miniONNXRuntime

A minimal ONNX Runtime implementation that walks through the full inference pipeline: model loading, graph construction, and operator execution.

// repository documentation