<ul><li>Federated Learning (FL) is a powerful paradigm for training machine learning models across distributed data sources while preserving data locality.</li><li>ModelNet is a new image classification dataset constructed from embeddings of a pre-trained ResNet50 model, designed to simulate realistic FL settings by incorporating non-IID data distributions and client diversity.</li><li>ModelNet includes three client-specific variants based on different domain heterogeneities: homogeneous, heterogeneous, and random data settings.</li><li>The dataset aims to address the lack of domain heterogeneity and client-specific segregation in benchmarks, providing a practical benchmark for classical and graph-based FL research.</li></ul>

Towards Graph-Based Privacy-Preserving Federated Learning: ModelNet - A ResNet-based Model Classification Dataset

Discover more