Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Nov 12, 2025

• [XPU]Fix crash due to removed VLLM_USE_V1 attribute (#28520)

  Show description for a4730c1
  chaojun-zhang

  authored
  a4730c1

  Copy full SHA for a4730c1

• [Model] fix glm4_moe_mtp load weights with GLM-4.6 checkpoint. (#27597)

  Show description for d3ade61
  

  wuyaoxuehun

  and

  wuao.scotty

  authored
  d3ade61

  Copy full SHA for d3ade61

• [BugFix]: --enable-lora with model granite-4.0-micro crash (#27733)

  Show description for 1761dea
  yyzxw

  authored
  1761dea

  Copy full SHA for 1761dea

• [CI] Introduce autorun_on_main feature (#27836)

  Show description for c748355
  hl475

  authored
  c748355

  Copy full SHA for c748355

• [CI/Build] Fix crash due to removed VLLM_USE_V1 attribute in EPD (#28521)

  Show description for 91864b7
  

  3 peopleauthored
  91864b7

  Copy full SHA for 91864b7

• [Core] Cache `vllm_is_batch_invariant` (#28304)

  Show description for ac0bb2c
  lgeiger

  authored
  ac0bb2c

  Copy full SHA for ac0bb2c

• [Benchmark] Add retry support to fix workload bias in multi-turn benchmark (#28493)

  ai-jz

  authored
  f31419e

  Copy full SHA for f31419e

• [BugFix] Add fallback path in `apply_rotary_pos_emb_flashattn` for non-cuda platforms (#28447)

  Show description for b9ce9a3
  faaany

  authored
  b9ce9a3

  Copy full SHA for b9ce9a3

• [Core] Encoder separation for Encode-Prefill-Decode Disaggregation (#25233)

  Show description for 4ccffe5
  

  7 peopleauthored
  4ccffe5

  Copy full SHA for 4ccffe5

• [Model][Qwen3VL] Simplify `get_mrope_input_positions` using numpy (#28302)

  Show description for cbb799e
  lgeiger

  authored
  cbb799e

  Copy full SHA for cbb799e

• `VLLM_USE_TRITON_FLASH_ATTN` V0 variable deprecation (#27611)

  Show description for 9f0247c
  AndreasKaratzas

  authored
  9f0247c

  Copy full SHA for 9f0247c

• [CPU] Refactor CPU attention backend (#27954)

  Show description for 7f829be
  bigPYJ1151

  authored
  7f829be

  Copy full SHA for 7f829be

• [[V0 deprecation]]Remove VLLM_USE_V1 env (#28204)

  Show description for e171039
  wangxiyuan

  authored
  e171039

  Copy full SHA for e171039

• [Performance] Cache loaded custom logitsprocs to avoid overheads (#28462)

  Show description for 3f770f4
  Isotr0py

  authored
  3f770f4

  Copy full SHA for 3f770f4

• [Frontend] Change CompilationMode to a proper Enum (#28165)

  Show description for 48c8793
  gmagogsfm

  authored
  48c8793

  Copy full SHA for 48c8793

• [BugFix] Graceful handling of torch symm mem errors. (#27671)

  Show description for 1788aa1
  

  ilmarkov

  and

  mgoin

  authored
  1788aa1

  Copy full SHA for 1788aa1

• Use FLASHINFER MLA backend when testing fp8_kv_scale_compile (#28491)

  Show description for d235395
  adabeyta

  authored
  d235395

  Copy full SHA for d235395

Commits on Nov 11, 2025

• [Feature] Allow configuring FlashInfer workspace size (#28269)

  Show description for 412e153
  

  maxyanghu

  and

  gemini-code-assist[bot]

  authored
  412e153

  Copy full SHA for 412e153

• [Bugfix] Disable shared expert overlap if Marlin MoE is used (#28410)

  Show description for e5f599d
  mgoin

  authored
  e5f599d

  Copy full SHA for e5f599d

• Add Zurich vLLM Meetup (#28488)

  Show description for 28534b9
  mgoin

  authored
  28534b9

  Copy full SHA for 28534b9

• [Misc] Cleanup Executor interface (#28441)

  Show description for d4902ba
  wangxiyuan

  authored
  d4902ba

  Copy full SHA for d4902ba

• [TPU] Rename path to tpu platform (#28452)

  Show description for df4d3a4
  kyuyeunk

  authored
  df4d3a4

  Copy full SHA for df4d3a4

• [LoRA][1/N]Remove LoRA extra vocab (#28382)

  Show description for 9d1c474
  jeejeelee

  authored
  9d1c474

  Copy full SHA for 9d1c474

• [Misc] fix typo in DCP comment (#28389)

  Show description for 8c32c6e
  Livinfly

  authored
  8c32c6e

  Copy full SHA for 8c32c6e

• [V0 deprecation] Clean up num_prefill_tokens logic for V0 (#28203)

  Show description for de120bc
  gcanlin

  authored
  de120bc

  Copy full SHA for de120bc

• [Perf] Use np.ndarray instead of list[list[int]] to reduce GC overhead (#28245)

  Show description for 4228be7
  Jialin

  authored
  4228be7

  Copy full SHA for 4228be7

• [Misc] Remove unused attention prefix prefill ops functions (#26971)

  Show description for 76e4dcf
  lgeiger

  authored
  76e4dcf

  Copy full SHA for 76e4dcf

• [BugFix] Fix Siglip2Attention on XPU (#28448)

  Show description for d5edcb8
  faaany

  authored
  d5edcb8

  Copy full SHA for d5edcb8

• [Kernel] Optimize rms_norm kernel (#27931)

  Show description for 6c3c0f8
  xyang16

  authored
  6c3c0f8

  Copy full SHA for 6c3c0f8

• Prefer FlashAttention MLA as default over FlashMLA (#27363)

  Show description for 684f254
  MatthewBonanni

  authored
  684f254

  Copy full SHA for 684f254

• [CI/Build] Refactor Attention backend for test_prefix_prefill from xformers to SDPA (#28424)

  Show description for e553424
  

  zhewenl

  and

  ywang96

  authored
  e553424

  Copy full SHA for e553424

• [Quantization] fix attention quantization of gpt_oss model (#27334)

  Show description for 5a1271d
  xuebwang-amd

  authored
  5a1271d

  Copy full SHA for 5a1271d

• [ROCm][Quantization] extend AMD Quark to support mixed-precision quantized model (#24239)

  Show description for 05576df
  

  3 peopleauthored
  05576df

  Copy full SHA for 05576df

• [Kernel][Perf] fuse QK Norm and RoPE into one cuda kernel for Qwen Model (#27165)

  Show description for 68c09ef
  izhuhaoran

  authored
  68c09ef

  Copy full SHA for 68c09ef

• [NIXL] Generalize block-first backend layouts (FlashInfer-like) (#28282)

  NickLucche

  authored
  a7ef3eb

  Copy full SHA for a7ef3eb

Pagination

Previous

Next