开头申明：训练素材来源于网图，模型是Chilloutmix_Ni训练。训练后的Lora成品不推荐用于商业用途。如果你用于商业用途，一切风险由你自己承担，我不承担任何责任。

使用方法：推荐使用Chilloutmix_Ni大模型，lora权重推荐0.9，VAE推荐真实感840000。（远景遇到面部五官不太行的情况推荐打开ADetailer面部修复），添加更多关键词与搭配其他lora食用效果更佳。

生图规格推荐：512*720或512*六40

目前精度：泛化性OK，无论是真人模型还是动漫模型都能很好地保持角色特征，近景效果不错，手部不太行，加折扇关键词后要抽卡，fullbody面部欠佳。后续会更新2.0，继续强化全身和其他方面。

封面图的生成参数；模型：Chilloutmix_Ni，VAE：真实感840000，lora权重：0.9，步数：50，采样：DPM++ 2M Karras，分辨率：512*720，种子：915963000

正面tag：1girl,solo,white hair,pink eyes,flower,bangs,(chinese clothes:0.8),lips,masterpiece,best quality,(pink dress:0.8),chinese style,Hair Accessories,Jewelry,See Through,lantern,Tindal effect,particle effects,illuminatio,impressionist art,real light and shadow,looking at viewer,(loli:1.1),double bun,happy,open mouth,(multicolored ribbon:1.3),(garden:1.3),open mouth,happy,(upper_body:1.3),cuteg,(night:1.8),realistic,ultra realistic 8k cg,(shadow:1.4),laughing,photorealistic,concept art,rock,

负面tag：(worst quality:2),(low quality:2),(normal quality:2),lowres,normal quality,skin spots,acnes,skin blemishes,age spot,(ugly:1.331),(duplicate:1.331),(morbid:1.21),(mutilated:1.21),(tranny:1.331),mutated hands,(poorly drawn hands:1.5),blurry,(bad anatomy:1.21),(bad proportions:1.331),extra limbs,(disfigured:1.331),(missing arms:1.331),(extra legs:1.331),(fused fingers:1.61051),(too many fingers:1.61051),(unclear eyes:1.331),EasyNegative,SimpleNegative_AnimeV1,坏图修复veryBadImageNegative_veryBadImageNegative_v1.3,

QQ讨论AI群：735666909，欢迎各位一起讨论学习和返图！

Statement at the beginning: The training material comes from the network graph, and the model is Chilloutmix_Ni training. The finished Lora after training is not recommended for commercial use. If you use it for commercial purposes, all risks are at your own risk, and I do not assume any responsibility.

How to use: It is recommended to use Chilloutmix_Ni large model, lora weight is recommended to be 0.9, and VAE is recommended to be realism 840000. (If you encounter a situation where your facial features are not good, it is recommended to open ADetailer facial rejuvenation), add more keywords and use it with other lora to eat better.

Recommended specifications: 512*720 or 512*640

Current accuracy:Generalization OK,Whether it's a live-action model or an anime model, the character characteristics can be well maintained,The close-up effect is good,The hands are not very good,After adding the folding fan keyword, you have to draw a card,Fullbody face is not good。 In the future, 2.0 will be updated to continue to strengthen the whole body and other aspects.

Parameters for generating cover images; Model: Chilloutmix_Ni, VAE: realism 840000, lora weight: 0.9, number of steps: 50, sample: DPM++ 2M Karras, resolution: 512*720, seed: 915963000

Front tag:1girl,solo,white hair,pink eyes,flower,bangs,(Chinese clothes:0.8),lips,masterpiece,best quality,(pink dress:0.8),Chinese style,Hair Accessories,Jewelry,See Through, lantern,Tindal effect,particle effects,illuminatio,impressionist art,real light and shadow,looking at viewer,(loli:1.1),double bun,happy,open mouth,(multicolored ribbon:1.3),( garden:1.3),open mouth,happy,(upper_body:1.3),cuteg,(night:1.8),realistic,ultra realistic 8k cg,(shadow:1.4),laughing,photorealistic,concept art,rock,

Negative tag:(worst quality:2),(low quality:2),(normal quality:2),lowres,normal quality,skin spots,acnes,skin blemishes,age spot,(ugly:1.331),(duplicate:1.331),(morbid:1.21),( mutilated:1.21),(tranny:1.331),mutated hands,(poorly drawn hands:1.5),blurry,(bad anatomy:1.21),(bad proportions:1.331),extra limbs,(disfigured:1.331),(missing arms:1.331),( extra legs:1.331),(fused fingers:1.61051),(too many fingers:1.61051),(unclear eyes:1.331),EasyNegative,SimpleNegative_AnimeV1,Bad image fixveryBadImageNegative_veryBadImageNegative_v1.3,

QQ Discussion AI Group: 735666909, welcome everyone to discuss learning and return to the map!