Very fast Clustering with Kohonen Learning Procedure KMEANS based on NUMPY, SCIPY or THEANO - View it on GitHub