pytorch创建模型方式

1.继承自nn.Module的方式

from torch import nn
import  torch.nn.functional as F
'''继承自nn.Module'''

class LModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.L1 = nn.Linear(10,10)
        self.L2 = nn.Linear(10,64)
        self.L3 = nn.Linear(64,10)
        self.L4 = nn.Linear(10,5)
        self.L5 = nn.Linear(5,1)
        '''也可以不定义这个直接使用F低阶API接口'''
        self.relu = nn.ReLU()
        self.sig = nn.Sigmoid()

    def forward(self, input):
        x = self.L1(input)
        x = self.relu(x)
        '''x = F.relu(x)'''
        x = self.L2(x)
        x = self.sig(x)
        '''x = F.sigmod(x)'''
        x = self.L3(x)
        x = self.relu(x)
        '''x = F.relu(x)'''
        x = self.L4(x)
        x = self.relu(x)
        '''x = F.relu(x)'''
        x = self.L5(x)
        x = self.sig(x)
        '''x = F.sigmod(x)'''

        return x






'''使用模型办法'''
x = np.random.randn(10)
print(x)
x = torch.from_numpy(x).type(torch.float32)
print(x)
model = LModel()
print(model)
x = model(x)
print(x)

2.直接使用 nn.Sequential

    model = nn.Sequential(nn.Linear(10,64),
                          nn.ReLU(),
                          nn.Linear(64,25),
                          nn.Tanh(),
                          nn.Linear(25,10),
                          nn.ELU(),
                          nn.Linear(10,5),
                          nn.Sigmoid(),
                          nn.Softmax(dim=0))
    x = np.random.randn(10)
    print(x)
    x = torch.from_numpy(x).type(torch.float32)
    print(x)
    x = model(x)
    print(x)
    print(x.sum())

           定义模型需要用到的函数，基本都在nn这个模块里面有定义。