About

This workflow uses LSNet to perform artist classification on an input image. This workflow is not really for generating images per se, as it will simply output artist tags that are the most well-associated with a given input image. It is also not intended for 3D styles and realistic images, nor is it really intended for "traditional" art, it is purely for anime-style images.

It can be useful for finding the most likely artist tag(s) used in a prompt for an image that lacks metadata, or even if the image has a prompt, it can still be useful for seeing which artist tag(s) could yield similar results if the metadata lacks them.

Input image used: https://aibooru.online/posts/111018

(quick note for people who downloaded this prior to 11:45PM EST on 10/19/2025: made a mistake with the artist tags subgraph that made it operate in the wrong order, should be fine now.)