UiPath Documentation
document-understanding
2024.10
false
重要 :
新发布内容的本地化可能需要 1-2 周的时间才能完成。
UiPath logo, featuring letters U and I in white

Document Understanding 用户指南

上次更新日期 2026年4月6日

ML 包离线安装

入门指南

根据要使用的模型,您需要满足以下条件:

  • 对于 2022.10 及更高版本的模型:
    • Download the needed Document UnderstandingTM bundle. Here are the links for all the available bundles. The du bundle contains information about all models included into a specific version. For example, the dusemistructured-2024.10.0.tar.gz contains information about all out-of-the-box pre-trained ML Packages included in the 2024.10.0 version.
  • 对于 2022.4 及更早版本的模型(python37duv3python37duv4):
    • 所有 ML 包都以 .zip 文件的形式提供,并在 AI Center 中作为自定义包直接上传。要下载模型,请联系您的客户经理、CSM 或支持团队,以获取每个包的下载链接。
    • Download the needed Document Understanding bundle. Here are the links for all the available bundles.
备注:

Updated versions for the dulv bundles are no longer released. The latest version for the dulv bundles is 2023.10.4.

安装离线捆绑包

离线安装要求在命令行中将下载的 DU 捆绑包重命名为 du-ondemand.tar.gz。例如,如果您下载了名为 dusemistructured-2024.10.0.tar.gz 的 DU 捆绑包,则需要在安装时将其重命名为 du-ondemand.tar.gz

  1. 对于 Windows 计算机,请通过捆绑包链接直接下载,并将文件重命名为 du-ondemand.tar.gz

  2. 对于 Linux 计算机,请在可以访问互联网的计算机上使用以下命令下载所需的捆绑包:

    wget -O ~/<bundle-name.tar.gz> 'bundle-link'
    wget -O ~/<bundle-name.tar.gz> 'bundle-link'
    

    以下示例说明了如何下载适用于 Linux 的所需捆绑包:

    wget -O ~/du-ondemand.tar.gz 'https://download.uipath.com/automation-suite/2024.10.0/dusemistructured-2024.10.0.tar.gz'
    wget -O ~/du-ondemand.tar.gz 'https://download.uipath.com/automation-suite/2024.10.0/dusemistructured-2024.10.0.tar.gz'
    
  3. 将以下捆绑包复制到集群的主计算机(进行安装的位置)上的 /uipath/tmp 文件夹:

    scp ~/<bundle-name.tar.gz> <username>@<node dns>:/uipath/tmp/
    scp ~/<bundle-name.tar.gz> <username>@<node dns>:/uipath/tmp/
    
  4. 连接到此主计算机并加载捆绑包:

    ./configureUiPathAS.sh registry upload --optional-offline-bundle "/uipath/tmp/du.tar.gz" --offline-tmp-folder "/uipath/tmp"
    ./configureUiPathAS.sh registry upload --optional-offline-bundle "/uipath/tmp/du.tar.gz" --offline-tmp-folder "/uipath/tmp"
    

Upload the model to AI Center

After downloading and installing the models, follow the steps described in the ML packages offline installation page from the AI Center User Guide to upload them to AI Center. Both ML package zip and metadata json files are needed for this procedure.

表单提取程序和智能关键字分类器

Access Form Extractor and Intelligent Keyword Classifier, with the following public URL:

  • <FQDN>/du_/svc/formextractor
  • <FQDN>/du_/svc/intelligentkeywords
备注:

使用公共 URL 时,请用实际的环境信息替换<FQDN>占位符。例如,在工作流中使用时, <FQDN>/du_/svc/formextractor将变为https://servicefabricserver.domain.com/du_/svc/formextractor

将 Document Understanding™ 捆绑包上传到外部 Docker 注册表

请按照以下步骤将 Document Understanding 捆绑包上传到外部 Docker 注册表:

  • Pull the desired DU image from the UiPath® registry hosted on registry.uipath.com
  • 根据您的 Docker 注册表名称重命名映像主机。
  • 将映像推送到外部 Docker 注册表

