Initialise the weights of the conv layers with the weights of a pretrained autoencoder

vision
1

Hi,
I have trained an autoencoder to get the lower dimensional representation of a image dataset.
Now I have saved the model. Then I load it and I want to remove the weights of the decoders. With the weights of the encoder I want to intiatlise the conv layers of a CNN classifier. This CNN classifier has same structure with the encoder of the autoencoder. Can anyone help me?

Basically I want to implement this project(implemented in keras)in pytorch:

2 
class Autoencoder(nn.Module):
    def __init__(self):
        super(Autoencoder, self).__init__()
          # encoder layers
        self.enc1 = nn.Conv2d(3, 96, kernel_size=11, stride=4)
        self.bn1 = nn.BatchNorm2d(96)
        self.enc2 = nn.Conv2d(96, 256, kernel_size=5, stride=1, padding=2)
        self.bn2 = nn.BatchNorm2d(256)
        self.enc3 = nn.Conv2d(256, 384, kernel_size=3, stride=1, padding=1)
        self.pool = nn.MaxPool2d(3, 2)

        # decoder layers
        self.dec1 = nn.ConvTranspose2d(384, 256, kernel_size=3, stride=2)
        self.dec2 = nn.ConvTranspose2d(256, 96, kernel_size=3, stride=2)
        self.dec3 = nn.ConvTranspose2d(96, 3, kernel_size=11, stride=4)

        def forward(self, x):

            # encoder
            x = F.relu(self.bn1(self.enc1(x)))
            x = self.pool(x)
            x = F.relu(self.bn2(self.enc2(x)))
            x = self.pool(x)
            x = F.relu(self.enc3(x))  # the latent space representation

            # decoder
            x = F.relu(self.dec1(x))
            x = F.relu(self.dec2(x))
            x = F.relu(self.dec3(x))
            # x = F.relu(self.dec4(x))
            # x = F.sigmoid(self.out(x))

            return x
class CNN(nn.Module):
    def __init__(self):
        super(CNN, self).__init__()

        # Conv layers
        self.enc1 = nn.Conv2d(3, 96, kernel_size=11, stride=4)
        self.bn1 = nn.BatchNorm2d(96)
        self.enc2 = nn.Conv2d(96, 256, kernel_size=5, stride=1, padding=2)
        self.bn2 = nn.BatchNorm2d(256)
        self.enc3 = nn.Conv2d(256, 384, kernel_size=3, stride=1, padding=1)
        self.pool = nn.MaxPool2d(3, 2)
        # FC layers
        
x=Autoencoder()
y=CNN()
print(x.enc1.weight)
x.enc1.weight=y.enc1.weight

print(x.enc1.weight)