1 paper
Jianquan Li, Xiaokang Liu, Honghong Zhao +3
Pre-trained language models (e.g., BERT) have achieved significant success in various natural language processing (NLP) tasks. However, high storage and computational costs obstruc…