This download contains a diverse benchmark spectroscopy dataset of around 62k large molecules extracted from organic crystals of the Cambridge Structural Database. Equilibrium molecular structures were obtained at the PBE level of density functional theory (DFT) for all contained molecules. Additionally, total energies and orbital eigenvalues are included, computed at the (i) PBE and PBE0 level of DFT for all molecules in vacuum, (ii) PBE0 level of DFT for a subset of around 31k molecules in water and (iii) PBE0 level of the more accurate GW method for a subset of 5k molecules in vacuum. The dataset and its subsets are provided in pandas dataframes, accessable from the python programming language, as explained in the provided tutorial file. The dataset may be used to develop and benchmark machine learning methods in quantum chemistry.
«
This download contains a diverse benchmark spectroscopy dataset of around 62k large molecules extracted from organic crystals of the Cambridge Structural Database. Equilibrium molecular structures were obtained at the PBE level of density functional theory (DFT) for all contained molecules. Additionally, total energies and orbital eigenvalues are included, computed at the (i) PBE and PBE0 level of DFT for all molecules in vacuum, (ii) PBE0 level of DFT for a subset of around 31k molecules in wat...
»