Geekbench ML Score

385
TensorFlow Lite GPU Inference Score
Geekbench ML 0.5.2 for iOS AArch64

Result Information

Upload Date September 07 2024 12:51 PM
Views 1

System Information

System Information
Operating System iOS 15.2
Model iPad Air 2
Model IDiPad5,4
Motherboard J82AP
CPU Information
Name Apple A8X
Topology 1 Processor, 3 Cores
Identifier ARM
Base Frequency 1.50 GHz
L1 Instruction Cache64.0 KB x 1
L1 Data Cache64.0 KB x 1
L2 Cache2.00 MB x 1
Memory Information
Size 1.94 GB
Inference Information
Framework TensorFlow Lite
Backend GPU
Device

Inference Performance

Workload Accuracy Score
Image Classification (F32)
100% 508
54.4 IPS
 
Image Classification (F16)
100% 535
57.3 IPS
 
Image Classification (I8)
81% 149
16.0 IPS
 
Image Segmentation (F32)
100% 887
5.48 IPS
 
Image Segmentation (F16)
100% 1136
7.02 IPS
 
Image Segmentation (I8)
89% 278
1.72 IPS
 
Pose Estimation (F32)
100% 310
1.25 IPS
 
Pose Estimation (F16)
100% 543
2.20 IPS
 
Pose Estimation (I8)
80% 52
0.21 IPS
 
Object Detection (F32)
100% 623
21.9 IPS
 
Object Detection (F16)
100% 809
28.4 IPS
 
Object Detection (I8)
52% 212
7.43 IPS
 
Face Detection (F32)
100% 852
9.79 IPS
 
Face Detection (F16)
97% 1212
13.9 IPS
 
Face Detection (I8)
75% 206
2.37 IPS
 
Text Classification (F32)
100% 330
130.6 IPS
 
Text Classification (F16)
98% 391
154.7 IPS
 
Machine Translation (F32)
100% 215
33.1 IPS
 
Machine Translation (F16)
100% 218
33.5 IPS
 

Compare

Set Baseline Geekbench 6 now available