Accuracy is not the only Metric that matters: Estimating the Energy Consumption of Deep Learning Models

Johannes Getzner, Bertrand Charpentier, Stephan Günnemann
4/3/2023
cs.LG

No official code found. Best implementation: JohannesGetzner/dl-energy-estimator (2 stars, last commit 2026-01-27).

Abstract

Modern machine learning models have started to consume incredible amounts of energy, thus incurring large carbon footprints (Strubell et al., 2019). To address this issue, we have created an energy estimation pipeline1, which allows practitioners to estimate the energy needs of their models in advance, without actually running or training them. We accomplished this, by collecting high-quality energy data and building a first baseline model, capable of predicting the energy consumption of DL models by accumulating their estimated layer-wise energies.

View on arXivView PDF

Code Implementations(1)

JohannesGetzner/dl-energy-estimatorCommunity60%

ICLR 2023 Publication: Accuracy is not the only Metric that matters: Estimating the Energy Consumption of Deep Learning Models (Spotlight Paper for the Tackling Climate Change with Machine Learning Workshop)

21Shell, PythonOct 22, 20228 months ago
climate-changecnndeep-learningenergy-consumptionmachine-learning+1 more

Similar Papers

Estimating Deep Learning energy consumption based on model architecture and training environment

91% match1 repo2023

Watt For What: Rethinking Deep Learning's Energy-Performance Relationship

89% match2023

How to estimate carbon footprint when training deep learning models? A guide and review

89% match2023

Uncovering Energy-Efficient Practices in Deep Learning Training: Preliminary Steps Towards Green AI

88% match2023

Minimizing Energy Consumption of Deep Learning Models by Energy-Aware Training

88% match2023

Towards energy-efficient Deep Learning: An overview of energy-efficient approaches along the Deep Learning Lifecycle

87% match2023

Cite this paper

@article{getzner2023accuracy,
  title  = {Accuracy is not the only Metric that matters: Estimating the Energy Consumption of Deep Learning Models},
  author = {Johannes Getzner and Bertrand Charpentier and Stephan Günnemann},
  year   = {2023},
  eprint = {2304.00897},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2304.00897v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2304.00897.svg)](https://codeofpaper.com/paper/2304.00897)

Discussion