VisualRWKV: Exploring Recurrent Neural Networks for Visual Language Models

TL;DR

VisualRWKV introduces a linear RNN-based multimodal model that efficiently processes visual language tasks, achieving competitive performance with significant speed and memory advantages over Transformer-based models.

Contribution

It is the first to apply a linear RNN architecture to multimodal visual language modeling, integrating novel recurrence and image processing mechanisms.

Findings

Achieves performance comparable to Transformer-based models like LLaVA-1.5.

Provides 3.98x faster inference speed and 54% GPU memory savings.

Demonstrates effectiveness of linear RNNs in visual language tasks.

Abstract

Visual Language Models (VLMs) have rapidly progressed with the recent success of large language models. However, there have been few attempts to incorporate efficient linear Recurrent Neural Networks (RNNs) architectures into VLMs. In this study, we introduce VisualRWKV, the first application of a linear RNN model to multimodal learning tasks, leveraging the pre-trained RWKV language model. We propose a data-dependent recurrence and sandwich prompts to enhance our modeling capabilities, along with a 2D image scanning mechanism to enrich the processing of visual sequences. Extensive experiments demonstrate that VisualRWKV achieves competitive performance compared to Transformer-based models like LLaVA-1.5 on various benchmarks. Compared to LLaVA-1.5, VisualRWKV has a speed advantage of 3.98 times and can save 54% of GPU memory when reaching an inference length of 24K tokens. To facilitate further research and analysis, we have made the checkpoints and the associated code publicly accessible at the following GitHub repository: see https://github.com/howard-hou/VisualRWKV.