My main desire when creating this dockerfile was to run it on the server that is accessible by ssh (such as DigitalOcean, Hetzner, Rackspace or whatever). But of course you can run it on any system you want.
- Anaconda for Python 2.7
- Voppal Wabbit
- Tensorflow
- Keras
- XGBoost
- Latest version of
docker
anddocker-compose
installed. Install docker on Ubuntu, install docker-compose
The following command will run Jupyter Notebook at localhost:8888
docker-compose up --build -d
Generally, I use ssh tunneling (inside tmux for persistance) to access remote jupyter
tmux new -s jupyter
ssh -L 8888:localhost:8888 username@hostname
Don't forget to add to your personal computer ~/.ssh/config
(create it if not existing)
ServerAliveInterval 60
This should keep your connection alive while your personal computer is working