Commits

Branch selector

 ddee88d

User selector

All users

Datepicker

All time

Commit History

Commits on Jan 28, 2025

• [Neuron][Kernel] NKI-based flash-attention kernel with paged KV cache (vllm-project#11277)

  Show description for ddee88d
  

  liangfu

  and

  JF-D

  authored
  ddee88d

  Copy full SHA for ddee88d

• Update pre-commit hooks (vllm-project#12475)

  Show description for 823ab79
  hmellor

  authored
  823ab79

  Copy full SHA for 823ab79

Commits on Jan 27, 2025

• [Feature] [Spec decode]: Enable MLPSpeculator/Medusa and prompt_logprobs with ChunkedPrefill (vllm-project#10132)

  Show description for 6116ca8
  

  NickLucche

  and

  wallashss

  authored
  6116ca8

  Copy full SHA for 6116ca8

• [FlashInfer] Upgrade to 0.2.0 (vllm-project#11194)

  Show description for 2bc3fbb
  

  abmfy

  and

  youkaichao

  authored
  2bc3fbb

  Copy full SHA for 2bc3fbb

• [V1][CI/Test] Do basic test for top-p & top-k sampling (vllm-project#12469)

  Show description for 3f1fc74
  WoosukKwon

  authored
  3f1fc74

  Copy full SHA for 3f1fc74

• [V1][Metrics] Add initial Prometheus logger (vllm-project#12416)

  Show description for 01ba927
  markmc

  authored
  01ba927

  Copy full SHA for 01ba927

• [Build] Only build 9.0a for scaled_mm and sparse kernels (vllm-project#12339)

  Show description for 103bd17
  LucasWilkinson

  authored
  103bd17

  Copy full SHA for 103bd17

• [Bugfix] Fix gpt2 GGUF inference (vllm-project#12467)

  Show description for ce69f7f
  Isotr0py

  authored
  ce69f7f

  Copy full SHA for ce69f7f

• [V1][Minor] Minor optimizations for update_from_output (vllm-project#12454)

  Show description for 624a1e4
  WoosukKwon

  authored
  624a1e4

  Copy full SHA for 624a1e4

• [Bugfix] Fix missing seq_start_loc in xformers prefill metadata (vllm-project#12464)

  Show description for 372bf08
  Isotr0py

  authored
  372bf08

  Copy full SHA for 372bf08

• [Bugfix] Fix Granite 3.0 MoE model loading (vllm-project#12446)

  Show description for 5204ff5
  DarkLight1337

  authored
  5204ff5

  Copy full SHA for 5204ff5

• [Frontend] Support scores endpoint in run_batch (vllm-project#12430)

  Show description for 0cc6b38
  pooyadavoodi

  authored
  0cc6b38

  Copy full SHA for 0cc6b38

• [V1] Avoid list creation in input preparation (vllm-project#12457)

  Show description for 28e0750
  WoosukKwon

  authored
  28e0750

  Copy full SHA for 28e0750

• [DOC] Add link to vLLM blog (vllm-project#12460)

  Show description for 582cf78
  terrytangyuan

  authored
  582cf78

  Copy full SHA for 582cf78

• [Frontend] Rerank API (Jina- and Cohere-compatible API) (vllm-project#12376)

  Show description for 0034b09
  K-Mistele

  authored
  0034b09

  Copy full SHA for 0034b09

Commits on Jan 26, 2025

• [Build/CI] Fix libcuda.so linkage (vllm-project#12424)

  Show description for 72bac73
  tlrmchlsmth

  authored
  72bac73

  Copy full SHA for 72bac73

• [Bugfix][Kernel] Fix perf regression caused by PR vllm-project#12405 (vllm-project#12434)

  Show description for 68f1114
  LucasWilkinson

  authored
  68f1114

  Copy full SHA for 68f1114

• [Bugfix/CI] Fix broken kernels/test_mha.py (vllm-project#12450)

  tlrmchlsmth

  authored
  72f4880

  Copy full SHA for 72f4880

• [Bugfix] Disable w16a16 2of4 sparse CompressedTensors24 (vllm-project#12417)

  Show description for aa2cd2c
  

  tlrmchlsmth

  and

  mgoin

  authored
  aa2cd2c

  Copy full SHA for aa2cd2c

• [Frontend] generation_config.json for maximum tokens(vllm-project#12242)

  Show description for 9ddc352
  

  9 peopleauthored
  9ddc352

  Copy full SHA for 9ddc352

• [Misc] Revert FA on ViT vllm-project#12355 and vllm-project#12435 (vllm-project#12445)

  ywang96

  authored
  a525527

  Copy full SHA for a525527

• [V1][Bugfix] Fix assertion when mm hashing is turned off (vllm-project#12439)

  Show description for 0ee349b
  ywang96

  authored
  0ee349b

  Copy full SHA for 0ee349b

• [V1][Perf] Reduce scheduling overhead in model runner after cuda sync (vllm-project#12094)

  Show description for fa63e71
  youngkent

  authored
  fa63e71

  Copy full SHA for fa63e71

• [Misc][Bugfix] FA3 support to ViT MHA layer (vllm-project#12435)

  Show description for 2a0309a
  

  ywang96

  and

  Isotr0py

  authored
  2a0309a

  Copy full SHA for 2a0309a

Commits on Jan 25, 2025

• [TPU][CI] Update torchxla version in requirement-tpu.txt (vllm-project#12422)

  Show description for 324960a
  lsy323

  authored
  324960a

  Copy full SHA for 324960a

• [Misc] Add FA2 support to ViT MHA layer (vllm-project#12355)

  Show description for f1fc051
  Isotr0py

  authored
  f1fc051

  Copy full SHA for f1fc051

• [ROCm][MoE] MI300 tuned configs Mixtral-8x(7B,22B) | fp16, fp8 (vllm-project#12408)

  Show description for bf21481
  divakar-amd

  authored
  bf21481

  Copy full SHA for bf21481

• [Bugfix] Fix BLIP-2 processing (vllm-project#12412)

  Show description for fb30ee9
  DarkLight1337

  authored
  fb30ee9

  Copy full SHA for fb30ee9

• [Bugfix][Kernel] Fix moe align block issue for mixtral (vllm-project#12413)

  ElizaWszola

  authored
  221d388

  Copy full SHA for 221d388

Commits on Jan 24, 2025

• [Bugfix][Kernel] FA3 Fix - RuntimeError: This flash attention build only supports pack_gqa (for build size reasons). (vllm-project#12405)

  Show description for 3132a93
  LucasWilkinson

  authored
  3132a93

  Copy full SHA for 3132a93

• [Misc] Remove deprecated code (vllm-project#12383)

  Show description for df5dafa
  DarkLight1337

  authored
  df5dafa

  Copy full SHA for df5dafa

• [Bugfix][Kernel] Fix CUDA 11.8 being broken by FA3 build (vllm-project#12375)

  Show description for ab5bbf5
  LucasWilkinson

  authored
  ab5bbf5

  Copy full SHA for ab5bbf5

• [Misc] Enable proxy support in benchmark script (vllm-project#12356)

  Show description for 3bb8e2c
  jsato8094

  authored
  3bb8e2c

  Copy full SHA for 3bb8e2c

• [ci/build] sync default value for wheel size (vllm-project#12398)

  Show description for e784c6b
  youkaichao

  authored
  e784c6b

  Copy full SHA for e784c6b

• [Hardware][Gaudi][Doc] Add missing step in setup instructions (vllm-project#12382)

  MohitIntel

  authored
  9a0f3bd

  Copy full SHA for 9a0f3bd

Pagination

Previous

Next