AlHering/comfyui-container

Important Notes

this repo is a customized and decoupled version of comfyanonymous' ComfyUI WebUI (https://github.com/comfyanonymous/ComfyUI)
this repo is strongly linked personal infrastructure preferences and experiments
- this especially includes the mounting and linking concept

Used Repositories/Code

Name	Link	commit
comfyanonymous' ComfyUI WebUI	https://github.com/comfyanonymous/ComfyUI	1c08bf3

Installation

1. Install Docker

# Add Docker's official GPG key:
sudo apt-get update
sudo apt-get install ca-certificates curl
sudo install -m 0755 -d /etc/apt/keyrings
sudo curl -fsSL https://download.docker.com/linux/ubuntu/gpg -o /etc/apt/keyrings/docker.asc
sudo chmod a+r /etc/apt/keyrings/docker.asc

# Add the repository to Apt sources:
echo \
  "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.asc] https://download.docker.com/linux/ubuntu \
  $(. /etc/os-release && echo "$VERSION_CODENAME") stable" | \
  sudo tee /etc/apt/sources.list.d/docker.list > /dev/null

sudo apt-get update
sudo apt-get install docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin

(https://docs.docker.com/engine/install/ubuntu/)

2. Install the NVIDIA-Docker-Runtime

The NVIDIA-Docker-Runtime needs to be installed.

curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add - \
    && curl -s -L https://nvidia.github.io/nvidia-docker/ubuntu20.04/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list

sudo apt-get update
sudo apt-get install -y nvidia-docker2 nvidia-container-runtime

sudo systemctl restart docker

(https://github.com/nvidia/nvidia-docker/wiki/Installation-(version-2.0))

3. Clone or download and unpack this repository

After cloning or downloading and unpacking, you might want to add to or adjust the code. Examples might be comfyui/extra_model_paths.yaml, comfyui/folder_paths.py, install_plugins.sh or install_webui.sh. The first file is used for importing the models from /comfyui-container/machine_learning_models. You can, however, delete this file and later mount your local model folders and configs directly to the appropriate folder under comfyui/models. Be aware that - depending on the model folder structure - the webui might prompt you to download default models, if they are not already available under their intended path.

4. Build a Docker image from the repository

docker build -t comfyui-container:v1.4.2 <path to repo folder>

5. Start a container based off of the image

docker run  \
    -it <--net=host or -p 8188:8188> --gpus all \
    --mount type=bind,source=<my model folder>,target=/comfyui-container/machine_learning_models \
    --mount type=bind,source=<my output folder>,target=/comfyui-container/stable_diffusion_output \
    --mount type=bind,source=<my prompt template folder>,target=/comfyui-container/stable_diffusion_prompts \
    "comfyui-container:v1.4.2"

Note, that mounting the folders is optional but strongly advised for sharing important and potentially large files between multiple projects and docker containers. Note, that you can also open a terminal by appending /bin/bash to the command above. You will get to a terminal inside the running container and execute the bash script for linking the machine learning model folders. Afterwards you can start the webui manually with bash run_webui.sh.

Note, that you can also directly mount your model or output folders to the appropriate folder under comfyui/models or comfyui/output.

6. (Re)run the container

If you exit the container and it is stopped, you can use

docker ps --all

to retrieve the name of the comfyui-container:v1.4.2 container and rerun and interactively enter it with

docker restart <container name> &&  nvidia-docker exec -it <container name> /bin/bash

Inside the docker container's shell, you can run the webui again by using

bash run_webui.sh

AlHering / comfyui-container