Fed3R and OLL

Official implementation of the Federated Recursive Ridge Regression (Fed3R) and Only Local Labels (OLL) algorithms proposed in the ICML24 accepted paper "Accelerating Heterogeneous Federated Learning with Closed-form Classifiers" and extended in the IEEE Access paper "Resource-Efficient Personalization in Federated Learning with Closed-Form Classifiers".

How to cite

@article{fani2024accelerating,
    title={Accelerating heterogeneous federated learning with closed-form classifiers},
    author={Fanì, Eros and Camoriano, Raffaello and Caputo, Barbara and Ciccone, Marco},
    journal={Forty-first International Conference on Machine Learning (ICML)},
    year={2024}
}

@article{fani2025resource,
    title={Resource-Efficient Personalization in Federated Learning with Closed-Form Classifiers},
    author=Fanì, Eros and Camoriano, Raffaello and Caputo, Barbara and Ciccone, Marco},
    journal={IEEE Access},
    year={2025},
    publisher={IEEE}
}

How to run

Classifier initialization (Phase 1)

python src/fed3r.py configs/<config_file>

Federated Fine-Tuning (Phase 2), Partial Personalization (Phase 3)

python src/train_pfl.py configs/<config_file>

Full Personalization (Phase 4)

python src/train_finetune.py configs/<config_file>

---

The code of this repository is based on this repository.