The Visio Divina Dataset: A.I. generated images using biblical text

Welcome to the Visio Divina Dataset (VVD). The VVD dataset is an output of the project Seeing the Words, conducted jointly at the Eep Talstra Centre for Bible and Computer (ETCBC) and the Computer Science department, Vrije Universiteit Amsterdam. Prof. Prof. Willem van Peursen led the project. Hidde Makimei was the student who wrote the code and generated the images. Shuai Wang was involved as a daily supervisor of the student and the data steward. The images were generated using different tools including - Dall E (in the folder ./images/dall_e_2/). - Midjourney (in the folder ./images/mid_journey) - Stable Diffusion (various versions). We used some popular open-source releases: SG161222 (SG in short), runwayml (RW), CompVis (CV), stabilityai, (SAI), prompthero (PH), nitrosocke (NS), and dreamlike-art (DA). There are in total 7,116 images. We generated 500 images using Dall E 2, and 616 images using Midjourney. For Stable Diffusion, we generated one thousand images each for RW, CV, SAI, PH, and SG. Additionally, we generated 500 images using SG and 500 images using DA. Please check the ETCBC website for the latest version and related resources.