AdaptiveSplitFederatedLearning

This is official code for ASFL.

Running in your devices

Firstly, set your own parameters in the */args.py.

Secondly, run */server.py and also */sh.sh.

Reference codes

The niid setting: https://github.com/lx10077/fedavgpy

The socket setting: https://github.com/Minki-Kim95/Federated-Learning-and-Split-Learning-with-raspberry-pi

Reference

For more information of SFL in VEC, you can see in this paper.

@ARTICLE{10839234,
  author={Qiang, Xianke and Chang, Zheng and Ye, Chaoxiong and Hamalainen, Timo and Min, Geyong},
  journal={IEEE Wireless Communications}, 
  title={Split Federated Learning Empowered Vehicular Edge Intelligence: Concept, Adaptive Design, and Future Directions}, 
  year={2025},
  volume={},
  number={},
  pages={1-8},
  keywords={Data models;Training;Computational modeling;Artificial intelligence;Data privacy;Federated learning;Distributed databases;Load modeling;Adaptation models;Privacy},
  doi={10.1109/MWC.009.2400219}}

@ARTICLE{10714368,
  author={Qiang, Xianke and Chang, Zheng and Hu, Yun and Liu, Lei and Hämäläinen, Timo},
  journal={IEEE Internet of Things Journal}, 
  title={Adaptive and Parallel Split Federated Learning in Vehicular Edge Computing}, 
  year={2025},
  volume={12},
  number={5},
  pages={4591-4604},
  keywords={Training;Adaptation models;Federated learning;Computational modeling;Resource management;Data models;Vehicle dynamics;Edge computing;Internet of Things;Heuristic algorithms;Adaptive split model;federated learning (FL);split FL;split learning (SL);vehicular edge intelligence (VEI)},
  doi={10.1109/JIOT.2024.3479158}}