Examples were made with anime model but it should work with any model.For prompt and settings just drop image you like to PNG info.Resolution for txt2img: 512x768ControlNet settings:Preprocessor: none Model: openpose.Control mode: Balanced.Control weight: 0.8 - 1.

[OpenPose] From above

To fix some of the issues with background you can use additional depth map or MLSD to get better perspective (as a second ControlNet unit)

Files with SHA256: 99d1fdab5db7f014f89558f0888a901d9472800801d5b1c4beff95a7e33c0c01