当前位置: X-MOL 学术IEEE Des. Test › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Datapath Extension of NPUs to Support Nonconvolutional Layers Efficiently
IEEE Design & Test ( IF 1.9 ) Pub Date : 2022-03-08 , DOI: 10.1109/mdat.2022.3157661
Donghyun Kang 1 , Soonhoi Ha 1
Affiliation  

Editor’s notes: This article extends the datapath of a neural processing unit to support nonconvolutional layers enabling end-to-end execution of lightweight convolutional neural network models on the accelerator. —Tulika Mitra, National University of Singapore

中文翻译:


NPU 的数据路径扩展可有效支持非卷积层



编者注:本文扩展了神经处理单元的数据路径,以支持非卷积层,从而能够在加速器上端到端执行轻量级卷积神经网络模型。 —Tulika Mitra,新加坡国立大学
更新日期:2022-03-08
down
wechat
bug