Unable to install ultralytics in python3.6 in Jetpack4.6

Unable to install Ultralytics in jetson nano board with Jetack 4.6. with python 3.6

Please check Ultralytics install - Jetson & Embedded Systems / Jetson Nano - NVIDIA Developer Forums to see if can help. Thanks

The recommendation is to install Python version 3.7 or newer. However, upon attempting installation with Python 3.10.12, an error occurs while running the Python script, specifically reporting an “illegal instruction core dumped” issue.

I need to install OpenCV, NumPy, Ultralytics for YOLOv8, and Matplotlib. However, during the installation process, I encountered numerous version dependency errors. Could you please provide me with a recommended version list that is compatible with Jetson Nano development boards?


Have you checked with the Ultralytics team to see if they can support Python 3.6?

Hi, ultralytics need python >= 3.7 to run, unfortunately upgrade python version in jetson nano will have torch version issue. If you really want to use yolo maybe consider another version out side ultralytics.

Yes, it’s correct that if i upgrade python version above 3.6.9 .Their is present of torch and torchvision for jetpack 5 and above , which doesn’t supported by jetson nano board. So, can anyone suggest me the alternate methodS.

Well either use yolov3 alexy version or use ssd mobilenet in Nvidia tutorial

Ok that true but my python script doesn’t satisfy the requirements then.
I need ultralytics.


Have you tried the below link?

It seems that the yolov5 repo can work on Nano.
But it’s recommended to confirm this with the Ultralytics team.


Yes, I gone with this but it not working at all.


Maybe you can check with the Ultralytics team.
They should know more about the compatibility limitation.


Ultralytics saying they need python3.7 or above.
But when I install Python 3.8 and run python script it shows illegal instruction core dumped and also torch and torch vision are not provided for the jetpack 4.6.4 for jetson nano with python 3.8. So how i can use ultralytics with torch and torchvision with cuda support and yolo.

And also how i can use deepstream and tensorRT in jetson nano baord .


Unfortunately, we only provide Python3.6 packages for the JetPack 4 environment.
To make it work with Ultralytics easier, you can consider changing to our Orin series device that can use newer software.


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