microsoft
/
Windows-Machine-Learning
зеркало из
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность

Fix per iteration perf output functionality that was deleted in refactoring work

This commit is contained in:
Ryan Lai 2019-04-17 14:55:33 -07:00
Родитель 31a0ec2c03
Коммит 9ccd65377b
3 изменённых файлов: 54 добавлений и 43 удалений
Показать статистику Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы

19
Tools/WinMLRunner/src/CommandLineArgs.cpp
Просмотреть файл

@ -218,13 +218,7 @@ CommandLineArgs::CommandLineArgs(const std::vector<std::wstring>& args)
}
catch (...)
{
// set to default path
auto time = std::time(nullptr);
struct tm localTime;
localtime_s(&localTime, &time);
std::wostringstream oss;
oss << std::put_time(&localTime, L"%Y-%m-%d_%H.%M.%S");
SetTensorOutputPath(L"\\PerIterationRun[" + oss.str() + L"]");
// Will Set Default Path after argument checks
}
}
else if (_wcsicmp(args[i].c_str(), L"-version") == 0)
@ -318,6 +312,17 @@ CommandLineArgs::CommandLineArgs(const std::vector<std::wstring>& args)
throw hresult_invalid_argument(msg.c_str());
}
}
// set default path for per iteration / tensor output
if (this->TensorOutputPath().empty())
{
auto time = std::time(nullptr);
struct tm localTime;
localtime_s(&localTime, &time);
std::wostringstream oss;
oss << std::put_time(&localTime, L"%Y-%m-%d_%H.%M.%S");
SetTensorOutputPath(L".\\PerIterationRun[" + oss.str() + L"]");
}
CheckForInvalidArguments();
}

2
Tools/WinMLRunner/src/OutputHelper.h
Просмотреть файл

@ -596,7 +596,7 @@ public:
void SetCSVFileName(const std::wstring& fileName) { m_csvFileName = fileName; }
void WritePerIterationPerformance(const CommandLineArgs& args, std::wstring model, std::wstring img)
void WritePerIterationPerformance(const CommandLineArgs& args, const std::wstring& model, const std::wstring& img)
{
if (m_csvFileNamePerIterationSummary.length() > 0)
{

