Neural network that can approximate both single variable and multivariable continuous functions. Neural Network has an input layer, two hidden layers and an output layer, with an arbitrary amount of nodes (dictated by the amount of weights in the checkpoints file). Layer 1 nodes use the function (X^W1 + B1, X is the sum of inputs, W1 is the weight of a Layer 1 node, and B1 is a bias of a Layer 1 node) and Layer 2 nodes use the function (X*W2 + B1, X is the sum of Layer 1 nodes, W2 is the weight of a Layer 2 node, and B2 is a bias of a Layer 2 node).