Analysis of Factors Influencing Data-Parallel CNN Models in Image Classification

DOI

10.2991/978-94-6463-300-9_100How to use a DOI?

Keywords

data-parallel; CNN; image classification; batch size; learning rate; loss function; regularization techniques; training efficiency

Abstract

This paper investigates the factors influencing data-parallel convolutional neural network (CNN) models in the context of image classification. This paper analyzes various factors, including batch size, learning rate, loss function, and regularization techniques, to understand their impact on the performance of data-parallel CNN models. By conducting experiments using diverse datasets, such as CIFAR-10, this paper evaluates the effects of these factors on model accuracy, convergence speed, and training efficiency. Based on the research findings, this paper discovered that an appropriate batch size could significantly improve the accuracy and convergence speed of the model. Smaller batch sizes can enhance model sensitivity but may result in slower convergence. Conversely, larger batch sizes can expedite convergence but might lead to overfitting. Additionally, this paper observed that appropriately adjusting the learning rate can further enhance model accuracy and training efficiency. Finally, using regularization techniques, such as L1 or L2 regularization, effectively controls model overfitting, thereby improving its generalization capabilities. These findings provide practical guidance for optimizing data-parallel CNN models and can assist researchers and practitioners in designing more efficient and accurate deep learning systems.

Copyright

© 2023 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Hao Yu
PY  - 2023
DA  - 2023/11/27
TI  - Analysis of Factors Influencing Data-Parallel CNN Models in Image Classification
BT  - Proceedings of the 2023 International Conference on Image, Algorithms and Artificial Intelligence (ICIAAI 2023)
PB  - Atlantis Press
SP  - 991
EP  - 1004
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-300-9_100
DO  - 10.2991/978-94-6463-300-9_100
ID  - Yu2023
ER  -