codecarbon

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

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.

VS Code extension for CodeCarbon

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

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

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

Public submission hub for PUMA local-LLM benchmark results

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

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

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

Reproducible ML pipeline for crop recommendation — FastAPI REST API, DVC data versioning, MLflow experiment tracking, Docker Compose deployment, Great Expectations validation, Locust load testing, Prometheus/Grafana monitoring, GitHub Actions CI/CD. Deployed on AWS EC2

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.

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).