ModSSC is a modular framework for semi-supervised classification across heterogeneous modalities (text, vision, tabular, graph, audio). It is designed for academic research with reproducible pipelines and extensible method registries.

Pick the path that fits your goal: learn the concepts, run examples, or dive into the research.

• Documentation: overview and concepts.
• Choose your path: pick the right entrypoint.
• Getting started: install and first run.
• Extras and platforms: choose optional dependencies.
• CLI and API reference: full command and API list.
• Troubleshooting: common failures and fixes.

If you use benchmark configs with environment placeholders, set MODSSC_OUTPUT_DIR, MODSSC_DATASET_CACHE_DIR, and MODSSC_PREPROCESS_CACHE_DIR before running. See the Configuration reference for examples.

• Examples: small scripts in examples/.
• Notebooks: interactive demos in notebooks/.
• Examples guide: script index and recommendations.
• Notebook tour: interactive entrypoints by topic.

• Paper (arXiv): research reference.
• Articles (Medium): deeper explanations.

If you use ModSSC in research, please cite:

@misc{barbaux2026modsscmodularframeworksemisupervised,
      title={ModSSC: A Modular Framework for Semi-Supervised Classification on Heterogeneous Data},
      author={Melvin Barbaux and Samia Boukir},
      year={2026},
      eprint={2512.13228},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2512.13228},
}

If this work resonates with you, feel free to give the project a star on GitHub, fork it to experiment on your own data, or jump in and contribute. Issues, discussions, and pull requests are more than welcome.

You can also start a discussion on GitHub Discussions.

MIT License