// SPDX-License-Identifier: Apache-2.0
//
+#include <transformations/op_conversions/convert_batch_to_space.hpp>
+#include <transformations/op_conversions/convert_space_to_batch.hpp>
+
#include "layer_test_utils.hpp"
+#include "plugin_config.hpp"
namespace LayerTestsUtils {
void LayerTestsCommon::Run() {
SKIP_IF_CURRENT_TEST_IS_DISABLED()
- ConfigurePlugin();
LoadNetwork();
Infer();
Validate();
}
-LayerTestsCommon::~LayerTestsCommon() {
- if (!configuration.empty()) {
- PluginCache::get().reset();
- }
-}
-
InferenceEngine::Blob::Ptr LayerTestsCommon::GenerateInput(const InferenceEngine::InputInfo &info) const {
return FuncTestUtils::createAndFillBlob(info.getTensorDesc());
}
const auto actualBuffer = lockedMemory.as<const std::uint8_t *>();
const auto &precision = actual->getTensorDesc().getPrecision();
- auto bufferSize = actual->size();
- // With dynamic batch, you need to size
- if (configuration.count(InferenceEngine::PluginConfigParams::KEY_DYN_BATCH_ENABLED)) {
- auto batchSize = actual->getTensorDesc().getDims()[0];
- auto halfBatchSize = batchSize > 1 ? batchSize/ 2 : 1;
- bufferSize = (actual->size() * halfBatchSize / batchSize);
- }
- const auto &size = bufferSize;
+ const auto &size = actual->size();
switch (precision) {
case InferenceEngine::Precision::FP32:
Compare(reinterpret_cast<const float *>(expectedBuffer), reinterpret_cast<const float *>(actualBuffer),
}
}
-void LayerTestsCommon::ConfigurePlugin() {
- if (!configuration.empty()) {
- core->SetConfig(configuration, targetDevice);
+void LayerTestsCommon::Compare(const InferenceEngine::Blob::Ptr &expected, const InferenceEngine::Blob::Ptr &actual) {
+ auto get_raw_buffer = [] (const InferenceEngine::Blob::Ptr &blob) {
+ auto memory = InferenceEngine::as<InferenceEngine::MemoryBlob>(blob);
+ IE_ASSERT(memory);
+ const auto lockedMemory = memory->wmap();
+ return lockedMemory.as<const std::uint8_t *>();
+ };
+ const auto expectedBuffer = get_raw_buffer(expected);
+ const auto actualBuffer = get_raw_buffer(actual);
+
+ const auto &precision = actual->getTensorDesc().getPrecision();
+ const auto &size = actual->size();
+ switch (precision) {
+ case InferenceEngine::Precision::FP32:
+ Compare(reinterpret_cast<const float *>(expectedBuffer), reinterpret_cast<const float *>(actualBuffer),
+ size, threshold);
+ break;
+ case InferenceEngine::Precision::I32:
+ Compare(reinterpret_cast<const std::int32_t *>(expectedBuffer),
+ reinterpret_cast<const std::int32_t *>(actualBuffer), size, 0);
+ break;
+ default:
+ FAIL() << "Comparator for " << precision << " precision isn't supported";
}
}
void LayerTestsCommon::LoadNetwork() {
cnnNetwork = InferenceEngine::CNNNetwork{function};
+ PreparePluginConfiguration(this);
ConfigureNetwork();
- executableNetwork = core->LoadNetwork(cnnNetwork, targetDevice);
+ executableNetwork = core->LoadNetwork(cnnNetwork, targetDevice, configuration);
}
void LayerTestsCommon::Infer() {
inferRequest = executableNetwork.CreateInferRequest();
inputs.clear();
- for (const auto &input : cnnNetwork.getInputsInfo()) {
+ for (const auto &input : executableNetwork.GetInputsInfo()) {
const auto &info = input.second;
auto blob = GenerateInput(*info);
inferRequest.SetBlob(info->name(), blob);
}
if (configuration.count(InferenceEngine::PluginConfigParams::KEY_DYN_BATCH_ENABLED) &&
configuration.count(InferenceEngine::PluginConfigParams::YES)) {
- auto batchSize = cnnNetwork.getInputsInfo().begin()->second->getTensorDesc().getDims()[0] / 2;
+ auto batchSize = executableNetwork.GetInputsInfo().begin()->second->getTensorDesc().getDims()[0] / 2;
inferRequest.SetBatch(batchSize);
}
inferRequest.Infer();
std::copy(buffer, buffer + inputSize, referenceInput.data());
}
- const auto &actualOutputs = GetOutputs();
- const auto &convertType = FuncTestUtils::PrecisionUtils::convertIE2nGraphPrc(actualOutputs[0]->getTensorDesc().getPrecision());
+ auto ieOutPrc = outPrc;
+ if (outPrc == InferenceEngine::Precision::UNSPECIFIED) {
+ const auto &actualOutputs = GetOutputs();
+ ieOutPrc = actualOutputs[0]->getTensorDesc().getPrecision();
+ }
+
+ const auto &convertType = FuncTestUtils::PrecisionUtils::convertIE2nGraphPrc(ieOutPrc);
std::vector<std::vector<std::uint8_t>> expectedOutputs;
switch (refMode) {
case INTERPRETER: {
// reference inference on device with other options and nGraph function has to be implemented here
break;
}
+ case INTERPRETER_TRANSFORMATIONS: {
+ auto cloned_function = ngraph::clone_function(*function);
+
+ // todo: add functionality to configure the necessary transformations for each test separately
+ ngraph::pass::Manager m;
+ m.register_pass<ngraph::pass::ConvertSpaceToBatch>();
+ m.register_pass<ngraph::pass::ConvertBatchToSpace>();
+ m.run_passes(cloned_function);
+ expectedOutputs = ngraph::helpers::interpreterFunction(cloned_function, referenceInputs, convertType);
+ break;
+ }
}
return expectedOutputs;
std::vector<InferenceEngine::Blob::Ptr> LayerTestsCommon::GetOutputs() {
auto outputs = std::vector<InferenceEngine::Blob::Ptr>{};
- for (const auto &output : cnnNetwork.getOutputsInfo()) {
+ for (const auto &output : executableNetwork.GetOutputsInfo()) {
const auto &name = output.first;
outputs.push_back(inferRequest.GetBlob(name));
}
void LayerTestsCommon::SetRefMode(RefMode mode) {
refMode = mode;
}
+
+std::shared_ptr<ngraph::Function> LayerTestsCommon::GetFunction() {
+ return function;
+}
+
+std::map<std::string, std::string>& LayerTestsCommon::GetConfiguration() {
+ return configuration;
+}
} // namespace LayerTestsUtils