huggingface/transformers

Pull Requests Commits

fix(rope): read original_max_position_embeddings from yarn validator's argument (#45887)

bzantium committed 18 days ago

Verified dfcb1a36

Fix deepseek v4 (#45892)

ArthurZucker committed 18 days ago

Verified a1b77cca

Fix/pe audio video bugs (#45886)

massimilianoviola committed 18 days ago

Verified a4c91a1d

Forward `revision` to `list_repo_files` in tokenizer loading (#45908)

nealwu committed 18 days ago

Verified 84db914a

Remove `_checkpoint_conversion_mapping` from model attributes (#45906)

Cyrilvallez committed 18 days ago

Verified fd2c3f95

Use `_keys_to_ignore_on_load_unexpected/missing` recursively from children (#45342)

Cyrilvallez committed 18 days ago

Verified d4b91100

Do not keep refs to submodules internally (#45889)

Cyrilvallez committed 18 days ago

Verified 02262036

[docs] paper cuts (#45798)

stevhliu committed 18 days ago

Verified 0aad7e1f

[Weight converter] Revert unnecessary changes to `rename_source_key` (#45828)

yonigozlan committed 19 days ago

Verified f15fc1e8

fix(minicpmv4_6): skip invalid failing tests (#45836)

kaixuanliu committed 19 days ago

Verified b75feb2a

audio tester class (#45391)

tarekziade committed 19 days ago

Verified 83f33cda

Remove deprecation cycle for inputs embeds (#45885)

Cyrilvallez committed 19 days ago

Verified 2d6815e4

Warn about forgetting attention mask functions (#45811)

TuringTux committed 19 days ago

Verified b890df15

Remove deprecation cycle for `cache_position` in masking primitives (#45884)

Cyrilvallez committed 19 days ago

Verified 6c66de3f

Revert "[deepseek_v4] fix CSA per-query masking in eager path"

ArthurZucker committed 19 days ago

436ba195

Revert "Merge branch 'main' of github.com:huggingface/transformers into fix-deepseek-v4-csa-per-query-mask"

ArthurZucker committed 19 days ago

d996062a

Merge branch 'main' of github.com:huggingface/transformers into fix-deepseek-v4-csa-per-query-mask

ArthurZucker committed 19 days ago

e9fd66d1

[deepseek_v4] fix CSA per-query masking in eager path

ArthurZucker committed 19 days ago

ac372e10

fix(testing_utils): guard get_device_capability with torch.cuda.is_available() (#45351)

RudrenduPaul committed 19 days ago

Verified 1ee03164

🚨 Refactor ViT to updated standards (#41693)

yonigozlan committed 22 days ago

Verified c9de1097

fix(rf_detr): fix failing tests (#45845)

kaixuanliu committed 22 days ago

Verified 381032b7

fix(granite4_vision): auto-fix failing tests (#45844)

kaixuanliu committed 22 days ago

Verified 4f076084

fix(laguna): fix failing tests (#45842)

kaixuanliu committed 22 days ago

Verified 9f0b8aee

:rotating_light: Generic Sequence Classifier works for multimodal models (#44664)

zucchini-nlp committed 22 days ago

Verified 1d5296f2

Add Qwen3.5 support for token classification (#45833)

Ratwood98 committed 22 days ago

Verified 0314464f

Fix import error in moe.py by providing explicit schema to custom_op (#45804)

rraghavkaushik committed 22 days ago

Verified a5cc128d

fix: correct typo 'seperate' -> 'separate' in comments across multipl… (#45824)

Caromisc committed 22 days ago

Verified 4b23cfc9

🚨 [Fuyu] Remove FuyuBatchFeature subclass, use BatchFeature with skip_tensor_conversion (#45808)

Abineshabee committed 22 days ago

Verified 44a2e7a4

Keep deleting (#45802)

zucchini-nlp committed 22 days ago

Verified 9ea2afbc

Fix gemma4 with multi-gpu setup (#45826)

zucchini-nlp committed 22 days ago

Verified 95933eb6

Newer Older