Add interactive frontend UI

Update runtime requirements

Merge origin/main into main

fix: add root status endpoint

docs: trim readme to submission essentials

docs: restore hugging face space metadata

Update README.md by removing unnecessary metadata

Remove author submission line from BLOG.md

docs: replace private wandb link with exported logs

docs: remove internal checklist from readme

docs: add separate mini blog writeup

docs: rewrite readme as mini blog

docs: add final submission readme and training logs

fix: load model snapshots from writable cache

fix: load peft base model with explicit cache

fix: pass explicit cache dir to model loader

fix: set model cache dirs before torch imports

feat: add optional trained model inference to env space

fix: reduce grpo training runtime

fix: parse chat completions in rewards

fix: emit prompt column for grpo trainer

fix: add grpo warnings compatibility attr

fix: choose writable cache root at space startup

fix: use writable runtime dirs in training space

fix: patch torchao dtype imports for unsloth

fix: shim torch._pytree.register_constant for torchao compat

fix: use PyTorch nightly for torchao compat

fix: switch to cu124 wheel index for torch 2.6.0

fix: pin torch==2.6.0 to satisfy torchao torch.int1 requirement

fix: verbose per-step logs and BaseException catch in _run_training

fix: capture transformers/TRL/unsloth logs in Gradio UI

fix: add unsloth_zoo as explicit dependency before unsloth

fix: install unsloth in Docker image at build time

fix: chunk-read pip output splitting on \r and \n

feat: stream pip install output live to UI logs

fix: install unsloth to /app/pkgs to bypass HF Space permission issue

fix: robust unsloth install — PyPI first, git fallback, log errors

feat: add A100 training Space with manual-trigger Gradio UI

feat: optimize Space — dockerignore, dotenv secrets, multi-worker uvicorn

fix: use lean server requirements in Dockerfile (exclude unsloth/torch)

chore: add HF Space frontmatter and README draft

feat: bulletproof GRPO training script + Colab notebook

feat: rewrite env to be fully openenv-core compliant

chore: update baseline results with real facts (450 QA pairs)

feat: implement data pipeline

feat: baseline eval and GRPO training script

Updated .gitignore

chore: add CLAUDE.md, README, update gitignore

feat: implement actions, reward, env, server

chore: pin dependencies