README.md

metadata

title: Audio Lyrics Extractor
emoji: 🎵
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 4.0.2
app_file: app.py
pinned: false

Audio Lyrics Extractor

This application combines Demucs for vocal isolation and Whisper for lyrics transcription. Upload any song to:

1. Extract vocals from the music
2. Transcribe the lyrics

Features

• Vocal isolation using Demucs
• Lyrics transcription using Whisper
• Support for various audio formats
• GPU acceleration when available

Models

• Demucs (htdemucs) for vocal separation
• Whisper (medium) for transcription

Usage

1. Upload an audio file
2. Select Whisper model size
3. Get isolated vocals and transcribed lyrics