Geekbench ML Score

89
TensorFlow Lite NNAPI Inference Score
Geekbench ML 0.5.0 for Android AArch64

Result Information

User GyBHard
Upload Date January 20 2022 05:31 AM
Views 5

System Information

System Information
Operating System Android 12
Model Raspberry Raspberry Pi 4
Model IDRaspberry Raspberry Pi 4
Motherboard
Governorperformance
CPU Information
Name ARM BCM2835
Topology 1 Processor, 4 Cores
Identifier ARM implementer 65 architecture 8 variant 0 part 3336 revision 3
Base Frequency 1.50 GHz
Cluster 14 Cores @ 1.50 GHz
Memory Information
Size 3.70 GB
Inference Information
Framework TensorFlow Lite
Backend NNAPI
Device

Inference Performance

Workload Accuracy Score
Image Classification (F32)
100% 78
8.36 IPS
 
Image Classification (F16)
100% 78
8.37 IPS
 
Image Classification (I8)
81% 87
9.37 IPS
 
Image Segmentation (F32)
100% 135
0.83 IPS
 
Image Segmentation (F16)
100% 135
0.83 IPS
 
Image Segmentation (I8)
89% 157
0.97 IPS
 
Pose Estimation (F32)
100% 53
0.21 IPS
 
Pose Estimation (F16)
100% 53
0.22 IPS
 
Pose Estimation (I8)
80% 27
0.11 IPS
 
Object Detection (F32)
100% 100
3.49 IPS
 
Object Detection (F16)
100% 100
3.49 IPS
 
Object Detection (I8)
52% 108
3.78 IPS
 
Face Detection (F32)
100% 105
1.21 IPS
 
Face Detection (F16)
100% 107
1.23 IPS
 
Face Detection (I8)
75% 117
1.34 IPS
 
Text Classification (F32)
100% 132
52.3 IPS
 
Text Classification (F16)
100% 132
52.2 IPS
 
Machine Translation (F32)
100% 69
10.6 IPS
 
Machine Translation (F16)
100% 68
10.5 IPS
 

Compare

Set Baseline Geekbench 6 now available