The ONNX Runtime has the following features:
- Supports ONNX Opset version 18, ONNX Runtime 1.16.0 and ONNX version 1.13
- Supports C++ and Python APIs (Python version 3)
- In addition to the Vitis AI EP, you can incorporate other execution providers such as ACL EP to accelerate the inference with AMD DPU
- Supports computation on the Arm® 64 Cortex-A72 cores and the supported target is VEK280 in Vitis AI 3.5
The ONNX Runtime offers the following advantages:
- Versatility
- You could deploy subgraphs on the AMD DPU while using other execution providers such as the Arm NN and Arm ACL for additional operators. This flexibility enables the deployment of models that may not be directly supported on targets boards.
- Improved performance
- By leveraging specialized execution providers such as the AMD DPU for certain operations and using other providers for the remaining operators, you can achieve optimized performance for their models.
- Expanded model support
- Enhancing ONNX Runtime enables the deployment of models with operators that are not natively supported by the DPU. By incorporating additional execution providers, you can execute a wide range of models, including those from the ONNX model zoo.