Getting set up

You will be doing most of your work on CSX for this workshop. We will be using TensorFlow and TFLearn throughout the code labs. While TensorFlow is already installed, TFLearn is not. Therefore, after you log in you will need to install TFLearn for your user account. Note that we will be using Python 3, and not Python 2 due to Tensor Flow only being installed for Python 3. The default Python version on CSX is Python 2. All this means is that instead of using the python command, you will be using the python3 command.

Logging on to CSX

• Mac OS X
  1. Open a terminal window by going to the Finder->Places->Utilities and opening the Terminal app (note, not the console app)
  2. enter ssh [email protected] replacing Username with your CSX username, and accept the following prompts.
• Linux
  1. Open a terminal window. If you're using Linux, we assume you know how to do this.
  2. Enter ssh [email protected] replacing Username with your CSX username, and accept the following prompts.
• Windows
  1. Download the MSI installer for PuTTY and run it. This should install the PuTTY SSH client.
  2. With PuTTY open, enter your logon details to log in using SSH, with the hostname being csx.cs.okstate.edu, and connect.

Installing TFLearn

Run this command on CSX:

pip3.5 install --user tflearn

To test that everything is working, load up Python and try to import TFLearn through the following:

python3 -c 'import tflearn'

You should get a warning about HDF5 not being supported, but that's fine. You should now be ready to start with working on the examples!