This is a LoRa model that trained by 100+ img generated with NijiV5,and it's the multires noise version.

This model can modify the stroke of the contour and has a certain impact on the artistic style.

weight 0.6-0.8 maybe generates better imgs,,trigger is kamibana.