Commit History

  1. 替换权重:S0B/S1B 底座重跑 SFT->CPO->GRPO(3000 prompt),recipe 修复 GRPO_LR 发散

    Verified signature
    1eb5314
  2. Upload README.md with huggingface_hub

    Verified signature
    0e76354

  1. Add files using upload-large-folder tool

    Verified signature
    03d337a
  2. code + tokenizer + model card

    Verified signature
    a4251f4
  3. initial commit

    Verified signature
    756396c