Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

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

• [doc] deepspeed universal checkpoint (#35015)

  Show description for c9c682d
  

  huyiwen

  and

  stevhliu

  authored
  c9c682d

  Copy full SHA for c9c682d

• Refactor/fix Cohere2 (#35594)

  Show description for 3a4ae6e
  Cyrilvallez

  authored
  3a4ae6e

  Copy full SHA for 3a4ae6e

• [tokenizers] Ensure that add_prefix_space is propagated to backend_tokenizer.pre_tokenizer (#35593)

  Show description for 32e0db8
  tomaarsen

  authored
  32e0db8

  Copy full SHA for 32e0db8

• Fix modular edge case + modular sorting order (#35562)

  Show description for 46276f9
  Cyrilvallez

  authored
  46276f9

  Copy full SHA for 46276f9

• PR for Issue #22694: Fixed Training Evaluation table display for VSCode (#35557)

  dedsec995

  authored
  d3fe9fa

  Copy full SHA for d3fe9fa

• Small fix rope kwargs (#35589)

  Show description for 395b114
  molbap

  authored
  395b114

  Copy full SHA for 395b114

• Fix flaky SwitchTransformersModelTest::test_training_gradient (#35587)

  Show description for 82dd6c1
  

  ydshieh

  and

  ArthurZucker

  authored
  82dd6c1

  Copy full SHA for 82dd6c1

• tokenizer train from iterator without pre_tokenizers (#35396)

  Show description for eb4579c
  ArthurZucker

  authored
  eb4579c

  Copy full SHA for eb4579c

• feat: add TP plan for granite (#35573)

  Show description for 320512d
  kmehant

  authored
  320512d

  Copy full SHA for 320512d

• [Idefics3] Move image features to same device as input embeds (#35100)

  Show description for 633da1b
  

  3 peopleauthored
  633da1b

  Copy full SHA for 633da1b

• Add inputs_embeds param to ModernBertModel (#35373)

  Show description for 832c619
  

  jxmorris12

  and

  tomaarsen

  authored
  832c619

  Copy full SHA for 832c619

• Fix flaky test_batching_equivalence (#35564)

  Show description for 1b2f942
  ydshieh

  authored
  1b2f942

  Copy full SHA for 1b2f942

• Setup loss_type in config at model init time (#34616)

  Show description for 4adc415
  

  ChanderG

  and

  ArthurZucker

  authored
  4adc415

  Copy full SHA for 4adc415

• Re-add missing __all__ for Cohere and Phi3 (#35578)

  Show description for c8ab6ce
  Cyrilvallez

  authored
  c8ab6ce

  Copy full SHA for c8ab6ce

• Minor fix in video text 2 text docs (#35546)

  Show description for 487c31a
  merveenoyan

  authored
  487c31a

  Copy full SHA for 487c31a

• More model refactoring! (#35359)

  Show description for 965a2fb
  Cyrilvallez

  authored
  965a2fb

  Copy full SHA for 965a2fb

• Don't show warning for inv_freq buffers (#35255)

  Show description for 137965c
  zucchini-nlp

  authored
  137965c

  Copy full SHA for 137965c

Pagination

Previous

Next