AI/ML Interoperability Standard
20.7k 2026-04-27
onnx/onnx
An open standard and format for machine learning models, enabling interoperability across different AI frameworks and hardware.
Core Features
Provides an open-source, standardized format for AI models (deep learning and traditional ML).
Enables interoperability across diverse machine learning frameworks, tools, and hardware.
Defines an extensible computation graph model, built-in operators, and standard data types.
Offers programming utilities for graph optimization, shape/type inference, and version conversion.
Detailed Introduction
ONNX (Open Neural Network Exchange) is an open ecosystem that establishes a standardized, open-source format for both deep learning and traditional machine learning models. Its primary goal is to foster interoperability, allowing AI developers to effortlessly transfer and deploy models across various frameworks, tools, and hardware. By defining a common computation graph model, built-in operators, and standard data types, ONNX significantly streamlines the research-to-production workflow, accelerating innovation within the AI community, particularly for inferencing tasks.