microsoft
/
CNTK
зеркало из
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность

Changing %ws to %ls in *printf() within the MachineLearning/cn/* source

This commit is contained in:
Jasha Droppo 2014-11-05 03:23:48 +00:00
Родитель 4d46e5f8dd
Коммит d044c66fe8
8 изменённых файлов: 43 добавлений и 43 удалений
Показать статистику Скачать Patch файл Скачать Diff файл Показать все файлы Свернуть все файлы

14
MachineLearning/cn/CompositeComputationNode.h
Просмотреть файл

@ -56,7 +56,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
char str[4096];
sprintf(str, "[%lu,%lu] ", FunctionValues().GetNumRows(), FunctionValues().GetNumCols());
fstream << string(str);
sprintf(str, "HasComputed=%ws", HasComputed()? L"true" : L"false");
sprintf(str, "HasComputed=%ls", HasComputed()? L"true" : L"false");
fstream << string(str);
PrintNodeValuesToFile(printValues, fstream);
@ -1057,7 +1057,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
if (m_children[0]->FunctionValues().GetNumCols() != weightCols || m_children[0]->FunctionValues().GetNumRows() != m_outputChannels)
{
msra::strfun::strprintf msg("convolutionWeight matrix %ws should have dimension [%d, %d] which is [outputChannels, kernelWidth * kernelHeight * inputChannels]",
msra::strfun::strprintf msg("convolutionWeight matrix %ls should have dimension [%d, %d] which is [outputChannels, kernelWidth * kernelHeight * inputChannels]",
m_children[0]->NodeName().c_str(), m_outputChannels, weightCols);
throw std::logic_error(msg.c_str());
}
@ -1070,7 +1070,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
if (m_children[1]->FunctionValues().GetNumRows() != inputDim)
{
msra::strfun::strprintf msg("each column of input to the convolution node %ws is a sample and should have dimension %d, which is inputWidth * inputHeight * inputChannels",
msra::strfun::strprintf msg("each column of input to the convolution node %ls is a sample and should have dimension %d, which is inputWidth * inputHeight * inputChannels",
NodeName().c_str(), inputDim);
throw std::logic_error(msg.c_str());
}
@ -1132,7 +1132,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
fstream << string(str);
sprintf(str, "Output[Width:%lu, Height:%lu, Channels:%lu] \n", m_outputWidth, m_outputHeight, m_outputChannels);
fstream << string(str);
sprintf(str, "ZeroPadding=%ws maxTempMemSizeInSamples=%lu\n", m_zeroPadding? L"true" : L"false", m_maxTempMemSizeInSamples);
sprintf(str, "ZeroPadding=%ls maxTempMemSizeInSamples=%lu\n", m_zeroPadding? L"true" : L"false", m_maxTempMemSizeInSamples);
fstream << string(str);
}
@ -1471,7 +1471,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
if (m_children[0]->FunctionValues().GetNumRows() != m_inputSizePerSample)
{
msra::strfun::strprintf msg("each column of input to the MaxPooling node %ws is a sample and should have dimension %d, which is inputWidth * inputHeight * inputChannels",
msra::strfun::strprintf msg("each column of input to the MaxPooling node %ls is a sample and should have dimension %d, which is inputWidth * inputHeight * inputChannels",
NodeName().c_str(), m_inputSizePerSample);
throw std::logic_error(msg.c_str());
}
@ -1732,7 +1732,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
if (m_children[0]->FunctionValues().GetNumRows() != m_inputSizePerSample)
{
msra::strfun::strprintf msg("each column of input to the AveragePooling node %ws is a sample and should have dimension %d, which is inputWidth * inputHeight * inputChannels",
msra::strfun::strprintf msg("each column of input to the AveragePooling node %ls is a sample and should have dimension %d, which is inputWidth * inputHeight * inputChannels",
NodeName().c_str(), m_inputSizePerSample);
throw std::logic_error(msg.c_str());
}
@ -2240,4 +2240,4 @@ namespace Microsoft { namespace MSR { namespace CNTK {
template class GMMLogLikelihoodNode<float>;
template class GMMLogLikelihoodNode<double>;
}}}
}}}

