Neural style transfer


I am not able to understand what these lines of code are for??Can someone explain this?

You are constructing an actual model by selecting some layers from the given vgg model. Then you invoke that model with two different inputs: the content image and the style image.