Non-Intrusive Adaptation: Input-Centric Parameter-efficient Fine-Tuning for Versatile Multimodal Modeling

TL;DR

This paper introduces AdaLink, a non-intrusive parameter-efficient fine-tuning method for large multimodal models, achieving competitive performance while maintaining the model's internal architecture unchanged, simplifying adaptation and deployment.

Contribution

The paper proposes AdaLink, a novel non-intrusive PEFT technique that matches state-of-the-art intrusive methods in performance for multimodal models.

Findings

AdaLink performs comparably to LoRA and full fine-tuning.

It is effective across text-only and multimodal tasks.

AdaLink simplifies model adaptation by avoiding internal architecture modifications.

Abstract

Large language models (LLMs) and vision language models (VLMs) demonstrate excellent performance on a wide range of tasks by scaling up parameter counts from O(10^9) to O(10^{12}) levels and further beyond. These large scales make it impossible to adapt and deploy fully specialized models given a task of interest. Parameter-efficient fine-tuning (PEFT) emerges as a promising direction to tackle the adaptation and serving challenges for such large models. We categorize PEFT techniques into two types: intrusive and non-intrusive. Intrusive PEFT techniques directly change a model's internal architecture. Though more flexible, they introduce significant complexities for training and serving. Non-intrusive PEFT techniques leave the internal architecture unchanged and only adapt model-external parameters, such as embeddings for input. In this work, we describe AdaLink as a non-intrusive PEFT technique that achieves competitive performance compared to SoTA intrusive PEFT (LoRA) and full model fine-tuning (FT) on various tasks. We evaluate using both text-only and multimodal tasks, with experiments that account for both parameter-count scaling and training regime (with and without instruction tuning).