codecarbon

A Satellite Semantic Segmentation Project using Unet and Attention Unet with Pytorch,

VS Code extension for CodeCarbon

Mini workshop to provide a peak of what’s happening under the hood of models currently at the frontier of the AI revolution, and about how we can track the emissions of our own code.

A tool to measure and compare the energy consumption of code variants.

Energy Consumption of various Machine Learning and Deep Learning Models using codecarbon

Make an impact with a single API call — plant trees, clean oceans, capture CO₂, and donate to global causes.

Apache Airflow plugin that tracks the energy utilization and carbon footprints of data workflows.

Make an impact with a single API call — plant trees, clean oceans, capture CO₂, and donate to global causes.

This project develops forecasting models for monitoring forest health, focusing on Larch Casebearer damage using Yolov8 models, with a focus on evaluating the environmental impact of the training process

Public submission hub for PUMA local-LLM benchmark results

Secure AWS workload profiler with Terraform, CodeCarbon, CloudWatch, and lease-aware auto-stop safety.

Codebase for the MLCost application developed for my thesis for the Telecommunications Enginnering bachelor, Universidad Rey Juan Carlos

End-to-end ML pipeline for predicting house prices using the Ames Housing dataset.

The system tracks the emissions of a given recommendation algorithm on a given dataset.

Diving into the world of Tracking CO2 Emissions from our software or code. Code Carbon is a lightweight open-source Python Library that lets you track the Co2 emissions produced by running code.

Carbon-aware machine learning benchmarking framework evaluating predictive performance alongside energy consumption and CO₂ emissions.

One of my beloved academic works that is exploring the roots and effects of an unequilibred trade-off between development expenses and gained performance has its basis residing in the current repository, which provides a full implementation of a DL model that can carry out semantic segmentation tasks while staying compact (12.2 M parm; 39 MB).

MakeImpact is a Python SDK that helps you integrate environmental actions into your applications. 🌱 With simple commands, you can easily contribute to sustainability efforts, like planting trees and cleaning oceans. 🐙

This study analyzes the energy efficiency (E(n)) of MergeSort and QuickSort on Apple M2 architecture. While QuickSort is faster for large datasets (N=50,000), it consumes significantly more energy (3.75 J) than MergeSort (2.50 J) due to higher instantaneous power draw. It highlights the need for energy-aware algorithm selection.

An end-to-end machine learning project predicting employee burnout risk (No Risk / At Risk / Burned Out) using 8,500 samples. Covers EDA, feature engineering, SMOTE-Tomek imbalance handling, 5 classifiers with GridSearchCV optimization, and carbon emission tracking via CodeCarbon.