This is an ESP32 board with OLED display. It's cool.
I was trying to program this board by Zerynth Studio, but the display on this board uses I2C, while the libraries for Zerynth Studio works only with SPI interface. In order to avoid to manage the I2C protocol from the base class, I had to move to Arduino IDE and forget Python.
Has been a long time since I used Arduino Ide and I discovered an interesting neural network implementation, it is distributed as Arduino library: Neurona.
This library implements a supervised neural network for Arduino boards and also it works for ESP32, my crush board. The usage of the library is quite simple, you initialize it with the number of inputs, the number of hidden layers and the number of outputs. You have to provide the trained weights, it does not train on Arduino, you have to use an external software to train the network and obtains the weight, then you pass the weight to the library.
