Commits · 913249f9078076ff2f11f38ce09f80b5d08de68c · OpenDAS / vllm_cscc · GitLab

30 Sep, 2025 3 commits
- DeepSeek-R1-Channel-INT8调用rmsquant融合 · 9f087f8b
  zhuwenwen authored Sep 30, 2025
  
  9f087f8b
- [kernels] add rotary_embedding_deepseek_fuse · 1e911dbd
  zhuwenwen authored Sep 30, 2025
```
off rotary_embedding_deepseek_fuse
```
  1e911dbd
- feat: moe_align_block_size 更新lightop 接口,加入对ep的支持 · 63f1c793
  zhuwenwen authored Sep 30, 2025
  
  63f1c793
26 Sep, 2025 3 commits
- update op · dfe1a844
  zhuwenwen authored Sep 26, 2025
  
  dfe1a844
- update layer.py · cbaa3d62
  zhuwenwen authored Sep 26, 2025
  
  cbaa3d62
- update VLLM_USE_LIGHTOP and VLLM_USE_OPT_CAT · f8cf43ae
  zhuwenwen authored Sep 26, 2025
  
  f8cf43ae
25 Sep, 2025 1 commit
- update VLLM_USE_LIGHTOP and VLLM_USE_OPT_CAT · f70d0ec6
  zhuwenwen authored Sep 25, 2025
  
  f70d0ec6
23 Sep, 2025 3 commits
- 修复W4A8 PP8崩溃问题 · d70425b2
  zhuwenwen authored Sep 23, 2025
  
  d70425b2
- deepseek-r1-w4a8 mlp/moe调用silu-mul-quant融合 · 3155bf9e
  zhuwenwen authored Sep 23, 2025
  
  3155bf9e
- Update layernorm.py · 5dc2dc0d
  zhuwenwen authored Sep 23, 2025
  
  5dc2dc0d
22 Sep, 2025 1 commit
- update glm4.5v conv layout · dcaa837c
  zhuwenwen authored Sep 22, 2025
  
  dcaa837c
18 Sep, 2025 1 commit
- [Model] Add LongCat-Flash · 50281e36
  zhuwenwen authored Sep 18, 2025
  
  50281e36
16 Sep, 2025 1 commit
- update activation args · e808fae2
  zhuwenwen authored Sep 16, 2025
  
  e808fae2
15 Sep, 2025 2 commits
- update moe_align_block_size · 8e2b79b4
  zhuwenwen authored Sep 15, 2025
  
  8e2b79b4
- deepseek-r1-w4a8使用rmsquant融合算子及横向融合 · 9f201bc1
  zhuwenwen authored Sep 15, 2025
  
  9f201bc1
14 Sep, 2025 1 commit
- update deepseek_v2.py · 2743daa4
  zhuwenwen authored Sep 14, 2025
  
