Given groups=1 weight of size [64 64 3 3] expected input[4 67 256 256] to have 64 channels but 67

我知道是网络层发生的问题，查看了网络层，但是实在不懂到底是哪里多了3个channels,急！

def init(self, ngf2=64, num_res_blocks=16, num_upsample=2):
super(Generator, self).init()

    self.conv0 = nn.Conv2d(in_channels=3,out_channels=ngf2, kernel_size=3, padding=1)
    # First layer
    self.conv1 = nn.Conv2d(in_channels=ngf2,out_channels=ngf2, kernel_size=3, padding=1, stride=1)
    # Residual blocks
    self.res_blocks = nn.Sequential(*[ResidualInResidualDenseBlock(ngf2) for _ in range(num_res_blocks)])
    # Second conv layer post residual blocks
    self.conv2 = nn.Conv2d(in_channels=ngf2, out_channels = ngf2, kernel_size=3, stride=1, padding=1)
    # Upsampling layers
    upsample_layers = []
    for _ in range(num_upsample):
        upsample_layers += [
            nn.Conv2d(ngf2, ngf2 * 4, kernel_size=1, stride=2, padding=1),
            nn.LeakyReLU(),
            nn.PixelShuffle(upscale_factor=2),
        ]
    self.upsampling = nn.Sequential(*upsample_layers)
    # Final output block
    self.conv3 = nn.Sequential(
        nn.Conv2d(in_channels=ngf2, out_channels = ngf2, kernel_size=5, stride=1, padding=1),
        nn.LeakyReLU(),
        nn.Conv2d(in_channels=ngf2, out_channels =64, kernel_size=5, padding=1, stride=1),
    )
    #self.conv4 = nn.Conv2d(out_channels=64, kernel_size=3, padding=1, in_channels=6)

def forward(self, x):
    out0 = self.conv0(x)
    out1 = self.conv1(out0)
    out = self.res_blocks(out1)
    out2 = self.conv2(out)
    out = torch.add(out1, out2)
    out = self.upsampling(out)
    out = self.conv3(out)
   # out = self.conv4(out)
    return out