Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Feb 1, 2024

• [docs] Backbone (huggingface#28739)

  Show description for abbffc4
  stevhliu

  authored
  abbffc4

• Add models from deit (huggingface#28302)

  Show description for 23ea674
  

  rajveer43

  and

  stevhliu

  authored
  23ea674

• [docs] fix some bugs about parameter description (huggingface#28806)

  Show description for d98591a
  

  zspo

  and

  p_spozzhang

  authored
  d98591a

• enable graident checkpointing in DetaObjectDetection and add tests in Swin/Donut_Swin (huggingface#28615)

  Show description for e19c12e
  SangbumChoi

  authored
  e19c12e

• Add tip on setting tokenizer attributes (huggingface#28764)

  Show description for 7bc6d76
  Rocketknight1

  authored
  7bc6d76

• Fix symbolic_trace with kv cache (huggingface#28724)

  Show description for 709dc43
  fxmarty

  authored
  709dc43

• Make is_torch_bf16_available_on_device more strict (huggingface#28796)

  Show description for eb8e7a0
  ydshieh

  authored
  eb8e7a0

• Adding [T5/MT5/UMT5]ForTokenClassification (huggingface#28443)

  Show description for 0d26abd
  hackyon

  authored
  0d26abd

Commits on Jan 31, 2024

• [docs] Correct the statement in the docstirng of compute_transition_scores in generation/utils.py (huggingface#28786)

  Ki-Seki

  authored
  7b2bd1f

• Split daily CI using 2 level matrix (huggingface#28773)

  Show description for 4735866
  ydshieh

  authored
  4735866

• Add artifact name in job step to maintain job / artifact correspondence (huggingface#28682)

  Show description for 95346e9
  ydshieh

  authored
  95346e9

• DeepSpeed: hardcode torch.arange dtype on float usage to avoid incorrect initialization (huggingface#28760)

  gante

  authored
  beb2a09

• Flax mistral (huggingface#26943)

  Show description for f7076cd
  kiansierra

  authored
  f7076cd

• Wrap Keras methods to support BatchEncoding (huggingface#28734)

  Show description for 7a49610
  Rocketknight1

  authored
  7a49610

• canonical repos moves (huggingface#28795)

  Show description for 721e2d9
  

  julien-c

  and

  LysandreJik

  authored
  721e2d9

• Resolve DeepSpeed cannot resume training with PeftModel (huggingface#28746)

  Show description for bebeeee
  lh0x00

  authored
  bebeeee

• [Whisper] Refactor forced_decoder_ids & prompt ids (huggingface#28687)

  Show description for 65a926e
  

  patrickvonplaten

  and

  sanchit-gandhi

  authored
  65a926e

• [HFQuantizer] Remove check_packages_compatibility logic (huggingface#28789)

  Show description for f9f1f2a
  younesbelkada

  authored
  f9f1f2a

• don't initialize the output embeddings if we're going to tie them to input embeddings (huggingface#28192)

  Show description for ae0c27a
  tom-p-reichel

  authored
  ae0c27a

• Prevent MLflow exception from disrupting training (huggingface#28779)

  Show description for a937425
  

  codiceSpaghetti

  and

  codiceSpaghetti

  authored
  a937425

• [bnb] Fix bnb slow tests (huggingface#28788)

  Show description for d703eaa
  younesbelkada

  authored
  d703eaa

Commits on Jan 30, 2024

• Pin Torch to <2.2.0 (huggingface#28785)

  Show description for 74c9cfe
  

  Rocketknight1

  and

  ydshieh

  authored
  74c9cfe

• Add tf_keras imports to prepare for Keras 3 (huggingface#28588)

  Show description for 415e9a0
  Rocketknight1

  authored
  415e9a0

• Task-specific pipeline init args (huggingface#28439)

  Show description for 1d489b3
  amyeroberts

  authored
  1d489b3

• [Backbone] Use load_backbone instead of AutoBackbone.from_config (huggingface#28661)

  Show description for 2fa1c80
  amyeroberts

  authored
  2fa1c80

• Further pin pytest version (in a temporary way) (huggingface#28780)

  Show description for c24c524
  ydshieh

  authored
  c24c524

• Fix transformers.utils.fx compatibility with torch<2.0 (huggingface#28774)

  Show description for 6f7d5db
  fxmarty

  authored
  6f7d5db

• Use Conv1d for TDNN (huggingface#25728)

  Show description for 5c8d941
  gau-nernst

  authored
  5c8d941

• [HfQuantizer] Move it to "Developper guides" (huggingface#28768)

  Show description for 866253f
  younesbelkada

  authored
  866253f

• HfQuantizer class for quantization-related stuff in modeling_utils.py (huggingface#26610)

  Show description for d78e78a
  

  5 peopleauthored
  d78e78a

• Move CLIP _no_split_modules to CLIPPreTrainedModel (huggingface#27841)

  Show description for 1f5590d
  lz1oceani

  authored
  1f5590d

• Don't allow passing load_in_8bit and load_in_4bit at the same time (huggingface#28266)

  Show description for a989c6c
  

  osanseviero

  and

  younesbelkada

  authored
  a989c6c

Commits on Jan 29, 2024

• Add French translation: french README.md (huggingface#28696)

  Show description for cd2eb8c
  

  ThibaultLengagne

  and

  Sarapuce

  authored
  cd2eb8c

• Support saving only PEFT adapter in checkpoints when using PEFT + FSDP (huggingface#28297)

  Show description for a055d09
  AjayP13

  authored
  a055d09

• [Whisper] Make tokenizer normalization public (huggingface#28136)

  Show description for da3c79b
  sanchit-gandhi

  authored
  da3c79b

Pagination

PreviousNext