自己写模型

今天下午自己在哪里倒腾着，不小心写了个Module出来，当然没有写具体的运算。吓写的，但可以从这个不能运行的程序中，学会写模型应该怎么写

class mylayer(torch.nn.Module):
    def __init__(self, inn=3, out=4):
        super(mylayer, self).__init__()
        self.inn = inn
        self.out = out
        self.weight = torch.nn.Parameter(torch.Tensor(3, 4))
        self.bias = torch.nn.Parameter(torch.Tensor(4))
    def forward(self,x):
        return 2*x

class tt(torch.nn.Module):
    def __init__(self):
        super(tt, self).__init__()
        self.linear = torch.nn.Linear(3,4)
        self.mylayer = mylayer(inn=4,out=5)
    def forward(self,x):
        out = self.linear(x)
        out = self.mylayer(out)
        return out

aa = tt()
print(aa)

tt(
(linear): Linear(in_features=3, out_features=4, bias=True)
(mylayer): mylayer()
)

aa.state_dict(