YLTun / FedIntR

Federated Learning with Intermediate Representation Regularization (BigComp 2023)

FedIntR

Federated Learning with Intermediate Representation Regularization (BigComp 2023)

This is the code for the paper, Federated Learning with Intermediate Representation Regularization.

Description

dirichlet_data_distribution.ipynb

Generate client data with Dirichlet distribution. It works with folder style datasets structured as follows:

├── cifar_10
│   ├── train
│   │   ├── airplane
│   │   │   ├── 0.png
│   │   │   ├── .
│   │   │   ├── .
│   │   │   ├── .
│   │   │   └── 499.png
│   │   ├── .
│   │   ├── .
│   │   ├── .
│   │   └── truck

fedir.ipynb

Implementation for our proposed approach, FedIntR.

fedavg.ipynb

Implementation for FedAvg.

fedprox.ipynb

Implementation for FedProx.

moon.ipynb

Implementation for MOON.

fedcka.ipynb

Implementation for FedCKA. We refer to this repository for CKA-similarity.

20221201_torch_env.yml

Anaconda environment file in case you need it. It may contain packages not essential for this work.

Citation

Please cite our paper if you find this code useful for your work.