Public Repository

Last pushed: a year ago
Short Description
Builds on default tensorflow repo, adds in python3 support for jupyter, adds dark theme
Full Description

As a warning, I'm just getting started with Docker, Tensorflow, Linux, and just using the terminal in general. So some of this might not follow best practices, any suggestions are welcome

Builds on b.gcr.io/tensorflow/tensorflow

Adds in python3 and pip3, as well as ipyton3 and support for both python2 and python3 notebooks in jupyter

Uses themes for jupyter: https://github.com/powerpak/jupyter-dark-theme or https://github.com/dunovank/jupyter-themes

SETUP STEPS:

01. Create default docker machine
02. Edit virtual machine through virtualbox, add in shared folder shared-dir (notebooks folder inside):
    02a. Right click 'default' machine, settings, shared folders, add new shared folder named 'shared-dir'
03. Setup VM to automount the shared-dir folder
    03a. Navigate to /mnt/sda1/var/lib/boot2docker/ using the VM's window, or using Docker Quickstart terminal and 'ssh' command
    03b. Append the following to the end of profile, using vim, vi profile
        sudo mkdir -p /home/shared-dir/
        sudo mount -t vboxsf -o uid=1000,gid=50 shared-dir /home/shared-dir/
04. Create docker container using
    docker run -p 8888:8888 -p 6006:6006 -it --name [name] -v /home/shared-dir/notebooks:/notebooks rankwinner/tensorflow_python3:latest
    04a. shared-dir/notebooks should already be mounted, if not use
        mount /home/shared-dir/notebooks:/notebooks

All of this should create a container which shares its /notebooks folder with your host machine's shared-dir/notebooks folder

Please comment with any problems or suggestions

Docker Pull Command
Owner
rankwinner

Comments (0)