Finetune on new language data (10 hours for 2 speaker)

[cod3r0k]

Hi, could you help me in finetune it into my own dataset?
my dataset follow below template:

wavs/ ...
metadata.csv

[nshmyrev]

Sure, what is the issue you have?

[cod3r0k]

I am seeking a step-by-step approach to do that. I don't see any comments about tokenization tricks, configurations, or detailed guidelines for that.

[nshmyrev]

Vosk-TTS is mostly R&D project, we do not have step-by-step guides for now, sorry.