huggingface/transformers

Pull Requests Commits

try to relax the tie_weights method

molbap committed 155 days ago

66b665dd

use the enable_gqa param in torch.nn.functional.scaled_dot_product_at… (#39412)

sywangyi committed 155 days ago

Verified 9323d087

Fix missing initializations for models created in 2023 (#39239)

bvantuan committed 155 days ago

Verified 6b3a1f2f

Raise `TypeError` instead of ValueError for invalid types (#38660)

Sai-Suraj-27 committed 155 days ago

Verified 970d9a75

Fix pylint warnings (#39477)

cyyever committed 155 days ago

Verified 822c5e45

Fix Qwen Omni integration test (#39553)

Cyrilvallez committed 155 days ago

Verified dc017cd7

🚨🚨🚨 [Trainer] Enable `average_tokens_across_devices` by default in `TrainingArguments` (#39395)

Krish0909 committed 155 days ago

Verified fdc0566e

Rename `_supports_flash_attn_2` in examples and tests (#39471)

zucchini-nlp committed 155 days ago

Verified 8c102e2e

Fix the check in flex test (#39548)

Cyrilvallez committed 155 days ago

Verified 3a152e3a

Fix bad tensor shape in failing Hubert test. (#39502)

ebezzam committed 155 days ago

Verified 78fb2d27

GLM-4 Update (#39393)

zRzRzRzRzRzRzR committed 155 days ago

Verified 39ba5f3c

[qwen2 vl] fix packing with all attentions (#39447)

zucchini-nlp committed 155 days ago

Verified 344012b3

[gemma3] support sequence classification task (#39465)

zucchini-nlp committed 155 days ago

Verified e42681b4

Fix placeholders replacement logic in auto_docstring (#39433)

yonigozlan committed 158 days ago

Verified 34133d0a

Update SAM/SAM HQ attention implementation + fix Cuda sync issues (#39386)

yonigozlan committed 158 days ago

Verified 433d2a23

Improve @auto_docstring doc and rename `args_doc.py` to `auto_docstring.py` (#39439)

yonigozlan committed 158 days ago

Verified 541bed22

Add fast image processor SAM (#39385)

yonigozlan committed 158 days ago

Verified de0dd313

Fix BatchEncoding.to() for nested elements (#38985)

eginhard committed 158 days ago

Verified 561a79a2

[gemma3] Fix do_convert_rgb in image processors. (#39438)

MohitIntel committed 158 days ago

Verified f4d07656

[chat template] return assistant mask in processors (#38545)

zucchini-nlp committed 158 days ago

Verified bcc00919

[dependencies] Update `datasets` pin (#39500)

gante committed 158 days ago

Verified 328ca9cf

Slack CI bot: set default result for non-existing artifacts (#39499)

ahadnagy committed 158 days ago

Verified fb583777

🚨🚨 Fix and simplify attention implementation dispatch and subconfigs handling (#39423)

Cyrilvallez committed 158 days ago

Verified 4ded9a41

[dependencies] temporary pyarrow pin (#39496)

gante committed 158 days ago

Verified 2b819ba4

Add voxtral (#39429)

eustlb committed 159 days ago

Verified 96704508

Fix typing order (#39467)

Tavish9 committed 159 days ago

Verified 73869f2e

Add unified logits_to_keep support to LLMClass (#39472)

hellopahe committed 159 days ago

Verified bda75b40

[serve] Add speech to text (`/v1/audio/transcriptions`) (#39434)

gante committed 159 days ago

Verified bf6c9976

Update integration_utils.py (#39469)

zhaiji0727 committed 159 days ago

Verified 8b3de61a

fix: ImageTextToTextPipeline handles user-defined generation_config (#39374)

peteryschneider committed 159 days ago

Verified 7fd60047

Older