biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity
13,032 Commits 2 Branches 140 Tags
709502558c3efa658b2b49c7019f9dc67929764d
Commit Graph

1731 Commits

Author SHA1 Message Date
Matthew Bonanni
bcf2333cd6 [CI] Fix LM Eval Large Models (H100) (#32423) 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
 2026-01-16 00:52:49 +00:00
Michael Goin
83239ff19a Add thread_n=64 support to Marlin MoE (#32360) 
Signed-off-by: mgoin <mgoin64@gmail.com>
 2026-01-15 16:45:44 -08:00
TomerBN-Nvidia
c277fbdf31 [Feat] Support non-gated MoE with Marlin, NVFP4 CUTLASS, FP8, INT8, compressed-tensors (#32257) 
Signed-off-by: Tomer Natan <tbarnatan@computelab-frontend-8.nvidia.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: Tomer Natan <tbarnatan@computelab-frontend-8.nvidia.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
Co-authored-by: Tomer Natan <tbarnatan@ipp1-1429.ipp1a1.colossus.nvidia.com>
 2026-01-15 16:15:05 -08:00
Yongye Zhu
31c29257c8 [MoE Refactor][17/N] Apply Refactor to Bf16 (#31827) 
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
 2026-01-15 12:53:40 -08:00
Aleksandr Malyshev
8c11001ba2 [ROCM] DSfp4 mla projection gemms weight dynamic quantization (#32238) 
Signed-off-by: Aleksandr Malyshev <maleksan@amd.com>
Co-authored-by: Aleksandr Malyshev <maleksan@amd.com>
 2026-01-15 14:13:08 -06:00
Lucas Wilkinson
c36ba69bda [BugFix] Fix assert x_s.shape[-1] == x_q.shape[-1] // group_shape[1] in Blackwell Quantized MoE Test (#32362) 
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
 2026-01-15 10:19:12 -08:00
Dipika Sikka
361dfdc9d8 [Quant] Support MXFP4 W4A16 for compressed-tensors MoE models (#32285) 
Signed-off-by: Dipika Sikka <dipikasikka1@gmail.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
 2026-01-15 07:25:55 -08:00
Matthew Bonanni
8ebfacaa75 [Attention][MLA] Make FLASHINFER_MLA the default MLA backend on Blackwell, and TRTLLM the default prefill (#32339) 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
 2026-01-15 09:49:57 -05:00
brian033
b89275d018 [ROCm] Improve error handling while loading quantized model on gfx120… (#31715) 
Signed-off-by: brian033 <85883730+brian033@users.noreply.github.com>
Co-authored-by: TJian <tunjian.tan@embeddedllm.com>
 2026-01-15 04:16:00 -08:00
Lucas Wilkinson
2c9b4cf5bf [BugFix] Fix DeepSeek-V3.1 + DeepGEMM incompatible scale shapes (#32361) 
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: Eldar Kurtić <8884008+eldarkurtic@users.noreply.github.com>
 2026-01-15 06:32:22 +00:00
rasmith
3c2685645e [CI][AMD][Quantization][BugFix] Fix fp8 max in quant_utils.py and update test_fp8_quant.::test_static_fp8_quant_group_2d to use correct fp8 dtype and adjust atol/rtol (#32201) 
Signed-off-by: Randall Smith <ransmith@amd.com>
 2026-01-15 05:04:34 +00:00
Michael Goin
6388b50058 [Docs] Add docs about OOT Quantization Plugins (#32035) 
Signed-off-by: mgoin <mgoin64@gmail.com>
 2026-01-14 15:25:45 +08:00
Yi Liu
50632adc58 Consolidate Intel Quantization Toolkit Integration in vLLM (#31716) 
Signed-off-by: yiliu30 <yi4.liu@intel.com>
 2026-01-14 07:11:30 +00:00
Roberto L. Castro
8ef50d9a6b [Kernel][Performance] Enable smaller Scaling Factor tiling for NVFP4 small-batch decoding (#30885) 
Signed-off-by: LopezCastroRoberto <roberto.lopez.castro@udc.es>
Signed-off-by: Roberto L. Castro <38211239+LopezCastroRoberto@users.noreply.github.com>
Signed-off-by: LopezCastroRoberto <rocastro@redhat.com>
 2026-01-13 15:22:53 -08:00
