HoVLE: Unleashing the Power of Monolithic Vision-Language Models with Holistic Vision-Language Embedding

TL;DR

HoVLE introduces a holistic embedding module for monolithic vision-language models, enabling effective processing of visual and textual data in a shared space, and achieves performance close to leading models.

Contribution

This paper proposes a novel holistic embedding module and multi-stage training strategy for monolithic VLMs, significantly improving their performance without degrading language capabilities.

Findings

Achieves near state-of-the-art results on multiple benchmarks.

Outperforms previous monolithic VLMs by a large margin.

Effectively aligns visual and textual embeddings in a shared space.

Abstract

The rapid advance of Large Language Models (LLMs) has catalyzed the development of Vision-Language Models (VLMs). Monolithic VLMs, which avoid modality-specific encoders, offer a promising alternative to the compositional ones but face the challenge of inferior performance. Most existing monolithic VLMs require tuning pre-trained LLMs to acquire vision abilities, which may degrade their language capabilities. To address this dilemma, this paper presents a novel high-performance monolithic VLM named HoVLE. We note that LLMs have been shown capable of interpreting images, when image embeddings are aligned with text embeddings. The challenge for current monolithic VLMs actually lies in the lack of a holistic embedding module for both vision and language inputs. Therefore, HoVLE introduces a holistic embedding module that converts visual and textual inputs into a shared space, allowing LLMs to process images in the same way as texts. Furthermore, a multi-stage training strategy is carefully designed to empower the holistic embedding module. It is first trained to distill visual features from a pre-trained vision encoder and text embeddings from the LLM, enabling large-scale training with unpaired random images and text tokens. The whole model further undergoes next-token prediction on multi-modal data to align the embeddings. Finally, an instruction-tuning stage is incorporated. Our experiments show that HoVLE achieves performance close to leading compositional models on various benchmarks, outperforming previous monolithic models by a large margin. Model available at https://huggingface.co/OpenGVLab/HoVLE.