Earthmover and Hugging Face collaborated to simplify running open-source AI weather forecasting models by streaming large meteorological datasets rather than requiring full downloads, using Earthmover’s cloud data platform built on Zarr and Icechunk. The article provides a tutorial showing how to initialize Microsoft’s Aurora model with ERA5 reanalysis data, run 24-hour forecasts, and validate results against observations. It also demonstrates deployment on Hugging Face’s computational infrastructure for users without local GPU access.
