Update Links to documentation

treo · treo · commit ccedcd48734d · 2020-06-16T11:01:51.000+02:00
Signed-off-by: Paul Dubs &lt;paul.dubs@gmail.com&gt;
diff --git a/cuda-specific-examples/README.md b/cuda-specific-examples/README.md
@@ -1,8 +1,8 @@
 ## Eclipse Deeplearning4j: CUDA Specific Examples
 
-Switching from a CPU only backend to a GPU backend is as simple as changing one dependency - one line in the pom.xml file for Maven users. Instead of specifying the nd4j-native-platform module specify the nd4j-cuda-X-platform where X indicated the version of CUDA. It is recommended to install cuDNN for better GPU performance. Runs will log warnings if cuDNN is not found. For more information, please refer to documentation [here](https://deeplearning4j.org/docs/latest/deeplearning4j-config-cudnn)
+Switching from a CPU only backend to a GPU backend is as simple as changing one dependency - one line in the pom.xml file for Maven users. Instead of specifying the nd4j-native-platform module specify the nd4j-cuda-X-platform where X indicated the version of CUDA. It is recommended to install cuDNN for better GPU performance. Runs will log warnings if cuDNN is not found. For more information, please refer to documentation [here](https://deeplearning4j.konduit.ai/config/backends/config-cudnn#using-deeplearning-4-j-with-cudnn)
 
-Users with acces to multiple gpus systems can use DL4J to further speed up the training process by training the models in parallel on them. Ideally these GPUs have the same speed and networking capabilities. This project contains a set of examples that demonstrate how to leverage performance from a multiple gpus setup. More documentation can be found [here](https://deeplearning4j.konduit.ai/getting-started/tutorials/using-multiple-gpus)
+Users with access to systems with multiple gpus can use DL4J to further speed up the training process by training the models in parallel on them. Ideally these GPUs have the same speed and networking capabilities. This project contains a set of examples that demonstrate how to leverage performance from a multiple gpus setup. More documentation can be found [here](https://deeplearning4j.konduit.ai/getting-started/tutorials/using-multiple-gpus)
 
 [Go back](../README.md) to the main repository page to explore other features/functionality of the **Eclipse Deeplearning4J** ecosystem. File an issue [here](https://github.com/eclipse/deeplearning4j-examples/issues) to request new features.
 
diff --git a/cuda-specific-examples/src/main/java/org/deeplearning4j/examples/multigpu/advanced/charmodelling/CharacterIterator.java b/cuda-specific-examples/src/main/java/org/deeplearning4j/examples/multigpu/advanced/charmodelling/CharacterIterator.java
@@ -163,7 +163,6 @@ public DataSet next(int num) {
         // dimension 0 = number of examples in minibatch
         // dimension 1 = size of each vector (i.e., number of characters)
         // dimension 2 = length of each time series/example
-        //Why 'f' order here? See http://deeplearning4j.org/usingrnns.html#data section "Alternative: Implementing a custom DataSetIterator"
 		INDArray input = Nd4j.create(new int[]{currMinibatchSize,validCharacters.length,exampleLength}, 'f');
 		INDArray labels = Nd4j.create(new int[]{currMinibatchSize,validCharacters.length,exampleLength}, 'f');
 
diff --git a/cuda-specific-examples/src/main/java/org/deeplearning4j/examples/multigpu/advanced/charmodelling/GenerateTxtModel.java b/cuda-specific-examples/src/main/java/org/deeplearning4j/examples/multigpu/advanced/charmodelling/GenerateTxtModel.java
@@ -58,9 +58,7 @@
 	from Project Gutenberg. Training on other text sources should be relatively easy to implement.
 
     For more details on RNNs in DL4J, see the following:
-    http://deeplearning4j.org/usingrnns
-    http://deeplearning4j.org/lstm
-    http://deeplearning4j.org/recurrentnetwork
+    https://deeplearning4j.konduit.ai/models/recurrent
  */
 public class GenerateTxtModel {
 	public static void main( String[] args ) throws Exception {
diff --git a/cuda-specific-examples/src/main/java/org/deeplearning4j/examples/multigpu/advanced/transferlearning/vgg16/README.md b/cuda-specific-examples/src/main/java/org/deeplearning4j/examples/multigpu/advanced/transferlearning/vgg16/README.md
@@ -1,5 +1,5 @@
 ##### TransferLearning
-Demonstrates use of the dl4j transfer learning API which allows users to construct a model based off an existing model by modifying the architecture, freezing certain parts selectively and then fine tuning parameters. Read the documentation for the Transfer Learning API at [https://deeplearning4j.org/transfer-learning](https://deeplearning4j.org/transfer-learning).  
+Demonstrates use of the dl4j transfer learning API which allows users to construct a model based off an existing model by modifying the architecture, freezing certain parts selectively and then fine tuning parameters. Read the documentation for the Transfer Learning API at [https://deeplearning4j.konduit.ai/tuning-and-training/transfer-learning](https://deeplearning4j.konduit.ai/tuning-and-training/transfer-learning).
 
