vllm-project/vllm

Pull Requests Commits

[Build] Download CUDA 12.9 nvcc/ptxas/cicc for better codegen in Docker build

tlrmchlsmth committed 140 days ago

c7b1080a

[Build] Downgrade default CUDA version to 12.8 for PyTorch cu128 compatibility

tlrmchlsmth committed 140 days ago

7ed39716

[CI] Remove unused precompiled wheel args from image build (#34767)

Amr Mahdi committed 141 days ago

Verified df3f537a

[Attention] Refactor `check_and_update_config` (#33600)

MatthewBonanni committed 141 days ago

Verified 77431529

[Feature] Decode Context Parallel support for GPU model runner v2 (#34179)

yewentao256 committed 141 days ago

Verified ab33d2a6

[Model Runner V2] Further simplification for PP (#34724)

WoosukKwon committed 141 days ago

Verified be3af2d2

[Kernel] Triton-based Top-k and Top-p sampler kernels (#33538)

cakeng committed 141 days ago

Verified c656ba3b

[Bugfix][MTP][Sparse MLA] Allow sparse MLA with MTP to run with FULL cudagraphs (#34457)

MatthewBonanni committed 141 days ago

Verified dc5fa77a

[CI] Fix flaky test_parsable_context (#34717)

sfeng33 committed 141 days ago

Verified 1e4a084c

[BugFix] Fix sp tests (#34716)

zou3519 committed 141 days ago

Verified 7967e854

Fixed whisper CPU test that does not spawn properly. (#34324)

almayne committed 141 days ago

Verified 6bd6d0c3

[CI][Nixl] Add CrossLayer KV layout tests (#34615)

NickLucche committed 141 days ago

Verified 8e962fef

[Renderer] Move InputPreprocessor into Renderer (2/2) (#34560)

DarkLight1337 committed 141 days ago

Verified 574fe752

[CI][BugFix] ShellCheck cleanup to remove baseline and preserve runtime behavior (#34514)

junuxyz committed 141 days ago

Verified c61a98f5

Fix docs build warning (#34686)

hmellor committed 141 days ago

Verified 28bffe94

[Bugfix] Fix 'remove_instance_endpoint' method logic in disagg_proxy_demo (#32922)

ChenqianCao committed 141 days ago

Verified ad65177a

Remove dead bitsandbytes CxB code from 8-bit inference path (#34633)

TimDettmers committed 141 days ago

Verified d44a5b6c

Revert "[Models] Fuse Qwen3.5 GDN's qkvz_proj and ba_proj" (#34683)

ZJY0516 committed 141 days ago

Verified 1d65283e

[Ray] Propagate third-party env vars to Ray workers via prefix matching (#34383)

kouroshHakha committed 141 days ago

Verified c464b573

[CI] Fix bake config artifact path for AMI rebuild pipeline (#34656)

Amr Mahdi committed 141 days ago

Verified c5c38e15

[Model Runner V2] Minor refactoring for penalties (#34662)

WoosukKwon committed 142 days ago

Verified d00df624

[Model Runner V2] Minor simplification for BadWordsState (#34669)

WoosukKwon committed 142 days ago

Verified 9752da9d

[Model Runner V2] Minor cleanup for PP (#34666)

WoosukKwon committed 142 days ago

Verified 04925b22

[Model Runner V2] Fix unintended CPU-GPU sync in make_dummy (#34667)

WoosukKwon committed 142 days ago

Verified d74278fb

[Bugfix] Fix fused MoE int32 overflow in stride*offset without perf regression (#34507)

haosdent committed 142 days ago

Verified b68fd899

[CI] Enable mypy import following for vllm/v1/kv_offload (#34639)

aneeshkp committed 142 days ago

Verified 0b5f9b72

[Core] Pipeline Parallel support for Model Runner V2 (#33960)

ZhanqiuHu committed 142 days ago

Verified 9a8853f7

[Model Runner V2] support bad_words sampling param (#33433)

izhuhaoran committed 142 days ago

Verified 387a1898

[NemotronH] Do not force router to run in fp32 (#34582)

roikoren755 committed 142 days ago

Verified 3b30e615

Targeting the MI355 agent pool with all existing tests (#34629)

Alexei-V-Ivanov-AMD committed 142 days ago

Verified 824f9e8f

Older