Commits · 1dc8a70b6d4e8ba4e139f1ddb86a166694f42f21 · OpenDAS / vllm_cscc

07 Aug, 2025 6 commits

[Attention] Support multiple attention metadata builders per kv_cache_spec +... · 1dc8a70b

Lucas Wilkinson authored Aug 06, 2025


[Attention] Support multiple attention metadata builders per kv_cache_spec  + proper local attention no hybrid kv cache fix (#21588)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>

1dc8a70b

[model] Support MiniCPM-V 4.0 (#22166) · 41b67f42
tc-mb authored Aug 07, 2025
```
Co-authored-by: imning3 <hbning@pku.edu.cn>
```
41b67f42

Fix trtllm-gen attention env and add attention sink (#22378) · 9a3835aa

Lain authored Aug 06, 2025


Signed-off-by: Siyuan Fu <siyuanf@nvidia.com>
Signed-off-by: Lain <fusiyuan2000@hotmail.com>
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
Co-authored-by: Yongye Zhu <zyy1102000@gmail.com>

9a3835aa

[gpt-oss] fix model config with hf_config (#22401) · 5c7cc33f
Yongye Zhu authored Aug 06, 2025
```
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
```
5c7cc33f
[Bug] Fix B200 DeepGEMM E8M0 Accuracy Issue (#22399) · eec890c1
Wentao Ye authored Aug 06, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
eec890c1
[v1] - Mamba1 Attention Metadata (#21249) · 46a13949
Asaf Joseph Gardin authored Aug 07, 2025
```
Signed-off-by: asafg <asafg@ai21.com>
Co-authored-by: asafg <asafg@ai21.com>
```
46a13949

06 Aug, 2025 5 commits

[gpt-oss] flashinfer mxfp4 (#22339) · 31f09c61

Yongye Zhu authored Aug 06, 2025


Signed-off-by: simon-mo <xmo@berkeley.edu>
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
Co-authored-by: simon-mo <xmo@berkeley.edu>

31f09c61

[GptOss] Add GptOss reasoning parser to support structure output (#22322) · a47e6ffe

Chen Zhang authored Aug 05, 2025


Signed-off-by: Chen Zhang <zhangch99@outlook.com>
Co-authored-by: LiuXiaoxuanPKU <lilyliupku@gmail.com>
Co-authored-by: simon-mo <xmo@berkeley.edu>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Co-authored-by: Hongxia Yang <62075498+hongxiayang@users.noreply.github.com>
Co-authored-by: Minseok Lee <47620120+minseokl@users.noreply.github.com>
Co-authored-by: Yongye Zhu <zyy1102000@gmail.com>

a47e6ffe

Add GPT-OSS model code and config [1/N] (#22327) · de98252f
Woosuk Kwon authored Aug 05, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
de98252f

Update transformers to `v4.55` (#21931) · 796bae07

Harry Mellor authored Aug 06, 2025


Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Signed-off-by: Isotr0py <2037008807@qq.com>
Signed-off-by: isotr0py <2037008807@qq.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: DarkLight1337 <tlleungac@connect.ust.hk>
Co-authored-by: Isotr0py <2037008807@qq.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>

796bae07

[Bugfix] Fix MoE BNB version (#22260) · 8e6c7e87
Jee Jee Li authored Aug 06, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
8e6c7e87

05 Aug, 2025 7 commits
- Migrate KimiVLImagePixelInputs to TensorSchema (#21769) · 05fae021
  Benji Beck authored Aug 05, 2025
```
Signed-off-by: Benji Beck <benjibeck@meta.com>
Co-authored-by: Isotr0py <2037008807@qq.com>
```
  05fae021
- [Model] Pooling model activation supports per request control by PoolingParams (#20538) · 586f2867
  wang.yuqi authored Aug 05, 2025
```
Signed-off-by: wang.yuqi <noooop@126.com>
```
  586f2867
- Optimize configuration access with LRU cache in custom ops (#22204) · 4b3e4474
  ZiTian.Zhao authored Aug 05, 2025
```
Signed-off-by: zitian zhao <zitian.zhao@tencentmusic.com>
```
  4b3e4474
- [Log] DeepGEMM Update Log for Unaligned Problem Size (#22208) · d7b28f34
  Wentao Ye authored Aug 04, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  d7b28f34
- self.gate dtype update for GLM-4.5 (#22203) · 6fa41e0c
  Yuxuan Zhang authored Aug 05, 2025
```
Signed-off-by: zRzRzRzRzRzRzR <2448370773@qq.com>
```
  6fa41e0c
- [FEAT] Refactor ROPE into module (#22192) · 6ad6b8e1
  TJian authored Aug 04, 2025
```
Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>
```
  6ad6b8e1
- [NVIDIA] Auto detect modelopt quant and fix DSR1-FP4 weight loading (#22073) · bdcb42e4
  Po-Han Huang (NVIDIA) authored Aug 05, 2025
  
