TARA
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/ernie-research/tool-augmented-reward-model
下载链接
链接失效反馈官方服务：
资源简介：
该数据集专为图像字幕生成而设计，重点关注图像的详细描述，包括时间、地点等上下文信息。此外，该数据集被用于评估自然语言生成优化模型（NLSOM）在生成图像描述性字幕方面的能力。其主要任务为图像字幕生成。
This dataset is specifically designed for image captioning, with a focus on detailed image descriptions including contextual information such as time and location. Additionally, it is used to evaluate the performance of the Natural Language Generation Optimization Model (NLSOM) in generating descriptive captions for images. Its core task is image captioning.
