Commits

Branch selector

 main

User selector

All users

All time

Commit History

Commits on Aug 15, 2025

• Merge branch 'huggingface:main' into main

  danielhanchen

  authored
  35acfa8

  Copy full SHA for 35acfa8

• Add repr to EncoderDecoderCache (#40195)

  Show description for 5068fcd
  Cyrilvallez

  authored
  5068fcd

  Copy full SHA for 5068fcd

• Fix fsdp for generic-task models (#40191)

  Show description for 4211756
  Cyrilvallez

  authored
  4211756

  Copy full SHA for 4211756

• Add swiglu_limit = 7.0

  danielhanchen

  authored
  411ab01

  Copy full SHA for 411ab01

• fix to avoid modifying a view in place (#40162)

  Show description for 4912d5b
  3outeille

  authored
  4912d5b

  Copy full SHA for 4912d5b

• make model doc device agnostic (#40143)

  Show description for cc99978
  

  yao-matrix

  and

  stevhliu

  authored
  cc99978

  Copy full SHA for cc99978

• [MINOR:TYPO] Update base.py (#40169)

  Show description for 85fce2e
  cakiki

  authored
  85fce2e

  Copy full SHA for 85fce2e

Commits on Aug 14, 2025

• Update dynamic attnt setter for multimodals (#39908)

  Show description for 52c6c1b
  

  zucchini-nlp

  and

  Cyrilvallez

  authored
  52c6c1b

  Copy full SHA for 52c6c1b

• Pin torch to 2.7.1 on CircleCI for now (#40174)

  Show description for 31b6e6e
  ydshieh

  authored
  31b6e6e

  Copy full SHA for 31b6e6e

• Add dates to the model docs (#39320)

  Show description for b02f2d8
  MHRDYN7

  authored
  b02f2d8

  Copy full SHA for b02f2d8

• Standardize BARTpho model card: badges, new examples, fixed broken im… (#40051)

  Show description for 8a658ac
  

  eshwanthkartitr

  and

  stevhliu

  authored
  8a658ac

  Copy full SHA for 8a658ac

• Add GptOssForSequenceClassification for GPT-OSS models (#40043)

  Show description for 2b6cbed
  

  zyfedward

  and

  Yuefeng Zhan

  authored
  2b6cbed

  Copy full SHA for 2b6cbed

• build: Add fast image processor tvp (#39529)

  Show description for b834cb8
  

  adutchengineer

  and

  yonigozlan

  authored
  b834cb8

  Copy full SHA for b834cb8

• Fix docs typo (#40167)

  Show description for 6f259bc
  

  qubvel

  and

  cijose

  authored
  6f259bc

  Copy full SHA for 6f259bc

• [bugfix] fix flash-attention2 unavailable error for Ascend NPU (#40151)

  Show description for 41980ce
  FightingZhen

  authored
  41980ce

  Copy full SHA for 41980ce

• [FA2] Fix it finally - revert fa kwargs preparation (#40161)

  Show description for eba1d62
  Cyrilvallez

  authored
  eba1d62

  Copy full SHA for eba1d62

• Replace `self.tokenizer` by `self.processing_class` (#40119)

  qgallouedec

  authored
  1c5d2f7

  Copy full SHA for 1c5d2f7

• [Continous Batching] set head_dim when config.head_dim is None (#40159)

  Show description for cfe52ff
  kashif

  authored
  cfe52ff

  Copy full SHA for cfe52ff

• Fix CI: Use correct import in SAM for torchvision InterpolationMode (#40160)

  Show description for c47544b
  manueldeprada

  authored
  c47544b

  Copy full SHA for c47544b

• [efficientloftr] fix bugs and follow original cross attn implementation strictly (#40141)

  Show description for 22e89e5
  

  sbucaille

  and

  qubvel

  authored
  22e89e5

  Copy full SHA for 22e89e5

• [Cohere2Vision] remove unused arg (#40103)

  Show description for 252364f
  zucchini-nlp

  authored
  252364f

  Copy full SHA for 252364f

Commits on Aug 13, 2025

• Create self-scheduled-amd-mi355-caller.yml (#40134)

  glegendre01

  authored
  e446372

  Copy full SHA for e446372

• Update Dockerfiles to install packages inside a virtual environment (#39098)

  Show description for be1ab51
  

  Sai-Suraj-27

  and

  ydshieh

  authored
  be1ab51

  Copy full SHA for be1ab51

• Add pytest marker: `torch_compile_test` and `torch_export_test` (#39950)

  Show description for 591708d
  ydshieh

  authored
  591708d

  Copy full SHA for 591708d

• Fix quantized cache with only cache_implementation in generate (#40144)

  Show description for 12e49cd
  Cyrilvallez

  authored
  12e49cd

  Copy full SHA for 12e49cd

• 🌐 [i18n-KO] Translated `gemma3.md` to Korean (#39865)

  Show description for e651ae0
  

  seopp

  and

  chhaewxn

  authored
  e651ae0

  Copy full SHA for e651ae0

• updated visualBERT modelcard (#40057)

  Show description for 0f9c259
  Anil-Red

  authored
  0f9c259

  Copy full SHA for 0f9c259

• Remove an old badly designed test (#40142)

  Show description for 412c9c3
  Cyrilvallez

  authored
  412c9c3

  Copy full SHA for 412c9c3

• [docs] Fix ko toctree (#40138)

  Show description for eb5768a
  stevhliu

  authored
  eb5768a

  Copy full SHA for eb5768a

• Add Segment Anything 2 (SAM2) (#32317)

  Show description for 68a13cd
  

  6 peopleauthored
  68a13cd

  Copy full SHA for 68a13cd

• Fix Janus (#40140)

  Show description for 25ad9c8
  Cyrilvallez

  authored
  25ad9c8

  Copy full SHA for 25ad9c8

• gpt oss is important (#40139)

  ArthurZucker

  authored
  bec6926

  Copy full SHA for bec6926

• 🌐 [i18n-KO] Translated `pipelines.md` to Korean (#39577)

  Show description for ab91085
  

  4 peopleauthored
  ab91085

  Copy full SHA for ab91085

• 🚨 Use lru_cache for sine pos embeddings MaskFormer (#40007)

  Show description for 20c6b47
  yonigozlan

  authored
  20c6b47

  Copy full SHA for 20c6b47

• 🌐 [i18n-KO] Translated grounding-dino.md to Korean (#39861)

  Show description for 6b728f1
  

  4 peopleauthored
  6b728f1

  Copy full SHA for 6b728f1