Long time lurker here, I saw there are a couple other mura style LORAs - go check them out!Since I wasn't really satisfied with the available options I went to train my own, this is a result of a hand selected database and a merge with another cartoon LORA - hope you like it!The LORA works best at strength 1, with the trigger "murastyle".Faces and overall results might come out a little unbaked since the style in its core is a bit more rough, I found through trial and error that the best generation settings are:[v] Hires, fix:<ul><li>Upscaler = R-ESRGAN 4x+Anime6B</li><li>Upscale by = 1.5</li><li>Hires steps = 20</li><li>Denoising strength = 0.45</li></ul>[v] ADetailer:<ul><li>(default settings)</li></ul>The hires fix takes care of the rough lines and ADetailer handles the faces, the result in the preview were made using these settings.

Files with SHA256: 2996beb83792ea70ac212f136536946b165796b3b54fc999ad58233f9d21bd2d