Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Jun 20, 2025

• Update blip model card (#38513)

  Show description for 2166b6b
  

  3 peopleauthored
  2166b6b

  Copy full SHA for 2166b6b

• Fix custom generate from local directory (#38916)

  Show description for 166e823
  manueldeprada

  authored
  166e823

  Copy full SHA for 166e823

• Switch to use A10 progressively (#38936)

  Show description for 3d34b92
  ydshieh

  authored
  3d34b92

  Copy full SHA for 3d34b92

• Fix more flaky `test_initialization` (#38932)

  Show description for b8059e1
  ydshieh

  authored
  b8059e1

  Copy full SHA for b8059e1

• Correctly raise error for awq quantization (#38945)

  Show description for 5ee60f9
  Cyrilvallez

  authored
  5ee60f9

  Copy full SHA for 5ee60f9

• Pin PyTorch extras for AMD containers (#38941)

  Show description for 8ac2d75
  

  ahadnagy

  and

  ivarflakstad

  authored
  8ac2d75

  Copy full SHA for 8ac2d75

• Add kwargs for timm.create_model in TimmWrapper (#38860)

  Show description for 9120567
  qubvel

  authored
  9120567

  Copy full SHA for 9120567

• [static cache] fix device map per layer in VLMs (#38488)

  Show description for ff95974
  zucchini-nlp

  authored
  ff95974

  Copy full SHA for ff95974

• Remove `ALL_LAYERNORM_LAYERS` (#38922)

  Show description for aa42987
  Cyrilvallez

  authored
  aa42987

  Copy full SHA for aa42987

• add pytorch-xpu Dockerfile (#38875)

  Show description for 38a9b70
  yao-matrix

  authored
  38a9b70

  Copy full SHA for 38a9b70

• Modernbert fixes (#38912)

  Show description for 9bcdd5c
  

  3 peopleauthored
  9bcdd5c

  Copy full SHA for 9bcdd5c

• Skip some tests for now (#38931)

  Show description for 31d30b7
  ydshieh

  authored
  31d30b7

  Copy full SHA for 31d30b7

Commits on Jun 19, 2025

• Remove deprecated classes in modeling_utils.py (#38919)

  Show description for 0725cd6
  Cyrilvallez

  authored
  0725cd6

  Copy full SHA for 0725cd6

• feat: add flexible Liger Kernel configuration to TrainingArguments (#38911)

  Show description for 797860c
  hamza-hcompany

  authored
  797860c

  Copy full SHA for 797860c

• Allow make-fixup on main branch, albeit slowly (#38892)

  Show description for 89b35be
  Rocketknight1

  authored
  89b35be

  Copy full SHA for 89b35be

• feat: Add granite architectures to auto tokenizer name mappings (#38802)

  Show description for 9a02e76
  gabe-l-hart

  authored
  9a02e76

  Copy full SHA for 9a02e76

• Fix ReDOS in tokenizer digit substitution (#38844)

  Show description for 54a0216
  Rocketknight1

  authored
  54a0216

  Copy full SHA for 54a0216

• Skip sdpa tests if submodule does not support sdpa (#38907)

  ivarflakstad

  authored
  af6120b

  Copy full SHA for af6120b

• Fix `FalconMambaIntegrationTests` (#38566)

  Show description for 5d26a38
  ydshieh

  authored
  5d26a38

  Copy full SHA for 5d26a38

• align xpu's autocast behavior w/ cuda by using device agnostic torch APIs (#38284)

  Show description for a9ce8c6
  

  yao-matrix

  and

  ydshieh

  authored
  a9ce8c6

  Copy full SHA for a9ce8c6

• Fix unnecessary super calls (#38897)

  Show description for 0a53df1
  cyyever

  authored
  0a53df1

  Copy full SHA for 0a53df1

• Fix `fsmt` tests (#38904)

  Show description for b949747
  ydshieh

  authored
  b949747

  Copy full SHA for b949747

• [phi-4] use mel filters from audio utils (#36966)

  Show description for 11738f8
  

  3 peopleauthored
  11738f8

  Copy full SHA for 11738f8

• Use `raise from e` in `hub.py` utility (#37241)

  Show description for f7b2182
  

  Wauplin

  and

  ArthurZucker

  authored
  f7b2182

  Copy full SHA for f7b2182

• Add support for specifying revisions when pushing to Hub via internal Trainer call (#36852)

  Show description for 3756bf1
  

  IsaacBreen

  and

  ArthurZucker

  authored
  3756bf1

  Copy full SHA for 3756bf1

Commits on Jun 18, 2025

• Update bamba model card (#38853)

  Show description for 458e0b3
  

  druvdub

  and

  stevhliu

  authored
  458e0b3

  Copy full SHA for 458e0b3

• [video processor] fix slow tests (#38881)

  Show description for ea01334
  

  zucchini-nlp

  and

  ydshieh

  authored
  ea01334

  Copy full SHA for ea01334

• 36978 | Fast image processor for DPT model (#37481)

  Show description for b922b22
  

  3 peopleauthored
  b922b22

  Copy full SHA for b922b22

• Docs: Add custom fine-tuning tutorial to TrOCR model page (#38847)

  Show description for c27f628
  

  Ashutosh-4485

  and

  stevhliu

  authored
  c27f628

  Copy full SHA for c27f628

• log: Add logging when using split_batches and per_device_train_batch_size (#38633)

  Show description for 0a289d1
  

  KeshavSingh29

  and

  SunMarc

  authored
  0a289d1

  Copy full SHA for 0a289d1

• [bugfix] fix ATTN_MASK_NPU device mismatch error on multi-device NPU … (#38876)

  Show description for c55d806
  qykong

  authored
  c55d806

  Copy full SHA for c55d806

• Fix loop var naming (#38885)

  Rocketknight1

  authored
  9cd7570

  Copy full SHA for 9cd7570

• More PYUP fixes (#38883)

  Show description for 1fc67a2
  cyyever

  authored
  1fc67a2

  Copy full SHA for 1fc67a2

• null deepspeed_plugin in args for wandb callback fake trainer (#38867)

  winglian

  authored
  12d4c5b

  Copy full SHA for 12d4c5b

• Fixed markdown for BertTokenizer's '[CLS]' token. (#38506)

  eu90h

  authored
  3620b32

  Copy full SHA for 3620b32

Pagination

Previous

Next