Research paper · 2019-10-21
Quantifying the Carbon Emissions of Machine Learning
Alexandre Lacoste, Alexandra Luccioni, Victor Schmidt, Thomas Dandres
Abstract
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.
Details
- arXiv identifier
- 1910.09700
- Published
- 2019-10-21
- Authors
- Alexandre Lacoste, Alexandra Luccioni, Victor Schmidt, Thomas Dandres
Models That Cite This Paper
- Described byt5-small
- Described byt5-base
- Described bydistilgpt2
- Described bystable-diffusion-v1-5
- Described bydistilroberta-base
- Described byflan-t5-base
- Described bygpt2-large
- Described bystable-diffusion-safety-checker
- Described bystable-diffusion-v1-4
- Described bydistilbert-base-multilingual-cased
- Described bydistilbert-base-uncased-mnli
- Described byflan-t5-small
- Described byflan-t5-large
- Described byt5-large
- Described byt5-3b
- Described bydetr-doc-table-detection
- Described byroberta-large-mnli
- Described bydistilbert-base-cased-distilled-squad
- Described byflan-t5-xl
- Described bystable-diffusion-inpainting
- Described byspeecht5_tts
- Described bydistilbert-base-uncased-distilled-squad
- Described byt5-11b
- Described bybiobert-large-cased-v1.1-squad
- Described bydiscogs-maest-10s-pw-129e
- Described bypothole_detection
- Described byMoA-155M
- Described byMini-oss-0.6b
- Described byDistilQwen3-1.7B-uncensored
- Described byDNA-50M
- Described byDNA-175M
- Described bygemma-270m-math-reasoner
- Described byjano-tts
- Described byDiscovered
- Described bytiny-audio-granite-qwen
- Described bykinyarwanda-mms-vits-finetune
- Described bytrade-idp-qwen3vl-model-v1
- Described byQwen3.6-35B-A3B-heretic-lora
- Described byDeepSeek-R1-7B-MathPhysics-v7-lora
- Described byqwen3-medical-qlora