Quantifying the Carbon Emissions of Machine Learning
This work addresses environmental concerns for machine learning practitioners and organizations by providing a tool to estimate and reduce carbon emissions, though it is incremental as it builds on existing awareness without introducing new methods.
The paper tackles the problem of quantifying carbon emissions from training neural networks by identifying key factors like server location, training duration, and hardware, and presents the Machine Learning Emissions Calculator to help the community understand and mitigate environmental impact.
From an environmental standpoint, there are a few crucial aspects of training a neural network that have a major impact on the quantity of carbon that it emits. These factors include: the location of the server used for training and the energy grid that it uses, the length of the training procedure, and even the make and model of hardware on which the training takes place. In order to approximate these emissions, we present our Machine Learning Emissions Calculator, a tool for our community to better understand the environmental impact of training ML models. We accompany this tool with an explanation of the factors cited above, as well as concrete actions that individual practitioners and organizations can take to mitigate their carbon emissions.