site_name: PaddleOCR 文档 site_url: https://paddlepaddle.github.io/PaddleOCR/ site_author: PaddleOCR PMC site_description: Awesome multilingual OCR toolkits based on PaddlePaddle (practical ultra lightweight OCR system, support 80+ languages recognition, provide data annotation and synthesis tools, support training and deployment among server, mobile, embedded and IoT devices) repo_name: PaddlePaddle/PaddleOCR repo_url: https://github.com/PaddlePaddle/PaddleOCR copyright: Copyright © 2024 Maintained by PaddleOCR PMC. edit_uri: edit/main/docs/ theme: name: material logo: ./static/images/logo.jpg favicon: ./static/images/logo.jpg custom_dir: overrides features: - announce.dismiss - content.tooltips - content.code.copy - content.tabs.link - content.footnote.tooltips - content.action.edit - content.action.view - navigation.expand # 默认打开所有的字节 - navigation.tabs # 顶级索引被作为tab - navigation.tabs.sticky # tab始终可见 - navigation.top # 开启顶部导航栏 - navigation.tracking # 导航栏跟踪 - navigation.footer - navigation.indexes - search.highlight # 搜索高亮 - search.share # 搜索分享 - search.suggest # 搜索建议 - toc.follow # 目录跟踪-页面右侧的小目录 palette: - media: "(prefers-color-scheme: light)" # 浅色 scheme: default primary: indigo accent: indigo toggle: icon: material/brightness-7 name: Switch to dark mode - media: "(prefers-color-scheme: dark)" # 深色 scheme: slate primary: black accent: indigo toggle: icon: material/brightness-4 name: Switch to system preference icon: logo: logo previous: fontawesome/solid/angle-left next: fontawesome/solid/angle-right repo: fontawesome/brands/github edit: material/pencil view: material/eye tag: default-tag: fontawesome/solid/tag hardware-tag: fontawesome/solid/microchip software-tag: fontawesome/solid/laptop-code admonition: note: octicons/tag-16 abstract: octicons/checklist-16 info: octicons/info-16 tip: octicons/squirrel-16 success: octicons/check-16 question: octicons/question-16 warning: octicons/alert-16 failure: octicons/x-circle-16 danger: octicons/zap-16 bug: octicons/bug-16 example: octicons/beaker-16 quote: octicons/quote-16 plugins: - tags - offline - search: separator: '[\s\u200b\-_,:!=\[\: )"`/]+|\.(?!\d)|&[lg]t;|(?!\b)(?=[A-Z][a-z])' - i18n: docs_structure: suffix fallback_to_default: true reconfigure_material: true reconfigure_search: true languages: - locale: zh name: 简体中文 default: true build: true - locale: en name: English site_name: PaddleOCR Documentation link: /en/ nav_translations: Home: Home 快速开始: Quick Start 近期更新: Recently Update 模型: Model 概述: Overview PP-OCR 文本检测识别: PP-OCR 概述: Overview 快速开始: Quick Start 快速安装: Quick Installation 效果展示: Visualization 运行环境: Environment Preparation 模型库: Model 模型训练: Model Training 基本概念: Basic concepts 文本检测: Text Detection 文本识别: Text Recognition 文本方向分类器: Text Angle Classification 关键信息提取: Key Information Extraction 模型微调: Fine-tune 模型压缩: Model Compression 模型量化: Model Quantization 模型裁剪: Model Prune 知识蒸馏: Knowledge Distillation 推理部署: Model Deploy 概述: Overview 基于Python预测引擎推理: Python Inference 基于C++预测引擎推理: CPP Inference Visual Studio 2019 Community CMake 