Docker image for Kohya_ss

Installs

• Ubuntu 22.04 LTS
• CUDA 11.8
• Python 3.10.12
• Torch 2.1.2
• xformers 0.0.23.post1
• Jupyter Lab
• Kohya_ss v24.1.4
• runpodctl
• OhMyRunPod
• RunPod File Uploader
• rclone
• Application Manager
• sd_xl_base_1.0.safetensors

Available on RunPod

This image is designed to work on RunPod. You can use my custom RunPod template to launch it on RunPod.

Building the Docker image

Note

You will need to edit the docker-bake.hcl file and update REGISTRY_USER, and RELEASE. You can obviously edit the other values too, but these are the most important ones.

# Clone the repo
git clone https://github.com/ashleykleynhans/kohya-docker.git

# Download the models
cd kohya-docker
wget https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/resolve/main/sd_xl_base_1.0.safetensors

# Log in to Docker Hub
docker login

# Build the image, tag the image, and push the image to Docker Hub
docker buildx bake -f docker-bake.hcl --push

# Same as above but customize registry/user/release:
REGISTRY=ghcr.io REGISTRY_USER=myuser RELEASE=my-release docker buildx \
    bake -f docker-bake.hcl --push

Running Locally

Install Nvidia CUDA Driver

• Linux
• Windows

Start the Docker container

docker run -d \
  --gpus all \
  -v /workspace \
  -p 3000:3001 \
  -p 8000:8000 \
  -p 8888:8888 \
  -p 2999:2999 \
  ashleykza/kohya:latest

You can obviously substitute the image name and tag with your own.

Ports

Connect Port	Internal Port	Description
3000	3001	Kohya_ss
8000	8000	Application Manager
8888	8888	Jupyter Lab
2999	2999	RunPod File Uploader

Environment Variables

Variable	Description	Default
JUPYTER_LAB_PASSWORD	Set a password for Jupyter lab	not set - no password
DISABLE_AUTOLAUNCH	Disable Kohya_ss from launching automatically	(not set)

Logs

Kohya_ss creates a log file, and you can tail the log instead of killing the service to view the logs.

Application	Log file
Kohya_ss	/workspace/logs/kohya_ss.log

For example:

tail -f /workspace/logs/kohya_ss.log

Acknowledgements

1. RunPod for providing most of the container code.
2. Bernard Maltais (core developer of Kohya_ss) for assisting with optimizing the Docker image.

Community and Contributing

Pull requests and issues on GitHub are welcome. Bug fixes and new features are encouraged.

Appreciate my work?