onnxconverter-common

What it is and what it does

onnxconverter-common is a utility library that sits at the intersection of multiple machine learning framework converters, providing shared functions and abstractions for translating models to ONNX format. Rather than being a converter itself, it acts as a foundation that different framework-specific converters build upon—allowing them to work together when a model combines components from multiple frameworks (like a scikit-learn pipeline that includes an xgboost model). The package depends on numpy, onnx, packaging, and protobuf, making it a lightweight addition to existing ONNX workflows.

The library is maintained by Microsoft, actively developed, and supports Python 3.8 through 3.13 on Linux, Windows, and macOS. It carries an MIT License, making it suitable for both open-source and commercial use. With nearly 2 million monthly downloads and a position in the top 5000 PyPI packages, it is a stable, widely-adopted component in the ONNX ecosystem.

Use it for:

• Build a converter that translates scikit-learn models to ONNX using shared utility functions
• Enable a pipeline that combines xgboost and scikit-learn models and exports them together to ONNX
• Develop a custom framework converter that leverages common ONNX conversion patterns and helpers
• Optimize model export workflows by reusing standardized conversion logic across multiple frameworks
• Integrate model format conversion into a production ML serving pipeline that targets ONNX runtime

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Provides common utilities and functions for converting machine learning models from various AI frameworks to ONNX format, enabling interoperability between different framework converters.

Yes. onnxconverter-common is a stable, actively maintained library with low install friction, no known vulnerabilities, and permissive licensing. Install it if you are building or using ONNX converters, especially when working with multiple frameworks or planning to leverage existing converter ecosystems. It is a foundational dependency rather than a standalone tool, so evaluate it in the context of your specific converter needs.

Install

onnxconverter-common on PyPI

pip install onnxconverter-common

uv add onnxconverter-common

poetry add onnxconverter-common

Installing onnxconverter-common

Before you install

Low install friction with a pure-wheel distribution. Actively maintained with a recent release (351 days ago) and ongoing repository activity. Supports current Python versions from 3.8 through 3.13.

License in practice

MIT License permits unrestricted use, modification, and distribution for both commercial and private projects with minimal restrictions—only requiring preservation of copyright and license notices.

Quickstart

pip install onnxconverter-common

import onnxconverter_common
from onnx import load

# Use converter utilities with your ONNX model
model = load('model.onnx')

Verify before relying

• Specific converter frameworks supported beyond the general statement of 'various AI frameworks'
• Whether mixed-framework conversion (e.g. scikit-learn + xgboost) is production-ready or experimental
• Performance characteristics or optimization capabilities beyond the reference to float16 documentation

Package facts

Evidence: onnxconverter_common-1.16.0-py2.py3-none-any.whl

Tags

More Artificial Intelligence packages