C++ version

[cduch9]

Hi, this work is interesting.

I would like to try and benchmark this algorithm from a C++ progam. The solution I have now is not performing to my satisfaction and yours might be better. Do you have a recommandation on the way to do this performance test?

Thanks.

[lxxue]

Hi, I did the benchmark in python with fvcore benchmark. An example of this can be found in this script. I hope this helps.

[cduch9]

I want to try in C++ instead of Python.

[lxxue]

Then I guess you need to rewrite frnn.py with C++ to have a C++ standalone implementation.

[cduch9]

Yes, and we must link with the C++ Distributions of PyTorch (https://pytorch.org/cppdocs/installing.html). I was kind of hoping that someone has done it before.

I will stick to Python and convert my test data instead. You can close the issue if you want. Thanks.