Skip to content

Add bidirectional attention mask for EmbeddingGemma #2382

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.

Sign up for GitHub

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

Jump to bottom
Merged
merged 1 commit into from
Sep 10, 2025
Merged

Add bidirectional attention mask for EmbeddingGemma #2382

merged 1 commit into from
Sep 10, 2025

Conversation

@abheesht17
Copy link
Collaborator

@abheesht17 abheesht17 commented Sep 5, 2025 •
edited
Loading

The attention mask looks like this (for sequence length = 10, sliding window length = 4), and last 4 tokens padded:

[[[ True,  True,  True, False, False, False, False, False, False, False],
  [ True,  True,  True,  True, False, False, False, False, False, False],
  [ False,  True,  True,  True,  True, False, False, False, False, False],
  [False,  False,  True,  True,  True,  True, False, False, False, False],
  [False, False,  False,  True,  True,  True,  False, False, False, False],
  [False, False, False,  False,  True,  True,  False,  False, False, False],
  [False, False, False, False,  False,  False,  False,  False,  False, False],
  [False, False, False, False, False,  False,  False,  False,  False,  False],
  [False, False, False, False, False, False,  False,  False,  False,  False],
  [False, False, False, False, False, False, False,  False,  False,  False]]]

@github-actions github-actions bot added the Gemma Gemma model specific issues label Sep 5, 2025
@abheesht17 abheesht17 changed the title Add bidirectional masking for EmbeddingGemma Add bidirectional attention mask for EmbeddingGemma Sep 5, 2025
@abheesht17 abheesht17 marked this pull request as ready for review September 5, 2025 06:03
mattdangerw
mattdangerw approved these changes Sep 10, 2025
View reviewed changes
Copy link
Member

@mattdangerw mattdangerw left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks!

@abheesht17 abheesht17 merged commit 5b13fbf into keras-team:master Sep 10, 2025
9 of 10 checks passed
amitsrivastava78 pushed a commit to amitsrivastava78/keras-hub that referenced this pull request Oct 8, 2025
@abheesht17 @amitsrivastava78
Add bidirectional masking for EmbeddingGemma (keras-team#2382)
2576589
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@mattdangerw mattdangerw mattdangerw approved these changes

@divyashreepathihalli divyashreepathihalli Awaiting requested review from divyashreepathihalli

Assignees

No one assigned

Labels

Gemma Gemma model specific issues

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@abheesht17 @mattdangerw