data = loadmat('ex4weights.mat')
Theta = [data['Theta1'], data['Theta2']]
theta = np.block([t.reshape(t.size, order='F') for t in Theta])
nn = NeuralNetwork(X, y, Theta, layer_dims, 0)
print('========== Part 1.3: Feedforward and Cost Function ==========')
J = nn.cost(theta)
print('Cost: {0:0.6f} (expected: 0.287629)'.format(J))
print('========== Part 1.4: Regularized Cost Function ==========')
nn.update_lambda(1)
J = nn.cost(theta)
print('Regularized cost: {0:0.6f} (expected: 0.383770)'.format(J))
print('========== Part 2.1: Sigmoid Gradient ==========')
g = nn.sigmoid_gradient(np.array([-1.0, -0.5, 0, 0.5, 1.0]))
e_g = [0.19661193, 0.23500371, 0.250000, 0.23500371, 0.19661193]
print('---------------------')
print(' Actual | Expected')
print('---------------------')
for i, j in zip(g, e_g):
print('{0: .6f} | {1: .6f}'.format(i, j))
diff = np.linalg.norm(e_g - g) / np.linalg.norm(e_g + g)
print('Relative difference: {0}'.format(diff))
print('========== Part 2.4: Gradient Checking ==========')
sample_dims = [3, 5, 3]
sample_m = 5
X_s = nn.initialize_weights(sample_dims[0] - 1, sample_m, None, debug=True)
y_s = 1 + (np.arange(1, sample_m + 1) % sample_dims[-1])
Theta = [
