---
datasets:
- litagin/moe-speech
language:
- ja
---
# Style-Bert-VITS2の300m相当の学習

![](mos_results\/mos_moe-speech-300m.png)