Do we need Label Regularization to Fine-tune Pre-trained Language Models?

Open in new window