This is the second version of the model, but the first was never released to the public, so conventionally it is version 1.0





Warning!
This model is being raided by zoophiles who put 1 star because I put bad reviews on their zoophile Loras.
I give them bad ratings because the Civitai administration doesn't want to implement the functionality to completely exclude zoophile content from the feed.
This LoRa is trained on studio photos of cars from the 80s and 90s, as well as a little bit of the 70s.
You can nice concepts for inspiration. It can also be used with people and landscapes, but the results are not guaranteed.
It is better to add AutoStudio at the beginning of the prompt, this is the keyword for the LoRa.
To create unusual colours and lighting, it's recommended to use prompts like [colour1:colour2:step]. For example [red:green:15], which means that the first 15 steps will use red and after that green. This also applies to body types, car brands and other things.
In general, it works about the same on all models. For sure you can get variety of styles on different models as well.
Recommended parameters:
Resolution: 768 x 512
Samplers: DPM++ 2 and DPM SDE series. You may try others on different models.
25-100 steps.
Hires.fix is recommended, especially 4x-UltraSharp.
CFG Scale Fix: CFG Scale 14-20, mimic scale 7-10. For more information, read the extension description here.
You can see examples of prompts in the gallery.
This is the second version of the model, but the first was never released to the public, so conventionally it is version 1.0
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
