the model of Azula from the avatar cartoon. The recommended value is 0.55 - 0.7 . LoRA is not too flexible, just repeats the style of the original cartoon, but maybe you will get something better.
Description
FAQ
Comments (4)
Each and every Azula model I try is rigid af. How does this compare?
i guess this lora is trained with the model u are using to generate the images, cuz it's like almost impossible to get cool result with other checkpoints
yes, all my LoRAs are made on divineelegancemix_V6/5, because nai does not give exactly the result I want. I just share what I do for myself
matthewlgavin425Mar 16, 2024
Try "EDmfcG". Works pretty well, honestly. Use the Seed from one of the images posted here, and it works great.