Exact exchange-correlation potentials and energies for learning density functionals

This directory contains the training data used to develop machine-learning based exchange-correlation (XC) functionals in density functional theory (DFT) (https://doi.org/10.48550/arXiv.2409.06498). It also contains the .ptc model files afor the MLXC functionals and the python scripts to use them. It also contains aggregated benchmark results for the MLXCs.   

In terms of the training data, it contains the density (and its higher order derivatives) for several atoms and molecules obtained using highly accurate configuration interaction (CI) calculations. Further, it contains the exact XC potential and XC energy obtained from the CI densities through accurate inverse DFT calculations.