Fix attention sink implementation in flex attention #41083
Fix attention sink implementation in flex attention
ea9afc16
fix dim
4c17ba6c
fix
3e64462f
Remove print
6f2e4c9a
SamuelBarryCS
marked this pull request as ready for review 244 days ago
SamuelBarryCS
changed the title [WIP] Fix attention sink in flex attention Fix attention sink implementation in flex attention 244 days ago
raisae error when return_lse is False yet s_aux is providewd
f37738fb
Merge branch 'main' into fix-attention-sink
c94b1e27
Clean test files for merge
cdf33258
Merge branch 'fix-attention-sink' of github.com:SamuelBarryCS/transfo…
9c90c3d3
Update src/transformers/integrations/flex_attention.py
696bdd76
force return lse
2e18a7ee
Merge branch 'fix-attention-sink' of github.com:SamuelBarryCS/transfo…
f59dde4e
Add to doc
58799e23
Merge branch 'main' into fix-attention-sink
a2c147dc
SunMarc
enabled auto-merge (squash) 238 days ago
SunMarc
approved these changes
on 2025-09-29
SunMarc
merged
52cbc7c8
into main 238 days ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub