This is a LORA of the VTuber Tozura,

this LORA was trained with the model "anything4.5merge" I recommend that you use this model so that the image is generated correctly...