vllm-project/vllm

Pull Requests Commits

[Core] Add explicit layer parallel plans

LucasWilkinson committed 96 days ago

6bf03e0d

[Model] Optimize Qwen3.5 on H20 (#48350)

zzt93 committed 97 days ago

Verified 2595d5ce

[ROCm][MiniMax-M3] Add AITER sparse paged attention (#47287)

tanpinsiang committed 97 days ago

Verified ee5a89f4

[Kernel] Implement CUDA kernel for ReLUSquaredActivation (relu^2) (#39058)

tanish-malekar committed 97 days ago

Verified e26264f3

[Bugfix][KV Offloading] Fix stale transfer_jobs after reset_cache + harden job completion (#48102)

Alex-ai-future committed 97 days ago

Verified 4c81772e

[CI][Rust Frontend] Pin cargo tool versions (#48222)

BugenZhao committed 97 days ago

Verified 27c3e579

[EC Connector] Add EC Transfer Params (#42433)

omerpaz95 committed 97 days ago

Verified 8df14cfc

[CI][2/N] reduce CI time (#48394)

ZJY0516 committed 97 days ago

Verified 370b678a

Make tiering offload region DP-replica aware (#47987)

liranschour committed 97 days ago

Verified 5c0c987c

[Bugfix] Guard mixed-dtype allreduce RMSNorm quant fusions (#48330)

hugo-cen committed 97 days ago

Verified 5f8e73cb

[Frontend] Add /abort_requests to the RLHF dev API router (#47173)

aoshen02 committed 98 days ago

Verified 83762b77

[Perf][Qwen] Replace MOE all-reduce with reduce-scatter (#47006)

gcanlin committed 98 days ago

Verified a02984ed

Runtime Draft Weight Update for Speculative Decoding (#46725)

vx120 committed 98 days ago

Verified fc1c5480

[2/N][Core] support partial prefix cache hit for hybrid model (#46384)

ZJY0516 committed 98 days ago

Verified 481e481b

[CI][CPU] Add Qwen2-VL multimodal tests for CPU backend and fix incompatibilities (#48072)

zhejiangxiaomai committed 98 days ago

Verified 8e981630

[Bugfix][LoRA] Support ark_linear base layer in _get_lora_device (#47690)

AlejandroParedesLT committed 98 days ago

Verified 9a48eef8

[CI] split tests to reduce CI time (#48219)

ZJY0516 committed 98 days ago

Verified 1ef1c7eb

fix(processor): route MiMo-V2-Omni media fetch through MediaConnector (#43117)

ibondarenko1 committed 98 days ago

Verified 54503ece

fix(entrypoints): stop resolve_items leaking in-flight media fetch tasks on partial failure (#48333)

ErenAta16 committed 98 days ago

Verified 00673115

[2/N][KV-Cache Layout Refactor] Pack K/V into the content dim across attention backends (#44455)

LucasWilkinson committed 98 days ago

Verified 51878e5b

[XPU][UT]Fix InternS1ProForConditionalGeneration AssertionError (#48232)

Yejing-Lai committed 98 days ago

Verified 76fedaa2

FP32 router GEMV optimization (#48335)

jeejeelee committed 98 days ago

Verified 19069bcb

[CI] Point CI at Transformers release rather than release branch (#48328)

hmellor committed 98 days ago

Verified 1bd8f80a

[XPU]remove is_xxx from moe class and bump up kernels (#48079)

mayuyuace committed 98 days ago

Verified 0b6636cb

Bump Transformers version to 5.13.0 (#47867)

hmellor committed 99 days ago

Verified 4a6440ac

[Revert] [Build] Update vllm ...builds FA3 with torch stable API (#48269)

LucasWilkinson committed 99 days ago

Verified bec0a4ed

[Logs] DP Supervisor Log Improvement (#48278)

robertgshaw2-redhat committed 99 days ago

Verified 3d99b049

[Misc] Use meta tensor for KV cache stride calculation (#47316)

LucasWilkinson committed 99 days ago

Verified 04d553f3

[BugFix] Fix packed HND KV cache reshape for FlashAttention (#47314)

LucasWilkinson committed 99 days ago

Verified 9c18e90f

[BugFix] weights processing peak memory reduction for nvfp4 MoE layers (#46276)

thisisjimmyfb committed 99 days ago

Verified 09238796

Older