1 paper
Chao Yu, Wenhao Zhu, Chaoming Liu +2
The paradigm of pre-training followed by fine-tuning on downstream tasks has become the mainstream method in natural language processing tasks. Although pre-trained models have the…