Dinosaur
A spectral dynamical core for global atmospheric modeling written in JAX.
https://github.com/neuralgcm/dinosaur
Keywords from Contributors
convolutional-neural-network transformers weather neuralgcm jax bayesian-methods probabilistic-programming climate cfd hyperparameter-tuning
Last synced: 1 day ago
JSON representation
Acceptance Criteria
- Revelant topics? false
- External users? true
- Open source license? true
- Active? true
- Fork? false
Repository metadata
NeuralGCM's differentiable dycore
- Host: GitHub
- URL: https://github.com/neuralgcm/dinosaur
- Owner: neuralgcm
- License: apache-2.0
- Created: 2023-12-08T14:43:35.000Z (almost 3 years ago)
- Default Branch: main
- Last Pushed: 2026-10-06T21:20:37.000Z (12 days ago)
- Last Synced: 2026-10-10T00:24:13.161Z (9 days ago)
- Language: Jupyter Notebook
- Homepage:
- Size: 26.8 MB
- Stars: 337
- Watchers: 9
- Forks: 41
- Open Issues: 7
- Releases: 17
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
README.md
Dinosaur: Differentiable Dynamics for Global Atmospheric Modeling 🦖
Authors: Jamie A. Smith, Dmitrii Kochkov, Peter Norgaard, Janni Yuval, Stephan Hoyer
Dinosaur is an old-fashioned (some might say prehistoric) dynamical core for global atmospheric modeling, re-written in JAX to meet the needs of modern AI weather/climate models:
- Dynamics: Dinosaur uses spectral methods to solve the shallow water equations and the primitive equations (moist and dry) on sigma coordinates.
- Auto-diff: Dinosaur supports both forward- and backward-mode automatic differentiation in JAX. This enables "online training" of hybrid AI/physics models.
- Acceleration: Dinosaur is designed to run efficiently on modern accelerator
hardware (GPU/TPU), including parallelization across multiple devices.
For more details, see our paper Neural General Circulation Models for Weather and Climate.
Usage instructions
Dinosaur is an experimental research project that we are still working on
documenting.
We currently have three notebooks illustrating how to use Dinosaur:
Each also has a semi-Lagrangian
variant demonstrating semi-Lagrangian advection at 6-12x longer time steps:
We recommend running them using Google Colab with a GPU runtime.
You can also install Dinosaur locally: pip install dinosaur
See also
If you like Dinosaur, you might also like
SpeedyWeather.jl, which
solves similar equations in Julia.
Contributing
See CONTRIBUTING.md for details. We are open to user
contributions, but please reach out (either on GitHub or by email) to coordinate
before starting significant work.
License
Apache 2.0; see LICENSE for details.
Owner metadata
- Name: neuralgcm
- Login: neuralgcm
- Email:
- Kind: organization
- Description:
- Website:
- Location:
- Twitter:
- Company:
- Icon url: https://avatars.githubusercontent.com/u/190404709?v=4
- Repositories: 5
- Last ynced at: 2026-10-16T04:13:05.688Z
- Profile URL: https://github.com/neuralgcm
GitHub Events
Total
- Create event: 55
- Delete event: 34
- Fork event: 11
- Issue comment event: 10
- Issues event: 6
- Pull request event: 86
- Push event: 184
- Release event: 11
- Watch event: 73
Last Year
- Create event: 12
- Delete event: 7
- Fork event: 1
- Issue comment event: 2
- Issues event: 2
- Pull request event: 18
- Push event: 52
- Release event: 2
- Watch event: 7
Committers metadata
Last synced: 3 days ago
Total Commits: 219
Total Committers: 12
Avg Commits per committer: 18.25
Development Distribution Score (DDS): 0.251
Commits in past year: 143
Committers in past year: 8
Avg Commits per committer in past year: 17.875
Development Distribution Score (DDS) in past year: 0.21
| Name | Commits | |
|---|---|---|
| Stephan Hoyer | s****r@g****m | 164 |
| Dmitrii Kochkov | d****v@g****m | 32 |
| Janni Yuval | j****l@g****m | 11 |
| Dinosaur authors | n****y@g****m | 2 |
| Amer Elsheikh | a****e@g****m | 2 |
| Jake VanderPlas | v****s@g****m | 2 |
| jaeminoh | j****h@g****m | 1 |
| Rebecca Chen | r****n@g****m | 1 |
| Peter Hawkins | p****s@g****m | 1 |
| Matej Aleksandrov | m****v@g****m | 1 |
| Ian Langmore | l****e@g****m | 1 |
| Hana Joo | h****o@g****m | 1 |
Committer domains:
- google.com: 10
Issue and Pull Request metadata
Last synced: 3 days ago
Total issues: 8
Total pull requests: 147
Average time to close issues: 13 days
Average time to close pull requests: 2 days
Total issue authors: 5
Total pull request authors: 4
Average comments per issue: 2.0
Average comments per pull request: 0.07
Merged pull request: 77
Bot issues: 0
Bot pull requests: 140
Past year issues: 3
Past year pull requests: 31
Past year average time to close issues: N/A
Past year average time to close pull requests: 6 days
Past year issue authors: 2
Past year pull request authors: 2
Past year average comments per issue: 0.33
Past year average comments per pull request: 0.1
Past year merged pull request: 16
Past year bot issues: 0
Past year bot pull requests: 28
Top Issue Authors
- shoyer (2)
- duncanwp (2)
- whpy (2)
- sit23 (1)
- milankl (1)
Top Pull Request Authors
- copybara-service[bot] (140)
- shoyer (5)
- jeromebarre (1)
- deven367 (1)
Top Issue Labels
- documentation (1)
- help wanted (1)
- enhancement (1)
Top Pull Request Labels
- pull ready (2)
Package metadata
- Total packages: 1
-
Total downloads:
- pypi: 9,921 last-month
- Total dependent packages: 0
- Total dependent repositories: 1
- Total versions: 13
- Total maintainers: 2
pypi.org: dinosaur
Differentiable Dynamics for Global Atmospheric Modeling 🦖
- Homepage:
- Documentation: https://dinosaur.readthedocs.io/
- Licenses: Apache-2.0
- Latest release: 1.5.1 (published 3 days ago)
- Last Synced: 2026-10-16T01:03:56.535Z (3 days ago)
- Versions: 13
- Dependent Packages: 0
- Dependent Repositories: 1
- Downloads: 9,921 Last month
-
Rankings:
- Dependent packages count: 10.095%
- Dependent repos count: 21.615%
- Average: 27.43%
- Downloads: 50.579%
- Maintainers (2)
Dependencies
- actions/cache v2 composite
- actions/checkout v2 composite
- actions/setup-python v2 composite
- styfle/cancel-workflow-action 0.7.0 composite
- fsspec *
- jax *
- jaxlib *
- numpy *
- pandas *
- pint *
- scikit-learn *
- scipy *
- tree-math *
- xarray *
- xarray-tensorstore *
Score: 17.52825965197309