Deep learning is all the rage right now, and it is powering some amazing tech, from self-driving cars to board game playing computers. However, most deep learning models, and actually most artificial intelligence models in general, are far from portable. They have a ton of dependencies, and, even if we use Docker to package our models, machine learning images could take up 2GB or more of disk space.
In this project, I will explore shrink a Deep Learning model for image to image translation and deploy it to a Pocket Chip that is integrated with a Movidius NCS. This model will allow me to draw doodles of cats and immediately translate those doodles to realistic pictures of cats. All from the convenience of a pocket sized computer.