Commits vllm-project/vllm

Helion KDA backend port from sglang https://github.com/sgl-project/sglang/pull/32593

alexm-redhat committed 24 days ago

7de0d6ea

[Bugfix][MoE] Tune FlashInfer experts to scheduler token limit (#52989)

mgoin committed 25 days ago

Verified bfb6c134

[DSV4][Kernel] Fuse shared experts into MegaMoE (#53040)

gcanlin committed 25 days ago

Verified 4f6885ff

[CI/Build][ROCm] Keep the CUDA-only kernel tests out of the ROCm run (#53113)

stefankoncarevic committed 25 days ago

Verified 3b829cf1

[Bugfix] Support MistralCommonBackend tokenizers in structured output (#52720)

thanhpt1110 committed 25 days ago

Verified d56bbf39

[Bugfix][MiniMax-M3] Keep FP8 query allocation stable across CUDA graph replay (#51203)

kyleliang-nv committed 25 days ago

Verified 7c8b68b9

[Misc] Don't allow language-model-only used with encoder CG together (#53127)

Isotr0py committed 25 days ago

Verified 00f7f258

upgrade tpu-inference to v0.27.0 (#53088)

meiyeh123 committed 25 days ago

Verified ae256289

Reduce `AutoWeightsLoader` kwargs (#53106)

hmellor committed 25 days ago

Verified 1fe3a157

Fix Transformers modelling backend `RMSNormFuser.fuse` performance (#52766)

hmellor committed 25 days ago

Verified 2dd17225

[Feature][Model Runner V2] Support extract_hidden_states speculation (#49811)

zupengwang committed 25 days ago

Verified 01af92e1

[Core][Multimodal] Skip redundant placeholder scan when token match succeeds (#52925)

Yiqin-17 committed 25 days ago

Verified cb09dd74

[Kernel] Add FlashInfer TRTLLM MXFP8 linear backend (#52204)

seonjinn committed 25 days ago

Verified bd8865a2

[Bugfix] Skip MM processor cache inserts larger than capacity (#53016)

Prudhvivuda committed 25 days ago

Verified de216b6e

[Docker] Update to nixl-1.3.2 (#51777)

sandeep-maddipatla committed 25 days ago

Verified 4b7cb949

[Model] Add tower and connector LoRA support for LFM2-VL (#51498)

zupengwang committed 25 days ago

Verified df137690

[Docs] Use incremental builds for C++ changes in `AGENTS.md` (#53098)

gau-nernst committed 25 days ago

Verified 6259572b

[Bugfix][GDN] Reset speculative decode count for an empty draft schedule (#53077)

khluu committed 25 days ago

Verified 6df7adc1

Fix weight tying (#51665)

hmellor committed 25 days ago

Verified f0c14b4f

[3/N] Harden Transformers modelling backend multi-modal path (#51827)

hmellor committed 25 days ago

Verified 6e85feb1

[CI] replace shellcheck script with shellcheck-py hook (#52572)

wjabbour committed 25 days ago

Verified 1eab6fef

[BugFix][Mooncake] Fix Mooncake saves from sparse Mamba block tables (#51362)

ZeldaHuang committed 25 days ago

Verified 5b1e7a81

[Rust Frontend] Fix Qwen parser auto-detection (#51169)

sagearc committed 25 days ago

Verified 727274a7

[Bugfix] Record non-ImportError attention backend probe failures instead of crashing engine init (#51703)

Eoin-Houstoun committed 25 days ago

Verified 30e2394c

[ROCm] Fix DeepSeek V4 indexer numerics and coverage (#50803)

AndreasKaratzas committed 25 days ago

Verified 6b68db44

[XPU][Tests] Make tests device-agnostic (#51968)

pmanczak committed 25 days ago

Verified 44cf3f04

[Bugfix] Return HTTP 400 instead of 501 for unknown chat roles in DeepSeek encoders (#53071)

JC-ut0 committed 25 days ago

Verified 38e9cefd

[Kernel][Kimi] fused vision q/k roper kernel (#50400)

lengrongfu committed 25 days ago

Verified c8de5199

[CI] Fix nonexistent dependency for data-parallel example test selection (#53026)

taneem-ibrahim committed 25 days ago

Verified 5d4d4704

[Bugfix][Security] Guard _load_ov2_processor with resolve_trust_remote_code (#52952)

jperezdealgaba committed 25 days ago

Verified c0a25c08