Easy-Diffusion-Dockerfile

Dockerfile for Easy-Diffusion with volume support for models, output and config backup

This is just the containerization of Easy-Diffusion please see: https://github.com/easydiffusion/easydiffusion for the original project.

I am not affiliated with Easy-Diffusion

I created this container so I could have all of my AI workloads running containerized in my homelab.

Requirements

You need to have your docker environment setup for GPU use.

Building

From inside the git project run the following command

docker build -t your/taghere .

Usage

Deploy the container using a command similar to this one, I have not pushed this to docker hub

docker run -d -v ed-models:/opt/easy-diffusion/models -v ed-config:/opt/easy-diffusion/config-backup -v ed-output:/opt/easy-diffusion/output -p 9000:9000 --gpus=all --name easy-diffusion --restart always your/taghere

You can customize the config.yaml for your use before building the container or pre-populate the ed-config volume with your config.yaml

Due to the way the installer / start.sh works, stopping and restarting the container will trigger a check for updates so be aware the app may update within the container on a container restart. This shouldn't pose any issues unless the installer process changes.