Google 最新(19.10)提出的预训练模型T5(Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer) 最近放出了源代码及paper.
连续刷榜GLUE SuperGLUE两大benchmark:https://gluebenchmark.com/leaderboard
paper地址: https://arxiv.org/abs/1910.10683
代码模型:https://github.com/google-resea
原文:https://www.cnblogs.com/jiayibing2333/p/12122181.html