ConvTranspose error with DirectML
Okay, here's the updated reply, including a request for a proper DirectML conversion:
Subject: Re: ConvTranspose error with DirectML
Hi everyone,
I'm experiencing the same issue as others in this thread: a Non-zero status code returned while running ConvTranspose node error when using the DirectML execution provider. Everything works fine with CPU, but DirectML consistently fails at the /N.1/pool/ConvTranspose node.
Error Message: Non-zero status code returned while running ConvTranspose node. Name:'/N.1/pool/ConvTranspose' Status Message: D:\a_work\1\s\onnxruntime\core\providers\dml\DmlExecutionProvider\src\MLOperatorAuthorImpl.cpp(2778)\onnxruntime_pybind11_state.pyd!00007FF861BA79AA: (caller: 00007FF861B737E5) Exception(2) tid(3de0) 80070057 The parameter is incorrect.
Here are my details:
Model Opset: 17
DirectML Version: 1.19.2
Hardware: AMD GPU (Unable to verify CUDA functionality in WSL2)
OS: Windows
It seems there's an incompatibility between the input parameters of the ConvTranspose operator and the DirectML execution provider in my environment. if you anyone have a workaround or a fix highly aprreciate it. thanks