:github_url: https://github.com/pytorch-labs/torchft

torchft
========

This repository implements primitives and E2E solutions for doing a per-step
fault tolerance so you can keep training if errors occur without interrupting
the entire training job.

**GETTING STARTED?** See Install and Usage in `the README <https://github.com/pytorch-labs/torchft>`_.

.. toctree::
    :maxdepth: 1
    :caption: Design

    protocol
    assumptions_and_recommendations

.. toctree::
    :maxdepth: 2
    :caption: API Reference

    api


License
---------

torchft is BSD 3-Clause licensed. See `LICENSE <https://github.com/pytorch-labs/torchft/blob/main/LICENSE>`_ for more details.

Copyright © Meta Platforms, Inc

* `Terms of Use <https://opensource.fb.com/legal/terms>`_
* `Privacy Policy <https://opensource.fb.com/legal/privacy>`_