actually, I didn’t assign any character-specific tags. what I trained is simply an animation style. you can use this style together with character tags to generate images, provided that the base model recognizes the character.