risashinoda/BioVITA

BioVITA是一个用于视觉-文本-声学对齐的生物数据集，包含多种生物物种的媒体数据（如音频记录）及其分类信息。数据集提供了物种的科学名称、常用名称、分类层级（如类、目、科、属）、记录ID、来源、许可证、质量等级、文件名、下载链接和源页面URL等特征字段。数据分为训练集、测试集和未见集，用于支持多模态对齐研究。BioVITA is a biological dataset for visual-textual-acoustic alignment, containing media data (e.g., audio recordings) of various biological species along with their taxonomic information. The dataset includes features such as scientific name, common name, taxonomic hierarchy (e.g., class, order, family, genus), recording ID, source, license, quality grade, file name, download URL, and source page URL. It is split into train, test, and unseen sets to support multimodal alignment research.