Post-training quantization speeds up autoregressive weather models with little quality loss
Post-Training Quantization of Autoregressive Weather Models
Machine Learning
Summary
Weather forecasts use computer models that are often slow and require a lot of computing power. The authors looked at a way to make these models run faster and use less energy by simplifying how the computers handle numbers after the models are already trained. They tried this on two weather prediction models and found that forecasts for short time periods still made sense. This means weather models can become more efficient without losing much accuracy.
What this means in practice
- •For cloud infrastructure teams: Deploy weather emulators with post-training quantization to reduce GPU resource usage and power consumption during forecast computations.
- •For mobile device developers: Implement quantized weather forecasting models on edge devices to enable localized short-range weather predictions with efficient computation.
Authors
Ananyo Bhattacharya, Swastik Bhattacharya, Christiane Jablonowski
Abstract
Advancements in high-resolution numerical weather prediction (NWP) and data assimilation (DA) have shaped the developments in deep learning (DL) architectures emulating atmospheric dynamics. Emulators for weather forecasting exhibit forecast quality comparable to physics based models at forecast horizon scaling from few days to subseasonal time scales. The emulators are driven by hardware-accelerated matrix multiplication in autoregressive inferences, significantly reducing the computation time and resources required for NWP. Optimization of the matrix multiplication processes in GPU architectures provides opportunities to scale towards high-resolution domain, and offers implementation of out of the box solutions. Post-training quantization (PTQ) has been demonstrated across multiple DL architectures to accelerate and increase the number of computations in unit time while consuming less power, enabling applications on edge hardware. In this study, we investigate the effect of PTQ on pre-trained AI emulators for global-scale weather forecasting. We implement PTQ algorithms in Deep Learning Weather Prediction (DLWP) and FourCastNet (FCN) models as a proof of concept for geophysical fluid dynamics applications. We systematically investigate the effect of PTQ on emulator inferences over short-range forecast horizons. Evaluation of PTQ configurations using simulated quantization hints at qualitatively meaningful forecasts over short-time horizons. These results provide a first benchmark of PTQ for autoregressive weather emulators and a basis for quantization-based optimization of DL models for dynamical systems.