PerceptionBenchmark / .gitignore
DariusGiannoli
data: add Middlebury stereo dataset (10 scenes, PNG+PFM via LFS)
ad7e281
# --- 1. THE BRAINS (Heavy Model Weights) ---
# Never commit large binary files. Users should download them via scripts.
models/*.pt
models/*.pth
models/*.onnx
# --- 2. THE DATA (Datasets) ---
# Middlebury stereo benchmark scenes are tracked via Git LFS
# Everything else in data/ (raw custom uploads etc.) stays local
data/*
!data/middlebury/
!data/README.md
# --- 3. THE RESULTS (Logs & Plots) ---
# You might want to keep the final report, but not every test run.
results/logs/
results/yolo/
# Optional: Uncomment if you don't want to track generated plots
results/plots/
# --- 3b. TRAINING RUNS ---
# YOLO training artifacts: batch images, curves, weights checkpoints
training/runs/
# --- 4. PYTHON JUNK (Generated files) ---
# These are compiled Python files that appear automatically.
__pycache__/
*.pyc
*.pyo
*.pyd
# --- 5. SYSTEM & IDE JUNK ---
# Mac OS specific annoying file
.DS_Store
# VS Code / PyCharm settings (Optional, but recommended)
.vscode/
.idea/
# Environment variables (Secrets)
.env
venv/
env/