weathernext

Google Research and Google DeepMind also build their models for Google's own TPU hardware. It's only natural for them, but weather centres can't buy TPUs and can't / don't want to be locked to Google's cloud offerings. For Gencast ('WeatherNext Gen', I believe), the repository provides instructions and caveats (https://github.com/google-deepmind/graphcast/blob/main/docs/...) for inference on GPU, and it's generally slower and more memory intensive. I imagine that FGN/WeatherNext 2 would also have similar surprises. Training is also harder. DeepMind has only open-sourced the inference code for its first two models, and getting a working, reasonably-performant training loop written is not trivial. NOAA hasn't retrained its weights from scratch, but the fine-tuning they did re: GFS inputs still requires the full training apparatus.

Graphcast (the model this is based on) has been validated in weather models for a while[1]. It uses transformers, much like LLMs. Transformers are really impressive at modeling a variety of things and have become very common throughout a lot of ML models, there's no reason to besmirch these methods as "integrating an LLM into a weather model" [1] https://github.com/google-deepmind/graphcast

blaylockbk/Herbie: https://github.com/blaylockbk/Herbie : > Download numerical weather prediction datasets (HRRR, RAP, GFS, IFS, etc.) from NOMADS, NODD partners (Amazon, Google, Microsoft), ECMWF open data, and the Pando Archive System The Herbie docs mention a "GFS GraphCast" but not yet GenCast? https://herbie.readthedocs.io/en/stable/gallery/noaa_models/... "GenCast predicts weather and the risks of extreme conditions with state-of-the-art accuracy" (2024) https://deepmind.google/discover/blog/gencast-predicts-weath... "Probabilistic weather forecasting with machine learning" (2024) ; GenCast paper https://www.nature.com/articles/s41586-024-08252-9 google-deepmind/graphcast: https://github.com/google-deepmind/graphcast Are there error and cost benchmarks for these predictive models?