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

Sharing the checkpoint #1

Closed
Cuistiano opened this issue Jun 1, 2023 · 9 comments
Closed

Sharing the checkpoint #1

Cuistiano opened this issue Jun 1, 2023 · 9 comments

Comments

@Cuistiano
Copy link

Hi there! Thanks for your great work! Do you have any plan to release the checkpoint in the near future?

@Parskatt
Copy link
Owner

Parskatt commented Jun 1, 2023

I think it should load automatically when you load from the model zoo? You can find the raw urls hardcoded in the repo.

@Parskatt
Copy link
Owner

Parskatt commented Jun 1, 2023

See here https://github.com/Parskatt/RoMa/blob/main/roma/models/model_zoo/__init__.py

@Cuistiano
Copy link
Author

I get it. Thanks!

@YJonmo
Copy link

YJonmo commented Oct 10, 2023

Hi there,

Thanks for this great work. Is there TensorRT version of the weights available somewhere?

Cheers,
Yaqub

@Parskatt
Copy link
Owner

Hi there,

Thanks for this great work. Is there TensorRT version of the weights available somewhere?

Cheers,
Yaqub

There is not. How can I make that conversion, is there a simple way?

@YJonmo
Copy link

YJonmo commented Oct 10, 2023

There are several ways of doing it:

  1. convert the pth to onnx and then onnx to trt format. For example How to convert custom .pth Model to .onnx? danielgatis/rembg#193

  2. convert the pth directly to trt format: https://forums.developer.nvidia.com/t/how-to-directly-convert-a-trained-pytorch-model-into-tensorrt-model-there-are-already-pth-files-ready-after-training/240929/2

@Parskatt
Copy link
Owner

Cool. Ill give it a try when I have time. Therr is some slightly complicated logic at infrrencr time thst might make it difficult to convert.

@YJonmo
Copy link

YJonmo commented Oct 11, 2023

Thank you. I knew it was not possible to convert the LoFTR into the .trt formant. I hope this is not the same.

@Parskatt
Copy link
Owner

Parskatt commented Oct 11, 2023 via email

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