Cuda enable using pytorch

[Keshav.A]

Jetson Nano 4gb developer kit

• Jetpack : 4.5.1
• Deepstream : 5.1
• Ubuntu : 18.04
• Cuda Version : 10.2.89
• TensorRt - 7.1.3.1
• Python version - 3.10
• pyTorch version - 2.0.1 Installed from official website using
  
  

I am unable to access GPU using pytorch

import torch
torch.cuda.is_available()

False

torch.version.cuda

None

torch.cuda.current_device()

AssertionError: Torch not compiled with CUDA enabled

How can i use pytorch to use GPU in order to perform detection using yolov6 and tracking using deepsort ?

[AastaLLL]

Hi,

PyTorch doesn’t support the Jetson GPU officially.

Please try the prebuilt package from our website.
You will need to install a package that is built with the same JetPack version as your environment.
https://docs.nvidia.com/deeplearning/frameworks/install-pytorch-jetson-platform/index.html

Thanks.

[Keshav.A]

I follow the steps from the link and install pytroch but the output results are same as i mentioned in the post. Please help, i want to use GPU with pytorch.

[dusty_nv]

@Keshav.A uninstall PyTorch first, and you can find more detailed install instructions for PyTorch on Jetson Nano in this topic:

PyTorch for Jetson

Below are pre-built PyTorch pip wheel installers for Jetson Nano, TX1/TX2, Xavier, and Orin with JetPack 4.2 and newer. Download one of the PyTorch binaries from below for your version of JetPack, and see the installation instructions to run on your Jetson. These pip wheels are built for ARM aarch64 architecture, so run these commands on your Jetson (not on a host PC). You can also use the pre-built l4t-pytorch and l4t-ml container images and Dockerfiles. PyTorch pip wheels JetPack 5

PyTo…

If you still are having problems after that, I would recommend trying the l4t-pytorch container which already has PyTorch with CUDA support installed:

NVIDIA L4T PyTorch | NVIDIA NGC