Google Git
Sign in
android / platform / external / armnn / 44179c37 / . / src / backends / backendsCommon / test / layerTests / ConcatTestImpl.cpp
blob: 1e40b42dcfb918075298f11383b84316941fb97a [file] [log] [blame]
//
// Copyright © 2017 Arm Ltd. All rights reserved.
// SPDX-License-Identifier: MIT
//
#include "ConcatTestImpl.hpp"
#include <QuantizeHelper.hpp>
#include <ResolveType.hpp>
#include <armnnUtils/Permute.hpp>
#include <backendsCommon/test/TensorCopyUtils.hpp>
#include <backendsCommon/test/WorkloadTestUtils.hpp>
#include <test/TensorHelpers.hpp>
using namespace armnn;
using namespace armnnUtils;
//
// Helper functions and templates
//
OriginsDescriptor CreateDescriptorForConcat(
const std::vector<TensorInfo> & inputTensorInfos,
unsigned int concatDim)
{
std::vector<TensorShape> shapes;
shapes.reserve(inputTensorInfos.size());
for (const TensorInfo& it: inputTensorInfos)
{
shapes.push_back(it.GetShape());
}
return CreateDescriptorForConcatenation(shapes.begin(), shapes.end(), concatDim);
}
//
// Concat is only supported for N and C dimensions for NCHW and the inner most dimension
// In case of <4 dimensions we need to make sure that the concat dimensions are at least
// the 3rd slowest iterating one or the inner most dimension.
//
bool NeedPermuteForConcat(
const std::vector<TensorInfo> & inputTensorInfos,
unsigned int concatDim)
{
// See note above. Additionally we expect the input shapes to have the
// same number of dimensions.
unsigned int nDimensions = 0;
// Determine the number of dimensions as well as sanity check them
// agains test implementation issues.
for (auto && tensorInfo : inputTensorInfos)
{
if (!nDimensions)
{
nDimensions = tensorInfo.GetShape().GetNumDimensions();
}
else
{
BOOST_ASSERT_MSG(nDimensions == tensorInfo.GetShape().GetNumDimensions(),
"Input shapes must have the same number of dimensions");
}
}
return (nDimensions < 3 || (nDimensions == 3 && (nDimensions-concatDim) < 3 && (nDimensions-concatDim) != 1));
}
TensorShape ExpandTensorShapeTo3dForPermute(const TensorShape & inputShape)
{
unsigned int numDims = inputShape.GetNumDimensions();
if (numDims >= 3)
{
// Nothing to do if the inputShape has at least 3 dimensions.
return inputShape;
}
std::vector<unsigned int> newDims(size_t(3), 1u);
unsigned int expandedBy = 3 - numDims;
for (unsigned int i=0; i<numDims; ++i)
{
newDims[expandedBy+i] = inputShape[i];
}
return TensorShape(3u, &newDims[0]);
}
void Generate3dPermuteVectorForConcat(
unsigned int numDimensions,
unsigned int & concatDim,
std::pair<PermutationVector, PermutationVector> & permutations)
{
BOOST_ASSERT_MSG(numDimensions <= 3,
"Only dimensions 1,2 and 3 are supported by this helper");
unsigned int expandedBy = 3 - numDimensions;
unsigned int expandedConcatAxis = concatDim + expandedBy;
if (expandedConcatAxis == 2)
{
concatDim = 0;
PermutationVector forwardPermutation({1, 2, 0});
PermutationVector reversePermutation({2, 0, 1});
permutations = std::make_pair(forwardPermutation, reversePermutation);
}
else if (expandedConcatAxis == 1)
{
concatDim = 0;
PermutationVector forwardPermutation({2, 0, 1});
PermutationVector reversePermutation({1, 2, 0});
permutations = std::make_pair(forwardPermutation, reversePermutation);
}
else
{
BOOST_ASSERT(expandedConcatAxis == 0);
concatDim = 0;
}
}
template<typename T> void PermuteTensorData(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
const PermutationVector& mappings,
TensorInfo & inputTensorInfo,
const T * inputData,
std::vector<T>& outputData)
{
IgnoreUnused(memoryManager);
BOOST_ASSERT_MSG(inputData != nullptr, "inputData must not be null");
if (inputData == nullptr)
{
// Nullptr is an error in the test. By returning without doing the concatenation
// I expect the caller to fail the test. It still makes sense to report this as
// an assert for Debug builds.
return;
}
TensorInfo outputTensorInfo = armnnUtils::Permuted(inputTensorInfo, mappings);
std::unique_ptr<ITensorHandle> inputHandle = workloadFactory.CreateTensorHandle(inputTensorInfo);
std::unique_ptr<ITensorHandle> outputHandle = workloadFactory.CreateTensorHandle(outputTensorInfo);
PermuteQueueDescriptor queueDescriptor;
queueDescriptor.m_Parameters = PermuteDescriptor{mappings};
WorkloadInfo workloadInfo;
AddInputToWorkload(queueDescriptor, workloadInfo, inputTensorInfo, inputHandle.get());
AddOutputToWorkload(queueDescriptor, workloadInfo, outputTensorInfo, outputHandle.get());
std::unique_ptr<IWorkload> workload = workloadFactory.CreatePermute(queueDescriptor, workloadInfo);
inputHandle->Allocate();
outputHandle->Allocate();
CopyDataToITensorHandle(inputHandle.get(), inputData);
workload->PostAllocationConfigure();
workload->Execute();
outputData.resize(outputTensorInfo.GetNumElements());
CopyDataFromITensorHandle(&outputData[0], outputHandle.get());
inputTensorInfo = outputTensorInfo;
}
//
// Permute the input tensors so we can do a supported concatenation.
// Also treat lower than 3d tensors as 3d by adding dummy 1 dimensions
// at the front. Finally this function tells what the output shape
// of the permuted concatenated tensor is going to be.
//
template<typename T> void PermuteInputsForConcat(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
std::vector<TensorInfo> & inputTensorInfos,
std::vector<T *> & inputData,
std::vector<std::vector<T>> & inputDataStorage,
PermutationVector & permuteVector,
unsigned int & concatDim,
TensorInfo & outputTensorInfo)
{
IgnoreUnused(memoryManager);
BOOST_ASSERT_MSG(inputTensorInfos.size() > 1,
"Expecting more than one tensor to be concatenated here");
unsigned int numDims = 0;
unsigned int nthInput = 0;
const PermutationVector identity({0, 1, 2});
std::pair<PermutationVector, PermutationVector> permutations =
std::make_pair(identity, identity);
inputDataStorage.resize(inputData.size());
for (auto && tensorInfo : inputTensorInfos)
{
if (numDims == 0)
{
numDims = tensorInfo.GetShape().GetNumDimensions();
Generate3dPermuteVectorForConcat(numDims, concatDim, permutations);
// Store the reverese permutation.
permuteVector = permutations.second;
BOOST_ASSERT_MSG(!permuteVector.IsEqual(identity),
"Test logic error, we don't need permutation, so we shouldn't arrive here");
}
else
{
BOOST_ASSERT_MSG(numDims == tensorInfo.GetShape().GetNumDimensions(),
"All inputs must have the same number of dimensions");
}
TensorInfo newTensorInfo = tensorInfo;
newTensorInfo.SetShape(ExpandTensorShapeTo3dForPermute(tensorInfo.GetShape()));
PermuteTensorData<T>(workloadFactory,
memoryManager,
permutations.first,
newTensorInfo,
inputData[nthInput],
inputDataStorage[nthInput]);
inputData[nthInput] = inputDataStorage[nthInput].data();
inputTensorInfos[nthInput] = newTensorInfo;
++nthInput;
}
outputTensorInfo.SetShape(
armnnUtils::Permuted(
ExpandTensorShapeTo3dForPermute(outputTensorInfo.GetShape()),
permutations.first));
}
//
// This is the pair of PermuteInputsForConcat(...) which permutes back
// the output of the concatenation so we can check it against an expected
// output.
//
template <typename T> void PermuteOutputForConcat(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
const TensorInfo & tensorInfo,
const PermutationVector & permuteVector,
std::unique_ptr<ITensorHandle> && inputDataHandle,
T * data)
{
BOOST_ASSERT_MSG(data != nullptr, "data must not be null");
if (data == nullptr)
{
// Nullptr is an error in the test. By returning without doing the permutation
// I expect the caller to fail the test. It still makes sense to report this as
// an assert for Debug builds.
return;
}
TensorInfo resultTensorInfo = tensorInfo;
std::vector<T> inputData(tensorInfo.GetNumElements());
std::vector<T> outputData;
CopyDataFromITensorHandle(&inputData[0], inputDataHandle.get());
PermuteTensorData<T>(workloadFactory,
memoryManager,
permuteVector,
resultTensorInfo,
&inputData[0],
outputData);
::memcpy(data, &outputData[0], sizeof(T)*outputData.size());
}
template<typename T> void Concatenate(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
std::initializer_list<const TensorInfo> inputTensorInfosOrig,
std::initializer_list<T *> inputsOrig,
const TensorInfo& outputTensorInfoOrig,
T * output,
unsigned int concatDim,
bool useSubtensor)
{
BOOST_ASSERT_MSG(output != nullptr, "output must not be null");
if (output == nullptr)
{
// Nullptr is an error in the test. By returning without doing the permutation
// I expect the caller to fail the test. It still makes sense to report this as
// an assert for Debug builds.
return;
}
// Saves a copy of the parameters which we might need to change.
std::vector<TensorInfo> inputTensorInfos(inputTensorInfosOrig.begin(), inputTensorInfosOrig.end());
std::vector<T *> inputs = inputsOrig;
TensorInfo outputTensorInfo = outputTensorInfoOrig;
PermutationVector permuteVector{0, 1, 2};
// Holds and automatically releases memory for the reshaped input data.
std::vector<std::vector<T>> tmpInputDataStorage;
const size_t inputCount = inputTensorInfos.size();
bool needPermuteForConcat = NeedPermuteForConcat(inputTensorInfos, concatDim);
if (needPermuteForConcat)
{
//
// We need to permute the inputs, because concatenation along
// the requested axis is not supported.
//
PermuteInputsForConcat<T>(workloadFactory,
memoryManager,
inputTensorInfos,
inputs,
tmpInputDataStorage,
permuteVector,
concatDim,
outputTensorInfo);
}
WorkloadInfo workloadInfo;
std::vector<std::unique_ptr<ITensorHandle>> inputHandles;
inputHandles.reserve(inputCount);
std::unique_ptr<ITensorHandle> outputHandle = workloadFactory.CreateTensorHandle(outputTensorInfo);
ConcatQueueDescriptor queueDescriptor;
OriginsDescriptor viewsDescriptor = CreateDescriptorForConcat(inputTensorInfos, concatDim);
queueDescriptor.m_Parameters = viewsDescriptor;
if (useSubtensor)
{
queueDescriptor.m_ViewOrigins.reserve(viewsDescriptor.GetNumViews());
for (unsigned int i = 0; i < viewsDescriptor.GetNumViews(); ++i)
{
queueDescriptor.m_ViewOrigins.emplace_back(std::vector<unsigned int>(viewsDescriptor.GetViewOrigin(i),
viewsDescriptor.GetViewOrigin(i) + viewsDescriptor.GetNumDimensions()));
}
outputHandle = workloadFactory.CreateTensorHandle(outputTensorInfo);
const bool subTensorsSupported = workloadFactory.SupportsSubTensors();
for (unsigned int i = 0; i < inputCount; ++i)
{
const TensorInfo& inputTensorInfo = inputTensorInfos[i];
std::unique_ptr<ITensorHandle> inputHandle =
subTensorsSupported ?
workloadFactory.CreateSubTensorHandle(*outputHandle,
inputTensorInfo.GetShape(),
queueDescriptor.m_ViewOrigins[i].m_Origin.data()) :
workloadFactory.CreateTensorHandle(inputTensorInfo);
inputHandles.emplace_back(std::move(inputHandle));
}
}
else
{
for (unsigned int i = 0; i < inputCount; ++i)
{
std::unique_ptr<ITensorHandle> inputHandle = workloadFactory.CreateTensorHandle(inputTensorInfos[i]);
inputHandles.emplace_back(std::move(inputHandle));
}
}
for (unsigned int i = 0; i < inputCount; ++i)
{
AddInputToWorkload(queueDescriptor, workloadInfo, inputTensorInfos[i], inputHandles[i].get());
}
AddOutputToWorkload(queueDescriptor, workloadInfo, outputTensorInfo, outputHandle.get());
std::unique_ptr<IWorkload> workload = workloadFactory.CreateConcat(queueDescriptor, workloadInfo);
for (auto& inputHandle : inputHandles)
{
inputHandle->Allocate();
}
outputHandle->Allocate();
unsigned int nextInputId = 0;
for (auto& inputHandle : inputHandles)
{
CopyDataToITensorHandle(inputHandle.get(), inputs[nextInputId]);
++nextInputId;
}
workload->PostAllocationConfigure();
workload->Execute();
if (needPermuteForConcat)
{
PermuteOutputForConcat<T>(workloadFactory,
memoryManager,
outputTensorInfo,
permuteVector,
std::move(outputHandle),
output);
}
else
{
CopyDataFromITensorHandle(output, outputHandle.get());
}
}
//
// Implementation templates
//
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 1> Concat1dTestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
TensorInfo inputTensorInfo({ 3 }, ArmnnType, qScale, qOffset);
auto input0 = MakeTensor<T, 1>(inputTensorInfo, QuantizedVector<T>({ 1.0f, 2.0f, 3.0f }, qScale, qOffset));
auto input1 = MakeTensor<T, 1>(inputTensorInfo, QuantizedVector<T>({ 4.0f, 5.0f, 6.0f }, qScale, qOffset));
auto input2 = MakeTensor<T, 1>(inputTensorInfo, QuantizedVector<T>({ 7.0f, 8.0f, 9.0f }, qScale, qOffset));
TensorInfo outputTensorInfo({ 9 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 1> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory, memoryManager,
{ inputTensorInfo, inputTensorInfo, inputTensorInfo },
{ input0.data(), input1.data(), input2.data() },
outputTensorInfo,
output.data(),
0,
true);
result.output = MakeTensor<T, 1>(outputTensorInfo, output);
result.outputExpected = MakeTensor<T, 1>(outputTensorInfo, QuantizedVector<T>(
{
1.0f, 2.0f, 3.0f, 4.0f, 5.0f, 6.0f, 7.0f, 8.0f, 9.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 2> Concat2dTestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
const TensorInfo& outputTensorInfo,
unsigned int dimension,
const float qScale,
const int32_t qOffset)
{
TensorInfo inputTensorInfo({ 2, 3 }, ArmnnType, qScale, qOffset);
auto input0 = MakeTensor<T, 2>(inputTensorInfo, QuantizedVector<T>(
{
// Batch 0
1.0f, 2.0f, 3.0f,
// Batch 1
10.0f, 11.0f, 12.0f,
},
qScale, qOffset));
auto input1 = MakeTensor<T, 2>(inputTensorInfo, QuantizedVector<T>(
{
// Batch 0
4.0f, 5.0f, 6.0f,
// Batch 1
13.0f, 14.0f, 15.0f,
},
qScale, qOffset));
auto input2 = MakeTensor<T, 2>(inputTensorInfo, QuantizedVector<T>(
{
// Batch 0
7.0f, 8.0f, 9.0f,
// Batch 1
16.0f, 17.0f, 18.0f,
},
qScale, qOffset));
LayerTestResult<T, 2> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory, memoryManager,
{ inputTensorInfo, inputTensorInfo, inputTensorInfo },
{ input0.data(), input1.data(), input2.data() },
outputTensorInfo,
output.data(),
dimension,
true);
result.output = MakeTensor<T, 2>(outputTensorInfo, output);
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 2> Concat2dDim0TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
TensorInfo outputTensorInfo({ 6, 3 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 2> result = Concat2dTestImpl<ArmnnType>(
workloadFactory, memoryManager, outputTensorInfo, 0, qScale, qOffset);
result.outputExpected = MakeTensor<T, 2>(outputTensorInfo, QuantizedVector<T>(
{
// Batch 0
1.0f, 2.0f, 3.0f,
// Batch 1
10.0f, 11.0f, 12.0f,
// Batch 2
4.0f, 5.0f, 6.0f,
// Batch 3
13.0f, 14.0f, 15.0f,
// Batch 4
7.0f, 8.0f, 9.0f,
// Batch 5
16.0f, 17.0f, 18.0f,
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 2> Concat2dDim1TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
TensorInfo outputTensorInfo({ 2, 9 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 2> result = Concat2dTestImpl<ArmnnType>(
workloadFactory, memoryManager, outputTensorInfo, 1, qScale, qOffset);
result.outputExpected = MakeTensor<T, 2>(outputTensorInfo, QuantizedVector<T>(
{
// Batch 0
1.0f, 2.0f, 3.0f, 4.0f, 5.0f, 6.0f, 7.0f, 8.0f, 9.0f,
// Batch 1
10.0f, 11.0f, 12.0f, 13.0f, 14.0f, 15.0f, 16.0f, 17.0f, 18.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 2> Concat2dDim0DiffInputDimsTestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
TensorInfo input0TensorInfo({ 2, 3 }, ArmnnType, qScale, qOffset);
auto input0 = MakeTensor<T, 2>(input0TensorInfo, QuantizedVector<T>(
{
// Batch 0
1.0f, 2.0f, 3.0f,
// Batch 1
10.0f, 11.0f, 12.0f,
},
qScale, qOffset));
TensorInfo input1TensorInfo({ 3, 3 }, ArmnnType, qScale, qOffset);
auto input1 = MakeTensor<T, 2>(input1TensorInfo, QuantizedVector<T>(
{
// Batch 0
4.0f, 5.0f, 6.0f,
// Batch 1
13.0f, 14.0f, 15.0f,
// Batch 0
7.0f, 8.0f, 9.0f,
},
qScale, qOffset));
TensorInfo input2TensorInfo({ 1, 3 }, ArmnnType, qScale, qOffset);
auto input2 = MakeTensor<T, 2>(input2TensorInfo, QuantizedVector<T>(
{
// Batch 1
16.0f, 17.0f, 18.0f,
},
qScale, qOffset));
TensorInfo outputTensorInfo({ 6, 3 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 2> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory, memoryManager,
{ input0TensorInfo, input1TensorInfo, input2TensorInfo },
{ input0.data(), input1.data(), input2.data() },
outputTensorInfo,
output.data(),
0,
true);
result.output = MakeTensor<T, 2>(outputTensorInfo, output);
result.outputExpected = MakeTensor<T, 2>(outputTensorInfo, QuantizedVector<T>(
{
// Batch 0
1.0f, 2.0f, 3.0f,
// Batch 1
10.0f, 11.0f, 12.0f,
// Batch 2
4.0f, 5.0f, 6.0f,
// Batch 3
13.0f, 14.0f, 15.0f,
// Batch 4
7.0f, 8.0f, 9.0f,
// Batch 5
16.0f, 17.0f, 18.0f,
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 2> Concat2dDim1DiffInputDimsTestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
