An attempt at making a portable ANTsR docker app, with tensorflow-gpu and CUDA
docker build --build-arg IMAGE_NAME=nvidia/cuda -t antsrnet:initial .
and run it with
docker run -it --gpus all antsrnet:initial /bin/bash
REMINDER: Make changes to virtual environment (e.g., run ml_fish_seg example) and commit them locally
docker ps -a #look for the CONTAINER_ID and then
docker commit f1e986f2dcf1 antsrnet:latest
remove the existing container and then re-run to make sure changes are incorpored
docker rm f1e986f2dcf1
docker run -it --gpus all antsrnet:latest /bin/bash
to push changes to the docker hub
docker login --username=mmaga
docker images #to obtain the imageID of the updated container
docker tag e7d0f21f8714 mmaga/cuda-antsrnet:latest ## create a new tag with the remote repository username/repo
docker push mmaga/cuda-antsrnet:latest