Leveraging Visual Tokens for Extended Text Contexts in Multi-Modal Learning

TL;DR

This paper introduces VisInContext, a method that uses visual tokens to efficiently extend in-context text length in multi-modal models, reducing resource costs and improving downstream task performance.

Contribution

It presents a novel technique that significantly increases in-context text length in multi-modal models using visual tokens, with minimal additional computational costs.

Findings

Expands in-context text length from 256 to 2048 tokens.

Achieves superior downstream benchmark performance.

Enhances document understanding and retrieval capabilities.

Abstract

Training models with longer in-context lengths is a significant challenge for multimodal model due to substantial GPU memory and computational costs. This exploratory study does not present state-of-the-art models; rather, it introduces an innovative method designed to increase in-context text length in multi-modality large language models (MLLMs) efficiently. We present Visualized In-Context Text Processing (VisInContext), which processes long in-context text using visual tokens. This technique significantly reduces GPU memory usage and floating point operations (FLOPs) for both training and inferenceing stage. For instance, our method expands the pre-training in-context text length from 256 to 2048 tokens with nearly same FLOPs for a 56 billion parameter MOE model. Experimental results demonstrate that model trained with VisInContext delivers superior performance on common downstream benchmarks for in-context few-shot evaluation. Additionally, VisInContext is complementary to existing methods for increasing in-context text length and enhances document understanding capabilities, showing great potential in document QA tasks and sequential document retrieval.