This dataset consists of charge densities of individual snapshots from a molecular dynamics trajectory (DFT simulations?). We insert 8 ethylene carbonate molecules in the simulation box. To quickly explore a large part of the configurational space we put Hookean constraints on the molecular bonds (to maintain molecular identity such that molecules are not torn apart at such high temperature) and run Langevin molecular dynamics with thermostat temperature of 3000 K. The simulation was run for 12380 steps of 0.5 fs.

The resulting CHGCAR files have been compressed with lz4 compression and packed in non-compressed tar archives with up to 1000 structures in each.

The datasplits json files contain the indices (0-index) of the train, validation and test sets used in the paper "Equivariant Graph neural networks for fast electron density estimation of molecules, liquids, and solids"

To only use equilibrated structures the first 1000 time steps are not used in the paper. Index 0 in the split files corresponds to index 1000 of the CHGCAR files (01000.CHGCAR.lz4).