Nvidia Jetson AGX Orin not able to use GPU

Robotics & Edge Computing Jetson Systems Jetson AGX Orin
,
1

Recently, I tried training my custom YOLO model, but the training process was very slow because my GPU wasn’t being utilized. I’ve installed CUDA and tried multiple approaches found online to enable GPU acceleration, but none have worked. Even though CUDA is installed and available, the GPU remains idle during training, significantly increasing the training time. Can you help me resolve this issue?
image

image
image510×124 16 KB

image
image842×441 40.4 KB

2

Hello,

Thanks for visiting the NVIDIA Developer forums! Your topic will be best served in the Jetson category.

I will move this post over for visibility.

Cheers,
Tom

4

Thanks Tom !!!

6

Hi,

Please find below the link for the PyTorch package with CUDA support:
https://elinux.org/Jetson/L4T/TRT_Customized_Example#YoloV11_using_Pytorch

Thanks.

7

Hello,
Do I have to setup a virtual environment to install torch-2.5.0 torchvision-0.20.0 ?

9

Hi,

No. The step is optional.
It’s easy for some users who want to try multiple PyTorch versions.

Thanks.