oshizo/japanese-text-image-retrieval-train
收藏Hugging Face2025-01-26 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/oshizo/japanese-text-image-retrieval-train
下载链接
链接失效反馈官方服务：
资源简介：
JDocQA数据集是一个将shunk031/JDocQA的train split中的PDF数据转换为图像，并使用NDLOCR进行OCR处理得到的文本与图像配对的数据集。图像根据页面文字数量调整尺寸，并为每个OCR文本生成三个问题，随机选取一个问题作为数据集的一部分。数据集遵循cc-by-sa-4.0许可。
The JDocQA dataset is a collection of PDF data from the train split of shunk031/JDocQA, converted into images and paired with text obtained from OCR processing using NDLOCR. The images are resized based on the number of characters on the page, and three questions are generated for each OCR text, with one question randomly selected to be included in the dataset. The dataset follows the cc-by-sa-4.0 license.
提供机构：
oshizo
