The tinyML examples for the ESP32 written in C++
All tinyML examples consist of two parts:
- Creating of the model
- The application on the micro-controller using the model.
In general the micro-controller application will create input data to the model, e.g. by reading its hardware devices like the camera, the microphone or the accelerometer. These data serve as input to the TensorFlow model. Data output by the model are converted into some output data. In case of the person detection demo an LED might be lit if a person is seen by the camera while it is switched off when no person is in sight. For the wake word detection a different LED may be lit for each word detected.
In the following we go through each of the examples, building them step be step, starting from the creation of the TensorFlow model to the implementation of the micro-controller application.
--
Uli Raich - 2023-09-01
Comments