Pull Requests vllm-project/vllm

implement issue #20711

#20879 opened 2025-07-13 11:51 by kiscad

[Refactor][V1] Move outlines utils for V1 imports structured-output ready v1

#20878 opened 2025-07-13 11:51 by aarnphm

[Benchmark] Add expert parallel support to MoE benchmark performance

#20876 opened 2025-07-13 09:12 by Chen-zexi

[Frontend] OpenAI Responses API supports Tool/Function calling frontend v1 tool-calling llama deepseek

#20874 opened 2025-07-13 08:00 by chaunceyjiang

[Bugfix] Fix: Fix multi loras with tp >=2 and LRU cache

#20873 opened 2025-07-13 07:38 by charent

[Feature][EPLB] Add EPLB support for OLMoE

#20872 opened 2025-07-13 06:59 by ztang2370

[EPLB] Add EPLB support for dots1

#20870 opened 2025-07-13 06:32 by wenchen76

Allow serving Llama4ForCausalLM directly new-model llama

#20868 opened 2025-07-13 05:09 by sarckk

nit llama qwen

#20864 opened 2025-07-12 22:30 by py-andy-c

Fused moe tuning ep performance

#20863 opened 2025-07-12 22:04 by robertgshaw2-redhat

[Feature] limit thinking tokens frontend v1 deepseek

#20859 opened 2025-07-12 09:22 by llsj14

[Misc] Relax translations tests

#20856 opened 2025-07-12 07:49 by NickLucche

[Nvidia] Integrate cudnn prefill paged attention kernel for head_dim == 128 models, like Llama family needs-rebase v1 llama

#20850 opened 2025-07-12 00:19 by elfiegg

[WIP] Enable xpu sleep mode v1

#20848 opened 2025-07-11 23:09 by yangw1234

[Docs] Update supported models documentation with missing models documentation

#20844 opened 2025-07-11 22:07 by luccafong

[V1] [Hybrid] Refactor mamba state shape calculation; enable V1 via cli v1

#20840 opened 2025-07-11 20:04 by tdoublep

[Model] Replace Mamba2 RMSNorm Gated with Fused Triton Kernel

#20839 opened 2025-07-11 19:38 by cyang49

[BugFix] fix two issues: using metadata for causal-conv1d and init_states in v0

#20838 opened 2025-07-11 19:30 by thoangtrvn

[Frontend] Add chunked processing to handle long inputs in embedding models documentation frontend

#20837 opened 2025-07-11 18:58 by x22x22

[compile][startup] Disable C++ compilation of symbolic shapes

#20836 opened 2025-07-11 18:49 by anijain2305

Add DeepSeek V2/V3 model family to PP tests deepseek

#20831 opened 2025-07-11 16:59 by eicherseiji

Switching attention backend to correctly complete LoRA and Quantized Model Tests rocm ci/build

#20829 opened 2025-07-11 16:55 by Alexei-V-Ivanov-AMD

[Bugfix] Fix Qwen2 audio chat template for old version transformers compatibility frontend qwen

#20826 opened 2025-07-11 16:05 by Isotr0py

[Bugfix] Fix the bug in Hermes streaming parsing frontend ci/build tool-calling

#20824 opened 2025-07-11 15:39 by pre-master

[Model] Add ToolParser for Hunyuan A13B. documentation performance frontend tool-calling

#20820 opened 2025-07-11 14:05 by kzjeef

[Feature][EPLB] Add eplb support for Qwen3 qwen

#20815 opened 2025-07-11 12:11 by aladerran

[v1][core]Support for attention free models v1

#20811 opened 2025-07-11 10:55 by christian-pinto

[Model] Add support for Jina Embeddings V4 documentation performance new-model multi-modality

#20802 opened 2025-07-11 07:01 by sigridjineth

[FIX] bump mistral common to support devstral 2507 ci/build

#20801 opened 2025-07-11 06:28 by ptS10011

[Feature] Add command tool parser for Command-A model frontend tool-calling

#20800 opened 2025-07-11 06:19 by gjgjos