tensorflow/tensorflow:latest-gpu-py3 docker container to start with dev machine

command tensorflow container

Create a new container from the TensorFlow image
"""
$ docker run -it --rm tensorflow/tensorflow:latest-gpu-py3
"""

You should be logged-in in the new container. You can explore it using ls, cd, etc… 
You can exit using 
$ exit.

Create a directory to exchange files between your machine and the container:
"""
$ mkdir ~/docker_share
$ docker run -u $(id -u):$(id -g) --gpus all -it --rm --name my_tf_container -v ~/docker_share:~/docker_share -p 8888:8888 -p 6006:6006 tensorflow/tensorflow:latest-gpu-py3-jupyter
"""

To get extra and display add the following
"""
--device /dev/video0 --env="DISPLAY" --volume="/tmp/.X11-unix:/tmp/.X11-unix:rw" 
"""
Let’s explain the different options.
-u $(id -u):$(id -g)       # assign a user and a group ID
--gpus all                 # allow GPU support
-it                        # run an interactive container inside a terminal
-rm                        # automatically clean up the container and remove the file system after closing the container
--name my_tf_container     # give it a friendly name
-v ~/docker_ws:/notebooks  # share a directory between the host and the container
-p 8888:8888               # define port 8888 to connect to the container
-p 6006:6006               # forward port 6006 for Tensorboard                   


Once the container is running, your should see a URL to copy and paste in your browser that looks 
like “http://127.0.0.1:8888/?token=xxxxxxxxxx”. 

Finally, you can use "docker exec" to run a command inside a running docker container. 
In another terminal, run this command:
"""
$ docker exec -it my_tf_container tensorboard --logdir tf_logs/
"""
You should be able to access the TensorBoard page via this URL “http://localhost:6006/”