76
Tools/WinMLRunner/src/Run.cpp
Просмотреть файл

@ -80,12 +80,12 @@ HRESULT BindInputFeatures(const LearningModel& model, const LearningModelBinding
return S_OK;
}
HRESULT LoadModel(LearningModel &model, const std::wstring& path, bool capturePerf, OutputHelper& output, const CommandLineArgs& args,
uint32_t iterationNum, Profiler<WINML_MODEL_TEST_PERF>& profiler)
HRESULT LoadModel(LearningModel& model, const std::wstring& path, bool capturePerf, OutputHelper& output,
const CommandLineArgs& args, uint32_t iterationNum, Profiler<WINML_MODEL_TEST_PERF>& profiler)
{
try
{
output.PrintLoadingInfo(path);
output.PrintLoadingInfo(path);
if (capturePerf)
{
WINML_PROFILING_START(profiler, WINML_MODEL_TEST_PERF::LOAD_MODEL);
@ -111,9 +111,9 @@ HRESULT LoadModel(LearningModel &model, const std::wstring& path, bool capturePe
return S_OK;
}
HRESULT CreateSession(LearningModelSession& session, IDirect3DDevice& winrtDevice, LearningModel& model, CommandLineArgs& args, OutputHelper& output,
DeviceType deviceType, DeviceCreationLocation deviceCreationLocation,
Profiler<WINML_MODEL_TEST_PERF>& profiler)
HRESULT CreateSession(LearningModelSession& session, IDirect3DDevice& winrtDevice, LearningModel& model,
CommandLineArgs& args, OutputHelper& output, DeviceType deviceType,
DeviceCreationLocation deviceCreationLocation, Profiler<WINML_MODEL_TEST_PERF>& profiler)
{
if (model == nullptr)
{
@ -209,12 +209,14 @@ HRESULT CreateSession(LearningModelSession& session, IDirect3DDevice& winrtDevic
return S_OK;
}
HRESULT BindInputs(LearningModelBinding &context, const LearningModel& model, const LearningModelSession& session, OutputHelper& output,
DeviceType deviceType, const CommandLineArgs& args, InputBindingType inputBindingType, InputDataType inputDataType,
const IDirect3DDevice& winrtDevice, DeviceCreationLocation deviceCreationLocation, uint32_t iteration,
HRESULT BindInputs(LearningModelBinding& context, const LearningModel& model, const LearningModelSession& session,
OutputHelper& output, DeviceType deviceType, const CommandLineArgs& args,
InputBindingType inputBindingType, InputDataType inputDataType, const IDirect3DDevice& winrtDevice,
DeviceCreationLocation deviceCreationLocation, uint32_t iteration,
Profiler<WINML_MODEL_TEST_PERF>& profiler)
{
if (deviceType == DeviceType::CPU && inputDataType == InputDataType::Tensor && inputBindingType == InputBindingType::GPU)
if (deviceType == DeviceType::CPU && inputDataType == InputDataType::Tensor &&
inputBindingType == InputBindingType::GPU)
{
std::cout << "Cannot create D3D12 device on client if CPU device type is selected." << std::endl;
return E_INVALIDARG;
@ -238,7 +240,6 @@ HRESULT BindInputs(LearningModelBinding &context, const LearningModel& model, co
return hr.code();
}
HRESULT bindInputResult =
BindInputFeatures(model, context, inputFeatures, args, output, captureIterationPerf, iteration, profiler);
@ -277,16 +278,16 @@ std::vector<std::wstring> GetModelsInDirectory(CommandLineArgs& args, OutputHelp
return modelPaths;
}
HRESULT CheckIfModelAndConfigurationsAreSupported(LearningModel& model, const std::wstring& modelPath,
const DeviceType deviceType,
const std::vector<InputDataType>& inputDataTypes,
const std::vector<DeviceCreationLocation>& deviceCreationLocations)
{
// Does user want image as input binding
bool hasInputBindingImage = std::any_of(inputDataTypes.begin(), inputDataTypes.end(), [](const InputDataType inputDataType){
bool hasInputBindingImage =
std::any_of(inputDataTypes.begin(), inputDataTypes.end(), [](const InputDataType inputDataType) {
return inputDataType == InputDataType::ImageBGR || inputDataType == InputDataType::ImageRGB;
});
});
for (auto inputFeature : model.InputFeatures())
{
@ -345,7 +346,6 @@ HRESULT EvaluateModel(LearningModelEvaluationResult& result, const LearningModel
if (args.IsPerIterationCapture())
{
output.SaveEvalPerformance(profiler, iterationNum);
}
}
}
@ -496,18 +496,19 @@ int run(CommandLineArgs& args, Profiler<WINML_MODEL_TEST_PERF>& profiler) try
ConcurrentLoadModel(modelPaths, args.NumThreads(), args.ThreadInterval(), true);
return 0;
}
for (const auto& path: modelPaths)
for (const auto& path : modelPaths)
{
LearningModel model = nullptr;
LoadModel(model, path, args.IsPerformanceCapture() || args.IsPerIterationCapture(), output, args, 0, profiler);
LoadModel(model, path, args.IsPerformanceCapture() || args.IsPerIterationCapture(), output, args, 0,
profiler);
for (auto deviceType : deviceTypes)
{
lastHr = CheckIfModelAndConfigurationsAreSupported(model, path, deviceType, inputDataTypes,
deviceCreationLocations);
deviceCreationLocations);
if (FAILED(lastHr))
{
continue;
continue;
}
for (auto deviceCreationLocation : deviceCreationLocations)
{
@ -523,27 +524,28 @@ int run(CommandLineArgs& args, Profiler<WINML_MODEL_TEST_PERF>& profiler) try
LearningModelSession session = nullptr;
IDirect3DDevice winrtDevice = nullptr;
lastHr = CreateSession(session, winrtDevice, model, args, output, deviceType,
deviceCreationLocation, profiler);
deviceCreationLocation, profiler);
if (FAILED(lastHr))
{
continue;
}
for (auto inputDataType : inputDataTypes)
{
for (auto inputBindingType : inputBindingTypes)
{
// Clear up bind, eval performance metrics after iteration
if (args.IsPerformanceCapture() || args.IsPerIterationCapture())
{
// Resets all values from profiler for bind and evaluate.
profiler.Reset(WINML_MODEL_TEST_PERF::BIND_VALUE, WINML_MODEL_TEST_PERF::COUNT);
}
for (auto inputBindingType : inputBindingTypes)
{
// Clear up bind, eval performance metrics after iteration
if (args.IsPerformanceCapture() || args.IsPerIterationCapture())
{
// Resets all values from profiler for bind and evaluate.
profiler.Reset(WINML_MODEL_TEST_PERF::BIND_VALUE, WINML_MODEL_TEST_PERF::COUNT);
}
for (uint32_t i = 0; i < args.NumIterations(); i++)
{
#if defined(_AMD64_)
// PIX markers only work on AMD64
// If PIX tool was attached then capture already began for the first iteration before session creation.
// This is to begin PIX capture for each iteration after the first iteration.
// If PIX tool was attached then capture already began for the first iteration before
// session creation. This is to begin PIX capture for each iteration after the first
// iteration.
if (i > 0 && output.GetGraphicsAnalysis())
{
output.GetGraphicsAnalysis()->BeginCapture();
@ -551,15 +553,15 @@ int run(CommandLineArgs& args, Profiler<WINML_MODEL_TEST_PERF>& profiler) try
#endif
LearningModelBinding context(session);
lastHr = BindInputs(context, model, session, output, deviceType, args, inputBindingType,
inputDataType, winrtDevice, deviceCreationLocation, i, profiler);
inputDataType, winrtDevice, deviceCreationLocation, i, profiler);
if (FAILED(lastHr))
{
break;
}
LearningModelEvaluationResult result = nullptr;
bool capture_perf = args.IsPerformanceCapture() || args.IsPerIterationCapture();
lastHr = EvaluateModel(result, model, context, session, args, output,
capture_perf, i, profiler);
lastHr = EvaluateModel(result, model, context, session, args, output, capture_perf, i,
profiler);
if (FAILED(lastHr))
{
output.PrintEvaluatingInfo(i + 1, deviceType, inputBindingType, inputDataType,
@ -604,8 +606,12 @@ int run(CommandLineArgs& args, Profiler<WINML_MODEL_TEST_PERF>& profiler) try
inputBindingTypeStringified,
deviceCreationLocationStringified);
}
else if (args.IsPerIterationCapture() && args.IsImageInput())
{
output.WritePerIterationPerformance(args, args.ModelPath(), args.ImagePath());
}
}
}
}
}
}
}