Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

ONNX Version of Models #22

Closed
amblamps opened this issue Oct 13, 2024 · 4 comments
Closed

ONNX Version of Models #22

amblamps opened this issue Oct 13, 2024 · 4 comments

Comments

@amblamps
Copy link

Hi, thank you for sharing these models, the quality of the output is amazing! Is there any chance of an ONNX release?

@SWivid
Copy link
Owner

SWivid commented Oct 13, 2024

Not sure if onnx inference would be slow, and it is very possible that we will publish updated ckpt (e.g. more language support)

If anyone is familiar with onnx and interested, welcome to pr

@amblamps
Copy link
Author

amblamps commented Oct 13, 2024

ONNX does not yet support ISTFT needed for Vocos. There are a couple of workarounds.

gemelo-ai/vocos#38

onnx/onnx#4777

@mush42
Copy link

mush42 commented Oct 16, 2024

Hi

For ONNX support we can use WaveNeXt which is based on Vocos but it does not use the istft component.

Unless the tokenizer is separated, exporting to ONNX will not be that useful. The whole point is to have a minimal inference-only package that does not require pytorch.

@SWivid
Copy link
Owner

SWivid commented Oct 22, 2024

#214 @DakeQQ got a onnx version.
Will move to there with onnx discussion, closing this issue~

@SWivid SWivid closed this as completed Oct 22, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants