This model is another attempt to train a style based on a single training epoch sample image, except it turned out much better than I expected.