Commits vllm-project/vllm

[ROCm][DSv4.1][Perf] Run the fused mHC seams as AITER's single-launch HIP Mega-MHC (#65992)

davzhuAMD committed 1 hour ago

Verified 9fbe012b

[Perf] Allow DeepGEMM/Triton block-FP8 MoE with FlashInfer NVLink all2all (#67592)

aoshen02 committed 2 hours ago

Verified 8657860c

[Core][MM] Make the multimodal encoder cache size configurable (#65534)

HangruiCao2000 committed 2 hours ago

Verified 261042ba

[CI][Test] Keep dummy FP8 weights within +-1 in the HiSparse e2e tests (#67753)

khluu committed 3 hours ago

Verified 1489434b

[KV Offload][Bugfix] Hold P2P e2e test ports until the tiers bind them (#67607)

AndreasKaratzas committed 3 hours ago

Verified ca8952a9

[CI] Require batch invariance for derender parity tests (#67742)

gty111 committed 5 hours ago

Verified 5039f736

[Bugfix][Muse Spark] Load the model classes lazily so importing the package doesn't initialize CUDA (#67691)

njhill committed 5 hours ago

Verified 6a587fed

[ROCm][PP] Stage hand-off on SDMA copy engines (--pp-sdma-handoff) (#65658)

benenzhu committed 6 hours ago

Verified 4244a09e

[CI] Fix NIXL EP import test (#67574)

itayalroy committed 6 hours ago

Verified 730194d2

[Bugfix][KV Offload] Report only written blocks from fs tier store jobs (#63524)

AdvayMonga committed 7 hours ago

Verified b575d74f

[Rust Frontend] Support n > 1 on /inference/v1/generate (#66878)

devshah-cohere committed 7 hours ago

Verified 85217ffe

[XPU] Flatten N-D input in XPU MXFP8 linear kernel (#66494)

Liangliang-Ma committed 9 hours ago

Verified 9bc16ae6

[CI] Load gte-Qwen2 without remote code in test_non_causal_models (#67716)

hmellor committed 9 hours ago

Verified a58360ca

[Bugfix][Audio] Treat zero decode limits as disabled (#67611)

Woolgathererer committed 10 hours ago

Verified 4388bec9

[Rust Frontend] Bump llm-multimodal version (#67670)

Isotr0py committed 10 hours ago

Verified 6c7e4804

[CI] Start multi_process_parallel's Ray cluster with one CPU per rank (#67391)

wzhao18 committed 12 hours ago

Verified 27a441ca

[Perf][ROCm][DSv4.1] DeepSeek-V4.1-Flash: dense MXFP8 GEMMs and wo_a on AITER's tuned FlyDSL kernels, and a faster MXFP8 activation quant (MI355X) (#67377)

limou102 committed 12 hours ago

Verified 9d8c9806

[Model Loader] Fill dummy FP8 weights with random codes across the range (#67693)

WoosukKwon committed 15 hours ago

Verified a16bfee1

[Core][DBO] Dual batch overlap under pure tensor parallelism (split PR of M3.1) (#63483)

benenzhu committed 18 hours ago

Verified 5e51c0ca

[CI][Test] Stub pipeline_relay in the engram DP-shard model forward test (#67636)

khluu committed 18 hours ago

Verified cb71f581

[Bugfix][Core] Keep dummy-run requests within max_model_len (#67545)

simon-veitner-redhat committed 19 hours ago

Verified 640c349c

[XPU][MLA] Register XPUFlashAttnMLABackend for FLASH_ATTN_MLA in XPU (#64255)

ZhouXiaoya25 committed 21 hours ago

Verified bb23c0a8

[Bugfix][MRV2] Allow non 8-bytes aligned `--max-num-batched-tokens` for DeepSeek-v4 (#66401)

wtdcode committed 22 hours ago

Verified d2f92859

[Transformers backend] Serve Qwen2 with the Transformers modeling backend (#66821)

hmellor committed 22 hours ago

Verified 0431c0f7

[CI] Bump Transformers version to 5.23.0 (#67502)

hmellor committed 22 hours ago

Verified 0edc2ea1

[MM] Support mm_device_do_normalize for LLaVA-OneVision (#67505)

xs1128 committed 22 hours ago

Verified 76a4fedf

[Chore] Update agent skill for vLLM (#67273)

yewentao256 committed 23 hours ago

Verified 57802683

[MM] Support mm_device_do_normalize for Qwen3-Omni (#67426)

gordan12 committed 23 hours ago

Verified 0e73a584

[CI][Test] Run test_mm_prefix_lm_pipeline_parallel with spawned workers (#67637)

khluu committed 1 day ago

Verified 865cf8c1

[CI][XPU] Bound DeepStack test profiling to avoid OOM (#67601)

hlin99 committed 1 day ago

Verified 78233b8f