamp": "2025-07-14T12:34:47.838535"
  },
  {
    "epoch": 2055,
    "batch_idx": 160,
    "loss": 0.19937288761138916,
    "learning_rate": 1.8648117535293745e-06,
    "timestep_mean": 556.0625,
    "epsilon_norm": 9.230964660644531,
    "prediction_norm": 8.280115127563477,
    "timestamp": "2025-07-14T12:34:47.843740"
  },
  {
    "epoch": 2055,
    "batch_idx": 192,
    "loss": 0.08208762854337692,
    "learning_rate": 1.8648117535293745e-06,
    "timestep_mean": 644.0,
    "epsilon_norm": 3.595198154449463,
    "prediction_norm": 3.0106170177459717,
    "timestamp": "2025-07-14T12:34:47.848833"
  },
  {
    "epoch": 2056,
    "batch_idx": 0,
    "loss": 0.25819194316864014,
    "learning_rate": 1.93319542841375e-06,
    "timestep_mean": 467.21875,
    "epsilon_norm": 8.459670066833496,
    "prediction_norm": 7.283213138580322,
    "timestamp": "2025-07-14T12:34:47.854099"
  },
  {
    "epoch": 2056,
    "batch_idx": 32,
    "loss": 0.2930455505847931,
    "learning_rate": 1.93319542841375e-06,
    "timestep_mean": 528.46875,
    "epsilon_norm": 6.82397985458374,
    "prediction_norm": 5.263978004455566,
    "timestamp": "2025-07-14T12:34:47.859289"
  },
  {
    "epoch": 2056,
    "batch_idx": 64,
    "loss": 0.38681453466415405,
    "learning_rate": 1.93319542841375e-06,
    "timestep_mean": 491.0,
    "epsilon_norm": 9.063182830810547,
    "prediction_norm": 7.260396480560303,
    "timestamp": "2025-07-14T12:34:47.864443"
  },
  {
    "epoch": 2056,
    "batch_idx": 96,
    "loss": 0.22799403965473175,
    "learning_rate": 1.93319542841375e-06,
    "timestep_mean": 530.4375,
    "epsilon_norm": 7.428537368774414,
    "prediction_norm": 6.802665710449219,
    "timestamp": "2025-07-14T12:34:47.869589"
  },
  {
    "epoch": 2056,
    "batch_idx": 128,
    "loss": 0.2618926167488098,
    "learning_rate": 1.93319542841375e-06,
    "timestep_mean": 507.15625,
    "epsilon_norm": 6.831376075744629,
    "prediction_norm": 5.4619903564453125,
    "timestamp": "2025-07-14T12:34:47.874519"
  },
  {
    "epoch": 2056,
    "batch_idx": 160,
    "loss": 0.18987445533275604,
    "learning_rate": 1.93319542841375e-06,
    "timestep_mean": 453.875,
    "epsilon_norm": 7.954996109008789,
    "prediction_norm": 6.972548007965088,
    "timestamp": "2025-07-14T12:34:47.879713"
  },
  {
    "epoch": 2056,
    "batch_idx": 192,
    "loss": 0.12548398971557617,
    "learning_rate": 1.93319542841375e-06,
    "timestep_mean": 421.125,
    "epsilon_norm": 2.8793110847473145,
    "prediction_norm": 2.8473966121673584,
    "timestamp": "2025-07-14T12:34:47.884794"
  },
  {
    "epoch": 2057,
    "batch_idx": 0,
    "loss": 0.2561458647251129,
    "learning_rate": 2.002808033627126e-06,
    "timestep_mean": 475.5625,
    "epsilon_norm": 6.482403755187988,
    "prediction_norm": 5.454984188079834,
    "timestamp": "2025-07-14T12:34:47.889878"
  },
  {
    "epoch": 2057,
    "batch_idx": 32,
    "loss": 0.19868402183055878,
    "learning_rate": 2.002808033627126e-06,
    "timestep_mean": 473.53125,
    "epsilon_norm": 6.948184013366699,
    "prediction_norm": 5.814813137054443,
    "timestamp": "2025-07-14T12:34:47.895173"
  },
  {
    "epoch": 2057,
    "batch_idx": 64,
    "loss": 0.1797671616077423,
    "learning_rate": 2.002808033627126e-06,
    "timestep_mean": 501.0625,
    "epsilon_norm": 6.8253936767578125,
    "prediction_norm": 6.453794956207275,
    "timestamp": "2025-07-14T12:34:47.900442"
  },
  {
    "epoch": 2057,
    "batch_idx": 96,
    "loss": 0.135679692029953,
    "learning_rate": 2.002808033627126e-06,
    "timestep_mean": 627.1875,
    "epsilon_norm": 8.19377326965332,
    "prediction_norm": 7.768622875213623,
    "timestamp": "2025-07-14T12:34:47.905597"
  },
  {
    "epoch": 2057,
    "batch_idx": 128,
    "loss": 0.16425666213035583,
    "learning_rate": 2.002808033627126e-06,
    "timestep_mean": 593.34375,
    "epsilon_norm": 7.8928751945495605,
    "prediction_norm": 6.9740471839904785,
    "timestamp": "2025-07-14T12:34:47.910723"
  },
  {
    "epoch": 2057,
    "batch_idx": 160,
    "loss": 0.25694823265075684,
    "learning_rate": 2.002808033627126e-06,
    "timestep_mean": 494.15625,
    "epsilon_norm": 6.903834342956543,
    "prediction_norm": 5.802910804748535,
    "timestamp": "2025-07-14T12:34:47.915865"
  },
  {
    "epoch": 2057,
    "batch_idx": 192,
    "loss": 0.28111135959625244,
    "learning_rate": 2.002808033627126e-06,
    "timestep_mean": 369.375,
    "epsilon_norm": 2.570354461669922,
    "prediction_norm": 2.415121555328369,
    "timestamp": "2025-07-14T12:34:47.921050"
  },
  {
    "epoch": 2058,
    "batch_idx": 0,
    "loss": 0.3199327290058136,
    "learning_rate": 2.0736493974071203e-06,
    "timestep_mean": 433.375,
    "epsilon_norm": 7.936645984649658,
    "prediction_norm": 7.191863059997559,
    "timestamp": "2025-07-14T12:34:47.926233"
  },
  {
    "epoch": 2058,
    "batch_idx": 32,
    "loss": 0.4013752043247223,
    "learning_rate": 2.0736493974071203e-06,
    "timestep_mean": 477.78125,
    "epsilon_norm": 8.408870697021484,
    "prediction_norm": 6.822955131530762,
    "timestamp": "2025-07-14T12:34:47.931480"
  },
  {
    "epoch": 2058,
    "batch_idx": 64,
    "loss": 0.15424656867980957,
    "learning_rate": 2.0736493974071203e-06,
    "timestep_mean": 447.59375,
    "epsilon_norm": 7.671952247619629,
    "prediction_norm": 6.7270636558532715,
    "timestamp": "2025-07-14T12:34:47.936623"
  },
  {
    "epoch": 2058,
    "batch_idx": 96,
    "loss": 0.2559577226638794,
    "learning_rate": 2.0736493974071203e-06,
    "timestep_mean": 497.46875,
    "epsilon_norm": 7.974028587341309,
    "prediction_norm": 6.557868003845215,
    "timestamp": "2025-07-14T12:34:47.941954"
  },
  {
    "epoch": 2058,
    "batch_idx": 128,
    "loss": 0.26038461923599243,
    "learning_rate": 2.0736493974071203e-06,
    "timestep_mean": 471.03125,
    "epsilon_norm": 7.987565040588379,
    "prediction_norm": 6.699836730957031,
    "timestamp": "2025-07-14T12:34:47.947150"
  },
  {
    "epoch": 2058,
    "batch_idx": 160,
    "loss": 0.39006543159484863,
    "learning_rate": 2.0736493974071203e-06,
    "timestep_mean": 377.03125,
    "epsilon_norm": 7.808325290679932,
    "prediction_norm": 6.170558929443359,
    "timestamp": "2025-07-14T12:34:47.952255"
  },
  {
    "epoch": 2058,
    "batch_idx": 192,
    "loss": 0.10073492676019669,
    "learning_rate": 2.0736493974071203e-06,
    "timestep_mean": 474.125,
    "epsilon_norm": 3.413006544113159,
    "prediction_norm": 3.0473899841308594,
    "timestamp": "2025-07-14T12:34:47.957314"
  },
  {
    "epoch": 2059,
    "batch_idx": 0,
    "loss": 0.12942847609519958,
    "learning_rate": 2.1457193449599424e-06,
    "timestep_mean": 534.59375,
    "epsilon_norm": 7.014190673828125,
    "prediction_norm": 5.640792369842529,
    "timestamp": "2025-07-14T12:34:47.962505"
  },
  {
    "epoch": 2059,
    "batch_idx": 32,
    "loss": 0.11659064888954163,
    "learning_rate": 2.1457193449599424e-06,
    "timestep_mean": 588.09375,
    "epsilon_norm": 8.805484771728516,
    "prediction_norm": 8.009563446044922,
    "timestamp": "2025-07-14T12:34:47.967854"
  },
  {
    "epoch": 2059,
    "batch_idx": 64,
    "loss": 0.19839045405387878,
    "learning_rate": 2.1457193449599424e-06,
    "timestep_mean": 585.3125,
    "epsilon_norm": 8.025755882263184,
    "prediction_norm": 7.532150745391846,
    "timestamp": "2025-07-14T12:34:47.973004"
  },
  {
    "epoch": 2059,
    "batch_idx": 96,
    "loss": 0.18329501152038574,
    "learning_rate": 2.1457193449599424e-06,
    "timestep_mean": 590.125,
    "epsilon_norm": 7.71552848815918,
    "prediction_norm": 6.753676414489746,
    "timestamp": "2025-07-14T12:34:47.978129"
  },
  {
    "epoch": 2059,
    "batch_idx": 128,
    "loss": 0.16031332314014435,
    "learning_rate": 2.1457193449599424e-06,
    "timestep_mean": 568.4375,
    "epsilon_norm": 8.031384468078613,
    "prediction_norm": 7.2714080810546875,
    "timestamp": "2025-07-14T12:34:47.983290"
  },
  {
    "epoch": 2059,
    "batch_idx": 160,
    "loss": 0.29733365774154663,
    "learning_rate": 2.1457193449599424e-06,
    "timestep_mean": 567.8125,
    "epsilon_norm": 8.487512588500977,
    "prediction_norm": 6.786543846130371,
    "timestamp": "2025-07-14T12:34:47.988660"
  },
  {
    "epoch": 2059,
    "batch_idx": 192,
    "loss": 0.48528409004211426,
    "learning_rate": 2.1457193449599424e-06,
    "timestep_mean": 287.0,
    "epsilon_norm": 3.668187379837036,
    "prediction_norm": 2.4178051948547363,
    "timestamp": "2025-07-14T12:34:47.993749"
  },
  {
    "epoch": 2060,
    "batch_idx": 0,
    "loss": 0.2003629207611084,
    "learning_rate": 2.2190176984600044e-06,
    "timestep_mean": 494.09375,
    "epsilon_norm": 8.2265625,
    "prediction_norm": 7.471122741699219,
    "timestamp": "2025-07-14T12:34:47.998916"
  },
  {
    "epoch": 2060,
    "batch_idx": 32,
    "loss": 0.2876073122024536,
    "learning_rate": 2.2190176984600044e-06,
    "timestep_mean": 431.875,
    "epsilon_norm": 8.123126983642578,
    "prediction_norm": 6.560241222381592,
    "timestamp": "2025-07-14T12:34:48.004042"
  },
  {
    "epoch": 2060,
    "batch_idx": 64,
    "loss": 0.2670089900493622,
    "learning_rate": 2.2190176984600044e-06,
    "timestep_mean": 483.1875,
    "epsilon_norm": 8.246298789978027,
    "prediction_norm": 7.2649664878845215,
    "timestamp": "2025-07-14T12:34:48.009269"
  },
  {
    "epoch": 2060,
    "batch_idx": 96,
    "loss": 0.11825115978717804,
    "learning_rate": 2.2190176984600044e-06,
    "timestep_mean": 540.0,
    "epsilon_norm": 7.84596061706543,
    "prediction_norm": 7.2131757736206055,
    "timestamp": "2025-07-14T12:34:48.014360"
  },
  {
    "epoch": 2060,
    "batch_idx": 128,
    "loss": 0.280396044254303,
    "learning_rate": 2.2190176984600044e-06,
    "timestep_mean": 411.15625,
    "epsilon_norm": 7.147276878356934,
    "prediction_norm": 5.217519283294678,
    "timestamp": "2025-07-14T12:34:48.019571"
  },
  {
    "epoch": 2060,
    "batch_idx": 160,
    "loss": 0.2600330710411072,
    "learning_rate": 2.2190176984600044e-06,
    "timestep_mean": 463.9375,
    "epsilon_norm": 7.224470138549805,
    "prediction_norm": 6.210673809051514,
    "timestamp": "2025-07-14T12:34:48.024721"
  },
  {
    "epoch": 2060,
    "batch_idx": 192,
    "loss": 0.15669777989387512,
    "learning_rate": 2.2190176984600044e-06,
    "timestep_mean": 494.0,
    "epsilon_norm": 2.778733968734741,
    "prediction_norm": 2.639707088470459,
    "timestamp": "2025-07-14T12:34:48.029993"
  },
  {
    "epoch": 2061,
    "batch_idx": 0,
    "loss": 0.2550591230392456,
    "learning_rate": 2.2935442770509212e-06,
    "timestep_mean": 439.625,
    "epsilon_norm": 7.831594467163086,
    "prediction_norm": 7.050889015197754,
    "timestamp": "2025-07-14T12:34:48.034990"
  },
  {
    "epoch": 2061,
    "batch_idx": 32,
    "loss": 0.34498876333236694,
    "learning_rate": 2.2935442770509212e-06,
    "timestep_mean": 417.21875,
    "epsilon_norm": 8.683622360229492,
    "prediction_norm": 7.0531907081604,
    "timestamp": "2025-07-14T12:34:48.040009"
  },
  {
    "epoch": 2061,
    "batch_idx": 64,
    "loss": 0.18419155478477478,
    "learning_rate": 2.2935442770509212e-06,
    "timestep_mean": 537.03125,
    "epsilon_norm": 7.89256477355957,
    "prediction_norm": 7.086679458618164,
    "timestamp": "2025-07-14T12:34:48.045094"
  },
  {
    "epoch": 2061,
    "batch_idx": 96,
    "loss": 0.21376848220825195,
    "learning_rate": 2.2935442770509212e-06,
    "timestep_mean": 429.25,
    "epsilon_norm": 6.779323577880859,
    "prediction_norm": 5.71073579788208,
    "timestamp": "2025-07-14T12:34:48.050233"
  },
  {
    "epoch": 2061,
    "batch_idx": 128,
    "loss": 0.21236452460289001,
    "learning_rate": 2.2935442770509212e-06,
    "timestep_mean": 474.9375,
    "epsilon_norm": 7.597095966339111,
    "prediction_norm": 6.788601875305176,
    "timestamp": "2025-07-14T12:34:48.055347"
  },
  {
    "epoch": 2061,
    "batch_idx": 160,
    "loss": 0.28851449489593506,
    "learning_rate": 2.2935442770509212e-06,
    "timestep_mean": 446.34375,
    "epsilon_norm": 8.038883209228516,
    "prediction_norm": 7.050688743591309,
    "timestamp": "2025-07-14T12:34:48.060357"
  },
  {
    "epoch": 2061,
    "batch_idx": 192,
    "loss": 0.287028968334198,
    "learning_rate": 2.2935442770509212e-06,
    "timestep_mean": 328.75,
    "epsilon_norm": 4.277395248413086,
    "prediction_norm": 3.154029130935669,
    "timestamp": "2025-07-14T12:34:48.065292"
  },
  {
    "epoch": 2062,
    "batch_idx": 0,
    "loss": 0.41632550954818726,
    "learning_rate": 2.3692988968458426e-06,
    "timestep_mean": 490.8125,
    "epsilon_norm": 7.690074920654297,
    "prediction_norm": 5.642953872680664,
    "timestamp": "2025-07-14T12:34:48.070454"
  },
  {
    "epoch": 2062,
    "batch_idx": 32,
    "loss": 0.23808367550373077,
    "learning_rate": 2.3692988968458426e-06,
    "timestep_mean": 541.78125,
    "epsilon_norm": 7.592187404632568,
    "prediction_norm": 6.3731489181518555,
    "timestamp": "2025-07-14T12:34:48.075693"
  },
  {
    "epoch": 2062,
    "batch_idx": 64,
    "loss": 0.19615918397903442,
    "learning_rate": 2.3692988968458426e-06,
    "timestep_mean": 538.4375,
    "epsilon_norm": 8.284687995910645,
    "prediction_norm": 7.020633697509766,
    "timestamp": "2025-07-14T12:34:48.080815"
  },
  {
    "epoch": 2062,
    "batch_idx": 96,
    "loss": 0.3722975254058838,
    "learning_rate": 2.3692988968458426e-06,
    "timestep_mean": 505.78125,
    "epsilon_norm": 7.377840518951416,
    "prediction_norm": 5.578924655914307,
    "timestamp": "2025-07-14T12:34:48.085912"
  },
  {
    "epoch": 2062,
    "batch_idx": 128,
    "loss": 0.21883639693260193,
    "learning_rate": 2.3692988968458426e-06,
    "timestep_mean": 544.71875,
    "epsilon_norm": 7.73562479019165,
    "prediction_norm": 6.822762966156006,
    "timestamp": "2025-07-14T12:34:48.090770"
  },
  {
    "epoch": 2062,
    "batch_idx": 160,
    "loss": 0.12789961695671082,
    "learning_rate": 2.3692988968458426e-06,
    "timestep_mean": 568.6875,
    "epsilon_norm": 8.351130485534668,
    "prediction_norm": 8.071052551269531,
    "timestamp": "2025-07-14T12:34:48.095560"
  },
  {
    "epoch": 2062,
    "batch_idx": 192,
    "loss": 0.131077840924263,
    "learning_rate": 2.3692988968458426e-06,
    "timestep_mean": 491.125,
    "epsilon_norm": 3.3248450756073,
    "prediction_norm": 2.968311071395874,
    "timestamp": "2025-07-14T12:34:48.100509"
  },
  {
    "epoch": 2063,
    "batch_idx": 0,
    "loss": 0.23516309261322021,
    "learning_rate": 2.4462813709276753e-06,
    "timestep_mean": 497.46875,
    "epsilon_norm": 8.2920503616333,
    "prediction_norm": 6.711633682250977,
    "timestamp": "2025-07-14T12:34:48.105334"
  },
  {
    "epoch": 2063,
    "batch_idx": 32,
    "loss": 0.15997186303138733,
    "learning_rate": 2.4462813709276753e-06,
    "timestep_mean": 523.0625,
    "epsilon_norm": 7.21133279800415,
    "prediction_norm": 5.812430381774902,
    "timestamp": "2025-07-14T12:34:48.110183"
  },
  {
    "epoch": 2063,
    "batch_idx": 64,
    "loss": 0.29438918828964233,
    "learning_rate": 2.4462813709276753e-06,
    "timestep_mean": 463.5625,
    "epsilon_norm": 8.748876571655273,
    "prediction_norm": 7.255599021911621,
    "timestamp": "2025-07-14T12:34:48.114990"
  },
  {
    "epoch": 2063,
    "batch_idx": 96,
    "loss": 0.3005639314651489,
    "learning_rate": 2.4462813709276753e-06,
    "timestep_mean": 511.25,
    "epsilon_norm": 7.767084121704102,
    "prediction_norm": 6.219977855682373,
    "timestamp": "2025-07-14T12:34:48.119982"
  },
  {
    "epoch": 2063,
    "batch_idx": 128,
    "loss": 0.28485387563705444,
    "learning_rate": 2.4462813709276753e-06,
    "timestep_mean": 556.40625,
    "epsilon_norm": 7.5922770500183105,
    "prediction_norm": 6.54973840713501,
    "timestamp": "2025-07-14T12:34:48.125164"
  },
  {
    "epoch": 2063,
    "batch_idx": 160,
    "loss": 0.34106284379959106,
    "learning_rate": 2.4462813709276753e-06,
    "timestep_mean": 449.21875,
    "epsilon_norm": 8.607512474060059,
    "prediction_norm": 6.423291206359863,
    "timestamp": "2025-07-14T12:34:48.130312"
  },
  {
    "epoch": 2063,
    "batch_idx": 192,
    "loss": 1.3085163831710815,
    "learning_rate": 2.4462813709276753e-06,
    "timestep_mean": 356.875,
    "epsilon_norm": 5.412893295288086,
    "prediction_norm": 2.948316812515259,
    "timestamp": "2025-07-14T12:34:48.135374"
  },
  {
    "epoch": 2064,
    "batch_idx": 0,
    "loss": 0.24210605025291443,
    "learning_rate": 2.5244915093499162e-06,
    "timestep_mean": 504.53125,
    "epsilon_norm": 7.539015293121338,
    "prediction_norm": 6.331808090209961,
    "timestamp": "2025-07-14T12:34:48.140410"
  },
  {
    "epoch": 2064,
    "batch_idx": 32,
    "loss": 0.12451991438865662,
    "learning_rate": 2.5244915093499162e-06,
    "timestep_mean": 523.78125,
    "epsilon_norm": 6.904049396514893,
    "prediction_norm": 6.515994071960449,
    "timestamp": "2025-07-14T12:34:48.145562"
  },
  {
    "epoch": 2064,
    "batch_idx": 64,
    "loss": 0.33081284165382385,
    "learning_rate": 2.5244915093499162e-06,
    "timestep_mean": 487.5625,
    "epsilon_norm": 7.795504570007324,
    "prediction_norm": 5.761646747589111,
    "timestamp": "2025-07-14T12:34:48.150721"
  },
  {
    "epoch": 2064,
    "batch_idx": 96,
    "loss": 0.342678427696228,
    "learning_rate": 2.5244915093499162e-06,
    "timestep_mean": 478.6875,
    "epsilon_norm": 8.85317611694336,
    "prediction_norm": 7.07001256942749,
    "timestamp": "2025-07-14T12:34:48.155772"
  },
  {
    "epoch": 2064,
    "batch_idx": 128,
    "loss": 0.3480519652366638,
    "learning_rate": 2.5244915093499162e-06,
    "timestep_mean": 387.5625,
    "epsilon_norm": 8.232259750366211,
    "prediction_norm": 6.6791605949401855,
    "timestamp": "2025-07-14T12:34:48.160865"
  },
  {
    "epoch": 2064,
    "batch_idx": 160,
    "loss": 0.32823446393013,
    "learning_rate": 2.5244915093499162e-06,
    "timestep_mean": 455.375,
    "epsilon_norm": 7.713239669799805,
    "prediction_norm": 6.282871723175049,
    "timestamp": "2025-07-14T12:34:48.165974"
  },
  {
    "epoch": 2064,
    "batch_idx": 192,
    "loss": 0.22932149469852448,
    "learning_rate": 2.5244915093499162e-06,
    "timestep_mean": 591.875,
    "epsilon_norm": 4.253199577331543,
    "prediction_norm": 3.492340087890625,
    "timestamp": "2025-07-14T12:34:48.171103"
  },
  {
    "epoch": 2065,
    "batch_idx": 0,
    "loss": 0.3417445421218872,
    "learning_rate": 2.603929119136764e-06,
    "timestep_mean": 578.3125,
    "epsilon_norm": 8.699322700500488,
    "prediction_norm": 7.567726135253906,
    "timestamp": "2025-07-14T12:34:48.176310"
  },
  {
    "epoch": 2065,
    "batch_idx": 32,
    "loss": 0.20828622579574585,
    "learning_rate": 2.603929119136764e-06,
    "timestep_mean": 554.875,
    "epsilon_norm": 7.822032451629639,
    "prediction_norm": 6.745060443878174,
    "timestamp": "2025-07-14T12:34:48.181386"
  },
  {
    "epoch": 2065,
    "batch_idx": 64,
    "loss": 0.2390090376138687,
    "learning_rate": 2.603929119136764e-06,
    "timestep_mean": 527.6875,
    "epsilon_norm": 7.388711452484131,
    "prediction_norm": 6.277665615081787,
    "timestamp": "2025-07-14T12:34:48.186573"
  },
  {
    "epoch": 2065,
    "batch_idx": 96,
    "loss": 0.3546162247657776,
    "learning_rate": 2.603929119136764e-06,
    "timestep_mean": 471.75,
    "epsilon_norm": 7.267488956451416,
    "prediction_norm": 6.377961158752441,
    "timestamp": "2025-07-14T12:34:48.191703"
  },
  {
    "epoch": 2065,
    "batch_idx": 128,
    "loss": 0.2094452977180481,
    "learning_rate": 2.603929119136764e-06,
    "timestep_mean": 450.59375,
    "epsilon_norm": 6.791502952575684,
    "prediction_norm": 6.088021278381348,
    "timestamp": "2025-07-14T12:34:48.196983"
  },
  {
    "epoch": 2065,
    "batch_idx": 160,
    "loss": 0.3820571005344391,
    "learning_rate": 2.603929119136764e-06,
    "timestep_mean": 489.3125,
    "epsilon_norm": 8.753180503845215,
    "prediction_norm": 6.864710330963135,
    "timestamp": "2025-07-14T12:34:48.202154"
  },
  {
    "epoch": 2065,
    "batch_idx": 192,
    "loss": 0.38808146119117737,
    "learning_rate": 2.603929119136764e-06,
    "timestep_mean": 217.625,
    "epsilon_norm": 4.382523059844971,
    "prediction_norm": 3.7037668228149414,
    "timestamp": "2025-07-14T12:34:48.207131"
  },
  {
    "epoch": 2066,
    "batch_idx": 0,
    "loss": 0.27902287244796753,
    "learning_rate": 2.684594004283783e-06,
    "timestep_mean": 471.3125,
    "epsilon_norm": 8.287980079650879,
    "prediction_norm": 6.680689811706543,
    "timestamp": "2025-07-14T12:34:48.212404"
  },
  {
    "epoch": 2066,
    "batch_idx": 32,
    "loss": 0.3863959014415741,
    "learning_rate": 2.684594004283783e-06,
    "timestep_mean": 424.375,
    "epsilon_norm": 7.751304626464844,
    "prediction_norm": 6.285149097442627,
    "timestamp": "2025-07-14T12:34:48.217626"
  },
  {
    "epoch": 2066,
    "batch_idx": 64,
    "loss": 0.14867831766605377,
    "learning_rate": 2.684594004283783e-06,
    "timestep_mean": 554.8125,
    "epsilon_norm": 7.309908866882324,
    "prediction_norm": 6.8195013999938965,
    "timestamp": "2025-07-14T12:34:48.222726"
  },
  {
    "epoch": 2066,
    "batch_idx": 96,
    "loss": 0.13875797390937805,
    "learning_rate": 2.684594004283783e-06,
    "timestep_mean": 578.15625,
    "epsilon_norm": 7.898808002471924,
    "prediction_norm": 7.4576544761657715,
    "timestamp": "2025-07-14T12:34:48.227937"
  },
  {
    "epoch": 2066,
    "batch_idx": 128,
    "loss": 0.36216604709625244,
    "learning_rate": 2.684594004283783e-06,
    "timestep_mean": 484.84375,
    "epsilon_norm": 7.849037170410156,
    "prediction_norm": 5.603626251220703,
    "timestamp": "2025-07-14T12:34:48.233090"
  },
  {
    "epoch": 2066,
    "batch_idx": 160,
    "loss": 0.3392987549304962,
    "learning_rate": 2.684594004283783e-06,
    "timestep_mean": 490.03125,
    "epsilon_norm": 8.402709007263184,
    "prediction_norm": 6.974860668182373,
    "timestamp": "2025-07-14T12:34:48.238456"
  },
  {
    "epoch": 2066,
    "batch_idx": 192,
    "loss": 0.5510070323944092,
    "learning_rate": 2.684594004283783e-06,
    "timestep_mean": 400.75,
    "epsilon_norm": 4.295605182647705,
    "prediction_norm": 2.735041856765747,
    "timestamp": "2025-07-14T12:34:48.243672"
  },
  {
    "epoch": 2067,
    "batch_idx": 0,
    "loss": 0.2557034492492676,
    "learning_rate": 2.766485965758517e-06,
    "timestep_mean": 432.03125,
    "epsilon_norm": 7.635479927062988,
    "prediction_norm": 6.030629634857178,
    "timestamp": "2025-07-14T12:34:48.248793"
  },
  {
    "epoch": 2067,
    "batch_idx": 32,
    "loss": 0.2663704752922058,
    "learning_rate": 2.766485965758517e-06,
    "timestep_mean": 563.15625,
    "epsilon_norm": 8.102474212646484,
    "prediction_norm": 7.094755172729492,
    "timestamp": "2025-07-14T12:34:48.254044"
  },
  {
    "epoch": 2067,
    "batch_idx": 64,
    "loss": 0.20743310451507568,
    "learning_rate": 2.766485965758517e-06,
    "timestep_mean": 529.1875,
    "epsilon_norm": 8.465314865112305,
    "prediction_norm": 7.220807075500488,
    "timestamp": "2025-07-14T12:34:48.259290"
  },
  {
    "epoch": 2067,
    "batch_idx": 96,
    "loss": 0.3140336871147156,
    "learning_rate": 2.766485965758517e-06,
    "timestep_mean": 543.46875,
    "epsilon_norm": 8.503060340881348,
    "prediction_norm": 7.528323173522949,
    "timestamp": "2025-07-14T12:34:48.264467"
  },
  {
    "epoch": 2067,
    "batch_idx": 128,
    "loss": 0.11486665159463882,
    "learning_rate": 2.766485965758517e-06,
    "timestep_mean": 564.0,
    "epsilon_norm": 8.484648704528809,
    "prediction_norm": 8.161876678466797,
    "timestamp": "2025-07-14T12:34:48.269643"
  },
  {
    "epoch": 2067,
    "batch_idx": 160,
    "loss": 0.23316653072834015,
    "learning_rate": 2.766485965758517e-06,
    "timestep_mean": 586.0625,
    "epsilon_norm": 8.556766510009766,
    "prediction_norm": 7.512660503387451,
    "timestamp": "2025-07-14T12:34:48.274608"
  },
  {
    "epoch": 2067,
    "batch_idx": 192,
    "loss": 0.15564891695976257,
    "learning_rate": 2.766485965758517e-06,
    "timestep_mean": 431.75,
    "epsilon_norm": 3.571366310119629,
    "prediction_norm": 3.533787250518799,
    "timestamp": "2025-07-14T12:34:48.279913"
  },
  {
    "epoch": 2068,
    "batch_idx": 0,
    "loss": 0.30017006397247314,
    "learning_rate": 2.8496048015005416e-06,
    "timestep_mean": 501.6875,
    "epsilon_norm": 8.205002784729004,
    "prediction_norm": 7.126302719116211,
    "timestamp": "2025-07-14T12:34:48.284995"
  },
  {
    "epoch": 2068,
    "batch_idx": 32,
    "loss": 0.38990747928619385,
    "learning_rate": 2.8496048015005416e-06,
    "timestep_mean": 583.375,
    "epsilon_norm": 8.763049125671387,
    "prediction_norm": 7.579427242279053,
    "timestamp": "2025-07-14T12:34:48.290032"
  },
  {
    "epoch": 2068,
    "batch_idx": 64,
    "loss": 0.16155776381492615,
    "learning_rate": 2.8496048015005416e-06,
    "timestep_mean": 507.9375,
    "epsilon_norm": 7.256934642791748,
    "prediction_norm": 6.106733322143555,
    "timestamp": "2025-07-14T12:34:48.294942"
  },
  {
    "epoch": 2068,
    "batch_idx": 96,
    "loss": 0.5069133043289185,
    "learning_rate": 2.8496048015005416e-06,
    "timestep_mean": 531.375,
    "epsilon_norm": 8.409050941467285,
    "prediction_norm": 6.169069290161133,
    "timestamp": "2025-07-14T12:34:48.299991"
  },
  {
    "epoch": 2068,
    "batch_idx": 128,
    "loss": 0.2728324234485626,
    "learning_rate": 2.8496048015005416e-06,
    "timestep_mean": 532.8125,
    "epsilon_norm": 8.248457908630371,
    "prediction_norm": 6.790984153747559,
    "timestamp": "2025-07-14T12:34:48.305263"
  },
  {
    "epoch": 2068,
    "batch_idx": 160,
    "loss": 0.42241743206977844,
    "learning_rate": 2.8496048015005416e-06,
    "timestep_mean": 442.5625,
    "epsilon_norm": 8.670236587524414,
    "prediction_norm": 7.113818645477295,
    "timestamp": "2025-07-14T12:34:48.310387"
  },
  {
    "epoch": 2068,
    "batch_idx": 192,
    "loss": 0.1877475529909134,
    "learning_rate": 2.8496048015005416e-06,
    "timestep_mean": 430.125,
    "epsilon_norm": 3.7901508808135986,
    "prediction_norm": 3.238313674926758,
    "timestamp": "2025-07-14T12:34:48.315472"
  },
  {
    "epoch": 2069,
    "batch_idx": 0,
    "loss": 0.13625650107860565,
    "learning_rate": 2.933950306422465e-06,
    "timestep_mean": 619.40625,
    "epsilon_norm": 8.306520462036133,
    "prediction_norm": 8.160019874572754,
    "timestamp": "2025-07-14T12:34:48.320678"
  },
  {
    "epoch": 2069,
    "batch_idx": 32,
    "loss": 0.2090483009815216,
    "learning_rate": 2.933950306422465e-06,
    "timestep_mean": 491.90625,
    "epsilon_norm": 7.901076793670654,
    "prediction_norm": 6.195827007293701,
    "timestamp": "2025-07-14T12:34:48.325833"
  },
  {
    "epoch": 2069,
    "batch_idx": 64,
    "loss": 0.34694749116897583,
    "learning_rate": 2.933950306422465e-06,
    "timestep_mean": 591.375,
    "epsilon_norm": 7.559941291809082,
    "prediction_norm": 5.79580020904541,
    "timestamp": "2025-07-14T12:34:48.330903"
  },
  {
    "epoch": 2069,
    "batch_idx": 96,
    "loss": 0.1424245536327362,
    "learning_rate": 2.933950306422465e-06,
    "timestep_mean": 470.21875,
    "epsilon_norm": 6.689529895782471,
    "prediction_norm": 5.875269889831543,
    "timestamp": "2025-07-14T12:34:48.336097"
  },
  {
    "epoch": 2069,
    "batch_idx": 128,
    "loss": 0.22997614741325378,
    "learning_rate": 2.933950306422465e-06,
    "timestep_mean": 444.0625,
    "epsilon_norm": 7.1994147300720215,
    "prediction_norm": 5.939748764038086,
    "timestamp": "2025-07-14T12:34:48.341222"
  },
  {
    "epoch": 2069,
    "batch_idx": 160,
    "loss": 0.2827039361000061,
    "learning_rate": 2.933950306422465e-06,
    "timestep_mean": 392.78125,
    "epsilon_norm": 7.4046854972839355,
    "prediction_norm": 6.738635540008545,
    "timestamp": "2025-07-14T12:34:48.346570"
  },
  {
    "epoch": 2069,
    "batch_idx": 192,
    "loss": 0.11878149956464767,
    "learning_rate": 2.933950306422465e-06,
    "timestep_mean": 660.0,
    "epsilon_norm": 3.396148204803467,
    "prediction_norm": 3.256359577178955,
    "timestamp": "2025-07-14T12:34:48.351665"
  },
  {
    "epoch": 2070,
    "batch_idx": 0,
    "loss": 0.2529388666152954,
    "learning_rate": 3.0195222724101507e-06,
    "timestep_mean": 511.40625,
    "epsilon_norm": 8.78860855102539,
    "prediction_norm": 7.622913837432861,
    "timestamp": "2025-07-14T12:34:48.356755"
  },
  {
    "epoch": 2070,
    "batch_idx": 32,
    "loss": 0.26789793372154236,
    "learning_rate": 3.0195222724101507e-06,
    "timestep_mean": 508.625,
    "epsilon_norm": 8.484403610229492,
    "prediction_norm": 7.0509033203125,
    "timestamp": "2025-07-14T12:34:48.361991"
  },
  {
    "epoch": 2070,
    "batch_idx": 64,
    "loss": 0.1822848916053772,
    "learning_rate": 3.0195222724101507e-06,
    "timestep_mean": 469.59375,
    "epsilon_norm": 7.587179660797119,
    "prediction_norm": 6.985199451446533,
    "timestamp": "2025-07-14T12:34:48.367261"
  },
  {
    "epoch": 2070,
    "batch_idx": 96,
    "loss": 0.18447577953338623,
    "learning_rate": 3.0195222724101507e-06,
    "timestep_mean": 445.78125,
    "epsilon_norm": 7.599787712097168,
    "prediction_norm": 6.176187992095947,
    "timestamp": "2025-07-14T12:34:48.372380"
  },
  {
    "epoch": 2070,
    "batch_idx": 128,
    "loss": 0.1414252519607544,
    "learning_rate": 3.0195222724101507e-06,
    "timestep_mean": 568.21875,
    "epsilon_norm": 7.999392509460449,
    "prediction_norm": 7.743504524230957,
    "timestamp": "2025-07-14T12:34:48.377253"
  },
  {
    "epoch": 2070,
    "batch_idx": 160,
    "loss": 0.275942325592041,
    "learning_rate": 3.0195222724101507e-06,
    "timestep_mean": 481.78125,
    "epsilon_norm": 8.242849349975586,
    "prediction_norm": 7.202282428741455,
    "timestamp": "2025-07-14T12:34:48.382252"
  },
  {
    "epoch": 2070,
    "batch_idx": 192,
    "loss": 0.4525328278541565,
    "learning_rate": 3.0195222724101507e-06,
    "timestep_mean": 412.5,
    "epsilon_norm": 4.292324542999268,
    "prediction_norm": 3.177804946899414,
    "timestamp": "2025-07-14T12:34:48.387317"
  },
  {
    "epoch": 2071,
    "batch_idx": 0,
    "loss": 0.14535731077194214,
    "learning_rate": 3.1063204883232706e-06,
    "timestep_mean": 624.65625,
    "epsilon_norm": 7.910338401794434,
    "prediction_norm": 7.486793041229248,
    "timestamp": "2025-07-14T12:34:48.392602"
  },
  {
    "epoch": 2071,
    "batch_idx": 32,
    "loss": 0.4100560247898102,
    "learning_rate": 3.1063204883232706e-06,
    "timestep_mean": 489.375,
    "epsilon_norm": 8.300671577453613,
    "prediction_norm": 6.98557186126709,
    "timestamp": "2025-07-14T12:34:48.397729"
  },
  {
    "epoch": 2071,
    "batch_idx": 64,
    "loss": 0.676537036895752,
    "learning_rate": 3.1063204883232706e-06,
    "timestep_mean": 450.59375,
    "epsilon_norm": 9.909399032592773,
    "prediction_norm": 6.974056720733643,
    "timestamp": "2025-07-14T12:34:48.402626"
  },
  {
    "epoch": 2071,
    "batch_idx": 96,
    "loss": 0.286587655544281,
    "learning_rate": 3.1063204883232706e-06,
    "timestep_mean": 572.65625,
    "epsilon_norm": 7.914554595947266,
    "prediction_norm": 7.16878604888916,
    "timestamp": "2025-07-14T12:34:48.407537"
  },
  {
    "epoch": 2071,
    "batch_idx": 128,
    "loss": 0.16984249651432037,
    "learning_rate": 3.1063204883232706e-06,
    "timestep_mean": 452.5,
    "epsilon_norm": 8.274645805358887,
    "prediction_norm": 7.615505218505859,
    "timestamp": "2025-07-14T12:34:48.412351"
  },
  {
    "epoch": 2071,
    "batch_idx": 160,
    "loss": 0.30813390016555786,
    "learning_rate": 3.1063204883232706e-06,
    "timestep_mean": 518.28125,
    "epsilon_norm": 7.239374160766602,
    "prediction_norm": 5.695877552032471,
    "timestamp": "2025-07-14T12:34:48.417564"
  },
  {
    "epoch": 2071,
    "batch_idx": 192,
    "loss": 0.42090529203414917,
    "learning_rate": 3.1063204883232706e-06,
    "timestep_mean": 460.875,
    "epsilon_norm": 4.725797176361084,
    "prediction_norm": 3.018012762069702,
    "timestamp": "2025-07-14T12:34:48.422617"
  },
  {
    "epoch": 2072,
    "batch_idx": 0,
    "loss": 0.2665477693080902,
    "learning_rate": 3.1943447399958065e-06,
    "timestep_mean": 530.625,
    "epsilon_norm": 9.533895492553711,
    "prediction_norm": 7.980255603790283,
    "timestamp": "2025-07-14T12:34:48.427741"
  },
  {
    "epoch": 2072,
    "batch_idx": 32,
    "loss": 0.29801368713378906,
    "learning_rate": 3.1943447399958065e-06,
    "timestep_mean": 564.5,
    "epsilon_norm": 8.367046356201172,
    "prediction_norm": 7.655715465545654,
    "timestamp": "2025-07-14T12:34:48.432756"
  },
  {
    "epoch": 2072,
    "batch_idx": 64,
    "loss": 0.20207396149635315,
    "learning_rate": 3.1943447399958065e-06,
    "timestep_mean": 477.59375,
    "epsilon_norm": 8.28209400177002,
    "prediction_norm": 7.647837162017822,
    "timestamp": "2025-07-14T12:34:48.437925"
  },
  {
    "epoch": 2072,
    "batch_idx": 96,
    "loss": 0.227277934551239,
    "learning_rate": 3.1943447399958065e-06,
    "timestep_mean": 527.78125,
    "epsilon_norm": 7.869251251220703,
    "prediction_norm": 6.706872463226318,
    "timestamp": "2025-07-14T12:34:48.443043"
  },
  {
    "epoch": 2072,
    "batch_idx": 128,
    "loss": 0.2211451381444931,
    "learning_rate": 3.1943447399958065e-06,
    "timestep_mean": 613.15625,
    "epsilon_norm": 7.561808109283447,
    "prediction_norm": 6.86402702331543,
    "timestamp": "2025-07-14T12:34:48.448086"
  },
  {
    "epoch": 2072,
    "batch_idx": 160,
    "loss": 0.222611665725708,
    "learning_rate": 3.1943447399958065e-06,
    "timestep_mean": 474.09375,
    "epsilon_norm": 7.65997838973999,
    "prediction_norm": 6.620434761047363,
    "timestamp": "2025-07-14T12:34:48.453209"
  },
  {
    "epoch": 2072,
    "batch_idx": 192,
    "loss": 0.22114840149879456,
    "learning_rate": 3.1943447399958065e-06,
    "timestep_mean": 514.0,
    "epsilon_norm": 4.894209861755371,
    "prediction_norm": 4.428941249847412,
    "timestamp": "2025-07-14T12:34:48.458521"
  },
  {
    "epoch": 2073,
    "batch_idx": 0,
    "loss": 0.49283164739608765,
    "learning_rate": 3.2835948102367163e-06,
    "timestep_mean": 458.75,
    "epsilon_norm": 9.006216049194336,
    "prediction_norm": 6.804218292236328,
    "timestamp": "2025-07-14T12:34:48.463647"
  },
  {
    "epoch": 2073,
    "batch_idx": 32,
    "loss": 0.19251984357833862,
    "learning_rate": 3.2835948102367163e-06,
    "timestep_mean": 454.09375,
    "epsilon_norm": 8.064576148986816,
    "prediction_norm": 7.028006553649902,
    "timestamp": "2025-07-14T12:34:48.468807"
  },
  {
    "epoch": 2073,
    "batch_idx": 64,
    "loss": 0.39655736088752747,
    "learning_rate": 3.2835948102367163e-06,
    "timestep_mean": 489.75,
    "epsilon_norm": 8.179819107055664,
    "prediction_norm": 6.509337425231934,
    "timestamp": "2025-07-14T12:34:48.473852"
  },
  {
    "epoch": 2073,
    "batch_idx": 96,
    "loss": 0.3343738317489624,
    "learning_rate": 3.2835948102367163e-06,
    "timestep_mean": 521.78125,
    "epsilon_norm": 7.583380222320557,
    "prediction_norm": 5.984016418457031,
    "timestamp": "2025-07-14T12:34:48.479157"
  },
  {
    "epoch": 2073,
    "batch_idx": 128,
    "loss": 0.17948028445243835,
    "learning_rate": 3.2835948102367163e-06,
    "timestep_mean": 530.46875,
    "epsilon_norm": 7.542649745941162,
    "prediction_norm": 6.465890884399414,
    "timestamp": "2025-07-14T12:34:48.484522"
  },
  {
    "epoch": 2073,
    "batch_idx": 160,
    "loss": 0.19393596053123474,
    "learning_rate": 3.2835948102367163e-06,
    "timestep_mean": 466.8125,
    "epsilon_norm": 7.034757614135742,
    "prediction_norm": 5.98081636428833,
    "timestamp": "2025-07-14T12:34:48.489625"
  },
  {
    "epoch": 2073,
    "batch_idx": 192,
    "loss": 0.42027971148490906,
    "learning_rate": 3.2835948102367163e-06,
    "timestep_mean": 517.625,
    "epsilon_norm": 4.6440019607543945,
    "prediction_norm": 3.3476688861846924,
    "timestamp": "2025-07-14T12:34:48.494767"
  },
  {
    "epoch": 2074,
    "batch_idx": 0,
    "loss": 0.24514973163604736,
    "learning_rate": 3.37407047883032e-06,
    "timestep_mean": 588.9375,
    "epsilon_norm": 8.120176315307617,
    "prediction_norm": 7.019924640655518,
    "timestamp": "2025-07-14T12:34:48.499885"
  },
  {
    "epoch": 2074,
    "batch_idx": 32,
    "loss": 0.1539304554462433,
    "learning_rate": 3.37407047883032e-06,
    "timestep_mean": 517.03125,
    "epsilon_norm": 8.66347599029541,
    "prediction_norm": 8.24375057220459,
    "timestamp": "2025-07-14T12:34:48.505215"
  },
  {
    "epoch": 2074,
    "batch_idx": 64,
    "loss": 0.5726816654205322,
    "learning_rate": 3.37407047883032e-06,
    "timestep_mean": 432.59375,
    "epsilon_norm": 8.220808982849121,
    "prediction_norm": 5.501026630401611,
    "timestamp": "2025-07-14T12:34:48.510386"
  },
  {
    "epoch": 2074,
    "batch_idx": 96,
    "loss": 0.4423021674156189,
    "learning_rate": 3.37407047883032e-06,
    "timestep_mean": 434.46875,
    "epsilon_norm": 7.500818729400635,
    "prediction_norm": 5.664423942565918,
    "timestamp": "2025-07-14T12:34:48.515387"
  },
  {
    "epoch": 2074,
    "batch_idx": 128,
    "loss": 0.31070709228515625,
    "learning_rate": 3.37407047883032e-06,
    "timestep_mean": 466.9375,
    "epsilon_norm": 7.937017440795898,
    "prediction_norm": 6.466262340545654,
    "timestamp": "2025-07-14T12:34:48.520590"
  },
  {
    "epoch": 2074,
    "batch_idx": 160,
    "loss": 0.5266275405883789,
    "learning_rate": 3.37407047883032e-06,
    "timestep_mean": 506.09375,
    "epsilon_norm": 9.01425552368164,
    "prediction_norm": 7.053743362426758,
    "timestamp": "2025-07-14T12:34:48.525631"
  },
  {
    "epoch": 2074,
    "batch_idx": 192,
    "loss": 0.07104353606700897,
    "learning_rate": 3.37407047883032e-06,
    "timestep_mean": 559.625,
    "epsilon_norm": 2.6033992767333984,
    "prediction_norm": 2.6961371898651123,
    "timestamp": "2025-07-14T12:34:48.530612"
  },
  {
    "epoch": 2075,
    "batch_idx": 0,
    "loss": 0.4914439618587494,
    "learning_rate": 3.465771522536858e-06,
    "timestep_mean": 432.71875,
    "epsilon_norm": 9.168079376220703,
    "prediction_norm": 6.631424427032471,
    "timestamp": "2025-07-14T12:34:48.535663"
  },
  {
    "epoch": 2075,
    "batch_idx": 32,
    "loss": 0.21484701335430145,
    "learning_rate": 3.465771522536858e-06,
    "timestep_mean": 447.46875,
    "epsilon_norm": 8.13545036315918,
    "prediction_norm": 7.421809196472168,
    "timestamp": "2025-07-14T12:34:48.540700"
  },
  {
    "epoch": 2075,
    "batch_idx": 64,
    "loss": 0.2809286117553711,
    "learning_rate": 3.465771522536858e-06,
    "timestep_mean": 510.65625,
    "epsilon_norm": 7.849559307098389,
    "prediction_norm": 6.413572311401367,
    "timestamp": "2025-07-14T12:34:48.545912"
  },
  {
    "epoch": 2075,
    "batch_idx": 96,
    "loss": 0.40158116817474365,
    "learning_rate": 3.465771522536858e-06,
    "timestep_mean": 469.71875,
    "epsilon_norm": 7.632244110107422,
    "prediction_norm": 6.0164408683776855,
    "timestamp": "2025-07-14T12:34:48.551094"
  },
  {
    "epoch": 2075,
    "batch_idx": 128,
    "loss": 0.23605278134346008,
    "learning_rate": 3.465771522536858e-06,
    "timestep_mean": 515.25,
    "epsilon_norm": 7.953861713409424,
    "prediction_norm": 6.853907585144043,
    "timestamp": "2025-07-14T12:34:48.556206"
  },
  {
    "epoch": 2075,
    "batch_idx": 160,
    "loss": 0.16592209041118622,
    "learning_rate": 3.465771522536858e-06,
    "timestep_mean": 521.0625,
    "epsilon_norm": 8.525311470031738,
    "prediction_norm": 7.550526142120361,
    "timestamp": "2025-07-14T12:34:48.561324"
  },
  {
    "epoch": 2075,
    "batch_idx": 192,
    "loss": 0.17247968912124634,
    "learning_rate": 3.465771522536858e-06,
    "timestep_mean": 540.75,
    "epsilon_norm": 4.360125541687012,
    "prediction_norm": 3.6940622329711914,
    "timestamp": "2025-07-14T12:34:48.566312"
  },
  {
    "epoch": 2076,
    "batch_idx": 0,
    "loss": 0.21791347861289978,
    "learning_rate": 3.5586977150931556e-06,
    "timestep_mean": 532.21875,
    "epsilon_norm": 8.060613632202148,
    "prediction_norm": 6.668852806091309,
    "timestamp": "2025-07-14T12:34:48.571639"
  },
  {
    "epoch": 2076,
    "batch_idx": 32,
    "loss": 0.2607469856739044,
    "learning_rate": 3.5586977150931556e-06,
    "timestep_mean": 444.09375,
    "epsilon_norm": 7.283718109130859,
    "prediction_norm": 5.977640628814697,
    "timestamp": "2025-07-14T12:34:48.576683"
  },
  {
    "epoch": 2076,
    "batch_idx": 64,
    "loss": 0.3512851893901825,
    "learning_rate": 3.5586977150931556e-06,
    "timestep_mean": 438.6875,
    "epsilon_norm": 7.683563232421875,
    "prediction_norm": 6.125763416290283,
    "timestamp": "2025-07-14T12:34:48.581748"
  },
  {
    "epoch": 2076,
    "batch_idx": 96,
    "loss": 0.18430587649345398,
    "learning_rate": 3.5586977150931556e-06,
    "timestep_mean": 458.65625,
    "epsilon_norm": 6.441440582275391,
    "prediction_norm": 5.137763023376465,
    "timestamp": "2025-07-14T12:34:48.586939"
  },
  {
    "epoch": 2076,
    "batch_idx": 128,
    "loss": 0.22397297620773315,
    "learning_rate": 3.5586977150931556e-06,
    "timestep_mean": 503.71875,
    "epsilon_norm": 7.8046875,
    "prediction_norm": 6.45200777053833,
    "timestamp": "2025-07-14T12:34:48.592135"
  },
  {
    "epoch": 2076,
    "batch_idx": 160,
    "loss": 0.16571322083473206,
    "learning_rate": 3.5586977150931556e-06,
    "timestep_mean": 571.0625,
    "epsilon_norm": 6.8660712242126465,
    "prediction_norm": 5.873371124267578,
    "timestamp": "2025-07-14T12:34:48.597292"
  },
  {
    "epoch": 2076,
    "batch_idx": 192,
    "loss": 0.36991602182388306,
    "learning_rate": 3.5586977150931556e-06,
    "timestep_mean": 568.375,
    "epsilon_norm": 4.186411380767822,
    "prediction_norm": 3.07829213142395,
    "timestamp": "2025-07-14T12:34:48.602375"
  },
  {
    "epoch": 2077,
    "batch_idx": 0,
    "loss": 0.28217393159866333,
    "learning_rate": 3.6528488272130667e-06,
    "timestep_mean": 618.28125,
    "epsilon_norm": 7.39035701751709,
    "prediction_norm": 6.175076007843018,
    "timestamp": "2025-07-14T12:34:48.607386"
  },
  {
    "epoch": 2077,
    "batch_idx": 32,
    "loss": 0.33525264263153076,
    "learning_rate": 3.6528488272130667e-06,
    "timestep_mean": 419.0625,
    "epsilon_norm": 7.7396087646484375,
    "prediction_norm": 6.0495147705078125,
    "timestamp": "2025-07-14T12:34:48.612566"
  },
  {
    "epoch": 2077,
    "batch_idx": 64,
    "loss": 0.13440604507923126,
    "learning_rate": 3.6528488272130667e-06,
    "timestep_mean": 480.5,
    "epsilon_norm": 6.817437171936035,
    "prediction_norm": 6.695012092590332,
    "timestamp": "2025-07-14T12:34:48.617745"
  },
  {
    "epoch": 2077,
    "batch_idx": 96,
    "loss": 0.17572757601737976,
    "learning_rate": 3.6528488272130667e-06,
    "timestep_mean": 565.28125,
    "epsilon_norm": 7.427039623260498,
    "prediction_norm": 6.7197442054748535,
    "timestamp": "2025-07-14T12:34:48.622857"
  },
  {
    "epoch": 2077,
    "batch_idx": 128,
    "loss": 0.22179512679576874,
    "learning_rate": 3.6528488272130667e-06,
    "timestep_mean": 493.78125,
    "epsilon_norm": 8.940971374511719,
    "prediction_norm": 7.718297481536865,
    "timestamp": "2025-07-14T12:34:48.627987"
  },
  {
    "epoch": 2077,
    "batch_idx": 160,
    "loss": 0.15178020298480988,
    "learning_rate": 3.6528488272130667e-06,
    "timestep_mean": 524.5,
    "epsilon_norm": 7.722970485687256,
    "prediction_norm": 6.941417217254639,
    "timestamp": "2025-07-14T12:34:48.633041"
  },
  {
    "epoch": 2077,
    "batch_idx": 192,
    "loss": 0.20498010516166687,
    "learning_rate": 3.6528488272130667e-06,
    "timestep_mean": 731.0,
    "epsilon_norm": 3.6798837184906006,
    "prediction_norm": 2.9377200603485107,
    "timestamp": "2025-07-14T12:34:48.638306"
  },
  {
    "epoch": 2078,
    "batch_idx": 0,
    "loss": 0.2816975712776184,
    "learning_rate": 3.7482246265881413e-06,
    "timestep_mean": 496.1875,
    "epsilon_norm": 7.369171619415283,
    "prediction_norm": 6.02598237991333,
    "timestamp": "2025-07-14T12:34:48.643381"
  },
  {
    "epoch": 2078,
    "batch_idx": 32,
    "loss": 0.2291814088821411,
    "learning_rate": 3.7482246265881413e-06,
    "timestep_mean": 645.25,
    "epsilon_norm": 8.821771621704102,
    "prediction_norm": 7.889098644256592,
    "timestamp": "2025-07-14T12:34:48.648480"
  },
  {
    "epoch": 2078,
    "batch_idx": 64,
    "loss": 0.38102245330810547,
    "learning_rate": 3.7482246265881413e-06,
    "timestep_mean": 467.90625,
    "epsilon_norm": 7.52417516708374,
    "prediction_norm": 6.253537654876709,
    "timestamp": "2025-07-14T12:34:48.653672"
  },
  {
    "epoch": 2078,
    "batch_idx": 96,
    "loss": 0.19557756185531616,
    "learning_rate": 3.7482246265881413e-06,
    "timestep_mean": 503.6875,
    "epsilon_norm": 7.432689189910889,
    "prediction_norm": 6.592925548553467,
    "timestamp": "2025-07-14T12:34:48.658736"
  },
  {
    "epoch": 2078,
    "batch_idx": 128,
    "loss": 0.36401981115341187,
    "learning_rate": 3.7482246265881413e-06,
    "timestep_mean": 480.15625,
    "epsilon_norm": 7.6028361320495605,
    "prediction_norm": 6.227372646331787,
    "timestamp": "2025-07-14T12:34:48.663915"
  },
  {
    "epoch": 2078,
    "batch_idx": 160,
    "loss": 0.20486114919185638,
    "learning_rate": 3.7482246265881413e-06,
    "timestep_mean": 553.21875,
    "epsilon_norm": 7.5352959632873535,
    "prediction_norm": 6.837242603302002,
    "timestamp": "2025-07-14T12:34:48.669074"
  },
  {
    "epoch": 2078,
    "batch_idx": 192,
    "loss": 0.23301643133163452,
    "learning_rate": 3.7482246265881413e-06,
    "timestep_mean": 598.25,
    "epsilon_norm": 4.678531169891357,
    "prediction_norm": 4.113453388214111,
    "timestamp": "2025-07-14T12:34:48.674108"
  },
  {
    "epoch": 2079,
    "batch_idx": 0,
    "loss": 0.23611056804656982,
    "learning_rate": 3.844824877887958e-06,
    "timestep_mean": 471.71875,
    "epsilon_norm": 7.299243927001953,
    "prediction_norm": 5.890957832336426,
    "timestamp": "2025-07-14T12:34:48.679273"
  },
  {
    "epoch": 2079,
    "batch_idx": 32,
    "loss": 0.2105276882648468,
    "learning_rate": 3.844824877887958e-06,
    "timestep_mean": 486.125,
    "epsilon_norm": 7.767651081085205,
    "prediction_norm": 7.168363571166992,
    "timestamp": "2025-07-14T12:34:48.684430"
  },
  {
    "epoch": 2079,
    "batch_idx": 64,
    "loss": 0.36129337549209595,
    "learning_rate": 3.844824877887958e-06,
    "timestep_mean": 528.6875,
    "epsilon_norm": 7.822125434875488,
    "prediction_norm": 6.455286502838135,
    "timestamp": "2025-07-14T12:34:48.689528"
  },
  {
    "epoch": 2079,
    "batch_idx": 96,
    "loss": 0.11711852252483368,
    "learning_rate": 3.844824877887958e-06,
    "timestep_mean": 582.9375,
    "epsilon_norm": 7.448644638061523,
    "prediction_norm": 7.03138542175293,
    "timestamp": "2025-07-14T12:34:48.694620"
  },
  {
    "epoch": 2079,
    "batch_idx": 128,
    "loss": 0.25710225105285645,
    "learning_rate": 3.844824877887958e-06,
    "timestep_mean": 495.5625,
    "epsilon_norm": 7.647492408752441,
    "prediction_norm": 6.5482282638549805,
    "timestamp": "2025-07-14T12:34:48.699599"
  },
  {
    "epoch": 2079,
    "batch_idx": 160,
    "loss": 0.17053869366645813,
    "learning_rate": 3.844824877887958e-06,
    "timestep_mean": 508.03125,
    "epsilon_norm": 8.649863243103027,
    "prediction_norm": 7.6358771324157715,
    "timestamp": "2025-07-14T12:34:48.704799"
  },
  {
    "epoch": 2079,
    "batch_idx": 192,
    "loss": 0.5555723309516907,
    "learning_rate": 3.844824877887958e-06,
    "timestep_mean": 536.875,
    "epsilon_norm": 5.657909870147705,
    "prediction_norm": 4.904521942138672,
    "timestamp": "2025-07-14T12:34:48.709567"
  },
  {
    "epoch": 2080,
    "batch_idx": 0,
    "loss": 0.3053268492221832,
    "learning_rate": 3.942649342761066e-06,
    "timestep_mean": 484.03125,
    "epsilon_norm": 6.723114490509033,
    "prediction_norm": 5.71791934967041,
    "timestamp": "2025-07-14T12:34:48.714409"
  },
  {
    "epoch": 2080,
    "batch_idx": 32,
    "loss": 0.2047596275806427,
    "learning_rate": 3.942649342761066e-06,
    "timestep_mean": 501.625,
    "epsilon_norm": 8.312225341796875,
    "prediction_norm": 7.700724124908447,
    "timestamp": "2025-07-14T12:34:48.719269"
  },
  {
    "epoch": 2080,
    "batch_idx": 64,
    "loss": 0.2273777425289154,
    "learning_rate": 3.942649342761066e-06,
    "timestep_mean": 544.25,
    "epsilon_norm": 9.435603141784668,
    "prediction_norm": 8.567902565002441,
    "timestamp": "2025-07-14T12:34:48.724184"
  },
  {
    "epoch": 2080,
    "batch_idx": 96,
    "loss": 0.2799607813358307,
    "learning_rate": 3.942649342761066e-06,
    "timestep_mean": 512.34375,
    "epsilon_norm": 7.503625392913818,
    "prediction_norm": 6.367336273193359,
    "timestamp": "2025-07-14T12:34:48.729222"
  },
  {
    "epoch": 2080,
    "batch_idx": 128,
    "loss": 0.3626571297645569,
    "learning_rate": 3.942649342761066e-06,
    "timestep_mean": 511.59375,
    "epsilon_norm": 8.546154975891113,
    "prediction_norm": 7.135107040405273,
    "timestamp": "2025-07-14T12:34:48.734434"
  },
  {
    "epoch": 2080,
    "batch_idx": 160,
    "loss": 0.13848960399627686,
    "learning_rate": 3.942649342761066e-06,
    "timestep_mean": 642.0625,
    "epsilon_norm": 9.147595405578613,
    "prediction_norm": 9.039735794067383,
    "timestamp": "2025-07-14T12:34:48.739525"
  },
  {
    "epoch": 2080,
    "batch_idx": 192,
    "loss": 0.3153694272041321,
    "learning_rate": 3.942649342761066e-06,
    "timestep_mean": 407.875,
    "epsilon_norm": 4.998055458068848,
    "prediction_norm": 4.001641273498535,
    "timestamp": "2025-07-14T12:34:48.744570"
  },
  {
    "epoch": 2081,
    "batch_idx": 0,
    "loss": 0.18203169107437134,
    "learning_rate": 4.041697779835322e-06,
    "timestep_mean": 510.34375,
    "epsilon_norm": 7.547521591186523,
    "prediction_norm": 6.649340629577637,
    "timestamp": "2025-07-14T12:34:48.749454"
  },
  {
    "epoch": 2081,
    "batch_idx": 32,
    "loss": 0.1518702208995819,
    "learning_rate": 4.041697779835322e-06,
    "timestep_mean": 532.3125,
    "epsilon_norm": 7.974281311035156,
    "prediction_norm": 7.226606369018555,
    "timestamp": "2025-07-14T12:34:48.754636"
  },
  {
    "epoch": 2081,
    "batch_idx": 64,
    "loss": 0.3521069884300232,
    "learning_rate": 4.041697779835322e-06,
    "timestep_mean": 515.78125,
    "epsilon_norm": 9.266505241394043,
    "prediction_norm": 7.330082893371582,
    "timestamp": "2025-07-14T12:34:48.759623"
  },
  {
    "epoch": 2081,
    "batch_idx": 96,
    "loss": 0.2872830629348755,
    "learning_rate": 4.041697779835322e-06,
    "timestep_mean": 452.6875,
    "epsilon_norm": 7.798683166503906,
    "prediction_norm": 6.163485527038574,
    "timestamp": "2025-07-14T12:34:48.764658"
  },
  {
    "epoch": 2081,
    "batch_idx": 128,
    "loss": 0.24636539816856384,
    "learning_rate": 4.041697779835322e-06,
    "timestep_mean": 616.71875,
    "epsilon_norm": 8.995327949523926,
    "prediction_norm": 8.065260887145996,
    "timestamp": "2025-07-14T12:34:48.769740"
  },
  {
    "epoch": 2081,
    "batch_idx": 160,
    "loss": 0.3226718604564667,
    "learning_rate": 4.041697779835322e-06,
    "timestep_mean": 504.5625,
    "epsilon_norm": 8.27943229675293,
    "prediction_norm": 7.054220676422119,
    "timestamp": "2025-07-14T12:34:48.774672"
  },
  {
    "epoch": 2081,
    "batch_idx": 192,
    "loss": 0.7899718880653381,
    "learning_rate": 4.041697779835322e-06,
    "timestep_mean": 463.625,
    "epsilon_norm": 4.4333415031433105,
    "prediction_norm": 3.144331455230713,
    "timestamp": "2025-07-14T12:34:48.779571"
  },
  {
    "epoch": 2082,
    "batch_idx": 0,
    "loss": 0.4571727514266968,
    "learning_rate": 4.141969944718553e-06,
    "timestep_mean": 495.15625,
    "epsilon_norm": 7.915322780609131,
    "prediction_norm": 5.892182350158691,
    "timestamp": "2025-07-14T12:34:48.784413"
  },
  {
    "epoch": 2082,
    "batch_idx": 32,
    "loss": 0.22082477807998657,
    "learning_rate": 4.141969944718553e-06,
    "timestep_mean": 485.15625,
    "epsilon_norm": 7.879044532775879,
    "prediction_norm": 7.207665920257568,
    "timestamp": "2025-07-14T12:34:48.789269"
  },
  {
    "epoch": 2082,
    "batch_idx": 64,
    "loss": 0.18506821990013123,
    "learning_rate": 4.141969944718553e-06,
    "timestep_mean": 407.78125,
    "epsilon_norm": 7.63810396194458,
    "prediction_norm": 6.889399528503418,
    "timestamp": "2025-07-14T12:34:48.794137"
  },
  {
    "epoch": 2082,
    "batch_idx": 96,
    "loss": 0.16630905866622925,
    "learning_rate": 4.141969944718553e-06,
    "timestep_mean": 443.34375,
    "epsilon_norm": 8.051286697387695,
    "prediction_norm": 7.6063385009765625,
    "timestamp": "2025-07-14T12:34:48.799002"
  },
  {
    "epoch": 2082,
    "batch_idx": 128,
    "loss": 0.16826757788658142,
    "learning_rate": 4.141969944718553e-06,
    "timestep_mean": 500.84375,
    "epsilon_norm": 7.158509731292725,
    "prediction_norm": 6.107629776000977,
    "timestamp": "2025-07-14T12:34:48.803844"
  },
  {
    "epoch": 2082,
    "batch_idx": 160,
    "loss": 0.1610797941684723,
    "learning_rate": 4.141969944718553e-06,
    "timestep_mean": 531.8125,
    "epsilon_norm": 7.470571041107178,
    "prediction_norm": 6.662116527557373,
    "timestamp": "2025-07-14T12:34:48.808754"
  },
  {
    "epoch": 2082,
    "batch_idx": 192,
    "loss": 0.22743187844753265,
    "learning_rate": 4.141969944718553e-06,
    "timestep_mean": 524.75,
    "epsilon_norm": 3.5711820125579834,
    "prediction_norm": 3.3185386657714844,
    "timestamp": "2025-07-14T12:34:48.813582"
  },
  {
    "epoch": 2083,
    "batch_idx": 0,
    "loss": 0.3878869414329529,
    "learning_rate": 4.243465589999167e-06,
    "timestep_mean": 495.65625,
    "epsilon_norm": 7.670616626739502,
    "prediction_norm": 6.456746578216553,
    "timestamp": "2025-07-14T12:34:48.818488"
  },
  {
    "epoch": 2083,
    "batch_idx": 32,
    "loss": 0.220300555229187,
    "learning_rate": 4.243465589999167e-06,
    "timestep_mean": 449.125,
    "epsilon_norm": 8.510854721069336,
    "prediction_norm": 7.557823181152344,
    "timestamp": "2025-07-14T12:34:48.823336"
  },
  {
    "epoch": 2083,
    "batch_idx": 64,
    "loss": 0.25192055106163025,
    "learning_rate": 4.243465589999167e-06,
    "timestep_mean": 543.625,
    "epsilon_norm": 7.828601837158203,
    "prediction_norm": 6.785226345062256,
    "timestamp": "2025-07-14T12:34:48.828184"
  },
  {
    "epoch": 2083,
    "batch_idx": 96,
    "loss": 0.19369009137153625,
    "learning_rate": 4.243465589999167e-06,
    "timestep_mean": 567.21875,
    "epsilon_norm": 7.056777000427246,
    "prediction_norm": 5.815238952636719,
    "timestamp": "2025-07-14T12:34:48.832989"
  },
  {
    "epoch": 2083,
    "batch_idx": 128,
    "loss": 0.21956029534339905,
    "learning_rate": 4.243465589999167e-06,
    "timestep_mean": 492.71875,
    "epsilon_norm": 10.541792869567871,
    "prediction_norm": 9.294447898864746,
    "timestamp": "2025-07-14T12:34:48.837951"
  },
  {
    "epoch": 2083,
    "batch_idx": 160,
    "loss": 0.2382972240447998,
    "learning_rate": 4.243465589999167e-06,
    "timestep_mean": 512.0,
    "epsilon_norm": 8.542764663696289,
    "prediction_norm": 7.5996551513671875,
    "timestamp": "2025-07-14T12:34:48.842775"
  },
  {
    "epoch": 2083,
    "batch_idx": 192,
    "loss": 0.056601740419864655,
    "learning_rate": 4.243465589999167e-06,
    "timestep_mean": 631.25,
    "epsilon_norm": 5.206565856933594,
    "prediction_norm": 5.356545925140381,
    "timestamp": "2025-07-14T12:34:48.847578"
  },
  {
    "epoch": 2084,
    "batch_idx": 0,
    "loss": 0.21130605041980743,
    "learning_rate": 4.34618446524671e-06,
    "timestep_mean": 511.40625,
    "epsilon_norm": 8.078359603881836,
    "prediction_norm": 6.823399066925049,
    "timestamp": "2025-07-14T12:34:48.852398"
  },
  {
    "epoch": 2084,
    "batch_idx": 32,
    "loss": 0.15271008014678955,
    "learning_rate": 4.34618446524671e-06,
    "timestep_mean": 514.6875,
    "epsilon_norm": 8.735849380493164,
    "prediction_norm": 8.136554718017578,
    "timestamp": "2025-07-14T12:34:48.857272"
  },
  {
    "epoch": 2084,
    "batch_idx": 64,
    "loss": 0.2807818055152893,
    "learning_rate": 4.34618446524671e-06,
    "timestep_mean": 453.5625,
    "epsilon_norm": 9.021180152893066,
    "prediction_norm": 7.262690544128418,
    "timestamp": "2025-07-14T12:34:48.862138"
  },
  {
    "epoch": 2084,
    "batch_idx": 96,
    "loss": 0.2911376953125,
    "learning_rate": 4.34618446524671e-06,
    "timestep_mean": 467.625,
    "epsilon_norm": 8.674670219421387,
    "prediction_norm": 7.148521423339844,
    "timestamp": "2025-07-14T12:34:48.867059"
  },
  {
    "epoch": 2084,
    "batch_idx": 128,
    "loss": 0.29466089606285095,
    "learning_rate": 4.34618446524671e-06,
    "timestep_mean": 418.28125,
    "epsilon_norm": 7.909583568572998,
    "prediction_norm": 6.096256732940674,
    "timestamp": "2025-07-14T12:34:48.871943"
  },
  {
    "epoch": 2084,
    "batch_idx": 160,
    "loss": 0.27453622221946716,
    "learning_rate": 4.34618446524671e-06,
    "timestep_mean": 487.34375,
    "epsilon_norm": 7.630361557006836,
    "prediction_norm": 6.127861022949219,
    "timestamp": "2025-07-14T12:34:48.876812"
  },
  {
    "epoch": 2084,
    "batch_idx": 192,
    "loss": 0.1348453164100647,
    "learning_rate": 4.34618446524671e-06,
    "timestep_mean": 610.875,
    "epsilon_norm": 4.0463995933532715,
    "prediction_norm": 3.584794044494629,
    "timestamp": "2025-07-14T12:34:48.881633"
  },
  {
    "epoch": 2085,
    "batch_idx": 0,
    "loss": 0.38057318329811096,
    "learning_rate": 4.4501263170125865e-06,
    "timestep_mean": 462.78125,
    "epsilon_norm": 8.055301666259766,
    "prediction_norm": 6.119548797607422,
    "timestamp": "2025-07-14T12:34:48.886499"
  },
  {
    "epoch": 2085,
    "batch_idx": 32,
    "loss": 0.41474199295043945,
    "learning_rate": 4.4501263170125865e-06,
    "timestep_mean": 431.71875,
    "epsilon_norm": 7.238062858581543,
    "prediction_norm": 5.75759744644165,
    "timestamp": "2025-07-14T12:34:48.891332"
  },
  {
    "epoch": 2085,
    "batch_idx": 64,
    "loss": 0.4290291965007782,
    "learning_rate": 4.4501263170125865e-06,
    "timestep_mean": 332.25,
    "epsilon_norm": 8.41179370880127,
    "prediction_norm": 6.411358833312988,
    "timestamp": "2025-07-14T12:34:48.896233"
  },
  {
    "epoch": 2085,
    "batch_idx": 96,
    "loss": 0.28718799352645874,
    "learning_rate": 4.4501263170125865e-06,
    "timestep_mean": 396.125,
    "epsilon_norm": 7.294833660125732,
    "prediction_norm": 5.602402687072754,
    "timestamp": "2025-07-14T12:34:48.901125"
  },
  {
    "epoch": 2085,
    "batch_idx": 128,
    "loss": 0.2105642855167389,
    "learning_rate": 4.4501263170125865e-06,
    "timestep_mean": 460.96875,
    "epsilon_norm": 7.276248455047607,
    "prediction_norm": 6.134605407714844,
    "timestamp": "2025-07-14T12:34:48.906030"
  },
  {
    "epoch": 2085,
    "batch_idx": 160,
    "loss": 0.35962456464767456,
    "learning_rate": 4.4501263170125865e-06,
    "timestep_mean": 558.9375,
    "epsilon_norm": 8.604203224182129,
    "prediction_norm": 7.03373384475708,
    "timestamp": "2025-07-14T12:34:48.910867"
  },
  {
    "epoch": 2085,
    "batch_idx": 192,
    "loss": 0.6497508883476257,
    "learning_rate": 4.4501263170125865e-06,
    "timestep_mean": 402.625,
    "epsilon_norm": 4.779829025268555,
    "prediction_norm": 3.9291744232177734,
    "timestamp": "2025-07-14T12:34:48.915767"
  },
  {
    "epoch": 2086,
    "batch_idx": 0,
    "loss": 0.24163833260536194,
    "learning_rate": 4.55529088883067e-06,
    "timestep_mean": 510.21875,
    "epsilon_norm": 6.558856010437012,
    "prediction_norm": 5.7142157554626465,
    "timestamp": "2025-07-14T12:34:48.920609"
  },
  {
    "epoch": 2086,
    "batch_idx": 32,
    "loss": 0.3450802266597748,
    "learning_rate": 4.55529088883067e-06,
    "timestep_mean": 485.125,
    "epsilon_norm": 8.499987602233887,
    "prediction_norm": 7.179579257965088,
    "timestamp": "2025-07-14T12:34:48.925561"
  },
  {
    "epoch": 2086,
    "batch_idx": 64,
    "loss": 0.22310036420822144,
    "learning_rate": 4.55529088883067e-06,
    "timestep_mean": 438.9375,
    "epsilon_norm": 7.51213264465332,
    "prediction_norm": 5.784492015838623,
    "timestamp": "2025-07-14T12:34:48.930433"
  },
  {
    "epoch": 2086,
    "batch_idx": 96,
    "loss": 0.3227665424346924,
    "learning_rate": 4.55529088883067e-06,
    "timestep_mean": 476.96875,
    "epsilon_norm": 8.669855117797852,
    "prediction_norm": 6.564530849456787,
    "timestamp": "2025-07-14T12:34:48.935307"
  },
  {
    "epoch": 2086,
    "batch_idx": 128,
    "loss": 0.3666141629219055,
    "learning_rate": 4.55529088883067e-06,
    "timestep_mean": 513.59375,
    "epsilon_norm": 7.599697113037109,
    "prediction_norm": 5.780829906463623,
    "timestamp": "2025-07-14T12:34:48.940206"
  },
  {
    "epoch": 2086,
    "batch_idx": 160,
    "loss": 0.3866719901561737,
    "learning_rate": 4.55529088883067e-06,
    "timestep_mean": 549.6875,
    "epsilon_norm": 9.085549354553223,
    "prediction_norm": 7.9634270668029785,
    "timestamp": "2025-07-14T12:34:48.945245"
  },
  {
    "epoch": 2086,
    "batch_idx": 192,
    "loss": 0.007633927743881941,
    "learning_rate": 4.55529088883067e-06,
    "timestep_mean": 797.375,
    "epsilon_norm": 3.672964334487915,
    "prediction_norm": 3.5147571563720703,
    "timestamp": "2025-07-14T12:34:48.950219"
  },
  {
    "epoch": 2087,
    "batch_idx": 0,
    "loss": 0.4144037067890167,
    "learning_rate": 4.661677921217694e-06,
    "timestep_mean": 437.25,
    "epsilon_norm": 6.94329833984375,
    "prediction_norm": 5.0943989753723145,
    "timestamp": "2025-07-14T12:34:48.955094"
  },
  {
    "epoch": 2087,
    "batch_idx": 32,
    "loss": 0.2235366255044937,
    "learning_rate": 4.661677921217694e-06,
    "timestep_mean": 435.59375,
    "epsilon_norm": 7.430109977722168,
    "prediction_norm": 6.153697967529297,
    "timestamp": "2025-07-14T12:34:48.959955"
  },
  {
    "epoch": 2087,
    "batch_idx": 64,
    "loss": 0.25509995222091675,
    "learning_rate": 4.661677921217694e-06,
    "timestep_mean": 527.53125,
    "epsilon_norm": 9.000880241394043,
    "prediction_norm": 7.054610252380371,
    "timestamp": "2025-07-14T12:34:48.964826"
  },
  {
    "epoch": 2087,
    "batch_idx": 96,
    "loss": 0.3322666883468628,
    "learning_rate": 4.661677921217694e-06,
    "timestep_mean": 519.03125,
    "epsilon_norm": 7.887640476226807,
    "prediction_norm": 6.884598255157471,
    "timestamp": "2025-07-14T12:34:48.969694"
  },
  {
    "epoch": 2087,
    "batch_idx": 128,
    "loss": 0.38519811630249023,
    "learning_rate": 4.661677921217694e-06,
    "timestep_mean": 493.25,
    "epsilon_norm": 8.541568756103516,
    "prediction_norm": 6.995584011077881,
    "timestamp": "2025-07-14T12:34:48.974543"
  },
  {
    "epoch": 2087,
    "batch_idx": 160,
    "loss": 0.19524893164634705,
    "learning_rate": 4.661677921217694e-06,
    "timestep_mean": 578.625,
    "epsilon_norm": 8.068251609802246,
    "prediction_norm": 7.380081653594971,
    "timestamp": "2025-07-14T12:34:48.979426"
  },
  {
    "epoch": 2087,
    "batch_idx": 192,
    "loss": 0.2671193480491638,
    "learning_rate": 4.661677921217694e-06,
    "timestep_mean": 413.0,
    "epsilon_norm": 3.9934699535369873,
    "prediction_norm": 3.2965567111968994,
    "timestamp": "2025-07-14T12:34:48.984277"
  },
  {
    "epoch": 2088,
    "batch_idx": 0,
    "loss": 0.15800069272518158,
    "learning_rate": 4.769287151674356e-06,
    "timestep_mean": 605.3125,
    "epsilon_norm": 7.5178961753845215,
    "prediction_norm": 6.977625846862793,
    "timestamp": "2025-07-14T12:34:48.989114"
  },
  {
    "epoch": 2088,
    "batch_idx": 32,
    "loss": 0.21972927451133728,
    "learning_rate": 4.769287151674356e-06,
    "timestep_mean": 457.65625,
    "epsilon_norm": 8.025032997131348,
    "prediction_norm": 6.839761734008789,
    "timestamp": "2025-07-14T12:34:48.993968"
  },
  {
    "epoch": 2088,
    "batch_idx": 64,
    "loss": 0.19544321298599243,
    "learning_rate": 4.769287151674356e-06,
    "timestep_mean": 512.875,
    "epsilon_norm": 7.497654914855957,
    "prediction_norm": 6.567048072814941,
    "timestamp": "2025-07-14T12:34:48.998781"
  },
  {
    "epoch": 2088,
    "batch_idx": 96,
    "loss": 0.170052632689476,
    "learning_rate": 4.769287151674356e-06,
    "timestep_mean": 474.625,
    "epsilon_norm": 7.641039848327637,
    "prediction_norm": 7.043814182281494,
    "timestamp": "2025-07-14T12:34:49.003595"
  },
  {
    "epoch": 2088,
    "batch_idx": 128,
    "loss": 0.3561028838157654,
    "learning_rate": 4.769287151674356e-06,
    "timestep_mean": 526.5625,
    "epsilon_norm": 7.482975482940674,
    "prediction_norm": 6.4735941886901855,
    "timestamp": "2025-07-14T12:34:49.008533"
  },
  {
    "epoch": 2088,
    "batch_idx": 160,
    "loss": 0.28129658102989197,
    "learning_rate": 4.769287151674356e-06,
    "timestep_mean": 497.375,
    "epsilon_norm": 8.095708847045898,
    "prediction_norm": 7.0995049476623535,
    "timestamp": "2025-07-14T12:34:49.013353"
  },
  {
    "epoch": 2088,
    "batch_idx": 192,
    "loss": 0.14358189702033997,
    "learning_rate": 4.769287151674356e-06,
    "timestep_mean": 561.25,
    "epsilon_norm": 3.999765396118164,
    "prediction_norm": 3.8751072883605957,
    "timestamp": "2025-07-14T12:34:49.018133"
  },
  {
    "epoch": 2089,
    "batch_idx": 0,
    "loss": 0.2609080672264099,
    "learning_rate": 4.878118314685548e-06,
    "timestep_mean": 505.96875,
    "epsilon_norm": 8.408554077148438,
    "prediction_norm": 7.856872081756592,
    "timestamp": "2025-07-14T12:34:49.023006"
  },
  {
    "epoch": 2089,
    "batch_idx": 32,
    "loss": 0.26184579730033875,
    "learning_rate": 4.878118314685548e-06,
    "timestep_mean": 504.34375,
    "epsilon_norm": 7.356343746185303,
    "prediction_norm": 6.698517322540283,
    "timestamp": "2025-07-14T12:34:49.027827"
  },
  {
    "epoch": 2089,
    "batch_idx": 64,
    "loss": 0.5352712869644165,
    "learning_rate": 4.878118314685548e-06,
    "timestep_mean": 428.21875,
    "epsilon_norm": 8.270716667175293,
    "prediction_norm": 5.941254615783691,
    "timestamp": "2025-07-14T12:34:49.032660"
  },
  {
    "epoch": 2089,
    "batch_idx": 96,
    "loss": 0.1045103520154953,
    "learning_rate": 4.878118314685548e-06,
    "timestep_mean": 560.8125,
    "epsilon_norm": 8.167433738708496,
    "prediction_norm": 7.750609397888184,
    "timestamp": "2025-07-14T12:34:49.037439"
  },
  {
    "epoch": 2089,
    "batch_idx": 128,
    "loss": 0.3217247724533081,
    "learning_rate": 4.878118314685548e-06,
    "timestep_mean": 474.59375,
    "epsilon_norm": 7.8904595375061035,
    "prediction_norm": 6.366423606872559,
    "timestamp": "2025-07-14T12:34:49.042596"
  },
  {
    "epoch": 2089,
    "batch_idx": 160,
    "loss": 0.4176468849182129,
    "learning_rate": 4.878118314685548e-06,
    "timestep_mean": 493.25,
    "epsilon_norm": 8.625753402709961,
    "prediction_norm": 7.071091651916504,
    "timestamp": "2025-07-14T12:34:49.047680"
  },
  {
    "epoch": 2089,
    "batch_idx": 192,
    "loss": 0.09558883309364319,
    "learning_rate": 4.878118314685548e-06,
    "timestep_mean": 553.25,
    "epsilon_norm": 3.0087954998016357,
    "prediction_norm": 2.7812819480895996,
    "timestamp": "2025-07-14T12:34:49.052817"
  },
  {
    "epoch": 2090,
    "batch_idx": 0,
    "loss": 0.18285229802131653,
    "learning_rate": 4.988171141721238e-06,
    "timestep_mean": 496.6875,
    "epsilon_norm": 8.221632957458496,
    "prediction_norm": 8.036340713500977,
    "timestamp": "2025-07-14T12:34:49.058003"
  },
  {
    "epoch": 2090,
    "batch_idx": 32,
    "loss": 0.23399314284324646,
    "learning_rate": 4.988171141721238e-06,
    "timestep_mean": 456.5,
    "epsilon_norm": 6.3564324378967285,
    "prediction_norm": 5.919695854187012,
    "timestamp": "2025-07-14T12:34:49.063213"
  },
  {
    "epoch": 2090,
    "batch_idx": 64,
    "loss": 0.28229522705078125,
    "learning_rate": 4.988171141721238e-06,
    "timestep_mean": 419.75,
    "epsilon_norm": 6.750603199005127,
    "prediction_norm": 5.8422417640686035,
    "timestamp": "2025-07-14T12:34:49.068365"
  },
  {
    "epoch": 2090,
    "batch_idx": 96,
    "loss": 0.21042290329933167,
    "learning_rate": 4.988171141721238e-06,
    "timestep_mean": 522.96875,
    "epsilon_norm": 7.743922710418701,
    "prediction_norm": 6.683525085449219,
    "timestamp": "2025-07-14T12:34:49.073407"
  },
  {
    "epoch": 2090,
    "batch_idx": 128,
    "loss": 0.13874953985214233,
    "learning_rate": 4.988171141721238e-06,
    "timestep_mean": 532.21875,
    "epsilon_norm": 7.3382368087768555,
    "prediction_norm": 6.046087741851807,
    "timestamp": "2025-07-14T12:34:49.078571"
  },
  {
    "epoch": 2090,
    "batch_idx": 160,
    "loss": 0.175969660282135,
    "learning_rate": 4.988171141721238e-06,
    "timestep_mean": 547.0625,
    "epsilon_norm": 6.5651326179504395,
    "prediction_norm": 5.698395252227783,
    "timestamp": "2025-07-14T12:34:49.083840"
  },
  {
    "epoch": 2090,
    "batch_idx": 192,
    "loss": 0.20394626259803772,
    "learning_rate": 4.988171141721238e-06,
    "timestep_mean": 472.75,
    "epsilon_norm": 4.341278553009033,
    "prediction_norm": 3.4508960247039795,
    "timestamp": "2025-07-14T12:34:49.088911"
  },
  {
    "epoch": 2091,
    "batch_idx": 0,
    "loss": 0.1752658188343048,
    "learning_rate": 5.0994453612368604e-06,
    "timestep_mean": 485.59375,
    "epsilon_norm": 8.250218391418457,
    "prediction_norm": 7.731131076812744,
    "timestamp": "2025-07-14T12:34:49.094075"
  },
  {
    "epoch": 2091,
    "batch_idx": 32,
    "loss": 0.31584230065345764,
    "learning_rate": 5.0994453612368604e-06,
    "timestep_mean": 547.21875,
    "epsilon_norm": 7.814690113067627,
    "prediction_norm": 5.97505521774292,
    "timestamp": "2025-07-14T12:34:49.099166"
  },
  {
    "epoch": 2091,
    "batch_idx": 64,
    "loss": 0.3349187970161438,
    "learning_rate": 5.0994453612368604e-06,
    "timestep_mean": 466.125,
    "epsilon_norm": 8.903419494628906,
    "prediction_norm": 7.394201755523682,
    "timestamp": "2025-07-14T12:34:49.104343"
  },
  {
    "epoch": 2091,
    "batch_idx": 96,
    "loss": 0.25762057304382324,
    "learning_rate": 5.0994453612368604e-06,
    "timestep_mean": 457.65625,
    "epsilon_norm": 8.399690628051758,
    "prediction_norm": 7.533636569976807,
    "timestamp": "2025-07-14T12:34:49.109492"
  },
  {
    "epoch": 2091,
    "batch_idx": 128,
    "loss": 0.26574328541755676,
    "learning_rate": 5.0994453612368604e-06,
    "timestep_mean": 514.25,
    "epsilon_norm": 8.273477554321289,
    "prediction_norm": 6.792495250701904,
    "timestamp": "2025-07-14T12:34:49.114582"
  },
  {
    "epoch": 2091,
    "batch_idx": 160,
    "loss": 0.15020619332790375,
    "learning_rate": 5.0994453612368604e-06,
    "timestep_mean": 573.8125,
    "epsilon_norm": 8.78208065032959,
    "prediction_norm": 8.146120071411133,
    "timestamp": "2025-07-14T12:34:49.119871"
  },
  {
    "epoch": 2091,
    "batch_idx": 192,
    "loss": 0.06693895161151886,
    "learning_rate": 5.0994453612368604e-06,
    "timestep_mean": 381.625,
    "epsilon_norm": 3.1687097549438477,
    "prediction_norm": 3.0756428241729736,
    "timestamp": "2025-07-14T12:34:49.124910"
  },
  {
    "epoch": 2092,
    "batch_idx": 0,
    "loss": 0.27849841117858887,
    "learning_rate": 5.211940698674485e-06,
    "timestep_mean": 524.65625,
    "epsilon_norm": 8.030695915222168,
    "prediction_norm": 6.570830821990967,
    "timestamp": "2025-07-14T12:34:49.130218"
  },
  {
    "epoch": 2092,
    "batch_idx": 32,
    "loss": 0.27209484577178955,
    "learning_rate": 5.211940698674485e-06,
    "timestep_mean": 527.78125,
    "epsilon_norm": 7.572181701660156,
    "prediction_norm": 6.3561248779296875,
    "timestamp": "2025-07-14T12:34:49.135427"
  },
  {
    "epoch": 2092,
    "batch_idx": 64,
    "loss": 0.34591853618621826,
    "learning_rate": 5.211940698674485e-06,
    "timestep_mean": 392.25,
    "epsilon_norm": 7.558387279510498,
    "prediction_norm": 6.041871070861816,
    "timestamp": "2025-07-14T12:34:49.140565"
  },
  {
    "epoch": 2092,
    "batch_idx": 96,
    "loss": 0.3628217577934265,
    "learning_rate": 5.211940698674485e-06,
    "timestep_mean": 494.5625,
    "epsilon_norm": 8.373015403747559,
    "prediction_norm": 7.064360618591309,
    "timestamp": "2025-07-14T12:34:49.145778"
  },
  {
    "epoch": 2092,
    "batch_idx": 128,
    "loss": 0.11397404223680496,
    "learning_rate": 5.211940698674485e-06,
    "timestep_mean": 549.40625,
    "epsilon_norm": 8.143403053283691,
    "prediction_norm": 7.266425609588623,
    "timestamp": "2025-07-14T12:34:49.150613"
  },
  {
    "epoch": 2092,
    "batch_idx": 160,
    "loss": 0.40052399039268494,
    "learning_rate": 5.211940698674485e-06,
    "timestep_mean": 349.125,
    "epsilon_norm": 7.786913871765137,
    "prediction_norm": 5.967092037200928,
    "timestamp": "2025-07-14T12:34:49.155754"
  },
  {
    "epoch": 2092,
    "batch_idx": 192,
    "loss": 0.1098267212510109,
    "learning_rate": 5.211940698674485e-06,
    "timestep_mean": 379.625,
    "epsilon_norm": 3.313725471496582,
    "prediction_norm": 2.997598886489868,
    "timestamp": "2025-07-14T12:34:49.160830"
  },
  {
    "epoch": 2093,
    "batch_idx": 0,
    "loss": 0.2004818171262741,
    "learning_rate": 5.325656876463033e-06,
    "timestep_mean": 557.15625,
    "epsilon_norm": 8.755416870117188,
    "prediction_norm": 8.002159118652344,
    "timestamp": "2025-07-14T12:34:49.165909"
  },
  {
    "epoch": 2093,
    "batch_idx": 32,
    "loss": 0.213534414768219,
    "learning_rate": 5.325656876463033e-06,
    "timestep_mean": 569.125,
    "epsilon_norm": 7.558409690856934,
    "prediction_norm": 6.484984397888184,
    "timestamp": "2025-07-14T12:34:49.171098"
  },
  {
    "epoch": 2093,
    "batch_idx": 64,
    "loss": 0.15814359486103058,
    "learning_rate": 5.325656876463033e-06,
    "timestep_mean": 507.84375,
    "epsilon_norm": 8.189188003540039,
    "prediction_norm": 7.34099006652832,
    "timestamp": "2025-07-14T12:34:49.176240"
  },
  {
    "epoch": 2093,
    "batch_idx": 96,
    "loss": 0.36891210079193115,
    "learning_rate": 5.325656876463033e-06,
    "timestep_mean": 430.6875,
    "epsilon_norm": 8.350152969360352,
    "prediction_norm": 7.235630989074707,
    "timestamp": "2025-07-14T12:34:49.181218"
  },
  {
    "epoch": 2093,
    "batch_idx": 128,
    "loss": 0.21987196803092957,
    "learning_rate": 5.325656876463033e-06,
    "timestep_mean": 476.1875,
    "epsilon_norm": 7.786680221557617,
    "prediction_norm": 6.843603610992432,
    "timestamp": "2025-07-14T12:34:49.186274"
  },
  {
    "epoch": 2093,
    "batch_idx": 160,
    "loss": 0.20974613726139069,
    "learning_rate": 5.325656876463033e-06,
    "timestep_mean": 523.84375,
    "epsilon_norm": 7.3225417137146,
    "prediction_norm": 6.385613441467285,
    "timestamp": "2025-07-14T12:34:49.191408"
  },
  {
    "epoch": 2093,
    "batch_idx": 192,
    "loss": 0.29894059896469116,
    "learning_rate": 5.325656876463033e-06,
    "timestep_mean": 477.375,
    "epsilon_norm": 3.494184732437134,
    "prediction_norm": 2.77005672454834,
    "timestamp": "2025-07-14T12:34:49.196642"
  },
  {
    "epoch": 2094,
    "batch_idx": 0,
    "loss": 0.06518390029668808,
    "learning_rate": 5.440593614019059e-06,
    "timestep_mean": 549.8125,
    "epsilon_norm": 6.70557975769043,
    "prediction_norm": 6.409613132476807,
    "timestamp": "2025-07-14T12:34:49.201726"
  },
  {
    "epoch": 2094,
    "batch_idx": 32,
    "loss": 0.18918129801750183,
    "learning_rate": 5.440593614019059e-06,
    "timestep_mean": 501.0625,
    "epsilon_norm": 7.6266703605651855,
    "prediction_norm": 6.846685409545898,
    "timestamp": "2025-07-14T12:34:49.206793"
  },
  {
    "epoch": 2094,
    "batch_idx": 64,
    "loss": 0.22885265946388245,
    "learning_rate": 5.440593614019059e-06,
    "timestep_mean": 469.0,
    "epsilon_norm": 7.1267781257629395,
    "prediction_norm": 5.759581089019775,
    "timestamp": "2025-07-14T12:34:49.211965"
  },
  {
    "epoch": 2094,
    "batch_idx": 96,
    "loss": 0.35833561420440674,
    "learning_rate": 5.440593614019059e-06,
    "timestep_mean": 414.375,
    "epsilon_norm": 7.888171672821045,
    "prediction_norm": 6.287069797515869,
    "timestamp": "2025-07-14T12:34:49.217093"
  },
  {
    "epoch": 2094,
    "batch_idx": 128,
    "loss": 0.35073375701904297,
    "learning_rate": 5.440593614019059e-06,
    "timestep_mean": 475.375,
    "epsilon_norm": 7.864226341247559,
    "prediction_norm": 6.463277339935303,
    "timestamp": "2025-07-14T12:34:49.222245"
  },
  {
    "epoch": 2094,
    "batch_idx": 160,
    "loss": 0.2140653431415558,
    "learning_rate": 5.440593614019059e-06,
    "timestep_mean": 594.65625,
    "epsilon_norm": 7.604952335357666,
    "prediction_norm": 7.099820137023926,
    "timestamp": "2025-07-14T12:34:49.227333"
  },
  {
    "epoch": 2094,
    "batch_idx": 192,
    "loss": 0.4396846294403076,
    "learning_rate": 5.440593614019059e-06,
    "timestep_mean": 386.5,
    "epsilon_norm": 4.361995697021484,
    "prediction_norm": 3.3204267024993896,
    "timestamp": "2025-07-14T12:34:49.232383"
  },
  {
    "epoch": 2095,
    "batch_idx": 0,
    "loss": 0.11013079434633255,
    "learning_rate": 5.556750627747694e-06,
    "timestep_mean": 584.0625,
    "epsilon_norm": 8.183952331542969,
    "prediction_norm": 7.729267120361328,
    "timestamp": "2025-07-14T12:34:49.237631"
  },
  {
    "epoch": 2095,
    "batch_idx": 32,
    "loss": 0.3292742073535919,
    "learning_rate": 5.556750627747694e-06,
    "timestep_mean": 506.53125,
    "epsilon_norm": 9.611791610717773,
    "prediction_norm": 8.356029510498047,
    "timestamp": "2025-07-14T12:34:49.242808"
  },
  {
    "epoch": 2095,
    "batch_idx": 64,
    "loss": 0.16103124618530273,
    "learning_rate": 5.556750627747694e-06,
    "timestep_mean": 487.40625,
    "epsilon_norm": 7.00653600692749,
    "prediction_norm": 6.351606845855713,
    "timestamp": "2025-07-14T12:34:49.247914"
  },
  {
    "epoch": 2095,
    "batch_idx": 96,
    "loss": 0.1376928687095642,
    "learning_rate": 5.556750627747694e-06,
    "timestep_mean": 502.71875,
    "epsilon_norm": 7.00252628326416,
    "prediction_norm": 6.3879852294921875,
    "timestamp": "2025-07-14T12:34:49.252938"
  },
  {
    "epoch": 2095,
    "batch_idx": 128,
    "loss": 0.29513654112815857,
    "learning_rate": 5.556750627747694e-06,
    "timestep_mean": 434.71875,
    "epsilon_norm": 7.69420051574707,
    "prediction_norm": 6.386261940002441,
    "timestamp": "2025-07-14T12:34:49.257933"
  },
  {
    "epoch": 2095,
    "batch_idx": 160,
    "loss": 0.25345900654792786,
    "learning_rate": 5.556750627747694e-06,
    "timestep_mean": 533.1875,
    "epsilon_norm": 7.2182512283325195,
    "prediction_norm": 6.480485916137695,
    "timestamp": "2025-07-14T12:34:49.262853"
  },
  {
    "epoch": 2095,
    "batch_idx": 192,
    "loss": 0.3066675364971161,
    "learning_rate": 5.556750627747694e-06,
    "timestep_mean": 269.375,
    "epsilon_norm": 3.7248809337615967,
    "prediction_norm": 2.911342144012451,
    "timestamp": "2025-07-14T12:34:49.267684"
  },
  {
    "epoch": 2096,
    "batch_idx": 0,
    "loss": 0.5605467557907104,
    "learning_rate": 5.674127631043032e-06,
    "timestep_mean": 348.5625,
    "epsilon_norm": 7.959635257720947,
    "prediction_norm": 5.131824493408203,
    "timestamp": "2025-07-14T12:34:49.272537"
  },
  {
    "epoch": 2096,
    "batch_idx": 32,
    "loss": 0.15360304713249207,
    "learning_rate": 5.674127631043032e-06,
    "timestep_mean": 443.96875,
    "epsilon_norm": 6.939497470855713,
    "prediction_norm": 5.991028308868408,
    "timestamp": "2025-07-14T12:34:49.277608"
  },
  {
    "epoch": 2096,
    "batch_idx": 64,
    "loss": 0.3676488995552063,
    "learning_rate": 5.674127631043032e-06,
    "timestep_mean": 523.40625,
    "epsilon_norm": 8.205381393432617,
    "prediction_norm": 6.358651638031006,
    "timestamp": "2025-07-14T12:34:49.282690"
  },
  {
    "epoch": 2096,
    "batch_idx": 96,
    "loss": 0.3474425971508026,
    "learning_rate": 5.674127631043032e-06,
    "timestep_mean": 513.59375,
    "epsilon_norm": 7.739007949829102,
    "prediction_norm": 6.754108905792236,
    "timestamp": "2025-07-14T12:34:49.287776"
  },
  {
    "epoch": 2096,
    "batch_idx": 128,
    "loss": 0.15532532334327698,
    "learning_rate": 5.674127631043032e-06,
    "timestep_mean": 542.71875,
    "epsilon_norm": 7.8454060554504395,
    "prediction_norm": 6.786777019500732,
    "timestamp": "2025-07-14T12:34:49.292894"
  },
  {
    "epoch": 2096,
    "batch_idx": 160,
    "loss": 0.46317577362060547,
    "learning_rate": 5.674127631043032e-06,
    "timestep_mean": 495.25,
    "epsilon_norm": 9.015693664550781,
    "prediction_norm": 6.813936233520508,
    "timestamp": "2025-07-14T12:34:49.297964"
  },
  {
    "epoch": 2096,
    "batch_idx": 192,
    "loss": 0.6043339967727661,
    "learning_rate": 5.674127631043032e-06,
    "timestep_mean": 532.0,
    "epsilon_norm": 3.9885687828063965,
    "prediction_norm": 3.0721683502197266,
    "timestamp": "2025-07-14T12:34:49.303026"
  },
  {
    "epoch": 2097,
    "batch_idx": 0,
    "loss": 0.24993614852428436,
    "learning_rate": 5.7927243342888535e-06,
    "timestep_mean": 478.15625,
    "epsilon_norm": 7.382826328277588,
    "prediction_norm": 5.988236427307129,
    "timestamp": "2025-07-14T12:34:49.308177"
  },
  {
    "epoch": 2097,
    "batch_idx": 32,
    "loss": 0.3026113510131836,
    "learning_rate": 5.7927243342888535e-06,
    "timestep_mean": 548.375,
    "epsilon_norm": 8.209766387939453,
    "prediction_norm": 6.978698253631592,
    "timestamp": "2025-07-14T12:34:49.313378"
  },
  {
    "epoch": 2097,
    "batch_idx": 64,
    "loss": 0.4581376910209656,
    "learning_rate": 5.7927243342888535e-06,
    "timestep_mean": 419.09375,
    "epsilon_norm": 8.102608680725098,
    "prediction_norm": 5.80554723739624,
    "timestamp": "2025-07-14T12:34:49.318509"
  },
  {
    "epoch": 2097,
    "batch_idx": 96,
    "loss": 0.20574551820755005,
    "learning_rate": 5.7927243342888535e-06,
    "timestep_mean": 534.34375,
    "epsilon_norm": 7.653617858886719,
    "prediction_norm": 6.790566921234131,
    "timestamp": "2025-07-14T12:34:49.323582"
  },
  {
    "epoch": 2097,
    "batch_idx": 128,
    "loss": 0.19191959500312805,
    "learning_rate": 5.7927243342888535e-06,
    "timestep_mean": 496.8125,
    "epsilon_norm": 7.181809902191162,
    "prediction_norm": 6.444633483886719,
    "timestamp": "2025-07-14T12:34:49.328780"
  },
  {
    "epoch": 2097,
    "batch_idx": 160,
    "loss": 0.33922290802001953,
    "learning_rate": 5.7927243342888535e-06,
    "timestep_mean": 493.34375,
    "epsilon_norm": 7.868530750274658,
    "prediction_norm": 6.531527042388916,
    "timestamp": "2025-07-14T12:34:49.334068"
  },
  {
    "epoch": 2097,
    "batch_idx": 192,
    "loss": 0.4417344331741333,
    "learning_rate": 5.7927243342888535e-06,
    "timestep_mean": 388.0,
    "epsilon_norm": 5.3827104568481445,
    "prediction_norm": 4.220565319061279,
    "timestamp": "2025-07-14T12:34:49.339160"
  },
  {
    "epoch": 2098,
    "batch_idx": 0,
    "loss": 0.20877106487751007,
    "learning_rate": 5.9125404448597904e-06,
    "timestep_mean": 491.71875,
    "epsilon_norm": 8.317601203918457,
    "prediction_norm": 7.181787490844727,
    "timestamp": "2025-07-14T12:34:49.344848"
  },
  {
    "epoch": 2098,
    "batch_idx": 32,
    "loss": 0.30230921506881714,
    "learning_rate": 5.9125404448597904e-06,
    "timestep_mean": 519.96875,
    "epsilon_norm": 7.9766411781311035,
    "prediction_norm": 5.659332752227783,
    "timestamp": "2025-07-14T12:34:49.350478"
  },
  {
    "epoch": 2098,
    "batch_idx": 64,
    "loss": 0.23616228997707367,
    "learning_rate": 5.9125404448597904e-06,
    "timestep_mean": 497.96875,
    "epsilon_norm": 9.13421630859375,
    "prediction_norm": 7.292116165161133,
    "timestamp": "2025-07-14T12:34:49.355742"
  },
  {
    "epoch": 2098,
    "batch_idx": 96,
    "loss": 0.4025583267211914,
    "learning_rate": 5.9125404448597904e-06,
    "timestep_mean": 434.5625,
    "epsilon_norm": 7.3364434242248535,
    "prediction_norm": 6.192653656005859,
    "timestamp": "2025-07-14T12:34:49.361029"
  },
  {
    "epoch": 2098,
    "batch_idx": 128,
    "loss": 0.3079070448875427,
    "learning_rate": 5.9125404448597904e-06,
    "timestep_mean": 453.34375,
    "epsilon_norm": 9.024998664855957,
    "prediction_norm": 7.484658241271973,
    "timestamp": "2025-07-14T12:34:49.366415"
  },
  {
    "epoch": 2098,
    "batch_idx": 160,
    "loss": 0.21396145224571228,
    "learning_rate": 5.9125404448597904e-06,
    "timestep_mean": 521.03125,
    "epsilon_norm": 7.57412052154541,
    "prediction_norm": 6.940417766571045,
    "timestamp": "2025-07-14T12:34:49.371949"
  },
  {
    "epoch": 2098,
    "batch_idx": 192,
    "loss": 0.04968607798218727,
    "learning_rate": 5.9125404448597904e-06,
    "timestep_mean": 494.875,
    "epsilon_norm": 3.846881628036499,
    "prediction_norm": 3.2948172092437744,
    "timestamp": "2025-07-14T12:34:49.377045"
  },
  {
    "epoch": 2099,
    "batch_idx": 0,
    "loss": 0.2771025598049164,
    "learning_rate": 6.033575667121269e-06,
    "timestep_mean": 527.21875,
    "epsilon_norm": 9.493009567260742,
    "prediction_norm": 8.46825122833252,
    "timestamp": "2025-07-14T12:34:49.382074"
  },
  {
    "epoch": 2099,
    "batch_idx": 32,
    "loss": 0.21062667667865753,
    "learning_rate": 6.033575667121269e-06,
    "timestep_mean": 512.3125,
    "epsilon_norm": 7.212937831878662,
    "prediction_norm": 6.752081871032715,
    "timestamp": "2025-07-14T12:34:49.386999"
  },
  {
    "epoch": 2099,
    "batch_idx": 64,
    "loss": 0.178972989320755,
    "learning_rate": 6.033575667121269e-06,
    "timestep_mean": 554.78125,
    "epsilon_norm": 7.39589262008667,
    "prediction_norm": 6.106764316558838,
    "timestamp": "2025-07-14T12:34:49.392001"
  },
  {
    "epoch": 2099,
    "batch_idx": 96,
    "loss": 0.24138742685317993,
    "learning_rate": 6.033575667121269e-06,
    "timestep_mean": 467.375,
    "epsilon_norm": 6.74979305267334,
    "prediction_norm": 5.4151692390441895,
    "timestamp": "2025-07-14T12:34:49.396873"
  },
  {
    "epoch": 2099,
    "batch_idx": 128,
    "loss": 0.41805747151374817,
    "learning_rate": 6.033575667121269e-06,
    "timestep_mean": 494.53125,
    "epsilon_norm": 8.13135051727295,
    "prediction_norm": 6.909759521484375,
    "timestamp": "2025-07-14T12:34:49.401737"
  },
  {
    "epoch": 2099,
    "batch_idx": 160,
    "loss": 0.24709507822990417,
    "learning_rate": 6.033575667121269e-06,
    "timestep_mean": 401.3125,
    "epsilon_norm": 7.874423980712891,
    "prediction_norm": 6.65081262588501,
    "timestamp": "2025-07-14T12:34:49.406616"
  },
  {
    "epoch": 2099,
    "batch_idx": 192,
    "loss": 0.04611313343048096,
    "learning_rate": 6.033575667121269e-06,
    "timestep_mean": 613.875,
    "epsilon_norm": 3.9623076915740967,
    "prediction_norm": 4.060735702514648,
    "timestamp": "2025-07-14T12:34:49.411447"
  },
  {
    "epoch": 2100,
    "batch_idx": 0,
    "loss": 0.4740096628665924,
    "learning_rate": 6.155829702431124e-06,
    "timestep_mean": 452.9375,
    "epsilon_norm": 8.795160293579102,
    "prediction_norm": 7.602890968322754,
    "timestamp": "2025-07-14T12:34:49.416854"
  },
  {
    "epoch": 2100,
    "batch_idx": 32,
    "loss": 0.1552344113588333,
    "learning_rate": 6.155829702431124e-06,
    "timestep_mean": 520.78125,
    "epsilon_norm": 9.532439231872559,
    "prediction_norm": 8.576472282409668,
    "timestamp": "2025-07-14T12:34:49.422169"
  },
  {
    "epoch": 2100,
    "batch_idx": 64,
    "loss": 0.20078927278518677,
    "learning_rate": 6.155829702431124e-06,
    "timestep_mean": 504.46875,
    "epsilon_norm": 7.833801746368408,
    "prediction_norm": 7.148561000823975,
    "timestamp": "2025-07-14T12:34:49.427465"
  },
  {
    "epoch": 2100,
    "batch_idx": 96,
    "loss": 0.26920896768569946,
    "learning_rate": 6.155829702431124e-06,
    "timestep_mean": 544.4375,
    "epsilon_norm": 8.120203018188477,
    "prediction_norm": 6.577528953552246,
    "timestamp": "2025-07-14T12:34:49.432938"
  },
  {
    "epoch": 2100,
    "batch_idx": 128,
    "loss": 0.11967171728610992,
    "learning_rate": 6.155829702431124e-06,
    "timestep_mean": 596.125,
    "epsilon_norm": 8.322478294372559,
    "prediction_norm": 7.913330554962158,
    "timestamp": "2025-07-14T12:34:49.438349"
  },
  {
    "epoch": 2100,
    "batch_idx": 160,
    "loss": 0.2718680500984192,
    "learning_rate": 6.155829702431124e-06,
    "timestep_mean": 505.21875,
    "epsilon_norm": 7.928926944732666,
    "prediction_norm": 6.421165943145752,
    "timestamp": "2025-07-14T12:34:49.443627"
  },
  {
    "epoch": 2100,
    "batch_idx": 192,
    "loss": 0.08657123893499374,
    "learning_rate": 6.155829702431124e-06,
    "timestep_mean": 503.125,
    "epsilon_norm": 4.512248992919922,
    "prediction_norm": 4.641605854034424,
    "timestamp": "2025-07-14T12:34:49.449154"
  },
  {
    "epoch": 2101,
    "batch_idx": 0,
    "loss": 0.47740882635116577,
    "learning_rate": 6.279302249139594e-06,
    "timestep_mean": 396.09375,
    "epsilon_norm": 8.31118106842041,
    "prediction_norm": 6.037235260009766,
    "timestamp": "2025-07-14T12:34:49.454495"
  },
  {
    "epoch": 2101,
    "batch_idx": 32,
    "loss": 0.31813621520996094,
    "learning_rate": 6.279302249139594e-06,
    "timestep_mean": 424.71875,
    "epsilon_norm": 9.413969993591309,
    "prediction_norm": 7.393683433532715,
    "timestamp": "2025-07-14T12:34:49.459807"
  },
  {
    "epoch": 2101,
    "batch_idx": 64,
    "loss": 0.17705342173576355,
    "learning_rate": 6.279302249139594e-06,
    "timestep_mean": 491.6875,
    "epsilon_norm": 6.850361347198486,
    "prediction_norm": 6.3313703536987305,
    "timestamp": "2025-07-14T12:34:49.465084"
  },
  {
    "epoch": 2101,
    "batch_idx": 96,
    "loss": 0.32525134086608887,
    "learning_rate": 6.279302249139594e-06,
    "timestep_mean": 525.59375,
    "epsilon_norm": 7.779392242431641,
    "prediction_norm": 6.56755256652832,
    "timestamp": "2025-07-14T12:34:49.470436"
  },
  {
    "epoch": 2101,
    "batch_idx": 128,
    "loss": 0.3400465250015259,
    "learning_rate": 6.279302249139594e-06,
    "timestep_mean": 417.78125,
    "epsilon_norm": 8.275511741638184,
    "prediction_norm": 7.109566688537598,
    "timestamp": "2025-07-14T12:34:49.475947"
  },
  {
    "epoch": 2101,
    "batch_idx": 160,
    "loss": 0.36699944734573364,
    "learning_rate": 6.279302249139594e-06,
    "timestep_mean": 510.65625,
    "epsilon_norm": 7.916501998901367,
    "prediction_norm": 5.88120174407959,
    "timestamp": "2025-07-14T12:34:49.481129"
  },
  {
    "epoch": 2101,
    "batch_idx": 192,
    "loss": 0.3953549265861511,
    "learning_rate": 6.279302249139594e-06,
    "timestep_mean": 451.625,
    "epsilon_norm": 3.495373249053955,
    "prediction_norm": 2.6366989612579346,
    "timestamp": "2025-07-14T12:34:49.486295"
  },
  {
    "epoch": 2102,
    "batch_idx": 0,
    "loss": 0.20861825346946716,
    "learning_rate": 6.4039930025903776e-06,
    "timestep_mean": 470.84375,
    "epsilon_norm": 8.9517822265625,
    "prediction_norm": 7.940133094787598,
    "timestamp": "2025-07-14T12:34:49.491564"
  },
  {
    "epoch": 2102,
    "batch_idx": 32,
    "loss": 0.19905205070972443,
    "learning_rate": 6.4039930025903776e-06,
    "timestep_mean": 518.84375,
    "epsilon_norm": 7.957282066345215,
    "prediction_norm": 7.019738674163818,
    "timestamp": "2025-07-14T12:34:49.496915"
  },
  {
    "epoch": 2102,
    "batch_idx": 64,
    "loss": 0.17159153521060944,
    "learning_rate": 6.4039930025903776e-06,
    "timestep_mean": 521.59375,
    "epsilon_norm": 7.697997093200684,
    "prediction_norm": 7.279221534729004,
    "timestamp": "2025-07-14T12:34:49.502097"
  },
  {
    "epoch": 2102,
    "batch_idx": 96,
    "loss": 0.29336678981781006,
    "learning_rate": 6.4039930025903776e-06,
    "timestep_mean": 577.9375,
    "epsilon_norm": 8.136344909667969,
    "prediction_norm": 7.0326457023620605,
    "timestamp": "2025-07-14T12:34:49.506933"
  },
  {
    "epoch": 2102,
    "batch_idx": 128,
    "loss": 0.2414238154888153,
    "learning_rate": 6.4039930025903776e-06,
    "timestep_mean": 512.6875,
    "epsilon_norm": 7.806125164031982,
    "prediction_norm": 6.738033771514893,
    "timestamp": "2025-07-14T12:34:49.511942"
  },
  {
    "epoch": 2102,
    "batch_idx": 160,
    "loss": 0.26754119992256165,
    "learning_rate": 6.4039930025903776e-06,
    "timestep_mean": 467.28125,
    "epsilon_norm": 6.900285720825195,
    "prediction_norm": 5.818763732910156,
    "timestamp": "2025-07-14T12:34:49.516959"
  },
  {
    "epoch": 2102,
    "batch_idx": 192,
    "loss": 0.036226630210876465,
    "learning_rate": 6.4039930025903776e-06,
    "timestep_mean": 730.125,
    "epsilon_norm": 4.843022346496582,
    "prediction_norm": 4.6464362144470215,
    "timestamp": "2025-07-14T12:34:49.521826"
  },
  {
    "epoch": 2103,
    "batch_idx": 0,
    "loss": 0.41206175088882446,
    "learning_rate": 6.529901655121581e-06,
    "timestep_mean": 370.5,
    "epsilon_norm": 9.516865730285645,
    "prediction_norm": 7.489861965179443,
    "timestamp": "2025-07-14T12:34:49.526831"
  },
  {
    "epoch": 2103,
    "batch_idx": 32,
    "loss": 0.25893720984458923,
    "learning_rate": 6.529901655121581e-06,
    "timestep_mean": 485.5,
    "epsilon_norm": 7.463970184326172,
    "prediction_norm": 6.245032787322998,
    "timestamp": "2025-07-14T12:34:49.531933"
  },
  {
    "epoch": 2103,
    "batch_idx": 64,
    "loss": 0.22668012976646423,
    "learning_rate": 6.529901655121581e-06,
    "timestep_mean": 535.5,
    "epsilon_norm": 8.623735427856445,
    "prediction_norm": 7.333003044128418,
    "timestamp": "2025-07-14T12:34:49.537115"
  },
  {
    "epoch": 2103,
    "batch_idx": 96,
    "loss": 0.255912721157074,
    "learning_rate": 6.529901655121581e-06,
    "timestep_mean": 488.65625,
    "epsilon_norm": 7.467203140258789,
    "prediction_norm": 6.70416784286499,
    "timestamp": "2025-07-14T12:34:49.542377"
  },
  {
    "epoch": 2103,
    "batch_idx": 128,
    "loss": 0.23970767855644226,
    "learning_rate": 6.529901655121581e-06,
    "timestep_mean": 515.09375,
    "epsilon_norm": 8.19015884399414,
    "prediction_norm": 7.447140693664551,
    "timestamp": "2025-07-14T12:34:49.547586"
  },
  {
    "epoch": 2103,
    "batch_idx": 160,
    "loss": 0.2432488352060318,
    "learning_rate": 6.529901655121581e-06,
    "timestep_mean": 500.46875,
    "epsilon_norm": 6.908351898193359,
    "prediction_norm": 5.599534034729004,
    "timestamp": "2025-07-14T12:34:49.552788"
  },
  {
    "epoch": 2103,
    "batch_idx": 192,
    "loss": 0.29067468643188477,
    "learning_rate": 6.529901655121581e-06,
    "timestep_mean": 302.125,
    "epsilon_norm": 4.783220291137695,
    "prediction_norm": 4.351367950439453,
    "timestamp": "2025-07-14T12:34:49.557969"
  },
  {
    "epoch": 2104,
    "batch_idx": 0,
    "loss": 0.2432597279548645,
    "learning_rate": 6.65702789606599e-06,
    "timestep_mean": 492.59375,
    "epsilon_norm": 7.733436584472656,
    "prediction_norm": 6.393580913543701,
    "timestamp": "2025-07-14T12:34:49.563171"
  },
  {
    "epoch": 2104,
    "batch_idx": 32,
    "loss": 0.31446972489356995,
    "learning_rate": 6.65702789606599e-06,
    "timestep_mean": 576.15625,
    "epsilon_norm": 7.899984836578369,
    "prediction_norm": 6.223567485809326,
    "timestamp": "2025-07-14T12:34:49.568263"
  },
  {
    "epoch": 2104,
    "batch_idx": 64,
    "loss": 0.2730160355567932,
    "learning_rate": 6.65702789606599e-06,
    "timestep_mean": 533.84375,
    "epsilon_norm": 8.648172378540039,
    "prediction_norm": 7.183347225189209,
    "timestamp": "2025-07-14T12:34:49.573426"
  },
  {
    "epoch": 2104,
    "batch_idx": 96,
    "loss": 0.43712443113327026,
    "learning_rate": 6.65702789606599e-06,
    "timestep_mean": 501.53125,
    "epsilon_norm": 8.370342254638672,
    "prediction_norm": 6.162170886993408,
    "timestamp": "2025-07-14T12:34:49.578666"
  },
  {
    "epoch": 2104,
    "batch_idx": 128,
    "loss": 0.3894612193107605,
    "learning_rate": 6.65702789606599e-06,
    "timestep_mean": 445.875,
    "epsilon_norm": 7.971068859100342,
    "prediction_norm": 6.024843692779541,
    "timestamp": "2025-07-14T12:34:49.583979"
  },
  {
    "epoch": 2104,
    "batch_idx": 160,
    "loss": 0.15488740801811218,
    "learning_rate": 6.65702789606599e-06,
    "timestep_mean": 489.25,
    "epsilon_norm": 7.895293712615967,
    "prediction_norm": 7.711811065673828,
    "timestamp": "2025-07-14T12:34:49.589121"
  },
  {
    "epoch": 2104,
    "batch_idx": 192,
    "loss": 0.18669648468494415,
    "learning_rate": 6.65702789606599e-06,
    "timestep_mean": 514.625,
    "epsilon_norm": 4.006534576416016,
    "prediction_norm": 3.9666943550109863,
    "timestamp": "2025-07-14T12:34:49.594280"
  },
  {
    "epoch": 2105,
    "batch_idx": 0,
    "loss": 0.21478478610515594,
    "learning_rate": 6.785371411752237e-06,
    "timestep_mean": 492.59375,
    "epsilon_norm": 7.11856746673584,
    "prediction_norm": 6.707346439361572,
    "timestamp": "2025-07-14T12:34:49.599352"
  },
  {
    "epoch": 2105,
    "batch_idx": 32,
    "loss": 0.21294213831424713,
    "learning_rate": 6.785371411752237e-06,
    "timestep_mean": 534.40625,
    "epsilon_norm": 7.277726650238037,
    "prediction_norm": 6.066823959350586,
    "timestamp": "2025-07-14T12:34:49.604394"
  },
  {
    "epoch": 2105,
    "batch_idx": 64,
    "loss": 0.260387659072876,
    "learning_rate": 6.785371411752237e-06,
    "timestep_mean": 473.875,
    "epsilon_norm": 7.5526347160339355,
    "prediction_norm": 6.795938968658447,
    "timestamp": "2025-07-14T12:34:49.609441"
  },
  {
    "epoch": 2105,
    "batch_idx": 96,
    "loss": 0.18242916464805603,
    "learning_rate": 6.785371411752237e-06,
    "timestep_mean": 568.4375,
    "epsilon_norm": 7.23154878616333,
    "prediction_norm": 6.748368263244629,
    "timestamp": "2025-07-14T12:34:49.614599"
  },
  {
    "epoch": 2105,
    "batch_idx": 128,
    "loss": 0.3139030635356903,
    "learning_rate": 6.785371411752237e-06,
    "timestep_mean": 543.1875,
    "epsilon_norm": 9.078805923461914,
    "prediction_norm": 7.478227138519287,
    "timestamp": "2025-07-14T12:34:49.619743"
  },
  {
    "epoch": 2105,
    "batch_idx": 160,
    "loss": 0.5170981884002686,
    "learning_rate": 6.785371411752237e-06,
    "timestep_mean": 459.0,
    "epsilon_norm": 7.607950210571289,
    "prediction_norm": 5.511213302612305,
    "timestamp": "2025-07-14T12:34:49.624736"
  },
  {
    "epoch": 2105,
    "batch_idx": 192,
    "loss": 0.2712671160697937,
    "learning_rate": 6.785371411752237e-06,
    "timestep_mean": 549.0,
    "epsilon_norm": 3.1258034706115723,
    "prediction_norm": 1.9382973909378052,
    "timestamp": "2025-07-14T12:34:49.629682"
  },
  {
    "epoch": 2106,
    "batch_idx": 0,
    "loss": 0.19709739089012146,
    "learning_rate": 6.91493188550558e-06,
    "timestep_mean": 526.9375,
    "epsilon_norm": 7.468574523925781,
    "prediction_norm": 6.7696638107299805,
    "timestamp": "2025-07-14T12:34:49.634601"
  },
  {
    "epoch": 2106,
    "batch_idx": 32,
    "loss": 0.21645738184452057,
    "learning_rate": 6.91493188550558e-06,
    "timestep_mean": 516.25,
    "epsilon_norm": 8.465944290161133,
    "prediction_norm": 7.159270286560059,
    "timestamp": "2025-07-14T12:34:49.639567"
  },
  {
    "epoch": 2106,
    "batch_idx": 64,
    "loss": 0.21278059482574463,
    "learning_rate": 6.91493188550558e-06,
    "timestep_mean": 543.375,
    "epsilon_norm": 9.024538040161133,
    "prediction_norm": 8.320239067077637,
    "timestamp": "2025-07-14T12:34:49.644433"
  },
  {
    "epoch": 2106,
    "batch_idx": 96,
    "loss": 0.22917801141738892,
    "learning_rate": 6.91493188550558e-06,
    "timestep_mean": 583.71875,
    "epsilon_norm": 7.770428657531738,
    "prediction_norm": 6.7758588790893555,
    "timestamp": "2025-07-14T12:34:49.649317"
  },
  {
    "epoch": 2106,
    "batch_idx": 128,
    "loss": 0.36065176129341125,
    "learning_rate": 6.91493188550558e-06,
    "timestep_mean": 508.1875,
    "epsilon_norm": 7.335065841674805,
    "prediction_norm": 6.717140197753906,
    "timestamp": "2025-07-14T12:34:49.654394"
  },
  {
    "epoch": 2106,
    "batch_idx": 160,
    "loss": 0.17885848879814148,
    "learning_rate": 6.91493188550558e-06,
    "timestep_mean": 516.6875,
    "epsilon_norm": 7.798805236816406,
    "prediction_norm": 7.432170391082764,
    "timestamp": "2025-07-14T12:34:49.659443"
  },
  {
    "epoch": 2106,
    "batch_idx": 192,
    "loss": 0.5267356634140015,
    "learning_rate": 6.91493188550558e-06,
    "timestep_mean": 403.25,
    "epsilon_norm": 3.860264301300049,
    "prediction_norm": 2.8043875694274902,
    "timestamp": "2025-07-14T12:34:49.664429"
  },
  {
    "epoch": 2107,
    "batch_idx": 0,
    "loss": 0.1763809621334076,
    "learning_rate": 7.045708997648346e-06,
    "timestep_mean": 518.15625,
    "epsilon_norm": 7.366932392120361,
    "prediction_norm": 6.829412937164307,
    "timestamp": "2025-07-14T12:34:49.669301"
  },
  {
    "epoch": 2107,
    "batch_idx": 32,
    "loss": 0.31136664748191833,
    "learning_rate": 7.045708997648346e-06,
    "timestep_mean": 416.25,
    "epsilon_norm": 8.447383880615234,
    "prediction_norm": 7.267314434051514,
    "timestamp": "2025-07-14T12:34:49.674353"
  },
  {
    "epoch": 2107,
    "batch_idx": 64,
    "loss": 0.37044936418533325,
    "learning_rate": 7.045708997648346e-06,
    "timestep_mean": 493.34375,
    "epsilon_norm": 8.594968795776367,
    "prediction_norm": 6.929251670837402,
    "timestamp": "2025-07-14T12:34:49.679436"
  },
  {
    "epoch": 2107,
    "batch_idx": 96,
    "loss": 0.40707334876060486,
    "learning_rate": 7.045708997648346e-06,
    "timestep_mean": 483.4375,
    "epsilon_norm": 7.390455722808838,
    "prediction_norm": 5.4650774002075195,
    "timestamp": "2025-07-14T12:34:49.684354"
  },
  {
    "epoch": 2107,
    "batch_idx": 128,
    "loss": 0.3523852229118347,
    "learning_rate": 7.045708997648346e-06,
    "timestep_mean": 459.625,
    "epsilon_norm": 9.013060569763184,
    "prediction_norm": 7.472143650054932,
    "timestamp": "2025-07-14T12:34:49.689264"
  },
  {
    "epoch": 2107,
    "batch_idx": 160,
    "loss": 0.22606244683265686,
    "learning_rate": 7.045708997648346e-06,
    "timestep_mean": 512.5625,
    "epsilon_norm": 7.8287224769592285,
    "prediction_norm": 6.623382568359375,
    "timestamp": "2025-07-14T12:34:49.694148"
  },
  {
    "epoch": 2107,
    "batch_idx": 192,
    "loss": 0.642211377620697,
    "learning_rate": 7.045708997648346e-06,
    "timestep_mean": 533.375,
    "epsilon_norm": 4.2330546379089355,
    "prediction_norm": 3.301281452178955,
    "timestamp": "2025-07-14T12:34:49.699073"
  },
  {
    "epoch": 2108,
    "batch_idx": 0,
    "loss": 0.3118034899234772,
    "learning_rate": 7.177702425500985e-06,
    "timestep_mean": 400.375,
    "epsilon_norm": 8.170194625854492,
    "prediction_norm": 6.212399482727051,
    "timestamp": "2025-07-14T12:34:49.704015"
  },
  {
    "epoch": 2108,
    "batch_idx": 32,
    "loss": 0.21522772312164307,
    "learning_rate": 7.177702425500985e-06,
    "timestep_mean": 480.03125,
    "epsilon_norm": 8.19485855102539,
    "prediction_norm": 7.872549533843994,
    "timestamp": "2025-07-14T12:34:49.712826"
  },
  {
    "epoch": 2108,
    "batch_idx": 64,
    "loss": 0.4453587830066681,
    "learning_rate": 7.177702425500985e-06,
    "timestep_mean": 329.40625,
    "epsilon_norm": 7.916881561279297,
    "prediction_norm": 6.284536838531494,
    "timestamp": "2025-07-14T12:34:49.718469"
  },
  {
    "epoch": 2108,
    "batch_idx": 96,
    "loss": 0.2803206443786621,
    "learning_rate": 7.177702425500985e-06,
    "timestep_mean": 422.1875,
    "epsilon_norm": 8.173490524291992,
    "prediction_norm": 7.253833770751953,
    "timestamp": "2025-07-14T12:34:49.723723"
  },
  {
    "epoch": 2108,
    "batch_idx": 128,
    "loss": 0.44764307141304016,
    "learning_rate": 7.177702425500985e-06,
    "timestep_mean": 563.25,
    "epsilon_norm": 9.441732406616211,
    "prediction_norm": 7.542068958282471,
    "timestamp": "2025-07-14T12:34:49.728977"
  },
  {
    "epoch": 2108,
    "batch_idx": 160,
    "loss": 0.2683737874031067,
    "learning_rate": 7.177702425500985e-06,
    "timestep_mean": 548.09375,
    "epsilon_norm": 8.605374336242676,
    "prediction_norm": 7.393214702606201,
    "timestamp": "2025-07-14T12:34:49.734193"
  },
  {
    "epoch": 2108,
    "batch_idx": 192,
    "loss": 0.11092934012413025,
    "learning_rate": 7.177702425500985e-06,
    "timestep_mean": 611.5,
    "epsilon_norm": 4.324313163757324,
    "prediction_norm": 3.829854965209961,
    "timestamp": "2025-07-14T12:34:49.739510"
  },
  {
    "epoch": 2109,
    "batch_idx": 0,
    "loss": 0.22669988870620728,
    "learning_rate": 7.310911843382904e-06,
    "timestep_mean": 491.21875,
    "epsilon_norm": 8.431713104248047,
    "prediction_norm": 7.974897861480713,
    "timestamp": "2025-07-14T12:34:49.744734"
  },
  {
    "epoch": 2109,
    "batch_idx": 32,
    "loss": 0.13679566979408264,
    "learning_rate": 7.310911843382904e-06,
    "timestep_mean": 537.28125,
    "epsilon_norm": 7.36201810836792,
    "prediction_norm": 6.81635046005249,
    "timestamp": "2025-07-14T12:34:49.749969"
  },
  {
    "epoch": 2109,
    "batch_idx": 64,
    "loss": 0.28200972080230713,
    "learning_rate": 7.310911843382904e-06,
    "timestep_mean": 586.59375,
    "epsilon_norm": 8.294234275817871,
    "prediction_norm": 6.898576259613037,
    "timestamp": "2025-07-14T12:34:49.755242"
  },
  {
    "epoch": 2109,
    "batch_idx": 96,
    "loss": 0.1911783367395401,
    "learning_rate": 7.310911843382904e-06,
    "timestep_mean": 501.65625,
    "epsilon_norm": 7.233455657958984,
    "prediction_norm": 6.527870178222656,
    "timestamp": "2025-07-14T12:34:49.760461"
  },
  {
    "epoch": 2109,
    "batch_idx": 128,
    "loss": 0.2180316150188446,
    "learning_rate": 7.310911843382904e-06,
    "timestep_mean": 503.6875,
    "epsilon_norm": 7.786250591278076,
    "prediction_norm": 6.500145435333252,
    "timestamp": "2025-07-14T12:34:49.765713"
  },
  {
    "epoch": 2109,
    "batch_idx": 160,
    "loss": 0.14514559507369995,
    "learning_rate": 7.310911843382904e-06,
    "timestep_mean": 543.65625,
    "epsilon_norm": 7.801000118255615,
    "prediction_norm": 6.729719638824463,
    "timestamp": "2025-07-14T12:34:49.770911"
  },
  {
    "epoch": 2109,
    "batch_idx": 192,
    "loss": 0.1346787065267563,
    "learning_rate": 7.310911843382904e-06,
    "timestep_mean": 491.625,
    "epsilon_norm": 2.929720640182495,
    "prediction_norm": 2.08567214012146,
    "timestamp": "2025-07-14T12:34:49.776088"
  },
  {
    "epoch": 2110,
    "batch_idx": 0,
    "loss": 0.15706585347652435,
    "learning_rate": 7.445336922613021e-06,
    "timestep_mean": 528.65625,
    "epsilon_norm": 7.801656723022461,
    "prediction_norm": 7.2441582679748535,
    "timestamp": "2025-07-14T12:34:49.781334"
  },
  {
    "epoch": 2110,
    "batch_idx": 32,
    "loss": 0.16741931438446045,
    "learning_rate": 7.445336922613021e-06,
    "timestep_mean": 538.90625,
    "epsilon_norm": 7.305515289306641,
    "prediction_norm": 6.655752182006836,
    "timestamp": "2025-07-14T12:34:49.786608"
  },
  {
    "epoch": 2110,
    "batch_idx": 64,
    "loss": 0.11097636818885803,
    "learning_rate": 7.445336922613021e-06,
    "timestep_mean": 565.15625,
    "epsilon_norm": 6.346826553344727,
    "prediction_norm": 6.377520561218262,
    "timestamp": "2025-07-14T12:34:49.791902"
  },
  {
    "epoch": 2110,
    "batch_idx": 96,
    "loss": 0.2831442654132843,
    "learning_rate": 7.445336922613021e-06,
    "timestep_mean": 499.84375,
    "epsilon_norm": 8.474340438842773,
    "prediction_norm": 7.427798271179199,
    "timestamp": "2025-07-14T12:34:49.797060"
  },
  {
    "epoch": 2110,
    "batch_idx": 128,
    "loss": 0.43554916977882385,
    "learning_rate": 7.445336922613021e-06,
    "timestep_mean": 490.71875,
    "epsilon_norm": 8.457306861877441,
    "prediction_norm": 5.986133575439453,
    "timestamp": "2025-07-14T12:34:49.802301"
  },
  {
    "epoch": 2110,
    "batch_idx": 160,
    "loss": 0.5395121574401855,
    "learning_rate": 7.445336922613021e-06,
    "timestep_mean": 448.125,
    "epsilon_norm": 8.581637382507324,
    "prediction_norm": 6.264869213104248,
    "timestamp": "2025-07-14T12:34:49.807467"
  },
  {
    "epoch": 2110,
    "batch_idx": 192,
    "loss": 0.5350221395492554,
    "learning_rate": 7.445336922613021e-06,
    "timestep_mean": 540.75,
    "epsilon_norm": 3.4128806591033936,
    "prediction_norm": 2.119846820831299,
    "timestamp": "2025-07-14T12:34:49.812592"
  },
  {
    "epoch": 2111,
    "batch_idx": 0,
    "loss": 0.22767050564289093,
    "learning_rate": 7.580977331510929e-06,
    "timestep_mean": 569.90625,
    "epsilon_norm": 8.1082181930542,
    "prediction_norm": 7.585517883300781,
    "timestamp": "2025-07-14T12:34:49.817783"
  },
  {
    "epoch": 2111,
    "batch_idx": 32,
    "loss": 0.3919672966003418,
    "learning_rate": 7.580977331510929e-06,
    "timestep_mean": 466.5625,
    "epsilon_norm": 9.507607460021973,
    "prediction_norm": 8.139986038208008,
    "timestamp": "2025-07-14T12:34:49.823004"
  },
  {
    "epoch": 2111,
    "batch_idx": 64,
    "loss": 0.18500638008117676,
    "learning_rate": 7.580977331510929e-06,
    "timestep_mean": 513.34375,
    "epsilon_norm": 7.93490743637085,
    "prediction_norm": 7.216535568237305,
    "timestamp": "2025-07-14T12:34:49.828195"
  },
  {
    "epoch": 2111,
    "batch_idx": 96,
    "loss": 0.24412357807159424,
    "learning_rate": 7.580977331510929e-06,
    "timestep_mean": 472.5,
    "epsilon_norm": 8.45362663269043,
    "prediction_norm": 7.3346099853515625,
    "timestamp": "2025-07-14T12:34:49.833146"
  },
  {
    "epoch": 2111,
    "batch_idx": 128,
    "loss": 0.39405280351638794,
    "learning_rate": 7.580977331510929e-06,
    "timestep_mean": 476.0,
    "epsilon_norm": 7.322546482086182,
    "prediction_norm": 6.3071393966674805,
    "timestamp": "2025-07-14T12:34:49.838422"
  },
  {
    "epoch": 2111,
    "batch_idx": 160,
    "loss": 0.29096126556396484,
    "learning_rate": 7.580977331510929e-06,
    "timestep_mean": 501.53125,
    "epsilon_norm": 7.638833999633789,
    "prediction_norm": 6.894816875457764,
    "timestamp": "2025-07-14T12:34:49.843524"
  },
  {
    "epoch": 2111,
    "batch_idx": 192,
    "loss": 0.28687602281570435,
    "learning_rate": 7.580977331510929e-06,
    "timestep_mean": 435.625,
    "epsilon_norm": 2.786146402359009,
    "prediction_norm": 2.5782289505004883,
    "timestamp": "2025-07-14T12:34:49.848651"
  },
  {
    "epoch": 2112,
    "batch_idx": 0,
    "loss": 0.35091495513916016,
    "learning_rate": 7.717832735397344e-06,
    "timestep_mean": 560.21875,
    "epsilon_norm": 8.827651977539062,
    "prediction_norm": 7.503089904785156,
    "timestamp": "2025-07-14T12:34:49.853867"
  },
  {
    "epoch": 2112,
    "batch_idx": 32,
    "loss": 0.2999611496925354,
    "learning_rate": 7.717832735397344e-06,
    "timestep_mean": 448.5625,
    "epsilon_norm": 7.242712497711182,
    "prediction_norm": 6.201826095581055,
    "timestamp": "2025-07-14T12:34:49.859230"
  },
  {
    "epoch": 2112,
    "batch_idx": 64,
    "loss": 0.3384704291820526,
    "learning_rate": 7.717832735397344e-06,
    "timestep_mean": 505.78125,
    "epsilon_norm": 7.5200395584106445,
    "prediction_norm": 6.325972080230713,
    "timestamp": "2025-07-14T12:34:49.864339"
  },
  {
    "epoch": 2112,
    "batch_idx": 96,
    "loss": 0.1353955864906311,
    "learning_rate": 7.717832735397344e-06,
    "timestep_mean": 546.3125,
    "epsilon_norm": 7.499261856079102,
    "prediction_norm": 6.85169792175293,
    "timestamp": "2025-07-14T12:34:49.869560"
  },
  {
    "epoch": 2112,
    "batch_idx": 128,
    "loss": 0.3983702063560486,
    "learning_rate": 7.717832735397344e-06,
    "timestep_mean": 453.375,
    "epsilon_norm": 10.093677520751953,
    "prediction_norm": 7.973222732543945,
    "timestamp": "2025-07-14T12:34:49.874835"
  },
  {
    "epoch": 2112,
    "batch_idx": 160,
    "loss": 0.13582393527030945,
    "learning_rate": 7.717832735397344e-06,
    "timestep_mean": 528.03125,
    "epsilon_norm": 6.7410173416137695,
    "prediction_norm": 5.8765106201171875,
    "timestamp": "2025-07-14T12:34:49.880030"
  },
  {
    "epoch": 2112,
    "batch_idx": 192,
    "loss": 0.08207820355892181,
    "learning_rate": 7.717832735397344e-06,
    "timestep_mean": 520.5,
    "epsilon_norm": 2.9298672676086426,
    "prediction_norm": 2.926783800125122,
    "timestamp": "2025-07-14T12:34:49.885119"
  },
  {
    "epoch": 2113,
    "batch_idx": 0,
    "loss": 0.14019432663917542,
    "learning_rate": 7.855902796595048e-06,
    "timestep_mean": 537.8125,
    "epsilon_norm": 7.660694122314453,
    "prediction_norm": 7.4534807205200195,
    "timestamp": "2025-07-14T12:34:49.890272"
  },
  {
    "epoch": 2113,
    "batch_idx": 32,
    "loss": 0.14345699548721313,
    "learning_rate": 7.855902796595048e-06,
    "timestep_mean": 625.78125,
    "epsilon_norm": 8.307413101196289,
    "prediction_norm": 7.6704792976379395,
    "timestamp": "2025-07-14T12:34:49.895186"
  },
  {
    "epoch": 2113,
    "batch_idx": 64,
    "loss": 0.3910417854785919,
    "learning_rate": 7.855902796595048e-06,
    "timestep_mean": 471.0625,
    "epsilon_norm": 7.555991172790527,
    "prediction_norm": 5.6147613525390625,
    "timestamp": "2025-07-14T12:34:49.900130"
  },
  {
    "epoch": 2113,
    "batch_idx": 96,
    "loss": 0.15333929657936096,
    "learning_rate": 7.855902796595048e-06,
    "timestep_mean": 579.71875,
    "epsilon_norm": 6.660264015197754,
    "prediction_norm": 5.841708183288574,
    "timestamp": "2025-07-14T12:34:49.904948"
  },
  {
    "epoch": 2113,
    "batch_idx": 128,
    "loss": 0.3529295325279236,
    "learning_rate": 7.855902796595048e-06,
    "timestep_mean": 436.65625,
    "epsilon_norm": 8.517189979553223,
    "prediction_norm": 7.313319206237793,
    "timestamp": "2025-07-14T12:34:49.909812"
  },
  {
    "epoch": 2113,
    "batch_idx": 160,
    "loss": 0.25181224942207336,
    "learning_rate": 7.855902796595048e-06,
    "timestep_mean": 411.4375,
    "epsilon_norm": 7.82291841506958,
    "prediction_norm": 6.338637351989746,
    "timestamp": "2025-07-14T12:34:49.914731"
  },
  {
    "epoch": 2113,
    "batch_idx": 192,
    "loss": 0.2280181646347046,
    "learning_rate": 7.855902796595048e-06,
    "timestep_mean": 456.5,
    "epsilon_norm": 4.921783447265625,
    "prediction_norm": 4.291921138763428,
    "timestamp": "2025-07-14T12:34:49.919602"
  },
  {
    "epoch": 2114,
    "batch_idx": 0,
    "loss": 0.31649667024612427,
    "learning_rate": 7.995187174430107e-06,
    "timestep_mean": 437.6875,
    "epsilon_norm": 7.150686264038086,
    "prediction_norm": 5.753782749176025,
    "timestamp": "2025-07-14T12:34:49.924555"
  },
  {
    "epoch": 2114,
    "batch_idx": 32,
    "loss": 0.23192480206489563,
    "learning_rate": 7.995187174430107e-06,
    "timestep_mean": 491.40625,
    "epsilon_norm": 7.712403774261475,
    "prediction_norm": 6.941391944885254,
    "timestamp": "2025-07-14T12:34:49.929485"
  },
  {
    "epoch": 2114,
    "batch_idx": 64,
    "loss": 0.3068283200263977,
    "learning_rate": 7.995187174430107e-06,
    "timestep_mean": 516.875,
    "epsilon_norm": 8.358342170715332,
    "prediction_norm": 7.433631896972656,
    "timestamp": "2025-07-14T12:34:49.934297"
  },
  {
    "epoch": 2114,
    "batch_idx": 96,
    "loss": 0.08850494027137756,
    "learning_rate": 7.995187174430107e-06,
    "timestep_mean": 589.75,
    "epsilon_norm": 7.2735772132873535,
    "prediction_norm": 6.967632293701172,
    "timestamp": "2025-07-14T12:34:49.939137"
  },
  {
    "epoch": 2114,
    "batch_idx": 128,
    "loss": 0.4254732131958008,
    "learning_rate": 7.995187174430107e-06,
    "timestep_mean": 479.25,
    "epsilon_norm": 8.022714614868164,
    "prediction_norm": 6.619296073913574,
    "timestamp": "2025-07-14T12:34:49.943990"
  },
  {
    "epoch": 2114,
    "batch_idx": 160,
    "loss": 0.3695945143699646,
    "learning_rate": 7.995187174430107e-06,
    "timestep_mean": 412.21875,
    "epsilon_norm": 8.444351196289062,
    "prediction_norm": 6.373759746551514,
    "timestamp": "2025-07-14T12:34:49.948852"
  },
  {
    "epoch": 2114,
    "batch_idx": 192,
    "loss": 1.0111470222473145,
    "learning_rate": 7.995187174430107e-06,
    "timestep_mean": 382.125,
    "epsilon_norm": 4.969861030578613,
    "prediction_norm": 2.399181604385376,
    "timestamp": "2025-07-14T12:34:49.953681"
  },
  {
    "epoch": 2115,
    "batch_idx": 0,
    "loss": 0.22931954264640808,
    "learning_rate": 8.135685525232094e-06,
    "timestep_mean": 480.03125,
    "epsilon_norm": 7.797474384307861,
    "prediction_norm": 6.6937031745910645,
    "timestamp": "2025-07-14T12:34:49.958635"
  },
  {
    "epoch": 2115,
    "batch_idx": 32,
    "loss": 0.22648176550865173,
    "learning_rate": 8.135685525232094e-06,
    "timestep_mean": 509.21875,
    "epsilon_norm": 7.3355817794799805,
    "prediction_norm": 5.835866928100586,
    "timestamp": "2025-07-14T12:34:49.963437"
  },
  {
    "epoch": 2115,
    "batch_idx": 64,
    "loss": 0.43945300579071045,
    "learning_rate": 8.135685525232094e-06,
    "timestep_mean": 442.96875,
    "epsilon_norm": 7.595931529998779,
    "prediction_norm": 6.098330497741699,
    "timestamp": "2025-07-14T12:34:49.968217"
  },
  {
    "epoch": 2115,
    "batch_idx": 96,
    "loss": 0.33647677302360535,
    "learning_rate": 8.135685525232094e-06,
    "timestep_mean": 505.96875,
    "epsilon_norm": 8.510014533996582,
    "prediction_norm": 7.3573784828186035,
    "timestamp": "2025-07-14T12:34:49.973082"
  },
  {
    "epoch": 2115,
    "batch_idx": 128,
    "loss": 0.5041690468788147,
    "learning_rate": 8.135685525232094e-06,
    "timestep_mean": 406.125,
    "epsilon_norm": 8.6204195022583,
    "prediction_norm": 6.351252555847168,
    "timestamp": "2025-07-14T12:34:49.977982"
  },
  {
    "epoch": 2115,
    "batch_idx": 160,
    "loss": 0.18879248201847076,
    "learning_rate": 8.135685525232094e-06,
    "timestep_mean": 497.375,
    "epsilon_norm": 8.846552848815918,
    "prediction_norm": 7.659143447875977,
    "timestamp": "2025-07-14T12:34:49.982818"
  },
  {
    "epoch": 2115,
    "batch_idx": 192,
    "loss": 0.08971301466226578,
    "learning_rate": 8.135685525232094e-06,
    "timestep_mean": 483.875,
    "epsilon_norm": 3.0607848167419434,
    "prediction_norm": 3.1595640182495117,
    "timestamp": "2025-07-14T12:34:49.987878"
  },
  {
    "epoch": 2116,
    "batch_idx": 0,
    "loss": 0.12340810149908066,
    "learning_rate": 8.277397502335149e-06,
    "timestep_mean": 531.0,
    "epsilon_norm": 7.955563545227051,
    "prediction_norm": 7.577998638153076,
    "timestamp": "2025-07-14T12:34:49.993015"
  },
  {
    "epoch": 2116,
    "batch_idx": 32,
    "loss": 0.159318208694458,
    "learning_rate": 8.277397502335149e-06,
    "timestep_mean": 522.78125,
    "epsilon_norm": 6.565911769866943,
    "prediction_norm": 6.1138997077941895,
    "timestamp": "2025-07-14T12:34:49.998167"
  },
  {
    "epoch": 2116,
    "batch_idx": 64,
    "loss": 0.3276192247867584,
    "learning_rate": 8.277397502335149e-06,
    "timestep_mean": 473.71875,
    "epsilon_norm": 8.547083854675293,
    "prediction_norm": 6.605646133422852,
    "timestamp": "2025-07-14T12:34:50.003355"
  },
  {
    "epoch": 2116,
    "batch_idx": 96,
    "loss": 0.35738861560821533,
    "learning_rate": 8.277397502335149e-06,
    "timestep_mean": 537.59375,
    "epsilon_norm": 8.465614318847656,
    "prediction_norm": 6.73137903213501,
    "timestamp": "2025-07-14T12:34:50.008729"
  },
  {
    "epoch": 2116,
    "batch_idx": 128,
    "loss": 0.2628602385520935,
    "learning_rate": 8.277397502335149e-06,
    "timestep_mean": 459.34375,
    "epsilon_norm": 7.682980537414551,
    "prediction_norm": 7.335683345794678,
    "timestamp": "2025-07-14T12:34:50.013725"
  },
  {
    "epoch": 2116,
    "batch_idx": 160,
    "loss": 0.10232055187225342,
    "learning_rate": 8.277397502335149e-06,
    "timestep_mean": 639.21875,
    "epsilon_norm": 7.0466156005859375,
    "prediction_norm": 6.1504034996032715,
    "timestamp": "2025-07-14T12:34:50.018747"
  },
  {
    "epoch": 2116,
    "batch_idx": 192,
    "loss": 0.5351417064666748,
    "learning_rate": 8.277397502335149e-06,
    "timestep_mean": 496.375,
    "epsilon_norm": 4.571544170379639,
    "prediction_norm": 2.971487522125244,
    "timestamp": "2025-07-14T12:34:50.023940"
  },
  {
    "epoch": 2117,
    "batch_idx": 0,
    "loss": 0.21427331864833832,
    "learning_rate": 8.420322756079247e-06,
    "timestep_mean": 486.625,
    "epsilon_norm": 6.88764762878418,
    "prediction_norm": 6.303803443908691,
    "timestamp": "2025-07-14T12:34:50.029193"
  },
  {
    "epoch": 2117,
    "batch_idx": 32,
    "loss": 0.5043656826019287,
    "learning_rate": 8.420322756079247e-06,
    "timestep_mean": 424.5625,
    "epsilon_norm": 8.636711120605469,
    "prediction_norm": 5.831522464752197,
    "timestamp": "2025-07-14T12:34:50.034377"
  },
  {
    "epoch": 2117,
    "batch_idx": 64,
    "loss": 0.23233814537525177,
    "learning_rate": 8.420322756079247e-06,
    "timestep_mean": 534.65625,
    "epsilon_norm": 7.869096279144287,
    "prediction_norm": 6.676340579986572,
    "timestamp": "2025-07-14T12:34:50.039561"
  },
  {
    "epoch": 2117,
    "batch_idx": 96,
    "loss": 0.37535619735717773,
    "learning_rate": 8.420322756079247e-06,
    "timestep_mean": 437.46875,
    "epsilon_norm": 6.312993049621582,
    "prediction_norm": 5.072848796844482,
    "timestamp": "2025-07-14T12:34:50.044830"
  },
  {
    "epoch": 2117,
    "batch_idx": 128,
    "loss": 0.20540671050548553,
    "learning_rate": 8.420322756079247e-06,
    "timestep_mean": 495.25,
    "epsilon_norm": 7.259200096130371,
    "prediction_norm": 6.357961177825928,
    "timestamp": "2025-07-14T12:34:50.050049"
  },
  {
    "epoch": 2117,
    "batch_idx": 160,
    "loss": 0.22473357617855072,
    "learning_rate": 8.420322756079247e-06,
    "timestep_mean": 545.6875,
    "epsilon_norm": 8.70098876953125,
    "prediction_norm": 7.559520721435547,
    "timestamp": "2025-07-14T12:34:50.055193"
  },
  {
    "epoch": 2117,
    "batch_idx": 192,
    "loss": 0.06866813451051712,
    "learning_rate": 8.420322756079247e-06,
    "timestep_mean": 470.75,
    "epsilon_norm": 3.3621459007263184,
    "prediction_norm": 3.1903090476989746,
    "timestamp": "2025-07-14T12:34:50.060301"
  },
  {
    "epoch": 2118,
    "batch_idx": 0,
    "loss": 0.2683389186859131,
    "learning_rate": 8.564460933810371e-06,
    "timestep_mean": 516.65625,
    "epsilon_norm": 8.717577934265137,
    "prediction_norm": 7.087712287902832,
    "timestamp": "2025-07-14T12:34:50.065493"
  },
  {
    "epoch": 2118,
    "batch_idx": 32,
    "loss": 0.24614068865776062,
    "learning_rate": 8.564460933810371e-06,
    "timestep_mean": 486.96875,
    "epsilon_norm": 7.608020782470703,
    "prediction_norm": 7.138520240783691,
    "timestamp": "2025-07-14T12:34:50.070710"
  },
  {
    "epoch": 2118,
    "batch_idx": 64,
    "loss": 0.2274044156074524,
    "learning_rate": 8.564460933810371e-06,
    "timestep_mean": 507.375,
    "epsilon_norm": 7.926602840423584,
    "prediction_norm": 6.494617462158203,
    "timestamp": "2025-07-14T12:34:50.075823"
  },
  {
    "epoch": 2118,
    "batch_idx": 96,
    "loss": 0.1955072581768036,
    "learning_rate": 8.564460933810371e-06,
    "timestep_mean": 516.9375,
    "epsilon_norm": 7.626676559448242,
    "prediction_norm": 6.988178253173828,
    "timestamp": "2025-07-14T12:34:50.080925"
  },
  {
    "epoch": 2118,
    "batch_idx": 128,
    "loss": 0.35040369629859924,
    "learning_rate": 8.564460933810371e-06,
    "timestep_mean": 434.53125,
    "epsilon_norm": 9.726753234863281,
    "prediction_norm": 8.665580749511719,
    "timestamp": "2025-07-14T12:34:50.086134"
  },
  {
    "epoch": 2118,
    "batch_idx": 160,
    "loss": 0.3763939440250397,
    "learning_rate": 8.564460933810371e-06,
    "timestep_mean": 611.9375,
    "epsilon_norm": 8.403205871582031,
    "prediction_norm": 6.493884086608887,
    "timestamp": "2025-07-14T12:34:50.091365"
  },
  {
    "epoch": 2118,
    "batch_idx": 192,
    "loss": 0.2772829830646515,
    "learning_rate": 8.564460933810371e-06,
    "timestep_mean": 438.375,
    "epsilon_norm": 3.4423484802246094,
    "prediction_norm": 2.290379047393799,
    "timestamp": "2025-07-14T12:34:50.096703"
  },
  {
    "epoch": 2119,
    "batch_idx": 0,
    "loss": 0.39266490936279297,
    "learning_rate": 8.709811679882063e-06,
    "timestep_mean": 513.4375,
    "epsilon_norm": 8.07037353515625,
    "prediction_norm": 6.281092643737793,
    "timestamp": "2025-07-14T12:34:50.101898"
  },
  {
    "epoch": 2119,
    "batch_idx": 32,
    "loss": 0.2370031327009201,
    "learning_rate": 8.709811679882063e-06,
    "timestep_mean": 501.90625,
    "epsilon_norm": 7.882142066955566,
    "prediction_norm": 6.5924072265625,
    "timestamp": "2025-07-14T12:34:50.107058"
  },
  {
    "epoch": 2119,
    "batch_idx": 64,
    "loss": 0.3821852207183838,
    "learning_rate": 8.709811679882063e-06,
    "timestep_mean": 483.15625,
    "epsilon_norm": 7.649198532104492,
    "prediction_norm": 5.557452201843262,
    "timestamp": "2025-07-14T12:34:50.112221"
  },
  {
    "epoch": 2119,
    "batch_idx": 96,
    "loss": 0.26920223236083984,
    "learning_rate": 8.709811679882063e-06,
    "timestep_mean": 512.1875,
    "epsilon_norm": 8.211034774780273,
    "prediction_norm": 7.400902271270752,
    "timestamp": "2025-07-14T12:34:50.117564"
  },
  {
    "epoch": 2119,
    "batch_idx": 128,
    "loss": 0.2970668375492096,
    "learning_rate": 8.709811679882063e-06,
    "timestep_mean": 437.28125,
    "epsilon_norm": 8.149015426635742,
    "prediction_norm": 6.781374931335449,
    "timestamp": "2025-07-14T12:34:50.122737"
  },
  {
    "epoch": 2119,
    "batch_idx": 160,
    "loss": 0.3590554893016815,
    "learning_rate": 8.709811679882063e-06,
    "timestep_mean": 532.0,
    "epsilon_norm": 8.553105354309082,
    "prediction_norm": 6.703334808349609,
    "timestamp": "2025-07-14T12:34:50.127947"
  },
  {
    "epoch": 2119,
    "batch_idx": 192,
    "loss": 0.12387789785861969,
    "learning_rate": 8.709811679882063e-06,
    "timestep_mean": 536.375,
    "epsilon_norm": 4.940398693084717,
    "prediction_norm": 4.5279541015625,
    "timestamp": "2025-07-14T12:34:50.133182"
  },
  {
    "epoch": 2120,
    "batch_idx": 0,
    "loss": 0.15341521799564362,
    "learning_rate": 8.856374635655707e-06,
    "timestep_mean": 552.9375,
    "epsilon_norm": 7.134097576141357,
    "prediction_norm": 6.025279521942139,
    "timestamp": "2025-07-14T12:34:50.138513"
  },
  {
    "epoch": 2120,
    "batch_idx": 32,
    "loss": 0.28300726413726807,
    "learning_rate": 8.856374635655707e-06,
    "timestep_mean": 496.28125,
    "epsilon_norm": 7.895658016204834,
    "prediction_norm": 6.577466011047363,
    "timestamp": "2025-07-14T12:34:50.143741"
  },
  {
    "epoch": 2120,
    "batch_idx": 64,
    "loss": 0.2727735638618469,
    "learning_rate": 8.856374635655707e-06,
    "timestep_mean": 496.96875,
    "epsilon_norm": 7.750472068786621,
    "prediction_norm": 6.366668224334717,
    "timestamp": "2025-07-14T12:34:50.148919"
  },
  {
    "epoch": 2120,
    "batch_idx": 96,
    "loss": 0.1924160122871399,
    "learning_rate": 8.856374635655707e-06,
    "timestep_mean": 583.875,
    "epsilon_norm": 7.204117774963379,
    "prediction_norm": 6.805584907531738,
    "timestamp": "2025-07-14T12:34:50.154123"
  },
  {
    "epoch": 2120,
    "batch_idx": 128,
    "loss": 0.3727147579193115,
    "learning_rate": 8.856374635655707e-06,
    "timestep_mean": 480.15625,
    "epsilon_norm": 8.631113052368164,
    "prediction_norm": 6.836287498474121,
    "timestamp": "2025-07-14T12:34:50.159525"
  },
  {
    "epoch": 2120,
    "batch_idx": 160,
    "loss": 0.4079018831253052,
    "learning_rate": 8.856374635655707e-06,
    "timestep_mean": 434.03125,
    "epsilon_norm": 8.810155868530273,
    "prediction_norm": 6.745543479919434,
    "timestamp": "2025-07-14T12:34:50.164738"
  },
  {
    "epoch": 2120,
    "batch_idx": 192,
    "loss": 0.8316023945808411,
    "learning_rate": 8.856374635655707e-06,
    "timestep_mean": 280.375,
    "epsilon_norm": 4.1766533851623535,
    "prediction_norm": 2.912996292114258,
    "timestamp": "2025-07-14T12:34:50.169915"
  },
  {
    "epoch": 2121,
    "batch_idx": 0,
    "loss": 0.29122698307037354,
    "learning_rate": 9.004149439501686e-06,
    "timestep_mean": 490.9375,
    "epsilon_norm": 8.166501998901367,
    "prediction_norm": 7.151699066162109,
    "timestamp": "2025-07-14T12:34:50.175025"
  },
  {
    "epoch": 2121,
    "batch_idx": 32,
    "loss": 0.41996920108795166,
    "learning_rate": 9.004149439501686e-06,
    "timestep_mean": 482.375,
    "epsilon_norm": 9.474522590637207,
    "prediction_norm": 7.2676777839660645,
    "timestamp": "2025-07-14T12:34:50.180329"
  },
  {
    "epoch": 2121,
    "batch_idx": 64,
    "loss": 0.4672568440437317,
    "learning_rate": 9.004149439501686e-06,
    "timestep_mean": 449.875,
    "epsilon_norm": 7.9044108390808105,
    "prediction_norm": 6.197859287261963,
    "timestamp": "2025-07-14T12:34:50.185544"
  },
  {
    "epoch": 2121,
    "batch_idx": 96,
    "loss": 0.43578627705574036,
    "learning_rate": 9.004149439501686e-06,
    "timestep_mean": 526.4375,
    "epsilon_norm": 10.132819175720215,
    "prediction_norm": 8.686002731323242,
    "timestamp": "2025-07-14T12:34:50.190707"
  },
  {
    "epoch": 2121,
    "batch_idx": 128,
    "loss": 0.4733714461326599,
    "learning_rate": 9.004149439501686e-06,
    "timestep_mean": 538.5,
    "epsilon_norm": 8.928503036499023,
    "prediction_norm": 7.443879127502441,
    "timestamp": "2025-07-14T12:34:50.195888"
  },
  {
    "epoch": 2121,
    "batch_idx": 160,
    "loss": 0.14325843751430511,
    "learning_rate": 9.004149439501686e-06,
    "timestep_mean": 526.78125,
    "epsilon_norm": 8.67637825012207,
    "prediction_norm": 8.138477325439453,
    "timestamp": "2025-07-14T12:34:50.201111"
  },
  {
    "epoch": 2121,
    "batch_idx": 192,
    "loss": 0.1807892620563507,
    "learning_rate": 9.004149439501686e-06,
    "timestep_mean": 330.5,
    "epsilon_norm": 3.3540873527526855,
    "prediction_norm": 2.784849166870117,
    "timestamp": "2025-07-14T12:34:50.206256"
  },
  {
    "epoch": 2122,
    "batch_idx": 0,
    "loss": 0.15376682579517365,
    "learning_rate": 9.153135726800555e-06,
    "timestep_mean": 515.3125,
    "epsilon_norm": 7.549628734588623,
    "prediction_norm": 6.955234527587891,
    "timestamp": "2025-07-14T12:34:50.211118"
  },
  {
    "epoch": 2122,
    "batch_idx": 32,
    "loss": 0.404519259929657,
    "learning_rate": 9.153135726800555e-06,
    "timestep_mean": 475.5625,
    "epsilon_norm": 8.060371398925781,
    "prediction_norm": 5.953460216522217,
    "timestamp": "2025-07-14T12:34:50.215965"
  },
  {
    "epoch": 2122,
    "batch_idx": 64,
    "loss": 0.18436819314956665,
    "learning_rate": 9.153135726800555e-06,
    "timestep_mean": 565.25,
    "epsilon_norm": 10.032946586608887,
    "prediction_norm": 9.14859390258789,
    "timestamp": "2025-07-14T12:34:50.220828"
  },
  {
    "epoch": 2122,
    "batch_idx": 96,
    "loss": 0.19597363471984863,
    "learning_rate": 9.153135726800555e-06,
    "timestep_mean": 499.5,
    "epsilon_norm": 7.870100975036621,
    "prediction_norm": 6.979944229125977,
    "timestamp": "2025-07-14T12:34:50.225793"
  },
  {
    "epoch": 2122,
    "batch_idx": 128,
    "loss": 0.17660430073738098,
    "learning_rate": 9.153135726800555e-06,
    "timestep_mean": 553.34375,
    "epsilon_norm": 7.708338737487793,
    "prediction_norm": 6.971002578735352,
    "timestamp": "2025-07-14T12:34:50.230589"
  },
  {
    "epoch": 2122,
    "batch_idx": 160,
    "loss": 0.31036126613616943,
    "learning_rate": 9.153135726800555e-06,
    "timestep_mean": 474.375,
    "epsilon_norm": 7.050073623657227,
    "prediction_norm": 5.617257595062256,
    "timestamp": "2025-07-14T12:34:50.235416"
  },
  {
    "epoch": 2122,
    "batch_idx": 192,
    "loss": 0.9134119749069214,
    "learning_rate": 9.153135726800555e-06,
    "timestep_mean": 405.0,
    "epsilon_norm": 5.489104270935059,
    "prediction_norm": 4.035102844238281,
    "timestamp": "2025-07-14T12:34:50.240201"
  },
  {
    "epoch": 2123,
    "batch_idx": 0,
    "loss": 0.35079723596572876,
    "learning_rate": 9.30333312994337e-06,
    "timestep_mean": 531.09375,
    "epsilon_norm": 8.973335266113281,
    "prediction_norm": 7.452481269836426,
    "timestamp": "2025-07-14T12:34:50.245069"
  },
  {
    "epoch": 2123,
    "batch_idx": 32,
    "loss": 0.2648282051086426,
    "learning_rate": 9.30333312994337e-06,
    "timestep_mean": 481.09375,
    "epsilon_norm": 8.440678596496582,
    "prediction_norm": 7.525950908660889,
    "timestamp": "2025-07-14T12:34:50.249892"
  },
  {
    "epoch": 2123,
    "batch_idx": 64,
    "loss": 0.14945265650749207,
    "learning_rate": 9.30333312994337e-06,
    "timestep_mean": 587.09375,
    "epsilon_norm": 6.432621002197266,
    "prediction_norm": 6.192357063293457,
    "timestamp": "2025-07-14T12:34:50.254829"
  },
  {
    "epoch": 2123,
    "batch_idx": 96,
    "loss": 0.3372567892074585,
    "learning_rate": 9.30333312994337e-06,
    "timestep_mean": 478.34375,
    "epsilon_norm": 9.071242332458496,
    "prediction_norm": 7.802985191345215,
    "timestamp": "2025-07-14T12:34:50.259650"
  },
  {
    "epoch": 2123,
    "batch_idx": 128,
    "loss": 0.19437313079833984,
    "learning_rate": 9.30333312994337e-06,
    "timestep_mean": 529.28125,
    "epsilon_norm": 7.972926139831543,
    "prediction_norm": 6.854885101318359,
    "timestamp": "2025-07-14T12:34:50.264559"
  },
  {
    "epoch": 2123,
    "batch_idx": 160,
    "loss": 0.4129912853240967,
    "learning_rate": 9.30333312994337e-06,
    "timestep_mean": 476.375,
    "epsilon_norm": 8.51399040222168,
    "prediction_norm": 6.610404014587402,
    "timestamp": "2025-07-14T12:34:50.269462"
  },
  {
    "epoch": 2123,
    "batch_idx": 192,
    "loss": 0.5660603046417236,
    "learning_rate": 9.30333312994337e-06,
    "timestep_mean": 479.25,
    "epsilon_norm": 3.9759111404418945,
    "prediction_norm": 2.938480854034424,
    "timestamp": "2025-07-14T12:34:50.274321"
  },
  {
    "epoch": 2124,
    "batch_idx": 0,
    "loss": 0.32824641466140747,
    "learning_rate": 9.454741278332968e-06,
    "timestep_mean": 463.84375,
    "epsilon_norm": 8.355541229248047,
    "prediction_norm": 6.690584659576416,
    "timestamp": "2025-07-14T12:34:50.279233"
  },
  {
    "epoch": 2124,
    "batch_idx": 32,
    "loss": 0.2657858431339264,
    "learning_rate": 9.454741278332968e-06,
    "timestep_mean": 564.59375,
    "epsilon_norm": 7.376384258270264,
    "prediction_norm": 6.029972076416016,
    "timestamp": "2025-07-14T12:34:50.284109"
  },
  {
    "epoch": 2124,
    "batch_idx": 64,
    "loss": 0.20275211334228516,
    "learning_rate": 9.454741278332968e-06,
    "timestep_mean": 510.125,
    "epsilon_norm": 7.10624885559082,
    "prediction_norm": 6.644125461578369,
    "timestamp": "2025-07-14T12:34:50.289525"
  },
  {
    "epoch": 2124,
    "batch_idx": 96,
    "loss": 0.36578792333602905,
    "learning_rate": 9.454741278332968e-06,
    "timestep_mean": 473.3125,
    "epsilon_norm": 8.198054313659668,
    "prediction_norm": 6.377968788146973,
    "timestamp": "2025-07-14T12:34:50.294938"
  },
  {
    "epoch": 2124,
    "batch_idx": 128,
    "loss": 0.20704635977745056,
    "learning_rate": 9.454741278332968e-06,
    "timestep_mean": 578.375,
    "epsilon_norm": 8.045835494995117,
    "prediction_norm": 6.857718467712402,
    "timestamp": "2025-07-14T12:34:50.300431"
  },
  {
    "epoch": 2124,
    "batch_idx": 160,
    "loss": 0.21148771047592163,
    "learning_rate": 9.454741278332968e-06,
    "timestep_mean": 523.21875,
    "epsilon_norm": 7.891017913818359,
    "prediction_norm": 6.5719404220581055,
    "timestamp": "2025-07-14T12:34:50.305859"
  },
  {
    "epoch": 2124,
    "batch_idx": 192,
    "loss": 0.602849006652832,
    "learning_rate": 9.454741278332968e-06,
    "timestep_mean": 471.375,
    "epsilon_norm": 4.459496974945068,
    "prediction_norm": 3.103893518447876,
    "timestamp": "2025-07-14T12:34:50.311213"
  },
  {
    "epoch": 2125,
    "batch_idx": 0,
    "loss": 0.4275817275047302,
    "learning_rate": 9.607359798384742e-06,
    "timestep_mean": 571.34375,
    "epsilon_norm": 9.573867797851562,
    "prediction_norm": 8.292726516723633,
    "timestamp": "2025-07-14T12:34:50.316182"
  },
  {
    "epoch": 2125,
    "batch_idx": 32,
    "loss": 0.2776530086994171,
    "learning_rate": 9.607359798384742e-06,
    "timestep_mean": 501.59375,
    "epsilon_norm": 7.784280776977539,
    "prediction_norm": 6.5653605461120605,
    "timestamp": "2025-07-14T12:34:50.321050"
  },
  {
    "epoch": 2125,
    "batch_idx": 64,
    "loss": 0.3450450301170349,
    "learning_rate": 9.607359798384742e-06,
    "timestep_mean": 506.9375,
    "epsilon_norm": 9.333932876586914,
    "prediction_norm": 7.4192070960998535,
    "timestamp": "2025-07-14T12:34:50.325948"
  },
  {
    "epoch": 2125,
    "batch_idx": 96,
    "loss": 0.2655147910118103,
    "learning_rate": 9.607359798384742e-06,
    "timestep_mean": 454.21875,
    "epsilon_norm": 7.357254505157471,
    "prediction_norm": 6.450386047363281,
    "timestamp": "2025-07-14T12:34:50.330902"
  },
  {
    "epoch": 2125,
    "batch_idx": 128,
    "loss": 0.2715366780757904,
    "learning_rate": 9.607359798384742e-06,
    "timestep_mean": 457.28125,
    "epsilon_norm": 7.9475884437561035,
    "prediction_norm": 6.269521713256836,
    "timestamp": "2025-07-14T12:34:50.335773"
  },
  {
    "epoch": 2125,
    "batch_idx": 160,
    "loss": 0.19835129380226135,
    "learning_rate": 9.607359798384742e-06,
    "timestep_mean": 553.5,
    "epsilon_norm": 6.764613628387451,
    "prediction_norm": 5.890599250793457,
    "timestamp": "2025-07-14T12:34:50.340820"
  },
  {
    "epoch": 2125,
    "batch_idx": 192,
    "loss": 0.0878055989742279,
    "learning_rate": 9.607359798384742e-06,
    "timestep_mean": 636.375,
    "epsilon_norm": 4.428382873535156,
    "prediction_norm": 3.758183717727661,
    "timestamp": "2025-07-14T12:34:50.346040"
  },
  {
    "epoch": 2126,
    "batch_idx": 0,
    "loss": 0.12537524104118347,
    "learning_rate": 9.761188313527803e-06,
    "timestep_mean": 485.0,
    "epsilon_norm": 6.858992576599121,
    "prediction_norm": 6.531896114349365,
    "timestamp": "2025-07-14T12:34:50.350940"
  },
  {
    "epoch": 2126,
    "batch_idx": 32,
    "loss": 0.2799603343009949,
    "learning_rate": 9.761188313527803e-06,
    "timestep_mean": 471.65625,
    "epsilon_norm": 9.936033248901367,
    "prediction_norm": 8.204341888427734,
    "timestamp": "2025-07-14T12:34:50.355769"
  },
  {
    "epoch": 2126,
    "batch_idx": 64,
    "loss": 0.22908706963062286,
    "learning_rate": 9.761188313527803e-06,
    "timestep_mean": 541.9375,
    "epsilon_norm": 7.634427547454834,
    "prediction_norm": 6.855406284332275,
    "timestamp": "2025-07-14T12:34:50.360652"
  },
  {
    "epoch": 2126,
    "batch_idx": 96,
    "loss": 0.12799639999866486,
    "learning_rate": 9.761188313527803e-06,
    "timestep_mean": 518.15625,
    "epsilon_norm": 7.741695880889893,
    "prediction_norm": 6.817145824432373,
    "timestamp": "2025-07-14T12:34:50.365456"
  },
  {
    "epoch": 2126,
    "batch_idx": 128,
    "loss": 0.21483658254146576,
    "learning_rate": 9.761188313527803e-06,
    "timestep_mean": 541.96875,
    "epsilon_norm": 6.924168586730957,
    "prediction_norm": 6.302657127380371,
    "timestamp": "2025-07-14T12:34:50.370337"
  },
  {
    "epoch": 2126,
    "batch_idx": 160,
    "loss": 0.16056817770004272,
    "learning_rate": 9.761188313527803e-06,
    "timestep_mean": 570.75,
    "epsilon_norm": 9.122461318969727,
    "prediction_norm": 8.143000602722168,
    "timestamp": "2025-07-14T12:34:50.375342"
  },
  {
    "epoch": 2126,
    "batch_idx": 192,
    "loss": 0.35516807436943054,
    "learning_rate": 9.761188313527803e-06,
    "timestep_mean": 467.375,
    "epsilon_norm": 3.7671220302581787,
    "prediction_norm": 2.5869712829589844,
    "timestamp": "2025-07-14T12:34:50.380191"
  },
  {
    "epoch": 2127,
    "batch_idx": 0,
    "loss": 0.3219453990459442,
    "learning_rate": 9.916226444205434e-06,
    "timestep_mean": 528.3125,
    "epsilon_norm": 8.99734878540039,
    "prediction_norm": 7.154966354370117,
    "timestamp": "2025-07-14T12:34:50.385025"
  },
  {
    "epoch": 2127,
    "batch_idx": 32,
    "loss": 0.14495503902435303,
    "learning_rate": 9.916226444205434e-06,
    "timestep_mean": 541.9375,
    "epsilon_norm": 8.343046188354492,
    "prediction_norm": 7.458523750305176,
    "timestamp": "2025-07-14T12:34:50.389894"
  },
  {
    "epoch": 2127,
    "batch_idx": 64,
    "loss": 0.3249036371707916,
    "learning_rate": 9.916226444205434e-06,
    "timestep_mean": 408.3125,
    "epsilon_norm": 7.500273704528809,
    "prediction_norm": 5.49415397644043,
    "timestamp": "2025-07-14T12:34:50.394819"
  },
  {
    "epoch": 2127,
    "batch_idx": 96,
    "loss": 0.24123263359069824,
    "learning_rate": 9.916226444205434e-06,
    "timestep_mean": 586.8125,
    "epsilon_norm": 6.927382946014404,
    "prediction_norm": 5.806008815765381,
    "timestamp": "2025-07-14T12:34:50.399739"
  },
  {
    "epoch": 2127,
    "batch_idx": 128,
    "loss": 0.1310543417930603,
    "learning_rate": 9.916226444205434e-06,
    "timestep_mean": 531.90625,
    "epsilon_norm": 8.998549461364746,
    "prediction_norm": 8.575228691101074,
    "timestamp": "2025-07-14T12:34:50.404629"
  },
  {
    "epoch": 2127,
    "batch_idx": 160,
    "loss": 0.2016347348690033,
    "learning_rate": 9.916226444205434e-06,
    "timestep_mean": 482.375,
    "epsilon_norm": 6.291799545288086,
    "prediction_norm": 5.315706729888916,
    "timestamp": "2025-07-14T12:34:50.409448"
  },
  {
    "epoch": 2127,
    "batch_idx": 192,
    "loss": 0.23077797889709473,
    "learning_rate": 9.916226444205434e-06,
    "timestep_mean": 331.375,
    "epsilon_norm": 2.929678440093994,
    "prediction_norm": 2.9887704849243164,
    "timestamp": "2025-07-14T12:34:50.414448"
  },
  {
    "epoch": 2128,
    "batch_idx": 0,
    "loss": 0.25078830122947693,
    "learning_rate": 1.0072473807876581e-05,
    "timestep_mean": 540.75,
    "epsilon_norm": 8.873737335205078,
    "prediction_norm": 7.554908752441406,
    "timestamp": "2025-07-14T12:34:50.420063"
  },
  {
    "epoch": 2128,
    "batch_idx": 32,
    "loss": 0.2210025191307068,
    "learning_rate": 1.0072473807876581e-05,
    "timestep_mean": 528.34375,
    "epsilon_norm": 7.318855285644531,
    "prediction_norm": 6.701930999755859,
    "timestamp": "2025-07-14T12:34:50.425618"
  },
  {
    "epoch": 2128,
    "batch_idx": 64,
    "loss": 0.36026662588119507,
    "learning_rate": 1.0072473807876581e-05,
    "timestep_mean": 463.84375,
    "epsilon_norm": 7.998106956481934,
    "prediction_norm": 6.465228080749512,
    "timestamp": "2025-07-14T12:34:50.431041"
  },
  {
    "epoch": 2128,
    "batch_idx": 96,
    "loss": 0.3302868604660034,
    "learning_rate": 1.0072473807876581e-05,
    "timestep_mean": 491.6875,
    "epsilon_norm": 7.33000373840332,
    "prediction_norm": 6.092495441436768,
    "timestamp": "2025-07-14T12:34:50.436437"
  },
  {
    "epoch": 2128,
    "batch_idx": 128,
    "loss": 0.32250893115997314,
    "learning_rate": 1.0072473807876581e-05,
    "timestep_mean": 464.28125,
    "epsilon_norm": 8.588395118713379,
    "prediction_norm": 7.547336101531982,
    "timestamp": "2025-07-14T12:34:50.441418"
  },
  {
    "epoch": 2128,
    "batch_idx": 160,
    "loss": 0.2914985716342926,
    "learning_rate": 1.0072473807876581e-05,
    "timestep_mean": 519.125,
    "epsilon_norm": 8.945962905883789,
    "prediction_norm": 7.915607452392578,
    "timestamp": "2025-07-14T12:34:50.446473"
  },
  {
    "epoch": 2128,
    "batch_idx": 192,
    "loss": 0.18594121932983398,
    "learning_rate": 1.0072473807876581e-05,
    "timestep_mean": 608.5,
    "epsilon_norm": 3.6899666786193848,
    "prediction_norm": 3.0461795330047607,
    "timestamp": "2025-07-14T12:34:50.451458"
  },
  {
    "epoch": 2129,
    "batch_idx": 0,
    "loss": 0.1982463002204895,
    "learning_rate": 1.0229930019016353e-05,
    "timestep_mean": 476.5625,
    "epsilon_norm": 8.064504623413086,
    "prediction_norm": 6.878756046295166,
    "timestamp": "2025-07-14T12:34:50.456436"
  },
  {
    "epoch": 2129,
    "batch_idx": 32,
    "loss": 0.27828556299209595,
    "learning_rate": 1.0229930019016353e-05,
    "timestep_mean": 443.875,
    "epsilon_norm": 7.014949321746826,
    "prediction_norm": 5.936335563659668,
    "timestamp": "2025-07-14T12:34:50.461539"
  },
  {
    "epoch": 2129,
    "batch_idx": 64,
    "loss": 0.35975658893585205,
    "learning_rate": 1.0229930019016353e-05,
    "timestep_mean": 543.625,
    "epsilon_norm": 8.123896598815918,
    "prediction_norm": 6.443610191345215,
    "timestamp": "2025-07-14T12:34:50.466386"
  },
  {
    "epoch": 2129,
    "batch_idx": 96,
    "loss": 0.36641716957092285,
    "learning_rate": 1.0229930019016353e-05,
    "timestep_mean": 465.3125,
    "epsilon_norm": 8.042379379272461,
    "prediction_norm": 7.04482889175415,
    "timestamp": "2025-07-14T12:34:50.471285"
  },
  {
    "epoch": 2129,
    "batch_idx": 128,
    "loss": 0.26569902896881104,
    "learning_rate": 1.0229930019016353e-05,
    "timestep_mean": 502.9375,
    "epsilon_norm": 7.945163726806641,
    "prediction_norm": 6.687344074249268,
    "timestamp": "2025-07-14T12:34:50.476095"
  },
  {
    "epoch": 2129,
    "batch_idx": 160,
    "loss": 0.34459981322288513,
    "learning_rate": 1.0229930019016353e-05,
    "timestep_mean": 462.84375,
    "epsilon_norm": 8.169031143188477,
    "prediction_norm": 7.029288291931152,
    "timestamp": "2025-07-14T12:34:50.481184"
  },
  {
    "epoch": 2129,
    "batch_idx": 192,
    "loss": 0.2858179211616516,
    "learning_rate": 1.0229930019016353e-05,
    "timestep_mean": 383.625,
    "epsilon_norm": 3.397822380065918,
    "prediction_norm": 2.4862635135650635,
    "timestamp": "2025-07-14T12:34:50.486493"
  },
  {
    "epoch": 2130,
    "batch_idx": 0,
    "loss": 0.22073926031589508,
    "learning_rate": 1.038859468911708e-05,
    "timestep_mean": 469.90625,
    "epsilon_norm": 8.19937801361084,
    "prediction_norm": 7.243435859680176,
    "timestamp": "2025-07-14T12:34:50.492032"
  },
  {
    "epoch": 2130,
    "batch_idx": 32,
    "loss": 0.2018914371728897,
    "learning_rate": 1.038859468911708e-05,
    "timestep_mean": 442.15625,
    "epsilon_norm": 6.797720909118652,
    "prediction_norm": 5.629621982574463,
    "timestamp": "2025-07-14T12:34:50.497717"
  },
  {
    "epoch": 2130,
    "batch_idx": 64,
    "loss": 0.20215846598148346,
    "learning_rate": 1.038859468911708e-05,
    "timestep_mean": 486.5,
    "epsilon_norm": 7.712428569793701,
    "prediction_norm": 6.499995708465576,
    "timestamp": "2025-07-14T12:34:50.503003"
  },
  {
    "epoch": 2130,
    "batch_idx": 96,
    "loss": 0.4300670027732849,
    "learning_rate": 1.038859468911708e-05,
    "timestep_mean": 484.875,
    "epsilon_norm": 8.772724151611328,
    "prediction_norm": 7.630157470703125,
    "timestamp": "2025-07-14T12:34:50.508253"
  },
  {
    "epoch": 2130,
    "batch_idx": 128,
    "loss": 0.24800419807434082,
    "learning_rate": 1.038859468911708e-05,
    "timestep_mean": 538.15625,
    "epsilon_norm": 7.4062581062316895,
    "prediction_norm": 6.005038738250732,
    "timestamp": "2025-07-14T12:34:50.513684"
  },
  {
    "epoch": 2130,
    "batch_idx": 160,
    "loss": 0.24234867095947266,
    "learning_rate": 1.038859468911708e-05,
    "timestep_mean": 491.3125,
    "epsilon_norm": 6.967911243438721,
    "prediction_norm": 6.2325005531311035,
    "timestamp": "2025-07-14T12:34:50.519234"
  },
  {
    "epoch": 2130,
    "batch_idx": 192,
    "loss": 0.05992549657821655,
    "learning_rate": 1.038859468911708e-05,
    "timestep_mean": 566.75,
    "epsilon_norm": 4.681318283081055,
    "prediction_norm": 4.753829002380371,
    "timestamp": "2025-07-14T12:34:50.524550"
  },
  {
    "epoch": 2131,
    "batch_idx": 0,
    "loss": 0.2191329002380371,
    "learning_rate": 1.0548467426689702e-05,
    "timestep_mean": 485.375,
    "epsilon_norm": 5.948108196258545,
    "prediction_norm": 4.752809524536133,
    "timestamp": "2025-07-14T12:34:50.529915"
  },
  {
    "epoch": 2131,
    "batch_idx": 32,
    "loss": 0.1906876266002655,
    "learning_rate": 1.0548467426689702e-05,
    "timestep_mean": 440.46875,
    "epsilon_norm": 6.625113487243652,
    "prediction_norm": 6.287996768951416,
    "timestamp": "2025-07-14T12:34:50.534905"
  },
  {
    "epoch": 2131,
    "batch_idx": 64,
    "loss": 0.2930314242839813,
    "learning_rate": 1.0548467426689702e-05,
    "timestep_mean": 502.625,
    "epsilon_norm": 8.47133731842041,
    "prediction_norm": 7.029069423675537,
    "timestamp": "2025-07-14T12:34:50.539910"
  },
  {
    "epoch": 2131,
    "batch_idx": 96,
    "loss": 0.21723823249340057,
    "learning_rate": 1.0548467426689702e-05,
    "timestep_mean": 585.46875,
    "epsilon_norm": 7.943328380584717,
    "prediction_norm": 6.861044883728027,
    "timestamp": "2025-07-14T12:34:50.545326"
  },
  {
    "epoch": 2131,
    "batch_idx": 128,
    "loss": 0.329262375831604,
    "learning_rate": 1.0548467426689702e-05,
    "timestep_mean": 496.625,
    "epsilon_norm": 9.02530288696289,
    "prediction_norm": 7.757320404052734,
    "timestamp": "2025-07-14T12:34:50.550722"
  },
  {
    "epoch": 2131,
    "batch_idx": 160,
    "loss": 0.3291139602661133,
    "learning_rate": 1.0548467426689702e-05,
    "timestep_mean": 433.59375,
    "epsilon_norm": 9.08447551727295,
    "prediction_norm": 7.397952079772949,
    "timestamp": "2025-07-14T12:34:50.556089"
  },
  {
    "epoch": 2131,
    "batch_idx": 192,
    "loss": 0.19047218561172485,
    "learning_rate": 1.0548467426689702e-05,
    "timestep_mean": 581.875,
    "epsilon_norm": 4.488326549530029,
    "prediction_norm": 3.688300848007202,
    "timestamp": "2025-07-14T12:34:50.561249"
  },
  {
    "epoch": 2132,
    "batch_idx": 0,
    "loss": 0.21936047077178955,
    "learning_rate": 1.0709547837263923e-05,
    "timestep_mean": 470.9375,
    "epsilon_norm": 7.678422451019287,
    "prediction_norm": 7.143368721008301,
    "timestamp": "2025-07-14T12:34:50.566635"
  },
  {
    "epoch": 2132,
    "batch_idx": 32,
    "loss": 0.18828114867210388,
    "learning_rate": 1.0709547837263923e-05,
    "timestep_mean": 505.03125,
    "epsilon_norm": 6.6643877029418945,
    "prediction_norm": 6.656837463378906,
    "timestamp": "2025-07-14T12:34:50.571882"
  },
  {
    "epoch": 2132,
    "batch_idx": 64,
    "loss": 0.2681642770767212,
    "learning_rate": 1.0709547837263923e-05,
    "timestep_mean": 511.59375,
    "epsilon_norm": 8.485969543457031,
    "prediction_norm": 7.069472312927246,
    "timestamp": "2025-07-14T12:34:50.577093"
  },
  {
    "epoch": 2132,
    "batch_idx": 96,
    "loss": 0.4003208875656128,
    "learning_rate": 1.0709547837263923e-05,
    "timestep_mean": 492.21875,
    "epsilon_norm": 8.701203346252441,
    "prediction_norm": 6.409393310546875,
    "timestamp": "2025-07-14T12:34:50.582274"
  },
  {
    "epoch": 2132,
    "batch_idx": 128,
    "loss": 0.2107619047164917,
    "learning_rate": 1.0709547837263923e-05,
    "timestep_mean": 525.90625,
    "epsilon_norm": 9.821513175964355,
    "prediction_norm": 9.361956596374512,
    "timestamp": "2025-07-14T12:34:50.587454"
  },
  {
    "epoch": 2132,
    "batch_idx": 160,
    "loss": 0.2937752902507782,
    "learning_rate": 1.0709547837263923e-05,
    "timestep_mean": 575.84375,
    "epsilon_norm": 8.11043930053711,
    "prediction_norm": 7.240612506866455,
    "timestamp": "2025-07-14T12:34:50.592805"
  },
  {
    "epoch": 2132,
    "batch_idx": 192,
    "loss": 0.11761291325092316,
    "learning_rate": 1.0709547837263923e-05,
    "timestep_mean": 537.125,
    "epsilon_norm": 3.518327474594116,
    "prediction_norm": 3.0159261226654053,
    "timestamp": "2025-07-14T12:34:50.597953"
  },
  {
    "epoch": 2133,
    "batch_idx": 0,
    "loss": 0.3560531735420227,
    "learning_rate": 1.0871835523390009e-05,
    "timestep_mean": 524.5625,
    "epsilon_norm": 9.022808074951172,
    "prediction_norm": 6.887698650360107,
    "timestamp": "2025-07-14T12:34:50.603157"
  },
  {
    "epoch": 2133,
    "batch_idx": 32,
    "loss": 0.3009014129638672,
    "learning_rate": 1.0871835523390009e-05,
    "timestep_mean": 438.25,
    "epsilon_norm": 7.5341339111328125,
    "prediction_norm": 5.956254482269287,
    "timestamp": "2025-07-14T12:34:50.608362"
  },
  {
    "epoch": 2133,
    "batch_idx": 64,
    "loss": 0.521744430065155,
    "learning_rate": 1.0871835523390009e-05,
    "timestep_mean": 486.0,
    "epsilon_norm": 8.382073402404785,
    "prediction_norm": 6.0499982833862305,
    "timestamp": "2025-07-14T12:34:50.613582"
  },
  {
    "epoch": 2133,
    "batch_idx": 96,
    "loss": 0.20463711023330688,
    "learning_rate": 1.0871835523390009e-05,
    "timestep_mean": 584.90625,
    "epsilon_norm": 8.434441566467285,
    "prediction_norm": 7.60009241104126,
    "timestamp": "2025-07-14T12:34:50.618781"
  },
  {
    "epoch": 2133,
    "batch_idx": 128,
    "loss": 0.0928824245929718,
    "learning_rate": 1.0871835523390009e-05,
    "timestep_mean": 622.6875,
    "epsilon_norm": 6.952813148498535,
    "prediction_norm": 6.56473970413208,
    "timestamp": "2025-07-14T12:34:50.624057"
  },
  {
    "epoch": 2133,
    "batch_idx": 160,
    "loss": 0.43543314933776855,
    "learning_rate": 1.0871835523390009e-05,
    "timestep_mean": 527.0,
    "epsilon_norm": 8.791211128234863,
    "prediction_norm": 7.262517929077148,
    "timestamp": "2025-07-14T12:34:50.629260"
  },
  {
    "epoch": 2133,
    "batch_idx": 192,
    "loss": 0.1155029684305191,
    "learning_rate": 1.0871835523390009e-05,
    "timestep_mean": 407.625,
    "epsilon_norm": 3.994727611541748,
    "prediction_norm": 3.3201675415039062,
    "timestamp": "2025-07-14T12:34:50.634605"
  },
  {
    "epoch": 2134,
    "batch_idx": 0,
    "loss": 0.46228256821632385,
    "learning_rate": 1.1035330084639098e-05,
    "timestep_mean": 449.0,
    "epsilon_norm": 9.56423568725586,
    "prediction_norm": 7.552985668182373,
    "timestamp": "2025-07-14T12:34:50.639867"
  },
  {
    "epoch": 2134,
    "batch_idx": 32,
    "loss": 0.40020495653152466,
    "learning_rate": 1.1035330084639098e-05,
    "timestep_mean": 447.125,
    "epsilon_norm": 8.2273588180542,
    "prediction_norm": 6.60250997543335,
    "timestamp": "2025-07-14T12:34:50.645026"
  },
  {
    "epoch": 2134,
    "batch_idx": 64,
    "loss": 0.2751712203025818,
    "learning_rate": 1.1035330084639098e-05,
    "timestep_mean": 489.34375,
    "epsilon_norm": 8.072710990905762,
    "prediction_norm": 6.475829601287842,
    "timestamp": "2025-07-14T12:34:50.650384"
  },
  {
    "epoch": 2134,
    "batch_idx": 96,
    "loss": 0.2050061970949173,
    "learning_rate": 1.1035330084639098e-05,
    "timestep_mean": 465.375,
    "epsilon_norm": 7.846796035766602,
    "prediction_norm": 6.595571994781494,
    "timestamp": "2025-07-14T12:34:50.655574"
  },
  {
    "epoch": 2134,
    "batch_idx": 128,
    "loss": 0.3434141278266907,
    "learning_rate": 1.1035330084639098e-05,
    "timestep_mean": 477.8125,
    "epsilon_norm": 9.261918067932129,
    "prediction_norm": 7.748049736022949,
    "timestamp": "2025-07-14T12:34:50.660795"
  },
  {
    "epoch": 2134,
    "batch_idx": 160,
    "loss": 0.38498952984809875,
    "learning_rate": 1.1035330084639098e-05,
    "timestep_mean": 488.25,
    "epsilon_norm": 9.248785972595215,
    "prediction_norm": 6.982617378234863,
    "timestamp": "2025-07-14T12:34:50.665975"
  },
  {
    "epoch": 2134,
    "batch_idx": 192,
    "loss": 0.13866646587848663,
    "learning_rate": 1.1035330084639098e-05,
    "timestep_mean": 524.75,
    "epsilon_norm": 3.659647226333618,
    "prediction_norm": 3.230952262878418,
    "timestamp": "2025-07-14T12:34:50.671147"
  },
  {
    "epoch": 2135,
    "batch_idx": 0,
    "loss": 0.18903766572475433,
    "learning_rate": 1.1200031117604607e-05,
    "timestep_mean": 497.75,
    "epsilon_norm": 6.355045795440674,
    "prediction_norm": 5.497159004211426,
    "timestamp": "2025-07-14T12:34:50.676409"
  },
  {
    "epoch": 2135,
    "batch_idx": 32,
    "loss": 0.4896126985549927,
    "learning_rate": 1.1200031117604607e-05,
    "timestep_mean": 453.5,
    "epsilon_norm": 8.602578163146973,
    "prediction_norm": 6.46823263168335,
    "timestamp": "2025-07-14T12:34:50.681620"
  },
  {
    "epoch": 2135,
    "batch_idx": 64,
    "loss": 0.39006608724594116,
    "learning_rate": 1.1200031117604607e-05,
    "timestep_mean": 446.8125,
    "epsilon_norm": 7.492772579193115,
    "prediction_norm": 6.1834797859191895,
    "timestamp": "2025-07-14T12:34:50.686877"
  },
  {
    "epoch": 2135,
    "batch_idx": 96,
    "loss": 0.20859861373901367,
    "learning_rate": 1.1200031117604607e-05,
    "timestep_mean": 516.75,
    "epsilon_norm": 8.03946590423584,
    "prediction_norm": 7.3640336990356445,
    "timestamp": "2025-07-14T12:34:50.692154"
  },
  {
    "epoch": 2135,
    "batch_idx": 128,
    "loss": 0.22271820902824402,
    "learning_rate": 1.1200031117604607e-05,
    "timestep_mean": 470.90625,
    "epsilon_norm": 8.293465614318848,
    "prediction_norm": 7.5074849128723145,
    "timestamp": "2025-07-14T12:34:50.697310"
  },
  {
    "epoch": 2135,
    "batch_idx": 160,
    "loss": 0.18984846770763397,
    "learning_rate": 1.1200031117604607e-05,
    "timestep_mean": 566.75,
    "epsilon_norm": 7.8195109367370605,
    "prediction_norm": 6.901498794555664,
    "timestamp": "2025-07-14T12:34:50.702525"
  },
  {
    "epoch": 2135,
    "batch_idx": 192,
    "loss": 0.20104467868804932,
    "learning_rate": 1.1200031117604607e-05,
    "timestep_mean": 515.5,
    "epsilon_norm": 4.07381010055542,
    "prediction_norm": 3.7772040367126465,
    "timestamp": "2025-07-14T12:34:50.707646"
  },
  {
    "epoch": 2136,
    "batch_idx": 0,
    "loss": 0.34560880064964294,
    "learning_rate": 1.1365938215903274e-05,
    "timestep_mean": 490.6875,
    "epsilon_norm": 8.444351196289062,
    "prediction_norm": 7.075429916381836,
    "timestamp": "2025-07-14T12:34:50.713087"
  },
  {
    "epoch": 2136,
    "batch_idx": 32,
    "loss": 0.1021970883011818,
    "learning_rate": 1.1365938215903274e-05,
    "timestep_mean": 543.28125,
    "epsilon_norm": 7.553903102874756,
    "prediction_norm": 6.866037368774414,
    "timestamp": "2025-07-14T12:34:50.718210"
  },
  {
    "epoch": 2136,
    "batch_idx": 64,
    "loss": 0.28501957654953003,
    "learning_rate": 1.1365938215903274e-05,
    "timestep_mean": 487.4375,
    "epsilon_norm": 6.836187839508057,
    "prediction_norm": 5.18303918838501,
    "timestamp": "2025-07-14T12:34:50.723378"
  },
  {
    "epoch": 2136,
    "batch_idx": 96,
    "loss": 0.17118403315544128,
    "learning_rate": 1.1365938215903274e-05,
    "timestep_mean": 454.625,
    "epsilon_norm": 7.1685686111450195,
    "prediction_norm": 6.32431697845459,
    "timestamp": "2025-07-14T12:34:50.728545"
  },
  {
    "epoch": 2136,
    "batch_idx": 128,
    "loss": 0.3401031792163849,
    "learning_rate": 1.1365938215903274e-05,
    "timestep_mean": 517.1875,
    "epsilon_norm": 8.119568824768066,
    "prediction_norm": 6.547407150268555,
    "timestamp": "2025-07-14T12:34:50.733823"
  },
  {
    "epoch": 2136,
    "batch_idx": 160,
    "loss": 0.28391581773757935,
    "learning_rate": 1.1365938215903274e-05,
    "timestep_mean": 422.28125,
    "epsilon_norm": 7.113687515258789,
    "prediction_norm": 5.7819108963012695,
    "timestamp": "2025-07-14T12:34:50.738999"
  },
  {
    "epoch": 2136,
    "batch_idx": 192,
    "loss": 0.28933843970298767,
    "learning_rate": 1.1365938215903274e-05,
    "timestep_mean": 471.0,
    "epsilon_norm": 4.686058044433594,
    "prediction_norm": 3.3128671646118164,
    "timestamp": "2025-07-14T12:34:50.744174"
  },
  {
    "epoch": 2137,
    "batch_idx": 0,
    "loss": 0.37800753116607666,
    "learning_rate": 1.1533050970175666e-05,
    "timestep_mean": 511.21875,
    "epsilon_norm": 8.553091049194336,
    "prediction_norm": 7.074580669403076,
    "timestamp": "2025-07-14T12:34:50.749370"
  },
  {
    "epoch": 2137,
    "batch_idx": 32,
    "loss": 0.2299950271844864,
    "learning_rate": 1.1533050970175666e-05,
    "timestep_mean": 499.3125,
    "epsilon_norm": 7.661386966705322,
    "prediction_norm": 6.943528175354004,
    "timestamp": "2025-07-14T12:34:50.754574"
  },
  {
    "epoch": 2137,
    "batch_idx": 64,
    "loss": 0.27066802978515625,
    "learning_rate": 1.1533050970175666e-05,
    "timestep_mean": 413.21875,
    "epsilon_norm": 7.716713905334473,
    "prediction_norm": 6.772509574890137,
    "timestamp": "2025-07-14T12:34:50.759758"
  },
  {
    "epoch": 2137,
    "batch_idx": 96,
    "loss": 0.16690750420093536,
    "learning_rate": 1.1533050970175666e-05,
    "timestep_mean": 571.0,
    "epsilon_norm": 6.75479793548584,
    "prediction_norm": 6.065530776977539,
    "timestamp": "2025-07-14T12:34:50.764931"
  },
  {
    "epoch": 2137,
    "batch_idx": 128,
    "loss": 0.3092663288116455,
    "learning_rate": 1.1533050970175666e-05,
    "timestep_mean": 479.65625,
    "epsilon_norm": 7.155519485473633,
    "prediction_norm": 5.788748264312744,
    "timestamp": "2025-07-14T12:34:50.770333"
  },
  {
    "epoch": 2137,
    "batch_idx": 160,
    "loss": 0.252305269241333,
    "learning_rate": 1.1533050970175666e-05,
    "timestep_mean": 520.15625,
    "epsilon_norm": 8.753080368041992,
    "prediction_norm": 7.601300239562988,
    "timestamp": "2025-07-14T12:34:50.775457"
  },
  {
    "epoch": 2137,
    "batch_idx": 192,
    "loss": 0.29571568965911865,
    "learning_rate": 1.1533050970175666e-05,
    "timestep_mean": 521.75,
    "epsilon_norm": 4.585338115692139,
    "prediction_norm": 4.361526966094971,
    "timestamp": "2025-07-14T12:34:50.780619"
  },
  {
    "epoch": 2138,
    "batch_idx": 0,
    "loss": 0.40559038519859314,
    "learning_rate": 1.1701368968087671e-05,
    "timestep_mean": 478.34375,
    "epsilon_norm": 9.809584617614746,
    "prediction_norm": 7.546932697296143,
    "timestamp": "2025-07-14T12:34:50.785857"
  },
  {
    "epoch": 2138,
    "batch_idx": 32,
    "loss": 0.1585310846567154,
    "learning_rate": 1.1701368968087671e-05,
    "timestep_mean": 515.1875,
    "epsilon_norm": 6.834320068359375,
    "prediction_norm": 5.9334211349487305,
    "timestamp": "2025-07-14T12:34:50.791045"
  },
  {
    "epoch": 2138,
    "batch_idx": 64,
    "loss": 0.27087682485580444,
    "learning_rate": 1.1701368968087671e-05,
    "timestep_mean": 582.6875,
    "epsilon_norm": 8.859013557434082,
    "prediction_norm": 7.626240253448486,
    "timestamp": "2025-07-14T12:34:50.796384"
  },
  {
    "epoch": 2138,
    "batch_idx": 96,
    "loss": 0.34417837858200073,
    "learning_rate": 1.1701368968087671e-05,
    "timestep_mean": 505.0,
    "epsilon_norm": 8.301148414611816,
    "prediction_norm": 6.942723274230957,
    "timestamp": "2025-07-14T12:34:50.801557"
  },
  {
    "epoch": 2138,
    "batch_idx": 128,
    "loss": 0.20464912056922913,
    "learning_rate": 1.1701368968087671e-05,
    "timestep_mean": 545.5,
    "epsilon_norm": 8.37087345123291,
    "prediction_norm": 7.683331489562988,
    "timestamp": "2025-07-14T12:34:50.806789"
  },
  {
    "epoch": 2138,
    "batch_idx": 160,
    "loss": 0.395097941160202,
    "learning_rate": 1.1701368968087671e-05,
    "timestep_mean": 469.21875,
    "epsilon_norm": 9.054869651794434,
    "prediction_norm": 6.500072956085205,
    "timestamp": "2025-07-14T12:34:50.811861"
  },
  {
    "epoch": 2138,
    "batch_idx": 192,
    "loss": 0.17824235558509827,
    "learning_rate": 1.1701368968087671e-05,
    "timestep_mean": 528.0,
    "epsilon_norm": 3.2394356727600098,
    "prediction_norm": 2.8694660663604736,
    "timestamp": "2025-07-14T12:34:50.817182"
  },
  {
    "epoch": 2139,
    "batch_idx": 0,
    "loss": 0.16267640888690948,
    "learning_rate": 1.1870891794331501e-05,
    "timestep_mean": 484.15625,
    "epsilon_norm": 6.862362861633301,
    "prediction_norm": 6.171002388000488,
    "timestamp": "2025-07-14T12:34:50.822513"
  },
  {
    "epoch": 2139,
    "batch_idx": 32,
    "loss": 0.1831919252872467,
    "learning_rate": 1.1870891794331501e-05,
    "timestep_mean": 522.125,
    "epsilon_norm": 8.229825019836426,
    "prediction_norm": 7.512200355529785,
    "timestamp": "2025-07-14T12:34:50.827743"
  },
  {
    "epoch": 2139,
    "batch_idx": 64,
    "loss": 0.34586021304130554,
    "learning_rate": 1.1870891794331501e-05,
    "timestep_mean": 462.28125,
    "epsilon_norm": 8.08860969543457,
    "prediction_norm": 6.7513957023620605,
    "timestamp": "2025-07-14T12:34:50.832924"
  },
  {
    "epoch": 2139,
    "batch_idx": 96,
    "loss": 0.29709625244140625,
    "learning_rate": 1.1870891794331501e-05,
    "timestep_mean": 519.46875,
    "epsilon_norm": 9.320913314819336,
    "prediction_norm": 7.558339595794678,
    "timestamp": "2025-07-14T12:34:50.838429"
  },
  {
    "epoch": 2139,
    "batch_idx": 128,
    "loss": 0.272855281829834,
    "learning_rate": 1.1870891794331501e-05,
    "timestep_mean": 463.125,
    "epsilon_norm": 7.976613521575928,
    "prediction_norm": 6.174221515655518,
    "timestamp": "2025-07-14T12:34:50.843664"
  },
  {
    "epoch": 2139,
    "batch_idx": 160,
    "loss": 0.3971004784107208,
    "learning_rate": 1.1870891794331501e-05,
    "timestep_mean": 501.59375,
    "epsilon_norm": 8.522483825683594,
    "prediction_norm": 6.42505407333374,
    "timestamp": "2025-07-14T12:34:50.848994"
  },
  {
    "epoch": 2139,
    "batch_idx": 192,
    "loss": 0.18480896949768066,
    "learning_rate": 1.1870891794331501e-05,
    "timestep_mean": 626.625,
    "epsilon_norm": 4.470332622528076,
    "prediction_norm": 3.701702356338501,
    "timestamp": "2025-07-14T12:34:50.854145"
  },
  {
    "epoch": 2140,
    "batch_idx": 0,
    "loss": 0.18304061889648438,
    "learning_rate": 1.2041619030626301e-05,
    "timestep_mean": 465.78125,
    "epsilon_norm": 7.401337146759033,
    "prediction_norm": 6.400012016296387,
    "timestamp": "2025-07-14T12:34:50.859500"
  },
  {
    "epoch": 2140,
    "batch_idx": 32,
    "loss": 0.48502737283706665,
    "learning_rate": 1.2041619030626301e-05,
    "timestep_mean": 421.15625,
    "epsilon_norm": 9.401082038879395,
    "prediction_norm": 7.17733907699585,
    "timestamp": "2025-07-14T12:34:50.864699"
  },
  {
    "epoch": 2140,
    "batch_idx": 64,
    "loss": 0.209544375538826,
    "learning_rate": 1.2041619030626301e-05,
    "timestep_mean": 481.625,
    "epsilon_norm": 6.979682445526123,
    "prediction_norm": 5.9839982986450195,
    "timestamp": "2025-07-14T12:34:50.869896"
  },
  {
    "epoch": 2140,
    "batch_idx": 96,
    "loss": 0.28183674812316895,
    "learning_rate": 1.2041619030626301e-05,
    "timestep_mean": 487.25,
    "epsilon_norm": 7.862729072570801,
    "prediction_norm": 6.833395957946777,
    "timestamp": "2025-07-14T12:34:50.875049"
  },
  {
    "epoch": 2140,
    "batch_idx": 128,
    "loss": 0.41076645255088806,
    "learning_rate": 1.2041619030626301e-05,
    "timestep_mean": 487.25,
    "epsilon_norm": 8.304055213928223,
    "prediction_norm": 6.6769232749938965,
    "timestamp": "2025-07-14T12:34:50.880251"
  },
  {
    "epoch": 2140,
    "batch_idx": 160,
    "loss": 0.31684672832489014,
    "learning_rate": 1.2041619030626301e-05,
    "timestep_mean": 436.875,
    "epsilon_norm": 8.61583423614502,
    "prediction_norm": 7.197770595550537,
    "timestamp": "2025-07-14T12:34:50.885564"
  },
  {
    "epoch": 2140,
    "batch_idx": 192,
    "loss": 0.08675391227006912,
    "learning_rate": 1.2041619030626301e-05,
    "timestep_mean": 356.625,
    "epsilon_norm": 2.6888394355773926,
    "prediction_norm": 2.620723009109497,
    "timestamp": "2025-07-14T12:34:50.890865"
  },
  {
    "epoch": 2141,
    "batch_idx": 0,
    "loss": 0.4150298833847046,
    "learning_rate": 1.2213550255719648e-05,
    "timestep_mean": 529.84375,
    "epsilon_norm": 7.8382649421691895,
    "prediction_norm": 6.4866180419921875,
    "timestamp": "2025-07-14T12:34:50.896151"
  },
  {
    "epoch": 2141,
    "batch_idx": 32,
    "loss": 0.18215037882328033,
    "learning_rate": 1.2213550255719648e-05,
    "timestep_mean": 474.5625,
    "epsilon_norm": 7.038408279418945,
    "prediction_norm": 6.488996982574463,
    "timestamp": "2025-07-14T12:34:50.901337"
  },
  {
    "epoch": 2141,
    "batch_idx": 64,
    "loss": 0.14146070182323456,
    "learning_rate": 1.2213550255719648e-05,
    "timestep_mean": 500.6875,
    "epsilon_norm": 7.516141414642334,
    "prediction_norm": 7.079586029052734,
    "timestamp": "2025-07-14T12:34:50.906293"
  },
  {
    "epoch": 2141,
    "batch_idx": 96,
    "loss": 0.16844795644283295,
    "learning_rate": 1.2213550255719648e-05,
    "timestep_mean": 500.34375,
    "epsilon_norm": 8.176137924194336,
    "prediction_norm": 7.101306438446045,
    "timestamp": "2025-07-14T12:34:50.911334"
  },
  {
    "epoch": 2141,
    "batch_idx": 128,
    "loss": 0.15623393654823303,
    "learning_rate": 1.2213550255719648e-05,
    "timestep_mean": 479.25,
    "epsilon_norm": 8.231707572937012,
    "prediction_norm": 7.760531902313232,
    "timestamp": "2025-07-14T12:34:50.916578"
  },
  {
    "epoch": 2141,
    "batch_idx": 160,
    "loss": 0.3774990439414978,
    "learning_rate": 1.2213550255719648e-05,
    "timestep_mean": 457.5625,
    "epsilon_norm": 8.518333435058594,
    "prediction_norm": 6.7961273193359375,
    "timestamp": "2025-07-14T12:34:50.921776"
  },
  {
    "epoch": 2141,
    "batch_idx": 192,
    "loss": 0.12375891953706741,
    "learning_rate": 1.2213550255719648e-05,
    "timestep_mean": 569.0,
    "epsilon_norm": 3.222365617752075,
    "prediction_norm": 3.3267428874969482,
    "timestamp": "2025-07-14T12:34:50.926701"
  },
  {
    "epoch": 2142,
    "batch_idx": 0,
    "loss": 0.18406376242637634,
    "learning_rate": 1.2386685045388276e-05,
    "timestep_mean": 544.6875,
    "epsilon_norm": 8.730074882507324,
    "prediction_norm": 7.668691635131836,
    "timestamp": "2025-07-14T12:34:50.931518"
  },
  {
    "epoch": 2142,
    "batch_idx": 32,
    "loss": 0.3106866776943207,
    "learning_rate": 1.2386685045388276e-05,
    "timestep_mean": 444.3125,
    "epsilon_norm": 8.14495849609375,
    "prediction_norm": 6.937127113342285,
    "timestamp": "2025-07-14T12:34:50.936487"
  },
  {
    "epoch": 2142,
    "batch_idx": 64,
    "loss": 0.31505846977233887,
    "learning_rate": 1.2386685045388276e-05,
    "timestep_mean": 492.75,
    "epsilon_norm": 7.749640464782715,
    "prediction_norm": 6.548361301422119,
    "timestamp": "2025-07-14T12:34:50.941373"
  },
  {
    "epoch": 2142,
    "batch_idx": 96,
    "loss": 0.049897804856300354,
    "learning_rate": 1.2386685045388276e-05,
    "timestep_mean": 667.96875,
    "epsilon_norm": 7.197867393493652,
    "prediction_norm": 7.190432071685791,
    "timestamp": "2025-07-14T12:34:50.946267"
  },
  {
    "epoch": 2142,
    "batch_idx": 128,
    "loss": 0.4807611107826233,
    "learning_rate": 1.2386685045388276e-05,
    "timestep_mean": 447.03125,
    "epsilon_norm": 8.859783172607422,
    "prediction_norm": 7.119596481323242,
    "timestamp": "2025-07-14T12:34:50.951379"
  },
  {
    "epoch": 2142,
    "batch_idx": 160,
    "loss": 0.32316362857818604,
    "learning_rate": 1.2386685045388276e-05,
    "timestep_mean": 531.9375,
    "epsilon_norm": 8.872220039367676,
    "prediction_norm": 7.433311939239502,
    "timestamp": "2025-07-14T12:34:50.956564"
  },
  {
    "epoch": 2142,
    "batch_idx": 192,
    "loss": 0.17662495374679565,
    "learning_rate": 1.2386685045388276e-05,
    "timestep_mean": 456.25,
    "epsilon_norm": 4.522340774536133,
    "prediction_norm": 4.003837585449219,
    "timestamp": "2025-07-14T12:34:50.961698"
  },
  {
    "epoch": 2143,
    "batch_idx": 0,
    "loss": 0.25853365659713745,
    "learning_rate": 1.2561022972439349e-05,
    "timestep_mean": 465.28125,
    "epsilon_norm": 7.33000373840332,
    "prediction_norm": 6.093363285064697,
    "timestamp": "2025-07-14T12:34:50.967060"
  },
  {
    "epoch": 2143,
    "batch_idx": 32,
    "loss": 0.12611377239227295,
    "learning_rate": 1.2561022972439349e-05,
    "timestep_mean": 598.03125,
    "epsilon_norm": 7.975232124328613,
    "prediction_norm": 7.294130325317383,
    "timestamp": "2025-07-14T12:34:50.972246"
  },
  {
    "epoch": 2143,
    "batch_idx": 64,
    "loss": 0.340210497379303,
    "learning_rate": 1.2561022972439349e-05,
    "timestep_mean": 537.9375,
    "epsilon_norm": 8.859639167785645,
    "prediction_norm": 7.249686241149902,
    "timestamp": "2025-07-14T12:34:50.977474"
  },
  {
    "epoch": 2143,
    "batch_idx": 96,
    "loss": 0.2872927188873291,
    "learning_rate": 1.2561022972439349e-05,
    "timestep_mean": 456.46875,
    "epsilon_norm": 9.236178398132324,
    "prediction_norm": 7.594180583953857,
    "timestamp": "2025-07-14T12:34:50.982709"
  },
  {
    "epoch": 2143,
    "batch_idx": 128,
    "loss": 0.18720290064811707,
    "learning_rate": 1.2561022972439349e-05,
    "timestep_mean": 469.34375,
    "epsilon_norm": 7.65188455581665,
    "prediction_norm": 7.111563205718994,
    "timestamp": "2025-07-14T12:34:50.987960"
  },
  {
    "epoch": 2143,
    "batch_idx": 160,
    "loss": 0.06461896002292633,
    "learning_rate": 1.2561022972439349e-05,
    "timestep_mean": 629.75,
    "epsilon_norm": 7.503754615783691,
    "prediction_norm": 7.3245649337768555,
    "timestamp": "2025-07-14T12:34:50.992802"
  },
  {
    "epoch": 2143,
    "batch_idx": 192,
    "loss": 0.14722731709480286,
    "learning_rate": 1.2561022972439349e-05,
    "timestep_mean": 512.375,
    "epsilon_norm": 3.4554049968719482,
    "prediction_norm": 2.909423351287842,
    "timestamp": "2025-07-14T12:34:50.997630"
  },
  {
    "epoch": 2144,
    "batch_idx": 0,
    "loss": 0.20797349512577057,
    "learning_rate": 1.2736563606711458e-05,
    "timestep_mean": 484.03125,
    "epsilon_norm": 8.010984420776367,
    "prediction_norm": 6.9858574867248535,
    "timestamp": "2025-07-14T12:34:51.002624"
  },
  {
    "epoch": 2144,
    "batch_idx": 32,
    "loss": 0.3093520998954773,
    "learning_rate": 1.2736563606711458e-05,
    "timestep_mean": 425.34375,
    "epsilon_norm": 7.9517669677734375,
    "prediction_norm": 6.245305061340332,
    "timestamp": "2025-07-14T12:34:51.007699"
  },
  {
    "epoch": 2144,
    "batch_idx": 64,
    "loss": 0.21793901920318604,
    "learning_rate": 1.2736563606711458e-05,
    "timestep_mean": 512.59375,
    "epsilon_norm": 7.825323104858398,
    "prediction_norm": 7.303248405456543,
    "timestamp": "2025-07-14T12:34:51.012851"
  },
  {
    "epoch": 2144,
    "batch_idx": 96,
    "loss": 0.290433794260025,
    "learning_rate": 1.2736563606711458e-05,
    "timestep_mean": 449.75,
    "epsilon_norm": 8.012585639953613,
    "prediction_norm": 6.911420822143555,
    "timestamp": "2025-07-14T12:34:51.017887"
  },
  {
    "epoch": 2144,
    "batch_idx": 128,
    "loss": 0.19180980324745178,
    "learning_rate": 1.2736563606711458e-05,
    "timestep_mean": 517.3125,
    "epsilon_norm": 6.373985767364502,
    "prediction_norm": 5.897381782531738,
    "timestamp": "2025-07-14T12:34:51.023089"
  },
  {
    "epoch": 2144,
    "batch_idx": 160,
    "loss": 0.3112238645553589,
    "learning_rate": 1.2736563606711458e-05,
    "timestep_mean": 431.9375,
    "epsilon_norm": 8.59748649597168,
    "prediction_norm": 6.883862495422363,
    "timestamp": "2025-07-14T12:34:51.028161"
  },
  {
    "epoch": 2144,
    "batch_idx": 192,
    "loss": 0.16164560616016388,
    "learning_rate": 1.2736563606711458e-05,
    "timestep_mean": 433.5,
    "epsilon_norm": 3.728062391281128,
    "prediction_norm": 3.2824902534484863,
    "timestamp": "2025-07-14T12:34:51.033222"
  },
  {
    "epoch": 2145,
    "batch_idx": 0,
    "loss": 0.36064714193344116,
    "learning_rate": 1.2913306515075348e-05,
    "timestep_mean": 447.5625,
    "epsilon_norm": 6.522423267364502,
    "prediction_norm": 5.343704700469971,
    "timestamp": "2025-07-14T12:34:51.038365"
  },
  {
    "epoch": 2145,
    "batch_idx": 32,
    "loss": 0.3115735650062561,
    "learning_rate": 1.2913306515075348e-05,
    "timestep_mean": 506.375,
    "epsilon_norm": 8.916783332824707,
    "prediction_norm": 8.177826881408691,
    "timestamp": "2025-07-14T12:34:51.043432"
  },
  {
    "epoch": 2145,
    "batch_idx": 64,
    "loss": 0.17636609077453613,
    "learning_rate": 1.2913306515075348e-05,
    "timestep_mean": 504.25,
    "epsilon_norm": 7.554908275604248,
    "prediction_norm": 6.534919261932373,
    "timestamp": "2025-07-14T12:34:51.048729"
  },
  {
    "epoch": 2145,
    "batch_idx": 96,
    "loss": 0.4964934289455414,
    "learning_rate": 1.2913306515075348e-05,
    "timestep_mean": 481.84375,
    "epsilon_norm": 8.850945472717285,
    "prediction_norm": 6.426416873931885,
    "timestamp": "2025-07-14T12:34:51.053969"
  },
  {
    "epoch": 2145,
    "batch_idx": 128,
    "loss": 0.5185216665267944,
    "learning_rate": 1.2913306515075348e-05,
    "timestep_mean": 470.375,
    "epsilon_norm": 7.900918960571289,
    "prediction_norm": 6.206233024597168,
    "timestamp": "2025-07-14T12:34:51.059261"
  },
  {
    "epoch": 2145,
    "batch_idx": 160,
    "loss": 0.3615390658378601,
    "learning_rate": 1.2913306515075348e-05,
    "timestep_mean": 562.0,
    "epsilon_norm": 7.964675426483154,
    "prediction_norm": 6.55560302734375,
    "timestamp": "2025-07-14T12:34:51.064385"
  },
  {
    "epoch": 2145,
    "batch_idx": 192,
    "loss": 0.7040794491767883,
    "learning_rate": 1.2913306515075348e-05,
    "timestep_mean": 333.625,
    "epsilon_norm": 4.441646575927734,
    "prediction_norm": 2.4166324138641357,
    "timestamp": "2025-07-14T12:34:51.069566"
  },
  {
    "epoch": 2146,
    "batch_idx": 0,
    "loss": 0.3600936532020569,
    "learning_rate": 1.3091251261435529e-05,
    "timestep_mean": 570.4375,
    "epsilon_norm": 7.568099498748779,
    "prediction_norm": 6.3121232986450195,
    "timestamp": "2025-07-14T12:34:51.074766"
  },
  {
    "epoch": 2146,
    "batch_idx": 32,
    "loss": 0.3139128088951111,
    "learning_rate": 1.3091251261435529e-05,
    "timestep_mean": 503.875,
    "epsilon_norm": 8.882414817810059,
    "prediction_norm": 7.27664041519165,
    "timestamp": "2025-07-14T12:34:51.080097"
  },
  {
    "epoch": 2146,
    "batch_idx": 64,
    "loss": 0.4470493197441101,
    "learning_rate": 1.3091251261435529e-05,
    "timestep_mean": 434.53125,
    "epsilon_norm": 8.789830207824707,
    "prediction_norm": 6.292790412902832,
    "timestamp": "2025-07-14T12:34:51.085247"
  },
  {
    "epoch": 2146,
    "batch_idx": 96,
    "loss": 0.2203410267829895,
    "learning_rate": 1.3091251261435529e-05,
    "timestep_mean": 432.6875,
    "epsilon_norm": 7.502026081085205,
    "prediction_norm": 5.809362888336182,
    "timestamp": "2025-07-14T12:34:51.090540"
  },
  {
    "epoch": 2146,
    "batch_idx": 128,
    "loss": 0.35931864380836487,
    "learning_rate": 1.3091251261435529e-05,
    "timestep_mean": 559.84375,
    "epsilon_norm": 7.749159336090088,
    "prediction_norm": 6.376730442047119,
    "timestamp": "2025-07-14T12:34:51.095833"
  },
  {
    "epoch": 2146,
    "batch_idx": 160,
    "loss": 0.2408868968486786,
    "learning_rate": 1.3091251261435529e-05,
    "timestep_mean": 571.78125,
    "epsilon_norm": 8.195171356201172,
    "prediction_norm": 7.465574741363525,
    "timestamp": "2025-07-14T12:34:51.101114"
  },
  {
    "epoch": 2146,
    "batch_idx": 192,
    "loss": 0.20472580194473267,
    "learning_rate": 1.3091251261435529e-05,
    "timestep_mean": 516.375,
    "epsilon_norm": 3.7695305347442627,
    "prediction_norm": 3.727574586868286,
    "timestamp": "2025-07-14T12:34:51.106236"
  },
  {
    "epoch": 2147,
    "batch_idx": 0,
    "loss": 0.3773801326751709,
    "learning_rate": 1.3270397406731102e-05,
    "timestep_mean": 446.90625,
    "epsilon_norm": 8.074379920959473,
    "prediction_norm": 6.278836727142334,
    "timestamp": "2025-07-14T12:34:51.111413"
  },
  {
    "epoch": 2147,
    "batch_idx": 32,
    "loss": 0.3851838707923889,
    "learning_rate": 1.3270397406731102e-05,
    "timestep_mean": 483.375,
    "epsilon_norm": 8.623769760131836,
    "prediction_norm": 6.803372859954834,
    "timestamp": "2025-07-14T12:34:51.116610"
  },
  {
    "epoch": 2147,
    "batch_idx": 64,
    "loss": 0.27900439500808716,
    "learning_rate": 1.3270397406731102e-05,
    "timestep_mean": 432.65625,
    "epsilon_norm": 9.18710994720459,
    "prediction_norm": 7.516537189483643,
    "timestamp": "2025-07-14T12:34:51.122036"
  },
  {
    "epoch": 2147,
    "batch_idx": 96,
    "loss": 0.31352537870407104,
    "learning_rate": 1.3270397406731102e-05,
    "timestep_mean": 487.84375,
    "epsilon_norm": 7.330381393432617,
    "prediction_norm": 6.771271228790283,
    "timestamp": "2025-07-14T12:34:51.127252"
  },
  {
    "epoch": 2147,
    "batch_idx": 128,
    "loss": 0.24850976467132568,
    "learning_rate": 1.3270397406731102e-05,
    "timestep_mean": 516.125,
    "epsilon_norm": 7.4591193199157715,
    "prediction_norm": 6.476779460906982,
    "timestamp": "2025-07-14T12:34:51.132589"
  },
  {
    "epoch": 2147,
    "batch_idx": 160,
    "loss": 0.21931827068328857,
    "learning_rate": 1.3270397406731102e-05,
    "timestep_mean": 474.125,
    "epsilon_norm": 7.612268924713135,
    "prediction_norm": 6.486813068389893,
    "timestamp": "2025-07-14T12:34:51.137804"
  },
  {
    "epoch": 2147,
    "batch_idx": 192,
    "loss": 0.1685105264186859,
    "learning_rate": 1.3270397406731102e-05,
    "timestep_mean": 479.875,
    "epsilon_norm": 3.8688013553619385,
    "prediction_norm": 3.4316468238830566,
    "timestamp": "2025-07-14T12:34:51.142923"
  },
  {
    "epoch": 2148,
    "batch_idx": 0,
    "loss": 0.38374629616737366,
    "learning_rate": 1.345074450893676e-05,
    "timestep_mean": 447.53125,
    "epsilon_norm": 8.733125686645508,
    "prediction_norm": 7.524079322814941,
    "timestamp": "2025-07-14T12:34:51.148161"
  },
  {
    "epoch": 2148,
    "batch_idx": 32,
    "loss": 0.2254759967327118,
    "learning_rate": 1.345074450893676e-05,
    "timestep_mean": 538.0625,
    "epsilon_norm": 8.241887092590332,
    "prediction_norm": 7.262294769287109,
    "timestamp": "2025-07-14T12:34:51.153337"
  },
  {
    "epoch": 2148,
    "batch_idx": 64,
    "loss": 0.16144081950187683,
    "learning_rate": 1.345074450893676e-05,
    "timestep_mean": 495.0625,
    "epsilon_norm": 6.9336018562316895,
    "prediction_norm": 6.1581902503967285,
    "timestamp": "2025-07-14T12:34:51.158721"
  },
  {
    "epoch": 2148,
    "batch_idx": 96,
    "loss": 0.37834203243255615,
    "learning_rate": 1.345074450893676e-05,
    "timestep_mean": 546.96875,
    "epsilon_norm": 8.348073959350586,
    "prediction_norm": 7.134807586669922,
    "timestamp": "2025-07-14T12:34:51.164052"
  },
  {
    "epoch": 2148,
    "batch_idx": 128,
    "loss": 0.6188904047012329,
    "learning_rate": 1.345074450893676e-05,
    "timestep_mean": 390.625,
    "epsilon_norm": 7.895516872406006,
    "prediction_norm": 5.305697441101074,
    "timestamp": "2025-07-14T12:34:51.169362"
  },
  {
    "epoch": 2148,
    "batch_idx": 160,
    "loss": 0.275999516248703,
    "learning_rate": 1.345074450893676e-05,
    "timestep_mean": 475.71875,
    "epsilon_norm": 8.079197883605957,
    "prediction_norm": 7.028848171234131,
    "timestamp": "2025-07-14T12:34:51.174223"
  },
  {
    "epoch": 2148,
    "batch_idx": 192,
    "loss": 0.14380313456058502,
    "learning_rate": 1.345074450893676e-05,
    "timestep_mean": 701.0,
    "epsilon_norm": 4.826711177825928,
    "prediction_norm": 4.529826641082764,
    "timestamp": "2025-07-14T12:34:51.178992"
  },
  {
    "epoch": 2149,
    "batch_idx": 0,
    "loss": 0.1475415974855423,
    "learning_rate": 1.3632292123063792e-05,
    "timestep_mean": 521.9375,
    "epsilon_norm": 7.824590682983398,
    "prediction_norm": 7.5829925537109375,
    "timestamp": "2025-07-14T12:34:51.183911"
  },
  {
    "epoch": 2149,
    "batch_idx": 32,
    "loss": 0.5849155187606812,
    "learning_rate": 1.3632292123063792e-05,
    "timestep_mean": 414.84375,
    "epsilon_norm": 8.759785652160645,
    "prediction_norm": 6.557613849639893,
    "timestamp": "2025-07-14T12:34:51.188860"
  },
  {
    "epoch": 2149,
    "batch_idx": 64,
    "loss": 0.3192638158798218,
    "learning_rate": 1.3632292123063792e-05,
    "timestep_mean": 421.71875,
    "epsilon_norm": 9.144681930541992,
    "prediction_norm": 7.287606716156006,
    "timestamp": "2025-07-14T12:34:51.193900"
  },
  {
    "epoch": 2149,
    "batch_idx": 96,
    "loss": 0.4635290503501892,
    "learning_rate": 1.3632292123063792e-05,
    "timestep_mean": 393.9375,
    "epsilon_norm": 8.862519264221191,
    "prediction_norm": 6.791038513183594,
    "timestamp": "2025-07-14T12:34:51.198801"
  },
  {
    "epoch": 2149,
    "batch_idx": 128,
    "loss": 0.13950124382972717,
    "learning_rate": 1.3632292123063792e-05,
    "timestep_mean": 495.5625,
    "epsilon_norm": 9.057908058166504,
    "prediction_norm": 8.575935363769531,
    "timestamp": "2025-07-14T12:34:51.203779"
  },
  {
    "epoch": 2149,
    "batch_idx": 160,
    "loss": 0.2820079028606415,
    "learning_rate": 1.3632292123063792e-05,
    "timestep_mean": 489.53125,
    "epsilon_norm": 7.426655292510986,
    "prediction_norm": 6.613450050354004,
    "timestamp": "2025-07-14T12:34:51.208729"
  },
  {
    "epoch": 2149,
    "batch_idx": 192,
    "loss": 0.07053107023239136,
    "learning_rate": 1.3632292123063792e-05,
    "timestep_mean": 804.25,
    "epsilon_norm": 3.3567564487457275,
    "prediction_norm": 3.3081085681915283,
    "timestamp": "2025-07-14T12:34:51.213552"
  },
  {
    "epoch": 2150,
    "batch_idx": 0,
    "loss": 0.27770888805389404,
    "learning_rate": 1.3815039801161687e-05,
    "timestep_mean": 537.9375,
    "epsilon_norm": 8.55631160736084,
    "prediction_norm": 7.281513690948486,
    "timestamp": "2025-07-14T12:34:51.218559"
  },
  {
    "epoch": 2150,
    "batch_idx": 32,
    "loss": 0.33813390135765076,
    "learning_rate": 1.3815039801161687e-05,
    "timestep_mean": 435.96875,
    "epsilon_norm": 8.0995512008667,
    "prediction_norm": 6.801182746887207,
    "timestamp": "2025-07-14T12:34:51.223441"
  },
  {
    "epoch": 2150,
    "batch_idx": 64,
    "loss": 0.2877862751483917,
    "learning_rate": 1.3815039801161687e-05,
    "timestep_mean": 468.0625,
    "epsilon_norm": 9.874449729919434,
    "prediction_norm": 8.556660652160645,
    "timestamp": "2025-07-14T12:34:51.228387"
  },
  {
    "epoch": 2150,
    "batch_idx": 96,
    "loss": 0.20119716227054596,
    "learning_rate": 1.3815039801161687e-05,
    "timestep_mean": 560.5,
    "epsilon_norm": 7.671741008758545,
    "prediction_norm": 6.92549467086792,
    "timestamp": "2025-07-14T12:34:51.233202"
  },
  {
    "epoch": 2150,
    "batch_idx": 128,
    "loss": 0.26425009965896606,
    "learning_rate": 1.3815039801161687e-05,
    "timestep_mean": 496.4375,
    "epsilon_norm": 8.766454696655273,
    "prediction_norm": 7.295600891113281,
    "timestamp": "2025-07-14T12:34:51.238082"
  },
  {
    "epoch": 2150,
    "batch_idx": 160,
    "loss": 0.16571345925331116,
    "learning_rate": 1.3815039801161687e-05,
    "timestep_mean": 461.125,
    "epsilon_norm": 8.586068153381348,
    "prediction_norm": 8.193183898925781,
    "timestamp": "2025-07-14T12:34:51.242945"
  },
  {
    "epoch": 2150,
    "batch_idx": 192,
    "loss": 0.1462572067975998,
    "learning_rate": 1.3815039801161687e-05,
    "timestep_mean": 558.0,
    "epsilon_norm": 3.6142077445983887,
    "prediction_norm": 3.2112181186676025,
    "timestamp": "2025-07-14T12:34:51.247814"
  },
  {
    "epoch": 2151,
    "batch_idx": 0,
    "loss": 0.2847355008125305,
    "learning_rate": 1.3998987092318694e-05,
    "timestep_mean": 492.21875,
    "epsilon_norm": 8.484983444213867,
    "prediction_norm": 7.164156913757324,
    "timestamp": "2025-07-14T12:34:51.252726"
  },
  {
    "epoch": 2151,
    "batch_idx": 32,
    "loss": 0.24711790680885315,
    "learning_rate": 1.3998987092318694e-05,
    "timestep_mean": 547.0625,
    "epsilon_norm": 6.889727592468262,
    "prediction_norm": 5.885591506958008,
    "timestamp": "2025-07-14T12:34:51.257607"
  },
  {
    "epoch": 2151,
    "batch_idx": 64,
    "loss": 0.2217911034822464,
    "learning_rate": 1.3998987092318694e-05,
    "timestep_mean": 540.21875,
    "epsilon_norm": 7.974112510681152,
    "prediction_norm": 6.739559650421143,
    "timestamp": "2025-07-14T12:34:51.262579"
  },
  {
    "epoch": 2151,
    "batch_idx": 96,
    "loss": 0.23648959398269653,
    "learning_rate": 1.3998987092318694e-05,
    "timestep_mean": 489.8125,
    "epsilon_norm": 7.560051918029785,
    "prediction_norm": 6.730294704437256,
    "timestamp": "2025-07-14T12:34:51.267533"
  },
  {
    "epoch": 2151,
    "batch_idx": 128,
    "loss": 0.2873896360397339,
    "learning_rate": 1.3998987092318694e-05,
    "timestep_mean": 454.375,
    "epsilon_norm": 8.409611701965332,
    "prediction_norm": 6.940802097320557,
    "timestamp": "2025-07-14T12:34:51.272385"
  },
  {
    "epoch": 2151,
    "batch_idx": 160,
    "loss": 0.13474182784557343,
    "learning_rate": 1.3998987092318694e-05,
    "timestep_mean": 542.8125,
    "epsilon_norm": 7.085373401641846,
    "prediction_norm": 6.4447922706604,
    "timestamp": "2025-07-14T12:34:51.277291"
  },
  {
    "epoch": 2151,
    "batch_idx": 192,
    "loss": 0.11466261744499207,
    "learning_rate": 1.3998987092318694e-05,
    "timestep_mean": 451.75,
    "epsilon_norm": 3.5381968021392822,
    "prediction_norm": 3.3820765018463135,
    "timestamp": "2025-07-14T12:34:51.282174"
  },
  {
    "epoch": 2152,
    "batch_idx": 0,
    "loss": 0.23184272646903992,
    "learning_rate": 1.4184133542663037e-05,
    "timestep_mean": 500.09375,
    "epsilon_norm": 8.077536582946777,
    "prediction_norm": 6.763825416564941,
    "timestamp": "2025-07-14T12:34:51.287022"
  },
  {
    "epoch": 2152,
    "batch_idx": 32,
    "loss": 0.2364913672208786,
    "learning_rate": 1.4184133542663037e-05,
    "timestep_mean": 525.03125,
    "epsilon_norm": 7.779599666595459,
    "prediction_norm": 6.668905735015869,
    "timestamp": "2025-07-14T12:34:51.291906"
  },
  {
    "epoch": 2152,
    "batch_idx": 64,
    "loss": 0.24763572216033936,
    "learning_rate": 1.4184133542663037e-05,
    "timestep_mean": 514.0625,
    "epsilon_norm": 6.773046493530273,
    "prediction_norm": 5.853155612945557,
    "timestamp": "2025-07-14T12:34:51.296739"
  },
  {
    "epoch": 2152,
    "batch_idx": 96,
    "loss": 0.2315618395805359,
    "learning_rate": 1.4184133542663037e-05,
    "timestep_mean": 454.53125,
    "epsilon_norm": 8.09958553314209,
    "prediction_norm": 6.910659313201904,
    "timestamp": "2025-07-14T12:34:51.301506"
  },
  {
    "epoch": 2152,
    "batch_idx": 128,
    "loss": 0.24919229745864868,
    "learning_rate": 1.4184133542663037e-05,
    "timestep_mean": 479.4375,
    "epsilon_norm": 8.797607421875,
    "prediction_norm": 7.580145359039307,
    "timestamp": "2025-07-14T12:34:51.306453"
  },
  {
    "epoch": 2152,
    "batch_idx": 160,
    "loss": 0.30629581212997437,
    "learning_rate": 1.4184133542663037e-05,
    "timestep_mean": 550.0625,
    "epsilon_norm": 8.380078315734863,
    "prediction_norm": 6.276772499084473,
    "timestamp": "2025-07-14T12:34:51.311297"
  },
  {
    "epoch": 2152,
    "batch_idx": 192,
    "loss": 0.08968634158372879,
    "learning_rate": 1.4184133542663037e-05,
    "timestep_mean": 632.75,
    "epsilon_norm": 4.466448783874512,
    "prediction_norm": 3.863757610321045,
    "timestamp": "2025-07-14T12:34:51.316164"
  },
  {
    "epoch": 2153,
    "batch_idx": 0,
    "loss": 0.2318599820137024,
    "learning_rate": 1.4370478695364367e-05,
    "timestep_mean": 464.25,
    "epsilon_norm": 7.2192840576171875,
    "prediction_norm": 6.5649943351745605,
    "timestamp": "2025-07-14T12:34:51.321105"
  },
  {
    "epoch": 2153,
    "batch_idx": 32,
    "loss": 0.3485102653503418,
    "learning_rate": 1.4370478695364367e-05,
    "timestep_mean": 460.28125,
    "epsilon_norm": 8.306107521057129,
    "prediction_norm": 6.4870100021362305,
    "timestamp": "2025-07-14T12:34:51.326238"
  },
  {
    "epoch": 2153,
    "batch_idx": 64,
    "loss": 0.12338481843471527,
    "learning_rate": 1.4370478695364367e-05,
    "timestep_mean": 506.28125,
    "epsilon_norm": 7.626993179321289,
    "prediction_norm": 7.104329586029053,
    "timestamp": "2025-07-14T12:34:51.331436"
  },
  {
    "epoch": 2153,
    "batch_idx": 96,
    "loss": 0.17900028824806213,
    "learning_rate": 1.4370478695364367e-05,
    "timestep_mean": 555.0,
    "epsilon_norm": 7.803394317626953,
    "prediction_norm": 6.581977367401123,
    "timestamp": "2025-07-14T12:34:51.336749"
  },
  {
    "epoch": 2153,
    "batch_idx": 128,
    "loss": 0.2734527587890625,
    "learning_rate": 1.4370478695364367e-05,
    "timestep_mean": 467.6875,
    "epsilon_norm": 7.915470123291016,
    "prediction_norm": 6.547241687774658,
    "timestamp": "2025-07-14T12:34:51.342172"
  },
  {
    "epoch": 2153,
    "batch_idx": 160,
    "loss": 0.20443855226039886,
    "learning_rate": 1.4370478695364367e-05,
    "timestep_mean": 514.65625,
    "epsilon_norm": 7.837214469909668,
    "prediction_norm": 6.669092178344727,
    "timestamp": "2025-07-14T12:34:51.347339"
  },
  {
    "epoch": 2153,
    "batch_idx": 192,
    "loss": 0.11412516236305237,
    "learning_rate": 1.4370478695364367e-05,
    "timestep_mean": 545.875,
    "epsilon_norm": 3.866840362548828,
    "prediction_norm": 3.5245413780212402,
    "timestamp": "2025-07-14T12:34:51.352522"
  },
  {
    "epoch": 2154,
    "batch_idx": 0,
    "loss": 0.1774621605873108,
    "learning_rate": 1.4558022090634474e-05,
    "timestep_mean": 505.28125,
    "epsilon_norm": 8.081624984741211,
    "prediction_norm": 7.206718444824219,
    "timestamp": "2025-07-14T12:34:51.357864"
  },
  {
    "epoch": 2154,
    "batch_idx": 32,
    "loss": 0.276852548122406,
    "learning_rate": 1.4558022090634474e-05,
    "timestep_mean": 430.3125,
    "epsilon_norm": 7.960000038146973,
    "prediction_norm": 6.523802757263184,
    "timestamp": "2025-07-14T12:34:51.363287"
  },
  {
    "epoch": 2154,
    "batch_idx": 64,
    "loss": 0.39453721046447754,
    "learning_rate": 1.4558022090634474e-05,
    "timestep_mean": 424.8125,
    "epsilon_norm": 7.131932258605957,
    "prediction_norm": 6.0105509757995605,
    "timestamp": "2025-07-14T12:34:51.368378"
  },
  {
    "epoch": 2154,
    "batch_idx": 96,
    "loss": 0.32225334644317627,
    "learning_rate": 1.4558022090634474e-05,
    "timestep_mean": 441.9375,
    "epsilon_norm": 9.07175350189209,
    "prediction_norm": 7.2921833992004395,
    "timestamp": "2025-07-14T12:34:51.373626"
  },
  {
    "epoch": 2154,
    "batch_idx": 128,
    "loss": 0.12421167641878128,
    "learning_rate": 1.4558022090634474e-05,
    "timestep_mean": 491.6875,
    "epsilon_norm": 6.8157782554626465,
    "prediction_norm": 5.855624675750732,
    "timestamp": "2025-07-14T12:34:51.378914"
  },
  {
    "epoch": 2154,
    "batch_idx": 160,
    "loss": 0.1460043042898178,
    "learning_rate": 1.4558022090634474e-05,
    "timestep_mean": 587.65625,
    "epsilon_norm": 7.592757701873779,
    "prediction_norm": 6.793826580047607,
    "timestamp": "2025-07-14T12:34:51.384333"
  },
  {
    "epoch": 2154,
    "batch_idx": 192,
    "loss": 0.11430475860834122,
    "learning_rate": 1.4558022090634474e-05,
    "timestep_mean": 470.125,
    "epsilon_norm": 3.058878183364868,
    "prediction_norm": 3.0231199264526367,
    "timestamp": "2025-07-14T12:34:51.389349"
  },
  {
    "epoch": 2155,
    "batch_idx": 0,
    "loss": 0.2593843638896942,
    "learning_rate": 1.4746763265728796e-05,
    "timestep_mean": 497.96875,
    "epsilon_norm": 7.051959037780762,
    "prediction_norm": 5.594873905181885,
    "timestamp": "2025-07-14T12:34:51.394549"
  },
  {
    "epoch": 2155,
    "batch_idx": 32,
    "loss": 0.2460172325372696,
    "learning_rate": 1.4746763265728796e-05,
    "timestep_mean": 473.15625,
    "epsilon_norm": 8.636272430419922,
    "prediction_norm": 7.77877140045166,
    "timestamp": "2025-07-14T12:34:51.399792"
  },
  {
    "epoch": 2155,
    "batch_idx": 64,
    "loss": 0.1804153472185135,
    "learning_rate": 1.4746763265728796e-05,
    "timestep_mean": 540.5625,
    "epsilon_norm": 7.2276225090026855,
    "prediction_norm": 6.316305637359619,
    "timestamp": "2025-07-14T12:34:51.405029"
  },
  {
    "epoch": 2155,
    "batch_idx": 96,
    "loss": 0.13059405982494354,
    "learning_rate": 1.4746763265728796e-05,
    "timestep_mean": 537.5,
    "epsilon_norm": 7.687329292297363,
    "prediction_norm": 6.902548313140869,
    "timestamp": "2025-07-14T12:34:51.409940"
  },
  {
    "epoch": 2155,
    "batch_idx": 128,
    "loss": 0.16827869415283203,
    "learning_rate": 1.4746763265728796e-05,
    "timestep_mean": 602.5,
    "epsilon_norm": 6.7654500007629395,
    "prediction_norm": 6.032044887542725,
    "timestamp": "2025-07-14T12:34:51.414856"
  },
  {
    "epoch": 2155,
    "batch_idx": 160,
    "loss": 0.24330376088619232,
    "learning_rate": 1.4746763265728796e-05,
    "timestep_mean": 450.3125,
    "epsilon_norm": 8.386655807495117,
    "prediction_norm": 7.3455071449279785,
    "timestamp": "2025-07-14T12:34:51.420175"
  },
  {
    "epoch": 2155,
    "batch_idx": 192,
    "loss": 0.149021714925766,
    "learning_rate": 1.4746763265728796e-05,
    "timestep_mean": 381.5,
    "epsilon_norm": 3.492110013961792,
    "prediction_norm": 3.1002330780029297,
    "timestamp": "2025-07-14T12:34:51.425463"
  },
  {
    "epoch": 2156,
    "batch_idx": 0,
    "loss": 0.30650457739830017,
    "learning_rate": 1.493670175494713e-05,
    "timestep_mean": 459.40625,
    "epsilon_norm": 7.208836555480957,
    "prediction_norm": 5.702390670776367,
    "timestamp": "2025-07-14T12:34:51.430754"
  },
  {
    "epoch": 2156,
    "batch_idx": 32,
    "loss": 0.23306135833263397,
    "learning_rate": 1.493670175494713e-05,
    "timestep_mean": 496.84375,
    "epsilon_norm": 7.756511211395264,
    "prediction_norm": 6.806015491485596,
    "timestamp": "2025-07-14T12:34:51.435969"
  },
  {
    "epoch": 2156,
    "batch_idx": 64,
    "loss": 0.26175785064697266,
    "learning_rate": 1.493670175494713e-05,
    "timestep_mean": 531.28125,
    "epsilon_norm": 7.818967342376709,
    "prediction_norm": 6.577602863311768,
    "timestamp": "2025-07-14T12:34:51.441186"
  },
  {
    "epoch": 2156,
    "batch_idx": 96,
    "loss": 0.3510562479496002,
    "learning_rate": 1.493670175494713e-05,
    "timestep_mean": 416.5625,
    "epsilon_norm": 7.8576765060424805,
    "prediction_norm": 6.568607807159424,
    "timestamp": "2025-07-14T12:34:51.446330"
  },
  {
    "epoch": 2156,
    "batch_idx": 128,
    "loss": 0.3029389977455139,
    "learning_rate": 1.493670175494713e-05,
    "timestep_mean": 453.25,
    "epsilon_norm": 8.63419246673584,
    "prediction_norm": 6.723848342895508,
    "timestamp": "2025-07-14T12:34:51.451649"
  },
  {
    "epoch": 2156,
    "batch_idx": 160,
    "loss": 0.1700618863105774,
    "learning_rate": 1.493670175494713e-05,
    "timestep_mean": 549.8125,
    "epsilon_norm": 7.7807464599609375,
    "prediction_norm": 6.921867370605469,
    "timestamp": "2025-07-14T12:34:51.456947"
  },
  {
    "epoch": 2156,
    "batch_idx": 192,
    "loss": 0.5208197236061096,
    "learning_rate": 1.493670175494713e-05,
    "timestep_mean": 376.625,
    "epsilon_norm": 4.764187812805176,
    "prediction_norm": 3.665783166885376,
    "timestamp": "2025-07-14T12:34:51.462127"
  },
  {
    "epoch": 2157,
    "batch_idx": 0,
    "loss": 0.27895480394363403,
    "learning_rate": 1.5127837089635082e-05,
    "timestep_mean": 500.375,
    "epsilon_norm": 7.630695819854736,
    "prediction_norm": 6.423348426818848,
    "timestamp": "2025-07-14T12:34:51.467445"
  },
  {
    "epoch": 2157,
    "batch_idx": 32,
    "loss": 0.4255842864513397,
    "learning_rate": 1.5127837089635082e-05,
    "timestep_mean": 513.1875,
    "epsilon_norm": 8.747718811035156,
    "prediction_norm": 6.566196441650391,
    "timestamp": "2025-07-14T12:34:51.472574"
  },
  {
    "epoch": 2157,
    "batch_idx": 64,
    "loss": 0.24829742312431335,
    "learning_rate": 1.5127837089635082e-05,
    "timestep_mean": 567.875,
    "epsilon_norm": 7.565797328948975,
    "prediction_norm": 6.455365180969238,
    "timestamp": "2025-07-14T12:34:51.477770"
  },
  {
    "epoch": 2157,
    "batch_idx": 96,
    "loss": 0.5588312149047852,
    "learning_rate": 1.5127837089635082e-05,
    "timestep_mean": 420.0,
    "epsilon_norm": 7.675083637237549,
    "prediction_norm": 5.488874912261963,
    "timestamp": "2025-07-14T12:34:51.483039"
  },
  {
    "epoch": 2157,
    "batch_idx": 128,
    "loss": 0.1961124837398529,
    "learning_rate": 1.5127837089635082e-05,
    "timestep_mean": 475.25,
    "epsilon_norm": 7.649675369262695,
    "prediction_norm": 7.345437526702881,
    "timestamp": "2025-07-14T12:34:51.488408"
  },
  {
    "epoch": 2157,
    "batch_idx": 160,
    "loss": 0.277915358543396,
    "learning_rate": 1.5127837089635082e-05,
    "timestep_mean": 433.625,
    "epsilon_norm": 7.083102703094482,
    "prediction_norm": 5.960955619812012,
    "timestamp": "2025-07-14T12:34:51.493430"
  },
  {
    "epoch": 2157,
    "batch_idx": 192,
    "loss": 0.1954113245010376,
    "learning_rate": 1.5127837089635082e-05,
    "timestep_mean": 424.0,
    "epsilon_norm": 4.9807233810424805,
    "prediction_norm": 4.317471027374268,
    "timestamp": "2025-07-14T12:34:51.498236"
  },
  {
    "epoch": 2158,
    "batch_idx": 0,
    "loss": 0.22781449556350708,
    "learning_rate": 1.5320168798185403e-05,
    "timestep_mean": 442.84375,
    "epsilon_norm": 7.45710563659668,
    "prediction_norm": 6.392763614654541,
    "timestamp": "2025-07-14T12:34:51.503182"
  },
  {
    "epoch": 2158,
    "batch_idx": 32,
    "loss": 0.31512361764907837,
    "learning_rate": 1.5320168798185403e-05,
    "timestep_mean": 463.0625,
    "epsilon_norm": 7.634487628936768,
    "prediction_norm": 5.521396160125732,
    "timestamp": "2025-07-14T12:34:51.508029"
  },
  {
    "epoch": 2158,
    "batch_idx": 64,
    "loss": 0.17438533902168274,
    "learning_rate": 1.5320168798185403e-05,
    "timestep_mean": 512.25,
    "epsilon_norm": 7.765189170837402,
    "prediction_norm": 7.182044506072998,
    "timestamp": "2025-07-14T12:34:51.512936"
  },
  {
    "epoch": 2158,
    "batch_idx": 96,
    "loss": 0.44419199228286743,
    "learning_rate": 1.5320168798185403e-05,
    "timestep_mean": 546.0,
    "epsilon_norm": 8.429550170898438,
    "prediction_norm": 6.618148326873779,
    "timestamp": "2025-07-14T12:34:51.517902"
  },
  {
    "epoch": 2158,
    "batch_idx": 128,
    "loss": 0.3244333863258362,
    "learning_rate": 1.5320168798185403e-05,
    "timestep_mean": 501.5625,
    "epsilon_norm": 7.549532413482666,
    "prediction_norm": 6.135956764221191,
    "timestamp": "2025-07-14T12:34:51.523116"
  },
  {
    "epoch": 2158,
    "batch_idx": 160,
    "loss": 0.13296467065811157,
    "learning_rate": 1.5320168798185403e-05,
    "timestep_mean": 533.09375,
    "epsilon_norm": 7.547886371612549,
    "prediction_norm": 6.345503807067871,
    "timestamp": "2025-07-14T12:34:51.528217"
  },
  {
    "epoch": 2158,
    "batch_idx": 192,
    "loss": 0.270500510931015,
    "learning_rate": 1.5320168798185403e-05,
    "timestep_mean": 352.5,
    "epsilon_norm": 3.7272651195526123,
    "prediction_norm": 2.683633804321289,
    "timestamp": "2025-07-14T12:34:51.533306"
  },
  {
    "epoch": 2159,
    "batch_idx": 0,
    "loss": 0.18874791264533997,
    "learning_rate": 1.551369640603853e-05,
    "timestep_mean": 453.3125,
    "epsilon_norm": 6.610904693603516,
    "prediction_norm": 5.591344356536865,
    "timestamp": "2025-07-14T12:34:51.538626"
  },
  {
    "epoch": 2159,
    "batch_idx": 32,
    "loss": 0.09150257706642151,
    "learning_rate": 1.551369640603853e-05,
    "timestep_mean": 640.25,
    "epsilon_norm": 7.844418525695801,
    "prediction_norm": 7.32066535949707,
    "timestamp": "2025-07-14T12:34:51.543764"
  },
  {
    "epoch": 2159,
    "batch_idx": 64,
    "loss": 0.36383557319641113,
    "learning_rate": 1.551369640603853e-05,
    "timestep_mean": 490.6875,
    "epsilon_norm": 7.670957565307617,
    "prediction_norm": 5.636868476867676,
    "timestamp": "2025-07-14T12:34:51.548940"
  },
  {
    "epoch": 2159,
    "batch_idx": 96,
    "loss": 0.19938090443611145,
    "learning_rate": 1.551369640603853e-05,
    "timestep_mean": 487.65625,
    "epsilon_norm": 7.235546588897705,
    "prediction_norm": 6.244507789611816,
    "timestamp": "2025-07-14T12:34:51.554215"
  },
  {
    "epoch": 2159,
    "batch_idx": 128,
    "loss": 0.19359999895095825,
    "learning_rate": 1.551369640603853e-05,
    "timestep_mean": 528.3125,
    "epsilon_norm": 8.61150074005127,
    "prediction_norm": 7.593090534210205,
    "timestamp": "2025-07-14T12:34:51.559419"
  },
  {
    "epoch": 2159,
    "batch_idx": 160,
    "loss": 0.3290872573852539,
    "learning_rate": 1.551369640603853e-05,
    "timestep_mean": 468.5625,
    "epsilon_norm": 7.4456787109375,
    "prediction_norm": 6.0433807373046875,
    "timestamp": "2025-07-14T12:34:51.564572"
  },
  {
    "epoch": 2159,
    "batch_idx": 192,
    "loss": 0.3126959502696991,
    "learning_rate": 1.551369640603853e-05,
    "timestep_mean": 547.5,
    "epsilon_norm": 4.097448825836182,
    "prediction_norm": 3.415461540222168,
    "timestamp": "2025-07-14T12:34:51.569743"
  },
  {
    "epoch": 2160,
    "batch_idx": 0,
    "loss": 0.43805694580078125,
    "learning_rate": 1.5708419435684436e-05,
    "timestep_mean": 452.875,
    "epsilon_norm": 7.843635082244873,
    "prediction_norm": 5.887305736541748,
    "timestamp": "2025-07-14T12:34:51.574935"
  },
  {
    "epoch": 2160,
    "batch_idx": 32,
    "loss": 0.14979666471481323,
    "learning_rate": 1.5708419435684436e-05,
    "timestep_mean": 520.5,
    "epsilon_norm": 7.700666427612305,
    "prediction_norm": 6.703777313232422,
    "timestamp": "2025-07-14T12:34:51.580289"
  },
  {
    "epoch": 2160,
    "batch_idx": 64,
    "loss": 0.32994335889816284,
    "learning_rate": 1.5708419435684436e-05,
    "timestep_mean": 493.46875,
    "epsilon_norm": 8.656874656677246,
    "prediction_norm": 6.49025821685791,
    "timestamp": "2025-07-14T12:34:51.585461"
  },
  {
    "epoch": 2160,
    "batch_idx": 96,
    "loss": 0.328868567943573,
    "learning_rate": 1.5708419435684436e-05,
    "timestep_mean": 524.46875,
    "epsilon_norm": 7.719762325286865,
    "prediction_norm": 6.3020453453063965,
    "timestamp": "2025-07-14T12:34:51.590653"
  },
  {
    "epoch": 2160,
    "batch_idx": 128,
    "loss": 0.28523191809654236,
    "learning_rate": 1.5708419435684436e-05,
    "timestep_mean": 457.1875,
    "epsilon_norm": 7.603763103485107,
    "prediction_norm": 6.260280132293701,
    "timestamp": "2025-07-14T12:34:51.595784"
  },
  {
    "epoch": 2160,
    "batch_idx": 160,
    "loss": 0.2222372442483902,
    "learning_rate": 1.5708419435684436e-05,
    "timestep_mean": 505.375,
    "epsilon_norm": 8.440210342407227,
    "prediction_norm": 7.235265731811523,
    "timestamp": "2025-07-14T12:34:51.601137"
  },
  {
    "epoch": 2160,
    "batch_idx": 192,
    "loss": 0.27580976486206055,
    "learning_rate": 1.5708419435684436e-05,
    "timestep_mean": 528.0,
    "epsilon_norm": 4.480764865875244,
    "prediction_norm": 4.504003524780273,
    "timestamp": "2025-07-14T12:34:51.606256"
  },
  {
    "epoch": 2161,
    "batch_idx": 0,
    "loss": 0.3484012484550476,
    "learning_rate": 1.590433740666339e-05,
    "timestep_mean": 441.5,
    "epsilon_norm": 7.697351455688477,
    "prediction_norm": 6.012073040008545,
    "timestamp": "2025-07-14T12:34:51.611412"
  },
  {
    "epoch": 2161,
    "batch_idx": 32,
    "loss": 0.2806788384914398,
    "learning_rate": 1.590433740666339e-05,
    "timestep_mean": 404.96875,
    "epsilon_norm": 7.755016803741455,
    "prediction_norm": 6.68959379196167,
    "timestamp": "2025-07-14T12:34:51.616506"
  },
  {
    "epoch": 2161,
    "batch_idx": 64,
    "loss": 0.2620134949684143,
    "learning_rate": 1.590433740666339e-05,
    "timestep_mean": 430.875,
    "epsilon_norm": 7.831623554229736,
    "prediction_norm": 6.25811767578125,
    "timestamp": "2025-07-14T12:34:51.621743"
  },
  {
    "epoch": 2161,
    "batch_idx": 96,
    "loss": 0.2547423243522644,
    "learning_rate": 1.590433740666339e-05,
    "timestep_mean": 443.28125,
    "epsilon_norm": 7.996334075927734,
    "prediction_norm": 6.744884490966797,
    "timestamp": "2025-07-14T12:34:51.627009"
  },
  {
    "epoch": 2161,
    "batch_idx": 128,
    "loss": 0.2099992036819458,
    "learning_rate": 1.590433740666339e-05,
    "timestep_mean": 492.5625,
    "epsilon_norm": 7.834962368011475,
    "prediction_norm": 7.273488998413086,
    "timestamp": "2025-07-14T12:34:51.632098"
  },
  {
    "epoch": 2161,
    "batch_idx": 160,
    "loss": 0.39279672503471375,
    "learning_rate": 1.590433740666339e-05,
    "timestep_mean": 421.15625,
    "epsilon_norm": 8.131887435913086,
    "prediction_norm": 5.965307235717773,
    "timestamp": "2025-07-14T12:34:51.637257"
  },
  {
    "epoch": 2161,
    "batch_idx": 192,
    "loss": 0.047414593398571014,
    "learning_rate": 1.590433740666339e-05,
    "timestep_mean": 538.0,
    "epsilon_norm": 2.686572551727295,
    "prediction_norm": 2.3192007541656494,
    "timestamp": "2025-07-14T12:34:51.642378"
  },
  {
    "epoch": 2162,
    "batch_idx": 0,
    "loss": 0.28605765104293823,
    "learning_rate": 1.61014498355673e-05,
    "timestep_mean": 452.1875,
    "epsilon_norm": 8.848002433776855,
    "prediction_norm": 7.684994220733643,
    "timestamp": "2025-07-14T12:34:51.647600"
  },
  {
    "epoch": 2162,
    "batch_idx": 32,
    "loss": 0.31243887543678284,
    "learning_rate": 1.61014498355673e-05,
    "timestep_mean": 486.28125,
    "epsilon_norm": 9.667539596557617,
    "prediction_norm": 7.888558387756348,
    "timestamp": "2025-07-14T12:34:51.652743"
  },
  {
    "epoch": 2162,
    "batch_idx": 64,
    "loss": 0.3164973258972168,
    "learning_rate": 1.61014498355673e-05,
    "timestep_mean": 491.6875,
    "epsilon_norm": 7.70410680770874,
    "prediction_norm": 6.105097770690918,
    "timestamp": "2025-07-14T12:34:51.657842"
  },
  {
    "epoch": 2162,
    "batch_idx": 96,
    "loss": 0.21658290922641754,
    "learning_rate": 1.61014498355673e-05,
    "timestep_mean": 456.46875,
    "epsilon_norm": 7.18804407119751,
    "prediction_norm": 5.89359712600708,
    "timestamp": "2025-07-14T12:34:51.663294"
  },
  {
    "epoch": 2162,
    "batch_idx": 128,
    "loss": 0.36208030581474304,
    "learning_rate": 1.61014498355673e-05,
    "timestep_mean": 473.5625,
    "epsilon_norm": 8.69564151763916,
    "prediction_norm": 7.483022689819336,
    "timestamp": "2025-07-14T12:34:51.668425"
  },
  {
    "epoch": 2162,
    "batch_idx": 160,
    "loss": 0.3185265064239502,
    "learning_rate": 1.61014498355673e-05,
    "timestep_mean": 434.3125,
    "epsilon_norm": 7.141158580780029,
    "prediction_norm": 5.82009744644165,
    "timestamp": "2025-07-14T12:34:51.673653"
  },
  {
    "epoch": 2162,
    "batch_idx": 192,
    "loss": 0.06422924250364304,
    "learning_rate": 1.61014498355673e-05,
    "timestep_mean": 582.75,
    "epsilon_norm": 4.311668872833252,
    "prediction_norm": 4.351570129394531,
    "timestamp": "2025-07-14T12:34:51.678823"
  },
  {
    "epoch": 2163,
    "batch_idx": 0,
    "loss": 0.20952649414539337,
    "learning_rate": 1.6299756236040773e-05,
    "timestep_mean": 484.96875,
    "epsilon_norm": 8.256346702575684,
    "prediction_norm": 7.327296257019043,
    "timestamp": "2025-07-14T12:34:51.684077"
  },
  {
    "epoch": 2163,
    "batch_idx": 32,
    "loss": 0.283454954624176,
    "learning_rate": 1.6299756236040773e-05,
    "timestep_mean": 451.65625,
    "epsilon_norm": 7.735880374908447,
    "prediction_norm": 6.178924560546875,
    "timestamp": "2025-07-14T12:34:51.689403"
  },
  {
    "epoch": 2163,
    "batch_idx": 64,
    "loss": 0.2334427535533905,
    "learning_rate": 1.6299756236040773e-05,
    "timestep_mean": 467.21875,
    "epsilon_norm": 8.597877502441406,
    "prediction_norm": 7.7526984214782715,
    "timestamp": "2025-07-14T12:34:51.694542"
  },
  {
    "epoch": 2163,
    "batch_idx": 96,
    "loss": 0.23454000055789948,
    "learning_rate": 1.6299756236040773e-05,
    "timestep_mean": 592.25,
    "epsilon_norm": 7.718217372894287,
    "prediction_norm": 7.470861434936523,
    "timestamp": "2025-07-14T12:34:51.699801"
  },
  {
    "epoch": 2163,
    "batch_idx": 128,
    "loss": 0.34206241369247437,
    "learning_rate": 1.6299756236040773e-05,
    "timestep_mean": 456.34375,
    "epsilon_norm": 8.574616432189941,
    "prediction_norm": 6.764487266540527,
    "timestamp": "2025-07-14T12:34:51.705225"
  },
  {
    "epoch": 2163,
    "batch_idx": 160,
    "loss": 0.3170262277126312,
    "learning_rate": 1.6299756236040773e-05,
    "timestep_mean": 415.0,
    "epsilon_norm": 7.708319664001465,
    "prediction_norm": 6.133222579956055,
    "timestamp": "2025-07-14T12:34:51.710421"
  },
  {
    "epoch": 2163,
    "batch_idx": 192,
    "loss": 0.11894029378890991,
    "learning_rate": 1.6299756236040773e-05,
    "timestep_mean": 452.125,
    "epsilon_norm": 2.8499279022216797,
    "prediction_norm": 2.9590044021606445,
    "timestamp": "2025-07-14T12:34:51.715565"
  },
  {
    "epoch": 2164,
    "batch_idx": 0,
    "loss": 0.2899167537689209,
    "learning_rate": 1.649925611878248e-05,
    "timestep_mean": 492.625,
    "epsilon_norm": 7.443445205688477,
    "prediction_norm": 6.375082969665527,
    "timestamp": "2025-07-14T12:34:51.720810"
  },
  {
    "epoch": 2164,
    "batch_idx": 32,
    "loss": 0.18226638436317444,
    "learning_rate": 1.649925611878248e-05,
    "timestep_mean": 422.5,
    "epsilon_norm": 7.6310882568359375,
    "prediction_norm": 6.781177043914795,
    "timestamp": "2025-07-14T12:34:51.726124"
  },
  {
    "epoch": 2164,
    "batch_idx": 64,
    "loss": 0.2682037949562073,
    "learning_rate": 1.649925611878248e-05,
    "timestep_mean": 437.25,
    "epsilon_norm": 7.393654823303223,
    "prediction_norm": 5.680342197418213,
    "timestamp": "2025-07-14T12:34:51.731375"
  },
  {
    "epoch": 2164,
    "batch_idx": 96,
    "loss": 0.15672123432159424,
    "learning_rate": 1.649925611878248e-05,
    "timestep_mean": 482.875,
    "epsilon_norm": 7.4443511962890625,
    "prediction_norm": 6.810126304626465,
    "timestamp": "2025-07-14T12:34:51.736546"
  },
  {
    "epoch": 2164,
    "batch_idx": 128,
    "loss": 0.5240998864173889,
    "learning_rate": 1.649925611878248e-05,
    "timestep_mean": 398.0625,
    "epsilon_norm": 8.766148567199707,
    "prediction_norm": 6.875383377075195,
    "timestamp": "2025-07-14T12:34:51.741891"
  },
  {
    "epoch": 2164,
    "batch_idx": 160,
    "loss": 0.2315877228975296,
    "learning_rate": 1.649925611878248e-05,
    "timestep_mean": 504.875,
    "epsilon_norm": 7.507619380950928,
    "prediction_norm": 6.7205939292907715,
    "timestamp": "2025-07-14T12:34:51.747177"
  },
  {
    "epoch": 2164,
    "batch_idx": 192,
    "loss": 0.08690164238214493,
    "learning_rate": 1.649925611878248e-05,
    "timestep_mean": 747.5,
    "epsilon_norm": 3.7038681507110596,
    "prediction_norm": 3.577583074569702,
    "timestamp": "2025-07-14T12:34:51.752340"
  },
  {
    "epoch": 2165,
    "batch_idx": 0,
    "loss": 0.29123371839523315,
    "learning_rate": 1.6699948991546343e-05,
    "timestep_mean": 600.40625,
    "epsilon_norm": 8.489806175231934,
    "prediction_norm": 7.533815383911133,
    "timestamp": "2025-07-14T12:34:51.757587"
  },
  {
    "epoch": 2165,
    "batch_idx": 32,
    "loss": 0.18508771061897278,
    "learning_rate": 1.6699948991546343e-05,
    "timestep_mean": 474.15625,
    "epsilon_norm": 6.89445161819458,
    "prediction_norm": 6.175425052642822,
    "timestamp": "2025-07-14T12:34:51.762812"
  },
  {
    "epoch": 2165,
    "batch_idx": 64,
    "loss": 0.36143958568573,
    "learning_rate": 1.6699948991546343e-05,
    "timestep_mean": 515.59375,
    "epsilon_norm": 8.327386856079102,
    "prediction_norm": 6.919075965881348,
    "timestamp": "2025-07-14T12:34:51.767976"
  },
  {
    "epoch": 2165,
    "batch_idx": 96,
    "loss": 0.20725244283676147,
    "learning_rate": 1.6699948991546343e-05,
    "timestep_mean": 520.09375,
    "epsilon_norm": 7.617985248565674,
    "prediction_norm": 6.436886787414551,
    "timestamp": "2025-07-14T12:34:51.773259"
  },
  {
    "epoch": 2165,
    "batch_idx": 128,
    "loss": 0.20572206377983093,
    "learning_rate": 1.6699948991546343e-05,
    "timestep_mean": 551.28125,
    "epsilon_norm": 9.16019058227539,
    "prediction_norm": 8.20396900177002,
    "timestamp": "2025-07-14T12:34:51.778493"
  },
  {
    "epoch": 2165,
    "batch_idx": 160,
    "loss": 0.26229774951934814,
    "learning_rate": 1.6699948991546343e-05,
    "timestep_mean": 456.84375,
    "epsilon_norm": 7.058103084564209,
    "prediction_norm": 6.0115580558776855,
    "timestamp": "2025-07-14T12:34:51.783824"
  },
  {
    "epoch": 2165,
    "batch_idx": 192,
    "loss": 0.08042273670434952,
    "learning_rate": 1.6699948991546343e-05,
    "timestep_mean": 634.375,
    "epsilon_norm": 3.7392196655273438,
    "prediction_norm": 3.598600149154663,
    "timestamp": "2025-07-14T12:34:51.789025"
  },
  {
    "epoch": 2166,
    "batch_idx": 0,
    "loss": 0.2695144712924957,
    "learning_rate": 1.6901834359142645e-05,
    "timestep_mean": 551.90625,
    "epsilon_norm": 9.366888999938965,
    "prediction_norm": 8.048216819763184,
    "timestamp": "2025-07-14T12:34:51.794191"
  },
  {
    "epoch": 2166,
    "batch_idx": 32,
    "loss": 0.2073528915643692,
    "learning_rate": 1.6901834359142645e-05,
    "timestep_mean": 569.46875,
    "epsilon_norm": 7.086822032928467,
    "prediction_norm": 6.429475784301758,
    "timestamp": "2025-07-14T12:34:51.799445"
  },
  {
    "epoch": 2166,
    "batch_idx": 64,
    "loss": 0.22214481234550476,
    "learning_rate": 1.6901834359142645e-05,
    "timestep_mean": 465.53125,
    "epsilon_norm": 8.662089347839355,
    "prediction_norm": 6.989349365234375,
    "timestamp": "2025-07-14T12:34:51.804683"
  },
  {
    "epoch": 2166,
    "batch_idx": 96,
    "loss": 0.47339263558387756,
    "learning_rate": 1.6901834359142645e-05,
    "timestep_mean": 448.375,
    "epsilon_norm": 9.65898609161377,
    "prediction_norm": 7.558119773864746,
    "timestamp": "2025-07-14T12:34:51.809918"
  },
  {
    "epoch": 2166,
    "batch_idx": 128,
    "loss": 0.4225825071334839,
    "learning_rate": 1.6901834359142645e-05,
    "timestep_mean": 453.28125,
    "epsilon_norm": 8.94582748413086,
    "prediction_norm": 7.048630714416504,
    "timestamp": "2025-07-14T12:34:51.815116"
  },
  {
    "epoch": 2166,
    "batch_idx": 160,
    "loss": 0.2550293803215027,
    "learning_rate": 1.6901834359142645e-05,
    "timestep_mean": 522.21875,
    "epsilon_norm": 8.133209228515625,
    "prediction_norm": 6.616236686706543,
    "timestamp": "2025-07-14T12:34:51.820265"
  },
  {
    "epoch": 2166,
    "batch_idx": 192,
    "loss": 0.10595592856407166,
    "learning_rate": 1.6901834359142645e-05,
    "timestep_mean": 575.875,
    "epsilon_norm": 3.8324198722839355,
    "prediction_norm": 3.32940673828125,
    "timestamp": "2025-07-14T12:34:51.825690"
  },
  {
    "epoch": 2167,
    "batch_idx": 0,
    "loss": 0.2511686384677887,
    "learning_rate": 1.710491172343929e-05,
    "timestep_mean": 503.0,
    "epsilon_norm": 8.063349723815918,
    "prediction_norm": 6.9376935958862305,
    "timestamp": "2025-07-14T12:34:51.830794"
  },
  {
    "epoch": 2167,
    "batch_idx": 32,
    "loss": 0.3525545597076416,
    "learning_rate": 1.710491172343929e-05,
    "timestep_mean": 527.5625,
    "epsilon_norm": 8.671412467956543,
    "prediction_norm": 6.736069679260254,
    "timestamp": "2025-07-14T12:34:51.835852"
  },
  {
    "epoch": 2167,
    "batch_idx": 64,
    "loss": 0.19727276265621185,
    "learning_rate": 1.710491172343929e-05,
    "timestep_mean": 472.8125,
    "epsilon_norm": 7.267811298370361,
    "prediction_norm": 6.156298637390137,
    "timestamp": "2025-07-14T12:34:51.840819"
  },
  {
    "epoch": 2167,
    "batch_idx": 96,
    "loss": 0.27716219425201416,
    "learning_rate": 1.710491172343929e-05,
    "timestep_mean": 472.15625,
    "epsilon_norm": 8.44787883758545,
    "prediction_norm": 7.5335540771484375,
    "timestamp": "2025-07-14T12:34:51.845646"
  },
  {
    "epoch": 2167,
    "batch_idx": 128,
    "loss": 0.28636908531188965,
    "learning_rate": 1.710491172343929e-05,
    "timestep_mean": 483.0625,
    "epsilon_norm": 8.650873184204102,
    "prediction_norm": 7.010990619659424,
    "timestamp": "2025-07-14T12:34:51.850662"
  },
  {
    "epoch": 2167,
    "batch_idx": 160,
    "loss": 0.20057587325572968,
    "learning_rate": 1.710491172343929e-05,
    "timestep_mean": 543.90625,
    "epsilon_norm": 8.596329689025879,
    "prediction_norm": 7.348715782165527,
    "timestamp": "2025-07-14T12:34:51.855482"
  },
  {
    "epoch": 2167,
    "batch_idx": 192,
    "loss": 0.20745685696601868,
    "learning_rate": 1.710491172343929e-05,
    "timestep_mean": 458.0,
    "epsilon_norm": 3.8837766647338867,
    "prediction_norm": 3.4666972160339355,
    "timestamp": "2025-07-14T12:34:51.860509"
  },
  {
    "epoch": 2168,
    "batch_idx": 0,
    "loss": 0.27892830967903137,
    "learning_rate": 1.7309180583363096e-05,
    "timestep_mean": 576.21875,
    "epsilon_norm": 9.151808738708496,
    "prediction_norm": 7.363502502441406,
    "timestamp": "2025-07-14T12:34:51.865678"
  },
  {
    "epoch": 2168,
    "batch_idx": 32,
    "loss": 0.3220862150192261,
    "learning_rate": 1.7309180583363096e-05,
    "timestep_mean": 563.5,
    "epsilon_norm": 8.908356666564941,
    "prediction_norm": 7.9392924308776855,
    "timestamp": "2025-07-14T12:34:51.870836"
  },
  {
    "epoch": 2168,
    "batch_idx": 64,
    "loss": 0.32992035150527954,
    "learning_rate": 1.7309180583363096e-05,
    "timestep_mean": 506.65625,
    "epsilon_norm": 7.861856937408447,
    "prediction_norm": 6.332399368286133,
    "timestamp": "2025-07-14T12:34:51.875907"
  },
  {
    "epoch": 2168,
    "batch_idx": 96,
    "loss": 0.3501097559928894,
    "learning_rate": 1.7309180583363096e-05,
    "timestep_mean": 433.34375,
    "epsilon_norm": 7.881492614746094,
    "prediction_norm": 6.623791217803955,
    "timestamp": "2025-07-14T12:34:51.881047"
  },
  {
    "epoch": 2168,
    "batch_idx": 128,
    "loss": 0.5146650075912476,
    "learning_rate": 1.7309180583363096e-05,
    "timestep_mean": 390.65625,
    "epsilon_norm": 7.826972007751465,
    "prediction_norm": 5.752448081970215,
    "timestamp": "2025-07-14T12:34:51.886195"
  },
  {
    "epoch": 2168,
    "batch_idx": 160,
    "loss": 0.4155172109603882,
    "learning_rate": 1.7309180583363096e-05,
    "timestep_mean": 540.15625,
    "epsilon_norm": 9.68301010131836,
    "prediction_norm": 8.106039047241211,
    "timestamp": "2025-07-14T12:34:51.891383"
  },
  {
    "epoch": 2168,
    "batch_idx": 192,
    "loss": 0.06346841156482697,
    "learning_rate": 1.7309180583363096e-05,
    "timestep_mean": 480.75,
    "epsilon_norm": 4.342888355255127,
    "prediction_norm": 4.261517524719238,
    "timestamp": "2025-07-14T12:34:51.896725"
  },
  {
    "epoch": 2169,
    "batch_idx": 0,
    "loss": 0.4816534221172333,
    "learning_rate": 1.7514640434900897e-05,
    "timestep_mean": 418.9375,
    "epsilon_norm": 8.456855773925781,
    "prediction_norm": 6.821827411651611,
    "timestamp": "2025-07-14T12:34:51.902002"
  },
  {
    "epoch": 2169,
    "batch_idx": 32,
    "loss": 0.2785204350948334,
    "learning_rate": 1.7514640434900897e-05,
    "timestep_mean": 453.53125,
    "epsilon_norm": 7.8023152351379395,
    "prediction_norm": 6.6039838790893555,
    "timestamp": "2025-07-14T12:34:51.907085"
  },
  {
    "epoch": 2169,
    "batch_idx": 64,
    "loss": 0.33302128314971924,
    "learning_rate": 1.7514640434900897e-05,
    "timestep_mean": 529.84375,
    "epsilon_norm": 8.720260620117188,
    "prediction_norm": 7.6014485359191895,
    "timestamp": "2025-07-14T12:34:51.912289"
  },
  {
    "epoch": 2169,
    "batch_idx": 96,
    "loss": 0.2730433940887451,
    "learning_rate": 1.7514640434900897e-05,
    "timestep_mean": 576.5625,
    "epsilon_norm": 7.748023986816406,
    "prediction_norm": 6.688211441040039,
    "timestamp": "2025-07-14T12:34:51.917628"
  },
  {
    "epoch": 2169,
    "batch_idx": 128,
    "loss": 0.3613714575767517,
    "learning_rate": 1.7514640434900897e-05,
    "timestep_mean": 475.96875,
    "epsilon_norm": 8.523433685302734,
    "prediction_norm": 6.879965305328369,
    "timestamp": "2025-07-14T12:34:51.922430"
  },
  {
    "epoch": 2169,
    "batch_idx": 160,
    "loss": 0.19658900797367096,
    "learning_rate": 1.7514640434900897e-05,
    "timestep_mean": 541.34375,
    "epsilon_norm": 7.795179843902588,
    "prediction_norm": 6.833507537841797,
    "timestamp": "2025-07-14T12:34:51.927627"
  },
  {
    "epoch": 2169,
    "batch_idx": 192,
    "loss": 0.8244989514350891,
    "learning_rate": 1.7514640434900897e-05,
    "timestep_mean": 500.75,
    "epsilon_norm": 4.845831871032715,
    "prediction_norm": 2.8589541912078857,
    "timestamp": "2025-07-14T12:34:51.932786"
  },
  {
    "epoch": 2170,
    "batch_idx": 0,
    "loss": 0.30296218395233154,
    "learning_rate": 1.7721290771100996e-05,
    "timestep_mean": 517.875,
    "epsilon_norm": 8.981515884399414,
    "prediction_norm": 7.017187118530273,
    "timestamp": "2025-07-14T12:34:51.937927"
  },
  {
    "epoch": 2170,
    "batch_idx": 32,
    "loss": 0.46197882294654846,
    "learning_rate": 1.7721290771100996e-05,
    "timestep_mean": 447.5625,
    "epsilon_norm": 7.864575386047363,
    "prediction_norm": 6.039435386657715,
    "timestamp": "2025-07-14T12:34:51.943067"
  },
  {
    "epoch": 2170,
    "batch_idx": 64,
    "loss": 0.2244015634059906,
    "learning_rate": 1.7721290771100996e-05,
    "timestep_mean": 429.40625,
    "epsilon_norm": 7.472254276275635,
    "prediction_norm": 6.303593158721924,
    "timestamp": "2025-07-14T12:34:51.948386"
  },
  {
    "epoch": 2170,
    "batch_idx": 96,
    "loss": 0.23874914646148682,
    "learning_rate": 1.7721290771100996e-05,
    "timestep_mean": 489.53125,
    "epsilon_norm": 7.162150859832764,
    "prediction_norm": 6.13884973526001,
    "timestamp": "2025-07-14T12:34:51.953629"
  },
  {
    "epoch": 2170,
    "batch_idx": 128,
    "loss": 0.16994896531105042,
    "learning_rate": 1.7721290771100996e-05,
    "timestep_mean": 603.0625,
    "epsilon_norm": 7.696821689605713,
    "prediction_norm": 7.135982036590576,
    "timestamp": "2025-07-14T12:34:51.959080"
  },
  {
    "epoch": 2170,
    "batch_idx": 160,
    "loss": 0.41474267840385437,
    "learning_rate": 1.7721290771100996e-05,
    "timestep_mean": 447.375,
    "epsilon_norm": 8.277512550354004,
    "prediction_norm": 6.94473123550415,
    "timestamp": "2025-07-14T12:34:51.964311"
  },
  {
    "epoch": 2170,
    "batch_idx": 192,
    "loss": 0.844351053237915,
    "learning_rate": 1.7721290771100996e-05,
    "timestep_mean": 472.625,
    "epsilon_norm": 4.265419006347656,
    "prediction_norm": 3.0258030891418457,
    "timestamp": "2025-07-14T12:34:51.969412"
  },
  {
    "epoch": 2171,
    "batch_idx": 0,
    "loss": 0.5527151823043823,
    "learning_rate": 1.7929131082074144e-05,
    "timestep_mean": 478.21875,
    "epsilon_norm": 8.379168510437012,
    "prediction_norm": 6.234854221343994,
    "timestamp": "2025-07-14T12:34:51.974625"
  },
  {
    "epoch": 2171,
    "batch_idx": 32,
    "loss": 0.34936410188674927,
    "learning_rate": 1.7929131082074144e-05,
    "timestep_mean": 519.4375,
    "epsilon_norm": 6.9393086433410645,
    "prediction_norm": 5.249261856079102,
    "timestamp": "2025-07-14T12:34:51.979998"
  },
  {
    "epoch": 2171,
    "batch_idx": 64,
    "loss": 0.2064521312713623,
    "learning_rate": 1.7929131082074144e-05,
    "timestep_mean": 540.3125,
    "epsilon_norm": 6.950139999389648,
    "prediction_norm": 6.104191303253174,
    "timestamp": "2025-07-14T12:34:51.985234"
  },
  {
    "epoch": 2171,
    "batch_idx": 96,
    "loss": 0.35036224126815796,
    "learning_rate": 1.7929131082074144e-05,
    "timestep_mean": 503.84375,
    "epsilon_norm": 8.183829307556152,
    "prediction_norm": 6.8496880531311035,
    "timestamp": "2025-07-14T12:34:51.990259"
  },
  {
    "epoch": 2171,
    "batch_idx": 128,
    "loss": 0.08158816397190094,
    "learning_rate": 1.7929131082074144e-05,
    "timestep_mean": 555.21875,
    "epsilon_norm": 7.267029285430908,
    "prediction_norm": 6.882390975952148,
    "timestamp": "2025-07-14T12:34:51.995303"
  },
  {
    "epoch": 2171,
    "batch_idx": 160,
    "loss": 0.3518032431602478,
    "learning_rate": 1.7929131082074144e-05,
    "timestep_mean": 598.53125,
    "epsilon_norm": 8.227252006530762,
    "prediction_norm": 7.071324825286865,
    "timestamp": "2025-07-14T12:34:52.000703"
  },
  {
    "epoch": 2171,
    "batch_idx": 192,
    "loss": 0.04699074849486351,
    "learning_rate": 1.7929131082074144e-05,
    "timestep_mean": 572.5,
    "epsilon_norm": 3.6327192783355713,
    "prediction_norm": 3.556485891342163,
    "timestamp": "2025-07-14T12:34:52.005901"
  },
  {
    "epoch": 2172,
    "batch_idx": 0,
    "loss": 0.184954434633255,
    "learning_rate": 1.8138160854995175e-05,
    "timestep_mean": 527.3125,
    "epsilon_norm": 7.611405372619629,
    "prediction_norm": 7.032492160797119,
    "timestamp": "2025-07-14T12:34:52.011096"
  },
  {
    "epoch": 2172,
    "batch_idx": 32,
    "loss": 0.1483188271522522,
    "learning_rate": 1.8138160854995175e-05,
    "timestep_mean": 643.8125,
    "epsilon_norm": 8.266639709472656,
    "prediction_norm": 7.9278059005737305,
    "timestamp": "2025-07-14T12:34:52.016286"
  },
  {
    "epoch": 2172,
    "batch_idx": 64,
    "loss": 0.3292641043663025,
    "learning_rate": 1.8138160854995175e-05,
    "timestep_mean": 484.375,
    "epsilon_norm": 7.824622631072998,
    "prediction_norm": 5.648508548736572,
    "timestamp": "2025-07-14T12:34:52.021624"
  },
  {
    "epoch": 2172,
    "batch_idx": 96,
    "loss": 0.47821497917175293,
    "learning_rate": 1.8138160854995175e-05,
    "timestep_mean": 492.25,
    "epsilon_norm": 8.33144760131836,
    "prediction_norm": 7.143098831176758,
    "timestamp": "2025-07-14T12:34:52.026858"
  },
  {
    "epoch": 2172,
    "batch_idx": 128,
    "loss": 0.1937970370054245,
    "learning_rate": 1.8138160854995175e-05,
    "timestep_mean": 494.84375,
    "epsilon_norm": 7.262972354888916,
    "prediction_norm": 5.979762554168701,
    "timestamp": "2025-07-14T12:34:52.032003"
  },
  {
    "epoch": 2172,
    "batch_idx": 160,
    "loss": 0.4126243591308594,
    "learning_rate": 1.8138160854995175e-05,
    "timestep_mean": 567.71875,
    "epsilon_norm": 7.953196048736572,
    "prediction_norm": 6.483335971832275,
    "timestamp": "2025-07-14T12:34:52.037229"
  },
  {
    "epoch": 2172,
    "batch_idx": 192,
    "loss": 0.32447177171707153,
    "learning_rate": 1.8138160854995175e-05,
    "timestep_mean": 471.75,
    "epsilon_norm": 4.020171165466309,
    "prediction_norm": 3.4888663291931152,
    "timestamp": "2025-07-14T12:34:52.042616"
  },
  {
    "epoch": 2173,
    "batch_idx": 0,
    "loss": 0.13099601864814758,
    "learning_rate": 1.8348379574103764e-05,
    "timestep_mean": 581.8125,
    "epsilon_norm": 8.582120895385742,
    "prediction_norm": 8.124003410339355,
    "timestamp": "2025-07-14T12:34:52.047888"
  },
  {
    "epoch": 2173,
    "batch_idx": 32,
    "loss": 0.20728114247322083,
    "learning_rate": 1.8348379574103764e-05,
    "timestep_mean": 415.625,
    "epsilon_norm": 6.824642658233643,
    "prediction_norm": 5.785641670227051,
    "timestamp": "2025-07-14T12:34:52.053039"
  },
  {
    "epoch": 2173,
    "batch_idx": 64,
    "loss": 0.29381394386291504,
    "learning_rate": 1.8348379574103764e-05,
    "timestep_mean": 507.53125,
    "epsilon_norm": 8.498411178588867,
    "prediction_norm": 7.622684001922607,
    "timestamp": "2025-07-14T12:34:52.058253"
  },
  {
    "epoch": 2173,
    "batch_idx": 96,
    "loss": 0.44681069254875183,
    "learning_rate": 1.8348379574103764e-05,
    "timestep_mean": 542.1875,
    "epsilon_norm": 9.206171035766602,
    "prediction_norm": 7.19915246963501,
    "timestamp": "2025-07-14T12:34:52.063553"
  },
  {
    "epoch": 2173,
    "batch_idx": 128,
