This is a CONFUSED LORA. Wouldn't it be better if it focused on a single image? Peeps wanting any form of multi-image output can trivially achieve this composite using Photoshop (etc), or special nodes in ComfyUI (etc).