Description
A simple feedforward neural network built with PyTorch to classify handwritten digits (0-9) using the MNIST dataset. This project demonstrates data preprocessing, model training, evaluation, and visualization of predictions.
Features
- Preprocess and normalize MNIST images
- Train a simple feedforward neural network
- Evaluate model accuracy on test data
- Visualize sample predictions
- Save and reload trained model parameters
Dependencies
- Python 3.x
- PyTorch
- torchvision
- NumPy
- Matplotlib