VisualPerceptron
2019, Oct 01
The VisualPerceptron is a project meant for visualization of the learning process of a Perceptron using gradient descent. It was originally designed to be used as an example of such algorithms in a class I taught about Neural Networks.
The original implementation uses PyQT to provide a GUI, in which the user can tune some parameters, feed some data visually and watch the Perceptron to learn (or not) how to classify the given data.
All the Perceptron algorithm is implemented using no library other than Numpy, as to be a mathematical display of how neural networks work.
You can check it in this repository.