huggingface/transformers

Pull Requests Commits

SunMarc committed 78 days ago

3582fccc

Keep FP8 weights dense when quantizing on the fly

SunMarc committed 78 days ago

4478eb2d

Fix AttributeError in FP8 update_tp_plan when the impl has no TP overrides

SunMarc committed 78 days ago

b925a262

Create the continuous batching CPU group with local synchronization (#48302)

qgallouedec committed 78 days ago

Verified dabae5fc

Resolve continuous batching config against the text config for composite models (#48299)

qgallouedec committed 78 days ago

Verified 598d8ba8

[Glm 5.3 Flash] GLM 5.3 Flash Support (#48342)

Dovis01 committed 78 days ago

Verified eb4d9e2a

Fix kernel commit and repo paths for ESMFold2 (#48186)

Rocketknight1 committed 78 days ago

Verified ba5cb08d

[`CI`] Unblock fast CI for now (failing tests) (#48344)

vasqu committed 78 days ago

Verified 1e452995

[qwen4_exp] disable torch/onnx export tests due to data-dependent control flow (#48345)

ydshieh committed 78 days ago

Verified 0f0f7e1a

[debug] Trace previous CI run selection in get_previous_daily_ci.py (#48338)

ydshieh committed 78 days ago

Verified 36bc98ef

Normalize HunYuanVL's legacy field aliases via attribute_map (#48261)

hmellor committed 78 days ago

Verified 2c889bf2

Add Qwen4Exp model (#48337)

Cyrilvallez committed 78 days ago

Verified fc5c5bde

Restore BC for the tensor-parallel API (#48300)

ArthurZucker committed 78 days ago

Verified f0e42efb

Fix video-llama modular conversion (#48336)

zucchini-nlp committed 78 days ago

Verified 7d337843

CI: gate the hunyuan-moe slow test (#48330)

tarekziade committed 78 days ago

Verified 949e301b

Add a regression test for force_accelerate_hooks signature preservation (#48260)

wtdcode committed 78 days ago

Verified 5fcf605f

Add an opt-in per-frame pixel cap (cap_pixels_per_frame) to the Qwen3-VL video processor (#48071)

dkrisman committed 78 days ago

Verified 27e7f6c8

Fix `scores` type in stopping criteria docstrings (#47676)

qgallouedec committed 78 days ago

Verified 53e3152d

Docstring check didn't match some file - fix it (#48121)

zucchini-nlp committed 78 days ago

Verified b5416997

Add shared ImageProcessingTester (#47745)

guarin committed 78 days ago

Verified b064ba69

fix bug for blt model parallel bug (#48327)

kaixuanliu committed 78 days ago

Verified 2514a5e9

Cpmant fix use cache (#48013)

jiqing-feng committed 78 days ago

Verified 0f93813a

Fix AutoTokenizer returning TokenizersBackend for DeepSeek-R1-Distill-Qwen models (#48211)

ydshieh committed 78 days ago

Verified 51edaf85

Stop pairing two-word pretokenized documents in the python tokenizer (#48279)

VaggelisGian committed 79 days ago

Verified 6ec0f835

[docs] Fix failing doctests (#47687)

stevhliu committed 79 days ago

Verified 2c1376cf

Fix build_2d_sinusoidal_position_embedding on MPS (#47897)

guarin committed 79 days ago

Verified 562cfd94

Refactor audio buffer creation to use NumPy (#48295)

gsaon committed 79 days ago

Verified efdc97d8

[docs] Kernel supported models (#48258)

stevhliu committed 79 days ago

Verified ea1410d7

[Gemma4] Investigate flaky test_generation_beyond_sliding_window_1_eager (#48236)

ydshieh committed 79 days ago

Verified 50adfa19

Let gradient checkpointing skip layers with every_n_layers (#48200)

qgallouedec committed 79 days ago

Verified 2b4e2b3e

Older