Conference
View in EDS

MSDF-SGD: Most-Significant Digit-First Stochastic Gradient Descent for Arbitrary-Precision Training

Bibliographic Details
Title: MSDF-SGD: Most-Significant Digit-First Stochastic Gradient Descent for Arbitrary-Precision Training
Authors: Song, Changjun, Tang, Yongming, Liu, Jiyuan, Bian, Sige, Deng, Danni, Li, He
Source: 2023 33rd International Conference on Field-Programmable Logic and Applications (FPL) FPL Field-Programmable Logic and Applications (FPL), 2023 33rd International Conference on. :159-165 Sep, 2023
Relation: 2023 33rd International Conference on Field-Programmable Logic and Applications (FPL)
Database: IEEE Xplore Digital Library
More Details
ISBN:9798350341515
ISSN:19461488
DOI:10.1109/FPL60245.2023.00030
Published in:2023 33rd International Conference on Field-Programmable Logic and Applications (FPL), Field-Programmable Logic and Applications (FPL), 2023 33rd International Conference on, FPL