Infrastructure / Compilers & Model Optimization

TensorRT

TensorRT is NVIDIA's SDK for compiling trained networks into optimized inference engines for NVIDIA GPUs, applying layer fusion, kernel selection and explicit quantization. This repository holds the open-source components: the ONNX parser, the plugin library, the Python bindings and samples. The optimizing builder itself ships as a prebuilt binary release that the repository links against.

Verified 2026-08-18 via the GitHub API and the LICENSE body.

Openness

2 high confidence

2.0

license: Apache-2.0(OSI)
source: partial

Apache-2.0 license body confirmed. Source is `partial` rather than `public`: the Apache-2.0 license covers the parsers, plugins, bindings and samples in this repository, and the README's build instructions require downloading a prebuilt TensorRT release tarball and pointing TRT_LIBPATH at it, so the optimizing builder that is the product is not in the published tree. The formula settles a partial source on `source` alone, so nothing about the license can lift it.

https://github.com/NVIDIA/TensorRT/blob/main/LICENSE recorded 2026-08-18
LICENSE file is the verbatim Apache-2.0 text
https://api.github.com/repos/NVIDIA/TensorRT recorded 2026-08-18
Repo metadata - license spdx_id Apache-2.0, private false, archived false, default branch main - for NVIDIA/TensorRT.
https://github.com/NVIDIA/TensorRT/blob/main/README.md recorded 2026-08-18
README's build instructions require downloading a prebuilt TensorRT release tarball from developer.nvidia.com and pointing TRT_LIBPATH at its lib directory, so the optimizing builder is linked as a binary rather than built from this repository, which holds the parsers, plugins, bindings and samples.

Adoption

3 high confidence

3.0

189,754 PyPI downloads of `tensorrt` in the trailing 30 days, which lands in the 100K-1M band of the software usage scale, level 3.

https://pypistats.org/api/packages/tensorrt/recent recorded 2026-08-18
last_month downloads = 189,754 for tensorrt

Capability

4 medium confidence

4.0

Banded on the category feature matrix as full compilation narrowed to one vendor's targets. Placed one band below the apache-tvm anchor, on a matrix that bands on how much of the model-to-hardware transformation pipeline a product performs, over how many inputs and targets.

https://github.com/NVIDIA/TensorRT/blob/main/README.md recorded 2026-08-18
README still documents the open source components of the TensorRT SDK, and build instructions that require downloading a prebuilt TensorRT release tarball and setting TRT_LIBPATH.

Verified 2026-08-18