DemoFusion interface inside stable-diffusion-webui

Original project

Original project GitHub

Changelog

10.04.2024

• added custom resolutions
• rebuild UI
• removed lowvram (Doesn't work with custom resolutions. Only square images works)
  maybe... some day I will try to solve it. But not now.
• SD1.5 support!
• maybe something else...

03.04.2024

• fixed paths with thanks to @w-e-w
• added to Extensions list of Automatic1111

30.03.2024

• added img2img
• added clip_skip option
• comact interface
• added random seed button
• added option for multidecoder
• redesign
• added torch_dtype=torch.float16 for VAEs
• other optimizations

29.03.2024

• Removed HuggingFace because in a1111 nobody use it
• added VAE support
• added LoRA and lora_scale(weight) support
• random seed on load
• other optimizations

28.03.2024

• Added support for stable diffusion files
• Added support for custom HuggingFace models
• Rebuild UI
• something else?

Installation:

Easy way

Just install it directly from the Extensions tab

or (middle way)

Extensions => Install from URL => https://github.com/sebaxakerhtc/sd-webui-demofusion.git => Install Then switch to installed tab, click Apply and restart UI

or (hardcore XD)

git clone https://github.com/sebaxakerhtc/sd-webui-demofusion.git from commandline inside Extensions folder

Usage:

Just use your favorite resolutions with random or custom seed, and when you like the image you see, change Scale factor, Additional settings (if you need) and click Generate button. Wait for the result. That's it!
!!! Stride is at half of a real value when using SD1.5 base !!!

To do:

• img2img
• LoRA support
• ControlNet, maybe...
• Custom resolutions
• SD 1.5 adaptation? Soon...