大模型目标检测微调数据集

# TextVQA validation set with grounding truth bounding box The dataset used in the paper [MLLMs Know Where to Look: Training-free Perception of Small Visual Details with Multimodal LLMs](https://arxiv.org/pdf/2502.17422) for studying MLLMs' attention patterns. The dataset is sourced from [TextVQA](https://textvqa.org/dataset/) and annotated **manually** with ground-truth bounding boxes. We consider questions with a single area of interest in the image so that 4370 out of 5000 samples are kept. ## Citation If you find our paper and code useful for your research and applications, please cite using this BibTeX: ``` @article{zhang2025mllms, title={MLLMs know where to look: Training-free perception of small visual details with multimodal LLMs}, author={Zhang, Jiarui and Khayatkhoei, Mahyar and Chhikara, Prateek and Ilievski, Filip}, journal={arXiv preprint arXiv:2502.17422}, year={2025} } ```# 带真值边界框的TextVQA验证集 本数据集用于论文《MLLMs Know Where to Look: Training-free Perception of Small Visual Details with Multimodal LLMs》（arXiv:2502.17422），以研究多模态大语言模型（Multimodal Large Language Model, MLLM）的注意力模式。 该数据集源自[TextVQA](https://textvqa.org/dataset/)，并经人工手动标注了真值边界框（ground-truth bounding box）。 我们筛选出图像中仅包含单一感兴趣区域的问题，最终从5000个样本中保留了4370个。 ## 引用说明 若您的研究与应用工作中用到了本文及代码，请通过以下BibTeX格式进行引用： @article{zhang2025mllms, title={MLLMs know where to look: Training-free perception of small visual details with multimodal LLMs}, author={Zhang, Jiarui and Khayatkhoei, Mahyar and Chhikara, Prateek and Ilievski, Filip}, journal={arXiv preprint arXiv:2502.17422}, year={2025} }

搜集汇总

背景与挑战

背景概述

该数据集是TextVQA验证集的子集，包含4370个样本，用于研究多模态大语言模型（MLLMs）对小视觉细节的注意力模式。数据集中的每个样本都经过手动标注，包含真实边界框，适用于目标检测微调任务。