TensorInfo input0TensorInfo({ 2, 3 }, ArmnnType, qScale, qOffset);
auto input0 = MakeTensor<T, 2>(input0TensorInfo, QuantizedVector<T>(
{
// Batch 0
1.0f, 2.0f, 3.0f,
// Batch 1
10.0f, 11.0f, 12.0f,
},
qScale, qOffset));
TensorInfo input1TensorInfo({ 2, 5 }, ArmnnType, qScale, qOffset);
auto input1 = MakeTensor<T, 2>(input1TensorInfo, QuantizedVector<T>(
{
// Batch 0
4.0f, 5.0f, 6.0f, 7.0f, 8.0f,
// Batch 1
13.0f, 14.0f, 15.0f, 16.0f, 17.0f,
},
qScale, qOffset));
TensorInfo input2TensorInfo({ 2, 1 }, ArmnnType, qScale, qOffset);
auto input2 = MakeTensor<T, 2>(input2TensorInfo, QuantizedVector<T>(
{
// Batch 0
9.0f,
// Batch 1
18.0f
},
qScale, qOffset));
TensorInfo outputTensorInfo({ 2, 9 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 2> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory, memoryManager,
{ input0TensorInfo, input1TensorInfo, input2TensorInfo },
{ input0.data(), input1.data(), input2.data() },
outputTensorInfo,
output.data(),
1,
true);
result.output = MakeTensor<T, 2>(outputTensorInfo, output);
result.outputExpected = MakeTensor<T, 2>(outputTensorInfo, QuantizedVector<T>(
{
// Batch 0
1.0f, 2.0f, 3.0f, 4.0f, 5.0f, 6.0f, 7.0f, 8.0f, 9.0f,
// Batch 1
10.0f, 11.0f, 12.0f, 13.0f, 14.0f, 15.0f, 16.0f, 17.0f, 18.0f,
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 3> Concat3dTestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
const TensorInfo& outputTensorInfo,
unsigned int dimension,
bool useSubtensor,
float qScale,
int32_t qOffset)
{
TensorInfo inputTensorInfo({ 2, 3, 2 }, ArmnnType, qScale, qOffset);
auto input0 = MakeTensor<T, 3>(inputTensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
1.0f, 2.0f,
// Batch 0, Channel 1
3.0f, 4.0f,
// Batch 0, Channel 2
5.0f, 6.0f,
// Batch 1, Channel 0
19.0f, 20.0f,
// Batch 1, Channel 1
21.0f, 22.0f,
// Batch 1, Channel 2
23.0f, 24.0f
},
qScale, qOffset));
auto input1 = MakeTensor<T, 3>(inputTensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
7.0f, 8.0f,
// Batch 0, Channel 1
9.0f, 10.0f,
// Batch 0, Channel 2
11.0f, 12.0f,
// Batch 1, Channel 0
25.0f, 26.0f,
// Batch 1, Channel 1
27.0f, 28.0f,
// Batch 1, Channel 2
29.0f, 30.0f
},
qScale, qOffset));
auto input2 = MakeTensor<T, 3>(inputTensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
13.0f, 14.0f,
// Batch 0, Channel 1
15.0f, 16.0f,
// Batch 0, Channel 2
17.0f, 18.0f,
// Batch 1, Channel 0
31.0f, 32.0f,
// Batch 1, Channel 1
33.0f, 34.0f,
// Batch 1, Channel 2
35.0f, 36.0f
},
qScale, qOffset));
LayerTestResult<T, 3> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory, memoryManager,
{ inputTensorInfo, inputTensorInfo, inputTensorInfo },
{ input0.data(), input1.data(), input2.data() },
outputTensorInfo,
output.data(),
dimension,
useSubtensor);
result.output = MakeTensor<T, 3>(outputTensorInfo, output);
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 3> Concat3dDim0TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
TensorInfo outputTensorInfo({ 6, 3, 2 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 3> result = Concat3dTestImpl<ArmnnType>(
workloadFactory, memoryManager, outputTensorInfo, 0, true, qScale, qOffset);
result.outputExpected = MakeTensor<T, 3>(outputTensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
1.0f, 2.0f,
// Batch 0, Channel 1
3.0f, 4.0f,
// Batch 0, Channel 2
5.0f, 6.0f,
// Batch 1, Channel 0
19.0f, 20.0f,
// Batch 1, Channel 1
21.0f, 22.0f,
// Batch 1, Channel 2
23.0f, 24.0f,
// Batch 2, Channel 0
7.0f, 8.0f,
// Batch 2, Channel 1
9.0f, 10.0f,
// Batch 2, Channel 2
11.0f, 12.0f,
// Batch 3, Channel 0
25.0f, 26.0f,
// Batch 3, Channel 1
27.0f, 28.0f,
// Batch 3, Channel 2
29.0f, 30.0f,
// Batch 4, Channel 0
13.0f, 14.0f,
// Batch 4, Channel 1
15.0f, 16.0f,
// Batch 4, Channel 2
17.0f, 18.0f,
// Batch 5, Channel 0
31.0f, 32.0f,
// Batch 5, Channel 1
33.0f, 34.0f,
// Batch 5, Channel 2
35.0f, 36.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 3> Concat3dDim1TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
TensorInfo outputTensorInfo({ 2, 9, 2 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 3> result = Concat3dTestImpl<ArmnnType>(
workloadFactory, memoryManager, outputTensorInfo, 1, true, qScale, qOffset);
result.outputExpected = MakeTensor<T, 3>(outputTensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
1.0f, 2.0f,
// Batch 0, Channel 1
3.0f, 4.0f,
// Batch 0, Channel 2
5.0f, 6.0f,
// Batch 0, Channel 3
7.0f, 8.0f,
// Batch 0, Channel 4
9.0f, 10.0f,
// Batch 0, Channel 5
11.0f, 12.0f,
// Batch 0, Channel 6
13.0f, 14.0f,
// Batch 0, Channel 7
15.0f, 16.0f,
// Batch 0, Channel 8
17.0f, 18.0f,
// Batch 1, Channel 0
19.0f, 20.0f,
// Batch 1, Channel 1
21.0f, 22.0f,
// Batch 1, Channel 2
23.0f, 24.0f,
// Batch 1, Channel 3
25.0f, 26.0f,
// Batch 1, Channel 4
27.0f, 28.0f,
// Batch 1, Channel 5
29.0f, 30.0f,
// Batch 1, Channel 6
31.0f, 32.0f,
// Batch 1, Channel 7
33.0f, 34.0f,
// Batch 1, Channel 8
35.0f, 36.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 3> Concat3dDim2TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
bool useSubtensor,
float qScale,
int32_t qOffset)
{
TensorInfo outputTensorInfo({ 2, 3, 6 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 3> result = Concat3dTestImpl<ArmnnType>(
workloadFactory, memoryManager, outputTensorInfo, 2, useSubtensor, qScale, qOffset);
result.outputExpected = MakeTensor<T, 3>(outputTensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
1.0f, 2.0f, 7.0f, 8.0f, 13.0f, 14.0f,
// Batch 0, Channel 1
3.0f, 4.0f, 9.0f, 10.0f, 15.0f, 16.0f,
// Batch 0, Channel 2
5.0f, 6.0f, 11.0f, 12.0f, 17.0f, 18.0f,
// Batch 1, Channel 0
19.0f, 20.0f, 25.0f, 26.0f, 31.0f, 32.0f,
// Batch 1, Channel 1
21.0f, 22.0f, 27.0f, 28.0f, 33.0f, 34.0f,
// Batch 1, Channel 2
23.0f, 24.0f, 29.0f, 30.0f, 35.0f, 36.0f,
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 3> Concat3dDim0DiffInputDimsTestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
TensorInfo input0TensorInfo({ 2, 3, 2 }, ArmnnType);
auto input0 = MakeTensor<T, 3>(input0TensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
1.0f, 2.0f,
// Batch 0, Channel 1
3.0f, 4.0f,
// Batch 0, Channel 2
5.0f, 6.0f,
// Batch 1, Channel 0
19.0f, 20.0f,
// Batch 1, Channel 1
21.0f, 22.0f,
// Batch 1, Channel 2
23.0f, 24.0f
},
qScale, qOffset));
TensorInfo input1TensorInfo({ 1, 3, 2 }, ArmnnType);
auto input1 = MakeTensor<T, 3>(input1TensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
7.0f, 8.0f,
// Batch 0, Channel 1
9.0f, 10.0f,
// Batch 0, Channel 2
11.0f, 12.0f,
},
qScale, qOffset));
TensorInfo input2TensorInfo({ 3, 3, 2 }, ArmnnType);
auto input2 = MakeTensor<T, 3>(input2TensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
25.0f, 26.0f,
// Batch 0, Channel 1
27.0f, 28.0f,
// Batch 0, Channel 2
29.0f, 30.0f,
// Batch 1, Channel 0
13.0f, 14.0f,
// Batch 1, Channel 1
15.0f, 16.0f,
// Batch 1, Channel 2
17.0f, 18.0f,
// Batch 2, Channel 0
31.0f, 32.0f,
// Batch 2, Channel 1
33.0f, 34.0f,
// Batch 2, Channel 2
35.0f, 36.0f
},
qScale, qOffset));
TensorInfo outputTensorInfo({ 6, 3, 2 }, ArmnnType);
LayerTestResult<T, 3> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory, memoryManager,
{ input0TensorInfo, input1TensorInfo, input2TensorInfo },
{ input0.data(), input1.data(), input2.data() },
outputTensorInfo,
output.data(),
0,
true);
result.output = MakeTensor<T, 3>(outputTensorInfo, output);
