I am satisfied with the results when I…
I am satisfied with the results when I create images based on reference drawings. There can be up to 5 of them with some models, which is sufficient. As for the creation of images, some are very nice. The animation is usually weaker, or rather the original objects are simplified and distorted, especially faces. What works with very little success is editing images, where the AI often does not understand the description. Moreover, editing is very expensive in terms of credits, so it is generally not worth editing a generated image. I have been using Deevid for several months and it looks like I will be for some time.