6
MachineLearning/cn/ComputationNetwork.cpp
Просмотреть файл

@ -88,7 +88,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
newNode = new LookupTableNode<ElemType>(fstream, m_deviceId, nodeName);
else
{
fprintf(stderr, "Error creating new ComputationNode of type %ws, with name %ws\n", nodeType.c_str(), nodeName.c_str());
fprintf(stderr, "Error creating new ComputationNode of type %ls, with name %ls\n", nodeType.c_str(), nodeName.c_str());
throw std::invalid_argument("Invalid node type.");
}
@ -162,7 +162,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
newNode = new LookupTableNode<ElemType>(m_deviceId, nodeName);
else
{
fprintf(stderr, "Error creating new ComputationNode of type %ws, with name %ws\n", nodeType.c_str(), nodeName.c_str());
fprintf(stderr, "Error creating new ComputationNode of type %ls, with name %ls\n", nodeType.c_str(), nodeName.c_str());
throw std::invalid_argument("Invalid node type.");
}
@ -171,4 +171,4 @@ namespace Microsoft { namespace MSR { namespace CNTK {
}
}}}
}}}

18
MachineLearning/cn/ComputationNetwork.h
Просмотреть файл

@ -112,7 +112,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
}
else //node name is not found, dump all nodes
{
fprintf (stderr, "Warning: node name %ws does not exist in the network. dumping all nodes.\n", nodeName.c_str());
fprintf (stderr, "Warning: node name %ls does not exist in the network. dumping all nodes.\n", nodeName.c_str());
DumpAllNodesToFile(printValues, outputFile);
}
}
@ -780,7 +780,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
newNode = new GMMLogLikelihoodNode<ElemType>(fstream, modelVersion, m_deviceId, nodeName);
else
{
fprintf(stderr, "Error creating new ComputationNode of type %ws, with name %ws\n", nodeType.c_str(), nodeName.c_str());
fprintf(stderr, "Error creating new ComputationNode of type %ls, with name %ls\n", nodeType.c_str(), nodeName.c_str());
throw std::invalid_argument("Invalid node type.");
}
@ -934,7 +934,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
newNode = new GMMLogLikelihoodNode<ElemType>(m_deviceId, nodeName);
else
{
fprintf(stderr, "Error creating new ComputationNode of type %ws, with name %ws\n", nodeType.c_str(), nodeName.c_str());
fprintf(stderr, "Error creating new ComputationNode of type %ls, with name %ls\n", nodeType.c_str(), nodeName.c_str());
throw std::invalid_argument("Invalid node type.");
}
@ -1428,7 +1428,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
fprintf (stderr, "Evaluate Node: %s\n",(msra::strfun::utf8 ((*nodeIter)->NodeName())).c_str());
#endif
#if DUMPOUTPUT
fprintf(stderr,"Forward_%ws\n",(*nodeIter)->NodeName().c_str());
fprintf(stderr,"Forward_%ls\n",(*nodeIter)->NodeName().c_str());
#endif
(*nodeIter)->EvaluateThisNode(); // we manage time stamp here so that derived classes don't need to worry about it
(*nodeIter)->UpdateEvalTimeStamp();
@ -1776,7 +1776,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
void ValidateNetwork(const ComputationNodePtr rootNode)
{
fprintf(stderr, "\n\nValidating node %ws \n", rootNode->NodeName().c_str());
fprintf(stderr, "\n\nValidating node %ls \n", rootNode->NodeName().c_str());
std::list<ComputationNodePtr>& nodes = GetEvalOrder(rootNode);
@ -1908,7 +1908,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
fprintf(stderr, " nodes in the recurrent loops : \n");
for (auto itr = (*iter).m_recurrentNodes.begin(); itr != (*iter).m_recurrentNodes.end(); itr++)
{
fprintf (stderr, "%ws\t", (*itr)->NodeName().c_str() );
fprintf (stderr, "%ls\t", (*itr)->NodeName().c_str() );
}
}
@ -1939,7 +1939,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
fprintf(stderr, " nodes in the recurrent loops : \n");
for (auto itr = (*iter).m_recurrentNodes.begin(); itr != (*iter).m_recurrentNodes.end(); itr++)
{
fprintf (stderr, "%ws\t", (*itr)->NodeName().c_str() );
fprintf (stderr, "%ls\t", (*itr)->NodeName().c_str() );
}
}
}
@ -2047,7 +2047,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
size_t max_visitedOrderInLoop = 0;
for (auto itr = (*iter).m_recurrentNodes.begin(); itr != (*iter).m_recurrentNodes.end(); itr++)
{
fprintf (stderr, "%ws\t", (*itr)->NodeName().c_str() );
fprintf (stderr, "%ls\t", (*itr)->NodeName().c_str() );
if (max_visitedOrderInLoop < (*itr)->GetVisitedOrder())
{
max_visitedOrderInLoop = (*itr)->GetVisitedOrder();
@ -2135,7 +2135,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
fprintf(stderr, "Reordered nodes\n");
for (auto itr = nodes.begin(); itr != nodes.end(); itr++)
{
fprintf (stderr, "%ws\n", (*itr)->NodeName().c_str() );
fprintf (stderr, "%ls\n", (*itr)->NodeName().c_str() );
}
#endif
}

