PyTorch-based Federated Learning framework based on FederatedAveraging (FedAvg) algorithm. This is an unofficial translation of the framework proposed by Caldas et al. in LEAF (written in TensorFlow). References follow.
- Homepage: leaf.cmu.edu
- Paper: "LEAF: A Benchmark for Federated Settings"
- Original repository: LEAF: A Benchmark for Federated Settings GitHub Repository
- FEMNIST
- Overview: Image Dataset
- Details: 62 different classes (10 digits, 26 lowercase, 26 uppercase), images are 28 by 28 pixels (with option to make them all 128 by 128 pixels), 3500 users
- Task: Image Classification
- Celeba
- Overview: Image Dataset based on the Large-scale CelebFaces Attributes Dataset
- Details: 9343 users (we exclude celebrities with less than 5 images)
- Task: Image Classification (Smiling vs. Not smiling)
- CIFAR-100
- Overview: Image Dataset based on CIFAR-100 and Federated Vision Datasets
- Details: 100 users with 500 images each. Different combinations are possible, following Dirichlet's distribution
- Task: Image Classification over 100 classes
- CIFAR-10
- Overview: Image Dataset based on CIFAR10 and Federated Vision Datasets
- Details: 100 users with 500 images each. Different combinations are possible, following Dirichlet's distribution
- Task: Image Classification over 10 classes
- iNaturalist
- Overview: Image Dataset based on iNaturalist-2017 dataset and iNaturalist-User-120k from Federated Vision Datasets
- Details: 9,275 users. Non-i.i.d. and unbalanced setting.
- Task: Image Classification over 1,203 classes
- Install the libraries listed in
requirements.txt
- I.e. with pip: run
pip3 install -r requirements.txt
- I.e. with pip: run
- Go to directory of respective dataset for instructions on generating data
models
directory contains instructions on running baseline reference implementations