I have tlt model and converted to trt engine using deepstream sdk.
Now deepstream sdk has some issue to run successfully.
So I like to use the converted engine in TRT.
Since the engine was converted using the same device (AGX Xavier), I can directly load the engine to TRT program.
In TRT , need to know input layer name
float* hostDataBuffer = static_cast<float*>(buffers.getHostBuffer(mParams.inputNodeName));
What is the input layer name for the network Resnet18 trained with DetectNet V2 ?