ONNX - AshokBhat/ml GitHub Wiki

About

• Stands for "Open Neural Network Exchange"
• Original developers: Microsoft and Meta

ONNX Format

• Open format to represent deep learning models
• Enables models to be trained in one framework and transferred to another for inference
• Supported in Microsoft Cognitive Toolkit, MXNet, and PyTorch

Governance

• Graduate-level project in Linux Foundation
• Steering committee : Microsoft, Meta, Amazon, NVIDIA and Intel

ONNX runtime

• Inference engine for models in the ONNX format
• Open source project, released by Microsoft

See Also

• Frameworks: [TensorFlow]] ](/AshokBhat/ml/wiki/[PyTorch) | [MXNET]] | ONNX
• [Inference engine]]s: [[ArmNN]] ](/AshokBhat/ml/wiki/[OpenVINO) | [TensorRT]] | Core ML