Discrete Audio Pretrain Data Collection Pretraining Discrete Audio Data -- Interleaved Sequence using Mimi (8 codebooks) • 3 items • Updated about 4 hours ago
SODA Models Collection SODA models trained on Yodas+Emilia+Nemotron for 500B tokens from scratch • 6 items • Updated about 4 hours ago
SODA Warm-Start (Qwen3-based) Models Collection SODA models trained on Yodas+Emilia+Nemotron for 500B tokens with a Qwen3 initialization • 3 items • Updated about 4 hours ago
Discrete Audio IsoFLOP Models Collection IsoFLOP Models trained on Yodas+Emilia+Nemotron from budgets of 3e18 to 3e20 • 64 items • Updated about 4 hours ago