multitask-learning

A semi-parametric prompt tuning method improving multitask generalization for parameter-efficient fine-tuning with cross-task zero-shot generalization.