Geekbench ML Score

483
TensorFlow Lite CPU Inference Score
Geekbench ML 0.6.0 for Android AArch64

Result Information

Upload Date July 27 2024 01:05 AM
Views 1

System Information

System Information
Operating System Android 12
Model Xiaomi Mi 10T 5G
Model IDXiaomi M2007J3SY
Motherboard apollo
Governorschedutil
CPU Information
Name ARM Qualcomm
Topology 1 Processor, 8 Cores
Identifier ARM implementer 65 architecture 8 variant 1 part 3341 revision 0
Base Frequency 1.80 GHz
Cluster 14 Cores @ 1.80 GHz
Cluster 23 Cores @ 2.42 GHz
Cluster 31 Core @ 2.84 GHz
Memory Information
Size 7.43 GB
Inference Information
Framework TensorFlow Lite
Backend CPU
Device Default

Inference Performance

Workload Accuracy Score
Image Classification (F32)
100% 259
48.5 IPS
 
Image Classification (F16)
100% 259
48.5 IPS
 
Image Classification (I8)
97% 563
105.4 IPS
 
Image Segmentation (F32)
100% 371
6.19 IPS
 
Image Segmentation (F16)
100% 369
6.16 IPS
 
Image Segmentation (I8)
98% 393
6.57 IPS
 
Pose Estimation (F32)
100% 546
0.66 IPS
 
Pose Estimation (F16)
100% 539
0.65 IPS
 
Pose Estimation (I8)
100% 2206
2.67 IPS
 
Object Detection (F32)
100% 273
20.4 IPS
 
Object Detection (F16)
100% 268
20.0 IPS
 
Object Detection (I8)
61% 734
54.8 IPS
 
Face Detection (F32)
100% 593
7.05 IPS
 
Face Detection (F16)
100% 589
7.00 IPS
 
Face Detection (I8)
86% 1246
14.8 IPS
 
Depth Estimation (F32)
100% 502
3.90 IPS
 
Depth Estimation (F16)
100% 492
3.82 IPS
 
Depth Estimation (I8)
95% 991
7.68 IPS
 
Style Transfer (F32)
100% 833
1.10 IPS
 
Style Transfer (F16)
100% 822
1.08 IPS
 
Style Transfer (I8)
98% 1951
2.57 IPS
 
Image Super-Resolution (F32)
100% 289
10.3 IPS
 
Image Super-Resolution (F16)
100% 298
10.6 IPS
 
Image Super-Resolution (I8)
98% 1124
40.1 IPS
 
Text Classification (F32)
100% 76
109.8 IPS
 
Text Classification (F16)
100% 280
402.9 IPS
 
Text Classification (I8)
92% 236
339.5 IPS
 
Machine Translation (F32)
100% 578
10.6 IPS
 
Machine Translation (F16)
100% 594
10.9 IPS
 
Machine Translation (I8)
64% 201
3.69 IPS
 

Compare

Set Baseline Geekbench 6 now available