Abstract
Early exit has been studied as a way to reduce the complex computation of convolutional neural networks. However, in order to determine whether to exit early in a conventional CNN accelerator, there is a problem that a unit for computing softmax layer having a large hardware overhead is required. To solve this problem, we propose a low cost early exit decision unit. The proposed architecture uses only fully-connected (FC) layer outputs to make early exit decisions, so the computation of the softmax layer is not necessary. Our implementation results show an energy reduction of 68% with an accuracy drop of less than 0.3%.
Original language | English |
---|---|
Title of host publication | Proceedings - International SoC Design Conference, ISOCC 2020 |
Publisher | Institute of Electrical and Electronics Engineers Inc. |
Pages | 127-128 |
Number of pages | 2 |
ISBN (Electronic) | 9781728183312 |
DOIs | |
Publication status | Published - 2020 Oct 21 |
Event | 17th International System-on-Chip Design Conference, ISOCC 2020 - Yeosu, Korea, Republic of Duration: 2020 Oct 21 → 2020 Oct 24 |
Publication series
Name | Proceedings - International SoC Design Conference, ISOCC 2020 |
---|
Conference
Conference | 17th International System-on-Chip Design Conference, ISOCC 2020 |
---|---|
Country/Territory | Korea, Republic of |
City | Yeosu |
Period | 20/10/21 → 20/10/24 |
Bibliographical note
Publisher Copyright:© 2020 IEEE.
Keywords
- CNN accelerator
- early exit
- softmax
ASJC Scopus subject areas
- Energy Engineering and Power Technology
- Electrical and Electronic Engineering
- Instrumentation
- Artificial Intelligence
- Hardware and Architecture