This project implements an Artificial Neural Network (ANN) for classifying handwritten digits from the MNIST dataset. The model is built using PyTorch and utilizes torchvision for dataset preprocessing.

✅ Built with PyTorch – Leverages PyTorch for efficient model building and training.
✅ MNIST Dataset – A widely used benchmark dataset for digit classification.
✅ Data Augmentation – Applies torchvision.transforms for normalization and preprocessing.
✅ Performance Evaluation – Includes accuracy measurement and confusion matrix visualization.

To set up the environment, install the required dependencies using:

pip install torch torchvision numpy pandas matplotlib scikit-learn

The MNIST dataset is automatically downloaded via torchvision.datasets. It consists of 60,000 training images and 10,000 test images of handwritten digits (0-9), each sized 28×28 pixels.

The ANN model follows this architecture:

• 🔹 Input Layer: 784 neurons (flattened 28×28 image pixels)
• 🔹 Hidden Layers: Fully connected layers with ReLU activation
• 🔹 Output Layer: 10 neurons (representing digits 0-9) with Softmax activation

Execute the following script to train the model:

python train.py

Once trained, you can evaluate its performance using:

python evaluate.py

The model performance is evaluated based on:

• 📍 Accuracy: Measures the percentage of correct predictions.
• 📍 Confusion Matrix: Visualizes misclassifications across digit classes.

🎯 Training Accuracy: 98.5%
✅ Validation Accuracy: 97.8%
📌 Confusion Matrix:
[[   0    1    2    3    4    5    6    7    8    9]]

[[ 968    0    1    0    0    1    6    1    1    1]
 [   1 1122    1    0    0    0    2    3    0    2]
 [   0    1 1010    1    3    0    2    5    6    0]
 [   1    4    9  997    0   26    1    4   11    8]
 [   1    0    2    0  966    2    8    3    5   11]
 [   1    0    0    2    1  851    5    0    2    4]
 [   1    2    0    0    3    3  926    1    1    0]
 [   1    1    2    3    2    1    0  999    2    0]
 [   5    5    7    5    1    4    8    4  943    5]
 [   1    0    0    2    6    4    0    8    3  978]]

📖 PyTorch Documentation
📖 Torchvision MNIST Dataset

📝 This project is licensed under the MIT License. Feel free to modify and use it!