How to perform finetuning in Pytorch?

[apaszke (Adam Paszke)]

class MyModel(nn.Module):
    def __init__(self, pretrained_model):
        self.pretrained_model = pretrained_model
        self.last_layer = ... # create layer

    def forward(self, x):
        return self.last_layer(self.pretrained_model(x))

pretrained_model = torchvision.models.resnet18(pretrained=True)
model = MyModel(pretrained_model)