ONNX & TensorRT

WIP deep-learning deep-learning deployment-serving dl dl-optimization inference onnx optimization optimization-compute tensorrt 1 min read

Exporting and optimizing deep learning models for production

Overview

An open standard format for representing ML models, allowing interoperability between PyTorch, TensorFlow, etc.

NVIDIA’s SDK for high-performance deep learning inference, performing graph optimizations and layer fusion.

TODO: Add examples of exporting PyTorch to ONNX and compiling to TensorRT.