transformers
Fix convert_and_export_with_cache failures for GPU models
#38976

Merged

Fix convert_and_export_with_cache failures for GPU models #38976

gante merged 11 commits into huggingface:main from Stonepia:tongsu/export_fix

Add the `device` option for `generate()`

41def460

Add device for default tensors to avoid tensor mismatch

e18187c1

[test] Enable test_static_cache_exportability for torch_device

c3e9914c

Stonepia marked this pull request as draft 1 year ago

liangan1 commented on 2025-06-24

Stonepia commented on 2025-06-24

infer device from the prompt_token_ids

51ec0301

Add device for generated tensor

6458ff35

[Test] Make `test_export_static_cache` tests to run on devices rather…

37a3e8a9

Merge branch 'huggingface:main' into tongsu/export_fix

dad68b61

fix format

9b474755

Stonepia marked this pull request as ready for review 1 year ago

Merge branch 'main' into tongsu/export_fix

8a6985c5

gante approved these changes on 2025-07-08

infer device from the model

8fd0d52f

Merge branch 'main' into tongsu/export_fix

129a4944

Stonepia requested a review from

gante 1 year ago

gante approved these changes on 2025-07-17

gante enabled auto-merge (squash) 1 year ago

gante merged fc700c2a into main 1 year ago

Reviewers

gante

yao-matrix

liangan1

Assignees

No one assigned

Labels

None yet

Milestone

No milestone