[英]Google Document AI Labeling Task
Here everyone, I am fairly new google cloud console.大家好,我是相当新的谷歌云控制台。 I am trying to customize a google document ai model that will learn to extract different sections of document to various data.
我正在尝试自定义一个谷歌文档 ai model,它将学习将文档的不同部分提取到各种数据。 As you can on the image that it fails to train the model and I have been running the Labeling Task for several days now I have not seen progress.
正如您在图像上看到的那样,它无法训练 model,而且我已经运行标记任务好几天了,现在我还没有看到任何进展。 Can you please assist in telling what is the right way to customize google document ia model enter image description here
您能否协助说明自定义谷歌文档的正确方法是什么 ia model enter image description here
I have tried to manually label the different sections of the document, it took me a while so I did around 20 test and training dataset which I think the model to not train then I decided to do the Labeling Task as an alternative to manually labeling the dataset.我尝试手动 label 文档的不同部分,这花了我一段时间所以我做了大约 20 个测试和训练数据集,我认为 model 不训练然后我决定做标签任务作为手动标记的替代方法数据集。
Here is the information about Labeling Tasks for Document AI这是有关 Document AI 标签任务的信息
Labeling Tasks use Human-in-the-Loop to have human labelers label documents for training data or production review. Labeling Tasks 使用Human-in-the-Loop让人工标记器 label 文档用于训练数据或生产审查。 You can either set up your own labeling team or apply for access to theGoogle-Managed Workforce .
您可以建立自己的标签团队或申请访问Google 管理的员工队伍。
However, it doesn't seem like this is the correct course of action for what you are trying to do, since the labeling has already been completed.但是,这似乎不是您尝试执行的正确操作过程,因为标签已经完成。 Could you provide more clarification on what you are trying to accomplish with the Custom Document Extractor?
您能否就您尝试使用自定义文档提取器完成的工作提供更多说明?
Note: Some of the error messages that are output from Document AI Workbench are not very descriptive (eg Internal Error Occurred
) but the product development team is working to surface more helpful errors when possible.注意:来自 Document AI Workbench 的一些错误消息 output 的描述性不强(例如,
Internal Error Occurred
),但产品开发团队正在努力尽可能地发现更有帮助的错误。
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.