| 22 | |
| 23 | |
| 24 | class Upsample(nn.Module): |
| 25 | def __init__(self, in_channels, with_conv): |
| 26 | super().__init__() |
| 27 | self.with_conv = with_conv |
| 28 | if self.with_conv: |
| 29 | self.conv = torch.nn.Conv2d( |
| 30 | in_channels, in_channels, kernel_size=3, stride=1, padding=1 |
| 31 | ) |
| 32 | |
| 33 | def forward(self, x): |
| 34 | x = torch.nn.functional.interpolate(x, scale_factor=2.0, mode="nearest") |
| 35 | if self.with_conv: |
| 36 | x = self.conv(x) |
| 37 | return x |
| 38 | |
| 39 | |
| 40 | class Downsample(nn.Module): |