Creative Hardware
TinyML: Getting Started With Running a Model on a Microcontroller
Putting a trained model on an ESP32 or an RP2040 is a five-step pipeline with one genuinely new concept in it — the tensor arena, a block of RAM you allocate by hand and tune by trial. This tutorial walks the whole path, explains why int8 quantisation is non-negotiable rather than an optimisation, and untangles the TensorFlow Lite / LiteRT / tflite-micro naming, which currently misleads almost everyone who searches for it.