error in plotting in the visualization tutorial

[Ahmed Abuali]

Hello everyone,

I just started this course, and I am not expert in python

 

I tried the visualization code for the one hidden layer with 3 neurons

but I get this error

ValueError: cannot reshape array of size 1 into shape (100,100)

I think this is line that produce the error

img=ax[1].imshow(np.reshape(y_out,[M,M]),origin='lower', extent=[y0range[0],y0range[1],y1range[0],y1range[1]])

Can you help me resolve this error?

[Ahmed Abuali]

here is the call

visualize_network(weights=[ [
    [0.2,0.9],   # weights of 2 input neurons for 1st hidden neuron
    [-0.5,0.3],  # weights of 2 input neurons for 2nd hidden
    [0.8,-1.3]   # weights of 2 input neurons for 3rd hidden
    ],
    [
        [-0.3,0.7,0.5] # weights of 3 hidden neurons for output
    ] ],
    biases=[
        [0.1,-0.5,-0.5], # biases of 3 hidden neurons
        [-.2] # bias for output neuron
            ],
    activations=['jump',# activation for hidden
                 'linear' # activation for output
                ],
    y0range=[-3,3], y1range=[-3,3])

[Ahmed Abuali]

I also discovered that the code is working if I have

activations=['jump',# activation for hidden

                 'jump' # activation for output

instead of

activations=['jump',# activation for hidden
                 'linear' # activation for output

[Ahmed Abuali]

and if I use

activations=['jump',# activation for hidden

                 'sigmoid' # activation for output

it gives the behavior expected from 

activations=['jump',# activation for hidden
                 'linear' # activation for output