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

make sure the leading and tailing word boundary exists. #1019

Open
wants to merge 4 commits into
base: main
Choose a base branch
from

Conversation

bfs18
Copy link
Contributor

@bfs18 bfs18 commented Jan 26, 2025

Fix #1016
On the wav2vec alignment tutorial page, word boundary symbols are present at the beginning and end. During the whisperX alignment process, the tailing word boundary might be missing, causing the tailing silence to be absorbed into the last word. Adding the tailing word boundary would resolve this issue.

@bfs18 bfs18 mentioned this pull request Jan 26, 2025
@lesca
Copy link

lesca commented Jan 31, 2025

@bfs18 I cloned the latest main branch (44e8bf5) and patched with this PR. From my recent transcriptions on serveral videos, this PR works well and I no longer see sublitles overlap. Thanks for your great works!

@bfs18
Copy link
Contributor Author

bfs18 commented Jan 31, 2025

Hi @lesca I am glad to hear that. Thank you for your feedback.

@bfs18
Copy link
Contributor Author

bfs18 commented Feb 7, 2025

Hi @Barabazs @m-bain
I believe this fix works. It produces correct results for the test case mentioned in issue #1016 (though the original poster has not yet confirmed). Additionally, it should resolve the related issue #999.

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

Successfully merging this pull request may close these issues.

Alignment issues after #986 (support timestamp for numbers)
3 participants