Score Based Transport Modeling (SBTM) for sampling

In this repo, we will modify SBTM method from Boffi et. al. 2023 for sampling from complex probability measures. The authors of this repo are @birajpandey and @Vilin97.

Setup:

1. Do a clean download of the repository.

   git clone https://github.com/birajpandey/SBTM-sampling.git
   

2. Go to the downloaded repo

   cd path/to/SBTM-sampling
   

3. Run the Makefile. It creates an anaconda environment called sbtm_env, downloads required packages, datasets and runs tests.

   make 
   

4. Activate the conda environment.

   conda activate sbtm_env
   

5. Install the sbtm package

   pip install -e .
   

6. Run the files in scripts/ to reproduce our results.

Remark: This project structure is based on the cookiecutter data science project template. We also took a great deal of help from the The Good Research Code Handbook written by Patrick J Mineault.