If I am utilizing Sagemaker for training a model, (deployed or not - doesn\'t matter) writing predictions, what are the pros and cons of using Sagemaker\'s XGBoost vs. open
