Geekbench ML Score

649
TensorFlow Lite GPU Inference Score
Geekbench ML 0.6.0 for Android AArch64

Result Information

Upload Date October 21 2024 04:44 PM
Views 1

System Information

System Information
Operating System Android 14
Model Xiaomi 2201122C
Model IDXiaomi 2201122C
Motherboard taro
Governorwalt
CPU Information
Name ARM ARMv8
Topology 1 Processor, 8 Cores
Identifier ARM implementer 65 architecture 8 variant 2 part 3400 revision 0
Base Frequency 1.78 GHz
Cluster 14 Cores @ 1.79 GHz
Cluster 23 Cores @ 2.50 GHz
Cluster 31 Core @ 3.00 GHz
Memory Information
Size 10.93 GB
Inference Information
Framework TensorFlow Lite
Backend GPU
Device Default

Inference Performance

Workload Accuracy Score
Image Classification (F32)
100% 351
65.7 IPS
 
Image Classification (F16)
100% 510
95.4 IPS
 
Image Classification (I8)
97% 476
89.2 IPS
 
Image Segmentation (F32)
100% 638
10.6 IPS
 
Image Segmentation (F16)
100% 943
15.8 IPS
 
Image Segmentation (I8)
98% 884
14.8 IPS
 
Pose Estimation (F32)
100% 561
0.68 IPS
 
Pose Estimation (F16)
100% 1041
1.26 IPS
 
Pose Estimation (I8)
100% 983
1.19 IPS
 
Object Detection (F32)
100% 257
19.2 IPS
 
Object Detection (F16)
100% 409
30.5 IPS
 
Object Detection (I8)
62% 364
27.2 IPS
 
Face Detection (F32)
100% 1198
14.2 IPS
 
Face Detection (F16)
97% 1796
21.4 IPS
 
Face Detection (I8)
87% 1628
19.4 IPS
 
Depth Estimation (F32)
100% 851
6.60 IPS
 
Depth Estimation (F16)
100% 1544
12.0 IPS
 
Depth Estimation (I8)
95% 1516
11.8 IPS
 
Style Transfer (F32)
100% 1866
2.45 IPS
 
Style Transfer (F16)
100% 3792
4.99 IPS
 
Style Transfer (I8)
98% 3762
4.95 IPS
 
Image Super-Resolution (F32)
100% 295
10.5 IPS
 
Image Super-Resolution (F16)
100% 378
13.5 IPS
 
Image Super-Resolution (I8)
98% 449
16.0 IPS
 
Text Classification (F32)
100% 166
239.1 IPS
 
Text Classification (F16)
100% 166
238.1 IPS
 
Text Classification (I8)
92% 296
426.1 IPS
 
Machine Translation (F32)
100% 300
5.52 IPS
 
Machine Translation (F16)
100% 303
5.57 IPS
 
Machine Translation (I8)
64% 355
6.53 IPS
 

Compare

Set Baseline Geekbench 6 now available