notebook: embed executed GRPO run outputs + plot

deploy: full Space sync — app + server + training + GRPO run artifacts

make GRPOConfig kwargs version-tolerant

add HF Jobs GRPO training script

wire Colab notebook to TRL GRPOTrainer (real LoRA weight updates)

hackathon submission: theme-aligned README, refreshed training, Qwen vs random plots

Add CUDA backend (transformers) for HF Jobs A10G runs

Deploy ClaimSense adjudication gym