# Runtime artifacts runs/ logs/ checkpoints/ outputs/ wandb/ tensorboard/ *.log *.pt *.ckpt *.safetensors *.bin *.core core.* # Large generated pretraining corpus: keep code/manifests in git, keep corpus local. dataset/pretrain/data/* !dataset/pretrain/data/.gitkeep # Python/cache/editor __pycache__/ *.py[cod] .DS_Store .env .venv/