result.outputExpected = MakeTensor<T, 3>(outputTensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
1.0f, 2.0f,
// Batch 0, Channel 1
3.0f, 4.0f,
// Batch 0, Channel 2
5.0f, 6.0f,
// Batch 1, Channel 0
19.0f, 20.0f,
// Batch 1, Channel 1
21.0f, 22.0f,
// Batch 1, Channel 2
23.0f, 24.0f,
// Batch 2, Channel 0
7.0f, 8.0f,
// Batch 2, Channel 1
9.0f, 10.0f,
// Batch 2, Channel 2
11.0f, 12.0f,
// Batch 3, Channel 0
25.0f, 26.0f,
// Batch 3, Channel 1
27.0f, 28.0f,
// Batch 3, Channel 2
29.0f, 30.0f,
// Batch 4, Channel 0
13.0f, 14.0f,
// Batch 4, Channel 1
15.0f, 16.0f,
// Batch 4, Channel 2
17.0f, 18.0f,
// Batch 5, Channel 0
31.0f, 32.0f,
// Batch 5, Channel 1
33.0f, 34.0f,
// Batch 5, Channel 2
35.0f, 36.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 3> Concat3dDim1DiffInputDimsTestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
TensorInfo input0TensorInfo({ 2, 3, 2 }, ArmnnType, qScale, qOffset);
auto input0 = MakeTensor<T, 3>(input0TensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
1.0f, 2.0f,
// Batch 0, Channel 1
3.0f, 4.0f,
// Batch 0, Channel 2
5.0f, 6.0f,
// Batch 1, Channel 0
19.0f, 20.0f,
// Batch 1, Channel 1
21.0f, 22.0f,
// Batch 1, Channel 2
23.0f, 24.0f
},
qScale, qOffset));
TensorInfo input1TensorInfo({ 2, 4, 2 }, ArmnnType, qScale, qOffset);
auto input1 = MakeTensor<T, 3>(input1TensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
7.0f, 8.0f,
// Batch 0, Channel 1
9.0f, 10.0f,
// Batch 0, Channel 2
11.0f, 12.0f,
// Batch 0, Channel 3
25.0f, 26.0f,
// Batch 1, Channel 0
27.0f, 28.0f,
// Batch 1, Channel 1
29.0f, 30.0f,
// Batch 1, Channel 2
13.0f, 14.0f,
// Batch 1, Channel 3
15.0f, 16.0f,
},
qScale, qOffset));
TensorInfo input2TensorInfo({ 2, 1, 2 }, ArmnnType, qScale, qOffset);
auto input2 = MakeTensor<T, 3>(input2TensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
17.0f, 18.0f,
// Batch 1, Channel 0
31.0f, 32.0f,
},
qScale, qOffset));
TensorInfo outputTensorInfo({ 2, 8, 2 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 3> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory, memoryManager,
{ input0TensorInfo, input1TensorInfo, input2TensorInfo },
{ input0.data(), input1.data(), input2.data() },
outputTensorInfo,
output.data(),
1,
true);
result.output = MakeTensor<T, 3>(outputTensorInfo, output);
result.outputExpected = MakeTensor<T, 3>(outputTensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
1.0f, 2.0f,
// Batch 0, Channel 1
3.0f, 4.0f,
// Batch 0, Channel 2
5.0f, 6.0f,
// Batch 0, Channel 3
7.0f, 8.0f,
// Batch 0, Channel 4
9.0f, 10.0f,
// Batch 0, Channel 5
11.0f, 12.0f,
// Batch 0, Channel 6
25.0f, 26.0f,
// Batch 0, Channel 7
17.0f, 18.0f,
// Batch 1, Channel 0
19.0f, 20.0f,
// Batch 1, Channel 1
21.0f, 22.0f,
// Batch 1, Channel 2
23.0f, 24.0f,
// Batch 1, Channel 3
27.0f, 28.0f,
// Batch 1, Channel 4
29.0f, 30.0f,
// Batch 1, Channel 5
13.0f, 14.0f,
// Batch 1, Channel 6
15.0f, 16.0f,
// Batch 1, Channel 7
31.0f, 32.0f,
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 3> Concat3dDim2DiffInputDimsTestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
bool useSubtensor,
float qScale,
int32_t qOffset)
{
TensorInfo input0TensorInfo({ 2, 3, 2 }, ArmnnType, qScale, qOffset);
auto input0 = MakeTensor<T, 3>(input0TensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
1.0f, 2.0f,
// Batch 0, Channel 1
3.0f, 4.0f,
// Batch 0, Channel 2
5.0f, 6.0f,
// Batch 1, Channel 0
19.0f, 20.0f,
// Batch 1, Channel 1
21.0f, 22.0f,
// Batch 1, Channel 2
23.0f, 24.0f
},
qScale, qOffset));
TensorInfo input1TensorInfo({ 2, 3, 1 }, ArmnnType, qScale, qOffset);
auto input1 = MakeTensor<T, 3>(input1TensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
7.0f,
// Batch 0, Channel 1
9.0f,
// Batch 0, Channel 2
11.0f,
// Batch 1, Channel 0
25.0f,
// Batch 1, Channel 1
27.0f,
// Batch 1, Channel 2
29.0f
},
qScale, qOffset));
TensorInfo input2TensorInfo({ 2, 3, 3 }, ArmnnType, qScale, qOffset);
auto input2 = MakeTensor<T, 3>(input2TensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
13.0f, 14.0f, 50.0f,
// Batch 0, Channel 1
15.0f, 16.0f, 51.0f,
// Batch 0, Channel 2
17.0f, 18.0f, 52.0f,
// Batch 1, Channel 0
31.0f, 32.0f, 53.0f,
// Batch 1, Channel 1
33.0f, 34.0f, 54.0f,
// Batch 1, Channel 2
35.0f, 36.0f, 55.0f,
},
qScale, qOffset));
TensorInfo outputTensorInfo({ 2, 3, 6 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 3> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory, memoryManager,
{ input0TensorInfo, input1TensorInfo, input2TensorInfo },
{ input0.data(), input1.data(), input2.data() },
outputTensorInfo,
output.data(),
2,
useSubtensor);
result.output = MakeTensor<T, 3>(outputTensorInfo, output);
result.outputExpected = MakeTensor<T, 3>(outputTensorInfo, QuantizedVector<T>(
{
// Batch 0, Channel 0
1.0f, 2.0f, 7.0f, 13.0f, 14.0f, 50.0f,
// Batch 0, Channel 1
3.0f, 4.0f, 9.0f, 15.0f, 16.0f, 51.0f,
// Batch 0, Channel 2
5.0f, 6.0f, 11.0f, 17.0f, 18.0f, 52.0f,
// Batch 1, Channel 0
19.0f, 20.0f, 25.0f, 31.0f, 32.0f, 53.0f,
// Batch 1, Channel 1
21.0f, 22.0f, 27.0f, 33.0f, 34.0f, 54.0f,
// Batch 1, Channel 2
23.0f, 24.0f, 29.0f, 35.0f, 36.0f, 55.0f,
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 4> Concat4dTestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
const TensorInfo& outputTensorInfo,
unsigned int dimension,
bool useSubtensor,
float qScale,
int32_t qOffset)
{
TensorInfo inputTensorInfo({ 1, 3, 2, 2 }, ArmnnType, qScale, qOffset);
auto input0 = MakeTensor<T, 4>(inputTensorInfo, QuantizedVector<T>(
{
1.0f, 2.0f,
3.0f, 4.0f,
5.0f, 6.0f,
7.0f, 8.0f,
9.0f, 10.0f,
11.0f, 12.0f
},
qScale, qOffset));
auto input1 = MakeTensor<T, 4>(inputTensorInfo, QuantizedVector<T>(
{
11.0f, 12.0f,
13.0f, 14.0f,
15.0f, 16.0f,
17.0f, 18.0f,
19.0f, 20.0f,
21.0f, 22.0f
},
qScale, qOffset));
auto input2 = MakeTensor<T, 4>(inputTensorInfo, QuantizedVector<T>(
{
21.0f, 22.0f,
23.0f, 24.0f,
25.0f, 26.0f,
27.0f, 28.0f,
29.0f, 30.0f,
31.0f, 32.0f
},
qScale, qOffset));
LayerTestResult<T, 4> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory,
memoryManager,
{inputTensorInfo, inputTensorInfo, inputTensorInfo},
{input0.data(), input1.data(), input2.data()},
outputTensorInfo,
output.data(),
dimension,
useSubtensor);
result.output = MakeTensor<T, 4>(outputTensorInfo, output);
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 4> Concat4dDim0TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
TensorInfo outputTensorInfo({ 3, 3, 2, 2 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 4> result = Concat4dTestImpl<ArmnnType>(
workloadFactory, memoryManager, outputTensorInfo, 0, true, qScale, qOffset);
result.outputExpected = MakeTensor<T, 4>(outputTensorInfo, QuantizedVector<T>(
{
1.0f, 2.0f,
3.0f, 4.0f,
5.0f, 6.0f,
7.0f, 8.0f,
9.0f, 10.0f,
11.0f, 12.0f,
11.0f, 12.0f,
13.0f, 14.0f,
15.0f, 16.0f,
17.0f, 18.0f,
19.0f, 20.0f,
21.0f, 22.0f,
21.0f, 22.0f,
23.0f, 24.0f,
25.0f, 26.0f,
27.0f, 28.0f,
29.0f, 30.0f,
31.0f, 32.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 4> Concat4dDim1TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
TensorInfo outputTensorInfo({ 1, 9, 2, 2 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 4> result = Concat4dTestImpl<ArmnnType>(
workloadFactory, memoryManager, outputTensorInfo, 1, true, qScale, qOffset);
result.outputExpected = MakeTensor<T, 4>(outputTensorInfo, QuantizedVector<T>(
{
