A Neural Network implementation using numpy. The current program automatically
plots the loss for training for AND, OR, and XOR functions. You can assign
different input or number of layers for the network, and you can also assign 
different nodes for each hidden layers. The momentum is enabled. To do this,
you just need to assign momentum to the agent init function.

To initialize an agent, call constructor for NeuralMMAgent. 

Key functions for debugging:

test_agent.set_weights(): sets weights accordingly. Each layer is one element
in the list. Each element contains all of the weights in that layer. For 
instance, 
test_agent.set_weights([[-.37, .26, .10, -0.24], [-0.01, -0.05]])
represents the example in class.

test_agent.set_thetas(): sets thetas. 
test_agent.set_thetas([[.0, .0], [.0]]) represents the example in class. 

train_net(self, input_list, output_list, max_num_epoch=100000,
                  max_sse=0.1):
        ''' Trains neural net using incremental learning
            (update once per input-output pair)
            Arguments:
                input_list -- 2D list of inputs
                output_list -- 2D list of outputs matching inputs
                max_num_epoch -- int for which the nn stops training
                max_sse -- float for which the nn stops training
        '''
        
print_weights_thetas_deltas():
    prints necessary numbers including weights, thetas, change in these values
    and delta.