详情
推薦
v1.1
v1.0
Truality-Divas

Truality-Divas

3.2K
12
9
#名人

It's the Truality engine (pro version) https://civitai.com/models/158621/the-truality-engine by Tower13Studios

merged with and Divas (1.0 version) https://civitai.com/models/113337/divas by alexds9

It does celebrities really well. Most without even using a lora. It does do ??????, but I haven't seen it do anything ???????? without a separate lora for that, then it will.


Version 1.1 is a merge of the same 2 models with consistent factor (elucid) added to it. by consistentfactor444. https://civitai.com/models/9114/consistent-factor-euclid This improves face likeness of most lora models but it seems to have taken away some background capability. Hence why I named it 1.1 instead of 2.0. it may be worse then it was. I'm still playing with merges trying to make my perfect SD1.5. this is not it yet.


-To make my faces look good in full body photos I use the Adetailer extension

-Adetailer enabled (just face_yolov8n.pt which is the default setting in Adetailer) which is the default setting) but i leave the prompt box blank so it uses my main prompt.


I'm using Stable-diffusion-webui-Forge because i have less the 4gb Vram and this works fantastic for that. But you don't need that unless your computer is a potato like mine.


Below are the settings and base prompts I use to get good results with this model.

I use DPM++2M Karras

sample steps 25

CFG 7.5

720x1080

(my good computer died, larger photos just take too long.. even using forge)


To make my faces look good in full body photos I use the Adetailer extention

-Adetailer enabled (just face_yolov8n.pt) but i leave the prompt box blank so it uses my main prompt.

yes i wrote this twice because it REALLY helps, for those that do not know it's a simple extension you get from the extensions tab, basically you render a picture like normal, then it does a second pass on just the face and renders that like inpaint.. but automatically done for each photo. it also has setting for bodies and stuff but i only messed around with the faces.


positive prompts

Best quality,4k,Realistic,highly detailed,


---negative prompts

Watermark,Text,child,censored,deformed,bad anatomy,disfigured,poorly drawn face,mutated,extra limb,ugly,poorly drawn hands,missing limb,floating limbs,disconnected limbs,disconnected head,malformed hands,long neck,mutated hands and fingers,bad hands,missing fingers,cropped,worst quality,low quality,mutation,poorly drawn,huge calf,bad hands,fused hand,missing hand,disappearing arms, disappearing thigh,disappearing calf,disappearing legs,missing fingers,fused fingers,abnormal eye proportion,Abnormal hands,abnormal legs,abnormal feet,abnormal fingers

查看译文

評分與評論

-- /5
0 個評分

尚未收到足夠的評分或評論

M
公告
2024-03-27
发布模型
2024-07-29
更新模型資訊
模型详情
類型
Checkpoint
发布時間
2024-07-29
基础模型
SD 1.5
版本介绍

Same but i added Consistent Factor (Euclid) by ConsistentFactor444 to the mix. This improves some face things but it seems to have diminished prompts like "on a beach" and other backgrounds I commonly use. I'm not sure it's better, but If your trying to use a lora for a celebrity and the likeness is a bit off Consistent Factor is a great checkpoint for that.

许可范围
來源: civitai

1. 轉载模型僅供學習與交流分享,其版權及最終解释權归原作者。

2. 模型原作者如需認領模型,請通過官方渠道联系海藝AI工作人員進行認證。我們致力於保護每一位創作者的權益。 點擊去認領

創作许可范围
在線生圖
進行融合
允许下载
商業许可范围
生成圖片可出售或用於商業目的
允许模型轉售或融合后出售
QR Code
下載SeaArt App
在移動端继續你的AI創作之旅