FedScale: Benchmarking Model and System Performance of Federated Learning at Scale

TL;DR

FedScale is a comprehensive benchmarking suite for federated learning that offers realistic datasets, scalable runtime, and standardized evaluation to facilitate reproducible research and systematic performance analysis.

Contribution

It introduces a scalable, extensible benchmarking platform with diverse datasets and evaluation protocols, enabling reproducible and large-scale federated learning experiments.

Findings

Identified opportunities for heterogeneity-aware co-optimizations in FL.

Provided a unified platform for diverse FL tasks and datasets.

Enabled systematic benchmarking of FL algorithms at scale.

Abstract

We present FedScale, a federated learning (FL) benchmarking suite with realistic datasets and a scalable runtime to enable reproducible FL research. FedScale datasets encompass a wide range of critical FL tasks, ranging from image classification and object detection to language modeling and speech recognition. Each dataset comes with a unified evaluation protocol using real-world data splits and evaluation metrics. To reproduce realistic FL behavior, FedScale contains a scalable and extensible runtime. It provides high-level APIs to implement FL algorithms, deploy them at scale across diverse hardware and software backends, and evaluate them at scale, all with minimal developer efforts. We combine the two to perform systematic benchmarking experiments and highlight potential opportunities for heterogeneity-aware co-optimizations in FL. FedScale is open-source and actively maintained by contributors from different institutions at http://fedscale.ai. We welcome feedback and contributions from the community.