详情
推荐
v2.0
v1.1
DarkVision

DarkVision

28.3K
152
327
#逼真的照片
#女性
#analog
#基础模型
#男性
#写实
#女性
#checkpoint trained
#逼真的照片
#女性
#analog
#基础模型
#男性
#写实
#女性
#checkpoint trained

All the preview images are generated using a 568 x 767 resolution, with a 2X upscale, 4X ultra-sharp upscaler, and minimal prompts with after-detailer extension.

For the best quality, I highly recommend that you use the same negative prompt that is used in the example images and modify it.

V.1

Vae is baked into the model

Mainly focused on Photo-realism

Recommended LORAs that I tested to enhance the quality:

  • I highly recommend using Adetailer.

Detail Slider - use it at strength of 0.4 - 0.6 for significant boost in realism

PerfectEyes - strength 0.7

Surreal/Freckled Skin Slider - it increases the details on skin, use it at the strength 1.0 - 1.25

hairdetailer - strength 0.6

epiCRealLife Enhancer - strength 0.3

Recommended settings:

Positive -- (film grain:1.5),8K raw photo, Best quality, masterpiece, ultra high res, cowboy shot, (photorealistic:1.4), professional lighting, 1girl, solo, long hair, looking at viewer, natural skin texture, realistic eye and face details, rouge \(makeup\), parted lips, 

Negative -- ((((tattoo, top down, hat, three legs, nail polish, necklace, toes))), building, reflect face, fishnet, one leg, curtain, sketches, (worst quality:2),paintings, sketches, (low quality:2), (normal quality:2), lowres, ((monochrome)), ((grayscale)), skin blemishes,extra fingers,fewer fingers, badhandv4 ng_deepnegative_v1_75t, badhandv4 (bad-picture-chill-75v:2), easynegative, ng_deepnegative_v1_75t, (low quality:1.3),(deformed:1.3),(malformed hands:1.4),(signature:1.2),(artist name:1.2),(watermark:1.2)
  • Sampling method: Euler A, Euler, DPM++ 2M SDE Karras, DPM++ 2M Karras, DPM++ SDE Karras.

  • Sampling steps: 26 - 32.

  • Width x Height: 512 x 512 or 512 x 768.

  • CFG Scale: 6 - 8.

  • Clip skip: 2.

Hires, fix.

查看译文

评分与评论

3.5 /5
0个评分

尚未收到足够的评分或评论

C
Cake_
29
319
公告
2023-10-11
发布模型
2023-12-10
更新模型信息
模型详情
类型
Checkpoint
发布时间
2023-12-10
基础模型
SD 1.5
许可范围
来源: civitai

1. 转载模型仅供学习与交流分享,其版权及最终解释权归原作者。

2. 模型原作者如需认领模型,请通过官方渠道联系海艺AI工作人员进行认证。我们致力于保护每一位创作者的权益。 点击去认领

创作许可范围
在线生图
进行融合
允许下载
商业许可范围
生成图片可出售或用于商业目的
允许模型转售或融合后出售
QR Code
下载SeaArt App
在移动端继续你的AI创作之旅