This repository contains the dataset generated for training machine learning models in the paper:

“Automatic generation of input files with optimised k-point meshes for Quantum ESPRESSO self-consistent field total energy calculations.”

The dataset consists of:

summary.csv

A high-level overview of k-point convergence results for all materials, including:

• Goldilocks-optimised k-point meshes
• MC3D reference k-point meshes
• K-point distance metrics
• different converge levels including medium (2 ~ 4 meV), well (1 ~ 2meV), ultra ( (<= 1 meV)

structure_calc_details/

A collection of all structure files (.cif) and detailed calculation outputs, including:

• Chemical formula
• Source ID (traceable to MC3D)
• Number of atoms
• MC3D k-points Mesh
• Goldilocks_kpoints
• Total Energy (eV)
• wall_time
• energy_xc
• rho_cutoff
• wfc_cutoff
• energy_ewald
• fermi_energy
• energy_hartree
• energy_smearing
• energy_one_electron
• EleNumber of bands, electrons, k-points
• number_of_bravais_symmetries
• number_of_atomic_wfc
• number_of_symmetries
• energy_one_center_paw
• Convergence Notes
• Goldilocks_k_distance