Check in models version for info, I did not edit this part for so long!
Support me on Kofi:
https://ko-fi.com/ziggart If you like my work a coffee is always welcome ! :)
Edit: V22.3 a huge thanks to users like https://civarchive.com/user/YeiYeiArt with his wonderful loras that I merged into this model as well as, https://civarchive.com/user/tappy for his overwatch pôny Lora and his disney 3d pack Lora
This models was intended to reproduce Hentai animation nsfw. It can produce other stuff.
Resolution tested 832x1216 (I always use that one both side)
You may also join my discord server to discuss models and share generation.
Banned from Twitch, so hello to Kick, will do a live stream easter weekend to do some comfyui tutorials !
You can find my workflows here: https://civarchive.com/articles/4338
I believe there is some animagine and pony deep inside my merge so License should follow.
Description
Here is a new base for my next hentai model to come, it's basically a mix of 6 of my best hentai models: 22.3 / 23.3 / Nextgen V1 / V2 / DarkHentai V1 and LightHentai V1
Hope you like it, use your usual settings it works with almost everythings....
FAQ
Comments (5)
Why are you not showing generation data for your sample images?
Samplers and number of steps is a crucial information for evaluating Hyper models!
already explained why several time, my workflow in comfyui save it all in the png but when uploaded to civitai it does not take it in, I have to use some special tricks that make it really more slow, so sometime I'm lazzy but you can get it all either by asking or coming on our discord, all is there.
for sampling methode / steps / cfg I suggest you try your usual first, it usually works ;p) my sampling methode actually is:
Euler Ancestral cfg pp / AYS sdxl - 25steps - cfg 2 with perturbed guidance at 1.3 and guidance limiter cfg rescale at 0.3
@mskiller51
Would it really kill you to manually add full generation data manually to ONE image for each model?
P.S. 25 steps for Hyper? Are you for real?
Hyper is supposed to be small steps model, 25 steps doesn't tell anything about the quality of its generation in the modes people actually would try to use it.
@ateist yeah it would lol, seriously, try all your settings before complaining, most of my models are fine with most settings, IO also help people around on discord, so stop complaining and have fun with it, yes 25 steps are my settings, but it aslo works as an hyper in 8 steps...Also your comment appears on main, so how I know which model are you speaking about ? and for my settings, I believe I always wrote them into details of the model version....so you have it for one, and I did push civitai save on my latest model for everyone to get all gen data....Did you ever spent time creating a model, posting it and explain all that ? I don't think so, so please I do believe your comment is a little arsh and out !
@mskiller51 https://github.com/nkchocoai/ComfyUI-SaveImageWithMetaData
1 node that replaces the standard image saver and everything the image has in its generation data from Comfy is saved on Civitai