Charting nanocluster structures via convolutional neural networks

The repository contains a notebook for the training of the autoencoder for the RDFs for structural classification. The notebook describes the procedure going from RDFs calculation to clustering of the reduced space. In the folder are contained Au147 structures, together with the associated pretrained AE, the 3D chart and the different clustering performed varying mean shift bandwidth.

Files:

-  ChartAu147.ipynb: notebook

- Configurations: directory with the dataset divided according to the CNA classification of the structures, xyz format with no headers, every 147 lines is a single structure

- Libraries: directory with functions imported in the notebook

- Precomputed: directory with the precomputed outputs

             - rdfs.npy: preocmputed RDFs of the data stored in configurations, npy format to load with NumPy

             - labels.npy: CNA labels of the RDFs, npy format to load with NumPy

             - model_au147.pth:  pretrained model for au147

             - scaler_au147.pkl: minmax scaler of the RDFs

             - chart_3d.dat: 3d space generated via the encoder on the au147 dataset

             - ae_reconstructions.npy: reconstructions of the rdfs of the model (model_au147.pth)

             - MSscanbw: pretrained mean shift clustering with different bandwidths, the file "clus_vs_bw.dat"  reports the number of clusters associated to each  bandwidth