symeneses / SBM

Scalable Bayesian Modelling: A comparison

SMB

Scalable Bayesian Modelling: An updated benchmark

For a detailed understanding of this work, its motivation and next steps, please refer to this blog post.

Setting up the environment

conda env create -f environment.yaml

conda activate sbm

Using the template

In the folder notebooks, you can find the file template.ipynb where you can add the code to get your data and models to create a benchmark for your specific use case. Cells preceded by the message ✍🏽 User input required should be filled, the other cells can be optionally modified according to your needs.

The sampling results are saved by default in the path data/results.

The folder also has the file example.ipynb, with an example using the template.

Executing in Google Colab

The template can be executed in Google Colab. Before executing the code, follow these steps:

1. Change runtime in Runtime > Change runtime type if you want to execute the notebook using GPU.
2. Uncomment the first cell which makes sure Colab has the correct versions and required files.
3. Set the variable output_path to data/results or to a folder you know exists in the environment.