A fully connected neural network built from scratch in C++
Implemented via two class types so far. Class 1 = neural network class. Recieves a topology of number of layers and number of neurons per layer and provides the forward propagation and back propagation methods.
Class 2 = neuron class. All calculations are implemented via methods owned by the neuron class such as the calculation of derivatives and changing weights.
to do: Implementation of training loop and create a training data file. May implement this via a training data class.