| 11 | from torch.utils.data import DataLoader |
| 12 | |
| 13 | class Discriminator(torch.nn.Module): |
| 14 | def __init__(self,features = 64): |
| 15 | super(Discriminator, self).__init__() |
| 16 | self.in_features = features |
| 17 | self.conv = torch.nn.Sequential( |
| 18 | torch.nn.Conv2d(in_channels = 3,out_channels = features,kernel_size = (4,4), |
| 19 | stride=(2,2),padding = (1,1),bias = False), |
| 20 | torch.nn.BatchNorm2d(num_features= features), |
| 21 | torch.nn.LeakyReLU(negative_slope = 0.2,inplace = True),#inplace，其作用是：该nn.Relu() 函数计算得到的输出是否更新传入的输出。 |
| 22 | |
| 23 | torch.nn.Conv2d(in_channels=features , out_channels=features * 2, kernel_size=(4, 4), |
| 24 | stride=(2, 2), padding=(1,1), bias=False), |
| 25 | torch.nn.BatchNorm2d(num_features=features * 2), |
| 26 | torch.nn.LeakyReLU(negative_slope = 0.2,inplace = True), |
| 27 | |
| 28 | torch.nn.Conv2d(in_channels=features * 2, out_channels=features * 4, kernel_size=(4, 4), |
| 29 | stride=(2, 2), padding=(1,1), bias=False), |
| 30 | torch.nn.BatchNorm2d(num_features=features * 4), |
| 31 | torch.nn.LeakyReLU(negative_slope = 0.2,inplace = True), |
| 32 | |
| 33 | torch.nn.Conv2d(in_channels=features * 4, out_channels=features * 8, kernel_size=(4, 4), |
| 34 | stride=(2, 2), padding=(1,1), bias=False), |
| 35 | torch.nn.BatchNorm2d(num_features=features * 8), |
| 36 | torch.nn.LeakyReLU(negative_slope = 0.2,inplace = True), |
| 37 | |
| 38 | torch.nn.Conv2d(in_channels=features * 8, out_channels=1, kernel_size=(4, 4), |
| 39 | stride=(1,1), padding=(0,0), bias=False), |
| 40 | torch.nn.Sigmoid() |
| 41 | ) |
| 42 | def forward(self,input): |
| 43 | x = self.conv(input) |
| 44 | return x |
| 45 | |
| 46 | if __name__ == '__main__': |
| 47 | dis = Discriminator() |