拉取所需的 Document Understanding 映像

Pull the images from UiPath® registry by running these commands:

docker pull <uipath_registry_server>/<image_name>
docker pull <uipath_registry_server>/<image_name>

以下示例说明了如何从名为 registry.uipath.com 的注册表中拉取 UiPath 文档 OCR 捆绑包的映像:

docker pull registry.uipath.com/aicenter/du-doc-ocr:v24.10-10.3-rc02
docker pull registry.uipath.com/aicenter/du-doc-ocr:v24.10-10.3-rc02

重命名映像主机

通过运行以下命令,重命名映像主机:

docker tag <uipath_registry_server>/<image_name> <your_registry_server>/<image_name>
docker tag <uipath_registry_server>/<image_name> <your_registry_server>/<image_name>

以下示例说明了如何将 UiPath 文档 OCR 捆绑包的映像从名为 registry.uipath.com 的注册表重命名为名为 registry.mycompany.com 的注册表:

docker tag registry.uipath.com/aicenter/du-doc-ocr:v24.10-10.3-rc02 registory.mycompany.com/aicenter/du-doc-ocr:v23.10.0
docker tag registry.uipath.com/aicenter/du-doc-ocr:v24.10-10.3-rc02 registory.mycompany.com/aicenter/du-doc-ocr:v23.10.0

将映像推送到外部 Docker 注册表

通过运行以下命令,将映像推送到外部 Docker 注册表:

docker push <your_registry_server>/<image_name>
docker push <your_registry_server>/<image_name>

以下示例说明了如何将 UiPath 文档 OCR 捆绑包的映像推送到外部 Docker 注册表:

docker push registory.mycompany.com/aicenter/du-doc-ocr:v24.10.0
docker push registory.mycompany.com/aicenter/du-doc-ocr:v24.10.0

每个 Document Understanding 捆绑包的映像

2024.10.8

Document Understanding 捆绑包图像
UiPath 文档 OCRaicenter/du-doc-ocr:v24.10-2.24-rc07
UiPathDocumentOCR_CPUaicenter/du-doc-ocr-cpu:v24.10-2.24-rc07
扩展语言 OCRdu/du-extended-ocr-proxy:v24.10-3.10-rc03 du/uipath-ocr-extended:v24.10-3.10-rc03 du/du-extended-ocr-reporting:v24.10-3.10-rc03
文档分类器aicenter/du-ml-document-type-text-classifier:v24.10-2.24-rc07
Out-of-the-box Pre-trained ML Packagesaicenter/du-semistructured:v24.10-2.24-rc07

2024.10.7

Document Understanding 捆绑包图像
UiPath 文档 OCRaicenter/du-doc-ocr:v24.10-11.18-rc02
UiPathDocumentOCR_CPUaicenter/du-doc-ocr-cpu:v24.10-11.18-rc02
扩展语言 OCRdu/du-extended-ocr-proxy:v24.10-12.10-rc02 du/uipath-ocr-extended:v24.10-12.10-rc02 du/du-extended-ocr-reporting:v24.10-12.10-rc02
文档分类器aicenter/du-ml-document-type-text-classifier:v24.10-11.18-rc02
Out-of-the-box Pre-trained ML Packagesaicenter/du-semistructured:v24.10-11.18-rc02

2024.10.6

Document Understanding 捆绑包图像
扩展语言 OCRdu/du-extended-ocr-proxy:v24.10-10.31-rc02 du/uipath-ocr-extended:v24.10-10.31-rc02 du/du-extended-ocr-reporting:v24.10-10.31-rc02

2024.10.5

Document Understanding 捆绑包图像
UiPath 文档 OCRaicenter/du-doc-ocr:v24.10-8.25-rc02
UiPathDocumentOCR_CPUaicenter/du-doc-ocr-cpu:v24.10-8.25-rc02
扩展语言 OCRdu/du-extended-ocr-proxy:v24.10-9.11-rc02 du/uipath-ocr-extended:v24.10-9.11-rc02 du/du-extended-ocr-reporting:v24.10-9.11-rc02
文档分类器aicenter/du-ml-document-type-text-classifier:v24.10-8.25-rc02
Out-of-the-box Pre-trained ML Packagesaicenter/du-semistructured:v24.10-8.25-rc02

