biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity
11,727 Commits 2 Branches 140 Tags
e5a621b724e5570aaffc4bbf9c5f6ec9bca63333
Commit Graph

1522 Commits

Author SHA1 Message Date
Matthew Bonanni
fc1d8be3dc [Attention] Update attention imports (#29540) 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
 2025-11-27 11:19:09 -05:00
Jinzhen Lin
a67dec7cba [Bugfix] fix IMA issue in certain cases of the moe marlin kernel (#28619) 
Signed-off-by: Jinzhen Lin <jinzhen.ljz@antgroup.com>
Co-authored-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
 2025-11-26 19:02:21 -08:00
HDCharles
df01eda4dc [Bugfix] Make compressed-tensors MoEs respect ignored layers (#28878) 
Signed-off-by: HDCharles <charlesdavidhernandez@gmail.com>
 2025-11-26 21:35:13 -05:00
Matthew Bonanni
430dd4d9eb [Attention] Remove imports from vllm/attention/__init__.py (#29342) 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
 2025-11-26 10:53:15 -07:00
HDCharles
e603129505 [refactor] CTConfig methods to static/class methods (#28870) 
Signed-off-by: HDCharles <charlesdavidhernandez@gmail.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
 2025-11-26 17:21:58 +00:00
Wentao Ye
0b0aa874e8 [Perf] Optimize batch invariant BMM, 18.1% Throughput improvement, 10.7% TTFT improvement (#29345) 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Signed-off-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
 2025-11-26 09:38:52 -07:00
Huamin Li
70d5953f82 Revert "[Bugfix] Fix GPT-OSS AR+NORM fusion (#28841)" (#29483) 
Signed-off-by: Huamin Li <3ericli@gmail.com>
 2025-11-26 22:27:26 +08:00
Xin Yang
53d7f1f601 [Kernel] Use pre-allocated output buffer for triton kernel fused_experts (#29219) 
Signed-off-by: Xin Yang <xyangx@amazon.com>
 2025-11-26 10:21:00 +08:00
Michael Goin
7df0289782 Change warning logs to debug for unimplemented MXFP4 Linear/Attention (#29441) 
Signed-off-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
 2025-11-25 22:52:31 +00:00
Michael Goin
e502098643 [Kernel] Add NVFP4 MoE CUTLASS support for SM120 (#29242) 
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
 2025-11-25 06:59:07 -08:00
elvischenv
6330f9477d [Bugfix] Fix GPT-OSS AR+NORM fusion (#28841) 
Signed-off-by: elvischenv <219235043+elvischenv@users.noreply.github.com>
 2025-11-25 07:59:40 +00:00
Fadi Arafeh
98caeadd54 [fix][cpu] Use a SwigluOAI impl which supports interleaved gate-up wei (#29273) 
Signed-off-by: Fadi Arafeh <fadi.arafeh@arm.com>
 2025-11-25 15:11:11 +08:00
Isotr0py
92effb07a4 [Model] Add HunyuanOCR support (#29327) 
Signed-off-by: manayang <jackmanayang@gmail.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: sergeywang <sergeywang@tencent.com>
Co-authored-by: manayang <jackmanayang@gmail.com>
Co-authored-by: manayang <manayang@tencent.com>
Co-authored-by: Roger Wang <hey@rogerw.io>
 2025-11-25 03:28:51 +00:00
Michael Goin
6f1355a1b7 [Perf] Disable DeepGEMM MoE by default when TP=8 is used (#29346) 
Signed-off-by: mgoin <mgoin64@gmail.com>
 2025-11-24 19:01:40 -07:00
Wentao Ye
699bca76c0 [UX] Raise error for attn backend of batch invariant (#29348) 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
 2025-11-24 17:49:01 -07:00
Michael Goin
c17610e2ba [Bugfix] Only use triton_kernels for MXFP4 on SM90 and SM100 (#29339) 
Signed-off-by: mgoin <mgoin64@gmail.com>
 2025-11-24 18:22:46 -05:00
bnellnm
