search

coqui-ai / TTS

🐸💬 - a deep learning toolkit for Text-to-Speech, battle-tested in research and production
http://coqui.ai
Mozilla Public License 2.0
34.34k stars 4.16k forks source link

[Feature request] Vall-e implementation #2269

Closed iamkhalidbashir closed 1 year ago

iamkhalidbashir commented 1 year ago

Vall-e by microsoft seems to have out performed the yourTTS model in zero short tts. Link: https://arxiv.org/abs/2301.02111

Solution

An implementation of it would be amazing

tekakutli commented 1 year ago

An independent PyTorch implementation https://github.com/enhuiz/VALL-E

erogol commented 1 year ago

We can pull it from there. I hope they will be able to train the model. In my own implementation, the model is prohibitively large and consuming to train.

stale[bot] commented 1 year ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. You might also look our discussion channels.

stanleyshly commented 1 year ago

@erogol Someone appears to have trained a model, with some success: lifeiteng/vall-e/issues/58. It took a lot of compute(4 days on 8 A100s).

jav-ed commented 1 year ago

There is now also the X version, see: https://vallex-demo.github.io/