vllm-project/vllm

Pull Requests Commits

Merge branch 'main' into glm5-router

jeejeelee committed 21 hours ago

Verified f9887a1c

[CI] Fix mypy for `vllm/v1/ops` (#39219)

yewentao256 committed 21 hours ago

Verified aec18492

Fix Responses JSON schema alias serialization (#38519)

noobHappylife committed 22 hours ago

Verified 2a49284c

[Model] Update ColModernVBERT to support latest HF checkpoint (#39307)

ieBoytsov committed 22 hours ago

Verified d37b3787

[Bug] Fix routing bias dtype for trtllm per-block fp8 moe (#38989)

wzhao18 committed 22 hours ago

Verified 92fbec39

[Bugfix] Fix cpu-offload-gb assertion with non-default block sizes (#36461)

AjAnubolu committed 22 hours ago

Verified 2f41d6c0

[Gemma4] Support quantized MoE (#39045)

dsikka committed 23 hours ago

Verified 3aecdf08

[UX] Integrate DeepGEMM into vLLM wheel via CMake (#37980)

mgoin committed 23 hours ago

Verified eb4205fe

[XPU][UT] update UTs in CI (#39296)

zhenwei-intel committed 23 hours ago

Verified 83aea214

[W8A8 Block Linear Refactor][2/N] Remove W8A8Fp8BlockLinearOp and adopt Fp8 block linear kernel selections. (#33892)

maralbahari committed 1 day ago

Verified 2e9034c9

[Bugfix] FlashInfer MXINT4 MoE crashes, missing do_finalize (#39315)

benchislett committed 1 day ago

Verified 8332078c

[torch.compile] Allow usage of Opaque Objects in PyTorch 2.11 (#39286)

zou3519 committed 1 day ago

Verified ba4a78eb

[Bugfix]Fix EP precision for Qwen3.5, Qwen3-Next (#39181)

USTCKAY committed 1 day ago

Verified f3c7941e

[CI Bug] Fix pre-commit issue in main (#39347)

yewentao256 committed 1 day ago

Verified 3352bf8b

[BugFix] --max-model-len=-1 causes over-limit requests to hang and starve the entire service (#39102)

triangleXIV committed 1 day ago

Verified 7c94ae16

`tests/v1/e2e/spec_decode`: assert async scheduling is used (#39206)

puririshi98 committed 1 day ago

Verified ad05edfb

[Feature] Batch invariant nvfp4 linear support (#39322)

yewentao256 committed 1 day ago

Verified 20181372

[MoE] Move DEEP_GEMM into experts/ subdirectory (#39005)

Jackmin801 committed 1 day ago

Verified a776a48b

[Tool] `adjust_request` to reasoning parser, and Gemma4 fixes (#39027)

bbrowning committed 1 day ago

Verified 8477fe42

[Attention] relax the head dim 512 and paged kv for sm90+FA4 (#38835)

IwakuraRein committed 1 day ago

Verified e24e0a43

[Perf][Kernel] Persistent TopK scheduler: unified CUDAGraph-safe kernel with dynamic per-row dispatch - DeepSeek-V3.2 DSA decode (#37421)

LopezCastroRoberto committed 1 day ago

Verified b55d830e

[Feature] NUMA binding support for GPU workers (#38635)

Harry-Chen committed 1 day ago

Verified 75e01a39

[kv_offload+HMA][5/N]: Track group block hashes and block IDs (#37109)

orozery committed 1 day ago

Verified 512c5eb4

[Bugfix] Fix Gemma4 streaming tool call corruption for split boolean/number values (#39114)

sfeng33 committed 1 day ago

Verified 13151a4d

[ROCm] Enable fused_silu_mul_block_quant on ROCm (#38817)

gshtras committed 1 day ago

Verified 56c976c1

[Core] Use tuple_return in split_module for tuple-conformant subgraphs (#38752)

frgossen committed 1 day ago

Verified d74a306c

[ROCm][CI-Build] Cherry pick triton BUFFER_OPS fix and update AITER (#38580)

gshtras committed 1 day ago

Verified 0e9f0a51

[Bugfix] Fix V1 logprobs empty strings for multi-byte UTF-8 tokens when logprobs > 0 (#34875)

haosdent committed 1 day ago

Verified 8904fc4d

[Bugfix] Add missing ASRDataset import and CLI args in benchmarks/throughput.py (#38114)

nemanjaudovic committed 1 day ago

Verified 1a2c1763

Merge branch 'main' into glm5-router

jeejeelee committed 1 day ago

Verified 36782451

Older