ATP-LLaVA: Adaptive Token Pruning for Large Vision Language Models

TL;DR

ATP-LLaVA introduces an adaptive token pruning method for large vision language models, reducing computational costs by 75% with minimal performance loss through layer-wise and instance-wise strategies.

Contribution

It proposes a novel adaptive token pruning module and spatial augmented pruning strategy that dynamically adjust token retention based on input and layer, improving efficiency.

Findings

Reduces token count by 75% on average

Maintains performance with only 1.9% degradation

Effective across seven benchmark datasets

Abstract

Large Vision Language Models (LVLMs) have achieved significant success across multi-modal tasks. However, the computational cost of processing long visual tokens can be prohibitively expensive on resource-limited devices. Previous methods have identified redundancy in visual tokens within the Large Language Model (LLM) decoder layers and have mitigated this by pruning tokens using a pre-defined or fixed ratio, thereby reducing computational overhead. Nonetheless, we observe that the impact of pruning ratio varies across different LLM layers and instances (image-prompt pairs). Therefore, it is essential to develop a layer-wise and instance-wise vision token pruning strategy to balance computational cost and model performance effectively. We propose ATP-LLaVA, a novel approach that adaptively determines instance-specific token pruning ratios for each LLM layer. Specifically, we introduce an Adaptive Token Pruning (ATP) module, which computes the importance score and pruning threshold based on input instance adaptively. The ATP module can be seamlessly integrated between any two LLM layers with negligible computational overhead. Additionally, we develop a Spatial Augmented Pruning (SAP) strategy that prunes visual tokens with both token redundancy and spatial modeling perspectives. Our approach reduces the average token count by 75% while maintaining performance, with only a minimal 1.9% degradation across seven widely used benchmarks. The project page can be accessed via https://yxxxb.github.io/ATP-LLaVA-page/.