    "loss": 0.21090367436408997,
    "learning_rate": 1.8348379574103764e-05,
    "timestep_mean": 470.09375,
    "epsilon_norm": 8.122694969177246,
    "prediction_norm": 6.932271957397461,
    "timestamp": "2025-07-14T12:34:52.068771"
  },
  {
    "epoch": 2173,
    "batch_idx": 160,
    "loss": 0.10112286359071732,
    "learning_rate": 1.8348379574103764e-05,
    "timestep_mean": 503.0625,
    "epsilon_norm": 7.066711902618408,
    "prediction_norm": 6.712032794952393,
    "timestamp": "2025-07-14T12:34:52.074067"
  },
  {
    "epoch": 2173,
    "batch_idx": 192,
    "loss": 0.3708447217941284,
    "learning_rate": 1.8348379574103764e-05,
    "timestep_mean": 486.75,
    "epsilon_norm": 4.47624397277832,
    "prediction_norm": 3.9703211784362793,
    "timestamp": "2025-07-14T12:34:52.079269"
  },
  {
    "epoch": 2174,
    "batch_idx": 0,
    "loss": 0.2288626879453659,
    "learning_rate": 1.8559786720706216e-05,
    "timestep_mean": 486.0625,
    "epsilon_norm": 7.157936096191406,
    "prediction_norm": 6.472746849060059,
    "timestamp": "2025-07-14T12:34:52.084446"
  },
  {
    "epoch": 2174,
    "batch_idx": 32,
    "loss": 0.17688851058483124,
    "learning_rate": 1.8559786720706216e-05,
    "timestep_mean": 529.71875,
    "epsilon_norm": 7.0631585121154785,
    "prediction_norm": 5.899823188781738,
    "timestamp": "2025-07-14T12:34:52.089704"
  },
  {
    "epoch": 2174,
    "batch_idx": 64,
    "loss": 0.27366048097610474,
    "learning_rate": 1.8559786720706216e-05,
    "timestep_mean": 566.96875,
    "epsilon_norm": 7.432652473449707,
    "prediction_norm": 6.311312198638916,
    "timestamp": "2025-07-14T12:34:52.094806"
  },
  {
    "epoch": 2174,
    "batch_idx": 96,
    "loss": 0.40988993644714355,
    "learning_rate": 1.8559786720706216e-05,
    "timestep_mean": 538.09375,
    "epsilon_norm": 7.352718830108643,
    "prediction_norm": 5.975734710693359,
    "timestamp": "2025-07-14T12:34:52.099747"
  },
  {
    "epoch": 2174,
    "batch_idx": 128,
    "loss": 0.23438584804534912,
    "learning_rate": 1.8559786720706216e-05,
    "timestep_mean": 560.375,
    "epsilon_norm": 8.897902488708496,
    "prediction_norm": 7.391615867614746,
    "timestamp": "2025-07-14T12:34:52.104899"
  },
  {
    "epoch": 2174,
    "batch_idx": 160,
    "loss": 0.1352299004793167,
    "learning_rate": 1.8559786720706216e-05,
    "timestep_mean": 614.78125,
    "epsilon_norm": 7.552880764007568,
    "prediction_norm": 7.345639705657959,
    "timestamp": "2025-07-14T12:34:52.110012"
  },
  {
    "epoch": 2174,
    "batch_idx": 192,
    "loss": 0.1896727830171585,
    "learning_rate": 1.8559786720706216e-05,
    "timestep_mean": 514.375,
    "epsilon_norm": 3.6288115978240967,
    "prediction_norm": 3.39898681640625,
    "timestamp": "2025-07-14T12:34:52.115125"
  },
  {
    "epoch": 2175,
    "batch_idx": 0,
    "loss": 0.4456954300403595,
    "learning_rate": 1.8772381773176395e-05,
    "timestep_mean": 523.4375,
    "epsilon_norm": 7.843331336975098,
    "prediction_norm": 5.898974418640137,
    "timestamp": "2025-07-14T12:34:52.120186"
  },
  {
    "epoch": 2175,
    "batch_idx": 32,
    "loss": 0.22204338014125824,
    "learning_rate": 1.8772381773176395e-05,
    "timestep_mean": 589.8125,
    "epsilon_norm": 7.827902317047119,
    "prediction_norm": 6.838617324829102,
    "timestamp": "2025-07-14T12:34:52.125166"
  },
  {
    "epoch": 2175,
    "batch_idx": 64,
    "loss": 0.37694746255874634,
    "learning_rate": 1.8772381773176395e-05,
    "timestep_mean": 469.625,
    "epsilon_norm": 8.176628112792969,
    "prediction_norm": 6.896281719207764,
    "timestamp": "2025-07-14T12:34:52.130187"
  },
  {
    "epoch": 2175,
    "batch_idx": 96,
    "loss": 0.33765822649002075,
    "learning_rate": 1.8772381773176395e-05,
    "timestep_mean": 496.5,
    "epsilon_norm": 8.044559478759766,
    "prediction_norm": 6.382333755493164,
    "timestamp": "2025-07-14T12:34:52.135038"
  },
  {
    "epoch": 2175,
    "batch_idx": 128,
    "loss": 0.24540764093399048,
    "learning_rate": 1.8772381773176395e-05,
    "timestep_mean": 496.71875,
    "epsilon_norm": 7.89316463470459,
    "prediction_norm": 6.9236602783203125,
    "timestamp": "2025-07-14T12:34:52.139850"
  },
  {
    "epoch": 2175,
    "batch_idx": 160,
    "loss": 0.379181444644928,
    "learning_rate": 1.8772381773176395e-05,
    "timestep_mean": 500.0625,
    "epsilon_norm": 8.90821361541748,
    "prediction_norm": 7.419593811035156,
    "timestamp": "2025-07-14T12:34:52.144701"
  },
  {
    "epoch": 2175,
    "batch_idx": 192,
    "loss": 0.268798828125,
    "learning_rate": 1.8772381773176395e-05,
    "timestep_mean": 483.125,
    "epsilon_norm": 3.986116886138916,
    "prediction_norm": 3.1682403087615967,
    "timestamp": "2025-07-14T12:34:52.149573"
  },
  {
    "epoch": 2176,
    "batch_idx": 0,
    "loss": 0.32924801111221313,
    "learning_rate": 1.898616420695707e-05,
    "timestep_mean": 519.28125,
    "epsilon_norm": 7.17288875579834,
    "prediction_norm": 5.704233169555664,
    "timestamp": "2025-07-14T12:34:52.154443"
  },
  {
    "epoch": 2176,
    "batch_idx": 32,
    "loss": 0.17759370803833008,
    "learning_rate": 1.898616420695707e-05,
    "timestep_mean": 516.34375,
    "epsilon_norm": 8.092864990234375,
    "prediction_norm": 7.757628917694092,
    "timestamp": "2025-07-14T12:34:52.159327"
  },
  {
    "epoch": 2176,
    "batch_idx": 64,
    "loss": 0.3236343562602997,
    "learning_rate": 1.898616420695707e-05,
    "timestep_mean": 479.71875,
    "epsilon_norm": 7.897994518280029,
    "prediction_norm": 6.372277736663818,
    "timestamp": "2025-07-14T12:34:52.164148"
  },
  {
    "epoch": 2176,
    "batch_idx": 96,
    "loss": 0.15905511379241943,
    "learning_rate": 1.898616420695707e-05,
    "timestep_mean": 548.6875,
    "epsilon_norm": 8.933649063110352,
    "prediction_norm": 7.87653923034668,
    "timestamp": "2025-07-14T12:34:52.169032"
  },
  {
    "epoch": 2176,
    "batch_idx": 128,
    "loss": 0.14174842834472656,
    "learning_rate": 1.898616420695707e-05,
    "timestep_mean": 592.78125,
    "epsilon_norm": 8.009371757507324,
    "prediction_norm": 7.817914009094238,
    "timestamp": "2025-07-14T12:34:52.173934"
  },
  {
    "epoch": 2176,
    "batch_idx": 160,
    "loss": 0.1661665439605713,
    "learning_rate": 1.898616420695707e-05,
    "timestep_mean": 496.3125,
    "epsilon_norm": 8.996539115905762,
    "prediction_norm": 7.924474239349365,
    "timestamp": "2025-07-14T12:34:52.178860"
  },
  {
    "epoch": 2176,
    "batch_idx": 192,
    "loss": 0.20768357813358307,
    "learning_rate": 1.898616420695707e-05,
    "timestep_mean": 501.625,
    "epsilon_norm": 4.094457149505615,
    "prediction_norm": 3.1715712547302246,
    "timestamp": "2025-07-14T12:34:52.183894"
  },
  {
    "epoch": 2177,
    "batch_idx": 0,
    "loss": 0.30133822560310364,
    "learning_rate": 1.920113349456141e-05,
    "timestep_mean": 490.5625,
    "epsilon_norm": 9.305891036987305,
    "prediction_norm": 8.056876182556152,
    "timestamp": "2025-07-14T12:34:52.188806"
  },
  {
    "epoch": 2177,
    "batch_idx": 32,
    "loss": 0.271940141916275,
    "learning_rate": 1.920113349456141e-05,
    "timestep_mean": 495.71875,
    "epsilon_norm": 7.563194751739502,
    "prediction_norm": 6.18908166885376,
    "timestamp": "2025-07-14T12:34:52.193679"
  },
  {
    "epoch": 2177,
    "batch_idx": 64,
    "loss": 0.16257572174072266,
    "learning_rate": 1.920113349456141e-05,
    "timestep_mean": 517.40625,
    "epsilon_norm": 8.489784240722656,
    "prediction_norm": 7.360576629638672,
    "timestamp": "2025-07-14T12:34:52.198597"
  },
  {
    "epoch": 2177,
    "batch_idx": 96,
    "loss": 0.2752782106399536,
    "learning_rate": 1.920113349456141e-05,
    "timestep_mean": 450.1875,
    "epsilon_norm": 8.780040740966797,
    "prediction_norm": 8.174510955810547,
    "timestamp": "2025-07-14T12:34:52.203640"
  },
  {
    "epoch": 2177,
    "batch_idx": 128,
    "loss": 0.24073290824890137,
    "learning_rate": 1.920113349456141e-05,
    "timestep_mean": 541.78125,
    "epsilon_norm": 8.75977897644043,
    "prediction_norm": 7.851526260375977,
    "timestamp": "2025-07-14T12:34:52.208571"
  },
  {
    "epoch": 2177,
    "batch_idx": 160,
    "loss": 0.260109007358551,
    "learning_rate": 1.920113349456141e-05,
    "timestep_mean": 487.15625,
    "epsilon_norm": 6.359407424926758,
    "prediction_norm": 5.119139194488525,
    "timestamp": "2025-07-14T12:34:52.213492"
  },
  {
    "epoch": 2177,
    "batch_idx": 192,
    "loss": 0.25820234417915344,
    "learning_rate": 1.920113349456141e-05,
    "timestep_mean": 418.125,
    "epsilon_norm": 4.019469261169434,
    "prediction_norm": 3.7332184314727783,
    "timestamp": "2025-07-14T12:34:52.218331"
  },
  {
    "epoch": 2178,
    "batch_idx": 0,
    "loss": 0.3716176152229309,
    "learning_rate": 1.941728910557404e-05,
    "timestep_mean": 437.15625,
    "epsilon_norm": 8.368897438049316,
    "prediction_norm": 6.531007766723633,
    "timestamp": "2025-07-14T12:34:52.223319"
  },
  {
    "epoch": 2178,
    "batch_idx": 32,
    "loss": 0.3467559218406677,
    "learning_rate": 1.941728910557404e-05,
    "timestep_mean": 485.125,
    "epsilon_norm": 8.598114967346191,
    "prediction_norm": 6.97335958480835,
    "timestamp": "2025-07-14T12:34:52.228176"
  },
  {
    "epoch": 2178,
    "batch_idx": 64,
    "loss": 0.4271639585494995,
    "learning_rate": 1.941728910557404e-05,
    "timestep_mean": 370.15625,
    "epsilon_norm": 8.397027015686035,
    "prediction_norm": 6.407516956329346,
    "timestamp": "2025-07-14T12:34:52.233061"
  },
  {
    "epoch": 2178,
    "batch_idx": 96,
    "loss": 0.23975825309753418,
    "learning_rate": 1.941728910557404e-05,
    "timestep_mean": 426.3125,
    "epsilon_norm": 7.027950286865234,
    "prediction_norm": 6.029178142547607,
    "timestamp": "2025-07-14T12:34:52.237984"
  },
  {
    "epoch": 2178,
    "batch_idx": 128,
    "loss": 0.15801939368247986,
    "learning_rate": 1.941728910557404e-05,
    "timestep_mean": 461.84375,
    "epsilon_norm": 6.976150989532471,
    "prediction_norm": 6.754568099975586,
    "timestamp": "2025-07-14T12:34:52.242833"
  },
  {
    "epoch": 2178,
    "batch_idx": 160,
    "loss": 0.1706281453371048,
    "learning_rate": 1.941728910557404e-05,
    "timestep_mean": 527.5,
    "epsilon_norm": 8.248348236083984,
    "prediction_norm": 7.438488960266113,
    "timestamp": "2025-07-14T12:34:52.247703"
  },
  {
    "epoch": 2178,
    "batch_idx": 192,
    "loss": 0.08339321613311768,
    "learning_rate": 1.941728910557404e-05,
    "timestep_mean": 758.0,
    "epsilon_norm": 3.120605230331421,
    "prediction_norm": 3.084135055541992,
    "timestamp": "2025-07-14T12:34:52.252594"
  },
  {
    "epoch": 2179,
    "batch_idx": 0,
    "loss": 0.24046944081783295,
    "learning_rate": 1.9634630506652468e-05,
    "timestep_mean": 473.3125,
    "epsilon_norm": 7.999782085418701,
    "prediction_norm": 7.111046314239502,
    "timestamp": "2025-07-14T12:34:52.257704"
  },
  {
    "epoch": 2179,
    "batch_idx": 32,
    "loss": 0.26984208822250366,
    "learning_rate": 1.9634630506652468e-05,
    "timestep_mean": 533.40625,
    "epsilon_norm": 7.5159711837768555,
    "prediction_norm": 6.983894348144531,
    "timestamp": "2025-07-14T12:34:52.262877"
  },
  {
    "epoch": 2179,
    "batch_idx": 64,
    "loss": 0.16315320134162903,
    "learning_rate": 1.9634630506652468e-05,
    "timestep_mean": 446.90625,
    "epsilon_norm": 7.92042875289917,
    "prediction_norm": 6.81173849105835,
    "timestamp": "2025-07-14T12:34:52.267969"
  },
  {
    "epoch": 2179,
    "batch_idx": 96,
    "loss": 0.24407222867012024,
    "learning_rate": 1.9634630506652468e-05,
    "timestep_mean": 502.0625,
    "epsilon_norm": 6.866214752197266,
    "prediction_norm": 5.887734889984131,
    "timestamp": "2025-07-14T12:34:52.273172"
  },
  {
    "epoch": 2179,
    "batch_idx": 128,
    "loss": 0.45304861664772034,
    "learning_rate": 1.9634630506652468e-05,
    "timestep_mean": 469.09375,
    "epsilon_norm": 8.945252418518066,
    "prediction_norm": 7.062152862548828,
    "timestamp": "2025-07-14T12:34:52.278448"
  },
  {
    "epoch": 2179,
    "batch_idx": 160,
    "loss": 0.4255104064941406,
    "learning_rate": 1.9634630506652468e-05,
    "timestep_mean": 498.4375,
    "epsilon_norm": 9.95467472076416,
    "prediction_norm": 8.102116584777832,
    "timestamp": "2025-07-14T12:34:52.283777"
  },
  {
    "epoch": 2179,
    "batch_idx": 192,
    "loss": 0.14857327938079834,
    "learning_rate": 1.9634630506652468e-05,
    "timestep_mean": 563.5,
    "epsilon_norm": 4.4078803062438965,
    "prediction_norm": 3.8186912536621094,
    "timestamp": "2025-07-14T12:34:52.289012"
  },
  {
    "epoch": 2180,
    "batch_idx": 0,
    "loss": 0.2816099524497986,
    "learning_rate": 1.9853157161528506e-05,
    "timestep_mean": 474.6875,
    "epsilon_norm": 7.4126362800598145,
    "prediction_norm": 6.361053943634033,
    "timestamp": "2025-07-14T12:34:52.294214"
  },
  {
    "epoch": 2180,
    "batch_idx": 32,
    "loss": 0.24588511884212494,
    "learning_rate": 1.9853157161528506e-05,
    "timestep_mean": 515.84375,
    "epsilon_norm": 7.834414958953857,
    "prediction_norm": 6.858976364135742,
    "timestamp": "2025-07-14T12:34:52.299388"
  },
  {
    "epoch": 2180,
    "batch_idx": 64,
    "loss": 0.3328118622303009,
    "learning_rate": 1.9853157161528506e-05,
    "timestep_mean": 562.21875,
    "epsilon_norm": 8.202354431152344,
    "prediction_norm": 6.7922043800354,
    "timestamp": "2025-07-14T12:34:52.304659"
  },
  {
    "epoch": 2180,
    "batch_idx": 96,
    "loss": 0.20582249760627747,
    "learning_rate": 1.9853157161528506e-05,
    "timestep_mean": 541.46875,
    "epsilon_norm": 7.118279933929443,
    "prediction_norm": 5.774844646453857,
    "timestamp": "2025-07-14T12:34:52.309877"
  },
  {
    "epoch": 2180,
    "batch_idx": 128,
    "loss": 0.2633928060531616,
    "learning_rate": 1.9853157161528506e-05,
    "timestep_mean": 544.15625,
    "epsilon_norm": 6.72363805770874,
    "prediction_norm": 5.920017242431641,
    "timestamp": "2025-07-14T12:34:52.315157"
  },
  {
    "epoch": 2180,
    "batch_idx": 160,
    "loss": 0.1866278350353241,
    "learning_rate": 1.9853157161528506e-05,
    "timestep_mean": 496.75,
    "epsilon_norm": 7.468534469604492,
    "prediction_norm": 6.780176162719727,
    "timestamp": "2025-07-14T12:34:52.320440"
  },
  {
    "epoch": 2180,
    "batch_idx": 192,
    "loss": 0.05566026270389557,
    "learning_rate": 1.9853157161528506e-05,
    "timestep_mean": 644.5,
    "epsilon_norm": 3.7944867610931396,
    "prediction_norm": 4.028587341308594,
    "timestamp": "2025-07-14T12:34:52.325718"
  },
  {
    "epoch": 2181,
    "batch_idx": 0,
    "loss": 0.42559656500816345,
    "learning_rate": 2.0072868531009188e-05,
    "timestep_mean": 542.0625,
    "epsilon_norm": 8.505624771118164,
    "prediction_norm": 7.074125289916992,
    "timestamp": "2025-07-14T12:34:52.330876"
  },
  {
    "epoch": 2181,
    "batch_idx": 32,
    "loss": 0.11671706289052963,
    "learning_rate": 2.0072868531009188e-05,
    "timestep_mean": 506.875,
    "epsilon_norm": 7.88755989074707,
    "prediction_norm": 6.580897331237793,
    "timestamp": "2025-07-14T12:34:52.336118"
  },
  {
    "epoch": 2181,
    "batch_idx": 64,
    "loss": 0.1989613175392151,
    "learning_rate": 2.0072868531009188e-05,
    "timestep_mean": 542.21875,
    "epsilon_norm": 6.894886493682861,
    "prediction_norm": 6.437304973602295,
    "timestamp": "2025-07-14T12:34:52.341266"
  },
  {
    "epoch": 2181,
    "batch_idx": 96,
    "loss": 0.4974159598350525,
    "learning_rate": 2.0072868531009188e-05,
    "timestep_mean": 409.09375,
    "epsilon_norm": 8.289423942565918,
    "prediction_norm": 6.195432662963867,
    "timestamp": "2025-07-14T12:34:52.346696"
  },
  {
    "epoch": 2181,
    "batch_idx": 128,
    "loss": 0.19648948311805725,
    "learning_rate": 2.0072868531009188e-05,
    "timestep_mean": 488.78125,
    "epsilon_norm": 8.773835182189941,
    "prediction_norm": 7.325461387634277,
    "timestamp": "2025-07-14T12:34:52.351907"
  },
  {
    "epoch": 2181,
    "batch_idx": 160,
    "loss": 0.4044589102268219,
    "learning_rate": 2.0072868531009188e-05,
    "timestep_mean": 505.1875,
    "epsilon_norm": 9.260281562805176,
    "prediction_norm": 7.5266642570495605,
    "timestamp": "2025-07-14T12:34:52.357081"
  },
  {
    "epoch": 2181,
    "batch_idx": 192,
    "loss": 0.30420711636543274,
    "learning_rate": 2.0072868531009188e-05,
    "timestep_mean": 338.75,
    "epsilon_norm": 4.2108988761901855,
    "prediction_norm": 3.3469886779785156,
    "timestamp": "2025-07-14T12:34:52.362184"
  },
  {
    "epoch": 2182,
    "batch_idx": 0,
    "loss": 0.16677439212799072,
    "learning_rate": 2.0293764072978604e-05,
    "timestep_mean": 560.9375,
    "epsilon_norm": 7.906322002410889,
    "prediction_norm": 7.244874954223633,
    "timestamp": "2025-07-14T12:34:52.367607"
  },
  {
    "epoch": 2182,
    "batch_idx": 32,
    "loss": 0.18604379892349243,
    "learning_rate": 2.0293764072978604e-05,
    "timestep_mean": 542.8125,
    "epsilon_norm": 7.423675537109375,
    "prediction_norm": 6.053855895996094,
    "timestamp": "2025-07-14T12:34:52.372822"
  },
  {
    "epoch": 2182,
    "batch_idx": 64,
    "loss": 0.15977710485458374,
    "learning_rate": 2.0293764072978604e-05,
    "timestep_mean": 487.90625,
    "epsilon_norm": 7.1892499923706055,
    "prediction_norm": 6.395656108856201,
    "timestamp": "2025-07-14T12:34:52.378010"
  },
  {
    "epoch": 2182,
    "batch_idx": 96,
    "loss": 0.2864280641078949,
    "learning_rate": 2.0293764072978604e-05,
    "timestep_mean": 526.875,
    "epsilon_norm": 8.37593936920166,
    "prediction_norm": 7.29725980758667,
    "timestamp": "2025-07-14T12:34:52.383257"
  },
  {
    "epoch": 2182,
    "batch_idx": 128,
    "loss": 0.13657565414905548,
    "learning_rate": 2.0293764072978604e-05,
    "timestep_mean": 482.75,
    "epsilon_norm": 7.649993896484375,
    "prediction_norm": 6.744636535644531,
    "timestamp": "2025-07-14T12:34:52.388627"
  },
  {
    "epoch": 2182,
    "batch_idx": 160,
    "loss": 0.19222049415111542,
    "learning_rate": 2.0293764072978604e-05,
    "timestep_mean": 512.96875,
    "epsilon_norm": 7.707124710083008,
    "prediction_norm": 6.7124247550964355,
    "timestamp": "2025-07-14T12:34:52.393937"
  },
  {
    "epoch": 2182,
    "batch_idx": 192,
    "loss": 0.1608370542526245,
    "learning_rate": 2.0293764072978604e-05,
    "timestep_mean": 473.125,
    "epsilon_norm": 3.004455327987671,
    "prediction_norm": 3.0277867317199707,
    "timestamp": "2025-07-14T12:34:52.399128"
  },
  {
    "epoch": 2183,
    "batch_idx": 0,
    "loss": 0.23045086860656738,
    "learning_rate": 2.0515843242398954e-05,
    "timestep_mean": 525.375,
    "epsilon_norm": 7.888483047485352,
    "prediction_norm": 7.301064491271973,
    "timestamp": "2025-07-14T12:34:52.404063"
  },
  {
    "epoch": 2183,
    "batch_idx": 32,
    "loss": 0.25251272320747375,
    "learning_rate": 2.0515843242398954e-05,
    "timestep_mean": 479.4375,
    "epsilon_norm": 8.409499168395996,
    "prediction_norm": 6.788095474243164,
    "timestamp": "2025-07-14T12:34:52.408996"
  },
  {
    "epoch": 2183,
    "batch_idx": 64,
    "loss": 0.11880958080291748,
    "learning_rate": 2.0515843242398954e-05,
    "timestep_mean": 464.375,
    "epsilon_norm": 7.208565711975098,
    "prediction_norm": 6.924386024475098,
    "timestamp": "2025-07-14T12:34:52.413893"
  },
  {
    "epoch": 2183,
    "batch_idx": 96,
    "loss": 0.3894112706184387,
    "learning_rate": 2.0515843242398954e-05,
    "timestep_mean": 489.65625,
    "epsilon_norm": 8.82320499420166,
    "prediction_norm": 7.151989459991455,
    "timestamp": "2025-07-14T12:34:52.419249"
  },
  {
    "epoch": 2183,
    "batch_idx": 128,
    "loss": 0.2929450571537018,
    "learning_rate": 2.0515843242398954e-05,
    "timestep_mean": 509.78125,
    "epsilon_norm": 8.050325393676758,
    "prediction_norm": 7.056917667388916,
    "timestamp": "2025-07-14T12:34:52.424516"
  },
  {
    "epoch": 2183,
    "batch_idx": 160,
    "loss": 0.3419138789176941,
    "learning_rate": 2.0515843242398954e-05,
    "timestep_mean": 510.4375,
    "epsilon_norm": 8.66558837890625,
    "prediction_norm": 6.819061279296875,
    "timestamp": "2025-07-14T12:34:52.429840"
  },
  {
    "epoch": 2183,
    "batch_idx": 192,
    "loss": 0.4072505235671997,
    "learning_rate": 2.0515843242398954e-05,
    "timestep_mean": 587.25,
    "epsilon_norm": 5.16084098815918,
    "prediction_norm": 3.9404687881469727,
    "timestamp": "2025-07-14T12:34:52.435066"
  },
  {
    "epoch": 2184,
    "batch_idx": 0,
    "loss": 0.2573315501213074,
    "learning_rate": 2.0739105491312126e-05,
    "timestep_mean": 552.40625,
    "epsilon_norm": 7.612368583679199,
    "prediction_norm": 6.608809471130371,
    "timestamp": "2025-07-14T12:34:52.440377"
  },
  {
    "epoch": 2184,
    "batch_idx": 32,
    "loss": 0.11437202990055084,
    "learning_rate": 2.0739105491312126e-05,
    "timestep_mean": 684.78125,
    "epsilon_norm": 7.675146102905273,
    "prediction_norm": 7.381552219390869,
    "timestamp": "2025-07-14T12:34:52.445611"
  },
  {
    "epoch": 2184,
    "batch_idx": 64,
    "loss": 0.16420310735702515,
    "learning_rate": 2.0739105491312126e-05,
    "timestep_mean": 584.40625,
    "epsilon_norm": 8.817402839660645,
    "prediction_norm": 8.12049388885498,
    "timestamp": "2025-07-14T12:34:52.450930"
  },
  {
    "epoch": 2184,
    "batch_idx": 96,
    "loss": 0.2993697226047516,
    "learning_rate": 2.0739105491312126e-05,
    "timestep_mean": 424.46875,
    "epsilon_norm": 8.019493103027344,
    "prediction_norm": 6.978116989135742,
    "timestamp": "2025-07-14T12:34:52.456277"
  },
  {
    "epoch": 2184,
    "batch_idx": 128,
    "loss": 0.23860204219818115,
    "learning_rate": 2.0739105491312126e-05,
    "timestep_mean": 441.6875,
    "epsilon_norm": 7.0433220863342285,
    "prediction_norm": 6.1255974769592285,
    "timestamp": "2025-07-14T12:34:52.461453"
  },
  {
    "epoch": 2184,
    "batch_idx": 160,
    "loss": 0.40946346521377563,
    "learning_rate": 2.0739105491312126e-05,
    "timestep_mean": 464.0625,
    "epsilon_norm": 7.353335380554199,
    "prediction_norm": 5.837271690368652,
    "timestamp": "2025-07-14T12:34:52.466883"
  },
  {
    "epoch": 2184,
    "batch_idx": 192,
    "loss": 0.22509172558784485,
    "learning_rate": 2.0739105491312126e-05,
    "timestep_mean": 588.625,
    "epsilon_norm": 4.4679436683654785,
    "prediction_norm": 3.471466541290283,
    "timestamp": "2025-07-14T12:34:52.472085"
  },
  {
    "epoch": 2185,
    "batch_idx": 0,
    "loss": 0.31972846388816833,
    "learning_rate": 2.096355026884038e-05,
    "timestep_mean": 458.5625,
    "epsilon_norm": 8.131094932556152,
    "prediction_norm": 6.340129375457764,
    "timestamp": "2025-07-14T12:34:52.477386"
  },
  {
    "epoch": 2185,
    "batch_idx": 32,
    "loss": 0.3197355568408966,
    "learning_rate": 2.096355026884038e-05,
    "timestep_mean": 480.03125,
    "epsilon_norm": 8.014897346496582,
    "prediction_norm": 6.962751388549805,
    "timestamp": "2025-07-14T12:34:52.482643"
  },
  {
    "epoch": 2185,
    "batch_idx": 64,
    "loss": 0.27535414695739746,
    "learning_rate": 2.096355026884038e-05,
    "timestep_mean": 512.1875,
    "epsilon_norm": 7.941730976104736,
    "prediction_norm": 6.539032936096191,
    "timestamp": "2025-07-14T12:34:52.487887"
  },
  {
    "epoch": 2185,
    "batch_idx": 96,
    "loss": 0.20812202990055084,
    "learning_rate": 2.096355026884038e-05,
    "timestep_mean": 474.40625,
    "epsilon_norm": 7.669785499572754,
    "prediction_norm": 6.246921062469482,
    "timestamp": "2025-07-14T12:34:52.493163"
  },
  {
    "epoch": 2185,
    "batch_idx": 128,
    "loss": 0.2152085304260254,
    "learning_rate": 2.096355026884038e-05,
    "timestep_mean": 521.53125,
    "epsilon_norm": 9.075818061828613,
    "prediction_norm": 8.069375038146973,
    "timestamp": "2025-07-14T12:34:52.498402"
  },
  {
    "epoch": 2185,
    "batch_idx": 160,
    "loss": 0.15894779562950134,
    "learning_rate": 2.096355026884038e-05,
    "timestep_mean": 477.125,
    "epsilon_norm": 6.498726844787598,
    "prediction_norm": 6.135801315307617,
    "timestamp": "2025-07-14T12:34:52.503639"
  },
  {
    "epoch": 2185,
    "batch_idx": 192,
    "loss": 0.20767803490161896,
    "learning_rate": 2.096355026884038e-05,
    "timestep_mean": 573.875,
    "epsilon_norm": 5.126979827880859,
    "prediction_norm": 4.43813943862915,
    "timestamp": "2025-07-14T12:34:52.508820"
  },
  {
    "epoch": 2186,
    "batch_idx": 0,
    "loss": 0.2719791829586029,
    "learning_rate": 2.1189177021188933e-05,
    "timestep_mean": 448.9375,
    "epsilon_norm": 7.761468410491943,
    "prediction_norm": 5.908076763153076,
    "timestamp": "2025-07-14T12:34:52.514073"
  },
  {
    "epoch": 2186,
    "batch_idx": 32,
    "loss": 0.14141865074634552,
    "learning_rate": 2.1189177021188933e-05,
    "timestep_mean": 488.625,
    "epsilon_norm": 7.059420108795166,
    "prediction_norm": 6.43751859664917,
    "timestamp": "2025-07-14T12:34:52.519228"
  },
  {
    "epoch": 2186,
    "batch_idx": 64,
    "loss": 0.19223347306251526,
    "learning_rate": 2.1189177021188933e-05,
    "timestep_mean": 502.96875,
    "epsilon_norm": 7.2246623039245605,
    "prediction_norm": 6.745064735412598,
    "timestamp": "2025-07-14T12:34:52.524457"
  },
  {
    "epoch": 2186,
    "batch_idx": 96,
    "loss": 0.10568632930517197,
    "learning_rate": 2.1189177021188933e-05,
    "timestep_mean": 527.78125,
    "epsilon_norm": 7.02882194519043,
    "prediction_norm": 6.698810577392578,
    "timestamp": "2025-07-14T12:34:52.529594"
  },
  {
    "epoch": 2186,
    "batch_idx": 128,
    "loss": 0.20640994608402252,
    "learning_rate": 2.1189177021188933e-05,
    "timestep_mean": 394.625,
    "epsilon_norm": 7.689428806304932,
    "prediction_norm": 6.673061370849609,
    "timestamp": "2025-07-14T12:34:52.534764"
  },
  {
    "epoch": 2186,
    "batch_idx": 160,
    "loss": 0.3680320084095001,
    "learning_rate": 2.1189177021188933e-05,
    "timestep_mean": 452.125,
    "epsilon_norm": 8.861320495605469,
    "prediction_norm": 6.968521595001221,
    "timestamp": "2025-07-14T12:34:52.539951"
  },
  {
    "epoch": 2186,
    "batch_idx": 192,
    "loss": 0.10783340036869049,
    "learning_rate": 2.1189177021188933e-05,
    "timestep_mean": 592.0,
    "epsilon_norm": 4.030434608459473,
    "prediction_norm": 4.165464401245117,
    "timestamp": "2025-07-14T12:34:52.545094"
  },
  {
    "epoch": 2187,
    "batch_idx": 0,
    "loss": 0.25508126616477966,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 504.15625,
    "epsilon_norm": 9.219528198242188,
    "prediction_norm": 8.245926856994629,
    "timestamp": "2025-07-14T12:34:52.550506"
  },
  {
    "epoch": 2187,
    "batch_idx": 32,
    "loss": 0.11720538884401321,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 598.09375,
    "epsilon_norm": 7.480005264282227,
    "prediction_norm": 6.7502827644348145,
    "timestamp": "2025-07-14T12:34:52.555714"
  },
  {
    "epoch": 2187,
    "batch_idx": 64,
    "loss": 0.22666245698928833,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 532.21875,
    "epsilon_norm": 9.265141487121582,
    "prediction_norm": 8.159310340881348,
    "timestamp": "2025-07-14T12:34:52.560897"
  },
  {
    "epoch": 2187,
    "batch_idx": 96,
    "loss": 0.2963740825653076,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 497.5,
    "epsilon_norm": 7.204514503479004,
    "prediction_norm": 6.135149955749512,
    "timestamp": "2025-07-14T12:34:52.566109"
  },
  {
    "epoch": 2187,
    "batch_idx": 128,
    "loss": 0.3687809705734253,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 496.75,
    "epsilon_norm": 8.42252254486084,
    "prediction_norm": 7.139288425445557,
    "timestamp": "2025-07-14T12:34:52.571296"
  },
  {
    "epoch": 2187,
    "batch_idx": 160,
    "loss": 0.22980770468711853,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 508.4375,
    "epsilon_norm": 7.902979850769043,
    "prediction_norm": 6.887965679168701,
    "timestamp": "2025-07-14T12:34:52.576532"
  },
  {
    "epoch": 2187,
    "batch_idx": 192,
    "loss": 0.27348005771636963,
    "learning_rate": 2.141598519164594e-05,
    "timestep_mean": 504.125,
    "epsilon_norm": 3.0207440853118896,
    "prediction_norm": 2.310088872909546,
    "timestamp": "2025-07-14T12:34:52.581753"
  },
  {
    "epoch": 2188,
    "batch_idx": 0,
    "loss": 0.17269986867904663,
    "learning_rate": 2.1643974220584774e-05,
    "timestep_mean": 511.0625,
    "epsilon_norm": 6.939423561096191,
    "prediction_norm": 5.741427898406982,
    "timestamp": "2025-07-14T12:34:52.586991"
  },
  {
    "epoch": 2188,
    "batch_idx": 32,
    "loss": 0.17599397897720337,
    "learning_rate": 2.1643974220584774e-05,
    "timestep_mean": 523.46875,
    "epsilon_norm": 7.675107002258301,
    "prediction_norm": 7.101795673370361,
    "timestamp": "2025-07-14T12:34:52.592407"
  },
  {
    "epoch": 2188,
    "batch_idx": 64,
    "loss": 0.14010709524154663,
    "learning_rate": 2.1643974220584774e-05,
    "timestep_mean": 501.21875,
    "epsilon_norm": 8.452479362487793,
    "prediction_norm": 7.521107196807861,
    "timestamp": "2025-07-14T12:34:52.597629"
  },
  {
    "epoch": 2188,
    "batch_idx": 96,
    "loss": 0.4158724546432495,
    "learning_rate": 2.1643974220584774e-05,
    "timestep_mean": 466.625,
    "epsilon_norm": 7.526451110839844,
    "prediction_norm": 6.119387149810791,
    "timestamp": "2025-07-14T12:34:52.602781"
  },
  {
    "epoch": 2188,
    "batch_idx": 128,
    "loss": 0.29652273654937744,
    "learning_rate": 2.1643974220584774e-05,
    "timestep_mean": 454.9375,
    "epsilon_norm": 8.239230155944824,
    "prediction_norm": 6.750655651092529,
    "timestamp": "2025-07-14T12:34:52.608015"
  },
  {
    "epoch": 2188,
    "batch_idx": 160,
    "loss": 0.26549312472343445,
    "learning_rate": 2.1643974220584774e-05,
    "timestep_mean": 488.90625,
    "epsilon_norm": 9.122186660766602,
    "prediction_norm": 7.409173965454102,
    "timestamp": "2025-07-14T12:34:52.613245"
  },
  {
    "epoch": 2188,
    "batch_idx": 192,
    "loss": 0.20948752760887146,
    "learning_rate": 2.1643974220584774e-05,
    "timestep_mean": 458.625,
    "epsilon_norm": 2.8838775157928467,
    "prediction_norm": 2.544539451599121,
    "timestamp": "2025-07-14T12:34:52.618338"
  },
  {
    "epoch": 2189,
    "batch_idx": 0,
    "loss": 0.1811600923538208,
    "learning_rate": 2.187314354546537e-05,
    "timestep_mean": 523.03125,
    "epsilon_norm": 8.438946723937988,
    "prediction_norm": 7.694382667541504,
    "timestamp": "2025-07-14T12:34:52.623447"
  },
  {
    "epoch": 2189,
    "batch_idx": 32,
    "loss": 0.25758254528045654,
    "learning_rate": 2.187314354546537e-05,
    "timestep_mean": 502.59375,
    "epsilon_norm": 8.442204475402832,
    "prediction_norm": 6.702645301818848,
    "timestamp": "2025-07-14T12:34:52.628708"
  },
  {
    "epoch": 2189,
    "batch_idx": 64,
    "loss": 0.31396299600601196,
    "learning_rate": 2.187314354546537e-05,
    "timestep_mean": 470.15625,
    "epsilon_norm": 7.861839771270752,
    "prediction_norm": 6.234204292297363,
    "timestamp": "2025-07-14T12:34:52.634023"
  },
  {
    "epoch": 2189,
    "batch_idx": 96,
    "loss": 0.3758126199245453,
    "learning_rate": 2.187314354546537e-05,
    "timestep_mean": 445.125,
    "epsilon_norm": 7.2144341468811035,
    "prediction_norm": 6.036584377288818,
    "timestamp": "2025-07-14T12:34:52.639227"
  },
  {
    "epoch": 2189,
    "batch_idx": 128,
    "loss": 0.2821516990661621,
    "learning_rate": 2.187314354546537e-05,
    "timestep_mean": 498.1875,
    "epsilon_norm": 7.840387344360352,
    "prediction_norm": 6.643072128295898,
    "timestamp": "2025-07-14T12:34:52.644419"
  },
  {
    "epoch": 2189,
    "batch_idx": 160,
    "loss": 0.19200465083122253,
    "learning_rate": 2.187314354546537e-05,
    "timestep_mean": 520.90625,
    "epsilon_norm": 8.36264705657959,
    "prediction_norm": 8.160398483276367,
    "timestamp": "2025-07-14T12:34:52.649697"
  },
  {
    "epoch": 2189,
    "batch_idx": 192,
    "loss": 0.24037614464759827,
    "learning_rate": 2.187314354546537e-05,
    "timestep_mean": 644.75,
    "epsilon_norm": 3.8663744926452637,
    "prediction_norm": 2.990659713745117,
    "timestamp": "2025-07-14T12:34:52.655017"
  },
  {
    "epoch": 2190,
    "batch_idx": 0,
    "loss": 0.4472411274909973,
    "learning_rate": 2.210349260083493e-05,
    "timestep_mean": 549.125,
    "epsilon_norm": 8.548251152038574,
    "prediction_norm": 6.698482036590576,
    "timestamp": "2025-07-14T12:34:52.660260"
  },
  {
    "epoch": 2190,
    "batch_idx": 32,
    "loss": 0.21983729302883148,
    "learning_rate": 2.210349260083493e-05,
    "timestep_mean": 561.78125,
    "epsilon_norm": 7.411257743835449,
    "prediction_norm": 5.908211708068848,
    "timestamp": "2025-07-14T12:34:52.665505"
  },
  {
    "epoch": 2190,
    "batch_idx": 64,
    "loss": 0.2989407777786255,
    "learning_rate": 2.210349260083493e-05,
    "timestep_mean": 414.65625,
    "epsilon_norm": 6.520695686340332,
    "prediction_norm": 4.5943498611450195,
    "timestamp": "2025-07-14T12:34:52.670577"
  },
  {
    "epoch": 2190,
    "batch_idx": 96,
    "loss": 0.2131771445274353,
    "learning_rate": 2.210349260083493e-05,
    "timestep_mean": 584.375,
    "epsilon_norm": 8.020705223083496,
    "prediction_norm": 6.977303504943848,
    "timestamp": "2025-07-14T12:34:52.675552"
  },
  {
    "epoch": 2190,
    "batch_idx": 128,
    "loss": 0.27429914474487305,
    "learning_rate": 2.210349260083493e-05,
    "timestep_mean": 512.78125,
    "epsilon_norm": 8.007208824157715,
    "prediction_norm": 6.593364238739014,
    "timestamp": "2025-07-14T12:34:52.680382"
  },
  {
    "epoch": 2190,
    "batch_idx": 160,
    "loss": 0.21279853582382202,
    "learning_rate": 2.210349260083493e-05,
    "timestep_mean": 534.53125,
    "epsilon_norm": 8.59683895111084,
    "prediction_norm": 7.688830852508545,
    "timestamp": "2025-07-14T12:34:52.685286"
  },
  {
    "epoch": 2190,
    "batch_idx": 192,
    "loss": 0.14982742071151733,
    "learning_rate": 2.210349260083493e-05,
    "timestep_mean": 532.75,
    "epsilon_norm": 4.656580448150635,
    "prediction_norm": 4.292455196380615,
    "timestamp": "2025-07-14T12:34:52.690150"
  },
  {
    "epoch": 2191,
    "batch_idx": 0,
    "loss": 0.24142692983150482,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 524.3125,
    "epsilon_norm": 8.813408851623535,
    "prediction_norm": 7.58466911315918,
    "timestamp": "2025-07-14T12:34:52.695123"
  },
  {
    "epoch": 2191,
    "batch_idx": 32,
    "loss": 0.26718056201934814,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 513.65625,
    "epsilon_norm": 8.474478721618652,
    "prediction_norm": 7.215795993804932,
    "timestamp": "2025-07-14T12:34:52.700473"
  },
  {
    "epoch": 2191,
    "batch_idx": 64,
    "loss": 0.16944390535354614,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 462.125,
    "epsilon_norm": 6.993288993835449,
    "prediction_norm": 6.595801830291748,
    "timestamp": "2025-07-14T12:34:52.705676"
  },
  {
    "epoch": 2191,
    "batch_idx": 96,
    "loss": 0.201607346534729,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 491.125,
    "epsilon_norm": 7.19844388961792,
    "prediction_norm": 6.040660858154297,
    "timestamp": "2025-07-14T12:34:52.710830"
  },
  {
    "epoch": 2191,
    "batch_idx": 128,
    "loss": 0.21185074746608734,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 493.75,
    "epsilon_norm": 6.938266754150391,
    "prediction_norm": 6.083101272583008,
    "timestamp": "2025-07-14T12:34:52.715827"
  },
  {
    "epoch": 2191,
    "batch_idx": 160,
    "loss": 0.09588924050331116,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 597.34375,
    "epsilon_norm": 7.742969036102295,
    "prediction_norm": 7.930240154266357,
    "timestamp": "2025-07-14T12:34:52.720906"
  },
  {
    "epoch": 2191,
    "batch_idx": 192,
    "loss": 0.4644085764884949,
    "learning_rate": 2.2335020818330325e-05,
    "timestep_mean": 510.375,
    "epsilon_norm": 4.320060729980469,
    "prediction_norm": 3.636561870574951,
    "timestamp": "2025-07-14T12:34:52.726076"
  },
  {
    "epoch": 2192,
    "batch_idx": 0,
    "loss": 0.13410687446594238,
    "learning_rate": 2.2567727626678574e-05,
    "timestep_mean": 581.40625,
    "epsilon_norm": 6.574648857116699,
    "prediction_norm": 5.686865329742432,
    "timestamp": "2025-07-14T12:34:52.731278"
  },
  {
    "epoch": 2192,
    "batch_idx": 32,
    "loss": 0.16801658272743225,
    "learning_rate": 2.2567727626678574e-05,
    "timestep_mean": 582.28125,
    "epsilon_norm": 7.6502461433410645,
    "prediction_norm": 7.096632480621338,
    "timestamp": "2025-07-14T12:34:52.736213"
  },
  {
    "epoch": 2192,
    "batch_idx": 64,
    "loss": 0.29496634006500244,
    "learning_rate": 2.2567727626678574e-05,
    "timestep_mean": 472.6875,
    "epsilon_norm": 8.017606735229492,
    "prediction_norm": 7.2989654541015625,
    "timestamp": "2025-07-14T12:34:52.741259"
  },
  {
    "epoch": 2192,
    "batch_idx": 96,
    "loss": 0.36154791712760925,
    "learning_rate": 2.2567727626678574e-05,
    "timestep_mean": 468.25,
    "epsilon_norm": 9.012605667114258,
    "prediction_norm": 7.522678375244141,
    "timestamp": "2025-07-14T12:34:52.746483"
  },
  {
    "epoch": 2192,
    "batch_idx": 128,
    "loss": 0.20184572041034698,
    "learning_rate": 2.2567727626678574e-05,
    "timestep_mean": 506.65625,
    "epsilon_norm": 7.9696197509765625,
    "prediction_norm": 6.583491802215576,
    "timestamp": "2025-07-14T12:34:52.751748"
  },
  {
    "epoch": 2192,
    "batch_idx": 160,
    "loss": 0.15808013081550598,
    "learning_rate": 2.2567727626678574e-05,
    "timestep_mean": 551.34375,
    "epsilon_norm": 7.775021076202393,
    "prediction_norm": 7.299780368804932,
    "timestamp": "2025-07-14T12:34:52.756995"
  },
  {
    "epoch": 2192,
    "batch_idx": 192,
    "loss": 0.1802867352962494,
    "learning_rate": 2.2567727626678574e-05,
    "timestep_mean": 439.875,
    "epsilon_norm": 4.4849653244018555,
    "prediction_norm": 3.986362934112549,
    "timestamp": "2025-07-14T12:34:52.762065"
  },
  {
    "epoch": 2193,
    "batch_idx": 0,
    "loss": 0.31488728523254395,
    "learning_rate": 2.2801612451698698e-05,
    "timestep_mean": 544.34375,
    "epsilon_norm": 8.964825630187988,
    "prediction_norm": 7.152093410491943,
    "timestamp": "2025-07-14T12:34:52.767338"
  },
  {
    "epoch": 2193,
    "batch_idx": 32,
    "loss": 0.11891355365514755,
    "learning_rate": 2.2801612451698698e-05,
    "timestep_mean": 603.03125,
    "epsilon_norm": 6.949802398681641,
    "prediction_norm": 6.547406196594238,
    "timestamp": "2025-07-14T12:34:52.772554"
  },
  {
    "epoch": 2193,
    "batch_idx": 64,
    "loss": 0.2587519586086273,
    "learning_rate": 2.2801612451698698e-05,
    "timestep_mean": 406.6875,
    "epsilon_norm": 7.307175636291504,
    "prediction_norm": 5.356732368469238,
    "timestamp": "2025-07-14T12:34:52.777706"
  },
  {
    "epoch": 2193,
    "batch_idx": 96,
    "loss": 0.19628700613975525,
    "learning_rate": 2.2801612451698698e-05,
    "timestep_mean": 514.4375,
    "epsilon_norm": 8.728967666625977,
    "prediction_norm": 7.639105796813965,
    "timestamp": "2025-07-14T12:34:52.782861"
  },
  {
    "epoch": 2193,
    "batch_idx": 128,
    "loss": 0.5207676291465759,
    "learning_rate": 2.2801612451698698e-05,
    "timestep_mean": 423.0,
    "epsilon_norm": 7.849643230438232,
    "prediction_norm": 5.613712310791016,
    "timestamp": "2025-07-14T12:34:52.788113"
  },
  {
    "epoch": 2193,
    "batch_idx": 160,
    "loss": 0.2247721403837204,
    "learning_rate": 2.2801612451698698e-05,
    "timestep_mean": 539.34375,
    "epsilon_norm": 7.408513069152832,
    "prediction_norm": 6.155703544616699,
    "timestamp": "2025-07-14T12:34:52.793360"
  },
  {
    "epoch": 2193,
    "batch_idx": 192,
    "loss": 0.2364455908536911,
    "learning_rate": 2.2801612451698698e-05,
    "timestep_mean": 556.125,
    "epsilon_norm": 3.066934585571289,
    "prediction_norm": 2.4808402061462402,
    "timestamp": "2025-07-14T12:34:52.798571"
  },
  {
    "epoch": 2194,
    "batch_idx": 0,
    "loss": 0.2898847758769989,
    "learning_rate": 2.3036674716303266e-05,
    "timestep_mean": 595.875,
    "epsilon_norm": 8.155513763427734,
    "prediction_norm": 7.205878257751465,
    "timestamp": "2025-07-14T12:34:52.803670"
  },
  {
    "epoch": 2194,
    "batch_idx": 32,
    "loss": 0.262993723154068,
    "learning_rate": 2.3036674716303266e-05,
    "timestep_mean": 502.0625,
    "epsilon_norm": 8.086234092712402,
    "prediction_norm": 6.90316915512085,
    "timestamp": "2025-07-14T12:34:52.809094"
  },
  {
    "epoch": 2194,
    "batch_idx": 64,
    "loss": 0.22570057213306427,
    "learning_rate": 2.3036674716303266e-05,
    "timestep_mean": 606.6875,
    "epsilon_norm": 8.710153579711914,
    "prediction_norm": 7.619816303253174,
    "timestamp": "2025-07-14T12:34:52.814151"
  },
  {
    "epoch": 2194,
    "batch_idx": 96,
    "loss": 0.29628247022628784,
    "learning_rate": 2.3036674716303266e-05,
    "timestep_mean": 438.96875,
    "epsilon_norm": 8.380583763122559,
    "prediction_norm": 7.010447025299072,
    "timestamp": "2025-07-14T12:34:52.819206"
  },
  {
    "epoch": 2194,
    "batch_idx": 128,
    "loss": 0.17506803572177887,
    "learning_rate": 2.3036674716303266e-05,
    "timestep_mean": 560.03125,
    "epsilon_norm": 8.346404075622559,
    "prediction_norm": 7.147733211517334,
    "timestamp": "2025-07-14T12:34:52.824318"
  },
  {
    "epoch": 2194,
    "batch_idx": 160,
    "loss": 0.29144614934921265,
    "learning_rate": 2.3036674716303266e-05,
    "timestep_mean": 450.71875,
    "epsilon_norm": 6.882853984832764,
    "prediction_norm": 5.775864124298096,
    "timestamp": "2025-07-14T12:34:52.829466"
  },
  {
    "epoch": 2194,
    "batch_idx": 192,
    "loss": 0.20277655124664307,
    "learning_rate": 2.3036674716303266e-05,
    "timestep_mean": 631.625,
    "epsilon_norm": 3.6485002040863037,
    "prediction_norm": 2.8395111560821533,
    "timestamp": "2025-07-14T12:34:52.834693"
  },
  {
    "epoch": 2195,
    "batch_idx": 0,
    "loss": 0.25791847705841064,
    "learning_rate": 2.3272913840499393e-05,
    "timestep_mean": 516.4375,
    "epsilon_norm": 7.419556140899658,
    "prediction_norm": 6.127612590789795,
    "timestamp": "2025-07-14T12:34:52.839843"
  },
  {
    "epoch": 2195,
    "batch_idx": 32,
    "loss": 0.34494125843048096,
    "learning_rate": 2.3272913840499393e-05,
    "timestep_mean": 436.9375,
    "epsilon_norm": 7.875904083251953,
    "prediction_norm": 6.953400611877441,
    "timestamp": "2025-07-14T12:34:52.845003"
  },
  {
    "epoch": 2195,
    "batch_idx": 64,
    "loss": 0.46998825669288635,
    "learning_rate": 2.3272913840499393e-05,
    "timestep_mean": 418.46875,
    "epsilon_norm": 8.714264869689941,
    "prediction_norm": 6.70344877243042,
    "timestamp": "2025-07-14T12:34:52.850271"
  },
  {
    "epoch": 2195,
    "batch_idx": 96,
    "loss": 0.5376214385032654,
    "learning_rate": 2.3272913840499393e-05,
    "timestep_mean": 478.46875,
    "epsilon_norm": 9.370373725891113,
    "prediction_norm": 7.427541732788086,
    "timestamp": "2025-07-14T12:34:52.855496"
  },
  {
    "epoch": 2195,
    "batch_idx": 128,
    "loss": 0.2502790093421936,
    "learning_rate": 2.3272913840499393e-05,
    "timestep_mean": 564.34375,
    "epsilon_norm": 8.212752342224121,
    "prediction_norm": 7.593293190002441,
    "timestamp": "2025-07-14T12:34:52.860711"
  },
  {
    "epoch": 2195,
    "batch_idx": 160,
    "loss": 0.1814364641904831,
    "learning_rate": 2.3272913840499393e-05,
    "timestep_mean": 469.4375,
    "epsilon_norm": 7.43546724319458,
    "prediction_norm": 6.384286880493164,
    "timestamp": "2025-07-14T12:34:52.865853"
  },
  {
    "epoch": 2195,
    "batch_idx": 192,
    "loss": 0.18450963497161865,
    "learning_rate": 2.3272913840499393e-05,
    "timestep_mean": 507.875,
    "epsilon_norm": 3.840590238571167,
    "prediction_norm": 3.416107416152954,
    "timestamp": "2025-07-14T12:34:52.870953"
  },
  {
    "epoch": 2196,
    "batch_idx": 0,
    "loss": 0.41996079683303833,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 431.84375,
    "epsilon_norm": 7.738505840301514,
    "prediction_norm": 5.979503631591797,
    "timestamp": "2025-07-14T12:34:52.876227"
  },
  {
    "epoch": 2196,
    "batch_idx": 32,
    "loss": 0.1748020350933075,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 477.8125,
    "epsilon_norm": 8.752484321594238,
    "prediction_norm": 7.968251705169678,
    "timestamp": "2025-07-14T12:34:52.881443"
  },
  {
    "epoch": 2196,
    "batch_idx": 64,
    "loss": 0.3124067783355713,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 550.625,
    "epsilon_norm": 8.93731689453125,
    "prediction_norm": 7.994266510009766,
    "timestamp": "2025-07-14T12:34:52.886577"
  },
  {
    "epoch": 2196,
    "batch_idx": 96,
    "loss": 0.26169639825820923,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 481.1875,
    "epsilon_norm": 8.057638168334961,
    "prediction_norm": 6.941879749298096,
    "timestamp": "2025-07-14T12:34:52.891848"
  },
  {
    "epoch": 2196,
    "batch_idx": 128,
    "loss": 0.26752516627311707,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 479.84375,
    "epsilon_norm": 8.834824562072754,
    "prediction_norm": 7.118525505065918,
    "timestamp": "2025-07-14T12:34:52.897074"
  },
  {
    "epoch": 2196,
    "batch_idx": 160,
    "loss": 0.18756335973739624,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 488.8125,
    "epsilon_norm": 6.9262237548828125,
    "prediction_norm": 6.28331184387207,
    "timestamp": "2025-07-14T12:34:52.902330"
  },
  {
    "epoch": 2196,
    "batch_idx": 192,
    "loss": 0.6112973690032959,
    "learning_rate": 2.3510329241390624e-05,
    "timestep_mean": 358.0,
    "epsilon_norm": 3.6176507472991943,
    "prediction_norm": 2.918023109436035,
    "timestamp": "2025-07-14T12:34:52.907454"
  },
  {
    "epoch": 2197,
    "batch_idx": 0,
    "loss": 0.20882731676101685,
    "learning_rate": 2.374892033317799e-05,
    "timestep_mean": 534.625,
    "epsilon_norm": 7.194891452789307,
    "prediction_norm": 5.930565357208252,
    "timestamp": "2025-07-14T12:34:52.912579"
  },
  {
    "epoch": 2197,
    "batch_idx": 32,
    "loss": 0.26776039600372314,
    "learning_rate": 2.374892033317799e-05,
    "timestep_mean": 490.84375,
    "epsilon_norm": 7.526581287384033,
    "prediction_norm": 6.696505546569824,
    "timestamp": "2025-07-14T12:34:52.917841"
  },
  {
    "epoch": 2197,
    "batch_idx": 64,
    "loss": 0.17692884802818298,
    "learning_rate": 2.374892033317799e-05,
    "timestep_mean": 491.78125,
    "epsilon_norm": 7.953813076019287,
    "prediction_norm": 6.8170270919799805,
    "timestamp": "2025-07-14T12:34:52.923053"
  },
  {
    "epoch": 2197,
    "batch_idx": 96,
    "loss": 0.26733818650245667,
    "learning_rate": 2.374892033317799e-05,
    "timestep_mean": 467.78125,
    "epsilon_norm": 8.266151428222656,
    "prediction_norm": 6.8999199867248535,
    "timestamp": "2025-07-14T12:34:52.928237"
  },
  {
    "epoch": 2197,
    "batch_idx": 128,
    "loss": 0.26115772128105164,
    "learning_rate": 2.374892033317799e-05,
    "timestep_mean": 506.46875,
    "epsilon_norm": 5.989043712615967,
    "prediction_norm": 5.245241641998291,
    "timestamp": "2025-07-14T12:34:52.933186"
  },
  {
    "epoch": 2197,
    "batch_idx": 160,
    "loss": 0.15957307815551758,
    "learning_rate": 2.374892033317799e-05,
    "timestep_mean": 462.65625,
    "epsilon_norm": 7.723184108734131,
    "prediction_norm": 6.712148189544678,
    "timestamp": "2025-07-14T12:34:52.938329"
  },
  {
    "epoch": 2197,
    "batch_idx": 192,
    "loss": 0.23557382822036743,
    "learning_rate": 2.374892033317799e-05,
    "timestep_mean": 559.5,
    "epsilon_norm": 4.809893608093262,
    "prediction_norm": 3.878241777420044,
    "timestamp": "2025-07-14T12:34:52.943457"
  },
  {
    "epoch": 2198,
    "batch_idx": 0,
    "loss": 0.15377670526504517,
    "learning_rate": 2.398868652716174e-05,
    "timestep_mean": 530.75,
    "epsilon_norm": 6.746760368347168,
    "prediction_norm": 6.2050886154174805,
    "timestamp": "2025-07-14T12:34:52.948713"
  },
  {
    "epoch": 2198,
    "batch_idx": 32,
    "loss": 0.3901583254337311,
    "learning_rate": 2.398868652716174e-05,
    "timestep_mean": 482.9375,
    "epsilon_norm": 8.560282707214355,
    "prediction_norm": 6.358182907104492,
    "timestamp": "2025-07-14T12:34:52.953873"
  },
  {
    "epoch": 2198,
    "batch_idx": 64,
    "loss": 0.4398749768733978,
    "learning_rate": 2.398868652716174e-05,
    "timestep_mean": 451.09375,
    "epsilon_norm": 8.416654586791992,
    "prediction_norm": 6.077602863311768,
    "timestamp": "2025-07-14T12:34:52.959279"
  },
  {
    "epoch": 2198,
    "batch_idx": 96,
    "loss": 0.2591845691204071,
    "learning_rate": 2.398868652716174e-05,
    "timestep_mean": 539.0625,
    "epsilon_norm": 7.8323140144348145,
    "prediction_norm": 6.999878883361816,
    "timestamp": "2025-07-14T12:34:52.964473"
  },
  {
    "epoch": 2198,
    "batch_idx": 128,
    "loss": 0.18256689608097076,
    "learning_rate": 2.398868652716174e-05,
    "timestep_mean": 477.0,
    "epsilon_norm": 6.525547027587891,
    "prediction_norm": 5.223465919494629,
    "timestamp": "2025-07-14T12:34:52.969590"
  },
  {
    "epoch": 2198,
    "batch_idx": 160,
    "loss": 0.36050742864608765,
    "learning_rate": 2.398868652716174e-05,
    "timestep_mean": 571.40625,
    "epsilon_norm": 8.064201354980469,
    "prediction_norm": 6.899285316467285,
    "timestamp": "2025-07-14T12:34:52.974697"
  },
  {
    "epoch": 2198,
    "batch_idx": 192,
    "loss": 0.13983842730522156,
    "learning_rate": 2.398868652716174e-05,
    "timestep_mean": 454.625,
    "epsilon_norm": 4.748592853546143,
    "prediction_norm": 4.374463081359863,
    "timestamp": "2025-07-14T12:34:52.979851"
  },
  {
    "epoch": 2199,
    "batch_idx": 0,
    "loss": 0.20361661911010742,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 559.1875,
    "epsilon_norm": 7.384793281555176,
    "prediction_norm": 6.520464897155762,
    "timestamp": "2025-07-14T12:34:52.984906"
  },
  {
    "epoch": 2199,
    "batch_idx": 32,
    "loss": 0.2460823953151703,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 499.65625,
    "epsilon_norm": 7.561460018157959,
    "prediction_norm": 6.118812561035156,
    "timestamp": "2025-07-14T12:34:52.990086"
  },
  {
    "epoch": 2199,
    "batch_idx": 64,
    "loss": 0.30321404337882996,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 416.0625,
    "epsilon_norm": 7.0407586097717285,
    "prediction_norm": 6.419000148773193,
    "timestamp": "2025-07-14T12:34:52.995201"
  },
  {
    "epoch": 2199,
    "batch_idx": 96,
    "loss": 0.3406631350517273,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 432.4375,
    "epsilon_norm": 8.960240364074707,
    "prediction_norm": 6.606173992156982,
    "timestamp": "2025-07-14T12:34:53.000503"
  },
  {
    "epoch": 2199,
    "batch_idx": 128,
    "loss": 0.2179836928844452,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 469.3125,
    "epsilon_norm": 6.583925724029541,
    "prediction_norm": 5.960675239562988,
    "timestamp": "2025-07-14T12:34:53.005674"
  },
  {
    "epoch": 2199,
    "batch_idx": 160,
    "loss": 0.19992460310459137,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 481.96875,
    "epsilon_norm": 7.304600238800049,
    "prediction_norm": 6.7754316329956055,
    "timestamp": "2025-07-14T12:34:53.010849"
  },
  {
    "epoch": 2199,
    "batch_idx": 192,
    "loss": 0.2867327630519867,
    "learning_rate": 2.4229627231742542e-05,
    "timestep_mean": 480.125,
    "epsilon_norm": 3.7422070503234863,
    "prediction_norm": 2.9980480670928955,
    "timestamp": "2025-07-14T12:34:53.015979"
  },
  {
    "epoch": 2200,
    "batch_idx": 0,
    "loss": 0.3647127151489258,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 500.53125,
    "epsilon_norm": 8.519854545593262,
    "prediction_norm": 7.222990989685059,
    "timestamp": "2025-07-14T12:34:53.021077"
  },
  {
    "epoch": 2200,
    "batch_idx": 32,
    "loss": 0.24170471727848053,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 465.96875,
    "epsilon_norm": 7.96533203125,
    "prediction_norm": 7.0692315101623535,
    "timestamp": "2025-07-14T12:34:53.026215"
  },
  {
    "epoch": 2200,
    "batch_idx": 64,
    "loss": 0.09728825837373734,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 555.875,
    "epsilon_norm": 6.94049072265625,
    "prediction_norm": 6.3111891746521,
    "timestamp": "2025-07-14T12:34:53.031363"
  },
  {
    "epoch": 2200,
    "batch_idx": 96,
    "loss": 0.1431676149368286,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 509.34375,
    "epsilon_norm": 7.71006441116333,
    "prediction_norm": 6.639052391052246,
    "timestamp": "2025-07-14T12:34:53.036569"
  },
  {
    "epoch": 2200,
    "batch_idx": 128,
    "loss": 0.3305332660675049,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 514.15625,
    "epsilon_norm": 6.9947829246521,
    "prediction_norm": 6.131302833557129,
    "timestamp": "2025-07-14T12:34:53.041714"
  },
  {
    "epoch": 2200,
    "batch_idx": 160,
    "loss": 0.3146766126155853,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 501.1875,
    "epsilon_norm": 7.528791904449463,
    "prediction_norm": 6.668147087097168,
    "timestamp": "2025-07-14T12:34:53.047032"
  },
  {
    "epoch": 2200,
    "batch_idx": 192,
    "loss": 0.564306378364563,
    "learning_rate": 2.447174185242323e-05,
    "timestep_mean": 389.125,
    "epsilon_norm": 4.442625522613525,
    "prediction_norm": 2.857279062271118,
    "timestamp": "2025-07-14T12:34:53.052117"
  },
  {
    "epoch": 2201,
    "batch_idx": 0,
    "loss": 0.33968594670295715,
    "learning_rate": 2.4715029791810057e-05,
    "timestep_mean": 551.59375,
    "epsilon_norm": 8.675765037536621,
    "prediction_norm": 7.296939373016357,
    "timestamp": "2025-07-14T12:34:53.057287"
  },
  {
    "epoch": 2201,
    "batch_idx": 32,
    "loss": 0.17944636940956116,
    "learning_rate": 2.4715029791810057e-05,
    "timestep_mean": 573.4375,
    "epsilon_norm": 8.487741470336914,
    "prediction_norm": 7.793777942657471,
    "timestamp": "2025-07-14T12:34:53.062477"
  },
  {
    "epoch": 2201,
    "batch_idx": 64,
    "loss": 0.1607649326324463,
    "learning_rate": 2.4715029791810057e-05,
    "timestep_mean": 463.8125,
    "epsilon_norm": 6.202402591705322,
    "prediction_norm": 5.652889251708984,
    "timestamp": "2025-07-14T12:34:53.067631"
  },
  {
    "epoch": 2201,
    "batch_idx": 96,
    "loss": 0.30393165349960327,
    "learning_rate": 2.4715029791810057e-05,
    "timestep_mean": 434.0,
    "epsilon_norm": 6.404499530792236,
    "prediction_norm": 4.857313632965088,
    "timestamp": "2025-07-14T12:34:53.072823"
  },
  {
    "epoch": 2201,
    "batch_idx": 128,
    "loss": 0.2423395812511444,
    "learning_rate": 2.4715029791810057e-05,
    "timestep_mean": 452.5625,
    "epsilon_norm": 5.819914817810059,
    "prediction_norm": 4.736772060394287,
    "timestamp": "2025-07-14T12:34:53.077964"
  },
  {
    "epoch": 2201,
    "batch_idx": 160,
    "loss": 0.4449005722999573,
    "learning_rate": 2.4715029791810057e-05,
    "timestep_mean": 462.75,
    "epsilon_norm": 9.010825157165527,
    "prediction_norm": 7.227792739868164,
    "timestamp": "2025-07-14T12:34:53.083129"
  },
  {
    "epoch": 2201,
    "batch_idx": 192,
    "loss": 0.1938360333442688,
    "learning_rate": 2.4715029791810057e-05,
    "timestep_mean": 465.75,
    "epsilon_norm": 4.313343524932861,
    "prediction_norm": 3.813416004180908,
    "timestamp": "2025-07-14T12:34:53.088422"
  },
  {
    "epoch": 2202,
    "batch_idx": 0,
    "loss": 0.3610699474811554,
    "learning_rate": 2.4959490449614148e-05,
    "timestep_mean": 512.0,
    "epsilon_norm": 8.94509506225586,
    "prediction_norm": 6.8694963455200195,
    "timestamp": "2025-07-14T12:34:53.093587"
  },
  {
    "epoch": 2202,
    "batch_idx": 32,
    "loss": 0.2115841656923294,
    "learning_rate": 2.4959490449614148e-05,
    "timestep_mean": 508.875,
    "epsilon_norm": 8.711915016174316,
    "prediction_norm": 7.943767547607422,
    "timestamp": "2025-07-14T12:34:53.098800"
  },
  {
    "epoch": 2202,
    "batch_idx": 64,
    "loss": 0.4311275780200958,
    "learning_rate": 2.4959490449614148e-05,
    "timestep_mean": 528.90625,
    "epsilon_norm": 8.986180305480957,
    "prediction_norm": 6.745502948760986,
    "timestamp": "2025-07-14T12:34:53.104052"
  },
  {
    "epoch": 2202,
    "batch_idx": 96,
    "loss": 0.1528301239013672,
    "learning_rate": 2.4959490449614148e-05,
    "timestep_mean": 558.96875,
    "epsilon_norm": 7.398559093475342,
    "prediction_norm": 6.387452602386475,
    "timestamp": "2025-07-14T12:34:53.109262"
  },
  {
    "epoch": 2202,
    "batch_idx": 128,
    "loss": 0.41861337423324585,
    "learning_rate": 2.4959490449614148e-05,
    "timestep_mean": 546.5625,
    "epsilon_norm": 8.261319160461426,
    "prediction_norm": 6.039865970611572,
    "timestamp": "2025-07-14T12:34:53.114380"
  },
  {
    "epoch": 2202,
    "batch_idx": 160,
    "loss": 0.31505027413368225,
    "learning_rate": 2.4959490449614148e-05,
    "timestep_mean": 512.59375,
    "epsilon_norm": 7.137465953826904,
    "prediction_norm": 6.082203388214111,
    "timestamp": "2025-07-14T12:34:53.119547"
  },
  {
    "epoch": 2202,
    "batch_idx": 192,
    "loss": 0.3577115535736084,
    "learning_rate": 2.4959490449614148e-05,
    "timestep_mean": 479.5,
    "epsilon_norm": 4.075109481811523,
    "prediction_norm": 3.170503616333008,
    "timestamp": "2025-07-14T12:34:53.124572"
  },
  {
    "epoch": 2203,
    "batch_idx": 0,
    "loss": 0.11556559801101685,
    "learning_rate": 2.5205123222653218e-05,
    "timestep_mean": 564.84375,
    "epsilon_norm": 7.315559387207031,
    "prediction_norm": 6.714468002319336,
    "timestamp": "2025-07-14T12:34:53.130036"
  },
  {
    "epoch": 2203,
    "batch_idx": 32,
    "loss": 0.17272347211837769,
    "learning_rate": 2.5205123222653218e-05,
    "timestep_mean": 504.65625,
    "epsilon_norm": 8.622879981994629,
    "prediction_norm": 7.237459659576416,
    "timestamp": "2025-07-14T12:34:53.135240"
  },
  {
    "epoch": 2203,
    "batch_idx": 64,
    "loss": 0.25593507289886475,
    "learning_rate": 2.5205123222653218e-05,
    "timestep_mean": 477.96875,
    "epsilon_norm": 7.50203800201416,
    "prediction_norm": 6.535653114318848,
    "timestamp": "2025-07-14T12:34:53.140344"
  },
  {
    "epoch": 2203,
    "batch_idx": 96,
    "loss": 0.4169916808605194,
    "learning_rate": 2.5205123222653218e-05,
    "timestep_mean": 440.28125,
    "epsilon_norm": 9.381975173950195,
    "prediction_norm": 6.844491958618164,
    "timestamp": "2025-07-14T12:34:53.145456"
  },
  {
    "epoch": 2203,
    "batch_idx": 128,
    "loss": 0.5542733669281006,
    "learning_rate": 2.5205123222653218e-05,
    "timestep_mean": 391.78125,
    "epsilon_norm": 7.928072452545166,
    "prediction_norm": 5.655940532684326,
    "timestamp": "2025-07-14T12:34:53.150442"
  },
  {
    "epoch": 2203,
    "batch_idx": 160,
    "loss": 0.2654082179069519,
    "learning_rate": 2.5205123222653218e-05,
    "timestep_mean": 513.21875,
    "epsilon_norm": 7.12026834487915,
    "prediction_norm": 6.229788303375244,
    "timestamp": "2025-07-14T12:34:53.155258"
  },
  {
    "epoch": 2203,
    "batch_idx": 192,
    "loss": 0.06197842210531235,
    "learning_rate": 2.5205123222653218e-05,
    "timestep_mean": 665.25,
    "epsilon_norm": 4.75935173034668,
    "prediction_norm": 4.681443691253662,
    "timestamp": "2025-07-14T12:34:53.160191"
  },
  {
    "epoch": 2204,
    "batch_idx": 0,
    "loss": 0.29749172925949097,
    "learning_rate": 2.545192750485275e-05,
    "timestep_mean": 486.96875,
    "epsilon_norm": 7.709157466888428,
    "prediction_norm": 6.470574378967285,
    "timestamp": "2025-07-14T12:34:53.165075"
  },
  {
    "epoch": 2204,
    "batch_idx": 32,
    "loss": 0.4681704342365265,
    "learning_rate": 2.545192750485275e-05,
    "timestep_mean": 529.71875,
    "epsilon_norm": 8.569024085998535,
    "prediction_norm": 6.4499006271362305,
    "timestamp": "2025-07-14T12:34:53.169902"
  },
  {
    "epoch": 2204,
    "batch_idx": 64,
    "loss": 0.44157612323760986,
    "learning_rate": 2.545192750485275e-05,
    "timestep_mean": 469.125,
    "epsilon_norm": 7.727487087249756,
    "prediction_norm": 6.734127521514893,
    "timestamp": "2025-07-14T12:34:53.174817"
  },
  {
    "epoch": 2204,
    "batch_idx": 96,
    "loss": 0.19910027086734772,
    "learning_rate": 2.545192750485275e-05,
    "timestep_mean": 470.6875,
    "epsilon_norm": 7.479338645935059,
    "prediction_norm": 6.701475143432617,
    "timestamp": "2025-07-14T12:34:53.179743"
  },
  {
    "epoch": 2204,
    "batch_idx": 128,
    "loss": 0.199438214302063,
    "learning_rate": 2.545192750485275e-05,
    "timestep_mean": 515.1875,
    "epsilon_norm": 7.219829082489014,
    "prediction_norm": 5.953857421875,
    "timestamp": "2025-07-14T12:34:53.184598"
  },
  {
    "epoch": 2204,
    "batch_idx": 160,
    "loss": 0.3223423659801483,
    "learning_rate": 2.545192750485275e-05,
    "timestep_mean": 431.375,
    "epsilon_norm": 8.269857406616211,
    "prediction_norm": 6.732019424438477,
    "timestamp": "2025-07-14T12:34:53.189445"
  },
  {
    "epoch": 2204,
    "batch_idx": 192,
    "loss": 0.1000852882862091,
    "learning_rate": 2.545192750485275e-05,
    "timestep_mean": 587.25,
    "epsilon_norm": 2.921645164489746,
    "prediction_norm": 2.3096187114715576,
    "timestamp": "2025-07-14T12:34:53.194277"
  },
  {
    "epoch": 2205,
    "batch_idx": 0,
    "loss": 0.34715208411216736,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 396.125,
    "epsilon_norm": 8.16454029083252,
    "prediction_norm": 6.2517170906066895,
    "timestamp": "2025-07-14T12:34:53.199171"
  },
  {
    "epoch": 2205,
    "batch_idx": 32,
    "loss": 0.18414819240570068,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 517.53125,
    "epsilon_norm": 7.247180938720703,
    "prediction_norm": 6.56971549987793,
    "timestamp": "2025-07-14T12:34:53.203998"
  },
  {
    "epoch": 2205,
    "batch_idx": 64,
    "loss": 0.5146229863166809,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 384.0625,
    "epsilon_norm": 8.225543975830078,
    "prediction_norm": 5.7425665855407715,
    "timestamp": "2025-07-14T12:34:53.208962"
  },
  {
    "epoch": 2205,
    "batch_idx": 96,
    "loss": 0.11312033236026764,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 566.53125,
    "epsilon_norm": 7.938047409057617,
    "prediction_norm": 7.391602039337158,
    "timestamp": "2025-07-14T12:34:53.213921"
  },
  {
    "epoch": 2205,
    "batch_idx": 128,
    "loss": 0.5566213726997375,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 440.6875,
    "epsilon_norm": 8.477093696594238,
    "prediction_norm": 6.21066951751709,
    "timestamp": "2025-07-14T12:34:53.218794"
  },
  {
    "epoch": 2205,
    "batch_idx": 160,
    "loss": 0.3102791905403137,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 483.46875,
    "epsilon_norm": 9.361098289489746,
    "prediction_norm": 8.213301658630371,
    "timestamp": "2025-07-14T12:34:53.223673"
  },
  {
    "epoch": 2205,
    "batch_idx": 192,
    "loss": 0.18651801347732544,
    "learning_rate": 2.5699902687247696e-05,
    "timestep_mean": 414.25,
    "epsilon_norm": 3.221388101577759,
    "prediction_norm": 2.8313348293304443,
    "timestamp": "2025-07-14T12:34:53.228482"
  },
  {
    "epoch": 2206,
    "batch_idx": 0,
    "loss": 0.35428962111473083,
    "learning_rate": 2.5949048157984096e-05,
    "timestep_mean": 512.5625,
    "epsilon_norm": 8.02587604522705,
    "prediction_norm": 6.488957405090332,
    "timestamp": "2025-07-14T12:34:53.233529"
  },
  {
    "epoch": 2206,
    "batch_idx": 32,
    "loss": 0.21894647181034088,
    "learning_rate": 2.5949048157984096e-05,
    "timestep_mean": 471.59375,
    "epsilon_norm": 6.826401710510254,
    "prediction_norm": 5.496298313140869,
    "timestamp": "2025-07-14T12:34:53.238339"
  },
  {
    "epoch": 2206,
    "batch_idx": 64,
    "loss": 0.22970443964004517,
    "learning_rate": 2.5949048157984096e-05,
    "timestep_mean": 488.78125,
    "epsilon_norm": 7.384426593780518,
    "prediction_norm": 6.251358985900879,
    "timestamp": "2025-07-14T12:34:53.243182"
  },
  {
    "epoch": 2206,
    "batch_idx": 96,
    "loss": 0.13540393114089966,
    "learning_rate": 2.5949048157984096e-05,
    "timestep_mean": 609.0,
    "epsilon_norm": 8.58399486541748,
    "prediction_norm": 7.920112133026123,
    "timestamp": "2025-07-14T12:34:53.247962"
  },
  {
    "epoch": 2206,
    "batch_idx": 128,
    "loss": 0.18327933549880981,
    "learning_rate": 2.5949048157984096e-05,
    "timestep_mean": 545.0,
    "epsilon_norm": 7.219777584075928,
    "prediction_norm": 6.7437968254089355,
    "timestamp": "2025-07-14T12:34:53.252865"
  },
  {
    "epoch": 2206,
    "batch_idx": 160,
    "loss": 0.48608726263046265,
    "learning_rate": 2.5949048157984096e-05,
    "timestep_mean": 447.5625,
    "epsilon_norm": 7.53237771987915,
    "prediction_norm": 5.98953914642334,
    "timestamp": "2025-07-14T12:34:53.257719"
  },
  {
    "epoch": 2206,
    "batch_idx": 192,
    "loss": 0.1467445343732834,
    "learning_rate": 2.5949048157984096e-05,
    "timestep_mean": 621.625,
    "epsilon_norm": 3.271299362182617,
    "prediction_norm": 2.670886754989624,
    "timestamp": "2025-07-14T12:34:53.262718"
  },
  {
    "epoch": 2207,
    "batch_idx": 0,
    "loss": 0.27098771929740906,
    "learning_rate": 2.6199363302319975e-05,
    "timestep_mean": 557.4375,
    "epsilon_norm": 6.633547306060791,
    "prediction_norm": 6.109094619750977,
    "timestamp": "2025-07-14T12:34:53.267857"
  },
  {
    "epoch": 2207,
    "batch_idx": 32,
    "loss": 0.25452813506126404,
    "learning_rate": 2.6199363302319975e-05,
    "timestep_mean": 497.0,
    "epsilon_norm": 7.732132911682129,
    "prediction_norm": 6.395445823669434,
    "timestamp": "2025-07-14T12:34:53.272838"
  },
  {
    "epoch": 2207,
    "batch_idx": 64,
    "loss": 0.1673239767551422,
    "learning_rate": 2.6199363302319975e-05,
    "timestep_mean": 557.9375,
    "epsilon_norm": 8.091897010803223,
    "prediction_norm": 7.155696868896484,
    "timestamp": "2025-07-14T12:34:53.277975"
  },
  {
    "epoch": 2207,
    "batch_idx": 96,
    "loss": 0.16461214423179626,
    "learning_rate": 2.6199363302319975e-05,
    "timestep_mean": 551.125,
    "epsilon_norm": 7.750988483428955,
    "prediction_norm": 6.944802284240723,
    "timestamp": "2025-07-14T12:34:53.283168"
  },
  {
    "epoch": 2207,
    "batch_idx": 128,
    "loss": 0.2111918181180954,
    "learning_rate": 2.6199363302319975e-05,
    "timestep_mean": 428.625,
    "epsilon_norm": 7.142619609832764,
    "prediction_norm": 5.603428840637207,
    "timestamp": "2025-07-14T12:34:53.288457"
  },
  {
    "epoch": 2207,
    "batch_idx": 160,
    "loss": 0.42609989643096924,
    "learning_rate": 2.6199363302319975e-05,
    "timestep_mean": 471.09375,
    "epsilon_norm": 8.6094331741333,
    "prediction_norm": 7.304285049438477,
    "timestamp": "2025-07-14T12:34:53.293621"
  },
  {
    "epoch": 2207,
    "batch_idx": 192,
    "loss": 0.1997685730457306,
    "learning_rate": 2.6199363302319975e-05,
    "timestep_mean": 477.75,
    "epsilon_norm": 3.0001957416534424,
    "prediction_norm": 2.6024882793426514,
    "timestamp": "2025-07-14T12:34:53.298707"
  },
  {
    "epoch": 2208,
    "batch_idx": 0,
    "loss": 0.15825463831424713,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 517.625,
    "epsilon_norm": 7.341708660125732,
    "prediction_norm": 6.763637542724609,
    "timestamp": "2025-07-14T12:34:53.303842"
  },
  {
    "epoch": 2208,
    "batch_idx": 32,
    "loss": 0.27669525146484375,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 544.4375,
    "epsilon_norm": 7.211450099945068,
    "prediction_norm": 6.424366474151611,
    "timestamp": "2025-07-14T12:34:53.309061"
  },
  {
    "epoch": 2208,
    "batch_idx": 64,
    "loss": 0.3324038088321686,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 395.28125,
    "epsilon_norm": 7.329998970031738,
    "prediction_norm": 5.5506591796875,
    "timestamp": "2025-07-14T12:34:53.314281"
  },
  {
    "epoch": 2208,
    "batch_idx": 96,
    "loss": 0.17882892489433289,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 613.59375,
    "epsilon_norm": 7.846573829650879,
    "prediction_norm": 6.832331657409668,
    "timestamp": "2025-07-14T12:34:53.319456"
  },
  {
    "epoch": 2208,
    "batch_idx": 128,
    "loss": 0.2306869626045227,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 504.65625,
    "epsilon_norm": 7.6641669273376465,
    "prediction_norm": 6.75932502746582,
    "timestamp": "2025-07-14T12:34:53.324659"
  },
  {
    "epoch": 2208,
    "batch_idx": 160,
    "loss": 0.5954721570014954,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 465.78125,
    "epsilon_norm": 8.329652786254883,
    "prediction_norm": 5.984157085418701,
    "timestamp": "2025-07-14T12:34:53.329876"
  },
  {
    "epoch": 2208,
    "batch_idx": 192,
    "loss": 0.14963868260383606,
    "learning_rate": 2.645084750262788e-05,
    "timestep_mean": 349.375,
    "epsilon_norm": 2.932621955871582,
    "prediction_norm": 2.4201719760894775,
    "timestamp": "2025-07-14T12:34:53.335041"
  },
  {
    "epoch": 2209,
    "batch_idx": 0,
    "loss": 0.40574029088020325,
    "learning_rate": 2.670350013839545e-05,
    "timestep_mean": 494.96875,
    "epsilon_norm": 8.503983497619629,
    "prediction_norm": 6.618249416351318,
    "timestamp": "2025-07-14T12:34:53.340184"
  },
  {
    "epoch": 2209,
    "batch_idx": 32,
    "loss": 0.15618008375167847,
    "learning_rate": 2.670350013839545e-05,
    "timestep_mean": 604.9375,
    "epsilon_norm": 8.227355003356934,
    "prediction_norm": 7.476046085357666,
    "timestamp": "2025-07-14T12:34:53.345406"
  },
  {
    "epoch": 2209,
    "batch_idx": 64,
    "loss": 0.2231331765651703,
    "learning_rate": 2.670350013839545e-05,
    "timestep_mean": 495.25,
    "epsilon_norm": 7.253669261932373,
    "prediction_norm": 6.455113887786865,
    "timestamp": "2025-07-14T12:34:53.350637"
  },
  {
    "epoch": 2209,
    "batch_idx": 96,
    "loss": 0.575523853302002,
    "learning_rate": 2.670350013839545e-05,
    "timestep_mean": 437.9375,
    "epsilon_norm": 8.669798851013184,
    "prediction_norm": 6.422379493713379,
    "timestamp": "2025-07-14T12:34:53.355754"
  },
  {
    "epoch": 2209,
    "batch_idx": 128,
    "loss": 0.4392017722129822,
    "learning_rate": 2.670350013839545e-05,
    "timestep_mean": 535.78125,
    "epsilon_norm": 8.277353286743164,
    "prediction_norm": 5.9822678565979,
    "timestamp": "2025-07-14T12:34:53.360945"
  },
  {
    "epoch": 2209,
    "batch_idx": 160,
    "loss": 0.42775264382362366,
    "learning_rate": 2.670350013839545e-05,
    "timestep_mean": 399.8125,
    "epsilon_norm": 8.588887214660645,
    "prediction_norm": 6.291223526000977,
    "timestamp": "2025-07-14T12:34:53.366102"
  },
  {
    "epoch": 2209,
    "batch_idx": 192,
    "loss": 0.3763583302497864,
    "learning_rate": 2.670350013839545e-05,
    "timestep_mean": 523.375,
    "epsilon_norm": 3.7010903358459473,
    "prediction_norm": 3.2068352699279785,
    "timestamp": "2025-07-14T12:34:53.371310"
  },
  {
    "epoch": 2210,
    "batch_idx": 0,
    "loss": 0.15794594585895538,
    "learning_rate": 2.6957320586227347e-05,
    "timestep_mean": 545.09375,
    "epsilon_norm": 7.876202583312988,
    "prediction_norm": 7.015949726104736,
    "timestamp": "2025-07-14T12:34:53.376174"
  },
  {
    "epoch": 2210,
    "batch_idx": 32,
    "loss": 0.274738609790802,
    "learning_rate": 2.6957320586227347e-05,
    "timestep_mean": 514.65625,
    "epsilon_norm": 8.088163375854492,
    "prediction_norm": 7.198956489562988,
    "timestamp": "2025-07-14T12:34:53.381133"
  },
  {
    "epoch": 2210,
    "batch_idx": 64,
    "loss": 0.25709956884384155,
    "learning_rate": 2.6957320586227347e-05,
    "timestep_mean": 460.21875,
    "epsilon_norm": 8.06120491027832,
    "prediction_norm": 6.588230609893799,
    "timestamp": "2025-07-14T12:34:53.386296"
  },
  {
    "epoch": 2210,
    "batch_idx": 96,
    "loss": 0.1983485221862793,
    "learning_rate": 2.6957320586227347e-05,
    "timestep_mean": 468.125,
    "epsilon_norm": 8.124066352844238,
    "prediction_norm": 7.030069351196289,
    "timestamp": "2025-07-14T12:34:53.391475"
  },
  {
    "epoch": 2210,
    "batch_idx": 128,
    "loss": 0.2416432499885559,
    "learning_rate": 2.6957320586227347e-05,
    "timestep_mean": 503.78125,
    "epsilon_norm": 9.194517135620117,
    "prediction_norm": 8.308497428894043,
    "timestamp": "2025-07-14T12:34:53.396708"
  },
  {
    "epoch": 2210,
    "batch_idx": 160,
    "loss": 0.31118160486221313,
    "learning_rate": 2.6957320586227347e-05,
    "timestep_mean": 496.03125,
    "epsilon_norm": 8.505849838256836,
    "prediction_norm": 7.203174114227295,
    "timestamp": "2025-07-14T12:34:53.401917"
  },
  {
    "epoch": 2210,
    "batch_idx": 192,
    "loss": 0.15892143547534943,
    "learning_rate": 2.6957320586227347e-05,
    "timestep_mean": 445.75,
    "epsilon_norm": 3.5827479362487793,
    "prediction_norm": 3.1803934574127197,
    "timestamp": "2025-07-14T12:34:53.406918"
  },
  {
    "epoch": 2211,
    "batch_idx": 0,
    "loss": 0.12663677334785461,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 508.1875,
    "epsilon_norm": 6.548434257507324,
    "prediction_norm": 6.012162685394287,
    "timestamp": "2025-07-14T12:34:53.411857"
  },
  {
    "epoch": 2211,
    "batch_idx": 32,
    "loss": 0.2835755944252014,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 351.3125,
    "epsilon_norm": 8.288808822631836,
    "prediction_norm": 6.85789155960083,
    "timestamp": "2025-07-14T12:34:53.417203"
  },
  {
    "epoch": 2211,
    "batch_idx": 64,
    "loss": 0.1750076413154602,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 544.03125,
    "epsilon_norm": 8.614826202392578,
    "prediction_norm": 8.012920379638672,
    "timestamp": "2025-07-14T12:34:53.422280"
  },
  {
    "epoch": 2211,
    "batch_idx": 96,
    "loss": 0.12101738899946213,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 528.21875,
    "epsilon_norm": 7.318935394287109,
    "prediction_norm": 6.628942966461182,
    "timestamp": "2025-07-14T12:34:53.427400"
  },
  {
    "epoch": 2211,
    "batch_idx": 128,
    "loss": 0.26635032892227173,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 519.8125,
    "epsilon_norm": 8.662078857421875,
    "prediction_norm": 6.623335838317871,
    "timestamp": "2025-07-14T12:34:53.432557"
  },
  {
    "epoch": 2211,
    "batch_idx": 160,
    "loss": 0.4504164457321167,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 483.03125,
    "epsilon_norm": 8.762279510498047,
    "prediction_norm": 6.45037317276001,
    "timestamp": "2025-07-14T12:34:53.437593"
  },
  {
    "epoch": 2211,
    "batch_idx": 192,
    "loss": 0.2759242057800293,
    "learning_rate": 2.7212308219846987e-05,
    "timestep_mean": 549.25,
    "epsilon_norm": 4.4665374755859375,
    "prediction_norm": 3.233306884765625,
    "timestamp": "2025-07-14T12:34:53.442918"
  },
  {
    "epoch": 2212,
    "batch_idx": 0,
    "loss": 0.13890564441680908,
    "learning_rate": 2.7468462410097587e-05,
    "timestep_mean": 539.34375,
    "epsilon_norm": 7.432728290557861,
    "prediction_norm": 6.943068504333496,
    "timestamp": "2025-07-14T12:34:53.448130"
  },
  {
    "epoch": 2212,
    "batch_idx": 32,
    "loss": 0.2539270520210266,
    "learning_rate": 2.7468462410097587e-05,
    "timestep_mean": 488.65625,
    "epsilon_norm": 8.019551277160645,
    "prediction_norm": 6.479624271392822,
    "timestamp": "2025-07-14T12:34:53.453356"
  },
  {
    "epoch": 2212,
    "batch_idx": 64,
    "loss": 0.3894622027873993,
    "learning_rate": 2.7468462410097587e-05,
    "timestep_mean": 416.84375,
    "epsilon_norm": 8.375821113586426,
    "prediction_norm": 5.908703804016113,
    "timestamp": "2025-07-14T12:34:53.458298"
  },
  {
    "epoch": 2212,
    "batch_idx": 96,
    "loss": 0.12848246097564697,
    "learning_rate": 2.7468462410097587e-05,
    "timestep_mean": 508.09375,
    "epsilon_norm": 7.736937522888184,
    "prediction_norm": 6.886041641235352,
    "timestamp": "2025-07-14T12:34:53.463488"
  },
  {
    "epoch": 2212,
    "batch_idx": 128,
    "loss": 0.2134033739566803,
    "learning_rate": 2.7468462410097587e-05,
    "timestep_mean": 523.5625,
    "epsilon_norm": 6.985106945037842,
    "prediction_norm": 6.436432838439941,
    "timestamp": "2025-07-14T12:34:53.468706"
  },
  {
    "epoch": 2212,
    "batch_idx": 160,
    "loss": 0.2759246528148651,
    "learning_rate": 2.7468462410097587e-05,
    "timestep_mean": 458.28125,
    "epsilon_norm": 7.9771623611450195,
    "prediction_norm": 6.647865295410156,
    "timestamp": "2025-07-14T12:34:53.473839"
  },
  {
    "epoch": 2212,
    "batch_idx": 192,
    "loss": 0.39362865686416626,
    "learning_rate": 2.7468462410097587e-05,
    "timestep_mean": 540.875,
    "epsilon_norm": 4.810657501220703,
    "prediction_norm": 3.519749402999878,
    "timestamp": "2025-07-14T12:34:53.479142"
  },
  {
    "epoch": 2213,
    "batch_idx": 0,
    "loss": 0.17138686776161194,
    "learning_rate": 2.772578252494428e-05,
    "timestep_mean": 557.9375,
    "epsilon_norm": 7.52223014831543,
    "prediction_norm": 6.750734329223633,
    "timestamp": "2025-07-14T12:34:53.484526"
  },
  {
    "epoch": 2213,
    "batch_idx": 32,
    "loss": 0.40520673990249634,
    "learning_rate": 2.772578252494428e-05,
    "timestep_mean": 482.6875,
    "epsilon_norm": 7.940892219543457,
    "prediction_norm": 6.130568027496338,
    "timestamp": "2025-07-14T12:34:53.489663"
  },
  {
    "epoch": 2213,
    "batch_idx": 64,
    "loss": 0.32575222849845886,
    "learning_rate": 2.772578252494428e-05,
    "timestep_mean": 527.65625,
    "epsilon_norm": 8.488447189331055,
    "prediction_norm": 7.2661824226379395,
    "timestamp": "2025-07-14T12:34:53.494952"
  },
  {
    "epoch": 2213,
    "batch_idx": 96,
    "loss": 0.2531246542930603,
    "learning_rate": 2.772578252494428e-05,
    "timestep_mean": 461.34375,
    "epsilon_norm": 8.112624168395996,
    "prediction_norm": 7.384342193603516,
    "timestamp": "2025-07-14T12:34:53.500060"
  },
  {
    "epoch": 2213,
    "batch_idx": 128,
    "loss": 0.36354774236679077,
    "learning_rate": 2.772578252494428e-05,
    "timestep_mean": 390.84375,
    "epsilon_norm": 7.051604270935059,
    "prediction_norm": 4.93308162689209,
    "timestamp": "2025-07-14T12:34:53.505195"
  },
  {
    "epoch": 2213,
    "batch_idx": 160,
    "loss": 0.23767244815826416,
    "learning_rate": 2.772578252494428e-05,
    "timestep_mean": 466.9375,
    "epsilon_norm": 7.402740001678467,
    "prediction_norm": 6.45417594909668,
    "timestamp": "2025-07-14T12:34:53.510361"
  },
  {
    "epoch": 2213,
    "batch_idx": 192,
    "loss": 0.13067223131656647,
    "learning_rate": 2.772578252494428e-05,
    "timestep_mean": 346.375,
    "epsilon_norm": 2.603217601776123,
    "prediction_norm": 2.6228222846984863,
    "timestamp": "2025-07-14T12:34:53.515507"
  },
  {
    "epoch": 2214,
    "batch_idx": 0,
    "loss": 0.542833685874939,
    "learning_rate": 2.798426792947516e-05,
    "timestep_mean": 441.09375,
    "epsilon_norm": 8.274090766906738,
    "prediction_norm": 6.218810081481934,
    "timestamp": "2025-07-14T12:34:53.520882"
  },
  {
    "epoch": 2214,
    "batch_idx": 32,
    "loss": 0.1571744680404663,
    "learning_rate": 2.798426792947516e-05,
    "timestep_mean": 552.65625,
    "epsilon_norm": 7.0560760498046875,
    "prediction_norm": 6.796413898468018,
    "timestamp": "2025-07-14T12:34:53.526096"
  },
  {
    "epoch": 2214,
    "batch_idx": 64,
    "loss": 0.3056919574737549,
    "learning_rate": 2.798426792947516e-05,
    "timestep_mean": 490.0,
    "epsilon_norm": 7.572585105895996,
    "prediction_norm": 6.3696088790893555,
    "timestamp": "2025-07-14T12:34:53.531215"
  },
  {
    "epoch": 2214,
    "batch_idx": 96,
    "loss": 0.2958627939224243,
    "learning_rate": 2.798426792947516e-05,
    "timestep_mean": 471.5625,
    "epsilon_norm": 7.50717306137085,
    "prediction_norm": 6.196311950683594,
    "timestamp": "2025-07-14T12:34:53.536476"
  },
  {
    "epoch": 2214,
    "batch_idx": 128,
    "loss": 0.14915814995765686,
    "learning_rate": 2.798426792947516e-05,
    "timestep_mean": 501.46875,
    "epsilon_norm": 6.384957790374756,
    "prediction_norm": 5.686193943023682,
    "timestamp": "2025-07-14T12:34:53.541652"
  },
  {
    "epoch": 2214,
    "batch_idx": 160,
    "loss": 0.6048247218132019,
    "learning_rate": 2.798426792947516e-05,
    "timestep_mean": 477.78125,
    "epsilon_norm": 8.876545906066895,
    "prediction_norm": 6.323363304138184,
    "timestamp": "2025-07-14T12:34:53.546967"
  },
  {
    "epoch": 2214,
    "batch_idx": 192,
    "loss": 1.0048648118972778,
    "learning_rate": 2.798426792947516e-05,
    "timestep_mean": 410.25,
    "epsilon_norm": 5.2525129318237305,
    "prediction_norm": 3.9861652851104736,
    "timestamp": "2025-07-14T12:34:53.552136"
  },
  {
    "epoch": 2215,
    "batch_idx": 0,
    "loss": 0.31068161129951477,
    "learning_rate": 2.8243917985903188e-05,
    "timestep_mean": 478.65625,
    "epsilon_norm": 8.462591171264648,
    "prediction_norm": 7.183691501617432,
    "timestamp": "2025-07-14T12:34:53.557371"
  },
  {
    "epoch": 2215,
    "batch_idx": 32,
    "loss": 0.4749608039855957,
    "learning_rate": 2.8243917985903188e-05,
    "timestep_mean": 413.375,
    "epsilon_norm": 8.951654434204102,
    "prediction_norm": 7.112136363983154,
    "timestamp": "2025-07-14T12:34:53.562571"
  },
  {
    "epoch": 2215,
    "batch_idx": 64,
    "loss": 0.21090835332870483,
    "learning_rate": 2.8243917985903188e-05,
    "timestep_mean": 548.375,
    "epsilon_norm": 6.7870001792907715,
    "prediction_norm": 5.706225872039795,
    "timestamp": "2025-07-14T12:34:53.567816"
  },
  {
    "epoch": 2215,
    "batch_idx": 96,
    "loss": 0.3916577696800232,
    "learning_rate": 2.8243917985903188e-05,
    "timestep_mean": 402.875,
    "epsilon_norm": 8.921283721923828,
    "prediction_norm": 6.936136722564697,
    "timestamp": "2025-07-14T12:34:53.572844"
  },
  {
    "epoch": 2215,
    "batch_idx": 128,
    "loss": 0.4626834988594055,
    "learning_rate": 2.8243917985903188e-05,
    "timestep_mean": 477.5625,
    "epsilon_norm": 8.743738174438477,
    "prediction_norm": 6.010957717895508,
    "timestamp": "2025-07-14T12:34:53.578126"
  },
  {
    "epoch": 2215,
    "batch_idx": 160,
    "loss": 0.5326500535011292,
    "learning_rate": 2.8243917985903188e-05,
    "timestep_mean": 529.9375,
    "epsilon_norm": 8.7833833694458,
    "prediction_norm": 6.595823287963867,
    "timestamp": "2025-07-14T12:34:53.583287"
  },
  {
    "epoch": 2215,
    "batch_idx": 192,
    "loss": 0.10145246237516403,
    "learning_rate": 2.8243917985903188e-05,
    "timestep_mean": 381.0,
    "epsilon_norm": 3.0888352394104004,
    "prediction_norm": 3.2072668075561523,
    "timestamp": "2025-07-14T12:34:53.588581"
  },
  {
    "epoch": 2216,
    "batch_idx": 0,
    "loss": 0.20869886875152588,
    "learning_rate": 2.8504732053567837e-05,
    "timestep_mean": 488.0,
    "epsilon_norm": 7.937518119812012,
    "prediction_norm": 7.178534507751465,
    "timestamp": "2025-07-14T12:34:53.593722"
  },
  {
    "epoch": 2216,
    "batch_idx": 32,
    "loss": 0.3224305808544159,
    "learning_rate": 2.8504732053567837e-05,
    "timestep_mean": 499.5625,
    "epsilon_norm": 7.309872627258301,
    "prediction_norm": 6.06897497177124,
    "timestamp": "2025-07-14T12:34:53.598874"
  },
  {
    "epoch": 2216,
    "batch_idx": 64,
    "loss": 0.19913600385189056,
    "learning_rate": 2.8504732053567837e-05,
    "timestep_mean": 511.0,
    "epsilon_norm": 7.871572017669678,
    "prediction_norm": 6.576340675354004,
    "timestamp": "2025-07-14T12:34:53.604138"
  },
  {
    "epoch": 2216,
    "batch_idx": 96,
    "loss": 0.20583677291870117,
    "learning_rate": 2.8504732053567837e-05,
    "timestep_mean": 515.1875,
    "epsilon_norm": 7.553702354431152,
    "prediction_norm": 6.986380577087402,
    "timestamp": "2025-07-14T12:34:53.609527"
  },
  {
    "epoch": 2216,
    "batch_idx": 128,
    "loss": 0.38022342324256897,
    "learning_rate": 2.8504732053567837e-05,
    "timestep_mean": 441.40625,
    "epsilon_norm": 8.008853912353516,
    "prediction_norm": 6.909347057342529,
    "timestamp": "2025-07-14T12:34:53.614630"
  },
  {
    "epoch": 2216,
    "batch_idx": 160,
    "loss": 0.25063979625701904,
    "learning_rate": 2.8504732053567837e-05,
    "timestep_mean": 496.90625,
    "epsilon_norm": 6.536674976348877,
    "prediction_norm": 5.326143264770508,
    "timestamp": "2025-07-14T12:34:53.619763"
  },
  {
    "epoch": 2216,
    "batch_idx": 192,
    "loss": 0.38141703605651855,
    "learning_rate": 2.8504732053567837e-05,
    "timestep_mean": 357.375,
    "epsilon_norm": 2.9390578269958496,
    "prediction_norm": 1.838440179824829,
    "timestamp": "2025-07-14T12:34:53.624986"
  },
  {
    "epoch": 2217,
    "batch_idx": 0,
    "loss": 0.272659033536911,
    "learning_rate": 2.876670948893605e-05,
    "timestep_mean": 574.8125,
    "epsilon_norm": 8.237401962280273,
    "prediction_norm": 6.970982551574707,
    "timestamp": "2025-07-14T12:34:53.630366"
  },
  {
    "epoch": 2217,
    "batch_idx": 32,
    "loss": 0.24894605576992035,
    "learning_rate": 2.876670948893605e-05,
    "timestep_mean": 486.90625,
    "epsilon_norm": 7.548069953918457,
    "prediction_norm": 5.936403751373291,
    "timestamp": "2025-07-14T12:34:53.635579"
  },
  {
    "epoch": 2217,
    "batch_idx": 64,
    "loss": 0.15285328030586243,
    "learning_rate": 2.876670948893605e-05,
    "timestep_mean": 543.46875,
    "epsilon_norm": 7.26837682723999,
    "prediction_norm": 6.336068153381348,
    "timestamp": "2025-07-14T12:34:53.640765"
  },
  {
    "epoch": 2217,
    "batch_idx": 96,
    "loss": 0.32271674275398254,
    "learning_rate": 2.876670948893605e-05,
    "timestep_mean": 397.40625,
    "epsilon_norm": 6.93418550491333,
    "prediction_norm": 5.9704389572143555,
    "timestamp": "2025-07-14T12:34:53.645996"
  },
  {
    "epoch": 2217,
    "batch_idx": 128,
    "loss": 0.22717520594596863,
    "learning_rate": 2.876670948893605e-05,
    "timestep_mean": 533.0625,
    "epsilon_norm": 8.599355697631836,
    "prediction_norm": 7.431723594665527,
    "timestamp": "2025-07-14T12:34:53.651102"
  },
  {
    "epoch": 2217,
    "batch_idx": 160,
    "loss": 0.25185292959213257,
    "learning_rate": 2.876670948893605e-05,
    "timestep_mean": 463.65625,
    "epsilon_norm": 7.860836982727051,
    "prediction_norm": 6.573348522186279,
    "timestamp": "2025-07-14T12:34:53.656213"
  },
  {
    "epoch": 2217,
    "batch_idx": 192,
    "loss": 0.6302995681762695,
    "learning_rate": 2.876670948893605e-05,
    "timestep_mean": 488.625,
    "epsilon_norm": 4.5764288902282715,
    "prediction_norm": 3.4285521507263184,
    "timestamp": "2025-07-14T12:34:53.661408"
  },
  {
    "epoch": 2218,
    "batch_idx": 0,
    "loss": 0.6780067086219788,
    "learning_rate": 2.9029849645604723e-05,
    "timestep_mean": 417.78125,
    "epsilon_norm": 8.344869613647461,
    "prediction_norm": 5.681224822998047,
    "timestamp": "2025-07-14T12:34:53.666688"
  },
  {
    "epoch": 2218,
    "batch_idx": 32,
    "loss": 0.3154445290565491,
    "learning_rate": 2.9029849645604723e-05,
    "timestep_mean": 469.65625,
    "epsilon_norm": 6.929445743560791,
    "prediction_norm": 5.676628589630127,
    "timestamp": "2025-07-14T12:34:53.671875"
  },
  {
    "epoch": 2218,
    "batch_idx": 64,
    "loss": 0.17420774698257446,
    "learning_rate": 2.9029849645604723e-05,
    "timestep_mean": 543.375,
    "epsilon_norm": 7.739081859588623,
    "prediction_norm": 7.112042427062988,
    "timestamp": "2025-07-14T12:34:53.677011"
  },
  {
    "epoch": 2218,
    "batch_idx": 96,
    "loss": 0.21567758917808533,
    "learning_rate": 2.9029849645604723e-05,
    "timestep_mean": 539.03125,
    "epsilon_norm": 8.776124954223633,
    "prediction_norm": 7.66267728805542,
    "timestamp": "2025-07-14T12:34:53.682204"
  },
  {
    "epoch": 2218,
    "batch_idx": 128,
    "loss": 0.2923697531223297,
    "learning_rate": 2.9029849645604723e-05,
    "timestep_mean": 513.53125,
    "epsilon_norm": 9.48176097869873,
    "prediction_norm": 8.49461841583252,
    "timestamp": "2025-07-14T12:34:53.687427"
  },
  {
    "epoch": 2218,
    "batch_idx": 160,
    "loss": 0.20972248911857605,
    "learning_rate": 2.9029849645604723e-05,
    "timestep_mean": 571.5625,
    "epsilon_norm": 7.3772382736206055,
    "prediction_norm": 6.184971809387207,
    "timestamp": "2025-07-14T12:34:53.692804"
  },
  {
    "epoch": 2218,
    "batch_idx": 192,
    "loss": 0.17234168946743011,
    "learning_rate": 2.9029849645604723e-05,
    "timestep_mean": 436.5,
    "epsilon_norm": 3.7946345806121826,
    "prediction_norm": 3.210249900817871,
    "timestamp": "2025-07-14T12:34:53.697920"
  },
  {
    "epoch": 2219,
    "batch_idx": 0,
    "loss": 0.2830689549446106,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 513.40625,
    "epsilon_norm": 8.231283187866211,
    "prediction_norm": 6.841257572174072,
    "timestamp": "2025-07-14T12:34:53.703164"
  },
  {
    "epoch": 2219,
    "batch_idx": 32,
    "loss": 0.2865230143070221,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 501.0,
    "epsilon_norm": 7.188407897949219,
    "prediction_norm": 6.125839710235596,
    "timestamp": "2025-07-14T12:34:53.708532"
  },
  {
    "epoch": 2219,
    "batch_idx": 64,
    "loss": 0.3115965723991394,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 447.15625,
    "epsilon_norm": 8.261123657226562,
    "prediction_norm": 6.778960704803467,
    "timestamp": "2025-07-14T12:34:53.713706"
  },
  {
    "epoch": 2219,
    "batch_idx": 96,
    "loss": 0.4867200553417206,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 485.15625,
    "epsilon_norm": 7.790428161621094,
    "prediction_norm": 5.920174598693848,
    "timestamp": "2025-07-14T12:34:53.718853"
  },
  {
    "epoch": 2219,
    "batch_idx": 128,
    "loss": 0.25697988271713257,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 490.5625,
    "epsilon_norm": 7.501555442810059,
    "prediction_norm": 6.500092506408691,
    "timestamp": "2025-07-14T12:34:53.724105"
  },
  {
    "epoch": 2219,
    "batch_idx": 160,
    "loss": 0.3226238489151001,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 488.59375,
    "epsilon_norm": 8.447952270507812,
    "prediction_norm": 6.476746082305908,
    "timestamp": "2025-07-14T12:34:53.729318"
  },
  {
    "epoch": 2219,
    "batch_idx": 192,
    "loss": 0.1698884665966034,
    "learning_rate": 2.9294151874301615e-05,
    "timestep_mean": 594.875,
    "epsilon_norm": 3.8422129154205322,
    "prediction_norm": 3.527493715286255,
    "timestamp": "2025-07-14T12:34:53.734490"
  },
  {
    "epoch": 2220,
    "batch_idx": 0,
    "loss": 0.33715635538101196,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 512.34375,
    "epsilon_norm": 8.760738372802734,
    "prediction_norm": 6.837008953094482,
    "timestamp": "2025-07-14T12:34:53.739426"
  },
  {
    "epoch": 2220,
    "batch_idx": 32,
    "loss": 0.308926522731781,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 557.78125,
    "epsilon_norm": 7.80560302734375,
    "prediction_norm": 6.83754825592041,
    "timestamp": "2025-07-14T12:34:53.744314"
  },
  {
    "epoch": 2220,
    "batch_idx": 64,
    "loss": 0.2590237259864807,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 450.75,
    "epsilon_norm": 8.650702476501465,
    "prediction_norm": 7.133548259735107,
    "timestamp": "2025-07-14T12:34:53.749208"
  },
  {
    "epoch": 2220,
    "batch_idx": 96,
    "loss": 0.2971794903278351,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 403.0625,
    "epsilon_norm": 6.531971454620361,
    "prediction_norm": 5.476563930511475,
    "timestamp": "2025-07-14T12:34:53.754028"
  },
  {
    "epoch": 2220,
    "batch_idx": 128,
    "loss": 0.4062615633010864,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 548.5625,
    "epsilon_norm": 8.157083511352539,
    "prediction_norm": 6.45265007019043,
    "timestamp": "2025-07-14T12:34:53.759001"
  },
  {
    "epoch": 2220,
    "batch_idx": 160,
    "loss": 0.5416846871376038,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 401.5,
    "epsilon_norm": 8.635560989379883,
    "prediction_norm": 6.770707130432129,
    "timestamp": "2025-07-14T12:34:53.763841"
  },
  {
    "epoch": 2220,
    "batch_idx": 192,
    "loss": 0.5111027956008911,
    "learning_rate": 2.955961552288732e-05,
    "timestep_mean": 365.5,
    "epsilon_norm": 4.35472297668457,
    "prediction_norm": 2.8763537406921387,
    "timestamp": "2025-07-14T12:34:53.768709"
  },
  {
    "epoch": 2221,
    "batch_idx": 0,
    "loss": 0.6839481592178345,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 520.4375,
    "epsilon_norm": 9.66470718383789,
    "prediction_norm": 7.027004718780518,
    "timestamp": "2025-07-14T12:34:53.773790"
  },
  {
    "epoch": 2221,
    "batch_idx": 32,
    "loss": 0.19286638498306274,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 496.03125,
    "epsilon_norm": 7.740805149078369,
    "prediction_norm": 6.697580814361572,
    "timestamp": "2025-07-14T12:34:53.778924"
  },
  {
    "epoch": 2221,
    "batch_idx": 64,
    "loss": 0.1148926243185997,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 543.625,
    "epsilon_norm": 6.612951278686523,
    "prediction_norm": 6.114381313323975,
    "timestamp": "2025-07-14T12:34:53.784075"
  },
  {
    "epoch": 2221,
    "batch_idx": 96,
    "loss": 0.35367485880851746,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 488.78125,
    "epsilon_norm": 7.722715377807617,
    "prediction_norm": 6.038073539733887,
    "timestamp": "2025-07-14T12:34:53.789192"
  },
  {
    "epoch": 2221,
    "batch_idx": 128,
    "loss": 0.4491443336009979,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 448.6875,
    "epsilon_norm": 8.407196044921875,
    "prediction_norm": 6.920974254608154,
    "timestamp": "2025-07-14T12:34:53.794339"
  },
  {
    "epoch": 2221,
    "batch_idx": 160,
    "loss": 0.5331017971038818,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 413.46875,
    "epsilon_norm": 8.85998249053955,
    "prediction_norm": 6.895301818847656,
    "timestamp": "2025-07-14T12:34:53.799454"
  },
  {
    "epoch": 2221,
    "batch_idx": 192,
    "loss": 0.18617761135101318,
    "learning_rate": 2.9826239936356518e-05,
    "timestep_mean": 520.375,
    "epsilon_norm": 2.9252843856811523,
    "prediction_norm": 2.6808013916015625,
    "timestamp": "2025-07-14T12:34:53.804510"
  },
  {
    "epoch": 2222,
    "batch_idx": 0,
    "loss": 0.16548389196395874,
    "learning_rate": 3.009402445684022e-05,
    "timestep_mean": 568.34375,
    "epsilon_norm": 7.540687084197998,
    "prediction_norm": 6.86620569229126,
    "timestamp": "2025-07-14T12:34:53.809593"
  },
  {
    "epoch": 2222,
    "batch_idx": 32,
    "loss": 0.1321890950202942,
    "learning_rate": 3.009402445684022e-05,
    "timestep_mean": 536.71875,
    "epsilon_norm": 7.813183784484863,
    "prediction_norm": 7.214004993438721,
    "timestamp": "2025-07-14T12:34:53.814715"
  },
  {
    "epoch": 2222,
    "batch_idx": 64,
    "loss": 0.26595303416252136,
    "learning_rate": 3.009402445684022e-05,
    "timestep_mean": 510.0625,
    "epsilon_norm": 8.931087493896484,
    "prediction_norm": 7.674853801727295,
    "timestamp": "2025-07-14T12:34:53.819837"
  },
  {
    "epoch": 2222,
    "batch_idx": 96,
    "loss": 0.4916531443595886,
    "learning_rate": 3.009402445684022e-05,
    "timestep_mean": 528.59375,
    "epsilon_norm": 9.48644733428955,
    "prediction_norm": 7.671720504760742,
    "timestamp": "2025-07-14T12:34:53.824773"
  },
  {
    "epoch": 2222,
    "batch_idx": 128,
    "loss": 0.1380702257156372,
    "learning_rate": 3.009402445684022e-05,
    "timestep_mean": 484.65625,
    "epsilon_norm": 7.909388542175293,
    "prediction_norm": 7.5729756355285645,
    "timestamp": "2025-07-14T12:34:53.829725"
  },
  {
    "epoch": 2222,
    "batch_idx": 160,
    "loss": 0.22332638502120972,
    "learning_rate": 3.009402445684022e-05,
    "timestep_mean": 507.40625,
    "epsilon_norm": 7.3813934326171875,
    "prediction_norm": 6.454699993133545,
    "timestamp": "2025-07-14T12:34:53.834597"
  },
  {
    "epoch": 2222,
    "batch_idx": 192,
    "loss": 0.2491838037967682,
    "learning_rate": 3.009402445684022e-05,
    "timestep_mean": 497.25,
    "epsilon_norm": 3.9839231967926025,
    "prediction_norm": 3.5773766040802,
    "timestamp": "2025-07-14T12:34:53.839452"
  },
  {
    "epoch": 2223,
    "batch_idx": 0,
    "loss": 0.24145108461380005,
    "learning_rate": 3.036296842360652e-05,
    "timestep_mean": 521.1875,
    "epsilon_norm": 8.421630859375,
    "prediction_norm": 7.884946346282959,
    "timestamp": "2025-07-14T12:34:53.844528"
  },
  {
    "epoch": 2223,
    "batch_idx": 32,
    "loss": 0.24748092889785767,
    "learning_rate": 3.036296842360652e-05,
    "timestep_mean": 464.0,
    "epsilon_norm": 7.309399127960205,
    "prediction_norm": 6.402005195617676,
    "timestamp": "2025-07-14T12:34:53.849459"
  },
  {
    "epoch": 2223,
    "batch_idx": 64,
    "loss": 0.2725299000740051,
    "learning_rate": 3.036296842360652e-05,
    "timestep_mean": 488.40625,
    "epsilon_norm": 8.084582328796387,
    "prediction_norm": 7.0592570304870605,
    "timestamp": "2025-07-14T12:34:53.854274"
  },
  {
    "epoch": 2223,
    "batch_idx": 96,
    "loss": 0.2555997371673584,
    "learning_rate": 3.036296842360652e-05,
    "timestep_mean": 646.84375,
    "epsilon_norm": 8.106060028076172,
    "prediction_norm": 6.627750873565674,
    "timestamp": "2025-07-14T12:34:53.859290"
  },
  {
    "epoch": 2223,
    "batch_idx": 128,
    "loss": 0.302367627620697,
    "learning_rate": 3.036296842360652e-05,
    "timestep_mean": 436.9375,
    "epsilon_norm": 8.0029878616333,
    "prediction_norm": 6.456353664398193,
    "timestamp": "2025-07-14T12:34:53.864130"
  },
  {
    "epoch": 2223,
    "batch_idx": 160,
    "loss": 0.5974730849266052,
    "learning_rate": 3.036296842360652e-05,
    "timestep_mean": 494.9375,
    "epsilon_norm": 8.701942443847656,
    "prediction_norm": 6.961742877960205,
    "timestamp": "2025-07-14T12:34:53.868961"
  },
  {
    "epoch": 2223,
    "batch_idx": 192,
    "loss": 0.19200414419174194,
    "learning_rate": 3.036296842360652e-05,
    "timestep_mean": 538.75,
    "epsilon_norm": 4.546763896942139,
    "prediction_norm": 4.964007377624512,
    "timestamp": "2025-07-14T12:34:53.873782"
  },
  {
    "epoch": 2224,
    "batch_idx": 0,
    "loss": 0.13167159259319305,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 596.5,
    "epsilon_norm": 8.107913970947266,
    "prediction_norm": 7.870992183685303,
    "timestamp": "2025-07-14T12:34:53.878773"
  },
  {
    "epoch": 2224,
    "batch_idx": 32,
    "loss": 0.3601100742816925,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 507.34375,
    "epsilon_norm": 8.120083808898926,
    "prediction_norm": 6.368712902069092,
    "timestamp": "2025-07-14T12:34:53.883685"
  },
  {
    "epoch": 2224,
    "batch_idx": 64,
    "loss": 0.3791634142398834,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 411.6875,
    "epsilon_norm": 7.427974700927734,
    "prediction_norm": 5.353906631469727,
    "timestamp": "2025-07-14T12:34:53.888631"
  },
  {
    "epoch": 2224,
    "batch_idx": 96,
    "loss": 0.27555567026138306,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 488.0,
    "epsilon_norm": 8.277490615844727,
    "prediction_norm": 7.254385471343994,
    "timestamp": "2025-07-14T12:34:53.893540"
  },
  {
    "epoch": 2224,
    "batch_idx": 128,
    "loss": 0.280915230512619,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 532.65625,
    "epsilon_norm": 8.000364303588867,
    "prediction_norm": 7.281966209411621,
    "timestamp": "2025-07-14T12:34:53.900516"
  },
  {
    "epoch": 2224,
    "batch_idx": 160,
    "loss": 0.4164067804813385,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 377.09375,
    "epsilon_norm": 9.003479957580566,
    "prediction_norm": 7.30103063583374,
    "timestamp": "2025-07-14T12:34:53.907055"
  },
  {
    "epoch": 2224,
    "batch_idx": 192,
    "loss": 0.1781989336013794,
    "learning_rate": 3.063307117306296e-05,
    "timestep_mean": 529.75,
    "epsilon_norm": 3.7997426986694336,
    "prediction_norm": 3.5631518363952637,
    "timestamp": "2025-07-14T12:34:53.913288"
  },
  {
    "epoch": 2225,
    "batch_idx": 0,
    "loss": 0.16432100534439087,
    "learning_rate": 3.090433203875802e-05,
    "timestep_mean": 471.28125,
    "epsilon_norm": 6.728714466094971,
    "prediction_norm": 5.708161354064941,
    "timestamp": "2025-07-14T12:34:53.920367"
  },
  {
    "epoch": 2225,
    "batch_idx": 32,
    "loss": 0.2483980655670166,
    "learning_rate": 3.090433203875802e-05,
    "timestep_mean": 536.21875,
    "epsilon_norm": 7.841914653778076,
    "prediction_norm": 6.591660022735596,
    "timestamp": "2025-07-14T12:34:53.927214"
  },
  {
    "epoch": 2225,
    "batch_idx": 64,
    "loss": 0.38082659244537354,
    "learning_rate": 3.090433203875802e-05,
    "timestep_mean": 507.46875,
    "epsilon_norm": 8.773406982421875,
    "prediction_norm": 6.779058933258057,
    "timestamp": "2025-07-14T12:34:53.933195"
  },
  {
    "epoch": 2225,
    "batch_idx": 96,
    "loss": 0.40436315536499023,
    "learning_rate": 3.090433203875802e-05,
    "timestep_mean": 425.21875,
    "epsilon_norm": 7.3678693771362305,
    "prediction_norm": 5.997736930847168,
    "timestamp": "2025-07-14T12:34:53.939730"
  },
  {
    "epoch": 2225,
    "batch_idx": 128,
    "loss": 0.29557931423187256,
    "learning_rate": 3.090433203875802e-05,
    "timestep_mean": 478.40625,
    "epsilon_norm": 7.815457820892334,
    "prediction_norm": 6.67259407043457,
    "timestamp": "2025-07-14T12:34:53.946036"
  },
  {
    "epoch": 2225,
    "batch_idx": 160,
    "loss": 0.17191669344902039,
    "learning_rate": 3.090433203875802e-05,
    "timestep_mean": 496.40625,
    "epsilon_norm": 7.3093647956848145,
    "prediction_norm": 5.9858198165893555,
    "timestamp": "2025-07-14T12:34:53.951702"
  },
  {
    "epoch": 2225,
    "batch_idx": 192,
    "loss": 0.15368513762950897,
    "learning_rate": 3.090433203875802e-05,
    "timestep_mean": 523.0,
    "epsilon_norm": 3.851060152053833,
    "prediction_norm": 3.7270796298980713,
    "timestamp": "2025-07-14T12:34:53.957602"
  },
  {
    "epoch": 2226,
    "batch_idx": 0,
    "loss": 0.5139524340629578,
    "learning_rate": 3.1176750351382165e-05,
    "timestep_mean": 440.0,
    "epsilon_norm": 9.358441352844238,
    "prediction_norm": 6.784777641296387,
    "timestamp": "2025-07-14T12:34:53.963118"
  },
  {
    "epoch": 2226,
    "batch_idx": 32,
    "loss": 0.17226853966712952,
    "learning_rate": 3.1176750351382165e-05,
    "timestep_mean": 515.5625,
    "epsilon_norm": 7.415907859802246,
    "prediction_norm": 6.538266181945801,
    "timestamp": "2025-07-14T12:34:53.968421"
  },
  {
    "epoch": 2226,
    "batch_idx": 64,
    "loss": 0.13229712843894958,
    "learning_rate": 3.1176750351382165e-05,
    "timestep_mean": 464.71875,
    "epsilon_norm": 7.624476432800293,
    "prediction_norm": 6.909841537475586,
    "timestamp": "2025-07-14T12:34:53.973991"
  },
  {
    "epoch": 2226,
    "batch_idx": 96,
    "loss": 0.16971731185913086,
    "learning_rate": 3.1176750351382165e-05,
    "timestep_mean": 523.09375,
    "epsilon_norm": 6.935450077056885,
    "prediction_norm": 6.218749523162842,
    "timestamp": "2025-07-14T12:34:53.979422"
  },
  {
    "epoch": 2226,
    "batch_idx": 128,
    "loss": 0.27532297372817993,
    "learning_rate": 3.1176750351382165e-05,
    "timestep_mean": 461.5625,
    "epsilon_norm": 9.250070571899414,
    "prediction_norm": 7.813875675201416,
    "timestamp": "2025-07-14T12:34:53.984836"
  },
  {
    "epoch": 2226,
    "batch_idx": 160,
    "loss": 0.1396060734987259,
    "learning_rate": 3.1176750351382165e-05,
    "timestep_mean": 509.5,
    "epsilon_norm": 7.075045585632324,
    "prediction_norm": 6.619739532470703,
    "timestamp": "2025-07-14T12:34:53.989967"
  },
  {
    "epoch": 2226,
    "batch_idx": 192,
    "loss": 0.32206517457962036,
    "learning_rate": 3.1176750351382165e-05,
    "timestep_mean": 383.25,
    "epsilon_norm": 4.377369403839111,
    "prediction_norm": 3.6349034309387207,
    "timestamp": "2025-07-14T12:34:53.995326"
  },
  {
    "epoch": 2227,
    "batch_idx": 0,
    "loss": 0.21541985869407654,
    "learning_rate": 3.1450325438770617e-05,
    "timestep_mean": 530.71875,
    "epsilon_norm": 8.356369018554688,
    "prediction_norm": 7.087202072143555,
    "timestamp": "2025-07-14T12:34:54.000738"
  },
  {
    "epoch": 2227,
    "batch_idx": 32,
    "loss": 0.18598206341266632,
    "learning_rate": 3.1450325438770617e-05,
    "timestep_mean": 517.90625,
    "epsilon_norm": 6.832719802856445,
    "prediction_norm": 6.287952899932861,
    "timestamp": "2025-07-14T12:34:54.005983"
  },
  {
    "epoch": 2227,
    "batch_idx": 64,
    "loss": 0.212731271982193,
    "learning_rate": 3.1450325438770617e-05,
    "timestep_mean": 515.09375,
    "epsilon_norm": 8.405594825744629,
    "prediction_norm": 7.736879825592041,
    "timestamp": "2025-07-14T12:34:54.011344"
  },
  {
    "epoch": 2227,
    "batch_idx": 96,
    "loss": 0.4900147020816803,
    "learning_rate": 3.1450325438770617e-05,
    "timestep_mean": 393.4375,
    "epsilon_norm": 8.555418014526367,
    "prediction_norm": 5.545672416687012,
    "timestamp": "2025-07-14T12:34:54.016606"
  },
  {
    "epoch": 2227,
    "batch_idx": 128,
    "loss": 0.31117820739746094,
    "learning_rate": 3.1450325438770617e-05,
    "timestep_mean": 471.8125,
    "epsilon_norm": 8.329824447631836,
    "prediction_norm": 6.837131977081299,
    "timestamp": "2025-07-14T12:34:54.021943"
  },
  {
    "epoch": 2227,
    "batch_idx": 160,
    "loss": 0.13671427965164185,
    "learning_rate": 3.1450325438770617e-05,
    "timestep_mean": 587.21875,
    "epsilon_norm": 8.505577087402344,
    "prediction_norm": 8.089658737182617,
    "timestamp": "2025-07-14T12:34:54.027213"
  },
  {
    "epoch": 2227,
    "batch_idx": 192,
    "loss": 0.7819651365280151,
    "learning_rate": 3.1450325438770617e-05,
    "timestep_mean": 452.625,
    "epsilon_norm": 4.885234355926514,
    "prediction_norm": 4.000937461853027,
    "timestamp": "2025-07-14T12:34:54.032561"
  },
  {
    "epoch": 2228,
    "batch_idx": 0,
    "loss": 0.19585266709327698,
    "learning_rate": 3.1725056625903895e-05,
    "timestep_mean": 504.59375,
    "epsilon_norm": 7.715805530548096,
    "prediction_norm": 6.937972545623779,
    "timestamp": "2025-07-14T12:34:54.037889"
  },
  {
    "epoch": 2228,
    "batch_idx": 32,
    "loss": 0.3125315308570862,
    "learning_rate": 3.1725056625903895e-05,
    "timestep_mean": 519.3125,
    "epsilon_norm": 7.726285934448242,
    "prediction_norm": 6.408675670623779,
    "timestamp": "2025-07-14T12:34:54.043160"
  },
  {
    "epoch": 2228,
    "batch_idx": 64,
    "loss": 0.17720875144004822,
    "learning_rate": 3.1725056625903895e-05,
    "timestep_mean": 543.21875,
    "epsilon_norm": 8.247659683227539,
    "prediction_norm": 7.266658782958984,
    "timestamp": "2025-07-14T12:34:54.048403"
  },
  {
    "epoch": 2228,
    "batch_idx": 96,
    "loss": 0.19963809847831726,
    "learning_rate": 3.1725056625903895e-05,
    "timestep_mean": 442.0625,
    "epsilon_norm": 7.814929485321045,
    "prediction_norm": 6.635021686553955,
    "timestamp": "2025-07-14T12:34:54.053744"
  },
  {
    "epoch": 2228,
    "batch_idx": 128,
    "loss": 0.3089563250541687,
    "learning_rate": 3.1725056625903895e-05,
    "timestep_mean": 514.4375,
    "epsilon_norm": 9.377185821533203,
    "prediction_norm": 8.296544075012207,
    "timestamp": "2025-07-14T12:34:54.059093"
  },
  {
    "epoch": 2228,
    "batch_idx": 160,
    "loss": 0.23997315764427185,
    "learning_rate": 3.1725056625903895e-05,
    "timestep_mean": 487.875,
    "epsilon_norm": 7.502061367034912,
    "prediction_norm": 6.488371849060059,
    "timestamp": "2025-07-14T12:34:54.064376"
  },
  {
    "epoch": 2228,
    "batch_idx": 192,
    "loss": 0.39189624786376953,
    "learning_rate": 3.1725056625903895e-05,
    "timestep_mean": 364.125,
    "epsilon_norm": 3.6635639667510986,
    "prediction_norm": 3.063375234603882,
    "timestamp": "2025-07-14T12:34:54.069755"
  },
  {
    "epoch": 2229,
    "batch_idx": 0,
    "loss": 0.24522119760513306,
    "learning_rate": 3.200094323490997e-05,
    "timestep_mean": 505.46875,
    "epsilon_norm": 8.189184188842773,
    "prediction_norm": 6.617629051208496,
    "timestamp": "2025-07-14T12:34:54.075215"
  },
  {
    "epoch": 2229,
    "batch_idx": 32,
    "loss": 0.11912688612937927,
    "learning_rate": 3.200094323490997e-05,
    "timestep_mean": 461.9375,
    "epsilon_norm": 8.937309265136719,
    "prediction_norm": 8.188594818115234,
    "timestamp": "2025-07-14T12:34:54.080565"
  },
  {
    "epoch": 2229,
    "batch_idx": 64,
    "loss": 0.34322553873062134,
    "learning_rate": 3.200094323490997e-05,
    "timestep_mean": 488.03125,
    "epsilon_norm": 8.316449165344238,
    "prediction_norm": 6.907989025115967,
    "timestamp": "2025-07-14T12:34:54.085766"
  },
  {
    "epoch": 2229,
    "batch_idx": 96,
    "loss": 0.1540735363960266,
    "learning_rate": 3.200094323490997e-05,
    "timestep_mean": 564.8125,
    "epsilon_norm": 8.725082397460938,
    "prediction_norm": 7.925055027008057,
    "timestamp": "2025-07-14T12:34:54.090975"
  },
  {
    "epoch": 2229,
    "batch_idx": 128,
    "loss": 0.43676894903182983,
    "learning_rate": 3.200094323490997e-05,
    "timestep_mean": 380.65625,
    "epsilon_norm": 8.011820793151855,
    "prediction_norm": 5.658731937408447,
    "timestamp": "2025-07-14T12:34:54.096208"
  },
  {
    "epoch": 2229,
    "batch_idx": 160,
    "loss": 0.21100807189941406,
    "learning_rate": 3.200094323490997e-05,
    "timestep_mean": 436.21875,
    "epsilon_norm": 7.202768325805664,
    "prediction_norm": 6.296750068664551,
    "timestamp": "2025-07-14T12:34:54.101315"
  },
  {
    "epoch": 2229,
    "batch_idx": 192,
    "loss": 0.12621773779392242,
    "learning_rate": 3.200094323490997e-05,
    "timestep_mean": 431.0,
    "epsilon_norm": 4.074199676513672,
    "prediction_norm": 3.844569444656372,
    "timestamp": "2025-07-14T12:34:54.106366"
  },
  {
    "epoch": 2230,
    "batch_idx": 0,
    "loss": 0.18016497790813446,
    "learning_rate": 3.227798458506641e-05,
    "timestep_mean": 603.375,
    "epsilon_norm": 7.38355016708374,
    "prediction_norm": 7.21296501159668,
    "timestamp": "2025-07-14T12:34:54.111351"
  },
  {
    "epoch": 2230,
    "batch_idx": 32,
    "loss": 0.1438663750886917,
    "learning_rate": 3.227798458506641e-05,
    "timestep_mean": 455.8125,
    "epsilon_norm": 7.726790904998779,
    "prediction_norm": 6.666227340698242,
    "timestamp": "2025-07-14T12:34:54.116556"
  },
  {
    "epoch": 2230,
    "batch_idx": 64,
    "loss": 0.1913192868232727,
    "learning_rate": 3.227798458506641e-05,
    "timestep_mean": 535.25,
    "epsilon_norm": 6.684840202331543,
    "prediction_norm": 6.2921905517578125,
    "timestamp": "2025-07-14T12:34:54.121784"
  },
  {
    "epoch": 2230,
    "batch_idx": 96,
    "loss": 0.21573813259601593,
    "learning_rate": 3.227798458506641e-05,
    "timestep_mean": 455.15625,
    "epsilon_norm": 7.415350437164307,
    "prediction_norm": 6.46640157699585,
    "timestamp": "2025-07-14T12:34:54.127023"
  },
  {
    "epoch": 2230,
    "batch_idx": 128,
    "loss": 0.24773117899894714,
    "learning_rate": 3.227798458506641e-05,
    "timestep_mean": 524.28125,
    "epsilon_norm": 9.652990341186523,
    "prediction_norm": 7.917211532592773,
    "timestamp": "2025-07-14T12:34:54.132319"
  },
  {
    "epoch": 2230,
    "batch_idx": 160,
    "loss": 0.3188943564891815,
    "learning_rate": 3.227798458506641e-05,
    "timestep_mean": 534.96875,
    "epsilon_norm": 7.863807201385498,
    "prediction_norm": 6.476862907409668,
    "timestamp": "2025-07-14T12:34:54.137446"
  },
  {
    "epoch": 2230,
    "batch_idx": 192,
    "loss": 0.8794666528701782,
    "learning_rate": 3.227798458506641e-05,
    "timestep_mean": 394.0,
    "epsilon_norm": 4.269529819488525,
    "prediction_norm": 2.682957649230957,
    "timestamp": "2025-07-14T12:34:54.142674"
  },
  {
    "epoch": 2231,
    "batch_idx": 0,
    "loss": 0.18676035106182098,
    "learning_rate": 3.255617999280084e-05,
    "timestep_mean": 530.15625,
    "epsilon_norm": 7.842097759246826,
    "prediction_norm": 7.267899513244629,
    "timestamp": "2025-07-14T12:34:54.147934"
  },
  {
    "epoch": 2231,
    "batch_idx": 32,
    "loss": 0.16384050250053406,
    "learning_rate": 3.255617999280084e-05,
    "timestep_mean": 517.46875,
    "epsilon_norm": 6.380690097808838,
    "prediction_norm": 5.834822177886963,
    "timestamp": "2025-07-14T12:34:54.153197"
  },
  {
    "epoch": 2231,
    "batch_idx": 64,
    "loss": 0.3571212887763977,
    "learning_rate": 3.255617999280084e-05,
    "timestep_mean": 466.96875,
    "epsilon_norm": 8.251877784729004,
    "prediction_norm": 6.6199870109558105,
    "timestamp": "2025-07-14T12:34:54.158377"
  },
  {
    "epoch": 2231,
    "batch_idx": 96,
    "loss": 0.23532506823539734,
    "learning_rate": 3.255617999280084e-05,
    "timestep_mean": 481.03125,
    "epsilon_norm": 6.955652713775635,
    "prediction_norm": 5.9129204750061035,
    "timestamp": "2025-07-14T12:34:54.163547"
  },
  {
    "epoch": 2231,
    "batch_idx": 128,
    "loss": 0.38032394647598267,
    "learning_rate": 3.255617999280084e-05,
    "timestep_mean": 398.71875,
    "epsilon_norm": 7.783803939819336,
    "prediction_norm": 6.400778293609619,
    "timestamp": "2025-07-14T12:34:54.168692"
  },
  {
    "epoch": 2231,
    "batch_idx": 160,
    "loss": 0.377116322517395,
    "learning_rate": 3.255617999280084e-05,
    "timestep_mean": 446.21875,
    "epsilon_norm": 7.9967546463012695,
    "prediction_norm": 6.33218240737915,
    "timestamp": "2025-07-14T12:34:54.173958"
  },
  {
    "epoch": 2231,
    "batch_idx": 192,
    "loss": 0.1378319263458252,
    "learning_rate": 3.255617999280084e-05,
    "timestep_mean": 489.5,
    "epsilon_norm": 4.463637828826904,
    "prediction_norm": 4.380124568939209,
    "timestamp": "2025-07-14T12:34:54.179135"
  },
  {
    "epoch": 2232,
    "batch_idx": 0,
    "loss": 0.3358113169670105,
    "learning_rate": 3.2835528771693965e-05,
    "timestep_mean": 463.15625,
    "epsilon_norm": 8.2366943359375,
    "prediction_norm": 7.934234142303467,
    "timestamp": "2025-07-14T12:34:54.184380"
  },
  {
    "epoch": 2232,
    "batch_idx": 32,
    "loss": 0.5710706114768982,
    "learning_rate": 3.2835528771693965e-05,
    "timestep_mean": 473.9375,
    "epsilon_norm": 8.954916954040527,
    "prediction_norm": 6.819049835205078,
    "timestamp": "2025-07-14T12:34:54.189672"
  },
  {
    "epoch": 2232,
    "batch_idx": 64,
    "loss": 0.377946674823761,
    "learning_rate": 3.2835528771693965e-05,
    "timestep_mean": 435.1875,
    "epsilon_norm": 7.81030797958374,
    "prediction_norm": 5.793206691741943,
    "timestamp": "2025-07-14T12:34:54.194882"
  },
  {
    "epoch": 2232,
    "batch_idx": 96,
    "loss": 0.13784722983837128,
    "learning_rate": 3.2835528771693965e-05,
    "timestep_mean": 584.96875,
    "epsilon_norm": 7.251975059509277,
    "prediction_norm": 6.750802993774414,
    "timestamp": "2025-07-14T12:34:54.200018"
  },
  {
    "epoch": 2232,
    "batch_idx": 128,
    "loss": 0.22009748220443726,
    "learning_rate": 3.2835528771693965e-05,
    "timestep_mean": 579.375,
    "epsilon_norm": 7.956126689910889,
    "prediction_norm": 6.5896687507629395,
    "timestamp": "2025-07-14T12:34:54.205356"
  },
  {
    "epoch": 2232,
    "batch_idx": 160,
    "loss": 0.08239846676588058,
    "learning_rate": 3.2835528771693965e-05,
    "timestep_mean": 553.46875,
    "epsilon_norm": 7.87455940246582,
    "prediction_norm": 7.909854888916016,
    "timestamp": "2025-07-14T12:34:54.210591"
  },
  {
    "epoch": 2232,
    "batch_idx": 192,
    "loss": 0.3171009421348572,
    "learning_rate": 3.2835528771693965e-05,
    "timestep_mean": 544.125,
    "epsilon_norm": 3.2836945056915283,
    "prediction_norm": 3.0766327381134033,
    "timestamp": "2025-07-14T12:34:54.215808"
  },
  {
    "epoch": 2233,
    "batch_idx": 0,
    "loss": 0.15864062309265137,
    "learning_rate": 3.311603023248042e-05,
    "timestep_mean": 517.375,
    "epsilon_norm": 7.14200496673584,
    "prediction_norm": 6.687327861785889,
    "timestamp": "2025-07-14T12:34:54.221069"
  },
  {
    "epoch": 2233,
    "batch_idx": 32,
    "loss": 0.2356976568698883,
    "learning_rate": 3.311603023248042e-05,
    "timestep_mean": 408.3125,
    "epsilon_norm": 7.333929061889648,
    "prediction_norm": 6.10205602645874,
    "timestamp": "2025-07-14T12:34:54.226250"
  },
  {
    "epoch": 2233,
    "batch_idx": 64,
    "loss": 0.2675802409648895,
    "learning_rate": 3.311603023248042e-05,
    "timestep_mean": 514.59375,
    "epsilon_norm": 8.588674545288086,
    "prediction_norm": 7.589129447937012,
    "timestamp": "2025-07-14T12:34:54.231571"
  },
  {
    "epoch": 2233,
    "batch_idx": 96,
    "loss": 0.19491644203662872,
    "learning_rate": 3.311603023248042e-05,
    "timestep_mean": 554.0625,
    "epsilon_norm": 8.155125617980957,
    "prediction_norm": 6.905085563659668,
    "timestamp": "2025-07-14T12:34:54.236816"
  },
  {
    "epoch": 2233,
    "batch_idx": 128,
    "loss": 0.3429412245750427,
    "learning_rate": 3.311603023248042e-05,
    "timestep_mean": 396.90625,
    "epsilon_norm": 7.308386325836182,
    "prediction_norm": 6.332974433898926,
    "timestamp": "2025-07-14T12:34:54.242053"
  },
  {
    "epoch": 2233,
    "batch_idx": 160,
    "loss": 0.2293587625026703,
    "learning_rate": 3.311603023248042e-05,
    "timestep_mean": 589.25,
    "epsilon_norm": 7.877262115478516,
    "prediction_norm": 6.630975723266602,
    "timestamp": "2025-07-14T12:34:54.247309"
  },
  {
    "epoch": 2233,
    "batch_idx": 192,
    "loss": 0.18957240879535675,
    "learning_rate": 3.311603023248042e-05,
    "timestep_mean": 598.375,
    "epsilon_norm": 3.249281883239746,
    "prediction_norm": 3.1490259170532227,
    "timestamp": "2025-07-14T12:34:54.252478"
  },
  {
    "epoch": 2234,
    "batch_idx": 0,
    "loss": 0.24671271443367004,
    "learning_rate": 3.3397683683050656e-05,
    "timestep_mean": 506.78125,
    "epsilon_norm": 8.164361000061035,
    "prediction_norm": 6.790595531463623,
    "timestamp": "2025-07-14T12:34:54.257715"
  },
  {
    "epoch": 2234,
    "batch_idx": 32,
    "loss": 0.18613366782665253,
    "learning_rate": 3.3397683683050656e-05,
    "timestep_mean": 539.125,
    "epsilon_norm": 8.009099006652832,
    "prediction_norm": 7.389301300048828,
    "timestamp": "2025-07-14T12:34:54.262914"
  },
  {
    "epoch": 2234,
    "batch_idx": 64,
    "loss": 0.35818207263946533,
    "learning_rate": 3.3397683683050656e-05,
    "timestep_mean": 509.90625,
    "epsilon_norm": 9.18249797821045,
    "prediction_norm": 7.849409580230713,
    "timestamp": "2025-07-14T12:34:54.268065"
  },
  {
    "epoch": 2234,
    "batch_idx": 96,
    "loss": 0.19324937462806702,
    "learning_rate": 3.3397683683050656e-05,
    "timestep_mean": 499.9375,
    "epsilon_norm": 6.674978733062744,
    "prediction_norm": 5.713037967681885,
    "timestamp": "2025-07-14T12:34:54.273322"
  },
  {
    "epoch": 2234,
    "batch_idx": 128,
    "loss": 0.27772772312164307,
    "learning_rate": 3.3397683683050656e-05,
    "timestep_mean": 402.1875,
    "epsilon_norm": 7.3824920654296875,
    "prediction_norm": 6.015865802764893,
    "timestamp": "2025-07-14T12:34:54.278477"
  },
  {
    "epoch": 2234,
    "batch_idx": 160,
    "loss": 0.278206467628479,
    "learning_rate": 3.3397683683050656e-05,
    "timestep_mean": 534.53125,
    "epsilon_norm": 7.953442573547363,
    "prediction_norm": 6.704947471618652,
    "timestamp": "2025-07-14T12:34:54.283590"
  },
  {
    "epoch": 2234,
    "batch_idx": 192,
    "loss": 0.3150547444820404,
    "learning_rate": 3.3397683683050656e-05,
    "timestep_mean": 524.75,
    "epsilon_norm": 3.0290565490722656,
    "prediction_norm": 2.328510046005249,
    "timestamp": "2025-07-14T12:34:54.288907"
  },
  {
    "epoch": 2235,
    "batch_idx": 0,
    "loss": 0.22718477249145508,
    "learning_rate": 3.368048842845298e-05,
    "timestep_mean": 542.375,
    "epsilon_norm": 8.362796783447266,
    "prediction_norm": 7.509525775909424,
    "timestamp": "2025-07-14T12:34:54.294123"
  },
  {
    "epoch": 2235,
    "batch_idx": 32,
    "loss": 0.26147088408470154,
    "learning_rate": 3.368048842845298e-05,
    "timestep_mean": 467.9375,
    "epsilon_norm": 9.102692604064941,
    "prediction_norm": 8.084898948669434,
    "timestamp": "2025-07-14T12:34:54.299331"
  },
  {
    "epoch": 2235,
    "batch_idx": 64,
    "loss": 0.32414504885673523,
    "learning_rate": 3.368048842845298e-05,
    "timestep_mean": 492.9375,
    "epsilon_norm": 8.448881149291992,
    "prediction_norm": 7.314060688018799,
    "timestamp": "2025-07-14T12:34:54.304479"
  },
  {
    "epoch": 2235,
    "batch_idx": 96,
    "loss": 0.1884695589542389,
    "learning_rate": 3.368048842845298e-05,
    "timestep_mean": 494.625,
    "epsilon_norm": 9.003670692443848,
    "prediction_norm": 7.723452091217041,
    "timestamp": "2025-07-14T12:34:54.309636"
  },
  {
    "epoch": 2235,
    "batch_idx": 128,
    "loss": 0.21212337911128998,
    "learning_rate": 3.368048842845298e-05,
    "timestep_mean": 484.75,
    "epsilon_norm": 8.092551231384277,
    "prediction_norm": 7.0006103515625,
    "timestamp": "2025-07-14T12:34:54.314868"
  },
  {
    "epoch": 2235,
    "batch_idx": 160,
    "loss": 0.2278386801481247,
    "learning_rate": 3.368048842845298e-05,
    "timestep_mean": 517.9375,
    "epsilon_norm": 7.96276330947876,
    "prediction_norm": 7.168148517608643,
    "timestamp": "2025-07-14T12:34:54.320077"
  },
  {
    "epoch": 2235,
    "batch_idx": 192,
    "loss": 0.067511647939682,
    "learning_rate": 3.368048842845298e-05,
    "timestep_mean": 613.75,
    "epsilon_norm": 2.653123617172241,
    "prediction_norm": 2.3992483615875244,
    "timestamp": "2025-07-14T12:34:54.325184"
  },
  {
    "epoch": 2236,
    "batch_idx": 0,
    "loss": 0.2596079707145691,
    "learning_rate": 3.396444377089456e-05,
    "timestep_mean": 513.09375,
    "epsilon_norm": 9.012786865234375,
    "prediction_norm": 7.37339448928833,
    "timestamp": "2025-07-14T12:34:54.330189"
  },
  {
    "epoch": 2236,
    "batch_idx": 32,
    "loss": 0.16808068752288818,
    "learning_rate": 3.396444377089456e-05,
    "timestep_mean": 461.53125,
    "epsilon_norm": 7.306617736816406,
    "prediction_norm": 6.405406951904297,
    "timestamp": "2025-07-14T12:34:54.335025"
  },
  {
    "epoch": 2236,
    "batch_idx": 64,
    "loss": 0.29477617144584656,
    "learning_rate": 3.396444377089456e-05,
    "timestep_mean": 409.875,
    "epsilon_norm": 7.714208126068115,
    "prediction_norm": 6.514294147491455,
    "timestamp": "2025-07-14T12:34:54.339920"
  },
  {
    "epoch": 2236,
    "batch_idx": 96,
    "loss": 0.1839294731616974,
    "learning_rate": 3.396444377089456e-05,
    "timestep_mean": 624.59375,
    "epsilon_norm": 9.375025749206543,
    "prediction_norm": 8.873270034790039,
    "timestamp": "2025-07-14T12:34:54.344777"
  },
  {
    "epoch": 2236,
    "batch_idx": 128,
    "loss": 0.15421637892723083,
    "learning_rate": 3.396444377089456e-05,
    "timestep_mean": 497.5625,
    "epsilon_norm": 6.667425155639648,
    "prediction_norm": 5.4203033447265625,
    "timestamp": "2025-07-14T12:34:54.349666"
  },
  {
    "epoch": 2236,
    "batch_idx": 160,
    "loss": 0.15043146908283234,
    "learning_rate": 3.396444377089456e-05,
    "timestep_mean": 495.875,
    "epsilon_norm": 8.071182250976562,
    "prediction_norm": 7.377492904663086,
    "timestamp": "2025-07-14T12:34:54.354841"
  },
  {
    "epoch": 2236,
    "batch_idx": 192,
    "loss": 0.08401845395565033,
    "learning_rate": 3.396444377089456e-05,
    "timestep_mean": 670.0,
    "epsilon_norm": 3.541943311691284,
    "prediction_norm": 3.1691486835479736,
    "timestamp": "2025-07-14T12:34:54.359780"
  },
  {
    "epoch": 2237,
    "batch_idx": 0,
    "loss": 0.24435359239578247,
    "learning_rate": 3.424954900974384e-05,
    "timestep_mean": 492.46875,
    "epsilon_norm": 9.008461952209473,
    "prediction_norm": 7.002222537994385,
    "timestamp": "2025-07-14T12:34:54.364948"
  },
  {
    "epoch": 2237,
    "batch_idx": 32,
    "loss": 0.4647822976112366,
    "learning_rate": 3.424954900974384e-05,
    "timestep_mean": 524.65625,
    "epsilon_norm": 8.850948333740234,
    "prediction_norm": 6.642287731170654,
    "timestamp": "2025-07-14T12:34:54.370092"
  },
  {
    "epoch": 2237,
    "batch_idx": 64,
    "loss": 0.3304463028907776,
    "learning_rate": 3.424954900974384e-05,
    "timestep_mean": 467.8125,
    "epsilon_norm": 8.248984336853027,
    "prediction_norm": 7.442234516143799,
    "timestamp": "2025-07-14T12:34:54.375265"
  },
  {
    "epoch": 2237,
    "batch_idx": 96,
    "loss": 0.2623928189277649,
    "learning_rate": 3.424954900974384e-05,
    "timestep_mean": 456.59375,
    "epsilon_norm": 7.391290664672852,
    "prediction_norm": 6.453682899475098,
    "timestamp": "2025-07-14T12:34:54.380437"
  },
  {
    "epoch": 2237,
    "batch_idx": 128,
    "loss": 0.2063201367855072,
    "learning_rate": 3.424954900974384e-05,
    "timestep_mean": 562.1875,
    "epsilon_norm": 7.417323589324951,
    "prediction_norm": 6.746199131011963,
    "timestamp": "2025-07-14T12:34:54.385501"
  },
  {
    "epoch": 2237,
    "batch_idx": 160,
    "loss": 0.3744586110115051,
    "learning_rate": 3.424954900974384e-05,
    "timestep_mean": 458.9375,
    "epsilon_norm": 8.199090003967285,
    "prediction_norm": 6.734641075134277,
    "timestamp": "2025-07-14T12:34:54.390583"
  },
  {
    "epoch": 2237,
    "batch_idx": 192,
    "loss": 0.37812623381614685,
    "learning_rate": 3.424954900974384e-05,
    "timestep_mean": 411.5,
    "epsilon_norm": 4.4198899269104,
    "prediction_norm": 3.3181936740875244,
    "timestamp": "2025-07-14T12:34:54.395652"
  },
  {
    "epoch": 2238,
    "batch_idx": 0,
    "loss": 0.2432897835969925,
    "learning_rate": 3.4535803441532096e-05,
    "timestep_mean": 462.03125,
    "epsilon_norm": 7.112498760223389,
    "prediction_norm": 5.762885570526123,
    "timestamp": "2025-07-14T12:34:54.400842"
  },
  {
    "epoch": 2238,
    "batch_idx": 32,
    "loss": 0.6010847091674805,
    "learning_rate": 3.4535803441532096e-05,
    "timestep_mean": 466.15625,
    "epsilon_norm": 9.064562797546387,
    "prediction_norm": 6.49866247177124,
    "timestamp": "2025-07-14T12:34:54.405846"
  },
  {
    "epoch": 2238,
    "batch_idx": 64,
    "loss": 0.3150775134563446,
    "learning_rate": 3.4535803441532096e-05,
    "timestep_mean": 531.28125,
    "epsilon_norm": 8.064898490905762,
    "prediction_norm": 6.559620380401611,
    "timestamp": "2025-07-14T12:34:54.410698"
  },
  {
    "epoch": 2238,
    "batch_idx": 96,
    "loss": 0.37331509590148926,
    "learning_rate": 3.4535803441532096e-05,
    "timestep_mean": 544.09375,
    "epsilon_norm": 8.462730407714844,
    "prediction_norm": 7.285738468170166,
    "timestamp": "2025-07-14T12:34:54.415647"
  },
  {
    "epoch": 2238,
    "batch_idx": 128,
    "loss": 0.22142446041107178,
    "learning_rate": 3.4535803441532096e-05,
    "timestep_mean": 562.375,
    "epsilon_norm": 7.934715747833252,
    "prediction_norm": 6.640902996063232,
    "timestamp": "2025-07-14T12:34:54.420920"
  },
  {
    "epoch": 2238,
    "batch_idx": 160,
    "loss": 0.21162766218185425,
    "learning_rate": 3.4535803441532096e-05,
    "timestep_mean": 514.9375,
    "epsilon_norm": 6.968100070953369,
    "prediction_norm": 6.043918132781982,
    "timestamp": "2025-07-14T12:34:54.426113"
  },
  {
    "epoch": 2238,
    "batch_idx": 192,
    "loss": 0.026423228904604912,
    "learning_rate": 3.4535803441532096e-05,
    "timestep_mean": 616.5,
    "epsilon_norm": 2.8020071983337402,
    "prediction_norm": 2.6263437271118164,
    "timestamp": "2025-07-14T12:34:54.431442"
  },
  {
    "epoch": 2239,
    "batch_idx": 0,
    "loss": 0.5778588056564331,
    "learning_rate": 3.482320635995504e-05,
    "timestep_mean": 381.25,
    "epsilon_norm": 8.554940223693848,
    "prediction_norm": 6.121712684631348,
    "timestamp": "2025-07-14T12:34:54.436689"
  },
  {
    "epoch": 2239,
    "batch_idx": 32,
    "loss": 0.35998186469078064,
    "learning_rate": 3.482320635995504e-05,
    "timestep_mean": 436.125,
    "epsilon_norm": 7.216463088989258,
    "prediction_norm": 5.370848178863525,
    "timestamp": "2025-07-14T12:34:54.441918"
  },
  {
    "epoch": 2239,
    "batch_idx": 64,
    "loss": 0.44744789600372314,
    "learning_rate": 3.482320635995504e-05,
    "timestep_mean": 446.40625,
    "epsilon_norm": 7.612955093383789,
    "prediction_norm": 5.634605407714844,
    "timestamp": "2025-07-14T12:34:54.446967"
  },
  {
    "epoch": 2239,
    "batch_idx": 96,
    "loss": 0.2903003394603729,
    "learning_rate": 3.482320635995504e-05,
    "timestep_mean": 444.46875,
    "epsilon_norm": 6.636932849884033,
    "prediction_norm": 5.222820281982422,
    "timestamp": "2025-07-14T12:34:54.452014"
  },
  {
    "epoch": 2239,
    "batch_idx": 128,
    "loss": 0.3168455958366394,
    "learning_rate": 3.482320635995504e-05,
    "timestep_mean": 477.71875,
    "epsilon_norm": 9.15849494934082,
    "prediction_norm": 7.486791610717773,
    "timestamp": "2025-07-14T12:34:54.457212"
  },
  {
    "epoch": 2239,
    "batch_idx": 160,
    "loss": 0.26722049713134766,
    "learning_rate": 3.482320635995504e-05,
    "timestep_mean": 553.8125,
    "epsilon_norm": 9.214658737182617,
    "prediction_norm": 7.329102993011475,
    "timestamp": "2025-07-14T12:34:54.462310"
  },
  {
    "epoch": 2239,
    "batch_idx": 192,
    "loss": 0.3478684425354004,
    "learning_rate": 3.482320635995504e-05,
    "timestep_mean": 446.875,
    "epsilon_norm": 4.1789727210998535,
    "prediction_norm": 2.661884307861328,
    "timestamp": "2025-07-14T12:34:54.467628"
  },
  {
    "epoch": 2240,
    "batch_idx": 0,
    "loss": 0.4000331461429596,
    "learning_rate": 3.5111757055874305e-05,
    "timestep_mean": 476.0625,
    "epsilon_norm": 8.097359657287598,
    "prediction_norm": 5.912138938903809,
    "timestamp": "2025-07-14T12:34:54.472812"
  },
  {
    "epoch": 2240,
    "batch_idx": 32,
    "loss": 0.30885016918182373,
    "learning_rate": 3.5111757055874305e-05,
    "timestep_mean": 513.125,
    "epsilon_norm": 8.782479286193848,
    "prediction_norm": 7.4273881912231445,
    "timestamp": "2025-07-14T12:34:54.477714"
  },
  {
    "epoch": 2240,
    "batch_idx": 64,
    "loss": 0.23365581035614014,
    "learning_rate": 3.5111757055874305e-05,
    "timestep_mean": 522.71875,
    "epsilon_norm": 8.465680122375488,
    "prediction_norm": 7.92960262298584,
    "timestamp": "2025-07-14T12:34:54.482630"
  },
  {
    "epoch": 2240,
    "batch_idx": 96,
    "loss": 0.25984564423561096,
    "learning_rate": 3.5111757055874305e-05,
    "timestep_mean": 498.6875,
    "epsilon_norm": 8.905847549438477,
    "prediction_norm": 7.76190710067749,
    "timestamp": "2025-07-14T12:34:54.487495"
  },
  {
    "epoch": 2240,
    "batch_idx": 128,
    "loss": 0.4964340925216675,
    "learning_rate": 3.5111757055874305e-05,
    "timestep_mean": 477.65625,
    "epsilon_norm": 9.100629806518555,
    "prediction_norm": 6.923920631408691,
    "timestamp": "2025-07-14T12:34:54.492495"
  },
  {
    "epoch": 2240,
    "batch_idx": 160,
    "loss": 0.2924179434776306,
    "learning_rate": 3.5111757055874305e-05,
    "timestep_mean": 450.625,
    "epsilon_norm": 6.601587295532227,
    "prediction_norm": 5.7440385818481445,
    "timestamp": "2025-07-14T12:34:54.497351"
  },
  {
    "epoch": 2240,
    "batch_idx": 192,
    "loss": 0.6948790550231934,
    "learning_rate": 3.5111757055874305e-05,
    "timestep_mean": 178.0,
    "epsilon_norm": 2.9896764755249023,
    "prediction_norm": 2.1172966957092285,
    "timestamp": "2025-07-14T12:34:54.502229"
  },
  {
    "epoch": 2241,
    "batch_idx": 0,
    "loss": 0.2485763132572174,
    "learning_rate": 3.540145481731991e-05,
    "timestep_mean": 493.09375,
    "epsilon_norm": 7.753688812255859,
    "prediction_norm": 6.764188766479492,
    "timestamp": "2025-07-14T12:34:54.507135"
  },
  {
    "epoch": 2241,
    "batch_idx": 32,
    "loss": 0.3653874397277832,
    "learning_rate": 3.540145481731991e-05,
    "timestep_mean": 406.0625,
    "epsilon_norm": 7.786439418792725,
    "prediction_norm": 5.942874431610107,
    "timestamp": "2025-07-14T12:34:54.512091"
  },
  {
    "epoch": 2241,
    "batch_idx": 64,
    "loss": 0.30241629481315613,
    "learning_rate": 3.540145481731991e-05,
    "timestep_mean": 585.375,
    "epsilon_norm": 8.222394943237305,
    "prediction_norm": 7.645702838897705,
    "timestamp": "2025-07-14T12:34:54.517022"
  },
  {
    "epoch": 2241,
    "batch_idx": 96,
    "loss": 0.2822916507720947,
    "learning_rate": 3.540145481731991e-05,
    "timestep_mean": 499.25,
    "epsilon_norm": 7.845022201538086,
    "prediction_norm": 6.2147393226623535,
    "timestamp": "2025-07-14T12:34:54.522015"
  },
  {
    "epoch": 2241,
    "batch_idx": 128,
    "loss": 0.28105610609054565,
    "learning_rate": 3.540145481731991e-05,
    "timestep_mean": 444.5625,
    "epsilon_norm": 8.256911277770996,
    "prediction_norm": 7.31200647354126,
    "timestamp": "2025-07-14T12:34:54.526839"
  },
  {
    "epoch": 2241,
    "batch_idx": 160,
    "loss": 0.08051231503486633,
    "learning_rate": 3.540145481731991e-05,
    "timestep_mean": 604.5,
    "epsilon_norm": 7.093473434448242,
    "prediction_norm": 7.108767032623291,
    "timestamp": "2025-07-14T12:34:54.531731"
  },
  {
    "epoch": 2241,
    "batch_idx": 192,
    "loss": 0.16263964772224426,
    "learning_rate": 3.540145481731991e-05,
    "timestep_mean": 516.875,
    "epsilon_norm": 3.535644054412842,
    "prediction_norm": 2.7200441360473633,
    "timestamp": "2025-07-14T12:34:54.536652"
  },
  {
    "epoch": 2242,
    "batch_idx": 0,
    "loss": 0.2536538243293762,
    "learning_rate": 3.569229892949142e-05,
    "timestep_mean": 518.5625,
    "epsilon_norm": 8.508498191833496,
    "prediction_norm": 7.6753435134887695,
    "timestamp": "2025-07-14T12:34:54.541527"
  },
  {
    "epoch": 2242,
    "batch_idx": 32,
    "loss": 0.28184086084365845,
    "learning_rate": 3.569229892949142e-05,
    "timestep_mean": 444.34375,
    "epsilon_norm": 7.622731685638428,
    "prediction_norm": 5.671923637390137,
    "timestamp": "2025-07-14T12:34:54.546752"
  },
  {
    "epoch": 2242,
    "batch_idx": 64,
    "loss": 0.08973803371191025,
    "learning_rate": 3.569229892949142e-05,
    "timestep_mean": 541.46875,
    "epsilon_norm": 6.555177688598633,
    "prediction_norm": 6.343901634216309,
    "timestamp": "2025-07-14T12:34:54.551916"
  },
  {
    "epoch": 2242,
    "batch_idx": 96,
    "loss": 0.2917419672012329,
    "learning_rate": 3.569229892949142e-05,
    "timestep_mean": 430.375,
    "epsilon_norm": 6.988022327423096,
    "prediction_norm": 4.842586517333984,
    "timestamp": "2025-07-14T12:34:54.556998"
  },
  {
    "epoch": 2242,
    "batch_idx": 128,
    "loss": 0.2333815097808838,
    "learning_rate": 3.569229892949142e-05,
    "timestep_mean": 416.03125,
    "epsilon_norm": 8.01762866973877,
    "prediction_norm": 6.803735733032227,
    "timestamp": "2025-07-14T12:34:54.562084"
  },
  {
    "epoch": 2242,
    "batch_idx": 160,
    "loss": 0.2397238314151764,
    "learning_rate": 3.569229892949142e-05,
    "timestep_mean": 526.84375,
    "epsilon_norm": 6.7600626945495605,
    "prediction_norm": 5.42765474319458,
    "timestamp": "2025-07-14T12:34:54.567301"
  },
  {
    "epoch": 2242,
    "batch_idx": 192,
    "loss": 0.7802858352661133,
    "learning_rate": 3.569229892949142e-05,
    "timestep_mean": 361.125,
    "epsilon_norm": 4.4346699714660645,
    "prediction_norm": 2.4622690677642822,
    "timestamp": "2025-07-14T12:34:54.572432"
  },
  {
    "epoch": 2243,
    "batch_idx": 0,
    "loss": 0.11703721433877945,
    "learning_rate": 3.598428867475971e-05,
    "timestep_mean": 578.5,
    "epsilon_norm": 7.358001232147217,
    "prediction_norm": 6.997682571411133,
    "timestamp": "2025-07-14T12:34:54.577533"
  },
  {
    "epoch": 2243,
    "batch_idx": 32,
    "loss": 0.3245941400527954,
    "learning_rate": 3.598428867475971e-05,
    "timestep_mean": 546.3125,
    "epsilon_norm": 7.853048801422119,
    "prediction_norm": 6.159085750579834,
    "timestamp": "2025-07-14T12:34:54.582718"
  },
  {
    "epoch": 2243,
    "batch_idx": 64,
    "loss": 0.29851987957954407,
    "learning_rate": 3.598428867475971e-05,
    "timestep_mean": 447.09375,
    "epsilon_norm": 7.400120735168457,
    "prediction_norm": 5.30532169342041,
    "timestamp": "2025-07-14T12:34:54.587878"
  },
  {
    "epoch": 2243,
    "batch_idx": 96,
    "loss": 0.3788057863712311,
    "learning_rate": 3.598428867475971e-05,
    "timestep_mean": 457.28125,
    "epsilon_norm": 8.079605102539062,
    "prediction_norm": 5.986160755157471,
    "timestamp": "2025-07-14T12:34:54.593056"
  },
  {
    "epoch": 2243,
    "batch_idx": 128,
    "loss": 0.3339563310146332,
    "learning_rate": 3.598428867475971e-05,
    "timestep_mean": 522.6875,
    "epsilon_norm": 8.034605979919434,
    "prediction_norm": 6.600541591644287,
    "timestamp": "2025-07-14T12:34:54.598236"
  },
  {
    "epoch": 2243,
    "batch_idx": 160,
    "loss": 0.2394244372844696,
    "learning_rate": 3.598428867475971e-05,
    "timestep_mean": 484.65625,
    "epsilon_norm": 8.390380859375,
    "prediction_norm": 7.424227714538574,
    "timestamp": "2025-07-14T12:34:54.603375"
  },
  {
    "epoch": 2243,
    "batch_idx": 192,
    "loss": 0.6715225577354431,
    "learning_rate": 3.598428867475971e-05,
    "timestep_mean": 459.5,
    "epsilon_norm": 4.523532867431641,
    "prediction_norm": 3.5542705059051514,
    "timestamp": "2025-07-14T12:34:54.608617"
  },
  {
    "epoch": 2244,
    "batch_idx": 0,
    "loss": 0.25567197799682617,
    "learning_rate": 3.627742333266941e-05,
    "timestep_mean": 437.1875,
    "epsilon_norm": 7.844997406005859,
    "prediction_norm": 6.550068378448486,
    "timestamp": "2025-07-14T12:34:54.613795"
  },
  {
    "epoch": 2244,
    "batch_idx": 32,
    "loss": 0.11945488303899765,
    "learning_rate": 3.627742333266941e-05,
    "timestep_mean": 523.125,
    "epsilon_norm": 7.409799575805664,
    "prediction_norm": 6.729986667633057,
    "timestamp": "2025-07-14T12:34:54.619039"
  },
  {
    "epoch": 2244,
    "batch_idx": 64,
    "loss": 0.4344465136528015,
    "learning_rate": 3.627742333266941e-05,
    "timestep_mean": 519.3125,
    "epsilon_norm": 8.063835144042969,
    "prediction_norm": 6.3706583976745605,
    "timestamp": "2025-07-14T12:34:54.624139"
  },
  {
    "epoch": 2244,
    "batch_idx": 96,
    "loss": 0.24888384342193604,
    "learning_rate": 3.627742333266941e-05,
    "timestep_mean": 527.28125,
    "epsilon_norm": 8.001361846923828,
    "prediction_norm": 7.160218715667725,
    "timestamp": "2025-07-14T12:34:54.629272"
  },
  {
    "epoch": 2244,
    "batch_idx": 128,
    "loss": 0.5435963869094849,
    "learning_rate": 3.627742333266941e-05,
    "timestep_mean": 367.15625,
    "epsilon_norm": 9.512633323669434,
    "prediction_norm": 6.812366008758545,
    "timestamp": "2025-07-14T12:34:54.634484"
  },
  {
    "epoch": 2244,
    "batch_idx": 160,
    "loss": 0.35245752334594727,
    "learning_rate": 3.627742333266941e-05,
    "timestep_mean": 548.53125,
    "epsilon_norm": 7.94796085357666,
    "prediction_norm": 5.92977237701416,
    "timestamp": "2025-07-14T12:34:54.639680"
  },
  {
    "epoch": 2244,
    "batch_idx": 192,
    "loss": 0.1945418417453766,
    "learning_rate": 3.627742333266941e-05,
    "timestep_mean": 434.25,
    "epsilon_norm": 3.5506434440612793,
    "prediction_norm": 2.729224920272827,
    "timestamp": "2025-07-14T12:34:54.644799"
  },
  {
    "epoch": 2245,
    "batch_idx": 0,
    "loss": 0.236861914396286,
    "learning_rate": 3.657170217993953e-05,
    "timestep_mean": 471.96875,
    "epsilon_norm": 8.668155670166016,
    "prediction_norm": 7.762551307678223,
    "timestamp": "2025-07-14T12:34:54.649972"
  },
  {
    "epoch": 2245,
    "batch_idx": 32,
    "loss": 0.28502124547958374,
    "learning_rate": 3.657170217993953e-05,
    "timestep_mean": 554.78125,
    "epsilon_norm": 8.950772285461426,
    "prediction_norm": 7.3418402671813965,
    "timestamp": "2025-07-14T12:34:54.655364"
  },
  {
    "epoch": 2245,
    "batch_idx": 64,
    "loss": 0.27343568205833435,
    "learning_rate": 3.657170217993953e-05,
    "timestep_mean": 495.75,
    "epsilon_norm": 7.797773838043213,
    "prediction_norm": 7.37345027923584,
    "timestamp": "2025-07-14T12:34:54.660519"
  },
  {
    "epoch": 2245,
    "batch_idx": 96,
    "loss": 0.1657916009426117,
    "learning_rate": 3.657170217993953e-05,
    "timestep_mean": 504.5,
    "epsilon_norm": 7.249541759490967,
    "prediction_norm": 6.633604526519775,
    "timestamp": "2025-07-14T12:34:54.665709"
  },
  {
    "epoch": 2245,
    "batch_idx": 128,
    "loss": 0.2197287678718567,
    "learning_rate": 3.657170217993953e-05,
    "timestep_mean": 524.9375,
    "epsilon_norm": 7.837684631347656,
    "prediction_norm": 6.85270357131958,
    "timestamp": "2025-07-14T12:34:54.670935"
  },
  {
    "epoch": 2245,
    "batch_idx": 160,
    "loss": 0.30829057097435,
    "learning_rate": 3.657170217993953e-05,
    "timestep_mean": 508.25,
    "epsilon_norm": 7.874268531799316,
    "prediction_norm": 6.587205410003662,
    "timestamp": "2025-07-14T12:34:54.676113"
  },
  {
    "epoch": 2245,
    "batch_idx": 192,
    "loss": 0.1384146511554718,
    "learning_rate": 3.657170217993953e-05,
    "timestep_mean": 600.0,
    "epsilon_norm": 4.110231876373291,
    "prediction_norm": 4.017587184906006,
    "timestamp": "2025-07-14T12:34:54.681116"
  },
  {
    "epoch": 2246,
    "batch_idx": 0,
    "loss": 0.11824025213718414,
    "learning_rate": 3.686712449046668e-05,
    "timestep_mean": 552.25,
    "epsilon_norm": 7.207492828369141,
    "prediction_norm": 6.505537033081055,
    "timestamp": "2025-07-14T12:34:54.686268"
  },
  {
    "epoch": 2246,
    "batch_idx": 32,
    "loss": 0.24929678440093994,
    "learning_rate": 3.686712449046668e-05,
    "timestep_mean": 483.5625,
    "epsilon_norm": 8.317717552185059,
    "prediction_norm": 6.910686016082764,
    "timestamp": "2025-07-14T12:34:54.691414"
  },
  {
    "epoch": 2246,
    "batch_idx": 64,
    "loss": 0.21929217875003815,
    "learning_rate": 3.686712449046668e-05,
    "timestep_mean": 565.78125,
    "epsilon_norm": 8.813887596130371,
    "prediction_norm": 7.311326026916504,
    "timestamp": "2025-07-14T12:34:54.696722"
  },
  {
    "epoch": 2246,
    "batch_idx": 96,
    "loss": 0.19742530584335327,
    "learning_rate": 3.686712449046668e-05,
    "timestep_mean": 499.96875,
    "epsilon_norm": 6.945078372955322,
    "prediction_norm": 6.301903247833252,
    "timestamp": "2025-07-14T12:34:54.701864"
  },
  {
    "epoch": 2246,
    "batch_idx": 128,
    "loss": 0.2687329053878784,
    "learning_rate": 3.686712449046668e-05,
    "timestep_mean": 523.40625,
    "epsilon_norm": 7.986067295074463,
    "prediction_norm": 7.0191731452941895,
    "timestamp": "2025-07-14T12:34:54.706991"
  },
  {
    "epoch": 2246,
    "batch_idx": 160,
    "loss": 0.25693997740745544,
    "learning_rate": 3.686712449046668e-05,
    "timestep_mean": 485.1875,
    "epsilon_norm": 8.067399024963379,
    "prediction_norm": 7.341372489929199,
    "timestamp": "2025-07-14T12:34:54.712264"
  },
  {
    "epoch": 2246,
    "batch_idx": 192,
    "loss": 0.07019909471273422,
    "learning_rate": 3.686712449046668e-05,
    "timestep_mean": 575.875,
    "epsilon_norm": 3.6341254711151123,
    "prediction_norm": 3.5998694896698,
    "timestamp": "2025-07-14T12:34:54.717598"
  },
  {
    "epoch": 2247,
    "batch_idx": 0,
    "loss": 0.4395676851272583,
    "learning_rate": 3.7163689535325434e-05,
    "timestep_mean": 461.125,
    "epsilon_norm": 8.85250473022461,
    "prediction_norm": 6.100472927093506,
    "timestamp": "2025-07-14T12:34:54.722766"
  },
  {
    "epoch": 2247,
    "batch_idx": 32,
    "loss": 0.4113169312477112,
    "learning_rate": 3.7163689535325434e-05,
    "timestep_mean": 507.4375,
    "epsilon_norm": 8.856773376464844,
    "prediction_norm": 6.335731029510498,
    "timestamp": "2025-07-14T12:34:54.727931"
  },
  {
    "epoch": 2247,
    "batch_idx": 64,
    "loss": 0.5833123326301575,
    "learning_rate": 3.7163689535325434e-05,
    "timestep_mean": 498.625,
    "epsilon_norm": 9.523521423339844,
    "prediction_norm": 6.865653991699219,
    "timestamp": "2025-07-14T12:34:54.733053"
  },
  {
    "epoch": 2247,
    "batch_idx": 96,
    "loss": 0.11707447469234467,
    "learning_rate": 3.7163689535325434e-05,
    "timestep_mean": 537.46875,
    "epsilon_norm": 8.173850059509277,
    "prediction_norm": 7.359706878662109,
    "timestamp": "2025-07-14T12:34:54.738243"
  },
  {
    "epoch": 2247,
    "batch_idx": 128,
    "loss": 0.11836274713277817,
    "learning_rate": 3.7163689535325434e-05,
    "timestep_mean": 500.5625,
    "epsilon_norm": 7.545854568481445,
    "prediction_norm": 7.5511088371276855,
    "timestamp": "2025-07-14T12:34:54.743255"
  },
  {
    "epoch": 2247,
    "batch_idx": 160,
    "loss": 0.3476064205169678,
    "learning_rate": 3.7163689535325434e-05,
    "timestep_mean": 531.53125,
    "epsilon_norm": 8.044357299804688,
    "prediction_norm": 6.231298923492432,
    "timestamp": "2025-07-14T12:34:54.748442"
  },
  {
    "epoch": 2247,
    "batch_idx": 192,
    "loss": 0.11262296140193939,
    "learning_rate": 3.7163689535325434e-05,
    "timestep_mean": 563.75,
    "epsilon_norm": 4.1364593505859375,
    "prediction_norm": 3.5840137004852295,
    "timestamp": "2025-07-14T12:34:54.753657"
  },
  {
    "epoch": 2248,
    "batch_idx": 0,
    "loss": 0.2064899504184723,
    "learning_rate": 3.746139658277096e-05,
    "timestep_mean": 467.09375,
    "epsilon_norm": 5.952351093292236,
    "prediction_norm": 5.11627721786499,
    "timestamp": "2025-07-14T12:34:54.758820"
  },
  {
    "epoch": 2248,
    "batch_idx": 32,
    "loss": 0.46277689933776855,
    "learning_rate": 3.746139658277096e-05,
    "timestep_mean": 533.1875,
    "epsilon_norm": 9.105569839477539,
    "prediction_norm": 6.615347862243652,
    "timestamp": "2025-07-14T12:34:54.764034"
  },
  {
    "epoch": 2248,
    "batch_idx": 64,
    "loss": 0.22546371817588806,
    "learning_rate": 3.746139658277096e-05,
    "timestep_mean": 581.875,
    "epsilon_norm": 7.6203436851501465,
    "prediction_norm": 6.881222724914551,
    "timestamp": "2025-07-14T12:34:54.769227"
  },
  {
    "epoch": 2248,
    "batch_idx": 96,
    "loss": 0.14175966382026672,
    "learning_rate": 3.746139658277096e-05,
    "timestep_mean": 473.1875,
    "epsilon_norm": 7.685400485992432,
    "prediction_norm": 7.028383255004883,
    "timestamp": "2025-07-14T12:34:54.774274"
  },
  {
    "epoch": 2248,
    "batch_idx": 128,
    "loss": 0.21350261569023132,
    "learning_rate": 3.746139658277096e-05,
    "timestep_mean": 599.53125,
    "epsilon_norm": 7.893570899963379,
    "prediction_norm": 6.74318790435791,
    "timestamp": "2025-07-14T12:34:54.779514"
  },
  {
    "epoch": 2248,
    "batch_idx": 160,
    "loss": 0.28898605704307556,
    "learning_rate": 3.746139658277096e-05,
    "timestep_mean": 538.59375,
    "epsilon_norm": 8.286945343017578,
    "prediction_norm": 6.86256217956543,
    "timestamp": "2025-07-14T12:34:54.784728"
  },
  {
    "epoch": 2248,
    "batch_idx": 192,
    "loss": 0.6091001033782959,
    "learning_rate": 3.746139658277096e-05,
    "timestep_mean": 493.25,
    "epsilon_norm": 4.289463520050049,
    "prediction_norm": 2.8321077823638916,
    "timestamp": "2025-07-14T12:34:54.789810"
  },
  {
    "epoch": 2249,
    "batch_idx": 0,
    "loss": 0.2114276885986328,
    "learning_rate": 3.776024489824091e-05,
    "timestep_mean": 566.25,
    "epsilon_norm": 8.219229698181152,
    "prediction_norm": 7.260115623474121,
    "timestamp": "2025-07-14T12:34:54.794968"
  },
  {
    "epoch": 2249,
    "batch_idx": 32,
    "loss": 0.39364898204803467,
    "learning_rate": 3.776024489824091e-05,
    "timestep_mean": 487.75,
    "epsilon_norm": 8.793462753295898,
    "prediction_norm": 7.454946517944336,
    "timestamp": "2025-07-14T12:34:54.800241"
  },
  {
    "epoch": 2249,
    "batch_idx": 64,
    "loss": 0.14865171909332275,
    "learning_rate": 3.776024489824091e-05,
    "timestep_mean": 488.9375,
    "epsilon_norm": 6.231723785400391,
    "prediction_norm": 5.749694347381592,
    "timestamp": "2025-07-14T12:34:54.805488"
  },
  {
    "epoch": 2249,
    "batch_idx": 96,
    "loss": 0.23975928127765656,
    "learning_rate": 3.776024489824091e-05,
    "timestep_mean": 562.0,
    "epsilon_norm": 7.1625471115112305,
    "prediction_norm": 6.242585182189941,
    "timestamp": "2025-07-14T12:34:54.810668"
  },
  {
    "epoch": 2249,
    "batch_idx": 128,
    "loss": 0.4671946167945862,
    "learning_rate": 3.776024489824091e-05,
    "timestep_mean": 416.90625,
    "epsilon_norm": 8.430971145629883,
    "prediction_norm": 6.129302978515625,
    "timestamp": "2025-07-14T12:34:54.815830"
  },
  {
    "epoch": 2249,
    "batch_idx": 160,
    "loss": 0.16467402875423431,
    "learning_rate": 3.776024489824091e-05,
    "timestep_mean": 573.21875,
    "epsilon_norm": 8.05367660522461,
    "prediction_norm": 7.152191638946533,
    "timestamp": "2025-07-14T12:34:54.821103"
  },
  {
    "epoch": 2249,
    "batch_idx": 192,
    "loss": 0.2566176652908325,
    "learning_rate": 3.776024489824091e-05,
    "timestep_mean": 357.75,
    "epsilon_norm": 3.452233076095581,
    "prediction_norm": 2.9765524864196777,
    "timestamp": "2025-07-14T12:34:54.826124"
  },
  {
    "epoch": 2250,
    "batch_idx": 0,
    "loss": 0.2924368381500244,
    "learning_rate": 3.806023374435662e-05,
    "timestep_mean": 480.375,
    "epsilon_norm": 7.181312084197998,
    "prediction_norm": 5.954043865203857,
    "timestamp": "2025-07-14T12:34:54.831149"
  },
  {
    "epoch": 2250,
    "batch_idx": 32,
    "loss": 0.5138927102088928,
    "learning_rate": 3.806023374435662e-05,
    "timestep_mean": 459.59375,
    "epsilon_norm": 8.35940933227539,
    "prediction_norm": 6.919773101806641,
    "timestamp": "2025-07-14T12:34:54.836016"
  },
  {
    "epoch": 2250,
    "batch_idx": 64,
    "loss": 0.3788318932056427,
    "learning_rate": 3.806023374435662e-05,
    "timestep_mean": 505.96875,
    "epsilon_norm": 8.558385848999023,
    "prediction_norm": 6.907095432281494,
    "timestamp": "2025-07-14T12:34:54.840880"
  },
  {
    "epoch": 2250,
    "batch_idx": 96,
    "loss": 0.2676374912261963,
    "learning_rate": 3.806023374435662e-05,
    "timestep_mean": 412.03125,
    "epsilon_norm": 8.288777351379395,
    "prediction_norm": 7.206384181976318,
    "timestamp": "2025-07-14T12:34:54.845741"
  },
  {
    "epoch": 2250,
    "batch_idx": 128,
    "loss": 0.24631819128990173,
    "learning_rate": 3.806023374435662e-05,
    "timestep_mean": 473.03125,
    "epsilon_norm": 7.064560413360596,
    "prediction_norm": 6.217043399810791,
    "timestamp": "2025-07-14T12:34:54.850649"
  },
  {
    "epoch": 2250,
    "batch_idx": 160,
    "loss": 0.18155688047409058,
    "learning_rate": 3.806023374435662e-05,
    "timestep_mean": 545.0,
    "epsilon_norm": 7.27325963973999,
    "prediction_norm": 6.230165481567383,
    "timestamp": "2025-07-14T12:34:54.855604"
  },
  {
    "epoch": 2250,
    "batch_idx": 192,
    "loss": 0.011663476005196571,
    "learning_rate": 3.806023374435662e-05,
    "timestep_mean": 751.75,
    "epsilon_norm": 3.7259654998779297,
    "prediction_norm": 3.6580727100372314,
    "timestamp": "2025-07-14T12:34:54.860407"
  },
  {
    "epoch": 2251,
    "batch_idx": 0,
    "loss": 0.36989036202430725,
    "learning_rate": 3.836136238092551e-05,
    "timestep_mean": 351.21875,
    "epsilon_norm": 6.733765602111816,
    "prediction_norm": 4.92358922958374,
    "timestamp": "2025-07-14T12:34:54.865301"
  },
  {
    "epoch": 2251,
    "batch_idx": 32,
    "loss": 0.46230071783065796,
    "learning_rate": 3.836136238092551e-05,
    "timestep_mean": 457.15625,
    "epsilon_norm": 7.897520065307617,
    "prediction_norm": 6.704988956451416,
    "timestamp": "2025-07-14T12:34:54.870204"
  },
  {
    "epoch": 2251,
    "batch_idx": 64,
    "loss": 0.16369903087615967,
    "learning_rate": 3.836136238092551e-05,
    "timestep_mean": 516.1875,
    "epsilon_norm": 6.985003471374512,
    "prediction_norm": 6.20591402053833,
    "timestamp": "2025-07-14T12:34:54.875278"
  },
  {
    "epoch": 2251,
    "batch_idx": 96,
    "loss": 0.43686944246292114,
    "learning_rate": 3.836136238092551e-05,
    "timestep_mean": 465.03125,
    "epsilon_norm": 8.454835891723633,
    "prediction_norm": 5.840087413787842,
    "timestamp": "2025-07-14T12:34:54.880172"
  },
  {
    "epoch": 2251,
    "batch_idx": 128,
    "loss": 0.28291162848472595,
    "learning_rate": 3.836136238092551e-05,
    "timestep_mean": 542.625,
    "epsilon_norm": 7.840846538543701,
    "prediction_norm": 6.7212395668029785,
    "timestamp": "2025-07-14T12:34:54.885009"
  },
  {
    "epoch": 2251,
    "batch_idx": 160,
    "loss": 0.36199522018432617,
    "learning_rate": 3.836136238092551e-05,
    "timestep_mean": 419.78125,
    "epsilon_norm": 7.913239479064941,
    "prediction_norm": 6.667163848876953,
    "timestamp": "2025-07-14T12:34:54.889895"
  },
  {
    "epoch": 2251,
    "batch_idx": 192,
    "loss": 0.07962408661842346,
    "learning_rate": 3.836136238092551e-05,
    "timestep_mean": 673.5,
    "epsilon_norm": 4.5414652824401855,
    "prediction_norm": 4.241066932678223,
    "timestamp": "2025-07-14T12:34:54.894750"
  },
  {
    "epoch": 2252,
    "batch_idx": 0,
    "loss": 0.28299587965011597,
    "learning_rate": 3.8663630064942655e-05,
    "timestep_mean": 470.78125,
    "epsilon_norm": 9.411832809448242,
    "prediction_norm": 8.058300018310547,
    "timestamp": "2025-07-14T12:34:54.899649"
  },
  {
    "epoch": 2252,
    "batch_idx": 32,
    "loss": 0.17456156015396118,
    "learning_rate": 3.8663630064942655e-05,
    "timestep_mean": 521.0625,
    "epsilon_norm": 8.211044311523438,
    "prediction_norm": 6.721195697784424,
    "timestamp": "2025-07-14T12:34:54.904641"
  },
  {
    "epoch": 2252,
    "batch_idx": 64,
    "loss": 0.20011352002620697,
    "learning_rate": 3.8663630064942655e-05,
    "timestep_mean": 505.28125,
    "epsilon_norm": 7.821759223937988,
    "prediction_norm": 6.708611488342285,
    "timestamp": "2025-07-14T12:34:54.909514"
  },
  {
    "epoch": 2252,
    "batch_idx": 96,
    "loss": 0.4807145297527313,
    "learning_rate": 3.8663630064942655e-05,
    "timestep_mean": 376.53125,
    "epsilon_norm": 8.494495391845703,
    "prediction_norm": 6.683429718017578,
    "timestamp": "2025-07-14T12:34:54.914519"
  },
  {
    "epoch": 2252,
    "batch_idx": 128,
    "loss": 0.21619057655334473,
    "learning_rate": 3.8663630064942655e-05,
    "timestep_mean": 510.84375,
    "epsilon_norm": 7.891923427581787,
    "prediction_norm": 6.862485408782959,
    "timestamp": "2025-07-14T12:34:54.919329"
  },
  {
    "epoch": 2252,
    "batch_idx": 160,
    "loss": 0.2564331293106079,
    "learning_rate": 3.8663630064942655e-05,
    "timestep_mean": 542.09375,
    "epsilon_norm": 9.10219955444336,
    "prediction_norm": 8.34325885772705,
    "timestamp": "2025-07-14T12:34:54.924185"
  },
  {
    "epoch": 2252,
    "batch_idx": 192,
    "loss": 0.07659096270799637,
    "learning_rate": 3.8663630064942655e-05,
    "timestep_mean": 554.0,
    "epsilon_norm": 2.6961514949798584,
    "prediction_norm": 2.373689889907837,
    "timestamp": "2025-07-14T12:34:54.929030"
  },
  {
    "epoch": 2253,
    "batch_idx": 0,
    "loss": 0.23707516491413116,
    "learning_rate": 3.896703605059242e-05,
    "timestep_mean": 491.75,
    "epsilon_norm": 8.356762886047363,
    "prediction_norm": 6.915239334106445,
    "timestamp": "2025-07-14T12:34:54.934025"
  },
  {
    "epoch": 2253,
    "batch_idx": 32,
    "loss": 0.23001165688037872,
    "learning_rate": 3.896703605059242e-05,
    "timestep_mean": 465.875,
    "epsilon_norm": 8.072443008422852,
    "prediction_norm": 7.322483062744141,
    "timestamp": "2025-07-14T12:34:54.938895"
  },
  {
    "epoch": 2253,
    "batch_idx": 64,
    "loss": 0.2663753032684326,
    "learning_rate": 3.896703605059242e-05,
    "timestep_mean": 489.625,
    "epsilon_norm": 8.571586608886719,
    "prediction_norm": 7.4936442375183105,
    "timestamp": "2025-07-14T12:34:54.943759"
  },
  {
    "epoch": 2253,
    "batch_idx": 96,
    "loss": 0.2100771963596344,
    "learning_rate": 3.896703605059242e-05,
    "timestep_mean": 540.8125,
    "epsilon_norm": 7.934042930603027,
    "prediction_norm": 6.995185375213623,
    "timestamp": "2025-07-14T12:34:54.948658"
  },
  {
    "epoch": 2253,
    "batch_idx": 128,
    "loss": 0.16884705424308777,
    "learning_rate": 3.896703605059242e-05,
    "timestep_mean": 533.5625,
    "epsilon_norm": 7.496562480926514,
    "prediction_norm": 6.667675495147705,
    "timestamp": "2025-07-14T12:34:54.953506"
  },
  {
    "epoch": 2253,
    "batch_idx": 160,
    "loss": 0.24815794825553894,
    "learning_rate": 3.896703605059242e-05,
    "timestep_mean": 589.71875,
    "epsilon_norm": 6.778853893280029,
    "prediction_norm": 5.48440408706665,
    "timestamp": "2025-07-14T12:34:54.958411"
  },
  {
    "epoch": 2253,
    "batch_idx": 192,
    "loss": 0.3393203914165497,
    "learning_rate": 3.896703605059242e-05,
    "timestep_mean": 446.125,
    "epsilon_norm": 5.412655830383301,
    "prediction_norm": 4.106726169586182,
    "timestamp": "2025-07-14T12:34:54.963352"
  },
  {
    "epoch": 2254,
    "batch_idx": 0,
    "loss": 0.13527944684028625,
    "learning_rate": 3.927157958925081e-05,
    "timestep_mean": 568.28125,
    "epsilon_norm": 8.230788230895996,
    "prediction_norm": 7.266050338745117,
    "timestamp": "2025-07-14T12:34:54.968210"
  },
  {
    "epoch": 2254,
    "batch_idx": 32,
    "loss": 0.17342990636825562,
    "learning_rate": 3.927157958925081e-05,
    "timestep_mean": 517.6875,
    "epsilon_norm": 7.681870937347412,
    "prediction_norm": 7.0978007316589355,
    "timestamp": "2025-07-14T12:34:54.973157"
  },
  {
    "epoch": 2254,
    "batch_idx": 64,
    "loss": 0.4924009144306183,
    "learning_rate": 3.927157958925081e-05,
    "timestep_mean": 491.5625,
    "epsilon_norm": 8.33731746673584,
    "prediction_norm": 6.836748123168945,
    "timestamp": "2025-07-14T12:34:54.978312"
  },
  {
    "epoch": 2254,
    "batch_idx": 96,
    "loss": 0.21410346031188965,
    "learning_rate": 3.927157958925081e-05,
    "timestep_mean": 462.1875,
    "epsilon_norm": 7.560249328613281,
    "prediction_norm": 6.573585510253906,
    "timestamp": "2025-07-14T12:34:54.983190"
  },
  {
    "epoch": 2254,
    "batch_idx": 128,
    "loss": 0.22189506888389587,
    "learning_rate": 3.927157958925081e-05,
    "timestep_mean": 512.46875,
    "epsilon_norm": 8.49448299407959,
    "prediction_norm": 8.115022659301758,
    "timestamp": "2025-07-14T12:34:54.988124"
  },
  {
    "epoch": 2254,
    "batch_idx": 160,
    "loss": 0.17749513685703278,
    "learning_rate": 3.927157958925081e-05,
    "timestep_mean": 416.625,
    "epsilon_norm": 7.4621758460998535,
    "prediction_norm": 6.4277801513671875,
    "timestamp": "2025-07-14T12:34:54.992954"
  },
  {
    "epoch": 2254,
    "batch_idx": 192,
    "loss": 0.23606330156326294,
    "learning_rate": 3.927157958925081e-05,
    "timestep_mean": 422.625,
    "epsilon_norm": 2.7029194831848145,
    "prediction_norm": 2.543163776397705,
    "timestamp": "2025-07-14T12:34:54.997851"
  },
  {
    "epoch": 2255,
    "batch_idx": 0,
    "loss": 0.26163071393966675,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 459.90625,
    "epsilon_norm": 9.16381549835205,
    "prediction_norm": 7.676699638366699,
    "timestamp": "2025-07-14T12:34:55.002733"
  },
  {
    "epoch": 2255,
    "batch_idx": 32,
    "loss": 0.35569101572036743,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 519.78125,
    "epsilon_norm": 7.570328235626221,
    "prediction_norm": 6.546938896179199,
    "timestamp": "2025-07-14T12:34:55.007606"
  },
  {
    "epoch": 2255,
    "batch_idx": 64,
    "loss": 0.4111352264881134,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 491.84375,
    "epsilon_norm": 8.330445289611816,
    "prediction_norm": 7.27789831161499,
    "timestamp": "2025-07-14T12:34:55.012461"
  },
  {
    "epoch": 2255,
    "batch_idx": 96,
    "loss": 0.11589781939983368,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 541.625,
    "epsilon_norm": 7.232936859130859,
    "prediction_norm": 6.778360843658447,
    "timestamp": "2025-07-14T12:34:55.017368"
  },
  {
    "epoch": 2255,
    "batch_idx": 128,
    "loss": 0.13775435090065002,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 561.21875,
    "epsilon_norm": 7.867006301879883,
    "prediction_norm": 7.4125823974609375,
    "timestamp": "2025-07-14T12:34:55.022190"
  },
  {
    "epoch": 2255,
    "batch_idx": 160,
    "loss": 0.2854459285736084,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 526.8125,
    "epsilon_norm": 9.24251651763916,
    "prediction_norm": 7.4957804679870605,
    "timestamp": "2025-07-14T12:34:55.027071"
  },
  {
    "epoch": 2255,
    "batch_idx": 192,
    "loss": 0.593308687210083,
    "learning_rate": 3.957725992948691e-05,
    "timestep_mean": 440.625,
    "epsilon_norm": 4.154024600982666,
    "prediction_norm": 3.265355110168457,
    "timestamp": "2025-07-14T12:34:55.031986"
  },
  {
    "epoch": 2256,
    "batch_idx": 0,
    "loss": 0.22282084822654724,
    "learning_rate": 3.988407631706481e-05,
    "timestep_mean": 517.3125,
    "epsilon_norm": 8.566350936889648,
    "prediction_norm": 7.4407148361206055,
    "timestamp": "2025-07-14T12:34:55.036835"
  },
  {
    "epoch": 2256,
    "batch_idx": 32,
    "loss": 0.42023026943206787,
    "learning_rate": 3.988407631706481e-05,
    "timestep_mean": 444.75,
    "epsilon_norm": 7.903234958648682,
    "prediction_norm": 5.801369667053223,
    "timestamp": "2025-07-14T12:34:55.041706"
  },
  {
    "epoch": 2256,
    "batch_idx": 64,
    "loss": 0.3111203908920288,
    "learning_rate": 3.988407631706481e-05,
    "timestep_mean": 447.15625,
    "epsilon_norm": 7.2033467292785645,
    "prediction_norm": 5.836050987243652,
    "timestamp": "2025-07-14T12:34:55.046716"
  },
  {
    "epoch": 2256,
    "batch_idx": 96,
    "loss": 0.38678231835365295,
    "learning_rate": 3.988407631706481e-05,
    "timestep_mean": 451.375,
    "epsilon_norm": 10.032175064086914,
    "prediction_norm": 8.067089080810547,
    "timestamp": "2025-07-14T12:34:55.051567"
  },
  {
    "epoch": 2256,
    "batch_idx": 128,
    "loss": 0.3787781596183777,
    "learning_rate": 3.988407631706481e-05,
    "timestep_mean": 417.21875,
    "epsilon_norm": 7.376856327056885,
    "prediction_norm": 5.442153453826904,
    "timestamp": "2025-07-14T12:34:55.056472"
  },
  {
    "epoch": 2256,
    "batch_idx": 160,
    "loss": 0.21387270092964172,
    "learning_rate": 3.988407631706481e-05,
    "timestep_mean": 478.0,
    "epsilon_norm": 7.451067924499512,
    "prediction_norm": 6.3570556640625,
    "timestamp": "2025-07-14T12:34:55.061301"
  },
  {
    "epoch": 2256,
    "batch_idx": 192,
    "loss": 0.32217535376548767,
    "learning_rate": 3.988407631706481e-05,
    "timestep_mean": 535.75,
    "epsilon_norm": 3.865957260131836,
    "prediction_norm": 3.0145742893218994,
    "timestamp": "2025-07-14T12:34:55.066118"
  },
  {
    "epoch": 2257,
    "batch_idx": 0,
    "loss": 0.24051256477832794,
    "learning_rate": 4.01920279949457e-05,
    "timestep_mean": 518.5,
    "epsilon_norm": 7.300107002258301,
    "prediction_norm": 6.365264415740967,
    "timestamp": "2025-07-14T12:34:55.071134"
  },
  {
    "epoch": 2257,
    "batch_idx": 32,
    "loss": 0.27289149165153503,
    "learning_rate": 4.01920279949457e-05,
    "timestep_mean": 480.03125,
    "epsilon_norm": 6.779027462005615,
    "prediction_norm": 5.471419334411621,
    "timestamp": "2025-07-14T12:34:55.076377"
  },
  {
    "epoch": 2257,
    "batch_idx": 64,
    "loss": 0.2731449007987976,
    "learning_rate": 4.01920279949457e-05,
    "timestep_mean": 554.4375,
    "epsilon_norm": 7.866480350494385,
    "prediction_norm": 6.820722579956055,
    "timestamp": "2025-07-14T12:34:55.081565"
  },
  {
    "epoch": 2257,
    "batch_idx": 96,
    "loss": 0.3929733633995056,
    "learning_rate": 4.01920279949457e-05,
    "timestep_mean": 555.90625,
    "epsilon_norm": 8.535656929016113,
    "prediction_norm": 7.023194313049316,
    "timestamp": "2025-07-14T12:34:55.086728"
  },
  {
    "epoch": 2257,
    "batch_idx": 128,
    "loss": 0.21782249212265015,
    "learning_rate": 4.01920279949457e-05,
    "timestep_mean": 503.03125,
    "epsilon_norm": 7.883387088775635,
    "prediction_norm": 7.254192352294922,
    "timestamp": "2025-07-14T12:34:55.091970"
  },
  {
    "epoch": 2257,
    "batch_idx": 160,
    "loss": 0.2097162902355194,
    "learning_rate": 4.01920279949457e-05,
    "timestep_mean": 486.71875,
    "epsilon_norm": 6.642462253570557,
    "prediction_norm": 5.979885578155518,
    "timestamp": "2025-07-14T12:34:55.097095"
  },
  {
    "epoch": 2257,
    "batch_idx": 192,
    "loss": 0.1770516186952591,
    "learning_rate": 4.01920279949457e-05,
    "timestep_mean": 325.625,
    "epsilon_norm": 4.190246105194092,
    "prediction_norm": 3.486340045928955,
    "timestamp": "2025-07-14T12:34:55.102253"
  },
  {
    "epoch": 2258,
    "batch_idx": 0,
    "loss": 0.2008463442325592,
    "learning_rate": 4.050111420328939e-05,
    "timestep_mean": 524.53125,
    "epsilon_norm": 8.990028381347656,
    "prediction_norm": 7.754715919494629,
    "timestamp": "2025-07-14T12:34:55.107418"
  },
  {
    "epoch": 2258,
    "batch_idx": 32,
    "loss": 0.496120810508728,
    "learning_rate": 4.050111420328939e-05,
    "timestep_mean": 451.03125,
    "epsilon_norm": 8.333712577819824,
    "prediction_norm": 5.647116184234619,
    "timestamp": "2025-07-14T12:34:55.112563"
  },
  {
    "epoch": 2258,
    "batch_idx": 64,
    "loss": 0.2639821767807007,
    "learning_rate": 4.050111420328939e-05,
    "timestep_mean": 495.8125,
    "epsilon_norm": 7.604011535644531,
    "prediction_norm": 6.590063095092773,
    "timestamp": "2025-07-14T12:34:55.117840"
  },
  {
    "epoch": 2258,
    "batch_idx": 96,
    "loss": 0.1943517029285431,
    "learning_rate": 4.050111420328939e-05,
    "timestep_mean": 503.40625,
    "epsilon_norm": 6.827128887176514,
    "prediction_norm": 5.728496074676514,
    "timestamp": "2025-07-14T12:34:55.123044"
  },
  {
    "epoch": 2258,
    "batch_idx": 128,
    "loss": 0.22819657623767853,
    "learning_rate": 4.050111420328939e-05,
    "timestep_mean": 389.5625,
    "epsilon_norm": 7.58146333694458,
    "prediction_norm": 7.217460632324219,
    "timestamp": "2025-07-14T12:34:55.128241"
  },
  {
    "epoch": 2258,
    "batch_idx": 160,
    "loss": 0.5804325938224792,
    "learning_rate": 4.050111420328939e-05,
    "timestep_mean": 542.25,
    "epsilon_norm": 9.473995208740234,
    "prediction_norm": 6.981513500213623,
    "timestamp": "2025-07-14T12:34:55.133590"
  },
  {
    "epoch": 2258,
    "batch_idx": 192,
    "loss": 0.13508594036102295,
    "learning_rate": 4.050111420328939e-05,
    "timestep_mean": 630.25,
    "epsilon_norm": 3.8342373371124268,
    "prediction_norm": 3.308882236480713,
    "timestamp": "2025-07-14T12:34:55.138632"
  },
  {
    "epoch": 2259,
    "batch_idx": 0,
    "loss": 0.19086456298828125,
    "learning_rate": 4.0811334179456256e-05,
    "timestep_mean": 530.34375,
    "epsilon_norm": 8.910337448120117,
    "prediction_norm": 8.064870834350586,
    "timestamp": "2025-07-14T12:34:55.143650"
  },
  {
    "epoch": 2259,
    "batch_idx": 32,
    "loss": 0.26444104313850403,
    "learning_rate": 4.0811334179456256e-05,
    "timestep_mean": 435.65625,
    "epsilon_norm": 7.884471893310547,
    "prediction_norm": 7.1781325340271,
    "timestamp": "2025-07-14T12:34:55.148747"
  },
  {
    "epoch": 2259,
    "batch_idx": 64,
    "loss": 0.3479577898979187,
    "learning_rate": 4.0811334179456256e-05,
    "timestep_mean": 503.40625,
    "epsilon_norm": 8.753474235534668,
    "prediction_norm": 7.650266647338867,
    "timestamp": "2025-07-14T12:34:55.153916"
  },
  {
    "epoch": 2259,
    "batch_idx": 96,
    "loss": 0.11925777792930603,
    "learning_rate": 4.0811334179456256e-05,
    "timestep_mean": 555.4375,
    "epsilon_norm": 6.83831787109375,
    "prediction_norm": 6.088875770568848,
    "timestamp": "2025-07-14T12:34:55.159108"
  },
  {
    "epoch": 2259,
    "batch_idx": 128,
    "loss": 0.24027717113494873,
    "learning_rate": 4.0811334179456256e-05,
    "timestep_mean": 499.625,
    "epsilon_norm": 8.135740280151367,
    "prediction_norm": 6.802454471588135,
    "timestamp": "2025-07-14T12:34:55.164345"
  },
  {
    "epoch": 2259,
    "batch_idx": 160,
    "loss": 0.3144495487213135,
    "learning_rate": 4.0811334179456256e-05,
    "timestep_mean": 499.65625,
    "epsilon_norm": 8.35534381866455,
    "prediction_norm": 6.785764694213867,
    "timestamp": "2025-07-14T12:34:55.169549"
  },
  {
    "epoch": 2259,
    "batch_idx": 192,
    "loss": 0.42571812868118286,
    "learning_rate": 4.0811334179456256e-05,
    "timestep_mean": 484.875,
    "epsilon_norm": 3.7550649642944336,
    "prediction_norm": 2.7316415309906006,
    "timestamp": "2025-07-14T12:34:55.174687"
  },
  {
    "epoch": 2260,
    "batch_idx": 0,
    "loss": 0.30144596099853516,
    "learning_rate": 4.1122687158009425e-05,
    "timestep_mean": 422.75,
    "epsilon_norm": 8.444559097290039,
    "prediction_norm": 6.854128837585449,
    "timestamp": "2025-07-14T12:34:55.179807"
  },
  {
    "epoch": 2260,
    "batch_idx": 32,
    "loss": 0.3276829421520233,
    "learning_rate": 4.1122687158009425e-05,
    "timestep_mean": 520.1875,
    "epsilon_norm": 9.040712356567383,
    "prediction_norm": 7.9399566650390625,
    "timestamp": "2025-07-14T12:34:55.184995"
  },
  {
    "epoch": 2260,
    "batch_idx": 64,
    "loss": 0.2043658196926117,
    "learning_rate": 4.1122687158009425e-05,
    "timestep_mean": 494.84375,
    "epsilon_norm": 7.566270351409912,
    "prediction_norm": 6.724692344665527,
    "timestamp": "2025-07-14T12:34:55.190202"
  },
  {
    "epoch": 2260,
    "batch_idx": 96,
    "loss": 0.25210198760032654,
    "learning_rate": 4.1122687158009425e-05,
    "timestep_mean": 514.09375,
    "epsilon_norm": 8.184015274047852,
    "prediction_norm": 6.866322040557861,
    "timestamp": "2025-07-14T12:34:55.195242"
  },
  {
    "epoch": 2260,
    "batch_idx": 128,
    "loss": 0.38956254720687866,
    "learning_rate": 4.1122687158009425e-05,
    "timestep_mean": 359.59375,
    "epsilon_norm": 8.306954383850098,
    "prediction_norm": 6.122776031494141,
    "timestamp": "2025-07-14T12:34:55.200532"
  },
  {
    "epoch": 2260,
    "batch_idx": 160,
    "loss": 0.23016975820064545,
    "learning_rate": 4.1122687158009425e-05,
    "timestep_mean": 557.625,
    "epsilon_norm": 7.909029006958008,
    "prediction_norm": 7.026235103607178,
    "timestamp": "2025-07-14T12:34:55.205754"
  },
  {
    "epoch": 2260,
    "batch_idx": 192,
    "loss": 0.23562708497047424,
    "learning_rate": 4.1122687158009425e-05,
    "timestep_mean": 556.375,
    "epsilon_norm": 4.422583103179932,
    "prediction_norm": 3.5820231437683105,
    "timestamp": "2025-07-14T12:34:55.210895"
  },
  {
    "epoch": 2261,
    "batch_idx": 0,
    "loss": 0.2581288516521454,
    "learning_rate": 4.143517237071653e-05,
    "timestep_mean": 467.4375,
    "epsilon_norm": 6.785281658172607,
    "prediction_norm": 5.424283981323242,
    "timestamp": "2025-07-14T12:34:55.216065"
  },
  {
    "epoch": 2261,
    "batch_idx": 32,
    "loss": 0.33430036902427673,
    "learning_rate": 4.143517237071653e-05,
    "timestep_mean": 483.28125,
    "epsilon_norm": 8.364336013793945,
    "prediction_norm": 7.349791526794434,
    "timestamp": "2025-07-14T12:34:55.221185"
  },
  {
    "epoch": 2261,
    "batch_idx": 64,
    "loss": 0.12455205619335175,
    "learning_rate": 4.143517237071653e-05,
    "timestep_mean": 507.84375,
    "epsilon_norm": 7.003073215484619,
    "prediction_norm": 6.616370677947998,
    "timestamp": "2025-07-14T12:34:55.226376"
  },
  {
    "epoch": 2261,
    "batch_idx": 96,
    "loss": 0.2660471498966217,
    "learning_rate": 4.143517237071653e-05,
    "timestep_mean": 486.84375,
    "epsilon_norm": 7.714317321777344,
    "prediction_norm": 5.680781364440918,
    "timestamp": "2025-07-14T12:34:55.231600"
  },
  {
    "epoch": 2261,
    "batch_idx": 128,
    "loss": 0.2590675354003906,
    "learning_rate": 4.143517237071653e-05,
    "timestep_mean": 520.375,
    "epsilon_norm": 7.46929931640625,
    "prediction_norm": 6.041719913482666,
    "timestamp": "2025-07-14T12:34:55.236804"
  },
  {
    "epoch": 2261,
    "batch_idx": 160,
    "loss": 0.19539177417755127,
    "learning_rate": 4.143517237071653e-05,
    "timestep_mean": 427.65625,
    "epsilon_norm": 7.27937126159668,
    "prediction_norm": 6.301098346710205,
    "timestamp": "2025-07-14T12:34:55.242109"
  },
  {
    "epoch": 2261,
    "batch_idx": 192,
    "loss": 0.21078410744667053,
    "learning_rate": 4.143517237071653e-05,
    "timestep_mean": 263.125,
    "epsilon_norm": 4.125210285186768,
    "prediction_norm": 3.557614803314209,
    "timestamp": "2025-07-14T12:34:55.247264"
  },
  {
    "epoch": 2262,
    "batch_idx": 0,
    "loss": 0.23357489705085754,
    "learning_rate": 4.174878904655099e-05,
    "timestep_mean": 440.1875,
    "epsilon_norm": 7.969254016876221,
    "prediction_norm": 6.673079967498779,
    "timestamp": "2025-07-14T12:34:55.252508"
  },
  {
    "epoch": 2262,
    "batch_idx": 32,
    "loss": 0.16830462217330933,
    "learning_rate": 4.174878904655099e-05,
    "timestep_mean": 511.84375,
    "epsilon_norm": 7.424383163452148,
    "prediction_norm": 6.774330139160156,
    "timestamp": "2025-07-14T12:34:55.257701"
  },
  {
    "epoch": 2262,
    "batch_idx": 64,
    "loss": 0.2588847279548645,
    "learning_rate": 4.174878904655099e-05,
    "timestep_mean": 558.59375,
    "epsilon_norm": 8.519488334655762,
    "prediction_norm": 7.279325008392334,
    "timestamp": "2025-07-14T12:34:55.262875"
  },
  {
    "epoch": 2262,
    "batch_idx": 96,
    "loss": 0.23963136970996857,
    "learning_rate": 4.174878904655099e-05,
    "timestep_mean": 473.15625,
    "epsilon_norm": 7.5099616050720215,
    "prediction_norm": 6.035778045654297,
    "timestamp": "2025-07-14T12:34:55.268095"
  },
  {
    "epoch": 2262,
    "batch_idx": 128,
    "loss": 0.38154107332229614,
    "learning_rate": 4.174878904655099e-05,
    "timestep_mean": 486.96875,
    "epsilon_norm": 8.487607955932617,
    "prediction_norm": 6.237751483917236,
    "timestamp": "2025-07-14T12:34:55.273304"
  },
  {
    "epoch": 2262,
    "batch_idx": 160,
    "loss": 0.15053445100784302,
    "learning_rate": 4.174878904655099e-05,
    "timestep_mean": 538.59375,
    "epsilon_norm": 7.177816867828369,
    "prediction_norm": 6.202805995941162,
    "timestamp": "2025-07-14T12:34:55.278399"
  },
  {
    "epoch": 2262,
    "batch_idx": 192,
    "loss": 0.37624770402908325,
    "learning_rate": 4.174878904655099e-05,
    "timestep_mean": 463.625,
    "epsilon_norm": 3.8032965660095215,
    "prediction_norm": 2.852452516555786,
    "timestamp": "2025-07-14T12:34:55.283661"
  },
  {
    "epoch": 2263,
    "batch_idx": 0,
    "loss": 0.3850690424442291,
    "learning_rate": 4.206353641169525e-05,
    "timestep_mean": 401.1875,
    "epsilon_norm": 8.524744987487793,
    "prediction_norm": 6.220920562744141,
    "timestamp": "2025-07-14T12:34:55.288864"
  },
  {
    "epoch": 2263,
    "batch_idx": 32,
    "loss": 0.35843127965927124,
    "learning_rate": 4.206353641169525e-05,
    "timestep_mean": 497.0,
    "epsilon_norm": 9.462667465209961,
    "prediction_norm": 8.168371200561523,
    "timestamp": "2025-07-14T12:34:55.294062"
  },
  {
    "epoch": 2263,
    "batch_idx": 64,
    "loss": 0.3779991865158081,
    "learning_rate": 4.206353641169525e-05,
    "timestep_mean": 505.84375,
    "epsilon_norm": 7.463738918304443,
    "prediction_norm": 6.151611804962158,
    "timestamp": "2025-07-14T12:34:55.299221"
  },
  {
    "epoch": 2263,
    "batch_idx": 96,
    "loss": 0.3910899758338928,
    "learning_rate": 4.206353641169525e-05,
    "timestep_mean": 513.3125,
    "epsilon_norm": 7.907359600067139,
    "prediction_norm": 5.774411201477051,
    "timestamp": "2025-07-14T12:34:55.304418"
  },
  {
    "epoch": 2263,
    "batch_idx": 128,
    "loss": 0.12125150114297867,
    "learning_rate": 4.206353641169525e-05,
    "timestep_mean": 464.1875,
    "epsilon_norm": 7.645977020263672,
    "prediction_norm": 6.907222270965576,
    "timestamp": "2025-07-14T12:34:55.309601"
  },
  {
    "epoch": 2263,
    "batch_idx": 160,
    "loss": 0.35465770959854126,
    "learning_rate": 4.206353641169525e-05,
    "timestep_mean": 432.6875,
    "epsilon_norm": 7.876417636871338,
    "prediction_norm": 6.231918811798096,
    "timestamp": "2025-07-14T12:34:55.314748"
  },
  {
    "epoch": 2263,
    "batch_idx": 192,
    "loss": 0.05157887190580368,
    "learning_rate": 4.206353641169525e-05,
    "timestep_mean": 564.75,
    "epsilon_norm": 3.2538559436798096,
    "prediction_norm": 3.295552968978882,
    "timestamp": "2025-07-14T12:34:55.319750"
  },
  {
    "epoch": 2264,
    "batch_idx": 0,
    "loss": 0.18163779377937317,
    "learning_rate": 4.2379413689541345e-05,
    "timestep_mean": 505.0625,
    "epsilon_norm": 8.863926887512207,
    "prediction_norm": 8.091323852539062,
    "timestamp": "2025-07-14T12:34:55.324920"
  },
  {
    "epoch": 2264,
    "batch_idx": 32,
    "loss": 0.3573707938194275,
    "learning_rate": 4.2379413689541345e-05,
    "timestep_mean": 522.0,
    "epsilon_norm": 8.317889213562012,
    "prediction_norm": 7.257349491119385,
    "timestamp": "2025-07-14T12:34:55.329912"
  },
  {
    "epoch": 2264,
    "batch_idx": 64,
    "loss": 0.1752571165561676,
    "learning_rate": 4.2379413689541345e-05,
    "timestep_mean": 519.65625,
    "epsilon_norm": 7.485922336578369,
    "prediction_norm": 6.865139484405518,
    "timestamp": "2025-07-14T12:34:55.335063"
  },
  {
    "epoch": 2264,
    "batch_idx": 96,
    "loss": 0.43004387617111206,
    "learning_rate": 4.2379413689541345e-05,
    "timestep_mean": 399.28125,
    "epsilon_norm": 8.729315757751465,
    "prediction_norm": 6.791995048522949,
    "timestamp": "2025-07-14T12:34:55.340179"
  },
  {
    "epoch": 2264,
    "batch_idx": 128,
    "loss": 0.18407486379146576,
    "learning_rate": 4.2379413689541345e-05,
    "timestep_mean": 618.28125,
    "epsilon_norm": 7.725772857666016,
    "prediction_norm": 6.877103328704834,
    "timestamp": "2025-07-14T12:34:55.345292"
  },
  {
    "epoch": 2264,
    "batch_idx": 160,
    "loss": 0.3201867341995239,
    "learning_rate": 4.2379413689541345e-05,
    "timestep_mean": 467.21875,
    "epsilon_norm": 7.921690464019775,
    "prediction_norm": 6.763804912567139,
    "timestamp": "2025-07-14T12:34:55.350690"
  },
  {
    "epoch": 2264,
    "batch_idx": 192,
    "loss": 0.10959415137767792,
    "learning_rate": 4.2379413689541345e-05,
    "timestep_mean": 508.875,
    "epsilon_norm": 3.953188896179199,
    "prediction_norm": 3.448012113571167,
    "timestamp": "2025-07-14T12:34:55.355837"
  },
  {
    "epoch": 2265,
    "batch_idx": 0,
    "loss": 0.41980651021003723,
    "learning_rate": 4.269642010069319e-05,
    "timestep_mean": 481.375,
    "epsilon_norm": 8.509687423706055,
    "prediction_norm": 7.122633457183838,
    "timestamp": "2025-07-14T12:34:55.361101"
  },
  {
    "epoch": 2265,
    "batch_idx": 32,
    "loss": 0.18702976405620575,
    "learning_rate": 4.269642010069319e-05,
    "timestep_mean": 475.5,
    "epsilon_norm": 6.976423263549805,
    "prediction_norm": 6.4038496017456055,
    "timestamp": "2025-07-14T12:34:55.366308"
  },
  {
    "epoch": 2265,
    "batch_idx": 64,
    "loss": 0.33879533410072327,
    "learning_rate": 4.269642010069319e-05,
    "timestep_mean": 430.71875,
    "epsilon_norm": 8.41781234741211,
    "prediction_norm": 6.485308647155762,
    "timestamp": "2025-07-14T12:34:55.371627"
  },
  {
    "epoch": 2265,
    "batch_idx": 96,
    "loss": 0.33834531903266907,
    "learning_rate": 4.269642010069319e-05,
    "timestep_mean": 503.8125,
    "epsilon_norm": 8.676396369934082,
    "prediction_norm": 7.426053524017334,
    "timestamp": "2025-07-14T12:34:55.376788"
  },
  {
    "epoch": 2265,
    "batch_idx": 128,
    "loss": 0.2923581600189209,
    "learning_rate": 4.269642010069319e-05,
    "timestep_mean": 546.03125,
    "epsilon_norm": 8.842938423156738,
    "prediction_norm": 7.050233840942383,
    "timestamp": "2025-07-14T12:34:55.381861"
  },
  {
    "epoch": 2265,
    "batch_idx": 160,
    "loss": 0.22518029808998108,
    "learning_rate": 4.269642010069319e-05,
    "timestep_mean": 564.71875,
    "epsilon_norm": 7.366386890411377,
    "prediction_norm": 6.671931743621826,
    "timestamp": "2025-07-14T12:34:55.386919"
  },
  {
    "epoch": 2265,
    "batch_idx": 192,
    "loss": 0.27622464299201965,
    "learning_rate": 4.269642010069319e-05,
    "timestep_mean": 450.125,
    "epsilon_norm": 3.7075419425964355,
    "prediction_norm": 2.7169535160064697,
    "timestamp": "2025-07-14T12:34:55.392201"
  },
  {
    "epoch": 2266,
    "batch_idx": 0,
    "loss": 0.19780370593070984,
    "learning_rate": 4.301455486296946e-05,
    "timestep_mean": 493.9375,
    "epsilon_norm": 7.437999248504639,
    "prediction_norm": 6.2856831550598145,
    "timestamp": "2025-07-14T12:34:55.397294"
  },
  {
    "epoch": 2266,
    "batch_idx": 32,
    "loss": 0.4350222945213318,
    "learning_rate": 4.301455486296946e-05,
    "timestep_mean": 419.65625,
    "epsilon_norm": 8.322531700134277,
    "prediction_norm": 6.304026126861572,
    "timestamp": "2025-07-14T12:34:55.402365"
  },
  {
    "epoch": 2266,
    "batch_idx": 64,
    "loss": 0.22991183400154114,
    "learning_rate": 4.301455486296946e-05,
    "timestep_mean": 520.46875,
    "epsilon_norm": 9.373101234436035,
    "prediction_norm": 8.261371612548828,
    "timestamp": "2025-07-14T12:34:55.407304"
  },
  {
    "epoch": 2266,
    "batch_idx": 96,
    "loss": 0.15534603595733643,
    "learning_rate": 4.301455486296946e-05,
    "timestep_mean": 532.90625,
    "epsilon_norm": 7.941470623016357,
    "prediction_norm": 7.371337890625,
    "timestamp": "2025-07-14T12:34:55.412176"
  },
  {
    "epoch": 2266,
    "batch_idx": 128,
    "loss": 0.09079859405755997,
    "learning_rate": 4.301455486296946e-05,
    "timestep_mean": 525.9375,
    "epsilon_norm": 7.107966899871826,
    "prediction_norm": 6.860215187072754,
    "timestamp": "2025-07-14T12:34:55.417473"
  },
  {
    "epoch": 2266,
    "batch_idx": 160,
    "loss": 0.34895119071006775,
    "learning_rate": 4.301455486296946e-05,
    "timestep_mean": 452.5625,
    "epsilon_norm": 7.873839378356934,
    "prediction_norm": 6.643303394317627,
    "timestamp": "2025-07-14T12:34:55.422720"
  },
  {
    "epoch": 2266,
    "batch_idx": 192,
    "loss": 0.10363918542861938,
    "learning_rate": 4.301455486296946e-05,
    "timestep_mean": 602.0,
    "epsilon_norm": 3.9437379837036133,
    "prediction_norm": 3.8250808715820312,
    "timestamp": "2025-07-14T12:34:55.427782"
  },
  {
    "epoch": 2267,
    "batch_idx": 0,
    "loss": 0.3664977550506592,
    "learning_rate": 4.333381719140384e-05,
    "timestep_mean": 489.65625,
    "epsilon_norm": 8.102888107299805,
    "prediction_norm": 7.310081958770752,
    "timestamp": "2025-07-14T12:34:55.432958"
  },
  {
    "epoch": 2267,
    "batch_idx": 32,
    "loss": 0.5173467993736267,
    "learning_rate": 4.333381719140384e-05,
    "timestep_mean": 519.28125,
    "epsilon_norm": 9.083805084228516,
    "prediction_norm": 7.090732097625732,
    "timestamp": "2025-07-14T12:34:55.438305"
  },
  {
    "epoch": 2267,
    "batch_idx": 64,
    "loss": 0.31863489747047424,
    "learning_rate": 4.333381719140384e-05,
    "timestep_mean": 469.3125,
    "epsilon_norm": 6.850121021270752,
    "prediction_norm": 5.134404182434082,
    "timestamp": "2025-07-14T12:34:55.443383"
  },
  {
    "epoch": 2267,
    "batch_idx": 96,
    "loss": 0.2423776537179947,
    "learning_rate": 4.333381719140384e-05,
    "timestep_mean": 496.84375,
    "epsilon_norm": 6.508511066436768,
    "prediction_norm": 5.899868011474609,
    "timestamp": "2025-07-14T12:34:55.448504"
  },
  {
    "epoch": 2267,
    "batch_idx": 128,
    "loss": 0.20562893152236938,
    "learning_rate": 4.333381719140384e-05,
    "timestep_mean": 545.0,
    "epsilon_norm": 7.849404335021973,
    "prediction_norm": 7.006850719451904,
    "timestamp": "2025-07-14T12:34:55.453637"
  },
  {
    "epoch": 2267,
    "batch_idx": 160,
    "loss": 0.24169868230819702,
    "learning_rate": 4.333381719140384e-05,
    "timestep_mean": 482.71875,
    "epsilon_norm": 7.825197219848633,
    "prediction_norm": 7.183524131774902,
    "timestamp": "2025-07-14T12:34:55.458806"
  },
  {
    "epoch": 2267,
    "batch_idx": 192,
    "loss": 0.5786234140396118,
    "learning_rate": 4.333381719140384e-05,
    "timestep_mean": 306.125,
    "epsilon_norm": 4.652295112609863,
    "prediction_norm": 2.701195240020752,
    "timestamp": "2025-07-14T12:34:55.463710"
  },
  {
    "epoch": 2268,
    "batch_idx": 0,
    "loss": 0.2841465473175049,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 518.78125,
    "epsilon_norm": 8.375869750976562,
    "prediction_norm": 7.233330249786377,
    "timestamp": "2025-07-14T12:34:55.468877"
  },
  {
    "epoch": 2268,
    "batch_idx": 32,
    "loss": 0.2854868769645691,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 465.0625,
    "epsilon_norm": 7.719232559204102,
    "prediction_norm": 6.437976360321045,
    "timestamp": "2025-07-14T12:34:55.474214"
  },
  {
    "epoch": 2268,
    "batch_idx": 64,
    "loss": 0.2879801392555237,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 448.25,
    "epsilon_norm": 8.115580558776855,
    "prediction_norm": 6.880031585693359,
    "timestamp": "2025-07-14T12:34:55.479288"
  },
  {
    "epoch": 2268,
    "batch_idx": 96,
    "loss": 0.2074078619480133,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 546.59375,
    "epsilon_norm": 7.595890045166016,
    "prediction_norm": 6.844090461730957,
    "timestamp": "2025-07-14T12:34:55.484556"
  },
  {
    "epoch": 2268,
    "batch_idx": 128,
    "loss": 0.3457184135913849,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 562.96875,
    "epsilon_norm": 7.96477746963501,
    "prediction_norm": 6.871219635009766,
    "timestamp": "2025-07-14T12:34:55.489740"
  },
  {
    "epoch": 2268,
    "batch_idx": 160,
    "loss": 0.25777333974838257,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 541.5,
    "epsilon_norm": 8.090559959411621,
    "prediction_norm": 6.9203715324401855,
    "timestamp": "2025-07-14T12:34:55.494875"
  },
  {
    "epoch": 2268,
    "batch_idx": 192,
    "loss": 0.0308566614985466,
    "learning_rate": 4.365420629824863e-05,
    "timestep_mean": 652.125,
    "epsilon_norm": 3.1523988246917725,
    "prediction_norm": 3.273062229156494,
    "timestamp": "2025-07-14T12:34:55.499795"
  },
  {
    "epoch": 2269,
    "batch_idx": 0,
    "loss": 0.18892233073711395,
    "learning_rate": 4.39757213929753e-05,
    "timestep_mean": 495.75,
    "epsilon_norm": 7.8040266036987305,
    "prediction_norm": 7.029459476470947,
    "timestamp": "2025-07-14T12:34:55.504787"
  },
  {
    "epoch": 2269,
    "batch_idx": 32,
    "loss": 0.2110409140586853,
    "learning_rate": 4.39757213929753e-05,
    "timestep_mean": 524.03125,
    "epsilon_norm": 7.366105556488037,
    "prediction_norm": 6.3579254150390625,
    "timestamp": "2025-07-14T12:34:55.509644"
  },
  {
    "epoch": 2269,
    "batch_idx": 64,
    "loss": 0.14125022292137146,
    "learning_rate": 4.39757213929753e-05,
    "timestep_mean": 508.5625,
    "epsilon_norm": 7.736150741577148,
    "prediction_norm": 6.66897439956665,
    "timestamp": "2025-07-14T12:34:55.514517"
  },
  {
    "epoch": 2269,
    "batch_idx": 96,
    "loss": 0.2823236584663391,
    "learning_rate": 4.39757213929753e-05,
    "timestep_mean": 551.78125,
    "epsilon_norm": 8.85250473022461,
    "prediction_norm": 7.9717230796813965,
    "timestamp": "2025-07-14T12:34:55.519345"
  },
  {
    "epoch": 2269,
    "batch_idx": 128,
    "loss": 0.21029962599277496,
    "learning_rate": 4.39757213929753e-05,
    "timestep_mean": 511.03125,
    "epsilon_norm": 8.089865684509277,
    "prediction_norm": 7.362054824829102,
    "timestamp": "2025-07-14T12:34:55.524230"
  },
  {
    "epoch": 2269,
    "batch_idx": 160,
    "loss": 0.2305423468351364,
    "learning_rate": 4.39757213929753e-05,
    "timestep_mean": 504.71875,
    "epsilon_norm": 8.388534545898438,
    "prediction_norm": 7.413569450378418,
    "timestamp": "2025-07-14T12:34:55.529226"
  },
  {
    "epoch": 2269,
    "batch_idx": 192,
    "loss": 0.1282385289669037,
    "learning_rate": 4.39757213929753e-05,
    "timestep_mean": 608.375,
    "epsilon_norm": 3.1402978897094727,
    "prediction_norm": 2.5248863697052,
    "timestamp": "2025-07-14T12:34:55.534483"
  },
  {
    "epoch": 2270,
    "batch_idx": 0,
    "loss": 0.312701016664505,
    "learning_rate": 4.429836168227738e-05,
    "timestep_mean": 536.71875,
    "epsilon_norm": 7.993875503540039,
    "prediction_norm": 6.363151550292969,
    "timestamp": "2025-07-14T12:34:55.539564"
  },
  {
    "epoch": 2270,
    "batch_idx": 32,
    "loss": 0.1759813278913498,
    "learning_rate": 4.429836168227738e-05,
    "timestep_mean": 463.09375,
    "epsilon_norm": 7.127710819244385,
    "prediction_norm": 6.1596293449401855,
    "timestamp": "2025-07-14T12:34:55.544646"
  },
  {
    "epoch": 2270,
    "batch_idx": 64,
    "loss": 0.4210478663444519,
    "learning_rate": 4.429836168227738e-05,
    "timestep_mean": 485.875,
    "epsilon_norm": 9.244304656982422,
    "prediction_norm": 7.845112323760986,
    "timestamp": "2025-07-14T12:34:55.549847"
  },
  {
    "epoch": 2270,
    "batch_idx": 96,
    "loss": 0.3907362222671509,
    "learning_rate": 4.429836168227738e-05,
    "timestep_mean": 502.84375,
    "epsilon_norm": 8.352508544921875,
    "prediction_norm": 7.305070877075195,
    "timestamp": "2025-07-14T12:34:55.555039"
  },
  {
    "epoch": 2270,
    "batch_idx": 128,
    "loss": 0.21794536709785461,
    "learning_rate": 4.429836168227738e-05,
    "timestep_mean": 474.5,
    "epsilon_norm": 7.79155969619751,
    "prediction_norm": 7.2029643058776855,
    "timestamp": "2025-07-14T12:34:55.560179"
  },
  {
    "epoch": 2270,
    "batch_idx": 160,
    "loss": 0.11503006517887115,
    "learning_rate": 4.429836168227738e-05,
    "timestep_mean": 523.875,
    "epsilon_norm": 7.034976005554199,
    "prediction_norm": 6.649857044219971,
    "timestamp": "2025-07-14T12:34:55.565329"
  },
  {
    "epoch": 2270,
    "batch_idx": 192,
    "loss": 0.17735883593559265,
    "learning_rate": 4.429836168227738e-05,
    "timestep_mean": 585.125,
    "epsilon_norm": 3.12606143951416,
    "prediction_norm": 3.0861806869506836,
    "timestamp": "2025-07-14T12:34:55.570214"
  },
  {
    "epoch": 2271,
    "batch_idx": 0,
    "loss": 0.1541169285774231,
    "learning_rate": 4.462212637007215e-05,
    "timestep_mean": 560.59375,
    "epsilon_norm": 7.353816032409668,
    "prediction_norm": 6.50228214263916,
    "timestamp": "2025-07-14T12:34:55.575248"
  },
  {
    "epoch": 2271,
    "batch_idx": 32,
    "loss": 0.32156902551651,
    "learning_rate": 4.462212637007215e-05,
    "timestep_mean": 463.90625,
    "epsilon_norm": 7.037036895751953,
    "prediction_norm": 6.153407573699951,
    "timestamp": "2025-07-14T12:34:55.580117"
  },
  {
    "epoch": 2271,
    "batch_idx": 64,
    "loss": 0.3550753593444824,
    "learning_rate": 4.462212637007215e-05,
    "timestep_mean": 481.90625,
    "epsilon_norm": 8.717803955078125,
    "prediction_norm": 6.9431023597717285,
    "timestamp": "2025-07-14T12:34:55.585016"
  },
  {
    "epoch": 2271,
    "batch_idx": 96,
    "loss": 0.17335256934165955,
    "learning_rate": 4.462212637007215e-05,
    "timestep_mean": 488.375,
    "epsilon_norm": 7.024582386016846,
    "prediction_norm": 5.936232089996338,
    "timestamp": "2025-07-14T12:34:55.589850"
  },
  {
    "epoch": 2271,
    "batch_idx": 128,
    "loss": 0.16175076365470886,
    "learning_rate": 4.462212637007215e-05,
    "timestep_mean": 459.9375,
    "epsilon_norm": 6.230231761932373,
    "prediction_norm": 5.59136962890625,
    "timestamp": "2025-07-14T12:34:55.594684"
  },
  {
    "epoch": 2271,
    "batch_idx": 160,
    "loss": 0.33269762992858887,
    "learning_rate": 4.462212637007215e-05,
    "timestep_mean": 514.71875,
    "epsilon_norm": 7.972800254821777,
    "prediction_norm": 6.914881706237793,
    "timestamp": "2025-07-14T12:34:55.599585"
  },
  {
    "epoch": 2271,
    "batch_idx": 192,
    "loss": 0.13475528359413147,
    "learning_rate": 4.462212637007215e-05,
    "timestep_mean": 533.0,
    "epsilon_norm": 4.09325647354126,
    "prediction_norm": 3.750880002975464,
    "timestamp": "2025-07-14T12:34:55.604647"
  },
  {
    "epoch": 2272,
    "batch_idx": 0,
    "loss": 0.28913047909736633,
    "learning_rate": 4.494701465750225e-05,
    "timestep_mean": 489.90625,
    "epsilon_norm": 8.217937469482422,
    "prediction_norm": 7.0659990310668945,
    "timestamp": "2025-07-14T12:34:55.609775"
  },
  {
    "epoch": 2272,
    "batch_idx": 32,
    "loss": 0.26182690262794495,
    "learning_rate": 4.494701465750225e-05,
    "timestep_mean": 448.375,
    "epsilon_norm": 7.704532623291016,
    "prediction_norm": 6.453441143035889,
    "timestamp": "2025-07-14T12:34:55.614890"
  },
  {
    "epoch": 2272,
    "batch_idx": 64,
    "loss": 0.27610889077186584,
    "learning_rate": 4.494701465750225e-05,
    "timestep_mean": 550.09375,
    "epsilon_norm": 7.7201948165893555,
    "prediction_norm": 6.490997314453125,
    "timestamp": "2025-07-14T12:34:55.620033"
  },
  {
    "epoch": 2272,
    "batch_idx": 96,
    "loss": 0.14575836062431335,
    "learning_rate": 4.494701465750225e-05,
    "timestep_mean": 492.25,
    "epsilon_norm": 6.722649574279785,
    "prediction_norm": 6.476539611816406,
    "timestamp": "2025-07-14T12:34:55.625255"
  },
  {
    "epoch": 2272,
    "batch_idx": 128,
    "loss": 0.3118267059326172,
    "learning_rate": 4.494701465750225e-05,
    "timestep_mean": 501.4375,
    "epsilon_norm": 8.075765609741211,
    "prediction_norm": 6.742362976074219,
    "timestamp": "2025-07-14T12:34:55.630416"
  },
  {
    "epoch": 2272,
    "batch_idx": 160,
    "loss": 0.2351049929857254,
    "learning_rate": 4.494701465750225e-05,
    "timestep_mean": 563.34375,
    "epsilon_norm": 8.721096992492676,
    "prediction_norm": 8.11351490020752,
    "timestamp": "2025-07-14T12:34:55.635490"
  },
  {
    "epoch": 2272,
    "batch_idx": 192,
    "loss": 0.040511585772037506,
    "learning_rate": 4.494701465750225e-05,
    "timestep_mean": 614.625,
    "epsilon_norm": 3.6243393421173096,
    "prediction_norm": 3.8781137466430664,
    "timestamp": "2025-07-14T12:34:55.640457"
  },
  {
    "epoch": 2273,
    "batch_idx": 0,
    "loss": 0.2365170270204544,
    "learning_rate": 4.527302574293815e-05,
    "timestep_mean": 514.625,
    "epsilon_norm": 8.058401107788086,
    "prediction_norm": 6.752457141876221,
    "timestamp": "2025-07-14T12:34:55.645643"
  },
  {
    "epoch": 2273,
    "batch_idx": 32,
    "loss": 0.4633254408836365,
    "learning_rate": 4.527302574293815e-05,
    "timestep_mean": 467.28125,
    "epsilon_norm": 9.429852485656738,
    "prediction_norm": 8.106760025024414,
    "timestamp": "2025-07-14T12:34:55.650961"
  },
  {
    "epoch": 2273,
    "batch_idx": 64,
    "loss": 0.2869855463504791,
    "learning_rate": 4.527302574293815e-05,
    "timestep_mean": 475.0,
    "epsilon_norm": 8.571935653686523,
    "prediction_norm": 7.688783168792725,
    "timestamp": "2025-07-14T12:34:55.656109"
  },
  {
    "epoch": 2273,
    "batch_idx": 96,
    "loss": 0.2467835545539856,
    "learning_rate": 4.527302574293815e-05,
    "timestep_mean": 572.6875,
    "epsilon_norm": 8.181572914123535,
    "prediction_norm": 7.345608234405518,
    "timestamp": "2025-07-14T12:34:55.661308"
  },
  {
    "epoch": 2273,
    "batch_idx": 128,
    "loss": 0.18251846730709076,
    "learning_rate": 4.527302574293815e-05,
    "timestep_mean": 552.625,
    "epsilon_norm": 7.436464786529541,
    "prediction_norm": 7.223295211791992,
    "timestamp": "2025-07-14T12:34:55.666541"
  },
  {
    "epoch": 2273,
    "batch_idx": 160,
    "loss": 0.25520142912864685,
    "learning_rate": 4.527302574293815e-05,
    "timestep_mean": 535.5625,
    "epsilon_norm": 8.70730972290039,
    "prediction_norm": 7.7005462646484375,
    "timestamp": "2025-07-14T12:34:55.671817"
  },
  {
    "epoch": 2273,
    "batch_idx": 192,
    "loss": 0.8512018918991089,
    "learning_rate": 4.527302574293815e-05,
    "timestep_mean": 484.375,
    "epsilon_norm": 4.927829742431641,
    "prediction_norm": 3.3042471408843994,
    "timestamp": "2025-07-14T12:34:55.676884"
  },
  {
    "epoch": 2274,
    "batch_idx": 0,
    "loss": 0.16275259852409363,
    "learning_rate": 4.560015882197996e-05,
    "timestep_mean": 544.75,
    "epsilon_norm": 8.304033279418945,
    "prediction_norm": 7.656909465789795,
    "timestamp": "2025-07-14T12:34:55.681963"
  },
  {
    "epoch": 2274,
    "batch_idx": 32,
    "loss": 0.30880048871040344,
    "learning_rate": 4.560015882197996e-05,
    "timestep_mean": 488.40625,
    "epsilon_norm": 7.589737415313721,
    "prediction_norm": 6.146371841430664,
    "timestamp": "2025-07-14T12:34:55.687099"
  },
  {
    "epoch": 2274,
    "batch_idx": 64,
    "loss": 0.1582985818386078,
    "learning_rate": 4.560015882197996e-05,
    "timestep_mean": 492.4375,
    "epsilon_norm": 7.671225070953369,
    "prediction_norm": 7.08497428894043,
    "timestamp": "2025-07-14T12:34:55.692955"
  },
  {
    "epoch": 2274,
    "batch_idx": 96,
    "loss": 0.3487001657485962,
    "learning_rate": 4.560015882197996e-05,
    "timestep_mean": 469.78125,
    "epsilon_norm": 7.650694370269775,
    "prediction_norm": 6.397622108459473,
    "timestamp": "2025-07-14T12:34:55.698298"
  },
  {
    "epoch": 2274,
    "batch_idx": 128,
    "loss": 0.2755580246448517,
    "learning_rate": 4.560015882197996e-05,
    "timestep_mean": 495.0,
    "epsilon_norm": 7.945795059204102,
    "prediction_norm": 6.454809665679932,
    "timestamp": "2025-07-14T12:34:55.703758"
  },
  {
    "epoch": 2274,
    "batch_idx": 160,
    "loss": 0.19262191653251648,
    "learning_rate": 4.560015882197996e-05,
    "timestep_mean": 488.84375,
    "epsilon_norm": 8.067846298217773,
    "prediction_norm": 7.167858123779297,
    "timestamp": "2025-07-14T12:34:55.709139"
  },
  {
    "epoch": 2274,
    "batch_idx": 192,
    "loss": 0.144901305437088,
    "learning_rate": 4.560015882197996e-05,
    "timestep_mean": 584.125,
    "epsilon_norm": 3.7216761112213135,
    "prediction_norm": 3.443676710128784,
    "timestamp": "2025-07-14T12:34:55.714240"
  },
  {
    "epoch": 2275,
    "batch_idx": 0,
    "loss": 0.3111084997653961,
    "learning_rate": 4.592841308745946e-05,
    "timestep_mean": 507.0625,
    "epsilon_norm": 8.022066116333008,
    "prediction_norm": 6.781866550445557,
    "timestamp": "2025-07-14T12:34:55.719470"
  },
  {
    "epoch": 2275,
    "batch_idx": 32,
    "loss": 0.20697595179080963,
    "learning_rate": 4.592841308745946e-05,
    "timestep_mean": 476.71875,
    "epsilon_norm": 7.411568641662598,
    "prediction_norm": 6.189743995666504,
    "timestamp": "2025-07-14T12:34:55.724655"
  },
  {
    "epoch": 2275,
    "batch_idx": 64,
    "loss": 0.2531285881996155,
    "learning_rate": 4.592841308745946e-05,
    "timestep_mean": 512.3125,
    "epsilon_norm": 7.8288893699646,
    "prediction_norm": 6.737571716308594,
    "timestamp": "2025-07-14T12:34:55.729964"
  },
  {
    "epoch": 2275,
    "batch_idx": 96,
    "loss": 0.17704597115516663,
    "learning_rate": 4.592841308745946e-05,
    "timestep_mean": 471.6875,
    "epsilon_norm": 7.68312931060791,
    "prediction_norm": 6.731707572937012,
    "timestamp": "2025-07-14T12:34:55.735071"
  },
  {
    "epoch": 2275,
    "batch_idx": 128,
    "loss": 0.3587039113044739,
    "learning_rate": 4.592841308745946e-05,
    "timestep_mean": 472.71875,
    "epsilon_norm": 8.295510292053223,
    "prediction_norm": 6.661843776702881,
    "timestamp": "2025-07-14T12:34:55.740236"
  },
  {
    "epoch": 2275,
    "batch_idx": 160,
    "loss": 0.35500723123550415,
    "learning_rate": 4.592841308745946e-05,
    "timestep_mean": 453.09375,
    "epsilon_norm": 8.604976654052734,
    "prediction_norm": 6.806029319763184,
    "timestamp": "2025-07-14T12:34:55.745269"
  },
  {
    "epoch": 2275,
    "batch_idx": 192,
    "loss": 0.42604732513427734,
    "learning_rate": 4.592841308745946e-05,
    "timestep_mean": 560.375,
    "epsilon_norm": 4.019279479980469,
    "prediction_norm": 3.003419876098633,
    "timestamp": "2025-07-14T12:34:55.750531"
  },
  {
    "epoch": 2276,
    "batch_idx": 0,
    "loss": 0.26712268590927124,
    "learning_rate": 4.625778772944164e-05,
    "timestep_mean": 487.28125,
    "epsilon_norm": 7.719787120819092,
    "prediction_norm": 5.971258640289307,
    "timestamp": "2025-07-14T12:34:55.755639"
  },
  {
    "epoch": 2276,
    "batch_idx": 32,
    "loss": 0.17193660140037537,
    "learning_rate": 4.625778772944164e-05,
    "timestep_mean": 518.0,
    "epsilon_norm": 7.0790019035339355,
    "prediction_norm": 6.470602512359619,
    "timestamp": "2025-07-14T12:34:55.760774"
  },
  {
    "epoch": 2276,
    "batch_idx": 64,
    "loss": 0.23105494678020477,
    "learning_rate": 4.625778772944164e-05,
    "timestep_mean": 581.96875,
    "epsilon_norm": 8.978827476501465,
    "prediction_norm": 8.047407150268555,
    "timestamp": "2025-07-14T12:34:55.765982"
  },
  {
    "epoch": 2276,
    "batch_idx": 96,
    "loss": 0.39342015981674194,
    "learning_rate": 4.625778772944164e-05,
    "timestep_mean": 390.78125,
    "epsilon_norm": 8.545708656311035,
    "prediction_norm": 6.894431114196777,
    "timestamp": "2025-07-14T12:34:55.771246"
  },
  {
    "epoch": 2276,
    "batch_idx": 128,
    "loss": 0.35378730297088623,
    "learning_rate": 4.625778772944164e-05,
    "timestep_mean": 372.5625,
    "epsilon_norm": 8.853596687316895,
    "prediction_norm": 7.631654262542725,
    "timestamp": "2025-07-14T12:34:55.776427"
  },
  {
    "epoch": 2276,
    "batch_idx": 160,
    "loss": 0.48196738958358765,
    "learning_rate": 4.625778772944164e-05,
    "timestep_mean": 471.46875,
    "epsilon_norm": 8.644253730773926,
    "prediction_norm": 5.765547275543213,
    "timestamp": "2025-07-14T12:34:55.781540"
  },
  {
    "epoch": 2276,
    "batch_idx": 192,
    "loss": 0.16622725129127502,
    "learning_rate": 4.625778772944164e-05,
    "timestep_mean": 542.75,
    "epsilon_norm": 4.376646518707275,
    "prediction_norm": 3.8018672466278076,
    "timestamp": "2025-07-14T12:34:55.786776"
  },
  {
    "epoch": 2277,
    "batch_idx": 0,
    "loss": 0.14567483961582184,
    "learning_rate": 4.658828193522738e-05,
    "timestep_mean": 462.9375,
    "epsilon_norm": 6.976831436157227,
    "prediction_norm": 5.836800575256348,
    "timestamp": "2025-07-14T12:34:55.792476"
  },
  {
    "epoch": 2277,
    "batch_idx": 32,
    "loss": 0.20020759105682373,
    "learning_rate": 4.658828193522738e-05,
    "timestep_mean": 509.28125,
    "epsilon_norm": 6.894193649291992,
    "prediction_norm": 5.875478744506836,
    "timestamp": "2025-07-14T12:34:55.797541"
  },
  {
    "epoch": 2277,
    "batch_idx": 64,
    "loss": 0.12361107766628265,
    "learning_rate": 4.658828193522738e-05,
    "timestep_mean": 528.46875,
    "epsilon_norm": 7.497558116912842,
    "prediction_norm": 7.208860397338867,
    "timestamp": "2025-07-14T12:34:55.802618"
  },
  {
    "epoch": 2277,
    "batch_idx": 96,
    "loss": 0.22512832283973694,
    "learning_rate": 4.658828193522738e-05,
    "timestep_mean": 431.5,
    "epsilon_norm": 7.531618118286133,
    "prediction_norm": 6.306921005249023,
    "timestamp": "2025-07-14T12:34:55.807671"
  },
  {
    "epoch": 2277,
    "batch_idx": 128,
    "loss": 0.2755523920059204,
    "learning_rate": 4.658828193522738e-05,
    "timestep_mean": 408.25,
    "epsilon_norm": 7.587795734405518,
    "prediction_norm": 6.501263618469238,
    "timestamp": "2025-07-14T12:34:55.812575"
  },
  {
    "epoch": 2277,
    "batch_idx": 160,
    "loss": 0.5496054291725159,
    "learning_rate": 4.658828193522738e-05,
    "timestep_mean": 337.4375,
    "epsilon_norm": 8.099444389343262,
    "prediction_norm": 6.101632595062256,
    "timestamp": "2025-07-14T12:34:55.817513"
  },
  {
    "epoch": 2277,
    "batch_idx": 192,
    "loss": 0.1260148584842682,
    "learning_rate": 4.658828193522738e-05,
    "timestep_mean": 625.875,
    "epsilon_norm": 4.298640251159668,
    "prediction_norm": 3.978738307952881,
    "timestamp": "2025-07-14T12:34:55.822432"
  },
  {
    "epoch": 2278,
    "batch_idx": 0,
    "loss": 0.1757003515958786,
    "learning_rate": 4.691989488935513e-05,
    "timestep_mean": 454.8125,
    "epsilon_norm": 7.689565181732178,
    "prediction_norm": 6.867109775543213,
    "timestamp": "2025-07-14T12:34:55.827305"
  },
  {
    "epoch": 2278,
    "batch_idx": 32,
    "loss": 0.4528547525405884,
    "learning_rate": 4.691989488935513e-05,
    "timestep_mean": 500.6875,
    "epsilon_norm": 8.570882797241211,
    "prediction_norm": 6.154420375823975,
    "timestamp": "2025-07-14T12:34:55.832224"
  },
  {
    "epoch": 2278,
    "batch_idx": 64,
    "loss": 0.31565868854522705,
    "learning_rate": 4.691989488935513e-05,
    "timestep_mean": 592.875,
    "epsilon_norm": 9.002718925476074,
    "prediction_norm": 8.136733055114746,
    "timestamp": "2025-07-14T12:34:55.837066"
  },
  {
    "epoch": 2278,
    "batch_idx": 96,
    "loss": 0.21303200721740723,
    "learning_rate": 4.691989488935513e-05,
    "timestep_mean": 455.03125,
    "epsilon_norm": 6.358805179595947,
    "prediction_norm": 5.330104827880859,
    "timestamp": "2025-07-14T12:34:55.842054"
  },
  {
    "epoch": 2278,
    "batch_idx": 128,
    "loss": 0.12514808773994446,
    "learning_rate": 4.691989488935513e-05,
    "timestep_mean": 531.34375,
    "epsilon_norm": 6.5006022453308105,
    "prediction_norm": 5.568521976470947,
    "timestamp": "2025-07-14T12:34:55.846898"
  },
  {
    "epoch": 2278,
    "batch_idx": 160,
    "loss": 0.255453497171402,
    "learning_rate": 4.691989488935513e-05,
    "timestep_mean": 408.0625,
    "epsilon_norm": 7.503598213195801,
    "prediction_norm": 6.58780574798584,
    "timestamp": "2025-07-14T12:34:55.851753"
  },
  {
    "epoch": 2278,
    "batch_idx": 192,
    "loss": 0.3530977964401245,
    "learning_rate": 4.691989488935513e-05,
    "timestep_mean": 503.5,
    "epsilon_norm": 5.495978832244873,
    "prediction_norm": 5.033856391906738,
    "timestamp": "2025-07-14T12:34:55.856622"
  },
  {
    "epoch": 2279,
    "batch_idx": 0,
    "loss": 0.1560976505279541,
    "learning_rate": 4.7252625773602865e-05,
    "timestep_mean": 524.0625,
    "epsilon_norm": 6.086991310119629,
    "prediction_norm": 5.758582592010498,
    "timestamp": "2025-07-14T12:34:55.861509"
  },
  {
    "epoch": 2279,
    "batch_idx": 32,
    "loss": 0.37073981761932373,
    "learning_rate": 4.7252625773602865e-05,
    "timestep_mean": 521.90625,
    "epsilon_norm": 9.928285598754883,
    "prediction_norm": 8.158583641052246,
    "timestamp": "2025-07-14T12:34:55.866350"
  },
  {
    "epoch": 2279,
    "batch_idx": 64,
    "loss": 0.10971613228321075,
    "learning_rate": 4.7252625773602865e-05,
    "timestep_mean": 652.8125,
    "epsilon_norm": 7.9764251708984375,
    "prediction_norm": 7.017088413238525,
    "timestamp": "2025-07-14T12:34:55.871296"
  },
  {
    "epoch": 2279,
    "batch_idx": 96,
    "loss": 0.20990215241909027,
    "learning_rate": 4.7252625773602865e-05,
    "timestep_mean": 561.03125,
    "epsilon_norm": 8.72106647491455,
    "prediction_norm": 7.133674144744873,
    "timestamp": "2025-07-14T12:34:55.876199"
  },
  {
    "epoch": 2279,
    "batch_idx": 128,
    "loss": 0.26896828413009644,
    "learning_rate": 4.7252625773602865e-05,
    "timestep_mean": 473.15625,
    "epsilon_norm": 7.621956825256348,
    "prediction_norm": 6.799654483795166,
    "timestamp": "2025-07-14T12:34:55.881124"
  },
  {
    "epoch": 2279,
    "batch_idx": 160,
    "loss": 0.41773873567581177,
    "learning_rate": 4.7252625773602865e-05,
    "timestep_mean": 519.0625,
    "epsilon_norm": 8.273050308227539,
    "prediction_norm": 7.00136137008667,
    "timestamp": "2025-07-14T12:34:55.885980"
  },
  {
    "epoch": 2279,
    "batch_idx": 192,
    "loss": 0.12387998402118683,
    "learning_rate": 4.7252625773602865e-05,
    "timestep_mean": 597.625,
    "epsilon_norm": 4.5511932373046875,
    "prediction_norm": 4.127173900604248,
    "timestamp": "2025-07-14T12:34:55.890848"
  },
  {
    "epoch": 2280,
    "batch_idx": 0,
    "loss": 0.26304346323013306,
    "learning_rate": 4.758647376699035e-05,
    "timestep_mean": 524.9375,
    "epsilon_norm": 7.8408918380737305,
    "prediction_norm": 6.561542987823486,
    "timestamp": "2025-07-14T12:34:55.895682"
  },
  {
    "epoch": 2280,
    "batch_idx": 32,
    "loss": 0.2141161859035492,
    "learning_rate": 4.758647376699035e-05,
    "timestep_mean": 508.03125,
    "epsilon_norm": 7.459468841552734,
    "prediction_norm": 6.492082118988037,
    "timestamp": "2025-07-14T12:34:55.900610"
  },
  {
    "epoch": 2280,
    "batch_idx": 64,
    "loss": 0.308706670999527,
    "learning_rate": 4.758647376699035e-05,
    "timestep_mean": 463.65625,
    "epsilon_norm": 8.329398155212402,
    "prediction_norm": 6.933418273925781,
    "timestamp": "2025-07-14T12:34:55.905503"
  },
  {
    "epoch": 2280,
    "batch_idx": 96,
    "loss": 0.17159691452980042,
    "learning_rate": 4.758647376699035e-05,
    "timestep_mean": 532.28125,
    "epsilon_norm": 7.5002312660217285,
    "prediction_norm": 7.041795253753662,
    "timestamp": "2025-07-14T12:34:55.910362"
  },
  {
    "epoch": 2280,
    "batch_idx": 128,
    "loss": 0.2739919424057007,
    "learning_rate": 4.758647376699035e-05,
    "timestep_mean": 537.53125,
    "epsilon_norm": 7.469657897949219,
    "prediction_norm": 6.012845039367676,
    "timestamp": "2025-07-14T12:34:55.915293"
  },
  {
    "epoch": 2280,
    "batch_idx": 160,
    "loss": 0.25673437118530273,
    "learning_rate": 4.758647376699035e-05,
    "timestep_mean": 472.9375,
    "epsilon_norm": 7.178996562957764,
    "prediction_norm": 6.052853107452393,
    "timestamp": "2025-07-14T12:34:55.920253"
  },
  {
    "epoch": 2280,
    "batch_idx": 192,
    "loss": 0.08226963132619858,
    "learning_rate": 4.758647376699035e-05,
    "timestep_mean": 516.5,
    "epsilon_norm": 3.8556618690490723,
    "prediction_norm": 3.8499534130096436,
    "timestamp": "2025-07-14T12:34:55.925125"
  },
  {
    "epoch": 2281,
    "batch_idx": 0,
    "loss": 0.17946693301200867,
    "learning_rate": 4.792143804578045e-05,
    "timestep_mean": 604.5625,
    "epsilon_norm": 8.466662406921387,
    "prediction_norm": 7.39377498626709,
    "timestamp": "2025-07-14T12:34:55.930018"
  },
  {
    "epoch": 2281,
    "batch_idx": 32,
    "loss": 0.2848159074783325,
    "learning_rate": 4.792143804578045e-05,
    "timestep_mean": 475.84375,
    "epsilon_norm": 7.905838489532471,
    "prediction_norm": 6.7819132804870605,
    "timestamp": "2025-07-14T12:34:55.934859"
  },
  {
    "epoch": 2281,
    "batch_idx": 64,
    "loss": 0.35313329100608826,
    "learning_rate": 4.792143804578045e-05,
    "timestep_mean": 355.46875,
    "epsilon_norm": 8.406563758850098,
    "prediction_norm": 6.6520609855651855,
    "timestamp": "2025-07-14T12:34:55.939731"
  },
  {
    "epoch": 2281,
    "batch_idx": 96,
    "loss": 0.3796162009239197,
    "learning_rate": 4.792143804578045e-05,
    "timestep_mean": 457.9375,
    "epsilon_norm": 7.901422023773193,
    "prediction_norm": 6.301486968994141,
    "timestamp": "2025-07-14T12:34:55.944546"
  },
  {
    "epoch": 2281,
    "batch_idx": 128,
    "loss": 0.22046734392642975,
    "learning_rate": 4.792143804578045e-05,
    "timestep_mean": 506.125,
    "epsilon_norm": 8.12814712524414,
    "prediction_norm": 7.140345573425293,
    "timestamp": "2025-07-14T12:34:55.949409"
  },
  {
    "epoch": 2281,
    "batch_idx": 160,
    "loss": 0.16818925738334656,
    "learning_rate": 4.792143804578045e-05,
    "timestep_mean": 495.5,
    "epsilon_norm": 7.913567543029785,
    "prediction_norm": 7.312149524688721,
    "timestamp": "2025-07-14T12:34:55.954256"
  },
  {
    "epoch": 2281,
    "batch_idx": 192,
    "loss": 0.26578715443611145,
    "learning_rate": 4.792143804578045e-05,
    "timestep_mean": 552.625,
    "epsilon_norm": 4.55648136138916,
    "prediction_norm": 4.053032875061035,
    "timestamp": "2025-07-14T12:34:55.959092"
  },
  {
    "epoch": 2282,
    "batch_idx": 0,
    "loss": 0.2058205008506775,
    "learning_rate": 4.8257517783482616e-05,
    "timestep_mean": 606.5,
    "epsilon_norm": 8.723034858703613,
    "prediction_norm": 7.187193393707275,
    "timestamp": "2025-07-14T12:34:55.963906"
  },
  {
    "epoch": 2282,
    "batch_idx": 32,
    "loss": 0.17044246196746826,
    "learning_rate": 4.8257517783482616e-05,
    "timestep_mean": 484.375,
    "epsilon_norm": 8.107230186462402,
    "prediction_norm": 7.137983798980713,
    "timestamp": "2025-07-14T12:34:55.968752"
  },
  {
    "epoch": 2282,
    "batch_idx": 64,
    "loss": 0.24783168733119965,
    "learning_rate": 4.8257517783482616e-05,
    "timestep_mean": 457.71875,
    "epsilon_norm": 6.822952747344971,
    "prediction_norm": 5.980834007263184,
    "timestamp": "2025-07-14T12:34:55.973573"
  },
  {
    "epoch": 2282,
    "batch_idx": 96,
    "loss": 0.23706303536891937,
    "learning_rate": 4.8257517783482616e-05,
    "timestep_mean": 480.5625,
    "epsilon_norm": 7.854062080383301,
    "prediction_norm": 6.274496555328369,
    "timestamp": "2025-07-14T12:34:55.978407"
  },
  {
    "epoch": 2282,
    "batch_idx": 128,
    "loss": 0.204517662525177,
    "learning_rate": 4.8257517783482616e-05,
    "timestep_mean": 510.3125,
    "epsilon_norm": 8.229681968688965,
    "prediction_norm": 7.194408416748047,
    "timestamp": "2025-07-14T12:34:55.983260"
  },
  {
    "epoch": 2282,
    "batch_idx": 160,
    "loss": 0.18948696553707123,
    "learning_rate": 4.8257517783482616e-05,
    "timestep_mean": 428.09375,
    "epsilon_norm": 7.82216739654541,
    "prediction_norm": 6.805300235748291,
    "timestamp": "2025-07-14T12:34:55.988197"
  },
  {
    "epoch": 2282,
    "batch_idx": 192,
    "loss": 0.0918477475643158,
    "learning_rate": 4.8257517783482616e-05,
    "timestep_mean": 472.125,
    "epsilon_norm": 3.5021040439605713,
    "prediction_norm": 3.0067191123962402,
    "timestamp": "2025-07-14T12:34:55.993015"
  },
  {
    "epoch": 2283,
    "batch_idx": 0,
    "loss": 0.3496863842010498,
    "learning_rate": 4.859471215085323e-05,
    "timestep_mean": 384.875,
    "epsilon_norm": 8.555559158325195,
    "prediction_norm": 6.65933895111084,
    "timestamp": "2025-07-14T12:34:55.997878"
  },
  {
    "epoch": 2283,
    "batch_idx": 32,
    "loss": 0.48779240250587463,
    "learning_rate": 4.859471215085323e-05,
    "timestep_mean": 457.125,
    "epsilon_norm": 8.18569564819336,
    "prediction_norm": 6.436384201049805,
    "timestamp": "2025-07-14T12:34:56.002680"
  },
  {
    "epoch": 2283,
    "batch_idx": 64,
    "loss": 0.3343515396118164,
    "learning_rate": 4.859471215085323e-05,
    "timestep_mean": 390.53125,
    "epsilon_norm": 9.114313125610352,
    "prediction_norm": 7.664881706237793,
    "timestamp": "2025-07-14T12:34:56.007523"
  },
  {
    "epoch": 2283,
    "batch_idx": 96,
    "loss": 0.2852006256580353,
    "learning_rate": 4.859471215085323e-05,
    "timestep_mean": 478.84375,
    "epsilon_norm": 8.341557502746582,
    "prediction_norm": 7.296973705291748,
    "timestamp": "2025-07-14T12:34:56.012369"
  },
  {
    "epoch": 2283,
    "batch_idx": 128,
    "loss": 0.2886160910129547,
    "learning_rate": 4.859471215085323e-05,
    "timestep_mean": 492.875,
    "epsilon_norm": 7.143287181854248,
    "prediction_norm": 5.762724876403809,
    "timestamp": "2025-07-14T12:34:56.017303"
  },
  {
    "epoch": 2283,
    "batch_idx": 160,
    "loss": 0.11146443337202072,
    "learning_rate": 4.859471215085323e-05,
    "timestep_mean": 526.90625,
    "epsilon_norm": 7.928872585296631,
    "prediction_norm": 7.5266008377075195,
    "timestamp": "2025-07-14T12:34:56.022306"
  },
  {
    "epoch": 2283,
    "batch_idx": 192,
    "loss": 1.2751952409744263,
    "learning_rate": 4.859471215085323e-05,
    "timestep_mean": 470.875,
    "epsilon_norm": 5.578515529632568,
    "prediction_norm": 3.152505874633789,
    "timestamp": "2025-07-14T12:34:56.027272"
  },
  {
    "epoch": 2284,
    "batch_idx": 0,
    "loss": 0.2426401972770691,
    "learning_rate": 4.893302031589861e-05,
    "timestep_mean": 541.03125,
    "epsilon_norm": 7.439604759216309,
    "prediction_norm": 6.100200176239014,
    "timestamp": "2025-07-14T12:34:56.032202"
  },
  {
    "epoch": 2284,
    "batch_idx": 32,
    "loss": 0.15117532014846802,
    "learning_rate": 4.893302031589861e-05,
    "timestep_mean": 569.9375,
    "epsilon_norm": 9.023855209350586,
    "prediction_norm": 8.097776412963867,
    "timestamp": "2025-07-14T12:34:56.037098"
  },
  {
    "epoch": 2284,
    "batch_idx": 64,
    "loss": 0.1325233429670334,
    "learning_rate": 4.893302031589861e-05,
    "timestep_mean": 655.875,
    "epsilon_norm": 8.345637321472168,
    "prediction_norm": 7.709322452545166,
    "timestamp": "2025-07-14T12:34:56.041975"
  },
  {
    "epoch": 2284,
    "batch_idx": 96,
    "loss": 0.2650221884250641,
    "learning_rate": 4.893302031589861e-05,
    "timestep_mean": 559.21875,
    "epsilon_norm": 7.583968639373779,
    "prediction_norm": 6.752447128295898,
    "timestamp": "2025-07-14T12:34:56.046888"
  },
  {
    "epoch": 2284,
    "batch_idx": 128,
    "loss": 0.30508852005004883,
    "learning_rate": 4.893302031589861e-05,
    "timestep_mean": 497.125,
    "epsilon_norm": 8.293889045715332,
    "prediction_norm": 6.5362420082092285,
    "timestamp": "2025-07-14T12:34:56.051852"
  },
  {
    "epoch": 2284,
    "batch_idx": 160,
    "loss": 0.21364013850688934,
    "learning_rate": 4.893302031589861e-05,
    "timestep_mean": 531.875,
    "epsilon_norm": 8.287521362304688,
    "prediction_norm": 7.871005058288574,
    "timestamp": "2025-07-14T12:34:56.056890"
  },
  {
    "epoch": 2284,
    "batch_idx": 192,
    "loss": 0.1404946744441986,
    "learning_rate": 4.893302031589861e-05,
    "timestep_mean": 492.75,
    "epsilon_norm": 3.2628016471862793,
    "prediction_norm": 3.4325966835021973,
    "timestamp": "2025-07-14T12:34:56.061991"
  },
  {
    "epoch": 2285,
    "batch_idx": 0,
    "loss": 0.26683881878852844,
    "learning_rate": 4.927244144387724e-05,
    "timestep_mean": 586.6875,
    "epsilon_norm": 7.9991865158081055,
    "prediction_norm": 6.87736177444458,
    "timestamp": "2025-07-14T12:34:56.067075"
  },
  {
    "epoch": 2285,
    "batch_idx": 32,
    "loss": 0.24914194643497467,
    "learning_rate": 4.927244144387724e-05,
    "timestep_mean": 460.25,
    "epsilon_norm": 7.757663249969482,
    "prediction_norm": 6.6162109375,
    "timestamp": "2025-07-14T12:34:56.072203"
  },
  {
    "epoch": 2285,
    "batch_idx": 64,
    "loss": 0.2290855497121811,
    "learning_rate": 4.927244144387724e-05,
    "timestep_mean": 475.15625,
    "epsilon_norm": 8.093914985656738,
    "prediction_norm": 6.9924187660217285,
    "timestamp": "2025-07-14T12:34:56.077282"
  },
  {
    "epoch": 2285,
    "batch_idx": 96,
    "loss": 0.3357902467250824,
    "learning_rate": 4.927244144387724e-05,
    "timestep_mean": 514.53125,
    "epsilon_norm": 8.07552719116211,
    "prediction_norm": 6.61718225479126,
    "timestamp": "2025-07-14T12:34:56.082339"
  },
  {
    "epoch": 2285,
    "batch_idx": 128,
    "loss": 0.205536350607872,
    "learning_rate": 4.927244144387724e-05,
    "timestep_mean": 526.96875,
    "epsilon_norm": 7.1899847984313965,
    "prediction_norm": 6.521572113037109,
    "timestamp": "2025-07-14T12:34:56.087438"
  },
  {
    "epoch": 2285,
    "batch_idx": 160,
    "loss": 0.22616742551326752,
    "learning_rate": 4.927244144387724e-05,
    "timestep_mean": 530.59375,
    "epsilon_norm": 7.673978805541992,
    "prediction_norm": 6.494936943054199,
    "timestamp": "2025-07-14T12:34:56.092439"
  },
  {
    "epoch": 2285,
    "batch_idx": 192,
    "loss": 0.6378246545791626,
    "learning_rate": 4.927244144387724e-05,
    "timestep_mean": 584.0,
    "epsilon_norm": 4.554463863372803,
    "prediction_norm": 3.3177618980407715,
    "timestamp": "2025-07-14T12:34:56.097398"
  },
  {
    "epoch": 2286,
    "batch_idx": 0,
    "loss": 0.3004917502403259,
    "learning_rate": 4.9612974697301e-05,
    "timestep_mean": 473.5625,
    "epsilon_norm": 8.452743530273438,
    "prediction_norm": 7.3616623878479,
    "timestamp": "2025-07-14T12:34:56.102499"
  },
  {
    "epoch": 2286,
    "batch_idx": 32,
    "loss": 0.28449344635009766,
    "learning_rate": 4.9612974697301e-05,
    "timestep_mean": 505.84375,
    "epsilon_norm": 7.179260730743408,
    "prediction_norm": 5.3822245597839355,
    "timestamp": "2025-07-14T12:34:56.107661"
  },
  {
    "epoch": 2286,
    "batch_idx": 64,
    "loss": 0.2457275092601776,
    "learning_rate": 4.9612974697301e-05,
    "timestep_mean": 528.8125,
    "epsilon_norm": 8.001341819763184,
    "prediction_norm": 6.329434871673584,
    "timestamp": "2025-07-14T12:34:56.112960"
  },
  {
    "epoch": 2286,
    "batch_idx": 96,
    "loss": 0.17725595831871033,
    "learning_rate": 4.9612974697301e-05,
    "timestep_mean": 534.0625,
    "epsilon_norm": 6.659813404083252,
    "prediction_norm": 5.7636823654174805,
    "timestamp": "2025-07-14T12:34:56.118015"
  },
  {
    "epoch": 2286,
    "batch_idx": 128,
    "loss": 0.3457549810409546,
    "learning_rate": 4.9612974697301e-05,
    "timestep_mean": 430.34375,
    "epsilon_norm": 7.328707695007324,
    "prediction_norm": 5.721480369567871,
    "timestamp": "2025-07-14T12:34:56.122932"
  },
  {
    "epoch": 2286,
    "batch_idx": 160,
    "loss": 0.2683712840080261,
    "learning_rate": 4.9612974697301e-05,
    "timestep_mean": 490.125,
    "epsilon_norm": 8.2217378616333,
    "prediction_norm": 6.651997089385986,
    "timestamp": "2025-07-14T12:34:56.127917"
  },
  {
    "epoch": 2286,
    "batch_idx": 192,
    "loss": 0.3283061683177948,
    "learning_rate": 4.9612974697301e-05,
    "timestep_mean": 379.875,
    "epsilon_norm": 2.761931896209717,
    "prediction_norm": 2.227968215942383,
    "timestamp": "2025-07-14T12:34:56.132843"
  },
  {
    "epoch": 2287,
    "batch_idx": 0,
    "loss": 0.41710156202316284,
    "learning_rate": 4.995461923593811e-05,
    "timestep_mean": 487.71875,
    "epsilon_norm": 8.120024681091309,
    "prediction_norm": 6.096289157867432,
    "timestamp": "2025-07-14T12:34:56.137989"
  },
  {
    "epoch": 2287,
    "batch_idx": 32,
    "loss": 0.3204594850540161,
    "learning_rate": 4.995461923593811e-05,
    "timestep_mean": 469.4375,
    "epsilon_norm": 6.787332057952881,
    "prediction_norm": 5.79341983795166,
    "timestamp": "2025-07-14T12:34:56.143043"
  },
  {
    "epoch": 2287,
    "batch_idx": 64,
    "loss": 0.4046591520309448,
    "learning_rate": 4.995461923593811e-05,
    "timestep_mean": 530.09375,
    "epsilon_norm": 9.664551734924316,
    "prediction_norm": 8.34450912475586,
    "timestamp": "2025-07-14T12:34:56.148063"
  },
  {
    "epoch": 2287,
    "batch_idx": 96,
    "loss": 0.1941850185394287,
    "learning_rate": 4.995461923593811e-05,
    "timestep_mean": 522.90625,
    "epsilon_norm": 8.10787582397461,
    "prediction_norm": 6.933799743652344,
    "timestamp": "2025-07-14T12:34:56.153224"
  },
  {
    "epoch": 2287,
    "batch_idx": 128,
    "loss": 0.09467068314552307,
    "learning_rate": 4.995461923593811e-05,
    "timestep_mean": 597.96875,
    "epsilon_norm": 8.422752380371094,
    "prediction_norm": 8.329290390014648,
    "timestamp": "2025-07-14T12:34:56.158171"
  },
  {
    "epoch": 2287,
    "batch_idx": 160,
    "loss": 0.4986458122730255,
    "learning_rate": 4.995461923593811e-05,
    "timestep_mean": 414.59375,
    "epsilon_norm": 8.604537963867188,
    "prediction_norm": 6.22605037689209,
    "timestamp": "2025-07-14T12:34:56.163396"
  },
  {
    "epoch": 2287,
    "batch_idx": 192,
    "loss": 0.2902880311012268,
    "learning_rate": 4.995461923593811e-05,
    "timestep_mean": 573.625,
    "epsilon_norm": 4.632201671600342,
    "prediction_norm": 4.0935211181640625,
    "timestamp": "2025-07-14T12:34:56.168423"
  },
  {
    "epoch": 2288,
    "batch_idx": 0,
    "loss": 0.3951537609100342,
    "learning_rate": 5.029737421681461e-05,
    "timestep_mean": 506.875,
    "epsilon_norm": 9.171256065368652,
    "prediction_norm": 6.790978908538818,
    "timestamp": "2025-07-14T12:34:56.173490"
  },
  {
    "epoch": 2288,
    "batch_idx": 32,
    "loss": 0.23916015028953552,
    "learning_rate": 5.029737421681461e-05,
    "timestep_mean": 462.75,
    "epsilon_norm": 8.373416900634766,
    "prediction_norm": 7.198589324951172,
    "timestamp": "2025-07-14T12:34:56.178637"
  },
  {
    "epoch": 2288,
    "batch_idx": 64,
    "loss": 0.27649688720703125,
    "learning_rate": 5.029737421681461e-05,
    "timestep_mean": 482.03125,
    "epsilon_norm": 8.030891418457031,
    "prediction_norm": 6.372375965118408,
    "timestamp": "2025-07-14T12:34:56.183728"
  },
  {
    "epoch": 2288,
    "batch_idx": 96,
    "loss": 0.25103050470352173,
    "learning_rate": 5.029737421681461e-05,
    "timestep_mean": 549.71875,
    "epsilon_norm": 7.685946941375732,
    "prediction_norm": 7.018520832061768,
    "timestamp": "2025-07-14T12:34:56.188930"
  },
  {
    "epoch": 2288,
    "batch_idx": 128,
    "loss": 0.19799670577049255,
    "learning_rate": 5.029737421681461e-05,
    "timestep_mean": 487.4375,
    "epsilon_norm": 7.139477729797363,
    "prediction_norm": 5.601950645446777,
    "timestamp": "2025-07-14T12:34:56.194084"
  },
  {
    "epoch": 2288,
    "batch_idx": 160,
    "loss": 0.1982724666595459,
    "learning_rate": 5.029737421681461e-05,
    "timestep_mean": 508.875,
    "epsilon_norm": 7.103512763977051,
    "prediction_norm": 6.633593559265137,
    "timestamp": "2025-07-14T12:34:56.199180"
  },
  {
    "epoch": 2288,
    "batch_idx": 192,
    "loss": 0.17608335614204407,
    "learning_rate": 5.029737421681461e-05,
    "timestep_mean": 725.125,
    "epsilon_norm": 3.7738959789276123,
    "prediction_norm": 3.383071184158325,
    "timestamp": "2025-07-14T12:34:56.204471"
  },
  {
    "epoch": 2289,
    "batch_idx": 0,
    "loss": 0.36282575130462646,
    "learning_rate": 5.0641238794216434e-05,
    "timestep_mean": 575.28125,
    "epsilon_norm": 7.90952730178833,
    "prediction_norm": 6.477133274078369,
    "timestamp": "2025-07-14T12:34:56.209621"
  },
  {
    "epoch": 2289,
    "batch_idx": 32,
    "loss": 0.3335084319114685,
    "learning_rate": 5.0641238794216434e-05,
    "timestep_mean": 544.40625,
    "epsilon_norm": 8.02588939666748,
    "prediction_norm": 6.671430587768555,
    "timestamp": "2025-07-14T12:34:56.214695"
  },
  {
    "epoch": 2289,
    "batch_idx": 64,
    "loss": 0.31628042459487915,
    "learning_rate": 5.0641238794216434e-05,
    "timestep_mean": 479.75,
    "epsilon_norm": 8.109726905822754,
    "prediction_norm": 6.318804740905762,
    "timestamp": "2025-07-14T12:34:56.219886"
  },
  {
    "epoch": 2289,
    "batch_idx": 96,
    "loss": 0.36900433897972107,
    "learning_rate": 5.0641238794216434e-05,
    "timestep_mean": 457.96875,
    "epsilon_norm": 9.078258514404297,
    "prediction_norm": 7.73287296295166,
    "timestamp": "2025-07-14T12:34:56.224955"
  },
  {
    "epoch": 2289,
    "batch_idx": 128,
    "loss": 0.19247658550739288,
    "learning_rate": 5.0641238794216434e-05,
    "timestep_mean": 553.5,
    "epsilon_norm": 9.165231704711914,
    "prediction_norm": 8.909075736999512,
    "timestamp": "2025-07-14T12:34:56.230076"
  },
  {
    "epoch": 2289,
    "batch_idx": 160,
    "loss": 0.17627516388893127,
    "learning_rate": 5.0641238794216434e-05,
    "timestep_mean": 554.21875,
    "epsilon_norm": 8.181530952453613,
    "prediction_norm": 7.278285026550293,
    "timestamp": "2025-07-14T12:34:56.235166"
  },
  {
    "epoch": 2289,
    "batch_idx": 192,
    "loss": 0.07322780042886734,
    "learning_rate": 5.0641238794216434e-05,
    "timestep_mean": 452.125,
    "epsilon_norm": 3.129676580429077,
    "prediction_norm": 3.2843470573425293,
    "timestamp": "2025-07-14T12:34:56.240221"
  },
  {
    "epoch": 2290,
    "batch_idx": 0,
    "loss": 0.18284092843532562,
    "learning_rate": 5.098621211969227e-05,
    "timestep_mean": 591.78125,
    "epsilon_norm": 7.999612808227539,
    "prediction_norm": 7.363385200500488,
    "timestamp": "2025-07-14T12:34:56.245439"
  },
  {
    "epoch": 2290,
    "batch_idx": 32,
    "loss": 0.08475710451602936,
    "learning_rate": 5.098621211969227e-05,
    "timestep_mean": 598.4375,
    "epsilon_norm": 7.660425186157227,
    "prediction_norm": 7.519084930419922,
    "timestamp": "2025-07-14T12:34:56.250733"
  },
  {
    "epoch": 2290,
    "batch_idx": 64,
    "loss": 0.5028380751609802,
    "learning_rate": 5.098621211969227e-05,
    "timestep_mean": 533.75,
    "epsilon_norm": 8.497733116149902,
    "prediction_norm": 6.066587924957275,
    "timestamp": "2025-07-14T12:34:56.255849"
  },
  {
    "epoch": 2290,
    "batch_idx": 96,
    "loss": 0.2083314210176468,
    "learning_rate": 5.098621211969227e-05,
    "timestep_mean": 497.40625,
    "epsilon_norm": 8.042678833007812,
    "prediction_norm": 6.5305705070495605,
    "timestamp": "2025-07-14T12:34:56.260892"
  },
  {
    "epoch": 2290,
    "batch_idx": 128,
    "loss": 0.2750145196914673,
    "learning_rate": 5.098621211969227e-05,
    "timestep_mean": 574.28125,
    "epsilon_norm": 7.859809875488281,
    "prediction_norm": 6.896103858947754,
    "timestamp": "2025-07-14T12:34:56.265935"
  },
  {
    "epoch": 2290,
    "batch_idx": 160,
    "loss": 0.3406352996826172,
    "learning_rate": 5.098621211969227e-05,
    "timestep_mean": 443.125,
    "epsilon_norm": 7.290460586547852,
    "prediction_norm": 6.053735256195068,
    "timestamp": "2025-07-14T12:34:56.271118"
  },
  {
    "epoch": 2290,
    "batch_idx": 192,
    "loss": 0.10636777430772781,
    "learning_rate": 5.098621211969227e-05,
    "timestep_mean": 600.625,
    "epsilon_norm": 3.4567956924438477,
    "prediction_norm": 3.1478230953216553,
    "timestamp": "2025-07-14T12:34:56.276241"
  },
  {
    "epoch": 2291,
    "batch_idx": 0,
    "loss": 0.10322851687669754,
    "learning_rate": 5.133229334205445e-05,
    "timestep_mean": 576.625,
    "epsilon_norm": 7.669719696044922,
    "prediction_norm": 7.123345851898193,
    "timestamp": "2025-07-14T12:34:56.281555"
  },
  {
    "epoch": 2291,
    "batch_idx": 32,
    "loss": 0.2055833339691162,
    "learning_rate": 5.133229334205445e-05,
    "timestep_mean": 485.8125,
    "epsilon_norm": 8.356353759765625,
    "prediction_norm": 6.838517189025879,
    "timestamp": "2025-07-14T12:34:56.286647"
  },
  {
    "epoch": 2291,
    "batch_idx": 64,
    "loss": 0.10773633420467377,
    "learning_rate": 5.133229334205445e-05,
    "timestep_mean": 576.625,
    "epsilon_norm": 7.711887359619141,
    "prediction_norm": 7.167654991149902,
    "timestamp": "2025-07-14T12:34:56.291819"
  },
  {
    "epoch": 2291,
    "batch_idx": 96,
    "loss": 0.2880651354789734,
    "learning_rate": 5.133229334205445e-05,
    "timestep_mean": 510.6875,
    "epsilon_norm": 8.229646682739258,
    "prediction_norm": 6.741303443908691,
    "timestamp": "2025-07-14T12:34:56.297007"
  },
  {
    "epoch": 2291,
    "batch_idx": 128,
    "loss": 0.24880921840667725,
    "learning_rate": 5.133229334205445e-05,
    "timestep_mean": 467.65625,
    "epsilon_norm": 8.360461235046387,
    "prediction_norm": 6.831578254699707,
    "timestamp": "2025-07-14T12:34:56.302133"
  },
  {
    "epoch": 2291,
    "batch_idx": 160,
    "loss": 0.47358059883117676,
    "learning_rate": 5.133229334205445e-05,
    "timestep_mean": 429.40625,
    "epsilon_norm": 8.872891426086426,
    "prediction_norm": 7.158336639404297,
    "timestamp": "2025-07-14T12:34:56.307202"
  },
  {
    "epoch": 2291,
    "batch_idx": 192,
    "loss": 0.6500961780548096,
    "learning_rate": 5.133229334205445e-05,
    "timestep_mean": 380.25,
    "epsilon_norm": 3.9571290016174316,
    "prediction_norm": 2.4597978591918945,
    "timestamp": "2025-07-14T12:34:56.312350"
  },
  {
    "epoch": 2292,
    "batch_idx": 0,
    "loss": 0.3084409832954407,
    "learning_rate": 5.16794816073821e-05,
