Hugging Face's logo

Datasets:
boavizta
/
open_data_boamps

License:
Dataset card Files
xet
 Community
18

Upload 10 files

#30
by Nivose - opened
base: refs/heads/main
←
from: refs/pr/30
Discussion Files changed
+900
-0
Files changed (10) hide show
  1. data/report_SSG_inference_question-answering_privateCloud_35348ab0-1abf-4839-9f2c-33c5e9b10fe4.json +90 -0
  2. data/report_SSG_inference_question-answering_privateCloud_372309f6-196b-4286-98a2-67ddb959ebe5.json +90 -0
  3. data/report_SSG_inference_question-answering_privateCloud_3adf12f2-31a3-454a-a9d7-6e1710a412f2.json +90 -0
  4. data/report_SSG_inference_question-answering_privateCloud_3d11ef38-e5f9-43f0-ac25-fcdd61115133.json +90 -0
  5. data/report_SSG_inference_question-answering_privateCloud_66b9b016-e026-4d02-ae96-b83e978897c7.json +90 -0
  6. data/report_SSG_inference_question-answering_privateCloud_70d6ebc5-7244-49c4-9603-6522d78db536.json +90 -0
  7. data/report_SSG_inference_question-answering_privateCloud_9237143f-fd40-45e6-bf72-51d45f5401fa.json +90 -0
  8. data/report_SSG_inference_question-answering_privateCloud_a3342146-0d82-4501-acf4-47cb7d9fda17.json +90 -0
  9. data/report_SSG_inference_question-answering_privateCloud_b502f765-2c20-4f4a-9265-5126c5b4eba0.json +90 -0
  10. data/report_SSG_inference_question-answering_privateCloud_bd01b8bb-5c3f-44ef-b1f8-81e7bb3305ed.json +90 -0
data/report_SSG_inference_question-answering_privateCloud_35348ab0-1abf-4839-9f2c-33c5e9b10fe4.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "header": {
3
+ "reportId":"35348ab0-1abf-4839-9f2c-33c5e9b10fe4",
4
+ "reportDatetime":"2024-09-30T08:28:17",
5
+ "reportStatus":"final",
6
+ "publisher": {
7
+ "name":"SopraSteriaGroup",
8
+ "division":"sustAIn",
9
+ "projectName":"GA2I",
10
+ "confidentialityLevel":"public"
11
+ }
12
+ },
13
+ "task": {
14
+ "taskFamily":"question-answering",
15
+ "taskStage":"inference",
16
+ "algorithms": [
17
+ {
18
+ "algorithmType":"llm",
19
+ "foundationModelName":"llama2-13b",
20
+ "parametersNumber":13.0,
21
+ "framework":"vllm",
22
+ "quantization":"q16"
23
+ }
24
+ ],
25
+ "dataset": [
26
+ {
27
+ "dataUsage":"input",
28
+ "dataType":"token",
29
+ "dataQuantity":3225,
30
+ "source":"private",
31
+ "owner":"SSG"
32
+ },
33
+ {
34
+ "dataUsage":"output",
35
+ "dataType":"token",
36
+ "dataQuantity":39539,
37
+ "source":"private",
38
+ "owner":"SSG"
39
+ }
40
+ ],
41
+ "estimatedAccuracy":"good",
42
+ "taskDescription":"Simple question answer with 103 questions from simple ones to contextualised one. Dataset : 91 simple questions (Average: 11 Tokens) and 12 contextualized questions (Question: average 15 Tokens - Context: average 160 Tokens) All tests have been made on a single GPU Even if the node have 2 available GPU."
