Skip to content

Latest commit

 

History

History
27 lines (20 loc) · 1.38 KB

README.md

File metadata and controls

27 lines (20 loc) · 1.38 KB
Raw

granite-wxc

This repository contains code and examples to apply the Prithvi WxC foundation model to downscaling tasks. In particular, the repository contains both code and instructions for generic fine-tuning tasks as well as fine-tuned models for MERRA2 2m temperature data as reference.

6x downscaling of MERRA-2 2m temperature

Getting started

  1. Create a virtual environment

  2. Clone this repository as well as that of the foundation model. Install both in the virtual environment.:

    git clone https://github.com/NASA-IMPACT/Prithvi-WxC
git clone https://github.com/IBM/granite-wxc.git
cd Prithvi-WxC
pip install '.[examples]'
cd ../granite-wxc
pip install '.[examples]'

  3. Run the notebook in the examples directory.

    This notebook will download model weights as well as sample data for basic illustration from Hugging Face.

Fine-tuned model

The fine-tuned model for MERRA-2 2m temperature data is available via Hugging Face. For an application to EURO-CORDEX data please refer to the paper.