StableLLaVA: Enhanced Visual Instruction Tuning with Synthesized Image-Dialogue Data

TL;DR

StableLLaVA introduces a novel method for visual instruction tuning by synthesizing diverse image-dialogue datasets using generative models, significantly improving multimodal model capabilities and achieving state-of-the-art results.

Contribution

The paper presents a new data collection approach that synthesizes images and dialogues, reducing domain bias and enabling scalable, diverse datasets for enhanced multimodal instruction tuning.

Findings

Significant improvements in over ten capabilities.

State-of-the-art results on multiple benchmarks.

Flexible dataset scaling with generative models.

Abstract

The remarkable multimodal capabilities demonstrated by OpenAI's GPT-4 have sparked significant interest in the development of multimodal Large Language Models (LLMs). A primary research objective of such models is to align visual and textual modalities effectively while comprehending human instructions. Current methodologies often rely on annotations derived from benchmark datasets to construct image-dialogue datasets for training purposes, akin to instruction tuning in LLMs. However, these datasets often exhibit domain bias, potentially constraining the generative capabilities of the models. In an effort to mitigate these limitations, we propose a novel data collection methodology that synchronously synthesizes images and dialogues for visual instruction tuning. This approach harnesses the power of generative models, marrying the abilities of ChatGPT and text-to-image generative models to yield a diverse and controllable dataset with varied image content. Additionally, datasets can be arbitrarily scaled. This not only provides greater flexibility compared to existing methodologies but also significantly enhances several model capabilities. Our research includes comprehensive experiments conducted on various datasets. The results emphasize substantial enhancements in more than ten commonly assessed capabilities. Additionally, our model achieves state-of-the-art results across multiple widely recognized multimodal benchmarks.