---
pipeline_tag: audio-to-audio
tags:
- roformer
- speech
---
Specially trained to separate the speaker's voice from background music that contains vocals.