2024.10.4

Document Understanding 捆绑包图像
UiPath 文档 OCRaicenter/du-doc-ocr:v24.10-5.15-rc03
UiPathDocumentOCR_CPUaicenter/du-doc-ocr-cpu:v24.10-5.15-rc03
扩展语言 OCRdu/du-extended-ocr-proxy:v24.10-6.05-rc02 du/uipath-ocr-extended:v24.10-6.05-rc02 du/du-extended-ocr-reporting:v24.10-6.05-rc02
文档分类器aicenter/du-ml-document-type-text-classifier:v24.10-5.15-rc03
Out-of-the-box Pre-trained ML Packagesaicenter/du-semistructured:v24.10-5.15-rc03

2024.10.3

Document Understanding 捆绑包图像
UiPath 文档 OCRaicenter/du-doc-ocr:v24.10-3.11-rc01
UiPathDocumentOCR_CPUaicenter/du-doc-ocr-cpu:v24.10-3.11-rc01
扩展语言 OCRdu/du-extended-ocr-proxy:v24.10-4.20-rc02 du/uipath-ocr-extended:v24.10-4.20-rc02 du/du-extended-ocr-reporting:v24.10-4.20-rc02
文档分类器aicenter/du-ml-document-type-text-classifier:v24.10-3.11-rc01
Out-of-the-box Pre-trained ML Packagesaicenter/du-semistructured:v24.10-3.11-rc01

2024.10.2

Document Understanding 捆绑包图像
UiPath 文档 OCRaicenter/du-doc-ocr:v24.10-1.27-rc02
UiPathDocumentOCR_CPUaicenter/du-doc-ocr-cpu:v24.10-1.27-rc02
扩展语言 OCRdu/du-extended-ocr-proxy:v24.10-2.10-rc01 du/uipath-ocr-extended:v24.10-2.10-rc01 du/du-extended-ocr-reporting:v24.10-2.10-rc01
文档分类器aicenter/du-ml-document-type-text-classifier:v24.10-1.27-rc02
Out-of-the-box Pre-trained ML Packagesaicenter/du-semistructured:v24.10-1.27-rc02

2024.10.1

Document Understanding 捆绑包图像
UiPath 文档 OCRaicenter/du-doc-ocr:v24.10-11.21-rc12
UiPathDocumentOCR_CPUaicenter/du-doc-ocr-cpu:v24.10-11.21-rc12
扩展语言 OCRdu/du-extended-ocr-proxy:v24.10-12.03-rc04 du/uipath-ocr-extended:v24.10-12.03-rc04 du/du-extended-ocr-reporting:v24.10-12.03-rc04
文档分类器aicenter/du-ml-document-type-text-classifier:v24.10-11.21-rc12
Out-of-the-box Pre-trained ML Packagesaicenter/du-semistructured:v24.10-11.21-rc12

2024.10.0

Document Understanding 捆绑包图像
UiPath 文档 OCRaicenter/du-doc-ocr:v24.10-10.03-rc02
UiPathDocumentOCR_CPUaicenter/du-doc-ocr-cpu:v24.10-10.03-rc02
扩展语言 OCRdu/du-extended-ocr-proxy:v24.10-10.26-rc01 du/uipath-ocr-extended:v24.10-10.26-rc01 du/du-extended-ocr-reporting:v24.10-10.26-rc01
文档分类器aicenter/du-ml-document-type-text-classifier:v24.10-10.03-rc02
Out-of-the-box Pre-trained ML Packagesaicenter/du-semistructured:v24.10-10.17-rc02

此页面有帮助吗?

连接

需要帮助? 支持

想要了解详细内容? UiPath Academy

有问题? UiPath 论坛

保持更新