Commits · 961cce86addec7274d13cf8671e60922855f6728 · OpenDAS / vllm_cscc

"tests/v1/engine/test_parallel_sampling.py" did not exist on "97d1c99302df6f7eadc0d0b32ec174db69cb4421"

31 Jul, 2025 2 commits
- 去除多余的w4a8参数 · 961cce86
  gaoqiong authored Jul 31, 2025
  
  961cce86
- 增加fused moe文件中w4a8的相关修改 · f5a7f12c
  gaoqiong authored Jul 31, 2025
  
  f5a7f12c
29 Jul, 2025 1 commit
- fix: 修复W8A8读config路径错误，删除int8_utils.py文件 · 7e5fb6fe
  jujl1 authored Jul 29, 2025
  
  7e5fb6fe
28 Jul, 2025 3 commits
- fix: 修复W8A8INT8读config问题 · c6187ade
  jujl1 authored Jul 28, 2025
  
  c6187ade
- 修改W4A8 以及W8A8量化量化092接口 · 7017f30c
  gaoqiong authored Jul 28, 2025
  
  7017f30c
- add v1 cpm support · 98958aed
  zhuwenwen authored Jul 28, 2025
  
  98958aed
26 Jul, 2025 1 commit
- 增加moe awq-marlin的支持 · c2bcb0ab
  yangql authored Jul 26, 2025
  
  c2bcb0ab
25 Jul, 2025 6 commits
- 修复block-wise的错误参数，以及增加dpsk-fp16-dtype的性能选择功能 · cb37537e
  yangql authored Jul 25, 2025
  
  cb37537e
- Merge branch 'v0.9.2-dev-wm' into 'v0.9.2-dev' · 4d479e7e
  zhuwenwen authored Jul 25, 2025
```
[fix]解决deepseek报错

See merge request dcutoolkit/deeplearing/vllm!162
```
  4d479e7e
- [fix]解决deepseek报错 · cf4c84b5
  王敏 authored Jul 25, 2025
  
  cf4c84b5
- [Fix] MLA only supports decode-only full CUDAGraph capture. Make sure all... · b35835a1
  zhuwenwen authored Jul 25, 2025
```
[Fix] MLA only supports decode-only full CUDAGraph capture. Make sure all cudagraph capture sizes <= max_num_seq.
```
  b35835a1
- set VLLM_USE_V1=1 · e4a84fdc
  zhuwenwen authored Jul 25, 2025
  
  e4a84fdc
- Triton-fused DeepseekScalingRotaryEmbedding · 7151fd54
  zhuwenwen authored Jul 25, 2025
  
  7151fd54
24 Jul, 2025 4 commits
- support full_cuda_graph · 8da1c576
  zhuwenwen authored Jul 24, 2025
  
  8da1c576
- update fa v1 interface · f9408aff
  zhuwenwen authored Jul 24, 2025
  
  f9408aff
- add fused_experts_impl of int8 · 41e6d686
  zhuwenwen authored Jul 24, 2025
  
  41e6d686
- skip fp8 and ActivationQuantFusionPass · a246d08c
  zhuwenwen authored Jul 24, 2025
  
  a246d08c
22 Jul, 2025 2 commits
- update max_seq_len_to_capture to option int · d560429c
  zhuwenwen authored Jul 22, 2025
  
  d560429c
- update max_seq_len_to_capture to int · 988fc31c
  zhuwenwen authored Jul 22, 2025
  
  988fc31c
21 Jul, 2025 2 commits
- update max_seq_len_to_capture · 72273242
  zhuwenwen authored Jul 21, 2025
  
  72273242
- update List · 267cc5ff
  zhuwenwen authored Jul 21, 2025
  
  267cc5ff
18 Jul, 2025 4 commits
- update fused_moe.py · b1babea8
  zhuwenwen authored Jul 18, 2025
  
  b1babea8
- fix run error · 5c004388
  zhuwenwen authored Jul 18, 2025
  
  5c004388
- remove unused sgl_moe_align_block_size · fd46df33
  zhuwenwen authored Jul 18, 2025
  
  fd46df33
- Merge tag 'v0.9.2' into v0.9.2-dev · a40a133c
  zhuwenwen authored Jul 18, 2025
  
  a40a133c
17 Jul, 2025 5 commits
- add fa-pa benchmark · 1a9a61d7
  zhuwenwen authored Jul 17, 2025
  
  1a9a61d7
- 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
16 Jul, 2025 1 commit
- support rms_norm_dynamic_per_token_quant · 91feb245
  zhuwenwen authored Jul 16, 2025
  
  91feb245
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 3 commits
- Merge branch 'v0.9.1-dev' of http://10.16.6.30/dcutoolkit/deeplearing/vllm into v0.9.1-dev · 7bf6c98f
  zhuwenwen authored Jul 14, 2025
  
  7bf6c98f
- update blockwise_int8.py · 5d35160b
  zhuwenwen authored Jul 14, 2025
  
  5d35160b
- Merge branch 'v0.9.1-dev-wm' into 'v0.9.1-dev' · 0a9b4a7f
  zhuwenwen authored Jul 14, 2025
```
[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

See merge request dcutoolkit/deeplearing/vllm!158
```
  0a9b4a7f
12 Jul, 2025 4 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
- Merge remote-tracking branch 'origin/v0.9.1-dev_whl' into v0.9.1-dev · c73d2c3c
  zhuwenwen authored Jul 12, 2025
  
  c73d2c3c
- fix rmsnrom precision · ac61d69b
  wanghl6 authored Jul 12, 2025
  
  ac61d69b
11 Jul, 2025 1 commit
- Merge branch 'v0.9.1-dev_lzg' into 'v0.9.1-dev' · e522713c
  zhuwenwen authored Jul 11, 2025
```
tbo debug fix error

See merge request dcutoolkit/deeplearing/vllm!156
```
  e522713c