A simple implementation of auto differentiation system. I have made this before, however, the old project is poorly structured and full of bugs so I decided to create a brand new project.
This is the updated version of NeuralNetwork.
The final goal is to create a deep learning "library" that is able to perform training (and feedforward of course) on Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN), possibly LSTM and GRU.
All of this is going to be based on using a computational graph If you want to see the whole project plan you can hop to TODO list.