vllm-project/vllm

Pull Requests Commits

[Feat] Iteration-level profiling for Torch and CUDA profiler (#28987)

benchislett committed 19 days ago

Verified fcbcba6c

[cpu][ci] Add initial set of tests for Arm CPUs (#28657)

fadara01 committed 19 days ago

Verified 3168285f

[AMD] Use Decoupled Kernel Block Size to Support AITER MLA block_size=1 (#27715)

zq1997 committed 19 days ago

Verified 3fb0d909

[DeepSeek + LMCache Multiprocess] handle MLA for deepseek model + LMCache Multiprocess connector (#29039)

KuntaiDu committed 19 days ago

Verified 05c2dee7

[torchao] fix safetensors for sharding (#28169)

liangel-02 committed 19 days ago

Verified 1d642872

[Misc] Colorize logs (#29017)

njhill committed 19 days ago

Verified 9ccef8e3

[GC Debugger] Simply and improve GC Debugger Utils (#29029)

Jialin committed 19 days ago

Verified 537cc635

[Refactor] Optimize `select_experts` (#28069)

yewentao256 committed 19 days ago

Verified 5031cd5d

[Performance] Reduce DeepGEMM N dim restriction from 128 to 64 multiplier (#28687)

alexm-redhat committed 19 days ago

Verified 3aaa94ac

[Feature] EPLB on Qwen3VLMoe and CompressedTensorsWNA16MoEMethod (#28849)

JartX committed 19 days ago

Verified 8e38e998

[CI] Fix precommit `rope_theta` issue (#29040)

yewentao256 committed 19 days ago

Verified 0075bfff

[Bugfix] Move flashinfer kernel check into ```__init__``` function of ```FusedMoE``` (#29018)

maxyanghu committed 19 days ago

Verified cb0a7b4b

[BugFix] Fix false assertion with spec-decode=[2,4,..] and TP>2 (#29036)

LucasWilkinson committed 19 days ago

Verified 8f4f77a7

[ROCm][CI] Fix Weight Loading With Multiple GPU Tests on ROCm (#28984)

micah-wil committed 19 days ago

Verified 22e44ad5

[DeepSeek] Fix DeepSeek V3.2 Rope Embedding (#28968)

zyongye committed 19 days ago

Verified 88f5b19f

[MoE] Nvfp4 Masked Gemm: Add flashinfer grouped_gemm_nt_masked (#25990)

wenscarl committed 19 days ago

Verified 613abb50

[BugFix] Ray with multiple nodes (#28873)

juliendenize committed 19 days ago

Verified cdeec2e6

[Bug] Fix Batch Invariant MLA test (#28967)

yewentao256 committed 19 days ago

Verified 1607e664

[CI/Build] Fix test_prefix_prefill for AMD (#28905)

rjrock committed 19 days ago

Verified 68d72319

[Feature] Prefill Context Parallel (PCP) basic support (#28718)

pisceskkk committed 19 days ago

Verified 2fd893b4

Eagle: MM Cuda Graphs with MRope (#28896)

IzzyPutterman committed 19 days ago

Verified 02f5903b

Upstreaming aiter triton attention backend as a new backend (#28701)

maleksan85 committed 19 days ago

Verified ac10fd3c

[Bugfix] Fix precision corruption when shared_experts_stream=None (#28942)

zhyajie committed 19 days ago

Verified 9d2d5612

[Kernels] Improve H200 Fused MoE Config (#28992)

robertgshaw2-redhat committed 19 days ago

Verified fe69f331

[Core] Reuse created spec tokens lists to mitigate GC cost (#28917)

Jialin committed 19 days ago

Verified 3319a493

Re-enable FlashInfer for Llama4 on Blackwell in e2e fusion tests (#28966)

Copilot committed 19 days ago

Verified 61728cd1

GLM-V video segmentation solution adjustment (#28941)

zRzRzRzRzRzRzR committed 19 days ago

Verified 0c80efd9

Update `rope_scaling` to `rope_parameters` in preparation for Transformers v5 (#28542)

hmellor committed 19 days ago

Verified a8b70304

[Model][Mamba] Add selector for mamba attention backend and make it pluggable for other device (#26487)

shen-shanshan committed 19 days ago

Verified d44e9df7

[BugFix] Fix async-scheduling + FlashAttn MLA (#28990)

LucasWilkinson committed 19 days ago

Verified 48fc8b1e

Newer Older