Leveraging Cloud Resource for Hyperparameter Tuning in Deep Learning Models

Main Article Content

Krishnateja Shiva, Pradeep Etikani, Vijaya Venkata Sri Rama Bhaskar, Ashok Choppadandi, Arth Dave

Abstract

The act of tweaking the hyperparameters is very vital in the enhancements of deep learning models, although it is expensive in terms of computational complexity. The following paper aims to examine the possibility of using cloud resources to invest in hyperparameter tuning. We discuss several popular cloud-based platforms and services, review results of their benchmarking, and provide proofs of concept for real-world use-cases on a convolutional neural network (CNN) model for image classification. We found that cloud resources show a large advantage in time and cost of hyperparameter tuning making it possible for deep learning practitioners to adopt the solution (Li, Wang, & Zhang, 2021).

Article Details

How to Cite
Krishnateja Shiva. (2022). Leveraging Cloud Resource for Hyperparameter Tuning in Deep Learning Models. International Journal on Recent and Innovation Trends in Computing and Communication, 10(2), 30–35. Retrieved from https://www.ijritcc.org/index.php/ijritcc/article/view/10980
Section
Articles