Commits · 19cc9468fd0fa1701e7cb74b5928b329a1d16cf1 · OpenDAS / vllm_cscc · GitLab

22 Dec, 2025 1 commit
- [Feature]: Support NVIDIA ModelOpt HF FP8 variants FP8_PER_CHANNEL_PER_TOKEN... · 19cc9468
  CedricHuang authored Dec 22, 2025
```
[Feature]: Support NVIDIA ModelOpt HF FP8 variants FP8_PER_CHANNEL_PER_TOKEN and FP8_PB_WO  in vLLM (#30957)
```
  19cc9468
21 Dec, 2025 1 commit
- [MoE Refactor][4/N] Marlin Fp8 Mk (#31036) · b471092d
  Robert Shaw authored Dec 21, 2025
  
  b471092d
20 Dec, 2025 5 commits

[Quantization] add marlin w4a8/w8a8 check (#31061) · 7c73ceb5
Jinzhen Lin authored Dec 21, 2025
```
Signed-off-by: Jinzhen Lin <jinzhen.ljz@antgroup.com>
```
7c73ceb5

[Quantization] support logical_widths for fp8 marlin (#30962) · ee52d990

Jinzhen Lin authored Dec 21, 2025

Signed-off-by: Jinzhen Lin <jinzhen.ljz@antgroup.com>
Signed-off-by: Jinzhen Lin <linjinzhen@hotmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

ee52d990

[MoE Refactor][5/N] Isolate zero expert to LongCatFlash (#28891) · 54c89243

baonudesifeizhai authored Dec 20, 2025

Signed-off-by: baonudesifeizhai <85092850+baonudesifeizhai@users.noreply.github.com>
Signed-off-by: Dongjie Zou <85092850+baonudesifeizhai@users.noreply.github.com>
Signed-off-by: baonudesifeizhai <baonudesifeizhai@gmail.com>
Signed-off-by: Robert Shaw <robertgshaw2@gmail.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robertgshaw2@gmail.com>

54c89243

[XPU] enable fp8 online streaming quantization (#30944) · 560ae963
Yan Ma authored Dec 20, 2025
```
Signed-off-by: Yan Ma <yan.ma@intel.com>
```
560ae963
GLM-4.7 Tool Parser and Doc Update (#30876) · 8a7a4143
Yuxuan Zhang authored Dec 20, 2025
```
Signed-off-by: zRzRzRzRzRzRzR <2448370773@qq.com>
```
8a7a4143

19 Dec, 2025 10 commits
- [MoE Refactor][2/N] Use Modular Kernels for Fp8 (#30825) · 95befecc
  Robert Shaw authored Dec 19, 2025
```
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
```
  95befecc
- [Bug] Fix `error 'Dynamo failed to run FX node with fake tensors` for Deepseek V3.2 (#31046) · 4cf94298
  Wentao Ye authored Dec 19, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  4cf94298
- [MoE Refactor][3/N] Deprecate cutlass block quant fp8 (b200) (#30990) · 83a317f6
  Robert Shaw authored Dec 19, 2025
```
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
```
  83a317f6
- [Refactor] Refactor for `DeepGemmQuantScaleFMT` using cache (#30898) · 3bd8335b
  Wentao Ye authored Dec 19, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  3bd8335b
- [Model] Add MiMo-V2-Flash support (#30836) · 969bbc7c
  Zhonghua Deng authored Dec 20, 2025
```
Signed-off-by: Abatom <abzhonghua@gmail.com>
Signed-off-by: Jumiar <liuanqim10@126.com>
Signed-off-by: Zyann7 <zyann7@outlook.com>
Co-authored-by: Jumiar <liuanqim10@126.com>
Co-authored-by: Zyann7 <zyann7@outlook.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>
```
  969bbc7c
- [Quantization] fix marlin w8a8 check (#30961) · 5fbfa8d9
  Jinzhen Lin authored Dec 19, 2025
```
Signed-off-by: Jinzhen Lin <jinzhen.ljz@antgroup.com>
```
  5fbfa8d9
- [CustomOp][Refactor] Extract common methods for ApplyRotaryEmb CustomOp (#31021) · 23a1946e
  Shanshan Shen authored Dec 19, 2025
```
Signed-off-by: shen-shanshan <467638484@qq.com>
```
  23a1946e
- [Quantization] enable compressed-tensors marlin support for turing (2) (#31008) · 9187de9f
  Jinzhen Lin authored Dec 19, 2025
```
Signed-off-by: Jinzhen Lin <jinzhen.ljz@antgroup.com>
```
  9187de9f
- [Quantization] enable compressed-tensors marlin support for turing (#31000) · de08b8f6
  Jinzhen Lin authored Dec 19, 2025
```
Signed-off-by: Jinzhen Lin <jinzhen.ljz@antgroup.com>
```
  de08b8f6
- [ROCm][CI][Bugfix] Multi-Modal Model Support Fixes and Attention Backend Improvements (#30270) · 7b43db21
  Andreas Karatzas authored Dec 18, 2025
```
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
```
  7b43db21
18 Dec, 2025 17 commits
- [Bug] Fix compressed tensor not using deepgemm (#30820) · 97000a2b
  Wentao Ye authored Dec 18, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  97000a2b
- tuned fused configs for B300 (#30629) · b8c477c1
  navmarri14 authored Dec 18, 2025
  
