C++ API

This is the API documentation for the NVIDIA TensorRT library. The NVIDIA TensorRT C++ API allows developers to import, calibrate, generate, and deploy networks using C++. Networks can be imported directly from ONNX. They can also be created programmatically by instantiating individual layers and setting parameters and weights directly.

Interface - TensorRT C++ API

See also

Using the C++ API

Developer guide with end-to-end examples for building and running engines.

NVIDIA TensorRT Migration Guide

Primary path: TensorRT 10.x to 11.x; appendix for TensorRT 8.x to 10.x.

Troubleshooting

Common errors, FAQs, and debugging guidance.

Python API

The Python equivalent of this API reference.