Carbon emissions of a CNN model: A comparative study of colab and Jupyter notebook with an example
Author Identifier (ORCID)
Laizah Sashah Mutasa: https://orcid.org/0000-0003-1377-2862
Abstract
Green computing, also known as green technology, is an environmentally sustainable method of using computers and the various resources associated with them, such as monitors, printers, storage devices, networking, and communication systems, efficiently and effectively while having the least amount of impact, if any, on the environment. Green computing also encompasses choosing sustainably sourced raw materials, reducing electronic waste, and promoting sustainability through the use of renewable resources. Potholes detection is an important aspect of road maintenance and repair, as it allows for timely and effective intervention to prevent further vehicle damage and safety hazards for road users. In recent years, advances in computer vision and deep learning technologies have motivated the use of Convolutional Neural Networks (CNNs) for pothole detection. This study seeks to raise consciousness about carbon emissions by monitoring the emissions generated by a CNN model running on both a cloud platform and an outdated device. By tracking and comparing the carbon emissions of these two different computing setups, the study shed light on the environmental impact associated with utilizing modern cloud platforms versus other devices. The method utilized in this study consists of four key stages: data pre-processing, model training, evaluation, and carbon emissions tracking for both Google Colab and a laptop. The results highlight the environmental benefits of utilizing cloud computing for model execution. Organizations and individuals seeking to deploy resource-intensive models like CNNs for tasks such as pothole detection are encouraged to consider cloud-based platforms to reduce their carbon footprint.
Document Type
Conference Proceeding
Date of Publication
1-1-2026
Volume
2723 CCIS
Publication Title
Communications in Computer and Information Science
Publisher
Springer
School
School of Business and Law
Copyright
subscription content
First Page
94
Last Page
115
Comments
Kyei, E. A., Asare, J. W., Ujakpa, M. M., Mutasa, L. S., Gavua, E. K., Freeman, E., Acquah-Brown, W. L., & Lempogo, F. (2026). Carbon emissions of a CNN model: A comparative study of colab and Jupyter notebook with an example. In Communications in Computer and Information Science (Vol. 2723, pp. 94–115). Springer. https://doi.org/10.1007/978-3-032-13056-3_9