Commits · 48ddb02b79d7e22e2eefbf5294bf70de50afd1b2 · OpenDAS / vllm_cscc

25 Nov, 2025 3 commits

[Hybrid Allocator] Support KV cache groups with different block_size (#29143) · 48ddb02b
Yifan Qiao authored Nov 25, 2025
```
Signed-off-by: Yifan Qiao <yifanqiao@berkeley.edu>
Co-authored-by: Chen Zhang <zhangch99@outlook.com>
```
48ddb02b

[Model] Add HunyuanOCR support (#29327) · 92effb07

Isotr0py authored Nov 25, 2025


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>

92effb07

[Spec Decode] Add support for EAGLE3 heads that do not use_aux_hidden_states (#27688) · 5f9679a4

Hanjie Qiu authored Nov 24, 2025


Signed-off-by: hjjq <hanjieq@nvidia.com>
Signed-off-by: Benjamin Chislett <bchislett@nvidia.com>
Co-authored-by: Benjamin Chislett <bchislett@nvidia.com>

5f9679a4

24 Nov, 2025 7 commits

[XPU]fix Kimi-VL-A3B-thinking on xpu (#29309) · 3cfa63ad
Yan Ma authored Nov 25, 2025
```
Signed-off-by: Yan Ma <yan.ma@intel.com>
```
3cfa63ad

[Multimodal][Qwen3 Omni] Make Qwen3 Omni work with audio-in-video inputs in V1 engine. (#27721) · 839c6b7b

Chenheli Hua authored Nov 24, 2025


Signed-off-by: Chenheli Hua <huachenheli@outlook.com>
Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Roger Wang <hey@rogerw.io>

839c6b7b

Add option to use unbacked, and backed size obl dynamic shapes for more sounds... · 7a228b53

Laith Sakka authored Nov 24, 2025


Add option to use unbacked, and backed size obl dynamic shapes for more sounds compilation. (#26199)
Signed-off-by: Laith Sakka <lsakka@meta.com>

7a228b53

[Bugfix][Rocm] Fix shared expert weight loading failure in DeepSeek-MTP (#27563) · 8005e606
杰兮 authored Nov 24, 2025
```
Signed-off-by: zhyajie <yajizhan@amd.com>
Co-authored-by: zhyajie <yajizhan@amd.com>
```
8005e606
[Core] Deprecate `xformers` (#29262) · 0ff70821
Roger Wang authored Nov 23, 2025
```
Signed-off-by: Roger Wang <hey@rogerw.io>
```
0ff70821

[Model] Add OpenCUA-7B support (#29068) · 30854783

Zero authored Nov 24, 2025

Signed-off-by: lim4349 <rockmanzero@naver.com>
Signed-off-by: Zero <rockmanzero@naver.com>
Co-authored-by: Cloud User <ubuntu@a100-80g-4.novalocal>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

30854783

[LoRA] Optimize 3D MoE logic (#29222) · 1073ba68
Jee Jee Li authored Nov 24, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
1073ba68

22 Nov, 2025 1 commit

Fix EVS crash when using `video_embeds` inputs in Qwen2.5-VL (#29232) · d84d8f44

ZiTian Zhao authored Nov 22, 2025


Signed-off-by: zitian.zhao <zitian.zhao@tencentmusic.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>

d84d8f44

21 Nov, 2025 5 commits

[Core] Align whisper closer to other multimodal models (#27292) · cca2d2cd
Russell Bryant authored Nov 21, 2025
```
Signed-off-by: Russell Bryant <rbryant@redhat.com>
```
cca2d2cd
[V0 deprecation] Remove more V0 references (#29088) · aab0102a
Cyrus Leung authored Nov 21, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
aab0102a

[CI Failure] Fix Gemma3 RoPE configuration for sliding attention layers (#29111) · 8ac3a414

Huamin Li authored Nov 20, 2025


Signed-off-by: Huamin Li <3ericli@gmail.com>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>

8ac3a414

[Bugfix] Fix Plamo3 rope handling (#29092) · 0e741c12

Cyrus Leung authored Nov 21, 2025

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

0e741c12

[LoRA][2/2]Remove LoRA extra vocab (#28545) · 9875be64
Jee Jee Li authored Nov 21, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
9875be64

