Is it possible to call any ONNX Runtime Execution provider through WinML API? I'm able to run custom DML through WinML but unable to find steps for WinML to onnxruntime EP.
Asked
Active
Viewed 194 times
1 Answers
0
thanks for the question. The supported path for the Windows ML API is to use the DML EP. This diagram illustrates how Windows ML uses the ONNX Runtime for CPU inferencing, and the DirectML EP for GPU acceleration across all Windows devices.
Adele
- 51
- 2
-
Is there way to use other EP through WinML stack as it also uses ORT as backend? – ms2579 Jan 29 '21 at 08:08