    "timestep_mean": 492.3125,
    "epsilon_norm": 8.881193161010742,
    "prediction_norm": 7.693029880523682,
    "timestamp": "2025-07-14T12:34:56.317571"
  },
  {
    "epoch": 2292,
    "batch_idx": 32,
    "loss": 0.2484499216079712,
    "learning_rate": 5.16794816073821e-05,
    "timestep_mean": 549.8125,
    "epsilon_norm": 6.4747724533081055,
    "prediction_norm": 5.410682678222656,
    "timestamp": "2025-07-14T12:34:56.322550"
  },
  {
    "epoch": 2292,
    "batch_idx": 64,
    "loss": 0.2045523226261139,
    "learning_rate": 5.16794816073821e-05,
    "timestep_mean": 550.53125,
    "epsilon_norm": 7.82608699798584,
    "prediction_norm": 7.087782382965088,
    "timestamp": "2025-07-14T12:34:56.327616"
  },
  {
    "epoch": 2292,
    "batch_idx": 96,
    "loss": 0.2472662776708603,
    "learning_rate": 5.16794816073821e-05,
    "timestep_mean": 547.3125,
    "epsilon_norm": 8.012730598449707,
    "prediction_norm": 7.448741912841797,
    "timestamp": "2025-07-14T12:34:56.332849"
  },
  {
    "epoch": 2292,
    "batch_idx": 128,
    "loss": 0.18890805542469025,
    "learning_rate": 5.16794816073821e-05,
    "timestep_mean": 530.96875,
    "epsilon_norm": 6.912955284118652,
    "prediction_norm": 6.0032854080200195,
    "timestamp": "2025-07-14T12:34:56.338222"
  },
  {
    "epoch": 2292,
    "batch_idx": 160,
    "loss": 0.12202130258083344,
    "learning_rate": 5.16794816073821e-05,
    "timestep_mean": 537.4375,
    "epsilon_norm": 7.716407775878906,
    "prediction_norm": 7.283902168273926,
    "timestamp": "2025-07-14T12:34:56.343274"
  },
  {
    "epoch": 2292,
    "batch_idx": 192,
    "loss": 0.3796064257621765,
    "learning_rate": 5.16794816073821e-05,
    "timestep_mean": 358.375,
    "epsilon_norm": 3.509958028793335,
    "prediction_norm": 2.0784454345703125,
    "timestamp": "2025-07-14T12:34:56.348344"
  },
  {
    "epoch": 2293,
    "batch_idx": 0,
    "loss": 0.2511433959007263,
    "learning_rate": 5.2027776059022636e-05,
    "timestep_mean": 484.0,
    "epsilon_norm": 7.517512798309326,
    "prediction_norm": 6.237220287322998,
    "timestamp": "2025-07-14T12:34:56.353560"
  },
  {
    "epoch": 2293,
    "batch_idx": 32,
    "loss": 0.26815393567085266,
    "learning_rate": 5.2027776059022636e-05,
    "timestep_mean": 583.90625,
    "epsilon_norm": 8.744573593139648,
    "prediction_norm": 7.849466323852539,
    "timestamp": "2025-07-14T12:34:56.358698"
  },
  {
    "epoch": 2293,
    "batch_idx": 64,
    "loss": 0.1713567078113556,
    "learning_rate": 5.2027776059022636e-05,
    "timestep_mean": 549.53125,
    "epsilon_norm": 8.637301445007324,
    "prediction_norm": 7.861236095428467,
    "timestamp": "2025-07-14T12:34:56.363852"
  },
  {
    "epoch": 2293,
    "batch_idx": 96,
    "loss": 0.18915939331054688,
    "learning_rate": 5.2027776059022636e-05,
    "timestep_mean": 492.71875,
    "epsilon_norm": 7.876934051513672,
    "prediction_norm": 6.829017639160156,
    "timestamp": "2025-07-14T12:34:56.369013"
  },
  {
    "epoch": 2293,
    "batch_idx": 128,
    "loss": 0.35084956884384155,
    "learning_rate": 5.2027776059022636e-05,
    "timestep_mean": 428.15625,
    "epsilon_norm": 7.677058219909668,
    "prediction_norm": 6.4916276931762695,
    "timestamp": "2025-07-14T12:34:56.374045"
  },
  {
    "epoch": 2293,
    "batch_idx": 160,
    "loss": 0.3541925847530365,
    "learning_rate": 5.2027776059022636e-05,
    "timestep_mean": 524.71875,
    "epsilon_norm": 8.787951469421387,
    "prediction_norm": 7.554197788238525,
    "timestamp": "2025-07-14T12:34:56.379259"
  },
  {
    "epoch": 2293,
    "batch_idx": 192,
    "loss": 0.3034401535987854,
    "learning_rate": 5.2027776059022636e-05,
    "timestep_mean": 487.125,
    "epsilon_norm": 4.817756175994873,
    "prediction_norm": 4.147158145904541,
    "timestamp": "2025-07-14T12:34:56.384391"
  },
  {
    "epoch": 2294,
    "batch_idx": 0,
    "loss": 0.3238581120967865,
    "learning_rate": 5.237717583759426e-05,
    "timestep_mean": 461.4375,
    "epsilon_norm": 7.8777008056640625,
    "prediction_norm": 6.077123641967773,
    "timestamp": "2025-07-14T12:34:56.389603"
  },
  {
    "epoch": 2294,
    "batch_idx": 32,
    "loss": 0.23402482271194458,
    "learning_rate": 5.237717583759426e-05,
    "timestep_mean": 499.5625,
    "epsilon_norm": 8.83247184753418,
    "prediction_norm": 8.502943992614746,
    "timestamp": "2025-07-14T12:34:56.394803"
  },
  {
    "epoch": 2294,
    "batch_idx": 64,
    "loss": 0.162684828042984,
    "learning_rate": 5.237717583759426e-05,
    "timestep_mean": 560.625,
    "epsilon_norm": 7.358288288116455,
    "prediction_norm": 6.570675849914551,
    "timestamp": "2025-07-14T12:34:56.399909"
  },
  {
    "epoch": 2294,
    "batch_idx": 96,
    "loss": 0.32098376750946045,
    "learning_rate": 5.237717583759426e-05,
    "timestep_mean": 535.53125,
    "epsilon_norm": 7.724733829498291,
    "prediction_norm": 6.6076812744140625,
    "timestamp": "2025-07-14T12:34:56.405071"
  },
  {
    "epoch": 2294,
    "batch_idx": 128,
    "loss": 0.20859143137931824,
    "learning_rate": 5.237717583759426e-05,
    "timestep_mean": 412.25,
    "epsilon_norm": 7.338778018951416,
    "prediction_norm": 6.255876064300537,
    "timestamp": "2025-07-14T12:34:56.409980"
  },
  {
    "epoch": 2294,
    "batch_idx": 160,
    "loss": 0.426358699798584,
    "learning_rate": 5.237717583759426e-05,
    "timestep_mean": 427.75,
    "epsilon_norm": 7.636579513549805,
    "prediction_norm": 5.867249965667725,
    "timestamp": "2025-07-14T12:34:56.414924"
  },
  {
    "epoch": 2294,
    "batch_idx": 192,
    "loss": 0.34536948800086975,
    "learning_rate": 5.237717583759426e-05,
    "timestep_mean": 499.75,
    "epsilon_norm": 3.432359457015991,
    "prediction_norm": 2.761964797973633,
    "timestamp": "2025-07-14T12:34:56.420137"
  },
  {
    "epoch": 2295,
    "batch_idx": 0,
    "loss": 0.47136324644088745,
    "learning_rate": 5.272768008098754e-05,
    "timestep_mean": 512.875,
    "epsilon_norm": 7.965235233306885,
    "prediction_norm": 5.28540563583374,
    "timestamp": "2025-07-14T12:34:56.425385"
  },
  {
    "epoch": 2295,
    "batch_idx": 32,
    "loss": 0.3334987163543701,
    "learning_rate": 5.272768008098754e-05,
    "timestep_mean": 577.8125,
    "epsilon_norm": 9.101564407348633,
    "prediction_norm": 7.519506931304932,
    "timestamp": "2025-07-14T12:34:56.430479"
  },
  {
    "epoch": 2295,
    "batch_idx": 64,
    "loss": 0.32891249656677246,
    "learning_rate": 5.272768008098754e-05,
    "timestep_mean": 462.59375,
    "epsilon_norm": 7.779142379760742,
    "prediction_norm": 6.401009559631348,
    "timestamp": "2025-07-14T12:34:56.435607"
  },
  {
    "epoch": 2295,
    "batch_idx": 96,
    "loss": 0.3926718235015869,
    "learning_rate": 5.272768008098754e-05,
    "timestep_mean": 464.09375,
    "epsilon_norm": 7.663327217102051,
    "prediction_norm": 6.231934547424316,
    "timestamp": "2025-07-14T12:34:56.440706"
  },
  {
    "epoch": 2295,
    "batch_idx": 128,
    "loss": 0.23599711060523987,
    "learning_rate": 5.272768008098754e-05,
    "timestep_mean": 527.4375,
    "epsilon_norm": 7.977023124694824,
    "prediction_norm": 6.718702793121338,
    "timestamp": "2025-07-14T12:34:56.445873"
  },
  {
    "epoch": 2295,
    "batch_idx": 160,
    "loss": 0.44170522689819336,
    "learning_rate": 5.272768008098754e-05,
    "timestep_mean": 381.9375,
    "epsilon_norm": 6.756991863250732,
    "prediction_norm": 5.003628730773926,
    "timestamp": "2025-07-14T12:34:56.451324"
  },
  {
    "epoch": 2295,
    "batch_idx": 192,
    "loss": 0.33669134974479675,
    "learning_rate": 5.272768008098754e-05,
    "timestep_mean": 426.0,
    "epsilon_norm": 3.9663643836975098,
    "prediction_norm": 3.140420913696289,
    "timestamp": "2025-07-14T12:34:56.456329"
  },
  {
    "epoch": 2296,
    "batch_idx": 0,
    "loss": 0.37196314334869385,
    "learning_rate": 5.307928792436816e-05,
    "timestep_mean": 488.03125,
    "epsilon_norm": 8.385790824890137,
    "prediction_norm": 6.870016098022461,
    "timestamp": "2025-07-14T12:34:56.461419"
  },
  {
    "epoch": 2296,
    "batch_idx": 32,
    "loss": 0.2807857096195221,
    "learning_rate": 5.307928792436816e-05,
    "timestep_mean": 470.375,
    "epsilon_norm": 7.1967949867248535,
    "prediction_norm": 5.641582012176514,
    "timestamp": "2025-07-14T12:34:56.466579"
  },
  {
    "epoch": 2296,
    "batch_idx": 64,
    "loss": 0.18783064186573029,
    "learning_rate": 5.307928792436816e-05,
    "timestep_mean": 559.15625,
    "epsilon_norm": 8.91834545135498,
    "prediction_norm": 7.5453572273254395,
    "timestamp": "2025-07-14T12:34:56.471483"
  },
  {
    "epoch": 2296,
    "batch_idx": 96,
    "loss": 0.24903196096420288,
    "learning_rate": 5.307928792436816e-05,
    "timestep_mean": 539.875,
    "epsilon_norm": 7.511750221252441,
    "prediction_norm": 6.5327277183532715,
    "timestamp": "2025-07-14T12:34:56.476627"
  },
  {
    "epoch": 2296,
    "batch_idx": 128,
    "loss": 0.36571717262268066,
    "learning_rate": 5.307928792436816e-05,
    "timestep_mean": 510.96875,
    "epsilon_norm": 7.6319899559021,
    "prediction_norm": 5.856987476348877,
    "timestamp": "2025-07-14T12:34:56.481723"
  },
  {
    "epoch": 2296,
    "batch_idx": 160,
    "loss": 0.21781831979751587,
    "learning_rate": 5.307928792436816e-05,
    "timestep_mean": 494.65625,
    "epsilon_norm": 7.178933620452881,
    "prediction_norm": 6.419722080230713,
    "timestamp": "2025-07-14T12:34:56.486786"
  },
  {
    "epoch": 2296,
    "batch_idx": 192,
    "loss": 0.21126699447631836,
    "learning_rate": 5.307928792436816e-05,
    "timestep_mean": 671.625,
    "epsilon_norm": 2.672891855239868,
    "prediction_norm": 2.1316521167755127,
    "timestamp": "2025-07-14T12:34:56.491901"
  },
  {
    "epoch": 2297,
    "batch_idx": 0,
    "loss": 0.33762556314468384,
    "learning_rate": 5.343199850017887e-05,
    "timestep_mean": 465.5,
    "epsilon_norm": 7.568611145019531,
    "prediction_norm": 6.239960670471191,
    "timestamp": "2025-07-14T12:34:56.496834"
  },
  {
    "epoch": 2297,
    "batch_idx": 32,
    "loss": 0.3136017322540283,
    "learning_rate": 5.343199850017887e-05,
    "timestep_mean": 470.8125,
    "epsilon_norm": 7.21991491317749,
    "prediction_norm": 6.329634189605713,
    "timestamp": "2025-07-14T12:34:56.501721"
  },
  {
    "epoch": 2297,
    "batch_idx": 64,
    "loss": 0.28057539463043213,
    "learning_rate": 5.343199850017887e-05,
    "timestep_mean": 544.21875,
    "epsilon_norm": 7.195251941680908,
    "prediction_norm": 5.696454048156738,
    "timestamp": "2025-07-14T12:34:56.506518"
  },
  {
    "epoch": 2297,
    "batch_idx": 96,
    "loss": 0.1979541778564453,
    "learning_rate": 5.343199850017887e-05,
    "timestep_mean": 533.8125,
    "epsilon_norm": 7.8929243087768555,
    "prediction_norm": 7.295807838439941,
    "timestamp": "2025-07-14T12:34:56.511443"
  },
  {
    "epoch": 2297,
    "batch_idx": 128,
    "loss": 0.32255586981773376,
    "learning_rate": 5.343199850017887e-05,
    "timestep_mean": 464.28125,
    "epsilon_norm": 8.557991027832031,
    "prediction_norm": 7.0151801109313965,
    "timestamp": "2025-07-14T12:34:56.516325"
  },
  {
    "epoch": 2297,
    "batch_idx": 160,
    "loss": 0.18132302165031433,
    "learning_rate": 5.343199850017887e-05,
    "timestep_mean": 511.03125,
    "epsilon_norm": 8.288433074951172,
    "prediction_norm": 7.979154586791992,
    "timestamp": "2025-07-14T12:34:56.521349"
  },
  {
    "epoch": 2297,
    "batch_idx": 192,
    "loss": 0.2810071110725403,
    "learning_rate": 5.343199850017887e-05,
    "timestep_mean": 439.25,
    "epsilon_norm": 3.931173086166382,
    "prediction_norm": 2.7961692810058594,
    "timestamp": "2025-07-14T12:34:56.526192"
  },
  {
    "epoch": 2298,
    "batch_idx": 0,
    "loss": 0.2528868019580841,
    "learning_rate": 5.378581093814104e-05,
    "timestep_mean": 543.375,
    "epsilon_norm": 8.601573944091797,
    "prediction_norm": 7.231271743774414,
    "timestamp": "2025-07-14T12:34:56.531133"
  },
  {
    "epoch": 2298,
    "batch_idx": 32,
    "loss": 0.19775688648223877,
    "learning_rate": 5.378581093814104e-05,
    "timestep_mean": 524.78125,
    "epsilon_norm": 7.5997796058654785,
    "prediction_norm": 6.988284587860107,
    "timestamp": "2025-07-14T12:34:56.536102"
  },
  {
    "epoch": 2298,
    "batch_idx": 64,
    "loss": 0.315101683139801,
    "learning_rate": 5.378581093814104e-05,
    "timestep_mean": 492.53125,
    "epsilon_norm": 8.980037689208984,
    "prediction_norm": 7.930654048919678,
    "timestamp": "2025-07-14T12:34:56.541233"
  },
  {
    "epoch": 2298,
    "batch_idx": 96,
    "loss": 0.34187161922454834,
    "learning_rate": 5.378581093814104e-05,
    "timestep_mean": 505.5,
    "epsilon_norm": 8.690655708312988,
    "prediction_norm": 7.218707084655762,
    "timestamp": "2025-07-14T12:34:56.546354"
  },
  {
    "epoch": 2298,
    "batch_idx": 128,
    "loss": 0.1957627534866333,
    "learning_rate": 5.378581093814104e-05,
    "timestep_mean": 478.84375,
    "epsilon_norm": 7.438674449920654,
    "prediction_norm": 6.7918219566345215,
    "timestamp": "2025-07-14T12:34:56.551545"
  },
  {
    "epoch": 2298,
    "batch_idx": 160,
    "loss": 0.10583242774009705,
    "learning_rate": 5.378581093814104e-05,
    "timestep_mean": 499.96875,
    "epsilon_norm": 7.321800231933594,
    "prediction_norm": 6.492275238037109,
    "timestamp": "2025-07-14T12:34:56.556567"
  },
  {
    "epoch": 2298,
    "batch_idx": 192,
    "loss": 0.45607492327690125,
    "learning_rate": 5.378581093814104e-05,
    "timestep_mean": 463.5,
    "epsilon_norm": 3.593371868133545,
    "prediction_norm": 2.813030958175659,
    "timestamp": "2025-07-14T12:34:56.561608"
  },
  {
    "epoch": 2299,
    "batch_idx": 0,
    "loss": 0.24017514288425446,
    "learning_rate": 5.4140724365258124e-05,
    "timestep_mean": 487.0,
    "epsilon_norm": 7.54452657699585,
    "prediction_norm": 6.598568916320801,
    "timestamp": "2025-07-14T12:34:56.566763"
  },
  {
    "epoch": 2299,
    "batch_idx": 32,
    "loss": 0.232599675655365,
    "learning_rate": 5.4140724365258124e-05,
    "timestep_mean": 540.65625,
    "epsilon_norm": 7.6171040534973145,
    "prediction_norm": 6.772517204284668,
    "timestamp": "2025-07-14T12:34:56.571952"
  },
  {
    "epoch": 2299,
    "batch_idx": 64,
    "loss": 0.310390830039978,
    "learning_rate": 5.4140724365258124e-05,
    "timestep_mean": 420.25,
    "epsilon_norm": 7.996092796325684,
    "prediction_norm": 6.603540420532227,
    "timestamp": "2025-07-14T12:34:56.577047"
  },
  {
    "epoch": 2299,
    "batch_idx": 96,
    "loss": 0.12799875438213348,
    "learning_rate": 5.4140724365258124e-05,
    "timestep_mean": 570.125,
    "epsilon_norm": 7.256709575653076,
    "prediction_norm": 6.703498840332031,
    "timestamp": "2025-07-14T12:34:56.582099"
  },
  {
    "epoch": 2299,
    "batch_idx": 128,
    "loss": 0.1992713063955307,
    "learning_rate": 5.4140724365258124e-05,
    "timestep_mean": 466.25,
    "epsilon_norm": 7.813490867614746,
    "prediction_norm": 6.616788387298584,
    "timestamp": "2025-07-14T12:34:56.587275"
  },
  {
    "epoch": 2299,
    "batch_idx": 160,
    "loss": 0.22881701588630676,
    "learning_rate": 5.4140724365258124e-05,
    "timestep_mean": 523.4375,
    "epsilon_norm": 7.941833019256592,
    "prediction_norm": 6.642348289489746,
    "timestamp": "2025-07-14T12:34:56.592549"
  },
  {
    "epoch": 2299,
    "batch_idx": 192,
    "loss": 0.06355670839548111,
    "learning_rate": 5.4140724365258124e-05,
    "timestep_mean": 493.0,
    "epsilon_norm": 4.665925025939941,
    "prediction_norm": 4.8871049880981445,
    "timestamp": "2025-07-14T12:34:56.597604"
  },
  {
    "epoch": 2300,
    "batch_idx": 0,
    "loss": 0.23421573638916016,
    "learning_rate": 5.449673790581616e-05,
    "timestep_mean": 565.125,
    "epsilon_norm": 8.327892303466797,
    "prediction_norm": 7.791943073272705,
    "timestamp": "2025-07-14T12:34:56.602657"
  },
  {
    "epoch": 2300,
    "batch_idx": 32,
    "loss": 0.1422623097896576,
    "learning_rate": 5.449673790581616e-05,
    "timestep_mean": 546.15625,
    "epsilon_norm": 7.573051929473877,
    "prediction_norm": 6.922206401824951,
    "timestamp": "2025-07-14T12:34:56.607869"
  },
  {
    "epoch": 2300,
    "batch_idx": 64,
    "loss": 0.17482918500900269,
    "learning_rate": 5.449673790581616e-05,
    "timestep_mean": 547.71875,
    "epsilon_norm": 8.501592636108398,
    "prediction_norm": 7.5965986251831055,
    "timestamp": "2025-07-14T12:34:56.612769"
  },
  {
    "epoch": 2300,
    "batch_idx": 96,
    "loss": 0.3093993067741394,
    "learning_rate": 5.449673790581616e-05,
    "timestep_mean": 434.90625,
    "epsilon_norm": 7.414426326751709,
    "prediction_norm": 6.4286723136901855,
    "timestamp": "2025-07-14T12:34:56.617679"
  },
  {
    "epoch": 2300,
    "batch_idx": 128,
    "loss": 0.20444107055664062,
    "learning_rate": 5.449673790581616e-05,
    "timestep_mean": 473.65625,
    "epsilon_norm": 7.0442023277282715,
    "prediction_norm": 6.086751937866211,
    "timestamp": "2025-07-14T12:34:56.622623"
  },
  {
    "epoch": 2300,
    "batch_idx": 160,
    "loss": 0.3666726350784302,
    "learning_rate": 5.449673790581616e-05,
    "timestep_mean": 497.53125,
    "epsilon_norm": 7.962136268615723,
    "prediction_norm": 6.506361961364746,
    "timestamp": "2025-07-14T12:34:56.627558"
  },
  {
    "epoch": 2300,
    "batch_idx": 192,
    "loss": 0.19691795110702515,
    "learning_rate": 5.449673790581616e-05,
    "timestep_mean": 563.875,
    "epsilon_norm": 4.198229789733887,
    "prediction_norm": 3.8745410442352295,
    "timestamp": "2025-07-14T12:34:56.632717"
  },
  {
    "epoch": 2301,
    "batch_idx": 0,
    "loss": 0.21510060131549835,
    "learning_rate": 5.485385068138721e-05,
    "timestep_mean": 555.0,
    "epsilon_norm": 8.325373649597168,
    "prediction_norm": 6.850228786468506,
    "timestamp": "2025-07-14T12:34:56.637912"
  },
  {
    "epoch": 2301,
    "batch_idx": 32,
    "loss": 0.3395877778530121,
    "learning_rate": 5.485385068138721e-05,
    "timestep_mean": 526.125,
    "epsilon_norm": 7.585537433624268,
    "prediction_norm": 6.924656391143799,
    "timestamp": "2025-07-14T12:34:56.642983"
  },
  {
    "epoch": 2301,
    "batch_idx": 64,
    "loss": 0.44198495149612427,
    "learning_rate": 5.485385068138721e-05,
    "timestep_mean": 504.6875,
    "epsilon_norm": 9.344759941101074,
    "prediction_norm": 7.548184871673584,
    "timestamp": "2025-07-14T12:34:56.647993"
  },
  {
    "epoch": 2301,
    "batch_idx": 96,
    "loss": 0.21497374773025513,
    "learning_rate": 5.485385068138721e-05,
    "timestep_mean": 483.0625,
    "epsilon_norm": 6.633162975311279,
    "prediction_norm": 5.864450931549072,
    "timestamp": "2025-07-14T12:34:56.653092"
  },
  {
    "epoch": 2301,
    "batch_idx": 128,
    "loss": 0.49085384607315063,
    "learning_rate": 5.485385068138721e-05,
    "timestep_mean": 520.4375,
    "epsilon_norm": 8.83629322052002,
    "prediction_norm": 6.578785419464111,
    "timestamp": "2025-07-14T12:34:56.658218"
  },
  {
    "epoch": 2301,
    "batch_idx": 160,
    "loss": 0.3192002773284912,
    "learning_rate": 5.485385068138721e-05,
    "timestep_mean": 490.5,
    "epsilon_norm": 7.31128454208374,
    "prediction_norm": 6.163556098937988,
    "timestamp": "2025-07-14T12:34:56.663418"
  },
  {
    "epoch": 2301,
    "batch_idx": 192,
    "loss": 0.11141620576381683,
    "learning_rate": 5.485385068138721e-05,
    "timestep_mean": 563.5,
    "epsilon_norm": 3.615007162094116,
    "prediction_norm": 3.1967787742614746,
    "timestamp": "2025-07-14T12:34:56.668561"
  },
  {
    "epoch": 2302,
    "batch_idx": 0,
    "loss": 0.35361218452453613,
    "learning_rate": 5.5212061810831165e-05,
    "timestep_mean": 550.875,
    "epsilon_norm": 8.548005104064941,
    "prediction_norm": 7.24234676361084,
    "timestamp": "2025-07-14T12:34:56.673689"
  },
  {
    "epoch": 2302,
    "batch_idx": 32,
    "loss": 0.2435085028409958,
    "learning_rate": 5.5212061810831165e-05,
    "timestep_mean": 337.1875,
    "epsilon_norm": 6.104724407196045,
    "prediction_norm": 4.838100910186768,
    "timestamp": "2025-07-14T12:34:56.678863"
  },
  {
    "epoch": 2302,
    "batch_idx": 64,
    "loss": 0.33077359199523926,
    "learning_rate": 5.5212061810831165e-05,
    "timestep_mean": 460.875,
    "epsilon_norm": 8.130565643310547,
    "prediction_norm": 6.316916465759277,
    "timestamp": "2025-07-14T12:34:56.684095"
  },
  {
    "epoch": 2302,
    "batch_idx": 96,
    "loss": 0.2359398603439331,
    "learning_rate": 5.5212061810831165e-05,
    "timestep_mean": 486.0625,
    "epsilon_norm": 7.071300506591797,
    "prediction_norm": 5.566501140594482,
    "timestamp": "2025-07-14T12:34:56.689225"
  },
  {
    "epoch": 2302,
    "batch_idx": 128,
    "loss": 0.186222106218338,
    "learning_rate": 5.5212061810831165e-05,
    "timestep_mean": 470.0,
    "epsilon_norm": 6.878737449645996,
    "prediction_norm": 6.321247100830078,
    "timestamp": "2025-07-14T12:34:56.694322"
  },
  {
    "epoch": 2302,
    "batch_idx": 160,
    "loss": 0.3840372562408447,
    "learning_rate": 5.5212061810831165e-05,
    "timestep_mean": 448.40625,
    "epsilon_norm": 7.4119720458984375,
    "prediction_norm": 6.018802642822266,
    "timestamp": "2025-07-14T12:34:56.699533"
  },
  {
    "epoch": 2302,
    "batch_idx": 192,
    "loss": 0.14095550775527954,
    "learning_rate": 5.5212061810831165e-05,
    "timestep_mean": 474.0,
    "epsilon_norm": 3.36295747756958,
    "prediction_norm": 3.112976551055908,
    "timestamp": "2025-07-14T12:34:56.704453"
  },
  {
    "epoch": 2303,
    "batch_idx": 0,
    "loss": 0.3266443908214569,
    "learning_rate": 5.557137041029732e-05,
    "timestep_mean": 478.71875,
    "epsilon_norm": 8.167207717895508,
    "prediction_norm": 6.520300388336182,
    "timestamp": "2025-07-14T12:34:56.709511"
  },
  {
    "epoch": 2303,
    "batch_idx": 32,
    "loss": 0.2790628671646118,
    "learning_rate": 5.557137041029732e-05,
    "timestep_mean": 521.71875,
    "epsilon_norm": 7.613072395324707,
    "prediction_norm": 6.586374282836914,
    "timestamp": "2025-07-14T12:34:56.714581"
  },
  {
    "epoch": 2303,
    "batch_idx": 64,
    "loss": 0.3713696599006653,
    "learning_rate": 5.557137041029732e-05,
    "timestep_mean": 393.125,
    "epsilon_norm": 9.692357063293457,
    "prediction_norm": 7.662054061889648,
    "timestamp": "2025-07-14T12:34:56.719697"
  },
  {
    "epoch": 2303,
    "batch_idx": 96,
    "loss": 0.3562510311603546,
    "learning_rate": 5.557137041029732e-05,
    "timestep_mean": 505.71875,
    "epsilon_norm": 8.416666984558105,
    "prediction_norm": 6.865705490112305,
    "timestamp": "2025-07-14T12:34:56.724926"
  },
  {
    "epoch": 2303,
    "batch_idx": 128,
    "loss": 0.24618777632713318,
    "learning_rate": 5.557137041029732e-05,
    "timestep_mean": 503.5,
    "epsilon_norm": 8.430724143981934,
    "prediction_norm": 6.848703861236572,
    "timestamp": "2025-07-14T12:34:56.730123"
  },
  {
    "epoch": 2303,
    "batch_idx": 160,
    "loss": 0.2963210940361023,
    "learning_rate": 5.557137041029732e-05,
    "timestep_mean": 495.59375,
    "epsilon_norm": 7.9826531410217285,
    "prediction_norm": 7.012240886688232,
    "timestamp": "2025-07-14T12:34:56.735250"
  },
  {
    "epoch": 2303,
    "batch_idx": 192,
    "loss": 0.1573958545923233,
    "learning_rate": 5.557137041029732e-05,
    "timestep_mean": 462.125,
    "epsilon_norm": 3.349440574645996,
    "prediction_norm": 3.633692979812622,
    "timestamp": "2025-07-14T12:34:56.740356"
  },
  {
    "epoch": 2304,
    "batch_idx": 0,
    "loss": 0.23629924654960632,
    "learning_rate": 5.593177559322783e-05,
    "timestep_mean": 522.84375,
    "epsilon_norm": 8.919337272644043,
    "prediction_norm": 7.729024410247803,
    "timestamp": "2025-07-14T12:34:56.745452"
  },
  {
    "epoch": 2304,
    "batch_idx": 32,
    "loss": 0.5262116193771362,
    "learning_rate": 5.593177559322783e-05,
    "timestep_mean": 462.75,
    "epsilon_norm": 7.768518447875977,
    "prediction_norm": 6.269116401672363,
    "timestamp": "2025-07-14T12:34:56.750768"
  },
  {
    "epoch": 2304,
    "batch_idx": 64,
    "loss": 0.22882604598999023,
    "learning_rate": 5.593177559322783e-05,
    "timestep_mean": 451.78125,
    "epsilon_norm": 8.479118347167969,
    "prediction_norm": 6.976922988891602,
    "timestamp": "2025-07-14T12:34:56.755824"
  },
  {
    "epoch": 2304,
    "batch_idx": 96,
    "loss": 0.22251129150390625,
    "learning_rate": 5.593177559322783e-05,
    "timestep_mean": 574.8125,
    "epsilon_norm": 8.300270080566406,
    "prediction_norm": 7.301779747009277,
    "timestamp": "2025-07-14T12:34:56.760903"
  },
  {
    "epoch": 2304,
    "batch_idx": 128,
    "loss": 0.29076263308525085,
    "learning_rate": 5.593177559322783e-05,
    "timestep_mean": 619.53125,
    "epsilon_norm": 8.215445518493652,
    "prediction_norm": 6.817623138427734,
    "timestamp": "2025-07-14T12:34:56.766091"
  },
  {
    "epoch": 2304,
    "batch_idx": 160,
    "loss": 0.2138909250497818,
    "learning_rate": 5.593177559322783e-05,
    "timestep_mean": 455.6875,
    "epsilon_norm": 7.535299301147461,
    "prediction_norm": 6.839357852935791,
    "timestamp": "2025-07-14T12:34:56.771402"
  },
  {
    "epoch": 2304,
    "batch_idx": 192,
    "loss": 0.410680890083313,
    "learning_rate": 5.593177559322783e-05,
    "timestep_mean": 539.25,
    "epsilon_norm": 4.6940155029296875,
    "prediction_norm": 3.991727828979492,
    "timestamp": "2025-07-14T12:34:56.776530"
  },
  {
    "epoch": 2305,
    "batch_idx": 0,
    "loss": 0.2604905664920807,
    "learning_rate": 5.629327647035849e-05,
    "timestep_mean": 388.8125,
    "epsilon_norm": 7.0014872550964355,
    "prediction_norm": 5.2818193435668945,
    "timestamp": "2025-07-14T12:34:56.781648"
  },
  {
    "epoch": 2305,
    "batch_idx": 32,
    "loss": 0.3793606162071228,
    "learning_rate": 5.629327647035849e-05,
    "timestep_mean": 471.8125,
    "epsilon_norm": 8.618230819702148,
    "prediction_norm": 6.835339069366455,
    "timestamp": "2025-07-14T12:34:56.786778"
  },
  {
    "epoch": 2305,
    "batch_idx": 64,
    "loss": 0.3011966347694397,
    "learning_rate": 5.629327647035849e-05,
    "timestep_mean": 370.28125,
    "epsilon_norm": 6.824827671051025,
    "prediction_norm": 5.2316179275512695,
    "timestamp": "2025-07-14T12:34:56.791960"
  },
  {
    "epoch": 2305,
    "batch_idx": 96,
    "loss": 0.23044666647911072,
    "learning_rate": 5.629327647035849e-05,
    "timestep_mean": 511.3125,
    "epsilon_norm": 7.844704627990723,
    "prediction_norm": 7.020326137542725,
    "timestamp": "2025-07-14T12:34:56.797140"
  },
  {
    "epoch": 2305,
    "batch_idx": 128,
    "loss": 0.4270758032798767,
    "learning_rate": 5.629327647035849e-05,
    "timestep_mean": 542.375,
    "epsilon_norm": 8.3297758102417,
    "prediction_norm": 6.8074951171875,
    "timestamp": "2025-07-14T12:34:56.802272"
  },
  {
    "epoch": 2305,
    "batch_idx": 160,
    "loss": 0.28802490234375,
    "learning_rate": 5.629327647035849e-05,
    "timestep_mean": 427.375,
    "epsilon_norm": 8.386656761169434,
    "prediction_norm": 6.900585651397705,
    "timestamp": "2025-07-14T12:34:56.807433"
  },
  {
    "epoch": 2305,
    "batch_idx": 192,
    "loss": 0.6698997020721436,
    "learning_rate": 5.629327647035849e-05,
    "timestep_mean": 493.25,
    "epsilon_norm": 3.8836896419525146,
    "prediction_norm": 2.76696515083313,
    "timestamp": "2025-07-14T12:34:56.812583"
  },
  {
    "epoch": 2306,
    "batch_idx": 0,
    "loss": 0.22578342258930206,
    "learning_rate": 5.6655872149721806e-05,
    "timestep_mean": 475.375,
    "epsilon_norm": 8.31353759765625,
    "prediction_norm": 7.310543537139893,
    "timestamp": "2025-07-14T12:34:56.817850"
  },
  {
    "epoch": 2306,
    "batch_idx": 32,
    "loss": 0.21456106007099152,
    "learning_rate": 5.6655872149721806e-05,
    "timestep_mean": 531.5625,
    "epsilon_norm": 9.483259201049805,
    "prediction_norm": 8.70571517944336,
    "timestamp": "2025-07-14T12:34:56.822926"
  },
  {
    "epoch": 2306,
    "batch_idx": 64,
    "loss": 0.2249952107667923,
    "learning_rate": 5.6655872149721806e-05,
    "timestep_mean": 549.5,
    "epsilon_norm": 7.6240010261535645,
    "prediction_norm": 6.867846488952637,
    "timestamp": "2025-07-14T12:34:56.828056"
  },
  {
    "epoch": 2306,
    "batch_idx": 96,
    "loss": 0.35199007391929626,
    "learning_rate": 5.6655872149721806e-05,
    "timestep_mean": 457.3125,
    "epsilon_norm": 8.161930084228516,
    "prediction_norm": 6.66502571105957,
    "timestamp": "2025-07-14T12:34:56.833211"
  },
  {
    "epoch": 2306,
    "batch_idx": 128,
    "loss": 0.27311012148857117,
    "learning_rate": 5.6655872149721806e-05,
    "timestep_mean": 506.59375,
    "epsilon_norm": 8.428510665893555,
    "prediction_norm": 6.999026298522949,
    "timestamp": "2025-07-14T12:34:56.838439"
  },
  {
    "epoch": 2306,
    "batch_idx": 160,
    "loss": 0.403739869594574,
    "learning_rate": 5.6655872149721806e-05,
    "timestep_mean": 520.3125,
    "epsilon_norm": 9.644853591918945,
    "prediction_norm": 7.879380702972412,
    "timestamp": "2025-07-14T12:34:56.843635"
  },
  {
    "epoch": 2306,
    "batch_idx": 192,
    "loss": 0.35527071356773376,
    "learning_rate": 5.6655872149721806e-05,
    "timestep_mean": 453.25,
    "epsilon_norm": 3.0964090824127197,
    "prediction_norm": 2.388110399246216,
    "timestamp": "2025-07-14T12:34:56.848736"
  },
  {
    "epoch": 2307,
    "batch_idx": 0,
    "loss": 0.1500071883201599,
    "learning_rate": 5.70195617366492e-05,
    "timestep_mean": 575.3125,
    "epsilon_norm": 7.632837772369385,
    "prediction_norm": 6.550416946411133,
    "timestamp": "2025-07-14T12:34:56.853890"
  },
  {
    "epoch": 2307,
    "batch_idx": 32,
    "loss": 0.18095538020133972,
    "learning_rate": 5.70195617366492e-05,
    "timestep_mean": 547.0,
    "epsilon_norm": 8.358981132507324,
    "prediction_norm": 7.205241680145264,
    "timestamp": "2025-07-14T12:34:56.859295"
  },
  {
    "epoch": 2307,
    "batch_idx": 64,
    "loss": 0.2658618688583374,
    "learning_rate": 5.70195617366492e-05,
    "timestep_mean": 511.71875,
    "epsilon_norm": 7.619632244110107,
    "prediction_norm": 6.4953436851501465,
    "timestamp": "2025-07-14T12:34:56.864354"
  },
  {
    "epoch": 2307,
    "batch_idx": 96,
    "loss": 0.20361711084842682,
    "learning_rate": 5.70195617366492e-05,
    "timestep_mean": 469.21875,
    "epsilon_norm": 7.03495454788208,
    "prediction_norm": 6.2596611976623535,
    "timestamp": "2025-07-14T12:34:56.869486"
  },
  {
    "epoch": 2307,
    "batch_idx": 128,
    "loss": 0.20123614370822906,
    "learning_rate": 5.70195617366492e-05,
    "timestep_mean": 503.28125,
    "epsilon_norm": 8.304200172424316,
    "prediction_norm": 7.319944381713867,
    "timestamp": "2025-07-14T12:34:56.874685"
  },
  {
    "epoch": 2307,
    "batch_idx": 160,
    "loss": 0.29393261671066284,
    "learning_rate": 5.70195617366492e-05,
    "timestep_mean": 585.03125,
    "epsilon_norm": 9.088807106018066,
    "prediction_norm": 7.594192981719971,
    "timestamp": "2025-07-14T12:34:56.879993"
  },
  {
    "epoch": 2307,
    "batch_idx": 192,
    "loss": 0.17931562662124634,
    "learning_rate": 5.70195617366492e-05,
    "timestep_mean": 525.625,
    "epsilon_norm": 2.8328802585601807,
    "prediction_norm": 2.484346389770508,
    "timestamp": "2025-07-14T12:34:56.885140"
  },
  {
    "epoch": 2308,
    "batch_idx": 0,
    "loss": 0.2337368130683899,
    "learning_rate": 5.73843443337725e-05,
    "timestep_mean": 493.65625,
    "epsilon_norm": 8.239911079406738,
    "prediction_norm": 7.502532005310059,
    "timestamp": "2025-07-14T12:34:56.890282"
  },
  {
    "epoch": 2308,
    "batch_idx": 32,
    "loss": 0.22239141166210175,
    "learning_rate": 5.73843443337725e-05,
    "timestep_mean": 558.125,
    "epsilon_norm": 8.581391334533691,
    "prediction_norm": 7.315526962280273,
    "timestamp": "2025-07-14T12:34:56.895357"
  },
  {
    "epoch": 2308,
    "batch_idx": 64,
    "loss": 0.1626381278038025,
    "learning_rate": 5.73843443337725e-05,
    "timestep_mean": 508.46875,
    "epsilon_norm": 8.101266860961914,
    "prediction_norm": 7.170721054077148,
    "timestamp": "2025-07-14T12:34:56.900604"
  },
  {
    "epoch": 2308,
    "batch_idx": 96,
    "loss": 0.22035256028175354,
    "learning_rate": 5.73843443337725e-05,
    "timestep_mean": 507.0,
    "epsilon_norm": 7.824528217315674,
    "prediction_norm": 6.134221076965332,
    "timestamp": "2025-07-14T12:34:56.905687"
  },
  {
    "epoch": 2308,
    "batch_idx": 128,
    "loss": 0.41746824979782104,
    "learning_rate": 5.73843443337725e-05,
    "timestep_mean": 468.1875,
    "epsilon_norm": 8.10893726348877,
    "prediction_norm": 7.345935344696045,
    "timestamp": "2025-07-14T12:34:56.910743"
  },
  {
    "epoch": 2308,
    "batch_idx": 160,
    "loss": 0.20874768495559692,
    "learning_rate": 5.73843443337725e-05,
    "timestep_mean": 455.6875,
    "epsilon_norm": 7.6239094734191895,
    "prediction_norm": 6.714413166046143,
    "timestamp": "2025-07-14T12:34:56.915938"
  },
  {
    "epoch": 2308,
    "batch_idx": 192,
    "loss": 0.41036221385002136,
    "learning_rate": 5.73843443337725e-05,
    "timestep_mean": 409.875,
    "epsilon_norm": 5.027305603027344,
    "prediction_norm": 3.4873833656311035,
    "timestamp": "2025-07-14T12:34:56.921103"
  },
  {
    "epoch": 2309,
    "batch_idx": 0,
    "loss": 0.17439347505569458,
    "learning_rate": 5.7750219041027084e-05,
    "timestep_mean": 498.78125,
    "epsilon_norm": 8.304269790649414,
    "prediction_norm": 7.512249946594238,
    "timestamp": "2025-07-14T12:34:56.926266"
  },
  {
    "epoch": 2309,
    "batch_idx": 32,
    "loss": 0.13606131076812744,
    "learning_rate": 5.7750219041027084e-05,
    "timestep_mean": 567.1875,
    "epsilon_norm": 7.45844841003418,
    "prediction_norm": 7.1123046875,
    "timestamp": "2025-07-14T12:34:56.931359"
  },
  {
    "epoch": 2309,
    "batch_idx": 64,
    "loss": 0.2900939881801605,
    "learning_rate": 5.7750219041027084e-05,
    "timestep_mean": 469.875,
    "epsilon_norm": 8.436182022094727,
    "prediction_norm": 7.055172443389893,
    "timestamp": "2025-07-14T12:34:56.936483"
  },
  {
    "epoch": 2309,
    "batch_idx": 96,
    "loss": 0.21032527089118958,
    "learning_rate": 5.7750219041027084e-05,
    "timestep_mean": 519.65625,
    "epsilon_norm": 8.530324935913086,
    "prediction_norm": 7.980008602142334,
    "timestamp": "2025-07-14T12:34:56.941672"
  },
  {
    "epoch": 2309,
    "batch_idx": 128,
    "loss": 0.29843175411224365,
    "learning_rate": 5.7750219041027084e-05,
    "timestep_mean": 500.9375,
    "epsilon_norm": 8.421255111694336,
    "prediction_norm": 6.991377353668213,
    "timestamp": "2025-07-14T12:34:56.946807"
  },
  {
    "epoch": 2309,
    "batch_idx": 160,
    "loss": 0.12118580937385559,
    "learning_rate": 5.7750219041027084e-05,
    "timestep_mean": 509.9375,
    "epsilon_norm": 7.692203521728516,
    "prediction_norm": 7.378689289093018,
    "timestamp": "2025-07-14T12:34:56.951930"
  },
  {
    "epoch": 2309,
    "batch_idx": 192,
    "loss": 0.4392373263835907,
    "learning_rate": 5.7750219041027084e-05,
    "timestep_mean": 378.0,
    "epsilon_norm": 3.5821096897125244,
    "prediction_norm": 2.602323293685913,
    "timestamp": "2025-07-14T12:34:56.956998"
  },
  {
    "epoch": 2310,
    "batch_idx": 0,
    "loss": 0.4317787289619446,
    "learning_rate": 5.811718495565334e-05,
    "timestep_mean": 458.34375,
    "epsilon_norm": 8.09827995300293,
    "prediction_norm": 6.340600490570068,
    "timestamp": "2025-07-14T12:34:56.962147"
  },
  {
    "epoch": 2310,
    "batch_idx": 32,
    "loss": 0.21280154585838318,
    "learning_rate": 5.811718495565334e-05,
    "timestep_mean": 511.0625,
    "epsilon_norm": 7.231136322021484,
    "prediction_norm": 6.3109612464904785,
    "timestamp": "2025-07-14T12:34:56.967402"
  },
  {
    "epoch": 2310,
    "batch_idx": 64,
    "loss": 0.1637110412120819,
    "learning_rate": 5.811718495565334e-05,
    "timestep_mean": 514.875,
    "epsilon_norm": 6.91690731048584,
    "prediction_norm": 5.894900321960449,
    "timestamp": "2025-07-14T12:34:56.972499"
  },
  {
    "epoch": 2310,
    "batch_idx": 96,
    "loss": 0.2946496307849884,
    "learning_rate": 5.811718495565334e-05,
    "timestep_mean": 521.90625,
    "epsilon_norm": 8.925753593444824,
    "prediction_norm": 7.589571475982666,
    "timestamp": "2025-07-14T12:34:56.977470"
  },
  {
    "epoch": 2310,
    "batch_idx": 128,
    "loss": 0.3564727306365967,
    "learning_rate": 5.811718495565334e-05,
    "timestep_mean": 530.28125,
    "epsilon_norm": 9.070578575134277,
    "prediction_norm": 8.17890453338623,
    "timestamp": "2025-07-14T12:34:56.982506"
  },
  {
    "epoch": 2310,
    "batch_idx": 160,
    "loss": 0.32867616415023804,
    "learning_rate": 5.811718495565334e-05,
    "timestep_mean": 458.5,
    "epsilon_norm": 7.729340553283691,
    "prediction_norm": 5.972033500671387,
    "timestamp": "2025-07-14T12:34:56.987754"
  },
  {
    "epoch": 2310,
    "batch_idx": 192,
    "loss": 0.5802619457244873,
    "learning_rate": 5.811718495565334e-05,
    "timestep_mean": 374.875,
    "epsilon_norm": 3.5019567012786865,
    "prediction_norm": 2.8761677742004395,
    "timestamp": "2025-07-14T12:34:56.992900"
  },
  {
    "epoch": 2311,
    "batch_idx": 0,
    "loss": 0.30087047815322876,
    "learning_rate": 5.848524117219951e-05,
    "timestep_mean": 473.375,
    "epsilon_norm": 7.433311462402344,
    "prediction_norm": 6.617002010345459,
    "timestamp": "2025-07-14T12:34:56.997928"
  },
  {
    "epoch": 2311,
    "batch_idx": 32,
    "loss": 0.2339792549610138,
    "learning_rate": 5.848524117219951e-05,
    "timestep_mean": 495.71875,
    "epsilon_norm": 7.928303241729736,
    "prediction_norm": 7.46336030960083,
    "timestamp": "2025-07-14T12:34:57.003039"
  },
  {
    "epoch": 2311,
    "batch_idx": 64,
    "loss": 0.29572510719299316,
    "learning_rate": 5.848524117219951e-05,
    "timestep_mean": 560.25,
    "epsilon_norm": 7.130633354187012,
    "prediction_norm": 5.5321221351623535,
    "timestamp": "2025-07-14T12:34:57.008310"
  },
  {
    "epoch": 2311,
    "batch_idx": 96,
    "loss": 0.2586069107055664,
    "learning_rate": 5.848524117219951e-05,
    "timestep_mean": 501.9375,
    "epsilon_norm": 9.35132122039795,
    "prediction_norm": 8.265488624572754,
    "timestamp": "2025-07-14T12:34:57.013471"
  },
  {
    "epoch": 2311,
    "batch_idx": 128,
    "loss": 0.3928084075450897,
    "learning_rate": 5.848524117219951e-05,
    "timestep_mean": 460.625,
    "epsilon_norm": 8.817337989807129,
    "prediction_norm": 7.621615409851074,
    "timestamp": "2025-07-14T12:34:57.018551"
  },
  {
    "epoch": 2311,
    "batch_idx": 160,
    "loss": 0.41471877694129944,
    "learning_rate": 5.848524117219951e-05,
    "timestep_mean": 435.53125,
    "epsilon_norm": 7.256616115570068,
    "prediction_norm": 5.316674709320068,
    "timestamp": "2025-07-14T12:34:57.023571"
  },
  {
    "epoch": 2311,
    "batch_idx": 192,
    "loss": 0.15100570023059845,
    "learning_rate": 5.848524117219951e-05,
    "timestep_mean": 434.0,
    "epsilon_norm": 3.2066283226013184,
    "prediction_norm": 3.0021283626556396,
    "timestamp": "2025-07-14T12:34:57.028610"
  },
  {
    "epoch": 2312,
    "batch_idx": 0,
    "loss": 0.16466036438941956,
    "learning_rate": 5.8854386782523486e-05,
    "timestep_mean": 475.6875,
    "epsilon_norm": 7.847996234893799,
    "prediction_norm": 6.928371906280518,
    "timestamp": "2025-07-14T12:34:57.033740"
  },
  {
    "epoch": 2312,
    "batch_idx": 32,
    "loss": 0.31008949875831604,
    "learning_rate": 5.8854386782523486e-05,
    "timestep_mean": 491.15625,
    "epsilon_norm": 9.32980728149414,
    "prediction_norm": 8.196922302246094,
    "timestamp": "2025-07-14T12:34:57.038863"
  },
  {
    "epoch": 2312,
    "batch_idx": 64,
    "loss": 0.270814448595047,
    "learning_rate": 5.8854386782523486e-05,
    "timestep_mean": 466.25,
    "epsilon_norm": 6.227094650268555,
    "prediction_norm": 5.93980598449707,
    "timestamp": "2025-07-14T12:34:57.043925"
  },
  {
    "epoch": 2312,
    "batch_idx": 96,
    "loss": 0.5304243564605713,
    "learning_rate": 5.8854386782523486e-05,
    "timestep_mean": 549.09375,
    "epsilon_norm": 9.778135299682617,
    "prediction_norm": 7.080318927764893,
    "timestamp": "2025-07-14T12:34:57.049096"
  },
  {
    "epoch": 2312,
    "batch_idx": 128,
    "loss": 0.2497459053993225,
    "learning_rate": 5.8854386782523486e-05,
    "timestep_mean": 513.1875,
    "epsilon_norm": 8.069856643676758,
    "prediction_norm": 6.858660697937012,
    "timestamp": "2025-07-14T12:34:57.054233"
  },
  {
    "epoch": 2312,
    "batch_idx": 160,
    "loss": 0.2641713619232178,
    "learning_rate": 5.8854386782523486e-05,
    "timestep_mean": 555.875,
    "epsilon_norm": 6.633773326873779,
    "prediction_norm": 5.262491226196289,
    "timestamp": "2025-07-14T12:34:57.059527"
  },
  {
    "epoch": 2312,
    "batch_idx": 192,
    "loss": 0.09135336428880692,
    "learning_rate": 5.8854386782523486e-05,
    "timestep_mean": 529.875,
    "epsilon_norm": 3.8062620162963867,
    "prediction_norm": 3.4418792724609375,
    "timestamp": "2025-07-14T12:34:57.064575"
  },
  {
    "epoch": 2313,
    "batch_idx": 0,
    "loss": 0.4503939151763916,
    "learning_rate": 5.9224620875794964e-05,
    "timestep_mean": 508.125,
    "epsilon_norm": 9.738930702209473,
    "prediction_norm": 7.5365447998046875,
    "timestamp": "2025-07-14T12:34:57.069716"
  },
  {
    "epoch": 2313,
    "batch_idx": 32,
    "loss": 0.27551454305648804,
    "learning_rate": 5.9224620875794964e-05,
    "timestep_mean": 478.6875,
    "epsilon_norm": 7.425381660461426,
    "prediction_norm": 6.444515228271484,
    "timestamp": "2025-07-14T12:34:57.074957"
  },
  {
    "epoch": 2313,
    "batch_idx": 64,
    "loss": 0.12116910517215729,
    "learning_rate": 5.9224620875794964e-05,
    "timestep_mean": 526.125,
    "epsilon_norm": 7.668247222900391,
    "prediction_norm": 7.169154167175293,
    "timestamp": "2025-07-14T12:34:57.080156"
  },
  {
    "epoch": 2313,
    "batch_idx": 96,
    "loss": 0.2380559891462326,
    "learning_rate": 5.9224620875794964e-05,
    "timestep_mean": 426.4375,
    "epsilon_norm": 8.185561180114746,
    "prediction_norm": 7.007161617279053,
    "timestamp": "2025-07-14T12:34:57.085258"
  },
  {
    "epoch": 2313,
    "batch_idx": 128,
    "loss": 0.3605426549911499,
    "learning_rate": 5.9224620875794964e-05,
    "timestep_mean": 427.84375,
    "epsilon_norm": 7.478887557983398,
    "prediction_norm": 5.352866172790527,
    "timestamp": "2025-07-14T12:34:57.090500"
  },
  {
    "epoch": 2313,
    "batch_idx": 160,
    "loss": 0.37210190296173096,
    "learning_rate": 5.9224620875794964e-05,
    "timestep_mean": 407.1875,
    "epsilon_norm": 8.480535507202148,
    "prediction_norm": 6.610911846160889,
    "timestamp": "2025-07-14T12:34:57.095667"
  },
  {
    "epoch": 2313,
    "batch_idx": 192,
    "loss": 0.40523892641067505,
    "learning_rate": 5.9224620875794964e-05,
    "timestep_mean": 434.125,
    "epsilon_norm": 4.409087181091309,
    "prediction_norm": 3.7371416091918945,
    "timestamp": "2025-07-14T12:34:57.100936"
  },
  {
    "epoch": 2314,
    "batch_idx": 0,
    "loss": 0.40515750646591187,
    "learning_rate": 5.9595942538498274e-05,
    "timestep_mean": 486.3125,
    "epsilon_norm": 9.019683837890625,
    "prediction_norm": 6.8201823234558105,
    "timestamp": "2025-07-14T12:34:57.106047"
  },
  {
    "epoch": 2314,
    "batch_idx": 32,
    "loss": 0.21273373067378998,
    "learning_rate": 5.9595942538498274e-05,
    "timestep_mean": 545.8125,
    "epsilon_norm": 6.154026985168457,
    "prediction_norm": 5.367265701293945,
    "timestamp": "2025-07-14T12:34:57.111115"
  },
  {
    "epoch": 2314,
    "batch_idx": 64,
    "loss": 0.19797658920288086,
    "learning_rate": 5.9595942538498274e-05,
    "timestep_mean": 511.0625,
    "epsilon_norm": 7.872810363769531,
    "prediction_norm": 6.793405532836914,
    "timestamp": "2025-07-14T12:34:57.116333"
  },
  {
    "epoch": 2314,
    "batch_idx": 96,
    "loss": 0.6642511487007141,
    "learning_rate": 5.9595942538498274e-05,
    "timestep_mean": 508.96875,
    "epsilon_norm": 8.69543743133545,
    "prediction_norm": 5.896773815155029,
    "timestamp": "2025-07-14T12:34:57.121550"
  },
  {
    "epoch": 2314,
    "batch_idx": 128,
    "loss": 0.18063503503799438,
    "learning_rate": 5.9595942538498274e-05,
    "timestep_mean": 580.28125,
    "epsilon_norm": 6.95790433883667,
    "prediction_norm": 6.411848545074463,
    "timestamp": "2025-07-14T12:34:57.126669"
  },
  {
    "epoch": 2314,
    "batch_idx": 160,
    "loss": 0.21358084678649902,
    "learning_rate": 5.9595942538498274e-05,
    "timestep_mean": 505.09375,
    "epsilon_norm": 7.529633045196533,
    "prediction_norm": 6.393919944763184,
    "timestamp": "2025-07-14T12:34:57.131784"
  },
  {
    "epoch": 2314,
    "batch_idx": 192,
    "loss": 0.16210444271564484,
    "learning_rate": 5.9595942538498274e-05,
    "timestep_mean": 662.125,
    "epsilon_norm": 4.202343463897705,
    "prediction_norm": 3.671386480331421,
    "timestamp": "2025-07-14T12:34:57.136863"
  },
  {
    "epoch": 2315,
    "batch_idx": 0,
    "loss": 0.17298483848571777,
    "learning_rate": 5.996835085443411e-05,
    "timestep_mean": 512.8125,
    "epsilon_norm": 7.142756462097168,
    "prediction_norm": 6.51680850982666,
    "timestamp": "2025-07-14T12:34:57.142117"
  },
  {
    "epoch": 2315,
    "batch_idx": 32,
    "loss": 0.21388095617294312,
    "learning_rate": 5.996835085443411e-05,
    "timestep_mean": 517.625,
    "epsilon_norm": 8.32733154296875,
    "prediction_norm": 7.336871147155762,
    "timestamp": "2025-07-14T12:34:57.147264"
  },
  {
    "epoch": 2315,
    "batch_idx": 64,
    "loss": 0.48144155740737915,
    "learning_rate": 5.996835085443411e-05,
    "timestep_mean": 453.9375,
    "epsilon_norm": 9.11301326751709,
    "prediction_norm": 7.152636528015137,
    "timestamp": "2025-07-14T12:34:57.152289"
  },
  {
    "epoch": 2315,
    "batch_idx": 96,
    "loss": 0.24920547008514404,
    "learning_rate": 5.996835085443411e-05,
    "timestep_mean": 500.5625,
    "epsilon_norm": 8.217897415161133,
    "prediction_norm": 7.235884189605713,
    "timestamp": "2025-07-14T12:34:57.157479"
  },
  {
    "epoch": 2315,
    "batch_idx": 128,
    "loss": 0.2742263674736023,
    "learning_rate": 5.996835085443411e-05,
    "timestep_mean": 517.96875,
    "epsilon_norm": 7.225286960601807,
    "prediction_norm": 5.8996262550354,
    "timestamp": "2025-07-14T12:34:57.162604"
  },
  {
    "epoch": 2315,
    "batch_idx": 160,
    "loss": 0.2553727626800537,
    "learning_rate": 5.996835085443411e-05,
    "timestep_mean": 500.3125,
    "epsilon_norm": 8.092751502990723,
    "prediction_norm": 7.415218353271484,
    "timestamp": "2025-07-14T12:34:57.167736"
  },
  {
    "epoch": 2315,
    "batch_idx": 192,
    "loss": 0.6314641237258911,
    "learning_rate": 5.996835085443411e-05,
    "timestep_mean": 632.125,
    "epsilon_norm": 5.133471488952637,
    "prediction_norm": 3.719308614730835,
    "timestamp": "2025-07-14T12:34:57.172887"
  },
  {
    "epoch": 2316,
    "batch_idx": 0,
    "loss": 0.2788761854171753,
    "learning_rate": 6.034184490472203e-05,
    "timestep_mean": 602.46875,
    "epsilon_norm": 8.054217338562012,
    "prediction_norm": 6.883970260620117,
    "timestamp": "2025-07-14T12:34:57.177977"
  },
  {
    "epoch": 2316,
    "batch_idx": 32,
    "loss": 0.2864234447479248,
    "learning_rate": 6.034184490472203e-05,
    "timestep_mean": 493.625,
    "epsilon_norm": 8.672637939453125,
    "prediction_norm": 7.145856857299805,
    "timestamp": "2025-07-14T12:34:57.183273"
  },
  {
    "epoch": 2316,
    "batch_idx": 64,
    "loss": 0.2612839341163635,
    "learning_rate": 6.034184490472203e-05,
    "timestep_mean": 520.6875,
    "epsilon_norm": 7.708687782287598,
    "prediction_norm": 6.232870101928711,
    "timestamp": "2025-07-14T12:34:57.188495"
  },
  {
    "epoch": 2316,
    "batch_idx": 96,
    "loss": 0.18963474035263062,
    "learning_rate": 6.034184490472203e-05,
    "timestep_mean": 524.625,
    "epsilon_norm": 7.732512474060059,
    "prediction_norm": 6.688458442687988,
    "timestamp": "2025-07-14T12:34:57.193642"
  },
  {
    "epoch": 2316,
    "batch_idx": 128,
    "loss": 0.6606190204620361,
    "learning_rate": 6.034184490472203e-05,
    "timestep_mean": 377.5625,
    "epsilon_norm": 9.710110664367676,
    "prediction_norm": 6.685187339782715,
    "timestamp": "2025-07-14T12:34:57.198757"
  },
  {
    "epoch": 2316,
    "batch_idx": 160,
    "loss": 0.3240993022918701,
    "learning_rate": 6.034184490472203e-05,
    "timestep_mean": 401.34375,
    "epsilon_norm": 8.152352333068848,
    "prediction_norm": 6.638093948364258,
    "timestamp": "2025-07-14T12:34:57.203878"
  },
  {
    "epoch": 2316,
    "batch_idx": 192,
    "loss": 0.05626455694437027,
    "learning_rate": 6.034184490472203e-05,
    "timestep_mean": 693.75,
    "epsilon_norm": 3.3957645893096924,
    "prediction_norm": 3.165371894836426,
    "timestamp": "2025-07-14T12:34:57.209056"
  },
  {
    "epoch": 2317,
    "batch_idx": 0,
    "loss": 0.4264826774597168,
    "learning_rate": 6.071642376780232e-05,
    "timestep_mean": 532.125,
    "epsilon_norm": 8.317656517028809,
    "prediction_norm": 6.638001918792725,
    "timestamp": "2025-07-14T12:34:57.214011"
  },
  {
    "epoch": 2317,
    "batch_idx": 32,
    "loss": 0.18129149079322815,
    "learning_rate": 6.071642376780232e-05,
    "timestep_mean": 541.9375,
    "epsilon_norm": 8.062922477722168,
    "prediction_norm": 6.9894118309021,
    "timestamp": "2025-07-14T12:34:57.218891"
  },
  {
    "epoch": 2317,
    "batch_idx": 64,
    "loss": 0.25144660472869873,
    "learning_rate": 6.071642376780232e-05,
    "timestep_mean": 566.625,
    "epsilon_norm": 8.033342361450195,
    "prediction_norm": 6.907781600952148,
    "timestamp": "2025-07-14T12:34:57.223809"
  },
  {
    "epoch": 2317,
    "batch_idx": 96,
    "loss": 0.12711432576179504,
    "learning_rate": 6.071642376780232e-05,
    "timestep_mean": 519.46875,
    "epsilon_norm": 6.048383712768555,
    "prediction_norm": 5.295257091522217,
    "timestamp": "2025-07-14T12:34:57.228945"
  },
  {
    "epoch": 2317,
    "batch_idx": 128,
    "loss": 0.37702739238739014,
    "learning_rate": 6.071642376780232e-05,
    "timestep_mean": 429.5,
    "epsilon_norm": 7.794785499572754,
    "prediction_norm": 6.135797023773193,
    "timestamp": "2025-07-14T12:34:57.234082"
  },
  {
    "epoch": 2317,
    "batch_idx": 160,
    "loss": 0.18956615030765533,
    "learning_rate": 6.071642376780232e-05,
    "timestep_mean": 531.25,
    "epsilon_norm": 9.255977630615234,
    "prediction_norm": 8.102110862731934,
    "timestamp": "2025-07-14T12:34:57.239095"
  },
  {
    "epoch": 2317,
    "batch_idx": 192,
    "loss": 0.06679028272628784,
    "learning_rate": 6.071642376780232e-05,
    "timestep_mean": 610.25,
    "epsilon_norm": 3.7285187244415283,
    "prediction_norm": 3.302846670150757,
    "timestamp": "2025-07-14T12:34:57.244119"
  },
  {
    "epoch": 2318,
    "batch_idx": 0,
    "loss": 0.1550755798816681,
    "learning_rate": 6.10920865194393e-05,
    "timestep_mean": 547.53125,
    "epsilon_norm": 7.077977657318115,
    "prediction_norm": 6.273893356323242,
    "timestamp": "2025-07-14T12:34:57.249155"
  },
  {
    "epoch": 2318,
    "batch_idx": 32,
    "loss": 0.33266517519950867,
    "learning_rate": 6.10920865194393e-05,
    "timestep_mean": 483.25,
    "epsilon_norm": 8.35998249053955,
    "prediction_norm": 7.198187828063965,
    "timestamp": "2025-07-14T12:34:57.254443"
  },
  {
    "epoch": 2318,
    "batch_idx": 64,
    "loss": 0.3055300712585449,
    "learning_rate": 6.10920865194393e-05,
    "timestep_mean": 512.03125,
    "epsilon_norm": 8.673705101013184,
    "prediction_norm": 7.64107608795166,
    "timestamp": "2025-07-14T12:34:57.259533"
  },
  {
    "epoch": 2318,
    "batch_idx": 96,
    "loss": 0.16838401556015015,
    "learning_rate": 6.10920865194393e-05,
    "timestep_mean": 504.84375,
    "epsilon_norm": 6.970674991607666,
    "prediction_norm": 6.212276935577393,
    "timestamp": "2025-07-14T12:34:57.264645"
  },
  {
    "epoch": 2318,
    "batch_idx": 128,
    "loss": 0.21750834584236145,
    "learning_rate": 6.10920865194393e-05,
    "timestep_mean": 426.78125,
    "epsilon_norm": 7.13454008102417,
    "prediction_norm": 5.3422064781188965,
    "timestamp": "2025-07-14T12:34:57.269768"
  },
  {
    "epoch": 2318,
    "batch_idx": 160,
    "loss": 0.4649025499820709,
    "learning_rate": 6.10920865194393e-05,
    "timestep_mean": 402.71875,
    "epsilon_norm": 7.775268077850342,
    "prediction_norm": 5.37280797958374,
    "timestamp": "2025-07-14T12:34:57.274888"
  },
  {
    "epoch": 2318,
    "batch_idx": 192,
    "loss": 0.41173499822616577,
    "learning_rate": 6.10920865194393e-05,
    "timestep_mean": 470.375,
    "epsilon_norm": 3.732727289199829,
    "prediction_norm": 2.969921350479126,
    "timestamp": "2025-07-14T12:34:57.280177"
  },
  {
    "epoch": 2319,
    "batch_idx": 0,
    "loss": 0.19206704199314117,
    "learning_rate": 6.146883223272219e-05,
    "timestep_mean": 466.28125,
    "epsilon_norm": 7.467243671417236,
    "prediction_norm": 6.749215126037598,
    "timestamp": "2025-07-14T12:34:57.285283"
  },
  {
    "epoch": 2319,
    "batch_idx": 32,
    "loss": 0.14809435606002808,
    "learning_rate": 6.146883223272219e-05,
    "timestep_mean": 511.1875,
    "epsilon_norm": 7.039935111999512,
    "prediction_norm": 6.263057708740234,
    "timestamp": "2025-07-14T12:34:57.290242"
  },
  {
    "epoch": 2319,
    "batch_idx": 64,
    "loss": 0.07679963111877441,
    "learning_rate": 6.146883223272219e-05,
    "timestep_mean": 618.5,
    "epsilon_norm": 8.490795135498047,
    "prediction_norm": 7.924170970916748,
    "timestamp": "2025-07-14T12:34:57.295370"
  },
  {
    "epoch": 2319,
    "batch_idx": 96,
    "loss": 0.14656323194503784,
    "learning_rate": 6.146883223272219e-05,
    "timestep_mean": 477.46875,
    "epsilon_norm": 6.8927998542785645,
    "prediction_norm": 6.644947528839111,
    "timestamp": "2025-07-14T12:34:57.300502"
  },
  {
    "epoch": 2319,
    "batch_idx": 128,
    "loss": 0.4129752218723297,
    "learning_rate": 6.146883223272219e-05,
    "timestep_mean": 479.15625,
    "epsilon_norm": 7.441012859344482,
    "prediction_norm": 6.101332187652588,
    "timestamp": "2025-07-14T12:34:57.305610"
  },
  {
    "epoch": 2319,
    "batch_idx": 160,
    "loss": 0.24124634265899658,
    "learning_rate": 6.146883223272219e-05,
    "timestep_mean": 421.40625,
    "epsilon_norm": 7.677645206451416,
    "prediction_norm": 6.166684150695801,
    "timestamp": "2025-07-14T12:34:57.310769"
  },
  {
    "epoch": 2319,
    "batch_idx": 192,
    "loss": 0.04362237825989723,
    "learning_rate": 6.146883223272219e-05,
    "timestep_mean": 650.125,
    "epsilon_norm": 4.984726428985596,
    "prediction_norm": 4.615880012512207,
    "timestamp": "2025-07-14T12:34:57.315613"
  },
  {
    "epoch": 2320,
    "batch_idx": 0,
    "loss": 0.2829315662384033,
    "learning_rate": 6.184665997806831e-05,
    "timestep_mean": 520.65625,
    "epsilon_norm": 8.574651718139648,
    "prediction_norm": 7.794095516204834,
    "timestamp": "2025-07-14T12:34:57.320499"
  },
  {
    "epoch": 2320,
    "batch_idx": 32,
    "loss": 0.2552570402622223,
    "learning_rate": 6.184665997806831e-05,
    "timestep_mean": 501.90625,
    "epsilon_norm": 7.831101417541504,
    "prediction_norm": 6.337144374847412,
    "timestamp": "2025-07-14T12:34:57.325453"
  },
  {
    "epoch": 2320,
    "batch_idx": 64,
    "loss": 0.2414025217294693,
    "learning_rate": 6.184665997806831e-05,
    "timestep_mean": 482.3125,
    "epsilon_norm": 7.789974689483643,
    "prediction_norm": 6.6442952156066895,
    "timestamp": "2025-07-14T12:34:57.330279"
  },
  {
    "epoch": 2320,
    "batch_idx": 96,
    "loss": 0.2883634567260742,
    "learning_rate": 6.184665997806831e-05,
    "timestep_mean": 497.375,
    "epsilon_norm": 8.695488929748535,
    "prediction_norm": 7.522851943969727,
    "timestamp": "2025-07-14T12:34:57.335059"
  },
  {
    "epoch": 2320,
    "batch_idx": 128,
    "loss": 0.23161913454532623,
    "learning_rate": 6.184665997806831e-05,
    "timestep_mean": 436.8125,
    "epsilon_norm": 7.392833709716797,
    "prediction_norm": 5.991995334625244,
    "timestamp": "2025-07-14T12:34:57.339882"
  },
  {
    "epoch": 2320,
    "batch_idx": 160,
    "loss": 0.2648783326148987,
    "learning_rate": 6.184665997806831e-05,
    "timestep_mean": 509.625,
    "epsilon_norm": 7.201797962188721,
    "prediction_norm": 6.430239677429199,
    "timestamp": "2025-07-14T12:34:57.344707"
  },
  {
    "epoch": 2320,
    "batch_idx": 192,
    "loss": 0.2613098621368408,
    "learning_rate": 6.184665997806831e-05,
    "timestep_mean": 442.625,
    "epsilon_norm": 3.9331159591674805,
    "prediction_norm": 2.9031822681427,
    "timestamp": "2025-07-14T12:34:57.349471"
  },
  {
    "epoch": 2321,
    "batch_idx": 0,
    "loss": 0.2959994673728943,
    "learning_rate": 6.222556882322555e-05,
    "timestep_mean": 457.65625,
    "epsilon_norm": 7.424436092376709,
    "prediction_norm": 5.637040138244629,
    "timestamp": "2025-07-14T12:34:57.354263"
  },
  {
    "epoch": 2321,
    "batch_idx": 32,
    "loss": 0.23197296261787415,
    "learning_rate": 6.222556882322555e-05,
    "timestep_mean": 583.8125,
    "epsilon_norm": 8.245222091674805,
    "prediction_norm": 7.47544527053833,
    "timestamp": "2025-07-14T12:34:57.359163"
  },
  {
    "epoch": 2321,
    "batch_idx": 64,
    "loss": 0.23296231031417847,
    "learning_rate": 6.222556882322555e-05,
    "timestep_mean": 470.0,
    "epsilon_norm": 7.792771816253662,
    "prediction_norm": 6.940490245819092,
    "timestamp": "2025-07-14T12:34:57.363989"
  },
  {
    "epoch": 2321,
    "batch_idx": 96,
    "loss": 0.4323135316371918,
    "learning_rate": 6.222556882322555e-05,
    "timestep_mean": 448.40625,
    "epsilon_norm": 7.87877893447876,
    "prediction_norm": 6.326810836791992,
    "timestamp": "2025-07-14T12:34:57.368826"
  },
  {
    "epoch": 2321,
    "batch_idx": 128,
    "loss": 0.20298701524734497,
    "learning_rate": 6.222556882322555e-05,
    "timestep_mean": 472.75,
    "epsilon_norm": 8.786515235900879,
    "prediction_norm": 7.333720684051514,
    "timestamp": "2025-07-14T12:34:57.373634"
  },
  {
    "epoch": 2321,
    "batch_idx": 160,
    "loss": 0.1975918412208557,
    "learning_rate": 6.222556882322555e-05,
    "timestep_mean": 521.34375,
    "epsilon_norm": 6.681341171264648,
    "prediction_norm": 6.083439350128174,
    "timestamp": "2025-07-14T12:34:57.378504"
  },
  {
    "epoch": 2321,
    "batch_idx": 192,
    "loss": 0.18061204254627228,
    "learning_rate": 6.222556882322555e-05,
    "timestep_mean": 628.0,
    "epsilon_norm": 4.014612674713135,
    "prediction_norm": 3.8759286403656006,
    "timestamp": "2025-07-14T12:34:57.383311"
  },
  {
    "epoch": 2322,
    "batch_idx": 0,
    "loss": 0.42087090015411377,
    "learning_rate": 6.260555783327366e-05,
    "timestep_mean": 485.25,
    "epsilon_norm": 8.06099796295166,
    "prediction_norm": 6.43060302734375,
    "timestamp": "2025-07-14T12:34:57.388329"
  },
  {
    "epoch": 2322,
    "batch_idx": 32,
    "loss": 0.21123185753822327,
    "learning_rate": 6.260555783327366e-05,
    "timestep_mean": 502.75,
    "epsilon_norm": 7.038102626800537,
    "prediction_norm": 5.888309001922607,
    "timestamp": "2025-07-14T12:34:57.393185"
  },
  {
    "epoch": 2322,
    "batch_idx": 64,
    "loss": 0.2203320860862732,
    "learning_rate": 6.260555783327366e-05,
    "timestep_mean": 540.0625,
    "epsilon_norm": 8.46215534210205,
    "prediction_norm": 8.033326148986816,
    "timestamp": "2025-07-14T12:34:57.398053"
  },
  {
    "epoch": 2322,
    "batch_idx": 96,
    "loss": 0.4003908038139343,
    "learning_rate": 6.260555783327366e-05,
    "timestep_mean": 492.1875,
    "epsilon_norm": 9.306109428405762,
    "prediction_norm": 7.705846309661865,
    "timestamp": "2025-07-14T12:34:57.402905"
  },
  {
    "epoch": 2322,
    "batch_idx": 128,
    "loss": 0.2065642774105072,
    "learning_rate": 6.260555783327366e-05,
    "timestep_mean": 459.1875,
    "epsilon_norm": 9.134407997131348,
    "prediction_norm": 7.910583019256592,
    "timestamp": "2025-07-14T12:34:57.407698"
  },
  {
    "epoch": 2322,
    "batch_idx": 160,
    "loss": 0.2419578731060028,
    "learning_rate": 6.260555783327366e-05,
    "timestep_mean": 541.8125,
    "epsilon_norm": 8.097131729125977,
    "prediction_norm": 7.071118354797363,
    "timestamp": "2025-07-14T12:34:57.412550"
  },
  {
    "epoch": 2322,
    "batch_idx": 192,
    "loss": 0.14648377895355225,
    "learning_rate": 6.260555783327366e-05,
    "timestep_mean": 566.75,
    "epsilon_norm": 3.824521064758301,
    "prediction_norm": 3.5562338829040527,
    "timestamp": "2025-07-14T12:34:57.417646"
  },
  {
    "epoch": 2323,
    "batch_idx": 0,
    "loss": 0.2612101435661316,
    "learning_rate": 6.29866260706279e-05,
    "timestep_mean": 485.1875,
    "epsilon_norm": 8.743351936340332,
    "prediction_norm": 7.383854389190674,
    "timestamp": "2025-07-14T12:34:57.422616"
  },
  {
    "epoch": 2323,
    "batch_idx": 32,
    "loss": 0.20314094424247742,
    "learning_rate": 6.29866260706279e-05,
    "timestep_mean": 431.28125,
    "epsilon_norm": 8.587079048156738,
    "prediction_norm": 7.291380882263184,
    "timestamp": "2025-07-14T12:34:57.427556"
  },
  {
    "epoch": 2323,
    "batch_idx": 64,
    "loss": 0.11255975812673569,
    "learning_rate": 6.29866260706279e-05,
    "timestep_mean": 551.5625,
    "epsilon_norm": 7.469566345214844,
    "prediction_norm": 6.913637161254883,
    "timestamp": "2025-07-14T12:34:57.432364"
  },
  {
    "epoch": 2323,
    "batch_idx": 96,
    "loss": 0.15044370293617249,
    "learning_rate": 6.29866260706279e-05,
    "timestep_mean": 539.625,
    "epsilon_norm": 7.444936275482178,
    "prediction_norm": 6.78450870513916,
    "timestamp": "2025-07-14T12:34:57.437211"
  },
  {
    "epoch": 2323,
    "batch_idx": 128,
    "loss": 0.19133079051971436,
    "learning_rate": 6.29866260706279e-05,
    "timestep_mean": 506.65625,
    "epsilon_norm": 8.071383476257324,
    "prediction_norm": 7.376504898071289,
    "timestamp": "2025-07-14T12:34:57.442121"
  },
  {
    "epoch": 2323,
    "batch_idx": 160,
    "loss": 0.23643594980239868,
    "learning_rate": 6.29866260706279e-05,
    "timestep_mean": 446.96875,
    "epsilon_norm": 7.455453872680664,
    "prediction_norm": 6.591015338897705,
    "timestamp": "2025-07-14T12:34:57.446990"
  },
  {
    "epoch": 2323,
    "batch_idx": 192,
    "loss": 0.36577656865119934,
    "learning_rate": 6.29866260706279e-05,
    "timestep_mean": 472.5,
    "epsilon_norm": 5.2170515060424805,
    "prediction_norm": 4.166147708892822,
    "timestamp": "2025-07-14T12:34:57.451986"
  },
  {
    "epoch": 2324,
    "batch_idx": 0,
    "loss": 0.2367558479309082,
    "learning_rate": 6.336877259504016e-05,
    "timestep_mean": 641.625,
    "epsilon_norm": 7.931528568267822,
    "prediction_norm": 7.3088698387146,
    "timestamp": "2025-07-14T12:34:57.456839"
  },
  {
    "epoch": 2324,
    "batch_idx": 32,
    "loss": 0.15682318806648254,
    "learning_rate": 6.336877259504016e-05,
    "timestep_mean": 481.125,
    "epsilon_norm": 8.895655632019043,
    "prediction_norm": 8.302385330200195,
    "timestamp": "2025-07-14T12:34:57.461681"
  },
  {
    "epoch": 2324,
    "batch_idx": 64,
    "loss": 0.2691548466682434,
    "learning_rate": 6.336877259504016e-05,
    "timestep_mean": 561.0625,
    "epsilon_norm": 8.561511993408203,
    "prediction_norm": 7.816417217254639,
    "timestamp": "2025-07-14T12:34:57.466513"
  },
  {
    "epoch": 2324,
    "batch_idx": 96,
    "loss": 0.391865074634552,
    "learning_rate": 6.336877259504016e-05,
    "timestep_mean": 467.09375,
    "epsilon_norm": 8.267420768737793,
    "prediction_norm": 7.008575439453125,
    "timestamp": "2025-07-14T12:34:57.471479"
  },
  {
    "epoch": 2324,
    "batch_idx": 128,
    "loss": 0.4372074007987976,
    "learning_rate": 6.336877259504016e-05,
    "timestep_mean": 532.71875,
    "epsilon_norm": 7.7483978271484375,
    "prediction_norm": 6.821485996246338,
    "timestamp": "2025-07-14T12:34:57.476324"
  },
  {
    "epoch": 2324,
    "batch_idx": 160,
    "loss": 0.21502655744552612,
    "learning_rate": 6.336877259504016e-05,
    "timestep_mean": 556.9375,
    "epsilon_norm": 8.775203704833984,
    "prediction_norm": 8.048592567443848,
    "timestamp": "2025-07-14T12:34:57.481170"
  },
  {
    "epoch": 2324,
    "batch_idx": 192,
    "loss": 0.3573340177536011,
    "learning_rate": 6.336877259504016e-05,
    "timestep_mean": 478.0,
    "epsilon_norm": 6.0777082443237305,
    "prediction_norm": 4.882181644439697,
    "timestamp": "2025-07-14T12:34:57.485960"
  },
  {
    "epoch": 2325,
    "batch_idx": 0,
    "loss": 0.2537763714790344,
    "learning_rate": 6.375199646360154e-05,
    "timestep_mean": 522.84375,
    "epsilon_norm": 8.041436195373535,
    "prediction_norm": 6.651388168334961,
    "timestamp": "2025-07-14T12:34:57.490857"
  },
  {
    "epoch": 2325,
    "batch_idx": 32,
    "loss": 0.26602375507354736,
    "learning_rate": 6.375199646360154e-05,
    "timestep_mean": 448.15625,
    "epsilon_norm": 8.133333206176758,
    "prediction_norm": 6.46613883972168,
    "timestamp": "2025-07-14T12:34:57.495666"
  },
  {
    "epoch": 2325,
    "batch_idx": 64,
    "loss": 0.2158314287662506,
    "learning_rate": 6.375199646360154e-05,
    "timestep_mean": 536.09375,
    "epsilon_norm": 8.577704429626465,
    "prediction_norm": 7.557621955871582,
    "timestamp": "2025-07-14T12:34:57.500563"
  },
  {
    "epoch": 2325,
    "batch_idx": 96,
    "loss": 0.3568858504295349,
    "learning_rate": 6.375199646360154e-05,
    "timestep_mean": 496.3125,
    "epsilon_norm": 8.224413871765137,
    "prediction_norm": 6.843560218811035,
    "timestamp": "2025-07-14T12:34:57.505404"
  },
  {
    "epoch": 2325,
    "batch_idx": 128,
    "loss": 0.2933494448661804,
    "learning_rate": 6.375199646360154e-05,
    "timestep_mean": 507.5625,
    "epsilon_norm": 7.731497287750244,
    "prediction_norm": 6.661837577819824,
    "timestamp": "2025-07-14T12:34:57.510272"
  },
  {
    "epoch": 2325,
    "batch_idx": 160,
    "loss": 0.24868500232696533,
    "learning_rate": 6.375199646360154e-05,
    "timestep_mean": 523.84375,
    "epsilon_norm": 6.672670364379883,
    "prediction_norm": 5.703871250152588,
    "timestamp": "2025-07-14T12:34:57.515151"
  },
  {
    "epoch": 2325,
    "batch_idx": 192,
    "loss": 0.7836129665374756,
    "learning_rate": 6.375199646360154e-05,
    "timestep_mean": 285.875,
    "epsilon_norm": 4.415475368499756,
    "prediction_norm": 2.2884466648101807,
    "timestamp": "2025-07-14T12:34:57.520004"
  },
  {
    "epoch": 2326,
    "batch_idx": 0,
    "loss": 0.3127497136592865,
    "learning_rate": 6.413629673074574e-05,
    "timestep_mean": 480.03125,
    "epsilon_norm": 7.804495334625244,
    "prediction_norm": 6.778661251068115,
    "timestamp": "2025-07-14T12:34:57.524934"
  },
  {
    "epoch": 2326,
    "batch_idx": 32,
    "loss": 0.2801635265350342,
    "learning_rate": 6.413629673074574e-05,
    "timestep_mean": 550.0,
    "epsilon_norm": 7.118198394775391,
    "prediction_norm": 6.216061592102051,
    "timestamp": "2025-07-14T12:34:57.529870"
  },
  {
    "epoch": 2326,
    "batch_idx": 64,
    "loss": 0.1779804825782776,
    "learning_rate": 6.413629673074574e-05,
    "timestep_mean": 498.28125,
    "epsilon_norm": 8.487290382385254,
    "prediction_norm": 8.267691612243652,
    "timestamp": "2025-07-14T12:34:57.534812"
  },
  {
    "epoch": 2326,
    "batch_idx": 96,
    "loss": 0.28262728452682495,
    "learning_rate": 6.413629673074574e-05,
    "timestep_mean": 410.46875,
    "epsilon_norm": 7.1156182289123535,
    "prediction_norm": 5.732677936553955,
    "timestamp": "2025-07-14T12:34:57.539672"
  },
  {
    "epoch": 2326,
    "batch_idx": 128,
    "loss": 0.2351142019033432,
    "learning_rate": 6.413629673074574e-05,
    "timestep_mean": 550.5625,
    "epsilon_norm": 7.648615837097168,
    "prediction_norm": 6.755463600158691,
    "timestamp": "2025-07-14T12:34:57.544682"
  },
  {
    "epoch": 2326,
    "batch_idx": 160,
    "loss": 0.23855528235435486,
    "learning_rate": 6.413629673074574e-05,
    "timestep_mean": 593.4375,
    "epsilon_norm": 8.764935493469238,
    "prediction_norm": 7.714417457580566,
    "timestamp": "2025-07-14T12:34:57.549526"
  },
  {
    "epoch": 2326,
    "batch_idx": 192,
    "loss": 0.2522164583206177,
    "learning_rate": 6.413629673074574e-05,
    "timestep_mean": 422.875,
    "epsilon_norm": 3.7017533779144287,
    "prediction_norm": 2.5681960582733154,
    "timestamp": "2025-07-14T12:34:57.554469"
  },
  {
    "epoch": 2327,
    "batch_idx": 0,
    "loss": 0.26743069291114807,
    "learning_rate": 6.452167244824971e-05,
    "timestep_mean": 499.90625,
    "epsilon_norm": 7.242513656616211,
    "prediction_norm": 6.289880752563477,
    "timestamp": "2025-07-14T12:34:57.559436"
  },
  {
    "epoch": 2327,
    "batch_idx": 32,
    "loss": 0.2909875512123108,
    "learning_rate": 6.452167244824971e-05,
    "timestep_mean": 561.28125,
    "epsilon_norm": 7.355870246887207,
    "prediction_norm": 6.280580043792725,
    "timestamp": "2025-07-14T12:34:57.564335"
  },
  {
    "epoch": 2327,
    "batch_idx": 64,
    "loss": 0.2775050699710846,
    "learning_rate": 6.452167244824971e-05,
    "timestep_mean": 493.75,
    "epsilon_norm": 7.965914726257324,
    "prediction_norm": 6.683830261230469,
    "timestamp": "2025-07-14T12:34:57.569308"
  },
  {
    "epoch": 2327,
    "batch_idx": 96,
    "loss": 0.3200501501560211,
    "learning_rate": 6.452167244824971e-05,
    "timestep_mean": 481.4375,
    "epsilon_norm": 7.724353313446045,
    "prediction_norm": 5.596395015716553,
    "timestamp": "2025-07-14T12:34:57.574252"
  },
  {
    "epoch": 2327,
    "batch_idx": 128,
    "loss": 0.22277040779590607,
    "learning_rate": 6.452167244824971e-05,
    "timestep_mean": 487.78125,
    "epsilon_norm": 8.184353828430176,
    "prediction_norm": 6.784726619720459,
    "timestamp": "2025-07-14T12:34:57.579151"
  },
  {
    "epoch": 2327,
    "batch_idx": 160,
    "loss": 0.26472941040992737,
    "learning_rate": 6.452167244824971e-05,
    "timestep_mean": 585.53125,
    "epsilon_norm": 6.507142543792725,
    "prediction_norm": 5.322240352630615,
    "timestamp": "2025-07-14T12:34:57.584189"
  },
  {
    "epoch": 2327,
    "batch_idx": 192,
    "loss": 0.05500739440321922,
    "learning_rate": 6.452167244824971e-05,
    "timestep_mean": 703.625,
    "epsilon_norm": 4.637051105499268,
    "prediction_norm": 4.438155651092529,
    "timestamp": "2025-07-14T12:34:57.589091"
  },
  {
    "epoch": 2328,
    "batch_idx": 0,
    "loss": 0.21309427917003632,
    "learning_rate": 6.490812266523728e-05,
    "timestep_mean": 472.15625,
    "epsilon_norm": 7.707463264465332,
    "prediction_norm": 7.269193172454834,
    "timestamp": "2025-07-14T12:34:57.593970"
  },
  {
    "epoch": 2328,
    "batch_idx": 32,
    "loss": 0.28338855504989624,
    "learning_rate": 6.490812266523728e-05,
    "timestep_mean": 456.21875,
    "epsilon_norm": 7.574084281921387,
    "prediction_norm": 6.078338146209717,
    "timestamp": "2025-07-14T12:34:57.598828"
  },
  {
    "epoch": 2328,
    "batch_idx": 64,
    "loss": 0.19693265855312347,
    "learning_rate": 6.490812266523728e-05,
    "timestep_mean": 524.6875,
    "epsilon_norm": 6.732393264770508,
    "prediction_norm": 6.011006832122803,
    "timestamp": "2025-07-14T12:34:57.603725"
  },
  {
    "epoch": 2328,
    "batch_idx": 96,
    "loss": 0.22410161793231964,
    "learning_rate": 6.490812266523728e-05,
    "timestep_mean": 511.1875,
    "epsilon_norm": 7.5229620933532715,
    "prediction_norm": 6.1662397384643555,
    "timestamp": "2025-07-14T12:34:57.608566"
  },
  {
    "epoch": 2328,
    "batch_idx": 128,
    "loss": 0.2608516216278076,
    "learning_rate": 6.490812266523728e-05,
    "timestep_mean": 560.09375,
    "epsilon_norm": 8.632309913635254,
    "prediction_norm": 7.544564247131348,
    "timestamp": "2025-07-14T12:34:57.613552"
  },
  {
    "epoch": 2328,
    "batch_idx": 160,
    "loss": 0.2273615598678589,
    "learning_rate": 6.490812266523728e-05,
    "timestep_mean": 533.15625,
    "epsilon_norm": 8.246702194213867,
    "prediction_norm": 7.016002655029297,
    "timestamp": "2025-07-14T12:34:57.618430"
  },
  {
    "epoch": 2328,
    "batch_idx": 192,
    "loss": 0.1692979633808136,
    "learning_rate": 6.490812266523728e-05,
    "timestep_mean": 556.25,
    "epsilon_norm": 3.5277764797210693,
    "prediction_norm": 3.222500801086426,
    "timestamp": "2025-07-14T12:34:57.623343"
  },
  {
    "epoch": 2329,
    "batch_idx": 0,
    "loss": 0.30155372619628906,
    "learning_rate": 6.5295646428181e-05,
    "timestep_mean": 536.625,
    "epsilon_norm": 8.52889633178711,
    "prediction_norm": 6.675055027008057,
    "timestamp": "2025-07-14T12:34:57.628207"
  },
  {
    "epoch": 2329,
    "batch_idx": 32,
    "loss": 0.09114916622638702,
    "learning_rate": 6.5295646428181e-05,
    "timestep_mean": 532.375,
    "epsilon_norm": 6.543186664581299,
    "prediction_norm": 6.394688606262207,
    "timestamp": "2025-07-14T12:34:57.633155"
  },
  {
    "epoch": 2329,
    "batch_idx": 64,
    "loss": 0.3836280107498169,
    "learning_rate": 6.5295646428181e-05,
    "timestep_mean": 523.3125,
    "epsilon_norm": 8.072906494140625,
    "prediction_norm": 6.498769760131836,
    "timestamp": "2025-07-14T12:34:57.638133"
  },
  {
    "epoch": 2329,
    "batch_idx": 96,
    "loss": 0.44650328159332275,
    "learning_rate": 6.5295646428181e-05,
    "timestep_mean": 456.96875,
    "epsilon_norm": 8.186004638671875,
    "prediction_norm": 5.879494667053223,
    "timestamp": "2025-07-14T12:34:57.642998"
  },
  {
    "epoch": 2329,
    "batch_idx": 128,
    "loss": 0.30867478251457214,
    "learning_rate": 6.5295646428181e-05,
    "timestep_mean": 416.5,
    "epsilon_norm": 7.692599296569824,
    "prediction_norm": 6.389824867248535,
    "timestamp": "2025-07-14T12:34:57.647810"
  },
  {
    "epoch": 2329,
    "batch_idx": 160,
    "loss": 0.1751001626253128,
    "learning_rate": 6.5295646428181e-05,
    "timestep_mean": 507.90625,
    "epsilon_norm": 8.104238510131836,
    "prediction_norm": 7.816421985626221,
    "timestamp": "2025-07-14T12:34:57.652719"
  },
  {
    "epoch": 2329,
    "batch_idx": 192,
    "loss": 0.20971976220607758,
    "learning_rate": 6.5295646428181e-05,
    "timestep_mean": 599.75,
    "epsilon_norm": 3.72830867767334,
    "prediction_norm": 3.1970765590667725,
    "timestamp": "2025-07-14T12:34:57.657557"
  },
  {
    "epoch": 2330,
    "batch_idx": 0,
    "loss": 0.3436214327812195,
    "learning_rate": 6.568424278090458e-05,
    "timestep_mean": 457.4375,
    "epsilon_norm": 8.601008415222168,
    "prediction_norm": 7.050410270690918,
    "timestamp": "2025-07-14T12:34:57.662494"
  },
  {
    "epoch": 2330,
    "batch_idx": 32,
    "loss": 0.3934636116027832,
    "learning_rate": 6.568424278090458e-05,
    "timestep_mean": 504.5,
    "epsilon_norm": 8.815237998962402,
    "prediction_norm": 7.238414287567139,
    "timestamp": "2025-07-14T12:34:57.667429"
  },
  {
    "epoch": 2330,
    "batch_idx": 64,
    "loss": 0.20984379947185516,
    "learning_rate": 6.568424278090458e-05,
    "timestep_mean": 492.4375,
    "epsilon_norm": 7.760919094085693,
    "prediction_norm": 7.527183532714844,
    "timestamp": "2025-07-14T12:34:57.672264"
  },
  {
    "epoch": 2330,
    "batch_idx": 96,
    "loss": 0.2952757775783539,
    "learning_rate": 6.568424278090458e-05,
    "timestep_mean": 456.125,
    "epsilon_norm": 7.2164130210876465,
    "prediction_norm": 5.80937385559082,
    "timestamp": "2025-07-14T12:34:57.677095"
  },
  {
    "epoch": 2330,
    "batch_idx": 128,
    "loss": 0.6243233680725098,
    "learning_rate": 6.568424278090458e-05,
    "timestep_mean": 423.5625,
    "epsilon_norm": 8.244739532470703,
    "prediction_norm": 5.904461860656738,
    "timestamp": "2025-07-14T12:34:57.681987"
  },
  {
    "epoch": 2330,
    "batch_idx": 160,
    "loss": 0.33884042501449585,
    "learning_rate": 6.568424278090458e-05,
    "timestep_mean": 533.71875,
    "epsilon_norm": 8.901700973510742,
    "prediction_norm": 7.566984176635742,
    "timestamp": "2025-07-14T12:34:57.686791"
  },
  {
    "epoch": 2330,
    "batch_idx": 192,
    "loss": 0.2574407458305359,
    "learning_rate": 6.568424278090458e-05,
    "timestep_mean": 421.875,
    "epsilon_norm": 3.721430540084839,
    "prediction_norm": 3.1818575859069824,
    "timestamp": "2025-07-14T12:34:57.691627"
  },
  {
    "epoch": 2331,
    "batch_idx": 0,
    "loss": 0.4846542477607727,
    "learning_rate": 6.607391076458493e-05,
    "timestep_mean": 405.46875,
    "epsilon_norm": 9.107399940490723,
    "prediction_norm": 7.380254745483398,
    "timestamp": "2025-07-14T12:34:57.696557"
  },
  {
    "epoch": 2331,
    "batch_idx": 32,
    "loss": 0.1893896460533142,
    "learning_rate": 6.607391076458493e-05,
    "timestep_mean": 495.0,
    "epsilon_norm": 7.8314619064331055,
    "prediction_norm": 6.592811107635498,
    "timestamp": "2025-07-14T12:34:57.701426"
  },
  {
    "epoch": 2331,
    "batch_idx": 64,
    "loss": 0.20494617521762848,
    "learning_rate": 6.607391076458493e-05,
    "timestep_mean": 416.90625,
    "epsilon_norm": 6.840648651123047,
    "prediction_norm": 6.255077362060547,
    "timestamp": "2025-07-14T12:34:57.706267"
  },
  {
    "epoch": 2331,
    "batch_idx": 96,
    "loss": 0.2286999523639679,
    "learning_rate": 6.607391076458493e-05,
    "timestep_mean": 459.40625,
    "epsilon_norm": 6.900765895843506,
    "prediction_norm": 6.091866493225098,
    "timestamp": "2025-07-14T12:34:57.711089"
  },
  {
    "epoch": 2331,
    "batch_idx": 128,
    "loss": 0.1633562594652176,
    "learning_rate": 6.607391076458493e-05,
    "timestep_mean": 488.71875,
    "epsilon_norm": 7.8383588790893555,
    "prediction_norm": 7.461233139038086,
    "timestamp": "2025-07-14T12:34:57.715929"
  },
  {
    "epoch": 2331,
    "batch_idx": 160,
    "loss": 0.5089648962020874,
    "learning_rate": 6.607391076458493e-05,
    "timestep_mean": 416.53125,
    "epsilon_norm": 6.898227214813232,
    "prediction_norm": 5.707096576690674,
    "timestamp": "2025-07-14T12:34:57.720751"
  },
  {
    "epoch": 2331,
    "batch_idx": 192,
    "loss": 0.5449007749557495,
    "learning_rate": 6.607391076458493e-05,
    "timestep_mean": 535.375,
    "epsilon_norm": 4.504918098449707,
    "prediction_norm": 2.7308006286621094,
    "timestamp": "2025-07-14T12:34:57.725611"
  },
  {
    "epoch": 2332,
    "batch_idx": 0,
    "loss": 0.2015867680311203,
    "learning_rate": 6.64646494177551e-05,
    "timestep_mean": 427.90625,
    "epsilon_norm": 6.807071685791016,
    "prediction_norm": 5.697406768798828,
    "timestamp": "2025-07-14T12:34:57.730435"
  },
  {
    "epoch": 2332,
    "batch_idx": 32,
    "loss": 0.20300671458244324,
    "learning_rate": 6.64646494177551e-05,
    "timestep_mean": 459.9375,
    "epsilon_norm": 7.684482097625732,
    "prediction_norm": 6.825615406036377,
    "timestamp": "2025-07-14T12:34:57.735322"
  },
  {
    "epoch": 2332,
    "batch_idx": 64,
    "loss": 0.33822381496429443,
    "learning_rate": 6.64646494177551e-05,
    "timestep_mean": 537.96875,
    "epsilon_norm": 8.6515531539917,
    "prediction_norm": 7.052870273590088,
    "timestamp": "2025-07-14T12:34:57.740219"
  },
  {
    "epoch": 2332,
    "batch_idx": 96,
    "loss": 0.23660919070243835,
    "learning_rate": 6.64646494177551e-05,
    "timestep_mean": 479.90625,
    "epsilon_norm": 7.714196681976318,
    "prediction_norm": 7.236475467681885,
    "timestamp": "2025-07-14T12:34:57.745064"
  },
  {
    "epoch": 2332,
    "batch_idx": 128,
    "loss": 0.213359534740448,
    "learning_rate": 6.64646494177551e-05,
    "timestep_mean": 499.25,
    "epsilon_norm": 7.099191188812256,
    "prediction_norm": 6.226813793182373,
    "timestamp": "2025-07-14T12:34:57.749982"
  },
  {
    "epoch": 2332,
    "batch_idx": 160,
    "loss": 0.4258815050125122,
    "learning_rate": 6.64646494177551e-05,
    "timestep_mean": 452.375,
    "epsilon_norm": 8.517329216003418,
    "prediction_norm": 6.249880790710449,
    "timestamp": "2025-07-14T12:34:57.754872"
  },
  {
    "epoch": 2332,
    "batch_idx": 192,
    "loss": 0.2769383192062378,
    "learning_rate": 6.64646494177551e-05,
    "timestep_mean": 570.875,
    "epsilon_norm": 3.175957441329956,
    "prediction_norm": 2.4138131141662598,
    "timestamp": "2025-07-14T12:34:57.759650"
  },
  {
    "epoch": 2333,
    "batch_idx": 0,
    "loss": 0.2697252035140991,
    "learning_rate": 6.68564577763064e-05,
    "timestep_mean": 469.78125,
    "epsilon_norm": 8.301620483398438,
    "prediction_norm": 7.064389228820801,
    "timestamp": "2025-07-14T12:34:57.764499"
  },
  {
    "epoch": 2333,
    "batch_idx": 32,
    "loss": 0.27036505937576294,
    "learning_rate": 6.68564577763064e-05,
    "timestep_mean": 494.09375,
    "epsilon_norm": 8.101887702941895,
    "prediction_norm": 7.382775783538818,
    "timestamp": "2025-07-14T12:34:57.769310"
  },
  {
    "epoch": 2333,
    "batch_idx": 64,
    "loss": 0.38109514117240906,
    "learning_rate": 6.68564577763064e-05,
    "timestep_mean": 449.6875,
    "epsilon_norm": 7.150235176086426,
    "prediction_norm": 5.031374931335449,
    "timestamp": "2025-07-14T12:34:57.774178"
  },
  {
    "epoch": 2333,
    "batch_idx": 96,
    "loss": 0.22294773161411285,
    "learning_rate": 6.68564577763064e-05,
    "timestep_mean": 555.15625,
    "epsilon_norm": 6.705972194671631,
    "prediction_norm": 5.430328845977783,
    "timestamp": "2025-07-14T12:34:57.778970"
  },
  {
    "epoch": 2333,
    "batch_idx": 128,
    "loss": 0.3166052997112274,
    "learning_rate": 6.68564577763064e-05,
    "timestep_mean": 553.5,
    "epsilon_norm": 8.450716972351074,
    "prediction_norm": 7.081153869628906,
    "timestamp": "2025-07-14T12:34:57.784030"
  },
  {
    "epoch": 2333,
    "batch_idx": 160,
    "loss": 0.2911432385444641,
    "learning_rate": 6.68564577763064e-05,
    "timestep_mean": 457.25,
    "epsilon_norm": 6.9895172119140625,
    "prediction_norm": 6.082958698272705,
    "timestamp": "2025-07-14T12:34:57.789192"
  },
  {
    "epoch": 2333,
    "batch_idx": 192,
    "loss": 0.11743539571762085,
    "learning_rate": 6.68564577763064e-05,
    "timestep_mean": 620.875,
    "epsilon_norm": 3.671741485595703,
    "prediction_norm": 2.8784427642822266,
    "timestamp": "2025-07-14T12:34:57.794255"
  },
  {
    "epoch": 2334,
    "batch_idx": 0,
    "loss": 0.17244914174079895,
    "learning_rate": 6.724933487349061e-05,
    "timestep_mean": 626.875,
    "epsilon_norm": 8.066089630126953,
    "prediction_norm": 7.374524116516113,
    "timestamp": "2025-07-14T12:34:57.799378"
  },
  {
    "epoch": 2334,
    "batch_idx": 32,
    "loss": 0.32854190468788147,
    "learning_rate": 6.724933487349061e-05,
    "timestep_mean": 458.1875,
    "epsilon_norm": 8.925772666931152,
    "prediction_norm": 7.273942470550537,
    "timestamp": "2025-07-14T12:34:57.804647"
  },
  {
    "epoch": 2334,
    "batch_idx": 64,
    "loss": 0.309942364692688,
    "learning_rate": 6.724933487349061e-05,
    "timestep_mean": 580.03125,
    "epsilon_norm": 7.413466453552246,
    "prediction_norm": 5.793613433837891,
    "timestamp": "2025-07-14T12:34:57.809630"
  },
  {
    "epoch": 2334,
    "batch_idx": 96,
    "loss": 0.30352699756622314,
    "learning_rate": 6.724933487349061e-05,
    "timestep_mean": 545.4375,
    "epsilon_norm": 8.562617301940918,
    "prediction_norm": 6.857859134674072,
    "timestamp": "2025-07-14T12:34:57.814502"
  },
  {
    "epoch": 2334,
    "batch_idx": 128,
    "loss": 0.5996087193489075,
    "learning_rate": 6.724933487349061e-05,
    "timestep_mean": 457.53125,
    "epsilon_norm": 11.704970359802246,
    "prediction_norm": 9.718184471130371,
    "timestamp": "2025-07-14T12:34:57.819286"
  },
  {
    "epoch": 2334,
    "batch_idx": 160,
    "loss": 0.2535025477409363,
    "learning_rate": 6.724933487349061e-05,
    "timestep_mean": 569.84375,
    "epsilon_norm": 8.564227104187012,
    "prediction_norm": 7.877154350280762,
    "timestamp": "2025-07-14T12:34:57.824092"
  },
  {
    "epoch": 2334,
    "batch_idx": 192,
    "loss": 0.1620946228504181,
    "learning_rate": 6.724933487349061e-05,
    "timestep_mean": 546.125,
    "epsilon_norm": 4.353763580322266,
    "prediction_norm": 3.6225664615631104,
    "timestamp": "2025-07-14T12:34:57.828907"
  },
  {
    "epoch": 2335,
    "batch_idx": 0,
    "loss": 0.5076959729194641,
    "learning_rate": 6.764327973992262e-05,
    "timestep_mean": 435.375,
    "epsilon_norm": 8.967374801635742,
    "prediction_norm": 6.812005519866943,
    "timestamp": "2025-07-14T12:34:57.833865"
  },
  {
    "epoch": 2335,
    "batch_idx": 32,
    "loss": 0.1658667027950287,
    "learning_rate": 6.764327973992262e-05,
    "timestep_mean": 504.0625,
    "epsilon_norm": 6.758995056152344,
    "prediction_norm": 6.348489761352539,
    "timestamp": "2025-07-14T12:34:57.838800"
  },
  {
    "epoch": 2335,
    "batch_idx": 64,
    "loss": 0.20156408846378326,
    "learning_rate": 6.764327973992262e-05,
    "timestep_mean": 530.8125,
    "epsilon_norm": 7.607810974121094,
    "prediction_norm": 6.782864093780518,
    "timestamp": "2025-07-14T12:34:57.843817"
  },
  {
    "epoch": 2335,
    "batch_idx": 96,
    "loss": 0.13640160858631134,
    "learning_rate": 6.764327973992262e-05,
    "timestep_mean": 510.21875,
    "epsilon_norm": 7.076061725616455,
    "prediction_norm": 6.440375804901123,
    "timestamp": "2025-07-14T12:34:57.848843"
  },
  {
    "epoch": 2335,
    "batch_idx": 128,
    "loss": 0.2651442587375641,
    "learning_rate": 6.764327973992262e-05,
    "timestep_mean": 494.625,
    "epsilon_norm": 7.561825275421143,
    "prediction_norm": 6.621974468231201,
    "timestamp": "2025-07-14T12:34:57.854426"
  },
  {
    "epoch": 2335,
    "batch_idx": 160,
    "loss": 0.19708743691444397,
    "learning_rate": 6.764327973992262e-05,
    "timestep_mean": 501.78125,
    "epsilon_norm": 8.056509017944336,
    "prediction_norm": 7.095658302307129,
    "timestamp": "2025-07-14T12:34:57.859555"
  },
  {
    "epoch": 2335,
    "batch_idx": 192,
    "loss": 0.5439905524253845,
    "learning_rate": 6.764327973992262e-05,
    "timestep_mean": 594.75,
    "epsilon_norm": 4.725417137145996,
    "prediction_norm": 3.790689468383789,
    "timestamp": "2025-07-14T12:34:57.864599"
  },
  {
    "epoch": 2336,
    "batch_idx": 0,
    "loss": 0.26421624422073364,
    "learning_rate": 6.803829140358247e-05,
    "timestep_mean": 471.25,
    "epsilon_norm": 7.838181495666504,
    "prediction_norm": 6.167932987213135,
    "timestamp": "2025-07-14T12:34:57.869769"
  },
  {
    "epoch": 2336,
    "batch_idx": 32,
    "loss": 0.24416020512580872,
    "learning_rate": 6.803829140358247e-05,
    "timestep_mean": 490.09375,
    "epsilon_norm": 8.525781631469727,
    "prediction_norm": 7.3016276359558105,
    "timestamp": "2025-07-14T12:34:57.874914"
  },
  {
    "epoch": 2336,
    "batch_idx": 64,
    "loss": 0.31133562326431274,
    "learning_rate": 6.803829140358247e-05,
    "timestep_mean": 507.96875,
    "epsilon_norm": 8.331616401672363,
    "prediction_norm": 6.4611382484436035,
    "timestamp": "2025-07-14T12:34:57.880105"
  },
  {
    "epoch": 2336,
    "batch_idx": 96,
    "loss": 0.4084312915802002,
    "learning_rate": 6.803829140358247e-05,
    "timestep_mean": 472.0625,
    "epsilon_norm": 7.940348148345947,
    "prediction_norm": 5.796875953674316,
    "timestamp": "2025-07-14T12:34:57.885231"
  },
  {
    "epoch": 2336,
    "batch_idx": 128,
    "loss": 0.23035992681980133,
    "learning_rate": 6.803829140358247e-05,
    "timestep_mean": 585.59375,
    "epsilon_norm": 7.56472110748291,
    "prediction_norm": 7.059087753295898,
    "timestamp": "2025-07-14T12:34:57.890252"
  },
  {
    "epoch": 2336,
    "batch_idx": 160,
    "loss": 0.16287559270858765,
    "learning_rate": 6.803829140358247e-05,
    "timestep_mean": 554.96875,
    "epsilon_norm": 6.644585132598877,
    "prediction_norm": 5.73565673828125,
    "timestamp": "2025-07-14T12:34:57.895441"
  },
  {
    "epoch": 2336,
    "batch_idx": 192,
    "loss": 0.2622920274734497,
    "learning_rate": 6.803829140358247e-05,
    "timestep_mean": 545.0,
    "epsilon_norm": 3.1929209232330322,
    "prediction_norm": 2.3347389698028564,
    "timestamp": "2025-07-14T12:34:57.900520"
  },
  {
    "epoch": 2337,
    "batch_idx": 0,
    "loss": 0.3308570981025696,
    "learning_rate": 6.843436888981824e-05,
    "timestep_mean": 414.46875,
    "epsilon_norm": 8.035297393798828,
    "prediction_norm": 6.4529852867126465,
    "timestamp": "2025-07-14T12:34:57.905678"
  },
  {
    "epoch": 2337,
    "batch_idx": 32,
    "loss": 0.3072686791419983,
    "learning_rate": 6.843436888981824e-05,
    "timestep_mean": 511.6875,
    "epsilon_norm": 8.085023880004883,
    "prediction_norm": 6.776006698608398,
    "timestamp": "2025-07-14T12:34:57.910782"
  },
  {
    "epoch": 2337,
    "batch_idx": 64,
    "loss": 0.2634466290473938,
    "learning_rate": 6.843436888981824e-05,
    "timestep_mean": 525.90625,
    "epsilon_norm": 8.207374572753906,
    "prediction_norm": 7.014005661010742,
    "timestamp": "2025-07-14T12:34:57.915878"
  },
  {
    "epoch": 2337,
    "batch_idx": 96,
    "loss": 0.16975542902946472,
    "learning_rate": 6.843436888981824e-05,
    "timestep_mean": 598.78125,
    "epsilon_norm": 8.159932136535645,
    "prediction_norm": 7.272671222686768,
    "timestamp": "2025-07-14T12:34:57.921121"
  },
  {
    "epoch": 2337,
    "batch_idx": 128,
    "loss": 0.1299845576286316,
    "learning_rate": 6.843436888981824e-05,
    "timestep_mean": 563.6875,
    "epsilon_norm": 8.355256080627441,
    "prediction_norm": 7.584962368011475,
    "timestamp": "2025-07-14T12:34:57.926249"
  },
  {
    "epoch": 2337,
    "batch_idx": 160,
    "loss": 0.22998939454555511,
    "learning_rate": 6.843436888981824e-05,
    "timestep_mean": 546.09375,
    "epsilon_norm": 8.719764709472656,
    "prediction_norm": 7.929693222045898,
    "timestamp": "2025-07-14T12:34:57.931362"
  },
  {
    "epoch": 2337,
    "batch_idx": 192,
    "loss": 0.17631790041923523,
    "learning_rate": 6.843436888981824e-05,
    "timestep_mean": 361.875,
    "epsilon_norm": 3.8090779781341553,
    "prediction_norm": 3.6154348850250244,
    "timestamp": "2025-07-14T12:34:57.936592"
  },
  {
    "epoch": 2338,
    "batch_idx": 0,
    "loss": 0.2152813822031021,
    "learning_rate": 6.883151122134823e-05,
    "timestep_mean": 546.375,
    "epsilon_norm": 9.032333374023438,
    "prediction_norm": 8.044633865356445,
    "timestamp": "2025-07-14T12:34:57.941926"
  },
  {
    "epoch": 2338,
    "batch_idx": 32,
    "loss": 0.21353980898857117,
    "learning_rate": 6.883151122134823e-05,
    "timestep_mean": 522.75,
    "epsilon_norm": 8.093015670776367,
    "prediction_norm": 6.9729719161987305,
    "timestamp": "2025-07-14T12:34:57.947156"
  },
  {
    "epoch": 2338,
    "batch_idx": 64,
    "loss": 0.21602056920528412,
    "learning_rate": 6.883151122134823e-05,
    "timestep_mean": 536.59375,
    "epsilon_norm": 7.612515926361084,
    "prediction_norm": 6.648777008056641,
    "timestamp": "2025-07-14T12:34:57.952310"
  },
  {
    "epoch": 2338,
    "batch_idx": 96,
    "loss": 0.3002612888813019,
    "learning_rate": 6.883151122134823e-05,
    "timestep_mean": 529.4375,
    "epsilon_norm": 9.495866775512695,
    "prediction_norm": 7.918484687805176,
    "timestamp": "2025-07-14T12:34:57.957423"
  },
  {
    "epoch": 2338,
    "batch_idx": 128,
    "loss": 0.15607304871082306,
    "learning_rate": 6.883151122134823e-05,
    "timestep_mean": 480.75,
    "epsilon_norm": 7.888047695159912,
    "prediction_norm": 6.942271709442139,
    "timestamp": "2025-07-14T12:34:57.962666"
  },
  {
    "epoch": 2338,
    "batch_idx": 160,
    "loss": 0.14016175270080566,
    "learning_rate": 6.883151122134823e-05,
    "timestep_mean": 511.3125,
    "epsilon_norm": 8.778751373291016,
    "prediction_norm": 8.418913841247559,
    "timestamp": "2025-07-14T12:34:57.967877"
  },
  {
    "epoch": 2338,
    "batch_idx": 192,
    "loss": 0.8146020174026489,
    "learning_rate": 6.883151122134823e-05,
    "timestep_mean": 502.5,
    "epsilon_norm": 4.135654926300049,
    "prediction_norm": 2.802375078201294,
    "timestamp": "2025-07-14T12:34:57.972968"
  },
  {
    "epoch": 2339,
    "batch_idx": 0,
    "loss": 0.18922443687915802,
    "learning_rate": 6.922971741826284e-05,
    "timestep_mean": 481.96875,
    "epsilon_norm": 8.549164772033691,
    "prediction_norm": 7.666717052459717,
    "timestamp": "2025-07-14T12:34:57.978241"
  },
  {
    "epoch": 2339,
    "batch_idx": 32,
    "loss": 0.273676335811615,
    "learning_rate": 6.922971741826284e-05,
    "timestep_mean": 448.125,
    "epsilon_norm": 8.762920379638672,
    "prediction_norm": 7.182923793792725,
    "timestamp": "2025-07-14T12:34:57.983460"
  },
  {
    "epoch": 2339,
    "batch_idx": 64,
    "loss": 0.14714765548706055,
    "learning_rate": 6.922971741826284e-05,
    "timestep_mean": 541.46875,
    "epsilon_norm": 6.814379692077637,
    "prediction_norm": 6.411179542541504,
    "timestamp": "2025-07-14T12:34:57.988617"
  },
  {
    "epoch": 2339,
    "batch_idx": 96,
    "loss": 0.15480700135231018,
    "learning_rate": 6.922971741826284e-05,
    "timestep_mean": 468.5625,
    "epsilon_norm": 7.489253044128418,
    "prediction_norm": 6.877716541290283,
    "timestamp": "2025-07-14T12:34:57.993507"
  },
  {
    "epoch": 2339,
    "batch_idx": 128,
    "loss": 0.1313694715499878,
    "learning_rate": 6.922971741826284e-05,
    "timestep_mean": 544.84375,
    "epsilon_norm": 7.887709617614746,
    "prediction_norm": 6.984988212585449,
    "timestamp": "2025-07-14T12:34:57.998600"
  },
  {
    "epoch": 2339,
    "batch_idx": 160,
    "loss": 0.31438302993774414,
    "learning_rate": 6.922971741826284e-05,
    "timestep_mean": 475.15625,
    "epsilon_norm": 8.3297758102417,
    "prediction_norm": 7.151134967803955,
    "timestamp": "2025-07-14T12:34:58.003745"
  },
  {
    "epoch": 2339,
    "batch_idx": 192,
    "loss": 0.20410913228988647,
    "learning_rate": 6.922971741826284e-05,
    "timestep_mean": 408.0,
    "epsilon_norm": 2.873120069503784,
    "prediction_norm": 2.7914388179779053,
    "timestamp": "2025-07-14T12:34:58.008907"
  },
  {
    "epoch": 2340,
    "batch_idx": 0,
    "loss": 0.35878854990005493,
    "learning_rate": 6.962898649802836e-05,
    "timestep_mean": 469.625,
    "epsilon_norm": 7.299165725708008,
    "prediction_norm": 6.069883823394775,
    "timestamp": "2025-07-14T12:34:58.013997"
  },
  {
    "epoch": 2340,
    "batch_idx": 32,
    "loss": 0.18877360224723816,
    "learning_rate": 6.962898649802836e-05,
    "timestep_mean": 491.75,
    "epsilon_norm": 8.096796989440918,
    "prediction_norm": 7.403500080108643,
    "timestamp": "2025-07-14T12:34:58.019115"
  },
  {
    "epoch": 2340,
    "batch_idx": 64,
    "loss": 0.10330824553966522,
    "learning_rate": 6.962898649802836e-05,
    "timestep_mean": 520.9375,
    "epsilon_norm": 8.154767036437988,
    "prediction_norm": 7.847299575805664,
    "timestamp": "2025-07-14T12:34:58.024137"
  },
  {
    "epoch": 2340,
    "batch_idx": 96,
    "loss": 0.27144068479537964,
    "learning_rate": 6.962898649802836e-05,
    "timestep_mean": 451.53125,
    "epsilon_norm": 8.858363151550293,
    "prediction_norm": 7.008274555206299,
    "timestamp": "2025-07-14T12:34:58.029335"
  },
  {
    "epoch": 2340,
    "batch_idx": 128,
    "loss": 0.22138121724128723,
    "learning_rate": 6.962898649802836e-05,
    "timestep_mean": 512.875,
    "epsilon_norm": 5.893664836883545,
    "prediction_norm": 5.2401957511901855,
    "timestamp": "2025-07-14T12:34:58.034495"
  },
  {
    "epoch": 2340,
    "batch_idx": 160,
    "loss": 0.15038001537322998,
    "learning_rate": 6.962898649802836e-05,
    "timestep_mean": 550.53125,
    "epsilon_norm": 7.590279579162598,
    "prediction_norm": 7.092535495758057,
    "timestamp": "2025-07-14T12:34:58.039563"
  },
  {
    "epoch": 2340,
    "batch_idx": 192,
    "loss": 0.31072014570236206,
    "learning_rate": 6.962898649802836e-05,
    "timestep_mean": 385.125,
    "epsilon_norm": 3.833937406539917,
    "prediction_norm": 3.072547435760498,
    "timestamp": "2025-07-14T12:34:58.044689"
  },
  {
    "epoch": 2341,
    "batch_idx": 0,
    "loss": 0.3630355894565582,
    "learning_rate": 7.002931747548765e-05,
    "timestep_mean": 464.875,
    "epsilon_norm": 8.611199378967285,
    "prediction_norm": 6.830338954925537,
    "timestamp": "2025-07-14T12:34:58.049854"
  },
  {
    "epoch": 2341,
    "batch_idx": 32,
    "loss": 0.2267688810825348,
    "learning_rate": 7.002931747548765e-05,
    "timestep_mean": 436.25,
    "epsilon_norm": 6.943152904510498,
    "prediction_norm": 6.455704212188721,
    "timestamp": "2025-07-14T12:34:58.055133"
  },
  {
    "epoch": 2341,
    "batch_idx": 64,
    "loss": 0.252002090215683,
    "learning_rate": 7.002931747548765e-05,
    "timestep_mean": 543.5,
    "epsilon_norm": 8.647828102111816,
    "prediction_norm": 6.9901580810546875,
    "timestamp": "2025-07-14T12:34:58.060192"
  },
  {
    "epoch": 2341,
    "batch_idx": 96,
    "loss": 0.182511106133461,
    "learning_rate": 7.002931747548765e-05,
    "timestep_mean": 574.28125,
    "epsilon_norm": 8.313523292541504,
    "prediction_norm": 7.385649681091309,
    "timestamp": "2025-07-14T12:34:58.065329"
  },
  {
    "epoch": 2341,
    "batch_idx": 128,
    "loss": 0.2912749946117401,
    "learning_rate": 7.002931747548765e-05,
    "timestep_mean": 442.375,
    "epsilon_norm": 6.952081203460693,
    "prediction_norm": 5.797149181365967,
    "timestamp": "2025-07-14T12:34:58.070503"
  },
  {
    "epoch": 2341,
    "batch_idx": 160,
    "loss": 0.27578282356262207,
    "learning_rate": 7.002931747548765e-05,
    "timestep_mean": 419.8125,
    "epsilon_norm": 8.032363891601562,
    "prediction_norm": 6.269959449768066,
    "timestamp": "2025-07-14T12:34:58.075791"
  },
  {
    "epoch": 2341,
    "batch_idx": 192,
    "loss": 0.8734056353569031,
    "learning_rate": 7.002931747548765e-05,
    "timestep_mean": 436.375,
    "epsilon_norm": 4.311412811279297,
    "prediction_norm": 2.9873650074005127,
    "timestamp": "2025-07-14T12:34:58.080861"
  },
  {
    "epoch": 2342,
    "batch_idx": 0,
    "loss": 0.33109983801841736,
    "learning_rate": 7.04307093628639e-05,
    "timestep_mean": 554.3125,
    "epsilon_norm": 7.8059186935424805,
    "prediction_norm": 6.613068103790283,
    "timestamp": "2025-07-14T12:34:58.085963"
  },
  {
    "epoch": 2342,
    "batch_idx": 32,
    "loss": 0.18178310990333557,
    "learning_rate": 7.04307093628639e-05,
    "timestep_mean": 603.53125,
    "epsilon_norm": 6.527848720550537,
    "prediction_norm": 5.565823554992676,
    "timestamp": "2025-07-14T12:34:58.091084"
  },
  {
    "epoch": 2342,
    "batch_idx": 64,
    "loss": 0.2516022026538849,
    "learning_rate": 7.04307093628639e-05,
    "timestep_mean": 501.4375,
    "epsilon_norm": 7.899616241455078,
    "prediction_norm": 6.749431133270264,
    "timestamp": "2025-07-14T12:34:58.096343"
  },
  {
    "epoch": 2342,
    "batch_idx": 96,
    "loss": 0.15989287197589874,
    "learning_rate": 7.04307093628639e-05,
    "timestep_mean": 485.625,
    "epsilon_norm": 7.404068946838379,
    "prediction_norm": 6.484110355377197,
    "timestamp": "2025-07-14T12:34:58.101448"
  },
  {
    "epoch": 2342,
    "batch_idx": 128,
    "loss": 0.5026401281356812,
    "learning_rate": 7.04307093628639e-05,
    "timestep_mean": 503.28125,
    "epsilon_norm": 8.380474090576172,
    "prediction_norm": 6.532435894012451,
    "timestamp": "2025-07-14T12:34:58.106576"
  },
  {
    "epoch": 2342,
    "batch_idx": 160,
    "loss": 0.2305489480495453,
    "learning_rate": 7.04307093628639e-05,
    "timestep_mean": 535.40625,
    "epsilon_norm": 7.582152366638184,
    "prediction_norm": 6.317266941070557,
    "timestamp": "2025-07-14T12:34:58.111864"
  },
  {
    "epoch": 2342,
    "batch_idx": 192,
    "loss": 0.31525611877441406,
    "learning_rate": 7.04307093628639e-05,
    "timestep_mean": 376.5,
    "epsilon_norm": 4.0184173583984375,
    "prediction_norm": 3.580413818359375,
    "timestamp": "2025-07-14T12:34:58.116959"
  },
  {
    "epoch": 2343,
    "batch_idx": 0,
    "loss": 0.19449207186698914,
    "learning_rate": 7.083316116976275e-05,
    "timestep_mean": 538.375,
    "epsilon_norm": 8.17616081237793,
    "prediction_norm": 7.406454086303711,
    "timestamp": "2025-07-14T12:34:58.122193"
  },
  {
    "epoch": 2343,
    "batch_idx": 32,
    "loss": 0.13638879358768463,
    "learning_rate": 7.083316116976275e-05,
    "timestep_mean": 563.375,
    "epsilon_norm": 8.190308570861816,
    "prediction_norm": 7.444912910461426,
    "timestamp": "2025-07-14T12:34:58.127390"
  },
  {
    "epoch": 2343,
    "batch_idx": 64,
    "loss": 0.3528350591659546,
    "learning_rate": 7.083316116976275e-05,
    "timestep_mean": 447.9375,
    "epsilon_norm": 8.382843971252441,
    "prediction_norm": 6.88813591003418,
    "timestamp": "2025-07-14T12:34:58.132547"
  },
  {
    "epoch": 2343,
    "batch_idx": 96,
    "loss": 0.1835789829492569,
    "learning_rate": 7.083316116976275e-05,
    "timestep_mean": 555.59375,
    "epsilon_norm": 7.688449382781982,
    "prediction_norm": 7.034073829650879,
    "timestamp": "2025-07-14T12:34:58.137801"
  },
  {
    "epoch": 2343,
    "batch_idx": 128,
    "loss": 0.3151312470436096,
    "learning_rate": 7.083316116976275e-05,
    "timestep_mean": 556.1875,
    "epsilon_norm": 8.465173721313477,
    "prediction_norm": 6.57277774810791,
    "timestamp": "2025-07-14T12:34:58.142954"
  },
  {
    "epoch": 2343,
    "batch_idx": 160,
    "loss": 0.25743991136550903,
    "learning_rate": 7.083316116976275e-05,
    "timestep_mean": 442.90625,
    "epsilon_norm": 8.33230209350586,
    "prediction_norm": 6.69269323348999,
    "timestamp": "2025-07-14T12:34:58.148138"
  },
  {
    "epoch": 2343,
    "batch_idx": 192,
    "loss": 0.26975005865097046,
    "learning_rate": 7.083316116976275e-05,
    "timestep_mean": 412.625,
    "epsilon_norm": 3.4880456924438477,
    "prediction_norm": 3.5190529823303223,
    "timestamp": "2025-07-14T12:34:58.153200"
  },
  {
    "epoch": 2344,
    "batch_idx": 0,
    "loss": 0.09298470616340637,
    "learning_rate": 7.123667190317396e-05,
    "timestep_mean": 540.4375,
    "epsilon_norm": 7.335867404937744,
    "prediction_norm": 7.010931015014648,
    "timestamp": "2025-07-14T12:34:58.158383"
  },
  {
    "epoch": 2344,
    "batch_idx": 32,
    "loss": 0.2746508717536926,
    "learning_rate": 7.123667190317396e-05,
    "timestep_mean": 545.375,
    "epsilon_norm": 8.148682594299316,
    "prediction_norm": 6.838063716888428,
    "timestamp": "2025-07-14T12:34:58.163642"
  },
  {
    "epoch": 2344,
    "batch_idx": 64,
    "loss": 0.1584855616092682,
    "learning_rate": 7.123667190317396e-05,
    "timestep_mean": 550.625,
    "epsilon_norm": 7.988854885101318,
    "prediction_norm": 7.152002811431885,
    "timestamp": "2025-07-14T12:34:58.168742"
  },
  {
    "epoch": 2344,
    "batch_idx": 96,
    "loss": 0.188024640083313,
    "learning_rate": 7.123667190317396e-05,
    "timestep_mean": 591.4375,
    "epsilon_norm": 7.957474231719971,
    "prediction_norm": 6.878891468048096,
    "timestamp": "2025-07-14T12:34:58.173854"
  },
  {
    "epoch": 2344,
    "batch_idx": 128,
    "loss": 0.1838163286447525,
    "learning_rate": 7.123667190317396e-05,
    "timestep_mean": 539.875,
    "epsilon_norm": 9.18596363067627,
    "prediction_norm": 8.643683433532715,
    "timestamp": "2025-07-14T12:34:58.179123"
  },
  {
    "epoch": 2344,
    "batch_idx": 160,
    "loss": 0.6412092447280884,
    "learning_rate": 7.123667190317396e-05,
    "timestep_mean": 490.71875,
    "epsilon_norm": 9.25178050994873,
    "prediction_norm": 7.458720684051514,
    "timestamp": "2025-07-14T12:34:58.184335"
  },
  {
    "epoch": 2344,
    "batch_idx": 192,
    "loss": 0.38221263885498047,
    "learning_rate": 7.123667190317396e-05,
    "timestep_mean": 384.5,
    "epsilon_norm": 4.2130913734436035,
    "prediction_norm": 3.737029552459717,
    "timestamp": "2025-07-14T12:34:58.189514"
  },
  {
    "epoch": 2345,
    "batch_idx": 0,
    "loss": 0.2352771908044815,
    "learning_rate": 7.164124056747535e-05,
    "timestep_mean": 489.4375,
    "epsilon_norm": 8.272148132324219,
    "prediction_norm": 7.218472957611084,
    "timestamp": "2025-07-14T12:34:58.194656"
  },
  {
    "epoch": 2345,
    "batch_idx": 32,
    "loss": 0.18721561133861542,
    "learning_rate": 7.164124056747535e-05,
    "timestep_mean": 517.59375,
    "epsilon_norm": 7.9044108390808105,
    "prediction_norm": 7.164804935455322,
    "timestamp": "2025-07-14T12:34:58.199705"
  },
  {
    "epoch": 2345,
    "batch_idx": 64,
    "loss": 0.24399897456169128,
    "learning_rate": 7.164124056747535e-05,
    "timestep_mean": 550.25,
    "epsilon_norm": 7.642945766448975,
    "prediction_norm": 6.915175914764404,
    "timestamp": "2025-07-14T12:34:58.204909"
  },
  {
    "epoch": 2345,
    "batch_idx": 96,
    "loss": 0.26848098635673523,
    "learning_rate": 7.164124056747535e-05,
    "timestep_mean": 464.15625,
    "epsilon_norm": 6.92387580871582,
    "prediction_norm": 5.758957862854004,
    "timestamp": "2025-07-14T12:34:58.209951"
  },
  {
    "epoch": 2345,
    "batch_idx": 128,
    "loss": 0.2689669728279114,
    "learning_rate": 7.164124056747535e-05,
    "timestep_mean": 538.96875,
    "epsilon_norm": 7.8539228439331055,
    "prediction_norm": 6.554204940795898,
    "timestamp": "2025-07-14T12:34:58.215083"
  },
  {
    "epoch": 2345,
    "batch_idx": 160,
    "loss": 0.30517953634262085,
    "learning_rate": 7.164124056747535e-05,
    "timestep_mean": 551.34375,
    "epsilon_norm": 8.037046432495117,
    "prediction_norm": 6.381122589111328,
    "timestamp": "2025-07-14T12:34:58.220247"
  },
  {
    "epoch": 2345,
    "batch_idx": 192,
    "loss": 0.34540170431137085,
    "learning_rate": 7.164124056747535e-05,
    "timestep_mean": 423.125,
    "epsilon_norm": 3.6440861225128174,
    "prediction_norm": 2.8034427165985107,
    "timestamp": "2025-07-14T12:34:58.225447"
  },
  {
    "epoch": 2346,
    "batch_idx": 0,
    "loss": 0.20401638746261597,
    "learning_rate": 7.204686616443363e-05,
    "timestep_mean": 489.71875,
    "epsilon_norm": 7.950667858123779,
    "prediction_norm": 6.712552070617676,
    "timestamp": "2025-07-14T12:34:58.230703"
  },
  {
    "epoch": 2346,
    "batch_idx": 32,
    "loss": 0.235398069024086,
    "learning_rate": 7.204686616443363e-05,
    "timestep_mean": 567.4375,
    "epsilon_norm": 7.236448764801025,
    "prediction_norm": 5.958629608154297,
    "timestamp": "2025-07-14T12:34:58.235737"
  },
  {
    "epoch": 2346,
    "batch_idx": 64,
    "loss": 0.17986413836479187,
    "learning_rate": 7.204686616443363e-05,
    "timestep_mean": 561.375,
    "epsilon_norm": 7.126068115234375,
    "prediction_norm": 6.07232141494751,
    "timestamp": "2025-07-14T12:34:58.240809"
  },
  {
    "epoch": 2346,
    "batch_idx": 96,
    "loss": 0.20907139778137207,
    "learning_rate": 7.204686616443363e-05,
    "timestep_mean": 507.75,
    "epsilon_norm": 7.867208957672119,
    "prediction_norm": 6.513485431671143,
    "timestamp": "2025-07-14T12:34:58.246044"
  },
  {
    "epoch": 2346,
    "batch_idx": 128,
    "loss": 0.3945178687572479,
    "learning_rate": 7.204686616443363e-05,
    "timestep_mean": 574.09375,
    "epsilon_norm": 7.596435070037842,
    "prediction_norm": 6.291482925415039,
    "timestamp": "2025-07-14T12:34:58.251218"
  },
  {
    "epoch": 2346,
    "batch_idx": 160,
    "loss": 0.15394245088100433,
    "learning_rate": 7.204686616443363e-05,
    "timestep_mean": 523.28125,
    "epsilon_norm": 6.996439456939697,
    "prediction_norm": 6.379874229431152,
    "timestamp": "2025-07-14T12:34:58.256353"
  },
  {
    "epoch": 2346,
    "batch_idx": 192,
    "loss": 0.11912186443805695,
    "learning_rate": 7.204686616443363e-05,
    "timestep_mean": 644.75,
    "epsilon_norm": 3.433292865753174,
    "prediction_norm": 2.952523946762085,
    "timestamp": "2025-07-14T12:34:58.261433"
  },
  {
    "epoch": 2347,
    "batch_idx": 0,
    "loss": 0.21356560289859772,
    "learning_rate": 7.245354769320812e-05,
    "timestep_mean": 576.6875,
    "epsilon_norm": 7.679319858551025,
    "prediction_norm": 6.703714370727539,
    "timestamp": "2025-07-14T12:34:58.266562"
  },
  {
    "epoch": 2347,
    "batch_idx": 32,
    "loss": 0.19128753244876862,
    "learning_rate": 7.245354769320812e-05,
    "timestep_mean": 467.3125,
    "epsilon_norm": 8.51722240447998,
    "prediction_norm": 7.379830837249756,
    "timestamp": "2025-07-14T12:34:58.271797"
  },
  {
    "epoch": 2347,
    "batch_idx": 64,
    "loss": 0.2186601758003235,
    "learning_rate": 7.245354769320812e-05,
    "timestep_mean": 418.53125,
    "epsilon_norm": 8.050436019897461,
    "prediction_norm": 6.867757797241211,
    "timestamp": "2025-07-14T12:34:58.276898"
  },
  {
    "epoch": 2347,
    "batch_idx": 96,
    "loss": 0.286632776260376,
    "learning_rate": 7.245354769320812e-05,
    "timestep_mean": 479.9375,
    "epsilon_norm": 8.06539535522461,
    "prediction_norm": 6.505840301513672,
    "timestamp": "2025-07-14T12:34:58.281971"
  },
  {
    "epoch": 2347,
    "batch_idx": 128,
    "loss": 0.6170433759689331,
    "learning_rate": 7.245354769320812e-05,
    "timestep_mean": 456.78125,
    "epsilon_norm": 9.926843643188477,
    "prediction_norm": 7.515768527984619,
    "timestamp": "2025-07-14T12:34:58.287173"
  },
  {
    "epoch": 2347,
    "batch_idx": 160,
    "loss": 0.321554958820343,
    "learning_rate": 7.245354769320812e-05,
    "timestep_mean": 475.03125,
    "epsilon_norm": 8.012322425842285,
    "prediction_norm": 6.678215503692627,
    "timestamp": "2025-07-14T12:34:58.292378"
  },
  {
    "epoch": 2347,
    "batch_idx": 192,
    "loss": 0.1280384361743927,
    "learning_rate": 7.245354769320812e-05,
    "timestep_mean": 412.75,
    "epsilon_norm": 2.508409261703491,
    "prediction_norm": 2.5273983478546143,
    "timestamp": "2025-07-14T12:34:58.297515"
  },
  {
    "epoch": 2348,
    "batch_idx": 0,
    "loss": 0.2841694951057434,
    "learning_rate": 7.28612841503526e-05,
    "timestep_mean": 525.46875,
    "epsilon_norm": 8.809432983398438,
    "prediction_norm": 7.678553104400635,
    "timestamp": "2025-07-14T12:34:58.302641"
  },
  {
    "epoch": 2348,
    "batch_idx": 32,
    "loss": 0.22271086275577545,
    "learning_rate": 7.28612841503526e-05,
    "timestep_mean": 495.875,
    "epsilon_norm": 8.130072593688965,
    "prediction_norm": 7.012895584106445,
    "timestamp": "2025-07-14T12:34:58.307744"
  },
  {
    "epoch": 2348,
    "batch_idx": 64,
    "loss": 0.3828314542770386,
    "learning_rate": 7.28612841503526e-05,
    "timestep_mean": 494.0625,
    "epsilon_norm": 8.371291160583496,
    "prediction_norm": 6.5919647216796875,
    "timestamp": "2025-07-14T12:34:58.312957"
  },
  {
    "epoch": 2348,
    "batch_idx": 96,
    "loss": 0.20910361409187317,
    "learning_rate": 7.28612841503526e-05,
    "timestep_mean": 453.75,
    "epsilon_norm": 8.113737106323242,
    "prediction_norm": 7.231832504272461,
    "timestamp": "2025-07-14T12:34:58.318069"
  },
  {
    "epoch": 2348,
    "batch_idx": 128,
    "loss": 0.14741834998130798,
    "learning_rate": 7.28612841503526e-05,
    "timestep_mean": 516.46875,
    "epsilon_norm": 8.178075790405273,
    "prediction_norm": 7.459428310394287,
    "timestamp": "2025-07-14T12:34:58.323030"
  },
  {
    "epoch": 2348,
    "batch_idx": 160,
    "loss": 0.14513429999351501,
    "learning_rate": 7.28612841503526e-05,
    "timestep_mean": 510.84375,
    "epsilon_norm": 7.974229335784912,
    "prediction_norm": 7.625633239746094,
    "timestamp": "2025-07-14T12:34:58.328190"
  },
  {
    "epoch": 2348,
    "batch_idx": 192,
    "loss": 0.3907456398010254,
    "learning_rate": 7.28612841503526e-05,
    "timestep_mean": 484.875,
    "epsilon_norm": 3.960996627807617,
    "prediction_norm": 2.979231357574463,
    "timestamp": "2025-07-14T12:34:58.333321"
  },
  {
    "epoch": 2349,
    "batch_idx": 0,
    "loss": 0.4714549779891968,
    "learning_rate": 7.327007452981778e-05,
    "timestep_mean": 527.8125,
    "epsilon_norm": 7.887132167816162,
    "prediction_norm": 5.750972270965576,
    "timestamp": "2025-07-14T12:34:58.338640"
  },
  {
    "epoch": 2349,
    "batch_idx": 32,
    "loss": 0.4006792902946472,
    "learning_rate": 7.327007452981778e-05,
    "timestep_mean": 431.5,
    "epsilon_norm": 7.19498348236084,
    "prediction_norm": 5.168997764587402,
    "timestamp": "2025-07-14T12:34:58.343733"
  },
  {
    "epoch": 2349,
    "batch_idx": 64,
    "loss": 0.2743619382381439,
    "learning_rate": 7.327007452981778e-05,
    "timestep_mean": 484.625,
    "epsilon_norm": 7.667065620422363,
    "prediction_norm": 6.5853962898254395,
    "timestamp": "2025-07-14T12:34:58.348764"
  },
  {
    "epoch": 2349,
    "batch_idx": 96,
    "loss": 0.3158882260322571,
    "learning_rate": 7.327007452981778e-05,
    "timestep_mean": 566.125,
    "epsilon_norm": 8.09742546081543,
    "prediction_norm": 6.705521583557129,
    "timestamp": "2025-07-14T12:34:58.353962"
  },
  {
    "epoch": 2349,
    "batch_idx": 128,
    "loss": 0.2737727761268616,
    "learning_rate": 7.327007452981778e-05,
    "timestep_mean": 451.875,
    "epsilon_norm": 8.853933334350586,
    "prediction_norm": 7.376881122589111,
    "timestamp": "2025-07-14T12:34:58.359133"
  },
  {
    "epoch": 2349,
    "batch_idx": 160,
    "loss": 0.42880260944366455,
    "learning_rate": 7.327007452981778e-05,
    "timestep_mean": 412.0,
    "epsilon_norm": 8.273590087890625,
    "prediction_norm": 6.61895751953125,
    "timestamp": "2025-07-14T12:34:58.364271"
  },
  {
    "epoch": 2349,
    "batch_idx": 192,
    "loss": 0.27058088779449463,
    "learning_rate": 7.327007452981778e-05,
    "timestep_mean": 510.625,
    "epsilon_norm": 4.774011135101318,
    "prediction_norm": 4.90500020980835,
    "timestamp": "2025-07-14T12:34:58.369388"
  },
  {
    "epoch": 2350,
    "batch_idx": 0,
    "loss": 0.34718772768974304,
    "learning_rate": 7.367991782295404e-05,
    "timestep_mean": 384.6875,
    "epsilon_norm": 7.573118686676025,
    "prediction_norm": 5.526468276977539,
    "timestamp": "2025-07-14T12:34:58.374468"
  },
  {
    "epoch": 2350,
    "batch_idx": 32,
    "loss": 0.438260555267334,
    "learning_rate": 7.367991782295404e-05,
    "timestep_mean": 389.78125,
    "epsilon_norm": 8.685976028442383,
    "prediction_norm": 7.288602828979492,
    "timestamp": "2025-07-14T12:34:58.379642"
  },
  {
    "epoch": 2350,
    "batch_idx": 64,
    "loss": 0.15786108374595642,
    "learning_rate": 7.367991782295404e-05,
    "timestep_mean": 535.125,
    "epsilon_norm": 7.623051166534424,
    "prediction_norm": 6.894251346588135,
    "timestamp": "2025-07-14T12:34:58.384715"
  },
  {
    "epoch": 2350,
    "batch_idx": 96,
    "loss": 0.39687517285346985,
    "learning_rate": 7.367991782295404e-05,
    "timestep_mean": 473.71875,
    "epsilon_norm": 8.333761215209961,
    "prediction_norm": 7.251819610595703,
    "timestamp": "2025-07-14T12:34:58.389706"
  },
  {
    "epoch": 2350,
    "batch_idx": 128,
    "loss": 0.19648905098438263,
    "learning_rate": 7.367991782295404e-05,
    "timestep_mean": 586.03125,
    "epsilon_norm": 7.764552116394043,
    "prediction_norm": 6.604463577270508,
    "timestamp": "2025-07-14T12:34:58.394894"
  },
  {
    "epoch": 2350,
    "batch_idx": 160,
    "loss": 0.3021000921726227,
    "learning_rate": 7.367991782295404e-05,
    "timestep_mean": 464.40625,
    "epsilon_norm": 7.7968878746032715,
    "prediction_norm": 6.873538494110107,
    "timestamp": "2025-07-14T12:34:58.400002"
  },
  {
    "epoch": 2350,
    "batch_idx": 192,
    "loss": 0.22772400081157684,
    "learning_rate": 7.367991782295404e-05,
    "timestep_mean": 396.125,
    "epsilon_norm": 3.349428415298462,
    "prediction_norm": 2.5928995609283447,
    "timestamp": "2025-07-14T12:34:58.405071"
  },
  {
    "epoch": 2351,
    "batch_idx": 0,
    "loss": 0.29912418127059937,
    "learning_rate": 7.409081301851386e-05,
    "timestep_mean": 485.03125,
    "epsilon_norm": 8.6463623046875,
    "prediction_norm": 8.197110176086426,
    "timestamp": "2025-07-14T12:34:58.409863"
  },
  {
    "epoch": 2351,
    "batch_idx": 32,
    "loss": 0.4389459788799286,
    "learning_rate": 7.409081301851386e-05,
    "timestep_mean": 462.1875,
    "epsilon_norm": 8.786883354187012,
    "prediction_norm": 6.863731861114502,
    "timestamp": "2025-07-14T12:34:58.414706"
  },
  {
    "epoch": 2351,
    "batch_idx": 64,
    "loss": 0.4101848602294922,
    "learning_rate": 7.409081301851386e-05,
    "timestep_mean": 398.3125,
    "epsilon_norm": 7.9043049812316895,
    "prediction_norm": 6.345711708068848,
    "timestamp": "2025-07-14T12:34:58.419985"
  },
  {
    "epoch": 2351,
    "batch_idx": 96,
    "loss": 0.25864118337631226,
    "learning_rate": 7.409081301851386e-05,
    "timestep_mean": 422.375,
    "epsilon_norm": 7.283708095550537,
    "prediction_norm": 6.2918500900268555,
    "timestamp": "2025-07-14T12:34:58.425071"
  },
  {
    "epoch": 2351,
    "batch_idx": 128,
    "loss": 0.15450143814086914,
    "learning_rate": 7.409081301851386e-05,
    "timestep_mean": 497.59375,
    "epsilon_norm": 7.124826431274414,
    "prediction_norm": 6.494288444519043,
    "timestamp": "2025-07-14T12:34:58.430312"
  },
  {
    "epoch": 2351,
    "batch_idx": 160,
    "loss": 0.2584601640701294,
    "learning_rate": 7.409081301851386e-05,
    "timestep_mean": 490.96875,
    "epsilon_norm": 8.719779014587402,
    "prediction_norm": 7.527294635772705,
    "timestamp": "2025-07-14T12:34:58.435249"
  },
  {
    "epoch": 2351,
    "batch_idx": 192,
    "loss": 0.5107377171516418,
    "learning_rate": 7.409081301851386e-05,
    "timestep_mean": 491.25,
    "epsilon_norm": 4.178795337677002,
    "prediction_norm": 3.254101037979126,
    "timestamp": "2025-07-14T12:34:58.440387"
  },
  {
    "epoch": 2352,
    "batch_idx": 0,
    "loss": 0.1873399317264557,
    "learning_rate": 7.450275910265427e-05,
    "timestep_mean": 539.875,
    "epsilon_norm": 7.4389543533325195,
    "prediction_norm": 6.392887115478516,
    "timestamp": "2025-07-14T12:34:58.445550"
  },
  {
    "epoch": 2352,
    "batch_idx": 32,
    "loss": 0.24947898089885712,
    "learning_rate": 7.450275910265427e-05,
    "timestep_mean": 488.90625,
    "epsilon_norm": 7.291805267333984,
    "prediction_norm": 6.037703514099121,
    "timestamp": "2025-07-14T12:34:58.450709"
  },
  {
    "epoch": 2352,
    "batch_idx": 64,
    "loss": 0.2812609076499939,
    "learning_rate": 7.450275910265427e-05,
    "timestep_mean": 503.90625,
    "epsilon_norm": 8.25572681427002,
    "prediction_norm": 7.166568756103516,
    "timestamp": "2025-07-14T12:34:58.455910"
  },
  {
    "epoch": 2352,
    "batch_idx": 96,
    "loss": 0.3495723009109497,
    "learning_rate": 7.450275910265427e-05,
    "timestep_mean": 480.03125,
    "epsilon_norm": 6.71419095993042,
    "prediction_norm": 5.972560882568359,
    "timestamp": "2025-07-14T12:34:58.461043"
  },
  {
    "epoch": 2352,
    "batch_idx": 128,
    "loss": 0.2010752260684967,
    "learning_rate": 7.450275910265427e-05,
    "timestep_mean": 482.65625,
    "epsilon_norm": 7.508735179901123,
    "prediction_norm": 6.112357139587402,
    "timestamp": "2025-07-14T12:34:58.466141"
  },
  {
    "epoch": 2352,
    "batch_idx": 160,
    "loss": 0.19390463829040527,
    "learning_rate": 7.450275910265427e-05,
    "timestep_mean": 471.5625,
    "epsilon_norm": 7.72269868850708,
    "prediction_norm": 6.697182655334473,
    "timestamp": "2025-07-14T12:34:58.471301"
  },
  {
    "epoch": 2352,
    "batch_idx": 192,
    "loss": 0.11435334384441376,
    "learning_rate": 7.450275910265427e-05,
    "timestep_mean": 493.125,
    "epsilon_norm": 4.514087677001953,
    "prediction_norm": 4.5461015701293945,
    "timestamp": "2025-07-14T12:34:58.476307"
  },
  {
    "epoch": 2353,
    "batch_idx": 0,
    "loss": 0.1593419909477234,
    "learning_rate": 7.491575505893895e-05,
    "timestep_mean": 490.75,
    "epsilon_norm": 8.28780746459961,
    "prediction_norm": 7.342077255249023,
    "timestamp": "2025-07-14T12:34:58.481437"
  },
  {
    "epoch": 2353,
    "batch_idx": 32,
    "loss": 0.24550873041152954,
    "learning_rate": 7.491575505893895e-05,
    "timestep_mean": 549.40625,
    "epsilon_norm": 6.965670585632324,
    "prediction_norm": 5.998417854309082,
    "timestamp": "2025-07-14T12:34:58.486550"
  },
  {
    "epoch": 2353,
    "batch_idx": 64,
    "loss": 0.5325779914855957,
    "learning_rate": 7.491575505893895e-05,
    "timestep_mean": 434.375,
    "epsilon_norm": 8.628934860229492,
    "prediction_norm": 6.278731346130371,
    "timestamp": "2025-07-14T12:34:58.491713"
  },
  {
    "epoch": 2353,
    "batch_idx": 96,
    "loss": 0.3912431597709656,
    "learning_rate": 7.491575505893895e-05,
    "timestep_mean": 464.90625,
    "epsilon_norm": 7.946142673492432,
    "prediction_norm": 6.654686450958252,
    "timestamp": "2025-07-14T12:34:58.496681"
  },
  {
    "epoch": 2353,
    "batch_idx": 128,
    "loss": 0.12612885236740112,
    "learning_rate": 7.491575505893895e-05,
    "timestep_mean": 561.0625,
    "epsilon_norm": 6.983274936676025,
    "prediction_norm": 6.111461162567139,
    "timestamp": "2025-07-14T12:34:58.501621"
  },
  {
    "epoch": 2353,
    "batch_idx": 160,
    "loss": 0.380440354347229,
    "learning_rate": 7.491575505893895e-05,
    "timestep_mean": 478.03125,
    "epsilon_norm": 7.79957914352417,
    "prediction_norm": 6.263989448547363,
    "timestamp": "2025-07-14T12:34:58.506674"
  },
  {
    "epoch": 2353,
    "batch_idx": 192,
    "loss": 0.11961393803358078,
    "learning_rate": 7.491575505893895e-05,
    "timestep_mean": 598.375,
    "epsilon_norm": 2.03178071975708,
    "prediction_norm": 2.15254545211792,
    "timestamp": "2025-07-14T12:34:58.511819"
  },
  {
    "epoch": 2354,
    "batch_idx": 0,
    "loss": 0.25647106766700745,
    "learning_rate": 7.532979986834189e-05,
    "timestep_mean": 551.40625,
    "epsilon_norm": 8.42772102355957,
    "prediction_norm": 7.1182379722595215,
    "timestamp": "2025-07-14T12:34:58.517184"
  },
  {
    "epoch": 2354,
    "batch_idx": 32,
    "loss": 0.20551899075508118,
    "learning_rate": 7.532979986834189e-05,
    "timestep_mean": 482.34375,
    "epsilon_norm": 8.229451179504395,
    "prediction_norm": 7.372935771942139,
    "timestamp": "2025-07-14T12:34:58.522213"
  },
  {
    "epoch": 2354,
    "batch_idx": 64,
    "loss": 0.17000678181648254,
    "learning_rate": 7.532979986834189e-05,
    "timestep_mean": 499.125,
    "epsilon_norm": 7.683428764343262,
    "prediction_norm": 6.901005268096924,
    "timestamp": "2025-07-14T12:34:58.527232"
  },
  {
    "epoch": 2354,
    "batch_idx": 96,
    "loss": 0.2714840769767761,
    "learning_rate": 7.532979986834189e-05,
    "timestep_mean": 587.65625,
    "epsilon_norm": 8.0098237991333,
    "prediction_norm": 6.673581600189209,
    "timestamp": "2025-07-14T12:34:58.532357"
  },
  {
    "epoch": 2354,
    "batch_idx": 128,
    "loss": 0.42507505416870117,
    "learning_rate": 7.532979986834189e-05,
    "timestep_mean": 479.71875,
    "epsilon_norm": 9.751275062561035,
    "prediction_norm": 7.648978233337402,
    "timestamp": "2025-07-14T12:34:58.537443"
  },
  {
    "epoch": 2354,
    "batch_idx": 160,
    "loss": 0.21598151326179504,
    "learning_rate": 7.532979986834189e-05,
    "timestep_mean": 566.96875,
    "epsilon_norm": 7.300812244415283,
    "prediction_norm": 6.490701198577881,
    "timestamp": "2025-07-14T12:34:58.542598"
  },
  {
    "epoch": 2354,
    "batch_idx": 192,
    "loss": 0.36811062693595886,
    "learning_rate": 7.532979986834189e-05,
    "timestep_mean": 448.5,
    "epsilon_norm": 5.6613287925720215,
    "prediction_norm": 5.165212631225586,
    "timestamp": "2025-07-14T12:34:58.547676"
  },
  {
    "epoch": 2355,
    "batch_idx": 0,
    "loss": 0.2201211154460907,
    "learning_rate": 7.574489250924828e-05,
    "timestep_mean": 509.28125,
    "epsilon_norm": 8.315731048583984,
    "prediction_norm": 6.782807350158691,
    "timestamp": "2025-07-14T12:34:58.552806"
  },
  {
    "epoch": 2355,
    "batch_idx": 32,
    "loss": 0.26332998275756836,
    "learning_rate": 7.574489250924828e-05,
    "timestep_mean": 529.5,
    "epsilon_norm": 7.854105472564697,
    "prediction_norm": 6.942696571350098,
    "timestamp": "2025-07-14T12:34:58.557971"
  },
  {
    "epoch": 2355,
    "batch_idx": 64,
    "loss": 0.20291218161582947,
    "learning_rate": 7.574489250924828e-05,
    "timestep_mean": 412.15625,
    "epsilon_norm": 6.8422441482543945,
    "prediction_norm": 6.173833847045898,
    "timestamp": "2025-07-14T12:34:58.563029"
  },
  {
    "epoch": 2355,
    "batch_idx": 96,
    "loss": 0.35900822281837463,
    "learning_rate": 7.574489250924828e-05,
    "timestep_mean": 392.0625,
    "epsilon_norm": 7.85186243057251,
    "prediction_norm": 5.750199794769287,
    "timestamp": "2025-07-14T12:34:58.567939"
  },
  {
    "epoch": 2355,
    "batch_idx": 128,
    "loss": 0.159473717212677,
    "learning_rate": 7.574489250924828e-05,
    "timestep_mean": 520.03125,
    "epsilon_norm": 7.670528888702393,
    "prediction_norm": 7.381417751312256,
    "timestamp": "2025-07-14T12:34:58.572796"
  },
  {
    "epoch": 2355,
    "batch_idx": 160,
    "loss": 0.2054952085018158,
    "learning_rate": 7.574489250924828e-05,
    "timestep_mean": 520.125,
    "epsilon_norm": 8.600761413574219,
    "prediction_norm": 7.036421775817871,
    "timestamp": "2025-07-14T12:34:58.577671"
  },
  {
    "epoch": 2355,
    "batch_idx": 192,
    "loss": 0.13361769914627075,
    "learning_rate": 7.574489250924828e-05,
    "timestep_mean": 469.375,
    "epsilon_norm": 2.161914348602295,
    "prediction_norm": 2.1557955741882324,
    "timestamp": "2025-07-14T12:34:58.582499"
  },
  {
    "epoch": 2356,
    "batch_idx": 0,
    "loss": 0.40200382471084595,
    "learning_rate": 7.616103195745852e-05,
    "timestep_mean": 575.46875,
    "epsilon_norm": 8.843893051147461,
    "prediction_norm": 7.645884990692139,
    "timestamp": "2025-07-14T12:34:58.587348"
  },
  {
    "epoch": 2356,
    "batch_idx": 32,
    "loss": 0.5600368976593018,
    "learning_rate": 7.616103195745852e-05,
    "timestep_mean": 502.34375,
    "epsilon_norm": 8.597790718078613,
    "prediction_norm": 6.63147497177124,
    "timestamp": "2025-07-14T12:34:58.592335"
  },
  {
    "epoch": 2356,
    "batch_idx": 64,
    "loss": 0.2765396237373352,
    "learning_rate": 7.616103195745852e-05,
    "timestep_mean": 420.625,
    "epsilon_norm": 7.591140270233154,
    "prediction_norm": 6.327201843261719,
    "timestamp": "2025-07-14T12:34:58.597434"
  },
  {
    "epoch": 2356,
    "batch_idx": 96,
    "loss": 0.16526076197624207,
    "learning_rate": 7.616103195745852e-05,
    "timestep_mean": 545.5625,
    "epsilon_norm": 7.352038383483887,
    "prediction_norm": 6.446715831756592,
    "timestamp": "2025-07-14T12:34:58.602360"
  },
  {
    "epoch": 2356,
    "batch_idx": 128,
    "loss": 0.31030040979385376,
    "learning_rate": 7.616103195745852e-05,
    "timestep_mean": 470.8125,
    "epsilon_norm": 8.693499565124512,
    "prediction_norm": 7.080893039703369,
    "timestamp": "2025-07-14T12:34:58.607366"
  },
  {
    "epoch": 2356,
    "batch_idx": 160,
    "loss": 0.2803078889846802,
    "learning_rate": 7.616103195745852e-05,
    "timestep_mean": 443.34375,
    "epsilon_norm": 8.360725402832031,
    "prediction_norm": 6.771475315093994,
    "timestamp": "2025-07-14T12:34:58.612283"
  },
  {
    "epoch": 2356,
    "batch_idx": 192,
    "loss": 0.0977686196565628,
    "learning_rate": 7.616103195745852e-05,
    "timestep_mean": 643.125,
    "epsilon_norm": 4.123141765594482,
    "prediction_norm": 3.6338655948638916,
    "timestamp": "2025-07-14T12:34:58.617348"
  },
  {
    "epoch": 2357,
    "batch_idx": 0,
    "loss": 0.28562915325164795,
    "learning_rate": 7.657821718618989e-05,
    "timestep_mean": 539.0625,
    "epsilon_norm": 7.694952487945557,
    "prediction_norm": 6.736022472381592,
    "timestamp": "2025-07-14T12:34:58.622310"
  },
  {
    "epoch": 2357,
    "batch_idx": 32,
    "loss": 0.37530362606048584,
    "learning_rate": 7.657821718618989e-05,
    "timestep_mean": 409.53125,
    "epsilon_norm": 8.78956127166748,
    "prediction_norm": 6.879211902618408,
    "timestamp": "2025-07-14T12:34:58.627226"
  },
  {
    "epoch": 2357,
    "batch_idx": 64,
    "loss": 0.25870615243911743,
    "learning_rate": 7.657821718618989e-05,
    "timestep_mean": 514.75,
    "epsilon_norm": 9.366050720214844,
    "prediction_norm": 9.155725479125977,
    "timestamp": "2025-07-14T12:34:58.632115"
  },
  {
    "epoch": 2357,
    "batch_idx": 96,
    "loss": 0.7017176151275635,
    "learning_rate": 7.657821718618989e-05,
    "timestep_mean": 422.0,
    "epsilon_norm": 9.190298080444336,
    "prediction_norm": 6.913138389587402,
    "timestamp": "2025-07-14T12:34:58.637033"
  },
  {
    "epoch": 2357,
    "batch_idx": 128,
    "loss": 0.25233158469200134,
    "learning_rate": 7.657821718618989e-05,
    "timestep_mean": 520.8125,
    "epsilon_norm": 8.946146011352539,
    "prediction_norm": 7.60496187210083,
    "timestamp": "2025-07-14T12:34:58.641950"
  },
  {
    "epoch": 2357,
    "batch_idx": 160,
    "loss": 0.4297603666782379,
    "learning_rate": 7.657821718618989e-05,
    "timestep_mean": 448.59375,
    "epsilon_norm": 7.926474094390869,
    "prediction_norm": 6.048633575439453,
    "timestamp": "2025-07-14T12:34:58.646801"
  },
  {
    "epoch": 2357,
    "batch_idx": 192,
    "loss": 0.2793649435043335,
    "learning_rate": 7.657821718618989e-05,
    "timestep_mean": 367.875,
    "epsilon_norm": 3.164400577545166,
    "prediction_norm": 2.62312912940979,
    "timestamp": "2025-07-14T12:34:58.651645"
  },
  {
    "epoch": 2358,
    "batch_idx": 0,
    "loss": 0.3981107473373413,
    "learning_rate": 7.6996447166079e-05,
    "timestep_mean": 466.78125,
    "epsilon_norm": 8.849044799804688,
    "prediction_norm": 7.324704647064209,
    "timestamp": "2025-07-14T12:34:58.656583"
  },
  {
    "epoch": 2358,
    "batch_idx": 32,
    "loss": 0.28561317920684814,
    "learning_rate": 7.6996447166079e-05,
    "timestep_mean": 505.34375,
    "epsilon_norm": 6.832458019256592,
    "prediction_norm": 5.745936870574951,
    "timestamp": "2025-07-14T12:34:58.661491"
  },
  {
    "epoch": 2358,
    "batch_idx": 64,
    "loss": 0.1502707451581955,
    "learning_rate": 7.6996447166079e-05,
    "timestep_mean": 473.40625,
    "epsilon_norm": 6.131302833557129,
    "prediction_norm": 5.447673320770264,
    "timestamp": "2025-07-14T12:34:58.666408"
  },
  {
    "epoch": 2358,
    "batch_idx": 96,
    "loss": 0.22420160472393036,
    "learning_rate": 7.6996447166079e-05,
    "timestep_mean": 490.75,
    "epsilon_norm": 8.433197975158691,
    "prediction_norm": 7.313695430755615,
    "timestamp": "2025-07-14T12:34:58.671384"
  },
  {
    "epoch": 2358,
    "batch_idx": 128,
    "loss": 0.2167392075061798,
    "learning_rate": 7.6996447166079e-05,
    "timestep_mean": 522.40625,
    "epsilon_norm": 7.4358978271484375,
    "prediction_norm": 6.5822272300720215,
    "timestamp": "2025-07-14T12:34:58.676294"
  },
  {
    "epoch": 2358,
    "batch_idx": 160,
    "loss": 0.32162654399871826,
    "learning_rate": 7.6996447166079e-05,
    "timestep_mean": 473.15625,
    "epsilon_norm": 7.585084915161133,
    "prediction_norm": 6.703200817108154,
    "timestamp": "2025-07-14T12:34:58.681191"
  },
  {
    "epoch": 2358,
    "batch_idx": 192,
    "loss": 0.26748499274253845,
    "learning_rate": 7.6996447166079e-05,
    "timestep_mean": 287.5,
    "epsilon_norm": 3.8394765853881836,
    "prediction_norm": 2.416386365890503,
    "timestamp": "2025-07-14T12:34:58.686170"
  },
  {
    "epoch": 2359,
    "batch_idx": 0,
    "loss": 0.2628152072429657,
    "learning_rate": 7.741572086518545e-05,
    "timestep_mean": 480.53125,
    "epsilon_norm": 7.8568243980407715,
    "prediction_norm": 6.47075891494751,
    "timestamp": "2025-07-14T12:34:58.691081"
  },
  {
    "epoch": 2359,
    "batch_idx": 32,
    "loss": 0.12171226739883423,
    "learning_rate": 7.741572086518545e-05,
    "timestep_mean": 609.9375,
    "epsilon_norm": 7.90642786026001,
    "prediction_norm": 7.097747325897217,
    "timestamp": "2025-07-14T12:34:58.695918"
  },
  {
    "epoch": 2359,
    "batch_idx": 64,
    "loss": 0.26274430751800537,
    "learning_rate": 7.741572086518545e-05,
    "timestep_mean": 563.3125,
    "epsilon_norm": 8.91799259185791,
    "prediction_norm": 7.635969161987305,
    "timestamp": "2025-07-14T12:34:58.700879"
  },
  {
    "epoch": 2359,
    "batch_idx": 96,
    "loss": 0.16225144267082214,
    "learning_rate": 7.741572086518545e-05,
    "timestep_mean": 556.0625,
    "epsilon_norm": 7.194034576416016,
    "prediction_norm": 6.034626007080078,
    "timestamp": "2025-07-14T12:34:58.705685"
  },
  {
    "epoch": 2359,
    "batch_idx": 128,
    "loss": 0.24529235064983368,
    "learning_rate": 7.741572086518545e-05,
    "timestep_mean": 482.03125,
    "epsilon_norm": 7.513716697692871,
    "prediction_norm": 6.7288713455200195,
    "timestamp": "2025-07-14T12:34:58.710591"
  },
  {
    "epoch": 2359,
    "batch_idx": 160,
    "loss": 0.2708270847797394,
    "learning_rate": 7.741572086518545e-05,
    "timestep_mean": 456.28125,
    "epsilon_norm": 8.847877502441406,
    "prediction_norm": 7.236850261688232,
    "timestamp": "2025-07-14T12:34:58.715429"
  },
  {
    "epoch": 2359,
    "batch_idx": 192,
    "loss": 0.24103562533855438,
    "learning_rate": 7.741572086518545e-05,
    "timestep_mean": 684.75,
    "epsilon_norm": 3.4509594440460205,
    "prediction_norm": 2.486999034881592,
    "timestamp": "2025-07-14T12:34:58.720417"
  },
  {
    "epoch": 2360,
    "batch_idx": 0,
    "loss": 0.09341603517532349,
    "learning_rate": 7.783603724899274e-05,
    "timestep_mean": 492.28125,
    "epsilon_norm": 5.6180925369262695,
    "prediction_norm": 5.644660949707031,
    "timestamp": "2025-07-14T12:34:58.725490"
  },
  {
    "epoch": 2360,
    "batch_idx": 32,
    "loss": 0.20530611276626587,
    "learning_rate": 7.783603724899274e-05,
    "timestep_mean": 473.5,
    "epsilon_norm": 8.832252502441406,
    "prediction_norm": 7.646117210388184,
    "timestamp": "2025-07-14T12:34:58.730472"
  },
  {
    "epoch": 2360,
    "batch_idx": 64,
    "loss": 0.4542701840400696,
    "learning_rate": 7.783603724899274e-05,
    "timestep_mean": 561.90625,
    "epsilon_norm": 8.21780776977539,
    "prediction_norm": 6.293549537658691,
    "timestamp": "2025-07-14T12:34:58.735436"
  },
  {
    "epoch": 2360,
    "batch_idx": 96,
    "loss": 0.3827953338623047,
    "learning_rate": 7.783603724899274e-05,
    "timestep_mean": 463.125,
    "epsilon_norm": 8.260335922241211,
    "prediction_norm": 6.733423233032227,
    "timestamp": "2025-07-14T12:34:58.740346"
  },
  {
    "epoch": 2360,
    "batch_idx": 128,
    "loss": 0.41214174032211304,
    "learning_rate": 7.783603724899274e-05,
    "timestep_mean": 507.53125,
    "epsilon_norm": 8.903351783752441,
    "prediction_norm": 7.2895612716674805,
    "timestamp": "2025-07-14T12:34:58.745195"
  },
  {
    "epoch": 2360,
    "batch_idx": 160,
    "loss": 0.2532437741756439,
    "learning_rate": 7.783603724899274e-05,
    "timestep_mean": 552.90625,
    "epsilon_norm": 7.869970321655273,
    "prediction_norm": 6.817338466644287,
    "timestamp": "2025-07-14T12:34:58.750054"
  },
  {
    "epoch": 2360,
    "batch_idx": 192,
    "loss": 0.1979389190673828,
    "learning_rate": 7.783603724899274e-05,
    "timestep_mean": 421.625,
    "epsilon_norm": 3.486884117126465,
    "prediction_norm": 2.7155306339263916,
    "timestamp": "2025-07-14T12:34:58.755065"
  },
  {
    "epoch": 2361,
    "batch_idx": 0,
    "loss": 0.2469456046819687,
    "learning_rate": 7.82573952804118e-05,
    "timestep_mean": 556.46875,
    "epsilon_norm": 8.998811721801758,
    "prediction_norm": 7.980412006378174,
    "timestamp": "2025-07-14T12:34:58.759852"
  },
  {
    "epoch": 2361,
    "batch_idx": 32,
    "loss": 0.29753005504608154,
    "learning_rate": 7.82573952804118e-05,
    "timestep_mean": 584.78125,
    "epsilon_norm": 7.6356000900268555,
    "prediction_norm": 6.140786647796631,
    "timestamp": "2025-07-14T12:34:58.764704"
  },
  {
    "epoch": 2361,
    "batch_idx": 64,
    "loss": 0.5883547067642212,
    "learning_rate": 7.82573952804118e-05,
    "timestep_mean": 384.96875,
    "epsilon_norm": 8.361528396606445,
    "prediction_norm": 5.128452301025391,
    "timestamp": "2025-07-14T12:34:58.769574"
  },
  {
    "epoch": 2361,
    "batch_idx": 96,
    "loss": 0.14551305770874023,
    "learning_rate": 7.82573952804118e-05,
    "timestep_mean": 596.4375,
    "epsilon_norm": 7.386944770812988,
    "prediction_norm": 6.295836448669434,
    "timestamp": "2025-07-14T12:34:58.774513"
  },
  {
    "epoch": 2361,
    "batch_idx": 128,
    "loss": 0.1629774421453476,
    "learning_rate": 7.82573952804118e-05,
    "timestep_mean": 540.625,
    "epsilon_norm": 6.602689266204834,
    "prediction_norm": 6.070123195648193,
    "timestamp": "2025-07-14T12:34:58.779389"
  },
  {
    "epoch": 2361,
    "batch_idx": 160,
    "loss": 0.24409566819667816,
    "learning_rate": 7.82573952804118e-05,
    "timestep_mean": 476.3125,
    "epsilon_norm": 7.506642818450928,
    "prediction_norm": 6.058985233306885,
    "timestamp": "2025-07-14T12:34:58.784199"
  },
  {
    "epoch": 2361,
    "batch_idx": 192,
    "loss": 0.31343066692352295,
    "learning_rate": 7.82573952804118e-05,
    "timestep_mean": 487.875,
    "epsilon_norm": 3.373142719268799,
    "prediction_norm": 2.650332450866699,
    "timestamp": "2025-07-14T12:34:58.789022"
  },
  {
    "epoch": 2362,
    "batch_idx": 0,
    "loss": 0.14153531193733215,
    "learning_rate": 7.867979391978408e-05,
    "timestep_mean": 544.15625,
    "epsilon_norm": 7.167873382568359,
    "prediction_norm": 6.630909442901611,
    "timestamp": "2025-07-14T12:34:58.793974"
  },
  {
    "epoch": 2362,
    "batch_idx": 32,
    "loss": 0.3816888928413391,
    "learning_rate": 7.867979391978408e-05,
    "timestep_mean": 456.90625,
    "epsilon_norm": 8.040770530700684,
    "prediction_norm": 5.966794490814209,
    "timestamp": "2025-07-14T12:34:58.798839"
  },
  {
    "epoch": 2362,
    "batch_idx": 64,
    "loss": 0.3156399726867676,
    "learning_rate": 7.867979391978408e-05,
    "timestep_mean": 517.1875,
    "epsilon_norm": 7.648488998413086,
    "prediction_norm": 6.1751251220703125,
    "timestamp": "2025-07-14T12:34:58.803924"
  },
  {
    "epoch": 2362,
    "batch_idx": 96,
    "loss": 0.14956729114055634,
    "learning_rate": 7.867979391978408e-05,
    "timestep_mean": 547.375,
    "epsilon_norm": 6.801197052001953,
    "prediction_norm": 6.2413787841796875,
    "timestamp": "2025-07-14T12:34:58.809116"
  },
  {
    "epoch": 2362,
    "batch_idx": 128,
    "loss": 0.20284360647201538,
    "learning_rate": 7.867979391978408e-05,
    "timestep_mean": 546.09375,
    "epsilon_norm": 8.013557434082031,
    "prediction_norm": 7.143161773681641,
    "timestamp": "2025-07-14T12:34:58.814304"
  },
  {
    "epoch": 2362,
    "batch_idx": 160,
    "loss": 0.26801085472106934,
    "learning_rate": 7.867979391978408e-05,
    "timestep_mean": 461.5,
    "epsilon_norm": 7.735660076141357,
    "prediction_norm": 6.719997882843018,
    "timestamp": "2025-07-14T12:34:58.819351"
  },
  {
    "epoch": 2362,
    "batch_idx": 192,
    "loss": 0.03306486830115318,
    "learning_rate": 7.867979391978408e-05,
    "timestep_mean": 642.375,
    "epsilon_norm": 4.413960933685303,
    "prediction_norm": 4.444421768188477,
    "timestamp": "2025-07-14T12:34:58.824507"
  },
  {
    "epoch": 2363,
    "batch_idx": 0,
    "loss": 0.2559950649738312,
    "learning_rate": 7.910323212488258e-05,
    "timestep_mean": 489.3125,
    "epsilon_norm": 8.395410537719727,
    "prediction_norm": 6.8371429443359375,
    "timestamp": "2025-07-14T12:34:58.829594"
  },
  {
    "epoch": 2363,
    "batch_idx": 32,
    "loss": 0.20815640687942505,
    "learning_rate": 7.910323212488258e-05,
    "timestep_mean": 473.59375,
    "epsilon_norm": 9.322007179260254,
    "prediction_norm": 7.929483413696289,
    "timestamp": "2025-07-14T12:34:58.834689"
  },
  {
    "epoch": 2363,
    "batch_idx": 64,
    "loss": 0.21904689073562622,
    "learning_rate": 7.910323212488258e-05,
    "timestep_mean": 521.4375,
    "epsilon_norm": 7.683909893035889,
    "prediction_norm": 6.699164867401123,
    "timestamp": "2025-07-14T12:34:58.839646"
  },
  {
    "epoch": 2363,
    "batch_idx": 96,
    "loss": 0.40811747312545776,
    "learning_rate": 7.910323212488258e-05,
    "timestep_mean": 448.90625,
    "epsilon_norm": 8.012341499328613,
    "prediction_norm": 6.03804874420166,
    "timestamp": "2025-07-14T12:34:58.844734"
  },
  {
    "epoch": 2363,
    "batch_idx": 128,
    "loss": 0.34539109468460083,
    "learning_rate": 7.910323212488258e-05,
    "timestep_mean": 367.9375,
    "epsilon_norm": 8.040770530700684,
    "prediction_norm": 5.876420497894287,
    "timestamp": "2025-07-14T12:34:58.849826"
  },
  {
    "epoch": 2363,
    "batch_idx": 160,
    "loss": 0.18779988586902618,
    "learning_rate": 7.910323212488258e-05,
    "timestep_mean": 503.03125,
    "epsilon_norm": 7.455063819885254,
    "prediction_norm": 6.2837815284729,
    "timestamp": "2025-07-14T12:34:58.855186"
  },
  {
    "epoch": 2363,
    "batch_idx": 192,
    "loss": 0.15002040565013885,
    "learning_rate": 7.910323212488258e-05,
    "timestep_mean": 471.0,
    "epsilon_norm": 2.6634795665740967,
    "prediction_norm": 2.4112722873687744,
    "timestamp": "2025-07-14T12:34:58.860185"
  },
  {
    "epoch": 2364,
    "batch_idx": 0,
    "loss": 0.2921133041381836,
    "learning_rate": 7.952770885091565e-05,
    "timestep_mean": 518.59375,
    "epsilon_norm": 7.570130825042725,
    "prediction_norm": 6.610939025878906,
    "timestamp": "2025-07-14T12:34:58.865240"
  },
  {
    "epoch": 2364,
    "batch_idx": 32,
    "loss": 0.23454223573207855,
    "learning_rate": 7.952770885091565e-05,
    "timestep_mean": 508.15625,
    "epsilon_norm": 7.755198955535889,
    "prediction_norm": 6.751988410949707,
    "timestamp": "2025-07-14T12:34:58.870456"
  },
  {
    "epoch": 2364,
    "batch_idx": 64,
    "loss": 0.18679626286029816,
    "learning_rate": 7.952770885091565e-05,
    "timestep_mean": 548.9375,
    "epsilon_norm": 8.485301971435547,
    "prediction_norm": 7.760500907897949,
    "timestamp": "2025-07-14T12:34:58.875548"
  },
  {
    "epoch": 2364,
    "batch_idx": 96,
    "loss": 0.18242086470127106,
    "learning_rate": 7.952770885091565e-05,
    "timestep_mean": 488.90625,
    "epsilon_norm": 7.531964302062988,
    "prediction_norm": 6.586057186126709,
    "timestamp": "2025-07-14T12:34:58.880631"
  },
  {
    "epoch": 2364,
    "batch_idx": 128,
    "loss": 0.4094304144382477,
    "learning_rate": 7.952770885091565e-05,
    "timestep_mean": 496.15625,
    "epsilon_norm": 8.963748931884766,
    "prediction_norm": 7.197203636169434,
    "timestamp": "2025-07-14T12:34:58.885729"
  },
  {
    "epoch": 2364,
    "batch_idx": 160,
    "loss": 0.18930701911449432,
    "learning_rate": 7.952770885091565e-05,
    "timestep_mean": 528.6875,
    "epsilon_norm": 6.344852924346924,
    "prediction_norm": 6.248086452484131,
    "timestamp": "2025-07-14T12:34:58.890734"
  },
  {
    "epoch": 2364,
    "batch_idx": 192,
    "loss": 0.2127210944890976,
    "learning_rate": 7.952770885091565e-05,
    "timestep_mean": 545.625,
    "epsilon_norm": 5.22558069229126,
    "prediction_norm": 4.738268852233887,
    "timestamp": "2025-07-14T12:34:58.896005"
  },
  {
    "epoch": 2365,
    "batch_idx": 0,
    "loss": 0.16994798183441162,
    "learning_rate": 7.995322305052922e-05,
    "timestep_mean": 523.5625,
    "epsilon_norm": 8.02548599243164,
    "prediction_norm": 7.904359817504883,
    "timestamp": "2025-07-14T12:34:58.901160"
  },
  {
    "epoch": 2365,
    "batch_idx": 32,
    "loss": 0.2105807363986969,
    "learning_rate": 7.995322305052922e-05,
    "timestep_mean": 491.90625,
    "epsilon_norm": 7.064164161682129,
    "prediction_norm": 5.454349040985107,
    "timestamp": "2025-07-14T12:34:58.906132"
  },
  {
    "epoch": 2365,
    "batch_idx": 64,
    "loss": 0.22264118492603302,
    "learning_rate": 7.995322305052922e-05,
    "timestep_mean": 477.09375,
    "epsilon_norm": 8.077670097351074,
    "prediction_norm": 6.804741859436035,
    "timestamp": "2025-07-14T12:34:58.911262"
  },
  {
    "epoch": 2365,
    "batch_idx": 96,
    "loss": 0.22986820340156555,
    "learning_rate": 7.995322305052922e-05,
    "timestep_mean": 423.71875,
    "epsilon_norm": 8.849818229675293,
    "prediction_norm": 8.263355255126953,
    "timestamp": "2025-07-14T12:34:58.916326"
  },
  {
    "epoch": 2365,
    "batch_idx": 128,
    "loss": 0.31421735882759094,
    "learning_rate": 7.995322305052922e-05,
    "timestep_mean": 501.53125,
    "epsilon_norm": 7.582183361053467,
    "prediction_norm": 6.366546154022217,
    "timestamp": "2025-07-14T12:34:58.921426"
  },
  {
    "epoch": 2365,
    "batch_idx": 160,
    "loss": 0.24667561054229736,
    "learning_rate": 7.995322305052922e-05,
    "timestep_mean": 481.625,
    "epsilon_norm": 7.139087200164795,
    "prediction_norm": 6.065492153167725,
    "timestamp": "2025-07-14T12:34:58.926404"
  },
  {
    "epoch": 2365,
    "batch_idx": 192,
    "loss": 0.08145023882389069,
    "learning_rate": 7.995322305052922e-05,
    "timestep_mean": 545.375,
    "epsilon_norm": 3.450852394104004,
    "prediction_norm": 3.2021842002868652,
    "timestamp": "2025-07-14T12:34:58.931465"
  },
  {
    "epoch": 2366,
    "batch_idx": 0,
    "loss": 0.19605252146720886,
    "learning_rate": 8.037977367380929e-05,
    "timestep_mean": 541.1875,
    "epsilon_norm": 7.620492935180664,
    "prediction_norm": 6.476161956787109,
    "timestamp": "2025-07-14T12:34:58.936724"
  },
  {
    "epoch": 2366,
    "batch_idx": 32,
    "loss": 0.2084159255027771,
    "learning_rate": 8.037977367380929e-05,
    "timestep_mean": 446.0625,
    "epsilon_norm": 8.800457954406738,
    "prediction_norm": 8.041715621948242,
    "timestamp": "2025-07-14T12:34:58.941969"
  },
  {
    "epoch": 2366,
    "batch_idx": 64,
    "loss": 0.12685886025428772,
    "learning_rate": 8.037977367380929e-05,
    "timestep_mean": 632.8125,
    "epsilon_norm": 8.287811279296875,
    "prediction_norm": 7.503299236297607,
    "timestamp": "2025-07-14T12:34:58.947032"
  },
  {
    "epoch": 2366,
    "batch_idx": 96,
    "loss": 0.17193055152893066,
    "learning_rate": 8.037977367380929e-05,
    "timestep_mean": 514.90625,
    "epsilon_norm": 7.041402339935303,
    "prediction_norm": 5.71710205078125,
    "timestamp": "2025-07-14T12:34:58.952241"
  },
  {
    "epoch": 2366,
    "batch_idx": 128,
    "loss": 0.28217950463294983,
    "learning_rate": 8.037977367380929e-05,
    "timestep_mean": 499.9375,
    "epsilon_norm": 8.897944450378418,
    "prediction_norm": 6.96989631652832,
    "timestamp": "2025-07-14T12:34:58.957431"
  },
  {
    "epoch": 2366,
    "batch_idx": 160,
    "loss": 0.26996803283691406,
    "learning_rate": 8.037977367380929e-05,
    "timestep_mean": 508.71875,
    "epsilon_norm": 8.229081153869629,
    "prediction_norm": 7.214666366577148,
    "timestamp": "2025-07-14T12:34:58.962678"
  },
  {
    "epoch": 2366,
    "batch_idx": 192,
    "loss": 0.4025009274482727,
    "learning_rate": 8.037977367380929e-05,
    "timestep_mean": 286.0,
    "epsilon_norm": 3.751009464263916,
    "prediction_norm": 2.3967602252960205,
    "timestamp": "2025-07-14T12:34:58.967831"
  },
  {
    "epoch": 2367,
    "batch_idx": 0,
    "loss": 0.2056063413619995,
    "learning_rate": 8.080735966828474e-05,
    "timestep_mean": 479.53125,
    "epsilon_norm": 8.416074752807617,
    "prediction_norm": 7.213422775268555,
    "timestamp": "2025-07-14T12:34:58.972942"
  },
  {
    "epoch": 2367,
    "batch_idx": 32,
    "loss": 0.2692316174507141,
    "learning_rate": 8.080735966828474e-05,
    "timestep_mean": 511.28125,
    "epsilon_norm": 9.17152214050293,
    "prediction_norm": 7.128687858581543,
    "timestamp": "2025-07-14T12:34:58.978145"
  },
  {
    "epoch": 2367,
    "batch_idx": 64,
    "loss": 0.16348785161972046,
    "learning_rate": 8.080735966828474e-05,
    "timestep_mean": 627.1875,
    "epsilon_norm": 8.491381645202637,
    "prediction_norm": 7.9757890701293945,
    "timestamp": "2025-07-14T12:34:58.983302"
  },
  {
    "epoch": 2367,
    "batch_idx": 96,
    "loss": 0.18355348706245422,
    "learning_rate": 8.080735966828474e-05,
    "timestep_mean": 524.28125,
    "epsilon_norm": 7.558879375457764,
    "prediction_norm": 6.967663288116455,
    "timestamp": "2025-07-14T12:34:58.988555"
  },
  {
    "epoch": 2367,
    "batch_idx": 128,
    "loss": 0.3638421297073364,
    "learning_rate": 8.080735966828474e-05,
    "timestep_mean": 479.3125,
    "epsilon_norm": 8.056992530822754,
    "prediction_norm": 6.863397121429443,
    "timestamp": "2025-07-14T12:34:58.993607"
  },
  {
    "epoch": 2367,
    "batch_idx": 160,
    "loss": 0.3396070599555969,
    "learning_rate": 8.080735966828474e-05,
    "timestep_mean": 489.0,
    "epsilon_norm": 8.215734481811523,
    "prediction_norm": 7.163444519042969,
    "timestamp": "2025-07-14T12:34:58.998807"
  },
  {
    "epoch": 2367,
    "batch_idx": 192,
    "loss": 0.3009742200374603,
    "learning_rate": 8.080735966828474e-05,
    "timestep_mean": 417.25,
    "epsilon_norm": 4.400979518890381,
    "prediction_norm": 3.770738124847412,
    "timestamp": "2025-07-14T12:34:59.003754"
  },
  {
    "epoch": 2368,
    "batch_idx": 0,
    "loss": 0.1290939748287201,
    "learning_rate": 8.123597997892936e-05,
    "timestep_mean": 472.03125,
    "epsilon_norm": 7.150495529174805,
    "prediction_norm": 6.460949897766113,
    "timestamp": "2025-07-14T12:34:59.008829"
  },
  {
    "epoch": 2368,
    "batch_idx": 32,
    "loss": 0.5216185450553894,
    "learning_rate": 8.123597997892936e-05,
    "timestep_mean": 476.46875,
    "epsilon_norm": 9.475135803222656,
    "prediction_norm": 7.3758864402771,
    "timestamp": "2025-07-14T12:34:59.013735"
  },
  {
    "epoch": 2368,
    "batch_idx": 64,
    "loss": 0.15066127479076385,
    "learning_rate": 8.123597997892936e-05,
    "timestep_mean": 569.28125,
    "epsilon_norm": 7.392168998718262,
    "prediction_norm": 6.927484512329102,
    "timestamp": "2025-07-14T12:34:59.018612"
  },
  {
    "epoch": 2368,
    "batch_idx": 96,
    "loss": 0.19112429022789001,
    "learning_rate": 8.123597997892936e-05,
    "timestep_mean": 489.75,
    "epsilon_norm": 7.728661060333252,
    "prediction_norm": 6.496013164520264,
    "timestamp": "2025-07-14T12:34:59.023473"
  },
  {
    "epoch": 2368,
    "batch_idx": 128,
    "loss": 0.25708791613578796,
    "learning_rate": 8.123597997892936e-05,
    "timestep_mean": 644.84375,
    "epsilon_norm": 7.816168308258057,
    "prediction_norm": 6.463644504547119,
    "timestamp": "2025-07-14T12:34:59.028429"
  },
  {
    "epoch": 2368,
    "batch_idx": 160,
    "loss": 0.17449504137039185,
    "learning_rate": 8.123597997892936e-05,
    "timestep_mean": 533.3125,
    "epsilon_norm": 8.835925102233887,
    "prediction_norm": 7.853769779205322,
    "timestamp": "2025-07-14T12:34:59.033282"
  },
  {
    "epoch": 2368,
    "batch_idx": 192,
    "loss": 0.11763256043195724,
    "learning_rate": 8.123597997892936e-05,
    "timestep_mean": 489.875,
    "epsilon_norm": 2.9386048316955566,
    "prediction_norm": 2.6694793701171875,
    "timestamp": "2025-07-14T12:34:59.038487"
  },
  {
    "epoch": 2369,
    "batch_idx": 0,
    "loss": 0.2510939836502075,
    "learning_rate": 8.166563354816528e-05,
    "timestep_mean": 407.34375,
    "epsilon_norm": 7.787644386291504,
    "prediction_norm": 6.694823265075684,
    "timestamp": "2025-07-14T12:34:59.047016"
  },
  {
    "epoch": 2369,
    "batch_idx": 32,
    "loss": 0.2558152377605438,
    "learning_rate": 8.166563354816528e-05,
    "timestep_mean": 481.15625,
    "epsilon_norm": 8.135311126708984,
    "prediction_norm": 7.125664710998535,
    "timestamp": "2025-07-14T12:34:59.052550"
  },
  {
    "epoch": 2369,
    "batch_idx": 64,
    "loss": 0.19623634219169617,
    "learning_rate": 8.166563354816528e-05,
    "timestep_mean": 437.375,
    "epsilon_norm": 7.217682361602783,
    "prediction_norm": 6.69939661026001,
    "timestamp": "2025-07-14T12:34:59.057794"
  },
  {
    "epoch": 2369,
    "batch_idx": 96,
    "loss": 0.18083611130714417,
    "learning_rate": 8.166563354816528e-05,
    "timestep_mean": 499.34375,
    "epsilon_norm": 7.387584686279297,
    "prediction_norm": 6.626479625701904,
    "timestamp": "2025-07-14T12:34:59.063110"
  },
  {
    "epoch": 2369,
    "batch_idx": 128,
    "loss": 0.21259859204292297,
    "learning_rate": 8.166563354816528e-05,
    "timestep_mean": 519.34375,
    "epsilon_norm": 8.1701021194458,
    "prediction_norm": 7.273044109344482,
    "timestamp": "2025-07-14T12:34:59.068296"
  },
  {
    "epoch": 2369,
    "batch_idx": 160,
    "loss": 0.1046789139509201,
    "learning_rate": 8.166563354816528e-05,
    "timestep_mean": 596.15625,
    "epsilon_norm": 6.239630699157715,
    "prediction_norm": 5.957879066467285,
    "timestamp": "2025-07-14T12:34:59.073481"
  },
  {
    "epoch": 2369,
    "batch_idx": 192,
    "loss": 0.13942116498947144,
    "learning_rate": 8.166563354816528e-05,
    "timestep_mean": 634.25,
    "epsilon_norm": 5.019131183624268,
    "prediction_norm": 4.783416271209717,
    "timestamp": "2025-07-14T12:34:59.078773"
  },
  {
    "epoch": 2370,
    "batch_idx": 0,
    "loss": 0.28216010332107544,
    "learning_rate": 8.2096319315865e-05,
    "timestep_mean": 359.375,
    "epsilon_norm": 7.589948654174805,
    "prediction_norm": 6.416508674621582,
    "timestamp": "2025-07-14T12:34:59.084039"
  },
  {
    "epoch": 2370,
    "batch_idx": 32,
    "loss": 0.2289191633462906,
    "learning_rate": 8.2096319315865e-05,
    "timestep_mean": 464.5,
    "epsilon_norm": 6.958115100860596,
    "prediction_norm": 6.2039666175842285,
    "timestamp": "2025-07-14T12:34:59.089121"
  },
  {
    "epoch": 2370,
    "batch_idx": 64,
    "loss": 0.17798474431037903,
    "learning_rate": 8.2096319315865e-05,
    "timestep_mean": 513.125,
    "epsilon_norm": 6.604887008666992,
    "prediction_norm": 6.210693836212158,
    "timestamp": "2025-07-14T12:34:59.094263"
  },
  {
    "epoch": 2370,
    "batch_idx": 96,
    "loss": 0.3342112898826599,
    "learning_rate": 8.2096319315865e-05,
    "timestep_mean": 528.84375,
    "epsilon_norm": 7.990871429443359,
    "prediction_norm": 6.5362749099731445,
    "timestamp": "2025-07-14T12:34:59.099419"
  },
  {
    "epoch": 2370,
    "batch_idx": 128,
    "loss": 0.22240044176578522,
    "learning_rate": 8.2096319315865e-05,
    "timestep_mean": 484.875,
    "epsilon_norm": 7.881281852722168,
    "prediction_norm": 6.620209693908691,
    "timestamp": "2025-07-14T12:34:59.104657"
  },
  {
    "epoch": 2370,
    "batch_idx": 160,
    "loss": 0.18623459339141846,
    "learning_rate": 8.2096319315865e-05,
    "timestep_mean": 531.21875,
    "epsilon_norm": 6.995840549468994,
    "prediction_norm": 6.505038261413574,
    "timestamp": "2025-07-14T12:34:59.109712"
  },
  {
    "epoch": 2370,
    "batch_idx": 192,
    "loss": 0.10570130497217178,
    "learning_rate": 8.2096319315865e-05,
    "timestep_mean": 345.25,
    "epsilon_norm": 3.570509195327759,
    "prediction_norm": 3.294731378555298,
    "timestamp": "2025-07-14T12:34:59.114828"
  },
  {
    "epoch": 2371,
    "batch_idx": 0,
    "loss": 0.42986583709716797,
    "learning_rate": 8.252803621935443e-05,
    "timestep_mean": 467.21875,
    "epsilon_norm": 8.55772590637207,
    "prediction_norm": 6.998203277587891,
    "timestamp": "2025-07-14T12:34:59.120061"
  },
  {
    "epoch": 2371,
    "batch_idx": 32,
    "loss": 0.27496010065078735,
    "learning_rate": 8.252803621935443e-05,
    "timestep_mean": 494.53125,
    "epsilon_norm": 7.804380416870117,
    "prediction_norm": 6.2197136878967285,
    "timestamp": "2025-07-14T12:34:59.125325"
  },
  {
    "epoch": 2371,
    "batch_idx": 64,
    "loss": 0.28013187646865845,
    "learning_rate": 8.252803621935443e-05,
    "timestep_mean": 452.875,
    "epsilon_norm": 7.837759494781494,
    "prediction_norm": 6.811629295349121,
    "timestamp": "2025-07-14T12:34:59.130563"
  },
  {
    "epoch": 2371,
    "batch_idx": 96,
    "loss": 0.348503977060318,
    "learning_rate": 8.252803621935443e-05,
    "timestep_mean": 404.96875,
    "epsilon_norm": 9.19572925567627,
    "prediction_norm": 7.39091157913208,
    "timestamp": "2025-07-14T12:34:59.135668"
  },
  {
    "epoch": 2371,
    "batch_idx": 128,
    "loss": 0.269893616437912,
    "learning_rate": 8.252803621935443e-05,
    "timestep_mean": 397.96875,
    "epsilon_norm": 7.342459201812744,
    "prediction_norm": 6.038568496704102,
    "timestamp": "2025-07-14T12:34:59.140752"
  },
  {
    "epoch": 2371,
    "batch_idx": 160,
    "loss": 0.19817547500133514,
    "learning_rate": 8.252803621935443e-05,
    "timestep_mean": 522.96875,
    "epsilon_norm": 7.75047492980957,
    "prediction_norm": 6.849184513092041,
    "timestamp": "2025-07-14T12:34:59.146013"
  },
  {
    "epoch": 2371,
    "batch_idx": 192,
    "loss": 0.1058184951543808,
    "learning_rate": 8.252803621935443e-05,
    "timestep_mean": 354.625,
    "epsilon_norm": 3.4922397136688232,
    "prediction_norm": 3.749642848968506,
    "timestamp": "2025-07-14T12:34:59.151151"
  },
  {
    "epoch": 2372,
    "batch_idx": 0,
    "loss": 0.2720367908477783,
    "learning_rate": 8.296078319341463e-05,
    "timestep_mean": 548.71875,
    "epsilon_norm": 9.151565551757812,
    "prediction_norm": 8.222357749938965,
    "timestamp": "2025-07-14T12:34:59.156304"
  },
  {
    "epoch": 2372,
    "batch_idx": 32,
    "loss": 0.2545775771141052,
    "learning_rate": 8.296078319341463e-05,
    "timestep_mean": 456.09375,
    "epsilon_norm": 6.7147064208984375,
    "prediction_norm": 4.618917465209961,
    "timestamp": "2025-07-14T12:34:59.161375"
  },
  {
    "epoch": 2372,
    "batch_idx": 64,
    "loss": 0.24946022033691406,
    "learning_rate": 8.296078319341463e-05,
    "timestep_mean": 466.90625,
    "epsilon_norm": 8.2108793258667,
    "prediction_norm": 6.848982334136963,
    "timestamp": "2025-07-14T12:34:59.166610"
  },
  {
    "epoch": 2372,
    "batch_idx": 96,
    "loss": 0.3692946135997772,
    "learning_rate": 8.296078319341463e-05,
    "timestep_mean": 461.0,
    "epsilon_norm": 8.35029411315918,
    "prediction_norm": 6.788488864898682,
    "timestamp": "2025-07-14T12:34:59.171666"
  },
  {
    "epoch": 2372,
    "batch_idx": 128,
    "loss": 0.2088559865951538,
    "learning_rate": 8.296078319341463e-05,
    "timestep_mean": 484.96875,
    "epsilon_norm": 8.689533233642578,
    "prediction_norm": 7.1375226974487305,
    "timestamp": "2025-07-14T12:34:59.176542"
  },
  {
    "epoch": 2372,
    "batch_idx": 160,
    "loss": 0.49594253301620483,
    "learning_rate": 8.296078319341463e-05,
    "timestep_mean": 460.90625,
    "epsilon_norm": 8.86892032623291,
    "prediction_norm": 6.787295341491699,
    "timestamp": "2025-07-14T12:34:59.181502"
  },
  {
    "epoch": 2372,
    "batch_idx": 192,
    "loss": 0.5615200996398926,
    "learning_rate": 8.296078319341463e-05,
    "timestep_mean": 443.0,
    "epsilon_norm": 4.75324010848999,
    "prediction_norm": 3.0290191173553467,
    "timestamp": "2025-07-14T12:34:59.186400"
  },
  {
    "epoch": 2373,
    "batch_idx": 0,
    "loss": 0.37147900462150574,
    "learning_rate": 8.339455917028567e-05,
    "timestep_mean": 462.15625,
    "epsilon_norm": 8.691527366638184,
    "prediction_norm": 7.164719581604004,
    "timestamp": "2025-07-14T12:34:59.191340"
  },
  {
    "epoch": 2373,
    "batch_idx": 32,
    "loss": 0.22994808852672577,
    "learning_rate": 8.339455917028567e-05,
    "timestep_mean": 519.9375,
    "epsilon_norm": 7.484551429748535,
    "prediction_norm": 6.910852432250977,
    "timestamp": "2025-07-14T12:34:59.196432"
  },
  {
    "epoch": 2373,
    "batch_idx": 64,
    "loss": 0.18930160999298096,
    "learning_rate": 8.339455917028567e-05,
    "timestep_mean": 511.5625,
    "epsilon_norm": 8.294885635375977,
    "prediction_norm": 7.550780296325684,
    "timestamp": "2025-07-14T12:34:59.201465"
  },
  {
    "epoch": 2373,
    "batch_idx": 96,
    "loss": 0.2983759045600891,
    "learning_rate": 8.339455917028567e-05,
    "timestep_mean": 425.75,
    "epsilon_norm": 8.057979583740234,
    "prediction_norm": 6.725199222564697,
    "timestamp": "2025-07-14T12:34:59.206570"
  },
  {
    "epoch": 2373,
    "batch_idx": 128,
    "loss": 0.31063902378082275,
    "learning_rate": 8.339455917028567e-05,
    "timestep_mean": 500.53125,
    "epsilon_norm": 9.947676658630371,
    "prediction_norm": 8.51491928100586,
    "timestamp": "2025-07-14T12:34:59.211776"
  },
  {
    "epoch": 2373,
    "batch_idx": 160,
    "loss": 0.31690603494644165,
    "learning_rate": 8.339455917028567e-05,
    "timestep_mean": 523.96875,
    "epsilon_norm": 7.952665328979492,
    "prediction_norm": 6.405287265777588,
    "timestamp": "2025-07-14T12:34:59.216973"
  },
  {
    "epoch": 2373,
    "batch_idx": 192,
    "loss": 0.08480872958898544,
    "learning_rate": 8.339455917028567e-05,
    "timestep_mean": 681.125,
    "epsilon_norm": 4.846102237701416,
    "prediction_norm": 4.632143974304199,
    "timestamp": "2025-07-14T12:34:59.222017"
  },
  {
    "epoch": 2374,
    "batch_idx": 0,
    "loss": 0.17148813605308533,
    "learning_rate": 8.382936307966856e-05,
    "timestep_mean": 543.5,
    "epsilon_norm": 6.242519855499268,
    "prediction_norm": 4.601989269256592,
    "timestamp": "2025-07-14T12:34:59.227104"
  },
  {
    "epoch": 2374,
    "batch_idx": 32,
    "loss": 0.31126028299331665,
    "learning_rate": 8.382936307966856e-05,
    "timestep_mean": 492.125,
    "epsilon_norm": 7.048488140106201,
    "prediction_norm": 5.667203903198242,
    "timestamp": "2025-07-14T12:34:59.231961"
  },
  {
    "epoch": 2374,
    "batch_idx": 64,
    "loss": 0.26572126150131226,
    "learning_rate": 8.382936307966856e-05,
    "timestep_mean": 530.0625,
    "epsilon_norm": 9.473857879638672,
    "prediction_norm": 8.579809188842773,
    "timestamp": "2025-07-14T12:34:59.236865"
  },
  {
    "epoch": 2374,
    "batch_idx": 96,
    "loss": 0.16084808111190796,
    "learning_rate": 8.382936307966856e-05,
    "timestep_mean": 489.3125,
    "epsilon_norm": 7.258864879608154,
    "prediction_norm": 7.4918928146362305,
    "timestamp": "2025-07-14T12:34:59.241735"
  },
  {
    "epoch": 2374,
    "batch_idx": 128,
    "loss": 0.25867486000061035,
    "learning_rate": 8.382936307966856e-05,
    "timestep_mean": 557.15625,
    "epsilon_norm": 8.169147491455078,
    "prediction_norm": 6.964104175567627,
    "timestamp": "2025-07-14T12:34:59.246896"
  },
  {
    "epoch": 2374,
    "batch_idx": 160,
    "loss": 0.22103866934776306,
    "learning_rate": 8.382936307966856e-05,
    "timestep_mean": 466.84375,
    "epsilon_norm": 7.0012383460998535,
    "prediction_norm": 6.369564533233643,
    "timestamp": "2025-07-14T12:34:59.251783"
  },
  {
    "epoch": 2374,
    "batch_idx": 192,
    "loss": 0.25704097747802734,
    "learning_rate": 8.382936307966856e-05,
    "timestep_mean": 444.125,
    "epsilon_norm": 4.979886531829834,
    "prediction_norm": 4.3754191398620605,
    "timestamp": "2025-07-14T12:34:59.256642"
  },
  {
    "epoch": 2375,
    "batch_idx": 0,
    "loss": 0.3478740453720093,
    "learning_rate": 8.426519384872751e-05,
    "timestep_mean": 403.5,
    "epsilon_norm": 8.822572708129883,
    "prediction_norm": 6.7028679847717285,
    "timestamp": "2025-07-14T12:34:59.261551"
  },
  {
    "epoch": 2375,
    "batch_idx": 32,
    "loss": 0.29314255714416504,
    "learning_rate": 8.426519384872751e-05,
    "timestep_mean": 442.78125,
    "epsilon_norm": 7.932023048400879,
    "prediction_norm": 6.456984996795654,
    "timestamp": "2025-07-14T12:34:59.266384"
  },
  {
    "epoch": 2375,
    "batch_idx": 64,
    "loss": 0.3624971807003021,
    "learning_rate": 8.426519384872751e-05,
    "timestep_mean": 444.40625,
    "epsilon_norm": 8.253486633300781,
    "prediction_norm": 7.192625045776367,
    "timestamp": "2025-07-14T12:34:59.271426"
  },
  {
    "epoch": 2375,
    "batch_idx": 96,
    "loss": 0.30698493123054504,
    "learning_rate": 8.426519384872751e-05,
    "timestep_mean": 479.28125,
    "epsilon_norm": 8.579573631286621,
    "prediction_norm": 7.378668785095215,
    "timestamp": "2025-07-14T12:34:59.276341"
  },
  {
    "epoch": 2375,
    "batch_idx": 128,
    "loss": 0.2573508024215698,
    "learning_rate": 8.426519384872751e-05,
    "timestep_mean": 426.0,
    "epsilon_norm": 7.445715427398682,
    "prediction_norm": 6.324853897094727,
    "timestamp": "2025-07-14T12:34:59.281242"
  },
  {
    "epoch": 2375,
    "batch_idx": 160,
    "loss": 0.33027273416519165,
    "learning_rate": 8.426519384872751e-05,
    "timestep_mean": 471.03125,
    "epsilon_norm": 8.739507675170898,
    "prediction_norm": 6.726714611053467,
    "timestamp": "2025-07-14T12:34:59.286339"
  },
  {
    "epoch": 2375,
    "batch_idx": 192,
    "loss": 0.33412158489227295,
    "learning_rate": 8.426519384872751e-05,
    "timestep_mean": 478.125,
    "epsilon_norm": 3.47428035736084,
    "prediction_norm": 2.5780832767486572,
    "timestamp": "2025-07-14T12:34:59.291267"
  },
  {
    "epoch": 2376,
    "batch_idx": 0,
    "loss": 0.33396703004837036,
    "learning_rate": 8.47020504020939e-05,
    "timestep_mean": 412.3125,
    "epsilon_norm": 8.267247200012207,
    "prediction_norm": 6.592773914337158,
    "timestamp": "2025-07-14T12:34:59.296360"
  },
  {
    "epoch": 2376,
    "batch_idx": 32,
    "loss": 0.06875438988208771,
    "learning_rate": 8.47020504020939e-05,
    "timestep_mean": 534.09375,
    "epsilon_norm": 8.179062843322754,
    "prediction_norm": 7.556854248046875,
    "timestamp": "2025-07-14T12:34:59.301399"
  },
  {
    "epoch": 2376,
    "batch_idx": 64,
    "loss": 0.3039657771587372,
    "learning_rate": 8.47020504020939e-05,
    "timestep_mean": 572.09375,
    "epsilon_norm": 8.341418266296387,
    "prediction_norm": 7.545889377593994,
    "timestamp": "2025-07-14T12:34:59.306567"
  },
  {
    "epoch": 2376,
    "batch_idx": 96,
    "loss": 0.3702332079410553,
    "learning_rate": 8.47020504020939e-05,
    "timestep_mean": 463.0,
    "epsilon_norm": 7.651941776275635,
    "prediction_norm": 6.585123538970947,
    "timestamp": "2025-07-14T12:34:59.311734"
  },
  {
    "epoch": 2376,
    "batch_idx": 128,
    "loss": 0.324445903301239,
    "learning_rate": 8.47020504020939e-05,
    "timestep_mean": 425.40625,
    "epsilon_norm": 8.580079078674316,
    "prediction_norm": 7.6919169425964355,
    "timestamp": "2025-07-14T12:34:59.316964"
  },
  {
    "epoch": 2376,
    "batch_idx": 160,
    "loss": 0.15751081705093384,
    "learning_rate": 8.47020504020939e-05,
    "timestep_mean": 505.25,
    "epsilon_norm": 7.238541126251221,
    "prediction_norm": 6.997550964355469,
    "timestamp": "2025-07-14T12:34:59.321951"
  },
  {
    "epoch": 2376,
    "batch_idx": 192,
    "loss": 0.2079392820596695,
    "learning_rate": 8.47020504020939e-05,
    "timestep_mean": 470.75,
    "epsilon_norm": 4.201381206512451,
    "prediction_norm": 3.6051182746887207,
    "timestamp": "2025-07-14T12:34:59.326911"
  },
  {
    "epoch": 2377,
    "batch_idx": 0,
    "loss": 0.23974597454071045,
    "learning_rate": 8.513993166186713e-05,
    "timestep_mean": 585.96875,
    "epsilon_norm": 8.37971305847168,
    "prediction_norm": 7.997890472412109,
    "timestamp": "2025-07-14T12:34:59.332147"
  },
  {
    "epoch": 2377,
    "batch_idx": 32,
    "loss": 0.37161940336227417,
    "learning_rate": 8.513993166186713e-05,
    "timestep_mean": 457.9375,
    "epsilon_norm": 8.608221054077148,
    "prediction_norm": 7.357931613922119,
    "timestamp": "2025-07-14T12:34:59.337329"
  },
  {
    "epoch": 2377,
    "batch_idx": 64,
    "loss": 0.36683881282806396,
    "learning_rate": 8.513993166186713e-05,
    "timestep_mean": 496.5,
    "epsilon_norm": 7.392222881317139,
    "prediction_norm": 6.0890889167785645,
    "timestamp": "2025-07-14T12:34:59.342472"
  },
  {
    "epoch": 2377,
    "batch_idx": 96,
    "loss": 0.3227965831756592,
    "learning_rate": 8.513993166186713e-05,
    "timestep_mean": 505.375,
    "epsilon_norm": 8.073009490966797,
    "prediction_norm": 7.088560581207275,
    "timestamp": "2025-07-14T12:34:59.347687"
  },
  {
    "epoch": 2377,
    "batch_idx": 128,
    "loss": 0.2387627512216568,
    "learning_rate": 8.513993166186713e-05,
    "timestep_mean": 573.9375,
    "epsilon_norm": 7.579224586486816,
    "prediction_norm": 6.9214701652526855,
    "timestamp": "2025-07-14T12:34:59.352908"
  },
  {
    "epoch": 2377,
    "batch_idx": 160,
    "loss": 0.23821240663528442,
    "learning_rate": 8.513993166186713e-05,
    "timestep_mean": 546.375,
    "epsilon_norm": 7.526379585266113,
    "prediction_norm": 6.294125080108643,
    "timestamp": "2025-07-14T12:34:59.358110"
  },
  {
    "epoch": 2377,
    "batch_idx": 192,
    "loss": 0.3536105155944824,
    "learning_rate": 8.513993166186713e-05,
    "timestep_mean": 564.25,
    "epsilon_norm": 3.8771026134490967,
    "prediction_norm": 2.7148237228393555,
    "timestamp": "2025-07-14T12:34:59.363351"
  },
  {
    "epoch": 2378,
    "batch_idx": 0,
    "loss": 0.17977550625801086,
    "learning_rate": 8.557883654761925e-05,
    "timestep_mean": 506.0,
    "epsilon_norm": 7.704241752624512,
    "prediction_norm": 7.458139896392822,
    "timestamp": "2025-07-14T12:34:59.368572"
  },
  {
    "epoch": 2378,
    "batch_idx": 32,
    "loss": 0.1979721188545227,
    "learning_rate": 8.557883654761925e-05,
    "timestep_mean": 453.71875,
    "epsilon_norm": 7.871231555938721,
    "prediction_norm": 7.2261247634887695,
    "timestamp": "2025-07-14T12:34:59.373763"
  },
  {
    "epoch": 2378,
    "batch_idx": 64,
    "loss": 0.3082878887653351,
    "learning_rate": 8.557883654761925e-05,
    "timestep_mean": 500.625,
    "epsilon_norm": 7.618923187255859,
    "prediction_norm": 6.274471759796143,
    "timestamp": "2025-07-14T12:34:59.379069"
  },
  {
    "epoch": 2378,
    "batch_idx": 96,
    "loss": 0.46000075340270996,
    "learning_rate": 8.557883654761925e-05,
    "timestep_mean": 543.90625,
    "epsilon_norm": 7.993251323699951,
    "prediction_norm": 6.583596229553223,
    "timestamp": "2025-07-14T12:34:59.384334"
  },
  {
    "epoch": 2378,
    "batch_idx": 128,
    "loss": 0.3535255193710327,
    "learning_rate": 8.557883654761925e-05,
    "timestep_mean": 468.96875,
    "epsilon_norm": 7.907281875610352,
    "prediction_norm": 6.864411354064941,
    "timestamp": "2025-07-14T12:34:59.389512"
  },
  {
    "epoch": 2378,
    "batch_idx": 160,
    "loss": 0.14521929621696472,
    "learning_rate": 8.557883654761925e-05,
    "timestep_mean": 564.3125,
    "epsilon_norm": 6.127816200256348,
    "prediction_norm": 5.990741729736328,
    "timestamp": "2025-07-14T12:34:59.394712"
  },
  {
    "epoch": 2378,
    "batch_idx": 192,
    "loss": 0.269273966550827,
    "learning_rate": 8.557883654761925e-05,
    "timestep_mean": 592.75,
    "epsilon_norm": 3.8080854415893555,
    "prediction_norm": 3.8537182807922363,
    "timestamp": "2025-07-14T12:34:59.399871"
  },
  {
    "epoch": 2379,
    "batch_idx": 0,
    "loss": 0.2795536518096924,
    "learning_rate": 8.601876397639571e-05,
    "timestep_mean": 542.5625,
    "epsilon_norm": 8.15129566192627,
    "prediction_norm": 6.93568229675293,
    "timestamp": "2025-07-14T12:34:59.405119"
  },
  {
    "epoch": 2379,
    "batch_idx": 32,
    "loss": 0.18333855271339417,
    "learning_rate": 8.601876397639571e-05,
    "timestep_mean": 550.0625,
    "epsilon_norm": 7.69091796875,
    "prediction_norm": 7.183048248291016,
    "timestamp": "2025-07-14T12:34:59.409955"
  },
  {
    "epoch": 2379,
    "batch_idx": 64,
    "loss": 0.4373568892478943,
    "learning_rate": 8.601876397639571e-05,
    "timestep_mean": 603.65625,
    "epsilon_norm": 7.8228230476379395,
    "prediction_norm": 6.3818159103393555,
    "timestamp": "2025-07-14T12:34:59.414862"
  },
  {
    "epoch": 2379,
    "batch_idx": 96,
    "loss": 0.5100704431533813,
    "learning_rate": 8.601876397639571e-05,
    "timestep_mean": 476.0,
    "epsilon_norm": 8.899321556091309,
    "prediction_norm": 6.859896183013916,
    "timestamp": "2025-07-14T12:34:59.420131"
  },
  {
    "epoch": 2379,
    "batch_idx": 128,
    "loss": 0.43685227632522583,
    "learning_rate": 8.601876397639571e-05,
    "timestep_mean": 450.8125,
    "epsilon_norm": 8.214682579040527,
    "prediction_norm": 6.368797302246094,
    "timestamp": "2025-07-14T12:34:59.425311"
  },
  {
    "epoch": 2379,
    "batch_idx": 160,
    "loss": 0.3926326334476471,
    "learning_rate": 8.601876397639571e-05,
    "timestep_mean": 521.4375,
    "epsilon_norm": 8.94177532196045,
    "prediction_norm": 7.6093363761901855,
    "timestamp": "2025-07-14T12:34:59.430502"
  },
  {
    "epoch": 2379,
    "batch_idx": 192,
    "loss": 1.020910382270813,
    "learning_rate": 8.601876397639571e-05,
    "timestep_mean": 415.0,
    "epsilon_norm": 4.683910369873047,
    "prediction_norm": 3.533245325088501,
    "timestamp": "2025-07-14T12:34:59.435636"
  },
  {
    "epoch": 2380,
    "batch_idx": 0,
    "loss": 0.4402579665184021,
    "learning_rate": 8.645971286271923e-05,
    "timestep_mean": 468.4375,
    "epsilon_norm": 9.378101348876953,
    "prediction_norm": 7.866067886352539,
    "timestamp": "2025-07-14T12:34:59.441104"
  },
  {
    "epoch": 2380,
    "batch_idx": 32,
    "loss": 0.3343866765499115,
    "learning_rate": 8.645971286271923e-05,
    "timestep_mean": 431.375,
    "epsilon_norm": 8.735370635986328,
    "prediction_norm": 6.946041107177734,
    "timestamp": "2025-07-14T12:34:59.446290"
  },
  {
    "epoch": 2380,
    "batch_idx": 64,
    "loss": 0.19131632149219513,
    "learning_rate": 8.645971286271923e-05,
    "timestep_mean": 574.375,
    "epsilon_norm": 7.851778984069824,
    "prediction_norm": 7.318297386169434,
    "timestamp": "2025-07-14T12:34:59.451323"
  },
  {
    "epoch": 2380,
    "batch_idx": 96,
    "loss": 0.2876605987548828,
    "learning_rate": 8.645971286271923e-05,
    "timestep_mean": 513.21875,
    "epsilon_norm": 9.245539665222168,
    "prediction_norm": 7.942542552947998,
    "timestamp": "2025-07-14T12:34:59.456688"
  },
  {
    "epoch": 2380,
    "batch_idx": 128,
    "loss": 0.27054494619369507,
    "learning_rate": 8.645971286271923e-05,
    "timestep_mean": 446.15625,
    "epsilon_norm": 6.846044063568115,
    "prediction_norm": 5.8349738121032715,
    "timestamp": "2025-07-14T12:34:59.461945"
  },
  {
    "epoch": 2380,
    "batch_idx": 160,
    "loss": 0.3151155710220337,
    "learning_rate": 8.645971286271923e-05,
    "timestep_mean": 447.34375,
    "epsilon_norm": 7.992830276489258,
    "prediction_norm": 6.541781425476074,
    "timestamp": "2025-07-14T12:34:59.467183"
  },
  {
    "epoch": 2380,
    "batch_idx": 192,
    "loss": 0.18695621192455292,
    "learning_rate": 8.645971286271923e-05,
    "timestep_mean": 499.875,
    "epsilon_norm": 4.078987121582031,
    "prediction_norm": 3.66648530960083,
    "timestamp": "2025-07-14T12:34:59.472439"
  },
  {
    "epoch": 2381,
    "batch_idx": 0,
    "loss": 0.31265783309936523,
    "learning_rate": 8.690168211859265e-05,
    "timestep_mean": 483.65625,
    "epsilon_norm": 9.057414054870605,
    "prediction_norm": 7.659865856170654,
    "timestamp": "2025-07-14T12:34:59.477626"
  },
  {
    "epoch": 2381,
    "batch_idx": 32,
    "loss": 0.2903836965560913,
    "learning_rate": 8.690168211859265e-05,
    "timestep_mean": 453.375,
    "epsilon_norm": 8.712005615234375,
    "prediction_norm": 7.263728141784668,
    "timestamp": "2025-07-14T12:34:59.482861"
  },
  {
    "epoch": 2381,
    "batch_idx": 64,
    "loss": 0.2654973566532135,
    "learning_rate": 8.690168211859265e-05,
    "timestep_mean": 592.78125,
    "epsilon_norm": 7.829588413238525,
    "prediction_norm": 6.62098503112793,
    "timestamp": "2025-07-14T12:34:59.488075"
  },
  {
    "epoch": 2381,
    "batch_idx": 96,
    "loss": 0.12749256193637848,
    "learning_rate": 8.690168211859265e-05,
    "timestep_mean": 601.96875,
    "epsilon_norm": 7.585014820098877,
    "prediction_norm": 6.623146057128906,
    "timestamp": "2025-07-14T12:34:59.493225"
  },
  {
    "epoch": 2381,
    "batch_idx": 128,
    "loss": 0.1697300225496292,
    "learning_rate": 8.690168211859265e-05,
    "timestep_mean": 589.625,
    "epsilon_norm": 8.120800971984863,
    "prediction_norm": 7.278961658477783,
    "timestamp": "2025-07-14T12:34:59.498458"
  },
  {
    "epoch": 2381,
    "batch_idx": 160,
    "loss": 0.4013943672180176,
    "learning_rate": 8.690168211859265e-05,
    "timestep_mean": 499.1875,
    "epsilon_norm": 8.944657325744629,
    "prediction_norm": 7.683764457702637,
    "timestamp": "2025-07-14T12:34:59.503560"
  },
  {
    "epoch": 2381,
    "batch_idx": 192,
    "loss": 0.3165929317474365,
    "learning_rate": 8.690168211859265e-05,
    "timestep_mean": 544.0,
    "epsilon_norm": 3.54494047164917,
    "prediction_norm": 2.394244432449341,
    "timestamp": "2025-07-14T12:34:59.508656"
  },
  {
    "epoch": 2382,
    "batch_idx": 0,
    "loss": 0.4124433994293213,
    "learning_rate": 8.734467065350042e-05,
    "timestep_mean": 377.0625,
    "epsilon_norm": 7.586155414581299,
    "prediction_norm": 5.494206428527832,
    "timestamp": "2025-07-14T12:34:59.513808"
  },
  {
    "epoch": 2382,
    "batch_idx": 32,
    "loss": 0.2230398952960968,
    "learning_rate": 8.734467065350042e-05,
    "timestep_mean": 490.34375,
    "epsilon_norm": 8.349485397338867,
    "prediction_norm": 7.740927219390869,
    "timestamp": "2025-07-14T12:34:59.518934"
  },
  {
    "epoch": 2382,
    "batch_idx": 64,
    "loss": 0.4518193006515503,
    "learning_rate": 8.734467065350042e-05,
    "timestep_mean": 451.96875,
    "epsilon_norm": 8.41973876953125,
    "prediction_norm": 6.681028842926025,
    "timestamp": "2025-07-14T12:34:59.524293"
  },
  {
    "epoch": 2382,
    "batch_idx": 96,
    "loss": 0.29649001359939575,
    "learning_rate": 8.734467065350042e-05,
    "timestep_mean": 497.78125,
    "epsilon_norm": 8.184505462646484,
    "prediction_norm": 7.592154502868652,
    "timestamp": "2025-07-14T12:34:59.529445"
  },
  {
    "epoch": 2382,
    "batch_idx": 128,
    "loss": 0.2875494360923767,
    "learning_rate": 8.734467065350042e-05,
    "timestep_mean": 472.21875,
    "epsilon_norm": 8.429878234863281,
    "prediction_norm": 7.334597587585449,
    "timestamp": "2025-07-14T12:34:59.534725"
  },
  {
    "epoch": 2382,
    "batch_idx": 160,
    "loss": 0.20333519577980042,
    "learning_rate": 8.734467065350042e-05,
    "timestep_mean": 467.90625,
    "epsilon_norm": 7.56797981262207,
    "prediction_norm": 6.970881462097168,
    "timestamp": "2025-07-14T12:34:59.539874"
  },
  {
    "epoch": 2382,
    "batch_idx": 192,
    "loss": 0.06214156001806259,
    "learning_rate": 8.734467065350042e-05,
    "timestep_mean": 588.75,
    "epsilon_norm": 3.275557279586792,
    "prediction_norm": 3.3072361946105957,
    "timestamp": "2025-07-14T12:34:59.545099"
  },
  {
    "epoch": 2383,
    "batch_idx": 0,
    "loss": 0.4345071315765381,
    "learning_rate": 8.778867737441252e-05,
    "timestep_mean": 450.25,
    "epsilon_norm": 8.770218849182129,
    "prediction_norm": 7.4415812492370605,
    "timestamp": "2025-07-14T12:34:59.550445"
  },
  {
    "epoch": 2383,
    "batch_idx": 32,
    "loss": 0.3757393956184387,
    "learning_rate": 8.778867737441252e-05,
    "timestep_mean": 432.25,
    "epsilon_norm": 7.870388507843018,
    "prediction_norm": 5.705860614776611,
    "timestamp": "2025-07-14T12:34:59.555728"
  },
  {
    "epoch": 2383,
    "batch_idx": 64,
    "loss": 0.5507671236991882,
    "learning_rate": 8.778867737441252e-05,
    "timestep_mean": 482.46875,
    "epsilon_norm": 8.223687171936035,
    "prediction_norm": 5.654468536376953,
    "timestamp": "2025-07-14T12:34:59.560969"
  },
  {
    "epoch": 2383,
    "batch_idx": 96,
    "loss": 0.20107552409172058,
    "learning_rate": 8.778867737441252e-05,
    "timestep_mean": 475.65625,
    "epsilon_norm": 8.330632209777832,
    "prediction_norm": 7.389891147613525,
    "timestamp": "2025-07-14T12:34:59.566127"
  },
  {
    "epoch": 2383,
    "batch_idx": 128,
    "loss": 0.1983865201473236,
    "learning_rate": 8.778867737441252e-05,
    "timestep_mean": 494.34375,
    "epsilon_norm": 7.138798713684082,
    "prediction_norm": 5.975395202636719,
    "timestamp": "2025-07-14T12:34:59.571405"
  },
  {
    "epoch": 2383,
    "batch_idx": 160,
    "loss": 0.28967729210853577,
    "learning_rate": 8.778867737441252e-05,
    "timestep_mean": 506.21875,
    "epsilon_norm": 7.623650550842285,
    "prediction_norm": 6.642980575561523,
    "timestamp": "2025-07-14T12:34:59.576505"
  },
  {
    "epoch": 2383,
    "batch_idx": 192,
    "loss": 0.3084498345851898,
    "learning_rate": 8.778867737441252e-05,
    "timestep_mean": 380.375,
    "epsilon_norm": 3.205707311630249,
    "prediction_norm": 2.188695192337036,
    "timestamp": "2025-07-14T12:34:59.581666"
  },
  {
    "epoch": 2384,
    "batch_idx": 0,
    "loss": 0.26569852232933044,
    "learning_rate": 8.823370118578658e-05,
    "timestep_mean": 614.625,
    "epsilon_norm": 9.088409423828125,
    "prediction_norm": 7.8944926261901855,
    "timestamp": "2025-07-14T12:34:59.586896"
  },
  {
    "epoch": 2384,
    "batch_idx": 32,
    "loss": 0.11120212078094482,
    "learning_rate": 8.823370118578658e-05,
    "timestep_mean": 523.8125,
    "epsilon_norm": 6.718257904052734,
    "prediction_norm": 6.278723239898682,
    "timestamp": "2025-07-14T12:34:59.592194"
  },
  {
    "epoch": 2384,
    "batch_idx": 64,
    "loss": 0.3819153308868408,
    "learning_rate": 8.823370118578658e-05,
    "timestep_mean": 472.4375,
    "epsilon_norm": 9.033886909484863,
    "prediction_norm": 7.32882022857666,
    "timestamp": "2025-07-14T12:34:59.597300"
  },
  {
    "epoch": 2384,
    "batch_idx": 96,
    "loss": 0.16054809093475342,
    "learning_rate": 8.823370118578658e-05,
    "timestep_mean": 551.3125,
    "epsilon_norm": 7.405679225921631,
    "prediction_norm": 6.607224941253662,
    "timestamp": "2025-07-14T12:34:59.602467"
  },
  {
    "epoch": 2384,
    "batch_idx": 128,
    "loss": 0.28865063190460205,
    "learning_rate": 8.823370118578658e-05,
    "timestep_mean": 499.34375,
    "epsilon_norm": 8.86595344543457,
    "prediction_norm": 7.74613094329834,
    "timestamp": "2025-07-14T12:34:59.607683"
  },
  {
    "epoch": 2384,
    "batch_idx": 160,
    "loss": 0.243470698595047,
    "learning_rate": 8.823370118578658e-05,
    "timestep_mean": 578.28125,
    "epsilon_norm": 8.527664184570312,
    "prediction_norm": 7.70947790145874,
    "timestamp": "2025-07-14T12:34:59.612807"
  },
  {
    "epoch": 2384,
    "batch_idx": 192,
    "loss": 0.3322330117225647,
    "learning_rate": 8.823370118578658e-05,
    "timestep_mean": 503.625,
    "epsilon_norm": 3.390611171722412,
    "prediction_norm": 2.6470742225646973,
    "timestamp": "2025-07-14T12:34:59.618071"
  },
  {
    "epoch": 2385,
    "batch_idx": 0,
    "loss": 0.2986152172088623,
    "learning_rate": 8.867974098957034e-05,
    "timestep_mean": 495.71875,
    "epsilon_norm": 8.369446754455566,
    "prediction_norm": 6.783434867858887,
    "timestamp": "2025-07-14T12:34:59.623258"
  },
  {
    "epoch": 2385,
    "batch_idx": 32,
    "loss": 0.19458355009555817,
    "learning_rate": 8.867974098957034e-05,
    "timestep_mean": 593.53125,
    "epsilon_norm": 6.7583770751953125,
    "prediction_norm": 5.911473274230957,
    "timestamp": "2025-07-14T12:34:59.628509"
  },
  {
    "epoch": 2385,
    "batch_idx": 64,
    "loss": 0.2214096486568451,
    "learning_rate": 8.867974098957034e-05,
    "timestep_mean": 467.625,
    "epsilon_norm": 7.873051166534424,
    "prediction_norm": 6.9223246574401855,
    "timestamp": "2025-07-14T12:34:59.633896"
  },
  {
    "epoch": 2385,
    "batch_idx": 96,
    "loss": 0.34392544627189636,
    "learning_rate": 8.867974098957034e-05,
    "timestep_mean": 535.40625,
    "epsilon_norm": 8.013049125671387,
    "prediction_norm": 6.523928642272949,
    "timestamp": "2025-07-14T12:34:59.639105"
  },
  {
    "epoch": 2385,
    "batch_idx": 128,
    "loss": 0.22084996104240417,
    "learning_rate": 8.867974098957034e-05,
    "timestep_mean": 450.78125,
    "epsilon_norm": 7.547440052032471,
    "prediction_norm": 6.336249351501465,
    "timestamp": "2025-07-14T12:34:59.644204"
  },
  {
    "epoch": 2385,
    "batch_idx": 160,
    "loss": 0.15119357407093048,
    "learning_rate": 8.867974098957034e-05,
    "timestep_mean": 471.4375,
    "epsilon_norm": 7.1199774742126465,
    "prediction_norm": 6.596835613250732,
    "timestamp": "2025-07-14T12:34:59.649453"
  },
  {
    "epoch": 2385,
    "batch_idx": 192,
    "loss": 0.14335376024246216,
    "learning_rate": 8.867974098957034e-05,
    "timestep_mean": 453.375,
    "epsilon_norm": 3.201162099838257,
    "prediction_norm": 2.578760862350464,
    "timestamp": "2025-07-14T12:34:59.654665"
  },
  {
    "epoch": 2386,
    "batch_idx": 0,
    "loss": 0.24817299842834473,
    "learning_rate": 8.912679568520513e-05,
    "timestep_mean": 397.21875,
    "epsilon_norm": 7.106723308563232,
    "prediction_norm": 5.418825149536133,
    "timestamp": "2025-07-14T12:34:59.659828"
  },
  {
    "epoch": 2386,
    "batch_idx": 32,
    "loss": 0.31690606474876404,
    "learning_rate": 8.912679568520513e-05,
    "timestep_mean": 492.65625,
    "epsilon_norm": 6.962954521179199,
    "prediction_norm": 5.914402008056641,
    "timestamp": "2025-07-14T12:34:59.665157"
  },
  {
    "epoch": 2386,
    "batch_idx": 64,
    "loss": 0.22179658710956573,
    "learning_rate": 8.912679568520513e-05,
    "timestep_mean": 598.75,
    "epsilon_norm": 8.399781227111816,
    "prediction_norm": 6.819432258605957,
    "timestamp": "2025-07-14T12:34:59.670321"
  },
  {
    "epoch": 2386,
    "batch_idx": 96,
    "loss": 0.21044212579727173,
    "learning_rate": 8.912679568520513e-05,
    "timestep_mean": 477.21875,
    "epsilon_norm": 7.1745285987854,
    "prediction_norm": 5.98386287689209,
    "timestamp": "2025-07-14T12:34:59.675509"
  },
  {
    "epoch": 2386,
    "batch_idx": 128,
    "loss": 0.2630760371685028,
    "learning_rate": 8.912679568520513e-05,
    "timestep_mean": 441.8125,
    "epsilon_norm": 7.097589015960693,
    "prediction_norm": 6.061302661895752,
    "timestamp": "2025-07-14T12:34:59.680752"
  },
  {
    "epoch": 2386,
    "batch_idx": 160,
    "loss": 0.47407424449920654,
    "learning_rate": 8.912679568520513e-05,
    "timestep_mean": 472.40625,
    "epsilon_norm": 8.455484390258789,
    "prediction_norm": 6.928846836090088,
    "timestamp": "2025-07-14T12:34:59.685992"
  },
  {
    "epoch": 2386,
    "batch_idx": 192,
    "loss": 0.3950304388999939,
    "learning_rate": 8.912679568520513e-05,
    "timestep_mean": 592.125,
    "epsilon_norm": 4.4300031661987305,
    "prediction_norm": 3.478811502456665,
    "timestamp": "2025-07-14T12:34:59.691141"
  },
  {
    "epoch": 2387,
    "batch_idx": 0,
    "loss": 0.5189442038536072,
    "learning_rate": 8.957486416962783e-05,
    "timestep_mean": 399.53125,
    "epsilon_norm": 8.741748809814453,
    "prediction_norm": 6.22404146194458,
    "timestamp": "2025-07-14T12:34:59.696498"
  },
  {
    "epoch": 2387,
    "batch_idx": 32,
    "loss": 0.24366194009780884,
    "learning_rate": 8.957486416962783e-05,
    "timestep_mean": 446.09375,
    "epsilon_norm": 7.114530563354492,
    "prediction_norm": 6.010345458984375,
    "timestamp": "2025-07-14T12:34:59.701729"
  },
  {
    "epoch": 2387,
    "batch_idx": 64,
    "loss": 0.3012968599796295,
    "learning_rate": 8.957486416962783e-05,
    "timestep_mean": 429.6875,
    "epsilon_norm": 7.514357566833496,
    "prediction_norm": 5.664200305938721,
    "timestamp": "2025-07-14T12:34:59.706959"
  },
  {
    "epoch": 2387,
    "batch_idx": 96,
    "loss": 0.35975831747055054,
    "learning_rate": 8.957486416962783e-05,
    "timestep_mean": 530.65625,
    "epsilon_norm": 9.284873008728027,
    "prediction_norm": 7.832146167755127,
    "timestamp": "2025-07-14T12:34:59.712126"
  },
  {
    "epoch": 2387,
    "batch_idx": 128,
    "loss": 0.25597184896469116,
    "learning_rate": 8.957486416962783e-05,
    "timestep_mean": 483.21875,
    "epsilon_norm": 7.58732795715332,
    "prediction_norm": 6.160111427307129,
    "timestamp": "2025-07-14T12:34:59.717381"
  },
  {
    "epoch": 2387,
    "batch_idx": 160,
    "loss": 0.11961319297552109,
    "learning_rate": 8.957486416962783e-05,
    "timestep_mean": 608.0,
    "epsilon_norm": 7.483068943023682,
    "prediction_norm": 6.956782817840576,
    "timestamp": "2025-07-14T12:34:59.722527"
  },
  {
    "epoch": 2387,
    "batch_idx": 192,
    "loss": 0.22897683084011078,
    "learning_rate": 8.957486416962783e-05,
    "timestep_mean": 498.5,
    "epsilon_norm": 3.2557661533355713,
    "prediction_norm": 2.7397713661193848,
    "timestamp": "2025-07-14T12:34:59.727692"
  },
  {
    "epoch": 2388,
    "batch_idx": 0,
    "loss": 0.2586011290550232,
    "learning_rate": 9.002394533727403e-05,
    "timestep_mean": 541.125,
    "epsilon_norm": 8.269492149353027,
    "prediction_norm": 7.0495405197143555,
    "timestamp": "2025-07-14T12:34:59.732896"
  },
  {
    "epoch": 2388,
    "batch_idx": 32,
    "loss": 0.16630782186985016,
    "learning_rate": 9.002394533727403e-05,
    "timestep_mean": 542.5625,
    "epsilon_norm": 6.247360706329346,
    "prediction_norm": 5.200317859649658,
    "timestamp": "2025-07-14T12:34:59.738115"
  },
  {
    "epoch": 2388,
    "batch_idx": 64,
    "loss": 0.2552184760570526,
    "learning_rate": 9.002394533727403e-05,
    "timestep_mean": 497.5625,
    "epsilon_norm": 8.204461097717285,
    "prediction_norm": 7.064277172088623,
    "timestamp": "2025-07-14T12:34:59.743252"
  },
  {
    "epoch": 2388,
    "batch_idx": 96,
    "loss": 0.24208670854568481,
    "learning_rate": 9.002394533727403e-05,
    "timestep_mean": 534.625,
    "epsilon_norm": 7.491002559661865,
    "prediction_norm": 6.683372974395752,
    "timestamp": "2025-07-14T12:34:59.748271"
  },
  {
    "epoch": 2388,
    "batch_idx": 128,
    "loss": 0.28647422790527344,
    "learning_rate": 9.002394533727403e-05,
    "timestep_mean": 497.1875,
    "epsilon_norm": 8.35143756866455,
    "prediction_norm": 7.481151103973389,
    "timestamp": "2025-07-14T12:34:59.753071"
  },
  {
    "epoch": 2388,
    "batch_idx": 160,
    "loss": 0.4443265199661255,
    "learning_rate": 9.002394533727403e-05,
    "timestep_mean": 423.5625,
    "epsilon_norm": 9.090641021728516,
    "prediction_norm": 6.747748374938965,
    "timestamp": "2025-07-14T12:34:59.758044"
  },
  {
    "epoch": 2388,
    "batch_idx": 192,
    "loss": 0.7117441296577454,
    "learning_rate": 9.002394533727403e-05,
    "timestep_mean": 297.25,
    "epsilon_norm": 4.441194534301758,
    "prediction_norm": 2.4315435886383057,
    "timestamp": "2025-07-14T12:34:59.763015"
  },
  {
    "epoch": 2389,
    "batch_idx": 0,
    "loss": 0.23174723982810974,
    "learning_rate": 9.047403808008057e-05,
    "timestep_mean": 526.625,
    "epsilon_norm": 9.298574447631836,
    "prediction_norm": 8.227627754211426,
    "timestamp": "2025-07-14T12:34:59.767940"
  },
  {
    "epoch": 2389,
    "batch_idx": 32,
    "loss": 0.1390460729598999,
    "learning_rate": 9.047403808008057e-05,
    "timestep_mean": 538.96875,
    "epsilon_norm": 7.156113147735596,
    "prediction_norm": 6.807897090911865,
    "timestamp": "2025-07-14T12:34:59.772798"
  },
  {
    "epoch": 2389,
    "batch_idx": 64,
    "loss": 0.5750184059143066,
    "learning_rate": 9.047403808008057e-05,
    "timestep_mean": 444.0,
    "epsilon_norm": 9.540287017822266,
    "prediction_norm": 6.987914085388184,
    "timestamp": "2025-07-14T12:34:59.777624"
  },
  {
    "epoch": 2389,
    "batch_idx": 96,
    "loss": 0.2397710531949997,
    "learning_rate": 9.047403808008057e-05,
    "timestep_mean": 561.28125,
    "epsilon_norm": 8.023459434509277,
    "prediction_norm": 6.885962963104248,
    "timestamp": "2025-07-14T12:34:59.782572"
  },
  {
    "epoch": 2389,
    "batch_idx": 128,
    "loss": 0.2936360538005829,
    "learning_rate": 9.047403808008057e-05,
    "timestep_mean": 446.90625,
    "epsilon_norm": 7.659297943115234,
    "prediction_norm": 6.611606597900391,
    "timestamp": "2025-07-14T12:34:59.787404"
  },
  {
    "epoch": 2389,
    "batch_idx": 160,
    "loss": 0.167734295129776,
    "learning_rate": 9.047403808008057e-05,
    "timestep_mean": 564.90625,
    "epsilon_norm": 9.010936737060547,
    "prediction_norm": 8.137989044189453,
    "timestamp": "2025-07-14T12:34:59.792369"
  },
  {
    "epoch": 2389,
    "batch_idx": 192,
    "loss": 0.3756919503211975,
    "learning_rate": 9.047403808008057e-05,
    "timestep_mean": 295.375,
    "epsilon_norm": 4.1370320320129395,
    "prediction_norm": 2.7099955081939697,
    "timestamp": "2025-07-14T12:34:59.797229"
  },
  {
    "epoch": 2390,
    "batch_idx": 0,
    "loss": 0.49810677766799927,
    "learning_rate": 9.09251412874885e-05,
    "timestep_mean": 498.34375,
    "epsilon_norm": 8.830632209777832,
    "prediction_norm": 6.5566534996032715,
    "timestamp": "2025-07-14T12:34:59.802134"
  },
  {
    "epoch": 2390,
    "batch_idx": 32,
    "loss": 0.14794650673866272,
    "learning_rate": 9.09251412874885e-05,
    "timestep_mean": 488.28125,
    "epsilon_norm": 8.723057746887207,
    "prediction_norm": 7.381062030792236,
    "timestamp": "2025-07-14T12:34:59.806932"
  },
  {
    "epoch": 2390,
    "batch_idx": 64,
    "loss": 0.35372138023376465,
    "learning_rate": 9.09251412874885e-05,
    "timestep_mean": 429.03125,
    "epsilon_norm": 7.996156692504883,
    "prediction_norm": 6.591353416442871,
    "timestamp": "2025-07-14T12:34:59.811747"
  },
  {
    "epoch": 2390,
    "batch_idx": 96,
    "loss": 0.1947585642337799,
    "learning_rate": 9.09251412874885e-05,
    "timestep_mean": 509.65625,
    "epsilon_norm": 6.102043151855469,
    "prediction_norm": 5.265927314758301,
    "timestamp": "2025-07-14T12:34:59.816543"
  },
  {
    "epoch": 2390,
    "batch_idx": 128,
    "loss": 0.4210399389266968,
    "learning_rate": 9.09251412874885e-05,
    "timestep_mean": 462.21875,
    "epsilon_norm": 7.110367774963379,
    "prediction_norm": 5.373936176300049,
    "timestamp": "2025-07-14T12:34:59.821581"
  },
  {
    "epoch": 2390,
    "batch_idx": 160,
    "loss": 0.40665528178215027,
    "learning_rate": 9.09251412874885e-05,
    "timestep_mean": 540.125,
    "epsilon_norm": 7.73843240737915,
    "prediction_norm": 6.047184944152832,
    "timestamp": "2025-07-14T12:34:59.826394"
  },
  {
    "epoch": 2390,
    "batch_idx": 192,
    "loss": 0.12009097635746002,
    "learning_rate": 9.09251412874885e-05,
    "timestep_mean": 530.625,
    "epsilon_norm": 3.3356282711029053,
    "prediction_norm": 3.266533851623535,
    "timestamp": "2025-07-14T12:34:59.831199"
  },
  {
    "epoch": 2391,
    "batch_idx": 0,
    "loss": 0.1641879826784134,
    "learning_rate": 9.137725384644528e-05,
    "timestep_mean": 468.84375,
    "epsilon_norm": 7.30170202255249,
    "prediction_norm": 6.130490303039551,
    "timestamp": "2025-07-14T12:34:59.836050"
  },
  {
    "epoch": 2391,
    "batch_idx": 32,
    "loss": 0.2137342393398285,
    "learning_rate": 9.137725384644528e-05,
    "timestep_mean": 455.53125,
    "epsilon_norm": 7.954261779785156,
    "prediction_norm": 6.417543411254883,
    "timestamp": "2025-07-14T12:34:59.840854"
  },
  {
    "epoch": 2391,
    "batch_idx": 64,
    "loss": 0.33820027112960815,
    "learning_rate": 9.137725384644528e-05,
    "timestep_mean": 635.0,
    "epsilon_norm": 8.230476379394531,
    "prediction_norm": 6.978365898132324,
    "timestamp": "2025-07-14T12:34:59.845743"
  },
  {
    "epoch": 2391,
    "batch_idx": 96,
    "loss": 0.22470524907112122,
    "learning_rate": 9.137725384644528e-05,
    "timestep_mean": 429.53125,
    "epsilon_norm": 7.916198253631592,
    "prediction_norm": 7.049777507781982,
    "timestamp": "2025-07-14T12:34:59.850752"
  },
  {
    "epoch": 2391,
    "batch_idx": 128,
    "loss": 0.2649575173854828,
    "learning_rate": 9.137725384644528e-05,
    "timestep_mean": 482.125,
    "epsilon_norm": 7.151547908782959,
    "prediction_norm": 6.6005940437316895,
    "timestamp": "2025-07-14T12:34:59.855933"
  },
  {
    "epoch": 2391,
    "batch_idx": 160,
    "loss": 0.09624582529067993,
    "learning_rate": 9.137725384644528e-05,
    "timestep_mean": 594.34375,
    "epsilon_norm": 8.124444007873535,
    "prediction_norm": 7.910253047943115,
    "timestamp": "2025-07-14T12:34:59.861151"
  },
  {
    "epoch": 2391,
    "batch_idx": 192,
    "loss": 0.6928715705871582,
    "learning_rate": 9.137725384644528e-05,
    "timestep_mean": 555.5,
    "epsilon_norm": 4.566211700439453,
    "prediction_norm": 3.086838722229004,
    "timestamp": "2025-07-14T12:34:59.866414"
  },
  {
    "epoch": 2392,
    "batch_idx": 0,
    "loss": 0.22753436863422394,
    "learning_rate": 9.183037464140823e-05,
    "timestep_mean": 497.375,
    "epsilon_norm": 7.138063907623291,
    "prediction_norm": 5.306772708892822,
    "timestamp": "2025-07-14T12:34:59.871565"
  },
  {
    "epoch": 2392,
    "batch_idx": 32,
    "loss": 0.37390992045402527,
    "learning_rate": 9.183037464140823e-05,
    "timestep_mean": 493.03125,
    "epsilon_norm": 7.881432056427002,
    "prediction_norm": 6.192555904388428,
    "timestamp": "2025-07-14T12:34:59.876704"
  },
  {
    "epoch": 2392,
    "batch_idx": 64,
    "loss": 0.1969689577817917,
    "learning_rate": 9.183037464140823e-05,
    "timestep_mean": 567.8125,
    "epsilon_norm": 7.448990345001221,
    "prediction_norm": 6.306550979614258,
    "timestamp": "2025-07-14T12:34:59.881882"
  },
  {
    "epoch": 2392,
    "batch_idx": 96,
    "loss": 0.43230342864990234,
    "learning_rate": 9.183037464140823e-05,
    "timestep_mean": 516.40625,
    "epsilon_norm": 7.8855719566345215,
    "prediction_norm": 6.514018535614014,
    "timestamp": "2025-07-14T12:34:59.887067"
  },
  {
    "epoch": 2392,
    "batch_idx": 128,
    "loss": 0.42338427901268005,
    "learning_rate": 9.183037464140823e-05,
    "timestep_mean": 404.875,
    "epsilon_norm": 9.50648307800293,
    "prediction_norm": 7.819517135620117,
    "timestamp": "2025-07-14T12:34:59.892614"
  },
  {
    "epoch": 2392,
    "batch_idx": 160,
    "loss": 0.2883586585521698,
    "learning_rate": 9.183037464140823e-05,
    "timestep_mean": 540.75,
    "epsilon_norm": 7.953446388244629,
    "prediction_norm": 6.828961372375488,
    "timestamp": "2025-07-14T12:34:59.897776"
  },
  {
    "epoch": 2392,
    "batch_idx": 192,
    "loss": 1.1223102807998657,
    "learning_rate": 9.183037464140823e-05,
    "timestep_mean": 432.625,
    "epsilon_norm": 5.032527446746826,
    "prediction_norm": 3.377608060836792,
    "timestamp": "2025-07-14T12:34:59.903069"
  },
  {
    "epoch": 2393,
    "batch_idx": 0,
    "loss": 0.1676015406847,
    "learning_rate": 9.228450255434688e-05,
    "timestep_mean": 482.71875,
    "epsilon_norm": 6.3658342361450195,
    "prediction_norm": 5.288539409637451,
    "timestamp": "2025-07-14T12:34:59.908329"
  },
  {
    "epoch": 2393,
    "batch_idx": 32,
    "loss": 0.3354477882385254,
    "learning_rate": 9.228450255434688e-05,
    "timestep_mean": 496.9375,
    "epsilon_norm": 7.815113544464111,
    "prediction_norm": 7.363254547119141,
    "timestamp": "2025-07-14T12:34:59.913597"
  },
  {
    "epoch": 2393,
    "batch_idx": 64,
    "loss": 0.3938809037208557,
    "learning_rate": 9.228450255434688e-05,
    "timestep_mean": 520.96875,
    "epsilon_norm": 8.275296211242676,
    "prediction_norm": 6.719467639923096,
    "timestamp": "2025-07-14T12:34:59.918829"
  },
  {
    "epoch": 2393,
    "batch_idx": 96,
    "loss": 0.36855652928352356,
    "learning_rate": 9.228450255434688e-05,
    "timestep_mean": 440.21875,
    "epsilon_norm": 8.463804244995117,
    "prediction_norm": 6.7430195808410645,
    "timestamp": "2025-07-14T12:34:59.924041"
  },
  {
    "epoch": 2393,
    "batch_idx": 128,
    "loss": 0.28985387086868286,
    "learning_rate": 9.228450255434688e-05,
    "timestep_mean": 427.15625,
    "epsilon_norm": 8.00035572052002,
    "prediction_norm": 6.981853485107422,
    "timestamp": "2025-07-14T12:34:59.929359"
  },
  {
    "epoch": 2393,
    "batch_idx": 160,
    "loss": 0.1254262775182724,
    "learning_rate": 9.228450255434688e-05,
    "timestep_mean": 586.15625,
    "epsilon_norm": 8.41002368927002,
    "prediction_norm": 7.970906734466553,
    "timestamp": "2025-07-14T12:34:59.934527"
  },
  {
    "epoch": 2393,
    "batch_idx": 192,
    "loss": 0.2189932018518448,
    "learning_rate": 9.228450255434688e-05,
    "timestep_mean": 520.75,
    "epsilon_norm": 3.5749945640563965,
    "prediction_norm": 3.3072285652160645,
    "timestamp": "2025-07-14T12:34:59.939626"
  },
  {
    "epoch": 2394,
    "batch_idx": 0,
    "loss": 0.3057331144809723,
    "learning_rate": 9.273963646474546e-05,
    "timestep_mean": 488.96875,
    "epsilon_norm": 8.029166221618652,
    "prediction_norm": 7.125597953796387,
    "timestamp": "2025-07-14T12:34:59.944805"
  },
  {
    "epoch": 2394,
    "batch_idx": 32,
    "loss": 0.4697943925857544,
    "learning_rate": 9.273963646474546e-05,
    "timestep_mean": 490.3125,
    "epsilon_norm": 8.39651107788086,
    "prediction_norm": 6.43516731262207,
    "timestamp": "2025-07-14T12:34:59.950022"
  },
  {
    "epoch": 2394,
    "batch_idx": 64,
    "loss": 0.18908575177192688,
    "learning_rate": 9.273963646474546e-05,
    "timestep_mean": 587.375,
    "epsilon_norm": 6.580326557159424,
    "prediction_norm": 5.884589195251465,
    "timestamp": "2025-07-14T12:34:59.955348"
  },
  {
    "epoch": 2394,
    "batch_idx": 96,
    "loss": 0.23325634002685547,
    "learning_rate": 9.273963646474546e-05,
    "timestep_mean": 633.78125,
    "epsilon_norm": 7.84213399887085,
    "prediction_norm": 6.415457725524902,
    "timestamp": "2025-07-14T12:34:59.960488"
  },
  {
    "epoch": 2394,
    "batch_idx": 128,
    "loss": 0.24400503933429718,
    "learning_rate": 9.273963646474546e-05,
    "timestep_mean": 607.15625,
    "epsilon_norm": 7.255229949951172,
    "prediction_norm": 6.244943141937256,
    "timestamp": "2025-07-14T12:34:59.965721"
  },
  {
    "epoch": 2394,
    "batch_idx": 160,
    "loss": 0.23340705037117004,
    "learning_rate": 9.273963646474546e-05,
    "timestep_mean": 564.125,
    "epsilon_norm": 8.606372833251953,
    "prediction_norm": 7.505452632904053,
    "timestamp": "2025-07-14T12:34:59.970968"
  },
  {
    "epoch": 2394,
    "batch_idx": 192,
    "loss": 0.25907406210899353,
    "learning_rate": 9.273963646474546e-05,
    "timestep_mean": 287.875,
    "epsilon_norm": 2.9175961017608643,
    "prediction_norm": 2.135852813720703,
    "timestamp": "2025-07-14T12:34:59.976193"
  },
  {
    "epoch": 2395,
    "batch_idx": 0,
    "loss": 0.29193252325057983,
    "learning_rate": 9.319577524960674e-05,
    "timestep_mean": 489.0,
    "epsilon_norm": 7.168989181518555,
    "prediction_norm": 6.075035095214844,
    "timestamp": "2025-07-14T12:34:59.981451"
  },
  {
    "epoch": 2395,
    "batch_idx": 32,
    "loss": 0.4475664496421814,
    "learning_rate": 9.319577524960674e-05,
    "timestep_mean": 400.09375,
    "epsilon_norm": 7.987569332122803,
    "prediction_norm": 5.642583847045898,
    "timestamp": "2025-07-14T12:34:59.986416"
  },
  {
    "epoch": 2395,
    "batch_idx": 64,
    "loss": 0.28667551279067993,
    "learning_rate": 9.319577524960674e-05,
    "timestep_mean": 409.0625,
    "epsilon_norm": 7.137690544128418,
    "prediction_norm": 6.174527645111084,
    "timestamp": "2025-07-14T12:34:59.991585"
  },
  {
    "epoch": 2395,
    "batch_idx": 96,
    "loss": 0.17264539003372192,
    "learning_rate": 9.319577524960674e-05,
    "timestep_mean": 511.40625,
    "epsilon_norm": 6.5557732582092285,
    "prediction_norm": 5.6559882164001465,
    "timestamp": "2025-07-14T12:34:59.996846"
  },
  {
    "epoch": 2395,
    "batch_idx": 128,
    "loss": 0.11386953294277191,
    "learning_rate": 9.319577524960674e-05,
    "timestep_mean": 650.375,
    "epsilon_norm": 9.124640464782715,
    "prediction_norm": 8.52232551574707,
    "timestamp": "2025-07-14T12:35:00.001872"
  },
  {
    "epoch": 2395,
    "batch_idx": 160,
    "loss": 0.18219204246997833,
    "learning_rate": 9.319577524960674e-05,
    "timestep_mean": 512.65625,
    "epsilon_norm": 7.076167106628418,
    "prediction_norm": 6.808674335479736,
    "timestamp": "2025-07-14T12:35:00.006794"
  },
  {
    "epoch": 2395,
    "batch_idx": 192,
    "loss": 0.10953603684902191,
    "learning_rate": 9.319577524960674e-05,
    "timestep_mean": 550.5,
    "epsilon_norm": 3.6773672103881836,
    "prediction_norm": 2.888139247894287,
    "timestamp": "2025-07-14T12:35:00.017651"
  },
  {
    "epoch": 2396,
    "batch_idx": 0,
    "loss": 0.41936689615249634,
    "learning_rate": 9.365291778345334e-05,
    "timestep_mean": 505.875,
    "epsilon_norm": 7.403255462646484,
    "prediction_norm": 5.561443328857422,
    "timestamp": "2025-07-14T12:35:00.022720"
  },
  {
    "epoch": 2396,
    "batch_idx": 32,
    "loss": 0.19752134382724762,
    "learning_rate": 9.365291778345334e-05,
    "timestep_mean": 512.34375,
    "epsilon_norm": 7.2194318771362305,
    "prediction_norm": 6.167787551879883,
    "timestamp": "2025-07-14T12:35:00.027694"
  },
  {
    "epoch": 2396,
    "batch_idx": 64,
    "loss": 0.2645219564437866,
    "learning_rate": 9.365291778345334e-05,
    "timestep_mean": 452.125,
    "epsilon_norm": 8.319036483764648,
    "prediction_norm": 7.524710178375244,
    "timestamp": "2025-07-14T12:35:00.032717"
  },
  {
    "epoch": 2396,
    "batch_idx": 96,
    "loss": 0.36744755506515503,
    "learning_rate": 9.365291778345334e-05,
    "timestep_mean": 522.28125,
    "epsilon_norm": 8.218457221984863,
    "prediction_norm": 6.777249813079834,
    "timestamp": "2025-07-14T12:35:00.037753"
  },
  {
    "epoch": 2396,
    "batch_idx": 128,
    "loss": 0.2082040160894394,
    "learning_rate": 9.365291778345334e-05,
    "timestep_mean": 503.84375,
    "epsilon_norm": 7.7159104347229,
    "prediction_norm": 6.798738956451416,
    "timestamp": "2025-07-14T12:35:00.042677"
  },
  {
    "epoch": 2396,
    "batch_idx": 160,
    "loss": 0.3843958377838135,
    "learning_rate": 9.365291778345334e-05,
    "timestep_mean": 400.21875,
    "epsilon_norm": 7.714481353759766,
    "prediction_norm": 6.011960506439209,
    "timestamp": "2025-07-14T12:35:00.047704"
  },
  {
    "epoch": 2396,
    "batch_idx": 192,
    "loss": 0.04140826314687729,
    "learning_rate": 9.365291778345334e-05,
    "timestep_mean": 686.25,
    "epsilon_norm": 3.3228189945220947,
    "prediction_norm": 3.089404582977295,
    "timestamp": "2025-07-14T12:35:00.052648"
  },
  {
    "epoch": 2397,
    "batch_idx": 0,
    "loss": 0.21715430915355682,
    "learning_rate": 9.41110629383313e-05,
    "timestep_mean": 501.65625,
    "epsilon_norm": 7.796413898468018,
    "prediction_norm": 6.421536922454834,
    "timestamp": "2025-07-14T12:35:00.057647"
  },
  {
    "epoch": 2397,
    "batch_idx": 32,
    "loss": 0.18559050559997559,
    "learning_rate": 9.41110629383313e-05,
    "timestep_mean": 513.5,
    "epsilon_norm": 7.71799373626709,
    "prediction_norm": 6.4284348487854,
    "timestamp": "2025-07-14T12:35:00.062588"
  },
  {
    "epoch": 2397,
    "batch_idx": 64,
    "loss": 0.23876355588436127,
    "learning_rate": 9.41110629383313e-05,
    "timestep_mean": 462.34375,
    "epsilon_norm": 6.480154514312744,
    "prediction_norm": 5.292003154754639,
    "timestamp": "2025-07-14T12:35:00.067769"
  },
  {
    "epoch": 2397,
    "batch_idx": 96,
    "loss": 0.2012060582637787,
    "learning_rate": 9.41110629383313e-05,
    "timestep_mean": 463.5625,
    "epsilon_norm": 7.981535911560059,
    "prediction_norm": 6.678328514099121,
    "timestamp": "2025-07-14T12:35:00.072741"
  },
  {
    "epoch": 2397,
    "batch_idx": 128,
    "loss": 0.34608349204063416,
    "learning_rate": 9.41110629383313e-05,
    "timestep_mean": 490.65625,
    "epsilon_norm": 7.313973426818848,
    "prediction_norm": 5.72075080871582,
    "timestamp": "2025-07-14T12:35:00.077803"
  },
  {
    "epoch": 2397,
    "batch_idx": 160,
    "loss": 0.3937615156173706,
    "learning_rate": 9.41110629383313e-05,
    "timestep_mean": 395.0,
    "epsilon_norm": 9.756921768188477,
    "prediction_norm": 8.261597633361816,
    "timestamp": "2025-07-14T12:35:00.082827"
  },
  {
    "epoch": 2397,
    "batch_idx": 192,
    "loss": 0.10709112882614136,
    "learning_rate": 9.41110629383313e-05,
    "timestep_mean": 571.375,
    "epsilon_norm": 4.298872947692871,
    "prediction_norm": 4.216396331787109,
    "timestamp": "2025-07-14T12:35:00.087795"
  },
  {
    "epoch": 2398,
    "batch_idx": 0,
    "loss": 0.320448100566864,
    "learning_rate": 9.457020958381353e-05,
    "timestep_mean": 473.90625,
    "epsilon_norm": 8.058271408081055,
    "prediction_norm": 6.788157939910889,
    "timestamp": "2025-07-14T12:35:00.092802"
  },
  {
    "epoch": 2398,
    "batch_idx": 32,
    "loss": 0.2637390196323395,
    "learning_rate": 9.457020958381353e-05,
    "timestep_mean": 471.0625,
    "epsilon_norm": 7.664769172668457,
    "prediction_norm": 6.275193214416504,
    "timestamp": "2025-07-14T12:35:00.097868"
  },
  {
    "epoch": 2398,
    "batch_idx": 64,
    "loss": 0.19349387288093567,
    "learning_rate": 9.457020958381353e-05,
    "timestep_mean": 514.71875,
    "epsilon_norm": 7.4937543869018555,
    "prediction_norm": 6.483509540557861,
    "timestamp": "2025-07-14T12:35:00.102968"
  },
  {
    "epoch": 2398,
    "batch_idx": 96,
    "loss": 0.1638682782649994,
    "learning_rate": 9.457020958381353e-05,
    "timestep_mean": 508.125,
    "epsilon_norm": 8.495077133178711,
    "prediction_norm": 7.776979446411133,
    "timestamp": "2025-07-14T12:35:00.108035"
  },
  {
    "epoch": 2398,
    "batch_idx": 128,
    "loss": 0.29012393951416016,
    "learning_rate": 9.457020958381353e-05,
    "timestep_mean": 583.875,
    "epsilon_norm": 9.178364753723145,
    "prediction_norm": 8.06697940826416,
    "timestamp": "2025-07-14T12:35:00.113154"
  },
  {
    "epoch": 2398,
    "batch_idx": 160,
    "loss": 0.4426286220550537,
    "learning_rate": 9.457020958381353e-05,
    "timestep_mean": 472.09375,
    "epsilon_norm": 8.01071834564209,
    "prediction_norm": 6.314879894256592,
    "timestamp": "2025-07-14T12:35:00.118123"
  },
  {
    "epoch": 2398,
    "batch_idx": 192,
    "loss": 0.6688932776451111,
    "learning_rate": 9.457020958381353e-05,
    "timestep_mean": 674.5,
    "epsilon_norm": 4.670690536499023,
    "prediction_norm": 3.3551368713378906,
    "timestamp": "2025-07-14T12:35:00.123047"
  },
  {
    "epoch": 2399,
    "batch_idx": 0,
    "loss": 0.3566102683544159,
    "learning_rate": 9.503035658700072e-05,
    "timestep_mean": 575.6875,
    "epsilon_norm": 8.212982177734375,
    "prediction_norm": 7.142906188964844,
    "timestamp": "2025-07-14T12:35:00.128057"
  },
  {
    "epoch": 2399,
    "batch_idx": 32,
    "loss": 0.13165323436260223,
    "learning_rate": 9.503035658700072e-05,
    "timestep_mean": 582.34375,
    "epsilon_norm": 8.567400932312012,
    "prediction_norm": 7.757423400878906,
    "timestamp": "2025-07-14T12:35:00.133078"
  },
  {
    "epoch": 2399,
    "batch_idx": 64,
    "loss": 0.11009294539690018,
    "learning_rate": 9.503035658700072e-05,
    "timestep_mean": 582.40625,
    "epsilon_norm": 8.215134620666504,
    "prediction_norm": 7.648421764373779,
    "timestamp": "2025-07-14T12:35:00.138240"
  },
  {
    "epoch": 2399,
    "batch_idx": 96,
    "loss": 0.17029322683811188,
    "learning_rate": 9.503035658700072e-05,
    "timestep_mean": 496.375,
    "epsilon_norm": 8.015851974487305,
    "prediction_norm": 7.289095401763916,
    "timestamp": "2025-07-14T12:35:00.143382"
  },
  {
    "epoch": 2399,
    "batch_idx": 128,
    "loss": 0.3735344409942627,
    "learning_rate": 9.503035658700072e-05,
    "timestep_mean": 539.90625,
    "epsilon_norm": 7.774883270263672,
    "prediction_norm": 6.289248943328857,
    "timestamp": "2025-07-14T12:35:00.148329"
  },
  {
    "epoch": 2399,
    "batch_idx": 160,
    "loss": 0.3220352530479431,
    "learning_rate": 9.503035658700072e-05,
    "timestep_mean": 582.875,
    "epsilon_norm": 8.559515953063965,
    "prediction_norm": 6.82865047454834,
    "timestamp": "2025-07-14T12:35:00.153519"
  },
  {
    "epoch": 2399,
    "batch_idx": 192,
    "loss": 0.15359249711036682,
    "learning_rate": 9.503035658700072e-05,
    "timestep_mean": 628.375,
    "epsilon_norm": 3.3826260566711426,
    "prediction_norm": 3.0102922916412354,
    "timestamp": "2025-07-14T12:35:00.158727"
  },
  {
    "epoch": 2400,
    "batch_idx": 0,
    "loss": 0.2362864911556244,
    "learning_rate": 9.54915028125265e-05,
    "timestep_mean": 490.625,
    "epsilon_norm": 8.166481018066406,
    "prediction_norm": 7.229464530944824,
    "timestamp": "2025-07-14T12:35:00.163957"
  },
  {
    "epoch": 2400,
    "batch_idx": 32,
    "loss": 0.34128159284591675,
    "learning_rate": 9.54915028125265e-05,
    "timestep_mean": 520.125,
    "epsilon_norm": 9.23103141784668,
    "prediction_norm": 7.382731914520264,
    "timestamp": "2025-07-14T12:35:00.169212"
  },
  {
    "epoch": 2400,
    "batch_idx": 64,
    "loss": 0.1258830726146698,
    "learning_rate": 9.54915028125265e-05,
    "timestep_mean": 505.4375,
    "epsilon_norm": 7.489856719970703,
    "prediction_norm": 7.177438259124756,
    "timestamp": "2025-07-14T12:35:00.174486"
  },
  {
    "epoch": 2400,
    "batch_idx": 96,
    "loss": 0.15658457577228546,
    "learning_rate": 9.54915028125265e-05,
    "timestep_mean": 550.78125,
    "epsilon_norm": 7.173396110534668,
    "prediction_norm": 5.96738862991333,
    "timestamp": "2025-07-14T12:35:00.179690"
  },
  {
    "epoch": 2400,
    "batch_idx": 128,
    "loss": 0.42913687229156494,
    "learning_rate": 9.54915028125265e-05,
    "timestep_mean": 467.6875,
    "epsilon_norm": 8.982348442077637,
    "prediction_norm": 7.035050392150879,
    "timestamp": "2025-07-14T12:35:00.184897"
  },
  {
    "epoch": 2400,
    "batch_idx": 160,
    "loss": 0.219707190990448,
    "learning_rate": 9.54915028125265e-05,
    "timestep_mean": 492.65625,
    "epsilon_norm": 8.84351921081543,
    "prediction_norm": 7.606037139892578,
    "timestamp": "2025-07-14T12:35:00.190078"
  },
  {
    "epoch": 2400,
    "batch_idx": 192,
    "loss": 0.44624781608581543,
    "learning_rate": 9.54915028125265e-05,
    "timestep_mean": 372.0,
    "epsilon_norm": 3.65718412399292,
    "prediction_norm": 2.628098964691162,
    "timestamp": "2025-07-14T12:35:00.195179"
  },
  {
    "epoch": 2401,
    "batch_idx": 0,
    "loss": 0.14990898966789246,
    "learning_rate": 9.59536471225579e-05,
    "timestep_mean": 541.1875,
    "epsilon_norm": 8.442797660827637,
    "prediction_norm": 7.700395107269287,
    "timestamp": "2025-07-14T12:35:00.200508"
  },
  {
    "epoch": 2401,
    "batch_idx": 32,
    "loss": 0.3626123070716858,
    "learning_rate": 9.59536471225579e-05,
    "timestep_mean": 438.3125,
    "epsilon_norm": 8.35329818725586,
    "prediction_norm": 7.052323341369629,
    "timestamp": "2025-07-14T12:35:00.205778"
  },
  {
    "epoch": 2401,
    "batch_idx": 64,
    "loss": 0.24455764889717102,
    "learning_rate": 9.59536471225579e-05,
    "timestep_mean": 468.125,
    "epsilon_norm": 6.600005626678467,
    "prediction_norm": 5.366993427276611,
    "timestamp": "2025-07-14T12:35:00.210937"
  },
  {
    "epoch": 2401,
    "batch_idx": 96,
    "loss": 0.2452974170446396,
    "learning_rate": 9.59536471225579e-05,
    "timestep_mean": 507.78125,
    "epsilon_norm": 7.302652359008789,
    "prediction_norm": 6.355494976043701,
    "timestamp": "2025-07-14T12:35:00.216155"
  },
  {
    "epoch": 2401,
    "batch_idx": 128,
    "loss": 0.2908717095851898,
    "learning_rate": 9.59536471225579e-05,
    "timestep_mean": 395.0625,
    "epsilon_norm": 7.282307147979736,
    "prediction_norm": 5.992715358734131,
    "timestamp": "2025-07-14T12:35:00.221392"
  },
  {
    "epoch": 2401,
    "batch_idx": 160,
    "loss": 0.26326704025268555,
    "learning_rate": 9.59536471225579e-05,
    "timestep_mean": 405.28125,
    "epsilon_norm": 7.245398044586182,
    "prediction_norm": 6.341165542602539,
    "timestamp": "2025-07-14T12:35:00.226561"
  },
  {
    "epoch": 2401,
    "batch_idx": 192,
    "loss": 0.35660919547080994,
    "learning_rate": 9.59536471225579e-05,
    "timestep_mean": 402.5,
    "epsilon_norm": 3.3854730129241943,
    "prediction_norm": 2.5380403995513916,
    "timestamp": "2025-07-14T12:35:00.231722"
  },
  {
    "epoch": 2402,
    "batch_idx": 0,
    "loss": 0.36678534746170044,
    "learning_rate": 9.641678837680003e-05,
    "timestep_mean": 406.6875,
    "epsilon_norm": 7.592114448547363,
    "prediction_norm": 6.180432319641113,
    "timestamp": "2025-07-14T12:35:00.236960"
  },
  {
    "epoch": 2402,
    "batch_idx": 32,
    "loss": 0.296930193901062,
    "learning_rate": 9.641678837680003e-05,
    "timestep_mean": 495.46875,
    "epsilon_norm": 7.561378479003906,
    "prediction_norm": 5.916155815124512,
    "timestamp": "2025-07-14T12:35:00.242221"
  },
  {
    "epoch": 2402,
    "batch_idx": 64,
    "loss": 0.3595848083496094,
    "learning_rate": 9.641678837680003e-05,
    "timestep_mean": 434.71875,
    "epsilon_norm": 7.6923441886901855,
    "prediction_norm": 5.840794086456299,
    "timestamp": "2025-07-14T12:35:00.247407"
  },
  {
    "epoch": 2402,
    "batch_idx": 96,
    "loss": 0.2557028532028198,
    "learning_rate": 9.641678837680003e-05,
    "timestep_mean": 539.15625,
    "epsilon_norm": 7.5881195068359375,
    "prediction_norm": 6.275747776031494,
    "timestamp": "2025-07-14T12:35:00.252603"
  },
  {
    "epoch": 2402,
    "batch_idx": 128,
    "loss": 0.5562667846679688,
    "learning_rate": 9.641678837680003e-05,
    "timestep_mean": 438.53125,
    "epsilon_norm": 8.672845840454102,
    "prediction_norm": 6.795953750610352,
    "timestamp": "2025-07-14T12:35:00.257766"
  },
  {
    "epoch": 2402,
    "batch_idx": 160,
    "loss": 0.265838623046875,
    "learning_rate": 9.641678837680003e-05,
    "timestep_mean": 396.15625,
    "epsilon_norm": 7.779478073120117,
    "prediction_norm": 6.022846698760986,
    "timestamp": "2025-07-14T12:35:00.262940"
  },
  {
    "epoch": 2402,
    "batch_idx": 192,
    "loss": 0.611807107925415,
    "learning_rate": 9.641678837680003e-05,
    "timestep_mean": 414.375,
    "epsilon_norm": 4.267873764038086,
    "prediction_norm": 3.057692289352417,
    "timestamp": "2025-07-14T12:35:00.267958"
  },
  {
    "epoch": 2403,
    "batch_idx": 0,
    "loss": 0.21102815866470337,
    "learning_rate": 9.68809254324979e-05,
    "timestep_mean": 515.9375,
    "epsilon_norm": 8.253138542175293,
    "prediction_norm": 6.899427890777588,
    "timestamp": "2025-07-14T12:35:00.273142"
  },
  {
    "epoch": 2403,
    "batch_idx": 32,
    "loss": 0.1561553180217743,
    "learning_rate": 9.68809254324979e-05,
    "timestep_mean": 566.875,
    "epsilon_norm": 6.655187606811523,
    "prediction_norm": 5.7854461669921875,
    "timestamp": "2025-07-14T12:35:00.278412"
  },
  {
    "epoch": 2403,
    "batch_idx": 64,
    "loss": 0.25237560272216797,
    "learning_rate": 9.68809254324979e-05,
    "timestep_mean": 497.625,
    "epsilon_norm": 7.032680034637451,
    "prediction_norm": 5.7342987060546875,
    "timestamp": "2025-07-14T12:35:00.283666"
  },
  {
    "epoch": 2403,
    "batch_idx": 96,
    "loss": 0.5101523995399475,
    "learning_rate": 9.68809254324979e-05,
    "timestep_mean": 496.15625,
    "epsilon_norm": 7.898317813873291,
    "prediction_norm": 5.914909362792969,
    "timestamp": "2025-07-14T12:35:00.288878"
  },
  {
    "epoch": 2403,
    "batch_idx": 128,
    "loss": 0.37748658657073975,
    "learning_rate": 9.68809254324979e-05,
    "timestep_mean": 456.96875,
    "epsilon_norm": 7.469834804534912,
    "prediction_norm": 5.935799598693848,
    "timestamp": "2025-07-14T12:35:00.294116"
  },
  {
    "epoch": 2403,
    "batch_idx": 160,
    "loss": 0.13302412629127502,
    "learning_rate": 9.68809254324979e-05,
    "timestep_mean": 516.25,
    "epsilon_norm": 8.335744857788086,
    "prediction_norm": 7.477677345275879,
    "timestamp": "2025-07-14T12:35:00.299348"
  },
  {
    "epoch": 2403,
    "batch_idx": 192,
    "loss": 0.21807020902633667,
    "learning_rate": 9.68809254324979e-05,
    "timestep_mean": 437.5,
    "epsilon_norm": 3.36832332611084,
    "prediction_norm": 2.7754430770874023,
    "timestamp": "2025-07-14T12:35:00.304529"
  },
  {
    "epoch": 2404,
    "batch_idx": 0,
    "loss": 0.35610532760620117,
    "learning_rate": 9.734605714443925e-05,
    "timestep_mean": 431.625,
    "epsilon_norm": 7.934976100921631,
    "prediction_norm": 6.475448131561279,
    "timestamp": "2025-07-14T12:35:00.309650"
  },
  {
    "epoch": 2404,
    "batch_idx": 32,
    "loss": 0.26466551423072815,
    "learning_rate": 9.734605714443925e-05,
    "timestep_mean": 539.96875,
    "epsilon_norm": 7.0501251220703125,
    "prediction_norm": 5.686648845672607,
    "timestamp": "2025-07-14T12:35:00.314835"
  },
  {
    "epoch": 2404,
    "batch_idx": 64,
    "loss": 0.27076634764671326,
    "learning_rate": 9.734605714443925e-05,
    "timestep_mean": 466.9375,
    "epsilon_norm": 7.786022186279297,
    "prediction_norm": 6.297595500946045,
    "timestamp": "2025-07-14T12:35:00.320069"
  },
  {
    "epoch": 2404,
    "batch_idx": 96,
    "loss": 0.22285941243171692,
    "learning_rate": 9.734605714443925e-05,
    "timestep_mean": 519.625,
    "epsilon_norm": 8.584948539733887,
    "prediction_norm": 7.207002639770508,
    "timestamp": "2025-07-14T12:35:00.325331"
  },
  {
    "epoch": 2404,
    "batch_idx": 128,
    "loss": 0.08788339048624039,
    "learning_rate": 9.734605714443925e-05,
    "timestep_mean": 549.0,
    "epsilon_norm": 7.844188213348389,
    "prediction_norm": 7.245180606842041,
    "timestamp": "2025-07-14T12:35:00.330542"
  },
  {
    "epoch": 2404,
    "batch_idx": 160,
    "loss": 0.16806551814079285,
    "learning_rate": 9.734605714443925e-05,
    "timestep_mean": 483.96875,
    "epsilon_norm": 5.9058918952941895,
    "prediction_norm": 4.706003665924072,
    "timestamp": "2025-07-14T12:35:00.335753"
  },
  {
    "epoch": 2404,
    "batch_idx": 192,
    "loss": 0.10685719549655914,
    "learning_rate": 9.734605714443925e-05,
    "timestep_mean": 468.5,
    "epsilon_norm": 3.4549946784973145,
    "prediction_norm": 3.01800537109375,
    "timestamp": "2025-07-14T12:35:00.340727"
  },
  {
    "epoch": 2405,
    "batch_idx": 0,
    "loss": 0.09495363384485245,
    "learning_rate": 9.781218236495794e-05,
    "timestep_mean": 522.28125,
    "epsilon_norm": 7.1437249183654785,
    "prediction_norm": 6.665027618408203,
    "timestamp": "2025-07-14T12:35:00.345980"
  },
  {
    "epoch": 2405,
    "batch_idx": 32,
    "loss": 0.12613919377326965,
    "learning_rate": 9.781218236495794e-05,
    "timestep_mean": 556.65625,
    "epsilon_norm": 7.115521430969238,
    "prediction_norm": 6.802919864654541,
    "timestamp": "2025-07-14T12:35:00.351194"
  },
  {
    "epoch": 2405,
    "batch_idx": 64,
    "loss": 0.1877603828907013,
    "learning_rate": 9.781218236495794e-05,
    "timestep_mean": 517.84375,
    "epsilon_norm": 6.931029319763184,
    "prediction_norm": 6.564073085784912,
    "timestamp": "2025-07-14T12:35:00.356332"
  },
  {
    "epoch": 2405,
    "batch_idx": 96,
    "loss": 0.1467858850955963,
    "learning_rate": 9.781218236495794e-05,
    "timestep_mean": 532.65625,
    "epsilon_norm": 7.527373313903809,
    "prediction_norm": 6.925187110900879,
    "timestamp": "2025-07-14T12:35:00.361650"
  },
  {
    "epoch": 2405,
    "batch_idx": 128,
    "loss": 0.2560865879058838,
    "learning_rate": 9.781218236495794e-05,
    "timestep_mean": 554.96875,
    "epsilon_norm": 8.555680274963379,
    "prediction_norm": 7.952111721038818,
    "timestamp": "2025-07-14T12:35:00.367043"
  },
  {
    "epoch": 2405,
    "batch_idx": 160,
    "loss": 0.18744011223316193,
    "learning_rate": 9.781218236495794e-05,
    "timestep_mean": 531.5625,
    "epsilon_norm": 7.788218975067139,
    "prediction_norm": 6.6500701904296875,
    "timestamp": "2025-07-14T12:35:00.372286"
  },
  {
    "epoch": 2405,
    "batch_idx": 192,
    "loss": 0.5131803750991821,
    "learning_rate": 9.781218236495794e-05,
    "timestep_mean": 456.625,
    "epsilon_norm": 5.361766815185547,
    "prediction_norm": 3.818894624710083,
    "timestamp": "2025-07-14T12:35:00.377460"
  },
  {
    "epoch": 2406,
    "batch_idx": 0,
    "loss": 0.2621583044528961,
    "learning_rate": 9.827929994393632e-05,
    "timestep_mean": 460.78125,
    "epsilon_norm": 6.613762378692627,
    "prediction_norm": 5.476378440856934,
    "timestamp": "2025-07-14T12:35:00.382755"
  },
  {
    "epoch": 2406,
    "batch_idx": 32,
    "loss": 0.27487021684646606,
    "learning_rate": 9.827929994393632e-05,
    "timestep_mean": 478.375,
    "epsilon_norm": 7.8762969970703125,
    "prediction_norm": 6.536018371582031,
    "timestamp": "2025-07-14T12:35:00.387897"
  },
  {
    "epoch": 2406,
    "batch_idx": 64,
    "loss": 0.17184653878211975,
    "learning_rate": 9.827929994393632e-05,
    "timestep_mean": 481.53125,
    "epsilon_norm": 8.088603019714355,
    "prediction_norm": 6.855894565582275,
    "timestamp": "2025-07-14T12:35:00.392888"
  },
  {
    "epoch": 2406,
    "batch_idx": 96,
    "loss": 0.4898458421230316,
    "learning_rate": 9.827929994393632e-05,
    "timestep_mean": 437.625,
    "epsilon_norm": 8.583139419555664,
    "prediction_norm": 6.160802841186523,
    "timestamp": "2025-07-14T12:35:00.397931"
  },
  {
    "epoch": 2406,
    "batch_idx": 128,
    "loss": 0.23632198572158813,
    "learning_rate": 9.827929994393632e-05,
    "timestep_mean": 518.6875,
    "epsilon_norm": 6.789729595184326,
    "prediction_norm": 5.463912487030029,
    "timestamp": "2025-07-14T12:35:00.402843"
  },
  {
    "epoch": 2406,
    "batch_idx": 160,
    "loss": 0.18088793754577637,
    "learning_rate": 9.827929994393632e-05,
    "timestep_mean": 647.625,
    "epsilon_norm": 8.35126781463623,
    "prediction_norm": 7.002140045166016,
    "timestamp": "2025-07-14T12:35:00.407909"
  },
  {
    "epoch": 2406,
    "batch_idx": 192,
    "loss": 0.2514043152332306,
    "learning_rate": 9.827929994393632e-05,
    "timestep_mean": 465.0,
    "epsilon_norm": 3.577836751937866,
    "prediction_norm": 3.385434865951538,
    "timestamp": "2025-07-14T12:35:00.414078"
  },
  {
    "epoch": 2407,
    "batch_idx": 0,
    "loss": 0.29287803173065186,
    "learning_rate": 9.874740872880839e-05,
    "timestep_mean": 480.5,
    "epsilon_norm": 7.229179859161377,
    "prediction_norm": 5.834659576416016,
    "timestamp": "2025-07-14T12:35:00.419309"
  },
  {
    "epoch": 2407,
    "batch_idx": 32,
    "loss": 0.3411112427711487,
    "learning_rate": 9.874740872880839e-05,
    "timestep_mean": 508.34375,
    "epsilon_norm": 7.571539878845215,
    "prediction_norm": 6.052134037017822,
    "timestamp": "2025-07-14T12:35:00.430863"
  },
  {
    "epoch": 2407,
    "batch_idx": 64,
    "loss": 0.17446836829185486,
    "learning_rate": 9.874740872880839e-05,
    "timestep_mean": 660.5,
    "epsilon_norm": 9.068800926208496,
    "prediction_norm": 8.476424217224121,
    "timestamp": "2025-07-14T12:35:00.436200"
  },
  {
    "epoch": 2407,
    "batch_idx": 96,
    "loss": 0.5861158967018127,
    "learning_rate": 9.874740872880839e-05,
    "timestep_mean": 499.71875,
    "epsilon_norm": 8.24388313293457,
    "prediction_norm": 6.531285762786865,
    "timestamp": "2025-07-14T12:35:00.441427"
  },
  {
    "epoch": 2407,
    "batch_idx": 128,
    "loss": 0.23379120230674744,
    "learning_rate": 9.874740872880839e-05,
    "timestep_mean": 483.8125,
    "epsilon_norm": 8.100141525268555,
    "prediction_norm": 6.7463836669921875,
    "timestamp": "2025-07-14T12:35:00.446587"
  },
  {
    "epoch": 2407,
    "batch_idx": 160,
    "loss": 0.09938260912895203,
    "learning_rate": 9.874740872880839e-05,
    "timestep_mean": 541.28125,
    "epsilon_norm": 6.675723552703857,
    "prediction_norm": 6.179121494293213,
    "timestamp": "2025-07-14T12:35:00.451783"
  },
  {
    "epoch": 2407,
    "batch_idx": 192,
    "loss": 0.12963399291038513,
    "learning_rate": 9.874740872880839e-05,
    "timestep_mean": 492.0,
    "epsilon_norm": 3.7196290493011475,
    "prediction_norm": 3.3475990295410156,
    "timestamp": "2025-07-14T12:35:00.456993"
  },
  {
    "epoch": 2408,
    "batch_idx": 0,
    "loss": 0.347258061170578,
    "learning_rate": 9.92165075645619e-05,
    "timestep_mean": 494.0625,
    "epsilon_norm": 8.195162773132324,
    "prediction_norm": 6.256433963775635,
    "timestamp": "2025-07-14T12:35:00.462158"
  },
  {
    "epoch": 2408,
    "batch_idx": 32,
    "loss": 0.48986178636550903,
    "learning_rate": 9.92165075645619e-05,
    "timestep_mean": 423.78125,
    "epsilon_norm": 8.96296215057373,
    "prediction_norm": 6.80976676940918,
    "timestamp": "2025-07-14T12:35:00.467616"
  },
  {
    "epoch": 2408,
    "batch_idx": 64,
    "loss": 0.129503071308136,
    "learning_rate": 9.92165075645619e-05,
    "timestep_mean": 605.125,
    "epsilon_norm": 8.113945960998535,
    "prediction_norm": 7.504753112792969,
    "timestamp": "2025-07-14T12:35:00.472613"
  },
  {
    "epoch": 2408,
    "batch_idx": 96,
    "loss": 0.1172897219657898,
    "learning_rate": 9.92165075645619e-05,
    "timestep_mean": 542.71875,
    "epsilon_norm": 7.306836128234863,
    "prediction_norm": 6.670239448547363,
    "timestamp": "2025-07-14T12:35:00.477767"
  },
  {
    "epoch": 2408,
    "batch_idx": 128,
    "loss": 0.2871885895729065,
    "learning_rate": 9.92165075645619e-05,
    "timestep_mean": 498.15625,
    "epsilon_norm": 7.811586856842041,
    "prediction_norm": 6.55630350112915,
    "timestamp": "2025-07-14T12:35:00.483001"
  },
  {
    "epoch": 2408,
    "batch_idx": 160,
    "loss": 0.10975408554077148,
    "learning_rate": 9.92165075645619e-05,
    "timestep_mean": 563.21875,
    "epsilon_norm": 7.1572980880737305,
    "prediction_norm": 6.74175500869751,
    "timestamp": "2025-07-14T12:35:00.488242"
  },
  {
    "epoch": 2408,
    "batch_idx": 192,
    "loss": 0.433189332485199,
    "learning_rate": 9.92165075645619e-05,
    "timestep_mean": 575.875,
    "epsilon_norm": 3.861309289932251,
    "prediction_norm": 3.130375623703003,
    "timestamp": "2025-07-14T12:35:00.493325"
  },
  {
    "epoch": 2409,
    "batch_idx": 0,
    "loss": 0.5723455548286438,
    "learning_rate": 9.968659529374235e-05,
    "timestep_mean": 476.15625,
    "epsilon_norm": 9.363676071166992,
    "prediction_norm": 6.450407981872559,
    "timestamp": "2025-07-14T12:35:00.498509"
  },
  {
    "epoch": 2409,
    "batch_idx": 32,
    "loss": 0.4346548616886139,
    "learning_rate": 9.968659529374235e-05,
    "timestep_mean": 601.0,
    "epsilon_norm": 9.077157974243164,
    "prediction_norm": 7.769267559051514,
    "timestamp": "2025-07-14T12:35:00.503717"
  },
  {
    "epoch": 2409,
    "batch_idx": 64,
    "loss": 0.23252002894878387,
    "learning_rate": 9.968659529374235e-05,
    "timestep_mean": 483.15625,
    "epsilon_norm": 8.216891288757324,
    "prediction_norm": 7.803765773773193,
    "timestamp": "2025-07-14T12:35:00.509137"
  },
  {
    "epoch": 2409,
    "batch_idx": 96,
    "loss": 0.2492111325263977,
    "learning_rate": 9.968659529374235e-05,
    "timestep_mean": 392.84375,
    "epsilon_norm": 7.181349277496338,
    "prediction_norm": 6.328484058380127,
    "timestamp": "2025-07-14T12:35:00.514303"
  },
  {
    "epoch": 2409,
    "batch_idx": 128,
    "loss": 0.20306123793125153,
    "learning_rate": 9.968659529374235e-05,
    "timestep_mean": 484.65625,
    "epsilon_norm": 6.992309093475342,
    "prediction_norm": 6.153535842895508,
    "timestamp": "2025-07-14T12:35:00.519402"
  },
  {
    "epoch": 2409,
    "batch_idx": 160,
    "loss": 0.3775705397129059,
    "learning_rate": 9.968659529374235e-05,
    "timestep_mean": 458.8125,
    "epsilon_norm": 8.354131698608398,
    "prediction_norm": 6.853310585021973,
    "timestamp": "2025-07-14T12:35:00.524453"
  },
  {
    "epoch": 2409,
    "batch_idx": 192,
    "loss": 0.30494415760040283,
    "learning_rate": 9.968659529374235e-05,
    "timestep_mean": 402.5,
    "epsilon_norm": 3.7565574645996094,
    "prediction_norm": 3.3302676677703857,
    "timestamp": "2025-07-14T12:35:00.529655"
  },
  {
    "epoch": 2410,
    "batch_idx": 0,
    "loss": 0.16268789768218994,
    "learning_rate": 0.00010015767075645488,
    "timestep_mean": 510.40625,
    "epsilon_norm": 7.233957290649414,
    "prediction_norm": 6.765772819519043,
    "timestamp": "2025-07-14T12:35:00.534837"
  },
  {
    "epoch": 2410,
    "batch_idx": 32,
    "loss": 0.3172990381717682,
    "learning_rate": 0.00010015767075645488,
