Skip to content

splinter21bak/bert-vits

 
 

Repository files navigation

bert-vits

vits with bert

put pytorch_model.bin to bert/roberta_wwm_ext_large/
Then resample all audio to 44100Hz,generate filelists.
Please read the code in preprocess_text.py to generate the specified format filelist.
Run preprocess_text.py
Run spec_gen.py
Run train_ms.py(Only multi speaker training is supported)

Releases

No releases published

Packages

No packages published

Languages

  • Python 99.2%
  • Cython 0.8%