Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Sep 23, 2025

• [Bugfix] Fix for the import error from #24588 (#25481)

  Show description for 690f948
  gshtras

  authored
  690f948

  Copy full SHA for 690f948

• [Build] Update Xgrammar to 0.1.25 (#25467)

  Show description for 08275ec
  chaunceyjiang

  authored
  08275ec

  Copy full SHA for 08275ec

• [Bugfix] gpt-oss container tool output bug (#25485)

  Show description for c828d1b
  

  alecsolder

  and

  Alec Solder

  authored
  c828d1b

  Copy full SHA for c828d1b

• [CI] Fix Pre-commit Issue (#25497)

  Show description for 8b8a8af
  yewentao256

  authored
  8b8a8af

  Copy full SHA for 8b8a8af

• Enable symmetric memory all reduce by default only enabling for TP (#25070)

  Show description for 8bdd8b5
  

  ilmarkov

  and

  mgoin

  authored
  8bdd8b5

  Copy full SHA for 8bdd8b5

• [Bugfix] Lower gpt-oss max cudagraph size to 992 to be compatible with FA3 (#25508)

  Show description for a8ffc4f
  mgoin

  authored
  a8ffc4f

  Copy full SHA for a8ffc4f

• [Speculators][Speculative Decoding] Fix gpt-oss eagle3 accuracy issue (#25406)

  Show description for d5944d5
  jiahanc

  authored
  d5944d5

  Copy full SHA for d5944d5

• [Perf] Change default CUDAGraphMode from PIECEWISE to FULL_AND_PIECEWISE (#25444)

  Show description for 24fab45
  mgoin

  authored
  24fab45

  Copy full SHA for 24fab45

• [Performance] Move apply_w8a8_block_fp8_linear to an op class (#24666)

  Show description for 6340025
  

  3 peopleauthored
  6340025

  Copy full SHA for 6340025

• [core] add nccl symmetric memory for all reduce (#24532)

  Show description for 8c1c81a
  

  Amir-19

  and

  mgoin

  authored
  8c1c81a

  Copy full SHA for 8c1c81a

• [ROCm] Add skinny gemm bias support for dtypes fp16,bf16,fp8 (#24988)

  Show description for a3a7828
  amd-hhashemi

  authored
  a3a7828

  Copy full SHA for a3a7828

• [Core] Ensure LoRA linear respect the base_layer's tp_size and tp_rank (#25487)

  Show description for 5abb117
  jeejeelee

  authored
  5abb117

  Copy full SHA for 5abb117

• [Spec Decode][CI] Add e2e test for `examples/spec_decode.py` and prevent breaking Acceptance Length (#24531)

  Show description for 867ecdd
  

  ekagra-ranjan

  and

  ywang96

  authored
  867ecdd

  Copy full SHA for 867ecdd

• [Misc] Reduce initialization time of auto_tune (#23682)

  Show description for 24e8222
  wdhongtw

  authored
  24e8222

  Copy full SHA for 24e8222

• [V1][Kernel] Add triton implementation for `reshape_and_cache_flash` (#24503)

  Show description for 100b630
  

  3 peopleauthored
  100b630

  Copy full SHA for 100b630

• Use macro guard CUDA functions for back compatibility in grouped_topk_kernel.cu (#25346)

  Show description for 527821d
  

  6 peopleauthored
  527821d

  Copy full SHA for 527821d

• [Log] Optimize kv cache memory log from Bytes to GiB (#25204)

  Show description for 846197f
  yewentao256

  authored
  846197f

  Copy full SHA for 846197f

• [BugFix] Fix UB in per_token_group_quant.cu (#24913)

  Show description for 2357480
  rivos-shreeasish

  authored
  2357480

  Copy full SHA for 2357480

• [Kernels] Support blocked fp8 quantization for compressed tensors MoE (#25219)

  Show description for f11e3c5
  

  bnellnm

  and

  mgoin

  authored
  f11e3c5

  Copy full SHA for f11e3c5

• Add backward compatibility for `GuidedDecodingParams` (#25422)

  Show description for 875d6de
  hmellor

  authored
  875d6de

  Copy full SHA for 875d6de

• [Core/DBO][2/N] Dual-Batch Overlap add DeepEP High Throughput support and Prefill support (#24845)

  Show description for cc1dc7e
  

  4 peopleauthored
  cc1dc7e

  Copy full SHA for cc1dc7e

• [V1] Remove V0 code paths for Hybrid models (#25400)

  Show description for a903669
  tdoublep

  authored
  a903669

  Copy full SHA for a903669

• [UX] Change kv-cache-memory log level to debug (#25479)

  Show description for 2c58742
  mgoin

  authored
  2c58742

  Copy full SHA for 2c58742

• [XPU] Fix MOE DP accuracy issue on XPU (#25465)

  faaany

  authored
  4c966e4

  Copy full SHA for 4c966e4

• [Docs] NixlConnector quickstart guide (#24249)

  Show description for da5e7e4
  

  panpan0000

  and

  NickLucche

  authored
  da5e7e4

  Copy full SHA for da5e7e4

• [P/D] Support NIXL connector to disconnect during a clean shutdown (#24423)

  Show description for f05a4f0
  

  chaunceyjiang

  and

  markmc

  authored
  f05a4f0

  Copy full SHA for f05a4f0

• [BugFix] Register expert_map as named buffer for wake_up and sleep (#25458)

  Show description for 61d1b35
  

  wuxibin89

  and

  youkaichao

  authored
  61d1b35

  Copy full SHA for 61d1b35

• [CI/Build] Fix disabled v1 attention backend selection test (#25471)

  Show description for b6a136b
  Isotr0py

  authored
  b6a136b

  Copy full SHA for b6a136b

• [docs] Benchmark Serving Incorrect Arg (#25474)

  Show description for 0d9fe26
  vllmellm

  authored
  0d9fe26

  Copy full SHA for 0d9fe26

• [Core] Optimize LoRA weight loading (#25403)

  Show description for 273690a
  jeejeelee

  authored
  273690a

  Copy full SHA for 273690a

• [Bugfix] Fix idefics3 `tie_word_embeddings` (#25454)

  Show description for 231c2c6
  Isotr0py

  authored
  231c2c6

  Copy full SHA for 231c2c6

• [Test]: Hermes tool parser stream output error in Qwen3 case (#25203)

  Show description for 4322c55
  ahartel

  authored
  4322c55

  Copy full SHA for 4322c55

• [Misc] Move DP for ViT code inside model executor dir (#25459)

  Show description for babad6e
  DarkLight1337

  authored
  babad6e

  Copy full SHA for babad6e

• [Frontend] Add a new xml-based tool parser for qwen3-coder (#25028)

  Show description for 9383cd6
  Zhikaiiii

  authored
  9383cd6

  Copy full SHA for 9383cd6

• Handle triton kernel import exception (#25319)

  Show description for ba8d216
  minosfuture

  authored
  ba8d216

  Copy full SHA for ba8d216

Pagination

Previous

Next