huggingface/transformers

Pull Requests Commits

ydshieh committed 1 year ago

22b7f7e0

ydshieh committed 1 year ago

9a781a49

ydshieh committed 1 year ago

def8ea56

ydshieh committed 1 year ago

fc3f99f8

ydshieh committed 1 year ago

42f0074d

Fix Bad Outputs in Fast Path for GraniteMoeHybrid (#39033)

alex-jw-brooks committed 1 year ago

Verified 5995cfa0

Granite speech speedup + model saving bugfix (#39028)

avihu111 committed 1 year ago

Verified 22b0a898

[tests] remove TF tests (uses of `require_tf`) (#38944)

gante committed 1 year ago

Verified 1d45d90e

Two ReDOS fixes (#39013)

Rocketknight1 committed 1 year ago

Verified d37f7517

[Kyutai-STT] correct model type + model id (#39035)

eustlb committed 1 year ago

Verified 551e48f1

Add SmolLM3 (#38755)

anton-l committed 1 year ago

Verified dad0e87c

refactor: remove custom BarkLayerNorm (#39003)

eginhard committed 1 year ago

Verified 3233e9b7

Fix grammatical error in models documentation (#39019)

marcndo committed 1 year ago

Verified 3c1d4dfb

Remove script datasets in tests (#38940)

lhoestq committed 1 year ago

Verified 858f9b71

fix gemma3 grad acc (#37208)

SunMarc committed 1 year ago

Verified 3c322c9c

fix: astronomical loss with ModernBERT when using gradient checkpointing (#38982) (#38983)

umarbutler committed 1 year ago

Verified 860b898d

Support for Flash Attention 3 (#38972)

EduardDurech committed 1 year ago

Verified a2eb75c8

Fix the seamless_m4t cannot work on Gaudi (#38363)

yuanwu2017 committed 1 year ago

Verified de98fb25

[Model] add dots1 (#38143)

redmoe-moutain committed 1 year ago

Verified 7503cb91

Encoder-Decoder Gemma (#38332)

bzhangGo committed 1 year ago

Verified 3ef88969

GLM-4.1V Model support (#38431)

zRzRzRzRzRzRzR committed 1 year ago

Verified af987026

Drop unnecessary tokens in GPT2Model generation (#39016)

null-pointer-access committed 1 year ago

Verified 7b380738

[video processor] support torchcodec and decrease cuda memory usage (#38880)

zucchini-nlp committed 1 year ago

Verified e212ff9e

[AutoModelForMaskGeneration] Remove duplicate code (#38622)

NielsRogge committed 1 year ago

Verified 11d0feac

Fix graph break in torch.compile when using FA2 with attention_mask=None and batch size > 1 (#37332)

efsotr committed 1 year ago

Verified 3ee72af6

Add zero dim tensor check when using flash_attention (#38280)

ranzhejiang committed 1 year ago

Verified ae32f1ad

[LightGlue] Fixed attribute usage from descriptor_dim to keypoint_detector_descriptor_dim (#39021)

sbucaille committed 1 year ago

Verified ca402e21

Add Hugging Face authentication procedure for IDEs (PyCharm, VS Code,… (#38954)

marcndo committed 1 year ago

Verified 48b6ef02

[HPU][Critical Issue Fix] ThreadPool instead of Pool for parallel pre-processing (#39002)

dsmertin committed 1 year ago

Verified ea9a3092

Skip sdpa dispatch on flash test due to unsupported head dims (#39010)

ivarflakstad committed 1 year ago

Verified 995666ed

Older