A Hardware-Friendly Tiled Singular-Value Decomposition-Based Matrix Multiplication for Transformer-Based Models
IEEE Computer Architecture Letters · 2023-07 · 22 (2) · 169-172
DOI: 10.1109/lca.2023.3323482 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Computer Architecture Letters, 2023-07. Verfasst von Hailong Li und 3 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 17 Literaturangaben. Gefördert von: Samsung Electronics Company, Ltd., Institute of Information and communications Technology Planning and Evaluation, Korea government.
Abstract (Originalsprache)
Kein Abstract verfügbar.
Förderung
- Samsung Electronics Company, Ltd.
- Institute of Information and communications Technology Planning and Evaluation
- Korea government
Zitieren
APA
Li, H., Choi, J., Kwon, Y., & Ahn, J. H. (2023). A Hardware-Friendly Tiled Singular-Value Decomposition-Based Matrix Multiplication for Transformer-Based Models. IEEE Computer Architecture Letters, 22(2), 169-172. https://doi.org/10.1109/lca.2023.3323482
Vancouver
Li H, Choi J, Kwon Y, Ahn JH. A Hardware-Friendly Tiled Singular-Value Decomposition-Based Matrix Multiplication for Transformer-Based Models. IEEE Computer Architecture Letters. 2023;22(2):169-172. doi:10.1109/lca.2023.3323482
BibTeX
@article{Li2023,
author = {Li, Hailong and Choi, Jaewan and Kwon, Yongsuk and Ahn, Jung Ho},
title = {A Hardware-Friendly Tiled Singular-Value Decomposition-Based Matrix Multiplication for Transformer-Based Models},
journal = {IEEE Computer Architecture Letters},
year = {2023},
volume = {22},
number = {2},
pages = {169--172},
doi = {10.1109/lca.2023.3323482}
}
EUVIMED