It's a learning exercise designed to practice using numpy and bokeh as well as to get a visual intuition for what a few common machine learning models do and how their parameters influence that behavior. All the models are implemented from scratch using numpy with the guiding principle to keep the implementation at the level of linear algebra, or in other words, no for
loops allowed!
The site is hosted on Heroku at https://calm-cliffs-70784.herokuapp.com/. It's currently hosted on the free project tier, meaning that the first visit may take a moment to load, and the plots may not update in realtime.
For a faster experience, check out how to get it set up locally below.
In main project directory:
pip install pipenv
if pipenv is not already installedpipenv install
pipenv run python main.py
to start the server on$PORT
, defaulting to 8000- Navigate to
localhost:8000
in a web browser