Found a GIGANTIC Library of poses. (Patreon)

There are just over 2000 poses included in an almost 6GB zip file.

If you are new to ControlNet, let me introduce you to this world. For reference: The input image provided to Controlnet is shown in the first image, while the output from Controlnet's DW Pose Preprocessor (DW being a newer preprocessor with improved body/face/hand pose detection) is shown in the second image. The stick figure image generated by Controlnet can be used to guide the Stable Diffusion Model to pose any desired character from the preprocessed image.

Meaning: You can position them in any way you like. There are 3D and 2D GUI programs that allow you to create poses manually and use the preprocessed image as the desired pose for your character.

LINK to Online GUI (It's free): https://zhuyu1997.github.io/open-pose-editor/

BTW, This gives me a fun Idea for Patreon Members, But again I don't want you guys to work to make a pose for me to use (it would be included in cheapest tier, no way I would charge extra for you guys to do the work).
But request Poses in the way of using the GUI on your own time and then sending me the 4 Images(bottom right corner in image below) and I can use those to make the characters in poses you like.
This is a maybe and kinda may defeat the purpose of it being a benefit since you guys end up working for it xD

EDIT: I finally remember how Depth Map hands work, it's required to have depth map control net after open pose if you want the hands to look correct, DW pose hands work well, but their is no depth so SD doesn't know how it looks when seeing it in stick figure mode, so depth map helps with that since it will see the creases and the look of the images, Also when you set the hands up they are aligned with the Stick figure! that's why Depth hands go after the pose.

Here is what it looks like: (See above image and then look at this: (I have other images but they are long and it will extend the post much longer xD........oh yeah...their's Normal Maps...the kind that you find for video game textures or that go with it! xD

This process can be time-consuming but useful if you want a specific pose to work. I will be testing this out on characters for now and see how things turn out. The advantage is that controlnet won't affect the style and appearance; it acts like a under the hood thing on the image you want.

The original was from earlier this year, around April. As you can tell, it's not perfect. Finally the last image, I tested elsa on this and it's quite good! Obviously I can alter the pose and make her hands posed better to not look as messed up, this is an old image anyways :), This is the result of using ADetailer for the breasts and face (I didn't modify the vagina because it was fine). The models were improved after many iterations of SD model improvements and Dynamic LoRa to keep her pose.

This alone is just one of MANY things you can do with Controlnet xD All thanks to Stable Diffusion being open source! This is why I have a bone to pick with Midjourney and Dall-e, because most AI art haters think all we do is just type in prompts, and they only do that because they haven't heard what SD really is, since it's not that well known. Well, at least people who know AI art may have heard of it.

I won't work on much but when I'm back On Tuesday I will get to this, I already have 2 sets being scheduled for that day also!

Extra Results: (I will add more if I get the time)