12.1 ONNX

ONNX : Open Neural Network Exchange

Created Date: 2025-06-26

ONNX is an open format built to represent machine learning models. ONNX defines a common set of operators - the building blocks of machine learning and deep learning models - and a common file format to enable AI developers to use models with a variety of frameworks, tools, runtimes, and compilers.

12.1.1 Build Model

Building an ONNX model primarily involves defining a computational graph using ONNX operators and then serializing it into the ONNX format. While you typically train a model in a framework like PyTorch or TensorFlow and then export it to ONNX, you can also construct an ONNX graph directly using the onnx Python library.

12.1.1.1 Construct Directly

Here's how you can conceptualize building an ONNX model:

1. Define the Computational Graph:

   12.1.2 Export to ONNX Format

   12.1.3 Inference using ONNX

   12.1.4 Export to Another Framework

   Prev: 11.5 Dispatcher Next: 12.2 ExecuTorch