AI-ANNE (2.0)

(A) (N)eural (N)et...

AI-ANNE enables resource-efficient deep learning models on microcontrollers as embedded systems. After pre-training with TensorFlow and Keras, the parameters in MicroPython can be transferred to AI-ANNE for offline and real-time monitoring.

...for (E)xploration

AI-ANNE is also a didactic tool for universities, vocational schools and general schools. The aim is to convey the functionality of neural networks (activation functions, parameters, etc.) in a simple way using classic data science examples..


MIT License. Information and free download:

Introduction to Neural Networks

Already equipped with two functional neural networks, AI-ANNE can be flexibly adapted for practical use as well as for teaching and learning settings. This is about the neurons, layers and functions of neural networks.

Mathematical Basics

AI-ANNE provides a clear introduction to the mathematical principles of neural networks and explains how they work using practical examples. Sigmoid, Tanh, ReLU, Leaky ReLU and Softmax can be tried out directly as functions.

Code with MicroPython

With AI-ANNE, the mathematical principles can be easily programmed in MicroPython and implemented directly for use on a microcontroller. For example, a matrix can be transposed and functions programmed.

  • LinkedIn
  • Xing
  • GitHub
  • YouTube
  • UTB

STATISTICAL THINKING (2025) | LEGAL DISCLOSURE (PDF)
PRIVACY POLICY (PDF) | DESIGN: HTML5 UP (LINK) | No Cookies. No Tracking.