43
+ },
44
+ "measures": [
45
+ {
46
+ "measurementMethod":"codecarbon",
47
+ "version":"2.5.0",
48
+ "cpuTrackingMode":"machine",
49
+ "gpuTrackingMode":"machine",
50
+ "powerConsumption":0.0016147225036887,
51
+ "measurementDuration":11.497305870056152,
52
+ "measurementDateTime":"2024-09-30T08:28:17"
53
+ }
54
+ ],
55
+ "system": {
56
+ "os":"Linux-5.15.0-116-generic-x86_64-with-glibc2.35"
57
+ },
58
+ "software": {
59
+ "language":"python",
60
+ "version":"3.10.12"
61
+ },
62
+ "infrastructure": {
63
+ "infraType":"privateCloud",
64
+ "cloudProvider":"OVH-v100s",
65
+ "components": [
66
+ {
67
+ "componentName":"Intel(R) Xeon(R) Gold 6226R CPU @ 2.90GHz",
68
+ "componentType":"cpu",
69
+ "nbComponent":30
70
+ },
71
+ {
72
+ "componentName":"2 x Tesla V100S-PCIE-32GB",
73
+ "componentType":"gpu",
74
+ "nbComponent":2
75
+ },
76
+ {
77
+ "componentName":"RAM",
78
+ "componentType":"ram",
79
+ "nbComponent":1,
80
+ "memorySize":86.30359268188477
81
+ }
82
+ ]
83
+ },
84
+ "environment": {
85
+ "country":"France",
86
+ "latitude":48.8582,
87
+ "longitude":2.3387
88
+ },
89
+ "quality":"high"
90
+ }
data/report_SSG_inference_question-answering_privateCloud_372309f6-196b-4286-98a2-67ddb959ebe5.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "header": {
3
+ "reportId":"372309f6-196b-4286-98a2-67ddb959ebe5",
4
+ "reportDatetime":"2024-10-16T07:26:12",
5
+ "reportStatus":"final",
6
+ "publisher": {
7
+ "name":"SopraSteriaGroup",
8
+ "division":"sustAIn",
9
+ "projectName":"GA2I",
10
+ "confidentialityLevel":"public"
11
+ }
12
+ },
13
+ "task": {
14
+ "taskFamily":"question-answering",
15
+ "taskStage":"inference",
16
+ "algorithms": [
17
+ {
18
+ "algorithmType":"llm",
19
+ "foundationModelName":"llama3-8b",
20
+ "parametersNumber":8.0,
21
+ "framework":"vllm",
22
+ "quantization":"q16"
23
+ }
24
+ ],
25
+ "dataset": [
26
+ {
27
+ "dataUsage":"input",
28
+ "dataType":"token",
29
+ "dataQuantity":2987,
30
+ "source":"private",
31
+ "owner":"SSG"
32
+ },
33
+ {
34
+ "dataUsage":"output",
35
+ "dataType":"token",
36
+ "dataQuantity":63066,
37
+ "source":"private",
38
+ "owner":"SSG"
39
+ }
40
+ ],
41
+ "estimatedAccuracy":"good",
42
+ "taskDescription":"Simple question answer with 103 questions from simple ones to contextualised one. Dataset : 91 simple questions (Average: 11 Tokens) and 12 contextualized questions (Question: average 15 Tokens - Context: average 160 Tokens) All tests have been made on a single GPU Even if the node have 2 available GPU."
43
+ },
44
+ "measures": [
45
+ {
46
+ "measurementMethod":"codecarbon",
47
+ "version":"2.5.0",
48
+ "cpuTrackingMode":"machine",
49
+ "gpuTrackingMode":"machine",
50
+ "powerConsumption":0.0006877453643107,
51
+ "measurementDuration":5.442811965942383,
52
+ "measurementDateTime":"2024-10-16T07:26:12"
53
+ }
54
+ ],
55
+ "system": {
56
+ "os":"Linux-5.15.0-116-generic-x86_64-with-glibc2.35"
57
+ },
58
+ "software": {
59
+ "language":"python",
60
+ "version":"3.10.12"
61
+ },
62
+ "infrastructure": {
63
+ "infraType":"privateCloud",
64
+ "cloudProvider":"OVH-v100s",
65
+ "components": [
66
+ {
67
+ "componentName":"Intel(R) Xeon(R) Gold 6226R CPU @ 2.90GHz",
68
+ "componentType":"cpu",
69
+ "nbComponent":30
70
+ },
71
+ {
72
+ "componentName":"2 x Tesla V100S-PCIE-32GB",
73
+ "componentType":"gpu",
74
+ "nbComponent":2
75
+ },
76
+ {
77
+ "componentName":"RAM",
78
+ "componentType":"ram",
79
+ "nbComponent":1,
80
+ "memorySize":86.3035888671875
81
+ }
82
+ ]
83
+ },
84
+ "environment": {
85
+ "country":"France",
86
+ "latitude":48.8582,
87
+ "longitude":2.3387
88
+ },
89
+ "quality":"high"
90
+ }
data/report_SSG_inference_question-answering_privateCloud_3adf12f2-31a3-454a-a9d7-6e1710a412f2.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "header": {
