Making AI Less ''Thirsty'': Uncovering and Addressing the Secret Water Footprint of AI

Pengfei Li, Jianyi Yang, Mohammad A. Islam and Shaolei Ren*

(*corresponding author)

Note

This is the official implementation of the arxiv paper [PDF]

Requirements

• python>=3.6

Installation

• Clone this repo:

git clone https://github.com/Ren-Research/Making-AI-Less-Thirsty.git
cd Making-AI-Less-Thirsty

• Install dependencies:

pip install -r requirements.txt

Results

• Hourly carbon efficiency and on-site WUE for the first week of August 2022.

• Estimated water and carbon footprints of training LaMDA with different starting months in 2022.

Citation

@article{ren2023water,
  title={Making AI Less "Thirsty": Uncovering and Addressing the Secret Water Footprint of AI Models},
  author={Li, Pengfei and Yang, Jianyi and Islam, Mohammad A. and Ren, Shaolei},
  journal={arXiv preprint arXiv:2304.03271},
  year={2023}
}

Acknowledgement

• Energy data from EIA Opendata Project.
• Weather data from Iowa Environmental Mesonet.