Heron-Bench: A Benchmark for Evaluating Vision Language Models in Japanese

TL;DR

Heron-Bench is a new benchmark designed to evaluate Japanese vision-language models, addressing the language gap in multimodal understanding and providing insights into model strengths and limitations.

Contribution

The paper introduces Japanese Heron-Bench, a novel benchmark and baseline VLM for Japanese, filling a critical gap in multilingual multimodal evaluation.

Findings

Heron-Bench effectively measures Japanese VLM capabilities.

Baseline Japanese VLM shows specific strengths and limitations.

Identifies capability gaps between GPT-4V and baseline models.

Abstract

Vision Language Models (VLMs) have undergone a rapid evolution, giving rise to significant advancements in the realm of multimodal understanding tasks. However, the majority of these models are trained and evaluated on English-centric datasets, leaving a gap in the development and evaluation of VLMs for other languages, such as Japanese. This gap can be attributed to the lack of methodologies for constructing VLMs and the absence of benchmarks to accurately measure their performance. To address this issue, we introduce a novel benchmark, Japanese Heron-Bench, for evaluating Japanese capabilities of VLMs. The Japanese Heron-Bench consists of a variety of imagequestion answer pairs tailored to the Japanese context. Additionally, we present a baseline Japanese VLM that has been trained with Japanese visual instruction tuning datasets. Our Heron-Bench reveals the strengths and limitations of the proposed VLM across various ability dimensions. Furthermore, we clarify the capability gap between strong closed models like GPT-4V and the baseline model, providing valuable insights for future research in this domain. We release the benchmark dataset and training code to facilitate further developments in Japanese VLM research.