Commits

Branch selector

 main

User selector

All users

All time

Commit History

Commits on May 12, 2026

• [Hybrid] Warmup Mamba2 SSD kernel (#39822)

  Show description for 8f89381
  

  tdoublep

  and

  claude

  authored
  8f89381

  Copy full SHA for 8f89381

• [MXFP4] Support for linear layers + compressed-tensors integration (#41664)

  dsikka

  authored
  a7b801e

  Copy full SHA for a7b801e

• [XPU] bump up vllm-xpu-kernels to v0.1.8 (#42410)

  Show description for 4df1be9
  jikunshang

  authored
  4df1be9

  Copy full SHA for 4df1be9

• [XPU] keep generator state of sycl kernel align with pytorch (#41771)

  Show description for bc03f28
  

  3 peopleauthored
  bc03f28

  Copy full SHA for bc03f28

• [Doc] Fix typo in llm-d documentation link (#42397)

  Show description for 9971329
  

  woernfl

  and

  DarkLight1337

  authored
  9971329

  Copy full SHA for 9971329

• [CI] De-flake Language Models Test (Extended Generation) test_models(False-False-5-32-bigcode/starcoder2-3b) (#42392)

  Show description for fc8bf6e
  haosdent

  authored
  fc8bf6e

  Copy full SHA for fc8bf6e

• [UT][XPU] fix test_parallel_sampling due to global random state (#42388)

  Show description for 07a40ed
  zhenwei-intel

  authored
  07a40ed

  Copy full SHA for 07a40ed

• [CI] Move DockerHub and PyPI publish steps to end of release pipeline (#42355)

  Show description for e1c8776
  khluu

  authored
  e1c8776

  Copy full SHA for e1c8776

• [CI] Migrate remaining B200 jobs to b200-k8s with test fixes (#42387)

  Show description for 1ff9d33
  khluu

  authored
  1ff9d33

  Copy full SHA for 1ff9d33

• [Bugfix] Fix empty channel/recipient in harmony for /v1/responses (#35540)

  Show description for 7f65f84
  

  kg6-sleipnir

  and

  chaunceyjiang

  authored
  7f65f84

  Copy full SHA for 7f65f84

• [Bugfix] Fix double reduce in flashinfer_nvlink_two_sided and flashinfer_nvlink_one_sided backends (#41382)

  Show description for ef34592
  amitz-nv

  authored
  ef34592

  Copy full SHA for ef34592

• [CI] Migrate more B200 jobs to b200-k8s queue (#42356)

  Show description for f69644c
  khluu

  authored
  f69644c

  Copy full SHA for f69644c

• [Frontend] Consolidate Speech to Text entrypoints. (#42370)

  Show description for d37e25f
  

  noooop

  and

  gemini-code-assist[bot]

  authored
  d37e25f

  Copy full SHA for d37e25f

• [XPU] update dp rank w/o env-var isolation (#39856)

  Show description for 8517cda
  

  zhenwei-intel

  and

  jikunshang

  authored
  8517cda

  Copy full SHA for 8517cda

• [Bugfix][Performance Improvement] Improve penalties triton kernel performance (#40657)

  Show description for 4e498b5
  Lucaskabela

  authored
  4e498b5

  Copy full SHA for 4e498b5

• Implement custom dataset class for ASR benchmarking (#41576)

  Show description for 28ee78a
  

  ymoslem

  and

  gemini-code-assist[bot]

  authored
  28ee78a

  Copy full SHA for 28ee78a

• [Fix] Gemma4 Mixed-Resolution Image Co-Batching Crash (#42217)

  Show description for 630492d
  skyloevil

  authored
  630492d

  Copy full SHA for 630492d

• [Bugifx] [Qwen3CoderTool] Restore supports_required_and_named for required tool_choice (#42292)

  Show description for 920bf3e
  chaunceyjiang

  authored
  920bf3e

  Copy full SHA for 920bf3e

Commits on May 11, 2026

• Add VLLM_USE_SPINLOOP_EXT to use more efficient busy polling (#36517)

  Show description for 39dff5f
  pschlan-amd

  authored
  39dff5f

  Copy full SHA for 39dff5f

• [Model Runner V2] Bug fix: logprob dtype int64/int32 issue (#41761)

  Show description for d7af6b3
  

  3 peopleauthored
  d7af6b3

  Copy full SHA for d7af6b3

• [Perf][1/n] Eliminate various GPU<->CPU syncs (#41429)

  Show description for bbee532
  

  njhill

  and

  mergify[bot]

  authored
  bbee532

  Copy full SHA for bbee532

• [Bugfix] Fix DSV4 swiglu_limit on marlin backend (#42287)

  Show description for 5318138
  

  jeejeelee

  and

  zyongye

  authored
  5318138

  Copy full SHA for 5318138

• [CI] Consolidate Speech to Text tests (#42274)

  Show description for a0dc7a0
  noooop

  authored
  a0dc7a0

  Copy full SHA for a0dc7a0

• [BugFix] Prevent orphaned process on NCCL destroy (#39846)

  Show description for 56e5810
  

  jeffreywang88

  and

  tlrmchlsmth

  authored
  56e5810

  Copy full SHA for 56e5810

• [CI] Add tests/parser to CI coverage (#41877)

  Show description for 639cbfd
  sfeng33

  authored
  639cbfd

  Copy full SHA for 639cbfd

• [ROCm][Perf] Fix RMSNorm+Quant fusion for gfx950 (non-fnuz) (#41825)

  Show description for a721315
  

  6 peopleauthored
  a721315

  Copy full SHA for a721315

• [Bugfix] Fix int32 overflow in DeepGEMM SiLU/mul FP8 Triton kernel (#42201)

  Show description for 6fdb493
  

  Flink-ddd

  and

  gemini-code-assist[bot]

  authored
  6fdb493

  Copy full SHA for 6fdb493

• [Docs] Add Apple Silicon documentation for vLLM-Metal GPU support (#41987)

  Show description for cf0d279
  

  alexagriffith

  and

  claude

  authored
  cf0d279

  Copy full SHA for cf0d279

• Add documentation about vLLM FIPS compliance (#42190)

  Show description for 5497ffb
  

  vrdn-23

  and

  russellb

  authored
  5497ffb

  Copy full SHA for 5497ffb

• [Model Runner V2] Fix `seq_lens_cpu_upper_bound` (#42202)

  Show description for 9af6a5e
  njhill

  authored
  9af6a5e

  Copy full SHA for 9af6a5e

• [ROCm][DSv4] implement flash sparse mla with triton kernels (#41812)

  Show description for 7863fff
  whx-sjtu

  authored
  7863fff

  Copy full SHA for 7863fff

• [Perf] Batch invariance with Cutlass fp8 support, 28.9% E2E latency improvement (#40408)

  Show description for 0d453e2
  yewentao256

  authored
  0d453e2

  Copy full SHA for 0d453e2

• [Bug] Fix kimi dtype issue with `mm_projector_forward` (#42081)

  Show description for 3f9c0c2
  yewentao256

  authored
  3f9c0c2

  Copy full SHA for 3f9c0c2

• [Bugfix] Accept canonicalized `modelopt_*` quant_method in `_extract_modelopt_quant_algo` (#42181)

  Show description for a2e776d
  vadiklyutiy

  authored
  a2e776d

  Copy full SHA for a2e776d

• [kv_offload] Move `FilterReusedOffloadingManager` logic to `CPUOffloadingManager` (#41727)

  Show description for 4955990
  

  hickeyma

  and

  gemini-code-assist[bot]

  authored
  4955990

  Copy full SHA for 4955990