Commits · c277fbdf31f49c15049fe361ef92a7e34abc7b74 · OpenDAS / vllm_cscc

16 Jan, 2026 1 commit

[Feat] Support non-gated MoE with Marlin, NVFP4 CUTLASS, FP8, INT8, compressed-tensors (#32257) · c277fbdf

TomerBN-Nvidia authored Jan 16, 2026


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>

c277fbdf

15 Jan, 2026 6 commits

[ROCM] DSfp4 mla projection gemms weight dynamic quantization (#32238) · 8c11001b
Aleksandr Malyshev authored Jan 15, 2026
```
Signed-off-by: Aleksandr Malyshev <maleksan@amd.com>
Co-authored-by: Aleksandr Malyshev <maleksan@amd.com>
```
8c11001b

[BugFix] Fix `assert x_s.shape[-1] == x_q.shape[-1] // group_shape[1]` in... · c36ba69b

Lucas Wilkinson authored Jan 15, 2026


[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>

c36ba69b

[Quant] Support MXFP4 W4A16 for compressed-tensors MoE models (#32285) · 361dfdc9
Dipika Sikka authored Jan 15, 2026
```
Signed-off-by: Dipika Sikka <dipikasikka1@gmail.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
```
361dfdc9
[ROCm] Improve error handling while loading quantized model on gfx120… (#31715) · b89275d0
brian033 authored Jan 15, 2026
```
Signed-off-by: brian033 <85883730+brian033@users.noreply.github.com>
Co-authored-by: TJian <tunjian.tan@embeddedllm.com>
```
b89275d0

[BugFix] Fix DeepSeek-V3.1 + DeepGEMM incompatible scale shapes (#32361) · 2c9b4cf5

Lucas Wilkinson authored Jan 14, 2026


Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Co-authored-by: Eldar Kurtić <8884008+eldarkurtic@users.noreply.github.com>

2c9b4cf5

[CI][AMD][Quantization][BugFix] Fix fp8 max in quant_utils.py and update... · 3c268564

rasmith authored Jan 14, 2026


[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>

3c268564

14 Jan, 2026 2 commits
- [Docs] Add docs about OOT Quantization Plugins (#32035) · 6388b500
  Michael Goin authored Jan 14, 2026
```
Signed-off-by: mgoin <mgoin64@gmail.com>
```
  6388b500
- Consolidate Intel Quantization Toolkit Integration in vLLM (#31716) · 50632adc
  Yi Liu authored Jan 14, 2026
```
Signed-off-by: yiliu30 <yi4.liu@intel.com>
```
  50632adc
13 Jan, 2026 2 commits

[Kernel][Performance] Enable smaller Scaling Factor tiling for NVFP4 small-batch decoding (#30885) · 8ef50d9a

Roberto L. Castro authored Jan 14, 2026

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>

8ef50d9a

[Quantization] fix: overflow with static per-tensor scaling (#29867) · a5bbbd2f

Mickaël Seznec authored Jan 13, 2026


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>

a5bbbd2f

11 Jan, 2026 1 commit
- [Hardware][AMD][CI][Bugfix] Fix AMD Quantization test group (#31713) · bde57ab2
  Matt authored Jan 11, 2026
```
Signed-off-by: Matthew Wong <Matthew.Wong2@amd.com>
```
  bde57ab2
10 Jan, 2026 2 commits
- [Bugfix][Quantization] Ensure input contiguity in per_token_quant_int8 (#31637) · 6ea001cf
  Vensen authored Jan 11, 2026
```
Signed-off-by: vensen <vensenmu@gmail.com>
```
  6ea001cf
- [Quant] Support MXFP4 W4A16 for compressed-tensors dense models (#31926) · e6c6f2c7
  Michael Goin authored Jan 10, 2026
```
Signed-off-by: mgoin <mgoin64@gmail.com>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
```
  e6c6f2c7
09 Jan, 2026 6 commits

[Quant] Make static quant support all group shapes (#30833) · 0a0aa077
Lucas Wilkinson authored Jan 09, 2026
```
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
```
0a0aa077
[fix] add cutedsl to global sf (#32001) · f9e2a75a
jiahanc authored Jan 09, 2026
```
Signed-off-by: jiahanc <173873397+jiahanc@users.noreply.github.com>
```
f9e2a75a
[Doc] Add developer guide for CustomOp (#30886) · 08d954f0
Shanshan Shen authored Jan 10, 2026
```
Signed-off-by: shen-shanshan <467638484@qq.com>
```
08d954f0

[Bugfix] Fix Typo from NVFP4 Refactor (#31977) · 0fa8dd24

Robert Shaw authored Jan 08, 2026


Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>

0fa8dd24

[Quantization] Deprecate Long Tail of Schemes (#31688) · 5825bbc1

Robert Shaw authored Jan 08, 2026

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>

5825bbc1

[MoE Refactoring][Bugfix]Wrap WNA16 Triton kernel into mk and change... · d62cfe54

Yongye Zhu authored Jan 08, 2026


[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>

d62cfe54

08 Jan, 2026 5 commits
- [Compressed-Tensors] Simplify NVFP4 Conditions, enable marlin support for NVFP4A16 MoEs (#30881) · 5d3b6097
  Dipika Sikka authored Jan 08, 2026
  
