Dataset and voxelized geometries for NiTi TPMS lattice CNN surrogate model

This repository contains the dataset used to voxelize NiTi TPMS lattice geometries from STL files and pair them with FEA stress–strain data in order to train a 3D convolutional neural network (CNN) surrogate model, via 5-fold cross-validation predicting point-wise pseudoelastic stress from voxelized geometry, strain, and loading state.