Real time object detection of DPU yolo3

Hi, I am now using DPU on ultra96 v2 pynq board.
DPU is vary awesome,so i want to run the testing data with real time camera on yolo3 model.
I have successfully change the testing data from only one picture of dog to my own real time webcam.
I capture the first 100 frames and try to test the image each.
But when i run the code, it always appear kernel restarting error.

Does someone meets the same problem?
Is that the calculation too large to do real time detection?
Any advice is welcome!
Thanks for your help~

You can check the jupyter log to find out the real reason why it is crashing. I think the location is in /var/log/jupyter.log or something similar.

OK, I will give it a try.
Thanks for your help!