Suppose we train an SVM with 1 billion training data points, and after we get the solution, we find that there are 5 support vectors. Subsequently, we train a new SVM with a new training dataset that contains only the 5 support vectors as training data. The two SVM models will be the same. True False In Lloyd's algorithm for K-Means, the final centroids after the algorithm converges will always be actual data points from our dataset. True False OO