Initial Setup



In [1]:

    
## Setup the path for our codebase
import sys
sys.path.append( '../code/' )



In [2]:

    
%matplotlib inline
import matplotlib.pyplot as plt

Example Data (centered Quadratic)



In [3]:

    
import neural_network.simple as simple



In [4]:

    
data = simple.generate_hill_data(100)



In [5]:

    
xs = map(lambda z: z[0], data)
ys = map(lambda z: z[1], data)
plt.plot( xs, ys )









    Out[5]:





[<matplotlib.lines.Line2D at 0x7c00cf8>]



In [6]:

    
nn = simple.SimpleScalarF_1Layer(hidden_layer_size=3)
simple.plot_fits( nn, data, train_epochs=200 )

Let's visualize the inputs to the final layer



In [7]:

    
nn.visualize_inputs_to_final_layer(data)









    





<matplotlib.figure.Figure at 0x7ad2eb8>



In [8]:

    
nn = simple.SimpleScalarF_1Layer(hidden_layer_size=10)
simple.plot_fits( nn, data, train_epochs=200 )



In [9]:

    
nn = simple.SimpleScalarF_1Layer(hidden_layer_size=10,learning_rate=0.001)
simple.plot_fits( nn, data, train_epochs=20 )



In [10]:

    
nn = simple.SimpleScalarF_1Layer(hidden_layer_size=10,learning_rate=0.001)
simple.plot_fits( nn, data, train_epochs=20 )



In [11]:

    
nn = simple.SimpleScalarF_1Layer(hidden_layer_size=10,learning_rate=0.001)
simple.plot_fits( nn, data, train_epochs=2000 )



In [12]:

    
nn.visualize_inputs_to_final_layer(data)









    





<matplotlib.figure.Figure at 0x8db1da0>