Meteor: Mamba-based Traversal of Rationale for Large Language and Vision Models

TL;DR

Meteor introduces an efficient Mamba-based traversal method for rationales that enhances large language and vision models' understanding and answering abilities across diverse tasks without increasing model size.

Contribution

The paper presents Meteor, a novel approach leveraging Mamba architecture for rationale traversal, improving performance of LLVMs on vision-language tasks without additional encoders or larger models.

Findings

Significant performance improvements on multiple benchmarks.

Efficient processing of lengthy rationales with linear time complexity.

Enhanced understanding of diverse visual and conceptual information.

Abstract

The rapid development of large language and vision models (LLVMs) has been driven by advances in visual instruction tuning. Recently, open-source LLVMs have curated high-quality visual instruction tuning datasets and utilized additional vision encoders or multiple computer vision models in order to narrow the performance gap with powerful closed-source LLVMs. These advancements are attributed to multifaceted information required for diverse capabilities, including fundamental image understanding, real-world knowledge about common-sense and non-object concepts (e.g., charts, diagrams, symbols, signs, and math problems), and step-by-step procedures for solving complex questions. Drawing from the multifaceted information, we present a new efficient LLVM, Mamba-based traversal of rationales (Meteor), which leverages multifaceted rationale to enhance understanding and answering capabilities. To embed lengthy rationales containing abundant information, we employ the Mamba architecture, capable of processing sequential data with linear time complexity. We introduce a new concept of traversal of rationale that facilitates efficient embedding of rationale. Subsequently, the backbone multimodal language model (MLM) is trained to generate answers with the aid of rationale. Through these steps, Meteor achieves significant improvements in vision language performances across multiple evaluation benchmarks requiring diverse capabilities, without scaling up the model size or employing additional vision encoders and computer vision models.