When the model version is updated, I’ll put new prompt in “About this version” if necessary.

V1.0

In SDXL, ControlNet performance is not as good as it’s in SD v1.5, and usually needs 2 controlnet to secure the output. To secure the pose could be reproduced in SDXL, the model will include 2 controlnet model files at least.

OpenPose bone structure (also json file), Depth map (lineart sometimes) and example image with prompt information in zip file.

Prompt: climbing,

Optional Prompt: mountain,

ControlNet weight: OpenPose 1, and 0.5 for others.

Welcome to share your creation here!

Note:

Key word “climbing” can generate nice poses.

However, it’s very difficult to get real “climbing” image because SD v1.5 cannot align subject and background.