Upload Date | October 24 2024 03:26 PM |
Views | 7 |
AI Information | |
---|---|
Framework | ONNX |
Backend | DirectML |
Device | NVIDIA GeForce RTX 4090 |
System Information | |
---|---|
Operating System | Microsoft Windows 10 Pro (64-bit) |
Model | WRX80D8-2T |
Motherboard | ASRockRack WRX80D8-2T |
Power Plan | Rovnovha |
CPU Information | |
---|---|
Name | AMD Ryzen Threadripper PRO 3975WX |
Topology | 1 Processor, 32 Cores, 64 Threads |
Identifier | AuthenticAMD Family 23 Model 49 Stepping 0 |
Base Frequency | 3.50 GHz |
Cluster 1 | 32 Cores |
Memory Information | |
---|---|
Size | 256.00 GB |
Workload | Accuracy | Score | |
---|---|---|---|
Image Classification (SP)
|
100% |
11199
2.08 KIPS |
|
Image Classification (HP)
|
99% |
14103
2.63 KIPS |
|
Image Classification (Q)
|
100% |
9585
1.78 KIPS |
|
Image Segmentation (SP)
|
100% |
12521
203.0 IPS |
|
Image Segmentation (HP)
|
100% |
13317
215.9 IPS |
|
Image Segmentation (Q)
|
98% |
11674
189.8 IPS |
|
Pose Estimation (SP)
|
100% |
281853
328.9 IPS |
|
Pose Estimation (HP)
|
100% |
437037
509.9 IPS |
|
Pose Estimation (Q)
|
97% |
248109
290.5 IPS |
|
Object Detection (SP)
|
100% |
14700
1.17 KIPS |
|
Object Detection (HP)
|
100% |
17832
1.41 KIPS |
|
Object Detection (Q)
|
89% |
12198
976.4 IPS |
|
Face Detection (SP)
|
100% |
30749
365.4 IPS |
|
Face Detection (HP)
|
100% |
35257
418.9 IPS |
|
Face Detection (Q)
|
97% |
27836
331.9 IPS |
|
Depth Estimation (SP)
|
100% |
84768
653.1 IPS |
|
Depth Estimation (HP)
|
99% |
136569
1.05 KIPS |
|
Depth Estimation (Q)
|
75% |
62445
503.7 IPS |
|
Style Transfer (SP)
|
100% |
498768
641.2 IPS |
|
Style Transfer (HP)
|
100% |
873053
1.12 KIPS |
|
Style Transfer (Q)
|
98% |
434326
560.1 IPS |
|
Image Super-Resolution (SP)
|
100% |
51258
1.89 KIPS |
|
Image Super-Resolution (HP)
|
100% |
65944
2.44 KIPS |
|
Image Super-Resolution (Q)
|
99% |
38962
1.44 KIPS |
|
Text Classification (SP)
|
100% |
3013
4.02 KIPS |
|
Text Classification (HP)
|
99% |
3576
4.77 KIPS |
|
Text Classification (Q)
|
97% |
1931
2.59 KIPS |
|
Machine Translation (SP)
|
100% |
3156
54.4 IPS |
|
Machine Translation (HP)
|
100% |
3163
54.5 IPS |
|
Machine Translation (Q)
|
60% |
1606
35.8 IPS |