Is with a great pleasure that i present this Lora.
It's Composition Lora
I recommend strength between 0.7 and 0.9
I recommend use the keywords ear pulling
FOR F/F:2girls, ear pulling
FOR M/F (Very hard but doable): 1boy , 1girl , ear pulling
Add `femdom` to negative
Reduce strength to `0.6`
FOR F/M1girl , 1boy , ear pulling , femdom
INPAINT
VERY GOOD FOR INPAINT HAND PULLING EAR
Max Res 1024x1024
Description
FAQ
Comments (6)
V2 is way less overfitting.
M/F is less hard but still not easy.
Might release V3 soon with better M/F balance
Any chance it's capable of M/M or even has some slight M/M training?
I am a novice; may I ask how to prevent the prompt words of two characters from interfering with each other?
MysticFrostwind Once there's more M/f , M/M will be easier too... since there's enoguth Male Dom , and Male Sub ... current model has too much male sub.
adogcat2021782 I usually seperate out sections in a prompt for each character. So I'd do;
First line is basic prompts, angles and views, locations, lighting, etc etc.
Second line is either specific style loras or will be character information.
Third line is second character, kept contained from the first.
masterpiece, high quality, outside, sunlit, volumetric lighting,
BREAK,
(character 1, details, blah blah),
BREAK,
(character 2, details, blah blah),
That's usually how I handle multiple characters. Seems to work well enough, with most models only giving a 10% chance of mixing things up or making errors with Illustrious checkpoints. Most of the time the model understands what you're trying to do and will be able to handle it. Can also try Forge Couple / Regional Prompter if you're locally generating.
spanking_cutie I can try in a bit to make some M/M content with this lora that you can use later in future training, if that would help.