onnxr: Bindings to 'ONNX' Runtime

Provides native access to the 'Open Neural Network Exchange' (ONNX) Runtime <https://onnxruntime.ai/>, which is a performant engine for running machine learning models that are saved to a standardized format. Rather than interfacing with 'ONNX' via 'Python', as in the official 'onnx' package, 'onnxr' directly interfaces with the runtime's 'C++' API via 'cpp11'. Models saved to '.onnx' files can be loaded and run on various backends, including CPUs and Apple's 'CoreML' library.

Version:	0.1.2
LinkingTo:	cpp11
Suggests:	jpeg, knitr, rmarkdown, testthat (≥ 3.0.0)
Published:	2026-06-08
DOI:	10.32614/CRAN.package.onnxr
Author:	Cory McCartan [aut, cre, cph], Caleb Carr [cph] (Author of 'nativeORT' package that is the basis for this package), Microsoft Corporation [cph] (Copyright holder of src/onnxruntime headers)
Maintainer:	Cory McCartan <mccartan at psu.edu>
BugReports:	https://github.com/CoryMcCartan/onnxr/issues
License:	MIT + file LICENSE
URL:	https://corymccartan.com/onnxr/, https://github.com/CoryMcCartan/onnxr
NeedsCompilation:	yes
Language:	en-US
Materials:	README, NEWS
CRAN checks:	onnxr results

Documentation:

Reference manual:	onnxr.html , onnxr.pdf
Vignettes:	End-to-end example (source, R code)

Downloads:

Package source:	onnxr_0.1.2.tar.gz
Windows binaries:	r-devel: onnxr_0.1.2.zip, r-release: onnxr_0.1.2.zip, r-oldrel: onnxr_0.1.2.zip
macOS binaries:	r-release (arm64): onnxr_0.1.2.tgz, r-oldrel (arm64): onnxr_0.1.2.tgz, r-release (x86_64): onnxr_0.1.2.tgz, r-oldrel (x86_64): onnxr_0.1.2.tgz

Linking:

Please use the canonical form https://CRAN.R-project.org/package=onnxr to link to this page.