Hi there, i have a jetson xavier NX and i installed the jetson sdk with version 5.1.1 jp.
I have a contianer running on this board in which i have ubuntu 20.04 and cuda 11.4.
I have to use Libtorch (c++ version of Pytorch ) which use also cuda 11.4, because in my container i use pointcloudtransport where i implemented my custom plugin that use torch in c++.
Now the question is: is there a way to install libtorch on my container which use cuda 11.4 and is supported for my jetson xavier NX?.
I successfull installed this version of pytorch but it’s releted to python and not in c++:
with version - torch-1.14.0a0+44dac51c.nv23.02-cp38-cp38-linux_aarch64.whl but in which way can i install and do i use api of torch for c++?
And last question if there is a way to use c++ version inside the python one installed by whl, in which way can i linking and use it in a CMAKETEXT where i have to use require(Torch) ?
Thank you in advance