12
MachineLearning/cn/ComputationNode.h
Просмотреть файл

@ -290,7 +290,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
//for debugging purpose
virtual void PrintSelf(bool printMatrices = false) const
{
fprintf(stderr, "\n%ws[%lu, %lu] = %ws", NodeName().c_str(), FunctionValues().GetNumRows(), FunctionValues().GetNumCols(), OperationName().c_str());
fprintf(stderr, "\n%ls[%lu, %lu] = %ls", NodeName().c_str(), FunctionValues().GetNumRows(), FunctionValues().GetNumCols(), OperationName().c_str());
if (!IsLeaf())
{
@ -299,7 +299,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
{
if (i > 0)
fprintf(stderr, ", ");
fprintf(stderr, "%ws[%lu, %lu]", Inputs(i)?Inputs(i)->NodeName().c_str():L"NULL", Inputs(i)->FunctionValues().GetNumRows(), Inputs(i)->FunctionValues().GetNumCols());
fprintf(stderr, "%ls[%lu, %lu]", Inputs(i)?Inputs(i)->NodeName().c_str():L"NULL", Inputs(i)->FunctionValues().GetNumRows(), Inputs(i)->FunctionValues().GetNumCols());
}
fprintf(stderr, ")");
}
@ -417,7 +417,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
(msra::strfun::utf8 (child->NodeName())).c_str());
#endif
#if DUMPOUTPUT
fprintf(stderr,"Backprop%d_%ws\n",i,NodeName().c_str());
fprintf(stderr,"Backprop%d_%ls\n",i,NodeName().c_str());
#endif
ComputeInputPartial(i); //this computes partial wrt to the child and sums the gradient value in the child
}
@ -600,7 +600,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
virtual void PrintSelfBeforeValidation(bool allowNulls=false) const
{
fprintf(stderr, "\nValidating --> %ws = %ws", NodeName().c_str(), OperationName().c_str());
fprintf(stderr, "\nValidating --> %ls = %ls", NodeName().c_str(), OperationName().c_str());
if (!IsLeaf())
{
@ -623,10 +623,10 @@ namespace Microsoft { namespace MSR { namespace CNTK {
if (IsChildAnImage(i)) //image
fprintf(stderr, "%ws[%lu {W=%lu, H=%lu, C=%lu}, %lu]", child->NodeName().c_str(), child->FunctionValues().GetNumRows(),
fprintf(stderr, "%ls[%lu {W=%lu, H=%lu, C=%lu}, %lu]", child->NodeName().c_str(), child->FunctionValues().GetNumRows(),
child->m_outputWidth, child->m_outputHeight, child->m_outputChannels, child->FunctionValues().GetNumCols());
else
fprintf(stderr, "%ws[%lu, %lu]", child->NodeName().c_str(), child->FunctionValues().GetNumRows(), child->FunctionValues().GetNumCols());
fprintf(stderr, "%ls[%lu, %lu]", child->NodeName().c_str(), child->FunctionValues().GetNumRows(), child->FunctionValues().GetNumCols());
}
fprintf(stderr, ")");

