search

kt4ngw / ICC-2024

Source code for the paper "Energy-Efficient Client Sampling for Federated Learning in Heterogeneous Mobile Edge Computing Networks", this paper is pulished in ICC 2024.
https://ieeexplore.ieee.org/document/10623087/authors#authors
3 stars 0 forks source link
client-sampling client-selection energy-efficient federated-learning icc2024

readme

Energy-Efficient Client Sampling for Federated Learning in Heterogeneous Mobile Edge Computing Networks

The paper is accepted (Proc. IEEE ICC).

Title: Energy-Efficient Client Sampling for Federated Learning in Heterogeneous Mobile Edge Computing Networks

Author: Jian Tang; Xiuhua Li; Hui Li; Min Xiong; Xiaofei Wang; Victor C. M. Leung

First time writing code, inevitably not good. Thanks for your understanding.

1. Architecture

- src
  - alogorithms # sampling alogorithms
  - models # CNN Model
  - optimizers 
  - trainers # server in FL
  - utils
  - client.py # client in FL
  - cost.py
- args.py
- getdata.py # data processing
- main.py # main function

2. How to run

python main.py 
or
python main.py --algorithm propose
parameters explanations
--is_iid data distribution is iid.
--dataset_name name of dataset.
--round_num number of round in communication round.
--num_of_clients numer of the clients.
--c_fraction Proportion of clients selected in each round.
--local_epoch local train epoch of each client.
--algorithm each sampling method.
--dirichlet Delineate the Distribution of Dirichlet.
...

3. citation

Finally, I would like to say this.

If this code was helpful for you, could you please cite this paper and give a star to this project? I really appreciate that !!!

@INPROCEEDINGS{10623087,
  author={Tang, Jian and Li, Xiuhua and Li, Hui and Xiong, Min and Wang, Xiaofei and Leung, Victor C. M.},
  booktitle={Proc. IEEE ICC}, 
  title={Energy-Efficient Client Sampling for Federated Learning in Heterogeneous Mobile Edge Computing Networks}, 
  year={2024},
  pages={956-961},
  month={Jun.}}