1.0f, 2.0f,
3.0f, 4.0f,
5.0f, 6.0f,
7.0f, 8.0f,
9.0f, 10.0f,
11.0f, 12.0f,
11.0f, 12.0f,
13.0f, 14.0f,
15.0f, 16.0f,
17.0f, 18.0f,
19.0f, 20.0f,
21.0f, 22.0f,
21.0f, 22.0f,
23.0f, 24.0f,
25.0f, 26.0f,
27.0f, 28.0f,
29.0f, 30.0f,
31.0f, 32.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 4> Concat4dDim2TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
TensorInfo outputTensorInfo({ 1, 3, 6, 2 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 4> result = Concat4dTestImpl<ArmnnType>(
workloadFactory, memoryManager, outputTensorInfo, 2, true, qScale, qOffset);
result.outputExpected = MakeTensor<T, 4>(outputTensorInfo, QuantizedVector<T>(
{
1.0f, 2.0f,
3.0f, 4.0f,
11.0f, 12.0f,
13.0f, 14.0f,
21.0f, 22.0f,
23.0f, 24.0f,
5.0f, 6.0f,
7.0f, 8.0f,
15.0f, 16.0f,
17.0f, 18.0f,
25.0f, 26.0f,
27.0f, 28.0f,
9.0f, 10.0f,
11.0f, 12.0f,
19.0f, 20.0f,
21.0f, 22.0f,
29.0f, 30.0f,
31.0f, 32.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 4> Concat4dDim3TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset,
bool useSubtensor)
{
TensorInfo outputTensorInfo({ 1, 3, 2, 6 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 4> result = Concat4dTestImpl<ArmnnType>(
workloadFactory, memoryManager, outputTensorInfo, 3, useSubtensor, qScale, qOffset);
result.outputExpected = MakeTensor<T, 4>(outputTensorInfo, QuantizedVector<T>(
{
1.0f, 2.0f,
11.0f, 12.0f,
21.0f, 22.0f,
3.0f, 4.0f,
13.0f, 14.0f,
23.0f, 24.0f,
5.0f, 6.0f,
15.0f, 16.0f,
25.0f, 26.0f,
7.0f, 8.0f,
17.0f, 18.0f,
27.0f, 28.0f,
9.0f, 10.0f,
19.0f, 20.0f,
29.0f, 30.0f,
11.0f, 12.0f,
21.0f, 22.0f,
31.0f, 32.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 4> Concat4dDiffShapeDim0TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
constexpr unsigned int dimension = 0u;
TensorInfo inputTensorInfo0({ 1, 3, 2, 2 }, ArmnnType, qScale, qOffset);
auto input0 = MakeTensor<T, 4>(inputTensorInfo0, QuantizedVector<T>(
{
1.0f, 2.0f,
3.0f, 4.0f,
5.0f, 6.0f,
7.0f, 8.0f,
9.0f, 10.0f,
11.0f, 12.0f
},
qScale, qOffset));
TensorInfo inputTensorInfo1({ 2, 3, 2, 2 }, ArmnnType, qScale, qOffset);
auto input1 = MakeTensor<T, 4>(inputTensorInfo1, QuantizedVector<T>(
{
11.0f, 12.0f,
13.0f, 14.0f,
15.0f, 16.0f,
17.0f, 18.0f,
19.0f, 20.0f,
21.0f, 22.0f,
21.0f, 22.0f,
23.0f, 24.0f,
25.0f, 26.0f,
27.0f, 28.0f,
29.0f, 30.0f,
31.0f, 32.0f
},
qScale, qOffset));
TensorInfo outputTensorInfo({ 3, 3, 2, 2 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 4> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory,
memoryManager,
{inputTensorInfo0, inputTensorInfo1},
{input0.data(), input1.data()},
outputTensorInfo,
output.data(),
dimension,
true);
result.output = MakeTensor<T, 4>(outputTensorInfo, output);
result.outputExpected = MakeTensor<T, 4>(outputTensorInfo, QuantizedVector<T>(
{
1.0f, 2.0f,
3.0f, 4.0f,
5.0f, 6.0f,
7.0f, 8.0f,
9.0f, 10.0f,
11.0f, 12.0f,
11.0f, 12.0f,
13.0f, 14.0f,
15.0f, 16.0f,
17.0f, 18.0f,
19.0f, 20.0f,
21.0f, 22.0f,
21.0f, 22.0f,
23.0f, 24.0f,
25.0f, 26.0f,
27.0f, 28.0f,
29.0f, 30.0f,
31.0f, 32.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 4> Concat4dDiffShapeDim1TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
constexpr unsigned int dimension = 1u;
TensorInfo inputTensorInfo0({ 1, 3, 2, 2 }, ArmnnType, qScale, qOffset);
auto input0 = MakeTensor<T, 4>(inputTensorInfo0, QuantizedVector<T>(
{
1.0f, 2.0f,
3.0f, 4.0f,
5.0f, 6.0f,
7.0f, 8.0f,
9.0f, 10.0f,
11.0f, 12.0f
},
qScale, qOffset));
TensorInfo inputTensorInfo1({ 1, 2, 2, 2 }, ArmnnType, qScale, qOffset);
auto input1 = MakeTensor<T, 4>(inputTensorInfo1, QuantizedVector<T>(
{
11.0f, 12.0f,
13.0f, 14.0f,
15.0f, 16.0f,
17.0f, 18.0f,
},
qScale, qOffset));
TensorInfo outputTensorInfo({ 1, 5, 2, 2 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 4> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory,
memoryManager,
{inputTensorInfo0, inputTensorInfo1},
{input0.data(), input1.data()},
outputTensorInfo,
output.data(),
dimension,
true);
result.output = MakeTensor<T, 4>(outputTensorInfo, output);
result.outputExpected = MakeTensor<T, 4>(outputTensorInfo, QuantizedVector<T>(
{
1.0f, 2.0f,
3.0f, 4.0f,
5.0f, 6.0f,
7.0f, 8.0f,
9.0f, 10.0f,
11.0f, 12.0f,
11.0f, 12.0f,
13.0f, 14.0f,
15.0f, 16.0f,
17.0f, 18.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 4> Concat4dDiffShapeDim2TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset)
{
constexpr unsigned int dimension = 2u;
TensorInfo inputTensorInfo0({ 1, 3, 2, 2 }, ArmnnType, qScale, qOffset);
auto input0 = MakeTensor<T, 4>(inputTensorInfo0, QuantizedVector<T>(
{
1.0f, 2.0f,
3.0f, 4.0f,
5.0f, 6.0f,
7.0f, 8.0f,
9.0f, 10.0f,
11.0f, 12.0f
},
qScale, qOffset));
TensorInfo inputTensorInfo1({ 1, 3, 3, 2 }, ArmnnType, qScale, qOffset);
auto input1 = MakeTensor<T, 4>(inputTensorInfo1, QuantizedVector<T>(
{
11.0f, 12.0f,
13.0f, 14.0f,
15.0f, 16.0f,
17.0f, 18.0f,
19.0f, 20.0f,
21.0f, 22.0f,
23.0f, 24.0f,
25.0f, 26.0f,
27.0f, 28.0f
},
qScale, qOffset));
TensorInfo outputTensorInfo({ 1, 3, 5, 2 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 4> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory,
memoryManager,
{inputTensorInfo0, inputTensorInfo1},
{input0.data(), input1.data()},
outputTensorInfo,
output.data(),
dimension,
true);
result.output = MakeTensor<T, 4>(outputTensorInfo, output);
result.outputExpected = MakeTensor<T, 4>(outputTensorInfo, QuantizedVector<T>(
{
1.0f, 2.0f,
3.0f, 4.0f,
11.0f, 12.0f,
13.0f, 14.0f,
15.0f, 16.0f,
5.0f, 6.0f,
7.0f, 8.0f,
17.0f, 18.0f,
19.0f, 20.0f,
21.0f, 22.0f,
9.0f, 10.0f,
11.0f, 12.0f,
23.0f, 24.0f,
25.0f, 26.0f,
27.0f, 28.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T = ResolveType<ArmnnType>>
LayerTestResult<T, 4> Concat4dDiffShapeDim3TestImpl(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
float qScale,
int32_t qOffset,
bool useSubtensor)
{
constexpr unsigned int dimension = 3u;
TensorInfo inputTensorInfo0({ 1, 3, 2, 2 }, ArmnnType, qScale, qOffset);
auto input0 = MakeTensor<T, 4>(inputTensorInfo0, QuantizedVector<T>(
{
1.0f, 2.0f,
3.0f, 4.0f,
5.0f, 6.0f,
7.0f, 8.0f,
9.0f, 10.0f,
11.0f, 12.0f
},
qScale, qOffset));
TensorInfo inputTensorInfo1({ 1, 3, 2, 3 }, ArmnnType, qScale, qOffset);
auto input1 = MakeTensor<T, 4>(inputTensorInfo1, QuantizedVector<T>(
{
11.0f, 12.0f, 13.0f,
14.0f, 15.0f, 16.0f,
17.0f, 18.0f, 19.0f,
20.0f, 21.0f, 22.0f,
23.0f, 24.0f, 25.0f,
26.0f, 27.0f, 28.0f
},
qScale, qOffset));
TensorInfo outputTensorInfo({ 1, 3, 2, 5 }, ArmnnType, qScale, qOffset);
LayerTestResult<T, 4> result(outputTensorInfo);
std::vector<T> output;
output.resize(outputTensorInfo.GetNumElements());
Concatenate<T>(workloadFactory,
memoryManager,
{inputTensorInfo0, inputTensorInfo1},
{input0.data(), input1.data()},
outputTensorInfo,
output.data(),
dimension,
useSubtensor);
result.output = MakeTensor<T, 4>(outputTensorInfo, output);
result.outputExpected = MakeTensor<T, 4>(outputTensorInfo, QuantizedVector<T>(
{
1.0f, 2.0f, 11.0f, 12.0f, 13.0f,
3.0f, 4.0f, 14.0f, 15.0f, 16.0f,
5.0f, 6.0f, 17.0f, 18.0f, 19.0f,
7.0f, 8.0f, 20.0f, 21.0f, 22.0f,
9.0f, 10.0f, 23.0f, 24.0f, 25.0f,
11.0f, 12.0f, 26.0f, 27.0f, 28.0f
},
qScale, qOffset));
return result;
}
template<DataType ArmnnType, typename T>
LayerTestResult<T, 3> ConcatDifferentInputOutputQParamTest(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
bool useSubtensor)
{
IgnoreUnused(memoryManager);
// Defines the tensor descriptors.
TensorInfo outputTensorInfo({ 3, 6, 3 }, ArmnnType);
TensorInfo inputTensorInfo1({ 3, 6, 2 }, ArmnnType);
TensorInfo inputTensorInfo2({ 3, 6, 1 }, ArmnnType);
std::vector<TensorShape> inputTensorShapes({inputTensorInfo1.GetShape(), inputTensorInfo2.GetShape()});
// Quantized input1 tensor.
const float inputScale1 = 0.5f;
const int32_t inputOffset1 = 5;
auto input1 = MakeTensor<T, 3>(inputTensorInfo1, std::vector<T>(
{
1, 2, 3,
4, 5, 6,
7, 8, 9,
10, 11, 12,
13, 14, 15,
16, 17, 18,
19, 20, 21,
22, 23, 24,
25, 26, 27,
28, 29, 30,
31, 32, 33,
34, 35, 36
}));
// Quatized input2 tensor.
const float inputScale2 = 0.2f;
const int32_t inputOffset2 = 10;
auto input2 = MakeTensor<T, 3>(inputTensorInfo2, std::vector<T>(
{
37, 38, 39,
40, 41, 42,
43, 44, 45,
46, 47, 48,
49, 50, 51,
52, 53, 54
}));
// Quantized output tensor.
const float outputScale = 0.1f;
const int32_t outputOffset = 20;
LayerTestResult<T, 3> ret(outputTensorInfo);
ret.outputExpected = MakeTensor<T, 3>(outputTensorInfo, std::vector<T>(
{
0, 5, 74,
10, 15, 76,
20, 25, 78,
30, 35, 80,
40, 45, 82,
50, 55, 84,
60, 65, 86,
70, 75, 88,
80, 85, 90,
90, 95, 92,
100, 105, 94,
110, 115, 96,
120, 125, 98,
130, 135, 100,
140, 145, 102,
150, 155, 104,
160, 165, 106,
170, 175, 108
}));
outputTensorInfo.SetQuantizationScale(outputScale);
outputTensorInfo.SetQuantizationOffset(outputOffset);
inputTensorInfo1.SetQuantizationScale(inputScale1);
inputTensorInfo1.SetQuantizationOffset(inputOffset1);
inputTensorInfo2.SetQuantizationScale(inputScale2);
inputTensorInfo2.SetQuantizationOffset(inputOffset2);
std::vector<unsigned int> wOrigin1 = { 0, 0, 0 }; //Extent of the window is defined by size of input[0].
ConcatQueueDescriptor::ViewOrigin window1(wOrigin1);
std::vector<unsigned int> wOrigin2 = { 0, 0, 2 }; //Extent of the window is defined by size of input[1].
ConcatQueueDescriptor::ViewOrigin window2(wOrigin2);
std::unique_ptr<ITensorHandle> outputHandle = workloadFactory.CreateTensorHandle(outputTensorInfo);
bool subTensorsSupported = useSubtensor && workloadFactory.SupportsSubTensors();
std::unique_ptr<ITensorHandle> inputHandle1 =
subTensorsSupported ?
workloadFactory.CreateSubTensorHandle(*outputHandle, inputTensorInfo1.GetShape(), wOrigin1.data()) :
workloadFactory.CreateTensorHandle(inputTensorInfo1);
std::unique_ptr<ITensorHandle> inputHandle2 =
subTensorsSupported ?
workloadFactory.CreateSubTensorHandle(*outputHandle, inputTensorInfo2.GetShape(), wOrigin2.data()) :
workloadFactory.CreateTensorHandle(inputTensorInfo2);
ConcatQueueDescriptor data;
OriginsDescriptor desc = CreateDescriptorForConcatenation(
inputTensorShapes.begin(),inputTensorShapes.end(), 2);
data.m_Parameters = desc;
WorkloadInfo info;
AddInputToWorkload(data, info, inputTensorInfo1, inputHandle1.get());
AddInputToWorkload(data, info, inputTensorInfo2, inputHandle2.get());
AddOutputToWorkload(data, info, outputTensorInfo, outputHandle.get());
data.m_ViewOrigins.push_back(window1);
data.m_ViewOrigins.push_back(window2);
std::unique_ptr<IWorkload> workload = workloadFactory.CreateConcat(data, info);
inputHandle1->Allocate();
inputHandle2->Allocate();
outputHandle->Allocate();
CopyDataToITensorHandle(inputHandle1.get(), &input1[0][0][0]);
CopyDataToITensorHandle(inputHandle2.get(), &input2[0][0][0]);
workload->PostAllocationConfigure();
workload->Execute();
CopyDataFromITensorHandle(&ret.output[0][0][0], outputHandle.get());
return ret;
}
//
// Explicit template specializations
//
template LayerTestResult<ResolveType<DataType::QAsymmU8>, 3>
ConcatDifferentInputOutputQParamTest<DataType::QAsymmU8>(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
bool useSubtensor);
template LayerTestResult<ResolveType<DataType::QSymmS16>, 3>
ConcatDifferentInputOutputQParamTest<DataType::QSymmS16>(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
bool useSubtensor);
//
// Implementation functions
//
LayerTestResult<float,3> ConcatTest(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
IgnoreUnused(memoryManager);
unsigned int outputWidth = 3;
unsigned int outputHeight = 6;
unsigned int outputChannels = 3;
unsigned int inputWidth1 = 3;
unsigned int inputHeight1 = 6;
unsigned int inputChannels1 = 2;
unsigned int inputWidth2 = 3;
unsigned int inputHeight2 = 6;
unsigned int inputChannels2 = 1;
// Define the tensor descriptors.
TensorInfo outputTensorInfo({ outputChannels, outputHeight, outputWidth }, DataType::Float32);
TensorInfo inputTensorInfo1({ inputChannels1, inputHeight1, inputWidth1 }, DataType::Float32);
TensorInfo inputTensorInfo2({ inputChannels2, inputHeight2, inputWidth2 }, DataType::Float32);
LayerTestResult<float,3> ret(outputTensorInfo);
ret.outputExpected = MakeTensor<float, 3>(outputTensorInfo, std::vector<float>(
{
1.0f, 2.0f, 3.0f,
4.0f, 5.0f, 6.0f,
7.0f, 8.0f, 9.0f,
10.0f, 11.0f, 12.0f,
13.0f, 14.0f, 15.0f,
16.0f, 17.0f, 18.0f,
19.0f, 20.0f, 21.0f,
22.0f, 23.0f, 24.0f,
25.0f, 26.0f, 27.0f,
28.0f, 29.0f, 30.0f,
31.0f, 32.0f, 33.0f,
34.0f, 35.0f, 36.0f,
37.0f, 38.0f, 39.0f,
40.0f, 41.0f, 42.0f,
43.0f, 44.0f, 45.0f,
46.0f, 47.0f, 48.0f,
49.0f, 50.0f, 51.0f,
52.0f, 53.0f, 54.0f,
})
);
auto input1 = MakeTensor<float, 3>(inputTensorInfo1, std::vector<float>(
{
1.0f, 2.0f, 3.0f,
4.0f, 5.0f, 6.0f,
7.0f, 8.0f, 9.0f,
10.0f, 11.0f, 12.0f,
13.0f, 14.0f, 15.0f,
16.0f, 17.0f, 18.0f,
19.0f, 20.0f, 21.0f,
22.0f, 23.0f, 24.0f,
25.0f, 26.0f, 27.0f,
28.0f, 29.0f, 30.0f,
31.0f, 32.0f, 33.0f,
34.0f, 35.0f, 36.0f,
})
);
auto input2 = MakeTensor<float, 3>(inputTensorInfo2, std::vector<float>(
{
37.0f, 38.0f, 39.0f,
40.0f, 41.0f, 42.0f,
43.0f, 44.0f, 45.0f,
46.0f, 47.0f, 48.0f,
49.0f, 50.0f, 51.0f,
52.0f, 53.0f, 54.0f,
})
);
std::vector<unsigned int> wOrigin1 = {0, 0, 0}; //Extent of the window is defined by size of input[0].
ConcatQueueDescriptor::ViewOrigin window1(wOrigin1);
std::vector<unsigned int> wOrigin2 = {2, 0, 0}; //Extent of the window is defined by size of input[1].
ConcatQueueDescriptor::ViewOrigin window2(wOrigin2);
std::unique_ptr<ITensorHandle> outputHandle = workloadFactory.CreateTensorHandle(outputTensorInfo);
bool subTensorsSupported = workloadFactory.SupportsSubTensors();
std::unique_ptr<ITensorHandle> inputHandle1 =
subTensorsSupported ?
workloadFactory.CreateSubTensorHandle(*outputHandle, inputTensorInfo1.GetShape(), wOrigin1.data()) :
workloadFactory.CreateTensorHandle(inputTensorInfo1);
std::unique_ptr<ITensorHandle> inputHandle2 =
subTensorsSupported ?
