For secondary inference model for DeepStream SDK,
I trained a classification model and export it with my own data set. After then, I exported it for Jetson Nano.
However, the result of secondary inference always outputs first entry of label.txt file.
The problems are same with same issue
My training environments are:
- TLT docker container : nvcr.io/nvidia/tlt-streamanalytics:v2.0_py3
- model: mobilenet_v1
Jetson Nano uses DeepStream 4.0