Anime XL v2, represents a significant increase in quality, continuity, prompt adherence, hands, faces, eyes, and overall aesthetic of the original I've completely re trained and merged this model. Original Fine tuned and based on Animagine and Kohaku Delta - two MASSIVE and I mean MASSIVE datasets. This gives the model incredible versatility and range. It has a unique drawing style akin to that of coloured manga or graphic novels, This one however is targeted at making characters and people. It can cover fantasy to realism pretty well, and if you try to FORCE realistic you'll get a 3d rendered effect. Fun things to try - Photorealistic - Hyper Realistic - 3D rendered - Ghibli Style - Kohaku Style - Animagine Style - potato mode <-- heh fun when the text encoder understands it otherwise you just get a random series of potatoes... also kinda fun Finally - Quality tags. Not "necessarily" needed. However do make a major difference. I've explicitly designed and fine tuned the model not to require quality tags as a mandatory thing to get a decent image. However , I have also allowed the use of quality tags. Meaning I've also tuned in Pony's quality tags score_up9 etc... So you can use a combination of Pony Quality up tags, or Down, as well as weight values and standard tags. Go nuts. FINALLY, if anyone can spot the easter egg I hid in the model. Bonus points to you! -_^ Take a guess in the comments and I'll let whoever gets it right know. Winner gets a tip from me. I'm broke so dont expect a lot hahaha Wish you all the best, Be your best h4 (b'.')b