Skip to content

blubrom/MLCA

 
 

Repository files navigation

MLCA: a tool for Machine Learning Life Cycle Assessment


MLCA is a tool for assessing the environmental footprint of computation. It has been described in the MLCA paper presented at the 2024 International Conference on ICT for Sustainability (ICT4S) in Stockholm Sweden.

This repository contains the code for the tool as weel as the code and results of all the experiments realised to validate the tool.

Run the tool

Prerequisite

Python 3, pipenv recommended

Setup pipenv

Install pipenv globally

$ sudo pip3 install pipenv

Install dependencies and create a python virtual environment.

$ pipenv install -d 
$ pipenv shell

Launch a local server

Once in the pipenv environment

Development server uses uvicorn and fastapi, you can launch development server with the uvicorn CLI.

$ uvicorn boaviztapi.main:app --host=localhost --port 5000

OpenAPI specification (Swagger)

Once API server is launched API swagger is available at http://localhost:5000/docs. Experiments can be run at this adress. The file with the experiments also detail how to call the tool automatically.

📜 License

GNU Affero General Public License v3.0

About

a tool to perform simplified LCA on ML processes based on BOAVIZTAPI

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 91.7%
  • TeX 7.7%
  • Other 0.6%