Download the zip file. Inside are 23 .png files used for poses. Requires the ControlNet extension for Automatic1111. If you don’t know what those 2 things are, check out some youtube tutorials on setting them up.
I’ve saved you some time and meticulously prepared 20 different heroic stance poses that the Stable Diffusion models can handle pretty well. Some easier than others. Mostly likely due to the fact that not many images of this type were trained into the model. So some models will have better results than others. Example, there are 2 “falling” poses that are an extreme challenge as the models don’t like generating people upside down. If you manage to get a good image from it, post it to show off your expert prompting skills. The sample images here were generated using the legendary A to Zovya RPG Artist’s Tools model.
How to use:
Download the zip file. Inside are the poses that you will instruct controlnet to use instead of generating one from an existing image. Craft your prompt as you usually would. Let the model know with prompts, this is a (hero / flying / fighting) stance pose to help it along. In the controlnet settings, choose “none” for the preprocessor and “openpose” for the model. Adjust weight depending on how difficult the model wants to be. The pose images provided are 512×768 aspect ratio of 2:3. You can crop or add to the existing images with your favorite program to customize your composition.
Do you have requests? I’ve been putting in many more hours lately with this. That’s my problem, not yours. But if you’d like to tip me, buy me a beer. Beer encourages me to ignore work and make AI models instead. Tip and make a request. I’ll give it a shot if I can. Here at Ko-Fi