README: point submission links to Hugging Face Space repo URLs.

Prepare hackathon submission: sandbox-default notebook, single README, committed results.

Add env.step safety net so checkpoint hallucinations never crash evaluation.

Harden checkpoint sanitizer against all sandbox/sim edge cases.

Fix HF Space no-docker sandbox launcher shutdown compatibility.

Finalize checkpoint-first evaluation and sandbox submission workflow.

Improve submission docs and training run transparency.

Improve checkpoint policy quality and environment robustness.

Fix checkpoint evaluation reliability and keep heuristic as strict fallback.

audit pass: hackathon-ready polish and truthful reporting

strengthen strict JSON reward shaping for GRPO

harden reward callback and add training diagnostics

tune GRPO batch and accumulation settings

harden GRPO reward shaping and stabilize updates

chore: stop ignoring private sandbox implementation doc

refactor: ignore private documentation and parameterize training batch size in train.py

Polish judge-facing docs and add checkpoint evaluation.

Harden training signal and clean generated artifacts.

Fix Unsloth GRPOTrainer keyword compatibility.

Fix TRL GRPO compatibility and Unsloth import ordering.

Harden hackathon readiness and surface reproducible evidence.

add nbformat cell ids to train_notebook.ipynb

revert train_notebook.ipynb to Colab-compatible baseline

revert train.py to Colab-compatible baseline

add secure HF token handling for DGX training

harden training for shared DGX GPUs

make Step 3 sanity check resilient on DGX GPUs

make DGX training resilient without bitsandbytes

upgrade bitsandbytes and wire torch CUDA runtime for DGX

repair broken aiohttp/charset stack for unsloth imports

force wheel-only DGX install path in notebook

fix DGX notebook dependency resolution failures

stabilize DGX notebook dependency installation

disable external report logging by default in GRPO training

auto-select fp16 on non-Ampere GPUs for GRPO training

fix GRPO config/trainer compatibility across TRL versions

enforce unsloth-first colab training and preflight checks

stabilize colab training deps with robust fallback path

fix colab GRPO dependency setup and error guidance

fix single-worker deployment for stable session state

fix colab training import by adding mergekit dependency

fix UI button initialization regression on Spaces

fix ci path handling for data-backed tests

updated EICC v2 environment, APIs, and training pipeline

updated Frontend

updated Readme.md

added error toasts

updated Readme.md

updated frontend

updated frontend