Geekbench ML Score

622
TensorFlow Lite NNAPI Inference Score
Geekbench ML 0.6.0 for Android AArch64

Result Information

Upload Date February 23 2024 01:17 AM
Views 1

System Information

System Information
Operating System Android 14
Model Samsung Galaxy S23 Ultra
Model IDsamsung SM-S918U1
Motherboard kalama
Governorwalt
CPU Information
Name ARM ARMv8
Topology 1 Processor, 8 Cores
Identifier ARM implementer 65 architecture 8 variant 1 part 3406 revision 0
Base Frequency 2.02 GHz
Cluster 13 Cores @ 2.02 GHz
Cluster 24 Cores @ 2.80 GHz
Cluster 31 Core @ 3.36 GHz
Memory Information
Size 10.77 GB
Inference Information
Framework TensorFlow Lite
Backend NNAPI
Device Default

Inference Performance

Workload Accuracy Score
Image Classification (F32)
100% 288
54.0 IPS
 
Image Classification (F16)
100% 284
53.2 IPS
 
Image Classification (I8)
97% 709
132.6 IPS
 
Image Segmentation (F32)
100% 347
5.80 IPS
 
Image Segmentation (F16)
100% 346
5.78 IPS
 
Image Segmentation (I8)
98% 713
11.9 IPS
 
Pose Estimation (F32)
100% 483
0.59 IPS
 
Pose Estimation (F16)
100% 476
0.58 IPS
 
Pose Estimation (I8)
100% 1991
2.41 IPS
 
Object Detection (F32)
100% 299
22.3 IPS
 
Object Detection (F16)
100% 293
21.9 IPS
 
Object Detection (I8)
61% 813
60.7 IPS
 
Face Detection (F32)
100% 676
8.03 IPS
 
Face Detection (F16)
100% 683
8.12 IPS
 
Face Detection (I8)
86% 1499
17.8 IPS
 
Depth Estimation (F32)
100% 596
4.62 IPS
 
Depth Estimation (F16)
100% 598
4.64 IPS
 
Depth Estimation (I8)
95% 1707
13.2 IPS
 
Style Transfer (F32)
100% 946
1.24 IPS
 
Style Transfer (F16)
100% 947
1.25 IPS
 
Style Transfer (I8)
98% 2406
3.16 IPS
 
Image Super-Resolution (F32)
100% 344
12.3 IPS
 
Image Super-Resolution (F16)
100% 333
11.9 IPS
 
Image Super-Resolution (I8)
98% 1199
42.8 IPS
 
Text Classification (F32)
100% 421
605.6 IPS
 
Text Classification (F16)
100% 415
596.2 IPS
 
Text Classification (I8)
92% 658
945.8 IPS
 
Machine Translation (F32)
100% 655
12.0 IPS
 
Machine Translation (F16)
100% 638
11.7 IPS
 
Machine Translation (I8)
64% 699
12.9 IPS
 

Compare

Set Baseline Geekbench 6 now available