vllm-project/vllm

Pull Requests Commits

[Bugfix] Add batched experts auto-upgrade for VLLM_FLASHINFER_MOE_BACKEND path

esmeetu committed 31 days ago

53a86e69

[Bugfix]: Fix MinimaxM2ToolParser missing tools parameter (#39683)

chaunceyjiang committed 33 days ago

Verified d2130a47

[LMCache][MP] optimize save when mla enabled (#38810)

chunxiaozheng committed 33 days ago

Verified c687bf22

[Model Runner V2] Add full cuda graph support for eagle prefill (#37588)

TheEpicDolphin committed 33 days ago

Verified ccf90ba7

ParakeetExtractor performance and UX enhancements (#39423)

netanel-haber committed 33 days ago

Verified 6adacfcb

[Refactor][Parser] Simplify parse_delta (#39728)

sfeng33 committed 33 days ago

Verified 14cb86c1

Bug/test eagle dp v0 (#38938)

Monishver11 committed 33 days ago

Verified 8213e8f8

[Bugfix][Pooling] Fix silent weight corruption with buffer-reusing iterators (#39650)

pedramr committed 33 days ago

Verified 3693f922

[Core][Metrics] expose waiting request breakdown via labeled metric (capacity/deferred) (#38435)

mukesh-hai committed 33 days ago

Verified 5c18b961

[Bugfix] Reject non-nvfp4 dtypes when using the flashinfer_nvlink_one_sided all2all backend (#39717)

tlrmchlsmth committed 33 days ago

Verified f72b2097

[Doc] Fix Python-only build 404 fallback guidance (#38052)

Yuyi-Ao committed 33 days ago

Verified 610a3efc

[Bugfix][Kernel][ROCm] Fix triton_w4a16 scales mismatch when BLOCK_K > group_size (#39705)

JartX committed 33 days ago

Verified f414f906

[Misc] Multi-turn benchmark output performance json (#39572)

NickLucche committed 33 days ago

Verified 8625ec26

[Bugfix] Use is_integrated to detect UMA GPUs for memory reporting (#35356)

haosdent committed 33 days ago

Verified 995e9a20

[Quantization] [Refactor] Create special "GptOssMxfp4MoeMethod" (#39604)

zyongye committed 33 days ago

Verified 739e5945

[Bugfix] Fix tensor shape mismatch in sparse attention with speculative decoding (#39542)

santiramos27 committed 33 days ago

Verified 4d042ed8

[CI][Metrics] Fix local_cache_hit assertion after prompt tokens metrics updates (#39709)

ZhanqiuHu committed 33 days ago

Verified 10d9872d

[Bug] Fix rocm sparse attn indexer issue (#39225)

yewentao256 committed 33 days ago

Verified ccd0d1d9

[Bugfix][CT] Fix KV cache scale handling (#39418)

yiliu30 committed 33 days ago

Verified d8ddb316

[Bugfix] stream failure when model name not in audio endpoints (#36679)

ekagra-ranjan committed 33 days ago

Verified 1ce0318c

[Bug] Fix TypeError when hf_config.architectures is None during model loading (#38849)

TihoElek committed 33 days ago

Verified 8d825b87

[MXFP8] [XPU] add a new compressed tensor schema and add a xpu mxfp8 gemm kernel (#38707)

zufangzhu committed 33 days ago

Verified 1b19bd75

[Bugfix] Fix Responses API instructions leaking through previous_response_id (#37727)

he-yufeng committed 33 days ago

Verified 200a727e

feat: add max_tokens_per_doc in rerank request. (#38827)

jefp committed 33 days ago

Verified edbc1abd

[Bugfix] Fix GLM tool parser streaming with MTP or stream interval (#39253)

sfeng33 committed 34 days ago

Verified 0e39202c

[XPU]Enhance environment collection for Intel XPU and optimize layout (#35698)

1643661061leo committed 34 days ago

Verified 9dd5ee01

feat: rename logit_bias/logit_scale to logit_mean/logit_sigma for affine score calibration (#39530)

jefp committed 34 days ago

Verified fa6ae311

fix(lmcache): correct store for cached requests and num_scheduled_tokens in lmcache_mp_connector.py (#39655)

maobaolong committed 34 days ago

Verified 2a3c32ce

fused qknorm+rope kernel optimization for SM9.0 (#37376)

EricccYang committed 34 days ago

Verified 4beeb068

[compile] Enable AOT compile with batch invariance mode. (#39201)

zhxchen17 committed 34 days ago

Verified cae98406

Older