Support Vectors Machines are generalizations of linear decision boundaries for linearly nonseparable cases. In that notebook its coded from scracht a simple svm and we study how it learns and converge.