Polish for hackathon submission: training evidence, two pipelines, UI, docs

Print every reward call so terminal shows continuous progress

Replace env-simulation reward with fast pure-heuristic to fix hang

Fix GRPO training: reward variance, batch/gen alignment, generation config

Update run_training.py and train_grpo.py, remove Dockerfile.training

fix: notebook uses compute_grpo_reward_env, updated hyperparams, no emojis

feat: curriculum training + Karnataka scenarios + repo cleanup

OpenGrid: Multi-agent POMDP power grid environment with GRPO training