


I've noticed that some of the models I have collected are being by their creators. One such LoRA was LowRA. So, I decided to try to make a similar LoRA myself.
Goals:
Dark images
Subtle lighting
Night images
Dramatic images
And so on...
I tried to minimize the network rank used in this LoRA so that the main focus of the image could be better prompted for, just make it more low-key. In my opinion, this LoRA achieves that. The results are more dark, dreamy, or subtly lit. All while retaining color.
Weights: -0.5 to 1 have worked the best for me. Negative weights will make the image more dreamy and light, whereas positive weights make the image darker. Within this range, you can think of the LoRA as a light slider. I recommend sticking with lower absolute weights if you are generating people. Scenes allow for more freedom of the weight, generally speaking.
All example images were generated using epiCRealism (because Civitai still won't let me enter that manually).
1. 재게시된 모델의 권리는 원 제작자에게 있습니다.
2. 모델 원작자가 모델을 인증받으려면 공식 채널을 통해 SeaArt.AI 직원에게 문의하세요. 저희는 모든 창작자의 권리를 보호하기 위해 노력합니다. 인증하러 이동
