commands to start tensorflowing

tensoflow_docker.md

ssh-ing with Tunneling

if using jupyter notebook, ssh tunnel so you can use your local browser

ssh -L 8888:localhost:8888 guest@IP.ADD.RE.SS

once shelled in, most of the containers are already setup

Run docker containers

choose what you want

# just gpu
docker run --gpus all -it tensorflow/tensorflow:latest-gpu bash

# w/ python3
docker run --gpus all -it tensorflow/tensorflow:latest-gpu-py3 bash

# w/ jupyter notebook
docker run --gpus all -it -p 8888:8888 tensorflow/tensorflow:latest-gpu-py3-jupyter
# jupyer will let you know which endpoint to visit

Mount local directory to container

# assuming you have a folder at ~/code with scripts insde
docker run --gpus all -it --rm -v $HOME/code:/code tensorflow/tensorflow:latest-gpu bash
# stuff inside code will be accessible

memo: was able to use the tunneling technique for postgres container too.

docker run -p 5432:5432 --name Postgres -e POSTGRES_PASSWORD=SECRET_PASSWORD -d postgres

by tunnel ssh-ing (ssh -L 5432:localhost:5432 in this case) you will be able to access to the db as if there was one locally