This project is part of my ongoing journey into Deep Learning, Neural Networks, and Artificial Intelligence. It focuses on building an Artificial Neural Network (ANN) to classify digits (0–9) using the popular MNIST dataset.
The goal of this project is to train an ANN that can accurately recognise and classify digits. By learning from thousands of labelled digit images, the model is able to generalise and make predictions on unseen data.
- Achieved 97.8% accuracy on the test set.
- Used 1 hidden layer with 128 nodes.
- Applied ReLU activation in the hidden layer for better learning performance.
- Used Softmax activation in the output layer (10 nodes) for multi-class classification.
- Built using TensorFlow and Keras.
- Efficient model architecture for fast convergence.
- Trained and evaluated on the MNIST dataset.
- Clean and well-commented code for better understanding and reproducibility.
- Language: Python
- Libraries & Frameworks:
TensorFlowKerasNumPyMatplotlib