NVIDIA® TensorRT™ is an SDK for high-performance deep learning inference on NVIDIA GPUs. This repository contains the open source components of TensorRT.

ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator

Libraries for applying sparsification recipes to neural networks with a few lines of code, enabling faster and smaller models

Sparsity-aware deep learning inference runtime for CPUs

OpenVINO™ is an open-source toolkit for optimizing and deploying AI inference