Jetpack Image

[tawfiq1200]

@n-snyder

While setting up YOLO V4 on the Jetson, I have understood that some packages such as opencv, cuda and cudnn which come preinstalled on the default Jetpack are missing. They are required to use the Jetson's GPU and build darknet which is required to run YOLO V4 which is our preferred computer vision algorithm. I have installed opencv which is not that difficult but I have not installed cuda and cudnn yet. I have not found an easy of doing it but we may eventually find. However, I wonder if we actually could use the image that has all these preinstalled. Did you use a custom image from Nvidia or did you actually remove those packages to save space?

I think it will be nice at this point for us to have instructions to flash the Jetson if it is needed. I know how to do it with the developer's kit but I do not know how to do it with the module. We may need it at any point so it is good to have those instructions. I can setup docker and everything else.

[n-snyder]

Jetpack 4.6 is installed with the base packages per nVidia's documentation, which probably does not include CUDA. I did not remove any packages.

You are right that you cannot simply use the dev kit image on a production module. The image must be compatible with the devices available on the carrier board.

[tawfiq1200]

I think we need to know how to flash the eMMC. I checked the Jetpack's about section. It has a GPU. I am not sure why the base package does not have CUDA as it is needed to make the GPU work.

…

On Sun, Aug 7, 2022 at 10:32 PM Nate ***@***.***> wrote: Jetpack 4.6 is installed with the base packages per nVidia's documentation, which probably does not include CUDA. I did not remove any packages. You are right that you cannot simply use the dev kit image on a production module. The image must be compatible with the devices available on the carrier board. — Reply to this email directly, view it on GitHub <#29 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADLRIJG6C6IX6BJUG33XGBLVYBWVHANCNFSM55ZVF22Q> . You are receiving this because you authored the thread.Message ID: ***@***.***>