OPTIMIZE_SPACE.md

Space Optimization Guide

Current Space Usage: 5.3GB

Breakdown:

• movies/: 3.4GB (video files - expected)
• venv/: 2.0GB (virtual environment - can be optimized)

Optimization Options

1. Remove Unused Dependencies (Save ~500MB)

These packages are installed but NOT in requirements.txt and NOT used in code:

• transformers (115MB) - Not used
• opencv-python (179MB) - Not used (was for old NSFW detection)
• sympy (74MB) - Not used
• onnxruntime (53MB) - Not used
• yt-dlp (25MB) - Not used

To remove:

source venv/bin/activate
pip uninstall transformers opencv-python opencv-python-headless sympy onnxruntime yt-dlp -y

2. Optimize PyTorch (Already Done ✓)

You're already using CPU-only PyTorch (2.9.1+cpu), which is optimal.

3. Move Processed Videos

After cleaning videos, you can:

• Delete original videos (keep only cleaned versions)
• Move videos to external storage
• Use symbolic links if videos are elsewhere

4. Clean Cache Regularly

Run the cleanup script:

./cleanup_space.sh

Recommended Actions

1. Remove unused packages (saves ~500MB):

   source venv/bin/activate
   pip uninstall transformers opencv-python opencv-python-headless sympy onnxruntime yt-dlp -y

2. Move processed videos (saves 1.7GB if you delete originals):

   • Keep only *_cleaned.mp4 files
   • Delete original videos after verification

3. Total potential savings: ~2.2GB (from 5.3GB to ~3.1GB)