  5d3b6097
- [Misc][Refactor] Add FusedMoERouter object (#30519) · e74698c2
  bnellnm authored Jan 08, 2026
```
Signed-off-by: Bill Nell <bnell@redhat.com>
```
  e74698c2
- [Bugfix] Fix vllm serve failure with Nemotron Nano V3 FP8 (#31960) · b8112c1d
  danisereb authored Jan 08, 2026
```
Signed-off-by: Daniel Serebrenik <daserebrenik@nvidia.com>
```
  b8112c1d
- [MoE Refactor][16/N] Apply Refactor to NVFP4 (#31692) · 9f6dcb71
  Robert Shaw authored Jan 07, 2026
```
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: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Pavani Majety <pmajety@nvidia.com>
```
  9f6dcb71
- [MoE Refactor][15/N] Apply Refactor to Fp8 (#31415) · 5dcd7ef1
  Robert Shaw authored Jan 07, 2026
  
  5dcd7ef1
07 Jan, 2026 4 commits

[Perf][Kernels] Enable FlashInfer DeepGEMM swapAB on SM90 (for W8A8 Linear Op) (#29213) · cc6dafae

Kate Cheng authored Jan 07, 2026


Signed-off-by: Kate Cheng <yunhsuanc@nvidia.com>
Signed-off-by: Jhao-Ting Chen <jhaotingc@nvidia.com>
Co-authored-by: Jhao-Ting Chen <jhaotingc@nvidia.com>

cc6dafae

[Misc] Improve error messages for unsupported types and parameters (#30593) · 0790f076

BlankR authored Jan 07, 2026


Signed-off-by: BlankR <hjyblanche@gmail.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>

0790f076

[Refactor][TPU] Remove torch_xla path and use tpu-inference (#30808) · e7596371

weiyu authored Jan 07, 2026


Signed-off-by: Wei-Yu Lin <weiyulin@google.com>
Signed-off-by: weiyu <62784299+weiyu0824@users.noreply.github.com>

e7596371

[Bugfix][Hardware][AMD] Consolidate FP8 min/max values helper function (#31106) · 4614c5a5

Kevin McKay authored Jan 07, 2026


Signed-off-by: c0de128 <kevin.mckay@outlook.com>
Signed-off-by: Kevin McKay <kevin@example.com>
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>

4614c5a5

06 Jan, 2026 2 commits

[Quantization][Refactor] Move CPU GPTQ kernel into MP linear (#31801) · 8becf146

Li, Jiang authored Jan 07, 2026

Signed-off-by: jiang1.li <jiang1.li@intel.com>
Signed-off-by: Li, Jiang <bigpyj64@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

8becf146

[MoE Refactor][14/N] Clean Up FI Quant Config Smuggling (#31593) · af8fd730
Robert Shaw authored Jan 06, 2026
```
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
```
af8fd730

05 Jan, 2026 1 commit

[Bugfix] Fix Broken ModelOpt NVFP4 MoE (#31742) · f6c0009a

Robert Shaw authored Jan 05, 2026


Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>

f6c0009a

03 Jan, 2026 1 commit

[MoE Refactor][13/N] Convert FI to Use PFNoEP (#31533) · 268b1c55

Robert Shaw authored Jan 03, 2026


Signed-off-by: Robert Shaw <robshaw@redhat.com>
Signed-off-by: Robert Shaw <robertgshaw2@gmail.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>

268b1c55

02 Jan, 2026 3 commits
- [MoE] Fix output_shape calculation in Attention layer to handle 3D query inputs (#31596) · 6ef770df
  Andreas Karatzas authored Jan 02, 2026
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  6ef770df
- Remove unused `use_marlin` variable in `Mxfp4MoEMethod` (#31549) · 1f43c121
  Vaibhav Sourirajan authored Jan 02, 2026
```
Signed-off-by: vaibhav sourirajan <vs2787@columbia.edu>
```
  1f43c121
- [Bugfix] Fix activation quantization for compressed-tensors W4A16 (#31572) · ca179d0f
  Tmn07 authored Jan 02, 2026
```
Signed-off-by: Tmn07 <tmn0796@gmail.com>
```
  ca179d0f
30 Dec, 2025 1 commit
- [BugFix] add select_gemm_impl on CompressedTensorsWNA16MoEMethod to support LoRA (#31453) · 07728bf5
  JartX authored Dec 30, 2025
```
Signed-off-by: JartX <sagformas@epdcenter.es>
```
  07728bf5
29 Dec, 2025 2 commits

[Bugfix][ROCm] Fix Static Quant Issue (#31502) · 56f51625

Robert Shaw authored Dec 29, 2025


Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: TJian <tunjian.tan@embeddedllm.com>

56f51625

[MoE Refactor][12/N] Marlin Fp8 MoE Pure Function (#31499) · 9152a30d
Robert Shaw authored Dec 29, 2025
```
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
```
9152a30d

27 Dec, 2025 1 commit
- [MoE Refactor][10/N] Cleanup Fp8 Process Weights After Loading (#31169) · 727c41f3
  Robert Shaw authored Dec 27, 2025
```
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
```
  727c41f3