Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Jan 14, 2025

• Enhance DataCollatorForLanguageModeling with Configurable Token Replacement Probabilities (#35251)

  Show description for c61fcde
  mahdibaghbanzadeh

  authored
  c61fcde

  Copy full SHA for c61fcde

• Enhanced Installation Section in README.md (#35094)

  Show description for b0cdbd9
  egojoseph

  authored
  b0cdbd9

  Copy full SHA for b0cdbd9

• Fix : add require_read_token for gemma2 gated model (#35687)

  Show description for a11041f
  MekkCyber

  authored
  a11041f

  Copy full SHA for a11041f

• Fix expected output for ggml test (#35686)

  Show description for df2a812
  MekkCyber

  authored
  df2a812

  Copy full SHA for df2a812

• Fix : HQQ config when hqq not available (#35655)

  Show description for 0506365
  MekkCyber

  authored
  0506365

  Copy full SHA for 0506365

• Update torchao.md: use auto-compilation (#35490)

  Show description for 715fdd6
  

  martin0258

  and

  SunMarc

  authored
  715fdd6

  Copy full SHA for 715fdd6

• Fix : adding einops lib in the CI docker for some bitsandbytes tests (#35652)

  Show description for 4b8d1f7
  MekkCyber

  authored
  4b8d1f7

  Copy full SHA for 4b8d1f7

Commits on Jan 13, 2025

• Fix zero_shot_image_classification documentation guide link in SigLIP (#35671)

  aretrace

  authored
  34f76bb

  Copy full SHA for 34f76bb

• Add-helium (#35669)

  Show description for c23a1c1
  

  ArthurZucker

  and

  LaurentMazare

  authored
  c23a1c1

  Copy full SHA for c23a1c1

• [i18n-ar] Translated file : docs/source/ar/tasks/token_classification.md into Arabic (#35193)

  Show description for a3f8232
  

  AhmedAlmaghz

  and

  abodacs

  authored
  a3f8232

  Copy full SHA for a3f8232

• [tests] make cuda-only tests device-agnostic (#35607)

  Show description for 2fa876d
  faaany

  authored
  2fa876d

  Copy full SHA for 2fa876d

• [Compile] Only test compiling model forward pass (#35658)

  Show description for e6f9b03
  ArthurZucker

  authored
  e6f9b03

  Copy full SHA for e6f9b03

• Enable different torch dtype in sub models (#34873)

  Show description for 84a6789
  zucchini-nlp

  authored
  84a6789

  Copy full SHA for 84a6789

• [Phi] bias should be True (#35650)

  Show description for 8708917
  ArthurZucker

  authored
  8708917

  Copy full SHA for 8708917

• Removed some duplicated code (#35637)

  Show description for 91f14f1
  

  Sai-Suraj-27

  and

  molbap

  authored
  91f14f1

  Copy full SHA for 91f14f1

• Fix whisper compile (#35413)

  Show description for b8c34d9
  jiqing-feng

  authored
  b8c34d9

  Copy full SHA for b8c34d9

• Fix device in rope module when using dynamic updates (#35608)

  Show description for cd44bdb
  Cyrilvallez

  authored
  cd44bdb

  Copy full SHA for cd44bdb

Commits on Jan 10, 2025

• Update codeowners with individual model owners (#35595)

  Show description for 15bd3e6
  

  3 peopleauthored
  15bd3e6

  Copy full SHA for 15bd3e6

• Skip MobileNetV1ModelTest::test_batching_equivalence for now (#35614)

  Show description for 1e3c6c1
  ydshieh

  authored
  1e3c6c1

  Copy full SHA for 1e3c6c1

• Fix flaky test_beam_search_low_memory (#35611)

  Show description for 04eae98
  ydshieh

  authored
  04eae98

  Copy full SHA for 04eae98

• Let EarlyStoppingCallback not require load_best_model_at_end (#35101)

  Show description for b02828e
  muellerzr

  authored
  b02828e

  Copy full SHA for b02828e

• Added error when sequence length is bigger than max_position_embeddings (#32156)

  Show description for 0aaf124
  Taha1506

  authored
  0aaf124

  Copy full SHA for 0aaf124

• Use inherit tempdir makers for tests + fix failing DS tests (#35600)

  Show description for 1211e61
  muellerzr

  authored
  1211e61

  Copy full SHA for 1211e61

• Fix flaky test_custom_4d_attention_mask (#35606)

  Show description for bbc0004
  ydshieh

  authored
  bbc0004

  Copy full SHA for bbc0004

• v4.49.0-dev

  ArthurZucker

  committed
  f63829c

  Copy full SHA for f63829c

• [WIP] Emu3: add model (#33770)

  Show description for 52e1f87
  

  3 peopleauthored
  52e1f87

  Copy full SHA for 52e1f87

• Fix flex_attention in training mode (#35605)

  Show description for ccc0381
  Cyrilvallez

  authored
  ccc0381

  Copy full SHA for ccc0381

• Remove benchmark.py after #34275

  ArthurZucker

  committed
  a9bd1e6

  Copy full SHA for a9bd1e6

• Chat template: return vectorized output in processors (#34275)

  Show description for e0646f3
  

  3 peopleauthored
  e0646f3

  Copy full SHA for e0646f3

• Add Moonshine (#34784)

  Show description for 5f087d1
  

  eustlb

  and

  xenova

  authored
  5f087d1

  Copy full SHA for 5f087d1

• Skip torchscript tests if a cache object is in model's outputs (#35596)

  Show description for 6f127d3
  ydshieh

  authored
  6f127d3

  Copy full SHA for 6f127d3

• ModernBert: reuse GemmaRotaryEmbedding via modular + Integration tests (#35459)

  Show description for 6b73ee8
  tomaarsen

  authored
  6b73ee8

  Copy full SHA for 6b73ee8

Commits on Jan 9, 2025

• Add flex_attn to diffllama (#35601)

  Show description for 8de7b1b
  muellerzr

  authored
  8de7b1b

  Copy full SHA for 8de7b1b

• ModernBERT bug fixes (#35404)

  Show description for 1e3ddcb
  

  warner-benjamin

  and

  tomaarsen

  authored
  1e3ddcb

  Copy full SHA for 1e3ddcb

• add _supports_flex_attn = True for models that do support it (#35598)

  Show description for e97d7a5
  ArthurZucker

  authored
  e97d7a5

  Copy full SHA for e97d7a5

Pagination

Previous

Next