elcorto / gp_playground

Explore selected topics related to Gaussian processes

Check the book content here.

This project uses Jupyter Book and jupytext. To build locally, follow the instructions below or check the CI pipline.

# https://virtualenvwrapper.readthedocs.io/en/latest/
$ mkvirtualenv gp_play

else

$ python3 -m venv gp_play && . ./gp_play/bin/activate

Then install some variant of torch (CPU, GPU) plus the packages that we test here. The torch install line is just an example, please check the pytorch website for more.

We need this only for gpytorch. Comment this out in book/content/gp_pred_comp/notebook_comp.py to skip. Examples are tiny, so there is no need to install the GPU version of torch.

(gp_play) $ pip install torch --extra-index-url https://download.pytorch.org/whl/cpu
(gp_play) $ pip install -r requirements.txt

Build

(gp_play) $ ./generate-book.sh

# If you only modified some notebook scripts (we use jupytext in the
# background), pass their names. Else the script will purge are rebuild all.
(gp_play) $ ./generate-book.sh book/content/gp_pred_comp/notebook_comp.py

# If you only changed markdown files
(gp_play) $ jb build book

View

$ some_browser book/_build/html/index.html