PR #39788 Fix re-compilations for cross attention cache

Fix re-compilations for cross attention cache #39788

zucchini-nlp merged 1 commit into huggingface:main from zucchini-nlp:cache-cross-attn-compile

fix recompilations for cross attn cache

018bfe8d

zucchini-nlp requested a review from

manueldeprada 297 days ago

zucchini-nlp requested a review from

gante 297 days ago

manueldeprada approved these changes on 2025-07-30

zucchini-nlp merged 8e077a3e into main 297 days ago

Reviewers

manueldeprada

gante

Assignees

No one assigned

Labels

None yet

Milestone

No milestone