Crate axonml_onnx 

Axonml ONNX - ONNX Import/Export for ML Models

This crate provides support for importing and exporting models in the ONNX (Open Neural Network Exchange) format, enabling interoperability with PyTorch, TensorFlow, and other ML frameworks.

Features

• Import ONNX models and convert to Axonml modules
• Export Axonml models to ONNX format
• Support for common operators (Conv, MatMul, ReLU, etc.)
• Weight loading from ONNX initializers

Example

use axonml_onnx::{OnnxModel, import_onnx};

// Import an ONNX model
let model = import_onnx("model.onnx")?;

// Run inference
let output = model.forward(&input);

// Export back to ONNX
model.export_onnx("model_out.onnx")?;

@version 0.1.0 @author AutomataNexus Development Team

Re-exports

pub use error::OnnxError;

pub use error::OnnxResult;

pub use model::OnnxModel;

pub use parser::import_onnx;

pub use parser::import_onnx_bytes;

pub use export::export_onnx;

Modules

error

ONNX Error Types

export

ONNX Export

model

ONNX Model Representation

operators

ONNX Operator Implementations

parser

ONNX Parser

proto

ONNX Protocol Buffer Definitions

Constants

ONNX_IR_VERSION

ONNX IR version.

SUPPORTED_OPSET_VERSION

ONNX opset version supported by this crate.