ML Playground

This directory is for exploratory and model-preparation work only.

Current contents include:

• CLIP/: CLIP/mobileclip notebooks and experiments.
• YOLOv5Face/: YOLOv5Face notebooks and related assets.
• data/: local sample images used by notebooks.

Running notebooks

1. Install uv.
2. From repo root, run uv sync --project infra/ml.
3. In VS Code/Jupyter, use kernel infra/ml/.venv/bin/python.

Notebook hygiene

Clear notebook outputs before committing to keep diffs readable and stable.