https://github.com/sdpython/onnxcustom

Tutorial on how to convert machine learned models into ONNX
https://github.com/sdpython/onnxcustom

Last synced: 18 days ago
JSON representation

Tutorial on how to convert machine learned models into ONNX

• Host: GitHub
• URL: https://github.com/sdpython/onnxcustom
• Owner: sdpython
• License: mit
• Created: 2020-06-30T11:33:46.000Z (almost 5 years ago)
• Default Branch: master
• Last Pushed: 2023-03-11T20:52:31.000Z (about 2 years ago)
• Last Synced: 2025-05-06T22:13:46.315Z (18 days ago)
• Language: Python
• Homepage: http://www.xavierdupre.fr/app/onnxcustom/helpsphinx/index.html
• Size: 6.04 MB
• Stars: 16
• Watchers: 3
• Forks: 3
• Open Issues: 5
• Metadata Files:
  • Readme: README.rst
  • Changelog: HISTORY.rst
  • License: LICENSE.txt

Awesome Lists containing this project

README

        

.. image:: https://circleci.com/gh/sdpython/onnxcustom/tree/master.svg?style=svg

    :target: https://circleci.com/gh/sdpython/onnxcustom/tree/master

.. image:: https://travis-ci.com/sdpython/onnxcustom.svg?branch=master

    :target: https://app.travis-ci.com/github/sdpython/onnxcustom

    :alt: Build status

.. image:: https://ci.appveyor.com/api/projects/status/a3sn45a2fayoxb5q?svg=true

    :target: https://ci.appveyor.com/project/sdpython/onnxcustom

    :alt: Build Status Windows

.. image:: https://codecov.io/gh/sdpython/onnxcustom/branch/master/graph/badge.svg

    :target: https://codecov.io/gh/sdpython/onnxcustom

.. image:: https://badge.fury.io/py/onnxcustom.svg

    :target: http://badge.fury.io/py/onnxcustom

.. image:: http://img.shields.io/github/issues/sdpython/onnxcustom.png

    :alt: GitHub Issues

    :target: https://github.com/sdpython/onnxcustom/issues

.. image:: https://img.shields.io/badge/license-MIT-blue.svg

    :alt: MIT License

    :target: http://opensource.org/licenses/MIT

.. image:: https://pepy.tech/badge/onnxcustom/month

    :target: https://pepy.tech/project/onnxcustom/month

    :alt: Downloads

.. image:: https://img.shields.io/github/forks/sdpython/onnxcustom.svg

    :target: https://github.com/sdpython/onnxcustom/

    :alt: Forks

.. image:: https://img.shields.io/github/stars/sdpython/onnxcustom.svg

    :target: https://github.com/sdpython/onnxcustom/

    :alt: Stars

.. image:: https://img.shields.io/github/repo-size/sdpython/onnxcustom

    :target: https://github.com/sdpython/onnxcustom/

    :alt: size

onnxcustom: custom ONNX

=======================

.. image:: https://raw.githubusercontent.com/sdpython/onnxcustom/master/_doc/sphinxdoc/source/_static/project_ico.png

    :width: 50

`documentation `_

Examples, tutorial on how to convert machine learned models into ONNX,

implement your own converter or runtime, or even train with ONNX / onnxruntime.

The function *check* or the command line ``python -m onnxcustom check``

checks the module is properly installed and returns processing

time for a couple of functions or simply:

::

    import onnxcustom

    onnxcustom.check()

The documentation also introduces *onnx*, *onnxruntime* for

inference and training.

The tutorial related to *scikit-learn*

has been merged into `sklearn-onnx documentation

`_.

Among the tools this package implements, you may find:

* a tool to convert NVidia Profilder logs into a dataframe,

* a SGD optimizer similar to what *scikit-learn* implements but

  based on *onnxruntime-training* and able to train an CPU and GPU,

* functions to manipulate *onnx* graph.

**Installation of onnxruntime-training**

onnxruntime-training is only available on Linux. The CPU

can be installed with the following instruction.

::

    pip install onnxruntime-training --extra-index-url https://download.onnxruntime.ai/onnxruntime_nightly_cpu.html

Versions using GPU with CUDA or ROCm are available. Check

`download.onnxruntime.ai `_

to find a specific version.

You can use it on Windows

inside WSL (Windows Linux Subsystem) or compile it for CPU:

::

    python tools\ci_build\build.py --skip_tests --build_dir .\build\Windows --config Release --build_shared_lib --build_wheel --numpy_version= --cmake_generator="Visual Studio 16 2019" --enable_training --enable_training_ops

GPU versions work better on WSL, see `Build onnxruntime on WSL (Windows Linux Subsystem)

`_.

*onnxcustom* can be installed from pypi.