The combined versions lean most heavily toward realism.
I am generating images in more cartoon/amine styles and using those to balance the datasets one concept at a time.
Version 2.0 and Refined both do fisting really well. It was the easiest concept to train, so I will likely wait to do self-fisting until last.
Recommended settings:
Cfg: 5
Steps: 25-28
Weight of 1 or 0.9.
Hi-res your image with a lower weight 0.6 or 0.7.
This Lora does a lot of different things. Look at the prompts to see how to group triggers, and please share your posts to my model.
Description
This version will be much kinder to faces, and less style opinionated. Although, I haven't been able to make it as transparent as I'd like.
Do not just copy paste all the trigger words.
Look at sample pics for ways to combine the triggers.
If you're getting bad faces, try this in negative: "pixelated face"
Note: I had grander plans to release this as it's own lora called "Anal Show", so that's why you see that in the sample pics.
Here are a couple fun examples of things that work:
1girl, solo, lying on back, (hips lifted:1.5), 4nal_f1sting, s3lf_f1sting, feet on floor, arched back, wet pussy, sexy body, beautiful face,
2girls, dark-skinned female, 23 years old, thin and toned body, thick thighs, big ass, thin waist, long hair, curly hair, side view, profile view, dramatic angle, bent over, arm support, butt, pussy, panties pulled down, around thighs, skindentations, (Str3tching_another:1.4), 4nal_str3tching, analingus, licking, small anal prolapse, fingers in asshole, large breasts, bent over, chest press, counter, coffee bar, dimples of Venus,
Additionally, there are some stubborn tags that I didn't list in the trigger words, but they are trained, and are cool if you can get them to work.
Example below:
1girl, solo, ((Hands_front_and_back)), 4nal_str3tching, on knees, spread legs, head on pillow, nude, visually interesting, sexy, rear view, brown hair, looking back,
Please try it out and let me know what you think.
What I did for this version:
Culled some source images for quality
Added about 100 additional images, many of which were anime or painting style.
Cropped or censored faces that were having too strong an impact on the lora output
Lowered network dim and alpha