共 20 个 commit，涉及 176 个文件，+3645/-1112 行变动。

概要

统计项 数值
Commit 数 20
变更文件 176
新增行数 +3645
删除行数 -1112

Commit 列表

🧪 CI/Tests

  • 27c3e57 #48222 — [CI][Rust Frontend] Pin cargo tool versions (#48222)
    • 作者: Bugen Zhao | +35/-46 | 1 个文件

    This PR pins the tool versions installed in cargo CI for Rust frontend to enhance reproducibility, to prevent issues like #48170 from happening again. The versions here are extracted from the latest successful run. For tools installed by cargo binstall, use semantic version requirements so that patch updates are still applied. ## Test Result —

  • 370b678 #48394 — [CI][2/N] reduce CI time (#48394)
    • 作者: Jiangyun Zhu | +8/-4 | 2 个文件

    After https://github.com/vllm-project/vllm/pull/48219, the CI wall time is as follows(based on https://buildkite.com/vllm/ci/builds/77652/waterfall): This PR reduces CI wall time by parallelizing long-running kernel and attention tests - Kernels Core Operation：parallelism: 3 - 1:43:07 -> 40min - Kernels Helion：parallelism: 2 - 1:36:13 -> 36min - V1 attention H100：parallelism: 2 - **1:07:14…

  • 8e98163 #48072 — [CI][CPU] Add Qwen2-VL multimodal tests for CPU backend and fix incompatibilities (#48072)
    • 作者: zhao, zhenhui | +33/-7 | 2 个文件

    Support CPU backend for multimodal embedding reduction. Fixes two critical issues: 1. pin_memory incompatibility on CPU devices 2. Local function serialization failure in cross-process RPC Enables Qwen2-VL and Qwen2.5-VL multimodal inference on CPU-only vLLM deployments. bash VLLM_CI_ENV=0 pytest -x -v -s tests/models/multimodal/generation/test_qwen2_5_vl.py

  • 1ef1c7e #48219 — [CI] split tests to reduce CI time (#48219)
    • 作者: Jiangyun Zhu | +4/-3 | 2 个文件

    Split tests to reduce CI time ## Test Result —

  • 1bd8f80 #48328 — [CI] Point CI at Transformers release rather than release branch (#48328)
    • 作者: Harry Mellor | +8/-8 | 8 个文件

📦 Other

  • 8df14cf #42433 — [EC Connector] Add EC Transfer Params (#42433)
    • 作者: omerpaz95 | +416/-12 | 55 个文件

    Add ec_transfer_params as a top-level dict[str, Any] | None field to vLLM’s OpenAI-compatible protocol — both on requests (input routing) and responses (output) — mirroring the existing kv_transfer_params pattern exactly. This also wires ECConnectorBase.request_finished() into the scheduler’s _free_request(), so EC connectors can emit transfer params when a request completes. The field propagates …

  • 5c0c987 #47987 — Make tiering offload region DP-replica aware (#47987)
    • 作者: liranschour | +28/-18 | 4 个文件

    Each data-parallel replica is a separate engine with its own KV blocks, so its /dev/shm offload region must be distinct. Derive a per-replica instance id (instance_id_dp{index}) for both the scheduler-side and worker-side SharedOffloadRegion, mirroring the P2P tier’s base_port + data_parallel_index scheme. Also fold the global physical device index back into the replica-local [0, world_size) range…

  • fc1c548 #46725 — Runtime Draft Weight Update for Speculative Decoding (#46725)
    • 作者: vx120 | +129/-1 | 12 个文件

    Co-authored-by: crp0128 191679376@qq.com ## Background ### Target-model weight update today vLLM exposes three weight update paths for RL training loops: | Path | Entry points | Transport | |——|————-|———–| | Checkpoint / layerwise | start_weight_update → update_weights → finish_weight_update | IPC or NCCL | | Direct reload | reload_weights | disk / iterator | | Kernel-format |…

  • 481e481 #46384 — [2/N][Core] support partial prefix cache hit for hybrid model (#46384)
    • 作者: Jiangyun Zhu | +1651/-175 | 21 个文件

    This PR builds on the partial prefix-cache primitives from #45939 and completes fine-grained prefix-cache hits for aligned full-attention + Mamba hybrid models. The overall design is described in RFC #45702. The PR adds: - a public prefix_match_unit that selects prefix-match granularity independently from physical KV-cache block sizes; - fine-grained lookup in the full-attention and Mamba cache ma…

  • 54503ec #43117 — fix(processor): route MiMo-V2-Omni media fetch through MediaConnector (#43117)
    • 作者: Ievgen Bondarenko | +22/-76 | 1 个文件

    What this PR does Fixes SSRF + arbitrary local file read in MiMoV2OmniMultiModalProcessor by routing caller-supplied media strings through MediaConnector instead of raw requests.get / Image.open calls. vllm/transformers_utils/processors/mimo_v2_omni.py had two sinks: - _fetch_image() at line 231 - issued requests.get(src, timeout=30) directly on any http:// or https:// string, did Image.open(s…

