MNIST Digit Recognition using PyTorch

Overview

This project implements a simple digit recognition model using the MNIST dataset. The model is built using PyTorch and follows a feedforward neural network architecture. The network is trained to classify handwritten digits from 0 to 9 based on images of size 28x28 pixels.

Table of Contents

• Installation
• Model Architecture
• Training and Evaluation
• Logging and Visualization

Installation

To run this project, you need to have Python and the following packages installed:

• PyTorch
• torchvision
• matplotlib
• Pillow
• TensorBoard

Model Architecture

The model is defined in the MyFirstNetwork class, which consists of the following layers:

• Input Layer: Accepts input images of size 28x28 pixels, flattened to a vector of size 784.
• Hidden Layer: Contains 100 neurons with ReLU activation.
• Output Layer: Outputs predictions for 10 classes (digits 0-9).

Training and Evaluation

The training process includes the following steps:

• Data Loading: The MNIST dataset is downloaded and split into training, validation, and test sets.
• Training Loop: For each epoch, the model performs a forward pass, computes loss, performs backpropagation, and updates model parameters.
• Evaluation: The model's performance is evaluated on the validation and test datasets at the end of each epoch.

Logging and Visualization

Logging is handled through Python's built-in logging library, which saves training progress and metrics to a file. Additionally, TensorBoard is used to visualize training loss and accuracy over epochs.