It should be a type of collage art, and specific elements can be controlled through vocabulary
The trigger word is "split composition" because I didn't quite understand it during training.
The recommended weight is between 0.6-0.8. When the base model uses a two-dimensional painting style, there is a chance that the effect will be similar to the different painting styles on both sides (because the training materials are all realistic)
The usage method can be: trigger word + character type (girl, boy, etc.) + element (color, item, etc.) + background (if not added, it will be similar to a simple background)
