Pytorch 多层感知机--TypeError

1. 错误提示

Traceback (most recent call last):
  File "D:/Python_practice/CIRFAR-10/daily_test4.py", line 33, in <module>
    perception = Perception(3,4,1)
  File "D:/Python_practice/CIRFAR-10/daily_test4.py", line 27, in __init__
    self.layer1 = Linear(in_features,hidden_features)
  File "D:/Python_practice/CIRFAR-10/daily_test4.py", line 10, in __init__
    nn.Module().__init__(self)
TypeError: __init__() takes 1 positional argument but 2 were given

2.我的代码

import torch as t
from torch import nn
from torch.autograd import Variable as V

'''单个全连接层'''
class Linear(nn.Module): # 继承nn.Module
    #在构造函数中__init__（）中必须自己定义可学习的参数，并封住成Parameter
    def __init__(self,in_features,out_features):
        # super(Linear,self).__init__() #与nn.Module().__init__(self)
        nn.Module().__init__(self)
        self.w = nn.Parameter(t.randn(in_features,out_features))
        self.b = nn.Parameter(t.randn(out_features))
    def forward(self,x): #单词 错了？？？ forward
        x = x.mm(self.w)
        return x + self.b.expand_as(x)
'''感知机'''
class Perception(nn.Module):
    def __init__(self,in_features,hidden_features,out_features):
        nn.Module.__init__(self)
        # super(Perception,self).__init__()
        self.layer1 = Linear(in_features,hidden_features)
        self.layer2 = Linear(hidden_features,out_features)
    def forward(self,x):
        x = self.layer1(x)
        x = t.sigmoid(x)
        return self.layer2(x)
perception = Perception(3,4,1)
for name,param in perception.named_parameters():
    print(name,param.size())

3.错误原因
感觉是super方法调用nn.Module方法会包括Linear类中的自定义一些初始参数传递，而用nn.Module().__init__(self)方法调用仅仅是调用了该模块的自带的初始参数。所以会出现传参位数不匹配错误。

以上只是个人看法，如果有不错欢迎指正，谢谢各位大佬！

4. 修改方法

将Linear类中__init__()方法中的调用nn.Module模块中的构造函数形式换为super(Linear,self).__init__() 形式即可。

建议可以用super(classname,self).__init)__()方法来调用nn.Module()的构造函数，避免出错。