codearena-rl / server

Commit History

  1. Add Ollama fallback support to ai_fixer

    havinashpatil
    05d943b
  2. Add Hugging Face Inference API fallback for AI fixer

    havinashpatil
    18261cc
  3. Fix builtin syntax fixer for inline one-line headers and validate TGI fixes

    havinashpatil
    0c0a8ff
  4. Fix TGI endpoint plumbing and use full TGI runtime image in Dockerfile

    havinashpatil
    a8bc575
  5. Add TGI integration for cloud LLM serving⁠

    havinashpatil
    434afdf
  6. Serve React frontend from FastAPI backend for one-click access

    havinashpatil
    62c7e95
  7. Finalizing CodeArena RL Benchmark: frontend improvements, GRPO training scripts, and cleaned environment

    havinashpatil
    03a7eb9

  1. Complete all tasks: Adaptive curriculum, GRPO, React frontend, LLM-as-a-judge

    havinashpatil
    a448db8
  2. Upload meta

    havinashpatil
    03defc2

  1. fix: reset task_id parsing, grader tuple crash fallback, and inference score output

    646409d
  2. fix: clamp scores to strict (0.001, 0.999) range

    b3485c6
  3. fix: openenv.yaml tasks graders + strict score bounds

    9967cb5
  4. fix: add safe_reward() clamp at every reward return point

    f14f8d9
  5. Add normalize_reward(); enforce strict (0,1) on all paths

    b53855f
  6. Update grading logic to strict (0,1) ratio fallback

    78b7e6b
  7. fix: OpenEnv multi-mode compliance — add main() entrypoint and uv.lock

    e92bfc1
  8. Production-ready: add server/app.py with fallback-safe /reset, fix Dockerfile, add HF metadata, add task JSON files

    dcc8fa3
  9. Initial commit

    54a19c9