As the model was trained primarily on manga-style illustrations, it tends to produce a strong "flat color" look.

Try combining it with other style LoRAs or experimenting by lowering the LoRA weight to 0.6–0.7 as needed.