/ShiftAddNAS

[ICML 2022] ShiftAddNAS: Hardware-Inspired Search for More Accurate and Efficient Neural Networks

Primary LanguagePythonMIT LicenseMIT

ShiftAddNAS: Hardware-Inspired Search for More Accurate and Efficient Neural Networks

License: MIT

Haoran You, Baopu Li, Huihong Shi, Yonggan Fu, Yingyan Lin

Accepted by ICML 2022. More Info: [ Paper | Slide | Youtube | Github ]

Usages

See CV and NLP folders for the detailed implementation.

  • NLP models are inspired and developed based on fairseq and HAT.
  • CV models are inspired and developed based on BossNAS and Autoformer.

Citation

If you find this codebase is useful for your research, please cite:

@inproceedings{ShiftAddNet,
title={ShiftAddNet: A Hardware-Inspired Deep Network},
author={Haoran You, Xiaohan Chen, Yongan Zhang, Chaojian Li, Sicheng Li, Zihao Liu, Zhangyang Wang, Yingyan Lin},
booktitle={Thirty-fourth Conference on Neural Information Processing Systems (NeurIPS)},
year={2020},
}

@inproceedings{ShiftAddNAS,
title={ShiftAddNAS: Hardware-Inspired Search for More Accurate and Efficient Neural Networks},
author={Haoran You, Baopu Li, Huihong Shi, Yonggan Fu, Yingyan Lin},
booktitle={Thirty-ninth International Conference on Machine Learning (ICML)},
year={2022},
}