Installing TensorRT in a Conda env on Jetson Orin Nano

Hi everyone,
I’m using a Jetson Orin Nano 8GB, I observed that the Jetpack consists TRT and the corresponding CUDA and PyCuda packages and are available for use after flashing.
I wish to use TRT from a miniconda env but TRT and PyCuda are not discoverable when running a program from the env, hence do I need to install TRT and PyCuda again in this env or is there a way to use the already installed packages in the Conda env as well.

Please let me know if my understanding of this is correct.

Any advise in this direction would be very helpful.

Hi,

Do you use the same Python version in the virtual environment?
If yes, you can create the env with -system-site-packages.

Thanks.

Hi @AastaLLL ,
My apologies for the delayed response. But how should I enable this for a conda env? I didn’t find a way to do that

Hi,

Please see below for the info:

Thanks.

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