The training objective is designed to dynamically select and update prompts from the prompt pool to learn tasks sequentially given a pretrained backbone model.
確定！ 回上一頁