 More more examples refer to the section in the dl4j-example repo [here](../../../../../../../../../../../dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/features/transferlearning/README.md)
 
diff --git a/dl4j-distributed-training-examples/scripts/patentExampleTrain.sh b/dl4j-distributed-training-examples/scripts/patentExampleTrain.sh
@@ -31,7 +31,7 @@ NUM_NODES=...                                   #Number of nodes in the cluster
 SPARKSUBMIT=/opt/spark/bin/spark-submit
 MASTER_PORT=7077                                #Port for the spark master. Default is 7077
 MINIBATCH=32                                    #Minibatch size for preprocessed datasets
-    # For memory config, see https://deeplearning4j.org/memory
+    # For memory config, see https://deeplearning4j.konduit.ai/config/config-memory
 JAVA_HEAP_MEM=10G
 OFFHEAP_MEM_JAVACPP=20G
 OFFHEAP_JAVACPP_MAX_PHYS=30G
diff --git a/dl4j-distributed-training-examples/src/main/java/org/deeplearning4j/distributedtrainingexamples/patent/README.md b/dl4j-distributed-training-examples/src/main/java/org/deeplearning4j/distributedtrainingexamples/patent/README.md
@@ -15,7 +15,7 @@ Number of classes: 398
 Number of documents/examples (after preprocessing): approx. 5.7 million (training set) plus approx. 170000 (test set)
 
 Dataset size: approx. 86 GB (zip format), 464 GB raw text. Note the example performs preprocessing from the compressed ZIP format.
-Requires an additional 20GB of storage space for preprocessing  
+Requires an additional 20GB of storage space for preprocessing
 
 **Neural Network**: a CNN classifier for text classification. Approximately 600,000 parameters
 
@@ -70,7 +70,7 @@ MASTER_IP=...
 AZURE_STORAGE_ACCT=...
 AZURE_STORAGE_ACCT_KEY=...
 AZURE_CONTAINER_ZIPS=patentzips
-AZURE_CONTAINER_PREPROC=patentExamplePreproc 
+AZURE_CONTAINER_PREPROC=patentExamplePreproc
 ```
 
 Note that some clusters may have the master already configured.
@@ -93,7 +93,7 @@ is pointed to the same value for ```AZURE_CONTAINER_PREPROC```.
 **Alternatively to setting storage account**
 
 You can set the storage account credentials in your Hadoop core-site.xml file. See "Configuring Credentials" in this guide for details: [https://hadoop.apache.org/docs/current/hadoop-azure/index.html](https://hadoop.apache.org/docs/current/hadoop-azure/index.html)
- 
+
 
 **Second: Run the Script**
 
@@ -108,7 +108,7 @@ After preprocessing is complete, you will have:
     2. For HTTP access (if enabled): ```https://AZURE_STORAGE_ACCT.blob.core.windows.net/AZURE_CONTAINER_ZIPS/```
 2. Preprocessed training and test data (with default sequence length of 1000 and minibatch size of 32)
    1. For Spark access: ```wasbs://AZURE_CONTAINER_PREPROC@AZURE_STORAGE_ACCT.blob.core.windows.net/seqLength1000_mb32/```
-   2. For HTTP access (if enabled): ```https://AZURE_STORAGE_ACCT.blob.core.windows.net/AZURE_CONTAINER_PREPROC/```  
+   2. For HTTP access (if enabled): ```https://AZURE_STORAGE_ACCT.blob.core.windows.net/AZURE_CONTAINER_PREPROC/```
 
 Note that the preprocessed directory will have ```train``` and ```test``` subdirectories.
 The format of the files in those train/test directories is a custom format designed to be loaded
@@ -131,7 +131,7 @@ Set the following required arguments to the same values used for the preprocessi
 MASTER_IP=...
 AZURE_STORAGE_ACCT=...
 AZURE_STORAGE_ACCT_KEY=...
-AZURE_CONTAINER_PREPROC=patentExamplePreproc 
+AZURE_CONTAINER_PREPROC=patentExamplePreproc
 ```
 
 The following configuration options also need to be set:
@@ -142,7 +142,7 @@ LOCAL_SAVE_DIR
 
 Your network mask should be set to the network used for spark communication. For example, [10.0.0.0/16]
 See the following links for further details:
-* [DL4J Distributed Training - Netmask](https://deeplearning4j.org/distributed#netmask)
+* [DL4J Distributed Training - Netmask](https://deeplearning4j.konduit.ai/distributed-deep-learning/parameter-server#netmask)
 * [How to Find the IP Address, Subnet Mask & Gateway of a Computer](https://yourbusiness.azcentral.com/ip-address-subnet-mask-gateway-computer-14563.html)
 * [What is a Subnet Mask](https://www.iplocation.net/subnet-mask)
 
diff --git a/dl4j-distributed-training-examples/src/main/java/org/deeplearning4j/distributedtrainingexamples/tinyimagenet/TrainSpark.java b/dl4j-distributed-training-examples/src/main/java/org/deeplearning4j/distributedtrainingexamples/tinyimagenet/TrainSpark.java
@@ -78,7 +78,7 @@
  * a larger network, better selection of hyperparameters, and more epochs.
  *
  * For further details on DL4J's Spark implementation, see the "Distributed Deep Learning" pages at:
- * https://deeplearning4j.org/docs/latest/
+ * https://deeplearning4j.konduit.ai/distributed-deep-learning/intro
  *
  * A local (single machine) version of this example is available in TrainLocal
  *
@@ -145,7 +145,7 @@ protected void entryPoint(String[] args) throws Exception {
         //Set up TrainingMaster for gradient sharing training
         VoidConfiguration voidConfiguration = VoidConfiguration.builder()
             .unicastPort(port)                          // Should be open for IN/OUT communications on all Spark nodes
-            .networkMask(networkMask)                   // Local network mask - for example, 10.0.0.0/16 - see https://deeplearning4j.org/docs/latest/deeplearning4j-scaleout-parameter-server
+            .networkMask(networkMask)                   // Local network mask - for example, 10.0.0.0/16 - see https://deeplearning4j.konduit.ai/distributed-deep-learning/parameter-server#netmask
             .controllerAddress(masterIP)                // IP address of the master/driver node
             .meshBuildMode(MeshBuildMode.PLAIN)
             .build();
diff --git a/dl4j-examples/README.md b/dl4j-examples/README.md
@@ -166,7 +166,7 @@ Trace where data from each example comes from and get metadata on prediction err
 Train a MultiLayerNetwork where the errors come from an external source, instead of using an Output layer and a labels array.
 
 ##### TransferLearning
-Demonstrates use of the dl4j transfer learning API which allows users to construct a model based off an existing model by modifying the architecture, freezing certain parts selectively and then fine tuning parameters. Read the documentation for the Transfer Learning API at [https://deeplearning4j.org/transfer-learning](https://deeplearning4j.org/transfer-learning).
+Demonstrates use of the dl4j transfer learning API which allows users to construct a model based off an existing model by modifying the architecture, freezing certain parts selectively and then fine tuning parameters. Read the documentation for the Transfer Learning API at [https://deeplearning4j.konduit.ai/tuning-and-training/transfer-learning](https://deeplearning4j.konduit.ai/tuning-and-training/transfer-learning).
 * [EditLastLayerOthersFrozen.java](./src/main/java/org/deeplearning4j/examples/advanced/features/transferlearning/editlastlayer/EditLastLayerOthersFrozen.java)
 Modifies just the last layer in vgg16, freezes the rest and trains the network on the flower dataset.
 * [FeaturizedPreSave.java](./src/main/java/org/deeplearning4j/examples/advanced/features/transferlearning/editlastlayer/presave/FeaturizedPreSave.java) & [FitFromFeaturized.java](./src/main/java/org/deeplearning4j/examples/advanced/features/transferlearning/editlastlayer/presave/FitFromFeaturized.java)
diff --git a/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/features/transferlearning/README.md b/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/features/transferlearning/README.md
@@ -1,10 +1,10 @@
 ##### TransferLearning
-Demonstrates use of the dl4j transfer learning API which allows users to construct a model based off an existing model by modifying the architecture, freezing certain parts selectively and then fine tuning parameters. Read the documentation for the Transfer Learning API at [https://deeplearning4j.org/transfer-learning](https://deeplearning4j.org/transfer-learning).  
-* [EditLastLayerOthersFrozen.java](./editlastlayer/EditLastLayerOthersFrozen.java)  
+Demonstrates use of the dl4j transfer learning API which allows users to construct a model based off an existing model by modifying the architecture, freezing certain parts selectively and then fine tuning parameters. Read the documentation for the Transfer Learning API at [https://deeplearning4j.konduit.ai/tuning-and-training/transfer-learning](https://deeplearning4j.konduit.ai/tuning-and-training/transfer-learning).
+* [EditLastLayerOthersFrozen.java](./editlastlayer/EditLastLayerOthersFrozen.java)
 Modifies just the last layer in vgg16, freezes the rest and trains the network on the flower dataset.
 * [FeaturizedPreSave.java](./editlastlayer/presave/FeaturizedPreSave.java) & [FitFromFeaturized.java](./editlastlayer/presave/FitFromFeaturized.java)
 Save time on the forward pass during multiple epochs by "featurizing" the datasets. FeaturizedPreSave saves the output at the last frozen layer and FitFromFeaturize fits to the presaved data so you can iterate quicker with different learning parameters.
-* [EditAtBottleneckOthersFrozen.java](./editfrombottleneck/EditAtBottleneckOthersFrozen.java)  
+* [EditAtBottleneckOthersFrozen.java](./editfrombottleneck/EditAtBottleneckOthersFrozen.java)
 A more complex example of modifying model architecure by adding/removing vertices
-* [FineTuneFromBlockFour.java](./finetuneonly/FineTuneFromBlockFour.java)  
+* [FineTuneFromBlockFour.java](./finetuneonly/FineTuneFromBlockFour.java)
 Reads in a saved model (training information and all) and fine tunes it by overriding its training information with what is specified
diff --git a/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/charmodelling/generatetext/GenerateTxtCharCompGraphModel.java b/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/charmodelling/generatetext/GenerateTxtCharCompGraphModel.java
@@ -38,7 +38,7 @@
 /**
  * This example is almost identical to the LSTMCharModellingExample, except that it utilizes the ComputationGraph
  * architecture instead of MultiLayerNetwork architecture. See the javadoc in that example for details.
- * For more details on the ComputationGraph architecture, see http://deeplearning4j.org/compgraph
+ * For more details on the ComputationGraph architecture, see https://deeplearning4j.konduit.ai/models/computationgraph
  *
  * In addition to the use of the ComputationGraph a, this version has skip connections between the first and output layers,
  * in order to show how this configuration is done. In practice, this means we have the following types of connections:
diff --git a/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/charmodelling/generatetext/GenerateTxtModel.java b/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/charmodelling/generatetext/GenerateTxtModel.java
@@ -54,9 +54,7 @@
 	from Project Gutenberg. Training on other text sources should be relatively easy to implement.
 
     For more details on RNNs in DL4J, see the following:
-    http://deeplearning4j.org/usingrnns
-    http://deeplearning4j.org/lstm
-    http://deeplearning4j.org/recurrentnetwork
+    https://deeplearning4j.konduit.ai/models/recurrent
  */
 @SuppressWarnings("DuplicatedCode")
 public class GenerateTxtModel {
diff --git a/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/charmodelling/utils/CharacterIterator.java b/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/charmodelling/utils/CharacterIterator.java
@@ -187,7 +187,6 @@ public DataSet next(int num) {
         // dimension 0 = number of examples in minibatch
         // dimension 1 = size of each vector (i.e., number of characters)
         // dimension 2 = length of each time series/example
-        //Why 'f' order here? See http://deeplearning4j.org/usingrnns.html#data section "Alternative: Implementing a custom DataSetIterator"
 		INDArray input = Nd4j.create(new int[]{currMinibatchSize,validCharacters.length,exampleLength}, 'f');
 		INDArray labels = Nd4j.create(new int[]{currMinibatchSize,validCharacters.length,exampleLength}, 'f');
 
diff --git a/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/embeddingsfromcorpus/word2vec/Word2VecRawTextExample.java b/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/embeddingsfromcorpus/word2vec/Word2VecRawTextExample.java
@@ -34,7 +34,7 @@
  * Created by agibsonccc on 10/9/14.
  *
  * Neural net that processes text into wordvectors. See below url for an in-depth explanation.
- * https://deeplearning4j.org/word2vec.html
+ * https://deeplearning4j.konduit.ai/language-processing/word2vec
  */
 public class Word2VecRawTextExample {
 
diff --git a/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/objectdetection/TinyYoloHouseNumberDetection.java b/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/objectdetection/TinyYoloHouseNumberDetection.java
@@ -70,7 +70,7 @@
  * - YOLO: Real-Time Object Detection: https://pjreddie.com/darknet/yolo/ <br>
  * - The Street View House Numbers (SVHN) Dataset: http://ufldl.stanford.edu/housenumbers/ <br>
  * <p>
- * Please note, cuDNN should be used to obtain reasonable performance: https://deeplearning4j.org/cudnn
+ * Please note, cuDNN should be used to obtain reasonable performance: https://deeplearning4j.konduit.ai/config/backends/config-cudnn#using-deeplearning-4-j-with-cudnn
  *
  * @author saudet
  */
diff --git a/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/seq2seq/AdditionModelWithSeq2Seq.java b/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/seq2seq/AdditionModelWithSeq2Seq.java
@@ -49,9 +49,9 @@ public class AdditionModelWithSeq2Seq {
 
         This example is build using a computation graph with RNN layers.
         Refer here for more details on computation graphs in dl4j
-            https://deeplearning4j.org/compgraph
+            https://deeplearning4j.konduit.ai/models/computationgraph
         And here for RNNs
-            https://deeplearning4j.org/usingrnns
+            https://deeplearning4j.konduit.ai/models/recurrent
 
         There are two RNN layers to this computation graph. The inputs to them are as follows,
         During training:
diff --git a/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/textclassification/pretrainedword2vec/ImdbReviewClassificationRNN.java b/dl4j-examples/src/main/java/org/deeplearning4j/examples/advanced/modelling/textclassification/pretrainedword2vec/ImdbReviewClassificationRNN.java
@@ -94,7 +94,7 @@ public static void main(String[] args) throws Exception {
         int truncateReviewsToLength = 256;  //Truncate reviews with length (# words) greater than this
         final int seed = 0;     //Seed for reproducibility
 
-        Nd4j.getMemoryManager().setAutoGcWindow(10000);  //https://deeplearning4j.org/workspaces
+        Nd4j.getMemoryManager().setAutoGcWindow(10000);  //https://deeplearning4j.konduit.ai/config/config-memory/config-workspaces#garbage-collector
 
         //Set up network configuration
         MultiLayerConfiguration conf = new NeuralNetConfiguration.Builder()
diff --git a/dl4j-examples/src/main/java/org/deeplearning4j/examples/quickstart/features/earlystopping/EarlyStoppingMNIST.java b/dl4j-examples/src/main/java/org/deeplearning4j/examples/quickstart/features/earlystopping/EarlyStoppingMNIST.java
@@ -53,7 +53,7 @@
  * have the minimum test set loss
  * This is an over-simplified example, but the principles used here should apply in more realistic cases.
  *
- * For further details on early stopping, see http://deeplearning4j.org/earlystopping.html
+ * For further details on early stopping, see https://deeplearning4j.konduit.ai/getting-started/tutorials/early-stopping
  *
  * @author Alex Black
  */
diff --git a/dl4j-examples/src/main/java/org/deeplearning4j/examples/quickstart/modeling/feedforward/classification/ModelXOR.java b/dl4j-examples/src/main/java/org/deeplearning4j/examples/quickstart/modeling/feedforward/classification/ModelXOR.java
@@ -112,11 +112,9 @@ public static void main(String[] args) {
             .updater(new Sgd(0.1))
             .seed(seed)
             .biasInit(0) // init the bias with 0 - empirical value, too
-            // from "http://deeplearning4j.org/architecture": The networks can
-            // process the input more quickly and more accurately by ingesting
+            // The networks can process the input more quickly and more accurately by ingesting
             // minibatches 5-10 elements at a time in parallel.
-            // this example runs better without, because the dataset is smaller than
-            // the mini batch size
+            // This example runs better without, because the dataset is smaller than the mini batch size
             .miniBatch(false)
             .list()
             .layer(new DenseLayer.Builder()
diff --git a/dl4j-examples/src/main/java/org/deeplearning4j/examples/quickstart/modeling/recurrent/UCISequenceClassification.java b/dl4j-examples/src/main/java/org/deeplearning4j/examples/quickstart/modeling/recurrent/UCISequenceClassification.java
@@ -75,7 +75,7 @@
  *
  * 2. Load the training data using CSVSequenceRecordReader (to load/parse the CSV files) and SequenceRecordReaderDataSetIterator
  *    (to convert it to DataSet objects, ready to train)
- *    For more details on this step, see: http://deeplearning4j.org/usingrnns#data
+ *    For more details on this step, see: https://deeplearning4j.konduit.ai/models/recurrent#data-for-rnns
  *
  * 3. Normalize the data. The raw data contain values that are too large for effective training, and need to be normalized.
  *    Normalization is conducted using NormalizerStandardize, based on statistics (mean, st.dev) collected on the training
diff --git a/mvn-project-template/pom.xml b/mvn-project-template/pom.xml
@@ -64,7 +64,7 @@
         -->
 
         <!-- Optional, but recommended: if you use CUDA, also use CuDNN. To use this, CuDNN must also be installed -->
-        <!-- See: https://deeplearning4j.org/cudnn -->
+        <!-- See: https://deeplearning4j.konduit.ai/config/backends/config-cudnn#using-deeplearning-4-j-with-cudnn -->
         <!--
         <dependency>
             <groupId>org.deeplearning4j</groupId>
diff --git a/tensorflow-keras-import-examples/src/main/java/org/deeplearning4j/modelimportexamples/keras/advanced/deepmoji/ImportDeepMoji.java b/tensorflow-keras-import-examples/src/main/java/org/deeplearning4j/modelimportexamples/keras/advanced/deepmoji/ImportDeepMoji.java
@@ -49,6 +49,7 @@ public static void main(String[] args) throws Exception {
         File directory = new File(DATA_PATH);
         if(!directory.exists()) directory.mkdir();
 
+        // TODO: fix link
         String modelUrl = "http://blob.deeplearning4j.org/models/deepmoji.h5";
         String downloadPath = DATA_PATH + "deepmoji_model.h5";
         File cachedKerasFile = new File(downloadPath);
diff --git a/tensorflow-keras-import-examples/src/main/java/org/deeplearning4j/modelimportexamples/tf/advanced/mobilenet/MobileNetTransferLearningExample.md b/tensorflow-keras-import-examples/src/main/java/org/deeplearning4j/modelimportexamples/tf/advanced/mobilenet/MobileNetTransferLearningExample.md

Original file line number	Diff line number	Diff line change
`@@ -78,7 +78,7 @@`
`78`	`78`	`* a larger network, better selection of hyperparameters, and more epochs.`
`79`	`79`	`*`
`80`	`80`	`* For further details on DL4J's Spark implementation, see the "Distributed Deep Learning" pages at:`
`81`		`- * https://deeplearning4j.org/docs/latest/`
	`81`	`+ * https://deeplearning4j.konduit.ai/distributed-deep-learning/intro`
`82`	`82`	`*`
`83`	`83`	`* A local (single machine) version of this example is available in TrainLocal`
`84`	`84`	`*`
`@@ -145,7 +145,7 @@ protected void entryPoint(String[] args) throws Exception {`
`145`	`145`	`//Set up TrainingMaster for gradient sharing training`
`146`	`146`	`VoidConfiguration voidConfiguration = VoidConfiguration.builder()`
`147`	`147`	`.unicastPort(port) // Should be open for IN/OUT communications on all Spark nodes`
`148`		`- .networkMask(networkMask) // Local network mask - for example, 10.0.0.0/16 - see https://deeplearning4j.org/docs/latest/deeplearning4j-scaleout-parameter-server`
	`148`	`+ .networkMask(networkMask) // Local network mask - for example, 10.0.0.0/16 - see https://deeplearning4j.konduit.ai/distributed-deep-learning/parameter-server#netmask`
`149`	`149`	`.controllerAddress(masterIP) // IP address of the master/driver node`
`150`	`150`	`.meshBuildMode(MeshBuildMode.PLAIN)`
`151`	`151`	`.build();`
Original file line number	Diff line number	Diff line change
`@@ -38,7 +38,7 @@`
`38`	`38`	`/**`
`39`	`39`	`* This example is almost identical to the LSTMCharModellingExample, except that it utilizes the ComputationGraph`
`40`	`40`	`* architecture instead of MultiLayerNetwork architecture. See the javadoc in that example for details.`
`41`		`- * For more details on the ComputationGraph architecture, see http://deeplearning4j.org/compgraph`
	`41`	`+ * For more details on the ComputationGraph architecture, see https://deeplearning4j.konduit.ai/models/computationgraph`
`42`	`42`	`*`
`43`	`43`	`* In addition to the use of the ComputationGraph a, this version has skip connections between the first and output layers,`
`44`	`44`	`* in order to show how this configuration is done. In practice, this means we have the following types of connections:`