Calibrated Seq2seq Models for Efficient and Generalizable Ultra-fine Entity Typing

Open in new window