Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in
Toggle navigation
Menu
Open sidebar
OpenDAS
vllm_cscc
Commits
0a130908
Commit
0a130908
authored
Feb 28, 2025
by
zhuwenwen
Browse files
Merge branch 'v0.7.2-dev-w8a8' into 'v0.7.2-dev'
V0.7.2 dev w8a8 See merge request dcutoolkit/deeplearing/vllm!76
parents
6b7651af
6fd8c21b
Changes
37
Hide whitespace changes
Inline
Side-by-side
Showing
20 changed files
with
0 additions
and
8387 deletions
+0
-8387
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_12288_4096_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_12288_4096_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_1280_8192_K100_AI.json
...ers/quantization/configs/w8a8/W8A8_1280_8192_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_13824_5120_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_13824_5120_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_14336_8192_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_14336_8192_K100_AI.json
+0
-427
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_15360_5120_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_15360_5120_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_22016_4096_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_22016_4096_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_2560_8192_K100_AI.json
...ers/quantization/configs/w8a8/W8A8_2560_8192_K100_AI.json
+0
-436
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_27648_5120_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_27648_5120_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_28672_4096_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_28672_4096_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_28672_8192_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_28672_8192_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_32000_4096_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_32000_4096_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_3584_18944_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_3584_18944_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_3584_3584_K100_AI.json
...ers/quantization/configs/w8a8/W8A8_3584_3584_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_37888_3584_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_37888_3584_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_4096_11008_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_4096_11008_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_4096_14336_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_4096_14336_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_4096_4096_K100_AI.json
...ers/quantization/configs/w8a8/W8A8_4096_4096_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_4608_3584_K100_AI.json
...ers/quantization/configs/w8a8/W8A8_4608_3584_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_5120_13824_K100_AI.json
...rs/quantization/configs/w8a8/W8A8_5120_13824_K100_AI.json
+0
-418
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_5120_2560_K100_AI.json
...ers/quantization/configs/w8a8/W8A8_5120_2560_K100_AI.json
+0
-418
No files found.
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_12288_4096_K100_AI.json
deleted
100755 → 0
View file @
6b7651af
{
"12288_4096"
:
{
"20"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"24"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"28"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"32"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"36"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"40"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"44"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"48"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"72"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"80"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"88"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"96"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"1"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"2"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"3"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"4"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"9"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"10"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"14"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"15"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"256"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_1280_8192_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"1280_8192"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"3"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"4"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"9"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"14"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"20"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"24"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"28"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"32"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"40"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"44"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"48"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"72"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"80"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"88"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"96"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"112"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"120"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"128"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"256"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"512"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"1024"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"2048"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
1
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_13824_5120_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"13824_5120"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"3"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"4"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"5"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"6"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"7"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"8"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"9"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"11"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"12"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"13"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"14"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"15"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"16"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"20"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"24"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"28"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"32"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"40"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"44"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"48"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"80"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"88"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"96"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"256"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_14336_8192_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"14336_8192"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
2
},
"2"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"3"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"4"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
2
},
"5"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
2
},
"6"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"7"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"8"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"9"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
2
},
"10"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
2
},
"11"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"12"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
2
},
"13"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"14"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
2
},
"15"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"16"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
2
},
"17"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"20"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"24"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"28"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"32"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"40"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"44"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"48"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"80"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"88"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"96"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"256"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_15360_5120_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"15360_5120"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"2"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"3"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"9"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"10"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"14"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"20"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"24"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"28"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"32"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"40"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"44"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"48"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"52"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"56"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"60"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"80"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"88"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"96"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"256"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"512"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"1024"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_22016_4096_K100_AI.json
deleted
100755 → 0
View file @
6b7651af
{
"22016_4096"
:
{
"20"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"24"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"28"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"32"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"36"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"40"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"44"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"48"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"56"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"60"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"64"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"72"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"80"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"88"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"96"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"1"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"3"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"4"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"5"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"6"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"7"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"8"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"9"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"11"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"12"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"13"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"14"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"15"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"16"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"256"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_2560_8192_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"2560_8192"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"3"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"4"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"9"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"14"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"17"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"20"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"24"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"28"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"32"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"40"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"44"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"48"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"80"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"88"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"96"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"136"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"144"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"152"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"160"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"256"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"512"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2048"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"4096"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"16384"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_27648_5120_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"27648_5120"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
4
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"3"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"9"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"10"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"14"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"20"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"24"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"28"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"32"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"40"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"44"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"48"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"80"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"88"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"96"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"256"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_28672_4096_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"28672_4096"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"2"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"3"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"5"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"6"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"7"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"9"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"10"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"11"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"12"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"13"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"14"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"15"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"16"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"20"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"24"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"28"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"32"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"40"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"44"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"48"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"72"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"80"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"88"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"96"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"256"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_28672_8192_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"28672_8192"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"2"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"3"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"5"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"6"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"7"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"9"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"10"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"11"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"12"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"13"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"14"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"15"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"16"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"20"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"24"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"28"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"32"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"40"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"44"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"48"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"52"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"56"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"60"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"64"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"72"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"80"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"88"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"96"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"256"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_32000_4096_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"32000_4096"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"2"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"3"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"5"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"6"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"7"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"8"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"9"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"11"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"12"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"13"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"14"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"15"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"16"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"20"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"24"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"28"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"32"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"40"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"44"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"48"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"72"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"80"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"88"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"96"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"256"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_3584_18944_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"3584_18944"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"3"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"4"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"6"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"7"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"8"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"9"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"14"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"20"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"24"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"28"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"32"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"40"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"44"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"48"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"52"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"56"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"60"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"80"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"88"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"96"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"256"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"1024"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_3584_3584_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"3584_3584"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"2"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"3"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"4"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"9"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"14"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"20"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"24"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"28"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"32"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"40"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"44"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"48"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"80"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"88"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"96"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"256"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"512"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"1024"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_37888_3584_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"37888_3584"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"3"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"4"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"5"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"9"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"14"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"20"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"24"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"28"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"32"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"40"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"44"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"48"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"80"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"88"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"96"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"256"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_4096_11008_K100_AI.json
deleted
100755 → 0
View file @
6b7651af
{
"4096_11008"
:
{
"20"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"24"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"28"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"32"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"40"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"44"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"48"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"52"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"80"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"88"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"96"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"1"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"2"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"3"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"4"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
2
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"9"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"11"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"14"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"16"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"256"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"1024"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_4096_14336_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"4096_14336"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"2"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"3"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"4"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"9"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"14"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
2
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"20"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"24"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"28"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"32"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"40"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"44"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
16
},
"48"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
16
},
"52"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
16
},
"56"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"80"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"88"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
16
},
"96"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
512
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
16
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
16
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
16
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
16
},
"256"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"1024"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_4096_4096_K100_AI.json
deleted
100755 → 0
View file @
6b7651af
{
"4096_4096"
:
{
"20"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"24"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"28"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"32"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"40"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"44"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"48"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"80"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"88"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"96"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"1"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"2"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"3"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"4"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"9"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"14"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"256"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"512"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_4608_3584_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"4608_3584"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"2"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"3"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"4"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"9"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"14"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"20"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"24"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"28"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"32"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"40"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"44"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"48"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"80"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"88"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"96"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"256"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"512"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2048"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"4096"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
1
,
"num_warps"
:
4
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_5120_13824_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"5120_13824"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"2"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"3"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"4"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"9"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"10"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"14"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"20"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"24"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"28"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"32"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"40"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"44"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"48"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"52"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"56"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
2
,
"num_stages"
:
2
,
"num_warps"
:
8
},
"80"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"88"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"96"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"256"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"512"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"1024"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"2048"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
vllm/model_executor/layers/quantization/configs/w8a8/W8A8_5120_2560_K100_AI.json
deleted
100644 → 0
View file @
6b7651af
{
"5120_2560"
:
{
"1"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"2"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"3"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"4"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"5"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"6"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"7"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"8"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"9"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"10"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"11"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"12"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"13"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"14"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
2
},
"15"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"16"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
16
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
2
},
"20"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"24"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"28"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"32"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"36"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"40"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"44"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"48"
:
{
"BLOCK_SIZE_M"
:
16
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"52"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"56"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"60"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"64"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
32
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"72"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"80"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"88"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"96"
:
{
"BLOCK_SIZE_M"
:
32
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"104"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"112"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"120"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"128"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
16
},
"136"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"144"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
},
"152"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"160"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
256
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
16
},
"256"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
64
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"512"
:
{
"BLOCK_SIZE_M"
:
64
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
4
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"1024"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
2
,
"num_warps"
:
4
},
"2048"
:
{
"BLOCK_SIZE_M"
:
128
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
128
,
"GROUP_SIZE_M"
:
8
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"4096"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
128
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
4
},
"8192"
:
{
"BLOCK_SIZE_M"
:
256
,
"BLOCK_SIZE_N"
:
256
,
"BLOCK_SIZE_K"
:
64
,
"GROUP_SIZE_M"
:
2
,
"SPLIT_K"
:
1
,
"num_stages"
:
0
,
"num_warps"
:
8
}
}
}
\ No newline at end of file
Prev
1
2
Next
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment