由于网络通常非常深，我们一层一层写会耗费很大的精力。而且修改起来也比较麻烦，不方便维护。网络通常由一些相似的块组成，我们写出一个块，循环生成整个网络。

nn.Sequential

def__init__(self, num_layers=1, activation=torch.nn.ReLU()):super(Classifier3, self).__init__()
    self.num_layers= num_layers
    self.dense1= torch.nn.Sequential(
        torch.nn.utils.spectral_norm(torch.nn.Linear(512,256)),
        activation)
    self.dense= self._make_layers(num_layers,activation)
    self.dense2= torch.nn.Sequential(
        torch.nn.utils.spectral_norm(torch.nn.Linear(256,128)),
        activation)
    self.dense3= torch.nn.Sequential(
        torch.nn.utils.spectral_norm(torch.nn.Linear(128,64)),
        activation,
        torch.nn.Linear(64,2),
        torch.nn.LogSoftmax(dim=1))def_make_layers(self, num_layers=1, activation=torch.nn.ReLU()):
    dense=[]for iinrange(num_layers):
        dense.append(torch.nn.Sequential(
            torch.nn.utils.spectral_norm(torch.nn.Linear(256,256)),
            activation))return torch.nn.Sequential(*dense)

nn.ModuleList

def__init__(self, num_layers=1, activation=torch.nn.ReLU()):super(Classifier3, self).__init__()
    self.num_layers= num_layers
    self.dense1= torch.nn.Sequential(
        torch.nn.utils.spectral_norm(torch.nn.Linear(512,256)),
        activation)
    dense=[]for iinrange(num_layers):
        dense.append(torch.nn.Sequential(
            torch.nn.utils.spectral_norm(torch.nn.Linear(256,256)),
            activation))
    self.dense_block= torch.nn.ModuleList(dense)
    self.dense2= torch.nn.Sequential(
        torch.nn.utils.spectral_norm(torch.nn.Linear(256,128)),
        activation)
    self.dense3= torch.nn.Sequential(
        torch.nn.utils.spectral_norm(torch.nn.Linear(128,64)),
        activation,
        torch.nn.Linear(64,2),
        torch.nn.LogSoftmax(dim=1))