vllm-project/vllm

Pull Requests Commits

quick fix for 37665 (#37722)

xuechendi committed 53 days ago

Verified 583f6071

NickLucche committed 54 days ago

fbfdf1d9

[Model Runner V2] Fix draft logits not populated during cudagraph replay (#37639)

TheEpicDolphin committed 54 days ago

Verified dcee9be9

[CI] Removing deprecated rlhf examples reference (#37585)

AndreasKaratzas committed 54 days ago

Verified bd8c4c07

[Bug] Fix FlashInfer allreduce fusion workspace uninitialized error (#37461)

wzhao18 committed 54 days ago

Verified 0140eafb

[XPU] bump vllm-xpu-kernels to v0.1.4 (#37641)

jikunshang committed 54 days ago

Verified bdf6a0a5

[PluggableLayer][MM] Add PluggableLayer for CustomQwen2Decoder (#37293)

Wangbei25 committed 54 days ago

Verified 0674d1fe

[Model] Refactor Step3-VL processor to HF style (#37579)

DarkLight1337 committed 54 days ago

Verified 30108fc8

[Refactor] Relocate entrypoint tests to match serving code structure (#37593)

sfeng33 committed 54 days ago

Verified e2d1c8b5

[XPU] Automatically detect target platform as XPU in build. (#37634)

ccrhx4 committed 54 days ago

Verified 6951fcd4

[Model Runner V2] fix draft attention metadata generation (#37364)

TheEpicDolphin committed 54 days ago

Verified 39474513

fix(xpu): Re-compute compile ranges after platform-specific config updates (#37523)

Liangyx2 committed 54 days ago

Verified 638a872d

[V0 Deprecation] Deprecate --disable-frontend-multiprocessing (#37612)

sfeng33 committed 54 days ago

Verified 9040151f

[Bugfix][LoRA] Fix Qwen35 LoRA (#36976)

jeejeelee committed 54 days ago

Verified 8fbe3f30

[compile][graph_partition]Add tensor size handling (#36038)

fxdawnn committed 54 days ago

Verified ea2c148f

[Feat] Enable CompressedTensorW4A8Int for XPU (#37207)

tianmu-li committed 54 days ago

Verified 47b7af0d

fix: disambiguate multimodal prefix cache keys (#36708)

tianshu-Michael-yu committed 54 days ago

Verified 269bf46d

[CI] Update mergify tool-calling label paths (#37478)

sfeng33 committed 54 days ago

Verified e5a77a50

Fix DP coordinator ZMQ port TOCTOU (#37452)

itayalroy committed 54 days ago

Verified ca1ac1a4

[ROCm][Bugfix] fix cache block size mismatch for aiter unified attention (#37606)

divakar-amd committed 54 days ago

Verified 4ca3fa6b

[Bugfix] Fix Deepseekv32 tool parser when stream interval > 1 (#36056)

sfeng33 committed 54 days ago

Verified be12afd2

[Bug] Fix EmbedIOprocessor "classify" <-> "embed" (#37573)

yewentao256 committed 54 days ago

Verified df3c0291

[Refactor] Remove dead code in pooling model (#37572)

yewentao256 committed 54 days ago

Verified 2be1a0f7

Fix AttributeError in Qwen3.5 GDN layers with quantized models (#37448)

jhsmith409 committed 54 days ago

Verified 4120a05f

[CI][BugFix][AMD] Don't set VLLM_ROCM_USE_AITER anymore in test_rocm_aiter_topk since its not necessary (#36996)

rasmith committed 54 days ago

Verified 98ff0429

[Performance] Enable Triton autotuning disk cache by default (#37188)

arpera committed 54 days ago

Verified b55156ea

test Qwen/Qwen3-4B-Instruct-2507 for unbacked (#36064)

laithsakka committed 54 days ago

Verified 112944fa

[MoE Refactor] Rename "naive" all2all backend (#36294)

bnellnm committed 54 days ago

Verified 91be5f9b

Comment fix for async rl example (#35244)

hao-aaron committed 54 days ago

Verified 4ee847e4

[ROCm][CI] Cleaning and restructuring amd-ci legacy pipeline (#34839)

AndreasKaratzas committed 54 days ago

Verified 040a505f

Older