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

自己训练的新的模型导入后无法运行 #59

Open
ZMxyxyxy opened this issue Apr 17, 2024 · 9 comments
Open

自己训练的新的模型导入后无法运行 #59

ZMxyxyxy opened this issue Apr 17, 2024 · 9 comments

Comments

@ZMxyxyxy
Copy link

我们自己训练的新的模型导入后无法运行,我们采用的是syenet,也是基于ncnn的,导出了param和bin文件,以同样的方式导入。
my

@tumuyan
Copy link
Owner

tumuyan commented Apr 17, 2024

你好,是这样的,虽然同样是ncnn格式的模型,但是不同项目的模型可能需要重新实现推理程序。
你可以参考下nihui做的这些项目,https://github.com/tumuyan/RealSR-NCNN-Android?tab=readme-ov-file#ncnn-projects-and-models

整体的过程是图片分片,然后预处理,推理后再做后处理,最后把分片图片合并并完成存储。
虽然我做了平台移植以及凑数的gui,但是事实上并不能理解nihui的这些项目到底是怎么做到的。

@tumuyan
Copy link
Owner

tumuyan commented Apr 17, 2024

看了下你的截图,你的模型到底是几倍的?从命名上讲是一倍的,但是程序读取的时候似乎并没有正确找到1倍的模型。
如截图是我尝试运行一个正常的x1外置模型的画面
放置了 /storage/emulated/0/realsr/models-ESRGAN-Fatality_NoiseToner-sharpen_denoise/x1.param 和 /storage/emulated/0/realsr/models-ESRGAN-Fatality_NoiseToner-sharpen_denoise/x1.bin 两个文件

Screenshot_2024-04-17-11-32-08-758_com tumuyan ncnn realsr

@ZMxyxyxy
Copy link
Author

ZMxyxyxy commented Apr 17, 2024 via email

@ZMxyxyxy
Copy link
Author

ZMxyxyxy commented Apr 17, 2024 via email

@tumuyan
Copy link
Owner

tumuyan commented Apr 17, 2024

虽然希望不大,但是不应该在这一步出错。可以把文件名改为x2试试,ncnn模型可以打包上传一下我看看有没有机会。

@ZMxyxyxy
Copy link
Author

ZMxyxyxy commented Apr 17, 2024 via email

@tumuyan
Copy link
Owner

tumuyan commented Apr 17, 2024

似乎你并没有上传文件…

@ZMxyxyxy
Copy link
Author

您好,我通过邮箱发给您了,不知您是否收到

@tumuyan
Copy link
Owner

tumuyan commented Apr 17, 2024

没有收到邮件呢

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants