Commits · c8a63b38329805dcffdd7a6f580a18616972dc45 · OpenDAS / vllm_cscc

25 Mar, 2025 1 commit
- add custom allreduce check · c8a63b38
  xiabo authored Mar 25, 2025
  
  c8a63b38
22 Mar, 2025 1 commit
- add custom allreduce cudagraph · a8c92908
  xiabo authored Mar 22, 2025
  
  a8c92908
17 Mar, 2025 1 commit
- add custom allreduce · 2d6bccd9
  xiabo authored Mar 17, 2025
  
  2d6bccd9
15 Mar, 2025 6 commits
- update fused_moe.py · 8a7c4451
  zhuwenwen authored Mar 15, 2025
  
  8a7c4451
- update qwen2-vl and qwen2.5-vl layout · 37ef8dd3
  zhuwenwen authored Mar 15, 2025
  
  37ef8dd3
- set llama and qwen tn layout · b845811e
  zhuwenwen authored Mar 15, 2025
  
  b845811e
- 调整pa tc和非tc调用关系 · 0d2ccb8d
  zhangshao authored Mar 15, 2025
  
  0d2ccb8d
- 支持fusemoe对int4的scale合zero合并读取操作 · 50424a23
  yangql authored Mar 15, 2025
  
  50424a23
- 支持fusemoe对int4的scale合zero合并读取操作 · f78a0510
  yangql authored Mar 15, 2025
  
  f78a0510
14 Mar, 2025 3 commits
- 增加线性int8 gemm配置 · 268d8a77
  gaoqiong authored Mar 14, 2025
  
  268d8a77
- [feat]add gen_mla_pa_tables.py and optimize triton config configuration · 71c60bd5
  zhuwenwen authored Mar 14, 2025
  
  71c60bd5
- 增加deepseek v3 block-int8量化支持 · 92545504
  gaoqiong authored Mar 14, 2025
  
  92545504
13 Mar, 2025 1 commit
- update mla to obtain the optimal configuration from config · 146eb9d3
  zhuwenwen authored Mar 13, 2025
  
  146eb9d3
12 Mar, 2025 1 commit
- [feat]添加BW deepseek_v3 fused_moe configs · ef27eec0
  王敏 authored Mar 12, 2025
  
  ef27eec0
11 Mar, 2025 1 commit
- [feat]优化fused_moe tuning，上传K100_AI的configs · 79a9bc29
  王敏 authored Mar 11, 2025
  
  79a9bc29
06 Mar, 2025 1 commit
- update mla optest · e4a9c2cd
  zhuwenwen authored Mar 06, 2025
  
  e4a9c2cd
03 Mar, 2025 2 commits
- update triton mla · 52121d00
  zhuwenwen authored Mar 03, 2025
  
  52121d00
- 解决PA部分size计算错误的问题 · 228a714a
  zhangshao authored Mar 03, 2025
  
  228a714a
28 Feb, 2025 3 commits
- add VLLM_USE_TRITON_OPT_MLA to use optimized MLA attention · 40083064
  zhuwenwen authored Feb 28, 2025
  
  40083064
- 删除vllm的w8a8 config · 6fd8c21b
  gaoqiong authored Feb 28, 2025
  
  6fd8c21b
- 修复w8a8找不到对应config的bug · 044076a1
  gaoqiong authored Feb 28, 2025
  
  044076a1
26 Feb, 2025 6 commits
- [fix]修复ds3量化代码引入的报错 · bcce8974
  王敏 authored Feb 26, 2025
  
  bcce8974
- [feat]合入ds3 MTP功能 · 1fbdf957
  王敏 authored Feb 26, 2025
  
  1fbdf957
- 解决cudagraph计算错误的问题 · d365012a
  zhangshao authored Feb 26, 2025
  
  d365012a
- update deepseek-v2 · 96802ca0
  zhuwenwen authored Feb 26, 2025
  
  96802ca0
- add cudagraph · 4d4c6fe3
  zhuwenwen authored Feb 26, 2025
  
  4d4c6fe3
- allow to set --enforce-eager · 537e2d9c
  zhuwenwen authored Feb 26, 2025
  
  537e2d9c
25 Feb, 2025 1 commit
- update version · 76621631
  zhuwenwen authored Feb 25, 2025
  
  76621631
22 Feb, 2025 3 commits
- update qwen2_vl nn lyaout · 437e6aef
  zhuwenwen authored Feb 22, 2025
  
  437e6aef
- set enforce eager · 7f881746
  zhuwenwen authored Feb 22, 2025
  
  7f881746
- support qwen2_5-vl · 0a89b8a3
  zhuwenwen authored Feb 22, 2025
  
  0a89b8a3
20 Feb, 2025 2 commits
- 适配deepseekv3\v2 moe awq的推理支持 · 47bd229c
  yangql authored Feb 20, 2025
  
  47bd229c
- skip fp8 fusion · 4a734b9d
  zhuwenwen authored Feb 20, 2025
  
  4a734b9d
19 Feb, 2025 2 commits
- add deepseek-v2 tp2 fused_moe_config · 177520a9
  zhuwenwen authored Feb 19, 2025
  
  177520a9
- update lm_head tn layout for awq · f0e7d72d
  zhuwenwen authored Feb 19, 2025
  
  f0e7d72d
18 Feb, 2025 3 commits
- add deepseek-v2 tp1 fused_moe_config · f0ac5f23
  zhuwenwen authored Feb 18, 2025
  
  f0ac5f23
- 增加lm_nn的量化控控控制,控制为tn · 41d0696e
  gaoqiong authored Feb 18, 2025
  
  41d0696e
- 添加w8a8 关于0.7.2的修改 · 8c61b2ed
  gaoqiong authored Feb 18, 2025
  
  8c61b2ed
17 Feb, 2025 1 commit
- [fix]1.解决deepseekv2/v3 nn layout加载报错问题;2.上传ep1和ep2的fused_moe config · 249fca2a
  王敏 authored Feb 17, 2025
  
  249fca2a
13 Feb, 2025 1 commit
- support deepseek_v2 nn layout · a8b2d878
  zhuwenwen authored Feb 13, 2025
  
  a8b2d878