Skip to main content

> ML_STANDARD // ONNX-OPEN-NEURAL-NETWORK-EXCHANGE-IR_v1.0

ONNX (Open Neural Network Exchange) Intermediate Representation Specification

LF AI & Data Foundation (Linux Foundation) · International Open Source Standard · active

specificationInternational Open Source Standardactive
Official StandardBack to Standards

Regulatory & Technical Framework Summary

Open standard format for representing machine learning models, enabling model interoperability across frameworks (PyTorch, TensorFlow, Scikit-Learn) and execution runtimes.

Key Compliance Requirements

  • Standardized computational graph model based on Protocol Buffers (protobuf) schema
  • Comprehensive standard operator set (Opset) specifying mathematical semantics of layers
  • Support for traditional ML models via ONNX-ML extension (trees, linear models, scalers)
  • Extensibility mechanisms for custom domain-specific hardware operators

Applicable Sectors & Tasks

Affected Sectors:
technologyautomotivemanufacturingcloud computing
Affected Tasks:
model compilationimage classificationtext generation

TinyCTO provides regulatory summaries and technical engineering alignment for informational purposes only. This content does not constitute formal legal advice or regulatory compliance certification.