Not working custom YOLOv3 object detection with DeepStream 6.0.1

[vlad88]

My board:
Nvidia Jetson Nano 2 GB
JetPack 4.6.1
DeepStream 6.0.1

I followed the instructions from here : Custom YOLO Model in the DeepStream YOLO App — DeepStream 6.1.1 Release documentation for YOLOv3, also I followed instructions from /opt/nvidia/deepstream/deepstream-6.0/sources/objectDetector_Yolo/README.txt. So, all is working good with downloaded config file and weight file for yolov3.
But, when I run DeepStream with my yolo configs and weights, it can’t detect trained objects. (Note: before, my yolo weights works good with DeepStream, but now I reflash my board, install all again, deepstream 6.0.1, and now it not working. Also, I tried my config and yolo weights to run with darknet framework and it is working. So, I think the problem is in DeepStream).

Now, I downloaded DeepStream 6.0.1 for Jetson, maybe I should try DeepStream 6.0 Early Access for Jetson ?(but this is for JetPack 4.5.1)

[kayccc]

Your topic was posted in the wrong category. I am moving this to the DeepStream SDK category for visibility.

[Amycao]

before, my yolo weights works good with DeepStream

On which deepstream version?

I tried my config and yolo weights to run with darknet framework and it is working.

Which TensorRT version you used?

[vlad88]

I solved the problem.
I reinstall DeepStream, but now with Method 2 : Using the DeepStream tar package:
And it is working now
Thank you