Add concise STRUCTURES25 model card

README.md

@@ -1,3 +1,51 @@
----
-license: cc0-1.0
----
+---
+license: cc0-1.0
+library_name: pytorch
+language:
+  - en
+tags:
+  - chemistry
+  - density-functional-theory
+  - orbital-free-dft
+  - electron-density
+  - graph-neural-network
+  - equivariant
+  - pytorch-lightning
+---
+
+<p align="center">
+  <a href="https://sciai-lab.org/"><img src="https://sciai-lab.org/images/logo.svg" alt="SciAI Lab" width="72" style="background-color: white; padding: 8px; border-radius: 12px;"></a>
+</p>
+
+# STRUCTURES25
+
+**Machine-learned orbital-free density functional theory**
+
+Pretrained models for optimizing molecular electron densities and energies, developed by **SciAI Lab, Heidelberg University**. Equivariant graph neural networks learn the kinetic-plus-exchange-correlation energy functional from reference DFT data.
+
+[GitHub](https://github.com/sciai-lab/structures25) · [Documentation](https://sciai-lab.github.io/structures25/) · [Paper](https://doi.org/10.1021/jacs.5c06219) · [Data](https://doi.org/10.5061/dryad.0cfxpnwcs)
+
+![STRUCTURES25 benchmark comparison and workflow: atom-centered electron densities, an equivariant neural energy functional, and iterative density optimization.](https://github.com/user-attachments/assets/00abb696-95e3-4aaa-857b-2b7548d45646)
+
+## Available models
+
+| Model | Training molecules | CLI name |
+| --- | --- | --- |
+| [QM9](https://huggingface.co/sciai-lab/structures25/tree/main/trained-on-qm9) | QM9 | `str25_qm9` |
+| [QMugs](https://huggingface.co/sciai-lab/structures25/tree/main/trained-on-qmugs) | Small-molecule QMugs subset | `str25_qmugs` |
+
+Both checkpoints support **H, C, N, O, and F** and use training data augmented with perturbed external potentials. Intended for molecular OF-DFT research within the chemical scope of the training data; accuracy and convergence on new systems require validation.
+
+## Get started
+
+Follow the [installation and setup guide](https://sciai-lab.github.io/structures25/installation.html) to install `mldft`, download the models, and configure their locations. Then run:
+
+```bash
+mldft example.xyz --model str25_qm9
+```
+
+Use `--model str25_qmugs` for the QMugs checkpoint. See the [usage guide](https://sciai-lab.github.io/structures25/usage.html) for options and the [replication guide](https://github.com/sciai-lab/structures25/blob/main/REPLICATION_GUIDE.md) for technical details and benchmarks.
+
+## Reference
+
+Remme et al., **Stable and Accurate Orbital-Free Density Functional Theory Powered by Machine Learning**, *J. Am. Chem. Soc.* **147**, 28851–28859 (2025). [DOI](https://doi.org/10.1021/jacs.5c06219) · [BibTeX](https://sciai-lab.github.io/structures25/#structures25-documentation)