ppstructure 中的KIE 和 PaddleNLP中的UIE 哪个更适合用来提取版式不相同的发票关键信息 #14863
Unanswered
Kyo1234567
asked this question in
Q&A
Replies: 1 comment 1 reply
-
在你的场景——提取版式不同的发票的关键信息(字段名称和位置均不同)下,PaddleNLP 的 UIE(通用信息抽取模型) 更加适合,而 PP-Structure 中的 KIE 主要用于特定版式的文档信息提取。 对比分析
结论
推荐方案
如果你的目标是处理不同版式的发票,建议使用 UIE,因为它的能力更通用,泛化能力更强且易于应用。 Response generated by 🤖 feifei-bot | chatgpt-4o-latest |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
我发现ppstructure 中的KIE和PaddleNLP中的UIE似乎都可以用来提取发票、合同中的关键信息。如果我需要提取版式不同的发票信息,每种发票都有发票编号、日期、金额等信息,但是字段名称和位置都可能不相同。请问这种场景下,KIE 和 UIE哪个更合适?哪个精度更高?
Beta Was this translation helpful? Give feedback.
All reactions