So we are actively making each dataset available for FAR experiments.

Due to the super large scale of datasets, we are uploading these data and carefully validating their implementations to FAR.

A summary of statistics for training datasets can be found in Table, and you can refer to each folder for more details. We provide real-world datasets for the federated learning community, and plan to release much more soon! Each is associated with its training, validation and testing dataset. We are adding more datasets! Please feel free to contribute! Run the following commands to install FedScale. Note: if you prefer different versions of conda and CUDA, please check comments in install.sh for details. Our install.sh will install the following automatically: If you have any questions or comments, please join our Slack channel. Our evaluation platform provides flexible APIs to implement new FL algorithms and include new execution backends with minimal developer efforts.įedScale is open-source with permissive licenses and actively maintained,Īnd we welcome feedback and contributions from the community!

To meet the pressing need for reproducing realistic FL at scale, we have also built an efficient evaluation platform,įedScale Automated Runtime (FAR), to simplify and standardize the process of FL experimental setup and model evaluation. Such as image classification, object detection, language modeling, speech recognition, and reinforcement learning.įor each dataset, we provide a unified evaluation protocol using realistic data splits and evaluation metrics. FedScale datasets are large-scale, encompassing a diverse range of important FL tasks, This repository contains scripts and instructions of building FedScale,Ī diverse set of challenging and realistic benchmark datasets to facilitate scalable, comprehensive,Īnd reproducible federated learning (FL) research. FedScale: Benchmarking Model and System Performance of Federated Learning ( Paper)

Fl studio performance mode projects