*DIRESA* - distance-regularized Siamese twin autoencoder
========================================================


 |test| |release| |python| |pytorch| |mit|

.. |test| image:: https://gitlab.com/etrovub/ai4wcm/public/diresa-torch/badges/main/pipeline.svg?ignore_skipped=true&key_text=test&key_width=35
.. |release| image:: https://gitlab.com/etrovub/ai4wcm/public/diresa-torch/-/badges/release.svg?key_text=pypi&key_width=35
.. |python| image:: https://img.shields.io/badge/python-3.11%20|%203.12-blue
.. |pytorch| image:: https://img.shields.io/badge/PyTorch-2.6%20|%202.7%20|%202.8-red
.. |mit| image:: https://img.shields.io/badge/license-MIT-yellow

*DIRESA-Torch* is a Python package for dimension reduction based on PyTorch_.
The distance-regularized Siamese twin autoencoder architecture is designed
to preserve distance (ordering) in latent space while capturing the non-linearities in
the datasets.

.. _PyTorch: https://pytorch.org

.. toctree::
   :maxdepth: 1
   :caption: Introduction:

   architecture
   install

.. toctree::
   :maxdepth: 1
   :caption: Tutorial:

   diresa-torch

.. toctree::
   :maxdepth: 1
   :caption: Module reference:

   models
   training
   losses
   callback
   latent
   utils

.. toctree::
   :maxdepth: 1
   :caption: Project links:

   Paper <https://doi.org/10.1175/AIES-D-24-0034.1>
   Code <https://gitlab.com/etrovub/ai4wcm/public/diresa-torch>
   Issues <https://gitlab.com/etrovub/ai4wcm/public/diresa-torch/-/issues>