Matthew Bonanni
2263d44b68 [4/N][Attention] Move MLA common to model_executor (#32060) 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
 2026-01-13 09:08:45 -08:00
Matthew Bonanni
98f60e5acb [6/N][Attention] Move utils to more appropriate locations (#32215) 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
 2026-01-13 05:38:52 -08:00
Mickaël Seznec
a5bbbd2f24 [Quantization] fix: overflow with static per-tensor scaling (#29867) 
Signed-off-by: Mickael Seznec <mickael@mistral.ai>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
 2026-01-13 12:56:01 +00:00
Cyrus Leung
232214b2ae [Bugfix] Replace PoolingParams.normalize with use_activation (#32243) 
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
 2026-01-13 10:45:42 +00:00
Andreas Karatzas
11b6af5280 [ROCm][Bugfix] Fix Mamba batched decode producing incorrect output (#32099) 
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
 2026-01-13 05:46:53 +00:00
xuebwang-amd
629584bfc9 [Kernel][MoE] fix computation order of MoE weight multiplication and improve flow (#31962) 
Signed-off-by: xuebwang-amd <xuebwang@amd.com>
 2026-01-12 17:17:30 -05:00
Lucas Kabela
ad8818bb5e [Misc][BE] Type coverage for vllm/compilation [3/3] (#31748) 
Signed-off-by: Lucas Kabela <lucaskabela@meta.com>
 2026-01-12 19:24:38 +00:00
Matthew Bonanni
20228cb851 [3/N][Attention] Move AttentionMetadata-related code from utils.py to backend.py (#32054) 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
 2026-01-12 09:13:56 -08:00
Cyrus Leung
8863c2b25c [Model] Standardize pooling heads (#32148) 
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
 2026-01-12 17:01:49 +00:00
danielafrimi
3f72639d36 [FIX] Add NO_MUL activation support for modular kernel path (#31528) 
Signed-off-by: dafrimi <dafrimi@nvidia.com>
Signed-off-by: <>
Co-authored-by: root <root@gpu-267.slurm-workers-slurm.slurm.svc.cluster.local>
Co-authored-by: root <root@gpu-537.slurm-workers-slurm.slurm.svc.cluster.local>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: root <root@pool0-01777.cm.cluster>
 2026-01-12 11:55:49 -05:00
Hongxin Xu
49e6b86c91 [Feature] Support recording expert indices for rollout router replay (#28284) 
Signed-off-by: xhx1022 <1737006628@qq.com>
Signed-off-by: Hongxin Xu <70438206+xhx1022@users.noreply.github.com>
Signed-off-by: arlenxu <arlenxu@tencent.com>
Co-authored-by: 22quinn <33176974+22quinn@users.noreply.github.com>
Co-authored-by: arlenxu <arlenxu@tencent.com>
 2026-01-12 06:23:04 -08:00
Matt
bde57ab2ed [Hardware][AMD][CI][Bugfix] Fix AMD Quantization test group (#31713) 
Signed-off-by: Matthew Wong <Matthew.Wong2@amd.com>
 2026-01-10 23:19:46 -08:00
Vensen
6ea001cfb7 [Bugfix][Quantization] Ensure input contiguity in per_token_quant_int8 (#31637) 
Signed-off-by: vensen <vensenmu@gmail.com>
 2026-01-10 12:40:02 -08:00
gnovack
d1fd802fa3 fused_moe_kernel - cast accumulator after applying router weights (#32002) 
Signed-off-by: gnovack <gnovack@amazon.com>
 2026-01-11 04:36:45 +08:00
Michael Goin
e6c6f2c79d [Quant] Support MXFP4 W4A16 for compressed-tensors dense models (#31926) 
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
 2026-01-10 06:44:35 -08:00
Cyrus Leung
583a90e005 [Refactor] Separate sequence and token pooling types (#32026) 
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
 2026-01-10 04:53:24 +00:00
maang
52d428295d [Core] Refactor ColumnParallelLinear: remove unused parameter and optimize forward (#31939) 