编译指南: Visual Studio 2019 Community CMake Compilation Guide 服务化部署: Sever Deployment Jetson部署: Jetson Deployment 端侧部署: Device-side Deployment 网页前端部署: Paddle.js Web Deployment Paddle2ONNX模型转化与预测: Paddle2ONNX 云上飞桨部署工具: Paddle Cloud Benchmark: Benchmark 博客: Blog paddleocr package使用说明: Paddleocr Package Instructions 多语言模型: Multi-language model 动手学OCR: Dive into OCR 切片操作: Slice PaddleOCR模型推理参数解释: PaddleOCR Model Inference Parameter Explanation 分布式训练: Distributed training 项目克隆: Project Clone 配置文件内容与生成: Configuration 如何生产自定义超轻量模型？: How To Make Your own lightweight OCR model? PP-Structure文档分析: PP-Structure 概述: Overview 快速开始: Quick Start 模型库: Model 模型训练: Model Training 基本概念: Basic concepts 版面分析: Layout Analysis 版面恢复: Recovery To Doc 表格识别: Table Recognition 关键信息提取: Key Information Extraction 推理部署: Deploy 概述: Overview 基于Python预测引擎推理: Python Inference 基于C++预测引擎推理: CPP Inference 服务化部署: Sever Deployment 博客: Blog 返回识别位置: Return Recognition Location 怎样完成基于图像数据的信息抽取任务: Key Information Extraction Pipeline 前沿算法与模型: Academic Algorithms 概述: algorithm/overview.md 文本检测算法: Text Detection Algorithms DB与DB++: DB and DB++ 文本识别算法: Text Recognition Algorithms 文本超分辨率算法: Text Super-Resolution Algorithm 公式识别算法: Formulat Recognition 端到端OCR算法: End-to-End OCR Algorithms 表格识别算法: Table Recognition Algorithms 关键信息抽取算法: Key Information Extraction Algorithms 使用PaddleOCR架构添加新算法: Add new algorithms 场景应用: Applications 数据标注与合成: Data Annotation and Synthesis 概述: Overview 其它数据标注工具: Other data annotation tools 其它数据合成工具: Others data synthesis tools 数据集: Datasets 通用中英文OCR数据集: General Chinese and English OCR dataset 手写中文OCR数据集: Handwritten Chinese OCR Dataset 垂类多语言OCR数据集: Vertical multi-language OCR dataset 版面分析数据集: Layout Analysis Dataset 表格识别数据集: Table recognition dataset 关键信息提取数据集: Key Information Extraction Dataset FAQ: FAQ 社区: Community 社区贡献: Community Contribution 附录: Appendix - locale: "null" name: Help translating build: false fixed_link: "https://github.com/PaddlePaddle/PaddleOCR/discussions/13374" - git-committers: repository: PaddlePaddle/PaddleOCR branch: main token: !!python/object/apply:os.getenv ["MKDOCS_GIT_COMMITTERS_APIKEY"] - git-revision-date-localized: enable_creation_date: true markdown_extensions: - abbr - attr_list - pymdownx.snippets - pymdownx.critic - pymdownx.caret - pymdownx.keys - pymdownx.mark - pymdownx.tilde - footnotes - def_list - md_in_html - pymdownx.tasklist: custom_checkbox: true - toc: permalink: true - pymdownx.superfences: custom_fences: - name: mermaid class: mermaid format: !!python/name:pymdownx.superfences.fence_code_format - pymdownx.emoji: emoji_index: !!python/name:material.extensions.emoji.twemoji emoji_generator: !!python/name:material.extensions.emoji.to_svg - pymdownx.highlight: anchor_linenums: true line_spans: __span pygments_lang_class: true - pymdownx.inlinehilite - pymdownx.tabbed: alternate_style: true - admonition - pymdownx.details - pymdownx.superfences - pymdownx.arithmatex: generic: true extra: social: - icon: fontawesome/brands/github link: https://github.com/PaddlePaddle/PaddleOCR - icon: fontawesome/brands/python link: https://pypi.org/project/paddleocr/ extra_javascript: - javascripts/katex.min.js - https://unpkg.com/katex@0/dist/katex.min.js - https://unpkg.com/katex@0/dist/contrib/auto-render.min.js extra_css: - https://unpkg.com/katex@0/dist/katex.min.css nav: - Home: index.md - 快速开始: quick_start.md - 近期更新: update.md - 模型: - 概览: model/index.md - 多硬件安装飞桨: - 多硬件安装飞桨: model/hardware/install_other_devices.md - 支持硬件列表: model/hardware/supported_models.md - PP-OCR 文本检测识别: - 概述: ppocr/overview.md - 快速开始: ppocr/quick_start.md - 快速安装: ppocr/installation.md - 效果展示: ppocr/visualization.md - 运行环境: ppocr/environment.md - 模型库: ppocr/model_list.md - 模型训练: - 基本概念: ppocr/model_train/training.md - 文本检测: ppocr/model_train/detection.md - 文本识别: ppocr/model_train/recognition.md - 文本方向分类器: ppocr/model_train/angle_class.md - 关键信息提取: ppocr/model_train/kie.md - 模型微调: ppocr/model_train/finetune.md - 模型压缩: - 模型量化: ppocr/model_compress/quantization.md - 模型裁剪: ppocr/model_compress/prune.md - 知识蒸馏: ppocr/model_compress/knowledge_distillation.md - 推理部署: - 概述: ppocr/infer_deploy/index.md - 基于Python预测引擎推理: ppocr/infer_deploy/python_infer.md - 基于C++预测引擎推理: ppocr/infer_deploy/cpp_infer.md - Visual Studio 2019 Community CMake 编译指南: ppocr/infer_deploy/windows_vs2019_build.md - 服务化部署: ppocr/infer_deploy/paddle_server.md - Jetson部署: ppocr/infer_deploy/Jetson_infer.md - 端侧部署: ppocr/infer_deploy/lite.md - 网页前端部署: ppocr/infer_deploy/paddle_js.md - Paddle2ONNX模型转化与预测: ppocr/infer_deploy/paddle2onnx.md - 云上飞桨部署工具: ppocr/infer_deploy/paddle_cloud.md - Benchmark: ppocr/infer_deploy/benchmark.md - 博客: - PP-OCRv3技术报告: ppocr/blog/PP-OCRv3_introduction.md - PP-OCRv4技术报告: ppocr/blog/PP-OCRv4_introduction.md - paddleocr package使用说明: ppocr/blog/whl.md - 多语言模型: ppocr/blog/multi_languages.md - 动手学OCR: ppocr/blog/ocr_book.md - Enhanced CTC Loss: ppocr/blog/enhanced_ctc_loss.md - 切片操作: ppocr/blog/slice.md - PaddleOCR模型推理参数解释: ppocr/blog/inference_args.md - 分布式训练: ppocr/blog/distributed_training.md - 项目克隆: ppocr/blog/clone.md - 配置文件内容与生成: ppocr/blog/config.md - 如何生产自定义超轻量模型？: ppocr/blog/customize.md - PP-Structure文档分析: - 概述: ppstructure/overview.md - 快速开始: ppstructure/quick_start.md - 模型库: ppstructure/models_list.md - 模型训练: - 基本概念: ppstructure/model_train/training.md - 版面分析: ppstructure/model_train/train_layout.md - 表格识别: ppstructure/model_train/train_table.md - 版面恢复: ppstructure/model_train/recovery_to_doc.md - 关键信息提取: ppstructure/model_train/train_kie.md - 推理部署: - 概述: ppstructure/infer_deploy/index.md - 基于Python预测引擎推理: ppstructure/infer_deploy/python_infer.md - 基于C++预测引擎推理: ppstructure/infer_deploy/cpp_infer.md - 服务化部署: ppstructure/infer_deploy/paddle_server.md - 博客: - 返回识别位置: ppstructure/blog/return_word_pos.md - 怎样完成基于图像数据的信息抽取任务: ppstructure/blog/how_to_do_kie.md - 前沿算法与模型: - 概述: algorithm/overview.md - 文本检测算法: - DB与DB++: algorithm/text_detection/algorithm_det_db.md - EAST: algorithm/text_detection/algorithm_det_east.md - SAST: algorithm/text_detection/algorithm_det_sast.md - PSENet: