1 paper
Huan Zhang, Shenghua Fan, Shuyu Dong +3
Continual Learning with Pre-trained Models holds great promise for efficient adaptation across sequential tasks. However, most existing approaches freeze PTMs and rely on auxiliary…