Wuhua is a traditional bondage technique from ancient China. This LoRA is designed to reproduce the characteristic Wuhua binding pose and rope arrangement.

Increasing the LoRA weight can improve pose accuracy and consistency. In most cases, I recommend keeping the weight at 1.5 or below, as excessively high values may begin to affect overall image quality.

Side-view (from side,) compositions may be less reliable than front or back views. For characters with long hair, back-view generations may also automatically split or move the hair aside to make the bound arms and rope structure more visible.

If you want the hands to remain visible in a front view composition, adding:

clenched hand,

may help.

A significant portion of the training dataset consists of 3D/3DCG images. In rare cases, this may slightly influence the rendering style. If the result looks too 3D or realistic, try adding the following to the negative prompt:

3d, realistic,