N.N.N. V.0.1-alpha
Added: Gradient descent; 4 input nodes, 2 hidden layers each with 5 nodes, 2 output nodes (67 parameters); changeable learning rate, tolerance, and time limit (more accurate).
Note: Play around with different values, see what happens. We'll be switching to a more efficient method of node generation in the next version (hopefully).
Documentation can be found here.