Error while running deepstream_nvdsanalytics.py

Hi,
I have trained YOLOv3 object detection model with resnet10 backbone in my ubuntu machine by following the nvidia documentation: (TLT Pretrained Object Detection | NVIDIA NGC). I have exported the model to int8 calibration and was trying to use the output .etlt file on my jetson nano for running inference using nvdsanalytics plugin. The inference was working fine in my ubuntu machine using the same .etlt file.
But in my jetson nano with deepstream 5.1, while I am running inference using deepstream_nvdsanalytics.py as following command:
“python3 deepstream_nvdsanalytics.py file:///home/mahesh-jetson/deepstream-samples/TestVideo/test.mp4”
it is showing the following error:

The config file used for the purpose:


Can you please help me to solve the above mentioned error?

please check Tlt-converter error > UffParser : Unsupported number of Graph 0 - #7 by Morganh for error - “UffParser : Unsupported number of Graph 0”

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