---
title: XFUND_DoclingDocument
canonical_url: "https://www.modelscope.cn/datasets/ds4sd/XFUND_DoclingDocument"
md_url: "https://www.modelscope.cn/datasets/ds4sd/XFUND_DoclingDocument.md"
repository: ds4sd/XFUND_DoclingDocument
last_updated: 2025-03-13
license: "Apache License 2.0"
storage_size: "3.7 GB"
downloads: 21
stars: 0
---

# XFUND_DoclingDocument

> XFUND_DoclingDocument - ds4sd 在 ModelScope 开源的数据集。XFUND DoclingDocument Dataset

ds4sd/XFUND_DoclingDocument 是 ModelScope 魔搭社区上的数据集，存储大小 3.7 GB，采用 Apache License 2.0 许可。

- **Repository**: ds4sd/XFUND_DoclingDocument
- **License**: Apache License 2.0
- **Storage size**: 3.7 GB
- **Downloads**: 21
- **Stars**: 0
- **Last updated**: 2025-03-13

Source: https://www.modelscope.cn/datasets/ds4sd/XFUND_DoclingDocument

---

# XFUND DoclingDocument Dataset

This repository contains a private version of the XFUND dataset processed into the [**DoclingDocument**](https://ds4sd.github.io/docling/concepts/docling_document/) format using [Docling](https://github.com/DS4SD/docling).

## Dataset Overview

- **Total Samples:** 1393 annotated forms
- **Splits:**
  - **train:** 1043 samples
  - **validation:** 350 samples
- **Features:**
  - `docling_version`: Version of the Docling document format with which the document is processed.
  - `document_id`: Unique identifier for each document.
  - `GroundTruthDocument`: The **DoclingDocument** object containing the annotations.
  - `GroundTruthPageImages`: Images corresponding to each page.
  - `GroundTruthPictures`: Pictures present in the document.
  - `BinaryDocument`: Image of the form in binary format.
  - `mimetype`: MIME type of the binary document.

## Source

The original XFUND dataset was introduced in the paper: [XFUND](https://aclanthology.org/2022.findings-acl.253.pdf). It consists of annotated forms in 7 languages: `DE, FR, IT, JA, ZH, PT, ES`.

## Docling Framework

The conversion to the DoclingDocument format was performed using the [Docling](https://github.com/DS4SD/docling), which provides a standardized way to represent a document.