3
+ "reportId":"3adf12f2-31a3-454a-a9d7-6e1710a412f2",
4
+ "reportDatetime":"2024-09-26T10:47:04",
5
+ "reportStatus":"final",
6
+ "publisher": {
7
+ "name":"SopraSteriaGroup",
8
+ "division":"sustAIn",
9
+ "projectName":"GA2I",
10
+ "confidentialityLevel":"public"
11
+ }
12
+ },
13
+ "task": {
14
+ "taskFamily":"question-answering",
15
+ "taskStage":"inference",
16
+ "algorithms": [
17
+ {
18
+ "algorithmType":"llm",
19
+ "foundationModelName":"llama2-13b",
20
+ "parametersNumber":13.0,
21
+ "framework":"vllm",
22
+ "quantization":"q16"
23
+ }
24
+ ],
25
+ "dataset": [
26
+ {
27
+ "dataUsage":"input",
28
+ "dataType":"token",
29
+ "dataQuantity":3225,
30
+ "source":"private",
31
+ "owner":"SSG"
32
+ },
33
+ {
34
+ "dataUsage":"output",
35
+ "dataType":"token",
36
+ "dataQuantity":12170,
37
+ "source":"private",
38
+ "owner":"SSG"
39
+ }
40
+ ],
41
+ "estimatedAccuracy":"good",
42
+ "taskDescription":"Simple question answer with 103 questions from simple ones to contextualised one. Dataset : 91 simple questions (Average: 11 Tokens) and 12 contextualized questions (Question: average 15 Tokens - Context: average 160 Tokens) All tests have been made on a single GPU Even if the node have 2 available GPU."
43
+ },
44
+ "measures": [
45
+ {
46
+ "measurementMethod":"codecarbon",
47
+ "version":"2.5.0",
48
+ "cpuTrackingMode":"machine",
49
+ "gpuTrackingMode":"machine",
50
+ "powerConsumption":0.0004894143530043,
51
+ "measurementDuration":3.4514710903167725,
52
+ "measurementDateTime":"2024-09-26T10:47:04"
53
+ }
54
+ ],
55
+ "system": {
56
+ "os":"Linux-5.15.0-116-generic-x86_64-with-glibc2.35"
57
+ },
58
+ "software": {
59
+ "language":"python",
60
+ "version":"3.10.12"
61
+ },
62
+ "infrastructure": {
63
+ "infraType":"privateCloud",
64
+ "cloudProvider":"OVH-v100s",
65
+ "components": [
66
+ {
67
+ "componentName":"Intel(R) Xeon(R) Gold 6226R CPU @ 2.90GHz",
68
+ "componentType":"cpu",
69
+ "nbComponent":30
70
+ },
71
+ {
72
+ "componentName":"2 x Tesla V100S-PCIE-32GB",
73
+ "componentType":"gpu",
74
+ "nbComponent":2
75
+ },
76
+ {
77
+ "componentName":"RAM",
78
+ "componentType":"ram",
79
+ "nbComponent":1,
80
+ "memorySize":86.3035888671875
81
+ }
82
+ ]
83
+ },
84
+ "environment": {
85
+ "country":"France",
86
+ "latitude":48.8582,
87
+ "longitude":2.3387
88
+ },
89
+ "quality":"high"
90
+ }
data/report_SSG_inference_question-answering_privateCloud_3d11ef38-e5f9-43f0-ac25-fcdd61115133.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "header": {
3
+ "reportId":"3d11ef38-e5f9-43f0-ac25-fcdd61115133",
4
+ "reportDatetime":"2024-11-06T10:15:23",
5
+ "reportStatus":"final",
6
+ "publisher": {
7
+ "name":"SopraSteriaGroup",
8
+ "division":"sustAIn",
9
+ "projectName":"GA2I",
10
+ "confidentialityLevel":"public"
11
+ }
12
+ },
13
+ "task": {
14
+ "taskFamily":"question-answering",
15
+ "taskStage":"inference",
16
+ "algorithms": [
17
+ {
18
+ "algorithmType":"llm",
19
+ "foundationModelName":"meta-llama-3-13b-instruct-gguf",
20
+ "parametersNumber":13.0,
21
+ "framework":"llamacpp",
22
+ "quantization":"8bit0"
23
+ }
24
+ ],
25
+ "dataset": [
26
+ {
27
+ "dataUsage":"input",
28
+ "dataType":"token",
29
+ "dataQuantity":5930,
30
+ "source":"private",
31
+ "owner":"SSG"
32
+ },
33
+ {
34
+ "dataUsage":"output",
35
+ "dataType":"token",
36
+ "dataQuantity":28559,
37
+ "source":"private",
38
+ "owner":"SSG"
39
+ }
40
+ ],
41
+ "estimatedAccuracy":"good",
42
+ "taskDescription":"Simple question answer with 103 questions from simple ones to contextualised one. Dataset : 91 simple questions (Average: 11 Tokens) and 12 contextualized questions (Question: average 15 Tokens - Context: average 160 Tokens) All tests have been made on a single GPU Even if the node have 2 available GPU."
