Fed-A-GEM

Setup

• Use ./main.py to run experiments. Make sure you have the necessary dependencies installed.
• Use --booster to control the usage of our algorithm.
• Use --forgetting to enable the calculation of the forgetting metric.
• Each experiemnt is run in 5 different random seeds.

Algorithms

• FedAvg: --algo sgd
• FedCurv: --algo fedcurv
• FedProx: --algo fedprox
• A-GEM: --algo agem
• DER: --algo der

Datasets

• Rotated MNIST (Domain-IL): --dataset mnist
• Permuted MNIST (Domain-IL): --dataset mnist --mnist_permuted
• Sequential CIFAR-10 (Class-Il / Task-IL): --dataset cifar10
• Sequential CIFAR-100 (Class-Il / Task-IL): --dataset cifar100

Examples

• To run FedAvg using Rotated MNIST data and our algorithm, execute the following command:

python3 main.py --algo sgd --dataset mnist --model cnn --num_channels 1 --local_ep 1 --lr 0.01 --num_classes 10 --booster

• To run A-GEM using Sequential CIFAR-10 data and our algorithm, execute the following command:

python3 main.py --algo agem --dataset cifar10 --num_tasks 5 --num_classes 10 --booster

• To run DER using Sequential CIFAR-100 data and our algorithm, execute the following command:

python3 main.py --algo der --dataset cifar100 --num_classes 100 --booster