Add --hub-resume and --step-offset to allow resuming directly from HF Hub PEFT checkpoints

Revert speed optimizations to prioritize real-world model quality

Optimize training speed for T4 GPUs

Fix HF_TOKEN scope issue for CheckpointPushCallback

Made changes to train_grpo

Fix links to match correct usernames/orgs

Fix checkpoint persistence, add leaderboard and update HF links

fix: use GitHub raw URLs for images so README renders on HF Space

Added blog post

Added blog post

Added readme again

Added readme

Update: Added final imporvements for hackathon

chore: clean up local dev files and temporary virtual environments

chore: clean up local dev files and temporary virtual environments

Update: Triggering the full run

Fix: batch%num_generations math

Cuda returns false fixed

COMPUTE_DRIVE fix

Fix: Removed BitsandBytes

Fix: Fixed dependancy issues

Revert "Fix: Dockerfile"

Fix: Dockerfile

Fix: Fixed again again

Fix: Fixed again

Fix: Fixing Again

Fix: Fixing

Fix: Trying to fix dependency issues

Fix: Fixed file

fix: serialize bug_metadata as JSON to fix pyarrow mixed-type error

fix: upgrade bitsandbytes>=0.49.0 (triton.ops), switch to Qwen2.5-Coder-3B

fix: torch at build time, remove mergekit (conflicts accelerate/peft/trl)

fix: empty requirements.txt, install training deps at runtime

fix: remove wandb - click conflict with gradio causes resolution-too-deep

fix: resolve pip dependency conflicts for HF Spaces build

Fix: loosen strict dependencies to prevent pip backtracking

fix: remove hardcoded torch from requirements for HF space

chore: normalize dataset inputs and fix mergekit dependency for TRL 0.14.0

Add HANDOVER.md: full project state, deps, training instructions, known fixes

Auto-detect GPU: bfloat16+batch2+gen8 on A100, float16+batch1+gen4 on T4 — same script works on both

Reduce max_completion_length to 160 for T4 speed: target 1000 steps in <8hrs

Fix: bump bitsandbytes to 0.45.3 for CUDA 12.x support on Kaggle T4

Optimize for Kaggle P100: float16, batch=1, grad_accum=8, num_gen=4, max_completion=256, lora_r=8

Fix GRPOConfig: rename max_new_tokens to max_completion_length for trl==0.14.0

Update: Added testing

Align gradio version with Hugging Face Space builder2

Add dockerignore to reduce Space build context

Stabilize Space runtime: pin ML deps and disable runtime package drift

Pin torch to cu121 build + use model.device instead of hardcoded cuda string

Replace unsloth with bitsandbytes+peft: fixes CUDA driver incompatibility on HF A100