Reset It and Forget It: Relearning Last-Layer Weights Improves Continual and Transfer Learning

Open in new window