  b8c477c1
- [Perf] enable flashinfer rotary_embedding custom ops in DeepSeek rotary (#30729) · 53ad423f
  jiahanc authored Dec 18, 2025
```
Signed-off-by: jiahanc <173873397+jiahanc@users.noreply.github.com>
```
  53ad423f
- [MM Encoder]: Migrate legacy ViT `MultiHeadAttention` to new... · 700a5ad6
  Isotr0py authored Dec 19, 2025
```
[MM Encoder]: Migrate legacy ViT `MultiHeadAttention` to new `MMEncoderAttention` interface (#30684)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  700a5ad6
- fix fp8 online quantization streaming with tp > 1 (#30900) · f4ee2c3d
  Vasiliy Kuznetsov authored Dec 18, 2025
```
Signed-off-by: vasiliy <vasiliy@fb.com>
```
  f4ee2c3d
- [LoRA] Set default MXFP4 LoRA backend to Marlin (#30598) · 9a5e9652
  Xin Yang authored Dec 18, 2025
```
Signed-off-by: Xin Yang <xyangx@amazon.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
```
  9a5e9652
- adds jais 2 support (#30188) · 28d15ab5
  sarathc-cerebras authored Dec 18, 2025
```
Signed-off-by: sarathc-cerebras <sarath.chandran@cerebras.net>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
```
  28d15ab5
- [Bug] Fix batch invariant in torch 2.10 (#30907) · 66287582
  Wentao Ye authored Dec 18, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Signed-off-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
```
  66287582
- Filter safetensors files to download if .safetensors.index.json exists (#30537) · 100f93d2
  Michael Goin authored Dec 18, 2025
```
Signed-off-by: mgoin <mgoin64@gmail.com>
```
  100f93d2
- [moe] Use enable_chunking func (to support disabling chunking) (#29935) · 8372be28
  Ming Yang authored Dec 18, 2025
```
Signed-off-by: Ming Yang <minos.future@gmail.com>
```
  8372be28
- [CPU] Refactor CPU fused MOE (#30531) · e3ab93c8
  Li, Jiang authored Dec 18, 2025
```
Signed-off-by: jiang1.li <jiang1.li@intel.com>
```
  e3ab93c8
- [PERF] Qwen3-next. Add fp8 cutlass MoE tuned configs. `chmod -x *MI308X.json` (#29553) · 717ac33d
  Vadim Gimpelson authored Dec 18, 2025
```
Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>
```
  717ac33d
- [refactor] Add prefix support to embed_tokens in DeepSeek MTP (#30788) · b166ef20
  zzhxxx authored Dec 18, 2025
```
Signed-off-by: zzhx1 <zzh_201018@outlook.com>
```
  b166ef20
- [UX] Reduce DeepGEMM warmup log output to single progress bar (#30903) · 4a8412f7
  Matthew Bonanni authored Dec 17, 2025
```
Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
```
  4a8412f7
- [Quantization] Support Quark int4-fp8 w4a8 for MoE (#30071) · 0c738b58
  Bowen Bao authored Dec 17, 2025
```
Signed-off-by: Bowen Bao <bowenbao@amd.com>
```
  0c738b58
- [Chore] Remove v0 dead code for Qwen2.5-omni (#30883) · 6fe58876
  Isotr0py authored Dec 18, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  6fe58876
- [v1] Add PrefixLM support to TritonAttention backend (#30386) · 74a1ac38
  Isotr0py authored Dec 18, 2025
  
  74a1ac38
17 Dec, 2025 6 commits
- [BugFix] Workspace allocation during profile run : DeepEPHighThroughput + DeepGEMM (#30899) · e3fc374a
  Varun Sundar Rabindranath authored Dec 17, 2025
  
  e3fc374a
- 2.9.1 PyTorch release update (#28495) · e06d0bf0
  Andrey Talman authored Dec 17, 2025
  
  e06d0bf0
- [Model] Gemma3: Support untied word embeddings (#30827) · 196cdc32
  KimHyemin authored Dec 18, 2025
```
Signed-off-by: www-spam <panmahm@naver.com>
```
  196cdc32
- [Bugfix] deepseek-V3.2 self.weights_proj has no bias (#30841) · 84896fda
  baoqian426 authored Dec 17, 2025
```
Signed-off-by: baoqian <1354987947@qq.com>
Signed-off-by: baoqian426 <1354987947@qq.com>
```
  84896fda
- [Bug] Fix AttributeError: 'ColumnParallelLinear' object has no attribute... · f284d7bd
  Wentao Ye authored Dec 17, 2025
```
[Bug] Fix AttributeError: 'ColumnParallelLinear' object has no attribute `weight_scale_inv` (#30823)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  f284d7bd
- CustomOp: grouped topk (#29575) · 3b1d440e
  Xinyu Chen authored Dec 17, 2025
```
Signed-off-by: Xinyu Chen <xinyu1.chen@intel.com>
```
  3b1d440e