__pycache__/
*.pyc
.DS_Store
Thumbs.db

# Cloned CLDF source repositories (large, not ours to commit)
sources/

# Training data (too large for git, regenerated from scripts)
data/training/lexicons/
data/training/cognate_pairs/

# Large raw downloads (regenerated by ingest scripts)
data/training/raw/linear_b/UnicodeData.txt
# data/training/validation/ — now tracked via Git LFS