This is python implementation of K-means clustering algorithm for random datasets using transfer learning VGG-16
Initial weights loaded from already trained VGG_16 architecture based on IMAGENET datasets give more accuracy than single K-means, hence, k-means with VGG16 produces more robust result