Understanding neural style transfer

Neural style transfer is a technique where you blend a content image and a styled image by matching their feature distributions in order to generate a final image that's similar to the content image but that's artistically painted with the style of the styled image. Style transfer can be done in TensorFlow in two different ways:

• Using the pre-trained model in TensorFlow Hub. This is where you upload your image and style and the kit will generate your styled output. You can upload your image at https://colab.research.google.com/github/tensorflow/hub/blob/master/examples/colab/tf2_arbitrary_image_stylization.ipynb. Note that TensorFlow Hub is a source of many pre-trained networks.
• Develop your own model by training the neural network. To do this, follow these steps:
  1. Select the VGG19 network – it has five convolution (Conv2D...