Dataset and pretrained MLIP (and QHA scripts) for thermosalient molecular crystal N‑2-Propylidene-4hydroxybenzohydrazide

model is a deployed NequIP model in the TorchScript format. dataset.extxyz is the dataset containing 1507 structures (three forms and isolated molecules) of N-2-Propylidene-4hydroxybenzohydrazid. The dataset was used to train the model. The dataset is based on DFT calculations that have been performed using the Vienna ab-initio simulation package, VASP.6.3.0 employing r2SCAN exchange-correlation functional. training_script_and_model_parameters.yaml is the training script with all the hyperparameters used to create the model. ssneb_transition_path_structures.zip is a directory containing 28 structures of the transition path calculated in ssneb, from structure 0 (form I) to structure 27 (form III) QHA files are to run/reproduce the Gibbs free energy calculations as described in the paper step1 creates the energy-volume curve structures step2 calculates the thermal properties in the HA (harmonic approximation) step3 calculate and plot Gibbs free energy The version of all necessary packages: - ase 3.22.1 - nequip 0.5.5 - python 3.10.4 For QHA: phonopy version 2.18.0