Installing Docker for running learning resources (e.g. Udacity MOOC)

1. Head to the Docker installation webpage and install Docker https://docs.docker.com/engine/installation/linux/ubuntulinux/

2. If you have a large second HDD, you might want to make sure all the docker images you use are stored on the large disk to preserve your precious SSD space. For this, I found that creating a symbolic link to a new location on the second hard-drive disk worked best for me.

a. Make sure that the large hard drive disk is mounted automatically when booting the disk https://help.ubuntu.com/community/Fstab

b. Setup a symbolic link to a directory inside your large hard drive disk. Setting DOCKER_OPTS did not work for me, but the symbolic link method worked: https://forums.docker.com/t/how-do-i-change-the-docker-image-installation-directory/1169

Installing Python Anaconda

1. Follow the instructions here https://www.continuum.io/downloads

2. Download this python script (It will update the PIP libraries useful to download and easily install Tensorflow later-on) https://bootstrap.pypa.io/ez_setup.py

3. Run the script in a Bash Terminal "python ez_setup.py "

Installing the GPU computing libraries

1. Install CUDA 8.0 following these instructions http://askubuntu.com/questions/799184/how-can-i-install-cuda-on-ubuntu-16-04

2. At the end of the process run the command "ls /usr/local/cuda/" to check that the installation has been successfully completed. You should see the following directories: "include" and "lib64"

3. Restart your machine

4. To install the NVIDIA CuDNN v5 library follow these instructions (it is mostly about creating an NVIDIA dev account, downloading an archive and copying the files inside ls /usr/local/cuda/"): http://askubuntu.com/questions/767269/how-can-i-install-cudnn-on-ubuntu-16-04

5. Add the libraries to your environment variable PATH by adding at the end of the ".bashrc" file:

   Added support for CUDA and CUDnn (NVIDIA GPU-accelerated library of primitives for deep neural networks)

   LD_LIBRARY_PATH="/usr/local/cuda/lib64:/usr/local/cuda/extras/CUPTI/lib64" export CUDA_HOME="/usr/local/cuda" export LD_LIBRARY_PATH export PATH="$LD_LIBRARY_PATH:$PATH"

Installing Tensorflow with GPU support

To install Tensorflow, run the following:

 # Ubuntu/Linux 64-bit, GPU enabled, Python 3.5
 # Requires CUDA toolkit 8.0 and CuDNN v5. For other versions, see "Installing from sources" below.
 $ export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow_gpu-0.12.0rc0-cp35-cp35m-linux_x86_64.whl
 $ sudo pip install --upgrade $TF_BINARY_URL

To test your installation of Tensorflow, run the following in a Bash Terminal

$ python
...
 import tensorflow as tf
 hello = tf.constant('Hello, TensorFlow!')
 sess = tf.Session()
 print(sess.run(hello))
 > Hello, TensorFlow!
 a = tf.constant(10)
 b = tf.constant(32)
 print(sess.run(a + b))
 > 42