This Lora is trained by some art paintings and stills of my favorite movie --- Blade Runner 2049 (See under for further information and how to use)
此lora运用了我最爱的电影 -银翼杀手2049 的美术绘画以及部分剧照训练而成。(使用方法与简介见下文)
The training base model is sd1.4, but it also has many unexpected effects on other base models. have a try!
If the image glitch, try to increase the number of sampling steps or use the img2img with tile upscale to optimize.
This Lora can be used alone as a scene generation, but the generated content may be relatively simple,
The best way to use this Lora is to use it with other subjects, or to mix it with other Lora: add this Lora to the subject you want to generate, for example, add this Lora when generating girls, cars, and indoor scenes ;
It can make the original picture have more intense contrast and colorful light and shadow effects, (cyberpunk's iconic contrast between warm and cold light), more atmospheric perspective effects, MORE light effects, and more storytelling screen light and shadow division …
训练底模是sd1.4, 但搭配其他底模也能有很多意想不到效果,试着尝试搭配不同的底模吧!
如果画面崩坏,试着调高生成步数,或者后续用图生图tile放大去优化回来
此lora可以单独作为场景生成使用,但是生成内容可能会比较单一,
这个lora最好的使用方式是搭配其他主体使用,或者是与其他lora混合:在你想生成的主体物上再额外加上这个lora,譬如说在生成女孩,汽车,室内场景时加上此lora;
它能够使原本的画面拥有 对比度更加强烈和多彩的光影效果,(赛博朋克风标志性的冷暖光对比),更多的大气透视效果,更丰富的光效,更具故事感的画面光影分割......
The difference before and after using this lora
使用此lora的前后区别