Signed-off-by: maang <maang_h@163.com>
 2026-01-10 04:19:49 +00:00
Lucas Kabela
ea6d067a2a [Misc][LLaMa4] Compile LLaMa Vision Encoder (#30709) 
Signed-off-by: Lucas Kabela <lucaskabela@meta.com>
 2026-01-09 22:01:38 -05:00
Matthew Bonanni
2612ba9285 [1/N][Attention] Restructure attention: move files (#31916) 
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
 2026-01-09 13:10:24 -08:00
Lucas Wilkinson
0a0aa07747 [Quant] Make static quant support all group shapes (#30833) 
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
 2026-01-09 12:49:27 -08:00
jiahanc
f9e2a75a1e [fix] add cutedsl to global sf (#32001) 
Signed-off-by: jiahanc <173873397+jiahanc@users.noreply.github.com>
 2026-01-09 12:03:02 -08:00
Runkai Tao
a4d5d663e2 Add unpermute-aware fused MoE path and small-batch fallback (#29354) 
Signed-off-by: Runkai Tao <rt572@physics.rutgers.edu>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
 2026-01-09 12:58:39 -07:00
Wentao Ye
308feab33f [Perf] Optimize cutlass moe problem size calculation, 5.3% E2E Throughput improvement, 2.2% TTFT improvement (#31830) 
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Signed-off-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
 2026-01-09 11:13:43 -08:00
Michael Goin
d5ec6c056f [UX] Add vLLM model inspection view (#29450) 
Signed-off-by: mgoin <mgoin64@gmail.com>
 2026-01-09 10:12:35 -07:00
Shanshan Shen
08d954f036 [Doc] Add developer guide for CustomOp (#30886) 
Signed-off-by: shen-shanshan <467638484@qq.com>
 2026-01-09 16:21:11 +00:00
Michael Goin
34cd32fe30 [Perf][Kernel] Fused SiLU+Mul+Quant kernel for NVFP4 cutlass_moe (#31832) 
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
 2026-01-09 07:40:33 -07:00
Xin Yang
e7b68f4d6c [Bugfix] Fix Triton FusedMoE LoRA (#30585) 
Signed-off-by: Xin Yang <xyangx@amazon.com>
 2026-01-09 11:46:59 +00:00
Cyrus Leung
c8ed39b9dd [Model] Reorganize pooling layers (#31973) 
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
 2026-01-09 11:02:14 +00:00
Robert Shaw
0fa8dd24d2 [Bugfix] Fix Typo from NVFP4 Refactor (#31977) 
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
 2026-01-08 16:18:50 -08:00
Robert Shaw
5825bbc1f7 [Quantization] Deprecate Long Tail of Schemes (#31688) 
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Signed-off-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
 2026-01-08 19:07:45 -05:00
Yongye Zhu
d62cfe546d [MoE Refactoring][Bugfix]Wrap WNA16 Triton kernel into mk and change compressed tensor kernel selection (#31752) 
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
 2026-01-08 19:01:30 -05:00
Lucas Wilkinson
6cdf015c3c [Misc] Fix Current vLLM config is not set. warnings, assert to avoid issues in the future (#31747) 
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
Co-authored-by: Luka Govedič <ProExpertProg@users.noreply.github.com>
 2026-01-08 15:20:49 -08:00
Dipika Sikka
5d3b6097ad [Compressed-Tensors] Simplify NVFP4 Conditions, enable marlin support for NVFP4A16 MoEs (#30881) 2026-01-08 17:45:17 -05:00
bnellnm
e74698c27a [Misc][Refactor] Add FusedMoERouter object (#30519) 
Signed-off-by: Bill Nell <bnell@redhat.com>
 2026-01-08 20:52:55 +00:00
Michael Goin
87e07a6b46 Revert "feat(moe): Add is_act_and_mul=False support for Triton MoE kernels" (#31978) 2026-01-08 11:31:53 -08:00
danisereb
b8112c1d85 [Bugfix] Fix vllm serve failure with Nemotron Nano V3 FP8 (#31960) 
Signed-off-by: Daniel Serebrenik <daserebrenik@nvidia.com>
 2026-01-08 16:08:37 +00:00