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

GRLIR blind resolution ported to onnx. Correctly load in fp32, not fp16. Any advice for conversion? #114

Open
zelenooki87 opened this issue Oct 21, 2024 · 2 comments

Comments

@zelenooki87
Copy link

zelenooki87 commented Oct 21, 2024

Hi. Could you check why my GRLIR large blind super resolution ported model in ONNX format produces green artifacts in fp16 mode? It works correctly in fp32. I tried the parameters from holywu's vsgrlir port. I used this .safetensors model file and this conversion script. Afterward, I simplified the model with onnxsim commands.
Could you help me with conversion? because this is really nice model. :) Thanks
https://mega.nz/folder/fRlSFLxD#e4riv3cHoD5rYXM5E5_v3w

EDIT: Conversion is not possible in Chainner. have tried that first

@WolframRhodium
Copy link
Contributor

WolframRhodium commented Oct 21, 2024

Sorry, I understand the significance of the issue, but I don't have time to do that.

@zelenooki87
Copy link
Author

I understand.
If someone other can take I look at conversion process, that would be nice.

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