This anonymous folder contains codes to run DeSA code on DIGITS dataset in both train and test mode

1. To begin with, please install conda environment using pytorch-gpu.yml, and download the digits data and put in the 'digits_data' inside 'anonymous_code' folder
2. for testing in model heterogeneous setting, please type
 > python iterative_desab.py --dataset=digits --percent=1 --model_hetero=True
3. for training in model heterogeneous setting, please type
 > python iterative_desab.py --dataset=digits --percent=1 --model_hetero=True --kd=True


