This is a basic geometries dataset with 810 samples, it includes 50x50 images with triangles, circles and squares. #Example code to extract dataset import h5py from sklearn.model_selection import train_test_split hf = h5py.File('dataset.h5', 'r') X = np.array(hf.get('X')) y = np.array(hf.get('y')) #Split dataset 20% for Testing and 80% for Training X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=0)