By the author:

First version of the fine-tuned Stable Diffusion model (based on v1.5) trained on screenshots from 80s music video Take on me (from A-HA band). Use the token takeON in your prompts to use the style.