43
+ },
44
+ "measures": [
45
+ {
46
+ "measurementMethod":"codecarbon",
47
+ "version":"2.5.0",
48
+ "cpuTrackingMode":"machine",
49
+ "gpuTrackingMode":"machine",
50
+ "powerConsumption":0.0007014954565571,
51
+ "measurementDuration":10.449934482574465,
52
+ "measurementDateTime":"2024-11-06T10:15:23"
53
+ }
54
+ ],
55
+ "system": {
56
+ "os":"Linux-5.15.0-124-generic-x86_64-with-glibc2.35"
57
+ },
58
+ "software": {
59
+ "language":"python",
60
+ "version":"3.10.12"
61
+ },
62
+ "infrastructure": {
63
+ "infraType":"privateCloud",
64
+ "cloudProvider":"OVH-v100s",
65
+ "components": [
66
+ {
67
+ "componentName":"Intel(R) Xeon(R) Gold 6226R CPU @ 2.90GHz",
68
+ "componentType":"cpu",
69
+ "nbComponent":30
70
+ },
71
+ {
72
+ "componentName":"2 x Tesla V100S-PCIE-32GB",
73
+ "componentType":"gpu",
74
+ "nbComponent":2
75
+ },
76
+ {
77
+ "componentName":"RAM",
78
+ "componentType":"ram",
79
+ "nbComponent":1,
80
+ "memorySize":86.30358123779297
81
+ }
82
+ ]
83
+ },
84
+ "environment": {
85
+ "country":"France",
86
+ "latitude":48.8582,
87
+ "longitude":2.3387
88
+ },
89
+ "quality":"high"
90
+ }
data/report_SSG_inference_question-answering_privateCloud_66b9b016-e026-4d02-ae96-b83e978897c7.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "header": {
3
+ "reportId":"66b9b016-e026-4d02-ae96-b83e978897c7",
4
+ "reportDatetime":"2024-10-30T10:12:49",
5
+ "reportStatus":"final",
6
+ "publisher": {
7
+ "name":"SopraSteriaGroup",
8
+ "division":"sustAIn",
9
+ "projectName":"GA2I",
10
+ "confidentialityLevel":"public"
11
+ }
12
+ },
13
+ "task": {
14
+ "taskFamily":"question-answering",
15
+ "taskStage":"inference",
16
+ "algorithms": [
17
+ {
18
+ "algorithmType":"llm",
19
+ "foundationModelName":"mistral-7b-instruct",
20
+ "parametersNumber":7.0,
21
+ "framework":"vllm",
22
+ "quantization":"q16"
23
+ }
24
+ ],
25
+ "dataset": [
26
+ {
27
+ "dataUsage":"input",
28
+ "dataType":"token",
29
+ "dataQuantity":4101,
30
+ "source":"private",
31
+ "owner":"SSG"
32
+ },
33
+ {
34
+ "dataUsage":"output",
35
+ "dataType":"token",
36
+ "dataQuantity":39458,
37
+ "source":"private",
38
+ "owner":"SSG"
39
+ }
40
+ ],
41
+ "estimatedAccuracy":"good",
42
+ "taskDescription":"Simple question answer with 103 questions from simple ones to contextualised one. Dataset : 91 simple questions (Average: 11 Tokens) and 12 contextualized questions (Question: average 15 Tokens - Context: average 160 Tokens) All tests have been made on a single GPU Even if the node have 2 available GPU."
