Skip to content

About

๐Ÿงฎ EcoLogits Calculator estimate the environmental footprint of your GenAI usage

Resources

Contributing

Stars

8 stars

Watchers

0 watching

Forks

Latest commit

ย 

History

243 Commits

Folders and files

Repository files navigation

EcoLogits Calculator

EcoLogits

Measure the Environmental Impact of Generative AI

License Python Streamlit


๐ŸŒ About

EcoLogits Calculator is an open-source interactive tool for estimating the energy consumption and environmental footprint of generative AI models. Developed by the non-profit CodeCarbon, this calculator helps individuals, researchers, and organizations understand and evaluate the sustainability of AI usage.

The calculator provides:

  • Energy consumption estimates for AI model inference
  • Environmental impact assessments including carbon emissions and water usage
  • Real-world scaling analysis to understand the broader impact
  • Support for multiple AI providers and models
  • Educational insights into sustainable AI practices

โœจ Features

  • ๐Ÿงฎ Interactive Calculator: Easily estimate environmental impacts by selecting a provider, model, and example usage
  • ๐Ÿค“ Expert Mode: Advanced options for users who want granular control over calculations
  • ๐Ÿข Company Mode: Estimate the aggregate environmental footprint of AI usage across an organisation
  • ๐Ÿ‘ฝ Expert Company Mode: Multi-model, multi-scenario analysis of token volumes and aggregated environmental impacts.
  • ๐Ÿช™ Token Estimator: Understand tokenization and estimate token counts for your inputs
  • ๐Ÿ“Š Visualization: Interactive charts showing environmental equivalences and scaling projections
  • ๐ŸŒ Multi-Provider Support: Evaluate impacts across different AI providers and models
  • ๐Ÿ“– Methodology Documentation: Transparent, science-backed calculation methods
  • ๐Ÿ“ฑ Web-Based Interface: No installation neededโ€”access directly through the web

๐Ÿš€ Quick Start

The Calculator is available at calculator.ecologits.ai.

If you want to run the calculator locally, follow the instructions below.

Prerequisites

  • Python 3.13+
  • uv package manager

Installation

# Clone the repository
git clone https://github.com/mlco2/ecologits-calculator.git
cd ecologits-calculator

# Install dependencies using uv
uv sync

# Run the application
uv run streamlit run app.py

The calculator will open in your browser at http://localhost:8501

๐Ÿ“š How It Works

The basic workflow of the EcoLogits Calculator involves the following steps:

  1. Select Model: Choose an AI provider and model from the available options
  2. Provide Input: Enter example prompts or text that you want to evaluate
  3. View Results: Instantly see energy consumption and environmental impact estimates
  4. Explore Equivalences: Understand impacts in relatable terms (e.g., "equivalent to driving X km")
  5. Analyze Scaling: See how impacts multiply when scaled to larger populations

Key Metrics

  • Energy Consumption: Measured in kWh per inference
  • Carbon Emissions: Estimated COโ‚‚ equivalent based on energy grid composition
  • Other Environmental Impacts: Water usage, resource consumption, etc.
  • Scaling Analysis: Projections for 1% of world population using the same prompt daily for one year

๐Ÿ“– Learning More

๐Ÿค Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines on how to:

  • Report bugs
  • Suggest features
  • Submit pull requests
  • Set up your development environment

๐Ÿ“œ License

This project is licensed under the CC BY-SA 4.0 License.

๐Ÿ™ Support

If you find this tool helpful, please consider:

  • โญ Starring this repository
  • ๐Ÿ’ฌ Sharing feedback and suggestions
  • ๐Ÿค Contributing to the project
  • ๐Ÿ’ Supporting CodeCarbon on their mission to make AI more sustainable

๐Ÿ“ฌ Questions?

Have questions or feedback? Feel free to:


Developed with โค๏ธ by CodeCarbon

About

๐Ÿงฎ EcoLogits Calculator estimate the environmental footprint of your GenAI usage

Resources

Contributing

Stars

8 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages