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

重构和精简 wenet 的解码部分,未来结构更清晰,代码更紧凑,解码效率更高。 #2066

Closed
4 tasks done
robin1001 opened this issue Oct 20, 2023 · 2 comments

Comments

@robin1001
Copy link
Collaborator

robin1001 commented Oct 20, 2023

@robin1001
Copy link
Collaborator Author

The lines of the core files:

  • wenet/transformer/asr_model.py, 1000+ -> 400+
  • wenet/bin/recognize.py, -> 400+ -> 200+

@robin1001
Copy link
Collaborator Author

Test result on AIShell-1:
image

We only support batch size 1 in the old system in decoding, add add batch support in #2059. Compared with the old system with batch size 1, there is about 4 times speedup.

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