  bdcb42e4
04 Aug, 2025 4 commits
- Fix Arcee model weight loading: Add custom load_weights (#21725) · a5fff3bd
  Raghav Ravishankar authored Aug 04, 2025
```
Signed-off-by: alyosha-swamy <raghav@arcee.ai>
```
  a5fff3bd
- [feat] move WEIGHT_SCALE_SUPPORTED into raise block to accelerate RLHF weight loading (#21164) · c1b4eb04
  Weixiao Huang authored Aug 04, 2025
```
Signed-off-by: huangweixiao <huangweixiao@msh.team>
```
  c1b4eb04
- [Misc] Modify the organization of GLM series (#22171) · a7b8788d
  Jee Jee Li authored Aug 04, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
  a7b8788d
- Remove index_put from MM embeddings merging (#22105) · e5949e5a
  Chenxi Yang authored Aug 03, 2025
```
Co-authored-by: Chenxi Yang <cxyang@meta.com>
```
  e5949e5a
03 Aug, 2025 4 commits
- fuse fp32 for GLM-4.5 e_score_correction_bias (#22143) · d3c18c9c
  Yuxuan Zhang authored Aug 04, 2025
```
Signed-off-by: zRzRzRzRzRzRzR <2448370773@qq.com>
```
  d3c18c9c
- [CI/Build][Bugfix] Fix Qwen2.5 tests in CPU CI via fallback silu_and_mul to... · b5dfb94f
  Li, Jiang authored Aug 03, 2025
```
[CI/Build][Bugfix] Fix Qwen2.5 tests in CPU CI via fallback silu_and_mul to torch native implementation (#22145)
Signed-off-by: jiang1.li <jiang1.li@intel.com>
```
  b5dfb94f
- [Misc] Add tensor schema test coverage for multimodal models (#21754) · 3dddbf1f
  Isotr0py authored Aug 03, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Isotr0py <2037008807@qq.com>
```
  3dddbf1f
- [Fix] Fix llama4 modelopt weight loading error (#22107) · 337eb23b
  jiahanc authored Aug 03, 2025
```
Signed-off-by: jiahanc <173873397+jiahanc@users.noreply.github.com>
Co-authored-by: mgoin <mgoin64@gmail.com>
```
  337eb23b
02 Aug, 2025 9 commits

[xpu]support moe models on XPU platform (#21643) · 73e1b9b1
Yan Ma authored Aug 02, 2025
```
Signed-off-by: yan <yan.ma@intel.com>
Signed-off-by: Yan Ma <yan.ma@intel.com>
```
73e1b9b1

[Model] Mamba2 preallocate SSM output tensor to avoid d2d copy overhead (#21075) · b690e348

Chih-Chieh Yang authored Aug 02, 2025


Signed-off-by: Chih-Chieh Yang <7364402+cyang49@users.noreply.github.com>
Signed-off-by: Chih-Chieh-Yang <7364402+cyang49@users.noreply.github.com>

b690e348

for glm-4.1V update (#22000) · 25373b6c

Yuxuan Zhang authored Aug 02, 2025


Signed-off-by: Isotr0py <2037008807@qq.com>
Signed-off-by: zRzRzRzRzRzRzR <2448370773@qq.com>
Co-authored-by: Isotr0py <2037008807@qq.com>

25373b6c

[Bugfix] Mamba2 remove bugged initial state condition in chunk scan (#22034) · c64861d6
Chih-Chieh Yang authored Aug 02, 2025
```
Signed-off-by: Chih-Chieh-Yang <7364402+cyang49@users.noreply.github.com>
```
c64861d6

[FEAT][ROCm] Enable running Flash Attention as ViT attn backend for Qwen-VL... · d3a6f212

vllmellm authored Aug 02, 2025


[FEAT][ROCm] Enable running Flash Attention as ViT attn backend for Qwen-VL models on ROCm platform. (#22069)
Signed-off-by: tjtanaavllm <tunjian.tan@amd.com>
Signed-off-by: vllmellm <vllm.ellm@embeddedllm.com>
Co-authored-by: tjtanaavllm <tunjian.tan@amd.com>

d3a6f212

[Speculators][Speculative Decoding] Add Qwen Eagle3 Support (#21835) · 9f9c38c3
Dipika Sikka authored Aug 01, 2025
```
Signed-off-by: Dipika Sikka <dipikasikka1@gmail.com>
```
9f9c38c3
[Misc] DeepGemmExperts : Avoid JIT generation in the hot-path (#21955) · a65f46be
Varun Sundar Rabindranath authored Aug 02, 2025
```
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
```
a65f46be

[Model] Qwen2.5 VL SiLU-and-Mul (#22066) · ee2eb6ec

vllmellm authored Aug 02, 2025


Signed-off-by: kf <kuanfu.liu@embeddedllm.com>
Signed-off-by: vllmellm <vllm.ellm@embeddedllm.com>
Co-authored-by: kf <kuanfu.liu@embeddedllm.com>

ee2eb6ec

feat: Add Support GPTQ Quantization MOE on ROCM vllm serve (#21733) · 3654847d
JartX authored Aug 02, 2025

3654847d

01 Aug, 2025 5 commits
- Enable headless models for pooling in the Transformers backend (#21767) · 38c8bce8
  Harry Mellor authored Aug 01, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  38c8bce8
- [Bugfix] [Performance] DeepEPHighThroughput + DeepSeek : Quant before Dispatch (#21837) · ac45c44d
  Varun Sundar Rabindranath authored Aug 01, 2025
```
Signed-off-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
Co-authored-by: Varun Sundar Rabindranath <vsundarr@redhat.com>
```
  ac45c44d
- [Bugfix] Fix glm4.1v video inference issue (#22067) · 3f8e9521
  Isotr0py authored Aug 02, 2025
```
Signed-off-by: Isotr0py <2037008807@qq.com>
```
  3f8e9521
- [Speculative Decoding] Add `speculators` config support (#21345) · dfbc1f88
  Dipika Sikka authored Aug 01, 2025
  
  dfbc1f88
- Revert "Update sampling_metadata.py (#21937)" (#22088) · 87c94bc8
  Harry Mellor authored Aug 01, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  87c94bc8