Recommended weight: 0.3 - 0.7
Recommended models: VXP 2.2 Hyper, Sunflower.
This is my first LoRA attempt. It's a LoRA designed to add details and a japanese urban-looking style to your images. It was trained on the base SDXL 1.0 model.
The lower versions such as v1.1, 1.2, 1.5, are less modern than the v1.10 and v1.11 versions, but v1.10 and v1.11 are the best in my opinion.
This LoRA may affect poses and framing as the weight increases. If these aspects are important to you, I’d suggest keeping the weight between 0.3 and 0.5 and improving the weight of your poses and framing tags.
If you’re looking for more details, you can use a weight between 0.7 and 1.
Prompt example:
jurbansc3nery, outdoors, building, scenery, road, street, sign, sky, day, tree, road sign, ground vehicle, city, real world location, power lines, lamppost, utility pole, motor vehicle, car, crosswalk, traffic light, cityscape, blue sky, cherry blossoms,
During my tests i had no issues, but it seems that on PixAI some models will ignore the character and just process the background, i will try to fix that, but in the meantime if that happens to you, just lower the Lora's weight and remove some words out of the prompt like below:
sign, day, tree, road sign, power lines, lamppost, utility pole, car, city, traffic light, cherry blossom,
