DONE

.gitignore

@@ -0,0 +1,38 @@
+target/
+!.mvn/wrapper/maven-wrapper.jar
+!**/src/main/**/target/
+!**/src/test/**/target/
+
+### IntelliJ IDEA ###
+.idea/modules.xml
+.idea/jarRepositories.xml
+.idea/compiler.xml
+.idea/libraries/
+*.iws
+*.iml
+*.ipr
+
+### Eclipse ###
+.apt_generated
+.classpath
+.factorypath
+.project
+.settings
+.springBeans
+.sts4-cache
+
+### NetBeans ###
+/nbproject/private/
+/nbbuild/
+/dist/
+/nbdist/
+/.nb-gradle/
+build/
+!**/src/main/**/build/
+!**/src/test/**/build/
+
+### VS Code ###
+.vscode/
+
+### Mac OS ###
+.DS_Store

Layer1_Biases.txt

@@ -0,0 +1 @@
+-0.30779064 -0.11988339 0.21006837 0.48366025 0.027321858 -0.3208103 0.20687543 0.4597395 0.24228075 0.46249318 -0.4809847 -0.35293418 -0.22552863 -0.33045918 0.21397136 -0.35107967 0.23396465 0.47500616 0.18155238 0.4322363 

Layer2_Biases.txt

@@ -0,0 +1 @@
+0.14082673 -0.40868983 -0.024709726 -0.3233395 0.35455197 0.097625956 -0.16446406 -0.3674031 -0.0673621 -0.2148308 

Layer2_Weights.txt

@@ -0,0 +1,10 @@
+0.36638698 -0.111338705 -0.28196928 -0.27613354 0.30107862 -0.028524958 -0.64342284 0.35270464 -0.44135 0.071591415 0.12277257 0.348394 -0.008241653 -0.046460714 -0.30686167 -0.61905825 -0.12674114 0.38351807 -0.27559012 -0.2511327 
+0.4936019 -0.5685425 0.116411686 0.20801988 -0.1977276 -0.0012066597 0.34004936 -0.5168037 0.49804622 0.5259802 -0.18001957 -0.271705 0.07692847 -0.5320646 -0.51504177 0.58450395 -0.48857433 -0.60006136 0.6742939 -0.25381067 
+0.40631506 0.1401135 0.6673145 -0.5735027 0.22399487 0.28001013 0.40400422 0.18887623 -0.058334175 -0.01644563 -0.016555214 -0.35288614 -0.16616665 0.3261699 -0.1806511 -0.08278414 -0.03714714 -0.31807142 -0.5358126 -0.38093758 
+0.13797599 -0.14424857 0.29825753 -0.03976522 0.6498257 -0.2890022 -0.07700998 -0.19703439 0.51549846 -0.08980761 -0.017724093 -0.065522134 0.4251227 0.14015302 -0.42326257 -0.094496764 0.21350513 -0.30841625 0.3195116 0.15583916 
+-0.23721714 -0.06378106 0.14189686 0.0135865575 -0.16137446 -0.31988472 -0.027995463 -0.4917153 -0.07798508 -0.055689104 -0.07299139 0.3358422 -0.31224555 0.483767 0.6839953 0.080406845 -0.05582057 -0.461042 -0.46382022 0.2974709 
+0.13445474 -0.23427147 -0.2389834 0.20653479 0.33018672 -0.39512363 -0.01938143 -0.3572667 0.27603877 0.108057015 -0.010054774 0.49647596 0.17828695 -0.03504971 -0.064571954 -0.40656537 -0.4288492 0.41101915 0.21771714 -0.25679678 
+-0.06744745 -0.08651667 0.27657387 0.023940165 -0.34873477 -0.25213164 0.1468239 -0.12327071 -0.4478091 0.6570352 0.5582282 0.6778385 -0.22575188 0.3505717 -0.21985711 -0.20484298 -0.13206692 -0.5086645 0.1869471 -0.2988697 
+0.17438063 0.36547676 -0.07429125 0.16180484 -0.22732061 0.4348774 -0.24492918 -0.3274857 -0.35651952 0.38184467 -0.109175995 -0.14720117 0.56861883 -0.38205588 0.2642675 0.5543364 0.5023515 -0.23207542 -0.03440018 -0.12204064 
+-0.24434838 0.19652593 0.15170915 -0.06536993 -0.1742611 0.23247719 0.35724777 -0.07121579 0.36941552 -0.37764698 -0.32478374 0.2852862 -0.4487918 0.02041369 -0.33575803 0.16593426 -0.38837287 0.056614436 -0.012501329 -0.082749896 
+-0.41855156 0.0029425975 0.28064233 -0.015640926 -0.3093709 0.12042929 0.0812075 0.009912889 0.118127234 -0.33111233 -0.165428 0.3494854 0.278411 -0.48556912 0.5982413 0.1723589 0.2813808 -0.5124341 -0.4347027 0.023336934 

