Add README and comments for RNN examples (recurrent folder)

Author: NA-V10

dl4j-examples/src/main/java/org/deeplearning4j/examples/quickstart/modeling/recurrent/MemorizeSequence.java

@@ -17,6 +17,10 @@
  * SPDX-License-Identifier: Apache-2.0
  ******************************************************************************/
 
+  // MemorizeSequence.java
+ // A simple RNN example where the network learns to memorize and reproduce a short sequence.
+// Demonstrates basic RNN training and backpropagation-through-time.
+
 package org.deeplearning4j.examples.quickstart.modeling.recurrent;
 
 import org.deeplearning4j.nn.conf.MultiLayerConfiguration;
@@ -87,6 +91,8 @@ public static void main(String[] args) {
 			hiddenLayerBuilder.activation(Activation.TANH);
 			listBuilder.layer(i, hiddenLayerBuilder.build());
 		}
+   
+		// Build a simple RNN with one recurrent layer to memorize the sequence
 
 		// we need to use RnnOutputLayer for our RNN
 		RnnOutputLayer.Builder outputLayerBuilder = new RnnOutputLayer.Builder(LossFunction.MCXENT);
@@ -96,7 +102,8 @@ public static void main(String[] args) {
 		outputLayerBuilder.nIn(HIDDEN_LAYER_WIDTH);
 		outputLayerBuilder.nOut(LEARNSTRING_CHARS.size());
 		listBuilder.layer(HIDDEN_LAYER_CONT, outputLayerBuilder.build());
-
+        
+		
 		// create network
 		MultiLayerConfiguration conf = listBuilder.build();
 		MultiLayerNetwork net = new MultiLayerNetwork(conf);
@@ -122,6 +129,9 @@ public static void main(String[] args) {
 			labels.putScalar(new int[] { 0, LEARNSTRING_CHARS_LIST.indexOf(nextChar), samplePos }, 1);
 			samplePos++;
 		}
+
+		// Train the RNN to output the same sequence it receives as input
+
 		DataSet trainingData = new DataSet(input, labels);
 
 		// some epochs

dl4j-examples/src/main/java/org/deeplearning4j/examples/quickstart/modeling/recurrent/README.md

@@ -0,0 +1,67 @@
+# Recurrent Neural Network (RNN) Examples – DeepLearning4J
+
+This folder contains simple recurrent neural network (RNN) examples using LSTM and embedding layers.  
+These examples demonstrate how to work with sequential data such as signals, sequences, and text-like inputs.
+
+---
+
+## 🔁 MemorizeSequence.java
+A minimal RNN example where the network learns to memorize and reproduce a fixed sequence.
+
+### What this example teaches
+- How RNNs store information across time steps  
+- How backpropagation-through-time works  
+- How sequence learning differs from feedforward networks  
+
+### Expected Behavior
+The network eventually outputs the same sequence it was trained on.
+
+---
+
+## 🔡 RNNEmbedding.java
+Demonstrates the use of an **EmbeddingLayer** followed by RNN layers.
+
+### Key Concepts
+- Turning integer-encoded inputs into dense vectors  
+- Word/token embedding  
+- Passing embedded sequences into RNN layers  
+
+This is a useful template for NLP-style models.
+
+---
+
+## 📊 UCISequenceClassification.java
+Sequence classification on a dataset from the UCI machine learning repository.
+
+### What this example shows
+- Loading sequential datasets  
+- Recurrent classification (predict a label for a whole sequence)  
+- Time-series preprocessing and normalization  
+
+---
+
+## ✔ How to Run Any Example
+
+Use the following command:
+
+mvn -q exec:java -Dexec.mainClass="org.deeplearning4j.examples.quickstart.modeling.recurrent.<ClassName>"
+
+
+Example:
+
+
+
+mvn -q exec:java -Dexec.mainClass="org.deeplearning4j.examples.quickstart.modeling.recurrent.MemorizeSequence"
+
+
+---
+
+## 🙌 Why This README Helps
+
+This folder previously had no documentation.  
+This README explains:
+- What each RNN example does  
+- What concepts it teaches  
+- How to run each file  
+
+This improves clarity for beginners working with sequential models in DL4J.

dl4j-examples/src/main/java/org/deeplearning4j/examples/quickstart/modeling/recurrent/RNNEmbedding.java

@@ -17,6 +17,11 @@
  * SPDX-License-Identifier: Apache-2.0
  ******************************************************************************/
 
+
+  // RNNEmbedding.java
+ // Demonstrates how to use an EmbeddingLayer + RNN layers for sequence modeling.
+// Useful for NLP-style integer token inputs.
+
 package org.deeplearning4j.examples.quickstart.modeling.recurrent;
 
 import org.deeplearning4j.nn.conf.MultiLayerConfiguration;
@@ -45,6 +50,9 @@
  *
  * @author Alex Black
  */
+
+ // Convert integer token IDs into dense embedding vectors
+
 public class RNNEmbedding {
     public static void main(String[] args) {
 
@@ -64,6 +72,10 @@ public static void main(String[] args) {
             }
         }
 
+        
+      // Feed embedded vectors into an LSTM/RNN to capture sequence structure
+
+
         MultiLayerConfiguration conf = new NeuralNetConfiguration.Builder()
             .activation(Activation.RELU)
             .list()

dl4j-examples/src/main/java/org/deeplearning4j/examples/quickstart/modeling/recurrent/UCISequenceClassification.java

@@ -17,6 +17,10 @@
  * SPDX-License-Identifier: Apache-2.0
  ******************************************************************************/
 
+
+ // UCISequenceClassification.java
+// Demonstrates sequence classification using an RNN on UCI dataset sequences.
+
 package org.deeplearning4j.examples.quickstart.modeling.recurrent;
 
 import org.apache.commons.io.FileUtils;