8f066146c3 [MoE][Refactor] Make select_experts a non-static method (#29067) 
Signed-off-by: Bill Nell <bnell@redhat.com>
 2025-11-24 13:38:04 -05:00
jiahanc
5f96c00c55 [Fix] Add SM check to flashinfer MOE backend (#29144) 
Signed-off-by: jiahanc <173873397+jiahanc@users.noreply.github.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
 2025-11-23 00:39:30 +00:00
Federico
f55c76c2b3 chore: add RTX_PRO_6000 GLM4.6-FP8 kernel tuning (#29240) 2025-11-22 08:42:48 -08:00
Bram Wasti
5f7209a793 [tiny] Remove unsupported TRITON_MLA backend from batch invariance (#28832) 
Signed-off-by: Bram Wasti <bwasti@meta.com>
Signed-off-by: Bram Wasti <bwasti@fb.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
 2025-11-22 21:00:50 +08:00
jinghanhu
988ee66b0d Handle triton kernel import exception (#29062) 2025-11-22 10:07:50 +00:00
FlintyLemming
052950e5b3 Add fused MoE config for H200 E160 N192 fp8 (#29182) 
Signed-off-by: FlintyLemming <admin@flinty.moe>
 2025-11-21 17:37:51 -08:00
Lukas Geiger
d045e22dfe [Model][Qwen3VL] Tune Triton w8a8 block fp8 kernel for L40s (#29217) 
Signed-off-by: Lukas Geiger <lukas.geiger94@gmail.com>
 2025-11-21 17:30:55 -08:00
Varun Sundar Rabindranath
3137991f55 [BugFix] EPLB + B200 + DeepGEMM : Handle column-major scales tensor (#29162) 
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
 2025-11-21 14:28:17 -08:00
Lucas Wilkinson
1840c5cb18 [BugFix] Make sure to allocate worst case MoE workspace during profile run in the DP + EP case (#27426) 
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
 2025-11-21 11:41:52 -08:00
Mingyuan Ma
b4c8fbaae2 Add TRTLLM MoE NVFP4 kernel to CompressedTensorsW4A4MoeMethod (#28892) 
Signed-off-by: mingyuanm <mingyuanm@nvidia.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
 2025-11-21 09:54:11 -07:00
rasmith
e99e467384 [CI/Build][Kernel][AMD] Move extra dim to after load in _fwd_kv_parallel in lighting_attn.py (#29132) 
Signed-off-by: Randall Smith <ransmith@amd.com>
Co-authored-by: Randall Smith <ransmith@amd.com>
 2025-11-21 11:53:09 -05:00
Wentao Ye
a42ab317ac [Log] Optimize startup log (#28948) 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Signed-off-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: Nick Hill <nhill@redhat.com>
 2025-11-21 08:46:20 -08:00
Aleksandr Malyshev
b7f1f490a6 Upstream triton fp4 weight preshuffle (#28888) 
Signed-off-by: Aleksandr Malyshev <maleksan@amd.com>
Co-authored-by: Aleksandr Malyshev <maleksan@amd.com>
 2025-11-21 11:34:46 -05:00
Cyrus Leung
aab0102a26 [V0 deprecation] Remove more V0 references (#29088) 
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
 2025-11-21 11:56:59 +00:00
Hongxia Yang
3f5f36da3f [ROCm] Fix for import when building with upstream triton for gfx1100 for gpt-oss serving (#29127) 
Signed-off-by: Hongxia Yang <hongxia.yang@amd.com>
 2025-11-21 03:30:07 +00:00
Wentao Ye
e1eefa4c40 [Bug] Fix torch warning of tf32 usage (#29112) 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
 2025-11-21 01:54:59 +00:00
Wentao Ye
df44df0143 [Feature] Shared Experts Overlap with FI deepgemm swap kernel, 2.2% throughput improvement and 3.6% TTFT improvement (#28879) 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
 2025-11-20 18:41:49 -07:00
Anna Shors
6eb745d9bd Add truncate arg to yarn to match openai implementation of gpt-oss (#28244) 
Signed-off-by: ashors1 <ashors@nvidia.com>
