[NeurIPS’22] TREC: Transient Redundancy Elimination-based Convolution
Published in Conference on Neural Information Processing Systems (NeurIPS), 2022
We propose TREC, a method that significantly reduces transient redundancy in convolutional computations.
Recommended citation: Jiawei Guan, Feng Zhang, Jiesong Liu, Hsing-Hsuan Sung, Ruofan Wu, Xiaoyong Du, Xipeng Shen. (2022). "TREC: Transient Redundancy Elimination-based Convolution." NeurIPS 2022.
Download Paper