    "timestep_mean": 513.9375,
    "epsilon_norm": 7.4063801765441895,
    "prediction_norm": 6.044169902801514,
    "timestamp": "2025-07-14T12:35:00.540026"
  },
  {
    "epoch": 2410,
    "batch_idx": 64,
    "loss": 0.27084532380104065,
    "learning_rate": 0.00010015767075645488,
    "timestep_mean": 560.78125,
    "epsilon_norm": 7.915310859680176,
    "prediction_norm": 6.682384014129639,
    "timestamp": "2025-07-14T12:35:00.545262"
  },
  {
    "epoch": 2410,
    "batch_idx": 96,
    "loss": 0.13829925656318665,
    "learning_rate": 0.00010015767075645488,
    "timestep_mean": 440.65625,
    "epsilon_norm": 7.309453964233398,
    "prediction_norm": 6.550492763519287,
    "timestamp": "2025-07-14T12:35:00.550551"
  },
  {
    "epoch": 2410,
    "batch_idx": 128,
    "loss": 0.28250497579574585,
    "learning_rate": 0.00010015767075645488,
    "timestep_mean": 559.125,
    "epsilon_norm": 8.206602096557617,
    "prediction_norm": 6.956653594970703,
    "timestamp": "2025-07-14T12:35:00.555690"
  },
  {
    "epoch": 2410,
    "batch_idx": 160,
    "loss": 0.3939563035964966,
    "learning_rate": 0.00010015767075645488,
    "timestep_mean": 454.4375,
    "epsilon_norm": 8.591116905212402,
    "prediction_norm": 6.975970268249512,
    "timestamp": "2025-07-14T12:35:00.560903"
  },
  {
    "epoch": 2410,
    "batch_idx": 192,
    "loss": 0.12881752848625183,
    "learning_rate": 0.00010015767075645488,
    "timestep_mean": 529.25,
    "epsilon_norm": 3.636275053024292,
    "prediction_norm": 3.20619797706604,
    "timestamp": "2025-07-14T12:35:00.566148"
  },
  {
    "epoch": 2411,
    "batch_idx": 0,
    "loss": 0.2842608690261841,
    "learning_rate": 0.0001006297327903677,
    "timestep_mean": 611.25,
    "epsilon_norm": 7.93076753616333,
    "prediction_norm": 7.265724182128906,
    "timestamp": "2025-07-14T12:35:00.571261"
  },
  {
    "epoch": 2411,
    "batch_idx": 32,
    "loss": 0.2939344346523285,
    "learning_rate": 0.0001006297327903677,
    "timestep_mean": 502.28125,
    "epsilon_norm": 8.120996475219727,
    "prediction_norm": 7.469336986541748,
    "timestamp": "2025-07-14T12:35:00.576411"
  },
  {
    "epoch": 2411,
    "batch_idx": 64,
    "loss": 0.2554376721382141,
    "learning_rate": 0.0001006297327903677,
    "timestep_mean": 534.0,
    "epsilon_norm": 7.894801616668701,
    "prediction_norm": 6.5591654777526855,
    "timestamp": "2025-07-14T12:35:00.581693"
  },
  {
    "epoch": 2411,
    "batch_idx": 96,
    "loss": 0.2415708601474762,
    "learning_rate": 0.0001006297327903677,
    "timestep_mean": 576.03125,
    "epsilon_norm": 8.329829216003418,
    "prediction_norm": 7.451364517211914,
    "timestamp": "2025-07-14T12:35:00.586930"
  },
  {
    "epoch": 2411,
    "batch_idx": 128,
    "loss": 0.24919435381889343,
    "learning_rate": 0.0001006297327903677,
    "timestep_mean": 406.5625,
    "epsilon_norm": 8.185635566711426,
    "prediction_norm": 7.260318756103516,
    "timestamp": "2025-07-14T12:35:00.592170"
  },
  {
    "epoch": 2411,
    "batch_idx": 160,
    "loss": 0.2812030613422394,
    "learning_rate": 0.0001006297327903677,
    "timestep_mean": 516.40625,
    "epsilon_norm": 8.481144905090332,
    "prediction_norm": 6.874217987060547,
    "timestamp": "2025-07-14T12:35:00.597145"
  },
  {
    "epoch": 2411,
    "batch_idx": 192,
    "loss": 0.20410692691802979,
    "learning_rate": 0.0001006297327903677,
    "timestep_mean": 415.25,
    "epsilon_norm": 3.188474655151367,
    "prediction_norm": 2.822157144546509,
    "timestamp": "2025-07-14T12:35:00.602242"
  },
  {
    "epoch": 2412,
    "batch_idx": 0,
    "loss": 0.24992021918296814,
    "learning_rate": 0.00010110278023071473,
    "timestep_mean": 478.15625,
    "epsilon_norm": 8.232169151306152,
    "prediction_norm": 7.019160747528076,
    "timestamp": "2025-07-14T12:35:00.607252"
  },
  {
    "epoch": 2412,
    "batch_idx": 32,
    "loss": 0.3662160336971283,
    "learning_rate": 0.00010110278023071473,
    "timestep_mean": 445.6875,
    "epsilon_norm": 8.733657836914062,
    "prediction_norm": 7.132213592529297,
    "timestamp": "2025-07-14T12:35:00.612424"
  },
  {
    "epoch": 2412,
    "batch_idx": 64,
    "loss": 0.5094950199127197,
    "learning_rate": 0.00010110278023071473,
    "timestep_mean": 471.21875,
    "epsilon_norm": 8.726168632507324,
    "prediction_norm": 6.402667999267578,
    "timestamp": "2025-07-14T12:35:00.617622"
  },
  {
    "epoch": 2412,
    "batch_idx": 96,
    "loss": 0.3111457824707031,
    "learning_rate": 0.00010110278023071473,
    "timestep_mean": 523.03125,
    "epsilon_norm": 8.626708030700684,
    "prediction_norm": 7.221590995788574,
    "timestamp": "2025-07-14T12:35:00.622728"
  },
  {
    "epoch": 2412,
    "batch_idx": 128,
    "loss": 0.2641681432723999,
    "learning_rate": 0.00010110278023071473,
    "timestep_mean": 531.125,
    "epsilon_norm": 7.923740863800049,
    "prediction_norm": 6.769379615783691,
    "timestamp": "2025-07-14T12:35:00.627947"
  },
  {
    "epoch": 2412,
    "batch_idx": 160,
    "loss": 0.2767602801322937,
    "learning_rate": 0.00010110278023071473,
    "timestep_mean": 477.25,
    "epsilon_norm": 7.9221673011779785,
    "prediction_norm": 6.592893123626709,
    "timestamp": "2025-07-14T12:35:00.633193"
  },
  {
    "epoch": 2412,
    "batch_idx": 192,
    "loss": 0.1850133240222931,
    "learning_rate": 0.00010110278023071473,
    "timestep_mean": 529.0,
    "epsilon_norm": 3.890768527984619,
    "prediction_norm": 3.4443750381469727,
    "timestamp": "2025-07-14T12:35:00.638533"
  },
  {
    "epoch": 2413,
    "batch_idx": 0,
    "loss": 0.2096146047115326,
    "learning_rate": 0.00010157681191029813,
    "timestep_mean": 442.8125,
    "epsilon_norm": 7.610322952270508,
    "prediction_norm": 6.766741752624512,
    "timestamp": "2025-07-14T12:35:00.643585"
  },
  {
    "epoch": 2413,
    "batch_idx": 32,
    "loss": 0.17380757629871368,
    "learning_rate": 0.00010157681191029813,
    "timestep_mean": 507.34375,
    "epsilon_norm": 7.021859645843506,
    "prediction_norm": 6.386390209197998,
    "timestamp": "2025-07-14T12:35:00.648557"
  },
  {
    "epoch": 2413,
    "batch_idx": 64,
    "loss": 0.13073503971099854,
    "learning_rate": 0.00010157681191029813,
    "timestep_mean": 546.125,
    "epsilon_norm": 7.748810768127441,
    "prediction_norm": 6.902493953704834,
    "timestamp": "2025-07-14T12:35:00.653460"
  },
  {
    "epoch": 2413,
    "batch_idx": 96,
    "loss": 0.13694778084754944,
    "learning_rate": 0.00010157681191029813,
    "timestep_mean": 532.4375,
    "epsilon_norm": 7.032306671142578,
    "prediction_norm": 6.533278942108154,
    "timestamp": "2025-07-14T12:35:00.658349"
  },
  {
    "epoch": 2413,
    "batch_idx": 128,
    "loss": 0.4928664565086365,
    "learning_rate": 0.00010157681191029813,
    "timestep_mean": 497.21875,
    "epsilon_norm": 9.898423194885254,
    "prediction_norm": 8.01759147644043,
    "timestamp": "2025-07-14T12:35:00.663289"
  },
  {
    "epoch": 2413,
    "batch_idx": 160,
    "loss": 0.25087276101112366,
    "learning_rate": 0.00010157681191029813,
    "timestep_mean": 511.625,
    "epsilon_norm": 8.788289070129395,
    "prediction_norm": 7.254578113555908,
    "timestamp": "2025-07-14T12:35:00.668116"
  },
  {
    "epoch": 2413,
    "batch_idx": 192,
    "loss": 0.18280184268951416,
    "learning_rate": 0.00010157681191029813,
    "timestep_mean": 430.75,
    "epsilon_norm": 4.88336706161499,
    "prediction_norm": 4.518512725830078,
    "timestamp": "2025-07-14T12:35:00.673026"
  },
  {
    "epoch": 2414,
    "batch_idx": 0,
    "loss": 0.3578554391860962,
    "learning_rate": 0.00010205182665949227,
    "timestep_mean": 509.0,
    "epsilon_norm": 8.457587242126465,
    "prediction_norm": 6.705007076263428,
    "timestamp": "2025-07-14T12:35:00.677890"
  },
  {
    "epoch": 2414,
    "batch_idx": 32,
    "loss": 0.3117564916610718,
    "learning_rate": 0.00010205182665949227,
    "timestep_mean": 502.625,
    "epsilon_norm": 8.027305603027344,
    "prediction_norm": 6.729553699493408,
    "timestamp": "2025-07-14T12:35:00.682809"
  },
  {
    "epoch": 2414,
    "batch_idx": 64,
    "loss": 0.28802716732025146,
    "learning_rate": 0.00010205182665949227,
    "timestep_mean": 416.90625,
    "epsilon_norm": 7.806787490844727,
    "prediction_norm": 6.5346503257751465,
    "timestamp": "2025-07-14T12:35:00.687649"
  },
  {
    "epoch": 2414,
    "batch_idx": 96,
    "loss": 0.24438880383968353,
    "learning_rate": 0.00010205182665949227,
    "timestep_mean": 566.71875,
    "epsilon_norm": 7.637166976928711,
    "prediction_norm": 6.503207206726074,
    "timestamp": "2025-07-14T12:35:00.692543"
  },
  {
    "epoch": 2414,
    "batch_idx": 128,
    "loss": 0.2207646518945694,
    "learning_rate": 0.00010205182665949227,
    "timestep_mean": 517.71875,
    "epsilon_norm": 7.850177764892578,
    "prediction_norm": 6.4156622886657715,
    "timestamp": "2025-07-14T12:35:00.697421"
  },
  {
    "epoch": 2414,
    "batch_idx": 160,
    "loss": 0.2535873353481293,
    "learning_rate": 0.00010205182665949227,
    "timestep_mean": 471.125,
    "epsilon_norm": 7.4066925048828125,
    "prediction_norm": 6.271114349365234,
    "timestamp": "2025-07-14T12:35:00.702372"
  },
  {
    "epoch": 2414,
    "batch_idx": 192,
    "loss": 0.16573360562324524,
    "learning_rate": 0.00010205182665949227,
    "timestep_mean": 513.5,
    "epsilon_norm": 4.051086902618408,
    "prediction_norm": 4.5083160400390625,
    "timestamp": "2025-07-14T12:35:00.707278"
  },
  {
    "epoch": 2415,
    "batch_idx": 0,
    "loss": 0.20204591751098633,
    "learning_rate": 0.00010252782330624528,
    "timestep_mean": 554.8125,
    "epsilon_norm": 8.408102035522461,
    "prediction_norm": 7.255411148071289,
    "timestamp": "2025-07-14T12:35:00.712133"
  },
  {
    "epoch": 2415,
    "batch_idx": 32,
    "loss": 0.4052521288394928,
    "learning_rate": 0.00010252782330624528,
    "timestep_mean": 466.96875,
    "epsilon_norm": 8.493265151977539,
    "prediction_norm": 6.986995697021484,
    "timestamp": "2025-07-14T12:35:00.717121"
  },
  {
    "epoch": 2415,
    "batch_idx": 64,
    "loss": 0.4041377305984497,
    "learning_rate": 0.00010252782330624528,
    "timestep_mean": 489.75,
    "epsilon_norm": 7.7070746421813965,
    "prediction_norm": 6.1074419021606445,
    "timestamp": "2025-07-14T12:35:00.722093"
  },
  {
    "epoch": 2415,
    "batch_idx": 96,
    "loss": 0.3156033158302307,
    "learning_rate": 0.00010252782330624528,
    "timestep_mean": 460.09375,
    "epsilon_norm": 7.67562198638916,
    "prediction_norm": 6.919717788696289,
    "timestamp": "2025-07-14T12:35:00.727100"
  },
  {
    "epoch": 2415,
    "batch_idx": 128,
    "loss": 0.1417873501777649,
    "learning_rate": 0.00010252782330624528,
    "timestep_mean": 607.3125,
    "epsilon_norm": 7.68931245803833,
    "prediction_norm": 7.591168403625488,
    "timestamp": "2025-07-14T12:35:00.731931"
  },
  {
    "epoch": 2415,
    "batch_idx": 160,
    "loss": 0.19366134703159332,
    "learning_rate": 0.00010252782330624528,
    "timestep_mean": 531.1875,
    "epsilon_norm": 8.322529792785645,
    "prediction_norm": 7.626806735992432,
    "timestamp": "2025-07-14T12:35:00.736730"
  },
  {
    "epoch": 2415,
    "batch_idx": 192,
    "loss": 0.024209626019001007,
    "learning_rate": 0.00010252782330624528,
    "timestep_mean": 663.625,
    "epsilon_norm": 3.3698151111602783,
    "prediction_norm": 3.2439959049224854,
    "timestamp": "2025-07-14T12:35:00.741537"
  },
  {
    "epoch": 2416,
    "batch_idx": 0,
    "loss": 0.1171661764383316,
    "learning_rate": 0.00010300480067608243,
    "timestep_mean": 548.875,
    "epsilon_norm": 8.347457885742188,
    "prediction_norm": 7.790773391723633,
    "timestamp": "2025-07-14T12:35:00.746491"
  },
  {
    "epoch": 2416,
    "batch_idx": 32,
    "loss": 0.40110522508621216,
    "learning_rate": 0.00010300480067608243,
    "timestep_mean": 501.90625,
    "epsilon_norm": 7.796219348907471,
    "prediction_norm": 6.634026527404785,
    "timestamp": "2025-07-14T12:35:00.751324"
  },
  {
    "epoch": 2416,
    "batch_idx": 64,
    "loss": 0.194474458694458,
    "learning_rate": 0.00010300480067608243,
    "timestep_mean": 501.75,
    "epsilon_norm": 8.010973930358887,
    "prediction_norm": 6.80350923538208,
    "timestamp": "2025-07-14T12:35:00.756112"
  },
  {
    "epoch": 2416,
    "batch_idx": 96,
    "loss": 0.2446603924036026,
    "learning_rate": 0.00010300480067608243,
    "timestep_mean": 430.3125,
    "epsilon_norm": 8.665923118591309,
    "prediction_norm": 7.595573902130127,
    "timestamp": "2025-07-14T12:35:00.760982"
  },
  {
    "epoch": 2416,
    "batch_idx": 128,
    "loss": 0.18640166521072388,
    "learning_rate": 0.00010300480067608243,
    "timestep_mean": 531.15625,
    "epsilon_norm": 7.460865497589111,
    "prediction_norm": 7.230530261993408,
    "timestamp": "2025-07-14T12:35:00.765842"
  },
  {
    "epoch": 2416,
    "batch_idx": 160,
    "loss": 0.33848685026168823,
    "learning_rate": 0.00010300480067608243,
    "timestep_mean": 464.4375,
    "epsilon_norm": 7.555706977844238,
    "prediction_norm": 6.007552623748779,
    "timestamp": "2025-07-14T12:35:00.770654"
  },
  {
    "epoch": 2416,
    "batch_idx": 192,
    "loss": 0.24480299651622772,
    "learning_rate": 0.00010300480067608243,
    "timestep_mean": 422.125,
    "epsilon_norm": 4.245275020599365,
    "prediction_norm": 4.271617889404297,
    "timestamp": "2025-07-14T12:35:00.775495"
  },
  {
    "epoch": 2417,
    "batch_idx": 0,
    "loss": 0.3011549711227417,
    "learning_rate": 0.00010348275759211015,
    "timestep_mean": 486.9375,
    "epsilon_norm": 7.213940143585205,
    "prediction_norm": 5.613903999328613,
    "timestamp": "2025-07-14T12:35:00.780632"
  },
  {
    "epoch": 2417,
    "batch_idx": 32,
    "loss": 0.2510504126548767,
    "learning_rate": 0.00010348275759211015,
    "timestep_mean": 453.3125,
    "epsilon_norm": 6.727745532989502,
    "prediction_norm": 6.423182964324951,
    "timestamp": "2025-07-14T12:35:00.785735"
  },
  {
    "epoch": 2417,
    "batch_idx": 64,
    "loss": 0.4512869715690613,
    "learning_rate": 0.00010348275759211015,
    "timestep_mean": 487.21875,
    "epsilon_norm": 9.281281471252441,
    "prediction_norm": 7.203700065612793,
    "timestamp": "2025-07-14T12:35:00.790892"
  },
  {
    "epoch": 2417,
    "batch_idx": 96,
    "loss": 0.2770358920097351,
    "learning_rate": 0.00010348275759211015,
    "timestep_mean": 557.6875,
    "epsilon_norm": 7.779147148132324,
    "prediction_norm": 6.994396209716797,
    "timestamp": "2025-07-14T12:35:00.796002"
  },
  {
    "epoch": 2417,
    "batch_idx": 128,
    "loss": 0.10182555019855499,
    "learning_rate": 0.00010348275759211015,
    "timestep_mean": 556.75,
    "epsilon_norm": 7.292599678039551,
    "prediction_norm": 6.540582656860352,
    "timestamp": "2025-07-14T12:35:00.801220"
  },
  {
    "epoch": 2417,
    "batch_idx": 160,
    "loss": 0.3200371563434601,
    "learning_rate": 0.00010348275759211015,
    "timestep_mean": 468.46875,
    "epsilon_norm": 7.736733913421631,
    "prediction_norm": 6.2944865226745605,
    "timestamp": "2025-07-14T12:35:00.806370"
  },
  {
    "epoch": 2417,
    "batch_idx": 192,
    "loss": 0.20093931257724762,
    "learning_rate": 0.00010348275759211015,
    "timestep_mean": 384.75,
    "epsilon_norm": 2.610215902328491,
    "prediction_norm": 2.3410844802856445,
    "timestamp": "2025-07-14T12:35:00.811474"
  },
  {
    "epoch": 2418,
    "batch_idx": 0,
    "loss": 0.1723795235157013,
    "learning_rate": 0.0001039616928750167,
    "timestep_mean": 485.5,
    "epsilon_norm": 7.64994478225708,
    "prediction_norm": 7.043252944946289,
    "timestamp": "2025-07-14T12:35:00.816709"
  },
  {
    "epoch": 2418,
    "batch_idx": 32,
    "loss": 0.4849444627761841,
    "learning_rate": 0.0001039616928750167,
    "timestep_mean": 439.1875,
    "epsilon_norm": 8.579978942871094,
    "prediction_norm": 5.852020263671875,
    "timestamp": "2025-07-14T12:35:00.821944"
  },
  {
    "epoch": 2418,
    "batch_idx": 64,
    "loss": 0.31604671478271484,
    "learning_rate": 0.0001039616928750167,
    "timestep_mean": 419.0625,
    "epsilon_norm": 8.765023231506348,
    "prediction_norm": 7.576828479766846,
    "timestamp": "2025-07-14T12:35:00.827091"
  },
  {
    "epoch": 2418,
    "batch_idx": 96,
    "loss": 0.2559482455253601,
    "learning_rate": 0.0001039616928750167,
    "timestep_mean": 540.21875,
    "epsilon_norm": 8.75033950805664,
    "prediction_norm": 7.2810893058776855,
    "timestamp": "2025-07-14T12:35:00.832326"
  },
  {
    "epoch": 2418,
    "batch_idx": 128,
    "loss": 0.22994837164878845,
    "learning_rate": 0.0001039616928750167,
    "timestep_mean": 474.6875,
    "epsilon_norm": 7.346006870269775,
    "prediction_norm": 6.554113864898682,
    "timestamp": "2025-07-14T12:35:00.837596"
  },
  {
    "epoch": 2418,
    "batch_idx": 160,
    "loss": 0.17746290564537048,
    "learning_rate": 0.0001039616928750167,
    "timestep_mean": 487.34375,
    "epsilon_norm": 8.382678031921387,
    "prediction_norm": 7.278899192810059,
    "timestamp": "2025-07-14T12:35:00.842779"
  },
  {
    "epoch": 2418,
    "batch_idx": 192,
    "loss": 0.2771818935871124,
    "learning_rate": 0.0001039616928750167,
    "timestep_mean": 405.25,
    "epsilon_norm": 4.3794732093811035,
    "prediction_norm": 3.6342601776123047,
    "timestamp": "2025-07-14T12:35:00.847852"
  },
  {
    "epoch": 2419,
    "batch_idx": 0,
    "loss": 0.30147722363471985,
    "learning_rate": 0.0001044416053430771,
    "timestep_mean": 532.0625,
    "epsilon_norm": 7.857204437255859,
    "prediction_norm": 6.847645282745361,
    "timestamp": "2025-07-14T12:35:00.852954"
  },
  {
    "epoch": 2419,
    "batch_idx": 32,
    "loss": 0.24195532500743866,
    "learning_rate": 0.0001044416053430771,
    "timestep_mean": 544.9375,
    "epsilon_norm": 7.348923683166504,
    "prediction_norm": 5.751437187194824,
    "timestamp": "2025-07-14T12:35:00.858056"
  },
  {
    "epoch": 2419,
    "batch_idx": 64,
    "loss": 0.21939192712306976,
    "learning_rate": 0.0001044416053430771,
    "timestep_mean": 588.90625,
    "epsilon_norm": 7.24507999420166,
    "prediction_norm": 6.517640590667725,
    "timestamp": "2025-07-14T12:35:00.863315"
  },
  {
    "epoch": 2419,
    "batch_idx": 96,
    "loss": 0.18307191133499146,
    "learning_rate": 0.0001044416053430771,
    "timestep_mean": 416.625,
    "epsilon_norm": 8.37633991241455,
    "prediction_norm": 7.606081008911133,
    "timestamp": "2025-07-14T12:35:00.868457"
  },
  {
    "epoch": 2419,
    "batch_idx": 128,
    "loss": 0.3432878851890564,
    "learning_rate": 0.0001044416053430771,
    "timestep_mean": 485.5625,
    "epsilon_norm": 8.10605525970459,
    "prediction_norm": 6.347482204437256,
    "timestamp": "2025-07-14T12:35:00.873621"
  },
  {
    "epoch": 2419,
    "batch_idx": 160,
    "loss": 0.1375764161348343,
    "learning_rate": 0.0001044416053430771,
    "timestep_mean": 523.34375,
    "epsilon_norm": 7.999377727508545,
    "prediction_norm": 7.438833236694336,
    "timestamp": "2025-07-14T12:35:00.878856"
  },
  {
    "epoch": 2419,
    "batch_idx": 192,
    "loss": 0.1413775384426117,
    "learning_rate": 0.0001044416053430771,
    "timestep_mean": 609.625,
    "epsilon_norm": 3.6861157417297363,
    "prediction_norm": 3.240532636642456,
    "timestamp": "2025-07-14T12:35:00.884100"
  },
  {
    "epoch": 2420,
    "batch_idx": 0,
    "loss": 0.17361538112163544,
    "learning_rate": 0.0001049224938121551,
    "timestep_mean": 523.28125,
    "epsilon_norm": 5.964662551879883,
    "prediction_norm": 5.927833080291748,
    "timestamp": "2025-07-14T12:35:00.889233"
  },
  {
    "epoch": 2420,
    "batch_idx": 32,
    "loss": 0.15384332835674286,
    "learning_rate": 0.0001049224938121551,
    "timestep_mean": 521.875,
    "epsilon_norm": 7.301962375640869,
    "prediction_norm": 6.536772727966309,
    "timestamp": "2025-07-14T12:35:00.894616"
  },
  {
    "epoch": 2420,
    "batch_idx": 64,
    "loss": 0.3425341248512268,
    "learning_rate": 0.0001049224938121551,
    "timestep_mean": 453.03125,
    "epsilon_norm": 8.308273315429688,
    "prediction_norm": 6.476617336273193,
    "timestamp": "2025-07-14T12:35:00.899814"
  },
  {
    "epoch": 2420,
    "batch_idx": 96,
    "loss": 0.356563001871109,
    "learning_rate": 0.0001049224938121551,
    "timestep_mean": 449.46875,
    "epsilon_norm": 7.741757869720459,
    "prediction_norm": 5.865738868713379,
    "timestamp": "2025-07-14T12:35:00.905062"
  },
  {
    "epoch": 2420,
    "batch_idx": 128,
    "loss": 0.2615455389022827,
    "learning_rate": 0.0001049224938121551,
    "timestep_mean": 489.9375,
    "epsilon_norm": 8.2242431640625,
    "prediction_norm": 7.0548176765441895,
    "timestamp": "2025-07-14T12:35:00.910223"
  },
  {
    "epoch": 2420,
    "batch_idx": 160,
    "loss": 0.2069585621356964,
    "learning_rate": 0.0001049224938121551,
    "timestep_mean": 463.8125,
    "epsilon_norm": 7.6016106605529785,
    "prediction_norm": 6.316904544830322,
    "timestamp": "2025-07-14T12:35:00.915446"
  },
  {
    "epoch": 2420,
    "batch_idx": 192,
    "loss": 0.31457453966140747,
    "learning_rate": 0.0001049224938121551,
    "timestep_mean": 480.5,
    "epsilon_norm": 3.641913890838623,
    "prediction_norm": 2.6047050952911377,
    "timestamp": "2025-07-14T12:35:00.920633"
  },
  {
    "epoch": 2421,
    "batch_idx": 0,
    "loss": 0.18022102117538452,
    "learning_rate": 0.00010540435709570588,
    "timestep_mean": 598.15625,
    "epsilon_norm": 7.901449680328369,
    "prediction_norm": 6.49053430557251,
    "timestamp": "2025-07-14T12:35:00.925974"
  },
  {
    "epoch": 2421,
    "batch_idx": 32,
    "loss": 0.27084264159202576,
    "learning_rate": 0.00010540435709570588,
    "timestep_mean": 525.28125,
    "epsilon_norm": 7.21329927444458,
    "prediction_norm": 6.515506267547607,
    "timestamp": "2025-07-14T12:35:00.931109"
  },
  {
    "epoch": 2421,
    "batch_idx": 64,
    "loss": 0.24470189213752747,
    "learning_rate": 0.00010540435709570588,
    "timestep_mean": 437.53125,
    "epsilon_norm": 8.207477569580078,
    "prediction_norm": 7.302070617675781,
    "timestamp": "2025-07-14T12:35:00.936283"
  },
  {
    "epoch": 2421,
    "batch_idx": 96,
    "loss": 0.37046098709106445,
    "learning_rate": 0.00010540435709570588,
    "timestep_mean": 411.40625,
    "epsilon_norm": 7.201938152313232,
    "prediction_norm": 6.076343059539795,
    "timestamp": "2025-07-14T12:35:00.941344"
  },
  {
    "epoch": 2421,
    "batch_idx": 128,
    "loss": 0.36442285776138306,
    "learning_rate": 0.00010540435709570588,
    "timestep_mean": 557.6875,
    "epsilon_norm": 8.926284790039062,
    "prediction_norm": 7.247888088226318,
    "timestamp": "2025-07-14T12:35:00.946287"
  },
  {
    "epoch": 2421,
    "batch_idx": 160,
    "loss": 0.38261544704437256,
    "learning_rate": 0.00010540435709570588,
    "timestep_mean": 453.9375,
    "epsilon_norm": 8.735871315002441,
    "prediction_norm": 7.0390238761901855,
    "timestamp": "2025-07-14T12:35:00.951154"
  },
  {
    "epoch": 2421,
    "batch_idx": 192,
    "loss": 0.17003239691257477,
    "learning_rate": 0.00010540435709570588,
    "timestep_mean": 375.75,
    "epsilon_norm": 2.818582534790039,
    "prediction_norm": 2.271937608718872,
    "timestamp": "2025-07-14T12:35:00.955914"
  },
  {
    "epoch": 2422,
    "batch_idx": 0,
    "loss": 0.24151462316513062,
    "learning_rate": 0.00010588719400478023,
    "timestep_mean": 467.46875,
    "epsilon_norm": 8.048005104064941,
    "prediction_norm": 6.7261857986450195,
    "timestamp": "2025-07-14T12:35:00.960811"
  },
  {
    "epoch": 2422,
    "batch_idx": 32,
    "loss": 0.3247361481189728,
    "learning_rate": 0.00010588719400478023,
    "timestep_mean": 401.96875,
    "epsilon_norm": 7.995122909545898,
    "prediction_norm": 6.41689395904541,
    "timestamp": "2025-07-14T12:35:00.965660"
  },
  {
    "epoch": 2422,
    "batch_idx": 64,
    "loss": 0.4574877619743347,
    "learning_rate": 0.00010588719400478023,
    "timestep_mean": 470.1875,
    "epsilon_norm": 8.380542755126953,
    "prediction_norm": 6.06109619140625,
    "timestamp": "2025-07-14T12:35:00.970525"
  },
  {
    "epoch": 2422,
    "batch_idx": 96,
    "loss": 0.23984581232070923,
    "learning_rate": 0.00010588719400478023,
    "timestep_mean": 454.09375,
    "epsilon_norm": 7.026443004608154,
    "prediction_norm": 5.824692249298096,
    "timestamp": "2025-07-14T12:35:00.975475"
  },
  {
    "epoch": 2422,
    "batch_idx": 128,
    "loss": 0.39432910084724426,
    "learning_rate": 0.00010588719400478023,
    "timestep_mean": 428.15625,
    "epsilon_norm": 7.092758655548096,
    "prediction_norm": 5.581791877746582,
    "timestamp": "2025-07-14T12:35:00.980311"
  },
  {
    "epoch": 2422,
    "batch_idx": 160,
    "loss": 0.27073606848716736,
    "learning_rate": 0.00010588719400478023,
    "timestep_mean": 476.75,
    "epsilon_norm": 7.108043193817139,
    "prediction_norm": 5.8293867111206055,
    "timestamp": "2025-07-14T12:35:00.985398"
  },
  {
    "epoch": 2422,
    "batch_idx": 192,
    "loss": 0.2090006321668625,
    "learning_rate": 0.00010588719400478023,
    "timestep_mean": 563.0,
    "epsilon_norm": 3.5866339206695557,
    "prediction_norm": 3.5154592990875244,
    "timestamp": "2025-07-14T12:35:00.990479"
  },
  {
    "epoch": 2423,
    "batch_idx": 0,
    "loss": 0.12450288981199265,
    "learning_rate": 0.00010637100334802561,
    "timestep_mean": 581.25,
    "epsilon_norm": 7.5315728187561035,
    "prediction_norm": 7.010408878326416,
    "timestamp": "2025-07-14T12:35:00.995713"
  },
  {
    "epoch": 2423,
    "batch_idx": 32,
    "loss": 0.30240026116371155,
    "learning_rate": 0.00010637100334802561,
    "timestep_mean": 462.1875,
    "epsilon_norm": 8.037276268005371,
    "prediction_norm": 6.539255619049072,
    "timestamp": "2025-07-14T12:35:01.000995"
  },
  {
    "epoch": 2423,
    "batch_idx": 64,
    "loss": 0.26158928871154785,
    "learning_rate": 0.00010637100334802561,
    "timestep_mean": 560.0,
    "epsilon_norm": 8.043583869934082,
    "prediction_norm": 6.706326961517334,
    "timestamp": "2025-07-14T12:35:01.006068"
  },
  {
    "epoch": 2423,
    "batch_idx": 96,
    "loss": 0.1465427279472351,
    "learning_rate": 0.00010637100334802561,
    "timestep_mean": 513.84375,
    "epsilon_norm": 8.12032413482666,
    "prediction_norm": 7.235977649688721,
    "timestamp": "2025-07-14T12:35:01.011403"
  },
  {
    "epoch": 2423,
    "batch_idx": 128,
    "loss": 0.2680352032184601,
    "learning_rate": 0.00010637100334802561,
    "timestep_mean": 497.9375,
    "epsilon_norm": 8.065757751464844,
    "prediction_norm": 6.683845043182373,
    "timestamp": "2025-07-14T12:35:01.016746"
  },
  {
    "epoch": 2423,
    "batch_idx": 160,
    "loss": 0.19547611474990845,
    "learning_rate": 0.00010637100334802561,
    "timestep_mean": 573.4375,
    "epsilon_norm": 7.718096733093262,
    "prediction_norm": 6.64691686630249,
    "timestamp": "2025-07-14T12:35:01.022031"
  },
  {
    "epoch": 2423,
    "batch_idx": 192,
    "loss": 0.3681740164756775,
    "learning_rate": 0.00010637100334802561,
    "timestep_mean": 535.75,
    "epsilon_norm": 4.564181804656982,
    "prediction_norm": 3.194556951522827,
    "timestamp": "2025-07-14T12:35:01.027049"
  },
  {
    "epoch": 2424,
    "batch_idx": 0,
    "loss": 0.2982277572154999,
    "learning_rate": 0.00010685578393169074,
    "timestep_mean": 471.625,
    "epsilon_norm": 8.776469230651855,
    "prediction_norm": 7.6976213455200195,
    "timestamp": "2025-07-14T12:35:01.032196"
  },
  {
    "epoch": 2424,
    "batch_idx": 32,
    "loss": 0.36565470695495605,
    "learning_rate": 0.00010685578393169074,
    "timestep_mean": 516.53125,
    "epsilon_norm": 8.487404823303223,
    "prediction_norm": 7.0672173500061035,
    "timestamp": "2025-07-14T12:35:01.037780"
  },
  {
    "epoch": 2424,
    "batch_idx": 64,
    "loss": 0.4300960898399353,
    "learning_rate": 0.00010685578393169074,
    "timestep_mean": 418.375,
    "epsilon_norm": 8.326248168945312,
    "prediction_norm": 5.964868068695068,
    "timestamp": "2025-07-14T12:35:01.042960"
  },
  {
    "epoch": 2424,
    "batch_idx": 96,
    "loss": 0.2073325216770172,
    "learning_rate": 0.00010685578393169074,
    "timestep_mean": 452.5,
    "epsilon_norm": 8.231049537658691,
    "prediction_norm": 6.889402389526367,
    "timestamp": "2025-07-14T12:35:01.047820"
  },
  {
    "epoch": 2424,
    "batch_idx": 128,
    "loss": 0.2726808786392212,
    "learning_rate": 0.00010685578393169074,
    "timestep_mean": 407.46875,
    "epsilon_norm": 6.475539684295654,
    "prediction_norm": 5.559649467468262,
    "timestamp": "2025-07-14T12:35:01.052791"
  },
  {
    "epoch": 2424,
    "batch_idx": 160,
    "loss": 0.4183835983276367,
    "learning_rate": 0.00010685578393169074,
    "timestep_mean": 495.96875,
    "epsilon_norm": 8.631197929382324,
    "prediction_norm": 6.684288501739502,
    "timestamp": "2025-07-14T12:35:01.057994"
  },
  {
    "epoch": 2424,
    "batch_idx": 192,
    "loss": 0.25261855125427246,
    "learning_rate": 0.00010685578393169074,
    "timestep_mean": 583.875,
    "epsilon_norm": 3.494394540786743,
    "prediction_norm": 3.1087398529052734,
    "timestamp": "2025-07-14T12:35:01.063315"
  },
  {
    "epoch": 2425,
    "batch_idx": 0,
    "loss": 0.3812910318374634,
    "learning_rate": 0.00010734153455962767,
    "timestep_mean": 438.875,
    "epsilon_norm": 8.182028770446777,
    "prediction_norm": 6.598721027374268,
    "timestamp": "2025-07-14T12:35:01.068430"
  },
  {
    "epoch": 2425,
    "batch_idx": 32,
    "loss": 0.20458769798278809,
    "learning_rate": 0.00010734153455962767,
    "timestep_mean": 538.78125,
    "epsilon_norm": 6.8447651863098145,
    "prediction_norm": 5.912698268890381,
    "timestamp": "2025-07-14T12:35:01.073559"
  },
  {
    "epoch": 2425,
    "batch_idx": 64,
    "loss": 0.5795919895172119,
    "learning_rate": 0.00010734153455962767,
    "timestep_mean": 472.53125,
    "epsilon_norm": 7.688173770904541,
    "prediction_norm": 6.006072044372559,
    "timestamp": "2025-07-14T12:35:01.078742"
  },
  {
    "epoch": 2425,
    "batch_idx": 96,
    "loss": 0.37662234902381897,
    "learning_rate": 0.00010734153455962767,
    "timestep_mean": 516.1875,
    "epsilon_norm": 8.5101957321167,
    "prediction_norm": 7.948591232299805,
    "timestamp": "2025-07-14T12:35:01.083942"
  },
  {
    "epoch": 2425,
    "batch_idx": 128,
    "loss": 0.3364994525909424,
    "learning_rate": 0.00010734153455962767,
    "timestep_mean": 520.71875,
    "epsilon_norm": 7.8766326904296875,
    "prediction_norm": 6.474616527557373,
    "timestamp": "2025-07-14T12:35:01.089001"
  },
  {
    "epoch": 2425,
    "batch_idx": 160,
    "loss": 0.1534607708454132,
    "learning_rate": 0.00010734153455962767,
    "timestep_mean": 523.84375,
    "epsilon_norm": 7.131972789764404,
    "prediction_norm": 6.6766676902771,
    "timestamp": "2025-07-14T12:35:01.094236"
  },
  {
    "epoch": 2425,
    "batch_idx": 192,
    "loss": 0.05703877657651901,
    "learning_rate": 0.00010734153455962767,
    "timestep_mean": 689.875,
    "epsilon_norm": 5.150404453277588,
    "prediction_norm": 5.001392841339111,
    "timestamp": "2025-07-14T12:35:01.099410"
  },
  {
    "epoch": 2426,
    "batch_idx": 0,
    "loss": 0.4644671678543091,
    "learning_rate": 0.00010782825403329524,
    "timestep_mean": 462.5,
    "epsilon_norm": 9.086153984069824,
    "prediction_norm": 7.213189601898193,
    "timestamp": "2025-07-14T12:35:01.104742"
  },
  {
    "epoch": 2426,
    "batch_idx": 32,
    "loss": 0.3128061592578888,
    "learning_rate": 0.00010782825403329524,
    "timestep_mean": 506.84375,
    "epsilon_norm": 7.8275980949401855,
    "prediction_norm": 5.412469387054443,
    "timestamp": "2025-07-14T12:35:01.109932"
  },
  {
    "epoch": 2426,
    "batch_idx": 64,
    "loss": 0.40740227699279785,
    "learning_rate": 0.00010782825403329524,
    "timestep_mean": 447.9375,
    "epsilon_norm": 7.844934463500977,
    "prediction_norm": 6.198150634765625,
    "timestamp": "2025-07-14T12:35:01.115094"
  },
  {
    "epoch": 2426,
    "batch_idx": 96,
    "loss": 0.4787270128726959,
    "learning_rate": 0.00010782825403329524,
    "timestep_mean": 428.46875,
    "epsilon_norm": 8.730457305908203,
    "prediction_norm": 6.293288707733154,
    "timestamp": "2025-07-14T12:35:01.120329"
  },
  {
    "epoch": 2426,
    "batch_idx": 128,
    "loss": 0.1297225058078766,
    "learning_rate": 0.00010782825403329524,
    "timestep_mean": 558.59375,
    "epsilon_norm": 6.7462615966796875,
    "prediction_norm": 6.176027774810791,
    "timestamp": "2025-07-14T12:35:01.125560"
  },
  {
    "epoch": 2426,
    "batch_idx": 160,
    "loss": 0.20840588212013245,
    "learning_rate": 0.00010782825403329524,
    "timestep_mean": 478.5625,
    "epsilon_norm": 7.155029296875,
    "prediction_norm": 5.876713752746582,
    "timestamp": "2025-07-14T12:35:01.130691"
  },
  {
    "epoch": 2426,
    "batch_idx": 192,
    "loss": 0.2150895595550537,
    "learning_rate": 0.00010782825403329524,
    "timestep_mean": 506.375,
    "epsilon_norm": 3.6078133583068848,
    "prediction_norm": 3.149801731109619,
    "timestamp": "2025-07-14T12:35:01.135901"
  },
  {
    "epoch": 2427,
    "batch_idx": 0,
    "loss": 0.14314720034599304,
    "learning_rate": 0.00010831594115176116,
    "timestep_mean": 586.28125,
    "epsilon_norm": 6.501591205596924,
    "prediction_norm": 6.041738510131836,
    "timestamp": "2025-07-14T12:35:01.141022"
  },
  {
    "epoch": 2427,
    "batch_idx": 32,
    "loss": 0.19483253359794617,
    "learning_rate": 0.00010831594115176116,
    "timestep_mean": 418.65625,
    "epsilon_norm": 7.2568135261535645,
    "prediction_norm": 6.1580400466918945,
    "timestamp": "2025-07-14T12:35:01.146140"
  },
  {
    "epoch": 2427,
    "batch_idx": 64,
    "loss": 0.36371347308158875,
    "learning_rate": 0.00010831594115176116,
    "timestep_mean": 586.1875,
    "epsilon_norm": 8.624940872192383,
    "prediction_norm": 7.1364946365356445,
    "timestamp": "2025-07-14T12:35:01.151305"
  },
  {
    "epoch": 2427,
    "batch_idx": 96,
    "loss": 0.22194677591323853,
    "learning_rate": 0.00010831594115176116,
    "timestep_mean": 558.5,
    "epsilon_norm": 8.161293029785156,
    "prediction_norm": 7.033824920654297,
    "timestamp": "2025-07-14T12:35:01.156420"
  },
  {
    "epoch": 2427,
    "batch_idx": 128,
    "loss": 0.3675253093242645,
    "learning_rate": 0.00010831594115176116,
    "timestep_mean": 449.1875,
    "epsilon_norm": 8.034912109375,
    "prediction_norm": 6.8403730392456055,
    "timestamp": "2025-07-14T12:35:01.161598"
  },
  {
    "epoch": 2427,
    "batch_idx": 160,
    "loss": 0.22639769315719604,
    "learning_rate": 0.00010831594115176116,
    "timestep_mean": 483.15625,
    "epsilon_norm": 8.0400390625,
    "prediction_norm": 6.842855453491211,
    "timestamp": "2025-07-14T12:35:01.166731"
  },
  {
    "epoch": 2427,
    "batch_idx": 192,
    "loss": 0.22923821210861206,
    "learning_rate": 0.00010831594115176116,
    "timestep_mean": 636.125,
    "epsilon_norm": 5.238209247589111,
    "prediction_norm": 4.115729331970215,
    "timestamp": "2025-07-14T12:35:01.171794"
  },
  {
    "epoch": 2428,
    "batch_idx": 0,
    "loss": 0.2854178547859192,
    "learning_rate": 0.00010880459471170615,
    "timestep_mean": 478.40625,
    "epsilon_norm": 8.06322193145752,
    "prediction_norm": 6.602194786071777,
    "timestamp": "2025-07-14T12:35:01.176858"
  },
  {
    "epoch": 2428,
    "batch_idx": 32,
    "loss": 0.11483180522918701,
    "learning_rate": 0.00010880459471170615,
    "timestep_mean": 477.96875,
    "epsilon_norm": 6.542126178741455,
    "prediction_norm": 6.310992240905762,
    "timestamp": "2025-07-14T12:35:01.181986"
  },
  {
    "epoch": 2428,
    "batch_idx": 64,
    "loss": 0.28545111417770386,
    "learning_rate": 0.00010880459471170615,
    "timestep_mean": 420.78125,
    "epsilon_norm": 7.789060115814209,
    "prediction_norm": 6.745673179626465,
    "timestamp": "2025-07-14T12:35:01.187186"
  },
  {
    "epoch": 2428,
    "batch_idx": 96,
    "loss": 0.17926821112632751,
    "learning_rate": 0.00010880459471170615,
    "timestep_mean": 518.3125,
    "epsilon_norm": 7.1831464767456055,
    "prediction_norm": 6.761910438537598,
    "timestamp": "2025-07-14T12:35:01.192456"
  },
  {
    "epoch": 2428,
    "batch_idx": 128,
    "loss": 0.38215747475624084,
    "learning_rate": 0.00010880459471170615,
    "timestep_mean": 557.0625,
    "epsilon_norm": 7.274737358093262,
    "prediction_norm": 5.856569290161133,
    "timestamp": "2025-07-14T12:35:01.197555"
  },
  {
    "epoch": 2428,
    "batch_idx": 160,
    "loss": 0.15393763780593872,
    "learning_rate": 0.00010880459471170615,
    "timestep_mean": 602.0625,
    "epsilon_norm": 9.973426818847656,
    "prediction_norm": 8.938879013061523,
    "timestamp": "2025-07-14T12:35:01.202664"
  },
  {
    "epoch": 2428,
    "batch_idx": 192,
    "loss": 0.17168861627578735,
    "learning_rate": 0.00010880459471170615,
    "timestep_mean": 531.375,
    "epsilon_norm": 3.533090591430664,
    "prediction_norm": 3.3970208168029785,
    "timestamp": "2025-07-14T12:35:01.207831"
  },
  {
    "epoch": 2429,
    "batch_idx": 0,
    "loss": 0.17981958389282227,
    "learning_rate": 0.00010929421350742622,
    "timestep_mean": 490.625,
    "epsilon_norm": 6.463651180267334,
    "prediction_norm": 5.7172651290893555,
    "timestamp": "2025-07-14T12:35:01.213054"
  },
  {
    "epoch": 2429,
    "batch_idx": 32,
    "loss": 0.1863868534564972,
    "learning_rate": 0.00010929421350742622,
    "timestep_mean": 574.65625,
    "epsilon_norm": 8.572179794311523,
    "prediction_norm": 7.6800537109375,
    "timestamp": "2025-07-14T12:35:01.218197"
  },
  {
    "epoch": 2429,
    "batch_idx": 64,
    "loss": 0.4723183810710907,
    "learning_rate": 0.00010929421350742622,
    "timestep_mean": 449.75,
    "epsilon_norm": 8.577592849731445,
    "prediction_norm": 6.627254962921143,
    "timestamp": "2025-07-14T12:35:01.223328"
  },
  {
    "epoch": 2429,
    "batch_idx": 96,
    "loss": 0.23202356696128845,
    "learning_rate": 0.00010929421350742622,
    "timestep_mean": 506.0,
    "epsilon_norm": 8.17085075378418,
    "prediction_norm": 6.749828338623047,
    "timestamp": "2025-07-14T12:35:01.228481"
  },
  {
    "epoch": 2429,
    "batch_idx": 128,
    "loss": 0.20799878239631653,
    "learning_rate": 0.00010929421350742622,
    "timestep_mean": 459.40625,
    "epsilon_norm": 7.678086280822754,
    "prediction_norm": 6.5595197677612305,
    "timestamp": "2025-07-14T12:35:01.233648"
  },
  {
    "epoch": 2429,
    "batch_idx": 160,
    "loss": 0.3122156858444214,
    "learning_rate": 0.00010929421350742622,
    "timestep_mean": 422.65625,
    "epsilon_norm": 8.031291961669922,
    "prediction_norm": 6.739512920379639,
    "timestamp": "2025-07-14T12:35:01.238730"
  },
  {
    "epoch": 2429,
    "batch_idx": 192,
    "loss": 0.18532736599445343,
    "learning_rate": 0.00010929421350742622,
    "timestep_mean": 494.25,
    "epsilon_norm": 5.127477645874023,
    "prediction_norm": 4.653006553649902,
    "timestamp": "2025-07-14T12:35:01.243802"
  },
  {
    "epoch": 2430,
    "batch_idx": 0,
    "loss": 0.3914834260940552,
    "learning_rate": 0.00010978479633083525,
    "timestep_mean": 549.5625,
    "epsilon_norm": 8.794403076171875,
    "prediction_norm": 6.879510879516602,
    "timestamp": "2025-07-14T12:35:01.249065"
  },
  {
    "epoch": 2430,
    "batch_idx": 32,
    "loss": 0.1582680344581604,
    "learning_rate": 0.00010978479633083525,
    "timestep_mean": 524.28125,
    "epsilon_norm": 7.51963996887207,
    "prediction_norm": 6.647556781768799,
    "timestamp": "2025-07-14T12:35:01.254290"
  },
  {
    "epoch": 2430,
    "batch_idx": 64,
    "loss": 0.2749083340167999,
    "learning_rate": 0.00010978479633083525,
    "timestep_mean": 473.09375,
    "epsilon_norm": 7.093822479248047,
    "prediction_norm": 6.087417125701904,
    "timestamp": "2025-07-14T12:35:01.259476"
  },
  {
    "epoch": 2430,
    "batch_idx": 96,
    "loss": 0.2297828048467636,
    "learning_rate": 0.00010978479633083525,
    "timestep_mean": 536.21875,
    "epsilon_norm": 7.595180988311768,
    "prediction_norm": 6.648374080657959,
    "timestamp": "2025-07-14T12:35:01.264619"
  },
  {
    "epoch": 2430,
    "batch_idx": 128,
    "loss": 0.3374040722846985,
    "learning_rate": 0.00010978479633083525,
    "timestep_mean": 431.40625,
    "epsilon_norm": 8.154936790466309,
    "prediction_norm": 6.230671405792236,
    "timestamp": "2025-07-14T12:35:01.269830"
  },
  {
    "epoch": 2430,
    "batch_idx": 160,
    "loss": 0.27776044607162476,
    "learning_rate": 0.00010978479633083525,
    "timestep_mean": 567.5625,
    "epsilon_norm": 8.942309379577637,
    "prediction_norm": 8.077136993408203,
    "timestamp": "2025-07-14T12:35:01.275094"
  },
  {
    "epoch": 2430,
    "batch_idx": 192,
    "loss": 0.06167189031839371,
    "learning_rate": 0.00010978479633083525,
    "timestep_mean": 547.0,
    "epsilon_norm": 2.7685272693634033,
    "prediction_norm": 2.448857307434082,
    "timestamp": "2025-07-14T12:35:01.280158"
  },
  {
    "epoch": 2431,
    "batch_idx": 0,
    "loss": 0.22810481488704681,
    "learning_rate": 0.00011027634197146949,
    "timestep_mean": 534.9375,
    "epsilon_norm": 8.203697204589844,
    "prediction_norm": 7.316473484039307,
    "timestamp": "2025-07-14T12:35:01.285326"
  },
  {
    "epoch": 2431,
    "batch_idx": 32,
    "loss": 0.24188891053199768,
    "learning_rate": 0.00011027634197146949,
    "timestep_mean": 507.21875,
    "epsilon_norm": 7.081907749176025,
    "prediction_norm": 6.041755199432373,
    "timestamp": "2025-07-14T12:35:01.290529"
  },
  {
    "epoch": 2431,
    "batch_idx": 64,
    "loss": 0.17648610472679138,
    "learning_rate": 0.00011027634197146949,
    "timestep_mean": 519.0625,
    "epsilon_norm": 7.195754051208496,
    "prediction_norm": 6.99506139755249,
    "timestamp": "2025-07-14T12:35:01.295698"
  },
  {
    "epoch": 2431,
    "batch_idx": 96,
    "loss": 0.4176795184612274,
    "learning_rate": 0.00011027634197146949,
    "timestep_mean": 446.34375,
    "epsilon_norm": 8.93125057220459,
    "prediction_norm": 7.472376823425293,
    "timestamp": "2025-07-14T12:35:01.301022"
  },
  {
    "epoch": 2431,
    "batch_idx": 128,
    "loss": 0.3971090316772461,
    "learning_rate": 0.00011027634197146949,
    "timestep_mean": 476.78125,
    "epsilon_norm": 8.12651538848877,
    "prediction_norm": 6.145733833312988,
    "timestamp": "2025-07-14T12:35:01.306188"
  },
  {
    "epoch": 2431,
    "batch_idx": 160,
    "loss": 0.2362777590751648,
    "learning_rate": 0.00011027634197146949,
    "timestep_mean": 392.0,
    "epsilon_norm": 7.531215667724609,
    "prediction_norm": 6.7052202224731445,
    "timestamp": "2025-07-14T12:35:01.311278"
  },
  {
    "epoch": 2431,
    "batch_idx": 192,
    "loss": 0.12042434513568878,
    "learning_rate": 0.00011027634197146949,
    "timestep_mean": 521.75,
    "epsilon_norm": 4.145560264587402,
    "prediction_norm": 3.813537359237671,
    "timestamp": "2025-07-14T12:35:01.316349"
  },
  {
    "epoch": 2432,
    "batch_idx": 0,
    "loss": 0.22727897763252258,
    "learning_rate": 0.00011076884921648846,
    "timestep_mean": 464.625,
    "epsilon_norm": 7.329292297363281,
    "prediction_norm": 6.51146125793457,
    "timestamp": "2025-07-14T12:35:01.321549"
  },
  {
    "epoch": 2432,
    "batch_idx": 32,
    "loss": 0.29405316710472107,
    "learning_rate": 0.00011076884921648846,
    "timestep_mean": 530.78125,
    "epsilon_norm": 7.880647659301758,
    "prediction_norm": 7.3458662033081055,
    "timestamp": "2025-07-14T12:35:01.326422"
  },
  {
    "epoch": 2432,
    "batch_idx": 64,
    "loss": 0.21825936436653137,
    "learning_rate": 0.00011076884921648846,
    "timestep_mean": 446.90625,
    "epsilon_norm": 8.163544654846191,
    "prediction_norm": 7.421761512756348,
    "timestamp": "2025-07-14T12:35:01.331291"
  },
  {
    "epoch": 2432,
    "batch_idx": 96,
    "loss": 0.12755879759788513,
    "learning_rate": 0.00011076884921648846,
    "timestep_mean": 550.0,
    "epsilon_norm": 8.400712013244629,
    "prediction_norm": 7.7212982177734375,
    "timestamp": "2025-07-14T12:35:01.336161"
  },
  {
    "epoch": 2432,
    "batch_idx": 128,
    "loss": 0.14156755805015564,
    "learning_rate": 0.00011076884921648846,
    "timestep_mean": 489.75,
    "epsilon_norm": 7.21070671081543,
    "prediction_norm": 7.108170032501221,
    "timestamp": "2025-07-14T12:35:01.340999"
  },
  {
    "epoch": 2432,
    "batch_idx": 160,
    "loss": 0.2218513786792755,
    "learning_rate": 0.00011076884921648846,
    "timestep_mean": 520.59375,
    "epsilon_norm": 8.168035507202148,
    "prediction_norm": 7.160158157348633,
    "timestamp": "2025-07-14T12:35:01.345880"
  },
  {
    "epoch": 2432,
    "batch_idx": 192,
    "loss": 0.22266337275505066,
    "learning_rate": 0.00011076884921648846,
    "timestep_mean": 387.75,
    "epsilon_norm": 3.796708106994629,
    "prediction_norm": 3.414203643798828,
    "timestamp": "2025-07-14T12:35:01.350864"
  },
  {
    "epoch": 2433,
    "batch_idx": 0,
    "loss": 0.21140122413635254,
    "learning_rate": 0.00011126231685067969,
    "timestep_mean": 552.90625,
    "epsilon_norm": 7.611698627471924,
    "prediction_norm": 6.678571701049805,
    "timestamp": "2025-07-14T12:35:01.355669"
  },
  {
    "epoch": 2433,
    "batch_idx": 32,
    "loss": 0.16633902490139008,
    "learning_rate": 0.00011126231685067969,
    "timestep_mean": 597.65625,
    "epsilon_norm": 8.933040618896484,
    "prediction_norm": 7.996068954467773,
    "timestamp": "2025-07-14T12:35:01.360532"
  },
  {
    "epoch": 2433,
    "batch_idx": 64,
    "loss": 0.15771153569221497,
    "learning_rate": 0.00011126231685067969,
    "timestep_mean": 551.4375,
    "epsilon_norm": 8.175945281982422,
    "prediction_norm": 7.415250301361084,
    "timestamp": "2025-07-14T12:35:01.365481"
  },
  {
    "epoch": 2433,
    "batch_idx": 96,
    "loss": 0.21178306639194489,
    "learning_rate": 0.00011126231685067969,
    "timestep_mean": 528.25,
    "epsilon_norm": 7.028938293457031,
    "prediction_norm": 6.207857131958008,
    "timestamp": "2025-07-14T12:35:01.370293"
  },
  {
    "epoch": 2433,
    "batch_idx": 128,
    "loss": 0.2094087302684784,
    "learning_rate": 0.00011126231685067969,
    "timestep_mean": 508.9375,
    "epsilon_norm": 6.544493675231934,
    "prediction_norm": 5.4487080574035645,
    "timestamp": "2025-07-14T12:35:01.375129"
  },
  {
    "epoch": 2433,
    "batch_idx": 160,
    "loss": 0.24180467426776886,
    "learning_rate": 0.00011126231685067969,
    "timestep_mean": 582.90625,
    "epsilon_norm": 9.090378761291504,
    "prediction_norm": 7.933424949645996,
    "timestamp": "2025-07-14T12:35:01.380040"
  },
  {
    "epoch": 2433,
    "batch_idx": 192,
    "loss": 0.22099965810775757,
    "learning_rate": 0.00011126231685067969,
    "timestep_mean": 471.5,
    "epsilon_norm": 4.471932888031006,
    "prediction_norm": 3.894029140472412,
    "timestamp": "2025-07-14T12:35:01.384838"
  },
  {
    "epoch": 2434,
    "batch_idx": 0,
    "loss": 0.20034857094287872,
    "learning_rate": 0.00011175674365646096,
    "timestep_mean": 569.75,
    "epsilon_norm": 7.743480682373047,
    "prediction_norm": 6.643539905548096,
    "timestamp": "2025-07-14T12:35:01.389681"
  },
  {
    "epoch": 2434,
    "batch_idx": 32,
    "loss": 0.20822495222091675,
    "learning_rate": 0.00011175674365646096,
    "timestep_mean": 577.0,
    "epsilon_norm": 7.2366719245910645,
    "prediction_norm": 6.580783367156982,
    "timestamp": "2025-07-14T12:35:01.394507"
  },
  {
    "epoch": 2434,
    "batch_idx": 64,
    "loss": 0.2569900155067444,
    "learning_rate": 0.00011175674365646096,
    "timestep_mean": 464.0,
    "epsilon_norm": 7.69481086730957,
    "prediction_norm": 6.67152738571167,
    "timestamp": "2025-07-14T12:35:01.399417"
  },
  {
    "epoch": 2434,
    "batch_idx": 96,
    "loss": 0.28749188780784607,
    "learning_rate": 0.00011175674365646096,
    "timestep_mean": 460.71875,
    "epsilon_norm": 7.236572265625,
    "prediction_norm": 6.051323413848877,
    "timestamp": "2025-07-14T12:35:01.404237"
  },
  {
    "epoch": 2434,
    "batch_idx": 128,
    "loss": 0.24678300321102142,
    "learning_rate": 0.00011175674365646096,
    "timestep_mean": 499.09375,
    "epsilon_norm": 7.418979644775391,
    "prediction_norm": 6.234973430633545,
    "timestamp": "2025-07-14T12:35:01.409172"
  },
  {
    "epoch": 2434,
    "batch_idx": 160,
    "loss": 0.2702164649963379,
    "learning_rate": 0.00011175674365646096,
    "timestep_mean": 541.0625,
    "epsilon_norm": 8.041633605957031,
    "prediction_norm": 6.802690029144287,
    "timestamp": "2025-07-14T12:35:01.414054"
  },
  {
    "epoch": 2434,
    "batch_idx": 192,
    "loss": 0.1497611105442047,
    "learning_rate": 0.00011175674365646096,
    "timestep_mean": 483.125,
    "epsilon_norm": 3.1082847118377686,
    "prediction_norm": 3.0307483673095703,
    "timestamp": "2025-07-14T12:35:01.419048"
  },
  {
    "epoch": 2435,
    "batch_idx": 0,
    "loss": 0.08249035477638245,
    "learning_rate": 0.0001122521284138829,
    "timestep_mean": 585.5625,
    "epsilon_norm": 7.950531005859375,
    "prediction_norm": 7.826835632324219,
    "timestamp": "2025-07-14T12:35:01.424245"
  },
  {
    "epoch": 2435,
    "batch_idx": 32,
    "loss": 0.41484400629997253,
    "learning_rate": 0.0001122521284138829,
    "timestep_mean": 408.34375,
    "epsilon_norm": 8.167344093322754,
    "prediction_norm": 6.390890121459961,
    "timestamp": "2025-07-14T12:35:01.429475"
  },
  {
    "epoch": 2435,
    "batch_idx": 64,
    "loss": 0.1732112020254135,
    "learning_rate": 0.0001122521284138829,
    "timestep_mean": 547.625,
    "epsilon_norm": 8.27012825012207,
    "prediction_norm": 7.298457145690918,
    "timestamp": "2025-07-14T12:35:01.434694"
  },
  {
    "epoch": 2435,
    "batch_idx": 96,
    "loss": 0.24144867062568665,
    "learning_rate": 0.0001122521284138829,
    "timestep_mean": 579.34375,
    "epsilon_norm": 8.135031700134277,
    "prediction_norm": 7.072062015533447,
    "timestamp": "2025-07-14T12:35:01.439842"
  },
  {
    "epoch": 2435,
    "batch_idx": 128,
    "loss": 0.16008973121643066,
    "learning_rate": 0.0001122521284138829,
    "timestep_mean": 464.125,
    "epsilon_norm": 7.047712326049805,
    "prediction_norm": 6.247639179229736,
    "timestamp": "2025-07-14T12:35:01.444912"
  },
  {
    "epoch": 2435,
    "batch_idx": 160,
    "loss": 0.32431238889694214,
    "learning_rate": 0.0001122521284138829,
    "timestep_mean": 551.34375,
    "epsilon_norm": 8.46070384979248,
    "prediction_norm": 6.913598537445068,
    "timestamp": "2025-07-14T12:35:01.450038"
  },
  {
    "epoch": 2435,
    "batch_idx": 192,
    "loss": 0.18606089055538177,
    "learning_rate": 0.0001122521284138829,
    "timestep_mean": 625.75,
    "epsilon_norm": 4.151069641113281,
    "prediction_norm": 3.722560167312622,
    "timestamp": "2025-07-14T12:35:01.455181"
  },
  {
    "epoch": 2436,
    "batch_idx": 0,
    "loss": 0.23153537511825562,
    "learning_rate": 0.00011274846990063328,
    "timestep_mean": 508.46875,
    "epsilon_norm": 6.8516926765441895,
    "prediction_norm": 5.531863212585449,
    "timestamp": "2025-07-14T12:35:01.460386"
  },
  {
    "epoch": 2436,
    "batch_idx": 32,
    "loss": 0.19670596718788147,
    "learning_rate": 0.00011274846990063328,
    "timestep_mean": 438.59375,
    "epsilon_norm": 7.228754043579102,
    "prediction_norm": 6.246798992156982,
    "timestamp": "2025-07-14T12:35:01.465577"
  },
  {
    "epoch": 2436,
    "batch_idx": 64,
    "loss": 0.19180333614349365,
    "learning_rate": 0.00011274846990063328,
    "timestep_mean": 481.46875,
    "epsilon_norm": 7.796327590942383,
    "prediction_norm": 6.644285678863525,
    "timestamp": "2025-07-14T12:35:01.470791"
  },
  {
    "epoch": 2436,
    "batch_idx": 96,
    "loss": 0.14608515799045563,
    "learning_rate": 0.00011274846990063328,
    "timestep_mean": 552.6875,
    "epsilon_norm": 8.037858009338379,
    "prediction_norm": 7.523731231689453,
    "timestamp": "2025-07-14T12:35:01.475993"
  },
  {
    "epoch": 2436,
    "batch_idx": 128,
    "loss": 0.25937405228614807,
    "learning_rate": 0.00011274846990063328,
    "timestep_mean": 477.0625,
    "epsilon_norm": 7.325802803039551,
    "prediction_norm": 6.032642841339111,
    "timestamp": "2025-07-14T12:35:01.481204"
  },
  {
    "epoch": 2436,
    "batch_idx": 160,
    "loss": 0.22377178072929382,
    "learning_rate": 0.00011274846990063328,
    "timestep_mean": 483.78125,
    "epsilon_norm": 8.203620910644531,
    "prediction_norm": 7.107266426086426,
    "timestamp": "2025-07-14T12:35:01.486135"
  },
  {
    "epoch": 2436,
    "batch_idx": 192,
    "loss": 0.13976603746414185,
    "learning_rate": 0.00011274846990063328,
    "timestep_mean": 630.25,
    "epsilon_norm": 4.4313273429870605,
    "prediction_norm": 3.931638717651367,
    "timestamp": "2025-07-14T12:35:01.490930"
  },
  {
    "epoch": 2437,
    "batch_idx": 0,
    "loss": 0.16899268329143524,
    "learning_rate": 0.00011324576689203855,
    "timestep_mean": 508.25,
    "epsilon_norm": 7.2503461837768555,
    "prediction_norm": 6.145979881286621,
    "timestamp": "2025-07-14T12:35:01.495773"
  },
  {
    "epoch": 2437,
    "batch_idx": 32,
    "loss": 0.33584532141685486,
    "learning_rate": 0.00011324576689203855,
    "timestep_mean": 500.375,
    "epsilon_norm": 7.675613880157471,
    "prediction_norm": 5.885526180267334,
    "timestamp": "2025-07-14T12:35:01.500716"
  },
  {
    "epoch": 2437,
    "batch_idx": 64,
    "loss": 0.20335105061531067,
    "learning_rate": 0.00011324576689203855,
    "timestep_mean": 465.4375,
    "epsilon_norm": 8.147993087768555,
    "prediction_norm": 7.001682281494141,
    "timestamp": "2025-07-14T12:35:01.505827"
  },
  {
    "epoch": 2437,
    "batch_idx": 96,
    "loss": 0.18747735023498535,
    "learning_rate": 0.00011324576689203855,
    "timestep_mean": 515.34375,
    "epsilon_norm": 8.32800579071045,
    "prediction_norm": 7.148656368255615,
    "timestamp": "2025-07-14T12:35:01.510919"
  },
  {
    "epoch": 2437,
    "batch_idx": 128,
    "loss": 0.2658109664916992,
    "learning_rate": 0.00011324576689203855,
    "timestep_mean": 536.34375,
    "epsilon_norm": 8.275074005126953,
    "prediction_norm": 7.54617166519165,
    "timestamp": "2025-07-14T12:35:01.516114"
  },
  {
    "epoch": 2437,
    "batch_idx": 160,
    "loss": 0.3829159736633301,
    "learning_rate": 0.00011324576689203855,
    "timestep_mean": 499.5,
    "epsilon_norm": 8.138551712036133,
    "prediction_norm": 6.99308967590332,
    "timestamp": "2025-07-14T12:35:01.521266"
  },
  {
    "epoch": 2437,
    "batch_idx": 192,
    "loss": 0.1525241732597351,
    "learning_rate": 0.00011324576689203855,
    "timestep_mean": 566.625,
    "epsilon_norm": 4.315263271331787,
    "prediction_norm": 3.8509726524353027,
    "timestamp": "2025-07-14T12:35:01.526118"
  },
  {
    "epoch": 2438,
    "batch_idx": 0,
    "loss": 0.14006777107715607,
    "learning_rate": 0.00011374401816106796,
    "timestep_mean": 558.59375,
    "epsilon_norm": 7.6695146560668945,
    "prediction_norm": 7.306829452514648,
    "timestamp": "2025-07-14T12:35:01.531000"
  },
  {
    "epoch": 2438,
    "batch_idx": 32,
    "loss": 0.345823734998703,
    "learning_rate": 0.00011374401816106796,
    "timestep_mean": 463.53125,
    "epsilon_norm": 8.201701164245605,
    "prediction_norm": 6.640706539154053,
    "timestamp": "2025-07-14T12:35:01.535849"
  },
  {
    "epoch": 2438,
    "batch_idx": 64,
    "loss": 0.32829323410987854,
    "learning_rate": 0.00011374401816106796,
    "timestep_mean": 540.09375,
    "epsilon_norm": 8.224358558654785,
    "prediction_norm": 7.213184356689453,
    "timestamp": "2025-07-14T12:35:01.540771"
  },
  {
    "epoch": 2438,
    "batch_idx": 96,
    "loss": 0.29845869541168213,
    "learning_rate": 0.00011374401816106796,
    "timestep_mean": 478.59375,
    "epsilon_norm": 7.6034417152404785,
    "prediction_norm": 6.029084205627441,
    "timestamp": "2025-07-14T12:35:01.545629"
  },
  {
    "epoch": 2438,
    "batch_idx": 128,
    "loss": 0.12383396178483963,
    "learning_rate": 0.00011374401816106796,
    "timestep_mean": 580.375,
    "epsilon_norm": 7.672112464904785,
    "prediction_norm": 7.010031223297119,
    "timestamp": "2025-07-14T12:35:01.550689"
  },
  {
    "epoch": 2438,
    "batch_idx": 160,
    "loss": 0.17598553001880646,
    "learning_rate": 0.00011374401816106796,
    "timestep_mean": 552.375,
    "epsilon_norm": 7.203334808349609,
    "prediction_norm": 6.009214401245117,
    "timestamp": "2025-07-14T12:35:01.555738"
  },
  {
    "epoch": 2438,
    "batch_idx": 192,
    "loss": 0.8294931650161743,
    "learning_rate": 0.00011374401816106796,
    "timestep_mean": 299.375,
    "epsilon_norm": 5.149986743927002,
    "prediction_norm": 3.5075812339782715,
    "timestamp": "2025-07-14T12:35:01.560858"
  },
  {
    "epoch": 2439,
    "batch_idx": 0,
    "loss": 0.2683534324169159,
    "learning_rate": 0.00011424322247833604,
    "timestep_mean": 542.96875,
    "epsilon_norm": 8.226052284240723,
    "prediction_norm": 6.845093250274658,
    "timestamp": "2025-07-14T12:35:01.566029"
  },
  {
    "epoch": 2439,
    "batch_idx": 32,
    "loss": 0.22832077741622925,
    "learning_rate": 0.00011424322247833604,
    "timestep_mean": 487.03125,
    "epsilon_norm": 8.500219345092773,
    "prediction_norm": 7.3384928703308105,
    "timestamp": "2025-07-14T12:35:01.571273"
  },
  {
    "epoch": 2439,
    "batch_idx": 64,
    "loss": 0.2013191282749176,
    "learning_rate": 0.00011424322247833604,
    "timestep_mean": 515.09375,
    "epsilon_norm": 6.647977828979492,
    "prediction_norm": 5.1211042404174805,
    "timestamp": "2025-07-14T12:35:01.576473"
  },
  {
    "epoch": 2439,
    "batch_idx": 96,
    "loss": 0.2353399693965912,
    "learning_rate": 0.00011424322247833604,
    "timestep_mean": 509.8125,
    "epsilon_norm": 8.28689956665039,
    "prediction_norm": 6.975816249847412,
    "timestamp": "2025-07-14T12:35:01.581692"
  },
  {
    "epoch": 2439,
    "batch_idx": 128,
    "loss": 0.2030852884054184,
    "learning_rate": 0.00011424322247833604,
    "timestep_mean": 482.0625,
    "epsilon_norm": 7.8138837814331055,
    "prediction_norm": 7.364253520965576,
    "timestamp": "2025-07-14T12:35:01.586828"
  },
  {
    "epoch": 2439,
    "batch_idx": 160,
    "loss": 0.30449938774108887,
    "learning_rate": 0.00011424322247833604,
    "timestep_mean": 543.84375,
    "epsilon_norm": 8.066267013549805,
    "prediction_norm": 6.501518726348877,
    "timestamp": "2025-07-14T12:35:01.592030"
  },
  {
    "epoch": 2439,
    "batch_idx": 192,
    "loss": 0.054512668401002884,
    "learning_rate": 0.00011424322247833604,
    "timestep_mean": 676.75,
    "epsilon_norm": 3.7226803302764893,
    "prediction_norm": 3.3582448959350586,
    "timestamp": "2025-07-14T12:35:01.597132"
  },
  {
    "epoch": 2440,
    "batch_idx": 0,
    "loss": 0.2583111822605133,
    "learning_rate": 0.00011474337861210564,
    "timestep_mean": 466.375,
    "epsilon_norm": 7.783985137939453,
    "prediction_norm": 6.408865928649902,
    "timestamp": "2025-07-14T12:35:01.602419"
  },
  {
    "epoch": 2440,
    "batch_idx": 32,
    "loss": 0.28008222579956055,
    "learning_rate": 0.00011474337861210564,
    "timestep_mean": 449.0,
    "epsilon_norm": 9.519798278808594,
    "prediction_norm": 8.193316459655762,
    "timestamp": "2025-07-14T12:35:01.607632"
  },
  {
    "epoch": 2440,
    "batch_idx": 64,
    "loss": 0.18679118156433105,
    "learning_rate": 0.00011474337861210564,
    "timestep_mean": 500.34375,
    "epsilon_norm": 7.746826171875,
    "prediction_norm": 6.597161769866943,
    "timestamp": "2025-07-14T12:35:01.612992"
  },
  {
    "epoch": 2440,
    "batch_idx": 96,
    "loss": 0.17203021049499512,
    "learning_rate": 0.00011474337861210564,
    "timestep_mean": 543.625,
    "epsilon_norm": 7.8908305168151855,
    "prediction_norm": 7.509161949157715,
    "timestamp": "2025-07-14T12:35:01.618161"
  },
  {
    "epoch": 2440,
    "batch_idx": 128,
    "loss": 0.4362209439277649,
    "learning_rate": 0.00011474337861210564,
    "timestep_mean": 419.375,
    "epsilon_norm": 8.336905479431152,
    "prediction_norm": 6.491598606109619,
    "timestamp": "2025-07-14T12:35:01.623388"
  },
  {
    "epoch": 2440,
    "batch_idx": 160,
    "loss": 0.3045197129249573,
    "learning_rate": 0.00011474337861210564,
    "timestep_mean": 557.5,
    "epsilon_norm": 8.237517356872559,
    "prediction_norm": 7.262665271759033,
    "timestamp": "2025-07-14T12:35:01.628623"
  },
  {
    "epoch": 2440,
    "batch_idx": 192,
    "loss": 0.12615209817886353,
    "learning_rate": 0.00011474337861210564,
    "timestep_mean": 720.25,
    "epsilon_norm": 3.8331258296966553,
    "prediction_norm": 3.6696107387542725,
    "timestamp": "2025-07-14T12:35:01.633745"
  },
  {
    "epoch": 2441,
    "batch_idx": 0,
    "loss": 0.19842490553855896,
    "learning_rate": 0.00011524448532829124,
    "timestep_mean": 529.5625,
    "epsilon_norm": 7.2961106300354,
    "prediction_norm": 6.351016044616699,
    "timestamp": "2025-07-14T12:35:01.638937"
  },
  {
    "epoch": 2441,
    "batch_idx": 32,
    "loss": 0.17072361707687378,
    "learning_rate": 0.00011524448532829124,
    "timestep_mean": 502.03125,
    "epsilon_norm": 7.946102142333984,
    "prediction_norm": 6.815952777862549,
    "timestamp": "2025-07-14T12:35:01.644150"
  },
  {
    "epoch": 2441,
    "batch_idx": 64,
    "loss": 0.12002622336149216,
    "learning_rate": 0.00011524448532829124,
    "timestep_mean": 618.40625,
    "epsilon_norm": 8.820160865783691,
    "prediction_norm": 8.346170425415039,
    "timestamp": "2025-07-14T12:35:01.649416"
  },
  {
    "epoch": 2441,
    "batch_idx": 96,
    "loss": 0.16103875637054443,
    "learning_rate": 0.00011524448532829124,
    "timestep_mean": 466.09375,
    "epsilon_norm": 7.531209468841553,
    "prediction_norm": 6.454870700836182,
    "timestamp": "2025-07-14T12:35:01.654694"
  },
  {
    "epoch": 2441,
    "batch_idx": 128,
    "loss": 0.13814213871955872,
    "learning_rate": 0.00011524448532829124,
    "timestep_mean": 587.34375,
    "epsilon_norm": 7.786414623260498,
    "prediction_norm": 7.551868915557861,
    "timestamp": "2025-07-14T12:35:01.659959"
  },
  {
    "epoch": 2441,
    "batch_idx": 160,
    "loss": 0.3866683542728424,
    "learning_rate": 0.00011524448532829124,
    "timestep_mean": 503.375,
    "epsilon_norm": 7.647299766540527,
    "prediction_norm": 5.4589996337890625,
    "timestamp": "2025-07-14T12:35:01.665113"
  },
  {
    "epoch": 2441,
    "batch_idx": 192,
    "loss": 0.42643406987190247,
    "learning_rate": 0.00011524448532829124,
    "timestep_mean": 557.0,
    "epsilon_norm": 4.399971008300781,
    "prediction_norm": 3.7029807567596436,
    "timestamp": "2025-07-14T12:35:01.670335"
  },
  {
    "epoch": 2442,
    "batch_idx": 0,
    "loss": 0.22312641143798828,
    "learning_rate": 0.00011574654139046193,
    "timestep_mean": 540.6875,
    "epsilon_norm": 7.663031101226807,
    "prediction_norm": 6.893436431884766,
    "timestamp": "2025-07-14T12:35:01.675516"
  },
  {
    "epoch": 2442,
    "batch_idx": 32,
    "loss": 0.17888516187667847,
    "learning_rate": 0.00011574654139046193,
    "timestep_mean": 511.625,
    "epsilon_norm": 7.4375810623168945,
    "prediction_norm": 6.1665778160095215,
    "timestamp": "2025-07-14T12:35:01.680609"
  },
  {
    "epoch": 2442,
    "batch_idx": 64,
    "loss": 0.18456143140792847,
    "learning_rate": 0.00011574654139046193,
    "timestep_mean": 519.46875,
    "epsilon_norm": 7.3689656257629395,
    "prediction_norm": 6.344102382659912,
    "timestamp": "2025-07-14T12:35:01.685831"
  },
  {
    "epoch": 2442,
    "batch_idx": 96,
    "loss": 0.2748505175113678,
    "learning_rate": 0.00011574654139046193,
    "timestep_mean": 534.15625,
    "epsilon_norm": 7.885654926300049,
    "prediction_norm": 6.904541492462158,
    "timestamp": "2025-07-14T12:35:01.691019"
  },
  {
    "epoch": 2442,
    "batch_idx": 128,
    "loss": 0.2959240674972534,
    "learning_rate": 0.00011574654139046193,
    "timestep_mean": 472.1875,
    "epsilon_norm": 8.942230224609375,
    "prediction_norm": 7.7943501472473145,
    "timestamp": "2025-07-14T12:35:01.696360"
  },
  {
    "epoch": 2442,
    "batch_idx": 160,
    "loss": 0.30888718366622925,
    "learning_rate": 0.00011574654139046193,
    "timestep_mean": 473.78125,
    "epsilon_norm": 7.455313205718994,
    "prediction_norm": 6.333010673522949,
    "timestamp": "2025-07-14T12:35:01.701587"
  },
  {
    "epoch": 2442,
    "batch_idx": 192,
    "loss": 0.06651313602924347,
    "learning_rate": 0.00011574654139046193,
    "timestep_mean": 455.25,
    "epsilon_norm": 3.6286513805389404,
    "prediction_norm": 3.7790772914886475,
    "timestamp": "2025-07-14T12:35:01.706719"
  },
  {
    "epoch": 2443,
    "batch_idx": 0,
    "loss": 0.2762299180030823,
    "learning_rate": 0.00011624954555984431,
    "timestep_mean": 375.90625,
    "epsilon_norm": 7.293359279632568,
    "prediction_norm": 5.940226078033447,
    "timestamp": "2025-07-14T12:35:01.711992"
  },
  {
    "epoch": 2443,
    "batch_idx": 32,
    "loss": 0.19896841049194336,
    "learning_rate": 0.00011624954555984431,
    "timestep_mean": 564.03125,
    "epsilon_norm": 9.026312828063965,
    "prediction_norm": 8.103549003601074,
    "timestamp": "2025-07-14T12:35:01.717144"
  },
  {
    "epoch": 2443,
    "batch_idx": 64,
    "loss": 0.19237640500068665,
    "learning_rate": 0.00011624954555984431,
    "timestep_mean": 528.25,
    "epsilon_norm": 7.812472820281982,
    "prediction_norm": 7.1443681716918945,
    "timestamp": "2025-07-14T12:35:01.722227"
  },
  {
    "epoch": 2443,
    "batch_idx": 96,
    "loss": 0.26920920610427856,
    "learning_rate": 0.00011624954555984431,
    "timestep_mean": 430.34375,
    "epsilon_norm": 7.23307991027832,
    "prediction_norm": 6.074281215667725,
    "timestamp": "2025-07-14T12:35:01.727406"
  },
  {
    "epoch": 2443,
    "batch_idx": 128,
    "loss": 0.349540114402771,
    "learning_rate": 0.00011624954555984431,
    "timestep_mean": 515.09375,
    "epsilon_norm": 8.383432388305664,
    "prediction_norm": 7.245818614959717,
    "timestamp": "2025-07-14T12:35:01.732555"
  },
  {
    "epoch": 2443,
    "batch_idx": 160,
    "loss": 0.4105603098869324,
    "learning_rate": 0.00011624954555984431,
    "timestep_mean": 411.03125,
    "epsilon_norm": 7.6998491287231445,
    "prediction_norm": 6.064333438873291,
    "timestamp": "2025-07-14T12:35:01.737839"
  },
  {
    "epoch": 2443,
    "batch_idx": 192,
    "loss": 0.12353186309337616,
    "learning_rate": 0.00011624954555984431,
    "timestep_mean": 581.25,
    "epsilon_norm": 3.6807973384857178,
    "prediction_norm": 3.7422025203704834,
    "timestamp": "2025-07-14T12:35:01.742920"
  },
  {
    "epoch": 2444,
    "batch_idx": 0,
    "loss": 0.3151393532752991,
    "learning_rate": 0.00011675349659532537,
    "timestep_mean": 517.15625,
    "epsilon_norm": 7.463373184204102,
    "prediction_norm": 6.170055866241455,
    "timestamp": "2025-07-14T12:35:01.748096"
  },
  {
    "epoch": 2444,
    "batch_idx": 32,
    "loss": 0.23101720213890076,
    "learning_rate": 0.00011675349659532537,
    "timestep_mean": 480.34375,
    "epsilon_norm": 6.809889793395996,
    "prediction_norm": 5.621973991394043,
    "timestamp": "2025-07-14T12:35:01.753369"
  },
  {
    "epoch": 2444,
    "batch_idx": 64,
    "loss": 0.29423317313194275,
    "learning_rate": 0.00011675349659532537,
    "timestep_mean": 441.8125,
    "epsilon_norm": 7.687393665313721,
    "prediction_norm": 6.229788303375244,
    "timestamp": "2025-07-14T12:35:01.758578"
  },
  {
    "epoch": 2444,
    "batch_idx": 96,
    "loss": 0.3777468204498291,
    "learning_rate": 0.00011675349659532537,
    "timestep_mean": 517.34375,
    "epsilon_norm": 8.200098037719727,
    "prediction_norm": 6.237639904022217,
    "timestamp": "2025-07-14T12:35:01.763814"
  },
  {
    "epoch": 2444,
    "batch_idx": 128,
    "loss": 0.14662858843803406,
    "learning_rate": 0.00011675349659532537,
    "timestep_mean": 513.46875,
    "epsilon_norm": 7.038150787353516,
    "prediction_norm": 6.636003017425537,
    "timestamp": "2025-07-14T12:35:01.769081"
  },
  {
    "epoch": 2444,
    "batch_idx": 160,
    "loss": 0.10520614683628082,
    "learning_rate": 0.00011675349659532537,
    "timestep_mean": 594.71875,
    "epsilon_norm": 7.812488079071045,
    "prediction_norm": 7.41293478012085,
    "timestamp": "2025-07-14T12:35:01.774242"
  },
  {
    "epoch": 2444,
    "batch_idx": 192,
    "loss": 0.4970133304595947,
    "learning_rate": 0.00011675349659532537,
    "timestep_mean": 683.625,
    "epsilon_norm": 3.482367753982544,
    "prediction_norm": 2.9527359008789062,
    "timestamp": "2025-07-14T12:35:01.779499"
  },
  {
    "epoch": 2445,
    "batch_idx": 0,
    "loss": 0.20443248748779297,
    "learning_rate": 0.00011725839325345616,
    "timestep_mean": 529.65625,
    "epsilon_norm": 8.459090232849121,
    "prediction_norm": 7.930041313171387,
    "timestamp": "2025-07-14T12:35:01.784739"
  },
  {
    "epoch": 2445,
    "batch_idx": 32,
    "loss": 0.2399139702320099,
    "learning_rate": 0.00011725839325345616,
    "timestep_mean": 516.03125,
    "epsilon_norm": 8.056048393249512,
    "prediction_norm": 7.026374340057373,
    "timestamp": "2025-07-14T12:35:01.789945"
  },
  {
    "epoch": 2445,
    "batch_idx": 64,
    "loss": 0.19921502470970154,
    "learning_rate": 0.00011725839325345616,
    "timestep_mean": 512.59375,
    "epsilon_norm": 7.79231071472168,
    "prediction_norm": 6.582043647766113,
    "timestamp": "2025-07-14T12:35:01.795182"
  },
  {
    "epoch": 2445,
    "batch_idx": 96,
    "loss": 0.17261143028736115,
    "learning_rate": 0.00011725839325345616,
    "timestep_mean": 588.90625,
    "epsilon_norm": 7.146630764007568,
    "prediction_norm": 6.639585971832275,
    "timestamp": "2025-07-14T12:35:01.800335"
  },
  {
    "epoch": 2445,
    "batch_idx": 128,
    "loss": 0.2941754460334778,
    "learning_rate": 0.00011725839325345616,
    "timestep_mean": 452.625,
    "epsilon_norm": 8.0795316696167,
    "prediction_norm": 6.259666919708252,
    "timestamp": "2025-07-14T12:35:01.805478"
  },
  {
    "epoch": 2445,
    "batch_idx": 160,
    "loss": 0.3569191098213196,
    "learning_rate": 0.00011725839325345616,
    "timestep_mean": 506.09375,
    "epsilon_norm": 8.53960132598877,
    "prediction_norm": 6.909661293029785,
    "timestamp": "2025-07-14T12:35:01.810736"
  },
  {
    "epoch": 2445,
    "batch_idx": 192,
    "loss": 0.29849773645401,
    "learning_rate": 0.00011725839325345616,
    "timestep_mean": 515.75,
    "epsilon_norm": 3.771517515182495,
    "prediction_norm": 3.4530956745147705,
    "timestamp": "2025-07-14T12:35:01.815929"
  },
  {
    "epoch": 2446,
    "batch_idx": 0,
    "loss": 0.3386671245098114,
    "learning_rate": 0.0001177642342884544,
    "timestep_mean": 500.5,
    "epsilon_norm": 8.612578392028809,
    "prediction_norm": 6.745696544647217,
    "timestamp": "2025-07-14T12:35:01.821173"
  },
  {
    "epoch": 2446,
    "batch_idx": 32,
    "loss": 0.13108614087104797,
    "learning_rate": 0.0001177642342884544,
    "timestep_mean": 575.09375,
    "epsilon_norm": 7.56110954284668,
    "prediction_norm": 6.658706188201904,
    "timestamp": "2025-07-14T12:35:01.826323"
  },
  {
    "epoch": 2446,
    "batch_idx": 64,
    "loss": 0.2846083641052246,
    "learning_rate": 0.0001177642342884544,
    "timestep_mean": 453.875,
    "epsilon_norm": 8.341777801513672,
    "prediction_norm": 7.24361515045166,
    "timestamp": "2025-07-14T12:35:01.831479"
  },
  {
    "epoch": 2446,
    "batch_idx": 96,
    "loss": 0.1873682737350464,
    "learning_rate": 0.0001177642342884544,
    "timestep_mean": 632.90625,
    "epsilon_norm": 7.752226829528809,
    "prediction_norm": 7.400733947753906,
    "timestamp": "2025-07-14T12:35:01.836689"
  },
  {
    "epoch": 2446,
    "batch_idx": 128,
    "loss": 0.4129239022731781,
    "learning_rate": 0.0001177642342884544,
    "timestep_mean": 463.46875,
    "epsilon_norm": 7.129493713378906,
    "prediction_norm": 5.599897384643555,
    "timestamp": "2025-07-14T12:35:01.841815"
  },
  {
    "epoch": 2446,
    "batch_idx": 160,
    "loss": 0.43138018250465393,
    "learning_rate": 0.0001177642342884544,
    "timestep_mean": 437.65625,
    "epsilon_norm": 8.561220169067383,
    "prediction_norm": 6.615943908691406,
    "timestamp": "2025-07-14T12:35:01.846913"
  },
  {
    "epoch": 2446,
    "batch_idx": 192,
    "loss": 0.2794128656387329,
    "learning_rate": 0.0001177642342884544,
    "timestep_mean": 305.875,
    "epsilon_norm": 3.887022018432617,
    "prediction_norm": 3.8563761711120605,
    "timestamp": "2025-07-14T12:35:01.852134"
  },
  {
    "epoch": 2447,
    "batch_idx": 0,
    "loss": 0.23504197597503662,
    "learning_rate": 0.00011827101845220762,
    "timestep_mean": 451.4375,
    "epsilon_norm": 7.711744785308838,
    "prediction_norm": 6.642643451690674,
    "timestamp": "2025-07-14T12:35:01.857348"
  },
  {
    "epoch": 2447,
    "batch_idx": 32,
    "loss": 0.30629870295524597,
    "learning_rate": 0.00011827101845220762,
    "timestep_mean": 422.90625,
    "epsilon_norm": 7.9787468910217285,
    "prediction_norm": 6.65847635269165,
    "timestamp": "2025-07-14T12:35:01.862546"
  },
  {
    "epoch": 2447,
    "batch_idx": 64,
    "loss": 0.19280436635017395,
    "learning_rate": 0.00011827101845220762,
    "timestep_mean": 509.75,
    "epsilon_norm": 7.424790859222412,
    "prediction_norm": 6.73682975769043,
    "timestamp": "2025-07-14T12:35:01.867919"
  },
  {
    "epoch": 2447,
    "batch_idx": 96,
    "loss": 0.35943007469177246,
    "learning_rate": 0.00011827101845220762,
    "timestep_mean": 419.9375,
    "epsilon_norm": 7.246563911437988,
    "prediction_norm": 6.011947154998779,
    "timestamp": "2025-07-14T12:35:01.873141"
  },
  {
    "epoch": 2447,
    "batch_idx": 128,
    "loss": 0.2611705958843231,
    "learning_rate": 0.00011827101845220762,
    "timestep_mean": 534.6875,
    "epsilon_norm": 7.927059650421143,
    "prediction_norm": 7.157565593719482,
    "timestamp": "2025-07-14T12:35:01.878323"
  },
  {
    "epoch": 2447,
    "batch_idx": 160,
    "loss": 0.31842148303985596,
    "learning_rate": 0.00011827101845220762,
    "timestep_mean": 471.125,
    "epsilon_norm": 8.917255401611328,
    "prediction_norm": 7.4441938400268555,
    "timestamp": "2025-07-14T12:35:01.883442"
  },
  {
    "epoch": 2447,
    "batch_idx": 192,
    "loss": 0.044643692672252655,
    "learning_rate": 0.00011827101845220762,
    "timestep_mean": 643.25,
    "epsilon_norm": 3.6134886741638184,
    "prediction_norm": 3.7695744037628174,
    "timestamp": "2025-07-14T12:35:01.888603"
  },
  {
    "epoch": 2448,
    "batch_idx": 0,
    "loss": 0.15033569931983948,
    "learning_rate": 0.00011877874449427633,
    "timestep_mean": 489.84375,
    "epsilon_norm": 7.9823198318481445,
    "prediction_norm": 7.258111476898193,
    "timestamp": "2025-07-14T12:35:01.893987"
  },
  {
    "epoch": 2448,
    "batch_idx": 32,
    "loss": 0.3117706775665283,
    "learning_rate": 0.00011877874449427633,
    "timestep_mean": 541.21875,
    "epsilon_norm": 8.170528411865234,
    "prediction_norm": 6.583456993103027,
    "timestamp": "2025-07-14T12:35:01.899232"
  },
  {
    "epoch": 2448,
    "batch_idx": 64,
    "loss": 0.2793086767196655,
    "learning_rate": 0.00011877874449427633,
    "timestep_mean": 441.46875,
    "epsilon_norm": 7.606079578399658,
    "prediction_norm": 5.987842559814453,
    "timestamp": "2025-07-14T12:35:01.904063"
  },
  {
    "epoch": 2448,
    "batch_idx": 96,
    "loss": 0.26339030265808105,
    "learning_rate": 0.00011877874449427633,
    "timestep_mean": 422.53125,
    "epsilon_norm": 8.191171646118164,
    "prediction_norm": 6.743055820465088,
    "timestamp": "2025-07-14T12:35:01.909034"
  },
  {
    "epoch": 2448,
    "batch_idx": 128,
    "loss": 0.1648806780576706,
    "learning_rate": 0.00011877874449427633,
    "timestep_mean": 474.65625,
    "epsilon_norm": 6.10366153717041,
    "prediction_norm": 5.262767314910889,
    "timestamp": "2025-07-14T12:35:01.913868"
  },
  {
    "epoch": 2448,
    "batch_idx": 160,
    "loss": 0.305125892162323,
    "learning_rate": 0.00011877874449427633,
    "timestep_mean": 463.28125,
    "epsilon_norm": 8.706451416015625,
    "prediction_norm": 7.145885944366455,
    "timestamp": "2025-07-14T12:35:01.918773"
  },
  {
    "epoch": 2448,
    "batch_idx": 192,
    "loss": 0.16742563247680664,
    "learning_rate": 0.00011877874449427633,
    "timestep_mean": 470.75,
    "epsilon_norm": 3.7074339389801025,
    "prediction_norm": 3.8757290840148926,
    "timestamp": "2025-07-14T12:35:01.923642"
  },
  {
    "epoch": 2449,
    "batch_idx": 0,
    "loss": 0.26448583602905273,
    "learning_rate": 0.00011928741116189667,
    "timestep_mean": 445.6875,
    "epsilon_norm": 7.435739040374756,
    "prediction_norm": 6.037808418273926,
    "timestamp": "2025-07-14T12:35:01.928558"
  },
  {
    "epoch": 2449,
    "batch_idx": 32,
    "loss": 0.3040798306465149,
    "learning_rate": 0.00011928741116189667,
    "timestep_mean": 460.875,
    "epsilon_norm": 7.924905776977539,
    "prediction_norm": 6.712686538696289,
    "timestamp": "2025-07-14T12:35:01.933460"
  },
  {
    "epoch": 2449,
    "batch_idx": 64,
    "loss": 0.22828778624534607,
    "learning_rate": 0.00011928741116189667,
    "timestep_mean": 465.34375,
    "epsilon_norm": 7.9705424308776855,
    "prediction_norm": 6.609892845153809,
    "timestamp": "2025-07-14T12:35:01.938345"
  },
  {
    "epoch": 2449,
    "batch_idx": 96,
    "loss": 0.26654431223869324,
    "learning_rate": 0.00011928741116189667,
    "timestep_mean": 427.28125,
    "epsilon_norm": 8.649686813354492,
    "prediction_norm": 7.111255168914795,
    "timestamp": "2025-07-14T12:35:01.943154"
  },
  {
    "epoch": 2449,
    "batch_idx": 128,
    "loss": 0.4146871864795685,
    "learning_rate": 0.00011928741116189667,
    "timestep_mean": 445.3125,
    "epsilon_norm": 6.280467510223389,
    "prediction_norm": 4.660019397735596,
    "timestamp": "2025-07-14T12:35:01.948005"
  },
  {
    "epoch": 2449,
    "batch_idx": 160,
    "loss": 0.12448625266551971,
    "learning_rate": 0.00011928741116189667,
    "timestep_mean": 578.5625,
    "epsilon_norm": 7.9583330154418945,
    "prediction_norm": 7.255730152130127,
    "timestamp": "2025-07-14T12:35:01.952882"
  },
  {
    "epoch": 2449,
    "batch_idx": 192,
    "loss": 0.12866398692131042,
    "learning_rate": 0.00011928741116189667,
    "timestep_mean": 406.375,
    "epsilon_norm": 3.4218060970306396,
    "prediction_norm": 3.4788780212402344,
    "timestamp": "2025-07-14T12:35:01.957681"
  },
  {
    "epoch": 2450,
    "batch_idx": 0,
    "loss": 0.28828567266464233,
    "learning_rate": 0.00011979701719998477,
    "timestep_mean": 502.34375,
    "epsilon_norm": 8.228168487548828,
    "prediction_norm": 6.946310043334961,
    "timestamp": "2025-07-14T12:35:01.962499"
  },
  {
    "epoch": 2450,
    "batch_idx": 32,
    "loss": 0.25345107913017273,
    "learning_rate": 0.00011979701719998477,
    "timestep_mean": 477.25,
    "epsilon_norm": 7.612403869628906,
    "prediction_norm": 5.872077941894531,
    "timestamp": "2025-07-14T12:35:01.967399"
  },
  {
    "epoch": 2450,
    "batch_idx": 64,
    "loss": 0.2660129964351654,
    "learning_rate": 0.00011979701719998477,
    "timestep_mean": 528.625,
    "epsilon_norm": 8.739164352416992,
    "prediction_norm": 7.176982402801514,
    "timestamp": "2025-07-14T12:35:01.972293"
  },
  {
    "epoch": 2450,
    "batch_idx": 96,
    "loss": 0.2872915267944336,
    "learning_rate": 0.00011979701719998477,
    "timestep_mean": 453.96875,
    "epsilon_norm": 7.618276119232178,
    "prediction_norm": 6.1062421798706055,
    "timestamp": "2025-07-14T12:35:01.977307"
  },
  {
    "epoch": 2450,
    "batch_idx": 128,
    "loss": 0.2881648540496826,
    "learning_rate": 0.00011979701719998477,
    "timestep_mean": 455.25,
    "epsilon_norm": 6.602522373199463,
    "prediction_norm": 5.292232513427734,
    "timestamp": "2025-07-14T12:35:01.982471"
  },
  {
    "epoch": 2450,
    "batch_idx": 160,
    "loss": 0.30915626883506775,
    "learning_rate": 0.00011979701719998477,
    "timestep_mean": 439.90625,
    "epsilon_norm": 8.083280563354492,
    "prediction_norm": 6.571258068084717,
    "timestamp": "2025-07-14T12:35:01.987596"
  },
  {
    "epoch": 2450,
    "batch_idx": 192,
    "loss": 0.2526637017726898,
    "learning_rate": 0.00011979701719998477,
    "timestep_mean": 440.25,
    "epsilon_norm": 4.361123561859131,
    "prediction_norm": 3.4958369731903076,
    "timestamp": "2025-07-14T12:35:01.992850"
  },
  {
    "epoch": 2451,
    "batch_idx": 0,
    "loss": 0.08840682357549667,
    "learning_rate": 0.00012030756135113802,
    "timestep_mean": 612.28125,
    "epsilon_norm": 8.09911060333252,
    "prediction_norm": 7.430865287780762,
    "timestamp": "2025-07-14T12:35:01.997948"
  },
  {
    "epoch": 2451,
    "batch_idx": 32,
    "loss": 0.24222087860107422,
    "learning_rate": 0.00012030756135113802,
    "timestep_mean": 498.34375,
    "epsilon_norm": 7.790632247924805,
    "prediction_norm": 6.8545637130737305,
    "timestamp": "2025-07-14T12:35:02.003066"
  },
  {
    "epoch": 2451,
    "batch_idx": 64,
    "loss": 0.39488983154296875,
    "learning_rate": 0.00012030756135113802,
    "timestep_mean": 472.34375,
    "epsilon_norm": 8.412405014038086,
    "prediction_norm": 6.13068962097168,
    "timestamp": "2025-07-14T12:35:02.008188"
  },
  {
    "epoch": 2451,
    "batch_idx": 96,
    "loss": 0.43873506784439087,
    "learning_rate": 0.00012030756135113802,
    "timestep_mean": 449.71875,
    "epsilon_norm": 8.130945205688477,
    "prediction_norm": 5.707583904266357,
    "timestamp": "2025-07-14T12:35:02.013441"
  },
  {
    "epoch": 2451,
    "batch_idx": 128,
    "loss": 0.470216304063797,
    "learning_rate": 0.00012030756135113802,
    "timestep_mean": 456.96875,
    "epsilon_norm": 8.479568481445312,
    "prediction_norm": 7.19067907333374,
    "timestamp": "2025-07-14T12:35:02.018665"
  },
  {
    "epoch": 2451,
    "batch_idx": 160,
    "loss": 0.34113410115242004,
    "learning_rate": 0.00012030756135113802,
    "timestep_mean": 471.40625,
    "epsilon_norm": 7.1538872718811035,
    "prediction_norm": 5.520803928375244,
    "timestamp": "2025-07-14T12:35:02.023930"
  },
  {
    "epoch": 2451,
    "batch_idx": 192,
    "loss": 0.2958083152770996,
    "learning_rate": 0.00012030756135113802,
    "timestep_mean": 420.0,
    "epsilon_norm": 3.4698123931884766,
    "prediction_norm": 2.892650604248047,
    "timestamp": "2025-07-14T12:35:02.029090"
  },
  {
    "epoch": 2452,
    "batch_idx": 0,
    "loss": 0.3336867690086365,
    "learning_rate": 0.00012081904235563924,
    "timestep_mean": 481.875,
    "epsilon_norm": 9.013975143432617,
    "prediction_norm": 6.953216552734375,
    "timestamp": "2025-07-14T12:35:02.034322"
  },
  {
    "epoch": 2452,
    "batch_idx": 32,
    "loss": 0.1575891077518463,
    "learning_rate": 0.00012081904235563924,
    "timestep_mean": 555.5,
    "epsilon_norm": 8.479726791381836,
    "prediction_norm": 7.067138195037842,
    "timestamp": "2025-07-14T12:35:02.039312"
  },
  {
    "epoch": 2452,
    "batch_idx": 64,
    "loss": 0.34678536653518677,
    "learning_rate": 0.00012081904235563924,
    "timestep_mean": 487.15625,
    "epsilon_norm": 7.254392623901367,
    "prediction_norm": 5.6236491203308105,
    "timestamp": "2025-07-14T12:35:02.044394"
  },
  {
    "epoch": 2452,
    "batch_idx": 96,
    "loss": 0.3017466068267822,
    "learning_rate": 0.00012081904235563924,
    "timestep_mean": 538.3125,
    "epsilon_norm": 8.005066871643066,
    "prediction_norm": 6.714557647705078,
    "timestamp": "2025-07-14T12:35:02.049536"
  },
  {
    "epoch": 2452,
    "batch_idx": 128,
    "loss": 0.2718156576156616,
    "learning_rate": 0.00012081904235563924,
    "timestep_mean": 482.46875,
    "epsilon_norm": 7.944308280944824,
    "prediction_norm": 6.356270790100098,
    "timestamp": "2025-07-14T12:35:02.054644"
  },
  {
    "epoch": 2452,
    "batch_idx": 160,
    "loss": 0.32511022686958313,
    "learning_rate": 0.00012081904235563924,
    "timestep_mean": 488.59375,
    "epsilon_norm": 6.870295524597168,
    "prediction_norm": 5.932827472686768,
    "timestamp": "2025-07-14T12:35:02.059536"
  },
  {
    "epoch": 2452,
    "batch_idx": 192,
    "loss": 0.2622101306915283,
    "learning_rate": 0.00012081904235563924,
    "timestep_mean": 546.5,
    "epsilon_norm": 5.277894496917725,
    "prediction_norm": 4.270263195037842,
    "timestamp": "2025-07-14T12:35:02.064413"
  },
  {
    "epoch": 2453,
    "batch_idx": 0,
    "loss": 0.10667730122804642,
    "learning_rate": 0.0001213314589514605,
    "timestep_mean": 457.0,
    "epsilon_norm": 6.704412937164307,
    "prediction_norm": 6.173722267150879,
    "timestamp": "2025-07-14T12:35:02.069276"
  },
  {
    "epoch": 2453,
    "batch_idx": 32,
    "loss": 0.21566739678382874,
    "learning_rate": 0.0001213314589514605,
    "timestep_mean": 451.9375,
    "epsilon_norm": 7.155894756317139,
    "prediction_norm": 5.881345272064209,
    "timestamp": "2025-07-14T12:35:02.074246"
  },
  {
    "epoch": 2453,
    "batch_idx": 64,
    "loss": 0.3368580937385559,
    "learning_rate": 0.0001213314589514605,
    "timestep_mean": 462.3125,
    "epsilon_norm": 7.651257038116455,
    "prediction_norm": 6.201758861541748,
    "timestamp": "2025-07-14T12:35:02.079236"
  },
  {
    "epoch": 2453,
    "batch_idx": 96,
    "loss": 0.17710748314857483,
    "learning_rate": 0.0001213314589514605,
    "timestep_mean": 515.71875,
    "epsilon_norm": 7.705346584320068,
    "prediction_norm": 7.245475769042969,
    "timestamp": "2025-07-14T12:35:02.084321"
  },
  {
    "epoch": 2453,
    "batch_idx": 128,
    "loss": 0.3241856098175049,
    "learning_rate": 0.0001213314589514605,
    "timestep_mean": 450.65625,
    "epsilon_norm": 7.987658977508545,
    "prediction_norm": 6.595913887023926,
    "timestamp": "2025-07-14T12:35:02.089441"
  },
  {
    "epoch": 2453,
    "batch_idx": 160,
    "loss": 0.2533065378665924,
    "learning_rate": 0.0001213314589514605,
    "timestep_mean": 480.3125,
    "epsilon_norm": 7.660891532897949,
    "prediction_norm": 6.152774810791016,
    "timestamp": "2025-07-14T12:35:02.094561"
  },
  {
    "epoch": 2453,
    "batch_idx": 192,
    "loss": 0.3701131343841553,
    "learning_rate": 0.0001213314589514605,
    "timestep_mean": 509.875,
    "epsilon_norm": 5.46207857131958,
    "prediction_norm": 4.2425217628479,
    "timestamp": "2025-07-14T12:35:02.099665"
  },
  {
    "epoch": 2454,
    "batch_idx": 0,
    "loss": 0.21584096550941467,
    "learning_rate": 0.00012184480987426431,
    "timestep_mean": 468.4375,
    "epsilon_norm": 6.368538856506348,
    "prediction_norm": 5.281220436096191,
    "timestamp": "2025-07-14T12:35:02.104889"
  },
  {
    "epoch": 2454,
    "batch_idx": 32,
    "loss": 0.12907299399375916,
    "learning_rate": 0.00012184480987426431,
    "timestep_mean": 650.0625,
    "epsilon_norm": 6.643771171569824,
    "prediction_norm": 6.013856410980225,
    "timestamp": "2025-07-14T12:35:02.110051"
  },
  {
    "epoch": 2454,
    "batch_idx": 64,
    "loss": 0.25211501121520996,
    "learning_rate": 0.00012184480987426431,
    "timestep_mean": 528.6875,
    "epsilon_norm": 7.4034199714660645,
    "prediction_norm": 6.4757161140441895,
    "timestamp": "2025-07-14T12:35:02.115176"
  },
  {
    "epoch": 2454,
    "batch_idx": 96,
    "loss": 0.29583877325057983,
    "learning_rate": 0.00012184480987426431,
    "timestep_mean": 484.15625,
    "epsilon_norm": 8.56266975402832,
    "prediction_norm": 7.433845520019531,
    "timestamp": "2025-07-14T12:35:02.120388"
  },
  {
    "epoch": 2454,
    "batch_idx": 128,
    "loss": 0.40132391452789307,
    "learning_rate": 0.00012184480987426431,
    "timestep_mean": 420.75,
    "epsilon_norm": 9.30656623840332,
    "prediction_norm": 7.328925609588623,
    "timestamp": "2025-07-14T12:35:02.125736"
  },
  {
    "epoch": 2454,
    "batch_idx": 160,
    "loss": 0.4157812297344208,
    "learning_rate": 0.00012184480987426431,
    "timestep_mean": 480.5625,
    "epsilon_norm": 10.14206314086914,
    "prediction_norm": 8.822874069213867,
    "timestamp": "2025-07-14T12:35:02.130970"
  },
  {
    "epoch": 2454,
    "batch_idx": 192,
    "loss": 0.3211105465888977,
    "learning_rate": 0.00012184480987426431,
    "timestep_mean": 528.625,
    "epsilon_norm": 4.748392105102539,
    "prediction_norm": 3.792182683944702,
    "timestamp": "2025-07-14T12:35:02.136137"
  },
  {
    "epoch": 2455,
    "batch_idx": 0,
    "loss": 0.2845459282398224,
    "learning_rate": 0.00012235909385740844,
    "timestep_mean": 463.75,
    "epsilon_norm": 6.857561111450195,
    "prediction_norm": 5.647461891174316,
    "timestamp": "2025-07-14T12:35:02.141298"
  },
  {
    "epoch": 2455,
    "batch_idx": 32,
    "loss": 0.29362332820892334,
    "learning_rate": 0.00012235909385740844,
    "timestep_mean": 480.71875,
    "epsilon_norm": 9.259476661682129,
    "prediction_norm": 8.036994934082031,
    "timestamp": "2025-07-14T12:35:02.146639"
  },
  {
    "epoch": 2455,
    "batch_idx": 64,
    "loss": 0.20062188804149628,
    "learning_rate": 0.00012235909385740844,
    "timestep_mean": 546.78125,
    "epsilon_norm": 7.243244647979736,
    "prediction_norm": 6.970165252685547,
    "timestamp": "2025-07-14T12:35:02.151786"
  },
  {
    "epoch": 2455,
    "batch_idx": 96,
    "loss": 0.3742256164550781,
    "learning_rate": 0.00012235909385740844,
    "timestep_mean": 531.15625,
    "epsilon_norm": 7.8802666664123535,
    "prediction_norm": 6.969174385070801,
    "timestamp": "2025-07-14T12:35:02.157045"
  },
  {
    "epoch": 2455,
    "batch_idx": 128,
    "loss": 0.35159704089164734,
    "learning_rate": 0.00012235909385740844,
    "timestep_mean": 454.46875,
    "epsilon_norm": 7.944408893585205,
    "prediction_norm": 6.301984786987305,
    "timestamp": "2025-07-14T12:35:02.162209"
  },
  {
    "epoch": 2455,
    "batch_idx": 160,
    "loss": 0.22014541923999786,
    "learning_rate": 0.00012235909385740844,
    "timestep_mean": 495.53125,
    "epsilon_norm": 8.154997825622559,
    "prediction_norm": 7.356765270233154,
    "timestamp": "2025-07-14T12:35:02.167515"
  },
  {
    "epoch": 2455,
    "batch_idx": 192,
    "loss": 0.14778771996498108,
    "learning_rate": 0.00012235909385740844,
    "timestep_mean": 399.75,
    "epsilon_norm": 4.529670238494873,
    "prediction_norm": 4.127006530761719,
    "timestamp": "2025-07-14T12:35:02.172604"
  },
  {
    "epoch": 2456,
    "batch_idx": 0,
    "loss": 0.32527095079421997,
    "learning_rate": 0.0001228743096319485,
    "timestep_mean": 588.34375,
    "epsilon_norm": 9.645957946777344,
    "prediction_norm": 8.838693618774414,
    "timestamp": "2025-07-14T12:35:02.177846"
  },
  {
    "epoch": 2456,
    "batch_idx": 32,
    "loss": 0.35148146748542786,
    "learning_rate": 0.0001228743096319485,
    "timestep_mean": 524.46875,
    "epsilon_norm": 9.453474998474121,
    "prediction_norm": 7.952197074890137,
    "timestamp": "2025-07-14T12:35:02.183043"
  },
  {
    "epoch": 2456,
    "batch_idx": 64,
    "loss": 0.2622497081756592,
    "learning_rate": 0.0001228743096319485,
    "timestep_mean": 561.75,
    "epsilon_norm": 7.932262897491455,
    "prediction_norm": 6.7969889640808105,
    "timestamp": "2025-07-14T12:35:02.188348"
  },
  {
    "epoch": 2456,
    "batch_idx": 96,
    "loss": 0.2203179895877838,
    "learning_rate": 0.0001228743096319485,
    "timestep_mean": 519.46875,
    "epsilon_norm": 7.908144474029541,
    "prediction_norm": 6.9540276527404785,
    "timestamp": "2025-07-14T12:35:02.193516"
  },
  {
    "epoch": 2456,
    "batch_idx": 128,
    "loss": 0.2590848207473755,
    "learning_rate": 0.0001228743096319485,
    "timestep_mean": 548.4375,
    "epsilon_norm": 7.921508312225342,
    "prediction_norm": 6.701135158538818,
    "timestamp": "2025-07-14T12:35:02.198747"
  },
  {
    "epoch": 2456,
    "batch_idx": 160,
    "loss": 0.2732360363006592,
    "learning_rate": 0.0001228743096319485,
    "timestep_mean": 494.1875,
    "epsilon_norm": 8.309477806091309,
    "prediction_norm": 7.163332939147949,
    "timestamp": "2025-07-14T12:35:02.203719"
  },
  {
    "epoch": 2456,
    "batch_idx": 192,
    "loss": 0.1567811667919159,
    "learning_rate": 0.0001228743096319485,
    "timestep_mean": 450.25,
    "epsilon_norm": 3.1767685413360596,
    "prediction_norm": 2.8750078678131104,
    "timestamp": "2025-07-14T12:35:02.208897"
  },
  {
    "epoch": 2457,
    "batch_idx": 0,
    "loss": 0.35666579008102417,
    "learning_rate": 0.00012339045592664021,
    "timestep_mean": 510.84375,
    "epsilon_norm": 7.815249443054199,
    "prediction_norm": 6.718178749084473,
    "timestamp": "2025-07-14T12:35:02.214088"
  },
  {
    "epoch": 2457,
    "batch_idx": 32,
    "loss": 0.3708006739616394,
    "learning_rate": 0.00012339045592664021,
    "timestep_mean": 430.34375,
    "epsilon_norm": 9.238631248474121,
    "prediction_norm": 8.082058906555176,
    "timestamp": "2025-07-14T12:35:02.219298"
  },
  {
    "epoch": 2457,
    "batch_idx": 64,
    "loss": 0.2548602223396301,
    "learning_rate": 0.00012339045592664021,
    "timestep_mean": 459.84375,
    "epsilon_norm": 7.952418327331543,
    "prediction_norm": 6.637882709503174,
    "timestamp": "2025-07-14T12:35:02.224481"
  },
  {
    "epoch": 2457,
    "batch_idx": 96,
    "loss": 0.31199342012405396,
    "learning_rate": 0.00012339045592664021,
    "timestep_mean": 343.625,
    "epsilon_norm": 7.469438076019287,
    "prediction_norm": 6.361302852630615,
    "timestamp": "2025-07-14T12:35:02.229692"
  },
  {
    "epoch": 2457,
    "batch_idx": 128,
    "loss": 0.253950297832489,
    "learning_rate": 0.00012339045592664021,
    "timestep_mean": 454.96875,
    "epsilon_norm": 7.349654674530029,
    "prediction_norm": 6.468444347381592,
    "timestamp": "2025-07-14T12:35:02.234847"
  },
  {
    "epoch": 2457,
    "batch_idx": 160,
    "loss": 0.2453647255897522,
    "learning_rate": 0.00012339045592664021,
    "timestep_mean": 482.78125,
    "epsilon_norm": 7.075421333312988,
    "prediction_norm": 6.177896976470947,
    "timestamp": "2025-07-14T12:35:02.240107"
  },
  {
    "epoch": 2457,
    "batch_idx": 192,
    "loss": 0.26393964886665344,
    "learning_rate": 0.00012339045592664021,
    "timestep_mean": 553.875,
    "epsilon_norm": 3.5231754779815674,
    "prediction_norm": 2.754969596862793,
    "timestamp": "2025-07-14T12:35:02.245220"
  },
  {
    "epoch": 2458,
    "batch_idx": 0,
    "loss": 0.38016313314437866,
    "learning_rate": 0.00012390753146794458,
    "timestep_mean": 497.8125,
    "epsilon_norm": 8.215028762817383,
    "prediction_norm": 6.792055130004883,
    "timestamp": "2025-07-14T12:35:02.250602"
  },
  {
    "epoch": 2458,
    "batch_idx": 32,
    "loss": 0.33460092544555664,
    "learning_rate": 0.00012390753146794458,
    "timestep_mean": 515.40625,
    "epsilon_norm": 8.641776084899902,
    "prediction_norm": 7.181250095367432,
    "timestamp": "2025-07-14T12:35:02.255846"
  },
  {
    "epoch": 2458,
    "batch_idx": 64,
    "loss": 0.1625322550535202,
    "learning_rate": 0.00012390753146794458,
    "timestep_mean": 456.40625,
    "epsilon_norm": 6.787511825561523,
    "prediction_norm": 5.966763973236084,
    "timestamp": "2025-07-14T12:35:02.261019"
  },
  {
    "epoch": 2458,
    "batch_idx": 96,
    "loss": 0.1729167103767395,
    "learning_rate": 0.00012390753146794458,
    "timestep_mean": 508.9375,
    "epsilon_norm": 7.659267425537109,
    "prediction_norm": 6.809955596923828,
    "timestamp": "2025-07-14T12:35:02.266197"
  },
  {
    "epoch": 2458,
    "batch_idx": 128,
    "loss": 0.23170851171016693,
    "learning_rate": 0.00012390753146794458,
    "timestep_mean": 459.71875,
    "epsilon_norm": 8.14124584197998,
    "prediction_norm": 6.8590779304504395,
    "timestamp": "2025-07-14T12:35:02.271404"
  },
  {
    "epoch": 2458,
    "batch_idx": 160,
    "loss": 0.22832924127578735,
    "learning_rate": 0.00012390753146794458,
    "timestep_mean": 445.78125,
    "epsilon_norm": 6.942399978637695,
    "prediction_norm": 6.301459312438965,
    "timestamp": "2025-07-14T12:35:02.276589"
  },
  {
    "epoch": 2458,
    "batch_idx": 192,
    "loss": 0.2782389223575592,
    "learning_rate": 0.00012390753146794458,
    "timestep_mean": 482.125,
    "epsilon_norm": 3.394028902053833,
    "prediction_norm": 3.1958067417144775,
    "timestamp": "2025-07-14T12:35:02.281679"
  },
  {
    "epoch": 2459,
    "batch_idx": 0,
    "loss": 0.21800105273723602,
    "learning_rate": 0.00012442553498002873,
    "timestep_mean": 555.75,
    "epsilon_norm": 8.47199535369873,
    "prediction_norm": 7.22228479385376,
    "timestamp": "2025-07-14T12:35:02.286839"
  },
  {
    "epoch": 2459,
    "batch_idx": 32,
    "loss": 0.3183520436286926,
    "learning_rate": 0.00012442553498002873,
    "timestep_mean": 407.1875,
    "epsilon_norm": 8.214045524597168,
    "prediction_norm": 7.346040725708008,
    "timestamp": "2025-07-14T12:35:02.292235"
  },
  {
    "epoch": 2459,
    "batch_idx": 64,
    "loss": 0.20575249195098877,
    "learning_rate": 0.00012442553498002873,
    "timestep_mean": 444.75,
    "epsilon_norm": 7.089128017425537,
    "prediction_norm": 6.0320305824279785,
    "timestamp": "2025-07-14T12:35:02.297482"
  },
  {
    "epoch": 2459,
    "batch_idx": 96,
    "loss": 0.2704889476299286,
    "learning_rate": 0.00012442553498002873,
    "timestep_mean": 499.5625,
    "epsilon_norm": 7.95125675201416,
    "prediction_norm": 7.00594425201416,
    "timestamp": "2025-07-14T12:35:02.302631"
  },
  {
    "epoch": 2459,
    "batch_idx": 128,
    "loss": 0.2899557948112488,
    "learning_rate": 0.00012442553498002873,
    "timestep_mean": 454.0625,
    "epsilon_norm": 7.699699401855469,
    "prediction_norm": 5.64699649810791,
    "timestamp": "2025-07-14T12:35:02.307849"
  },
  {
    "epoch": 2459,
    "batch_idx": 160,
    "loss": 0.4150916039943695,
    "learning_rate": 0.00012442553498002873,
    "timestep_mean": 502.84375,
    "epsilon_norm": 8.394464492797852,
    "prediction_norm": 7.010791778564453,
    "timestamp": "2025-07-14T12:35:02.313014"
  },
  {
    "epoch": 2459,
    "batch_idx": 192,
    "loss": 0.2902872860431671,
    "learning_rate": 0.00012442553498002873,
    "timestep_mean": 504.5,
    "epsilon_norm": 2.404489040374756,
    "prediction_norm": 1.5947539806365967,
    "timestamp": "2025-07-14T12:35:02.318119"
  },
  {
    "epoch": 2460,
    "batch_idx": 0,
    "loss": 0.15847156941890717,
    "learning_rate": 0.00012494446518477047,
    "timestep_mean": 476.4375,
    "epsilon_norm": 8.270755767822266,
    "prediction_norm": 7.386885643005371,
    "timestamp": "2025-07-14T12:35:02.323227"
  },
  {
    "epoch": 2460,
    "batch_idx": 32,
    "loss": 0.1891942024230957,
    "learning_rate": 0.00012494446518477047,
    "timestep_mean": 537.1875,
    "epsilon_norm": 7.681728363037109,
    "prediction_norm": 7.038736343383789,
    "timestamp": "2025-07-14T12:35:02.328443"
  },
  {
    "epoch": 2460,
    "batch_idx": 64,
    "loss": 0.23902229964733124,
    "learning_rate": 0.00012494446518477047,
    "timestep_mean": 537.875,
    "epsilon_norm": 7.097433567047119,
    "prediction_norm": 6.036543369293213,
    "timestamp": "2025-07-14T12:35:02.333705"
  },
  {
    "epoch": 2460,
    "batch_idx": 96,
    "loss": 0.4041679799556732,
    "learning_rate": 0.00012494446518477047,
    "timestep_mean": 438.03125,
    "epsilon_norm": 7.204825401306152,
    "prediction_norm": 5.794173717498779,
    "timestamp": "2025-07-14T12:35:02.338953"
  },
  {
    "epoch": 2460,
    "batch_idx": 128,
    "loss": 0.2269434630870819,
    "learning_rate": 0.00012494446518477047,
    "timestep_mean": 495.96875,
    "epsilon_norm": 7.898746013641357,
    "prediction_norm": 6.469267845153809,
    "timestamp": "2025-07-14T12:35:02.344154"
  },
  {
    "epoch": 2460,
    "batch_idx": 160,
    "loss": 0.31423118710517883,
    "learning_rate": 0.00012494446518477047,
    "timestep_mean": 537.3125,
    "epsilon_norm": 8.11247444152832,
    "prediction_norm": 7.143446445465088,
    "timestamp": "2025-07-14T12:35:02.349330"
  },
  {
    "epoch": 2460,
    "batch_idx": 192,
    "loss": 0.19194790720939636,
    "learning_rate": 0.00012494446518477047,
    "timestep_mean": 669.125,
    "epsilon_norm": 4.652910232543945,
    "prediction_norm": 4.335396766662598,
    "timestamp": "2025-07-14T12:35:02.354522"
  },
  {
    "epoch": 2461,
    "batch_idx": 0,
    "loss": 0.28565114736557007,
    "learning_rate": 0.00012546432080176123,
    "timestep_mean": 417.5,
    "epsilon_norm": 6.505520820617676,
    "prediction_norm": 4.729513645172119,
    "timestamp": "2025-07-14T12:35:02.359736"
  },
  {
    "epoch": 2461,
    "batch_idx": 32,
    "loss": 0.19647252559661865,
    "learning_rate": 0.00012546432080176123,
    "timestep_mean": 483.1875,
    "epsilon_norm": 8.114607810974121,
    "prediction_norm": 6.957645416259766,
    "timestamp": "2025-07-14T12:35:02.364907"
  },
  {
    "epoch": 2461,
    "batch_idx": 64,
    "loss": 0.23667854070663452,
    "learning_rate": 0.00012546432080176123,
    "timestep_mean": 583.65625,
    "epsilon_norm": 8.418851852416992,
    "prediction_norm": 7.94557523727417,
    "timestamp": "2025-07-14T12:35:02.370143"
  },
  {
    "epoch": 2461,
    "batch_idx": 96,
    "loss": 0.4015994966030121,
    "learning_rate": 0.00012546432080176123,
    "timestep_mean": 542.90625,
    "epsilon_norm": 8.558807373046875,
    "prediction_norm": 6.9804534912109375,
    "timestamp": "2025-07-14T12:35:02.375496"
  },
  {
    "epoch": 2461,
    "batch_idx": 128,
    "loss": 0.16521607339382172,
    "learning_rate": 0.00012546432080176123,
    "timestep_mean": 488.8125,
    "epsilon_norm": 6.932889938354492,
    "prediction_norm": 6.874530792236328,
    "timestamp": "2025-07-14T12:35:02.380710"
  },
  {
    "epoch": 2461,
    "batch_idx": 160,
    "loss": 0.24555885791778564,
    "learning_rate": 0.00012546432080176123,
    "timestep_mean": 464.0,
    "epsilon_norm": 8.780512809753418,
    "prediction_norm": 7.92612886428833,
    "timestamp": "2025-07-14T12:35:02.385928"
  },
  {
    "epoch": 2461,
    "batch_idx": 192,
    "loss": 0.2655514180660248,
    "learning_rate": 0.00012546432080176123,
    "timestep_mean": 559.75,
    "epsilon_norm": 4.523105144500732,
    "prediction_norm": 3.8043053150177,
    "timestamp": "2025-07-14T12:35:02.391116"
  },
  {
    "epoch": 2462,
    "batch_idx": 0,
    "loss": 0.22817310690879822,
    "learning_rate": 0.00012598510054830908,
    "timestep_mean": 568.84375,
    "epsilon_norm": 7.771637439727783,
    "prediction_norm": 7.518332004547119,
    "timestamp": "2025-07-14T12:35:02.396256"
  },
  {
    "epoch": 2462,
    "batch_idx": 32,
    "loss": 0.39793020486831665,
    "learning_rate": 0.00012598510054830908,
    "timestep_mean": 490.0625,
    "epsilon_norm": 7.971259593963623,
    "prediction_norm": 5.949859142303467,
    "timestamp": "2025-07-14T12:35:02.401335"
  },
  {
    "epoch": 2462,
    "batch_idx": 64,
    "loss": 0.23074546456336975,
    "learning_rate": 0.00012598510054830908,
    "timestep_mean": 431.15625,
    "epsilon_norm": 7.367683410644531,
    "prediction_norm": 6.589780330657959,
    "timestamp": "2025-07-14T12:35:02.406173"
  },
  {
    "epoch": 2462,
    "batch_idx": 96,
    "loss": 0.3603018820285797,
    "learning_rate": 0.00012598510054830908,
    "timestep_mean": 442.25,
    "epsilon_norm": 8.310344696044922,
    "prediction_norm": 6.160833835601807,
    "timestamp": "2025-07-14T12:35:02.411069"
  },
  {
    "epoch": 2462,
    "batch_idx": 128,
    "loss": 0.22059397399425507,
    "learning_rate": 0.00012598510054830908,
    "timestep_mean": 520.21875,
    "epsilon_norm": 7.9344282150268555,
    "prediction_norm": 6.357058525085449,
    "timestamp": "2025-07-14T12:35:02.416171"
  },
  {
    "epoch": 2462,
    "batch_idx": 160,
    "loss": 0.5547961592674255,
    "learning_rate": 0.00012598510054830908,
    "timestep_mean": 440.09375,
    "epsilon_norm": 8.2705717086792,
    "prediction_norm": 6.337059497833252,
    "timestamp": "2025-07-14T12:35:02.421152"
  },
  {
    "epoch": 2462,
    "batch_idx": 192,
    "loss": 0.06680111587047577,
    "learning_rate": 0.00012598510054830908,
    "timestep_mean": 656.25,
    "epsilon_norm": 2.610085964202881,
    "prediction_norm": 2.448606014251709,
    "timestamp": "2025-07-14T12:35:02.426328"
  },
  {
    "epoch": 2463,
    "batch_idx": 0,
    "loss": 0.2132261097431183,
    "learning_rate": 0.00012650680313944145,
    "timestep_mean": 529.03125,
    "epsilon_norm": 7.665284156799316,
    "prediction_norm": 6.016353607177734,
    "timestamp": "2025-07-14T12:35:02.431288"
  },
  {
    "epoch": 2463,
    "batch_idx": 32,
    "loss": 0.28109851479530334,
    "learning_rate": 0.00012650680313944145,
    "timestep_mean": 491.59375,
    "epsilon_norm": 8.157841682434082,
    "prediction_norm": 7.062737464904785,
    "timestamp": "2025-07-14T12:35:02.436360"
  },
  {
    "epoch": 2463,
    "batch_idx": 64,
    "loss": 0.23179973661899567,
    "learning_rate": 0.00012650680313944145,
    "timestep_mean": 449.21875,
    "epsilon_norm": 7.529074192047119,
    "prediction_norm": 6.897960662841797,
    "timestamp": "2025-07-14T12:35:02.441249"
  },
  {
    "epoch": 2463,
    "batch_idx": 96,
    "loss": 0.4019234776496887,
    "learning_rate": 0.00012650680313944145,
    "timestep_mean": 409.1875,
    "epsilon_norm": 8.803893089294434,
    "prediction_norm": 6.423601150512695,
    "timestamp": "2025-07-14T12:35:02.446152"
  },
  {
    "epoch": 2463,
    "batch_idx": 128,
    "loss": 0.3254210352897644,
    "learning_rate": 0.00012650680313944145,
    "timestep_mean": 405.4375,
    "epsilon_norm": 8.996136665344238,
    "prediction_norm": 7.737043380737305,
    "timestamp": "2025-07-14T12:35:02.451072"
  },
  {
    "epoch": 2463,
    "batch_idx": 160,
    "loss": 0.29478874802589417,
    "learning_rate": 0.00012650680313944145,
    "timestep_mean": 495.96875,
    "epsilon_norm": 7.62346076965332,
    "prediction_norm": 6.505724906921387,
    "timestamp": "2025-07-14T12:35:02.455935"
  },
  {
    "epoch": 2463,
    "batch_idx": 192,
    "loss": 0.300939679145813,
    "learning_rate": 0.00012650680313944145,
    "timestep_mean": 517.25,
    "epsilon_norm": 4.036101341247559,
    "prediction_norm": 3.211333751678467,
    "timestamp": "2025-07-14T12:35:02.460780"
  },
  {
    "epoch": 2464,
    "batch_idx": 0,
    "loss": 0.14342650771141052,
    "learning_rate": 0.0001270294272879092,
    "timestep_mean": 526.5,
    "epsilon_norm": 7.402267932891846,
    "prediction_norm": 6.841665267944336,
    "timestamp": "2025-07-14T12:35:02.465639"
  },
  {
    "epoch": 2464,
    "batch_idx": 32,
    "loss": 0.17418715357780457,
    "learning_rate": 0.0001270294272879092,
    "timestep_mean": 572.90625,
    "epsilon_norm": 8.109241485595703,
    "prediction_norm": 7.2850446701049805,
    "timestamp": "2025-07-14T12:35:02.470547"
  },
  {
    "epoch": 2464,
    "batch_idx": 64,
    "loss": 0.2731817960739136,
    "learning_rate": 0.0001270294272879092,
    "timestep_mean": 534.09375,
    "epsilon_norm": 8.551458358764648,
    "prediction_norm": 7.8247270584106445,
    "timestamp": "2025-07-14T12:35:02.475520"
  },
  {
    "epoch": 2464,
    "batch_idx": 96,
    "loss": 0.5750815868377686,
    "learning_rate": 0.0001270294272879092,
    "timestep_mean": 462.5,
    "epsilon_norm": 9.419811248779297,
    "prediction_norm": 6.837067604064941,
    "timestamp": "2025-07-14T12:35:02.480571"
  },
  {
    "epoch": 2464,
    "batch_idx": 128,
    "loss": 0.3250967860221863,
    "learning_rate": 0.0001270294272879092,
    "timestep_mean": 485.5625,
    "epsilon_norm": 7.54243278503418,
    "prediction_norm": 6.52068567276001,
    "timestamp": "2025-07-14T12:35:02.485795"
  },
  {
    "epoch": 2464,
    "batch_idx": 160,
    "loss": 0.2847030460834503,
    "learning_rate": 0.0001270294272879092,
    "timestep_mean": 402.1875,
    "epsilon_norm": 6.426549911499023,
    "prediction_norm": 4.536596775054932,
    "timestamp": "2025-07-14T12:35:02.490947"
  },
  {
    "epoch": 2464,
    "batch_idx": 192,
    "loss": 0.3265303373336792,
    "learning_rate": 0.0001270294272879092,
    "timestep_mean": 596.5,
    "epsilon_norm": 4.457396507263184,
    "prediction_norm": 3.449284315109253,
    "timestamp": "2025-07-14T12:35:02.496030"
  },
  {
    "epoch": 2465,
    "batch_idx": 0,
    "loss": 0.31776779890060425,
    "learning_rate": 0.0001275529717041893,
    "timestep_mean": 450.0,
    "epsilon_norm": 7.467044353485107,
    "prediction_norm": 5.734023094177246,
    "timestamp": "2025-07-14T12:35:02.501162"
  },
  {
    "epoch": 2465,
    "batch_idx": 32,
    "loss": 0.3782263696193695,
    "learning_rate": 0.0001275529717041893,
    "timestep_mean": 480.96875,
    "epsilon_norm": 7.611216068267822,
    "prediction_norm": 5.502845287322998,
    "timestamp": "2025-07-14T12:35:02.506293"
  },
  {
    "epoch": 2465,
    "batch_idx": 64,
    "loss": 0.2778942584991455,
    "learning_rate": 0.0001275529717041893,
    "timestep_mean": 502.03125,
    "epsilon_norm": 8.187853813171387,
    "prediction_norm": 6.7346086502075195,
    "timestamp": "2025-07-14T12:35:02.511419"
  },
  {
    "epoch": 2465,
    "batch_idx": 96,
    "loss": 0.37105703353881836,
    "learning_rate": 0.0001275529717041893,
    "timestep_mean": 457.28125,
    "epsilon_norm": 7.944967269897461,
    "prediction_norm": 5.595866680145264,
    "timestamp": "2025-07-14T12:35:02.516600"
  },
  {
    "epoch": 2465,
    "batch_idx": 128,
    "loss": 0.30602705478668213,
    "learning_rate": 0.0001275529717041893,
    "timestep_mean": 475.21875,
    "epsilon_norm": 8.266146659851074,
    "prediction_norm": 6.552256107330322,
    "timestamp": "2025-07-14T12:35:02.521859"
  },
  {
    "epoch": 2465,
    "batch_idx": 160,
    "loss": 0.18049749732017517,
    "learning_rate": 0.0001275529717041893,
    "timestep_mean": 392.09375,
    "epsilon_norm": 8.483166694641113,
    "prediction_norm": 7.656795024871826,
    "timestamp": "2025-07-14T12:35:02.527071"
  },
  {
    "epoch": 2465,
    "batch_idx": 192,
    "loss": 0.6413092613220215,
    "learning_rate": 0.0001275529717041893,
    "timestep_mean": 351.625,
    "epsilon_norm": 3.5448853969573975,
    "prediction_norm": 2.731257915496826,
    "timestamp": "2025-07-14T12:35:02.532233"
  },
  {
    "epoch": 2466,
    "batch_idx": 0,
    "loss": 0.1409115344285965,
    "learning_rate": 0.00012807743509648763,
    "timestep_mean": 569.90625,
    "epsilon_norm": 7.671026706695557,
    "prediction_norm": 6.518370628356934,
    "timestamp": "2025-07-14T12:35:02.537477"
  },
  {
    "epoch": 2466,
    "batch_idx": 32,
    "loss": 0.3108883798122406,
    "learning_rate": 0.00012807743509648763,
    "timestep_mean": 535.15625,
    "epsilon_norm": 8.45020866394043,
    "prediction_norm": 7.171254634857178,
    "timestamp": "2025-07-14T12:35:02.542618"
  },
  {
    "epoch": 2466,
    "batch_idx": 64,
    "loss": 0.36232447624206543,
    "learning_rate": 0.00012807743509648763,
    "timestep_mean": 496.6875,
    "epsilon_norm": 8.939718246459961,
    "prediction_norm": 7.318530082702637,
    "timestamp": "2025-07-14T12:35:02.547633"
  },
  {
    "epoch": 2466,
    "batch_idx": 96,
    "loss": 0.336953341960907,
    "learning_rate": 0.00012807743509648763,
    "timestep_mean": 501.71875,
    "epsilon_norm": 7.812839508056641,
    "prediction_norm": 6.501004695892334,
    "timestamp": "2025-07-14T12:35:02.552845"
  },
  {
    "epoch": 2466,
    "batch_idx": 128,
    "loss": 0.20321610569953918,
    "learning_rate": 0.00012807743509648763,
    "timestep_mean": 410.375,
    "epsilon_norm": 7.43761682510376,
    "prediction_norm": 6.147639751434326,
    "timestamp": "2025-07-14T12:35:02.557950"
  },
  {
    "epoch": 2466,
    "batch_idx": 160,
    "loss": 0.2257676124572754,
    "learning_rate": 0.00012807743509648763,
    "timestep_mean": 572.3125,
    "epsilon_norm": 6.937847137451172,
    "prediction_norm": 5.870375633239746,
    "timestamp": "2025-07-14T12:35:02.562941"
  },
  {
    "epoch": 2466,
    "batch_idx": 192,
    "loss": 0.055504895746707916,
    "learning_rate": 0.00012807743509648763,
    "timestep_mean": 622.25,
    "epsilon_norm": 3.887300729751587,
    "prediction_norm": 3.7753262519836426,
    "timestamp": "2025-07-14T12:35:02.568073"
  },
  {
    "epoch": 2467,
    "batch_idx": 0,
    "loss": 0.45630159974098206,
    "learning_rate": 0.0001286028161707434,
    "timestep_mean": 493.40625,
    "epsilon_norm": 7.832283020019531,
    "prediction_norm": 6.697312831878662,
    "timestamp": "2025-07-14T12:35:02.573226"
  },
  {
    "epoch": 2467,
    "batch_idx": 32,
    "loss": 0.2562323808670044,
    "learning_rate": 0.0001286028161707434,
    "timestep_mean": 499.8125,
    "epsilon_norm": 7.757646083831787,
    "prediction_norm": 6.366115093231201,
    "timestamp": "2025-07-14T12:35:02.578273"
  },
  {
    "epoch": 2467,
    "batch_idx": 64,
    "loss": 0.33603984117507935,
    "learning_rate": 0.0001286028161707434,
    "timestep_mean": 472.28125,
    "epsilon_norm": 7.723268985748291,
    "prediction_norm": 6.425026893615723,
    "timestamp": "2025-07-14T12:35:02.583286"
  },
  {
    "epoch": 2467,
    "batch_idx": 96,
    "loss": 0.3413185775279999,
    "learning_rate": 0.0001286028161707434,
    "timestep_mean": 461.375,
    "epsilon_norm": 8.090889930725098,
    "prediction_norm": 7.128986358642578,
    "timestamp": "2025-07-14T12:35:02.588503"
  },
  {
    "epoch": 2467,
    "batch_idx": 128,
    "loss": 0.15170744061470032,
    "learning_rate": 0.0001286028161707434,
    "timestep_mean": 471.15625,
    "epsilon_norm": 7.929049968719482,
    "prediction_norm": 7.484124183654785,
    "timestamp": "2025-07-14T12:35:02.593728"
  },
  {
    "epoch": 2467,
    "batch_idx": 160,
    "loss": 0.20157459378242493,
    "learning_rate": 0.0001286028161707434,
    "timestep_mean": 508.34375,
    "epsilon_norm": 8.807543754577637,
    "prediction_norm": 7.874764919281006,
    "timestamp": "2025-07-14T12:35:02.598963"
  },
  {
    "epoch": 2467,
    "batch_idx": 192,
    "loss": 0.7057549953460693,
    "learning_rate": 0.0001286028161707434,
    "timestep_mean": 380.625,
    "epsilon_norm": 3.305925130844116,
    "prediction_norm": 1.937769889831543,
    "timestamp": "2025-07-14T12:35:02.604094"
  },
  {
    "epoch": 2468,
    "batch_idx": 0,
    "loss": 0.15090930461883545,
    "learning_rate": 0.00012912911363063064,
    "timestep_mean": 501.46875,
    "epsilon_norm": 7.764732360839844,
    "prediction_norm": 7.0898661613464355,
    "timestamp": "2025-07-14T12:35:02.609366"
  },
  {
    "epoch": 2468,
    "batch_idx": 32,
    "loss": 0.28862130641937256,
    "learning_rate": 0.00012912911363063064,
    "timestep_mean": 462.53125,
    "epsilon_norm": 7.802978992462158,
    "prediction_norm": 6.380351543426514,
    "timestamp": "2025-07-14T12:35:02.615219"
  },
  {
    "epoch": 2468,
    "batch_idx": 64,
    "loss": 0.18457795679569244,
    "learning_rate": 0.00012912911363063064,
    "timestep_mean": 494.84375,
    "epsilon_norm": 8.023331642150879,
    "prediction_norm": 7.803031921386719,
    "timestamp": "2025-07-14T12:35:02.620771"
  },
  {
    "epoch": 2468,
    "batch_idx": 96,
    "loss": 0.1990348845720291,
    "learning_rate": 0.00012912911363063064,
    "timestep_mean": 514.6875,
    "epsilon_norm": 7.525053977966309,
    "prediction_norm": 6.994021415710449,
    "timestamp": "2025-07-14T12:35:02.626290"
  },
  {
    "epoch": 2468,
    "batch_idx": 128,
    "loss": 0.3126566410064697,
    "learning_rate": 0.00012912911363063064,
    "timestep_mean": 509.3125,
    "epsilon_norm": 8.835524559020996,
    "prediction_norm": 7.146361351013184,
    "timestamp": "2025-07-14T12:35:02.631842"
  },
  {
    "epoch": 2468,
    "batch_idx": 160,
    "loss": 0.20201168954372406,
    "learning_rate": 0.00012912911363063064,
    "timestep_mean": 489.4375,
    "epsilon_norm": 7.415812969207764,
    "prediction_norm": 7.191676139831543,
    "timestamp": "2025-07-14T12:35:02.637126"
  },
  {
    "epoch": 2468,
    "batch_idx": 192,
    "loss": 0.38505950570106506,
    "learning_rate": 0.00012912911363063064,
    "timestep_mean": 459.25,
    "epsilon_norm": 3.834172487258911,
    "prediction_norm": 2.6997056007385254,
    "timestamp": "2025-07-14T12:35:02.642401"
  },
  {
    "epoch": 2469,
    "batch_idx": 0,
    "loss": 0.21157220005989075,
    "learning_rate": 0.0001296563261775628,
    "timestep_mean": 489.3125,
    "epsilon_norm": 8.035001754760742,
    "prediction_norm": 6.842315673828125,
    "timestamp": "2025-07-14T12:35:02.647388"
  },
  {
    "epoch": 2469,
    "batch_idx": 32,
    "loss": 0.09649749100208282,
    "learning_rate": 0.0001296563261775628,
    "timestep_mean": 607.0,
    "epsilon_norm": 8.531026840209961,
    "prediction_norm": 7.957900524139404,
    "timestamp": "2025-07-14T12:35:02.652257"
  },
  {
    "epoch": 2469,
    "batch_idx": 64,
    "loss": 0.317222535610199,
    "learning_rate": 0.0001296563261775628,
    "timestep_mean": 455.40625,
    "epsilon_norm": 8.027421951293945,
    "prediction_norm": 5.884765148162842,
    "timestamp": "2025-07-14T12:35:02.657181"
  },
  {
    "epoch": 2469,
    "batch_idx": 96,
    "loss": 0.3613821864128113,
    "learning_rate": 0.0001296563261775628,
    "timestep_mean": 470.71875,
    "epsilon_norm": 7.910857677459717,
    "prediction_norm": 5.945577144622803,
    "timestamp": "2025-07-14T12:35:02.662043"
  },
  {
    "epoch": 2469,
    "batch_idx": 128,
    "loss": 0.20149321854114532,
    "learning_rate": 0.0001296563261775628,
    "timestep_mean": 448.5625,
    "epsilon_norm": 7.682455062866211,
    "prediction_norm": 6.741691589355469,
    "timestamp": "2025-07-14T12:35:02.667004"
  },
  {
    "epoch": 2469,
    "batch_idx": 160,
    "loss": 0.19929364323616028,
    "learning_rate": 0.0001296563261775628,
    "timestep_mean": 485.15625,
    "epsilon_norm": 8.285888671875,
    "prediction_norm": 7.080260276794434,
    "timestamp": "2025-07-14T12:35:02.672089"
  },
  {
    "epoch": 2469,
    "batch_idx": 192,
    "loss": 0.41538164019584656,
    "learning_rate": 0.0001296563261775628,
    "timestep_mean": 507.875,
    "epsilon_norm": 4.7103190422058105,
    "prediction_norm": 3.3492915630340576,
    "timestamp": "2025-07-14T12:35:02.677085"
  },
  {
    "epoch": 2470,
    "batch_idx": 0,
    "loss": 0.31402668356895447,
    "learning_rate": 0.00013018445251069543,
    "timestep_mean": 562.0625,
    "epsilon_norm": 8.110188484191895,
    "prediction_norm": 7.356996536254883,
    "timestamp": "2025-07-14T12:35:02.682232"
  },
  {
    "epoch": 2470,
    "batch_idx": 32,
    "loss": 0.27243122458457947,
    "learning_rate": 0.00013018445251069543,
    "timestep_mean": 510.9375,
    "epsilon_norm": 8.02099323272705,
    "prediction_norm": 6.362903118133545,
    "timestamp": "2025-07-14T12:35:02.687349"
  },
  {
    "epoch": 2470,
    "batch_idx": 64,
    "loss": 0.2197023183107376,
    "learning_rate": 0.00013018445251069543,
    "timestep_mean": 553.75,
    "epsilon_norm": 9.441044807434082,
    "prediction_norm": 8.183226585388184,
    "timestamp": "2025-07-14T12:35:02.692496"
  },
  {
    "epoch": 2470,
    "batch_idx": 96,
    "loss": 0.23778721690177917,
    "learning_rate": 0.00013018445251069543,
    "timestep_mean": 496.0625,
    "epsilon_norm": 6.222710609436035,
    "prediction_norm": 5.397019863128662,
    "timestamp": "2025-07-14T12:35:02.697566"
  },
  {
    "epoch": 2470,
    "batch_idx": 128,
    "loss": 0.24253995716571808,
    "learning_rate": 0.00013018445251069543,
    "timestep_mean": 537.84375,
    "epsilon_norm": 8.554749488830566,
    "prediction_norm": 7.127801418304443,
    "timestamp": "2025-07-14T12:35:02.702739"
  },
  {
    "epoch": 2470,
    "batch_idx": 160,
    "loss": 0.10125931352376938,
    "learning_rate": 0.00013018445251069543,
    "timestep_mean": 530.125,
    "epsilon_norm": 6.620540618896484,
    "prediction_norm": 6.276020050048828,
    "timestamp": "2025-07-14T12:35:02.708231"
  },
  {
    "epoch": 2470,
    "batch_idx": 192,
    "loss": 0.13394218683242798,
    "learning_rate": 0.00013018445251069543,
    "timestep_mean": 317.75,
    "epsilon_norm": 3.1647868156433105,
    "prediction_norm": 2.6704182624816895,
    "timestamp": "2025-07-14T12:35:02.713687"
  },
  {
    "epoch": 2471,
    "batch_idx": 0,
    "loss": 0.2551928758621216,
    "learning_rate": 0.000130713491326929,
    "timestep_mean": 569.96875,
    "epsilon_norm": 8.534028053283691,
    "prediction_norm": 7.254890441894531,
    "timestamp": "2025-07-14T12:35:02.718976"
  },
  {
    "epoch": 2471,
    "batch_idx": 32,
    "loss": 0.26418131589889526,
    "learning_rate": 0.000130713491326929,
    "timestep_mean": 489.0625,
    "epsilon_norm": 6.824492454528809,
    "prediction_norm": 5.981907367706299,
    "timestamp": "2025-07-14T12:35:02.724143"
  },
  {
    "epoch": 2471,
    "batch_idx": 64,
    "loss": 0.4398559033870697,
    "learning_rate": 0.000130713491326929,
    "timestep_mean": 435.625,
    "epsilon_norm": 8.558247566223145,
    "prediction_norm": 7.000121116638184,
    "timestamp": "2025-07-14T12:35:02.729280"
  },
  {
    "epoch": 2471,
    "batch_idx": 96,
    "loss": 0.18245437741279602,
    "learning_rate": 0.000130713491326929,
    "timestep_mean": 509.5,
    "epsilon_norm": 7.057462215423584,
    "prediction_norm": 6.32728385925293,
    "timestamp": "2025-07-14T12:35:02.734544"
  },
  {
    "epoch": 2471,
    "batch_idx": 128,
    "loss": 0.35867762565612793,
    "learning_rate": 0.000130713491326929,
    "timestep_mean": 425.40625,
    "epsilon_norm": 7.778858184814453,
    "prediction_norm": 6.077085018157959,
    "timestamp": "2025-07-14T12:35:02.739737"
  },
  {
    "epoch": 2471,
    "batch_idx": 160,
    "loss": 0.23616923391819,
    "learning_rate": 0.000130713491326929,
    "timestep_mean": 540.875,
    "epsilon_norm": 8.448354721069336,
    "prediction_norm": 6.994546890258789,
    "timestamp": "2025-07-14T12:35:02.744868"
  },
  {
    "epoch": 2471,
    "batch_idx": 192,
    "loss": 0.08721444010734558,
    "learning_rate": 0.000130713491326929,
    "timestep_mean": 547.125,
    "epsilon_norm": 3.086021661758423,
    "prediction_norm": 2.409471273422241,
    "timestamp": "2025-07-14T12:35:02.750046"
  },
  {
    "epoch": 2472,
    "batch_idx": 0,
    "loss": 0.20154133439064026,
    "learning_rate": 0.00013124344132091332,
    "timestep_mean": 511.375,
    "epsilon_norm": 8.072370529174805,
    "prediction_norm": 7.091299533843994,
    "timestamp": "2025-07-14T12:35:02.755383"
  },
  {
    "epoch": 2472,
    "batch_idx": 32,
    "loss": 0.09146825969219208,
    "learning_rate": 0.00013124344132091332,
    "timestep_mean": 535.4375,
    "epsilon_norm": 7.463101387023926,
    "prediction_norm": 7.1882710456848145,
    "timestamp": "2025-07-14T12:35:02.760434"
  },
  {
    "epoch": 2472,
    "batch_idx": 64,
    "loss": 0.24557630717754364,
    "learning_rate": 0.00013124344132091332,
    "timestep_mean": 550.4375,
    "epsilon_norm": 8.03048324584961,
    "prediction_norm": 7.1402153968811035,
    "timestamp": "2025-07-14T12:35:02.765660"
  },
  {
    "epoch": 2472,
    "batch_idx": 96,
    "loss": 0.21031129360198975,
    "learning_rate": 0.00013124344132091332,
    "timestep_mean": 528.28125,
    "epsilon_norm": 7.648778438568115,
    "prediction_norm": 6.398907661437988,
    "timestamp": "2025-07-14T12:35:02.770820"
  },
  {
    "epoch": 2472,
    "batch_idx": 128,
    "loss": 0.15954114496707916,
    "learning_rate": 0.00013124344132091332,
    "timestep_mean": 501.5,
    "epsilon_norm": 8.642417907714844,
    "prediction_norm": 8.130517959594727,
    "timestamp": "2025-07-14T12:35:02.776031"
  },
  {
    "epoch": 2472,
    "batch_idx": 160,
    "loss": 0.31774598360061646,
    "learning_rate": 0.00013124344132091332,
    "timestep_mean": 524.34375,
    "epsilon_norm": 7.635526657104492,
    "prediction_norm": 5.939126014709473,
    "timestamp": "2025-07-14T12:35:02.781286"
  },
  {
    "epoch": 2472,
    "batch_idx": 192,
    "loss": 0.029192257672548294,
    "learning_rate": 0.00013124344132091332,
    "timestep_mean": 719.25,
    "epsilon_norm": 3.931673526763916,
    "prediction_norm": 3.5576133728027344,
    "timestamp": "2025-07-14T12:35:02.786475"
  },
  {
    "epoch": 2473,
    "batch_idx": 0,
    "loss": 0.2514365315437317,
    "learning_rate": 0.00013177430118504896,
    "timestep_mean": 492.84375,
    "epsilon_norm": 9.258237838745117,
    "prediction_norm": 7.749692440032959,
    "timestamp": "2025-07-14T12:35:02.791656"
  },
  {
    "epoch": 2473,
    "batch_idx": 32,
    "loss": 0.15564420819282532,
    "learning_rate": 0.00013177430118504896,
    "timestep_mean": 490.0625,
    "epsilon_norm": 8.27140998840332,
    "prediction_norm": 7.429356098175049,
    "timestamp": "2025-07-14T12:35:02.797028"
  },
  {
    "epoch": 2473,
    "batch_idx": 64,
    "loss": 0.20023180544376373,
    "learning_rate": 0.00013177430118504896,
    "timestep_mean": 537.65625,
    "epsilon_norm": 8.635010719299316,
    "prediction_norm": 7.996188640594482,
    "timestamp": "2025-07-14T12:35:02.802222"
  },
  {
    "epoch": 2473,
    "batch_idx": 96,
    "loss": 0.20857584476470947,
    "learning_rate": 0.00013177430118504896,
    "timestep_mean": 538.34375,
    "epsilon_norm": 7.4386677742004395,
    "prediction_norm": 6.517767429351807,
    "timestamp": "2025-07-14T12:35:02.807395"
  },
  {
    "epoch": 2473,
    "batch_idx": 128,
    "loss": 0.6017513871192932,
    "learning_rate": 0.00013177430118504896,
    "timestep_mean": 542.75,
    "epsilon_norm": 9.09156322479248,
    "prediction_norm": 7.505527019500732,
    "timestamp": "2025-07-14T12:35:02.812556"
  },
  {
    "epoch": 2473,
    "batch_idx": 160,
    "loss": 0.23980680108070374,
    "learning_rate": 0.00013177430118504896,
    "timestep_mean": 526.59375,
    "epsilon_norm": 6.952910900115967,
    "prediction_norm": 6.661520004272461,
    "timestamp": "2025-07-14T12:35:02.817719"
  },
  {
    "epoch": 2473,
    "batch_idx": 192,
    "loss": 0.6029407978057861,
    "learning_rate": 0.00013177430118504896,
    "timestep_mean": 484.25,
    "epsilon_norm": 4.2922492027282715,
    "prediction_norm": 2.748063564300537,
    "timestamp": "2025-07-14T12:35:02.822852"
  },
  {
    "epoch": 2474,
    "batch_idx": 0,
    "loss": 0.2804161310195923,
    "learning_rate": 0.00013230606960949227,
    "timestep_mean": 521.21875,
    "epsilon_norm": 8.602264404296875,
    "prediction_norm": 7.735194206237793,
    "timestamp": "2025-07-14T12:35:02.828083"
  },
  {
    "epoch": 2474,
    "batch_idx": 32,
    "loss": 0.31716665625572205,
    "learning_rate": 0.00013230606960949227,
    "timestep_mean": 443.53125,
    "epsilon_norm": 7.255444049835205,
    "prediction_norm": 5.737146854400635,
    "timestamp": "2025-07-14T12:35:02.833240"
  },
  {
    "epoch": 2474,
    "batch_idx": 64,
    "loss": 0.3040432631969452,
    "learning_rate": 0.00013230606960949227,
    "timestep_mean": 466.5,
    "epsilon_norm": 8.36712646484375,
    "prediction_norm": 6.76095724105835,
    "timestamp": "2025-07-14T12:35:02.838486"
  },
  {
    "epoch": 2474,
    "batch_idx": 96,
    "loss": 0.3636035919189453,
    "learning_rate": 0.00013230606960949227,
    "timestep_mean": 539.03125,
    "epsilon_norm": 7.395891189575195,
    "prediction_norm": 6.139933109283447,
    "timestamp": "2025-07-14T12:35:02.843683"
  },
  {
    "epoch": 2474,
    "batch_idx": 128,
    "loss": 0.36319059133529663,
    "learning_rate": 0.00013230606960949227,
    "timestep_mean": 484.59375,
    "epsilon_norm": 7.98886775970459,
    "prediction_norm": 6.017143249511719,
    "timestamp": "2025-07-14T12:35:02.848796"
  },
  {
    "epoch": 2474,
    "batch_idx": 160,
    "loss": 0.312364786863327,
    "learning_rate": 0.00013230606960949227,
    "timestep_mean": 505.96875,
    "epsilon_norm": 8.988664627075195,
    "prediction_norm": 7.72844934463501,
    "timestamp": "2025-07-14T12:35:02.853847"
  },
  {
    "epoch": 2474,
    "batch_idx": 192,
    "loss": 0.41219136118888855,
    "learning_rate": 0.00013230606960949227,
    "timestep_mean": 453.75,
    "epsilon_norm": 3.876091718673706,
    "prediction_norm": 3.344066858291626,
    "timestamp": "2025-07-14T12:35:02.859103"
  },
  {
    "epoch": 2475,
    "batch_idx": 0,
    "loss": 0.2950696051120758,
    "learning_rate": 0.00013283874528215758,
    "timestep_mean": 441.125,
    "epsilon_norm": 7.232916355133057,
    "prediction_norm": 6.15603494644165,
    "timestamp": "2025-07-14T12:35:02.864258"
  },
  {
    "epoch": 2475,
    "batch_idx": 32,
    "loss": 0.4551185965538025,
    "learning_rate": 0.00013283874528215758,
    "timestep_mean": 458.03125,
    "epsilon_norm": 8.131584167480469,
    "prediction_norm": 6.69944953918457,
    "timestamp": "2025-07-14T12:35:02.869499"
  },
  {
    "epoch": 2475,
    "batch_idx": 64,
    "loss": 0.2943689525127411,
    "learning_rate": 0.00013283874528215758,
    "timestep_mean": 494.0625,
    "epsilon_norm": 7.792794227600098,
    "prediction_norm": 6.412148475646973,
    "timestamp": "2025-07-14T12:35:02.874725"
  },
  {
    "epoch": 2475,
    "batch_idx": 96,
    "loss": 0.21358484029769897,
    "learning_rate": 0.00013283874528215758,
    "timestep_mean": 507.84375,
    "epsilon_norm": 8.218936920166016,
    "prediction_norm": 6.8713178634643555,
    "timestamp": "2025-07-14T12:35:02.879970"
  },
  {
    "epoch": 2475,
    "batch_idx": 128,
    "loss": 0.23517762124538422,
    "learning_rate": 0.00013283874528215758,
    "timestep_mean": 544.34375,
    "epsilon_norm": 7.86557149887085,
    "prediction_norm": 6.70982027053833,
    "timestamp": "2025-07-14T12:35:02.885102"
  },
  {
    "epoch": 2475,
    "batch_idx": 160,
    "loss": 0.22673766314983368,
    "learning_rate": 0.00013283874528215758,
    "timestep_mean": 475.09375,
    "epsilon_norm": 8.303882598876953,
    "prediction_norm": 7.450920581817627,
    "timestamp": "2025-07-14T12:35:02.890315"
  },
  {
    "epoch": 2475,
    "batch_idx": 192,
    "loss": 0.2717088460922241,
    "learning_rate": 0.00013283874528215758,
    "timestep_mean": 529.625,
    "epsilon_norm": 4.854861736297607,
    "prediction_norm": 4.082205295562744,
    "timestamp": "2025-07-14T12:35:02.895595"
  },
  {
    "epoch": 2476,
    "batch_idx": 0,
    "loss": 0.24589958786964417,
    "learning_rate": 0.00013337232688872029,
    "timestep_mean": 436.40625,
    "epsilon_norm": 7.634149074554443,
    "prediction_norm": 6.358457565307617,
    "timestamp": "2025-07-14T12:35:02.900840"
  },
  {
    "epoch": 2476,
    "batch_idx": 32,
    "loss": 0.32293933629989624,
    "learning_rate": 0.00013337232688872029,
    "timestep_mean": 410.53125,
    "epsilon_norm": 8.536508560180664,
    "prediction_norm": 7.1962409019470215,
    "timestamp": "2025-07-14T12:35:02.905945"
  },
  {
    "epoch": 2476,
    "batch_idx": 64,
    "loss": 0.10959376394748688,
    "learning_rate": 0.00013337232688872029,
    "timestep_mean": 597.375,
    "epsilon_norm": 6.3271870613098145,
    "prediction_norm": 6.184228897094727,
    "timestamp": "2025-07-14T12:35:02.911066"
  },
  {
    "epoch": 2476,
    "batch_idx": 96,
    "loss": 0.16936740279197693,
    "learning_rate": 0.00013337232688872029,
    "timestep_mean": 505.875,
    "epsilon_norm": 6.582100868225098,
    "prediction_norm": 5.971499443054199,
    "timestamp": "2025-07-14T12:35:02.916233"
  },
  {
    "epoch": 2476,
    "batch_idx": 128,
    "loss": 0.4509130120277405,
    "learning_rate": 0.00013337232688872029,
    "timestep_mean": 477.40625,
    "epsilon_norm": 8.719120979309082,
    "prediction_norm": 6.675444602966309,
    "timestamp": "2025-07-14T12:35:02.921428"
  },
  {
    "epoch": 2476,
    "batch_idx": 160,
    "loss": 0.09979517012834549,
    "learning_rate": 0.00013337232688872029,
    "timestep_mean": 625.6875,
    "epsilon_norm": 8.353899002075195,
    "prediction_norm": 8.203680038452148,
    "timestamp": "2025-07-14T12:35:02.926496"
  },
  {
    "epoch": 2476,
    "batch_idx": 192,
    "loss": 0.3315950632095337,
    "learning_rate": 0.00013337232688872029,
    "timestep_mean": 406.375,
    "epsilon_norm": 4.9635186195373535,
    "prediction_norm": 4.335996627807617,
    "timestamp": "2025-07-14T12:35:02.931643"
  },
  {
    "epoch": 2477,
    "batch_idx": 0,
    "loss": 0.25484511256217957,
    "learning_rate": 0.00013390681311262103,
    "timestep_mean": 426.15625,
    "epsilon_norm": 8.346965789794922,
    "prediction_norm": 7.264099597930908,
    "timestamp": "2025-07-14T12:35:02.936734"
  },
  {
    "epoch": 2477,
    "batch_idx": 32,
    "loss": 0.2015586495399475,
    "learning_rate": 0.00013390681311262103,
    "timestep_mean": 577.46875,
    "epsilon_norm": 8.860860824584961,
    "prediction_norm": 8.352065086364746,
    "timestamp": "2025-07-14T12:35:02.941540"
  },
  {
    "epoch": 2477,
    "batch_idx": 64,
    "loss": 0.41574567556381226,
    "learning_rate": 0.00013390681311262103,
    "timestep_mean": 447.5625,
    "epsilon_norm": 8.190567970275879,
    "prediction_norm": 6.507313251495361,
    "timestamp": "2025-07-14T12:35:02.946501"
  },
  {
    "epoch": 2477,
    "batch_idx": 96,
    "loss": 0.3157165050506592,
    "learning_rate": 0.00013390681311262103,
    "timestep_mean": 518.8125,
    "epsilon_norm": 8.706130027770996,
    "prediction_norm": 7.0436577796936035,
    "timestamp": "2025-07-14T12:35:02.951322"
  },
  {
    "epoch": 2477,
    "batch_idx": 128,
    "loss": 0.36215686798095703,
    "learning_rate": 0.00013390681311262103,
    "timestep_mean": 577.78125,
    "epsilon_norm": 8.45276927947998,
    "prediction_norm": 7.214011192321777,
    "timestamp": "2025-07-14T12:35:02.956362"
  },
  {
    "epoch": 2477,
    "batch_idx": 160,
    "loss": 0.23094362020492554,
    "learning_rate": 0.00013390681311262103,
    "timestep_mean": 523.0,
    "epsilon_norm": 7.576618671417236,
    "prediction_norm": 6.844470024108887,
    "timestamp": "2025-07-14T12:35:02.961281"
  },
  {
    "epoch": 2477,
    "batch_idx": 192,
    "loss": 0.17409320175647736,
    "learning_rate": 0.00013390681311262103,
    "timestep_mean": 444.125,
    "epsilon_norm": 3.0450475215911865,
    "prediction_norm": 2.3316500186920166,
    "timestamp": "2025-07-14T12:35:02.966111"
  },
  {
    "epoch": 2478,
    "batch_idx": 0,
    "loss": 0.27023813128471375,
    "learning_rate": 0.00013444220263506815,
    "timestep_mean": 549.75,
    "epsilon_norm": 7.4395060539245605,
    "prediction_norm": 6.060009479522705,
    "timestamp": "2025-07-14T12:35:02.970961"
  },
  {
    "epoch": 2478,
    "batch_idx": 32,
    "loss": 0.3056318759918213,
    "learning_rate": 0.00013444220263506815,
    "timestep_mean": 453.3125,
    "epsilon_norm": 7.8698320388793945,
    "prediction_norm": 6.602395057678223,
    "timestamp": "2025-07-14T12:35:02.975896"
  },
  {
    "epoch": 2478,
    "batch_idx": 64,
    "loss": 0.11432226747274399,
    "learning_rate": 0.00013444220263506815,
    "timestep_mean": 515.28125,
    "epsilon_norm": 7.462225437164307,
    "prediction_norm": 6.948734760284424,
    "timestamp": "2025-07-14T12:35:02.981069"
  },
  {
    "epoch": 2478,
    "batch_idx": 96,
    "loss": 0.19293059408664703,
    "learning_rate": 0.00013444220263506815,
    "timestep_mean": 495.21875,
    "epsilon_norm": 8.242130279541016,
    "prediction_norm": 7.025914669036865,
    "timestamp": "2025-07-14T12:35:02.992117"
  },
  {
    "epoch": 2478,
    "batch_idx": 128,
    "loss": 0.3258441686630249,
    "learning_rate": 0.00013444220263506815,
    "timestep_mean": 504.6875,
    "epsilon_norm": 7.887675762176514,
    "prediction_norm": 6.28341817855835,
    "timestamp": "2025-07-14T12:35:02.997344"
  },
  {
    "epoch": 2478,
    "batch_idx": 160,
    "loss": 0.42050841450691223,
    "learning_rate": 0.00013444220263506815,
    "timestep_mean": 487.3125,
    "epsilon_norm": 8.318543434143066,
    "prediction_norm": 6.585225582122803,
    "timestamp": "2025-07-14T12:35:03.002202"
  },
  {
    "epoch": 2478,
    "batch_idx": 192,
    "loss": 0.2709266245365143,
    "learning_rate": 0.00013444220263506815,
    "timestep_mean": 588.875,
    "epsilon_norm": 4.319713592529297,
    "prediction_norm": 4.116884231567383,
    "timestamp": "2025-07-14T12:35:03.006999"
  },
  {
    "epoch": 2479,
    "batch_idx": 0,
    "loss": 0.23271474242210388,
    "learning_rate": 0.0001349784941350413,
    "timestep_mean": 573.375,
    "epsilon_norm": 7.3672661781311035,
    "prediction_norm": 6.725098133087158,
    "timestamp": "2025-07-14T12:35:03.011724"
  },
  {
    "epoch": 2479,
    "batch_idx": 32,
    "loss": 0.46411973237991333,
    "learning_rate": 0.0001349784941350413,
    "timestep_mean": 496.28125,
    "epsilon_norm": 8.572110176086426,
    "prediction_norm": 6.999968528747559,
    "timestamp": "2025-07-14T12:35:03.016415"
  },
  {
    "epoch": 2479,
    "batch_idx": 64,
    "loss": 0.22461506724357605,
    "learning_rate": 0.0001349784941350413,
    "timestep_mean": 520.71875,
    "epsilon_norm": 7.953521251678467,
    "prediction_norm": 6.361349582672119,
    "timestamp": "2025-07-14T12:35:03.021299"
  },
  {
    "epoch": 2479,
    "batch_idx": 96,
    "loss": 0.12693655490875244,
    "learning_rate": 0.0001349784941350413,
    "timestep_mean": 619.0,
    "epsilon_norm": 7.761019706726074,
    "prediction_norm": 7.496781826019287,
    "timestamp": "2025-07-14T12:35:03.026186"
  },
  {
    "epoch": 2479,
    "batch_idx": 128,
    "loss": 0.23636232316493988,
    "learning_rate": 0.0001349784941350413,
    "timestep_mean": 598.625,
    "epsilon_norm": 7.80369758605957,
    "prediction_norm": 6.456759452819824,
    "timestamp": "2025-07-14T12:35:03.031117"
  },
  {
    "epoch": 2479,
    "batch_idx": 160,
    "loss": 0.2528071105480194,
    "learning_rate": 0.0001349784941350413,
    "timestep_mean": 456.78125,
    "epsilon_norm": 6.596374988555908,
    "prediction_norm": 5.394485950469971,
    "timestamp": "2025-07-14T12:35:03.036038"
  },
  {
    "epoch": 2479,
    "batch_idx": 192,
    "loss": 0.08029940724372864,
    "learning_rate": 0.0001349784941350413,
    "timestep_mean": 701.125,
    "epsilon_norm": 4.539905548095703,
    "prediction_norm": 4.550182342529297,
    "timestamp": "2025-07-14T12:35:03.040993"
  },
  {
    "epoch": 2480,
    "batch_idx": 0,
    "loss": 0.3345530927181244,
    "learning_rate": 0.00013551568628929455,
    "timestep_mean": 451.96875,
    "epsilon_norm": 8.671577453613281,
    "prediction_norm": 7.337407112121582,
    "timestamp": "2025-07-14T12:35:03.045915"
  },
  {
    "epoch": 2480,
    "batch_idx": 32,
    "loss": 0.25142186880111694,
    "learning_rate": 0.00013551568628929455,
    "timestep_mean": 396.0,
    "epsilon_norm": 6.840582847595215,
    "prediction_norm": 6.178186893463135,
    "timestamp": "2025-07-14T12:35:03.050713"
  },
  {
    "epoch": 2480,
    "batch_idx": 64,
    "loss": 0.10443995893001556,
    "learning_rate": 0.00013551568628929455,
    "timestep_mean": 545.4375,
    "epsilon_norm": 7.413768768310547,
    "prediction_norm": 7.199765682220459,
    "timestamp": "2025-07-14T12:35:03.055564"
  },
  {
    "epoch": 2480,
    "batch_idx": 96,
    "loss": 0.17259931564331055,
    "learning_rate": 0.00013551568628929455,
    "timestep_mean": 546.78125,
    "epsilon_norm": 7.376158714294434,
    "prediction_norm": 6.8159074783325195,
    "timestamp": "2025-07-14T12:35:03.060461"
  },
  {
    "epoch": 2480,
    "batch_idx": 128,
    "loss": 0.20105203986167908,
    "learning_rate": 0.00013551568628929455,
    "timestep_mean": 559.65625,
    "epsilon_norm": 8.109054565429688,
    "prediction_norm": 6.891699314117432,
    "timestamp": "2025-07-14T12:35:03.065303"
  },
  {
    "epoch": 2480,
    "batch_idx": 160,
    "loss": 0.24635347723960876,
    "learning_rate": 0.00013551568628929455,
    "timestep_mean": 502.40625,
    "epsilon_norm": 8.41385269165039,
    "prediction_norm": 7.141746520996094,
    "timestamp": "2025-07-14T12:35:03.070307"
  },
  {
    "epoch": 2480,
    "batch_idx": 192,
    "loss": 0.6470379829406738,
    "learning_rate": 0.00013551568628929455,
    "timestep_mean": 537.0,
    "epsilon_norm": 4.885053634643555,
    "prediction_norm": 4.028631210327148,
    "timestamp": "2025-07-14T12:35:03.075419"
  },
  {
    "epoch": 2481,
    "batch_idx": 0,
    "loss": 0.3138203024864197,
    "learning_rate": 0.00013605377777235945,
    "timestep_mean": 431.15625,
    "epsilon_norm": 8.71963882446289,
    "prediction_norm": 6.767062187194824,
    "timestamp": "2025-07-14T12:35:03.080341"
  },
  {
    "epoch": 2481,
    "batch_idx": 32,
    "loss": 0.18830029666423798,
    "learning_rate": 0.00013605377777235945,
    "timestep_mean": 561.78125,
    "epsilon_norm": 8.209183692932129,
    "prediction_norm": 7.0769548416137695,
    "timestamp": "2025-07-14T12:35:03.085411"
  },
  {
    "epoch": 2481,
    "batch_idx": 64,
    "loss": 0.15702462196350098,
    "learning_rate": 0.00013605377777235945,
    "timestep_mean": 507.125,
    "epsilon_norm": 6.0120391845703125,
    "prediction_norm": 5.115957736968994,
    "timestamp": "2025-07-14T12:35:03.090384"
  },
  {
    "epoch": 2481,
    "batch_idx": 96,
    "loss": 0.4354192614555359,
    "learning_rate": 0.00013605377777235945,
    "timestep_mean": 461.375,
    "epsilon_norm": 8.813349723815918,
    "prediction_norm": 6.689050674438477,
    "timestamp": "2025-07-14T12:35:03.095339"
  },
  {
    "epoch": 2481,
    "batch_idx": 128,
    "loss": 0.15433013439178467,
    "learning_rate": 0.00013605377777235945,
    "timestep_mean": 486.15625,
    "epsilon_norm": 6.303890705108643,
    "prediction_norm": 5.743016242980957,
    "timestamp": "2025-07-14T12:35:03.100459"
  },
  {
    "epoch": 2481,
    "batch_idx": 160,
    "loss": 0.14399626851081848,
    "learning_rate": 0.00013605377777235945,
    "timestep_mean": 555.375,
    "epsilon_norm": 7.794036388397217,
    "prediction_norm": 6.651461124420166,
    "timestamp": "2025-07-14T12:35:03.105359"
  },
  {
    "epoch": 2481,
    "batch_idx": 192,
    "loss": 0.3446664810180664,
    "learning_rate": 0.00013605377777235945,
    "timestep_mean": 500.5,
    "epsilon_norm": 4.105491638183594,
    "prediction_norm": 3.738538980484009,
    "timestamp": "2025-07-14T12:35:03.110308"
  },
  {
    "epoch": 2482,
    "batch_idx": 0,
    "loss": 0.1564212143421173,
    "learning_rate": 0.00013659276725654885,
    "timestep_mean": 548.625,
    "epsilon_norm": 7.889684200286865,
    "prediction_norm": 6.992867469787598,
    "timestamp": "2025-07-14T12:35:03.115209"
  },
  {
    "epoch": 2482,
    "batch_idx": 32,
    "loss": 0.2888135313987732,
    "learning_rate": 0.00013659276725654885,
    "timestep_mean": 549.90625,
    "epsilon_norm": 8.645620346069336,
    "prediction_norm": 7.470658779144287,
    "timestamp": "2025-07-14T12:35:03.120333"
  },
  {
    "epoch": 2482,
    "batch_idx": 64,
    "loss": 0.10545480996370316,
    "learning_rate": 0.00013659276725654885,
    "timestep_mean": 435.46875,
    "epsilon_norm": 7.627228736877441,
    "prediction_norm": 6.861410140991211,
    "timestamp": "2025-07-14T12:35:03.125397"
  },
  {
    "epoch": 2482,
    "batch_idx": 96,
    "loss": 0.32049715518951416,
    "learning_rate": 0.00013659276725654885,
    "timestep_mean": 440.6875,
    "epsilon_norm": 8.327753067016602,
    "prediction_norm": 7.066197872161865,
    "timestamp": "2025-07-14T12:35:03.130574"
  },
  {
    "epoch": 2482,
    "batch_idx": 128,
    "loss": 0.2115209698677063,
    "learning_rate": 0.00013659276725654885,
    "timestep_mean": 489.875,
    "epsilon_norm": 7.755948066711426,
    "prediction_norm": 6.877328395843506,
    "timestamp": "2025-07-14T12:35:03.135679"
  },
  {
    "epoch": 2482,
    "batch_idx": 160,
    "loss": 0.2753922641277313,
    "learning_rate": 0.00013659276725654885,
    "timestep_mean": 439.78125,
    "epsilon_norm": 7.034081935882568,
    "prediction_norm": 5.677992820739746,
    "timestamp": "2025-07-14T12:35:03.140849"
  },
  {
    "epoch": 2482,
    "batch_idx": 192,
    "loss": 0.2974547743797302,
    "learning_rate": 0.00013659276725654885,
    "timestep_mean": 411.0,
    "epsilon_norm": 3.68444561958313,
    "prediction_norm": 3.042238712310791,
    "timestamp": "2025-07-14T12:35:03.146008"
  },
  {
    "epoch": 2483,
    "batch_idx": 0,
    "loss": 0.47456634044647217,
    "learning_rate": 0.0001371326534119598,
    "timestep_mean": 525.375,
    "epsilon_norm": 8.888434410095215,
    "prediction_norm": 6.5860748291015625,
    "timestamp": "2025-07-14T12:35:03.151194"
  },
  {
    "epoch": 2483,
    "batch_idx": 32,
    "loss": 0.46136337518692017,
    "learning_rate": 0.0001371326534119598,
    "timestep_mean": 464.875,
    "epsilon_norm": 8.136119842529297,
    "prediction_norm": 6.708712100982666,
    "timestamp": "2025-07-14T12:35:03.156370"
  },
  {
    "epoch": 2483,
    "batch_idx": 64,
    "loss": 0.32835984230041504,
    "learning_rate": 0.0001371326534119598,
    "timestep_mean": 482.53125,
    "epsilon_norm": 8.263559341430664,
    "prediction_norm": 6.328837871551514,
    "timestamp": "2025-07-14T12:35:03.161564"
  },
  {
    "epoch": 2483,
    "batch_idx": 96,
    "loss": 0.42875152826309204,
    "learning_rate": 0.0001371326534119598,
    "timestep_mean": 431.9375,
    "epsilon_norm": 9.63547134399414,
    "prediction_norm": 7.736879348754883,
    "timestamp": "2025-07-14T12:35:03.166612"
  },
  {
    "epoch": 2483,
    "batch_idx": 128,
    "loss": 0.3381727635860443,
    "learning_rate": 0.0001371326534119598,
    "timestep_mean": 331.875,
    "epsilon_norm": 8.044754981994629,
    "prediction_norm": 6.210080146789551,
    "timestamp": "2025-07-14T12:35:03.171849"
  },
  {
    "epoch": 2483,
    "batch_idx": 160,
    "loss": 0.20531880855560303,
    "learning_rate": 0.0001371326534119598,
    "timestep_mean": 449.3125,
    "epsilon_norm": 7.189584255218506,
    "prediction_norm": 5.854247093200684,
    "timestamp": "2025-07-14T12:35:03.176968"
  },
  {
    "epoch": 2483,
    "batch_idx": 192,
    "loss": 0.15244504809379578,
    "learning_rate": 0.0001371326534119598,
    "timestep_mean": 540.5,
    "epsilon_norm": 4.9821062088012695,
    "prediction_norm": 4.361471176147461,
    "timestamp": "2025-07-14T12:35:03.181980"
  },
  {
    "epoch": 2484,
    "batch_idx": 0,
    "loss": 0.21585214138031006,
    "learning_rate": 0.00013767343490647706,
    "timestep_mean": 531.34375,
    "epsilon_norm": 7.594202518463135,
    "prediction_norm": 7.284721851348877,
    "timestamp": "2025-07-14T12:35:03.187182"
  },
  {
    "epoch": 2484,
    "batch_idx": 32,
    "loss": 0.30521342158317566,
    "learning_rate": 0.00013767343490647706,
    "timestep_mean": 538.0625,
    "epsilon_norm": 7.342010498046875,
    "prediction_norm": 6.308984756469727,
    "timestamp": "2025-07-14T12:35:03.192451"
  },
  {
    "epoch": 2484,
    "batch_idx": 64,
    "loss": 0.24103721976280212,
    "learning_rate": 0.00013767343490647706,
    "timestep_mean": 473.125,
    "epsilon_norm": 8.525586128234863,
    "prediction_norm": 7.988442897796631,
    "timestamp": "2025-07-14T12:35:03.197507"
  },
  {
    "epoch": 2484,
    "batch_idx": 96,
    "loss": 0.1568308174610138,
    "learning_rate": 0.00013767343490647706,
    "timestep_mean": 492.8125,
    "epsilon_norm": 6.815915584564209,
    "prediction_norm": 6.138838768005371,
    "timestamp": "2025-07-14T12:35:03.202598"
  },
  {
    "epoch": 2484,
    "batch_idx": 128,
    "loss": 0.354189932346344,
    "learning_rate": 0.00013767343490647706,
    "timestep_mean": 581.0,
    "epsilon_norm": 8.597336769104004,
    "prediction_norm": 7.705024242401123,
    "timestamp": "2025-07-14T12:35:03.207602"
  },
  {
    "epoch": 2484,
    "batch_idx": 160,
    "loss": 0.38476496934890747,
    "learning_rate": 0.00013767343490647706,
    "timestep_mean": 443.03125,
    "epsilon_norm": 7.837372303009033,
    "prediction_norm": 5.8005805015563965,
    "timestamp": "2025-07-14T12:35:03.212526"
  },
  {
    "epoch": 2484,
    "batch_idx": 192,
    "loss": 0.5780763030052185,
    "learning_rate": 0.00013767343490647706,
    "timestep_mean": 574.75,
    "epsilon_norm": 4.7537522315979,
    "prediction_norm": 3.4268360137939453,
    "timestamp": "2025-07-14T12:35:03.217411"
  },
  {
    "epoch": 2485,
    "batch_idx": 0,
    "loss": 0.3104483485221863,
    "learning_rate": 0.00013821511040577552,
    "timestep_mean": 492.21875,
    "epsilon_norm": 8.792015075683594,
    "prediction_norm": 7.689204216003418,
    "timestamp": "2025-07-14T12:35:03.222264"
  },
  {
    "epoch": 2485,
    "batch_idx": 32,
    "loss": 0.23306629061698914,
    "learning_rate": 0.00013821511040577552,
    "timestep_mean": 463.65625,
    "epsilon_norm": 8.684575080871582,
    "prediction_norm": 7.705559253692627,
    "timestamp": "2025-07-14T12:35:03.227107"
  },
  {
    "epoch": 2485,
    "batch_idx": 64,
    "loss": 0.4260860085487366,
    "learning_rate": 0.00013821511040577552,
    "timestep_mean": 460.71875,
    "epsilon_norm": 8.773367881774902,
    "prediction_norm": 7.124134063720703,
    "timestamp": "2025-07-14T12:35:03.232163"
  },
  {
    "epoch": 2485,
    "batch_idx": 96,
    "loss": 0.22069895267486572,
    "learning_rate": 0.00013821511040577552,
    "timestep_mean": 576.875,
    "epsilon_norm": 7.307972431182861,
    "prediction_norm": 6.852117538452148,
    "timestamp": "2025-07-14T12:35:03.237319"
  },
  {
    "epoch": 2485,
    "batch_idx": 128,
    "loss": 0.4088863730430603,
    "learning_rate": 0.00013821511040577552,
    "timestep_mean": 495.1875,
    "epsilon_norm": 8.407933235168457,
    "prediction_norm": 6.408193111419678,
    "timestamp": "2025-07-14T12:35:03.242428"
  },
  {
    "epoch": 2485,
    "batch_idx": 160,
    "loss": 0.2766481041908264,
    "learning_rate": 0.00013821511040577552,
    "timestep_mean": 432.625,
    "epsilon_norm": 5.75813627243042,
    "prediction_norm": 4.794893264770508,
    "timestamp": "2025-07-14T12:35:03.247551"
  },
  {
    "epoch": 2485,
    "batch_idx": 192,
    "loss": 0.1574118435382843,
    "learning_rate": 0.00013821511040577552,
    "timestep_mean": 739.875,
    "epsilon_norm": 4.247633934020996,
    "prediction_norm": 3.553699493408203,
    "timestamp": "2025-07-14T12:35:03.252583"
  },
  {
    "epoch": 2486,
    "batch_idx": 0,
    "loss": 0.16415847837924957,
    "learning_rate": 0.0001387576785733254,
    "timestep_mean": 590.25,
    "epsilon_norm": 7.139174461364746,
    "prediction_norm": 6.037951469421387,
    "timestamp": "2025-07-14T12:35:03.257690"
  },
  {
    "epoch": 2486,
    "batch_idx": 32,
    "loss": 0.18173755705356598,
    "learning_rate": 0.0001387576785733254,
    "timestep_mean": 600.0625,
    "epsilon_norm": 7.484649658203125,
    "prediction_norm": 7.042386054992676,
    "timestamp": "2025-07-14T12:35:03.262765"
  },
  {
    "epoch": 2486,
    "batch_idx": 64,
    "loss": 0.14529454708099365,
    "learning_rate": 0.0001387576785733254,
    "timestep_mean": 533.78125,
    "epsilon_norm": 7.35304594039917,
    "prediction_norm": 6.50050687789917,
    "timestamp": "2025-07-14T12:35:03.267938"
  },
  {
    "epoch": 2486,
    "batch_idx": 96,
    "loss": 0.11696209013462067,
    "learning_rate": 0.0001387576785733254,
    "timestep_mean": 595.03125,
    "epsilon_norm": 6.780542373657227,
    "prediction_norm": 6.397825241088867,
    "timestamp": "2025-07-14T12:35:03.273030"
  },
  {
    "epoch": 2486,
    "batch_idx": 128,
    "loss": 0.21633413434028625,
    "learning_rate": 0.0001387576785733254,
    "timestep_mean": 610.875,
    "epsilon_norm": 8.38533878326416,
    "prediction_norm": 7.422599792480469,
    "timestamp": "2025-07-14T12:35:03.278135"
  },
  {
    "epoch": 2486,
    "batch_idx": 160,
    "loss": 0.1603374034166336,
    "learning_rate": 0.0001387576785733254,
    "timestep_mean": 535.3125,
    "epsilon_norm": 8.08910846710205,
    "prediction_norm": 7.355768203735352,
    "timestamp": "2025-07-14T12:35:03.283303"
  },
  {
    "epoch": 2486,
    "batch_idx": 192,
    "loss": 0.09536126255989075,
    "learning_rate": 0.0001387576785733254,
    "timestep_mean": 630.25,
    "epsilon_norm": 2.7494454383850098,
    "prediction_norm": 2.734504461288452,
    "timestamp": "2025-07-14T12:35:03.288425"
  },
  {
    "epoch": 2487,
    "batch_idx": 0,
    "loss": 0.29008203744888306,
    "learning_rate": 0.00013930113807039335,
    "timestep_mean": 529.5,
    "epsilon_norm": 7.242708206176758,
    "prediction_norm": 6.353573799133301,
    "timestamp": "2025-07-14T12:35:03.293528"
  },
  {
    "epoch": 2487,
    "batch_idx": 32,
    "loss": 0.23945897817611694,
    "learning_rate": 0.00013930113807039335,
    "timestep_mean": 448.0,
    "epsilon_norm": 7.694332599639893,
    "prediction_norm": 6.686412334442139,
    "timestamp": "2025-07-14T12:35:03.298611"
  },
  {
    "epoch": 2487,
    "batch_idx": 64,
    "loss": 0.38751667737960815,
    "learning_rate": 0.00013930113807039335,
    "timestep_mean": 544.96875,
    "epsilon_norm": 8.825759887695312,
    "prediction_norm": 7.921204090118408,
    "timestamp": "2025-07-14T12:35:03.303752"
  },
  {
    "epoch": 2487,
    "batch_idx": 96,
    "loss": 0.27493172883987427,
    "learning_rate": 0.00013930113807039335,
    "timestep_mean": 464.1875,
    "epsilon_norm": 8.368866920471191,
    "prediction_norm": 7.203143119812012,
    "timestamp": "2025-07-14T12:35:03.308928"
  },
  {
    "epoch": 2487,
    "batch_idx": 128,
    "loss": 0.14941617846488953,
    "learning_rate": 0.00013930113807039335,
    "timestep_mean": 489.65625,
    "epsilon_norm": 7.876015663146973,
    "prediction_norm": 6.8991498947143555,
    "timestamp": "2025-07-14T12:35:03.314098"
  },
  {
    "epoch": 2487,
    "batch_idx": 160,
    "loss": 0.26582109928131104,
    "learning_rate": 0.00013930113807039335,
    "timestep_mean": 537.65625,
    "epsilon_norm": 8.068822860717773,
    "prediction_norm": 6.8230299949646,
    "timestamp": "2025-07-14T12:35:03.319328"
  },
  {
    "epoch": 2487,
    "batch_idx": 192,
    "loss": 0.25562748312950134,
    "learning_rate": 0.00013930113807039335,
    "timestep_mean": 418.625,
    "epsilon_norm": 3.361398220062256,
    "prediction_norm": 2.9076223373413086,
    "timestamp": "2025-07-14T12:35:03.324401"
  },
  {
    "epoch": 2488,
    "batch_idx": 0,
    "loss": 0.20407995581626892,
    "learning_rate": 0.00013984548755604676,
    "timestep_mean": 539.78125,
    "epsilon_norm": 8.6873197555542,
    "prediction_norm": 7.644851207733154,
    "timestamp": "2025-07-14T12:35:03.329461"
  },
  {
    "epoch": 2488,
    "batch_idx": 32,
    "loss": 0.37432926893234253,
    "learning_rate": 0.00013984548755604676,
    "timestep_mean": 523.96875,
    "epsilon_norm": 8.023653984069824,
    "prediction_norm": 6.853068828582764,
    "timestamp": "2025-07-14T12:35:03.334641"
  },
  {
    "epoch": 2488,
    "batch_idx": 64,
    "loss": 0.302086740732193,
    "learning_rate": 0.00013984548755604676,
    "timestep_mean": 522.46875,
    "epsilon_norm": 7.487246513366699,
    "prediction_norm": 5.8702592849731445,
    "timestamp": "2025-07-14T12:35:03.339774"
  },
  {
    "epoch": 2488,
    "batch_idx": 96,
    "loss": 0.2282802164554596,
    "learning_rate": 0.00013984548755604676,
    "timestep_mean": 546.59375,
    "epsilon_norm": 7.944234371185303,
    "prediction_norm": 6.560833930969238,
    "timestamp": "2025-07-14T12:35:03.344750"
  },
  {
    "epoch": 2488,
    "batch_idx": 128,
    "loss": 0.3815096616744995,
    "learning_rate": 0.00013984548755604676,
    "timestep_mean": 508.75,
    "epsilon_norm": 7.8231120109558105,
    "prediction_norm": 6.100518226623535,
    "timestamp": "2025-07-14T12:35:03.349882"
  },
  {
    "epoch": 2488,
    "batch_idx": 160,
    "loss": 0.14825603365898132,
    "learning_rate": 0.00013984548755604676,
    "timestep_mean": 500.625,
    "epsilon_norm": 8.030657768249512,
    "prediction_norm": 7.326493740081787,
    "timestamp": "2025-07-14T12:35:03.355147"
  },
  {
    "epoch": 2488,
    "batch_idx": 192,
    "loss": 0.07364095747470856,
    "learning_rate": 0.00013984548755604676,
    "timestep_mean": 511.5,
    "epsilon_norm": 4.588879585266113,
    "prediction_norm": 4.248805999755859,
    "timestamp": "2025-07-14T12:35:03.360176"
  },
  {
    "epoch": 2489,
    "batch_idx": 0,
    "loss": 0.3923007845878601,
    "learning_rate": 0.00014039072568715805,
    "timestep_mean": 472.21875,
    "epsilon_norm": 8.810840606689453,
    "prediction_norm": 7.0131049156188965,
    "timestamp": "2025-07-14T12:35:03.365361"
  },
  {
    "epoch": 2489,
    "batch_idx": 32,
    "loss": 0.21711736917495728,
    "learning_rate": 0.00014039072568715805,
    "timestep_mean": 541.375,
    "epsilon_norm": 10.097018241882324,
    "prediction_norm": 8.704668998718262,
    "timestamp": "2025-07-14T12:35:03.370592"
  },
  {
    "epoch": 2489,
    "batch_idx": 64,
    "loss": 0.25397631525993347,
    "learning_rate": 0.00014039072568715805,
    "timestep_mean": 507.65625,
    "epsilon_norm": 7.706237316131592,
    "prediction_norm": 6.591097354888916,
    "timestamp": "2025-07-14T12:35:03.375819"
  },
  {
    "epoch": 2489,
    "batch_idx": 96,
    "loss": 0.32078754901885986,
    "learning_rate": 0.00014039072568715805,
    "timestep_mean": 472.78125,
    "epsilon_norm": 8.381884574890137,
    "prediction_norm": 6.535946369171143,
    "timestamp": "2025-07-14T12:35:03.380954"
  },
  {
    "epoch": 2489,
    "batch_idx": 128,
    "loss": 0.28533077239990234,
    "learning_rate": 0.00014039072568715805,
    "timestep_mean": 525.5,
    "epsilon_norm": 7.830981254577637,
    "prediction_norm": 6.3718461990356445,
    "timestamp": "2025-07-14T12:35:03.386058"
  },
  {
    "epoch": 2489,
    "batch_idx": 160,
    "loss": 0.28453078866004944,
    "learning_rate": 0.00014039072568715805,
    "timestep_mean": 540.0625,
    "epsilon_norm": 7.541029930114746,
    "prediction_norm": 6.17069673538208,
    "timestamp": "2025-07-14T12:35:03.391140"
  },
  {
    "epoch": 2489,
    "batch_idx": 192,
    "loss": 0.11776641756296158,
    "learning_rate": 0.00014039072568715805,
    "timestep_mean": 529.625,
    "epsilon_norm": 4.7798542976379395,
    "prediction_norm": 4.733554363250732,
    "timestamp": "2025-07-14T12:35:03.396307"
  },
  {
    "epoch": 2490,
    "batch_idx": 0,
    "loss": 0.27142226696014404,
    "learning_rate": 0.00014093685111840577,
    "timestep_mean": 509.71875,
    "epsilon_norm": 7.53572940826416,
    "prediction_norm": 7.122150421142578,
    "timestamp": "2025-07-14T12:35:03.401434"
  },
  {
    "epoch": 2490,
    "batch_idx": 32,
    "loss": 0.2949046790599823,
    "learning_rate": 0.00014093685111840577,
    "timestep_mean": 614.25,
    "epsilon_norm": 9.127152442932129,
    "prediction_norm": 7.922408103942871,
    "timestamp": "2025-07-14T12:35:03.406452"
  },
  {
    "epoch": 2490,
    "batch_idx": 64,
    "loss": 0.21542976796627045,
    "learning_rate": 0.00014093685111840577,
    "timestep_mean": 485.40625,
    "epsilon_norm": 7.151165962219238,
    "prediction_norm": 6.993461608886719,
    "timestamp": "2025-07-14T12:35:03.411286"
  },
  {
    "epoch": 2490,
    "batch_idx": 96,
    "loss": 0.21273818612098694,
    "learning_rate": 0.00014093685111840577,
    "timestep_mean": 354.8125,
    "epsilon_norm": 7.138125896453857,
    "prediction_norm": 6.926883697509766,
    "timestamp": "2025-07-14T12:35:03.416449"
  },
  {
    "epoch": 2490,
    "batch_idx": 128,
    "loss": 0.3985888361930847,
    "learning_rate": 0.00014093685111840577,
    "timestep_mean": 473.5,
    "epsilon_norm": 7.621734619140625,
    "prediction_norm": 6.2324347496032715,
    "timestamp": "2025-07-14T12:35:03.422613"
  },
  {
    "epoch": 2490,
    "batch_idx": 160,
    "loss": 0.23086079955101013,
    "learning_rate": 0.00014093685111840577,
    "timestep_mean": 516.0625,
    "epsilon_norm": 8.570892333984375,
    "prediction_norm": 7.400290012359619,
    "timestamp": "2025-07-14T12:35:03.427974"
  },
  {
    "epoch": 2490,
    "batch_idx": 192,
    "loss": 0.6229596138000488,
    "learning_rate": 0.00014093685111840577,
    "timestep_mean": 499.0,
    "epsilon_norm": 4.411616802215576,
    "prediction_norm": 3.1720099449157715,
    "timestamp": "2025-07-14T12:35:03.433195"
  },
  {
    "epoch": 2491,
    "batch_idx": 0,
    "loss": 0.44626861810684204,
    "learning_rate": 0.0001414838625022802,
    "timestep_mean": 406.84375,
    "epsilon_norm": 8.123222351074219,
    "prediction_norm": 6.420974254608154,
    "timestamp": "2025-07-14T12:35:03.438467"
  },
  {
    "epoch": 2491,
    "batch_idx": 32,
    "loss": 0.2810367941856384,
    "learning_rate": 0.0001414838625022802,
    "timestep_mean": 465.9375,
    "epsilon_norm": 8.508156776428223,
    "prediction_norm": 7.091092586517334,
    "timestamp": "2025-07-14T12:35:03.443622"
  },
  {
    "epoch": 2491,
    "batch_idx": 64,
    "loss": 0.2268475890159607,
    "learning_rate": 0.0001414838625022802,
    "timestep_mean": 445.4375,
    "epsilon_norm": 6.683987140655518,
    "prediction_norm": 5.844172477722168,
    "timestamp": "2025-07-14T12:35:03.448771"
  },
  {
    "epoch": 2491,
    "batch_idx": 96,
    "loss": 0.2904268503189087,
    "learning_rate": 0.0001414838625022802,
    "timestep_mean": 519.5,
    "epsilon_norm": 8.296141624450684,
    "prediction_norm": 7.384227752685547,
    "timestamp": "2025-07-14T12:35:03.453891"
  },
  {
    "epoch": 2491,
    "batch_idx": 128,
    "loss": 0.13750113546848297,
    "learning_rate": 0.0001414838625022802,
    "timestep_mean": 624.03125,
    "epsilon_norm": 9.325722694396973,
    "prediction_norm": 8.952949523925781,
    "timestamp": "2025-07-14T12:35:03.459127"
  },
  {
    "epoch": 2491,
    "batch_idx": 160,
    "loss": 0.33248621225357056,
    "learning_rate": 0.0001414838625022802,
    "timestep_mean": 504.625,
    "epsilon_norm": 8.399617195129395,
    "prediction_norm": 7.144741535186768,
    "timestamp": "2025-07-14T12:35:03.464084"
  },
  {
    "epoch": 2491,
    "batch_idx": 192,
    "loss": 0.13036996126174927,
    "learning_rate": 0.0001414838625022802,
    "timestep_mean": 429.75,
    "epsilon_norm": 4.180614948272705,
    "prediction_norm": 3.671684980392456,
    "timestamp": "2025-07-14T12:35:03.469241"
  },
  {
    "epoch": 2492,
    "batch_idx": 0,
    "loss": 0.5046107769012451,
    "learning_rate": 0.00014203175848908486,
    "timestep_mean": 471.1875,
    "epsilon_norm": 8.273518562316895,
    "prediction_norm": 6.421188831329346,
    "timestamp": "2025-07-14T12:35:03.474566"
  },
  {
    "epoch": 2492,
    "batch_idx": 32,
    "loss": 0.5245593786239624,
    "learning_rate": 0.00014203175848908486,
    "timestep_mean": 439.4375,
    "epsilon_norm": 8.201804161071777,
    "prediction_norm": 6.371432781219482,
    "timestamp": "2025-07-14T12:35:03.480063"
  },
  {
    "epoch": 2492,
    "batch_idx": 64,
    "loss": 0.23245030641555786,
    "learning_rate": 0.00014203175848908486,
    "timestep_mean": 500.25,
    "epsilon_norm": 7.524296760559082,
    "prediction_norm": 6.772346496582031,
    "timestamp": "2025-07-14T12:35:03.485143"
  },
  {
    "epoch": 2492,
    "batch_idx": 96,
    "loss": 0.21255949139595032,
    "learning_rate": 0.00014203175848908486,
    "timestep_mean": 492.59375,
    "epsilon_norm": 7.036540508270264,
    "prediction_norm": 6.0598931312561035,
    "timestamp": "2025-07-14T12:35:03.490475"
  },
  {
    "epoch": 2492,
    "batch_idx": 128,
    "loss": 0.293917179107666,
    "learning_rate": 0.00014203175848908486,
    "timestep_mean": 497.28125,
    "epsilon_norm": 7.679465293884277,
    "prediction_norm": 5.509167671203613,
    "timestamp": "2025-07-14T12:35:03.495637"
  },
  {
    "epoch": 2492,
    "batch_idx": 160,
    "loss": 0.33928802609443665,
    "learning_rate": 0.00014203175848908486,
    "timestep_mean": 540.46875,
    "epsilon_norm": 8.050829887390137,
    "prediction_norm": 6.5443644523620605,
    "timestamp": "2025-07-14T12:35:03.500930"
  },
  {
    "epoch": 2492,
    "batch_idx": 192,
    "loss": 0.08175274729728699,
    "learning_rate": 0.00014203175848908486,
    "timestep_mean": 696.125,
    "epsilon_norm": 4.175130844116211,
    "prediction_norm": 3.9485373497009277,
    "timestamp": "2025-07-14T12:35:03.506125"
  },
  {
    "epoch": 2493,
    "batch_idx": 0,
    "loss": 0.17419975996017456,
    "learning_rate": 0.00014258053772694054,
    "timestep_mean": 541.28125,
    "epsilon_norm": 9.633528709411621,
    "prediction_norm": 8.463903427124023,
    "timestamp": "2025-07-14T12:35:03.511343"
  },
  {
    "epoch": 2493,
    "batch_idx": 32,
    "loss": 0.2915361523628235,
    "learning_rate": 0.00014258053772694054,
    "timestep_mean": 495.0625,
    "epsilon_norm": 7.3386640548706055,
    "prediction_norm": 5.537668228149414,
    "timestamp": "2025-07-14T12:35:03.516487"
  },
  {
    "epoch": 2493,
    "batch_idx": 64,
    "loss": 0.17154857516288757,
    "learning_rate": 0.00014258053772694054,
    "timestep_mean": 401.28125,
    "epsilon_norm": 6.966693878173828,
    "prediction_norm": 6.278832912445068,
    "timestamp": "2025-07-14T12:35:03.521880"
  },
  {
    "epoch": 2493,
    "batch_idx": 96,
    "loss": 0.26465263962745667,
    "learning_rate": 0.00014258053772694054,
    "timestep_mean": 482.09375,
    "epsilon_norm": 8.225379943847656,
    "prediction_norm": 6.963505744934082,
    "timestamp": "2025-07-14T12:35:03.527118"
  },
  {
    "epoch": 2493,
    "batch_idx": 128,
    "loss": 0.2882494330406189,
    "learning_rate": 0.00014258053772694054,
    "timestep_mean": 523.84375,
    "epsilon_norm": 8.410022735595703,
    "prediction_norm": 7.109362602233887,
    "timestamp": "2025-07-14T12:35:03.532337"
  },
  {
    "epoch": 2493,
    "batch_idx": 160,
    "loss": 0.1684795767068863,
    "learning_rate": 0.00014258053772694054,
    "timestep_mean": 617.75,
    "epsilon_norm": 8.857270240783691,
    "prediction_norm": 8.290929794311523,
    "timestamp": "2025-07-14T12:35:03.537428"
  },
  {
    "epoch": 2493,
    "batch_idx": 192,
    "loss": 0.2892446517944336,
    "learning_rate": 0.00014258053772694054,
    "timestep_mean": 374.875,
    "epsilon_norm": 4.425792694091797,
    "prediction_norm": 3.5669126510620117,
    "timestamp": "2025-07-14T12:35:03.542692"
  },
  {
    "epoch": 2494,
    "batch_idx": 0,
    "loss": 0.2733086049556732,
    "learning_rate": 0.0001431301988617897,
    "timestep_mean": 518.875,
    "epsilon_norm": 8.778904914855957,
    "prediction_norm": 7.508351802825928,
    "timestamp": "2025-07-14T12:35:03.547879"
  },
  {
    "epoch": 2494,
    "batch_idx": 32,
    "loss": 0.2833247780799866,
    "learning_rate": 0.0001431301988617897,
    "timestep_mean": 460.75,
    "epsilon_norm": 8.080291748046875,
    "prediction_norm": 6.91098690032959,
    "timestamp": "2025-07-14T12:35:03.553148"
  },
  {
    "epoch": 2494,
    "batch_idx": 64,
    "loss": 0.16103339195251465,
    "learning_rate": 0.0001431301988617897,
    "timestep_mean": 473.125,
    "epsilon_norm": 7.970174312591553,
    "prediction_norm": 7.300382137298584,
    "timestamp": "2025-07-14T12:35:03.558136"
  },
  {
    "epoch": 2494,
    "batch_idx": 96,
    "loss": 0.17798030376434326,
    "learning_rate": 0.0001431301988617897,
    "timestep_mean": 528.15625,
    "epsilon_norm": 8.308112144470215,
    "prediction_norm": 7.264129161834717,
    "timestamp": "2025-07-14T12:35:03.563062"
  },
  {
    "epoch": 2494,
    "batch_idx": 128,
    "loss": 0.3319351375102997,
    "learning_rate": 0.0001431301988617897,
    "timestep_mean": 492.78125,
    "epsilon_norm": 8.51191234588623,
    "prediction_norm": 7.014512062072754,
    "timestamp": "2025-07-14T12:35:03.568033"
  },
  {
    "epoch": 2494,
    "batch_idx": 160,
    "loss": 0.25545984506607056,
    "learning_rate": 0.0001431301988617897,
    "timestep_mean": 429.75,
    "epsilon_norm": 7.243889331817627,
    "prediction_norm": 6.830824851989746,
    "timestamp": "2025-07-14T12:35:03.572883"
  },
  {
    "epoch": 2494,
    "batch_idx": 192,
    "loss": 0.08666640520095825,
    "learning_rate": 0.0001431301988617897,
    "timestep_mean": 488.625,
    "epsilon_norm": 3.96148419380188,
    "prediction_norm": 3.3511459827423096,
    "timestamp": "2025-07-14T12:35:03.577871"
  },
  {
    "epoch": 2495,
    "batch_idx": 0,
    "loss": 0.1601230651140213,
    "learning_rate": 0.00014368074053739763,
    "timestep_mean": 514.9375,
    "epsilon_norm": 7.0422163009643555,
    "prediction_norm": 6.110082626342773,
    "timestamp": "2025-07-14T12:35:03.582769"
  },
  {
    "epoch": 2495,
    "batch_idx": 32,
    "loss": 0.15999914705753326,
    "learning_rate": 0.00014368074053739763,
    "timestep_mean": 588.65625,
    "epsilon_norm": 7.793012619018555,
    "prediction_norm": 6.588131427764893,
    "timestamp": "2025-07-14T12:35:03.587652"
  },
  {
    "epoch": 2495,
    "batch_idx": 64,
    "loss": 0.3150564432144165,
    "learning_rate": 0.00014368074053739763,
    "timestep_mean": 516.0,
    "epsilon_norm": 8.636029243469238,
    "prediction_norm": 7.0256524085998535,
    "timestamp": "2025-07-14T12:35:03.592743"
  },
  {
    "epoch": 2495,
    "batch_idx": 96,
    "loss": 0.2615763247013092,
    "learning_rate": 0.00014368074053739763,
    "timestep_mean": 496.375,
    "epsilon_norm": 8.160380363464355,
    "prediction_norm": 6.7083635330200195,
    "timestamp": "2025-07-14T12:35:03.597808"
  },
  {
    "epoch": 2495,
    "batch_idx": 128,
    "loss": 0.3584456443786621,
    "learning_rate": 0.00014368074053739763,
    "timestep_mean": 423.25,
    "epsilon_norm": 8.337554931640625,
    "prediction_norm": 6.514702796936035,
    "timestamp": "2025-07-14T12:35:03.602969"
  },
  {
    "epoch": 2495,
    "batch_idx": 160,
    "loss": 0.3145826458930969,
    "learning_rate": 0.00014368074053739763,
    "timestep_mean": 533.75,
    "epsilon_norm": 8.395081520080566,
    "prediction_norm": 7.227078914642334,
    "timestamp": "2025-07-14T12:35:03.608059"
  },
  {
    "epoch": 2495,
    "batch_idx": 192,
    "loss": 0.30706238746643066,
    "learning_rate": 0.00014368074053739763,
    "timestep_mean": 589.875,
    "epsilon_norm": 4.0157790184021,
    "prediction_norm": 3.0820298194885254,
    "timestamp": "2025-07-14T12:35:03.613248"
  },
  {
    "epoch": 2496,
    "batch_idx": 0,
    "loss": 0.36747533082962036,
    "learning_rate": 0.00014423216139535766,
    "timestep_mean": 478.25,
    "epsilon_norm": 9.078516006469727,
    "prediction_norm": 7.803555488586426,
    "timestamp": "2025-07-14T12:35:03.618434"
  },
  {
    "epoch": 2496,
    "batch_idx": 32,
    "loss": 0.23857006430625916,
    "learning_rate": 0.00014423216139535766,
    "timestep_mean": 540.71875,
    "epsilon_norm": 7.768341064453125,
    "prediction_norm": 6.353668689727783,
    "timestamp": "2025-07-14T12:35:03.623486"
  },
  {
    "epoch": 2496,
    "batch_idx": 64,
    "loss": 0.19706851243972778,
    "learning_rate": 0.00014423216139535766,
    "timestep_mean": 522.40625,
    "epsilon_norm": 8.07772159576416,
    "prediction_norm": 7.373962879180908,
    "timestamp": "2025-07-14T12:35:03.628539"
  },
  {
    "epoch": 2496,
    "batch_idx": 96,
    "loss": 0.4106443226337433,
    "learning_rate": 0.00014423216139535766,
    "timestep_mean": 407.65625,
    "epsilon_norm": 8.48241901397705,
    "prediction_norm": 7.147467136383057,
    "timestamp": "2025-07-14T12:35:03.633822"
  },
  {
    "epoch": 2496,
    "batch_idx": 128,
    "loss": 0.31698375940322876,
    "learning_rate": 0.00014423216139535766,
    "timestep_mean": 542.15625,
    "epsilon_norm": 8.737563133239746,
    "prediction_norm": 7.447155475616455,
    "timestamp": "2025-07-14T12:35:03.638913"
  },
  {
    "epoch": 2496,
    "batch_idx": 160,
    "loss": 0.2239939570426941,
    "learning_rate": 0.00014423216139535766,
    "timestep_mean": 502.0625,
    "epsilon_norm": 7.566422939300537,
    "prediction_norm": 6.300479888916016,
    "timestamp": "2025-07-14T12:35:03.644135"
  },
  {
    "epoch": 2496,
    "batch_idx": 192,
    "loss": 0.08518686890602112,
    "learning_rate": 0.00014423216139535766,
    "timestep_mean": 703.375,
    "epsilon_norm": 3.444385528564453,
    "prediction_norm": 3.4923458099365234,
    "timestamp": "2025-07-14T12:35:03.649309"
  },
  {
    "epoch": 2497,
    "batch_idx": 0,
    "loss": 0.5519790053367615,
    "learning_rate": 0.00014478446007509354,
    "timestep_mean": 495.84375,
    "epsilon_norm": 7.885746955871582,
    "prediction_norm": 5.3153910636901855,
    "timestamp": "2025-07-14T12:35:03.654439"
  },
  {
    "epoch": 2497,
    "batch_idx": 32,
    "loss": 0.4026184380054474,
    "learning_rate": 0.00014478446007509354,
    "timestep_mean": 457.59375,
    "epsilon_norm": 8.685914993286133,
    "prediction_norm": 7.3320841789245605,
    "timestamp": "2025-07-14T12:35:03.659616"
  },
  {
    "epoch": 2497,
    "batch_idx": 64,
    "loss": 0.13002774119377136,
    "learning_rate": 0.00014478446007509354,
    "timestep_mean": 602.21875,
    "epsilon_norm": 8.552839279174805,
    "prediction_norm": 8.255414009094238,
    "timestamp": "2025-07-14T12:35:03.664677"
  },
  {
    "epoch": 2497,
    "batch_idx": 96,
    "loss": 0.2861374318599701,
    "learning_rate": 0.00014478446007509354,
    "timestep_mean": 577.75,
    "epsilon_norm": 9.575898170471191,
    "prediction_norm": 8.359771728515625,
    "timestamp": "2025-07-14T12:35:03.669875"
  },
  {
    "epoch": 2497,
    "batch_idx": 128,
    "loss": 0.2553732097148895,
    "learning_rate": 0.00014478446007509354,
    "timestep_mean": 543.65625,
    "epsilon_norm": 8.720442771911621,
    "prediction_norm": 7.556944847106934,
    "timestamp": "2025-07-14T12:35:03.674982"
  },
  {
    "epoch": 2497,
    "batch_idx": 160,
    "loss": 0.2800235152244568,
    "learning_rate": 0.00014478446007509354,
    "timestep_mean": 461.0,
    "epsilon_norm": 8.963597297668457,
    "prediction_norm": 7.641712188720703,
    "timestamp": "2025-07-14T12:35:03.680339"
  },
  {
    "epoch": 2497,
    "batch_idx": 192,
    "loss": 0.13331590592861176,
    "learning_rate": 0.00014478446007509354,
    "timestep_mean": 360.0,
    "epsilon_norm": 2.6849629878997803,
    "prediction_norm": 2.3860621452331543,
    "timestamp": "2025-07-14T12:35:03.685499"
  },
  {
    "epoch": 2498,
    "batch_idx": 0,
    "loss": 0.15399564802646637,
    "learning_rate": 0.00014533763521386344,
    "timestep_mean": 532.875,
    "epsilon_norm": 8.829889297485352,
    "prediction_norm": 8.591224670410156,
    "timestamp": "2025-07-14T12:35:03.690722"
  },
  {
    "epoch": 2498,
    "batch_idx": 32,
    "loss": 0.18800124526023865,
    "learning_rate": 0.00014533763521386344,
    "timestep_mean": 479.90625,
    "epsilon_norm": 6.97492790222168,
    "prediction_norm": 6.34386682510376,
    "timestamp": "2025-07-14T12:35:03.695923"
  },
  {
    "epoch": 2498,
    "batch_idx": 64,
    "loss": 0.2500699460506439,
    "learning_rate": 0.00014533763521386344,
    "timestep_mean": 520.9375,
    "epsilon_norm": 7.222539901733398,
    "prediction_norm": 6.296646595001221,
    "timestamp": "2025-07-14T12:35:03.701142"
  },
  {
    "epoch": 2498,
    "batch_idx": 96,
    "loss": 0.28840717673301697,
    "learning_rate": 0.00014533763521386344,
    "timestep_mean": 408.03125,
    "epsilon_norm": 8.336517333984375,
    "prediction_norm": 7.287313461303711,
    "timestamp": "2025-07-14T12:35:03.706373"
  },
  {
    "epoch": 2498,
    "batch_idx": 128,
    "loss": 0.28339898586273193,
    "learning_rate": 0.00014533763521386344,
    "timestep_mean": 513.53125,
    "epsilon_norm": 7.333807468414307,
    "prediction_norm": 6.137838363647461,
    "timestamp": "2025-07-14T12:35:03.711341"
  },
  {
    "epoch": 2498,
    "batch_idx": 160,
    "loss": 0.2791751027107239,
    "learning_rate": 0.00014533763521386344,
    "timestep_mean": 468.53125,
    "epsilon_norm": 8.1774320602417,
    "prediction_norm": 7.0902886390686035,
    "timestamp": "2025-07-14T12:35:03.716350"
  },
  {
    "epoch": 2498,
    "batch_idx": 192,
    "loss": 0.2102772295475006,
    "learning_rate": 0.00014533763521386344,
    "timestep_mean": 597.5,
    "epsilon_norm": 4.589688301086426,
    "prediction_norm": 4.284668445587158,
    "timestamp": "2025-07-14T12:35:03.721439"
  },
  {
    "epoch": 2499,
    "batch_idx": 0,
    "loss": 0.22910906374454498,
    "learning_rate": 0.00014589168544676238,
    "timestep_mean": 426.5,
    "epsilon_norm": 8.36862850189209,
    "prediction_norm": 7.4888691902160645,
    "timestamp": "2025-07-14T12:35:03.726575"
  },
  {
    "epoch": 2499,
    "batch_idx": 32,
    "loss": 0.24116770923137665,
    "learning_rate": 0.00014589168544676238,
    "timestep_mean": 511.78125,
    "epsilon_norm": 8.649043083190918,
    "prediction_norm": 8.494921684265137,
    "timestamp": "2025-07-14T12:35:03.731696"
  },
  {
    "epoch": 2499,
    "batch_idx": 64,
    "loss": 0.15436725318431854,
    "learning_rate": 0.00014589168544676238,
    "timestep_mean": 549.03125,
    "epsilon_norm": 7.493612289428711,
    "prediction_norm": 7.463479518890381,
    "timestamp": "2025-07-14T12:35:03.736939"
  },
  {
    "epoch": 2499,
    "batch_idx": 96,
    "loss": 0.32802948355674744,
    "learning_rate": 0.00014589168544676238,
    "timestep_mean": 514.125,
    "epsilon_norm": 8.090195655822754,
    "prediction_norm": 7.038028240203857,
    "timestamp": "2025-07-14T12:35:03.742257"
  },
  {
    "epoch": 2499,
    "batch_idx": 128,
    "loss": 0.29403749108314514,
    "learning_rate": 0.00014589168544676238,
    "timestep_mean": 533.3125,
    "epsilon_norm": 7.590181350708008,
    "prediction_norm": 6.659061431884766,
    "timestamp": "2025-07-14T12:35:03.747414"
  },
  {
    "epoch": 2499,
    "batch_idx": 160,
    "loss": 0.3367689251899719,
    "learning_rate": 0.00014589168544676238,
    "timestep_mean": 525.9375,
    "epsilon_norm": 6.754665851593018,
    "prediction_norm": 5.834183692932129,
    "timestamp": "2025-07-14T12:35:03.752547"
  },
  {
    "epoch": 2499,
    "batch_idx": 192,
    "loss": 0.25048425793647766,
    "learning_rate": 0.00014589168544676238,
    "timestep_mean": 538.0,
    "epsilon_norm": 3.2208635807037354,
    "prediction_norm": 2.8519344329833984,
    "timestamp": "2025-07-14T12:35:03.757720"
  },
  {
    "epoch": 2500,
    "batch_idx": 0,
    "loss": 0.2461269646883011,
    "learning_rate": 0.00014644660940672655,
    "timestep_mean": 557.96875,
    "epsilon_norm": 7.900681972503662,
    "prediction_norm": 7.6965556144714355,
    "timestamp": "2025-07-14T12:35:03.762882"
  },
  {
    "epoch": 2500,
    "batch_idx": 32,
    "loss": 0.22254331409931183,
    "learning_rate": 0.00014644660940672655,
    "timestep_mean": 482.53125,
    "epsilon_norm": 7.600135803222656,
    "prediction_norm": 6.957341194152832,
    "timestamp": "2025-07-14T12:35:03.768149"
  },
  {
    "epoch": 2500,
    "batch_idx": 64,
    "loss": 0.23286956548690796,
    "learning_rate": 0.00014644660940672655,
    "timestep_mean": 453.8125,
    "epsilon_norm": 8.49081802368164,
    "prediction_norm": 7.5115461349487305,
    "timestamp": "2025-07-14T12:35:03.773329"
  },
  {
    "epoch": 2500,
    "batch_idx": 96,
    "loss": 0.21913835406303406,
    "learning_rate": 0.00014644660940672655,
    "timestep_mean": 451.5625,
    "epsilon_norm": 7.6619181632995605,
    "prediction_norm": 6.812696933746338,
    "timestamp": "2025-07-14T12:35:03.778507"
  },
  {
    "epoch": 2500,
    "batch_idx": 128,
    "loss": 0.2880181670188904,
    "learning_rate": 0.00014644660940672655,
    "timestep_mean": 414.9375,
    "epsilon_norm": 8.462583541870117,
    "prediction_norm": 7.562308311462402,
    "timestamp": "2025-07-14T12:35:03.783874"
  },
  {
    "epoch": 2500,
    "batch_idx": 160,
    "loss": 0.4912565052509308,
    "learning_rate": 0.00014644660940672655,
    "timestep_mean": 535.875,
    "epsilon_norm": 7.381770610809326,
    "prediction_norm": 5.417759418487549,
    "timestamp": "2025-07-14T12:35:03.789189"
  },
  {
    "epoch": 2500,
    "batch_idx": 192,
    "loss": 0.1675070822238922,
    "learning_rate": 0.00014644660940672655,
    "timestep_mean": 449.125,
    "epsilon_norm": 4.420810222625732,
    "prediction_norm": 3.8530960083007812,
    "timestamp": "2025-07-14T12:35:03.794262"
  },
  {
    "epoch": 2501,
    "batch_idx": 0,
    "loss": 0.22945620119571686,
    "learning_rate": 0.0001470024057245364,
    "timestep_mean": 491.5625,
    "epsilon_norm": 8.850382804870605,
    "prediction_norm": 8.213741302490234,
    "timestamp": "2025-07-14T12:35:03.799446"
  },
  {
    "epoch": 2501,
    "batch_idx": 32,
    "loss": 0.17535530030727386,
    "learning_rate": 0.0001470024057245364,
    "timestep_mean": 570.65625,
    "epsilon_norm": 7.264306545257568,
    "prediction_norm": 6.733546733856201,
    "timestamp": "2025-07-14T12:35:03.804618"
  },
  {
    "epoch": 2501,
    "batch_idx": 64,
    "loss": 0.2638823390007019,
    "learning_rate": 0.0001470024057245364,
    "timestep_mean": 490.90625,
    "epsilon_norm": 6.593379974365234,
    "prediction_norm": 5.121376037597656,
    "timestamp": "2025-07-14T12:35:03.809717"
  },
  {
    "epoch": 2501,
    "batch_idx": 96,
    "loss": 0.25896191596984863,
    "learning_rate": 0.0001470024057245364,
    "timestep_mean": 508.46875,
    "epsilon_norm": 8.46417236328125,
    "prediction_norm": 6.972978591918945,
    "timestamp": "2025-07-14T12:35:03.814713"
  },
  {
    "epoch": 2501,
    "batch_idx": 128,
    "loss": 0.6324543952941895,
    "learning_rate": 0.0001470024057245364,
    "timestep_mean": 500.8125,
    "epsilon_norm": 9.76114273071289,
    "prediction_norm": 7.644247055053711,
    "timestamp": "2025-07-14T12:35:03.819690"
  },
  {
    "epoch": 2501,
    "batch_idx": 160,
    "loss": 0.35966652631759644,
    "learning_rate": 0.0001470024057245364,
    "timestep_mean": 465.46875,
    "epsilon_norm": 9.594791412353516,
    "prediction_norm": 7.784883975982666,
    "timestamp": "2025-07-14T12:35:03.824541"
  },
  {
    "epoch": 2501,
    "batch_idx": 192,
    "loss": 0.20072264969348907,
    "learning_rate": 0.0001470024057245364,
    "timestep_mean": 559.375,
    "epsilon_norm": 4.181789398193359,
    "prediction_norm": 3.833665132522583,
    "timestamp": "2025-07-14T12:35:03.829459"
  },
  {
    "epoch": 2502,
    "batch_idx": 0,
    "loss": 0.38855820894241333,
    "learning_rate": 0.00014755907302881958,
    "timestep_mean": 503.0625,
    "epsilon_norm": 8.875051498413086,
    "prediction_norm": 7.52620792388916,
    "timestamp": "2025-07-14T12:35:03.834559"
  },
  {
    "epoch": 2502,
    "batch_idx": 32,
    "loss": 0.22159725427627563,
    "learning_rate": 0.00014755907302881958,
    "timestep_mean": 524.34375,
    "epsilon_norm": 7.564513206481934,
    "prediction_norm": 6.587836265563965,
    "timestamp": "2025-07-14T12:35:03.839712"
  },
  {
    "epoch": 2502,
    "batch_idx": 64,
    "loss": 0.4775802493095398,
    "learning_rate": 0.00014755907302881958,
    "timestep_mean": 423.3125,
    "epsilon_norm": 8.02832317352295,
    "prediction_norm": 5.043471336364746,
    "timestamp": "2025-07-14T12:35:03.844840"
  },
  {
    "epoch": 2502,
    "batch_idx": 96,
    "loss": 0.16312871873378754,
    "learning_rate": 0.00014755907302881958,
    "timestep_mean": 491.1875,
    "epsilon_norm": 6.7340240478515625,
    "prediction_norm": 5.483918190002441,
    "timestamp": "2025-07-14T12:35:03.849932"
  },
  {
    "epoch": 2502,
    "batch_idx": 128,
    "loss": 0.1586132049560547,
    "learning_rate": 0.00014755907302881958,
    "timestep_mean": 594.96875,
    "epsilon_norm": 8.019619941711426,
    "prediction_norm": 7.065737247467041,
    "timestamp": "2025-07-14T12:35:03.855151"
  },
  {
    "epoch": 2502,
    "batch_idx": 160,
    "loss": 0.7302449941635132,
    "learning_rate": 0.00014755907302881958,
    "timestep_mean": 429.5,
    "epsilon_norm": 9.271408081054688,
    "prediction_norm": 6.7579450607299805,
    "timestamp": "2025-07-14T12:35:03.860259"
  },
  {
    "epoch": 2502,
    "batch_idx": 192,
    "loss": 0.1465587317943573,
    "learning_rate": 0.00014755907302881958,
    "timestep_mean": 550.375,
    "epsilon_norm": 3.4786202907562256,
    "prediction_norm": 3.089637517929077,
    "timestamp": "2025-07-14T12:35:03.865315"
  },
  {
    "epoch": 2503,
    "batch_idx": 0,
    "loss": 0.21337459981441498,
    "learning_rate": 0.0001481166099460551,
    "timestep_mean": 427.1875,
    "epsilon_norm": 7.1554951667785645,
    "prediction_norm": 6.080084800720215,
    "timestamp": "2025-07-14T12:35:03.870457"
  },
  {
    "epoch": 2503,
    "batch_idx": 32,
    "loss": 0.32924604415893555,
    "learning_rate": 0.0001481166099460551,
    "timestep_mean": 457.4375,
    "epsilon_norm": 8.76717472076416,
    "prediction_norm": 7.241843223571777,
    "timestamp": "2025-07-14T12:35:03.875749"
  },
  {
    "epoch": 2503,
    "batch_idx": 64,
    "loss": 0.18474654853343964,
    "learning_rate": 0.0001481166099460551,
    "timestep_mean": 480.28125,
    "epsilon_norm": 7.570378303527832,
    "prediction_norm": 6.87298059463501,
    "timestamp": "2025-07-14T12:35:03.880800"
  },
  {
    "epoch": 2503,
    "batch_idx": 96,
    "loss": 0.38892117142677307,
    "learning_rate": 0.0001481166099460551,
    "timestep_mean": 480.28125,
    "epsilon_norm": 9.337646484375,
    "prediction_norm": 6.748590469360352,
    "timestamp": "2025-07-14T12:35:03.885976"
  },
  {
    "epoch": 2503,
    "batch_idx": 128,
    "loss": 0.2591772675514221,
    "learning_rate": 0.0001481166099460551,
    "timestep_mean": 473.28125,
    "epsilon_norm": 8.289641380310059,
    "prediction_norm": 6.358345031738281,
    "timestamp": "2025-07-14T12:35:03.891180"
  },
  {
    "epoch": 2503,
    "batch_idx": 160,
    "loss": 0.20135249197483063,
    "learning_rate": 0.0001481166099460551,
    "timestep_mean": 514.65625,
    "epsilon_norm": 8.492837905883789,
    "prediction_norm": 7.316190242767334,
    "timestamp": "2025-07-14T12:35:03.896315"
  },
  {
    "epoch": 2503,
    "batch_idx": 192,
    "loss": 1.159481167793274,
    "learning_rate": 0.0001481166099460551,
    "timestep_mean": 347.75,
    "epsilon_norm": 5.508967399597168,
    "prediction_norm": 3.9091291427612305,
    "timestamp": "2025-07-14T12:35:03.901542"
  },
  {
    "epoch": 2504,
    "batch_idx": 0,
    "loss": 0.23382610082626343,
    "learning_rate": 0.00014867501510057573,
    "timestep_mean": 484.5625,
    "epsilon_norm": 7.132769584655762,
    "prediction_norm": 5.875900745391846,
    "timestamp": "2025-07-14T12:35:03.906622"
  },
  {
    "epoch": 2504,
    "batch_idx": 32,
    "loss": 0.2832069396972656,
    "learning_rate": 0.00014867501510057573,
    "timestep_mean": 504.03125,
    "epsilon_norm": 7.653716564178467,
    "prediction_norm": 6.133142948150635,
    "timestamp": "2025-07-14T12:35:03.911665"
  },
  {
    "epoch": 2504,
    "batch_idx": 64,
    "loss": 0.21813228726387024,
    "learning_rate": 0.00014867501510057573,
    "timestep_mean": 530.90625,
    "epsilon_norm": 8.170828819274902,
    "prediction_norm": 7.502066135406494,
    "timestamp": "2025-07-14T12:35:03.917002"
  },
  {
    "epoch": 2504,
    "batch_idx": 96,
    "loss": 0.5122686624526978,
    "learning_rate": 0.00014867501510057573,
    "timestep_mean": 471.375,
    "epsilon_norm": 8.826098442077637,
    "prediction_norm": 7.22197151184082,
    "timestamp": "2025-07-14T12:35:03.922203"
  },
  {
    "epoch": 2504,
    "batch_idx": 128,
    "loss": 0.4370266795158386,
    "learning_rate": 0.00014867501510057573,
    "timestep_mean": 529.5625,
    "epsilon_norm": 8.789656639099121,
    "prediction_norm": 7.098198890686035,
    "timestamp": "2025-07-14T12:35:03.927357"
  },
  {
    "epoch": 2504,
    "batch_idx": 160,
    "loss": 0.30433011054992676,
    "learning_rate": 0.00014867501510057573,
    "timestep_mean": 502.625,
    "epsilon_norm": 6.50470495223999,
    "prediction_norm": 5.450772285461426,
    "timestamp": "2025-07-14T12:35:03.932562"
  },
  {
    "epoch": 2504,
    "batch_idx": 192,
    "loss": 0.16261716187000275,
    "learning_rate": 0.00014867501510057573,
    "timestep_mean": 643.375,
    "epsilon_norm": 3.4525532722473145,
    "prediction_norm": 3.382550001144409,
    "timestamp": "2025-07-14T12:35:03.937695"
  },
  {
    "epoch": 2505,
    "batch_idx": 0,
    "loss": 0.25936591625213623,
    "learning_rate": 0.00014923428711457247,
    "timestep_mean": 470.3125,
    "epsilon_norm": 8.17185115814209,
    "prediction_norm": 7.0839152336120605,
    "timestamp": "2025-07-14T12:35:03.943040"
  },
  {
    "epoch": 2505,
    "batch_idx": 32,
    "loss": 0.23335584998130798,
    "learning_rate": 0.00014923428711457247,
    "timestep_mean": 484.34375,
    "epsilon_norm": 9.344503402709961,
    "prediction_norm": 8.482669830322266,
    "timestamp": "2025-07-14T12:35:03.948257"
  },
  {
    "epoch": 2505,
    "batch_idx": 64,
    "loss": 0.2162831425666809,
    "learning_rate": 0.00014923428711457247,
    "timestep_mean": 453.6875,
    "epsilon_norm": 7.4540839195251465,
    "prediction_norm": 6.436909198760986,
    "timestamp": "2025-07-14T12:35:03.953441"
  },
  {
    "epoch": 2505,
    "batch_idx": 96,
    "loss": 0.18085341155529022,
    "learning_rate": 0.00014923428711457247,
    "timestep_mean": 425.34375,
    "epsilon_norm": 7.795283794403076,
    "prediction_norm": 6.8057637214660645,
    "timestamp": "2025-07-14T12:35:03.958726"
  },
  {
    "epoch": 2505,
    "batch_idx": 128,
    "loss": 0.3226756751537323,
    "learning_rate": 0.00014923428711457247,
    "timestep_mean": 501.59375,
    "epsilon_norm": 7.815010070800781,
    "prediction_norm": 6.531152248382568,
    "timestamp": "2025-07-14T12:35:03.963947"
  },
  {
    "epoch": 2505,
    "batch_idx": 160,
    "loss": 0.4745544195175171,
    "learning_rate": 0.00014923428711457247,
    "timestep_mean": 428.28125,
    "epsilon_norm": 7.842282772064209,
    "prediction_norm": 6.149234294891357,
    "timestamp": "2025-07-14T12:35:03.969070"
  },
  {
    "epoch": 2505,
    "batch_idx": 192,
    "loss": 0.14382366836071014,
    "learning_rate": 0.00014923428711457247,
    "timestep_mean": 580.375,
    "epsilon_norm": 3.835392475128174,
    "prediction_norm": 3.6732590198516846,
    "timestamp": "2025-07-14T12:35:03.974149"
  },
  {
    "epoch": 2506,
    "batch_idx": 0,
    "loss": 0.266357958316803,
    "learning_rate": 0.0001497944246080973,
    "timestep_mean": 469.03125,
    "epsilon_norm": 7.605609893798828,
    "prediction_norm": 6.526388645172119,
    "timestamp": "2025-07-14T12:35:03.979320"
  },
  {
    "epoch": 2506,
    "batch_idx": 32,
    "loss": 0.4257199764251709,
    "learning_rate": 0.0001497944246080973,
    "timestep_mean": 401.8125,
    "epsilon_norm": 8.72609806060791,
    "prediction_norm": 5.996417999267578,
    "timestamp": "2025-07-14T12:35:03.984500"
  },
  {
    "epoch": 2506,
    "batch_idx": 64,
    "loss": 0.4667559862136841,
    "learning_rate": 0.0001497944246080973,
    "timestep_mean": 489.28125,
    "epsilon_norm": 7.399178981781006,
    "prediction_norm": 6.4496378898620605,
    "timestamp": "2025-07-14T12:35:03.989506"
  },
  {
    "epoch": 2506,
    "batch_idx": 96,
    "loss": 0.3328298330307007,
    "learning_rate": 0.0001497944246080973,
    "timestep_mean": 391.4375,
    "epsilon_norm": 8.483150482177734,
    "prediction_norm": 6.592792510986328,
    "timestamp": "2025-07-14T12:35:03.994754"
  },
  {
    "epoch": 2506,
    "batch_idx": 128,
    "loss": 0.25809526443481445,
    "learning_rate": 0.0001497944246080973,
    "timestep_mean": 497.3125,
    "epsilon_norm": 8.092830657958984,
    "prediction_norm": 6.912051200866699,
    "timestamp": "2025-07-14T12:35:03.999841"
  },
  {
    "epoch": 2506,
    "batch_idx": 160,
    "loss": 0.29891955852508545,
    "learning_rate": 0.0001497944246080973,
    "timestep_mean": 500.625,
    "epsilon_norm": 7.834741115570068,
    "prediction_norm": 6.288322448730469,
    "timestamp": "2025-07-14T12:35:04.004934"
  },
  {
    "epoch": 2506,
    "batch_idx": 192,
    "loss": 0.5999538898468018,
    "learning_rate": 0.0001497944246080973,
    "timestep_mean": 375.25,
    "epsilon_norm": 5.486845016479492,
    "prediction_norm": 3.6607751846313477,
    "timestamp": "2025-07-14T12:35:04.009909"
  },
  {
    "epoch": 2507,
    "batch_idx": 0,
    "loss": 0.2064300775527954,
    "learning_rate": 0.00015035542619906613,
    "timestep_mean": 540.15625,
    "epsilon_norm": 7.964887619018555,
    "prediction_norm": 7.386631965637207,
    "timestamp": "2025-07-14T12:35:04.015059"
  },
  {
    "epoch": 2507,
    "batch_idx": 32,
    "loss": 0.23830483853816986,
    "learning_rate": 0.00015035542619906613,
    "timestep_mean": 593.03125,
    "epsilon_norm": 8.350096702575684,
    "prediction_norm": 7.645432472229004,
    "timestamp": "2025-07-14T12:35:04.020224"
  },
  {
    "epoch": 2507,
    "batch_idx": 64,
    "loss": 0.22903874516487122,
    "learning_rate": 0.00015035542619906613,
    "timestep_mean": 448.90625,
    "epsilon_norm": 6.928468227386475,
    "prediction_norm": 6.243250846862793,
    "timestamp": "2025-07-14T12:35:04.025522"
  },
  {
    "epoch": 2507,
    "batch_idx": 96,
    "loss": 0.364856481552124,
    "learning_rate": 0.00015035542619906613,
    "timestep_mean": 567.71875,
    "epsilon_norm": 8.7251615524292,
    "prediction_norm": 6.909605979919434,
    "timestamp": "2025-07-14T12:35:04.030681"
  },
  {
    "epoch": 2507,
    "batch_idx": 128,
    "loss": 0.1937781125307083,
    "learning_rate": 0.00015035542619906613,
    "timestep_mean": 524.625,
    "epsilon_norm": 7.691380023956299,
    "prediction_norm": 6.6331257820129395,
    "timestamp": "2025-07-14T12:35:04.035800"
  },
  {
    "epoch": 2507,
    "batch_idx": 160,
    "loss": 0.2820751965045929,
    "learning_rate": 0.00015035542619906613,
    "timestep_mean": 477.65625,
    "epsilon_norm": 7.69860315322876,
    "prediction_norm": 6.435400485992432,
    "timestamp": "2025-07-14T12:35:04.040806"
  },
  {
    "epoch": 2507,
    "batch_idx": 192,
    "loss": 0.834517240524292,
    "learning_rate": 0.00015035542619906613,
    "timestep_mean": 369.625,
    "epsilon_norm": 4.472218990325928,
    "prediction_norm": 2.534841299057007,
    "timestamp": "2025-07-14T12:35:04.045877"
  },
  {
    "epoch": 2508,
    "batch_idx": 0,
    "loss": 0.33330121636390686,
    "learning_rate": 0.00015091729050326403,
    "timestep_mean": 492.34375,
    "epsilon_norm": 7.314996242523193,
    "prediction_norm": 5.590113639831543,
    "timestamp": "2025-07-14T12:35:04.051055"
  },
  {
    "epoch": 2508,
    "batch_idx": 32,
    "loss": 0.1693151593208313,
    "learning_rate": 0.00015091729050326403,
    "timestep_mean": 540.46875,
    "epsilon_norm": 7.998927116394043,
    "prediction_norm": 7.578150272369385,
    "timestamp": "2025-07-14T12:35:04.056148"
  },
  {
    "epoch": 2508,
    "batch_idx": 64,
    "loss": 0.1555628776550293,
    "learning_rate": 0.00015091729050326403,
    "timestep_mean": 621.1875,
    "epsilon_norm": 6.819228649139404,
    "prediction_norm": 6.0904622077941895,
    "timestamp": "2025-07-14T12:35:04.061275"
  },
  {
    "epoch": 2508,
    "batch_idx": 96,
    "loss": 0.21772699058055878,
    "learning_rate": 0.00015091729050326403,
    "timestep_mean": 491.6875,
    "epsilon_norm": 7.473680019378662,
    "prediction_norm": 6.707371234893799,
    "timestamp": "2025-07-14T12:35:04.066141"
  },
  {
    "epoch": 2508,
    "batch_idx": 128,
    "loss": 0.3186650574207306,
    "learning_rate": 0.00015091729050326403,
    "timestep_mean": 439.4375,
    "epsilon_norm": 7.1225810050964355,
    "prediction_norm": 5.465721130371094,
    "timestamp": "2025-07-14T12:35:04.071269"
  },
  {
    "epoch": 2508,
    "batch_idx": 160,
    "loss": 0.34157970547676086,
    "learning_rate": 0.00015091729050326403,
    "timestep_mean": 446.0625,
    "epsilon_norm": 8.48257827758789,
    "prediction_norm": 6.843090057373047,
    "timestamp": "2025-07-14T12:35:04.076363"
  },
  {
    "epoch": 2508,
    "batch_idx": 192,
    "loss": 0.21483340859413147,
    "learning_rate": 0.00015091729050326403,
    "timestep_mean": 537.0,
    "epsilon_norm": 3.6423113346099854,
    "prediction_norm": 2.9924967288970947,
    "timestamp": "2025-07-14T12:35:04.081549"
  },
  {
    "epoch": 2509,
    "batch_idx": 0,
    "loss": 0.44032108783721924,
    "learning_rate": 0.00015148001613434623,
    "timestep_mean": 458.125,
    "epsilon_norm": 7.7345147132873535,
    "prediction_norm": 5.807488441467285,
    "timestamp": "2025-07-14T12:35:04.086678"
  },
  {
    "epoch": 2509,
    "batch_idx": 32,
    "loss": 0.3301548957824707,
    "learning_rate": 0.00015148001613434623,
    "timestep_mean": 442.0,
    "epsilon_norm": 9.313339233398438,
    "prediction_norm": 7.6780924797058105,
    "timestamp": "2025-07-14T12:35:04.091774"
  },
  {
    "epoch": 2509,
    "batch_idx": 64,
    "loss": 0.16679400205612183,
    "learning_rate": 0.00015148001613434623,
    "timestep_mean": 537.65625,
    "epsilon_norm": 7.851639270782471,
    "prediction_norm": 6.904145240783691,
    "timestamp": "2025-07-14T12:35:04.097039"
  },
  {
    "epoch": 2509,
    "batch_idx": 96,
    "loss": 0.24945731461048126,
    "learning_rate": 0.00015148001613434623,
    "timestep_mean": 464.15625,
    "epsilon_norm": 7.508335113525391,
    "prediction_norm": 7.165384292602539,
    "timestamp": "2025-07-14T12:35:04.102180"
  },
  {
    "epoch": 2509,
    "batch_idx": 128,
    "loss": 0.16796207427978516,
    "learning_rate": 0.00015148001613434623,
    "timestep_mean": 593.0625,
    "epsilon_norm": 8.297513008117676,
    "prediction_norm": 7.632187843322754,
    "timestamp": "2025-07-14T12:35:04.107134"
  },
  {
    "epoch": 2509,
    "batch_idx": 160,
    "loss": 0.3336697816848755,
    "learning_rate": 0.00015148001613434623,
    "timestep_mean": 523.71875,
    "epsilon_norm": 8.544392585754395,
    "prediction_norm": 7.367135047912598,
    "timestamp": "2025-07-14T12:35:04.112071"
  },
  {
    "epoch": 2509,
    "batch_idx": 192,
    "loss": 0.052077606320381165,
    "learning_rate": 0.00015148001613434623,
    "timestep_mean": 659.5,
    "epsilon_norm": 3.8382654190063477,
    "prediction_norm": 3.809453248977661,
    "timestamp": "2025-07-14T12:35:04.116978"
  },
  {
    "epoch": 2510,
    "batch_idx": 0,
    "loss": 0.25585830211639404,
    "learning_rate": 0.00015204360170384314,
    "timestep_mean": 492.46875,
    "epsilon_norm": 7.060697078704834,
    "prediction_norm": 6.5512471199035645,
    "timestamp": "2025-07-14T12:35:04.121827"
  },
  {
    "epoch": 2510,
    "batch_idx": 32,
    "loss": 0.1680036038160324,
    "learning_rate": 0.00015204360170384314,
    "timestep_mean": 479.84375,
    "epsilon_norm": 8.450247764587402,
    "prediction_norm": 7.39198112487793,
    "timestamp": "2025-07-14T12:35:04.126652"
  },
  {
    "epoch": 2510,
    "batch_idx": 64,
    "loss": 0.33290010690689087,
    "learning_rate": 0.00015204360170384314,
    "timestep_mean": 445.5625,
    "epsilon_norm": 7.946394920349121,
    "prediction_norm": 6.561738967895508,
    "timestamp": "2025-07-14T12:35:04.131470"
  },
  {
    "epoch": 2510,
    "batch_idx": 96,
    "loss": 0.3506229519844055,
    "learning_rate": 0.00015204360170384314,
    "timestep_mean": 590.34375,
    "epsilon_norm": 7.9217963218688965,
    "prediction_norm": 6.4652485847473145,
    "timestamp": "2025-07-14T12:35:04.136311"
  },
  {
    "epoch": 2510,
    "batch_idx": 128,
    "loss": 0.3674173653125763,
    "learning_rate": 0.00015204360170384314,
    "timestep_mean": 475.8125,
    "epsilon_norm": 8.068182945251465,
    "prediction_norm": 6.089923858642578,
    "timestamp": "2025-07-14T12:35:04.141144"
  },
  {
    "epoch": 2510,
    "batch_idx": 160,
    "loss": 0.2034687101840973,
    "learning_rate": 0.00015204360170384314,
    "timestep_mean": 528.21875,
    "epsilon_norm": 7.285874843597412,
    "prediction_norm": 6.13275146484375,
    "timestamp": "2025-07-14T12:35:04.146021"
  },
  {
    "epoch": 2510,
    "batch_idx": 192,
    "loss": 0.26845860481262207,
    "learning_rate": 0.00015204360170384314,
    "timestep_mean": 576.25,
    "epsilon_norm": 3.354905605316162,
    "prediction_norm": 3.3770997524261475,
    "timestamp": "2025-07-14T12:35:04.150880"
  },
  {
    "epoch": 2511,
    "batch_idx": 0,
    "loss": 0.39671918749809265,
    "learning_rate": 0.0001526080458211638,
    "timestep_mean": 456.21875,
    "epsilon_norm": 8.384957313537598,
    "prediction_norm": 6.996165752410889,
    "timestamp": "2025-07-14T12:35:04.155769"
  },
  {
    "epoch": 2511,
    "batch_idx": 32,
    "loss": 0.12218070775270462,
    "learning_rate": 0.0001526080458211638,
    "timestep_mean": 493.0,
    "epsilon_norm": 6.274343967437744,
    "prediction_norm": 6.010973930358887,
    "timestamp": "2025-07-14T12:35:04.160590"
  },
  {
    "epoch": 2511,
    "batch_idx": 64,
    "loss": 0.14039914309978485,
    "learning_rate": 0.0001526080458211638,
    "timestep_mean": 536.09375,
    "epsilon_norm": 7.636865615844727,
    "prediction_norm": 7.096255302429199,
    "timestamp": "2025-07-14T12:35:04.165435"
  },
  {
    "epoch": 2511,
    "batch_idx": 96,
    "loss": 0.2536470890045166,
    "learning_rate": 0.0001526080458211638,
    "timestep_mean": 504.90625,
    "epsilon_norm": 8.631160736083984,
    "prediction_norm": 7.277017116546631,
    "timestamp": "2025-07-14T12:35:04.170351"
  },
  {
    "epoch": 2511,
    "batch_idx": 128,
    "loss": 0.23636192083358765,
    "learning_rate": 0.0001526080458211638,
    "timestep_mean": 430.4375,
    "epsilon_norm": 7.728488922119141,
    "prediction_norm": 5.982442378997803,
    "timestamp": "2025-07-14T12:35:04.175334"
  },
  {
    "epoch": 2511,
    "batch_idx": 160,
    "loss": 0.2795808017253876,
    "learning_rate": 0.0001526080458211638,
    "timestep_mean": 475.75,
    "epsilon_norm": 8.279967308044434,
    "prediction_norm": 7.444126605987549,
    "timestamp": "2025-07-14T12:35:04.180214"
  },
  {
    "epoch": 2511,
    "batch_idx": 192,
    "loss": 0.22845980525016785,
    "learning_rate": 0.0001526080458211638,
    "timestep_mean": 487.375,
    "epsilon_norm": 3.5188424587249756,
    "prediction_norm": 2.669529914855957,
    "timestamp": "2025-07-14T12:35:04.185030"
  },
  {
    "epoch": 2512,
    "batch_idx": 0,
    "loss": 0.4161507785320282,
    "learning_rate": 0.0001531733470935978,
    "timestep_mean": 431.28125,
    "epsilon_norm": 8.353080749511719,
    "prediction_norm": 6.930656433105469,
    "timestamp": "2025-07-14T12:35:04.189899"
  },
  {
    "epoch": 2512,
    "batch_idx": 32,
    "loss": 0.2540954351425171,
    "learning_rate": 0.0001531733470935978,
    "timestep_mean": 508.75,
    "epsilon_norm": 8.07315444946289,
    "prediction_norm": 7.476938247680664,
    "timestamp": "2025-07-14T12:35:04.194791"
  },
  {
    "epoch": 2512,
    "batch_idx": 64,
    "loss": 0.3748920261859894,
    "learning_rate": 0.0001531733470935978,
    "timestep_mean": 449.4375,
    "epsilon_norm": 8.55517864227295,
    "prediction_norm": 6.78749418258667,
    "timestamp": "2025-07-14T12:35:04.199706"
  },
  {
    "epoch": 2512,
    "batch_idx": 96,
    "loss": 0.2005980908870697,
    "learning_rate": 0.0001531733470935978,
    "timestep_mean": 584.71875,
    "epsilon_norm": 7.765892028808594,
    "prediction_norm": 6.984908580780029,
    "timestamp": "2025-07-14T12:35:04.204696"
  },
  {
    "epoch": 2512,
    "batch_idx": 128,
    "loss": 0.21376296877861023,
    "learning_rate": 0.0001531733470935978,
    "timestep_mean": 470.09375,
    "epsilon_norm": 7.818544387817383,
    "prediction_norm": 6.2703447341918945,
    "timestamp": "2025-07-14T12:35:04.209755"
  },
  {
    "epoch": 2512,
    "batch_idx": 160,
    "loss": 0.15572749078273773,
    "learning_rate": 0.0001531733470935978,
    "timestep_mean": 489.0625,
    "epsilon_norm": 7.723334312438965,
    "prediction_norm": 7.139255523681641,
    "timestamp": "2025-07-14T12:35:04.214670"
  },
  {
    "epoch": 2512,
    "batch_idx": 192,
    "loss": 0.34821009635925293,
    "learning_rate": 0.0001531733470935978,
    "timestep_mean": 489.5,
    "epsilon_norm": 3.661313533782959,
    "prediction_norm": 2.3920137882232666,
    "timestamp": "2025-07-14T12:35:04.219566"
  },
  {
    "epoch": 2513,
    "batch_idx": 0,
    "loss": 0.2960204482078552,
    "learning_rate": 0.00015373950412632116,
    "timestep_mean": 472.0,
    "epsilon_norm": 8.277203559875488,
    "prediction_norm": 7.019554615020752,
    "timestamp": "2025-07-14T12:35:04.224437"
  },
  {
    "epoch": 2513,
    "batch_idx": 32,
    "loss": 0.5122361183166504,
    "learning_rate": 0.00015373950412632116,
    "timestep_mean": 405.0625,
    "epsilon_norm": 8.639867782592773,
    "prediction_norm": 6.029783248901367,
    "timestamp": "2025-07-14T12:35:04.229269"
  },
  {
    "epoch": 2513,
    "batch_idx": 64,
    "loss": 0.5085663199424744,
    "learning_rate": 0.00015373950412632116,
    "timestep_mean": 453.625,
    "epsilon_norm": 9.318441390991211,
    "prediction_norm": 6.052178382873535,
    "timestamp": "2025-07-14T12:35:04.234258"
  },
  {
    "epoch": 2513,
    "batch_idx": 96,
    "loss": 0.4624277651309967,
    "learning_rate": 0.00015373950412632116,
    "timestep_mean": 453.40625,
    "epsilon_norm": 8.721738815307617,
    "prediction_norm": 6.5094990730285645,
    "timestamp": "2025-07-14T12:35:04.239068"
  },
  {
    "epoch": 2513,
    "batch_idx": 128,
    "loss": 0.42230671644210815,
    "learning_rate": 0.00015373950412632116,
    "timestep_mean": 462.65625,
    "epsilon_norm": 8.085079193115234,
    "prediction_norm": 6.005842685699463,
    "timestamp": "2025-07-14T12:35:04.243931"
  },
  {
    "epoch": 2513,
    "batch_idx": 160,
    "loss": 0.3462216854095459,
    "learning_rate": 0.00015373950412632116,
    "timestep_mean": 500.53125,
    "epsilon_norm": 7.026319980621338,
    "prediction_norm": 5.542911529541016,
    "timestamp": "2025-07-14T12:35:04.248735"
  },
  {
    "epoch": 2513,
    "batch_idx": 192,
    "loss": 0.16759389638900757,
    "learning_rate": 0.00015373950412632116,
    "timestep_mean": 499.5,
    "epsilon_norm": 4.373785018920898,
    "prediction_norm": 3.7309188842773438,
    "timestamp": "2025-07-14T12:35:04.253535"
  },
  {
    "epoch": 2514,
    "batch_idx": 0,
    "loss": 0.32439228892326355,
    "learning_rate": 0.00015430651552239717,
    "timestep_mean": 525.8125,
    "epsilon_norm": 8.76138973236084,
    "prediction_norm": 7.252854347229004,
    "timestamp": "2025-07-14T12:35:04.258387"
  },
  {
    "epoch": 2514,
    "batch_idx": 32,
    "loss": 0.2942959666252136,
    "learning_rate": 0.00015430651552239717,
    "timestep_mean": 488.53125,
    "epsilon_norm": 7.0708513259887695,
    "prediction_norm": 5.701213836669922,
    "timestamp": "2025-07-14T12:35:04.263313"
  },
  {
    "epoch": 2514,
    "batch_idx": 64,
    "loss": 0.1914277970790863,
    "learning_rate": 0.00015430651552239717,
    "timestep_mean": 528.71875,
    "epsilon_norm": 7.159046173095703,
    "prediction_norm": 6.522822380065918,
    "timestamp": "2025-07-14T12:35:04.268145"
  },
  {
    "epoch": 2514,
    "batch_idx": 96,
    "loss": 0.3006390333175659,
    "learning_rate": 0.00015430651552239717,
    "timestep_mean": 449.15625,
    "epsilon_norm": 7.403606414794922,
    "prediction_norm": 6.284820079803467,
    "timestamp": "2025-07-14T12:35:04.273000"
  },
  {
    "epoch": 2514,
    "batch_idx": 128,
    "loss": 0.20368032157421112,
    "learning_rate": 0.00015430651552239717,
    "timestep_mean": 564.5,
    "epsilon_norm": 8.349793434143066,
    "prediction_norm": 7.333000659942627,
    "timestamp": "2025-07-14T12:35:04.277870"
  },
  {
    "epoch": 2514,
    "batch_idx": 160,
    "loss": 0.2317228466272354,
    "learning_rate": 0.00015430651552239717,
    "timestep_mean": 530.875,
    "epsilon_norm": 8.535700798034668,
    "prediction_norm": 7.052305698394775,
    "timestamp": "2025-07-14T12:35:04.282770"
  },
  {
    "epoch": 2514,
    "batch_idx": 192,
    "loss": 0.2734092175960541,
    "learning_rate": 0.00015430651552239717,
    "timestep_mean": 533.875,
    "epsilon_norm": 4.351255416870117,
    "prediction_norm": 3.6871304512023926,
    "timestamp": "2025-07-14T12:35:04.287784"
  },
  {
    "epoch": 2515,
    "batch_idx": 0,
    "loss": 0.30294787883758545,
    "learning_rate": 0.00015487437988278175,
    "timestep_mean": 540.84375,
    "epsilon_norm": 8.132319450378418,
    "prediction_norm": 6.910858631134033,
    "timestamp": "2025-07-14T12:35:04.292896"
  },
  {
    "epoch": 2515,
    "batch_idx": 32,
    "loss": 0.2762514650821686,
    "learning_rate": 0.00015487437988278175,
    "timestep_mean": 575.15625,
    "epsilon_norm": 7.607439994812012,
    "prediction_norm": 6.593010425567627,
    "timestamp": "2025-07-14T12:35:04.298003"
  },
  {
    "epoch": 2515,
    "batch_idx": 64,
    "loss": 0.08524493128061295,
    "learning_rate": 0.00015487437988278175,
    "timestep_mean": 576.15625,
    "epsilon_norm": 7.172493934631348,
    "prediction_norm": 6.872750759124756,
    "timestamp": "2025-07-14T12:35:04.303135"
  },
  {
    "epoch": 2515,
    "batch_idx": 96,
    "loss": 0.22527354955673218,
    "learning_rate": 0.00015487437988278175,
    "timestep_mean": 407.0625,
    "epsilon_norm": 7.355942726135254,
    "prediction_norm": 6.394262313842773,
    "timestamp": "2025-07-14T12:35:04.308313"
  },
  {
    "epoch": 2515,
    "batch_idx": 128,
    "loss": 0.24898159503936768,
    "learning_rate": 0.00015487437988278175,
    "timestep_mean": 426.4375,
    "epsilon_norm": 8.034892082214355,
    "prediction_norm": 7.09135103225708,
    "timestamp": "2025-07-14T12:35:04.313623"
  },
  {
    "epoch": 2515,
    "batch_idx": 160,
    "loss": 0.24045689404010773,
    "learning_rate": 0.00015487437988278175,
    "timestep_mean": 571.65625,
    "epsilon_norm": 8.018367767333984,
    "prediction_norm": 7.220431327819824,
    "timestamp": "2025-07-14T12:35:04.318867"
  },
  {
    "epoch": 2515,
    "batch_idx": 192,
    "loss": 0.14223378896713257,
    "learning_rate": 0.00015487437988278175,
    "timestep_mean": 571.125,
    "epsilon_norm": 5.426266193389893,
    "prediction_norm": 5.085038661956787,
    "timestamp": "2025-07-14T12:35:04.323929"
  },
  {
    "epoch": 2516,
    "batch_idx": 0,
    "loss": 0.5028868913650513,
    "learning_rate": 0.00015544309580632623,
    "timestep_mean": 459.0625,
    "epsilon_norm": 8.269289016723633,
    "prediction_norm": 5.7654290199279785,
    "timestamp": "2025-07-14T12:35:04.329092"
  },
  {
    "epoch": 2516,
    "batch_idx": 32,
    "loss": 0.2613736391067505,
    "learning_rate": 0.00015544309580632623,
    "timestep_mean": 557.5625,
    "epsilon_norm": 8.05581283569336,
    "prediction_norm": 7.064676284790039,
    "timestamp": "2025-07-14T12:35:04.334381"
  },
  {
    "epoch": 2516,
    "batch_idx": 64,
    "loss": 0.21423503756523132,
    "learning_rate": 0.00015544309580632623,
    "timestep_mean": 534.1875,
    "epsilon_norm": 8.394817352294922,
    "prediction_norm": 7.453101634979248,
    "timestamp": "2025-07-14T12:35:04.339504"
  },
  {
    "epoch": 2516,
    "batch_idx": 96,
    "loss": 0.1501552313566208,
    "learning_rate": 0.00015544309580632623,
    "timestep_mean": 503.15625,
    "epsilon_norm": 8.588580131530762,
    "prediction_norm": 8.246833801269531,
    "timestamp": "2025-07-14T12:35:04.344595"
  },
  {
    "epoch": 2516,
    "batch_idx": 128,
    "loss": 0.33175790309906006,
    "learning_rate": 0.00015544309580632623,
    "timestep_mean": 469.5,
    "epsilon_norm": 9.365771293640137,
    "prediction_norm": 7.662922382354736,
    "timestamp": "2025-07-14T12:35:04.349810"
  },
  {
    "epoch": 2516,
    "batch_idx": 160,
    "loss": 0.13301673531532288,
    "learning_rate": 0.00015544309580632623,
    "timestep_mean": 546.03125,
    "epsilon_norm": 7.447661399841309,
    "prediction_norm": 6.902908802032471,
    "timestamp": "2025-07-14T12:35:04.354982"
  },
  {
    "epoch": 2516,
    "batch_idx": 192,
    "loss": 0.06869583576917648,
    "learning_rate": 0.00015544309580632623,
    "timestep_mean": 537.125,
    "epsilon_norm": 3.5506012439727783,
    "prediction_norm": 3.8873190879821777,
    "timestamp": "2025-07-14T12:35:04.360037"
  },
  {
    "epoch": 2517,
    "batch_idx": 0,
    "loss": 0.4620335102081299,
    "learning_rate": 0.00015601266188978026,
    "timestep_mean": 410.5,
    "epsilon_norm": 8.432029724121094,
    "prediction_norm": 6.440574645996094,
    "timestamp": "2025-07-14T12:35:04.365197"
  },
  {
    "epoch": 2517,
    "batch_idx": 32,
    "loss": 0.3488716185092926,
    "learning_rate": 0.00015601266188978026,
    "timestep_mean": 457.25,
    "epsilon_norm": 8.373826026916504,
    "prediction_norm": 7.637530326843262,
    "timestamp": "2025-07-14T12:35:04.370352"
  },
  {
    "epoch": 2517,
    "batch_idx": 64,
    "loss": 0.18681243062019348,
    "learning_rate": 0.00015601266188978026,
    "timestep_mean": 455.59375,
    "epsilon_norm": 7.343937873840332,
    "prediction_norm": 5.9423112869262695,
    "timestamp": "2025-07-14T12:35:04.375581"
  },
  {
    "epoch": 2517,
    "batch_idx": 96,
    "loss": 0.3190948963165283,
    "learning_rate": 0.00015601266188978026,
    "timestep_mean": 482.8125,
    "epsilon_norm": 8.769983291625977,
    "prediction_norm": 7.621682643890381,
    "timestamp": "2025-07-14T12:35:04.380797"
  },
  {
    "epoch": 2517,
    "batch_idx": 128,
    "loss": 0.3453601598739624,
    "learning_rate": 0.00015601266188978026,
    "timestep_mean": 413.375,
    "epsilon_norm": 8.44444751739502,
    "prediction_norm": 6.4504218101501465,
    "timestamp": "2025-07-14T12:35:04.385941"
  },
  {
    "epoch": 2517,
    "batch_idx": 160,
    "loss": 0.2721438705921173,
    "learning_rate": 0.00015601266188978026,
    "timestep_mean": 455.9375,
    "epsilon_norm": 7.483682155609131,
    "prediction_norm": 6.3983917236328125,
    "timestamp": "2025-07-14T12:35:04.390807"
  },
  {
    "epoch": 2517,
    "batch_idx": 192,
    "loss": 0.1995115429162979,
    "learning_rate": 0.00015601266188978026,
    "timestep_mean": 461.75,
    "epsilon_norm": 3.906752824783325,
    "prediction_norm": 3.4471852779388428,
    "timestamp": "2025-07-14T12:35:04.395644"
  },
  {
    "epoch": 2518,
    "batch_idx": 0,
    "loss": 0.28592705726623535,
    "learning_rate": 0.00015658307672779628,
    "timestep_mean": 545.5625,
    "epsilon_norm": 7.7221269607543945,
    "prediction_norm": 6.5193705558776855,
    "timestamp": "2025-07-14T12:35:04.400514"
  },
  {
    "epoch": 2518,
    "batch_idx": 32,
    "loss": 0.22960276901721954,
    "learning_rate": 0.00015658307672779628,
    "timestep_mean": 510.65625,
    "epsilon_norm": 6.9729132652282715,
    "prediction_norm": 5.964707374572754,
    "timestamp": "2025-07-14T12:35:04.405371"
  },
  {
    "epoch": 2518,
    "batch_idx": 64,
    "loss": 0.18395578861236572,
    "learning_rate": 0.00015658307672779628,
    "timestep_mean": 504.78125,
    "epsilon_norm": 8.396729469299316,
    "prediction_norm": 7.737396240234375,
    "timestamp": "2025-07-14T12:35:04.410223"
  },
  {
    "epoch": 2518,
    "batch_idx": 96,
    "loss": 0.17379456758499146,
    "learning_rate": 0.00015658307672779628,
    "timestep_mean": 558.46875,
    "epsilon_norm": 7.107521057128906,
    "prediction_norm": 6.150170803070068,
    "timestamp": "2025-07-14T12:35:04.415335"
  },
  {
    "epoch": 2518,
    "batch_idx": 128,
    "loss": 0.2359408736228943,
    "learning_rate": 0.00015658307672779628,
    "timestep_mean": 466.65625,
    "epsilon_norm": 7.75942325592041,
    "prediction_norm": 7.21380615234375,
    "timestamp": "2025-07-14T12:35:04.420486"
  },
  {
    "epoch": 2518,
    "batch_idx": 160,
    "loss": 0.2522905468940735,
    "learning_rate": 0.00015658307672779628,
    "timestep_mean": 520.21875,
    "epsilon_norm": 8.383798599243164,
    "prediction_norm": 7.456960678100586,
    "timestamp": "2025-07-14T12:35:04.425756"
  },
  {
    "epoch": 2518,
    "batch_idx": 192,
    "loss": 0.23484131693840027,
    "learning_rate": 0.00015658307672779628,
    "timestep_mean": 530.0,
    "epsilon_norm": 3.9372670650482178,
    "prediction_norm": 3.3316802978515625,
    "timestamp": "2025-07-14T12:35:04.430954"
  },
  {
    "epoch": 2519,
    "batch_idx": 0,
    "loss": 0.13137966394424438,
    "learning_rate": 0.00015715433891293245,
    "timestep_mean": 515.96875,
    "epsilon_norm": 7.300027847290039,
    "prediction_norm": 6.549612998962402,
    "timestamp": "2025-07-14T12:35:04.436108"
  },
  {
    "epoch": 2519,
    "batch_idx": 32,
    "loss": 0.2575674057006836,
    "learning_rate": 0.00015715433891293245,
    "timestep_mean": 472.375,
    "epsilon_norm": 8.215129852294922,
    "prediction_norm": 6.461212158203125,
    "timestamp": "2025-07-14T12:35:04.441218"
  },
  {
    "epoch": 2519,
    "batch_idx": 64,
    "loss": 0.20899873971939087,
    "learning_rate": 0.00015715433891293245,
    "timestep_mean": 441.1875,
    "epsilon_norm": 8.119277000427246,
    "prediction_norm": 7.144996166229248,
    "timestamp": "2025-07-14T12:35:04.446418"
  },
  {
    "epoch": 2519,
    "batch_idx": 96,
    "loss": 0.291084885597229,
    "learning_rate": 0.00015715433891293245,
    "timestep_mean": 510.84375,
    "epsilon_norm": 7.906520366668701,
    "prediction_norm": 6.524580478668213,
    "timestamp": "2025-07-14T12:35:04.451521"
  },
  {
    "epoch": 2519,
    "batch_idx": 128,
    "loss": 0.15068915486335754,
    "learning_rate": 0.00015715433891293245,
    "timestep_mean": 543.34375,
    "epsilon_norm": 8.05223560333252,
    "prediction_norm": 7.56025505065918,
    "timestamp": "2025-07-14T12:35:04.456602"
  },
  {
    "epoch": 2519,
    "batch_idx": 160,
    "loss": 0.23741905391216278,
    "learning_rate": 0.00015715433891293245,
    "timestep_mean": 539.28125,
    "epsilon_norm": 6.861361026763916,
    "prediction_norm": 5.674314022064209,
    "timestamp": "2025-07-14T12:35:04.461734"
  },
  {
    "epoch": 2519,
    "batch_idx": 192,
    "loss": 0.3194864094257355,
    "learning_rate": 0.00015715433891293245,
    "timestep_mean": 432.5,
    "epsilon_norm": 4.087325096130371,
    "prediction_norm": 3.268059730529785,
    "timestamp": "2025-07-14T12:35:04.467023"
  },
  {
    "epoch": 2520,
    "batch_idx": 0,
    "loss": 0.2545055150985718,
    "learning_rate": 0.00015772644703565617,
    "timestep_mean": 534.53125,
    "epsilon_norm": 8.17855453491211,
    "prediction_norm": 6.706248760223389,
    "timestamp": "2025-07-14T12:35:04.472189"
  },
  {
    "epoch": 2520,
    "batch_idx": 32,
    "loss": 0.29207029938697815,
    "learning_rate": 0.00015772644703565617,
    "timestep_mean": 417.65625,
    "epsilon_norm": 9.111102104187012,
    "prediction_norm": 7.7489142417907715,
    "timestamp": "2025-07-14T12:35:04.477331"
  },
  {
    "epoch": 2520,
    "batch_idx": 64,
    "loss": 0.15885230898857117,
    "learning_rate": 0.00015772644703565617,
    "timestep_mean": 515.625,
    "epsilon_norm": 8.309988021850586,
    "prediction_norm": 7.4143476486206055,
    "timestamp": "2025-07-14T12:35:04.482389"
  },
  {
    "epoch": 2520,
    "batch_idx": 96,
    "loss": 0.24913479387760162,
    "learning_rate": 0.00015772644703565617,
    "timestep_mean": 501.65625,
    "epsilon_norm": 7.853616714477539,
    "prediction_norm": 6.4813456535339355,
    "timestamp": "2025-07-14T12:35:04.487514"
  },
  {
    "epoch": 2520,
    "batch_idx": 128,
    "loss": 0.17138820886611938,
    "learning_rate": 0.00015772644703565617,
    "timestep_mean": 508.65625,
    "epsilon_norm": 7.690619945526123,
    "prediction_norm": 6.943209171295166,
    "timestamp": "2025-07-14T12:35:04.492641"
  },
  {
    "epoch": 2520,
    "batch_idx": 160,
    "loss": 0.15812817215919495,
    "learning_rate": 0.00015772644703565617,
    "timestep_mean": 541.71875,
    "epsilon_norm": 8.273398399353027,
    "prediction_norm": 7.575256824493408,
    "timestamp": "2025-07-14T12:35:04.497890"
  },
  {
    "epoch": 2520,
    "batch_idx": 192,
    "loss": 0.09761379659175873,
    "learning_rate": 0.00015772644703565617,
    "timestep_mean": 577.125,
    "epsilon_norm": 2.945631980895996,
    "prediction_norm": 2.6769697666168213,
    "timestamp": "2025-07-14T12:35:04.503065"
  },
  {
    "epoch": 2521,
    "batch_idx": 0,
    "loss": 0.2527492642402649,
    "learning_rate": 0.00015829939968434713,
    "timestep_mean": 465.34375,
    "epsilon_norm": 8.029847145080566,
    "prediction_norm": 6.764098644256592,
    "timestamp": "2025-07-14T12:35:04.508238"
  },
  {
    "epoch": 2521,
    "batch_idx": 32,
    "loss": 0.39237889647483826,
    "learning_rate": 0.00015829939968434713,
    "timestep_mean": 551.09375,
    "epsilon_norm": 7.986727714538574,
    "prediction_norm": 6.13102912902832,
    "timestamp": "2025-07-14T12:35:04.513546"
  },
  {
    "epoch": 2521,
    "batch_idx": 64,
    "loss": 0.22064974904060364,
    "learning_rate": 0.00015829939968434713,
    "timestep_mean": 487.0625,
    "epsilon_norm": 7.96185302734375,
    "prediction_norm": 6.979194641113281,
    "timestamp": "2025-07-14T12:35:04.518773"
  },
  {
    "epoch": 2521,
    "batch_idx": 96,
    "loss": 0.21318884193897247,
    "learning_rate": 0.00015829939968434713,
    "timestep_mean": 485.71875,
    "epsilon_norm": 8.326593399047852,
    "prediction_norm": 7.0627946853637695,
    "timestamp": "2025-07-14T12:35:04.524021"
  },
  {
    "epoch": 2521,
    "batch_idx": 128,
    "loss": 0.34453457593917847,
    "learning_rate": 0.00015829939968434713,
    "timestep_mean": 456.9375,
    "epsilon_norm": 8.947720527648926,
    "prediction_norm": 7.720624923706055,
    "timestamp": "2025-07-14T12:35:04.529266"
  },
  {
    "epoch": 2521,
    "batch_idx": 160,
    "loss": 0.19632269442081451,
    "learning_rate": 0.00015829939968434713,
    "timestep_mean": 432.6875,
    "epsilon_norm": 7.3008317947387695,
    "prediction_norm": 6.430189609527588,
    "timestamp": "2025-07-14T12:35:04.534612"
  },
  {
    "epoch": 2521,
    "batch_idx": 192,
    "loss": 0.49891185760498047,
    "learning_rate": 0.00015829939968434713,
    "timestep_mean": 635.125,
    "epsilon_norm": 4.986059188842773,
    "prediction_norm": 3.2420668601989746,
    "timestamp": "2025-07-14T12:35:04.539690"
  },
  {
    "epoch": 2522,
    "batch_idx": 0,
    "loss": 0.20352943241596222,
    "learning_rate": 0.00015887319544530224,
    "timestep_mean": 530.84375,
    "epsilon_norm": 8.16058349609375,
    "prediction_norm": 7.136009216308594,
    "timestamp": "2025-07-14T12:35:04.544879"
  },
  {
    "epoch": 2522,
    "batch_idx": 32,
    "loss": 0.1937314122915268,
    "learning_rate": 0.00015887319544530224,
    "timestep_mean": 530.8125,
    "epsilon_norm": 7.276880741119385,
    "prediction_norm": 6.519293785095215,
    "timestamp": "2025-07-14T12:35:04.550088"
  },
  {
    "epoch": 2522,
    "batch_idx": 64,
    "loss": 0.2929069399833679,
    "learning_rate": 0.00015887319544530224,
    "timestep_mean": 459.53125,
    "epsilon_norm": 8.665210723876953,
    "prediction_norm": 6.916376113891602,
    "timestamp": "2025-07-14T12:35:04.555319"
  },
  {
    "epoch": 2522,
    "batch_idx": 96,
    "loss": 0.7884179949760437,
    "learning_rate": 0.00015887319544530224,
    "timestep_mean": 363.5625,
    "epsilon_norm": 9.588573455810547,
    "prediction_norm": 5.965827941894531,
    "timestamp": "2025-07-14T12:35:04.560386"
  },
  {
    "epoch": 2522,
    "batch_idx": 128,
    "loss": 0.3620957136154175,
    "learning_rate": 0.00015887319544530224,
    "timestep_mean": 459.65625,
    "epsilon_norm": 7.646910190582275,
    "prediction_norm": 5.84000301361084,
    "timestamp": "2025-07-14T12:35:04.565638"
  },
  {
    "epoch": 2522,
    "batch_idx": 160,
    "loss": 0.1827506572008133,
    "learning_rate": 0.00015887319544530224,
    "timestep_mean": 546.34375,
    "epsilon_norm": 7.387204170227051,
    "prediction_norm": 6.630688667297363,
    "timestamp": "2025-07-14T12:35:04.570503"
  },
  {
    "epoch": 2522,
    "batch_idx": 192,
    "loss": 0.562537431716919,
    "learning_rate": 0.00015887319544530224,
    "timestep_mean": 321.25,
    "epsilon_norm": 4.367272853851318,
    "prediction_norm": 2.9848482608795166,
    "timestamp": "2025-07-14T12:35:04.575537"
  },
  {
    "epoch": 2523,
    "batch_idx": 0,
    "loss": 0.1843261420726776,
    "learning_rate": 0.000159447832902737,
    "timestep_mean": 607.8125,
    "epsilon_norm": 7.472573757171631,
    "prediction_norm": 7.01745080947876,
    "timestamp": "2025-07-14T12:35:04.580765"
  },
  {
    "epoch": 2523,
    "batch_idx": 32,
    "loss": 0.2636216878890991,
    "learning_rate": 0.000159447832902737,
    "timestep_mean": 484.0625,
    "epsilon_norm": 6.482326030731201,
    "prediction_norm": 5.946146011352539,
    "timestamp": "2025-07-14T12:35:04.585974"
  },
  {
    "epoch": 2523,
    "batch_idx": 64,
    "loss": 0.2195310890674591,
    "learning_rate": 0.000159447832902737,
    "timestep_mean": 492.40625,
    "epsilon_norm": 8.388704299926758,
    "prediction_norm": 7.356735706329346,
    "timestamp": "2025-07-14T12:35:04.591079"
  },
  {
    "epoch": 2523,
    "batch_idx": 96,
    "loss": 0.43675869703292847,
    "learning_rate": 0.000159447832902737,
    "timestep_mean": 419.3125,
    "epsilon_norm": 8.578009605407715,
    "prediction_norm": 6.898054122924805,
    "timestamp": "2025-07-14T12:35:04.596212"
  },
  {
    "epoch": 2523,
    "batch_idx": 128,
    "loss": 0.42414870858192444,
    "learning_rate": 0.000159447832902737,
    "timestep_mean": 401.84375,
    "epsilon_norm": 8.090485572814941,
    "prediction_norm": 5.792302131652832,
    "timestamp": "2025-07-14T12:35:04.601415"
  },
  {
    "epoch": 2523,
    "batch_idx": 160,
    "loss": 0.25053662061691284,
    "learning_rate": 0.000159447832902737,
    "timestep_mean": 488.8125,
    "epsilon_norm": 8.42497444152832,
    "prediction_norm": 7.211361408233643,
    "timestamp": "2025-07-14T12:35:04.606500"
  },
  {
    "epoch": 2523,
    "batch_idx": 192,
    "loss": 0.7706550359725952,
    "learning_rate": 0.000159447832902737,
    "timestep_mean": 361.625,
    "epsilon_norm": 3.660881280899048,
    "prediction_norm": 2.315432548522949,
    "timestamp": "2025-07-14T12:35:04.611657"
  },
  {
    "epoch": 2524,
    "batch_idx": 0,
    "loss": 0.2642638683319092,
    "learning_rate": 0.00016002331063879075,
    "timestep_mean": 490.3125,
    "epsilon_norm": 7.018298625946045,
    "prediction_norm": 6.452786922454834,
    "timestamp": "2025-07-14T12:35:04.616944"
  },
  {
    "epoch": 2524,
    "batch_idx": 32,
    "loss": 0.29194867610931396,
    "learning_rate": 0.00016002331063879075,
    "timestep_mean": 459.96875,
    "epsilon_norm": 7.4503679275512695,
    "prediction_norm": 7.0556488037109375,
    "timestamp": "2025-07-14T12:35:04.622036"
  },
  {
    "epoch": 2524,
    "batch_idx": 64,
    "loss": 0.33023953437805176,
    "learning_rate": 0.00016002331063879075,
    "timestep_mean": 458.65625,
    "epsilon_norm": 9.069907188415527,
    "prediction_norm": 7.596406936645508,
    "timestamp": "2025-07-14T12:35:04.627163"
  },
  {
    "epoch": 2524,
    "batch_idx": 96,
    "loss": 0.34436237812042236,
    "learning_rate": 0.00016002331063879075,
    "timestep_mean": 498.40625,
    "epsilon_norm": 8.719305992126465,
    "prediction_norm": 7.863966941833496,
    "timestamp": "2025-07-14T12:35:04.632310"
  },
  {
    "epoch": 2524,
    "batch_idx": 128,
    "loss": 0.1868250072002411,
    "learning_rate": 0.00016002331063879075,
    "timestep_mean": 495.375,
    "epsilon_norm": 8.02930736541748,
    "prediction_norm": 7.589868068695068,
    "timestamp": "2025-07-14T12:35:04.637352"
  },
  {
    "epoch": 2524,
    "batch_idx": 160,
    "loss": 0.2915339469909668,
    "learning_rate": 0.00016002331063879075,
    "timestep_mean": 524.09375,
    "epsilon_norm": 7.518911838531494,
    "prediction_norm": 6.785571575164795,
    "timestamp": "2025-07-14T12:35:04.642574"
  },
  {
    "epoch": 2524,
    "batch_idx": 192,
    "loss": 0.37633436918258667,
    "learning_rate": 0.00016002331063879075,
    "timestep_mean": 456.375,
    "epsilon_norm": 2.783675193786621,
    "prediction_norm": 2.08247971534729,
    "timestamp": "2025-07-14T12:35:04.647537"
  },
  {
    "epoch": 2525,
    "batch_idx": 0,
    "loss": 0.21570464968681335,
    "learning_rate": 0.00016059962723352964,
    "timestep_mean": 478.46875,
    "epsilon_norm": 7.217769622802734,
    "prediction_norm": 6.572176933288574,
    "timestamp": "2025-07-14T12:35:04.652602"
  },
  {
    "epoch": 2525,
    "batch_idx": 32,
    "loss": 0.2293812781572342,
    "learning_rate": 0.00016059962723352964,
    "timestep_mean": 445.78125,
    "epsilon_norm": 7.382665634155273,
    "prediction_norm": 6.2992753982543945,
    "timestamp": "2025-07-14T12:35:04.657755"
  },
  {
    "epoch": 2525,
    "batch_idx": 64,
    "loss": 0.358437716960907,
    "learning_rate": 0.00016059962723352964,
    "timestep_mean": 479.0,
    "epsilon_norm": 7.3873748779296875,
    "prediction_norm": 6.220217704772949,
    "timestamp": "2025-07-14T12:35:04.662714"
  },
  {
    "epoch": 2525,
    "batch_idx": 96,
    "loss": 0.4020565152168274,
    "learning_rate": 0.00016059962723352964,
    "timestep_mean": 453.6875,
    "epsilon_norm": 7.627417087554932,
    "prediction_norm": 6.474416255950928,
    "timestamp": "2025-07-14T12:35:04.667850"
  },
  {
    "epoch": 2525,
    "batch_idx": 128,
    "loss": 0.14763793349266052,
    "learning_rate": 0.00016059962723352964,
    "timestep_mean": 507.03125,
    "epsilon_norm": 7.82549524307251,
    "prediction_norm": 7.508721351623535,
    "timestamp": "2025-07-14T12:35:04.672935"
  },
  {
    "epoch": 2525,
    "batch_idx": 160,
    "loss": 0.2138974964618683,
    "learning_rate": 0.00016059962723352964,
    "timestep_mean": 479.9375,
    "epsilon_norm": 7.9019036293029785,
    "prediction_norm": 7.1178178787231445,
    "timestamp": "2025-07-14T12:35:04.678071"
  },
  {
    "epoch": 2525,
    "batch_idx": 192,
    "loss": 0.05987328663468361,
    "learning_rate": 0.00016059962723352964,
    "timestep_mean": 727.75,
    "epsilon_norm": 4.133465766906738,
    "prediction_norm": 3.9465744495391846,
    "timestamp": "2025-07-14T12:35:04.683158"
  },
  {
    "epoch": 2526,
    "batch_idx": 0,
    "loss": 0.14576925337314606,
    "learning_rate": 0.00016117678126494918,
    "timestep_mean": 565.53125,
    "epsilon_norm": 7.704497337341309,
    "prediction_norm": 7.317131519317627,
    "timestamp": "2025-07-14T12:35:04.688151"
  },
  {
    "epoch": 2526,
    "batch_idx": 32,
    "loss": 0.3146767020225525,
    "learning_rate": 0.00016117678126494918,
    "timestep_mean": 476.90625,
    "epsilon_norm": 7.37579345703125,
    "prediction_norm": 6.593120098114014,
    "timestamp": "2025-07-14T12:35:04.693070"
  },
  {
    "epoch": 2526,
    "batch_idx": 64,
    "loss": 0.2332051694393158,
    "learning_rate": 0.00016117678126494918,
    "timestep_mean": 464.53125,
    "epsilon_norm": 6.628536224365234,
    "prediction_norm": 5.3132452964782715,
    "timestamp": "2025-07-14T12:35:04.697887"
  },
  {
    "epoch": 2526,
    "batch_idx": 96,
    "loss": 0.44580185413360596,
    "learning_rate": 0.00016117678126494918,
    "timestep_mean": 456.21875,
    "epsilon_norm": 8.272505760192871,
    "prediction_norm": 6.675274848937988,
    "timestamp": "2025-07-14T12:35:04.702755"
  },
  {
    "epoch": 2526,
    "batch_idx": 128,
    "loss": 0.2377987802028656,
    "learning_rate": 0.00016117678126494918,
    "timestep_mean": 522.09375,
    "epsilon_norm": 8.387731552124023,
    "prediction_norm": 7.747990131378174,
    "timestamp": "2025-07-14T12:35:04.707663"
  },
  {
    "epoch": 2526,
    "batch_idx": 160,
    "loss": 0.15477631986141205,
    "learning_rate": 0.00016117678126494918,
    "timestep_mean": 535.625,
    "epsilon_norm": 7.565483093261719,
    "prediction_norm": 6.5329790115356445,
    "timestamp": "2025-07-14T12:35:04.712672"
  },
  {
    "epoch": 2526,
    "batch_idx": 192,
    "loss": 0.3004949390888214,
    "learning_rate": 0.00016117678126494918,
    "timestep_mean": 299.75,
    "epsilon_norm": 2.347757577896118,
    "prediction_norm": 1.26796555519104,
    "timestamp": "2025-07-14T12:35:04.717534"
  },
  {
    "epoch": 2527,
    "batch_idx": 0,
    "loss": 0.23872831463813782,
    "learning_rate": 0.0001617547713089798,
    "timestep_mean": 464.09375,
    "epsilon_norm": 8.033726692199707,
    "prediction_norm": 6.898017883300781,
    "timestamp": "2025-07-14T12:35:04.722393"
  },
  {
    "epoch": 2527,
    "batch_idx": 32,
    "loss": 0.19419468939304352,
    "learning_rate": 0.0001617547713089798,
    "timestep_mean": 459.4375,
    "epsilon_norm": 7.532804489135742,
    "prediction_norm": 6.325260162353516,
    "timestamp": "2025-07-14T12:35:04.727292"
  },
  {
    "epoch": 2527,
    "batch_idx": 64,
    "loss": 0.25906577706336975,
    "learning_rate": 0.0001617547713089798,
    "timestep_mean": 490.75,
    "epsilon_norm": 7.511559009552002,
    "prediction_norm": 5.959120273590088,
    "timestamp": "2025-07-14T12:35:04.732164"
  },
  {
    "epoch": 2527,
    "batch_idx": 96,
    "loss": 0.1223943904042244,
    "learning_rate": 0.0001617547713089798,
    "timestep_mean": 592.625,
    "epsilon_norm": 6.891476631164551,
    "prediction_norm": 6.524792194366455,
    "timestamp": "2025-07-14T12:35:04.736968"
  },
  {
    "epoch": 2527,
    "batch_idx": 128,
    "loss": 0.3491623103618622,
    "learning_rate": 0.0001617547713089798,
    "timestep_mean": 494.03125,
    "epsilon_norm": 7.310731887817383,
    "prediction_norm": 5.9182448387146,
    "timestamp": "2025-07-14T12:35:04.741941"
  },
  {
    "epoch": 2527,
    "batch_idx": 160,
    "loss": 0.19918636977672577,
    "learning_rate": 0.0001617547713089798,
    "timestep_mean": 535.78125,
    "epsilon_norm": 8.022697448730469,
    "prediction_norm": 6.90111780166626,
    "timestamp": "2025-07-14T12:35:04.746800"
  },
  {
    "epoch": 2527,
    "batch_idx": 192,
    "loss": 0.3325067162513733,
    "learning_rate": 0.0001617547713089798,
    "timestep_mean": 459.25,
    "epsilon_norm": 3.4707655906677246,
