Space T4 + real HF path: split requirements, GPU defaults, API reset behavior

Finalize eval-friendly defaults: offline baseline, deterministic API reset, docs cleanup

fix(api): Pydantic v2 step payload (exclude_none); env candidate_solutions; README Space API curl guide; Gradio API tab

fix(space): fastapi>=0.115.2 for Gradio 5.25; Pydantic v2 model_dump in API

Final hardening for 100% submission readiness

Final cleanup for FORGE-v4: Colab entrypoint, OpenEnv API, 10x optimization, and Judge Narrative generation