X-LINUX-AI application samples zoo

Applicable for

STM32MP13x lines, STM32MP15x lines, STM32MP25x lines


Image classification using TensorFlow Lite C++ Image classification using TensorFlow Lite Python	Image classification using Coral Edge TPU TensorFlow Lite C++ Image classification using Coral Edge TPU TensorFlow Lite Python	Image classification using ONNX Python runtime	Image classification using OpenVX C++ API (NBG)
Object detection using TensorFlow Lite C++ Object detection using TensorFlow Lite Python	Object detection using Coral Edge TPU TensorFlow Lite C++ Object detection using Coral Edge TPU TensorFlow Lite Python	Object detection using ONNX C++ runtime Object detection using ONNX Python runtime
Face recognition using TensorFlow Lite C++
Human Pose Estimation using TensorFlow Lite Python
Semantic Segmentation using TensorFlow Lite Python

1. Performance Comparison[edit | edit source]↑

	STM32MP15x CPU	STM32MP25 Alpha NPU	STM32MP25 Beta NPU
MobilenetV1_0.5_128_quant.tflite	22 FPS	255 FPS	515 FPS
SSD_MobilenetV1_1.0_300_quant.tflite	2 FPS	14 FPS	40 FPS
STModel_FaceReco	5 FPS	10 FPS	15 FPS
Movenet_SinglePose_Lightning_int8_4.tflite	1.8 FPS	7 FPS (executed on GPU only)	14 FPS (executed on GPU only)
DeepLabV3_quant.tflite	1 FPS	5 FPS	16 FPS