Benchmarking GFN Methods for ML-Ready OPV Database Generation - Supporting Data

This dataset contains all computational data supporting the manuscript “Benchmarking GFN Methods for ML-Ready Organic Photovoltaic Database Generation” submitted to the Journal of Chemical Information and Modeling. The dataset includes optimized molecular geometries, electronic properties, and machine learning validation results for 140 organic molecules (76 from CEP database, 64 from QM9 database) computed using four GFN methods (GFN-FF, GFN0-xTB, GFN1-xTB, GFN2-xTB) and DFT reference calculations. Contents: - 670 optimized geometries (XYZ format) - Molecular identifiers (SMILES strings) - Structural and electronic properties (CSV format) - Complete machine learning validation package - Analysis scripts (Python) See README.md for detailed documentation.