Starstruk Logo MarkStarstruk

v3.0.0

PaddlePaddle/PaddleOCRv3.0.0May 20, 2025by cuicheng01

AI Summary

A major version release introducing PP-OCRv5, PP-StructureV3, and PP-ChatOCRv4. It integrates PaddleX3.0 capabilities, upgrades to Paddle 3.0, and unifies deployment interfaces and model naming conventions.

Key Highlights

  • Released PP-OCRv5: Single model supports 5 text types and complex handwriting, improving accuracy by 13%.
  • Released PP-StructureV3: Supports multi-scenario, multi-layout PDF high-precision parsing.
  • Released PP-ChatOCRv4: Native support for ERNIE 4.5 Turbo, improving accuracy by 15%.
  • Integrated PaddleX3.0: Unified deployment capabilities covering high-performance inference, service deployment, and edge deployment.
  • Upgraded to Paddle 3.0: Compatible with CINN compiler; static model files renamed from `.pdmodel` to `.json`.

Breaking Changes

  • Static graph model storage filename changed from `xxx.pdmodel` to `xxx.json`.
  • Unified and optimized Python API and CLI interfaces.
  • Adopted a new unified model naming system.

New Features

  • PP-OCRv5 model support
  • PP-StructureV3 support
  • PP-ChatOCRv4 support
  • PaddleX3.0 deployment integration
  • Paddle 3.0 compatibility (CINN)

Full Release Notes

- **发布全场景文字识别模型PP-OCRv5：** 单模型支持五种文字类型和复杂手写体识别；整体识别精度相比上一代提升13个百分点。
 
- **发布通用文档解析方案PP-StructureV3：** 支持多场景、多版式 PDF 高精度解析，在公开评测集中领先众多开源和闭源方案。

- **发布智能文档理解方案PP-ChatOCRv4：** 原生支持文心大模型4.5 Turbo，精度相比上一代提升15个百分点。

- **重构部署能力，统一推理接口：** PaddleOCR 3.0 融合了飞桨 PaddleX3.0 工具的底层能力，全面升级推理、部署模块，优化 2.x 版本的设计，统一并优化了 Python API 和命令行接口（CLI）。部署能力现覆盖高性能推理、服务化部署及端侧部署三大场景。

- **适配飞桨框架 3.0，优化训练流程：** 新版本已兼容飞桨 3.0 的 CINN 编译器等最新特性，静态图模型存储文件名由 `xxx.pdmodel` 改为 `xxx.json`。

- **统一模型名称：** 对PaddleOCR3.0支持的模型命名体系进行了更新，采用更规范、统一的命名规则，为后续迭代与维护奠定基础。