PaddleOCR 3.0 Technical Report

TL;DR

PaddleOCR 3.0 is an open-source OCR toolkit that offers advanced multilingual recognition, document parsing, and information extraction solutions with high efficiency, rivaling larger models.

Contribution

It introduces three major solutions for OCR and document understanding, with models that are smaller yet competitive in accuracy and efficiency.

Findings

Models with fewer than 100 million parameters achieve competitive accuracy.

Provides efficient tools for training, inference, and deployment.

Supports heterogeneous hardware acceleration for versatile deployment.

Abstract

This technical report introduces PaddleOCR 3.0, an Apache-licensed open-source toolkit for OCR and document parsing. To address the growing demand for document understanding in the era of large language models, PaddleOCR 3.0 presents three major solutions: (1) PP-OCRv5 for multilingual text recognition, (2) PP-StructureV3 for hierarchical document parsing, and (3) PP-ChatOCRv4 for key information extraction. Compared to mainstream vision-language models (VLMs), these models with fewer than 100 million parameters achieve competitive accuracy and efficiency, rivaling billion-parameter VLMs. In addition to offering a high-quality OCR model library, PaddleOCR 3.0 provides efficient tools for training, inference, and deployment, supports heterogeneous hardware acceleration, and enables developers to easily build intelligent document applications.