Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Apr 11, 2025

• [Bugfix][VLM] Fix failing Phi-4-MM multi-images tests and add vision-speech test (#16424)

  Show description for 9319514
  Isotr0py

  authored
  9319514

  Copy full SHA for 9319514

• Update supported_hardware.md for TPU INT8 (#16437)

  mgoin

  authored
  ed37599

  Copy full SHA for ed37599

• [Llama4] Enable attention temperature tuning by default for long context (>32k) (#16439)

  Show description for 99ef59c
  

  sarckk

  and

  yeqcharlotte

  authored
  99ef59c

  Copy full SHA for 99ef59c

• update benchmark_serving_structured_output to include auto backend (#16438)

  Show description for d544d14
  Chenyaaang

  authored
  d544d14

  Copy full SHA for d544d14

• check input length of sonnet samples (#16423)

  Show description for 3e397a9
  alexey-belyakov

  authored
  3e397a9

  Copy full SHA for 3e397a9

Commits on Apr 10, 2025

• Fix range_ratio Bug in RandomDataset (#16126)

  Show description for 268c325
  jadewang21

  authored
  268c325

  Copy full SHA for 268c325

• [TPU][V1] Disable per-request seed/Generator (#16172)

  Show description for 3cc9af8
  NickLucche

  authored
  3cc9af8

  Copy full SHA for 3cc9af8

• [Bugfix] Fix output token length check logic (#16419)

  Show description for 7cd0bd7
  eeslook

  authored
  7cd0bd7

  Copy full SHA for 7cd0bd7

• [VLM] Avoid unnecessary dummy multimodal data during processing (#16416)

  Show description for 56d4aef
  DarkLight1337

  authored
  56d4aef

  Copy full SHA for 56d4aef

• [V1] Zero-copy tensor/ndarray serialization/transmission (#13790)

  Show description for dd143ef
  njhill

  authored
  dd143ef

  Copy full SHA for dd143ef

• [Model] Reduce redundant computations in mamba2 blocks for Bamba-9B (#15423)

  Show description for daefed0
  

  cyang49

  and

  fabianlim

  authored
  daefed0

  Copy full SHA for daefed0

• [Bugfix] Fix bug when dataset is json (#15899)

  Show description for 5fbab20
  Chenyaaang

  authored
  5fbab20

  Copy full SHA for 5fbab20

• [V1][Spec Decode] Eagle Model loading (#16035)

  Show description for e8224f3
  LiuXiaoxuanPKU

  authored
  e8224f3

  Copy full SHA for e8224f3

• [V1] Set structured output backend to `auto` by default (#15724)

  Show description for 9665313
  russellb

  authored
  9665313

  Copy full SHA for 9665313

• Improve configs - `ParallelConfig` (#16332)

  Show description for 0c54fc7
  hmellor

  authored
  0c54fc7

  Copy full SHA for 0c54fc7

• [TPU][V1] Use `language_model` interface for getting text backbone in MM (#16410)

  Show description for c1b5785
  NickLucche

  authored
  c1b5785

  Copy full SHA for c1b5785

• [VLM] Remove `BaseProcessingInfo.get_mm_max_tokens_per_item` (#16408)

  Show description for 83b824c
  DarkLight1337

  authored
  83b824c

  Copy full SHA for 83b824c

• Fix the torch version parsing logic (#15857)

  houseroad

  authored
  7678fcd

  Copy full SHA for 7678fcd

• [CI] Add auto update workflow for Dockerfile graph (#11879)

  Show description for 8661c02
  WineChord

  authored
  8661c02

  Copy full SHA for 8661c02

• [doc] update the wrong link (#16401)

  Show description for ce8d6b7
  reidliu41

  authored
  ce8d6b7

  Copy full SHA for ce8d6b7

• [Model] Remove image mm limit for LLaMa4 (#16365)

  Show description for 61de3ef
  yeqcharlotte

  authored
  61de3ef

  Copy full SHA for 61de3ef

• Update Numba to 0.61.2 (#16376)

  Show description for ec1f9c8
  cyyever

  authored
  ec1f9c8

  Copy full SHA for ec1f9c8

• [doc] add download model tips (#16389)

  Show description for 65e0909
  reidliu41

  authored
  65e0909

  Copy full SHA for 65e0909

• [Kernel] Use moe_wna16 kernel for compressed tensors wna16 moe models (#16038)

  Show description for c70cf0f
  mgoin

  authored
  c70cf0f

  Copy full SHA for c70cf0f

• [Bugfix] Fix validation error for text-only Mllama 3.2 (#16377)

  Show description for a5d11a5
  DarkLight1337

  authored
  a5d11a5

  Copy full SHA for a5d11a5

• [Misc] Update transformers version limits of multi-modal tests (#16381)

  Show description for 3d4c877
  DarkLight1337

  authored
  3d4c877

  Copy full SHA for 3d4c877

• [Model] use AutoWeightsLoader for deepseek_v2, internlm2 (#16383)

  Show description for a9bd832
  aaron-ang

  authored
  a9bd832

  Copy full SHA for a9bd832

• fix sonnet dataset sample when prefix len is very small (#16379)

  Show description for 417bcef
  Chenyaaang

  authored
  417bcef

  Copy full SHA for 417bcef

• [Bugfix][TPU] Fix TPU validate_request (#16369)

  Show description for baada0e
  mgoin

  authored
  baada0e

  Copy full SHA for baada0e

• [misc] use tqdm.auto where appropriate (#16290)

  Show description for 82eb61d
  

  Benjamin Kitor

  authored
  82eb61d

  Copy full SHA for 82eb61d

• [CI][Bugfix] Pin triton version for CPU (#16384)

  Show description for 0d4d06f
  ywang96

  authored
  0d4d06f

  Copy full SHA for 0d4d06f

• [bugfix] Avoid the time consumption caused by creating dummy videos. (#16371)

  Jintao-Huang

  authored
  4aed0ca

  Copy full SHA for 4aed0ca

• [TPU] Fix dummy loading OOM (#16372)

  Show description for 1621b25
  yaochengji

  authored
  1621b25

  Copy full SHA for 1621b25

• [Model] use AutoWeightsLoader for granite, granitemoe, granitemoeshared, grok1, mixtral (#16325)

  Show description for a564797
  aaron-ang

  authored
  a564797

  Copy full SHA for a564797

• [Bugfix]: do not shutdown server if `skip_special_use=False` for MistralTokenizer (#14094)

  Show description for 1da6a09
  gcalmettes

  authored
  1da6a09

  Copy full SHA for 1da6a09

Pagination

Previous

Next