使用介绍
A full-character LoRA model based on the anime “Magic Girl Ilya”
This is LoRA style, so if you want to retrieve the character, please use the tag that is most relevant to the character to guide the model.
For example:
1girl,busty, long_hair, blonde_hair, red_eyes, brown_eyes, very_long_hair, hair_ribbon, drill_hair
You’ll get Louisa
1girl, smile, short_hair, open_mouth, blonde_hair, brown_eyes, hair_bun, double_bun,
Will get Gakumazawa Tatsuko.
1girl, solo, long_hair, bangs, brown_hair, black_hair, brown_eyes,
You’ll get Miyu
1girl, solo, long_hair, blonde_hair, red_eyes, white_hair, grey_hair
Will get Illya
1girl,dark_skin,pink_hair
Will get Kuro
etc
If you are trapped in how to manipulate the character input into the automatic tag, the example image will be helpful.
The model is trained by using the tag constructed by smiling wolf wd-v1-4-swinv2-tagger-v2 under the threshold of 0.25. The output tag is likely to match the training data.
The basic model suggested for lora is the 7th generation animeC.
基于动漫《魔法少女伊利亚》的全角色劳拉模型
这是LoRA风格,所以如果要检索角色,请使用与角色最相关的标签来引导模型。
例如:
1女孩,胸部丰满,长发,金发,红眼睛,棕色眼睛,长发,发带,钻发
会得到露维亚
1女孩,微笑,短发,张嘴,金发,棕色眼睛,发髻,双髻,
会得到岳间泽龙子
1女孩,solo,长发,刘海,棕色头发,黑色头发,棕色眼睛,
会得到米语
1女孩,solo,长发,金发,红眼,白发,灰发
会得到伊利亚
1女孩,深色皮肤,粉红色头发
会得到李书行
等等
如果被困在如何操纵字符输入到自动标记的例子图像将有所帮助。
该模型用在. 25阈值下使用微笑狼wd-v1-4-swinv2-tagger-v2构建的标签进行训练。由此输出的标签很可能与训练数据相匹配。
建议用于lora的基本模型是第7代animeC
触发词 Trigger Words: | |
使用提示: | CLIP SKIP: 1 |
作品参照
1girl,busty, long_hair, blonde_hair, red_eyes, brown_eyes, very_long_hair, hair_ribbon, drill_hair
tag:动漫,角色,命运
Warning: A non-numeric value encountered in /var/www/html/wp-content/plugins/erphpdown/addon/tuan/index.php on line 209
评论0