Sound-VECaps

This is the dataset for Sound-VECaps, a large-scale audio dataset with visual-enhanced captions. We also release the dataset for AudioCaps-Enhanced, the visual-enhanced AudioCaps testing dataset as the new benchmark.本数据集为Sound-VECaps专用数据集，是一款搭载视觉增强字幕的大规模音频数据集。我们同时发布了AudioCaps-Enhanced专用数据集，该数据集为经过视觉增强的AudioCaps测试数据集，可作为全新的基准测试集。