SVL-Caption
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/CAMMA-public/Surg-FTDA
下载链接
链接失效反馈官方服务：
资源简介：
该数据集名为SVL-Caption，包含了用于训练图像标注任务的图像-文本对。此外，该数据集还用于评估生成性任务，在这些任务中，模型会通过标准的标注评价指标如BLEU、METEOR和CIDEr来进行评估。该数据集的规模包括1351个图像-文本对，其专注于图像标注的任务。
This dataset, designated SVL-Caption, comprises image-text pairs intended for training image captioning tasks. Furthermore, it is employed to evaluate generative tasks, where models are assessed using standard captioning evaluation metrics including BLEU, METEOR, and CIDEr. With a total of 1,351 image-text pairs, this dataset is specifically focused on image captioning tasks.
