A backend which allows the usage of ONNX MLIR compiled models (model.so) with the Triton Inference Server. - View it on GitHub