ONNX
An open standard format for neural networks so you can train in one framework and run inference in another.
Think of it like
Like PDF for neural networks — save once, read anywhere.
Example
Train a model in PyTorch, export it as ONNX, and run it in C++ or JavaScript without rewriting.
How it actually works
ONNX is a graph format that represents computation as an acyclic graph of operations. Runtimes like ONNX Runtime, TensorRT, and others can optimize and run it. It doesn't lock you to one framework, though exporting and importing always involves some friction.
For product teams
ONNX lets you decouple model development from deployment technology choices.
For engineers
ONNX is solid for deploying PyTorch and TensorFlow models; gaps still exist for newest ops.
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome