Hugging Face's logo

Spaces:
arcee-ai
/
Benchmarks
Running

App Files Community
Benchmarks
File size: 3,235 Bytes 
10318fd
 
 
 
 
86d92de
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f011c00
 
 
 
 
 
 
 
86d92de
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
10318fd
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
 
"""Module containing performance results for the Virtuoso-Medium model."""

results_virtuoso_medium = {
    "name": "Virtuoso-Medium",
    "modelType": "Qwen2.5 32B",
    "configurations": [
        {
            "instanceType": "r8g.4xlarge",
            "quantization": "Q4_0_4_8",
            "container": "llama.cpp 11/27/24",
            "status": "OK",
            "tokensPerSecond": "10.5",
            "notes": "-fa",
        },
        {
            "instanceType": "g5.12xlarge",
            "quantization": "none",
            "container": "LMI 0.30+vLLM 0.6.2",
            "status": "OK",
            "tokensPerSecond": "23",
            "notes": '"OPTION_MAX_MODEL_LEN": "16384",\n"TENSOR_PARALLEL_DEGREE": "max",',
        },
        {
            "instanceType": "g6.12xlarge",
            "quantization": "none",
            "container": "LMI 0.30+vLLM 0.6.2",
            "status": "OK",
            "tokensPerSecond": "14",
            "notes": '"OPTION_MAX_MODEL_LEN": "16384",\n"TENSOR_PARALLEL_DEGREE": "max",',
        },
        {
            "instanceType": "g6e.12xlarge",
            "quantization": "none",
            "container": "vLLM0.6.4.post1",
            "status": "OK",
            "tokensPerSecond": "36",
            "notes": "--tensor-parallel-size 4 --max-model-len 16384",
        },
        {
            "instanceType": "g6e.12xlarge (2 GPUs)",
            "quantization": "none",
            "container": "vLLM 0.6.4.post1",
            "status": "OK",
            "tokensPerSecond": "21",
            "notes": "--tensor-parallel-size 2 --max-model-len 16384",
        },
        {
            "instanceType": "g6e.48xlarge",
            "quantization": "none",
            "container": "LMI 0.30+vLLM 0.6.2",
            "status": "OK",
            "tokensPerSecond": "48",
            "notes": "",
        },
        {
            "instanceType": "p4d.24xlarge",
            "quantization": "none",
            "container": "LMI 0.30+vLLM 0.6.2",
            "status": "OK",
            "tokensPerSecond": "72.5",
            "notes": '"OPTION_MAX_MODEL_LEN": "32768",\n"TENSOR_PARALLEL_DEGREE": "max",',
        },
        {
            "instanceType": "p5.48xlarge",
            "quantization": "none",
            "container": "vLLM 0.6.4.post1",
            "status": "OK",
            "tokensPerSecond": "117",
            "notes": "--tensor-parallel-size 8",
        },
        {
            "instanceType": "p5.48xlarge (4 GPUs)",
            "quantization": "none",
            "container": "vLLM 0.6.4.post1",
            "status": "OK",
            "tokensPerSecond": "88",
            "notes": "--tensor-parallel-size 4",
        },
        {
            "instanceType": "p5.48xlarge (2 GPUs)",
            "quantization": "none",
            "container": "vLLM 0.6.4.post1",
            "status": "OK",
            "tokensPerSecond": "58",
            "notes": "--tensor-parallel-size 2",
        },
        {
            "instanceType": "p5.48xlarge (1 GPU)",
            "quantization": "none",
            "container": "vLLM 0.6.4.post1",
            "status": "OK",
            "tokensPerSecond": "38",
            "notes": "--tensor-parallel-size 1",
        },
    ],
}