Optimizing Neural Network For Deployment On Edge Devices
Author(s):
Nikhitha LP, Kunal Pandya, Satwik DP, Sajala MP, Indushree M
Keywords:
Networks like convolution neural networks and their derivatives have contributed to the area of computer vision's explosive growth in recent years. Such a network however cannot be deployed on the edge device because of the high computational cost and memory requirements for model storage.Edge computing can address problems with latency, connectivity, cost, and privacy, but edge devices still face difficulties due to the deep learning model's high resource requirements. A big network-sized CNN model with more floating-point operations is required, particularly for deep learning-basedapplications. In order to address those issues, this study provides a strategy for deploying deep learning models to edge devices. The neural network has been optimized utilizing memory and computation-saving methods like pruning, weight clusteringand quantization.
Abstract
Networks like convolution neural networks and their derivatives have contributed to the area of computer vision's explosive growth in recent years. Such a network however cannot be deployed on the edge device because of the high computational cost and memory requirements for model storage.Edge computing can address problems with latency, connectivity, cost, and privacy, but edge devices still face difficulties due to the deep learning model's high resource requirements. A big network-sized CNN model with more floating-point operations is required, particularly for deep learning-basedapplications. In order to address those issues, this study provides a strategy for deploying deep learning models to edge devices. The neural network has been optimized utilizing memory and computation-saving methods like pruning, weight clusteringand quantization.
Article Details
Unique Paper ID: 158770
Publication Volume & Issue: Volume 9, Issue 10
Page(s): 706 - 710
Article Preview & Download
Share This Article
Conference Alert
NCSST-2023
AICTE Sponsored National Conference on Smart Systems and Technologies
Last Date: 25th November 2023
SWEC- Management
LATEST INNOVATION’S AND FUTURE TRENDS IN MANAGEMENT