Upload Date | August 30 2024 04:09 AM |
Views | 1 |
System Information | |
---|---|
Operating System | iOS 16.6 |
Model | iPad Pro (12.9-inch 3rd Generation) |
Model ID | iPad8,5 |
Motherboard | J320AP |
CPU Information | |
---|---|
Name | Apple A12X Bionic |
Topology | 1 Processor, 8 Cores |
Identifier | ARM |
Base Frequency | 2.49 GHz |
Cluster 1 | 4 Cores |
Cluster 2 | 4 Cores |
L1 Instruction Cache | 48.0 KB x 1 |
L1 Data Cache | 32.0 KB x 1 |
L2 Cache | 2.00 MB x 1 |
Memory Information | |
---|---|
Size | 3.66 GB |
Inference Information | |
---|---|
Framework | Core ML |
Backend | Neural Engine |
Device | Default |
Workload | Accuracy | Score | |
---|---|---|---|
Image Classification (F32)
|
100% |
867
162.3 IPS |
|
Image Classification (F16)
|
100% |
2258
422.6 IPS |
|
Image Classification (I8)
|
99% |
2270
424.7 IPS |
|
Image Segmentation (F32)
|
100% |
352
5.87 IPS |
|
Image Segmentation (F16)
|
100% |
393
6.56 IPS |
|
Image Segmentation (I8)
|
100% |
394
6.58 IPS |
|
Pose Estimation (F32)
|
100% |
3014
3.65 IPS |
|
Pose Estimation (F16)
|
100% |
25462
30.8 IPS |
|
Pose Estimation (I8)
|
100% |
25555
30.9 IPS |
|
Object Detection (F32)
|
100% |
873
65.2 IPS |
|
Object Detection (F16)
|
100% |
1091
81.5 IPS |
|
Object Detection (I8)
|
97% |
887
66.2 IPS |
|
Face Detection (F32)
|
100% |
3394
40.4 IPS |
|
Face Detection (F16)
|
97% |
3923
46.6 IPS |
|
Face Detection (I8)
|
96% |
3733
44.4 IPS |
|
Depth Estimation (F32)
|
100% |
3197
24.8 IPS |
|
Depth Estimation (F16)
|
100% |
19931
154.6 IPS |
|
Depth Estimation (I8)
|
99% |
20243
157.0 IPS |
|
Style Transfer (F32)
|
24% |
8138
10.7 IPS |
|
Style Transfer (F16)
|
30% |
11199
14.7 IPS |
|
Style Transfer (I8)
|
30% |
11172
14.7 IPS |
|
Image Super-Resolution (F32)
|
100% |
1691
60.4 IPS |
|
Image Super-Resolution (F16)
|
100% |
3807
135.9 IPS |
|
Image Super-Resolution (I8)
|
100% |
3755
134.1 IPS |
|
Text Classification (F32)
|
100% |
218
313.1 IPS |
|
Text Classification (F16)
|
100% |
259
371.8 IPS |
|
Text Classification (I8)
|
96% |
259
372.2 IPS |
|
Machine Translation (F32)
|
100% |
500
9.21 IPS |
|
Machine Translation (F16)
|
100% |
458
8.42 IPS |
|
Machine Translation (I8)
|
99% |
500
9.21 IPS |