DeepCNT-22 - Model and Dataset

This dataset supports our publication Dynamics of Growing Carbon Nanotube Interfaces Probed by Machine Learning-Enabled Molecular Simulations. It includes the DeepCNT-22 model, a machine learning force field (MLFF) developed for simulating CNT growth on iron catalysts, along with the data and files used in the training and simulation processes.

DeepCNT-22 model (deepcnt-22.pb): The fully trained MLFF model used for production simulations in our study.
DeepCNT-22 dataset (dataset.zip): Contains the training, validation, and test datasets for DeepCNT-22, in both ASE-db format (training.db, validation.db, test.db) and DeePMD format (training.zip, validation.zip, test.zip). Each image is labeled with total energy (image.data.energy) in eV, force (image.data.forces) in eV/Å, and stress (image.data.stress) in eV/ų.
Conversion script (db_to_deepmd.py): A Python script for converting data from ASE-db to DeePMD format.
DeePMD files (deepmd.zip): Input files utilized for training the DeepCNT-22 model using DeePMD.
LAMMPS files (lammps.zip): Input files for simulating CNT growth in LAMMPS.
VASP files (vasp.zip): Input files used for labeling data with VASP.
CNT growth trajectory (6,5_traj.dump): The full trajectory of the (6,5) CNT growth dumped every 2 ps, contains atomic energy (c_pe_atom) and C-C coordination number (c_coord) for each carbon atom.