Hugging Face's logo

Datasets:
optimum-benchmark
/
cpu

Dataset card Viewer Files Community
IlyasMoutawwakil HF staff commited on
Commit
a6d5405
•
1 Parent(s): a7dee38

Upload cpu_inference_transformers_text-generation_openai-community/gpt2/benchmark_report.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. cpu_inference_transformers_text-generation_openai-community/gpt2/benchmark_report.json +87 -93
cpu_inference_transformers_text-generation_openai-community/gpt2/benchmark_report.json CHANGED
@@ -2,7 +2,7 @@
2
  "load": {
3
  "memory": {
4
  "unit": "MB",
5
- "max_ram": 1110.257664,
6
  "max_global_vram": null,
7
  "max_process_vram": null,
8
  "max_reserved": null,
@@ -11,31 +11,31 @@
11
  "latency": {
12
  "unit": "s",
13
  "count": 1,
14
- "total": 4.741065211999995,
15
- "mean": 4.741065211999995,
16
  "stdev": 0.0,
17
- "p50": 4.741065211999995,
18
- "p90": 4.741065211999995,
19
- "p95": 4.741065211999995,
20
- "p99": 4.741065211999995,
21
  "values": [
22
- 4.741065211999995
23
  ]
24
  },
25
  "throughput": null,
26
  "energy": {
27
  "unit": "kWh",
28
- "cpu": 6.719550953971016e-05,
29
- "ram": 2.8085806865067295e-06,
30
  "gpu": 0,
31
- "total": 7.000409022621688e-05
32
  },
33
  "efficiency": null
34
  },
35
  "prefill": {
36
  "memory": {
37
  "unit": "MB",
38
- "max_ram": 963.878912,
39
  "max_global_vram": null,
40
  "max_process_vram": null,
41
  "max_reserved": null,
@@ -43,52 +43,50 @@
43
  },
44
  "latency": {
45
  "unit": "s",
46
- "count": 15,
47
- "total": 0.6285618450000925,
48
- "mean": 0.04190412300000616,
49
- "stdev": 0.0005770860447191328,
50
- "p50": 0.041817378000018834,
51
- "p90": 0.042688608799983284,
52
- "p95": 0.04302177559999905,
53
- "p99": 0.043183375919992384,
54
  "values": [
55
- 0.04205661399998917,
56
- 0.04231871599995429,
57
- 0.04141679200000681,
58
- 0.04139818800001649,
59
- 0.042935204000002614,
60
- 0.04322377599999072,
61
- 0.04174381600000743,
62
- 0.04222742499996457,
63
- 0.042024744000002556,
64
- 0.04170702800001891,
65
- 0.04189196500004755,
66
- 0.041817378000018834,
67
- 0.04149982000001273,
68
- 0.04107729700001528,
69
- 0.04122308200004454
70
  ]
71
  },
72
  "throughput": {
73
  "unit": "tokens/s",
74
- "value": 47.72800041656297
75
  },
76
  "energy": {
77
  "unit": "kWh",
78
- "cpu": 1.5938092867533368e-06,
79
- "ram": 6.660810903831589e-08,
80
  "gpu": 0.0,
81
