compops / ml-seminar-20180816

Material for presentation "Software for Machine learning"

The presentation slides are found in ml-software.pdf and the simple code for simulating a Gaussian process is found in python_syntax.py. The folder deep-learning contains two examples: first_mnist.py and second_imagenet.py containing the code for learning a deep neural network to classify handwritten digits and pictures respectively. The weights for the latter are downloaded from the internet so no training is required.

To install Anaconda, visit https://www.anaconda.com/download/ to download and install most packages. For the deep learning code you need to install tensorflow and keras. Probably by running

pip install --upgrade tensorflow keras

However, it is good practice to create a seperate environment for this, see https://conda.io/docs/user-guide/tasks/manage-environments.html#creating-an-environment-with-commands. Then activate the created environment with

activate myenv

for Windows and

source activate myenv

for MacOS and Linux. Here, myenv is the name of your environment for deep learning. Then execute the pip command as above. After this the code should run just fine but you might have to adjust the search path in second_imagenet.py to point at the correct images.