workloadFactory.CreateSubTensorHandle(*outputHandle, inputTensorInfo2.GetShape(), wOrigin2.data()) :
workloadFactory.CreateTensorHandle(inputTensorInfo2);
ConcatQueueDescriptor data;
WorkloadInfo info;
AddInputToWorkload(data, info, inputTensorInfo1, inputHandle1.get());
AddInputToWorkload(data, info, inputTensorInfo2, inputHandle2.get());
AddOutputToWorkload(data, info, outputTensorInfo, outputHandle.get());
data.m_ViewOrigins.push_back(window1);
data.m_ViewOrigins.push_back(window2);
std::unique_ptr<IWorkload> workload = workloadFactory.CreateConcat(data, info);
inputHandle1->Allocate();
inputHandle2->Allocate();
outputHandle->Allocate();
CopyDataToITensorHandle(inputHandle1.get(), &input1[0][0][0]);
CopyDataToITensorHandle(inputHandle2.get(), &input2[0][0][0]);
workload->PostAllocationConfigure();
workload->Execute();
CopyDataFromITensorHandle(&ret.output[0][0][0], outputHandle.get());
return ret;
}
LayerTestResult<float, 1> Concat1dTest(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat1dTestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 2> Concat2dDim0Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat2dDim0TestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 2> Concat2dDim1Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat2dDim1TestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 2> Concat2dDim0DiffInputDimsTest(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat2dDim0DiffInputDimsTestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 2> Concat2dDim1DiffInputDimsTest(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat2dDim1DiffInputDimsTestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 3> Concat3dDim0Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat3dDim0TestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 3> Concat3dDim1Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat3dDim1TestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 3> Concat3dDim2Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
bool useSubtensor)
{
return Concat3dDim2TestImpl<DataType::Float32>(workloadFactory, memoryManager, useSubtensor, 0.0f, 0);
}
LayerTestResult<float, 3> Concat3dDim0DiffInputDimsTest(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat3dDim0DiffInputDimsTestImpl<DataType::Float32>(
workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 3> Concat3dDim1DiffInputDimsTest(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat3dDim1DiffInputDimsTestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 3> Concat3dDim2DiffInputDimsTest(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
bool useSubtensor)
{
return Concat3dDim2DiffInputDimsTestImpl<DataType::Float32>(
workloadFactory, memoryManager, useSubtensor, 0.0f, 0);
}
LayerTestResult<float, 4> Concat4dDim0Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat4dDim0TestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 4> Concat4dDim1Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat4dDim1TestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 4> Concat4dDim2Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat4dDim2TestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 4> Concat4dDim3Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
bool useSubtensor)
{
return Concat4dDim3TestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0, useSubtensor);
}
LayerTestResult<float, 4> Concat4dDiffShapeDim0Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat4dDiffShapeDim0TestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 4> Concat4dDiffShapeDim1Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat4dDiffShapeDim1TestImpl<DataType::Float32>(
workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 4> Concat4dDiffShapeDim2Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat4dDiffShapeDim2TestImpl<DataType::Float32>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<float, 4> Concat4dDiffShapeDim3Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
bool useSubtensor)
{
return Concat4dDiffShapeDim3TestImpl<DataType::Float32>(
workloadFactory, memoryManager, 0.0f, 0, useSubtensor);
}
LayerTestResult<Half, 3> ConcatFloat16Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat3dDim1TestImpl<DataType::Float16>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<BFloat16, 3> ConcatBFloat16Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat3dDim1TestImpl<DataType::BFloat16>(workloadFactory, memoryManager, 0.0f, 0);
}
LayerTestResult<uint8_t, 3> ConcatUint8DifferentQParamsTest(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
IgnoreUnused(memoryManager);
unsigned int outputWidth = 3;
unsigned int outputHeight = 6;
unsigned int outputChannels = 3;
unsigned int inputWidth1 = 3;
unsigned int inputHeight1 = 6;
unsigned int inputChannels1 = 2;
unsigned int inputWidth2 = 3;
unsigned int inputHeight2 = 6;
unsigned int inputChannels2 = 1;
// Defines the tensor descriptors.
TensorInfo outputTensorInfo({ outputChannels, outputHeight, outputWidth }, DataType::QAsymmU8);
TensorInfo inputTensorInfo1({ inputChannels1, inputHeight1, inputWidth1 }, DataType::QAsymmU8);
TensorInfo inputTensorInfo2({ inputChannels2, inputHeight2, inputWidth2 }, DataType::QAsymmU8);
// Quantized input1 tensor. Range [-3, 1]
const float inputScale1 = 0.015686f;
const int32_t inputOffset1 = 192;
auto input1 = MakeTensor<uint8_t, 3>(inputTensorInfo1, std::vector<uint8_t>(
{
1, 2, 3,
4, 5, 6,
7, 8, 9,
10, 11, 12,
13, 14, 15,
16, 17, 18,
19, 20, 21,
22, 23, 24,
25, 26, 27,
28, 29, 30,
31, 32, 33,
34, 35, 36,
})
);
// Quatized input2 tensor. Range [-1, 4]
const float inputScale2 = 0.019608f;
const int32_t inputOffset2 = 50;
auto input2 = MakeTensor<uint8_t, 3>(inputTensorInfo2, std::vector<uint8_t>(
{
37, 38, 39,
40, 41, 42,
43, 44, 45,
46, 47, 48,
49, 50, 51,
52, 53, 54,
})
);
// Output has the same quantization parameters than input1,
// so that only the requantization of input2 is required
const float outputScale = 0.015686f;
const int32_t outputOffset = 192;
LayerTestResult<uint8_t, 3> ret(outputTensorInfo);
ret.outputExpected = MakeTensor<uint8_t, 3>(outputTensorInfo, std::vector<uint8_t>(
{
1, 2, 3,
4, 5, 6,
7, 8, 9,
10, 11, 12,
13, 14, 15,
16, 17, 18,
19, 20, 21,
22, 23, 24,
25, 26, 27,
28, 29, 30,
31, 32, 33,
34, 35, 36,
176, 177, 178,
179, 181, 182,
183, 184, 186,
187, 188, 189,
191, 192, 193,
195, 196, 197,
})
);
outputTensorInfo.SetQuantizationScale(outputScale);
outputTensorInfo.SetQuantizationOffset(outputOffset);
inputTensorInfo1.SetQuantizationScale(inputScale1);
inputTensorInfo1.SetQuantizationOffset(inputOffset1);
inputTensorInfo2.SetQuantizationScale(inputScale2);
inputTensorInfo2.SetQuantizationOffset(inputOffset2);
std::vector<unsigned int> wOrigin1 = { 0, 0, 0 }; //Extent of the window is defined by size of input[0].
ConcatQueueDescriptor::ViewOrigin window1(wOrigin1);
std::vector<unsigned int> wOrigin2 = { 2, 0, 0 }; //Extent of the window is defined by size of input[1].
ConcatQueueDescriptor::ViewOrigin window2(wOrigin2);
std::unique_ptr<ITensorHandle> outputHandle = workloadFactory.CreateTensorHandle(outputTensorInfo);
bool subTensorsSupported = workloadFactory.SupportsSubTensors();
std::unique_ptr<ITensorHandle> inputHandle1 =
subTensorsSupported ?
workloadFactory.CreateSubTensorHandle(*outputHandle, inputTensorInfo1.GetShape(), wOrigin1.data()) :
workloadFactory.CreateTensorHandle(inputTensorInfo1);
std::unique_ptr<ITensorHandle> inputHandle2 =
subTensorsSupported ?
workloadFactory.CreateSubTensorHandle(*outputHandle, inputTensorInfo2.GetShape(), wOrigin2.data()) :
workloadFactory.CreateTensorHandle(inputTensorInfo2);
ConcatQueueDescriptor data;
WorkloadInfo info;
AddInputToWorkload(data, info, inputTensorInfo1, inputHandle1.get());
AddInputToWorkload(data, info, inputTensorInfo2, inputHandle2.get());
AddOutputToWorkload(data, info, outputTensorInfo, outputHandle.get());
data.m_ViewOrigins.push_back(window1);
data.m_ViewOrigins.push_back(window2);
std::unique_ptr<IWorkload> workload = workloadFactory.CreateConcat(data, info);
inputHandle1->Allocate();
inputHandle2->Allocate();
outputHandle->Allocate();
CopyDataToITensorHandle(inputHandle1.get(), &input1[0][0][0]);
CopyDataToITensorHandle(inputHandle2.get(), &input2[0][0][0]);
workload->PostAllocationConfigure();
workload->Execute();
CopyDataFromITensorHandle(&ret.output[0][0][0], outputHandle.get());
return ret;
}
LayerTestResult<uint8_t, 3> ConcatUint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
IgnoreUnused(memoryManager);
unsigned int outputWidth = 3;
unsigned int outputHeight = 6;
unsigned int outputChannels = 3;
unsigned int inputWidth1 = 3;
unsigned int inputHeight1 = 6;
unsigned int inputChannels1 = 2;
unsigned int inputWidth2 = 3;
unsigned int inputHeight2 = 6;
unsigned int inputChannels2 = 1;
// Defines the tensor descriptors.
TensorInfo outputTensorInfo({ outputChannels, outputHeight, outputWidth }, DataType::QAsymmU8);
TensorInfo inputTensorInfo1({ inputChannels1, inputHeight1, inputWidth1 }, DataType::QAsymmU8);
TensorInfo inputTensorInfo2({ inputChannels2, inputHeight2, inputWidth2 }, DataType::QAsymmU8);
// Arbitrary scale and offsets. They don't really matter as the Concat operator doesn't dequantize/quantize them.
const float scale = 0.13497836f;
const int32_t offset = -7;
outputTensorInfo.SetQuantizationScale(scale);
outputTensorInfo.SetQuantizationOffset(offset);
inputTensorInfo1.SetQuantizationScale(scale);
inputTensorInfo1.SetQuantizationOffset(offset);
inputTensorInfo2.SetQuantizationScale(scale);
inputTensorInfo2.SetQuantizationOffset(offset);
LayerTestResult<uint8_t, 3> ret(outputTensorInfo);
ret.outputExpected = MakeTensor<uint8_t, 3>(outputTensorInfo, std::vector<uint8_t>(
{
1, 2, 3,
4, 5, 6,
7, 8, 9,
10, 11, 12,
13, 14, 15,
16, 17, 18,
19, 20, 21,
22, 23, 24,
25, 26, 27,
28, 29, 30,
31, 32, 33,
34, 35, 36,
37, 38, 39,
40, 41, 42,
43, 44, 45,
46, 47, 48,
49, 50, 51,
52, 53, 54,
})
);
auto input1 = MakeTensor<uint8_t, 3>(inputTensorInfo1, std::vector<uint8_t>(
{
1, 2, 3,
4, 5, 6,
7, 8, 9,
10, 11, 12,
13, 14, 15,
16, 17, 18,
19, 20, 21,
22, 23, 24,
25, 26, 27,
28, 29, 30,
31, 32, 33,
34, 35, 36,
})
);
auto input2 = MakeTensor<uint8_t, 3>(inputTensorInfo2, std::vector<uint8_t>(
{
37, 38, 39,
40, 41, 42,
43, 44, 45,
46, 47, 48,
49, 50, 51,
52, 53, 54,
})
);
std::vector<unsigned int> wOrigin1 = { 0, 0, 0 }; //Extent of the window is defined by size of input[0].
