Modeling martensitic transformation temperatures in Zirconia-Ceria solid solutions using machine learning interatomic potentials

This repository contains files necessary to train a machine learning interatomic potential for the martensitic region of the Zirconia-Ceria solid solution. We provide a trained and compressed DeepMD-kit model, which is the file frozen_model_c.pb. Training data is provided in a DeepMD-compatible format within the archive file Train_Data.tar.gz. A training script is also provided, input_mod.json, as well as a python file necessary to connect said script to the training data.Initial training data consisted of a number of relaxations at various Cerium contents, ab initio molecular dynamics data, and strained structures. Iterative learning was employed to grow the dataset.