ersky/Classification-Crystal-Symmetry-Perovskite-using-Deep-Learning

XRD Perovskite Crystal Symmetry Dataset

This repository hosts the dataset used in the research paper "Implementation of Convolutional Neural Network with Whale Optimization Algorithm on X-ray Diffraction Patterns for Crystal Symmetry Detection in Perovskite".

It contains the preprocessed X-ray diffraction (XRD) patterns, their labels, the metadata tables, and the t-SNE visualization results used for training, validation, and testing of the model.

Note: This repository only contains the data. The source code, including the scripts to train, validate, and run the model, is available on GitHub: GitHub Repository

Repository Contents

File	Type	Description
X_train_raw.npy	NumPy array	Raw XRD patterns for the training set
y_train.npy	NumPy array	Labels for the training set
X_val_raw.npy	NumPy array	Raw XRD patterns for the validation set
y_val.npy	NumPy array	Labels for the validation set
X_test_raw.npy	NumPy array	Raw XRD patterns for the test set
y_test.npy	NumPy array	Labels for the test set
df_train_meta_filtered (1).csv	CSV	Filtered metadata for the training set
df_val_meta_filtered (1).csv	CSV	Filtered metadata for the validation set
hasil_tsne_3d.csv	CSV	3D t-SNE embedding results used for data visualization

Usage

Loading the data

import numpy as np
import pandas as pd

# XRD patterns and labels
X_train = np.load("X_train_raw.npy")
y_train = np.load("y_train.npy")
X_val = np.load("X_val_raw.npy")
y_val = np.load("y_val.npy")
X_test = np.load("X_test_raw.npy")
y_test = np.load("y_test.npy")

# Metadata and t-SNE results
df_train_meta = pd.read_csv("df_train_meta_filtered (1).csv")
df_val_meta = pd.read_csv("df_val_meta_filtered (1).csv")
df_tsne = pd.read_csv("hasil_tsne_3d.csv")

print(X_train.shape, y_train.shape)

Downloading from the Hugging Face Hub

from huggingface_hub import snapshot_download

local_dir = snapshot_download(
    repo_id="ersky/Classification-Crystal-Symmetry-Perovskite-using-Deep-Learning",
    repo_type="dataset",
)

Running the program

To train, validate, or run inference with the model, please follow the instructions in the GitHub repository:

GitHub Repository

Download the files from this repository and place them in the data directory expected by the source code, as described in the GitHub README.

Contact

For questions about the data or the research, please open an issue in the GitHub repository or start a discussion in the Community tab of this repository.