Can i download pytorch 1.9 with jetpack 6.2.1 in jetson orin nano?

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

i have jetson orin nano with jetpack 6.2.1.
i want to install pytorch 1.9.0 in it with python 3.9 or lower.
can build it from the source github ?

2

*** Please note that this reply is generated by LLM automatically ***

To install PyTorch 1.9 on your Jetson Orin Nano with JetPack 6.2.1, you have a few options. Before we dive into the installation process, let’s ensure your device meets the necessary requirements.

Hardware Requirements:
Your Jetson Orin Nano should have a minimum of 8GB of memory and a 128GB (or larger) NVMe drive. Additionally, you’ll need a USB-C flashing cable, a monitor, keyboard, and mouse for the setup process.

Software Requirements:
You’ve mentioned that you’re running JetPack 6.2.1, which is compatible with PyTorch 1.9. You’ll also need to ensure you have the necessary dependencies installed.

Installing PyTorch 1.9:
There are two primary methods to install PyTorch 1.9 on your Jetson Orin Nano:

  1. Pre-built Wheels:
    You can try installing PyTorch using pre-built wheels, which are available for Jetson devices. This method is quicker and more straightforward. You can use the following command:
pip3 install --extra-index-url https://developer.download.nvidia.com/compute/redist/jp/v46/pytorch pytorch==1.9.0+cu11.4

This command installs PyTorch 1.9 with CUDA 11.4 support.

  1. Building from Source:
    If you prefer to build PyTorch from source, you can follow these steps:
# Install dependencies
sudo apt-get update
sudo apt-get install -y libopenblas-base libopenmpi-dev libomp-dev

# Clone the PyTorch repository
git clone --recursive https://github.com/pytorch/pytorch.git

# Checkout the PyTorch 1.9 branch
cd pytorch
git checkout v1.9.0

# Build PyTorch
python setup.py install

This process will take longer, as it builds PyTorch from source.

Comparison of Methods:
Both methods have their advantages. The pre-built wheel method is faster and more convenient, while building from source provides more control over the installation process.

Additional Tips:

  • Ensure you have the necessary dependencies installed before installing PyTorch.
  • If you encounter issues during the installation process, you can try cleaning the build directory and restarting the installation process.
  • You can verify the PyTorch installation by running a simple test script, such as:
import torch
print(torch.__version__)

This should output the PyTorch version you just installed.

In summary, you can install PyTorch 1.9 on your Jetson Orin Nano with JetPack 6.2.1 using either pre-built wheels or by building from source. Both methods have their advantages, and you can choose the one that best suits your needs.

*** Please give the thumbs up if you get answers you like. Or provide feedback to help us improve the answer. ***

4

Hi,

On JetPack 6.2.1, it by default uses Python 3.10 and PyTorch 2.6+.

But you can build it from the source manually.
The detailed steps are already shared in the comment above.

Thanks.

5

Hi

can i download Python 3.9 or lower and build it from the source manually ?
or i switch to JetPack 5.4.4

thanks

6

Hi,

You can give it a try.

We have users that upgrades to the custom Python 3.12.
So the downgrade should work but you might need to handle some compatibility issue manually.

Thanks.