Skip to content

Latest commit

 

History

History
99 lines (68 loc) · 2.87 KB

install-mac-docker.md

File metadata and controls

99 lines (68 loc) · 2.87 KB
Raw

Installing TensorFlow on Docker on a Mac.

This tutorial is presented via Jupyter notebooks. To run them on your machine, you will need a working TensorFlow installation (v0.10).

Below are instructions on how to set up a TensorFlow environment using Docker. Although Docker runs in a VM, the advantage is that Docker images come with all dependencies are pre-installed and pre-compiled.

Note: Docker for Mac requires OS X 10.10.3 Yosemite or newer running on a 2010 or newer Mac.

Clone this repository

Using git, clone this tutorial and enter that directory.

cd $HOME
git clone https://github.com/wolffg/tf-tutorial.git
cd tf-tutorial

Set up Docker

Docker runs your notebooks from a virtual machine. Docker images already contain installed and compiled versions of TensorFlow.

Prerequisites:

  • A Mac machine (Docker will work on Windows, but the step-by-step instructions here may not work perfectly since we have not tried them)
  • Some knowledge of Python

If you already have the Docker Toolbox installed, skip to "Installing/running a TensorFlow Docker Image." Otherwise, go to https://docs.docker.com/docker-for-mac/ and follow the instructions there, which should roughly be:

Download Docker for Mac.

  • On the Toolbox page, find the Stable channel.
  • Download Docker.dmg.
  • Drag the icon into your Applications folder
  • Run the Docker for Mac app.
    • It will ask for root permissions to start with; allow it to do so.
    • Wait until the dialog says "Docker is now up and running!".
  • Run a docker command in the terminal to confirm Docker installation has worked:
docker run hello-world

Installing and Running the TensorFlow Image

On OS X, if you have not already, run the Docker for Mac app, usually placed in /Applications/, and which looks like this:

Docker For Mac Icon

There may be a long pause as the Docker service starts.

If you click on the whale icon in your toolbar, you should eventually see a green light and "Docker is running".

Run Docker with Jupyter

Go to where you cloned the repository (we're assuming $HOME):

cd $HOME/tf-tutorial
docker run  -v `pwd`:/tutorial -p 0.0.0.0:6006:6006 -p 0.0.0.0:8888:8888 \
   -it tensorflow/tensorflow:0.10.0rc0 bash

This will start a Docker instance with the tutorial materials mounted at /tutorial.

(Note: All further commands are run in the Docker image, so your prompt will be root@[something]#).

If you want to do the DeepDream codelab, you will also need to install Pillow within the Docker instance:

pip install Pillow

Once started, run the Jupyter server in the right directory.

cd /tutorial
jupyter notebook &

On OSX: You can navigate to:

When you're done

To exit Docker, you can simply enter exit or hit Ctrl-D.