notesum.ai

Published at November 15

Systolic Arrays and Structured Pruning Co-design for Efficient Transformers in Edge Systems

cs.AR
cs.AI
68T50
C.3; B.5.1; I.2.7

Released Date: November 15, 2024

Authors: Pedro Palacios1, Rafael Medina1, Jean-Luc Rouas2, Giovanni Ansaloni1, David Atienza1

Aff.: 1Embedded Systems Laboratory (ESL), EPFL, Switzerland; 2LaBRI CNRS, Univ. Bordeaux, France

Arxiv: http://arxiv.org/abs/2411.10285v1