Geekbench ML Score

1986
TensorFlow Lite GPU Inference Score
Geekbench ML 0.6.0 for Android AArch64

Result Information

Upload Date November 08 2024 04:47 PM
Views 1

System Information

System Information
Operating System Android 15
Model asus ASUSAI2501E
Model IDasus ASUSAI2501E
Motherboard sun
Governorwalt
CPU Information
Name Qualcomm ARMv8
Topology 1 Processor, 8 Cores
Identifier ARM implementer 81 architecture 8 variant 3 part 1 revision 4
Base Frequency 3.53 GHz
Cluster 16 Cores @ 3.53 GHz
Cluster 22 Cores @ 4.32 GHz
Memory Information
Size 22.51 GB
Inference Information
Framework TensorFlow Lite
Backend GPU
Device Default

Inference Performance

Workload Accuracy Score
Image Classification (F32)
100% 700
130.9 IPS
 
Image Classification (F16)
100% 1906
356.6 IPS
 
Image Classification (I8)
100% 1826
341.6 IPS
 
Image Segmentation (F32)
100% 1565
26.1 IPS
 
Image Segmentation (F16)
100% 3340
55.8 IPS
 
Image Segmentation (I8)
98% 3189
53.3 IPS
 
Pose Estimation (F32)
100% 1531
1.85 IPS
 
Pose Estimation (F16)
100% 2372
2.87 IPS
 
Pose Estimation (I8)
100% 2390
2.89 IPS
 
Object Detection (F32)
100% 532
39.7 IPS
 
Object Detection (F16)
100% 1116
83.3 IPS
 
Object Detection (I8)
59% 1048
78.2 IPS
 
Face Detection (F32)
100% 2866
34.1 IPS
 
Face Detection (F16)
97% 6131
72.9 IPS
 
Face Detection (I8)
86% 5733
68.2 IPS
 
Depth Estimation (F32)
100% 1559
12.1 IPS
 
Depth Estimation (F16)
100% 4308
33.4 IPS
 
Depth Estimation (I8)
94% 4388
34.0 IPS
 
Style Transfer (F32)
100% 5687
7.48 IPS
 
Style Transfer (F16)
100% 11946
15.7 IPS
 
Style Transfer (I8)
98% 11946
15.7 IPS
 
Image Super-Resolution (F32)
100% 668
23.9 IPS
 
Image Super-Resolution (F16)
100% 2000
71.4 IPS
 
Image Super-Resolution (I8)
98% 1926
68.8 IPS
 
Text Classification (F32)
100% 660
948.4 IPS
 
Text Classification (F16)
100% 660
948.2 IPS
 
Text Classification (I8)
92% 1024
1.47 KIPS
 
Machine Translation (F32)
100% 1120
20.6 IPS
 
Machine Translation (F16)
100% 1106
20.3 IPS
 
Machine Translation (I8)
64% 1165
21.4 IPS
 

Compare

Set Baseline Geekbench 6 now available