The deployment environments of a machine learning (ML) model are changing. In recent years, we went from locally training models and running them on standalone scripts to deploying them in massive and specialized setups. However, the industry hasn’t been focusing only on large-scaled-productionized ML, but also its small, portable, and accessible counterpart — for machine learning has found a place in embedded systems.
Improving machine learning involves more than making the algorithms smarter and larger. As the field improves, there has also been an improvement in their speed, size, and computational efficiency. This advance led to TinyML, the subfield of machine learning concerning models in power-constraint devices like microprocessors.
This article explains how you can create a TensorFlow model for detecting the Pokémon Pikachu and Bulbasaur on an Arduino NANO 33 BLE Sense microprocessor. The content we will see here covers the data collection procedure (done on the Arduino), a brief data analysis, training the model, how to convert said model to a format understood by the Arduino’s TensorFlow Lite for Microcontrollers library, and how to deploy it on the device.

#machine-learning #towards-data-science #programming #tensorflow #arduino