Carbontracker: Tracking and Predicting the Carbon Footprint of Training Deep Learning Models

Deep learning (DL) can achieve impressive results across a wide variety of\ntasks, but this often comes at the cost of training models for extensive\nperiods on specialized hardware accelerators. This energy-intensive workload\nhas seen immense growth in recent years. Machine learning (ML) may become a\nsignificant contributor to climate change if this exponential trend continues.\nIf practitioners are aware of their energy and carbon footprint, then they may\nactively take steps to reduce it whenever possible. In this work, we present\nCarbontracker, a tool for tracking and predicting the energy and carbon\nfootprint of training DL models. We propose that energy and carbon footprint of\nmodel development and training is reported alongside performance metrics using\ntools like Carbontracker. We hope this will promote responsible computing in ML\nand encourage research into energy-efficient deep neural networks.\n