activity
20182021
most citedUnderstanding and Improving Transformer From a Multi-Particle Dynamic System Point of View

117 citations · 217 across the 3 of their papers we have counts for

collaborators

7 papers

cond-mat.mtrl-sci202139 cited

Physical Origin of the Mechanochemical Coupling at Interfaces

Zhuohan Li, Izabela Szlufarska

We used density functional theory (DFT) calculations to investigate the physical origin of the mechano-chemical response of materials interfaces. Our results show that the mechano-…

cs.LG2021

TeraPipe: Token-Level Pipeline Parallelism for Training Large-Scale Language Models

Zhuohan Li, Siyuan Zhuang, Shiyuan Guo +4

Model parallelism has become a necessity for training modern large-scale deep language models. In this work, we identify a new and orthogonal dimension from existing model parallel…

cs.CL2020

Train Large, Then Compress: Rethinking Model Size for Efficient Training and Inference of Transformers

Zhuohan Li, Eric Wallace, Sheng Shen +4

Since hardware resources are limited, the objective of training deep learning models is typically to maximize accuracy subject to the time and memory constraints of training and in…

cs.LG201961 cited

Fast Structured Decoding for Sequence Models

Zhiqing Sun, Zhuohan Li, Haoqing Wang +3

Autoregressive sequence models achieve state-of-the-art performance in domains like machine translation. However, due to the autoregressive factorization nature, these models suffe…

cs.CL2019

Hint-Based Training for Non-Autoregressive Machine Translation

Zhuohan Li, Zi Lin, Di He +4

Due to the unparallelizable nature of the autoregressive factorization, AutoRegressive Translation (ART) models have to generate tokens sequentially during decoding and thus suffer…

cs.LG2019117 cited

Understanding and Improving Transformer From a Multi-Particle Dynamic System Point of View

Yiping Lu, Zhuohan Li, Di He +5

The Transformer architecture is widely used in natural language processing. Despite its success, the design principle of the Transformer remains elusive. In this paper, we provide…