1 paper
Siqi Ping, Yuzhu Mao, Yang Liu +2
Although large-scale pre-trained models hold great potential for adapting to downstream tasks through fine-tuning, the performance of such fine-tuned models is often limited by the…