Cannot install PyTorch on Jetson Xavier NX Developer Kit

[PatrickS]

Hi,

I am trying to install pytorch on the Xavier.
I have followed the following instructions for the basic installation:

Jetson Xavier NX Dev Kit

I installed PyTorch according to these instructions (I get the same errors with the source files of PyTorch)

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…

And After trying to import pytorch with python3, the following error arises

import torch
Traceback (most recent call last):
File “”, line 1, in
File “/home/jetson/.local/lib/python3.6/site-packages/torch/init.py”, line 81, in
from torch._C import *
ImportError: libcudart.so.10.0: cannot open shared object file: No such file or directory

I have also read the following thread, but don’t know how to get JetPack 4.2 onto the developer kit:

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…

Thanks in advance

[dusty_nv]

@PatrickS

ImportError: libcudart.so.10.0: cannot open shared object file: No such file or directory

Hi @PatrickS, if you are getting that error, you downloaded a version of PyTorch for a previous version of JetPack.

Use this one for JetPack 4.4 production release (L4T R32.4.3)

• Python 3.6 - torch-1.6.0-cp36-cp36m-linux_aarch64.whl

[PatrickS]

Works! Thanks for your answer.

Is it still possible to install earlier versions of torch because they are required as dependencies for other software packages, such as v1.4.0?

[dusty_nv]

@PatrickS

Is it still possible to install earlier versions of torch because they are required as dependencies for other software packages, such as v1.4.0?

It’s not possible to install earlier versions of PyTorch < 1.6 on JetPack 4.4 production release (L4T R32.4.3), because there were changes that needed to be made in PyTorch 1.6 to support the latest cuDNN 8.0 release that is in JetPack 4.4.