Commits · fd46df331de3dd1832e601cf589204f424312241 · OpenDAS / vllm_cscc · GitLab

17 Jul, 2025 4 commits
- update adpaptive block_size · 937a3ec1
  zhuwenwen authored Jul 17, 2025
  
  937a3ec1
- update default block_size · 4a84480f
  zhuwenwen authored Jul 17, 2025
  
  4a84480f
- set default block_size and pa · b61f7a69
  zhuwenwen authored Jul 17, 2025
  
  b61f7a69
- add VLLM_USE_PA_PRINT_PARAM to print fa-pa size · 1092a467
  zhuwenwen authored Jul 17, 2025
  
  1092a467
15 Jul, 2025 1 commit
- 修复在fp16下的数值越界导致的精度问题 · 18f030d9
  zhuwenwen authored Jul 15, 2025
  
  18f030d9
14 Jul, 2025 1 commit
- update blockwise_int8.py · 5d35160b
  zhuwenwen authored Jul 14, 2025
  
  5d35160b
12 Jul, 2025 2 commits
- [fix]1.优化mtp代码，解决prefix-caching不兼容问题; 2.修复v1 engine从tokenizer config中读取max_model_len导致长文本输入报错 · b16f183c
  王敏 authored Jul 12, 2025
  
  b16f183c
- remove qk contiguous · e34d3444
  zhuwenwen authored Jul 12, 2025
  
  e34d3444
11 Jul, 2025 2 commits
- tbo debug fix error · 957096ce
  lizhigong authored Jul 11, 2025
  
  957096ce
- add qk contiguous · bcf6be4c
  zhuwenwen authored Jul 11, 2025
  
  bcf6be4c
10 Jul, 2025 5 commits
- add apex rmsnorm · 3e0a595d
  zhuwenwen authored Jul 10, 2025
  
  3e0a595d
- fix zero overhead to support chunk prefill · a495fc3b
  zhuwenwen authored Jul 10, 2025
  
  a495fc3b
- add two batch overlap decude support muti-stream cuda-graph · fe1c4016
  zhuwenwen authored Jul 10, 2025
  
  fe1c4016
- update bw vdim128 · d805c59c
  zhuwenwen authored Jul 10, 2025
  
  d805c59c
- [fix]修复091 v0 engine mtp报错问题 · ac182341
  王敏 authored Jul 10, 2025
  
  ac182341
09 Jul, 2025 1 commit
- [fix]修复零消耗引入的cudagraph模式数据准备cpu耗时问题 · cd6239dc
  zhuwenwen authored Jul 09, 2025
  
  cd6239dc
07 Jul, 2025 5 commits
- 还原误删代码 · f0acced0
  王敏 authored Jul 07, 2025
  
  f0acced0
- 删掉无效代码 · ee9ee960
  王敏 authored Jul 07, 2025
  
  ee9ee960
- [feat]支持v1 engine flashmla和mtp同时开启 · ac4dedb1
  王敏 authored Jul 07, 2025
  
  ac4dedb1
- 修改激活量化算子实现方式 · b2fa85ce
  zhuwenwen authored Jul 07, 2025
  
  b2fa85ce
- update vllm_flash_attn_varlen_func · ef6c0877
  zhuwenwen authored Jul 07, 2025
  
