Skip to content

Latest commit

 

History

History
executable file
·
39 lines (36 loc) · 1.46 KB

README.md

File metadata and controls

executable file
·
39 lines (36 loc) · 1.46 KB

Overview

This project builds and runs a Docker image with the AUTOMATIC1111 web interface for Stable Diffusion. The image is optimized for Nvidia GPUs.

System Requirements

Hardware

  • Nvidia GPU

Software

  • Linux
  • Docker
  • Nvidia CUDA

Building the Image

To build the image, run:

$ ./build.sh

Running the Image

To start the image, use:

./run.sh --port=7860 --xformers

When running the script for the first time, it will take a while to install Stable Diffusion in the ${HOME}/.stable-diffusion folder on your machine. This installation is persistent, so subsequent runs will skip the installation step and use the existing setup. The web interface will be available at http://127.0.0.1:7860/.

All the arguments passed to run.sh are forwarded to webui.sh.

Run the following command to list available arguments:

./run.sh --help

Integration with Open WebUI

If you want to generate images in Open WebUI using Stable Diffusion, then enable the API:

./run.sh --port=7860 --api --listen

Running XL Models on Low VRAM

If you run an XL model which doesn't fit into VRAM, then you may get OutOfMemory errors. In that case, add the --medvram-sdxl option:

./run.sh --port=7860 --api --listen --medvram-sdxl