feat: add attention mask support for padded token sequences#115
Open
richardk53 wants to merge 1 commit intomainfrom
Open
feat: add attention mask support for padded token sequences#115richardk53 wants to merge 1 commit intomainfrom
richardk53 wants to merge 1 commit intomainfrom
Conversation
richardk53
requested review from
HennerM,
MauritsBleeker and
Ndles
as code owners
|
All contributors have signed the CLA ✍️ ✅ |
|
richardk53
changed the title
Author
|
I have read the CLA Document and I hereby sign the CLA |
Author
|
recheck |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The mixed attention implementation currently does not support variable length sequences because attention masks are not supported. This PR adds support for masking that allows for padding sequences.
The main use-case is to batch multiple anchor token sequences of different lengths to sequences of the same length by padding them.
The mask is for keys only and broadcasted over queries, preventing attention to tokens that are not valid. For anchor attention, where anchors produce keys and values, this means no anchor or query can attend to anchors that are masked out.
I have read the CLA Document and I hereby sign the CLA