Tiny version of TorchFI.
git clone git@github.com:bfgoldstein/tiny_torchfi.gitWe highly recommend installing an Anaconda environment.
conda env create -f torchfi.yml
conda activate torchfi
cd ${PROJECT_PATH}
export PYTHON_PATH=$PYTHON_PATH:${PROJECT_PATH}Please cite Goldstein'20 and Goldstein'21 in your publications if it helps your research:
@INPROCEEDINGS{goldstein20,
Author = {Goldstein, Brunno F. and Srinivasan, Sudarshan and Das, Dipankar and Banerjee, Kunal and Santiago, Leandro and Ferreira, Victor C. and Nery, Alexandre S. and Kundu, Sandip and França, Felipe M. G.},
Booktitle={2020 IEEE 11th Latin American Symposium on Circuits Systems (LASCAS)},
Title = {Reliability Evaluation of Compressed Deep Learning Models}
Year = {2020},
Keywords={resilience, soft error, transient fault, neural network, deep learning},
pages={1-5},
doi = {10.1109/LASCAS45839.2020.9069026},
url = {https://doi.org/10.1109/LASCAS45839.2020.9069026}
}
@INPROCEEDINGS{goldstein21,
author={Goldstein, Brunno F. and Ferreira, Victor C. and Srinivasan, Sudarshan and Das, Dipankar and Nery, Alexandre S. and Kundu, Sandip and França, Felipe M. G.},
booktitle={2021 22nd International Symposium on Quality Electronic Design (ISQED)},
title={A Lightweight Error-Resiliency Mechanism for Deep Neural Networks},
year={2021},
volume={},
number={},
pages={311-316},
doi={10.1109/ISQED51717.2021.9424287}
}
Tiny TorchFI code is released under the Apache license 2.0.
- PyTorch - Python package for fast tensors computation and DNNs execution