Ai generations

About: All of the generations on this page was creating using The Big Sleep AI. Below you can find some of the better visualizations the program has generated. You can find the creator on his Twitter page, where all credit of the code is due. Most of the examples below have been turned into gifs of various revisions that the text-to-image outputs had went through. Some of the inputs are suggestions of others, which is indicated if there are initials next to the phrases.

How: Running this program with a Colab Pro subscription grants you use of a GPU thus greatly reducing the run time. However, running this program using the free version works fine as well. There are original instructions on reddit about running The Big Sleep using Google Colab. On the Python Repo you can find a simplified version of this notebook. 

Update (Dec. 2022): In recent times, AI text to image technology has grown exponentially in both quality and quantity! When comparing what I was working with around 2021, to the potential of DALL-E 2, it brings these outputs to shame. Nevertheless, am still very happy (and proud) of the abstract awesome-ness that was generated with this program. As a side note, it sort of makes me feel like an early adopter to this technology, before it blew up in popularity!