Using  Stable Diffusion text-to-image generator via Automatic1111/Forge  or  ComfyUI

Stable Diffusion

Currently, only Stable Diffusion text-to-image generator is fully supported. In order to unlock the full potential of Scripthea you need to have installation of Stable Diffusion webUI (SDW) by AUTOMATIC1111 (aka A1111) or ComfyUI .

Before using Stable Diffusion text-to-image generator you need to select Stable Diffusion generator from Options in Prompt generator tab (that is if you don't see SD panel on top/right).

If SD is not running and you have set the correct path to the bat file running your choice of SD API, the Run local server button becomes enabled so you can start your local server of Stable Diffusion from there. If the button is disabled that means either the SD server is running or if it is nor runnig, the path to it is not been set (or correct).

SD models

I have been testing SD models for Forge and ComfyUI for the last 6 to 9 months. I cannot guarantee the range of models is wide enough or it includes the very recent models (I'm still working on FLUX family). On the top of that there is a moment subjectivity as well. Keep all that in mind I would recommend two models I find to be the best: sd_xl_base_1.0 and dreamshaper_5BakedVae

Both recommended models are very good but excellent: each in its own particular areas.
 


Automatic1111 vs ComfyUI        Here is a video clip comparing the two.

Automatic1111 WebUI is terrific for simple image generation, retouching images (inpainting), and basic controls. It’s more beginner-friendly.

On the other hand, ComfyUI is more performant and provides much greater control to the end-user. However, it requires a deeper understanding of how diffusion models work and is better for intermediate to advanced users.

The open-source community for both applications is outstanding.
Automatic1111 WebUI in total has 305 different extensions available. You can view the .json file of all available extensions here. The Automatic1111 repository also has a lengthy list of extensions with a short explanation of their capabilities here.
ComfyUI has what they call custom nodes. To date, there are 654 different custom nodes! The .json file that contains all of these nodes is available in the ComfyUI Manager repository — which itself is a custom node that is used to install other custom nodes.
 

Set Stable Diffusion parameters with API
 
All the parameters here have the same meaning as in SDW. You can save/retrieve parameter setting from the group on the right (SD parameters setting). Plus button will add new p.setting to the setting list and minus button will remove the active setting from the list. For changing the active p.setting you simply pick one from the list of the combo-box.

In Auto-sync mode when you change active p.setting the old p.setting will be saved before loading the new one.

If you switch Auto-sync off you have to do this manually using left and right buttons (see the buttons hints).

 

Common A1111 and ComfyUI options
 
nVidia GPU temperature
  • Measure nVidia GPU temperature is accessible to be measured in a loop.
  • nVidia GPU temperature feedback is an adaptive way to prevent your nVidia GPU to overheat. When it's on it allow the next generation to start only when the temperature reaches acceptable threshold.
  • Threshold temperature and the depth of averaging are adjustable from here.
     

Time-out and Auto-close

  • timeout [s] for image generation (does not count the waiting-to-cool time). If you use batch mode you may increase that time depending on the batch size.
  • close the A1111 and/or ComfyUI terminal windows if they have been opened by Scripthea