ONNX

API to store large tensors outside ModelProto, avoiding protobuf serialization of large initializers.

ONNX backend abstraction for executing models with device support and test utilities.

Python API reference for ONNX 1.23.0, covering versioning, protobuf data structures, serialization helpers, and model utilities.

Python helper functions to create ONNX graph components.

Python SDK for serializing and deserializing ONNX models and protocol buffer objects.

Python utilities for managing external tensor data in ONNX models, including conversion, loading, and saving operations.

Utilities to merge ONNX graphs and models and add name prefixes.

Open standard for machine learning interoperability enabling model exchange between frameworks.

Python API for ONNX protobuf structures: ModelProto, GraphProto, NodeProto, TensorProto, and other message types.

Python utility library for inlining ONNX model-local and selected functions into a model graph.

Graph utilities for validating ONNX proto messages.

Python module for ONNX operator schema definitions and version management.