atmorep-itwinai-plugin
itwinai plugin for AtmoRep model
Science Score: 44.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
✓CITATION.cff file
Found CITATION.cff file -
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
○Academic publication links
-
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (12.4%) to scientific vocabulary
Repository
itwinai plugin for AtmoRep model
Basic Info
Statistics
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 1
Metadata Files
README.md
AtmoRep itwinai plugin
This repository is an extended version of AtmoRep, originally developed by the AtmoRep team. This version integrates AtmoRep into the itwinai framework, enhancing its usability within digital twin applications.
How to use this plugin
First of all, install the plugin.
[!NOTE] If you run on an HPC system remember to load the correct modules before.
activate_env.shprovides an example for JSC.
bash
pip install "atmorep-itwinai-plugin @ git+https://github.com/matbun/atmorep-itwinai-plugin"
Now you can import the AtmoRep trainer in your code using
```python from itwinai.plugins.atmorep.trainer import AtmoRepTrainer
my_trainer = AtmoRepTrainer(...) ```
Alternatively, you can launch training using the configuration file in this repository:
```bash itwinai exec-pipeline --config-name config.yaml
If you want to dynamically override some fields in the config file
itwinai exec-pipeline --config-name config.yaml \ epochs=2 \ trainingpipeline.steps.0.config.pathmodels my_models ```
[!NOTE] Consider that this model needs to be distributed on 4 GPUs, as it implements model-parallel distributed training. It can scale to multiple nodes and the SLURM jobs cript
slurm.jsc.shprovides an example on how to launch distributed training on HPC. You can adjust the number of nodes in that file.
Docker containers
This repository provides two examples of Dockerfiles, one for JupyterLab images and the other for simple Docker containers. Both Dockerfiles are based on the itwinai container image, which already provides most of the dependencies.
Owner
- Name: Matteo Bunino
- Login: matbun
- Kind: user
- Website: matbun.github.io
- Repositories: 1
- Profile: https://github.com/matbun
Fellow @ CERN Openlab. Former data Science student @ {EURECOM, Polytechnic University of Turin}
Citation (CITATION.cff)
cff-version: 1.2.0
title: "AtmoRep itwinai plugin"
message: >-
If you use this software, please cite both the original AtmoRep project,
the itwinai framework, and this modified version.
type: software
authors:
# Original AtmoRep Authors
- given-names: Christian
family-names: Lessig
email: christian.lessig@ecmwf.int
affiliation: European Centre for Medium-Range Weather Forecasts (ECMWF)
- given-names: Ilaria
family-names: Luise
email: ilaria.luise@cern.ch
affiliation: European Organization for Nuclear Research (CERN)
- given-names: Martin
family-names: Schultz
email: m.schultz@fz-juelich.de
orcid: 'https://orcid.org/0000-0003-3455-774X'
affiliation: Forschungszentrum Jülich (FZJ)
- given-names: Michael
family-names: Langguth
email: m.langguth@fz-juelich.de
orcid: 'https://orcid.org/0000-0003-3354-5333'
affiliation: Forschungszentrum Jülich (FZJ)
- given-names: Bing
family-names: Gong
affiliation: Forschungszentrum Jülich (FZJ)
- given-names: Scarlet
family-names: Stadler
affiliation: Forschungszentrum Jülich (FZJ)
# New Contributors to the Plugin
- given-names: Matteo
family-names: Bunino
email: "matteo.bunino@cern.ch"
affiliation: CERN
identifiers:
- type: url
value: 'https://arxiv.org/abs/2308.13280'
description: corresponding Preprint
repository-code: 'https://github.com/matbun/atmorep-itwinai-plugin'
url: 'https://github.com/matbun/atmorep-itwinai-plugin'
abstract: >-
This repository is an extended version of AtmoRep, originally developed
by the AtmoRep team. This version integrates AtmoRep into the itwinai
as a plugin, enhancing its usability within digital twin applications.
AtmoRep is a novel, task-independent stochastic computer
model of atmospheric dynamics that can provide skillful
results for a wide range of applications. AtmoRep uses
large-scale representation learning from artificial
intelligence to determine a general description of the
highly complex, stochastic dynamics of the atmosphere
from the best available estimate of the system's historical
trajectory as constrained by observations. This is enabled
by a novel self-supervised learning objective and a unique
ensemble that samples from the stochastic model with a
variability informed by the one in the historical record.
Our work establishes that large-scale neural networks can
provide skillful, task-independent models of atmospheric
dynamics. With this, they provide a novel means to make
the large record of atmospheric observations accessible
for applications and for scientific inquiry, complementing
existing simulations based on first principles.
license: Apache-2.0
# commit: your_commit_hash_here
# version: your_version_here
# date-released: '2025-XX-XX'
references:
- type: software
title: "AtmoRep (Original Version)"
authors:
- name: "Christian Lessig"
- name: "Ilaria Luise"
- name: "Martin Schultz"
- name: "Michael Langguth"
version: "2.0 (preprint)"
url: "https://github.com/clessig/atmorep"
license: "MIT"
- type: software
title: "itwinai"
authors:
- name: "Matteo Bunino"
url: "https://github.com/interTwin-eu/itwinai"
license: "Apache-2.0"
GitHub Events
Total
- Release event: 1
- Delete event: 1
- Member event: 1
- Push event: 15
- Pull request event: 4
- Create event: 6
Last Year
- Release event: 1
- Delete event: 1
- Member event: 1
- Push event: 15
- Pull request event: 4
- Create event: 6
Dependencies
- actions/checkout v4 composite
- gaurav-nelson/github-action-markdown-link-check v1 composite
- actions/checkout v4 composite
- github/super-linter/slim v7 composite
- actions/checkout v4 composite
- eosc-synergy/sqaaas-assessment-action v2 composite
- eosc-synergy/sqaaas-step-action v1 composite
- ghcr.io/intertwin-eu/itwinai torch-slim-latest build
- itwinai [torch]
- pytest >=8.3.4
- 158 dependencies