Upload DeepseekV3ForCausalLM
Browse files- config.json +5 -33
- model.safetensors +2 -2
config.json
CHANGED
|
@@ -4,15 +4,9 @@
|
|
| 4 |
],
|
| 5 |
"attention_bias": false,
|
| 6 |
"attention_dropout": 0.0,
|
| 7 |
-
"auto_map": {
|
| 8 |
-
"AutoConfig": "configuration_deepseek.DeepseekV3Config",
|
| 9 |
-
"AutoModel": "modeling_deepseek.DeepseekV3Model",
|
| 10 |
-
"AutoModelForCausalLM": "modeling_deepseek.DeepseekV3ForCausalLM"
|
| 11 |
-
},
|
| 12 |
"bos_token_id": 0,
|
| 13 |
-
"dtype": "
|
| 14 |
"eos_token_id": 1,
|
| 15 |
-
"ep_size": 1,
|
| 16 |
"first_k_dense_replace": 3,
|
| 17 |
"head_dim": 64,
|
| 18 |
"hidden_act": "silu",
|
|
@@ -20,10 +14,9 @@
|
|
| 20 |
"initializer_range": 0.02,
|
| 21 |
"intermediate_size": 32,
|
| 22 |
"kv_lora_rank": 512,
|
| 23 |
-
"max_position_embeddings":
|
| 24 |
"model_type": "deepseek_v3",
|
| 25 |
"moe_intermediate_size": 2048,
|
| 26 |
-
"moe_layer_freq": 1,
|
| 27 |
"n_group": 8,
|
| 28 |
"n_routed_experts": 256,
|
| 29 |
"n_shared_experts": 1,
|
|
@@ -32,41 +25,20 @@
|
|
| 32 |
"num_experts_per_tok": 8,
|
| 33 |
"num_hidden_layers": 2,
|
| 34 |
"num_key_value_heads": 2,
|
| 35 |
-
"num_nextn_predict_layers": 1,
|
| 36 |
"pretraining_tp": 1,
|
| 37 |
"q_lora_rank": 1536,
|
| 38 |
"qk_head_dim": 192,
|
| 39 |
"qk_nope_head_dim": 128,
|
| 40 |
"qk_rope_head_dim": 64,
|
| 41 |
-
"quantization_config": {
|
| 42 |
-
"activation_scheme": "dynamic",
|
| 43 |
-
"fmt": "e4m3",
|
| 44 |
-
"quant_method": "fp8",
|
| 45 |
-
"weight_block_size": [
|
| 46 |
-
128,
|
| 47 |
-
128
|
| 48 |
-
]
|
| 49 |
-
},
|
| 50 |
"rms_norm_eps": 1e-06,
|
| 51 |
"rope_interleave": true,
|
| 52 |
-
"rope_scaling":
|
| 53 |
-
|
| 54 |
-
"beta_slow": 1.0,
|
| 55 |
-
"factor": 40.0,
|
| 56 |
-
"mscale": 1.0,
|
| 57 |
-
"mscale_all_dim": 1.0,
|
| 58 |
-
"original_max_position_embeddings": 4096,
|
| 59 |
-
"rope_type": "yarn",
|
| 60 |
-
"type": "yarn"
|
| 61 |
-
},
|
| 62 |
-
"rope_theta": 10000,
|
| 63 |
"routed_scaling_factor": 2.5,
|
| 64 |
-
"scoring_func": "sigmoid",
|
| 65 |
"tie_word_embeddings": false,
|
| 66 |
"topk_group": 4,
|
| 67 |
-
"topk_method": "noaux_tc",
|
| 68 |
"transformers_version": "4.57.0.dev0",
|
| 69 |
"use_cache": true,
|
| 70 |
"v_head_dim": 128,
|
| 71 |
-
"vocab_size":
|
| 72 |
}
|
|
|
|
| 4 |
],
|
| 5 |
"attention_bias": false,
|
| 6 |
"attention_dropout": 0.0,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
"bos_token_id": 0,
|
| 8 |
+
"dtype": "float32",
|
| 9 |
"eos_token_id": 1,
|
|
|
|
| 10 |
"first_k_dense_replace": 3,
|
| 11 |
"head_dim": 64,
|
| 12 |
"hidden_act": "silu",
|
|
|
|
| 14 |
"initializer_range": 0.02,
|
| 15 |
"intermediate_size": 32,
|
| 16 |
"kv_lora_rank": 512,
|
| 17 |
+
"max_position_embeddings": 4096,
|
| 18 |
"model_type": "deepseek_v3",
|
| 19 |
"moe_intermediate_size": 2048,
|
|
|
|
| 20 |
"n_group": 8,
|
| 21 |
"n_routed_experts": 256,
|
| 22 |
"n_shared_experts": 1,
|
|
|
|
| 25 |
"num_experts_per_tok": 8,
|
| 26 |
"num_hidden_layers": 2,
|
| 27 |
"num_key_value_heads": 2,
|
|
|
|
| 28 |
"pretraining_tp": 1,
|
| 29 |
"q_lora_rank": 1536,
|
| 30 |
"qk_head_dim": 192,
|
| 31 |
"qk_nope_head_dim": 128,
|
| 32 |
"qk_rope_head_dim": 64,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
"rms_norm_eps": 1e-06,
|
| 34 |
"rope_interleave": true,
|
| 35 |
+
"rope_scaling": null,
|
| 36 |
+
"rope_theta": 10000.0,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
"routed_scaling_factor": 2.5,
|
|
|
|
| 38 |
"tie_word_embeddings": false,
|
| 39 |
"topk_group": 4,
|
|
|
|
| 40 |
"transformers_version": "4.57.0.dev0",
|
| 41 |
"use_cache": true,
|
| 42 |
"v_head_dim": 128,
|
| 43 |
+
"vocab_size": 128815
|
| 44 |
}
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4029cdf7a5f9a0a7844863ac6bcf33b70438c50a78d7960d3b7ba479c2557dc6
|
| 3 |
+
size 22069200
|