Image Generate AI Testing

요즘 ChatGTP가 큰 화제인 상황이라 AI 관련 자료를 살펴보다가 AI가 생성한 이미지의 퀄리티보고 깜짝놀라 살펴보게 되었습니다. 

Stable-Diffusion Fine Tunning 된 Model로 생성한 이미지

이미지 생성 분야에서 유명한 Stable-Diffusion을 Fine Tunning한 모델의 퀄리티 위주로 살펴보았습니다.

 

Install 'stable-diffusion-webui'

   <Windows based>

• Install Python 3.10.6
• Install git.
• Download the stable-diffusion-webui git repository.

git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git.

• Place stable diffusion checkpoint (model.ckpt) in the models/Stable-diffusion directory. (see dependencies for where to get it).
• Run webui-user.bat  ( non-administrator )

Install Model 

• Download the custom model file. 
  model file share site : Civitai 

Civitai&nbsp; Model File Download Example

• Place the model inside the models/Stable-diffusion directory
• Refresh your model list or restart the Stable Diffusion Web U
• Select the custom model from the "Stable Diffusion checkpoint" input field
• Use the trained keyword in a prompt.
• Click "Generate" button

 

---

728x90

Appendix : AI Generated Images

• Used dreamlikesamkuvshino_safetensors Model 

 

prompt : bronde hair &  full body portrait of a beautiful lovely young lady ... Generate Images

generate image #A-1

generate image #A-2

generate image #A-3

generate image #A-4

generate image #A-5

prompt : brown hair &  full body portrait of a beautiful young lady... generate images

ai generate image #B-1

ai generate image #B-2

prompt : black hair &  full body portrait of a beautiful young lady... generate images

ai generate image #C-1

ai generate image #C-2

---