IVGCVSW-5826 Change weights layout for depthwise to [1,H,W,I*M]

 * This change is necessary because tflite uses a [1,H,W,I*M] format
   and uses the I*M dimension for per axis quantization. Our previous
   layout [M,I,H,W] can't handle the correlating quantization scales.
 * Updates Onnx-, TfLiteParser and TfliteDelegate
 * Updates the CpuRef, CpuAcc and GpuAcc backends
 * Adjusts unit tests
 * Adds test to ensure models with old layout can still be read and
   executed
 * Adds conversion function to previous layout [1,H,W,I*M] --> [M,I,H,W]
   which can be used by backend developers

!android-nn-driver:5553

Signed-off-by: Jan Eilers <jan.eilers@arm.com>
Change-Id: Ifef23368b8c3702cf315a5838d214f7dc13c0152

1 files changed, 1 insertions, 1 deletions

diff --git a/src/armnn/test/optimizations/FoldPadTests.cpp b/src/armnn/test/optimizations/FoldPadTests.cpp
index 7b4ac4170f..11f09e80e0 100644
--- a/src/armnn/test/optimizations/FoldPadTests.cpp
+++ b/src/armnn/test/optimizations/FoldPadTests.cpp
@@ -687,7 +687,7 @@ TEST_CASE("FoldPadLayerIntoDepthwiseConv2dLayer_ExecuteInferenceWithAndWithoutOp
     // avoided. The output tensors of each should match.
     const unsigned int inputShape[]   = {1, 4, 4, 3}; // NHWCin
     const unsigned int paddedShape[]  = {1, 6, 6, 3};
-    const unsigned int weightsShape[] = {4, 3, 2, 2};  // MCinHW
+    const unsigned int weightsShape[] = {1, 2, 2, 12};  // 1HWCout
     const unsigned int outputShape[]  = {1, 5, 5, 12}; // NHWCout
 
     std::vector<float> inputData({2.0f, 2.0f, 6.0f, 6.0f,