search

huawei-noah / Speech-Backbones

This is the main repository of open-sourced speech technology by Huawei Noah's Ark Lab.
545 stars 113 forks source link

support for bigvgan #27

Closed eschmidbauer closed 1 year ago

eschmidbauer commented 1 year ago

Is it possible to add support for bigvgan instead of hifigan?

li1jkdaw commented 1 year ago

Hi, @eschmidbauer ! We don't plan to release new Grad-TTS models for this BigVGAN vocoder. But you can train a Grad-TTS model compatible with this vocoder yourself by just replacing a few lines in params.py. I believe the only parameters you need to change are n_feats, n_fft, sample_rate, hop_length, win_length, f_min and f_max, just make them equal to the corresponding parameters from json config of the BigVGAN version you want to use.

eschmidbauer commented 1 year ago

thanks! i was able to figure it out