This LoRA is trained primarily for the purpose of creating NSFW images.
If you are not comfortable with NSFW images, do not use this LoRA.
It is not yet fully trained and may not produce the desired image.
<<Note
We recommend a weight of 0.7 to 1.0. However, heavier weights will result in more reddish tones, so please add negative prompts such as “reddish, red skin, red light”.
For use with character LoRA and costume LoRA,
Add an outfit type to the negative prompt.
Set the weight of this LoRA to 1.0 to 1.1 to lighten the weight of the LoRA being used in conjunction.
Specify “from below” to make the composition from below.
Translated with DeepL.com (free version)
Description
V1.0 was messy, so I updated it.
Increased the number of study sheets to 400
Generation probability has been increased, but the positions and clothing are now limited.
The most likely positions to be generated are
Normal position
Doggy Style
Cowgirl
and not so much in the standing position.
The probability of being drawn in skirts and other clothing that exposes the lower half of the body increases, but leotards, panties, and other clothing with fabric extending to the crotch area are more likely to be drawn halfway up the body.
The color of tentacles is still difficult to be reflected.
We hope to improve this in future versions.