4/15 update:new version for AnimagineXLV3.1, it's stabler than 1.0, and you can even use it as a 2.5D or 3D stylizer. It's also compatible with other pose or character loras.(Due to the data, this more is not good at drawing bare feet, you might need gacha)Recommended weight: 0.9 - 1.1,Trigger: asanagi-figure, photorealistic, photo \(medium\),<hr />Trained specifically for figure works Illustrated by Asanagi, trained on AnimagineXLV3, but also works on v3.1 (though it seems that V3 performs a little bit better). Most data was collected on Twitter. Because of the lack of the data, it might need a little bit gacha to work well.Recommended weight: 0.9 - 1.1,lower the weight if faces are twisted, and specific character tags are helpful for making faces look good.Tigger: asanagi-figure,Recommended tags (Try to remove them if faces are weird): photorealistic, photo \(medium\),Support nsfw contents, but you may need to add extra tags if some parts are missing.<hr />专门为朝凪的手办化作品训练的画风，底模AnimagineXLV3，3.1也能用不过貌似3.0效果稍微好一点。数据来源主要是推特返图，因为能用的图很少所以这个模型用的时候可能要稍微抽下卡建议权重 0.9 - 1.1, 如果人物的脸扭曲了或者很怪就降低些权重，使用特定的角色tag可以比较有效的避免脸部抽风触发词：asanagi-figure,建议同时使用（如果脸部很怪的话也可以移除这些tag）： photorealistic, photo \(medium\),也可以画nsfw内容，不过需要额外的tag来补充丢失的部分

Files with SHA256: 6e6978302f867552bd5fec2b1d2916ae00ab6f7530caa24a068426ab9c903def