Spaces:
Running
Running
Julien Simon
commited on
Commit
•
f011c00
1
Parent(s):
3100549
Add g6e.48xlarge
Browse files- results.py +7 -0
- results_virtuoso_large.py +10 -1
- results_virtuoso_medium.py +8 -0
- results_virtuoso_small.py +10 -1
results.py
CHANGED
@@ -85,6 +85,13 @@ instance_type_mappings = {
|
|
85 |
"url": "https://instances.vantage.sh/aws/ec2/g6e.12xlarge",
|
86 |
"price": 5.2465,
|
87 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
88 |
"g4dn.12xlarge": {
|
89 |
"cloud": "AWS",
|
90 |
"gpu": "4xNVIDIA T4",
|
|
|
85 |
"url": "https://instances.vantage.sh/aws/ec2/g6e.12xlarge",
|
86 |
"price": 5.2465,
|
87 |
},
|
88 |
+
"g6e.48xlarge": {
|
89 |
+
"cloud": "AWS",
|
90 |
+
"gpu": "8xNVIDIA L40S",
|
91 |
+
"gpuRAM": "192 GB",
|
92 |
+
"url": "https://instances.vantage.sh/aws/ec2/g6e.48xlarge",
|
93 |
+
"price": 30.131,
|
94 |
+
},
|
95 |
"g4dn.12xlarge": {
|
96 |
"cloud": "AWS",
|
97 |
"gpu": "4xNVIDIA T4",
|
results_virtuoso_large.py
CHANGED
@@ -3,5 +3,14 @@
|
|
3 |
results_virtuoso_large = {
|
4 |
"name": "Virtuoso-Large",
|
5 |
"modelType": "Qwen2 72B", # 2.5
|
6 |
-
"configurations": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
}
|
|
|
3 |
results_virtuoso_large = {
|
4 |
"name": "Virtuoso-Large",
|
5 |
"modelType": "Qwen2 72B", # 2.5
|
6 |
+
"configurations": [
|
7 |
+
{
|
8 |
+
"instanceType": "g6e.48xlarge",
|
9 |
+
"quantization": "none",
|
10 |
+
"container": "LMI 0.30+vLLM 0.6.2",
|
11 |
+
"status": "OK",
|
12 |
+
"tokensPerSecond": "28",
|
13 |
+
"notes": "",
|
14 |
+
},
|
15 |
+
],
|
16 |
}
|
results_virtuoso_medium.py
CHANGED
@@ -44,6 +44,14 @@ results_virtuoso_medium = {
|
|
44 |
"tokensPerSecond": "21",
|
45 |
"notes": "--tensor-parallel-size 2 --max-model-len 16384",
|
46 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
47 |
{
|
48 |
"instanceType": "p4d.24xlarge",
|
49 |
"quantization": "none",
|
|
|
44 |
"tokensPerSecond": "21",
|
45 |
"notes": "--tensor-parallel-size 2 --max-model-len 16384",
|
46 |
},
|
47 |
+
{
|
48 |
+
"instanceType": "g6e.48xlarge",
|
49 |
+
"quantization": "none",
|
50 |
+
"container": "LMI 0.30+vLLM 0.6.2",
|
51 |
+
"status": "OK",
|
52 |
+
"tokensPerSecond": "48",
|
53 |
+
"notes": "",
|
54 |
+
},
|
55 |
{
|
56 |
"instanceType": "p4d.24xlarge",
|
57 |
"quantization": "none",
|
results_virtuoso_small.py
CHANGED
@@ -3,5 +3,14 @@
|
|
3 |
results_virtuoso_small = {
|
4 |
"name": "Virtuoso-Small",
|
5 |
"modelType": "Qwen2.5 14B",
|
6 |
-
"configurations": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
}
|
|
|
3 |
results_virtuoso_small = {
|
4 |
"name": "Virtuoso-Small",
|
5 |
"modelType": "Qwen2.5 14B",
|
6 |
+
"configurations": [
|
7 |
+
{
|
8 |
+
"instanceType": "g6e.48xlarge",
|
9 |
+
"quantization": "none",
|
10 |
+
"container": "LMI 0.30+vLLM 0.6.2",
|
11 |
+
"status": "OK",
|
12 |
+
"tokensPerSecond": "78",
|
13 |
+
"notes": "",
|
14 |
+
},
|
15 |
+
],
|
16 |
}
|