1 paper · 1 filter
Chenyu Huang, Peng Ye, Shenghe Zheng +4
Transformer-based models with the pretrain-finetune paradigm bring about significant progress, along with the heavy storage and deployment costs of finetuned models on multiple tas…