In the exp_informer, it is forgotten to add scale as an arg. This makes it so it will always go to default init argument which is True. So when you try to take False as an argument to the informer training, then it will still be the default value of True.
In the exp_informer, it is forgotten to add scale as an arg. This makes it so it will always go to default init argument which is True. So when you try to take False as an argument to the informer training, then it will still be the default value of True.