    "prediction_norm": 2.221080780029297,
    "timestamp": "2025-07-14T12:35:04.751567"
  },
  {
    "epoch": 2528,
    "batch_idx": 0,
    "loss": 0.2950529158115387,
    "learning_rate": 0.0001623335959394883,
    "timestep_mean": 462.71875,
    "epsilon_norm": 7.869575023651123,
    "prediction_norm": 7.180719375610352,
    "timestamp": "2025-07-14T12:35:04.756401"
  },
  {
    "epoch": 2528,
    "batch_idx": 32,
    "loss": 0.1978110671043396,
    "learning_rate": 0.0001623335959394883,
    "timestep_mean": 506.71875,
    "epsilon_norm": 7.444151878356934,
    "prediction_norm": 6.241314888000488,
    "timestamp": "2025-07-14T12:35:04.761261"
  },
  {
    "epoch": 2528,
    "batch_idx": 64,
    "loss": 0.22308990359306335,
    "learning_rate": 0.0001623335959394883,
    "timestep_mean": 531.25,
    "epsilon_norm": 8.158791542053223,
    "prediction_norm": 6.577489376068115,
    "timestamp": "2025-07-14T12:35:04.766118"
  },
  {
    "epoch": 2528,
    "batch_idx": 96,
    "loss": 0.2217121422290802,
    "learning_rate": 0.0001623335959394883,
    "timestep_mean": 473.3125,
    "epsilon_norm": 7.298884391784668,
    "prediction_norm": 6.358685493469238,
    "timestamp": "2025-07-14T12:35:04.771017"
  },
  {
    "epoch": 2528,
    "batch_idx": 128,
    "loss": 0.24559623003005981,
    "learning_rate": 0.0001623335959394883,
    "timestep_mean": 490.40625,
    "epsilon_norm": 8.635152816772461,
    "prediction_norm": 7.763047695159912,
    "timestamp": "2025-07-14T12:35:04.775888"
  },
  {
    "epoch": 2528,
    "batch_idx": 160,
    "loss": 0.2571219205856323,
    "learning_rate": 0.0001623335959394883,
    "timestep_mean": 550.21875,
    "epsilon_norm": 8.183111190795898,
    "prediction_norm": 6.95981502532959,
    "timestamp": "2025-07-14T12:35:04.780864"
  },
  {
    "epoch": 2528,
    "batch_idx": 192,
    "loss": 0.352299302816391,
    "learning_rate": 0.0001623335959394883,
    "timestep_mean": 333.375,
    "epsilon_norm": 3.7539045810699463,
    "prediction_norm": 2.512601375579834,
    "timestamp": "2025-07-14T12:35:04.785810"
  },
  {
    "epoch": 2529,
    "batch_idx": 0,
    "loss": 0.26837337017059326,
    "learning_rate": 0.00016291325372828205,
    "timestep_mean": 439.09375,
    "epsilon_norm": 8.318948745727539,
    "prediction_norm": 6.490785121917725,
    "timestamp": "2025-07-14T12:35:04.790983"
  },
  {
    "epoch": 2529,
    "batch_idx": 32,
    "loss": 0.46249836683273315,
    "learning_rate": 0.00016291325372828205,
    "timestep_mean": 465.40625,
    "epsilon_norm": 9.201986312866211,
    "prediction_norm": 6.438231945037842,
    "timestamp": "2025-07-14T12:35:04.796048"
  },
  {
    "epoch": 2529,
    "batch_idx": 64,
    "loss": 0.3950771689414978,
    "learning_rate": 0.00016291325372828205,
    "timestep_mean": 427.0,
    "epsilon_norm": 7.347455024719238,
    "prediction_norm": 5.940491199493408,
    "timestamp": "2025-07-14T12:35:04.801092"
  },
  {
    "epoch": 2529,
    "batch_idx": 96,
    "loss": 0.17324574291706085,
    "learning_rate": 0.00016291325372828205,
    "timestep_mean": 532.71875,
    "epsilon_norm": 6.802042484283447,
    "prediction_norm": 6.753670692443848,
    "timestamp": "2025-07-14T12:35:04.806245"
  },
  {
    "epoch": 2529,
    "batch_idx": 128,
    "loss": 0.30092567205429077,
    "learning_rate": 0.00016291325372828205,
    "timestep_mean": 456.5625,
    "epsilon_norm": 7.720818042755127,
    "prediction_norm": 6.0374250411987305,
    "timestamp": "2025-07-14T12:35:04.811364"
  },
  {
    "epoch": 2529,
    "batch_idx": 160,
    "loss": 0.19895994663238525,
    "learning_rate": 0.00016291325372828205,
    "timestep_mean": 525.25,
    "epsilon_norm": 7.381173133850098,
    "prediction_norm": 6.439419269561768,
    "timestamp": "2025-07-14T12:35:04.816451"
  },
  {
    "epoch": 2529,
    "batch_idx": 192,
    "loss": 0.1269294023513794,
    "learning_rate": 0.00016291325372828205,
    "timestep_mean": 463.25,
    "epsilon_norm": 4.031614780426025,
    "prediction_norm": 3.2713735103607178,
    "timestamp": "2025-07-14T12:35:04.821645"
  },
  {
    "epoch": 2530,
    "batch_idx": 0,
    "loss": 0.15916940569877625,
    "learning_rate": 0.00016349374324511374,
    "timestep_mean": 519.84375,
    "epsilon_norm": 7.55808162689209,
    "prediction_norm": 6.912642955780029,
    "timestamp": "2025-07-14T12:35:04.826804"
  },
  {
    "epoch": 2530,
    "batch_idx": 32,
    "loss": 0.2373787760734558,
    "learning_rate": 0.00016349374324511374,
    "timestep_mean": 546.1875,
    "epsilon_norm": 8.637886047363281,
    "prediction_norm": 7.235086441040039,
    "timestamp": "2025-07-14T12:35:04.831897"
  },
  {
    "epoch": 2530,
    "batch_idx": 64,
    "loss": 0.35097283124923706,
    "learning_rate": 0.00016349374324511374,
    "timestep_mean": 478.125,
    "epsilon_norm": 8.213900566101074,
    "prediction_norm": 5.9325642585754395,
    "timestamp": "2025-07-14T12:35:04.837128"
  },
  {
    "epoch": 2530,
    "batch_idx": 96,
    "loss": 0.37595537304878235,
    "learning_rate": 0.00016349374324511374,
    "timestep_mean": 537.40625,
    "epsilon_norm": 7.655056476593018,
    "prediction_norm": 6.196350574493408,
    "timestamp": "2025-07-14T12:35:04.842233"
  },
  {
    "epoch": 2530,
    "batch_idx": 128,
    "loss": 0.18477533757686615,
    "learning_rate": 0.00016349374324511374,
    "timestep_mean": 510.90625,
    "epsilon_norm": 9.196513175964355,
    "prediction_norm": 7.924676895141602,
    "timestamp": "2025-07-14T12:35:04.847346"
  },
  {
    "epoch": 2530,
    "batch_idx": 160,
    "loss": 0.2396390289068222,
    "learning_rate": 0.00016349374324511374,
    "timestep_mean": 513.90625,
    "epsilon_norm": 7.892491817474365,
    "prediction_norm": 6.5677080154418945,
    "timestamp": "2025-07-14T12:35:04.852201"
  },
  {
    "epoch": 2530,
    "batch_idx": 192,
    "loss": 0.12819227576255798,
    "learning_rate": 0.00016349374324511374,
    "timestep_mean": 583.625,
    "epsilon_norm": 4.596446514129639,
    "prediction_norm": 4.227573394775391,
    "timestamp": "2025-07-14T12:35:04.857060"
  },
  {
    "epoch": 2531,
    "batch_idx": 0,
    "loss": 0.22454755008220673,
    "learning_rate": 0.0001640750630576827,
    "timestep_mean": 513.6875,
    "epsilon_norm": 8.590349197387695,
    "prediction_norm": 7.6012654304504395,
    "timestamp": "2025-07-14T12:35:04.861907"
  },
  {
    "epoch": 2531,
    "batch_idx": 32,
    "loss": 0.3446597158908844,
    "learning_rate": 0.0001640750630576827,
    "timestep_mean": 512.5625,
    "epsilon_norm": 8.541435241699219,
    "prediction_norm": 7.004941463470459,
    "timestamp": "2025-07-14T12:35:04.866790"
  },
  {
    "epoch": 2531,
    "batch_idx": 64,
    "loss": 0.1790269911289215,
    "learning_rate": 0.0001640750630576827,
    "timestep_mean": 511.78125,
    "epsilon_norm": 7.457998752593994,
    "prediction_norm": 6.6702561378479,
    "timestamp": "2025-07-14T12:35:04.871976"
  },
  {
    "epoch": 2531,
    "batch_idx": 96,
    "loss": 0.4473174810409546,
    "learning_rate": 0.0001640750630576827,
    "timestep_mean": 453.09375,
    "epsilon_norm": 7.762426376342773,
    "prediction_norm": 5.6699419021606445,
    "timestamp": "2025-07-14T12:35:04.877228"
  },
  {
    "epoch": 2531,
    "batch_idx": 128,
    "loss": 0.2942512035369873,
    "learning_rate": 0.0001640750630576827,
    "timestep_mean": 429.40625,
    "epsilon_norm": 7.614993572235107,
    "prediction_norm": 6.157127380371094,
    "timestamp": "2025-07-14T12:35:04.882382"
  },
  {
    "epoch": 2531,
    "batch_idx": 160,
    "loss": 0.28264284133911133,
    "learning_rate": 0.0001640750630576827,
    "timestep_mean": 502.5625,
    "epsilon_norm": 7.435344696044922,
    "prediction_norm": 5.618740081787109,
    "timestamp": "2025-07-14T12:35:04.887479"
  },
  {
    "epoch": 2531,
    "batch_idx": 192,
    "loss": 0.23531539738178253,
    "learning_rate": 0.0001640750630576827,
    "timestep_mean": 595.75,
    "epsilon_norm": 3.161419153213501,
    "prediction_norm": 2.8947315216064453,
    "timestamp": "2025-07-14T12:35:04.892694"
  },
  {
    "epoch": 2532,
    "batch_idx": 0,
    "loss": 0.27919384837150574,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 551.78125,
    "epsilon_norm": 7.952969551086426,
    "prediction_norm": 6.505899906158447,
    "timestamp": "2025-07-14T12:35:04.898029"
  },
  {
    "epoch": 2532,
    "batch_idx": 32,
    "loss": 0.20337407290935516,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 533.0625,
    "epsilon_norm": 8.560729026794434,
    "prediction_norm": 7.70805025100708,
    "timestamp": "2025-07-14T12:35:04.903207"
  },
  {
    "epoch": 2532,
    "batch_idx": 64,
    "loss": 0.12238404899835587,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 543.3125,
    "epsilon_norm": 7.495289325714111,
    "prediction_norm": 7.558399677276611,
    "timestamp": "2025-07-14T12:35:04.908422"
  },
  {
    "epoch": 2532,
    "batch_idx": 96,
    "loss": 0.2322380244731903,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 500.8125,
    "epsilon_norm": 7.431098461151123,
    "prediction_norm": 6.540485382080078,
    "timestamp": "2025-07-14T12:35:04.913692"
  },
  {
    "epoch": 2532,
    "batch_idx": 128,
    "loss": 0.2749482989311218,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 496.09375,
    "epsilon_norm": 7.586455821990967,
    "prediction_norm": 6.744515419006348,
    "timestamp": "2025-07-14T12:35:04.918902"
  },
  {
    "epoch": 2532,
    "batch_idx": 160,
    "loss": 0.15517392754554749,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 528.53125,
    "epsilon_norm": 8.958069801330566,
    "prediction_norm": 9.069700241088867,
    "timestamp": "2025-07-14T12:35:04.924015"
  },
  {
    "epoch": 2532,
    "batch_idx": 192,
    "loss": 0.25922685861587524,
    "learning_rate": 0.00016465721173164032,
    "timestep_mean": 502.75,
    "epsilon_norm": 4.375656604766846,
    "prediction_norm": 3.690265655517578,
    "timestamp": "2025-07-14T12:35:04.929142"
  },
  {
    "epoch": 2533,
    "batch_idx": 0,
    "loss": 0.2545999586582184,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 498.15625,
    "epsilon_norm": 8.606210708618164,
    "prediction_norm": 7.129477500915527,
    "timestamp": "2025-07-14T12:35:04.934406"
  },
  {
    "epoch": 2533,
    "batch_idx": 32,
    "loss": 0.1222960501909256,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 541.375,
    "epsilon_norm": 7.162229061126709,
    "prediction_norm": 6.970528602600098,
    "timestamp": "2025-07-14T12:35:04.939602"
  },
  {
    "epoch": 2533,
    "batch_idx": 64,
    "loss": 0.3452037572860718,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 571.90625,
    "epsilon_norm": 8.969757080078125,
    "prediction_norm": 8.087492942810059,
    "timestamp": "2025-07-14T12:35:04.944693"
  },
  {
    "epoch": 2533,
    "batch_idx": 96,
    "loss": 0.4191349744796753,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 465.1875,
    "epsilon_norm": 8.209409713745117,
    "prediction_norm": 6.092219829559326,
    "timestamp": "2025-07-14T12:35:04.949884"
  },
  {
    "epoch": 2533,
    "batch_idx": 128,
    "loss": 0.39286935329437256,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 534.96875,
    "epsilon_norm": 10.256805419921875,
    "prediction_norm": 8.21541976928711,
    "timestamp": "2025-07-14T12:35:04.955233"
  },
  {
    "epoch": 2533,
    "batch_idx": 160,
    "loss": 0.47726500034332275,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 478.15625,
    "epsilon_norm": 8.124473571777344,
    "prediction_norm": 6.120998859405518,
    "timestamp": "2025-07-14T12:35:04.960388"
  },
  {
    "epoch": 2533,
    "batch_idx": 192,
    "loss": 0.1528981477022171,
    "learning_rate": 0.0001652401878305925,
    "timestep_mean": 657.875,
    "epsilon_norm": 4.025742530822754,
    "prediction_norm": 3.77679705619812,
    "timestamp": "2025-07-14T12:35:04.965339"
  },
  {
    "epoch": 2534,
    "batch_idx": 0,
    "loss": 0.472278356552124,
    "learning_rate": 0.00016582398991610373,
    "timestep_mean": 495.6875,
    "epsilon_norm": 8.897421836853027,
    "prediction_norm": 7.202846050262451,
    "timestamp": "2025-07-14T12:35:04.970215"
  },
  {
    "epoch": 2534,
    "batch_idx": 32,
    "loss": 0.23049184679985046,
    "learning_rate": 0.00016582398991610373,
    "timestep_mean": 481.59375,
    "epsilon_norm": 8.108875274658203,
    "prediction_norm": 7.186767578125,
    "timestamp": "2025-07-14T12:35:04.975035"
  },
  {
    "epoch": 2534,
    "batch_idx": 64,
    "loss": 0.3124682307243347,
    "learning_rate": 0.00016582398991610373,
    "timestep_mean": 468.96875,
    "epsilon_norm": 7.8205952644348145,
    "prediction_norm": 6.2730488777160645,
    "timestamp": "2025-07-14T12:35:04.979999"
  },
  {
    "epoch": 2534,
    "batch_idx": 96,
    "loss": 0.25874295830726624,
    "learning_rate": 0.00016582398991610373,
    "timestep_mean": 470.90625,
    "epsilon_norm": 8.008883476257324,
    "prediction_norm": 6.782132148742676,
    "timestamp": "2025-07-14T12:35:04.984838"
  },
  {
    "epoch": 2534,
    "batch_idx": 128,
    "loss": 0.4938640296459198,
    "learning_rate": 0.00016582398991610373,
    "timestep_mean": 509.90625,
    "epsilon_norm": 9.144731521606445,
    "prediction_norm": 8.131425857543945,
    "timestamp": "2025-07-14T12:35:04.989951"
  },
  {
    "epoch": 2534,
    "batch_idx": 160,
    "loss": 0.1633712500333786,
    "learning_rate": 0.00016582398991610373,
    "timestep_mean": 466.6875,
    "epsilon_norm": 7.617090702056885,
    "prediction_norm": 7.430108070373535,
    "timestamp": "2025-07-14T12:35:04.995000"
  },
  {
    "epoch": 2534,
    "batch_idx": 192,
    "loss": 0.4181551933288574,
    "learning_rate": 0.00016582398991610373,
    "timestep_mean": 419.0,
    "epsilon_norm": 5.8762078285217285,
    "prediction_norm": 4.705584526062012,
    "timestamp": "2025-07-14T12:35:05.000053"
  },
  {
    "epoch": 2535,
    "batch_idx": 0,
    "loss": 0.2746516466140747,
    "learning_rate": 0.00016640861654770053,
    "timestep_mean": 437.9375,
    "epsilon_norm": 6.925642013549805,
    "prediction_norm": 6.291478157043457,
    "timestamp": "2025-07-14T12:35:05.005283"
  },
  {
    "epoch": 2535,
    "batch_idx": 32,
    "loss": 0.3256620764732361,
    "learning_rate": 0.00016640861654770053,
    "timestep_mean": 430.1875,
    "epsilon_norm": 6.971202850341797,
    "prediction_norm": 5.802309513092041,
    "timestamp": "2025-07-14T12:35:05.010339"
  },
  {
    "epoch": 2535,
    "batch_idx": 64,
    "loss": 0.09599122405052185,
    "learning_rate": 0.00016640861654770053,
    "timestep_mean": 573.0625,
    "epsilon_norm": 7.326198101043701,
    "prediction_norm": 6.771934986114502,
    "timestamp": "2025-07-14T12:35:05.015409"
  },
  {
    "epoch": 2535,
    "batch_idx": 96,
    "loss": 0.21703974902629852,
    "learning_rate": 0.00016640861654770053,
    "timestep_mean": 507.96875,
    "epsilon_norm": 8.16306209564209,
    "prediction_norm": 6.92022180557251,
    "timestamp": "2025-07-14T12:35:05.020497"
  },
  {
    "epoch": 2535,
    "batch_idx": 128,
    "loss": 0.3288552761077881,
    "learning_rate": 0.00016640861654770053,
    "timestep_mean": 496.90625,
    "epsilon_norm": 7.510863780975342,
    "prediction_norm": 6.324323654174805,
    "timestamp": "2025-07-14T12:35:05.025720"
  },
  {
    "epoch": 2535,
    "batch_idx": 160,
    "loss": 0.30719491839408875,
    "learning_rate": 0.00016640861654770053,
    "timestep_mean": 480.28125,
    "epsilon_norm": 7.638733863830566,
    "prediction_norm": 6.171480655670166,
    "timestamp": "2025-07-14T12:35:05.030898"
  },
  {
    "epoch": 2535,
    "batch_idx": 192,
    "loss": 0.3579564094543457,
    "learning_rate": 0.00016640861654770053,
    "timestep_mean": 467.875,
    "epsilon_norm": 3.4968132972717285,
    "prediction_norm": 2.433774471282959,
    "timestamp": "2025-07-14T12:35:05.035929"
  },
  {
    "epoch": 2536,
    "batch_idx": 0,
    "loss": 0.3291148245334625,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 456.1875,
    "epsilon_norm": 8.920601844787598,
    "prediction_norm": 7.096893310546875,
    "timestamp": "2025-07-14T12:35:05.041168"
  },
  {
    "epoch": 2536,
    "batch_idx": 32,
    "loss": 0.17527981102466583,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 456.6875,
    "epsilon_norm": 7.360747814178467,
    "prediction_norm": 6.617707252502441,
    "timestamp": "2025-07-14T12:35:05.046383"
  },
  {
    "epoch": 2536,
    "batch_idx": 64,
    "loss": 0.1974940299987793,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 585.15625,
    "epsilon_norm": 7.645476818084717,
    "prediction_norm": 6.663661956787109,
    "timestamp": "2025-07-14T12:35:05.051507"
  },
  {
    "epoch": 2536,
    "batch_idx": 96,
    "loss": 0.2967801094055176,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 398.21875,
    "epsilon_norm": 7.086730480194092,
    "prediction_norm": 5.732670783996582,
    "timestamp": "2025-07-14T12:35:05.056602"
  },
  {
    "epoch": 2536,
    "batch_idx": 128,
    "loss": 0.4184887111186981,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 446.03125,
    "epsilon_norm": 9.07519817352295,
    "prediction_norm": 6.888942718505859,
    "timestamp": "2025-07-14T12:35:05.061843"
  },
  {
    "epoch": 2536,
    "batch_idx": 160,
    "loss": 0.12470424175262451,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 543.375,
    "epsilon_norm": 7.967958927154541,
    "prediction_norm": 7.204914093017578,
    "timestamp": "2025-07-14T12:35:05.067052"
  },
  {
    "epoch": 2536,
    "batch_idx": 192,
    "loss": 0.16751165688037872,
    "learning_rate": 0.00016699406628287451,
    "timestep_mean": 641.75,
    "epsilon_norm": 6.145849704742432,
    "prediction_norm": 5.560732841491699,
    "timestamp": "2025-07-14T12:35:05.072228"
  },
  {
    "epoch": 2537,
    "batch_idx": 0,
    "loss": 0.16227173805236816,
    "learning_rate": 0.00016758033767708675,
    "timestep_mean": 515.21875,
    "epsilon_norm": 7.295304298400879,
    "prediction_norm": 6.321382522583008,
    "timestamp": "2025-07-14T12:35:05.077109"
  },
  {
    "epoch": 2537,
    "batch_idx": 32,
    "loss": 0.38662415742874146,
    "learning_rate": 0.00016758033767708675,
    "timestep_mean": 564.96875,
    "epsilon_norm": 8.47984790802002,
    "prediction_norm": 6.875194072723389,
    "timestamp": "2025-07-14T12:35:05.085552"
  },
  {
    "epoch": 2537,
    "batch_idx": 64,
    "loss": 0.40926769375801086,
    "learning_rate": 0.00016758033767708675,
    "timestep_mean": 463.9375,
    "epsilon_norm": 8.373209953308105,
    "prediction_norm": 7.083155155181885,
    "timestamp": "2025-07-14T12:35:05.092246"
  },
  {
    "epoch": 2537,
    "batch_idx": 96,
    "loss": 0.13693872094154358,
    "learning_rate": 0.00016758033767708675,
    "timestep_mean": 600.03125,
    "epsilon_norm": 7.89731502532959,
    "prediction_norm": 7.275704383850098,
    "timestamp": "2025-07-14T12:35:05.098646"
  },
  {
    "epoch": 2537,
    "batch_idx": 128,
    "loss": 0.15304255485534668,
    "learning_rate": 0.00016758033767708675,
    "timestep_mean": 616.6875,
    "epsilon_norm": 7.436468601226807,
    "prediction_norm": 7.176504611968994,
    "timestamp": "2025-07-14T12:35:05.104606"
  },
  {
    "epoch": 2537,
    "batch_idx": 160,
    "loss": 0.25164586305618286,
    "learning_rate": 0.00016758033767708675,
    "timestep_mean": 545.40625,
    "epsilon_norm": 7.864161968231201,
    "prediction_norm": 6.454324722290039,
    "timestamp": "2025-07-14T12:35:05.109888"
  },
  {
    "epoch": 2537,
    "batch_idx": 192,
    "loss": 0.10616762191057205,
    "learning_rate": 0.00016758033767708675,
    "timestep_mean": 495.625,
    "epsilon_norm": 4.15558385848999,
    "prediction_norm": 4.0798540115356445,
    "timestamp": "2025-07-14T12:35:05.115209"
  },
  {
    "epoch": 2538,
    "batch_idx": 0,
    "loss": 0.26913222670555115,
    "learning_rate": 0.00016816742928377098,
    "timestep_mean": 438.75,
    "epsilon_norm": 8.031946182250977,
    "prediction_norm": 6.477512359619141,
    "timestamp": "2025-07-14T12:35:05.120545"
  },
  {
    "epoch": 2538,
    "batch_idx": 32,
    "loss": 0.2617172598838806,
    "learning_rate": 0.00016816742928377098,
    "timestep_mean": 525.84375,
    "epsilon_norm": 6.634580612182617,
    "prediction_norm": 5.184713840484619,
    "timestamp": "2025-07-14T12:35:05.125860"
  },
  {
    "epoch": 2538,
    "batch_idx": 64,
    "loss": 0.3741798996925354,
    "learning_rate": 0.00016816742928377098,
    "timestep_mean": 406.90625,
    "epsilon_norm": 7.744754791259766,
    "prediction_norm": 6.544876575469971,
    "timestamp": "2025-07-14T12:35:05.131052"
  },
  {
    "epoch": 2538,
    "batch_idx": 96,
    "loss": 0.23624306917190552,
    "learning_rate": 0.00016816742928377098,
    "timestep_mean": 570.0625,
    "epsilon_norm": 7.640351295471191,
    "prediction_norm": 6.789226055145264,
    "timestamp": "2025-07-14T12:35:05.136260"
  },
  {
    "epoch": 2538,
    "batch_idx": 128,
    "loss": 0.2403460144996643,
    "learning_rate": 0.00016816742928377098,
    "timestep_mean": 520.375,
    "epsilon_norm": 8.29751968383789,
    "prediction_norm": 7.705713272094727,
    "timestamp": "2025-07-14T12:35:05.141516"
  },
  {
    "epoch": 2538,
    "batch_idx": 160,
    "loss": 0.16605135798454285,
    "learning_rate": 0.00016816742928377098,
    "timestep_mean": 541.03125,
    "epsilon_norm": 7.580920219421387,
    "prediction_norm": 6.88102388381958,
    "timestamp": "2025-07-14T12:35:05.146903"
  },
  {
    "epoch": 2538,
    "batch_idx": 192,
    "loss": 0.188150554895401,
    "learning_rate": 0.00016816742928377098,
    "timestep_mean": 459.5,
    "epsilon_norm": 3.8282244205474854,
    "prediction_norm": 3.2552614212036133,
    "timestamp": "2025-07-14T12:35:05.152101"
  },
  {
    "epoch": 2539,
    "batch_idx": 0,
    "loss": 0.26838696002960205,
    "learning_rate": 0.00016875533965433708,
    "timestep_mean": 509.9375,
    "epsilon_norm": 8.474227905273438,
    "prediction_norm": 8.490238189697266,
    "timestamp": "2025-07-14T12:35:05.157367"
  },
  {
    "epoch": 2539,
    "batch_idx": 32,
    "loss": 0.2953757643699646,
    "learning_rate": 0.00016875533965433708,
    "timestep_mean": 505.90625,
    "epsilon_norm": 8.677626609802246,
    "prediction_norm": 7.427653789520264,
    "timestamp": "2025-07-14T12:35:05.162568"
  },
  {
    "epoch": 2539,
    "batch_idx": 64,
    "loss": 0.28488755226135254,
    "learning_rate": 0.00016875533965433708,
    "timestep_mean": 457.375,
    "epsilon_norm": 9.458307266235352,
    "prediction_norm": 8.197494506835938,
    "timestamp": "2025-07-14T12:35:05.167795"
  },
  {
    "epoch": 2539,
    "batch_idx": 96,
    "loss": 0.19057108461856842,
    "learning_rate": 0.00016875533965433708,
    "timestep_mean": 592.71875,
    "epsilon_norm": 7.679934024810791,
    "prediction_norm": 6.935659408569336,
    "timestamp": "2025-07-14T12:35:05.173138"
  },
  {
    "epoch": 2539,
    "batch_idx": 128,
    "loss": 0.33553770184516907,
    "learning_rate": 0.00016875533965433708,
    "timestep_mean": 478.9375,
    "epsilon_norm": 7.443027973175049,
    "prediction_norm": 5.921205520629883,
    "timestamp": "2025-07-14T12:35:05.178378"
  },
  {
    "epoch": 2539,
    "batch_idx": 160,
    "loss": 0.2697969377040863,
    "learning_rate": 0.00016875533965433708,
    "timestep_mean": 427.84375,
    "epsilon_norm": 8.235118865966797,
    "prediction_norm": 6.52573823928833,
    "timestamp": "2025-07-14T12:35:05.183624"
  },
  {
    "epoch": 2539,
    "batch_idx": 192,
    "loss": 0.13650286197662354,
    "learning_rate": 0.00016875533965433708,
    "timestep_mean": 391.25,
    "epsilon_norm": 4.235215663909912,
    "prediction_norm": 3.8549909591674805,
    "timestamp": "2025-07-14T12:35:05.188476"
  },
  {
    "epoch": 2540,
    "batch_idx": 0,
    "loss": 0.24042025208473206,
    "learning_rate": 0.00016934406733817435,
    "timestep_mean": 483.28125,
    "epsilon_norm": 6.801711559295654,
    "prediction_norm": 5.8289594650268555,
    "timestamp": "2025-07-14T12:35:05.193322"
  },
  {
    "epoch": 2540,
    "batch_idx": 32,
    "loss": 0.1837376058101654,
    "learning_rate": 0.00016934406733817435,
    "timestep_mean": 526.3125,
    "epsilon_norm": 7.901609420776367,
    "prediction_norm": 6.587040424346924,
    "timestamp": "2025-07-14T12:35:05.198260"
  },
  {
    "epoch": 2540,
    "batch_idx": 64,
    "loss": 0.329194039106369,
    "learning_rate": 0.00016934406733817435,
    "timestep_mean": 538.21875,
    "epsilon_norm": 9.129039764404297,
    "prediction_norm": 7.267017841339111,
    "timestamp": "2025-07-14T12:35:05.203160"
  },
  {
    "epoch": 2540,
    "batch_idx": 96,
    "loss": 0.1667553186416626,
    "learning_rate": 0.00016934406733817435,
    "timestep_mean": 540.75,
    "epsilon_norm": 8.345370292663574,
    "prediction_norm": 7.6949849128723145,
    "timestamp": "2025-07-14T12:35:05.208036"
  },
  {
    "epoch": 2540,
    "batch_idx": 128,
    "loss": 0.28976553678512573,
    "learning_rate": 0.00016934406733817435,
    "timestep_mean": 500.4375,
    "epsilon_norm": 6.977273941040039,
    "prediction_norm": 5.4258246421813965,
    "timestamp": "2025-07-14T12:35:05.213005"
  },
  {
    "epoch": 2540,
    "batch_idx": 160,
    "loss": 0.3187008202075958,
    "learning_rate": 0.00016934406733817435,
    "timestep_mean": 527.6875,
    "epsilon_norm": 8.136682510375977,
    "prediction_norm": 6.593501091003418,
    "timestamp": "2025-07-14T12:35:05.217908"
  },
  {
    "epoch": 2540,
    "batch_idx": 192,
    "loss": 0.2630597949028015,
    "learning_rate": 0.00016934406733817435,
    "timestep_mean": 519.375,
    "epsilon_norm": 2.7187018394470215,
    "prediction_norm": 2.039236068725586,
    "timestamp": "2025-07-14T12:35:05.222717"
  },
  {
    "epoch": 2541,
    "batch_idx": 0,
    "loss": 0.25757163763046265,
    "learning_rate": 0.00016993361088265597,
    "timestep_mean": 525.53125,
    "epsilon_norm": 8.408123016357422,
    "prediction_norm": 7.522886753082275,
    "timestamp": "2025-07-14T12:35:05.227648"
  },
  {
    "epoch": 2541,
    "batch_idx": 32,
    "loss": 0.2925368547439575,
    "learning_rate": 0.00016993361088265597,
    "timestep_mean": 462.5625,
    "epsilon_norm": 7.724408149719238,
    "prediction_norm": 6.608741283416748,
    "timestamp": "2025-07-14T12:35:05.232515"
  },
  {
    "epoch": 2541,
    "batch_idx": 64,
    "loss": 0.26123547554016113,
    "learning_rate": 0.00016993361088265597,
    "timestep_mean": 475.40625,
    "epsilon_norm": 7.568858623504639,
    "prediction_norm": 5.851210117340088,
    "timestamp": "2025-07-14T12:35:05.237362"
  },
  {
    "epoch": 2541,
    "batch_idx": 96,
    "loss": 0.18605750799179077,
    "learning_rate": 0.00016993361088265597,
    "timestep_mean": 488.5625,
    "epsilon_norm": 8.420845985412598,
    "prediction_norm": 7.174912929534912,
    "timestamp": "2025-07-14T12:35:05.242374"
  },
  {
    "epoch": 2541,
    "batch_idx": 128,
    "loss": 0.2722322940826416,
    "learning_rate": 0.00016993361088265597,
    "timestep_mean": 453.8125,
    "epsilon_norm": 8.027314186096191,
    "prediction_norm": 6.859435558319092,
    "timestamp": "2025-07-14T12:35:05.247234"
  },
  {
    "epoch": 2541,
    "batch_idx": 160,
    "loss": 0.2629297077655792,
    "learning_rate": 0.00016993361088265597,
    "timestep_mean": 426.8125,
    "epsilon_norm": 9.881673812866211,
    "prediction_norm": 8.831016540527344,
    "timestamp": "2025-07-14T12:35:05.252175"
  },
  {
    "epoch": 2541,
    "batch_idx": 192,
    "loss": 0.08981647342443466,
    "learning_rate": 0.00016993361088265597,
    "timestep_mean": 457.0,
    "epsilon_norm": 3.4516491889953613,
    "prediction_norm": 3.47255277633667,
    "timestamp": "2025-07-14T12:35:05.257105"
  },
  {
    "epoch": 2542,
    "batch_idx": 0,
    "loss": 0.17996056377887726,
    "learning_rate": 0.00017052396883314192,
    "timestep_mean": 613.3125,
    "epsilon_norm": 7.900116920471191,
    "prediction_norm": 7.335625171661377,
    "timestamp": "2025-07-14T12:35:05.262178"
  },
  {
    "epoch": 2542,
    "batch_idx": 32,
    "loss": 0.1886141449213028,
    "learning_rate": 0.00017052396883314192,
    "timestep_mean": 534.0625,
    "epsilon_norm": 7.414186477661133,
    "prediction_norm": 6.8658037185668945,
    "timestamp": "2025-07-14T12:35:05.267190"
  },
  {
    "epoch": 2542,
    "batch_idx": 64,
    "loss": 0.1289222538471222,
    "learning_rate": 0.00017052396883314192,
    "timestep_mean": 530.75,
    "epsilon_norm": 8.044144630432129,
    "prediction_norm": 7.8470611572265625,
    "timestamp": "2025-07-14T12:35:05.272340"
  },
  {
    "epoch": 2542,
    "batch_idx": 96,
    "loss": 0.20965833961963654,
    "learning_rate": 0.00017052396883314192,
    "timestep_mean": 516.15625,
    "epsilon_norm": 8.0732421875,
    "prediction_norm": 7.839020729064941,
    "timestamp": "2025-07-14T12:35:05.277420"
  },
  {
    "epoch": 2542,
    "batch_idx": 128,
    "loss": 0.16732239723205566,
    "learning_rate": 0.00017052396883314192,
    "timestep_mean": 427.3125,
    "epsilon_norm": 8.33337688446045,
    "prediction_norm": 7.0953145027160645,
    "timestamp": "2025-07-14T12:35:05.282465"
  },
  {
    "epoch": 2542,
    "batch_idx": 160,
    "loss": 0.31590592861175537,
    "learning_rate": 0.00017052396883314192,
    "timestep_mean": 402.125,
    "epsilon_norm": 7.481116771697998,
    "prediction_norm": 6.122381687164307,
    "timestamp": "2025-07-14T12:35:05.287624"
  },
  {
    "epoch": 2542,
    "batch_idx": 192,
    "loss": 0.25593575835227966,
    "learning_rate": 0.00017052396883314192,
    "timestep_mean": 497.0,
    "epsilon_norm": 3.740901231765747,
    "prediction_norm": 3.1130030155181885,
    "timestamp": "2025-07-14T12:35:05.292769"
  },
  {
    "epoch": 2543,
    "batch_idx": 0,
    "loss": 0.47430166602134705,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 412.4375,
    "epsilon_norm": 8.76781940460205,
    "prediction_norm": 6.409969329833984,
    "timestamp": "2025-07-14T12:35:05.297998"
  },
  {
    "epoch": 2543,
    "batch_idx": 32,
    "loss": 0.11932586133480072,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 584.0,
    "epsilon_norm": 7.179952621459961,
    "prediction_norm": 6.184137344360352,
    "timestamp": "2025-07-14T12:35:05.303234"
  },
  {
    "epoch": 2543,
    "batch_idx": 64,
    "loss": 0.37282824516296387,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 392.59375,
    "epsilon_norm": 8.058667182922363,
    "prediction_norm": 6.407229900360107,
    "timestamp": "2025-07-14T12:35:05.308363"
  },
  {
    "epoch": 2543,
    "batch_idx": 96,
    "loss": 0.18614336848258972,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 516.25,
    "epsilon_norm": 7.529827117919922,
    "prediction_norm": 6.949192047119141,
    "timestamp": "2025-07-14T12:35:05.313748"
  },
  {
    "epoch": 2543,
    "batch_idx": 128,
    "loss": 0.3449443280696869,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 422.21875,
    "epsilon_norm": 8.571609497070312,
    "prediction_norm": 7.030488014221191,
    "timestamp": "2025-07-14T12:35:05.318967"
  },
  {
    "epoch": 2543,
    "batch_idx": 160,
    "loss": 0.3063030540943146,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 510.875,
    "epsilon_norm": 8.123120307922363,
    "prediction_norm": 7.214145183563232,
    "timestamp": "2025-07-14T12:35:05.323970"
  },
  {
    "epoch": 2543,
    "batch_idx": 192,
    "loss": 0.31103670597076416,
    "learning_rate": 0.00017111513973298227,
    "timestep_mean": 572.375,
    "epsilon_norm": 3.437417984008789,
    "prediction_norm": 3.270955801010132,
    "timestamp": "2025-07-14T12:35:05.329123"
  },
  {
    "epoch": 2544,
    "batch_idx": 0,
    "loss": 0.25446563959121704,
    "learning_rate": 0.00017170712212352213,
    "timestep_mean": 528.25,
    "epsilon_norm": 6.657229423522949,
    "prediction_norm": 5.254910469055176,
    "timestamp": "2025-07-14T12:35:05.334414"
  },
  {
    "epoch": 2544,
    "batch_idx": 32,
    "loss": 0.34267711639404297,
    "learning_rate": 0.00017170712212352213,
    "timestep_mean": 513.53125,
    "epsilon_norm": 8.794872283935547,
    "prediction_norm": 7.381048679351807,
    "timestamp": "2025-07-14T12:35:05.339510"
  },
  {
    "epoch": 2544,
    "batch_idx": 64,
    "loss": 0.25238654017448425,
    "learning_rate": 0.00017170712212352213,
    "timestep_mean": 445.9375,
    "epsilon_norm": 8.78108024597168,
    "prediction_norm": 7.771168231964111,
    "timestamp": "2025-07-14T12:35:05.344731"
  },
  {
    "epoch": 2544,
    "batch_idx": 96,
    "loss": 0.23458662629127502,
    "learning_rate": 0.00017170712212352213,
    "timestep_mean": 564.8125,
    "epsilon_norm": 7.44115686416626,
    "prediction_norm": 6.406454086303711,
    "timestamp": "2025-07-14T12:35:05.349947"
  },
  {
    "epoch": 2544,
    "batch_idx": 128,
    "loss": 0.46142545342445374,
    "learning_rate": 0.00017170712212352213,
    "timestep_mean": 555.34375,
    "epsilon_norm": 8.402095794677734,
    "prediction_norm": 6.4645304679870605,
    "timestamp": "2025-07-14T12:35:05.355084"
  },
  {
    "epoch": 2544,
    "batch_idx": 160,
    "loss": 0.23797070980072021,
    "learning_rate": 0.00017170712212352213,
    "timestep_mean": 585.5,
    "epsilon_norm": 9.011613845825195,
    "prediction_norm": 7.887800216674805,
    "timestamp": "2025-07-14T12:35:05.360309"
  },
  {
    "epoch": 2544,
    "batch_idx": 192,
    "loss": 0.20195350050926208,
    "learning_rate": 0.00017170712212352213,
    "timestep_mean": 446.25,
    "epsilon_norm": 4.620898723602295,
    "prediction_norm": 4.165022850036621,
    "timestamp": "2025-07-14T12:35:05.365456"
  },
  {
    "epoch": 2545,
    "batch_idx": 0,
    "loss": 0.07004791498184204,
    "learning_rate": 0.00017229991454410315,
    "timestep_mean": 556.40625,
    "epsilon_norm": 6.900074481964111,
    "prediction_norm": 6.738485813140869,
    "timestamp": "2025-07-14T12:35:05.370552"
  },
  {
    "epoch": 2545,
    "batch_idx": 32,
    "loss": 0.16738811135292053,
    "learning_rate": 0.00017229991454410315,
    "timestep_mean": 531.3125,
    "epsilon_norm": 6.413774490356445,
    "prediction_norm": 5.673899173736572,
    "timestamp": "2025-07-14T12:35:05.375807"
  },
  {
    "epoch": 2545,
    "batch_idx": 64,
    "loss": 0.43227726221084595,
    "learning_rate": 0.00017229991454410315,
    "timestep_mean": 417.78125,
    "epsilon_norm": 8.638921737670898,
    "prediction_norm": 6.603812217712402,
    "timestamp": "2025-07-14T12:35:05.381014"
  },
  {
    "epoch": 2545,
    "batch_idx": 96,
    "loss": 0.3211405277252197,
    "learning_rate": 0.00017229991454410315,
    "timestep_mean": 552.0,
    "epsilon_norm": 8.258248329162598,
    "prediction_norm": 6.7967047691345215,
    "timestamp": "2025-07-14T12:35:05.386087"
  },
  {
    "epoch": 2545,
    "batch_idx": 128,
    "loss": 0.19604098796844482,
    "learning_rate": 0.00017229991454410315,
    "timestep_mean": 505.5,
    "epsilon_norm": 7.70637845993042,
    "prediction_norm": 6.922509670257568,
    "timestamp": "2025-07-14T12:35:05.391064"
  },
  {
    "epoch": 2545,
    "batch_idx": 160,
    "loss": 0.31500113010406494,
    "learning_rate": 0.00017229991454410315,
    "timestep_mean": 534.5,
    "epsilon_norm": 7.480550289154053,
    "prediction_norm": 6.483745574951172,
    "timestamp": "2025-07-14T12:35:05.396172"
  },
  {
    "epoch": 2545,
    "batch_idx": 192,
    "loss": 0.07525703310966492,
    "learning_rate": 0.00017229991454410315,
    "timestep_mean": 556.125,
    "epsilon_norm": 4.409329891204834,
    "prediction_norm": 4.518141746520996,
    "timestamp": "2025-07-14T12:35:05.401213"
  },
  {
    "epoch": 2546,
    "batch_idx": 0,
    "loss": 0.20000270009040833,
    "learning_rate": 0.00017289351553206972,
    "timestep_mean": 514.4375,
    "epsilon_norm": 8.501627922058105,
    "prediction_norm": 7.828860759735107,
    "timestamp": "2025-07-14T12:35:05.406255"
  },
  {
    "epoch": 2546,
    "batch_idx": 32,
    "loss": 0.5838682651519775,
    "learning_rate": 0.00017289351553206972,
    "timestep_mean": 450.9375,
    "epsilon_norm": 9.945601463317871,
    "prediction_norm": 7.329063415527344,
    "timestamp": "2025-07-14T12:35:05.411156"
  },
  {
    "epoch": 2546,
    "batch_idx": 64,
    "loss": 0.37769177556037903,
    "learning_rate": 0.00017289351553206972,
    "timestep_mean": 434.21875,
    "epsilon_norm": 8.618735313415527,
    "prediction_norm": 7.609251976013184,
    "timestamp": "2025-07-14T12:35:05.416254"
  },
  {
    "epoch": 2546,
    "batch_idx": 96,
    "loss": 0.23912595212459564,
    "learning_rate": 0.00017289351553206972,
    "timestep_mean": 494.625,
    "epsilon_norm": 8.139243125915527,
    "prediction_norm": 7.457884311676025,
    "timestamp": "2025-07-14T12:35:05.421565"
  },
  {
    "epoch": 2546,
    "batch_idx": 128,
    "loss": 0.14148679375648499,
    "learning_rate": 0.00017289351553206972,
    "timestep_mean": 514.6875,
    "epsilon_norm": 8.10601806640625,
    "prediction_norm": 6.970908164978027,
    "timestamp": "2025-07-14T12:35:05.426825"
  },
  {
    "epoch": 2546,
    "batch_idx": 160,
    "loss": 0.15173488855361938,
    "learning_rate": 0.00017289351553206972,
    "timestep_mean": 562.15625,
    "epsilon_norm": 6.252120494842529,
    "prediction_norm": 5.317942142486572,
    "timestamp": "2025-07-14T12:35:05.431960"
  },
  {
    "epoch": 2546,
    "batch_idx": 192,
    "loss": 0.7407585382461548,
    "learning_rate": 0.00017289351553206972,
    "timestep_mean": 300.375,
    "epsilon_norm": 4.4948344230651855,
    "prediction_norm": 3.449125289916992,
    "timestamp": "2025-07-14T12:35:05.437260"
  },
  {
    "epoch": 2547,
    "batch_idx": 0,
    "loss": 0.21558694541454315,
    "learning_rate": 0.00017348792362276982,
    "timestep_mean": 504.4375,
    "epsilon_norm": 7.553348064422607,
    "prediction_norm": 6.460716247558594,
    "timestamp": "2025-07-14T12:35:05.442525"
  },
  {
    "epoch": 2547,
    "batch_idx": 32,
    "loss": 0.3704793155193329,
    "learning_rate": 0.00017348792362276982,
    "timestep_mean": 514.0,
    "epsilon_norm": 8.366853713989258,
    "prediction_norm": 6.581555366516113,
    "timestamp": "2025-07-14T12:35:05.447629"
  },
  {
    "epoch": 2547,
    "batch_idx": 64,
    "loss": 0.2553923726081848,
    "learning_rate": 0.00017348792362276982,
    "timestep_mean": 584.5625,
    "epsilon_norm": 8.947479248046875,
    "prediction_norm": 8.177117347717285,
    "timestamp": "2025-07-14T12:35:05.452868"
  },
  {
    "epoch": 2547,
    "batch_idx": 96,
    "loss": 0.4639224410057068,
    "learning_rate": 0.00017348792362276982,
    "timestep_mean": 458.96875,
    "epsilon_norm": 8.524964332580566,
    "prediction_norm": 6.167996883392334,
    "timestamp": "2025-07-14T12:35:05.457926"
  },
  {
    "epoch": 2547,
    "batch_idx": 128,
    "loss": 0.1763494461774826,
    "learning_rate": 0.00017348792362276982,
    "timestep_mean": 584.75,
    "epsilon_norm": 6.979572772979736,
    "prediction_norm": 6.570249080657959,
    "timestamp": "2025-07-14T12:35:05.463080"
  },
  {
    "epoch": 2547,
    "batch_idx": 160,
    "loss": 0.18223929405212402,
    "learning_rate": 0.00017348792362276982,
    "timestep_mean": 540.84375,
    "epsilon_norm": 8.195638656616211,
    "prediction_norm": 7.426591873168945,
    "timestamp": "2025-07-14T12:35:05.468161"
  },
  {
    "epoch": 2547,
    "batch_idx": 192,
    "loss": 0.10216276347637177,
    "learning_rate": 0.00017348792362276982,
    "timestep_mean": 549.625,
    "epsilon_norm": 4.139496803283691,
    "prediction_norm": 3.9465456008911133,
    "timestamp": "2025-07-14T12:35:05.473322"
  },
  {
    "epoch": 2548,
    "batch_idx": 0,
    "loss": 0.239055797457695,
    "learning_rate": 0.00017408313734956082,
    "timestep_mean": 535.78125,
    "epsilon_norm": 8.39074993133545,
    "prediction_norm": 7.697727203369141,
    "timestamp": "2025-07-14T12:35:05.478559"
  },
  {
    "epoch": 2548,
    "batch_idx": 32,
    "loss": 0.1750960350036621,
    "learning_rate": 0.00017408313734956082,
    "timestep_mean": 513.9375,
    "epsilon_norm": 7.658435344696045,
    "prediction_norm": 6.938159942626953,
    "timestamp": "2025-07-14T12:35:05.483887"
  },
  {
    "epoch": 2548,
    "batch_idx": 64,
    "loss": 0.1254051774740219,
    "learning_rate": 0.00017408313734956082,
    "timestep_mean": 478.6875,
    "epsilon_norm": 7.388942718505859,
    "prediction_norm": 7.000092029571533,
    "timestamp": "2025-07-14T12:35:05.489003"
  },
  {
    "epoch": 2548,
    "batch_idx": 96,
    "loss": 0.1688770204782486,
    "learning_rate": 0.00017408313734956082,
    "timestep_mean": 538.625,
    "epsilon_norm": 8.933088302612305,
    "prediction_norm": 7.748711585998535,
    "timestamp": "2025-07-14T12:35:05.494079"
  },
  {
    "epoch": 2548,
    "batch_idx": 128,
    "loss": 0.22197535634040833,
    "learning_rate": 0.00017408313734956082,
    "timestep_mean": 558.53125,
    "epsilon_norm": 8.92559814453125,
    "prediction_norm": 7.90032434463501,
    "timestamp": "2025-07-14T12:35:05.498991"
  },
  {
    "epoch": 2548,
    "batch_idx": 160,
    "loss": 0.18764854967594147,
    "learning_rate": 0.00017408313734956082,
    "timestep_mean": 491.3125,
    "epsilon_norm": 7.3891143798828125,
    "prediction_norm": 6.4222517013549805,
    "timestamp": "2025-07-14T12:35:05.503844"
  },
  {
    "epoch": 2548,
    "batch_idx": 192,
    "loss": 0.2377956509590149,
    "learning_rate": 0.00017408313734956082,
    "timestep_mean": 614.0,
    "epsilon_norm": 3.9972288608551025,
    "prediction_norm": 3.0932626724243164,
    "timestamp": "2025-07-14T12:35:05.508852"
  },
  {
    "epoch": 2549,
    "batch_idx": 0,
    "loss": 0.3295496702194214,
    "learning_rate": 0.00017467915524381196,
    "timestep_mean": 536.09375,
    "epsilon_norm": 8.353142738342285,
    "prediction_norm": 7.475399017333984,
    "timestamp": "2025-07-14T12:35:05.513731"
  },
  {
    "epoch": 2549,
    "batch_idx": 32,
    "loss": 0.32556986808776855,
    "learning_rate": 0.00017467915524381196,
    "timestep_mean": 440.75,
    "epsilon_norm": 6.460733413696289,
    "prediction_norm": 5.1861958503723145,
    "timestamp": "2025-07-14T12:35:05.518694"
  },
  {
    "epoch": 2549,
    "batch_idx": 64,
    "loss": 0.2603335380554199,
    "learning_rate": 0.00017467915524381196,
    "timestep_mean": 457.1875,
    "epsilon_norm": 8.24511432647705,
    "prediction_norm": 7.061618328094482,
    "timestamp": "2025-07-14T12:35:05.523847"
  },
  {
    "epoch": 2549,
    "batch_idx": 96,
    "loss": 0.2170480489730835,
    "learning_rate": 0.00017467915524381196,
    "timestep_mean": 447.3125,
    "epsilon_norm": 8.933797836303711,
    "prediction_norm": 8.397701263427734,
    "timestamp": "2025-07-14T12:35:05.528979"
  },
  {
    "epoch": 2549,
    "batch_idx": 128,
    "loss": 0.26630839705467224,
    "learning_rate": 0.00017467915524381196,
    "timestep_mean": 536.0625,
    "epsilon_norm": 7.244319438934326,
    "prediction_norm": 6.70365571975708,
    "timestamp": "2025-07-14T12:35:05.534180"
  },
  {
    "epoch": 2549,
    "batch_idx": 160,
    "loss": 0.41601142287254333,
    "learning_rate": 0.00017467915524381196,
    "timestep_mean": 427.65625,
    "epsilon_norm": 7.626640319824219,
    "prediction_norm": 6.177137851715088,
    "timestamp": "2025-07-14T12:35:05.539352"
  },
  {
    "epoch": 2549,
    "batch_idx": 192,
    "loss": 0.30357930064201355,
    "learning_rate": 0.00017467915524381196,
    "timestep_mean": 500.625,
    "epsilon_norm": 4.5579142570495605,
    "prediction_norm": 3.3845057487487793,
    "timestamp": "2025-07-14T12:35:05.544478"
  },
  {
    "epoch": 2550,
    "batch_idx": 0,
    "loss": 0.17992162704467773,
    "learning_rate": 0.00017527597583490864,
    "timestep_mean": 504.90625,
    "epsilon_norm": 6.780935764312744,
    "prediction_norm": 6.0087056159973145,
    "timestamp": "2025-07-14T12:35:05.549560"
  },
  {
    "epoch": 2550,
    "batch_idx": 32,
    "loss": 0.3752526044845581,
    "learning_rate": 0.00017527597583490864,
    "timestep_mean": 432.34375,
    "epsilon_norm": 8.143881797790527,
    "prediction_norm": 6.235514163970947,
    "timestamp": "2025-07-14T12:35:05.554696"
  },
  {
    "epoch": 2550,
    "batch_idx": 64,
    "loss": 0.3104642629623413,
    "learning_rate": 0.00017527597583490864,
    "timestep_mean": 453.875,
    "epsilon_norm": 8.370893478393555,
    "prediction_norm": 6.899819850921631,
    "timestamp": "2025-07-14T12:35:05.559708"
  },
  {
    "epoch": 2550,
    "batch_idx": 96,
    "loss": 0.2549079656600952,
    "learning_rate": 0.00017527597583490864,
    "timestep_mean": 462.53125,
    "epsilon_norm": 8.218160629272461,
    "prediction_norm": 7.106475830078125,
    "timestamp": "2025-07-14T12:35:05.564846"
  },
  {
    "epoch": 2550,
    "batch_idx": 128,
    "loss": 0.1969151794910431,
    "learning_rate": 0.00017527597583490864,
    "timestep_mean": 498.34375,
    "epsilon_norm": 7.548306941986084,
    "prediction_norm": 6.391547679901123,
    "timestamp": "2025-07-14T12:35:05.570033"
  },
  {
    "epoch": 2550,
    "batch_idx": 160,
    "loss": 0.356044203042984,
    "learning_rate": 0.00017527597583490864,
    "timestep_mean": 418.75,
    "epsilon_norm": 7.694275856018066,
    "prediction_norm": 5.966114521026611,
    "timestamp": "2025-07-14T12:35:05.575305"
  },
  {
    "epoch": 2550,
    "batch_idx": 192,
    "loss": 0.6684372425079346,
    "learning_rate": 0.00017527597583490864,
    "timestep_mean": 443.0,
    "epsilon_norm": 3.7696619033813477,
    "prediction_norm": 1.7478235960006714,
    "timestamp": "2025-07-14T12:35:05.580408"
  },
  {
    "epoch": 2551,
    "batch_idx": 0,
    "loss": 0.21022441983222961,
    "learning_rate": 0.00017587359765025443,
    "timestep_mean": 537.3125,
    "epsilon_norm": 7.339840888977051,
    "prediction_norm": 6.722222328186035,
    "timestamp": "2025-07-14T12:35:05.585462"
  },
  {
    "epoch": 2551,
    "batch_idx": 32,
    "loss": 0.4316878020763397,
    "learning_rate": 0.00017587359765025443,
    "timestep_mean": 416.6875,
    "epsilon_norm": 7.299735069274902,
    "prediction_norm": 6.074991703033447,
    "timestamp": "2025-07-14T12:35:05.590633"
  },
  {
    "epoch": 2551,
    "batch_idx": 64,
    "loss": 0.255929172039032,
    "learning_rate": 0.00017587359765025443,
    "timestep_mean": 539.4375,
    "epsilon_norm": 7.351171016693115,
    "prediction_norm": 6.0412116050720215,
    "timestamp": "2025-07-14T12:35:05.595856"
  },
  {
    "epoch": 2551,
    "batch_idx": 96,
    "loss": 0.2900075912475586,
    "learning_rate": 0.00017587359765025443,
    "timestep_mean": 563.15625,
    "epsilon_norm": 8.539811134338379,
    "prediction_norm": 7.154183387756348,
    "timestamp": "2025-07-14T12:35:05.601011"
  },
  {
    "epoch": 2551,
    "batch_idx": 128,
    "loss": 0.3094828426837921,
    "learning_rate": 0.00017587359765025443,
    "timestep_mean": 493.0,
    "epsilon_norm": 8.820304870605469,
    "prediction_norm": 7.88101863861084,
    "timestamp": "2025-07-14T12:35:05.606116"
  },
  {
    "epoch": 2551,
    "batch_idx": 160,
    "loss": 0.33562493324279785,
    "learning_rate": 0.00017587359765025443,
    "timestep_mean": 396.6875,
    "epsilon_norm": 6.523937225341797,
    "prediction_norm": 4.889268398284912,
    "timestamp": "2025-07-14T12:35:05.611302"
  },
  {
    "epoch": 2551,
    "batch_idx": 192,
    "loss": 0.37437498569488525,
    "learning_rate": 0.00017587359765025443,
    "timestep_mean": 388.875,
    "epsilon_norm": 4.44681978225708,
    "prediction_norm": 3.0627403259277344,
    "timestamp": "2025-07-14T12:35:05.616466"
  },
  {
    "epoch": 2552,
    "batch_idx": 0,
    "loss": 0.3042863607406616,
    "learning_rate": 0.00017647201921527827,
    "timestep_mean": 498.0625,
    "epsilon_norm": 8.345189094543457,
    "prediction_norm": 6.789229869842529,
    "timestamp": "2025-07-14T12:35:05.621633"
  },
  {
    "epoch": 2552,
    "batch_idx": 32,
    "loss": 0.24088913202285767,
    "learning_rate": 0.00017647201921527827,
    "timestep_mean": 454.03125,
    "epsilon_norm": 8.147768020629883,
    "prediction_norm": 6.861971855163574,
    "timestamp": "2025-07-14T12:35:05.626772"
  },
  {
    "epoch": 2552,
    "batch_idx": 64,
    "loss": 0.26229357719421387,
    "learning_rate": 0.00017647201921527827,
    "timestep_mean": 457.84375,
    "epsilon_norm": 6.630508899688721,
    "prediction_norm": 5.738804817199707,
    "timestamp": "2025-07-14T12:35:05.631880"
  },
  {
    "epoch": 2552,
    "batch_idx": 96,
    "loss": 0.4917924404144287,
    "learning_rate": 0.00017647201921527827,
    "timestep_mean": 421.0625,
    "epsilon_norm": 8.587578773498535,
    "prediction_norm": 5.9326300621032715,
    "timestamp": "2025-07-14T12:35:05.636930"
  },
  {
    "epoch": 2552,
    "batch_idx": 128,
    "loss": 0.4489397704601288,
    "learning_rate": 0.00017647201921527827,
    "timestep_mean": 371.0,
    "epsilon_norm": 7.798394680023193,
    "prediction_norm": 6.224820137023926,
    "timestamp": "2025-07-14T12:35:05.642016"
  },
  {
    "epoch": 2552,
    "batch_idx": 160,
    "loss": 0.10932035744190216,
    "learning_rate": 0.00017647201921527827,
    "timestep_mean": 561.4375,
    "epsilon_norm": 7.350022315979004,
    "prediction_norm": 6.798325061798096,
    "timestamp": "2025-07-14T12:35:05.647145"
  },
  {
    "epoch": 2552,
    "batch_idx": 192,
    "loss": 0.2548971474170685,
    "learning_rate": 0.00017647201921527827,
    "timestep_mean": 368.125,
    "epsilon_norm": 3.616719961166382,
    "prediction_norm": 2.5605146884918213,
    "timestamp": "2025-07-14T12:35:05.652197"
  },
  {
    "epoch": 2553,
    "batch_idx": 0,
    "loss": 0.2851099967956543,
    "learning_rate": 0.00017707123905343335,
    "timestep_mean": 533.5625,
    "epsilon_norm": 8.634230613708496,
    "prediction_norm": 6.858168125152588,
    "timestamp": "2025-07-14T12:35:05.657420"
  },
  {
    "epoch": 2553,
    "batch_idx": 32,
    "loss": 0.1984904259443283,
    "learning_rate": 0.00017707123905343335,
    "timestep_mean": 515.34375,
    "epsilon_norm": 8.408045768737793,
    "prediction_norm": 6.959193706512451,
    "timestamp": "2025-07-14T12:35:05.662518"
  },
  {
    "epoch": 2553,
    "batch_idx": 64,
    "loss": 0.3822658061981201,
    "learning_rate": 0.00017707123905343335,
    "timestep_mean": 398.375,
    "epsilon_norm": 7.411473274230957,
    "prediction_norm": 5.779748439788818,
    "timestamp": "2025-07-14T12:35:05.667711"
  },
  {
    "epoch": 2553,
    "batch_idx": 96,
    "loss": 0.20165598392486572,
    "learning_rate": 0.00017707123905343335,
    "timestep_mean": 392.96875,
    "epsilon_norm": 7.319247245788574,
    "prediction_norm": 5.574121475219727,
    "timestamp": "2025-07-14T12:35:05.672925"
  },
  {
    "epoch": 2553,
    "batch_idx": 128,
    "loss": 0.1349097192287445,
    "learning_rate": 0.00017707123905343335,
    "timestep_mean": 504.09375,
    "epsilon_norm": 8.031024932861328,
    "prediction_norm": 7.439765930175781,
    "timestamp": "2025-07-14T12:35:05.678146"
  },
  {
    "epoch": 2553,
    "batch_idx": 160,
    "loss": 0.32641175389289856,
    "learning_rate": 0.00017707123905343335,
    "timestep_mean": 520.21875,
    "epsilon_norm": 8.74975872039795,
    "prediction_norm": 6.569629669189453,
    "timestamp": "2025-07-14T12:35:05.683297"
  },
  {
    "epoch": 2553,
    "batch_idx": 192,
    "loss": 0.16530519723892212,
    "learning_rate": 0.00017707123905343335,
    "timestep_mean": 361.125,
    "epsilon_norm": 2.6643824577331543,
    "prediction_norm": 2.2965950965881348,
    "timestamp": "2025-07-14T12:35:05.688524"
  },
  {
    "epoch": 2554,
    "batch_idx": 0,
    "loss": 0.16559851169586182,
    "learning_rate": 0.00017767125568620467,
    "timestep_mean": 593.1875,
    "epsilon_norm": 9.151310920715332,
    "prediction_norm": 8.031389236450195,
    "timestamp": "2025-07-14T12:35:05.693754"
  },
  {
    "epoch": 2554,
    "batch_idx": 32,
    "loss": 0.3931226134300232,
    "learning_rate": 0.00017767125568620467,
    "timestep_mean": 487.5,
    "epsilon_norm": 8.86948299407959,
    "prediction_norm": 6.900519847869873,
    "timestamp": "2025-07-14T12:35:05.698890"
  },
  {
    "epoch": 2554,
    "batch_idx": 64,
    "loss": 0.3289554715156555,
    "learning_rate": 0.00017767125568620467,
    "timestep_mean": 398.40625,
    "epsilon_norm": 7.531031608581543,
    "prediction_norm": 5.812903881072998,
    "timestamp": "2025-07-14T12:35:05.704086"
  },
  {
    "epoch": 2554,
    "batch_idx": 96,
    "loss": 0.23460015654563904,
    "learning_rate": 0.00017767125568620467,
    "timestep_mean": 540.40625,
    "epsilon_norm": 8.977707862854004,
    "prediction_norm": 7.750682830810547,
    "timestamp": "2025-07-14T12:35:05.709405"
  },
  {
    "epoch": 2554,
    "batch_idx": 128,
    "loss": 0.2775203585624695,
    "learning_rate": 0.00017767125568620467,
    "timestep_mean": 514.15625,
    "epsilon_norm": 7.932437419891357,
    "prediction_norm": 6.699092388153076,
    "timestamp": "2025-07-14T12:35:05.714529"
  },
  {
    "epoch": 2554,
    "batch_idx": 160,
    "loss": 0.2671721577644348,
    "learning_rate": 0.00017767125568620467,
    "timestep_mean": 583.8125,
    "epsilon_norm": 6.620089530944824,
    "prediction_norm": 5.559974670410156,
    "timestamp": "2025-07-14T12:35:05.719857"
  },
  {
    "epoch": 2554,
    "batch_idx": 192,
    "loss": 0.0858907550573349,
    "learning_rate": 0.00017767125568620467,
    "timestep_mean": 350.75,
    "epsilon_norm": 2.60408616065979,
    "prediction_norm": 2.6870031356811523,
    "timestamp": "2025-07-14T12:35:05.725010"
  },
  {
    "epoch": 2555,
    "batch_idx": 0,
    "loss": 0.15100891888141632,
    "learning_rate": 0.00017827206763311083,
    "timestep_mean": 629.09375,
    "epsilon_norm": 9.0944242477417,
    "prediction_norm": 8.809372901916504,
    "timestamp": "2025-07-14T12:35:05.730406"
  },
  {
    "epoch": 2555,
    "batch_idx": 32,
    "loss": 0.2977466881275177,
    "learning_rate": 0.00017827206763311083,
    "timestep_mean": 438.09375,
    "epsilon_norm": 7.009055137634277,
    "prediction_norm": 6.031764984130859,
    "timestamp": "2025-07-14T12:35:05.735654"
  },
  {
    "epoch": 2555,
    "batch_idx": 64,
    "loss": 0.1768413931131363,
    "learning_rate": 0.00017827206763311083,
    "timestep_mean": 531.8125,
    "epsilon_norm": 8.06821346282959,
    "prediction_norm": 7.486626148223877,
    "timestamp": "2025-07-14T12:35:05.740803"
  },
  {
    "epoch": 2555,
    "batch_idx": 96,
    "loss": 0.3273288905620575,
    "learning_rate": 0.00017827206763311083,
    "timestep_mean": 519.03125,
    "epsilon_norm": 8.871583938598633,
    "prediction_norm": 7.553252696990967,
    "timestamp": "2025-07-14T12:35:05.745910"
  },
  {
    "epoch": 2555,
    "batch_idx": 128,
    "loss": 0.36171412467956543,
    "learning_rate": 0.00017827206763311083,
    "timestep_mean": 510.65625,
    "epsilon_norm": 7.044261455535889,
    "prediction_norm": 4.838630676269531,
    "timestamp": "2025-07-14T12:35:05.751071"
  },
  {
    "epoch": 2555,
    "batch_idx": 160,
    "loss": 0.5085867643356323,
    "learning_rate": 0.00017827206763311083,
    "timestep_mean": 477.34375,
    "epsilon_norm": 8.252053260803223,
    "prediction_norm": 6.0275187492370605,
    "timestamp": "2025-07-14T12:35:05.755981"
  },
  {
    "epoch": 2555,
    "batch_idx": 192,
    "loss": 0.12597717344760895,
    "learning_rate": 0.00017827206763311083,
    "timestep_mean": 605.5,
    "epsilon_norm": 4.142990589141846,
    "prediction_norm": 3.892194986343384,
    "timestamp": "2025-07-14T12:35:05.760846"
  },
  {
    "epoch": 2556,
    "batch_idx": 0,
    "loss": 0.344230592250824,
    "learning_rate": 0.0001788736734117081,
    "timestep_mean": 396.28125,
    "epsilon_norm": 7.5570902824401855,
    "prediction_norm": 6.257689476013184,
    "timestamp": "2025-07-14T12:35:05.765710"
  },
  {
    "epoch": 2556,
    "batch_idx": 32,
    "loss": 0.14890463650226593,
    "learning_rate": 0.0001788736734117081,
    "timestep_mean": 573.0625,
    "epsilon_norm": 7.851339817047119,
    "prediction_norm": 7.413495063781738,
    "timestamp": "2025-07-14T12:35:05.770572"
  },
  {
    "epoch": 2556,
    "batch_idx": 64,
    "loss": 0.41806110739707947,
    "learning_rate": 0.0001788736734117081,
    "timestep_mean": 422.4375,
    "epsilon_norm": 8.559404373168945,
    "prediction_norm": 6.607265472412109,
    "timestamp": "2025-07-14T12:35:05.775604"
  },
  {
    "epoch": 2556,
    "batch_idx": 96,
    "loss": 0.18691088259220123,
    "learning_rate": 0.0001788736734117081,
    "timestep_mean": 534.1875,
    "epsilon_norm": 7.489044189453125,
    "prediction_norm": 6.572136878967285,
    "timestamp": "2025-07-14T12:35:05.780765"
  },
  {
    "epoch": 2556,
    "batch_idx": 128,
    "loss": 0.25593769550323486,
    "learning_rate": 0.0001788736734117081,
    "timestep_mean": 428.75,
    "epsilon_norm": 7.373432159423828,
    "prediction_norm": 6.846863269805908,
    "timestamp": "2025-07-14T12:35:05.785865"
  },
  {
    "epoch": 2556,
    "batch_idx": 160,
    "loss": 0.25716498494148254,
    "learning_rate": 0.0001788736734117081,
    "timestep_mean": 476.65625,
    "epsilon_norm": 9.33867073059082,
    "prediction_norm": 8.72839069366455,
    "timestamp": "2025-07-14T12:35:05.790923"
  },
  {
    "epoch": 2556,
    "batch_idx": 192,
    "loss": 0.12651968002319336,
    "learning_rate": 0.0001788736734117081,
    "timestep_mean": 540.125,
    "epsilon_norm": 3.0636589527130127,
    "prediction_norm": 3.0992071628570557,
    "timestamp": "2025-07-14T12:35:05.795962"
  },
  {
    "epoch": 2557,
    "batch_idx": 0,
    "loss": 0.24936668574810028,
    "learning_rate": 0.00017947607153759402,
    "timestep_mean": 536.6875,
    "epsilon_norm": 6.530949115753174,
    "prediction_norm": 5.399301528930664,
    "timestamp": "2025-07-14T12:35:05.801188"
  },
  {
    "epoch": 2557,
    "batch_idx": 32,
    "loss": 0.3820332884788513,
    "learning_rate": 0.00017947607153759402,
    "timestep_mean": 485.15625,
    "epsilon_norm": 8.540337562561035,
    "prediction_norm": 6.684327602386475,
    "timestamp": "2025-07-14T12:35:05.806284"
  },
  {
    "epoch": 2557,
    "batch_idx": 64,
    "loss": 0.46390092372894287,
    "learning_rate": 0.00017947607153759402,
    "timestep_mean": 491.25,
    "epsilon_norm": 8.8905611038208,
    "prediction_norm": 7.019575595855713,
    "timestamp": "2025-07-14T12:35:05.811418"
  },
  {
    "epoch": 2557,
    "batch_idx": 96,
    "loss": 0.3167015314102173,
    "learning_rate": 0.00017947607153759402,
    "timestep_mean": 447.15625,
    "epsilon_norm": 8.64529037475586,
    "prediction_norm": 7.547784805297852,
    "timestamp": "2025-07-14T12:35:05.816651"
  },
  {
    "epoch": 2557,
    "batch_idx": 128,
    "loss": 0.16300278902053833,
    "learning_rate": 0.00017947607153759402,
    "timestep_mean": 501.8125,
    "epsilon_norm": 8.781635284423828,
    "prediction_norm": 8.029624938964844,
    "timestamp": "2025-07-14T12:35:05.821948"
  },
  {
    "epoch": 2557,
    "batch_idx": 160,
    "loss": 0.33672767877578735,
    "learning_rate": 0.00017947607153759402,
    "timestep_mean": 474.25,
    "epsilon_norm": 8.14172077178955,
    "prediction_norm": 6.623349189758301,
    "timestamp": "2025-07-14T12:35:05.827177"
  },
  {
    "epoch": 2557,
    "batch_idx": 192,
    "loss": 0.25511449575424194,
    "learning_rate": 0.00017947607153759402,
    "timestep_mean": 434.0,
    "epsilon_norm": 4.130009651184082,
    "prediction_norm": 2.9770281314849854,
    "timestamp": "2025-07-14T12:35:05.832271"
  },
  {
    "epoch": 2558,
    "batch_idx": 0,
    "loss": 0.3352721631526947,
    "learning_rate": 0.00018007926052441135,
    "timestep_mean": 510.59375,
    "epsilon_norm": 8.460088729858398,
    "prediction_norm": 7.047835350036621,
    "timestamp": "2025-07-14T12:35:05.837388"
  },
  {
    "epoch": 2558,
    "batch_idx": 32,
    "loss": 0.3727551996707916,
    "learning_rate": 0.00018007926052441135,
    "timestep_mean": 487.53125,
    "epsilon_norm": 8.167702674865723,
    "prediction_norm": 7.085231304168701,
    "timestamp": "2025-07-14T12:35:05.842745"
  },
  {
    "epoch": 2558,
    "batch_idx": 64,
    "loss": 0.2528136372566223,
    "learning_rate": 0.00018007926052441135,
    "timestep_mean": 556.09375,
    "epsilon_norm": 7.366929531097412,
    "prediction_norm": 6.59584903717041,
    "timestamp": "2025-07-14T12:35:05.847928"
  },
  {
    "epoch": 2558,
    "batch_idx": 96,
    "loss": 0.33975750207901,
    "learning_rate": 0.00018007926052441135,
    "timestep_mean": 514.5,
    "epsilon_norm": 9.100724220275879,
    "prediction_norm": 7.158052444458008,
    "timestamp": "2025-07-14T12:35:05.853084"
  },
  {
    "epoch": 2558,
    "batch_idx": 128,
    "loss": 0.15761779248714447,
    "learning_rate": 0.00018007926052441135,
    "timestep_mean": 503.0625,
    "epsilon_norm": 7.858156681060791,
    "prediction_norm": 7.55063009262085,
    "timestamp": "2025-07-14T12:35:05.858150"
  },
  {
    "epoch": 2558,
    "batch_idx": 160,
    "loss": 0.4237499237060547,
    "learning_rate": 0.00018007926052441135,
    "timestep_mean": 559.84375,
    "epsilon_norm": 10.018636703491211,
    "prediction_norm": 8.675787925720215,
    "timestamp": "2025-07-14T12:35:05.863331"
  },
  {
    "epoch": 2558,
    "batch_idx": 192,
    "loss": 0.5075602531433105,
    "learning_rate": 0.00018007926052441135,
    "timestep_mean": 419.875,
    "epsilon_norm": 3.4553048610687256,
    "prediction_norm": 2.834622383117676,
    "timestamp": "2025-07-14T12:35:05.868314"
  },
  {
    "epoch": 2559,
    "batch_idx": 0,
    "loss": 0.34407150745391846,
    "learning_rate": 0.00018068323888385034,
    "timestep_mean": 445.46875,
    "epsilon_norm": 6.707778453826904,
    "prediction_norm": 5.2923665046691895,
    "timestamp": "2025-07-14T12:35:05.873548"
  },
  {
    "epoch": 2559,
    "batch_idx": 32,
    "loss": 0.4095715284347534,
    "learning_rate": 0.00018068323888385034,
    "timestep_mean": 480.5625,
    "epsilon_norm": 8.559953689575195,
    "prediction_norm": 6.465142726898193,
    "timestamp": "2025-07-14T12:35:05.878736"
  },
  {
    "epoch": 2559,
    "batch_idx": 64,
    "loss": 0.18459343910217285,
    "learning_rate": 0.00018068323888385034,
    "timestep_mean": 421.5,
    "epsilon_norm": 6.770418167114258,
    "prediction_norm": 5.957447052001953,
    "timestamp": "2025-07-14T12:35:05.883941"
  },
  {
    "epoch": 2559,
    "batch_idx": 96,
    "loss": 0.12799504399299622,
    "learning_rate": 0.00018068323888385034,
    "timestep_mean": 614.21875,
    "epsilon_norm": 7.689248561859131,
    "prediction_norm": 6.791248798370361,
    "timestamp": "2025-07-14T12:35:05.889215"
  },
  {
    "epoch": 2559,
    "batch_idx": 128,
    "loss": 0.3420025110244751,
    "learning_rate": 0.00018068323888385034,
    "timestep_mean": 503.75,
    "epsilon_norm": 7.831173419952393,
    "prediction_norm": 6.647311210632324,
    "timestamp": "2025-07-14T12:35:05.894358"
  },
  {
    "epoch": 2559,
    "batch_idx": 160,
    "loss": 0.19832420349121094,
    "learning_rate": 0.00018068323888385034,
    "timestep_mean": 543.90625,
    "epsilon_norm": 8.370987892150879,
    "prediction_norm": 7.225758075714111,
    "timestamp": "2025-07-14T12:35:05.899591"
  },
  {
    "epoch": 2559,
    "batch_idx": 192,
    "loss": 0.5407525897026062,
    "learning_rate": 0.00018068323888385034,
    "timestep_mean": 470.75,
    "epsilon_norm": 4.595473766326904,
    "prediction_norm": 3.516270399093628,
    "timestamp": "2025-07-14T12:35:05.904772"
  },
  {
    "epoch": 2560,
    "batch_idx": 0,
    "loss": 0.19836433231830597,
    "learning_rate": 0.0001812880051256557,
    "timestep_mean": 544.125,
    "epsilon_norm": 8.23812198638916,
    "prediction_norm": 7.1207475662231445,
    "timestamp": "2025-07-14T12:35:05.909911"
  },
  {
    "epoch": 2560,
    "batch_idx": 32,
    "loss": 0.4207206666469574,
    "learning_rate": 0.0001812880051256557,
    "timestep_mean": 530.125,
    "epsilon_norm": 7.2679619789123535,
    "prediction_norm": 5.490806579589844,
    "timestamp": "2025-07-14T12:35:05.915050"
  },
  {
    "epoch": 2560,
    "batch_idx": 64,
    "loss": 0.1658461093902588,
    "learning_rate": 0.0001812880051256557,
    "timestep_mean": 543.125,
    "epsilon_norm": 7.479222297668457,
    "prediction_norm": 6.938315391540527,
    "timestamp": "2025-07-14T12:35:05.920233"
  },
  {
    "epoch": 2560,
    "batch_idx": 96,
    "loss": 0.177408367395401,
    "learning_rate": 0.0001812880051256557,
    "timestep_mean": 622.4375,
    "epsilon_norm": 7.49793815612793,
    "prediction_norm": 6.929590225219727,
    "timestamp": "2025-07-14T12:35:05.925611"
  },
  {
    "epoch": 2560,
    "batch_idx": 128,
    "loss": 0.2677622437477112,
    "learning_rate": 0.0001812880051256557,
    "timestep_mean": 506.09375,
    "epsilon_norm": 6.915901184082031,
    "prediction_norm": 5.856128215789795,
    "timestamp": "2025-07-14T12:35:05.930726"
  },
  {
    "epoch": 2560,
    "batch_idx": 160,
    "loss": 0.4240643382072449,
    "learning_rate": 0.0001812880051256557,
    "timestep_mean": 482.21875,
    "epsilon_norm": 8.809770584106445,
    "prediction_norm": 8.143836975097656,
    "timestamp": "2025-07-14T12:35:05.935932"
  },
  {
    "epoch": 2560,
    "batch_idx": 192,
    "loss": 0.22392502427101135,
    "learning_rate": 0.0001812880051256557,
    "timestep_mean": 385.375,
    "epsilon_norm": 3.404040575027466,
    "prediction_norm": 2.5295920372009277,
    "timestamp": "2025-07-14T12:35:05.941092"
  },
  {
    "epoch": 2561,
    "batch_idx": 0,
    "loss": 0.3359279930591583,
    "learning_rate": 0.0001818935577576258,
    "timestep_mean": 475.625,
    "epsilon_norm": 7.9632182121276855,
    "prediction_norm": 6.713288307189941,
    "timestamp": "2025-07-14T12:35:05.946222"
  },
  {
    "epoch": 2561,
    "batch_idx": 32,
    "loss": 0.2145182490348816,
    "learning_rate": 0.0001818935577576258,
    "timestep_mean": 581.6875,
    "epsilon_norm": 7.839762210845947,
    "prediction_norm": 6.878598213195801,
    "timestamp": "2025-07-14T12:35:05.951223"
  },
  {
    "epoch": 2561,
    "batch_idx": 64,
    "loss": 0.3101619780063629,
    "learning_rate": 0.0001818935577576258,
    "timestep_mean": 453.21875,
    "epsilon_norm": 7.621942043304443,
    "prediction_norm": 6.152863502502441,
    "timestamp": "2025-07-14T12:35:05.956348"
  },
  {
    "epoch": 2561,
    "batch_idx": 96,
    "loss": 0.26596665382385254,
    "learning_rate": 0.0001818935577576258,
    "timestep_mean": 392.5625,
    "epsilon_norm": 8.083748817443848,
    "prediction_norm": 6.801965713500977,
    "timestamp": "2025-07-14T12:35:05.961473"
  },
  {
    "epoch": 2561,
    "batch_idx": 128,
    "loss": 0.36037927865982056,
    "learning_rate": 0.0001818935577576258,
    "timestep_mean": 480.34375,
    "epsilon_norm": 8.361063003540039,
    "prediction_norm": 6.665660381317139,
    "timestamp": "2025-07-14T12:35:05.966626"
  },
  {
    "epoch": 2561,
    "batch_idx": 160,
    "loss": 0.28041326999664307,
    "learning_rate": 0.0001818935577576258,
    "timestep_mean": 514.8125,
    "epsilon_norm": 9.090912818908691,
    "prediction_norm": 7.4718146324157715,
    "timestamp": "2025-07-14T12:35:05.971926"
  },
  {
    "epoch": 2561,
    "batch_idx": 192,
    "loss": 0.034124795347452164,
    "learning_rate": 0.0001818935577576258,
    "timestep_mean": 580.625,
    "epsilon_norm": 2.825474262237549,
    "prediction_norm": 2.9940176010131836,
    "timestamp": "2025-07-14T12:35:05.977044"
  },
  {
    "epoch": 2562,
    "batch_idx": 0,
    "loss": 0.2603377401828766,
    "learning_rate": 0.00018249989528562014,
    "timestep_mean": 504.25,
    "epsilon_norm": 8.193180084228516,
    "prediction_norm": 6.776431083679199,
    "timestamp": "2025-07-14T12:35:05.982170"
  },
  {
    "epoch": 2562,
    "batch_idx": 32,
    "loss": 0.23690369725227356,
    "learning_rate": 0.00018249989528562014,
    "timestep_mean": 470.96875,
    "epsilon_norm": 8.457039833068848,
    "prediction_norm": 7.478818416595459,
    "timestamp": "2025-07-14T12:35:05.987338"
  },
  {
    "epoch": 2562,
    "batch_idx": 64,
    "loss": 0.31521353125572205,
    "learning_rate": 0.00018249989528562014,
    "timestep_mean": 480.46875,
    "epsilon_norm": 7.74384880065918,
    "prediction_norm": 6.305521011352539,
    "timestamp": "2025-07-14T12:35:05.992620"
  },
  {
    "epoch": 2562,
    "batch_idx": 96,
    "loss": 0.20763057470321655,
    "learning_rate": 0.00018249989528562014,
    "timestep_mean": 553.4375,
    "epsilon_norm": 8.598501205444336,
    "prediction_norm": 7.50482177734375,
    "timestamp": "2025-07-14T12:35:05.997759"
  },
  {
    "epoch": 2562,
    "batch_idx": 128,
    "loss": 0.8012456893920898,
    "learning_rate": 0.00018249989528562014,
    "timestep_mean": 420.3125,
    "epsilon_norm": 9.388517379760742,
    "prediction_norm": 5.90970516204834,
    "timestamp": "2025-07-14T12:35:06.002949"
  },
  {
    "epoch": 2562,
    "batch_idx": 160,
    "loss": 0.17093726992607117,
    "learning_rate": 0.00018249989528562014,
    "timestep_mean": 474.34375,
    "epsilon_norm": 7.471890926361084,
    "prediction_norm": 6.362022399902344,
    "timestamp": "2025-07-14T12:35:06.008115"
  },
  {
    "epoch": 2562,
    "batch_idx": 192,
    "loss": 0.09212326258420944,
    "learning_rate": 0.00018249989528562014,
    "timestep_mean": 491.25,
    "epsilon_norm": 2.4394006729125977,
    "prediction_norm": 2.442014217376709,
    "timestamp": "2025-07-14T12:35:06.013405"
  },
  {
    "epoch": 2563,
    "batch_idx": 0,
    "loss": 0.273630827665329,
    "learning_rate": 0.0001831070162135611,
    "timestep_mean": 524.9375,
    "epsilon_norm": 7.904688835144043,
    "prediction_norm": 7.209224224090576,
    "timestamp": "2025-07-14T12:35:06.018297"
  },
  {
    "epoch": 2563,
    "batch_idx": 32,
    "loss": 0.1889302134513855,
    "learning_rate": 0.0001831070162135611,
    "timestep_mean": 509.96875,
    "epsilon_norm": 7.7109479904174805,
    "prediction_norm": 6.515531063079834,
    "timestamp": "2025-07-14T12:35:06.023151"
  },
  {
    "epoch": 2563,
    "batch_idx": 64,
    "loss": 0.17488974332809448,
    "learning_rate": 0.0001831070162135611,
    "timestep_mean": 499.0625,
    "epsilon_norm": 9.188652038574219,
    "prediction_norm": 7.897755146026611,
    "timestamp": "2025-07-14T12:35:06.028089"
  },
  {
    "epoch": 2563,
    "batch_idx": 96,
    "loss": 0.308907151222229,
    "learning_rate": 0.0001831070162135611,
    "timestep_mean": 507.84375,
    "epsilon_norm": 7.497011661529541,
    "prediction_norm": 5.681863784790039,
    "timestamp": "2025-07-14T12:35:06.032952"
  },
  {
    "epoch": 2563,
    "batch_idx": 128,
    "loss": 0.2530764043331146,
    "learning_rate": 0.0001831070162135611,
    "timestep_mean": 495.71875,
    "epsilon_norm": 7.947809219360352,
    "prediction_norm": 6.910260200500488,
    "timestamp": "2025-07-14T12:35:06.037777"
  },
  {
    "epoch": 2563,
    "batch_idx": 160,
    "loss": 0.5585393905639648,
    "learning_rate": 0.0001831070162135611,
    "timestep_mean": 489.6875,
    "epsilon_norm": 7.806746006011963,
    "prediction_norm": 5.8676862716674805,
    "timestamp": "2025-07-14T12:35:06.042833"
  },
  {
    "epoch": 2563,
    "batch_idx": 192,
    "loss": 0.1182851791381836,
    "learning_rate": 0.0001831070162135611,
    "timestep_mean": 684.625,
    "epsilon_norm": 3.602782964706421,
    "prediction_norm": 2.8023362159729004,
    "timestamp": "2025-07-14T12:35:06.047682"
  },
  {
    "epoch": 2564,
    "batch_idx": 0,
    "loss": 0.30015096068382263,
    "learning_rate": 0.00018371491904343822,
    "timestep_mean": 514.5,
    "epsilon_norm": 8.85490894317627,
    "prediction_norm": 7.3468122482299805,
    "timestamp": "2025-07-14T12:35:06.052557"
  },
  {
    "epoch": 2564,
    "batch_idx": 32,
    "loss": 0.20909658074378967,
    "learning_rate": 0.00018371491904343822,
    "timestep_mean": 476.6875,
    "epsilon_norm": 7.003144264221191,
    "prediction_norm": 5.872583389282227,
    "timestamp": "2025-07-14T12:35:06.057439"
  },
  {
    "epoch": 2564,
    "batch_idx": 64,
    "loss": 0.28639790415763855,
    "learning_rate": 0.00018371491904343822,
    "timestep_mean": 413.59375,
    "epsilon_norm": 8.464284896850586,
    "prediction_norm": 6.855059623718262,
    "timestamp": "2025-07-14T12:35:06.062415"
  },
  {
    "epoch": 2564,
    "batch_idx": 96,
    "loss": 0.18835517764091492,
    "learning_rate": 0.00018371491904343822,
    "timestep_mean": 499.53125,
    "epsilon_norm": 8.857308387756348,
    "prediction_norm": 7.807746410369873,
    "timestamp": "2025-07-14T12:35:06.067403"
  },
  {
    "epoch": 2564,
    "batch_idx": 128,
    "loss": 0.3363625407218933,
    "learning_rate": 0.00018371491904343822,
    "timestep_mean": 533.625,
    "epsilon_norm": 8.23567008972168,
    "prediction_norm": 6.639177322387695,
    "timestamp": "2025-07-14T12:35:06.072256"
  },
  {
    "epoch": 2564,
    "batch_idx": 160,
    "loss": 0.3865610361099243,
    "learning_rate": 0.00018371491904343822,
    "timestep_mean": 535.6875,
    "epsilon_norm": 8.235515594482422,
    "prediction_norm": 7.198394298553467,
    "timestamp": "2025-07-14T12:35:06.077125"
  },
  {
    "epoch": 2564,
    "batch_idx": 192,
    "loss": 0.6986762285232544,
    "learning_rate": 0.00018371491904343822,
    "timestep_mean": 533.125,
    "epsilon_norm": 4.735008239746094,
    "prediction_norm": 3.7526803016662598,
    "timestamp": "2025-07-14T12:35:06.081933"
  },
  {
    "epoch": 2565,
    "batch_idx": 0,
    "loss": 0.2859017550945282,
    "learning_rate": 0.0001843236022753115,
    "timestep_mean": 473.0625,
    "epsilon_norm": 8.699946403503418,
    "prediction_norm": 7.812735080718994,
    "timestamp": "2025-07-14T12:35:06.086824"
  },
  {
    "epoch": 2565,
    "batch_idx": 32,
    "loss": 0.2628697156906128,
    "learning_rate": 0.0001843236022753115,
    "timestep_mean": 483.09375,
    "epsilon_norm": 8.213516235351562,
    "prediction_norm": 7.58459997177124,
    "timestamp": "2025-07-14T12:35:06.091665"
  },
  {
    "epoch": 2565,
    "batch_idx": 64,
    "loss": 0.2534153461456299,
    "learning_rate": 0.0001843236022753115,
    "timestep_mean": 506.96875,
    "epsilon_norm": 8.853170394897461,
    "prediction_norm": 8.42774772644043,
    "timestamp": "2025-07-14T12:35:06.096508"
  },
  {
    "epoch": 2565,
    "batch_idx": 96,
    "loss": 0.32079052925109863,
    "learning_rate": 0.0001843236022753115,
    "timestep_mean": 456.3125,
    "epsilon_norm": 6.233621597290039,
    "prediction_norm": 5.3062825202941895,
    "timestamp": "2025-07-14T12:35:06.101388"
  },
  {
    "epoch": 2565,
    "batch_idx": 128,
    "loss": 0.4385542869567871,
    "learning_rate": 0.0001843236022753115,
    "timestep_mean": 484.84375,
    "epsilon_norm": 8.655245780944824,
    "prediction_norm": 7.454627513885498,
    "timestamp": "2025-07-14T12:35:06.106301"
  },
  {
    "epoch": 2565,
    "batch_idx": 160,
    "loss": 0.3051544725894928,
    "learning_rate": 0.0001843236022753115,
    "timestep_mean": 524.625,
    "epsilon_norm": 8.444456100463867,
    "prediction_norm": 6.988106727600098,
    "timestamp": "2025-07-14T12:35:06.111229"
  },
  {
    "epoch": 2565,
    "batch_idx": 192,
    "loss": 0.1897335946559906,
    "learning_rate": 0.0001843236022753115,
    "timestep_mean": 550.125,
    "epsilon_norm": 3.891916036605835,
    "prediction_norm": 4.314535140991211,
    "timestamp": "2025-07-14T12:35:06.116091"
  },
  {
    "epoch": 2566,
    "batch_idx": 0,
    "loss": 0.24599558115005493,
    "learning_rate": 0.00018493306440731574,
    "timestep_mean": 472.5,
    "epsilon_norm": 6.989068984985352,
    "prediction_norm": 5.663539886474609,
    "timestamp": "2025-07-14T12:35:06.120938"
  },
  {
    "epoch": 2566,
    "batch_idx": 32,
    "loss": 0.1252291202545166,
    "learning_rate": 0.00018493306440731574,
    "timestep_mean": 554.53125,
    "epsilon_norm": 7.600152492523193,
    "prediction_norm": 7.389296054840088,
    "timestamp": "2025-07-14T12:35:06.125897"
  },
  {
    "epoch": 2566,
    "batch_idx": 64,
    "loss": 0.23217374086380005,
    "learning_rate": 0.00018493306440731574,
    "timestep_mean": 500.84375,
    "epsilon_norm": 7.272479057312012,
    "prediction_norm": 6.081084251403809,
    "timestamp": "2025-07-14T12:35:06.130892"
  },
  {
    "epoch": 2566,
    "batch_idx": 96,
    "loss": 0.21816608309745789,
    "learning_rate": 0.00018493306440731574,
    "timestep_mean": 489.9375,
    "epsilon_norm": 8.526141166687012,
    "prediction_norm": 7.631485939025879,
    "timestamp": "2025-07-14T12:35:06.135930"
  },
  {
    "epoch": 2566,
    "batch_idx": 128,
    "loss": 0.33661648631095886,
    "learning_rate": 0.00018493306440731574,
    "timestep_mean": 468.59375,
    "epsilon_norm": 7.024056434631348,
    "prediction_norm": 5.0193047523498535,
    "timestamp": "2025-07-14T12:35:06.140974"
  },
  {
    "epoch": 2566,
    "batch_idx": 160,
    "loss": 0.2705666124820709,
    "learning_rate": 0.00018493306440731574,
    "timestep_mean": 536.90625,
    "epsilon_norm": 7.1836113929748535,
    "prediction_norm": 6.280017375946045,
    "timestamp": "2025-07-14T12:35:06.146152"
  },
  {
    "epoch": 2566,
    "batch_idx": 192,
    "loss": 0.0607362762093544,
    "learning_rate": 0.00018493306440731574,
    "timestep_mean": 422.75,
    "epsilon_norm": 3.1644930839538574,
    "prediction_norm": 3.3188741207122803,
    "timestamp": "2025-07-14T12:35:06.151203"
  },
  {
    "epoch": 2567,
    "batch_idx": 0,
    "loss": 0.2506650388240814,
    "learning_rate": 0.00018554330393566373,
    "timestep_mean": 495.5625,
    "epsilon_norm": 8.569981575012207,
    "prediction_norm": 7.184218406677246,
    "timestamp": "2025-07-14T12:35:06.156268"
  },
  {
    "epoch": 2567,
    "batch_idx": 32,
    "loss": 0.376244455575943,
    "learning_rate": 0.00018554330393566373,
    "timestep_mean": 462.3125,
    "epsilon_norm": 8.235170364379883,
    "prediction_norm": 5.993122100830078,
    "timestamp": "2025-07-14T12:35:06.161507"
  },
  {
    "epoch": 2567,
    "batch_idx": 64,
    "loss": 0.21390776336193085,
    "learning_rate": 0.00018554330393566373,
    "timestep_mean": 456.8125,
    "epsilon_norm": 8.001603126525879,
    "prediction_norm": 6.9042463302612305,
    "timestamp": "2025-07-14T12:35:06.166632"
  },
  {
    "epoch": 2567,
    "batch_idx": 96,
    "loss": 0.22625690698623657,
    "learning_rate": 0.00018554330393566373,
    "timestep_mean": 479.53125,
    "epsilon_norm": 7.987061023712158,
    "prediction_norm": 7.675024509429932,
    "timestamp": "2025-07-14T12:35:06.171818"
  },
  {
    "epoch": 2567,
    "batch_idx": 128,
    "loss": 0.2111734002828598,
    "learning_rate": 0.00018554330393566373,
    "timestep_mean": 491.0,
    "epsilon_norm": 7.5090532302856445,
    "prediction_norm": 6.542463302612305,
    "timestamp": "2025-07-14T12:35:06.176876"
  },
  {
    "epoch": 2567,
    "batch_idx": 160,
    "loss": 0.34182682633399963,
    "learning_rate": 0.00018554330393566373,
    "timestep_mean": 586.96875,
    "epsilon_norm": 8.545622825622559,
    "prediction_norm": 7.341506481170654,
    "timestamp": "2025-07-14T12:35:06.182080"
  },
  {
    "epoch": 2567,
    "batch_idx": 192,
    "loss": 0.06241054832935333,
    "learning_rate": 0.00018554330393566373,
    "timestep_mean": 582.875,
    "epsilon_norm": 4.521510124206543,
    "prediction_norm": 4.6175689697265625,
    "timestamp": "2025-07-14T12:35:06.187168"
  },
  {
    "epoch": 2568,
    "batch_idx": 0,
    "loss": 0.34772583842277527,
    "learning_rate": 0.00018615431935465,
    "timestep_mean": 559.8125,
    "epsilon_norm": 8.2457275390625,
    "prediction_norm": 6.395474433898926,
    "timestamp": "2025-07-14T12:35:06.192525"
  },
  {
    "epoch": 2568,
    "batch_idx": 32,
    "loss": 0.33148494362831116,
    "learning_rate": 0.00018615431935465,
    "timestep_mean": 512.84375,
    "epsilon_norm": 8.006996154785156,
    "prediction_norm": 6.4728851318359375,
    "timestamp": "2025-07-14T12:35:06.197743"
  },
  {
    "epoch": 2568,
    "batch_idx": 64,
    "loss": 0.15990100800991058,
    "learning_rate": 0.00018615431935465,
    "timestep_mean": 531.4375,
    "epsilon_norm": 6.643442153930664,
    "prediction_norm": 6.109797477722168,
    "timestamp": "2025-07-14T12:35:06.202899"
  },
  {
    "epoch": 2568,
    "batch_idx": 96,
    "loss": 0.4051966369152069,
    "learning_rate": 0.00018615431935465,
    "timestep_mean": 544.9375,
    "epsilon_norm": 8.088171005249023,
    "prediction_norm": 6.98663330078125,
    "timestamp": "2025-07-14T12:35:06.208093"
  },
  {
    "epoch": 2568,
    "batch_idx": 128,
    "loss": 0.36146923899650574,
    "learning_rate": 0.00018615431935465,
    "timestep_mean": 483.9375,
    "epsilon_norm": 7.971320629119873,
    "prediction_norm": 6.409653186798096,
    "timestamp": "2025-07-14T12:35:06.213231"
  },
  {
    "epoch": 2568,
    "batch_idx": 160,
    "loss": 0.22964373230934143,
    "learning_rate": 0.00018615431935465,
    "timestep_mean": 629.59375,
    "epsilon_norm": 7.7002644538879395,
    "prediction_norm": 6.420594215393066,
    "timestamp": "2025-07-14T12:35:06.218272"
  },
  {
    "epoch": 2568,
    "batch_idx": 192,
    "loss": 0.131782665848732,
    "learning_rate": 0.00018615431935465,
    "timestep_mean": 410.875,
    "epsilon_norm": 4.427653789520264,
    "prediction_norm": 4.098321914672852,
    "timestamp": "2025-07-14T12:35:06.223448"
  },
  {
    "epoch": 2569,
    "batch_idx": 0,
    "loss": 0.1888141632080078,
    "learning_rate": 0.00018676610915665514,
    "timestep_mean": 396.25,
    "epsilon_norm": 7.108650207519531,
    "prediction_norm": 6.517757892608643,
    "timestamp": "2025-07-14T12:35:06.228671"
  },
  {
    "epoch": 2569,
    "batch_idx": 32,
    "loss": 0.33518195152282715,
    "learning_rate": 0.00018676610915665514,
    "timestep_mean": 555.46875,
    "epsilon_norm": 7.412801742553711,
    "prediction_norm": 6.669182777404785,
    "timestamp": "2025-07-14T12:35:06.233936"
  },
  {
    "epoch": 2569,
    "batch_idx": 64,
    "loss": 0.31082069873809814,
    "learning_rate": 0.00018676610915665514,
    "timestep_mean": 479.875,
    "epsilon_norm": 6.8783278465271,
    "prediction_norm": 5.560155391693115,
    "timestamp": "2025-07-14T12:35:06.239139"
  },
  {
    "epoch": 2569,
    "batch_idx": 96,
    "loss": 0.38870152831077576,
    "learning_rate": 0.00018676610915665514,
    "timestep_mean": 484.4375,
    "epsilon_norm": 8.03989315032959,
    "prediction_norm": 6.005763530731201,
    "timestamp": "2025-07-14T12:35:06.244346"
  },
  {
    "epoch": 2569,
    "batch_idx": 128,
    "loss": 0.36579710245132446,
    "learning_rate": 0.00018676610915665514,
    "timestep_mean": 495.28125,
    "epsilon_norm": 7.953193187713623,
    "prediction_norm": 6.501386642456055,
    "timestamp": "2025-07-14T12:35:06.249563"
  },
  {
    "epoch": 2569,
    "batch_idx": 160,
    "loss": 0.21385441720485687,
    "learning_rate": 0.00018676610915665514,
    "timestep_mean": 450.28125,
    "epsilon_norm": 7.639504909515381,
    "prediction_norm": 5.950444221496582,
    "timestamp": "2025-07-14T12:35:06.254866"
  },
  {
    "epoch": 2569,
    "batch_idx": 192,
    "loss": 0.663650393486023,
    "learning_rate": 0.00018676610915665514,
    "timestep_mean": 436.125,
    "epsilon_norm": 3.977809429168701,
    "prediction_norm": 2.815059185028076,
    "timestamp": "2025-07-14T12:35:06.259880"
  },
  {
    "epoch": 2570,
    "batch_idx": 0,
    "loss": 0.2878696024417877,
    "learning_rate": 0.00018737867183214792,
    "timestep_mean": 436.6875,
    "epsilon_norm": 6.934229850769043,
    "prediction_norm": 5.678905010223389,
    "timestamp": "2025-07-14T12:35:06.265025"
  },
  {
    "epoch": 2570,
    "batch_idx": 32,
    "loss": 0.2981324791908264,
    "learning_rate": 0.00018737867183214792,
    "timestep_mean": 477.25,
    "epsilon_norm": 7.913716793060303,
    "prediction_norm": 6.355945110321045,
    "timestamp": "2025-07-14T12:35:06.270209"
  },
  {
    "epoch": 2570,
    "batch_idx": 64,
    "loss": 0.1909932643175125,
    "learning_rate": 0.00018737867183214792,
    "timestep_mean": 420.28125,
    "epsilon_norm": 6.661271572113037,
    "prediction_norm": 5.8251543045043945,
    "timestamp": "2025-07-14T12:35:06.275526"
  },
  {
    "epoch": 2570,
    "batch_idx": 96,
    "loss": 0.42417168617248535,
    "learning_rate": 0.00018737867183214792,
    "timestep_mean": 480.40625,
    "epsilon_norm": 8.983739852905273,
    "prediction_norm": 7.625970840454102,
    "timestamp": "2025-07-14T12:35:06.280637"
  },
  {
    "epoch": 2570,
    "batch_idx": 128,
    "loss": 0.29112428426742554,
    "learning_rate": 0.00018737867183214792,
    "timestep_mean": 472.21875,
    "epsilon_norm": 7.483885288238525,
    "prediction_norm": 6.132586479187012,
    "timestamp": "2025-07-14T12:35:06.285834"
  },
  {
    "epoch": 2570,
    "batch_idx": 160,
    "loss": 0.13893212378025055,
    "learning_rate": 0.00018737867183214792,
    "timestep_mean": 568.03125,
    "epsilon_norm": 6.5783843994140625,
    "prediction_norm": 6.192499160766602,
    "timestamp": "2025-07-14T12:35:06.291023"
  },
  {
    "epoch": 2570,
    "batch_idx": 192,
    "loss": 0.07889007776975632,
    "learning_rate": 0.00018737867183214792,
    "timestep_mean": 618.125,
    "epsilon_norm": 2.920804500579834,
    "prediction_norm": 2.9425034523010254,
    "timestamp": "2025-07-14T12:35:06.296106"
  },
  {
    "epoch": 2571,
    "batch_idx": 0,
    "loss": 0.2764931321144104,
    "learning_rate": 0.00018799200586969118,
    "timestep_mean": 441.53125,
    "epsilon_norm": 8.745859146118164,
    "prediction_norm": 6.800570487976074,
    "timestamp": "2025-07-14T12:35:06.301307"
  },
  {
    "epoch": 2571,
    "batch_idx": 32,
    "loss": 0.4053186774253845,
    "learning_rate": 0.00018799200586969118,
    "timestep_mean": 504.375,
    "epsilon_norm": 8.390807151794434,
    "prediction_norm": 6.909934997558594,
    "timestamp": "2025-07-14T12:35:06.306454"
  },
  {
    "epoch": 2571,
    "batch_idx": 64,
    "loss": 0.3090426027774811,
    "learning_rate": 0.00018799200586969118,
    "timestep_mean": 460.09375,
    "epsilon_norm": 8.138412475585938,
    "prediction_norm": 6.807693958282471,
    "timestamp": "2025-07-14T12:35:06.311514"
  },
  {
    "epoch": 2571,
    "batch_idx": 96,
    "loss": 0.2650738060474396,
    "learning_rate": 0.00018799200586969118,
    "timestep_mean": 490.9375,
    "epsilon_norm": 8.113868713378906,
    "prediction_norm": 6.566582202911377,
    "timestamp": "2025-07-14T12:35:06.316663"
  },
  {
    "epoch": 2571,
    "batch_idx": 128,
    "loss": 0.20081494748592377,
    "learning_rate": 0.00018799200586969118,
    "timestep_mean": 514.03125,
    "epsilon_norm": 7.904755115509033,
    "prediction_norm": 6.918307781219482,
    "timestamp": "2025-07-14T12:35:06.321847"
  },
  {
    "epoch": 2571,
    "batch_idx": 160,
    "loss": 0.2538817226886749,
    "learning_rate": 0.00018799200586969118,
    "timestep_mean": 436.34375,
    "epsilon_norm": 8.941120147705078,
    "prediction_norm": 7.674804210662842,
    "timestamp": "2025-07-14T12:35:06.326797"
  },
  {
    "epoch": 2571,
    "batch_idx": 192,
    "loss": 0.10791787505149841,
    "learning_rate": 0.00018799200586969118,
    "timestep_mean": 692.125,
    "epsilon_norm": 4.481023788452148,
    "prediction_norm": 4.07545280456543,
    "timestamp": "2025-07-14T12:35:06.331927"
  },
  {
    "epoch": 2572,
    "batch_idx": 0,
    "loss": 0.4423588514328003,
    "learning_rate": 0.00018860610975594418,
    "timestep_mean": 508.6875,
    "epsilon_norm": 9.299206733703613,
    "prediction_norm": 7.795053958892822,
    "timestamp": "2025-07-14T12:35:06.337178"
  },
  {
    "epoch": 2572,
    "batch_idx": 32,
    "loss": 0.22403131425380707,
    "learning_rate": 0.00018860610975594418,
    "timestep_mean": 480.875,
    "epsilon_norm": 7.3988776206970215,
    "prediction_norm": 6.160722732543945,
    "timestamp": "2025-07-14T12:35:06.342222"
  },
  {
    "epoch": 2572,
    "batch_idx": 64,
    "loss": 0.3188033699989319,
    "learning_rate": 0.00018860610975594418,
    "timestep_mean": 498.46875,
    "epsilon_norm": 8.014349937438965,
    "prediction_norm": 6.582448482513428,
    "timestamp": "2025-07-14T12:35:06.347462"
  },
  {
    "epoch": 2572,
    "batch_idx": 96,
    "loss": 0.21163217723369598,
    "learning_rate": 0.00018860610975594418,
    "timestep_mean": 561.34375,
    "epsilon_norm": 8.338017463684082,
    "prediction_norm": 7.285716533660889,
    "timestamp": "2025-07-14T12:35:06.352682"
  },
  {
    "epoch": 2572,
    "batch_idx": 128,
    "loss": 0.3207318186759949,
    "learning_rate": 0.00018860610975594418,
    "timestep_mean": 495.75,
    "epsilon_norm": 8.39914321899414,
    "prediction_norm": 7.659329414367676,
    "timestamp": "2025-07-14T12:35:06.357757"
  },
  {
    "epoch": 2572,
    "batch_idx": 160,
    "loss": 0.1303357183933258,
    "learning_rate": 0.00018860610975594418,
    "timestep_mean": 556.25,
    "epsilon_norm": 8.007883071899414,
    "prediction_norm": 7.762277126312256,
    "timestamp": "2025-07-14T12:35:06.362980"
  },
  {
    "epoch": 2572,
    "batch_idx": 192,
    "loss": 0.09229152649641037,
    "learning_rate": 0.00018860610975594418,
    "timestep_mean": 548.875,
    "epsilon_norm": 3.5074851512908936,
    "prediction_norm": 3.5044877529144287,
    "timestamp": "2025-07-14T12:35:06.368161"
  },
  {
    "epoch": 2573,
    "batch_idx": 0,
    "loss": 0.39694029092788696,
    "learning_rate": 0.00018922098197566656,
    "timestep_mean": 465.625,
    "epsilon_norm": 7.6575398445129395,
    "prediction_norm": 6.573429107666016,
    "timestamp": "2025-07-14T12:35:06.373259"
  },
  {
    "epoch": 2573,
    "batch_idx": 32,
    "loss": 0.3435020446777344,
    "learning_rate": 0.00018922098197566656,
    "timestep_mean": 542.34375,
    "epsilon_norm": 9.44642162322998,
    "prediction_norm": 8.029034614562988,
    "timestamp": "2025-07-14T12:35:06.378542"
  },
  {
    "epoch": 2573,
    "batch_idx": 64,
    "loss": 0.26408272981643677,
    "learning_rate": 0.00018922098197566656,
    "timestep_mean": 416.09375,
    "epsilon_norm": 7.085300922393799,
    "prediction_norm": 5.678864479064941,
    "timestamp": "2025-07-14T12:35:06.383887"
  },
  {
    "epoch": 2573,
    "batch_idx": 96,
    "loss": 0.14968913793563843,
    "learning_rate": 0.00018922098197566656,
    "timestep_mean": 512.5625,
    "epsilon_norm": 7.807744979858398,
    "prediction_norm": 6.747130870819092,
    "timestamp": "2025-07-14T12:35:06.388976"
  },
  {
    "epoch": 2573,
    "batch_idx": 128,
    "loss": 0.4722438454627991,
    "learning_rate": 0.00018922098197566656,
    "timestep_mean": 607.34375,
    "epsilon_norm": 8.436287879943848,
    "prediction_norm": 6.933566570281982,
    "timestamp": "2025-07-14T12:35:06.394232"
  },
  {
    "epoch": 2573,
    "batch_idx": 160,
    "loss": 0.3337976932525635,
    "learning_rate": 0.00018922098197566656,
    "timestep_mean": 418.09375,
    "epsilon_norm": 6.980874061584473,
    "prediction_norm": 5.846076011657715,
    "timestamp": "2025-07-14T12:35:06.399441"
  },
  {
    "epoch": 2573,
    "batch_idx": 192,
    "loss": 0.15923789143562317,
    "learning_rate": 0.00018922098197566656,
    "timestep_mean": 588.75,
    "epsilon_norm": 3.082578659057617,
    "prediction_norm": 3.3945744037628174,
    "timestamp": "2025-07-14T12:35:06.404445"
  },
  {
    "epoch": 2574,
    "batch_idx": 0,
    "loss": 0.17569994926452637,
    "learning_rate": 0.00018983662101172233,
    "timestep_mean": 464.625,
    "epsilon_norm": 7.732640266418457,
    "prediction_norm": 7.05971097946167,
    "timestamp": "2025-07-14T12:35:06.409434"
  },
  {
    "epoch": 2574,
    "batch_idx": 32,
    "loss": 0.5928763747215271,
    "learning_rate": 0.00018983662101172233,
    "timestep_mean": 413.28125,
    "epsilon_norm": 8.558317184448242,
    "prediction_norm": 5.856447696685791,
    "timestamp": "2025-07-14T12:35:06.414354"
  },
  {
    "epoch": 2574,
    "batch_idx": 64,
    "loss": 0.2873155474662781,
    "learning_rate": 0.00018983662101172233,
    "timestep_mean": 487.25,
    "epsilon_norm": 8.77896785736084,
    "prediction_norm": 7.379340171813965,
    "timestamp": "2025-07-14T12:35:06.419558"
  },
  {
    "epoch": 2574,
    "batch_idx": 96,
    "loss": 0.2498590350151062,
    "learning_rate": 0.00018983662101172233,
    "timestep_mean": 504.375,
    "epsilon_norm": 6.576781749725342,
    "prediction_norm": 5.59212064743042,
    "timestamp": "2025-07-14T12:35:06.424754"
  },
  {
    "epoch": 2574,
    "batch_idx": 128,
    "loss": 0.4240015149116516,
    "learning_rate": 0.00018983662101172233,
    "timestep_mean": 466.34375,
    "epsilon_norm": 7.801471710205078,
    "prediction_norm": 6.120508670806885,
    "timestamp": "2025-07-14T12:35:06.430086"
  },
  {
    "epoch": 2574,
    "batch_idx": 160,
    "loss": 0.1935909390449524,
    "learning_rate": 0.00018983662101172233,
    "timestep_mean": 530.625,
    "epsilon_norm": 8.480291366577148,
    "prediction_norm": 7.766456127166748,
    "timestamp": "2025-07-14T12:35:06.435447"
  },
  {
    "epoch": 2574,
    "batch_idx": 192,
    "loss": 0.1606535017490387,
    "learning_rate": 0.00018983662101172233,
    "timestep_mean": 561.0,
    "epsilon_norm": 3.4852707386016846,
    "prediction_norm": 3.732156991958618,
    "timestamp": "2025-07-14T12:35:06.440667"
  },
  {
    "epoch": 2575,
    "batch_idx": 0,
    "loss": 0.2476467490196228,
    "learning_rate": 0.0001904530253450833,
    "timestep_mean": 536.625,
    "epsilon_norm": 7.765101432800293,
    "prediction_norm": 6.988835334777832,
    "timestamp": "2025-07-14T12:35:06.445864"
  },
  {
    "epoch": 2575,
    "batch_idx": 32,
    "loss": 0.27974191308021545,
    "learning_rate": 0.0001904530253450833,
    "timestep_mean": 427.5,
    "epsilon_norm": 9.090456008911133,
    "prediction_norm": 7.398133277893066,
    "timestamp": "2025-07-14T12:35:06.450963"
  },
  {
    "epoch": 2575,
    "batch_idx": 64,
    "loss": 0.273654580116272,
    "learning_rate": 0.0001904530253450833,
    "timestep_mean": 489.0625,
    "epsilon_norm": 7.179629802703857,
    "prediction_norm": 6.075196266174316,
    "timestamp": "2025-07-14T12:35:06.456213"
  },
  {
    "epoch": 2575,
    "batch_idx": 96,
    "loss": 0.28632593154907227,
    "learning_rate": 0.0001904530253450833,
    "timestep_mean": 531.25,
    "epsilon_norm": 8.138750076293945,
    "prediction_norm": 7.3130950927734375,
    "timestamp": "2025-07-14T12:35:06.461559"
  },
  {
    "epoch": 2575,
    "batch_idx": 128,
    "loss": 0.3272053301334381,
    "learning_rate": 0.0001904530253450833,
    "timestep_mean": 552.9375,
    "epsilon_norm": 6.881589412689209,
    "prediction_norm": 5.683774471282959,
    "timestamp": "2025-07-14T12:35:06.466602"
  },
  {
    "epoch": 2575,
    "batch_idx": 160,
    "loss": 0.22911439836025238,
    "learning_rate": 0.0001904530253450833,
    "timestep_mean": 501.53125,
    "epsilon_norm": 7.646364212036133,
    "prediction_norm": 7.294094562530518,
    "timestamp": "2025-07-14T12:35:06.471927"
  },
  {
    "epoch": 2575,
    "batch_idx": 192,
    "loss": 0.47603726387023926,
    "learning_rate": 0.0001904530253450833,
    "timestep_mean": 406.875,
    "epsilon_norm": 4.813592910766602,
    "prediction_norm": 3.2994754314422607,
    "timestamp": "2025-07-14T12:35:06.477132"
  },
  {
    "epoch": 2576,
    "batch_idx": 0,
    "loss": 0.30128321051597595,
    "learning_rate": 0.00019107019345483307,
    "timestep_mean": 527.8125,
    "epsilon_norm": 7.6406965255737305,
    "prediction_norm": 6.703238487243652,
    "timestamp": "2025-07-14T12:35:06.482241"
  },
  {
    "epoch": 2576,
    "batch_idx": 32,
    "loss": 0.18382450938224792,
    "learning_rate": 0.00019107019345483307,
    "timestep_mean": 508.90625,
    "epsilon_norm": 8.586665153503418,
    "prediction_norm": 7.1613078117370605,
    "timestamp": "2025-07-14T12:35:06.487689"
  },
  {
    "epoch": 2576,
    "batch_idx": 64,
    "loss": 0.18563884496688843,
    "learning_rate": 0.00019107019345483307,
    "timestep_mean": 488.875,
    "epsilon_norm": 7.621758460998535,
    "prediction_norm": 7.270987033843994,
    "timestamp": "2025-07-14T12:35:06.492904"
  },
  {
    "epoch": 2576,
    "batch_idx": 96,
    "loss": 0.22433990240097046,
    "learning_rate": 0.00019107019345483307,
    "timestep_mean": 499.46875,
    "epsilon_norm": 8.247946739196777,
    "prediction_norm": 6.987267017364502,
    "timestamp": "2025-07-14T12:35:06.498010"
  },
  {
    "epoch": 2576,
    "batch_idx": 128,
    "loss": 0.27715015411376953,
    "learning_rate": 0.00019107019345483307,
    "timestep_mean": 465.8125,
    "epsilon_norm": 8.167800903320312,
    "prediction_norm": 7.1045613288879395,
    "timestamp": "2025-07-14T12:35:06.503238"
  },
  {
    "epoch": 2576,
    "batch_idx": 160,
    "loss": 0.34257644414901733,
    "learning_rate": 0.00019107019345483307,
    "timestep_mean": 514.1875,
    "epsilon_norm": 7.202985763549805,
    "prediction_norm": 5.640503406524658,
    "timestamp": "2025-07-14T12:35:06.508705"
  },
  {
    "epoch": 2576,
    "batch_idx": 192,
    "loss": 0.25087815523147583,
    "learning_rate": 0.00019107019345483307,
    "timestep_mean": 467.75,
    "epsilon_norm": 3.684434413909912,
    "prediction_norm": 2.798426389694214,
    "timestamp": "2025-07-14T12:35:06.513994"
  },
  {
    "epoch": 2577,
    "batch_idx": 0,
    "loss": 0.1381157487630844,
    "learning_rate": 0.00019168812381817107,
    "timestep_mean": 505.75,
    "epsilon_norm": 8.80427074432373,
    "prediction_norm": 8.51962947845459,
    "timestamp": "2025-07-14T12:35:06.519291"
  },
  {
    "epoch": 2577,
    "batch_idx": 32,
    "loss": 0.11781814694404602,
    "learning_rate": 0.00019168812381817107,
    "timestep_mean": 598.75,
    "epsilon_norm": 7.232795238494873,
    "prediction_norm": 6.797180652618408,
    "timestamp": "2025-07-14T12:35:06.524617"
  },
  {
    "epoch": 2577,
    "batch_idx": 64,
    "loss": 0.372938871383667,
    "learning_rate": 0.00019168812381817107,
    "timestep_mean": 497.78125,
    "epsilon_norm": 7.738291263580322,
    "prediction_norm": 5.957042217254639,
    "timestamp": "2025-07-14T12:35:06.529661"
  },
  {
    "epoch": 2577,
    "batch_idx": 96,
    "loss": 0.28333261609077454,
    "learning_rate": 0.00019168812381817107,
    "timestep_mean": 419.84375,
    "epsilon_norm": 7.84243106842041,
    "prediction_norm": 6.123945236206055,
    "timestamp": "2025-07-14T12:35:06.534972"
  },
  {
    "epoch": 2577,
    "batch_idx": 128,
    "loss": 0.2990618348121643,
    "learning_rate": 0.00019168812381817107,
    "timestep_mean": 571.1875,
    "epsilon_norm": 8.438413619995117,
    "prediction_norm": 7.022679328918457,
    "timestamp": "2025-07-14T12:35:06.540126"
  },
  {
    "epoch": 2577,
    "batch_idx": 160,
    "loss": 0.3295322060585022,
    "learning_rate": 0.00019168812381817107,
    "timestep_mean": 406.25,
    "epsilon_norm": 7.7702317237854,
    "prediction_norm": 5.923547744750977,
    "timestamp": "2025-07-14T12:35:06.545401"
  },
  {
    "epoch": 2577,
    "batch_idx": 192,
    "loss": 0.28886356949806213,
    "learning_rate": 0.00019168812381817107,
    "timestep_mean": 459.875,
    "epsilon_norm": 3.3160641193389893,
    "prediction_norm": 2.1838128566741943,
    "timestamp": "2025-07-14T12:35:06.550604"
  },
  {
    "epoch": 2578,
    "batch_idx": 0,
    "loss": 0.20309442281723022,
    "learning_rate": 0.00019230681491041446,
    "timestep_mean": 510.75,
    "epsilon_norm": 6.9586076736450195,
    "prediction_norm": 6.428369522094727,
    "timestamp": "2025-07-14T12:35:06.555858"
  },
  {
    "epoch": 2578,
    "batch_idx": 32,
    "loss": 0.17248836159706116,
    "learning_rate": 0.00019230681491041446,
    "timestep_mean": 524.21875,
    "epsilon_norm": 7.861295223236084,
    "prediction_norm": 7.1241607666015625,
    "timestamp": "2025-07-14T12:35:06.561050"
  },
  {
    "epoch": 2578,
    "batch_idx": 64,
    "loss": 0.20258843898773193,
    "learning_rate": 0.00019230681491041446,
    "timestep_mean": 572.59375,
    "epsilon_norm": 8.061808586120605,
    "prediction_norm": 7.055146217346191,
    "timestamp": "2025-07-14T12:35:06.566240"
  },
  {
    "epoch": 2578,
    "batch_idx": 96,
    "loss": 0.4865497946739197,
    "learning_rate": 0.00019230681491041446,
    "timestep_mean": 412.5,
    "epsilon_norm": 8.262227058410645,
    "prediction_norm": 5.708662509918213,
    "timestamp": "2025-07-14T12:35:06.571391"
  },
  {
    "epoch": 2578,
    "batch_idx": 128,
    "loss": 0.405394583940506,
    "learning_rate": 0.00019230681491041446,
    "timestep_mean": 487.78125,
    "epsilon_norm": 7.9177165031433105,
    "prediction_norm": 5.778106212615967,
    "timestamp": "2025-07-14T12:35:06.576608"
  },
  {
    "epoch": 2578,
    "batch_idx": 160,
    "loss": 0.2545865476131439,
    "learning_rate": 0.00019230681491041446,
    "timestep_mean": 492.625,
    "epsilon_norm": 7.724233627319336,
    "prediction_norm": 6.396035194396973,
    "timestamp": "2025-07-14T12:35:06.581918"
  },
  {
    "epoch": 2578,
    "batch_idx": 192,
    "loss": 0.7563146352767944,
    "learning_rate": 0.00019230681491041446,
    "timestep_mean": 304.875,
    "epsilon_norm": 3.862825393676758,
    "prediction_norm": 2.8386058807373047,
    "timestamp": "2025-07-14T12:35:06.587028"
  },
  {
    "epoch": 2579,
    "batch_idx": 0,
    "loss": 0.16009196639060974,
    "learning_rate": 0.00019292626520500583,
    "timestep_mean": 504.59375,
    "epsilon_norm": 7.114129066467285,
    "prediction_norm": 5.9631853103637695,
    "timestamp": "2025-07-14T12:35:06.592468"
  },
  {
    "epoch": 2579,
    "batch_idx": 32,
    "loss": 0.21355567872524261,
    "learning_rate": 0.00019292626520500583,
    "timestep_mean": 439.78125,
    "epsilon_norm": 7.038283348083496,
    "prediction_norm": 5.920884609222412,
    "timestamp": "2025-07-14T12:35:06.597335"
  },
  {
    "epoch": 2579,
    "batch_idx": 64,
    "loss": 0.4232339560985565,
    "learning_rate": 0.00019292626520500583,
    "timestep_mean": 464.875,
    "epsilon_norm": 7.7347731590271,
    "prediction_norm": 6.183567523956299,
    "timestamp": "2025-07-14T12:35:06.602496"
  },
  {
    "epoch": 2579,
    "batch_idx": 96,
    "loss": 0.18611067533493042,
    "learning_rate": 0.00019292626520500583,
    "timestep_mean": 440.0625,
    "epsilon_norm": 8.773716926574707,
    "prediction_norm": 7.631134510040283,
    "timestamp": "2025-07-14T12:35:06.607764"
  },
  {
    "epoch": 2579,
    "batch_idx": 128,
    "loss": 0.16577133536338806,
    "learning_rate": 0.00019292626520500583,
    "timestep_mean": 447.5,
    "epsilon_norm": 6.343648433685303,
    "prediction_norm": 5.290283203125,
    "timestamp": "2025-07-14T12:35:06.612880"
  },
  {
    "epoch": 2579,
    "batch_idx": 160,
    "loss": 0.2641013264656067,
    "learning_rate": 0.00019292626520500583,
    "timestep_mean": 553.3125,
    "epsilon_norm": 7.757979869842529,
    "prediction_norm": 6.678797245025635,
    "timestamp": "2025-07-14T12:35:06.618023"
  },
  {
    "epoch": 2579,
    "batch_idx": 192,
    "loss": 0.15351566672325134,
    "learning_rate": 0.00019292626520500583,
    "timestep_mean": 559.875,
    "epsilon_norm": 2.9760220050811768,
    "prediction_norm": 2.391561985015869,
    "timestamp": "2025-07-14T12:35:06.623319"
  },
  {
    "epoch": 2580,
    "batch_idx": 0,
    "loss": 0.29166823625564575,
    "learning_rate": 0.0001935464731735122,
    "timestep_mean": 525.15625,
    "epsilon_norm": 6.869678020477295,
    "prediction_norm": 5.460216522216797,
    "timestamp": "2025-07-14T12:35:06.628643"
  },
  {
    "epoch": 2580,
    "batch_idx": 32,
    "loss": 0.3589041233062744,
    "learning_rate": 0.0001935464731735122,
    "timestep_mean": 590.78125,
    "epsilon_norm": 8.17570972442627,
    "prediction_norm": 6.5857834815979,
    "timestamp": "2025-07-14T12:35:06.633893"
  },
  {
    "epoch": 2580,
    "batch_idx": 64,
    "loss": 0.35374364256858826,
    "learning_rate": 0.0001935464731735122,
    "timestep_mean": 523.71875,
    "epsilon_norm": 8.077407836914062,
    "prediction_norm": 6.84276008605957,
    "timestamp": "2025-07-14T12:35:06.639079"
  },
  {
    "epoch": 2580,
    "batch_idx": 96,
    "loss": 0.27432578802108765,
    "learning_rate": 0.0001935464731735122,
    "timestep_mean": 479.84375,
    "epsilon_norm": 7.099719047546387,
    "prediction_norm": 5.594128608703613,
    "timestamp": "2025-07-14T12:35:06.644213"
  },
  {
    "epoch": 2580,
    "batch_idx": 128,
    "loss": 0.21999038755893707,
    "learning_rate": 0.0001935464731735122,
    "timestep_mean": 503.8125,
    "epsilon_norm": 7.970950126647949,
    "prediction_norm": 6.443778038024902,
    "timestamp": "2025-07-14T12:35:06.649310"
  },
  {
    "epoch": 2580,
    "batch_idx": 160,
    "loss": 0.16033485531806946,
    "learning_rate": 0.0001935464731735122,
    "timestep_mean": 580.3125,
    "epsilon_norm": 8.755464553833008,
    "prediction_norm": 7.561862945556641,
    "timestamp": "2025-07-14T12:35:06.654200"
  },
  {
    "epoch": 2580,
    "batch_idx": 192,
    "loss": 0.21060198545455933,
    "learning_rate": 0.0001935464731735122,
    "timestep_mean": 500.5,
    "epsilon_norm": 4.833399772644043,
    "prediction_norm": 4.503920078277588,
    "timestamp": "2025-07-14T12:35:06.659085"
  },
  {
    "epoch": 2581,
    "batch_idx": 0,
    "loss": 0.179364413022995,
    "learning_rate": 0.00019416743728563237,
    "timestep_mean": 517.40625,
    "epsilon_norm": 8.055930137634277,
    "prediction_norm": 7.054459095001221,
    "timestamp": "2025-07-14T12:35:06.664035"
  },
  {
    "epoch": 2581,
    "batch_idx": 32,
    "loss": 0.42242366075515747,
    "learning_rate": 0.00019416743728563237,
    "timestep_mean": 470.375,
    "epsilon_norm": 7.769189834594727,
    "prediction_norm": 6.160725116729736,
    "timestamp": "2025-07-14T12:35:06.669025"
  },
  {
    "epoch": 2581,
    "batch_idx": 64,
    "loss": 0.3268629312515259,
    "learning_rate": 0.00019416743728563237,
    "timestep_mean": 419.0625,
    "epsilon_norm": 9.702651977539062,
    "prediction_norm": 7.821994781494141,
    "timestamp": "2025-07-14T12:35:06.674138"
  },
  {
    "epoch": 2581,
    "batch_idx": 96,
    "loss": 0.17232745885849,
    "learning_rate": 0.00019416743728563237,
    "timestep_mean": 514.78125,
    "epsilon_norm": 6.29472541809082,
    "prediction_norm": 5.454887390136719,
    "timestamp": "2025-07-14T12:35:06.679185"
  },
  {
    "epoch": 2581,
    "batch_idx": 128,
    "loss": 0.2678380608558655,
    "learning_rate": 0.00019416743728563237,
    "timestep_mean": 491.40625,
    "epsilon_norm": 7.52333927154541,
    "prediction_norm": 5.816824913024902,
    "timestamp": "2025-07-14T12:35:06.684501"
  },
  {
    "epoch": 2581,
    "batch_idx": 160,
    "loss": 0.2965434789657593,
    "learning_rate": 0.00019416743728563237,
    "timestep_mean": 496.09375,
    "epsilon_norm": 8.173611640930176,
    "prediction_norm": 6.9216461181640625,
    "timestamp": "2025-07-14T12:35:06.689640"
  },
  {
    "epoch": 2581,
    "batch_idx": 192,
    "loss": 0.17145290970802307,
    "learning_rate": 0.00019416743728563237,
    "timestep_mean": 512.375,
    "epsilon_norm": 3.0415778160095215,
    "prediction_norm": 2.7947821617126465,
    "timestamp": "2025-07-14T12:35:06.694650"
  },
  {
    "epoch": 2582,
    "batch_idx": 0,
    "loss": 0.5672897100448608,
    "learning_rate": 0.00019478915600919914,
    "timestep_mean": 361.8125,
    "epsilon_norm": 8.454520225524902,
    "prediction_norm": 5.102663040161133,
    "timestamp": "2025-07-14T12:35:06.699803"
  },
  {
    "epoch": 2582,
    "batch_idx": 32,
    "loss": 0.2065381407737732,
    "learning_rate": 0.00019478915600919914,
    "timestep_mean": 575.90625,
    "epsilon_norm": 8.163771629333496,
    "prediction_norm": 7.019143104553223,
    "timestamp": "2025-07-14T12:35:06.705220"
  },
  {
    "epoch": 2582,
    "batch_idx": 64,
    "loss": 0.4446408152580261,
    "learning_rate": 0.00019478915600919914,
    "timestep_mean": 439.4375,
    "epsilon_norm": 8.570626258850098,
    "prediction_norm": 6.60829496383667,
    "timestamp": "2025-07-14T12:35:06.710364"
  },
  {
    "epoch": 2582,
    "batch_idx": 96,
    "loss": 0.16893580555915833,
    "learning_rate": 0.00019478915600919914,
    "timestep_mean": 502.6875,
    "epsilon_norm": 7.422472953796387,
    "prediction_norm": 6.493878364562988,
    "timestamp": "2025-07-14T12:35:06.715533"
  },
  {
    "epoch": 2582,
    "batch_idx": 128,
    "loss": 0.27223557233810425,
    "learning_rate": 0.00019478915600919914,
    "timestep_mean": 472.15625,
    "epsilon_norm": 7.316401481628418,
    "prediction_norm": 5.770605564117432,
    "timestamp": "2025-07-14T12:35:06.720780"
  },
  {
    "epoch": 2582,
    "batch_idx": 160,
    "loss": 0.3168754577636719,
    "learning_rate": 0.00019478915600919914,
    "timestep_mean": 513.125,
    "epsilon_norm": 8.577197074890137,
    "prediction_norm": 7.514344215393066,
    "timestamp": "2025-07-14T12:35:06.726004"
  },
  {
    "epoch": 2582,
    "batch_idx": 192,
    "loss": 0.10102791339159012,
    "learning_rate": 0.00019478915600919914,
    "timestep_mean": 555.0,
    "epsilon_norm": 2.8644678592681885,
    "prediction_norm": 3.0463907718658447,
    "timestamp": "2025-07-14T12:35:06.731287"
  },
  {
    "epoch": 2583,
    "batch_idx": 0,
    "loss": 0.23426921665668488,
    "learning_rate": 0.0001954116278101833,
    "timestep_mean": 550.6875,
    "epsilon_norm": 8.573476791381836,
    "prediction_norm": 7.543852806091309,
    "timestamp": "2025-07-14T12:35:06.736579"
  },
  {
    "epoch": 2583,
    "batch_idx": 32,
    "loss": 0.25376397371292114,
    "learning_rate": 0.0001954116278101833,
    "timestep_mean": 484.53125,
    "epsilon_norm": 8.443565368652344,
    "prediction_norm": 7.443979263305664,
    "timestamp": "2025-07-14T12:35:06.741739"
  },
  {
    "epoch": 2583,
    "batch_idx": 64,
    "loss": 0.42588287591934204,
    "learning_rate": 0.0001954116278101833,
    "timestep_mean": 495.34375,
    "epsilon_norm": 7.940649032592773,
    "prediction_norm": 6.171278953552246,
    "timestamp": "2025-07-14T12:35:06.747117"
  },
  {
    "epoch": 2583,
    "batch_idx": 96,
    "loss": 0.1788286566734314,
    "learning_rate": 0.0001954116278101833,
    "timestep_mean": 572.71875,
    "epsilon_norm": 7.465880393981934,
    "prediction_norm": 6.415046215057373,
    "timestamp": "2025-07-14T12:35:06.752379"
  },
  {
    "epoch": 2583,
    "batch_idx": 128,
    "loss": 0.20763926208019257,
    "learning_rate": 0.0001954116278101833,
    "timestep_mean": 535.5,
    "epsilon_norm": 7.9696831703186035,
    "prediction_norm": 7.04575252532959,
    "timestamp": "2025-07-14T12:35:06.757574"
  },
  {
    "epoch": 2583,
    "batch_idx": 160,
    "loss": 0.2691458463668823,
    "learning_rate": 0.0001954116278101833,
    "timestep_mean": 546.5,
    "epsilon_norm": 7.438324928283691,
    "prediction_norm": 6.570666790008545,
    "timestamp": "2025-07-14T12:35:06.762845"
  },
  {
    "epoch": 2583,
    "batch_idx": 192,
    "loss": 0.06817881017923355,
    "learning_rate": 0.0001954116278101833,
    "timestep_mean": 579.125,
    "epsilon_norm": 3.5237038135528564,
    "prediction_norm": 3.366157293319702,
    "timestamp": "2025-07-14T12:35:06.767985"
  },
  {
    "epoch": 2584,
    "batch_idx": 0,
    "loss": 0.1491478681564331,
    "learning_rate": 0.0001960348511526976,
    "timestep_mean": 580.65625,
    "epsilon_norm": 7.656669616699219,
    "prediction_norm": 7.17155122756958,
    "timestamp": "2025-07-14T12:35:06.772889"
  },
  {
    "epoch": 2584,
    "batch_idx": 32,
    "loss": 0.20960386097431183,
    "learning_rate": 0.0001960348511526976,
    "timestep_mean": 491.25,
    "epsilon_norm": 7.175987720489502,
    "prediction_norm": 6.080323219299316,
    "timestamp": "2025-07-14T12:35:06.777981"
  },
  {
    "epoch": 2584,
    "batch_idx": 64,
    "loss": 0.4101221263408661,
    "learning_rate": 0.0001960348511526976,
    "timestep_mean": 525.8125,
    "epsilon_norm": 9.008152961730957,
    "prediction_norm": 7.683704376220703,
    "timestamp": "2025-07-14T12:35:06.783189"
  },
  {
    "epoch": 2584,
    "batch_idx": 96,
    "loss": 0.34910857677459717,
    "learning_rate": 0.0001960348511526976,
    "timestep_mean": 517.03125,
    "epsilon_norm": 9.411202430725098,
    "prediction_norm": 7.713069915771484,
    "timestamp": "2025-07-14T12:35:06.788531"
  },
  {
    "epoch": 2584,
    "batch_idx": 128,
    "loss": 0.13178834319114685,
    "learning_rate": 0.0001960348511526976,
    "timestep_mean": 560.4375,
    "epsilon_norm": 8.319625854492188,
    "prediction_norm": 7.751018524169922,
    "timestamp": "2025-07-14T12:35:06.793688"
  },
  {
    "epoch": 2584,
    "batch_idx": 160,
    "loss": 0.15004831552505493,
    "learning_rate": 0.0001960348511526976,
    "timestep_mean": 517.15625,
    "epsilon_norm": 7.363935470581055,
    "prediction_norm": 6.535730838775635,
    "timestamp": "2025-07-14T12:35:06.798848"
  },
  {
    "epoch": 2584,
    "batch_idx": 192,
    "loss": 0.03865250200033188,
    "learning_rate": 0.0001960348511526976,
    "timestep_mean": 648.125,
    "epsilon_norm": 3.9715042114257812,
    "prediction_norm": 3.9983150959014893,
    "timestamp": "2025-07-14T12:35:06.803959"
  },
  {
    "epoch": 2585,
    "batch_idx": 0,
    "loss": 0.32463160157203674,
    "learning_rate": 0.00019665882449900043,
    "timestep_mean": 534.8125,
    "epsilon_norm": 7.810574054718018,
    "prediction_norm": 6.245743274688721,
    "timestamp": "2025-07-14T12:35:06.809162"
  },
  {
    "epoch": 2585,
    "batch_idx": 32,
    "loss": 0.31529268622398376,
    "learning_rate": 0.00019665882449900043,
    "timestep_mean": 484.59375,
    "epsilon_norm": 7.137917518615723,
    "prediction_norm": 5.587958335876465,
    "timestamp": "2025-07-14T12:35:06.814301"
  },
  {
    "epoch": 2585,
    "batch_idx": 64,
    "loss": 0.17688708007335663,
    "learning_rate": 0.00019665882449900043,
    "timestep_mean": 480.34375,
    "epsilon_norm": 7.757621765136719,
    "prediction_norm": 6.319023132324219,
    "timestamp": "2025-07-14T12:35:06.819447"
  },
  {
    "epoch": 2585,
    "batch_idx": 96,
    "loss": 0.2957127094268799,
    "learning_rate": 0.00019665882449900043,
    "timestep_mean": 474.625,
    "epsilon_norm": 7.38686466217041,
    "prediction_norm": 5.935890197753906,
    "timestamp": "2025-07-14T12:35:06.824580"
  },
  {
    "epoch": 2585,
    "batch_idx": 128,
    "loss": 0.18924687802791595,
    "learning_rate": 0.00019665882449900043,
    "timestep_mean": 605.53125,
    "epsilon_norm": 7.702327728271484,
    "prediction_norm": 6.851327896118164,
    "timestamp": "2025-07-14T12:35:06.829552"
  },
  {
    "epoch": 2585,
    "batch_idx": 160,
    "loss": 0.4083409905433655,
    "learning_rate": 0.00019665882449900043,
    "timestep_mean": 493.84375,
    "epsilon_norm": 8.20853042602539,
    "prediction_norm": 6.190164089202881,
    "timestamp": "2025-07-14T12:35:06.834477"
  },
  {
    "epoch": 2585,
    "batch_idx": 192,
    "loss": 0.3469931483268738,
    "learning_rate": 0.00019665882449900043,
    "timestep_mean": 435.625,
    "epsilon_norm": 4.1796159744262695,
    "prediction_norm": 3.1600050926208496,
    "timestamp": "2025-07-14T12:35:06.839354"
  },
  {
    "epoch": 2586,
    "batch_idx": 0,
    "loss": 0.361774206161499,
    "learning_rate": 0.00019728354630949995,
    "timestep_mean": 451.625,
    "epsilon_norm": 7.722514629364014,
    "prediction_norm": 6.09512996673584,
    "timestamp": "2025-07-14T12:35:06.844257"
  },
  {
    "epoch": 2586,
    "batch_idx": 32,
    "loss": 0.3116561472415924,
    "learning_rate": 0.00019728354630949995,
    "timestep_mean": 454.25,
    "epsilon_norm": 7.709969520568848,
    "prediction_norm": 5.957077980041504,
    "timestamp": "2025-07-14T12:35:06.849118"
  },
  {
    "epoch": 2586,
    "batch_idx": 64,
    "loss": 0.23218272626399994,
    "learning_rate": 0.00019728354630949995,
    "timestep_mean": 476.625,
    "epsilon_norm": 8.515449523925781,
    "prediction_norm": 6.77662992477417,
    "timestamp": "2025-07-14T12:35:06.854268"
  },
  {
    "epoch": 2586,
    "batch_idx": 96,
    "loss": 0.32227200269699097,
    "learning_rate": 0.00019728354630949995,
    "timestep_mean": 537.40625,
    "epsilon_norm": 8.292439460754395,
    "prediction_norm": 6.869560718536377,
    "timestamp": "2025-07-14T12:35:06.859446"
  },
  {
    "epoch": 2586,
    "batch_idx": 128,
    "loss": 0.2217300683259964,
    "learning_rate": 0.00019728354630949995,
    "timestep_mean": 544.59375,
    "epsilon_norm": 7.289139747619629,
    "prediction_norm": 6.946004867553711,
    "timestamp": "2025-07-14T12:35:06.864561"
  },
  {
    "epoch": 2586,
    "batch_idx": 160,
    "loss": 0.27529895305633545,
    "learning_rate": 0.00019728354630949995,
    "timestep_mean": 511.71875,
    "epsilon_norm": 7.23256778717041,
    "prediction_norm": 6.041696071624756,
    "timestamp": "2025-07-14T12:35:06.869767"
  },
  {
    "epoch": 2586,
    "batch_idx": 192,
    "loss": 0.1719910055398941,
    "learning_rate": 0.00019728354630949995,
    "timestep_mean": 592.75,
    "epsilon_norm": 4.491568088531494,
    "prediction_norm": 4.08235502243042,
    "timestamp": "2025-07-14T12:35:06.874906"
  },
  {
    "epoch": 2587,
    "batch_idx": 0,
    "loss": 0.2639709711074829,
    "learning_rate": 0.000197909015042756,
    "timestep_mean": 484.09375,
    "epsilon_norm": 8.357023239135742,
    "prediction_norm": 7.32483434677124,
    "timestamp": "2025-07-14T12:35:06.880133"
  },
  {
    "epoch": 2587,
    "batch_idx": 32,
    "loss": 0.12041729688644409,
    "learning_rate": 0.000197909015042756,
    "timestep_mean": 589.0625,
    "epsilon_norm": 8.236588478088379,
    "prediction_norm": 7.418087005615234,
    "timestamp": "2025-07-14T12:35:06.885208"
  },
  {
    "epoch": 2587,
    "batch_idx": 64,
    "loss": 0.30627042055130005,
    "learning_rate": 0.000197909015042756,
    "timestep_mean": 559.46875,
    "epsilon_norm": 7.847442150115967,
    "prediction_norm": 7.291009902954102,
    "timestamp": "2025-07-14T12:35:06.890401"
  },
  {
    "epoch": 2587,
    "batch_idx": 96,
    "loss": 0.19187447428703308,
    "learning_rate": 0.000197909015042756,
    "timestep_mean": 515.03125,
    "epsilon_norm": 8.488249778747559,
    "prediction_norm": 7.523502826690674,
    "timestamp": "2025-07-14T12:35:06.895682"
  },
  {
    "epoch": 2587,
    "batch_idx": 128,
    "loss": 0.2797015309333801,
    "learning_rate": 0.000197909015042756,
    "timestep_mean": 484.625,
    "epsilon_norm": 8.984625816345215,
    "prediction_norm": 7.716957092285156,
    "timestamp": "2025-07-14T12:35:06.901025"
  },
  {
    "epoch": 2587,
    "batch_idx": 160,
    "loss": 0.1294359266757965,
    "learning_rate": 0.000197909015042756,
    "timestep_mean": 546.59375,
    "epsilon_norm": 7.769992351531982,
    "prediction_norm": 7.000330924987793,
    "timestamp": "2025-07-14T12:35:06.906275"
  },
  {
    "epoch": 2587,
    "batch_idx": 192,
    "loss": 0.0986979678273201,
    "learning_rate": 0.000197909015042756,
    "timestep_mean": 451.375,
    "epsilon_norm": 2.655683994293213,
    "prediction_norm": 2.435255289077759,
    "timestamp": "2025-07-14T12:35:06.911303"
  },
  {
    "epoch": 2588,
    "batch_idx": 0,
    "loss": 0.26234012842178345,
    "learning_rate": 0.0001985352291554882,
    "timestep_mean": 512.0625,
    "epsilon_norm": 7.592109680175781,
    "prediction_norm": 7.105137348175049,
    "timestamp": "2025-07-14T12:35:06.916401"
  },
  {
    "epoch": 2588,
    "batch_idx": 32,
    "loss": 0.19268761575222015,
    "learning_rate": 0.0001985352291554882,
    "timestep_mean": 450.15625,
    "epsilon_norm": 6.563536167144775,
    "prediction_norm": 5.382727146148682,
    "timestamp": "2025-07-14T12:35:06.921692"
  },
  {
    "epoch": 2588,
    "batch_idx": 64,
    "loss": 0.22070205211639404,
    "learning_rate": 0.0001985352291554882,
    "timestep_mean": 461.71875,
    "epsilon_norm": 7.125497817993164,
    "prediction_norm": 6.581768035888672,
    "timestamp": "2025-07-14T12:35:06.926818"
  },
  {
    "epoch": 2588,
    "batch_idx": 96,
    "loss": 0.09454295039176941,
    "learning_rate": 0.0001985352291554882,
    "timestep_mean": 509.40625,
    "epsilon_norm": 7.31599235534668,
    "prediction_norm": 7.368825435638428,
    "timestamp": "2025-07-14T12:35:06.931946"
  },
  {
    "epoch": 2588,
    "batch_idx": 128,
    "loss": 0.1305837780237198,
    "learning_rate": 0.0001985352291554882,
    "timestep_mean": 570.0625,
    "epsilon_norm": 7.027165412902832,
    "prediction_norm": 6.592820644378662,
    "timestamp": "2025-07-14T12:35:06.937062"
  },
  {
    "epoch": 2588,
    "batch_idx": 160,
    "loss": 0.2679690718650818,
    "learning_rate": 0.0001985352291554882,
    "timestep_mean": 500.09375,
    "epsilon_norm": 8.2637357711792,
    "prediction_norm": 7.056300163269043,
    "timestamp": "2025-07-14T12:35:06.942357"
  },
  {
    "epoch": 2588,
    "batch_idx": 192,
    "loss": 0.3361734449863434,
    "learning_rate": 0.0001985352291554882,
    "timestep_mean": 474.875,
    "epsilon_norm": 4.121449947357178,
    "prediction_norm": 3.124727249145508,
    "timestamp": "2025-07-14T12:35:06.947457"
  },
  {
    "epoch": 2589,
    "batch_idx": 0,
    "loss": 0.22999709844589233,
    "learning_rate": 0.00019916218710257436,
    "timestep_mean": 576.25,
    "epsilon_norm": 7.704235076904297,
    "prediction_norm": 7.055056095123291,
    "timestamp": "2025-07-14T12:35:06.952710"
  },
  {
    "epoch": 2589,
    "batch_idx": 32,
    "loss": 0.26178014278411865,
    "learning_rate": 0.00019916218710257436,
    "timestep_mean": 513.75,
    "epsilon_norm": 8.524575233459473,
    "prediction_norm": 6.912668228149414,
    "timestamp": "2025-07-14T12:35:06.957894"
  },
  {
    "epoch": 2589,
    "batch_idx": 64,
    "loss": 0.3468984365463257,
    "learning_rate": 0.00019916218710257436,
    "timestep_mean": 514.53125,
    "epsilon_norm": 9.550189018249512,
    "prediction_norm": 7.899431228637695,
    "timestamp": "2025-07-14T12:35:06.963092"
  },
  {
    "epoch": 2589,
    "batch_idx": 96,
    "loss": 0.1742125153541565,
    "learning_rate": 0.00019916218710257436,
    "timestep_mean": 524.625,
    "epsilon_norm": 8.407859802246094,
    "prediction_norm": 7.697617053985596,
    "timestamp": "2025-07-14T12:35:06.968323"
  },
  {
    "epoch": 2589,
    "batch_idx": 128,
    "loss": 0.12439453601837158,
    "learning_rate": 0.00019916218710257436,
    "timestep_mean": 480.5,
    "epsilon_norm": 6.880099773406982,
    "prediction_norm": 6.707657337188721,
    "timestamp": "2025-07-14T12:35:06.973532"
  },
  {
    "epoch": 2589,
    "batch_idx": 160,
    "loss": 0.22338402271270752,
    "learning_rate": 0.00019916218710257436,
    "timestep_mean": 520.15625,
    "epsilon_norm": 8.248078346252441,
    "prediction_norm": 6.936987400054932,
    "timestamp": "2025-07-14T12:35:06.978735"
  },
  {
    "epoch": 2589,
    "batch_idx": 192,
    "loss": 0.47031569480895996,
    "learning_rate": 0.00019916218710257436,
    "timestep_mean": 390.0,
    "epsilon_norm": 4.269523620605469,
    "prediction_norm": 2.8919568061828613,
    "timestamp": "2025-07-14T12:35:06.983953"
  },
  {
    "epoch": 2590,
    "batch_idx": 0,
    "loss": 0.23741555213928223,
    "learning_rate": 0.0001997898873370584,
    "timestep_mean": 543.34375,
    "epsilon_norm": 9.51647663116455,
    "prediction_norm": 8.26734733581543,
    "timestamp": "2025-07-14T12:35:06.989137"
  },
  {
    "epoch": 2590,
    "batch_idx": 32,
    "loss": 0.3768511414527893,
    "learning_rate": 0.0001997898873370584,
    "timestep_mean": 473.9375,
    "epsilon_norm": 8.875955581665039,
    "prediction_norm": 7.434665679931641,
    "timestamp": "2025-07-14T12:35:06.994425"
  },
  {
    "epoch": 2590,
    "batch_idx": 64,
    "loss": 0.24345946311950684,
    "learning_rate": 0.0001997898873370584,
    "timestep_mean": 527.78125,
    "epsilon_norm": 8.579848289489746,
    "prediction_norm": 7.632312774658203,
    "timestamp": "2025-07-14T12:35:06.999665"
  },
  {
    "epoch": 2590,
    "batch_idx": 96,
    "loss": 0.3297842741012573,
    "learning_rate": 0.0001997898873370584,
    "timestep_mean": 521.21875,
    "epsilon_norm": 8.266241073608398,
    "prediction_norm": 7.450596809387207,
    "timestamp": "2025-07-14T12:35:07.005024"
  },
  {
    "epoch": 2590,
    "batch_idx": 128,
    "loss": 0.4131711721420288,
    "learning_rate": 0.0001997898873370584,
    "timestep_mean": 527.375,
    "epsilon_norm": 9.3516263961792,
    "prediction_norm": 7.812020301818848,
    "timestamp": "2025-07-14T12:35:07.010381"
  },
  {
    "epoch": 2590,
    "batch_idx": 160,
    "loss": 0.1954975128173828,
    "learning_rate": 0.0001997898873370584,
    "timestep_mean": 478.375,
    "epsilon_norm": 7.572806358337402,
    "prediction_norm": 7.095065116882324,
    "timestamp": "2025-07-14T12:35:07.015640"
  },
  {
    "epoch": 2590,
    "batch_idx": 192,
    "loss": 0.13431388139724731,
    "learning_rate": 0.0001997898873370584,
    "timestep_mean": 439.125,
    "epsilon_norm": 3.3256731033325195,
    "prediction_norm": 3.4586446285247803,
    "timestamp": "2025-07-14T12:35:07.020750"
  },
  {
    "epoch": 2591,
    "batch_idx": 0,
    "loss": 0.26051610708236694,
    "learning_rate": 0.0002004183283101525,
    "timestep_mean": 496.875,
    "epsilon_norm": 6.4499077796936035,
    "prediction_norm": 5.754740238189697,
    "timestamp": "2025-07-14T12:35:07.025894"
  },
  {
    "epoch": 2591,
    "batch_idx": 32,
    "loss": 0.2431432008743286,
    "learning_rate": 0.0002004183283101525,
    "timestep_mean": 511.1875,
    "epsilon_norm": 8.248729705810547,
    "prediction_norm": 7.237701892852783,
    "timestamp": "2025-07-14T12:35:07.031039"
  },
  {
    "epoch": 2591,
    "batch_idx": 64,
    "loss": 0.3125978112220764,
    "learning_rate": 0.0002004183283101525,
    "timestep_mean": 466.46875,
    "epsilon_norm": 8.213200569152832,
    "prediction_norm": 6.974714279174805,
    "timestamp": "2025-07-14T12:35:07.036240"
  },
  {
    "epoch": 2591,
    "batch_idx": 96,
    "loss": 0.4763680696487427,
    "learning_rate": 0.0002004183283101525,
    "timestep_mean": 390.125,
    "epsilon_norm": 8.227800369262695,
    "prediction_norm": 6.028008460998535,
    "timestamp": "2025-07-14T12:35:07.041494"
  },
  {
    "epoch": 2591,
    "batch_idx": 128,
    "loss": 0.17117324471473694,
    "learning_rate": 0.0002004183283101525,
    "timestep_mean": 497.3125,
    "epsilon_norm": 6.79219388961792,
    "prediction_norm": 5.885694980621338,
    "timestamp": "2025-07-14T12:35:07.046817"
  },
  {
    "epoch": 2591,
    "batch_idx": 160,
    "loss": 0.20048832893371582,
    "learning_rate": 0.0002004183283101525,
    "timestep_mean": 481.84375,
    "epsilon_norm": 7.362441062927246,
    "prediction_norm": 6.413002967834473,
    "timestamp": "2025-07-14T12:35:07.051959"
  },
  {
    "epoch": 2591,
    "batch_idx": 192,
    "loss": 0.1639111191034317,
    "learning_rate": 0.0002004183283101525,
    "timestep_mean": 528.0,
    "epsilon_norm": 3.8618245124816895,
    "prediction_norm": 3.701887369155884,
    "timestamp": "2025-07-14T12:35:07.057220"
  },
  {
    "epoch": 2592,
    "batch_idx": 0,
    "loss": 0.2080419510602951,
    "learning_rate": 0.00020104750847124093,
    "timestep_mean": 489.75,
    "epsilon_norm": 7.48566198348999,
    "prediction_norm": 6.546515941619873,
    "timestamp": "2025-07-14T12:35:07.062416"
  },
  {
    "epoch": 2592,
    "batch_idx": 32,
    "loss": 0.13862556219100952,
    "learning_rate": 0.00020104750847124093,
    "timestep_mean": 524.5625,
    "epsilon_norm": 8.53866958618164,
    "prediction_norm": 8.142117500305176,
    "timestamp": "2025-07-14T12:35:07.067609"
  },
  {
    "epoch": 2592,
    "batch_idx": 64,
    "loss": 0.3292756974697113,
    "learning_rate": 0.00020104750847124093,
    "timestep_mean": 491.90625,
    "epsilon_norm": 8.435158729553223,
    "prediction_norm": 7.2031049728393555,
    "timestamp": "2025-07-14T12:35:07.072831"
  },
  {
    "epoch": 2592,
    "batch_idx": 96,
    "loss": 0.26985403895378113,
    "learning_rate": 0.00020104750847124093,
    "timestep_mean": 407.1875,
    "epsilon_norm": 7.9647040367126465,
    "prediction_norm": 6.954792499542236,
    "timestamp": "2025-07-14T12:35:07.078035"
  },
  {
    "epoch": 2592,
    "batch_idx": 128,
    "loss": 0.12845690548419952,
    "learning_rate": 0.00020104750847124093,
    "timestep_mean": 535.625,
    "epsilon_norm": 7.971428394317627,
    "prediction_norm": 7.214890003204346,
    "timestamp": "2025-07-14T12:35:07.083314"
  },
  {
    "epoch": 2592,
    "batch_idx": 160,
    "loss": 0.2510393559932709,
    "learning_rate": 0.00020104750847124093,
    "timestep_mean": 480.46875,
    "epsilon_norm": 8.416828155517578,
    "prediction_norm": 6.926393032073975,
    "timestamp": "2025-07-14T12:35:07.088520"
  },
  {
    "epoch": 2592,
    "batch_idx": 192,
    "loss": 0.264635294675827,
    "learning_rate": 0.00020104750847124093,
    "timestep_mean": 362.875,
    "epsilon_norm": 5.000491619110107,
    "prediction_norm": 4.189856052398682,
    "timestamp": "2025-07-14T12:35:07.093693"
  },
  {
    "epoch": 2593,
    "batch_idx": 0,
    "loss": 0.22333267331123352,
    "learning_rate": 0.0002016774262678842,
    "timestep_mean": 575.8125,
    "epsilon_norm": 7.908036231994629,
    "prediction_norm": 6.310751438140869,
    "timestamp": "2025-07-14T12:35:07.098935"
  },
  {
    "epoch": 2593,
    "batch_idx": 32,
    "loss": 0.2920975089073181,
    "learning_rate": 0.0002016774262678842,
    "timestep_mean": 473.78125,
    "epsilon_norm": 8.147306442260742,
    "prediction_norm": 6.93400239944458,
    "timestamp": "2025-07-14T12:35:07.103941"
  },
  {
    "epoch": 2593,
    "batch_idx": 64,
    "loss": 0.24906811118125916,
    "learning_rate": 0.0002016774262678842,
    "timestep_mean": 529.03125,
    "epsilon_norm": 8.763339042663574,
    "prediction_norm": 7.361141204833984,
    "timestamp": "2025-07-14T12:35:07.109194"
  },
  {
    "epoch": 2593,
    "batch_idx": 96,
    "loss": 0.2844417691230774,
    "learning_rate": 0.0002016774262678842,
    "timestep_mean": 573.5,
    "epsilon_norm": 8.28236198425293,
    "prediction_norm": 7.091528415679932,
    "timestamp": "2025-07-14T12:35:07.114440"
  },
  {
    "epoch": 2593,
    "batch_idx": 128,
    "loss": 0.1777457296848297,
    "learning_rate": 0.0002016774262678842,
    "timestep_mean": 494.53125,
    "epsilon_norm": 9.034770011901855,
    "prediction_norm": 7.845890045166016,
    "timestamp": "2025-07-14T12:35:07.119561"
  },
  {
    "epoch": 2593,
    "batch_idx": 160,
    "loss": 0.12371668219566345,
    "learning_rate": 0.0002016774262678842,
    "timestep_mean": 601.0625,
    "epsilon_norm": 6.968589782714844,
    "prediction_norm": 6.458606243133545,
    "timestamp": "2025-07-14T12:35:07.124681"
  },
  {
    "epoch": 2593,
    "batch_idx": 192,
    "loss": 0.7737966775894165,
    "learning_rate": 0.0002016774262678842,
    "timestep_mean": 470.5,
    "epsilon_norm": 4.505627155303955,
    "prediction_norm": 2.3669896125793457,
    "timestamp": "2025-07-14T12:35:07.129840"
  },
  {
    "epoch": 2594,
    "batch_idx": 0,
    "loss": 0.3508197069168091,
    "learning_rate": 0.00020230808014582314,
    "timestep_mean": 527.8125,
    "epsilon_norm": 8.45905590057373,
    "prediction_norm": 6.696197986602783,
    "timestamp": "2025-07-14T12:35:07.134966"
  },
  {
    "epoch": 2594,
    "batch_idx": 32,
    "loss": 0.3365575671195984,
    "learning_rate": 0.00020230808014582314,
    "timestep_mean": 556.15625,
    "epsilon_norm": 8.476825714111328,
    "prediction_norm": 7.083218574523926,
    "timestamp": "2025-07-14T12:35:07.140157"
  },
  {
    "epoch": 2594,
    "batch_idx": 64,
    "loss": 0.23205676674842834,
    "learning_rate": 0.00020230808014582314,
    "timestep_mean": 578.15625,
    "epsilon_norm": 7.274454116821289,
    "prediction_norm": 6.464146614074707,
    "timestamp": "2025-07-14T12:35:07.145386"
  },
  {
    "epoch": 2594,
    "batch_idx": 96,
    "loss": 0.27726858854293823,
    "learning_rate": 0.00020230808014582314,
    "timestep_mean": 413.8125,
    "epsilon_norm": 8.675945281982422,
    "prediction_norm": 7.132672309875488,
    "timestamp": "2025-07-14T12:35:07.150797"
  },
  {
    "epoch": 2594,
    "batch_idx": 128,
    "loss": 0.34606271982192993,
    "learning_rate": 0.00020230808014582314,
    "timestep_mean": 507.28125,
    "epsilon_norm": 7.5617451667785645,
    "prediction_norm": 5.773067951202393,
    "timestamp": "2025-07-14T12:35:07.156012"
  },
  {
    "epoch": 2594,
    "batch_idx": 160,
    "loss": 0.3375663459300995,
    "learning_rate": 0.00020230808014582314,
    "timestep_mean": 465.59375,
    "epsilon_norm": 8.559378623962402,
    "prediction_norm": 6.891452789306641,
    "timestamp": "2025-07-14T12:35:07.161242"
  },
  {
    "epoch": 2594,
    "batch_idx": 192,
    "loss": 0.1987847536802292,
    "learning_rate": 0.00020230808014582314,
    "timestep_mean": 537.625,
    "epsilon_norm": 4.095525741577148,
    "prediction_norm": 3.569164276123047,
    "timestamp": "2025-07-14T12:35:07.166419"
  },
  {
    "epoch": 2595,
    "batch_idx": 0,
    "loss": 0.1605444848537445,
    "learning_rate": 0.000202939468548981,
    "timestep_mean": 525.0,
    "epsilon_norm": 7.818703651428223,
    "prediction_norm": 7.314304351806641,
    "timestamp": "2025-07-14T12:35:07.171642"
  },
  {
    "epoch": 2595,
    "batch_idx": 32,
    "loss": 0.28434985876083374,
    "learning_rate": 0.000202939468548981,
    "timestep_mean": 533.84375,
    "epsilon_norm": 9.190733909606934,
    "prediction_norm": 7.753133773803711,
    "timestamp": "2025-07-14T12:35:07.176682"
  },
  {
    "epoch": 2595,
    "batch_idx": 64,
    "loss": 0.21004128456115723,
    "learning_rate": 0.000202939468548981,
    "timestep_mean": 525.46875,
    "epsilon_norm": 7.564507007598877,
    "prediction_norm": 6.231045246124268,
    "timestamp": "2025-07-14T12:35:07.181866"
  },
  {
    "epoch": 2595,
    "batch_idx": 96,
    "loss": 0.18660897016525269,
    "learning_rate": 0.000202939468548981,
    "timestep_mean": 528.40625,
    "epsilon_norm": 8.763903617858887,
    "prediction_norm": 8.304924964904785,
    "timestamp": "2025-07-14T12:35:07.186989"
  },
  {
    "epoch": 2595,
    "batch_idx": 128,
    "loss": 0.26693135499954224,
    "learning_rate": 0.000202939468548981,
    "timestep_mean": 404.28125,
    "epsilon_norm": 8.831907272338867,
    "prediction_norm": 7.548715114593506,
    "timestamp": "2025-07-14T12:35:07.192263"
  },
  {
    "epoch": 2595,
    "batch_idx": 160,
    "loss": 0.15643200278282166,
    "learning_rate": 0.000202939468548981,
    "timestep_mean": 535.9375,
    "epsilon_norm": 7.110021114349365,
    "prediction_norm": 6.63587760925293,
    "timestamp": "2025-07-14T12:35:07.197542"
  },
  {
    "epoch": 2595,
    "batch_idx": 192,
    "loss": 0.3181287348270416,
    "learning_rate": 0.000202939468548981,
    "timestep_mean": 589.5,
    "epsilon_norm": 3.320333957672119,
    "prediction_norm": 2.3327133655548096,
    "timestamp": "2025-07-14T12:35:07.202805"
  },
  {
    "epoch": 2596,
    "batch_idx": 0,
    "loss": 0.21023544669151306,
    "learning_rate": 0.00020357158991947093,
    "timestep_mean": 517.84375,
    "epsilon_norm": 7.759536266326904,
    "prediction_norm": 7.012597560882568,
    "timestamp": "2025-07-14T12:35:07.207972"
  },
  {
    "epoch": 2596,
    "batch_idx": 32,
    "loss": 0.2392772138118744,
    "learning_rate": 0.00020357158991947093,
    "timestep_mean": 499.40625,
    "epsilon_norm": 7.579478740692139,
    "prediction_norm": 6.459845066070557,
    "timestamp": "2025-07-14T12:35:07.213158"
  },
  {
    "epoch": 2596,
    "batch_idx": 64,
    "loss": 0.30983924865722656,
    "learning_rate": 0.00020357158991947093,
    "timestep_mean": 411.875,
    "epsilon_norm": 7.75200080871582,
    "prediction_norm": 5.872285842895508,
    "timestamp": "2025-07-14T12:35:07.218321"
  },
  {
    "epoch": 2596,
    "batch_idx": 96,
    "loss": 0.29106003046035767,
    "learning_rate": 0.00020357158991947093,
    "timestep_mean": 524.9375,
    "epsilon_norm": 8.02331829071045,
    "prediction_norm": 6.9224348068237305,
    "timestamp": "2025-07-14T12:35:07.223645"
  },
  {
    "epoch": 2596,
    "batch_idx": 128,
    "loss": 0.6237331628799438,
    "learning_rate": 0.00020357158991947093,
    "timestep_mean": 425.0625,
    "epsilon_norm": 8.194074630737305,
    "prediction_norm": 5.503403186798096,
    "timestamp": "2025-07-14T12:35:07.228839"
  },
  {
    "epoch": 2596,
    "batch_idx": 160,
    "loss": 0.20456735789775848,
    "learning_rate": 0.00020357158991947093,
    "timestep_mean": 466.65625,
    "epsilon_norm": 8.146828651428223,
    "prediction_norm": 7.579224586486816,
    "timestamp": "2025-07-14T12:35:07.234170"
  },
  {
    "epoch": 2596,
    "batch_idx": 192,
    "loss": 0.2561368942260742,
    "learning_rate": 0.00020357158991947093,
    "timestep_mean": 437.5,
    "epsilon_norm": 4.606024265289307,
    "prediction_norm": 4.480942726135254,
    "timestamp": "2025-07-14T12:35:07.239308"
  },
  {
    "epoch": 2597,
    "batch_idx": 0,
    "loss": 0.3681349754333496,
    "learning_rate": 0.0002042044426975953,
    "timestep_mean": 464.15625,
    "epsilon_norm": 8.266125679016113,
    "prediction_norm": 6.174590110778809,
    "timestamp": "2025-07-14T12:35:07.244506"
  },
  {
    "epoch": 2597,
    "batch_idx": 32,
    "loss": 0.157217875123024,
    "learning_rate": 0.0002042044426975953,
    "timestep_mean": 456.125,
    "epsilon_norm": 6.758499622344971,
    "prediction_norm": 6.195956707000732,
    "timestamp": "2025-07-14T12:35:07.249627"
  },
  {
    "epoch": 2597,
    "batch_idx": 64,
    "loss": 0.17090046405792236,
    "learning_rate": 0.0002042044426975953,
    "timestep_mean": 484.96875,
    "epsilon_norm": 8.076552391052246,
    "prediction_norm": 7.592926025390625,
    "timestamp": "2025-07-14T12:35:07.254506"
  },
  {
    "epoch": 2597,
    "batch_idx": 96,
    "loss": 0.19939807057380676,
    "learning_rate": 0.0002042044426975953,
    "timestep_mean": 474.59375,
    "epsilon_norm": 7.893746376037598,
    "prediction_norm": 7.546415328979492,
    "timestamp": "2025-07-14T12:35:07.259410"
  },
  {
    "epoch": 2597,
    "batch_idx": 128,
    "loss": 0.20922711491584778,
    "learning_rate": 0.0002042044426975953,
    "timestep_mean": 381.5,
    "epsilon_norm": 7.114967346191406,
    "prediction_norm": 5.695887565612793,
    "timestamp": "2025-07-14T12:35:07.264510"
  },
  {
    "epoch": 2597,
    "batch_idx": 160,
    "loss": 0.3026636838912964,
    "learning_rate": 0.0002042044426975953,
    "timestep_mean": 449.34375,
    "epsilon_norm": 7.675363540649414,
    "prediction_norm": 6.376710414886475,
    "timestamp": "2025-07-14T12:35:07.269734"
  },
  {
    "epoch": 2597,
    "batch_idx": 192,
    "loss": 0.3586350083351135,
    "learning_rate": 0.0002042044426975953,
    "timestep_mean": 466.125,
    "epsilon_norm": 4.827165126800537,
    "prediction_norm": 3.7477002143859863,
    "timestamp": "2025-07-14T12:35:07.274803"
  },
  {
    "epoch": 2598,
    "batch_idx": 0,
    "loss": 0.3438798785209656,
    "learning_rate": 0.0002048380253218532,
    "timestep_mean": 502.09375,
    "epsilon_norm": 7.639842510223389,
    "prediction_norm": 6.583810329437256,
    "timestamp": "2025-07-14T12:35:07.280180"
  },
  {
    "epoch": 2598,
    "batch_idx": 32,
    "loss": 0.3592732548713684,
    "learning_rate": 0.0002048380253218532,
    "timestep_mean": 505.46875,
    "epsilon_norm": 8.36009693145752,
    "prediction_norm": 6.409046173095703,
    "timestamp": "2025-07-14T12:35:07.285347"
  },
  {
    "epoch": 2598,
    "batch_idx": 64,
    "loss": 0.25032544136047363,
    "learning_rate": 0.0002048380253218532,
    "timestep_mean": 471.71875,
    "epsilon_norm": 7.39646577835083,
    "prediction_norm": 6.233573913574219,
    "timestamp": "2025-07-14T12:35:07.290344"
  },
  {
    "epoch": 2598,
    "batch_idx": 96,
    "loss": 0.26836085319519043,
    "learning_rate": 0.0002048380253218532,
    "timestep_mean": 487.625,
    "epsilon_norm": 8.142531394958496,
    "prediction_norm": 6.365863800048828,
    "timestamp": "2025-07-14T12:35:07.295546"
  },
  {
    "epoch": 2598,
    "batch_idx": 128,
    "loss": 0.3404584527015686,
    "learning_rate": 0.0002048380253218532,
    "timestep_mean": 474.46875,
    "epsilon_norm": 8.289748191833496,
    "prediction_norm": 6.119505882263184,
    "timestamp": "2025-07-14T12:35:07.300845"
  },
  {
    "epoch": 2598,
    "batch_idx": 160,
    "loss": 0.16565436124801636,
    "learning_rate": 0.0002048380253218532,
    "timestep_mean": 568.8125,
    "epsilon_norm": 7.250462532043457,
    "prediction_norm": 6.51507043838501,
    "timestamp": "2025-07-14T12:35:07.305937"
  },
  {
    "epoch": 2598,
    "batch_idx": 192,
    "loss": 0.19772274792194366,
    "learning_rate": 0.0002048380253218532,
    "timestep_mean": 591.75,
    "epsilon_norm": 4.673155307769775,
    "prediction_norm": 4.524019241333008,
    "timestamp": "2025-07-14T12:35:07.311021"
  },
  {
    "epoch": 2599,
    "batch_idx": 0,
    "loss": 0.20875036716461182,
    "learning_rate": 0.00020547233622894242,
    "timestep_mean": 555.375,
    "epsilon_norm": 8.610339164733887,
    "prediction_norm": 7.2992377281188965,
    "timestamp": "2025-07-14T12:35:07.316172"
  },
  {
    "epoch": 2599,
    "batch_idx": 32,
    "loss": 0.2638716399669647,
    "learning_rate": 0.00020547233622894242,
    "timestep_mean": 468.71875,
    "epsilon_norm": 8.128170013427734,
    "prediction_norm": 7.126020431518555,
    "timestamp": "2025-07-14T12:35:07.321202"
  },
  {
    "epoch": 2599,
    "batch_idx": 64,
    "loss": 0.37497425079345703,
    "learning_rate": 0.00020547233622894242,
    "timestep_mean": 513.0625,
    "epsilon_norm": 7.411756992340088,
    "prediction_norm": 5.865695953369141,
    "timestamp": "2025-07-14T12:35:07.326336"
  },
  {
    "epoch": 2599,
    "batch_idx": 96,
    "loss": 0.46654319763183594,
    "learning_rate": 0.00020547233622894242,
    "timestep_mean": 427.3125,
    "epsilon_norm": 8.387931823730469,
    "prediction_norm": 6.946965217590332,
    "timestamp": "2025-07-14T12:35:07.331544"
  },
  {
    "epoch": 2599,
    "batch_idx": 128,
    "loss": 0.2765794098377228,
    "learning_rate": 0.00020547233622894242,
    "timestep_mean": 490.5625,
    "epsilon_norm": 8.132113456726074,
    "prediction_norm": 6.49174165725708,
    "timestamp": "2025-07-14T12:35:07.336650"
  },
  {
    "epoch": 2599,
    "batch_idx": 160,
    "loss": 0.629682183265686,
    "learning_rate": 0.00020547233622894242,
    "timestep_mean": 480.625,
    "epsilon_norm": 8.577122688293457,
    "prediction_norm": 6.9750285148620605,
    "timestamp": "2025-07-14T12:35:07.342026"
  },
  {
    "epoch": 2599,
    "batch_idx": 192,
    "loss": 0.1291387379169464,
    "learning_rate": 0.00020547233622894242,
    "timestep_mean": 533.0,
    "epsilon_norm": 3.4266934394836426,
    "prediction_norm": 2.9185526371002197,
    "timestamp": "2025-07-14T12:35:07.346923"
  },
  {
    "epoch": 2600,
    "batch_idx": 0,
    "loss": 0.20010244846343994,
    "learning_rate": 0.00020610737385376383,
    "timestep_mean": 532.28125,
    "epsilon_norm": 7.501710414886475,
    "prediction_norm": 6.6204986572265625,
    "timestamp": "2025-07-14T12:35:07.351882"
  },
  {
    "epoch": 2600,
    "batch_idx": 32,
    "loss": 0.242605060338974,
    "learning_rate": 0.00020610737385376383,
    "timestep_mean": 406.4375,
    "epsilon_norm": 7.6380839347839355,
    "prediction_norm": 6.602970123291016,
    "timestamp": "2025-07-14T12:35:07.356987"
  },
  {
    "epoch": 2600,
    "batch_idx": 64,
    "loss": 0.18096113204956055,
    "learning_rate": 0.00020610737385376383,
    "timestep_mean": 510.78125,
    "epsilon_norm": 7.300822734832764,
    "prediction_norm": 6.976245403289795,
    "timestamp": "2025-07-14T12:35:07.362112"
  },
  {
    "epoch": 2600,
    "batch_idx": 96,
    "loss": 0.342972069978714,
    "learning_rate": 0.00020610737385376383,
    "timestep_mean": 509.78125,
    "epsilon_norm": 7.984821796417236,
    "prediction_norm": 7.551562786102295,
    "timestamp": "2025-07-14T12:35:07.367222"
  },
  {
    "epoch": 2600,
    "batch_idx": 128,
    "loss": 0.2929077744483948,
    "learning_rate": 0.00020610737385376383,
    "timestep_mean": 406.25,
    "epsilon_norm": 7.389698505401611,
    "prediction_norm": 6.216007232666016,
    "timestamp": "2025-07-14T12:35:07.372353"
  },
  {
    "epoch": 2600,
    "batch_idx": 160,
    "loss": 0.3788633942604065,
    "learning_rate": 0.00020610737385376383,
    "timestep_mean": 598.03125,
    "epsilon_norm": 8.189359664916992,
    "prediction_norm": 6.862740516662598,
    "timestamp": "2025-07-14T12:35:07.377380"
  },
  {
    "epoch": 2600,
    "batch_idx": 192,
    "loss": 0.22884848713874817,
    "learning_rate": 0.00020610737385376383,
    "timestep_mean": 662.75,
    "epsilon_norm": 5.567150592803955,
    "prediction_norm": 4.859776020050049,
    "timestamp": "2025-07-14T12:35:07.382409"
  },
  {
    "epoch": 2601,
    "batch_idx": 0,
    "loss": 0.4727652072906494,
    "learning_rate": 0.00020674313662942523,
    "timestep_mean": 541.71875,
    "epsilon_norm": 9.182760238647461,
    "prediction_norm": 7.403135299682617,
    "timestamp": "2025-07-14T12:35:07.387559"
  },
  {
    "epoch": 2601,
    "batch_idx": 32,
    "loss": 0.24348154664039612,
    "learning_rate": 0.00020674313662942523,
    "timestep_mean": 532.90625,
    "epsilon_norm": 7.411331653594971,
    "prediction_norm": 6.6155877113342285,
    "timestamp": "2025-07-14T12:35:07.392579"
  },
  {
    "epoch": 2601,
    "batch_idx": 64,
    "loss": 0.3629678189754486,
    "learning_rate": 0.00020674313662942523,
    "timestep_mean": 462.875,
    "epsilon_norm": 8.791544914245605,
    "prediction_norm": 6.559207916259766,
    "timestamp": "2025-07-14T12:35:07.397589"
  },
  {
    "epoch": 2601,
    "batch_idx": 96,
    "loss": 0.3101171851158142,
    "learning_rate": 0.00020674313662942523,
    "timestep_mean": 459.46875,
    "epsilon_norm": 8.146275520324707,
    "prediction_norm": 6.014950275421143,
    "timestamp": "2025-07-14T12:35:07.402653"
  },
  {
    "epoch": 2601,
    "batch_idx": 128,
    "loss": 0.334838330745697,
    "learning_rate": 0.00020674313662942523,
    "timestep_mean": 381.53125,
    "epsilon_norm": 6.901767730712891,
    "prediction_norm": 5.770787239074707,
    "timestamp": "2025-07-14T12:35:07.407625"
  },
  {
    "epoch": 2601,
    "batch_idx": 160,
    "loss": 0.3967885375022888,
    "learning_rate": 0.00020674313662942523,
    "timestep_mean": 450.75,
    "epsilon_norm": 8.361778259277344,
    "prediction_norm": 7.339305400848389,
    "timestamp": "2025-07-14T12:35:07.412483"
  },
  {
    "epoch": 2601,
    "batch_idx": 192,
    "loss": 0.030603205785155296,
    "learning_rate": 0.00020674313662942523,
    "timestep_mean": 467.625,
    "epsilon_norm": 3.519308090209961,
    "prediction_norm": 3.254438638687134,
    "timestamp": "2025-07-14T12:35:07.417659"
  },
  {
    "epoch": 2602,
    "batch_idx": 0,
    "loss": 0.3934880197048187,
    "learning_rate": 0.00020737962298724525,
    "timestep_mean": 503.53125,
    "epsilon_norm": 9.222681999206543,
    "prediction_norm": 7.266713619232178,
    "timestamp": "2025-07-14T12:35:07.422835"
  },
  {
    "epoch": 2602,
    "batch_idx": 32,
    "loss": 0.30945688486099243,
    "learning_rate": 0.00020737962298724525,
    "timestep_mean": 458.375,
    "epsilon_norm": 8.26923656463623,
    "prediction_norm": 6.473451137542725,
    "timestamp": "2025-07-14T12:35:07.428089"
  },
  {
    "epoch": 2602,
    "batch_idx": 64,
    "loss": 0.4061799645423889,
    "learning_rate": 0.00020737962298724525,
    "timestep_mean": 443.71875,
    "epsilon_norm": 7.6877760887146,
    "prediction_norm": 6.255301475524902,
    "timestamp": "2025-07-14T12:35:07.433373"
  },
  {
    "epoch": 2602,
    "batch_idx": 96,
    "loss": 0.1432241052389145,
    "learning_rate": 0.00020737962298724525,
    "timestep_mean": 467.1875,
    "epsilon_norm": 6.659401893615723,
    "prediction_norm": 5.9844183921813965,
    "timestamp": "2025-07-14T12:35:07.438909"
  },
  {
    "epoch": 2602,
    "batch_idx": 128,
    "loss": 0.29556673765182495,
    "learning_rate": 0.00020737962298724525,
    "timestep_mean": 446.1875,
    "epsilon_norm": 7.5659308433532715,
    "prediction_norm": 5.536943435668945,
    "timestamp": "2025-07-14T12:35:07.444092"
  },
  {
    "epoch": 2602,
    "batch_idx": 160,
    "loss": 0.27368372678756714,
    "learning_rate": 0.00020737962298724525,
    "timestep_mean": 429.0,
    "epsilon_norm": 7.864902019500732,
    "prediction_norm": 7.141323566436768,
    "timestamp": "2025-07-14T12:35:07.449367"
  },
  {
    "epoch": 2602,
    "batch_idx": 192,
    "loss": 0.3215177059173584,
    "learning_rate": 0.00020737962298724525,
    "timestep_mean": 522.875,
    "epsilon_norm": 4.026321887969971,
    "prediction_norm": 3.3137896060943604,
    "timestamp": "2025-07-14T12:35:07.454651"
  },
  {
    "epoch": 2603,
    "batch_idx": 0,
    "loss": 0.29898810386657715,
    "learning_rate": 0.000208016831356757,
    "timestep_mean": 424.28125,
    "epsilon_norm": 7.6566691398620605,
    "prediction_norm": 6.257069110870361,
    "timestamp": "2025-07-14T12:35:07.459887"
  },
  {
    "epoch": 2603,
    "batch_idx": 32,
    "loss": 0.1661590337753296,
    "learning_rate": 0.000208016831356757,
    "timestep_mean": 497.21875,
    "epsilon_norm": 8.190812110900879,
    "prediction_norm": 7.069427013397217,
    "timestamp": "2025-07-14T12:35:07.464964"
  },
  {
    "epoch": 2603,
    "batch_idx": 64,
    "loss": 0.18958358466625214,
    "learning_rate": 0.000208016831356757,
    "timestep_mean": 548.25,
    "epsilon_norm": 7.850128173828125,
    "prediction_norm": 6.890565395355225,
    "timestamp": "2025-07-14T12:35:07.470119"
  },
  {
    "epoch": 2603,
    "batch_idx": 96,
    "loss": 0.25364744663238525,
    "learning_rate": 0.000208016831356757,
    "timestep_mean": 541.15625,
    "epsilon_norm": 8.228944778442383,
    "prediction_norm": 7.101609230041504,
    "timestamp": "2025-07-14T12:35:07.475360"
  },
  {
    "epoch": 2603,
    "batch_idx": 128,
    "loss": 0.3848888874053955,
    "learning_rate": 0.000208016831356757,
    "timestep_mean": 519.9375,
    "epsilon_norm": 8.103217124938965,
    "prediction_norm": 6.611563682556152,
    "timestamp": "2025-07-14T12:35:07.480563"
  },
  {
    "epoch": 2603,
    "batch_idx": 160,
    "loss": 0.2089935541152954,
    "learning_rate": 0.000208016831356757,
    "timestep_mean": 475.09375,
    "epsilon_norm": 7.278067111968994,
    "prediction_norm": 6.2887701988220215,
    "timestamp": "2025-07-14T12:35:07.485850"
  },
  {
    "epoch": 2603,
    "batch_idx": 192,
    "loss": 0.20213639736175537,
    "learning_rate": 0.000208016831356757,
    "timestep_mean": 663.625,
    "epsilon_norm": 3.8883774280548096,
    "prediction_norm": 3.4408228397369385,
    "timestamp": "2025-07-14T12:35:07.491093"
  },
  {
    "epoch": 2604,
    "batch_idx": 0,
    "loss": 0.3511599004268646,
    "learning_rate": 0.00020865476016571218,
    "timestep_mean": 486.3125,
    "epsilon_norm": 7.940634727478027,
    "prediction_norm": 6.56646203994751,
    "timestamp": "2025-07-14T12:35:07.496419"
  },
  {
    "epoch": 2604,
    "batch_idx": 32,
    "loss": 0.20424184203147888,
    "learning_rate": 0.00020865476016571218,
    "timestep_mean": 412.1875,
    "epsilon_norm": 7.287938594818115,
    "prediction_norm": 6.314400672912598,
    "timestamp": "2025-07-14T12:35:07.501501"
  },
  {
    "epoch": 2604,
    "batch_idx": 64,
    "loss": 0.3501240015029907,
    "learning_rate": 0.00020865476016571218,
    "timestep_mean": 408.21875,
    "epsilon_norm": 8.018129348754883,
    "prediction_norm": 6.23103141784668,
    "timestamp": "2025-07-14T12:35:07.506479"
  },
  {
    "epoch": 2604,
    "batch_idx": 96,
    "loss": 0.24182021617889404,
    "learning_rate": 0.00020865476016571218,
    "timestep_mean": 499.40625,
    "epsilon_norm": 7.497298240661621,
    "prediction_norm": 6.436929702758789,
    "timestamp": "2025-07-14T12:35:07.511604"
  },
  {
    "epoch": 2604,
    "batch_idx": 128,
    "loss": 0.3823881149291992,
    "learning_rate": 0.00020865476016571218,
    "timestep_mean": 453.0625,
    "epsilon_norm": 7.164802551269531,
    "prediction_norm": 6.212358474731445,
    "timestamp": "2025-07-14T12:35:07.516701"
  },
  {
    "epoch": 2604,
    "batch_idx": 160,
    "loss": 0.206853985786438,
    "learning_rate": 0.00020865476016571218,
    "timestep_mean": 512.6875,
    "epsilon_norm": 6.581127166748047,
    "prediction_norm": 5.512667655944824,
    "timestamp": "2025-07-14T12:35:07.522031"
  },
  {
    "epoch": 2604,
    "batch_idx": 192,
    "loss": 0.2671283483505249,
    "learning_rate": 0.00020865476016571218,
    "timestep_mean": 505.25,
    "epsilon_norm": 3.992643117904663,
    "prediction_norm": 3.376612663269043,
    "timestamp": "2025-07-14T12:35:07.527198"
  },
  {
    "epoch": 2605,
    "batch_idx": 0,
    "loss": 0.24738360941410065,
    "learning_rate": 0.00020929340784008523,
    "timestep_mean": 515.8125,
    "epsilon_norm": 7.190219402313232,
    "prediction_norm": 5.700862884521484,
    "timestamp": "2025-07-14T12:35:07.532357"
  },
  {
    "epoch": 2605,
    "batch_idx": 32,
    "loss": 0.3141249418258667,
    "learning_rate": 0.00020929340784008523,
    "timestep_mean": 553.25,
    "epsilon_norm": 8.293429374694824,
    "prediction_norm": 6.73675012588501,
    "timestamp": "2025-07-14T12:35:07.537558"
  },
  {
    "epoch": 2605,
    "batch_idx": 64,
    "loss": 0.1652669608592987,
    "learning_rate": 0.00020929340784008523,
    "timestep_mean": 520.09375,
    "epsilon_norm": 7.039402961730957,
    "prediction_norm": 6.757987976074219,
    "timestamp": "2025-07-14T12:35:07.542972"
  },
  {
    "epoch": 2605,
    "batch_idx": 96,
    "loss": 0.13919171690940857,
    "learning_rate": 0.00020929340784008523,
    "timestep_mean": 504.125,
    "epsilon_norm": 8.008190155029297,
    "prediction_norm": 7.198121070861816,
    "timestamp": "2025-07-14T12:35:07.548115"
  },
  {
    "epoch": 2605,
    "batch_idx": 128,
    "loss": 0.4018106460571289,
    "learning_rate": 0.00020929340784008523,
    "timestep_mean": 437.15625,
    "epsilon_norm": 8.575068473815918,
    "prediction_norm": 7.060732841491699,
    "timestamp": "2025-07-14T12:35:07.553254"
  },
  {
    "epoch": 2605,
    "batch_idx": 160,
    "loss": 0.3337043523788452,
    "learning_rate": 0.00020929340784008523,
    "timestep_mean": 449.40625,
    "epsilon_norm": 7.905435085296631,
    "prediction_norm": 6.800910949707031,
    "timestamp": "2025-07-14T12:35:07.558709"
  },
  {
    "epoch": 2605,
    "batch_idx": 192,
    "loss": 0.2937650680541992,
    "learning_rate": 0.00020929340784008523,
    "timestep_mean": 350.875,
    "epsilon_norm": 3.8433706760406494,
    "prediction_norm": 2.4954285621643066,
    "timestamp": "2025-07-14T12:35:07.563934"
  },
  {
    "epoch": 2606,
    "batch_idx": 0,
    "loss": 0.23974871635437012,
    "learning_rate": 0.00020993277280407577,
    "timestep_mean": 583.3125,
    "epsilon_norm": 8.127534866333008,
    "prediction_norm": 7.353804111480713,
    "timestamp": "2025-07-14T12:35:07.569083"
  },
  {
    "epoch": 2606,
    "batch_idx": 32,
    "loss": 0.16469678282737732,
    "learning_rate": 0.00020993277280407577,
    "timestep_mean": 493.0625,
    "epsilon_norm": 7.030440807342529,
    "prediction_norm": 6.5775604248046875,
    "timestamp": "2025-07-14T12:35:07.574319"
  },
  {
    "epoch": 2606,
    "batch_idx": 64,
    "loss": 0.21199041604995728,
    "learning_rate": 0.00020993277280407577,
    "timestep_mean": 568.03125,
    "epsilon_norm": 9.034714698791504,
    "prediction_norm": 7.833465099334717,
    "timestamp": "2025-07-14T12:35:07.579413"
  },
  {
    "epoch": 2606,
    "batch_idx": 96,
    "loss": 0.1583690047264099,
    "learning_rate": 0.00020993277280407577,
    "timestep_mean": 634.96875,
    "epsilon_norm": 8.306596755981445,
    "prediction_norm": 7.584737300872803,
    "timestamp": "2025-07-14T12:35:07.584617"
  },
  {
    "epoch": 2606,
    "batch_idx": 128,
    "loss": 0.30727750062942505,
    "learning_rate": 0.00020993277280407577,
    "timestep_mean": 464.125,
    "epsilon_norm": 7.432719707489014,
    "prediction_norm": 5.59854793548584,
    "timestamp": "2025-07-14T12:35:07.589780"
  },
  {
    "epoch": 2606,
    "batch_idx": 160,
    "loss": 0.4074869155883789,
    "learning_rate": 0.00020993277280407577,
    "timestep_mean": 484.40625,
    "epsilon_norm": 8.385416030883789,
    "prediction_norm": 6.496545314788818,
    "timestamp": "2025-07-14T12:35:07.594948"
  },
  {
    "epoch": 2606,
    "batch_idx": 192,
    "loss": 0.39460140466690063,
    "learning_rate": 0.00020993277280407577,
    "timestep_mean": 401.125,
    "epsilon_norm": 3.603543758392334,
    "prediction_norm": 2.866175889968872,
    "timestamp": "2025-07-14T12:35:07.600135"
  },
  {
    "epoch": 2607,
    "batch_idx": 0,
    "loss": 0.3112082779407501,
    "learning_rate": 0.00021057285348011472,
    "timestep_mean": 505.625,
    "epsilon_norm": 7.737798690795898,
    "prediction_norm": 6.042008399963379,
    "timestamp": "2025-07-14T12:35:07.605409"
  },
  {
    "epoch": 2607,
    "batch_idx": 32,
    "loss": 0.19615577161312103,
    "learning_rate": 0.00021057285348011472,
    "timestep_mean": 532.65625,
    "epsilon_norm": 7.096879482269287,
    "prediction_norm": 6.302920818328857,
    "timestamp": "2025-07-14T12:35:07.610492"
  },
  {
    "epoch": 2607,
    "batch_idx": 64,
    "loss": 0.2234145998954773,
    "learning_rate": 0.00021057285348011472,
    "timestep_mean": 427.75,
    "epsilon_norm": 6.865592002868652,
    "prediction_norm": 5.730305194854736,
    "timestamp": "2025-07-14T12:35:07.615607"
  },
  {
    "epoch": 2607,
    "batch_idx": 96,
    "loss": 0.5524297952651978,
    "learning_rate": 0.00021057285348011472,
    "timestep_mean": 392.375,
    "epsilon_norm": 8.764808654785156,
    "prediction_norm": 6.866793155670166,
    "timestamp": "2025-07-14T12:35:07.620863"
  },
  {
    "epoch": 2607,
    "batch_idx": 128,
    "loss": 0.18217816948890686,
    "learning_rate": 0.00021057285348011472,
    "timestep_mean": 526.8125,
    "epsilon_norm": 8.152237892150879,
    "prediction_norm": 7.23264217376709,
    "timestamp": "2025-07-14T12:35:07.626062"
  },
  {
    "epoch": 2607,
    "batch_idx": 160,
    "loss": 0.28905919194221497,
    "learning_rate": 0.00021057285348011472,
    "timestep_mean": 479.0,
    "epsilon_norm": 8.068617820739746,
    "prediction_norm": 7.151655197143555,
    "timestamp": "2025-07-14T12:35:07.631123"
  },
  {
    "epoch": 2607,
    "batch_idx": 192,
    "loss": 0.4574774205684662,
    "learning_rate": 0.00021057285348011472,
    "timestep_mean": 467.125,
    "epsilon_norm": 4.552955627441406,
    "prediction_norm": 2.9669063091278076,
    "timestamp": "2025-07-14T12:35:07.636299"
  },
  {
    "epoch": 2608,
    "batch_idx": 0,
    "loss": 0.13875934481620789,
    "learning_rate": 0.00021121364828886663,
    "timestep_mean": 542.21875,
    "epsilon_norm": 8.771225929260254,
    "prediction_norm": 8.26491641998291,
    "timestamp": "2025-07-14T12:35:07.641408"
  },
  {
    "epoch": 2608,
    "batch_idx": 32,
    "loss": 0.2271653264760971,
    "learning_rate": 0.00021121364828886663,
    "timestep_mean": 468.75,
    "epsilon_norm": 7.783134937286377,
    "prediction_norm": 7.025710582733154,
    "timestamp": "2025-07-14T12:35:07.646707"
  },
  {
    "epoch": 2608,
    "batch_idx": 64,
    "loss": 0.179517924785614,
    "learning_rate": 0.00021121364828886663,
    "timestep_mean": 550.96875,
    "epsilon_norm": 8.331068992614746,
    "prediction_norm": 7.245820999145508,
    "timestamp": "2025-07-14T12:35:07.651598"
  },
  {
    "epoch": 2608,
    "batch_idx": 96,
    "loss": 0.1970176249742508,
    "learning_rate": 0.00021121364828886663,
    "timestep_mean": 489.65625,
    "epsilon_norm": 7.061214447021484,
    "prediction_norm": 6.441562652587891,
    "timestamp": "2025-07-14T12:35:07.656520"
  },
  {
    "epoch": 2608,
    "batch_idx": 128,
    "loss": 0.3168637752532959,
    "learning_rate": 0.00021121364828886663,
    "timestep_mean": 512.6875,
    "epsilon_norm": 7.218179225921631,
    "prediction_norm": 5.883127689361572,
    "timestamp": "2025-07-14T12:35:07.661633"
  },
  {
    "epoch": 2608,
    "batch_idx": 160,
    "loss": 0.2640344798564911,
    "learning_rate": 0.00021121364828886663,
    "timestep_mean": 518.71875,
    "epsilon_norm": 7.519020080566406,
    "prediction_norm": 6.319243431091309,
    "timestamp": "2025-07-14T12:35:07.666733"
  },
  {
    "epoch": 2608,
    "batch_idx": 192,
    "loss": 0.17931614816188812,
    "learning_rate": 0.00021121364828886663,
    "timestep_mean": 622.0,
    "epsilon_norm": 2.915834903717041,
    "prediction_norm": 2.8935532569885254,
    "timestamp": "2025-07-14T12:35:07.671851"
  },
  {
    "epoch": 2609,
    "batch_idx": 0,
    "loss": 0.16337013244628906,
    "learning_rate": 0.00021185515564923398,
    "timestep_mean": 531.15625,
    "epsilon_norm": 7.9972710609436035,
    "prediction_norm": 6.957895755767822,
    "timestamp": "2025-07-14T12:35:07.676946"
  },
  {
    "epoch": 2609,
    "batch_idx": 32,
    "loss": 0.24895575642585754,
    "learning_rate": 0.00021185515564923398,
    "timestep_mean": 499.6875,
    "epsilon_norm": 6.921679496765137,
    "prediction_norm": 5.5943603515625,
    "timestamp": "2025-07-14T12:35:07.682050"
  },
  {
    "epoch": 2609,
    "batch_idx": 64,
    "loss": 0.06205137073993683,
    "learning_rate": 0.00021185515564923398,
    "timestep_mean": 562.78125,
    "epsilon_norm": 8.108981132507324,
    "prediction_norm": 7.996636390686035,
    "timestamp": "2025-07-14T12:35:07.687156"
  },
  {
    "epoch": 2609,
    "batch_idx": 96,
    "loss": 0.4075441360473633,
    "learning_rate": 0.00021185515564923398,
    "timestep_mean": 399.15625,
    "epsilon_norm": 7.459129333496094,
    "prediction_norm": 5.821137428283691,
    "timestamp": "2025-07-14T12:35:07.692427"
  },
  {
    "epoch": 2609,
    "batch_idx": 128,
    "loss": 0.21801233291625977,
    "learning_rate": 0.00021185515564923398,
    "timestep_mean": 457.375,
    "epsilon_norm": 6.914969444274902,
    "prediction_norm": 6.484511852264404,
    "timestamp": "2025-07-14T12:35:07.697643"
  },
  {
    "epoch": 2609,
    "batch_idx": 160,
    "loss": 0.4107104539871216,
    "learning_rate": 0.00021185515564923398,
    "timestep_mean": 437.875,
    "epsilon_norm": 7.976863384246826,
    "prediction_norm": 6.666613578796387,
    "timestamp": "2025-07-14T12:35:07.702785"
  },
  {
    "epoch": 2609,
    "batch_idx": 192,
    "loss": 0.038807161152362823,
    "learning_rate": 0.00021185515564923398,
    "timestep_mean": 750.75,
    "epsilon_norm": 4.34796142578125,
    "prediction_norm": 4.528717994689941,
    "timestamp": "2025-07-14T12:35:07.707924"
  },
  {
    "epoch": 2610,
    "batch_idx": 0,
    "loss": 0.3594897985458374,
    "learning_rate": 0.00021249737397836108,
    "timestep_mean": 529.09375,
    "epsilon_norm": 7.905425071716309,
    "prediction_norm": 6.253667831420898,
    "timestamp": "2025-07-14T12:35:07.713106"
  },
  {
    "epoch": 2610,
    "batch_idx": 32,
    "loss": 0.25159841775894165,
    "learning_rate": 0.00021249737397836108,
    "timestep_mean": 571.0,
    "epsilon_norm": 7.7135090827941895,
    "prediction_norm": 6.369748592376709,
    "timestamp": "2025-07-14T12:35:07.718213"
  },
  {
    "epoch": 2610,
    "batch_idx": 64,
    "loss": 0.2387339025735855,
    "learning_rate": 0.00021249737397836108,
    "timestep_mean": 575.21875,
    "epsilon_norm": 7.785093784332275,
    "prediction_norm": 6.614022254943848,
    "timestamp": "2025-07-14T12:35:07.723511"
  },
  {
    "epoch": 2610,
    "batch_idx": 96,
    "loss": 0.18999792635440826,
    "learning_rate": 0.00021249737397836108,
    "timestep_mean": 499.46875,
    "epsilon_norm": 7.0918192863464355,
    "prediction_norm": 6.5325236320495605,
    "timestamp": "2025-07-14T12:35:07.728717"
  },
  {
    "epoch": 2610,
    "batch_idx": 128,
    "loss": 0.22659315168857574,
    "learning_rate": 0.00021249737397836108,
    "timestep_mean": 497.34375,
    "epsilon_norm": 7.890143394470215,
    "prediction_norm": 7.063417434692383,
    "timestamp": "2025-07-14T12:35:07.734085"
  },
  {
    "epoch": 2610,
    "batch_idx": 160,
    "loss": 0.3144068121910095,
    "learning_rate": 0.00021249737397836108,
    "timestep_mean": 475.25,
    "epsilon_norm": 8.307135581970215,
    "prediction_norm": 6.683547496795654,
    "timestamp": "2025-07-14T12:35:07.739264"
  },
  {
    "epoch": 2610,
    "batch_idx": 192,
    "loss": 0.4089105725288391,
    "learning_rate": 0.00021249737397836108,
    "timestep_mean": 484.875,
    "epsilon_norm": 5.280938148498535,
    "prediction_norm": 3.999006748199463,
    "timestamp": "2025-07-14T12:35:07.744504"
  },
  {
    "epoch": 2611,
    "batch_idx": 0,
    "loss": 0.19945421814918518,
    "learning_rate": 0.00021314030169163815,
    "timestep_mean": 459.8125,
    "epsilon_norm": 8.08784008026123,
    "prediction_norm": 6.896647930145264,
    "timestamp": "2025-07-14T12:35:07.749539"
  },
  {
    "epoch": 2611,
    "batch_idx": 32,
    "loss": 0.16931909322738647,
    "learning_rate": 0.00021314030169163815,
    "timestep_mean": 578.78125,
    "epsilon_norm": 7.638345718383789,
    "prediction_norm": 7.076290607452393,
    "timestamp": "2025-07-14T12:35:07.754642"
  },
  {
    "epoch": 2611,
    "batch_idx": 64,
    "loss": 0.26613226532936096,
    "learning_rate": 0.00021314030169163815,
    "timestep_mean": 561.34375,
    "epsilon_norm": 7.910120010375977,
    "prediction_norm": 7.259274005889893,
    "timestamp": "2025-07-14T12:35:07.759826"
  },
  {
    "epoch": 2611,
    "batch_idx": 96,
    "loss": 0.19891361892223358,
    "learning_rate": 0.00021314030169163815,
    "timestep_mean": 509.8125,
    "epsilon_norm": 6.644090175628662,
    "prediction_norm": 5.706547737121582,
    "timestamp": "2025-07-14T12:35:07.764885"
  },
  {
    "epoch": 2611,
    "batch_idx": 128,
    "loss": 0.303070992231369,
    "learning_rate": 0.00021314030169163815,
    "timestep_mean": 415.53125,
    "epsilon_norm": 7.436195373535156,
    "prediction_norm": 6.1880879402160645,
    "timestamp": "2025-07-14T12:35:07.770046"
  },
  {
    "epoch": 2611,
    "batch_idx": 160,
    "loss": 0.4606471061706543,
    "learning_rate": 0.00021314030169163815,
    "timestep_mean": 432.875,
    "epsilon_norm": 7.540909290313721,
    "prediction_norm": 5.657779216766357,
    "timestamp": "2025-07-14T12:35:07.775121"
  },
  {
    "epoch": 2611,
    "batch_idx": 192,
    "loss": 0.376437783241272,
    "learning_rate": 0.00021314030169163815,
    "timestep_mean": 466.625,
    "epsilon_norm": 3.4923009872436523,
    "prediction_norm": 3.4690823554992676,
    "timestamp": "2025-07-14T12:35:07.780411"
  },
  {
    "epoch": 2612,
    "batch_idx": 0,
    "loss": 0.24643835425376892,
    "learning_rate": 0.00021378393720270522,
    "timestep_mean": 537.40625,
    "epsilon_norm": 7.455161094665527,
    "prediction_norm": 6.318399906158447,
    "timestamp": "2025-07-14T12:35:07.785617"
  },
  {
    "epoch": 2612,
    "batch_idx": 32,
    "loss": 0.07835909724235535,
    "learning_rate": 0.00021378393720270522,
    "timestep_mean": 564.125,
    "epsilon_norm": 6.4466657638549805,
    "prediction_norm": 6.178186893463135,
    "timestamp": "2025-07-14T12:35:07.790858"
  },
  {
    "epoch": 2612,
    "batch_idx": 64,
    "loss": 0.1946597397327423,
    "learning_rate": 0.00021378393720270522,
    "timestep_mean": 533.9375,
    "epsilon_norm": 8.028775215148926,
    "prediction_norm": 6.6802215576171875,
    "timestamp": "2025-07-14T12:35:07.796090"
  },
  {
    "epoch": 2612,
    "batch_idx": 96,
    "loss": 0.26767420768737793,
    "learning_rate": 0.00021378393720270522,
    "timestep_mean": 466.5,
    "epsilon_norm": 7.286527633666992,
    "prediction_norm": 5.954735279083252,
    "timestamp": "2025-07-14T12:35:07.801312"
  },
  {
    "epoch": 2612,
    "batch_idx": 128,
    "loss": 0.40690988302230835,
    "learning_rate": 0.00021378393720270522,
    "timestep_mean": 554.25,
    "epsilon_norm": 8.264520645141602,
    "prediction_norm": 6.756015300750732,
    "timestamp": "2025-07-14T12:35:07.806571"
  },
  {
    "epoch": 2612,
    "batch_idx": 160,
    "loss": 0.39810559153556824,
    "learning_rate": 0.00021378393720270522,
    "timestep_mean": 554.625,
    "epsilon_norm": 8.482723236083984,
    "prediction_norm": 6.633872985839844,
    "timestamp": "2025-07-14T12:35:07.811872"
  },
  {
    "epoch": 2612,
    "batch_idx": 192,
    "loss": 0.31324848532676697,
    "learning_rate": 0.00021378393720270522,
    "timestep_mean": 549.125,
    "epsilon_norm": 3.2096989154815674,
    "prediction_norm": 2.5720951557159424,
    "timestamp": "2025-07-14T12:35:07.817019"
  },
  {
    "epoch": 2613,
    "batch_idx": 0,
    "loss": 0.2585686445236206,
    "learning_rate": 0.00021442827892345466,
    "timestep_mean": 550.53125,
    "epsilon_norm": 8.469603538513184,
    "prediction_norm": 7.254713535308838,
    "timestamp": "2025-07-14T12:35:07.822064"
  },
  {
    "epoch": 2613,
    "batch_idx": 32,
    "loss": 0.1700810045003891,
    "learning_rate": 0.00021442827892345466,
    "timestep_mean": 518.625,
    "epsilon_norm": 8.196917533874512,
    "prediction_norm": 6.675836563110352,
    "timestamp": "2025-07-14T12:35:07.827008"
  },
  {
    "epoch": 2613,
    "batch_idx": 64,
    "loss": 0.23582381010055542,
    "learning_rate": 0.00021442827892345466,
    "timestep_mean": 512.375,
    "epsilon_norm": 7.321093559265137,
    "prediction_norm": 6.176888465881348,
    "timestamp": "2025-07-14T12:35:07.832060"
  },
  {
    "epoch": 2613,
    "batch_idx": 96,
    "loss": 0.27637922763824463,
    "learning_rate": 0.00021442827892345466,
    "timestep_mean": 542.78125,
    "epsilon_norm": 8.486833572387695,
    "prediction_norm": 7.160052299499512,
    "timestamp": "2025-07-14T12:35:07.837250"
  },
  {
    "epoch": 2613,
    "batch_idx": 128,
    "loss": 0.3264016807079315,
    "learning_rate": 0.00021442827892345466,
    "timestep_mean": 504.15625,
    "epsilon_norm": 8.69902515411377,
    "prediction_norm": 6.780730247497559,
    "timestamp": "2025-07-14T12:35:07.842489"
  },
  {
    "epoch": 2613,
    "batch_idx": 160,
    "loss": 0.2637297213077545,
    "learning_rate": 0.00021442827892345466,
    "timestep_mean": 510.34375,
    "epsilon_norm": 7.510804653167725,
    "prediction_norm": 6.151458740234375,
    "timestamp": "2025-07-14T12:35:07.847545"
  },
  {
    "epoch": 2613,
    "batch_idx": 192,
    "loss": 0.08236665278673172,
    "learning_rate": 0.00021442827892345466,
    "timestep_mean": 628.875,
    "epsilon_norm": 4.282102108001709,
    "prediction_norm": 4.06465482711792,
    "timestamp": "2025-07-14T12:35:07.852604"
  },
  {
    "epoch": 2614,
    "batch_idx": 0,
    "loss": 0.37210017442703247,
    "learning_rate": 0.00021507332526403834,
    "timestep_mean": 466.1875,
    "epsilon_norm": 8.19683837890625,
    "prediction_norm": 6.506934642791748,
    "timestamp": "2025-07-14T12:35:07.857757"
  },
  {
    "epoch": 2614,
    "batch_idx": 32,
    "loss": 0.20360887050628662,
    "learning_rate": 0.00021507332526403834,
    "timestep_mean": 554.75,
    "epsilon_norm": 9.025307655334473,
    "prediction_norm": 8.13211441040039,
    "timestamp": "2025-07-14T12:35:07.862841"
  },
  {
    "epoch": 2614,
    "batch_idx": 64,
    "loss": 0.16706976294517517,
    "learning_rate": 0.00021507332526403834,
    "timestep_mean": 525.59375,
    "epsilon_norm": 7.416891574859619,
    "prediction_norm": 7.309587478637695,
    "timestamp": "2025-07-14T12:35:07.868071"
  },
  {
    "epoch": 2614,
    "batch_idx": 96,
    "loss": 0.18066297471523285,
    "learning_rate": 0.00021507332526403834,
    "timestep_mean": 546.0625,
    "epsilon_norm": 6.657347202301025,
    "prediction_norm": 6.022051811218262,
    "timestamp": "2025-07-14T12:35:07.873227"
  },
  {
    "epoch": 2614,
    "batch_idx": 128,
    "loss": 0.407161682844162,
    "learning_rate": 0.00021507332526403834,
    "timestep_mean": 482.34375,
    "epsilon_norm": 10.145938873291016,
    "prediction_norm": 8.50416374206543,
    "timestamp": "2025-07-14T12:35:07.878377"
  },
  {
    "epoch": 2614,
    "batch_idx": 160,
    "loss": 0.13391493260860443,
    "learning_rate": 0.00021507332526403834,
    "timestep_mean": 549.03125,
    "epsilon_norm": 6.824704170227051,
    "prediction_norm": 6.156678199768066,
    "timestamp": "2025-07-14T12:35:07.883711"
  },
  {
    "epoch": 2614,
    "batch_idx": 192,
    "loss": 0.2055634707212448,
    "learning_rate": 0.00021507332526403834,
    "timestep_mean": 431.625,
    "epsilon_norm": 5.009241580963135,
    "prediction_norm": 4.1972270011901855,
    "timestamp": "2025-07-14T12:35:07.888899"
  },
  {
    "epoch": 2615,
    "batch_idx": 0,
    "loss": 0.5075329542160034,
    "learning_rate": 0.00021571907463286858,
    "timestep_mean": 412.6875,
    "epsilon_norm": 8.82982349395752,
    "prediction_norm": 7.06589412689209,
    "timestamp": "2025-07-14T12:35:07.894081"
  },
  {
    "epoch": 2615,
    "batch_idx": 32,
    "loss": 0.194560244679451,
    "learning_rate": 0.00021571907463286858,
    "timestep_mean": 482.40625,
    "epsilon_norm": 8.71136474609375,
    "prediction_norm": 7.610128879547119,
    "timestamp": "2025-07-14T12:35:07.899313"
  },
  {
    "epoch": 2615,
    "batch_idx": 64,
    "loss": 0.36173346638679504,
    "learning_rate": 0.00021571907463286858,
    "timestep_mean": 504.5625,
    "epsilon_norm": 9.614278793334961,
    "prediction_norm": 7.956759452819824,
    "timestamp": "2025-07-14T12:35:07.904453"
  },
  {
    "epoch": 2615,
    "batch_idx": 96,
    "loss": 0.20172283053398132,
    "learning_rate": 0.00021571907463286858,
    "timestep_mean": 571.03125,
    "epsilon_norm": 9.218526840209961,
    "prediction_norm": 8.350691795349121,
    "timestamp": "2025-07-14T12:35:07.909616"
  },
  {
    "epoch": 2615,
    "batch_idx": 128,
    "loss": 0.2627203166484833,
    "learning_rate": 0.00021571907463286858,
    "timestep_mean": 504.78125,
    "epsilon_norm": 7.199901103973389,
    "prediction_norm": 6.055924892425537,
    "timestamp": "2025-07-14T12:35:07.914636"
  },
  {
    "epoch": 2615,
    "batch_idx": 160,
    "loss": 0.2529730796813965,
    "learning_rate": 0.00021571907463286858,
    "timestep_mean": 481.09375,
    "epsilon_norm": 7.834526538848877,
    "prediction_norm": 7.082894325256348,
    "timestamp": "2025-07-14T12:35:07.919872"
  },
  {
    "epoch": 2615,
    "batch_idx": 192,
    "loss": 0.10435186326503754,
    "learning_rate": 0.00021571907463286858,
    "timestep_mean": 561.625,
    "epsilon_norm": 4.076540946960449,
    "prediction_norm": 3.9172563552856445,
    "timestamp": "2025-07-14T12:35:07.924937"
  },
  {
    "epoch": 2616,
    "batch_idx": 0,
    "loss": 0.3867471218109131,
    "learning_rate": 0.00021636552543662226,
    "timestep_mean": 499.125,
    "epsilon_norm": 9.030305862426758,
    "prediction_norm": 7.832205295562744,
    "timestamp": "2025-07-14T12:35:07.930310"
  },
  {
    "epoch": 2616,
    "batch_idx": 32,
    "loss": 0.2859496474266052,
    "learning_rate": 0.00021636552543662226,
    "timestep_mean": 485.5,
    "epsilon_norm": 8.019511222839355,
    "prediction_norm": 7.153493881225586,
    "timestamp": "2025-07-14T12:35:07.935534"
  },
  {
    "epoch": 2616,
    "batch_idx": 64,
    "loss": 0.19267421960830688,
    "learning_rate": 0.00021636552543662226,
    "timestep_mean": 452.90625,
    "epsilon_norm": 7.416889190673828,
    "prediction_norm": 6.8004326820373535,
    "timestamp": "2025-07-14T12:35:07.940729"
  },
  {
    "epoch": 2616,
    "batch_idx": 96,
    "loss": 0.25771838426589966,
    "learning_rate": 0.00021636552543662226,
    "timestep_mean": 431.75,
    "epsilon_norm": 7.581114768981934,
    "prediction_norm": 6.639876365661621,
    "timestamp": "2025-07-14T12:35:07.945898"
  },
  {
    "epoch": 2616,
    "batch_idx": 128,
    "loss": 0.2781926393508911,
    "learning_rate": 0.00021636552543662226,
    "timestep_mean": 454.09375,
    "epsilon_norm": 8.826255798339844,
    "prediction_norm": 7.50563907623291,
    "timestamp": "2025-07-14T12:35:07.951289"
  },
  {
    "epoch": 2616,
    "batch_idx": 160,
    "loss": 0.4000979959964752,
    "learning_rate": 0.00021636552543662226,
    "timestep_mean": 477.46875,
    "epsilon_norm": 6.920144081115723,
    "prediction_norm": 5.455776214599609,
    "timestamp": "2025-07-14T12:35:07.956368"
  },
  {
    "epoch": 2616,
    "batch_idx": 192,
    "loss": 0.06800686568021774,
    "learning_rate": 0.00021636552543662226,
    "timestep_mean": 469.0,
    "epsilon_norm": 3.7983038425445557,
    "prediction_norm": 3.6640713214874268,
    "timestamp": "2025-07-14T12:35:07.961473"
  },
  {
    "epoch": 2617,
    "batch_idx": 0,
    "loss": 0.3174740970134735,
    "learning_rate": 0.00021701267608024663,
    "timestep_mean": 464.9375,
    "epsilon_norm": 8.25096607208252,
    "prediction_norm": 6.758501052856445,
    "timestamp": "2025-07-14T12:35:07.966771"
  },
  {
    "epoch": 2617,
    "batch_idx": 32,
    "loss": 0.32614725828170776,
    "learning_rate": 0.00021701267608024663,
    "timestep_mean": 436.875,
    "epsilon_norm": 9.012701988220215,
    "prediction_norm": 7.069516658782959,
    "timestamp": "2025-07-14T12:35:07.972007"
  },
  {
    "epoch": 2617,
    "batch_idx": 64,
    "loss": 0.3374253213405609,
    "learning_rate": 0.00021701267608024663,
    "timestep_mean": 447.25,
    "epsilon_norm": 9.077332496643066,
    "prediction_norm": 7.7942280769348145,
    "timestamp": "2025-07-14T12:35:07.977321"
  },
  {
    "epoch": 2617,
    "batch_idx": 96,
    "loss": 0.2242065966129303,
    "learning_rate": 0.00021701267608024663,
    "timestep_mean": 457.71875,
    "epsilon_norm": 8.172785758972168,
    "prediction_norm": 6.760287284851074,
    "timestamp": "2025-07-14T12:35:07.982524"
  },
  {
    "epoch": 2617,
    "batch_idx": 128,
    "loss": 0.2887705862522125,
    "learning_rate": 0.00021701267608024663,
    "timestep_mean": 508.59375,
    "epsilon_norm": 8.683385848999023,
    "prediction_norm": 7.3227128982543945,
    "timestamp": "2025-07-14T12:35:07.987628"
  },
  {
    "epoch": 2617,
    "batch_idx": 160,
    "loss": 0.3327980637550354,
    "learning_rate": 0.00021701267608024663,
    "timestep_mean": 543.0,
    "epsilon_norm": 9.660762786865234,
    "prediction_norm": 8.13430118560791,
    "timestamp": "2025-07-14T12:35:07.992904"
  },
  {
    "epoch": 2617,
    "batch_idx": 192,
    "loss": 0.3979127109050751,
    "learning_rate": 0.00021701267608024663,
    "timestep_mean": 493.25,
    "epsilon_norm": 3.6613969802856445,
    "prediction_norm": 2.932898759841919,
    "timestamp": "2025-07-14T12:35:07.998064"
  },
  {
    "epoch": 2618,
    "batch_idx": 0,
    "loss": 0.26493892073631287,
    "learning_rate": 0.0002176605249669619,
    "timestep_mean": 489.4375,
    "epsilon_norm": 8.638567924499512,
    "prediction_norm": 7.712391376495361,
    "timestamp": "2025-07-14T12:35:08.003193"
  },
  {
    "epoch": 2618,
    "batch_idx": 32,
    "loss": 0.3129507899284363,
    "learning_rate": 0.0002176605249669619,
    "timestep_mean": 494.9375,
    "epsilon_norm": 6.9411540031433105,
    "prediction_norm": 5.988248825073242,
    "timestamp": "2025-07-14T12:35:08.008667"
  },
  {
    "epoch": 2618,
    "batch_idx": 64,
    "loss": 0.27141696214675903,
    "learning_rate": 0.0002176605249669619,
    "timestep_mean": 586.375,
    "epsilon_norm": 7.794902801513672,
    "prediction_norm": 6.611043453216553,
    "timestamp": "2025-07-14T12:35:08.013837"
  },
  {
    "epoch": 2618,
    "batch_idx": 96,
    "loss": 0.24206221103668213,
    "learning_rate": 0.0002176605249669619,
    "timestep_mean": 517.34375,
    "epsilon_norm": 7.4765305519104,
    "prediction_norm": 6.530993461608887,
    "timestamp": "2025-07-14T12:35:08.018994"
  },
  {
    "epoch": 2618,
    "batch_idx": 128,
    "loss": 0.2239842712879181,
    "learning_rate": 0.0002176605249669619,
    "timestep_mean": 593.8125,
    "epsilon_norm": 8.174649238586426,
    "prediction_norm": 6.810091018676758,
    "timestamp": "2025-07-14T12:35:08.024299"
  },
  {
    "epoch": 2618,
    "batch_idx": 160,
    "loss": 0.21452419459819794,
    "learning_rate": 0.0002176605249669619,
    "timestep_mean": 529.40625,
    "epsilon_norm": 9.122286796569824,
    "prediction_norm": 7.983526706695557,
    "timestamp": "2025-07-14T12:35:08.029501"
  },
  {
    "epoch": 2618,
    "batch_idx": 192,
    "loss": 0.1363930106163025,
    "learning_rate": 0.0002176605249669619,
    "timestep_mean": 710.875,
    "epsilon_norm": 4.035601615905762,
    "prediction_norm": 3.6741957664489746,
    "timestamp": "2025-07-14T12:35:08.034633"
  },
  {
    "epoch": 2619,
    "batch_idx": 0,
    "loss": 0.24801476299762726,
    "learning_rate": 0.00021830907049826524,
    "timestep_mean": 454.90625,
    "epsilon_norm": 8.793107986450195,
    "prediction_norm": 7.648248672485352,
    "timestamp": "2025-07-14T12:35:08.039953"
  },
  {
    "epoch": 2619,
    "batch_idx": 32,
    "loss": 0.4267863631248474,
    "learning_rate": 0.00021830907049826524,
    "timestep_mean": 467.0,
    "epsilon_norm": 9.160626411437988,
    "prediction_norm": 7.241775035858154,
    "timestamp": "2025-07-14T12:35:08.045098"
  },
  {
    "epoch": 2619,
    "batch_idx": 64,
    "loss": 0.388383686542511,
    "learning_rate": 0.00021830907049826524,
    "timestep_mean": 493.28125,
    "epsilon_norm": 8.830245971679688,
    "prediction_norm": 6.806769371032715,
    "timestamp": "2025-07-14T12:35:08.050357"
  },
  {
    "epoch": 2619,
    "batch_idx": 96,
    "loss": 0.16427773237228394,
    "learning_rate": 0.00021830907049826524,
    "timestep_mean": 514.71875,
    "epsilon_norm": 7.943325996398926,
    "prediction_norm": 7.086945533752441,
    "timestamp": "2025-07-14T12:35:08.055643"
  },
  {
    "epoch": 2619,
    "batch_idx": 128,
    "loss": 0.2918981909751892,
    "learning_rate": 0.00021830907049826524,
    "timestep_mean": 404.96875,
    "epsilon_norm": 7.374935150146484,
    "prediction_norm": 6.133610725402832,
    "timestamp": "2025-07-14T12:35:08.060828"
  },
  {
    "epoch": 2619,
    "batch_idx": 160,
    "loss": 0.28795552253723145,
    "learning_rate": 0.00021830907049826524,
    "timestep_mean": 506.25,
    "epsilon_norm": 8.918401718139648,
    "prediction_norm": 7.313405990600586,
    "timestamp": "2025-07-14T12:35:08.065956"
  },
  {
    "epoch": 2619,
    "batch_idx": 192,
    "loss": 0.15844401717185974,
    "learning_rate": 0.00021830907049826524,
    "timestep_mean": 505.375,
    "epsilon_norm": 3.5933918952941895,
    "prediction_norm": 2.618748664855957,
    "timestamp": "2025-07-14T12:35:08.071191"
  },
  {
    "epoch": 2620,
    "batch_idx": 0,
    "loss": 0.18647558987140656,
    "learning_rate": 0.000218958311073935,
    "timestep_mean": 521.625,
    "epsilon_norm": 7.680990219116211,
    "prediction_norm": 7.121603488922119,
    "timestamp": "2025-07-14T12:35:08.076432"
  },
  {
    "epoch": 2620,
    "batch_idx": 32,
    "loss": 0.33783361315727234,
    "learning_rate": 0.000218958311073935,
    "timestep_mean": 401.625,
    "epsilon_norm": 8.657742500305176,
    "prediction_norm": 6.839423656463623,
    "timestamp": "2025-07-14T12:35:08.081585"
  },
  {
    "epoch": 2620,
    "batch_idx": 64,
    "loss": 0.13630537688732147,
    "learning_rate": 0.000218958311073935,
    "timestep_mean": 548.5625,
    "epsilon_norm": 7.82394552230835,
    "prediction_norm": 7.049737453460693,
    "timestamp": "2025-07-14T12:35:08.086876"
  },
  {
    "epoch": 2620,
    "batch_idx": 96,
    "loss": 0.22923141717910767,
    "learning_rate": 0.000218958311073935,
    "timestep_mean": 529.75,
    "epsilon_norm": 7.776688098907471,
    "prediction_norm": 6.980798244476318,
    "timestamp": "2025-07-14T12:35:08.092189"
  },
  {
    "epoch": 2620,
    "batch_idx": 128,
    "loss": 0.30490657687187195,
    "learning_rate": 0.000218958311073935,
    "timestep_mean": 494.21875,
    "epsilon_norm": 8.384552001953125,
    "prediction_norm": 6.915313243865967,
    "timestamp": "2025-07-14T12:35:08.097313"
  },
  {
    "epoch": 2620,
    "batch_idx": 160,
    "loss": 0.35224950313568115,
    "learning_rate": 0.000218958311073935,
    "timestep_mean": 499.1875,
    "epsilon_norm": 7.9511003494262695,
    "prediction_norm": 6.577526569366455,
    "timestamp": "2025-07-14T12:35:08.102513"
  },
  {
    "epoch": 2620,
    "batch_idx": 192,
    "loss": 0.3059859871864319,
    "learning_rate": 0.000218958311073935,
    "timestep_mean": 492.0,
    "epsilon_norm": 3.421724557876587,
    "prediction_norm": 2.542604923248291,
    "timestamp": "2025-07-14T12:35:08.107736"
  },
  {
    "epoch": 2621,
    "batch_idx": 0,
    "loss": 0.40438032150268555,
    "learning_rate": 0.0002196082450920347,
    "timestep_mean": 530.5,
    "epsilon_norm": 9.79992389678955,
    "prediction_norm": 9.077986717224121,
    "timestamp": "2025-07-14T12:35:08.112794"
  },
  {
    "epoch": 2621,
    "batch_idx": 32,
    "loss": 0.42717444896698,
    "learning_rate": 0.0002196082450920347,
    "timestep_mean": 496.65625,
    "epsilon_norm": 8.585848808288574,
    "prediction_norm": 7.5036234855651855,
    "timestamp": "2025-07-14T12:35:08.117978"
  },
  {
    "epoch": 2621,
    "batch_idx": 64,
    "loss": 0.530454695224762,
    "learning_rate": 0.0002196082450920347,
    "timestep_mean": 448.46875,
    "epsilon_norm": 8.697266578674316,
    "prediction_norm": 6.987112045288086,
    "timestamp": "2025-07-14T12:35:08.123211"
  },
  {
    "epoch": 2621,
    "batch_idx": 96,
    "loss": 0.2482222616672516,
    "learning_rate": 0.0002196082450920347,
    "timestep_mean": 482.625,
    "epsilon_norm": 7.760384559631348,
    "prediction_norm": 6.538848876953125,
    "timestamp": "2025-07-14T12:35:08.128494"
  },
  {
    "epoch": 2621,
    "batch_idx": 128,
    "loss": 0.21641486883163452,
    "learning_rate": 0.0002196082450920347,
    "timestep_mean": 502.59375,
    "epsilon_norm": 8.435503005981445,
    "prediction_norm": 7.5327558517456055,
    "timestamp": "2025-07-14T12:35:08.133883"
  },
  {
    "epoch": 2621,
    "batch_idx": 160,
    "loss": 0.1949012279510498,
    "learning_rate": 0.0002196082450920347,
    "timestep_mean": 483.15625,
    "epsilon_norm": 8.373353004455566,
    "prediction_norm": 7.2313337326049805,
    "timestamp": "2025-07-14T12:35:08.139049"
  },
  {
    "epoch": 2621,
    "batch_idx": 192,
    "loss": 0.1269359290599823,
    "learning_rate": 0.0002196082450920347,
    "timestep_mean": 528.375,
    "epsilon_norm": 3.225308656692505,
    "prediction_norm": 3.34588623046875,
    "timestamp": "2025-07-14T12:35:08.144188"
  },
  {
    "epoch": 2622,
    "batch_idx": 0,
    "loss": 0.25631874799728394,
    "learning_rate": 0.00022025887094891716,
    "timestep_mean": 503.84375,
    "epsilon_norm": 8.69653606414795,
    "prediction_norm": 7.83463191986084,
    "timestamp": "2025-07-14T12:35:08.149449"
  },
  {
    "epoch": 2622,
    "batch_idx": 32,
    "loss": 0.291897714138031,
    "learning_rate": 0.00022025887094891716,
    "timestep_mean": 428.03125,
    "epsilon_norm": 7.705698013305664,
    "prediction_norm": 6.398762226104736,
    "timestamp": "2025-07-14T12:35:08.154568"
  },
  {
    "epoch": 2622,
    "batch_idx": 64,
    "loss": 0.4793597459793091,
    "learning_rate": 0.00022025887094891716,
    "timestep_mean": 468.0625,
    "epsilon_norm": 8.174398422241211,
    "prediction_norm": 7.0057501792907715,
    "timestamp": "2025-07-14T12:35:08.159727"
  },
  {
    "epoch": 2622,
    "batch_idx": 96,
    "loss": 0.17958587408065796,
    "learning_rate": 0.00022025887094891716,
    "timestep_mean": 534.0,
    "epsilon_norm": 7.474579334259033,
    "prediction_norm": 6.7114667892456055,
    "timestamp": "2025-07-14T12:35:08.164924"
  },
  {
    "epoch": 2622,
    "batch_idx": 128,
    "loss": 0.19610095024108887,
    "learning_rate": 0.00022025887094891716,
    "timestep_mean": 473.28125,
    "epsilon_norm": 8.218151092529297,
    "prediction_norm": 7.182568550109863,
    "timestamp": "2025-07-14T12:35:08.169873"
  },
  {
    "epoch": 2622,
    "batch_idx": 160,
    "loss": 0.1858794093132019,
    "learning_rate": 0.00022025887094891716,
    "timestep_mean": 544.8125,
    "epsilon_norm": 7.981020450592041,
    "prediction_norm": 7.383313179016113,
    "timestamp": "2025-07-14T12:35:08.174787"
  },
  {
    "epoch": 2622,
    "batch_idx": 192,
    "loss": 0.07431825250387192,
    "learning_rate": 0.00022025887094891716,
    "timestep_mean": 513.75,
    "epsilon_norm": 3.3817622661590576,
    "prediction_norm": 3.286623477935791,
    "timestamp": "2025-07-14T12:35:08.179742"
  },
  {
    "epoch": 2623,
    "batch_idx": 0,
    "loss": 0.171286940574646,
    "learning_rate": 0.0002209101870392269,
    "timestep_mean": 587.125,
    "epsilon_norm": 8.44792652130127,
    "prediction_norm": 7.566708564758301,
    "timestamp": "2025-07-14T12:35:08.184676"
  },
  {
    "epoch": 2623,
    "batch_idx": 32,
    "loss": 0.5407925844192505,
    "learning_rate": 0.0002209101870392269,
    "timestep_mean": 466.96875,
    "epsilon_norm": 8.867061614990234,
    "prediction_norm": 6.986232757568359,
    "timestamp": "2025-07-14T12:35:08.189543"
  },
  {
    "epoch": 2623,
    "batch_idx": 64,
    "loss": 0.35929152369499207,
    "learning_rate": 0.0002209101870392269,
    "timestep_mean": 535.78125,
    "epsilon_norm": 7.451480388641357,
    "prediction_norm": 6.408701419830322,
    "timestamp": "2025-07-14T12:35:08.194488"
  },
  {
    "epoch": 2623,
    "batch_idx": 96,
    "loss": 0.3541179895401001,
    "learning_rate": 0.0002209101870392269,
    "timestep_mean": 471.28125,
    "epsilon_norm": 8.647561073303223,
    "prediction_norm": 7.812666893005371,
    "timestamp": "2025-07-14T12:35:08.199411"
  },
  {
    "epoch": 2623,
    "batch_idx": 128,
    "loss": 0.2572042942047119,
    "learning_rate": 0.0002209101870392269,
    "timestep_mean": 477.65625,
    "epsilon_norm": 8.272738456726074,
    "prediction_norm": 7.431492328643799,
    "timestamp": "2025-07-14T12:35:08.204262"
  },
  {
    "epoch": 2623,
    "batch_idx": 160,
    "loss": 0.3125804364681244,
    "learning_rate": 0.0002209101870392269,
    "timestep_mean": 463.8125,
    "epsilon_norm": 8.554793357849121,
    "prediction_norm": 7.092367172241211,
    "timestamp": "2025-07-14T12:35:08.209275"
  },
  {
    "epoch": 2623,
    "batch_idx": 192,
    "loss": 0.3823462128639221,
    "learning_rate": 0.0002209101870392269,
    "timestep_mean": 431.875,
    "epsilon_norm": 3.900689125061035,
    "prediction_norm": 3.369847536087036,
    "timestamp": "2025-07-14T12:35:08.214134"
  },
  {
    "epoch": 2624,
    "batch_idx": 0,
    "loss": 0.14497047662734985,
    "learning_rate": 0.0002215621917559062,
    "timestep_mean": 533.09375,
    "epsilon_norm": 7.8120856285095215,
    "prediction_norm": 7.494575500488281,
    "timestamp": "2025-07-14T12:35:08.219038"
  },
  {
    "epoch": 2624,
    "batch_idx": 32,
    "loss": 0.5920984148979187,
    "learning_rate": 0.0002215621917559062,
    "timestep_mean": 479.59375,
    "epsilon_norm": 8.103014945983887,
    "prediction_norm": 5.80477237701416,
    "timestamp": "2025-07-14T12:35:08.223997"
  },
  {
    "epoch": 2624,
    "batch_idx": 64,
    "loss": 0.20172978937625885,
    "learning_rate": 0.0002215621917559062,
    "timestep_mean": 486.84375,
    "epsilon_norm": 8.782331466674805,
    "prediction_norm": 7.421592712402344,
    "timestamp": "2025-07-14T12:35:08.228874"
  },
  {
    "epoch": 2624,
    "batch_idx": 96,
    "loss": 0.20883189141750336,
    "learning_rate": 0.0002215621917559062,
    "timestep_mean": 467.4375,
    "epsilon_norm": 7.016066551208496,
    "prediction_norm": 6.099623680114746,
    "timestamp": "2025-07-14T12:35:08.233849"
  },
  {
    "epoch": 2624,
    "batch_idx": 128,
    "loss": 0.20263585448265076,
    "learning_rate": 0.0002215621917559062,
    "timestep_mean": 547.65625,
    "epsilon_norm": 8.004969596862793,
    "prediction_norm": 6.856678485870361,
    "timestamp": "2025-07-14T12:35:08.238714"
  },
  {
    "epoch": 2624,
    "batch_idx": 160,
    "loss": 0.322068452835083,
    "learning_rate": 0.0002215621917559062,
    "timestep_mean": 459.21875,
    "epsilon_norm": 6.839826583862305,
    "prediction_norm": 5.612168788909912,
    "timestamp": "2025-07-14T12:35:08.243615"
  },
  {
    "epoch": 2624,
    "batch_idx": 192,
    "loss": 0.12018995732069016,
    "learning_rate": 0.0002215621917559062,
    "timestep_mean": 471.375,
    "epsilon_norm": 4.1350603103637695,
    "prediction_norm": 3.5676627159118652,
    "timestamp": "2025-07-14T12:35:08.248482"
  },
  {
    "epoch": 2625,
    "batch_idx": 0,
    "loss": 0.2859252691268921,
    "learning_rate": 0.00022221488349019944,
    "timestep_mean": 581.03125,
    "epsilon_norm": 8.524669647216797,
    "prediction_norm": 6.859720230102539,
    "timestamp": "2025-07-14T12:35:08.253333"
  },
  {
    "epoch": 2625,
    "batch_idx": 32,
    "loss": 0.16722483932971954,
    "learning_rate": 0.00022221488349019944,
    "timestep_mean": 459.875,
    "epsilon_norm": 7.906796932220459,
    "prediction_norm": 6.560916900634766,
    "timestamp": "2025-07-14T12:35:08.258275"
  },
  {
    "epoch": 2625,
    "batch_idx": 64,
    "loss": 0.15792283415794373,
    "learning_rate": 0.00022221488349019944,
    "timestep_mean": 485.125,
    "epsilon_norm": 7.083012580871582,
    "prediction_norm": 6.290162086486816,
    "timestamp": "2025-07-14T12:35:08.263203"
  },
  {
    "epoch": 2625,
    "batch_idx": 96,
    "loss": 0.1354244351387024,
    "learning_rate": 0.00022221488349019944,
    "timestep_mean": 568.125,
    "epsilon_norm": 7.089130878448486,
    "prediction_norm": 6.159505844116211,
    "timestamp": "2025-07-14T12:35:08.268078"
  },
  {
    "epoch": 2625,
    "batch_idx": 128,
    "loss": 0.22160732746124268,
    "learning_rate": 0.00022221488349019944,
    "timestep_mean": 516.0,
    "epsilon_norm": 7.4144415855407715,
    "prediction_norm": 6.367939472198486,
    "timestamp": "2025-07-14T12:35:08.273100"
  },
  {
    "epoch": 2625,
    "batch_idx": 160,
    "loss": 0.3977041244506836,
    "learning_rate": 0.00022221488349019944,
    "timestep_mean": 478.28125,
    "epsilon_norm": 8.86668586730957,
    "prediction_norm": 7.127101421356201,
    "timestamp": "2025-07-14T12:35:08.277970"
  },
  {
    "epoch": 2625,
    "batch_idx": 192,
    "loss": 0.3164575397968292,
    "learning_rate": 0.00022221488349019944,
    "timestep_mean": 459.5,
    "epsilon_norm": 4.068852424621582,
    "prediction_norm": 2.8520772457122803,
    "timestamp": "2025-07-14T12:35:08.282854"
  },
  {
    "epoch": 2626,
    "batch_idx": 0,
    "loss": 0.3221099376678467,
    "learning_rate": 0.00022286826063165347,
    "timestep_mean": 526.0,
    "epsilon_norm": 8.215266227722168,
    "prediction_norm": 6.702794551849365,
    "timestamp": "2025-07-14T12:35:08.287789"
  },
  {
    "epoch": 2626,
    "batch_idx": 32,
    "loss": 0.32775992155075073,
    "learning_rate": 0.00022286826063165347,
    "timestep_mean": 515.34375,
    "epsilon_norm": 8.008854866027832,
    "prediction_norm": 6.751509666442871,
    "timestamp": "2025-07-14T12:35:08.292804"
  },
  {
    "epoch": 2626,
    "batch_idx": 64,
    "loss": 0.199426531791687,
    "learning_rate": 0.00022286826063165347,
    "timestep_mean": 529.9375,
    "epsilon_norm": 7.922433376312256,
    "prediction_norm": 6.943694591522217,
    "timestamp": "2025-07-14T12:35:08.297767"
  },
  {
    "epoch": 2626,
    "batch_idx": 96,
    "loss": 0.23463594913482666,
    "learning_rate": 0.00022286826063165347,
    "timestep_mean": 598.03125,
    "epsilon_norm": 8.103002548217773,
    "prediction_norm": 6.527592658996582,
    "timestamp": "2025-07-14T12:35:08.302666"
  },
  {
    "epoch": 2626,
    "batch_idx": 128,
    "loss": 0.19930614531040192,
    "learning_rate": 0.00022286826063165347,
    "timestep_mean": 537.5,
    "epsilon_norm": 7.798983573913574,
    "prediction_norm": 6.676182270050049,
    "timestamp": "2025-07-14T12:35:08.307536"
  },
  {
    "epoch": 2626,
    "batch_idx": 160,
    "loss": 0.27714452147483826,
    "learning_rate": 0.00022286826063165347,
    "timestep_mean": 538.65625,
    "epsilon_norm": 7.084444046020508,
    "prediction_norm": 5.699814796447754,
    "timestamp": "2025-07-14T12:35:08.312436"
  },
  {
    "epoch": 2626,
    "batch_idx": 192,
    "loss": 0.10526002943515778,
    "learning_rate": 0.00022286826063165347,
    "timestep_mean": 480.125,
    "epsilon_norm": 4.209239959716797,
    "prediction_norm": 3.756028890609741,
    "timestamp": "2025-07-14T12:35:08.317388"
  },
  {
    "epoch": 2627,
    "batch_idx": 0,
    "loss": 0.23034429550170898,
    "learning_rate": 0.00022352232156812557,
    "timestep_mean": 485.3125,
    "epsilon_norm": 7.1727399826049805,
    "prediction_norm": 6.0658674240112305,
    "timestamp": "2025-07-14T12:35:08.322322"
  },
  {
    "epoch": 2627,
    "batch_idx": 32,
    "loss": 0.3501019775867462,
    "learning_rate": 0.00022352232156812557,
    "timestep_mean": 460.25,
    "epsilon_norm": 7.070601463317871,
    "prediction_norm": 5.233532905578613,
    "timestamp": "2025-07-14T12:35:08.327275"
  },
  {
    "epoch": 2627,
    "batch_idx": 64,
    "loss": 0.11687187850475311,
    "learning_rate": 0.00022352232156812557,
    "timestep_mean": 541.8125,
    "epsilon_norm": 7.431737422943115,
    "prediction_norm": 6.983073711395264,
    "timestamp": "2025-07-14T12:35:08.332166"
  },
  {
    "epoch": 2627,
    "batch_idx": 96,
    "loss": 0.1732606738805771,
    "learning_rate": 0.00022352232156812557,
    "timestep_mean": 468.3125,
    "epsilon_norm": 6.991524696350098,
    "prediction_norm": 6.960409164428711,
    "timestamp": "2025-07-14T12:35:08.337116"
  },
  {
    "epoch": 2627,
    "batch_idx": 128,
    "loss": 0.2750258445739746,
    "learning_rate": 0.00022352232156812557,
    "timestep_mean": 553.65625,
    "epsilon_norm": 8.104042053222656,
    "prediction_norm": 6.922895908355713,
    "timestamp": "2025-07-14T12:35:08.342101"
  },
  {
    "epoch": 2627,
    "batch_idx": 160,
    "loss": 0.23126515746116638,
    "learning_rate": 0.00022352232156812557,
    "timestep_mean": 511.125,
    "epsilon_norm": 7.367361068725586,
    "prediction_norm": 6.900721549987793,
    "timestamp": "2025-07-14T12:35:08.346993"
  },
  {
    "epoch": 2627,
    "batch_idx": 192,
    "loss": 0.2069612592458725,
    "learning_rate": 0.00022352232156812557,
    "timestep_mean": 577.375,
    "epsilon_norm": 3.4281442165374756,
    "prediction_norm": 3.138265371322632,
    "timestamp": "2025-07-14T12:35:08.351862"
  },
  {
    "epoch": 2628,
    "batch_idx": 0,
    "loss": 0.24362462759017944,
    "learning_rate": 0.00022417706468578534,
    "timestep_mean": 454.6875,
    "epsilon_norm": 7.424177646636963,
    "prediction_norm": 6.441556453704834,
    "timestamp": "2025-07-14T12:35:08.356706"
  },
  {
    "epoch": 2628,
    "batch_idx": 32,
    "loss": 0.2860483229160309,
    "learning_rate": 0.00022417706468578534,
    "timestep_mean": 501.46875,
    "epsilon_norm": 8.59549331665039,
    "prediction_norm": 7.554342269897461,
    "timestamp": "2025-07-14T12:35:08.361561"
  },
  {
    "epoch": 2628,
    "batch_idx": 64,
    "loss": 0.18805010616779327,
    "learning_rate": 0.00022417706468578534,
    "timestep_mean": 562.375,
    "epsilon_norm": 7.58079719543457,
    "prediction_norm": 6.7609357833862305,
    "timestamp": "2025-07-14T12:35:08.366591"
  },
  {
    "epoch": 2628,
    "batch_idx": 96,
    "loss": 0.24656686186790466,
    "learning_rate": 0.00022417706468578534,
    "timestep_mean": 392.5625,
    "epsilon_norm": 8.106099128723145,
    "prediction_norm": 6.679067611694336,
    "timestamp": "2025-07-14T12:35:08.371665"
  },
  {
    "epoch": 2628,
    "batch_idx": 128,
    "loss": 0.5729922652244568,
    "learning_rate": 0.00022417706468578534,
    "timestep_mean": 450.90625,
    "epsilon_norm": 8.666045188903809,
    "prediction_norm": 5.934648036956787,
    "timestamp": "2025-07-14T12:35:08.376604"
  },
  {
    "epoch": 2628,
    "batch_idx": 160,
    "loss": 0.3709293305873871,
    "learning_rate": 0.00022417706468578534,
    "timestep_mean": 522.625,
    "epsilon_norm": 8.930462837219238,
    "prediction_norm": 6.292953968048096,
    "timestamp": "2025-07-14T12:35:08.381554"
  },
  {
    "epoch": 2628,
    "batch_idx": 192,
    "loss": 0.7992579340934753,
    "learning_rate": 0.00022417706468578534,
    "timestep_mean": 494.5,
    "epsilon_norm": 2.782992362976074,
    "prediction_norm": 2.4306938648223877,
    "timestamp": "2025-07-14T12:35:08.386517"
  },
  {
    "epoch": 2629,
    "batch_idx": 0,
    "loss": 0.328677773475647,
    "learning_rate": 0.0002248324883691192,
    "timestep_mean": 493.3125,
    "epsilon_norm": 8.36913013458252,
    "prediction_norm": 6.8804802894592285,
    "timestamp": "2025-07-14T12:35:08.391454"
  },
  {
    "epoch": 2629,
    "batch_idx": 32,
    "loss": 0.2996734082698822,
    "learning_rate": 0.0002248324883691192,
    "timestep_mean": 523.0,
    "epsilon_norm": 8.580669403076172,
    "prediction_norm": 7.0168890953063965,
    "timestamp": "2025-07-14T12:35:08.396451"
  },
  {
    "epoch": 2629,
    "batch_idx": 64,
    "loss": 0.34080642461776733,
    "learning_rate": 0.0002248324883691192,
    "timestep_mean": 506.875,
    "epsilon_norm": 7.948028564453125,
    "prediction_norm": 6.718023777008057,
    "timestamp": "2025-07-14T12:35:08.401460"
  },
  {
    "epoch": 2629,
    "batch_idx": 96,
    "loss": 0.32848799228668213,
    "learning_rate": 0.0002248324883691192,
    "timestep_mean": 494.65625,
    "epsilon_norm": 9.015680313110352,
    "prediction_norm": 6.950428485870361,
    "timestamp": "2025-07-14T12:35:08.406353"
  },
  {
    "epoch": 2629,
    "batch_idx": 128,
    "loss": 0.17028522491455078,
    "learning_rate": 0.0002248324883691192,
    "timestep_mean": 467.15625,
    "epsilon_norm": 7.329233169555664,
    "prediction_norm": 6.896050930023193,
    "timestamp": "2025-07-14T12:35:08.411266"
  },
  {
    "epoch": 2629,
    "batch_idx": 160,
    "loss": 0.43025094270706177,
    "learning_rate": 0.0002248324883691192,
    "timestep_mean": 387.65625,
    "epsilon_norm": 7.619575500488281,
    "prediction_norm": 6.562523365020752,
    "timestamp": "2025-07-14T12:35:08.416361"
  },
  {
    "epoch": 2629,
    "batch_idx": 192,
    "loss": 0.3015330135822296,
    "learning_rate": 0.0002248324883691192,
    "timestep_mean": 601.0,
    "epsilon_norm": 3.4396753311157227,
    "prediction_norm": 3.330455780029297,
    "timestamp": "2025-07-14T12:35:08.421578"
  },
  {
    "epoch": 2630,
    "batch_idx": 0,
    "loss": 0.24970635771751404,
    "learning_rate": 0.00022548859100093445,
    "timestep_mean": 555.3125,
    "epsilon_norm": 9.333409309387207,
    "prediction_norm": 8.342268943786621,
    "timestamp": "2025-07-14T12:35:08.426597"
  },
  {
    "epoch": 2630,
    "batch_idx": 32,
    "loss": 0.2088419497013092,
    "learning_rate": 0.00022548859100093445,
    "timestep_mean": 492.875,
    "epsilon_norm": 7.821186065673828,
    "prediction_norm": 6.9353508949279785,
    "timestamp": "2025-07-14T12:35:08.431660"
  },
  {
    "epoch": 2630,
    "batch_idx": 64,
    "loss": 0.1952536702156067,
    "learning_rate": 0.00022548859100093445,
    "timestep_mean": 476.53125,
    "epsilon_norm": 6.2535552978515625,
    "prediction_norm": 5.000917911529541,
    "timestamp": "2025-07-14T12:35:08.436611"
  },
  {
    "epoch": 2630,
    "batch_idx": 96,
    "loss": 0.34724271297454834,
    "learning_rate": 0.00022548859100093445,
    "timestep_mean": 429.96875,
    "epsilon_norm": 8.012699127197266,
    "prediction_norm": 6.2268266677856445,
    "timestamp": "2025-07-14T12:35:08.441627"
  },
  {
    "epoch": 2630,
    "batch_idx": 128,
    "loss": 0.19105437397956848,
    "learning_rate": 0.00022548859100093445,
    "timestep_mean": 466.90625,
    "epsilon_norm": 7.506341457366943,
    "prediction_norm": 6.976118087768555,
    "timestamp": "2025-07-14T12:35:08.446583"
  },
  {
    "epoch": 2630,
    "batch_idx": 160,
    "loss": 0.23302297294139862,
    "learning_rate": 0.00022548859100093445,
    "timestep_mean": 517.78125,
    "epsilon_norm": 6.9546685218811035,
    "prediction_norm": 6.012906551361084,
    "timestamp": "2025-07-14T12:35:08.451539"
  },
  {
    "epoch": 2630,
    "batch_idx": 192,
    "loss": 0.061638399958610535,
    "learning_rate": 0.00022548859100093445,
    "timestep_mean": 488.5,
    "epsilon_norm": 3.8887362480163574,
    "prediction_norm": 3.798902750015259,
    "timestamp": "2025-07-14T12:35:08.456464"
  },
  {
    "epoch": 2631,
    "batch_idx": 0,
    "loss": 0.14611753821372986,
    "learning_rate": 0.00022614537096236304,
    "timestep_mean": 488.125,
    "epsilon_norm": 6.063697338104248,
    "prediction_norm": 5.540060520172119,
    "timestamp": "2025-07-14T12:35:08.461763"
  },
  {
    "epoch": 2631,
    "batch_idx": 32,
    "loss": 0.2428300529718399,
    "learning_rate": 0.00022614537096236304,
    "timestep_mean": 430.46875,
    "epsilon_norm": 7.879294395446777,
    "prediction_norm": 6.472160816192627,
    "timestamp": "2025-07-14T12:35:08.466976"
  },
  {
    "epoch": 2631,
    "batch_idx": 64,
    "loss": 0.24788035452365875,
    "learning_rate": 0.00022614537096236304,
    "timestep_mean": 552.0,
    "epsilon_norm": 7.411891460418701,
    "prediction_norm": 6.5113444328308105,
    "timestamp": "2025-07-14T12:35:08.471935"
  },
  {
    "epoch": 2631,
    "batch_idx": 96,
    "loss": 0.1773921698331833,
    "learning_rate": 0.00022614537096236304,
    "timestep_mean": 531.8125,
    "epsilon_norm": 8.174638748168945,
    "prediction_norm": 6.905652046203613,
    "timestamp": "2025-07-14T12:35:08.477024"
  },
  {
    "epoch": 2631,
    "batch_idx": 128,
    "loss": 0.26241734623908997,
    "learning_rate": 0.00022614537096236304,
    "timestep_mean": 465.84375,
    "epsilon_norm": 8.089801788330078,
    "prediction_norm": 5.779341220855713,
    "timestamp": "2025-07-14T12:35:08.481932"
  },
  {
    "epoch": 2631,
    "batch_idx": 160,
    "loss": 0.37897616624832153,
    "learning_rate": 0.00022614537096236304,
    "timestep_mean": 428.59375,
    "epsilon_norm": 7.786956310272217,
    "prediction_norm": 6.064553260803223,
    "timestamp": "2025-07-14T12:35:08.486893"
  },
  {
    "epoch": 2631,
    "batch_idx": 192,
    "loss": 0.28626716136932373,
    "learning_rate": 0.00022614537096236304,
    "timestep_mean": 453.75,
    "epsilon_norm": 3.6761224269866943,
    "prediction_norm": 3.3988265991210938,
    "timestamp": "2025-07-14T12:35:08.491733"
  },
  {
    "epoch": 2632,
    "batch_idx": 0,
    "loss": 0.22339928150177002,
    "learning_rate": 0.00022680282663286604,
    "timestep_mean": 540.375,
    "epsilon_norm": 9.099827766418457,
    "prediction_norm": 7.410799503326416,
    "timestamp": "2025-07-14T12:35:08.496737"
  },
  {
    "epoch": 2632,
    "batch_idx": 32,
    "loss": 0.3284231424331665,
    "learning_rate": 0.00022680282663286604,
    "timestep_mean": 469.71875,
    "epsilon_norm": 8.885111808776855,
    "prediction_norm": 7.448473930358887,
    "timestamp": "2025-07-14T12:35:08.501627"
  },
  {
    "epoch": 2632,
    "batch_idx": 64,
    "loss": 0.2884589433670044,
    "learning_rate": 0.00022680282663286604,
    "timestep_mean": 485.125,
    "epsilon_norm": 8.390707969665527,
    "prediction_norm": 6.979241847991943,
    "timestamp": "2025-07-14T12:35:08.506538"
  },
  {
    "epoch": 2632,
    "batch_idx": 96,
    "loss": 0.19097822904586792,
    "learning_rate": 0.00022680282663286604,
    "timestep_mean": 523.59375,
    "epsilon_norm": 8.011442184448242,
    "prediction_norm": 7.033374786376953,
    "timestamp": "2025-07-14T12:35:08.511394"
  },
  {
    "epoch": 2632,
    "batch_idx": 128,
    "loss": 0.21102169156074524,
    "learning_rate": 0.00022680282663286604,
    "timestep_mean": 505.03125,
    "epsilon_norm": 7.116857528686523,
    "prediction_norm": 6.204660892486572,
    "timestamp": "2025-07-14T12:35:08.516236"
  },
  {
    "epoch": 2632,
    "batch_idx": 160,
    "loss": 0.13083341717720032,
    "learning_rate": 0.00022680282663286604,
    "timestep_mean": 547.09375,
    "epsilon_norm": 7.637169361114502,
    "prediction_norm": 7.13623571395874,
    "timestamp": "2025-07-14T12:35:08.521120"
  },
  {
    "epoch": 2632,
    "batch_idx": 192,
    "loss": 0.16254115104675293,
    "learning_rate": 0.00022680282663286604,
    "timestep_mean": 519.875,
    "epsilon_norm": 4.14223575592041,
    "prediction_norm": 3.703483819961548,
    "timestamp": "2025-07-14T12:35:08.526101"
  },
  {
    "epoch": 2633,
    "batch_idx": 0,
    "loss": 0.32516926527023315,
    "learning_rate": 0.00022746095639023626,
    "timestep_mean": 519.25,
    "epsilon_norm": 7.2385077476501465,
    "prediction_norm": 6.1228179931640625,
    "timestamp": "2025-07-14T12:35:08.530997"
  },
  {
    "epoch": 2633,
    "batch_idx": 32,
    "loss": 0.22409313917160034,
    "learning_rate": 0.00022746095639023626,
    "timestep_mean": 464.4375,
    "epsilon_norm": 6.348351001739502,
    "prediction_norm": 5.314754486083984,
    "timestamp": "2025-07-14T12:35:08.535858"
  },
  {
    "epoch": 2633,
    "batch_idx": 64,
    "loss": 0.23406100273132324,
    "learning_rate": 0.00022746095639023626,
    "timestep_mean": 459.25,
    "epsilon_norm": 7.549500942230225,
    "prediction_norm": 6.370977401733398,
    "timestamp": "2025-07-14T12:35:08.540766"
  },
  {
    "epoch": 2633,
    "batch_idx": 96,
    "loss": 0.5037074685096741,
    "learning_rate": 0.00022746095639023626,
    "timestep_mean": 528.34375,
    "epsilon_norm": 8.589727401733398,
    "prediction_norm": 6.731612205505371,
    "timestamp": "2025-07-14T12:35:08.545955"
  },
  {
    "epoch": 2633,
    "batch_idx": 128,
    "loss": 0.22196899354457855,
    "learning_rate": 0.00022746095639023626,
    "timestep_mean": 488.71875,
    "epsilon_norm": 8.825445175170898,
    "prediction_norm": 7.494227886199951,
    "timestamp": "2025-07-14T12:35:08.551280"
  },
  {
    "epoch": 2633,
    "batch_idx": 160,
    "loss": 0.23221883177757263,
    "learning_rate": 0.00022746095639023626,
    "timestep_mean": 467.9375,
    "epsilon_norm": 7.796595096588135,
    "prediction_norm": 6.5068464279174805,
    "timestamp": "2025-07-14T12:35:08.556384"
  },
  {
    "epoch": 2633,
    "batch_idx": 192,
    "loss": 0.08356758207082748,
    "learning_rate": 0.00022746095639023626,
    "timestep_mean": 443.125,
    "epsilon_norm": 3.8697891235351562,
    "prediction_norm": 3.451521158218384,
    "timestamp": "2025-07-14T12:35:08.561491"
  },
  {
    "epoch": 2634,
    "batch_idx": 0,
    "loss": 0.11969197541475296,
    "learning_rate": 0.00022811975861060426,
    "timestep_mean": 509.875,
    "epsilon_norm": 7.390944004058838,
    "prediction_norm": 6.632081031799316,
    "timestamp": "2025-07-14T12:35:08.566775"
  },
  {
    "epoch": 2634,
    "batch_idx": 32,
    "loss": 0.12930938601493835,
    "learning_rate": 0.00022811975861060426,
    "timestep_mean": 580.25,
    "epsilon_norm": 6.807537078857422,
    "prediction_norm": 6.225086212158203,
    "timestamp": "2025-07-14T12:35:08.572030"
  },
  {
    "epoch": 2634,
    "batch_idx": 64,
    "loss": 0.5538630485534668,
    "learning_rate": 0.00022811975861060426,
    "timestep_mean": 420.75,
    "epsilon_norm": 7.415377140045166,
    "prediction_norm": 5.8727216720581055,
    "timestamp": "2025-07-14T12:35:08.577256"
  },
  {
    "epoch": 2634,
    "batch_idx": 96,
    "loss": 0.1253853142261505,
    "learning_rate": 0.00022811975861060426,
    "timestep_mean": 540.78125,
    "epsilon_norm": 7.34799337387085,
    "prediction_norm": 6.429063320159912,
    "timestamp": "2025-07-14T12:35:08.582556"
  },
  {
    "epoch": 2634,
    "batch_idx": 128,
    "loss": 0.33352407813072205,
    "learning_rate": 0.00022811975861060426,
    "timestep_mean": 419.9375,
    "epsilon_norm": 8.364142417907715,
    "prediction_norm": 6.773915767669678,
    "timestamp": "2025-07-14T12:35:08.587693"
  },
  {
    "epoch": 2634,
    "batch_idx": 160,
    "loss": 0.2083096206188202,
    "learning_rate": 0.00022811975861060426,
    "timestep_mean": 509.5625,
    "epsilon_norm": 7.06523323059082,
    "prediction_norm": 5.886107444763184,
    "timestamp": "2025-07-14T12:35:08.592776"
  },
  {
    "epoch": 2634,
    "batch_idx": 192,
    "loss": 0.27494385838508606,
    "learning_rate": 0.00022811975861060426,
    "timestep_mean": 478.875,
    "epsilon_norm": 4.044837474822998,
    "prediction_norm": 3.2465147972106934,
    "timestamp": "2025-07-14T12:35:08.597837"
  },
  {
    "epoch": 2635,
    "batch_idx": 0,
    "loss": 0.27367591857910156,
    "learning_rate": 0.00022877923166844067,
    "timestep_mean": 485.21875,
    "epsilon_norm": 7.927872657775879,
    "prediction_norm": 6.750054359436035,
    "timestamp": "2025-07-14T12:35:08.602925"
  },
  {
    "epoch": 2635,
    "batch_idx": 32,
    "loss": 0.24273993074893951,
    "learning_rate": 0.00022877923166844067,
    "timestep_mean": 521.0,
    "epsilon_norm": 6.850299835205078,
    "prediction_norm": 5.9056010246276855,
    "timestamp": "2025-07-14T12:35:08.608141"
  },
  {
    "epoch": 2635,
    "batch_idx": 64,
    "loss": 0.2925616204738617,
    "learning_rate": 0.00022877923166844067,
    "timestep_mean": 512.0625,
    "epsilon_norm": 7.499293804168701,
    "prediction_norm": 6.234986782073975,
    "timestamp": "2025-07-14T12:35:08.613519"
  },
  {
    "epoch": 2635,
    "batch_idx": 96,
    "loss": 0.23320989310741425,
    "learning_rate": 0.00022877923166844067,
    "timestep_mean": 492.78125,
    "epsilon_norm": 7.709217071533203,
    "prediction_norm": 6.556753158569336,
    "timestamp": "2025-07-14T12:35:08.618658"
  },
  {
    "epoch": 2635,
    "batch_idx": 128,
    "loss": 0.3498745560646057,
    "learning_rate": 0.00022877923166844067,
    "timestep_mean": 494.40625,
    "epsilon_norm": 8.352884292602539,
    "prediction_norm": 7.020403861999512,
    "timestamp": "2025-07-14T12:35:08.623698"
  },
  {
    "epoch": 2635,
    "batch_idx": 160,
    "loss": 0.14013417065143585,
    "learning_rate": 0.00022877923166844067,
    "timestep_mean": 534.5,
    "epsilon_norm": 7.314271450042725,
    "prediction_norm": 7.039313793182373,
    "timestamp": "2025-07-14T12:35:08.628794"
  },
  {
    "epoch": 2635,
    "batch_idx": 192,
    "loss": 0.2438930869102478,
    "learning_rate": 0.00022877923166844067,
    "timestep_mean": 616.75,
    "epsilon_norm": 4.184013843536377,
    "prediction_norm": 3.6378183364868164,
    "timestamp": "2025-07-14T12:35:08.633893"
  },
  {
    "epoch": 2636,
    "batch_idx": 0,
    "loss": 0.4129525423049927,
    "learning_rate": 0.00022943937393656252,
    "timestep_mean": 500.65625,
    "epsilon_norm": 7.996117115020752,
    "prediction_norm": 6.703450679779053,
    "timestamp": "2025-07-14T12:35:08.639070"
  },
  {
    "epoch": 2636,
    "batch_idx": 32,
    "loss": 0.3139006793498993,
    "learning_rate": 0.00022943937393656252,
    "timestep_mean": 463.6875,
    "epsilon_norm": 6.991089820861816,
    "prediction_norm": 6.051356315612793,
    "timestamp": "2025-07-14T12:35:08.644275"
  },
  {
    "epoch": 2636,
    "batch_idx": 64,
    "loss": 0.22761954367160797,
    "learning_rate": 0.00022943937393656252,
    "timestep_mean": 455.78125,
    "epsilon_norm": 7.625126838684082,
    "prediction_norm": 6.488024711608887,
    "timestamp": "2025-07-14T12:35:08.649315"
  },
  {
    "epoch": 2636,
    "batch_idx": 96,
    "loss": 0.3953193426132202,
    "learning_rate": 0.00022943937393656252,
    "timestep_mean": 449.71875,
    "epsilon_norm": 7.968552112579346,
    "prediction_norm": 6.337369918823242,
    "timestamp": "2025-07-14T12:35:08.654520"
  },
  {
    "epoch": 2636,
    "batch_idx": 128,
    "loss": 0.2384566366672516,
    "learning_rate": 0.00022943937393656252,
    "timestep_mean": 465.46875,
    "epsilon_norm": 8.732681274414062,
    "prediction_norm": 7.671258449554443,
    "timestamp": "2025-07-14T12:35:08.659651"
  },
  {
    "epoch": 2636,
    "batch_idx": 160,
    "loss": 0.2518836259841919,
    "learning_rate": 0.00022943937393656252,
    "timestep_mean": 496.625,
    "epsilon_norm": 8.844245910644531,
    "prediction_norm": 7.516427516937256,
    "timestamp": "2025-07-14T12:35:08.664941"
  },
  {
    "epoch": 2636,
    "batch_idx": 192,
    "loss": 0.32689642906188965,
    "learning_rate": 0.00022943937393656252,
    "timestep_mean": 458.625,
    "epsilon_norm": 3.9168272018432617,
    "prediction_norm": 3.1397390365600586,
    "timestamp": "2025-07-14T12:35:08.670018"
  },
  {
    "epoch": 2637,
    "batch_idx": 0,
    "loss": 0.16235610842704773,
    "learning_rate": 0.00023010018378613308,
    "timestep_mean": 580.75,
    "epsilon_norm": 8.732378959655762,
    "prediction_norm": 8.062057495117188,
    "timestamp": "2025-07-14T12:35:08.675327"
  },
  {
    "epoch": 2637,
    "batch_idx": 32,
    "loss": 0.204769104719162,
    "learning_rate": 0.00023010018378613308,
    "timestep_mean": 460.8125,
    "epsilon_norm": 6.379058361053467,
    "prediction_norm": 5.490294933319092,
    "timestamp": "2025-07-14T12:35:08.680446"
  },
  {
    "epoch": 2637,
    "batch_idx": 64,
    "loss": 0.2621746361255646,
    "learning_rate": 0.00023010018378613308,
    "timestep_mean": 449.5625,
    "epsilon_norm": 8.10798168182373,
    "prediction_norm": 7.584301948547363,
    "timestamp": "2025-07-14T12:35:08.685666"
  },
  {
    "epoch": 2637,
    "batch_idx": 96,
    "loss": 0.36273419857025146,
    "learning_rate": 0.00023010018378613308,
    "timestep_mean": 492.84375,
    "epsilon_norm": 8.011988639831543,
    "prediction_norm": 6.818753719329834,
    "timestamp": "2025-07-14T12:35:08.690763"
  },
  {
    "epoch": 2637,
    "batch_idx": 128,
    "loss": 0.4822218120098114,
    "learning_rate": 0.00023010018378613308,
    "timestep_mean": 575.8125,
    "epsilon_norm": 9.600595474243164,
    "prediction_norm": 7.875061511993408,
    "timestamp": "2025-07-14T12:35:08.695860"
  },
  {
    "epoch": 2637,
    "batch_idx": 160,
    "loss": 0.36733412742614746,
    "learning_rate": 0.00023010018378613308,
    "timestep_mean": 457.0,
    "epsilon_norm": 8.174757957458496,
    "prediction_norm": 6.280885696411133,
    "timestamp": "2025-07-14T12:35:08.701044"
  },
  {
    "epoch": 2637,
    "batch_idx": 192,
    "loss": 0.2535664439201355,
    "learning_rate": 0.00023010018378613308,
    "timestep_mean": 436.75,
    "epsilon_norm": 4.504487991333008,
    "prediction_norm": 3.833770513534546,
    "timestamp": "2025-07-14T12:35:08.706062"
  },
  {
    "epoch": 2638,
    "batch_idx": 0,
    "loss": 0.46056869626045227,
    "learning_rate": 0.00023076165958667023,
    "timestep_mean": 498.28125,
    "epsilon_norm": 9.579767227172852,
    "prediction_norm": 7.958021640777588,
    "timestamp": "2025-07-14T12:35:08.711163"
  },
  {
    "epoch": 2638,
    "batch_idx": 32,
    "loss": 0.3151835799217224,
    "learning_rate": 0.00023076165958667023,
    "timestep_mean": 495.8125,
    "epsilon_norm": 8.50423526763916,
    "prediction_norm": 7.246349811553955,
    "timestamp": "2025-07-14T12:35:08.716386"
  },
  {
    "epoch": 2638,
    "batch_idx": 64,
    "loss": 0.14370504021644592,
    "learning_rate": 0.00023076165958667023,
    "timestep_mean": 538.0625,
    "epsilon_norm": 7.269535541534424,
    "prediction_norm": 6.938503265380859,
    "timestamp": "2025-07-14T12:35:08.721581"
  },
  {
    "epoch": 2638,
    "batch_idx": 96,
    "loss": 0.3930460810661316,
    "learning_rate": 0.00023076165958667023,
    "timestep_mean": 462.0,
    "epsilon_norm": 7.967893600463867,
    "prediction_norm": 5.950300216674805,
    "timestamp": "2025-07-14T12:35:08.726717"
  },
  {
    "epoch": 2638,
    "batch_idx": 128,
    "loss": 0.27811378240585327,
    "learning_rate": 0.00023076165958667023,
    "timestep_mean": 456.34375,
    "epsilon_norm": 8.32241153717041,
    "prediction_norm": 6.580214023590088,
    "timestamp": "2025-07-14T12:35:08.731909"
  },
  {
    "epoch": 2638,
    "batch_idx": 160,
    "loss": 0.2503672242164612,
    "learning_rate": 0.00023076165958667023,
    "timestep_mean": 516.8125,
    "epsilon_norm": 7.531970024108887,
    "prediction_norm": 6.556264877319336,
    "timestamp": "2025-07-14T12:35:08.737057"
  },
  {
    "epoch": 2638,
    "batch_idx": 192,
    "loss": 0.5590323209762573,
    "learning_rate": 0.00023076165958667023,
    "timestep_mean": 422.875,
    "epsilon_norm": 3.699756383895874,
    "prediction_norm": 3.2481532096862793,
    "timestamp": "2025-07-14T12:35:08.742250"
  },
  {
    "epoch": 2639,
    "batch_idx": 0,
    "loss": 0.18853698670864105,
    "learning_rate": 0.00023142379970604856,
    "timestep_mean": 506.5,
    "epsilon_norm": 7.456322193145752,
    "prediction_norm": 7.027231216430664,
    "timestamp": "2025-07-14T12:35:08.747369"
  },
  {
    "epoch": 2639,
    "batch_idx": 32,
    "loss": 0.20271138846874237,
    "learning_rate": 0.00023142379970604856,
    "timestep_mean": 420.96875,
    "epsilon_norm": 8.028268814086914,
    "prediction_norm": 7.062244415283203,
    "timestamp": "2025-07-14T12:35:08.752459"
  },
  {
    "epoch": 2639,
    "batch_idx": 64,
    "loss": 0.19925051927566528,
    "learning_rate": 0.00023142379970604856,
    "timestep_mean": 527.75,
    "epsilon_norm": 6.795228481292725,
    "prediction_norm": 6.261525630950928,
    "timestamp": "2025-07-14T12:35:08.757702"
  },
  {
    "epoch": 2639,
    "batch_idx": 96,
    "loss": 0.1832362562417984,
    "learning_rate": 0.00023142379970604856,
    "timestep_mean": 459.90625,
    "epsilon_norm": 7.3855180740356445,
    "prediction_norm": 6.481873035430908,
    "timestamp": "2025-07-14T12:35:08.762880"
  },
  {
    "epoch": 2639,
    "batch_idx": 128,
    "loss": 0.15665653347969055,
    "learning_rate": 0.00023142379970604856,
    "timestep_mean": 522.71875,
    "epsilon_norm": 7.818240165710449,
    "prediction_norm": 7.008128643035889,
    "timestamp": "2025-07-14T12:35:08.768032"
  },
  {
    "epoch": 2639,
    "batch_idx": 160,
    "loss": 0.5702298283576965,
    "learning_rate": 0.00023142379970604856,
    "timestep_mean": 383.65625,
    "epsilon_norm": 8.030373573303223,
    "prediction_norm": 5.62046480178833,
    "timestamp": "2025-07-14T12:35:08.773194"
  },
  {
    "epoch": 2639,
    "batch_idx": 192,
    "loss": 0.20830576121807098,
    "learning_rate": 0.00023142379970604856,
    "timestep_mean": 476.0,
    "epsilon_norm": 5.686061382293701,
    "prediction_norm": 5.185127258300781,
    "timestamp": "2025-07-14T12:35:08.778156"
  },
  {
    "epoch": 2640,
    "batch_idx": 0,
    "loss": 0.3385891616344452,
    "learning_rate": 0.00023208660251050189,
    "timestep_mean": 476.1875,
    "epsilon_norm": 7.745277404785156,
    "prediction_norm": 6.281867504119873,
    "timestamp": "2025-07-14T12:35:08.783100"
  },
  {
    "epoch": 2640,
    "batch_idx": 32,
    "loss": 0.5825350284576416,
    "learning_rate": 0.00023208660251050189,
    "timestep_mean": 450.6875,
    "epsilon_norm": 10.20756721496582,
    "prediction_norm": 7.4452128410339355,
    "timestamp": "2025-07-14T12:35:08.788389"
  },
  {
    "epoch": 2640,
    "batch_idx": 64,
    "loss": 0.27340856194496155,
    "learning_rate": 0.00023208660251050189,
    "timestep_mean": 573.09375,
    "epsilon_norm": 7.066315174102783,
    "prediction_norm": 6.1572651863098145,
    "timestamp": "2025-07-14T12:35:08.793429"
  },
  {
    "epoch": 2640,
    "batch_idx": 96,
    "loss": 0.16562937200069427,
    "learning_rate": 0.00023208660251050189,
    "timestep_mean": 471.4375,
    "epsilon_norm": 7.891103267669678,
    "prediction_norm": 7.656499862670898,
    "timestamp": "2025-07-14T12:35:08.798479"
  },
  {
    "epoch": 2640,
    "batch_idx": 128,
    "loss": 0.25685203075408936,
    "learning_rate": 0.00023208660251050189,
    "timestep_mean": 492.5625,
    "epsilon_norm": 7.355466365814209,
    "prediction_norm": 6.087555885314941,
    "timestamp": "2025-07-14T12:35:08.803655"
  },
  {
    "epoch": 2640,
    "batch_idx": 160,
    "loss": 0.2954306900501251,
    "learning_rate": 0.00023208660251050189,
    "timestep_mean": 530.5625,
    "epsilon_norm": 7.932209491729736,
    "prediction_norm": 7.597373008728027,
    "timestamp": "2025-07-14T12:35:08.808883"
  },
  {
    "epoch": 2640,
    "batch_idx": 192,
    "loss": 0.07212062180042267,
    "learning_rate": 0.00023208660251050189,
    "timestep_mean": 514.0,
    "epsilon_norm": 3.0259628295898438,
    "prediction_norm": 2.8129119873046875,
    "timestamp": "2025-07-14T12:35:08.813865"
  },
  {
    "epoch": 2641,
    "batch_idx": 0,
    "loss": 0.13515013456344604,
    "learning_rate": 0.0002327500663646315,
    "timestep_mean": 526.46875,
    "epsilon_norm": 7.865428447723389,
    "prediction_norm": 7.125107288360596,
    "timestamp": "2025-07-14T12:35:08.819049"
  },
  {
    "epoch": 2641,
    "batch_idx": 32,
    "loss": 0.16248422861099243,
    "learning_rate": 0.0002327500663646315,
    "timestep_mean": 482.09375,
    "epsilon_norm": 7.490138530731201,
    "prediction_norm": 6.596179485321045,
    "timestamp": "2025-07-14T12:35:08.824266"
  },
  {
    "epoch": 2641,
    "batch_idx": 64,
    "loss": 0.13768060505390167,
    "learning_rate": 0.0002327500663646315,
    "timestep_mean": 456.21875,
    "epsilon_norm": 6.120216369628906,
    "prediction_norm": 5.90858793258667,
    "timestamp": "2025-07-14T12:35:08.829425"
  },
  {
    "epoch": 2641,
    "batch_idx": 96,
    "loss": 0.22696970403194427,
    "learning_rate": 0.0002327500663646315,
    "timestep_mean": 503.125,
    "epsilon_norm": 8.657533645629883,
    "prediction_norm": 7.935534477233887,
    "timestamp": "2025-07-14T12:35:08.834569"
  },
  {
    "epoch": 2641,
    "batch_idx": 128,
    "loss": 0.38077715039253235,
    "learning_rate": 0.0002327500663646315,
    "timestep_mean": 522.59375,
    "epsilon_norm": 8.346843719482422,
    "prediction_norm": 6.977074146270752,
    "timestamp": "2025-07-14T12:35:08.839503"
  },
  {
    "epoch": 2641,
    "batch_idx": 160,
    "loss": 0.23563933372497559,
    "learning_rate": 0.0002327500663646315,
    "timestep_mean": 471.1875,
    "epsilon_norm": 7.955017566680908,
    "prediction_norm": 6.611258029937744,
    "timestamp": "2025-07-14T12:35:08.847258"
  },
  {
    "epoch": 2641,
    "batch_idx": 192,
    "loss": 0.19464905560016632,
    "learning_rate": 0.0002327500663646315,
    "timestep_mean": 621.125,
    "epsilon_norm": 3.9434032440185547,
    "prediction_norm": 3.2196779251098633,
    "timestamp": "2025-07-14T12:35:08.853848"
  },
  {
    "epoch": 2642,
    "batch_idx": 0,
    "loss": 0.2883949279785156,
    "learning_rate": 0.0002334141896314062,
    "timestep_mean": 489.6875,
    "epsilon_norm": 8.225470542907715,
    "prediction_norm": 7.2960734367370605,
    "timestamp": "2025-07-14T12:35:08.860257"
  },
  {
    "epoch": 2642,
    "batch_idx": 32,
    "loss": 0.36240482330322266,
    "learning_rate": 0.0002334141896314062,
    "timestep_mean": 529.3125,
    "epsilon_norm": 8.509564399719238,
    "prediction_norm": 6.734214782714844,
    "timestamp": "2025-07-14T12:35:08.868544"
  },
  {
    "epoch": 2642,
    "batch_idx": 64,
    "loss": 0.27335676550865173,
    "learning_rate": 0.0002334141896314062,
    "timestep_mean": 505.71875,
    "epsilon_norm": 7.445626258850098,
    "prediction_norm": 6.312917709350586,
    "timestamp": "2025-07-14T12:35:08.876224"
  },
  {
    "epoch": 2642,
    "batch_idx": 96,
    "loss": 0.2583240866661072,
    "learning_rate": 0.0002334141896314062,
    "timestep_mean": 461.71875,
    "epsilon_norm": 7.236215591430664,
    "prediction_norm": 5.8080973625183105,
    "timestamp": "2025-07-14T12:35:08.882655"
  },
  {
    "epoch": 2642,
    "batch_idx": 128,
    "loss": 0.17411299049854279,
    "learning_rate": 0.0002334141896314062,
    "timestep_mean": 443.4375,
    "epsilon_norm": 7.437952518463135,
    "prediction_norm": 6.357672691345215,
    "timestamp": "2025-07-14T12:35:08.888840"
  },
  {
    "epoch": 2642,
    "batch_idx": 160,
    "loss": 0.3275087773799896,
    "learning_rate": 0.0002334141896314062,
    "timestep_mean": 453.71875,
    "epsilon_norm": 7.217880725860596,
    "prediction_norm": 6.017734050750732,
    "timestamp": "2025-07-14T12:35:08.894907"
  },
  {
    "epoch": 2642,
    "batch_idx": 192,
    "loss": 0.569143533706665,
    "learning_rate": 0.0002334141896314062,
    "timestep_mean": 431.125,
    "epsilon_norm": 4.7408671379089355,
    "prediction_norm": 3.3413002490997314,
    "timestamp": "2025-07-14T12:35:08.901104"
  },
  {
    "epoch": 2643,
    "batch_idx": 0,
    "loss": 0.38878491520881653,
    "learning_rate": 0.00023407897067216708,
    "timestep_mean": 528.09375,
    "epsilon_norm": 8.14560317993164,
    "prediction_norm": 6.178668975830078,
    "timestamp": "2025-07-14T12:35:08.906731"
  },
  {
    "epoch": 2643,
    "batch_idx": 32,
    "loss": 0.23626753687858582,
    "learning_rate": 0.00023407897067216708,
    "timestep_mean": 545.65625,
    "epsilon_norm": 8.526900291442871,
    "prediction_norm": 8.207036972045898,
    "timestamp": "2025-07-14T12:35:08.912234"
  },
  {
    "epoch": 2643,
    "batch_idx": 64,
    "loss": 0.334065318107605,
    "learning_rate": 0.00023407897067216708,
    "timestep_mean": 438.3125,
    "epsilon_norm": 7.352119445800781,
    "prediction_norm": 5.709316253662109,
    "timestamp": "2025-07-14T12:35:08.917746"
  },
  {
    "epoch": 2643,
    "batch_idx": 96,
    "loss": 0.4810871183872223,
    "learning_rate": 0.00023407897067216708,
    "timestep_mean": 518.625,
    "epsilon_norm": 7.675450801849365,
    "prediction_norm": 5.7673773765563965,
    "timestamp": "2025-07-14T12:35:08.923355"
  },
  {
    "epoch": 2643,
    "batch_idx": 128,
    "loss": 0.189284086227417,
    "learning_rate": 0.00023407897067216708,
    "timestep_mean": 489.53125,
    "epsilon_norm": 8.792694091796875,
    "prediction_norm": 7.527673244476318,
    "timestamp": "2025-07-14T12:35:08.928869"
  },
  {
    "epoch": 2643,
    "batch_idx": 160,
    "loss": 0.20422983169555664,
    "learning_rate": 0.00023407897067216708,
    "timestep_mean": 506.84375,
    "epsilon_norm": 7.7852606773376465,
    "prediction_norm": 7.587907791137695,
    "timestamp": "2025-07-14T12:35:08.934264"
  },
  {
    "epoch": 2643,
    "batch_idx": 192,
    "loss": 0.3865009546279907,
    "learning_rate": 0.00023407897067216708,
    "timestep_mean": 571.75,
    "epsilon_norm": 5.32975435256958,
    "prediction_norm": 5.081358909606934,
    "timestamp": "2025-07-14T12:35:08.939503"
  },
  {
    "epoch": 2644,
    "batch_idx": 0,
    "loss": 0.2232505977153778,
    "learning_rate": 0.00023474440784663337,
    "timestep_mean": 465.75,
    "epsilon_norm": 7.652193546295166,
    "prediction_norm": 6.996335506439209,
    "timestamp": "2025-07-14T12:35:08.944742"
  },
  {
    "epoch": 2644,
    "batch_idx": 32,
    "loss": 0.21445970237255096,
    "learning_rate": 0.00023474440784663337,
    "timestep_mean": 564.875,
    "epsilon_norm": 6.890057563781738,
    "prediction_norm": 6.212287902832031,
    "timestamp": "2025-07-14T12:35:08.950051"
  },
  {
    "epoch": 2644,
    "batch_idx": 64,
    "loss": 0.16268308460712433,
    "learning_rate": 0.00023474440784663337,
    "timestep_mean": 478.3125,
    "epsilon_norm": 6.9506964683532715,
    "prediction_norm": 6.158447265625,
    "timestamp": "2025-07-14T12:35:08.955853"
  },
  {
    "epoch": 2644,
    "batch_idx": 96,
    "loss": 0.19717290997505188,
    "learning_rate": 0.00023474440784663337,
    "timestep_mean": 422.9375,
    "epsilon_norm": 8.353225708007812,
    "prediction_norm": 7.285224437713623,
    "timestamp": "2025-07-14T12:35:08.961381"
  },
  {
    "epoch": 2644,
    "batch_idx": 128,
    "loss": 0.24145260453224182,
    "learning_rate": 0.00023474440784663337,
    "timestep_mean": 563.65625,
    "epsilon_norm": 8.698119163513184,
    "prediction_norm": 7.488811492919922,
    "timestamp": "2025-07-14T12:35:08.966804"
  },
  {
    "epoch": 2644,
    "batch_idx": 160,
    "loss": 0.2656393051147461,
    "learning_rate": 0.00023474440784663337,
    "timestep_mean": 457.28125,
    "epsilon_norm": 7.323793888092041,
    "prediction_norm": 5.96876859664917,
    "timestamp": "2025-07-14T12:35:08.972218"
  },
  {
    "epoch": 2644,
    "batch_idx": 192,
    "loss": 0.34793123602867126,
    "learning_rate": 0.00023474440784663337,
    "timestep_mean": 330.0,
    "epsilon_norm": 3.304436206817627,
    "prediction_norm": 2.6829562187194824,
    "timestamp": "2025-07-14T12:35:08.977531"
  },
  {
    "epoch": 2645,
    "batch_idx": 0,
    "loss": 0.41042280197143555,
    "learning_rate": 0.00023541049951290502,
    "timestep_mean": 480.21875,
    "epsilon_norm": 7.741564750671387,
    "prediction_norm": 6.050985336303711,
    "timestamp": "2025-07-14T12:35:08.982948"
  },
  {
    "epoch": 2645,
    "batch_idx": 32,
    "loss": 0.2894659638404846,
    "learning_rate": 0.00023541049951290502,
    "timestep_mean": 566.0625,
    "epsilon_norm": 7.805511474609375,
    "prediction_norm": 7.055494785308838,
    "timestamp": "2025-07-14T12:35:08.988202"
  },
  {
    "epoch": 2645,
    "batch_idx": 64,
    "loss": 0.31774160265922546,
    "learning_rate": 0.00023541049951290502,
    "timestep_mean": 560.9375,
    "epsilon_norm": 8.481454849243164,
    "prediction_norm": 7.20985221862793,
    "timestamp": "2025-07-14T12:35:08.993264"
  },
  {
    "epoch": 2645,
    "batch_idx": 96,
    "loss": 0.25887352228164673,
    "learning_rate": 0.00023541049951290502,
    "timestep_mean": 443.03125,
    "epsilon_norm": 8.956202507019043,
    "prediction_norm": 7.249505519866943,
    "timestamp": "2025-07-14T12:35:08.998365"
  },
  {
    "epoch": 2645,
    "batch_idx": 128,
    "loss": 0.19308282434940338,
    "learning_rate": 0.00023541049951290502,
    "timestep_mean": 452.28125,
    "epsilon_norm": 7.396365165710449,
    "prediction_norm": 6.363536834716797,
    "timestamp": "2025-07-14T12:35:09.003644"
  },
  {
    "epoch": 2645,
    "batch_idx": 160,
    "loss": 0.3209594488143921,
    "learning_rate": 0.00023541049951290502,
    "timestep_mean": 512.4375,
    "epsilon_norm": 7.232694149017334,
    "prediction_norm": 5.48622989654541,
    "timestamp": "2025-07-14T12:35:09.008798"
  },
  {
    "epoch": 2645,
    "batch_idx": 192,
    "loss": 0.0956735759973526,
    "learning_rate": 0.00023541049951290502,
    "timestep_mean": 459.5,
    "epsilon_norm": 2.8773701190948486,
    "prediction_norm": 2.8649165630340576,
    "timestamp": "2025-07-14T12:35:09.014030"
  },
  {
    "epoch": 2646,
    "batch_idx": 0,
    "loss": 0.4520857036113739,
    "learning_rate": 0.00023607724402746705,
    "timestep_mean": 470.65625,
    "epsilon_norm": 8.537043571472168,
    "prediction_norm": 6.671285152435303,
    "timestamp": "2025-07-14T12:35:09.019482"
  },
  {
    "epoch": 2646,
    "batch_idx": 32,
    "loss": 0.15461434423923492,
    "learning_rate": 0.00023607724402746705,
    "timestep_mean": 523.21875,
    "epsilon_norm": 8.275461196899414,
    "prediction_norm": 7.428297519683838,
    "timestamp": "2025-07-14T12:35:09.024778"
  },
  {
    "epoch": 2646,
    "batch_idx": 64,
    "loss": 0.27098217606544495,
    "learning_rate": 0.00023607724402746705,
    "timestep_mean": 421.46875,
    "epsilon_norm": 8.075126647949219,
    "prediction_norm": 6.735459804534912,
    "timestamp": "2025-07-14T12:35:09.030167"
  },
  {
    "epoch": 2646,
    "batch_idx": 96,
    "loss": 0.1573699414730072,
    "learning_rate": 0.00023607724402746705,
    "timestep_mean": 579.875,
    "epsilon_norm": 6.9909892082214355,
    "prediction_norm": 6.178641319274902,
    "timestamp": "2025-07-14T12:35:09.035314"
  },
  {
    "epoch": 2646,
    "batch_idx": 128,
    "loss": 0.23656073212623596,
    "learning_rate": 0.00023607724402746705,
    "timestep_mean": 533.5,
    "epsilon_norm": 7.348965167999268,
    "prediction_norm": 6.6046624183654785,
    "timestamp": "2025-07-14T12:35:09.040483"
  },
  {
    "epoch": 2646,
    "batch_idx": 160,
    "loss": 0.21786373853683472,
    "learning_rate": 0.00023607724402746705,
    "timestep_mean": 449.25,
    "epsilon_norm": 7.805365562438965,
    "prediction_norm": 7.003521919250488,
    "timestamp": "2025-07-14T12:35:09.045550"
  },
  {
    "epoch": 2646,
    "batch_idx": 192,
    "loss": 0.031361356377601624,
    "learning_rate": 0.00023607724402746705,
    "timestep_mean": 757.375,
    "epsilon_norm": 4.05674409866333,
    "prediction_norm": 3.885263442993164,
    "timestamp": "2025-07-14T12:35:09.050709"
  },
  {
    "epoch": 2647,
    "batch_idx": 0,
    "loss": 0.3710320293903351,
    "learning_rate": 0.0002367446397451937,
    "timestep_mean": 448.21875,
    "epsilon_norm": 8.730552673339844,
    "prediction_norm": 7.0426225662231445,
    "timestamp": "2025-07-14T12:35:09.055638"
  },
  {
    "epoch": 2647,
    "batch_idx": 32,
    "loss": 0.33022189140319824,
    "learning_rate": 0.0002367446397451937,
    "timestep_mean": 473.4375,
    "epsilon_norm": 8.034051895141602,
    "prediction_norm": 6.356667995452881,
    "timestamp": "2025-07-14T12:35:09.060498"
  },
  {
    "epoch": 2647,
    "batch_idx": 64,
    "loss": 0.39191120862960815,
    "learning_rate": 0.0002367446397451937,
    "timestep_mean": 539.46875,
    "epsilon_norm": 8.51201343536377,
    "prediction_norm": 6.728936672210693,
    "timestamp": "2025-07-14T12:35:09.065377"
  },
  {
    "epoch": 2647,
    "batch_idx": 96,
    "loss": 0.12218444049358368,
    "learning_rate": 0.0002367446397451937,
    "timestep_mean": 550.65625,
    "epsilon_norm": 7.4376373291015625,
    "prediction_norm": 6.8133225440979,
    "timestamp": "2025-07-14T12:35:09.070272"
  },
  {
    "epoch": 2647,
    "batch_idx": 128,
    "loss": 0.236565962433815,
    "learning_rate": 0.0002367446397451937,
    "timestep_mean": 537.03125,
    "epsilon_norm": 10.032585144042969,
    "prediction_norm": 8.511597633361816,
    "timestamp": "2025-07-14T12:35:09.075263"
  },
  {
    "epoch": 2647,
    "batch_idx": 160,
    "loss": 0.3299734592437744,
    "learning_rate": 0.0002367446397451937,
    "timestep_mean": 484.65625,
    "epsilon_norm": 7.901907444000244,
    "prediction_norm": 6.520301342010498,
    "timestamp": "2025-07-14T12:35:09.080252"
  },
  {
    "epoch": 2647,
    "batch_idx": 192,
    "loss": 0.1593743860721588,
    "learning_rate": 0.0002367446397451937,
    "timestep_mean": 602.0,
    "epsilon_norm": 3.9682698249816895,
    "prediction_norm": 3.4333391189575195,
    "timestamp": "2025-07-14T12:35:09.085157"
  },
  {
    "epoch": 2648,
    "batch_idx": 0,
    "loss": 0.19057859480381012,
    "learning_rate": 0.00023741268501935233,
    "timestep_mean": 545.375,
    "epsilon_norm": 8.460593223571777,
    "prediction_norm": 7.869350910186768,
    "timestamp": "2025-07-14T12:35:09.090011"
  },
  {
    "epoch": 2648,
    "batch_idx": 32,
    "loss": 0.1617230772972107,
    "learning_rate": 0.00023741268501935233,
    "timestep_mean": 563.9375,
    "epsilon_norm": 8.769465446472168,
    "prediction_norm": 8.683647155761719,
    "timestamp": "2025-07-14T12:35:09.094977"
  },
  {
    "epoch": 2648,
    "batch_idx": 64,
    "loss": 0.5230091214179993,
    "learning_rate": 0.00023741268501935233,
    "timestep_mean": 462.0625,
    "epsilon_norm": 8.322546005249023,
    "prediction_norm": 5.975551605224609,
    "timestamp": "2025-07-14T12:35:09.099817"
  },
  {
    "epoch": 2648,
    "batch_idx": 96,
    "loss": 0.2333136796951294,
    "learning_rate": 0.00023741268501935233,
    "timestep_mean": 475.75,
    "epsilon_norm": 6.342963695526123,
    "prediction_norm": 5.7551445960998535,
    "timestamp": "2025-07-14T12:35:09.104733"
  },
  {
    "epoch": 2648,
    "batch_idx": 128,
    "loss": 0.4010413885116577,
    "learning_rate": 0.00023741268501935233,
    "timestep_mean": 521.0625,
    "epsilon_norm": 7.6439290046691895,
    "prediction_norm": 6.683835983276367,
    "timestamp": "2025-07-14T12:35:09.109613"
  },
  {
    "epoch": 2648,
    "batch_idx": 160,
    "loss": 0.1396661102771759,
    "learning_rate": 0.00023741268501935233,
    "timestep_mean": 556.5,
    "epsilon_norm": 6.923811912536621,
    "prediction_norm": 6.393426895141602,
    "timestamp": "2025-07-14T12:35:09.114391"
  },
  {
    "epoch": 2648,
    "batch_idx": 192,
    "loss": 0.1856733113527298,
    "learning_rate": 0.00023741268501935233,
    "timestep_mean": 349.5,
    "epsilon_norm": 3.1749587059020996,
    "prediction_norm": 2.561049699783325,
    "timestamp": "2025-07-14T12:35:09.119338"
  },
  {
    "epoch": 2649,
    "batch_idx": 0,
    "loss": 0.18271702527999878,
    "learning_rate": 0.00023808137820160804,
    "timestep_mean": 453.9375,
    "epsilon_norm": 6.8328046798706055,
    "prediction_norm": 6.105884075164795,
    "timestamp": "2025-07-14T12:35:09.124251"
  },
  {
    "epoch": 2649,
    "batch_idx": 32,
    "loss": 0.20705841481685638,
    "learning_rate": 0.00023808137820160804,
    "timestep_mean": 413.53125,
    "epsilon_norm": 6.840617656707764,
    "prediction_norm": 5.474943161010742,
    "timestamp": "2025-07-14T12:35:09.129144"
  },
  {
    "epoch": 2649,
    "batch_idx": 64,
    "loss": 0.23111212253570557,
    "learning_rate": 0.00023808137820160804,
    "timestep_mean": 511.59375,
    "epsilon_norm": 8.260660171508789,
    "prediction_norm": 7.217526435852051,
    "timestamp": "2025-07-14T12:35:09.134072"
  },
  {
    "epoch": 2649,
    "batch_idx": 96,
    "loss": 0.1943122148513794,
    "learning_rate": 0.00023808137820160804,
    "timestep_mean": 502.4375,
    "epsilon_norm": 9.354153633117676,
    "prediction_norm": 8.211369514465332,
    "timestamp": "2025-07-14T12:35:09.138856"
  },
  {
    "epoch": 2649,
    "batch_idx": 128,
    "loss": 0.2733151316642761,
    "learning_rate": 0.00023808137820160804,
    "timestep_mean": 465.71875,
    "epsilon_norm": 8.443878173828125,
    "prediction_norm": 6.599242210388184,
    "timestamp": "2025-07-14T12:35:09.143751"
  },
  {
    "epoch": 2649,
    "batch_idx": 160,
    "loss": 0.1769079566001892,
    "learning_rate": 0.00023808137820160804,
    "timestep_mean": 570.75,
    "epsilon_norm": 8.072502136230469,
    "prediction_norm": 7.439724445343018,
    "timestamp": "2025-07-14T12:35:09.148658"
  },
  {
    "epoch": 2649,
    "batch_idx": 192,
    "loss": 0.08703514188528061,
    "learning_rate": 0.00023808137820160804,
    "timestep_mean": 468.125,
    "epsilon_norm": 3.331357479095459,
    "prediction_norm": 2.908447027206421,
    "timestamp": "2025-07-14T12:35:09.153502"
  },
  {
    "epoch": 2650,
    "batch_idx": 0,
    "loss": 0.4403759241104126,
    "learning_rate": 0.00023875071764202605,
    "timestep_mean": 462.0625,
    "epsilon_norm": 7.969573497772217,
    "prediction_norm": 5.767059326171875,
    "timestamp": "2025-07-14T12:35:09.158345"
  },
  {
    "epoch": 2650,
    "batch_idx": 32,
    "loss": 0.11498142033815384,
    "learning_rate": 0.00023875071764202605,
    "timestep_mean": 533.34375,
    "epsilon_norm": 6.783502101898193,
    "prediction_norm": 6.261068820953369,
    "timestamp": "2025-07-14T12:35:09.163365"
  },
  {
    "epoch": 2650,
    "batch_idx": 64,
    "loss": 0.17424973845481873,
    "learning_rate": 0.00023875071764202605,
    "timestep_mean": 455.59375,
    "epsilon_norm": 6.952246189117432,
    "prediction_norm": 5.837406158447266,
    "timestamp": "2025-07-14T12:35:09.168218"
  },
  {
    "epoch": 2650,
    "batch_idx": 96,
    "loss": 0.25005286931991577,
    "learning_rate": 0.00023875071764202605,
    "timestep_mean": 504.46875,
    "epsilon_norm": 7.467503547668457,
    "prediction_norm": 6.244004249572754,
    "timestamp": "2025-07-14T12:35:09.173100"
  },
  {
    "epoch": 2650,
    "batch_idx": 128,
    "loss": 0.3426434099674225,
    "learning_rate": 0.00023875071764202605,
    "timestep_mean": 474.28125,
    "epsilon_norm": 7.716974258422852,
    "prediction_norm": 6.10995626449585,
    "timestamp": "2025-07-14T12:35:09.177895"
  },
  {
    "epoch": 2650,
    "batch_idx": 160,
    "loss": 0.2920891046524048,
    "learning_rate": 0.00023875071764202605,
    "timestep_mean": 463.0625,
    "epsilon_norm": 6.748651504516602,
    "prediction_norm": 4.809463977813721,
    "timestamp": "2025-07-14T12:35:09.182837"
  },
  {
    "epoch": 2650,
    "batch_idx": 192,
    "loss": 0.0655541643500328,
    "learning_rate": 0.00023875071764202605,
    "timestep_mean": 711.75,
    "epsilon_norm": 3.124168634414673,
    "prediction_norm": 2.453504800796509,
    "timestamp": "2025-07-14T12:35:09.187644"
  },
  {
    "epoch": 2651,
    "batch_idx": 0,
    "loss": 0.2370184063911438,
    "learning_rate": 0.0002394207016890779,
    "timestep_mean": 589.65625,
    "epsilon_norm": 8.026140213012695,
    "prediction_norm": 6.549046039581299,
    "timestamp": "2025-07-14T12:35:09.192574"
  },
  {
    "epoch": 2651,
    "batch_idx": 32,
    "loss": 0.313335120677948,
    "learning_rate": 0.0002394207016890779,
    "timestep_mean": 476.28125,
    "epsilon_norm": 7.171789169311523,
    "prediction_norm": 5.388160705566406,
    "timestamp": "2025-07-14T12:35:09.197467"
  },
  {
    "epoch": 2651,
    "batch_idx": 64,
    "loss": 0.2732090950012207,
    "learning_rate": 0.0002394207016890779,
    "timestep_mean": 417.90625,
    "epsilon_norm": 7.207713603973389,
    "prediction_norm": 6.019779682159424,
    "timestamp": "2025-07-14T12:35:09.202347"
  },
  {
    "epoch": 2651,
    "batch_idx": 96,
    "loss": 0.2639146149158478,
    "learning_rate": 0.0002394207016890779,
    "timestep_mean": 480.71875,
    "epsilon_norm": 8.244611740112305,
    "prediction_norm": 7.1492228507995605,
    "timestamp": "2025-07-14T12:35:09.207174"
  },
  {
    "epoch": 2651,
    "batch_idx": 128,
    "loss": 0.19265973567962646,
    "learning_rate": 0.0002394207016890779,
    "timestep_mean": 558.3125,
    "epsilon_norm": 7.323587417602539,
    "prediction_norm": 6.414564609527588,
    "timestamp": "2025-07-14T12:35:09.212047"
  },
  {
    "epoch": 2651,
    "batch_idx": 160,
    "loss": 0.33908647298812866,
    "learning_rate": 0.0002394207016890779,
    "timestep_mean": 509.5,
    "epsilon_norm": 7.429620742797852,
    "prediction_norm": 6.0424041748046875,
    "timestamp": "2025-07-14T12:35:09.217020"
  },
  {
    "epoch": 2651,
    "batch_idx": 192,
    "loss": 0.13159403204917908,
    "learning_rate": 0.0002394207016890779,
    "timestep_mean": 507.75,
    "epsilon_norm": 3.4273009300231934,
    "prediction_norm": 3.1535916328430176,
    "timestamp": "2025-07-14T12:35:09.221911"
  },
  {
    "epoch": 2652,
    "batch_idx": 0,
    "loss": 0.20391908288002014,
    "learning_rate": 0.0002400913286896453,
    "timestep_mean": 451.53125,
    "epsilon_norm": 6.050333499908447,
    "prediction_norm": 5.46887731552124,
    "timestamp": "2025-07-14T12:35:09.226716"
  },
  {
    "epoch": 2652,
    "batch_idx": 32,
    "loss": 0.30929136276245117,
    "learning_rate": 0.0002400913286896453,
    "timestep_mean": 476.90625,
    "epsilon_norm": 7.658592700958252,
    "prediction_norm": 6.683139801025391,
    "timestamp": "2025-07-14T12:35:09.231551"
  },
  {
    "epoch": 2652,
    "batch_idx": 64,
    "loss": 0.08156071603298187,
    "learning_rate": 0.0002400913286896453,
    "timestep_mean": 564.4375,
    "epsilon_norm": 8.011640548706055,
    "prediction_norm": 7.9844441413879395,
    "timestamp": "2025-07-14T12:35:09.236478"
  },
  {
    "epoch": 2652,
    "batch_idx": 96,
    "loss": 0.4386414885520935,
    "learning_rate": 0.0002400913286896453,
    "timestep_mean": 436.5625,
    "epsilon_norm": 8.983415603637695,
    "prediction_norm": 7.0859222412109375,
    "timestamp": "2025-07-14T12:35:09.241288"
  },
  {
    "epoch": 2652,
    "batch_idx": 128,
    "loss": 0.1992904096841812,
    "learning_rate": 0.0002400913286896453,
    "timestep_mean": 573.1875,
    "epsilon_norm": 7.996476650238037,
    "prediction_norm": 6.890102863311768,
    "timestamp": "2025-07-14T12:35:09.246116"
  },
  {
    "epoch": 2652,
    "batch_idx": 160,
    "loss": 0.4589262008666992,
    "learning_rate": 0.0002400913286896453,
    "timestep_mean": 417.8125,
    "epsilon_norm": 7.704826354980469,
    "prediction_norm": 5.344571590423584,
    "timestamp": "2025-07-14T12:35:09.251071"
  },
  {
    "epoch": 2652,
    "batch_idx": 192,
    "loss": 0.19859294593334198,
    "learning_rate": 0.0002400913286896453,
    "timestep_mean": 455.5,
    "epsilon_norm": 3.5829148292541504,
    "prediction_norm": 3.574159860610962,
    "timestamp": "2025-07-14T12:35:09.255925"
  },
  {
    "epoch": 2653,
    "batch_idx": 0,
    "loss": 0.45105820894241333,
    "learning_rate": 0.00024076259698902275,
    "timestep_mean": 446.71875,
    "epsilon_norm": 8.141158103942871,
    "prediction_norm": 6.280887603759766,
    "timestamp": "2025-07-14T12:35:09.260777"
  },
  {
    "epoch": 2653,
    "batch_idx": 32,
    "loss": 0.11643136292695999,
    "learning_rate": 0.00024076259698902275,
    "timestep_mean": 522.09375,
    "epsilon_norm": 7.121772766113281,
    "prediction_norm": 6.932370185852051,
    "timestamp": "2025-07-14T12:35:09.265653"
  },
  {
    "epoch": 2653,
    "batch_idx": 64,
    "loss": 0.2789496183395386,
    "learning_rate": 0.00024076259698902275,
    "timestep_mean": 459.71875,
    "epsilon_norm": 7.35178804397583,
    "prediction_norm": 6.540693283081055,
    "timestamp": "2025-07-14T12:35:09.270509"
  },
  {
    "epoch": 2653,
    "batch_idx": 96,
    "loss": 0.22816748917102814,
    "learning_rate": 0.00024076259698902275,
    "timestep_mean": 531.4375,
    "epsilon_norm": 7.215235233306885,
    "prediction_norm": 6.941120624542236,
    "timestamp": "2025-07-14T12:35:09.275467"
  },
  {
    "epoch": 2653,
    "batch_idx": 128,
    "loss": 0.2955218255519867,
    "learning_rate": 0.00024076259698902275,
    "timestep_mean": 511.9375,
    "epsilon_norm": 9.044563293457031,
    "prediction_norm": 7.179278373718262,
    "timestamp": "2025-07-14T12:35:09.280365"
  },
  {
    "epoch": 2653,
    "batch_idx": 160,
    "loss": 0.14582645893096924,
    "learning_rate": 0.00024076259698902275,
    "timestep_mean": 497.75,
    "epsilon_norm": 8.561323165893555,
    "prediction_norm": 7.821224212646484,
    "timestamp": "2025-07-14T12:35:09.285270"
  },
  {
    "epoch": 2653,
    "batch_idx": 192,
    "loss": 0.15025927126407623,
    "learning_rate": 0.00024076259698902275,
    "timestep_mean": 418.75,
    "epsilon_norm": 3.0235769748687744,
    "prediction_norm": 2.3538272380828857,
    "timestamp": "2025-07-14T12:35:09.290136"
  },
  {
    "epoch": 2654,
    "batch_idx": 0,
    "loss": 0.18908625841140747,
    "learning_rate": 0.00024143450493092173,
    "timestep_mean": 464.5,
    "epsilon_norm": 8.542791366577148,
    "prediction_norm": 7.350783824920654,
    "timestamp": "2025-07-14T12:35:09.295035"
  },
  {
    "epoch": 2654,
    "batch_idx": 32,
    "loss": 0.2886667847633362,
    "learning_rate": 0.00024143450493092173,
    "timestep_mean": 468.3125,
    "epsilon_norm": 8.193846702575684,
    "prediction_norm": 7.037858486175537,
    "timestamp": "2025-07-14T12:35:09.299835"
  },
  {
    "epoch": 2654,
    "batch_idx": 64,
    "loss": 0.1578119397163391,
    "learning_rate": 0.00024143450493092173,
    "timestep_mean": 565.59375,
    "epsilon_norm": 7.640335559844971,
    "prediction_norm": 7.006779670715332,
    "timestamp": "2025-07-14T12:35:09.304742"
  },
  {
    "epoch": 2654,
    "batch_idx": 96,
    "loss": 0.40332573652267456,
    "learning_rate": 0.00024143450493092173,
    "timestep_mean": 527.875,
    "epsilon_norm": 8.212651252746582,
    "prediction_norm": 6.755824089050293,
    "timestamp": "2025-07-14T12:35:09.309635"
  },
  {
    "epoch": 2654,
    "batch_idx": 128,
    "loss": 0.14982089400291443,
    "learning_rate": 0.00024143450493092173,
    "timestep_mean": 604.28125,
    "epsilon_norm": 8.115066528320312,
    "prediction_norm": 7.652420997619629,
    "timestamp": "2025-07-14T12:35:09.314529"
  },
  {
    "epoch": 2654,
    "batch_idx": 160,
    "loss": 0.21859849989414215,
    "learning_rate": 0.00024143450493092173,
    "timestep_mean": 554.78125,
    "epsilon_norm": 6.755720138549805,
    "prediction_norm": 6.324615001678467,
    "timestamp": "2025-07-14T12:35:09.319348"
  },
  {
    "epoch": 2654,
    "batch_idx": 192,
    "loss": 0.19006232917308807,
    "learning_rate": 0.00024143450493092173,
    "timestep_mean": 335.625,
    "epsilon_norm": 4.15315055847168,
    "prediction_norm": 3.7135963439941406,
    "timestamp": "2025-07-14T12:35:09.324229"
  },
  {
    "epoch": 2655,
    "batch_idx": 0,
    "loss": 0.3649292588233948,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 501.84375,
    "epsilon_norm": 8.004205703735352,
    "prediction_norm": 6.685581207275391,
    "timestamp": "2025-07-14T12:35:09.329123"
  },
  {
    "epoch": 2655,
    "batch_idx": 32,
    "loss": 0.21955694258213043,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 523.53125,
    "epsilon_norm": 8.147526741027832,
    "prediction_norm": 6.952797889709473,
    "timestamp": "2025-07-14T12:35:09.334086"
  },
  {
    "epoch": 2655,
    "batch_idx": 64,
    "loss": 0.14854808151721954,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 536.40625,
    "epsilon_norm": 7.927879810333252,
    "prediction_norm": 7.621185302734375,
    "timestamp": "2025-07-14T12:35:09.338955"
  },
  {
    "epoch": 2655,
    "batch_idx": 96,
    "loss": 0.3801191449165344,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 507.75,
    "epsilon_norm": 8.494396209716797,
    "prediction_norm": 6.775640964508057,
    "timestamp": "2025-07-14T12:35:09.343795"
  },
  {
    "epoch": 2655,
    "batch_idx": 128,
    "loss": 0.18603283166885376,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 489.5,
    "epsilon_norm": 8.819783210754395,
    "prediction_norm": 7.897593021392822,
    "timestamp": "2025-07-14T12:35:09.348616"
  },
  {
    "epoch": 2655,
    "batch_idx": 160,
    "loss": 0.26727575063705444,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 469.0625,
    "epsilon_norm": 7.823031425476074,
    "prediction_norm": 7.253519058227539,
    "timestamp": "2025-07-14T12:35:09.353465"
  },
  {
    "epoch": 2655,
    "batch_idx": 192,
    "loss": 0.7791739106178284,
    "learning_rate": 0.00024210705085747657,
    "timestep_mean": 218.25,
    "epsilon_norm": 5.035378456115723,
    "prediction_norm": 3.4800913333892822,
    "timestamp": "2025-07-14T12:35:09.358276"
  },
  {
    "epoch": 2656,
    "batch_idx": 0,
    "loss": 0.25557005405426025,
    "learning_rate": 0.00024278023310924738,
    "timestep_mean": 551.125,
    "epsilon_norm": 7.1599931716918945,
    "prediction_norm": 6.453965663909912,
    "timestamp": "2025-07-14T12:35:09.363242"
  },
  {
    "epoch": 2656,
    "batch_idx": 32,
    "loss": 0.2726956009864807,
    "learning_rate": 0.00024278023310924738,
    "timestep_mean": 509.875,
    "epsilon_norm": 8.065149307250977,
    "prediction_norm": 6.7168731689453125,
    "timestamp": "2025-07-14T12:35:09.368069"
  },
  {
    "epoch": 2656,
    "batch_idx": 64,
    "loss": 0.25052422285079956,
    "learning_rate": 0.00024278023310924738,
    "timestep_mean": 560.46875,
    "epsilon_norm": 8.289119720458984,
    "prediction_norm": 7.205470561981201,
    "timestamp": "2025-07-14T12:35:09.372902"
  },
  {
    "epoch": 2656,
    "batch_idx": 96,
    "loss": 0.16177956759929657,
    "learning_rate": 0.00024278023310924738,
    "timestep_mean": 506.46875,
    "epsilon_norm": 7.977025032043457,
    "prediction_norm": 7.587975025177002,
    "timestamp": "2025-07-14T12:35:09.377994"
  },
  {
    "epoch": 2656,
    "batch_idx": 128,
    "loss": 0.16908785700798035,
    "learning_rate": 0.00024278023310924738,
    "timestep_mean": 578.40625,
    "epsilon_norm": 8.152379989624023,
    "prediction_norm": 7.160826683044434,
    "timestamp": "2025-07-14T12:35:09.383186"
  },
  {
    "epoch": 2656,
    "batch_idx": 160,
    "loss": 0.18212686479091644,
    "learning_rate": 0.00024278023310924738,
    "timestep_mean": 462.1875,
    "epsilon_norm": 6.824045658111572,
    "prediction_norm": 6.257467269897461,
    "timestamp": "2025-07-14T12:35:09.388543"
  },
  {
    "epoch": 2656,
    "batch_idx": 192,
    "loss": 0.03256198763847351,
    "learning_rate": 0.00024278023310924738,
    "timestep_mean": 653.0,
    "epsilon_norm": 3.0248162746429443,
    "prediction_norm": 3.2230513095855713,
    "timestamp": "2025-07-14T12:35:09.393829"
  },
  {
    "epoch": 2657,
    "batch_idx": 0,
    "loss": 0.3329651355743408,
    "learning_rate": 0.00024345405002522283,
    "timestep_mean": 500.21875,
    "epsilon_norm": 8.945387840270996,
    "prediction_norm": 7.290660858154297,
    "timestamp": "2025-07-14T12:35:09.399042"
  },
  {
    "epoch": 2657,
    "batch_idx": 32,
    "loss": 0.22508826851844788,
    "learning_rate": 0.00024345405002522283,
    "timestep_mean": 453.0625,
    "epsilon_norm": 6.696900844573975,
    "prediction_norm": 5.399508953094482,
    "timestamp": "2025-07-14T12:35:09.404221"
  },
  {
    "epoch": 2657,
    "batch_idx": 64,
    "loss": 0.2023785412311554,
    "learning_rate": 0.00024345405002522283,
    "timestep_mean": 511.375,
    "epsilon_norm": 8.04425048828125,
    "prediction_norm": 7.093931674957275,
    "timestamp": "2025-07-14T12:35:09.409110"
  },
  {
    "epoch": 2657,
    "batch_idx": 96,
    "loss": 0.300657719373703,
    "learning_rate": 0.00024345405002522283,
    "timestep_mean": 452.28125,
    "epsilon_norm": 8.566340446472168,
    "prediction_norm": 7.252604961395264,
    "timestamp": "2025-07-14T12:35:09.414280"
  },
  {
    "epoch": 2657,
    "batch_idx": 128,
    "loss": 0.3182084858417511,
    "learning_rate": 0.00024345405002522283,
    "timestep_mean": 520.875,
    "epsilon_norm": 8.281725883483887,
    "prediction_norm": 7.168179988861084,
    "timestamp": "2025-07-14T12:35:09.419730"
  },
  {
    "epoch": 2657,
    "batch_idx": 160,
    "loss": 0.26110509037971497,
    "learning_rate": 0.00024345405002522283,
    "timestep_mean": 482.28125,
    "epsilon_norm": 7.408390522003174,
    "prediction_norm": 5.861781597137451,
    "timestamp": "2025-07-14T12:35:09.425030"
  },
  {
    "epoch": 2657,
    "batch_idx": 192,
    "loss": 0.049607113003730774,
    "learning_rate": 0.00024345405002522283,
    "timestep_mean": 562.875,
    "epsilon_norm": 4.07003116607666,
    "prediction_norm": 3.5428099632263184,
    "timestamp": "2025-07-14T12:35:09.430400"
  },
  {
    "epoch": 2658,
    "batch_idx": 0,
    "loss": 0.1629890650510788,
    "learning_rate": 0.0002441284999428277,
    "timestep_mean": 489.4375,
    "epsilon_norm": 7.728753089904785,
    "prediction_norm": 6.267138481140137,
    "timestamp": "2025-07-14T12:35:09.435649"
  },
  {
    "epoch": 2658,
    "batch_idx": 32,
    "loss": 0.18678775429725647,
    "learning_rate": 0.0002441284999428277,
    "timestep_mean": 541.21875,
    "epsilon_norm": 6.790212631225586,
    "prediction_norm": 6.3182783126831055,
    "timestamp": "2025-07-14T12:35:09.440899"
  },
  {
    "epoch": 2658,
    "batch_idx": 64,
    "loss": 0.3357657194137573,
    "learning_rate": 0.0002441284999428277,
    "timestep_mean": 499.625,
    "epsilon_norm": 9.099068641662598,
    "prediction_norm": 7.4301042556762695,
    "timestamp": "2025-07-14T12:35:09.446090"
  },
  {
    "epoch": 2658,
    "batch_idx": 96,
    "loss": 0.3890707194805145,
    "learning_rate": 0.0002441284999428277,
    "timestep_mean": 521.625,
    "epsilon_norm": 8.099800109863281,
    "prediction_norm": 5.670412063598633,
    "timestamp": "2025-07-14T12:35:09.451253"
  },
  {
    "epoch": 2658,
    "batch_idx": 128,
    "loss": 0.3615701198577881,
    "learning_rate": 0.0002441284999428277,
    "timestep_mean": 521.96875,
    "epsilon_norm": 7.443570613861084,
    "prediction_norm": 6.4727463722229,
    "timestamp": "2025-07-14T12:35:09.456402"
  },
  {
    "epoch": 2658,
    "batch_idx": 160,
    "loss": 0.3064177632331848,
    "learning_rate": 0.0002441284999428277,
    "timestep_mean": 482.0625,
    "epsilon_norm": 8.73182487487793,
    "prediction_norm": 7.600752830505371,
    "timestamp": "2025-07-14T12:35:09.461607"
  },
  {
    "epoch": 2658,
    "batch_idx": 192,
    "loss": 0.4097107946872711,
    "learning_rate": 0.0002441284999428277,
    "timestep_mean": 454.0,
    "epsilon_norm": 4.537874698638916,
    "prediction_norm": 3.4089417457580566,
    "timestamp": "2025-07-14T12:35:09.466755"
  },
  {
    "epoch": 2659,
    "batch_idx": 0,
    "loss": 0.21045418083667755,
    "learning_rate": 0.00024480358119792354,
    "timestep_mean": 465.4375,
    "epsilon_norm": 7.238684177398682,
    "prediction_norm": 6.421060562133789,
    "timestamp": "2025-07-14T12:35:09.472109"
  },
  {
    "epoch": 2659,
    "batch_idx": 32,
    "loss": 0.48701679706573486,
    "learning_rate": 0.00024480358119792354,
    "timestep_mean": 485.875,
    "epsilon_norm": 7.321594715118408,
    "prediction_norm": 5.6363844871521,
    "timestamp": "2025-07-14T12:35:09.477105"
  },
  {
    "epoch": 2659,
    "batch_idx": 64,
    "loss": 0.248650923371315,
    "learning_rate": 0.00024480358119792354,
    "timestep_mean": 487.25,
    "epsilon_norm": 8.282910346984863,
    "prediction_norm": 6.724541664123535,
    "timestamp": "2025-07-14T12:35:09.482152"
  },
  {
    "epoch": 2659,
    "batch_idx": 96,
    "loss": 0.18113011121749878,
    "learning_rate": 0.00024480358119792354,
    "timestep_mean": 596.75,
    "epsilon_norm": 8.24117660522461,
    "prediction_norm": 7.937979221343994,
    "timestamp": "2025-07-14T12:35:09.487403"
  },
  {
    "epoch": 2659,
    "batch_idx": 128,
    "loss": 0.20936237275600433,
    "learning_rate": 0.00024480358119792354,
    "timestep_mean": 446.75,
    "epsilon_norm": 8.054435729980469,
    "prediction_norm": 7.288615703582764,
    "timestamp": "2025-07-14T12:35:09.492634"
  },
  {
    "epoch": 2659,
    "batch_idx": 160,
    "loss": 0.24474473297595978,
    "learning_rate": 0.00024480358119792354,
    "timestep_mean": 465.8125,
    "epsilon_norm": 8.033116340637207,
    "prediction_norm": 6.82771110534668,
    "timestamp": "2025-07-14T12:35:09.497854"
  },
  {
    "epoch": 2659,
    "batch_idx": 192,
    "loss": 0.15809306502342224,
    "learning_rate": 0.00024480358119792354,
    "timestep_mean": 495.375,
    "epsilon_norm": 3.6114015579223633,
    "prediction_norm": 3.6198670864105225,
    "timestamp": "2025-07-14T12:35:09.503148"
  },
  {
    "epoch": 2660,
    "batch_idx": 0,
    "loss": 0.31892329454421997,
    "learning_rate": 0.00024547929212481484,
    "timestep_mean": 424.5,
    "epsilon_norm": 7.793560981750488,
    "prediction_norm": 5.525451183319092,
    "timestamp": "2025-07-14T12:35:09.508397"
  },
  {
    "epoch": 2660,
    "batch_idx": 32,
    "loss": 0.340068519115448,
    "learning_rate": 0.00024547929212481484,
    "timestep_mean": 475.9375,
    "epsilon_norm": 8.11895751953125,
    "prediction_norm": 6.950584888458252,
    "timestamp": "2025-07-14T12:35:09.513722"
  },
  {
    "epoch": 2660,
    "batch_idx": 64,
    "loss": 0.27918756008148193,
    "learning_rate": 0.00024547929212481484,
    "timestep_mean": 535.5,
    "epsilon_norm": 8.25507640838623,
    "prediction_norm": 6.9045844078063965,
    "timestamp": "2025-07-14T12:35:09.518925"
  },
  {
    "epoch": 2660,
    "batch_idx": 96,
    "loss": 0.20416182279586792,
    "learning_rate": 0.00024547929212481484,
    "timestep_mean": 423.78125,
    "epsilon_norm": 6.2265095710754395,
    "prediction_norm": 5.28663969039917,
    "timestamp": "2025-07-14T12:35:09.524175"
  },
  {
    "epoch": 2660,
    "batch_idx": 128,
    "loss": 0.15366005897521973,
    "learning_rate": 0.00024547929212481484,
    "timestep_mean": 612.59375,
    "epsilon_norm": 7.6266889572143555,
    "prediction_norm": 6.72127103805542,
    "timestamp": "2025-07-14T12:35:09.529231"
  },
  {
    "epoch": 2660,
    "batch_idx": 160,
    "loss": 0.19479887187480927,
    "learning_rate": 0.00024547929212481484,
    "timestep_mean": 436.78125,
    "epsilon_norm": 6.8708319664001465,
    "prediction_norm": 6.5710272789001465,
    "timestamp": "2025-07-14T12:35:09.534074"
  },
  {
    "epoch": 2660,
    "batch_idx": 192,
    "loss": 0.3111024498939514,
    "learning_rate": 0.00024547929212481484,
    "timestep_mean": 318.125,
    "epsilon_norm": 4.262676239013672,
    "prediction_norm": 3.028123617172241,
    "timestamp": "2025-07-14T12:35:09.538911"
  },
  {
    "epoch": 2661,
    "batch_idx": 0,
    "loss": 0.2262372225522995,
    "learning_rate": 0.00024615563105625127,
    "timestep_mean": 500.71875,
    "epsilon_norm": 7.027482032775879,
    "prediction_norm": 6.151226997375488,
    "timestamp": "2025-07-14T12:35:09.543822"
  },
  {
    "epoch": 2661,
    "batch_idx": 32,
    "loss": 0.3794591426849365,
    "learning_rate": 0.00024615563105625127,
    "timestep_mean": 462.8125,
    "epsilon_norm": 8.977940559387207,
    "prediction_norm": 7.3974785804748535,
    "timestamp": "2025-07-14T12:35:09.548733"
  },
  {
    "epoch": 2661,
    "batch_idx": 64,
    "loss": 0.20419485867023468,
    "learning_rate": 0.00024615563105625127,
    "timestep_mean": 439.25,
    "epsilon_norm": 6.974889755249023,
    "prediction_norm": 6.017406463623047,
    "timestamp": "2025-07-14T12:35:09.553761"
  },
  {
    "epoch": 2661,
    "batch_idx": 96,
    "loss": 0.4463586211204529,
    "learning_rate": 0.00024615563105625127,
    "timestep_mean": 496.0,
    "epsilon_norm": 7.70155668258667,
    "prediction_norm": 6.002379894256592,
    "timestamp": "2025-07-14T12:35:09.559015"
  },
  {
    "epoch": 2661,
    "batch_idx": 128,
    "loss": 0.30476969480514526,
    "learning_rate": 0.00024615563105625127,
    "timestep_mean": 545.625,
    "epsilon_norm": 8.361854553222656,
    "prediction_norm": 7.067180633544922,
    "timestamp": "2025-07-14T12:35:09.564166"
  },
  {
    "epoch": 2661,
    "batch_idx": 160,
    "loss": 0.18482579290866852,
    "learning_rate": 0.00024615563105625127,
    "timestep_mean": 565.78125,
    "epsilon_norm": 7.437371253967285,
    "prediction_norm": 6.4815545082092285,
    "timestamp": "2025-07-14T12:35:09.569383"
  },
  {
    "epoch": 2661,
    "batch_idx": 192,
    "loss": 0.10639575868844986,
    "learning_rate": 0.00024615563105625127,
    "timestep_mean": 612.875,
    "epsilon_norm": 4.0074462890625,
    "prediction_norm": 4.115122318267822,
    "timestamp": "2025-07-14T12:35:09.574482"
  },
  {
    "epoch": 2662,
    "batch_idx": 0,
    "loss": 0.0994989275932312,
    "learning_rate": 0.00024683259632343377,
    "timestep_mean": 538.28125,
    "epsilon_norm": 7.260957717895508,
    "prediction_norm": 7.0084733963012695,
    "timestamp": "2025-07-14T12:35:09.579628"
  },
  {
    "epoch": 2662,
    "batch_idx": 32,
    "loss": 0.5205451250076294,
    "learning_rate": 0.00024683259632343377,
    "timestep_mean": 517.125,
    "epsilon_norm": 7.713242053985596,
    "prediction_norm": 6.090980052947998,
    "timestamp": "2025-07-14T12:35:09.584777"
  },
  {
    "epoch": 2662,
    "batch_idx": 64,
    "loss": 0.2700825333595276,
    "learning_rate": 0.00024683259632343377,
    "timestep_mean": 522.78125,
    "epsilon_norm": 8.060301780700684,
    "prediction_norm": 7.127340316772461,
    "timestamp": "2025-07-14T12:35:09.589921"
  },
  {
    "epoch": 2662,
    "batch_idx": 96,
    "loss": 0.551819920539856,
    "learning_rate": 0.00024683259632343377,
    "timestep_mean": 487.875,
    "epsilon_norm": 7.695342540740967,
    "prediction_norm": 5.187455654144287,
    "timestamp": "2025-07-14T12:35:09.595094"
  },
  {
    "epoch": 2662,
    "batch_idx": 128,
    "loss": 0.39311712980270386,
    "learning_rate": 0.00024683259632343377,
    "timestep_mean": 463.28125,
    "epsilon_norm": 8.011892318725586,
    "prediction_norm": 6.043863773345947,
    "timestamp": "2025-07-14T12:35:09.600387"
  },
  {
    "epoch": 2662,
    "batch_idx": 160,
    "loss": 0.1514432579278946,
    "learning_rate": 0.00024683259632343377,
    "timestep_mean": 564.75,
    "epsilon_norm": 7.477052211761475,
    "prediction_norm": 6.5496439933776855,
    "timestamp": "2025-07-14T12:35:09.605536"
  },
  {
    "epoch": 2662,
    "batch_idx": 192,
    "loss": 0.24995307624340057,
    "learning_rate": 0.00024683259632343377,
    "timestep_mean": 615.0,
    "epsilon_norm": 4.6751275062561035,
    "prediction_norm": 4.468510627746582,
    "timestamp": "2025-07-14T12:35:09.610574"
  },
  {
    "epoch": 2663,
    "batch_idx": 0,
    "loss": 0.5310746431350708,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 473.78125,
    "epsilon_norm": 8.772761344909668,
    "prediction_norm": 6.902012348175049,
    "timestamp": "2025-07-14T12:35:09.615732"
  },
  {
    "epoch": 2663,
    "batch_idx": 32,
    "loss": 0.253194659948349,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 513.25,
    "epsilon_norm": 8.090017318725586,
    "prediction_norm": 6.3512115478515625,
    "timestamp": "2025-07-14T12:35:09.620831"
  },
  {
    "epoch": 2663,
    "batch_idx": 64,
    "loss": 0.29091185331344604,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 444.21875,
    "epsilon_norm": 8.30327320098877,
    "prediction_norm": 6.4021124839782715,
    "timestamp": "2025-07-14T12:35:09.626021"
  },
  {
    "epoch": 2663,
    "batch_idx": 96,
    "loss": 0.3141031861305237,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 450.21875,
    "epsilon_norm": 8.42668628692627,
    "prediction_norm": 7.2978973388671875,
    "timestamp": "2025-07-14T12:35:09.630978"
  },
  {
    "epoch": 2663,
    "batch_idx": 128,
    "loss": 0.24746167659759521,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 434.40625,
    "epsilon_norm": 7.09684419631958,
    "prediction_norm": 6.244515419006348,
    "timestamp": "2025-07-14T12:35:09.636062"
  },
  {
    "epoch": 2663,
    "batch_idx": 160,
    "loss": 0.18387317657470703,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 542.84375,
    "epsilon_norm": 7.1912522315979,
    "prediction_norm": 6.002761363983154,
    "timestamp": "2025-07-14T12:35:09.641271"
  },
  {
    "epoch": 2663,
    "batch_idx": 192,
    "loss": 0.5719783306121826,
    "learning_rate": 0.00024751018625601896,
    "timestep_mean": 536.75,
    "epsilon_norm": 4.897846698760986,
    "prediction_norm": 3.5322630405426025,
    "timestamp": "2025-07-14T12:35:09.646373"
  },
  {
    "epoch": 2664,
    "batch_idx": 0,
    "loss": 0.3865174651145935,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 448.65625,
    "epsilon_norm": 8.954009056091309,
    "prediction_norm": 7.414597511291504,
    "timestamp": "2025-07-14T12:35:09.651576"
  },
  {
    "epoch": 2664,
    "batch_idx": 32,
    "loss": 0.23767855763435364,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 554.25,
    "epsilon_norm": 6.133820056915283,
    "prediction_norm": 5.411640644073486,
    "timestamp": "2025-07-14T12:35:09.656758"
  },
  {
    "epoch": 2664,
    "batch_idx": 64,
    "loss": 0.32215616106987,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 487.625,
    "epsilon_norm": 7.026090621948242,
    "prediction_norm": 6.016361713409424,
    "timestamp": "2025-07-14T12:35:09.661933"
  },
  {
    "epoch": 2664,
    "batch_idx": 96,
    "loss": 0.29563289880752563,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 525.59375,
    "epsilon_norm": 8.312374114990234,
    "prediction_norm": 7.229063034057617,
    "timestamp": "2025-07-14T12:35:09.667301"
  },
  {
    "epoch": 2664,
    "batch_idx": 128,
    "loss": 0.27665597200393677,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 338.5625,
    "epsilon_norm": 7.155155658721924,
    "prediction_norm": 5.628351211547852,
    "timestamp": "2025-07-14T12:35:09.672465"
  },
  {
    "epoch": 2664,
    "batch_idx": 160,
    "loss": 0.40764567255973816,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 578.59375,
    "epsilon_norm": 8.334010124206543,
    "prediction_norm": 7.353503227233887,
    "timestamp": "2025-07-14T12:35:09.677621"
  },
  {
    "epoch": 2664,
    "batch_idx": 192,
    "loss": 0.13147515058517456,
    "learning_rate": 0.0002481883991821199,
    "timestep_mean": 527.25,
    "epsilon_norm": 5.896798610687256,
    "prediction_norm": 5.66804838180542,
    "timestamp": "2025-07-14T12:35:09.682715"
  },
  {
    "epoch": 2665,
    "batch_idx": 0,
    "loss": 0.2927286624908447,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 528.6875,
    "epsilon_norm": 9.488201141357422,
    "prediction_norm": 8.459355354309082,
    "timestamp": "2025-07-14T12:35:09.687966"
  },
  {
    "epoch": 2665,
    "batch_idx": 32,
    "loss": 0.28502699732780457,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 489.90625,
    "epsilon_norm": 6.852500915527344,
    "prediction_norm": 5.815589427947998,
    "timestamp": "2025-07-14T12:35:09.693184"
  },
  {
    "epoch": 2665,
    "batch_idx": 64,
    "loss": 0.45440778136253357,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 488.84375,
    "epsilon_norm": 7.934108257293701,
    "prediction_norm": 6.485008239746094,
    "timestamp": "2025-07-14T12:35:09.698436"
  },
  {
    "epoch": 2665,
    "batch_idx": 96,
    "loss": 0.3107770085334778,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 496.90625,
    "epsilon_norm": 8.326984405517578,
    "prediction_norm": 6.7075724601745605,
    "timestamp": "2025-07-14T12:35:09.703634"
  },
  {
    "epoch": 2665,
    "batch_idx": 128,
    "loss": 0.31086453795433044,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 503.0625,
    "epsilon_norm": 7.595461845397949,
    "prediction_norm": 6.0203680992126465,
    "timestamp": "2025-07-14T12:35:09.708941"
  },
  {
    "epoch": 2665,
    "batch_idx": 160,
    "loss": 0.3921341300010681,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 451.53125,
    "epsilon_norm": 8.121318817138672,
    "prediction_norm": 6.7951226234436035,
    "timestamp": "2025-07-14T12:35:09.714123"
  },
  {
    "epoch": 2665,
    "batch_idx": 192,
    "loss": 0.16990037262439728,
    "learning_rate": 0.00024886723342831406,
    "timestep_mean": 536.0,
    "epsilon_norm": 3.858419895172119,
    "prediction_norm": 3.8570961952209473,
    "timestamp": "2025-07-14T12:35:09.719270"
  },
  {
    "epoch": 2666,
    "batch_idx": 0,
    "loss": 0.2670576870441437,
    "learning_rate": 0.0002495466873196457,
    "timestep_mean": 569.6875,
    "epsilon_norm": 8.054007530212402,
    "prediction_norm": 7.397524833679199,
    "timestamp": "2025-07-14T12:35:09.724463"
  },
  {
    "epoch": 2666,
    "batch_idx": 32,
    "loss": 0.2654135823249817,
    "learning_rate": 0.0002495466873196457,
    "timestep_mean": 546.03125,
    "epsilon_norm": 8.309099197387695,
    "prediction_norm": 6.866293907165527,
    "timestamp": "2025-07-14T12:35:09.729657"
  },
  {
    "epoch": 2666,
    "batch_idx": 64,
    "loss": 0.09924455732107162,
    "learning_rate": 0.0002495466873196457,
    "timestep_mean": 583.1875,
    "epsilon_norm": 7.5833258628845215,
    "prediction_norm": 7.739071846008301,
    "timestamp": "2025-07-14T12:35:09.734872"
  },
  {
    "epoch": 2666,
    "batch_idx": 96,
    "loss": 0.2304304540157318,
    "learning_rate": 0.0002495466873196457,
    "timestep_mean": 520.21875,
    "epsilon_norm": 7.14033317565918,
    "prediction_norm": 6.293220520019531,
    "timestamp": "2025-07-14T12:35:09.740065"
  },
  {
    "epoch": 2666,
    "batch_idx": 128,
    "loss": 0.3542332947254181,
    "learning_rate": 0.0002495466873196457,
    "timestep_mean": 498.46875,
    "epsilon_norm": 8.747846603393555,
    "prediction_norm": 7.407313346862793,
    "timestamp": "2025-07-14T12:35:09.745195"
  },
  {
    "epoch": 2666,
    "batch_idx": 160,
    "loss": 0.22663894295692444,
    "learning_rate": 0.0002495466873196457,
    "timestep_mean": 615.9375,
    "epsilon_norm": 7.591371059417725,
    "prediction_norm": 6.318873882293701,
    "timestamp": "2025-07-14T12:35:09.750495"
  },
  {
    "epoch": 2666,
    "batch_idx": 192,
    "loss": 0.41748329997062683,
    "learning_rate": 0.0002495466873196457,
    "timestep_mean": 605.375,
    "epsilon_norm": 4.82913875579834,
    "prediction_norm": 4.674259185791016,
    "timestamp": "2025-07-14T12:35:09.755669"
  },
  {
    "epoch": 2667,
    "batch_idx": 0,
    "loss": 0.24152728915214539,
    "learning_rate": 0.00025022675917962925,
    "timestep_mean": 507.6875,
    "epsilon_norm": 8.452825546264648,
    "prediction_norm": 7.192758560180664,
    "timestamp": "2025-07-14T12:35:09.760797"
  },
  {
    "epoch": 2667,
    "batch_idx": 32,
    "loss": 0.25581812858581543,
    "learning_rate": 0.00025022675917962925,
    "timestep_mean": 479.09375,
    "epsilon_norm": 8.123329162597656,
    "prediction_norm": 6.704435348510742,
    "timestamp": "2025-07-14T12:35:09.765637"
  },
  {
    "epoch": 2667,
    "batch_idx": 64,
    "loss": 0.29885464906692505,
    "learning_rate": 0.00025022675917962925,
    "timestep_mean": 376.875,
    "epsilon_norm": 8.619481086730957,
    "prediction_norm": 6.891423225402832,
    "timestamp": "2025-07-14T12:35:09.770708"
  },
  {
    "epoch": 2667,
    "batch_idx": 96,
    "loss": 0.20137566328048706,
    "learning_rate": 0.00025022675917962925,
    "timestep_mean": 476.09375,
    "epsilon_norm": 6.978641033172607,
    "prediction_norm": 6.145742416381836,
    "timestamp": "2025-07-14T12:35:09.775875"
  },
  {
    "epoch": 2667,
    "batch_idx": 128,
    "loss": 0.3651197850704193,
    "learning_rate": 0.00025022675917962925,
    "timestep_mean": 433.375,
    "epsilon_norm": 8.223230361938477,
    "prediction_norm": 7.024078845977783,
    "timestamp": "2025-07-14T12:35:09.780739"
  },
  {
    "epoch": 2667,
    "batch_idx": 160,
    "loss": 0.3967292904853821,
    "learning_rate": 0.00025022675917962925,
    "timestep_mean": 551.15625,
    "epsilon_norm": 8.859187126159668,
    "prediction_norm": 6.81881046295166,
    "timestamp": "2025-07-14T12:35:09.785631"
  },
  {
    "epoch": 2667,
    "batch_idx": 192,
    "loss": 0.12159957736730576,
    "learning_rate": 0.00025022675917962925,
    "timestep_mean": 466.375,
    "epsilon_norm": 4.117712497711182,
    "prediction_norm": 3.552664279937744,
    "timestamp": "2025-07-14T12:35:09.790502"
  },
  {
    "epoch": 2668,
    "batch_idx": 0,
    "loss": 0.2894207239151001,
    "learning_rate": 0.0002509074473302548,
    "timestep_mean": 538.0,
    "epsilon_norm": 9.405790328979492,
    "prediction_norm": 8.165678024291992,
    "timestamp": "2025-07-14T12:35:09.795378"
  },
  {
    "epoch": 2668,
    "batch_idx": 32,
    "loss": 0.29791557788848877,
    "learning_rate": 0.0002509074473302548,
    "timestep_mean": 457.53125,
    "epsilon_norm": 8.317778587341309,
    "prediction_norm": 6.487197399139404,
    "timestamp": "2025-07-14T12:35:09.800286"
  },
  {
    "epoch": 2668,
    "batch_idx": 64,
    "loss": 0.25615617632865906,
    "learning_rate": 0.0002509074473302548,
    "timestep_mean": 548.59375,
    "epsilon_norm": 8.774917602539062,
    "prediction_norm": 7.8429155349731445,
    "timestamp": "2025-07-14T12:35:09.805192"
  },
  {
    "epoch": 2668,
    "batch_idx": 96,
    "loss": 0.2363138198852539,
    "learning_rate": 0.0002509074473302548,
    "timestep_mean": 474.71875,
    "epsilon_norm": 8.015020370483398,
    "prediction_norm": 6.743356704711914,
    "timestamp": "2025-07-14T12:35:09.810056"
  },
  {
    "epoch": 2668,
    "batch_idx": 128,
    "loss": 0.4048398733139038,
    "learning_rate": 0.0002509074473302548,
    "timestep_mean": 512.9375,
    "epsilon_norm": 8.640499114990234,
    "prediction_norm": 7.005371570587158,
    "timestamp": "2025-07-14T12:35:09.814942"
  },
  {
    "epoch": 2668,
    "batch_idx": 160,
    "loss": 0.22173118591308594,
    "learning_rate": 0.0002509074473302548,
    "timestep_mean": 446.375,
    "epsilon_norm": 7.624094009399414,
    "prediction_norm": 6.56551456451416,
    "timestamp": "2025-07-14T12:35:09.819857"
  },
  {
    "epoch": 2668,
    "batch_idx": 192,
    "loss": 0.06300456076860428,
    "learning_rate": 0.0002509074473302548,
    "timestep_mean": 541.625,
    "epsilon_norm": 3.5726635456085205,
    "prediction_norm": 3.864743232727051,
    "timestamp": "2025-07-14T12:35:09.824679"
  },
  {
    "epoch": 2669,
    "batch_idx": 0,
    "loss": 0.3169128894805908,
    "learning_rate": 0.00025158875009199293,
    "timestep_mean": 511.625,
    "epsilon_norm": 8.56135368347168,
    "prediction_norm": 6.92836856842041,
    "timestamp": "2025-07-14T12:35:09.829572"
  },
  {
    "epoch": 2669,
    "batch_idx": 32,
    "loss": 0.4805067181587219,
    "learning_rate": 0.00025158875009199293,
    "timestep_mean": 535.375,
    "epsilon_norm": 9.449810981750488,
    "prediction_norm": 7.940526485443115,
    "timestamp": "2025-07-14T12:35:09.834403"
  },
  {
    "epoch": 2669,
    "batch_idx": 64,
    "loss": 0.14789485931396484,
    "learning_rate": 0.00025158875009199293,
    "timestep_mean": 500.6875,
    "epsilon_norm": 7.726297855377197,
    "prediction_norm": 7.117215156555176,
    "timestamp": "2025-07-14T12:35:09.839248"
  },
  {
    "epoch": 2669,
    "batch_idx": 96,
    "loss": 0.27625834941864014,
    "learning_rate": 0.00025158875009199293,
    "timestep_mean": 561.875,
    "epsilon_norm": 8.527228355407715,
    "prediction_norm": 7.996889114379883,
    "timestamp": "2025-07-14T12:35:09.844054"
  },
  {
    "epoch": 2669,
    "batch_idx": 128,
    "loss": 0.428963840007782,
    "learning_rate": 0.00025158875009199293,
    "timestep_mean": 474.4375,
    "epsilon_norm": 8.040230751037598,
    "prediction_norm": 6.629022121429443,
    "timestamp": "2025-07-14T12:35:09.848892"
  },
  {
    "epoch": 2669,
    "batch_idx": 160,
    "loss": 0.33683639764785767,
    "learning_rate": 0.00025158875009199293,
    "timestep_mean": 427.09375,
    "epsilon_norm": 8.019394874572754,
    "prediction_norm": 5.901986122131348,
    "timestamp": "2025-07-14T12:35:09.853712"
  },
  {
    "epoch": 2669,
    "batch_idx": 192,
    "loss": 0.2293827384710312,
    "learning_rate": 0.00025158875009199293,
    "timestep_mean": 373.125,
    "epsilon_norm": 4.102066516876221,
    "prediction_norm": 3.9655609130859375,
    "timestamp": "2025-07-14T12:35:09.858617"
  },
  {
    "epoch": 2670,
    "batch_idx": 0,
    "loss": 0.39915919303894043,
    "learning_rate": 0.00025227066578379677,
    "timestep_mean": 515.28125,
    "epsilon_norm": 9.410689353942871,
    "prediction_norm": 7.6948981285095215,
    "timestamp": "2025-07-14T12:35:09.863453"
  },
  {
    "epoch": 2670,
    "batch_idx": 32,
    "loss": 0.2531525492668152,
    "learning_rate": 0.00025227066578379677,
    "timestep_mean": 416.59375,
    "epsilon_norm": 8.266892433166504,
    "prediction_norm": 7.158390045166016,
    "timestamp": "2025-07-14T12:35:09.868250"
  },
  {
    "epoch": 2670,
    "batch_idx": 64,
    "loss": 0.20753251016139984,
    "learning_rate": 0.00025227066578379677,
    "timestep_mean": 505.75,
    "epsilon_norm": 7.136760234832764,
    "prediction_norm": 6.019306182861328,
    "timestamp": "2025-07-14T12:35:09.873133"
  },
  {
    "epoch": 2670,
    "batch_idx": 96,
    "loss": 0.32739734649658203,
    "learning_rate": 0.00025227066578379677,
    "timestep_mean": 439.5625,
    "epsilon_norm": 8.902519226074219,
    "prediction_norm": 7.546806812286377,
    "timestamp": "2025-07-14T12:35:09.878053"
  },
  {
    "epoch": 2670,
    "batch_idx": 128,
    "loss": 0.4389987587928772,
    "learning_rate": 0.00025227066578379677,
    "timestep_mean": 454.96875,
    "epsilon_norm": 7.767008304595947,
    "prediction_norm": 5.655491352081299,
    "timestamp": "2025-07-14T12:35:09.882987"
  },
  {
    "epoch": 2670,
    "batch_idx": 160,
    "loss": 0.316834419965744,
    "learning_rate": 0.00025227066578379677,
    "timestep_mean": 515.28125,
    "epsilon_norm": 8.1095609664917,
    "prediction_norm": 6.830541133880615,
    "timestamp": "2025-07-14T12:35:09.887990"
  },
  {
    "epoch": 2670,
    "batch_idx": 192,
    "loss": 0.46677064895629883,
    "learning_rate": 0.00025227066578379677,
    "timestep_mean": 600.5,
    "epsilon_norm": 3.6501381397247314,
    "prediction_norm": 3.3452301025390625,
    "timestamp": "2025-07-14T12:35:09.893213"
  },
  {
    "epoch": 2671,
    "batch_idx": 0,
    "loss": 0.20585858821868896,
    "learning_rate": 0.00025295319272310634,
    "timestep_mean": 566.125,
    "epsilon_norm": 7.481196880340576,
    "prediction_norm": 6.651782989501953,
    "timestamp": "2025-07-14T12:35:09.898244"
  },
  {
    "epoch": 2671,
    "batch_idx": 32,
    "loss": 0.28333479166030884,
    "learning_rate": 0.00025295319272310634,
    "timestep_mean": 499.59375,
    "epsilon_norm": 7.441585063934326,
    "prediction_norm": 6.860870838165283,
    "timestamp": "2025-07-14T12:35:09.903362"
  },
  {
    "epoch": 2671,
    "batch_idx": 64,
    "loss": 0.29182934761047363,
    "learning_rate": 0.00025295319272310634,
    "timestep_mean": 510.03125,
    "epsilon_norm": 6.831064701080322,
    "prediction_norm": 5.875319004058838,
    "timestamp": "2025-07-14T12:35:09.908479"
  },
  {
    "epoch": 2671,
    "batch_idx": 96,
    "loss": 0.2825280725955963,
    "learning_rate": 0.00025295319272310634,
    "timestep_mean": 526.9375,
    "epsilon_norm": 8.776946067810059,
    "prediction_norm": 7.721164226531982,
    "timestamp": "2025-07-14T12:35:09.913697"
  },
  {
    "epoch": 2671,
    "batch_idx": 128,
    "loss": 0.43735364079475403,
    "learning_rate": 0.00025295319272310634,
    "timestep_mean": 424.8125,
    "epsilon_norm": 7.653533458709717,
    "prediction_norm": 6.097060680389404,
    "timestamp": "2025-07-14T12:35:09.918890"
  },
  {
    "epoch": 2671,
    "batch_idx": 160,
    "loss": 0.37599819898605347,
    "learning_rate": 0.00025295319272310634,
    "timestep_mean": 511.71875,
    "epsilon_norm": 7.806410789489746,
    "prediction_norm": 5.876260280609131,
    "timestamp": "2025-07-14T12:35:09.924114"
  },
  {
    "epoch": 2671,
    "batch_idx": 192,
    "loss": 0.24380545318126678,
    "learning_rate": 0.00025295319272310634,
    "timestep_mean": 484.0,
    "epsilon_norm": 4.770796775817871,
    "prediction_norm": 3.922529697418213,
    "timestamp": "2025-07-14T12:35:09.929251"
  },
  {
    "epoch": 2672,
    "batch_idx": 0,
    "loss": 0.2940244972705841,
    "learning_rate": 0.0002536363292258547,
    "timestep_mean": 505.5,
    "epsilon_norm": 8.746936798095703,
    "prediction_norm": 7.924770832061768,
    "timestamp": "2025-07-14T12:35:09.934540"
  },
  {
    "epoch": 2672,
    "batch_idx": 32,
    "loss": 0.18912380933761597,
    "learning_rate": 0.0002536363292258547,
    "timestep_mean": 595.9375,
    "epsilon_norm": 7.128696918487549,
    "prediction_norm": 6.098597526550293,
    "timestamp": "2025-07-14T12:35:09.939744"
  },
  {
    "epoch": 2672,
    "batch_idx": 64,
    "loss": 0.3965049386024475,
    "learning_rate": 0.0002536363292258547,
    "timestep_mean": 435.5625,
    "epsilon_norm": 8.218925476074219,
    "prediction_norm": 6.849862098693848,
    "timestamp": "2025-07-14T12:35:09.944966"
  },
  {
    "epoch": 2672,
    "batch_idx": 96,
    "loss": 0.24972917139530182,
    "learning_rate": 0.0002536363292258547,
    "timestep_mean": 604.59375,
    "epsilon_norm": 8.462715148925781,
    "prediction_norm": 7.363104343414307,
    "timestamp": "2025-07-14T12:35:09.950351"
  },
  {
    "epoch": 2672,
    "batch_idx": 128,
    "loss": 0.21433518826961517,
    "learning_rate": 0.0002536363292258547,
    "timestep_mean": 520.375,
    "epsilon_norm": 6.775066375732422,
    "prediction_norm": 5.242578983306885,
    "timestamp": "2025-07-14T12:35:09.955466"
  },
  {
    "epoch": 2672,
    "batch_idx": 160,
    "loss": 0.3453831076622009,
    "learning_rate": 0.0002536363292258547,
    "timestep_mean": 459.8125,
    "epsilon_norm": 7.385923862457275,
    "prediction_norm": 5.580408096313477,
    "timestamp": "2025-07-14T12:35:09.960704"
  },
  {
    "epoch": 2672,
    "batch_idx": 192,
    "loss": 0.46817126870155334,
    "learning_rate": 0.0002536363292258547,
    "timestep_mean": 422.125,
    "epsilon_norm": 4.4296956062316895,
    "prediction_norm": 3.4308955669403076,
    "timestamp": "2025-07-14T12:35:09.965868"
  },
  {
    "epoch": 2673,
    "batch_idx": 0,
    "loss": 0.18500538170337677,
    "learning_rate": 0.0002543200736064704,
    "timestep_mean": 451.65625,
    "epsilon_norm": 7.705126762390137,
    "prediction_norm": 6.690311431884766,
    "timestamp": "2025-07-14T12:35:09.971086"
  },
  {
    "epoch": 2673,
    "batch_idx": 32,
    "loss": 0.250273197889328,
    "learning_rate": 0.0002543200736064704,
    "timestep_mean": 466.25,
    "epsilon_norm": 7.199347496032715,
    "prediction_norm": 6.292940139770508,
    "timestamp": "2025-07-14T12:35:09.976287"
  },
  {
    "epoch": 2673,
    "batch_idx": 64,
    "loss": 0.23596526682376862,
    "learning_rate": 0.0002543200736064704,
    "timestep_mean": 463.3125,
    "epsilon_norm": 6.908791542053223,
    "prediction_norm": 6.068994045257568,
    "timestamp": "2025-07-14T12:35:09.981436"
  },
  {
    "epoch": 2673,
    "batch_idx": 96,
    "loss": 0.22078925371170044,
    "learning_rate": 0.0002543200736064704,
    "timestep_mean": 518.84375,
    "epsilon_norm": 7.16991662979126,
    "prediction_norm": 6.11989688873291,
    "timestamp": "2025-07-14T12:35:09.986619"
  },
  {
    "epoch": 2673,
    "batch_idx": 128,
    "loss": 0.2364966869354248,
    "learning_rate": 0.0002543200736064704,
    "timestep_mean": 525.84375,
    "epsilon_norm": 8.53833293914795,
    "prediction_norm": 7.5214128494262695,
    "timestamp": "2025-07-14T12:35:09.991830"
  },
  {
    "epoch": 2673,
    "batch_idx": 160,
    "loss": 0.2832109034061432,
    "learning_rate": 0.0002543200736064704,
    "timestep_mean": 514.84375,
    "epsilon_norm": 9.022844314575195,
    "prediction_norm": 7.596867561340332,
    "timestamp": "2025-07-14T12:35:09.997056"
  },
  {
    "epoch": 2673,
    "batch_idx": 192,
    "loss": 0.08327743411064148,
    "learning_rate": 0.0002543200736064704,
    "timestep_mean": 539.25,
    "epsilon_norm": 4.1654157638549805,
    "prediction_norm": 3.755969285964966,
    "timestamp": "2025-07-14T12:35:10.002158"
  },
  {
    "epoch": 2674,
    "batch_idx": 0,
    "loss": 0.3177752196788788,
    "learning_rate": 0.00025500442417788214,
    "timestep_mean": 523.6875,
    "epsilon_norm": 6.793910026550293,
    "prediction_norm": 5.625229835510254,
    "timestamp": "2025-07-14T12:35:10.007314"
  },
  {
    "epoch": 2674,
    "batch_idx": 32,
    "loss": 0.2955228090286255,
    "learning_rate": 0.00025500442417788214,
    "timestep_mean": 536.09375,
    "epsilon_norm": 7.9130120277404785,
    "prediction_norm": 6.232884407043457,
    "timestamp": "2025-07-14T12:35:10.012273"
  },
  {
    "epoch": 2674,
    "batch_idx": 64,
    "loss": 0.23877614736557007,
    "learning_rate": 0.00025500442417788214,
    "timestep_mean": 479.71875,
    "epsilon_norm": 7.353796482086182,
    "prediction_norm": 6.152708530426025,
    "timestamp": "2025-07-14T12:35:10.017429"
  },
  {
    "epoch": 2674,
    "batch_idx": 96,
    "loss": 0.2833807170391083,
    "learning_rate": 0.00025500442417788214,
    "timestep_mean": 501.1875,
    "epsilon_norm": 8.492512702941895,
    "prediction_norm": 7.272289752960205,
    "timestamp": "2025-07-14T12:35:10.022566"
  },
  {
    "epoch": 2674,
    "batch_idx": 128,
    "loss": 0.22924712300300598,
    "learning_rate": 0.00025500442417788214,
    "timestep_mean": 533.34375,
    "epsilon_norm": 8.08671760559082,
    "prediction_norm": 6.783555030822754,
    "timestamp": "2025-07-14T12:35:10.027661"
  },
  {
    "epoch": 2674,
    "batch_idx": 160,
    "loss": 0.32897573709487915,
    "learning_rate": 0.00025500442417788214,
    "timestep_mean": 448.90625,
    "epsilon_norm": 8.021642684936523,
    "prediction_norm": 6.501882076263428,
    "timestamp": "2025-07-14T12:35:10.032922"
  },
  {
    "epoch": 2674,
    "batch_idx": 192,
    "loss": 0.10745338350534439,
    "learning_rate": 0.00025500442417788214,
    "timestep_mean": 570.75,
    "epsilon_norm": 4.3293561935424805,
    "prediction_norm": 4.305980682373047,
    "timestamp": "2025-07-14T12:35:10.038051"
  },
  {
    "epoch": 2675,
    "batch_idx": 0,
    "loss": 0.3311481773853302,
    "learning_rate": 0.0002556893792515229,
    "timestep_mean": 454.25,
    "epsilon_norm": 9.73495101928711,
    "prediction_norm": 7.669857025146484,
    "timestamp": "2025-07-14T12:35:10.043252"
  },
  {
    "epoch": 2675,
    "batch_idx": 32,
    "loss": 0.32187584042549133,
    "learning_rate": 0.0002556893792515229,
    "timestep_mean": 532.09375,
    "epsilon_norm": 8.482285499572754,
    "prediction_norm": 7.384340763092041,
    "timestamp": "2025-07-14T12:35:10.048440"
  },
  {
    "epoch": 2675,
    "batch_idx": 64,
    "loss": 0.19733577966690063,
    "learning_rate": 0.0002556893792515229,
    "timestep_mean": 553.71875,
    "epsilon_norm": 8.166426658630371,
    "prediction_norm": 7.670736312866211,
    "timestamp": "2025-07-14T12:35:10.053675"
  },
  {
    "epoch": 2675,
    "batch_idx": 96,
    "loss": 0.36733996868133545,
    "learning_rate": 0.0002556893792515229,
    "timestep_mean": 517.28125,
    "epsilon_norm": 8.772420883178711,
    "prediction_norm": 7.031889915466309,
    "timestamp": "2025-07-14T12:35:10.058935"
  },
  {
    "epoch": 2675,
    "batch_idx": 128,
    "loss": 0.4205206632614136,
    "learning_rate": 0.0002556893792515229,
    "timestep_mean": 471.6875,
    "epsilon_norm": 8.196248054504395,
    "prediction_norm": 5.211371421813965,
    "timestamp": "2025-07-14T12:35:10.064133"
  },
  {
    "epoch": 2675,
    "batch_idx": 160,
    "loss": 0.22641795873641968,
    "learning_rate": 0.0002556893792515229,
    "timestep_mean": 472.90625,
    "epsilon_norm": 8.134354591369629,
    "prediction_norm": 6.983774662017822,
    "timestamp": "2025-07-14T12:35:10.069377"
  },
  {
    "epoch": 2675,
    "batch_idx": 192,
    "loss": 0.7558340430259705,
    "learning_rate": 0.0002556893792515229,
    "timestep_mean": 510.0,
    "epsilon_norm": 4.522344589233398,
    "prediction_norm": 2.9977638721466064,
    "timestamp": "2025-07-14T12:35:10.074571"
  },
  {
    "epoch": 2676,
    "batch_idx": 0,
    "loss": 0.13060489296913147,
    "learning_rate": 0.0002563749371373342,
    "timestep_mean": 511.6875,
    "epsilon_norm": 8.052138328552246,
    "prediction_norm": 7.046951770782471,
    "timestamp": "2025-07-14T12:35:10.079654"
  },
  {
    "epoch": 2676,
    "batch_idx": 32,
    "loss": 0.2542461156845093,
    "learning_rate": 0.0002563749371373342,
    "timestep_mean": 436.3125,
    "epsilon_norm": 7.244311332702637,
    "prediction_norm": 5.783865451812744,
    "timestamp": "2025-07-14T12:35:10.084756"
  },
  {
    "epoch": 2676,
    "batch_idx": 64,
    "loss": 0.21635659039020538,
    "learning_rate": 0.0002563749371373342,
    "timestep_mean": 537.15625,
    "epsilon_norm": 7.380337238311768,
    "prediction_norm": 6.268470764160156,
    "timestamp": "2025-07-14T12:35:10.089956"
  },
  {
    "epoch": 2676,
    "batch_idx": 96,
    "loss": 0.38086265325546265,
    "learning_rate": 0.0002563749371373342,
    "timestep_mean": 438.96875,
    "epsilon_norm": 8.788824081420898,
    "prediction_norm": 6.939931392669678,
    "timestamp": "2025-07-14T12:35:10.095047"
  },
  {
    "epoch": 2676,
    "batch_idx": 128,
    "loss": 0.2205798178911209,
    "learning_rate": 0.0002563749371373342,
    "timestep_mean": 459.8125,
    "epsilon_norm": 8.344263076782227,
    "prediction_norm": 7.317910671234131,
    "timestamp": "2025-07-14T12:35:10.100396"
  },
  {
    "epoch": 2676,
    "batch_idx": 160,
    "loss": 0.47766438126564026,
    "learning_rate": 0.0002563749371373342,
    "timestep_mean": 464.53125,
    "epsilon_norm": 9.286141395568848,
    "prediction_norm": 6.303641319274902,
    "timestamp": "2025-07-14T12:35:10.105668"
  },
  {
    "epoch": 2676,
    "batch_idx": 192,
    "loss": 0.21157702803611755,
    "learning_rate": 0.0002563749371373342,
    "timestep_mean": 491.625,
    "epsilon_norm": 3.389653444290161,
    "prediction_norm": 2.516855239868164,
    "timestamp": "2025-07-14T12:35:10.110892"
  },
  {
    "epoch": 2677,
    "batch_idx": 0,
    "loss": 0.2835780382156372,
    "learning_rate": 0.00025706109614377035,
    "timestep_mean": 546.53125,
    "epsilon_norm": 7.368359565734863,
    "prediction_norm": 6.14233922958374,
    "timestamp": "2025-07-14T12:35:10.116094"
  },
  {
    "epoch": 2677,
    "batch_idx": 32,
    "loss": 0.2816609740257263,
    "learning_rate": 0.00025706109614377035,
    "timestep_mean": 483.125,
    "epsilon_norm": 8.174860954284668,
    "prediction_norm": 7.125665664672852,
    "timestamp": "2025-07-14T12:35:10.121112"
  },
  {
    "epoch": 2677,
    "batch_idx": 64,
    "loss": 0.3391128182411194,
    "learning_rate": 0.00025706109614377035,
    "timestep_mean": 438.90625,
    "epsilon_norm": 6.974105358123779,
    "prediction_norm": 5.801359176635742,
    "timestamp": "2025-07-14T12:35:10.126041"
  },
  {
    "epoch": 2677,
    "batch_idx": 96,
    "loss": 0.3693990707397461,
    "learning_rate": 0.00025706109614377035,
    "timestep_mean": 468.84375,
    "epsilon_norm": 9.190423965454102,
    "prediction_norm": 7.9555792808532715,
    "timestamp": "2025-07-14T12:35:10.130885"
  },
  {
    "epoch": 2677,
    "batch_idx": 128,
    "loss": 0.3095867931842804,
    "learning_rate": 0.00025706109614377035,
    "timestep_mean": 538.625,
    "epsilon_norm": 8.329357147216797,
    "prediction_norm": 7.667918682098389,
    "timestamp": "2025-07-14T12:35:10.135724"
  },
  {
    "epoch": 2677,
    "batch_idx": 160,
    "loss": 0.16091689467430115,
    "learning_rate": 0.00025706109614377035,
    "timestep_mean": 504.1875,
    "epsilon_norm": 7.116674900054932,
    "prediction_norm": 6.506007194519043,
    "timestamp": "2025-07-14T12:35:10.140556"
  },
  {
    "epoch": 2677,
    "batch_idx": 192,
    "loss": 0.497000515460968,
    "learning_rate": 0.00025706109614377035,
    "timestep_mean": 509.625,
    "epsilon_norm": 3.20127534866333,
    "prediction_norm": 3.541069984436035,
    "timestamp": "2025-07-14T12:35:10.145506"
  },
  {
    "epoch": 2678,
    "batch_idx": 0,
    "loss": 0.2507469356060028,
    "learning_rate": 0.00025774785457780165,
    "timestep_mean": 473.5,
    "epsilon_norm": 8.172463417053223,
    "prediction_norm": 7.202625274658203,
    "timestamp": "2025-07-14T12:35:10.150755"
  },
  {
    "epoch": 2678,
    "batch_idx": 32,
    "loss": 0.2726508677005768,
    "learning_rate": 0.00025774785457780165,
    "timestep_mean": 511.9375,
    "epsilon_norm": 8.589405059814453,
    "prediction_norm": 7.095109939575195,
    "timestamp": "2025-07-14T12:35:10.155846"
  },
  {
    "epoch": 2678,
    "batch_idx": 64,
    "loss": 0.20097842812538147,
    "learning_rate": 0.00025774785457780165,
    "timestep_mean": 576.3125,
    "epsilon_norm": 7.42119836807251,
    "prediction_norm": 6.693098545074463,
    "timestamp": "2025-07-14T12:35:10.160905"
  },
  {
    "epoch": 2678,
    "batch_idx": 96,
    "loss": 0.290752649307251,
    "learning_rate": 0.00025774785457780165,
    "timestep_mean": 485.28125,
    "epsilon_norm": 7.043761253356934,
    "prediction_norm": 5.905899524688721,
    "timestamp": "2025-07-14T12:35:10.166027"
  },
  {
    "epoch": 2678,
    "batch_idx": 128,
    "loss": 0.2767990827560425,
    "learning_rate": 0.00025774785457780165,
    "timestep_mean": 481.84375,
    "epsilon_norm": 7.018167972564697,
    "prediction_norm": 5.509868621826172,
    "timestamp": "2025-07-14T12:35:10.171165"
  },
  {
    "epoch": 2678,
    "batch_idx": 160,
    "loss": 0.17828746140003204,
    "learning_rate": 0.00025774785457780165,
    "timestep_mean": 499.34375,
    "epsilon_norm": 7.717826843261719,
    "prediction_norm": 7.100047588348389,
    "timestamp": "2025-07-14T12:35:10.176252"
  },
  {
    "epoch": 2678,
    "batch_idx": 192,
    "loss": 0.17630109190940857,
    "learning_rate": 0.00025774785457780165,
    "timestep_mean": 606.625,
    "epsilon_norm": 4.341456413269043,
    "prediction_norm": 4.237065315246582,
    "timestamp": "2025-07-14T12:35:10.181276"
  },
  {
    "epoch": 2679,
    "batch_idx": 0,
    "loss": 0.25820833444595337,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 445.3125,
    "epsilon_norm": 7.327202320098877,
    "prediction_norm": 6.270656585693359,
    "timestamp": "2025-07-14T12:35:10.186400"
  },
  {
    "epoch": 2679,
    "batch_idx": 32,
    "loss": 0.1414395272731781,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 619.90625,
    "epsilon_norm": 8.571027755737305,
    "prediction_norm": 7.377089500427246,
    "timestamp": "2025-07-14T12:35:10.191588"
  },
  {
    "epoch": 2679,
    "batch_idx": 64,
    "loss": 0.5140541791915894,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 435.46875,
    "epsilon_norm": 8.387203216552734,
    "prediction_norm": 6.012574195861816,
    "timestamp": "2025-07-14T12:35:10.196872"
  },
  {
    "epoch": 2679,
    "batch_idx": 96,
    "loss": 0.23549865186214447,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 510.03125,
    "epsilon_norm": 8.758916854858398,
    "prediction_norm": 7.489012718200684,
    "timestamp": "2025-07-14T12:35:10.202091"
  },
  {
    "epoch": 2679,
    "batch_idx": 128,
    "loss": 0.14191114902496338,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 561.5,
    "epsilon_norm": 8.198216438293457,
    "prediction_norm": 7.700588226318359,
    "timestamp": "2025-07-14T12:35:10.207278"
  },
  {
    "epoch": 2679,
    "batch_idx": 160,
    "loss": 0.2653878927230835,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 518.96875,
    "epsilon_norm": 8.66438102722168,
    "prediction_norm": 7.452767372131348,
    "timestamp": "2025-07-14T12:35:10.212487"
  },
  {
    "epoch": 2679,
    "batch_idx": 192,
    "loss": 0.39947235584259033,
    "learning_rate": 0.0002584352107449198,
    "timestep_mean": 548.5,
    "epsilon_norm": 3.8861610889434814,
    "prediction_norm": 2.900958776473999,
    "timestamp": "2025-07-14T12:35:10.217704"
  },
  {
    "epoch": 2680,
    "batch_idx": 0,
    "loss": 0.33753812313079834,
    "learning_rate": 0.00025912316294914294,
    "timestep_mean": 458.40625,
    "epsilon_norm": 9.664294242858887,
    "prediction_norm": 7.300981044769287,
    "timestamp": "2025-07-14T12:35:10.222900"
  },
  {
    "epoch": 2680,
    "batch_idx": 32,
    "loss": 0.26798680424690247,
    "learning_rate": 0.00025912316294914294,
    "timestep_mean": 542.03125,
    "epsilon_norm": 8.091692924499512,
    "prediction_norm": 7.287613868713379,
    "timestamp": "2025-07-14T12:35:10.228053"
  },
  {
    "epoch": 2680,
    "batch_idx": 64,
    "loss": 0.381750226020813,
    "learning_rate": 0.00025912316294914294,
    "timestep_mean": 464.875,
    "epsilon_norm": 9.190140724182129,
    "prediction_norm": 7.3698859214782715,
    "timestamp": "2025-07-14T12:35:10.233232"
  },
  {
    "epoch": 2680,
    "batch_idx": 96,
    "loss": 0.260128378868103,
    "learning_rate": 0.00025912316294914294,
    "timestep_mean": 545.09375,
    "epsilon_norm": 6.6712751388549805,
    "prediction_norm": 5.2417073249816895,
    "timestamp": "2025-07-14T12:35:10.238567"
  },
  {
    "epoch": 2680,
    "batch_idx": 128,
    "loss": 0.35339534282684326,
    "learning_rate": 0.00025912316294914294,
    "timestep_mean": 454.21875,
    "epsilon_norm": 8.34238338470459,
    "prediction_norm": 6.91002893447876,
    "timestamp": "2025-07-14T12:35:10.243770"
  },
  {
    "epoch": 2680,
    "batch_idx": 160,
    "loss": 0.3265138864517212,
    "learning_rate": 0.00025912316294914294,
    "timestep_mean": 518.0625,
    "epsilon_norm": 8.56612491607666,
    "prediction_norm": 6.924702167510986,
    "timestamp": "2025-07-14T12:35:10.249000"
  },
  {
    "epoch": 2680,
    "batch_idx": 192,
    "loss": 0.3998866081237793,
    "learning_rate": 0.00025912316294914294,
    "timestep_mean": 511.5,
    "epsilon_norm": 3.962578535079956,
    "prediction_norm": 3.390859842300415,
    "timestamp": "2025-07-14T12:35:10.254167"
  },
  {
    "epoch": 2681,
    "batch_idx": 0,
    "loss": 0.25167304277420044,
    "learning_rate": 0.00025981170949301624,
    "timestep_mean": 491.875,
    "epsilon_norm": 7.26314640045166,
    "prediction_norm": 6.020586967468262,
    "timestamp": "2025-07-14T12:35:10.259353"
  },
  {
    "epoch": 2681,
    "batch_idx": 32,
    "loss": 0.21390576660633087,
    "learning_rate": 0.00025981170949301624,
    "timestep_mean": 461.03125,
    "epsilon_norm": 7.015220642089844,
    "prediction_norm": 5.815682411193848,
    "timestamp": "2025-07-14T12:35:10.264410"
  },
  {
    "epoch": 2681,
    "batch_idx": 64,
    "loss": 0.11764788627624512,
    "learning_rate": 0.00025981170949301624,
    "timestep_mean": 558.71875,
    "epsilon_norm": 7.314868927001953,
    "prediction_norm": 6.803377151489258,
    "timestamp": "2025-07-14T12:35:10.269604"
  },
  {
    "epoch": 2681,
    "batch_idx": 96,
    "loss": 0.3122274577617645,
    "learning_rate": 0.00025981170949301624,
    "timestep_mean": 490.3125,
    "epsilon_norm": 7.3995680809021,
    "prediction_norm": 6.415419578552246,
    "timestamp": "2025-07-14T12:35:10.274764"
  },
  {
    "epoch": 2681,
    "batch_idx": 128,
    "loss": 0.25867384672164917,
    "learning_rate": 0.00025981170949301624,
    "timestep_mean": 458.0625,
    "epsilon_norm": 6.973231792449951,
    "prediction_norm": 5.905004501342773,
    "timestamp": "2025-07-14T12:35:10.279958"
  },
  {
    "epoch": 2681,
    "batch_idx": 160,
    "loss": 0.7120777368545532,
    "learning_rate": 0.00025981170949301624,
    "timestep_mean": 429.625,
    "epsilon_norm": 9.115654945373535,
    "prediction_norm": 6.555093765258789,
    "timestamp": "2025-07-14T12:35:10.285143"
  },
  {
    "epoch": 2681,
    "batch_idx": 192,
    "loss": 0.3482643663883209,
    "learning_rate": 0.00025981170949301624,
    "timestep_mean": 426.25,
    "epsilon_norm": 3.3015525341033936,
    "prediction_norm": 2.7581028938293457,
    "timestamp": "2025-07-14T12:35:10.290289"
  },
  {
    "epoch": 2682,
    "batch_idx": 0,
    "loss": 0.21209079027175903,
    "learning_rate": 0.00026050084867762,
    "timestep_mean": 492.5,
    "epsilon_norm": 8.283747673034668,
    "prediction_norm": 7.323930740356445,
    "timestamp": "2025-07-14T12:35:10.295534"
  },
  {
    "epoch": 2682,
    "batch_idx": 32,
    "loss": 0.4413093328475952,
    "learning_rate": 0.00026050084867762,
    "timestep_mean": 452.4375,
    "epsilon_norm": 9.020130157470703,
    "prediction_norm": 7.000239849090576,
    "timestamp": "2025-07-14T12:35:10.300871"
  },
  {
    "epoch": 2682,
    "batch_idx": 64,
    "loss": 0.2900657057762146,
    "learning_rate": 0.00026050084867762,
    "timestep_mean": 486.5625,
    "epsilon_norm": 8.680880546569824,
    "prediction_norm": 7.255784511566162,
    "timestamp": "2025-07-14T12:35:10.306026"
  },
  {
    "epoch": 2682,
    "batch_idx": 96,
    "loss": 0.48681536316871643,
    "learning_rate": 0.00026050084867762,
    "timestep_mean": 460.375,
    "epsilon_norm": 9.464667320251465,
    "prediction_norm": 7.721619129180908,
    "timestamp": "2025-07-14T12:35:10.311245"
  },
  {
    "epoch": 2682,
    "batch_idx": 128,
    "loss": 0.1984786093235016,
    "learning_rate": 0.00026050084867762,
    "timestep_mean": 513.53125,
    "epsilon_norm": 8.307849884033203,
    "prediction_norm": 6.8512959480285645,
    "timestamp": "2025-07-14T12:35:10.316416"
  },
  {
    "epoch": 2682,
    "batch_idx": 160,
    "loss": 0.1021428033709526,
    "learning_rate": 0.00026050084867762,
    "timestep_mean": 599.28125,
    "epsilon_norm": 7.17039680480957,
    "prediction_norm": 6.710853576660156,
    "timestamp": "2025-07-14T12:35:10.321582"
  },
  {
    "epoch": 2682,
    "batch_idx": 192,
    "loss": 0.36029019951820374,
    "learning_rate": 0.00026050084867762,
    "timestep_mean": 591.25,
    "epsilon_norm": 4.813546180725098,
    "prediction_norm": 4.4114670753479,
    "timestamp": "2025-07-14T12:35:10.326607"
  },
  {
    "epoch": 2683,
    "batch_idx": 0,
    "loss": 0.26471707224845886,
    "learning_rate": 0.0002611905788025717,
    "timestep_mean": 494.03125,
    "epsilon_norm": 7.50308084487915,
    "prediction_norm": 6.286699295043945,
    "timestamp": "2025-07-14T12:35:10.331808"
  },
  {
    "epoch": 2683,
    "batch_idx": 32,
    "loss": 0.365159809589386,
    "learning_rate": 0.0002611905788025717,
    "timestep_mean": 376.71875,
    "epsilon_norm": 7.759324073791504,
    "prediction_norm": 5.882182598114014,
    "timestamp": "2025-07-14T12:35:10.336944"
  },
  {
    "epoch": 2683,
    "batch_idx": 64,
    "loss": 0.5437343716621399,
    "learning_rate": 0.0002611905788025717,
    "timestep_mean": 398.4375,
    "epsilon_norm": 7.961664199829102,
    "prediction_norm": 5.228789806365967,
    "timestamp": "2025-07-14T12:35:10.342290"
  },
  {
    "epoch": 2683,
    "batch_idx": 96,
    "loss": 0.3398423194885254,
    "learning_rate": 0.0002611905788025717,
    "timestep_mean": 442.34375,
    "epsilon_norm": 7.396749496459961,
    "prediction_norm": 5.9737982749938965,
    "timestamp": "2025-07-14T12:35:10.347166"
  },
  {
    "epoch": 2683,
    "batch_idx": 128,
    "loss": 0.18109726905822754,
    "learning_rate": 0.0002611905788025717,
    "timestep_mean": 527.28125,
    "epsilon_norm": 7.320528984069824,
    "prediction_norm": 6.871097087860107,
    "timestamp": "2025-07-14T12:35:10.352281"
  },
  {
    "epoch": 2683,
    "batch_idx": 160,
    "loss": 0.26600104570388794,
    "learning_rate": 0.0002611905788025717,
    "timestep_mean": 569.5625,
    "epsilon_norm": 7.163193225860596,
    "prediction_norm": 6.460860729217529,
    "timestamp": "2025-07-14T12:35:10.357514"
  },
  {
    "epoch": 2683,
    "batch_idx": 192,
    "loss": 0.1372154951095581,
    "learning_rate": 0.0002611905788025717,
    "timestep_mean": 532.0,
    "epsilon_norm": 2.9391369819641113,
    "prediction_norm": 2.6951591968536377,
    "timestamp": "2025-07-14T12:35:10.362737"
  },
  {
    "epoch": 2684,
    "batch_idx": 0,
    "loss": 0.20388157665729523,
    "learning_rate": 0.00026188089816603084,
    "timestep_mean": 534.65625,
    "epsilon_norm": 7.235410690307617,
    "prediction_norm": 6.266132831573486,
    "timestamp": "2025-07-14T12:35:10.367952"
  },
  {
    "epoch": 2684,
    "batch_idx": 32,
    "loss": 0.20447379350662231,
    "learning_rate": 0.00026188089816603084,
    "timestep_mean": 525.875,
    "epsilon_norm": 7.570921421051025,
    "prediction_norm": 6.819692611694336,
    "timestamp": "2025-07-14T12:35:10.373016"
  },
  {
    "epoch": 2684,
    "batch_idx": 64,
    "loss": 0.4046342968940735,
    "learning_rate": 0.00026188089816603084,
    "timestep_mean": 459.1875,
    "epsilon_norm": 7.954190254211426,
    "prediction_norm": 6.169151782989502,
    "timestamp": "2025-07-14T12:35:10.377852"
  },
  {
    "epoch": 2684,
    "batch_idx": 96,
    "loss": 0.1906495988368988,
    "learning_rate": 0.00026188089816603084,
    "timestep_mean": 565.125,
    "epsilon_norm": 7.7920823097229,
    "prediction_norm": 6.962466716766357,
    "timestamp": "2025-07-14T12:35:10.382806"
  },
  {
    "epoch": 2684,
    "batch_idx": 128,
    "loss": 0.1979387402534485,
    "learning_rate": 0.00026188089816603084,
    "timestep_mean": 527.6875,
    "epsilon_norm": 7.8398661613464355,
    "prediction_norm": 6.8871541023254395,
    "timestamp": "2025-07-14T12:35:10.387831"
  },
  {
    "epoch": 2684,
    "batch_idx": 160,
    "loss": 0.29061633348464966,
    "learning_rate": 0.00026188089816603084,
    "timestep_mean": 379.90625,
    "epsilon_norm": 7.7339043617248535,
    "prediction_norm": 5.964300632476807,
    "timestamp": "2025-07-14T12:35:10.392713"
  },
  {
    "epoch": 2684,
    "batch_idx": 192,
    "loss": 0.9375468492507935,
    "learning_rate": 0.00026188089816603084,
    "timestep_mean": 293.875,
    "epsilon_norm": 5.328389644622803,
    "prediction_norm": 3.7095766067504883,
    "timestamp": "2025-07-14T12:35:10.397547"
  },
  {
    "epoch": 2685,
    "batch_idx": 0,
    "loss": 0.30944275856018066,
    "learning_rate": 0.000262571805064703,
    "timestep_mean": 532.53125,
    "epsilon_norm": 8.47894287109375,
    "prediction_norm": 7.310800552368164,
    "timestamp": "2025-07-14T12:35:10.402428"
  },
  {
    "epoch": 2685,
    "batch_idx": 32,
    "loss": 0.3726845383644104,
    "learning_rate": 0.000262571805064703,
    "timestep_mean": 484.75,
    "epsilon_norm": 8.457212448120117,
    "prediction_norm": 6.279011249542236,
    "timestamp": "2025-07-14T12:35:10.407364"
  },
  {
    "epoch": 2685,
    "batch_idx": 64,
    "loss": 0.3828754723072052,
    "learning_rate": 0.000262571805064703,
    "timestep_mean": 443.59375,
    "epsilon_norm": 8.908961296081543,
    "prediction_norm": 6.761874675750732,
    "timestamp": "2025-07-14T12:35:10.412264"
  },
  {
    "epoch": 2685,
    "batch_idx": 96,
    "loss": 0.2225559949874878,
    "learning_rate": 0.000262571805064703,
    "timestep_mean": 441.65625,
    "epsilon_norm": 7.274007797241211,
    "prediction_norm": 5.873460292816162,
    "timestamp": "2025-07-14T12:35:10.417402"
  },
  {
    "epoch": 2685,
    "batch_idx": 128,
    "loss": 0.42298489809036255,
    "learning_rate": 0.000262571805064703,
    "timestep_mean": 526.125,
    "epsilon_norm": 7.780068874359131,
    "prediction_norm": 5.4987921714782715,
    "timestamp": "2025-07-14T12:35:10.422545"
  },
  {
    "epoch": 2685,
    "batch_idx": 160,
    "loss": 0.1459818035364151,
    "learning_rate": 0.000262571805064703,
    "timestep_mean": 466.46875,
    "epsilon_norm": 7.1524338722229,
    "prediction_norm": 6.570042133331299,
    "timestamp": "2025-07-14T12:35:10.427462"
  },
  {
    "epoch": 2685,
    "batch_idx": 192,
    "loss": 0.2601349949836731,
    "learning_rate": 0.000262571805064703,
    "timestep_mean": 407.625,
    "epsilon_norm": 4.695985794067383,
    "prediction_norm": 4.0168585777282715,
    "timestamp": "2025-07-14T12:35:10.432365"
  },
  {
    "epoch": 2686,
    "batch_idx": 0,
    "loss": 0.21934130787849426,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 520.59375,
    "epsilon_norm": 9.014822959899902,
    "prediction_norm": 7.806188106536865,
    "timestamp": "2025-07-14T12:35:10.437302"
  },
  {
    "epoch": 2686,
    "batch_idx": 32,
    "loss": 0.16947954893112183,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 551.09375,
    "epsilon_norm": 8.446554183959961,
    "prediction_norm": 7.774326324462891,
    "timestamp": "2025-07-14T12:35:10.442208"
  },
  {
    "epoch": 2686,
    "batch_idx": 64,
    "loss": 0.19708094000816345,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 468.8125,
    "epsilon_norm": 6.8665571212768555,
    "prediction_norm": 6.097081184387207,
    "timestamp": "2025-07-14T12:35:10.447193"
  },
  {
    "epoch": 2686,
    "batch_idx": 96,
    "loss": 0.21842297911643982,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 578.375,
    "epsilon_norm": 6.784533977508545,
    "prediction_norm": 5.55800199508667,
    "timestamp": "2025-07-14T12:35:10.452273"
  },
  {
    "epoch": 2686,
    "batch_idx": 128,
    "loss": 0.24528372287750244,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 538.6875,
    "epsilon_norm": 8.359883308410645,
    "prediction_norm": 7.138495445251465,
    "timestamp": "2025-07-14T12:35:10.457214"
  },
  {
    "epoch": 2686,
    "batch_idx": 160,
    "loss": 0.3458932638168335,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 458.84375,
    "epsilon_norm": 7.808132648468018,
    "prediction_norm": 6.792186260223389,
    "timestamp": "2025-07-14T12:35:10.462354"
  },
  {
    "epoch": 2686,
    "batch_idx": 192,
    "loss": 0.04266420751810074,
    "learning_rate": 0.00026326329779384423,
    "timestep_mean": 620.625,
    "epsilon_norm": 4.443561553955078,
    "prediction_norm": 4.506227016448975,
    "timestamp": "2025-07-14T12:35:10.467476"
  },
  {
    "epoch": 2687,
    "batch_idx": 0,
    "loss": 0.33018064498901367,
    "learning_rate": 0.0002639553746472652,
    "timestep_mean": 408.28125,
    "epsilon_norm": 6.960780143737793,
    "prediction_norm": 4.972797870635986,
    "timestamp": "2025-07-14T12:35:10.472610"
  },
  {
    "epoch": 2687,
    "batch_idx": 32,
    "loss": 0.2721458971500397,
    "learning_rate": 0.0002639553746472652,
    "timestep_mean": 520.4375,
    "epsilon_norm": 8.419103622436523,
    "prediction_norm": 7.202348709106445,
    "timestamp": "2025-07-14T12:35:10.477675"
  },
  {
    "epoch": 2687,
    "batch_idx": 64,
    "loss": 0.14340238273143768,
    "learning_rate": 0.0002639553746472652,
    "timestep_mean": 470.46875,
    "epsilon_norm": 8.111739158630371,
    "prediction_norm": 7.048946857452393,
    "timestamp": "2025-07-14T12:35:10.482809"
  },
  {
    "epoch": 2687,
    "batch_idx": 96,
    "loss": 0.22110293805599213,
    "learning_rate": 0.0002639553746472652,
    "timestep_mean": 569.625,
    "epsilon_norm": 8.437490463256836,
    "prediction_norm": 7.2860589027404785,
    "timestamp": "2025-07-14T12:35:10.487970"
  },
  {
    "epoch": 2687,
    "batch_idx": 128,
    "loss": 0.4353514313697815,
    "learning_rate": 0.0002639553746472652,
    "timestep_mean": 460.5,
    "epsilon_norm": 9.472886085510254,
    "prediction_norm": 7.500611782073975,
    "timestamp": "2025-07-14T12:35:10.493091"
  },
  {
    "epoch": 2687,
    "batch_idx": 160,
    "loss": 0.36137545108795166,
    "learning_rate": 0.0002639553746472652,
    "timestep_mean": 561.9375,
    "epsilon_norm": 7.197239875793457,
    "prediction_norm": 5.406813621520996,
    "timestamp": "2025-07-14T12:35:10.498319"
  },
  {
    "epoch": 2687,
    "batch_idx": 192,
    "loss": 0.1365961879491806,
    "learning_rate": 0.0002639553746472652,
    "timestep_mean": 493.375,
    "epsilon_norm": 4.186094760894775,
    "prediction_norm": 4.196533679962158,
    "timestamp": "2025-07-14T12:35:10.503473"
  },
  {
    "epoch": 2688,
    "batch_idx": 0,
    "loss": 0.19461695849895477,
    "learning_rate": 0.00026464803391733436,
    "timestep_mean": 568.375,
    "epsilon_norm": 7.205427169799805,
    "prediction_norm": 5.561349391937256,
    "timestamp": "2025-07-14T12:35:10.508698"
  },
  {
    "epoch": 2688,
    "batch_idx": 32,
    "loss": 0.4811716079711914,
    "learning_rate": 0.00026464803391733436,
    "timestep_mean": 427.09375,
    "epsilon_norm": 9.704160690307617,
    "prediction_norm": 7.95940637588501,
    "timestamp": "2025-07-14T12:35:10.513845"
  },
  {
    "epoch": 2688,
    "batch_idx": 64,
    "loss": 0.21025395393371582,
    "learning_rate": 0.00026464803391733436,
    "timestep_mean": 431.6875,
    "epsilon_norm": 8.396324157714844,
    "prediction_norm": 7.216775417327881,
    "timestamp": "2025-07-14T12:35:10.519009"
  },
  {
    "epoch": 2688,
    "batch_idx": 96,
    "loss": 0.23911795020103455,
    "learning_rate": 0.00026464803391733436,
    "timestep_mean": 536.78125,
    "epsilon_norm": 7.070917129516602,
    "prediction_norm": 6.31442403793335,
    "timestamp": "2025-07-14T12:35:10.524215"
  },
  {
    "epoch": 2688,
    "batch_idx": 128,
    "loss": 0.3143073320388794,
    "learning_rate": 0.00026464803391733436,
    "timestep_mean": 539.90625,
    "epsilon_norm": 7.718664646148682,
    "prediction_norm": 6.26738166809082,
    "timestamp": "2025-07-14T12:35:10.529365"
  },
  {
    "epoch": 2688,
    "batch_idx": 160,
    "loss": 0.32032620906829834,
    "learning_rate": 0.00026464803391733436,
    "timestep_mean": 424.40625,
    "epsilon_norm": 7.6734418869018555,
    "prediction_norm": 6.857409477233887,
    "timestamp": "2025-07-14T12:35:10.534569"
  },
  {
    "epoch": 2688,
    "batch_idx": 192,
    "loss": 0.32867157459259033,
    "learning_rate": 0.00026464803391733436,
    "timestep_mean": 598.5,
    "epsilon_norm": 3.4910764694213867,
    "prediction_norm": 2.7828962802886963,
    "timestamp": "2025-07-14T12:35:10.539729"
  },
  {
    "epoch": 2689,
    "batch_idx": 0,
    "loss": 0.3148665130138397,
    "learning_rate": 0.0002653412738949841,
    "timestep_mean": 491.65625,
    "epsilon_norm": 7.493431091308594,
    "prediction_norm": 6.449307918548584,
    "timestamp": "2025-07-14T12:35:10.544958"
  },
  {
    "epoch": 2689,
    "batch_idx": 32,
    "loss": 0.2597542405128479,
    "learning_rate": 0.0002653412738949841,
    "timestep_mean": 563.5,
    "epsilon_norm": 8.671348571777344,
    "prediction_norm": 7.678160190582275,
    "timestamp": "2025-07-14T12:35:10.550096"
  },
  {
    "epoch": 2689,
    "batch_idx": 64,
    "loss": 0.36878836154937744,
    "learning_rate": 0.0002653412738949841,
    "timestep_mean": 470.65625,
    "epsilon_norm": 7.906249046325684,
    "prediction_norm": 5.72825813293457,
    "timestamp": "2025-07-14T12:35:10.555363"
  },
  {
    "epoch": 2689,
    "batch_idx": 96,
    "loss": 0.4418380856513977,
    "learning_rate": 0.0002653412738949841,
    "timestep_mean": 489.34375,
    "epsilon_norm": 7.762930870056152,
    "prediction_norm": 6.001647472381592,
    "timestamp": "2025-07-14T12:35:10.560575"
  },
  {
    "epoch": 2689,
    "batch_idx": 128,
    "loss": 0.39842158555984497,
    "learning_rate": 0.0002653412738949841,
    "timestep_mean": 416.5625,
    "epsilon_norm": 8.364191055297852,
    "prediction_norm": 6.574035167694092,
    "timestamp": "2025-07-14T12:35:10.565763"
  },
  {
    "epoch": 2689,
    "batch_idx": 160,
    "loss": 0.33049073815345764,
    "learning_rate": 0.0002653412738949841,
    "timestep_mean": 460.5625,
    "epsilon_norm": 7.427796363830566,
    "prediction_norm": 5.820879936218262,
    "timestamp": "2025-07-14T12:35:10.570912"
  },
  {
    "epoch": 2689,
    "batch_idx": 192,
    "loss": 0.3340640366077423,
    "learning_rate": 0.0002653412738949841,
    "timestep_mean": 369.875,
    "epsilon_norm": 4.159857273101807,
    "prediction_norm": 3.3418962955474854,
    "timestamp": "2025-07-14T12:35:10.576282"
  },
  {
    "epoch": 2690,
    "batch_idx": 0,
    "loss": 0.21152400970458984,
    "learning_rate": 0.00026603509286971384,
    "timestep_mean": 506.125,
    "epsilon_norm": 8.427199363708496,
    "prediction_norm": 7.384430885314941,
    "timestamp": "2025-07-14T12:35:10.581300"
  },
  {
    "epoch": 2690,
    "batch_idx": 32,
    "loss": 0.3538826107978821,
    "learning_rate": 0.00026603509286971384,
    "timestep_mean": 495.4375,
    "epsilon_norm": 7.861941814422607,
    "prediction_norm": 6.8766279220581055,
    "timestamp": "2025-07-14T12:35:10.586409"
  },
  {
    "epoch": 2690,
    "batch_idx": 64,
    "loss": 0.2512047290802002,
    "learning_rate": 0.00026603509286971384,
    "timestep_mean": 567.40625,
    "epsilon_norm": 7.045284271240234,
    "prediction_norm": 6.11135196685791,
    "timestamp": "2025-07-14T12:35:10.591641"
  },
  {
    "epoch": 2690,
    "batch_idx": 96,
    "loss": 0.32196712493896484,
    "learning_rate": 0.00026603509286971384,
    "timestep_mean": 435.96875,
    "epsilon_norm": 8.093518257141113,
    "prediction_norm": 6.18445348739624,
    "timestamp": "2025-07-14T12:35:10.596888"
  },
  {
    "epoch": 2690,
    "batch_idx": 128,
    "loss": 0.20280882716178894,
    "learning_rate": 0.00026603509286971384,
    "timestep_mean": 560.59375,
    "epsilon_norm": 8.171993255615234,
    "prediction_norm": 7.681651592254639,
    "timestamp": "2025-07-14T12:35:10.602147"
  },
  {
    "epoch": 2690,
    "batch_idx": 160,
    "loss": 0.34204575419425964,
    "learning_rate": 0.00026603509286971384,
    "timestep_mean": 524.875,
    "epsilon_norm": 9.087423324584961,
    "prediction_norm": 7.945301055908203,
    "timestamp": "2025-07-14T12:35:10.607344"
  },
  {
    "epoch": 2690,
    "batch_idx": 192,
    "loss": 0.2539633810520172,
    "learning_rate": 0.00026603509286971384,
    "timestep_mean": 443.75,
    "epsilon_norm": 4.661159038543701,
    "prediction_norm": 3.854123115539551,
    "timestamp": "2025-07-14T12:35:10.612546"
  },
  {
    "epoch": 2691,
    "batch_idx": 0,
    "loss": 0.4291539192199707,
    "learning_rate": 0.00026672948912959414,
    "timestep_mean": 410.5625,
    "epsilon_norm": 8.498247146606445,
    "prediction_norm": 6.529942512512207,
    "timestamp": "2025-07-14T12:35:10.617981"
  },
  {
    "epoch": 2691,
    "batch_idx": 32,
    "loss": 0.15785762667655945,
    "learning_rate": 0.00026672948912959414,
    "timestep_mean": 514.0625,
    "epsilon_norm": 7.66496467590332,
    "prediction_norm": 6.6147260665893555,
    "timestamp": "2025-07-14T12:35:10.623134"
  },
  {
    "epoch": 2691,
    "batch_idx": 64,
    "loss": 0.19448992609977722,
    "learning_rate": 0.00026672948912959414,
    "timestep_mean": 534.15625,
    "epsilon_norm": 7.4679484367370605,
    "prediction_norm": 6.6001996994018555,
    "timestamp": "2025-07-14T12:35:10.628232"
  },
  {
    "epoch": 2691,
    "batch_idx": 96,
    "loss": 0.17191003262996674,
    "learning_rate": 0.00026672948912959414,
    "timestep_mean": 554.1875,
    "epsilon_norm": 7.0105719566345215,
    "prediction_norm": 6.473509788513184,
    "timestamp": "2025-07-14T12:35:10.633432"
  },
  {
    "epoch": 2691,
    "batch_idx": 128,
    "loss": 0.2346067726612091,
    "learning_rate": 0.00026672948912959414,
    "timestep_mean": 614.625,
    "epsilon_norm": 8.890657424926758,
    "prediction_norm": 8.101970672607422,
    "timestamp": "2025-07-14T12:35:10.638663"
  },
  {
    "epoch": 2691,
    "batch_idx": 160,
    "loss": 0.23089461028575897,
    "learning_rate": 0.00026672948912959414,
    "timestep_mean": 508.84375,
    "epsilon_norm": 7.574051856994629,
    "prediction_norm": 6.540607452392578,
    "timestamp": "2025-07-14T12:35:10.643475"
  },
  {
    "epoch": 2691,
    "batch_idx": 192,
    "loss": 0.07478567957878113,
    "learning_rate": 0.00026672948912959414,
    "timestep_mean": 702.375,
    "epsilon_norm": 4.097662925720215,
    "prediction_norm": 3.6472344398498535,
    "timestamp": "2025-07-14T12:35:10.648287"
  },
  {
    "epoch": 2692,
    "batch_idx": 0,
    "loss": 0.39715462923049927,
    "learning_rate": 0.00026742446096127126,
    "timestep_mean": 438.1875,
    "epsilon_norm": 9.467643737792969,
    "prediction_norm": 7.8308258056640625,
    "timestamp": "2025-07-14T12:35:10.653070"
  },
  {
    "epoch": 2692,
    "batch_idx": 32,
    "loss": 0.1778915375471115,
    "learning_rate": 0.00026742446096127126,
    "timestep_mean": 519.84375,
    "epsilon_norm": 6.717376232147217,
    "prediction_norm": 6.1858439445495605,
    "timestamp": "2025-07-14T12:35:10.657930"
  },
  {
    "epoch": 2692,
    "batch_idx": 64,
    "loss": 0.3727570176124573,
    "learning_rate": 0.00026742446096127126,
    "timestep_mean": 487.28125,
    "epsilon_norm": 8.792901039123535,
    "prediction_norm": 6.956530570983887,
    "timestamp": "2025-07-14T12:35:10.662788"
  },
  {
    "epoch": 2692,
    "batch_idx": 96,
    "loss": 0.1046309620141983,
    "learning_rate": 0.00026742446096127126,
    "timestep_mean": 629.78125,
    "epsilon_norm": 7.655029773712158,
    "prediction_norm": 7.108457088470459,
    "timestamp": "2025-07-14T12:35:10.667782"
  },
  {
    "epoch": 2692,
    "batch_idx": 128,
    "loss": 0.2332575023174286,
    "learning_rate": 0.00026742446096127126,
    "timestep_mean": 471.1875,
    "epsilon_norm": 6.884936809539795,
    "prediction_norm": 5.898587703704834,
    "timestamp": "2025-07-14T12:35:10.672919"
  },
  {
    "epoch": 2692,
    "batch_idx": 160,
    "loss": 0.16250364482402802,
    "learning_rate": 0.00026742446096127126,
    "timestep_mean": 606.25,
    "epsilon_norm": 9.229393005371094,
    "prediction_norm": 8.181751251220703,
    "timestamp": "2025-07-14T12:35:10.677991"
  },
  {
    "epoch": 2692,
    "batch_idx": 192,
    "loss": 0.19747024774551392,
    "learning_rate": 0.00026742446096127126,
    "timestep_mean": 564.25,
    "epsilon_norm": 3.963611364364624,
    "prediction_norm": 3.696619987487793,
    "timestamp": "2025-07-14T12:35:10.683122"
  },
  {
    "epoch": 2693,
    "batch_idx": 0,
    "loss": 0.13857129216194153,
    "learning_rate": 0.0002681200066499713,
    "timestep_mean": 627.59375,
    "epsilon_norm": 7.11187219619751,
    "prediction_norm": 6.65517520904541,
    "timestamp": "2025-07-14T12:35:10.688289"
  },
  {
    "epoch": 2693,
    "batch_idx": 32,
    "loss": 0.1868671178817749,
    "learning_rate": 0.0002681200066499713,
    "timestep_mean": 526.34375,
    "epsilon_norm": 6.571587562561035,
    "prediction_norm": 5.715107440948486,
    "timestamp": "2025-07-14T12:35:10.693410"
  },
  {
    "epoch": 2693,
    "batch_idx": 64,
    "loss": 0.25837114453315735,
    "learning_rate": 0.0002681200066499713,
    "timestep_mean": 478.40625,
    "epsilon_norm": 6.956940174102783,
    "prediction_norm": 5.709863662719727,
    "timestamp": "2025-07-14T12:35:10.698521"
  },
  {
    "epoch": 2693,
    "batch_idx": 96,
    "loss": 0.23642395436763763,
    "learning_rate": 0.0002681200066499713,
    "timestep_mean": 510.15625,
    "epsilon_norm": 7.5648112297058105,
    "prediction_norm": 6.400847911834717,
    "timestamp": "2025-07-14T12:35:10.703618"
  },
  {
    "epoch": 2693,
    "batch_idx": 128,
    "loss": 0.41182976961135864,
    "learning_rate": 0.0002681200066499713,
    "timestep_mean": 497.5,
    "epsilon_norm": 8.35244083404541,
    "prediction_norm": 6.982086658477783,
    "timestamp": "2025-07-14T12:35:10.708910"
  },
  {
    "epoch": 2693,
    "batch_idx": 160,
    "loss": 0.4239972233772278,
    "learning_rate": 0.0002681200066499713,
    "timestep_mean": 409.8125,
    "epsilon_norm": 8.674455642700195,
    "prediction_norm": 7.311583042144775,
    "timestamp": "2025-07-14T12:35:10.714020"
  },
  {
    "epoch": 2693,
    "batch_idx": 192,
    "loss": 0.4357590675354004,
    "learning_rate": 0.0002681200066499713,
    "timestep_mean": 340.875,
    "epsilon_norm": 3.6702213287353516,
    "prediction_norm": 2.498121738433838,
    "timestamp": "2025-07-14T12:35:10.719135"
  },
  {
    "epoch": 2694,
    "batch_idx": 0,
    "loss": 0.1664341539144516,
    "learning_rate": 0.00026881612447950484,
    "timestep_mean": 431.8125,
    "epsilon_norm": 7.740805625915527,
    "prediction_norm": 6.917115688323975,
    "timestamp": "2025-07-14T12:35:10.724241"
  },
  {
    "epoch": 2694,
    "batch_idx": 32,
    "loss": 0.17903360724449158,
    "learning_rate": 0.00026881612447950484,
    "timestep_mean": 473.625,
    "epsilon_norm": 8.356191635131836,
    "prediction_norm": 7.0051589012146,
    "timestamp": "2025-07-14T12:35:10.729434"
  },
  {
    "epoch": 2694,
    "batch_idx": 64,
    "loss": 0.1036682054400444,
    "learning_rate": 0.00026881612447950484,
    "timestep_mean": 516.09375,
    "epsilon_norm": 7.56195592880249,
    "prediction_norm": 7.046565532684326,
    "timestamp": "2025-07-14T12:35:10.734665"
  },
  {
    "epoch": 2694,
    "batch_idx": 96,
    "loss": 0.5248501300811768,
    "learning_rate": 0.00026881612447950484,
    "timestep_mean": 489.3125,
    "epsilon_norm": 8.42809772491455,
    "prediction_norm": 6.036344528198242,
    "timestamp": "2025-07-14T12:35:10.739942"
  },
  {
    "epoch": 2694,
    "batch_idx": 128,
    "loss": 0.17760303616523743,
    "learning_rate": 0.00026881612447950484,
    "timestep_mean": 536.8125,
    "epsilon_norm": 8.403242111206055,
    "prediction_norm": 7.700807571411133,
    "timestamp": "2025-07-14T12:35:10.745113"
  },
  {
    "epoch": 2694,
    "batch_idx": 160,
    "loss": 0.1428465098142624,
    "learning_rate": 0.00026881612447950484,
    "timestep_mean": 555.0625,
    "epsilon_norm": 8.864875793457031,
    "prediction_norm": 7.607769966125488,
    "timestamp": "2025-07-14T12:35:10.750373"
  },
  {
    "epoch": 2694,
    "batch_idx": 192,
    "loss": 0.09099258482456207,
    "learning_rate": 0.00026881612447950484,
    "timestep_mean": 626.5,
    "epsilon_norm": 4.63279390335083,
    "prediction_norm": 4.347873210906982,
    "timestamp": "2025-07-14T12:35:10.755518"
  },
  {
    "epoch": 2695,
    "batch_idx": 0,
    "loss": 0.16515997052192688,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 526.21875,
    "epsilon_norm": 7.242062568664551,
    "prediction_norm": 6.434131145477295,
    "timestamp": "2025-07-14T12:35:10.760761"
  },
  {
    "epoch": 2695,
    "batch_idx": 32,
    "loss": 0.2861882448196411,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 436.8125,
    "epsilon_norm": 8.493108749389648,
    "prediction_norm": 7.045404434204102,
    "timestamp": "2025-07-14T12:35:10.765985"
  },
  {
    "epoch": 2695,
    "batch_idx": 64,
    "loss": 0.3166730999946594,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 441.53125,
    "epsilon_norm": 8.452400207519531,
    "prediction_norm": 6.807448863983154,
    "timestamp": "2025-07-14T12:35:10.771099"
  },
  {
    "epoch": 2695,
    "batch_idx": 96,
    "loss": 0.34996408224105835,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 493.84375,
    "epsilon_norm": 8.54775619506836,
    "prediction_norm": 7.21299934387207,
    "timestamp": "2025-07-14T12:35:10.776306"
  },
  {
    "epoch": 2695,
    "batch_idx": 128,
    "loss": 0.12281689047813416,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 570.6875,
    "epsilon_norm": 7.811450004577637,
    "prediction_norm": 7.683614730834961,
    "timestamp": "2025-07-14T12:35:10.781521"
  },
  {
    "epoch": 2695,
    "batch_idx": 160,
    "loss": 0.3275887668132782,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 542.625,
    "epsilon_norm": 8.49068832397461,
    "prediction_norm": 6.997288703918457,
    "timestamp": "2025-07-14T12:35:10.786784"
  },
  {
    "epoch": 2695,
    "batch_idx": 192,
    "loss": 0.07128509879112244,
    "learning_rate": 0.0002695128127322691,
    "timestep_mean": 580.875,
    "epsilon_norm": 2.736701488494873,
    "prediction_norm": 2.8681211471557617,
    "timestamp": "2025-07-14T12:35:10.792069"
  },
  {
    "epoch": 2696,
    "batch_idx": 0,
    "loss": 0.2010076344013214,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 521.75,
    "epsilon_norm": 7.363034725189209,
    "prediction_norm": 6.450082302093506,
    "timestamp": "2025-07-14T12:35:10.797303"
  },
  {
    "epoch": 2696,
    "batch_idx": 32,
    "loss": 0.222215473651886,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 572.65625,
    "epsilon_norm": 7.776243209838867,
    "prediction_norm": 6.9393134117126465,
    "timestamp": "2025-07-14T12:35:10.802452"
  },
  {
    "epoch": 2696,
    "batch_idx": 64,
    "loss": 0.14913201332092285,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 529.96875,
    "epsilon_norm": 7.18159818649292,
    "prediction_norm": 6.834776878356934,
    "timestamp": "2025-07-14T12:35:10.807685"
  },
  {
    "epoch": 2696,
    "batch_idx": 96,
    "loss": 0.32056716084480286,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 513.21875,
    "epsilon_norm": 8.423907279968262,
    "prediction_norm": 7.499921798706055,
    "timestamp": "2025-07-14T12:35:10.812857"
  },
  {
    "epoch": 2696,
    "batch_idx": 128,
    "loss": 0.2776404917240143,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 538.625,
    "epsilon_norm": 8.328999519348145,
    "prediction_norm": 7.018411636352539,
    "timestamp": "2025-07-14T12:35:10.817984"
  },
  {
    "epoch": 2696,
    "batch_idx": 160,
    "loss": 0.32533448934555054,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 485.6875,
    "epsilon_norm": 8.014042854309082,
    "prediction_norm": 6.375101566314697,
    "timestamp": "2025-07-14T12:35:10.823225"
  },
  {
    "epoch": 2696,
    "batch_idx": 192,
    "loss": 0.19032028317451477,
    "learning_rate": 0.00027021006968925627,
    "timestep_mean": 400.75,
    "epsilon_norm": 4.372537136077881,
    "prediction_norm": 3.5797722339630127,
    "timestamp": "2025-07-14T12:35:10.828381"
  },
  {
    "epoch": 2697,
    "batch_idx": 0,
    "loss": 0.3329415023326874,
    "learning_rate": 0.00027090789363005415,
    "timestep_mean": 421.96875,
    "epsilon_norm": 8.941572189331055,
    "prediction_norm": 7.3721022605896,
    "timestamp": "2025-07-14T12:35:10.833655"
  },
  {
    "epoch": 2697,
    "batch_idx": 32,
    "loss": 0.46370983123779297,
    "learning_rate": 0.00027090789363005415,
    "timestep_mean": 550.125,
    "epsilon_norm": 7.465625762939453,
    "prediction_norm": 5.524097919464111,
    "timestamp": "2025-07-14T12:35:10.838942"
  },
  {
    "epoch": 2697,
    "batch_idx": 64,
    "loss": 0.28142261505126953,
    "learning_rate": 0.00027090789363005415,
    "timestep_mean": 525.21875,
    "epsilon_norm": 8.310327529907227,
    "prediction_norm": 6.407751083374023,
    "timestamp": "2025-07-14T12:35:10.844193"
  },
  {
    "epoch": 2697,
    "batch_idx": 96,
    "loss": 0.239979088306427,
    "learning_rate": 0.00027090789363005415,
    "timestep_mean": 474.75,
    "epsilon_norm": 7.717958450317383,
    "prediction_norm": 6.07688045501709,
    "timestamp": "2025-07-14T12:35:10.849330"
  },
  {
    "epoch": 2697,
    "batch_idx": 128,
    "loss": 0.19665570557117462,
    "learning_rate": 0.00027090789363005415,
    "timestep_mean": 494.09375,
    "epsilon_norm": 7.8536224365234375,
    "prediction_norm": 6.694620132446289,
    "timestamp": "2025-07-14T12:35:10.854503"
  },
  {
    "epoch": 2697,
    "batch_idx": 160,
    "loss": 0.351950466632843,
    "learning_rate": 0.00027090789363005415,
    "timestep_mean": 511.9375,
    "epsilon_norm": 8.541635513305664,
    "prediction_norm": 6.973288536071777,
    "timestamp": "2025-07-14T12:35:10.859725"
  },
  {
    "epoch": 2697,
    "batch_idx": 192,
    "loss": 0.32500946521759033,
    "learning_rate": 0.00027090789363005415,
    "timestep_mean": 530.125,
    "epsilon_norm": 4.0329437255859375,
    "prediction_norm": 3.8758106231689453,
    "timestamp": "2025-07-14T12:35:10.864860"
  },
  {
    "epoch": 2698,
    "batch_idx": 0,
    "loss": 0.3526431620121002,
    "learning_rate": 0.0002716062828328508,
    "timestep_mean": 539.09375,
    "epsilon_norm": 8.115638732910156,
    "prediction_norm": 6.84365177154541,
    "timestamp": "2025-07-14T12:35:10.870033"
  },
  {
    "epoch": 2698,
    "batch_idx": 32,
    "loss": 0.28560295701026917,
    "learning_rate": 0.0002716062828328508,
    "timestep_mean": 513.96875,
    "epsilon_norm": 7.228762626647949,
    "prediction_norm": 6.121595859527588,
    "timestamp": "2025-07-14T12:35:10.875359"
  },
  {
    "epoch": 2698,
    "batch_idx": 64,
    "loss": 0.5146001577377319,
    "learning_rate": 0.0002716062828328508,
    "timestep_mean": 448.5,
    "epsilon_norm": 8.166963577270508,
    "prediction_norm": 5.563836097717285,
    "timestamp": "2025-07-14T12:35:10.880510"
  },
  {
    "epoch": 2698,
    "batch_idx": 96,
    "loss": 0.31730878353118896,
    "learning_rate": 0.0002716062828328508,
    "timestep_mean": 493.84375,
    "epsilon_norm": 8.61383056640625,
    "prediction_norm": 7.534697532653809,
    "timestamp": "2025-07-14T12:35:10.885665"
  },
  {
    "epoch": 2698,
    "batch_idx": 128,
    "loss": 0.32148823142051697,
    "learning_rate": 0.0002716062828328508,
    "timestep_mean": 526.125,
    "epsilon_norm": 7.923185348510742,
    "prediction_norm": 7.2765278816223145,
    "timestamp": "2025-07-14T12:35:10.890912"
  },
  {
    "epoch": 2698,
    "batch_idx": 160,
    "loss": 0.2892606854438782,
    "learning_rate": 0.0002716062828328508,
    "timestep_mean": 466.3125,
    "epsilon_norm": 7.89301872253418,
    "prediction_norm": 7.153820514678955,
    "timestamp": "2025-07-14T12:35:10.896119"
  },
  {
    "epoch": 2698,
    "batch_idx": 192,
    "loss": 0.1675155758857727,
    "learning_rate": 0.0002716062828328508,
    "timestep_mean": 548.25,
    "epsilon_norm": 3.433645725250244,
    "prediction_norm": 3.189927101135254,
    "timestamp": "2025-07-14T12:35:10.901309"
  },
  {
    "epoch": 2699,
    "batch_idx": 0,
    "loss": 0.21648122370243073,
    "learning_rate": 0.0002723052355744406,
    "timestep_mean": 439.875,
    "epsilon_norm": 7.534762382507324,
    "prediction_norm": 7.08860445022583,
    "timestamp": "2025-07-14T12:35:10.906516"
  },
  {
    "epoch": 2699,
    "batch_idx": 32,
    "loss": 0.171665757894516,
    "learning_rate": 0.0002723052355744406,
    "timestep_mean": 557.6875,
    "epsilon_norm": 7.347033500671387,
    "prediction_norm": 6.691325664520264,
    "timestamp": "2025-07-14T12:35:10.911661"
  },
  {
    "epoch": 2699,
    "batch_idx": 64,
    "loss": 0.18739154934883118,
    "learning_rate": 0.0002723052355744406,
    "timestep_mean": 484.59375,
    "epsilon_norm": 7.130026817321777,
    "prediction_norm": 6.776333332061768,
    "timestamp": "2025-07-14T12:35:10.916984"
  },
  {
    "epoch": 2699,
    "batch_idx": 96,
    "loss": 0.3336840271949768,
    "learning_rate": 0.0002723052355744406,
    "timestep_mean": 418.125,
    "epsilon_norm": 8.009521484375,
    "prediction_norm": 6.715677738189697,
    "timestamp": "2025-07-14T12:35:10.922225"
  },
  {
    "epoch": 2699,
    "batch_idx": 128,
    "loss": 0.6669488549232483,
    "learning_rate": 0.0002723052355744406,
    "timestep_mean": 352.3125,
    "epsilon_norm": 8.186964988708496,
    "prediction_norm": 5.929096698760986,
    "timestamp": "2025-07-14T12:35:10.927404"
  },
  {
    "epoch": 2699,
    "batch_idx": 160,
    "loss": 0.22732174396514893,
    "learning_rate": 0.0002723052355744406,
    "timestep_mean": 597.625,
    "epsilon_norm": 7.865618705749512,
    "prediction_norm": 7.656993865966797,
    "timestamp": "2025-07-14T12:35:10.932639"
  },
  {
    "epoch": 2699,
    "batch_idx": 192,
    "loss": 0.622252881526947,
    "learning_rate": 0.0002723052355744406,
    "timestep_mean": 366.0,
    "epsilon_norm": 4.7113566398620605,
    "prediction_norm": 3.5164318084716797,
    "timestamp": "2025-07-14T12:35:10.937799"
  },
  {
    "epoch": 2700,
    "batch_idx": 0,
    "loss": 0.22956764698028564,
    "learning_rate": 0.0002730047501302271,
    "timestep_mean": 493.6875,
    "epsilon_norm": 6.928017616271973,
    "prediction_norm": 5.576104640960693,
    "timestamp": "2025-07-14T12:35:10.943019"
  },
  {
    "epoch": 2700,
    "batch_idx": 32,
    "loss": 0.34010621905326843,
    "learning_rate": 0.0002730047501302271,
    "timestep_mean": 445.875,
    "epsilon_norm": 8.830029487609863,
    "prediction_norm": 6.637082099914551,
    "timestamp": "2025-07-14T12:35:10.948070"
  },
  {
    "epoch": 2700,
    "batch_idx": 64,
    "loss": 0.24812249839305878,
    "learning_rate": 0.0002730047501302271,
    "timestep_mean": 458.5625,
    "epsilon_norm": 7.523047924041748,
    "prediction_norm": 5.989935398101807,
    "timestamp": "2025-07-14T12:35:10.952953"
  },
  {
    "epoch": 2700,
    "batch_idx": 96,
    "loss": 0.3730820417404175,
    "learning_rate": 0.0002730047501302271,
    "timestep_mean": 463.65625,
    "epsilon_norm": 7.770921230316162,
    "prediction_norm": 6.402722358703613,
    "timestamp": "2025-07-14T12:35:10.957808"
  },
  {
    "epoch": 2700,
    "batch_idx": 128,
    "loss": 0.2667035460472107,
    "learning_rate": 0.0002730047501302271,
    "timestep_mean": 561.625,
    "epsilon_norm": 6.880871772766113,
    "prediction_norm": 5.64962100982666,
    "timestamp": "2025-07-14T12:35:10.962656"
  },
  {
    "epoch": 2700,
    "batch_idx": 160,
    "loss": 0.19730471074581146,
    "learning_rate": 0.0002730047501302271,
    "timestep_mean": 592.90625,
    "epsilon_norm": 7.138515949249268,
    "prediction_norm": 6.030799865722656,
    "timestamp": "2025-07-14T12:35:10.967579"
  },
  {
    "epoch": 2700,
    "batch_idx": 192,
    "loss": 0.17086948454380035,
    "learning_rate": 0.0002730047501302271,
    "timestep_mean": 495.875,
    "epsilon_norm": 5.071913242340088,
    "prediction_norm": 3.9639482498168945,
    "timestamp": "2025-07-14T12:35:10.972484"
  },
  {
    "epoch": 2701,
    "batch_idx": 0,
    "loss": 0.06588434427976608,
    "learning_rate": 0.0002737048247742278,
    "timestep_mean": 533.375,
    "epsilon_norm": 7.385450839996338,
    "prediction_norm": 6.916980743408203,
    "timestamp": "2025-07-14T12:35:10.977718"
  },
  {
    "epoch": 2701,
    "batch_idx": 32,
    "loss": 0.37139326333999634,
    "learning_rate": 0.0002737048247742278,
    "timestep_mean": 569.5,
    "epsilon_norm": 8.862767219543457,
    "prediction_norm": 7.1812028884887695,
    "timestamp": "2025-07-14T12:35:10.982859"
  },
  {
    "epoch": 2701,
    "batch_idx": 64,
    "loss": 0.29208487272262573,
    "learning_rate": 0.0002737048247742278,
    "timestep_mean": 471.78125,
    "epsilon_norm": 6.691104888916016,
    "prediction_norm": 5.81031608581543,
    "timestamp": "2025-07-14T12:35:10.988021"
  },
  {
    "epoch": 2701,
    "batch_idx": 96,
    "loss": 0.21625593304634094,
    "learning_rate": 0.0002737048247742278,
    "timestep_mean": 526.03125,
    "epsilon_norm": 6.375790596008301,
    "prediction_norm": 5.873763084411621,
    "timestamp": "2025-07-14T12:35:10.993046"
  },
  {
    "epoch": 2701,
    "batch_idx": 128,
    "loss": 0.30349430441856384,
    "learning_rate": 0.0002737048247742278,
    "timestep_mean": 429.125,
    "epsilon_norm": 8.06960678100586,
    "prediction_norm": 6.317164897918701,
    "timestamp": "2025-07-14T12:35:10.998050"
  },
  {
    "epoch": 2701,
    "batch_idx": 160,
    "loss": 0.19025857746601105,
    "learning_rate": 0.0002737048247742278,
    "timestep_mean": 512.875,
    "epsilon_norm": 7.424136161804199,
    "prediction_norm": 6.1810221672058105,
    "timestamp": "2025-07-14T12:35:11.003197"
  },
  {
    "epoch": 2701,
    "batch_idx": 192,
    "loss": 0.23875273764133453,
    "learning_rate": 0.0002737048247742278,
    "timestep_mean": 444.375,
    "epsilon_norm": 3.2867653369903564,
    "prediction_norm": 2.833505630493164,
    "timestamp": "2025-07-14T12:35:11.008333"
  },
  {
    "epoch": 2702,
    "batch_idx": 0,
    "loss": 0.2832881808280945,
    "learning_rate": 0.0002744054577790779,
    "timestep_mean": 499.6875,
    "epsilon_norm": 8.210071563720703,
    "prediction_norm": 7.0754313468933105,
    "timestamp": "2025-07-14T12:35:11.013645"
  },
  {
    "epoch": 2702,
    "batch_idx": 32,
    "loss": 0.23769420385360718,
    "learning_rate": 0.0002744054577790779,
    "timestep_mean": 487.78125,
    "epsilon_norm": 7.971916675567627,
    "prediction_norm": 6.760983943939209,
    "timestamp": "2025-07-14T12:35:11.018812"
  },
  {
    "epoch": 2702,
    "batch_idx": 64,
    "loss": 0.18465358018875122,
    "learning_rate": 0.0002744054577790779,
    "timestep_mean": 494.1875,
    "epsilon_norm": 7.265573978424072,
    "prediction_norm": 5.9828009605407715,
    "timestamp": "2025-07-14T12:35:11.024056"
  },
  {
    "epoch": 2702,
    "batch_idx": 96,
    "loss": 0.22907501459121704,
    "learning_rate": 0.0002744054577790779,
    "timestep_mean": 539.8125,
    "epsilon_norm": 8.585199356079102,
    "prediction_norm": 7.406546592712402,
    "timestamp": "2025-07-14T12:35:11.029052"
  },
  {
    "epoch": 2702,
    "batch_idx": 128,
    "loss": 0.3154056668281555,
    "learning_rate": 0.0002744054577790779,
    "timestep_mean": 546.28125,
    "epsilon_norm": 7.977010726928711,
    "prediction_norm": 6.280817031860352,
    "timestamp": "2025-07-14T12:35:11.034158"
  },
  {
    "epoch": 2702,
    "batch_idx": 160,
    "loss": 0.24979937076568604,
    "learning_rate": 0.0002744054577790779,
    "timestep_mean": 507.75,
    "epsilon_norm": 8.210677146911621,
    "prediction_norm": 6.940141677856445,
    "timestamp": "2025-07-14T12:35:11.039280"
  },
  {
    "epoch": 2702,
    "batch_idx": 192,
    "loss": 0.20260517299175262,
    "learning_rate": 0.0002744054577790779,
    "timestep_mean": 671.125,
    "epsilon_norm": 3.0085113048553467,
    "prediction_norm": 2.7560534477233887,
    "timestamp": "2025-07-14T12:35:11.044446"
  },
  {
    "epoch": 2703,
    "batch_idx": 0,
    "loss": 0.4194597005844116,
    "learning_rate": 0.00027510664741603534,
    "timestep_mean": 439.40625,
    "epsilon_norm": 8.272520065307617,
    "prediction_norm": 6.442867279052734,
    "timestamp": "2025-07-14T12:35:11.049660"
  },
  {
    "epoch": 2703,
    "batch_idx": 32,
    "loss": 0.19805939495563507,
    "learning_rate": 0.00027510664741603534,
    "timestep_mean": 576.34375,
    "epsilon_norm": 7.799119472503662,
    "prediction_norm": 6.936436176300049,
    "timestamp": "2025-07-14T12:35:11.055000"
  },
  {
    "epoch": 2703,
    "batch_idx": 64,
    "loss": 0.33407163619995117,
    "learning_rate": 0.00027510664741603534,
    "timestep_mean": 471.65625,
    "epsilon_norm": 8.515520095825195,
    "prediction_norm": 6.968239784240723,
    "timestamp": "2025-07-14T12:35:11.060155"
  },
  {
    "epoch": 2703,
    "batch_idx": 96,
    "loss": 0.4702942371368408,
    "learning_rate": 0.00027510664741603534,
    "timestep_mean": 439.25,
    "epsilon_norm": 8.496211051940918,
    "prediction_norm": 7.209918022155762,
    "timestamp": "2025-07-14T12:35:11.065386"
  },
  {
    "epoch": 2703,
    "batch_idx": 128,
    "loss": 0.23764219880104065,
    "learning_rate": 0.00027510664741603534,
    "timestep_mean": 588.8125,
    "epsilon_norm": 8.153959274291992,
    "prediction_norm": 7.011035919189453,
    "timestamp": "2025-07-14T12:35:11.070613"
  },
  {
    "epoch": 2703,
    "batch_idx": 160,
    "loss": 0.2611280381679535,
    "learning_rate": 0.00027510664741603534,
    "timestep_mean": 461.53125,
    "epsilon_norm": 8.432162284851074,
    "prediction_norm": 7.460681438446045,
    "timestamp": "2025-07-14T12:35:11.075881"
  },
  {
    "epoch": 2703,
    "batch_idx": 192,
    "loss": 0.1372649073600769,
    "learning_rate": 0.00027510664741603534,
    "timestep_mean": 677.0,
    "epsilon_norm": 4.040350437164307,
    "prediction_norm": 3.842589855194092,
    "timestamp": "2025-07-14T12:35:11.080910"
  },
  {
    "epoch": 2704,
    "batch_idx": 0,
    "loss": 0.20332351326942444,
    "learning_rate": 0.00027580839195498464,
    "timestep_mean": 467.875,
    "epsilon_norm": 6.599565029144287,
    "prediction_norm": 5.193033218383789,
    "timestamp": "2025-07-14T12:35:11.086114"
  },
  {
    "epoch": 2704,
    "batch_idx": 32,
    "loss": 0.1610766053199768,
    "learning_rate": 0.00027580839195498464,
    "timestep_mean": 536.09375,
    "epsilon_norm": 7.564566135406494,
    "prediction_norm": 7.117117404937744,
    "timestamp": "2025-07-14T12:35:11.091296"
  },
  {
    "epoch": 2704,
    "batch_idx": 64,
    "loss": 0.20977561175823212,
    "learning_rate": 0.00027580839195498464,
    "timestep_mean": 447.96875,
    "epsilon_norm": 7.577036380767822,
    "prediction_norm": 7.2730865478515625,
    "timestamp": "2025-07-14T12:35:11.096633"
  },
  {
    "epoch": 2704,
    "batch_idx": 96,
    "loss": 0.2394116371870041,
    "learning_rate": 0.00027580839195498464,
    "timestep_mean": 482.53125,
    "epsilon_norm": 7.288229465484619,
    "prediction_norm": 6.455117225646973,
    "timestamp": "2025-07-14T12:35:11.101864"
  },
  {
    "epoch": 2704,
    "batch_idx": 128,
    "loss": 0.30314502120018005,
    "learning_rate": 0.00027580839195498464,
    "timestep_mean": 429.5625,
    "epsilon_norm": 8.02270221710205,
    "prediction_norm": 6.418545722961426,
    "timestamp": "2025-07-14T12:35:11.107075"
  },
  {
    "epoch": 2704,
    "batch_idx": 160,
    "loss": 0.519220769405365,
    "learning_rate": 0.00027580839195498464,
    "timestep_mean": 459.84375,
    "epsilon_norm": 8.697999954223633,
    "prediction_norm": 6.742981910705566,
    "timestamp": "2025-07-14T12:35:11.112216"
  },
  {
    "epoch": 2704,
    "batch_idx": 192,
    "loss": 0.13212530314922333,
    "learning_rate": 0.00027580839195498464,
    "timestep_mean": 443.375,
    "epsilon_norm": 4.243997573852539,
    "prediction_norm": 3.744899272918701,
    "timestamp": "2025-07-14T12:35:11.117550"
  },
  {
    "epoch": 2705,
    "batch_idx": 0,
    "loss": 0.2219177484512329,
    "learning_rate": 0.0002765106896644402,
    "timestep_mean": 515.4375,
    "epsilon_norm": 7.786224365234375,
    "prediction_norm": 6.847344398498535,
    "timestamp": "2025-07-14T12:35:11.122808"
  },
  {
    "epoch": 2705,
    "batch_idx": 32,
    "loss": 0.20355373620986938,
    "learning_rate": 0.0002765106896644402,
    "timestep_mean": 511.40625,
    "epsilon_norm": 7.732751846313477,
    "prediction_norm": 6.921631336212158,
    "timestamp": "2025-07-14T12:35:11.128000"
  },
  {
    "epoch": 2705,
    "batch_idx": 64,
    "loss": 0.1435707062482834,
    "learning_rate": 0.0002765106896644402,
    "timestep_mean": 534.8125,
    "epsilon_norm": 7.731553554534912,
    "prediction_norm": 7.5496015548706055,
    "timestamp": "2025-07-14T12:35:11.133217"
  },
  {
    "epoch": 2705,
    "batch_idx": 96,
    "loss": 0.2914464771747589,
    "learning_rate": 0.0002765106896644402,
    "timestep_mean": 442.53125,
    "epsilon_norm": 8.669548988342285,
    "prediction_norm": 7.814416408538818,
    "timestamp": "2025-07-14T12:35:11.138562"
  },
  {
    "epoch": 2705,
    "batch_idx": 128,
    "loss": 0.48379427194595337,
    "learning_rate": 0.0002765106896644402,
    "timestep_mean": 503.65625,
    "epsilon_norm": 9.57298755645752,
    "prediction_norm": 7.719655990600586,
    "timestamp": "2025-07-14T12:35:11.143552"
  },
  {
    "epoch": 2705,
    "batch_idx": 160,
    "loss": 0.28802943229675293,
    "learning_rate": 0.0002765106896644402,
    "timestep_mean": 393.8125,
    "epsilon_norm": 6.86224889755249,
    "prediction_norm": 5.238869667053223,
    "timestamp": "2025-07-14T12:35:11.148409"
  },
  {
    "epoch": 2705,
    "batch_idx": 192,
    "loss": 0.09314389526844025,
    "learning_rate": 0.0002765106896644402,
    "timestep_mean": 200.625,
    "epsilon_norm": 1.9985231161117554,
    "prediction_norm": 1.665452241897583,
    "timestamp": "2025-07-14T12:35:11.153375"
  },
  {
    "epoch": 2706,
    "batch_idx": 0,
    "loss": 0.14156752824783325,
    "learning_rate": 0.0002772135388115522,
    "timestep_mean": 519.15625,
    "epsilon_norm": 8.079524993896484,
    "prediction_norm": 7.580836772918701,
    "timestamp": "2025-07-14T12:35:11.158704"
  },
  {
    "epoch": 2706,
    "batch_idx": 32,
    "loss": 0.31011587381362915,
    "learning_rate": 0.0002772135388115522,
    "timestep_mean": 432.53125,
    "epsilon_norm": 7.550835609436035,
    "prediction_norm": 6.065539836883545,
    "timestamp": "2025-07-14T12:35:11.163828"
  },
  {
    "epoch": 2706,
    "batch_idx": 64,
    "loss": 0.2507213354110718,
    "learning_rate": 0.0002772135388115522,
    "timestep_mean": 498.84375,
    "epsilon_norm": 8.4549560546875,
    "prediction_norm": 7.570783615112305,
    "timestamp": "2025-07-14T12:35:11.169007"
  },
  {
    "epoch": 2706,
    "batch_idx": 96,
    "loss": 0.27136069536209106,
    "learning_rate": 0.0002772135388115522,
    "timestep_mean": 518.5625,
    "epsilon_norm": 7.582152843475342,
    "prediction_norm": 6.449938774108887,
    "timestamp": "2025-07-14T12:35:11.174218"
  },
  {
    "epoch": 2706,
    "batch_idx": 128,
    "loss": 0.2780243158340454,
    "learning_rate": 0.0002772135388115522,
    "timestep_mean": 500.59375,
    "epsilon_norm": 7.912546634674072,
    "prediction_norm": 6.543354511260986,
    "timestamp": "2025-07-14T12:35:11.179227"
  },
  {
    "epoch": 2706,
    "batch_idx": 160,
    "loss": 0.18599992990493774,
    "learning_rate": 0.0002772135388115522,
    "timestep_mean": 488.6875,
    "epsilon_norm": 6.449869155883789,
    "prediction_norm": 5.554630279541016,
    "timestamp": "2025-07-14T12:35:11.184557"
  },
  {
    "epoch": 2706,
    "batch_idx": 192,
    "loss": 0.1750885397195816,
    "learning_rate": 0.0002772135388115522,
    "timestep_mean": 625.875,
    "epsilon_norm": 4.0299973487854,
    "prediction_norm": 3.434903383255005,
    "timestamp": "2025-07-14T12:35:11.189590"
  },
  {
    "epoch": 2707,
    "batch_idx": 0,
    "loss": 0.3103296160697937,
    "learning_rate": 0.000277916937662111,
    "timestep_mean": 476.4375,
    "epsilon_norm": 7.395664215087891,
    "prediction_norm": 5.914638042449951,
    "timestamp": "2025-07-14T12:35:11.194475"
  },
  {
    "epoch": 2707,
    "batch_idx": 32,
    "loss": 0.2056090384721756,
    "learning_rate": 0.000277916937662111,
    "timestep_mean": 463.375,
    "epsilon_norm": 7.669148921966553,
    "prediction_norm": 7.157207012176514,
    "timestamp": "2025-07-14T12:35:11.199484"
  },
  {
    "epoch": 2707,
    "batch_idx": 64,
    "loss": 0.29983454942703247,
    "learning_rate": 0.000277916937662111,
    "timestep_mean": 541.75,
    "epsilon_norm": 8.445955276489258,
    "prediction_norm": 7.16722297668457,
    "timestamp": "2025-07-14T12:35:11.204596"
  },
  {
    "epoch": 2707,
    "batch_idx": 96,
    "loss": 0.1598975509405136,
    "learning_rate": 0.000277916937662111,
    "timestep_mean": 501.1875,
    "epsilon_norm": 6.329832077026367,
    "prediction_norm": 5.427835464477539,
    "timestamp": "2025-07-14T12:35:11.209651"
  },
  {
    "epoch": 2707,
    "batch_idx": 128,
    "loss": 0.2526302933692932,
    "learning_rate": 0.000277916937662111,
    "timestep_mean": 498.03125,
    "epsilon_norm": 8.282161712646484,
    "prediction_norm": 6.798518180847168,
    "timestamp": "2025-07-14T12:35:11.221000"
  },
  {
    "epoch": 2707,
    "batch_idx": 160,
    "loss": 0.4265088438987732,
    "learning_rate": 0.000277916937662111,
    "timestep_mean": 422.34375,
    "epsilon_norm": 8.434602737426758,
    "prediction_norm": 6.065800189971924,
    "timestamp": "2025-07-14T12:35:11.227013"
  },
  {
    "epoch": 2707,
    "batch_idx": 192,
    "loss": 0.05446828156709671,
    "learning_rate": 0.000277916937662111,
    "timestep_mean": 532.125,
    "epsilon_norm": 3.9158976078033447,
    "prediction_norm": 3.5944180488586426,
    "timestamp": "2025-07-14T12:35:11.232418"
  },
  {
    "epoch": 2708,
    "batch_idx": 0,
    "loss": 0.21607264876365662,
    "learning_rate": 0.0002786208844805499,
    "timestep_mean": 548.78125,
    "epsilon_norm": 7.8988423347473145,
    "prediction_norm": 6.5903449058532715,
    "timestamp": "2025-07-14T12:35:11.237699"
  },
  {
    "epoch": 2708,
    "batch_idx": 32,
    "loss": 0.21590286493301392,
    "learning_rate": 0.0002786208844805499,
    "timestep_mean": 545.6875,
    "epsilon_norm": 8.26200008392334,
    "prediction_norm": 6.7143473625183105,
    "timestamp": "2025-07-14T12:35:11.242930"
  },
  {
    "epoch": 2708,
    "batch_idx": 64,
    "loss": 0.27481895685195923,
    "learning_rate": 0.0002786208844805499,
    "timestep_mean": 534.6875,
    "epsilon_norm": 7.88264799118042,
    "prediction_norm": 6.886811256408691,
    "timestamp": "2025-07-14T12:35:11.247965"
  },
  {
    "epoch": 2708,
    "batch_idx": 96,
    "loss": 0.3239594101905823,
    "learning_rate": 0.0002786208844805499,
    "timestep_mean": 440.09375,
    "epsilon_norm": 7.8418989181518555,
    "prediction_norm": 6.241430759429932,
    "timestamp": "2025-07-14T12:35:11.252830"
  },
  {
    "epoch": 2708,
    "batch_idx": 128,
    "loss": 0.23153294622898102,
    "learning_rate": 0.0002786208844805499,
    "timestep_mean": 470.625,
    "epsilon_norm": 7.804384231567383,
    "prediction_norm": 6.777354717254639,
    "timestamp": "2025-07-14T12:35:11.257696"
  },
  {
    "epoch": 2708,
    "batch_idx": 160,
    "loss": 0.29728662967681885,
    "learning_rate": 0.0002786208844805499,
    "timestep_mean": 474.4375,
    "epsilon_norm": 9.375673294067383,
    "prediction_norm": 7.552515983581543,
    "timestamp": "2025-07-14T12:35:11.262596"
  },
  {
    "epoch": 2708,
    "batch_idx": 192,
    "loss": 0.6577221155166626,
    "learning_rate": 0.0002786208844805499,
    "timestep_mean": 619.5,
    "epsilon_norm": 4.481027603149414,
    "prediction_norm": 3.3120200634002686,
    "timestamp": "2025-07-14T12:35:11.267689"
  },
  {
    "epoch": 2709,
    "batch_idx": 0,
    "loss": 0.24669817090034485,
    "learning_rate": 0.0002793253775299492,
    "timestep_mean": 511.21875,
    "epsilon_norm": 9.040815353393555,
    "prediction_norm": 8.405293464660645,
    "timestamp": "2025-07-14T12:35:11.272893"
  },
  {
    "epoch": 2709,
    "batch_idx": 32,
    "loss": 0.21229895949363708,
    "learning_rate": 0.0002793253775299492,
    "timestep_mean": 522.21875,
    "epsilon_norm": 7.944886684417725,
    "prediction_norm": 7.089852333068848,
    "timestamp": "2025-07-14T12:35:11.277966"
  },
  {
    "epoch": 2709,
    "batch_idx": 64,
    "loss": 0.1850304901599884,
    "learning_rate": 0.0002793253775299492,
    "timestep_mean": 469.34375,
    "epsilon_norm": 7.779906272888184,
    "prediction_norm": 6.711308002471924,
    "timestamp": "2025-07-14T12:35:11.283063"
  },
  {
    "epoch": 2709,
    "batch_idx": 96,
    "loss": 0.1941538155078888,
    "learning_rate": 0.0002793253775299492,
    "timestep_mean": 548.1875,
    "epsilon_norm": 7.2568464279174805,
    "prediction_norm": 6.1842803955078125,
    "timestamp": "2025-07-14T12:35:11.288292"
  },
  {
    "epoch": 2709,
    "batch_idx": 128,
    "loss": 0.20201152563095093,
    "learning_rate": 0.0002793253775299492,
    "timestep_mean": 537.46875,
    "epsilon_norm": 8.529163360595703,
    "prediction_norm": 7.701755523681641,
    "timestamp": "2025-07-14T12:35:11.293318"
  },
  {
    "epoch": 2709,
    "batch_idx": 160,
    "loss": 0.43614107370376587,
    "learning_rate": 0.0002793253775299492,
    "timestep_mean": 537.84375,
    "epsilon_norm": 9.049846649169922,
    "prediction_norm": 7.317127227783203,
    "timestamp": "2025-07-14T12:35:11.298508"
  },
  {
    "epoch": 2709,
    "batch_idx": 192,
    "loss": 0.9625905752182007,
    "learning_rate": 0.0002793253775299492,
    "timestep_mean": 477.0,
    "epsilon_norm": 5.0762715339660645,
    "prediction_norm": 3.341397523880005,
    "timestamp": "2025-07-14T12:35:11.303654"
  },
  {
    "epoch": 2710,
    "batch_idx": 0,
    "loss": 0.18219667673110962,
    "learning_rate": 0.00028003041507204294,
    "timestep_mean": 486.1875,
    "epsilon_norm": 6.849001884460449,
    "prediction_norm": 6.053954124450684,
    "timestamp": "2025-07-14T12:35:11.308917"
  },
  {
    "epoch": 2710,
    "batch_idx": 32,
    "loss": 0.224941223859787,
    "learning_rate": 0.00028003041507204294,
    "timestep_mean": 584.375,
    "epsilon_norm": 7.733216762542725,
    "prediction_norm": 6.665583610534668,
    "timestamp": "2025-07-14T12:35:11.313841"
  },
  {
    "epoch": 2710,
    "batch_idx": 64,
    "loss": 0.3070235252380371,
    "learning_rate": 0.00028003041507204294,
    "timestep_mean": 413.9375,
    "epsilon_norm": 7.500216960906982,
    "prediction_norm": 6.167961120605469,
    "timestamp": "2025-07-14T12:35:11.318696"
  },
  {
    "epoch": 2710,
    "batch_idx": 96,
    "loss": 0.23865960538387299,
    "learning_rate": 0.00028003041507204294,
    "timestep_mean": 506.03125,
    "epsilon_norm": 8.10956859588623,
    "prediction_norm": 6.686968803405762,
    "timestamp": "2025-07-14T12:35:11.323519"
  },
  {
    "epoch": 2710,
    "batch_idx": 128,
    "loss": 0.21460184454917908,
    "learning_rate": 0.00028003041507204294,
    "timestep_mean": 441.875,
    "epsilon_norm": 7.436044216156006,
    "prediction_norm": 6.4202046394348145,
    "timestamp": "2025-07-14T12:35:11.328372"
  },
  {
    "epoch": 2710,
    "batch_idx": 160,
    "loss": 0.37738344073295593,
    "learning_rate": 0.00028003041507204294,
    "timestep_mean": 461.875,
    "epsilon_norm": 7.815643310546875,
    "prediction_norm": 6.1793293952941895,
    "timestamp": "2025-07-14T12:35:11.333323"
  },
  {
    "epoch": 2710,
    "batch_idx": 192,
    "loss": 0.0406993106007576,
    "learning_rate": 0.00028003041507204294,
    "timestep_mean": 697.625,
    "epsilon_norm": 2.7651093006134033,
    "prediction_norm": 2.447859048843384,
    "timestamp": "2025-07-14T12:35:11.338309"
  },
  {
    "epoch": 2711,
    "batch_idx": 0,
    "loss": 0.5782768726348877,
    "learning_rate": 0.0002807359953672214,
    "timestep_mean": 449.5625,
    "epsilon_norm": 9.029013633728027,
    "prediction_norm": 6.956764221191406,
    "timestamp": "2025-07-14T12:35:11.343191"
  },
  {
    "epoch": 2711,
    "batch_idx": 32,
    "loss": 0.3506821095943451,
    "learning_rate": 0.0002807359953672214,
    "timestep_mean": 520.3125,
    "epsilon_norm": 8.409891128540039,
    "prediction_norm": 7.429883003234863,
    "timestamp": "2025-07-14T12:35:11.348036"
  },
  {
    "epoch": 2711,
    "batch_idx": 64,
    "loss": 0.2896241247653961,
    "learning_rate": 0.0002807359953672214,
    "timestep_mean": 503.625,
    "epsilon_norm": 8.267094612121582,
    "prediction_norm": 6.987692832946777,
    "timestamp": "2025-07-14T12:35:11.353039"
  },
  {
    "epoch": 2711,
    "batch_idx": 96,
    "loss": 0.23238220810890198,
    "learning_rate": 0.0002807359953672214,
    "timestep_mean": 524.625,
    "epsilon_norm": 8.03877067565918,
    "prediction_norm": 7.221758842468262,
    "timestamp": "2025-07-14T12:35:11.358387"
  },
  {
    "epoch": 2711,
    "batch_idx": 128,
    "loss": 0.41810333728790283,
    "learning_rate": 0.0002807359953672214,
    "timestep_mean": 471.90625,
    "epsilon_norm": 7.904587268829346,
    "prediction_norm": 6.024886131286621,
    "timestamp": "2025-07-14T12:35:11.363460"
  },
  {
    "epoch": 2711,
    "batch_idx": 160,
    "loss": 0.3585299253463745,
    "learning_rate": 0.0002807359953672214,
    "timestep_mean": 464.3125,
    "epsilon_norm": 9.045141220092773,
    "prediction_norm": 7.149115562438965,
    "timestamp": "2025-07-14T12:35:11.368317"
  },
  {
    "epoch": 2711,
    "batch_idx": 192,
    "loss": 0.12799882888793945,
    "learning_rate": 0.0002807359953672214,
    "timestep_mean": 566.25,
    "epsilon_norm": 4.323115348815918,
    "prediction_norm": 3.4763259887695312,
    "timestamp": "2025-07-14T12:35:11.373356"
  },
  {
    "epoch": 2712,
    "batch_idx": 0,
    "loss": 0.21908482909202576,
    "learning_rate": 0.000281442116674534,
    "timestep_mean": 570.34375,
    "epsilon_norm": 7.175106048583984,
    "prediction_norm": 6.323914051055908,
    "timestamp": "2025-07-14T12:35:11.378273"
  },
  {
    "epoch": 2712,
    "batch_idx": 32,
    "loss": 0.2938743233680725,
    "learning_rate": 0.000281442116674534,
    "timestep_mean": 471.96875,
    "epsilon_norm": 7.351150989532471,
    "prediction_norm": 5.859546661376953,
    "timestamp": "2025-07-14T12:35:11.383315"
  },
  {
    "epoch": 2712,
    "batch_idx": 64,
    "loss": 0.17143863439559937,
    "learning_rate": 0.000281442116674534,
    "timestep_mean": 500.09375,
    "epsilon_norm": 7.314036846160889,
    "prediction_norm": 6.646258354187012,
    "timestamp": "2025-07-14T12:35:11.388354"
  },
  {
    "epoch": 2712,
    "batch_idx": 96,
    "loss": 0.19498443603515625,
    "learning_rate": 0.000281442116674534,
    "timestep_mean": 517.75,
    "epsilon_norm": 7.453150272369385,
    "prediction_norm": 7.114269733428955,
    "timestamp": "2025-07-14T12:35:11.393258"
  },
  {
    "epoch": 2712,
    "batch_idx": 128,
    "loss": 0.269891619682312,
    "learning_rate": 0.000281442116674534,
    "timestep_mean": 493.125,
    "epsilon_norm": 7.831651210784912,
    "prediction_norm": 6.936928749084473,
    "timestamp": "2025-07-14T12:35:11.398175"
  },
  {
    "epoch": 2712,
    "batch_idx": 160,
    "loss": 0.22544333338737488,
    "learning_rate": 0.000281442116674534,
    "timestep_mean": 499.5625,
    "epsilon_norm": 8.385329246520996,
    "prediction_norm": 7.30985164642334,
    "timestamp": "2025-07-14T12:35:11.403266"
  },
  {
    "epoch": 2712,
    "batch_idx": 192,
    "loss": 0.1266089826822281,
    "learning_rate": 0.000281442116674534,
    "timestep_mean": 542.75,
    "epsilon_norm": 3.3055317401885986,
    "prediction_norm": 3.2011094093322754,
    "timestamp": "2025-07-14T12:35:11.408157"
  },
  {
    "epoch": 2713,
    "batch_idx": 0,
    "loss": 0.4152764678001404,
    "learning_rate": 0.0002821487772516979,
    "timestep_mean": 417.6875,
    "epsilon_norm": 7.7047247886657715,
    "prediction_norm": 7.117833137512207,
    "timestamp": "2025-07-14T12:35:11.413093"
  },
  {
    "epoch": 2713,
    "batch_idx": 32,
    "loss": 0.28999027609825134,
    "learning_rate": 0.0002821487772516979,
    "timestep_mean": 402.25,
    "epsilon_norm": 7.553164482116699,
    "prediction_norm": 6.758502006530762,
    "timestamp": "2025-07-14T12:35:11.418147"
  },
  {
    "epoch": 2713,
    "batch_idx": 64,
    "loss": 0.16940705478191376,
    "learning_rate": 0.0002821487772516979,
    "timestep_mean": 561.15625,
    "epsilon_norm": 8.708831787109375,
    "prediction_norm": 8.408267974853516,
    "timestamp": "2025-07-14T12:35:11.423137"
  },
  {
    "epoch": 2713,
    "batch_idx": 96,
    "loss": 0.37332192063331604,
    "learning_rate": 0.0002821487772516979,
    "timestep_mean": 548.125,
    "epsilon_norm": 8.664525032043457,
    "prediction_norm": 7.258508205413818,
    "timestamp": "2025-07-14T12:35:11.428169"
  },
  {
    "epoch": 2713,
    "batch_idx": 128,
    "loss": 0.16372235119342804,
    "learning_rate": 0.0002821487772516979,
    "timestep_mean": 523.4375,
    "epsilon_norm": 7.520029544830322,
    "prediction_norm": 7.16849946975708,
    "timestamp": "2025-07-14T12:35:11.433206"
  },
  {
    "epoch": 2713,
    "batch_idx": 160,
    "loss": 0.20718824863433838,
    "learning_rate": 0.0002821487772516979,
    "timestep_mean": 520.9375,
    "epsilon_norm": 7.371255397796631,
    "prediction_norm": 6.682831764221191,
    "timestamp": "2025-07-14T12:35:11.438454"
  },
  {
    "epoch": 2713,
    "batch_idx": 192,
    "loss": 0.08514411747455597,
    "learning_rate": 0.0002821487772516979,
    "timestep_mean": 497.0,
    "epsilon_norm": 3.144598960876465,
    "prediction_norm": 3.1404476165771484,
    "timestamp": "2025-07-14T12:35:11.443631"
  },
  {
    "epoch": 2714,
    "batch_idx": 0,
    "loss": 0.456402450799942,
    "learning_rate": 0.00028285597535509843,
    "timestep_mean": 418.625,
    "epsilon_norm": 8.319523811340332,
    "prediction_norm": 6.534943580627441,
    "timestamp": "2025-07-14T12:35:11.448822"
  },
  {
    "epoch": 2714,
    "batch_idx": 32,
    "loss": 0.2588707208633423,
    "learning_rate": 0.00028285597535509843,
    "timestep_mean": 524.90625,
    "epsilon_norm": 8.240996360778809,
    "prediction_norm": 6.980052947998047,
    "timestamp": "2025-07-14T12:35:11.454072"
  },
  {
    "epoch": 2714,
    "batch_idx": 64,
    "loss": 0.36732903122901917,
    "learning_rate": 0.00028285597535509843,
    "timestep_mean": 424.90625,
    "epsilon_norm": 8.338875770568848,
    "prediction_norm": 6.595972537994385,
    "timestamp": "2025-07-14T12:35:11.460302"
  },
  {
    "epoch": 2714,
    "batch_idx": 96,
    "loss": 0.12890537083148956,
    "learning_rate": 0.00028285597535509843,
    "timestep_mean": 606.3125,
    "epsilon_norm": 8.612077713012695,
    "prediction_norm": 7.40012788772583,
    "timestamp": "2025-07-14T12:35:11.466288"
  },
  {
    "epoch": 2714,
    "batch_idx": 128,
    "loss": 0.1438293308019638,
    "learning_rate": 0.00028285597535509843,
    "timestep_mean": 602.34375,
    "epsilon_norm": 7.922463417053223,
    "prediction_norm": 7.420124530792236,
    "timestamp": "2025-07-14T12:35:11.471867"
  },
  {
    "epoch": 2714,
    "batch_idx": 160,
    "loss": 0.22772827744483948,
    "learning_rate": 0.00028285597535509843,
    "timestep_mean": 511.5,
    "epsilon_norm": 8.545426368713379,
    "prediction_norm": 7.527981758117676,
    "timestamp": "2025-07-14T12:35:11.477414"
  },
  {
    "epoch": 2714,
    "batch_idx": 192,
    "loss": 0.41160842776298523,
    "learning_rate": 0.00028285597535509843,
    "timestep_mean": 526.75,
    "epsilon_norm": 3.9972152709960938,
    "prediction_norm": 3.6032516956329346,
    "timestamp": "2025-07-14T12:35:11.483092"
  },
  {
    "epoch": 2715,
    "batch_idx": 0,
    "loss": 0.3869895339012146,
    "learning_rate": 0.00028356370923979376,
    "timestep_mean": 513.375,
    "epsilon_norm": 7.818859577178955,
    "prediction_norm": 5.596786975860596,
    "timestamp": "2025-07-14T12:35:11.489003"
  },
  {
    "epoch": 2715,
    "batch_idx": 32,
    "loss": 0.21777194738388062,
    "learning_rate": 0.00028356370923979376,
    "timestep_mean": 564.875,
    "epsilon_norm": 8.362278938293457,
    "prediction_norm": 7.19069766998291,
    "timestamp": "2025-07-14T12:35:11.494806"
  },
  {
    "epoch": 2715,
    "batch_idx": 64,
    "loss": 0.24345287680625916,
    "learning_rate": 0.00028356370923979376,
    "timestep_mean": 516.375,
    "epsilon_norm": 8.84145450592041,
    "prediction_norm": 7.100184440612793,
    "timestamp": "2025-07-14T12:35:11.500147"
  },
  {
    "epoch": 2715,
    "batch_idx": 96,
    "loss": 0.4999408423900604,
    "learning_rate": 0.00028356370923979376,
    "timestep_mean": 454.9375,
    "epsilon_norm": 9.251930236816406,
    "prediction_norm": 7.222828388214111,
    "timestamp": "2025-07-14T12:35:11.505410"
  },
  {
    "epoch": 2715,
    "batch_idx": 128,
    "loss": 0.30262845754623413,
    "learning_rate": 0.00028356370923979376,
    "timestep_mean": 544.5625,
    "epsilon_norm": 8.053057670593262,
    "prediction_norm": 6.364201545715332,
    "timestamp": "2025-07-14T12:35:11.510673"
  },
  {
    "epoch": 2715,
    "batch_idx": 160,
    "loss": 0.16570329666137695,
    "learning_rate": 0.00028356370923979376,
    "timestep_mean": 562.53125,
    "epsilon_norm": 8.405550003051758,
    "prediction_norm": 7.920889377593994,
    "timestamp": "2025-07-14T12:35:11.515942"
  },
  {
    "epoch": 2715,
    "batch_idx": 192,
    "loss": 0.5538915991783142,
    "learning_rate": 0.00028356370923979376,
    "timestep_mean": 564.0,
    "epsilon_norm": 4.136748313903809,
    "prediction_norm": 2.7133259773254395,
    "timestamp": "2025-07-14T12:35:11.521248"
  },
  {
    "epoch": 2716,
    "batch_idx": 0,
    "loss": 0.2243383824825287,
    "learning_rate": 0.00028427197715952116,
    "timestep_mean": 540.125,
    "epsilon_norm": 7.7597832679748535,
    "prediction_norm": 6.796736240386963,
    "timestamp": "2025-07-14T12:35:11.526474"
  },
  {
    "epoch": 2716,
    "batch_idx": 32,
    "loss": 0.2461353838443756,
    "learning_rate": 0.00028427197715952116,
    "timestep_mean": 477.46875,
    "epsilon_norm": 7.540910243988037,
    "prediction_norm": 6.346969127655029,
    "timestamp": "2025-07-14T12:35:11.531781"
  },
  {
    "epoch": 2716,
    "batch_idx": 64,
    "loss": 0.32275086641311646,
    "learning_rate": 0.00028427197715952116,
    "timestep_mean": 473.4375,
    "epsilon_norm": 8.861319541931152,
    "prediction_norm": 7.288015365600586,
    "timestamp": "2025-07-14T12:35:11.537295"
  },
  {
    "epoch": 2716,
    "batch_idx": 96,
    "loss": 0.3777202367782593,
    "learning_rate": 0.00028427197715952116,
    "timestep_mean": 435.03125,
    "epsilon_norm": 9.656420707702637,
    "prediction_norm": 7.727221488952637,
    "timestamp": "2025-07-14T12:35:11.542735"
  },
  {
    "epoch": 2716,
    "batch_idx": 128,
    "loss": 0.2567696273326874,
    "learning_rate": 0.00028427197715952116,
    "timestep_mean": 429.6875,
    "epsilon_norm": 7.609942436218262,
    "prediction_norm": 6.881466865539551,
    "timestamp": "2025-07-14T12:35:11.548028"
  },
  {
    "epoch": 2716,
    "batch_idx": 160,
    "loss": 0.5673664808273315,
    "learning_rate": 0.00028427197715952116,
    "timestep_mean": 469.25,
    "epsilon_norm": 8.646637916564941,
    "prediction_norm": 6.458019733428955,
    "timestamp": "2025-07-14T12:35:11.553327"
  },
  {
    "epoch": 2716,
    "batch_idx": 192,
    "loss": 0.3760773241519928,
    "learning_rate": 0.00028427197715952116,
    "timestep_mean": 357.0,
    "epsilon_norm": 3.052072286605835,
    "prediction_norm": 2.1499886512756348,
    "timestamp": "2025-07-14T12:35:11.558710"
  },
  {
    "epoch": 2717,
    "batch_idx": 0,
    "loss": 0.494243323802948,
    "learning_rate": 0.00028498077736669966,
    "timestep_mean": 395.8125,
    "epsilon_norm": 8.423087120056152,
    "prediction_norm": 6.408344268798828,
    "timestamp": "2025-07-14T12:35:11.564039"
  },
  {
    "epoch": 2717,
    "batch_idx": 32,
    "loss": 0.19506880640983582,
    "learning_rate": 0.00028498077736669966,
    "timestep_mean": 538.53125,
    "epsilon_norm": 7.590133190155029,
    "prediction_norm": 7.118350028991699,
    "timestamp": "2025-07-14T12:35:11.569340"
  },
  {
    "epoch": 2717,
    "batch_idx": 64,
    "loss": 0.40903255343437195,
    "learning_rate": 0.00028498077736669966,
    "timestep_mean": 512.4375,
    "epsilon_norm": 8.567187309265137,
    "prediction_norm": 7.248913764953613,
    "timestamp": "2025-07-14T12:35:11.574563"
  },
  {
    "epoch": 2717,
    "batch_idx": 96,
    "loss": 0.18310540914535522,
    "learning_rate": 0.00028498077736669966,
    "timestep_mean": 579.0625,
    "epsilon_norm": 8.363579750061035,
    "prediction_norm": 7.4267144203186035,
    "timestamp": "2025-07-14T12:35:11.579644"
  },
  {
    "epoch": 2717,
    "batch_idx": 128,
    "loss": 0.12152178585529327,
    "learning_rate": 0.00028498077736669966,
    "timestep_mean": 530.4375,
    "epsilon_norm": 7.809650421142578,
    "prediction_norm": 6.846477508544922,
    "timestamp": "2025-07-14T12:35:11.584952"
  },
  {
    "epoch": 2717,
    "batch_idx": 160,
    "loss": 0.3615643382072449,
    "learning_rate": 0.00028498077736669966,
    "timestep_mean": 460.96875,
    "epsilon_norm": 7.493903636932373,
    "prediction_norm": 6.235385417938232,
    "timestamp": "2025-07-14T12:35:11.590181"
  },
  {
    "epoch": 2717,
    "batch_idx": 192,
    "loss": 0.07331356406211853,
    "learning_rate": 0.00028498077736669966,
    "timestep_mean": 577.5,
    "epsilon_norm": 3.4514670372009277,
    "prediction_norm": 3.445918083190918,
    "timestamp": "2025-07-14T12:35:11.595313"
  },
  {
    "epoch": 2718,
    "batch_idx": 0,
    "loss": 0.3807435631752014,
    "learning_rate": 0.00028569010811243644,
    "timestep_mean": 463.90625,
    "epsilon_norm": 7.387201309204102,
    "prediction_norm": 5.4619317054748535,
    "timestamp": "2025-07-14T12:35:11.600839"
  },
  {
    "epoch": 2718,
    "batch_idx": 32,
    "loss": 0.11226950585842133,
    "learning_rate": 0.00028569010811243644,
    "timestep_mean": 567.5625,
    "epsilon_norm": 7.177846431732178,
    "prediction_norm": 6.796790599822998,
    "timestamp": "2025-07-14T12:35:11.605993"
  },
  {
    "epoch": 2718,
    "batch_idx": 64,
    "loss": 0.19790905714035034,
    "learning_rate": 0.00028569010811243644,
    "timestep_mean": 535.75,
    "epsilon_norm": 7.499131202697754,
    "prediction_norm": 6.4729743003845215,
    "timestamp": "2025-07-14T12:35:11.611275"
  },
  {
    "epoch": 2718,
    "batch_idx": 96,
    "loss": 0.41068342328071594,
    "learning_rate": 0.00028569010811243644,
    "timestep_mean": 329.0,
    "epsilon_norm": 8.837507247924805,
    "prediction_norm": 7.055453300476074,
    "timestamp": "2025-07-14T12:35:11.616499"
  },
  {
    "epoch": 2718,
    "batch_idx": 128,
    "loss": 0.1252114474773407,
    "learning_rate": 0.00028569010811243644,
    "timestep_mean": 538.6875,
    "epsilon_norm": 7.209668159484863,
    "prediction_norm": 7.377226829528809,
    "timestamp": "2025-07-14T12:35:11.621972"
  },
  {
    "epoch": 2718,
    "batch_idx": 160,
    "loss": 0.20464293658733368,
    "learning_rate": 0.00028569010811243644,
    "timestep_mean": 598.9375,
    "epsilon_norm": 7.727376937866211,
    "prediction_norm": 7.283235549926758,
    "timestamp": "2025-07-14T12:35:11.627177"
  },
  {
    "epoch": 2718,
    "batch_idx": 192,
    "loss": 0.1769772171974182,
    "learning_rate": 0.00028569010811243644,
    "timestep_mean": 463.875,
    "epsilon_norm": 3.7695672512054443,
    "prediction_norm": 4.509830951690674,
    "timestamp": "2025-07-14T12:35:11.632467"
  },
  {
    "epoch": 2719,
    "batch_idx": 0,
    "loss": 0.1236172616481781,
    "learning_rate": 0.00028639996764652704,
    "timestep_mean": 511.90625,
    "epsilon_norm": 8.29455280303955,
    "prediction_norm": 7.326960563659668,
    "timestamp": "2025-07-14T12:35:11.637796"
  },
  {
    "epoch": 2719,
    "batch_idx": 32,
    "loss": 0.16388744115829468,
    "learning_rate": 0.00028639996764652704,
    "timestep_mean": 449.09375,
    "epsilon_norm": 8.176931381225586,
    "prediction_norm": 7.5124287605285645,
    "timestamp": "2025-07-14T12:35:11.643378"
  },
  {
    "epoch": 2719,
    "batch_idx": 64,
    "loss": 0.2549395263195038,
    "learning_rate": 0.00028639996764652704,
    "timestep_mean": 604.125,
    "epsilon_norm": 8.334054946899414,
    "prediction_norm": 7.392454624176025,
    "timestamp": "2025-07-14T12:35:11.648926"
  },
  {
    "epoch": 2719,
    "batch_idx": 96,
    "loss": 0.24984797835350037,
    "learning_rate": 0.00028639996764652704,
    "timestep_mean": 523.90625,
    "epsilon_norm": 7.470210552215576,
    "prediction_norm": 6.127495765686035,
    "timestamp": "2025-07-14T12:35:11.654208"
  },
  {
    "epoch": 2719,
    "batch_idx": 128,
    "loss": 0.3835567235946655,
    "learning_rate": 0.00028639996764652704,
    "timestep_mean": 468.59375,
    "epsilon_norm": 7.4286980628967285,
    "prediction_norm": 6.6835737228393555,
    "timestamp": "2025-07-14T12:35:11.659392"
  },
  {
    "epoch": 2719,
    "batch_idx": 160,
    "loss": 0.38505151867866516,
    "learning_rate": 0.00028639996764652704,
    "timestep_mean": 480.375,
    "epsilon_norm": 8.002781867980957,
    "prediction_norm": 6.628676414489746,
    "timestamp": "2025-07-14T12:35:11.664549"
  },
  {
    "epoch": 2719,
    "batch_idx": 192,
    "loss": 0.26488780975341797,
    "learning_rate": 0.00028639996764652704,
    "timestep_mean": 664.0,
    "epsilon_norm": 4.565814971923828,
    "prediction_norm": 4.200465202331543,
    "timestamp": "2025-07-14T12:35:11.669839"
  },
  {
    "epoch": 2720,
    "batch_idx": 0,
    "loss": 0.3377878963947296,
    "learning_rate": 0.00028711035421746415,
    "timestep_mean": 468.21875,
    "epsilon_norm": 8.138053894042969,
    "prediction_norm": 6.391225337982178,
    "timestamp": "2025-07-14T12:35:11.675595"
  },
  {
    "epoch": 2720,
    "batch_idx": 32,
    "loss": 0.20416045188903809,
    "learning_rate": 0.00028711035421746415,
    "timestep_mean": 444.46875,
    "epsilon_norm": 7.734118938446045,
    "prediction_norm": 6.0533127784729,
    "timestamp": "2025-07-14T12:35:11.680787"
  },
  {
    "epoch": 2720,
    "batch_idx": 64,
    "loss": 0.23058006167411804,
    "learning_rate": 0.00028711035421746415,
    "timestep_mean": 554.65625,
    "epsilon_norm": 7.791714191436768,
    "prediction_norm": 6.901988506317139,
    "timestamp": "2025-07-14T12:35:11.686009"
  },
  {
    "epoch": 2720,
    "batch_idx": 96,
    "loss": 0.35872703790664673,
    "learning_rate": 0.00028711035421746415,
    "timestep_mean": 527.96875,
    "epsilon_norm": 8.39980697631836,
    "prediction_norm": 7.078207492828369,
    "timestamp": "2025-07-14T12:35:11.691261"
  },
  {
    "epoch": 2720,
    "batch_idx": 128,
    "loss": 0.2523813843727112,
    "learning_rate": 0.00028711035421746415,
    "timestep_mean": 525.375,
    "epsilon_norm": 9.444893836975098,
    "prediction_norm": 8.057857513427734,
    "timestamp": "2025-07-14T12:35:11.696642"
  },
  {
    "epoch": 2720,
    "batch_idx": 160,
    "loss": 0.13032710552215576,
    "learning_rate": 0.00028711035421746415,
    "timestep_mean": 432.34375,
    "epsilon_norm": 8.189094543457031,
    "prediction_norm": 7.251801490783691,
    "timestamp": "2025-07-14T12:35:11.701825"
  },
  {
    "epoch": 2720,
    "batch_idx": 192,
    "loss": 0.5289075374603271,
    "learning_rate": 0.00028711035421746415,
    "timestep_mean": 464.125,
    "epsilon_norm": 3.9645769596099854,
    "prediction_norm": 2.9064626693725586,
    "timestamp": "2025-07-14T12:35:11.707224"
  },
  {
    "epoch": 2721,
    "batch_idx": 0,
    "loss": 0.1688583791255951,
    "learning_rate": 0.00028782126607243997,
    "timestep_mean": 518.0625,
    "epsilon_norm": 6.9186625480651855,
    "prediction_norm": 5.82302713394165,
    "timestamp": "2025-07-14T12:35:11.712482"
  },
  {
    "epoch": 2721,
    "batch_idx": 32,
    "loss": 0.2521098256111145,
    "learning_rate": 0.00028782126607243997,
    "timestep_mean": 484.1875,
    "epsilon_norm": 7.729548454284668,
    "prediction_norm": 6.680340766906738,
    "timestamp": "2025-07-14T12:35:11.717929"
  },
  {
    "epoch": 2721,
    "batch_idx": 64,
    "loss": 0.3137446343898773,
    "learning_rate": 0.00028782126607243997,
    "timestep_mean": 517.53125,
    "epsilon_norm": 7.890268802642822,
    "prediction_norm": 6.703780651092529,
    "timestamp": "2025-07-14T12:35:11.723299"
  },
  {
    "epoch": 2721,
    "batch_idx": 96,
    "loss": 0.1507568061351776,
    "learning_rate": 0.00028782126607243997,
    "timestep_mean": 622.75,
    "epsilon_norm": 8.21989631652832,
    "prediction_norm": 7.538218021392822,
    "timestamp": "2025-07-14T12:35:11.728517"
  },
  {
    "epoch": 2721,
    "batch_idx": 128,
    "loss": 0.22963550686836243,
    "learning_rate": 0.00028782126607243997,
    "timestep_mean": 522.78125,
    "epsilon_norm": 8.036293029785156,
    "prediction_norm": 7.004150390625,
    "timestamp": "2025-07-14T12:35:11.733950"
  },
  {
    "epoch": 2721,
    "batch_idx": 160,
    "loss": 0.26646944880485535,
    "learning_rate": 0.00028782126607243997,
    "timestep_mean": 424.9375,
    "epsilon_norm": 7.68265438079834,
    "prediction_norm": 6.125759124755859,
    "timestamp": "2025-07-14T12:35:11.739133"
  },
  {
    "epoch": 2721,
    "batch_idx": 192,
    "loss": 0.2699395418167114,
    "learning_rate": 0.00028782126607243997,
    "timestep_mean": 474.375,
    "epsilon_norm": 4.162384033203125,
    "prediction_norm": 3.3880980014801025,
    "timestamp": "2025-07-14T12:35:11.744468"
  },
  {
    "epoch": 2722,
    "batch_idx": 0,
    "loss": 0.2886505722999573,
    "learning_rate": 0.0002885327014573492,
    "timestep_mean": 464.0625,
    "epsilon_norm": 7.317790508270264,
    "prediction_norm": 6.07647180557251,
    "timestamp": "2025-07-14T12:35:11.749741"
  },
  {
    "epoch": 2722,
    "batch_idx": 32,
    "loss": 0.44937607645988464,
    "learning_rate": 0.0002885327014573492,
    "timestep_mean": 508.03125,
    "epsilon_norm": 9.098472595214844,
    "prediction_norm": 7.217231273651123,
    "timestamp": "2025-07-14T12:35:11.755134"
  },
  {
    "epoch": 2722,
    "batch_idx": 64,
    "loss": 0.39476680755615234,
    "learning_rate": 0.0002885327014573492,
    "timestep_mean": 422.0,
    "epsilon_norm": 7.7617034912109375,
    "prediction_norm": 5.7745842933654785,
    "timestamp": "2025-07-14T12:35:11.760426"
  },
  {
    "epoch": 2722,
    "batch_idx": 96,
    "loss": 0.30880656838417053,
    "learning_rate": 0.0002885327014573492,
    "timestep_mean": 451.21875,
    "epsilon_norm": 8.707998275756836,
    "prediction_norm": 7.65256929397583,
    "timestamp": "2025-07-14T12:35:11.765865"
  },
  {
    "epoch": 2722,
    "batch_idx": 128,
    "loss": 0.23905329406261444,
    "learning_rate": 0.0002885327014573492,
    "timestep_mean": 504.78125,
    "epsilon_norm": 7.8885931968688965,
    "prediction_norm": 6.832342624664307,
    "timestamp": "2025-07-14T12:35:11.771159"
  },
  {
    "epoch": 2722,
    "batch_idx": 160,
    "loss": 0.2813638746738434,
    "learning_rate": 0.0002885327014573492,
    "timestep_mean": 530.875,
    "epsilon_norm": 8.029061317443848,
    "prediction_norm": 7.031426429748535,
    "timestamp": "2025-07-14T12:35:11.776400"
  },
  {
    "epoch": 2722,
    "batch_idx": 192,
    "loss": 0.2884240448474884,
    "learning_rate": 0.0002885327014573492,
    "timestep_mean": 451.75,
    "epsilon_norm": 2.8854331970214844,
    "prediction_norm": 2.3855669498443604,
    "timestamp": "2025-07-14T12:35:11.781665"
  },
  {
    "epoch": 2723,
    "batch_idx": 0,
    "loss": 0.293866902589798,
    "learning_rate": 0.0002892446586167958,
    "timestep_mean": 502.5,
    "epsilon_norm": 7.956780910491943,
    "prediction_norm": 6.5211100578308105,
    "timestamp": "2025-07-14T12:35:11.786892"
  },
  {
    "epoch": 2723,
    "batch_idx": 32,
    "loss": 0.13922575116157532,
    "learning_rate": 0.0002892446586167958,
    "timestep_mean": 605.65625,
    "epsilon_norm": 8.350715637207031,
    "prediction_norm": 7.741962909698486,
    "timestamp": "2025-07-14T12:35:11.792301"
  },
  {
    "epoch": 2723,
    "batch_idx": 64,
    "loss": 0.2899615168571472,
    "learning_rate": 0.0002892446586167958,
    "timestep_mean": 485.15625,
    "epsilon_norm": 7.728533744812012,
    "prediction_norm": 6.716732978820801,
    "timestamp": "2025-07-14T12:35:11.797565"
  },
  {
    "epoch": 2723,
    "batch_idx": 96,
    "loss": 0.2912867069244385,
    "learning_rate": 0.0002892446586167958,
    "timestep_mean": 507.25,
    "epsilon_norm": 8.14166259765625,
    "prediction_norm": 6.288894176483154,
    "timestamp": "2025-07-14T12:35:11.802806"
  },
  {
    "epoch": 2723,
    "batch_idx": 128,
    "loss": 0.21204006671905518,
    "learning_rate": 0.0002892446586167958,
    "timestep_mean": 450.875,
    "epsilon_norm": 6.753207206726074,
    "prediction_norm": 6.249102592468262,
    "timestamp": "2025-07-14T12:35:11.807981"
  },
  {
    "epoch": 2723,
    "batch_idx": 160,
    "loss": 0.35062235593795776,
    "learning_rate": 0.0002892446586167958,
    "timestep_mean": 478.53125,
    "epsilon_norm": 7.500758647918701,
    "prediction_norm": 6.776313304901123,
    "timestamp": "2025-07-14T12:35:11.813399"
  },
  {
    "epoch": 2723,
    "batch_idx": 192,
    "loss": 0.5090321898460388,
    "learning_rate": 0.0002892446586167958,
    "timestep_mean": 442.0,
    "epsilon_norm": 4.258602619171143,
    "prediction_norm": 3.3781864643096924,
    "timestamp": "2025-07-14T12:35:11.818746"
  },
  {
    "epoch": 2724,
    "batch_idx": 0,
    "loss": 0.1781998723745346,
    "learning_rate": 0.0002899571357940972,
    "timestep_mean": 585.9375,
    "epsilon_norm": 7.867448806762695,
    "prediction_norm": 7.56074333190918,
    "timestamp": "2025-07-14T12:35:11.824034"
  },
  {
    "epoch": 2724,
    "batch_idx": 32,
    "loss": 0.21262037754058838,
    "learning_rate": 0.0002899571357940972,
    "timestep_mean": 460.90625,
    "epsilon_norm": 7.344151020050049,
    "prediction_norm": 6.591265678405762,
    "timestamp": "2025-07-14T12:35:11.829457"
  },
  {
    "epoch": 2724,
    "batch_idx": 64,
    "loss": 0.17074790596961975,
    "learning_rate": 0.0002899571357940972,
    "timestep_mean": 513.28125,
    "epsilon_norm": 8.08007526397705,
    "prediction_norm": 6.959892749786377,
    "timestamp": "2025-07-14T12:35:11.834809"
  },
  {
    "epoch": 2724,
    "batch_idx": 96,
    "loss": 0.24346382915973663,
    "learning_rate": 0.0002899571357940972,
    "timestep_mean": 492.34375,
    "epsilon_norm": 7.711663722991943,
    "prediction_norm": 6.571106433868408,
    "timestamp": "2025-07-14T12:35:11.840016"
  },
  {
    "epoch": 2724,
    "batch_idx": 128,
    "loss": 0.1986321359872818,
    "learning_rate": 0.0002899571357940972,
    "timestep_mean": 626.3125,
    "epsilon_norm": 7.7274088859558105,
    "prediction_norm": 6.700252056121826,
    "timestamp": "2025-07-14T12:35:11.844996"
  },
  {
    "epoch": 2724,
    "batch_idx": 160,
    "loss": 0.39689671993255615,
    "learning_rate": 0.0002899571357940972,
    "timestep_mean": 460.375,
    "epsilon_norm": 8.456070899963379,
    "prediction_norm": 7.229078769683838,
    "timestamp": "2025-07-14T12:35:11.850005"
  },
  {
    "epoch": 2724,
    "batch_idx": 192,
    "loss": 0.11874957382678986,
    "learning_rate": 0.0002899571357940972,
    "timestep_mean": 544.25,
    "epsilon_norm": 3.42502498626709,
    "prediction_norm": 2.973681926727295,
    "timestamp": "2025-07-14T12:35:11.854975"
  },
  {
    "epoch": 2725,
    "batch_idx": 0,
    "loss": 0.2905694246292114,
    "learning_rate": 0.0002906701312312866,
    "timestep_mean": 584.96875,
    "epsilon_norm": 7.946549892425537,
    "prediction_norm": 6.771467208862305,
    "timestamp": "2025-07-14T12:35:11.859854"
  },
  {
    "epoch": 2725,
    "batch_idx": 32,
    "loss": 0.34975817799568176,
    "learning_rate": 0.0002906701312312866,
    "timestep_mean": 367.59375,
    "epsilon_norm": 8.383858680725098,
    "prediction_norm": 6.496343612670898,
    "timestamp": "2025-07-14T12:35:11.864768"
  },
  {
    "epoch": 2725,
    "batch_idx": 64,
    "loss": 0.11884866654872894,
    "learning_rate": 0.0002906701312312866,
    "timestep_mean": 522.65625,
    "epsilon_norm": 8.158442497253418,
    "prediction_norm": 7.2631707191467285,
    "timestamp": "2025-07-14T12:35:11.869841"
  },
  {
    "epoch": 2725,
    "batch_idx": 96,
    "loss": 0.19366110861301422,
    "learning_rate": 0.0002906701312312866,
    "timestep_mean": 468.71875,
    "epsilon_norm": 7.938950061798096,
    "prediction_norm": 6.660523414611816,
    "timestamp": "2025-07-14T12:35:11.875029"
  },
  {
    "epoch": 2725,
    "batch_idx": 128,
    "loss": 0.22341439127922058,
    "learning_rate": 0.0002906701312312866,
    "timestep_mean": 481.71875,
    "epsilon_norm": 7.6040825843811035,
    "prediction_norm": 6.804411888122559,
    "timestamp": "2025-07-14T12:35:11.880335"
  },
  {
    "epoch": 2725,
    "batch_idx": 160,
    "loss": 0.2350074052810669,
    "learning_rate": 0.0002906701312312866,
    "timestep_mean": 539.1875,
    "epsilon_norm": 7.9099225997924805,
    "prediction_norm": 7.284181118011475,
    "timestamp": "2025-07-14T12:35:11.885538"
  },
  {
    "epoch": 2725,
    "batch_idx": 192,
    "loss": 0.06567974388599396,
    "learning_rate": 0.0002906701312312866,
    "timestep_mean": 548.625,
    "epsilon_norm": 4.179754257202148,
    "prediction_norm": 4.004157543182373,
    "timestamp": "2025-07-14T12:35:11.890787"
  },
  {
    "epoch": 2726,
    "batch_idx": 0,
    "loss": 0.17024825513362885,
    "learning_rate": 0.000291383643169118,
    "timestep_mean": 552.21875,
    "epsilon_norm": 7.295976638793945,
    "prediction_norm": 6.875308990478516,
    "timestamp": "2025-07-14T12:35:11.895909"
  },
  {
    "epoch": 2726,
    "batch_idx": 32,
    "loss": 0.10295816510915756,
    "learning_rate": 0.000291383643169118,
    "timestep_mean": 614.1875,
    "epsilon_norm": 7.799424171447754,
    "prediction_norm": 7.253322124481201,
    "timestamp": "2025-07-14T12:35:11.901239"
  },
  {
    "epoch": 2726,
    "batch_idx": 64,
    "loss": 0.45515838265419006,
    "learning_rate": 0.000291383643169118,
    "timestep_mean": 439.5,
    "epsilon_norm": 8.366921424865723,
    "prediction_norm": 6.783559322357178,
    "timestamp": "2025-07-14T12:35:11.906590"
  },
  {
    "epoch": 2726,
    "batch_idx": 96,
    "loss": 0.3365573585033417,
    "learning_rate": 0.000291383643169118,
    "timestep_mean": 459.15625,
    "epsilon_norm": 8.133856773376465,
    "prediction_norm": 6.976296424865723,
    "timestamp": "2025-07-14T12:35:11.911773"
  },
  {
    "epoch": 2726,
    "batch_idx": 128,
    "loss": 0.22789809107780457,
    "learning_rate": 0.000291383643169118,
    "timestep_mean": 505.78125,
    "epsilon_norm": 8.163863182067871,
    "prediction_norm": 6.866249084472656,
    "timestamp": "2025-07-14T12:35:11.917071"
  },
  {
    "epoch": 2726,
    "batch_idx": 160,
    "loss": 0.25897181034088135,
    "learning_rate": 0.000291383643169118,
    "timestep_mean": 502.9375,
    "epsilon_norm": 7.959042072296143,
    "prediction_norm": 6.720633506774902,
    "timestamp": "2025-07-14T12:35:11.922369"
  },
  {
    "epoch": 2726,
    "batch_idx": 192,
    "loss": 0.036323994398117065,
    "learning_rate": 0.000291383643169118,
    "timestep_mean": 768.75,
    "epsilon_norm": 4.051664352416992,
    "prediction_norm": 3.755953550338745,
    "timestamp": "2025-07-14T12:35:11.927640"
  },
  {
    "epoch": 2727,
    "batch_idx": 0,
    "loss": 0.3687651753425598,
    "learning_rate": 0.0002920976698470718,
    "timestep_mean": 392.46875,
    "epsilon_norm": 8.44378662109375,
    "prediction_norm": 6.600840091705322,
    "timestamp": "2025-07-14T12:35:11.932837"
  },
  {
    "epoch": 2727,
    "batch_idx": 32,
    "loss": 0.17939604818820953,
    "learning_rate": 0.0002920976698470718,
    "timestep_mean": 632.4375,
    "epsilon_norm": 8.061468124389648,
    "prediction_norm": 7.13669490814209,
    "timestamp": "2025-07-14T12:35:11.938085"
  },
  {
    "epoch": 2727,
    "batch_idx": 64,
    "loss": 0.4114612340927124,
    "learning_rate": 0.0002920976698470718,
    "timestep_mean": 438.96875,
    "epsilon_norm": 9.400784492492676,
    "prediction_norm": 7.501497268676758,
    "timestamp": "2025-07-14T12:35:11.943372"
  },
  {
    "epoch": 2727,
    "batch_idx": 96,
    "loss": 0.4537406265735626,
    "learning_rate": 0.0002920976698470718,
    "timestep_mean": 547.8125,
    "epsilon_norm": 10.002552032470703,
    "prediction_norm": 8.453648567199707,
    "timestamp": "2025-07-14T12:35:11.948673"
  },
  {
    "epoch": 2727,
    "batch_idx": 128,
    "loss": 0.2858741283416748,
    "learning_rate": 0.0002920976698470718,
    "timestep_mean": 544.5,
    "epsilon_norm": 7.664339065551758,
    "prediction_norm": 6.537908554077148,
    "timestamp": "2025-07-14T12:35:11.953897"
  },
  {
    "epoch": 2727,
    "batch_idx": 160,
    "loss": 0.2526792883872986,
    "learning_rate": 0.0002920976698470718,
    "timestep_mean": 460.84375,
    "epsilon_norm": 7.907192707061768,
    "prediction_norm": 6.790922164916992,
    "timestamp": "2025-07-14T12:35:11.959189"
  },
  {
    "epoch": 2727,
    "batch_idx": 192,
    "loss": 0.7159008383750916,
    "learning_rate": 0.0002920976698470718,
    "timestep_mean": 452.625,
    "epsilon_norm": 3.6504712104797363,
    "prediction_norm": 2.852196216583252,
    "timestamp": "2025-07-14T12:35:11.964357"
  },
  {
    "epoch": 2728,
    "batch_idx": 0,
    "loss": 0.23635590076446533,
    "learning_rate": 0.00029281220950335844,
    "timestep_mean": 479.125,
    "epsilon_norm": 7.193645477294922,
    "prediction_norm": 5.951791286468506,
    "timestamp": "2025-07-14T12:35:11.969597"
  },
  {
    "epoch": 2728,
    "batch_idx": 32,
    "loss": 0.4104077219963074,
    "learning_rate": 0.00029281220950335844,
    "timestep_mean": 476.8125,
    "epsilon_norm": 8.306208610534668,
    "prediction_norm": 6.528933048248291,
    "timestamp": "2025-07-14T12:35:11.974898"
  },
  {
    "epoch": 2728,
    "batch_idx": 64,
    "loss": 0.42751172184944153,
    "learning_rate": 0.00029281220950335844,
    "timestep_mean": 459.28125,
    "epsilon_norm": 8.413313865661621,
    "prediction_norm": 6.989593505859375,
    "timestamp": "2025-07-14T12:35:11.980325"
  },
  {
    "epoch": 2728,
    "batch_idx": 96,
    "loss": 0.277945339679718,
    "learning_rate": 0.00029281220950335844,
    "timestep_mean": 520.125,
    "epsilon_norm": 6.815000534057617,
    "prediction_norm": 5.498100757598877,
    "timestamp": "2025-07-14T12:35:11.985501"
  },
  {
    "epoch": 2728,
    "batch_idx": 128,
    "loss": 0.3435809016227722,
    "learning_rate": 0.00029281220950335844,
    "timestep_mean": 459.65625,
    "epsilon_norm": 8.714299201965332,
    "prediction_norm": 7.375535011291504,
    "timestamp": "2025-07-14T12:35:11.990744"
  },
  {
    "epoch": 2728,
    "batch_idx": 160,
    "loss": 0.35118797421455383,
    "learning_rate": 0.00029281220950335844,
    "timestep_mean": 490.65625,
    "epsilon_norm": 8.576698303222656,
    "prediction_norm": 7.396772861480713,
    "timestamp": "2025-07-14T12:35:11.995949"
  },
  {
    "epoch": 2728,
    "batch_idx": 192,
    "loss": 0.2708808183670044,
    "learning_rate": 0.00029281220950335844,
    "timestep_mean": 555.625,
    "epsilon_norm": 4.861920356750488,
    "prediction_norm": 3.746776819229126,
    "timestamp": "2025-07-14T12:35:12.001279"
  },
  {
    "epoch": 2729,
    "batch_idx": 0,
    "loss": 0.23819001019001007,
    "learning_rate": 0.0002935272603749221,
    "timestep_mean": 409.1875,
    "epsilon_norm": 6.734854698181152,
    "prediction_norm": 5.6937079429626465,
    "timestamp": "2025-07-14T12:35:12.006558"
  },
  {
    "epoch": 2729,
    "batch_idx": 32,
    "loss": 0.21124018728733063,
    "learning_rate": 0.0002935272603749221,
    "timestep_mean": 499.59375,
    "epsilon_norm": 7.490490913391113,
    "prediction_norm": 6.669551372528076,
    "timestamp": "2025-07-14T12:35:12.011793"
  },
  {
    "epoch": 2729,
    "batch_idx": 64,
    "loss": 0.2409709095954895,
    "learning_rate": 0.0002935272603749221,
    "timestep_mean": 410.78125,
    "epsilon_norm": 7.425663471221924,
    "prediction_norm": 6.211584091186523,
    "timestamp": "2025-07-14T12:35:12.017056"
  },
  {
    "epoch": 2729,
    "batch_idx": 96,
    "loss": 0.09468570351600647,
    "learning_rate": 0.0002935272603749221,
    "timestep_mean": 560.03125,
    "epsilon_norm": 7.606462478637695,
    "prediction_norm": 7.556482315063477,
    "timestamp": "2025-07-14T12:35:12.022332"
  },
  {
    "epoch": 2729,
    "batch_idx": 128,
    "loss": 0.2868369221687317,
    "learning_rate": 0.0002935272603749221,
    "timestep_mean": 430.5625,
    "epsilon_norm": 7.255140781402588,
    "prediction_norm": 5.840218544006348,
    "timestamp": "2025-07-14T12:35:12.027555"
  },
  {
    "epoch": 2729,
    "batch_idx": 160,
    "loss": 0.2546338737010956,
    "learning_rate": 0.0002935272603749221,
    "timestep_mean": 474.6875,
    "epsilon_norm": 7.1959428787231445,
    "prediction_norm": 6.933826923370361,
    "timestamp": "2025-07-14T12:35:12.032831"
  },
  {
    "epoch": 2729,
    "batch_idx": 192,
    "loss": 0.15726977586746216,
    "learning_rate": 0.0002935272603749221,
    "timestep_mean": 581.625,
    "epsilon_norm": 3.8613572120666504,
    "prediction_norm": 3.6739730834960938,
    "timestamp": "2025-07-14T12:35:12.038002"
  },
  {
    "epoch": 2730,
    "batch_idx": 0,
    "loss": 0.39276519417762756,
    "learning_rate": 0.0002942428206974461,
    "timestep_mean": 441.84375,
    "epsilon_norm": 7.685040473937988,
    "prediction_norm": 5.627835273742676,
    "timestamp": "2025-07-14T12:35:12.043163"
  },
  {
    "epoch": 2730,
    "batch_idx": 32,
    "loss": 0.2214607298374176,
    "learning_rate": 0.0002942428206974461,
    "timestep_mean": 415.84375,
    "epsilon_norm": 7.12129020690918,
    "prediction_norm": 6.051755428314209,
    "timestamp": "2025-07-14T12:35:12.048432"
  },
  {
    "epoch": 2730,
    "batch_idx": 64,
    "loss": 0.5065875053405762,
    "learning_rate": 0.0002942428206974461,
    "timestep_mean": 533.53125,
    "epsilon_norm": 8.826223373413086,
    "prediction_norm": 7.568869590759277,
    "timestamp": "2025-07-14T12:35:12.053677"
  },
  {
    "epoch": 2730,
    "batch_idx": 96,
    "loss": 0.2051253765821457,
    "learning_rate": 0.0002942428206974461,
    "timestep_mean": 520.6875,
    "epsilon_norm": 7.35471773147583,
    "prediction_norm": 6.207861423492432,
    "timestamp": "2025-07-14T12:35:12.058953"
  },
  {
    "epoch": 2730,
    "batch_idx": 128,
    "loss": 0.3268115520477295,
    "learning_rate": 0.0002942428206974461,
    "timestep_mean": 507.5,
    "epsilon_norm": 7.693264484405518,
    "prediction_norm": 6.548259258270264,
    "timestamp": "2025-07-14T12:35:12.064162"
  },
  {
    "epoch": 2730,
    "batch_idx": 160,
    "loss": 0.2345922589302063,
    "learning_rate": 0.0002942428206974461,
    "timestep_mean": 514.5625,
    "epsilon_norm": 7.512180328369141,
    "prediction_norm": 6.612739086151123,
    "timestamp": "2025-07-14T12:35:12.069432"
  },
  {
    "epoch": 2730,
    "batch_idx": 192,
    "loss": 0.1886088103055954,
    "learning_rate": 0.0002942428206974461,
    "timestep_mean": 426.25,
    "epsilon_norm": 3.736229658126831,
    "prediction_norm": 3.21258544921875,
    "timestamp": "2025-07-14T12:35:12.074594"
  },
  {
    "epoch": 2731,
    "batch_idx": 0,
    "loss": 0.18532735109329224,
    "learning_rate": 0.0002949588887053561,
    "timestep_mean": 488.375,
    "epsilon_norm": 7.047242164611816,
    "prediction_norm": 6.071500778198242,
    "timestamp": "2025-07-14T12:35:12.079853"
  },
  {
    "epoch": 2731,
    "batch_idx": 32,
    "loss": 0.1829790472984314,
    "learning_rate": 0.0002949588887053561,
    "timestep_mean": 609.5,
    "epsilon_norm": 8.083066940307617,
    "prediction_norm": 7.489924430847168,
    "timestamp": "2025-07-14T12:35:12.085025"
  },
  {
    "epoch": 2731,
    "batch_idx": 64,
    "loss": 0.25126370787620544,
    "learning_rate": 0.0002949588887053561,
    "timestep_mean": 542.53125,
    "epsilon_norm": 8.121150016784668,
    "prediction_norm": 6.8711347579956055,
    "timestamp": "2025-07-14T12:35:12.090256"
  },
  {
    "epoch": 2731,
    "batch_idx": 96,
    "loss": 0.12123546749353409,
    "learning_rate": 0.0002949588887053561,
    "timestep_mean": 584.625,
    "epsilon_norm": 8.02177906036377,
    "prediction_norm": 7.2850446701049805,
    "timestamp": "2025-07-14T12:35:12.095510"
  },
  {
    "epoch": 2731,
    "batch_idx": 128,
    "loss": 0.2879200577735901,
    "learning_rate": 0.0002949588887053561,
    "timestep_mean": 519.78125,
    "epsilon_norm": 7.043971061706543,
    "prediction_norm": 5.837575912475586,
    "timestamp": "2025-07-14T12:35:12.100850"
  },
  {
    "epoch": 2731,
    "batch_idx": 160,
    "loss": 0.15286991000175476,
    "learning_rate": 0.0002949588887053561,
    "timestep_mean": 467.9375,
    "epsilon_norm": 6.835423946380615,
    "prediction_norm": 6.21080207824707,
    "timestamp": "2025-07-14T12:35:12.106061"
  },
  {
    "epoch": 2731,
    "batch_idx": 192,
    "loss": 0.23637351393699646,
    "learning_rate": 0.0002949588887053561,
    "timestep_mean": 504.0,
    "epsilon_norm": 3.319946527481079,
    "prediction_norm": 2.334174394607544,
    "timestamp": "2025-07-14T12:35:12.111235"
  },
  {
    "epoch": 2732,
    "batch_idx": 0,
    "loss": 0.18693149089813232,
    "learning_rate": 0.00029567546263182626,
    "timestep_mean": 502.625,
    "epsilon_norm": 7.903512954711914,
    "prediction_norm": 7.211854457855225,
    "timestamp": "2025-07-14T12:35:12.116422"
  },
  {
    "epoch": 2732,
    "batch_idx": 32,
    "loss": 0.2550855278968811,
    "learning_rate": 0.00029567546263182626,
    "timestep_mean": 421.53125,
    "epsilon_norm": 7.944708824157715,
    "prediction_norm": 6.51574182510376,
    "timestamp": "2025-07-14T12:35:12.121788"
  },
  {
    "epoch": 2732,
    "batch_idx": 64,
    "loss": 0.3433932065963745,
    "learning_rate": 0.00029567546263182626,
    "timestep_mean": 455.125,
    "epsilon_norm": 8.493298530578613,
    "prediction_norm": 7.134037494659424,
    "timestamp": "2025-07-14T12:35:12.127091"
  },
  {
    "epoch": 2732,
    "batch_idx": 96,
    "loss": 0.15834689140319824,
    "learning_rate": 0.00029567546263182626,
    "timestep_mean": 550.875,
    "epsilon_norm": 8.420798301696777,
    "prediction_norm": 7.628335952758789,
    "timestamp": "2025-07-14T12:35:12.132279"
  },
  {
    "epoch": 2732,
    "batch_idx": 128,
    "loss": 0.1414462924003601,
    "learning_rate": 0.00029567546263182626,
    "timestep_mean": 452.28125,
    "epsilon_norm": 6.2696685791015625,
    "prediction_norm": 5.460509777069092,
    "timestamp": "2025-07-14T12:35:12.137476"
  },
  {
    "epoch": 2732,
    "batch_idx": 160,
    "loss": 0.2279573529958725,
    "learning_rate": 0.00029567546263182626,
    "timestep_mean": 436.25,
    "epsilon_norm": 8.232048034667969,
    "prediction_norm": 6.9768805503845215,
    "timestamp": "2025-07-14T12:35:12.142947"
  },
  {
    "epoch": 2732,
    "batch_idx": 192,
    "loss": 0.33609580993652344,
    "learning_rate": 0.00029567546263182626,
    "timestep_mean": 334.0,
    "epsilon_norm": 3.607851266860962,
    "prediction_norm": 2.705782890319824,
    "timestamp": "2025-07-14T12:35:12.148109"
  },
  {
    "epoch": 2733,
    "batch_idx": 0,
    "loss": 0.13706918060779572,
    "learning_rate": 0.0002963925407087806,
    "timestep_mean": 543.28125,
    "epsilon_norm": 7.9785566329956055,
    "prediction_norm": 7.358309268951416,
    "timestamp": "2025-07-14T12:35:12.153406"
  },
  {
    "epoch": 2733,
    "batch_idx": 32,
    "loss": 0.19798395037651062,
    "learning_rate": 0.0002963925407087806,
    "timestep_mean": 561.59375,
    "epsilon_norm": 8.00821590423584,
    "prediction_norm": 7.298537731170654,
    "timestamp": "2025-07-14T12:35:12.158730"
  },
  {
    "epoch": 2733,
    "batch_idx": 64,
    "loss": 0.26536762714385986,
    "learning_rate": 0.0002963925407087806,
    "timestep_mean": 407.21875,
    "epsilon_norm": 9.28608512878418,
    "prediction_norm": 7.730309963226318,
    "timestamp": "2025-07-14T12:35:12.163641"
  },
  {
    "epoch": 2733,
    "batch_idx": 96,
    "loss": 0.2530028820037842,
    "learning_rate": 0.0002963925407087806,
    "timestep_mean": 476.75,
    "epsilon_norm": 9.224687576293945,
    "prediction_norm": 7.609028339385986,
    "timestamp": "2025-07-14T12:35:12.168874"
  },
  {
    "epoch": 2733,
    "batch_idx": 128,
    "loss": 0.4277004599571228,
    "learning_rate": 0.0002963925407087806,
    "timestep_mean": 526.46875,
    "epsilon_norm": 8.278645515441895,
    "prediction_norm": 6.898491859436035,
    "timestamp": "2025-07-14T12:35:12.174146"
  },
  {
    "epoch": 2733,
    "batch_idx": 160,
    "loss": 0.6059252023696899,
    "learning_rate": 0.0002963925407087806,
    "timestep_mean": 391.25,
    "epsilon_norm": 9.633932113647461,
    "prediction_norm": 7.084443092346191,
    "timestamp": "2025-07-14T12:35:12.179221"
  },
  {
    "epoch": 2733,
    "batch_idx": 192,
    "loss": 0.05495483800768852,
    "learning_rate": 0.0002963925407087806,
    "timestep_mean": 670.625,
    "epsilon_norm": 3.415726661682129,
    "prediction_norm": 3.4866974353790283,
    "timestamp": "2025-07-14T12:35:12.184388"
  },
  {
    "epoch": 2734,
    "batch_idx": 0,
    "loss": 0.14826112985610962,
    "learning_rate": 0.0002971101211669003,
    "timestep_mean": 597.8125,
    "epsilon_norm": 7.991868019104004,
    "prediction_norm": 7.7578301429748535,
    "timestamp": "2025-07-14T12:35:12.189633"
  },
  {
    "epoch": 2734,
    "batch_idx": 32,
    "loss": 0.2884039878845215,
    "learning_rate": 0.0002971101211669003,
    "timestep_mean": 415.84375,
    "epsilon_norm": 8.671982765197754,
    "prediction_norm": 7.0075578689575195,
    "timestamp": "2025-07-14T12:35:12.194740"
  },
  {
    "epoch": 2734,
    "batch_idx": 64,
    "loss": 0.2143116295337677,
    "learning_rate": 0.0002971101211669003,
    "timestep_mean": 542.40625,
    "epsilon_norm": 7.4816813468933105,
    "prediction_norm": 6.568613052368164,
    "timestamp": "2025-07-14T12:35:12.199983"
  },
  {
    "epoch": 2734,
    "batch_idx": 96,
    "loss": 0.5082968473434448,
    "learning_rate": 0.0002971101211669003,
    "timestep_mean": 505.5625,
    "epsilon_norm": 8.93844985961914,
    "prediction_norm": 7.540999889373779,
    "timestamp": "2025-07-14T12:35:12.205403"
  },
  {
    "epoch": 2734,
    "batch_idx": 128,
    "loss": 0.4306928217411041,
    "learning_rate": 0.0002971101211669003,
    "timestep_mean": 495.46875,
    "epsilon_norm": 9.076579093933105,
    "prediction_norm": 7.509546756744385,
    "timestamp": "2025-07-14T12:35:12.210671"
  },
  {
    "epoch": 2734,
    "batch_idx": 160,
    "loss": 0.3011090159416199,
    "learning_rate": 0.0002971101211669003,
    "timestep_mean": 471.5625,
    "epsilon_norm": 7.689469814300537,
    "prediction_norm": 6.383995056152344,
    "timestamp": "2025-07-14T12:35:12.215928"
  },
  {
    "epoch": 2734,
    "batch_idx": 192,
    "loss": 0.29990440607070923,
    "learning_rate": 0.0002971101211669003,
    "timestep_mean": 328.875,
    "epsilon_norm": 3.2990543842315674,
    "prediction_norm": 2.485231637954712,
    "timestamp": "2025-07-14T12:35:12.220994"
  },
  {
    "epoch": 2735,
    "batch_idx": 0,
    "loss": 0.12754839658737183,
    "learning_rate": 0.00029782820223562813,
    "timestep_mean": 482.59375,
    "epsilon_norm": 8.020487785339355,
    "prediction_norm": 7.284704208374023,
    "timestamp": "2025-07-14T12:35:12.226306"
  },
  {
    "epoch": 2735,
    "batch_idx": 32,
    "loss": 0.41718626022338867,
    "learning_rate": 0.00029782820223562813,
    "timestep_mean": 519.625,
    "epsilon_norm": 9.435688018798828,
    "prediction_norm": 8.101693153381348,
    "timestamp": "2025-07-14T12:35:12.231453"
  },
  {
    "epoch": 2735,
    "batch_idx": 64,
    "loss": 0.5627073049545288,
    "learning_rate": 0.00029782820223562813,
    "timestep_mean": 435.375,
    "epsilon_norm": 8.242804527282715,
    "prediction_norm": 6.195561408996582,
    "timestamp": "2025-07-14T12:35:12.236639"
  },
  {
    "epoch": 2735,
    "batch_idx": 96,
    "loss": 0.4113769829273224,
    "learning_rate": 0.00029782820223562813,
    "timestep_mean": 485.75,
    "epsilon_norm": 7.843433856964111,
    "prediction_norm": 5.467298984527588,
    "timestamp": "2025-07-14T12:35:12.241851"
  },
  {
    "epoch": 2735,
    "batch_idx": 128,
    "loss": 0.18082848191261292,
    "learning_rate": 0.00029782820223562813,
    "timestep_mean": 471.75,
    "epsilon_norm": 7.6379780769348145,
    "prediction_norm": 7.165949821472168,
    "timestamp": "2025-07-14T12:35:12.247155"
  },
  {
    "epoch": 2735,
    "batch_idx": 160,
    "loss": 0.21279402077198029,
    "learning_rate": 0.00029782820223562813,
    "timestep_mean": 483.21875,
    "epsilon_norm": 7.2572479248046875,
    "prediction_norm": 6.768487930297852,
    "timestamp": "2025-07-14T12:35:12.252374"
  },
  {
    "epoch": 2735,
    "batch_idx": 192,
    "loss": 0.40090787410736084,
    "learning_rate": 0.00029782820223562813,
    "timestep_mean": 349.75,
    "epsilon_norm": 3.364891290664673,
    "prediction_norm": 2.425511360168457,
    "timestamp": "2025-07-14T12:35:12.257539"
  },
  {
    "epoch": 2736,
    "batch_idx": 0,
    "loss": 0.35014328360557556,
    "learning_rate": 0.0002985467821431693,
    "timestep_mean": 547.625,
    "epsilon_norm": 8.133074760437012,
    "prediction_norm": 7.720625877380371,
    "timestamp": "2025-07-14T12:35:12.262761"
  },
  {
    "epoch": 2736,
    "batch_idx": 32,
    "loss": 0.1499829739332199,
    "learning_rate": 0.0002985467821431693,
    "timestep_mean": 554.375,
    "epsilon_norm": 8.236236572265625,
    "prediction_norm": 7.58168363571167,
    "timestamp": "2025-07-14T12:35:12.267929"
  },
  {
    "epoch": 2736,
    "batch_idx": 64,
    "loss": 0.32644879817962646,
    "learning_rate": 0.0002985467821431693,
    "timestep_mean": 534.46875,
    "epsilon_norm": 8.85977840423584,
    "prediction_norm": 7.758255481719971,
    "timestamp": "2025-07-14T12:35:12.273148"
  },
  {
    "epoch": 2736,
    "batch_idx": 96,
    "loss": 0.27495628595352173,
    "learning_rate": 0.0002985467821431693,
    "timestep_mean": 400.8125,
    "epsilon_norm": 8.14005184173584,
    "prediction_norm": 7.081725120544434,
    "timestamp": "2025-07-14T12:35:12.278278"
  },
  {
    "epoch": 2736,
    "batch_idx": 128,
    "loss": 0.2135266214609146,
    "learning_rate": 0.0002985467821431693,
    "timestep_mean": 544.15625,
    "epsilon_norm": 7.981362819671631,
    "prediction_norm": 7.530584335327148,
    "timestamp": "2025-07-14T12:35:12.283773"
  },
  {
    "epoch": 2736,
    "batch_idx": 160,
    "loss": 0.30018728971481323,
    "learning_rate": 0.0002985467821431693,
    "timestep_mean": 569.21875,
    "epsilon_norm": 8.448198318481445,
    "prediction_norm": 7.666265964508057,
    "timestamp": "2025-07-14T12:35:12.289063"
  },
  {
    "epoch": 2736,
    "batch_idx": 192,
    "loss": 0.6226838827133179,
    "learning_rate": 0.0002985467821431693,
    "timestep_mean": 403.0,
    "epsilon_norm": 3.896672487258911,
    "prediction_norm": 2.8497445583343506,
    "timestamp": "2025-07-14T12:35:12.294296"
  },
  {
    "epoch": 2737,
    "batch_idx": 0,
    "loss": 0.1860799342393875,
    "learning_rate": 0.0002992658591164997,
    "timestep_mean": 526.09375,
    "epsilon_norm": 6.967533111572266,
    "prediction_norm": 6.489552974700928,
    "timestamp": "2025-07-14T12:35:12.299526"
  },
  {
    "epoch": 2737,
    "batch_idx": 32,
    "loss": 0.22133031487464905,
    "learning_rate": 0.0002992658591164997,
    "timestep_mean": 539.125,
    "epsilon_norm": 7.790953159332275,
    "prediction_norm": 7.354147434234619,
    "timestamp": "2025-07-14T12:35:12.304760"
  },
  {
    "epoch": 2737,
    "batch_idx": 64,
    "loss": 0.24082157015800476,
    "learning_rate": 0.0002992658591164997,
    "timestep_mean": 501.59375,
    "epsilon_norm": 8.932846069335938,
    "prediction_norm": 7.775537490844727,
    "timestamp": "2025-07-14T12:35:12.309961"
  },
  {
    "epoch": 2737,
    "batch_idx": 96,
    "loss": 0.49021416902542114,
    "learning_rate": 0.0002992658591164997,
    "timestep_mean": 506.78125,
    "epsilon_norm": 9.608001708984375,
    "prediction_norm": 7.357234001159668,
    "timestamp": "2025-07-14T12:35:12.315200"
  },
  {
    "epoch": 2737,
    "batch_idx": 128,
    "loss": 0.2204684317111969,
    "learning_rate": 0.0002992658591164997,
    "timestep_mean": 582.6875,
    "epsilon_norm": 8.762534141540527,
    "prediction_norm": 8.108187675476074,
    "timestamp": "2025-07-14T12:35:12.320447"
  },
  {
    "epoch": 2737,
    "batch_idx": 160,
    "loss": 0.12380839139223099,
    "learning_rate": 0.0002992658591164997,
    "timestep_mean": 515.5625,
    "epsilon_norm": 6.769586086273193,
    "prediction_norm": 6.395465850830078,
    "timestamp": "2025-07-14T12:35:12.325663"
  },
  {
    "epoch": 2737,
    "batch_idx": 192,
    "loss": 0.1479790061712265,
    "learning_rate": 0.0002992658591164997,
    "timestep_mean": 565.875,
    "epsilon_norm": 4.100225925445557,
    "prediction_norm": 3.4200451374053955,
    "timestamp": "2025-07-14T12:35:12.330831"
  },
  {
    "epoch": 2738,
    "batch_idx": 0,
    "loss": 0.23883849382400513,
    "learning_rate": 0.0002999854313813687,
    "timestep_mean": 420.75,
    "epsilon_norm": 6.423959255218506,
    "prediction_norm": 5.018407344818115,
    "timestamp": "2025-07-14T12:35:12.336053"
  },
  {
    "epoch": 2738,
    "batch_idx": 32,
    "loss": 0.25923752784729004,
    "learning_rate": 0.0002999854313813687,
    "timestep_mean": 567.375,
    "epsilon_norm": 8.631914138793945,
    "prediction_norm": 7.179948806762695,
    "timestamp": "2025-07-14T12:35:12.341269"
  },
  {
    "epoch": 2738,
    "batch_idx": 64,
    "loss": 0.21962526440620422,
    "learning_rate": 0.0002999854313813687,
    "timestep_mean": 496.53125,
    "epsilon_norm": 8.417679786682129,
    "prediction_norm": 7.304574489593506,
    "timestamp": "2025-07-14T12:35:12.346550"
  },
  {
    "epoch": 2738,
    "batch_idx": 96,
    "loss": 0.16477026045322418,
    "learning_rate": 0.0002999854313813687,
    "timestep_mean": 513.375,
    "epsilon_norm": 5.903262138366699,
    "prediction_norm": 5.13225793838501,
    "timestamp": "2025-07-14T12:35:12.351758"
  },
  {
    "epoch": 2738,
    "batch_idx": 128,
    "loss": 0.18883772194385529,
    "learning_rate": 0.0002999854313813687,
    "timestep_mean": 459.9375,
    "epsilon_norm": 7.933727264404297,
    "prediction_norm": 6.737781524658203,
    "timestamp": "2025-07-14T12:35:12.356864"
  },
  {
    "epoch": 2738,
    "batch_idx": 160,
    "loss": 0.3423699140548706,
    "learning_rate": 0.0002999854313813687,
    "timestep_mean": 428.5,
    "epsilon_norm": 7.999568939208984,
    "prediction_norm": 5.719336032867432,
    "timestamp": "2025-07-14T12:35:12.361715"
  },
  {
    "epoch": 2738,
    "batch_idx": 192,
    "loss": 0.3563629984855652,
    "learning_rate": 0.0002999854313813687,
    "timestep_mean": 511.375,
    "epsilon_norm": 4.526101589202881,
    "prediction_norm": 3.661917209625244,
    "timestamp": "2025-07-14T12:35:12.366579"
  },
  {
    "epoch": 2739,
    "batch_idx": 0,
    "loss": 0.25844380259513855,
    "learning_rate": 0.00030070549716230196,
    "timestep_mean": 469.03125,
    "epsilon_norm": 8.317163467407227,
    "prediction_norm": 7.640166282653809,
    "timestamp": "2025-07-14T12:35:12.371515"
  },
  {
    "epoch": 2739,
    "batch_idx": 32,
    "loss": 0.19284310936927795,
    "learning_rate": 0.00030070549716230196,
    "timestep_mean": 488.53125,
    "epsilon_norm": 7.168805122375488,
    "prediction_norm": 6.913932800292969,
    "timestamp": "2025-07-14T12:35:12.376413"
  },
  {
    "epoch": 2739,
    "batch_idx": 64,
    "loss": 0.14096547663211823,
    "learning_rate": 0.00030070549716230196,
    "timestep_mean": 496.5625,
    "epsilon_norm": 7.878046989440918,
    "prediction_norm": 7.1840667724609375,
    "timestamp": "2025-07-14T12:35:12.381380"
  },
  {
    "epoch": 2739,
    "batch_idx": 96,
    "loss": 0.13449399173259735,
    "learning_rate": 0.00030070549716230196,
    "timestep_mean": 541.625,
    "epsilon_norm": 9.198211669921875,
    "prediction_norm": 8.463831901550293,
    "timestamp": "2025-07-14T12:35:12.386498"
  },
  {
    "epoch": 2739,
    "batch_idx": 128,
    "loss": 0.3283657431602478,
    "learning_rate": 0.00030070549716230196,
    "timestep_mean": 493.0625,
    "epsilon_norm": 7.463912010192871,
    "prediction_norm": 5.815418720245361,
    "timestamp": "2025-07-14T12:35:12.391682"
  },
  {
    "epoch": 2739,
    "batch_idx": 160,
    "loss": 0.24501310288906097,
    "learning_rate": 0.00030070549716230196,
    "timestep_mean": 507.21875,
    "epsilon_norm": 9.05805778503418,
    "prediction_norm": 8.189667701721191,
    "timestamp": "2025-07-14T12:35:12.396845"
  },
  {
    "epoch": 2739,
    "batch_idx": 192,
    "loss": 0.12592652440071106,
    "learning_rate": 0.00030070549716230196,
    "timestep_mean": 534.0,
    "epsilon_norm": 3.951395273208618,
    "prediction_norm": 3.90531325340271,
    "timestamp": "2025-07-14T12:35:12.401900"
  },
  {
    "epoch": 2740,
    "batch_idx": 0,
    "loss": 0.170698881149292,
    "learning_rate": 0.00030142605468261015,
    "timestep_mean": 537.625,
    "epsilon_norm": 7.9558258056640625,
    "prediction_norm": 6.986293315887451,
    "timestamp": "2025-07-14T12:35:12.406909"
  },
  {
    "epoch": 2740,
    "batch_idx": 32,
    "loss": 0.21941694617271423,
    "learning_rate": 0.00030142605468261015,
    "timestep_mean": 589.5625,
    "epsilon_norm": 8.194866180419922,
    "prediction_norm": 7.244954586029053,
    "timestamp": "2025-07-14T12:35:12.411781"
  },
  {
    "epoch": 2740,
    "batch_idx": 64,
    "loss": 0.10408850014209747,
    "learning_rate": 0.00030142605468261015,
    "timestep_mean": 515.21875,
    "epsilon_norm": 6.393520832061768,
    "prediction_norm": 6.088836193084717,
    "timestamp": "2025-07-14T12:35:12.417201"
  },
  {
    "epoch": 2740,
    "batch_idx": 96,
    "loss": 0.17989163100719452,
    "learning_rate": 0.00030142605468261015,
    "timestep_mean": 496.71875,
    "epsilon_norm": 8.203225135803223,
    "prediction_norm": 7.102659225463867,
    "timestamp": "2025-07-14T12:35:12.422401"
  },
  {
    "epoch": 2740,
    "batch_idx": 128,
    "loss": 0.2391817569732666,
    "learning_rate": 0.00030142605468261015,
    "timestep_mean": 580.625,
    "epsilon_norm": 8.788708686828613,
    "prediction_norm": 7.305954933166504,
    "timestamp": "2025-07-14T12:35:12.427551"
  },
  {
    "epoch": 2740,
    "batch_idx": 160,
    "loss": 0.27507972717285156,
    "learning_rate": 0.00030142605468261015,
    "timestep_mean": 561.09375,
    "epsilon_norm": 6.684134006500244,
    "prediction_norm": 5.699765682220459,
    "timestamp": "2025-07-14T12:35:12.432761"
  },
  {
    "epoch": 2740,
    "batch_idx": 192,
    "loss": 0.41132915019989014,
    "learning_rate": 0.00030142605468261015,
    "timestep_mean": 302.625,
    "epsilon_norm": 3.6050517559051514,
    "prediction_norm": 2.75622296333313,
    "timestamp": "2025-07-14T12:35:12.438067"
  },
  {
    "epoch": 2741,
    "batch_idx": 0,
    "loss": 0.35074713826179504,
    "learning_rate": 0.0003021471021643889,
    "timestep_mean": 514.4375,
    "epsilon_norm": 8.162505149841309,
    "prediction_norm": 6.886428356170654,
    "timestamp": "2025-07-14T12:35:12.443228"
  },
  {
    "epoch": 2741,
    "batch_idx": 32,
    "loss": 0.4019492268562317,
    "learning_rate": 0.0003021471021643889,
    "timestep_mean": 396.34375,
    "epsilon_norm": 8.067313194274902,
    "prediction_norm": 7.32677698135376,
    "timestamp": "2025-07-14T12:35:12.448465"
  },
  {
    "epoch": 2741,
    "batch_idx": 64,
    "loss": 0.35421258211135864,
    "learning_rate": 0.0003021471021643889,
    "timestep_mean": 456.09375,
    "epsilon_norm": 8.012445449829102,
    "prediction_norm": 6.518980503082275,
    "timestamp": "2025-07-14T12:35:12.453420"
  },
  {
    "epoch": 2741,
    "batch_idx": 96,
    "loss": 0.19290266931056976,
    "learning_rate": 0.0003021471021643889,
    "timestep_mean": 513.15625,
    "epsilon_norm": 7.295529842376709,
    "prediction_norm": 6.482341766357422,
    "timestamp": "2025-07-14T12:35:12.458263"
  },
  {
    "epoch": 2741,
    "batch_idx": 128,
    "loss": 0.22212094068527222,
    "learning_rate": 0.0003021471021643889,
    "timestep_mean": 492.0625,
    "epsilon_norm": 7.841999530792236,
    "prediction_norm": 6.60170841217041,
    "timestamp": "2025-07-14T12:35:12.463327"
  },
  {
    "epoch": 2741,
    "batch_idx": 160,
    "loss": 0.38885194063186646,
    "learning_rate": 0.0003021471021643889,
    "timestep_mean": 431.75,
    "epsilon_norm": 8.787171363830566,
    "prediction_norm": 6.736327171325684,
    "timestamp": "2025-07-14T12:35:12.468396"
  },
  {
    "epoch": 2741,
    "batch_idx": 192,
    "loss": 0.37965503334999084,
    "learning_rate": 0.0003021471021643889,
    "timestep_mean": 609.375,
    "epsilon_norm": 4.573579788208008,
    "prediction_norm": 3.2651760578155518,
    "timestamp": "2025-07-14T12:35:12.473455"
  },
  {
    "epoch": 2742,
    "batch_idx": 0,
    "loss": 0.28758737444877625,
    "learning_rate": 0.0003028686378285254,
    "timestep_mean": 442.5,
    "epsilon_norm": 9.178075790405273,
    "prediction_norm": 7.9552154541015625,
    "timestamp": "2025-07-14T12:35:12.478626"
  },
  {
    "epoch": 2742,
    "batch_idx": 32,
    "loss": 0.22765979170799255,
    "learning_rate": 0.0003028686378285254,
    "timestep_mean": 468.03125,
    "epsilon_norm": 8.047087669372559,
    "prediction_norm": 7.301349639892578,
    "timestamp": "2025-07-14T12:35:12.483905"
  },
  {
    "epoch": 2742,
    "batch_idx": 64,
    "loss": 0.42450177669525146,
    "learning_rate": 0.0003028686378285254,
    "timestep_mean": 524.625,
    "epsilon_norm": 8.821756362915039,
    "prediction_norm": 6.737730026245117,
    "timestamp": "2025-07-14T12:35:12.488944"
  },
  {
    "epoch": 2742,
    "batch_idx": 96,
    "loss": 0.16840726137161255,
    "learning_rate": 0.0003028686378285254,
    "timestep_mean": 520.3125,
    "epsilon_norm": 8.1579008102417,
    "prediction_norm": 6.912408828735352,
    "timestamp": "2025-07-14T12:35:12.494042"
  },
  {
    "epoch": 2742,
    "batch_idx": 128,
    "loss": 0.6298199892044067,
    "learning_rate": 0.0003028686378285254,
    "timestep_mean": 482.96875,
    "epsilon_norm": 8.55378532409668,
    "prediction_norm": 6.0812482833862305,
    "timestamp": "2025-07-14T12:35:12.499214"
  },
  {
    "epoch": 2742,
    "batch_idx": 160,
    "loss": 0.3319235146045685,
    "learning_rate": 0.0003028686378285254,
    "timestep_mean": 460.21875,
    "epsilon_norm": 8.197848320007324,
    "prediction_norm": 7.001638889312744,
    "timestamp": "2025-07-14T12:35:12.504366"
  },
  {
    "epoch": 2742,
    "batch_idx": 192,
    "loss": 0.17994926869869232,
    "learning_rate": 0.0003028686378285254,
    "timestep_mean": 480.625,
    "epsilon_norm": 3.1659131050109863,
    "prediction_norm": 2.455670118331909,
    "timestamp": "2025-07-14T12:35:12.509590"
  },
  {
    "epoch": 2743,
    "batch_idx": 0,
    "loss": 0.2251410037279129,
    "learning_rate": 0.0003035906598947014,
    "timestep_mean": 436.0,
    "epsilon_norm": 8.189587593078613,
    "prediction_norm": 7.020808219909668,
    "timestamp": "2025-07-14T12:35:12.514847"
  },
  {
    "epoch": 2743,
    "batch_idx": 32,
    "loss": 0.3519812226295471,
    "learning_rate": 0.0003035906598947014,
    "timestep_mean": 513.53125,
    "epsilon_norm": 8.538168907165527,
    "prediction_norm": 6.4655561447143555,
    "timestamp": "2025-07-14T12:35:12.520082"
  },
  {
    "epoch": 2743,
    "batch_idx": 64,
    "loss": 0.36691582202911377,
    "learning_rate": 0.0003035906598947014,
    "timestep_mean": 557.6875,
    "epsilon_norm": 7.935049057006836,
    "prediction_norm": 6.510095596313477,
    "timestamp": "2025-07-14T12:35:12.525465"
  },
  {
    "epoch": 2743,
    "batch_idx": 96,
    "loss": 0.2425013780593872,
    "learning_rate": 0.0003035906598947014,
    "timestep_mean": 439.9375,
    "epsilon_norm": 8.088373184204102,
    "prediction_norm": 6.793920040130615,
    "timestamp": "2025-07-14T12:35:12.530623"
  },
  {
    "epoch": 2743,
    "batch_idx": 128,
    "loss": 0.31857365369796753,
    "learning_rate": 0.0003035906598947014,
    "timestep_mean": 507.65625,
    "epsilon_norm": 8.152142524719238,
    "prediction_norm": 6.7967634201049805,
    "timestamp": "2025-07-14T12:35:12.535849"
  },
  {
    "epoch": 2743,
    "batch_idx": 160,
    "loss": 0.1838846504688263,
    "learning_rate": 0.0003035906598947014,
    "timestep_mean": 555.15625,
    "epsilon_norm": 6.883500099182129,
    "prediction_norm": 6.460198879241943,
    "timestamp": "2025-07-14T12:35:12.541125"
  },
  {
    "epoch": 2743,
    "batch_idx": 192,
    "loss": 0.306485652923584,
    "learning_rate": 0.0003035906598947014,
    "timestep_mean": 494.625,
    "epsilon_norm": 4.162879943847656,
    "prediction_norm": 3.6151528358459473,
    "timestamp": "2025-07-14T12:35:12.546164"
  },
  {
    "epoch": 2744,
    "batch_idx": 0,
    "loss": 0.36662259697914124,
    "learning_rate": 0.00030431316658139957,
    "timestep_mean": 443.78125,
    "epsilon_norm": 8.283909797668457,
    "prediction_norm": 6.912137031555176,
    "timestamp": "2025-07-14T12:35:12.551317"
  },
  {
    "epoch": 2744,
    "batch_idx": 32,
    "loss": 0.12408818304538727,
    "learning_rate": 0.00030431316658139957,
    "timestep_mean": 511.5,
    "epsilon_norm": 7.387548446655273,
    "prediction_norm": 7.563137531280518,
    "timestamp": "2025-07-14T12:35:12.556574"
  },
  {
    "epoch": 2744,
    "batch_idx": 64,
    "loss": 0.1293589323759079,
    "learning_rate": 0.00030431316658139957,
    "timestep_mean": 526.21875,
    "epsilon_norm": 8.182808876037598,
    "prediction_norm": 7.650989532470703,
    "timestamp": "2025-07-14T12:35:12.561838"
  },
  {
    "epoch": 2744,
    "batch_idx": 96,
    "loss": 0.15089362859725952,
    "learning_rate": 0.00030431316658139957,
    "timestep_mean": 563.65625,
    "epsilon_norm": 8.314984321594238,
    "prediction_norm": 7.846529483795166,
    "timestamp": "2025-07-14T12:35:12.567239"
  },
  {
    "epoch": 2744,
    "batch_idx": 128,
    "loss": 0.4908795654773712,
    "learning_rate": 0.00030431316658139957,
    "timestep_mean": 491.40625,
    "epsilon_norm": 8.62099838256836,
    "prediction_norm": 6.505150318145752,
    "timestamp": "2025-07-14T12:35:12.572472"
  },
  {
    "epoch": 2744,
    "batch_idx": 160,
    "loss": 0.31436678767204285,
    "learning_rate": 0.00030431316658139957,
    "timestep_mean": 519.09375,
    "epsilon_norm": 7.688009738922119,
    "prediction_norm": 6.823684215545654,
    "timestamp": "2025-07-14T12:35:12.577707"
  },
  {
    "epoch": 2744,
    "batch_idx": 192,
    "loss": 0.4847266972064972,
    "learning_rate": 0.00030431316658139957,
    "timestep_mean": 307.5,
    "epsilon_norm": 3.623276710510254,
    "prediction_norm": 2.5884108543395996,
    "timestamp": "2025-07-14T12:35:12.582906"
  },
  {
    "epoch": 2745,
    "batch_idx": 0,
    "loss": 0.1921636164188385,
    "learning_rate": 0.0003050361561059066,
    "timestep_mean": 523.3125,
    "epsilon_norm": 7.8965253829956055,
    "prediction_norm": 7.2368340492248535,
    "timestamp": "2025-07-14T12:35:12.588069"
  },
  {
    "epoch": 2745,
    "batch_idx": 32,
    "loss": 0.27503499388694763,
    "learning_rate": 0.0003050361561059066,
    "timestep_mean": 374.09375,
    "epsilon_norm": 8.038745880126953,
    "prediction_norm": 7.1578688621521,
    "timestamp": "2025-07-14T12:35:12.593283"
  },
  {
    "epoch": 2745,
    "batch_idx": 64,
    "loss": 0.3540298342704773,
    "learning_rate": 0.0003050361561059066,
    "timestep_mean": 547.71875,
    "epsilon_norm": 8.267900466918945,
    "prediction_norm": 6.097001552581787,
    "timestamp": "2025-07-14T12:35:12.598423"
  },
  {
    "epoch": 2745,
    "batch_idx": 96,
    "loss": 0.21886014938354492,
    "learning_rate": 0.0003050361561059066,
    "timestep_mean": 539.8125,
    "epsilon_norm": 6.966179847717285,
    "prediction_norm": 6.33142614364624,
    "timestamp": "2025-07-14T12:35:12.603696"
  },
  {
    "epoch": 2745,
    "batch_idx": 128,
    "loss": 0.582358717918396,
    "learning_rate": 0.0003050361561059066,
    "timestep_mean": 528.34375,
    "epsilon_norm": 8.449727058410645,
    "prediction_norm": 6.4656829833984375,
    "timestamp": "2025-07-14T12:35:12.608991"
  },
  {
    "epoch": 2745,
    "batch_idx": 160,
    "loss": 0.2744169235229492,
    "learning_rate": 0.0003050361561059066,
    "timestep_mean": 432.3125,
    "epsilon_norm": 8.375447273254395,
    "prediction_norm": 7.0282487869262695,
    "timestamp": "2025-07-14T12:35:12.614232"
  },
  {
    "epoch": 2745,
    "batch_idx": 192,
    "loss": 0.14307144284248352,
    "learning_rate": 0.0003050361561059066,
    "timestep_mean": 446.5,
    "epsilon_norm": 3.789900541305542,
    "prediction_norm": 3.220377206802368,
    "timestamp": "2025-07-14T12:35:12.619407"
  },
  {
    "epoch": 2746,
    "batch_idx": 0,
    "loss": 0.2992916703224182,
    "learning_rate": 0.0003057596266843176,
    "timestep_mean": 476.40625,
    "epsilon_norm": 7.199051856994629,
    "prediction_norm": 5.457594394683838,
    "timestamp": "2025-07-14T12:35:12.624625"
  },
  {
    "epoch": 2746,
    "batch_idx": 32,
    "loss": 0.2688653767108917,
    "learning_rate": 0.0003057596266843176,
    "timestep_mean": 559.0625,
    "epsilon_norm": 7.116596221923828,
    "prediction_norm": 5.491275310516357,
    "timestamp": "2025-07-14T12:35:12.629862"
  },
  {
    "epoch": 2746,
    "batch_idx": 64,
    "loss": 0.1086321771144867,
    "learning_rate": 0.0003057596266843176,
    "timestep_mean": 525.71875,
    "epsilon_norm": 7.7993974685668945,
    "prediction_norm": 6.981699466705322,
    "timestamp": "2025-07-14T12:35:12.635105"
  },
  {
    "epoch": 2746,
    "batch_idx": 96,
    "loss": 0.4012335240840912,
    "learning_rate": 0.0003057596266843176,
    "timestep_mean": 540.34375,
    "epsilon_norm": 8.029783248901367,
    "prediction_norm": 5.997928142547607,
    "timestamp": "2025-07-14T12:35:12.640340"
  },
  {
    "epoch": 2746,
    "batch_idx": 128,
    "loss": 0.29879137873649597,
    "learning_rate": 0.0003057596266843176,
    "timestep_mean": 587.0625,
    "epsilon_norm": 7.462384223937988,
    "prediction_norm": 6.143448352813721,
    "timestamp": "2025-07-14T12:35:12.645479"
  },
  {
    "epoch": 2746,
    "batch_idx": 160,
    "loss": 0.3257998824119568,
    "learning_rate": 0.0003057596266843176,
    "timestep_mean": 532.5,
    "epsilon_norm": 9.071252822875977,
    "prediction_norm": 7.829283714294434,
    "timestamp": "2025-07-14T12:35:12.650645"
  },
  {
    "epoch": 2746,
    "batch_idx": 192,
    "loss": 0.148724764585495,
    "learning_rate": 0.0003057596266843176,
    "timestep_mean": 597.375,
    "epsilon_norm": 4.034237384796143,
    "prediction_norm": 3.660787582397461,
    "timestamp": "2025-07-14T12:35:12.655661"
  },
  {
    "epoch": 2747,
    "batch_idx": 0,
    "loss": 0.44372326135635376,
    "learning_rate": 0.0003064835765315409,
    "timestep_mean": 537.25,
    "epsilon_norm": 8.29173469543457,
    "prediction_norm": 6.507423400878906,
    "timestamp": "2025-07-14T12:35:12.660849"
  },
  {
    "epoch": 2747,
    "batch_idx": 32,
    "loss": 0.4575989842414856,
    "learning_rate": 0.0003064835765315409,
    "timestep_mean": 479.0625,
    "epsilon_norm": 8.301340103149414,
    "prediction_norm": 6.786902904510498,
    "timestamp": "2025-07-14T12:35:12.665829"
  },
  {
    "epoch": 2747,
    "batch_idx": 64,
    "loss": 0.2305385172367096,
    "learning_rate": 0.0003064835765315409,
    "timestep_mean": 543.40625,
    "epsilon_norm": 7.67746114730835,
    "prediction_norm": 5.6159796714782715,
    "timestamp": "2025-07-14T12:35:12.670830"
  },
  {
    "epoch": 2747,
    "batch_idx": 96,
    "loss": 0.14869369566440582,
    "learning_rate": 0.0003064835765315409,
    "timestep_mean": 570.0625,
    "epsilon_norm": 8.400230407714844,
    "prediction_norm": 7.959001541137695,
    "timestamp": "2025-07-14T12:35:12.676084"
  },
  {
    "epoch": 2747,
    "batch_idx": 128,
    "loss": 0.511452853679657,
    "learning_rate": 0.0003064835765315409,
    "timestep_mean": 462.5,
    "epsilon_norm": 8.88277816772461,
    "prediction_norm": 6.925545692443848,
    "timestamp": "2025-07-14T12:35:12.681139"
  },
  {
    "epoch": 2747,
    "batch_idx": 160,
    "loss": 0.17914791405200958,
    "learning_rate": 0.0003064835765315409,
    "timestep_mean": 577.40625,
    "epsilon_norm": 9.397286415100098,
    "prediction_norm": 8.71977424621582,
    "timestamp": "2025-07-14T12:35:12.686345"
  },
  {
    "epoch": 2747,
    "batch_idx": 192,
    "loss": 0.2342730164527893,
    "learning_rate": 0.0003064835765315409,
    "timestep_mean": 617.625,
    "epsilon_norm": 4.270842552185059,
    "prediction_norm": 4.1070237159729,
    "timestamp": "2025-07-14T12:35:12.691585"
  },
  {
    "epoch": 2748,
    "batch_idx": 0,
    "loss": 0.648564338684082,
    "learning_rate": 0.0003072080038613023,
    "timestep_mean": 441.09375,
    "epsilon_norm": 8.747335433959961,
    "prediction_norm": 6.859588146209717,
    "timestamp": "2025-07-14T12:35:12.696964"
  },
  {
    "epoch": 2748,
    "batch_idx": 32,
    "loss": 0.1559324562549591,
    "learning_rate": 0.0003072080038613023,
    "timestep_mean": 520.46875,
    "epsilon_norm": 8.267431259155273,
    "prediction_norm": 7.986385822296143,
    "timestamp": "2025-07-14T12:35:12.702072"
  },
  {
    "epoch": 2748,
    "batch_idx": 64,
    "loss": 0.2618151009082794,
    "learning_rate": 0.0003072080038613023,
    "timestep_mean": 389.46875,
    "epsilon_norm": 7.888316631317139,
    "prediction_norm": 6.8258233070373535,
    "timestamp": "2025-07-14T12:35:12.707262"
  },
  {
    "epoch": 2748,
    "batch_idx": 96,
    "loss": 0.26387426257133484,
    "learning_rate": 0.0003072080038613023,
    "timestep_mean": 518.8125,
    "epsilon_norm": 7.204933166503906,
    "prediction_norm": 6.097006320953369,
    "timestamp": "2025-07-14T12:35:12.712421"
  },
  {
    "epoch": 2748,
    "batch_idx": 128,
    "loss": 0.4379355311393738,
    "learning_rate": 0.0003072080038613023,
    "timestep_mean": 513.1875,
    "epsilon_norm": 7.3257317543029785,
    "prediction_norm": 5.02316427230835,
    "timestamp": "2025-07-14T12:35:12.717478"
  },
  {
    "epoch": 2748,
    "batch_idx": 160,
    "loss": 0.14495323598384857,
    "learning_rate": 0.0003072080038613023,
    "timestep_mean": 545.84375,
    "epsilon_norm": 7.647644519805908,
    "prediction_norm": 6.708899974822998,
    "timestamp": "2025-07-14T12:35:12.722373"
  },
  {
    "epoch": 2748,
    "batch_idx": 192,
    "loss": 0.23119881749153137,
    "learning_rate": 0.0003072080038613023,
    "timestep_mean": 376.25,
    "epsilon_norm": 4.543348789215088,
    "prediction_norm": 3.7681925296783447,
    "timestamp": "2025-07-14T12:35:12.727186"
  },
  {
    "epoch": 2749,
    "batch_idx": 0,
    "loss": 0.3771134912967682,
    "learning_rate": 0.0003079329068861496,
    "timestep_mean": 521.09375,
    "epsilon_norm": 8.103294372558594,
    "prediction_norm": 7.042061805725098,
    "timestamp": "2025-07-14T12:35:12.732081"
  },
  {
    "epoch": 2749,
    "batch_idx": 32,
    "loss": 0.15687543153762817,
    "learning_rate": 0.0003079329068861496,
    "timestep_mean": 527.46875,
    "epsilon_norm": 7.58311128616333,
    "prediction_norm": 7.118113040924072,
    "timestamp": "2025-07-14T12:35:12.736896"
  },
  {
    "epoch": 2749,
    "batch_idx": 64,
    "loss": 0.44197115302085876,
    "learning_rate": 0.0003079329068861496,
    "timestep_mean": 384.4375,
    "epsilon_norm": 7.487955093383789,
    "prediction_norm": 5.760039329528809,
    "timestamp": "2025-07-14T12:35:12.741795"
  },
  {
    "epoch": 2749,
    "batch_idx": 96,
    "loss": 0.32628756761550903,
    "learning_rate": 0.0003079329068861496,
    "timestep_mean": 482.78125,
    "epsilon_norm": 8.055049896240234,
    "prediction_norm": 6.848428249359131,
    "timestamp": "2025-07-14T12:35:12.746760"
  },
  {
    "epoch": 2749,
    "batch_idx": 128,
    "loss": 0.296376496553421,
    "learning_rate": 0.0003079329068861496,
    "timestep_mean": 500.0625,
    "epsilon_norm": 8.517050743103027,
    "prediction_norm": 6.721940040588379,
    "timestamp": "2025-07-14T12:35:12.751618"
  },
  {
    "epoch": 2749,
    "batch_idx": 160,
    "loss": 0.34239864349365234,
    "learning_rate": 0.0003079329068861496,
    "timestep_mean": 460.40625,
    "epsilon_norm": 8.383341789245605,
    "prediction_norm": 7.07834005355835,
    "timestamp": "2025-07-14T12:35:12.756454"
  },
  {
    "epoch": 2749,
    "batch_idx": 192,
    "loss": 0.16034972667694092,
    "learning_rate": 0.0003079329068861496,
    "timestep_mean": 450.875,
    "epsilon_norm": 3.7377846240997314,
    "prediction_norm": 3.4014201164245605,
    "timestamp": "2025-07-14T12:35:12.761318"
  },
  {
    "epoch": 2750,
    "batch_idx": 0,
    "loss": 0.22724470496177673,
    "learning_rate": 0.00030865828381745555,
    "timestep_mean": 538.40625,
    "epsilon_norm": 7.836726188659668,
    "prediction_norm": 6.274449348449707,
    "timestamp": "2025-07-14T12:35:12.766243"
  },
  {
    "epoch": 2750,
    "batch_idx": 32,
    "loss": 0.3223351836204529,
    "learning_rate": 0.00030865828381745555,
    "timestep_mean": 447.59375,
    "epsilon_norm": 8.205887794494629,
    "prediction_norm": 6.410923480987549,
    "timestamp": "2025-07-14T12:35:12.771158"
  },
  {
    "epoch": 2750,
    "batch_idx": 64,
    "loss": 0.24939094483852386,
    "learning_rate": 0.00030865828381745555,
    "timestep_mean": 496.28125,
    "epsilon_norm": 7.343498229980469,
    "prediction_norm": 6.492471694946289,
    "timestamp": "2025-07-14T12:35:12.776124"
  },
  {
    "epoch": 2750,
    "batch_idx": 96,
    "loss": 0.3411208987236023,
    "learning_rate": 0.00030865828381745555,
    "timestep_mean": 467.53125,
    "epsilon_norm": 8.58543586730957,
    "prediction_norm": 6.636263847351074,
    "timestamp": "2025-07-14T12:35:12.780958"
  },
  {
    "epoch": 2750,
    "batch_idx": 128,
    "loss": 0.5573896169662476,
    "learning_rate": 0.00030865828381745555,
    "timestep_mean": 532.0,
    "epsilon_norm": 8.258536338806152,
    "prediction_norm": 5.803524017333984,
    "timestamp": "2025-07-14T12:35:12.785793"
  },
  {
    "epoch": 2750,
    "batch_idx": 160,
    "loss": 0.38112878799438477,
    "learning_rate": 0.00030865828381745555,
    "timestep_mean": 504.09375,
    "epsilon_norm": 7.581249237060547,
    "prediction_norm": 6.141185283660889,
    "timestamp": "2025-07-14T12:35:12.790651"
  },
  {
    "epoch": 2750,
    "batch_idx": 192,
    "loss": 0.0909155085682869,
    "learning_rate": 0.00030865828381745555,
    "timestep_mean": 593.375,
    "epsilon_norm": 3.2906665802001953,
    "prediction_norm": 3.292249917984009,
    "timestamp": "2025-07-14T12:35:12.795509"
  },
  {
    "epoch": 2751,
    "batch_idx": 0,
    "loss": 0.28814297914505005,
    "learning_rate": 0.00030938413286542586,
    "timestep_mean": 426.21875,
    "epsilon_norm": 7.5619611740112305,
    "prediction_norm": 5.512753963470459,
    "timestamp": "2025-07-14T12:35:12.800518"
  },
  {
    "epoch": 2751,
    "batch_idx": 32,
    "loss": 0.34237778186798096,
    "learning_rate": 0.00030938413286542586,
    "timestep_mean": 458.40625,
    "epsilon_norm": 7.369487762451172,
    "prediction_norm": 6.1385273933410645,
    "timestamp": "2025-07-14T12:35:12.805458"
  },
  {
    "epoch": 2751,
    "batch_idx": 64,
    "loss": 0.3774062991142273,
    "learning_rate": 0.00030938413286542586,
    "timestep_mean": 439.75,
    "epsilon_norm": 8.350318908691406,
    "prediction_norm": 6.524415016174316,
    "timestamp": "2025-07-14T12:35:12.810343"
  },
  {
    "epoch": 2751,
    "batch_idx": 96,
    "loss": 0.2545405328273773,
    "learning_rate": 0.00030938413286542586,
    "timestep_mean": 501.25,
    "epsilon_norm": 8.307989120483398,
    "prediction_norm": 7.079388618469238,
    "timestamp": "2025-07-14T12:35:12.815171"
  },
  {
    "epoch": 2751,
    "batch_idx": 128,
    "loss": 0.3927995562553406,
    "learning_rate": 0.00030938413286542586,
    "timestep_mean": 416.5625,
    "epsilon_norm": 7.351868629455566,
    "prediction_norm": 5.386138916015625,
    "timestamp": "2025-07-14T12:35:12.820046"
  },
  {
    "epoch": 2751,
    "batch_idx": 160,
    "loss": 0.3697975277900696,
    "learning_rate": 0.00030938413286542586,
    "timestep_mean": 434.34375,
    "epsilon_norm": 9.201112747192383,
    "prediction_norm": 6.445219993591309,
    "timestamp": "2025-07-14T12:35:12.824951"
  },
  {
    "epoch": 2751,
    "batch_idx": 192,
    "loss": 0.16598035395145416,
    "learning_rate": 0.00030938413286542586,
    "timestep_mean": 730.5,
    "epsilon_norm": 4.609513759613037,
    "prediction_norm": 4.015619277954102,
    "timestamp": "2025-07-14T12:35:12.829902"
  },
  {
    "epoch": 2752,
    "batch_idx": 0,
    "loss": 0.3300776779651642,
    "learning_rate": 0.00031011045223910017,
    "timestep_mean": 508.59375,
    "epsilon_norm": 6.602243900299072,
    "prediction_norm": 6.307257175445557,
    "timestamp": "2025-07-14T12:35:12.834742"
  },
  {
    "epoch": 2752,
    "batch_idx": 32,
    "loss": 0.37588149309158325,
    "learning_rate": 0.00031011045223910017,
    "timestep_mean": 588.375,
    "epsilon_norm": 8.41149616241455,
    "prediction_norm": 6.9941020011901855,
    "timestamp": "2025-07-14T12:35:12.839626"
  },
  {
    "epoch": 2752,
    "batch_idx": 64,
    "loss": 0.3689539134502411,
    "learning_rate": 0.00031011045223910017,
    "timestep_mean": 438.6875,
    "epsilon_norm": 7.658833026885986,
    "prediction_norm": 5.692500591278076,
    "timestamp": "2025-07-14T12:35:12.844484"
  },
  {
    "epoch": 2752,
    "batch_idx": 96,
    "loss": 0.3168122470378876,
    "learning_rate": 0.00031011045223910017,
    "timestep_mean": 480.34375,
    "epsilon_norm": 6.940467357635498,
    "prediction_norm": 5.430583477020264,
    "timestamp": "2025-07-14T12:35:12.849351"
  },
  {
    "epoch": 2752,
    "batch_idx": 128,
    "loss": 0.24318717420101166,
    "learning_rate": 0.00031011045223910017,
    "timestep_mean": 532.09375,
    "epsilon_norm": 6.927077770233154,
    "prediction_norm": 5.7579240798950195,
    "timestamp": "2025-07-14T12:35:12.854197"
  },
  {
    "epoch": 2752,
    "batch_idx": 160,
    "loss": 0.41966766119003296,
    "learning_rate": 0.00031011045223910017,
    "timestep_mean": 416.46875,
    "epsilon_norm": 7.781075954437256,
    "prediction_norm": 5.7651472091674805,
    "timestamp": "2025-07-14T12:35:12.859380"
  },
  {
    "epoch": 2752,
    "batch_idx": 192,
    "loss": 0.1376832276582718,
    "learning_rate": 0.00031011045223910017,
    "timestep_mean": 530.875,
    "epsilon_norm": 5.024806976318359,
    "prediction_norm": 4.506110191345215,
    "timestamp": "2025-07-14T12:35:12.864309"
  },
  {
    "epoch": 2753,
    "batch_idx": 0,
    "loss": 0.14683902263641357,
    "learning_rate": 0.0003108372401463568,
    "timestep_mean": 518.53125,
    "epsilon_norm": 7.033002853393555,
    "prediction_norm": 6.06308126449585,
    "timestamp": "2025-07-14T12:35:12.869172"
  },
  {
    "epoch": 2753,
    "batch_idx": 32,
    "loss": 0.31499624252319336,
    "learning_rate": 0.0003108372401463568,
    "timestep_mean": 425.71875,
    "epsilon_norm": 7.740142345428467,
    "prediction_norm": 5.782537460327148,
    "timestamp": "2025-07-14T12:35:12.874040"
  },
  {
    "epoch": 2753,
    "batch_idx": 64,
    "loss": 0.20076116919517517,
    "learning_rate": 0.0003108372401463568,
    "timestep_mean": 492.15625,
    "epsilon_norm": 7.579403877258301,
    "prediction_norm": 6.119564056396484,
    "timestamp": "2025-07-14T12:35:12.878933"
  },
  {
    "epoch": 2753,
    "batch_idx": 96,
    "loss": 0.3354671895503998,
    "learning_rate": 0.0003108372401463568,
    "timestep_mean": 468.40625,
    "epsilon_norm": 8.347467422485352,
    "prediction_norm": 6.84245491027832,
    "timestamp": "2025-07-14T12:35:12.883905"
  },
  {
    "epoch": 2753,
    "batch_idx": 128,
    "loss": 0.3492346405982971,
    "learning_rate": 0.0003108372401463568,
    "timestep_mean": 482.5,
    "epsilon_norm": 10.845267295837402,
    "prediction_norm": 9.679200172424316,
    "timestamp": "2025-07-14T12:35:12.888857"
  },
  {
    "epoch": 2753,
    "batch_idx": 160,
    "loss": 0.38079318404197693,
    "learning_rate": 0.0003108372401463568,
    "timestep_mean": 476.5625,
    "epsilon_norm": 8.080018043518066,
    "prediction_norm": 6.606443881988525,
    "timestamp": "2025-07-14T12:35:12.893773"
  },
  {
    "epoch": 2753,
    "batch_idx": 192,
    "loss": 0.25842809677124023,
    "learning_rate": 0.0003108372401463568,
    "timestep_mean": 413.125,
    "epsilon_norm": 3.369260311126709,
    "prediction_norm": 2.1092114448547363,
    "timestamp": "2025-07-14T12:35:12.898694"
  },
  {
    "epoch": 2754,
    "batch_idx": 0,
    "loss": 0.3126519024372101,
    "learning_rate": 0.00031156449479391923,
    "timestep_mean": 471.53125,
    "epsilon_norm": 8.230133056640625,
    "prediction_norm": 6.558040618896484,
    "timestamp": "2025-07-14T12:35:12.903616"
  },
  {
    "epoch": 2754,
    "batch_idx": 32,
    "loss": 0.2480936348438263,
    "learning_rate": 0.00031156449479391923,
    "timestep_mean": 454.65625,
    "epsilon_norm": 7.686103820800781,
    "prediction_norm": 6.661493301391602,
    "timestamp": "2025-07-14T12:35:12.908481"
  },
  {
    "epoch": 2754,
    "batch_idx": 64,
    "loss": 0.1774633526802063,
    "learning_rate": 0.00031156449479391923,
    "timestep_mean": 524.21875,
    "epsilon_norm": 8.411256790161133,
    "prediction_norm": 8.276110649108887,
    "timestamp": "2025-07-14T12:35:12.913427"
  },
  {
    "epoch": 2754,
    "batch_idx": 96,
    "loss": 0.2568370997905731,
    "learning_rate": 0.00031156449479391923,
    "timestep_mean": 589.21875,
    "epsilon_norm": 7.747182369232178,
    "prediction_norm": 6.931166172027588,
    "timestamp": "2025-07-14T12:35:12.918265"
  },
  {
    "epoch": 2754,
    "batch_idx": 128,
    "loss": 0.283816933631897,
    "learning_rate": 0.00031156449479391923,
    "timestep_mean": 478.59375,
    "epsilon_norm": 8.728599548339844,
    "prediction_norm": 8.512173652648926,
    "timestamp": "2025-07-14T12:35:12.923202"
  },
  {
    "epoch": 2754,
    "batch_idx": 160,
    "loss": 0.09860093891620636,
    "learning_rate": 0.00031156449479391923,
    "timestep_mean": 534.25,
    "epsilon_norm": 7.210836410522461,
    "prediction_norm": 7.048659801483154,
    "timestamp": "2025-07-14T12:35:12.928065"
  },
  {
    "epoch": 2754,
    "batch_idx": 192,
    "loss": 0.3490939438343048,
    "learning_rate": 0.00031156449479391923,
    "timestep_mean": 484.375,
    "epsilon_norm": 3.2834692001342773,
    "prediction_norm": 2.6995630264282227,
    "timestamp": "2025-07-14T12:35:12.932933"
  },
  {
    "epoch": 2755,
    "batch_idx": 0,
    "loss": 0.13716626167297363,
    "learning_rate": 0.00031229221438735903,
    "timestep_mean": 492.78125,
    "epsilon_norm": 7.3626179695129395,
    "prediction_norm": 6.500908374786377,
    "timestamp": "2025-07-14T12:35:12.937942"
  },
  {
    "epoch": 2755,
    "batch_idx": 32,
    "loss": 0.19476768374443054,
    "learning_rate": 0.00031229221438735903,
    "timestep_mean": 473.84375,
    "epsilon_norm": 6.6991143226623535,
    "prediction_norm": 5.765030860900879,
    "timestamp": "2025-07-14T12:35:12.942845"
  },
  {
    "epoch": 2755,
    "batch_idx": 64,
    "loss": 0.35189419984817505,
    "learning_rate": 0.00031229221438735903,
    "timestep_mean": 521.5625,
    "epsilon_norm": 9.413869857788086,
    "prediction_norm": 8.065401077270508,
    "timestamp": "2025-07-14T12:35:12.947714"
  },
  {
    "epoch": 2755,
    "batch_idx": 96,
    "loss": 0.3006228506565094,
    "learning_rate": 0.00031229221438735903,
    "timestep_mean": 462.28125,
    "epsilon_norm": 8.752216339111328,
    "prediction_norm": 7.098045349121094,
    "timestamp": "2025-07-14T12:35:12.952577"
  },
  {
    "epoch": 2755,
    "batch_idx": 128,
    "loss": 0.1908474862575531,
    "learning_rate": 0.00031229221438735903,
    "timestep_mean": 619.4375,
    "epsilon_norm": 7.168038845062256,
    "prediction_norm": 6.5320587158203125,
    "timestamp": "2025-07-14T12:35:12.957521"
  },
  {
    "epoch": 2755,
    "batch_idx": 160,
    "loss": 0.14101195335388184,
    "learning_rate": 0.00031229221438735903,
    "timestep_mean": 524.40625,
    "epsilon_norm": 6.996711730957031,
    "prediction_norm": 6.357548236846924,
    "timestamp": "2025-07-14T12:35:12.962384"
  },
  {
    "epoch": 2755,
    "batch_idx": 192,
    "loss": 0.5210199356079102,
    "learning_rate": 0.00031229221438735903,
    "timestep_mean": 331.375,
    "epsilon_norm": 3.919367551803589,
    "prediction_norm": 2.7786924839019775,
    "timestamp": "2025-07-14T12:35:12.967396"
  },
  {
    "epoch": 2756,
    "batch_idx": 0,
    "loss": 0.4371401071548462,
    "learning_rate": 0.0003130203971311003,
    "timestep_mean": 436.65625,
    "epsilon_norm": 8.596152305603027,
    "prediction_norm": 6.781727313995361,
    "timestamp": "2025-07-14T12:35:12.972284"
  },
  {
    "epoch": 2756,
    "batch_idx": 32,
    "loss": 0.23771031200885773,
    "learning_rate": 0.0003130203971311003,
    "timestep_mean": 553.59375,
    "epsilon_norm": 8.189454078674316,
    "prediction_norm": 6.810307502746582,
    "timestamp": "2025-07-14T12:35:12.977165"
  },
  {
    "epoch": 2756,
    "batch_idx": 64,
    "loss": 0.2434241771697998,
    "learning_rate": 0.0003130203971311003,
    "timestep_mean": 497.03125,
    "epsilon_norm": 7.655543804168701,
    "prediction_norm": 6.3544158935546875,
    "timestamp": "2025-07-14T12:35:12.982008"
  },
  {
    "epoch": 2756,
    "batch_idx": 96,
    "loss": 0.15534693002700806,
    "learning_rate": 0.0003130203971311003,
    "timestep_mean": 538.65625,
    "epsilon_norm": 7.488132476806641,
    "prediction_norm": 7.035815238952637,
    "timestamp": "2025-07-14T12:35:12.986849"
  },
  {
    "epoch": 2756,
    "batch_idx": 128,
    "loss": 0.24901628494262695,
    "learning_rate": 0.0003130203971311003,
    "timestep_mean": 471.71875,
    "epsilon_norm": 7.8056111335754395,
    "prediction_norm": 6.928155422210693,
    "timestamp": "2025-07-14T12:35:12.991757"
  },
  {
    "epoch": 2756,
    "batch_idx": 160,
    "loss": 0.3125688433647156,
    "learning_rate": 0.0003130203971311003,
    "timestep_mean": 398.6875,
    "epsilon_norm": 7.605166912078857,
    "prediction_norm": 6.30134391784668,
    "timestamp": "2025-07-14T12:35:12.996738"
  },
  {
    "epoch": 2756,
    "batch_idx": 192,
    "loss": 0.5639404654502869,
    "learning_rate": 0.0003130203971311003,
    "timestep_mean": 381.75,
    "epsilon_norm": 4.707180976867676,
    "prediction_norm": 3.256667137145996,
    "timestamp": "2025-07-14T12:35:13.001562"
  },
  {
    "epoch": 2757,
    "batch_idx": 0,
    "loss": 0.23050788044929504,
    "learning_rate": 0.0003137490412284246,
    "timestep_mean": 465.75,
    "epsilon_norm": 8.21338939666748,
    "prediction_norm": 7.010676383972168,
    "timestamp": "2025-07-14T12:35:13.006446"
  },
  {
    "epoch": 2757,
    "batch_idx": 32,
    "loss": 0.39023011922836304,
    "learning_rate": 0.0003137490412284246,
    "timestep_mean": 589.375,
    "epsilon_norm": 7.756778717041016,
    "prediction_norm": 5.579031944274902,
    "timestamp": "2025-07-14T12:35:13.011318"
  },
  {
    "epoch": 2757,
    "batch_idx": 64,
    "loss": 0.18116918206214905,
    "learning_rate": 0.0003137490412284246,
    "timestep_mean": 487.34375,
    "epsilon_norm": 8.65919303894043,
    "prediction_norm": 7.878998756408691,
    "timestamp": "2025-07-14T12:35:13.016307"
  },
  {
    "epoch": 2757,
    "batch_idx": 96,
    "loss": 0.375845342874527,
    "learning_rate": 0.0003137490412284246,
    "timestep_mean": 447.03125,
    "epsilon_norm": 8.451910972595215,
    "prediction_norm": 7.032200813293457,
    "timestamp": "2025-07-14T12:35:13.021201"
  },
  {
    "epoch": 2757,
    "batch_idx": 128,
    "loss": 0.08012647181749344,
    "learning_rate": 0.0003137490412284246,
    "timestep_mean": 644.71875,
    "epsilon_norm": 6.508755683898926,
    "prediction_norm": 6.426044940948486,
    "timestamp": "2025-07-14T12:35:13.026140"
  },
  {
    "epoch": 2757,
    "batch_idx": 160,
    "loss": 0.14752721786499023,
    "learning_rate": 0.0003137490412284246,
    "timestep_mean": 514.625,
    "epsilon_norm": 7.449428558349609,
    "prediction_norm": 6.781367778778076,
    "timestamp": "2025-07-14T12:35:13.031164"
  },
  {
    "epoch": 2757,
    "batch_idx": 192,
    "loss": 0.0908515676856041,
    "learning_rate": 0.0003137490412284246,
    "timestep_mean": 566.125,
    "epsilon_norm": 4.378721237182617,
    "prediction_norm": 3.978148937225342,
    "timestamp": "2025-07-14T12:35:13.042140"
  },
  {
    "epoch": 2758,
    "batch_idx": 0,
    "loss": 0.36141473054885864,
    "learning_rate": 0.000314478144881475,
    "timestep_mean": 492.28125,
    "epsilon_norm": 6.786043643951416,
    "prediction_norm": 5.390735149383545,
    "timestamp": "2025-07-14T12:35:13.047905"
  },
  {
    "epoch": 2758,
    "batch_idx": 32,
    "loss": 0.48263800144195557,
    "learning_rate": 0.000314478144881475,
    "timestep_mean": 573.0625,
    "epsilon_norm": 8.728178024291992,
    "prediction_norm": 7.508421897888184,
    "timestamp": "2025-07-14T12:35:13.052735"
  },
  {
    "epoch": 2758,
    "batch_idx": 64,
    "loss": 0.35097259283065796,
    "learning_rate": 0.000314478144881475,
    "timestep_mean": 496.84375,
    "epsilon_norm": 9.611186027526855,
    "prediction_norm": 8.01053524017334,
    "timestamp": "2025-07-14T12:35:13.057681"
  },
  {
    "epoch": 2758,
    "batch_idx": 96,
    "loss": 0.14171777665615082,
    "learning_rate": 0.000314478144881475,
    "timestep_mean": 585.375,
    "epsilon_norm": 7.653998374938965,
    "prediction_norm": 7.4066033363342285,
    "timestamp": "2025-07-14T12:35:13.062814"
  },
  {
    "epoch": 2758,
    "batch_idx": 128,
    "loss": 0.5196382999420166,
    "learning_rate": 0.000314478144881475,
    "timestep_mean": 429.9375,
    "epsilon_norm": 8.294146537780762,
    "prediction_norm": 6.033381938934326,
    "timestamp": "2025-07-14T12:35:13.067931"
  },
  {
    "epoch": 2758,
    "batch_idx": 160,
    "loss": 0.2954971492290497,
    "learning_rate": 0.000314478144881475,
    "timestep_mean": 474.4375,
    "epsilon_norm": 7.307246685028076,
    "prediction_norm": 6.071164131164551,
    "timestamp": "2025-07-14T12:35:13.072919"
  },
  {
    "epoch": 2758,
    "batch_idx": 192,
    "loss": 0.091923788189888,
    "learning_rate": 0.000314478144881475,
    "timestep_mean": 475.875,
    "epsilon_norm": 3.856335163116455,
    "prediction_norm": 3.902751922607422,
    "timestamp": "2025-07-14T12:35:13.077922"
  },
  {
    "epoch": 2759,
    "batch_idx": 0,
    "loss": 0.35116589069366455,
    "learning_rate": 0.0003152077062912611,
    "timestep_mean": 431.90625,
    "epsilon_norm": 7.7659687995910645,
    "prediction_norm": 6.40423059463501,
    "timestamp": "2025-07-14T12:35:13.082923"
  },
  {
    "epoch": 2759,
    "batch_idx": 32,
    "loss": 0.43127644062042236,
    "learning_rate": 0.0003152077062912611,
    "timestep_mean": 479.125,
    "epsilon_norm": 7.382046222686768,
    "prediction_norm": 5.970429420471191,
    "timestamp": "2025-07-14T12:35:13.087950"
  },
  {
    "epoch": 2759,
    "batch_idx": 64,
    "loss": 0.30970513820648193,
    "learning_rate": 0.0003152077062912611,
    "timestep_mean": 500.03125,
    "epsilon_norm": 7.952265739440918,
    "prediction_norm": 6.781362533569336,
    "timestamp": "2025-07-14T12:35:13.092851"
  },
  {
    "epoch": 2759,
    "batch_idx": 96,
    "loss": 0.18537136912345886,
    "learning_rate": 0.0003152077062912611,
    "timestep_mean": 510.40625,
    "epsilon_norm": 7.320211410522461,
    "prediction_norm": 6.64670467376709,
    "timestamp": "2025-07-14T12:35:13.097795"
  },
  {
    "epoch": 2759,
    "batch_idx": 128,
    "loss": 0.26428747177124023,
    "learning_rate": 0.0003152077062912611,
    "timestep_mean": 476.71875,
    "epsilon_norm": 7.274385452270508,
    "prediction_norm": 5.773372173309326,
    "timestamp": "2025-07-14T12:35:13.102883"
  },
  {
    "epoch": 2759,
    "batch_idx": 160,
    "loss": 0.4658719599246979,
    "learning_rate": 0.0003152077062912611,
    "timestep_mean": 474.875,
    "epsilon_norm": 9.24948501586914,
    "prediction_norm": 6.964636325836182,
    "timestamp": "2025-07-14T12:35:13.107815"
  },
  {
    "epoch": 2759,
    "batch_idx": 192,
    "loss": 0.4611102044582367,
    "learning_rate": 0.0003152077062912611,
    "timestep_mean": 512.0,
    "epsilon_norm": 4.353046894073486,
    "prediction_norm": 3.8352932929992676,
    "timestamp": "2025-07-14T12:35:13.112668"
  },
  {
    "epoch": 2760,
    "batch_idx": 0,
    "loss": 0.2872786521911621,
    "learning_rate": 0.0003159377236576618,
    "timestep_mean": 478.4375,
    "epsilon_norm": 7.252808570861816,
    "prediction_norm": 5.4159159660339355,
    "timestamp": "2025-07-14T12:35:13.117723"
  },
  {
    "epoch": 2760,
    "batch_idx": 32,
    "loss": 0.1351032555103302,
    "learning_rate": 0.0003159377236576618,
    "timestep_mean": 488.59375,
    "epsilon_norm": 6.504266262054443,
    "prediction_norm": 5.888336181640625,
    "timestamp": "2025-07-14T12:35:13.122673"
  },
  {
    "epoch": 2760,
    "batch_idx": 64,
    "loss": 0.23206350207328796,
    "learning_rate": 0.0003159377236576618,
    "timestep_mean": 517.5,
    "epsilon_norm": 8.508328437805176,
    "prediction_norm": 7.272620677947998,
    "timestamp": "2025-07-14T12:35:13.127704"
  },
  {
    "epoch": 2760,
    "batch_idx": 96,
    "loss": 0.6931031942367554,
    "learning_rate": 0.0003159377236576618,
    "timestep_mean": 511.5625,
    "epsilon_norm": 8.903188705444336,
    "prediction_norm": 6.41179895401001,
    "timestamp": "2025-07-14T12:35:13.132679"
  },
  {
    "epoch": 2760,
    "batch_idx": 128,
    "loss": 0.25914424657821655,
    "learning_rate": 0.0003159377236576618,
    "timestep_mean": 549.53125,
    "epsilon_norm": 7.2941131591796875,
    "prediction_norm": 6.495093822479248,
    "timestamp": "2025-07-14T12:35:13.137675"
  },
  {
    "epoch": 2760,
    "batch_idx": 160,
    "loss": 0.10609057545661926,
    "learning_rate": 0.0003159377236576618,
    "timestep_mean": 597.125,
    "epsilon_norm": 7.324337959289551,
    "prediction_norm": 6.277848243713379,
    "timestamp": "2025-07-14T12:35:13.142865"
  },
  {
    "epoch": 2760,
    "batch_idx": 192,
    "loss": 0.23433950543403625,
    "learning_rate": 0.0003159377236576618,
    "timestep_mean": 508.75,
    "epsilon_norm": 3.9247305393218994,
    "prediction_norm": 3.7669758796691895,
    "timestamp": "2025-07-14T12:35:13.147786"
  },
  {
    "epoch": 2761,
    "batch_idx": 0,
    "loss": 0.7320655584335327,
    "learning_rate": 0.00031666819517943195,
    "timestep_mean": 408.21875,
    "epsilon_norm": 9.049976348876953,
    "prediction_norm": 6.25331974029541,
    "timestamp": "2025-07-14T12:35:13.152819"
  },
  {
    "epoch": 2761,
    "batch_idx": 32,
    "loss": 0.38162899017333984,
    "learning_rate": 0.00031666819517943195,
    "timestep_mean": 475.9375,
    "epsilon_norm": 8.229119300842285,
    "prediction_norm": 5.737638473510742,
    "timestamp": "2025-07-14T12:35:13.157888"
  },
  {
    "epoch": 2761,
    "batch_idx": 64,
    "loss": 0.3480455279350281,
    "learning_rate": 0.00031666819517943195,
    "timestep_mean": 462.3125,
    "epsilon_norm": 8.378713607788086,
    "prediction_norm": 6.436763763427734,
    "timestamp": "2025-07-14T12:35:13.162946"
  },
  {
    "epoch": 2761,
    "batch_idx": 96,
    "loss": 0.20565658807754517,
    "learning_rate": 0.00031666819517943195,
    "timestep_mean": 580.125,
    "epsilon_norm": 7.670192241668701,
    "prediction_norm": 6.17879056930542,
    "timestamp": "2025-07-14T12:35:13.168044"
  },
  {
    "epoch": 2761,
    "batch_idx": 128,
    "loss": 0.32374075055122375,
    "learning_rate": 0.00031666819517943195,
    "timestep_mean": 543.0625,
    "epsilon_norm": 7.5427703857421875,
    "prediction_norm": 6.223675727844238,
    "timestamp": "2025-07-14T12:35:13.172909"
  },
  {
    "epoch": 2761,
    "batch_idx": 160,
    "loss": 0.3904052972793579,
    "learning_rate": 0.00031666819517943195,
    "timestep_mean": 499.8125,
    "epsilon_norm": 8.119255065917969,
    "prediction_norm": 6.286143779754639,
    "timestamp": "2025-07-14T12:35:13.177862"
  },
  {
    "epoch": 2761,
    "batch_idx": 192,
    "loss": 0.41716617345809937,
    "learning_rate": 0.00031666819517943195,
    "timestep_mean": 408.0,
    "epsilon_norm": 4.173270225524902,
    "prediction_norm": 2.926262378692627,
    "timestamp": "2025-07-14T12:35:13.183059"
  },
  {
    "epoch": 2762,
    "batch_idx": 0,
    "loss": 0.23882345855236053,
    "learning_rate": 0.00031739911905420676,
    "timestep_mean": 492.375,
    "epsilon_norm": 8.520673751831055,
    "prediction_norm": 7.157929420471191,
    "timestamp": "2025-07-14T12:35:13.188096"
  },
  {
    "epoch": 2762,
    "batch_idx": 32,
    "loss": 0.3028966188430786,
    "learning_rate": 0.00031739911905420676,
    "timestep_mean": 400.78125,
    "epsilon_norm": 7.540225982666016,
    "prediction_norm": 5.133526802062988,
    "timestamp": "2025-07-14T12:35:13.193117"
  },
  {
    "epoch": 2762,
    "batch_idx": 64,
    "loss": 0.6919517517089844,
    "learning_rate": 0.00031739911905420676,
    "timestep_mean": 474.34375,
    "epsilon_norm": 9.022639274597168,
    "prediction_norm": 5.65972900390625,
    "timestamp": "2025-07-14T12:35:13.198346"
  },
  {
    "epoch": 2762,
    "batch_idx": 96,
    "loss": 0.26963523030281067,
    "learning_rate": 0.00031739911905420676,
    "timestep_mean": 491.5,
    "epsilon_norm": 8.706096649169922,
    "prediction_norm": 7.682635307312012,
    "timestamp": "2025-07-14T12:35:13.203526"
  },
  {
    "epoch": 2762,
    "batch_idx": 128,
    "loss": 0.2577434182167053,
    "learning_rate": 0.00031739911905420676,
    "timestep_mean": 471.125,
    "epsilon_norm": 7.311307907104492,
    "prediction_norm": 6.944314002990723,
    "timestamp": "2025-07-14T12:35:13.208809"
  },
  {
    "epoch": 2762,
    "batch_idx": 160,
    "loss": 0.3044002652168274,
    "learning_rate": 0.00031739911905420676,
    "timestep_mean": 376.90625,
    "epsilon_norm": 7.057257175445557,
    "prediction_norm": 6.144604206085205,
    "timestamp": "2025-07-14T12:35:13.214030"
  },
  {
    "epoch": 2762,
    "batch_idx": 192,
    "loss": 0.0736047700047493,
    "learning_rate": 0.00031739911905420676,
    "timestep_mean": 795.0,
    "epsilon_norm": 2.739034414291382,
    "prediction_norm": 2.8074769973754883,
    "timestamp": "2025-07-14T12:35:13.219114"
  },
  {
    "epoch": 2763,
    "batch_idx": 0,
    "loss": 0.28619903326034546,
    "learning_rate": 0.00031813049347850314,
    "timestep_mean": 537.84375,
    "epsilon_norm": 7.988597393035889,
    "prediction_norm": 7.345550537109375,
    "timestamp": "2025-07-14T12:35:13.224186"
  },
  {
    "epoch": 2763,
    "batch_idx": 32,
    "loss": 0.2745306193828583,
    "learning_rate": 0.00031813049347850314,
    "timestep_mean": 416.03125,
    "epsilon_norm": 8.413114547729492,
    "prediction_norm": 7.137787342071533,
    "timestamp": "2025-07-14T12:35:13.229343"
  },
  {
    "epoch": 2763,
    "batch_idx": 64,
    "loss": 0.28991633653640747,
    "learning_rate": 0.00031813049347850314,
    "timestep_mean": 397.53125,
    "epsilon_norm": 7.623501300811768,
    "prediction_norm": 7.11516809463501,
    "timestamp": "2025-07-14T12:35:13.234769"
  },
  {
    "epoch": 2763,
    "batch_idx": 96,
    "loss": 0.1419242024421692,
    "learning_rate": 0.00031813049347850314,
    "timestep_mean": 489.09375,
    "epsilon_norm": 6.703536510467529,
    "prediction_norm": 6.637431621551514,
    "timestamp": "2025-07-14T12:35:13.239894"
  },
  {
    "epoch": 2763,
    "batch_idx": 128,
    "loss": 0.2975730001926422,
    "learning_rate": 0.00031813049347850314,
    "timestep_mean": 546.1875,
    "epsilon_norm": 7.838454246520996,
    "prediction_norm": 6.195285797119141,
    "timestamp": "2025-07-14T12:35:13.245122"
  },
  {
    "epoch": 2763,
    "batch_idx": 160,
    "loss": 0.2616068720817566,
    "learning_rate": 0.00031813049347850314,
    "timestep_mean": 463.875,
    "epsilon_norm": 6.892326354980469,
    "prediction_norm": 6.58663272857666,
    "timestamp": "2025-07-14T12:35:13.250418"
  },
  {
    "epoch": 2763,
    "batch_idx": 192,
    "loss": 0.1269448846578598,
    "learning_rate": 0.00031813049347850314,
    "timestep_mean": 492.375,
    "epsilon_norm": 3.823793411254883,
    "prediction_norm": 3.299004554748535,
    "timestamp": "2025-07-14T12:35:13.255614"
  },
  {
    "epoch": 2764,
    "batch_idx": 0,
    "loss": 0.2724343538284302,
    "learning_rate": 0.00031886231664772773,
    "timestep_mean": 435.28125,
    "epsilon_norm": 9.367609024047852,
    "prediction_norm": 8.383609771728516,