20 Nov, 2025 9 commits
- [BugFix] Fix flash_attn import in `siglip2navit.py` (#29082) · a2e9ebe9
  Fanli Lin authored Nov 20, 2025
```
Signed-off-by: Fanli Lin <fanli.lin@intel.com>
```
  a2e9ebe9
- [Bugfix] Fix spec decode memory regression after #28549 (#28819) · 93c8672c
  Zhewen Li authored Nov 20, 2025
```
Signed-off-by: zhewenli <zhewenli@meta.com>
```
  93c8672c
- [MODEL] Implement plamo3 (#28834) · c9e09311
  Shinichi Hemmi authored Nov 20, 2025
```
Signed-off-by: Shinichi Hemmi <50256998+Alnusjaponica@users.noreply.github.com>
```
  c9e09311
- [ROCm] Add AMD GPU support on Deepseek v3.2 and SparseMLA (#26670) · 06c20c99
  Pleaplusone authored Nov 20, 2025
```
Signed-off-by: ganyi <ygan@amd.com>
```
  06c20c99
- Add truncate arg to yarn to match openai implementation of gpt-oss (#28244) · 6eb745d9
  Anna Shors authored Nov 20, 2025
```
Signed-off-by: ashors1 <ashors@nvidia.com>
Co-authored-by: Chen Zhang <zhangch99@outlook.com>
```
  6eb745d9
- [BugFix] Fix Llama4 Pipeline Parallelism Assert Error (#28577) · dc45efc8
  Dezhan authored Nov 20, 2025
```
Co-authored-by: Dezhan Tu <dztu@meta.com>
```
  dc45efc8
- [ROCm][BugFix] Fix shared expert loading error when disable... · 7218f839
  Pleaplusone authored Nov 20, 2025
```
[ROCm][BugFix] Fix shared expert loading error when disable `VLLM_ROCM_USE_AITER_FUSION_SHARED_EXPERTS` (#28633)
Signed-off-by: ganyi <ygan@amd.com>
```
  7218f839
- [Model][QwenVL] Replace `torch.repeat_interleave` with faster `np.repeat` (#28964) · a9705a29
  Lukas Geiger authored Nov 20, 2025
```
Signed-off-by: Lukas Geiger <lukas.geiger94@gmail.com>
```
  a9705a29
- [Bugfix] Revert custom attention mask for gemma3-mm (#28995) · 64192d56
  Isotr0py authored Nov 20, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
```
  64192d56
19 Nov, 2025 14 commits

[Refactor] Optimize `select_experts` (#28069) · 5031cd5d
Wentao Ye authored Nov 19, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
5031cd5d
[Feature] EPLB on Qwen3VLMoe and CompressedTensorsWNA16MoEMethod (#28849) · 8e38e998
JartX authored Nov 20, 2025

8e38e998
[CI] Fix precommit `rope_theta` issue (#29040) · 0075bfff
Wentao Ye authored Nov 19, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
0075bfff
[DeepSeek] Fix DeepSeek V3.2 Rope Embedding (#28968) · 88f5b19f
Yongye Zhu authored Nov 19, 2025
```
Signed-off-by: Yongye Zhu <zyy1102000@gmail.com>
```
88f5b19f

[Feature] Prefill Context Parallel (PCP) basic support (#28718) · 2fd893b4

Qiu authored Nov 20, 2025


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>

2fd893b4

Eagle: MM Cuda Graphs with MRope (#28896) · 02f5903b

Izzy Putterman authored Nov 19, 2025


Signed-off-by: Izzy Putterman <iputterman@nvidia.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>

02f5903b

GLM-V video segmentation solution adjustment (#28941) · 0c80efd9
Yuxuan Zhang authored Nov 20, 2025
```
Signed-off-by: zRzRzRzRzRzRzR <2448370773@qq.com>
```
0c80efd9
Update `rope_scaling` to `rope_parameters` in preparation for Transformers v5 (#28542) · a8b70304
Harry Mellor authored Nov 19, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
a8b70304

[Model][Mamba] Add selector for mamba attention backend and make it pluggable... · d44e9df7

Shanshan Shen authored Nov 20, 2025


[Model][Mamba] Add selector for mamba attention backend and make it pluggable for other device (#26487)
Signed-off-by: shen-shanshan <467638484@qq.com>

d44e9df7

Relax Transformers modeling backend MoE experts check (#28952) · 4f5299f7
Harry Mellor authored Nov 19, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
4f5299f7
[Model][QwenVL] Simplify cos/sin rotary embedding indexing (#28962) · 3d4e7d34
Lukas Geiger authored Nov 19, 2025
```
Signed-off-by: Lukas Geiger <lukas.geiger94@gmail.com>
```
3d4e7d34
[Bugfix] Fix typo in Qwen3 Next model executor (#28960) · 73ff872d
Gleb Kurchanov authored Nov 19, 2025
```
Signed-off-by: Gleb Kurchanov <nepherpitou@gmail.com>
```
73ff872d
Move online quantization to `model.load_weights` (#26327) · da94c7c0
Jerry Zhang authored Nov 18, 2025
```
Signed-off-by: Jerry Zhang <jerryzh168@gmail.com>
```
da94c7c0

[Hybrid][torch.compile] Refactor mamba2 forward to avoid obscuring linear... · 1395461f

tomeras91 authored Nov 19, 2025


[Hybrid][torch.compile] Refactor mamba2 forward to avoid obscuring linear projections under custom op (#28587)
Signed-off-by: Tomer Asida <57313761+tomeras91@users.noreply.github.com>

1395461f

18 Nov, 2025 1 commit
- [Models] Replace all `nn.Conv2d` with vLLM's Conv2dLayer (#28842) · e4bb2684
  Isotr0py authored Nov 19, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  e4bb2684