Improved Knowledge Distillation for Pre-trained Language Models via Knowledge Selection

Open in new window