43
+ },
44
+ "measures": [
45
+ {
46
+ "measurementMethod":"codecarbon",
47
+ "version":"2.5.0",
48
+ "cpuTrackingMode":"machine",
49
+ "gpuTrackingMode":"machine",
50
+ "powerConsumption":0.0008431335305842,
51
+ "measurementDuration":6.305508852005005,
52
+ "measurementDateTime":"2024-10-30T10:12:49"
53
+ }
54
+ ],
55
+ "system": {
56
+ "os":"Linux-5.15.0-124-generic-x86_64-with-glibc2.35"
57
+ },
58
+ "software": {
59
+ "language":"python",
60
+ "version":"3.10.12"
61
+ },
62
+ "infrastructure": {
63
+ "infraType":"privateCloud",
64
+ "cloudProvider":"OVH-v100s",
65
+ "components": [
66
+ {
67
+ "componentName":"Intel(R) Xeon(R) Gold 6226R CPU @ 2.90GHz",
68
+ "componentType":"cpu",
69
+ "nbComponent":30
70
+ },
71
+ {
72
+ "componentName":"2 x Tesla V100S-PCIE-32GB",
73
+ "componentType":"gpu",
74
+ "nbComponent":2
75
+ },
76
+ {
77
+ "componentName":"RAM",
78
+ "componentType":"ram",
79
+ "nbComponent":1,
80
+ "memorySize":86.3035774230957
81
+ }
82
+ ]
83
+ },
84
+ "environment": {
85
+ "country":"France",
86
+ "latitude":48.7898,
87
+ "longitude":2.5447
88
+ },
89
+ "quality":"high"
90
+ }
data/report_SSG_inference_question-answering_privateCloud_70d6ebc5-7244-49c4-9603-6522d78db536.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "header": {
3
+ "reportId":"70d6ebc5-7244-49c4-9603-6522d78db536",
4
+ "reportDatetime":"2024-10-10T09:20:16",
5
+ "reportStatus":"final",
6
+ "publisher": {
7
+ "name":"SopraSteriaGroup",
8
+ "division":"sustAIn",
9
+ "projectName":"GA2I",
10
+ "confidentialityLevel":"public"
11
+ }
12
+ },
13
+ "task": {
14
+ "taskFamily":"question-answering",
15
+ "taskStage":"inference",
16
+ "algorithms": [
17
+ {
18
+ "algorithmType":"llm",
19
+ "foundationModelName":"mistral-7b",
20
+ "parametersNumber":7.0,
21
+ "framework":"vllm",
22
+ "quantization":"q16"
23
+ }
24
+ ],
25
+ "dataset": [
26
+ {
27
+ "dataUsage":"input",
28
+ "dataType":"token",
29
+ "dataQuantity":3159,
30
+ "source":"private",
31
+ "owner":"SSG"
32
+ },
33
+ {
34
+ "dataUsage":"output",
35
+ "dataType":"token",
36
+ "dataQuantity":78793,
37
+ "source":"private",
38
+ "owner":"SSG"
39
+ }
40
+ ],
41
+ "estimatedAccuracy":"good",
42
+ "taskDescription":"Simple question answer with 103 questions from simple ones to contextualised one. Dataset : 91 simple questions (Average: 11 Tokens) and 12 contextualized questions (Question: average 15 Tokens - Context: average 160 Tokens) All tests have been made on a single GPU Even if the node have 2 available GPU."
43
+ },
44
+ "measures": [
45
+ {
46
+ "measurementMethod":"codecarbon",
47
+ "version":"2.5.0",
48
+ "cpuTrackingMode":"machine",
49
+ "gpuTrackingMode":"machine",
50
+ "powerConsumption":0.0005677864944745,
51
+ "measurementDuration":4.518536329269409,
52
+ "measurementDateTime":"2024-10-10T09:20:16"
53
+ }
54
+ ],
55
+ "system": {
56
+ "os":"Linux-5.15.0-116-generic-x86_64-with-glibc2.35"
57
+ },
58
+ "software": {
59
+ "language":"python",
60
+ "version":"3.10.12"
61
+ },
62
+ "infrastructure": {
63
+ "infraType":"privateCloud",
64
+ "cloudProvider":"OVH-v100s",
65
+ "components": [
66
+ {
67
+ "componentName":"Intel(R) Xeon(R) Gold 6226R CPU @ 2.90GHz",
68
+ "componentType":"cpu",
69
+ "nbComponent":30
70
+ },
71
+ {
72
+ "componentName":"2 x Tesla V100S-PCIE-32GB",
73
+ "componentType":"gpu",
74
+ "nbComponent":2
75
+ },
76
+ {
77
+ "componentName":"RAM",
78
+ "componentType":"ram",
79
+ "nbComponent":1,
80
+ "memorySize":86.3035888671875
81
+ }
82
+ ]
83
+ },
84
+ "environment": {
85
+ "country":"France",
86
+ "latitude":48.8582,
87
+ "longitude":2.3387
88
+ },
89
+ "quality":"high"
90
+ }
data/report_SSG_inference_question-answering_privateCloud_9237143f-fd40-45e6-bf72-51d45f5401fa.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "header": {