- "total": 1.660417395791653e-06
82
  },
83
  "efficiency": {
84
  "unit": "tokens/kWh",
85
- "value": 1204516.4095901567
86
  }
87
  },
88
  "decode": {
89
  "memory": {
90
  "unit": "MB",
91
- "max_ram": 964.4032,
92
  "max_global_vram": null,
93
  "max_process_vram": null,
94
  "max_reserved": null,
@@ -96,81 +94,77 @@
96
  },
97
  "latency": {
98
  "unit": "s",
99
- "count": 15,
100
- "total": 0.3874191449998534,
101
- "mean": 0.02582794299999023,
102
- "stdev": 0.0007405831397145795,
103
- "p50": 0.025715652999963368,
104
- "p90": 0.02617861680001852,
105
- "p95": 0.026861670100004172,
106
- "p99": 0.027994397219970324,
107
  "values": [
108
- 0.025751395999975557,
109
- 0.02586119300002565,
110
- 0.026027754999972785,
111
- 0.02557163900002024,
112
- 0.02625485200002231,
113
- 0.025560767999991185,
114
- 0.025383235999981935,
115
- 0.026064264000012827,
116
- 0.02550354099997776,
117
- 0.02827757899996186,
118
- 0.02595737299998291,
119
- 0.025475631999995585,
120
- 0.025715652999963368,
121
- 0.025017731999980697,
122
- 0.024996531999988747
123
  ]
124
  },
125
  "throughput": {
126
  "unit": "tokens/s",
127
- "value": 38.717756191438795
128
  },
129
  "energy": {
130
  "unit": "kWh",
131
- "cpu": 9.596641262372338e-07,
132
- "ram": 4.0107503515113744e-08,
133
  "gpu": 0.0,
134
- "total": 9.99771629752347e-07
135
  },
136
  "efficiency": {
137
  "unit": "tokens/kWh",
138
- "value": 1000228.4224125359
139
  }
140
  },
141
  "per_token": {
142
  "memory": null,
143
  "latency": {
144
  "unit": "s",
145
- "count": 15,
146
- "total": 0.38240176999994446,
147
- "mean": 0.02549345133332963,
148
- "stdev": 0.0006976654720593424,
149
- "p50": 0.025409474000014143,
150
- "p90": 0.025838972800011108,
151
- "p95": 0.026470589999991034,
152
- "p99": 0.027517633199992134,
153
  "values": [
154
- 0.025409474000014143,
155
- 0.025532496000039373,
156
- 0.025714666999988367,
157
- 0.025250615999993897,
158
- 0.025909673999990446,
159
- 0.02523040799997034,
160
- 0.02506068000002415,
161
- 0.025732921000042097,
162
- 0.0251756150000233,
163
- 0.027779393999992408,
164
- 0.025615390999973897,
165
- 0.025159086999963165,
166
- 0.02541822399996363,
167
- 0.024717326999962097,
168
- 0.024695796000003156
169
  ]
170
  },
171
  "throughput": {
172
  "unit": "tokens/s",
173
- "value": 39.225759859851536
174
  },
175
  "energy": null,
176
  "efficiency": null
 
