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

Training ASR with this model for another language #1

Closed
EmreOzkose opened this issue Oct 15, 2021 · 2 comments
Closed

Training ASR with this model for another language #1

EmreOzkose opened this issue Oct 15, 2021 · 2 comments

Comments

@EmreOzkose
Copy link

EmreOzkose commented Oct 15, 2021

Hi, thank you for sharing a pre-trained model.

Have you ever tried to use this pre-trained model for training of another language ASR? Do you think that it will be useful for training ASR with another language?

Of course, "lang_bpe" will be also changed.

Note: I will also so experiments with this setup and report results here.

@csukuangfj
Copy link
Owner

No, I have not used this pre-trained model for other languages. Not sure whether it is helpful or not.

There are recipes for other languages, instead of English, in icefall that uses similar model architectures.
For instance, there is an ongoing pull request for the AIShell dataset, which is Chinese. Please see
k2-fsa/icefall#30

@EmreOzkose
Copy link
Author

I have 100h Turkish data for training and 1h data for test. I did some experiments with pre-trained model and got 0,62 relative improvement with pre-trained model. Certainly starting with pre-trained model works !

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