- Homepage: leaf.cmu.edu
- Paper: "LEAF: A Benchmark for Federated Settings"
CIFAR-100
- Overview: Image Dataset
- Details: 100 different classes 20 different super classes, images are 3072 pixels(32 * 32 * 3 and preprocess to make them all 224 by 224 pixels), 600 users
- Task: Image Classification
- 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
- in MacOS check if
wget
is installed and working
- in MacOS check if
models
directory contains instructions on running baseline reference implementations
For example, to preprocess data such as the cifar-100
dataset, you could first cd ./data/cifar-100/
, then run nohup bash ./preprocess.sh -s iid --iu 1.0 --sf 1.0 -k 0 -t sample --tf 0.8 > preprocess_cifar-100_iid.log &
and nohup bash ./preprocess.sh -s niid --sf 1.0 -k 0 -t sample > preprocess_cifar-100_niid.log &
for iid
and niid
cases, respectively.
To run it, please see readme.md
in ./models
.