Skip to content

Latest commit

 

History

History
19 lines (15 loc) · 887 Bytes

README.md

File metadata and controls

19 lines (15 loc) · 887 Bytes
Raw

Deployment with TensorRT

We provide code for deployment with TensorRT python API. In general, once you use NVIDIA GPU in your applications, TensorRT is the best choice for deployment, rather than training frameworks like TensorFlow, PyTorch, MXNet, Caffe...

Prerequirements

Refer to inference_speed_evaluation for details.

Getting Started

  1. usr to_onnx.py to generate onnx model file
  2. run predict_tensorrt.py to do inference based on the generated model file
  3. after you fully understand the code, you may reform and merge it to your own project.

In most practical cases, C++ is the primary choice for efficient running. So you can rewrite the code according to the python code structure. In the future, we will provide C++ version.

NVIDIA Jetson NANO&TX2 Deployment Instructions

TBD