Upload Date | May 22 2024 02:51 PM |
Views | 1 |
System Information | |
---|---|
Operating System | Android 14 |
Model | Google Pixel 8 |
Model ID | Google Pixel 8 |
Motherboard | shiba |
Governor | sched_pixel |
CPU Information | |
---|---|
Name | ARM ARMv8 |
Topology | 1 Processor, 9 Cores |
Identifier | ARM implementer 65 architecture 8 variant 1 part 3406 revision 0 |
Base Frequency | 1.70 GHz |
Cluster 1 | 4 Cores @ 1.70 GHz |
Cluster 2 | 4 Cores @ 2.37 GHz |
Cluster 3 | 1 Core @ 2.91 GHz |
Memory Information | |
---|---|
Size | 7.42 GB |
Inference Information | |
---|---|
Framework | TensorFlow Lite |
Backend | GPU |
Device | Default |
Workload | Accuracy | Score | |
---|---|---|---|
Image Classification (F32)
|
100% |
401
75.1 IPS |
|
Image Classification (F16)
|
100% |
386
72.2 IPS |
|
Image Classification (I8)
|
97% |
310
58.0 IPS |
|
Image Segmentation (F32)
|
100% |
741
12.4 IPS |
|
Image Segmentation (F16)
|
100% |
817
13.7 IPS |
|
Image Segmentation (I8)
|
98% |
805
13.5 IPS |
|
Pose Estimation (F32)
|
100% |
461
0.56 IPS |
|
Pose Estimation (F16)
|
100% |
671
0.81 IPS |
|
Pose Estimation (I8)
|
100% |
673
0.82 IPS |
|
Object Detection (F32)
|
100% |
282
21.1 IPS |
|
Object Detection (F16)
|
100% |
346
25.9 IPS |
|
Object Detection (I8)
|
62% |
336
25.1 IPS |
|
Face Detection (F32)
|
100% |
1217
14.5 IPS |
|
Face Detection (F16)
|
98% |
1239
14.7 IPS |
|
Face Detection (I8)
|
86% |
1220
14.5 IPS |
|
Depth Estimation (F32)
|
100% |
928
7.20 IPS |
|
Depth Estimation (F16)
|
100% |
1132
8.78 IPS |
|
Depth Estimation (I8)
|
95% |
1151
8.93 IPS |
|
Style Transfer (F32)
|
100% |
2620
3.45 IPS |
|
Style Transfer (F16)
|
100% |
3226
4.24 IPS |
|
Style Transfer (I8)
|
98% |
3232
4.25 IPS |
|
Image Super-Resolution (F32)
|
100% |
541
19.3 IPS |
|
Image Super-Resolution (F16)
|
100% |
653
23.3 IPS |
|
Image Super-Resolution (I8)
|
98% |
640
22.9 IPS |
|
Text Classification (F32)
|
100% |
384
551.4 IPS |
|
Text Classification (F16)
|
100% |
381
547.8 IPS |
|
Text Classification (I8)
|
92% |
575
826.1 IPS |
|
Machine Translation (F32)
|
100% |
587
10.8 IPS |
|
Machine Translation (F16)
|
100% |
555
10.2 IPS |
|
Machine Translation (I8)
|
64% |
555
10.2 IPS |