MetOffice / MetOffice/ai4c_hackathon

Demonstrate corrdiff inference from earth2

Open
#7 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

nvidia earth2 offers a pretrained downscaling model (CMIP6 to ERA5), which could be used for demonstrating at the hackathon.

- [ ] set up environment on JASMIN
- [ ] download sample data
- [ ] create pet accessor
- [ ] create tutorial notebook for running model
- [ ] create PET pipeline
- [ ] visualise data
- [ ] calculate scores

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reviewing the earth2 downscaling model and the checklist in this issue, then determine the JASMIN environment and sample data needed. The work involves a pet accessor, tutorial notebook, PET pipeline, visualisation and score calculation; it is done when each listed demonstration step is completed.

Written by the indexing model from the issue text.

Assessment

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

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.