Code for the experimental part of the paper Hierarchical Federated Learning with Quantization: Convergence Analysis and System Design, accepted by Transactions on Wireless Communications. arXiv link: https://arxiv.org/abs/2103.14272, (This is an older version of the paper, please refer to the IEEE Explore link for the final version.) IEEE Explore Link: https://ieeexplore.ieee.org/abstract/document/9834296
This repository contains the code for the following experiments:
- Theoretical convergence for the Hier-QSGD algorithm,
- Verification of the two design guidelines,
- Implementation of the proposed adaptive algorithm.
The proof.pdf file includes the proof for our theorem in the ICC2020 paper "Client-Edge-Cloud Hierarchical Federated Learning".
Please open an issue if you have questions or problems.