transformers
fix to accept cumulative_seqlens from TransformersKwargs in FA
#40194
Merged

fix to accept cumulative_seqlens from TransformersKwargs in FA #40194

Kurt232
Kurt232 Kurt232 changed the title fix to accept cumulative_seqlens from TransformersKwargs in FA #40193 fix to accept cumulative_seqlens from TransformersKwargs in FA 204 days ago
Cyrilvallez
Kurt232
vasqu
Cyrilvallez
Cyrilvallez
vasqu
Kurt232
vasqu
Cyrilvallez
Kurt232 fix to the typings which are unmatched to FA function signature
dc0624db
Kurt232 Kurt232 force pushed from fedbf6d0 to dc0624db 198 days ago
Kurt232
Kurt232 format changes by ruff
a9cd0b22
vasqu
vasqu commented on 2025-08-21
Kurt232 Kurt232 changed the title fix to accept cumulative_seqlens from TransformersKwargs in FA 🚨 fix to accept cumulative_seqlens from TransformersKwargs in FA 198 days ago
Kurt232 Update src/transformers/integrations/flash_paged.py
a1415149
ArthurZucker
ArthurZucker commented on 2025-08-21
ArthurZucker
vasqu
Kurt232 revert continuous_batching signiture, which is more meaningful
d93cdd1e
Kurt232
Kurt232 commented on 2025-08-22
Kurt232 Kurt232 changed the title 🚨 fix to accept cumulative_seqlens from TransformersKwargs in FA fix to accept cumulative_seqlens from TransformersKwargs in FA 197 days ago
ArthurZucker
ArthurZucker approved these changes on 2025-08-25
ArthurZucker ArthurZucker merged 14b89fed into main 194 days ago
Kurt232 Kurt232 deleted the fix/args_in_flash_attention_forward branch 194 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone