X-CUBE-AI-Implementation-of-MobileNetv1-using-STM32CubeIDE

MobileNetv1

The model used in this repository is:

• Width multiplier: 0.25
• Image Resolution: 160
• Pretrained MobileNetv1: Tensorflow
• Quantization: 8-bit
• Paper: link

Requirements

The project is made on the following versions:

• STMCubeIDE 1.8.0
• X-CUBE-AI 7.1.0
• Board: NUCLEO STM32H743ZIT6U

Check the versions before implementing, and if you encounter any error please move to the above-mentioned versions. While, the board can be anyone that supports the x-cube-ai library. The official documentation for x-cube-ai is here.

About Project

The project contains the STM32CubeIDE project template files. The files description is as follows:

• Core: Contains the source and include files necessary for the project execution.
• IOC file: "Mobv1_160_0.25_XCUBEAI.ioc" contains the project configurations.
• Middlewares: A generated code by the project configurations in .ioc file for the Mobilenet.
• Validation results: Validation results from the project.
• TensorFlow Lite model: "mobilenet_v1_0.25_160_quant.tflite" imagenet quantized model.

The implementation of other versions of MobileNetv1 will be coming soon.