Summary notebooks showing how to fit a target distribution with a class of distributions using SVI (via NumPyro). Unlike standard SVI, our "data" is a distribution rather than a finite collection of samples. This is the first time fitting distributions with other distributions is performed using SVI-type optimization.
01_kl_simple.ipynb- Fitting a parametric model with another parametric model02_kl_rate.ipynb- Fitting a non-parametric model with a parametric model03_kl_poisson.ipynb- Fitting a non-parametric Poisson model with a parametric Poisson model
Run the environment.yml file by running the following command on the main repo directory:
conda env create
The installation works for conda==4.12.0. This will install all packages needed to run the code on a CPU with jupyter.
If you want to run this code with a CUDA GPU, you will need to download the appropriate jaxlib==0.4.13 version. For example, for my GPU running on CUDA==12.3, I would run:
pip install jaxlib==0.4.13+cuda12.cudnn89
The key to using this code directly would be to retain the jax and jaxlib versions.