nomic-ai/VisRAG-Ret-Train-In-domain-data-by-source-hn-mine
收藏Hugging Face2025-04-01 更新2025-04-12 收录
下载链接:
https://hf-mirror.com/datasets/nomic-ai/VisRAG-Ret-Train-In-domain-data-by-source-hn-mine
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是关于视觉问答的,包含了从不同领域(如ArxivQA、ChartQA、InfoVQA、MP_DocVQA、PlotQA和SlideVQA)收集的训练数据。数据集中的特征包括查询内容、数据来源、图像文件名、负例段落数量、负例文本、正例段落数量、是否在前100个结果中找到金标准答案以及语言类型。
This dataset is for visual question answering and includes training data from different domains such as ArxivQA, ChartQA, InfoVQA, MP_DocVQA, PlotQA, and SlideVQA. The features of the dataset include the query text, source, image filename, number of negative passages, negative text, number of positive passages, whether the golden standard answer is in the top 100 results, and the language type.
提供机构:
nomic-ai



