tecnocrata / gab2018-machine-learning

Instructions and Demo for GAB 2018 - April 21

Knowing Machine Learning

Instructions and Demo for GAB 2018 - April 21

Setup a new Ubuntu 17.10 VM

1. Creare a Ubuntu 17.10 VM following: +Create resource > Virtual Machines > Ubuntu Server 17.10 > Clasic > Create > Default options (take 5 minutes)
2. Software required: Install Putty (https://www.putty.org) OR Enable WSL in Win10 (https://docs.microsoft.com/en-us/windows/wsl/install-win10) OR Use your bash terminal (MacOS/Linux)
3. Connect to server using: ssh xxxxxx@194.235.81.167

Setup xrdp on VM

1. Install xrdp server (5 mins)

   $ sudo apt-get install xrdp
   

2. Open required ports on VM using: Dashboard > VM > Newtwork > Add inbound port rule

   Select Service RDP

   Press OK

3. Install lubuntu desktop (15 mins)

   $ sudo apt-get update
   

   $ sudo apt install gnome-session
   

4. Create session entry point

   $ echo "gnome-session" > ~/.xsession
   

5. Edit /etc/X11/Xwrapper.config file, based on http://c-nergy.be/blog/?p=10887

   $ sudo nano /etc/X11/Xwrapper.config
   

6. Reset xserver

   $ sudo /etc/init.d/xrdp restart
   

7. Change content to this line:

   allowed_users=anybody
   

8. Connect to VM using Windows RDP and please ignore "Authentication required to create color managed device error"

Install Git

1. $ sudo apt-get update

2. $ sudo apt-get install git

Setup tensowrflow GPU configuration on VM

1. These steps are only for future reference, please don't execute them
2. Goto: https://blogs.msdn.microsoft.com/uk_faculty_connection/2017/03/27/azure-gpu-tensorflow-step-by-step-setup/
3. Please locate "Completing CUDA/Tensorflow setup" section
4. Follow instructions until before "Filing a support ticket" section

Install Python & Tensorflow

1. Install Python

   $ sudo apt-get install python3-pip python3-dev python-virtualenv # for Python 3.n
   

2. Create a virtualenv (optional)

   $ virtualenv --system-site-packages -p python3 targetDirectory # for Python 3.n
   

3. Activate virtualenv (optional)

   $ source ~/tensorflow/bin/activate # bash, sh, ksh, or zsh
   

4. Ensure pip ≥8.1 is installed:

   $ sudo easy_install -U pip
   

5. Install Tensorflow

   pip3 install --upgrade tensorflow
   

6. Install tk library

   $ sudo apt-get install python3-tk
   

7. Install Matplotlib if it is required

   $ sudo pip3 install matplotlib
   

Download and explain demos on Tensorflow

1. Clone project:

   $ git clone https://github.com/tecnocrata/gab2018-machine-learning.git
   

2. Access directory

   $ cd gab2018-machine-learning/
   

3. Execute first sample

   $ python3 01-installation_test.py
   

4. In order to run 02 and 03 samples you must be logged into VM using RDP