commit 53ef79504b4c881c572735393c2eede5fa556c46
author Jan Eilers <jan.eilers@arm.com> Wed Jun 02 12:01:25 2021 +0100
committer Jan Eilers <jan.eilers@arm.com> Wed Jun 16 11:31:42 2021 +0000
tree f6e0cd27c4d03075fa154074c5b12d7c8c3149f7
parent 77fe76bfa8cb798943821d1f3e432c228e1cdee3

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

CMakeLists.txt

diff --git a/CMakeLists.txt b/CMakeLists.txt
index ad4c17f..17785a6 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -737,6 +737,7 @@
             src/armnnDeserializer/test/DeserializeConstant.cpp
             src/armnnDeserializer/test/DeserializeConvolution2d.cpp
             src/armnnDeserializer/test/DeserializeDepthToSpace.cpp
+            src/armnnDeserializer/test/DeserializeDepthwiseConv2d.cpp
             src/armnnDeserializer/test/DeserializeDivision.cpp
             src/armnnDeserializer/test/DeserializeFill.cpp
             src/armnnDeserializer/test/DeserializeFloor.cpp