v1.0 160mb:
Lora trained on PonyXL with 87 photos of Katheryn Winnick. Use it with AutismMix! Lora strength should be in most cases between 0.7-1.2. Ideal strength with keyword "katxwinnick" is 1.0-1.1. Can be used with following prompting template :
positive : score_9, score_8_up, score_7_up, {source style e.g. source_anime}, {character action and description}, katxwinnick, <lora:katwinnick_87img_ponyxl_autism_xl_1_standard_merger-18_54_71_025_03_045-katxwinnick:1.05>
negative : monochrome, child, simple background, loli, 3d style, render, {anti source style e.g source_photo}
Description
FAQ
Comments (2)
So far, I was unable to reproduce anything near the images here with this model. The body form seems to match, but I always have to use ADetailer for the face to get something halfway recognisable.
Thats weird. If you used the model and AutismMix with the prompt I used you should be able to get by.