WattLayer: Get Layers Right to Estimate Inference Energy of Neural Networks
Provides a standardized methodology for estimating AI inference energy consumption, addressing a critical need for energy-efficient AI design across diverse tasks and hardware.
Proposed a task-independent, layer-wise energy estimation model for neural networks, achieving a median error of 19.6% on over 100,000 layers across 295 architectures and 3 hardware platforms, outperforming state-of-the-art methods.
The widespread adoption of Artificial Intelligence (AI) has led to increasing concerns about energy consumption, yet there is a lack of standardized methodologies to accurately estimate AI inference energy consumption, particularly across various tasks and architectures. In this study, we propose a task independent, layer-wise energy estimation model for AI architectures. Our model is evaluated on a large dataset of more than 100,000 layers for 295 neural network architectures across 3 widely-used tasks and 3 distinct hardware platforms. Our approach achieves a median error of 19.6%, outperforming state-of-the-art methods. We further show that layer-wise decomposition generalize to new tasks without complete retraining, by leveraging shared layers across architectures. It offer tools, insights and a precise methodology to empower stakeholders in designing energy-efficient AI systems.