huggingface/transformers

Pull Requests Commits

Merge branch 'main' of github.com:huggingface/transformers into fix-tok-pega

ArthurZucker committed 4 days ago

fb77abab

[Performance] FP8 Grouped and Batched Matmuls (#44231)

IlyasMoutawwakil committed 4 days ago

Verified ff2ba441

Make paligemma embed tokens standard (#44432)

zucchini-nlp committed 4 days ago

Verified 0e7cb4e3

chore(typing): Add type checking to `src/transformers/quantizers` (#44412)

tarekziade committed 4 days ago

Verified 8464284d

Fix: AQLM quantizer to match updated replace_with_aqlm_linear signature (#44577)

tarekziade committed 4 days ago

Verified 42ff4f32

[device_map] Fix device_map computation by correctly adjusting memory available (#44565)

Cyrilvallez committed 4 days ago

Verified 4b442cd6

Fix PR comment CI for quantization job (#44579)

ydshieh committed 4 days ago

Verified 519c26f9

Fix error message label and docstring default in load_sharded_checkpoint (#44523)

jnMetaCode committed 4 days ago

Verified 1bd97f24

Correct Tapas initialization (#44575)

Rocketknight1 committed 5 days ago

Verified 3b6c11d4

fix pegasus conversion

ArthurZucker committed 5 days ago

5cd21427

[`fix`] Prevent crash with Apertus without xielu installed (#44567)

tomaarsen committed 5 days ago

Verified 1cbb9c2e

Speed create_import_structure up with os.scandir() (#44253)

Rocketknight1 committed 5 days ago

Verified 652f2f7a

Fix failing `MusicgenStereo` integration tests (#44527)

Sai-Suraj-27 committed 5 days ago

Verified c7648c8f

Fix zamba2 rotary embedding call when use_mem_rope is False (#44551)

echarlaix committed 5 days ago

Verified b7e4da04

[Bugfix] fix video inference of qwen3vl and qwen3.5 series (#44474)

JJJYmmm committed 5 days ago

Verified 5a098a1e

add XPU Expectations for `higgs_audio_v2` tests (#44482)

kaixuanliu committed 5 days ago

Verified 0e3be387

chameleon added to MODELS_WITH_INCORRECT_HUB_TOKENIZER_CLASS (#44475)

itazap committed 5 days ago

Verified 52c93cfb

[CB] Add paged_attention kernel (#44379)

remi-or committed 5 days ago

Verified a08aa52d

Revert "test merge queue 1" (#44552)

ydshieh committed 5 days ago

Verified 2da98fda

test merge queue 1 (#44529)

ydshieh2 committed 5 days ago

Verified f428423e

Fix: Remove references to `text2text-generation`, `summarization` and `translation` pipeline tasks (#44510)

math-hiyoko committed 5 days ago

Verified 138c1c81

fix(testing): Fix MoonshineEncoder UnboundLocalError and Florence2VisionBackbone dtype mismatch (#44503)

harshaljanjani committed 5 days ago

Verified 1a50a3b1

fix FSDP loading with meta devices (#44473)

winglian committed 5 days ago

Verified 0a0ac7a2

Fix: Remove references to transformers run command (#44513)

math-hiyoko committed 5 days ago

Verified a049c004

[LW-DETR] Fix training (#44441)

NielsRogge committed 5 days ago

Verified a50b099a

fix(tokenizer): Only strip Fast from class names in AutoTokenizer if used as a suffix (#44443)

harshaljanjani committed 6 days ago

Verified f8786de5

[tie weights] 🚨 If both weights are present with same weights, still tie them (#44497)

Cyrilvallez committed 6 days ago

Verified 09677910

Fix AMD Docker image build timeout by pinning Flash Attention commit (#44546)

Abdennacer-Badaoui committed 6 days ago

Verified ce7efd8f

Make `_prepare_input_fn` and `_prepare_output_fn` instance methods (#44499)

michaelbenayoun committed 6 days ago

Verified 86eafab5

Fix: Conditionally import `torch.distributed.fsdp` in `trainer_seq2seq.py` (#44507)

0xDELUXA committed 6 days ago

Verified adc0d9ae

Older