anima toony merge

The anima model is very promising. When a proper e621 fine tune is available, I plan to do more with it. Here is a LoRA merge fine tune with as much compute as I can throw at it at the moment. Let's call this a beta. This merge has a good amount of e621 tag knowledge and concepts, but e621 artists are barely present. including the tag 'deviantart' in inference is highly recommended.