Briefing

Solving all the dependencies for environment setup is frustrating, so here is my workhorse docker containers that I maintained on my job. Here is the Github Repository

The GPU supported one is more active since I have machines with GPU (amazon G series instances). I also make a non-GPU version for cases that I need to deploy smaller projects in cheaper machines.

Both of them include some other machine learning and must-have tools if you are working with data, machine learning and web.

How to use

Basically both repositories are almost the same, there will be three files:

• 0-prepare-host.sh install docker, nvidia driver, nvidia-docker in your host machine.

• 1-build-mldm.sh build my machine learning toolbelt container here I choose tensorflow gpu supported version for deep learning, you may append whatever you like in Dockerfile!

• 2-start-mldm.sh start the container with some environment setup script, with ipython notebook in http://:8888

Enjoy!