I’ve recently been on a AI image generation role, primarily using MidJourney to create landscapes, but there is a cost to that service (It’s a processing expensive task, so I’m OK with that).
But for the job of creating character portraits, I’ve taken to using Stable Diffusion which is an open sourced implementation which can be run locally. It’s pretty heavy on the hardware requirements, needing a GPU with 6+gb of ram, but if you can meet that the results are pretty good.
The following are some character portraits I’ve used it to create:
