Chibi风格的模型 使用Animagine XL V3.0训练 触发词是chibi 推荐权重0.9-0.7
具有较强的泛化性，对于各种人设（oc）tag表现良好。相对于其他的chibi模型而言，对于背景tag表现也比较出色。
ps:我在初学者的时期并没有把每张图片都进行手动ps处理，所以留下了大量有水印的素材。为此，请在负面添加：signature
Model trained in Chibi style using Animagine XL V3.0, with "chibi" as the trigger word. Recommended weights are between 0.9 and 0.7.
It has relatively strong generalization and performs well for various personal (OC) tags. Compared with other chibi models, it also performs relatively well on background tags.
PS: During my beginner phase, I didn't manually Photoshop every image, so a lot of the materials still have watermarks. For this reason, please add a "signature" on the negative prompt.