If you like what I do please consider supporting me on Patreon and contributing your ideas to my future projects!
Poses to use in OpenPose ControlNet modeled from professional nude model photographs.
I tried to pick the best ones that gave the best results. If your model looks mildly contorted it might be because it wants to show the front side when the pose was taken from the backside. Typing "show ass" and "view from back" into the prompt will usually fix that.
If your model is clipping outside of your image it might be because you are using a different aspect ratio then what was used to make the pose. Try using a similar aspect ratio. 768 width x 512 height works well.
If the pose isn't coming out the way you want try lowering the Ending Control Step.
For #37 you can type "dildo, vaginal insertion" into the prompt.
For #63 you can type "licking ass, cunnilingus" otherwise the other girl might rest her head on the girls ass like a pillow.
For #80 you can type "lesbian, yuri, kissing".
For #88 you can type "1woman, 1man, blowjob, penis, fellatio, man and woman should both be in view". You can get more accurate results with regional prompting. This one isn't always the most accurate as the penis likes to float in the air sometimes.
I used Majicmix for testing.
Invite to my public discord: https://discord.gg/UFkPTfTDWs
Description
FAQ
Comments (16)
Those are actually pretty good
Even though these pose are nice, but unfortunately, the fucking Openpose is dickhead!~
The number 36 foot will go wrong.
I have a lot of respect for any content creators for Stable Diffusion. But please, for heaven's sake, do the reference generation for the poses. I don't have Stable Diffusion built into my brain, I can't imagine what pose I see in a picture.
Maybe next version, felt like too much work with this many poses tbh. Think of it like a grab bag full of surprises or run them in batch mode so its different pose every time.
@getphat I would love to do that... Except that generating on a 1050 ti takes 7-13 seconds per iteration. And that turns into 210-390 seconds at 30 steps. That's over 4 minutes per image (on average).
Thanks for the answer, though :-)
@QwiziRAM holy fuck yeah I can't imagine that, I'm spitting images out in a hand full of seconds, I'll have to include them next time for you guys with lesser capable gpus I guess, I'll try to do another 100 in the next couple weeks
skill issue
@QwiziRAM Wow...when I saw that people were using GTX1070s with A1111 I thought I might have wasted my money getting a 3060...if this is what I would be dealing with, I see I did not make a mistake holy shit.
ok but... why is n 100 819x512?
It's giving errors due to not being in a valid dimension XD
odd, when i use controlnet its like its ignoring my original model and using only the openpose model any advice?
Nevermind i wasnt paying enough attention it was the resolution, lol
Good poses but it'd be nice to have some previews included :)
Done
For some reason, when using "detect from image" - the skeleton just stands still and the image is loaded as a background.
Wouldn't it be better to export these poses as a json file?
It works perfectly fine in controlnet when loaded without a preprocessor model