moondream/ia_ocr
收藏Hugging Face2025-03-30 更新2025-04-08 收录
下载链接：
https://hf-mirror.com/datasets/moondream/ia_ocr
下载链接
链接失效反馈官方服务：
资源简介：
互联网档案馆OCR数据集，包含从互联网档案馆转录的文档页面，由Pixtral进行转录。数据集用于预训练，转录准确性一般。数据集包含图像和文本信息，分为多个片段配置。
Internet Archive OCR Dataset, consisting of document pages transcribed from the Internet Archive, transcribed by Pixtral. The dataset is used for pre-training with general transcription accuracy. The dataset includes image and text information and is divided into multiple shard configurations.
提供机构：
moondream