ConcatQueueDescriptor::ViewOrigin window1(wOrigin1);
std::vector<unsigned int> wOrigin2 = { 2, 0, 0 }; //Extent of the window is defined by size of input[1].
ConcatQueueDescriptor::ViewOrigin window2(wOrigin2);
std::unique_ptr<ITensorHandle> outputHandle = workloadFactory.CreateTensorHandle(outputTensorInfo);
bool subTensorsSupported = workloadFactory.SupportsSubTensors();
std::unique_ptr<ITensorHandle> inputHandle1 =
subTensorsSupported ?
workloadFactory.CreateSubTensorHandle(*outputHandle, inputTensorInfo1.GetShape(), wOrigin1.data()) :
workloadFactory.CreateTensorHandle(inputTensorInfo1);
std::unique_ptr<ITensorHandle> inputHandle2 =
subTensorsSupported ?
workloadFactory.CreateSubTensorHandle(*outputHandle, inputTensorInfo2.GetShape(), wOrigin2.data()) :
workloadFactory.CreateTensorHandle(inputTensorInfo2);
ConcatQueueDescriptor data;
WorkloadInfo info;
AddInputToWorkload(data, info, inputTensorInfo1, inputHandle1.get());
AddInputToWorkload(data, info, inputTensorInfo2, inputHandle2.get());
AddOutputToWorkload(data, info, outputTensorInfo, outputHandle.get());
data.m_ViewOrigins.push_back(window1);
data.m_ViewOrigins.push_back(window2);
std::unique_ptr<IWorkload> workload = workloadFactory.CreateConcat(data, info);
inputHandle1->Allocate();
inputHandle2->Allocate();
outputHandle->Allocate();
CopyDataToITensorHandle(inputHandle1.get(), &input1[0][0][0]);
CopyDataToITensorHandle(inputHandle2.get(), &input2[0][0][0]);
workload->PostAllocationConfigure();
workload->Execute();
CopyDataFromITensorHandle(&ret.output[0][0][0], outputHandle.get());
return ret;
}
LayerTestResult<uint16_t, 3> ConcatUint16Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
IgnoreUnused(memoryManager);
unsigned int outputWidth = 3;
unsigned int outputHeight = 6;
unsigned int outputChannels = 3;
unsigned int inputWidth1 = 3;
unsigned int inputHeight1 = 6;
unsigned int inputChannels1 = 2;
unsigned int inputWidth2 = 3;
unsigned int inputHeight2 = 6;
unsigned int inputChannels2 = 1;
// Defines the tensor descriptors.
TensorInfo outputTensorInfo({ outputChannels, outputHeight, outputWidth }, DataType::QSymmS16);
TensorInfo inputTensorInfo1({ inputChannels1, inputHeight1, inputWidth1 }, DataType::QSymmS16);
TensorInfo inputTensorInfo2({ inputChannels2, inputHeight2, inputWidth2 }, DataType::QSymmS16);
// Arbitrary scale and offsets. They don't really matter as the Concat operator doesn't dequantize/quantize them.
const float scale = 0.13497836f;
const int32_t offset = -7;
outputTensorInfo.SetQuantizationScale(scale);
outputTensorInfo.SetQuantizationOffset(offset);
inputTensorInfo1.SetQuantizationScale(scale);
inputTensorInfo1.SetQuantizationOffset(offset);
inputTensorInfo2.SetQuantizationScale(scale);
inputTensorInfo2.SetQuantizationOffset(offset);
LayerTestResult<uint16_t, 3> ret(outputTensorInfo);
ret.outputExpected = MakeTensor<uint16_t, 3>(outputTensorInfo, std::vector<uint16_t>(
{
1, 2, 3,
4, 5, 6,
7, 8, 9,
10, 11, 12,
13, 14, 15,
16, 17, 18,
19, 20, 21,
22, 23, 24,
25, 26, 27,
28, 29, 30,
31, 32, 33,
34, 35, 36,
37, 38, 39,
40, 41, 42,
43, 44, 45,
46, 47, 48,
49, 50, 51,
52, 53, 54,
}));
auto input1 = MakeTensor<uint16_t, 3>(inputTensorInfo1, std::vector<uint16_t>(
{
1, 2, 3,
4, 5, 6,
7, 8, 9,
10, 11, 12,
13, 14, 15,
16, 17, 18,
19, 20, 21,
22, 23, 24,
25, 26, 27,
28, 29, 30,
31, 32, 33,
34, 35, 36,
}));
auto input2 = MakeTensor<uint16_t, 3>(inputTensorInfo2, std::vector<uint16_t>(
{
37, 38, 39,
40, 41, 42,
43, 44, 45,
46, 47, 48,
49, 50, 51,
52, 53, 54,
}));
std::vector<unsigned int> wOrigin1 = { 0, 0, 0 }; //Extent of the window is defined by size of input[0].
ConcatQueueDescriptor::ViewOrigin window1(wOrigin1);
std::vector<unsigned int> wOrigin2 = { 2, 0, 0 }; //Extent of the window is defined by size of input[1].
ConcatQueueDescriptor::ViewOrigin window2(wOrigin2);
std::unique_ptr<ITensorHandle> outputHandle = workloadFactory.CreateTensorHandle(outputTensorInfo);
bool subTensorsSupported = workloadFactory.SupportsSubTensors();
std::unique_ptr<ITensorHandle> inputHandle1 =
subTensorsSupported ?
workloadFactory.CreateSubTensorHandle(*outputHandle, inputTensorInfo1.GetShape(), wOrigin1.data()) :
workloadFactory.CreateTensorHandle(inputTensorInfo1);
std::unique_ptr<ITensorHandle> inputHandle2 =
subTensorsSupported ?
workloadFactory.CreateSubTensorHandle(*outputHandle, inputTensorInfo2.GetShape(), wOrigin2.data()) :
workloadFactory.CreateTensorHandle(inputTensorInfo2);
ConcatQueueDescriptor data;
WorkloadInfo info;
AddInputToWorkload(data, info, inputTensorInfo1, inputHandle1.get());
AddInputToWorkload(data, info, inputTensorInfo2, inputHandle2.get());
AddOutputToWorkload(data, info, outputTensorInfo, outputHandle.get());
data.m_ViewOrigins.push_back(window1);
data.m_ViewOrigins.push_back(window2);
std::unique_ptr<IWorkload> workload = workloadFactory.CreateConcat(data, info);
inputHandle1->Allocate();
inputHandle2->Allocate();
outputHandle->Allocate();
CopyDataToITensorHandle(inputHandle1.get(), &input1[0][0][0]);
CopyDataToITensorHandle(inputHandle2.get(), &input2[0][0][0]);
workload->PostAllocationConfigure();
workload->Execute();
CopyDataFromITensorHandle(&ret.output[0][0][0], outputHandle.get());
return ret;
}
LayerTestResult<uint8_t, 1> Concat1dUint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat1dTestImpl<DataType::QAsymmU8>(workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 2> Concat2dDim0Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat2dDim0TestImpl<DataType::QAsymmU8>(workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 2> Concat2dDim1Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat2dDim1TestImpl<DataType::QAsymmU8>(workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 2> Concat2dDim0DiffInputDimsUint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat2dDim0DiffInputDimsTestImpl<DataType::QAsymmU8>(
workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 2> Concat2dDim1DiffInputDimsUint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat2dDim1DiffInputDimsTestImpl<DataType::QAsymmU8>(
workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 3> Concat3dDim0Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat3dDim0TestImpl<DataType::QAsymmU8>(workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 3> Concat3dDim1Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat3dDim1TestImpl<DataType::QAsymmU8>(workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 3> Concat3dDim2Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
bool useSubtensor)
{
return Concat3dDim2TestImpl<DataType::QAsymmU8>(
workloadFactory, memoryManager, useSubtensor, 0.5f, -1);
}
LayerTestResult<uint8_t, 3> Concat3dDim0DiffInputDimsUint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat3dDim0TestImpl<DataType::QAsymmU8>(workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 3> Concat3dDim1DiffInputDimsUint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat3dDim1DiffInputDimsTestImpl<DataType::QAsymmU8>(
workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 3> Concat3dDim2DiffInputDimsUint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
bool useSubtensor)
{
return Concat3dDim2DiffInputDimsTestImpl<DataType::QAsymmU8>(
workloadFactory, memoryManager, useSubtensor, 0.5f, -1);
}
LayerTestResult<uint8_t, 4> Concat4dDim0Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat4dDim0TestImpl<DataType::QAsymmU8>(workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 4> Concat4dDim1Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat4dDim1TestImpl<DataType::QAsymmU8>(workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 4> Concat4dDim2Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat4dDim2TestImpl<DataType::QAsymmU8>(workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 4> Concat4dDim3Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager, bool useSubtensor)
{
return Concat4dDim3TestImpl<DataType::QAsymmU8>(
workloadFactory, memoryManager, 0.5f, -1, useSubtensor);
}
LayerTestResult<uint8_t, 4> Concat4dDiffShapeDim0Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat4dDiffShapeDim0TestImpl<DataType::QAsymmU8>(
workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 4> Concat4dDiffShapeDim1Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat4dDiffShapeDim1TestImpl<DataType::QAsymmU8>(
workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 4> Concat4dDiffShapeDim2Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager)
{
return Concat4dDiffShapeDim2TestImpl<DataType::QAsymmU8>(
workloadFactory, memoryManager, 0.5f, -1);
}
LayerTestResult<uint8_t, 4> Concat4dDiffShapeDim3Uint8Test(
IWorkloadFactory& workloadFactory,
const IBackendInternal::IMemoryManagerSharedPtr& memoryManager,
bool useSubtensor)
{
return Concat4dDiffShapeDim3TestImpl<DataType::QAsymmU8>(
workloadFactory, memoryManager, 0.5f, -1, useSubtensor);
}