(self, x, z)
| 82 | self.output_nc = out_channels*2 if self.skip_connect else out_channels |
| 83 | |
| 84 | def forward(self, x, z): |
| 85 | out = x.pop() if self.skip_connect else x |
| 86 | for i in range(self.encoder_layers-self.decoder_layers, self.encoder_layers)[::-1]: |
| 87 | model = getattr(self, 'decoder' + str(i)) |
| 88 | out = model(out, z) |
| 89 | out = torch.cat([out, x.pop()], 1) if self.skip_connect else out |
| 90 | return out |
| 91 | |
| 92 | class ADAINEncoderBlock(nn.Module): |
| 93 | def __init__(self, input_nc, output_nc, feature_nc, nonlinearity=nn.LeakyReLU(), use_spect=False): |
nothing calls this directly
no outgoing calls
no test coverage detected