Finally, my first Lora is complete. After many failures and retries, I made some major changes to the trigger words. Please see the details here for how to use them, or you can directly view my image.
However, too much data can make the image look cluttered; for a cleaner portrait, try V0.7.
Basic prompts for portrait：ff000,close-up, portrait, slime suit, glossy skin, viscous, tendrils, veins on face, organic armor, dripping, covered neck, tentacles on face, tentacles in mouth, tentacles in nose, live clothing, [color] substance, [color] bodysuit
1. covered jaw,
2. covered mouth, mouth mask,
3. live helmet,short hair,
4. live helmet, blindfold, no eyes, covered eyes, no hair
Basic prompts for full-body images：hf101, full body, glowing veins, tendrils, slime suit, glossy skin, viscosities, veins on face, organic armor, dripping, covered neck, tentacles on face, tentacles in nose, tentacles in mouth, live clothing, [color] substance, [color] bodysuit, glowing [color2] lines,
1. covered jaw,
Negative: covered_mouth,no_mouth,
2. covered mouth,mouth mask,
3. Living Helmet,Mask open,short hair,
4. blindfold,no eyes,covered eyes,no hair,
other
futanari,covered penis,covered testicles,
If necessary, I can streamline the dataset to create a cleaner full-body image.
Thank you for your interest in my clumsy work.
Description
Finally, my first LORA runs well under normal circumstances, but it's a bit large because I trained it with large parameters in pursuit of accuracy (which actually didn't work well), so the file size is also large.