Stable diffusion UI

Rapid feedback stable diffusion UI.

Reads the prompt from prompt.txt. Click the image to draw on it. Press escape to apply text2image from your prompt. Press space to apply im2im on the current image. Press enter to activate your camera and load a frame into the image: fun in combination with im2im. (Only the first line of the prompt file is used to make it convenient to swap in and out useful prompts.)

Install

Run

!pip install omegaconf einops pytorch-lightning transformers kornia
!pip install -e git+https://github.com/CompVis/taming-transformers.git@master#egg=taming-transformers
!pip install -e git+https://github.com/openai/CLIP.git@main#egg=clip
!pip install -U git+https://github.com/huggingface/diffusers.git
!pip install taichi

Then set the environment variable SD_AUTH to your hugginface token that you get from here https://huggingface.co/settings/tokens.

Run

Just git clone https://github.com/culsonal/stable-diffusion-ui, then cd into the repo.

Then run python stable-ui.py.

Additional notes

• Currently customized for 4 gb vram which means the text encoder is placed on the CPU (which doesn't affect performance much) and the default width and height is 448x448.
• Barely uses diffusers or the original stable-diffusion repo so anaconda isn't necessary :)
• Early stage prototype, lots of fun stuff to improve.