A general model with a slant towards classical paintings. It has some anime capacity, but it isn't a core competency.Includes lineage from MangledMerge, Cheyenne, FabledIllusion and c0rruptf33d, and technically manmaruMix. I used Hako Mikan's Elemental supermerger and tested large numbers of possible ways of combining the three primary models, using python scripts to attempt to pseudo train a good setup. I also pseudo trained in some c0rruptf33d, when I noticed the model was more versatile than it appears. Once I was somewhat satisfied, I decided to use all my buzz to train enhancement Loras to "bring together" the merge. The idea being that, the merge is close to correct on a lot of concepts, so with a little training, it should automatically select the proper model that understands the given concept. This, seems to have worked? At least if the concept is in the training data and one of the models, it seems to have picked up on that? At the least, the model seems competent for what it is, and I don't see an obvious way to improve it further, besides throwing more resources at it, at the moment.

Arc Naked Singularity

Files with SHA256: 5ff5ce73f4f14ea2676f73e328bf5af5a58dc2c4b8dd1054f2e51da98ed967e8