abhisheks008 · ashis2004 · Jul 21, 2024
diff --git a/Sudoku Solver/Dataset/sudoku_dataset.csv b/Sudoku Solver/Dataset/sudoku_dataset.csv
diff --git a/Sudoku Solver/images/GAN model result.png b/Sudoku Solver/images/GAN model result.png
diff --git a/Sudoku Solver/images/RNN model result.png b/Sudoku Solver/images/RNN model result.png
diff --git a/Sudoku Solver/images/cnn model result.png b/Sudoku Solver/images/cnn model result.png
diff --git a/Sudoku Solver/models/Sudoku_Solver_Using_CNN,_RNN_and_GAN.ipynb b/Sudoku Solver/models/Sudoku_Solver_Using_CNN,_RNN_and_GAN.ipynb
diff --git a/Sudoku Solver/models/sudoku_gan_generator.h5 b/Sudoku Solver/models/sudoku_gan_generator.h5
diff --git a/Sudoku Solver/models/sudoku_rnn_model.h5 b/Sudoku Solver/models/sudoku_rnn_model.h5
diff --git a/Sudoku Solver/readme.md b/Sudoku Solver/readme.md
@@ -0,0 +1,26 @@
+## Introduction
+Sudoku is a popular logic-based number puzzle. This project demonstrates how to use Convolutional Neural Networks (CNN) with TensorFlow to solve Sudoku puzzles by recognizing and filling in the missing numbers.
+
+## Project Overview
+This project aims to solve Sudoku puzzles using three different deep learning algorithms: Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and Generative Adversarial Networks (GAN). Each algorithm is implemented to demonstrate its effectiveness in solving Sudoku puzzles.
+
+
+### Algorithms Used
+# Convolutional Neural Network (CNN)
+
+Training Loss: 0.8336
+Training Accuracy: 80.30%
+Test Accuracy: 80.30%
+
+# Recurrent Neural Network (RNN)
+
+Training Loss: 0.2435
+Training Accuracy: 97.10%
+Test Accuracy: 97.10%
+
+# Generative Adversarial Network (GAN)
+
+Training Epochs: 300
+Discriminator Loss: 0.1164
+Discriminator Accuracy: 98.44%
+Generator Loss: 2.9346
diff --git a/Sudoku Solver/requirement.txt b/Sudoku Solver/requirement.txt
@@ -0,0 +1,7 @@
+## Prerequisites
+To run this project, you need to have Python installed along with the following packages:
+
+TensorFlow
+Keras
+NumPy
+Pandas