Update README.md

Update README.md

fix: remove all nested duplicate folders

fix: correct W&B metrics in README

sync: hf-space is now the single source of truth

fix: Gradio mount, verifier stage3, poc_history clear, real notebook

rewrite: training notebook - drop Unsloth, use transformers+peft+bitsandbytes

fix: STAGE_FILES points stages 2-3 to actual app files + update reward_curve (LFS) + results table

feat: GRPO training, fixed curriculum stages 2+3, AST verifier, leaderboard UI, openenv schema

docs: clarify inference eval vs GRPO training

fix: remove 3 debug cells from notebook

fix: clarify Y-axis raw vs normalized reward

feat: add real training notebook from Colab run

fix: png rename + remove placeholder + fix blog link + add baseline comparison

fix: rename reward curve to .png to match README

docs: real results + reward curve + W&B link

docs: add W&B link and results

docs: complete README with full project description

fix: add openenv stub, fix requirements, remove subfolder

fix: requirements, remove stub openenv, clean pycache, fix models

fix: correct import names + full 3-phase step logic

Fix Observation import in server.py

fix: add app.py FastAPI wrapper over OpenEnv class

fix: use requirements.txt + correct uvicorn entry

teammate: models + judge v3-fixed

fix: use OpenEnv serve() method instead of uvicorn app

fix: stub models+judge so Space boots — teammate will override

fix: use uvicorn directly instead of openenv CLI

readme: fix yaml header no emoji

readme: fix HF Space yaml header

readme: add HF Space yaml metadata header

initial deploy — jairaj files only, teammate files pending

initial deploy — jairaj files, teammate files pending

initial commit