https://github.com/atomicarchitects/nequix
Nequix: Training a foundation model for materials on a budget.
Science Score: 36.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
✓Academic publication links
Links to: arxiv.org -
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (9.7%) to scientific vocabulary
Repository
Nequix: Training a foundation model for materials on a budget.
Basic Info
- Host: GitHub
- Owner: atomicarchitects
- License: mit
- Language: Python
- Default Branch: main
- Homepage: https://arxiv.org/abs/2508.16067
- Size: 2.56 MB
Statistics
- Stars: 21
- Watchers: 0
- Forks: 1
- Open Issues: 1
- Releases: 3
Metadata Files
README.md
Nequix
See more information in our preprint.
Usage
Installation
bash
pip install nequix
ASE calculator
Using nequix.calculator.NequixCalculator, you can perform calculations in
ASE with a pre-trained Nequix model.
```python from nequix.calculator import NequixCalculator
atoms = ... atoms.calc = NequixCalculator("nequix-mp-1") ```
Training
Models are trained with the nequix_train command using a single .yml
configuration file:
bash
nequix_train <config>.yml
To reproduce the training of Nequix-MP-1, first clone the repo and sync the environment:
bash
git clone https://github.com/atomicarchitects/nequix.git
cd nequix
uv sync
Then download the MPtrj data from
https://figshare.com/files/43302033 into data/ then run the following to extract the data:
bash
bash data/download_mptrj.sh
Then start the training run. The first time this is run it will preprocess the data into HDF5 files:
bash
nequix_train configs/nequix-mp-1.yml
This will take less than 125 hours on a single 4 x A100 node. The batch_size in the
config is per-device, so you should be able to run this on any number of GPUs
(although hyperparameters like learning rate are often sensitive to global batch
size, so keep in mind).
Citation
bibtex
@article{koker2025training,
title={Training a foundation model for materials on a budget},
author={Koker, Teddy and Smidt, Tess},
journal={arXiv preprint arXiv:2508.16067},
year={2025}
}
Owner
- Name: The Atomic Architects
- Login: atomicarchitects
- Kind: organization
- Location: United States of America
- Website: https://atomicarchitects.github.io/
- Twitter: AtomArchitects
- Repositories: 2
- Profile: https://github.com/atomicarchitects
Research Group of Prof. Tess Smidt
GitHub Events
Total
- Release event: 3
- Watch event: 16
- Delete event: 1
- Issue comment event: 1
- Push event: 8
- Fork event: 3
- Create event: 4
Last Year
- Release event: 3
- Watch event: 16
- Delete event: 1
- Issue comment event: 1
- Push event: 8
- Fork event: 3
- Create event: 4
Packages
- Total packages: 1
-
Total downloads:
- pypi 195 last-month
- Total dependent packages: 0
- Total dependent repositories: 0
- Total versions: 3
- Total maintainers: 1
pypi.org: nequix
Nequix source code
- Homepage: https://pypi.org/project/nequix/
- Documentation: https://nequix.readthedocs.io/
- License: MIT License Copyright (c) 2025 Teddy Koker Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
-
Latest release: 0.2.0
published 10 months ago
Rankings
Maintainers (1)
Dependencies
- actions/checkout v4 composite
- astral-sh/setup-uv v5 composite
- actions/checkout v4 composite
- astral-sh/setup-uv v5 composite
- ase >=3.24.0
- cloudpickle >=3.1.1
- e3nn-jax >=0.20.7
- equinox >=0.11.11
- h5py >=3.14.0
- jax >=0.4.34; sys_platform == 'darwin'
- jax [cuda12]>=0.4.34; sys_platform == 'linux'
- jraph >=0.0.6.dev0
- matscipy >=1.1.1
- optax >=0.2.5
- pyyaml >=6.0.2
- tqdm >=4.67.1
- wandb >=0.19.11
- wandb-osh >=1.2.2
- 111 dependencies