Intended to create images of people with denser and darker tatoo patterns. I recommend using values between 0.5 and 1. You an try to prompt more localized tattoos using : On arms: arms_slvs , arm_slv On legs : legs_slvs, leg_slv And : tt_body_sut , tattooed neck But the results are never too consistent, sometimes clothes are confused with tattoos by the machine... I got better results using <a rel="ugc" href="https://civarchive.com/models/207597/hassansdxl">Hassan</a> when prompting that way. Trained mostly with heavily tattooed white women sporting the so called "blackwork" style. The dataset is quite NSW leaning. Enjoy

Files with SHA256: 639709a66605784ee0bd1257456b779504a4cb0f63d3ac6c2ae94accb5063a1b