X-PLUG/mPLUG-DocOwl

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

2411 stars · 154 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

mPLUG-DocOwl is a family of multimodal large language models from Alibaba for OCR-free document understanding, including DocOwl1.5 and DocOwl2 for parsing documents, tables, and charts. It provides models, training data, inference, and fine-tuning code released on HuggingFace and ModelScope.

Use cases

• answer questions about scanned documents without OCR
• extract information from tables in document images
• understand and query charts with a vision-language model
• build a document VQA chatbot
• fine-tune a multimodal LLM on my own document data
• parse multi-page PDFs into structured answers

When to choose

• you need document, chart, or table understanding from images without a separate OCR pipeline
• you want open weights plus training data and fine-tuning code
• you need a research-grade multimodal LLM with strong DocVQA/ChartQA benchmarks

When to avoid

• you need lightweight CPU-only text extraction rather than GPU-based deep learning
• you need production OCR with precise bounding boxes
• you cannot host 3B-8B parameter models

Facets

library · maturity active

machine-learning deep-learning ocr nlp image-processing artificial-intelligence large-language-models computer-vision pdf python multimodal-llm document-understanding ocr-free chart-understanding table-understanding visual-question-answering huggingface natural-language-processing gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="X-PLUG/mPLUG-DocOwl")