My implementation of Fast Style Transfer described in Perceptual Losses for Real-Time Style Transfer and Super-Resolution.
The model uses VGG19 to extract the style and content layers. The network is trained using the COCO 2014 Dataset. Here is a gif of the model's evaluation on an image of a puppy as it gets trained over time.