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 hifi-gan #1265

Closed
kyakuno opened this issue Oct 8, 2023 · 5 comments
Closed

ADD hifi-gan #1265

kyakuno opened this issue Oct 8, 2023 · 5 comments
Assignees

Comments

@kyakuno
Copy link
Collaborator

kyakuno commented Oct 8, 2023

https://github.com/jik876/hifi-gan
Fast MelSpectrum2Wave that also runs on CPU.

@kyakuno
Copy link
Collaborator Author

kyakuno commented Oct 8, 2023

Currently, tacotron2 uses waveglow, but the processing is heavy. Therefore, I would like to be able to use hifi-gan as an option.

@kyakuno kyakuno changed the title ADD highfi gan ADD hifi-gan Oct 8, 2023
@kyakuno
Copy link
Collaborator Author

kyakuno commented Oct 8, 2023

An explanatory article about hifi gan.
https://qiita.com/NoriakiOshita/items/02deaa9af5dbbca02add

@kyakuno
Copy link
Collaborator Author

kyakuno commented Oct 12, 2023

  1. Register as hifi gan on ailia-models as a single model.
  2. Allow it to be specified as an option to tacotron2.
    https://github.com/axinc-ai/ailia-models/tree/master/audio_processing/tacotron2

@YToleubay
Copy link
Contributor

  1. added HIFIGAN model (standalone) #1288 it is for task 1, Register hifi gan on ailia-models as a single model.
  2. added HIFI GAN model (tacotron2) #1287 it is for task 2, Allow it to be specified as an option to tacotron2.

@kyakuno
Copy link
Collaborator Author

kyakuno commented Oct 24, 2023

We have exported LJ_FT_T2_V1. I checked the paper, it was the model that provided the highest scores.

@kyakuno kyakuno closed this as completed Oct 24, 2023
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

2 participants