Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

crnn训练,怎么合并多种字体的模型比较好? #9821

Closed
nissansz opened this issue Apr 25, 2023 · 13 comments
Closed

crnn训练,怎么合并多种字体的模型比较好? #9821

nissansz opened this issue Apr 25, 2023 · 13 comments
Assignees
Labels
help wanted this issue needs help status/close training this is a training related issue triaged this issue has been looked, and triaged.

Comments

@nissansz
Copy link

请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem

  • 系统环境/System Environment:win10
  • 版本号/Version:Paddle:2.5 PaddleOCR: 问题相关组件/Related components:
  • 运行指令/Command Code:
  • 完整报错/Complete Error Message:

crnn训练,怎么合并多种字体的模型比较好?

有没有识别字体名的模型?

@ToddBear ToddBear added the good first issue Good for newcomers label Jun 30, 2023
@livingbody
Copy link
Contributor

可以多种字体一起训练,合并数据集。
字体识别模型还么有,我没见到。

@shiyutang
Copy link
Collaborator

已有多种模型的话,可以参考模型集成。

@nissansz
Copy link
Author

nissansz commented Jul 2, 2023

怎么集成?
印刷扫描体有什么好的模拟生成程序吗

@shiyutang
Copy link
Collaborator

可以使用styleText进行模拟生成:https://github.com/PaddlePaddle/PaddleOCR/tree/release/2.6/StyleText

@shiyutang shiyutang reopened this Jul 2, 2023
@nissansz
Copy link
Author

nissansz commented Jul 2, 2023

目前日文字幕模型,模拟字幕效果还可以。但用这个模型识别扫描文件,反而会有错误,有时是英文,数字错误。
不知道怎么改善

@shiyutang
Copy link
Collaborator

可以给出运行的指令和效果吗?

@nissansz
Copy link
Author

nissansz commented Jul 3, 2023

就是0O之类错误,或多字,少字,错字

@livingbody
Copy link
Contributor

就是0O之类错误,或多字,少字,错字

可以针对易错字构造数据集继续进行训练。

@nissansz
Copy link
Author

nissansz commented Jul 3, 2023

你有训练好的模型可以分享吗

@livingbody
Copy link
Contributor

我没训。

@nissansz
Copy link
Author

nissansz commented Jul 8, 2023

resnet18, resnet34等都能训练crnn?
哪个好?

@shiyutang
Copy link
Collaborator

模型越大精度越高,但是需要的数据也越多。

@jzhang533 jzhang533 added triaged this issue has been looked, and triaged. training this is a training related issue help wanted this issue needs help and removed good first issue Good for newcomers status/close labels Apr 10, 2024
@UserWangZz
Copy link
Collaborator

该issue长时间未更新,暂将此issue关闭,如有需要可重新开启。

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
help wanted this issue needs help status/close training this is a training related issue triaged this issue has been looked, and triaged.
Projects
None yet
Development

No branches or pull requests

8 participants