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:

• Upscaler = R-ESRGAN 4x+Anime6B

• Upscale by = 1.5

• Hires steps = 20

• Denoising strength = 0.45

[v] ADetailer:

• (default settings)

The hires fix takes care of the rough lines and ADetailer handles the faces, the result in the preview were made using these settings.