Readme.md

@@ -0,0 +1,45 @@
+# Neural Network Implementation for Handwritten Digit Recognition without Using Any Libraries in Java
+This Java project implements a neural network from scratch to recognize handwritten digits from 0 to 9, trained on the MNIST dataset. The network architecture, training process, and future enhancements are detailed below.
+
+## Project Overview
+
+### Network Architecture
+
+- **Input Layer:** 784 neurons (28x28 pixel images)
+- **Hidden Layer:** 10 neurons, activated by the Rectified Linear Unit (ReLU) function
+- **Output Layer:** 10 neurons, activated by the softmax function
+
+### Training Algorithm
+
+The neural network is trained using the Stochastic Gradient Descent (SGD) Backpropagation algorithm. It uses the Rectified Linear activation function for the hidden layer and the softmax activation function for the output layer.
+
+### Future Enhancements
+
+- **Graphical User Interface (GUI):**
+  - Implemented using JavaFX.
+  - Allows users to draw a digit on the screen, which the neural network will then predict.
+- **Utils_Multi Class Update:**
+  - The current implementation runs out of memory.
+  - An updated version will be released soon.
+
+
+### Additional Features
+
+- **Test Class:**
+    - Evaluates the neural network on the MNIST test dataset.
+    - Calculates and reports the accuracy of the neural network.
+
+## Technical Details
+
+- **Programming Language:** Java 22
+- **Build Tool:** Maven
+
+## Getting Started
+
+### Prerequisites
+
+Ensure you have the following installed:
+
+- Java 22
+- Maven
+- OS - Arch Linux x86_64

pom.xml

@@ -0,0 +1,24 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project xmlns="http://maven.apache.org/POM/4.0.0"
+         xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+         xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
+    <modelVersion>4.0.0</modelVersion>
+
+    <groupId>org.example</groupId>
+    <artifactId>Raw-NN-Digit-Recognition</artifactId>
+    <version>1.0-SNAPSHOT</version>
+
+        <dependencies>
+            <dependency>
+                <groupId>org.apache.commons</groupId>
+                <artifactId>commons-csv</artifactId>
+                <version>1.11.0</version>
+            </dependency>
+        </dependencies>
+    <properties>
+        <maven.compiler.source>22</maven.compiler.source>
+        <maven.compiler.target>22</maven.compiler.target>
+        <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
+    </properties>
+
+</project>

src/main/java/org/example/DataSet.java

@@ -0,0 +1,19 @@
+package org.example;
+
+public class DataSet {
+    private final float[][] features;
+    private final float[] labels;
+
+    public DataSet(float[][] features, float[] labels) {
+        this.features = features;
+        this.labels = labels;
+    }
+
+    public float[][] getFeatures() {
+        return features;
+    }
+
+    public float[] getLabels() {
+        return labels;
+    }
+}

src/main/java/org/example/LayerDense.java

@@ -0,0 +1,36 @@
+package org.example;
+
+public class LayerDense {
+    float[][] weights;
+    float[] biases;
+    private float[][] outputs;
+
+    public LayerDense(int nInputs, int nNeurons) {
+        this.weights = new float[nNeurons][nInputs];
+        this.biases = new float[nNeurons];
+        // Initialize weights and biases betwwen -0.5 and 0.5
+        for (int i = 0; i < nNeurons; i++) {
+            for (int j = 0; j < nInputs; j++) {
+                this.weights[i][j] = (float) (Math.random() - 0.5);
+            }
+            this.biases[i] = (float) (Math.random() - 0.5);
+        }
+    }
+
+    public void forward(float[][] inputs) {
+        this.outputs = new float[weights.length][inputs[0].length];
+        for (int i = 0; i < weights.length; i++) {
+            for (int j = 0; j < inputs[0].length; j++) {
+                float sum = 0;
+                for (int k = 0; k < inputs.length; k++) {
+                    sum += weights[i][k] * inputs[k][j];
+                }
+                this.outputs[i][j] = sum + biases[i];
+            }
+        }
+    }
+
+    public float[][] getOutputs() {
+        return outputs;
+    }
+}