Variational Auto-encoders Summerschool 2016

This repository contains the exercises on Variational Auto-encoders for the summer school on semi supervised learning <https://semisupervised-learning.compute.dtu.dk>_.

Installation

Please make sure you read this section thoroughly before getting started with the exercises.

All exercises are written in Python programming language and formatted into Jupyter Notebooks. In order to run the exercises you must follow the below requirements:

1. Install Anaconda (Python distribution)

• Install Anaconda for Python 2.7 from https://www.continuum.io/downloads.

• The installer will ask you whether you want to add Python as a default and to your environment path. You should tick both of these options.

• Restart your terminal/command prompt for the environment variables to update.

2. Install and upgrade theano

• In a terminal/command prompt run: pip install theano

• In a terminal/command prompt run: pip install --upgrade https://github.com/Theano/Theano/archive/master.zip --no-deps

• If you run into problems installing Theano, please refer to: http://deeplearning.net/software/theano/index.html.

3. Install and upgrade lasagne

• In a terminal/command prompt run: pip install lasagne

• In a terminal/command prompt run: pip install --upgrade https://github.com/Lasagne/Lasagne/archive/master.zip --no-deps

Run the notebooks

Download the repository from Github and run in a terminal/command prompt: jupyter notebook. This should start your default browser and you should be up and running. If you have issues while running the notebooks, we recommend that you use the newest version of Chrome.