Is Hard Attention Helping You?

We have played around with hard attention, like traind using aroand it's hard to get into worry. And i think there is often a trade off between performances some sense in solving a problem in a human like way. The image net models are biased towards using texture to classify objects much more than humans are. It would almost certainly be competitionaly more efficient, at least learning,. not fom aor we love i gote ad relatively hard footed.