  ef6c0877
06 Jul, 2025 11 commits
- Revert "[V0 deprecation] Remove V0 CPU/XPU/TPU backends (#20412)" · a5dd03c1
  simon-mo authored Jul 06, 2025
```
This reverts commit e202dd27.
```
  a5dd03c1
- [Bugfix] Add `use_cross_encoder` flag to use correct activation in `ClassifierPooler` (#20527) · c18b3b8e
  Cyrus Leung authored Jul 07, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  c18b3b8e
- [BugFix][Spec Decode] Fix spec token ids in model runner (#20530) · 9528e3a0
  Woosuk Kwon authored Jul 06, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
  9528e3a0
- [V1] Support any head size for FlexAttention backend (#20467) · 9fb52e52
  Cyrus Leung authored Jul 07, 2025
```
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  9fb52e52
- [V0 deprecation] Remove V0 CPU/XPU/TPU backends (#20412) · e202dd27
  Woosuk Kwon authored Jul 06, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Signed-off-by: jiang1.li <jiang1.li@intel.com>
Co-authored-by: Li, Jiang <jiang1.li@intel.com>
```
  e202dd27
- [Misc] call the pre-defined func (#20518) · 43813e63
  Reid authored Jul 06, 2025
```
Signed-off-by: reidliu41 <reid201711@gmail.com>
```
  43813e63
- [Benchmark] Add support for multiple batch size benchmark through CLI in... · cede942b
  Brayden Zhong authored Jul 06, 2025
```
[Benchmark] Add support for multiple batch size benchmark through CLI in `benchmark_moe.py` (#20516)
Signed-off-by: Brayden Zhong <b8zhong@uwaterloo.ca>
```
  cede942b
- [Frontend] Support image object in llm.chat (#19635) · fe1e9248
  Flora Feng authored Jul 05, 2025
```
Signed-off-by: sfeng33 <4florafeng@gmail.com>
Signed-off-by: Flora Feng <4florafeng@gmail.com>
```
  fe1e9248
- [TPU][Bugfix] fix the MoE OOM issue (#20339) · 4548c03c
  Chengji Yao authored Jul 05, 2025
```
Signed-off-by: Chengji Yao <chengjiyao@google.com>
```
  4548c03c
- [Bugfix] Fix missing per_act_token parameter in compressed_tensors_moe (#20509) · 43287082
  Lucia Fang authored Jul 06, 2025
```
Signed-off-by: Lu Fang <fanglu@fb.com>
```
  43287082
- [Misc] remove unused import (#20517) · 8d763cb8
  Reid authored Jul 06, 2025
```
Signed-off-by: reidliu41 <reid201711@gmail.com>
```
  8d763cb8
05 Jul, 2025 6 commits
- [Misc] Add logger.exception for TPU information collection failures (#20510) · cf4cd539
  Reid authored Jul 05, 2025
```
Signed-off-by: reidliu41 <reid201711@gmail.com>
```
  cf4cd539
- [v1] Re-add fp32 support to v1 engine through FlexAttention (#19754) · 32c9be22
  Isotr0py authored Jul 05, 2025
```
Signed-off-by: Isotr0py <2037008807@qq.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
```
  32c9be22
- Fix unknown attribute of topk_indices_dtype in CompressedTensorsW8A8Fp8MoECutlassMethod (#20507) · 8aeaa910
  Lucia Fang authored Jul 05, 2025
```
Co-authored-by: Lucia (Lu) Fang <fanglu@meta.com>
```
  8aeaa910
- [Misc] Remove the unused LoRA test code (#20494) · 906e05d8
  Jee Jee Li authored Jul 05, 2025
```
Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>
```
  906e05d8
- [Misc] Add security warning for development mode endpoints (#20508) · 7e908704
  Reid authored Jul 05, 2025
```
Signed-off-by: reidliu41 <reid201711@gmail.com>
```
  7e908704
- update fix_functionalization.py · 2ced1fb2
  zhuwenwen authored Jul 05, 2025
  
  2ced1fb2
04 Jul, 2025 2 commits
- [CI Bugfix] Fix pre-commit failures on main (#20502) · c108781c
  Michael Goin authored Jul 05, 2025
  
  c108781c
- [feat]: CUTLASS block scaled group gemm for SM100 (#19757) · 3d184b95
  Duncan Moss authored Jul 04, 2025
```
Signed-off-by: Duncan Moss <djm.moss@gmail.com>
Co-authored-by: Duncan Moss <dmoss@nvidia.com>
```
  3d184b95