This three-part webinar series will introduce you to Anemoi, an open-source software framework developed by ECMWF and European meteorological services to develop and implement data-driven weather forecasting and Earth systems models.
ECMWF experts will explain what Anemoi is, the motivation behind it, and how it is being used to develop cutting-edge global and regional AI weather models, as well as applications in downscaling, ensemble forecasting and more. Our speakers will cover the key components of Anemoi: datasets, graphs, models, training, and inference.
The webinar series is free and open to anyone, and is scheduled as follows.
Introduction + Anemoi Datasets
Tuesday 16th June 15:00 CEST | Ana Prieto Nemesio and Harrison Cook
This webinar will give an overview of the Anemoi framework, its different components, and how it is being applied in operational and research settings. In the second part of the webinar, we will explore Anemoi Datasets, an Anemoi component which enables efficient creation, loading and transformation of datasets ready for data-driven weather modelling.
Anemoi Graphs + Anemoi Models
Wednesday 17th June 15:00 CEST | Mario Santa Cruz
Since Anemoi is built around graph neural networks, our second webinar introduces two key Anemoi components for defining training experiments. Anemoi Graphs is a suite of tools for creating graphs for use in data-driven weather forecasting models, while Anemoi Models specifies the deep learning model architecture and includes tools for preprocessing, among other things.
Anemoi Training + Anemoi Inference
Thursday 18th June 15:00 CEST | Ana Prieto Nemesio and Harrison Cook
Our final webinar will begin by covering Anemoi Training, which provides the tools to train large-scale machine learning models, including profiling, efficient parallelisation, and diagnostic tools. Finally, we will take a tour of Anemoi Inference, which generates forecasts (performs inference) from a trained model, using specified initial conditions.
Want to find out more in advance? Watch our short introductory video on Anemoi, explore the Anemoi repo on GitHub, or take a look at the Anemoi Documentation.