Just click the green flag, and watch. Will automatically stop when it gets everything right. -SPACE- to /.terminate/ It will usually take around a couple of minutes to finish, but sometimes a lot more, and so I will try to optimize and possibly add some more features to try and fix that.
v.0.1: Just a plain backpropagation algorithm using the hyperbolic tangent (tanh) activation function that acts on a network of 5 neurons, 2 inputs. 2 hidden neurons, and 1 output node to solve the XOR problem. Using a TON of handmade variables, so optimization would be in the form of using lists instead, to hold the weights, biases, activations, gradients, etc. v.0.2: So far just changed the conditions to make them more accurate.
