人物lora为杀戮的天使中的女主：瑞吉儿也就是Ray(UwU)
使用Animagine XLV3训练推荐权重为0.9-0.8因为我不希望把人物的自然特征和外在特征绑定在一起（主要是想着可以给ray换装），所以提示词有如下几类：
主要提示词有：rachel gardner, satsuriku no tenshi
人物样貌特征词有：long hair, blonde hair, blue eyes, hair between eyes, collarbone,bangs, very long hair , very long hair
服装提示词有：striped shirt, shirt, striped, choker, jacket, white jacket, open jacket, open clothes, shorts,short shorts, bag,blackfootwear,boots,
在使用过程中可以自由组合达（比如更换服装）到自己想要的效果，比如：1girl, rachel gardner, satsuriku no tenshi, solo, long hair, blonde hair, blue eyes, white background, hair between eyes, necklace,looking at viewer, upper body,simple background, collarbone,closed mouth,wedding dress, very longhair,
题外话：ok这是我的第一个人物lora也是我第一次使用正则（感觉有时候玄学有时候很有用），还有就是关于Ray的全身素材实在是太少了，迫不得已才使用正则(所以也不是很推荐使用这个lora生成全身图)。
The character Lora is the female protagonist in "Angels of Death" known as Rachel Gardner, or Ray (UwU).
Trained using Animagine XL V3, the recommended weight range is 0.9-0.8. I do not wish to strictly bind the character's natural features and external appearance together (mainly considering the option to change Ray's outfits), thus the prompt words fall into several categories:
Main prompt words include: rachel gardner, satsuriku no tenshi.
Character appearance feature words include: long hair, blonde hair, blue eyes, hair between eyes, collarbone, bangs, very long hair.
Outfit prompt words include: striped shirt, shirt, striped, choker, jacket, white jacket, open jacket, open clothes, shorts,short shorts, bag,black footwear,boots,
During usage, combinations can be freely made to achieve the desired effect, such as: 1girl, rachel gardner, satsuriku no tenshi, solo, long hair, blonde hair, blue eyes, white background, hair between eyes, necklace, looking at viewer, upper body, simple background, collarbone, closed mouth, wedding dress, very long hair.
Side note: OK, this is my first character Lora and my first time using Regex (which I find sometimes mystifying yet sometimes very useful). Also, there's a real scarcity of full-body materials for Ray, which forced me to resort to using Reg(Therefore, I don't particularly recommend using this Lora for generating full-body images).