Question

运行机器学习代码时出现错误。

我刚刚开始探索神经网络和机器学习，但我不知道为什么会发生这种情况或其含义。

for iteration in range(20000):

    input_layer = training_inputs
    outputs = sigmoid(np.dot(input_layer, synaptic_weights))

    error = training_inputs - outputs
    adjustment = error * sigmoid_derivative(outputs)

    synaptic_weights += np.dot(input_layer.T, adjustment)#error occurs here

*编辑：这是完整的代码

import numpy as np

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

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

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

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

np.random.seed(1)

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

print ('random starting syanptic weights: ')
print (synaptic_weights)

for iteration in range(1):

    input_layer = training_inputs
    outputs = sigmoid(np.dot(input_layer, synaptic_weights))

    error = training_inputs - outputs
    adjustment = error * sigmoid_derivative(outputs)

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

print(' synaptic weights after training: ')
print (synaptic_weights)
print ('outputs after training: ')
print (outputs)

Answer 1

替换

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

使用

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

错误：形状为（3,1）的不可广播输出操作数与广播形状（3,3）不匹配

1 个答案: