huggingface/transformers

Pull Requests Commits

pcuenca committed 82 days ago

5d86e324

Fix issue loading ASR pipeline

pcuenca committed 82 days ago

757cb275

Fix tests trainer again (#42933)

SunMarc committed 83 days ago

Verified 70179949

[`Ernie 4.5`] Ernie VL models (#39585)

vasqu committed 83 days ago

Verified a8a22624

[`Generation`] Fix default overwrite for non-`None` defaults (#42958)

vasqu committed 83 days ago

Verified f218ed21

Fix infinity in JSON serialized files (#42959)

LysandreJik committed 83 days ago

Verified d14d99ea

Add runner specification to CodeQL workflow (#42955)

paulinebm committed 83 days ago

Verified d54d78f7

[loading] Really initialize on meta device for huge perf gains (#42941)

Cyrilvallez committed 83 days ago

Verified bb9357ff

fix error: 'BlockMask' object has no attribute 'dtype' for lasr model (#42802)

kaixuanliu committed 83 days ago

Verified 789226c1

fix concat order (#42946)

eustlb committed 83 days ago

Verified a4d62291

[kernels] Fix failling tests (#42953)

MekkCyber committed 83 days ago

Verified cf0f071e

[Quantization] CI green by end of year (#42951)

MekkCyber committed 83 days ago

Verified b5eea347

Updated `backbone_config` docstrings and type annotations (#42927)

simonreise committed 84 days ago

Verified 3e4baf8e

[docs] dtype (#42883)

stevhliu committed 84 days ago

Verified b93f2e3a

Load generation config from nested configs (#42922)

zucchini-nlp committed 84 days ago

Verified f2c6d2ad

🚨 Fix ConvNeXt image processor default interpolation to BICUBIC (#42934)

lukepayyapilli committed 84 days ago

Verified 60634caa

fix(tvp): add missing type_vocab_size parameter to TvpConfig (#42928)

majiayu000 committed 84 days ago

Verified b62e5b3e

[Quantization] Removing misleading int8 quantization in Finegrained FP8 (#42945)

MekkCyber committed 84 days ago

Verified 4dc60fd3

Fp8 dq (#42926)

SunMarc committed 84 days ago

Verified af91c0ba

Add buffers to `_init_weights` for ALL models (#42309)

Cyrilvallez committed 84 days ago

Verified 537c2e3d

rewrite _process_parameter_type in auto_docstring.py to improve usability (#42431)

Yacklin committed 84 days ago

Verified 728f34c3

Update `param_element_size` (#42818)

SunMarc committed 84 days ago

Verified dd8057af

[Tests] Fix CompressedTensors tests (#42935)

kylesayrs committed 84 days ago

Verified 0a846542

[CB] Allow block sharing in hybrid models (#42877)

remi-or committed 84 days ago

Verified 04e78e67

[Quantization] Misc tests fixes (#42940)

MekkCyber committed 84 days ago

Verified d7dd443a

[Quantization] rm _pre_quantization_dtype from quantization tests (#42939)

MekkCyber committed 84 days ago

Verified d3d4b629

🚨 Generation config defaults are now `None` (#42702)

zucchini-nlp committed 84 days ago

Verified a81e04a9

Document new default shard size + dropped unsafe serialization (#42904)

Wauplin committed 84 days ago

Verified bdaddb6f

Improve BatchFeature (.to() works on lists/nested lists of tensors, auto-skip non array conversion) (#42884)

yonigozlan committed 85 days ago

Verified 9f583b1b

[docs] optimization cleanup (#42827)

stevhliu committed 85 days ago

Verified 5ef16edd

Older