Fix typos

Documentation/Documents/Configuration Files.md

@@ -685,7 +685,7 @@ The following parameters can be used to customize the behavior of the reader:
 
 -   **randomize** – \[{Auto}, None, \#\] the randomization range (number of records to randomize across) for randomizing the input. This needs to be an integral factor of the epochSize and an integral multiple of minibatch size. Setting it to Auto will let CNTK find something that works.
 
--   **minibatchMode** – \[{Partial},Full\] the mode for minibatchs when the end of the epoch is reached. In partial minibatch mode, if the remaining records are less than a full minibatch, only those read will be returned (a partial minibatch). I Full minibatch mode, no partial minibatches will be returned, instead those records will be skipped.
+-   **minibatchMode** – \[{Partial},Full\] the mode for minibatches when the end of the epoch is reached. In partial minibatch mode, if the remaining records are less than a full minibatch, only those read will be returned (a partial minibatch). I Full minibatch mode, no partial minibatches will be returned, instead those records will be skipped.
 
 Each of the data record sub-sections have the following parameters:
 
@@ -717,7 +717,7 @@ For training and evaluation the following need to be defined:
 
 -   **framemode** – \[{true}, false\] is the reader reading frames, or utterances
 
--   **minibatchMode** – \[{Partial},Full\] the mode for minibatchs when the end of the epoch is reached. In partial minibatch mode, if the remaining records are less than a full minibatch, only those read will be returned (a partial minibatch). I Full minibatch mode, no partial minibatches will be returned, instead those records will be skipped.
+-   **minibatchMode** – \[{Partial},Full\] the mode for minibatches when the end of the epoch is reached. In partial minibatch mode, if the remaining records are less than a full minibatch, only those read will be returned (a partial minibatch). I Full minibatch mode, no partial minibatches will be returned, instead those records will be skipped.
 
 -   **readAhead** – \[true,{false}\] have the reader read ahead in another thread. NOTE: some known issues with this feature
 
@@ -893,7 +893,7 @@ The parameters used for the binary reader are quite simple as most of the requir
 
 The following parameters can be used to customize the behavior of the reader:
 
--   **minibatchMode** – \[{Partial},Full\] the mode for minibatchs when the end of the epoch is reached. In partial minibatch mode, if the remaining records are less than a full minibatch, only those read will be returned (a partial minibatch). I Full minibatch mode, no partial minibatches will be returned, instead those records will be skipped.
+-   **minibatchMode** – \[{Partial},Full\] the mode for minibatches when the end of the epoch is reached. In partial minibatch mode, if the remaining records are less than a full minibatch, only those read will be returned (a partial minibatch). I Full minibatch mode, no partial minibatches will be returned, instead those records will be skipped.
 
 -   **file** – array of files paths to load. Each file may contain one or more datasets. The dataset names used when the file was created will be used when the file is read.
 

Source/CNTKv2LibraryDll/DistributedLearnerBase.cpp

@@ -37,7 +37,7 @@ namespace CNTK
 
     void DistributedLearnerBase::PrepaireZeroGradients(std::unordered_map<Parameter, NDArrayViewPtr>& gradientValues, MinibatchInfo& info)
     {
-        // Need to intialize gradients to 0 in case when it is an empty minibatch.
+        // Need to initialize gradients to 0 in case when it is an empty minibatch.
         for (auto& g : gradientValues)
         {
             auto weights = g.first.Value();

Source/Common/DataReader.cpp

@@ -240,7 +240,7 @@ size_t DataReader::GetCurrentSamplePosition()
 // GetMinibatch - Get the next minibatch (features and labels)
 // matrices - [in] a map with named matrix types (i.e. 'features', 'labels') mapped to the corresponding matrix,
 //             [out] each matrix resized if necessary containing data.
-// returns - true if there are more minibatches, false if no more minibatchs remain
+// returns - true if there are more minibatches, false if no more minibatches remain
 bool DataReader::GetMinibatch(StreamMinibatchInputs& matrices)
 {
     /**
@@ -273,7 +273,7 @@ bool DataReader::GetMinibatch(StreamMinibatchInputs& matrices)
 // latticeinput - lattice for each utterances in this minibatch
 // uids - lables stored in size_t vector instead of ElemType matrix
 // boundary - phone boundaries
-// returns - true if there are more minibatches, false if no more minibatchs remain
+// returns - true if there are more minibatches, false if no more minibatches remain
 bool DataReader::GetMinibatch4SE(std::vector<shared_ptr<const msra::dbn::latticepair>>& latticeinput, vector<size_t>& uids, vector<size_t>& boundaries, vector<size_t>& extrauttmap)
 {
     bool bRet = true;

