ROCm/AMDMIGraphX

AMD's graph optimization engine.

What it solves

AMD MIGraphX is designed to accelerate machine learning model inference. It provides a graph inference engine that optimizes the execution of ML models on AMD hardware.

How it works

It functions as a graph inference engine, utilizing the ROCm software stack (including MIOpen, rocBLAS, and HIP) to execute models. It supports reading ONNX files via Protobuf and provides both C/C++ and Python APIs for integration into applications.

Who it’s for

This tool is for developers and machine learning engineers who need to optimize and run ML model inference on AMD GPUs.

Highlights

  • AMD Hardware Acceleration: Specifically optimized for the ROCm ecosystem.
  • Multi-Language Support: Offers both C/C++ and Python bindings.
  • ONNX Compatibility: Supports reading models in the ONNX format.
  • Flexible Deployment: Provides multiple build options including binaries, CMake, and Docker containers.