vllm-project/vllm

Pull Requests Commits

Copilot committed 4 days ago

Verified 22e847ca

[Bugfix][Parser] Honor a prompt-closed reasoning block in non-streaming parse and reasoning token counts (#61857)

albertoperdomo2 committed 4 days ago

Verified ae8f4123

[CI] Fix B200 FusedMoE Layer Kernels hang on NVSHMEM IB setup (#61695)

mgoin committed 4 days ago

Verified cffc0cfa

[CI][AMD] Move 17 test groups from MI300 to MI355 and test large DeepSeek/Nemotron evals on AITER (#61955)

AndreasKaratzas committed 4 days ago

Verified 706791d5

[Bugfix][Model] Honor quantization config in Inkling MTP (#61921)

soyr-redhat committed 4 days ago

Verified 76081b9e

Add @bnellnm to CODEOWNERS (#62025)

bnellnm committed 4 days ago

Verified 1a5b1870

[Refactor] Remove dead code multiple places (#61843)

yewentao256 committed 4 days ago

Verified 8e25bbca

[Bugfix][ROCm][GLM-5.2] Stop forcing skip_final_all_reduce on AMD DSA model path (#58589)

avininjamay8 committed 4 days ago

Verified 8a6c0449

[Perf] Share Triton JIT warmup compilation pool across registered kernels (#61422)

ayush1399 committed 4 days ago

Verified de7eeaa8

[Bugfix][Frontend] Keep the parser's tool call IDs in batch derender (#61852)

hickeyma committed 4 days ago

Verified 64175d38

[Bugfix][Frontend] Honor the constructor chat template in offline scoring (#61990)

mottopanikeiku committed 4 days ago

Verified 4ab99ff1

[Frontend] Resolve logprobs on the parsed streaming derender path (#60437)

shimib committed 4 days ago

Verified 024c5188

[Bugfix][Mooncake] Fix MultiConnector stats and proxy SSE responses (#61218)

whx-sjtu committed 4 days ago

Verified 9a16d43e

[XPU][Linear][MXFP4] enable torch as mxfp4 linear backend on xpu (#51965)

zufangzhu committed 4 days ago

Verified 07f3857f

[Feature][PCP] Add DeepSeek-V4.1-Flash support (#59857)

wangyicong52 committed 4 days ago

Verified fa75d345

[CI][Intel] Skip ROCm-only test_topk_gating_enabled_by_cdna on XPU (#61932)

chaojun-zhang committed 4 days ago

Verified 074c83a0

[Dependency] Upgrade TokenSpeed MLA to 2.0.3 (#61689)

wzhao18 committed 4 days ago

Verified e3657929

[Bugfix] Honor code_revision in config-based tokenizer loading (#61611)

dskwe committed 4 days ago

Verified 49cb2bbf

[Bugfix][Structured Output] Mask grammar-invalid implicit reasoning ends (#61574)

vMaroon committed 4 days ago

Verified 4ac1f8d2

[Bugfix] Honor OPT token embedding width for prompt embeddings (#61736)

maithilijoshi20 committed 4 days ago

Verified 92dc3dd3

[ROCm][Perf] Any-shape low-latency MoE router GEMM on MI355 (#61895)

AndreasKaratzas committed 4 days ago

Verified b466fbac

[Perf] Kimi K3 Specialized Decode: CuTeDSL Optimized GB300 (1.5-2x) (#61867)

benchislett committed 4 days ago

Verified 642c18d6

[XPU][Docker] Bump nixl to 1.5.0 (#60356)

sandeep-maddipatla committed 4 days ago

Verified 3ddc6702

[Perf][DSv4.1] Split fused_wqa_wkv by token across TP ranks on SM90 (#58872)

ShuoleiWang committed 4 days ago

Verified 05ad0a8a

[Feature] Add External LB Mode Support for Elastic EP Scaling (#43202)

fangyuchu committed 4 days ago

Verified dd9e07a2

[Bugfix][Kernel] silu_and_mul_quant: run the vectorized loop (#61716)

Monishver11 committed 4 days ago

Verified fb3fec97

[ROCm][Qwen4Exp] Support HC and MoE sequence parallelism on ROCm (#61471)

AndreasKaratzas committed 4 days ago

Verified 29345f42

[Bugfix][MoE] Keep the gate's hash_indices_table out of RoutedExperts parameters (#61930)

khluu committed 4 days ago

Verified a956a812

Update DeepGEMM commit (#61744)

jeejeelee committed 4 days ago

Verified bee95149

[Bugfix] remove duplicate hash-routing capability method (#61924)

tessapham committed 4 days ago

Verified 8f93c648

Older