rnn的一个例子

直接po代码，简单的rnn的adder，改自github，不用TensorFlow等框架，可实现多位（超过8位）。

import copy, numpy as np

np.random.seed(0)

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

def sigmoid_output_to_derivative(output):
    return output * (1 - output)

int2binary = {}
binary_dim = 16
binary = []
def generateInt2Binary(largest_number):
    for i in range(largest_number):
        c = (bin(i).replace('0b','').zfill(binary_dim))
        binary.append([int(a) for a in c])

largest_number = pow(2, binary_dim)
generateInt2Binary(largest_number)

for i in range(largest_number):
    int2binary[i] = binary[i]

alpha = 0.5
input_dim = 2
hidden_dim = 32
output_dim = 1

synapse_0 = 2 * np.random.random((input_dim, hidden_dim)) - 1
synapse_1 = 2 * np.random.random((hidden_dim, output_dim)) - 1
synapse_h = 2 * np.random.random((hidden_dim, hidden_dim)) - 1

synapse_0_update = np.zeros_like(synapse_0)
synapse_1_update = np.zeros_like(synapse_1)
synapse_h_update = np.zeros_like(synapse_h)

print('start train...')
for j in range(10000):

    a_int = np.random.randint(largest_number/2)
    a = int2binary[a_int]
    b_int = np.random.randint(largest_number/2)
    b = int2binary[b_int]

    c_int = a_int + b_int
    c = int2binary[c_int]

    d = np.zeros_like(c)

    overallError = 0

    layer_2_deltas = list()
    layer_1_values = list()
    layer_1_values.append(np.zeros(hidden_dim))

    for position in range(binary_dim):
        X = np.array([[a[binary_dim - position - 1], b[binary_dim - position - 1]]])
        y = np.array([[c[binary_dim - position - 1]]]).T

        layer_1 = sigmoid(np.dot(X, synapse_0) + np.dot(layer_1_values[-1], synapse_h))

        layer_2 = sigmoid(np.dot(layer_1, synapse_1))

        layer_2_error = y - layer_2
        layer_2_deltas.append((layer_2_error) * sigmoid_output_to_derivative(layer_2))
        overallError += np.abs(layer_2_error[0])

        d[binary_dim - position - 1] = np.round(layer_2[0][0])

        layer_1_values.append(copy.deepcopy(layer_1))

    future_layer_1_delta = np.zeros(hidden_dim)

    for position in range(binary_dim):
        X = np.array([[a[position], b[position]]])
        layer_1 = layer_1_values[-position - 1]
        prev_layer_1 = layer_1_values[-position - 2]

        layer_2_delta = layer_2_deltas[-position - 1]
        layer_1_delta = (future_layer_1_delta.dot(synapse_h.T) + layer_2_delta.dot(
            synapse_1.T)) * sigmoid_output_to_derivative(layer_1)

        synapse_1_update += np.atleast_2d(layer_1).T.dot(layer_2_delta)
        synapse_h_update += np.atleast_2d(prev_layer_1).T.dot(layer_1_delta)
        synapse_0_update += X.T.dot(layer_1_delta)

        future_layer_1_delta = layer_1_delta

    synapse_0 += synapse_0_update * alpha
    synapse_1 += synapse_1_update * alpha
    synapse_h += synapse_h_update * alpha

    synapse_0_update *= 0
    synapse_1_update *= 0
    synapse_h_update *= 0

    if (j % 1000 == 0):
        print('this is the %d times test...' % j)
        print("ErrorRate is: %.2f" % overallError)
        out = 0
        for index, x in enumerate(reversed(d)):
            out += x * pow(2, index)
        print(str(a_int) + " + " + str(b_int) + " = " + str(out))
        print('')

print('training is over...')
while True:
    layer_1_values = list()
    layer_1_values.append(np.zeros(hidden_dim))
    a_int = int(input('please input the first number: '))
    b_int = int(input('please input the first number: '))
    a = int2binary[a_int]
    b = int2binary[b_int]
    c_true_int = a_int + b_int
    c = int2binary[c_true_int]
    d_bin = np.zeros_like(c)

    for position in range(binary_dim):
        # generate input and output
        X = np.array([[a[binary_dim - position - 1], b[binary_dim - position - 1]]])
        y = np.array([[c[binary_dim - position - 1]]]).T

        layer_1 = sigmoid(np.dot(X, synapse_0) + np.dot(layer_1_values[-1], synapse_h))

        layer_2 = sigmoid(np.dot(layer_1, synapse_1))

        d_bin[binary_dim - position - 1] = np.round(layer_2[0][0])
        layer_1_values.append(copy.deepcopy(layer_1))

    out = 0

    for index, x in enumerate(reversed(d_bin)):
        out += x * pow(2, index)
    print(str(a_int) + " + " + str(b_int) + " = " )
    print('predict: ' + str(out))
    print('true: ' + str(c_true_int))