Jetpack 4.5 + tensorrt with python 3.8

Hi
so i have an Xavier NX with Jetpack 4.5 and tensorrt 7.1.3
One of the python libraries i want to use for the system i am working on requires python 3.8, so I installed python 3.8 in a virtualenv and the library worked fine but i am having a hard time running the rest of the system (which depends on Cuda and tensorrt) on python 3.8.

So my question is, is it possible to build tensorrt 7.1.3 on python 3.8 to be able to runmy system?

Hi,

You can build the Python wrapper on your own but please upgrade TensorRT to at least 8.2:

Thanks.

Ok so i just need to upgrade the current tensorrt version to 8.2 and build my own python wrapper.
Could you please advise on how to properly upgrade tensorrt? knowing that it’s not possible for me to upgrade the jetpack, i need to do it on jetpack 4.5.

Hi,

You will need to upgrade JetPack as well to meet the compatibility.

Thanks.

I can see that jetpack 4.6 has cuda 10.2, python 3.6 (same as jetpack 4.5) and tensorrt 8, shouldn’t this mean that tensorrt 8 can be built on jetpack 4.5 as well without upgrading?

Hi,

We don’t test this combination.
You might meet some compatible issues since the GPU driver version changes.

Thanks.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.