Deploy yolov5 on triton server for Jetson Xavier NX

Hi I want to deploy yolo v5 model on triton server. After converting the model to onxx format what steps I need to follow as in the documentation of triton peoplenet example is given but I want clarity how take make a build file for another model also there is a file peoplenet.cc file so how can I customize it for my use case.

Hi,

You can find an ONNXRuntime backend in the below folder.
Please update the model name and the corresponding input/output information to inference with Triton.

/opt/nvidia/deepstream/deepstream-6.0/samples/trtis_model_repo/densenet_onnx

Thanks.

How can I make build for my model before running triton?

Hi,

Please check if the warmup option can meet your requirement:

Thanks.