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

Add support for the speaker encoder training using torch spectrograms #1303

Closed
wants to merge 2 commits into from

Conversation

Edresson
Copy link
Contributor

@Edresson Edresson commented Feb 28, 2022

It enables the training with the H/ASP speaker encoder.

In the future, I think that it is very important we implement unitest for the speaker encoder training. For now, I can't do that because I need some changes that I have done in the dev branch. When we merge this in the dev I can work on that.

@Edresson Edresson requested a review from erogol February 28, 2022 13:00
@erogol
Copy link
Member

erogol commented Mar 1, 2022

Do you want to just adapt the old implementation or implement using the new API like VITS?

@Edresson
Copy link
Contributor Author

Edresson commented Mar 1, 2022

Do you want to just adapt the old implementation or implement using the new API like VITS?

For now, I want to implement the training support in an easy/fast way. For emotion encoder, I'm using the same implementation (The only thing that changes is the formatters for now).

@erogol
Copy link
Member

erogol commented Mar 1, 2022

Ok I'll merge it to the DatasetRefactor branch once I finish my merging of FactorTrainerAPI

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

Successfully merging this pull request may close these issues.

2 participants