The ESP32 folder contains the client and server-side codes.
The Simulations folder contains the main code for creating machine learning models, evaluating them, and then converting them into their tiny versions. Each folder contains its own subsample of the dataset. The fold6-complex refers to the TinyML robustness subsection of the paper, which aims to evaluate the effect of the TensorFlow Lite (TFLite) library on converting more complex models into their tiny versions.
All the results, including tables and plots, are located within the Results folder.