Commits huggingface/transformers

minimal fix

Cyrilvallez committed 25 days ago

Verified 5083afd4

[LLaVA] Fix pixtral integration tests for cuda sm_86 (#48166)

ydshieh committed 25 days ago

Verified 945dac91

Port ESMC and ESMFold2 to Transformers (#46419)

Rocketknight1 committed 25 days ago

Verified c8df3fc9

[Qwen2.5-Omni] Update stale expected values for cuda sm_86 (#48164)

ydshieh committed 25 days ago

Verified cdea840f

[Mistral3] Fix batched integration tests: padding_side=left + update expected values (#48161)

ydshieh committed 25 days ago

Verified dd6c64b7

Fix DeepSeek V2 default vocab size (#48159)

hmellor committed 26 days ago

Verified e4dd56a3

[InternVL] Fix stale expected values for Llama integration tests (cuda sm_80) (#48153)

ydshieh committed 26 days ago

Verified 0c928118

Retry transient network errors (RemoteDisconnected) in github_utils (#48124)

ydshieh committed 26 days ago

Verified 95119511

Revert "[Quantization]: Refactor is_quantization_compressed for format-based detection" (#48072)

subin9 committed 26 days ago

Verified 238a2dbd

[Whisper] Fix speculative decoding: preserve cleared suppress tokens through super().generate() (#48108)

ydshieh committed 26 days ago

Verified b6688043

🚨[wav2vec2] Support attn_implementation=sdpa dispatch (#46196)

YangKai0616 committed 26 days ago

Verified 2b296305

fix bugs for clvp model (#47127)

kaixuanliu committed 26 days ago

Verified 9ad4b858

Always tie embeddings for LongT5 and Pop2Piano (#47620)

jiqing-feng committed 26 days ago

Verified 5909a467

Use generator with seed for LengthGroupedSampler in Trainer._get_eval_sampler for deterministic eval order with per_device_eval_batch_size > 1 (#48025)

philipshurpik committed 27 days ago

Verified 94f09cfe

Enable mlinter findings artifact for inline PR reviews (#48117)

ydshieh committed 27 days ago

Verified decba1d2

Fix ASR pipeline mono conversion for channels-last audio (fixes #47886) (#47888)

Kayvan-Zahiri committed 27 days ago

Verified bb8f2353

Fix CpmAnt loading: size lm_head to vocab_size (#48012)

jiqing-feng committed 27 days ago

Verified e872ca25

Delete old mlinter review comments before posting new ones (#48107)

ydshieh committed 27 days ago

Verified de795d93

Fix Qwen2.5-Omni / Qwen3-Omni-MoE generation with a compilable cache (#47872)

jiqing-feng committed 27 days ago

Verified 279dbcf4

[Video] Warn and return all frames when num_frames exceeds total_num_frames (#48074)

carlszk committed 27 days ago

Verified b67f702d

Accept artifact dir as argument in post_mlinter_review.py (#48106)

ydshieh committed 27 days ago

Verified 604da442

Improve group images by shape (#47964)

guarin committed 27 days ago

Verified 0e2ce39f

Fix mlinter artifact path (#48088)

ydshieh committed 27 days ago

Verified f2ef0b94

[Video] Fix convert_to_rgb channel slicing and alpha blending for RGBA videos (#48053)

Ultron09 committed 27 days ago

Verified e7e8b7fa

[Whisper] Fix decoder position IDs for left-padded batches in longform generation (#48028)

ydshieh committed 27 days ago

Verified e12c79c5

[PE] Skip test_sdpa_can_dispatch_on_flash for TimmWrapper-backed models (#48064)

ydshieh committed 28 days ago

Verified f9b76f2d

feat: add nvfp4 quantization (#47883)

drbh committed 28 days ago

Verified 80c66744

fix(pipeline): preserve model.generation_config precedence over pipeline defaults (#47752) (#47953)

nithin42 committed 28 days ago

Verified bc777261

[Gemma3] Update integration test expected values for A10G (#48036)

ydshieh committed 28 days ago

Verified ff09c381

Fallback from 'lanczos' to 'bicubic' when on cuda (#48026)

zucchini-nlp committed 28 days ago

Verified e8018693