3D-Transformer: Learning Context-Dependent Molecular Representations from 3D Spatial Information - 42Papers