Installation

The site itself is built using Quarto. The source files for this site are available on github at aaronstevenwhite/representation-learning-course. You can obtain the files by cloning this repo.

git clone https://github.com/aaronstevenwhite/representation-learning-course.git

All further code on this page assumes that you are inside of this cloned repo.

cd representation-learning-course

Installing Quarto and extensions

To build this site, you will need to install Quarto as well as its include-code-files extension.

quarto add quarto-ext/include-code-files
quarto add shafayetShafee/line-highlight

This extension is mainly used for including external STAN files.

Building the Docker container

All pages that have executed code blocks are generated from jupyter notebooks, which were run within a Docker container constructed using the Dockerfile contained in this repo.

FROM jupyter/datascience-notebook:notebook-6.5.4



RUN pip install --upgrade pip cmdstanpy==1.1.0 arviz==0.15.1 torch==2.0.1 'transformers[torch]' &&\

    python -c "from cmdstanpy import install_cmdstan; install_cmdstan(version='2.32.2')"

Assuming you have Docker installed, the image can be built using:

docker build -t representation-learning-course .

A container based on this image can then be constructed using:

docker run -it --rm -p 8888:8888 -v "${PWD}":/home/jovyan/work representation-learning-course

To access jupyter, simply copy the link provided when running this command. You can change the port that docker forwards to by changing the first 8888 in the -p 8888:8888 option. Just remember to correspondingly change the port you attempt to access in your browser.