深度学习入门——乘法层（Mullayer）

• 上游的值乘以输入数据的“翻转值”
  
  

# coding:utf-8
# 将神经网络的“层”实现为一个类。这里的层是神经网络中功能的单位，例如负责sigmoid函数的sigmoid、负责矩阵乘法的Affine等，都以层
# 的单位及逆行实现。这里也以层的单位来实现乘法节点和加法节点

class Mullayer:
    """乘法层"""
    def __init__(self):
        self.x = None
        self.y = None

    def forward(self, x, y):
        """前向"""
        self.x = x
        self.y = y
        out = x * y

        return out

    def backward(self, dout):
        """反向传播，传播参数为导数dout"""
        dx = dout * self.y # 翻转x和y,导数dout乘以正向传播的翻转值即可
        dy = dout * self.x

        return dx, dy


# -----------------------------------实现买苹果的计算图，正向传播计算价格-------------------------------------------------
apple = 100 # 单价
apple_num = 2
tax = 1.1 # 税

# layer
mul_apple_layer = Mullayer()
mul_tax_layer = Mullayer()

# forward
apple_price = mul_apple_layer.forward(apple, apple_num)
price = mul_tax_layer.forward(apple_price, tax)

print(price) # 220

# --------------------------------------实现苹果计算图，反向传播，传播导数---------------------------------------
"""与前面的正向传播完全反过来，并且输入是apple_price的导数"""
dprice = 1
dapple_price, dtax = mul_tax_layer.backward(dprice)
dapple, dapple_num = mul_apple_layer.backward(dapple_price)
print(dapple, dapple_num, dtax) # 2.2, 110, 200