
Lora trained on 232 sfw images, that I personally chose to. It just so happens to be my favorite character, and it was hard to choose art for training among the thousands I have, haha..
After a lot of trying, I finally managed to put together lora that gives more or less good results. But I'm still not sure I've made it perfect yet. I would be happy to get any help!
Recommended positive:
1girl, solo, masterpiece, furry, rabbit ears, white fur, open mouth, smile, red eyes, white hair, short red sleeves, black dress, raboot
Negative:
(worst quality, low quality:1.4), white background, EasyNegative, ng_deepnegative_v1_75t
Configs:
CFG Scale: 7 ~ 12
Raboot Lora Weight: 0.8 ~ 1
Works good on the Furry girl lora.
VAE: anime.
Trigger world: raboot
Trained on Counterfeit v3.
描述:
训练词语: raboot
名称: [character] raboot_v1e20.safetensors
大小 (KB): 36995
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success