$\gamma-$MoD: Exploring Mixture-of-Depth Adaptation for Multimodal Large Language Models

TL;DR

This paper introduces $oldsymbol{ extgamma}$-MoD, a novel mixture-of-depth adaptation strategy for multimodal large language models that significantly reduces computational costs while maintaining performance.

Contribution

The paper proposes ARank, a new metric for identifying redundant layers, and introduces shared vision-language router and masked routing learning to maximize sparsity in MLLMs.

Findings

Over 90% of dense layers converted to MoD layers.

Reduces training time by 31% and inference time by 53%.

Maintains performance with only 1.5% accuracy drop.

Abstract

Despite the significant progress in multimodal large language models (MLLMs), their high computational cost remains a barrier to real-world deployment. Inspired by the mixture of depths (MoDs) in natural language processing, we aim to address this limitation from the perspective of ``activated tokens''. Our key insight is that if most tokens are redundant for the layer computation, then can be skipped directly via the MoD layer. However, directly converting the dense layers of MLLMs to MoD layers leads to substantial performance degradation. To address this issue, we propose an innovative MoD adaptation strategy for existing MLLMs called $\gamma$-MoD. In $\gamma$-MoD, a novel metric is proposed to guide the deployment of MoDs in the MLLM, namely rank of attention maps (ARank). Through ARank, we can effectively identify which layer is redundant and should be replaced with the MoD layer. Based on ARank, we further propose two novel designs to maximize the computational sparsity of MLLM while maintaining its performance, namely shared vision-language router and masked routing learning. With these designs, more than 90% dense layers of the MLLM can be effectively converted to the MoD ones. To validate our method, we apply it to three popular MLLMs, and conduct extensive experiments on 9 benchmark datasets. Experimental results not only validate the significant efficiency benefit of $\gamma$-MoD to existing MLLMs but also confirm its generalization ability on various MLLMs. For example, with a minor performance drop, i.e., -1.5%, $\gamma$-MoD can reduce the training and inference time of LLaVA-HR by 31.0% and 53.2%, respectively.