Learning Expressive Prompting With Residuals for Vision Transformers

TL;DR

EXPRES introduces residual learnable tokens for vision transformers, significantly improving prompt tuning efficiency and achieving state-of-the-art results across multiple vision tasks with less computational cost.

Contribution

The paper proposes a novel prompt learning method with residual tokens for vision transformers, enhancing adaptation efficiency and performance over existing prompt tuning approaches.

Findings

Achieves state-of-the-art prompt tuning on VTAB benchmark

Orders of magnitude more prompt efficient than existing baselines

Analytically demonstrates computational benefits over finetuning

Abstract

Prompt learning is an efficient approach to adapt transformers by inserting learnable set of parameters into the input and intermediate representations of a pre-trained model. In this work, we present Expressive Prompts with Residuals (EXPRES) which modifies the prompt learning paradigm specifically for effective adaptation of vision transformers (ViT). Out method constructs downstream representations via learnable ``output'' tokens, that are akin to the learned class tokens of the ViT. Further for better steering of the downstream representation processed by the frozen transformer, we introduce residual learnable tokens that are added to the output of various computations. We apply EXPRES for image classification, few shot learning, and semantic segmentation, and show our method is capable of achieving state of the art prompt tuning on 3/3 categories of the VTAB benchmark. In addition to strong performance, we observe that our approach is an order of magnitude more prompt efficient than existing visual prompting baselines. We analytically show the computational benefits of our approach over weight space adaptation techniques like finetuning. Lastly we systematically corroborate the architectural design of our method via a series of ablation experiments.