Memory-Space Visual Prompting for Efficient Vision-Language Fine-Tuning

Abstract: Current solutions for efficiently constructing large vision-language (VL) models follow a two-step paradigm: projecting the output of pre-trained vision encoders to the input space of pre-trained LLMs as visual prompts; and then transferring the models to downstream VL tasks via end-to-end parameter-efficient fine-tuning (PEFT). However, this paradigm still exhibits inefficiency since it significantly increases the input length of the LLMs. In this paper, in contrast to integrating visual prompts into inputs, we regard visual prompts as additional knowledge that facilitates LLMs in addressing tasks associated with visual information. Motivated by the finding that Feed-Forward Network (FFN) of LLMs acts as "key-value memory", we introduce a novel approach termed memory-space visual prompting (MemVP), wherein visual prompts are concatenated with the weights of FFN for visual knowledge injection. Experimental results across various VL tasks and LLMs reveal that MemVP significantly reduces the training time and inference latency of the finetuned VL models and surpasses the performance of previous PEFT methods. Code: https://github.com/JieShibo/MemVP