/onnx2trt

Sample Jupyter Notebook Explains How to Export ONNX Model to TensorRT Engine

Primary LanguageJupyter Notebook

onnx2trt

A sample Jupyter notebook to explain how to export ONNX model to NVIDIA TensorRT engine.

Target platform

  • NVIDIA Jetson

Prerequisites

  • NVIDIA Jetson Developer Kit with JetPack
    • Tested with Jetson Xavier NX Developer Kit with JetPack 4.6

Installation

  1. Clone this repository.

    git clone https://github.com/MACNICA-CLAVIS-NV/onnx2trt

  2. Build the Docker cotainer image contains the required softwares for this Jupyter notebook.

    ./scripts/docker_build.sh

Usage

  1. Run a container from the build image.

    ./scripts/docker_run.sh

  2. Launch Jupyter Lab with the instruction displyed at the above step.

  3. Open the onnx2trt.ipynb notebook from the /onnx2trt path in the Jupyter file browser.