import numpy as np

def sigmoid(x):
    return 1/(1+np.exp(-x))

def deriv_sigmoid(x):
    return x*(1-x)

# training

# input data

x = np.array([[0, 0, 1],
              [0, 1, 1],
              [1, 0, 1],
              [1, 1, 1]])

# output data

y = np.array([[0],
              [1],
              [1],
              [0]])

np.random.seed(1)

syn_0 = 2*np.random.random((3,2)) - 1
syn_1 = 2*np.random.random((2,1)) - 1

max_iter = 1000

for i in range(max_iter):

    l0 = x
    l1 = sigmoid(np.dot(l0, syn_0))
    l2 = sigmoid(np.dot(l1, syn_1))

    l2_error = y - l2

    if(i % 1000) == 0:
        print("Error: %s" %np.mean(np.abs(l2_error)))

    l2_delta = l2_error * deriv_sigmoid(l2)

    syn_1 += l1.T.dot(l2_delta)

    l1_error = l2_delta.dot(syn_1.T)

    l1_delta = l1_error * deriv_sigmoid(l1)

    syn_0 += l0.T.dot(l1_delta)


print("Output after training = %s" %l2)

# testing

x = np.array([[0, 1, 1]])

l0 = x
l1 = sigmoid(np.dot(l0, syn_0))
l2 = sigmoid(np.dot(l1, syn_1))

print("testing output = %s" %l2)