Co-authored-by: Chen Zhang <zhangch99@outlook.com>
 2025-11-20 18:53:50 +08:00
Wentao Ye
2c52c7fd9a [Bug] Fix torch dynamo warning Dynamo detected a call to a functools.lru_cache (#29038) 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
 2025-11-20 16:52:23 +08:00
Shengliang Xu
a8c536829c Consolidate Nvidia ModelOpt quant config handling for all quantization methods (#28076) 
Signed-off-by: Shengliang Xu <shengliangx@nvidia.com>
 2025-11-19 22:39:36 -05:00
Wentao Ye
5031cd5d55 [Refactor] Optimize select_experts (#28069) 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
 2025-11-19 18:53:15 -05:00
JartX
8e38e99829 [Feature] EPLB on Qwen3VLMoe and CompressedTensorsWNA16MoEMethod (#28849) 2025-11-19 18:30:08 -05:00
Max Hu
cb0a7b4bea [Bugfix] Move flashinfer kernel check into ``__init__` function of `FusedMoE`` (#29018) 
Signed-off-by: Max Hu <hyoung2991@gmail.com>
 2025-11-19 21:54:15 +00:00
Yongye Zhu
88f5b19f0b [DeepSeek] Fix DeepSeek V3.2 Rope Embedding (#28968) 
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
 2025-11-19 16:30:04 -05:00
Shu Wang
613abb50d5 [MoE] Nvfp4 Masked Gemm: Add flashinfer grouped_gemm_nt_masked (#25990) 
Signed-off-by: Shu Wang. <shuw@nvidia.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
 2025-11-19 13:29:06 -08:00
Wentao Ye
1607e664f0 [Bug] Fix Batch Invariant MLA test (#28967) 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
 2025-11-19 21:18:32 +00:00
Qiu
2fd893b4ce [Feature] Prefill Context Parallel (PCP) basic support (#28718) 
Signed-off-by: QiuChunshuo <qiuchunshuo@huawei.com>
Signed-off-by: FENP <yuanyongjie.yyj@antgroup.com>
Signed-off-by: LookAround <lixushi@huawei.com>
Signed-off-by: Jingchun Gao <gaojingchun1@huawei.com>
Signed-off-by: zhenwenqi2024 <zhenwenqi_2022@qq.com>
Co-authored-by: FENP <yuanyongjie.yyj@antgroup.com>
Co-authored-by: LookAround <lixushi@huawei.com>
Co-authored-by: Jingchun Gao <gaojingchun1@huawei.com>
Co-authored-by: zhenwenqi2024 <zhenwenqi_2022@qq.com>
Co-authored-by: Jingchun Gao <63247409+gjc0824@users.noreply.github.com>
 2025-11-19 15:52:44 -05:00
杰兮
9d2d561257 [Bugfix] Fix precision corruption when shared_experts_stream=None (#28942) 
Signed-off-by: zhyajie <yajizhan@amd.com>
Co-authored-by: zhyajie <yajizhan@amd.com>
 2025-11-19 19:30:57 +00:00
Robert Shaw
fe69f331f8 [Kernels] Improve H200 Fused MoE Config (#28992) 
Signed-off-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
 2025-11-19 19:23:54 +00:00
Harry Mellor
a8b70304d6 Update rope_scaling to rope_parameters in preparation for Transformers v5 (#28542) 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
 2025-11-19 09:06:36 -08:00
Shanshan Shen
d44e9df7d4 [Model][Mamba] Add selector for mamba attention backend and make it pluggable for other device (#26487) 
Signed-off-by: shen-shanshan <467638484@qq.com>
 2025-11-19 16:24:55 +00:00
Chen Bruce
da2f6800e0 [Feat][Perf] Enable deepep-low-latency with round-robin expert placement. (#28449) 
Signed-off-by: bruceszchen <bruceszchen@tencent.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
 2025-11-19 13:46:24 +01:00
Xin Yang
468a8d72ba [Bugfix] Fix FusedMoEModularKernel for triton backend (#28913) 
Signed-off-by: Xin Yang <xyangx@amazon.com>
 2025-11-19 13:05:22 +08:00
Li, Jiang
20852c8f4c [CPU] Refactor CPU WNA16 (#28826) 
Signed-off-by: jiang1.li <jiang1.li@intel.com>
 2025-11-19 10:32:00 +08:00