-title:GLU Variants Improve Transformerurl:http://arxiv.org/abs/2002.05202v1-title:Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformerurl:http://arxiv.org/abs/1910.10683v3Publisher:googleTask:-sub_tag:文本生成sub_tag_en:Text2Text Generationtag:自然语言处理tag_en:Natural Language Processing