Source/Common/Include/ASGDHelper.h

@@ -14,7 +14,7 @@ namespace Microsoft { namespace MSR { namespace CNTK {
 //       Providing option for DataParallelASGD training. so that every nodes
 //       could adjust learning rate every minibatch at first N epochs.
 // -----------------------------------------------------------------------
-// TODO: We can removed these options once we can adjust learning rate at minibatchs level
+// TODO: We can removed these options once we can adjust learning rate at minibatches level
 enum class AdjustLearningRateAtBeginning : int
 {
     None = 0,  // default, don't adjust learning rate

Source/Common/Include/DataReader.h

@@ -442,7 +442,7 @@ class DataReader : public IDataReader, protected Plugin, public ScriptableObject
     // GetMinibatch - Get the next minibatch (features and labels)
     // matrices - [in] a map with named matrix types (i.e. 'features', 'labels') mapped to the corresponding matrix,
     //             [out] each matrix resized if necessary containing data.
-    // returns - true if there are more minibatches, false if no more minibatchs remain
+    // returns - true if there are more minibatches, false if no more minibatches remain
     virtual bool GetMinibatch(StreamMinibatchInputs& matrices);
     virtual bool GetMinibatch4SE(std::vector<shared_ptr<const msra::dbn::latticepair>>& latticeinput, vector<size_t>& uids, vector<size_t>& boundaries, vector<size_t>& extrauttmap);
     virtual bool GetHmmData(msra::asr::simplesenonehmm* hmm);

Source/Common/Include/latticestorage.h

@@ -13,7 +13,7 @@
 #include <stdint.h>
 #include <cstdio>
 
-#undef INITIAL_STRANGE // [v-hansu] intialize structs to strange values
+#undef INITIAL_STRANGE // [v-hansu] initialize structs to strange values
 #define PARALLEL_SIL   // [v-hansu] process sil on CUDA, used in other files, please search this
 #define LOGZERO -1e30f
 

Source/ComputationNetworkLib/RecurrentNodes.cpp

@@ -235,7 +235,7 @@ template<class ElemType, int direction>
         //     - matrix column indices into the initial state
         //        - if initial-state sequence has >1 steps, then index from back
         //        - if 1 step, then broadcast that to all
-        //     - or -1 for non-boundary entires
+        //     - or -1 for non-boundary entries
 
         // our own output MB layout
         let& outMBLayout = GetMBLayout();

Source/EvalDll/EvalReader.h

@@ -112,7 +112,7 @@ class EvalReader : public DataReaderBase
     // TryGetMinibatch - Get the next minibatch (features and labels)
     // matrices - [in] a map with named matrix types (i.e. 'features', 'labels') mapped to the corresponding matrix,
     //             [out] each matrix resized if necessary containing data.
-    // returns - true if there are more minibatches, false if no more minibatchs remain
+    // returns - true if there are more minibatches, false if no more minibatches remain
     virtual bool TryGetMinibatch(StreamMinibatchInputs& matrices)
     {
         // how many records are we reading this time

Source/Readers/BinaryReader/BinaryReader.cpp

@@ -249,7 +249,7 @@ bool BinaryReader<ElemType>::CheckEndDataset(size_t actualmbsize)
 // GetMinibatch - Get the next minibatch (features and labels)
 // matrices - [in] a map with named matrix types (i.e. 'features', 'labels') mapped to the corresponding matrix,
 //             [out] each matrix resized if necessary containing data.
-// returns - true if there are more minibatches, false if no more minibatchs remain
+// returns - true if there are more minibatches, false if no more minibatches remain
 template <class ElemType>
 bool BinaryReader<ElemType>::TryGetMinibatch(StreamMinibatchInputs& matrices)
 {

Source/Readers/DSSMReader/DSSMReader.cpp

@@ -318,7 +318,7 @@ void DSSMReader<ElemType>::StoreLabel(ElemType& labelStore, const LabelType& lab
 // GetMinibatch - Get the next minibatch (features and labels)
 // matrices - [in] a map with named matrix types (i.e. 'features', 'labels') mapped to the corresponding matrix,
 //             [out] each matrix resized if necessary containing data.
-// returns - true if there are more minibatches, false if no more minibatchs remain
+// returns - true if there are more minibatches, false if no more minibatches remain
 template <class ElemType>
 bool DSSMReader<ElemType>::TryGetMinibatch(StreamMinibatchInputs& matrices)
 {

Source/Readers/HTKMLFReader/HTKMLFReader.cpp

@@ -933,7 +933,7 @@ bool HTKMLFReader<ElemType>::GetHmmData(msra::asr::simplesenonehmm* hmm)
 // GetMinibatch - Get the next minibatch (features and labels)
 // matrices - [in] a map with named matrix types (i.e. 'features', 'labels') mapped to the corresponding matrix,
 //             [out] each matrix resized if necessary containing data.
-// returns - true if there are more minibatches, false if no more minibatchs remain
+// returns - true if there are more minibatches, false if no more minibatches remain
 // TODO: Why do we have two read functions? Is one not a superset of the other?
 template <class ElemType>
 bool HTKMLFReader<ElemType>::TryGetMinibatch(StreamMinibatchInputs& matrices)