  • 0067311 #48333 — fix(entrypoints): stop resolve_items leaking in-flight media fetch tasks on partial failure (#48333)
    • 作者: ErenAta16 | +52/-4 | 2 个文件

    AsyncMultiModalItemTracker.resolve_items fans out each modality’s media fetches with plain asyncio.gather. When one fetch in a modality fails (bad URL, timeout, an SSRF/domain-allowlist rejection, a decode error), gather propagates that exception immediately, but it does not cancel or await the sibling fetches for that same modality that are still in flight. Those tasks keep running detached from …

  • 51878e5 #44455 — [2/N][KV-Cache Layout Refactor] Pack K/V into the content dim across attention backends (#44455)
    • 作者: Lucas Wilkinson | +664/-604 | 31 个文件

    PR #42374 (first part of RFC #42082) has been split into 4 PRs: #44454 [1/N][KV-Cache Layout Refactor] Refactor DSV4 KV cache config -> #44455 [2/N][KV-Cache Layout Refactor] Pack K/V into the content dim across attention backends #44456 [3/N][KV-Cache Layout Refactor] Standardize Mamba cache; drop get_transfer_cache_regions #44458 [4/N][KV-Cache Layout Refactor] Standardize KV cache layout Packs …

  • 76fedaa #48232 — [XPU][UT]Fix InternS1ProForConditionalGeneration AssertionError (#48232)
    • 作者: Yejing Lai | +6/-0 | 1 个文件

    Fix UT tests/models/test_initialization.py::test_can_initialize_large_subset[InternS1ProForConditionalGeneration] meet AssertionError: 2 cannot be divided by 8 Root cause: InternS1Pro uses router_n_groups instead of n_group.

  • 19069bc #48335 — FP32 router GEMV optimization (#48335)
    • 作者: Jee Jee Li | +284/-67 | 4 个文件

    Test Result —

  • 0b6636c #48079 — [XPU]remove is_xxx from moe class and bump up kernels (#48079)
    • 作者: Qiming Zhang | +7/-37 | 3 个文件

    Align with https://github.com/vllm-project/vllm-xpu-kernels/pull/376 Tested below model, results is meaningful. model=Qwen/Qwen3-30B-A3B model=Qwen/Qwen2.5-Omni-7B-AWQ model=openai/gpt-oss-20b model=Qwen/Qwen3-30B-A3B-FP8 model=INCModel/Qwen3-30B-A3B-Instruct-2507-MXFP8-LLMC Bump up xpu kernels.

  • 4a6440a #47867 — Bump Transformers version to 5.13.0 (#47867)
    • 作者: Harry Mellor | +30/-24 | 12 个文件

    PRs needed for Transformers patch: - https://github.com/huggingface/transformers/pull/47148 - https://github.com/huggingface/transformers/pull/47174 - https://github.com/huggingface/transformers/pull/47245 — Errors resolved: - AttributeError: ‘str’ object has no attribute ‘module’ - https://github.com/huggingface/transformers/pull/46876 added if key.module.startswith(“transformers.”): wh…

🐛 Bug Fix

  • 5f8e73c #48330 — [Bugfix] Guard mixed-dtype allreduce RMSNorm quant fusions (#48330)
    • 作者: Hugo Centeno | +44/-3 | 2 个文件

    Fixes #48324. The residual FlashInfer allreduce + RMSNorm + static-quantization patterns could match graphs where the activation and RMSNorm weight have different dtypes. This occurs with Qwen/Gemma-style RMSNorm in nvidia/Qwen3.6-27B-NVFP4: the residual stream is BF16, while the effective RMSNorm weight is FP32 due to the weight.float() + 1.0 computation. Selecting the fused quantized operation f…

  • 9a48eef #47690 — [Bugfix][LoRA] Support ark_linear base layer in _get_lora_device (#47690)
    • 作者: Alejandro Paredes La Torre | +42/-0 | 2 个文件

    Fixes bug 1 from #47650: _get_lora_device doesn’t recognize the ark_linear base-layer shape produced by INCARKLinearMethod (INC WNA16 / AutoRound scheme), so –enable-lora crashes at model load with ValueError: Unsupported base layer for any AutoRound/INC ark-backed checkpoint. INCARKLinearMethod.process_weights_after_loading (vllm/model_executor/layers/quantization/inc/schemes/inc_wna16_linear.py…

🦀 Rust Frontend

  • 83762b7 #47173 — [Frontend] Add /abort_requests to the RLHF dev API router (#47173)
    • 作者: aoshen02 | +69/-12 | 8 个文件

    Add POST /abort_requests to the RLHF dev API router (vllm/entrypoints/serve/dev/rlhf/api_router.py). It calls EngineClient.abort(request_ids) — removing queued requests from the waiting queue and finish-aborting running ones — without pausing the scheduler. When request_ids is empty/missing, it aborts all in-flight requests (the dev RL server runs AsyncLLM). RL rollout frameworks need to drop …

⚡ Performance