English | 简体中文
Join our training camp WeChat group for discussion and support.
This repository contains the course materials for the MinerU Training Camp. The curriculum covers MinerU product selection, local and containerized deployment, online API usage, model fine-tuning, OpenClaw Skill development, AI Agent building, and document parsing evaluation based on OmniDocBench, Dingo, and WebMainBench.
Course materials include:
- Course notes in Markdown for online reading, search, and maintenance
- Slide decks in PDF format for preview and sharing
- Video links to the Bilibili course collection
This course is designed for:
- Developers who want to quickly understand the MinerU product matrix and choose the right entry point
- Engineers who need to deploy MinerU locally, with Docker, or in domestic computing environments
- Developers who want to build document parsing, knowledge base, or information extraction applications with MinerU Open API
- Algorithm and platform teams interested in document parsing model fine-tuning, evaluation, and production workflows
- Practitioners who want to build Agents, Skills, or research tools with AI coding workflows
We recommend following the course order:
- Start with Lesson 01 to understand the MinerU product matrix and scenario-based selection.
- Continue with Lesson 02 to practice deployment for both the MinerU project and MinerU-HTML project.
- If you prefer a hosted, deployment-free workflow, study Lesson 03 on MinerU Online API.
- If you need model customization, study Lesson 04 on MinerU 1.2B fine-tuning.
- If you focus on ecosystem integration, study Lessons 05 and 06 on Skill and Agent development.
- Lessons 07 and 08 are optional. They are recommended for learners who need quality evaluation, model benchmarking, or model optimization.
| Lesson | Topic | Notes | Slides | Video |
|---|---|---|---|---|
| 00 | MinerU Training Camp Preview | Notes | - | Bilibili |
| 01 | MinerU Product Matrix: Quick Start and Scenario-Based Selection | Notes | Bilibili | |
| 02-1 | MinerU Project Deployment Practice | Notes | Bilibili | |
| 02-2 | MinerU-HTML Project Deployment Practice | Notes | Bilibili | |
| 03 | MinerU Online API Hands-on Tutorial | Notes | Bilibili | |
| 04 | MinerU 1.2B Model Fine-Tuning | Notes | Bilibili | |
| 05 | OpenClaw Skill Development: Flexible Document Q&A with MinerU | Notes | Bilibili | |
| 06 | Vibe Coding Practice: Building an AI Agent with MinerU | Notes | Bilibili | |
| 07 | MinerU-HTML Parsing Evaluation with Dingo | Notes | Bilibili | |
| 08 | OCR Benchmarking: Deep Evaluation of MinerU with OmniDocBench and Dingo | Notes | Bilibili |
.
├── 00课：MinerU实战训练营先导预告/
├── 01课：MinerU 全场景产品矩阵：快速上手与选型/
├── 02课：MinerU 多环境部署实践：从开源容器化到信创生态适配/
│ ├── 02-1：MinerU 项目/
│ └── 02-2：MinerU-HTML 项目/
├── 03课：MinerU 在线 API 实战教程/
├── 04课：MinerU 1.2B 模型微调/
├── 05课：OpenClaw Skill 开发实践：基于MinerU 的文档灵活问答/
├── 06课：Vibe Coding 实战：基于MinerU 搭建AI Agent/
├── 07课：MinerU-HTML 解析效能验证：基于 Dingo 的量化评测方法/
└── 08课：OCR模型对标：基于 OmniDocBench 与 Dingo 的 MinerU 模型深度评测/
- MinerU Website
- MinerU GitHub Repository
- MinerU Open API Documentation
- MinerU Ecosystem
- Training Camp WeChat Group
- OmniDocBench
- Dingo
- WebMainBench
- NanaDraw
Issues and pull requests are welcome for:
- Typos, broken links, or formatting issues
- Additional notes for deployment, API usage, and model fine-tuning
- Course practice cases, FAQs, and troubleshooting experience
- Evaluation data, evaluation scripts, or model comparison results
Copyright (c) 2026 OpenDataLab.
This course material is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