3
+ "reportId":"9237143f-fd40-45e6-bf72-51d45f5401fa",
4
+ "reportDatetime":"2024-10-10T06:55:22",
5
+ "reportStatus":"final",
6
+ "publisher": {
7
+ "name":"SopraSteriaGroup",
8
+ "division":"sustAIn",
9
+ "projectName":"GA2I",
10
+ "confidentialityLevel":"public"
11
+ }
12
+ },
13
+ "task": {
14
+ "taskFamily":"question-answering",
15
+ "taskStage":"inference",
16
+ "algorithms": [
17
+ {
18
+ "algorithmType":"llm",
19
+ "foundationModelName":"llama2-13b",
20
+ "parametersNumber":13.0,
21
+ "framework":"vllm",
22
+ "quantization":"q16"
23
+ }
24
+ ],
25
+ "dataset": [
26
+ {
27
+ "dataUsage":"input",
28
+ "dataType":"token",
29
+ "dataQuantity":1817,
30
+ "source":"private",
31
+ "owner":"SSG"
32
+ },
33
+ {
34
+ "dataUsage":"output",
35
+ "dataType":"token",
36
+ "dataQuantity":1008,
37
+ "source":"private",
38
+ "owner":"SSG"
39
+ }
40
+ ],
41
+ "estimatedAccuracy":"good",
42
+ "taskDescription":"Simple question answer with 103 questions from simple ones to contextualised one. Dataset : 91 simple questions (Average: 11 Tokens) and 12 contextualized questions (Question: average 15 Tokens - Context: average 160 Tokens) All tests have been made on a single GPU Even if the node have 2 available GPU."
43
+ },
44
+ "measures": [
45
+ {
46
+ "measurementMethod":"codecarbon",
47
+ "version":"2.5.0",
48
+ "cpuTrackingMode":"machine",
49
+ "gpuTrackingMode":"machine",
50
+ "powerConsumption":0.0006318521200859,
51
+ "measurementDuration":4.509745121002197,
52
+ "measurementDateTime":"2024-10-10T06:55:22"
53
+ }
54
+ ],
55
+ "system": {
56
+ "os":"Linux-5.15.0-116-generic-x86_64-with-glibc2.35"
57
+ },
58
+ "software": {
59
+ "language":"python",
60
+ "version":"3.10.12"
61
+ },
62
+ "infrastructure": {
63
+ "infraType":"privateCloud",
64
+ "cloudProvider":"OVH-v100s",
65
+ "components": [
66
+ {
67
+ "componentName":"Intel(R) Xeon(R) Gold 6226R CPU @ 2.90GHz",
68
+ "componentType":"cpu",
69
+ "nbComponent":30
70
+ },
71
+ {
72
+ "componentName":"2 x Tesla V100S-PCIE-32GB",
73
+ "componentType":"gpu",
74
+ "nbComponent":2
75
+ },
76
+ {
77
+ "componentName":"RAM",
78
+ "componentType":"ram",
79
+ "nbComponent":1,
80
+ "memorySize":86.3035888671875
81
+ }
82
+ ]
83
+ },
84
+ "environment": {
85
+ "country":"France",
86
+ "latitude":48.8582,
87
+ "longitude":2.3387
88
+ },
89
+ "quality":"high"
90
+ }
data/report_SSG_inference_question-answering_privateCloud_a3342146-0d82-4501-acf4-47cb7d9fda17.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "header": {
3
+ "reportId":"a3342146-0d82-4501-acf4-47cb7d9fda17",
4
+ "reportDatetime":"2024-10-22T05:44:53",
5
+ "reportStatus":"final",
6
+ "publisher": {
7
+ "name":"SopraSteriaGroup",
8
+ "division":"sustAIn",
9
+ "projectName":"GA2I",
10
+ "confidentialityLevel":"public"
11
+ }
12
+ },
13
+ "task": {
14
+ "taskFamily":"question-answering",
15
+ "taskStage":"inference",
16
+ "algorithms": [
17
+ {
18
+ "algorithmType":"llm",
19
+ "foundationModelName":"phi3med-14b",
20
+ "parametersNumber":14.0,
21
+ "framework":"vllm",
22
+ "quantization":"q16"
23
+ }
24
+ ],
25
+ "dataset": [
26
+ {
27
+ "dataUsage":"input",
28
+ "dataType":"token",
29
+ "dataQuantity":3533,
30
+ "source":"private",
31
+ "owner":"SSG"
32
+ },
33
+ {
34
+ "dataUsage":"output",
35
+ "dataType":"token",
36
+ "dataQuantity":202614,
37
+ "source":"private",
38
+ "owner":"SSG"
39
+ }
40
+ ],
41
+ "estimatedAccuracy":"good",
42
+ "taskDescription":"Simple question answer with 103 questions from simple ones to contextualised one. Dataset : 91 simple questions (Average: 11 Tokens) and 12 contextualized questions (Question: average 15 Tokens - Context: average 160 Tokens) All tests have been made on a single GPU Even if the node have 2 available GPU."
