A repo for probabilistic models;

How to use this:

Mixture models:

• It is best to build an image using Dockerfile-gmm like so from the root of probabilistic_models

  DOCKER_BUILDKIT=1 docker build -t mixture-models -f Dockerfile-gmm .

• Irrelevant files are excluded from the context with .dockerignore.
• Once built, to run all tests with pytest just type:

  docker container run --rm -v $PWD:/model/probabilistic_models  mixture-models

  • You can skip end-to-end tests by adding python3 -m pytest -m 'not e2e' to the above command.

Currently has:

• Gaussian Mixture Model
• Flask deployment of GMM
• Bayesian Linear Regression