algorithm/text_detection/algorithm_det_psenet.md - FCENet: algorithm/text_detection/algorithm_det_fcenet.md - DRRG: algorithm/text_detection/algorithm_det_drrg.md - CT: algorithm/text_detection/algorithm_det_ct.md - 文本识别算法: - CRNN: algorithm/text_recognition/algorithm_rec_crnn.md - Rosetta: algorithm/text_recognition/algorithm_rec_rosetta.md - STAR-Net: algorithm/text_recognition/algorithm_rec_starnet.md - RARE: algorithm/text_recognition/algorithm_rec_rare.md - SRN: algorithm/text_recognition/algorithm_rec_srn.md - NRTR: algorithm/text_recognition/algorithm_rec_nrtr.md - SAR: algorithm/text_recognition/algorithm_rec_sar.md - SEED: algorithm/text_recognition/algorithm_rec_seed.md - SVTR: algorithm/text_recognition/algorithm_rec_svtr.md - ViTSTR: algorithm/text_recognition/algorithm_rec_vitstr.md - ABINet: algorithm/text_recognition/algorithm_rec_abinet.md - VisionLAN: algorithm/text_recognition/algorithm_rec_visionlan.md - SPIN: algorithm/text_recognition/algorithm_rec_spin.md - RobustScanner: algorithm/text_recognition/algorithm_rec_robustscanner.md - RFL: algorithm/text_recognition/algorithm_rec_rfl.md - ParseQ: algorithm/text_recognition/algorithm_rec_parseq.md - CPPD: algorithm/text_recognition/algorithm_rec_cppd.md - SATRN: algorithm/text_recognition/algorithm_rec_satrn.md - 文本超分辨率算法: - Text Gestalt: algorithm/super_resolution/algorithm_sr_gestalt.md - Text Telescope: algorithm/super_resolution/algorithm_sr_telescope.md - 公式识别算法: - CAN: algorithm/formula_recognition/algorithm_rec_can.md - 端到端OCR算法: - PGNet: algorithm/end_to_end/algorithm_e2e_pgnet.md - 表格识别算法: - TableMaster: algorithm/table_recognition/algorithm_table_master.md - TableSLANet: algorithm/table_recognition/algorithm_table_slanet.md - 关键信息抽取算法: - VI-LayoutXLM: algorithm/kie/algorithm_kie_vi_layoutxlm.md - LayoutLM: algorithm/kie/algorithm_kie_layoutxlm.md # - LayoutLMv2: algorithm/kie/algorithm_kie_layoutxlm.md # - LayoutXLM: algorithm/kie/algorithm_kie_layoutxlm.md - SDMGR: algorithm/kie/./algorithm_kie_sdmgr.md - 使用PaddleOCR架构添加新算法: algorithm/add_new_algorithm.md - 场景应用: - 概述: applications/overview.md - 通用: - 高精度中文场景文本识别模型SVTR: applications/高精度中文识别模型.md - 手写体识别: applications/手写文字识别.md - 制造: - 数码管识别: applications/光功率计数码管字符识别.md - 液晶屏读数识别: applications/液晶屏读数识别.md - 包装生产日期: applications/包装生产日期识别.md - PCB文字识别: applications/PCB字符识别.md - 金融: - 表单VQA: applications/多模态表单识别.md - 增值税发票: applications/发票关键信息抽取.md - 印章检测与识别: applications/印章弯曲文字识别.md - 通用卡证识别: applications/快速构建卡证类OCR.md - 合同比对: applications/扫描合同关键信息提取.md - 交通: - 车牌识别: applications/轻量级车牌识别.md - 数据标注与合成: - 概述: data_anno_synth/overview.md - 其它数据标注工具: data_anno_synth/data_annotation.md - 其它数据合成工具: data_anno_synth/data_synthesis.md - 数据集: - 通用中英文OCR数据集: datasets/datasets.md - 手写中文OCR数据集: datasets/handwritten_datasets.md - 垂类多语言OCR数据集: datasets/vertical_and_multilingual_datasets.md - 版面分析数据集: datasets/layout_datasets.md - 表格识别数据集: datasets/table_datasets.md - 关键信息提取数据集: datasets/kie_datasets.md - FAQ: FAQ.md - 社区: - 社区贡献: community/community_contribution.md - 附录: community/code_and_doc.md