MetOffice / MetOffice/ai4c_hackathon

ACE2 demo notebook on JASMIN (PET)

Open
#8 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Jupyter Notebook
Stars
2
Forks
3
Avg merge
1m
Merged PRs (30d)
1

Description

ACE2 is an ML climate model developed by the Allen AI institute, with models and code available from hugging face. They also have ACE-hiro, a diffusion downscaler. We could create a notebook which runs ACE2 for climate and then does some downscaling

- [paper](https://www.nature.com/articles/s41612-025-01090-0)
- [ace2-era5](model https://huggingface.co/allenai/ACE2-ERA5)
- [ace2-hiro](diffusion downscaler https://huggingface.co/allenai/HiRO-ACE)

- [ ] Get model weights
- [ ] Get sample input data (ERA5?, CMIP6?)
- [ ] totrial notebook on ACE2
- [ ] tutorial notebook on ACE-hiro
- [ ] joint tutorial
- [ ] find way to run ACE2 for future climate? downscale CMIP6 future climate stuff

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reading the linked paper and the ACE2-ERA5 and HiRO-ACE model pages, then determine how to obtain model weights and sample ERA5 or CMIP6 input data. The work is complete when reproducible notebooks cover ACE2, ACE-hiro downscaling, their joint use, and the proposed future-climate workflow.

Written by the indexing model from the issue text.

Assessment

Tech stack
huggingface, jupyter-notebook, python, pytorch
Domain
data, machine-learning
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.