perceptron.py

import numpy as np


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


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


training_inputs = np.array([[0, 0, 1], [1, 1, 1], [1, 0, 1], [0, 1, 1]])
training_outputs = np.array([[0, 1, 1, 0]]).T

np.random.seed(1)

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

print('Synaptic weights in the beginning:')
print(synaptic_weights)

for i in range(20000):
    input_layer = training_inputs
    outputs = sigmoid(np.dot(input_layer, synaptic_weights))

    error = training_outputs - outputs
    adjustments = error * sigmoid_der(outputs)

    synaptic_weights += np.dot(input_layer.T, adjustments)

print('Synaptic weights after training:')
print(synaptic_weights)
print('After training, the Outputs are:')
print(outputs)