Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Nov 4, 2025

• [FPQuant] MXFP8 and MXFP4 backwards support (#41897)

  Show description for 020e713
  BlackSamorez

  authored
  020e713

  Copy full SHA for 020e713

• [v5] Deprecate Text2Text and related pipelines (#41996)

  Show description for 371ef0f
  Rocketknight1

  authored
  371ef0f

  Copy full SHA for 371ef0f

• [kernels] Fix XPU layernorm kernel (#41583)

  Show description for 6efc179
  

  MekkCyber

  and

  SunMarc

  authored
  6efc179

  Copy full SHA for 6efc179

• add fuyu fast image processors (#41817)

  Show description for 325810e
  

  DeXtAr47-oss

  and

  yonigozlan

  authored
  325810e

  Copy full SHA for 325810e

• Add GLPNImageProcessorFast (#41725)

  Show description for 9a19171
  

  Aravind-11

  and

  yonigozlan

  authored
  9a19171

  Copy full SHA for 9a19171

• Fix default image_rows and image_cols initialization in Idefics3 and SmolVLM processors (#41871)

  Show description for 26fca86
  MilkClouds

  authored
  26fca86

  Copy full SHA for 26fca86

• Fix issue with from pretrained and kwargs in image processors (#41997)

  Show description for 900cf9d
  yonigozlan

  authored
  900cf9d

  Copy full SHA for 900cf9d

• add back `logging_dir` (#42013)

  Show description for 154d510
  

  SunMarc

  and

  github-actions[bot]

  authored
  154d510

  Copy full SHA for 154d510

• Fix continuous batching tests (#42012)

  Show description for e3d4fa6
  Rocketknight1

  authored
  e3d4fa6

  Copy full SHA for e3d4fa6

• Reduce the number of benchmark in the CI (#42008)

  Show description for dd4e048
  remi-or

  authored
  dd4e048

  Copy full SHA for dd4e048

• Correct syntax error in trainer.md (#42001)

  Show description for 6ff4fab
  Yacklin

  authored
  6ff4fab

  Copy full SHA for 6ff4fab

• Fix `torch+deepspeed` docker file (#41985)

  Show description for 6d4450e
  ydshieh

  authored
  6d4450e

  Copy full SHA for 6d4450e

• DOC Fix typo in argument name: pseudoquant (#41994)

  Show description for aee5c23
  BenjaminBossan

  authored
  aee5c23

  Copy full SHA for aee5c23

• [kernels] change import time in KernelConfig (#42004)

  Show description for 5b6c209
  MekkCyber

  authored
  5b6c209

  Copy full SHA for 5b6c209

• Fix `run slow v2`: empty report when there is only one model (#42002)

  Show description for 258c76e
  ydshieh

  authored
  258c76e

  Copy full SHA for 258c76e

Commits on Nov 3, 2025

• Fixed wrong padding value in OWLv2 (#41938)

  Show description for 64397a8
  

  gjamesgoenawan

  and

  yonigozlan

  authored
  64397a8

  Copy full SHA for 64397a8

• Integrate colqwen2.5 using colqwen2 modelling code (#40600)

  Show description for cd30961
  

  4 peopleauthored
  cd30961

  Copy full SHA for cd30961

• fix 3 failed test cases for video_llama_3 model on Intel XPU (#41931)

  Show description for dd8f231
  kaixuanliu

  authored
  dd8f231

  Copy full SHA for dd8f231

• fix (CI): Refactor SSH runners (#41991)

  Show description for 1619a34
  glegendre01

  authored
  1619a34

  Copy full SHA for 1619a34

• More data in benchmarking (#41848)

  Show description for ff0f7d6
  remi-or

  authored
  ff0f7d6

  Copy full SHA for ff0f7d6

• Move the Mi355 to regular docker (#41989)

  Show description for 8030536
  remi-or

  authored
  8030536

  Copy full SHA for 8030536

• [kernels] Add Tests & CI for kernels (#41765)

  Show description for a623cda
  MekkCyber

  authored
  a623cda

  Copy full SHA for a623cda

• Fix `torchcodec` version in quantization docker file (#41988)

  Show description for 7d5160b
  ydshieh

  authored
  7d5160b

  Copy full SHA for 7d5160b

• docs: add continuous batching page (#41847)

  Show description for 22e39df
  McPatate

  authored
  22e39df

  Copy full SHA for 22e39df

• fix: dict[RopeParameters] to dict[str, RopeParameters] (#41963)

  RyanMullins

  authored
  63fbd50

  Copy full SHA for 63fbd50

• test tensor parallel: make tests for dense model more robust (#41968)

  Show description for b433ec8
  

  4 peopleauthored
  b433ec8

  Copy full SHA for b433ec8

• Use indices as position_ids in modernebert (#41789)

  Show description for 3c16c1a
  remi-or

  authored
  3c16c1a

  Copy full SHA for 3c16c1a

Commits on Nov 2, 2025

• add support for saving encoder only so any parakeet model can be loaded for inference (#41969)

  Show description for b9f90dc
  

  nithinraok

  and

  eustlb

  authored
  b9f90dc

  Copy full SHA for b9f90dc

• Fix `autoawq[kernels]` installation in quantization docker file (#41978)

  Show description for 37a6296
  ydshieh

  authored
  37a6296

  Copy full SHA for 37a6296

• Fix `detectron2` installation in docker files (#41975)

  Show description for 0ed6d51
  ydshieh

  authored
  0ed6d51

  Copy full SHA for 0ed6d51

Commits on Nov 1, 2025

• Run slow v2 (#41914)

  Show description for 8fb854c
  ydshieh

  authored
  8fb854c

  Copy full SHA for 8fb854c

Commits on Oct 31, 2025

• Fix typo in image_processing_lfm2_vl_fast (#41940)

  Show description for a0bf5a8
  yonigozlan

  authored
  a0bf5a8

  Copy full SHA for a0bf5a8

• make recurrent_gemma and voxtral cases pass on xpu (#41958)

  Show description for 6fb6d3c
  

  yao-matrix

  and

  ydshieh

  authored
  6fb6d3c

  Copy full SHA for 6fb6d3c

• [v5] Return a BatchEncoding dict from apply_chat_template by default (#41626)

  Show description for 5f8d02f
  Rocketknight1

  authored
  5f8d02f

  Copy full SHA for 5f8d02f

• V4.57.1 training ci: Refactor `test_tensor_parallel.py` (#41918)

  Show description for 4418728
  3outeille

  authored
  4418728

  Copy full SHA for 4418728

Pagination

Previous

Next