4
MachineLearning/cn/PTaskGraphBuilder.cpp
Просмотреть файл

@ -842,7 +842,7 @@ void PTaskGraphBuilder<ElemType>::CreateTaskDescriptorsForComputationNodes()
ComputationNodePtr node = *nodeIter;
std::wstring opName = node->OperationName();
if (m_verbosity >= 1) fprintf(stderr, " %ws(%ws): ",
if (m_verbosity >= 1) fprintf(stderr, " %ls(%ls): ",
opName.c_str(), node->NodeName().c_str());
// Learnable parameter node types.
@ -905,7 +905,7 @@ void PTaskGraphBuilder<ElemType>::CreateTaskDescriptorsForComputationNodes()
}
else
{
fprintf(stderr, "PTaskGraphBuilder does not (yet) support ComputationNode type %ws.\n",
fprintf(stderr, "PTaskGraphBuilder does not (yet) support ComputationNode type %ls.\n",
opName.c_str());
throw exception("Unsupported computation node type");
}

22
MachineLearning/cn/SGD.h
Просмотреть файл

@ -309,12 +309,12 @@ namespace Microsoft { namespace MSR { namespace CNTK {
if (startEpoch >= 0)
{
wstring modelFileName = GetModelNameForEpoch(int(startEpoch)-1);
fprintf(stderr, "Starting from checkpoint. Load Network From File %ws.\n", modelFileName.c_str());
fprintf(stderr, "Starting from checkpoint. Load Network From File %ls.\n", modelFileName.c_str());
net.LoadFromFile(modelFileName);
}
else
{
fprintf(stderr, "Load Network From the original model file %ws.\n", origModelFileName.c_str());
fprintf(stderr, "Load Network From the original model file %ls.\n", origModelFileName.c_str());
net.LoadFromFile(origModelFileName);
}
@ -324,7 +324,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
m_needRegularization = m_adaptationRegType != AdaptationRegType::None && m_adaptationRegWeight > 0;
if (m_needRegularization)
{
fprintf(stderr, "Load reference Network From the original model file %ws.\n", origModelFileName.c_str());
fprintf(stderr, "Load reference Network From the original model file %ls.\n", origModelFileName.c_str());
refNet.LoadFromFile(origModelFileName);
}
@ -355,7 +355,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
wstring modelFileName = GetModelNameForEpoch(int(startEpoch)-1);
if (startEpoch >= 0)
fprintf(stderr, "Starting from checkpoint. Load Network From File %ws.\n", modelFileName.c_str());
fprintf(stderr, "Starting from checkpoint. Load Network From File %ls.\n", modelFileName.c_str());
ComputationNetwork<ElemType>& net =
startEpoch<0? netBuilder->BuildNetworkFromDescription() : netBuilder->LoadNetworkFromFile(modelFileName);
startEpoch = max(startEpoch, 0);
@ -569,9 +569,9 @@ namespace Microsoft { namespace MSR { namespace CNTK {
for (size_t j = 0; j < epochEvalErrors.size(); j++)
fprintf(stderr, "[%lu]=%.8g ", j, epochEvalErrors[j]);
fprintf(stderr, "Ave Learn Rate Per Sample = %.10g Epoch Time=%.8g\n", learnRatePerSample, epochTime);
fprintf(stderr, "Finished Epoch[%lu]: Criterion Node [%ws] Per Sample = %.8g\n", i + 1, criterionNodes[0]->NodeName().c_str(), epochCriterion);
fprintf(stderr, "Finished Epoch[%lu]: Criterion Node [%ls] Per Sample = %.8g\n", i + 1, criterionNodes[0]->NodeName().c_str(), epochCriterion);
for (size_t j = 0; j < epochEvalErrors.size(); j++)
fprintf(stderr, "Finished Epoch[%lu]: Evaluation Node [%ws] Per Sample = %.8g\n", i + 1, evalNodeNames[j].c_str(), epochEvalErrors[j]);
fprintf(stderr, "Finished Epoch[%lu]: Evaluation Node [%ls] Per Sample = %.8g\n", i + 1, evalNodeNames[j].c_str(), epochEvalErrors[j]);
}
if (validationSetDataReader != trainSetDataReader && validationSetDataReader != nullptr)
@ -697,7 +697,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
for (auto nodeIter = nodes.begin(); nodeIter != nodes.end(); nodeIter++)
{
PreComputedNode<ElemType>* node = static_cast<PreComputedNode<ElemType>*> (*nodeIter);
fprintf(stderr, "\tNodeName: %ws\n", (node->NodeName()).c_str());
fprintf(stderr, "\tNodeName: %ls\n", (node->NodeName()).c_str());
}
//compute
@ -1143,7 +1143,7 @@ protected:
void UpdateWeights(const ComputationNodePtr node, Matrix<ElemType>& smoothedGradient, const ElemType learnRatePerSample, const size_t actualMBSize, const size_t expectedMBSize) const
{
#if DUMPOUTPUT
fprintf(stderr, "Update_%ws\n",node->NodeName().c_str());
fprintf(stderr, "Update_%ls\n",node->NodeName().c_str());
#endif
UpdateWeightsS(this, node->FunctionValues(), node->GradientValues(), smoothedGradient, learnRatePerSample, actualMBSize, expectedMBSize);
node->UpdateEvalTimeStamp();
@ -1330,7 +1330,7 @@ protected:
irow = max(0, irow);
icol = max(0, icol);
fprintf(stderr, "\n###### d%ws######\n", node->NodeName().c_str());
fprintf(stderr, "\n###### d%ls######\n", node->NodeName().c_str());
// node->FunctionValues().Print();
ElemType eOrg = node->FunctionValues()(irow,icol);
@ -1363,7 +1363,7 @@ protected:
bool wrong = (std::isnan(diff) || diff > threshold);
if (wrong)
{
fprintf (stderr, "\nd%ws Numeric gradient = %e, Error BP gradient = %e\n", node->NodeName().c_str(), eGradNum, eGradErr);
fprintf (stderr, "\nd%ls Numeric gradient = %e, Error BP gradient = %e\n", node->NodeName().c_str(), eGradNum, eGradErr);
return false;
}
}
@ -1453,4 +1453,4 @@ protected:
template class SGD<float>;
template class SGD<double>;
}}}
}}}

