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 code for conformer + ctc missed files #28

Open
superhg opened this issue Aug 16, 2022 · 1 comment
Open

training code for conformer + ctc missed files #28

superhg opened this issue Aug 16, 2022 · 1 comment

Comments

@superhg
Copy link

superhg commented Aug 16, 2022

hello,我对比了工程里的conformer实现代码,发现了一些与espnet对不上的,排除了一些版本的问题,发现有部分代码差别很大,如果直接用espnet最新的代码来训练conformer + ctc,有需要修改的地方吗?尤其是那个subsample这个地方

@superhg
Copy link
Author

superhg commented Aug 16, 2022

论文里面描述的mel 4倍的下采样是在bottle net feature extractor里面做的,然而,代码里面在bnf-prenet里面通过两层卷积(stride=2)也做了4倍下采样,这边不是很理解

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

1 participant