LoRA of character trained with HassakuModelV1.3 with 11 images and 2.86% loss when training up to 2200 steps.
That includes "pixel art" and "style parody" version damping Prompts. Use it as a negative Prompts.
I recommend in 0.7 if you want to keep the character true to the description, the suggested keywords comes to set a version. Main Prompts: paola spaghetti; Negative Prompts: style parody