2
  "load": {
3
  "memory": {
4
  "unit": "MB",
5
+ "max_ram": 1099.759616,
6
  "max_global_vram": null,
7
  "max_process_vram": null,
8
  "max_reserved": null,
 
11
  "latency": {
12
  "unit": "s",
13
  "count": 1,
14
+ "total": 5.033019491999994,
15
+ "mean": 5.033019491999994,
16
  "stdev": 0.0,
17
+ "p50": 5.033019491999994,
18
+ "p90": 5.033019491999994,
19
+ "p95": 5.033019491999994,
20
+ "p99": 5.033019491999994,
21
  "values": [
22
+ 5.033019491999994
23
  ]
24
  },
25
  "throughput": null,
26
  "energy": {
27
  "unit": "kWh",
28
+ "cpu": 6.804323328865899e-05,
29
+ "ram": 2.8438522789963372e-06,
30
  "gpu": 0,
31
+ "total": 7.088708556765532e-05
32
  },
33
  "efficiency": null
34
  },
35
  "prefill": {
36
  "memory": {
37
  "unit": "MB",
38
+ "max_ram": 952.991744,
39
  "max_global_vram": null,
40
  "max_process_vram": null,
41
  "max_reserved": null,
 
43
  },
44
  "latency": {
45
  "unit": "s",
46
+ "count": 13,
47
+ "total": 0.62556531000007,
48
+ "mean": 0.04812040846154385,
49
+ "stdev": 0.0010184667700024917,
50
+ "p50": 0.04828216200002089,
51
+ "p90": 0.049286218399998916,
52
+ "p95": 0.049384615800005346,
53
+ "p99": 0.0494577111600131,
54
  "values": [
55
+ 0.047916334000007055,
56
+ 0.049475985000015044,
57
+ 0.04727043700000877,
58
+ 0.04670981900000015,
59
+ 0.0482919550000247,
60
+ 0.04901934800000163,
61
+ 0.04828216200002089,
62
+ 0.049323702999998886,
63
+ 0.04773996800000191,
64
+ 0.04867056800000569,
65
+ 0.04913627999999903,
66
+ 0.04587774600000216,
67
+ 0.04785100499998407
 
 
68
  ]
69
  },
70
  "throughput": {
71
  "unit": "tokens/s",
72
+ "value": 41.562406969141385
73
  },
74
  "energy": {
75
  "unit": "kWh",
76
+ "cpu": 1.8891113775747795e-06,
77
+ "ram": 7.894644336001225e-08,
78
  "gpu": 0.0,
79
+ "total": 1.9680578209347916e-06
80
  },
81
  "efficiency": {
82
  "unit": "tokens/kWh",
83
+ "value": 1016230.3051899341
84
  }
85
  },
86
  "decode": {
87
  "memory": {
88
  "unit": "MB",
89
+ "max_ram": 954.04032,
90
  "max_global_vram": null,
91
  "max_process_vram": null,
92
  "max_reserved": null,
 
94
  },
95
  "latency": {
96
  "unit": "s",
97
+ "count": 13,
98
+ "total": 0.38152623399997765,
99
+ "mean": 0.029348171846152127,
100
+ "stdev": 0.0007770831253351279,
101
+ "p50": 0.029542226999978993,
102
+ "p90": 0.030372803800003113,
103
+ "p95": 0.030452681000002712,
104
+ "p99": 0.03049001299999986,
105
  "values": [
106
+ 0.029589063999992504,
107
+ 0.029542226999978993,
108
+ 0.030421571000005088,
109
+ 0.03049934599999915,
110
+ 0.028705965000000333,
111
+ 0.02836299099999451,
112
+ 0.02960677699999792,
113
+ 0.03017773499999521,
114
+ 0.030132738000020254,
115
+ 0.028703266999997368,
116
+ 0.029001905000001216,
117
+ 0.028671106999979656,
118
+ 0.02811154100001545
 
 
119
  ]
120
  },
121
  "throughput": {
122
  "unit": "tokens/s",
123
+ "value": 34.073672637674406
124
  },
125
  "energy": {
126
  "unit": "kWh",
127
+ "cpu": 1.0888817133726902e-06,
128
+ "ram": 4.550879494485929e-08,
129
  "gpu": 0.0,
130
+ "total": 1.1343905083175484e-06
131
  },
132
  "efficiency": {
133
  "unit": "tokens/kWh",
134
+ "value": 881530.6481038287
135
  }
136
  },
137
  "per_token": {
138
  "memory": null,
139
  "latency": {
140
  "unit": "s",
141
+ "count": 13,
142
+ "total": 0.3763213649999386,
143
+ "mean": 0.028947797307687588,
144
+ "stdev": 0.000764976234325758,
145
+ "p50": 0.029162888000001885,
146
+ "p90": 0.029980061399982105,
147
+ "p95": 0.030060573799994474,
148
+ "p99": 0.030100954760011973,
149
  "values": [
150
+ 0.029190839999984064,
151
+ 0.029162888000001885,
152
+ 0.030026922999979888,
153
+ 0.03011105000001635,
154
+ 0.02831931199997939,
155
+ 0.027940841999992472,
156
+ 0.02922570399999813,
157
+ 0.029792614999990974,
158
+ 0.029571789000016224,
159
+ 0.028314179999995304,
160
+ 0.02863348699997914,
161
+ 0.02829687799999192,
162
+ 0.02773485700001288
 
 
163
  ]
164
  },
165
  "throughput": {
166
  "unit": "tokens/s",
167
+ "value": 34.54494272468989
168
  },
169
  "energy": null,
170
  "efficiency": null