This model is used to generate diverse random Anime images with a more realistic and detailed look but still keeping the anime style. This is sort of a branch of my other model Rule34, it was made by merging Rule34_v10 x 0.65 + 0.35 x chaos_v12 by Kav.

The purpose of this model is the same as my Rule34 model, to create the sensation that you are browsing images from Rule34 and other booru image boards.

In my merges diversity and creativity takes priority even if it makes a lot of images with weird hands, feet etc.

To get advantage of the creativity I use DPM++2M with Steps 15 and CFG 2.

The model produces some realistic, non-anime type illustrations but that's okay with me, just adds more flair to it I guess.