2
MachineLearning/cn/SimpleEvaluator.h
Просмотреть файл

@ -334,7 +334,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
for (size_t i=0; i<evalResults.size(); i++)
{
fprintf(stderr, "%ws/Sample = %.8g ", evalNodes[i]->NodeName().c_str(), (evalResults[i]-evalResultsLastMBs[i])/numSamplesLastMBs);
fprintf(stderr, "%ls/Sample = %.8g ", evalNodes[i]->NodeName().c_str(), (evalResults[i]-evalResultsLastMBs[i])/numSamplesLastMBs);
}
fprintf(stderr, "\n");

8
MachineLearning/cn/cn.cpp
Просмотреть файл

@ -183,11 +183,11 @@ void DoCrossValidate(const ConfigParameters& config)
bool finalModelEvaluated = false;
for (size_t i=cvInterval[0]; i<=cvInterval[2]; i+=cvInterval[1])
{
wstring cvModelPath = msra::strfun::wstrprintf (L"%ws.%lld", modelPath.c_str(), i);
wstring cvModelPath = msra::strfun::wstrprintf (L"%ls.%lld", modelPath.c_str(), i);
if (!fexists (cvModelPath))
{
fprintf(stderr, "model %ws does not exist.\n", cvModelPath.c_str());
fprintf(stderr, "model %ls does not exist.\n", cvModelPath.c_str());
if (finalModelEvaluated || !fexists (modelPath))
continue; // file missing
else
@ -204,7 +204,7 @@ void DoCrossValidate(const ConfigParameters& config)
SimpleEvaluator<ElemType> eval(net, numMBsToShowResult, traceLevel);
fprintf(stderr, "model %ws --> \n",cvModelPath.c_str());
fprintf(stderr, "model %ls --> \n",cvModelPath.c_str());
std::vector<ElemType> evalErrors;
evalErrors = eval.Evaluate(cvDataReader, evalNodeNamesVector, mbSize[0], epochSize);
cvErrorResults.push_back(evalErrors);
@ -242,7 +242,7 @@ void DoCrossValidate(const ConfigParameters& config)
fprintf(stderr,"------------\n");
for (int i=0; i < minErrors.size(); ++i)
{
fprintf(stderr,"Based on Err[%d]: Best model = %ws with min err %.8g\n", i, cvModels[minErrIds[i]].c_str(), minErrors[i]);
fprintf(stderr,"Based on Err[%d]: Best model = %ls with min err %.8g\n", i, cvModels[minErrIds[i]].c_str(), minErrors[i]);
}
}