(self, in_channels, out_channels, *args, **kwargs)
| 836 | |
| 837 | class SimpleDecoder(nn.Module): |
| 838 | def __init__(self, in_channels, out_channels, *args, **kwargs): |
| 839 | super().__init__() |
| 840 | self.model = nn.ModuleList([nn.Conv2d(in_channels, in_channels, 1), |
| 841 | ResnetBlock(in_channels=in_channels, |
| 842 | out_channels=2 * in_channels, |
| 843 | temb_channels=0, dropout=0.0), |
| 844 | ResnetBlock(in_channels=2 * in_channels, |
| 845 | out_channels=4 * in_channels, |
| 846 | temb_channels=0, dropout=0.0), |
| 847 | ResnetBlock(in_channels=4 * in_channels, |
| 848 | out_channels=2 * in_channels, |
| 849 | temb_channels=0, dropout=0.0), |
| 850 | nn.Conv2d(2*in_channels, in_channels, 1), |
| 851 | Upsample(in_channels, with_conv=True)]) |
| 852 | # end |
| 853 | self.norm_out = Normalize(in_channels) |
| 854 | self.conv_out = torch.nn.Conv2d(in_channels, |
| 855 | out_channels, |
| 856 | kernel_size=3, |
| 857 | stride=1, |
| 858 | padding=1) |
| 859 | |
| 860 | def forward(self, x): |
| 861 | for i, layer in enumerate(self.model): |
nothing calls this directly
no test coverage detected