How to modify the first and last layer of pretrained network

vision
4 
class VGG(nn.Module):

    def __init__(self, features, num_classes=1000, init_weights=True):
        super(VGG, self).__init__()
        self.features = features
        self.classifier = nn.Sequential(
            nn.Linear(512 * 7 * 7, 4096),
            nn.ReLU(True),
            nn.Dropout(),
            nn.Linear(4096, 4096),
            nn.ReLU(True),
            nn.Dropout(),
            nn.Linear(4096, num_classes),
        )
        if init_weights:
            self._initialize_weights()

    def forward(self, x):
        x = self.features(x)
        x = x.view(x.size(0), -1)
        x = self.classifier(x)
        return x
    def _initialize_weights(self):
        for m in self.modules():
            if isinstance(m, nn.Conv2d):
                nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
                if m.bias is not None:
                    nn.init.constant_(m.bias, 0)
            elif isinstance(m, nn.BatchNorm2d):
                nn.init.constant_(m.weight, 1)
                nn.init.constant_(m.bias, 0)
            elif isinstance(m, nn.Linear):
                nn.init.normal_(m.weight, 0, 0.01)
                nn.init.constant_(m.bias, 0)

cfg = {
    'A': [64, 'M', 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M'],
    'B': [64, 64, 'M', 128, 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M'],
    'D': [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 'M', 512, 512, 512, 'M', 512, 512, 512, 'M'],
    'E': [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 256, 'M', 512, 512, 512, 512, 'M', 512, 512, 512, 512, 'M'],
}
def vgg19(pretrained=False, **kwargs):
    """VGG 19-layer model (configuration "E")
    Args:
        pretrained (bool): If True, returns a model pre-trained on ImageNet
    """
    if pretrained:
        kwargs['init_weights'] = False
    model = VGG(make_layers(cfg['E']), **kwargs)
    if pretrained:
        model.load_state_dict(model_zoo.load_url(model_urls['vgg19']))
    return model

def make_layers(cfg, batch_norm=False):
    layers = []
    in_channels = 6 #Define the number of channels
    for v in cfg:
        if v == 'M':
            layers += [nn.MaxPool2d(kernel_size=2, stride=2)]
        else:
            conv2d = nn.Conv2d(in_channels, v, kernel_size=3, padding=1)
            if batch_norm:
                layers += [conv2d, nn.BatchNorm2d(v), nn.ReLU(inplace=True)]
            else:
                layers += [conv2d, nn.ReLU(inplace=True)]
            in_channels = v
    return nn.Sequential(*layers)
vgg19(pretrained=False)

The Above is the VGG19 with 6 input channels.
For the weight you need to load it manually for every layer through iterating.

One solution would be us vgg19(pretrained=True) with input channels=3 then you will get net weights as ‘vgg19-dcbb9e9d.pth’ file then see my other reply

use flat_weights,shapes=flattenNetwork(vgg19_3channels)
x=unFlattenNetwork(flat_weights,shapes) --this will give you the numpy array for each layer.

then you modify the first x[0] which is 3 channel in the above to the 6 channels just by adding your weights or dummy weights.
then iterate through all layers and bias and modify the vgg19_6channel model with modified x you got above.

 vgg19_6channel.conv1.weight.data = torch.from_numpy(x[0])
 vgg19_6channel.conv1.bias.data =   torch.from_numpy(x[1]) ..etc
``
1 Like