简单神经网络实现 03

实现简单反向传播。

import numpy as npdef sigmoid(x):    """    Calculate sigmoid    """    return 1 / (1 + np.exp(-x))x = np.array([0.5, 0.1, -0.2])target = 0.6learnrate = 0.5weights_input_hidden = np.array([[0.5, -0.6],                                 [0.1, -0.2],                                 [0.1, 0.7]])weights_hidden_output = np.array([0.1, -0.3])## Forward passhidden_layer_input = np.dot(x, weights_input_hidden)hidden_layer_output = sigmoid(hidden_layer_input)output_layer_in = np.dot(hidden_layer_output, weights_hidden_output)output = sigmoid(output_layer_in)## Backwards pass## TODO: Calculate errorerror = target - outputprint('error',error)# TODO: Calculate error gradient for output layerdel_err_output = error * output * (1 - output)print('del_err_output',del_err_output)# TODO: Calculate error gradient for hidden layerdel_err_hidden = np.dot(del_err_output, weights_hidden_output) * hidden_layer_output * (1 - hidden_layer_output)print('del_err_hidden',del_err_hidden)# TODO: Calculate change in weights for hidden layer to output layerdelta_w_h_o = learnrate * del_err_output * hidden_layer_outputprint('delta_w_h_o',delta_w_h_o)# TODO: Calculate change in weights for input layer to hidden layerdelta_w_i_o = learnrate * del_err_hidden * x[:, None]print('delta_w_i_o',delta_w_i_o)print('Change in weights for hidden layer to output layer:')print(delta_w_h_o)print('Change in weights for input layer to hidden layer:')print(delta_w_i_o)