- Install: Double-click
install.batand wait - Run: Double-click
run.bat - Use: Browser opens automatically at http://localhost:7865
- Install: Run
./install.shin terminal - Run: Run
./run.shin terminal - Use: Open http://localhost:7865 in your browser
What happens during installation:
- Creates isolated Python environment
- Installs PyTorch and dependencies
- Takes 10-30 minutes depending on internet speed
First run (one-time):
- Downloads AI model (~14GB)
- Takes 5-15 minutes
- Model is cached for future use
- Go to "Single File" tab
- Upload your song
- Click "Generate Caption"
- Caption appears instantly!
- Go to "Batch Processing" tab
- Enter folder path:
C:\Music\MyAlbum - Click "Process Folder"
- All songs get captioned automatically!
Edit music_captioner.py, line at bottom:
app.launch(server_port=7865, share=False)Edit config.py:
DEFAULT_PROMPT = "Your custom prompt here"In batch mode, change extensions:
.mp3, .wav, .flac, .m4a
- GPU = Faster: If you have NVIDIA GPU, processing is much faster
- Batch = Efficient: Process multiple files at once
- Prompt = Control: Customize prompts to get different styles of captions
- Save = Automatic: Captions save automatically as .txt files
Q: Where are captions saved? A: Same folder as the audio file, same name with .txt extension
Q: Can I change the caption format? A: Yes! Modify the prompt to change what information is included
Q: How long does processing take? A: With GPU: 5-10 seconds per song Without GPU: 30-60 seconds per song
Q: Can I run multiple instances? A: Yes! Each instance needs a different port number
- Check
README.mdfor detailed troubleshooting - Make sure Python 3.9+ is installed
- Ensure you have 20GB free disk space
- Try running installer again if it failed
Enjoy your music captioning! 🎵