this model was trained on 14000 captioned images, it can pretty much do anything.

it's the same as the 1.5 version but gives better quality etc..

use this model with the lora https://civarchive.com/models/117259/anime-art-diffusion-xl

I included a custom node for prompting in comfyUI you can get the workflow that works best for me (at the moment) from there too.

https://github.com/nicolai256/comfyUI_Nodes_nicolai256

I made the custom node because it makes prompting for this lora super easy and I'm planning to add a couple nodes to create cards that fit with the prompts.

to get the Yu-Gi-Oh! style it's best to prompt these keywords into your prompt

glowing, yugioh style, yugioh monster, duel monster

you can also use these to enhance the style even more

level: 9 (change to a number u want)
Attribute: Aqua/Beast/Beast-Warrior/Cyberse/Dinosaur/Divine-Beast/Dragon/Fairy/Fiend/Fish/Insect/Machine/Plant/Psychic/
           Pyro/Reptile/Rock/Sea Serpent/Spellcaster/Thunder/Warrior/Winged Beast/Wyrm/Zombie
Type: Dark/Divine/Earth/Fire/Light/Water/Wind 
Attack: 2600 (change to a number u want)
Defense: 1900 (change to a number u want)
outline, digimon, pokemon, sparkle