LPD is based on TLT detectnet_v2 network. By default, users can run tlt detectnet_v2 infer against the tlt model or trt engine. Or users can deploy the etlt model or trt engine in deepstream and run the inference.
BTW, there is postprocess code which is exposed in C++ in /opt/nvidia/deepstream/deepstream/sources/libs/nvdsinfer_customparser/nvdsinfer_custombboxparser.cpp
And also some other user write python code to run inference against the trt engine. Run PeopleNet with tensorrt