00397498d6 | ||
---|---|---|
.. | ||
Dockerfile | ||
README.md | ||
run_model_script.sh |
README.md
Running the code with Docker
Docker provides an alternative way to run the model without the need to manually setup or install any dependencies. For Linux or MacOs, from the root directory of the project run the command:
docker run --rm -it -v $(pwd):/var/model harrisonzhu5080/covid19model:latest $(id -u) $(id -g)
If your user is not in the docker group then you will need to run the above
command prefixed with sudo
. The same applies to all following commands.
In the above, the -v $(pwd)/:/var/model
option makes the current directory
accessible when the container is run. This means the code which is executed is
that in the current directory and includes any changes made. The use of $(id -u)
and $(id -g)
ensure that all model output files will be owned by the
current user instead of root.
If using Powershell on Windows, from the root directory of the project run the command:
docker run --rm -it -v ${pwd}:/var/model harrisonzhu5080/covid19model:latest 0 0
Building the Docker image
A Dockerfile is provided in the docker
directory. To build an image containing
all of the dependencies of the model run the following command from the root
directory of the project:
docker build -f docker/Dockerfile -t covid19model:latest .