DFT datasets for training machine-learning potential to model lithium borosilicate glasses using DeePMD

Modifying ring structures in lithium borate glasses under compression: MD simulations using a machine-learning potential Shingo Urata, Aik Rui Tan, and Rafael Gómez-Bombarelli Phys. Rev. Materials 8, 033602 – Published 7 March 2024 DFT_DataSets.zip inlucudes atom configurations, energies, forces, box size, atom types, atom kinds, and virial in coord.raw, energy.raw, force.raw, type.raw, type_map.raw, and virial.raw, respectively. These data were divided into 10 sets in each forder named as set.000 to set.009. All DFT data were evaluated using PBE with a cutoff energy of 600 eV by VASP. SiO2-B2O3-Li2O.json is the input file for training DeePMD ver1.33. SiO2-B2O3-Li2O.json is the optimized potential model of DeePMD.