Unable to run Custom Model BEAGLEBOARD AI-64

Hi everyone!!!

I trained a custom model using modelmaker but when I try to run my model I keep getting this error. Has anyone had a similar issue? I trained the model then imported the model to modelzoo and use sudo app_edgeai to run it. I run the vision_apps scripts and get no errors.

Hi @karenulloar
There is something wrong the the Gstreamer pipeline.
What is the pipeline are you using ?

1 Like

Hello! Thank you for your response. I am using kmssink. I haven’t been able to run the model :confused: I am using a usb cameraa

gstreamer is awefull to work with. It might not fixe your issue but check how many Bytes your trying to output. Because the USB ports are limited to something like 400 MB and if your trying to output at high rate and/or high resolution, it can create issues.
You can check the USB port speed with “lsusb -t”. I couldn’t use multiple cameras at the same time because of it.

1 Like

Can you provide your yaml config file and command-line used to invoke app_edgeai? I’m looking for the path, environment and other details.

It would also be helpful to run sudo beagle-version and provide the output from that.

I will re-train and share the results, I think there is something wrong with my model.