The was trained based on 1,000 hd video game assets. So people using picture to picture or text to image while using a image refference. This model should help in assisting with old video game texture enhancments.

The model is not intended to act as a image shader, or character lora.