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

run problem #2

Open
Polarnight77 opened this issue Jun 7, 2024 · 4 comments
Open

run problem #2

Polarnight77 opened this issue Jun 7, 2024 · 4 comments

Comments

@Polarnight77
Copy link

Hello,I ran code with follow your instructions but had this problem in training "_pickle.UnpicklingError: invalid load key, '\xef'.
". May I ask if this has happened to you as well?

@nafiuny
Copy link
Owner

nafiuny commented Jun 7, 2024

Hello,I ran code with follow your instructions but had this problem in training "_pickle.UnpicklingError: invalid load key, '\xef'. ". May I ask if this has happened to you as well?

Hello, I had't had this problem. How did you run it? Please give me more information to help you. Which part did you get error?

@Polarnight77
Copy link
Author

Hello, I fixed it.This is when loading the melgan-neurips the parameters need to be modified.I would like to ask you if you have encountered that after the training is completed, during the test convert phase. For example, there are 50 files ready for the test conversion, but after the conversion there are only 49, and always one file will be missing.Thanks

@nafiuny
Copy link
Owner

nafiuny commented Jul 10, 2024

Hello, I fixed it.This is when loading the melgan-neurips the parameters need to be modified.I would like to ask you if you have encountered that after the training is completed, during the test convert phase. For example, there are 50 files ready for the test conversion, but after the conversion there are only 49, and always one file will be missing.Thanks

The number of converted sounds is equal to the number of samples of the first speaker.
The code was updated. load melgan was fixed. local melgan was modified. use it.

@Polarnight77
Copy link
Author

Thank you very much for your reply. The number of transformed voices is equal to the number of samples from the first speaker, is this meant to be the number of samples in evaluate dataset? When I train VCC2SF1 to transform VCC2TF1, the train samples are all 81 and the EVALUATE samples are all 51, but when I test the VCC2SF1 evaluation samples, the 51 samples change to 50. I do not know if you have encountered this problem?

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