Data for Physics-informed Hamiltonian learning for large-scale optoelectronic property prediction

This repository contains the source data supporting the results of the paper “Physics-informed Hamiltonian learning for large-scale optoelectronic property prediction.”

The dataset includes electronic structure data, molecular dynamics trajectories, and model optimization data for GaAs, CsPbBr₃, and MAPbBr₃. For each material, the data are organized in subfolders containing first-principles reference data,  inputs for the Hamster code (version 0.2.1), and associated outputs. Previously published first-principles data for GaAs are available at Zenodo (DOI: 10.5281/zenodo.10089542)

The repository is intended to enable reproduction of the results, benchmarking of Hamiltonian learning approaches, and reuse of the data for optoelectronic materials modeling. Detailed file formats and data structures are described in the accompanying README or the Hamster documentation (see GitHub).