Model Tailor: Mitigating Catastrophic Forgetting in Multi-modal Large Language Models

TL;DR

Model Tailor is a post-training adjustment technique that effectively mitigates catastrophic forgetting in multi-modal large language models by selectively preserving pre-trained parameters and introducing a compensation mechanism, ensuring high performance on both original and new tasks.

Contribution

The paper introduces Model Tailor, a novel post-training method that preserves pre-trained parameters while adapting to new tasks, significantly reducing catastrophic forgetting in MLLMs.

Findings

Maintains ~99% performance on original tasks.

Achieves ~97% performance on new tasks.

Effective in multi-task scenarios.

Abstract

Catastrophic forgetting emerges as a critical challenge when fine-tuning multi-modal large language models (MLLMs), where improving performance on unseen tasks often leads to a significant performance drop on the original tasks. This paper presents a comprehensive analysis of catastrophic forgetting in MLLMs and introduces a post-training adjustment method called Model Tailor. Our method primarily preserves the pre-trained parameters while replacing a small number ($\leq$ 10\%) of fine-tuned parameters, maintaining $\sim$ 99\% effectiveness on original tasks versus pre-training, and achieving $\sim$ 97\% on new tasks compared to standard fine-tuning. Specifically, we derive a sparse mask to identify the "model patch", based on a fusion strategy that integrates salience and sensitivity analysis. Subsequently, a compensation mechanism is introduced to "decorate the patch", enhancing the model's performance on both target and original tasks. Additionally, our method is adaptable to multi-task scenarios. Through extensive experiments on InstructBLIP and LLaVA-1.5 in both image captioning and visual question answering tasks, our approach demonstrates significant task adaptability while preserving inherent pre-trained capabilities.