I'm trying to generate images that would represent accessibility options for my clients who are disabled. The best case scenario would be the ability to upload my own image and then generate a modification that would show how the space could function better. For example, changing a single-level kitchen island to a bi-level island that would include work and eating spaces for people who use a wheelchair as well as family members who don't. Firefly doesn't seem to understand the words used with accessibility such as wheelchair, bi-level, ramp and so on. It would be great if the AI could learn such terms.