Overview Getting started Download and build the sample application Project structure Deploy the example to your STM32F7 Extend the program Retrain the machine learning model Troubleshooting Next steps Supplementary information: model training
Single Page Download PDF

Getting started

Before you begin

Here is what you will need to complete the guide:

  • Computer that supports Mbed CLI
  • STM32F7 discovery kit board
  • Mini-USB cable
  • Python 2.7. Using pyenv is recommended to manage python versions.
  • Serial port terminal programs. On Windows, examples are install programs like CoolTerm or Putty.

Getting started

TensorFlow Lite for Microcontrollers supports several devices out of the box, and is relatively easy to extend to new devices. For this guide, we will focus on the STM32F7 discovery kit.

STM32F7 discovery kit image

We will deploy a sample application that uses the microphone on the STM32F7 and a TensorFlow machine learning model to detect the words “yes” and “no”.

To do this, we will show you how to complete the following steps:

  1. Download and build the sample application
  2. Deploy the sample to your STM32F7
  3. Make some code changes to utilize the LCD display on the board
  4. Use new trained models to recognize different words
Previous Next