User | eahyndman3 |
Upload Date | November 02 2024 03:22 PM |
Views | 11 |
AI Information | |
---|---|
Framework | Core ML |
Backend | Neural Engine |
Device | ARM |
System Information | |
---|---|
Operating System | iOS 18.1 |
Model | iPad16,5 |
Model ID | iPad16,5 |
Motherboard | J720AP |
CPU Information | |
---|---|
Name | ARM |
Topology | 1 Processor, 10 Cores |
Identifier | ARM |
Base Frequency | 4.41 GHz |
Cluster 1 | 4 Cores |
Cluster 2 | 6 Cores |
Memory Information | |
---|---|
Size | 15.07 GB |
Workload | Accuracy | Score | |
---|---|---|---|
Image Classification (SP)
|
100% |
2019
375.5 IPS |
|
Image Classification (HP)
|
100% |
17118
3.18 KIPS |
|
Image Classification (Q)
|
97% |
24516
4.58 KIPS |
|
Image Segmentation (SP)
|
100% |
2136
34.6 IPS |
|
Image Segmentation (HP)
|
100% |
21855
354.3 IPS |
|
Image Segmentation (Q)
|
99% |
30364
492.2 IPS |
|
Pose Estimation (SP)
|
100% |
10044
11.7 IPS |
|
Pose Estimation (HP)
|
100% |
123499
144.1 IPS |
|
Pose Estimation (Q)
|
95% |
232499
272.5 IPS |
|
Object Detection (SP)
|
100% |
2239
177.6 IPS |
|
Object Detection (HP)
|
100% |
18129
1.44 KIPS |
|
Object Detection (Q)
|
91% |
25714
2.05 KIPS |
|
Face Detection (SP)
|
100% |
4015
47.7 IPS |
|
Face Detection (HP)
|
100% |
42948
510.3 IPS |
|
Face Detection (Q)
|
100% |
63832
758.5 IPS |
|
Depth Estimation (SP)
|
100% |
8314
64.1 IPS |
|
Depth Estimation (HP)
|
99% |
99785
768.8 IPS |
|
Depth Estimation (Q)
|
85% |
169360
1.32 KIPS |
|
Style Transfer (SP)
|
100% |
25200
32.4 IPS |
|
Style Transfer (HP)
|
100% |
230431
296.2 IPS |
|
Style Transfer (Q)
|
96% |
297551
384.0 IPS |
|
Image Super-Resolution (SP)
|
100% |
5883
217.2 IPS |
|
Image Super-Resolution (HP)
|
100% |
48629
1.80 KIPS |
|
Image Super-Resolution (Q)
|
99% |
82981
3.06 KIPS |
|
Text Classification (SP)
|
100% |
2922
3.90 KIPS |
|
Text Classification (HP)
|
100% |
5881
7.85 KIPS |
|
Text Classification (Q)
|
93% |
5992
8.04 KIPS |
|
Machine Translation (SP)
|
100% |
3902
67.2 IPS |
|
Machine Translation (HP)
|
100% |
16441
283.2 IPS |
|
Machine Translation (Q)
|
98% |
17981
310.7 IPS |