transformers
Deepseek v4 csa mask collapse
#45928
Merged

Deepseek v4 csa mask collapse #45928

ArthurZucker merged 1 commit into main from deepseek-v4-csa-mask-collapse
ArthurZucker
github-actions
ArthurZucker ArthurZucker force pushed from eacc6a1c to 65483d8c 9 days ago
ArthurZucker ArthurZucker force pushed from 65483d8c to 572de4ec 9 days ago
ArthurZucker ArthurZucker marked this pull request as ready for review 9 days ago
ArthurZucker [deepseek_v4] collapse CSA block_bias from [S, S*top_k] to [S, compre…
bd900875
ArthurZucker ArthurZucker force pushed from 572de4ec to bd900875 9 days ago
ArthurZucker ArthurZucker added for patch
ArthurZucker ArthurZucker merged 2ad5a9b8 into main 9 days ago
ArthurZucker ArthurZucker deleted the deepseek-v4-csa-mask-collapse branch 9 days ago
HuggingFaceDocBuilderDev
vadimkantorov
vadimkantorov commented on 2026-05-13

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone