Type | |
Stats | 199 132 2 |
Reviews | (15) |
Published | Mar 18, 2024 |
Base Model | |
Hash | AutoV2 AED7E703FC |
In order to adapt to more usage scenarios, it is recommended to read these reference usage methods.
1. If you want to retain the refined effect but need color photos, you can hang another portrait lora and adjust the weight of the two loras. You can also use ControlNet’s [Recolor] to convert black and white to color.
2. If you need to create an Asian girl’s face, you can use domestic large portrait models (墨幽人造人_MoYou_artificial_person, XXMix_9realistic, etc.), and just reduce the lora weight appropriately (between 0.65-0.8);
【Model introduction】
This LoRA model can generate refined portraits in black and white style with exquisite makeup, natural skin, and mainly side lighting. Generalization tests have been conducted on different hairstyles, clothes, accessories, etc., and all have good results. The training set consists of half European and American girls and half Asian girls. If you need to stably produce pictures of Asian girls, you can appropriately reduce the weight and use large portrait models of domestic big shots;
【Model Application】
The pictures are all refined portraits of girls, mainly close-up half-length photos, which are suitable for effects that require some facial expressions, such as
1. Close-up pictures of models in the beauty industry;
2. Models in the beauty industry;
3. Photography reference, etc.;
【Picture parameters】
Checkpoint: any portrait type is acceptable. For specific effects, please refer to the test pictures in the return area. If you have no special requirements for the face shape, ReVAnimated_v122 is recommended. For Asian face shapes, XXMix_9realistic_v4.0 is recommended.
VAE: Automatic or None
CLIP: 1-2
Sampling method: DPM++2M Karras
Sampling steps: 20-22
Image size: 768*1024, it is recommended to enable high-resolution restoration
ADetailer: no need
CFG scale: 7
lora weight: 0.6-0.9
Trigger Words：none
【Prompt】
prompt：
1girl,solo,monochrome,greyscale,looking at viewer,realistic,parted lips,lips,white shirt,long hair,close-up,exquisite makeup,upper body,masterpiece,best quality,8k,insane details,intricate details,high detail,((masterpiece)),((best quality)),
top prompt：monochrome，close-up,exquisite makeup；
Negative prompt：
NSFW,(worst quality:2),(low quality:2),(normal quality:2),lowres, normal quality,((monochrome)),((grayscale)),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:1.5),(too many fingers:1.5),(unclear eyes:1.331),lowers,bad hands,missing fingers,extra digit,(((extra arms and legs))),badhandv4,
---------------------------------------------【重要使用提醒】----------------------------------------------------
为适配更多使用场景，建议必读这些参考使用方法
1、想保留精修效果但是需要彩色照片的，可以再挂一个人像类的lora，调整两个lora的权重即可。也可使用ControlNet的【Recolor重上色】进行黑白转彩色效果
2、需要出亚洲女生脸型的，可使用国内的人像大模型（墨幽人造人，极氪写实等等），适当降低lora权重即可（0.65-0.8之间）；
【模型介绍】
本LoRA模型可生成妆容精致的黑白风格精修人像，磨皮自然，侧光为主。泛化测试了不同的发型，衣服，配饰等，均有不错的效果。训练集欧美和亚洲女生各一半，需要稳定出图亚洲女生的，可适当降低权重，用国内大佬的人像大模型；
【模型应用】
出图均为精修的女生人像，特写半身照为主，适合需要一些需要怼脸的效果，比如
1、美容行业的模特特写图；
2、美妆行业模特；
3、摄影参考等；
【出图参数】
大模型：人像类的均可，具体效果可参考返图区的测试图，对脸型无特别需求的，推荐ReVAnimated_v122，亚洲脸型推荐XXMix_9realistic_v4.0
VAE：自动或无均可
CLIP：1-2
采样方法：DPM++2M Karras
迭代步数：20-22
出图尺寸：768*1024，建议开启高分辨率修复
面部修复：无需
提示词引导系数：7
lora权重：0.6-0.9
【触发词】
无
【出图提示词】
正向提示词（参考）：1girl,solo,monochrome,greyscale,looking at viewer,realistic,parted lips,lips,white shirt,long hair,close-up,exquisite makeup,upper body,masterpiece,best quality,8k,insane details,intricate details,high detail,((masterpiece)),((best quality)),
建议必填提示词：monochrome，close-up,exquisite makeup；
负向提示词：
NSFW,(worst quality:2),(low quality:2),(normal quality:2),lowres, normal quality,((monochrome)),((grayscale)),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:1.5),(too many fingers:1.5),(unclear eyes:1.331),lowers,bad hands,missing fingers,extra digit,(((extra arms and legs))),badhandv4,