DL之DNN：自定义2层神经网络TwoLayerNet模型(封装为层级结构)利用MNIST数据集进行训练、GC对比

本文链接：https://blog.csdn.net/qq_41185868/article/details/88951733

导读
神经网络算法封装为层级结构的作用。在神经网络算法中，通过将神经网络的组成元素实现为层，可以高效地计算梯度（反向传播法）。通过比较数值微分和误差反向传播法的结果，可以确认误差反向传播法的实现是否正确（梯度确认）。

输出结果

设计思路

核心代码

 
 
 
class TwoLayerNet:
 
    def __init__(self, input_size, hidden_size, output_size, weight_init_std = 0.01):
 
        self.params = {}
        self.params['W1'] = weight_init_std * np.random.randn(input_size, hidden_size)
        self.params['b1'] = np.zeros(hidden_size)
        self.params['W2'] = weight_init_std * np.random.randn(hidden_size, output_size) 
        self.params['b2'] = np.zeros(output_size)
 
        self.layers = OrderedDict()
        self.layers['Affine1'] = Affine(self.params['W1'], self.params['b1'])
        self.layers['Relu1'] = Relu()
        self.layers['Affine2'] = Affine(self.params['W2'], self.params['b2'])
 
        self.lastLayer = SoftmaxWithLoss()
        
    def predict(self, x):
        for layer in self.layers.values():
            x = layer.forward(x)
        
        return x
        
    # x:输入数据, t:监督数据
    def loss(self, x, t):
        y = self.predict(x)
        return self.lastLayer.forward(y, t)
    
    def accuracy(self, x, t):
        y = self.predict(x)
        y = np.argmax(y, axis=1)
        if t.ndim != 1 : t = np.argmax(t, axis=1)
        
        accuracy = np.sum(y == t) / float(x.shape[0])
        return accuracy
        
        
    def gradient(self, x, t):
        self.loss(x, t)
 
 
        dout = 1
        dout = self.lastLayer.backward(dout)
        
        layers = list(self.layers.values())
        layers.reverse()
        for layer in layers:
            dout = layer.backward(dout)
 
        grads = {}
        grads['W1'], grads['b1'] = self.layers['Affine1'].dW, self.layers['Affine1'].db
        grads['W2'], grads['b2'] = self.layers['Affine2'].dW, self.layers['Affine2'].db
 
        return grads



network_batch = TwoLayerNet(input_size=784, hidden_size=50, output_size=10)

grad_numerical = network_batch.numerical_gradient_api(x_batch, t_batch)  
grad_backprop = network_batch.gradient(x_batch, t_batch)

代码实现过程错误记录

出现错误，待解决!!!

Traceback (most recent call last):
File "F:\File_Python\Python_daydayup\190316.py", line 281, in <module>
grad = network.gradient(x_batch, t_batch)
File "F:\File_Python\Python_daydayup\190316.py", line 222, in gradient
self.loss(x, t)
File "F:\File_Python\Python_daydayup\190316.py", line 193, in loss
return self.lastLayer.forward(y, t) # ☆★☆★☆★☆★☆★☆★☆★☆★☆★☆★☆★☆★☆★☆ ------部分更改
File "F:\File_Python\Python_daydayup\190316.py", line 132, in forward
self.loss = cross_entropy_error(self.y, self.t)
File "F:\File_Python\Python_daydayup\190316.py", line 39, in cross_entropy_error
return -np.sum(np.log(y[np.arange(batch_size), t.astype('int64')] + 1e-7)) / batch_size #t.astype('int64')
IndexError: shape mismatch: indexing arrays could not be broadcast together with shapes (100,) (100,10)