TRAINING
Get Started
This page helps new users understand the workflow, environment setup, required tools, and available learning resources.
FAQs
What is Global-AI?
Global-EAGLE is an experimental AI-based global weather prediction system using ECMWF’s Anemoi-Core architecture, NOAA data, and a full ML pipeline including preprocessing (ufs2arco), training, inference, and verification (wxvx).
What are the required inputs for running EAGLE?
GFS input: NOAA GFS BDP S3 bucket
HRRR input: NOAA HRRR BDP S3 bucket
PrepBufr: Grid-to-obs verification inputs
Required data files include:
data/global_one_degree.nc
data/hrrr_15km.nc
data/gfs.zarr
data/hrrr.zarr
What platforms does EAGLE support?
NOAA RDHPCS Ursa (GPU nodes)
Other HPCs using Singularity/Apptainer containers
