






LoRa for creating cubist painting (especially portraits)
The model is quite experimental and you will probably need to make some test before having what you want.
best prompts: solo, portrait, <color> skin color, facepaint, surreal, abstract
Recommand 1 to 0.7 weight for good picture. You can go below for more recognizable character, but going further than 1 generally don't produce good looking images.
neg-prompts and neg-TI can help but don't use too much of it (especially the ones that contradict elements of the model obviously)
Portrait are quite ok to make but full body or multiple characters can be really abstract and surreal. If this is the sort of result you want (like the last preview images), I suggest putting the model on full weight, and if you see some shapes and arrangements you like, try decrease the model's weight.
If you manage to have some great results or if you want to suggest some improvement for the model, you can send me feedbacks in the comment.
1. 轉载模型僅供學習與交流分享,其版權及最終解释權归原作者。
2. 模型原作者如需認領模型,請通過官方渠道联系海藝AI工作人員進行認證。我們致力於保護每一位創作者的權益。 點擊去認領
