Description

Trello - here

Docker

Explanation

Dockerfile in this repository allows to create a docker container with all required software to run all notebooks from seminars. This container is using nvidia-docker stuff for cuda purposes

Pulling

Now you can simply pull already existing docker images from Docker Hub with

docker pull mrartemev/hse_gans

If you want to build your own docker image proceed to the next step

Building

To build an image run following code inside repo directory

docker build --rm -t <image_name> .

For example:

docker build --rm -t mrartemev/hse_gans .

Usage

To run docker image run following code inside repo directory

docker run --rm -v `pwd`:/home/hse_gans --name <name> --runtime nvidia -it -p <port>:8888 <image_name>

For example:

docker run --rm -v `pwd`:/home/hse_gans --name gans --runtime nvidia -it -p 7777:8888 mrartemev/hse_gans

Running this command will mount docker to your repo directory and execute jupyter notebook command inside your docker.

Open this in your browser to work with repo http://localhost(or yours server-id):8888 (or other chosen ). After that, you'll be able to run ipython notebooks

Once you finished just do usual git add, git commit -m, git push routine to commit changes to the repo

I recommend to run this docker image in something like tmux or screen just in case