DynStabNet Dataset: Crystal Structures, Phonon-Derived Dynamic Stability Labels, and Screening Data for Semiconductor Materi

This dataset supports the DynStabNet study on dynamic stability prediction and screening for semiconductor materials. It contains crystal structure inputs, processed structure files, phonon-calculation-related outputs, derived dynamic-stability labels, and analysis metadata for a series of inorganic semiconductor systems. The dataset covers 19 material families, including AlAs, BAs, BN, CdS, CdTe, GaAs, GaN, GaP, Ge, In2Se3, In2Te3, InAs, InP, InSb, InSe, InTe, Si, SiC, and ZnS. The released data are intended to support reproducibility of dataset construction, phonon-based label generation, and machine-learning-based dynamic stability screening.