Exposing the Cost of Deep Learning Audio Development

2026-10-01

AI Summary

This study addresses the overlooked high energy consumption during the development phase of deep learning for audio. Leveraging Grid5000 activity logs, we estimate the energy usage of four projects conducted by the Multispeech team and perform a comparative analysis against model training energy costs. Our findings systematically reveal, for the first time, that total energy consumption during development is 3 to 256 times greater than that required for optimal model training, substantially exceeding the training phase itself. By challenging the prevailing paradigm that focuses exclusively on training energy, this work advocates for establishing comprehensive carbon footprint reporting mechanisms spanning the entire machine learning lifecycle, thereby providing critical empirical evidence to advance green AI practices.

Abstract

The environmental impact of deep learning has attracted increasing attention over the past decade. Existing studies mainly focus on the energy and carbon emissions of model training and inference, while the whole development phase is often overlooked. Yet, architecture prototyping and intensive experiments are conducted during this stage, which is highly energy-demanding. In this article, we propose a methodology to estimate these costs, based on activity logs from the Grid5000 shared computing platform used by the LORIA laboratory. As a case-study, we focus on audio projects developed in the Multispeech research team. We evaluate the overall energy cost of four projects, and we compare them to those of training the reported models. Our results show that the energy required for the development phase is 3 to 256 times greater than that required to train the best-performing model alone. These results advocate for a more systematic reporting of energy consumption across the entire life cycle of deep learning-based audio projects.

Problem

• Deep Learning
• Energy Consumption
• Audio Development
• Environmental Impact
• Carbon Emissions

Innovation

• Deep Learning
• Energy Consumption
• Audio Development
• Life Cycle Assessment
• Activity Logs