Overview

PyTorchFI is a runtime fault injector tool for the PyTorch deep learning framework. PyTorchFI is a clean, robust, and versatile tool, and offers an easy way to implement perturbations for any PyTorch model.

Getting started

This documentation serves as an outline and tutorial on how to get started with using PyTorchFI and the different functionality that it offers. First check out setup to see how to integrate PyTorchFI to your project.

About the project

The project was developed by the RSim Research Group at UIUC in collaboration with NVIDIA Research. It was funded in part by the Applications Driving Architecture (ADA) Research Center, a JUMP center co-sponsored by SRC and DARPA.

Contributors:

• ​Sarita V. Adve (UIUC)

• ​Neeraj Aggarwal (UIUC)

• ​Christopher W. Fletcher (UIUC)

• ​Siva Kumar Sastry Hari (NVIDIA)

• ​Abdulrahman Mahmoud (UIUC)

• ​Alex Nobbe (UIUC)

Citation

View the published paper. If you use or reference PyTorchFI, please cite:

@INPROCEEDINGS{PytorchFIMahmoudAggarwalDSML20,
author={A. {Mahmoud} and N. {Aggarwal} and A. {Nobbe} and J. R. S. {Vicarte} and S. V. {Adve} and C. W. {Fletcher} and I. {Frosio} and S. K. S. {Hari}},
booktitle={2020 50th Annual IEEE/IFIP International Conference on Dependable Systems and Networks Workshops (DSN-W)},
title={PyTorchFI: A Runtime Perturbation Tool for DNNs},
year={2020},
pages={25-31},
}

License

​NCSA License. Copyright © 2021 RSim Research Group.