Awesome-markdown-ebooks

Your GitHub PDFs, Now AI-Ready.

English | 简体中文

Project Introduction

The project gathers high-quality e-book repositories from GitHub and leverages MinerU 2.0 to transform the PDF content into Markdown format.

Each directory represents a repository originally hosted on GitHub.

If you possess any high-quality e-book resources that require conversion, you’re welcome to submit the links in an issue and we will assist with the PDF-to-Markdown extraction.

Our goal is to convert more high-quality knowledge data into AI-ready data.

Converted Repositories List

Repo url	Download
ChinaTextbook	opendatalab/awesome-markdown-ebooks/ChinaTextbook

Output File Structure Documentation (Based on MinerU2 vlm, Output File Structure)

1. Markdown Files and Images

• File Type: .md file + images/ folder
• Description: Final result of PDF to Markdown conversion
• Content: Document text content and image references

2. Model Output File

• Filename: model_output.txt
• Description: Intermediate inference data from VLM model
• Content: Model's visual understanding results of pages

3. Intermediate Processing File

• Filename: middle.json
• Description: Processed result from model_output.txt
• Content: Contains position information of text, images, formulas, tables, etc. in PDF

4. Content List File

• Filename: content_list.json
• Description: Final result converted from middle.json
• Content: Document conversion results segmented by elements, including page information

Star History