Commit History

  1. fix GRPOConfig: remove invalid max_length

    Verified signature
    6831b86
  2. GRPO compiler reward script

    Verified signature
    625c0f0

  1. SFT: 4 epochs larger curriculum

    Verified signature
    d5c7535
  2. fix DPOConfig: drop max_prompt_length

    Verified signature
    2ac14b9
  3. DPO: merge SFT LoRA then train

    Verified signature
    5fb8489
  4. fix Trackio crash: push at end only

    Verified signature
    927b117
  5. fix Trackio crash: push at end only

    Verified signature
    95f1e85
  6. DPO script

    Verified signature
    332a987
  7. SFT script

    Verified signature
    17a6842
  8. initial commit

    Verified signature
    a41fb61