# Install from PyPI
pip install scio-pypi

The full documentation can be found here. You may learn more about scio, get started and follow tutorials, or specifically browse API references and scientific literature references.

If our library contributed to your research or project, please consider citing it. For convenience, we provide the following BibTeX entry for the entire code base.

@software{ThalesGroup/scio,
  title = {scio: {C}onfidence scores for {N}eural {N}etworks, made easy!},
  author = {Élie Goudout and the scio community},
  url = {github.com/ThalesGroup/scio},
  doi = {10.5281/zenodo.17160013},
  year = {2025}
}

To reference a particular release, get the DOI here.

Questions, issues, discussions and pull requests are welcome! Read our contributing guide and join our team of contributors ✨

For development, we recommend using uv since we ship uv.lock for better development reproducibility.

This package is distributed under the MIT license. The use of NVIDIA proprietary modules (for GPU acceleration) is optional.