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

Instructions to Build Model Source into Binaries #39

Open
DaGooseYT opened this issue Jan 1, 2022 · 4 comments
Open

Instructions to Build Model Source into Binaries #39

DaGooseYT opened this issue Jan 1, 2022 · 4 comments

Comments

@DaGooseYT
Copy link

Hi,

Thank you for creating this port, It is very helpful
I see the information on how to compile the repo from source. But assuming that the information isn't classified, do you have instructions on how to build the RIFE models into binaries? I'm referring to the contextnet.bin , contextnet.param , etc. in /models/rife-vXX/ folders.

Thanks

@yavuzbehic
Copy link

ywah im looking for thiss too

@styler00dollar
Copy link

styler00dollar commented Nov 19, 2022

It requires manual work.

  • Edit PyTorch architecture to replace warp with another operator since grid_sample isn't supported in ncnn. Replace the warp function.
  • Either export into jit and then ncnn with pnnx or export with torch.onnx and then convert onnx into bin and param.
  • Manually modify the param file and swap the replaced warp with real warp.

@yavuzbehic
Copy link

It requires manual work.

* Edit PyTorch architecture to replace warp with another operator since grid_sample isn't supported in ncnn. Replace the warp function.

* Either export into jit and then ncnn with pnnx or export with torch.onnx and then convert onnx into bin and param.

* Manually modify the param file and swap the replaced warp with real warp.

how are these other people are using it directly ?

Also im using portable, didn't build anything. Can I still use the models ?

@styler00dollar
Copy link

The question was on how to create the models (bin/param) itself. For how to use the code you can refer to the readme.

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

3 participants