Datasets:

jiaheillu
/

sovits_audio_preview

# 胡桃语音模型训练参数



duration of dataset：40min 

epoch: 3500

batch_size :21

segments = 221

step:36800

Losses vs step

![截屏2023-04-05 02.38.41](./params.assets/%E6%88%AA%E5%B1%8F2023-04-05%2002.38.41.png)