  2743daa4
13 Sep, 2025 5 commits
- [Compilation Bug] Fix Inductor Graph Output with Shape Issue (#24772) · da3fa78d
  Wentao Ye authored Sep 12, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  da3fa78d
- Enable conversion of multimodal models to pooling tasks (#24451) · bbb70036
  Maximilien de Bayser authored Sep 12, 2025
```
Signed-off-by: Max de Bayser <mbayser@br.ibm.com>
```
  bbb70036
- [Qwen3-Next] MoE configs for H100 TP=1,2 and TP2/EP (#24739) · 01085b13
  Elvir Crnčević authored Sep 12, 2025
```
Signed-off-by: elvircrn <elvircrn@gmail.com>
```
  01085b13
- [BugFix] Fix Qwen3-Next PP (#24709) · 66160a99
  Nick Hill authored Sep 11, 2025
```
Signed-off-by: Nick Hill <nhill@redhat.com>
```
  66160a99
- [Qwen3-Next] MoE configs for H20 TP=1,2,4,8 (#24707) · eaca762c
  Jee Jee Li authored Sep 12, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
  eaca762c
12 Sep, 2025 2 commits
- [Bugfix] fixes the causal_conv1d_update kernel update non-speculative decoding cases (#24680) · 880c741b
  Tao He authored Sep 12, 2025
```
Signed-off-by: Tao He <linzhu.ht@alibaba-inc.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
```
  880c741b
- [Startup] Make DeepGEMM warmup scale with max-num-batched-tokens (#24693) · 2e6bc468
  Lucas Wilkinson authored Sep 11, 2025
```
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
```
  2e6bc468
11 Sep, 2025 17 commits
- [Bug] Fix Layer `weight_block_size` Assertion Issue (#24674) · fcba05c4
  Wentao Ye authored Sep 11, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  fcba05c4
- [Qwen3-Next] MOE configs for H100 TP4 (#24699) · f82f7a89
  Chen Zhang authored Sep 11, 2025
```
Signed-off-by: Chen Zhang <zhangch99@outlook.com>
```
  f82f7a89
- [Perf] Use upstream CUTLASS for SM90 Block FP8 kernel (#23280) · c3aea10d
  Michael Goin authored Sep 11, 2025
```
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
```
  c3aea10d
- [Qwen3-Next] Add B200 MoE configs for Qwen3-next (#24698) · 7a70a718
  Vadim Gimpelson authored Sep 12, 2025
```
Signed-off-by: Vadim Gimpelson <vadim.gimpelson@gmail.com>
```
  7a70a718
- [Qwen3-Next] MoE configs for H200 TP=1,2,4 (#24695) · 569bf1c9
  Woosuk Kwon authored Sep 11, 2025
```
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
```
  569bf1c9
- [Kernel][B200] `mxfp4` fused cutlass moe (#23696) · 074854b2
  Duncan Moss authored Sep 11, 2025
```
Signed-off-by: Duncan Moss <djm.moss@gmail.com>
Signed-off-by: Michael Goin <mgoin64@gmail.com>
Signed-off-by: mgoin <mgoin64@gmail.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
Co-authored-by: Michael Goin <mgoin64@gmail.com>
```
  074854b2
- [Qwen3-Next] Add MoE Config for H200 (#24688) · c733bd5e
  Woosuk Kwon authored Sep 11, 2025
```
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
```
  c733bd5e
- [Doc] Remove Useless Comments (#24687) · a892b259
  Wentao Ye authored Sep 11, 2025
```
Signed-off-by: yewentao256 <zhyanwentao@126.com>
```
  a892b259
- [Ultravox] Use wrapped_model_config to instantiate inner model (#24679) · 127ded0a
  Peter Salas authored Sep 11, 2025
```
Signed-off-by: Peter Salas <peter@fixie.ai>
```
  127ded0a
- [VLM] Migrate remain DP-supported ViT models to use `disable_tp` (#24363) · bb2b5126
  Isotr0py authored Sep 12, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  bb2b5126
- fix some typos (#24616) · e26fef83
  co63oc authored Sep 12, 2025
```
Signed-off-by: co63oc <co63oc@users.noreply.github.com>
```
  e26fef83
- [Bugfix] Fix platform-specific routing in CustomOp implementations (#24444) · 4aa23892
  Konrad Zawora authored Sep 11, 2025
```
Signed-off-by: Konrad Zawora <kzawora@habana.ai>
```
  4aa23892
- [VLM] Optimize GLM4.5-V-style video processing to only decode necessary frames (#24161) · bcbe2a4d
  Isotr0py authored Sep 12, 2025
```
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  bcbe2a4d
- [Bugifx] Fix qwen-next packed_modules_mapping (#24656) · 817beef7
  Jee Jee Li authored Sep 11, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
  817beef7
- Move `LoRAConfig` from `config/__init__.py` to `config/lora.py` (#24644) · 5f5271f1
  Harry Mellor authored Sep 11, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  5f5271f1
- Fix typing for `safetensors_load_strategy` (#24641) · d6249d06
  Harry Mellor authored Sep 11, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  d6249d06
- 新增dpsk-v3.1-awq的支持 · 2cfcb974
  zhuwenwen authored Sep 11, 2025
  
  2cfcb974