vlomme/Multi-Tacotron-Voice-Cloning

WaveGlow

cnstntn-kndrtv opened this issue · 3 comments

Is it possible to use WaveGlow?
And not to train your own model but use pretrained?

Yes. You must replace the vocoder

Can you provide some example?
Maybe we can create another branch.
Goal is to increase the quality of synthesised voice without long and expensive training (WaveGlow trained on 4 GPU for about 1, or 2 weeks). May be it would be helpful for you too.

Sorry, but I don’t have the desire and time right now.