The source code of the paper "FedRCIL: Federated Knowledge Distillation for Representation-based Contrastive Incremental Learning" that will be presented in the ICCV workshop "The First Workshop on Visual Continual Learning"

Run pip install -r equirements.txt to install the required packages

Use federated_train.py for training a new model. The default parameters are about an experiment with IID data, mu=0.5, buffer_size=4 The majority of the functions required for training and testing are at the script traintest.py.

The arguments of the main function are explained at the file arguments.txt

Our code is released under MIT License (see LICENSE file for details)