---
license: cc
pipeline_tag: text-to-audio
---

# Phonsa

This is the official weights of Phonsa, presented in [MPEcho: A Melody and Phoneme-Aware Generative Framework for Controllable Cover Song Generation](https://huggingface.co/papers/2607.26698).

The Phonsa is developed and trained by Hsuan-Yu Yeh.

Code: https://github.com/YatingMusic/Phonsa