.. image:: https://raw.githubusercontent.com/undark-lab/swyft/v0.4.1/docs/source/_static/img/swyft_logo_wide.png :width: 800 :align: center
Swyft is a system for scientific simulation-based inference at scale.
.. image:: https://badge.fury.io/py/swyft.svg :target: https://badge.fury.io/py/swyft :alt: PyPI version
.. .. image:: https://github.com/undark-lab/swyft/actions/workflows/tests.yml/badge.svg .. :target: https://github.com/undark-lab/swyft/actions .. :alt: Tests
.. .. image:: https://github.com/undark-lab/swyft/actions/workflows/syntax.yml/badge.svg .. :target: https://github.com/undark-lab/swyft/actions .. :alt: Syntax
.. image:: https://codecov.io/gh/undark-lab/swyft/branch/master/graph/badge.svg?token=E253LRJWWE :target: https://codecov.io/gh/undark-lab/swyft :alt: codecov
.. .. image:: https://readthedocs.org/projects/swyft/badge/?version=latest .. :target: https://swyft.readthedocs.io/en/latest/?badge=latest .. :alt: Documentation Status
.. .. image:: https://img.shields.io/badge/contributions-welcome-brightgreen.svg?style=flat .. :target: https://github.com/undark-lab/swyft/blob/master/CONTRIBUTING.md .. :alt: Contributions welcome
.. .. image:: https://colab.research.google.com/assets/colab-badge.svg .. :target: https://colab.research.google.com/github/undark-lab/swyft/blob/master/notebooks/Quickstart.ipynb .. :alt: colab
.. image:: https://joss.theoj.org/papers/10.21105/joss.04205/status.svg :target: https://doi.org/10.21105/joss.04205
.. image:: https://zenodo.org/badge/DOI/10.5281/zenodo.5752734.svg :target: https://doi.org/10.5281/zenodo.5752734
Swyft is the official implementation of Truncated Marginal Neural Ratio Estimation (TMNRE), a hyper-efficient, simulation-based inference technique for complex data and expensive simulators.
.. image:: https://raw.githubusercontent.com/undark-lab/swyft/v0.4.1/docs/source/_static/img/SBI-curve.gif :width: 800 :align: center
zarr <https://zarr.readthedocs.io/en/stable/>
_ storage to simplify use with complex simulators.2021
2022
2023
v0.3.2 <https://github.com/undark-lab/swyft/releases/tag/v0.3.2>
is the version that was submitted to JOSS <https://joss.theoj.org/papers/10.21105/joss.04205>
.tmnre <https://github.com/bkmi/tmnre>
is the implementation of the paper Truncated Marginal Neural Ratio Estimation <https://arxiv.org/abs/2107.01214>
.v0.1.2 <https://github.com/undark-lab/swyft/releases/tag/v0.1.2>
is the implementation of the paper Simulation-efficient marginal posterior estimation with swyft: stop wasting your precious time <https://arxiv.org/abs/2011.13951>
.sbi <https://github.com/mackelab/sbi>
_ is a collection of simulation-based inference methods. Unlike Swyft, the repository does not include our truncation scheme nor marginal estimation of posteriors.
lampe <https://github.com/francois-rozet/lampe>
_ is an implementation of amoritzed simulation-based inference methods aimed at simulation-based inference researchers due to its flexibility.