Hello,
I installed onnxruntime 1.10.0 on i.MX 8M-Plus. I'd like to executes inference on the NPU using Python. I see that the available providers are 'NnapiExecutionProvider', 'VsiNpuExecutionProvider', 'CPUExecutionProvider'. As follows:
Python 3.9.4 (default, Apr 4 2021, 18:23:51) [GCC 10.2.0] on linux Type "help", "copyright", "credits" or "license" for more information. >>> import onnxruntime as rt >>> rt.get_available_providers() ['NnapiExecutionProvider', 'VsiNpuExecutionProvider', 'CPUExecutionProvider']
When I use 'VsiNpuExecutionProvider', we get an error like: RuntimeError: Unknown Provider Type: VsiNpuExecutionProvider.
The example using " /usr/bin/onnxruntime-1.10.0/onnx_test_runner" with vsi_npu works well. How can I use VsiNpuExecutionProvider in python correctly?
Exact same problem here.
In case of the C++ API, the provider name 'vsi-npu' would be applied. but what about Python API? What is the exact name that it has been defined on the API?
I know the VSI NPU execution provider has limitations, but in my case I can't find VSI NPU execution provider. I would like to know how I can find and run it in Python. Thank you.