Tooling for AI environmental impact assessment

During one of Hub France IA’s webinars, many open-source tools dedicated to assessing the environmental footprint of AI were discussed:

• Green Algorithms - A framework to estimate the carbon footprint of computational science, presented by Loic Lannelongue.
• CodeCarbon - A Python tool to track the carbon emissions from compute tasks like AI training or inference, presented by Benoit Courty.
• AI Energy Score - Energy efficiency ratings for AI models, presented by Boris Gamazaychikov.
• EcoLogits - A set of tools to estimate the environmental footprint of generative AI at inference, presented by me.