Adapted Readme files wrt new directory structure

CNTK.sln

@@ -560,22 +560,22 @@ Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "VGG", "VGG", "{BC0D6DFF-80C
 		Examples\Image\Miscellaneous\ImageNet\VGG\VGG_E.ndl = Examples\Image\Miscellaneous\ImageNet\VGG\VGG_E.ndl
 	EndProjectSection
 EndProject
-Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Miscellaneous", "Miscellaneous", "{CCD56F12-BA17-4753-B5EE-4995FE682995}"
+Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Other", "Other", "{CCD56F12-BA17-4753-B5EE-4995FE682995}"
 EndProject
 Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Simple2d", "Simple2d", "{D2A060F1-128E-42A1-A0D0-3E3E1DFBC427}"
 	ProjectSection(SolutionItems) = preProject
-		Examples\Miscellaneous\Simple2d\README.md = Examples\Miscellaneous\Simple2d\README.md
+		Examples\Other\Simple2d\README.md = Examples\Other\Simple2d\README.md
 	EndProjectSection
 EndProject
 Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "NdlExamples", "NdlExamples", "{FC573A62-6DAE-40A4-8153-520C8571A007}"
 	ProjectSection(SolutionItems) = preProject
-		Examples\Miscellaneous\NdlExamples\NDLExamples.ndl = Examples\Miscellaneous\NdlExamples\NDLExamples.ndl
+		Examples\Other\NdlExamples\NDLExamples.ndl = Examples\Other\NdlExamples\NDLExamples.ndl
 	EndProjectSection
 EndProject
 Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Config", "Config", "{1E37CE40-556D-4693-B58C-F8D4CE349BB7}"
 	ProjectSection(SolutionItems) = preProject
-		Examples\Miscellaneous\Simple2d\Config\Multigpu.config = Examples\Miscellaneous\Simple2d\Config\Multigpu.config
-		Examples\Miscellaneous\Simple2d\Config\Simple.config = Examples\Miscellaneous\Simple2d\Config\Simple.config
+		Examples\Other\Simple2d\Config\Multigpu.config = Examples\Other\Simple2d\Config\Multigpu.config
+		Examples\Other\Simple2d\Config\Simple.config = Examples\Other\Simple2d\Config\Simple.config
 	EndProjectSection
 EndProject
 Project("{2150E333-8FDC-42A3-9474-1A3956D46DE8}") = "Miscellaneous", "Miscellaneous", "{BF1A621D-528B-4B84-AAFC-EF1455FC6830}"

Examples/Image/MNIST/README.md

@@ -2,13 +2,12 @@
 
 ## Overview
 
-| | |
-|:--------|:---|
-Data:     |The MNIST database (http://yann.lecun.com/exdb/mnist/) of handwritten digits.
-Purpose:  |This example demonstrates usage of NDL to train neural networks on MNIST dataset.
-Network:  |NDLNetworkBuilder, simple feed forward and convolutional networks, cross entropy with softmax.
-Training: |Stochastic gradient descent both with and without momentum.
-Comments: |There are two config files, details are provided below.
+|Data:     |The MNIST database (http://yann.lecun.com/exdb/mnist/) of handwritten digits.
+|:---------|:---
+|Purpose:  |This example demonstrates usage of the NDL (Network Description Language) to define networks.
+|Network:  |NDLNetworkBuilder, simple feed forward and convolutional networks, cross entropy with softmax.
+|Training: |Stochastic gradient descent both with and without momentum.
+|Comments: |There are two config files, details are provided below.
 
 ## Running the example
 
@@ -21,12 +20,12 @@ downloaded and converted by running the following command from the 'AdditionalFi
 
 The script will download all required files and convert them to CNTK-supported format. 
 The resulting files (Train-28x28.txt and Test-28x28.txt) will be stored in the 'Data' folder. 
-In case you don't have a Python installed, there are 2 options:
+In case you don't have Python installed, there are 2 options:
 
 1. Download and install latest version of Python 2.7 from: https://www.python.org/downloads/ 
-Then install numpy package by following instruction from: http://www.scipy.org/install.html#individual-packages
+Then install the numpy package by following instruction from: http://www.scipy.org/install.html#individual-packages
 
-2. Alternatively install Python Anaconda distribution which contains most of the 
+2. Alternatively install the Python Anaconda distribution which contains most of the 
 popular Python packages including numpy: http://continuum.io/downloads
 
 ### Setup

Examples/Image/README.md

@@ -1,12 +0,0 @@
-There are several examples for training different networks on image corpora 
-in the folder 'ExampleSetups/Image/' including:
-
-* MNIST
-* CIFAR-10
-* ImageNet
- 
-For those examples the data has to be downloaded and converted 
-(python scripts are provided in the corresponding folder).
-These examples also show how to train convolutional neural networks (CNN) with CNTK.
-
-We recommend to start with the MNIST example.

Examples/Other/Simple2d/README.md

@@ -2,19 +2,18 @@
 
 ## Overview
 
-| | |
-|:--------|:---|
-Data     |Two dimensional synthetic data
-Purpose  |Showcase how to train a simple CNTK network (CPU and GPU) and how to use it for scoring (decoding)
-Network  |SimpleNetworkBuilder, 2 hidden layers with 50 sigmoid nodes each, cross entropy with softmax
-Training |Stochastic gradient descent with momentum
-Comments |There are two config files: Simple.config uses a single CPU or GPU, Multigpu.config uses data-parallel SGD for training on multiple GPUs
+|Data     |Two dimensional synthetic data
+|:--------|:---
+|Purpose  |Showcase how to train a simple CNTK network (CPU and GPU) and how to use it for scoring (decoding)
+|Network  |SimpleNetworkBuilder, 2 hidden layers with 50 sigmoid nodes each, cross entropy with softmax
+|Training |Stochastic gradient descent with momentum
+|Comments |There are two config files: Simple.config uses a single CPU or GPU, Multigpu.config uses data-parallel SGD for training on multiple GPUs
 
 ## Running the example
 
 ### Getting the data
 
-The data for this example is already contained in the folder Demos/Simple2d/Data/.
+The data for this example is already contained in the folder Simple2d/Data/.
 
 ### Setup
 
@@ -30,16 +29,16 @@ or prefix the call to the cntk executable with the corresponding folder.
 
 ### Run
 
-Run the example from the Demos/Simple2d/Data folder using:
+Run the example from the Simple2d/Data folder using:
 
 `cntk configFile=../Config/Simple.config`
 
 or run from any folder and specify the Data folder as the `currentDirectory`, 
-e.g. running from the Demos/Simple2d folder using:
+e.g. running from the Simple2d folder using:
 
 `cntk configFile=Config/Simple.config currentDirectory=Data`
 
-The output folder will be created inside Demos/Simple2d/.
+The output folder will be created inside Simple2d/.
 
 ## Details
 
@@ -71,10 +70,13 @@ SimpleDemoDataReference.png shows a plot of the training data.
 
 ## Using a trained model
 
-The Test (e.g. Simple_Demo_Test) and the Output (e.g. Simple_Demo_Output) commands 
-specified in the config files use the trained model to compute labels for data specified in the SimpleDataTest.txt file. 
-The Test command computes prediction error, cross entropy and perplexity for the test set and outputs them to the console. 
-The Output command writes for each test instance the likelihood per label to a file `outputPath = $OutputDir$/SimpleOutput`. 
-To use the Output command either set `command=Simple_Demo_Output` in the config file or add it to the command line.
-The model that is used to compute the labels in these commands is defined 
-in the modelPath variable at the beginning of the file `modelPath=$modelDir$/simple.dnn`.
+The Test (Simple_Demo_Test) and the Output (Simple_Demo_Output) commands 
+specified in the config files use the trained model to compute labels for data 
+specified in the SimpleDataTest.txt file. The Test command computes prediction 
+error, cross entropy and perplexity for the test set and outputs them to the 
+console. The Output command writes for each test instance the likelihood per 
+label to a file `outputPath = $OutputDir$/SimpleOutput`. 
+To use the Output command either set `command=Simple_Demo_Output` in the config 
+file or add it to the command line. The model that is used to compute the labels 
+in these commands is defined in the modelPath variable at the beginning of the 
+file `modelPath=$modelDir$/simple.dnn`.

Examples/README.md

@@ -1,25 +1,20 @@
-# CNTK Demos and Example Setups
+# CNTK Examples
 
-This folder contains a few self-contained demos to get started with CNTK.
-The data for the demos is contained in the corresponding Data folders.
-Each demo folder has a Readme file that explains how to run it on Windows and Linux. 
-How to run the demos on Philly (https://phillywebportal.azurewebsites.net/index.aspx) is 
-explained in the Philly portal wiki (Philly is an internal GPU cluster for Microsoft production runs).
-The demos cover different domains such as Speech and Text classification 
-and show different types of networks including FF, CNN RNN and LSTM.
+This folder contains demos and examples to get started with CNTK.
+The examples are structured by topic into Image, Speech, Text and Other.
+The individual folders contain on the first level at least one self-contained example,
+which cover different types of networks including FF, CNN, RNN and LSTM.
+Further examples for for each category are provided in the corresponding Miscellaneous subfolder.
+Each folder contains a Readme file that explains how to run the example on Windows and Linux. 
+How to run the examples on Philly (https://philly) is explained in the Philly portal wiki 
+(Philly is an internal GPU cluster for Microsoft production runs).
 
-Further examples are provided in the folder 'ExampleSetups'. 
-A popular example is the MNIST handwritten digits classification task. 
-You can find this example in 'ExampleSetups/Images/MNIST'.
-The examples in 'ExampleSetups' might require downloading data which in some cases is not free of charge. 
-See individual folders for more information.
-
-The four examples shown in the table below provide a good introduction to CNTK.
-Additional more complex examples can be found in the 'ExampleSetups' folder.
+The examples shown in the table below provide a good introduction to CNTK.
+Please refer to the Readme file in the corresponding folder for further details.
 
 |Folder                   | Domain                                           | Network types   |
 |:------------------------|:-------------------------------------------------|:----------------|
-Demos/Simple2d            | Synthetic 2d data                                | FF (CPU and GPU)
-Demos/Speech              | Speech data (CMU AN4)                            | FF and LSTM
-Demos/Text                | Text data (penn treebank)                        | RNN
-ExampleSetups/Image/MNIST | Image data (MNIST handwritten digit recognition) | CNN 
+|Other/Simple2d           | Synthetic 2d data                                | FF (CPU and GPU)
+|Speech/AN4               | Speech data (CMU AN4)                            | FF and LSTM
+|Image/MNIST              | Image data (MNIST handwritten digit recognition) | CNN 
+|Text/PennTreebank        | Text data (penn treebank)                        | RNN

Examples/Speech/AN4/README.md

@@ -1,4 +1,4 @@
-# CNTK example: Speech 
+# CNTK example: Speech AN4
 
 ## License
 
@@ -9,19 +9,18 @@ This modified version of dataset is distributed under the terms of a AN4 license
 
 ## Overview
 
-| | |
-|:--------|:---|
-Data:     |Speech data from the CMU Audio Database aka AN4 (http://www.speech.cs.cmu.edu/databases/an4)
-Purpose:  |Showcase how to train feed forward and LSTM networks for speech data
-Network:  |SimpleNetworkBuilder for 2-layer FF, NdlNetworkBuilder for 3-layer LSTM network
-Training: |Data-parallel 1-Bit SGD with automatic mini batch rescaling (FF)
-Comments: |There are two config files: FeedForward.config and LSTM-NDL.config for FF and LSTM training respectively
+|Data:     |Speech data from the CMU Audio Database aka AN4 (http://www.speech.cs.cmu.edu/databases/an4)
+|:---------|:---|
+|Purpose:  |Showcase how to train feed forward and LSTM networks for speech data
+|Network:  |SimpleNetworkBuilder for 2-layer FF, NdlNetworkBuilder for 3-layer LSTM network
+|Training: |Data-parallel 1-Bit SGD with automatic mini batch rescaling (FF)
+|Comments: |There are two config files: FeedForward.config and LSTM-NDL.config for FF and LSTM training respectively
 
 ## Running the example
 
 ### Getting the data
 
-The data for this example is already contained in the folder Demos/Text/Data/.
+The data for this example is already contained in the folder AN4/Data/.
 
 ### Setup
 
@@ -37,16 +36,16 @@ or prefix the call to the cntk executable with the corresponding folder.
 
 ### Run
 
-Run the example from the Demos/Speech/Data folder using:
+Run the example from the Speech/Data folder using:
 
 `cntk configFile=../Config/FeedForward.config`
 
 or run from any folder and specify the Data folder as the `currentDirectory`, 
-e.g. running from the Demos/Speech folder using:
+e.g. running from the Speech folder using:
 
 `cntk configFile=Config/FeedForward.config currentDirectory=Data`
 
-The output folder will be created inside Demos/Speech/.
+The output folder will be created inside Speech/.
 
 ## Details
 
@@ -60,7 +59,7 @@ To run on CPU set `deviceId = -1`, to run on GPU set deviceId to "auto" or a spe
 
 The FeedForward.config file uses the SimpleNetworkBuilder to create a 2-layer 
 feed forward network with sigmoid nodes and a softmax layer.
-The LSTM-NDL.config file uses the NdlNetworkBuilder and refers to the lstmp-3layer_WithSelfStab.ndl file. 
+The LSTM-NDL.config file uses the NdlNetworkBuilder and refers to the lstmp-3layer-opt.ndl file. 
 In the ndl file an LSTM component is defined and used to create a 3-layer LSTM network with a softmax layer. 
 Both configuration only define and execute a single training task:
 

Examples/Text/PennTreebank/README.md

@@ -6,19 +6,19 @@ Note: The data is not checked into the repository currently since a license is r
 
 ## Overview
 
-| | |
-|:--------|:---|
-Data:     |The Penn Treebank Project (https://www.cis.upenn.edu/~treebank/) annotates naturally-occuring text for linguistic structure .
-Purpose:  |Showcase how to train a recurrent network for text data.
-Network:  |SimpleNetworkBuilder for recurrent network with two hidden layers.
-Training: |Stochastic gradient descent with adjusted learning rate.
-Comments: |The provided configuration file performs class based RNN training.
+|Data:     |The Penn Treebank Project (https://www.cis.upenn.edu/~treebank/) annotates naturally-occuring text for linguistic structure .
+|:---------|:---|
+|Purpose:  |Showcase how to train a recurrent network for text data.
+|Network:  |SimpleNetworkBuilder for recurrent network with two hidden layers.
+|Training: |Stochastic gradient descent with adjusted learning rate.
+|Comments: |The provided configuration file performs class based RNN training.
 
 ## Running the example
 
 ### Getting the data
 
-The data for this example is already contained in the folder Demos/Text/Data/.
+The data is not checked into the repository currently since a license is required for the penn treebank data.
+Please visit https://www.cis.upenn.edu/~treebank/
 
 ### Setup
 
@@ -34,16 +34,16 @@ or prefix the call to the cntk executable with the corresponding folder.
 
 ### Run
 
-Run the example from the Demos/Text/Data folder using:
+Run the example from the Text/Data folder using:
 
 `cntk configFile=../Config/rnn.config`
 
 or run from any folder and specify the Data folder as the `currentDirectory`, 
-e.g. running from the Demos/Text folder using:
+e.g. running from the Text folder using:
 
 `cntk configFile=Config/rnn.config currentDirectory=Data`
 
-The output folder will be created inside Demos/Text/.
+The output folder will be created inside Text/.
 
 ## Details
 

Tools/build-and-test

@@ -142,7 +142,7 @@ fi
 cd $CNTK_ROOT
 
 if ! [[ -f $CONF_FILE ]]; then
-    cp Examples/Miscellaneous/Simple2d/Config/Simple.config $CONF_FILE || exit $?
+    cp Examples/Other/Simple2d/Config/Simple.config $CONF_FILE || exit $?
 
     # This chmod is necessary due to restrictive Cygwin interpretation of Windows permissions.
     # Cygwin interprets Windows permissions as ----rwx---, which lacks read permissions for user.
@@ -213,8 +213,8 @@ fi
 if [[ $RUN == 1 ]]; then
 
     cd $PREFIX_DIR
-    echo "============ cp Examples/Miscellaneous/Simple2d/Config/Simple.config $CONF_FILE  ============"
-    echo "============ cd $CNTK_ROOT/Examples/Miscellaneous/Simple2d/Data  ============"
+    echo "============ cp Examples/Other/Simple2d/Config/Simple.config $CONF_FILE  ============"
+    echo "============ cd $CNTK_ROOT/Examples/Other/Simple2d/Data  ============"
 
     for TARGET in "${targetArray[@]}"
     do
@@ -250,7 +250,7 @@ if [[ $RUN == 1 ]]; then
                 echo "============ Running $BIN_PATH configFile=$CONF_FILE for ($FLAVOR) ($TARGET) ============"
             fi
             echo "============ output in ($OUT_FILE) ============"
-            cd $CNTK_ROOT/Examples/Miscellaneous/Simple2d/Data
+            cd $CNTK_ROOT/Examples/Other/Simple2d/Data
             rm -rf models
             if [[ $OS == "Windows_NT" ]]; then
                 # We have to use cygpath on Windows to modify the file paths into the format readable by cntk.