Classifier_meta_list is none in deepstream_test2.py

can state the exact commands the sample sgie file after training a custom model using tao, so basically I have a tlt model, how to convert it to a proper engine file or how to write a sgie file so that the script converts the etlt file to a engine file. @yingliu @yuweiw

You can use the trt model with deepstream directly. Nvinfer will generate engine files automatically if you configured the paras of the config file correctly. You can refer the link below:
https://docs.nvidia.com/metropolis/deepstream/dev-guide/text/DS_plugin_gst-nvinfer.html