huggingface/transformers

Pull Requests Commits

persimmon modular

itazap committed 23 days ago

49b2267c

itazap committed 24 days ago

6d907d6d

adding BC for custom toks accessing slow tok attrs deprecated in v5

ita.zaporozhets@huggingface.co committed 24 days ago

a1e85f48

[fix] DAC model: Apply STE in Dac.from_latents to match the forward pass (#43820)

harshaljanjani committed 24 days ago

Verified 3d24ffaa

Update hub metadata (#43892)

zucchini-nlp committed 24 days ago

Verified 15bada75

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

NielsRogge committed 24 days ago

Verified 60a5dbaa

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

tomaarsen committed 24 days ago

Verified 71d7fc58

Refactor trainer init (#43807)

SunMarc committed 24 days ago

Verified e12aa2d2

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

sywangyi committed 24 days ago

Verified 6d86e9a3

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

sywangyi committed 24 days ago

Verified b33235e3

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

SunMarc committed 24 days ago

Verified 760eae30

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

jp1924 committed 24 days ago

Verified af496d74

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

cyyever committed 24 days ago

Verified 520fad98

Update TrainingArguments (#43806)

SunMarc committed 24 days ago

Verified 44f92b63

Fix old tech stack in doc (#43879)

cyyever committed 24 days ago

Verified 6028dd2b

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

OiPunk committed 24 days ago

Verified 6df88fab

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

vasqu committed 24 days ago

Verified 9b4de431

[MistralCommonBackend] fix loading proc (#43887)

eustlb committed 24 days ago

Verified 3b2abb8f

Fix GlmMoeDsaConfig default mlp_layer_types in modular conversion (#43876)

OiPunk committed 24 days ago

Verified 476600a9

Fix init weights in remote code (#43768)

zucchini-nlp committed 24 days ago

Verified 884749a1

Improve PR comment CI feedback (#43852)

ydshieh committed 24 days ago

Verified b7b9d252

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

thecaptain789 committed 25 days ago

Verified b2028e77

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

Anri-Lombard committed 25 days ago

Verified a47a89a4

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

harshaljanjani committed 25 days ago

Verified 82a5b3d7

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

heathdutton committed 25 days ago

Verified d7ac5cd0

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

Rocketknight1 committed 25 days ago

Verified 9e4a8c44

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

vasqu committed 25 days ago

Verified 4b8ba25a

Fix video interpolation in pe_audio_video (#43811)

Rocketknight1 committed 25 days ago

Verified 7d5681b3

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

Mr-Neutr0n committed 25 days ago

Verified 018af00b

Update ModelType for Unigram tokenizer (#43860)

pavel-esir committed 25 days ago

Verified 1f62d4ed

Older