scugnizz-llama-training / agent_loop.py

Commit History

  1. Upload agent_loop.py with huggingface_hub

    Verified signature
    102ba2a

  1. thin agent_loop: no answer nudge (train post-tool answers)

    Verified signature
    9143c83
  2. agent_loop: same-tool re-call after success → answer nudge

    Verified signature
    959aa1a
  3. Hermes pass-through tool content (full payloads)

    Verified signature
    35fe76f
  4. Upload agent_loop.py with huggingface_hub

    Verified signature
    a0d3459

  1. harness: call-loop guard + answer nudge

    Verified signature
    38557ea
  2. Upload agent_loop.py with huggingface_hub

    Verified signature
    e5ec075