Using Deep Learning Models to Build a Grade All Rounder?

G b r t has shown great value across a variety types of data sets. It builds a grade, let's say, all rounder than many deep learning models focused on specific tasks. I think the message should be larger in the sense that this can be basically replicated for every field where we would assume that not strong enough base lines are delivered for these highly specialized tasks. So i think we will take with us on our way forward is definitely this critical eye on selecting useful base lines that are meaningful.