GPT vs BERT, under same computation and data resource, which one is better for downstream tasks like GLUE?
guotong1988 opened this issue · 1 comments
guotong1988 commented
Thank you very much.
EricLee8 commented
感觉BERT之类的用auto encoder模式来预训练的模型更适合GLUE吧,GPT比较擅长text generation