43
+ },
44
+ "measures": [
45
+ {
46
+ "measurementMethod":"codecarbon",
47
+ "version":"2.5.0",
48
+ "cpuTrackingMode":"machine",
49
+ "gpuTrackingMode":"machine",
50
+ "powerConsumption":0.0111332190261356,
51
+ "measurementDuration":84.0584135055542,
52
+ "measurementDateTime":"2024-10-22T05:44:53"
53
+ }
54
+ ],
55
+ "system": {
56
+ "os":"Linux-5.15.0-116-generic-x86_64-with-glibc2.35"
57
+ },
58
+ "software": {
59
+ "language":"python",
60
+ "version":"3.10.12"
61
+ },
62
+ "infrastructure": {
63
+ "infraType":"privateCloud",
64
+ "cloudProvider":"OVH-v100s",
65
+ "components": [
66
+ {
67
+ "componentName":"Intel(R) Xeon(R) Gold 6226R CPU @ 2.90GHz",
68
+ "componentType":"cpu",
69
+ "nbComponent":30
70
+ },
71
+ {
72
+ "componentName":"2 x Tesla V100S-PCIE-32GB",
73
+ "componentType":"gpu",
74
+ "nbComponent":2
75
+ },
76
+ {
77
+ "componentName":"RAM",
78
+ "componentType":"ram",
79
+ "nbComponent":1,
80
+ "memorySize":86.30360412597656
81
+ }
82
+ ]
83
+ },
84
+ "environment": {
85
+ "country":"France",
86
+ "latitude":48.8582,
87
+ "longitude":2.3387
88
+ },
89
+ "quality":"high"
90
+ }
data/report_SSG_inference_question-answering_privateCloud_b502f765-2c20-4f4a-9265-5126c5b4eba0.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "header": {
3
+ "reportId":"b502f765-2c20-4f4a-9265-5126c5b4eba0",
4
+ "reportDatetime":"2024-09-30T09:10:41",
5
+ "reportStatus":"final",
6
+ "publisher": {
7
+ "name":"SopraSteriaGroup",
8
+ "division":"sustAIn",
9
+ "projectName":"GA2I",
10
+ "confidentialityLevel":"public"
11
+ }
12
+ },
13
+ "task": {
14
+ "taskFamily":"question-answering",
15
+ "taskStage":"inference",
16
+ "algorithms": [
17
+ {
18
+ "algorithmType":"llm",
19
+ "foundationModelName":"llama2-13b",
20
+ "parametersNumber":13.0,
21
+ "framework":"vllm",
22
+ "quantization":"q16"
23
+ }
24
+ ],
25
+ "dataset": [
26
+ {
27
+ "dataUsage":"input",
28
+ "dataType":"token",
29
+ "dataQuantity":3225,
30
+ "source":"private",
31
+ "owner":"SSG"
32
+ },
33
+ {
34
+ "dataUsage":"output",
35
+ "dataType":"token",
36
+ "dataQuantity":88326,
37
+ "source":"private",
38
+ "owner":"SSG"
39
+ }
40
+ ],
41
+ "estimatedAccuracy":"good",
42
+ "taskDescription":"Simple question answer with 103 questions from simple ones to contextualised one. Dataset : 91 simple questions (Average: 11 Tokens) and 12 contextualized questions (Question: average 15 Tokens - Context: average 160 Tokens) All tests have been made on a single GPU Even if the node have 2 available GPU."
