v5: Better quality, added more images to the database, changed tags and changed base training model from Illustrious to Chenkin v0.1. Works on Illustrious and NAI models.

My first "Style" attempt of a LoRa based on Black Desert Online Ultra Remastered Graphics, i used 113 8K screenshots (Civitai downscales them so...) of scenery/landscape of multiple locations of the game with 5 different lighting types (base lighting of the zone, daylight, sunset, night and foggy night, 4gb of ultra quality images) so it shouldn't affect characters.

It doesn't need a trigger word to use, so you can merge it on your models. Works better on 2.5D, 3D and Realism Models but can be used on 2D and works on Illustrious, NoobAI and on some SDXL models. It changes output with same seeds, so test it.

It can add some stabilization to background, realism, add details and enhance postprocessing effects like depth of field, chromatic aberration, vignette, changes lighting of day, dark, sunset, night and fog, and changes contrast.

The images used in the model use the same SEED and doesn't use any type of Upscale, Hiresfix, CFG Rescale or ADetailer, so quality can be improved.

Yes, the main image is to show my Shai (is an adult so no worries).