WARNING: This is just a very rough test to get an estimate of how much I need to change for flux training vs sdxl/pdxl! This lora is able to reproduce the character, but not the style of the anime, so I am mainly just posting it for science reasons, not as a viable asset to make good images. Feel free to experiment with it, you might be able to get some nice results in combination with a good anime style (by prompt or a separate lora), but do not expect too much from it!

[EXPERIMENTAL AND VERY LOW STYLE UNDERSTANDING]Tarte from World's Finest Assassin and Anime style FLUX Test

EXPERIMENTAL just trying to figure out settings for flux lora trainer, might make an actually useable version later if I get the buzz for it and time to re-caption and upgrade my dataset, but no guarantees

Files with SHA256: 655eb51423ca229d01a6eb151df7212f32484ce6841af7cecf66ca3f2098a980