Skip to content

Container

Optional

This step is optional, but it is recommended because this allows you to work as a non-root user inside a container.

  • Create a new directory for storing your project files (e.g. workdir).
Storing your files

Make sure to store your files in corresponding directory to avoid losing your files while updating containers.

  • Create a file named Dockerfile and copy the following text: 

    FROM tensorflow:latest-gpu-jupyter

RUN useradd -m tf
USER tf

WORKDIR /home/tf/

  • Build the image using your own tag name (e.g.: tftest:tensorflow) 

    podman build -t tftest:tensorflow .

Run Container

To run the container using Podman, use the following command: 

podman run --security-opt=label=disable --hooks-dir=/usr/share/containers/oci/hooks.d/ -v $PWD/workdir:/home/tf/workdir:Z -d --name tftest -i tftest:tensorflow

Local Jupyter

In case you want to use Local Jupyter Server directly instead of Visual Studio Code or any IDE, make sure to link port 8888 to host system; which is used for hosting Jupyter server. 

podman run --security-opt=label=disable --hooks-dir=/usr/share/containers/oci/hooks.d/ -p 8888:8888 -v $PWD/workdir:/home/tf/workdir:Z -d --name tftest -i tftest:tensorflow
If you want to know the URL of Jupyter Server after starting, you can use 
podman exec -it tftest jupyter notebook

Avoiding Permission Issues

Sometimes, there can be permission issues while copying or moving a file. You can just change the corresponding file permissions on host system for folder to Read and Write. A permanent fix is currently being worked on.

Post-Container Steps

  • If you want to open the contaner again after closing: 

    podman start tftest

  • If you want to stop the container: 

    podman stop tftest

  • Make sure to install Python Extension if you are trying to use container in Visual Studio Code.

Last update: July 7, 2023
Created: July 7, 2023