transformers
[CP] Add attention_mask to the buffer when the mask is causal
#40619
Merged
Merged
[CP] Add attention_mask to the buffer when the mask is causal #40619
SunMarc
merged
3 commits into
main
from
kashif-patch-1
Fix attention mask validation for context parallelism
0dd1b041
SunMarc
approved these changes
on 2025-09-02
Merge branch 'main' into kashif-patch-1
9a8ae937
only split 2d attention masks
484321d7
Login to write a write a comment.
Login via GitHub