Calling scikit-learn functions from C++

Question

Is there way to call scikit-learn's functions from c++? I have the rest of most my code in C++ with opencv. I would like to be able use the classifiers scikit-learn provides. As far as I understand, there's no easy way - I need to use boost::python or swig. I also came across this project (https://github.com/spillai/numpy-opencv-converter) which shows interop between numpy arrays <==> cv::Mat objects, but I know only how to use to call C++ code from my python script, not the other way around.

Andreas Mueller · Accepted Answer · 2015-05-14 12:49:57Z

10

Have you looked at native C++ libraries like shogun and mlpack? These might be much easier for you to use.

answered May 14, 2015 at 12:49

Andreas Mueller

28.9k8 gold badges65 silver badges78 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Cloud Cho Over a year ago

Please, share if there are any new C++ tools since your answer in 2015, thanks.

Avrdan Over a year ago

Does Shogun support DBSCAN?

Andreas Mueller Over a year ago

@Avrdan it doesn't look like it but MLPack does: mlpack.org/doc/stable/cli_documentation.html#mlpack_dbscan just check both of their documentations.

The Nomadic Coder · Accepted Answer · 2022-05-24 13:00:10Z

8

Following up from the original answer, There seem to be two approaches today to solve the problem

A mechanism to wrap python into C++ and call it: A modern approach to do this is with PyBind11
Use native C++ ML libraries - MLPack seems to be the gold standard, and OpenCV for vision based ML.

answered May 24, 2022 at 13:00

The Nomadic Coder

6701 gold badge6 silver badges17 bronze badges

Comments

Ando Saabas · Accepted Answer · 2015-05-14 12:47:03Z

5

You can do it in a pretty straightforward way, by including python headers and just calling your python script and/or scikit methods via Py* wrappers.

See https://docs.python.org/2/extending/embedding.html#pure-embedding for a thorough example.

answered May 14, 2015 at 12:47

Ando Saabas

1,99715 silver badges12 bronze badges

Comments

Tacticus · Accepted Answer · 2023-12-17 22:59:01Z

0

For some supported models, it may be possible to convert them to ONNX runtime an run that via C++.

convert the sklearn model (pipeline) to ONNX (see here) and
run ONNX-runtime in C++ (see here)

edited Dec 17, 2023 at 22:59

answered Dec 17, 2023 at 22:36

Tacticus

57111 silver badges25 bronze badges

Collectives™ on Stack Overflow

Calling scikit-learn functions from C++

4 Answers 4

3 Comments

Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

3 Comments

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related