Source/Readers/Kaldi2Reader/HTKMLFReader.cpp

@@ -844,7 +844,7 @@ void HTKMLFReader<ElemType>::StartMinibatchLoopToWrite(size_t mbSize, size_t /*e
 // GetMinibatch - Get the next minibatch (features and labels)
 // matrices - [in] a map with named matrix types (i.e. 'features', 'labels') mapped to the corresponding matrix,
 //             [out] each matrix resized if necessary containing data.
-// returns - true if there are more minibatches, false if no more minibatchs remain
+// returns - true if there are more minibatches, false if no more minibatches remain
 template <class ElemType>
 bool HTKMLFReader<ElemType>::TryGetMinibatch(StreamMinibatchInputs& matrices)
 {

Source/Readers/SparsePCReader/SparsePCReader.cpp

@@ -205,7 +205,7 @@ void SparsePCReader<ElemType>::StartMinibatchLoop(size_t mbSize, size_t /*epoch*
 // GetMinibatch - Get the next minibatch (features and labels)
 // matrices - [in] a map with named matrix types (i.e. 'features', 'labels') mapped to the corresponding matrix,
 //             [out] each matrix resized if necessary containing data.
-// returns - true if there are more minibatches, false if no more minibatchs remain
+// returns - true if there are more minibatches, false if no more minibatches remain
 template <class ElemType>
 bool SparsePCReader<ElemType>::TryGetMinibatch(StreamMinibatchInputs& matrices)
 {

Source/Readers/UCIFastReader/UCIFastReader.cpp

@@ -763,7 +763,7 @@ void UCIFastReader<ElemType>::StoreLabel(ElemType& labelStore, const LabelType&
 // GetMinibatch - Get the next minibatch (features and labels)
 // matrices - [in] a map with named matrix types (i.e. 'features', 'labels') mapped to the corresponding matrix,
 //             [out] each matrix resized if necessary containing data.
-// returns - true if there are more minibatches, false if no more minibatchs remain
+// returns - true if there are more minibatches, false if no more minibatches remain
 template <class ElemType>
 bool UCIFastReader<ElemType>::TryGetMinibatch(StreamMinibatchInputs& matrices)
 {
@@ -825,7 +825,7 @@ bool UCIFastReader<ElemType>::TryGetMinibatch(StreamMinibatchInputs& matrices)
 // GetMinibatchImpl - The actual implementation of getting the next minibatch (features and labels)
 // matrices - [in] a map with named matrix types (i.e. 'features', 'labels') mapped to the corresponding matrix,
 //             [out] each matrix resized if necessary containing data.
-// returns - true if there are more minibatches, false if no more minibatchs remain
+// returns - true if there are more minibatches, false if no more minibatches remain
 template <class ElemType>
 bool UCIFastReader<ElemType>::GetMinibatchImpl(StreamMinibatchInputs& matrices)
 {

Source/SGDLib/SGD.cpp

@@ -3262,7 +3262,7 @@ SGDParams::SGDParams(const ConfigRecordType& configSGD, size_t sizeofElemType)
 #endif
             m_isAsyncBufferEnabled = configDataParallelASGD(L"UsePipeline", false);
             m_isSimulateMA = configDataParallelASGD(L"SimModelAverage", false); // using parameter server-based version of ModelAveragingSGD
-            if (configDataParallelASGD.Exists(L"AdjustLearningRateAtBeginning")) // adjust learning rate per m_adjustNumInBatch minibatchs until to original one,
+            if (configDataParallelASGD.Exists(L"AdjustLearningRateAtBeginning")) // adjust learning rate per m_adjustNumInBatch minibatches until to original one,
                                                                                  // this option could be used to takcle the unstableness of DataParallelASGD if you get a chance
             {
                 const ConfigRecordType & configAdjustLearningRateAtBeginning(configDataParallelASGD(L"AdjustLearningRateAtBeginning", ConfigRecordType::Record()));

Tutorials/CNTK_103C_MNIST_MultiLayerPerceptron.ipynb

@@ -67,7 +67,7 @@
     "- Data reading: We will use the CNTK Text reader \n",
     "- Data preprocessing: Covered in part A (suggested extension section). \n",
     "\n",
-    "There is a high overlap with CNTK 102. Though this tutorial we adapt the same model to work on MNIST data with 10 classses instead of the 2 classes we used in CNTK 102.\n"
+    "There is a high overlap with CNTK 102. Though this tutorial we adapt the same model to work on MNIST data with 10 classes instead of the 2 classes we used in CNTK 102.\n"
    ]
   },
   {