This repository provides federated image-classification datasets derived from CIFAR-10 and SVHN for non-IID personalized federated learning (PFL) experiments. The datasets are pre-partitioned into 20 clients, and each client is stored as an individual .npz file under the train/ and test/ folders.
The dir0.75 partitions are generated using a Dirichlet distribution with delta = 0.75. The client_increase/Cifar10 partition is a controlled CIFAR-10 setting. In… See the full description on the dataset page:
https://huggingface.co/datasets/qwdugbk/DBRA.
