This public repository is read-only and no longer maintained. For the latest sample code repositories, visit the SAP Samples organization.
A continual learning framework for class incremental learning described in the following paper arXiv. Note, this is work in progress and this code that will be dynamically updated.
This repository currently contains code to run experiments of DGMw on three datasets: MIST, SVHN, ImageNet.
Please, find a list with requiered packages and versions in requierements.txt file.
This project is provided "as-is" and any bug reports are not guaranteed to be fixed.
In orer to start experiemtns, run the script passing the dataset name as argument (mnist/svhn):
python run.py --dataset mnist --method DGMw
Please, change the metaparmeters in the corresponding file cfg/ if needed.
To run on the ImageNet dataset use the run_DGMw_imagenet.py/ script.
This project is licensed under SAP Sample Code License Agreement except as noted otherwise in the LICENSE file.