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

Implement #61

Closed
kam1107 opened this issue Nov 22, 2018 · 4 comments
Closed

Implement #61

kam1107 opened this issue Nov 22, 2018 · 4 comments

Comments

@kam1107
Copy link

kam1107 commented Nov 22, 2018

Hi, I noticed that nnet3 is implement in pykaldi. Is there any chance to also include nnet3bin (kaldi)? Furthermore, can xvector pipeline be implemented in pykaldi? Thank you very much.

@dogancan
Copy link
Member

We don't have any plans of implementing executables replicating the functionality of Kaldi command-line programs. Do you have a specific need that is not met by the Kaldi executables?

The xvector testing pipeline (not training) can (probably) be implemented in PyKaldi, but I don't know what it would bring to the table beyond the native implementation in Kaldi. Again, do you have a specific need that is not met by the Kaldi xvector pipeline?

@kam1107
Copy link
Author

kam1107 commented Nov 25, 2018

We plan to fine tune the network, or make some attempts on implementing a different loss function (e.g. center loss) to substitute softmax. We thought Python is more friendly for experiments like these.

@dogancan
Copy link
Member

What you are saying makes sense but I don’t quite follow how reimplementing kaldi executables in pykaldi would help you achieve that.

@dogancan
Copy link
Member

dogancan commented Dec 5, 2018

Hey I am closing this issue but feel free to reopen it if you think there is something we can do to help.

@dogancan dogancan closed this as completed Dec 5, 2018
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