AVQA

该数据集汇总了来自AVQA和MusicAVQA数据集的问答对，并通过指令调整模板进行了增强。该数据集支持对模型在回答需要多模态推理的复杂问题方面的评估工作，其任务旨在应对涵盖音频和视觉模态的问题解答。

This dataset aggregates question-answer pairs from the AVQA and MusicAVQA datasets, and enhances them using instruction tuning templates. This dataset supports the evaluation of models' performance on complex questions that require multimodal reasoning, with its targeted task focused on question answering covering both audio and visual modalities.