This is an implementation of a fully-connected neural network from scratch in python.

NeuralNetwork.py Contains the class of the neural network.

Hiddenlayer.py Contains the class of a hidden layer and the input weights of to the layer.

Inputlayer.py Contains the class of an input layer.

Currently, the neural network only supports the sigmoid and relu activationfunctions and the squared mean error.

To use the implementation, you only need Python version 3.7.4 (or later) and NumPy version 1.18.1 (or later).

An example on how to use the implementation can be seen in XOR.py, which solves the famous XOR-problem.