Neural networks and test datasets for distinguishing between LNO and NiO phases via diffraction patterns

This dataset contains multiple neural networks trained to differentiate between the LNO and NiO phases in battery materials. Multiple different network architectures are included. A convolutional neural network, a vision transformer and a swin transformer. Also a number of hybrid architectures combining the CNN and ViT architecture. Additionally, two CNN are included which are trained to predict the orientation of LNO and NiO respectively. The phase can then be predicted by comparing the confidences of theses seperate orientation network. Furthermore, the data for two experimental datasets are included to test the competency of the trained networks. The data from the sample labeled "TD022_LNO" is a pure LNO sample, while the sample named "LNO700" is a LNO sample, which exhibits a thin rocksalt layer on its edge due to the annealing process.