Jetson orin nx

Robotics & Edge Computing Jetson & Embedded Systems Jetson Orin NX
1

When I use YOLOV11 environment on Jetson and use YOLO’s built-in onnx to engine file, I get an error message saying “Attributeerror: module” tensorrt “has no attrivute” __version __ ". My environment: jetson linux36.2，jetpack6.0DP ，python3.10，torch2.2.0，torchvision0.17.1， Can use trtexec tool to convert file format, but cannot use YOLO for engine file inference

2

In addition, I am using the Conda virtual environment

4

Hi,

As JetPack 6 DP is a preview software, could you upgrade the environment to any GA version and try it again?

Thanks.

5

Thank you, do you mean I use jetpack 6.0 or 6.1？

6

Thank you very much for your prompt answer. If I have any new questions, can I continue to consult with you?

7

Hi,

You can file a topic anytime but is your issue fixed after upgrading to JetPack 6.0GA or 6.1GA?

Thanks.

8

Thank you. When I was trying to install Jetpack 6.0, I encountered a compilation error while uninstalling the built-in OpenCV and installing OpenCV with CUDA. Therefore, I am not sure if this issue still exists. Here, I would like to ask a question. If I do not uninstall the built-in OpenCV and install OpenCV with CUDA, but instead create a virtual environment and install Torch, can I use “print (torch. CUDA. is-available ())” in Python in the virtual environment to get “True”? If possible, I will not try to install OpenCV with CUDA again

9

Hi,

Support yes, you can still use GPU within the virtual environment.
You can check the below topic to get a PyTorch prebuilt with CUDA support:

Thanks.

11

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