A zero-dependency neural network implementation

Contains an example of training a XOR-problem and MNIST. Mnist dataset (csv) needs to be loaded into the root of the project from https://www.kaggle.com/datasets/oddrationale/mnist-in-csv

• not included in repo, since it is huge.

Also contains rough implementations for these optimizers:

• gradient descent
• momentum
• nesterov momentum
• AdaGrad
• AdaDelta
• RMSProp
• Adam