UnitTest Failed：TensorflowLiteParser

[xiaotongnii]

sysinfo: platform:anrdoid13 Arm cortex-A76+A55 armnn:22.02

TensorflowLiteParser DepthwiseConvolution2D. test failed and MaxPool2D Segmentation violation signal.

Orid:/data/local/tmp/ut # ./UnitTests --test-suite=TensorflowLiteParser_

[doctest] doctest version is "2.4.6" [doctest] run with "--help" for options

/home/armnnsrc/armnnTfLiteParser/test/DepthwiseConvolution2D.cpp:130: TEST SUITE: TensorflowLiteParser_DepthwiseConvolution2D TEST CASE: ParseDepthwiseConv2DSame

/home/armnnsrc/armnnTfLiteParser/test/ParserFlatbuffersFixture.hpp:385: ERROR: CHECK( result.m_Result ) is NOT correct! values: CHECK( false ) logged: 4 different values at: [0,0,1,0] (17 != 18), [0,1,0,0] (28 != 29), [0,1,2,0] (55 != 56), ... (and 1 other differences)

=============================================================================== /home/armnnsrc/armnnTfLiteParser/test/DepthwiseConvolution2D.cpp:182: TEST SUITE: TensorflowLiteParser_DepthwiseConvolution2D TEST CASE: ParseDepthwiseConv2DSameBias

/home/armnnsrc/armnnTfLiteParser/test/ParserFlatbuffersFixture.hpp:385: ERROR: CHECK( result.m_Result ) is NOT correct! values: CHECK( false ) logged: 4 different values at: [0,0,1,0] (22 != 23), [0,1,0,0] (33 != 34), [0,1,2,0] (60 != 61), ... (and 1 other differences)

=============================================================================== /home/armnnsrc/armnnTfLiteParser/test/DepthwiseConvolution2D.cpp:555: TEST SUITE: TensorflowLiteParser_DepthwiseConvolution2D TEST CASE: ParseDepthwiseConv2DFilterWeightsPerChannelQuant4

/home/armnnsrc/armnnTfLiteParser/test/ParserFlatbuffersFixture.hpp:385: ERROR: CHECK( result.m_Result ) is NOT correct! values: CHECK( false ) logged: 9 different values at: [0,0,0,7] (8 != 7), [0,0,1,7] (8 != 7), [0,0,2,7] (8 != 7), ... (and 6 other differences)

=============================================================================== /home/armnnsrc/armnnTfLiteParser/test/DepthwiseConvolution2D.cpp:821: TEST SUITE: TensorflowLiteParser_DepthwiseConvolution2D TEST CASE: ParseDepthwiseConv2DFilterWeightsPerChannelQuant4_2

/home/armnnsrc/armnnTfLiteParser/test/ParserFlatbuffersFixture.hpp:385: ERROR: CHECK( result.m_Result ) is NOT correct! values: CHECK( false ) logged: 27 different values at: [0,0,0,3] (7 != 6), [0,0,0,7] (21 != 20), [0,0,0,11] (26 != 25), ... (and 24 other differences)

=============================================================================== /home/armnnsrc/armnnTfLiteParser/test/DepthwiseConvolution2D.cpp:887: TEST SUITE: TensorflowLiteParser_DepthwiseConvolution2D TEST CASE: ParseDepthwiseConv2DFilterWeightsPerChannelQuant4_5

/home/armnnsrc/armnnTfLiteParser/test/ParserFlatbuffersFixture.hpp:385: ERROR: CHECK( result.m_Result ) is NOT correct! values: CHECK( false ) logged: 118 different values at: [0,0,0,2] (3 != 4), [0,0,0,4] (9 != 10), [0,0,0,5] (11 != 12), ... (and 115 other differences)

=============================================================================== /home/armnnsrc/armnnTfLiteParser/test/MaxPool2D.cpp:108: TEST SUITE: TensorflowLiteParser_MaxPool2D TEST CASE: MaxPoolLiteUint2DOutput

/home/armnnsrc/armnnTfLiteParser/test/MaxPool2D.cpp:108: FATAL ERROR: test case CRASHED: SIGSEGV - Segmentation violation signal

=============================================================================== [doctest] test cases: 149 | 143 passed | 6 failed | 4806 skipped [doctest] assertions: 975 | 970 passed | 5 failed | [doctest] Status: FAILURE! Segmentation fault (core dumped)

[ArmRyan]

Hello @Shelton-N ,

We don't officially support Android 13 yet. The first release that will likely support Android 13 will probably be 22.11.

Can I ask what NDK you are building with?

Thanks, Ryan

[xiaotongnii]

@ArmRyan , I compiled armnn22.02 with ndk-r20, Just test in Android13.