2. Acceleration

CUDA Execution Provider (Nvidia)

1. Uninstall dependencies:

pip uninstall onnxruntime onnxruntime-gpu

2. Install dependencies:

pip install onnxruntime onnxruntime-gpu

3. Usage in case the provider is available:

python run.py --execution-provider cuda

CoreML Execution Provider (Apple)

1. Uninstall dependencies:

pip uninstall onnxruntime onnxruntime-silicon

2. Install dependencies:

pip install onnxruntime onnxruntime-silicon

3. Usage in case the provider is available:

python run.py --execution-provider coreml

OpenVINO™ Execution Provider (Intel)

1. Uninstall dependencies:

pip uninstall onnxruntime onnxruntime-openvino

2. Install dependencies:

pip install onnxruntime onnxruntime-openvino

3. Usage in case the provider is available:

python run.py --execution-provider openvino

ROCm Execution Provider (AMD)

1. Build custom wheels from the onnxruntime source.

2. Install the custom wheels.

3. Usage in case the provider is available:

python run.py --execution-provider rocm