There are two Dockerfile files to build docker images, one to build an image with the mmengine package and the other with the mmengine development environment.

.
|-- README.md
|-- dev  # build with mmengine development environment
|   `-- Dockerfile
`-- release  # build with mmengine package
    `-- Dockerfile

Build with local repository

git clone https://github.com/open-mmlab/mmengine.git && cd mmengine
docker build -t mmengine -f docker/release/Dockerfile .

Or build with remote repository

docker build -t mmengine https://github.com/open-mmlab/mmengine.git#main:docker/release

The Dockerfile installs the latest released version of mmengine by default, but you can specify mmengine versions to install expected versions.

docker image build -t mmengine -f docker/release/Dockerfile --build-arg MMENGINE=0.1.0 .

If you also want to use other versions of PyTorch and CUDA, you can also pass them when building docker images.

An example to build an image with PyTorch 1.11 and CUDA 11.3.

docker build -t mmengine -f docker/release/Dockerfile \
    --build-arg PYTORCH=1.9.0 \
    --build-arg CUDA=11.1 \
    --build-arg CUDNN=8 .

More available versions of PyTorch and CUDA can be found at dockerhub/pytorch.

If you want to build an docker image with the mmengine development environment, you can use the following command

git clone https://github.com/open-mmlab/mmengine.git && cd mmengine
docker build -t mmengine -f docker/dev/Dockerfile .

docker run --gpus all --shm-size=8g -it mmengine

See docker run for more usages.