mllam/neural-lam

Add COSMO example notebook for onboarding

Open

#587 opened on Apr 10, 2026

View on GitHub
 (12 comments) (0 reactions) (2 assignees)Python (276 forks)auto 404
documentationhelp wanted

Repository metrics

Stars
 (282 stars)
PR merge metrics
 (PR metrics pending)

Description

Description

A hello_world_cosmo.ipynb notebook should be added to docs/notebooks/ to walk users through an end-to-end workflow with COSMO data, mirroring the DANRA example in #69.

The notebook already exists on Joel's research branch and should serve as the starting point: 👉 https://github.com/joeloskarsson/neural-lam-dev/blob/research/docs/reproduce_paper_sample.md

The in-progress PR is #392.

Main blocker

The full COSMO-2 dataset used for training is hosted at ETH and is multiple 100 GB, it will never be freely available. To make this notebook runnable out-of-the-box, a reduced COSMO sample needs to be prepared and hosted on an accessible platform (e.g. Hugging Face Datasets or Zenodo), similar to the small DANRA subset used in the DANRA notebook. Alternatively contact @leifdenby for available hosting options.

Related

  • Split from #69 (DANRA version lives there)
  • In-progress PR: #392

Contributor guide

Research direction
Adapt the existing notebook from Joel's research branch (https://github.com/joeloskarsson/neural lam dev/blob/research/docs/reproduce paper sample.md) to create a COSMO version. The main challenge is preparing a reduced COSMO dataset and hosting it publicly (e.g., on Hugging Face Datasets or Zenodo). Coordinate with @leifdenby or the PR #392 authors for hosting options and dataset details.
Tech stack
python
Domain
machine learningdata
Issue type
Documentation
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
Half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Mostly clear
Prerequisites
Python
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
65