Fix GPT-2 Flash Attention 2 generation with left-padding #41966
+10
−0
Conversation
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
Abdennacer-Badaoui
force-pushed
the
fix/test_flash_attn_2_generate_padding_left
branch
from
2877ab9 to
4f20a1f
Compare
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
|
[For maintainers] Suggested jobs to run (before merge) run-slow: decision_transformer, gpt2 |
|
@vasqu , ready for review :) |
remi-or
requested review from
Cyrilvallez and
vasqu
and removed request for
Cyrilvallez
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.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.
What does this PR do?
Fixes Flash Attention 2 generation with left-padding in GPT-2 by ensuring
is_causal=Trueis set even when padding masks are provided.Flash Attention 2 handles causal masking and padding independently, but the original code incorrectly disabled causal masking when any attention mask was present.
Fixes:
test_flash_attn_2_generate_padding_left