huggingface/transformers

Pull Requests Commits

Merge branch 'main' into cb-fix-k-race-bug

remi-or committed 100 days ago

Verified d863f6e3

Fix `kernelize()` crash for gpt_oss: missing `@use_kernel_func_from_hub` on `apply_rotary_pos_emb` (#45799)

jiqing-feng committed 100 days ago

Verified ebac5e52

Cache `merged_typed_dict` to not break `validate_typed_dict` caching (#45807)

lgeiger committed 100 days ago

Verified 95fdeb31

Get rid of deprecated use_return_dict call. (#45815)

rasmi committed 100 days ago

Verified 28ab024e

fix(qianfan_ocr): add XPU expectations (#45615)

kaixuanliu committed 100 days ago

Verified 15cef23c

Fix shared config mutation issue in flash_attn_from_config (#45678)

kaixuanliu committed 100 days ago

Verified 10f53971

Add RF-DETR (#36895)

sbucaille committed 100 days ago

Verified 81fe4f16

[Weight Converter] More fine-grained mappings on classes, scoping for every transforms (including weight converter) (#45661)

yonigozlan committed 100 days ago

Verified f397b9e6

refector: renamed file glob to cache to make it clearer (#45711)

tarekziade committed 101 days ago

Verified da6c53e4

Fix decorator order (#45806)

zucchini-nlp committed 101 days ago

Verified 0d28dffe

Fix WeightConverter regex incorrectly matching shared_experts as experts in DeepSeek V4 (#45782)

silencelamb committed 101 days ago

Verified 4c2daf7f

[`Granite 4.1 Vision`] Fixup integration tests (#45805)

vasqu committed 101 days ago

Verified f0e41a3e

fix: validate special token ids against attribute values (#45791)

CharlieKerfoot committed 101 days ago

Verified ec4d0aa6

Blockwise mask fn as opt arg in all masking functions (#45477)

zucchini-nlp committed 101 days ago

Verified 1825a374

[CB] Refactor any model-related code in a separate class (#45654)

remi-or committed 101 days ago

Verified 7f6419e6

fix: forward use_cache kwarg to attention mixer in nemotron_h (#45792)

CharlieKerfoot committed 102 days ago

Verified 41c3a5ac

Add Granite 4.1 Vision (granite4_vision) (#45597)

artem-spector committed 102 days ago

Verified 2871caff

fix: correct spelling in continuous_api docstring (#45749)

Dhruv908615 committed 102 days ago

Verified aaec1092

Fix link to modular transformers documentation (#45746)

SangbumChoi committed 102 days ago

Verified 7050d0e0

Gemma4: fix failed test cases (#45568)

kaixuanliu committed 102 days ago

Verified df2f2b53

First model (#45788)

SindhuRaghuram97 committed 102 days ago

Verified 2c7d3856

Fix CI: Allow more artifacts to be download in CI (#45785)

ydshieh committed 102 days ago

Verified a6ccf935

Add `concurrency` to `PR CI` workflow file (`pr-ci-caller.yml`) (#45786)

ydshieh committed 102 days ago

Verified 2c432d73

Reorder decorators for autodoc and dataclass (#45702)

zucchini-nlp committed 102 days ago

Verified 3db570f7

Unwrap `text_config` in `AutoModelFor*.from_config` (#45770)

jamesbraza committed 102 days ago

Verified 136befeb

deepseek r1 distilled tokenizer fix for qwen2 mapping (#45741)

itazap committed 102 days ago

Verified ffd36ed2

fix: Added Mps support in float fallback backends list (#45687)

rigen1048 committed 102 days ago

Verified d379ac18

Github Actions PR CI (caller) (#45476)

ydshieh committed 103 days ago

Verified d63bb4ac

Add EXAONE 4.5 implementations (#45471)

nuxlear committed 103 days ago

Verified a5b83a72

make sure we call check_auto in CI (#45775)

tarekziade committed 103 days ago

Verified 8c004ec6

Older