Is there a TensorRT version of DetectNet as well as a non-TensorRT version of it?

Hi everyone,

As the title states, I want to run DetectNet with and without TensorRT to see how effective the TensorRT conversion is. When you run ./detectnet-camera (https://github.com/dusty-nv/jetson-inference#locating-object-coordinates-using-detectnet) with the pretrained models like coco-dog, are they already converted?

Thank you,
Lorenzo