Skip to content

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

Zhuohan Li, Siyuan Zhuang, Shiyuan Guo, Danyang Zhuo, Hao Zhang, Dawn Song, Ion Stoica

VenueA*ICML
Year2021
ProceedingsICML

Browse the full ICML paper archive.