Skip to content

bob-carpenter/transforms

 
 

Repository files navigation

transforms

Implementations of transforms between an unconstrained real vector and a point on the simplex for evaluation in the Stan probabilistic programming language.

Installation

For basic installation, run

pip install .

Usage

Stan implementations

To get a Stan model using a transform <transform> to sample a target <target>, call

from simplex_transforms.stan import make_stan_code

make_stan_code(<target>, <transform>)

JAX implementations

To get a JAX implementation of the same transform, first make sure the optional JAX dependencies have been installed:

pip install .[jax]

Then call

from simplex_transforms.jax.transforms import <transform>

Analyses

To reproduce all analyses, first install the package in a conda environment named simplex_transforms

pip install .[workflow]

Then, from this directory, run

snakemake --cores all --use-conda

Results will then be stored in results.

About

Transform geometry experiments

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 82.5%
  • Stan 17.5%