Would appreciate images, feedback, requests, reviews, and of course comments in the comments. Please check out my other models if you like this one.
this is a Lora trained on images of multiple girl and guy paizuri images (so 2girls+2guys)
plenty of the images seemed to be flagged as of uploading, hopefully they get unflagged
don't expect out of this world results, this is really edging towards the limit of what's capable right now. got lazy so plenty of images are low quality, and some I just messed around with.
weights 0.8 - 1
optional tags: from_side, from_front, from_above, pov, straddling
2girls, 2boys OR 3girls, 3boys OR 4girls 4boys OR multiple_girls multiple_boys,
(not much art of 4 and beyond, 3 can work ok sometimes, 2 seems the best),
(pov is for pov of being the middle man, from_front is kinda like pov but it has a wider field of view, across_view is when viewer is receiving paizuri as well as a person in front of the viewer, from_side is usually an angled view from the right. Would probably suggest putting pov in negative for other variants) - [ terrible explanation and I don't even know if they work correctly ]
640x512, 768x512, 1024x512 were the dimensions that worked most consistently for me
Description
FAQ
Comments (4)
Can get good results with this by rolling the gacha until you get a decent starting point, then inpainting over each girl with character-specific LoRAs one at a time as if you were doing singles, then usual inpainting touch up job.
LoRA is very useful to get to that 'starting point'.
is it possible to recreate this model in PONY?
because as a model, I would find it interesting for PONY diffusion.
I just thought of making this for pony, should be up in a couple days