This LoRa model produces semi-realistic, Niji-style images.

Try using a strength of 0.8. If the output does not match your prompt, try adjusting your prompt by including terms like "semi-realistic," "Niji," or other keywords relevant to the type of image you want to create.