This is a first test of a model trained on my own artwork, so I can generate images that look like I actually drew them myself. I expect improvements in future versions.