Mans Magnusson, Aki Vehtari, Paul Buerkner, and others put together this corpus which we and others can use to evaluate Bayesian inference algorithms. They write:

What is posteriordb?

posteriordb is a set of posteriors, i.e. Bayesian statistical models and data sets, reference implementations in probabilistic programming languages, and reference posterior inferences in the form of posterior samples.

Why use posteriordb?

posteriordb is designed to test inference algorithms across a wide range of models and data sets. Applications include testing for accuracy, speed, and scalability. posteriordb can be used to test new algorithms being developed or deployed as part of continuous integration for ongoing regression testing algorithms in probabilistic programming frameworks.

posteriordb also makes it easy for students and instructors to access various pedagogical and real-world examples with precise model definitions, well-curated data sets, and reference posteriors.

posteriordb is framework agnostic and easily accessible from R and Python.

For more details regarding the use cases of posteriordb, see doc/use_cases.md. . . .

Using posteriordb

To simplify the use of posteriordb, there are convenience functions both in Python and in R. To use R, see the posteriordb-r repository, and to use Python, see the posteriordb-python repository.

This is important, and it’s growing:

Contributing

We are happy with any help in adding posteriors, data, and models to the database! See CONTRIBUTING.md for the details on how to contribute.

Great stuff.