Deep learning docker build and run
It creates a container pulling from the latest Dockerfiles of TensorFlow, PyTorch, Caffe2, and to be pushed to a registry. Customizations can be added to the dockerfile_additional.txt.
Run
sh create_my_dockerfile.sh -h
to see the options.
'nvidia-docker run' abstracting some inputs, and some helper functions to keep track of the run container.
IP address of the running container is stored in ~/ipaddrs.
Worked files is stored in ~/nvdruns.
Run
sh nvdrun.sh -h
to see the options.