The input tensor cannot be reshaped to the requested shape.

[NewtonLiuD]

Issue summary

WinRT originate error - 0x80004005 : 'Non-zero status code returned while running Reshape node. Name:'conv5-2_Reshape' Status Message: D:\a_work\1\s\onnxruntime\core\providers\cpu\tensor\reshape_helper.h:42 onnxruntime::ReshapeHelper::ReshapeHelper gsl::narrow_cast<int64_t>(input_shape.Size()) == size was false. The input tensor cannot be reshaped to the requested shape. Input shape:{128,1,128}, requested shape:{1,128}'。

Steps to reproduce

I used Windows ML to evaluate mtcnn det1.onnx passed, but det2. onnx,det3.onnx did not pass, and some methods found on the Internet did not solve the problem; The problem may occur during caffe's model convert to ONNX.
https://github.com/CongWeilin/mtcnn-caffe
https://github.com/htshinichi/caffe-onnx

Tried solutions

microsoft/onnxruntime#2779
https://nietras.com/2021/05/24/set-dynamic-batch-size-using-onnx-sharp/
microsoft/onnxruntime#2872

System configuration

• Operating system: windows10
• Compiler: Microsoft C++
• Visual Studio version: 2019
• CUDNN version (if applicable):
• Python version: 3.6