burn-onnx is a crate designed to simplify the process of importing models trained in other machine learning frameworks into the Burn framework via the ONNX format. It generates the Rust source code that aligns the imported model with Burn and converts tensor data into a compatible format.

burn-onnx

Library for importing ONNX models into the Burn deep learning framework.

Overview

burn-onnx converts ONNX models into native Burn code, enabling seamless integration of pre-trained models from popular ML frameworks (PyTorch, TensorFlow, etc.) into the Burn ecosystem.

Links

• Documentation
• Supported ONNX Operators
• Development Guide
• ONNX Tests
• Burn Book