43
+ },
44
+ "measures": [
45
+ {
46
+ "measurementMethod":"codecarbon",
47
+ "version":"2.5.0",
48
+ "cpuTrackingMode":"machine",
49
+ "gpuTrackingMode":"machine",
50
+ "powerConsumption":0.0048764027775579,
51
+ "measurementDuration":34.46843481063843,
52
+ "measurementDateTime":"2024-09-30T09:10:41"
53
+ }
54
+ ],
55
+ "system": {
56
+ "os":"Linux-5.15.0-116-generic-x86_64-with-glibc2.35"
57
+ },
58
+ "software": {
59
+ "language":"python",
60
+ "version":"3.10.12"
61
+ },
62
+ "infrastructure": {
63
+ "infraType":"privateCloud",
64
+ "cloudProvider":"OVH-v100s",
65
+ "components": [
66
+ {
67
+ "componentName":"Intel(R) Xeon(R) Gold 6226R CPU @ 2.90GHz",
68
+ "componentType":"cpu",
69
+ "nbComponent":30
70
+ },
71
+ {
72
+ "componentName":"2 x Tesla V100S-PCIE-32GB",
73
+ "componentType":"gpu",
74
+ "nbComponent":2
75
+ },
76
+ {
77
+ "componentName":"RAM",
78
+ "componentType":"ram",
79
+ "nbComponent":1,
80
+ "memorySize":86.30359268188477
81
+ }
82
+ ]
83
+ },
84
+ "environment": {
85
+ "country":"France",
86
+ "latitude":48.8582,
87
+ "longitude":2.3387
88
+ },
89
+ "quality":"high"
90
+ }
data/report_SSG_inference_question-answering_privateCloud_bd01b8bb-5c3f-44ef-b1f8-81e7bb3305ed.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "header": {
3
+ "reportId":"bd01b8bb-5c3f-44ef-b1f8-81e7bb3305ed",
4
+ "reportDatetime":"2024-10-23T08:02:59",
5
+ "reportStatus":"final",
6
+ "publisher": {
7
+ "name":"SopraSteriaGroup",
8
+ "division":"sustAIn",
9
+ "projectName":"GA2I",
10
+ "confidentialityLevel":"public"
11
+ }
12
+ },
13
+ "task": {
14
+ "taskFamily":"question-answering",
15
+ "taskStage":"inference",
16
+ "algorithms": [
17
+ {
18
+ "algorithmType":"llm",
19
+ "foundationModelName":"phi3med-14b",
20
+ "parametersNumber":14.0,
21
+ "framework":"vllm",
22
+ "quantization":"q16"
23
+ }
24
+ ],
25
+ "dataset": [
26
+ {
27
+ "dataUsage":"input",
28
+ "dataType":"token",
29
+ "dataQuantity":3533,
30
+ "source":"private",
31
+ "owner":"SSG"
32
+ },
33
+ {
34
+ "dataUsage":"output",
35
+ "dataType":"token",
36
+ "dataQuantity":329401,
37
+ "source":"private",
38
+ "owner":"SSG"
39
+ }
40
+ ],
41
+ "estimatedAccuracy":"good",
42
+ "taskDescription":"Simple question answer with 103 questions from simple ones to contextualised one. Dataset : 91 simple questions (Average: 11 Tokens) and 12 contextualized questions (Question: average 15 Tokens - Context: average 160 Tokens) All tests have been made on a single GPU Even if the node have 2 available GPU."
43
+ },
44
+ "measures": [
45
+ {
46
+ "measurementMethod":"codecarbon",
47
+ "version":"2.5.0",
48
+ "cpuTrackingMode":"machine",
49
+ "gpuTrackingMode":"machine",
50
+ "powerConsumption":0.0157068207062144,
51
+ "measurementDuration":119.17674136161804,
52
+ "measurementDateTime":"2024-10-23T08:02:59"
53
+ }
54
+ ],
55
+ "system": {
56
+ "os":"Linux-5.15.0-116-generic-x86_64-with-glibc2.35"
57
+ },
58
+ "software": {
59
+ "language":"python",
60
+ "version":"3.10.12"
61
+ },
62
+ "infrastructure": {
63
+ "infraType":"privateCloud",
64
+ "cloudProvider":"OVH-v100s",
65
+ "components": [
66
+ {
67
+ "componentName":"Intel(R) Xeon(R) Gold 6226R CPU @ 2.90GHz",
68
+ "componentType":"cpu",
69
+ "nbComponent":30
70
+ },
71
+ {
72
+ "componentName":"2 x Tesla V100S-PCIE-32GB",
73
+ "componentType":"gpu",
74
+ "nbComponent":2
75
+ },
76
+ {
77
+ "componentName":"RAM",
78
+ "componentType":"ram",
79
+ "nbComponent":1,
80
+ "memorySize":86.30359268188477
81
+ }
82
+ ]
83
+ },
84
+ "environment": {
85
+ "country":"France",
86
+ "latitude":48.7898,
87
+ "longitude":2.5447
88
+ },
89
+ "quality":"high"
90
+ }