Create a Program to Implement Cross Entropy in Python Assignment Solution.

import numpy as np

from keras.datasets import mnist

from keras.utils import np_utils

from dense import Dense

from softmax import Softmax

from convolutional import Convolutional

from reshape import Reshape

from activations import Tanh, Sigmoid

from losses import binary_cross_entropy, binary_cross_entropy_prime

def preprocess_data(x, y, limit):

'''

Will limit our data since using the whole thing will take forever on a cpu especially since we're

implementing this from scratch.

'''

x = x.reshape(len(x), 1, 28, 28)

x = x.astype("float32") / 255

y = np_utils.to_categorical(y)

y = y.reshape(len(y), 10, 1)

return x[:limit], y[:limit]

# load MNIST from server, limit to 100 images per class since we're not training on GPU

(x_train, y_train), (x_test, y_test) = mnist.load_data()

x_train, y_train = preprocess_data(x_train, y_train, 100)

x_test, y_test = preprocess_data(x_test, y_test, 100)

# TODO: Add our layers and the flow of input into this list.

network = [

Convolutional((1, 28, 28), 3, 5),

Sigmoid(),

Reshape((5, 26, 26), (5 * 26 * 26, 1)),

Dense(5 * 26 * 26, 100),

Sigmoid(),

Dense(100, 100),

Sigmoid(),

Dense(100, 10),

Softmax()

]

epochs = 100

learning_rate = 0.1

#train

for e in range(epochs):

error = 0

for x, y in zip(x_train, y_train):

# forward

output = x

for layer in network:

output = layer.forward(output)

# TODO: update our error

error += binary_cross_entropy(y,output)

grad = binary_cross_entropy_prime(y,output)

# TODO: perform back prop

for layer in reversed(network):

grad = layer.backward(grad,learning_rate)

error /= len(x_train)

print(f"{e + 1}/{epochs}, error={error}")

# TODO: run the test data through and print out your predictions

for x, y in zip(x_test, y_test):

output = x

for layer in network:

output = layer.forward(output)

print(f"y_pred: {np.argmax(output)}, y_true: {np.argmax(y)}")