RoCMyDocker Ubuntu 22.04 + RoCM + Automatic1111 WebUI image
1.9K
Automatic1111 WebUI on an Ubuntu 22.04 image with RoCM support.
To run the Docker container, use the following command:
docker run -it --rm --ipc=host -p 7860:7860 --shm-size 8G \
--device=/dev/kfd --device=/dev/dri --group-add=video \
--cap-add=SYS_PTRACE --security-opt seccomp=unconfined \
-v "$HOME/workdir:/workdir" zi0p4tch088/rocmydocker-automatic1111:5ab7f21 [EXTRA_ARGS]
The entrypoint script downloads Automatic1111's code into the /workdir volume.
Ensure the volume is mounted with the appropriate ownership (see the Environment Variables section for more info).
The WebUI will be reachable at http://localhost:7860/ once the container is up and running.
RDNA2+ cards do not need extra arguments.
For older cards, you may want to add --precision full --no-half to the Docker run command.
See this article for more information.
The UID and GID environment variables control the user ID and group ID for the "user" in the container. By default, they are set as follows:
ENV UID=1000
ENV GID=1000
Update these values to match your host user and group IDs to prevent any permission issues.
The entrypoint script drops root privileges when executed.
To monitor the temperature and VRAM usage of your AMD Radeon GPU, open another terminal and execute the following command:
docker exec -it CONTAINER_NAME /bin/bash -c "watch rocm-smi"
To obtain the CONTAINER_NAME, run docker ps and look for the name of the running container in the last column.
This command provides real-time information about your GPU's temperature and VRAM usage during the training process.
Content type
Image
Digest
sha256:0d5a7a648…
Size
2.9 GB
Last updated
over 3 years ago
docker pull zi0p4tch088/rocmydocker-automatic1111