huggingface/transformers

Pull Requests Commits

propagate modular changes

Rocketknight1 committed 7 hours ago

2ca52d79

Decorate cache updates with no_grad, just in case

Rocketknight1 committed 7 hours ago

84309d03

Update hub metadata (#43892)

zucchini-nlp committed 9 hours ago

Verified 15bada75

[Docs] Add docs for GLM-OCR and fix EomT-DINOv3 (#43710)

NielsRogge committed 9 hours ago

Verified 60a5dbaa

[`fix`] Use `last_hidden_state` key from `get_image_features` for llama4 (#43882)

tomaarsen committed 9 hours ago

Verified 71d7fc58

Refactor trainer init (#43807)

SunMarc committed 9 hours ago

Verified e12aa2d2

fix fbgemm fp8 multi-device load failure. (#43581)

sywangyi committed 10 hours ago

Verified 6d86e9a3

fix the error of tests/quantization/fbgemm_fp8/test_fbgemm_fp8.py::Fb… (#43547)

sywangyi committed 10 hours ago

Verified b33235e3

[Trainer] Move optimizer cls init to trainer_optimizer.py (#43738)

SunMarc committed 10 hours ago

Verified 760eae30

Make it possible to evaluate when using sequence parallel in HF Trainer (#43517)

jp1924 committed 10 hours ago

Verified af496d74

Remove unnecessary code or checks for PT 2.4+ (#43787)

cyyever committed 11 hours ago

Verified 520fad98

Update TrainingArguments (#43806)

SunMarc committed 11 hours ago

Verified 44f92b63

Fix old tech stack in doc (#43879)

cyyever committed 11 hours ago

Verified 6028dd2b

Fix SwanLab callback to forward resume init args (#43848)

OiPunk committed 12 hours ago

Verified 6df88fab

[`Jamba`] Fallback to slow path and warn instead of error out (#43889)

vasqu committed 12 hours ago

Verified 9b4de431

[MistralCommonBackend] fix loading proc (#43887)

eustlb committed 12 hours ago

Verified 3b2abb8f

Fix GlmMoeDsaConfig default mlp_layer_types in modular conversion (#43876)

OiPunk committed 12 hours ago

Verified 476600a9

Fix init weights in remote code (#43768)

zucchini-nlp committed 13 hours ago

Verified 884749a1

Improve PR comment CI feedback (#43852)

ydshieh committed 15 hours ago

Verified b7b9d252

docs: correct typo 'neccessary' to 'necessary' (#43868)

thecaptain789 committed 1 day ago

Verified b2028e77

[xLSTM] Fix bugs preventing small model training (#43209)

Anri-Lombard committed 1 day ago

Verified a47a89a4

docs: Add Switch Transformers docstring notes and update spectrogram comment (#43336)

harshaljanjani committed 1 day ago

Verified 82a5b3d7

Fix cardinality error for DETR models without explicit background class (#43513)

heathdutton committed 1 day ago

Verified d7ac5cd0

Look for the pad_token_id in the right place for Llama4 (#43539)

Rocketknight1 committed 1 day ago

Verified 9e4a8c44

:rotating_light: [`Attn`] New attn mask interface everywhere (#42848)

vasqu committed 1 day ago

Verified 4b8ba25a

Fix video interpolation in pe_audio_video (#43811)

Rocketknight1 committed 1 day ago

Verified 7d5681b3

[docs] Remove pipeline() examples from summarization/translation tasks (#43831)

Mr-Neutr0n committed 1 day ago

Verified 018af00b

Update ModelType for Unigram tokenizer (#43860)

pavel-esir committed 1 day ago

Verified 1f62d4ed

Add GlmMoeDsa (#43858)

Cyrilvallez committed 1 day ago

Verified 33a8e68d

delete unnecessary code to make moe compatible to full graph compile (#43855)

kaixuanliu committed 1 day ago

Verified 3b4a064d

Older