Deep Learning with PyTorch

Creating the optimizer

In general, we pass in the parameters of a network like VGG to be trained. But, in this example, we are using VGG models as feature extractors and, hence, we cannot pass the VGG parameters. Here, we will only provide the parameters of the opt_img variable that we will optimize to make the image have the required content and style. The following code creates the optimizer that optimizes its values:

optimizer = optim.LBFGS([opt_img]);

Now we have all the components for training.