    "timestamp": "2025-07-14T12:35:13.260876"
  },
  {
    "epoch": 2764,
    "batch_idx": 32,
    "loss": 0.17173704504966736,
    "learning_rate": 0.00031886231664772773,
    "timestep_mean": 475.5,
    "epsilon_norm": 7.813992023468018,
    "prediction_norm": 6.558722972869873,
    "timestamp": "2025-07-14T12:35:13.265967"
  },
  {
    "epoch": 2764,
    "batch_idx": 64,
    "loss": 0.25704145431518555,
    "learning_rate": 0.00031886231664772773,
    "timestep_mean": 465.40625,
    "epsilon_norm": 8.138840675354004,
    "prediction_norm": 6.940722942352295,
    "timestamp": "2025-07-14T12:35:13.271082"
  },
  {
    "epoch": 2764,
    "batch_idx": 96,
    "loss": 0.2348266839981079,
    "learning_rate": 0.00031886231664772773,
    "timestep_mean": 472.0625,
    "epsilon_norm": 7.2081475257873535,
    "prediction_norm": 5.945897579193115,
    "timestamp": "2025-07-14T12:35:13.276230"
  },
  {
    "epoch": 2764,
    "batch_idx": 128,
    "loss": 0.4747436046600342,
    "learning_rate": 0.00031886231664772773,
    "timestep_mean": 486.1875,
    "epsilon_norm": 8.104340553283691,
    "prediction_norm": 6.698108196258545,
    "timestamp": "2025-07-14T12:35:13.281348"
  },
  {
    "epoch": 2764,
    "batch_idx": 160,
    "loss": 0.3248756229877472,
    "learning_rate": 0.00031886231664772773,
    "timestep_mean": 512.46875,
    "epsilon_norm": 7.841084957122803,
    "prediction_norm": 6.369897365570068,
    "timestamp": "2025-07-14T12:35:13.286508"
  },
  {
    "epoch": 2764,
    "batch_idx": 192,
    "loss": 0.20507358014583588,
    "learning_rate": 0.00031886231664772773,
    "timestep_mean": 544.375,
    "epsilon_norm": 4.262502670288086,
    "prediction_norm": 4.215811252593994,
    "timestamp": "2025-07-14T12:35:13.291658"
  },
  {
    "epoch": 2765,
    "batch_idx": 0,
    "loss": 0.21534624695777893,
    "learning_rate": 0.0003195945867561796,
    "timestep_mean": 505.65625,
    "epsilon_norm": 7.132114410400391,
    "prediction_norm": 6.727038383483887,
    "timestamp": "2025-07-14T12:35:13.296886"
  },
  {
    "epoch": 2765,
    "batch_idx": 32,
    "loss": 0.2104208767414093,
    "learning_rate": 0.0003195945867561796,
    "timestep_mean": 408.375,
    "epsilon_norm": 6.569014549255371,
    "prediction_norm": 5.601286888122559,
    "timestamp": "2025-07-14T12:35:13.302158"
  },
  {
    "epoch": 2765,
    "batch_idx": 64,
    "loss": 0.5462130308151245,
    "learning_rate": 0.0003195945867561796,
    "timestep_mean": 462.125,
    "epsilon_norm": 8.386982917785645,
    "prediction_norm": 6.429222106933594,
    "timestamp": "2025-07-14T12:35:13.307358"
  },
  {
    "epoch": 2765,
    "batch_idx": 96,
    "loss": 0.26365455985069275,
    "learning_rate": 0.0003195945867561796,
    "timestep_mean": 527.125,
    "epsilon_norm": 7.950994491577148,
    "prediction_norm": 6.327492713928223,
    "timestamp": "2025-07-14T12:35:13.312517"
  },
  {
    "epoch": 2765,
    "batch_idx": 128,
    "loss": 0.17442548274993896,
    "learning_rate": 0.0003195945867561796,
    "timestep_mean": 597.96875,
    "epsilon_norm": 8.835627555847168,
    "prediction_norm": 7.6095709800720215,
    "timestamp": "2025-07-14T12:35:13.317789"
  },
  {
    "epoch": 2765,
    "batch_idx": 160,
    "loss": 0.31689363718032837,
    "learning_rate": 0.0003195945867561796,
    "timestep_mean": 486.0625,
    "epsilon_norm": 7.633610248565674,
    "prediction_norm": 5.442458629608154,
    "timestamp": "2025-07-14T12:35:13.322858"
  },
  {
    "epoch": 2765,
    "batch_idx": 192,
    "loss": 0.17893236875534058,
    "learning_rate": 0.0003195945867561796,
    "timestep_mean": 663.5,
    "epsilon_norm": 3.3755600452423096,
    "prediction_norm": 2.5965185165405273,
    "timestamp": "2025-07-14T12:35:13.327970"
  },
  {
    "epoch": 2766,
    "batch_idx": 0,
    "loss": 0.33954086899757385,
    "learning_rate": 0.0003203273019970555,
    "timestep_mean": 451.8125,
    "epsilon_norm": 8.962263107299805,
    "prediction_norm": 6.9159722328186035,
    "timestamp": "2025-07-14T12:35:13.333132"
  },
  {
    "epoch": 2766,
    "batch_idx": 32,
    "loss": 0.19613564014434814,
    "learning_rate": 0.0003203273019970555,
    "timestep_mean": 480.9375,
    "epsilon_norm": 7.898357391357422,
    "prediction_norm": 6.648434638977051,
    "timestamp": "2025-07-14T12:35:13.338427"
  },
  {
    "epoch": 2766,
    "batch_idx": 64,
    "loss": 0.24319273233413696,
    "learning_rate": 0.0003203273019970555,
    "timestep_mean": 481.40625,
    "epsilon_norm": 8.362334251403809,
    "prediction_norm": 7.470075607299805,
    "timestamp": "2025-07-14T12:35:13.343532"
  },
  {
    "epoch": 2766,
    "batch_idx": 96,
    "loss": 0.2855517864227295,
    "learning_rate": 0.0003203273019970555,
    "timestep_mean": 436.625,
    "epsilon_norm": 7.662215709686279,
    "prediction_norm": 6.191619873046875,
    "timestamp": "2025-07-14T12:35:13.348656"
  },
  {
    "epoch": 2766,
    "batch_idx": 128,
    "loss": 0.1768498420715332,
    "learning_rate": 0.0003203273019970555,
    "timestep_mean": 576.03125,
    "epsilon_norm": 6.6650776863098145,
    "prediction_norm": 5.281374454498291,
    "timestamp": "2025-07-14T12:35:13.353835"
  },
  {
    "epoch": 2766,
    "batch_idx": 160,
    "loss": 0.21274113655090332,
    "learning_rate": 0.0003203273019970555,
    "timestep_mean": 601.5,
    "epsilon_norm": 6.917214393615723,
    "prediction_norm": 6.026802062988281,
    "timestamp": "2025-07-14T12:35:13.359173"
  },
  {
    "epoch": 2766,
    "batch_idx": 192,
    "loss": 0.16972656548023224,
    "learning_rate": 0.0003203273019970555,
    "timestep_mean": 558.625,
    "epsilon_norm": 4.954835414886475,
    "prediction_norm": 4.2292914390563965,
    "timestamp": "2025-07-14T12:35:13.364227"
  },
  {
    "epoch": 2767,
    "batch_idx": 0,
    "loss": 0.48848026990890503,
    "learning_rate": 0.0003210604605624522,
    "timestep_mean": 539.5,
    "epsilon_norm": 9.29389476776123,
    "prediction_norm": 7.0257110595703125,
    "timestamp": "2025-07-14T12:35:13.369410"
  },
  {
    "epoch": 2767,
    "batch_idx": 32,
    "loss": 0.33916911482810974,
    "learning_rate": 0.0003210604605624522,
    "timestep_mean": 453.5,
    "epsilon_norm": 8.255023002624512,
    "prediction_norm": 6.692729949951172,
    "timestamp": "2025-07-14T12:35:13.374619"
  },
  {
    "epoch": 2767,
    "batch_idx": 64,
    "loss": 0.3238840699195862,
    "learning_rate": 0.0003210604605624522,
    "timestep_mean": 458.25,
    "epsilon_norm": 8.293601989746094,
    "prediction_norm": 6.696402072906494,
    "timestamp": "2025-07-14T12:35:13.379683"
  },
  {
    "epoch": 2767,
    "batch_idx": 96,
    "loss": 0.16797006130218506,
    "learning_rate": 0.0003210604605624522,
    "timestep_mean": 461.0625,
    "epsilon_norm": 8.329134941101074,
    "prediction_norm": 7.121342658996582,
    "timestamp": "2025-07-14T12:35:13.384898"
  },
  {
    "epoch": 2767,
    "batch_idx": 128,
    "loss": 0.16260655224323273,
    "learning_rate": 0.0003210604605624522,
    "timestep_mean": 560.0,
    "epsilon_norm": 7.933104038238525,
    "prediction_norm": 7.14213228225708,
    "timestamp": "2025-07-14T12:35:13.390044"
  },
  {
    "epoch": 2767,
    "batch_idx": 160,
    "loss": 0.31708022952079773,
    "learning_rate": 0.0003210604605624522,
    "timestep_mean": 483.65625,
    "epsilon_norm": 7.85885763168335,
    "prediction_norm": 7.257058620452881,
    "timestamp": "2025-07-14T12:35:13.395125"
  },
  {
    "epoch": 2767,
    "batch_idx": 192,
    "loss": 0.09057298302650452,
    "learning_rate": 0.0003210604605624522,
    "timestep_mean": 577.75,
    "epsilon_norm": 3.185239315032959,
    "prediction_norm": 3.1142263412475586,
    "timestamp": "2025-07-14T12:35:13.400388"
  },
  {
    "epoch": 2768,
    "batch_idx": 0,
    "loss": 0.2767346203327179,
    "learning_rate": 0.0003217940606433749,
    "timestep_mean": 456.0,
    "epsilon_norm": 7.481863498687744,
    "prediction_norm": 6.092569828033447,
    "timestamp": "2025-07-14T12:35:13.405873"
  },
  {
    "epoch": 2768,
    "batch_idx": 32,
    "loss": 0.2285967469215393,
    "learning_rate": 0.0003217940606433749,
    "timestep_mean": 474.09375,
    "epsilon_norm": 8.324151992797852,
    "prediction_norm": 6.905276298522949,
    "timestamp": "2025-07-14T12:35:13.411306"
  },
  {
    "epoch": 2768,
    "batch_idx": 64,
    "loss": 0.7529040575027466,
    "learning_rate": 0.0003217940606433749,
    "timestep_mean": 481.21875,
    "epsilon_norm": 8.519428253173828,
    "prediction_norm": 5.1868486404418945,
    "timestamp": "2025-07-14T12:35:13.417243"
  },
  {
    "epoch": 2768,
    "batch_idx": 96,
    "loss": 0.29453757405281067,
    "learning_rate": 0.0003217940606433749,
    "timestep_mean": 468.46875,
    "epsilon_norm": 7.769723892211914,
    "prediction_norm": 6.1007561683654785,
    "timestamp": "2025-07-14T12:35:13.422861"
  },
  {
    "epoch": 2768,
    "batch_idx": 128,
    "loss": 0.2982255220413208,
    "learning_rate": 0.0003217940606433749,
    "timestep_mean": 459.65625,
    "epsilon_norm": 7.809051036834717,
    "prediction_norm": 6.814284801483154,
    "timestamp": "2025-07-14T12:35:13.428210"
  },
  {
    "epoch": 2768,
    "batch_idx": 160,
    "loss": 0.3581658601760864,
    "learning_rate": 0.0003217940606433749,
    "timestep_mean": 422.65625,
    "epsilon_norm": 8.58014965057373,
    "prediction_norm": 6.5504150390625,
    "timestamp": "2025-07-14T12:35:13.433715"
  },
  {
    "epoch": 2768,
    "batch_idx": 192,
    "loss": 0.10369731485843658,
    "learning_rate": 0.0003217940606433749,
    "timestep_mean": 457.375,
    "epsilon_norm": 3.9160494804382324,
    "prediction_norm": 3.7118921279907227,
    "timestamp": "2025-07-14T12:35:13.439256"
  },
  {
    "epoch": 2769,
    "batch_idx": 0,
    "loss": 0.2937260866165161,
    "learning_rate": 0.00032252810042973857,
    "timestep_mean": 456.53125,
    "epsilon_norm": 8.35444164276123,
    "prediction_norm": 7.406572341918945,
    "timestamp": "2025-07-14T12:35:13.444752"
  },
  {
    "epoch": 2769,
    "batch_idx": 32,
    "loss": 0.19961418211460114,
    "learning_rate": 0.00032252810042973857,
    "timestep_mean": 536.6875,
    "epsilon_norm": 7.658763408660889,
    "prediction_norm": 6.635985851287842,
    "timestamp": "2025-07-14T12:35:13.450093"
  },
  {
    "epoch": 2769,
    "batch_idx": 64,
    "loss": 0.38798874616622925,
    "learning_rate": 0.00032252810042973857,
    "timestep_mean": 495.71875,
    "epsilon_norm": 7.074487209320068,
    "prediction_norm": 5.44218635559082,
    "timestamp": "2025-07-14T12:35:13.455641"
  },
  {
    "epoch": 2769,
    "batch_idx": 96,
    "loss": 0.39276987314224243,
    "learning_rate": 0.00032252810042973857,
    "timestep_mean": 482.65625,
    "epsilon_norm": 9.198297500610352,
    "prediction_norm": 8.007290840148926,
    "timestamp": "2025-07-14T12:35:13.461012"
  },
  {
    "epoch": 2769,
    "batch_idx": 128,
    "loss": 0.44257569313049316,
    "learning_rate": 0.00032252810042973857,
    "timestep_mean": 521.8125,
    "epsilon_norm": 8.539022445678711,
    "prediction_norm": 6.982123374938965,
    "timestamp": "2025-07-14T12:35:13.466417"
  },
  {
    "epoch": 2769,
    "batch_idx": 160,
    "loss": 0.16780970990657806,
    "learning_rate": 0.00032252810042973857,
    "timestep_mean": 535.6875,
    "epsilon_norm": 7.72463846206665,
    "prediction_norm": 6.804799556732178,
    "timestamp": "2025-07-14T12:35:13.472016"
  },
  {
    "epoch": 2769,
    "batch_idx": 192,
    "loss": 0.4710484445095062,
    "learning_rate": 0.00032252810042973857,
    "timestep_mean": 335.5,
    "epsilon_norm": 3.5252795219421387,
    "prediction_norm": 2.7676432132720947,
    "timestamp": "2025-07-14T12:35:13.477559"
  },
  {
    "epoch": 2770,
    "batch_idx": 0,
    "loss": 0.1254126876592636,
    "learning_rate": 0.00032326257811037207,
    "timestep_mean": 542.3125,
    "epsilon_norm": 7.008403778076172,
    "prediction_norm": 7.251054763793945,
    "timestamp": "2025-07-14T12:35:13.483067"
  },
  {
    "epoch": 2770,
    "batch_idx": 32,
    "loss": 0.29360783100128174,
    "learning_rate": 0.00032326257811037207,
    "timestep_mean": 527.5,
    "epsilon_norm": 8.6483736038208,
    "prediction_norm": 8.021699905395508,
    "timestamp": "2025-07-14T12:35:13.488527"
  },
  {
    "epoch": 2770,
    "batch_idx": 64,
    "loss": 0.16306385397911072,
    "learning_rate": 0.00032326257811037207,
    "timestep_mean": 569.46875,
    "epsilon_norm": 8.875187873840332,
    "prediction_norm": 8.381199836730957,
    "timestamp": "2025-07-14T12:35:13.493946"
  },
  {
    "epoch": 2770,
    "batch_idx": 96,
    "loss": 0.2776775360107422,
    "learning_rate": 0.00032326257811037207,
    "timestep_mean": 493.46875,
    "epsilon_norm": 8.986693382263184,
    "prediction_norm": 7.999520301818848,
    "timestamp": "2025-07-14T12:35:13.499338"
  },
  {
    "epoch": 2770,
    "batch_idx": 128,
    "loss": 0.10597044229507446,
    "learning_rate": 0.00032326257811037207,
    "timestep_mean": 556.96875,
    "epsilon_norm": 7.57631254196167,
    "prediction_norm": 7.375591278076172,
    "timestamp": "2025-07-14T12:35:13.504829"
  },
  {
    "epoch": 2770,
    "batch_idx": 160,
    "loss": 0.1376761496067047,
    "learning_rate": 0.00032326257811037207,
    "timestep_mean": 501.78125,
    "epsilon_norm": 7.148090839385986,
    "prediction_norm": 7.026288986206055,
    "timestamp": "2025-07-14T12:35:13.510215"
  },
  {
    "epoch": 2770,
    "batch_idx": 192,
    "loss": 0.1352933943271637,
    "learning_rate": 0.00032326257811037207,
    "timestep_mean": 574.75,
    "epsilon_norm": 3.5902559757232666,
    "prediction_norm": 3.016098976135254,
    "timestamp": "2025-07-14T12:35:13.515750"
  },
  {
    "epoch": 2771,
    "batch_idx": 0,
    "loss": 0.33896416425704956,
    "learning_rate": 0.0003239974918730251,
    "timestep_mean": 533.625,
    "epsilon_norm": 9.079172134399414,
    "prediction_norm": 7.857624053955078,
    "timestamp": "2025-07-14T12:35:13.521150"
  },
  {
    "epoch": 2771,
    "batch_idx": 32,
    "loss": 0.19479896128177643,
    "learning_rate": 0.0003239974918730251,
    "timestep_mean": 480.6875,
    "epsilon_norm": 7.619983196258545,
    "prediction_norm": 6.547856330871582,
    "timestamp": "2025-07-14T12:35:13.526552"
  },
  {
    "epoch": 2771,
    "batch_idx": 64,
    "loss": 0.26697999238967896,
    "learning_rate": 0.0003239974918730251,
    "timestep_mean": 455.0,
    "epsilon_norm": 8.141094207763672,
    "prediction_norm": 6.851504325866699,
    "timestamp": "2025-07-14T12:35:13.531943"
  },
  {
    "epoch": 2771,
    "batch_idx": 96,
    "loss": 0.24678157269954681,
    "learning_rate": 0.0003239974918730251,
    "timestep_mean": 518.03125,
    "epsilon_norm": 7.082847595214844,
    "prediction_norm": 6.444456100463867,
    "timestamp": "2025-07-14T12:35:13.537204"
  },
  {
    "epoch": 2771,
    "batch_idx": 128,
    "loss": 0.19647185504436493,
    "learning_rate": 0.0003239974918730251,
    "timestep_mean": 529.0,
    "epsilon_norm": 7.889230251312256,
    "prediction_norm": 6.854663372039795,
    "timestamp": "2025-07-14T12:35:13.542457"
  },
  {
    "epoch": 2771,
    "batch_idx": 160,
    "loss": 0.35544121265411377,
    "learning_rate": 0.0003239974918730251,
    "timestep_mean": 464.78125,
    "epsilon_norm": 8.064789772033691,
    "prediction_norm": 6.555693626403809,
    "timestamp": "2025-07-14T12:35:13.547801"
  },
  {
    "epoch": 2771,
    "batch_idx": 192,
    "loss": 0.42644429206848145,
    "learning_rate": 0.0003239974918730251,
    "timestep_mean": 461.875,
    "epsilon_norm": 3.590684652328491,
    "prediction_norm": 2.9855117797851562,
    "timestamp": "2025-07-14T12:35:13.553183"
  },
  {
    "epoch": 2772,
    "batch_idx": 0,
    "loss": 0.27837803959846497,
    "learning_rate": 0.0003247328399043706,
    "timestep_mean": 588.8125,
    "epsilon_norm": 8.215006828308105,
    "prediction_norm": 6.863149166107178,
    "timestamp": "2025-07-14T12:35:13.558662"
  },
  {
    "epoch": 2772,
    "batch_idx": 32,
    "loss": 0.21713587641716003,
    "learning_rate": 0.0003247328399043706,
    "timestep_mean": 507.03125,
    "epsilon_norm": 7.62130880355835,
    "prediction_norm": 6.399654865264893,
    "timestamp": "2025-07-14T12:35:13.563978"
  },
  {
    "epoch": 2772,
    "batch_idx": 64,
    "loss": 0.24549610912799835,
    "learning_rate": 0.0003247328399043706,
    "timestep_mean": 391.6875,
    "epsilon_norm": 6.994187831878662,
    "prediction_norm": 5.889540672302246,
    "timestamp": "2025-07-14T12:35:13.569379"
  },
  {
    "epoch": 2772,
    "batch_idx": 96,
    "loss": 0.17700369656085968,
    "learning_rate": 0.0003247328399043706,
    "timestep_mean": 511.96875,
    "epsilon_norm": 7.875685214996338,
    "prediction_norm": 7.153519630432129,
    "timestamp": "2025-07-14T12:35:13.574716"
  },
  {
    "epoch": 2772,
    "batch_idx": 128,
    "loss": 0.2532881200313568,
    "learning_rate": 0.0003247328399043706,
    "timestep_mean": 541.84375,
    "epsilon_norm": 7.328658103942871,
    "prediction_norm": 6.526907444000244,
    "timestamp": "2025-07-14T12:35:13.580230"
  },
  {
    "epoch": 2772,
    "batch_idx": 160,
    "loss": 0.45884624123573303,
    "learning_rate": 0.0003247328399043706,
    "timestep_mean": 497.46875,
    "epsilon_norm": 8.511283874511719,
    "prediction_norm": 6.451722145080566,
    "timestamp": "2025-07-14T12:35:13.585788"
  },
  {
    "epoch": 2772,
    "batch_idx": 192,
    "loss": 0.40665510296821594,
    "learning_rate": 0.0003247328399043706,
    "timestep_mean": 473.25,
    "epsilon_norm": 5.044209957122803,
    "prediction_norm": 3.908579111099243,
    "timestamp": "2025-07-14T12:35:13.591300"
  },
  {
    "epoch": 2773,
    "batch_idx": 0,
    "loss": 0.3963775634765625,
    "learning_rate": 0.00032546862039001175,
    "timestep_mean": 505.34375,
    "epsilon_norm": 9.09369945526123,
    "prediction_norm": 7.358882427215576,
    "timestamp": "2025-07-14T12:35:13.596771"
  },
  {
    "epoch": 2773,
    "batch_idx": 32,
    "loss": 0.2361435890197754,
    "learning_rate": 0.00032546862039001175,
    "timestep_mean": 439.75,
    "epsilon_norm": 7.042832374572754,
    "prediction_norm": 6.0026021003723145,
    "timestamp": "2025-07-14T12:35:13.602125"
  },
  {
    "epoch": 2773,
    "batch_idx": 64,
    "loss": 0.2038324475288391,
    "learning_rate": 0.00032546862039001175,
    "timestep_mean": 444.84375,
    "epsilon_norm": 7.047330379486084,
    "prediction_norm": 6.6988444328308105,
    "timestamp": "2025-07-14T12:35:13.607548"
  },
  {
    "epoch": 2773,
    "batch_idx": 96,
    "loss": 0.17743204534053802,
    "learning_rate": 0.00032546862039001175,
    "timestep_mean": 521.125,
    "epsilon_norm": 8.60423469543457,
    "prediction_norm": 7.3144145011901855,
    "timestamp": "2025-07-14T12:35:13.612905"
  },
  {
    "epoch": 2773,
    "batch_idx": 128,
    "loss": 0.30894985795021057,
    "learning_rate": 0.00032546862039001175,
    "timestep_mean": 581.96875,
    "epsilon_norm": 8.456622123718262,
    "prediction_norm": 6.729401111602783,
    "timestamp": "2025-07-14T12:35:13.618255"
  },
  {
    "epoch": 2773,
    "batch_idx": 160,
    "loss": 0.3018602728843689,
    "learning_rate": 0.00032546862039001175,
    "timestep_mean": 454.9375,
    "epsilon_norm": 7.373935222625732,
    "prediction_norm": 6.18588399887085,
    "timestamp": "2025-07-14T12:35:13.623598"
  },
  {
    "epoch": 2773,
    "batch_idx": 192,
    "loss": 0.3222726583480835,
    "learning_rate": 0.00032546862039001175,
    "timestep_mean": 540.5,
    "epsilon_norm": 3.071674346923828,
    "prediction_norm": 2.4035346508026123,
    "timestamp": "2025-07-14T12:35:13.629014"
  },
  {
    "epoch": 2774,
    "batch_idx": 0,
    "loss": 0.138652503490448,
    "learning_rate": 0.00032620483151448197,
    "timestep_mean": 561.53125,
    "epsilon_norm": 6.460405349731445,
    "prediction_norm": 5.6280741691589355,
    "timestamp": "2025-07-14T12:35:13.634539"
  },
  {
    "epoch": 2774,
    "batch_idx": 32,
    "loss": 0.2969805598258972,
    "learning_rate": 0.00032620483151448197,
    "timestep_mean": 503.96875,
    "epsilon_norm": 9.12575912475586,
    "prediction_norm": 7.435396194458008,
    "timestamp": "2025-07-14T12:35:13.639987"
  },
  {
    "epoch": 2774,
    "batch_idx": 64,
    "loss": 0.19402635097503662,
    "learning_rate": 0.00032620483151448197,
    "timestep_mean": 530.34375,
    "epsilon_norm": 7.878752708435059,
    "prediction_norm": 6.717524528503418,
    "timestamp": "2025-07-14T12:35:13.645338"
  },
  {
    "epoch": 2774,
    "batch_idx": 96,
    "loss": 0.47707414627075195,
    "learning_rate": 0.00032620483151448197,
    "timestep_mean": 441.3125,
    "epsilon_norm": 8.632107734680176,
    "prediction_norm": 6.5562262535095215,
    "timestamp": "2025-07-14T12:35:13.650810"
  },
  {
    "epoch": 2774,
    "batch_idx": 128,
    "loss": 0.17107030749320984,
    "learning_rate": 0.00032620483151448197,
    "timestep_mean": 490.0,
    "epsilon_norm": 6.966549873352051,
    "prediction_norm": 6.604525566101074,
    "timestamp": "2025-07-14T12:35:13.656219"
  },
  {
    "epoch": 2774,
    "batch_idx": 160,
    "loss": 0.28203830122947693,
    "learning_rate": 0.00032620483151448197,
    "timestep_mean": 510.40625,
    "epsilon_norm": 8.930621147155762,
    "prediction_norm": 8.527896881103516,
    "timestamp": "2025-07-14T12:35:13.661634"
  },
  {
    "epoch": 2774,
    "batch_idx": 192,
    "loss": 0.46636366844177246,
    "learning_rate": 0.00032620483151448197,
    "timestep_mean": 446.5,
    "epsilon_norm": 3.883749008178711,
    "prediction_norm": 3.077185869216919,
    "timestamp": "2025-07-14T12:35:13.667216"
  },
  {
    "epoch": 2775,
    "batch_idx": 0,
    "loss": 0.3258521556854248,
    "learning_rate": 0.0003269414714612539,
    "timestep_mean": 507.6875,
    "epsilon_norm": 8.288022994995117,
    "prediction_norm": 6.756896495819092,
    "timestamp": "2025-07-14T12:35:13.672734"
  },
  {
    "epoch": 2775,
    "batch_idx": 32,
    "loss": 0.22473488748073578,
    "learning_rate": 0.0003269414714612539,
    "timestep_mean": 445.15625,
    "epsilon_norm": 9.097834587097168,
    "prediction_norm": 8.14746379852295,
    "timestamp": "2025-07-14T12:35:13.678168"
  },
  {
    "epoch": 2775,
    "batch_idx": 64,
    "loss": 0.40671810507774353,
    "learning_rate": 0.0003269414714612539,
    "timestep_mean": 528.09375,
    "epsilon_norm": 7.883338928222656,
    "prediction_norm": 5.698857307434082,
    "timestamp": "2025-07-14T12:35:13.683363"
  },
  {
    "epoch": 2775,
    "batch_idx": 96,
    "loss": 0.0869467556476593,
    "learning_rate": 0.0003269414714612539,
    "timestep_mean": 568.3125,
    "epsilon_norm": 7.348570823669434,
    "prediction_norm": 6.686819076538086,
    "timestamp": "2025-07-14T12:35:13.688646"
  },
  {
    "epoch": 2775,
    "batch_idx": 128,
    "loss": 0.3244761824607849,
    "learning_rate": 0.0003269414714612539,
    "timestep_mean": 403.5,
    "epsilon_norm": 6.649723052978516,
    "prediction_norm": 4.742440223693848,
    "timestamp": "2025-07-14T12:35:13.693847"
  },
  {
    "epoch": 2775,
    "batch_idx": 160,
    "loss": 0.24157026410102844,
    "learning_rate": 0.0003269414714612539,
    "timestep_mean": 452.4375,
    "epsilon_norm": 7.772407531738281,
    "prediction_norm": 6.368978023529053,
    "timestamp": "2025-07-14T12:35:13.698991"
  },
  {
    "epoch": 2775,
    "batch_idx": 192,
    "loss": 0.16238781809806824,
    "learning_rate": 0.0003269414714612539,
    "timestep_mean": 400.375,
    "epsilon_norm": 3.5980453491210938,
    "prediction_norm": 3.0391337871551514,
    "timestamp": "2025-07-14T12:35:13.704567"
  },
  {
    "epoch": 2776,
    "batch_idx": 0,
    "loss": 0.3253301680088043,
    "learning_rate": 0.0003276785384127422,
    "timestep_mean": 516.4375,
    "epsilon_norm": 7.5528564453125,
    "prediction_norm": 6.836515426635742,
    "timestamp": "2025-07-14T12:35:13.709790"
  },
  {
    "epoch": 2776,
    "batch_idx": 32,
    "loss": 0.42123275995254517,
    "learning_rate": 0.0003276785384127422,
    "timestep_mean": 543.8125,
    "epsilon_norm": 8.488081932067871,
    "prediction_norm": 6.617332458496094,
    "timestamp": "2025-07-14T12:35:13.714997"
  },
  {
    "epoch": 2776,
    "batch_idx": 64,
    "loss": 0.15255536139011383,
    "learning_rate": 0.0003276785384127422,
    "timestep_mean": 509.46875,
    "epsilon_norm": 7.429451942443848,
    "prediction_norm": 7.155055046081543,
    "timestamp": "2025-07-14T12:35:13.720394"
  },
  {
    "epoch": 2776,
    "batch_idx": 96,
    "loss": 0.3091825842857361,
    "learning_rate": 0.0003276785384127422,
    "timestep_mean": 547.90625,
    "epsilon_norm": 9.261187553405762,
    "prediction_norm": 7.21327018737793,
    "timestamp": "2025-07-14T12:35:13.725666"
  },
  {
    "epoch": 2776,
    "batch_idx": 128,
    "loss": 0.3268201947212219,
    "learning_rate": 0.0003276785384127422,
    "timestep_mean": 473.8125,
    "epsilon_norm": 7.92460823059082,
    "prediction_norm": 6.409449577331543,
    "timestamp": "2025-07-14T12:35:13.730889"
  },
  {
    "epoch": 2776,
    "batch_idx": 160,
    "loss": 0.2257048487663269,
    "learning_rate": 0.0003276785384127422,
    "timestep_mean": 477.4375,
    "epsilon_norm": 7.446640968322754,
    "prediction_norm": 6.164243221282959,
    "timestamp": "2025-07-14T12:35:13.736047"
  },
  {
    "epoch": 2776,
    "batch_idx": 192,
    "loss": 0.20663732290267944,
    "learning_rate": 0.0003276785384127422,
    "timestep_mean": 444.125,
    "epsilon_norm": 3.0779173374176025,
    "prediction_norm": 2.5676429271698,
    "timestamp": "2025-07-14T12:35:13.741316"
  },
  {
    "epoch": 2777,
    "batch_idx": 0,
    "loss": 0.3853437006473541,
    "learning_rate": 0.00032841603055030664,
    "timestep_mean": 507.9375,
    "epsilon_norm": 8.388737678527832,
    "prediction_norm": 7.04024600982666,
    "timestamp": "2025-07-14T12:35:13.746413"
  },
  {
    "epoch": 2777,
    "batch_idx": 32,
    "loss": 0.353507399559021,
    "learning_rate": 0.00032841603055030664,
    "timestep_mean": 405.40625,
    "epsilon_norm": 8.123117446899414,
    "prediction_norm": 6.456214904785156,
    "timestamp": "2025-07-14T12:35:13.751526"
  },
  {
    "epoch": 2777,
    "batch_idx": 64,
    "loss": 0.3745254874229431,
    "learning_rate": 0.00032841603055030664,
    "timestep_mean": 453.5625,
    "epsilon_norm": 8.93835163116455,
    "prediction_norm": 7.04354190826416,
    "timestamp": "2025-07-14T12:35:13.756605"
  },
  {
    "epoch": 2777,
    "batch_idx": 96,
    "loss": 0.26444095373153687,
    "learning_rate": 0.00032841603055030664,
    "timestep_mean": 485.59375,
    "epsilon_norm": 9.043708801269531,
    "prediction_norm": 7.4693169593811035,
    "timestamp": "2025-07-14T12:35:13.761611"
  },
  {
    "epoch": 2777,
    "batch_idx": 128,
    "loss": 0.3931645154953003,
    "learning_rate": 0.00032841603055030664,
    "timestep_mean": 495.625,
    "epsilon_norm": 8.327997207641602,
    "prediction_norm": 6.548774719238281,
    "timestamp": "2025-07-14T12:35:13.766669"
  },
  {
    "epoch": 2777,
    "batch_idx": 160,
    "loss": 0.19407612085342407,
    "learning_rate": 0.00032841603055030664,
    "timestep_mean": 466.5,
    "epsilon_norm": 7.497767448425293,
    "prediction_norm": 6.547155380249023,
    "timestamp": "2025-07-14T12:35:13.771822"
  },
  {
    "epoch": 2777,
    "batch_idx": 192,
    "loss": 0.27772533893585205,
    "learning_rate": 0.00032841603055030664,
    "timestep_mean": 376.25,
    "epsilon_norm": 4.523063659667969,
    "prediction_norm": 4.104254245758057,
    "timestamp": "2025-07-14T12:35:13.776750"
  },
  {
    "epoch": 2778,
    "batch_idx": 0,
    "loss": 0.28878986835479736,
    "learning_rate": 0.00032915394605425853,
    "timestep_mean": 550.375,
    "epsilon_norm": 8.672019004821777,
    "prediction_norm": 8.027580261230469,
    "timestamp": "2025-07-14T12:35:13.781953"
  },
  {
    "epoch": 2778,
    "batch_idx": 32,
    "loss": 0.14731070399284363,
    "learning_rate": 0.00032915394605425853,
    "timestep_mean": 574.625,
    "epsilon_norm": 8.043998718261719,
    "prediction_norm": 7.428836345672607,
    "timestamp": "2025-07-14T12:35:13.786943"
  },
  {
    "epoch": 2778,
    "batch_idx": 64,
    "loss": 0.30950939655303955,
    "learning_rate": 0.00032915394605425853,
    "timestep_mean": 544.96875,
    "epsilon_norm": 9.376852035522461,
    "prediction_norm": 8.197612762451172,
    "timestamp": "2025-07-14T12:35:13.791795"
  },
  {
    "epoch": 2778,
    "batch_idx": 96,
    "loss": 0.17793095111846924,
    "learning_rate": 0.00032915394605425853,
    "timestep_mean": 510.25,
    "epsilon_norm": 7.362398147583008,
    "prediction_norm": 7.0253400802612305,
    "timestamp": "2025-07-14T12:35:13.796772"
  },
  {
    "epoch": 2778,
    "batch_idx": 128,
    "loss": 0.3317491412162781,
    "learning_rate": 0.00032915394605425853,
    "timestep_mean": 403.46875,
    "epsilon_norm": 7.631625652313232,
    "prediction_norm": 6.736538410186768,
    "timestamp": "2025-07-14T12:35:13.801616"
  },
  {
    "epoch": 2778,
    "batch_idx": 160,
    "loss": 0.20835208892822266,
    "learning_rate": 0.00032915394605425853,
    "timestep_mean": 482.46875,
    "epsilon_norm": 6.949294090270996,
    "prediction_norm": 6.298135280609131,
    "timestamp": "2025-07-14T12:35:13.806608"
  },
  {
    "epoch": 2778,
    "batch_idx": 192,
    "loss": 0.21793478727340698,
    "learning_rate": 0.00032915394605425853,
    "timestep_mean": 621.0,
    "epsilon_norm": 3.1118850708007812,
    "prediction_norm": 3.039407968521118,
    "timestamp": "2025-07-14T12:35:13.811585"
  },
  {
    "epoch": 2779,
    "batch_idx": 0,
    "loss": 0.20792031288146973,
    "learning_rate": 0.00032989228310386574,
    "timestep_mean": 429.09375,
    "epsilon_norm": 7.292995929718018,
    "prediction_norm": 6.327084064483643,
    "timestamp": "2025-07-14T12:35:13.816497"
  },
  {
    "epoch": 2779,
    "batch_idx": 32,
    "loss": 0.11672306805849075,
    "learning_rate": 0.00032989228310386574,
    "timestep_mean": 639.875,
    "epsilon_norm": 8.493301391601562,
    "prediction_norm": 7.757692813873291,
    "timestamp": "2025-07-14T12:35:13.821469"
  },
  {
    "epoch": 2779,
    "batch_idx": 64,
    "loss": 0.27028241753578186,
    "learning_rate": 0.00032989228310386574,
    "timestep_mean": 427.96875,
    "epsilon_norm": 7.853867053985596,
    "prediction_norm": 7.17383337020874,
    "timestamp": "2025-07-14T12:35:13.826365"
  },
  {
    "epoch": 2779,
    "batch_idx": 96,
    "loss": 0.24492889642715454,
    "learning_rate": 0.00032989228310386574,
    "timestep_mean": 422.3125,
    "epsilon_norm": 7.973180294036865,
    "prediction_norm": 6.890307426452637,
    "timestamp": "2025-07-14T12:35:13.831205"
  },
  {
    "epoch": 2779,
    "batch_idx": 128,
    "loss": 0.23535849153995514,
    "learning_rate": 0.00032989228310386574,
    "timestep_mean": 457.125,
    "epsilon_norm": 8.195989608764648,
    "prediction_norm": 6.469207286834717,
    "timestamp": "2025-07-14T12:35:13.836130"
  },
  {
    "epoch": 2779,
    "batch_idx": 160,
    "loss": 0.48152956366539,
    "learning_rate": 0.00032989228310386574,
    "timestep_mean": 422.90625,
    "epsilon_norm": 9.459831237792969,
    "prediction_norm": 7.865565776824951,
    "timestamp": "2025-07-14T12:35:13.841010"
  },
  {
    "epoch": 2779,
    "batch_idx": 192,
    "loss": 0.28460875153541565,
    "learning_rate": 0.00032989228310386574,
    "timestep_mean": 435.25,
    "epsilon_norm": 4.943689823150635,
    "prediction_norm": 3.968482732772827,
    "timestamp": "2025-07-14T12:35:13.845906"
  },
  {
    "epoch": 2780,
    "batch_idx": 0,
    "loss": 0.3629732131958008,
    "learning_rate": 0.0003306310398773549,
    "timestep_mean": 411.6875,
    "epsilon_norm": 7.295107364654541,
    "prediction_norm": 5.450940132141113,
    "timestamp": "2025-07-14T12:35:13.850991"
  },
  {
    "epoch": 2780,
    "batch_idx": 32,
    "loss": 0.2225831001996994,
    "learning_rate": 0.0003306310398773549,
    "timestep_mean": 469.375,
    "epsilon_norm": 7.793324947357178,
    "prediction_norm": 6.360358238220215,
    "timestamp": "2025-07-14T12:35:13.855877"
  },
  {
    "epoch": 2780,
    "batch_idx": 64,
    "loss": 0.25711846351623535,
    "learning_rate": 0.0003306310398773549,
    "timestep_mean": 406.625,
    "epsilon_norm": 8.05300521850586,
    "prediction_norm": 6.859641075134277,
    "timestamp": "2025-07-14T12:35:13.860717"
  },
  {
    "epoch": 2780,
    "batch_idx": 96,
    "loss": 0.18885470926761627,
    "learning_rate": 0.0003306310398773549,
    "timestep_mean": 480.625,
    "epsilon_norm": 9.586531639099121,
    "prediction_norm": 8.559972763061523,
    "timestamp": "2025-07-14T12:35:13.865548"
  },
  {
    "epoch": 2780,
    "batch_idx": 128,
    "loss": 0.3924827575683594,
    "learning_rate": 0.0003306310398773549,
    "timestep_mean": 507.125,
    "epsilon_norm": 8.951101303100586,
    "prediction_norm": 7.120509147644043,
    "timestamp": "2025-07-14T12:35:13.870437"
  },
  {
    "epoch": 2780,
    "batch_idx": 160,
    "loss": 0.3040020763874054,
    "learning_rate": 0.0003306310398773549,
    "timestep_mean": 474.28125,
    "epsilon_norm": 8.449298858642578,
    "prediction_norm": 6.783050060272217,
    "timestamp": "2025-07-14T12:35:13.875378"
  },
  {
    "epoch": 2780,
    "batch_idx": 192,
    "loss": 0.36208251118659973,
    "learning_rate": 0.0003306310398773549,
    "timestep_mean": 485.5,
    "epsilon_norm": 4.049394607543945,
    "prediction_norm": 3.6131393909454346,
    "timestamp": "2025-07-14T12:35:13.880278"
  },
  {
    "epoch": 2781,
    "batch_idx": 0,
    "loss": 0.3431078791618347,
    "learning_rate": 0.00033137021455191627,
    "timestep_mean": 455.5,
    "epsilon_norm": 8.018442153930664,
    "prediction_norm": 6.307980060577393,
    "timestamp": "2025-07-14T12:35:13.885228"
  },
  {
    "epoch": 2781,
    "batch_idx": 32,
    "loss": 0.4205883741378784,
    "learning_rate": 0.00033137021455191627,
    "timestep_mean": 454.71875,
    "epsilon_norm": 7.61041784286499,
    "prediction_norm": 5.349969863891602,
    "timestamp": "2025-07-14T12:35:13.890081"
  },
  {
    "epoch": 2781,
    "batch_idx": 64,
    "loss": 0.26477599143981934,
    "learning_rate": 0.00033137021455191627,
    "timestep_mean": 503.71875,
    "epsilon_norm": 7.11610221862793,
    "prediction_norm": 6.245423793792725,
    "timestamp": "2025-07-14T12:35:13.895052"
  },
  {
    "epoch": 2781,
    "batch_idx": 96,
    "loss": 0.39505645632743835,
    "learning_rate": 0.00033137021455191627,
    "timestep_mean": 438.875,
    "epsilon_norm": 7.397941589355469,
    "prediction_norm": 5.686349868774414,
    "timestamp": "2025-07-14T12:35:13.899925"
  },
  {
    "epoch": 2781,
    "batch_idx": 128,
    "loss": 0.35242199897766113,
    "learning_rate": 0.00033137021455191627,
    "timestep_mean": 458.1875,
    "epsilon_norm": 8.894253730773926,
    "prediction_norm": 6.878180503845215,
    "timestamp": "2025-07-14T12:35:13.904824"
  },
  {
    "epoch": 2781,
    "batch_idx": 160,
    "loss": 0.28433817625045776,
    "learning_rate": 0.00033137021455191627,
    "timestep_mean": 423.6875,
    "epsilon_norm": 7.8636369705200195,
    "prediction_norm": 6.823282241821289,
    "timestamp": "2025-07-14T12:35:13.909690"
  },
  {
    "epoch": 2781,
    "batch_idx": 192,
    "loss": 0.2859831154346466,
    "learning_rate": 0.00033137021455191627,
    "timestep_mean": 627.875,
    "epsilon_norm": 3.75116229057312,
    "prediction_norm": 2.677107572555542,
    "timestamp": "2025-07-14T12:35:13.914628"
  },
  {
    "epoch": 2782,
    "batch_idx": 0,
    "loss": 0.41234147548675537,
    "learning_rate": 0.00033210980530371025,
    "timestep_mean": 468.0625,
    "epsilon_norm": 8.398460388183594,
    "prediction_norm": 6.73493766784668,
    "timestamp": "2025-07-14T12:35:13.919745"
  },
  {
    "epoch": 2782,
    "batch_idx": 32,
    "loss": 0.19157211482524872,
    "learning_rate": 0.00033210980530371025,
    "timestep_mean": 530.875,
    "epsilon_norm": 7.528461933135986,
    "prediction_norm": 7.365387916564941,
    "timestamp": "2025-07-14T12:35:13.924776"
  },
  {
    "epoch": 2782,
    "batch_idx": 64,
    "loss": 0.2148844301700592,
    "learning_rate": 0.00033210980530371025,
    "timestep_mean": 527.59375,
    "epsilon_norm": 6.865739345550537,
    "prediction_norm": 6.215489864349365,
    "timestamp": "2025-07-14T12:35:13.929875"
  },
  {
    "epoch": 2782,
    "batch_idx": 96,
    "loss": 0.16615208983421326,
    "learning_rate": 0.00033210980530371025,
    "timestep_mean": 551.125,
    "epsilon_norm": 7.235555648803711,
    "prediction_norm": 6.941812992095947,
    "timestamp": "2025-07-14T12:35:13.934748"
  },
  {
    "epoch": 2782,
    "batch_idx": 128,
    "loss": 0.2740013599395752,
    "learning_rate": 0.00033210980530371025,
    "timestep_mean": 488.71875,
    "epsilon_norm": 8.263721466064453,
    "prediction_norm": 7.142472743988037,
    "timestamp": "2025-07-14T12:35:13.939623"
  },
  {
    "epoch": 2782,
    "batch_idx": 160,
    "loss": 0.3596401810646057,
    "learning_rate": 0.00033210980530371025,
    "timestep_mean": 499.875,
    "epsilon_norm": 8.967425346374512,
    "prediction_norm": 6.9126362800598145,
    "timestamp": "2025-07-14T12:35:13.944520"
  },
  {
    "epoch": 2782,
    "batch_idx": 192,
    "loss": 0.566093385219574,
    "learning_rate": 0.00033210980530371025,
    "timestep_mean": 348.625,
    "epsilon_norm": 3.591862916946411,
    "prediction_norm": 2.746882677078247,
    "timestamp": "2025-07-14T12:35:13.949430"
  },
  {
    "epoch": 2783,
    "batch_idx": 0,
    "loss": 0.30369675159454346,
    "learning_rate": 0.00033284981030787014,
    "timestep_mean": 491.90625,
    "epsilon_norm": 8.142921447753906,
    "prediction_norm": 7.051654815673828,
    "timestamp": "2025-07-14T12:35:13.954353"
  },
  {
    "epoch": 2783,
    "batch_idx": 32,
    "loss": 0.4332534074783325,
    "learning_rate": 0.00033284981030787014,
    "timestep_mean": 492.21875,
    "epsilon_norm": 8.999536514282227,
    "prediction_norm": 6.72325325012207,
    "timestamp": "2025-07-14T12:35:13.959298"
  },
  {
    "epoch": 2783,
    "batch_idx": 64,
    "loss": 0.409226655960083,
    "learning_rate": 0.00033284981030787014,
    "timestep_mean": 479.3125,
    "epsilon_norm": 7.429268836975098,
    "prediction_norm": 5.803380966186523,
    "timestamp": "2025-07-14T12:35:13.964232"
  },
  {
    "epoch": 2783,
    "batch_idx": 96,
    "loss": 0.2008223682641983,
    "learning_rate": 0.00033284981030787014,
    "timestep_mean": 597.5,
    "epsilon_norm": 8.513029098510742,
    "prediction_norm": 7.721036911010742,
    "timestamp": "2025-07-14T12:35:13.969064"
  },
  {
    "epoch": 2783,
    "batch_idx": 128,
    "loss": 0.2641516923904419,
    "learning_rate": 0.00033284981030787014,
    "timestep_mean": 397.8125,
    "epsilon_norm": 7.377753257751465,
    "prediction_norm": 5.282680988311768,
    "timestamp": "2025-07-14T12:35:13.973969"
  },
  {
    "epoch": 2783,
    "batch_idx": 160,
    "loss": 0.27811986207962036,
    "learning_rate": 0.00033284981030787014,
    "timestep_mean": 540.71875,
    "epsilon_norm": 9.31313419342041,
    "prediction_norm": 8.257301330566406,
    "timestamp": "2025-07-14T12:35:13.978790"
  },
  {
    "epoch": 2783,
    "batch_idx": 192,
    "loss": 0.5935746431350708,
    "learning_rate": 0.00033284981030787014,
    "timestep_mean": 470.0,
    "epsilon_norm": 4.608105182647705,
    "prediction_norm": 3.779940605163574,
    "timestamp": "2025-07-14T12:35:13.983846"
  },
  {
    "epoch": 2784,
    "batch_idx": 0,
    "loss": 0.29841870069503784,
    "learning_rate": 0.0003335902277385071,
    "timestep_mean": 500.65625,
    "epsilon_norm": 8.921022415161133,
    "prediction_norm": 7.196990013122559,
    "timestamp": "2025-07-14T12:35:13.988688"
  },
  {
    "epoch": 2784,
    "batch_idx": 32,
    "loss": 0.20168614387512207,
    "learning_rate": 0.0003335902277385071,
    "timestep_mean": 552.4375,
    "epsilon_norm": 7.32046365737915,
    "prediction_norm": 6.247675895690918,
    "timestamp": "2025-07-14T12:35:13.993795"
  },
  {
    "epoch": 2784,
    "batch_idx": 64,
    "loss": 0.19305968284606934,
    "learning_rate": 0.0003335902277385071,
    "timestep_mean": 495.0,
    "epsilon_norm": 8.43578052520752,
    "prediction_norm": 7.086343288421631,
    "timestamp": "2025-07-14T12:35:13.998739"
  },
  {
    "epoch": 2784,
    "batch_idx": 96,
    "loss": 0.4263732433319092,
    "learning_rate": 0.0003335902277385071,
    "timestep_mean": 528.90625,
    "epsilon_norm": 7.990440845489502,
    "prediction_norm": 5.917942523956299,
    "timestamp": "2025-07-14T12:35:14.003743"
  },
  {
    "epoch": 2784,
    "batch_idx": 128,
    "loss": 0.25247129797935486,
    "learning_rate": 0.0003335902277385071,
    "timestep_mean": 470.96875,
    "epsilon_norm": 8.295392036437988,
    "prediction_norm": 7.443105220794678,
    "timestamp": "2025-07-14T12:35:14.008717"
  },
  {
    "epoch": 2784,
    "batch_idx": 160,
    "loss": 0.2993132174015045,
    "learning_rate": 0.0003335902277385071,
    "timestep_mean": 539.75,
    "epsilon_norm": 8.519071578979492,
    "prediction_norm": 7.65943717956543,
    "timestamp": "2025-07-14T12:35:14.013648"
  },
  {
    "epoch": 2784,
    "batch_idx": 192,
    "loss": 0.19511106610298157,
    "learning_rate": 0.0003335902277385071,
    "timestep_mean": 642.5,
    "epsilon_norm": 3.699010133743286,
    "prediction_norm": 3.7413008213043213,
    "timestamp": "2025-07-14T12:35:14.018689"
  },
  {
    "epoch": 2785,
    "batch_idx": 0,
    "loss": 0.20044609904289246,
    "learning_rate": 0.0003343310557687149,
    "timestep_mean": 505.59375,
    "epsilon_norm": 6.709964752197266,
    "prediction_norm": 6.055840969085693,
    "timestamp": "2025-07-14T12:35:14.023718"
  },
  {
    "epoch": 2785,
    "batch_idx": 32,
    "loss": 0.3123837113380432,
    "learning_rate": 0.0003343310557687149,
    "timestep_mean": 408.46875,
    "epsilon_norm": 7.623050212860107,
    "prediction_norm": 6.078343868255615,
    "timestamp": "2025-07-14T12:35:14.028645"
  },
  {
    "epoch": 2785,
    "batch_idx": 64,
    "loss": 0.5760999321937561,
    "learning_rate": 0.0003343310557687149,
    "timestep_mean": 460.125,
    "epsilon_norm": 8.082413673400879,
    "prediction_norm": 6.215439796447754,
    "timestamp": "2025-07-14T12:35:14.033496"
  },
  {
    "epoch": 2785,
    "batch_idx": 96,
    "loss": 0.2537476420402527,
    "learning_rate": 0.0003343310557687149,
    "timestep_mean": 506.3125,
    "epsilon_norm": 8.080830574035645,
    "prediction_norm": 6.938091278076172,
    "timestamp": "2025-07-14T12:35:14.038493"
  },
  {
    "epoch": 2785,
    "batch_idx": 128,
    "loss": 0.10354065895080566,
    "learning_rate": 0.0003343310557687149,
    "timestep_mean": 566.78125,
    "epsilon_norm": 8.069707870483398,
    "prediction_norm": 7.622395992279053,
    "timestamp": "2025-07-14T12:35:14.043308"
  },
  {
    "epoch": 2785,
    "batch_idx": 160,
    "loss": 0.24849644303321838,
    "learning_rate": 0.0003343310557687149,
    "timestep_mean": 509.5625,
    "epsilon_norm": 7.803232669830322,
    "prediction_norm": 7.061293601989746,
    "timestamp": "2025-07-14T12:35:14.048235"
  },
  {
    "epoch": 2785,
    "batch_idx": 192,
    "loss": 0.5222986340522766,
    "learning_rate": 0.0003343310557687149,
    "timestep_mean": 404.0,
    "epsilon_norm": 4.209906101226807,
    "prediction_norm": 3.4879422187805176,
    "timestamp": "2025-07-14T12:35:14.053150"
  },
  {
    "epoch": 2786,
    "batch_idx": 0,
    "loss": 0.4105321168899536,
    "learning_rate": 0.00033507229257057433,
    "timestep_mean": 540.5,
    "epsilon_norm": 8.228628158569336,
    "prediction_norm": 7.137570858001709,
    "timestamp": "2025-07-14T12:35:14.058053"
  },
  {
    "epoch": 2786,
    "batch_idx": 32,
    "loss": 0.18830427527427673,
    "learning_rate": 0.00033507229257057433,
    "timestep_mean": 513.5625,
    "epsilon_norm": 7.959229946136475,
    "prediction_norm": 6.60762357711792,
    "timestamp": "2025-07-14T12:35:14.062885"
  },
  {
    "epoch": 2786,
    "batch_idx": 64,
    "loss": 0.3122606873512268,
    "learning_rate": 0.00033507229257057433,
    "timestep_mean": 555.59375,
    "epsilon_norm": 8.77207088470459,
    "prediction_norm": 7.293014049530029,
    "timestamp": "2025-07-14T12:35:14.067822"
  },
  {
    "epoch": 2786,
    "batch_idx": 96,
    "loss": 0.33097511529922485,
    "learning_rate": 0.00033507229257057433,
    "timestep_mean": 438.46875,
    "epsilon_norm": 8.005266189575195,
    "prediction_norm": 5.84778356552124,
    "timestamp": "2025-07-14T12:35:14.072707"
  },
  {
    "epoch": 2786,
    "batch_idx": 128,
    "loss": 0.39586323499679565,
    "learning_rate": 0.00033507229257057433,
    "timestep_mean": 535.5625,
    "epsilon_norm": 7.913194179534912,
    "prediction_norm": 5.962011814117432,
    "timestamp": "2025-07-14T12:35:14.077608"
  },
  {
    "epoch": 2786,
    "batch_idx": 160,
    "loss": 0.2972111701965332,
    "learning_rate": 0.00033507229257057433,
    "timestep_mean": 518.9375,
    "epsilon_norm": 9.064221382141113,
    "prediction_norm": 7.5804643630981445,
    "timestamp": "2025-07-14T12:35:14.082717"
  },
  {
    "epoch": 2786,
    "batch_idx": 192,
    "loss": 0.05630046874284744,
    "learning_rate": 0.00033507229257057433,
    "timestep_mean": 502.0,
    "epsilon_norm": 3.416548490524292,
    "prediction_norm": 3.464719533920288,
    "timestamp": "2025-07-14T12:35:14.087744"
  },
  {
    "epoch": 2787,
    "batch_idx": 0,
    "loss": 0.21944978833198547,
    "learning_rate": 0.00033581393631515645,
    "timestep_mean": 476.53125,
    "epsilon_norm": 6.763197898864746,
    "prediction_norm": 5.72951078414917,
    "timestamp": "2025-07-14T12:35:14.092956"
  },
  {
    "epoch": 2787,
    "batch_idx": 32,
    "loss": 0.2908744215965271,
    "learning_rate": 0.00033581393631515645,
    "timestep_mean": 508.6875,
    "epsilon_norm": 8.248270988464355,
    "prediction_norm": 7.269774913787842,
    "timestamp": "2025-07-14T12:35:14.098103"
  },
  {
    "epoch": 2787,
    "batch_idx": 64,
    "loss": 0.32266032695770264,
    "learning_rate": 0.00033581393631515645,
    "timestep_mean": 602.71875,
    "epsilon_norm": 8.206419944763184,
    "prediction_norm": 7.188443183898926,
    "timestamp": "2025-07-14T12:35:14.103245"
  },
  {
    "epoch": 2787,
    "batch_idx": 96,
    "loss": 0.27089136838912964,
    "learning_rate": 0.00033581393631515645,
    "timestep_mean": 537.90625,
    "epsilon_norm": 8.248921394348145,
    "prediction_norm": 7.261775970458984,
    "timestamp": "2025-07-14T12:35:14.108465"
  },
  {
    "epoch": 2787,
    "batch_idx": 128,
    "loss": 0.2808387577533722,
    "learning_rate": 0.00033581393631515645,
    "timestep_mean": 496.03125,
    "epsilon_norm": 7.272368431091309,
    "prediction_norm": 6.2884650230407715,
    "timestamp": "2025-07-14T12:35:14.113744"
  },
  {
    "epoch": 2787,
    "batch_idx": 160,
    "loss": 0.11506234854459763,
    "learning_rate": 0.00033581393631515645,
    "timestep_mean": 531.125,
    "epsilon_norm": 6.8643646240234375,
    "prediction_norm": 7.117947101593018,
    "timestamp": "2025-07-14T12:35:14.118733"
  },
  {
    "epoch": 2787,
    "batch_idx": 192,
    "loss": 0.07911211252212524,
    "learning_rate": 0.00033581393631515645,
    "timestep_mean": 487.5,
    "epsilon_norm": 3.318789005279541,
    "prediction_norm": 2.76686429977417,
    "timestamp": "2025-07-14T12:35:14.123634"
  },
  {
    "epoch": 2788,
    "batch_idx": 0,
    "loss": 0.14579874277114868,
    "learning_rate": 0.00033655598517252943,
    "timestep_mean": 565.96875,
    "epsilon_norm": 7.055203437805176,
    "prediction_norm": 6.154120922088623,
    "timestamp": "2025-07-14T12:35:14.128500"
  },
  {
    "epoch": 2788,
    "batch_idx": 32,
    "loss": 0.1747576743364334,
    "learning_rate": 0.00033655598517252943,
    "timestep_mean": 498.0625,
    "epsilon_norm": 8.73078441619873,
    "prediction_norm": 7.835709095001221,
    "timestamp": "2025-07-14T12:35:14.133386"
  },
  {
    "epoch": 2788,
    "batch_idx": 64,
    "loss": 0.2167808562517166,
    "learning_rate": 0.00033655598517252943,
    "timestep_mean": 529.59375,
    "epsilon_norm": 7.49971866607666,
    "prediction_norm": 6.708437442779541,
    "timestamp": "2025-07-14T12:35:14.138307"
  },
  {
    "epoch": 2788,
    "batch_idx": 96,
    "loss": 0.2886649966239929,
    "learning_rate": 0.00033655598517252943,
    "timestep_mean": 510.25,
    "epsilon_norm": 7.660767078399658,
    "prediction_norm": 6.669319152832031,
    "timestamp": "2025-07-14T12:35:14.143252"
  },
  {
    "epoch": 2788,
    "batch_idx": 128,
    "loss": 0.16097474098205566,
    "learning_rate": 0.00033655598517252943,
    "timestep_mean": 587.15625,
    "epsilon_norm": 7.4530110359191895,
    "prediction_norm": 6.4153242111206055,
    "timestamp": "2025-07-14T12:35:14.148291"
  },
  {
    "epoch": 2788,
    "batch_idx": 160,
    "loss": 0.25285667181015015,
    "learning_rate": 0.00033655598517252943,
    "timestep_mean": 471.71875,
    "epsilon_norm": 8.301258087158203,
    "prediction_norm": 7.089061737060547,
    "timestamp": "2025-07-14T12:35:14.153361"
  },
  {
    "epoch": 2788,
    "batch_idx": 192,
    "loss": 0.2700887620449066,
    "learning_rate": 0.00033655598517252943,
    "timestep_mean": 481.5,
    "epsilon_norm": 3.6291215419769287,
    "prediction_norm": 2.881251335144043,
    "timestamp": "2025-07-14T12:35:14.158446"
  },
  {
    "epoch": 2789,
    "batch_idx": 0,
    "loss": 0.21759964525699615,
    "learning_rate": 0.00033729843731176106,
    "timestep_mean": 506.59375,
    "epsilon_norm": 6.797236919403076,
    "prediction_norm": 6.090973854064941,
    "timestamp": "2025-07-14T12:35:14.163711"
  },
  {
    "epoch": 2789,
    "batch_idx": 32,
    "loss": 0.17161113023757935,
    "learning_rate": 0.00033729843731176106,
    "timestep_mean": 619.78125,
    "epsilon_norm": 8.779483795166016,
    "prediction_norm": 7.578819274902344,
    "timestamp": "2025-07-14T12:35:14.168862"
  },
  {
    "epoch": 2789,
    "batch_idx": 64,
    "loss": 0.3904307782649994,
    "learning_rate": 0.00033729843731176106,
    "timestep_mean": 550.625,
    "epsilon_norm": 8.419665336608887,
    "prediction_norm": 6.610352516174316,
    "timestamp": "2025-07-14T12:35:14.173925"
  },
  {
    "epoch": 2789,
    "batch_idx": 96,
    "loss": 0.26157647371292114,
    "learning_rate": 0.00033729843731176106,
    "timestep_mean": 548.9375,
    "epsilon_norm": 8.558334350585938,
    "prediction_norm": 7.0981950759887695,
    "timestamp": "2025-07-14T12:35:14.179079"
  },
  {
    "epoch": 2789,
    "batch_idx": 128,
    "loss": 0.26986899971961975,
    "learning_rate": 0.00033729843731176106,
    "timestep_mean": 575.1875,
    "epsilon_norm": 8.448899269104004,
    "prediction_norm": 6.837996482849121,
    "timestamp": "2025-07-14T12:35:14.184187"
  },
  {
    "epoch": 2789,
    "batch_idx": 160,
    "loss": 0.21850699186325073,
    "learning_rate": 0.00033729843731176106,
    "timestep_mean": 480.78125,
    "epsilon_norm": 8.277182579040527,
    "prediction_norm": 7.632463455200195,
    "timestamp": "2025-07-14T12:35:14.189075"
  },
  {
    "epoch": 2789,
    "batch_idx": 192,
    "loss": 0.6719680428504944,
    "learning_rate": 0.00033729843731176106,
    "timestep_mean": 500.75,
    "epsilon_norm": 5.374942302703857,
    "prediction_norm": 4.212756156921387,
    "timestamp": "2025-07-14T12:35:14.194037"
  },
  {
    "epoch": 2790,
    "batch_idx": 0,
    "loss": 0.4030669927597046,
    "learning_rate": 0.0003380412909009259,
    "timestep_mean": 506.28125,
    "epsilon_norm": 8.867907524108887,
    "prediction_norm": 6.608705520629883,
    "timestamp": "2025-07-14T12:35:14.199036"
  },
  {
    "epoch": 2790,
    "batch_idx": 32,
    "loss": 0.38812848925590515,
    "learning_rate": 0.0003380412909009259,
    "timestep_mean": 490.09375,
    "epsilon_norm": 9.603588104248047,
    "prediction_norm": 7.894231796264648,
    "timestamp": "2025-07-14T12:35:14.204099"
  },
  {
    "epoch": 2790,
    "batch_idx": 64,
    "loss": 0.29539865255355835,
    "learning_rate": 0.0003380412909009259,
    "timestep_mean": 552.0625,
    "epsilon_norm": 7.228529453277588,
    "prediction_norm": 5.62932014465332,
    "timestamp": "2025-07-14T12:35:14.209408"
  },
  {
    "epoch": 2790,
    "batch_idx": 96,
    "loss": 0.287547767162323,
    "learning_rate": 0.0003380412909009259,
    "timestep_mean": 446.78125,
    "epsilon_norm": 8.471049308776855,
    "prediction_norm": 6.717018127441406,
    "timestamp": "2025-07-14T12:35:14.214367"
  },
  {
    "epoch": 2790,
    "batch_idx": 128,
    "loss": 0.22410961985588074,
    "learning_rate": 0.0003380412909009259,
    "timestep_mean": 543.28125,
    "epsilon_norm": 7.634448528289795,
    "prediction_norm": 6.496231555938721,
    "timestamp": "2025-07-14T12:35:14.219253"
  },
  {
    "epoch": 2790,
    "batch_idx": 160,
    "loss": 0.20341061055660248,
    "learning_rate": 0.0003380412909009259,
    "timestep_mean": 485.125,
    "epsilon_norm": 7.077900409698486,
    "prediction_norm": 6.4126482009887695,
    "timestamp": "2025-07-14T12:35:14.224308"
  },
  {
    "epoch": 2790,
    "batch_idx": 192,
    "loss": 0.38806048035621643,
    "learning_rate": 0.0003380412909009259,
    "timestep_mean": 434.125,
    "epsilon_norm": 3.5579922199249268,
    "prediction_norm": 3.1766293048858643,
    "timestamp": "2025-07-14T12:35:14.229438"
  },
  {
    "epoch": 2791,
    "batch_idx": 0,
    "loss": 0.15484076738357544,
    "learning_rate": 0.00033878454410710503,
    "timestep_mean": 460.84375,
    "epsilon_norm": 7.0378923416137695,
    "prediction_norm": 6.6343159675598145,
    "timestamp": "2025-07-14T12:35:14.234556"
  },
  {
    "epoch": 2791,
    "batch_idx": 32,
    "loss": 0.18720164895057678,
    "learning_rate": 0.00033878454410710503,
    "timestep_mean": 560.4375,
    "epsilon_norm": 8.231945037841797,
    "prediction_norm": 7.173596382141113,
    "timestamp": "2025-07-14T12:35:14.239743"
  },
  {
    "epoch": 2791,
    "batch_idx": 64,
    "loss": 0.2416786253452301,
    "learning_rate": 0.00033878454410710503,
    "timestep_mean": 521.8125,
    "epsilon_norm": 8.242717742919922,
    "prediction_norm": 7.374215126037598,
    "timestamp": "2025-07-14T12:35:14.245014"
  },
  {
    "epoch": 2791,
    "batch_idx": 96,
    "loss": 0.39105117321014404,
    "learning_rate": 0.00033878454410710503,
    "timestep_mean": 489.375,
    "epsilon_norm": 7.994200229644775,
    "prediction_norm": 6.858853816986084,
    "timestamp": "2025-07-14T12:35:14.250112"
  },
  {
    "epoch": 2791,
    "batch_idx": 128,
    "loss": 0.2968360185623169,
    "learning_rate": 0.00033878454410710503,
    "timestep_mean": 537.78125,
    "epsilon_norm": 7.721371173858643,
    "prediction_norm": 6.652326583862305,
    "timestamp": "2025-07-14T12:35:14.255474"
  },
  {
    "epoch": 2791,
    "batch_idx": 160,
    "loss": 0.3136805593967438,
    "learning_rate": 0.00033878454410710503,
    "timestep_mean": 464.21875,
    "epsilon_norm": 7.862031936645508,
    "prediction_norm": 6.772212505340576,
    "timestamp": "2025-07-14T12:35:14.260626"
  },
  {
    "epoch": 2791,
    "batch_idx": 192,
    "loss": 0.12467014044523239,
    "learning_rate": 0.00033878454410710503,
    "timestep_mean": 646.875,
    "epsilon_norm": 4.0155930519104,
    "prediction_norm": 3.7341809272766113,
    "timestamp": "2025-07-14T12:35:14.265684"
  },
  {
    "epoch": 2792,
    "batch_idx": 0,
    "loss": 0.45781928300857544,
    "learning_rate": 0.00033952819509639566,
    "timestep_mean": 425.0625,
    "epsilon_norm": 7.338231086730957,
    "prediction_norm": 5.182127952575684,
    "timestamp": "2025-07-14T12:35:14.270908"
  },
  {
    "epoch": 2792,
    "batch_idx": 32,
    "loss": 0.23921117186546326,
    "learning_rate": 0.00033952819509639566,
    "timestep_mean": 469.65625,
    "epsilon_norm": 7.942672252655029,
    "prediction_norm": 6.497037410736084,
    "timestamp": "2025-07-14T12:35:14.276110"
  },
  {
    "epoch": 2792,
    "batch_idx": 64,
    "loss": 0.3579585552215576,
    "learning_rate": 0.00033952819509639566,
    "timestep_mean": 477.78125,
    "epsilon_norm": 7.537535190582275,
    "prediction_norm": 6.375535011291504,
    "timestamp": "2025-07-14T12:35:14.281255"
  },
  {
    "epoch": 2792,
    "batch_idx": 96,
    "loss": 0.428669810295105,
    "learning_rate": 0.00033952819509639566,
    "timestep_mean": 462.21875,
    "epsilon_norm": 8.063715934753418,
    "prediction_norm": 6.606743812561035,
    "timestamp": "2025-07-14T12:35:14.286184"
  },
  {
    "epoch": 2792,
    "batch_idx": 128,
    "loss": 0.17912116646766663,
    "learning_rate": 0.00033952819509639566,
    "timestep_mean": 539.21875,
    "epsilon_norm": 8.02685832977295,
    "prediction_norm": 7.605119705200195,
    "timestamp": "2025-07-14T12:35:14.291081"
  },
  {
    "epoch": 2792,
    "batch_idx": 160,
    "loss": 0.2511481046676636,
    "learning_rate": 0.00033952819509639566,
    "timestep_mean": 506.53125,
    "epsilon_norm": 7.8791937828063965,
    "prediction_norm": 7.294136047363281,
    "timestamp": "2025-07-14T12:35:14.296156"
  },
  {
    "epoch": 2792,
    "batch_idx": 192,
    "loss": 0.45131683349609375,
    "learning_rate": 0.00033952819509639566,
    "timestep_mean": 319.0,
    "epsilon_norm": 3.8999953269958496,
    "prediction_norm": 3.0333340167999268,
    "timestamp": "2025-07-14T12:35:14.301326"
  },
  {
    "epoch": 2793,
    "batch_idx": 0,
    "loss": 0.27230769395828247,
    "learning_rate": 0.00034027224203391334,
    "timestep_mean": 475.5,
    "epsilon_norm": 8.511917114257812,
    "prediction_norm": 7.877279758453369,
    "timestamp": "2025-07-14T12:35:14.306472"
  },
  {
    "epoch": 2793,
    "batch_idx": 32,
    "loss": 0.5093900561332703,
    "learning_rate": 0.00034027224203391334,
    "timestep_mean": 455.875,
    "epsilon_norm": 8.669785499572754,
    "prediction_norm": 6.433501720428467,
    "timestamp": "2025-07-14T12:35:14.311675"
  },
  {
    "epoch": 2793,
    "batch_idx": 64,
    "loss": 0.12065988779067993,
    "learning_rate": 0.00034027224203391334,
    "timestep_mean": 559.71875,
    "epsilon_norm": 7.202980995178223,
    "prediction_norm": 6.535956382751465,
    "timestamp": "2025-07-14T12:35:14.316875"
  },
  {
    "epoch": 2793,
    "batch_idx": 96,
    "loss": 0.1819462776184082,
    "learning_rate": 0.00034027224203391334,
    "timestep_mean": 549.28125,
    "epsilon_norm": 8.498437881469727,
    "prediction_norm": 7.310187816619873,
    "timestamp": "2025-07-14T12:35:14.321861"
  },
  {
    "epoch": 2793,
    "batch_idx": 128,
    "loss": 0.44815629720687866,
    "learning_rate": 0.00034027224203391334,
    "timestep_mean": 427.40625,
    "epsilon_norm": 8.57043170928955,
    "prediction_norm": 6.122495174407959,
    "timestamp": "2025-07-14T12:35:14.327052"
  },
  {
    "epoch": 2793,
    "batch_idx": 160,
    "loss": 0.3377358913421631,
    "learning_rate": 0.00034027224203391334,
    "timestep_mean": 442.3125,
    "epsilon_norm": 8.533319473266602,
    "prediction_norm": 6.795463562011719,
    "timestamp": "2025-07-14T12:35:14.332194"
  },
  {
    "epoch": 2793,
    "batch_idx": 192,
    "loss": 0.09900309145450592,
    "learning_rate": 0.00034027224203391334,
    "timestep_mean": 535.25,
    "epsilon_norm": 3.537510871887207,
    "prediction_norm": 3.4075238704681396,
    "timestamp": "2025-07-14T12:35:14.337389"
  },
  {
    "epoch": 2794,
    "batch_idx": 0,
    "loss": 0.33771783113479614,
    "learning_rate": 0.00034101668308379485,
    "timestep_mean": 488.25,
    "epsilon_norm": 7.819575309753418,
    "prediction_norm": 5.891109943389893,
    "timestamp": "2025-07-14T12:35:14.342769"
  },
  {
    "epoch": 2794,
    "batch_idx": 32,
    "loss": 0.2853253185749054,
    "learning_rate": 0.00034101668308379485,
    "timestep_mean": 459.625,
    "epsilon_norm": 8.077651023864746,
    "prediction_norm": 6.457541465759277,
    "timestamp": "2025-07-14T12:35:14.347917"
  },
  {
    "epoch": 2794,
    "batch_idx": 64,
    "loss": 0.2180309146642685,
    "learning_rate": 0.00034101668308379485,
    "timestep_mean": 530.5625,
    "epsilon_norm": 7.344140529632568,
    "prediction_norm": 6.88347053527832,
    "timestamp": "2025-07-14T12:35:14.353075"
  },
  {
    "epoch": 2794,
    "batch_idx": 96,
    "loss": 0.2825508117675781,
    "learning_rate": 0.00034101668308379485,
    "timestep_mean": 557.0625,
    "epsilon_norm": 8.348587989807129,
    "prediction_norm": 7.032145023345947,
    "timestamp": "2025-07-14T12:35:14.358333"
  },
  {
    "epoch": 2794,
    "batch_idx": 128,
    "loss": 0.4538354277610779,
    "learning_rate": 0.00034101668308379485,
    "timestep_mean": 429.1875,
    "epsilon_norm": 8.717256546020508,
    "prediction_norm": 6.8163557052612305,
    "timestamp": "2025-07-14T12:35:14.363607"
  },
  {
    "epoch": 2794,
    "batch_idx": 160,
    "loss": 0.21894562244415283,
    "learning_rate": 0.00034101668308379485,
    "timestep_mean": 557.34375,
    "epsilon_norm": 8.921124458312988,
    "prediction_norm": 8.013999938964844,
    "timestamp": "2025-07-14T12:35:14.368724"
  },
  {
    "epoch": 2794,
    "batch_idx": 192,
    "loss": 0.06357306241989136,
    "learning_rate": 0.00034101668308379485,
    "timestep_mean": 536.25,
    "epsilon_norm": 3.5484988689422607,
    "prediction_norm": 3.6238389015197754,
    "timestamp": "2025-07-14T12:35:14.373915"
  },
  {
    "epoch": 2795,
    "batch_idx": 0,
    "loss": 0.30198273062705994,
    "learning_rate": 0.0003417615164092073,
    "timestep_mean": 491.6875,
    "epsilon_norm": 7.863363265991211,
    "prediction_norm": 6.484491348266602,
    "timestamp": "2025-07-14T12:35:14.379197"
  },
  {
    "epoch": 2795,
    "batch_idx": 32,
    "loss": 0.24399031698703766,
    "learning_rate": 0.0003417615164092073,
    "timestep_mean": 524.40625,
    "epsilon_norm": 7.249258041381836,
    "prediction_norm": 5.940190315246582,
    "timestamp": "2025-07-14T12:35:14.384702"
  },
  {
    "epoch": 2795,
    "batch_idx": 64,
    "loss": 0.43119633197784424,
    "learning_rate": 0.0003417615164092073,
    "timestep_mean": 410.9375,
    "epsilon_norm": 8.058818817138672,
    "prediction_norm": 6.452996253967285,
    "timestamp": "2025-07-14T12:35:14.389788"
  },
  {
    "epoch": 2795,
    "batch_idx": 96,
    "loss": 0.2900916337966919,
    "learning_rate": 0.0003417615164092073,
    "timestep_mean": 458.09375,
    "epsilon_norm": 8.060502052307129,
    "prediction_norm": 6.881420612335205,
    "timestamp": "2025-07-14T12:35:14.395150"
  },
  {
    "epoch": 2795,
    "batch_idx": 128,
    "loss": 0.37463510036468506,
    "learning_rate": 0.0003417615164092073,
    "timestep_mean": 419.8125,
    "epsilon_norm": 7.566250324249268,
    "prediction_norm": 5.83877420425415,
    "timestamp": "2025-07-14T12:35:14.400553"
  },
  {
    "epoch": 2795,
    "batch_idx": 160,
    "loss": 0.8611267805099487,
    "learning_rate": 0.0003417615164092073,
    "timestep_mean": 387.90625,
    "epsilon_norm": 9.4675874710083,
    "prediction_norm": 6.940304279327393,
    "timestamp": "2025-07-14T12:35:14.405647"
  },
  {
    "epoch": 2795,
    "batch_idx": 192,
    "loss": 0.21789410710334778,
    "learning_rate": 0.0003417615164092073,
    "timestep_mean": 537.125,
    "epsilon_norm": 4.84628438949585,
    "prediction_norm": 4.453025817871094,
    "timestamp": "2025-07-14T12:35:14.410709"
  },
  {
    "epoch": 2796,
    "batch_idx": 0,
    "loss": 0.26402586698532104,
    "learning_rate": 0.00034250674017234785,
    "timestep_mean": 518.75,
    "epsilon_norm": 8.474573135375977,
    "prediction_norm": 7.292713642120361,
    "timestamp": "2025-07-14T12:35:14.415817"
  },
  {
    "epoch": 2796,
    "batch_idx": 32,
    "loss": 0.22514164447784424,
    "learning_rate": 0.00034250674017234785,
    "timestep_mean": 539.21875,
    "epsilon_norm": 7.342179775238037,
    "prediction_norm": 6.681300163269043,
    "timestamp": "2025-07-14T12:35:14.420886"
  },
  {
    "epoch": 2796,
    "batch_idx": 64,
    "loss": 0.20931945741176605,
    "learning_rate": 0.00034250674017234785,
    "timestep_mean": 598.0625,
    "epsilon_norm": 8.106704711914062,
    "prediction_norm": 7.5058183670043945,
    "timestamp": "2025-07-14T12:35:14.425943"
  },
  {
    "epoch": 2796,
    "batch_idx": 96,
    "loss": 0.33275532722473145,
    "learning_rate": 0.00034250674017234785,
    "timestep_mean": 506.8125,
    "epsilon_norm": 9.43847370147705,
    "prediction_norm": 8.109193801879883,
    "timestamp": "2025-07-14T12:35:14.430798"
  },
  {
    "epoch": 2796,
    "batch_idx": 128,
    "loss": 0.25117868185043335,
    "learning_rate": 0.00034250674017234785,
    "timestep_mean": 482.9375,
    "epsilon_norm": 8.321743965148926,
    "prediction_norm": 7.461702346801758,
    "timestamp": "2025-07-14T12:35:14.435711"
  },
  {
    "epoch": 2796,
    "batch_idx": 160,
    "loss": 0.23925496637821198,
    "learning_rate": 0.00034250674017234785,
    "timestep_mean": 557.5,
    "epsilon_norm": 8.134293556213379,
    "prediction_norm": 6.98995304107666,
    "timestamp": "2025-07-14T12:35:14.440684"
  },
  {
    "epoch": 2796,
    "batch_idx": 192,
    "loss": 0.17718671262264252,
    "learning_rate": 0.00034250674017234785,
    "timestep_mean": 425.75,
    "epsilon_norm": 3.16847825050354,
    "prediction_norm": 2.9858908653259277,
    "timestamp": "2025-07-14T12:35:14.445624"
  },
  {
    "epoch": 2797,
    "batch_idx": 0,
    "loss": 0.2926294207572937,
    "learning_rate": 0.00034325235253445155,
    "timestep_mean": 514.84375,
    "epsilon_norm": 8.570417404174805,
    "prediction_norm": 7.700107097625732,
    "timestamp": "2025-07-14T12:35:14.450588"
  },
  {
    "epoch": 2797,
    "batch_idx": 32,
    "loss": 0.21136397123336792,
    "learning_rate": 0.00034325235253445155,
    "timestep_mean": 542.53125,
    "epsilon_norm": 7.216522216796875,
    "prediction_norm": 6.591475486755371,
    "timestamp": "2025-07-14T12:35:14.455470"
  },
  {
    "epoch": 2797,
    "batch_idx": 64,
    "loss": 0.2638446092605591,
    "learning_rate": 0.00034325235253445155,
    "timestep_mean": 413.1875,
    "epsilon_norm": 7.8863749504089355,
    "prediction_norm": 6.532975196838379,
    "timestamp": "2025-07-14T12:35:14.460504"
  },
  {
    "epoch": 2797,
    "batch_idx": 96,
    "loss": 0.301457941532135,
    "learning_rate": 0.00034325235253445155,
    "timestep_mean": 530.03125,
    "epsilon_norm": 7.936946868896484,
    "prediction_norm": 7.162876605987549,
    "timestamp": "2025-07-14T12:35:14.465559"
  },
  {
    "epoch": 2797,
    "batch_idx": 128,
    "loss": 0.2755054533481598,
    "learning_rate": 0.00034325235253445155,
    "timestep_mean": 489.1875,
    "epsilon_norm": 7.914016246795654,
    "prediction_norm": 6.32029914855957,
    "timestamp": "2025-07-14T12:35:14.470729"
  },
  {
    "epoch": 2797,
    "batch_idx": 160,
    "loss": 0.13966648280620575,
    "learning_rate": 0.00034325235253445155,
    "timestep_mean": 572.40625,
    "epsilon_norm": 8.117023468017578,
    "prediction_norm": 7.594162464141846,
    "timestamp": "2025-07-14T12:35:14.475811"
  },
  {
    "epoch": 2797,
    "batch_idx": 192,
    "loss": 0.20379097759723663,
    "learning_rate": 0.00034325235253445155,
    "timestep_mean": 590.625,
    "epsilon_norm": 3.0044307708740234,
    "prediction_norm": 2.806788682937622,
    "timestamp": "2025-07-14T12:35:14.480866"
  },
  {
    "epoch": 2798,
    "batch_idx": 0,
    "loss": 0.35883843898773193,
    "learning_rate": 0.0003439983516557932,
    "timestep_mean": 531.34375,
    "epsilon_norm": 8.644397735595703,
    "prediction_norm": 7.580714702606201,
    "timestamp": "2025-07-14T12:35:14.485977"
  },
  {
    "epoch": 2798,
    "batch_idx": 32,
    "loss": 0.22315162420272827,
    "learning_rate": 0.0003439983516557932,
    "timestep_mean": 435.625,
    "epsilon_norm": 7.425307273864746,
    "prediction_norm": 6.239847183227539,
    "timestamp": "2025-07-14T12:35:14.491000"
  },
  {
    "epoch": 2798,
    "batch_idx": 64,
    "loss": 0.23498833179473877,
    "learning_rate": 0.0003439983516557932,
    "timestep_mean": 550.9375,
    "epsilon_norm": 7.887192249298096,
    "prediction_norm": 6.560179710388184,
    "timestamp": "2025-07-14T12:35:14.496067"
  },
  {
    "epoch": 2798,
    "batch_idx": 96,
    "loss": 0.1729157567024231,
    "learning_rate": 0.0003439983516557932,
    "timestep_mean": 525.8125,
    "epsilon_norm": 8.040485382080078,
    "prediction_norm": 7.2057414054870605,
    "timestamp": "2025-07-14T12:35:14.501236"
  },
  {
    "epoch": 2798,
    "batch_idx": 128,
    "loss": 0.26976943016052246,
    "learning_rate": 0.0003439983516557932,
    "timestep_mean": 462.0625,
    "epsilon_norm": 7.896828651428223,
    "prediction_norm": 6.992391586303711,
    "timestamp": "2025-07-14T12:35:14.506404"
  },
  {
    "epoch": 2798,
    "batch_idx": 160,
    "loss": 0.15655693411827087,
    "learning_rate": 0.0003439983516557932,
    "timestep_mean": 541.53125,
    "epsilon_norm": 7.4037957191467285,
    "prediction_norm": 5.937868118286133,
    "timestamp": "2025-07-14T12:35:14.511495"
  },
  {
    "epoch": 2798,
    "batch_idx": 192,
    "loss": 0.39352548122406006,
    "learning_rate": 0.0003439983516557932,
    "timestep_mean": 522.875,
    "epsilon_norm": 4.016345500946045,
    "prediction_norm": 3.415433406829834,
    "timestamp": "2025-07-14T12:35:14.516592"
  },
  {
    "epoch": 2799,
    "batch_idx": 0,
    "loss": 0.2728201746940613,
    "learning_rate": 0.0003447447356956944,
    "timestep_mean": 497.65625,
    "epsilon_norm": 8.030206680297852,
    "prediction_norm": 6.358059883117676,
    "timestamp": "2025-07-14T12:35:14.521953"
  },
  {
    "epoch": 2799,
    "batch_idx": 32,
    "loss": 0.37908580899238586,
    "learning_rate": 0.0003447447356956944,
    "timestep_mean": 457.71875,
    "epsilon_norm": 8.03520679473877,
    "prediction_norm": 5.858366012573242,
    "timestamp": "2025-07-14T12:35:14.527084"
  },
  {
    "epoch": 2799,
    "batch_idx": 64,
    "loss": 0.1837119311094284,
    "learning_rate": 0.0003447447356956944,
    "timestep_mean": 594.15625,
    "epsilon_norm": 7.591813087463379,
    "prediction_norm": 6.397414207458496,
    "timestamp": "2025-07-14T12:35:14.532158"
  },
  {
    "epoch": 2799,
    "batch_idx": 96,
    "loss": 0.38407859206199646,
    "learning_rate": 0.0003447447356956944,
    "timestep_mean": 466.625,
    "epsilon_norm": 7.718578338623047,
    "prediction_norm": 6.093908786773682,
    "timestamp": "2025-07-14T12:35:14.537244"
  },
  {
    "epoch": 2799,
    "batch_idx": 128,
    "loss": 0.23023271560668945,
    "learning_rate": 0.0003447447356956944,
    "timestep_mean": 489.0,
    "epsilon_norm": 7.68910551071167,
    "prediction_norm": 6.514288902282715,
    "timestamp": "2025-07-14T12:35:14.542504"
  },
  {
    "epoch": 2799,
    "batch_idx": 160,
    "loss": 0.09452600032091141,
    "learning_rate": 0.0003447447356956944,
    "timestep_mean": 567.1875,
    "epsilon_norm": 7.788435459136963,
    "prediction_norm": 7.515281677246094,
    "timestamp": "2025-07-14T12:35:14.547704"
  },
  {
    "epoch": 2799,
    "batch_idx": 192,
    "loss": 0.30915045738220215,
    "learning_rate": 0.0003447447356956944,
    "timestep_mean": 628.375,
    "epsilon_norm": 4.234306812286377,
    "prediction_norm": 3.8985326290130615,
    "timestamp": "2025-07-14T12:35:14.552917"
  },
  {
    "epoch": 2800,
    "batch_idx": 0,
    "loss": 0.30461904406547546,
    "learning_rate": 0.0003454915028125268,
    "timestep_mean": 447.125,
    "epsilon_norm": 7.750520706176758,
    "prediction_norm": 6.4728617668151855,
    "timestamp": "2025-07-14T12:35:14.558106"
  },
  {
    "epoch": 2800,
    "batch_idx": 32,
    "loss": 0.15610992908477783,
    "learning_rate": 0.0003454915028125268,
    "timestep_mean": 544.6875,
    "epsilon_norm": 7.360022068023682,
    "prediction_norm": 6.71543025970459,
    "timestamp": "2025-07-14T12:35:14.563461"
  },
  {
    "epoch": 2800,
    "batch_idx": 64,
    "loss": 0.14328835904598236,
    "learning_rate": 0.0003454915028125268,
    "timestep_mean": 562.21875,
    "epsilon_norm": 7.48799991607666,
    "prediction_norm": 6.750604629516602,
    "timestamp": "2025-07-14T12:35:14.568616"
  },
  {
    "epoch": 2800,
    "batch_idx": 96,
    "loss": 0.25532084703445435,
    "learning_rate": 0.0003454915028125268,
    "timestep_mean": 408.53125,
    "epsilon_norm": 7.8716206550598145,
    "prediction_norm": 6.6153483390808105,
    "timestamp": "2025-07-14T12:35:14.573793"
  },
  {
    "epoch": 2800,
    "batch_idx": 128,
    "loss": 0.2058514803647995,
    "learning_rate": 0.0003454915028125268,
    "timestep_mean": 589.375,
    "epsilon_norm": 7.204770088195801,
    "prediction_norm": 5.881476402282715,
    "timestamp": "2025-07-14T12:35:14.578997"
  },
  {
    "epoch": 2800,
    "batch_idx": 160,
    "loss": 0.09717258810997009,
    "learning_rate": 0.0003454915028125268,
    "timestep_mean": 459.8125,
    "epsilon_norm": 7.834768295288086,
    "prediction_norm": 7.7260026931762695,
    "timestamp": "2025-07-14T12:35:14.584250"
  },
  {
    "epoch": 2800,
    "batch_idx": 192,
    "loss": 0.8020739555358887,
    "learning_rate": 0.0003454915028125268,
    "timestep_mean": 452.5,
    "epsilon_norm": 4.263435363769531,
    "prediction_norm": 2.0106201171875,
    "timestamp": "2025-07-14T12:35:14.589398"
  },
  {
    "epoch": 2801,
    "batch_idx": 0,
    "loss": 0.2456967532634735,
    "learning_rate": 0.0003462386511637168,
    "timestep_mean": 514.3125,
    "epsilon_norm": 6.751231670379639,
    "prediction_norm": 6.274271488189697,
    "timestamp": "2025-07-14T12:35:14.594471"
  },
  {
    "epoch": 2801,
    "batch_idx": 32,
    "loss": 0.14472179114818573,
    "learning_rate": 0.0003462386511637168,
    "timestep_mean": 562.75,
    "epsilon_norm": 7.199399948120117,
    "prediction_norm": 6.082761287689209,
    "timestamp": "2025-07-14T12:35:14.599539"
  },
  {
    "epoch": 2801,
    "batch_idx": 64,
    "loss": 0.19451335072517395,
    "learning_rate": 0.0003462386511637168,
    "timestep_mean": 560.96875,
    "epsilon_norm": 9.088047981262207,
    "prediction_norm": 8.965811729431152,
    "timestamp": "2025-07-14T12:35:14.604330"
  },
  {
    "epoch": 2801,
    "batch_idx": 96,
    "loss": 0.2118264138698578,
    "learning_rate": 0.0003462386511637168,
    "timestep_mean": 477.96875,
    "epsilon_norm": 6.794744968414307,
    "prediction_norm": 5.981710433959961,
    "timestamp": "2025-07-14T12:35:14.609359"
  },
  {
    "epoch": 2801,
    "batch_idx": 128,
    "loss": 0.2864919900894165,
    "learning_rate": 0.0003462386511637168,
    "timestep_mean": 454.15625,
    "epsilon_norm": 8.159656524658203,
    "prediction_norm": 6.455173492431641,
    "timestamp": "2025-07-14T12:35:14.614281"
  },
  {
    "epoch": 2801,
    "batch_idx": 160,
    "loss": 0.19607463479042053,
    "learning_rate": 0.0003462386511637168,
    "timestep_mean": 592.5,
    "epsilon_norm": 7.699460506439209,
    "prediction_norm": 6.983944416046143,
    "timestamp": "2025-07-14T12:35:14.619172"
  },
  {
    "epoch": 2801,
    "batch_idx": 192,
    "loss": 0.21387401223182678,
    "learning_rate": 0.0003462386511637168,
    "timestep_mean": 419.375,
    "epsilon_norm": 3.5626814365386963,
    "prediction_norm": 3.412419557571411,
    "timestamp": "2025-07-14T12:35:14.624070"
  },
  {
    "epoch": 2802,
    "batch_idx": 0,
    "loss": 0.29383164644241333,
    "learning_rate": 0.0003469861789057501,
    "timestep_mean": 508.84375,
    "epsilon_norm": 7.339009761810303,
    "prediction_norm": 5.984430313110352,
    "timestamp": "2025-07-14T12:35:14.628956"
  },
  {
    "epoch": 2802,
    "batch_idx": 32,
    "loss": 0.1254328340291977,
    "learning_rate": 0.0003469861789057501,
    "timestep_mean": 537.84375,
    "epsilon_norm": 6.653263568878174,
    "prediction_norm": 6.6675286293029785,
    "timestamp": "2025-07-14T12:35:14.633915"
  },
  {
    "epoch": 2802,
    "batch_idx": 64,
    "loss": 0.21874956786632538,
    "learning_rate": 0.0003469861789057501,
    "timestep_mean": 410.9375,
    "epsilon_norm": 7.888552665710449,
    "prediction_norm": 7.282941818237305,
    "timestamp": "2025-07-14T12:35:14.638807"
  },
  {
    "epoch": 2802,
    "batch_idx": 96,
    "loss": 0.2577871084213257,
    "learning_rate": 0.0003469861789057501,
    "timestep_mean": 552.34375,
    "epsilon_norm": 8.650742530822754,
    "prediction_norm": 7.478903770446777,
    "timestamp": "2025-07-14T12:35:14.643705"
  },
  {
    "epoch": 2802,
    "batch_idx": 128,
    "loss": 0.32504549622535706,
    "learning_rate": 0.0003469861789057501,
    "timestep_mean": 445.9375,
    "epsilon_norm": 7.546457290649414,
    "prediction_norm": 5.928062438964844,
    "timestamp": "2025-07-14T12:35:14.648575"
  },
  {
    "epoch": 2802,
    "batch_idx": 160,
    "loss": 0.4138913154602051,
    "learning_rate": 0.0003469861789057501,
    "timestep_mean": 462.375,
    "epsilon_norm": 8.462407112121582,
    "prediction_norm": 6.267151355743408,
    "timestamp": "2025-07-14T12:35:14.653487"
  },
  {
    "epoch": 2802,
    "batch_idx": 192,
    "loss": 0.3200172185897827,
    "learning_rate": 0.0003469861789057501,
    "timestep_mean": 592.875,
    "epsilon_norm": 6.426757335662842,
    "prediction_norm": 6.104854106903076,
    "timestamp": "2025-07-14T12:35:14.658352"
  },
  {
    "epoch": 2803,
    "batch_idx": 0,
    "loss": 0.27807438373565674,
    "learning_rate": 0.0003477340841941763,
    "timestep_mean": 501.125,
    "epsilon_norm": 7.010939598083496,
    "prediction_norm": 5.746162414550781,
    "timestamp": "2025-07-14T12:35:14.663303"
  },
  {
    "epoch": 2803,
    "batch_idx": 32,
    "loss": 0.36688241362571716,
    "learning_rate": 0.0003477340841941763,
    "timestep_mean": 415.53125,
    "epsilon_norm": 7.400082111358643,
    "prediction_norm": 5.548248291015625,
    "timestamp": "2025-07-14T12:35:14.668104"
  },
  {
    "epoch": 2803,
    "batch_idx": 64,
    "loss": 0.09971034526824951,
    "learning_rate": 0.0003477340841941763,
    "timestep_mean": 636.96875,
    "epsilon_norm": 7.32047700881958,
    "prediction_norm": 6.74527645111084,
    "timestamp": "2025-07-14T12:35:14.673007"
  },
  {
    "epoch": 2803,
    "batch_idx": 96,
    "loss": 0.2245015799999237,
    "learning_rate": 0.0003477340841941763,
    "timestep_mean": 473.03125,
    "epsilon_norm": 6.682570934295654,
    "prediction_norm": 5.874456405639648,
    "timestamp": "2025-07-14T12:35:14.677827"
  },
  {
    "epoch": 2803,
    "batch_idx": 128,
    "loss": 0.38223928213119507,
    "learning_rate": 0.0003477340841941763,
    "timestep_mean": 489.0,
    "epsilon_norm": 8.902594566345215,
    "prediction_norm": 7.29548978805542,
    "timestamp": "2025-07-14T12:35:14.682708"
  },
  {
    "epoch": 2803,
    "batch_idx": 160,
    "loss": 0.3420226573944092,
    "learning_rate": 0.0003477340841941763,
    "timestep_mean": 503.8125,
    "epsilon_norm": 7.381524085998535,
    "prediction_norm": 6.652951717376709,
    "timestamp": "2025-07-14T12:35:14.687658"
  },
  {
    "epoch": 2803,
    "batch_idx": 192,
    "loss": 0.09916017949581146,
    "learning_rate": 0.0003477340841941763,
    "timestep_mean": 550.875,
    "epsilon_norm": 3.9636945724487305,
    "prediction_norm": 3.348327398300171,
    "timestamp": "2025-07-14T12:35:14.692613"
  },
  {
    "epoch": 2804,
    "batch_idx": 0,
    "loss": 0.48643937706947327,
    "learning_rate": 0.00034848236518361393,
    "timestep_mean": 458.25,
    "epsilon_norm": 8.854706764221191,
    "prediction_norm": 6.979076862335205,
    "timestamp": "2025-07-14T12:35:14.697612"
  },
  {
    "epoch": 2804,
    "batch_idx": 32,
    "loss": 0.29848647117614746,
    "learning_rate": 0.00034848236518361393,
    "timestep_mean": 439.21875,
    "epsilon_norm": 7.55680513381958,
    "prediction_norm": 6.184719562530518,
    "timestamp": "2025-07-14T12:35:14.702697"
  },
  {
    "epoch": 2804,
    "batch_idx": 64,
    "loss": 0.2995946407318115,
    "learning_rate": 0.00034848236518361393,
    "timestep_mean": 472.09375,
    "epsilon_norm": 7.447648525238037,
    "prediction_norm": 6.150735855102539,
    "timestamp": "2025-07-14T12:35:14.707815"
  },
  {
    "epoch": 2804,
    "batch_idx": 96,
    "loss": 0.22596746683120728,
    "learning_rate": 0.00034848236518361393,
    "timestep_mean": 435.5625,
    "epsilon_norm": 7.101670265197754,
    "prediction_norm": 5.6608357429504395,
    "timestamp": "2025-07-14T12:35:14.712938"
  },
  {
    "epoch": 2804,
    "batch_idx": 128,
    "loss": 0.2580578327178955,
    "learning_rate": 0.00034848236518361393,
    "timestep_mean": 455.3125,
    "epsilon_norm": 7.390200614929199,
    "prediction_norm": 6.298523426055908,
    "timestamp": "2025-07-14T12:35:14.717918"
  },
  {
    "epoch": 2804,
    "batch_idx": 160,
    "loss": 0.22002816200256348,
    "learning_rate": 0.00034848236518361393,
    "timestep_mean": 547.75,
    "epsilon_norm": 7.733804702758789,
    "prediction_norm": 6.878361701965332,
    "timestamp": "2025-07-14T12:35:14.723109"
  },
  {
    "epoch": 2804,
    "batch_idx": 192,
    "loss": 0.0447658970952034,
    "learning_rate": 0.00034848236518361393,
    "timestep_mean": 658.625,
    "epsilon_norm": 2.683258056640625,
    "prediction_norm": 2.746967315673828,
    "timestamp": "2025-07-14T12:35:14.728824"
  },
  {
    "epoch": 2805,
    "batch_idx": 0,
    "loss": 0.1795501708984375,
    "learning_rate": 0.0003492310200277526,
    "timestep_mean": 508.15625,
    "epsilon_norm": 8.117691993713379,
    "prediction_norm": 7.114039421081543,
    "timestamp": "2025-07-14T12:35:14.734481"
  },
  {
    "epoch": 2805,
    "batch_idx": 32,
    "loss": 0.141860693693161,
    "learning_rate": 0.0003492310200277526,
    "timestep_mean": 502.8125,
    "epsilon_norm": 8.035770416259766,
    "prediction_norm": 7.500452995300293,
    "timestamp": "2025-07-14T12:35:14.739622"
  },
  {
    "epoch": 2805,
    "batch_idx": 64,
    "loss": 0.3555212616920471,
    "learning_rate": 0.0003492310200277526,
    "timestep_mean": 443.78125,
    "epsilon_norm": 7.519465923309326,
    "prediction_norm": 6.197906494140625,
    "timestamp": "2025-07-14T12:35:14.744660"
  },
  {
    "epoch": 2805,
    "batch_idx": 96,
    "loss": 0.2672494053840637,
    "learning_rate": 0.0003492310200277526,
    "timestep_mean": 534.875,
    "epsilon_norm": 8.158228874206543,
    "prediction_norm": 7.091128349304199,
    "timestamp": "2025-07-14T12:35:14.749583"
  },
  {
    "epoch": 2805,
    "batch_idx": 128,
    "loss": 0.28655770421028137,
    "learning_rate": 0.0003492310200277526,
    "timestep_mean": 568.5625,
    "epsilon_norm": 8.43848705291748,
    "prediction_norm": 7.357412338256836,
    "timestamp": "2025-07-14T12:35:14.754668"
  },
  {
    "epoch": 2805,
    "batch_idx": 160,
    "loss": 0.16853995621204376,
    "learning_rate": 0.0003492310200277526,
    "timestep_mean": 483.71875,
    "epsilon_norm": 5.849072456359863,
    "prediction_norm": 4.797786712646484,
    "timestamp": "2025-07-14T12:35:14.759798"
  },
  {
    "epoch": 2805,
    "batch_idx": 192,
    "loss": 0.21376022696495056,
    "learning_rate": 0.0003492310200277526,
    "timestep_mean": 560.25,
    "epsilon_norm": 4.544502258300781,
    "prediction_norm": 4.033228874206543,
    "timestamp": "2025-07-14T12:35:14.764994"
  },
  {
    "epoch": 2806,
    "batch_idx": 0,
    "loss": 0.27062010765075684,
    "learning_rate": 0.0003499800468793623,
    "timestep_mean": 482.09375,
    "epsilon_norm": 7.994347095489502,
    "prediction_norm": 5.618127346038818,
    "timestamp": "2025-07-14T12:35:14.770233"
  },
  {
    "epoch": 2806,
    "batch_idx": 32,
    "loss": 0.26307058334350586,
    "learning_rate": 0.0003499800468793623,
    "timestep_mean": 429.21875,
    "epsilon_norm": 7.81379508972168,
    "prediction_norm": 6.326585292816162,
    "timestamp": "2025-07-14T12:35:14.775550"
  },
  {
    "epoch": 2806,
    "batch_idx": 64,
    "loss": 0.21276789903640747,
    "learning_rate": 0.0003499800468793623,
    "timestep_mean": 517.84375,
    "epsilon_norm": 6.145001411437988,
    "prediction_norm": 4.855018615722656,
    "timestamp": "2025-07-14T12:35:14.780748"
  },
  {
    "epoch": 2806,
    "batch_idx": 96,
    "loss": 0.23669031262397766,
    "learning_rate": 0.0003499800468793623,
    "timestep_mean": 465.53125,
    "epsilon_norm": 8.005106925964355,
    "prediction_norm": 6.92160701751709,
    "timestamp": "2025-07-14T12:35:14.785860"
  },
  {
    "epoch": 2806,
    "batch_idx": 128,
    "loss": 0.19920487701892853,
    "learning_rate": 0.0003499800468793623,
    "timestep_mean": 601.28125,
    "epsilon_norm": 7.8397135734558105,
    "prediction_norm": 7.071554183959961,
    "timestamp": "2025-07-14T12:35:14.790738"
  },
  {
    "epoch": 2806,
    "batch_idx": 160,
    "loss": 0.2549861669540405,
    "learning_rate": 0.0003499800468793623,
    "timestep_mean": 525.28125,
    "epsilon_norm": 7.5525360107421875,
    "prediction_norm": 6.148519515991211,
    "timestamp": "2025-07-14T12:35:14.795731"
  },
  {
    "epoch": 2806,
    "batch_idx": 192,
    "loss": 0.42056167125701904,
    "learning_rate": 0.0003499800468793623,
    "timestep_mean": 226.375,
    "epsilon_norm": 4.386645793914795,
    "prediction_norm": 2.8946595191955566,
    "timestamp": "2025-07-14T12:35:14.801020"
  },
  {
    "epoch": 2807,
    "batch_idx": 0,
    "loss": 0.28088831901550293,
    "learning_rate": 0.0003507294438902937,
    "timestep_mean": 490.9375,
    "epsilon_norm": 7.879375457763672,
    "prediction_norm": 6.673208236694336,
    "timestamp": "2025-07-14T12:35:14.806173"
  },
  {
    "epoch": 2807,
    "batch_idx": 32,
    "loss": 0.3363831043243408,
    "learning_rate": 0.0003507294438902937,
    "timestep_mean": 503.15625,
    "epsilon_norm": 7.397973537445068,
    "prediction_norm": 6.176229476928711,
    "timestamp": "2025-07-14T12:35:14.811184"
  },
  {
    "epoch": 2807,
    "batch_idx": 64,
    "loss": 0.24957320094108582,
    "learning_rate": 0.0003507294438902937,
    "timestep_mean": 455.28125,
    "epsilon_norm": 8.523836135864258,
    "prediction_norm": 7.84724235534668,
    "timestamp": "2025-07-14T12:35:14.816266"
  },
  {
    "epoch": 2807,
    "batch_idx": 96,
    "loss": 0.22135500609874725,
    "learning_rate": 0.0003507294438902937,
    "timestep_mean": 576.59375,
    "epsilon_norm": 9.320240020751953,
    "prediction_norm": 8.095232963562012,
    "timestamp": "2025-07-14T12:35:14.821448"
  },
  {
    "epoch": 2807,
    "batch_idx": 128,
    "loss": 0.19106045365333557,
    "learning_rate": 0.0003507294438902937,
    "timestep_mean": 510.875,
    "epsilon_norm": 7.919178485870361,
    "prediction_norm": 7.395359516143799,
    "timestamp": "2025-07-14T12:35:14.826539"
  },
  {
    "epoch": 2807,
    "batch_idx": 160,
    "loss": 0.46372976899147034,
    "learning_rate": 0.0003507294438902937,
    "timestep_mean": 452.375,
    "epsilon_norm": 8.097421646118164,
    "prediction_norm": 5.660942077636719,
    "timestamp": "2025-07-14T12:35:14.832108"
  },
  {
    "epoch": 2807,
    "batch_idx": 192,
    "loss": 0.17634332180023193,
    "learning_rate": 0.0003507294438902937,
    "timestep_mean": 506.625,
    "epsilon_norm": 4.787675857543945,
    "prediction_norm": 4.699638843536377,
    "timestamp": "2025-07-14T12:35:14.837541"
  },
  {
    "epoch": 2808,
    "batch_idx": 0,
    "loss": 0.23551440238952637,
    "learning_rate": 0.00035147920921148323,
    "timestep_mean": 448.75,
    "epsilon_norm": 7.39199686050415,
    "prediction_norm": 6.528806686401367,
    "timestamp": "2025-07-14T12:35:14.842684"
  },
  {
    "epoch": 2808,
    "batch_idx": 32,
    "loss": 0.31285005807876587,
    "learning_rate": 0.00035147920921148323,
    "timestep_mean": 458.375,
    "epsilon_norm": 8.099437713623047,
    "prediction_norm": 6.313084125518799,
    "timestamp": "2025-07-14T12:35:14.847872"
  },
  {
    "epoch": 2808,
    "batch_idx": 64,
    "loss": 0.3164752721786499,
    "learning_rate": 0.00035147920921148323,
    "timestep_mean": 565.09375,
    "epsilon_norm": 8.85306167602539,
    "prediction_norm": 7.799820899963379,
    "timestamp": "2025-07-14T12:35:14.853017"
  },
  {
    "epoch": 2808,
    "batch_idx": 96,
    "loss": 0.15170817077159882,
    "learning_rate": 0.00035147920921148323,
    "timestep_mean": 571.46875,
    "epsilon_norm": 7.946270942687988,
    "prediction_norm": 6.9636759757995605,
    "timestamp": "2025-07-14T12:35:14.858150"
  },
  {
    "epoch": 2808,
    "batch_idx": 128,
    "loss": 0.20253774523735046,
    "learning_rate": 0.00035147920921148323,
    "timestep_mean": 394.53125,
    "epsilon_norm": 8.092151641845703,
    "prediction_norm": 6.99570894241333,
    "timestamp": "2025-07-14T12:35:14.863469"
  },
  {
    "epoch": 2808,
    "batch_idx": 160,
    "loss": 0.3128155469894409,
    "learning_rate": 0.00035147920921148323,
    "timestep_mean": 552.15625,
    "epsilon_norm": 6.775115489959717,
    "prediction_norm": 5.673010349273682,
    "timestamp": "2025-07-14T12:35:14.868577"
  },
  {
    "epoch": 2808,
    "batch_idx": 192,
    "loss": 0.2834392189979553,
    "learning_rate": 0.00035147920921148323,
    "timestep_mean": 591.875,
    "epsilon_norm": 5.2577009201049805,
    "prediction_norm": 4.99235200881958,
    "timestamp": "2025-07-14T12:35:14.873729"
  },
  {
    "epoch": 2809,
    "batch_idx": 0,
    "loss": 0.47012948989868164,
    "learning_rate": 0.0003522293409929601,
    "timestep_mean": 445.34375,
    "epsilon_norm": 9.993647575378418,
    "prediction_norm": 7.950355529785156,
    "timestamp": "2025-07-14T12:35:14.878932"
  },
  {
    "epoch": 2809,
    "batch_idx": 32,
    "loss": 0.19062519073486328,
    "learning_rate": 0.0003522293409929601,
    "timestep_mean": 564.15625,
    "epsilon_norm": 7.568249225616455,
    "prediction_norm": 7.546050071716309,
    "timestamp": "2025-07-14T12:35:14.884250"
  },
  {
    "epoch": 2809,
    "batch_idx": 64,
    "loss": 0.1876157969236374,
    "learning_rate": 0.0003522293409929601,
    "timestep_mean": 521.9375,
    "epsilon_norm": 8.544949531555176,
    "prediction_norm": 7.248082637786865,
    "timestamp": "2025-07-14T12:35:14.889293"
  },
  {
    "epoch": 2809,
    "batch_idx": 96,
    "loss": 0.41835951805114746,
    "learning_rate": 0.0003522293409929601,
    "timestep_mean": 518.46875,
    "epsilon_norm": 8.905050277709961,
    "prediction_norm": 7.174017906188965,
    "timestamp": "2025-07-14T12:35:14.894645"
  },
  {
    "epoch": 2809,
    "batch_idx": 128,
    "loss": 0.4245690107345581,
    "learning_rate": 0.0003522293409929601,
    "timestep_mean": 467.6875,
    "epsilon_norm": 8.969151496887207,
    "prediction_norm": 6.983208656311035,
    "timestamp": "2025-07-14T12:35:14.899842"
  },
  {
    "epoch": 2809,
    "batch_idx": 160,
    "loss": 0.3674353361129761,
    "learning_rate": 0.0003522293409929601,
    "timestep_mean": 536.65625,
    "epsilon_norm": 7.347067356109619,
    "prediction_norm": 5.726297855377197,
    "timestamp": "2025-07-14T12:35:14.904976"
  },
  {
    "epoch": 2809,
    "batch_idx": 192,
    "loss": 0.18660317361354828,
    "learning_rate": 0.0003522293409929601,
    "timestep_mean": 423.0,
    "epsilon_norm": 2.8758182525634766,
    "prediction_norm": 3.1496639251708984,
    "timestamp": "2025-07-14T12:35:14.910191"
  },
  {
    "epoch": 2810,
    "batch_idx": 0,
    "loss": 0.40982627868652344,
    "learning_rate": 0.00035297983738384907,
    "timestep_mean": 378.40625,
    "epsilon_norm": 8.674379348754883,
    "prediction_norm": 7.051327228546143,
    "timestamp": "2025-07-14T12:35:14.915431"
  },
  {
    "epoch": 2810,
    "batch_idx": 32,
    "loss": 0.14827275276184082,
    "learning_rate": 0.00035297983738384907,
    "timestep_mean": 433.5,
    "epsilon_norm": 7.503171443939209,
    "prediction_norm": 6.288136005401611,
    "timestamp": "2025-07-14T12:35:14.920496"
  },
  {
    "epoch": 2810,
    "batch_idx": 64,
    "loss": 0.1794632375240326,
    "learning_rate": 0.00035297983738384907,
    "timestep_mean": 570.625,
    "epsilon_norm": 8.482467651367188,
    "prediction_norm": 7.8345489501953125,
    "timestamp": "2025-07-14T12:35:14.925843"
  },
  {
    "epoch": 2810,
    "batch_idx": 96,
    "loss": 0.21684329211711884,
    "learning_rate": 0.00035297983738384907,
    "timestep_mean": 499.78125,
    "epsilon_norm": 8.677844047546387,
    "prediction_norm": 7.753745079040527,
    "timestamp": "2025-07-14T12:35:14.930979"
  },
  {
    "epoch": 2810,
    "batch_idx": 128,
    "loss": 0.33391720056533813,
    "learning_rate": 0.00035297983738384907,
    "timestep_mean": 502.34375,
    "epsilon_norm": 8.624882698059082,
    "prediction_norm": 7.642632484436035,
    "timestamp": "2025-07-14T12:35:14.936061"
  },
  {
    "epoch": 2810,
    "batch_idx": 160,
    "loss": 0.32327812910079956,
    "learning_rate": 0.00035297983738384907,
    "timestep_mean": 496.71875,
    "epsilon_norm": 8.50749397277832,
    "prediction_norm": 7.625040054321289,
    "timestamp": "2025-07-14T12:35:14.941256"
  },
  {
    "epoch": 2810,
    "batch_idx": 192,
    "loss": 0.053058892488479614,
    "learning_rate": 0.00035297983738384907,
    "timestep_mean": 586.125,
    "epsilon_norm": 3.119921922683716,
    "prediction_norm": 3.113903522491455,
    "timestamp": "2025-07-14T12:35:14.946426"
  },
  {
    "epoch": 2811,
    "batch_idx": 0,
    "loss": 0.28202542662620544,
    "learning_rate": 0.0003537306965323736,
    "timestep_mean": 591.96875,
    "epsilon_norm": 8.370672225952148,
    "prediction_norm": 7.062357425689697,
    "timestamp": "2025-07-14T12:35:14.951494"
  },
  {
    "epoch": 2811,
    "batch_idx": 32,
    "loss": 0.4133903384208679,
    "learning_rate": 0.0003537306965323736,
    "timestep_mean": 416.46875,
    "epsilon_norm": 7.967806816101074,
    "prediction_norm": 6.194218158721924,
    "timestamp": "2025-07-14T12:35:14.956657"
  },
  {
    "epoch": 2811,
    "batch_idx": 64,
    "loss": 0.2340202033519745,
    "learning_rate": 0.0003537306965323736,
    "timestep_mean": 544.28125,
    "epsilon_norm": 7.307920455932617,
    "prediction_norm": 6.034980773925781,
    "timestamp": "2025-07-14T12:35:14.961825"
  },
  {
    "epoch": 2811,
    "batch_idx": 96,
    "loss": 0.224557027220726,
    "learning_rate": 0.0003537306965323736,
    "timestep_mean": 554.25,
    "epsilon_norm": 7.1458282470703125,
    "prediction_norm": 6.804294109344482,
    "timestamp": "2025-07-14T12:35:14.966880"
  },
  {
    "epoch": 2811,
    "batch_idx": 128,
    "loss": 0.27327850461006165,
    "learning_rate": 0.0003537306965323736,
    "timestep_mean": 572.78125,
    "epsilon_norm": 8.027556419372559,
    "prediction_norm": 6.821296215057373,
    "timestamp": "2025-07-14T12:35:14.972103"
  },
  {
    "epoch": 2811,
    "batch_idx": 160,
    "loss": 0.2207106351852417,
    "learning_rate": 0.0003537306965323736,
    "timestep_mean": 433.875,
    "epsilon_norm": 8.197972297668457,
    "prediction_norm": 7.221701622009277,
    "timestamp": "2025-07-14T12:35:14.977217"
  },
  {
    "epoch": 2811,
    "batch_idx": 192,
    "loss": 0.6299564838409424,
    "learning_rate": 0.0003537306965323736,
    "timestep_mean": 353.125,
    "epsilon_norm": 4.241063594818115,
    "prediction_norm": 2.6110007762908936,
    "timestamp": "2025-07-14T12:35:14.982339"
  },
  {
    "epoch": 2812,
    "batch_idx": 0,
    "loss": 0.2157491147518158,
    "learning_rate": 0.0003544819165858647,
    "timestep_mean": 541.96875,
    "epsilon_norm": 8.350915908813477,
    "prediction_norm": 7.554378509521484,
    "timestamp": "2025-07-14T12:35:14.987662"
  },
  {
    "epoch": 2812,
    "batch_idx": 32,
    "loss": 0.2806887626647949,
    "learning_rate": 0.0003544819165858647,
    "timestep_mean": 482.96875,
    "epsilon_norm": 7.919801712036133,
    "prediction_norm": 6.771703720092773,
    "timestamp": "2025-07-14T12:35:14.992996"
  },
  {
    "epoch": 2812,
    "batch_idx": 64,
    "loss": 0.29351168870925903,
    "learning_rate": 0.0003544819165858647,
    "timestep_mean": 397.375,
    "epsilon_norm": 6.80569314956665,
    "prediction_norm": 4.839521884918213,
    "timestamp": "2025-07-14T12:35:14.997917"
  },
  {
    "epoch": 2812,
    "batch_idx": 96,
    "loss": 0.2993026375770569,
    "learning_rate": 0.0003544819165858647,
    "timestep_mean": 518.59375,
    "epsilon_norm": 7.925017833709717,
    "prediction_norm": 6.693542003631592,
    "timestamp": "2025-07-14T12:35:15.002799"
  },
  {
    "epoch": 2812,
    "batch_idx": 128,
    "loss": 0.13997404277324677,
    "learning_rate": 0.0003544819165858647,
    "timestep_mean": 567.53125,
    "epsilon_norm": 8.005254745483398,
    "prediction_norm": 7.0185956954956055,
    "timestamp": "2025-07-14T12:35:15.007636"
  },
  {
    "epoch": 2812,
    "batch_idx": 160,
    "loss": 0.28980565071105957,
    "learning_rate": 0.0003544819165858647,
    "timestep_mean": 565.625,
    "epsilon_norm": 7.693258762359619,
    "prediction_norm": 6.518763065338135,
    "timestamp": "2025-07-14T12:35:15.012474"
  },
  {
    "epoch": 2812,
    "batch_idx": 192,
    "loss": 0.2390947937965393,
    "learning_rate": 0.0003544819165858647,
    "timestep_mean": 368.875,
    "epsilon_norm": 3.1443326473236084,
    "prediction_norm": 2.5237481594085693,
    "timestamp": "2025-07-14T12:35:15.017495"
  },
  {
    "epoch": 2813,
    "batch_idx": 0,
    "loss": 0.31940338015556335,
    "learning_rate": 0.00035523349569076094,
    "timestep_mean": 578.34375,
    "epsilon_norm": 7.388627052307129,
    "prediction_norm": 6.513986587524414,
    "timestamp": "2025-07-14T12:35:15.022330"
  },
  {
    "epoch": 2813,
    "batch_idx": 32,
    "loss": 0.23406276106834412,
    "learning_rate": 0.00035523349569076094,
    "timestep_mean": 455.28125,
    "epsilon_norm": 6.136587619781494,
    "prediction_norm": 4.575267791748047,
    "timestamp": "2025-07-14T12:35:15.027230"
  },
  {
    "epoch": 2813,
    "batch_idx": 64,
    "loss": 0.33707594871520996,
    "learning_rate": 0.00035523349569076094,
    "timestep_mean": 526.5,
    "epsilon_norm": 9.089438438415527,
    "prediction_norm": 7.336996555328369,
    "timestamp": "2025-07-14T12:35:15.032143"
  },
  {
    "epoch": 2813,
    "batch_idx": 96,
    "loss": 0.25849515199661255,
    "learning_rate": 0.00035523349569076094,
    "timestep_mean": 491.15625,
    "epsilon_norm": 7.7866997718811035,
    "prediction_norm": 6.89916467666626,
    "timestamp": "2025-07-14T12:35:15.036995"
  },
  {
    "epoch": 2813,
    "batch_idx": 128,
    "loss": 0.23071640729904175,
    "learning_rate": 0.00035523349569076094,
    "timestep_mean": 515.5,
    "epsilon_norm": 6.85491943359375,
    "prediction_norm": 5.72451114654541,
    "timestamp": "2025-07-14T12:35:15.041861"
  },
  {
    "epoch": 2813,
    "batch_idx": 160,
    "loss": 0.33182859420776367,
    "learning_rate": 0.00035523349569076094,
    "timestep_mean": 443.25,
    "epsilon_norm": 7.930274486541748,
    "prediction_norm": 6.285384654998779,
    "timestamp": "2025-07-14T12:35:15.046834"
  },
  {
    "epoch": 2813,
    "batch_idx": 192,
    "loss": 0.2901065945625305,
    "learning_rate": 0.00035523349569076094,
    "timestep_mean": 596.875,
    "epsilon_norm": 4.320590019226074,
    "prediction_norm": 3.179588556289673,
    "timestamp": "2025-07-14T12:35:15.051700"
  },
  {
    "epoch": 2814,
    "batch_idx": 0,
    "loss": 0.3053085505962372,
    "learning_rate": 0.0003559854319926163,
    "timestep_mean": 453.0625,
    "epsilon_norm": 7.337045192718506,
    "prediction_norm": 6.006092071533203,
    "timestamp": "2025-07-14T12:35:15.056549"
  },
  {
    "epoch": 2814,
    "batch_idx": 32,
    "loss": 0.6708931922912598,
    "learning_rate": 0.0003559854319926163,
    "timestep_mean": 402.0,
    "epsilon_norm": 8.66122817993164,
    "prediction_norm": 6.023171901702881,
    "timestamp": "2025-07-14T12:35:15.061462"
  },
  {
    "epoch": 2814,
    "batch_idx": 64,
    "loss": 0.37446409463882446,
    "learning_rate": 0.0003559854319926163,
    "timestep_mean": 518.71875,
    "epsilon_norm": 8.353815078735352,
    "prediction_norm": 5.94683837890625,
    "timestamp": "2025-07-14T12:35:15.066367"
  },
  {
    "epoch": 2814,
    "batch_idx": 96,
    "loss": 0.2128431499004364,
    "learning_rate": 0.0003559854319926163,
    "timestep_mean": 539.09375,
    "epsilon_norm": 9.204203605651855,
    "prediction_norm": 8.143853187561035,
    "timestamp": "2025-07-14T12:35:15.071457"
  },
  {
    "epoch": 2814,
    "batch_idx": 128,
    "loss": 0.21659345924854279,
    "learning_rate": 0.0003559854319926163,
    "timestep_mean": 493.3125,
    "epsilon_norm": 8.76522159576416,
    "prediction_norm": 7.752983093261719,
    "timestamp": "2025-07-14T12:35:15.076409"
  },
  {
    "epoch": 2814,
    "batch_idx": 160,
    "loss": 0.35588863492012024,
    "learning_rate": 0.0003559854319926163,
    "timestep_mean": 546.96875,
    "epsilon_norm": 8.360721588134766,
    "prediction_norm": 6.519289970397949,
    "timestamp": "2025-07-14T12:35:15.081383"
  },
  {
    "epoch": 2814,
    "batch_idx": 192,
    "loss": 0.226547971367836,
    "learning_rate": 0.0003559854319926163,
    "timestep_mean": 527.25,
    "epsilon_norm": 3.7553791999816895,
    "prediction_norm": 3.1177821159362793,
    "timestamp": "2025-07-14T12:35:15.086327"
  },
  {
    "epoch": 2815,
    "batch_idx": 0,
    "loss": 0.14488936960697174,
    "learning_rate": 0.00035673772363610165,
    "timestep_mean": 601.125,
    "epsilon_norm": 7.649186611175537,
    "prediction_norm": 7.121052265167236,
    "timestamp": "2025-07-14T12:35:15.091259"
  },
  {
    "epoch": 2815,
    "batch_idx": 32,
    "loss": 0.2101650834083557,
    "learning_rate": 0.00035673772363610165,
    "timestep_mean": 589.15625,
    "epsilon_norm": 7.147970199584961,
    "prediction_norm": 6.4216203689575195,
    "timestamp": "2025-07-14T12:35:15.096226"
  },
  {
    "epoch": 2815,
    "batch_idx": 64,
    "loss": 0.20512285828590393,
    "learning_rate": 0.00035673772363610165,
    "timestep_mean": 442.5625,
    "epsilon_norm": 7.602263450622559,
    "prediction_norm": 6.475302696228027,
    "timestamp": "2025-07-14T12:35:15.101180"
  },
  {
    "epoch": 2815,
    "batch_idx": 96,
    "loss": 0.13250119984149933,
    "learning_rate": 0.00035673772363610165,
    "timestep_mean": 586.15625,
    "epsilon_norm": 8.11543083190918,
    "prediction_norm": 7.82676887512207,
    "timestamp": "2025-07-14T12:35:15.106073"
  },
  {
    "epoch": 2815,
    "batch_idx": 128,
    "loss": 0.2285434454679489,
    "learning_rate": 0.00035673772363610165,
    "timestep_mean": 494.15625,
    "epsilon_norm": 8.504719734191895,
    "prediction_norm": 7.369950771331787,
    "timestamp": "2025-07-14T12:35:15.110977"
  },
  {
    "epoch": 2815,
    "batch_idx": 160,
    "loss": 0.22918999195098877,
    "learning_rate": 0.00035673772363610165,
    "timestep_mean": 429.84375,
    "epsilon_norm": 7.818742275238037,
    "prediction_norm": 7.389353275299072,
    "timestamp": "2025-07-14T12:35:15.116066"
  },
  {
    "epoch": 2815,
    "batch_idx": 192,
    "loss": 1.0630004405975342,
    "learning_rate": 0.00035673772363610165,
    "timestep_mean": 427.875,
    "epsilon_norm": 5.373281478881836,
    "prediction_norm": 3.375577449798584,
    "timestamp": "2025-07-14T12:35:15.121059"
  },
  {
    "epoch": 2816,
    "batch_idx": 0,
    "loss": 0.19744685292243958,
    "learning_rate": 0.00035749036876501225,
    "timestep_mean": 532.21875,
    "epsilon_norm": 8.488322257995605,
    "prediction_norm": 7.530005931854248,
    "timestamp": "2025-07-14T12:35:15.126068"
  },
  {
    "epoch": 2816,
    "batch_idx": 32,
    "loss": 0.3076416850090027,
    "learning_rate": 0.00035749036876501225,
    "timestep_mean": 463.625,
    "epsilon_norm": 6.693168640136719,
    "prediction_norm": 5.994275093078613,
    "timestamp": "2025-07-14T12:35:15.131012"
  },
  {
    "epoch": 2816,
    "batch_idx": 64,
    "loss": 0.27608412504196167,
    "learning_rate": 0.00035749036876501225,
    "timestep_mean": 460.21875,
    "epsilon_norm": 8.450093269348145,
    "prediction_norm": 8.056403160095215,
    "timestamp": "2025-07-14T12:35:15.135932"
  },
  {
    "epoch": 2816,
    "batch_idx": 96,
    "loss": 0.29396793246269226,
    "learning_rate": 0.00035749036876501225,
    "timestep_mean": 481.5,
    "epsilon_norm": 7.619867324829102,
    "prediction_norm": 6.692037105560303,
    "timestamp": "2025-07-14T12:35:15.140880"
  },
  {
    "epoch": 2816,
    "batch_idx": 128,
    "loss": 0.0924619510769844,
    "learning_rate": 0.00035749036876501225,
    "timestep_mean": 581.09375,
    "epsilon_norm": 7.026611804962158,
    "prediction_norm": 7.062796115875244,
    "timestamp": "2025-07-14T12:35:15.146015"
  },
  {
    "epoch": 2816,
    "batch_idx": 160,
    "loss": 0.19652758538722992,
    "learning_rate": 0.00035749036876501225,
    "timestep_mean": 475.53125,
    "epsilon_norm": 8.32120132446289,
    "prediction_norm": 8.021967887878418,
    "timestamp": "2025-07-14T12:35:15.151006"
  },
  {
    "epoch": 2816,
    "batch_idx": 192,
    "loss": 0.40247777104377747,
    "learning_rate": 0.00035749036876501225,
    "timestep_mean": 458.25,
    "epsilon_norm": 4.360269546508789,
    "prediction_norm": 3.684943914413452,
    "timestamp": "2025-07-14T12:35:15.155990"
  },
  {
    "epoch": 2817,
    "batch_idx": 0,
    "loss": 0.3456301689147949,
    "learning_rate": 0.00035824336552227234,
    "timestep_mean": 462.4375,
    "epsilon_norm": 7.739124298095703,
    "prediction_norm": 5.863372802734375,
    "timestamp": "2025-07-14T12:35:15.160871"
  },
  {
    "epoch": 2817,
    "batch_idx": 32,
    "loss": 0.20005479454994202,
    "learning_rate": 0.00035824336552227234,
    "timestep_mean": 567.40625,
    "epsilon_norm": 7.59184455871582,
    "prediction_norm": 6.847257137298584,
    "timestamp": "2025-07-14T12:35:15.165730"
  },
  {
    "epoch": 2817,
    "batch_idx": 64,
    "loss": 0.21259388327598572,
    "learning_rate": 0.00035824336552227234,
    "timestep_mean": 566.625,
    "epsilon_norm": 7.854424476623535,
    "prediction_norm": 6.589629173278809,
    "timestamp": "2025-07-14T12:35:15.170645"
  },
  {
    "epoch": 2817,
    "batch_idx": 96,
    "loss": 0.3348143994808197,
    "learning_rate": 0.00035824336552227234,
    "timestep_mean": 507.28125,
    "epsilon_norm": 7.810064792633057,
    "prediction_norm": 6.150611877441406,
    "timestamp": "2025-07-14T12:35:15.175592"
  },
  {
    "epoch": 2817,
    "batch_idx": 128,
    "loss": 0.22616112232208252,
    "learning_rate": 0.00035824336552227234,
    "timestep_mean": 465.875,
    "epsilon_norm": 8.905694961547852,
    "prediction_norm": 7.441643238067627,
    "timestamp": "2025-07-14T12:35:15.180469"
  },
  {
    "epoch": 2817,
    "batch_idx": 160,
    "loss": 0.39107227325439453,
    "learning_rate": 0.00035824336552227234,
    "timestep_mean": 439.90625,
    "epsilon_norm": 7.346696376800537,
    "prediction_norm": 4.990085601806641,
    "timestamp": "2025-07-14T12:35:15.185402"
  },
  {
    "epoch": 2817,
    "batch_idx": 192,
    "loss": 0.3123658299446106,
    "learning_rate": 0.00035824336552227234,
    "timestep_mean": 415.875,
    "epsilon_norm": 4.321906566619873,
    "prediction_norm": 2.858149766921997,
    "timestamp": "2025-07-14T12:35:15.190238"
  },
  {
    "epoch": 2818,
    "batch_idx": 0,
    "loss": 0.3367433249950409,
    "learning_rate": 0.000358996712049936,
    "timestep_mean": 506.03125,
    "epsilon_norm": 7.891545295715332,
    "prediction_norm": 6.561638832092285,
    "timestamp": "2025-07-14T12:35:15.195246"
  },
  {
    "epoch": 2818,
    "batch_idx": 32,
    "loss": 0.3514619767665863,
    "learning_rate": 0.000358996712049936,
    "timestep_mean": 459.5,
    "epsilon_norm": 7.355288505554199,
    "prediction_norm": 5.365832805633545,
    "timestamp": "2025-07-14T12:35:15.200092"
  },
  {
    "epoch": 2818,
    "batch_idx": 64,
    "loss": 0.43104490637779236,
    "learning_rate": 0.000358996712049936,
    "timestep_mean": 505.21875,
    "epsilon_norm": 8.186799049377441,
    "prediction_norm": 6.360253810882568,
    "timestamp": "2025-07-14T12:35:15.205135"
  },
  {
    "epoch": 2818,
    "batch_idx": 96,
    "loss": 0.2853888273239136,
    "learning_rate": 0.000358996712049936,
    "timestep_mean": 451.46875,
    "epsilon_norm": 7.9320759773254395,
    "prediction_norm": 6.216497898101807,
    "timestamp": "2025-07-14T12:35:15.209936"
  },
  {
    "epoch": 2818,
    "batch_idx": 128,
    "loss": 0.2682369351387024,
    "learning_rate": 0.000358996712049936,
    "timestep_mean": 570.9375,
    "epsilon_norm": 8.667231559753418,
    "prediction_norm": 7.225162029266357,
    "timestamp": "2025-07-14T12:35:15.214762"
  },
  {
    "epoch": 2818,
    "batch_idx": 160,
    "loss": 0.2742154896259308,
    "learning_rate": 0.000358996712049936,
    "timestep_mean": 554.0625,
    "epsilon_norm": 6.67658805847168,
    "prediction_norm": 5.897466659545898,
    "timestamp": "2025-07-14T12:35:15.219552"
  },
  {
    "epoch": 2818,
    "batch_idx": 192,
    "loss": 0.32070392370224,
    "learning_rate": 0.000358996712049936,
    "timestep_mean": 514.0,
    "epsilon_norm": 4.431947231292725,
    "prediction_norm": 3.6501052379608154,
    "timestamp": "2025-07-14T12:35:15.224394"
  },
  {
    "epoch": 2819,
    "batch_idx": 0,
    "loss": 0.32735663652420044,
    "learning_rate": 0.00035975040648919586,
    "timestep_mean": 489.59375,
    "epsilon_norm": 7.4924092292785645,
    "prediction_norm": 5.653780460357666,
    "timestamp": "2025-07-14T12:35:15.229251"
  },
  {
    "epoch": 2819,
    "batch_idx": 32,
    "loss": 0.4671667218208313,
    "learning_rate": 0.00035975040648919586,
    "timestep_mean": 521.875,
    "epsilon_norm": 8.210251808166504,
    "prediction_norm": 5.594256401062012,
    "timestamp": "2025-07-14T12:35:15.234103"
  },
  {
    "epoch": 2819,
    "batch_idx": 64,
    "loss": 0.201450914144516,
    "learning_rate": 0.00035975040648919586,
    "timestep_mean": 506.34375,
    "epsilon_norm": 6.7548956871032715,
    "prediction_norm": 6.212207317352295,
    "timestamp": "2025-07-14T12:35:15.239009"
  },
  {
    "epoch": 2819,
    "batch_idx": 96,
    "loss": 0.205514594912529,
    "learning_rate": 0.00035975040648919586,
    "timestep_mean": 455.8125,
    "epsilon_norm": 8.326798439025879,
    "prediction_norm": 7.591692924499512,
    "timestamp": "2025-07-14T12:35:15.243865"
  },
  {
    "epoch": 2819,
    "batch_idx": 128,
    "loss": 0.43403559923171997,
    "learning_rate": 0.00035975040648919586,
    "timestep_mean": 515.59375,
    "epsilon_norm": 8.39551067352295,
    "prediction_norm": 6.610182762145996,
    "timestamp": "2025-07-14T12:35:15.248788"
  },
  {
    "epoch": 2819,
    "batch_idx": 160,
    "loss": 0.29221516847610474,
    "learning_rate": 0.00035975040648919586,
    "timestep_mean": 406.09375,
    "epsilon_norm": 7.315358638763428,
    "prediction_norm": 5.885079860687256,
    "timestamp": "2025-07-14T12:35:15.253619"
  },
  {
    "epoch": 2819,
    "batch_idx": 192,
    "loss": 0.46031424403190613,
    "learning_rate": 0.00035975040648919586,
    "timestep_mean": 512.5,
    "epsilon_norm": 5.0091400146484375,
    "prediction_norm": 3.9810843467712402,
    "timestamp": "2025-07-14T12:35:15.258483"
  },
  {
    "epoch": 2820,
    "batch_idx": 0,
    "loss": 0.16315720975399017,
    "learning_rate": 0.00036050444698038596,
    "timestep_mean": 565.96875,
    "epsilon_norm": 7.0027995109558105,
    "prediction_norm": 6.264611721038818,
    "timestamp": "2025-07-14T12:35:15.263760"
  },
  {
    "epoch": 2820,
    "batch_idx": 32,
    "loss": 0.12933263182640076,
    "learning_rate": 0.00036050444698038596,
    "timestep_mean": 601.65625,
    "epsilon_norm": 7.910480976104736,
    "prediction_norm": 7.28444242477417,
    "timestamp": "2025-07-14T12:35:15.268883"
  },
  {
    "epoch": 2820,
    "batch_idx": 64,
    "loss": 0.28555697202682495,
    "learning_rate": 0.00036050444698038596,
    "timestep_mean": 450.15625,
    "epsilon_norm": 7.976048469543457,
    "prediction_norm": 6.6108880043029785,
    "timestamp": "2025-07-14T12:35:15.274009"
  },
  {
    "epoch": 2820,
    "batch_idx": 96,
    "loss": 0.2008906453847885,
    "learning_rate": 0.00036050444698038596,
    "timestep_mean": 527.75,
    "epsilon_norm": 7.621091365814209,
    "prediction_norm": 6.591189861297607,
    "timestamp": "2025-07-14T12:35:15.279130"
  },
  {
    "epoch": 2820,
    "batch_idx": 128,
    "loss": 0.2144896239042282,
    "learning_rate": 0.00036050444698038596,
    "timestep_mean": 516.3125,
    "epsilon_norm": 6.7293572425842285,
    "prediction_norm": 6.060027122497559,
    "timestamp": "2025-07-14T12:35:15.284373"
  },
  {
    "epoch": 2820,
    "batch_idx": 160,
    "loss": 0.1971498429775238,
    "learning_rate": 0.00036050444698038596,
    "timestep_mean": 464.15625,
    "epsilon_norm": 6.209665298461914,
    "prediction_norm": 5.921021938323975,
    "timestamp": "2025-07-14T12:35:15.289690"
  },
  {
    "epoch": 2820,
    "batch_idx": 192,
    "loss": 0.2793724536895752,
    "learning_rate": 0.00036050444698038596,
    "timestep_mean": 503.25,
    "epsilon_norm": 3.20853853225708,
    "prediction_norm": 3.0739269256591797,
    "timestamp": "2025-07-14T12:35:15.294706"
  },
  {
    "epoch": 2821,
    "batch_idx": 0,
    "loss": 0.3233935236930847,
    "learning_rate": 0.00036125883166298673,
    "timestep_mean": 495.15625,
    "epsilon_norm": 8.899632453918457,
    "prediction_norm": 7.466492176055908,
    "timestamp": "2025-07-14T12:35:15.299870"
  },
  {
    "epoch": 2821,
    "batch_idx": 32,
    "loss": 0.260248601436615,
    "learning_rate": 0.00036125883166298673,
    "timestep_mean": 496.875,
    "epsilon_norm": 7.970520973205566,
    "prediction_norm": 6.5938191413879395,
    "timestamp": "2025-07-14T12:35:15.305067"
  },
  {
    "epoch": 2821,
    "batch_idx": 64,
    "loss": 0.1933913677930832,
    "learning_rate": 0.00036125883166298673,
    "timestep_mean": 464.5,
    "epsilon_norm": 6.6796417236328125,
    "prediction_norm": 5.837567329406738,
    "timestamp": "2025-07-14T12:35:15.310347"
  },
  {
    "epoch": 2821,
    "batch_idx": 96,
    "loss": 0.29163381457328796,
    "learning_rate": 0.00036125883166298673,
    "timestep_mean": 456.5625,
    "epsilon_norm": 7.607641220092773,
    "prediction_norm": 6.267848491668701,
    "timestamp": "2025-07-14T12:35:15.315508"
  },
  {
    "epoch": 2821,
    "batch_idx": 128,
    "loss": 0.27299052476882935,
    "learning_rate": 0.00036125883166298673,
    "timestep_mean": 504.21875,
    "epsilon_norm": 8.380390167236328,
    "prediction_norm": 7.562715530395508,
    "timestamp": "2025-07-14T12:35:15.320728"
  },
  {
    "epoch": 2821,
    "batch_idx": 160,
    "loss": 0.3382997512817383,
    "learning_rate": 0.00036125883166298673,
    "timestep_mean": 502.09375,
    "epsilon_norm": 6.7928643226623535,
    "prediction_norm": 5.808373928070068,
    "timestamp": "2025-07-14T12:35:15.325731"
  },
  {
    "epoch": 2821,
    "batch_idx": 192,
    "loss": 0.14135697484016418,
    "learning_rate": 0.00036125883166298673,
    "timestep_mean": 741.125,
    "epsilon_norm": 2.9591610431671143,
    "prediction_norm": 2.6860668659210205,
    "timestamp": "2025-07-14T12:35:15.330825"
  },
  {
    "epoch": 2822,
    "batch_idx": 0,
    "loss": 0.31646108627319336,
    "learning_rate": 0.00036201355867562774,
    "timestep_mean": 506.90625,
    "epsilon_norm": 7.9810709953308105,
    "prediction_norm": 6.553152561187744,
    "timestamp": "2025-07-14T12:35:15.335844"
  },
  {
    "epoch": 2822,
    "batch_idx": 32,
    "loss": 0.41123124957084656,
    "learning_rate": 0.00036201355867562774,
    "timestep_mean": 389.6875,
    "epsilon_norm": 7.684757232666016,
    "prediction_norm": 6.261230945587158,
    "timestamp": "2025-07-14T12:35:15.340957"
  },
  {
    "epoch": 2822,
    "batch_idx": 64,
    "loss": 0.1454055905342102,
    "learning_rate": 0.00036201355867562774,
    "timestep_mean": 467.53125,
    "epsilon_norm": 7.10439920425415,
    "prediction_norm": 6.088613510131836,
    "timestamp": "2025-07-14T12:35:15.346154"
  },
  {
    "epoch": 2822,
    "batch_idx": 96,
    "loss": 0.1649501621723175,
    "learning_rate": 0.00036201355867562774,
    "timestep_mean": 538.25,
    "epsilon_norm": 6.751548767089844,
    "prediction_norm": 6.0324578285217285,
    "timestamp": "2025-07-14T12:35:15.351226"
  },
  {
    "epoch": 2822,
    "batch_idx": 128,
    "loss": 0.17940181493759155,
    "learning_rate": 0.00036201355867562774,
    "timestep_mean": 488.34375,
    "epsilon_norm": 7.666299343109131,
    "prediction_norm": 7.19480562210083,
    "timestamp": "2025-07-14T12:35:15.356464"
  },
  {
    "epoch": 2822,
    "batch_idx": 160,
    "loss": 0.24031062424182892,
    "learning_rate": 0.00036201355867562774,
    "timestep_mean": 511.96875,
    "epsilon_norm": 6.77171516418457,
    "prediction_norm": 5.73974609375,
    "timestamp": "2025-07-14T12:35:15.361650"
  },
  {
    "epoch": 2822,
    "batch_idx": 192,
    "loss": 0.12469466775655746,
    "learning_rate": 0.00036201355867562774,
    "timestep_mean": 352.0,
    "epsilon_norm": 2.7372021675109863,
    "prediction_norm": 2.3709967136383057,
    "timestamp": "2025-07-14T12:35:15.366662"
  },
  {
    "epoch": 2823,
    "batch_idx": 0,
    "loss": 0.19573360681533813,
    "learning_rate": 0.0003627686261560962,
    "timestep_mean": 514.0625,
    "epsilon_norm": 7.446896553039551,
    "prediction_norm": 6.768309593200684,
    "timestamp": "2025-07-14T12:35:15.371961"
  },
  {
    "epoch": 2823,
    "batch_idx": 32,
    "loss": 0.13461366295814514,
    "learning_rate": 0.0003627686261560962,
    "timestep_mean": 524.125,
    "epsilon_norm": 6.720388889312744,
    "prediction_norm": 6.524646759033203,
    "timestamp": "2025-07-14T12:35:15.377064"
  },
  {
    "epoch": 2823,
    "batch_idx": 64,
    "loss": 0.2259034961462021,
    "learning_rate": 0.0003627686261560962,
    "timestep_mean": 524.9375,
    "epsilon_norm": 7.542304515838623,
    "prediction_norm": 6.713181495666504,
    "timestamp": "2025-07-14T12:35:15.382251"
  },
  {
    "epoch": 2823,
    "batch_idx": 96,
    "loss": 0.26091068983078003,
    "learning_rate": 0.0003627686261560962,
    "timestep_mean": 463.40625,
    "epsilon_norm": 8.964670181274414,
    "prediction_norm": 7.837019443511963,
    "timestamp": "2025-07-14T12:35:15.387434"
  },
  {
    "epoch": 2823,
    "batch_idx": 128,
    "loss": 0.13074888288974762,
    "learning_rate": 0.0003627686261560962,
    "timestep_mean": 577.4375,
    "epsilon_norm": 8.778911590576172,
    "prediction_norm": 8.00892448425293,
    "timestamp": "2025-07-14T12:35:15.392759"
  },
  {
    "epoch": 2823,
    "batch_idx": 160,
    "loss": 0.1888250857591629,
    "learning_rate": 0.0003627686261560962,
    "timestep_mean": 554.0625,
    "epsilon_norm": 7.955528736114502,
    "prediction_norm": 7.0276994705200195,
    "timestamp": "2025-07-14T12:35:15.397869"
  },
  {
    "epoch": 2823,
    "batch_idx": 192,
    "loss": 0.05202432721853256,
    "learning_rate": 0.0003627686261560962,
    "timestep_mean": 587.125,
    "epsilon_norm": 4.161458969116211,
    "prediction_norm": 4.1450419425964355,
    "timestamp": "2025-07-14T12:35:15.402995"
  },
  {
    "epoch": 2824,
    "batch_idx": 0,
    "loss": 0.22682085633277893,
    "learning_rate": 0.0003635240322413383,
    "timestep_mean": 406.3125,
    "epsilon_norm": 8.939107894897461,
    "prediction_norm": 7.480410575866699,
    "timestamp": "2025-07-14T12:35:15.407945"
  },
  {
    "epoch": 2824,
    "batch_idx": 32,
    "loss": 0.3552604913711548,
    "learning_rate": 0.0003635240322413383,
    "timestep_mean": 466.34375,
    "epsilon_norm": 8.390230178833008,
    "prediction_norm": 6.59552526473999,
    "timestamp": "2025-07-14T12:35:15.412812"
  },
  {
    "epoch": 2824,
    "batch_idx": 64,
    "loss": 0.2552470266819,
    "learning_rate": 0.0003635240322413383,
    "timestep_mean": 535.96875,
    "epsilon_norm": 7.992809772491455,
    "prediction_norm": 6.447845935821533,
    "timestamp": "2025-07-14T12:35:15.418078"
  },
  {
    "epoch": 2824,
    "batch_idx": 96,
    "loss": 0.2898545563220978,
    "learning_rate": 0.0003635240322413383,
    "timestep_mean": 512.90625,
    "epsilon_norm": 7.65742826461792,
    "prediction_norm": 5.74272346496582,
    "timestamp": "2025-07-14T12:35:15.423359"
  },
  {
    "epoch": 2824,
    "batch_idx": 128,
    "loss": 0.32992982864379883,
    "learning_rate": 0.0003635240322413383,
    "timestep_mean": 448.9375,
    "epsilon_norm": 8.080282211303711,
    "prediction_norm": 6.128116130828857,
    "timestamp": "2025-07-14T12:35:15.428496"
  },
  {
    "epoch": 2824,
    "batch_idx": 160,
    "loss": 0.17920425534248352,
    "learning_rate": 0.0003635240322413383,
    "timestep_mean": 512.25,
    "epsilon_norm": 7.755082607269287,
    "prediction_norm": 6.657456398010254,
    "timestamp": "2025-07-14T12:35:15.433765"
  },
  {
    "epoch": 2824,
    "batch_idx": 192,
    "loss": 0.27617165446281433,
    "learning_rate": 0.0003635240322413383,
    "timestep_mean": 540.125,
    "epsilon_norm": 4.1219282150268555,
    "prediction_norm": 3.267077922821045,
    "timestamp": "2025-07-14T12:35:15.438897"
  },
  {
    "epoch": 2825,
    "batch_idx": 0,
    "loss": 0.1854557991027832,
    "learning_rate": 0.0003642797750674638,
    "timestep_mean": 539.3125,
    "epsilon_norm": 8.783342361450195,
    "prediction_norm": 7.727192401885986,
    "timestamp": "2025-07-14T12:35:15.444150"
  },
  {
    "epoch": 2825,
    "batch_idx": 32,
    "loss": 0.29621565341949463,
    "learning_rate": 0.0003642797750674638,
    "timestep_mean": 491.5,
    "epsilon_norm": 7.741011142730713,
    "prediction_norm": 6.707038879394531,
    "timestamp": "2025-07-14T12:35:15.449393"
  },
  {
    "epoch": 2825,
    "batch_idx": 64,
    "loss": 0.38869258761405945,
    "learning_rate": 0.0003642797750674638,
    "timestep_mean": 463.25,
    "epsilon_norm": 8.19125747680664,
    "prediction_norm": 6.573586940765381,
    "timestamp": "2025-07-14T12:35:15.454570"
  },
  {
    "epoch": 2825,
    "batch_idx": 96,
    "loss": 0.07176858186721802,
    "learning_rate": 0.0003642797750674638,
    "timestep_mean": 557.21875,
    "epsilon_norm": 7.458338260650635,
    "prediction_norm": 6.783841609954834,
    "timestamp": "2025-07-14T12:35:15.459679"
  },
  {
    "epoch": 2825,
    "batch_idx": 128,
    "loss": 0.307078093290329,
    "learning_rate": 0.0003642797750674638,
    "timestep_mean": 567.21875,
    "epsilon_norm": 8.543668746948242,
    "prediction_norm": 6.716522693634033,
    "timestamp": "2025-07-14T12:35:15.465006"
  },
  {
    "epoch": 2825,
    "batch_idx": 160,
    "loss": 0.20418381690979004,
    "learning_rate": 0.0003642797750674638,
    "timestep_mean": 524.25,
    "epsilon_norm": 6.4760565757751465,
    "prediction_norm": 5.584286212921143,
    "timestamp": "2025-07-14T12:35:15.470143"
  },
  {
    "epoch": 2825,
    "batch_idx": 192,
    "loss": 0.26885098218917847,
    "learning_rate": 0.0003642797750674638,
    "timestep_mean": 601.0,
    "epsilon_norm": 3.532167673110962,
    "prediction_norm": 2.9789159297943115,
    "timestamp": "2025-07-14T12:35:15.475407"
  },
  {
    "epoch": 2826,
    "batch_idx": 0,
    "loss": 0.25771021842956543,
    "learning_rate": 0.00036503585276975264,
    "timestep_mean": 518.75,
    "epsilon_norm": 8.32208251953125,
    "prediction_norm": 7.265315532684326,
    "timestamp": "2025-07-14T12:35:15.480678"
  },
  {
    "epoch": 2826,
    "batch_idx": 32,
    "loss": 0.2463449239730835,
    "learning_rate": 0.00036503585276975264,
    "timestep_mean": 525.3125,
    "epsilon_norm": 7.667903900146484,
    "prediction_norm": 6.833347797393799,
    "timestamp": "2025-07-14T12:35:15.485898"
  },
  {
    "epoch": 2826,
    "batch_idx": 64,
    "loss": 0.2581908404827118,
    "learning_rate": 0.00036503585276975264,
    "timestep_mean": 423.53125,
    "epsilon_norm": 7.420419692993164,
    "prediction_norm": 5.931341171264648,
    "timestamp": "2025-07-14T12:35:15.490932"
  },
  {
    "epoch": 2826,
    "batch_idx": 96,
    "loss": 0.12414959818124771,
    "learning_rate": 0.00036503585276975264,
    "timestep_mean": 594.1875,
    "epsilon_norm": 8.048896789550781,
    "prediction_norm": 7.260865211486816,
    "timestamp": "2025-07-14T12:35:15.496190"
  },
  {
    "epoch": 2826,
    "batch_idx": 128,
    "loss": 0.28008314967155457,
    "learning_rate": 0.00036503585276975264,
    "timestep_mean": 461.25,
    "epsilon_norm": 7.333484649658203,
    "prediction_norm": 6.639583110809326,
    "timestamp": "2025-07-14T12:35:15.501278"
  },
  {
    "epoch": 2826,
    "batch_idx": 160,
    "loss": 0.37188270688056946,
    "learning_rate": 0.00036503585276975264,
    "timestep_mean": 508.8125,
    "epsilon_norm": 8.301145553588867,
    "prediction_norm": 6.536392688751221,
    "timestamp": "2025-07-14T12:35:15.506334"
  },
  {
    "epoch": 2826,
    "batch_idx": 192,
    "loss": 0.4056047797203064,
    "learning_rate": 0.00036503585276975264,
    "timestep_mean": 496.0,
    "epsilon_norm": 4.300710201263428,
    "prediction_norm": 3.6136209964752197,
    "timestamp": "2025-07-14T12:35:15.511440"
  },
  {
    "epoch": 2827,
    "batch_idx": 0,
    "loss": 0.14266344904899597,
    "learning_rate": 0.00036579226348265847,
    "timestep_mean": 572.875,
    "epsilon_norm": 6.885909557342529,
    "prediction_norm": 5.944647789001465,
    "timestamp": "2025-07-14T12:35:15.516604"
  },
  {
    "epoch": 2827,
    "batch_idx": 32,
    "loss": 0.4947396516799927,
    "learning_rate": 0.00036579226348265847,
    "timestep_mean": 484.0625,
    "epsilon_norm": 8.986052513122559,
    "prediction_norm": 6.723264694213867,
    "timestamp": "2025-07-14T12:35:15.521914"
  },
  {
    "epoch": 2827,
    "batch_idx": 64,
    "loss": 0.4394788146018982,
    "learning_rate": 0.00036579226348265847,
    "timestep_mean": 430.09375,
    "epsilon_norm": 9.063060760498047,
    "prediction_norm": 7.183493614196777,
    "timestamp": "2025-07-14T12:35:15.527209"
  },
  {
    "epoch": 2827,
    "batch_idx": 96,
    "loss": 0.15122845768928528,
    "learning_rate": 0.00036579226348265847,
    "timestep_mean": 527.1875,
    "epsilon_norm": 7.683912754058838,
    "prediction_norm": 6.7319817543029785,
    "timestamp": "2025-07-14T12:35:15.532370"
  },
  {
    "epoch": 2827,
    "batch_idx": 128,
    "loss": 0.4561563730239868,
    "learning_rate": 0.00036579226348265847,
    "timestep_mean": 478.9375,
    "epsilon_norm": 8.83291244506836,
    "prediction_norm": 7.790671348571777,
    "timestamp": "2025-07-14T12:35:15.537420"
  },
  {
    "epoch": 2827,
    "batch_idx": 160,
    "loss": 0.2532883584499359,
    "learning_rate": 0.00036579226348265847,
    "timestep_mean": 461.75,
    "epsilon_norm": 7.6545329093933105,
    "prediction_norm": 6.838597297668457,
    "timestamp": "2025-07-14T12:35:15.542798"
  },
  {
    "epoch": 2827,
    "batch_idx": 192,
    "loss": 0.10932131856679916,
    "learning_rate": 0.00036579226348265847,
    "timestep_mean": 611.375,
    "epsilon_norm": 3.567699432373047,
    "prediction_norm": 3.6406009197235107,
    "timestamp": "2025-07-14T12:35:15.547785"
  },
  {
    "epoch": 2828,
    "batch_idx": 0,
    "loss": 0.4318356215953827,
    "learning_rate": 0.000366549005339813,
    "timestep_mean": 510.6875,
    "epsilon_norm": 7.8270792961120605,
    "prediction_norm": 6.202173233032227,
    "timestamp": "2025-07-14T12:35:15.552961"
  },
  {
    "epoch": 2828,
    "batch_idx": 32,
    "loss": 0.32149553298950195,
    "learning_rate": 0.000366549005339813,
    "timestep_mean": 506.28125,
    "epsilon_norm": 8.009523391723633,
    "prediction_norm": 7.014162063598633,
    "timestamp": "2025-07-14T12:35:15.558146"
  },
  {
    "epoch": 2828,
    "batch_idx": 64,
    "loss": 0.48118531703948975,
    "learning_rate": 0.000366549005339813,
    "timestep_mean": 439.0625,
    "epsilon_norm": 7.657478332519531,
    "prediction_norm": 5.373754024505615,
    "timestamp": "2025-07-14T12:35:15.563474"
  },
  {
    "epoch": 2828,
    "batch_idx": 96,
    "loss": 0.12283673882484436,
    "learning_rate": 0.000366549005339813,
    "timestep_mean": 610.375,
    "epsilon_norm": 8.615057945251465,
    "prediction_norm": 8.698806762695312,
    "timestamp": "2025-07-14T12:35:15.568603"
  },
  {
    "epoch": 2828,
    "batch_idx": 128,
    "loss": 0.4372815191745758,
    "learning_rate": 0.000366549005339813,
    "timestep_mean": 385.875,
    "epsilon_norm": 8.550633430480957,
    "prediction_norm": 6.6970930099487305,
    "timestamp": "2025-07-14T12:35:15.573861"
  },
  {
    "epoch": 2828,
    "batch_idx": 160,
    "loss": 0.2276567667722702,
    "learning_rate": 0.000366549005339813,
    "timestep_mean": 497.75,
    "epsilon_norm": 8.233470916748047,
    "prediction_norm": 7.106259822845459,
    "timestamp": "2025-07-14T12:35:15.579026"
  },
  {
    "epoch": 2828,
    "batch_idx": 192,
    "loss": 0.20275871455669403,
    "learning_rate": 0.000366549005339813,
    "timestep_mean": 572.875,
    "epsilon_norm": 4.442187786102295,
    "prediction_norm": 4.13268518447876,
    "timestamp": "2025-07-14T12:35:15.584199"
  },
  {
    "epoch": 2829,
    "batch_idx": 0,
    "loss": 0.530073881149292,
    "learning_rate": 0.0003673060764740308,
    "timestep_mean": 437.0625,
    "epsilon_norm": 8.09022331237793,
    "prediction_norm": 6.138092994689941,
    "timestamp": "2025-07-14T12:35:15.589407"
  },
  {
    "epoch": 2829,
    "batch_idx": 32,
    "loss": 0.2880713641643524,
    "learning_rate": 0.0003673060764740308,
    "timestep_mean": 588.78125,
    "epsilon_norm": 8.420394897460938,
    "prediction_norm": 7.383309364318848,
    "timestamp": "2025-07-14T12:35:15.594568"
  },
  {
    "epoch": 2829,
    "batch_idx": 64,
    "loss": 0.15897974371910095,
    "learning_rate": 0.0003673060764740308,
    "timestep_mean": 528.125,
    "epsilon_norm": 7.711885929107666,
    "prediction_norm": 7.0501275062561035,
    "timestamp": "2025-07-14T12:35:15.599626"
  },
  {
    "epoch": 2829,
    "batch_idx": 96,
    "loss": 0.47184300422668457,
    "learning_rate": 0.0003673060764740308,
    "timestep_mean": 419.1875,
    "epsilon_norm": 8.75765323638916,
    "prediction_norm": 6.581873416900635,
    "timestamp": "2025-07-14T12:35:15.604858"
  },
  {
    "epoch": 2829,
    "batch_idx": 128,
    "loss": 0.32655957341194153,
    "learning_rate": 0.0003673060764740308,
    "timestep_mean": 533.1875,
    "epsilon_norm": 9.419785499572754,
    "prediction_norm": 7.439058780670166,
    "timestamp": "2025-07-14T12:35:15.609965"
  },
  {
    "epoch": 2829,
    "batch_idx": 160,
    "loss": 0.2769778370857239,
    "learning_rate": 0.0003673060764740308,
    "timestep_mean": 497.34375,
    "epsilon_norm": 7.679352283477783,
    "prediction_norm": 6.697351932525635,
    "timestamp": "2025-07-14T12:35:15.615123"
  },
  {
    "epoch": 2829,
    "batch_idx": 192,
    "loss": 0.29680293798446655,
    "learning_rate": 0.0003673060764740308,
    "timestep_mean": 459.625,
    "epsilon_norm": 4.068239212036133,
    "prediction_norm": 3.3408477306365967,
    "timestamp": "2025-07-14T12:35:15.620313"
  },
  {
    "epoch": 2830,
    "batch_idx": 0,
    "loss": 0.46818122267723083,
    "learning_rate": 0.0003680634750173143,
    "timestep_mean": 481.59375,
    "epsilon_norm": 7.91200590133667,
    "prediction_norm": 6.1533427238464355,
    "timestamp": "2025-07-14T12:35:15.625671"
  },
  {
    "epoch": 2830,
    "batch_idx": 32,
    "loss": 0.26479166746139526,
    "learning_rate": 0.0003680634750173143,
    "timestep_mean": 532.96875,
    "epsilon_norm": 8.551898956298828,
    "prediction_norm": 7.596141815185547,
    "timestamp": "2025-07-14T12:35:15.630813"
  },
  {
    "epoch": 2830,
    "batch_idx": 64,
    "loss": 0.17004533112049103,
    "learning_rate": 0.0003680634750173143,
    "timestep_mean": 512.6875,
    "epsilon_norm": 7.750368118286133,
    "prediction_norm": 6.708733558654785,
    "timestamp": "2025-07-14T12:35:15.636065"
  },
  {
    "epoch": 2830,
    "batch_idx": 96,
    "loss": 0.36509281396865845,
    "learning_rate": 0.0003680634750173143,
    "timestep_mean": 530.5625,
    "epsilon_norm": 8.642413139343262,
    "prediction_norm": 6.808588027954102,
    "timestamp": "2025-07-14T12:35:15.641233"
  },
  {
    "epoch": 2830,
    "batch_idx": 128,
    "loss": 0.30282148718833923,
    "learning_rate": 0.0003680634750173143,
    "timestep_mean": 417.25,
    "epsilon_norm": 8.334139823913574,
    "prediction_norm": 6.642773151397705,
    "timestamp": "2025-07-14T12:35:15.646355"
  },
  {
    "epoch": 2830,
    "batch_idx": 160,
    "loss": 0.31030595302581787,
    "learning_rate": 0.0003680634750173143,
    "timestep_mean": 500.9375,
    "epsilon_norm": 8.283369064331055,
    "prediction_norm": 6.869301795959473,
    "timestamp": "2025-07-14T12:35:15.651598"
  },
  {
    "epoch": 2830,
    "batch_idx": 192,
    "loss": 0.7664011716842651,
    "learning_rate": 0.0003680634750173143,
    "timestep_mean": 314.0,
    "epsilon_norm": 5.049469947814941,
    "prediction_norm": 3.0145809650421143,
    "timestamp": "2025-07-14T12:35:15.656744"
  },
  {
    "epoch": 2831,
    "batch_idx": 0,
    "loss": 0.34439027309417725,
    "learning_rate": 0.00036882119910085806,
    "timestep_mean": 538.625,
    "epsilon_norm": 7.303602695465088,
    "prediction_norm": 5.824850559234619,
    "timestamp": "2025-07-14T12:35:15.661889"
  },
  {
    "epoch": 2831,
    "batch_idx": 32,
    "loss": 0.1685405820608139,
    "learning_rate": 0.00036882119910085806,
    "timestep_mean": 596.09375,
    "epsilon_norm": 8.54223918914795,
    "prediction_norm": 7.8636016845703125,
    "timestamp": "2025-07-14T12:35:15.667242"
  },
  {
    "epoch": 2831,
    "batch_idx": 64,
    "loss": 0.1386447548866272,
    "learning_rate": 0.00036882119910085806,
    "timestep_mean": 523.9375,
    "epsilon_norm": 7.308549880981445,
    "prediction_norm": 6.468632698059082,
    "timestamp": "2025-07-14T12:35:15.672437"
  },
  {
    "epoch": 2831,
    "batch_idx": 96,
    "loss": 0.1396632194519043,
    "learning_rate": 0.00036882119910085806,
    "timestep_mean": 516.125,
    "epsilon_norm": 7.262071132659912,
    "prediction_norm": 6.579044818878174,
    "timestamp": "2025-07-14T12:35:15.677540"
  },
  {
    "epoch": 2831,
    "batch_idx": 128,
    "loss": 0.34529000520706177,
    "learning_rate": 0.00036882119910085806,
    "timestep_mean": 499.9375,
    "epsilon_norm": 8.372709274291992,
    "prediction_norm": 7.606731414794922,
    "timestamp": "2025-07-14T12:35:15.682730"
  },
  {
    "epoch": 2831,
    "batch_idx": 160,
    "loss": 0.22856369614601135,
    "learning_rate": 0.00036882119910085806,
    "timestep_mean": 486.1875,
    "epsilon_norm": 8.73121166229248,
    "prediction_norm": 7.674673557281494,
    "timestamp": "2025-07-14T12:35:15.687903"
  },
  {
    "epoch": 2831,
    "batch_idx": 192,
    "loss": 0.38528159260749817,
    "learning_rate": 0.00036882119910085806,
    "timestep_mean": 391.75,
    "epsilon_norm": 3.7584805488586426,
    "prediction_norm": 3.450791358947754,
    "timestamp": "2025-07-14T12:35:15.692949"
  },
  {
    "epoch": 2832,
    "batch_idx": 0,
    "loss": 0.3399890065193176,
    "learning_rate": 0.00036957924685505255,
    "timestep_mean": 488.3125,
    "epsilon_norm": 8.410968780517578,
    "prediction_norm": 6.465134620666504,
    "timestamp": "2025-07-14T12:35:15.698165"
  },
  {
    "epoch": 2832,
    "batch_idx": 32,
    "loss": 0.3035859763622284,
    "learning_rate": 0.00036957924685505255,
    "timestep_mean": 548.53125,
    "epsilon_norm": 7.679002285003662,
    "prediction_norm": 6.451423168182373,
    "timestamp": "2025-07-14T12:35:15.703332"
  },
  {
    "epoch": 2832,
    "batch_idx": 64,
    "loss": 0.3361072540283203,
    "learning_rate": 0.00036957924685505255,
    "timestep_mean": 417.375,
    "epsilon_norm": 7.680271148681641,
    "prediction_norm": 6.018054008483887,
    "timestamp": "2025-07-14T12:35:15.708173"
  },
  {
    "epoch": 2832,
    "batch_idx": 96,
    "loss": 0.33462902903556824,
    "learning_rate": 0.00036957924685505255,
    "timestep_mean": 477.21875,
    "epsilon_norm": 7.157009124755859,
    "prediction_norm": 6.217471599578857,
    "timestamp": "2025-07-14T12:35:15.713176"
  },
  {
    "epoch": 2832,
    "batch_idx": 128,
    "loss": 0.1293288767337799,
    "learning_rate": 0.00036957924685505255,
    "timestep_mean": 594.25,
    "epsilon_norm": 9.277149200439453,
    "prediction_norm": 8.583126068115234,
    "timestamp": "2025-07-14T12:35:15.718049"
  },
  {
    "epoch": 2832,
    "batch_idx": 160,
    "loss": 0.15749382972717285,
    "learning_rate": 0.00036957924685505255,
    "timestep_mean": 517.1875,
    "epsilon_norm": 6.829709053039551,
    "prediction_norm": 6.157012939453125,
    "timestamp": "2025-07-14T12:35:15.722846"
  },
  {
    "epoch": 2832,
    "batch_idx": 192,
    "loss": 0.2224128693342209,
    "learning_rate": 0.00036957924685505255,
    "timestep_mean": 385.125,
    "epsilon_norm": 2.608511209487915,
    "prediction_norm": 2.606189250946045,
    "timestamp": "2025-07-14T12:35:15.727732"
  },
  {
    "epoch": 2833,
    "batch_idx": 0,
    "loss": 0.2155749499797821,
    "learning_rate": 0.0003703376164094903,
    "timestep_mean": 591.4375,
    "epsilon_norm": 7.827582836151123,
    "prediction_norm": 7.462778091430664,
    "timestamp": "2025-07-14T12:35:15.732614"
  },
  {
    "epoch": 2833,
    "batch_idx": 32,
    "loss": 0.19621901214122772,
    "learning_rate": 0.0003703376164094903,
    "timestep_mean": 553.5,
    "epsilon_norm": 7.422875881195068,
    "prediction_norm": 6.411311626434326,
    "timestamp": "2025-07-14T12:35:15.737464"
  },
  {
    "epoch": 2833,
    "batch_idx": 64,
    "loss": 0.21780163049697876,
    "learning_rate": 0.0003703376164094903,
    "timestep_mean": 470.8125,
    "epsilon_norm": 7.861071586608887,
    "prediction_norm": 6.35894775390625,
    "timestamp": "2025-07-14T12:35:15.742390"
  },
  {
    "epoch": 2833,
    "batch_idx": 96,
    "loss": 0.3227412700653076,
    "learning_rate": 0.0003703376164094903,
    "timestep_mean": 527.40625,
    "epsilon_norm": 8.065850257873535,
    "prediction_norm": 6.252252578735352,
    "timestamp": "2025-07-14T12:35:15.747418"
  },
  {
    "epoch": 2833,
    "batch_idx": 128,
    "loss": 0.0880739688873291,
    "learning_rate": 0.0003703376164094903,
    "timestep_mean": 555.125,
    "epsilon_norm": 8.413281440734863,
    "prediction_norm": 7.623112678527832,
    "timestamp": "2025-07-14T12:35:15.752300"
  },
  {
    "epoch": 2833,
    "batch_idx": 160,
    "loss": 0.5464649200439453,
    "learning_rate": 0.0003703376164094903,
    "timestep_mean": 528.59375,
    "epsilon_norm": 7.66322660446167,
    "prediction_norm": 5.593082427978516,
    "timestamp": "2025-07-14T12:35:15.757219"
  },
  {
    "epoch": 2833,
    "batch_idx": 192,
    "loss": 0.03240657597780228,
    "learning_rate": 0.0003703376164094903,
    "timestep_mean": 730.125,
    "epsilon_norm": 2.6697323322296143,
    "prediction_norm": 2.8406178951263428,
    "timestamp": "2025-07-14T12:35:15.762099"
  },
  {
    "epoch": 2834,
    "batch_idx": 0,
    "loss": 0.23558035492897034,
    "learning_rate": 0.000371096305892971,
    "timestep_mean": 459.5,
    "epsilon_norm": 7.8265838623046875,
    "prediction_norm": 6.315855979919434,
    "timestamp": "2025-07-14T12:35:15.767062"
  },
  {
    "epoch": 2834,
    "batch_idx": 32,
    "loss": 0.2073378562927246,
    "learning_rate": 0.000371096305892971,
    "timestep_mean": 492.4375,
    "epsilon_norm": 7.731133460998535,
    "prediction_norm": 6.522706508636475,
    "timestamp": "2025-07-14T12:35:15.771961"
  },
  {
    "epoch": 2834,
    "batch_idx": 64,
    "loss": 0.25521689653396606,
    "learning_rate": 0.000371096305892971,
    "timestep_mean": 539.875,
    "epsilon_norm": 6.506800651550293,
    "prediction_norm": 5.006854057312012,
    "timestamp": "2025-07-14T12:35:15.776786"
  },
  {
    "epoch": 2834,
    "batch_idx": 96,
    "loss": 0.1274530440568924,
    "learning_rate": 0.000371096305892971,
    "timestep_mean": 504.96875,
    "epsilon_norm": 7.955506324768066,
    "prediction_norm": 7.193660259246826,
    "timestamp": "2025-07-14T12:35:15.781661"
  },
  {
    "epoch": 2834,
    "batch_idx": 128,
    "loss": 0.21784205734729767,
    "learning_rate": 0.000371096305892971,
    "timestep_mean": 457.53125,
    "epsilon_norm": 7.683332920074463,
    "prediction_norm": 6.719261169433594,
    "timestamp": "2025-07-14T12:35:15.786552"
  },
  {
    "epoch": 2834,
    "batch_idx": 160,
    "loss": 0.3099585771560669,
    "learning_rate": 0.000371096305892971,
    "timestep_mean": 475.46875,
    "epsilon_norm": 8.29151439666748,
    "prediction_norm": 6.683414459228516,
    "timestamp": "2025-07-14T12:35:15.791417"
  },
  {
    "epoch": 2834,
    "batch_idx": 192,
    "loss": 0.2602689862251282,
    "learning_rate": 0.000371096305892971,
    "timestep_mean": 527.625,
    "epsilon_norm": 2.5360000133514404,
    "prediction_norm": 2.3367834091186523,
    "timestamp": "2025-07-14T12:35:15.796211"
  },
  {
    "epoch": 2835,
    "batch_idx": 0,
    "loss": 0.2754638195037842,
    "learning_rate": 0.0003718553134335026,
    "timestep_mean": 484.78125,
    "epsilon_norm": 7.481594562530518,
    "prediction_norm": 6.162106990814209,
    "timestamp": "2025-07-14T12:35:15.801148"
  },
  {
    "epoch": 2835,
    "batch_idx": 32,
    "loss": 0.267464280128479,
    "learning_rate": 0.0003718553134335026,
    "timestep_mean": 515.25,
    "epsilon_norm": 9.207518577575684,
    "prediction_norm": 8.371562004089355,
    "timestamp": "2025-07-14T12:35:15.805988"
  },
  {
    "epoch": 2835,
    "batch_idx": 64,
    "loss": 0.27010130882263184,
    "learning_rate": 0.0003718553134335026,
    "timestep_mean": 496.1875,
    "epsilon_norm": 8.60584545135498,
    "prediction_norm": 7.598509788513184,
    "timestamp": "2025-07-14T12:35:15.810863"
  },
  {
    "epoch": 2835,
    "batch_idx": 96,
    "loss": 0.28696271777153015,
    "learning_rate": 0.0003718553134335026,
    "timestep_mean": 535.9375,
    "epsilon_norm": 8.602255821228027,
    "prediction_norm": 7.323748588562012,
    "timestamp": "2025-07-14T12:35:15.815670"
  },
  {
    "epoch": 2835,
    "batch_idx": 128,
    "loss": 0.2492106705904007,
    "learning_rate": 0.0003718553134335026,
    "timestep_mean": 500.875,
    "epsilon_norm": 8.638238906860352,
    "prediction_norm": 6.879655361175537,
    "timestamp": "2025-07-14T12:35:15.820503"
  },
  {
    "epoch": 2835,
    "batch_idx": 160,
    "loss": 0.2506476938724518,
    "learning_rate": 0.0003718553134335026,
    "timestep_mean": 526.875,
    "epsilon_norm": 8.324348449707031,
    "prediction_norm": 6.972523212432861,
    "timestamp": "2025-07-14T12:35:15.825498"
  },
  {
    "epoch": 2835,
    "batch_idx": 192,
    "loss": 0.1244765967130661,
    "learning_rate": 0.0003718553134335026,
    "timestep_mean": 457.25,
    "epsilon_norm": 3.775470733642578,
    "prediction_norm": 3.4589250087738037,
    "timestamp": "2025-07-14T12:35:15.830413"
  },
  {
    "epoch": 2836,
    "batch_idx": 0,
    "loss": 0.28585389256477356,
    "learning_rate": 0.0003726146371583098,
    "timestep_mean": 502.28125,
    "epsilon_norm": 7.545804977416992,
    "prediction_norm": 6.24019193649292,
    "timestamp": "2025-07-14T12:35:15.835468"
  },
  {
    "epoch": 2836,
    "batch_idx": 32,
    "loss": 0.16172587871551514,
    "learning_rate": 0.0003726146371583098,
    "timestep_mean": 495.4375,
    "epsilon_norm": 8.473566055297852,
    "prediction_norm": 7.247059345245361,
    "timestamp": "2025-07-14T12:35:15.840552"
  },
  {
    "epoch": 2836,
    "batch_idx": 64,
    "loss": 0.20502391457557678,
    "learning_rate": 0.0003726146371583098,
    "timestep_mean": 546.71875,
    "epsilon_norm": 8.427193641662598,
    "prediction_norm": 7.621517181396484,
    "timestamp": "2025-07-14T12:35:15.845644"
  },
  {
    "epoch": 2836,
    "batch_idx": 96,
    "loss": 0.2248738706111908,
    "learning_rate": 0.0003726146371583098,
    "timestep_mean": 520.84375,
    "epsilon_norm": 7.280000686645508,
    "prediction_norm": 6.405783176422119,
    "timestamp": "2025-07-14T12:35:15.850837"
  },
  {
    "epoch": 2836,
    "batch_idx": 128,
    "loss": 0.35285675525665283,
    "learning_rate": 0.0003726146371583098,
    "timestep_mean": 457.65625,
    "epsilon_norm": 7.926976680755615,
    "prediction_norm": 6.325715065002441,
    "timestamp": "2025-07-14T12:35:15.855936"
  },
  {
    "epoch": 2836,
    "batch_idx": 160,
    "loss": 0.3589221239089966,
    "learning_rate": 0.0003726146371583098,
    "timestep_mean": 469.875,
    "epsilon_norm": 8.373812675476074,
    "prediction_norm": 7.253232955932617,
    "timestamp": "2025-07-14T12:35:15.861156"
  },
  {
    "epoch": 2836,
    "batch_idx": 192,
    "loss": 0.310052752494812,
    "learning_rate": 0.0003726146371583098,
    "timestep_mean": 514.375,
    "epsilon_norm": 4.988543510437012,
    "prediction_norm": 4.193291187286377,
    "timestamp": "2025-07-14T12:35:15.866236"
  },
  {
    "epoch": 2837,
    "batch_idx": 0,
    "loss": 0.2859623432159424,
    "learning_rate": 0.0003733742751938368,
    "timestep_mean": 495.625,
    "epsilon_norm": 7.692845821380615,
    "prediction_norm": 6.670872211456299,
    "timestamp": "2025-07-14T12:35:15.871370"
  },
  {
    "epoch": 2837,
    "batch_idx": 32,
    "loss": 0.24042852222919464,
    "learning_rate": 0.0003733742751938368,
    "timestep_mean": 496.8125,
    "epsilon_norm": 7.318385124206543,
    "prediction_norm": 5.932100296020508,
    "timestamp": "2025-07-14T12:35:15.876506"
  },
  {
    "epoch": 2837,
    "batch_idx": 64,
    "loss": 0.1603107899427414,
    "learning_rate": 0.0003733742751938368,
    "timestep_mean": 582.4375,
    "epsilon_norm": 7.704962730407715,
    "prediction_norm": 7.255630970001221,
    "timestamp": "2025-07-14T12:35:15.881716"
  },
  {
    "epoch": 2837,
    "batch_idx": 96,
    "loss": 0.2167091965675354,
    "learning_rate": 0.0003733742751938368,
    "timestep_mean": 480.8125,
    "epsilon_norm": 8.091485977172852,
    "prediction_norm": 7.286481857299805,
    "timestamp": "2025-07-14T12:35:15.886860"
  },
  {
    "epoch": 2837,
    "batch_idx": 128,
    "loss": 0.3248811662197113,
    "learning_rate": 0.0003733742751938368,
    "timestep_mean": 545.03125,
    "epsilon_norm": 7.99492883682251,
    "prediction_norm": 6.066136837005615,
    "timestamp": "2025-07-14T12:35:15.891856"
  },
  {
    "epoch": 2837,
    "batch_idx": 160,
    "loss": 0.3019312024116516,
    "learning_rate": 0.0003733742751938368,
    "timestep_mean": 493.21875,
    "epsilon_norm": 7.699489593505859,
    "prediction_norm": 6.301064968109131,
    "timestamp": "2025-07-14T12:35:15.896801"
  },
  {
    "epoch": 2837,
    "batch_idx": 192,
    "loss": 0.2167610228061676,
    "learning_rate": 0.0003733742751938368,
    "timestep_mean": 593.25,
    "epsilon_norm": 4.985299587249756,
    "prediction_norm": 4.247271537780762,
    "timestamp": "2025-07-14T12:35:15.901929"
  },
  {
    "epoch": 2838,
    "batch_idx": 0,
    "loss": 0.14232298731803894,
    "learning_rate": 0.00037413422566575267,
    "timestep_mean": 510.5625,
    "epsilon_norm": 7.5847907066345215,
    "prediction_norm": 7.74271297454834,
    "timestamp": "2025-07-14T12:35:15.907055"
  },
  {
    "epoch": 2838,
    "batch_idx": 32,
    "loss": 0.2197956144809723,
    "learning_rate": 0.00037413422566575267,
    "timestep_mean": 544.5,
    "epsilon_norm": 8.841917991638184,
    "prediction_norm": 8.239053726196289,
    "timestamp": "2025-07-14T12:35:15.912260"
  },
  {
    "epoch": 2838,
    "batch_idx": 64,
    "loss": 0.5106161832809448,
    "learning_rate": 0.00037413422566575267,
    "timestep_mean": 481.0625,
    "epsilon_norm": 8.712519645690918,
    "prediction_norm": 7.113710403442383,
    "timestamp": "2025-07-14T12:35:15.917487"
  },
  {
    "epoch": 2838,
    "batch_idx": 96,
    "loss": 0.43489712476730347,
    "learning_rate": 0.00037413422566575267,
    "timestep_mean": 515.125,
    "epsilon_norm": 8.031578063964844,
    "prediction_norm": 6.585275173187256,
    "timestamp": "2025-07-14T12:35:15.922659"
  },
  {
    "epoch": 2838,
    "batch_idx": 128,
    "loss": 0.3373972773551941,
    "learning_rate": 0.00037413422566575267,
    "timestep_mean": 438.71875,
    "epsilon_norm": 6.841410160064697,
    "prediction_norm": 5.617948055267334,
    "timestamp": "2025-07-14T12:35:15.927922"
  },
  {
    "epoch": 2838,
    "batch_idx": 160,
    "loss": 0.2165006697177887,
    "learning_rate": 0.00037413422566575267,
    "timestep_mean": 571.75,
    "epsilon_norm": 9.03567886352539,
    "prediction_norm": 8.321805000305176,
    "timestamp": "2025-07-14T12:35:15.933281"
  },
  {
    "epoch": 2838,
    "batch_idx": 192,
    "loss": 0.19435834884643555,
    "learning_rate": 0.00037413422566575267,
    "timestep_mean": 464.875,
    "epsilon_norm": 3.785753011703491,
    "prediction_norm": 3.393117904663086,
    "timestamp": "2025-07-14T12:35:15.938551"
  },
  {
    "epoch": 2839,
    "batch_idx": 0,
    "loss": 0.37926995754241943,
    "learning_rate": 0.0003748944866989539,
    "timestep_mean": 510.4375,
    "epsilon_norm": 6.852599620819092,
    "prediction_norm": 5.775565147399902,
    "timestamp": "2025-07-14T12:35:15.943919"
  },
  {
    "epoch": 2839,
    "batch_idx": 32,
    "loss": 0.3802815079689026,
    "learning_rate": 0.0003748944866989539,
    "timestep_mean": 423.0625,
    "epsilon_norm": 7.7044172286987305,
    "prediction_norm": 6.547565460205078,
    "timestamp": "2025-07-14T12:35:15.949152"
  },
  {
    "epoch": 2839,
    "batch_idx": 64,
    "loss": 0.17075061798095703,
    "learning_rate": 0.0003748944866989539,
    "timestep_mean": 517.0,
    "epsilon_norm": 7.574464797973633,
    "prediction_norm": 7.1047492027282715,
    "timestamp": "2025-07-14T12:35:15.954339"
  },
  {
    "epoch": 2839,
    "batch_idx": 96,
    "loss": 0.17642351984977722,
    "learning_rate": 0.0003748944866989539,
    "timestep_mean": 564.34375,
    "epsilon_norm": 7.038501262664795,
    "prediction_norm": 6.136530876159668,
    "timestamp": "2025-07-14T12:35:15.959590"
  },
  {
    "epoch": 2839,
    "batch_idx": 128,
    "loss": 0.29428285360336304,
    "learning_rate": 0.0003748944866989539,
    "timestep_mean": 464.75,
    "epsilon_norm": 7.454473972320557,
    "prediction_norm": 6.027009010314941,
    "timestamp": "2025-07-14T12:35:15.964668"
  },
  {
    "epoch": 2839,
    "batch_idx": 160,
    "loss": 0.19679643213748932,
    "learning_rate": 0.0003748944866989539,
    "timestep_mean": 502.1875,
    "epsilon_norm": 7.194705009460449,
    "prediction_norm": 6.316269874572754,
    "timestamp": "2025-07-14T12:35:15.969826"
  },
  {
    "epoch": 2839,
    "batch_idx": 192,
    "loss": 0.4387194514274597,
    "learning_rate": 0.0003748944866989539,
    "timestep_mean": 403.75,
    "epsilon_norm": 4.198215007781982,
    "prediction_norm": 3.4358739852905273,
    "timestamp": "2025-07-14T12:35:15.974982"
  },
  {
    "epoch": 2840,
    "batch_idx": 0,
    "loss": 0.45620226860046387,
    "learning_rate": 0.00037565505641757327,
    "timestep_mean": 562.6875,
    "epsilon_norm": 8.068690299987793,
    "prediction_norm": 5.8594183921813965,
    "timestamp": "2025-07-14T12:35:15.980234"
  },
  {
    "epoch": 2840,
    "batch_idx": 32,
    "loss": 0.21463294327259064,
    "learning_rate": 0.00037565505641757327,
    "timestep_mean": 437.96875,
    "epsilon_norm": 6.391630172729492,
    "prediction_norm": 6.002038955688477,
    "timestamp": "2025-07-14T12:35:15.985385"
  },
  {
    "epoch": 2840,
    "batch_idx": 64,
    "loss": 0.09338783472776413,
    "learning_rate": 0.00037565505641757327,
    "timestep_mean": 586.34375,
    "epsilon_norm": 7.963030815124512,
    "prediction_norm": 7.261363983154297,
    "timestamp": "2025-07-14T12:35:15.990517"
  },
  {
    "epoch": 2840,
    "batch_idx": 96,
    "loss": 0.23377718031406403,
    "learning_rate": 0.00037565505641757327,
    "timestep_mean": 475.375,
    "epsilon_norm": 7.832399845123291,
    "prediction_norm": 6.689317226409912,
    "timestamp": "2025-07-14T12:35:15.995606"
  },
  {
    "epoch": 2840,
    "batch_idx": 128,
    "loss": 0.42147332429885864,
    "learning_rate": 0.00037565505641757327,
    "timestep_mean": 464.78125,
    "epsilon_norm": 9.012925148010254,
    "prediction_norm": 7.1493916511535645,
    "timestamp": "2025-07-14T12:35:16.000658"
  },
  {
    "epoch": 2840,
    "batch_idx": 160,
    "loss": 0.6030181646347046,
    "learning_rate": 0.00037565505641757327,
    "timestep_mean": 371.09375,
    "epsilon_norm": 8.297748565673828,
    "prediction_norm": 6.008217811584473,
    "timestamp": "2025-07-14T12:35:16.005744"
  },
  {
    "epoch": 2840,
    "batch_idx": 192,
    "loss": 0.24181410670280457,
    "learning_rate": 0.00037565505641757327,
    "timestep_mean": 501.75,
    "epsilon_norm": 5.167621612548828,
    "prediction_norm": 4.045815944671631,
    "timestamp": "2025-07-14T12:35:16.010727"
  },
  {
    "epoch": 2841,
    "batch_idx": 0,
    "loss": 0.15836000442504883,
    "learning_rate": 0.0003764159329449806,
    "timestep_mean": 542.40625,
    "epsilon_norm": 8.6662015914917,
    "prediction_norm": 7.7289719581604,
    "timestamp": "2025-07-14T12:35:16.015886"
  },
  {
    "epoch": 2841,
    "batch_idx": 32,
    "loss": 0.48889413475990295,
    "learning_rate": 0.0003764159329449806,
    "timestep_mean": 567.1875,
    "epsilon_norm": 9.254778861999512,
    "prediction_norm": 7.617154121398926,
    "timestamp": "2025-07-14T12:35:16.021121"
  },
  {
    "epoch": 2841,
    "batch_idx": 64,
    "loss": 0.33000293374061584,
    "learning_rate": 0.0003764159329449806,
    "timestep_mean": 481.15625,
    "epsilon_norm": 8.449355125427246,
    "prediction_norm": 6.922522068023682,
    "timestamp": "2025-07-14T12:35:16.026289"
  },
  {
    "epoch": 2841,
    "batch_idx": 96,
    "loss": 0.29257702827453613,
    "learning_rate": 0.0003764159329449806,
    "timestep_mean": 469.21875,
    "epsilon_norm": 8.540426254272461,
    "prediction_norm": 7.440252780914307,
    "timestamp": "2025-07-14T12:35:16.031411"
  },
  {
    "epoch": 2841,
    "batch_idx": 128,
    "loss": 0.22519400715827942,
    "learning_rate": 0.0003764159329449806,
    "timestep_mean": 537.125,
    "epsilon_norm": 7.743433475494385,
    "prediction_norm": 7.410181522369385,
    "timestamp": "2025-07-14T12:35:16.036644"
  },
  {
    "epoch": 2841,
    "batch_idx": 160,
    "loss": 0.1905023157596588,
    "learning_rate": 0.0003764159329449806,
    "timestep_mean": 508.125,
    "epsilon_norm": 7.91876745223999,
    "prediction_norm": 7.190797328948975,
    "timestamp": "2025-07-14T12:35:16.041854"
  },
  {
    "epoch": 2841,
    "batch_idx": 192,
    "loss": 0.13520124554634094,
    "learning_rate": 0.0003764159329449806,
    "timestep_mean": 495.75,
    "epsilon_norm": 3.67924427986145,
    "prediction_norm": 3.2476155757904053,
    "timestamp": "2025-07-14T12:35:16.047080"
  },
  {
    "epoch": 2842,
    "batch_idx": 0,
    "loss": 0.22150063514709473,
    "learning_rate": 0.00037717711440378785,
    "timestep_mean": 359.78125,
    "epsilon_norm": 7.048937797546387,
    "prediction_norm": 5.651988506317139,
    "timestamp": "2025-07-14T12:35:16.052232"
  },
  {
    "epoch": 2842,
    "batch_idx": 32,
    "loss": 0.12435580044984818,
    "learning_rate": 0.00037717711440378785,
    "timestep_mean": 510.09375,
    "epsilon_norm": 7.231339931488037,
    "prediction_norm": 6.346043586730957,
    "timestamp": "2025-07-14T12:35:16.057375"
  },
  {
    "epoch": 2842,
    "batch_idx": 64,
    "loss": 0.1299346685409546,
    "learning_rate": 0.00037717711440378785,
    "timestep_mean": 537.34375,
    "epsilon_norm": 7.234219074249268,
    "prediction_norm": 7.365020751953125,
    "timestamp": "2025-07-14T12:35:16.062469"
  },
  {
    "epoch": 2842,
    "batch_idx": 96,
    "loss": 0.19867613911628723,
    "learning_rate": 0.00037717711440378785,
    "timestep_mean": 524.90625,
    "epsilon_norm": 7.417142868041992,
    "prediction_norm": 6.163015842437744,
    "timestamp": "2025-07-14T12:35:16.067730"
  },
  {
    "epoch": 2842,
    "batch_idx": 128,
    "loss": 0.39308303594589233,
    "learning_rate": 0.00037717711440378785,
    "timestep_mean": 394.1875,
    "epsilon_norm": 8.502232551574707,
    "prediction_norm": 6.72250509262085,
    "timestamp": "2025-07-14T12:35:16.072829"
  },
  {
    "epoch": 2842,
    "batch_idx": 160,
    "loss": 0.23900365829467773,
    "learning_rate": 0.00037717711440378785,
    "timestep_mean": 490.09375,
    "epsilon_norm": 7.938724517822266,
    "prediction_norm": 6.609907627105713,
    "timestamp": "2025-07-14T12:35:16.077888"
  },
  {
    "epoch": 2842,
    "batch_idx": 192,
    "loss": 0.027414482086896896,
    "learning_rate": 0.00037717711440378785,
    "timestep_mean": 655.875,
    "epsilon_norm": 2.5167460441589355,
    "prediction_norm": 1.9898285865783691,
    "timestamp": "2025-07-14T12:35:16.083106"
  },
  {
    "epoch": 2843,
    "batch_idx": 0,
    "loss": 0.38730669021606445,
    "learning_rate": 0.0003779385989158558,
    "timestep_mean": 428.09375,
    "epsilon_norm": 8.163278579711914,
    "prediction_norm": 6.2963104248046875,
    "timestamp": "2025-07-14T12:35:16.088407"
  },
  {
    "epoch": 2843,
    "batch_idx": 32,
    "loss": 0.4268268942832947,
    "learning_rate": 0.0003779385989158558,
    "timestep_mean": 454.4375,
    "epsilon_norm": 7.764284610748291,
    "prediction_norm": 6.326881408691406,
    "timestamp": "2025-07-14T12:35:16.093495"
  },
  {
    "epoch": 2843,
    "batch_idx": 64,
    "loss": 0.5428056716918945,
    "learning_rate": 0.0003779385989158558,
    "timestep_mean": 533.46875,
    "epsilon_norm": 8.76651668548584,
    "prediction_norm": 7.393239974975586,
    "timestamp": "2025-07-14T12:35:16.098735"
  },
  {
    "epoch": 2843,
    "batch_idx": 96,
    "loss": 0.3232958912849426,
    "learning_rate": 0.0003779385989158558,
    "timestep_mean": 441.25,
    "epsilon_norm": 8.140995025634766,
    "prediction_norm": 7.1580963134765625,
    "timestamp": "2025-07-14T12:35:16.103849"
  },
  {
    "epoch": 2843,
    "batch_idx": 128,
    "loss": 0.21572920680046082,
    "learning_rate": 0.0003779385989158558,
    "timestep_mean": 579.90625,
    "epsilon_norm": 8.044029235839844,
    "prediction_norm": 6.827436447143555,
    "timestamp": "2025-07-14T12:35:16.108984"
  },
  {
    "epoch": 2843,
    "batch_idx": 160,
    "loss": 0.29488009214401245,
    "learning_rate": 0.0003779385989158558,
    "timestep_mean": 407.46875,
    "epsilon_norm": 8.819718360900879,
    "prediction_norm": 7.170325756072998,
    "timestamp": "2025-07-14T12:35:16.114057"
  },
  {
    "epoch": 2843,
    "batch_idx": 192,
    "loss": 0.5538849830627441,
    "learning_rate": 0.0003779385989158558,
    "timestep_mean": 546.25,
    "epsilon_norm": 3.889902353286743,
    "prediction_norm": 2.651772975921631,
    "timestamp": "2025-07-14T12:35:16.119202"
  },
  {
    "epoch": 2844,
    "batch_idx": 0,
    "loss": 0.15552057325839996,
    "learning_rate": 0.0003787003846022967,
    "timestep_mean": 531.5,
    "epsilon_norm": 7.948922634124756,
    "prediction_norm": 7.331881999969482,
    "timestamp": "2025-07-14T12:35:16.124377"
  },
  {
    "epoch": 2844,
    "batch_idx": 32,
    "loss": 0.37828993797302246,
    "learning_rate": 0.0003787003846022967,
    "timestep_mean": 411.9375,
    "epsilon_norm": 7.412331581115723,
    "prediction_norm": 6.049206733703613,
    "timestamp": "2025-07-14T12:35:16.129528"
  },
  {
    "epoch": 2844,
    "batch_idx": 64,
    "loss": 0.35755831003189087,
    "learning_rate": 0.0003787003846022967,
    "timestep_mean": 531.5625,
    "epsilon_norm": 7.8328094482421875,
    "prediction_norm": 6.477061748504639,
    "timestamp": "2025-07-14T12:35:16.134687"
  },
  {
    "epoch": 2844,
    "batch_idx": 96,
    "loss": 0.2591893672943115,
    "learning_rate": 0.0003787003846022967,
    "timestep_mean": 466.96875,
    "epsilon_norm": 7.763804912567139,
    "prediction_norm": 7.014894008636475,
    "timestamp": "2025-07-14T12:35:16.139787"
  },
  {
    "epoch": 2844,
    "batch_idx": 128,
    "loss": 0.1624373197555542,
    "learning_rate": 0.0003787003846022967,
    "timestep_mean": 486.1875,
    "epsilon_norm": 7.26138162612915,
    "prediction_norm": 6.630845069885254,
    "timestamp": "2025-07-14T12:35:16.144940"
  },
  {
    "epoch": 2844,
    "batch_idx": 160,
    "loss": 0.2071891725063324,
    "learning_rate": 0.0003787003846022967,
    "timestep_mean": 528.53125,
    "epsilon_norm": 7.906866073608398,
    "prediction_norm": 7.5299391746521,
    "timestamp": "2025-07-14T12:35:16.150076"
  },
  {
    "epoch": 2844,
    "batch_idx": 192,
    "loss": 0.3130451738834381,
    "learning_rate": 0.0003787003846022967,
    "timestep_mean": 408.375,
    "epsilon_norm": 3.2246387004852295,
    "prediction_norm": 2.914536952972412,
    "timestamp": "2025-07-14T12:35:16.155138"
  },
  {
    "epoch": 2845,
    "batch_idx": 0,
    "loss": 0.20157860219478607,
    "learning_rate": 0.00037946246958348154,
    "timestep_mean": 550.5625,
    "epsilon_norm": 7.943063735961914,
    "prediction_norm": 7.8559770584106445,
    "timestamp": "2025-07-14T12:35:16.160386"
  },
  {
    "epoch": 2845,
    "batch_idx": 32,
    "loss": 0.3811344504356384,
    "learning_rate": 0.00037946246958348154,
    "timestep_mean": 406.15625,
    "epsilon_norm": 8.202865600585938,
    "prediction_norm": 5.996580123901367,
    "timestamp": "2025-07-14T12:35:16.165644"
  },
  {
    "epoch": 2845,
    "batch_idx": 64,
    "loss": 0.375450074672699,
    "learning_rate": 0.00037946246958348154,
    "timestep_mean": 433.28125,
    "epsilon_norm": 9.28679084777832,
    "prediction_norm": 7.626001834869385,
    "timestamp": "2025-07-14T12:35:16.170781"
  },
  {
    "epoch": 2845,
    "batch_idx": 96,
    "loss": 0.2746292054653168,
    "learning_rate": 0.00037946246958348154,
    "timestep_mean": 476.625,
    "epsilon_norm": 7.138264179229736,
    "prediction_norm": 5.901297092437744,
    "timestamp": "2025-07-14T12:35:16.176126"
  },
  {
    "epoch": 2845,
    "batch_idx": 128,
    "loss": 0.21335528790950775,
    "learning_rate": 0.00037946246958348154,
    "timestep_mean": 557.3125,
    "epsilon_norm": 7.275962829589844,
    "prediction_norm": 6.36779260635376,
    "timestamp": "2025-07-14T12:35:16.181222"
  },
  {
    "epoch": 2845,
    "batch_idx": 160,
    "loss": 0.3862181305885315,
    "learning_rate": 0.00037946246958348154,
    "timestep_mean": 443.53125,
    "epsilon_norm": 7.832514762878418,
    "prediction_norm": 6.001264572143555,
    "timestamp": "2025-07-14T12:35:16.186284"
  },
  {
    "epoch": 2845,
    "batch_idx": 192,
    "loss": 0.12613950669765472,
    "learning_rate": 0.00037946246958348154,
    "timestep_mean": 629.625,
    "epsilon_norm": 3.8151628971099854,
    "prediction_norm": 3.677863836288452,
    "timestamp": "2025-07-14T12:35:16.191385"
  },
  {
    "epoch": 2846,
    "batch_idx": 0,
    "loss": 0.15348824858665466,
    "learning_rate": 0.00038022485197903986,
    "timestep_mean": 611.09375,
    "epsilon_norm": 7.23082971572876,
    "prediction_norm": 6.448085784912109,
    "timestamp": "2025-07-14T12:35:16.196784"
  },
  {
    "epoch": 2846,
    "batch_idx": 32,
    "loss": 0.44127780199050903,
    "learning_rate": 0.00038022485197903986,
    "timestep_mean": 441.96875,
    "epsilon_norm": 8.455317497253418,
    "prediction_norm": 6.620364189147949,
    "timestamp": "2025-07-14T12:35:16.201896"
  },
  {
    "epoch": 2846,
    "batch_idx": 64,
    "loss": 0.21424192190170288,
    "learning_rate": 0.00038022485197903986,
    "timestep_mean": 533.28125,
    "epsilon_norm": 7.076302528381348,
    "prediction_norm": 5.850688934326172,
    "timestamp": "2025-07-14T12:35:16.207097"
  },
  {
    "epoch": 2846,
    "batch_idx": 96,
    "loss": 0.4780202805995941,
    "learning_rate": 0.00038022485197903986,
    "timestep_mean": 345.90625,
    "epsilon_norm": 8.347384452819824,
    "prediction_norm": 5.991203308105469,
    "timestamp": "2025-07-14T12:35:16.212078"
  },
  {
    "epoch": 2846,
    "batch_idx": 128,
    "loss": 0.22572942078113556,
    "learning_rate": 0.00038022485197903986,
    "timestep_mean": 524.8125,
    "epsilon_norm": 7.322014808654785,
    "prediction_norm": 5.940544128417969,
    "timestamp": "2025-07-14T12:35:16.217083"
  },
  {
    "epoch": 2846,
    "batch_idx": 160,
    "loss": 0.2198343127965927,
    "learning_rate": 0.00038022485197903986,
    "timestep_mean": 485.96875,
    "epsilon_norm": 8.51736068725586,
    "prediction_norm": 7.34538459777832,
    "timestamp": "2025-07-14T12:35:16.222245"
  },
  {
    "epoch": 2846,
    "batch_idx": 192,
    "loss": 0.18620696663856506,
    "learning_rate": 0.00038022485197903986,
    "timestep_mean": 518.75,
    "epsilon_norm": 4.656092643737793,
    "prediction_norm": 4.302711009979248,
    "timestamp": "2025-07-14T12:35:16.227335"
  },
  {
    "epoch": 2847,
    "batch_idx": 0,
    "loss": 0.2401328682899475,
    "learning_rate": 0.00038098752990786943,
    "timestep_mean": 430.9375,
    "epsilon_norm": 7.662734508514404,
    "prediction_norm": 6.753998756408691,
    "timestamp": "2025-07-14T12:35:16.232433"
  },
  {
    "epoch": 2847,
    "batch_idx": 32,
    "loss": 0.25087055563926697,
    "learning_rate": 0.00038098752990786943,
    "timestep_mean": 502.09375,
    "epsilon_norm": 7.986795902252197,
    "prediction_norm": 7.275176525115967,
    "timestamp": "2025-07-14T12:35:16.237665"
  },
  {
    "epoch": 2847,
    "batch_idx": 64,
    "loss": 0.22503282129764557,
    "learning_rate": 0.00038098752990786943,
    "timestep_mean": 509.625,
    "epsilon_norm": 8.605576515197754,
    "prediction_norm": 7.447218894958496,
    "timestamp": "2025-07-14T12:35:16.242973"
  },
  {
    "epoch": 2847,
    "batch_idx": 96,
    "loss": 0.28855329751968384,
    "learning_rate": 0.00038098752990786943,
    "timestep_mean": 517.40625,
    "epsilon_norm": 7.925400257110596,
    "prediction_norm": 6.952162265777588,
    "timestamp": "2025-07-14T12:35:16.248084"
  },
  {
    "epoch": 2847,
    "batch_idx": 128,
    "loss": 0.3909866213798523,
    "learning_rate": 0.00038098752990786943,
    "timestep_mean": 462.34375,
    "epsilon_norm": 7.915761947631836,
    "prediction_norm": 5.677051067352295,
    "timestamp": "2025-07-14T12:35:16.253287"
  },
  {
    "epoch": 2847,
    "batch_idx": 160,
    "loss": 0.15066729485988617,
    "learning_rate": 0.00038098752990786943,
    "timestep_mean": 475.0,
    "epsilon_norm": 6.798133850097656,
    "prediction_norm": 6.261990070343018,
    "timestamp": "2025-07-14T12:35:16.258408"
  },
  {
    "epoch": 2847,
    "batch_idx": 192,
    "loss": 0.3690226376056671,
    "learning_rate": 0.00038098752990786943,
    "timestep_mean": 522.875,
    "epsilon_norm": 4.284289360046387,
    "prediction_norm": 3.5617082118988037,
    "timestamp": "2025-07-14T12:35:16.263630"
  },
  {
    "epoch": 2848,
    "batch_idx": 0,
    "loss": 0.22032126784324646,
    "learning_rate": 0.00038175050148813876,
    "timestep_mean": 553.25,
    "epsilon_norm": 8.124917030334473,
    "prediction_norm": 7.411962032318115,
    "timestamp": "2025-07-14T12:35:16.268843"
  },
  {
    "epoch": 2848,
    "batch_idx": 32,
    "loss": 0.14173546433448792,
    "learning_rate": 0.00038175050148813876,
    "timestep_mean": 586.6875,
    "epsilon_norm": 7.712856769561768,
    "prediction_norm": 6.850135326385498,
    "timestamp": "2025-07-14T12:35:16.273970"
  },
  {
    "epoch": 2848,
    "batch_idx": 64,
    "loss": 0.24697384238243103,
    "learning_rate": 0.00038175050148813876,
    "timestep_mean": 560.4375,
    "epsilon_norm": 8.570298194885254,
    "prediction_norm": 7.30129337310791,
    "timestamp": "2025-07-14T12:35:16.279113"
  },
  {
    "epoch": 2848,
    "batch_idx": 96,
    "loss": 0.21927079558372498,
    "learning_rate": 0.00038175050148813876,
    "timestep_mean": 524.53125,
    "epsilon_norm": 7.657611846923828,
    "prediction_norm": 6.5144124031066895,
    "timestamp": "2025-07-14T12:35:16.284442"
  },
  {
    "epoch": 2848,
    "batch_idx": 128,
    "loss": 0.2069021463394165,
    "learning_rate": 0.00038175050148813876,
    "timestep_mean": 484.625,
    "epsilon_norm": 8.17072868347168,
    "prediction_norm": 7.060092449188232,
    "timestamp": "2025-07-14T12:35:16.289615"
  },
  {
    "epoch": 2848,
    "batch_idx": 160,
    "loss": 0.16152258217334747,
    "learning_rate": 0.00038175050148813876,
    "timestep_mean": 504.25,
    "epsilon_norm": 7.8776044845581055,
    "prediction_norm": 6.684619426727295,
    "timestamp": "2025-07-14T12:35:16.294601"
  },
  {
    "epoch": 2848,
    "batch_idx": 192,
    "loss": 0.19976568222045898,
    "learning_rate": 0.00038175050148813876,
    "timestep_mean": 517.25,
    "epsilon_norm": 3.961259126663208,
    "prediction_norm": 3.4930670261383057,
    "timestamp": "2025-07-14T12:35:16.299463"
  },
  {
    "epoch": 2849,
    "batch_idx": 0,
    "loss": 0.13577833771705627,
    "learning_rate": 0.0003825137648372904,
    "timestep_mean": 548.6875,
    "epsilon_norm": 7.746874809265137,
    "prediction_norm": 7.266533851623535,
    "timestamp": "2025-07-14T12:35:16.304381"
  },
  {
    "epoch": 2849,
    "batch_idx": 32,
    "loss": 0.47292864322662354,
    "learning_rate": 0.0003825137648372904,
    "timestep_mean": 449.03125,
    "epsilon_norm": 8.271241188049316,
    "prediction_norm": 6.224088668823242,
    "timestamp": "2025-07-14T12:35:16.309266"
  },
  {
    "epoch": 2849,
    "batch_idx": 64,
    "loss": 0.38153356313705444,
    "learning_rate": 0.0003825137648372904,
    "timestep_mean": 475.78125,
    "epsilon_norm": 8.482617378234863,
    "prediction_norm": 6.376030445098877,
    "timestamp": "2025-07-14T12:35:16.314074"
  },
  {
    "epoch": 2849,
    "batch_idx": 96,
    "loss": 0.1529615819454193,
    "learning_rate": 0.0003825137648372904,
    "timestep_mean": 641.4375,
    "epsilon_norm": 7.237194538116455,
    "prediction_norm": 6.571131229400635,
    "timestamp": "2025-07-14T12:35:16.318921"
  },
  {
    "epoch": 2849,
    "batch_idx": 128,
    "loss": 0.22541412711143494,
    "learning_rate": 0.0003825137648372904,
    "timestep_mean": 566.53125,
    "epsilon_norm": 7.937262535095215,
    "prediction_norm": 6.848958969116211,
    "timestamp": "2025-07-14T12:35:16.323801"
  },
  {
    "epoch": 2849,
    "batch_idx": 160,
    "loss": 0.15467634797096252,
    "learning_rate": 0.0003825137648372904,
    "timestep_mean": 484.34375,
    "epsilon_norm": 7.638070106506348,
    "prediction_norm": 6.9900407791137695,
    "timestamp": "2025-07-14T12:35:16.328643"
  },
  {
    "epoch": 2849,
    "batch_idx": 192,
    "loss": 0.21341615915298462,
    "learning_rate": 0.0003825137648372904,
    "timestep_mean": 649.375,
    "epsilon_norm": 4.183794975280762,
    "prediction_norm": 4.240729331970215,
    "timestamp": "2025-07-14T12:35:16.333440"
  },
  {
    "epoch": 2850,
    "batch_idx": 0,
    "loss": 0.11570943146944046,
    "learning_rate": 0.00038327731807204793,
    "timestep_mean": 504.875,
    "epsilon_norm": 7.005609512329102,
    "prediction_norm": 6.008133888244629,
    "timestamp": "2025-07-14T12:35:16.338413"
  },
  {
    "epoch": 2850,
    "batch_idx": 32,
    "loss": 0.11789646744728088,
    "learning_rate": 0.00038327731807204793,
    "timestep_mean": 461.75,
    "epsilon_norm": 8.103434562683105,
    "prediction_norm": 7.726100921630859,
    "timestamp": "2025-07-14T12:35:16.343219"
  },
  {
    "epoch": 2850,
    "batch_idx": 64,
    "loss": 0.27598750591278076,
    "learning_rate": 0.00038327731807204793,
    "timestep_mean": 423.96875,
    "epsilon_norm": 8.046613693237305,
    "prediction_norm": 6.200974941253662,
    "timestamp": "2025-07-14T12:35:16.348141"
  },
  {
    "epoch": 2850,
    "batch_idx": 96,
    "loss": 0.14072547852993011,
    "learning_rate": 0.00038327731807204793,
    "timestep_mean": 559.28125,
    "epsilon_norm": 7.82393217086792,
    "prediction_norm": 7.326687812805176,
    "timestamp": "2025-07-14T12:35:16.353207"
  },
  {
    "epoch": 2850,
    "batch_idx": 128,
    "loss": 0.3110651671886444,
    "learning_rate": 0.00038327731807204793,
    "timestep_mean": 530.34375,
    "epsilon_norm": 7.813004493713379,
    "prediction_norm": 5.769692420959473,
    "timestamp": "2025-07-14T12:35:16.358157"
  },
  {
    "epoch": 2850,
    "batch_idx": 160,
    "loss": 0.15241384506225586,
    "learning_rate": 0.00038327731807204793,
    "timestep_mean": 484.53125,
    "epsilon_norm": 8.300091743469238,
    "prediction_norm": 7.174940586090088,
    "timestamp": "2025-07-14T12:35:16.363401"
  },
  {
    "epoch": 2850,
    "batch_idx": 192,
    "loss": 0.06971321254968643,
    "learning_rate": 0.00038327731807204793,
    "timestep_mean": 676.75,
    "epsilon_norm": 3.909109592437744,
    "prediction_norm": 3.9764747619628906,
    "timestamp": "2025-07-14T12:35:16.368374"
  },
  {
    "epoch": 2851,
    "batch_idx": 0,
    "loss": 0.3516092896461487,
    "learning_rate": 0.00038404115930842046,
    "timestep_mean": 501.21875,
    "epsilon_norm": 7.908421993255615,
    "prediction_norm": 6.742178916931152,
    "timestamp": "2025-07-14T12:35:16.373362"
  },
  {
    "epoch": 2851,
    "batch_idx": 32,
    "loss": 0.2899525463581085,
    "learning_rate": 0.00038404115930842046,
    "timestep_mean": 502.65625,
    "epsilon_norm": 7.7644805908203125,
    "prediction_norm": 6.77819299697876,
    "timestamp": "2025-07-14T12:35:16.378422"
  },
  {
    "epoch": 2851,
    "batch_idx": 64,
    "loss": 0.2489771991968155,
    "learning_rate": 0.00038404115930842046,
    "timestep_mean": 538.5625,
    "epsilon_norm": 7.590417861938477,
    "prediction_norm": 7.156810760498047,
    "timestamp": "2025-07-14T12:35:16.383490"
  },
  {
    "epoch": 2851,
    "batch_idx": 96,
    "loss": 0.22033043205738068,
    "learning_rate": 0.00038404115930842046,
    "timestep_mean": 528.59375,
    "epsilon_norm": 7.305678844451904,
    "prediction_norm": 6.305896282196045,
    "timestamp": "2025-07-14T12:35:16.388587"
  },
  {
    "epoch": 2851,
    "batch_idx": 128,
    "loss": 0.6566540598869324,
    "learning_rate": 0.00038404115930842046,
    "timestep_mean": 388.5,
    "epsilon_norm": 9.025480270385742,
    "prediction_norm": 6.569200038909912,
    "timestamp": "2025-07-14T12:35:16.393759"
  },
  {
    "epoch": 2851,
    "batch_idx": 160,
    "loss": 0.5939830541610718,
    "learning_rate": 0.00038404115930842046,
    "timestep_mean": 404.90625,
    "epsilon_norm": 8.6392240524292,
    "prediction_norm": 6.902834415435791,
    "timestamp": "2025-07-14T12:35:16.398900"
  },
  {
    "epoch": 2851,
    "batch_idx": 192,
    "loss": 0.1248495802283287,
    "learning_rate": 0.00038404115930842046,
    "timestep_mean": 526.75,
    "epsilon_norm": 3.8330845832824707,
    "prediction_norm": 3.665935754776001,
    "timestamp": "2025-07-14T12:35:16.403891"
  },
  {
    "epoch": 2852,
    "batch_idx": 0,
    "loss": 0.13065916299819946,
    "learning_rate": 0.00038480528666170577,
    "timestep_mean": 632.21875,
    "epsilon_norm": 8.987627983093262,
    "prediction_norm": 7.855774402618408,
    "timestamp": "2025-07-14T12:35:16.408839"
  },
  {
    "epoch": 2852,
    "batch_idx": 32,
    "loss": 0.24549755454063416,
    "learning_rate": 0.00038480528666170577,
    "timestep_mean": 585.15625,
    "epsilon_norm": 8.6778564453125,
    "prediction_norm": 7.243473529815674,
    "timestamp": "2025-07-14T12:35:16.413787"
  },
  {
    "epoch": 2852,
    "batch_idx": 64,
    "loss": 0.2623676359653473,
    "learning_rate": 0.00038480528666170577,
    "timestep_mean": 550.65625,
    "epsilon_norm": 8.277352333068848,
    "prediction_norm": 7.795319557189941,
    "timestamp": "2025-07-14T12:35:16.418974"
  },
  {
    "epoch": 2852,
    "batch_idx": 96,
    "loss": 0.2770019769668579,
    "learning_rate": 0.00038480528666170577,
    "timestep_mean": 515.25,
    "epsilon_norm": 8.359333992004395,
    "prediction_norm": 6.879034519195557,
    "timestamp": "2025-07-14T12:35:16.424169"
  },
  {
    "epoch": 2852,
    "batch_idx": 128,
    "loss": 0.2191895991563797,
    "learning_rate": 0.00038480528666170577,
    "timestep_mean": 557.46875,
    "epsilon_norm": 8.075435638427734,
    "prediction_norm": 6.6511359214782715,
    "timestamp": "2025-07-14T12:35:16.429368"
  },
  {
    "epoch": 2852,
    "batch_idx": 160,
    "loss": 0.2646018862724304,
    "learning_rate": 0.00038480528666170577,
    "timestep_mean": 469.625,
    "epsilon_norm": 7.46146821975708,
    "prediction_norm": 6.994105339050293,
    "timestamp": "2025-07-14T12:35:16.434532"
  },
  {
    "epoch": 2852,
    "batch_idx": 192,
    "loss": 0.24009960889816284,
    "learning_rate": 0.00038480528666170577,
    "timestep_mean": 496.625,
    "epsilon_norm": 3.3335533142089844,
    "prediction_norm": 2.9689648151397705,
    "timestamp": "2025-07-14T12:35:16.439960"
  },
  {
    "epoch": 2853,
    "batch_idx": 0,
    "loss": 0.22935479879379272,
    "learning_rate": 0.0003855696982464945,
    "timestep_mean": 516.75,
    "epsilon_norm": 7.768913269042969,
    "prediction_norm": 7.442848205566406,
    "timestamp": "2025-07-14T12:35:16.445174"
  },
  {
    "epoch": 2853,
    "batch_idx": 32,
    "loss": 0.44248729944229126,
    "learning_rate": 0.0003855696982464945,
    "timestep_mean": 426.28125,
    "epsilon_norm": 8.430699348449707,
    "prediction_norm": 6.630124092102051,
    "timestamp": "2025-07-14T12:35:16.450439"
  },
  {
    "epoch": 2853,
    "batch_idx": 64,
    "loss": 0.29891282320022583,
    "learning_rate": 0.0003855696982464945,
    "timestep_mean": 497.375,
    "epsilon_norm": 6.726526737213135,
    "prediction_norm": 6.318124771118164,
    "timestamp": "2025-07-14T12:35:16.455658"
  },
  {
    "epoch": 2853,
    "batch_idx": 96,
    "loss": 0.36075279116630554,
    "learning_rate": 0.0003855696982464945,
    "timestep_mean": 451.59375,
    "epsilon_norm": 8.590883255004883,
    "prediction_norm": 7.341552257537842,
    "timestamp": "2025-07-14T12:35:16.460887"
  },
  {
    "epoch": 2853,
    "batch_idx": 128,
    "loss": 0.25971758365631104,
    "learning_rate": 0.0003855696982464945,
    "timestep_mean": 586.5,
    "epsilon_norm": 7.500260353088379,
    "prediction_norm": 6.635161876678467,
    "timestamp": "2025-07-14T12:35:16.466019"
  },
  {
    "epoch": 2853,
    "batch_idx": 160,
    "loss": 0.12131708860397339,
    "learning_rate": 0.0003855696982464945,
    "timestep_mean": 546.34375,
    "epsilon_norm": 7.53031587600708,
    "prediction_norm": 7.041249752044678,
    "timestamp": "2025-07-14T12:35:16.471291"
  },
  {
    "epoch": 2853,
    "batch_idx": 192,
    "loss": 0.3639792203903198,
    "learning_rate": 0.0003855696982464945,
    "timestep_mean": 475.875,
    "epsilon_norm": 3.5355124473571777,
    "prediction_norm": 2.859931468963623,
    "timestamp": "2025-07-14T12:35:16.476425"
  },
  {
    "epoch": 2854,
    "batch_idx": 0,
    "loss": 0.437894344329834,
    "learning_rate": 0.00038633439217667764,
    "timestep_mean": 397.28125,
    "epsilon_norm": 8.36280345916748,
    "prediction_norm": 5.156986236572266,
    "timestamp": "2025-07-14T12:35:16.481502"
  },
  {
    "epoch": 2854,
    "batch_idx": 32,
    "loss": 0.42325344681739807,
    "learning_rate": 0.00038633439217667764,
    "timestep_mean": 399.125,
    "epsilon_norm": 7.327053070068359,
    "prediction_norm": 5.3995561599731445,
    "timestamp": "2025-07-14T12:35:16.486721"
  },
  {
    "epoch": 2854,
    "batch_idx": 64,
    "loss": 0.20621293783187866,
    "learning_rate": 0.00038633439217667764,
    "timestep_mean": 500.84375,
    "epsilon_norm": 7.821346282958984,
    "prediction_norm": 6.259978771209717,
    "timestamp": "2025-07-14T12:35:16.492130"
  },
  {
    "epoch": 2854,
    "batch_idx": 96,
    "loss": 0.30273935198783875,
    "learning_rate": 0.00038633439217667764,
    "timestep_mean": 518.0,
    "epsilon_norm": 8.834086418151855,
    "prediction_norm": 7.7926025390625,
    "timestamp": "2025-07-14T12:35:16.497042"
  },
  {
    "epoch": 2854,
    "batch_idx": 128,
    "loss": 0.30657151341438293,
    "learning_rate": 0.00038633439217667764,
    "timestep_mean": 503.15625,
    "epsilon_norm": 7.458073139190674,
    "prediction_norm": 5.735569477081299,
    "timestamp": "2025-07-14T12:35:16.502134"
  },
  {
    "epoch": 2854,
    "batch_idx": 160,
    "loss": 0.2945014238357544,
    "learning_rate": 0.00038633439217667764,
    "timestep_mean": 453.46875,
    "epsilon_norm": 8.946622848510742,
    "prediction_norm": 6.859045028686523,
    "timestamp": "2025-07-14T12:35:16.507340"
  },
  {
    "epoch": 2854,
    "batch_idx": 192,
    "loss": 0.16570034623146057,
    "learning_rate": 0.00038633439217667764,
    "timestep_mean": 599.5,
    "epsilon_norm": 4.197937965393066,
    "prediction_norm": 3.626310348510742,
    "timestamp": "2025-07-14T12:35:16.512459"
  },
  {
    "epoch": 2855,
    "batch_idx": 0,
    "loss": 0.1780158281326294,
    "learning_rate": 0.000387099366565449,
    "timestep_mean": 521.96875,
    "epsilon_norm": 7.343600749969482,
    "prediction_norm": 6.777482032775879,
    "timestamp": "2025-07-14T12:35:16.517754"
  },
  {
    "epoch": 2855,
    "batch_idx": 32,
    "loss": 0.4536494016647339,
    "learning_rate": 0.000387099366565449,
    "timestep_mean": 449.21875,
    "epsilon_norm": 9.41794204711914,
    "prediction_norm": 7.189813613891602,
    "timestamp": "2025-07-14T12:35:16.522927"
  },
  {
    "epoch": 2855,
    "batch_idx": 64,
    "loss": 0.22539590299129486,
    "learning_rate": 0.000387099366565449,
    "timestep_mean": 547.25,
    "epsilon_norm": 7.611543655395508,
    "prediction_norm": 6.7726006507873535,
    "timestamp": "2025-07-14T12:35:16.528005"
  },
  {
    "epoch": 2855,
    "batch_idx": 96,
    "loss": 0.1710154116153717,
    "learning_rate": 0.000387099366565449,
    "timestep_mean": 609.71875,
    "epsilon_norm": 7.785595417022705,
    "prediction_norm": 6.890024662017822,
    "timestamp": "2025-07-14T12:35:16.533149"
  },
  {
    "epoch": 2855,
    "batch_idx": 128,
    "loss": 0.267988383769989,
    "learning_rate": 0.000387099366565449,
    "timestep_mean": 392.6875,
    "epsilon_norm": 7.223398685455322,
    "prediction_norm": 5.7250142097473145,
    "timestamp": "2025-07-14T12:35:16.538465"
  },
  {
    "epoch": 2855,
    "batch_idx": 160,
    "loss": 0.352899432182312,
    "learning_rate": 0.000387099366565449,
    "timestep_mean": 463.40625,
    "epsilon_norm": 8.640035629272461,
    "prediction_norm": 6.831094741821289,
    "timestamp": "2025-07-14T12:35:16.543447"
  },
  {
    "epoch": 2855,
    "batch_idx": 192,
    "loss": 0.15501850843429565,
    "learning_rate": 0.000387099366565449,
    "timestep_mean": 430.125,
    "epsilon_norm": 3.717717170715332,
    "prediction_norm": 3.0598418712615967,
    "timestamp": "2025-07-14T12:35:16.548714"
  },
  {
    "epoch": 2856,
    "batch_idx": 0,
    "loss": 0.4848950505256653,
    "learning_rate": 0.0003878646195253102,
    "timestep_mean": 504.75,
    "epsilon_norm": 8.901570320129395,
    "prediction_norm": 7.4085493087768555,
    "timestamp": "2025-07-14T12:35:16.553921"
  },
  {
    "epoch": 2856,
    "batch_idx": 32,
    "loss": 0.24906837940216064,
    "learning_rate": 0.0003878646195253102,
    "timestep_mean": 512.3125,
    "epsilon_norm": 9.239217758178711,
    "prediction_norm": 8.195843696594238,
    "timestamp": "2025-07-14T12:35:16.559106"
  },
  {
    "epoch": 2856,
    "batch_idx": 64,
    "loss": 0.29552289843559265,
    "learning_rate": 0.0003878646195253102,
    "timestep_mean": 489.3125,
    "epsilon_norm": 7.449398517608643,
    "prediction_norm": 6.331818580627441,
    "timestamp": "2025-07-14T12:35:16.564360"
  },
  {
    "epoch": 2856,
    "batch_idx": 96,
    "loss": 0.39700525999069214,
    "learning_rate": 0.0003878646195253102,
    "timestep_mean": 408.6875,
    "epsilon_norm": 7.800232410430908,
    "prediction_norm": 6.014807224273682,
    "timestamp": "2025-07-14T12:35:16.569559"
  },
  {
    "epoch": 2856,
    "batch_idx": 128,
    "loss": 0.43633586168289185,
    "learning_rate": 0.0003878646195253102,
    "timestep_mean": 511.9375,
    "epsilon_norm": 9.754369735717773,
    "prediction_norm": 8.16352653503418,
    "timestamp": "2025-07-14T12:35:16.574625"
  },
  {
    "epoch": 2856,
    "batch_idx": 160,
    "loss": 0.155268132686615,
    "learning_rate": 0.0003878646195253102,
    "timestep_mean": 502.90625,
    "epsilon_norm": 8.099445343017578,
    "prediction_norm": 7.282032489776611,
    "timestamp": "2025-07-14T12:35:16.579904"
  },
  {
    "epoch": 2856,
    "batch_idx": 192,
    "loss": 0.42558085918426514,
    "learning_rate": 0.0003878646195253102,
    "timestep_mean": 386.125,
    "epsilon_norm": 4.395755767822266,
    "prediction_norm": 3.69704270362854,
    "timestamp": "2025-07-14T12:35:16.585068"
  },
  {
    "epoch": 2857,
    "batch_idx": 0,
    "loss": 0.255370169878006,
    "learning_rate": 0.0003886301491680756,
    "timestep_mean": 438.78125,
    "epsilon_norm": 8.078361511230469,
    "prediction_norm": 7.212035179138184,
    "timestamp": "2025-07-14T12:35:16.590145"
  },
  {
    "epoch": 2857,
    "batch_idx": 32,
    "loss": 0.17709483206272125,
    "learning_rate": 0.0003886301491680756,
    "timestep_mean": 536.09375,
    "epsilon_norm": 7.920083045959473,
    "prediction_norm": 7.580747604370117,
    "timestamp": "2025-07-14T12:35:16.595401"
  },
  {
    "epoch": 2857,
    "batch_idx": 64,
    "loss": 0.14731921255588531,
    "learning_rate": 0.0003886301491680756,
    "timestep_mean": 508.21875,
    "epsilon_norm": 6.93784761428833,
    "prediction_norm": 6.387513637542725,
    "timestamp": "2025-07-14T12:35:16.600673"
  },
  {
    "epoch": 2857,
    "batch_idx": 96,
    "loss": 0.17577514052391052,
    "learning_rate": 0.0003886301491680756,
    "timestep_mean": 477.4375,
    "epsilon_norm": 8.350642204284668,
    "prediction_norm": 7.923811435699463,
    "timestamp": "2025-07-14T12:35:16.605756"
  },
  {
    "epoch": 2857,
    "batch_idx": 128,
    "loss": 0.2734403908252716,
    "learning_rate": 0.0003886301491680756,
    "timestep_mean": 512.875,
    "epsilon_norm": 6.974729537963867,
    "prediction_norm": 6.229692459106445,
    "timestamp": "2025-07-14T12:35:16.610930"
  },
  {
    "epoch": 2857,
    "batch_idx": 160,
    "loss": 0.1912284791469574,
    "learning_rate": 0.0003886301491680756,
    "timestep_mean": 478.8125,
    "epsilon_norm": 6.827847480773926,
    "prediction_norm": 6.406742572784424,
    "timestamp": "2025-07-14T12:35:16.616118"
  },
  {
    "epoch": 2857,
    "batch_idx": 192,
    "loss": 0.09433730691671371,
    "learning_rate": 0.0003886301491680756,
    "timestep_mean": 601.375,
    "epsilon_norm": 3.8711116313934326,
    "prediction_norm": 4.080130577087402,
    "timestamp": "2025-07-14T12:35:16.621286"
  },
  {
    "epoch": 2858,
    "batch_idx": 0,
    "loss": 0.17221765220165253,
    "learning_rate": 0.0003893959536048776,
    "timestep_mean": 504.21875,
    "epsilon_norm": 8.564513206481934,
    "prediction_norm": 8.142740249633789,
    "timestamp": "2025-07-14T12:35:16.626484"
  },
  {
    "epoch": 2858,
    "batch_idx": 32,
    "loss": 0.30688154697418213,
    "learning_rate": 0.0003893959536048776,
    "timestep_mean": 498.5,
    "epsilon_norm": 8.569497108459473,
    "prediction_norm": 6.879276275634766,
    "timestamp": "2025-07-14T12:35:16.631612"
  },
  {
    "epoch": 2858,
    "batch_idx": 64,
    "loss": 0.21206910908222198,
    "learning_rate": 0.0003893959536048776,
    "timestep_mean": 480.9375,
    "epsilon_norm": 7.204751014709473,
    "prediction_norm": 5.840057849884033,
    "timestamp": "2025-07-14T12:35:16.636745"
  },
  {
    "epoch": 2858,
    "batch_idx": 96,
    "loss": 0.4494507312774658,
    "learning_rate": 0.0003893959536048776,
    "timestep_mean": 512.4375,
    "epsilon_norm": 8.983444213867188,
    "prediction_norm": 6.726140022277832,
    "timestamp": "2025-07-14T12:35:16.642099"
  },
  {
    "epoch": 2858,
    "batch_idx": 128,
    "loss": 0.12304653227329254,
    "learning_rate": 0.0003893959536048776,
    "timestep_mean": 514.9375,
    "epsilon_norm": 7.019356727600098,
    "prediction_norm": 6.4587507247924805,
    "timestamp": "2025-07-14T12:35:16.647318"
  },
  {
    "epoch": 2858,
    "batch_idx": 160,
    "loss": 0.23443084955215454,
    "learning_rate": 0.0003893959536048776,
    "timestep_mean": 533.375,
    "epsilon_norm": 7.120704174041748,
    "prediction_norm": 6.091560363769531,
    "timestamp": "2025-07-14T12:35:16.652372"
  },
  {
    "epoch": 2858,
    "batch_idx": 192,
    "loss": 0.48703533411026,
    "learning_rate": 0.0003893959536048776,
    "timestep_mean": 410.25,
    "epsilon_norm": 3.7236721515655518,
    "prediction_norm": 2.5212831497192383,
    "timestamp": "2025-07-14T12:35:16.657438"
  },
  {
    "epoch": 2859,
    "batch_idx": 0,
    "loss": 0.29580631852149963,
    "learning_rate": 0.0003901620309461687,
    "timestep_mean": 512.4375,
    "epsilon_norm": 9.321309089660645,
    "prediction_norm": 7.908811569213867,
    "timestamp": "2025-07-14T12:35:16.662590"
  },
  {
    "epoch": 2859,
    "batch_idx": 32,
    "loss": 0.3006749451160431,
    "learning_rate": 0.0003901620309461687,
    "timestep_mean": 477.625,
    "epsilon_norm": 8.18144702911377,
    "prediction_norm": 6.056455612182617,
    "timestamp": "2025-07-14T12:35:16.667837"
  },
  {
    "epoch": 2859,
    "batch_idx": 64,
    "loss": 0.11929374933242798,
    "learning_rate": 0.0003901620309461687,
    "timestep_mean": 466.8125,
    "epsilon_norm": 7.583273887634277,
    "prediction_norm": 6.7178635597229,
    "timestamp": "2025-07-14T12:35:16.673094"
  },
  {
    "epoch": 2859,
    "batch_idx": 96,
    "loss": 0.32418292760849,
    "learning_rate": 0.0003901620309461687,
    "timestep_mean": 526.9375,
    "epsilon_norm": 7.780968189239502,
    "prediction_norm": 5.942628383636475,
    "timestamp": "2025-07-14T12:35:16.678241"
  },
  {
    "epoch": 2859,
    "batch_idx": 128,
    "loss": 0.25985127687454224,
    "learning_rate": 0.0003901620309461687,
    "timestep_mean": 499.5625,
    "epsilon_norm": 7.5945329666137695,
    "prediction_norm": 5.9864583015441895,
    "timestamp": "2025-07-14T12:35:16.683330"
  },
  {
    "epoch": 2859,
    "batch_idx": 160,
    "loss": 0.2467328906059265,
    "learning_rate": 0.0003901620309461687,
    "timestep_mean": 525.71875,
    "epsilon_norm": 7.12263822555542,
    "prediction_norm": 6.196779727935791,
    "timestamp": "2025-07-14T12:35:16.688522"
  },
  {
    "epoch": 2859,
    "batch_idx": 192,
    "loss": 0.7545676231384277,
    "learning_rate": 0.0003901620309461687,
    "timestep_mean": 400.25,
    "epsilon_norm": 4.607665061950684,
    "prediction_norm": 3.4170010089874268,
    "timestamp": "2025-07-14T12:35:16.693540"
  },
  {
    "epoch": 2860,
    "batch_idx": 0,
    "loss": 0.16084106266498566,
    "learning_rate": 0.00039092837930172975,
    "timestep_mean": 533.59375,
    "epsilon_norm": 8.341078758239746,
    "prediction_norm": 7.252120018005371,
    "timestamp": "2025-07-14T12:35:16.698752"
  },
  {
    "epoch": 2860,
    "batch_idx": 32,
    "loss": 0.2509039044380188,
    "learning_rate": 0.00039092837930172975,
    "timestep_mean": 489.625,
    "epsilon_norm": 7.475105285644531,
    "prediction_norm": 5.921060562133789,
    "timestamp": "2025-07-14T12:35:16.703973"
  },
  {
    "epoch": 2860,
    "batch_idx": 64,
    "loss": 0.36269837617874146,
    "learning_rate": 0.00039092837930172975,
    "timestep_mean": 470.3125,
    "epsilon_norm": 7.719167709350586,
    "prediction_norm": 5.813551425933838,
    "timestamp": "2025-07-14T12:35:16.709258"
  },
  {
    "epoch": 2860,
    "batch_idx": 96,
    "loss": 0.4097508490085602,
    "learning_rate": 0.00039092837930172975,
    "timestep_mean": 499.78125,
    "epsilon_norm": 7.1769866943359375,
    "prediction_norm": 4.98961877822876,
    "timestamp": "2025-07-14T12:35:16.714343"
  },
  {
    "epoch": 2860,
    "batch_idx": 128,
    "loss": 0.289257287979126,
    "learning_rate": 0.00039092837930172975,
    "timestep_mean": 462.5625,
    "epsilon_norm": 7.1933393478393555,
    "prediction_norm": 5.673839569091797,
    "timestamp": "2025-07-14T12:35:16.719364"
  },
  {
    "epoch": 2860,
    "batch_idx": 160,
    "loss": 0.18241336941719055,
    "learning_rate": 0.00039092837930172975,
    "timestep_mean": 598.125,
    "epsilon_norm": 6.941501140594482,
    "prediction_norm": 5.8753790855407715,
    "timestamp": "2025-07-14T12:35:16.724456"
  },
  {
    "epoch": 2860,
    "batch_idx": 192,
    "loss": 0.22650879621505737,
    "learning_rate": 0.00039092837930172975,
    "timestep_mean": 479.5,
    "epsilon_norm": 2.3307838439941406,
    "prediction_norm": 1.8526813983917236,
    "timestamp": "2025-07-14T12:35:16.729688"
  },
  {
    "epoch": 2861,
    "batch_idx": 0,
    "loss": 0.3744642734527588,
    "learning_rate": 0.000391694996780672,
    "timestep_mean": 490.78125,
    "epsilon_norm": 8.522732734680176,
    "prediction_norm": 6.524815559387207,
    "timestamp": "2025-07-14T12:35:16.734823"
  },
  {
    "epoch": 2861,
    "batch_idx": 32,
    "loss": 0.24775120615959167,
    "learning_rate": 0.000391694996780672,
    "timestep_mean": 537.125,
    "epsilon_norm": 7.381292343139648,
    "prediction_norm": 5.9525041580200195,
    "timestamp": "2025-07-14T12:35:16.740020"
  },
  {
    "epoch": 2861,
    "batch_idx": 64,
    "loss": 0.2745468020439148,
    "learning_rate": 0.000391694996780672,
    "timestep_mean": 430.5,
    "epsilon_norm": 7.125430583953857,
    "prediction_norm": 6.31449556350708,
    "timestamp": "2025-07-14T12:35:16.745119"
  },
  {
    "epoch": 2861,
    "batch_idx": 96,
    "loss": 0.24910613894462585,
    "learning_rate": 0.000391694996780672,
    "timestep_mean": 508.0,
    "epsilon_norm": 7.850162029266357,
    "prediction_norm": 6.993144989013672,
    "timestamp": "2025-07-14T12:35:16.750428"
  },
  {
    "epoch": 2861,
    "batch_idx": 128,
    "loss": 0.3252498507499695,
    "learning_rate": 0.000391694996780672,
    "timestep_mean": 476.59375,
    "epsilon_norm": 8.191744804382324,
    "prediction_norm": 5.7530412673950195,
    "timestamp": "2025-07-14T12:35:16.755648"
  },
  {
    "epoch": 2861,
    "batch_idx": 160,
    "loss": 0.2923770248889923,
    "learning_rate": 0.000391694996780672,
    "timestep_mean": 483.9375,
    "epsilon_norm": 7.6563615798950195,
    "prediction_norm": 5.823151588439941,
    "timestamp": "2025-07-14T12:35:16.760753"
  },
  {
    "epoch": 2861,
    "batch_idx": 192,
    "loss": 0.5563406944274902,
    "learning_rate": 0.000391694996780672,
    "timestep_mean": 273.5,
    "epsilon_norm": 3.9820549488067627,
    "prediction_norm": 2.536304473876953,
    "timestamp": "2025-07-14T12:35:16.766001"
  },
  {
    "epoch": 2862,
    "batch_idx": 0,
    "loss": 0.2979457378387451,
    "learning_rate": 0.0003924618814914442,
    "timestep_mean": 502.5625,
    "epsilon_norm": 8.932665824890137,
    "prediction_norm": 6.943233013153076,
    "timestamp": "2025-07-14T12:35:16.771212"
  },
  {
    "epoch": 2862,
    "batch_idx": 32,
    "loss": 0.16395695507526398,
    "learning_rate": 0.0003924618814914442,
    "timestep_mean": 511.65625,
    "epsilon_norm": 7.230569839477539,
    "prediction_norm": 6.231982231140137,
    "timestamp": "2025-07-14T12:35:16.776315"
  },
  {
    "epoch": 2862,
    "batch_idx": 64,
    "loss": 0.20900174975395203,
    "learning_rate": 0.0003924618814914442,
    "timestep_mean": 540.65625,
    "epsilon_norm": 7.415683269500732,
    "prediction_norm": 6.3259453773498535,
    "timestamp": "2025-07-14T12:35:16.781511"
  },
  {
    "epoch": 2862,
    "batch_idx": 96,
    "loss": 0.18106435239315033,
    "learning_rate": 0.0003924618814914442,
    "timestep_mean": 494.6875,
    "epsilon_norm": 6.717960357666016,
    "prediction_norm": 5.996159076690674,
    "timestamp": "2025-07-14T12:35:16.786693"
  },
  {
    "epoch": 2862,
    "batch_idx": 128,
    "loss": 0.3207513093948364,
    "learning_rate": 0.0003924618814914442,
    "timestep_mean": 484.4375,
    "epsilon_norm": 7.23450231552124,
    "prediction_norm": 6.339479923248291,
    "timestamp": "2025-07-14T12:35:16.791862"
  },
  {
    "epoch": 2862,
    "batch_idx": 160,
    "loss": 0.18066391348838806,
    "learning_rate": 0.0003924618814914442,
    "timestep_mean": 542.03125,
    "epsilon_norm": 7.6009416580200195,
    "prediction_norm": 6.839050769805908,
    "timestamp": "2025-07-14T12:35:16.797049"
  },
  {
    "epoch": 2862,
    "batch_idx": 192,
    "loss": 0.04414349049329758,
    "learning_rate": 0.0003924618814914442,
    "timestep_mean": 686.5,
    "epsilon_norm": 4.375096797943115,
    "prediction_norm": 4.293356895446777,
    "timestamp": "2025-07-14T12:35:16.802030"
  },
  {
    "epoch": 2863,
    "batch_idx": 0,
    "loss": 0.1176418885588646,
    "learning_rate": 0.0003932290315418334,
    "timestep_mean": 538.3125,
    "epsilon_norm": 7.123479843139648,
    "prediction_norm": 7.339165687561035,
    "timestamp": "2025-07-14T12:35:16.806918"
  },
  {
    "epoch": 2863,
    "batch_idx": 32,
    "loss": 0.1662212312221527,
    "learning_rate": 0.0003932290315418334,
    "timestep_mean": 595.53125,
    "epsilon_norm": 8.288677215576172,
    "prediction_norm": 7.670126438140869,
    "timestamp": "2025-07-14T12:35:16.811819"
  },
  {
    "epoch": 2863,
    "batch_idx": 64,
    "loss": 0.20413807034492493,
    "learning_rate": 0.0003932290315418334,
    "timestep_mean": 517.125,
    "epsilon_norm": 8.523303985595703,
    "prediction_norm": 7.732648849487305,
    "timestamp": "2025-07-14T12:35:16.816723"
  },
  {
    "epoch": 2863,
    "batch_idx": 96,
    "loss": 0.2117786705493927,
    "learning_rate": 0.0003932290315418334,
    "timestep_mean": 467.5,
    "epsilon_norm": 8.270966529846191,
    "prediction_norm": 7.048755645751953,
    "timestamp": "2025-07-14T12:35:16.821570"
  },
  {
    "epoch": 2863,
    "batch_idx": 128,
    "loss": 0.3378351926803589,
    "learning_rate": 0.0003932290315418334,
    "timestep_mean": 534.71875,
    "epsilon_norm": 7.373141288757324,
    "prediction_norm": 6.572807312011719,
    "timestamp": "2025-07-14T12:35:16.826411"
  },
  {
    "epoch": 2863,
    "batch_idx": 160,
    "loss": 0.18387679755687714,
    "learning_rate": 0.0003932290315418334,
    "timestep_mean": 572.0625,
    "epsilon_norm": 7.560541152954102,
    "prediction_norm": 6.644730567932129,
    "timestamp": "2025-07-14T12:35:16.831242"
  },
  {
    "epoch": 2863,
    "batch_idx": 192,
    "loss": 0.2556760907173157,
    "learning_rate": 0.0003932290315418334,
    "timestep_mean": 586.5,
    "epsilon_norm": 4.7952752113342285,
    "prediction_norm": 4.058026313781738,
    "timestamp": "2025-07-14T12:35:16.836120"
  },
  {
    "epoch": 2864,
    "batch_idx": 0,
    "loss": 0.45634499192237854,
    "learning_rate": 0.0003939964450389735,
    "timestep_mean": 468.125,
    "epsilon_norm": 9.396149635314941,
    "prediction_norm": 7.357842445373535,
    "timestamp": "2025-07-14T12:35:16.840948"
  },
  {
    "epoch": 2864,
    "batch_idx": 32,
    "loss": 0.2562717795372009,
    "learning_rate": 0.0003939964450389735,
    "timestep_mean": 585.34375,
    "epsilon_norm": 8.328107833862305,
    "prediction_norm": 7.2427167892456055,
    "timestamp": "2025-07-14T12:35:16.845748"
  },
  {
    "epoch": 2864,
    "batch_idx": 64,
    "loss": 0.3162686824798584,
    "learning_rate": 0.0003939964450389735,
    "timestep_mean": 453.78125,
    "epsilon_norm": 7.641247749328613,
    "prediction_norm": 5.815584182739258,
    "timestamp": "2025-07-14T12:35:16.850732"
  },
  {
    "epoch": 2864,
    "batch_idx": 96,
    "loss": 0.18807995319366455,
    "learning_rate": 0.0003939964450389735,
    "timestep_mean": 498.4375,
    "epsilon_norm": 8.191950798034668,
    "prediction_norm": 6.819892406463623,
    "timestamp": "2025-07-14T12:35:16.855545"
  },
  {
    "epoch": 2864,
    "batch_idx": 128,
    "loss": 0.2593575417995453,
    "learning_rate": 0.0003939964450389735,
    "timestep_mean": 457.59375,
    "epsilon_norm": 7.959728717803955,
    "prediction_norm": 6.409552574157715,
    "timestamp": "2025-07-14T12:35:16.860374"
  },
  {
    "epoch": 2864,
    "batch_idx": 160,
    "loss": 0.21988770365715027,
    "learning_rate": 0.0003939964450389735,
    "timestep_mean": 458.1875,
    "epsilon_norm": 7.722383975982666,
    "prediction_norm": 5.766551971435547,
    "timestamp": "2025-07-14T12:35:16.865246"
  },
  {
    "epoch": 2864,
    "batch_idx": 192,
    "loss": 0.11252724379301071,
    "learning_rate": 0.0003939964450389735,
    "timestep_mean": 629.75,
    "epsilon_norm": 4.655211448669434,
    "prediction_norm": 4.658327579498291,
    "timestamp": "2025-07-14T12:35:16.870084"
  },
  {
    "epoch": 2865,
    "batch_idx": 0,
    "loss": 0.2504968047142029,
    "learning_rate": 0.0003947641200893484,
    "timestep_mean": 598.96875,
    "epsilon_norm": 7.406142234802246,
    "prediction_norm": 6.354543685913086,
    "timestamp": "2025-07-14T12:35:16.874992"
  },
  {
    "epoch": 2865,
    "batch_idx": 32,
    "loss": 0.19982804358005524,
    "learning_rate": 0.0003947641200893484,
    "timestep_mean": 429.9375,
    "epsilon_norm": 7.783420562744141,
    "prediction_norm": 6.921262264251709,
    "timestamp": "2025-07-14T12:35:16.880152"
  },
  {
    "epoch": 2865,
    "batch_idx": 64,
    "loss": 0.12770019471645355,
    "learning_rate": 0.0003947641200893484,
    "timestep_mean": 555.78125,
    "epsilon_norm": 6.914035797119141,
    "prediction_norm": 6.535306453704834,
    "timestamp": "2025-07-14T12:35:16.885050"
  },
  {
    "epoch": 2865,
    "batch_idx": 96,
    "loss": 0.19927793741226196,
    "learning_rate": 0.0003947641200893484,
    "timestep_mean": 555.875,
    "epsilon_norm": 8.074556350708008,
    "prediction_norm": 6.93092679977417,
    "timestamp": "2025-07-14T12:35:16.890201"
  },
  {
    "epoch": 2865,
    "batch_idx": 128,
    "loss": 0.16787037253379822,
    "learning_rate": 0.0003947641200893484,
    "timestep_mean": 581.5,
    "epsilon_norm": 7.827860355377197,
    "prediction_norm": 7.318520545959473,
    "timestamp": "2025-07-14T12:35:16.895145"
  },
  {
    "epoch": 2865,
    "batch_idx": 160,
    "loss": 0.3084980845451355,
    "learning_rate": 0.0003947641200893484,
    "timestep_mean": 467.15625,
    "epsilon_norm": 8.393965721130371,
    "prediction_norm": 7.0917205810546875,
    "timestamp": "2025-07-14T12:35:16.900401"
  },
  {
    "epoch": 2865,
    "batch_idx": 192,
    "loss": 0.6246539950370789,
    "learning_rate": 0.0003947641200893484,
    "timestep_mean": 337.625,
    "epsilon_norm": 4.312653064727783,
    "prediction_norm": 2.403085708618164,
    "timestamp": "2025-07-14T12:35:16.905506"
  },
  {
    "epoch": 2866,
    "batch_idx": 0,
    "loss": 0.3157827854156494,
    "learning_rate": 0.0003955320547987949,
    "timestep_mean": 498.125,
    "epsilon_norm": 8.757195472717285,
    "prediction_norm": 6.983372688293457,
    "timestamp": "2025-07-14T12:35:16.910493"
  },
  {
    "epoch": 2866,
    "batch_idx": 32,
    "loss": 0.3006299138069153,
    "learning_rate": 0.0003955320547987949,
    "timestep_mean": 452.125,
    "epsilon_norm": 8.681656837463379,
    "prediction_norm": 7.5162787437438965,
    "timestamp": "2025-07-14T12:35:16.915683"
  },
  {
    "epoch": 2866,
    "batch_idx": 64,
    "loss": 0.33197659254074097,
    "learning_rate": 0.0003955320547987949,
    "timestep_mean": 479.09375,
    "epsilon_norm": 6.810145378112793,
    "prediction_norm": 5.692487716674805,
    "timestamp": "2025-07-14T12:35:16.920916"
  },
  {
    "epoch": 2866,
    "batch_idx": 96,
    "loss": 0.3856247663497925,
    "learning_rate": 0.0003955320547987949,
    "timestep_mean": 469.0,
    "epsilon_norm": 7.929693222045898,
    "prediction_norm": 5.561272144317627,
    "timestamp": "2025-07-14T12:35:16.926052"
  },
  {
    "epoch": 2866,
    "batch_idx": 128,
    "loss": 0.35577040910720825,
    "learning_rate": 0.0003955320547987949,
    "timestep_mean": 423.84375,
    "epsilon_norm": 8.410776138305664,
    "prediction_norm": 6.685318470001221,
    "timestamp": "2025-07-14T12:35:16.931059"
  },
  {
    "epoch": 2866,
    "batch_idx": 160,
    "loss": 0.15951040387153625,
    "learning_rate": 0.0003955320547987949,
    "timestep_mean": 541.3125,
    "epsilon_norm": 8.268913269042969,
    "prediction_norm": 7.923955917358398,
    "timestamp": "2025-07-14T12:35:16.936125"
  },
  {
    "epoch": 2866,
    "batch_idx": 192,
    "loss": 0.15471966564655304,
    "learning_rate": 0.0003955320547987949,
    "timestep_mean": 496.625,
    "epsilon_norm": 2.9957163333892822,
    "prediction_norm": 2.6584041118621826,
    "timestamp": "2025-07-14T12:35:16.941286"
  },
  {
    "epoch": 2867,
    "batch_idx": 0,
    "loss": 0.1753806620836258,
    "learning_rate": 0.00039630024727251175,
    "timestep_mean": 535.71875,
    "epsilon_norm": 8.026877403259277,
    "prediction_norm": 7.444442272186279,
    "timestamp": "2025-07-14T12:35:16.946514"
  },
  {
    "epoch": 2867,
    "batch_idx": 32,
    "loss": 0.5241762399673462,
    "learning_rate": 0.00039630024727251175,
    "timestep_mean": 488.09375,
    "epsilon_norm": 8.327548027038574,
    "prediction_norm": 6.51348352432251,
    "timestamp": "2025-07-14T12:35:16.951486"
  },
  {
    "epoch": 2867,
    "batch_idx": 64,
    "loss": 0.31039416790008545,
    "learning_rate": 0.00039630024727251175,
    "timestep_mean": 431.8125,
    "epsilon_norm": 7.67746114730835,
    "prediction_norm": 6.853114604949951,
    "timestamp": "2025-07-14T12:35:16.956592"
  },
  {
    "epoch": 2867,
    "batch_idx": 96,
    "loss": 0.43649619817733765,
    "learning_rate": 0.00039630024727251175,
    "timestep_mean": 501.46875,
    "epsilon_norm": 7.041027069091797,
    "prediction_norm": 5.3506059646606445,
    "timestamp": "2025-07-14T12:35:16.961758"
  },
  {
    "epoch": 2867,
    "batch_idx": 128,
    "loss": 0.31890085339546204,
    "learning_rate": 0.00039630024727251175,
    "timestep_mean": 511.6875,
    "epsilon_norm": 8.049062728881836,
    "prediction_norm": 7.12805700302124,
    "timestamp": "2025-07-14T12:35:16.966815"
  },
  {
    "epoch": 2867,
    "batch_idx": 160,
    "loss": 0.19180656969547272,
    "learning_rate": 0.00039630024727251175,
    "timestep_mean": 503.40625,
    "epsilon_norm": 7.301025390625,
    "prediction_norm": 6.30169153213501,
    "timestamp": "2025-07-14T12:35:16.972075"
  },
  {
    "epoch": 2867,
    "batch_idx": 192,
    "loss": 0.11262357234954834,
    "learning_rate": 0.00039630024727251175,
    "timestep_mean": 526.375,
    "epsilon_norm": 2.7495105266571045,
    "prediction_norm": 2.761610269546509,
    "timestamp": "2025-07-14T12:35:16.977264"
  },
  {
    "epoch": 2868,
    "batch_idx": 0,
    "loss": 0.1628769338130951,
    "learning_rate": 0.00039706869561506,
    "timestep_mean": 441.34375,
    "epsilon_norm": 7.188203811645508,
    "prediction_norm": 6.360656261444092,
    "timestamp": "2025-07-14T12:35:16.982301"
  },
  {
    "epoch": 2868,
    "batch_idx": 32,
    "loss": 0.20994427800178528,
    "learning_rate": 0.00039706869561506,
    "timestep_mean": 459.3125,
    "epsilon_norm": 8.555848121643066,
    "prediction_norm": 7.1730217933654785,
    "timestamp": "2025-07-14T12:35:16.987475"
  },
  {
    "epoch": 2868,
    "batch_idx": 64,
    "loss": 0.190446138381958,
    "learning_rate": 0.00039706869561506,
    "timestep_mean": 553.53125,
    "epsilon_norm": 8.007416725158691,
    "prediction_norm": 7.577669143676758,
    "timestamp": "2025-07-14T12:35:16.992771"
  },
  {
    "epoch": 2868,
    "batch_idx": 96,
    "loss": 0.1404159516096115,
    "learning_rate": 0.00039706869561506,
    "timestep_mean": 558.65625,
    "epsilon_norm": 9.430197715759277,
    "prediction_norm": 9.280523300170898,
    "timestamp": "2025-07-14T12:35:16.997841"
  },
  {
    "epoch": 2868,
    "batch_idx": 128,
    "loss": 0.22804605960845947,
    "learning_rate": 0.00039706869561506,
    "timestep_mean": 454.0625,
    "epsilon_norm": 7.2851433753967285,
    "prediction_norm": 6.4523606300354,
    "timestamp": "2025-07-14T12:35:17.003013"
  },
  {
    "epoch": 2868,
    "batch_idx": 160,
    "loss": 0.2956434190273285,
    "learning_rate": 0.00039706869561506,
    "timestep_mean": 493.28125,
    "epsilon_norm": 8.08760929107666,
    "prediction_norm": 6.780187129974365,
    "timestamp": "2025-07-14T12:35:17.008113"
  },
  {
    "epoch": 2868,
    "batch_idx": 192,
    "loss": 0.7182851433753967,
    "learning_rate": 0.00039706869561506,
    "timestep_mean": 463.625,
    "epsilon_norm": 4.068010330200195,
    "prediction_norm": 1.9113253355026245,
    "timestamp": "2025-07-14T12:35:17.013280"
  },
  {
    "epoch": 2869,
    "batch_idx": 0,
    "loss": 0.36437416076660156,
    "learning_rate": 0.0003978373979303702,
    "timestep_mean": 478.5,
    "epsilon_norm": 7.495201587677002,
    "prediction_norm": 6.092273712158203,
    "timestamp": "2025-07-14T12:35:17.018328"
  },
  {
    "epoch": 2869,
    "batch_idx": 32,
    "loss": 0.22874335944652557,
    "learning_rate": 0.0003978373979303702,
    "timestep_mean": 555.15625,
    "epsilon_norm": 8.52436351776123,
    "prediction_norm": 7.64765739440918,
    "timestamp": "2025-07-14T12:35:17.023385"
  },
  {
    "epoch": 2869,
    "batch_idx": 64,
    "loss": 0.21373923122882843,
    "learning_rate": 0.0003978373979303702,
    "timestep_mean": 582.96875,
    "epsilon_norm": 7.0018134117126465,
    "prediction_norm": 6.269330978393555,
    "timestamp": "2025-07-14T12:35:17.028573"
  },
  {
    "epoch": 2869,
    "batch_idx": 96,
    "loss": 0.22115638852119446,
    "learning_rate": 0.0003978373979303702,
    "timestep_mean": 479.25,
    "epsilon_norm": 7.44346809387207,
    "prediction_norm": 6.317440509796143,
    "timestamp": "2025-07-14T12:35:17.033931"
  },
  {
    "epoch": 2869,
    "batch_idx": 128,
    "loss": 0.30696770548820496,
    "learning_rate": 0.0003978373979303702,
    "timestep_mean": 499.875,
    "epsilon_norm": 8.96645450592041,
    "prediction_norm": 7.6717448234558105,
    "timestamp": "2025-07-14T12:35:17.039070"
  },
  {
    "epoch": 2869,
    "batch_idx": 160,
    "loss": 0.5104000568389893,
    "learning_rate": 0.0003978373979303702,
    "timestep_mean": 493.84375,
    "epsilon_norm": 8.972200393676758,
    "prediction_norm": 6.6121134757995605,
    "timestamp": "2025-07-14T12:35:17.044243"
  },
  {
    "epoch": 2869,
    "batch_idx": 192,
    "loss": 0.5380179286003113,
    "learning_rate": 0.0003978373979303702,
    "timestep_mean": 463.5,
    "epsilon_norm": 4.742398738861084,
    "prediction_norm": 3.5547266006469727,
    "timestamp": "2025-07-14T12:35:17.049356"
  },
  {
    "epoch": 2870,
    "batch_idx": 0,
    "loss": 0.22633692622184753,
    "learning_rate": 0.0003986063523217448,
    "timestep_mean": 427.71875,
    "epsilon_norm": 7.454401969909668,
    "prediction_norm": 6.184422016143799,
    "timestamp": "2025-07-14T12:35:17.054482"
  },
  {
    "epoch": 2870,
    "batch_idx": 32,
    "loss": 0.08590929210186005,
    "learning_rate": 0.0003986063523217448,
    "timestep_mean": 596.6875,
    "epsilon_norm": 7.512203693389893,
    "prediction_norm": 6.9165778160095215,
    "timestamp": "2025-07-14T12:35:17.059653"
  },
  {
    "epoch": 2870,
    "batch_idx": 64,
    "loss": 0.16201233863830566,
    "learning_rate": 0.0003986063523217448,
    "timestep_mean": 491.53125,
    "epsilon_norm": 7.731268405914307,
    "prediction_norm": 6.884397983551025,
    "timestamp": "2025-07-14T12:35:17.064745"
  },
  {
    "epoch": 2870,
    "batch_idx": 96,
    "loss": 0.2771527171134949,
    "learning_rate": 0.0003986063523217448,
    "timestep_mean": 556.53125,
    "epsilon_norm": 8.589244842529297,
    "prediction_norm": 6.902116298675537,
    "timestamp": "2025-07-14T12:35:17.069966"
  },
  {
    "epoch": 2870,
    "batch_idx": 128,
    "loss": 0.760400652885437,
    "learning_rate": 0.0003986063523217448,
    "timestep_mean": 428.5,
    "epsilon_norm": 9.489421844482422,
    "prediction_norm": 7.182444095611572,
    "timestamp": "2025-07-14T12:35:17.075107"
  },
  {
    "epoch": 2870,
    "batch_idx": 160,
    "loss": 0.18237265944480896,
    "learning_rate": 0.0003986063523217448,
    "timestep_mean": 480.59375,
    "epsilon_norm": 7.466843128204346,
    "prediction_norm": 6.608683109283447,
    "timestamp": "2025-07-14T12:35:17.080340"
  },
  {
    "epoch": 2870,
    "batch_idx": 192,
    "loss": 0.3718896806240082,
    "learning_rate": 0.0003986063523217448,
    "timestep_mean": 472.5,
    "epsilon_norm": 4.3738932609558105,
    "prediction_norm": 3.2379696369171143,
    "timestamp": "2025-07-14T12:35:17.085469"
  },
  {
    "epoch": 2871,
    "batch_idx": 0,
    "loss": 0.22827348113059998,
    "learning_rate": 0.0003993755568918658,
    "timestep_mean": 492.5,
    "epsilon_norm": 7.306509971618652,
    "prediction_norm": 6.578526496887207,
    "timestamp": "2025-07-14T12:35:17.090462"
  },
  {
    "epoch": 2871,
    "batch_idx": 32,
    "loss": 0.20491814613342285,
    "learning_rate": 0.0003993755568918658,
    "timestep_mean": 533.625,
    "epsilon_norm": 6.914144515991211,
    "prediction_norm": 6.310729026794434,
    "timestamp": "2025-07-14T12:35:17.095478"
  },
  {
    "epoch": 2871,
    "batch_idx": 64,
    "loss": 0.18176507949829102,
    "learning_rate": 0.0003993755568918658,
    "timestep_mean": 528.34375,
    "epsilon_norm": 7.275538921356201,
    "prediction_norm": 6.111683368682861,
    "timestamp": "2025-07-14T12:35:17.100609"
  },
  {
    "epoch": 2871,
    "batch_idx": 96,
    "loss": 0.32658979296684265,
    "learning_rate": 0.0003993755568918658,
    "timestep_mean": 417.84375,
    "epsilon_norm": 7.5822272300720215,
    "prediction_norm": 6.078466892242432,
    "timestamp": "2025-07-14T12:35:17.105833"
  },
  {
    "epoch": 2871,
    "batch_idx": 128,
    "loss": 0.159529447555542,
    "learning_rate": 0.0003993755568918658,
    "timestep_mean": 495.75,
    "epsilon_norm": 7.666587829589844,
    "prediction_norm": 6.31292200088501,
    "timestamp": "2025-07-14T12:35:17.111123"
  },
  {
    "epoch": 2871,
    "batch_idx": 160,
    "loss": 0.4650534987449646,
    "learning_rate": 0.0003993755568918658,
    "timestep_mean": 537.53125,
    "epsilon_norm": 8.382569313049316,
    "prediction_norm": 7.0023651123046875,
    "timestamp": "2025-07-14T12:35:17.116279"
  },
  {
    "epoch": 2871,
    "batch_idx": 192,
    "loss": 0.11987666040658951,
    "learning_rate": 0.0003993755568918658,
    "timestep_mean": 686.125,
    "epsilon_norm": 5.922033309936523,
    "prediction_norm": 5.375039100646973,
    "timestamp": "2025-07-14T12:35:17.121400"
  },
  {
    "epoch": 2872,
    "batch_idx": 0,
    "loss": 0.3369905948638916,
    "learning_rate": 0.00040014500974279736,
    "timestep_mean": 441.9375,
    "epsilon_norm": 8.600007057189941,
    "prediction_norm": 7.344152450561523,
    "timestamp": "2025-07-14T12:35:17.126556"
  },
  {
    "epoch": 2872,
    "batch_idx": 32,
    "loss": 0.20894008874893188,
    "learning_rate": 0.00040014500974279736,
    "timestep_mean": 454.0,
    "epsilon_norm": 8.278989791870117,
    "prediction_norm": 7.1935882568359375,
    "timestamp": "2025-07-14T12:35:17.131779"
  },
  {
    "epoch": 2872,
    "batch_idx": 64,
    "loss": 0.3383713364601135,
    "learning_rate": 0.00040014500974279736,
    "timestep_mean": 488.28125,
    "epsilon_norm": 8.374361038208008,
    "prediction_norm": 7.639798641204834,
    "timestamp": "2025-07-14T12:35:17.136695"
  },
  {
    "epoch": 2872,
    "batch_idx": 96,
    "loss": 0.3204832971096039,
    "learning_rate": 0.00040014500974279736,
    "timestep_mean": 558.0625,
    "epsilon_norm": 7.587261199951172,
    "prediction_norm": 6.066003322601318,
    "timestamp": "2025-07-14T12:35:17.141669"
  },
  {
    "epoch": 2872,
    "batch_idx": 128,
    "loss": 0.19391480088233948,
    "learning_rate": 0.00040014500974279736,
    "timestep_mean": 480.71875,
    "epsilon_norm": 6.481595993041992,
    "prediction_norm": 5.914731979370117,
    "timestamp": "2025-07-14T12:35:17.146609"
  },
  {
    "epoch": 2872,
    "batch_idx": 160,
    "loss": 0.20795556902885437,
    "learning_rate": 0.00040014500974279736,
    "timestep_mean": 558.1875,
    "epsilon_norm": 7.371255397796631,
    "prediction_norm": 6.713191986083984,
    "timestamp": "2025-07-14T12:35:17.151510"
  },
  {
    "epoch": 2872,
    "batch_idx": 192,
    "loss": 0.23567628860473633,
    "learning_rate": 0.00040014500974279736,
    "timestep_mean": 440.5,
    "epsilon_norm": 4.061174392700195,
    "prediction_norm": 3.3051936626434326,
    "timestamp": "2025-07-14T12:35:17.156640"
  },
  {
    "epoch": 2873,
    "batch_idx": 0,
    "loss": 0.43920570611953735,
    "learning_rate": 0.0004009147089759912,
    "timestep_mean": 464.625,
    "epsilon_norm": 8.726938247680664,
    "prediction_norm": 6.99160099029541,
    "timestamp": "2025-07-14T12:35:17.161737"
  },
  {
    "epoch": 2873,
    "batch_idx": 32,
    "loss": 0.25462791323661804,
    "learning_rate": 0.0004009147089759912,
    "timestep_mean": 529.75,
    "epsilon_norm": 8.884260177612305,
    "prediction_norm": 8.135729789733887,
    "timestamp": "2025-07-14T12:35:17.166881"
  },
  {
    "epoch": 2873,
    "batch_idx": 64,
    "loss": 0.2932382822036743,
    "learning_rate": 0.0004009147089759912,
    "timestep_mean": 483.75,
    "epsilon_norm": 8.466429710388184,
    "prediction_norm": 7.617396354675293,
    "timestamp": "2025-07-14T12:35:17.171964"
  },
  {
    "epoch": 2873,
    "batch_idx": 96,
    "loss": 0.25090986490249634,
    "learning_rate": 0.0004009147089759912,
    "timestep_mean": 475.25,
    "epsilon_norm": 7.196826934814453,
    "prediction_norm": 6.166693210601807,
    "timestamp": "2025-07-14T12:35:17.176978"
  },
  {
    "epoch": 2873,
    "batch_idx": 128,
    "loss": 0.3062501847743988,
    "learning_rate": 0.0004009147089759912,
    "timestep_mean": 391.21875,
    "epsilon_norm": 7.919154644012451,
    "prediction_norm": 7.122578144073486,
    "timestamp": "2025-07-14T12:35:17.182025"
  },
  {
    "epoch": 2873,
    "batch_idx": 160,
    "loss": 0.3646424412727356,
    "learning_rate": 0.0004009147089759912,
    "timestep_mean": 438.28125,
    "epsilon_norm": 7.557985782623291,
    "prediction_norm": 6.329957485198975,
    "timestamp": "2025-07-14T12:35:17.187169"
  },
  {
    "epoch": 2873,
    "batch_idx": 192,
    "loss": 0.1031375527381897,
    "learning_rate": 0.0004009147089759912,
    "timestep_mean": 554.375,
    "epsilon_norm": 3.7785074710845947,
    "prediction_norm": 3.652456760406494,
    "timestamp": "2025-07-14T12:35:17.192300"
  },
  {
    "epoch": 2874,
    "batch_idx": 0,
    "loss": 0.1353338360786438,
    "learning_rate": 0.0004016846526922908,
    "timestep_mean": 585.71875,
    "epsilon_norm": 7.635514259338379,
    "prediction_norm": 7.394085884094238,
    "timestamp": "2025-07-14T12:35:17.197352"
  },
  {
    "epoch": 2874,
    "batch_idx": 32,
    "loss": 0.656237006187439,
    "learning_rate": 0.0004016846526922908,
    "timestep_mean": 500.65625,
    "epsilon_norm": 8.068503379821777,
    "prediction_norm": 6.182132244110107,
    "timestamp": "2025-07-14T12:35:17.202444"
  },
  {
    "epoch": 2874,
    "batch_idx": 64,
    "loss": 0.21347805857658386,
    "learning_rate": 0.0004016846526922908,
    "timestep_mean": 516.8125,
    "epsilon_norm": 6.6035871505737305,
    "prediction_norm": 5.638702392578125,
    "timestamp": "2025-07-14T12:35:17.207551"
  },
  {
    "epoch": 2874,
    "batch_idx": 96,
    "loss": 0.25837084650993347,
    "learning_rate": 0.0004016846526922908,
    "timestep_mean": 461.65625,
    "epsilon_norm": 7.899489879608154,
    "prediction_norm": 7.212589740753174,
    "timestamp": "2025-07-14T12:35:17.212748"
  },
  {
    "epoch": 2874,
    "batch_idx": 128,
    "loss": 0.38235652446746826,
    "learning_rate": 0.0004016846526922908,
    "timestep_mean": 467.46875,
    "epsilon_norm": 8.355631828308105,
    "prediction_norm": 5.973780155181885,
    "timestamp": "2025-07-14T12:35:17.217926"
  },
  {
    "epoch": 2874,
    "batch_idx": 160,
    "loss": 0.21982203423976898,
    "learning_rate": 0.0004016846526922908,
    "timestep_mean": 556.9375,
    "epsilon_norm": 7.511907577514648,
    "prediction_norm": 6.616151332855225,
    "timestamp": "2025-07-14T12:35:17.222897"
  },
  {
    "epoch": 2874,
    "batch_idx": 192,
    "loss": 0.17115527391433716,
    "learning_rate": 0.0004016846526922908,
    "timestep_mean": 463.5,
    "epsilon_norm": 3.191321849822998,
    "prediction_norm": 2.693150758743286,
    "timestamp": "2025-07-14T12:35:17.227779"
  },
  {
    "epoch": 2875,
    "batch_idx": 0,
    "loss": 0.17019516229629517,
    "learning_rate": 0.0004024548389919367,
    "timestep_mean": 538.3125,
    "epsilon_norm": 8.077062606811523,
    "prediction_norm": 7.290188789367676,
    "timestamp": "2025-07-14T12:35:17.232920"
  },
  {
    "epoch": 2875,
    "batch_idx": 32,
    "loss": 0.3375042676925659,
    "learning_rate": 0.0004024548389919367,
    "timestep_mean": 516.875,
    "epsilon_norm": 9.40395450592041,
    "prediction_norm": 7.496068000793457,
    "timestamp": "2025-07-14T12:35:17.238106"
  },
  {
    "epoch": 2875,
    "batch_idx": 64,
    "loss": 0.3402116894721985,
    "learning_rate": 0.0004024548389919367,
    "timestep_mean": 475.1875,
    "epsilon_norm": 8.034979820251465,
    "prediction_norm": 6.416243553161621,
    "timestamp": "2025-07-14T12:35:17.243135"
  },
  {
    "epoch": 2875,
    "batch_idx": 96,
    "loss": 0.2741433382034302,
    "learning_rate": 0.0004024548389919367,
    "timestep_mean": 433.0,
    "epsilon_norm": 7.293503284454346,
    "prediction_norm": 5.922353267669678,
    "timestamp": "2025-07-14T12:35:17.248289"
  },
  {
    "epoch": 2875,
    "batch_idx": 128,
    "loss": 0.24091297388076782,
    "learning_rate": 0.0004024548389919367,
    "timestep_mean": 451.40625,
    "epsilon_norm": 7.312019348144531,
    "prediction_norm": 6.5648393630981445,
    "timestamp": "2025-07-14T12:35:17.253444"
  },
  {
    "epoch": 2875,
    "batch_idx": 160,
    "loss": 0.3335701823234558,
    "learning_rate": 0.0004024548389919367,
    "timestep_mean": 496.71875,
    "epsilon_norm": 8.632885932922363,
    "prediction_norm": 6.988450050354004,
    "timestamp": "2025-07-14T12:35:17.258699"
  },
  {
    "epoch": 2875,
    "batch_idx": 192,
    "loss": 0.08108673244714737,
    "learning_rate": 0.0004024548389919367,
    "timestep_mean": 601.375,
    "epsilon_norm": 3.4083073139190674,
    "prediction_norm": 3.047358512878418,
    "timestamp": "2025-07-14T12:35:17.263872"
  },
  {
    "epoch": 2876,
    "batch_idx": 0,
    "loss": 0.2674439251422882,
    "learning_rate": 0.00040322526597457115,
    "timestep_mean": 509.71875,
    "epsilon_norm": 8.337529182434082,
    "prediction_norm": 7.1914567947387695,
    "timestamp": "2025-07-14T12:35:17.268904"
  },
  {
    "epoch": 2876,
    "batch_idx": 32,
    "loss": 0.1884530484676361,
    "learning_rate": 0.00040322526597457115,
    "timestep_mean": 448.5625,
    "epsilon_norm": 7.36118745803833,
    "prediction_norm": 5.797445774078369,
    "timestamp": "2025-07-14T12:35:17.274014"
  },
  {
    "epoch": 2876,
    "batch_idx": 64,
    "loss": 0.36432015895843506,
    "learning_rate": 0.00040322526597457115,
    "timestep_mean": 472.125,
    "epsilon_norm": 9.024093627929688,
    "prediction_norm": 7.396352291107178,
    "timestamp": "2025-07-14T12:35:17.279237"
  },
  {
    "epoch": 2876,
    "batch_idx": 96,
    "loss": 0.32860177755355835,
    "learning_rate": 0.00040322526597457115,
    "timestep_mean": 474.53125,
    "epsilon_norm": 8.470349311828613,
    "prediction_norm": 6.846231460571289,
    "timestamp": "2025-07-14T12:35:17.284422"
  },
  {
    "epoch": 2876,
    "batch_idx": 128,
    "loss": 0.4082252085208893,
    "learning_rate": 0.00040322526597457115,
    "timestep_mean": 570.125,
    "epsilon_norm": 9.55308723449707,
    "prediction_norm": 7.641573905944824,
    "timestamp": "2025-07-14T12:35:17.289362"
  },
  {
    "epoch": 2876,
    "batch_idx": 160,
    "loss": 0.423237144947052,
    "learning_rate": 0.00040322526597457115,
    "timestep_mean": 372.9375,
    "epsilon_norm": 7.820786952972412,
    "prediction_norm": 6.307392120361328,
    "timestamp": "2025-07-14T12:35:17.294432"
  },
  {
    "epoch": 2876,
    "batch_idx": 192,
    "loss": 0.5844125747680664,
    "learning_rate": 0.00040322526597457115,
    "timestep_mean": 476.5,
    "epsilon_norm": 4.8054938316345215,
    "prediction_norm": 3.677456855773926,
    "timestamp": "2025-07-14T12:35:17.299262"
  },
  {
    "epoch": 2877,
    "batch_idx": 0,
    "loss": 0.3713071346282959,
    "learning_rate": 0.00040399593173924076,
    "timestep_mean": 454.625,
    "epsilon_norm": 7.898194789886475,
    "prediction_norm": 6.036605358123779,
    "timestamp": "2025-07-14T12:35:17.304204"
  },
  {
    "epoch": 2877,
    "batch_idx": 32,
    "loss": 0.4286918044090271,
    "learning_rate": 0.00040399593173924076,
    "timestep_mean": 385.5,
    "epsilon_norm": 7.969851016998291,
    "prediction_norm": 5.066624164581299,
    "timestamp": "2025-07-14T12:35:17.309396"
  },
  {
    "epoch": 2877,
    "batch_idx": 64,
    "loss": 0.18222980201244354,
    "learning_rate": 0.00040399593173924076,
    "timestep_mean": 515.46875,
    "epsilon_norm": 7.334891319274902,
    "prediction_norm": 6.42725133895874,
    "timestamp": "2025-07-14T12:35:17.314443"
  },
  {
    "epoch": 2877,
    "batch_idx": 96,
    "loss": 0.40661731362342834,
    "learning_rate": 0.00040399593173924076,
    "timestep_mean": 555.09375,
    "epsilon_norm": 7.0551323890686035,
    "prediction_norm": 5.415677547454834,
    "timestamp": "2025-07-14T12:35:17.319595"
  },
  {
    "epoch": 2877,
    "batch_idx": 128,
    "loss": 0.18413326144218445,
    "learning_rate": 0.00040399593173924076,
    "timestep_mean": 544.09375,
    "epsilon_norm": 7.590105056762695,
    "prediction_norm": 6.569890022277832,
    "timestamp": "2025-07-14T12:35:17.324707"
  },
  {
    "epoch": 2877,
    "batch_idx": 160,
    "loss": 0.3912166357040405,
    "learning_rate": 0.00040399593173924076,
    "timestep_mean": 567.09375,
    "epsilon_norm": 8.656451225280762,
    "prediction_norm": 7.506255626678467,
    "timestamp": "2025-07-14T12:35:17.329907"
  },
  {
    "epoch": 2877,
    "batch_idx": 192,
    "loss": 0.2136264592409134,
    "learning_rate": 0.00040399593173924076,
    "timestep_mean": 392.25,
    "epsilon_norm": 3.5271546840667725,
    "prediction_norm": 3.727993965148926,
    "timestamp": "2025-07-14T12:35:17.334882"
  },
  {
    "epoch": 2878,
    "batch_idx": 0,
    "loss": 0.30933719873428345,
    "learning_rate": 0.00040476683438440584,
    "timestep_mean": 551.5625,
    "epsilon_norm": 8.078731536865234,
    "prediction_norm": 7.020648956298828,
    "timestamp": "2025-07-14T12:35:17.339980"
  },
  {
    "epoch": 2878,
    "batch_idx": 32,
    "loss": 0.11192970722913742,
    "learning_rate": 0.00040476683438440584,
    "timestep_mean": 554.21875,
    "epsilon_norm": 6.2339935302734375,
    "prediction_norm": 6.240191459655762,
    "timestamp": "2025-07-14T12:35:17.345138"
  },
  {
    "epoch": 2878,
    "batch_idx": 64,
    "loss": 0.0816553384065628,
    "learning_rate": 0.00040476683438440584,
    "timestep_mean": 600.71875,
    "epsilon_norm": 7.3588128089904785,
    "prediction_norm": 7.183224201202393,
    "timestamp": "2025-07-14T12:35:17.350156"
  },
  {
    "epoch": 2878,
    "batch_idx": 96,
    "loss": 0.37517014145851135,
    "learning_rate": 0.00040476683438440584,
    "timestep_mean": 556.0,
    "epsilon_norm": 8.222159385681152,
    "prediction_norm": 6.624713897705078,
    "timestamp": "2025-07-14T12:35:17.355314"
  },
  {
    "epoch": 2878,
    "batch_idx": 128,
    "loss": 0.24391812086105347,
    "learning_rate": 0.00040476683438440584,
    "timestep_mean": 501.1875,
    "epsilon_norm": 8.326891899108887,
    "prediction_norm": 7.548203945159912,
    "timestamp": "2025-07-14T12:35:17.360468"
  },
  {
    "epoch": 2878,
    "batch_idx": 160,
    "loss": 0.3795700669288635,
    "learning_rate": 0.00040476683438440584,
    "timestep_mean": 510.375,
    "epsilon_norm": 7.215576648712158,
    "prediction_norm": 6.3173699378967285,
    "timestamp": "2025-07-14T12:35:17.365625"
  },
  {
    "epoch": 2878,
    "batch_idx": 192,
    "loss": 0.22136817872524261,
    "learning_rate": 0.00040476683438440584,
    "timestep_mean": 358.625,
    "epsilon_norm": 2.880297899246216,
    "prediction_norm": 2.3031368255615234,
    "timestamp": "2025-07-14T12:35:17.370830"
  },
  {
    "epoch": 2879,
    "batch_idx": 0,
    "loss": 0.19651319086551666,
    "learning_rate": 0.0004055379720079406,
    "timestep_mean": 471.375,
    "epsilon_norm": 7.530556678771973,
    "prediction_norm": 6.583160400390625,
    "timestamp": "2025-07-14T12:35:17.376047"
  },
  {
    "epoch": 2879,
    "batch_idx": 32,
    "loss": 0.3259055018424988,
    "learning_rate": 0.0004055379720079406,
    "timestep_mean": 515.28125,
    "epsilon_norm": 7.827208042144775,
    "prediction_norm": 6.218329906463623,
    "timestamp": "2025-07-14T12:35:17.381009"
  },
  {
    "epoch": 2879,
    "batch_idx": 64,
    "loss": 0.19291622936725616,
    "learning_rate": 0.0004055379720079406,
    "timestep_mean": 509.5625,
    "epsilon_norm": 8.660481452941895,
    "prediction_norm": 7.26046085357666,
    "timestamp": "2025-07-14T12:35:17.386014"
  },
  {
    "epoch": 2879,
    "batch_idx": 96,
    "loss": 0.5632153749465942,
    "learning_rate": 0.0004055379720079406,
    "timestep_mean": 397.40625,
    "epsilon_norm": 7.874530792236328,
    "prediction_norm": 5.9098801612854,
    "timestamp": "2025-07-14T12:35:17.391212"
  },
  {
    "epoch": 2879,
    "batch_idx": 128,
    "loss": 0.3426333963871002,
    "learning_rate": 0.0004055379720079406,
    "timestep_mean": 506.78125,
    "epsilon_norm": 8.226804733276367,
    "prediction_norm": 6.816798686981201,
    "timestamp": "2025-07-14T12:35:17.396267"
  },
  {
    "epoch": 2879,
    "batch_idx": 160,
    "loss": 0.20143485069274902,
    "learning_rate": 0.0004055379720079406,
    "timestep_mean": 524.375,
    "epsilon_norm": 7.499827861785889,
    "prediction_norm": 6.607202529907227,
    "timestamp": "2025-07-14T12:35:17.401388"
  },
  {
    "epoch": 2879,
    "batch_idx": 192,
    "loss": 0.7749242782592773,
    "learning_rate": 0.0004055379720079406,
    "timestep_mean": 405.375,
    "epsilon_norm": 4.975989818572998,
    "prediction_norm": 2.9611146450042725,
    "timestamp": "2025-07-14T12:35:17.406297"
  },
  {
    "epoch": 2880,
    "batch_idx": 0,
    "loss": 0.27640578150749207,
    "learning_rate": 0.00040630934270713875,
    "timestep_mean": 474.0,
    "epsilon_norm": 7.944700241088867,
    "prediction_norm": 7.012768745422363,
    "timestamp": "2025-07-14T12:35:17.411146"
  },
  {
    "epoch": 2880,
    "batch_idx": 32,
    "loss": 0.2712027430534363,
    "learning_rate": 0.00040630934270713875,
    "timestep_mean": 400.625,
    "epsilon_norm": 6.337338447570801,
    "prediction_norm": 4.952132701873779,
    "timestamp": "2025-07-14T12:35:17.416304"
  },
  {
    "epoch": 2880,
    "batch_idx": 64,
    "loss": 0.2012862265110016,
    "learning_rate": 0.00040630934270713875,
    "timestep_mean": 464.0625,
    "epsilon_norm": 8.040082931518555,
    "prediction_norm": 6.2152204513549805,
    "timestamp": "2025-07-14T12:35:17.421483"
  },
  {
    "epoch": 2880,
    "batch_idx": 96,
    "loss": 0.33080047369003296,
    "learning_rate": 0.00040630934270713875,
    "timestep_mean": 380.09375,
    "epsilon_norm": 8.631073951721191,
    "prediction_norm": 6.9945831298828125,
    "timestamp": "2025-07-14T12:35:17.426636"
  },
  {
    "epoch": 2880,
    "batch_idx": 128,
    "loss": 0.28737226128578186,
    "learning_rate": 0.00040630934270713875,
    "timestep_mean": 505.6875,
    "epsilon_norm": 7.399896144866943,
    "prediction_norm": 6.126368045806885,
    "timestamp": "2025-07-14T12:35:17.431645"
  },
  {
    "epoch": 2880,
    "batch_idx": 160,
    "loss": 0.27531617879867554,
    "learning_rate": 0.00040630934270713875,
    "timestep_mean": 537.9375,
    "epsilon_norm": 8.45030403137207,
    "prediction_norm": 6.941599369049072,
    "timestamp": "2025-07-14T12:35:17.436516"
  },
  {
    "epoch": 2880,
    "batch_idx": 192,
    "loss": 0.18866650760173798,
    "learning_rate": 0.00040630934270713875,
    "timestep_mean": 544.5,
    "epsilon_norm": 3.283210039138794,
    "prediction_norm": 2.2817556858062744,
    "timestamp": "2025-07-14T12:35:17.441369"
  },
  {
    "epoch": 2881,
    "batch_idx": 0,
    "loss": 0.30535104870796204,
    "learning_rate": 0.0004070809445787203,
    "timestep_mean": 534.53125,
    "epsilon_norm": 8.210487365722656,
    "prediction_norm": 6.952343940734863,
    "timestamp": "2025-07-14T12:35:17.446226"
  },
  {
    "epoch": 2881,
    "batch_idx": 32,
    "loss": 0.3604438900947571,
    "learning_rate": 0.0004070809445787203,
    "timestep_mean": 509.46875,
    "epsilon_norm": 8.339954376220703,
    "prediction_norm": 6.280716896057129,
    "timestamp": "2025-07-14T12:35:17.451162"
  },
  {
    "epoch": 2881,
    "batch_idx": 64,
    "loss": 0.2410988211631775,
    "learning_rate": 0.0004070809445787203,
    "timestep_mean": 497.9375,
    "epsilon_norm": 7.673226356506348,
    "prediction_norm": 6.4408860206604,
    "timestamp": "2025-07-14T12:35:17.456065"
  },
  {
    "epoch": 2881,
    "batch_idx": 96,
    "loss": 0.3201902508735657,
    "learning_rate": 0.0004070809445787203,
    "timestep_mean": 500.59375,
    "epsilon_norm": 7.300594329833984,
    "prediction_norm": 5.977999210357666,
    "timestamp": "2025-07-14T12:35:17.461197"
  },
  {
    "epoch": 2881,
    "batch_idx": 128,
    "loss": 0.247637078166008,
    "learning_rate": 0.0004070809445787203,
    "timestep_mean": 503.59375,
    "epsilon_norm": 8.073037147521973,
    "prediction_norm": 7.5560221672058105,
    "timestamp": "2025-07-14T12:35:17.466097"
  },
  {
    "epoch": 2881,
    "batch_idx": 160,
    "loss": 0.31155967712402344,
    "learning_rate": 0.0004070809445787203,
    "timestep_mean": 498.4375,
    "epsilon_norm": 8.705364227294922,
    "prediction_norm": 7.696987152099609,
    "timestamp": "2025-07-14T12:35:17.470993"
  },
  {
    "epoch": 2881,
    "batch_idx": 192,
    "loss": 0.08154653012752533,
    "learning_rate": 0.0004070809445787203,
    "timestep_mean": 483.0,
    "epsilon_norm": 3.0992510318756104,
    "prediction_norm": 2.7422914505004883,
    "timestamp": "2025-07-14T12:35:17.476002"
  },
  {
    "epoch": 2882,
    "batch_idx": 0,
    "loss": 0.17398995161056519,
    "learning_rate": 0.00040785277571883476,
