transformers
Avoid unnecessary device operations in loss computing
#36950

Merged

Avoid unnecessary device operations in loss computing #36950

Rocketknight1 merged 2 commits into huggingface:main from cyyever:improve_loss

github-actions marked this pull request as draft 266 days ago

cyyever marked this pull request as ready for review 266 days ago

github-actions requested a review from

ArthurZucker 266 days ago

cyyever force pushed from 99a4f493 to 4f11d5d9 265 days ago

Rocketknight1 approved these changes on 2025-03-26

cyyever force pushed from 4f11d5d9 to 04130152 265 days ago

cyyever requested a review from

Rocketknight1 265 days ago

Rocketknight1 removed review request from

ArthurZucker 264 days ago

cyyever force pushed from 04130152 to 1cb5f17e 264 days ago

cyyever force pushed from 1cb5f17e to 86ab5bd7 264 days ago

cyyever force pushed from 86ab5bd7 to cc11c3df 264 days ago

cyyever force pushed from cc11c3df to 8f9b9bd2 264 days ago

cyyever force pushed from 8f9b9bd2 to 87307413 264 days ago

cyyever force pushed from b92e242e to b856aa69 264 days ago

cyyever force pushed from 8735f3b0 to b856aa69 264 days ago

cyyever force pushed from b856aa69 to 00bd169f 264 days ago

cyyever force pushed from 00bd169f to cb572ae4 264 days ago

cyyever force pushed from cb572ae4 to f5618a27 264 days ago

cyyever changed the title ~~Avoid unnecessary tensor copy in loss computing~~ Avoid unnecessary device operations in loss computing 264 days ago

Avoid unnecessary tensor copy in loss computing

ac8ed1e4

Add type

f7aa29d8

cyyever force pushed from f5618a27 to f7aa29d8 264 days ago

Rocketknight1 approved these changes on 2025-03-27

Rocketknight1 merged 8c5e29ba into main 264 days ago

cyyever deleted the improve_loss branch 264 days ago

Reviewers

Rocketknight1

Assignees

No one assigned

Labels

None yet

Milestone

No milestone