Compatibility Issues with PyTorch and torchvision on Jetson Nano 2GB Developer Kit

I am currently working with the Jetson Nano 2GB Developer Kit and am running into compatibility issues when attempting to use PyTorch and torchvision with Python 3.8. I have successfully changed the default Python version to 3.8 on my Jetson Nano, but I’m encountering errors suggesting that PyTorch and torchvision are not supported or not compatible.

The specific issue is related to using these libraries with YOLOv8 from Ultralytics. I have downgraded torchvision , but I’m still facing challenges with PyTorch compatibility.

Has anyone successfully set up a similar environment on the Jetson Nano 2GB Developer Kit with Python 3.8, PyTorch, torchvision, and YOLOv8? If so, could you please share the steps you followed or any workarounds you implemented?

Additionally, any insights into the compatibility of PyTorch and torchvision with Python 3.8 on the Jetson Nano would be highly appreciated.

Hi,

We don’t have python3.8 PyTorch prebuilt for JetPack 4.
Could you share more about how you set up the environment?

What kind of issue do you meet when running Ultralytics’ YOLOv8 model?
Thanks.

Hi,
thanks for the reply . I got the error in my code is

Hi,

Which TorchVision version do you use?

Since your PyTorch is 1.8.0, please checkout the v0.9.0 for compatibility.
You can find the detailed building command in the below topic:

Thanks.

The yolov5 is now working on the python 3.6 i tried to run the yolov8 found problems

Hi,

So would you mind checking the TorchVision version in your environment?
Thanks.

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