Hello' ! o/
Requested LoRa, hope you'll like it. (Got pretty good results with it, sometimes it will take 1 or 2 images to get a good one as the AI struggle to understand this position but it's pretty ok)
Settings :
Works pretty well on different PONYXL checkpoints but you'll have the best results with WildCardX- XL PONY ! https://civarchive.com/models/333134/wildcardx-xl-pony
If you want more detailed images, don't hesitate to use https://civarchive.com/models/229213?modelVersionId=383563 with a strenght between 0.6 and 0.8, it's giving great results.
DPM++ 3M SDE Exponential / Euler A
20-40 Steps
CFG 2-4
PROMPT :
Use the trigger words "PsySmo", "crazy girl smile" , "breast smother" and don't forget to use the regular "score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up" at the beginning as it is a PONY model.
Also, don't hesitate to add "insanely detailed", "(masterpiece)" at the end of your prompt, I found that it was helping a lot. (At least with Wildcard's model)
NEGATIVE PROMPT - I'm using the one written on WildCardX-XL-PONY, which is : (worst quality, low quality, normal quality, lowres, low details, oversaturated, undersaturated, overexposed, underexposed, grayscale, bw, bad photo, bad photography, bad art:1.4), (watermark, signature, text font, username, error, logo, words, letters, digits, autograph, trademark, name:1.2)
PS : If you're struggling to reproduce the images I got, try addind "lying on bed" to your prompt. Did several tries and it helps the AI to understand the position as it's more "natural".
ENJOY !
Description
FAQ
Comments (3)
I've not tried your LoRa yet but as for your description things like prone bone, on stomach and lying should improve things.
Just tested, for Vivid's or Vendo's you need 0.8 at least (if realism) anything below that is a 10-20% change of getting a pose.
Also this is a great lora to force prone bone somehow.
Thanks for the feedback, hope it will help those who want to generate with those chekpoints ! Greatly appreciate it.







