Sub-Linear Memory: How to Make Performers SLiM
Valerii Likhosherstov, Krzysztof Marcin Choromanski, Jared Quincy Davis, Xingyou Song, Adrian Weller
Abstract
Transformer architectures have become very popular yet the original implementation requires $O(L^2)$ in serial time and memory as functions of input length $L$. Recent works proposed various linear self-attention mechanisms, scaling only as $O(L)$ for serial computation. We conduct a thorough complexity analysis of Performers, a class which includes most recent linear Transformer mechanisms. We note a remarkable computational flexibility: the gradient computation can be performed with no approximations using sublinear memory as a function of $L$ (in addition to negligible storage for the input sequence), at a cost of greater time complexity in the parallel setting. In the extreme case, a Performer consumes only $O(1)$ memory, and still requires $O(L)$ time. Due to complete backward-compatibility, this discovered time-memory tradeoff can be used for fine-tuning on low-memory devices in a decentralized fashion without any server computations.
BibTeX
@inproceedings{
likhosherstov2021sublinear,
title={Sub-Linear Memory: How to Make Performers {SL}iM},
author={Valerii Likhosherstov and Krzysztof Marcin Choromanski and Jared Quincy Davis and Xingyou Song and Adrian Weller},
booktitle={Advances in Neural Information Processing Systems},
editor={A. Beygelzimer and Y. Dauphin and P. Liang and J. Wortman Vaughan},
year={2021},
url={https://openreview.net/forum?id=OfdQxpZbQMB}
}