If you can use ComfyUI with my LoRAs i can now highly recommend using this workflow

I'm still testing it myself but using only the Layers 17-22 works great with my Body LoRAs as it keeps the original ZIT style but gives you the correct body.

Disclaimer:

I'm not satisfied with the training results, especially with the fact that mixing multiple LoRAs in ZIT doesn't really work. I hope the base version will give better results.

However, I'm bored and like to share my results.

The LoRA(s) did not learn any faces.

Trained on SFW and ???? images unless stated otherwise.

Flux1 Versions

Body List

https://mega.nz/file/Zbk1EbBA#epyDaCTsyDrIZgcdJ319SujW9WpEQ31cIVYfxFLYsGM

Work in Progress...

Version A

Version B

Version D

Trained ???? only

Version E

Version F

Version H

Version I

Training focused on Booty

Version K

Trained ???? only

Version L