    "timestep_mean": 550.75,
    "epsilon_norm": 8.008408546447754,
    "prediction_norm": 7.027590751647949,
    "timestamp": "2025-07-14T12:35:17.480865"
  },
  {
    "epoch": 2882,
    "batch_idx": 32,
    "loss": 0.33831149339675903,
    "learning_rate": 0.00040785277571883476,
    "timestep_mean": 508.0625,
    "epsilon_norm": 7.12093448638916,
    "prediction_norm": 5.622997283935547,
    "timestamp": "2025-07-14T12:35:17.485790"
  },
  {
    "epoch": 2882,
    "batch_idx": 64,
    "loss": 0.3575175702571869,
    "learning_rate": 0.00040785277571883476,
    "timestep_mean": 434.375,
    "epsilon_norm": 8.412751197814941,
    "prediction_norm": 6.688908576965332,
    "timestamp": "2025-07-14T12:35:17.490916"
  },
  {
    "epoch": 2882,
    "batch_idx": 96,
    "loss": 0.1408674120903015,
    "learning_rate": 0.00040785277571883476,
    "timestep_mean": 521.25,
    "epsilon_norm": 7.2882914543151855,
    "prediction_norm": 6.8614420890808105,
    "timestamp": "2025-07-14T12:35:17.496043"
  },
  {
    "epoch": 2882,
    "batch_idx": 128,
    "loss": 0.4371322691440582,
    "learning_rate": 0.00040785277571883476,
    "timestep_mean": 434.125,
    "epsilon_norm": 8.416224479675293,
    "prediction_norm": 6.775747776031494,
    "timestamp": "2025-07-14T12:35:17.501138"
  },
  {
    "epoch": 2882,
    "batch_idx": 160,
    "loss": 0.19453859329223633,
    "learning_rate": 0.00040785277571883476,
    "timestep_mean": 522.84375,
    "epsilon_norm": 7.884926795959473,
    "prediction_norm": 6.971549034118652,
    "timestamp": "2025-07-14T12:35:17.506195"
  },
  {
    "epoch": 2882,
    "batch_idx": 192,
    "loss": 0.5732965469360352,
    "learning_rate": 0.00040785277571883476,
    "timestep_mean": 467.25,
    "epsilon_norm": 5.059983253479004,
    "prediction_norm": 3.859576940536499,
    "timestamp": "2025-07-14T12:35:17.511310"
  },
  {
    "epoch": 2883,
    "batch_idx": 0,
    "loss": 0.15041834115982056,
    "learning_rate": 0.0004086248342230641,
    "timestep_mean": 488.6875,
    "epsilon_norm": 8.429364204406738,
    "prediction_norm": 7.10117769241333,
    "timestamp": "2025-07-14T12:35:17.516479"
  },
  {
    "epoch": 2883,
    "batch_idx": 32,
    "loss": 0.18807925283908844,
    "learning_rate": 0.0004086248342230641,
    "timestep_mean": 507.9375,
    "epsilon_norm": 6.761045455932617,
    "prediction_norm": 6.273315906524658,
    "timestamp": "2025-07-14T12:35:17.521628"
  },
  {
    "epoch": 2883,
    "batch_idx": 64,
    "loss": 0.35218337178230286,
    "learning_rate": 0.0004086248342230641,
    "timestep_mean": 526.84375,
    "epsilon_norm": 9.668487548828125,
    "prediction_norm": 8.066509246826172,
    "timestamp": "2025-07-14T12:35:17.526706"
  },
  {
    "epoch": 2883,
    "batch_idx": 96,
    "loss": 0.24741466343402863,
    "learning_rate": 0.0004086248342230641,
    "timestep_mean": 516.96875,
    "epsilon_norm": 7.054085731506348,
    "prediction_norm": 5.777506351470947,
    "timestamp": "2025-07-14T12:35:17.531982"
  },
  {
    "epoch": 2883,
    "batch_idx": 128,
    "loss": 0.21405968070030212,
    "learning_rate": 0.0004086248342230641,
    "timestep_mean": 517.84375,
    "epsilon_norm": 7.827713966369629,
    "prediction_norm": 6.874790668487549,
    "timestamp": "2025-07-14T12:35:17.537148"
  },
  {
    "epoch": 2883,
    "batch_idx": 160,
    "loss": 0.22155292332172394,
    "learning_rate": 0.0004086248342230641,
    "timestep_mean": 512.71875,
    "epsilon_norm": 7.61227560043335,
    "prediction_norm": 6.513280868530273,
    "timestamp": "2025-07-14T12:35:17.542448"
  },
  {
    "epoch": 2883,
    "batch_idx": 192,
    "loss": 0.2607647180557251,
    "learning_rate": 0.0004086248342230641,
    "timestep_mean": 362.75,
    "epsilon_norm": 2.420813798904419,
    "prediction_norm": 2.0194997787475586,
    "timestamp": "2025-07-14T12:35:17.547685"
  },
  {
    "epoch": 2884,
    "batch_idx": 0,
    "loss": 0.18045778572559357,
    "learning_rate": 0.00040939711818643214,
    "timestep_mean": 481.9375,
    "epsilon_norm": 7.297752857208252,
    "prediction_norm": 6.645872116088867,
    "timestamp": "2025-07-14T12:35:17.552888"
  },
  {
    "epoch": 2884,
    "batch_idx": 32,
    "loss": 0.3937235474586487,
    "learning_rate": 0.00040939711818643214,
    "timestep_mean": 451.125,
    "epsilon_norm": 7.648138046264648,
    "prediction_norm": 6.1165032386779785,
    "timestamp": "2025-07-14T12:35:17.558067"
  },
  {
    "epoch": 2884,
    "batch_idx": 64,
    "loss": 0.19123941659927368,
    "learning_rate": 0.00040939711818643214,
    "timestep_mean": 608.28125,
    "epsilon_norm": 8.112797737121582,
    "prediction_norm": 7.658148765563965,
    "timestamp": "2025-07-14T12:35:17.563152"
  },
  {
    "epoch": 2884,
    "batch_idx": 96,
    "loss": 0.1869913935661316,
    "learning_rate": 0.00040939711818643214,
    "timestep_mean": 616.78125,
    "epsilon_norm": 7.923400402069092,
    "prediction_norm": 7.275074005126953,
    "timestamp": "2025-07-14T12:35:17.568233"
  },
  {
    "epoch": 2884,
    "batch_idx": 128,
    "loss": 0.36854326725006104,
    "learning_rate": 0.00040939711818643214,
    "timestep_mean": 492.625,
    "epsilon_norm": 7.083496570587158,
    "prediction_norm": 5.4980692863464355,
    "timestamp": "2025-07-14T12:35:17.573568"
  },
  {
    "epoch": 2884,
    "batch_idx": 160,
    "loss": 0.1877475380897522,
    "learning_rate": 0.00040939711818643214,
    "timestep_mean": 531.53125,
    "epsilon_norm": 7.757334232330322,
    "prediction_norm": 6.865423679351807,
    "timestamp": "2025-07-14T12:35:17.578801"
  },
  {
    "epoch": 2884,
    "batch_idx": 192,
    "loss": 0.06875282526016235,
    "learning_rate": 0.00040939711818643214,
    "timestep_mean": 596.25,
    "epsilon_norm": 3.554292917251587,
    "prediction_norm": 4.140547275543213,
    "timestamp": "2025-07-14T12:35:17.583958"
  },
  {
    "epoch": 2885,
    "batch_idx": 0,
    "loss": 0.1152849942445755,
    "learning_rate": 0.0004101696257034045,
    "timestep_mean": 612.84375,
    "epsilon_norm": 7.049509048461914,
    "prediction_norm": 6.617589473724365,
    "timestamp": "2025-07-14T12:35:17.589121"
  },
  {
    "epoch": 2885,
    "batch_idx": 32,
    "loss": 0.16709628701210022,
    "learning_rate": 0.0004101696257034045,
    "timestep_mean": 543.0625,
    "epsilon_norm": 7.8701677322387695,
    "prediction_norm": 7.759156227111816,
    "timestamp": "2025-07-14T12:35:17.594251"
  },
  {
    "epoch": 2885,
    "batch_idx": 64,
    "loss": 0.2477080225944519,
    "learning_rate": 0.0004101696257034045,
    "timestep_mean": 460.5,
    "epsilon_norm": 7.83389949798584,
    "prediction_norm": 6.220908164978027,
    "timestamp": "2025-07-14T12:35:17.599471"
  },
  {
    "epoch": 2885,
    "batch_idx": 96,
    "loss": 0.31028324365615845,
    "learning_rate": 0.0004101696257034045,
    "timestep_mean": 496.15625,
    "epsilon_norm": 8.157504081726074,
    "prediction_norm": 6.5058698654174805,
    "timestamp": "2025-07-14T12:35:17.604519"
  },
  {
    "epoch": 2885,
    "batch_idx": 128,
    "loss": 0.2581648826599121,
    "learning_rate": 0.0004101696257034045,
    "timestep_mean": 457.46875,
    "epsilon_norm": 8.089604377746582,
    "prediction_norm": 7.52215576171875,
    "timestamp": "2025-07-14T12:35:17.609642"
  },
  {
    "epoch": 2885,
    "batch_idx": 160,
    "loss": 0.26872557401657104,
    "learning_rate": 0.0004101696257034045,
    "timestep_mean": 526.34375,
    "epsilon_norm": 7.335819244384766,
    "prediction_norm": 6.367391586303711,
    "timestamp": "2025-07-14T12:35:17.614746"
  },
  {
    "epoch": 2885,
    "batch_idx": 192,
    "loss": 0.05442255362868309,
    "learning_rate": 0.0004101696257034045,
    "timestep_mean": 813.5,
    "epsilon_norm": 4.661094665527344,
    "prediction_norm": 4.562492847442627,
    "timestamp": "2025-07-14T12:35:17.619827"
  },
  {
    "epoch": 2886,
    "batch_idx": 0,
    "loss": 0.4248414933681488,
    "learning_rate": 0.00041094235486789616,
    "timestep_mean": 503.90625,
    "epsilon_norm": 7.636394500732422,
    "prediction_norm": 5.815978527069092,
    "timestamp": "2025-07-14T12:35:17.624929"
  },
  {
    "epoch": 2886,
    "batch_idx": 32,
    "loss": 0.18157894909381866,
    "learning_rate": 0.00041094235486789616,
    "timestep_mean": 525.75,
    "epsilon_norm": 8.770524978637695,
    "prediction_norm": 7.402875900268555,
    "timestamp": "2025-07-14T12:35:17.630013"
  },
  {
    "epoch": 2886,
    "batch_idx": 64,
    "loss": 0.3264271020889282,
    "learning_rate": 0.00041094235486789616,
    "timestep_mean": 452.40625,
    "epsilon_norm": 8.924881935119629,
    "prediction_norm": 7.582577705383301,
    "timestamp": "2025-07-14T12:35:17.634922"
  },
  {
    "epoch": 2886,
    "batch_idx": 96,
    "loss": 0.24481500685214996,
    "learning_rate": 0.00041094235486789616,
    "timestep_mean": 525.0625,
    "epsilon_norm": 7.841598987579346,
    "prediction_norm": 7.337360382080078,
    "timestamp": "2025-07-14T12:35:17.639835"
  },
  {
    "epoch": 2886,
    "batch_idx": 128,
    "loss": 0.22728988528251648,
    "learning_rate": 0.00041094235486789616,
    "timestep_mean": 511.65625,
    "epsilon_norm": 9.36974811553955,
    "prediction_norm": 8.471085548400879,
    "timestamp": "2025-07-14T12:35:17.644687"
  },
  {
    "epoch": 2886,
    "batch_idx": 160,
    "loss": 0.356160044670105,
    "learning_rate": 0.00041094235486789616,
    "timestep_mean": 483.625,
    "epsilon_norm": 7.630453586578369,
    "prediction_norm": 5.7751336097717285,
    "timestamp": "2025-07-14T12:35:17.649558"
  },
  {
    "epoch": 2886,
    "batch_idx": 192,
    "loss": 0.6353338956832886,
    "learning_rate": 0.00041094235486789616,
    "timestep_mean": 354.75,
    "epsilon_norm": 4.663195610046387,
    "prediction_norm": 2.925320863723755,
    "timestamp": "2025-07-14T12:35:17.654423"
  },
  {
    "epoch": 2887,
    "batch_idx": 0,
    "loss": 0.23237574100494385,
    "learning_rate": 0.00041171530377327373,
    "timestep_mean": 482.84375,
    "epsilon_norm": 8.7134370803833,
    "prediction_norm": 7.87691068649292,
    "timestamp": "2025-07-14T12:35:17.659296"
  },
  {
    "epoch": 2887,
    "batch_idx": 32,
    "loss": 0.4543178677558899,
    "learning_rate": 0.00041171530377327373,
    "timestep_mean": 508.0625,
    "epsilon_norm": 8.066756248474121,
    "prediction_norm": 5.841052055358887,
    "timestamp": "2025-07-14T12:35:17.664234"
  },
  {
    "epoch": 2887,
    "batch_idx": 64,
    "loss": 0.15747639536857605,
    "learning_rate": 0.00041171530377327373,
    "timestep_mean": 560.8125,
    "epsilon_norm": 8.4563570022583,
    "prediction_norm": 8.02343463897705,
    "timestamp": "2025-07-14T12:35:17.669194"
  },
  {
    "epoch": 2887,
    "batch_idx": 96,
    "loss": 0.3107765018939972,
    "learning_rate": 0.00041171530377327373,
    "timestep_mean": 492.0,
    "epsilon_norm": 7.776337623596191,
    "prediction_norm": 6.637697696685791,
    "timestamp": "2025-07-14T12:35:17.674247"
  },
  {
    "epoch": 2887,
    "batch_idx": 128,
    "loss": 0.3265702426433563,
    "learning_rate": 0.00041171530377327373,
    "timestep_mean": 621.3125,
    "epsilon_norm": 8.128520965576172,
    "prediction_norm": 7.341191291809082,
    "timestamp": "2025-07-14T12:35:17.679069"
  },
  {
    "epoch": 2887,
    "batch_idx": 160,
    "loss": 0.145467147231102,
    "learning_rate": 0.00041171530377327373,
    "timestep_mean": 608.46875,
    "epsilon_norm": 7.217565059661865,
    "prediction_norm": 7.091174125671387,
    "timestamp": "2025-07-14T12:35:17.684028"
  },
  {
    "epoch": 2887,
    "batch_idx": 192,
    "loss": 0.26899227499961853,
    "learning_rate": 0.00041171530377327373,
    "timestep_mean": 513.625,
    "epsilon_norm": 3.834223985671997,
    "prediction_norm": 3.0985844135284424,
    "timestamp": "2025-07-14T12:35:17.688946"
  },
  {
    "epoch": 2888,
    "batch_idx": 0,
    "loss": 0.2295895516872406,
    "learning_rate": 0.00041248847051236266,
    "timestep_mean": 503.6875,
    "epsilon_norm": 8.556124687194824,
    "prediction_norm": 8.09472942352295,
    "timestamp": "2025-07-14T12:35:17.693832"
  },
  {
    "epoch": 2888,
    "batch_idx": 32,
    "loss": 0.18521052598953247,
    "learning_rate": 0.00041248847051236266,
    "timestep_mean": 501.8125,
    "epsilon_norm": 8.18779468536377,
    "prediction_norm": 7.940802097320557,
    "timestamp": "2025-07-14T12:35:17.698763"
  },
  {
    "epoch": 2888,
    "batch_idx": 64,
    "loss": 0.697382926940918,
    "learning_rate": 0.00041248847051236266,
    "timestep_mean": 436.34375,
    "epsilon_norm": 9.43880844116211,
    "prediction_norm": 7.698318004608154,
    "timestamp": "2025-07-14T12:35:17.703668"
  },
  {
    "epoch": 2888,
    "batch_idx": 96,
    "loss": 0.21878094971179962,
    "learning_rate": 0.00041248847051236266,
    "timestep_mean": 491.125,
    "epsilon_norm": 8.323469161987305,
    "prediction_norm": 7.626189708709717,
    "timestamp": "2025-07-14T12:35:17.708909"
  },
  {
    "epoch": 2888,
    "batch_idx": 128,
    "loss": 0.1973884105682373,
    "learning_rate": 0.00041248847051236266,
    "timestep_mean": 576.65625,
    "epsilon_norm": 7.4688639640808105,
    "prediction_norm": 7.0445733070373535,
    "timestamp": "2025-07-14T12:35:17.713752"
  },
  {
    "epoch": 2888,
    "batch_idx": 160,
    "loss": 0.11540675163269043,
    "learning_rate": 0.00041248847051236266,
    "timestep_mean": 580.0,
    "epsilon_norm": 8.414921760559082,
    "prediction_norm": 8.319820404052734,
    "timestamp": "2025-07-14T12:35:17.718717"
  },
  {
    "epoch": 2888,
    "batch_idx": 192,
    "loss": 0.08006042242050171,
    "learning_rate": 0.00041248847051236266,
    "timestep_mean": 584.0,
    "epsilon_norm": 3.1834218502044678,
    "prediction_norm": 2.9410505294799805,
    "timestamp": "2025-07-14T12:35:17.723603"
  },
  {
    "epoch": 2889,
    "batch_idx": 0,
    "loss": 0.23098090291023254,
    "learning_rate": 0.0004132618531774522,
    "timestep_mean": 519.0,
    "epsilon_norm": 7.055795192718506,
    "prediction_norm": 5.4845499992370605,
    "timestamp": "2025-07-14T12:35:17.728492"
  },
  {
    "epoch": 2889,
    "batch_idx": 32,
    "loss": 0.22738227248191833,
    "learning_rate": 0.0004132618531774522,
    "timestep_mean": 515.15625,
    "epsilon_norm": 7.090123176574707,
    "prediction_norm": 6.286265850067139,
    "timestamp": "2025-07-14T12:35:17.733312"
  },
  {
    "epoch": 2889,
    "batch_idx": 64,
    "loss": 0.24923543632030487,
    "learning_rate": 0.0004132618531774522,
    "timestep_mean": 510.21875,
    "epsilon_norm": 6.616640567779541,
    "prediction_norm": 5.153172016143799,
    "timestamp": "2025-07-14T12:35:17.738259"
  },
  {
    "epoch": 2889,
    "batch_idx": 96,
    "loss": 0.3054899573326111,
    "learning_rate": 0.0004132618531774522,
    "timestep_mean": 586.1875,
    "epsilon_norm": 9.303840637207031,
    "prediction_norm": 8.00839900970459,
    "timestamp": "2025-07-14T12:35:17.743085"
  },
  {
    "epoch": 2889,
    "batch_idx": 128,
    "loss": 0.2385319322347641,
    "learning_rate": 0.0004132618531774522,
    "timestep_mean": 527.15625,
    "epsilon_norm": 8.230361938476562,
    "prediction_norm": 6.762535572052002,
    "timestamp": "2025-07-14T12:35:17.747975"
  },
  {
    "epoch": 2889,
    "batch_idx": 160,
    "loss": 0.3182671070098877,
    "learning_rate": 0.0004132618531774522,
    "timestep_mean": 519.875,
    "epsilon_norm": 8.883537292480469,
    "prediction_norm": 6.6833038330078125,
    "timestamp": "2025-07-14T12:35:17.752803"
  },
  {
    "epoch": 2889,
    "batch_idx": 192,
    "loss": 0.3494243621826172,
    "learning_rate": 0.0004132618531774522,
    "timestep_mean": 531.625,
    "epsilon_norm": 3.8570733070373535,
    "prediction_norm": 2.763413429260254,
    "timestamp": "2025-07-14T12:35:17.757715"
  },
  {
    "epoch": 2890,
    "batch_idx": 0,
    "loss": 0.2799564003944397,
    "learning_rate": 0.0004140354498602963,
    "timestep_mean": 495.9375,
    "epsilon_norm": 8.288886070251465,
    "prediction_norm": 6.953462600708008,
    "timestamp": "2025-07-14T12:35:17.762568"
  },
  {
    "epoch": 2890,
    "batch_idx": 32,
    "loss": 0.3343757390975952,
    "learning_rate": 0.0004140354498602963,
    "timestep_mean": 480.8125,
    "epsilon_norm": 7.669986724853516,
    "prediction_norm": 5.74981689453125,
    "timestamp": "2025-07-14T12:35:17.767723"
  },
  {
    "epoch": 2890,
    "batch_idx": 64,
    "loss": 0.17982469499111176,
    "learning_rate": 0.0004140354498602963,
    "timestep_mean": 448.40625,
    "epsilon_norm": 6.994852066040039,
    "prediction_norm": 5.963651180267334,
    "timestamp": "2025-07-14T12:35:17.772738"
  },
  {
    "epoch": 2890,
    "batch_idx": 96,
    "loss": 0.19579362869262695,
    "learning_rate": 0.0004140354498602963,
    "timestep_mean": 536.1875,
    "epsilon_norm": 7.452767372131348,
    "prediction_norm": 6.533875942230225,
    "timestamp": "2025-07-14T12:35:17.777775"
  },
  {
    "epoch": 2890,
    "batch_idx": 128,
    "loss": 0.19239778816699982,
    "learning_rate": 0.0004140354498602963,
    "timestep_mean": 461.09375,
    "epsilon_norm": 7.954824924468994,
    "prediction_norm": 6.660279750823975,
    "timestamp": "2025-07-14T12:35:17.782935"
  },
  {
    "epoch": 2890,
    "batch_idx": 160,
    "loss": 0.2917221188545227,
    "learning_rate": 0.0004140354498602963,
    "timestep_mean": 496.65625,
    "epsilon_norm": 8.278956413269043,
    "prediction_norm": 6.011613845825195,
    "timestamp": "2025-07-14T12:35:17.788072"
  },
  {
    "epoch": 2890,
    "batch_idx": 192,
    "loss": 0.2742340862751007,
    "learning_rate": 0.0004140354498602963,
    "timestep_mean": 484.5,
    "epsilon_norm": 3.9052557945251465,
    "prediction_norm": 2.9675872325897217,
    "timestamp": "2025-07-14T12:35:17.793138"
  },
  {
    "epoch": 2891,
    "batch_idx": 0,
    "loss": 0.167128324508667,
    "learning_rate": 0.0004148092586521223,
    "timestep_mean": 550.53125,
    "epsilon_norm": 7.372208118438721,
    "prediction_norm": 6.658017158508301,
    "timestamp": "2025-07-14T12:35:17.798262"
  },
  {
    "epoch": 2891,
    "batch_idx": 32,
    "loss": 0.23624679446220398,
    "learning_rate": 0.0004148092586521223,
    "timestep_mean": 422.71875,
    "epsilon_norm": 7.456891059875488,
    "prediction_norm": 5.727545261383057,
    "timestamp": "2025-07-14T12:35:17.803515"
  },
  {
    "epoch": 2891,
    "batch_idx": 64,
    "loss": 0.31443601846694946,
    "learning_rate": 0.0004148092586521223,
    "timestep_mean": 558.15625,
    "epsilon_norm": 8.27632999420166,
    "prediction_norm": 6.698575973510742,
    "timestamp": "2025-07-14T12:35:17.808907"
  },
  {
    "epoch": 2891,
    "batch_idx": 96,
    "loss": 0.3244376480579376,
    "learning_rate": 0.0004148092586521223,
    "timestep_mean": 470.28125,
    "epsilon_norm": 7.650969982147217,
    "prediction_norm": 6.041589736938477,
    "timestamp": "2025-07-14T12:35:17.813941"
  },
  {
    "epoch": 2891,
    "batch_idx": 128,
    "loss": 0.22928068041801453,
    "learning_rate": 0.0004148092586521223,
    "timestep_mean": 507.625,
    "epsilon_norm": 7.510361194610596,
    "prediction_norm": 6.5502214431762695,
    "timestamp": "2025-07-14T12:35:17.819036"
  },
  {
    "epoch": 2891,
    "batch_idx": 160,
    "loss": 0.30960792303085327,
    "learning_rate": 0.0004148092586521223,
    "timestep_mean": 489.125,
    "epsilon_norm": 8.514092445373535,
    "prediction_norm": 7.594974994659424,
    "timestamp": "2025-07-14T12:35:17.823995"
  },
  {
    "epoch": 2891,
    "batch_idx": 192,
    "loss": 0.12422117590904236,
    "learning_rate": 0.0004148092586521223,
    "timestep_mean": 524.625,
    "epsilon_norm": 3.095242500305176,
    "prediction_norm": 2.785015344619751,
    "timestamp": "2025-07-14T12:35:17.829105"
  },
  {
    "epoch": 2892,
    "batch_idx": 0,
    "loss": 0.10347206890583038,
    "learning_rate": 0.00041558327764363407,
    "timestep_mean": 602.84375,
    "epsilon_norm": 6.854043483734131,
    "prediction_norm": 7.086852073669434,
    "timestamp": "2025-07-14T12:35:17.834234"
  },
  {
    "epoch": 2892,
    "batch_idx": 32,
    "loss": 0.32101815938949585,
    "learning_rate": 0.00041558327764363407,
    "timestep_mean": 551.3125,
    "epsilon_norm": 7.002908706665039,
    "prediction_norm": 5.630397319793701,
    "timestamp": "2025-07-14T12:35:17.839331"
  },
  {
    "epoch": 2892,
    "batch_idx": 64,
    "loss": 0.3017304539680481,
    "learning_rate": 0.00041558327764363407,
    "timestep_mean": 482.03125,
    "epsilon_norm": 7.658779621124268,
    "prediction_norm": 6.091226577758789,
    "timestamp": "2025-07-14T12:35:17.844502"
  },
  {
    "epoch": 2892,
    "batch_idx": 96,
    "loss": 0.254748672246933,
    "learning_rate": 0.00041558327764363407,
    "timestep_mean": 505.21875,
    "epsilon_norm": 6.392581939697266,
    "prediction_norm": 5.324066638946533,
    "timestamp": "2025-07-14T12:35:17.849727"
  },
  {
    "epoch": 2892,
    "batch_idx": 128,
    "loss": 0.2568373382091522,
    "learning_rate": 0.00041558327764363407,
    "timestep_mean": 597.78125,
    "epsilon_norm": 8.485753059387207,
    "prediction_norm": 8.105912208557129,
    "timestamp": "2025-07-14T12:35:17.855041"
  },
  {
    "epoch": 2892,
    "batch_idx": 160,
    "loss": 0.5037291646003723,
    "learning_rate": 0.00041558327764363407,
    "timestep_mean": 468.625,
    "epsilon_norm": 7.793978214263916,
    "prediction_norm": 5.489922523498535,
    "timestamp": "2025-07-14T12:35:17.860202"
  },
  {
    "epoch": 2892,
    "batch_idx": 192,
    "loss": 0.30362772941589355,
    "learning_rate": 0.00041558327764363407,
    "timestep_mean": 497.875,
    "epsilon_norm": 3.7297089099884033,
    "prediction_norm": 3.0224452018737793,
    "timestamp": "2025-07-14T12:35:17.865349"
  },
  {
    "epoch": 2893,
    "batch_idx": 0,
    "loss": 0.31647977232933044,
    "learning_rate": 0.000416357504925017,
    "timestep_mean": 507.03125,
    "epsilon_norm": 8.166707992553711,
    "prediction_norm": 6.730068206787109,
    "timestamp": "2025-07-14T12:35:17.870529"
  },
  {
    "epoch": 2893,
    "batch_idx": 32,
    "loss": 0.14973431825637817,
    "learning_rate": 0.000416357504925017,
    "timestep_mean": 547.4375,
    "epsilon_norm": 8.119842529296875,
    "prediction_norm": 7.476717472076416,
    "timestamp": "2025-07-14T12:35:17.875635"
  },
  {
    "epoch": 2893,
    "batch_idx": 64,
    "loss": 0.22408777475357056,
    "learning_rate": 0.000416357504925017,
    "timestep_mean": 561.59375,
    "epsilon_norm": 8.38562297821045,
    "prediction_norm": 7.604482173919678,
    "timestamp": "2025-07-14T12:35:17.880714"
  },
  {
    "epoch": 2893,
    "batch_idx": 96,
    "loss": 0.2819078862667084,
    "learning_rate": 0.000416357504925017,
    "timestep_mean": 415.3125,
    "epsilon_norm": 7.3455424308776855,
    "prediction_norm": 6.219442367553711,
    "timestamp": "2025-07-14T12:35:17.885942"
  },
  {
    "epoch": 2893,
    "batch_idx": 128,
    "loss": 0.18696406483650208,
    "learning_rate": 0.000416357504925017,
    "timestep_mean": 514.4375,
    "epsilon_norm": 9.739821434020996,
    "prediction_norm": 8.819525718688965,
    "timestamp": "2025-07-14T12:35:17.891064"
  },
  {
    "epoch": 2893,
    "batch_idx": 160,
    "loss": 0.3909326195716858,
    "learning_rate": 0.000416357504925017,
    "timestep_mean": 505.59375,
    "epsilon_norm": 8.967541694641113,
    "prediction_norm": 7.757875442504883,
    "timestamp": "2025-07-14T12:35:17.896393"
  },
  {
    "epoch": 2893,
    "batch_idx": 192,
    "loss": 0.14755260944366455,
    "learning_rate": 0.000416357504925017,
    "timestep_mean": 667.75,
    "epsilon_norm": 4.677928447723389,
    "prediction_norm": 4.5095014572143555,
    "timestamp": "2025-07-14T12:35:17.901620"
  },
  {
    "epoch": 2894,
    "batch_idx": 0,
    "loss": 0.36285877227783203,
    "learning_rate": 0.000417131938585941,
    "timestep_mean": 434.53125,
    "epsilon_norm": 6.706950664520264,
    "prediction_norm": 5.420599460601807,
    "timestamp": "2025-07-14T12:35:17.906739"
  },
  {
    "epoch": 2894,
    "batch_idx": 32,
    "loss": 0.283935546875,
    "learning_rate": 0.000417131938585941,
    "timestep_mean": 507.59375,
    "epsilon_norm": 7.980386734008789,
    "prediction_norm": 6.549922943115234,
    "timestamp": "2025-07-14T12:35:17.911972"
  },
  {
    "epoch": 2894,
    "batch_idx": 64,
    "loss": 0.477188378572464,
    "learning_rate": 0.000417131938585941,
    "timestep_mean": 420.1875,
    "epsilon_norm": 9.839055061340332,
    "prediction_norm": 8.255561828613281,
    "timestamp": "2025-07-14T12:35:17.917259"
  },
  {
    "epoch": 2894,
    "batch_idx": 96,
    "loss": 0.23790483176708221,
    "learning_rate": 0.000417131938585941,
    "timestep_mean": 477.90625,
    "epsilon_norm": 7.56069278717041,
    "prediction_norm": 6.494931697845459,
    "timestamp": "2025-07-14T12:35:17.922386"
  },
  {
    "epoch": 2894,
    "batch_idx": 128,
    "loss": 0.26415324211120605,
    "learning_rate": 0.000417131938585941,
    "timestep_mean": 419.3125,
    "epsilon_norm": 7.807750701904297,
    "prediction_norm": 6.140936374664307,
    "timestamp": "2025-07-14T12:35:17.927470"
  },
  {
    "epoch": 2894,
    "batch_idx": 160,
    "loss": 0.22178742289543152,
    "learning_rate": 0.000417131938585941,
    "timestep_mean": 472.40625,
    "epsilon_norm": 7.987759590148926,
    "prediction_norm": 7.49483585357666,
    "timestamp": "2025-07-14T12:35:17.932619"
  },
  {
    "epoch": 2894,
    "batch_idx": 192,
    "loss": 0.11277937889099121,
    "learning_rate": 0.000417131938585941,
    "timestep_mean": 561.0,
    "epsilon_norm": 3.8570733070373535,
    "prediction_norm": 3.873981475830078,
    "timestamp": "2025-07-14T12:35:17.937742"
  },
  {
    "epoch": 2895,
    "batch_idx": 0,
    "loss": 0.28298306465148926,
    "learning_rate": 0.0004179065767155694,
    "timestep_mean": 494.875,
    "epsilon_norm": 9.025156021118164,
    "prediction_norm": 7.2695207595825195,
    "timestamp": "2025-07-14T12:35:17.942882"
  },
  {
    "epoch": 2895,
    "batch_idx": 32,
    "loss": 0.2216198891401291,
    "learning_rate": 0.0004179065767155694,
    "timestep_mean": 528.0625,
    "epsilon_norm": 7.413363456726074,
    "prediction_norm": 6.761643886566162,
    "timestamp": "2025-07-14T12:35:17.948024"
  },
  {
    "epoch": 2895,
    "batch_idx": 64,
    "loss": 0.2986622452735901,
    "learning_rate": 0.0004179065767155694,
    "timestep_mean": 500.4375,
    "epsilon_norm": 7.45559024810791,
    "prediction_norm": 6.693334579467773,
    "timestamp": "2025-07-14T12:35:17.953199"
  },
  {
    "epoch": 2895,
    "batch_idx": 96,
    "loss": 0.42135387659072876,
    "learning_rate": 0.0004179065767155694,
    "timestep_mean": 496.15625,
    "epsilon_norm": 7.881925582885742,
    "prediction_norm": 6.23579216003418,
    "timestamp": "2025-07-14T12:35:17.958398"
  },
  {
    "epoch": 2895,
    "batch_idx": 128,
    "loss": 0.3574794828891754,
    "learning_rate": 0.0004179065767155694,
    "timestep_mean": 387.09375,
    "epsilon_norm": 8.76192569732666,
    "prediction_norm": 6.908168792724609,
    "timestamp": "2025-07-14T12:35:17.963590"
  },
  {
    "epoch": 2895,
    "batch_idx": 160,
    "loss": 0.38422733545303345,
    "learning_rate": 0.0004179065767155694,
    "timestep_mean": 466.21875,
    "epsilon_norm": 8.352592468261719,
    "prediction_norm": 6.840973377227783,
    "timestamp": "2025-07-14T12:35:17.968811"
  },
  {
    "epoch": 2895,
    "batch_idx": 192,
    "loss": 0.35346803069114685,
    "learning_rate": 0.0004179065767155694,
    "timestep_mean": 546.875,
    "epsilon_norm": 3.989590883255005,
    "prediction_norm": 3.4544780254364014,
    "timestamp": "2025-07-14T12:35:17.973979"
  },
  {
    "epoch": 2896,
    "batch_idx": 0,
    "loss": 0.15512579679489136,
    "learning_rate": 0.0004186814174025595,
    "timestep_mean": 496.40625,
    "epsilon_norm": 7.896182537078857,
    "prediction_norm": 7.927900314331055,
    "timestamp": "2025-07-14T12:35:17.979202"
  },
  {
    "epoch": 2896,
    "batch_idx": 32,
    "loss": 0.29059913754463196,
    "learning_rate": 0.0004186814174025595,
    "timestep_mean": 481.21875,
    "epsilon_norm": 7.916292190551758,
    "prediction_norm": 7.0877299308776855,
    "timestamp": "2025-07-14T12:35:17.984375"
  },
  {
    "epoch": 2896,
    "batch_idx": 64,
    "loss": 0.30035698413848877,
    "learning_rate": 0.0004186814174025595,
    "timestep_mean": 530.15625,
    "epsilon_norm": 8.580965042114258,
    "prediction_norm": 7.151876449584961,
    "timestamp": "2025-07-14T12:35:17.989227"
  },
  {
    "epoch": 2896,
    "batch_idx": 96,
    "loss": 0.38393011689186096,
    "learning_rate": 0.0004186814174025595,
    "timestep_mean": 516.21875,
    "epsilon_norm": 8.307808876037598,
    "prediction_norm": 6.487254619598389,
    "timestamp": "2025-07-14T12:35:17.994029"
  },
  {
    "epoch": 2896,
    "batch_idx": 128,
    "loss": 0.12949660420417786,
    "learning_rate": 0.0004186814174025595,
    "timestep_mean": 510.8125,
    "epsilon_norm": 7.288877010345459,
    "prediction_norm": 6.985658645629883,
    "timestamp": "2025-07-14T12:35:17.998895"
  },
  {
    "epoch": 2896,
    "batch_idx": 160,
    "loss": 0.38937127590179443,
    "learning_rate": 0.0004186814174025595,
    "timestep_mean": 523.4375,
    "epsilon_norm": 8.794434547424316,
    "prediction_norm": 6.647542476654053,
    "timestamp": "2025-07-14T12:35:18.003766"
  },
  {
    "epoch": 2896,
    "batch_idx": 192,
    "loss": 0.4731164574623108,
    "learning_rate": 0.0004186814174025595,
    "timestep_mean": 454.625,
    "epsilon_norm": 3.8744380474090576,
    "prediction_norm": 2.938922882080078,
    "timestamp": "2025-07-14T12:35:18.008655"
  },
  {
    "epoch": 2897,
    "batch_idx": 0,
    "loss": 0.16271832585334778,
    "learning_rate": 0.000419456458735068,
    "timestep_mean": 480.40625,
    "epsilon_norm": 8.083266258239746,
    "prediction_norm": 7.100708484649658,
    "timestamp": "2025-07-14T12:35:18.013700"
  },
  {
    "epoch": 2897,
    "batch_idx": 32,
    "loss": 0.2525728940963745,
    "learning_rate": 0.000419456458735068,
    "timestep_mean": 452.59375,
    "epsilon_norm": 6.401357650756836,
    "prediction_norm": 5.654067039489746,
    "timestamp": "2025-07-14T12:35:18.018796"
  },
  {
    "epoch": 2897,
    "batch_idx": 64,
    "loss": 0.23919379711151123,
    "learning_rate": 0.000419456458735068,
    "timestep_mean": 447.4375,
    "epsilon_norm": 7.746788024902344,
    "prediction_norm": 6.329591274261475,
    "timestamp": "2025-07-14T12:35:18.023917"
  },
  {
    "epoch": 2897,
    "batch_idx": 96,
    "loss": 0.33932042121887207,
    "learning_rate": 0.000419456458735068,
    "timestep_mean": 503.625,
    "epsilon_norm": 8.068778991699219,
    "prediction_norm": 6.085757732391357,
    "timestamp": "2025-07-14T12:35:18.029065"
  },
  {
    "epoch": 2897,
    "batch_idx": 128,
    "loss": 0.3141078054904938,
    "learning_rate": 0.000419456458735068,
    "timestep_mean": 519.46875,
    "epsilon_norm": 7.6037187576293945,
    "prediction_norm": 5.782103538513184,
    "timestamp": "2025-07-14T12:35:18.034312"
  },
  {
    "epoch": 2897,
    "batch_idx": 160,
    "loss": 0.2698579430580139,
    "learning_rate": 0.000419456458735068,
    "timestep_mean": 506.6875,
    "epsilon_norm": 8.119633674621582,
    "prediction_norm": 6.548745632171631,
    "timestamp": "2025-07-14T12:35:18.039312"
  },
  {
    "epoch": 2897,
    "batch_idx": 192,
    "loss": 0.45320791006088257,
    "learning_rate": 0.000419456458735068,
    "timestep_mean": 480.625,
    "epsilon_norm": 4.011654853820801,
    "prediction_norm": 2.8839879035949707,
    "timestamp": "2025-07-14T12:35:18.044386"
  },
  {
    "epoch": 2898,
    "batch_idx": 0,
    "loss": 0.27685651183128357,
    "learning_rate": 0.000420231698800758,
    "timestep_mean": 516.4375,
    "epsilon_norm": 8.781291007995605,
    "prediction_norm": 7.170007705688477,
    "timestamp": "2025-07-14T12:35:18.049571"
  },
  {
    "epoch": 2898,
    "batch_idx": 32,
    "loss": 0.30207765102386475,
    "learning_rate": 0.000420231698800758,
    "timestep_mean": 500.46875,
    "epsilon_norm": 7.359311580657959,
    "prediction_norm": 6.513112545013428,
    "timestamp": "2025-07-14T12:35:18.054771"
  },
  {
    "epoch": 2898,
    "batch_idx": 64,
    "loss": 0.1797879934310913,
    "learning_rate": 0.000420231698800758,
    "timestep_mean": 565.25,
    "epsilon_norm": 7.914876461029053,
    "prediction_norm": 6.443735122680664,
    "timestamp": "2025-07-14T12:35:18.059804"
  },
  {
    "epoch": 2898,
    "batch_idx": 96,
    "loss": 0.16025884449481964,
    "learning_rate": 0.000420231698800758,
    "timestep_mean": 482.5625,
    "epsilon_norm": 7.875601291656494,
    "prediction_norm": 6.838981628417969,
    "timestamp": "2025-07-14T12:35:18.065044"
  },
  {
    "epoch": 2898,
    "batch_idx": 128,
    "loss": 0.26235002279281616,
    "learning_rate": 0.000420231698800758,
    "timestep_mean": 500.90625,
    "epsilon_norm": 8.311929702758789,
    "prediction_norm": 7.417191028594971,
    "timestamp": "2025-07-14T12:35:18.070253"
  },
  {
    "epoch": 2898,
    "batch_idx": 160,
    "loss": 0.4953717589378357,
    "learning_rate": 0.000420231698800758,
    "timestep_mean": 483.25,
    "epsilon_norm": 8.254246711730957,
    "prediction_norm": 6.892363548278809,
    "timestamp": "2025-07-14T12:35:18.075401"
  },
  {
    "epoch": 2898,
    "batch_idx": 192,
    "loss": 0.18848799169063568,
    "learning_rate": 0.000420231698800758,
    "timestep_mean": 521.875,
    "epsilon_norm": 4.784523963928223,
    "prediction_norm": 4.393856048583984,
    "timestamp": "2025-07-14T12:35:18.080519"
  },
  {
    "epoch": 2899,
    "batch_idx": 0,
    "loss": 0.335358589887619,
    "learning_rate": 0.00042100713568680125,
    "timestep_mean": 520.03125,
    "epsilon_norm": 7.801929950714111,
    "prediction_norm": 6.072689533233643,
    "timestamp": "2025-07-14T12:35:18.085637"
  },
  {
    "epoch": 2899,
    "batch_idx": 32,
    "loss": 0.17048200964927673,
    "learning_rate": 0.00042100713568680125,
    "timestep_mean": 519.90625,
    "epsilon_norm": 8.109814643859863,
    "prediction_norm": 7.437401294708252,
    "timestamp": "2025-07-14T12:35:18.090913"
  },
  {
    "epoch": 2899,
    "batch_idx": 64,
    "loss": 0.24451123178005219,
    "learning_rate": 0.00042100713568680125,
    "timestep_mean": 507.59375,
    "epsilon_norm": 8.287282943725586,
    "prediction_norm": 7.460370063781738,
    "timestamp": "2025-07-14T12:35:18.096041"
  },
  {
    "epoch": 2899,
    "batch_idx": 96,
    "loss": 0.4455311894416809,
    "learning_rate": 0.00042100713568680125,
    "timestep_mean": 472.96875,
    "epsilon_norm": 8.098529815673828,
    "prediction_norm": 6.268576622009277,
    "timestamp": "2025-07-14T12:35:18.101301"
  },
  {
    "epoch": 2899,
    "batch_idx": 128,
    "loss": 0.1765369176864624,
    "learning_rate": 0.00042100713568680125,
    "timestep_mean": 490.0,
    "epsilon_norm": 7.330423355102539,
    "prediction_norm": 6.9157867431640625,
    "timestamp": "2025-07-14T12:35:18.106488"
  },
  {
    "epoch": 2899,
    "batch_idx": 160,
    "loss": 0.30630600452423096,
    "learning_rate": 0.00042100713568680125,
    "timestep_mean": 473.625,
    "epsilon_norm": 7.33781623840332,
    "prediction_norm": 5.809347629547119,
    "timestamp": "2025-07-14T12:35:18.111735"
  },
  {
    "epoch": 2899,
    "batch_idx": 192,
    "loss": 0.1767248809337616,
    "learning_rate": 0.00042100713568680125,
    "timestep_mean": 606.0,
    "epsilon_norm": 3.9758810997009277,
    "prediction_norm": 3.4915268421173096,
    "timestamp": "2025-07-14T12:35:18.116886"
  },
  {
    "epoch": 2900,
    "batch_idx": 0,
    "loss": 0.23253241181373596,
    "learning_rate": 0.0004217827674798856,
    "timestep_mean": 429.4375,
    "epsilon_norm": 7.4132843017578125,
    "prediction_norm": 6.360658645629883,
    "timestamp": "2025-07-14T12:35:18.122109"
  },
  {
    "epoch": 2900,
    "batch_idx": 32,
    "loss": 0.41044163703918457,
    "learning_rate": 0.0004217827674798856,
    "timestep_mean": 415.8125,
    "epsilon_norm": 7.48488712310791,
    "prediction_norm": 6.595294952392578,
    "timestamp": "2025-07-14T12:35:18.127141"
  },
  {
    "epoch": 2900,
    "batch_idx": 64,
    "loss": 0.593823254108429,
    "learning_rate": 0.0004217827674798856,
    "timestep_mean": 501.6875,
    "epsilon_norm": 8.719158172607422,
    "prediction_norm": 5.73227071762085,
    "timestamp": "2025-07-14T12:35:18.132381"
  },
  {
    "epoch": 2900,
    "batch_idx": 96,
    "loss": 0.3245505690574646,
    "learning_rate": 0.0004217827674798856,
    "timestep_mean": 554.34375,
    "epsilon_norm": 8.22205924987793,
    "prediction_norm": 6.697973728179932,
    "timestamp": "2025-07-14T12:35:18.137570"
  },
  {
    "epoch": 2900,
    "batch_idx": 128,
    "loss": 0.7052972316741943,
    "learning_rate": 0.0004217827674798856,
    "timestep_mean": 458.375,
    "epsilon_norm": 9.275603294372559,
    "prediction_norm": 6.675199508666992,
    "timestamp": "2025-07-14T12:35:18.142649"
  },
  {
    "epoch": 2900,
    "batch_idx": 160,
    "loss": 0.08552058786153793,
    "learning_rate": 0.0004217827674798856,
    "timestep_mean": 549.46875,
    "epsilon_norm": 7.8156938552856445,
    "prediction_norm": 7.5420942306518555,
    "timestamp": "2025-07-14T12:35:18.147624"
  },
  {
    "epoch": 2900,
    "batch_idx": 192,
    "loss": 1.1088249683380127,
    "learning_rate": 0.0004217827674798856,
    "timestep_mean": 355.25,
    "epsilon_norm": 4.7320876121521,
    "prediction_norm": 2.0816662311553955,
    "timestamp": "2025-07-14T12:35:18.152707"
  },
  {
    "epoch": 2901,
    "batch_idx": 0,
    "loss": 0.13854540884494781,
    "learning_rate": 0.0004225585922662154,
    "timestep_mean": 432.125,
    "epsilon_norm": 6.391837120056152,
    "prediction_norm": 6.313423156738281,
    "timestamp": "2025-07-14T12:35:18.157891"
  },
  {
    "epoch": 2901,
    "batch_idx": 32,
    "loss": 0.3733774423599243,
    "learning_rate": 0.0004225585922662154,
    "timestep_mean": 496.4375,
    "epsilon_norm": 8.708166122436523,
    "prediction_norm": 7.116513252258301,
    "timestamp": "2025-07-14T12:35:18.163164"
  },
  {
    "epoch": 2901,
    "batch_idx": 64,
    "loss": 0.23342479765415192,
    "learning_rate": 0.0004225585922662154,
    "timestep_mean": 593.6875,
    "epsilon_norm": 9.317978858947754,
    "prediction_norm": 7.951543807983398,
    "timestamp": "2025-07-14T12:35:18.168259"
  },
  {
    "epoch": 2901,
    "batch_idx": 96,
    "loss": 0.13374561071395874,
    "learning_rate": 0.0004225585922662154,
    "timestep_mean": 581.53125,
    "epsilon_norm": 8.054977416992188,
    "prediction_norm": 7.386089324951172,
    "timestamp": "2025-07-14T12:35:18.173394"
  },
  {
    "epoch": 2901,
    "batch_idx": 128,
    "loss": 0.2604220509529114,
    "learning_rate": 0.0004225585922662154,
    "timestep_mean": 467.78125,
    "epsilon_norm": 7.917669773101807,
    "prediction_norm": 6.866743087768555,
    "timestamp": "2025-07-14T12:35:18.178515"
  },
  {
    "epoch": 2901,
    "batch_idx": 160,
    "loss": 0.35285037755966187,
    "learning_rate": 0.0004225585922662154,
    "timestep_mean": 539.625,
    "epsilon_norm": 9.431815147399902,
    "prediction_norm": 7.533426761627197,
    "timestamp": "2025-07-14T12:35:18.183623"
  },
  {
    "epoch": 2901,
    "batch_idx": 192,
    "loss": 0.04540938511490822,
    "learning_rate": 0.0004225585922662154,
    "timestep_mean": 656.25,
    "epsilon_norm": 3.9227864742279053,
    "prediction_norm": 3.8531394004821777,
    "timestamp": "2025-07-14T12:35:18.188758"
  },
  {
    "epoch": 2902,
    "batch_idx": 0,
    "loss": 0.2254926562309265,
    "learning_rate": 0.00042333460813152065,
    "timestep_mean": 459.25,
    "epsilon_norm": 7.430685520172119,
    "prediction_norm": 6.866888046264648,
    "timestamp": "2025-07-14T12:35:18.193848"
  },
  {
    "epoch": 2902,
    "batch_idx": 32,
    "loss": 0.335921049118042,
    "learning_rate": 0.00042333460813152065,
    "timestep_mean": 512.09375,
    "epsilon_norm": 8.577263832092285,
    "prediction_norm": 7.743655681610107,
    "timestamp": "2025-07-14T12:35:18.199033"
  },
  {
    "epoch": 2902,
    "batch_idx": 64,
    "loss": 0.15038847923278809,
    "learning_rate": 0.00042333460813152065,
    "timestep_mean": 543.5,
    "epsilon_norm": 7.773843765258789,
    "prediction_norm": 7.378807067871094,
    "timestamp": "2025-07-14T12:35:18.204176"
  },
  {
    "epoch": 2902,
    "batch_idx": 96,
    "loss": 0.3151368200778961,
    "learning_rate": 0.00042333460813152065,
    "timestep_mean": 377.375,
    "epsilon_norm": 9.410783767700195,
    "prediction_norm": 7.748067378997803,
    "timestamp": "2025-07-14T12:35:18.209332"
  },
  {
    "epoch": 2902,
    "batch_idx": 128,
    "loss": 0.36398178339004517,
    "learning_rate": 0.00042333460813152065,
    "timestep_mean": 463.0,
    "epsilon_norm": 8.381891250610352,
    "prediction_norm": 7.188803195953369,
    "timestamp": "2025-07-14T12:35:18.214287"
  },
  {
    "epoch": 2902,
    "batch_idx": 160,
    "loss": 0.2829888164997101,
    "learning_rate": 0.00042333460813152065,
    "timestep_mean": 487.21875,
    "epsilon_norm": 8.232686042785645,
    "prediction_norm": 7.233038425445557,
    "timestamp": "2025-07-14T12:35:18.219165"
  },
  {
    "epoch": 2902,
    "batch_idx": 192,
    "loss": 0.09209012985229492,
    "learning_rate": 0.00042333460813152065,
    "timestep_mean": 614.625,
    "epsilon_norm": 4.986578464508057,
    "prediction_norm": 4.306352138519287,
    "timestamp": "2025-07-14T12:35:18.224009"
  },
  {
    "epoch": 2903,
    "batch_idx": 0,
    "loss": 0.25424519181251526,
    "learning_rate": 0.00042411081316105965,
    "timestep_mean": 557.5625,
    "epsilon_norm": 6.801502704620361,
    "prediction_norm": 6.178619384765625,
    "timestamp": "2025-07-14T12:35:18.228988"
  },
  {
    "epoch": 2903,
    "batch_idx": 32,
    "loss": 0.22227275371551514,
    "learning_rate": 0.00042411081316105965,
    "timestep_mean": 503.4375,
    "epsilon_norm": 8.10256290435791,
    "prediction_norm": 7.034272193908691,
    "timestamp": "2025-07-14T12:35:18.233988"
  },
  {
    "epoch": 2903,
    "batch_idx": 64,
    "loss": 0.26240551471710205,
    "learning_rate": 0.00042411081316105965,
    "timestep_mean": 518.9375,
    "epsilon_norm": 8.35122013092041,
    "prediction_norm": 7.035844802856445,
    "timestamp": "2025-07-14T12:35:18.238825"
  },
  {
    "epoch": 2903,
    "batch_idx": 96,
    "loss": 0.18801411986351013,
    "learning_rate": 0.00042411081316105965,
    "timestep_mean": 495.8125,
    "epsilon_norm": 7.467818737030029,
    "prediction_norm": 6.529819965362549,
    "timestamp": "2025-07-14T12:35:18.243612"
  },
  {
    "epoch": 2903,
    "batch_idx": 128,
    "loss": 0.43975400924682617,
    "learning_rate": 0.00042411081316105965,
    "timestep_mean": 545.3125,
    "epsilon_norm": 8.734613418579102,
    "prediction_norm": 6.942826271057129,
    "timestamp": "2025-07-14T12:35:18.248492"
  },
  {
    "epoch": 2903,
    "batch_idx": 160,
    "loss": 0.29251042008399963,
    "learning_rate": 0.00042411081316105965,
    "timestep_mean": 447.65625,
    "epsilon_norm": 7.343037128448486,
    "prediction_norm": 5.672811985015869,
    "timestamp": "2025-07-14T12:35:18.253365"
  },
  {
    "epoch": 2903,
    "batch_idx": 192,
    "loss": 0.44638270139694214,
    "learning_rate": 0.00042411081316105965,
    "timestep_mean": 407.5,
    "epsilon_norm": 4.771358489990234,
    "prediction_norm": 3.2016124725341797,
    "timestamp": "2025-07-14T12:35:18.258311"
  },
  {
    "epoch": 2904,
    "batch_idx": 0,
    "loss": 0.1538158655166626,
    "learning_rate": 0.0004248872054396229,
    "timestep_mean": 451.78125,
    "epsilon_norm": 6.144526481628418,
    "prediction_norm": 5.664215564727783,
    "timestamp": "2025-07-14T12:35:18.263305"
  },
  {
    "epoch": 2904,
    "batch_idx": 32,
    "loss": 0.2007514238357544,
    "learning_rate": 0.0004248872054396229,
    "timestep_mean": 532.21875,
    "epsilon_norm": 9.209380149841309,
    "prediction_norm": 8.527185440063477,
    "timestamp": "2025-07-14T12:35:18.268189"
  },
  {
    "epoch": 2904,
    "batch_idx": 64,
    "loss": 0.21697422862052917,
    "learning_rate": 0.0004248872054396229,
    "timestep_mean": 519.875,
    "epsilon_norm": 7.844576358795166,
    "prediction_norm": 7.08783483505249,
    "timestamp": "2025-07-14T12:35:18.273007"
  },
  {
    "epoch": 2904,
    "batch_idx": 96,
    "loss": 0.46208542585372925,
    "learning_rate": 0.0004248872054396229,
    "timestep_mean": 477.09375,
    "epsilon_norm": 9.270955085754395,
    "prediction_norm": 7.276562213897705,
    "timestamp": "2025-07-14T12:35:18.277874"
  },
  {
    "epoch": 2904,
    "batch_idx": 128,
    "loss": 0.21040183305740356,
    "learning_rate": 0.0004248872054396229,
    "timestep_mean": 443.9375,
    "epsilon_norm": 6.8195366859436035,
    "prediction_norm": 6.136470317840576,
    "timestamp": "2025-07-14T12:35:18.282784"
  },
  {
    "epoch": 2904,
    "batch_idx": 160,
    "loss": 0.15162017941474915,
    "learning_rate": 0.0004248872054396229,
    "timestep_mean": 547.09375,
    "epsilon_norm": 6.180505752563477,
    "prediction_norm": 4.982905864715576,
    "timestamp": "2025-07-14T12:35:18.287707"
  },
  {
    "epoch": 2904,
    "batch_idx": 192,
    "loss": 0.07536575198173523,
    "learning_rate": 0.0004248872054396229,
    "timestep_mean": 600.25,
    "epsilon_norm": 3.1000216007232666,
    "prediction_norm": 2.9258015155792236,
    "timestamp": "2025-07-14T12:35:18.292618"
  },
  {
    "epoch": 2905,
    "batch_idx": 0,
    "loss": 0.4639321565628052,
    "learning_rate": 0.0004256637830515394,
    "timestep_mean": 515.15625,
    "epsilon_norm": 8.925491333007812,
    "prediction_norm": 7.096874713897705,
    "timestamp": "2025-07-14T12:35:18.297463"
  },
  {
    "epoch": 2905,
    "batch_idx": 32,
    "loss": 0.15103469789028168,
    "learning_rate": 0.0004256637830515394,
    "timestep_mean": 529.03125,
    "epsilon_norm": 7.813577175140381,
    "prediction_norm": 6.867807388305664,
    "timestamp": "2025-07-14T12:35:18.302339"
  },
  {
    "epoch": 2905,
    "batch_idx": 64,
    "loss": 0.20886766910552979,
    "learning_rate": 0.0004256637830515394,
    "timestep_mean": 451.59375,
    "epsilon_norm": 7.936944007873535,
    "prediction_norm": 6.773038864135742,
    "timestamp": "2025-07-14T12:35:18.307279"
  },
  {
    "epoch": 2905,
    "batch_idx": 96,
    "loss": 0.44284912943840027,
    "learning_rate": 0.0004256637830515394,
    "timestep_mean": 444.28125,
    "epsilon_norm": 8.053625106811523,
    "prediction_norm": 5.358806133270264,
    "timestamp": "2025-07-14T12:35:18.312433"
  },
  {
    "epoch": 2905,
    "batch_idx": 128,
    "loss": 0.2798970341682434,
    "learning_rate": 0.0004256637830515394,
    "timestep_mean": 604.1875,
    "epsilon_norm": 9.115836143493652,
    "prediction_norm": 8.378021240234375,
    "timestamp": "2025-07-14T12:35:18.317538"
  },
  {
    "epoch": 2905,
    "batch_idx": 160,
    "loss": 0.2647879421710968,
    "learning_rate": 0.0004256637830515394,
    "timestep_mean": 537.09375,
    "epsilon_norm": 7.797919273376465,
    "prediction_norm": 7.142525672912598,
    "timestamp": "2025-07-14T12:35:18.322528"
  },
  {
    "epoch": 2905,
    "batch_idx": 192,
    "loss": 0.2762587070465088,
    "learning_rate": 0.0004256637830515394,
    "timestep_mean": 480.375,
    "epsilon_norm": 3.9609925746917725,
    "prediction_norm": 2.8287625312805176,
    "timestamp": "2025-07-14T12:35:18.327594"
  },
  {
    "epoch": 2906,
    "batch_idx": 0,
    "loss": 0.16551369428634644,
    "learning_rate": 0.0004264405440806826,
    "timestep_mean": 516.8125,
    "epsilon_norm": 7.97510290145874,
    "prediction_norm": 6.6641526222229,
    "timestamp": "2025-07-14T12:35:18.332685"
  },
  {
    "epoch": 2906,
    "batch_idx": 32,
    "loss": 0.26295679807662964,
    "learning_rate": 0.0004264405440806826,
    "timestep_mean": 547.15625,
    "epsilon_norm": 7.95758581161499,
    "prediction_norm": 7.104661464691162,
    "timestamp": "2025-07-14T12:35:18.337959"
  },
  {
    "epoch": 2906,
    "batch_idx": 64,
    "loss": 0.2969440817832947,
    "learning_rate": 0.0004264405440806826,
    "timestep_mean": 475.59375,
    "epsilon_norm": 9.470857620239258,
    "prediction_norm": 7.697181224822998,
    "timestamp": "2025-07-14T12:35:18.343102"
  },
  {
    "epoch": 2906,
    "batch_idx": 96,
    "loss": 0.3491917848587036,
    "learning_rate": 0.0004264405440806826,
    "timestep_mean": 460.96875,
    "epsilon_norm": 7.0056471824646,
    "prediction_norm": 5.203803539276123,
    "timestamp": "2025-07-14T12:35:18.348103"
  },
  {
    "epoch": 2906,
    "batch_idx": 128,
    "loss": 0.5242255926132202,
    "learning_rate": 0.0004264405440806826,
    "timestep_mean": 499.28125,
    "epsilon_norm": 8.756668090820312,
    "prediction_norm": 6.722695827484131,
    "timestamp": "2025-07-14T12:35:18.353340"
  },
  {
    "epoch": 2906,
    "batch_idx": 160,
    "loss": 0.33833733201026917,
    "learning_rate": 0.0004264405440806826,
    "timestep_mean": 506.3125,
    "epsilon_norm": 7.312482833862305,
    "prediction_norm": 5.985660552978516,
    "timestamp": "2025-07-14T12:35:18.358410"
  },
  {
    "epoch": 2906,
    "batch_idx": 192,
    "loss": 0.41144245862960815,
    "learning_rate": 0.0004264405440806826,
    "timestep_mean": 380.625,
    "epsilon_norm": 4.583573818206787,
    "prediction_norm": 3.6412646770477295,
    "timestamp": "2025-07-14T12:35:18.363767"
  },
  {
    "epoch": 2907,
    "batch_idx": 0,
    "loss": 0.4369593560695648,
    "learning_rate": 0.00042721748661047053,
    "timestep_mean": 509.90625,
    "epsilon_norm": 8.682124137878418,
    "prediction_norm": 6.2407073974609375,
    "timestamp": "2025-07-14T12:35:18.368792"
  },
  {
    "epoch": 2907,
    "batch_idx": 32,
    "loss": 0.1840338259935379,
    "learning_rate": 0.00042721748661047053,
    "timestep_mean": 577.28125,
    "epsilon_norm": 7.823132038116455,
    "prediction_norm": 6.977174758911133,
    "timestamp": "2025-07-14T12:35:18.373959"
  },
  {
    "epoch": 2907,
    "batch_idx": 64,
    "loss": 0.30585411190986633,
    "learning_rate": 0.00042721748661047053,
    "timestep_mean": 508.3125,
    "epsilon_norm": 7.583261966705322,
    "prediction_norm": 6.66485595703125,
    "timestamp": "2025-07-14T12:35:18.379193"
  },
  {
    "epoch": 2907,
    "batch_idx": 96,
    "loss": 0.5565590858459473,
    "learning_rate": 0.00042721748661047053,
    "timestep_mean": 466.9375,
    "epsilon_norm": 7.9948954582214355,
    "prediction_norm": 6.6110005378723145,
    "timestamp": "2025-07-14T12:35:18.384345"
  },
  {
    "epoch": 2907,
    "batch_idx": 128,
    "loss": 0.12395921349525452,
    "learning_rate": 0.00042721748661047053,
    "timestep_mean": 507.28125,
    "epsilon_norm": 6.101902008056641,
    "prediction_norm": 6.036503314971924,
    "timestamp": "2025-07-14T12:35:18.389317"
  },
  {
    "epoch": 2907,
    "batch_idx": 160,
    "loss": 0.5300198197364807,
    "learning_rate": 0.00042721748661047053,
    "timestep_mean": 522.3125,
    "epsilon_norm": 9.153253555297852,
    "prediction_norm": 7.1074981689453125,
    "timestamp": "2025-07-14T12:35:18.394488"
  },
  {
    "epoch": 2907,
    "batch_idx": 192,
    "loss": 0.5396448969841003,
    "learning_rate": 0.00042721748661047053,
    "timestep_mean": 451.0,
    "epsilon_norm": 4.2672648429870605,
    "prediction_norm": 4.1316304206848145,
    "timestamp": "2025-07-14T12:35:18.399629"
  },
  {
    "epoch": 2908,
    "batch_idx": 0,
    "loss": 0.12658119201660156,
    "learning_rate": 0.00042799460872387513,
    "timestep_mean": 603.09375,
    "epsilon_norm": 7.750802993774414,
    "prediction_norm": 7.383055210113525,
    "timestamp": "2025-07-14T12:35:18.404793"
  },
  {
    "epoch": 2908,
    "batch_idx": 32,
    "loss": 0.18890869617462158,
    "learning_rate": 0.00042799460872387513,
    "timestep_mean": 526.0625,
    "epsilon_norm": 8.09368896484375,
    "prediction_norm": 7.363936424255371,
    "timestamp": "2025-07-14T12:35:18.409701"
  },
  {
    "epoch": 2908,
    "batch_idx": 64,
    "loss": 0.21552833914756775,
    "learning_rate": 0.00042799460872387513,
    "timestep_mean": 478.78125,
    "epsilon_norm": 6.891531944274902,
    "prediction_norm": 5.5972113609313965,
    "timestamp": "2025-07-14T12:35:18.414558"
  },
  {
    "epoch": 2908,
    "batch_idx": 96,
    "loss": 0.21280714869499207,
    "learning_rate": 0.00042799460872387513,
    "timestep_mean": 473.34375,
    "epsilon_norm": 5.6854658126831055,
    "prediction_norm": 5.2907795906066895,
    "timestamp": "2025-07-14T12:35:18.419788"
  },
  {
    "epoch": 2908,
    "batch_idx": 128,
    "loss": 0.3310874402523041,
    "learning_rate": 0.00042799460872387513,
    "timestep_mean": 546.3125,
    "epsilon_norm": 8.437607765197754,
    "prediction_norm": 7.8558430671691895,
    "timestamp": "2025-07-14T12:35:18.424976"
  },
  {
    "epoch": 2908,
    "batch_idx": 160,
    "loss": 0.41167667508125305,
    "learning_rate": 0.00042799460872387513,
    "timestep_mean": 440.28125,
    "epsilon_norm": 8.01163101196289,
    "prediction_norm": 5.635100841522217,
    "timestamp": "2025-07-14T12:35:18.430212"
  },
  {
    "epoch": 2908,
    "batch_idx": 192,
    "loss": 0.09702407568693161,
    "learning_rate": 0.00042799460872387513,
    "timestep_mean": 350.0,
    "epsilon_norm": 4.149655818939209,
    "prediction_norm": 3.5607192516326904,
    "timestamp": "2025-07-14T12:35:18.435234"
  },
  {
    "epoch": 2909,
    "batch_idx": 0,
    "loss": 0.13875791430473328,
    "learning_rate": 0.00042877190850342493,
    "timestep_mean": 608.3125,
    "epsilon_norm": 7.508266925811768,
    "prediction_norm": 7.2551960945129395,
    "timestamp": "2025-07-14T12:35:18.440313"
  },
  {
    "epoch": 2909,
    "batch_idx": 32,
    "loss": 0.21265923976898193,
    "learning_rate": 0.00042877190850342493,
    "timestep_mean": 612.0,
    "epsilon_norm": 7.513050079345703,
    "prediction_norm": 6.724576950073242,
    "timestamp": "2025-07-14T12:35:18.445699"
  },
  {
    "epoch": 2909,
    "batch_idx": 64,
    "loss": 0.6031496524810791,
    "learning_rate": 0.00042877190850342493,
    "timestep_mean": 448.1875,
    "epsilon_norm": 7.392210483551025,
    "prediction_norm": 5.329018592834473,
    "timestamp": "2025-07-14T12:35:18.450893"
  },
  {
    "epoch": 2909,
    "batch_idx": 96,
    "loss": 0.25923001766204834,
    "learning_rate": 0.00042877190850342493,
    "timestep_mean": 448.15625,
    "epsilon_norm": 6.6238203048706055,
    "prediction_norm": 4.9418535232543945,
    "timestamp": "2025-07-14T12:35:18.455829"
  },
  {
    "epoch": 2909,
    "batch_idx": 128,
    "loss": 0.20301736891269684,
    "learning_rate": 0.00042877190850342493,
    "timestep_mean": 481.0,
    "epsilon_norm": 7.758335590362549,
    "prediction_norm": 6.497288703918457,
    "timestamp": "2025-07-14T12:35:18.460987"
  },
  {
    "epoch": 2909,
    "batch_idx": 160,
    "loss": 0.17790381610393524,
    "learning_rate": 0.00042877190850342493,
    "timestep_mean": 419.75,
    "epsilon_norm": 7.931620121002197,
    "prediction_norm": 6.766741752624512,
    "timestamp": "2025-07-14T12:35:18.466134"
  },
  {
    "epoch": 2909,
    "batch_idx": 192,
    "loss": 0.1044163927435875,
    "learning_rate": 0.00042877190850342493,
    "timestep_mean": 500.5,
    "epsilon_norm": 3.1877734661102295,
    "prediction_norm": 2.921140432357788,
    "timestamp": "2025-07-14T12:35:18.471442"
  },
  {
    "epoch": 2910,
    "batch_idx": 0,
    "loss": 0.38949939608573914,
    "learning_rate": 0.00042954938403120987,
    "timestep_mean": 468.65625,
    "epsilon_norm": 7.97959041595459,
    "prediction_norm": 5.535278797149658,
    "timestamp": "2025-07-14T12:35:18.476574"
  },
  {
    "epoch": 2910,
    "batch_idx": 32,
    "loss": 0.23451931774616241,
    "learning_rate": 0.00042954938403120987,
    "timestep_mean": 483.6875,
    "epsilon_norm": 7.405594825744629,
    "prediction_norm": 6.27996301651001,
    "timestamp": "2025-07-14T12:35:18.481710"
  },
  {
    "epoch": 2910,
    "batch_idx": 64,
    "loss": 0.24116602540016174,
    "learning_rate": 0.00042954938403120987,
    "timestep_mean": 534.25,
    "epsilon_norm": 8.266548156738281,
    "prediction_norm": 7.117774963378906,
    "timestamp": "2025-07-14T12:35:18.486924"
  },
  {
    "epoch": 2910,
    "batch_idx": 96,
    "loss": 0.17439857125282288,
    "learning_rate": 0.00042954938403120987,
    "timestep_mean": 518.375,
    "epsilon_norm": 7.655796527862549,
    "prediction_norm": 6.649548053741455,
    "timestamp": "2025-07-14T12:35:18.492085"
  },
  {
    "epoch": 2910,
    "batch_idx": 128,
    "loss": 0.3051905930042267,
    "learning_rate": 0.00042954938403120987,
    "timestep_mean": 451.78125,
    "epsilon_norm": 9.420562744140625,
    "prediction_norm": 8.091056823730469,
    "timestamp": "2025-07-14T12:35:18.497249"
  },
  {
    "epoch": 2910,
    "batch_idx": 160,
    "loss": 0.42091262340545654,
    "learning_rate": 0.00042954938403120987,
    "timestep_mean": 518.3125,
    "epsilon_norm": 9.483242988586426,
    "prediction_norm": 7.994411945343018,
    "timestamp": "2025-07-14T12:35:18.502412"
  },
  {
    "epoch": 2910,
    "batch_idx": 192,
    "loss": 0.4255707859992981,
    "learning_rate": 0.00042954938403120987,
    "timestep_mean": 590.125,
    "epsilon_norm": 4.928652286529541,
    "prediction_norm": 4.342209339141846,
    "timestamp": "2025-07-14T12:35:18.507470"
  },
  {
    "epoch": 2911,
    "batch_idx": 0,
    "loss": 0.3321763873100281,
    "learning_rate": 0.0004303270333888863,
    "timestep_mean": 466.46875,
    "epsilon_norm": 9.328003883361816,
    "prediction_norm": 7.859984397888184,
    "timestamp": "2025-07-14T12:35:18.512702"
  },
  {
    "epoch": 2911,
    "batch_idx": 32,
    "loss": 0.23799873888492584,
    "learning_rate": 0.0004303270333888863,
    "timestep_mean": 463.09375,
    "epsilon_norm": 7.851614475250244,
    "prediction_norm": 6.7516679763793945,
    "timestamp": "2025-07-14T12:35:18.517904"
  },
  {
    "epoch": 2911,
    "batch_idx": 64,
    "loss": 0.3372873365879059,
    "learning_rate": 0.0004303270333888863,
    "timestep_mean": 512.78125,
    "epsilon_norm": 8.397107124328613,
    "prediction_norm": 7.587241172790527,
    "timestamp": "2025-07-14T12:35:18.522933"
  },
  {
    "epoch": 2911,
    "batch_idx": 96,
    "loss": 0.48742491006851196,
    "learning_rate": 0.0004303270333888863,
    "timestep_mean": 352.0625,
    "epsilon_norm": 7.667863845825195,
    "prediction_norm": 6.2128376960754395,
    "timestamp": "2025-07-14T12:35:18.527970"
  },
  {
    "epoch": 2911,
    "batch_idx": 128,
    "loss": 0.11145702004432678,
    "learning_rate": 0.0004303270333888863,
    "timestep_mean": 562.90625,
    "epsilon_norm": 6.7554030418396,
    "prediction_norm": 6.443031311035156,
    "timestamp": "2025-07-14T12:35:18.532823"
  },
  {
    "epoch": 2911,
    "batch_idx": 160,
    "loss": 0.34577232599258423,
    "learning_rate": 0.0004303270333888863,
    "timestep_mean": 502.78125,
    "epsilon_norm": 8.370388984680176,
    "prediction_norm": 7.0625481605529785,
    "timestamp": "2025-07-14T12:35:18.537662"
  },
  {
    "epoch": 2911,
    "batch_idx": 192,
    "loss": 0.20961646735668182,
    "learning_rate": 0.0004303270333888863,
    "timestep_mean": 441.375,
    "epsilon_norm": 4.1848883628845215,
    "prediction_norm": 4.115463733673096,
    "timestamp": "2025-07-14T12:35:18.542613"
  },
  {
    "epoch": 2912,
    "batch_idx": 0,
    "loss": 0.3994547128677368,
    "learning_rate": 0.000431104854657682,
    "timestep_mean": 548.09375,
    "epsilon_norm": 7.676321029663086,
    "prediction_norm": 6.659666061401367,
    "timestamp": "2025-07-14T12:35:18.547484"
  },
  {
    "epoch": 2912,
    "batch_idx": 32,
    "loss": 0.26258188486099243,
    "learning_rate": 0.000431104854657682,
    "timestep_mean": 534.25,
    "epsilon_norm": 8.507407188415527,
    "prediction_norm": 7.883478164672852,
    "timestamp": "2025-07-14T12:35:18.552324"
  },
  {
    "epoch": 2912,
    "batch_idx": 64,
    "loss": 0.2797207832336426,
    "learning_rate": 0.000431104854657682,
    "timestep_mean": 469.875,
    "epsilon_norm": 8.071133613586426,
    "prediction_norm": 7.048168659210205,
    "timestamp": "2025-07-14T12:35:18.557202"
  },
  {
    "epoch": 2912,
    "batch_idx": 96,
    "loss": 0.4182586073875427,
    "learning_rate": 0.000431104854657682,
    "timestep_mean": 484.0625,
    "epsilon_norm": 8.149126052856445,
    "prediction_norm": 7.4386820793151855,
    "timestamp": "2025-07-14T12:35:18.562075"
  },
  {
    "epoch": 2912,
    "batch_idx": 128,
    "loss": 0.2465246617794037,
    "learning_rate": 0.000431104854657682,
    "timestep_mean": 443.15625,
    "epsilon_norm": 8.820902824401855,
    "prediction_norm": 7.865447998046875,
    "timestamp": "2025-07-14T12:35:18.566971"
  },
  {
    "epoch": 2912,
    "batch_idx": 160,
    "loss": 0.32786762714385986,
    "learning_rate": 0.000431104854657682,
    "timestep_mean": 447.6875,
    "epsilon_norm": 7.947319507598877,
    "prediction_norm": 6.687983512878418,
    "timestamp": "2025-07-14T12:35:18.572013"
  },
  {
    "epoch": 2912,
    "batch_idx": 192,
    "loss": 0.38594773411750793,
    "learning_rate": 0.000431104854657682,
    "timestep_mean": 357.0,
    "epsilon_norm": 4.458343982696533,
    "prediction_norm": 3.248262643814087,
    "timestamp": "2025-07-14T12:35:18.576978"
  },
  {
    "epoch": 2913,
    "batch_idx": 0,
    "loss": 0.46524301171302795,
    "learning_rate": 0.0004318828459184006,
    "timestep_mean": 485.0,
    "epsilon_norm": 7.766497611999512,
    "prediction_norm": 6.105404376983643,
    "timestamp": "2025-07-14T12:35:18.582092"
  },
  {
    "epoch": 2913,
    "batch_idx": 32,
    "loss": 0.13478302955627441,
    "learning_rate": 0.0004318828459184006,
    "timestep_mean": 577.15625,
    "epsilon_norm": 7.779409408569336,
    "prediction_norm": 7.377547740936279,
    "timestamp": "2025-07-14T12:35:18.587302"
  },
  {
    "epoch": 2913,
    "batch_idx": 64,
    "loss": 0.2514982223510742,
    "learning_rate": 0.0004318828459184006,
    "timestep_mean": 430.40625,
    "epsilon_norm": 7.3798370361328125,
    "prediction_norm": 6.346015453338623,
    "timestamp": "2025-07-14T12:35:18.592503"
  },
  {
    "epoch": 2913,
    "batch_idx": 96,
    "loss": 0.2191849946975708,
    "learning_rate": 0.0004318828459184006,
    "timestep_mean": 436.625,
    "epsilon_norm": 7.161368370056152,
    "prediction_norm": 5.874335289001465,
    "timestamp": "2025-07-14T12:35:18.597558"
  },
  {
    "epoch": 2913,
    "batch_idx": 128,
    "loss": 0.3857802748680115,
    "learning_rate": 0.0004318828459184006,
    "timestep_mean": 502.40625,
    "epsilon_norm": 7.313551425933838,
    "prediction_norm": 5.333539962768555,
    "timestamp": "2025-07-14T12:35:18.602622"
  },
  {
    "epoch": 2913,
    "batch_idx": 160,
    "loss": 0.19899487495422363,
    "learning_rate": 0.0004318828459184006,
    "timestep_mean": 520.375,
    "epsilon_norm": 7.733447551727295,
    "prediction_norm": 6.251636505126953,
    "timestamp": "2025-07-14T12:35:18.607580"
  },
  {
    "epoch": 2913,
    "batch_idx": 192,
    "loss": 0.426921546459198,
    "learning_rate": 0.0004318828459184006,
    "timestep_mean": 464.875,
    "epsilon_norm": 3.9555861949920654,
    "prediction_norm": 2.445723295211792,
    "timestamp": "2025-07-14T12:35:18.612887"
  },
  {
    "epoch": 2914,
    "batch_idx": 0,
    "loss": 0.2853725850582123,
    "learning_rate": 0.0004326610052514251,
    "timestep_mean": 495.625,
    "epsilon_norm": 7.3642778396606445,
    "prediction_norm": 5.76603364944458,
    "timestamp": "2025-07-14T12:35:18.618012"
  },
  {
    "epoch": 2914,
    "batch_idx": 32,
    "loss": 0.14130261540412903,
    "learning_rate": 0.0004326610052514251,
    "timestep_mean": 491.84375,
    "epsilon_norm": 7.484107971191406,
    "prediction_norm": 6.719111442565918,
    "timestamp": "2025-07-14T12:35:18.623002"
  },
  {
    "epoch": 2914,
    "batch_idx": 64,
    "loss": 0.2843016982078552,
    "learning_rate": 0.0004326610052514251,
    "timestep_mean": 514.9375,
    "epsilon_norm": 7.923332214355469,
    "prediction_norm": 6.751196384429932,
    "timestamp": "2025-07-14T12:35:18.628096"
  },
  {
    "epoch": 2914,
    "batch_idx": 96,
    "loss": 0.29432815313339233,
    "learning_rate": 0.0004326610052514251,
    "timestep_mean": 521.625,
    "epsilon_norm": 8.398868560791016,
    "prediction_norm": 7.293457984924316,
    "timestamp": "2025-07-14T12:35:18.633171"
  },
  {
    "epoch": 2914,
    "batch_idx": 128,
    "loss": 0.1263696253299713,
    "learning_rate": 0.0004326610052514251,
    "timestep_mean": 584.625,
    "epsilon_norm": 7.006469249725342,
    "prediction_norm": 5.892293453216553,
    "timestamp": "2025-07-14T12:35:18.638359"
  },
  {
    "epoch": 2914,
    "batch_idx": 160,
    "loss": 0.2487315535545349,
    "learning_rate": 0.0004326610052514251,
    "timestep_mean": 510.3125,
    "epsilon_norm": 7.9091339111328125,
    "prediction_norm": 6.158591270446777,
    "timestamp": "2025-07-14T12:35:18.643431"
  },
  {
    "epoch": 2914,
    "batch_idx": 192,
    "loss": 0.43463990092277527,
    "learning_rate": 0.0004326610052514251,
    "timestep_mean": 534.375,
    "epsilon_norm": 4.40601921081543,
    "prediction_norm": 3.7454071044921875,
    "timestamp": "2025-07-14T12:35:18.648358"
  },
  {
    "epoch": 2915,
    "batch_idx": 0,
    "loss": 0.181193009018898,
    "learning_rate": 0.00043343933073672515,
    "timestep_mean": 564.0,
    "epsilon_norm": 7.242203235626221,
    "prediction_norm": 6.623668670654297,
    "timestamp": "2025-07-14T12:35:18.653578"
  },
  {
    "epoch": 2915,
    "batch_idx": 32,
    "loss": 0.1795438528060913,
    "learning_rate": 0.00043343933073672515,
    "timestep_mean": 550.4375,
    "epsilon_norm": 6.587134838104248,
    "prediction_norm": 5.799042701721191,
    "timestamp": "2025-07-14T12:35:18.658717"
  },
  {
    "epoch": 2915,
    "batch_idx": 64,
    "loss": 0.35368984937667847,
    "learning_rate": 0.00043343933073672515,
    "timestep_mean": 596.0,
    "epsilon_norm": 8.503026962280273,
    "prediction_norm": 6.752050399780273,
    "timestamp": "2025-07-14T12:35:18.663825"
  },
  {
    "epoch": 2915,
    "batch_idx": 96,
    "loss": 0.3873986005783081,
    "learning_rate": 0.00043343933073672515,
    "timestep_mean": 408.40625,
    "epsilon_norm": 7.756714820861816,
    "prediction_norm": 6.801703929901123,
    "timestamp": "2025-07-14T12:35:18.668956"
  },
  {
    "epoch": 2915,
    "batch_idx": 128,
    "loss": 0.3127622902393341,
    "learning_rate": 0.00043343933073672515,
    "timestep_mean": 413.84375,
    "epsilon_norm": 7.249455451965332,
    "prediction_norm": 6.067992687225342,
    "timestamp": "2025-07-14T12:35:18.673961"
  },
  {
    "epoch": 2915,
    "batch_idx": 160,
    "loss": 0.3656221628189087,
    "learning_rate": 0.00043343933073672515,
    "timestep_mean": 507.5,
    "epsilon_norm": 8.650517463684082,
    "prediction_norm": 6.9854888916015625,
    "timestamp": "2025-07-14T12:35:18.679208"
  },
  {
    "epoch": 2915,
    "batch_idx": 192,
    "loss": 0.3572884798049927,
    "learning_rate": 0.00043343933073672515,
    "timestep_mean": 356.0,
    "epsilon_norm": 3.503541946411133,
    "prediction_norm": 2.4554336071014404,
    "timestamp": "2025-07-14T12:35:18.684270"
  },
  {
    "epoch": 2916,
    "batch_idx": 0,
    "loss": 0.4207212030887604,
    "learning_rate": 0.00043421782045385956,
    "timestep_mean": 445.6875,
    "epsilon_norm": 7.326844215393066,
    "prediction_norm": 5.694051742553711,
    "timestamp": "2025-07-14T12:35:18.689311"
  },
  {
    "epoch": 2916,
    "batch_idx": 32,
    "loss": 0.1820063591003418,
    "learning_rate": 0.00043421782045385956,
    "timestep_mean": 481.84375,
    "epsilon_norm": 7.487334728240967,
    "prediction_norm": 6.372810363769531,
    "timestamp": "2025-07-14T12:35:18.694443"
  },
  {
    "epoch": 2916,
    "batch_idx": 64,
    "loss": 0.1283349245786667,
    "learning_rate": 0.00043421782045385956,
    "timestep_mean": 536.40625,
    "epsilon_norm": 6.701922416687012,
    "prediction_norm": 6.444856643676758,
    "timestamp": "2025-07-14T12:35:18.699523"
  },
  {
    "epoch": 2916,
    "batch_idx": 96,
    "loss": 0.2812545895576477,
    "learning_rate": 0.00043421782045385956,
    "timestep_mean": 536.84375,
    "epsilon_norm": 8.627889633178711,
    "prediction_norm": 7.98626184463501,
    "timestamp": "2025-07-14T12:35:18.704861"
  },
  {
    "epoch": 2916,
    "batch_idx": 128,
    "loss": 0.19016940891742706,
    "learning_rate": 0.00043421782045385956,
    "timestep_mean": 502.375,
    "epsilon_norm": 7.913360595703125,
    "prediction_norm": 7.634756565093994,
    "timestamp": "2025-07-14T12:35:18.709919"
  },
  {
    "epoch": 2916,
    "batch_idx": 160,
    "loss": 0.2134525626897812,
    "learning_rate": 0.00043421782045385956,
    "timestep_mean": 526.375,
    "epsilon_norm": 8.67910099029541,
    "prediction_norm": 7.660254955291748,
    "timestamp": "2025-07-14T12:35:18.714960"
  },
  {
    "epoch": 2916,
    "batch_idx": 192,
    "loss": 0.3048272132873535,
    "learning_rate": 0.00043421782045385956,
    "timestep_mean": 380.25,
    "epsilon_norm": 3.848796844482422,
    "prediction_norm": 2.6542844772338867,
    "timestamp": "2025-07-14T12:35:18.719961"
  },
  {
    "epoch": 2917,
    "batch_idx": 0,
    "loss": 0.5139513611793518,
    "learning_rate": 0.0004349964724819837,
    "timestep_mean": 532.21875,
    "epsilon_norm": 8.622626304626465,
    "prediction_norm": 7.47634220123291,
    "timestamp": "2025-07-14T12:35:18.725098"
  },
  {
    "epoch": 2917,
    "batch_idx": 32,
    "loss": 0.27983400225639343,
    "learning_rate": 0.0004349964724819837,
    "timestep_mean": 422.21875,
    "epsilon_norm": 9.06634521484375,
    "prediction_norm": 7.747949600219727,
    "timestamp": "2025-07-14T12:35:18.730206"
  },
  {
    "epoch": 2917,
    "batch_idx": 64,
    "loss": 0.2850090265274048,
    "learning_rate": 0.0004349964724819837,
    "timestep_mean": 445.25,
    "epsilon_norm": 6.53227424621582,
    "prediction_norm": 5.2992329597473145,
    "timestamp": "2025-07-14T12:35:18.735070"
  },
  {
    "epoch": 2917,
    "batch_idx": 96,
    "loss": 0.31491976976394653,
    "learning_rate": 0.0004349964724819837,
    "timestep_mean": 482.90625,
    "epsilon_norm": 8.077369689941406,
    "prediction_norm": 7.117646217346191,
    "timestamp": "2025-07-14T12:35:18.739956"
  },
  {
    "epoch": 2917,
    "batch_idx": 128,
    "loss": 0.4322459101676941,
    "learning_rate": 0.0004349964724819837,
    "timestep_mean": 520.6875,
    "epsilon_norm": 8.946938514709473,
    "prediction_norm": 6.9145965576171875,
    "timestamp": "2025-07-14T12:35:18.744792"
  },
  {
    "epoch": 2917,
    "batch_idx": 160,
    "loss": 0.3804992139339447,
    "learning_rate": 0.0004349964724819837,
    "timestep_mean": 408.6875,
    "epsilon_norm": 8.622693061828613,
    "prediction_norm": 7.077010631561279,
    "timestamp": "2025-07-14T12:35:18.749634"
  },
  {
    "epoch": 2917,
    "batch_idx": 192,
    "loss": 0.33617687225341797,
    "learning_rate": 0.0004349964724819837,
    "timestep_mean": 276.375,
    "epsilon_norm": 3.5178909301757812,
    "prediction_norm": 2.489515542984009,
    "timestamp": "2025-07-14T12:35:18.754897"
  },
  {
    "epoch": 2918,
    "batch_idx": 0,
    "loss": 0.3472098410129547,
    "learning_rate": 0.0004357752848998497,
    "timestep_mean": 398.53125,
    "epsilon_norm": 8.290102005004883,
    "prediction_norm": 7.0205979347229,
    "timestamp": "2025-07-14T12:35:18.760015"
  },
  {
    "epoch": 2918,
    "batch_idx": 32,
    "loss": 0.30025744438171387,
    "learning_rate": 0.0004357752848998497,
    "timestep_mean": 461.8125,
    "epsilon_norm": 8.080190658569336,
    "prediction_norm": 6.174733638763428,
    "timestamp": "2025-07-14T12:35:18.765074"
  },
  {
    "epoch": 2918,
    "batch_idx": 64,
    "loss": 0.2562529742717743,
    "learning_rate": 0.0004357752848998497,
    "timestep_mean": 470.46875,
    "epsilon_norm": 7.6820478439331055,
    "prediction_norm": 6.060751914978027,
    "timestamp": "2025-07-14T12:35:18.770061"
  },
  {
    "epoch": 2918,
    "batch_idx": 96,
    "loss": 0.151725172996521,
    "learning_rate": 0.0004357752848998497,
    "timestep_mean": 529.96875,
    "epsilon_norm": 7.530941963195801,
    "prediction_norm": 6.480020999908447,
    "timestamp": "2025-07-14T12:35:18.775192"
  },
  {
    "epoch": 2918,
    "batch_idx": 128,
    "loss": 0.36596575379371643,
    "learning_rate": 0.0004357752848998497,
    "timestep_mean": 576.71875,
    "epsilon_norm": 8.398897171020508,
    "prediction_norm": 6.790395736694336,
    "timestamp": "2025-07-14T12:35:18.780338"
  },
  {
    "epoch": 2918,
    "batch_idx": 160,
    "loss": 0.15370187163352966,
    "learning_rate": 0.0004357752848998497,
    "timestep_mean": 552.71875,
    "epsilon_norm": 8.814689636230469,
    "prediction_norm": 7.938727855682373,
    "timestamp": "2025-07-14T12:35:18.785281"
  },
  {
    "epoch": 2918,
    "batch_idx": 192,
    "loss": 0.47273463010787964,
    "learning_rate": 0.0004357752848998497,
    "timestep_mean": 393.25,
    "epsilon_norm": 3.347499370574951,
    "prediction_norm": 2.0335958003997803,
    "timestamp": "2025-07-14T12:35:18.790173"
  },
  {
    "epoch": 2919,
    "batch_idx": 0,
    "loss": 0.23371000587940216,
    "learning_rate": 0.00043655425578581584,
    "timestep_mean": 494.46875,
    "epsilon_norm": 7.32125997543335,
    "prediction_norm": 6.647140979766846,
    "timestamp": "2025-07-14T12:35:18.795073"
  },
  {
    "epoch": 2919,
    "batch_idx": 32,
    "loss": 0.20773594081401825,
    "learning_rate": 0.00043655425578581584,
    "timestep_mean": 509.8125,
    "epsilon_norm": 7.239078521728516,
    "prediction_norm": 6.7787861824035645,
    "timestamp": "2025-07-14T12:35:18.799988"
  },
  {
    "epoch": 2919,
    "batch_idx": 64,
    "loss": 0.31445538997650146,
    "learning_rate": 0.00043655425578581584,
    "timestep_mean": 535.75,
    "epsilon_norm": 8.84007740020752,
    "prediction_norm": 6.681235313415527,
    "timestamp": "2025-07-14T12:35:18.804964"
  },
  {
    "epoch": 2919,
    "batch_idx": 96,
    "loss": 0.3005274832248688,
    "learning_rate": 0.00043655425578581584,
    "timestep_mean": 485.3125,
    "epsilon_norm": 6.313899517059326,
    "prediction_norm": 4.851322650909424,
    "timestamp": "2025-07-14T12:35:18.810061"
  },
  {
    "epoch": 2919,
    "batch_idx": 128,
    "loss": 0.3743075132369995,
    "learning_rate": 0.00043655425578581584,
    "timestep_mean": 498.8125,
    "epsilon_norm": 9.083246231079102,
    "prediction_norm": 7.64826774597168,
    "timestamp": "2025-07-14T12:35:18.815145"
  },
  {
    "epoch": 2919,
    "batch_idx": 160,
    "loss": 0.7186335325241089,
    "learning_rate": 0.00043655425578581584,
    "timestep_mean": 462.0,
    "epsilon_norm": 9.257606506347656,
    "prediction_norm": 6.4455790519714355,
    "timestamp": "2025-07-14T12:35:18.820115"
  },
  {
    "epoch": 2919,
    "batch_idx": 192,
    "loss": 0.4101034104824066,
    "learning_rate": 0.00043655425578581584,
    "timestep_mean": 249.25,
    "epsilon_norm": 3.992189884185791,
    "prediction_norm": 3.059359073638916,
    "timestamp": "2025-07-14T12:35:18.825271"
  },
  {
    "epoch": 2920,
    "batch_idx": 0,
    "loss": 0.31436657905578613,
    "learning_rate": 0.00043733338321784936,
    "timestep_mean": 450.4375,
    "epsilon_norm": 6.875187873840332,
    "prediction_norm": 5.661171913146973,
    "timestamp": "2025-07-14T12:35:18.830527"
  },
  {
    "epoch": 2920,
    "batch_idx": 32,
    "loss": 0.3945145606994629,
    "learning_rate": 0.00043733338321784936,
    "timestep_mean": 544.125,
    "epsilon_norm": 8.551074981689453,
    "prediction_norm": 6.88959264755249,
    "timestamp": "2025-07-14T12:35:18.835598"
  },
  {
    "epoch": 2920,
    "batch_idx": 64,
    "loss": 0.3299703598022461,
    "learning_rate": 0.00043733338321784936,
    "timestep_mean": 490.34375,
    "epsilon_norm": 9.0471773147583,
    "prediction_norm": 7.95961856842041,
    "timestamp": "2025-07-14T12:35:18.840870"
  },
  {
    "epoch": 2920,
    "batch_idx": 96,
    "loss": 0.16187015175819397,
    "learning_rate": 0.00043733338321784936,
    "timestep_mean": 489.5625,
    "epsilon_norm": 7.582394123077393,
    "prediction_norm": 7.3521552085876465,
    "timestamp": "2025-07-14T12:35:18.845897"
  },
  {
    "epoch": 2920,
    "batch_idx": 128,
    "loss": 0.23506148159503937,
    "learning_rate": 0.00043733338321784936,
    "timestep_mean": 541.9375,
    "epsilon_norm": 7.783771514892578,
    "prediction_norm": 7.271503925323486,
    "timestamp": "2025-07-14T12:35:18.851317"
  },
  {
    "epoch": 2920,
    "batch_idx": 160,
    "loss": 0.4651544392108917,
    "learning_rate": 0.00043733338321784936,
    "timestep_mean": 483.5,
    "epsilon_norm": 8.353391647338867,
    "prediction_norm": 6.389970779418945,
    "timestamp": "2025-07-14T12:35:18.856565"
  },
  {
    "epoch": 2920,
    "batch_idx": 192,
    "loss": 0.38015609979629517,
    "learning_rate": 0.00043733338321784936,
    "timestep_mean": 400.125,
    "epsilon_norm": 3.4426021575927734,
    "prediction_norm": 2.2010397911071777,
    "timestamp": "2025-07-14T12:35:18.861762"
  },
  {
    "epoch": 2921,
    "batch_idx": 0,
    "loss": 0.2876288890838623,
    "learning_rate": 0.0004381126652735294,
    "timestep_mean": 471.4375,
    "epsilon_norm": 9.05176067352295,
    "prediction_norm": 8.03453540802002,
    "timestamp": "2025-07-14T12:35:18.867000"
  },
  {
    "epoch": 2921,
    "batch_idx": 32,
    "loss": 0.22469396889209747,
    "learning_rate": 0.0004381126652735294,
    "timestep_mean": 482.5,
    "epsilon_norm": 7.443661689758301,
    "prediction_norm": 6.370287895202637,
    "timestamp": "2025-07-14T12:35:18.872230"
  },
  {
    "epoch": 2921,
    "batch_idx": 64,
    "loss": 0.2006150186061859,
    "learning_rate": 0.0004381126652735294,
    "timestep_mean": 503.40625,
    "epsilon_norm": 8.859478950500488,
    "prediction_norm": 7.91308069229126,
    "timestamp": "2025-07-14T12:35:18.877364"
  },
  {
    "epoch": 2921,
    "batch_idx": 96,
    "loss": 0.139887273311615,
    "learning_rate": 0.0004381126652735294,
    "timestep_mean": 508.125,
    "epsilon_norm": 7.306972503662109,
    "prediction_norm": 7.2467041015625,
    "timestamp": "2025-07-14T12:35:18.882430"
  },
  {
    "epoch": 2921,
    "batch_idx": 128,
    "loss": 0.22607412934303284,
    "learning_rate": 0.0004381126652735294,
    "timestep_mean": 488.625,
    "epsilon_norm": 7.537539482116699,
    "prediction_norm": 6.320183753967285,
    "timestamp": "2025-07-14T12:35:18.887493"
  },
  {
    "epoch": 2921,
    "batch_idx": 160,
    "loss": 0.28773993253707886,
    "learning_rate": 0.0004381126652735294,
    "timestep_mean": 473.625,
    "epsilon_norm": 7.667881488800049,
    "prediction_norm": 5.991764068603516,
    "timestamp": "2025-07-14T12:35:18.892697"
  },
  {
    "epoch": 2921,
    "batch_idx": 192,
    "loss": 0.1666325330734253,
    "learning_rate": 0.0004381126652735294,
    "timestep_mean": 604.25,
    "epsilon_norm": 3.940222978591919,
    "prediction_norm": 2.857959270477295,
    "timestamp": "2025-07-14T12:35:18.898110"
  },
  {
    "epoch": 2922,
    "batch_idx": 0,
    "loss": 0.22755815088748932,
    "learning_rate": 0.00043889210003005626,
    "timestep_mean": 499.125,
    "epsilon_norm": 7.748147010803223,
    "prediction_norm": 7.14603328704834,
    "timestamp": "2025-07-14T12:35:18.903336"
  },
  {
    "epoch": 2922,
    "batch_idx": 32,
    "loss": 0.28132736682891846,
    "learning_rate": 0.00043889210003005626,
    "timestep_mean": 461.4375,
    "epsilon_norm": 8.137939453125,
    "prediction_norm": 6.651984214782715,
    "timestamp": "2025-07-14T12:35:18.908651"
  },
  {
    "epoch": 2922,
    "batch_idx": 64,
    "loss": 0.3405567407608032,
    "learning_rate": 0.00043889210003005626,
    "timestep_mean": 383.34375,
    "epsilon_norm": 7.645474433898926,
    "prediction_norm": 6.3825531005859375,
    "timestamp": "2025-07-14T12:35:18.913664"
  },
  {
    "epoch": 2922,
    "batch_idx": 96,
    "loss": 0.29686135053634644,
    "learning_rate": 0.00043889210003005626,
    "timestep_mean": 484.84375,
    "epsilon_norm": 8.53649616241455,
    "prediction_norm": 7.0682830810546875,
    "timestamp": "2025-07-14T12:35:18.918766"
  },
  {
    "epoch": 2922,
    "batch_idx": 128,
    "loss": 0.4261552095413208,
    "learning_rate": 0.00043889210003005626,
    "timestep_mean": 523.0625,
    "epsilon_norm": 8.040149688720703,
    "prediction_norm": 6.20629358291626,
    "timestamp": "2025-07-14T12:35:18.924014"
  },
  {
    "epoch": 2922,
    "batch_idx": 160,
    "loss": 0.19741421937942505,
    "learning_rate": 0.00043889210003005626,
    "timestep_mean": 574.53125,
    "epsilon_norm": 8.5228853225708,
    "prediction_norm": 8.298259735107422,
    "timestamp": "2025-07-14T12:35:18.929103"
  },
  {
    "epoch": 2922,
    "batch_idx": 192,
    "loss": 0.14342331886291504,
    "learning_rate": 0.00043889210003005626,
    "timestep_mean": 499.375,
    "epsilon_norm": 3.348834276199341,
    "prediction_norm": 3.054044723510742,
    "timestamp": "2025-07-14T12:35:18.934229"
  },
  {
    "epoch": 2923,
    "batch_idx": 0,
    "loss": 0.2890079617500305,
    "learning_rate": 0.0004396716855642517,
    "timestep_mean": 539.21875,
    "epsilon_norm": 7.457131862640381,
    "prediction_norm": 6.8322248458862305,
    "timestamp": "2025-07-14T12:35:18.939446"
  },
  {
    "epoch": 2923,
    "batch_idx": 32,
    "loss": 0.2357179969549179,
    "learning_rate": 0.0004396716855642517,
    "timestep_mean": 514.75,
    "epsilon_norm": 7.8925065994262695,
    "prediction_norm": 7.168112754821777,
    "timestamp": "2025-07-14T12:35:18.944711"
  },
  {
    "epoch": 2923,
    "batch_idx": 64,
    "loss": 0.3204134702682495,
    "learning_rate": 0.0004396716855642517,
    "timestep_mean": 481.34375,
    "epsilon_norm": 7.389866828918457,
    "prediction_norm": 6.14035701751709,
    "timestamp": "2025-07-14T12:35:18.949665"
  },
  {
    "epoch": 2923,
    "batch_idx": 96,
    "loss": 0.2224154770374298,
    "learning_rate": 0.0004396716855642517,
    "timestep_mean": 507.28125,
    "epsilon_norm": 7.842660903930664,
    "prediction_norm": 7.153881549835205,
    "timestamp": "2025-07-14T12:35:18.954879"
  },
  {
    "epoch": 2923,
    "batch_idx": 128,
    "loss": 0.1833038181066513,
    "learning_rate": 0.0004396716855642517,
    "timestep_mean": 526.71875,
    "epsilon_norm": 7.459532737731934,
    "prediction_norm": 6.632122993469238,
    "timestamp": "2025-07-14T12:35:18.959960"
  },
  {
    "epoch": 2923,
    "batch_idx": 160,
    "loss": 0.4657382369041443,
    "learning_rate": 0.0004396716855642517,
    "timestep_mean": 366.25,
    "epsilon_norm": 8.946303367614746,
    "prediction_norm": 7.414240837097168,
    "timestamp": "2025-07-14T12:35:18.965108"
  },
  {
    "epoch": 2923,
    "batch_idx": 192,
    "loss": 0.19877983629703522,
    "learning_rate": 0.0004396716855642517,
    "timestep_mean": 605.0,
    "epsilon_norm": 2.9663476943969727,
    "prediction_norm": 2.585517168045044,
    "timestamp": "2025-07-14T12:35:18.970285"
  },
  {
    "epoch": 2924,
    "batch_idx": 0,
    "loss": 0.15995270013809204,
    "learning_rate": 0.0004404514199525664,
    "timestep_mean": 552.125,
    "epsilon_norm": 7.317228317260742,
    "prediction_norm": 7.503192901611328,
    "timestamp": "2025-07-14T12:35:18.975628"
  },
  {
    "epoch": 2924,
    "batch_idx": 32,
    "loss": 0.32216551899909973,
    "learning_rate": 0.0004404514199525664,
    "timestep_mean": 547.65625,
    "epsilon_norm": 7.622002124786377,
    "prediction_norm": 6.315568447113037,
    "timestamp": "2025-07-14T12:35:18.980771"
  },
  {
    "epoch": 2924,
    "batch_idx": 64,
    "loss": 0.2014485001564026,
    "learning_rate": 0.0004404514199525664,
    "timestep_mean": 553.0625,
    "epsilon_norm": 8.656588554382324,
    "prediction_norm": 8.282362937927246,
    "timestamp": "2025-07-14T12:35:18.985914"
  },
  {
    "epoch": 2924,
    "batch_idx": 96,
    "loss": 0.30628037452697754,
    "learning_rate": 0.0004404514199525664,
    "timestep_mean": 468.34375,
    "epsilon_norm": 7.163431167602539,
    "prediction_norm": 5.993596076965332,
    "timestamp": "2025-07-14T12:35:18.991087"
  },
  {
    "epoch": 2924,
    "batch_idx": 128,
    "loss": 0.28151190280914307,
    "learning_rate": 0.0004404514199525664,
    "timestep_mean": 465.84375,
    "epsilon_norm": 8.136589050292969,
    "prediction_norm": 7.019878387451172,
    "timestamp": "2025-07-14T12:35:18.996265"
  },
  {
    "epoch": 2924,
    "batch_idx": 160,
    "loss": 0.29621022939682007,
    "learning_rate": 0.0004404514199525664,
    "timestep_mean": 448.21875,
    "epsilon_norm": 7.726037979125977,
    "prediction_norm": 6.137922286987305,
    "timestamp": "2025-07-14T12:35:19.001388"
  },
  {
    "epoch": 2924,
    "batch_idx": 192,
    "loss": 0.2986603081226349,
    "learning_rate": 0.0004404514199525664,
    "timestep_mean": 391.25,
    "epsilon_norm": 4.014321327209473,
    "prediction_norm": 3.2345340251922607,
    "timestamp": "2025-07-14T12:35:19.006470"
  },
  {
    "epoch": 2925,
    "batch_idx": 0,
    "loss": 0.3000539243221283,
    "learning_rate": 0.00044123130127108235,
    "timestep_mean": 477.21875,
    "epsilon_norm": 7.522017002105713,
    "prediction_norm": 6.114412307739258,
    "timestamp": "2025-07-14T12:35:19.011579"
  },
  {
    "epoch": 2925,
    "batch_idx": 32,
    "loss": 0.25712472200393677,
    "learning_rate": 0.00044123130127108235,
    "timestep_mean": 524.84375,
    "epsilon_norm": 8.011162757873535,
    "prediction_norm": 6.372175216674805,
    "timestamp": "2025-07-14T12:35:19.017063"
  },
  {
    "epoch": 2925,
    "batch_idx": 64,
    "loss": 0.14878426492214203,
    "learning_rate": 0.00044123130127108235,
    "timestep_mean": 547.25,
    "epsilon_norm": 9.115245819091797,
    "prediction_norm": 7.64656925201416,
    "timestamp": "2025-07-14T12:35:19.022382"
  },
  {
    "epoch": 2925,
    "batch_idx": 96,
    "loss": 0.3549526333808899,
    "learning_rate": 0.00044123130127108235,
    "timestep_mean": 499.6875,
    "epsilon_norm": 8.517450332641602,
    "prediction_norm": 6.551684856414795,
    "timestamp": "2025-07-14T12:35:19.027568"
  },
  {
    "epoch": 2925,
    "batch_idx": 128,
    "loss": 0.3576595187187195,
    "learning_rate": 0.00044123130127108235,
    "timestep_mean": 477.71875,
    "epsilon_norm": 8.370450973510742,
    "prediction_norm": 6.546891689300537,
    "timestamp": "2025-07-14T12:35:19.032804"
  },
  {
    "epoch": 2925,
    "batch_idx": 160,
    "loss": 0.35055333375930786,
    "learning_rate": 0.00044123130127108235,
    "timestep_mean": 477.09375,
    "epsilon_norm": 7.258131980895996,
    "prediction_norm": 5.770954608917236,
    "timestamp": "2025-07-14T12:35:19.038045"
  },
  {
    "epoch": 2925,
    "batch_idx": 192,
    "loss": 0.06450024247169495,
    "learning_rate": 0.00044123130127108235,
    "timestep_mean": 644.625,
    "epsilon_norm": 3.4128377437591553,
    "prediction_norm": 2.992788314819336,
    "timestamp": "2025-07-14T12:35:19.043148"
  },
  {
    "epoch": 2926,
    "batch_idx": 0,
    "loss": 0.25931963324546814,
    "learning_rate": 0.0004420113275955205,
    "timestep_mean": 510.75,
    "epsilon_norm": 8.057334899902344,
    "prediction_norm": 6.477589130401611,
    "timestamp": "2025-07-14T12:35:19.048450"
  },
  {
    "epoch": 2926,
    "batch_idx": 32,
    "loss": 0.3910331428050995,
    "learning_rate": 0.0004420113275955205,
    "timestep_mean": 485.25,
    "epsilon_norm": 7.731637477874756,
    "prediction_norm": 5.804107666015625,
    "timestamp": "2025-07-14T12:35:19.053522"
  },
  {
    "epoch": 2926,
    "batch_idx": 64,
    "loss": 0.3742397427558899,
    "learning_rate": 0.0004420113275955205,
    "timestep_mean": 435.34375,
    "epsilon_norm": 7.93626594543457,
    "prediction_norm": 6.447458744049072,
    "timestamp": "2025-07-14T12:35:19.058429"
  },
  {
    "epoch": 2926,
    "batch_idx": 96,
    "loss": 0.32658374309539795,
    "learning_rate": 0.0004420113275955205,
    "timestep_mean": 448.53125,
    "epsilon_norm": 7.437930107116699,
    "prediction_norm": 5.488546371459961,
    "timestamp": "2025-07-14T12:35:19.063499"
  },
  {
    "epoch": 2926,
    "batch_idx": 128,
    "loss": 0.4729963541030884,
    "learning_rate": 0.0004420113275955205,
    "timestep_mean": 494.40625,
    "epsilon_norm": 7.701624870300293,
    "prediction_norm": 5.697486400604248,
    "timestamp": "2025-07-14T12:35:19.068396"
  },
  {
    "epoch": 2926,
    "batch_idx": 160,
    "loss": 0.3105567693710327,
    "learning_rate": 0.0004420113275955205,
    "timestep_mean": 519.59375,
    "epsilon_norm": 7.244688987731934,
    "prediction_norm": 6.525720119476318,
    "timestamp": "2025-07-14T12:35:19.073282"
  },
  {
    "epoch": 2926,
    "batch_idx": 192,
    "loss": 0.27246803045272827,
    "learning_rate": 0.0004420113275955205,
    "timestep_mean": 330.75,
    "epsilon_norm": 4.210777282714844,
    "prediction_norm": 3.2411701679229736,
    "timestamp": "2025-07-14T12:35:19.078377"
  },
  {
    "epoch": 2927,
    "batch_idx": 0,
    "loss": 0.24627164006233215,
    "learning_rate": 0.00044279149700124327,
    "timestep_mean": 544.25,
    "epsilon_norm": 8.027641296386719,
    "prediction_norm": 6.671591281890869,
    "timestamp": "2025-07-14T12:35:19.083483"
  },
  {
    "epoch": 2927,
    "batch_idx": 32,
    "loss": 0.34710049629211426,
    "learning_rate": 0.00044279149700124327,
    "timestep_mean": 491.40625,
    "epsilon_norm": 8.334296226501465,
    "prediction_norm": 6.774667263031006,
    "timestamp": "2025-07-14T12:35:19.088782"
  },
  {
    "epoch": 2927,
    "batch_idx": 64,
    "loss": 0.2958718240261078,
    "learning_rate": 0.00044279149700124327,
    "timestep_mean": 496.46875,
    "epsilon_norm": 8.57520866394043,
    "prediction_norm": 7.23888635635376,
    "timestamp": "2025-07-14T12:35:19.093890"
  },
  {
    "epoch": 2927,
    "batch_idx": 96,
    "loss": 0.2238696813583374,
    "learning_rate": 0.00044279149700124327,
    "timestep_mean": 475.03125,
    "epsilon_norm": 7.511922359466553,
    "prediction_norm": 6.412690162658691,
    "timestamp": "2025-07-14T12:35:19.098947"
  },
  {
    "epoch": 2927,
    "batch_idx": 128,
    "loss": 0.26867079734802246,
    "learning_rate": 0.00044279149700124327,
    "timestep_mean": 524.375,
    "epsilon_norm": 6.860057830810547,
    "prediction_norm": 5.837564945220947,
    "timestamp": "2025-07-14T12:35:19.104027"
  },
  {
    "epoch": 2927,
    "batch_idx": 160,
    "loss": 0.30545783042907715,
    "learning_rate": 0.00044279149700124327,
    "timestep_mean": 585.0625,
    "epsilon_norm": 7.674610614776611,
    "prediction_norm": 6.271712303161621,
    "timestamp": "2025-07-14T12:35:19.109229"
  },
  {
    "epoch": 2927,
    "batch_idx": 192,
    "loss": 0.19016824662685394,
    "learning_rate": 0.00044279149700124327,
    "timestep_mean": 523.625,
    "epsilon_norm": 4.341743469238281,
    "prediction_norm": 4.365907669067383,
    "timestamp": "2025-07-14T12:35:19.114350"
  },
  {
    "epoch": 2928,
    "batch_idx": 0,
    "loss": 0.21920731663703918,
    "learning_rate": 0.00044357180756326027,
    "timestep_mean": 481.875,
    "epsilon_norm": 8.06728458404541,
    "prediction_norm": 7.47230339050293,
    "timestamp": "2025-07-14T12:35:19.119569"
  },
  {
    "epoch": 2928,
    "batch_idx": 32,
    "loss": 0.23417063057422638,
    "learning_rate": 0.00044357180756326027,
    "timestep_mean": 526.34375,
    "epsilon_norm": 8.1704740524292,
    "prediction_norm": 6.9873127937316895,
    "timestamp": "2025-07-14T12:35:19.124834"
  },
  {
    "epoch": 2928,
    "batch_idx": 64,
    "loss": 0.26086127758026123,
    "learning_rate": 0.00044357180756326027,
    "timestep_mean": 504.0,
    "epsilon_norm": 8.203118324279785,
    "prediction_norm": 7.0190749168396,
    "timestamp": "2025-07-14T12:35:19.130093"
  },
  {
    "epoch": 2928,
    "batch_idx": 96,
    "loss": 0.22802124917507172,
    "learning_rate": 0.00044357180756326027,
    "timestep_mean": 480.78125,
    "epsilon_norm": 8.330172538757324,
    "prediction_norm": 7.599128246307373,
    "timestamp": "2025-07-14T12:35:19.135218"
  },
  {
    "epoch": 2928,
    "batch_idx": 128,
    "loss": 0.2996870279312134,
    "learning_rate": 0.00044357180756326027,
    "timestep_mean": 467.40625,
    "epsilon_norm": 7.9583516120910645,
    "prediction_norm": 6.937387943267822,
    "timestamp": "2025-07-14T12:35:19.140440"
  },
  {
    "epoch": 2928,
    "batch_idx": 160,
    "loss": 0.22391930222511292,
    "learning_rate": 0.00044357180756326027,
    "timestep_mean": 487.09375,
    "epsilon_norm": 7.570206642150879,
    "prediction_norm": 6.219746112823486,
    "timestamp": "2025-07-14T12:35:19.145650"
  },
  {
    "epoch": 2928,
    "batch_idx": 192,
    "loss": 0.14711472392082214,
    "learning_rate": 0.00044357180756326027,
    "timestep_mean": 381.875,
    "epsilon_norm": 3.226205825805664,
    "prediction_norm": 3.118502378463745,
    "timestamp": "2025-07-14T12:35:19.150863"
  },
  {
    "epoch": 2929,
    "batch_idx": 0,
    "loss": 0.18137508630752563,
    "learning_rate": 0.0004443522573562327,
    "timestep_mean": 511.5,
    "epsilon_norm": 6.794400215148926,
    "prediction_norm": 6.4143595695495605,
    "timestamp": "2025-07-14T12:35:19.156105"
  },
  {
    "epoch": 2929,
    "batch_idx": 32,
    "loss": 0.4177738130092621,
    "learning_rate": 0.0004443522573562327,
    "timestep_mean": 519.28125,
    "epsilon_norm": 8.02237606048584,
    "prediction_norm": 6.943900108337402,
    "timestamp": "2025-07-14T12:35:19.161349"
  },
  {
    "epoch": 2929,
    "batch_idx": 64,
    "loss": 0.5364894866943359,
    "learning_rate": 0.0004443522573562327,
    "timestep_mean": 421.6875,
    "epsilon_norm": 7.856091499328613,
    "prediction_norm": 5.664046287536621,
    "timestamp": "2025-07-14T12:35:19.166667"
  },
  {
    "epoch": 2929,
    "batch_idx": 96,
    "loss": 0.3535730242729187,
    "learning_rate": 0.0004443522573562327,
    "timestep_mean": 479.9375,
    "epsilon_norm": 7.3180999755859375,
    "prediction_norm": 5.592637062072754,
    "timestamp": "2025-07-14T12:35:19.171910"
  },
  {
    "epoch": 2929,
    "batch_idx": 128,
    "loss": 0.2584874629974365,
    "learning_rate": 0.0004443522573562327,
    "timestep_mean": 494.71875,
    "epsilon_norm": 7.9589033126831055,
    "prediction_norm": 6.724039554595947,
    "timestamp": "2025-07-14T12:35:19.177197"
  },
  {
    "epoch": 2929,
    "batch_idx": 160,
    "loss": 0.16941291093826294,
    "learning_rate": 0.0004443522573562327,
    "timestep_mean": 553.65625,
    "epsilon_norm": 9.151570320129395,
    "prediction_norm": 8.260702133178711,
    "timestamp": "2025-07-14T12:35:19.182362"
  },
  {
    "epoch": 2929,
    "batch_idx": 192,
    "loss": 0.13600951433181763,
    "learning_rate": 0.0004443522573562327,
    "timestep_mean": 587.875,
    "epsilon_norm": 4.328383445739746,
    "prediction_norm": 3.8282930850982666,
    "timestamp": "2025-07-14T12:35:19.187430"
  },
  {
    "epoch": 2930,
    "batch_idx": 0,
    "loss": 0.21499702334403992,
    "learning_rate": 0.0004451328444544788,
    "timestep_mean": 497.5625,
    "epsilon_norm": 7.760295867919922,
    "prediction_norm": 6.653627872467041,
    "timestamp": "2025-07-14T12:35:19.192837"
  },
  {
    "epoch": 2930,
    "batch_idx": 32,
    "loss": 0.36643677949905396,
    "learning_rate": 0.0004451328444544788,
    "timestep_mean": 496.09375,
    "epsilon_norm": 7.288800239562988,
    "prediction_norm": 5.2508440017700195,
    "timestamp": "2025-07-14T12:35:19.197999"
  },
  {
    "epoch": 2930,
    "batch_idx": 64,
    "loss": 0.16820888221263885,
    "learning_rate": 0.0004451328444544788,
    "timestep_mean": 507.03125,
    "epsilon_norm": 7.372514247894287,
    "prediction_norm": 6.6839680671691895,
    "timestamp": "2025-07-14T12:35:19.203023"
  },
  {
    "epoch": 2930,
    "batch_idx": 96,
    "loss": 0.5165111422538757,
    "learning_rate": 0.0004451328444544788,
    "timestep_mean": 538.40625,
    "epsilon_norm": 9.188333511352539,
    "prediction_norm": 8.04789924621582,
    "timestamp": "2025-07-14T12:35:19.208162"
  },
  {
    "epoch": 2930,
    "batch_idx": 128,
    "loss": 0.26096704602241516,
    "learning_rate": 0.0004451328444544788,
    "timestep_mean": 568.40625,
    "epsilon_norm": 7.694228172302246,
    "prediction_norm": 6.6585373878479,
    "timestamp": "2025-07-14T12:35:19.213587"
  },
  {
    "epoch": 2930,
    "batch_idx": 160,
    "loss": 0.20513014495372772,
    "learning_rate": 0.0004451328444544788,
    "timestep_mean": 556.6875,
    "epsilon_norm": 7.429027557373047,
    "prediction_norm": 6.448256969451904,
    "timestamp": "2025-07-14T12:35:19.218742"
  },
  {
    "epoch": 2930,
    "batch_idx": 192,
    "loss": 0.07710540294647217,
    "learning_rate": 0.0004451328444544788,
    "timestep_mean": 532.0,
    "epsilon_norm": 3.7990872859954834,
    "prediction_norm": 3.1791908740997314,
    "timestamp": "2025-07-14T12:35:19.223923"
  },
  {
    "epoch": 2931,
    "batch_idx": 0,
    "loss": 0.20513738691806793,
    "learning_rate": 0.0004459135669319765,
    "timestep_mean": 481.34375,
    "epsilon_norm": 6.7060394287109375,
    "prediction_norm": 5.7970290184021,
    "timestamp": "2025-07-14T12:35:19.229119"
  },
  {
    "epoch": 2931,
    "batch_idx": 32,
    "loss": 0.35767802596092224,
    "learning_rate": 0.0004459135669319765,
    "timestep_mean": 540.71875,
    "epsilon_norm": 8.74193286895752,
    "prediction_norm": 8.107956886291504,
    "timestamp": "2025-07-14T12:35:19.234250"
  },
  {
    "epoch": 2931,
    "batch_idx": 64,
    "loss": 0.21436432003974915,
    "learning_rate": 0.0004459135669319765,
    "timestep_mean": 606.625,
    "epsilon_norm": 8.274746894836426,
    "prediction_norm": 7.029448986053467,
    "timestamp": "2025-07-14T12:35:19.239493"
  },
  {
    "epoch": 2931,
    "batch_idx": 96,
    "loss": 0.25474557280540466,
    "learning_rate": 0.0004459135669319765,
    "timestep_mean": 526.28125,
    "epsilon_norm": 7.729105472564697,
    "prediction_norm": 6.443495273590088,
    "timestamp": "2025-07-14T12:35:19.244736"
  },
  {
    "epoch": 2931,
    "batch_idx": 128,
    "loss": 0.5303943157196045,
    "learning_rate": 0.0004459135669319765,
    "timestep_mean": 451.34375,
    "epsilon_norm": 8.160218238830566,
    "prediction_norm": 6.529788017272949,
    "timestamp": "2025-07-14T12:35:19.249707"
  },
  {
    "epoch": 2931,
    "batch_idx": 160,
    "loss": 0.18332917988300323,
    "learning_rate": 0.0004459135669319765,
    "timestep_mean": 542.71875,
    "epsilon_norm": 7.139559745788574,
    "prediction_norm": 6.4474968910217285,
    "timestamp": "2025-07-14T12:35:19.254778"
  },
  {
    "epoch": 2931,
    "batch_idx": 192,
    "loss": 0.5835814476013184,
    "learning_rate": 0.0004459135669319765,
    "timestep_mean": 364.75,
    "epsilon_norm": 4.210081577301025,
    "prediction_norm": 2.277270555496216,
    "timestamp": "2025-07-14T12:35:19.259893"
  },
  {
    "epoch": 2932,
    "batch_idx": 0,
    "loss": 0.6315401792526245,
    "learning_rate": 0.0004466944228623709,
    "timestep_mean": 470.375,
    "epsilon_norm": 8.605692863464355,
    "prediction_norm": 6.8699517250061035,
    "timestamp": "2025-07-14T12:35:19.265184"
  },
  {
    "epoch": 2932,
    "batch_idx": 32,
    "loss": 0.23714986443519592,
    "learning_rate": 0.0004466944228623709,
    "timestep_mean": 524.15625,
    "epsilon_norm": 7.510590076446533,
    "prediction_norm": 6.8845415115356445,
    "timestamp": "2025-07-14T12:35:19.270394"
  },
  {
    "epoch": 2932,
    "batch_idx": 64,
    "loss": 0.1167609691619873,
    "learning_rate": 0.0004466944228623709,
    "timestep_mean": 642.375,
    "epsilon_norm": 7.622704982757568,
    "prediction_norm": 7.251810073852539,
    "timestamp": "2025-07-14T12:35:19.275779"
  },
  {
    "epoch": 2932,
    "batch_idx": 96,
    "loss": 0.303701788187027,
    "learning_rate": 0.0004466944228623709,
    "timestep_mean": 451.75,
    "epsilon_norm": 8.93803882598877,
    "prediction_norm": 8.19127082824707,
    "timestamp": "2025-07-14T12:35:19.280957"
  },
  {
    "epoch": 2932,
    "batch_idx": 128,
    "loss": 0.12613990902900696,
    "learning_rate": 0.0004466944228623709,
    "timestep_mean": 490.90625,
    "epsilon_norm": 7.1111955642700195,
    "prediction_norm": 6.489922523498535,
    "timestamp": "2025-07-14T12:35:19.286040"
  },
  {
    "epoch": 2932,
    "batch_idx": 160,
    "loss": 0.28479641675949097,
    "learning_rate": 0.0004466944228623709,
    "timestep_mean": 535.71875,
    "epsilon_norm": 7.984126567840576,
    "prediction_norm": 6.483291149139404,
    "timestamp": "2025-07-14T12:35:19.291274"
  },
  {
    "epoch": 2932,
    "batch_idx": 192,
    "loss": 0.11627037078142166,
    "learning_rate": 0.0004466944228623709,
    "timestep_mean": 608.125,
    "epsilon_norm": 3.4384877681732178,
    "prediction_norm": 3.4028594493865967,
    "timestamp": "2025-07-14T12:35:19.296425"
  },
  {
    "epoch": 2933,
    "batch_idx": 0,
    "loss": 0.35709914565086365,
    "learning_rate": 0.0004474754103189783,
    "timestep_mean": 462.28125,
    "epsilon_norm": 7.867795944213867,
    "prediction_norm": 6.225615501403809,
    "timestamp": "2025-07-14T12:35:19.301604"
  },
  {
    "epoch": 2933,
    "batch_idx": 32,
    "loss": 0.34304869174957275,
    "learning_rate": 0.0004474754103189783,
    "timestep_mean": 498.84375,
    "epsilon_norm": 8.173819541931152,
    "prediction_norm": 6.564417839050293,
    "timestamp": "2025-07-14T12:35:19.306615"
  },
  {
    "epoch": 2933,
    "batch_idx": 64,
    "loss": 0.464775949716568,
    "learning_rate": 0.0004474754103189783,
    "timestep_mean": 481.375,
    "epsilon_norm": 7.925171375274658,
    "prediction_norm": 6.072270393371582,
    "timestamp": "2025-07-14T12:35:19.311894"
  },
  {
    "epoch": 2933,
    "batch_idx": 96,
    "loss": 0.16009247303009033,
    "learning_rate": 0.0004474754103189783,
    "timestep_mean": 584.34375,
    "epsilon_norm": 7.79653787612915,
    "prediction_norm": 6.701373100280762,
    "timestamp": "2025-07-14T12:35:19.317304"
  },
  {
    "epoch": 2933,
    "batch_idx": 128,
    "loss": 0.2792285084724426,
    "learning_rate": 0.0004474754103189783,
    "timestep_mean": 568.28125,
    "epsilon_norm": 8.520845413208008,
    "prediction_norm": 6.785283088684082,
    "timestamp": "2025-07-14T12:35:19.322441"
  },
  {
    "epoch": 2933,
    "batch_idx": 160,
    "loss": 0.22973886132240295,
    "learning_rate": 0.0004474754103189783,
    "timestep_mean": 612.875,
    "epsilon_norm": 9.572004318237305,
    "prediction_norm": 8.501908302307129,
    "timestamp": "2025-07-14T12:35:19.327491"
  },
  {
    "epoch": 2933,
    "batch_idx": 192,
    "loss": 0.586958646774292,
    "learning_rate": 0.0004474754103189783,
    "timestep_mean": 507.5,
    "epsilon_norm": 4.680159091949463,
    "prediction_norm": 3.6372439861297607,
    "timestamp": "2025-07-14T12:35:19.332470"
  },
  {
    "epoch": 2934,
    "batch_idx": 0,
    "loss": 0.3463824391365051,
    "learning_rate": 0.00044825652737478996,
    "timestep_mean": 528.1875,
    "epsilon_norm": 8.745882987976074,
    "prediction_norm": 7.88129997253418,
    "timestamp": "2025-07-14T12:35:19.337361"
  },
  {
    "epoch": 2934,
    "batch_idx": 32,
    "loss": 0.22377808392047882,
    "learning_rate": 0.00044825652737478996,
    "timestep_mean": 535.6875,
    "epsilon_norm": 8.352884292602539,
    "prediction_norm": 7.313117027282715,
    "timestamp": "2025-07-14T12:35:19.342274"
  },
  {
    "epoch": 2934,
    "batch_idx": 64,
    "loss": 0.46870607137680054,
    "learning_rate": 0.00044825652737478996,
    "timestep_mean": 426.21875,
    "epsilon_norm": 9.016417503356934,
    "prediction_norm": 6.97367000579834,
    "timestamp": "2025-07-14T12:35:19.347134"
  },
  {
    "epoch": 2934,
    "batch_idx": 96,
    "loss": 0.26341110467910767,
    "learning_rate": 0.00044825652737478996,
    "timestep_mean": 533.53125,
    "epsilon_norm": 8.076592445373535,
    "prediction_norm": 6.752857685089111,
    "timestamp": "2025-07-14T12:35:19.352012"
  },
  {
    "epoch": 2934,
    "batch_idx": 128,
    "loss": 0.1462019681930542,
    "learning_rate": 0.00044825652737478996,
    "timestep_mean": 600.71875,
    "epsilon_norm": 7.800099849700928,
    "prediction_norm": 6.684910774230957,
    "timestamp": "2025-07-14T12:35:19.356869"
  },
  {
    "epoch": 2934,
    "batch_idx": 160,
    "loss": 0.35557639598846436,
    "learning_rate": 0.00044825652737478996,
    "timestep_mean": 440.25,
    "epsilon_norm": 8.127541542053223,
    "prediction_norm": 6.806552886962891,
    "timestamp": "2025-07-14T12:35:19.361724"
  },
  {
    "epoch": 2934,
    "batch_idx": 192,
    "loss": 0.5411296486854553,
    "learning_rate": 0.00044825652737478996,
    "timestep_mean": 652.875,
    "epsilon_norm": 4.351536273956299,
    "prediction_norm": 3.227832555770874,
    "timestamp": "2025-07-14T12:35:19.366540"
  },
  {
    "epoch": 2935,
    "batch_idx": 0,
    "loss": 0.13015715777873993,
    "learning_rate": 0.0004490377721024762,
    "timestep_mean": 620.71875,
    "epsilon_norm": 9.709634780883789,
    "prediction_norm": 8.366503715515137,
    "timestamp": "2025-07-14T12:35:19.371484"
  },
  {
    "epoch": 2935,
    "batch_idx": 32,
    "loss": 0.21057502925395966,
    "learning_rate": 0.0004490377721024762,
    "timestep_mean": 531.1875,
    "epsilon_norm": 8.33199691772461,
    "prediction_norm": 7.423123836517334,
    "timestamp": "2025-07-14T12:35:19.376367"
  },
  {
    "epoch": 2935,
    "batch_idx": 64,
    "loss": 0.2693570852279663,
    "learning_rate": 0.0004490377721024762,
    "timestep_mean": 513.4375,
    "epsilon_norm": 9.087471008300781,
    "prediction_norm": 7.871659755706787,
    "timestamp": "2025-07-14T12:35:19.381196"
  },
  {
    "epoch": 2935,
    "batch_idx": 96,
    "loss": 0.2598322331905365,
    "learning_rate": 0.0004490377721024762,
    "timestep_mean": 511.375,
    "epsilon_norm": 8.782536506652832,
    "prediction_norm": 7.264800548553467,
    "timestamp": "2025-07-14T12:35:19.386084"
  },
  {
    "epoch": 2935,
    "batch_idx": 128,
    "loss": 0.2283351719379425,
    "learning_rate": 0.0004490377721024762,
    "timestep_mean": 464.03125,
    "epsilon_norm": 8.24482536315918,
    "prediction_norm": 7.0753021240234375,
    "timestamp": "2025-07-14T12:35:19.390965"
  },
  {
    "epoch": 2935,
    "batch_idx": 160,
    "loss": 0.39367082715034485,
    "learning_rate": 0.0004490377721024762,
    "timestep_mean": 443.84375,
    "epsilon_norm": 9.019367218017578,
    "prediction_norm": 6.626086235046387,
    "timestamp": "2025-07-14T12:35:19.395874"
  },
  {
    "epoch": 2935,
    "batch_idx": 192,
    "loss": 0.10047636926174164,
    "learning_rate": 0.0004490377721024762,
    "timestep_mean": 424.0,
    "epsilon_norm": 4.066774845123291,
    "prediction_norm": 3.511288642883301,
    "timestamp": "2025-07-14T12:35:19.400766"
  },
  {
    "epoch": 2936,
    "batch_idx": 0,
    "loss": 0.1975143700838089,
    "learning_rate": 0.0004498191425743937,
    "timestep_mean": 564.25,
    "epsilon_norm": 7.6195197105407715,
    "prediction_norm": 7.135593891143799,
    "timestamp": "2025-07-14T12:35:19.405708"
  },
  {
    "epoch": 2936,
    "batch_idx": 32,
    "loss": 0.13669291138648987,
    "learning_rate": 0.0004498191425743937,
    "timestep_mean": 579.8125,
    "epsilon_norm": 8.541180610656738,
    "prediction_norm": 8.21212100982666,
    "timestamp": "2025-07-14T12:35:19.410581"
  },
  {
    "epoch": 2936,
    "batch_idx": 64,
    "loss": 0.26388829946517944,
    "learning_rate": 0.0004498191425743937,
    "timestep_mean": 472.03125,
    "epsilon_norm": 6.720203876495361,
    "prediction_norm": 6.0990705490112305,
    "timestamp": "2025-07-14T12:35:19.415724"
  },
  {
    "epoch": 2936,
    "batch_idx": 96,
    "loss": 0.37432122230529785,
    "learning_rate": 0.0004498191425743937,
    "timestep_mean": 504.9375,
    "epsilon_norm": 9.253440856933594,
    "prediction_norm": 8.280036926269531,
    "timestamp": "2025-07-14T12:35:19.420974"
  },
  {
    "epoch": 2936,
    "batch_idx": 128,
    "loss": 0.29647594690322876,
    "learning_rate": 0.0004498191425743937,
    "timestep_mean": 512.4375,
    "epsilon_norm": 7.568197727203369,
    "prediction_norm": 6.9429030418396,
    "timestamp": "2025-07-14T12:35:19.426315"
  },
  {
    "epoch": 2936,
    "batch_idx": 160,
    "loss": 0.402337908744812,
    "learning_rate": 0.0004498191425743937,
    "timestep_mean": 450.8125,
    "epsilon_norm": 8.056709289550781,
    "prediction_norm": 5.831182956695557,
    "timestamp": "2025-07-14T12:35:19.431613"
  },
  {
    "epoch": 2936,
    "batch_idx": 192,
    "loss": 0.17459674179553986,
    "learning_rate": 0.0004498191425743937,
    "timestep_mean": 480.5,
    "epsilon_norm": 3.5655016899108887,
    "prediction_norm": 3.1306676864624023,
    "timestamp": "2025-07-14T12:35:19.436916"
  },
  {
    "epoch": 2937,
    "batch_idx": 0,
    "loss": 0.35709524154663086,
    "learning_rate": 0.0004506006368625891,
    "timestep_mean": 543.71875,
    "epsilon_norm": 9.344172477722168,
    "prediction_norm": 8.041050910949707,
    "timestamp": "2025-07-14T12:35:19.442238"
  },
  {
    "epoch": 2937,
    "batch_idx": 32,
    "loss": 0.3052046597003937,
    "learning_rate": 0.0004506006368625891,
    "timestep_mean": 469.78125,
    "epsilon_norm": 8.80759048461914,
    "prediction_norm": 7.813077449798584,
    "timestamp": "2025-07-14T12:35:19.447403"
  },
  {
    "epoch": 2937,
    "batch_idx": 64,
    "loss": 0.5022066235542297,
    "learning_rate": 0.0004506006368625891,
    "timestep_mean": 478.09375,
    "epsilon_norm": 8.701308250427246,
    "prediction_norm": 6.869323253631592,
    "timestamp": "2025-07-14T12:35:19.452665"
  },
  {
    "epoch": 2937,
    "batch_idx": 96,
    "loss": 0.223832905292511,
    "learning_rate": 0.0004506006368625891,
    "timestep_mean": 461.3125,
    "epsilon_norm": 9.260000228881836,
    "prediction_norm": 8.372527122497559,
    "timestamp": "2025-07-14T12:35:19.457805"
  },
  {
    "epoch": 2937,
    "batch_idx": 128,
    "loss": 0.23838479816913605,
    "learning_rate": 0.0004506006368625891,
    "timestep_mean": 518.34375,
    "epsilon_norm": 7.609281063079834,
    "prediction_norm": 6.470864772796631,
    "timestamp": "2025-07-14T12:35:19.463009"
  },
  {
    "epoch": 2937,
    "batch_idx": 160,
    "loss": 0.24090389907360077,
    "learning_rate": 0.0004506006368625891,
    "timestep_mean": 463.59375,
    "epsilon_norm": 8.2587890625,
    "prediction_norm": 7.037270545959473,
    "timestamp": "2025-07-14T12:35:19.468198"
  },
  {
    "epoch": 2937,
    "batch_idx": 192,
    "loss": 0.14221221208572388,
    "learning_rate": 0.0004506006368625891,
    "timestep_mean": 454.375,
    "epsilon_norm": 5.098321437835693,
    "prediction_norm": 4.568614482879639,
    "timestamp": "2025-07-14T12:35:19.473430"
  },
  {
    "epoch": 2938,
    "batch_idx": 0,
    "loss": 0.17217665910720825,
    "learning_rate": 0.0004513822530388014,
    "timestep_mean": 468.46875,
    "epsilon_norm": 6.615779399871826,
    "prediction_norm": 6.437990188598633,
    "timestamp": "2025-07-14T12:35:19.478705"
  },
  {
    "epoch": 2938,
    "batch_idx": 32,
    "loss": 0.2818872332572937,
    "learning_rate": 0.0004513822530388014,
    "timestep_mean": 531.15625,
    "epsilon_norm": 8.75085735321045,
    "prediction_norm": 7.612011909484863,
    "timestamp": "2025-07-14T12:35:19.484053"
  },
  {
    "epoch": 2938,
    "batch_idx": 64,
    "loss": 0.2747858166694641,
    "learning_rate": 0.0004513822530388014,
    "timestep_mean": 441.46875,
    "epsilon_norm": 8.854544639587402,
    "prediction_norm": 6.908243179321289,
    "timestamp": "2025-07-14T12:35:19.489256"
  },
  {
    "epoch": 2938,
    "batch_idx": 96,
    "loss": 0.30509909987449646,
    "learning_rate": 0.0004513822530388014,
    "timestep_mean": 498.0,
    "epsilon_norm": 7.853566646575928,
    "prediction_norm": 6.769389629364014,
    "timestamp": "2025-07-14T12:35:19.494405"
  },
  {
    "epoch": 2938,
    "batch_idx": 128,
    "loss": 0.22315019369125366,
    "learning_rate": 0.0004513822530388014,
    "timestep_mean": 490.6875,
    "epsilon_norm": 8.029634475708008,
    "prediction_norm": 7.572627067565918,
    "timestamp": "2025-07-14T12:35:19.499445"
  },
  {
    "epoch": 2938,
    "batch_idx": 160,
    "loss": 0.12706944346427917,
    "learning_rate": 0.0004513822530388014,
    "timestep_mean": 467.9375,
    "epsilon_norm": 6.766483306884766,
    "prediction_norm": 6.402101039886475,
    "timestamp": "2025-07-14T12:35:19.504411"
  },
  {
    "epoch": 2938,
    "batch_idx": 192,
    "loss": 0.08591745793819427,
    "learning_rate": 0.0004513822530388014,
    "timestep_mean": 697.125,
    "epsilon_norm": 4.219326496124268,
    "prediction_norm": 3.8707563877105713,
    "timestamp": "2025-07-14T12:35:19.509562"
  },
  {
    "epoch": 2939,
    "batch_idx": 0,
    "loss": 0.38158589601516724,
    "learning_rate": 0.00045216398917447193,
    "timestep_mean": 436.96875,
    "epsilon_norm": 8.137751579284668,
    "prediction_norm": 6.880005836486816,
    "timestamp": "2025-07-14T12:35:19.514769"
  },
  {
    "epoch": 2939,
    "batch_idx": 32,
    "loss": 0.24250343441963196,
    "learning_rate": 0.00045216398917447193,
    "timestep_mean": 453.4375,
    "epsilon_norm": 7.5971808433532715,
    "prediction_norm": 5.7228617668151855,
    "timestamp": "2025-07-14T12:35:19.520074"
  },
  {
    "epoch": 2939,
    "batch_idx": 64,
    "loss": 0.37447604537010193,
    "learning_rate": 0.00045216398917447193,
    "timestep_mean": 528.1875,
    "epsilon_norm": 8.45097827911377,
    "prediction_norm": 6.7932634353637695,
    "timestamp": "2025-07-14T12:35:19.525377"
  },
  {
    "epoch": 2939,
    "batch_idx": 96,
    "loss": 0.16757190227508545,
    "learning_rate": 0.00045216398917447193,
    "timestep_mean": 473.9375,
    "epsilon_norm": 8.888957023620605,
    "prediction_norm": 7.953481674194336,
    "timestamp": "2025-07-14T12:35:19.530364"
  },
  {
    "epoch": 2939,
    "batch_idx": 128,
    "loss": 0.1753121018409729,
    "learning_rate": 0.00045216398917447193,
    "timestep_mean": 555.25,
    "epsilon_norm": 7.946828365325928,
    "prediction_norm": 6.7472944259643555,
    "timestamp": "2025-07-14T12:35:19.535314"
  },
  {
    "epoch": 2939,
    "batch_idx": 160,
    "loss": 0.32784220576286316,
    "learning_rate": 0.00045216398917447193,
    "timestep_mean": 453.875,
    "epsilon_norm": 8.45180606842041,
    "prediction_norm": 7.121543884277344,
    "timestamp": "2025-07-14T12:35:19.540501"
  },
  {
    "epoch": 2939,
    "batch_idx": 192,
    "loss": 0.16527403891086578,
    "learning_rate": 0.00045216398917447193,
    "timestep_mean": 566.625,
    "epsilon_norm": 3.7984871864318848,
    "prediction_norm": 3.9428281784057617,
    "timestamp": "2025-07-14T12:35:19.545787"
  },
  {
    "epoch": 2940,
    "batch_idx": 0,
    "loss": 0.25433915853500366,
    "learning_rate": 0.00045294584334074387,
    "timestep_mean": 458.40625,
    "epsilon_norm": 7.486027717590332,
    "prediction_norm": 6.304356575012207,
    "timestamp": "2025-07-14T12:35:19.550919"
  },
  {
    "epoch": 2940,
    "batch_idx": 32,
    "loss": 0.2621614634990692,
    "learning_rate": 0.00045294584334074387,
    "timestep_mean": 454.25,
    "epsilon_norm": 8.175978660583496,
    "prediction_norm": 6.876945495605469,
    "timestamp": "2025-07-14T12:35:19.556147"
  },
  {
    "epoch": 2940,
    "batch_idx": 64,
    "loss": 0.19087541103363037,
    "learning_rate": 0.00045294584334074387,
    "timestep_mean": 540.15625,
    "epsilon_norm": 7.392985820770264,
    "prediction_norm": 6.386723518371582,
    "timestamp": "2025-07-14T12:35:19.561408"
  },
  {
    "epoch": 2940,
    "batch_idx": 96,
    "loss": 0.2048715353012085,
    "learning_rate": 0.00045294584334074387,
    "timestep_mean": 529.75,
    "epsilon_norm": 8.726763725280762,
    "prediction_norm": 7.252127170562744,
    "timestamp": "2025-07-14T12:35:19.566567"
  },
  {
    "epoch": 2940,
    "batch_idx": 128,
    "loss": 0.33498847484588623,
    "learning_rate": 0.00045294584334074387,
    "timestep_mean": 444.15625,
    "epsilon_norm": 7.729016304016113,
    "prediction_norm": 6.502425193786621,
    "timestamp": "2025-07-14T12:35:19.571859"
  },
  {
    "epoch": 2940,
    "batch_idx": 160,
    "loss": 0.14298216998577118,
    "learning_rate": 0.00045294584334074387,
    "timestep_mean": 579.625,
    "epsilon_norm": 6.7810750007629395,
    "prediction_norm": 5.879004001617432,
    "timestamp": "2025-07-14T12:35:19.577011"
  },
  {
    "epoch": 2940,
    "batch_idx": 192,
    "loss": 0.10173599421977997,
    "learning_rate": 0.00045294584334074387,
    "timestep_mean": 597.75,
    "epsilon_norm": 2.537369966506958,
    "prediction_norm": 2.653123378753662,
    "timestamp": "2025-07-14T12:35:19.582163"
  },
  {
    "epoch": 2941,
    "batch_idx": 0,
    "loss": 0.2826715111732483,
    "learning_rate": 0.00045372781360847027,
    "timestep_mean": 516.625,
    "epsilon_norm": 9.633820533752441,
    "prediction_norm": 8.546181678771973,
    "timestamp": "2025-07-14T12:35:19.587393"
  },
  {
    "epoch": 2941,
    "batch_idx": 32,
    "loss": 0.3221440315246582,
    "learning_rate": 0.00045372781360847027,
    "timestep_mean": 470.96875,
    "epsilon_norm": 8.192819595336914,
    "prediction_norm": 6.124181747436523,
    "timestamp": "2025-07-14T12:35:19.592603"
  },
  {
    "epoch": 2941,
    "batch_idx": 64,
    "loss": 0.23720940947532654,
    "learning_rate": 0.00045372781360847027,
    "timestep_mean": 538.875,
    "epsilon_norm": 7.182093620300293,
    "prediction_norm": 6.655923843383789,
    "timestamp": "2025-07-14T12:35:19.597750"
  },
  {
    "epoch": 2941,
    "batch_idx": 96,
    "loss": 0.5726550817489624,
    "learning_rate": 0.00045372781360847027,
    "timestep_mean": 483.75,
    "epsilon_norm": 9.393592834472656,
    "prediction_norm": 7.0882792472839355,
    "timestamp": "2025-07-14T12:35:19.603004"
  },
  {
    "epoch": 2941,
    "batch_idx": 128,
    "loss": 0.23225529491901398,
    "learning_rate": 0.00045372781360847027,
    "timestep_mean": 608.125,
    "epsilon_norm": 7.183887481689453,
    "prediction_norm": 6.446411609649658,
    "timestamp": "2025-07-14T12:35:19.608255"
  },
  {
    "epoch": 2941,
    "batch_idx": 160,
    "loss": 0.29857301712036133,
    "learning_rate": 0.00045372781360847027,
    "timestep_mean": 560.5,
    "epsilon_norm": 6.889885902404785,
    "prediction_norm": 5.723570823669434,
    "timestamp": "2025-07-14T12:35:19.613654"
  },
  {
    "epoch": 2941,
    "batch_idx": 192,
    "loss": 0.3179991841316223,
    "learning_rate": 0.00045372781360847027,
    "timestep_mean": 451.125,
    "epsilon_norm": 4.505697250366211,
    "prediction_norm": 3.785250186920166,
    "timestamp": "2025-07-14T12:35:19.618810"
  },
  {
    "epoch": 2942,
    "batch_idx": 0,
    "loss": 0.41574186086654663,
    "learning_rate": 0.00045450989804821643,
    "timestep_mean": 463.71875,
    "epsilon_norm": 7.505464553833008,
    "prediction_norm": 5.1897292137146,
    "timestamp": "2025-07-14T12:35:19.624142"
  },
  {
    "epoch": 2942,
    "batch_idx": 32,
    "loss": 0.2144545018672943,
    "learning_rate": 0.00045450989804821643,
    "timestep_mean": 527.03125,
    "epsilon_norm": 6.892685413360596,
    "prediction_norm": 6.241314888000488,
    "timestamp": "2025-07-14T12:35:19.629267"
  },
  {
    "epoch": 2942,
    "batch_idx": 64,
    "loss": 0.3707825541496277,
    "learning_rate": 0.00045450989804821643,
    "timestep_mean": 453.21875,
    "epsilon_norm": 7.436176300048828,
    "prediction_norm": 6.3021159172058105,
    "timestamp": "2025-07-14T12:35:19.634356"
  },
  {
    "epoch": 2942,
    "batch_idx": 96,
    "loss": 0.40635591745376587,
    "learning_rate": 0.00045450989804821643,
    "timestep_mean": 475.5,
    "epsilon_norm": 8.692740440368652,
    "prediction_norm": 6.659330368041992,
    "timestamp": "2025-07-14T12:35:19.639220"
  },
  {
    "epoch": 2942,
    "batch_idx": 128,
    "loss": 0.26991093158721924,
    "learning_rate": 0.00045450989804821643,
    "timestep_mean": 521.4375,
    "epsilon_norm": 7.58950138092041,
    "prediction_norm": 6.2837066650390625,
    "timestamp": "2025-07-14T12:35:19.644086"
  },
  {
    "epoch": 2942,
    "batch_idx": 160,
    "loss": 0.25007864832878113,
    "learning_rate": 0.00045450989804821643,
    "timestep_mean": 507.65625,
    "epsilon_norm": 9.210884094238281,
    "prediction_norm": 8.251469612121582,
    "timestamp": "2025-07-14T12:35:19.648938"
  },
  {
    "epoch": 2942,
    "batch_idx": 192,
    "loss": 0.340004026889801,
    "learning_rate": 0.00045450989804821643,
    "timestep_mean": 485.5,
    "epsilon_norm": 3.970006227493286,
    "prediction_norm": 3.4663496017456055,
    "timestamp": "2025-07-14T12:35:19.653954"
  },
  {
    "epoch": 2943,
    "batch_idx": 0,
    "loss": 0.1667892336845398,
    "learning_rate": 0.0004552920947302666,
    "timestep_mean": 662.65625,
    "epsilon_norm": 8.581916809082031,
    "prediction_norm": 7.7808308601379395,
    "timestamp": "2025-07-14T12:35:19.659157"
  },
  {
    "epoch": 2943,
    "batch_idx": 32,
    "loss": 0.5462517738342285,
    "learning_rate": 0.0004552920947302666,
    "timestep_mean": 495.3125,
    "epsilon_norm": 9.115219116210938,
    "prediction_norm": 7.6409687995910645,
    "timestamp": "2025-07-14T12:35:19.664415"
  },
  {
    "epoch": 2943,
    "batch_idx": 64,
    "loss": 0.4041401147842407,
    "learning_rate": 0.0004552920947302666,
    "timestep_mean": 463.28125,
    "epsilon_norm": 8.447723388671875,
    "prediction_norm": 6.472101211547852,
    "timestamp": "2025-07-14T12:35:19.669675"
  },
  {
    "epoch": 2943,
    "batch_idx": 96,
    "loss": 0.3217318058013916,
    "learning_rate": 0.0004552920947302666,
    "timestep_mean": 477.53125,
    "epsilon_norm": 7.194791316986084,
    "prediction_norm": 6.27049446105957,
    "timestamp": "2025-07-14T12:35:19.674843"
  },
  {
    "epoch": 2943,
    "batch_idx": 128,
    "loss": 0.40839388966560364,
    "learning_rate": 0.0004552920947302666,
    "timestep_mean": 490.21875,
    "epsilon_norm": 7.457673072814941,
    "prediction_norm": 6.384890556335449,
    "timestamp": "2025-07-14T12:35:19.680108"
  },
  {
    "epoch": 2943,
    "batch_idx": 160,
    "loss": 0.391564279794693,
    "learning_rate": 0.0004552920947302666,
    "timestep_mean": 460.9375,
    "epsilon_norm": 7.577538967132568,
    "prediction_norm": 6.408224582672119,
    "timestamp": "2025-07-14T12:35:19.685216"
  },
  {
    "epoch": 2943,
    "batch_idx": 192,
    "loss": 0.5214851498603821,
    "learning_rate": 0.0004552920947302666,
    "timestep_mean": 335.0,
    "epsilon_norm": 5.645705223083496,
    "prediction_norm": 5.8289618492126465,
    "timestamp": "2025-07-14T12:35:19.690393"
  },
  {
    "epoch": 2944,
    "batch_idx": 0,
    "loss": 0.2704547047615051,
    "learning_rate": 0.00045607440172462963,
    "timestep_mean": 524.125,
    "epsilon_norm": 7.365929126739502,
    "prediction_norm": 6.456141948699951,
    "timestamp": "2025-07-14T12:35:19.695479"
  },
  {
    "epoch": 2944,
    "batch_idx": 32,
    "loss": 0.28907448053359985,
    "learning_rate": 0.00045607440172462963,
    "timestep_mean": 557.875,
    "epsilon_norm": 8.808088302612305,
    "prediction_norm": 8.045557975769043,
    "timestamp": "2025-07-14T12:35:19.700658"
  },
  {
    "epoch": 2944,
    "batch_idx": 64,
    "loss": 0.26732689142227173,
    "learning_rate": 0.00045607440172462963,
    "timestep_mean": 441.65625,
    "epsilon_norm": 9.151989936828613,
    "prediction_norm": 8.084155082702637,
    "timestamp": "2025-07-14T12:35:19.705851"
  },
  {
    "epoch": 2944,
    "batch_idx": 96,
    "loss": 0.16997581720352173,
    "learning_rate": 0.00045607440172462963,
    "timestep_mean": 582.875,
    "epsilon_norm": 7.668690204620361,
    "prediction_norm": 7.4688286781311035,
    "timestamp": "2025-07-14T12:35:19.711046"
  },
  {
    "epoch": 2944,
    "batch_idx": 128,
    "loss": 0.2756105065345764,
    "learning_rate": 0.00045607440172462963,
    "timestep_mean": 540.53125,
    "epsilon_norm": 9.627598762512207,
    "prediction_norm": 8.14876651763916,
    "timestamp": "2025-07-14T12:35:19.716330"
  },
  {
    "epoch": 2944,
    "batch_idx": 160,
    "loss": 0.30354592204093933,
    "learning_rate": 0.00045607440172462963,
    "timestep_mean": 428.8125,
    "epsilon_norm": 7.752050399780273,
    "prediction_norm": 6.5733208656311035,
    "timestamp": "2025-07-14T12:35:19.721623"
  },
  {
    "epoch": 2944,
    "batch_idx": 192,
    "loss": 0.08154816180467606,
    "learning_rate": 0.00045607440172462963,
    "timestep_mean": 725.875,
    "epsilon_norm": 4.619140148162842,
    "prediction_norm": 4.569839000701904,
    "timestamp": "2025-07-14T12:35:19.726682"
  },
  {
    "epoch": 2945,
    "batch_idx": 0,
    "loss": 0.3877553343772888,
    "learning_rate": 0.00045685681710103945,
    "timestep_mean": 486.375,
    "epsilon_norm": 6.9094085693359375,
    "prediction_norm": 5.946578502655029,
    "timestamp": "2025-07-14T12:35:19.731905"
  },
  {
    "epoch": 2945,
    "batch_idx": 32,
    "loss": 0.41522619128227234,
    "learning_rate": 0.00045685681710103945,
    "timestep_mean": 524.625,
    "epsilon_norm": 9.443609237670898,
    "prediction_norm": 8.043841361999512,
    "timestamp": "2025-07-14T12:35:19.737138"
  },
  {
    "epoch": 2945,
    "batch_idx": 64,
    "loss": 0.30090010166168213,
    "learning_rate": 0.00045685681710103945,
    "timestep_mean": 503.8125,
    "epsilon_norm": 7.277950286865234,
    "prediction_norm": 6.161522388458252,
    "timestamp": "2025-07-14T12:35:19.742348"
  },
  {
    "epoch": 2945,
    "batch_idx": 96,
    "loss": 0.3328278064727783,
    "learning_rate": 0.00045685681710103945,
    "timestep_mean": 521.3125,
    "epsilon_norm": 9.465985298156738,
    "prediction_norm": 7.337752819061279,
    "timestamp": "2025-07-14T12:35:19.747596"
  },
  {
    "epoch": 2945,
    "batch_idx": 128,
    "loss": 0.656220018863678,
    "learning_rate": 0.00045685681710103945,
    "timestep_mean": 437.15625,
    "epsilon_norm": 8.351289749145508,
    "prediction_norm": 5.797370433807373,
    "timestamp": "2025-07-14T12:35:19.752812"
  },
  {
    "epoch": 2945,
    "batch_idx": 160,
    "loss": 0.47719693183898926,
    "learning_rate": 0.00045685681710103945,
    "timestep_mean": 424.0625,
    "epsilon_norm": 9.13447093963623,
    "prediction_norm": 7.471135139465332,
    "timestamp": "2025-07-14T12:35:19.757672"
  },
  {
    "epoch": 2945,
    "batch_idx": 192,
    "loss": 0.11760097742080688,
    "learning_rate": 0.00045685681710103945,
    "timestep_mean": 590.0,
    "epsilon_norm": 3.950878143310547,
    "prediction_norm": 3.724501609802246,
    "timestamp": "2025-07-14T12:35:19.762919"
  },
  {
    "epoch": 2946,
    "batch_idx": 0,
    "loss": 0.3688451945781708,
    "learning_rate": 0.00045763933892896435,
    "timestep_mean": 537.5625,
    "epsilon_norm": 8.787745475769043,
    "prediction_norm": 6.929105758666992,
    "timestamp": "2025-07-14T12:35:19.768221"
  },
  {
    "epoch": 2946,
    "batch_idx": 32,
    "loss": 0.19080686569213867,
    "learning_rate": 0.00045763933892896435,
    "timestep_mean": 491.375,
    "epsilon_norm": 7.438924789428711,
    "prediction_norm": 6.2319159507751465,
    "timestamp": "2025-07-14T12:35:19.773349"
  },
  {
    "epoch": 2946,
    "batch_idx": 64,
    "loss": 0.21981912851333618,
    "learning_rate": 0.00045763933892896435,
    "timestep_mean": 518.25,
    "epsilon_norm": 8.276206016540527,
    "prediction_norm": 7.7247748374938965,
    "timestamp": "2025-07-14T12:35:19.778527"
  },
  {
    "epoch": 2946,
    "batch_idx": 96,
    "loss": 0.2284996658563614,
    "learning_rate": 0.00045763933892896435,
    "timestep_mean": 478.25,
    "epsilon_norm": 7.735360145568848,
    "prediction_norm": 7.200433254241943,
    "timestamp": "2025-07-14T12:35:19.783955"
  },
  {
    "epoch": 2946,
    "batch_idx": 128,
    "loss": 0.2141333818435669,
    "learning_rate": 0.00045763933892896435,
    "timestep_mean": 553.15625,
    "epsilon_norm": 8.242319107055664,
    "prediction_norm": 6.961935520172119,
    "timestamp": "2025-07-14T12:35:19.789104"
  },
  {
    "epoch": 2946,
    "batch_idx": 160,
    "loss": 0.2921735644340515,
    "learning_rate": 0.00045763933892896435,
    "timestep_mean": 527.5,
    "epsilon_norm": 7.70729398727417,
    "prediction_norm": 6.340338230133057,
    "timestamp": "2025-07-14T12:35:19.794399"
  },
  {
    "epoch": 2946,
    "batch_idx": 192,
    "loss": 0.04971145838499069,
    "learning_rate": 0.00045763933892896435,
    "timestep_mean": 656.375,
    "epsilon_norm": 4.275179386138916,
    "prediction_norm": 4.271615982055664,
    "timestamp": "2025-07-14T12:35:19.799606"
  },
  {
    "epoch": 2947,
    "batch_idx": 0,
    "loss": 0.2752618193626404,
    "learning_rate": 0.00045842196527760956,
    "timestep_mean": 446.71875,
    "epsilon_norm": 7.556370258331299,
    "prediction_norm": 6.56424617767334,
    "timestamp": "2025-07-14T12:35:19.804837"
  },
  {
    "epoch": 2947,
    "batch_idx": 32,
    "loss": 0.34274041652679443,
    "learning_rate": 0.00045842196527760956,
    "timestep_mean": 505.90625,
    "epsilon_norm": 8.06804084777832,
    "prediction_norm": 6.200488567352295,
    "timestamp": "2025-07-14T12:35:19.810080"
  },
  {
    "epoch": 2947,
    "batch_idx": 64,
    "loss": 0.30410346388816833,
    "learning_rate": 0.00045842196527760956,
    "timestep_mean": 448.84375,
    "epsilon_norm": 7.821569442749023,
    "prediction_norm": 6.4467878341674805,
    "timestamp": "2025-07-14T12:35:19.815234"
  },
  {
    "epoch": 2947,
    "batch_idx": 96,
    "loss": 0.16234628856182098,
    "learning_rate": 0.00045842196527760956,
    "timestep_mean": 508.84375,
    "epsilon_norm": 8.13789176940918,
    "prediction_norm": 7.10626745223999,
    "timestamp": "2025-07-14T12:35:19.820498"
  },
  {
    "epoch": 2947,
    "batch_idx": 128,
    "loss": 0.35238245129585266,
    "learning_rate": 0.00045842196527760956,
    "timestep_mean": 527.1875,
    "epsilon_norm": 8.44962215423584,
    "prediction_norm": 6.286289215087891,
    "timestamp": "2025-07-14T12:35:19.825897"
  },
  {
    "epoch": 2947,
    "batch_idx": 160,
    "loss": 0.19737786054611206,
    "learning_rate": 0.00045842196527760956,
    "timestep_mean": 521.875,
    "epsilon_norm": 8.1908597946167,
    "prediction_norm": 7.652531623840332,
    "timestamp": "2025-07-14T12:35:19.831146"
  },
  {
    "epoch": 2947,
    "batch_idx": 192,
    "loss": 0.07163530588150024,
    "learning_rate": 0.00045842196527760956,
    "timestep_mean": 780.875,
    "epsilon_norm": 4.918907642364502,
    "prediction_norm": 4.162712574005127,
    "timestamp": "2025-07-14T12:35:19.836352"
  },
  {
    "epoch": 2948,
    "batch_idx": 0,
    "loss": 0.29790574312210083,
    "learning_rate": 0.00045920469421592276,
    "timestep_mean": 448.375,
    "epsilon_norm": 8.64832592010498,
    "prediction_norm": 7.3492045402526855,
    "timestamp": "2025-07-14T12:35:19.841663"
  },
  {
    "epoch": 2948,
    "batch_idx": 32,
    "loss": 0.1242518275976181,
    "learning_rate": 0.00045920469421592276,
    "timestep_mean": 622.5,
    "epsilon_norm": 7.6407647132873535,
    "prediction_norm": 7.820797920227051,
    "timestamp": "2025-07-14T12:35:19.847003"
  },
  {
    "epoch": 2948,
    "batch_idx": 64,
    "loss": 0.24657149612903595,
    "learning_rate": 0.00045920469421592276,
    "timestep_mean": 444.78125,
    "epsilon_norm": 8.248296737670898,
    "prediction_norm": 7.152194976806641,
    "timestamp": "2025-07-14T12:35:19.852251"
  },
  {
    "epoch": 2948,
    "batch_idx": 96,
    "loss": 0.1972711980342865,
    "learning_rate": 0.00045920469421592276,
    "timestep_mean": 468.46875,
    "epsilon_norm": 8.554264068603516,
    "prediction_norm": 7.042182445526123,
    "timestamp": "2025-07-14T12:35:19.857274"
  },
  {
    "epoch": 2948,
    "batch_idx": 128,
    "loss": 0.25550615787506104,
    "learning_rate": 0.00045920469421592276,
    "timestep_mean": 448.71875,
    "epsilon_norm": 7.913886547088623,
    "prediction_norm": 6.789611339569092,
    "timestamp": "2025-07-14T12:35:19.862084"
  },
  {
    "epoch": 2948,
    "batch_idx": 160,
    "loss": 0.26554274559020996,
    "learning_rate": 0.00045920469421592276,
    "timestep_mean": 502.5625,
    "epsilon_norm": 8.033601760864258,
    "prediction_norm": 7.538197994232178,
    "timestamp": "2025-07-14T12:35:19.866903"
  },
  {
    "epoch": 2948,
    "batch_idx": 192,
    "loss": 0.16022497415542603,
    "learning_rate": 0.00045920469421592276,
    "timestep_mean": 490.5,
    "epsilon_norm": 3.438899040222168,
    "prediction_norm": 3.1828043460845947,
    "timestamp": "2025-07-14T12:35:19.871893"
  },
  {
    "epoch": 2949,
    "batch_idx": 0,
    "loss": 0.26202327013015747,
    "learning_rate": 0.00045998752381259673,
    "timestep_mean": 437.0625,
    "epsilon_norm": 7.116046905517578,
    "prediction_norm": 5.659738063812256,
    "timestamp": "2025-07-14T12:35:19.876708"
  },
  {
    "epoch": 2949,
    "batch_idx": 32,
    "loss": 0.22369885444641113,
    "learning_rate": 0.00045998752381259673,
    "timestep_mean": 538.78125,
    "epsilon_norm": 7.7805681228637695,
    "prediction_norm": 6.629092693328857,
    "timestamp": "2025-07-14T12:35:19.881726"
  },
  {
    "epoch": 2949,
    "batch_idx": 64,
    "loss": 0.14225172996520996,
    "learning_rate": 0.00045998752381259673,
    "timestep_mean": 497.5,
    "epsilon_norm": 6.702772617340088,
    "prediction_norm": 6.072450160980225,
    "timestamp": "2025-07-14T12:35:19.886635"
  },
  {
    "epoch": 2949,
    "batch_idx": 96,
    "loss": 0.16691091656684875,
    "learning_rate": 0.00045998752381259673,
    "timestep_mean": 546.28125,
    "epsilon_norm": 7.269045829772949,
    "prediction_norm": 6.409688949584961,
    "timestamp": "2025-07-14T12:35:19.891519"
  },
  {
    "epoch": 2949,
    "batch_idx": 128,
    "loss": 0.17639270424842834,
    "learning_rate": 0.00045998752381259673,
    "timestep_mean": 547.4375,
    "epsilon_norm": 8.140816688537598,
    "prediction_norm": 6.95677375793457,
    "timestamp": "2025-07-14T12:35:19.896530"
  },
  {
    "epoch": 2949,
    "batch_idx": 160,
    "loss": 0.12226007878780365,
    "learning_rate": 0.00045998752381259673,
    "timestep_mean": 627.125,
    "epsilon_norm": 7.8025803565979,
    "prediction_norm": 7.509032249450684,
    "timestamp": "2025-07-14T12:35:19.901386"
  },
  {
    "epoch": 2949,
    "batch_idx": 192,
    "loss": 0.6626186966896057,
    "learning_rate": 0.00045998752381259673,
    "timestep_mean": 425.5,
    "epsilon_norm": 3.5843987464904785,
    "prediction_norm": 2.6841893196105957,
    "timestamp": "2025-07-14T12:35:19.906317"
  },
  {
    "epoch": 2950,
    "batch_idx": 0,
    "loss": 0.23218098282814026,
    "learning_rate": 0.00046077045213607857,
    "timestep_mean": 538.84375,
    "epsilon_norm": 8.100875854492188,
    "prediction_norm": 7.132857322692871,
    "timestamp": "2025-07-14T12:35:19.911264"
  },
  {
    "epoch": 2950,
    "batch_idx": 32,
    "loss": 0.4768502712249756,
    "learning_rate": 0.00046077045213607857,
    "timestep_mean": 389.90625,
    "epsilon_norm": 8.122482299804688,
    "prediction_norm": 5.8036017417907715,
    "timestamp": "2025-07-14T12:35:19.916209"
  },
  {
    "epoch": 2950,
    "batch_idx": 64,
    "loss": 0.21970245242118835,
    "learning_rate": 0.00046077045213607857,
    "timestep_mean": 515.4375,
    "epsilon_norm": 7.80385160446167,
    "prediction_norm": 6.9767327308654785,
    "timestamp": "2025-07-14T12:35:19.921077"
  },
  {
    "epoch": 2950,
    "batch_idx": 96,
    "loss": 0.19178426265716553,
    "learning_rate": 0.00046077045213607857,
    "timestep_mean": 504.0625,
    "epsilon_norm": 7.652960300445557,
    "prediction_norm": 6.420177936553955,
    "timestamp": "2025-07-14T12:35:19.926178"
  },
  {
    "epoch": 2950,
    "batch_idx": 128,
    "loss": 0.2581220865249634,
    "learning_rate": 0.00046077045213607857,
    "timestep_mean": 493.5,
    "epsilon_norm": 7.352241039276123,
    "prediction_norm": 5.974041938781738,
    "timestamp": "2025-07-14T12:35:19.931051"
  },
  {
    "epoch": 2950,
    "batch_idx": 160,
    "loss": 0.21458449959754944,
    "learning_rate": 0.00046077045213607857,
    "timestep_mean": 487.34375,
    "epsilon_norm": 7.440296649932861,
    "prediction_norm": 6.28354024887085,
    "timestamp": "2025-07-14T12:35:19.935914"
  },
  {
    "epoch": 2950,
    "batch_idx": 192,
    "loss": 0.603490948677063,
    "learning_rate": 0.00046077045213607857,
    "timestep_mean": 314.125,
    "epsilon_norm": 4.096111297607422,
    "prediction_norm": 2.6374030113220215,
    "timestamp": "2025-07-14T12:35:19.940705"
  },
  {
    "epoch": 2951,
    "batch_idx": 0,
    "loss": 0.3339986801147461,
    "learning_rate": 0.0004615534772545705,
    "timestep_mean": 445.90625,
    "epsilon_norm": 9.18792724609375,
    "prediction_norm": 7.519484519958496,
    "timestamp": "2025-07-14T12:35:19.945578"
  },
  {
    "epoch": 2951,
    "batch_idx": 32,
    "loss": 0.22362112998962402,
    "learning_rate": 0.0004615534772545705,
    "timestep_mean": 530.09375,
    "epsilon_norm": 7.6447272300720215,
    "prediction_norm": 6.786318778991699,
    "timestamp": "2025-07-14T12:35:19.950627"
  },
  {
    "epoch": 2951,
    "batch_idx": 64,
    "loss": 0.2684600353240967,
    "learning_rate": 0.0004615534772545705,
    "timestep_mean": 509.0625,
    "epsilon_norm": 7.340142726898193,
    "prediction_norm": 6.061273097991943,
    "timestamp": "2025-07-14T12:35:19.955516"
  },
  {
    "epoch": 2951,
    "batch_idx": 96,
    "loss": 0.31985217332839966,
    "learning_rate": 0.0004615534772545705,
    "timestep_mean": 509.5,
    "epsilon_norm": 8.273245811462402,
    "prediction_norm": 7.332018852233887,
    "timestamp": "2025-07-14T12:35:19.960465"
  },
  {
    "epoch": 2951,
    "batch_idx": 128,
    "loss": 0.19566470384597778,
    "learning_rate": 0.0004615534772545705,
    "timestep_mean": 443.4375,
    "epsilon_norm": 6.77827262878418,
    "prediction_norm": 6.773850440979004,
    "timestamp": "2025-07-14T12:35:19.965621"
  },
  {
    "epoch": 2951,
    "batch_idx": 160,
    "loss": 0.1282126009464264,
    "learning_rate": 0.0004615534772545705,
    "timestep_mean": 578.21875,
    "epsilon_norm": 7.432387351989746,
    "prediction_norm": 7.279216289520264,
    "timestamp": "2025-07-14T12:35:19.970673"
  },
  {
    "epoch": 2951,
    "batch_idx": 192,
    "loss": 0.22209663689136505,
    "learning_rate": 0.0004615534772545705,
    "timestep_mean": 364.125,
    "epsilon_norm": 3.3131730556488037,
    "prediction_norm": 3.120110511779785,
    "timestamp": "2025-07-14T12:35:19.975945"
  },
  {
    "epoch": 2952,
    "batch_idx": 0,
    "loss": 0.17239059507846832,
    "learning_rate": 0.00046233659723603503,
    "timestep_mean": 592.625,
    "epsilon_norm": 8.374259948730469,
    "prediction_norm": 7.89870023727417,
    "timestamp": "2025-07-14T12:35:19.981086"
  },
  {
    "epoch": 2952,
    "batch_idx": 32,
    "loss": 0.22859352827072144,
    "learning_rate": 0.00046233659723603503,
    "timestep_mean": 533.65625,
    "epsilon_norm": 7.428497791290283,
    "prediction_norm": 6.570047855377197,
    "timestamp": "2025-07-14T12:35:19.986227"
  },
  {
    "epoch": 2952,
    "batch_idx": 64,
    "loss": 0.29389989376068115,
    "learning_rate": 0.00046233659723603503,
    "timestep_mean": 466.625,
    "epsilon_norm": 7.588417053222656,
    "prediction_norm": 6.45626974105835,
    "timestamp": "2025-07-14T12:35:19.991405"
  },
  {
    "epoch": 2952,
    "batch_idx": 96,
    "loss": 0.19596073031425476,
    "learning_rate": 0.00046233659723603503,
    "timestep_mean": 527.3125,
    "epsilon_norm": 7.281318664550781,
    "prediction_norm": 6.060869216918945,
    "timestamp": "2025-07-14T12:35:19.996677"
  },
  {
    "epoch": 2952,
    "batch_idx": 128,
    "loss": 0.413284033536911,
    "learning_rate": 0.00046233659723603503,
    "timestep_mean": 503.6875,
    "epsilon_norm": 7.437617778778076,
    "prediction_norm": 5.480370998382568,
    "timestamp": "2025-07-14T12:35:20.001841"
  },
  {
    "epoch": 2952,
    "batch_idx": 160,
    "loss": 0.35137036442756653,
    "learning_rate": 0.00046233659723603503,
    "timestep_mean": 484.03125,
    "epsilon_norm": 8.300665855407715,
    "prediction_norm": 7.028388023376465,
    "timestamp": "2025-07-14T12:35:20.007036"
  },
  {
    "epoch": 2952,
    "batch_idx": 192,
    "loss": 0.4168958067893982,
    "learning_rate": 0.00046233659723603503,
    "timestep_mean": 624.375,
    "epsilon_norm": 4.255784511566162,
    "prediction_norm": 3.367233991622925,
    "timestamp": "2025-07-14T12:35:20.012242"
  },
  {
    "epoch": 2953,
    "batch_idx": 0,
    "loss": 0.1926603466272354,
    "learning_rate": 0.0004631198101482018,
    "timestep_mean": 479.78125,
    "epsilon_norm": 7.489234924316406,
    "prediction_norm": 6.635095119476318,
    "timestamp": "2025-07-14T12:35:20.017449"
  },
  {
    "epoch": 2953,
    "batch_idx": 32,
    "loss": 0.23570650815963745,
    "learning_rate": 0.0004631198101482018,
    "timestep_mean": 502.40625,
    "epsilon_norm": 7.275991916656494,
    "prediction_norm": 6.6580986976623535,
    "timestamp": "2025-07-14T12:35:20.022681"
  },
  {
    "epoch": 2953,
    "batch_idx": 64,
    "loss": 0.30348217487335205,
    "learning_rate": 0.0004631198101482018,
    "timestep_mean": 485.5,
    "epsilon_norm": 8.407902717590332,
    "prediction_norm": 6.7456817626953125,
    "timestamp": "2025-07-14T12:35:20.027889"
  },
  {
    "epoch": 2953,
    "batch_idx": 96,
    "loss": 0.3883105516433716,
    "learning_rate": 0.0004631198101482018,
    "timestep_mean": 386.9375,
    "epsilon_norm": 7.436578750610352,
    "prediction_norm": 5.2780842781066895,
    "timestamp": "2025-07-14T12:35:20.033074"
  },
  {
    "epoch": 2953,
    "batch_idx": 128,
    "loss": 0.18113642930984497,
    "learning_rate": 0.0004631198101482018,
    "timestep_mean": 494.53125,
    "epsilon_norm": 7.3987956047058105,
    "prediction_norm": 7.035022735595703,
    "timestamp": "2025-07-14T12:35:20.038450"
  },
  {
    "epoch": 2953,
    "batch_idx": 160,
    "loss": 0.1677088886499405,
    "learning_rate": 0.0004631198101482018,
    "timestep_mean": 519.8125,
    "epsilon_norm": 7.698200225830078,
    "prediction_norm": 6.530370712280273,
    "timestamp": "2025-07-14T12:35:20.043635"
  },
  {
    "epoch": 2953,
    "batch_idx": 192,
    "loss": 0.5192079544067383,
    "learning_rate": 0.0004631198101482018,
    "timestep_mean": 414.625,
    "epsilon_norm": 4.839073181152344,
    "prediction_norm": 3.8816566467285156,
    "timestamp": "2025-07-14T12:35:20.048774"
  },
  {
    "epoch": 2954,
    "batch_idx": 0,
    "loss": 0.4011314809322357,
    "learning_rate": 0.0004639031140585705,
    "timestep_mean": 601.21875,
    "epsilon_norm": 8.484572410583496,
    "prediction_norm": 6.107302665710449,
    "timestamp": "2025-07-14T12:35:20.054030"
  },
  {
    "epoch": 2954,
    "batch_idx": 32,
    "loss": 0.3003297448158264,
    "learning_rate": 0.0004639031140585705,
    "timestep_mean": 432.40625,
    "epsilon_norm": 7.723363399505615,
    "prediction_norm": 5.716816425323486,
    "timestamp": "2025-07-14T12:35:20.059277"
  },
  {
    "epoch": 2954,
    "batch_idx": 64,
    "loss": 0.3242120146751404,
    "learning_rate": 0.0004639031140585705,
    "timestep_mean": 447.75,
    "epsilon_norm": 7.963138580322266,
    "prediction_norm": 6.324161052703857,
    "timestamp": "2025-07-14T12:35:20.064453"
  },
  {
    "epoch": 2954,
    "batch_idx": 96,
    "loss": 0.389568030834198,
    "learning_rate": 0.0004639031140585705,
    "timestep_mean": 474.6875,
    "epsilon_norm": 7.4246602058410645,
    "prediction_norm": 5.28826379776001,
    "timestamp": "2025-07-14T12:35:20.069640"
  },
  {
    "epoch": 2954,
    "batch_idx": 128,
    "loss": 0.2869222164154053,
    "learning_rate": 0.0004639031140585705,
    "timestep_mean": 478.0,
    "epsilon_norm": 7.14135217666626,
    "prediction_norm": 5.912757873535156,
    "timestamp": "2025-07-14T12:35:20.074638"
  },
  {
    "epoch": 2954,
    "batch_idx": 160,
    "loss": 0.2581796944141388,
    "learning_rate": 0.0004639031140585705,
    "timestep_mean": 469.40625,
    "epsilon_norm": 7.831725120544434,
    "prediction_norm": 6.7329487800598145,
    "timestamp": "2025-07-14T12:35:20.079707"
  },
  {
    "epoch": 2954,
    "batch_idx": 192,
    "loss": 0.10600192844867706,
    "learning_rate": 0.0004639031140585705,
    "timestep_mean": 414.125,
    "epsilon_norm": 3.0609755516052246,
    "prediction_norm": 2.4488353729248047,
    "timestamp": "2025-07-14T12:35:20.084854"
  },
  {
    "epoch": 2955,
    "batch_idx": 0,
    "loss": 0.26190894842147827,
    "learning_rate": 0.0004646865070344179,
    "timestep_mean": 532.25,
    "epsilon_norm": 9.073163986206055,
    "prediction_norm": 7.57523250579834,
    "timestamp": "2025-07-14T12:35:20.090089"
  },
  {
    "epoch": 2955,
    "batch_idx": 32,
    "loss": 0.3578043580055237,
    "learning_rate": 0.0004646865070344179,
    "timestep_mean": 465.71875,
    "epsilon_norm": 7.779310703277588,
    "prediction_norm": 6.639773845672607,
    "timestamp": "2025-07-14T12:35:20.095265"
  },
  {
    "epoch": 2955,
    "batch_idx": 64,
    "loss": 0.23234474658966064,
    "learning_rate": 0.0004646865070344179,
    "timestep_mean": 586.75,
    "epsilon_norm": 8.184940338134766,
    "prediction_norm": 6.821191787719727,
    "timestamp": "2025-07-14T12:35:20.100626"
  },
  {
    "epoch": 2955,
    "batch_idx": 96,
    "loss": 0.20251698791980743,
    "learning_rate": 0.0004646865070344179,
    "timestep_mean": 572.5625,
    "epsilon_norm": 7.929272174835205,
    "prediction_norm": 7.343527317047119,
    "timestamp": "2025-07-14T12:35:20.105846"
  },
  {
    "epoch": 2955,
    "batch_idx": 128,
    "loss": 0.4583931565284729,
    "learning_rate": 0.0004646865070344179,
    "timestep_mean": 521.6875,
    "epsilon_norm": 8.36313533782959,
    "prediction_norm": 6.747030258178711,
    "timestamp": "2025-07-14T12:35:20.111080"
  },
  {
    "epoch": 2955,
    "batch_idx": 160,
    "loss": 0.33904141187667847,
    "learning_rate": 0.0004646865070344179,
    "timestep_mean": 464.46875,
    "epsilon_norm": 7.687126636505127,
    "prediction_norm": 6.150412559509277,
    "timestamp": "2025-07-14T12:35:20.116364"
  },
  {
    "epoch": 2955,
    "batch_idx": 192,
    "loss": 0.25488945841789246,
    "learning_rate": 0.0004646865070344179,
    "timestep_mean": 363.25,
    "epsilon_norm": 3.2397096157073975,
    "prediction_norm": 2.612417697906494,
    "timestamp": "2025-07-14T12:35:20.121556"
  },
  {
    "epoch": 2956,
    "batch_idx": 0,
    "loss": 0.19764012098312378,
    "learning_rate": 0.0004654699871427982,
    "timestep_mean": 529.40625,
    "epsilon_norm": 7.7091898918151855,
    "prediction_norm": 6.956932544708252,
    "timestamp": "2025-07-14T12:35:20.126855"
  },
  {
    "epoch": 2956,
    "batch_idx": 32,
    "loss": 0.18018397688865662,
    "learning_rate": 0.0004654699871427982,
    "timestep_mean": 534.78125,
    "epsilon_norm": 7.668220520019531,
    "prediction_norm": 6.696829795837402,
    "timestamp": "2025-07-14T12:35:20.132010"
  },
  {
    "epoch": 2956,
    "batch_idx": 64,
    "loss": 0.14493826031684875,
    "learning_rate": 0.0004654699871427982,
    "timestep_mean": 488.28125,
    "epsilon_norm": 7.460716247558594,
    "prediction_norm": 6.240663051605225,
    "timestamp": "2025-07-14T12:35:20.137231"
  },
  {
    "epoch": 2956,
    "batch_idx": 96,
    "loss": 0.39950111508369446,
    "learning_rate": 0.0004654699871427982,
    "timestep_mean": 450.96875,
    "epsilon_norm": 7.308255195617676,
    "prediction_norm": 5.176436424255371,
    "timestamp": "2025-07-14T12:35:20.142332"
  },
  {
    "epoch": 2956,
    "batch_idx": 128,
    "loss": 0.27437257766723633,
    "learning_rate": 0.0004654699871427982,
    "timestep_mean": 422.3125,
    "epsilon_norm": 7.291755676269531,
    "prediction_norm": 6.815122127532959,
    "timestamp": "2025-07-14T12:35:20.147527"
  },
  {
    "epoch": 2956,
    "batch_idx": 160,
    "loss": 0.27501413226127625,
    "learning_rate": 0.0004654699871427982,
    "timestep_mean": 453.03125,
    "epsilon_norm": 7.734771251678467,
    "prediction_norm": 6.445143699645996,
    "timestamp": "2025-07-14T12:35:20.152724"
  },
  {
    "epoch": 2956,
    "batch_idx": 192,
    "loss": 0.6770807504653931,
    "learning_rate": 0.0004654699871427982,
    "timestep_mean": 230.125,
    "epsilon_norm": 2.964404582977295,
    "prediction_norm": 1.4735496044158936,
    "timestamp": "2025-07-14T12:35:20.157918"
  },
  {
    "epoch": 2957,
    "batch_idx": 0,
    "loss": 0.5022616982460022,
    "learning_rate": 0.00046625355245055283,
    "timestep_mean": 430.0625,
    "epsilon_norm": 7.581923484802246,
    "prediction_norm": 5.93997049331665,
    "timestamp": "2025-07-14T12:35:20.163257"
  },
  {
    "epoch": 2957,
    "batch_idx": 32,
    "loss": 0.28970515727996826,
    "learning_rate": 0.00046625355245055283,
    "timestep_mean": 529.625,
    "epsilon_norm": 8.948412895202637,
    "prediction_norm": 7.468149662017822,
    "timestamp": "2025-07-14T12:35:20.168423"
  },
  {
    "epoch": 2957,
    "batch_idx": 64,
    "loss": 0.36465901136398315,
    "learning_rate": 0.00046625355245055283,
    "timestep_mean": 397.15625,
    "epsilon_norm": 7.621685981750488,
    "prediction_norm": 6.518438339233398,
    "timestamp": "2025-07-14T12:35:20.173688"
  },
  {
    "epoch": 2957,
    "batch_idx": 96,
    "loss": 0.17793984711170197,
    "learning_rate": 0.00046625355245055283,
    "timestep_mean": 493.5,
    "epsilon_norm": 8.476325988769531,
    "prediction_norm": 8.375962257385254,
    "timestamp": "2025-07-14T12:35:20.178708"
  },
  {
    "epoch": 2957,
    "batch_idx": 128,
    "loss": 0.3030662536621094,
    "learning_rate": 0.00046625355245055283,
    "timestep_mean": 498.1875,
    "epsilon_norm": 7.231207370758057,
    "prediction_norm": 6.234782695770264,
    "timestamp": "2025-07-14T12:35:20.183559"
  },
  {
    "epoch": 2957,
    "batch_idx": 160,
    "loss": 0.1688784509897232,
    "learning_rate": 0.00046625355245055283,
    "timestep_mean": 569.03125,
    "epsilon_norm": 8.802812576293945,
    "prediction_norm": 7.966907024383545,
    "timestamp": "2025-07-14T12:35:20.188449"
  },
  {
    "epoch": 2957,
    "batch_idx": 192,
    "loss": 0.2020488679409027,
    "learning_rate": 0.00046625355245055283,
    "timestep_mean": 562.375,
    "epsilon_norm": 4.075130462646484,
    "prediction_norm": 3.927081823348999,
    "timestamp": "2025-07-14T12:35:20.193310"
  },
  {
    "epoch": 2958,
    "batch_idx": 0,
    "loss": 0.33099251985549927,
    "learning_rate": 0.00046703720102431244,
    "timestep_mean": 516.90625,
    "epsilon_norm": 8.57001781463623,
    "prediction_norm": 7.609907627105713,
    "timestamp": "2025-07-14T12:35:20.198215"
  },
  {
    "epoch": 2958,
    "batch_idx": 32,
    "loss": 0.28383615612983704,
    "learning_rate": 0.00046703720102431244,
    "timestep_mean": 460.625,
    "epsilon_norm": 7.826822757720947,
    "prediction_norm": 6.186159610748291,
    "timestamp": "2025-07-14T12:35:20.203075"
  },
  {
    "epoch": 2958,
    "batch_idx": 64,
    "loss": 0.18140238523483276,
    "learning_rate": 0.00046703720102431244,
    "timestep_mean": 586.46875,
    "epsilon_norm": 7.610605716705322,
    "prediction_norm": 7.074911594390869,
    "timestamp": "2025-07-14T12:35:20.208182"
  },
  {
    "epoch": 2958,
    "batch_idx": 96,
    "loss": 0.3295302987098694,
    "learning_rate": 0.00046703720102431244,
    "timestep_mean": 489.65625,
    "epsilon_norm": 8.161697387695312,
    "prediction_norm": 6.580591678619385,
    "timestamp": "2025-07-14T12:35:20.213416"
  },
  {
    "epoch": 2958,
    "batch_idx": 128,
    "loss": 0.24982032179832458,
    "learning_rate": 0.00046703720102431244,
    "timestep_mean": 487.5,
    "epsilon_norm": 7.657015800476074,
    "prediction_norm": 7.563459873199463,
    "timestamp": "2025-07-14T12:35:20.218521"
  },
  {
    "epoch": 2958,
    "batch_idx": 160,
    "loss": 0.3781784772872925,
    "learning_rate": 0.00046703720102431244,
    "timestep_mean": 456.4375,
    "epsilon_norm": 8.735815048217773,
    "prediction_norm": 7.2676682472229,
    "timestamp": "2025-07-14T12:35:20.223680"
  },
  {
    "epoch": 2958,
    "batch_idx": 192,
    "loss": 0.24721169471740723,
    "learning_rate": 0.00046703720102431244,
    "timestep_mean": 507.875,
    "epsilon_norm": 3.6128692626953125,
    "prediction_norm": 2.691223621368408,
    "timestamp": "2025-07-14T12:35:20.228722"
  },
  {
    "epoch": 2959,
    "batch_idx": 0,
    "loss": 0.27700409293174744,
    "learning_rate": 0.00046782093093050144,
    "timestep_mean": 396.09375,
    "epsilon_norm": 8.537152290344238,
    "prediction_norm": 7.33444356918335,
    "timestamp": "2025-07-14T12:35:20.233794"
  },
  {
    "epoch": 2959,
    "batch_idx": 32,
    "loss": 0.3334738612174988,
    "learning_rate": 0.00046782093093050144,
    "timestep_mean": 496.03125,
    "epsilon_norm": 7.598211765289307,
    "prediction_norm": 6.0142669677734375,
    "timestamp": "2025-07-14T12:35:20.238953"
  },
  {
    "epoch": 2959,
    "batch_idx": 64,
    "loss": 0.3378656506538391,
    "learning_rate": 0.00046782093093050144,
    "timestep_mean": 505.84375,
    "epsilon_norm": 8.525232315063477,
    "prediction_norm": 7.187363147735596,
    "timestamp": "2025-07-14T12:35:20.244082"
  },
  {
    "epoch": 2959,
    "batch_idx": 96,
    "loss": 0.27852970361709595,
    "learning_rate": 0.00046782093093050144,
    "timestep_mean": 518.65625,
    "epsilon_norm": 8.902474403381348,
    "prediction_norm": 7.257601261138916,
    "timestamp": "2025-07-14T12:35:20.249242"
  },
  {
    "epoch": 2959,
    "batch_idx": 128,
    "loss": 0.22453556954860687,
    "learning_rate": 0.00046782093093050144,
    "timestep_mean": 495.71875,
    "epsilon_norm": 8.61784839630127,
    "prediction_norm": 7.481376647949219,
    "timestamp": "2025-07-14T12:35:20.254403"
  },
  {
    "epoch": 2959,
    "batch_idx": 160,
    "loss": 0.13482415676116943,
    "learning_rate": 0.00046782093093050144,
    "timestep_mean": 586.5625,
    "epsilon_norm": 7.445696830749512,
    "prediction_norm": 6.791595935821533,
    "timestamp": "2025-07-14T12:35:20.259524"
  },
  {
    "epoch": 2959,
    "batch_idx": 192,
    "loss": 0.255998432636261,
    "learning_rate": 0.00046782093093050144,
    "timestep_mean": 351.375,
    "epsilon_norm": 4.091495037078857,
    "prediction_norm": 3.376333475112915,
    "timestamp": "2025-07-14T12:35:20.264481"
  },
  {
    "epoch": 2960,
    "batch_idx": 0,
    "loss": 0.5435285568237305,
    "learning_rate": 0.0004686047402353441,
    "timestep_mean": 403.8125,
    "epsilon_norm": 9.235152244567871,
    "prediction_norm": 6.521090984344482,
    "timestamp": "2025-07-14T12:35:20.269742"
  },
  {
    "epoch": 2960,
    "batch_idx": 32,
    "loss": 0.2630443572998047,
    "learning_rate": 0.0004686047402353441,
    "timestep_mean": 498.40625,
    "epsilon_norm": 7.747584342956543,
    "prediction_norm": 6.638942241668701,
    "timestamp": "2025-07-14T12:35:20.275013"
  },
  {
    "epoch": 2960,
    "batch_idx": 64,
    "loss": 0.3997393250465393,
    "learning_rate": 0.0004686047402353441,
    "timestep_mean": 539.6875,
    "epsilon_norm": 8.208578109741211,
    "prediction_norm": 6.560158729553223,
    "timestamp": "2025-07-14T12:35:20.280337"
  },
  {
    "epoch": 2960,
    "batch_idx": 96,
    "loss": 0.34505102038383484,
    "learning_rate": 0.0004686047402353441,
    "timestep_mean": 401.40625,
    "epsilon_norm": 8.183991432189941,
    "prediction_norm": 7.162397384643555,
    "timestamp": "2025-07-14T12:35:20.285568"
  },
  {
    "epoch": 2960,
    "batch_idx": 128,
    "loss": 0.2890145182609558,
    "learning_rate": 0.0004686047402353441,
    "timestep_mean": 501.9375,
    "epsilon_norm": 8.35974407196045,
    "prediction_norm": 6.927242755889893,
    "timestamp": "2025-07-14T12:35:20.290764"
  },
  {
    "epoch": 2960,
    "batch_idx": 160,
    "loss": 0.19255927205085754,
    "learning_rate": 0.0004686047402353441,
    "timestep_mean": 515.78125,
    "epsilon_norm": 6.6201934814453125,
    "prediction_norm": 5.881331443786621,
    "timestamp": "2025-07-14T12:35:20.295958"
  },
  {
    "epoch": 2960,
    "batch_idx": 192,
    "loss": 0.7475595474243164,
    "learning_rate": 0.0004686047402353441,
    "timestep_mean": 464.25,
    "epsilon_norm": 4.612157344818115,
    "prediction_norm": 2.6950912475585938,
    "timestamp": "2025-07-14T12:35:20.301136"
  },
  {
    "epoch": 2961,
    "batch_idx": 0,
    "loss": 0.25663360953330994,
    "learning_rate": 0.0004693886270048703,
    "timestep_mean": 430.96875,
    "epsilon_norm": 7.196201324462891,
    "prediction_norm": 5.879426002502441,
    "timestamp": "2025-07-14T12:35:20.306184"
  },
  {
    "epoch": 2961,
    "batch_idx": 32,
    "loss": 0.3165016770362854,
    "learning_rate": 0.0004693886270048703,
    "timestep_mean": 423.0625,
    "epsilon_norm": 9.03332233428955,
    "prediction_norm": 7.968350410461426,
    "timestamp": "2025-07-14T12:35:20.311478"
  },
  {
    "epoch": 2961,
    "batch_idx": 64,
    "loss": 0.30762091279029846,
    "learning_rate": 0.0004693886270048703,
    "timestep_mean": 548.0,
    "epsilon_norm": 8.502781867980957,
    "prediction_norm": 7.426164150238037,
    "timestamp": "2025-07-14T12:35:20.316737"
  },
  {
    "epoch": 2961,
    "batch_idx": 96,
    "loss": 0.1855808049440384,
    "learning_rate": 0.0004693886270048703,
    "timestep_mean": 554.875,
    "epsilon_norm": 8.281332969665527,
    "prediction_norm": 7.12807035446167,
    "timestamp": "2025-07-14T12:35:20.321831"
  },
  {
    "epoch": 2961,
    "batch_idx": 128,
    "loss": 0.15488438308238983,
    "learning_rate": 0.0004693886270048703,
    "timestep_mean": 604.03125,
    "epsilon_norm": 6.993838787078857,
    "prediction_norm": 6.763953685760498,
    "timestamp": "2025-07-14T12:35:20.326941"
  },
  {
    "epoch": 2961,
    "batch_idx": 160,
    "loss": 0.18027544021606445,
    "learning_rate": 0.0004693886270048703,
    "timestep_mean": 486.90625,
    "epsilon_norm": 6.641837120056152,
    "prediction_norm": 5.587469100952148,
    "timestamp": "2025-07-14T12:35:20.332090"
  },
  {
    "epoch": 2961,
    "batch_idx": 192,
    "loss": 0.21785099804401398,
    "learning_rate": 0.0004693886270048703,
    "timestep_mean": 596.125,
    "epsilon_norm": 3.903034210205078,
    "prediction_norm": 3.4729771614074707,
    "timestamp": "2025-07-14T12:35:20.337311"
  },
  {
    "epoch": 2962,
    "batch_idx": 0,
    "loss": 0.17976874113082886,
    "learning_rate": 0.0004701725893049158,
    "timestep_mean": 596.875,
    "epsilon_norm": 8.37509536743164,
    "prediction_norm": 8.103321075439453,
    "timestamp": "2025-07-14T12:35:20.342629"
  },
  {
    "epoch": 2962,
    "batch_idx": 32,
    "loss": 0.22995609045028687,
    "learning_rate": 0.0004701725893049158,
    "timestep_mean": 497.75,
    "epsilon_norm": 8.589900970458984,
    "prediction_norm": 7.226408958435059,
    "timestamp": "2025-07-14T12:35:20.347627"
  },
  {
    "epoch": 2962,
    "batch_idx": 64,
    "loss": 0.27139487862586975,
    "learning_rate": 0.0004701725893049158,
    "timestep_mean": 490.09375,
    "epsilon_norm": 8.745498657226562,
    "prediction_norm": 7.921318054199219,
    "timestamp": "2025-07-14T12:35:20.352681"
  },
  {
    "epoch": 2962,
    "batch_idx": 96,
    "loss": 0.18276941776275635,
    "learning_rate": 0.0004701725893049158,
    "timestep_mean": 539.3125,
    "epsilon_norm": 7.446220397949219,
    "prediction_norm": 6.5331830978393555,
    "timestamp": "2025-07-14T12:35:20.357837"
  },
  {
    "epoch": 2962,
    "batch_idx": 128,
    "loss": 0.19083724915981293,
    "learning_rate": 0.0004701725893049158,
    "timestep_mean": 507.21875,
    "epsilon_norm": 6.762757301330566,
    "prediction_norm": 6.351639270782471,
    "timestamp": "2025-07-14T12:35:20.363014"
  },
  {
    "epoch": 2962,
    "batch_idx": 160,
    "loss": 0.26543647050857544,
    "learning_rate": 0.0004701725893049158,
    "timestep_mean": 465.96875,
    "epsilon_norm": 8.087113380432129,
    "prediction_norm": 6.659595966339111,
    "timestamp": "2025-07-14T12:35:20.368163"
  },
  {
    "epoch": 2962,
    "batch_idx": 192,
    "loss": 0.14980104565620422,
    "learning_rate": 0.0004701725893049158,
    "timestep_mean": 464.625,
    "epsilon_norm": 2.456848382949829,
    "prediction_norm": 1.7808091640472412,
    "timestamp": "2025-07-14T12:35:20.373482"
  },
  {
    "epoch": 2963,
    "batch_idx": 0,
    "loss": 0.20828011631965637,
    "learning_rate": 0.00047095662520113215,
    "timestep_mean": 540.96875,
    "epsilon_norm": 8.539678573608398,
    "prediction_norm": 7.563217639923096,
    "timestamp": "2025-07-14T12:35:20.378747"
  },
  {
    "epoch": 2963,
    "batch_idx": 32,
    "loss": 0.30862176418304443,
    "learning_rate": 0.00047095662520113215,
    "timestep_mean": 559.15625,
    "epsilon_norm": 7.59498405456543,
    "prediction_norm": 5.611134052276611,
    "timestamp": "2025-07-14T12:35:20.384174"
  },
  {
    "epoch": 2963,
    "batch_idx": 64,
    "loss": 0.3688628673553467,
    "learning_rate": 0.00047095662520113215,
    "timestep_mean": 506.375,
    "epsilon_norm": 8.206779479980469,
    "prediction_norm": 6.749758243560791,
    "timestamp": "2025-07-14T12:35:20.389417"
  },
  {
    "epoch": 2963,
    "batch_idx": 96,
    "loss": 0.1856522262096405,
    "learning_rate": 0.00047095662520113215,
    "timestep_mean": 587.5625,
    "epsilon_norm": 8.273608207702637,
    "prediction_norm": 7.255740642547607,
    "timestamp": "2025-07-14T12:35:20.394666"
  },
  {
    "epoch": 2963,
    "batch_idx": 128,
    "loss": 0.1874171644449234,
    "learning_rate": 0.00047095662520113215,
    "timestep_mean": 490.0,
    "epsilon_norm": 7.201178550720215,
    "prediction_norm": 6.372284412384033,
    "timestamp": "2025-07-14T12:35:20.399781"
  },
  {
    "epoch": 2963,
    "batch_idx": 160,
    "loss": 0.301398903131485,
    "learning_rate": 0.00047095662520113215,
    "timestep_mean": 559.65625,
    "epsilon_norm": 8.02126693725586,
    "prediction_norm": 6.4357452392578125,
    "timestamp": "2025-07-14T12:35:20.404986"
  },
  {
    "epoch": 2963,
    "batch_idx": 192,
    "loss": 0.05682806670665741,
    "learning_rate": 0.00047095662520113215,
    "timestep_mean": 632.625,
    "epsilon_norm": 2.9761717319488525,
    "prediction_norm": 2.7673957347869873,
    "timestamp": "2025-07-14T12:35:20.409860"
  },
  {
    "epoch": 2964,
    "batch_idx": 0,
    "loss": 0.19226565957069397,
    "learning_rate": 0.0004717407327589888,
    "timestep_mean": 559.90625,
    "epsilon_norm": 8.250650405883789,
    "prediction_norm": 7.165369987487793,
    "timestamp": "2025-07-14T12:35:20.414797"
  },
  {
    "epoch": 2964,
    "batch_idx": 32,
    "loss": 0.19997046887874603,
    "learning_rate": 0.0004717407327589888,
    "timestep_mean": 532.375,
    "epsilon_norm": 8.246508598327637,
    "prediction_norm": 7.323392868041992,
    "timestamp": "2025-07-14T12:35:20.420102"
  },
  {
    "epoch": 2964,
    "batch_idx": 64,
    "loss": 0.32857975363731384,
    "learning_rate": 0.0004717407327589888,
    "timestep_mean": 498.5,
    "epsilon_norm": 8.500916481018066,
    "prediction_norm": 6.749323844909668,
    "timestamp": "2025-07-14T12:35:20.425461"
  },
  {
    "epoch": 2964,
    "batch_idx": 96,
    "loss": 0.29623767733573914,
    "learning_rate": 0.0004717407327589888,
    "timestep_mean": 427.6875,
    "epsilon_norm": 8.104557991027832,
    "prediction_norm": 6.702045440673828,
    "timestamp": "2025-07-14T12:35:20.430668"
  },
  {
    "epoch": 2964,
    "batch_idx": 128,
    "loss": 0.18593266606330872,
    "learning_rate": 0.0004717407327589888,
    "timestep_mean": 525.15625,
    "epsilon_norm": 7.563093185424805,
    "prediction_norm": 7.387110710144043,
    "timestamp": "2025-07-14T12:35:20.435909"
  },
  {
    "epoch": 2964,
    "batch_idx": 160,
    "loss": 0.4088587462902069,
    "learning_rate": 0.0004717407327589888,
    "timestep_mean": 548.53125,
    "epsilon_norm": 8.557944297790527,
    "prediction_norm": 7.646717548370361,
    "timestamp": "2025-07-14T12:35:20.440965"
  },
  {
    "epoch": 2964,
    "batch_idx": 192,
    "loss": 0.10663922876119614,
    "learning_rate": 0.0004717407327589888,
    "timestep_mean": 515.625,
    "epsilon_norm": 3.7951011657714844,
    "prediction_norm": 3.7843973636627197,
    "timestamp": "2025-07-14T12:35:20.446312"
  },
  {
    "epoch": 2965,
    "batch_idx": 0,
    "loss": 0.23204997181892395,
    "learning_rate": 0.0004725249100437781,
    "timestep_mean": 553.34375,
    "epsilon_norm": 7.226472854614258,
    "prediction_norm": 6.505162239074707,
    "timestamp": "2025-07-14T12:35:20.451468"
  },
  {
    "epoch": 2965,
    "batch_idx": 32,
    "loss": 0.3028174638748169,
    "learning_rate": 0.0004725249100437781,
    "timestep_mean": 479.5,
    "epsilon_norm": 8.45064640045166,
    "prediction_norm": 7.4847307205200195,
    "timestamp": "2025-07-14T12:35:20.456727"
  },
  {
    "epoch": 2965,
    "batch_idx": 64,
    "loss": 0.19287855923175812,
    "learning_rate": 0.0004725249100437781,
    "timestep_mean": 527.09375,
    "epsilon_norm": 7.44333553314209,
    "prediction_norm": 6.4725260734558105,
    "timestamp": "2025-07-14T12:35:20.461967"
  },
  {
    "epoch": 2965,
    "batch_idx": 96,
    "loss": 0.23290906846523285,
    "learning_rate": 0.0004725249100437781,
    "timestep_mean": 503.78125,
    "epsilon_norm": 7.568780422210693,
    "prediction_norm": 7.212892055511475,
    "timestamp": "2025-07-14T12:35:20.467340"
  },
  {
    "epoch": 2965,
    "batch_idx": 128,
    "loss": 0.27975210547447205,
    "learning_rate": 0.0004725249100437781,
    "timestep_mean": 507.375,
    "epsilon_norm": 7.514129161834717,
    "prediction_norm": 6.457057952880859,
    "timestamp": "2025-07-14T12:35:20.472528"
  },
  {
    "epoch": 2965,
    "batch_idx": 160,
    "loss": 0.4184654951095581,
    "learning_rate": 0.0004725249100437781,
    "timestep_mean": 426.9375,
    "epsilon_norm": 8.174249649047852,
    "prediction_norm": 6.393367767333984,
    "timestamp": "2025-07-14T12:35:20.477718"
  },
  {
    "epoch": 2965,
    "batch_idx": 192,
    "loss": 0.5557408332824707,
    "learning_rate": 0.0004725249100437781,
    "timestep_mean": 516.125,
    "epsilon_norm": 4.585586071014404,
    "prediction_norm": 3.1208364963531494,
    "timestamp": "2025-07-14T12:35:20.482857"
  },
  {
    "epoch": 2966,
    "batch_idx": 0,
    "loss": 0.40361660718917847,
    "learning_rate": 0.0004733091551206207,
    "timestep_mean": 603.84375,
    "epsilon_norm": 7.5976176261901855,
    "prediction_norm": 6.007719993591309,
    "timestamp": "2025-07-14T12:35:20.488080"
  },
  {
    "epoch": 2966,
    "batch_idx": 32,
    "loss": 0.36719268560409546,
    "learning_rate": 0.0004733091551206207,
    "timestep_mean": 521.0625,
    "epsilon_norm": 8.319474220275879,
    "prediction_norm": 6.777656555175781,
    "timestamp": "2025-07-14T12:35:20.493297"
  },
  {
    "epoch": 2966,
    "batch_idx": 64,
    "loss": 0.22585618495941162,
    "learning_rate": 0.0004733091551206207,
    "timestep_mean": 501.15625,
    "epsilon_norm": 8.246185302734375,
    "prediction_norm": 6.973094940185547,
    "timestamp": "2025-07-14T12:35:20.498411"
  },
  {
    "epoch": 2966,
    "batch_idx": 96,
    "loss": 0.40591567754745483,
    "learning_rate": 0.0004733091551206207,
    "timestep_mean": 540.90625,
    "epsilon_norm": 8.075461387634277,
    "prediction_norm": 6.4037346839904785,
    "timestamp": "2025-07-14T12:35:20.503320"
  },
  {
    "epoch": 2966,
    "batch_idx": 128,
    "loss": 0.3087296485900879,
    "learning_rate": 0.0004733091551206207,
    "timestep_mean": 482.625,
    "epsilon_norm": 8.029031753540039,
    "prediction_norm": 6.760986804962158,
    "timestamp": "2025-07-14T12:35:20.508187"
  },
  {
    "epoch": 2966,
    "batch_idx": 160,
    "loss": 0.2394036501646042,
    "learning_rate": 0.0004733091551206207,
    "timestep_mean": 454.46875,
    "epsilon_norm": 7.505288124084473,
    "prediction_norm": 6.710402488708496,
    "timestamp": "2025-07-14T12:35:20.513174"
  },
  {
    "epoch": 2966,
    "batch_idx": 192,
    "loss": 0.10031908005475998,
    "learning_rate": 0.0004733091551206207,
    "timestep_mean": 611.375,
    "epsilon_norm": 5.453220844268799,
    "prediction_norm": 5.179184436798096,
    "timestamp": "2025-07-14T12:35:20.518055"
  },
  {
    "epoch": 2967,
    "batch_idx": 0,
    "loss": 0.24774491786956787,
    "learning_rate": 0.00047409346605446984,
    "timestep_mean": 471.65625,
    "epsilon_norm": 7.53014612197876,
    "prediction_norm": 6.439900875091553,
    "timestamp": "2025-07-14T12:35:20.522966"
  },
  {
    "epoch": 2967,
    "batch_idx": 32,
    "loss": 0.28798848390579224,
    "learning_rate": 0.00047409346605446984,
    "timestep_mean": 410.1875,
    "epsilon_norm": 7.317266941070557,
    "prediction_norm": 5.6584553718566895,
    "timestamp": "2025-07-14T12:35:20.528095"
  },
  {
    "epoch": 2967,
    "batch_idx": 64,
    "loss": 0.21333888173103333,
    "learning_rate": 0.00047409346605446984,
    "timestep_mean": 520.90625,
    "epsilon_norm": 6.8305230140686035,
    "prediction_norm": 5.613940715789795,
    "timestamp": "2025-07-14T12:35:20.533174"
  },
  {
    "epoch": 2967,
    "batch_idx": 96,
    "loss": 0.2945267856121063,
    "learning_rate": 0.00047409346605446984,
    "timestep_mean": 524.0625,
    "epsilon_norm": 7.0889739990234375,
    "prediction_norm": 6.449813365936279,
    "timestamp": "2025-07-14T12:35:20.538458"
  },
  {
    "epoch": 2967,
    "batch_idx": 128,
    "loss": 0.21997317671775818,
    "learning_rate": 0.00047409346605446984,
    "timestep_mean": 458.59375,
    "epsilon_norm": 7.789290904998779,
    "prediction_norm": 7.015056610107422,
    "timestamp": "2025-07-14T12:35:20.543532"
  },
  {
    "epoch": 2967,
    "batch_idx": 160,
    "loss": 0.3215697407722473,
    "learning_rate": 0.00047409346605446984,
    "timestep_mean": 538.90625,
    "epsilon_norm": 7.613544464111328,
    "prediction_norm": 6.297728538513184,
    "timestamp": "2025-07-14T12:35:20.548693"
  },
  {
    "epoch": 2967,
    "batch_idx": 192,
    "loss": 0.2577041983604431,
    "learning_rate": 0.00047409346605446984,
    "timestep_mean": 568.75,
    "epsilon_norm": 3.055670738220215,
    "prediction_norm": 2.1711156368255615,
    "timestamp": "2025-07-14T12:35:20.553815"
  },
  {
    "epoch": 2968,
    "batch_idx": 0,
    "loss": 0.1948610544204712,
    "learning_rate": 0.00047487784091011656,
    "timestep_mean": 545.71875,
    "epsilon_norm": 8.663798332214355,
    "prediction_norm": 7.588559150695801,
    "timestamp": "2025-07-14T12:35:20.559007"
  },
  {
    "epoch": 2968,
    "batch_idx": 32,
    "loss": 0.20987451076507568,
    "learning_rate": 0.00047487784091011656,
    "timestep_mean": 552.15625,
    "epsilon_norm": 7.435855865478516,
    "prediction_norm": 6.460585594177246,
    "timestamp": "2025-07-14T12:35:20.564161"
  },
  {
    "epoch": 2968,
    "batch_idx": 64,
    "loss": 0.2479703575372696,
    "learning_rate": 0.00047487784091011656,
    "timestep_mean": 637.21875,
    "epsilon_norm": 7.782617092132568,
    "prediction_norm": 6.609811305999756,
    "timestamp": "2025-07-14T12:35:20.569361"
  },
  {
    "epoch": 2968,
    "batch_idx": 96,
    "loss": 0.14267079532146454,
    "learning_rate": 0.00047487784091011656,
    "timestep_mean": 451.28125,
    "epsilon_norm": 5.669942378997803,
    "prediction_norm": 5.022044658660889,
    "timestamp": "2025-07-14T12:35:20.574545"
  },
  {
    "epoch": 2968,
    "batch_idx": 128,
    "loss": 0.1674618422985077,
    "learning_rate": 0.00047487784091011656,
    "timestep_mean": 556.0625,
    "epsilon_norm": 8.574563980102539,
    "prediction_norm": 7.384688377380371,
    "timestamp": "2025-07-14T12:35:20.579710"
  },
  {
    "epoch": 2968,
    "batch_idx": 160,
    "loss": 0.5369067788124084,
    "learning_rate": 0.00047487784091011656,
    "timestep_mean": 463.15625,
    "epsilon_norm": 9.158629417419434,
    "prediction_norm": 6.206899166107178,
    "timestamp": "2025-07-14T12:35:20.584849"
  },
  {
    "epoch": 2968,
    "batch_idx": 192,
    "loss": 0.09185187518596649,
    "learning_rate": 0.00047487784091011656,
    "timestep_mean": 529.25,
    "epsilon_norm": 3.40155029296875,
    "prediction_norm": 3.0104897022247314,
    "timestamp": "2025-07-14T12:35:20.589991"
  },
  {
    "epoch": 2969,
    "batch_idx": 0,
    "loss": 0.5197765827178955,
    "learning_rate": 0.00047566227775219306,
    "timestep_mean": 479.03125,
    "epsilon_norm": 8.888161659240723,
    "prediction_norm": 6.374247074127197,
    "timestamp": "2025-07-14T12:35:20.595172"
  },
  {
    "epoch": 2969,
    "batch_idx": 32,
    "loss": 0.21759134531021118,
    "learning_rate": 0.00047566227775219306,
    "timestep_mean": 517.9375,
    "epsilon_norm": 7.391624927520752,
    "prediction_norm": 5.803516387939453,
    "timestamp": "2025-07-14T12:35:20.600550"
  },
  {
    "epoch": 2969,
    "batch_idx": 64,
    "loss": 0.32600656151771545,
    "learning_rate": 0.00047566227775219306,
    "timestep_mean": 526.59375,
    "epsilon_norm": 8.188852310180664,
    "prediction_norm": 6.7972412109375,
    "timestamp": "2025-07-14T12:35:20.605675"
  },
  {
    "epoch": 2969,
    "batch_idx": 96,
    "loss": 0.3225146234035492,
    "learning_rate": 0.00047566227775219306,
    "timestep_mean": 443.4375,
    "epsilon_norm": 8.59328842163086,
    "prediction_norm": 6.918015956878662,
    "timestamp": "2025-07-14T12:35:20.610758"
  },
  {
    "epoch": 2969,
    "batch_idx": 128,
    "loss": 0.29567649960517883,
    "learning_rate": 0.00047566227775219306,
    "timestep_mean": 413.875,
    "epsilon_norm": 7.242198467254639,
    "prediction_norm": 5.35017728805542,
    "timestamp": "2025-07-14T12:35:20.615989"
  },
  {
    "epoch": 2969,
    "batch_idx": 160,
    "loss": 0.18585902452468872,
    "learning_rate": 0.00047566227775219306,
    "timestep_mean": 532.96875,
    "epsilon_norm": 8.046092987060547,
    "prediction_norm": 7.529190540313721,
    "timestamp": "2025-07-14T12:35:20.621207"
  },
  {
    "epoch": 2969,
    "batch_idx": 192,
    "loss": 0.5920725464820862,
    "learning_rate": 0.00047566227775219306,
    "timestep_mean": 388.125,
    "epsilon_norm": 4.463744640350342,
    "prediction_norm": 3.237595796585083,
    "timestamp": "2025-07-14T12:35:20.626360"
  },
  {
    "epoch": 2970,
    "batch_idx": 0,
    "loss": 0.41993775963783264,
    "learning_rate": 0.0004764467746451799,
    "timestep_mean": 440.5625,
    "epsilon_norm": 8.080021858215332,
    "prediction_norm": 6.4145708084106445,
    "timestamp": "2025-07-14T12:35:20.631662"
  },
  {
    "epoch": 2970,
    "batch_idx": 32,
    "loss": 0.1984606385231018,
    "learning_rate": 0.0004764467746451799,
    "timestep_mean": 495.15625,
    "epsilon_norm": 7.327513694763184,
    "prediction_norm": 6.683865547180176,
    "timestamp": "2025-07-14T12:35:20.636931"
  },
  {
    "epoch": 2970,
    "batch_idx": 64,
    "loss": 0.3627641499042511,
    "learning_rate": 0.0004764467746451799,
    "timestep_mean": 469.34375,
    "epsilon_norm": 8.35869026184082,
    "prediction_norm": 6.1323628425598145,
    "timestamp": "2025-07-14T12:35:20.642276"
  },
  {
    "epoch": 2970,
    "batch_idx": 96,
    "loss": 0.35097095370292664,
    "learning_rate": 0.0004764467746451799,
    "timestep_mean": 592.0,
    "epsilon_norm": 7.194218635559082,
    "prediction_norm": 5.737173557281494,
    "timestamp": "2025-07-14T12:35:20.647544"
  },
  {
    "epoch": 2970,
    "batch_idx": 128,
    "loss": 0.1838858723640442,
    "learning_rate": 0.0004764467746451799,
    "timestep_mean": 487.65625,
    "epsilon_norm": 7.7984538078308105,
    "prediction_norm": 6.536616802215576,
    "timestamp": "2025-07-14T12:35:20.652773"
  },
  {
    "epoch": 2970,
    "batch_idx": 160,
    "loss": 0.2338670790195465,
    "learning_rate": 0.0004764467746451799,
    "timestep_mean": 558.96875,
    "epsilon_norm": 7.2403883934021,
    "prediction_norm": 6.450489521026611,
    "timestamp": "2025-07-14T12:35:20.657977"
  },
  {
    "epoch": 2970,
    "batch_idx": 192,
    "loss": 0.17627428472042084,
    "learning_rate": 0.0004764467746451799,
    "timestep_mean": 521.875,
    "epsilon_norm": 3.3688433170318604,
    "prediction_norm": 2.1720004081726074,
    "timestamp": "2025-07-14T12:35:20.663143"
  },
  {
    "epoch": 2971,
    "batch_idx": 0,
    "loss": 0.3009882867336273,
    "learning_rate": 0.0004772313296534085,
    "timestep_mean": 515.0,
    "epsilon_norm": 7.640257835388184,
    "prediction_norm": 6.854029655456543,
    "timestamp": "2025-07-14T12:35:20.668320"
  },
  {
    "epoch": 2971,
    "batch_idx": 32,
    "loss": 0.23861515522003174,
    "learning_rate": 0.0004772313296534085,
    "timestep_mean": 481.21875,
    "epsilon_norm": 7.6239447593688965,
    "prediction_norm": 6.575891494750977,
    "timestamp": "2025-07-14T12:35:20.673579"
  },
  {
    "epoch": 2971,
    "batch_idx": 64,
    "loss": 0.17613667249679565,
    "learning_rate": 0.0004772313296534085,
    "timestep_mean": 464.34375,
    "epsilon_norm": 7.098949909210205,
    "prediction_norm": 6.006011962890625,
    "timestamp": "2025-07-14T12:35:20.678848"
  },
  {
    "epoch": 2971,
    "batch_idx": 96,
    "loss": 0.26811879873275757,
    "learning_rate": 0.0004772313296534085,
    "timestep_mean": 526.8125,
    "epsilon_norm": 9.929424285888672,
    "prediction_norm": 9.38409423828125,
    "timestamp": "2025-07-14T12:35:20.683960"
  },
  {
    "epoch": 2971,
    "batch_idx": 128,
    "loss": 0.29836416244506836,
    "learning_rate": 0.0004772313296534085,
    "timestep_mean": 587.71875,
    "epsilon_norm": 8.461345672607422,
    "prediction_norm": 8.248257637023926,
    "timestamp": "2025-07-14T12:35:20.688756"
  },
  {
    "epoch": 2971,
    "batch_idx": 160,
    "loss": 0.1734732687473297,
    "learning_rate": 0.0004772313296534085,
    "timestep_mean": 555.1875,
    "epsilon_norm": 7.120582103729248,
    "prediction_norm": 7.059817314147949,
    "timestamp": "2025-07-14T12:35:20.693794"
  },
  {
    "epoch": 2971,
    "batch_idx": 192,
    "loss": 0.05533549189567566,
    "learning_rate": 0.0004772313296534085,
    "timestep_mean": 486.625,
    "epsilon_norm": 4.486379623413086,
    "prediction_norm": 4.603307723999023,
    "timestamp": "2025-07-14T12:35:20.698716"
  },
  {
    "epoch": 2972,
    "batch_idx": 0,
    "loss": 0.21769820153713226,
    "learning_rate": 0.00047801594084106866,
    "timestep_mean": 475.21875,
    "epsilon_norm": 7.361020088195801,
    "prediction_norm": 6.809401512145996,
    "timestamp": "2025-07-14T12:35:20.703576"
  },
  {
    "epoch": 2972,
    "batch_idx": 32,
    "loss": 0.2764534652233124,
    "learning_rate": 0.00047801594084106866,
    "timestep_mean": 479.34375,
    "epsilon_norm": 7.838106632232666,
    "prediction_norm": 6.567056179046631,
    "timestamp": "2025-07-14T12:35:20.708477"
  },
  {
    "epoch": 2972,
    "batch_idx": 64,
    "loss": 0.42908528447151184,
    "learning_rate": 0.00047801594084106866,
    "timestep_mean": 551.3125,
    "epsilon_norm": 8.713433265686035,
    "prediction_norm": 7.062121868133545,
    "timestamp": "2025-07-14T12:35:20.713463"
  },
  {
    "epoch": 2972,
    "batch_idx": 96,
    "loss": 0.31713420152664185,
    "learning_rate": 0.00047801594084106866,
    "timestep_mean": 499.09375,
    "epsilon_norm": 7.95014762878418,
    "prediction_norm": 6.279283046722412,
    "timestamp": "2025-07-14T12:35:20.718297"
  },
  {
    "epoch": 2972,
    "batch_idx": 128,
    "loss": 0.27653926610946655,
    "learning_rate": 0.00047801594084106866,
    "timestep_mean": 547.21875,
    "epsilon_norm": 8.666956901550293,
    "prediction_norm": 6.989276885986328,
    "timestamp": "2025-07-14T12:35:20.723113"
  },
  {
    "epoch": 2972,
    "batch_idx": 160,
    "loss": 0.3176153898239136,
    "learning_rate": 0.00047801594084106866,
    "timestep_mean": 398.25,
    "epsilon_norm": 7.1894707679748535,
    "prediction_norm": 5.998682022094727,
    "timestamp": "2025-07-14T12:35:20.728026"
  },
  {
    "epoch": 2972,
    "batch_idx": 192,
    "loss": 0.7131906151771545,
    "learning_rate": 0.00047801594084106866,
    "timestep_mean": 528.5,
    "epsilon_norm": 4.891072750091553,
    "prediction_norm": 3.0948922634124756,
    "timestamp": "2025-07-14T12:35:20.732901"
  },
  {
    "epoch": 2973,
    "batch_idx": 0,
    "loss": 0.40414682030677795,
    "learning_rate": 0.000478800606272209,
    "timestep_mean": 534.09375,
    "epsilon_norm": 9.159119606018066,
    "prediction_norm": 7.708106517791748,
    "timestamp": "2025-07-14T12:35:20.737864"
  },
  {
    "epoch": 2973,
    "batch_idx": 32,
    "loss": 0.2545970678329468,
    "learning_rate": 0.000478800606272209,
    "timestep_mean": 427.09375,
    "epsilon_norm": 7.565605640411377,
    "prediction_norm": 6.642773628234863,
    "timestamp": "2025-07-14T12:35:20.742802"
  },
  {
    "epoch": 2973,
    "batch_idx": 64,
    "loss": 0.08543339371681213,
    "learning_rate": 0.000478800606272209,
    "timestep_mean": 578.875,
    "epsilon_norm": 6.730008602142334,
    "prediction_norm": 5.934453964233398,
    "timestamp": "2025-07-14T12:35:20.747705"
  },
  {
    "epoch": 2973,
    "batch_idx": 96,
    "loss": 0.157903254032135,
    "learning_rate": 0.000478800606272209,
    "timestep_mean": 458.09375,
    "epsilon_norm": 6.443411350250244,
    "prediction_norm": 5.7219624519348145,
    "timestamp": "2025-07-14T12:35:20.752540"
  },
  {
    "epoch": 2973,
    "batch_idx": 128,
    "loss": 0.27935534715652466,
    "learning_rate": 0.000478800606272209,
    "timestep_mean": 529.5625,
    "epsilon_norm": 8.837042808532715,
    "prediction_norm": 7.16818380355835,
    "timestamp": "2025-07-14T12:35:20.757374"
  },
  {
    "epoch": 2973,
    "batch_idx": 160,
    "loss": 0.2922505736351013,
    "learning_rate": 0.000478800606272209,
    "timestep_mean": 487.25,
    "epsilon_norm": 8.431525230407715,
    "prediction_norm": 6.96404504776001,
    "timestamp": "2025-07-14T12:35:20.762198"
  },
  {
    "epoch": 2973,
    "batch_idx": 192,
    "loss": 0.22782623767852783,
    "learning_rate": 0.000478800606272209,
    "timestep_mean": 505.5,
    "epsilon_norm": 2.0247960090637207,
    "prediction_norm": 1.8488632440567017,
    "timestamp": "2025-07-14T12:35:20.767095"
  },
  {
    "epoch": 2974,
    "batch_idx": 0,
    "loss": 0.3256818652153015,
    "learning_rate": 0.000479585324010746,
    "timestep_mean": 496.90625,
    "epsilon_norm": 7.942013263702393,
    "prediction_norm": 6.171290397644043,
    "timestamp": "2025-07-14T12:35:20.771984"
  },
  {
    "epoch": 2974,
    "batch_idx": 32,
    "loss": 0.33637088537216187,
    "learning_rate": 0.000479585324010746,
    "timestep_mean": 533.4375,
    "epsilon_norm": 7.384228706359863,
    "prediction_norm": 5.867968559265137,
    "timestamp": "2025-07-14T12:35:20.776764"
  },
  {
    "epoch": 2974,
    "batch_idx": 64,
    "loss": 0.29700514674186707,
    "learning_rate": 0.000479585324010746,
    "timestep_mean": 593.25,
    "epsilon_norm": 9.379009246826172,
    "prediction_norm": 7.823032379150391,
    "timestamp": "2025-07-14T12:35:20.781580"
  },
  {
    "epoch": 2974,
    "batch_idx": 96,
    "loss": 0.24085383117198944,
    "learning_rate": 0.000479585324010746,
    "timestep_mean": 513.28125,
    "epsilon_norm": 6.993546962738037,
    "prediction_norm": 5.523889541625977,
    "timestamp": "2025-07-14T12:35:20.786458"
  },
  {
    "epoch": 2974,
    "batch_idx": 128,
    "loss": 0.18608078360557556,
    "learning_rate": 0.000479585324010746,
    "timestep_mean": 421.8125,
    "epsilon_norm": 6.875402927398682,
    "prediction_norm": 5.925227165222168,
    "timestamp": "2025-07-14T12:35:20.791353"
  },
  {
    "epoch": 2974,
    "batch_idx": 160,
    "loss": 0.17215360701084137,
    "learning_rate": 0.000479585324010746,
    "timestep_mean": 564.9375,
    "epsilon_norm": 7.573339462280273,
    "prediction_norm": 6.683370113372803,
    "timestamp": "2025-07-14T12:35:20.796376"
  },
  {
    "epoch": 2974,
    "batch_idx": 192,
    "loss": 0.9507342576980591,
    "learning_rate": 0.000479585324010746,
    "timestep_mean": 416.5,
    "epsilon_norm": 4.529916763305664,
    "prediction_norm": 2.6927435398101807,
    "timestamp": "2025-07-14T12:35:20.801441"
  },
  {
    "epoch": 2975,
    "batch_idx": 0,
    "loss": 0.2215644270181656,
    "learning_rate": 0.00048037009212046716,
    "timestep_mean": 457.3125,
    "epsilon_norm": 7.595552921295166,
    "prediction_norm": 6.1983418464660645,
    "timestamp": "2025-07-14T12:35:20.806548"
  },
  {
    "epoch": 2975,
    "batch_idx": 32,
    "loss": 0.2490946650505066,
    "learning_rate": 0.00048037009212046716,
    "timestep_mean": 525.3125,
    "epsilon_norm": 8.731573104858398,
    "prediction_norm": 7.568502902984619,
    "timestamp": "2025-07-14T12:35:20.811707"
  },
  {
    "epoch": 2975,
    "batch_idx": 64,
    "loss": 0.3409096598625183,
    "learning_rate": 0.00048037009212046716,
    "timestep_mean": 483.9375,
    "epsilon_norm": 8.939919471740723,
    "prediction_norm": 7.682815074920654,
    "timestamp": "2025-07-14T12:35:20.816821"
  },
  {
    "epoch": 2975,
    "batch_idx": 96,
    "loss": 0.3576399087905884,
    "learning_rate": 0.00048037009212046716,
    "timestep_mean": 507.21875,
    "epsilon_norm": 7.574837684631348,
    "prediction_norm": 5.896481037139893,
    "timestamp": "2025-07-14T12:35:20.822069"
  },
  {
    "epoch": 2975,
    "batch_idx": 128,
    "loss": 0.13709703087806702,
    "learning_rate": 0.00048037009212046716,
    "timestep_mean": 474.28125,
    "epsilon_norm": 7.213697910308838,
    "prediction_norm": 6.549001216888428,
    "timestamp": "2025-07-14T12:35:20.827267"
  },
  {
    "epoch": 2975,
    "batch_idx": 160,
    "loss": 0.18727675080299377,
    "learning_rate": 0.00048037009212046716,
    "timestep_mean": 591.21875,
    "epsilon_norm": 8.143819808959961,
    "prediction_norm": 7.2086358070373535,
    "timestamp": "2025-07-14T12:35:20.832197"
  },
  {
    "epoch": 2975,
    "batch_idx": 192,
    "loss": 0.173342764377594,
    "learning_rate": 0.00048037009212046716,
    "timestep_mean": 490.25,
    "epsilon_norm": 3.590742588043213,
    "prediction_norm": 3.142094373703003,
    "timestamp": "2025-07-14T12:35:20.837243"
  },
  {
    "epoch": 2976,
    "batch_idx": 0,
    "loss": 0.37445276975631714,
    "learning_rate": 0.00048115490866503365,
    "timestep_mean": 494.875,
    "epsilon_norm": 9.050271034240723,
    "prediction_norm": 6.979881763458252,
    "timestamp": "2025-07-14T12:35:20.842579"
  },
  {
    "epoch": 2976,
    "batch_idx": 32,
    "loss": 0.2523346245288849,
    "learning_rate": 0.00048115490866503365,
    "timestep_mean": 496.5,
    "epsilon_norm": 7.1092729568481445,
    "prediction_norm": 6.020143985748291,
    "timestamp": "2025-07-14T12:35:20.847860"
  },
  {
    "epoch": 2976,
    "batch_idx": 64,
    "loss": 0.2847251892089844,
    "learning_rate": 0.00048115490866503365,
    "timestep_mean": 443.4375,
    "epsilon_norm": 7.007939338684082,
    "prediction_norm": 6.121615409851074,
    "timestamp": "2025-07-14T12:35:20.853065"
  },
  {
    "epoch": 2976,
    "batch_idx": 96,
    "loss": 0.28858429193496704,
    "learning_rate": 0.00048115490866503365,
    "timestep_mean": 494.625,
    "epsilon_norm": 6.9672627449035645,
    "prediction_norm": 6.374566078186035,
    "timestamp": "2025-07-14T12:35:20.858332"
  },
  {
    "epoch": 2976,
    "batch_idx": 128,
    "loss": 0.21175682544708252,
    "learning_rate": 0.00048115490866503365,
    "timestep_mean": 565.3125,
    "epsilon_norm": 7.594841003417969,
    "prediction_norm": 6.6959943771362305,
    "timestamp": "2025-07-14T12:35:20.863734"
  },
  {
    "epoch": 2976,
    "batch_idx": 160,
    "loss": 0.3291131556034088,
    "learning_rate": 0.00048115490866503365,
    "timestep_mean": 511.75,
    "epsilon_norm": 7.767269134521484,
    "prediction_norm": 6.761589050292969,
    "timestamp": "2025-07-14T12:35:20.868888"
  },
  {
    "epoch": 2976,
    "batch_idx": 192,
    "loss": 0.08027660846710205,
    "learning_rate": 0.00048115490866503365,
    "timestep_mean": 586.375,
    "epsilon_norm": 3.690972328186035,
    "prediction_norm": 3.6316018104553223,
    "timestamp": "2025-07-14T12:35:20.874094"
  },
  {
    "epoch": 2977,
    "batch_idx": 0,
    "loss": 0.34395167231559753,
    "learning_rate": 0.00048193977170799023,
    "timestep_mean": 444.71875,
    "epsilon_norm": 7.761231422424316,
    "prediction_norm": 6.556956768035889,
    "timestamp": "2025-07-14T12:35:20.879324"
  },
  {
    "epoch": 2977,
    "batch_idx": 32,
    "loss": 0.3435872197151184,
    "learning_rate": 0.00048193977170799023,
    "timestep_mean": 478.15625,
    "epsilon_norm": 7.938957691192627,
    "prediction_norm": 6.593257427215576,
    "timestamp": "2025-07-14T12:35:20.884678"
  },
  {
    "epoch": 2977,
    "batch_idx": 64,
    "loss": 0.0940597802400589,
    "learning_rate": 0.00048193977170799023,
    "timestep_mean": 524.53125,
    "epsilon_norm": 7.701530456542969,
    "prediction_norm": 6.791659355163574,
    "timestamp": "2025-07-14T12:35:20.889879"
  },
  {
    "epoch": 2977,
    "batch_idx": 96,
    "loss": 0.2514611482620239,
    "learning_rate": 0.00048193977170799023,
    "timestep_mean": 527.6875,
    "epsilon_norm": 7.510858058929443,
    "prediction_norm": 6.559613227844238,
    "timestamp": "2025-07-14T12:35:20.895135"
  },
  {
    "epoch": 2977,
    "batch_idx": 128,
    "loss": 0.3941793143749237,
    "learning_rate": 0.00048193977170799023,
    "timestep_mean": 512.9375,
    "epsilon_norm": 8.648784637451172,
    "prediction_norm": 7.36482048034668,
    "timestamp": "2025-07-14T12:35:20.900490"
  },
  {
    "epoch": 2977,
    "batch_idx": 160,
    "loss": 0.13391581177711487,
    "learning_rate": 0.00048193977170799023,
    "timestep_mean": 457.78125,
    "epsilon_norm": 6.977527618408203,
    "prediction_norm": 5.973277568817139,
    "timestamp": "2025-07-14T12:35:20.905697"
  },
  {
    "epoch": 2977,
    "batch_idx": 192,
    "loss": 0.645211935043335,
    "learning_rate": 0.00048193977170799023,
    "timestep_mean": 389.625,
    "epsilon_norm": 3.5631721019744873,
    "prediction_norm": 1.9198616743087769,
    "timestamp": "2025-07-14T12:35:20.910879"
  },
  {
    "epoch": 2978,
    "batch_idx": 0,
    "loss": 0.23572155833244324,
    "learning_rate": 0.00048272467931276524,
    "timestep_mean": 555.9375,
    "epsilon_norm": 7.741245269775391,
    "prediction_norm": 6.7111968994140625,
    "timestamp": "2025-07-14T12:35:20.916065"
  },
  {
    "epoch": 2978,
    "batch_idx": 32,
    "loss": 0.2898868918418884,
    "learning_rate": 0.00048272467931276524,
    "timestep_mean": 529.875,
    "epsilon_norm": 7.8709588050842285,
    "prediction_norm": 6.723044395446777,
    "timestamp": "2025-07-14T12:35:20.921248"
  },
  {
    "epoch": 2978,
    "batch_idx": 64,
    "loss": 0.20040854811668396,
    "learning_rate": 0.00048272467931276524,
    "timestep_mean": 505.84375,
    "epsilon_norm": 7.2756028175354,
    "prediction_norm": 6.298806190490723,
    "timestamp": "2025-07-14T12:35:20.926392"
  },
  {
    "epoch": 2978,
    "batch_idx": 96,
    "loss": 0.2563016414642334,
    "learning_rate": 0.00048272467931276524,
    "timestep_mean": 491.03125,
    "epsilon_norm": 7.791476726531982,
    "prediction_norm": 6.372869968414307,
    "timestamp": "2025-07-14T12:35:20.931605"
  },
  {
    "epoch": 2978,
    "batch_idx": 128,
    "loss": 0.313562273979187,
    "learning_rate": 0.00048272467931276524,
    "timestep_mean": 500.03125,
    "epsilon_norm": 7.9150776863098145,
    "prediction_norm": 6.10532808303833,
    "timestamp": "2025-07-14T12:35:20.936858"
  },
  {
    "epoch": 2978,
    "batch_idx": 160,
    "loss": 0.553678035736084,
    "learning_rate": 0.00048272467931276524,
    "timestep_mean": 463.09375,
    "epsilon_norm": 8.752395629882812,
    "prediction_norm": 6.98629903793335,
    "timestamp": "2025-07-14T12:35:20.942144"
  },
  {
    "epoch": 2978,
    "batch_idx": 192,
    "loss": 0.08595563471317291,
    "learning_rate": 0.00048272467931276524,
    "timestep_mean": 524.875,
    "epsilon_norm": 5.201702117919922,
    "prediction_norm": 5.280704975128174,
    "timestamp": "2025-07-14T12:35:20.947289"
  },
  {
    "epoch": 2979,
    "batch_idx": 0,
    "loss": 0.23592084646224976,
    "learning_rate": 0.00048350962954267626,
    "timestep_mean": 490.125,
    "epsilon_norm": 8.211618423461914,
    "prediction_norm": 7.259544849395752,
    "timestamp": "2025-07-14T12:35:20.952459"
  },
  {
    "epoch": 2979,
    "batch_idx": 32,
    "loss": 0.36602073907852173,
    "learning_rate": 0.00048350962954267626,
    "timestep_mean": 486.0,
    "epsilon_norm": 7.646996974945068,
    "prediction_norm": 5.839172840118408,
    "timestamp": "2025-07-14T12:35:20.957687"
  },
  {
    "epoch": 2979,
    "batch_idx": 64,
    "loss": 0.209777370095253,
    "learning_rate": 0.00048350962954267626,
    "timestep_mean": 530.375,
    "epsilon_norm": 8.157886505126953,
    "prediction_norm": 7.442111968994141,
    "timestamp": "2025-07-14T12:35:20.962770"
  },
  {
    "epoch": 2979,
    "batch_idx": 96,
    "loss": 0.18725407123565674,
    "learning_rate": 0.00048350962954267626,
    "timestep_mean": 465.03125,
    "epsilon_norm": 7.163614749908447,
    "prediction_norm": 7.0612568855285645,
    "timestamp": "2025-07-14T12:35:20.967980"
  },
  {
    "epoch": 2979,
    "batch_idx": 128,
    "loss": 0.3087123930454254,
    "learning_rate": 0.00048350962954267626,
    "timestep_mean": 495.65625,
    "epsilon_norm": 7.927812099456787,
    "prediction_norm": 7.524232864379883,
    "timestamp": "2025-07-14T12:35:20.973116"
  },
  {
    "epoch": 2979,
    "batch_idx": 160,
    "loss": 0.15142613649368286,
    "learning_rate": 0.00048350962954267626,
    "timestep_mean": 614.375,
    "epsilon_norm": 8.131691932678223,
    "prediction_norm": 7.785834312438965,
    "timestamp": "2025-07-14T12:35:20.978336"
  },
  {
    "epoch": 2979,
    "batch_idx": 192,
    "loss": 0.07802199572324753,
    "learning_rate": 0.00048350962954267626,
    "timestep_mean": 496.0,
    "epsilon_norm": 3.510608196258545,
    "prediction_norm": 3.239084243774414,
    "timestamp": "2025-07-14T12:35:20.983495"
  },
  {
    "epoch": 2980,
    "batch_idx": 0,
    "loss": 0.16333183646202087,
    "learning_rate": 0.00048429462046093715,
    "timestep_mean": 526.5,
    "epsilon_norm": 6.520833969116211,
    "prediction_norm": 6.1772074699401855,
    "timestamp": "2025-07-14T12:35:20.988712"
  },
  {
    "epoch": 2980,
    "batch_idx": 32,
    "loss": 0.22452425956726074,
    "learning_rate": 0.00048429462046093715,
    "timestep_mean": 500.0,
    "epsilon_norm": 8.057401657104492,
    "prediction_norm": 6.741139888763428,
    "timestamp": "2025-07-14T12:35:20.994019"
  },
  {
    "epoch": 2980,
    "batch_idx": 64,
    "loss": 0.3273266553878784,
    "learning_rate": 0.00048429462046093715,
    "timestep_mean": 542.96875,
    "epsilon_norm": 8.364092826843262,
    "prediction_norm": 6.609936714172363,
    "timestamp": "2025-07-14T12:35:20.999173"
  },
  {
    "epoch": 2980,
    "batch_idx": 96,
    "loss": 0.16040334105491638,
    "learning_rate": 0.00048429462046093715,
    "timestep_mean": 568.875,
    "epsilon_norm": 7.6827216148376465,
    "prediction_norm": 6.994644641876221,
    "timestamp": "2025-07-14T12:35:21.004385"
  },
  {
    "epoch": 2980,
    "batch_idx": 128,
    "loss": 0.2512896656990051,
    "learning_rate": 0.00048429462046093715,
    "timestep_mean": 515.25,
    "epsilon_norm": 7.908012866973877,
    "prediction_norm": 6.3424391746521,
    "timestamp": "2025-07-14T12:35:21.009714"
  },
  {
    "epoch": 2980,
    "batch_idx": 160,
    "loss": 0.16526886820793152,
    "learning_rate": 0.00048429462046093715,
    "timestep_mean": 558.84375,
    "epsilon_norm": 7.307777404785156,
    "prediction_norm": 6.482011795043945,
    "timestamp": "2025-07-14T12:35:21.014975"
  },
  {
    "epoch": 2980,
    "batch_idx": 192,
    "loss": 0.36611872911453247,
    "learning_rate": 0.00048429462046093715,
    "timestep_mean": 369.125,
    "epsilon_norm": 5.195552825927734,
    "prediction_norm": 4.211628437042236,
    "timestamp": "2025-07-14T12:35:21.020114"
  },
  {
    "epoch": 2981,
    "batch_idx": 0,
    "loss": 0.2481873631477356,
    "learning_rate": 0.00048507965013066086,
    "timestep_mean": 489.65625,
    "epsilon_norm": 8.670385360717773,
    "prediction_norm": 7.1569318771362305,
    "timestamp": "2025-07-14T12:35:21.025440"
  },
  {
    "epoch": 2981,
    "batch_idx": 32,
    "loss": 0.4718940854072571,
    "learning_rate": 0.00048507965013066086,
    "timestep_mean": 496.0625,
    "epsilon_norm": 8.344698905944824,
    "prediction_norm": 6.196201801300049,
    "timestamp": "2025-07-14T12:35:21.030655"
  },
  {
    "epoch": 2981,
    "batch_idx": 64,
    "loss": 0.239584282040596,
    "learning_rate": 0.00048507965013066086,
    "timestep_mean": 511.875,
    "epsilon_norm": 7.369384765625,
    "prediction_norm": 6.119410991668701,
    "timestamp": "2025-07-14T12:35:21.035863"
  },
  {
    "epoch": 2981,
    "batch_idx": 96,
    "loss": 0.17328616976737976,
    "learning_rate": 0.00048507965013066086,
    "timestep_mean": 464.6875,
    "epsilon_norm": 8.388676643371582,
    "prediction_norm": 7.884846210479736,
    "timestamp": "2025-07-14T12:35:21.041017"
  },
  {
    "epoch": 2981,
    "batch_idx": 128,
    "loss": 0.25909751653671265,
    "learning_rate": 0.00048507965013066086,
    "timestep_mean": 517.96875,
    "epsilon_norm": 7.389451026916504,
    "prediction_norm": 6.906265735626221,
    "timestamp": "2025-07-14T12:35:21.046179"
  },
  {
    "epoch": 2981,
    "batch_idx": 160,
    "loss": 0.4673529267311096,
    "learning_rate": 0.00048507965013066086,
    "timestep_mean": 493.65625,
    "epsilon_norm": 7.669323921203613,
    "prediction_norm": 6.420324325561523,
    "timestamp": "2025-07-14T12:35:21.051463"
  },
  {
    "epoch": 2981,
    "batch_idx": 192,
    "loss": 0.25241851806640625,
    "learning_rate": 0.00048507965013066086,
    "timestep_mean": 554.375,
    "epsilon_norm": 3.5242161750793457,
    "prediction_norm": 3.160816192626953,
    "timestamp": "2025-07-14T12:35:21.056648"
  },
  {
    "epoch": 2982,
    "batch_idx": 0,
    "loss": 0.1192527562379837,
    "learning_rate": 0.0004858647166148647,
    "timestep_mean": 555.78125,
    "epsilon_norm": 7.741724967956543,
    "prediction_norm": 7.165225028991699,
    "timestamp": "2025-07-14T12:35:21.061834"
  },
  {
    "epoch": 2982,
    "batch_idx": 32,
    "loss": 0.06991478055715561,
    "learning_rate": 0.0004858647166148647,
    "timestep_mean": 629.53125,
    "epsilon_norm": 6.856063365936279,
    "prediction_norm": 6.743958950042725,
    "timestamp": "2025-07-14T12:35:21.067153"
  },
  {
    "epoch": 2982,
    "batch_idx": 64,
    "loss": 0.286018967628479,
    "learning_rate": 0.0004858647166148647,
    "timestep_mean": 468.03125,
    "epsilon_norm": 7.480124473571777,
    "prediction_norm": 6.280297756195068,
    "timestamp": "2025-07-14T12:35:21.072425"
  },
  {
    "epoch": 2982,
    "batch_idx": 96,
    "loss": 0.1841467171907425,
    "learning_rate": 0.0004858647166148647,
    "timestep_mean": 604.65625,
    "epsilon_norm": 8.39228343963623,
    "prediction_norm": 8.3250732421875,
    "timestamp": "2025-07-14T12:35:21.077651"
  },
  {
    "epoch": 2982,
    "batch_idx": 128,
    "loss": 0.4121941328048706,
    "learning_rate": 0.0004858647166148647,
    "timestep_mean": 496.75,
    "epsilon_norm": 8.222513198852539,
    "prediction_norm": 6.481064319610596,
    "timestamp": "2025-07-14T12:35:21.082922"
  },
  {
    "epoch": 2982,
    "batch_idx": 160,
    "loss": 0.377694308757782,
    "learning_rate": 0.0004858647166148647,
    "timestep_mean": 485.9375,
    "epsilon_norm": 8.398736000061035,
    "prediction_norm": 6.576284408569336,
    "timestamp": "2025-07-14T12:35:21.088138"
  },
  {
    "epoch": 2982,
    "batch_idx": 192,
    "loss": 0.15445193648338318,
    "learning_rate": 0.0004858647166148647,
    "timestep_mean": 612.375,
    "epsilon_norm": 3.3772430419921875,
    "prediction_norm": 2.662580966949463,
    "timestamp": "2025-07-14T12:35:21.093263"
  },
  {
    "epoch": 2983,
    "batch_idx": 0,
    "loss": 0.20658624172210693,
    "learning_rate": 0.0004866498179764751,
    "timestep_mean": 519.5,
    "epsilon_norm": 8.406746864318848,
    "prediction_norm": 6.903425693511963,
    "timestamp": "2025-07-14T12:35:21.098477"
  },
  {
    "epoch": 2983,
    "batch_idx": 32,
    "loss": 0.3758442997932434,
    "learning_rate": 0.0004866498179764751,
    "timestep_mean": 513.625,
    "epsilon_norm": 7.507670879364014,
    "prediction_norm": 5.991235256195068,
    "timestamp": "2025-07-14T12:35:21.103657"
  },
  {
    "epoch": 2983,
    "batch_idx": 64,
    "loss": 0.5495884418487549,
    "learning_rate": 0.0004866498179764751,
    "timestep_mean": 416.65625,
    "epsilon_norm": 7.981182098388672,
    "prediction_norm": 5.09901237487793,
    "timestamp": "2025-07-14T12:35:21.108943"
  },
  {
    "epoch": 2983,
    "batch_idx": 96,
    "loss": 0.23473802208900452,
    "learning_rate": 0.0004866498179764751,
    "timestep_mean": 472.34375,
    "epsilon_norm": 7.458656311035156,
    "prediction_norm": 6.571764945983887,
    "timestamp": "2025-07-14T12:35:21.114183"
  },
  {
    "epoch": 2983,
    "batch_idx": 128,
    "loss": 0.2677616477012634,
    "learning_rate": 0.0004866498179764751,
    "timestep_mean": 423.53125,
    "epsilon_norm": 7.455811500549316,
    "prediction_norm": 6.637161731719971,
    "timestamp": "2025-07-14T12:35:21.119382"
  },
  {
    "epoch": 2983,
    "batch_idx": 160,
    "loss": 0.3815051317214966,
    "learning_rate": 0.0004866498179764751,
    "timestep_mean": 528.6875,
    "epsilon_norm": 8.401817321777344,
    "prediction_norm": 6.873074054718018,
    "timestamp": "2025-07-14T12:35:21.124588"
  },
  {
    "epoch": 2983,
    "batch_idx": 192,
    "loss": 0.43159618973731995,
    "learning_rate": 0.0004866498179764751,
    "timestep_mean": 266.125,
    "epsilon_norm": 3.2311806678771973,
    "prediction_norm": 1.982201099395752,
    "timestamp": "2025-07-14T12:35:21.129753"
  },
  {
    "epoch": 2984,
    "batch_idx": 0,
    "loss": 0.3291683495044708,
    "learning_rate": 0.00048743495227833244,
    "timestep_mean": 494.75,
    "epsilon_norm": 7.877328395843506,
    "prediction_norm": 5.774615287780762,
    "timestamp": "2025-07-14T12:35:21.134904"
  },
  {
    "epoch": 2984,
    "batch_idx": 32,
    "loss": 0.1744184046983719,
    "learning_rate": 0.00048743495227833244,
    "timestep_mean": 512.4375,
    "epsilon_norm": 7.117592811584473,
    "prediction_norm": 6.663741588592529,
    "timestamp": "2025-07-14T12:35:21.140141"
  },
  {
    "epoch": 2984,
    "batch_idx": 64,
    "loss": 0.2867804765701294,
    "learning_rate": 0.00048743495227833244,
    "timestep_mean": 530.40625,
    "epsilon_norm": 7.374436855316162,
    "prediction_norm": 6.225069999694824,
    "timestamp": "2025-07-14T12:35:21.145319"
  },
  {
    "epoch": 2984,
    "batch_idx": 96,
    "loss": 0.20697903633117676,
    "learning_rate": 0.00048743495227833244,
    "timestep_mean": 532.03125,
    "epsilon_norm": 6.858852863311768,
    "prediction_norm": 5.934372425079346,
    "timestamp": "2025-07-14T12:35:21.150673"
  },
  {
    "epoch": 2984,
    "batch_idx": 128,
    "loss": 0.29343709349632263,
    "learning_rate": 0.00048743495227833244,
    "timestep_mean": 491.9375,
    "epsilon_norm": 8.049551010131836,
    "prediction_norm": 6.8876166343688965,
    "timestamp": "2025-07-14T12:35:21.155860"
  },
  {
    "epoch": 2984,
    "batch_idx": 160,
    "loss": 0.20976094901561737,
    "learning_rate": 0.00048743495227833244,
    "timestep_mean": 486.5625,
    "epsilon_norm": 9.259451866149902,
    "prediction_norm": 7.686301231384277,
    "timestamp": "2025-07-14T12:35:21.160974"
  },
  {
    "epoch": 2984,
    "batch_idx": 192,
    "loss": 0.3876302242279053,
    "learning_rate": 0.00048743495227833244,
    "timestep_mean": 257.125,
    "epsilon_norm": 3.998518943786621,
    "prediction_norm": 2.758556365966797,
    "timestamp": "2025-07-14T12:35:21.166159"
  },
  {
    "epoch": 2985,
    "batch_idx": 0,
    "loss": 0.3307129740715027,
    "learning_rate": 0.00048822011758319653,
    "timestep_mean": 479.78125,
    "epsilon_norm": 8.184783935546875,
    "prediction_norm": 7.376732349395752,
    "timestamp": "2025-07-14T12:35:21.171387"
  },
  {
    "epoch": 2985,
    "batch_idx": 32,
    "loss": 0.18590641021728516,
    "learning_rate": 0.00048822011758319653,
    "timestep_mean": 491.09375,
    "epsilon_norm": 7.288467884063721,
    "prediction_norm": 6.617593765258789,
    "timestamp": "2025-07-14T12:35:21.176615"
  },
  {
    "epoch": 2985,
    "batch_idx": 64,
    "loss": 0.2960483431816101,
    "learning_rate": 0.00048822011758319653,
    "timestep_mean": 504.25,
    "epsilon_norm": 6.68831729888916,
    "prediction_norm": 5.226603984832764,
    "timestamp": "2025-07-14T12:35:21.181774"
  },
  {
    "epoch": 2985,
    "batch_idx": 96,
    "loss": 0.20957405865192413,
    "learning_rate": 0.00048822011758319653,
    "timestep_mean": 503.34375,
    "epsilon_norm": 7.439553737640381,
    "prediction_norm": 6.786667346954346,
    "timestamp": "2025-07-14T12:35:21.186869"
  },
  {
    "epoch": 2985,
    "batch_idx": 128,
    "loss": 0.24021495878696442,
    "learning_rate": 0.00048822011758319653,
    "timestep_mean": 525.9375,
    "epsilon_norm": 8.704955101013184,
    "prediction_norm": 7.362062931060791,
    "timestamp": "2025-07-14T12:35:21.192182"
  },
  {
    "epoch": 2985,
    "batch_idx": 160,
    "loss": 0.19208133220672607,
    "learning_rate": 0.00048822011758319653,
    "timestep_mean": 490.25,
    "epsilon_norm": 8.59175968170166,
    "prediction_norm": 8.065646171569824,
    "timestamp": "2025-07-14T12:35:21.197384"
  },
  {
    "epoch": 2985,
    "batch_idx": 192,
    "loss": 0.1768050491809845,
    "learning_rate": 0.00048822011758319653,
    "timestep_mean": 542.25,
    "epsilon_norm": 3.841315746307373,
    "prediction_norm": 2.9452548027038574,
    "timestamp": "2025-07-14T12:35:21.202515"
  },
  {
    "epoch": 2986,
    "batch_idx": 0,
    "loss": 0.2662525177001953,
    "learning_rate": 0.000489005311953749,
    "timestep_mean": 501.90625,
    "epsilon_norm": 7.196810245513916,
    "prediction_norm": 6.3198561668396,
    "timestamp": "2025-07-14T12:35:21.207697"
  },
  {
    "epoch": 2986,
    "batch_idx": 32,
    "loss": 0.2762281894683838,
    "learning_rate": 0.000489005311953749,
    "timestep_mean": 529.40625,
    "epsilon_norm": 9.699893951416016,
    "prediction_norm": 8.059789657592773,
    "timestamp": "2025-07-14T12:35:21.212945"
  },
  {
    "epoch": 2986,
    "batch_idx": 64,
    "loss": 0.20271269977092743,
    "learning_rate": 0.000489005311953749,
    "timestep_mean": 435.71875,
    "epsilon_norm": 7.253891468048096,
    "prediction_norm": 6.267143726348877,
    "timestamp": "2025-07-14T12:35:21.218200"
  },
  {
    "epoch": 2986,
    "batch_idx": 96,
    "loss": 0.4142965078353882,
    "learning_rate": 0.000489005311953749,
    "timestep_mean": 463.15625,
    "epsilon_norm": 8.771675109863281,
    "prediction_norm": 6.62258243560791,
    "timestamp": "2025-07-14T12:35:21.223441"
  },
  {
    "epoch": 2986,
    "batch_idx": 128,
    "loss": 0.2417205423116684,
    "learning_rate": 0.000489005311953749,
    "timestep_mean": 538.375,
    "epsilon_norm": 8.248398780822754,
    "prediction_norm": 7.047834873199463,
    "timestamp": "2025-07-14T12:35:21.228666"
  },
  {
    "epoch": 2986,
    "batch_idx": 160,
    "loss": 0.24148783087730408,
    "learning_rate": 0.000489005311953749,
    "timestep_mean": 609.9375,
    "epsilon_norm": 8.470728874206543,
    "prediction_norm": 7.907060146331787,
    "timestamp": "2025-07-14T12:35:21.233768"
  },
  {
    "epoch": 2986,
    "batch_idx": 192,
    "loss": 0.15897315740585327,
    "learning_rate": 0.000489005311953749,
    "timestep_mean": 590.875,
    "epsilon_norm": 3.300264358520508,
    "prediction_norm": 3.254173517227173,
    "timestamp": "2025-07-14T12:35:21.238604"
  },
  {
    "epoch": 2987,
    "batch_idx": 0,
    "loss": 0.5945943593978882,
    "learning_rate": 0.0004897905334526008,
    "timestep_mean": 464.0,
    "epsilon_norm": 7.966338634490967,
    "prediction_norm": 5.828166961669922,
    "timestamp": "2025-07-14T12:35:21.243338"
  },
  {
    "epoch": 2987,
    "batch_idx": 32,
    "loss": 0.2638327479362488,
    "learning_rate": 0.0004897905334526008,
    "timestep_mean": 549.59375,
    "epsilon_norm": 9.047115325927734,
    "prediction_norm": 8.273039817810059,
    "timestamp": "2025-07-14T12:35:21.248207"
  },
  {
    "epoch": 2987,
    "batch_idx": 64,
    "loss": 0.1947149634361267,
    "learning_rate": 0.0004897905334526008,
    "timestep_mean": 482.71875,
    "epsilon_norm": 6.37394905090332,
    "prediction_norm": 6.60264253616333,
    "timestamp": "2025-07-14T12:35:21.253106"
  },
  {
    "epoch": 2987,
    "batch_idx": 96,
    "loss": 0.4650334119796753,
    "learning_rate": 0.0004897905334526008,
    "timestep_mean": 447.03125,
    "epsilon_norm": 7.4043731689453125,
    "prediction_norm": 5.047101020812988,
    "timestamp": "2025-07-14T12:35:21.258109"
  },
  {
    "epoch": 2987,
    "batch_idx": 128,
    "loss": 0.3142544627189636,
    "learning_rate": 0.0004897905334526008,
    "timestep_mean": 421.0625,
    "epsilon_norm": 7.915049076080322,
    "prediction_norm": 7.13686990737915,
    "timestamp": "2025-07-14T12:35:21.263381"
  },
  {
    "epoch": 2987,
    "batch_idx": 160,
    "loss": 0.2055594027042389,
    "learning_rate": 0.0004897905334526008,
    "timestep_mean": 610.0,
    "epsilon_norm": 8.590412139892578,
    "prediction_norm": 8.279486656188965,
    "timestamp": "2025-07-14T12:35:21.268487"
  },
  {
    "epoch": 2987,
    "batch_idx": 192,
    "loss": 0.36425527930259705,
    "learning_rate": 0.0004897905334526008,
    "timestep_mean": 560.125,
    "epsilon_norm": 4.4867377281188965,
    "prediction_norm": 3.741710662841797,
    "timestamp": "2025-07-14T12:35:21.273388"
  },
  {
    "epoch": 2988,
    "batch_idx": 0,
    "loss": 0.17790824174880981,
    "learning_rate": 0.0004905757801422968,
    "timestep_mean": 528.3125,
    "epsilon_norm": 7.2627692222595215,
    "prediction_norm": 6.67598295211792,
    "timestamp": "2025-07-14T12:35:21.278472"
  },
  {
    "epoch": 2988,
    "batch_idx": 32,
    "loss": 0.19521749019622803,
    "learning_rate": 0.0004905757801422968,
    "timestep_mean": 522.03125,
    "epsilon_norm": 8.15127182006836,
    "prediction_norm": 7.520474910736084,
    "timestamp": "2025-07-14T12:35:21.283706"
  },
  {
    "epoch": 2988,
    "batch_idx": 64,
    "loss": 0.16597333550453186,
    "learning_rate": 0.0004905757801422968,
    "timestep_mean": 558.0625,
    "epsilon_norm": 7.458248138427734,
    "prediction_norm": 6.205073356628418,
    "timestamp": "2025-07-14T12:35:21.288833"
  },
  {
    "epoch": 2988,
    "batch_idx": 96,
    "loss": 0.242329940199852,
    "learning_rate": 0.0004905757801422968,
    "timestep_mean": 486.90625,
    "epsilon_norm": 7.574472427368164,
    "prediction_norm": 6.447376251220703,
    "timestamp": "2025-07-14T12:35:21.294002"
  },
  {
    "epoch": 2988,
    "batch_idx": 128,
    "loss": 0.30512407422065735,
    "learning_rate": 0.0004905757801422968,
    "timestep_mean": 436.90625,
    "epsilon_norm": 7.876684665679932,
    "prediction_norm": 6.56739616394043,
    "timestamp": "2025-07-14T12:35:21.299139"
  },
  {
    "epoch": 2988,
    "batch_idx": 160,
    "loss": 0.15977230668067932,
    "learning_rate": 0.0004905757801422968,
    "timestep_mean": 549.125,
    "epsilon_norm": 7.680596828460693,
    "prediction_norm": 7.038839340209961,
    "timestamp": "2025-07-14T12:35:21.304257"
  },
  {
    "epoch": 2988,
    "batch_idx": 192,
    "loss": 0.4629858732223511,
    "learning_rate": 0.0004905757801422968,
    "timestep_mean": 503.625,
    "epsilon_norm": 3.7651278972625732,
    "prediction_norm": 2.8133130073547363,
    "timestamp": "2025-07-14T12:35:21.309432"
  },
  {
    "epoch": 2989,
    "batch_idx": 0,
    "loss": 0.1554725170135498,
    "learning_rate": 0.0004913610500853191,
    "timestep_mean": 483.875,
    "epsilon_norm": 9.548379898071289,
    "prediction_norm": 8.070184707641602,
    "timestamp": "2025-07-14T12:35:21.314631"
  },
  {
    "epoch": 2989,
    "batch_idx": 32,
    "loss": 0.3298286199569702,
    "learning_rate": 0.0004913610500853191,
    "timestep_mean": 529.84375,
    "epsilon_norm": 9.210244178771973,
    "prediction_norm": 7.1561360359191895,
    "timestamp": "2025-07-14T12:35:21.319831"
  },
  {
    "epoch": 2989,
    "batch_idx": 64,
    "loss": 0.12848339974880219,
    "learning_rate": 0.0004913610500853191,
    "timestep_mean": 546.78125,
    "epsilon_norm": 8.296121597290039,
    "prediction_norm": 7.250399589538574,
    "timestamp": "2025-07-14T12:35:21.324901"
  },
  {
    "epoch": 2989,
    "batch_idx": 96,
    "loss": 0.1568593531847,
    "learning_rate": 0.0004913610500853191,
    "timestep_mean": 510.71875,
    "epsilon_norm": 7.5868401527404785,
    "prediction_norm": 6.962800979614258,
    "timestamp": "2025-07-14T12:35:21.330177"
  },
  {
    "epoch": 2989,
    "batch_idx": 128,
    "loss": 0.3340822458267212,
    "learning_rate": 0.0004913610500853191,
    "timestep_mean": 506.15625,
    "epsilon_norm": 8.019973754882812,
    "prediction_norm": 6.443509578704834,
    "timestamp": "2025-07-14T12:35:21.335033"
  },
  {
    "epoch": 2989,
    "batch_idx": 160,
    "loss": 0.21471837162971497,
    "learning_rate": 0.0004913610500853191,
    "timestep_mean": 545.4375,
    "epsilon_norm": 8.012619972229004,
    "prediction_norm": 6.572227954864502,
    "timestamp": "2025-07-14T12:35:21.339910"
  },
  {
    "epoch": 2989,
    "batch_idx": 192,
    "loss": 0.5190840363502502,
    "learning_rate": 0.0004913610500853191,
    "timestep_mean": 428.0,
    "epsilon_norm": 4.022797584533691,
    "prediction_norm": 3.2652435302734375,
    "timestamp": "2025-07-14T12:35:21.344728"
  },
  {
    "epoch": 2990,
    "batch_idx": 0,
    "loss": 0.19173116981983185,
    "learning_rate": 0.000492146341344091,
    "timestep_mean": 504.1875,
    "epsilon_norm": 7.302428245544434,
    "prediction_norm": 6.620484828948975,
    "timestamp": "2025-07-14T12:35:21.349587"
  },
  {
    "epoch": 2990,
    "batch_idx": 32,
    "loss": 0.4465166926383972,
    "learning_rate": 0.000492146341344091,
    "timestep_mean": 474.59375,
    "epsilon_norm": 8.6412992477417,
    "prediction_norm": 6.327339172363281,
    "timestamp": "2025-07-14T12:35:21.354810"
  },
  {
    "epoch": 2990,
    "batch_idx": 64,
    "loss": 0.23050257563591003,
    "learning_rate": 0.000492146341344091,
    "timestep_mean": 393.6875,
    "epsilon_norm": 6.370464324951172,
    "prediction_norm": 5.1935577392578125,
    "timestamp": "2025-07-14T12:35:21.359976"
  },
  {
    "epoch": 2990,
    "batch_idx": 96,
    "loss": 0.18247413635253906,
    "learning_rate": 0.000492146341344091,
    "timestep_mean": 579.1875,
    "epsilon_norm": 6.779207706451416,
    "prediction_norm": 5.957292079925537,
    "timestamp": "2025-07-14T12:35:21.365171"
  },
  {
    "epoch": 2990,
    "batch_idx": 128,
    "loss": 0.33962082862854004,
    "learning_rate": 0.000492146341344091,
    "timestep_mean": 472.3125,
    "epsilon_norm": 6.8654069900512695,
    "prediction_norm": 5.748068332672119,
    "timestamp": "2025-07-14T12:35:21.370165"
  },
  {
    "epoch": 2990,
    "batch_idx": 160,
    "loss": 0.24286898970603943,
    "learning_rate": 0.000492146341344091,
    "timestep_mean": 576.5,
    "epsilon_norm": 7.930336952209473,
    "prediction_norm": 7.06204891204834,
    "timestamp": "2025-07-14T12:35:21.375387"
  },
  {
    "epoch": 2990,
    "batch_idx": 192,
    "loss": 0.18972048163414001,
    "learning_rate": 0.000492146341344091,
    "timestep_mean": 543.0,
    "epsilon_norm": 4.07710599899292,
    "prediction_norm": 3.1408557891845703,
    "timestamp": "2025-07-14T12:35:21.380588"
  },
  {
    "epoch": 2991,
    "batch_idx": 0,
    "loss": 0.3822033107280731,
    "learning_rate": 0.0004929316519809849,
    "timestep_mean": 460.15625,
    "epsilon_norm": 6.671036243438721,
    "prediction_norm": 5.7353835105896,
    "timestamp": "2025-07-14T12:35:21.385697"
  },
  {
    "epoch": 2991,
    "batch_idx": 32,
    "loss": 0.18613770604133606,
    "learning_rate": 0.0004929316519809849,
    "timestep_mean": 509.03125,
    "epsilon_norm": 8.25390911102295,
    "prediction_norm": 7.324751377105713,
    "timestamp": "2025-07-14T12:35:21.390859"
  },
  {
    "epoch": 2991,
    "batch_idx": 64,
    "loss": 0.13679954409599304,
    "learning_rate": 0.0004929316519809849,
    "timestep_mean": 616.125,
    "epsilon_norm": 8.76680850982666,
    "prediction_norm": 8.09914493560791,
    "timestamp": "2025-07-14T12:35:21.396102"
  },
  {
    "epoch": 2991,
    "batch_idx": 96,
    "loss": 0.11775451898574829,
    "learning_rate": 0.0004929316519809849,
    "timestep_mean": 521.34375,
    "epsilon_norm": 7.4722900390625,
    "prediction_norm": 7.0114030838012695,
    "timestamp": "2025-07-14T12:35:21.401334"
  },
  {
    "epoch": 2991,
    "batch_idx": 128,
    "loss": 0.16686922311782837,
    "learning_rate": 0.0004929316519809849,
    "timestep_mean": 559.09375,
    "epsilon_norm": 8.257262229919434,
    "prediction_norm": 8.085390090942383,
    "timestamp": "2025-07-14T12:35:21.406400"
  },
  {
    "epoch": 2991,
    "batch_idx": 160,
    "loss": 0.1828424632549286,
    "learning_rate": 0.0004929316519809849,
    "timestep_mean": 600.34375,
    "epsilon_norm": 6.715932369232178,
    "prediction_norm": 6.298455715179443,
    "timestamp": "2025-07-14T12:35:21.411328"
  },
  {
    "epoch": 2991,
    "batch_idx": 192,
    "loss": 0.0737677812576294,
    "learning_rate": 0.0004929316519809849,
    "timestep_mean": 511.625,
    "epsilon_norm": 4.585106372833252,
    "prediction_norm": 4.510809421539307,
    "timestamp": "2025-07-14T12:35:21.416553"
  },
  {
    "epoch": 2992,
    "batch_idx": 0,
    "loss": 0.3088546395301819,
    "learning_rate": 0.0004937169800583253,
    "timestep_mean": 459.5,
    "epsilon_norm": 7.419503211975098,
    "prediction_norm": 6.4760823249816895,
    "timestamp": "2025-07-14T12:35:21.422055"
  },
  {
    "epoch": 2992,
    "batch_idx": 32,
    "loss": 0.3665986657142639,
    "learning_rate": 0.0004937169800583253,
    "timestep_mean": 435.21875,
    "epsilon_norm": 7.077846527099609,
    "prediction_norm": 5.803782939910889,
    "timestamp": "2025-07-14T12:35:21.427192"
  },
  {
    "epoch": 2992,
    "batch_idx": 64,
    "loss": 0.20043644309043884,
    "learning_rate": 0.0004937169800583253,
    "timestep_mean": 516.53125,
    "epsilon_norm": 8.196966171264648,
    "prediction_norm": 7.135966777801514,
    "timestamp": "2025-07-14T12:35:21.432402"
  },
  {
    "epoch": 2992,
    "batch_idx": 96,
    "loss": 0.30467355251312256,
    "learning_rate": 0.0004937169800583253,
    "timestep_mean": 397.84375,
    "epsilon_norm": 8.298198699951172,
    "prediction_norm": 6.216328144073486,
    "timestamp": "2025-07-14T12:35:21.437513"
  },
  {
    "epoch": 2992,
    "batch_idx": 128,
    "loss": 0.2438041865825653,
    "learning_rate": 0.0004937169800583253,
    "timestep_mean": 536.5625,
    "epsilon_norm": 7.43685245513916,
    "prediction_norm": 6.54144287109375,
    "timestamp": "2025-07-14T12:35:21.442819"
  },
  {
    "epoch": 2992,
    "batch_idx": 160,
    "loss": 0.2707754373550415,
    "learning_rate": 0.0004937169800583253,
    "timestep_mean": 560.59375,
    "epsilon_norm": 7.673951148986816,
    "prediction_norm": 6.70391321182251,
    "timestamp": "2025-07-14T12:35:21.447950"
  },
  {
    "epoch": 2992,
    "batch_idx": 192,
    "loss": 0.4779520630836487,
    "learning_rate": 0.0004937169800583253,
    "timestep_mean": 420.125,
    "epsilon_norm": 3.567262649536133,
    "prediction_norm": 2.8616788387298584,
    "timestamp": "2025-07-14T12:35:21.453094"
  },
  {
    "epoch": 2993,
    "batch_idx": 0,
    "loss": 0.31565988063812256,
    "learning_rate": 0.0004945023236383921,
    "timestep_mean": 592.46875,
    "epsilon_norm": 8.607421875,
    "prediction_norm": 7.150370121002197,
    "timestamp": "2025-07-14T12:35:21.458441"
  },
  {
    "epoch": 2993,
    "batch_idx": 32,
    "loss": 0.22085261344909668,
    "learning_rate": 0.0004945023236383921,
    "timestep_mean": 523.15625,
    "epsilon_norm": 7.495879650115967,
    "prediction_norm": 6.510871887207031,
    "timestamp": "2025-07-14T12:35:21.463725"
  },
  {
    "epoch": 2993,
    "batch_idx": 64,
    "loss": 0.46416109800338745,
    "learning_rate": 0.0004945023236383921,
    "timestep_mean": 477.8125,
    "epsilon_norm": 8.51966667175293,
    "prediction_norm": 6.535028457641602,
    "timestamp": "2025-07-14T12:35:21.468861"
  },
  {
    "epoch": 2993,
    "batch_idx": 96,
    "loss": 0.48066622018814087,
    "learning_rate": 0.0004945023236383921,
    "timestep_mean": 486.9375,
    "epsilon_norm": 8.298907279968262,
    "prediction_norm": 6.164234638214111,
    "timestamp": "2025-07-14T12:35:21.473947"
  },
  {
    "epoch": 2993,
    "batch_idx": 128,
    "loss": 0.29759353399276733,
    "learning_rate": 0.0004945023236383921,
    "timestep_mean": 540.15625,
    "epsilon_norm": 8.065595626831055,
    "prediction_norm": 6.457754611968994,
    "timestamp": "2025-07-14T12:35:21.479048"
  },
  {
    "epoch": 2993,
    "batch_idx": 160,
    "loss": 0.11343146860599518,
    "learning_rate": 0.0004945023236383921,
    "timestep_mean": 558.4375,
    "epsilon_norm": 8.808414459228516,
    "prediction_norm": 8.530016899108887,
    "timestamp": "2025-07-14T12:35:21.484228"
  },
  {
    "epoch": 2993,
    "batch_idx": 192,
    "loss": 0.5145835876464844,
    "learning_rate": 0.0004945023236383921,
    "timestep_mean": 465.125,
    "epsilon_norm": 3.9653520584106445,
    "prediction_norm": 3.3920271396636963,
    "timestamp": "2025-07-14T12:35:21.489331"
  },
  {
    "epoch": 2994,
    "batch_idx": 0,
    "loss": 0.15446943044662476,
    "learning_rate": 0.0004952876807834293,
    "timestep_mean": 534.84375,
    "epsilon_norm": 8.556221961975098,
    "prediction_norm": 8.234134674072266,
    "timestamp": "2025-07-14T12:35:21.494510"
  },
  {
    "epoch": 2994,
    "batch_idx": 32,
    "loss": 0.17445559799671173,
    "learning_rate": 0.0004952876807834293,
    "timestep_mean": 573.46875,
    "epsilon_norm": 8.15009880065918,
    "prediction_norm": 7.0107340812683105,
    "timestamp": "2025-07-14T12:35:21.499637"
  },
  {
    "epoch": 2994,
    "batch_idx": 64,
    "loss": 0.15544749796390533,
    "learning_rate": 0.0004952876807834293,
    "timestep_mean": 570.3125,
    "epsilon_norm": 6.824146747589111,
    "prediction_norm": 6.751227855682373,
    "timestamp": "2025-07-14T12:35:21.504527"
  },
  {
    "epoch": 2994,
    "batch_idx": 96,
    "loss": 0.16126325726509094,
    "learning_rate": 0.0004952876807834293,
    "timestep_mean": 551.96875,
    "epsilon_norm": 8.925497055053711,
    "prediction_norm": 8.646928787231445,
    "timestamp": "2025-07-14T12:35:21.509502"
  },
  {
    "epoch": 2994,
    "batch_idx": 128,
    "loss": 0.19841566681861877,
    "learning_rate": 0.0004952876807834293,
    "timestep_mean": 554.46875,
    "epsilon_norm": 8.246252059936523,
    "prediction_norm": 7.642999172210693,
    "timestamp": "2025-07-14T12:35:21.514372"
  },
  {
    "epoch": 2994,
    "batch_idx": 160,
    "loss": 0.25622618198394775,
    "learning_rate": 0.0004952876807834293,
    "timestep_mean": 534.09375,
    "epsilon_norm": 7.664261817932129,
    "prediction_norm": 6.475537300109863,
    "timestamp": "2025-07-14T12:35:21.519247"
  },
  {
    "epoch": 2994,
    "batch_idx": 192,
    "loss": 0.31681060791015625,
    "learning_rate": 0.0004952876807834293,
    "timestep_mean": 362.0,
    "epsilon_norm": 3.82137393951416,
    "prediction_norm": 3.084146499633789,
    "timestamp": "2025-07-14T12:35:21.524075"
  },
  {
    "epoch": 2995,
    "batch_idx": 0,
    "loss": 0.443589985370636,
    "learning_rate": 0.0004960730495556456,
    "timestep_mean": 431.9375,
    "epsilon_norm": 7.812287330627441,
    "prediction_norm": 5.882915019989014,
    "timestamp": "2025-07-14T12:35:21.529000"
  },
  {
    "epoch": 2995,
    "batch_idx": 32,
    "loss": 0.23147320747375488,
    "learning_rate": 0.0004960730495556456,
    "timestep_mean": 478.78125,
    "epsilon_norm": 7.757184982299805,
    "prediction_norm": 6.390911102294922,
    "timestamp": "2025-07-14T12:35:21.533943"
  },
  {
    "epoch": 2995,
    "batch_idx": 64,
    "loss": 0.24367687106132507,
    "learning_rate": 0.0004960730495556456,
    "timestep_mean": 593.03125,
    "epsilon_norm": 8.089072227478027,
    "prediction_norm": 6.627996444702148,
    "timestamp": "2025-07-14T12:35:21.538826"
  },
  {
    "epoch": 2995,
    "batch_idx": 96,
    "loss": 0.343964546918869,
    "learning_rate": 0.0004960730495556456,
    "timestep_mean": 537.34375,
    "epsilon_norm": 8.77531623840332,
    "prediction_norm": 6.827768325805664,
    "timestamp": "2025-07-14T12:35:21.543722"
  },
  {
    "epoch": 2995,
    "batch_idx": 128,
    "loss": 0.5310529470443726,
    "learning_rate": 0.0004960730495556456,
    "timestep_mean": 431.90625,
    "epsilon_norm": 8.8252592086792,
    "prediction_norm": 6.934788227081299,
    "timestamp": "2025-07-14T12:35:21.548612"
  },
  {
    "epoch": 2995,
    "batch_idx": 160,
    "loss": 0.5148910284042358,
    "learning_rate": 0.0004960730495556456,
    "timestep_mean": 476.78125,
    "epsilon_norm": 8.998339653015137,
    "prediction_norm": 6.680262565612793,
    "timestamp": "2025-07-14T12:35:21.553543"
  },
  {
    "epoch": 2995,
    "batch_idx": 192,
    "loss": 0.37893399596214294,
    "learning_rate": 0.0004960730495556456,
    "timestep_mean": 481.0,
    "epsilon_norm": 4.207242012023926,
    "prediction_norm": 2.877332925796509,
    "timestamp": "2025-07-14T12:35:21.558584"
  },
  {
    "epoch": 2996,
    "batch_idx": 0,
    "loss": 0.23038354516029358,
    "learning_rate": 0.0004968584280172221,
    "timestep_mean": 502.28125,
    "epsilon_norm": 7.4568586349487305,
    "prediction_norm": 6.200111389160156,
    "timestamp": "2025-07-14T12:35:21.563788"
  },
  {
    "epoch": 2996,
    "batch_idx": 32,
    "loss": 0.26398563385009766,
    "learning_rate": 0.0004968584280172221,
    "timestep_mean": 464.78125,
    "epsilon_norm": 8.081226348876953,
    "prediction_norm": 6.470036029815674,
    "timestamp": "2025-07-14T12:35:21.568917"
  },
  {
    "epoch": 2996,
    "batch_idx": 64,
    "loss": 0.14285969734191895,
    "learning_rate": 0.0004968584280172221,
    "timestep_mean": 494.3125,
    "epsilon_norm": 8.143577575683594,
    "prediction_norm": 7.469470977783203,
    "timestamp": "2025-07-14T12:35:21.574056"
  },
  {
    "epoch": 2996,
    "batch_idx": 96,
    "loss": 0.2879258096218109,
    "learning_rate": 0.0004968584280172221,
    "timestep_mean": 445.96875,
    "epsilon_norm": 8.542734146118164,
    "prediction_norm": 7.0719313621521,
    "timestamp": "2025-07-14T12:35:21.579153"
  },
  {
    "epoch": 2996,
    "batch_idx": 128,
    "loss": 0.2485143542289734,
    "learning_rate": 0.0004968584280172221,
    "timestep_mean": 479.65625,
    "epsilon_norm": 6.842309474945068,
    "prediction_norm": 5.759347915649414,
    "timestamp": "2025-07-14T12:35:21.584613"
  },
  {
    "epoch": 2996,
    "batch_idx": 160,
    "loss": 0.2730945348739624,
    "learning_rate": 0.0004968584280172221,
    "timestep_mean": 472.96875,
    "epsilon_norm": 8.926069259643555,
    "prediction_norm": 6.997396469116211,
    "timestamp": "2025-07-14T12:35:21.589736"
  },
  {
    "epoch": 2996,
    "batch_idx": 192,
    "loss": 0.1613008677959442,
    "learning_rate": 0.0004968584280172221,
    "timestep_mean": 357.0,
    "epsilon_norm": 3.5844085216522217,
    "prediction_norm": 2.9742891788482666,
    "timestamp": "2025-07-14T12:35:21.594901"
  },
  {
    "epoch": 2997,
    "batch_idx": 0,
    "loss": 0.30409467220306396,
    "learning_rate": 0.0004976438142303148,
    "timestep_mean": 427.6875,
    "epsilon_norm": 7.683530807495117,
    "prediction_norm": 7.418321132659912,
    "timestamp": "2025-07-14T12:35:21.600042"
  },
  {
    "epoch": 2997,
    "batch_idx": 32,
    "loss": 0.30976301431655884,
    "learning_rate": 0.0004976438142303148,
    "timestep_mean": 411.375,
    "epsilon_norm": 7.502396106719971,
    "prediction_norm": 6.16914701461792,
    "timestamp": "2025-07-14T12:35:21.605231"
  },
  {
    "epoch": 2997,
    "batch_idx": 64,
    "loss": 0.4413897693157196,
    "learning_rate": 0.0004976438142303148,
    "timestep_mean": 487.5,
    "epsilon_norm": 8.447099685668945,
    "prediction_norm": 6.215551853179932,
    "timestamp": "2025-07-14T12:35:21.610385"
  },
  {
    "epoch": 2997,
    "batch_idx": 96,
    "loss": 0.16940221190452576,
    "learning_rate": 0.0004976438142303148,
    "timestep_mean": 571.53125,
    "epsilon_norm": 6.746286392211914,
    "prediction_norm": 5.542312145233154,
    "timestamp": "2025-07-14T12:35:21.615525"
  },
  {
    "epoch": 2997,
    "batch_idx": 128,
    "loss": 0.24403218924999237,
    "learning_rate": 0.0004976438142303148,
    "timestep_mean": 423.5,
    "epsilon_norm": 8.515432357788086,
    "prediction_norm": 6.963890075683594,
    "timestamp": "2025-07-14T12:35:21.620766"
  },
  {
    "epoch": 2997,
    "batch_idx": 160,
    "loss": 0.30331408977508545,
    "learning_rate": 0.0004976438142303148,
    "timestep_mean": 502.59375,
    "epsilon_norm": 9.038805961608887,
    "prediction_norm": 7.837133884429932,
    "timestamp": "2025-07-14T12:35:21.625890"
  },
  {
    "epoch": 2997,
    "batch_idx": 192,
    "loss": 0.11838966608047485,
    "learning_rate": 0.0004976438142303148,
    "timestep_mean": 721.125,
    "epsilon_norm": 4.003679275512695,
    "prediction_norm": 3.191810131072998,
    "timestamp": "2025-07-14T12:35:21.630984"
  },
  {
    "epoch": 2998,
    "batch_idx": 0,
    "loss": 0.362581729888916,
    "learning_rate": 0.0004984292062570611,
    "timestep_mean": 565.65625,
    "epsilon_norm": 7.585656642913818,
    "prediction_norm": 6.280930519104004,
    "timestamp": "2025-07-14T12:35:21.635886"
  },
  {
    "epoch": 2998,
    "batch_idx": 32,
    "loss": 0.35926544666290283,
    "learning_rate": 0.0004984292062570611,
    "timestep_mean": 528.375,
    "epsilon_norm": 7.666529655456543,
    "prediction_norm": 6.441248893737793,
    "timestamp": "2025-07-14T12:35:21.640715"
  },
  {
    "epoch": 2998,
    "batch_idx": 64,
    "loss": 0.4063626527786255,
    "learning_rate": 0.0004984292062570611,
    "timestep_mean": 358.90625,
    "epsilon_norm": 7.919997692108154,
    "prediction_norm": 5.939764022827148,
    "timestamp": "2025-07-14T12:35:21.646027"
  },
  {
    "epoch": 2998,
    "batch_idx": 96,
    "loss": 0.2763356566429138,
    "learning_rate": 0.0004984292062570611,
    "timestep_mean": 520.53125,
    "epsilon_norm": 7.229475498199463,
    "prediction_norm": 5.994662761688232,
    "timestamp": "2025-07-14T12:35:21.651282"
  },
  {
    "epoch": 2998,
    "batch_idx": 128,
    "loss": 0.27328625321388245,
    "learning_rate": 0.0004984292062570611,
    "timestep_mean": 436.40625,
    "epsilon_norm": 8.23275089263916,
    "prediction_norm": 6.750453948974609,
    "timestamp": "2025-07-14T12:35:21.656289"
  },
  {
    "epoch": 2998,
    "batch_idx": 160,
    "loss": 0.16999423503875732,
    "learning_rate": 0.0004984292062570611,
    "timestep_mean": 542.09375,
    "epsilon_norm": 6.773622035980225,
    "prediction_norm": 5.997416019439697,
    "timestamp": "2025-07-14T12:35:21.661401"
  },
  {
    "epoch": 2998,
    "batch_idx": 192,
    "loss": 0.21079978346824646,
    "learning_rate": 0.0004984292062570611,
    "timestep_mean": 495.375,
    "epsilon_norm": 3.3540914058685303,
    "prediction_norm": 2.8537991046905518,
    "timestamp": "2025-07-14T12:35:21.666370"
  },
  {
    "epoch": 2999,
    "batch_idx": 0,
    "loss": 0.23268726468086243,
    "learning_rate": 0.0004992146021595859,
    "timestep_mean": 505.71875,
    "epsilon_norm": 7.9642181396484375,
    "prediction_norm": 6.948668003082275,
    "timestamp": "2025-07-14T12:35:21.671355"
  },
  {
    "epoch": 2999,
    "batch_idx": 32,
    "loss": 0.24486437439918518,
    "learning_rate": 0.0004992146021595859,
    "timestep_mean": 588.15625,
    "epsilon_norm": 7.800224304199219,
    "prediction_norm": 6.801667213439941,
    "timestamp": "2025-07-14T12:35:21.676298"
  },
  {
    "epoch": 2999,
    "batch_idx": 64,
    "loss": 0.39579343795776367,
    "learning_rate": 0.0004992146021595859,
    "timestep_mean": 497.5625,
    "epsilon_norm": 8.530603408813477,
    "prediction_norm": 7.044308662414551,
    "timestamp": "2025-07-14T12:35:21.681463"
  },
  {
    "epoch": 2999,
    "batch_idx": 96,
    "loss": 0.27419373393058777,
    "learning_rate": 0.0004992146021595859,
    "timestep_mean": 535.40625,
    "epsilon_norm": 9.405749320983887,
    "prediction_norm": 7.827719688415527,
    "timestamp": "2025-07-14T12:35:21.686390"
  },
  {
    "epoch": 2999,
    "batch_idx": 128,
    "loss": 0.16506758332252502,
    "learning_rate": 0.0004992146021595859,
    "timestep_mean": 568.25,
    "epsilon_norm": 6.952788352966309,
    "prediction_norm": 6.3306565284729,
    "timestamp": "2025-07-14T12:35:21.691285"
  },
  {
    "epoch": 2999,
    "batch_idx": 160,
    "loss": 0.3935713768005371,
    "learning_rate": 0.0004992146021595859,
    "timestep_mean": 498.25,
    "epsilon_norm": 7.050017833709717,
    "prediction_norm": 5.844621181488037,
    "timestamp": "2025-07-14T12:35:21.696175"
  },
  {
    "epoch": 2999,
    "batch_idx": 192,
    "loss": 0.29544326663017273,
    "learning_rate": 0.0004992146021595859,
    "timestep_mean": 616.625,
    "epsilon_norm": 4.658020496368408,
    "prediction_norm": 4.237834453582764,
    "timestamp": "2025-07-14T12:35:21.701171"
  }
]