HuggingFaceM4/Docmatix
收藏Hugging Face2024-08-26 更新2024-12-14 收录
下载链接：
https://hf-mirror.com/datasets/HuggingFaceM4/Docmatix
下载链接
链接失效反馈官方服务：
资源简介：
Docmatix是一个用于文档视觉问答（Document Visual Question Answering）的大规模数据集，是Idefics3模型微调的一部分。数据集包含图像和PDF两种格式，分别对应不同的配置（images和pdf）。每个样本包含图像和文本对话，文本对话包括用户和助手的问答对。数据集还提供了与其他DocVQA数据集的比较信息。
Docmatix is a massive dataset for Document Visual Question Answering, used for the fine-tuning of the vision-language model Idefics3. The dataset includes documents in image and PDF formats, along with conversations between a user and an assistant about the document content. The dataset configurations include images, PDF, and zero-shot experiments, each with different data files and features. The dataset is large, containing millions of images and question-answer pairs.
提供机构：
HuggingFaceM4
