Think of this as another way of achieving that. This theoretically has a higher ceiling of how much it can predict at a time. And more importantly is a lot more

Think of this as another way of achieving that. This theoretically has a higher ceiling of how much it can predict at a time. And more importantly is a lot more memory efficient during actual inference.