huggingface/transformers

Pull Requests Commits

temporary version fix for testing.

sayakpaul committed 23 days ago

d9f18d81

[CB] [Minor] Add data-parallel to overall script (#46490)

remi-or committed 23 days ago

Verified acc2cda7

[CI] fix and make less flaky (#46543)

zucchini-nlp committed 23 days ago

Verified 7d06b1a5

Fix hf_hub_download not placing file in current dir for url_to_local_path (#46545)

ydshieh committed 23 days ago

Verified 96fa1e5b

Only download hub cache for pipelines_torch job (#46544)

ydshieh committed 23 days ago

Verified 7c57402d

[docs] torchvision lanczos (#46528)

stevhliu committed 23 days ago

Verified 53ea1612

Add deepseek 3.2 exp (#41251)

ArthurZucker committed 23 days ago

Verified 513146c1

GPU go brr (#46540)

gante committed 23 days ago

Verified 6de66613

[docs] 4D attention masks (#46507)

stevhliu committed 23 days ago

Verified 33b7e4f8

Fix the offsets in processing (#46525)

zucchini-nlp committed 23 days ago

Verified b0a38eb6

Fix buggy action sha pin (#46534)

ydshieh committed 23 days ago

Verified c1b6e837

Fix trailing comma bug in DataCollatorForLanguageModeling example (#46527)

JemmaUZH committed 23 days ago

Verified 0095551f

Fix missing Gemma4Processor._compute_audio_num_tokens (#46416)

csantosbh committed 23 days ago

Verified 04cb96f3

Fix InternVL models (#46524)

hmellor committed 24 days ago

Verified d557ef55

fix(afmoe): reduce tokens in test_compile_static_cache to avoid flaky bfloat16 drift (#46521)

ydshieh committed 24 days ago

Verified d2094607

[CB] Add a "max_requests_per_batch" parameter (#46434)

remi-or committed 24 days ago

Verified 64f94504

Extended & simplified n-to-1 kernel fusion via KernelConfig (#46339)

michaelbenayoun committed 24 days ago

Verified 5047f086

Speed up modular conversion by reusing process pool (#46438)

guarin committed 25 days ago

Verified cc041637

revamp cv docs and fix rf-detr (#46219)

merveenoyan committed 25 days ago

Verified 1b7bc25c

Update hub metadata (#46379)

zucchini-nlp committed 25 days ago

Verified 1d185d43

extend DeepseekV4FlashIntegrationTest to non-cuda device (#46517)

sywangyi committed 25 days ago

Verified e3144397

[docs] deepgemm (#46361)

stevhliu committed 25 days ago

Verified 55115549

Fix `{image/video/audio}_token_ids` in `ProcessorMixin` (#46500)

hmellor committed 25 days ago

Verified d6a82ba8

[fix] regression introduced by #45534 (#46456)

eustlb committed 25 days ago

Verified b6bad752

[docs] tp for continuous batching (#46019)

stevhliu committed 25 days ago

Verified 153d6c16

Use torchvision's native LANCZOS interpolation instead of PIL fallback (#46496)

NicolasHug committed 25 days ago

Verified ff331709

skip quant cache tests for deepseek v4 (#46362)

kaixuanliu committed 25 days ago

Verified e569f888

revisit history parallel beam search tests to avoid unnecessary fix (#46495)

kaixuanliu committed 25 days ago

Verified 81fd7ae4

Add debugging info in `pr-ci-caller.yml` (#46505)

ydshieh committed 25 days ago

Verified e0779e46

Fix tests: 'Cohere2MoeModel' object has no attribute 'hf_device_map' (#46337)

kaixuanliu committed 25 days ago

Verified bfaf9216

Older