AMD Contributing MIGraphX/ROCm Back-End To Microsoft's ONNX Runtime For Machine Learning

AMD is adding a MIGraphX/ROCm back-end to Microsoft's ONNX run-time for machine learning inferencing to allow for Radeon GPU acceleration.

Microsoft's open-source ONNX Runtime as a cross-platform, high performance scoring engine for machine learning models is finally seeing AMD GPU support. This project has long supported NVIDIA TensorRT and CUDA along with Intel's APIs around DNNL / nGraph / OpenVINO, and even the ARM Compute Library. But now this Linux / Windows / macOS machine learning run-time will be able to support Radeon Open eCosystem (ROCm) for Radeon GPU acceleration on Linux.

The ONNX Runtime code from AMD is specifically targeting ROCm's MIGraphX graph optimization engine.

This AMD ROCm/MIGraphX back-end for ONNX is being reviewed here. Hopefully it will be released soon along with a new Radeon Open eCosystem release that formally adds in MIGraphX.

Those wanting to learn about the Microsoft ONNX Runtime itself can visit the GitHub project site.