Project Title: Fashion$-$MNIST Classification using a Simple Neural Network

Overview: The goal of this project is to implement a basic neural network for classifying images from the

Fashion$-$MNIST dataset. Fashion$-$MNIST is a dataset of small images of fashion items, and the task is to build

a neural network that can accurately classify these items into different categories.

Project Components:

Data Loading:

Explanation: In this step, students will load the Fashion$-$MNIST dataset. Explain that the

dataset contains grayscale images of $2828$ pixels, and each image corresponds to a specific

fashion item $($e$.$g$.,$ shoes, t$-$shirts, etc.$).$

a Hints: Use the provided DataLoader class to load images and labels from the dataset.

Neural Network Model:

Explanation: The neural network will have a simple architecture with one hidden layer.

Students will implement the forward pass using ReLU activation and sofimax for the output

layer.

Hints: Discuss the structure of the Mode$1$ class.

Training the Model:

Explanation: Students will train the neural network using the Stochastic Gradient Descent

$($SGD$)$ optimizer. They will implement the training loop, forward pass, and backpropagation.

Hints: Walk through the Trainer class.

Evaluation:

Explanation: After training, students will evaluate the model's performance on the test set.

Accuracy will be used as the metric.

Hints: Discuss the Evalua tor class.

Optional: Visualization and Further Improvements:

Explanation; Students can visualize misclassified examples to gain insights into the model's

performance. Additionally, they can explore potential improvements such us uning

hyperparameters, adjusting the model architecture, or using more advanced optimization

techniques.

Hints: Suggest using the Evaluator class for visualization. Encourage students to

experiment with different hyperparameters and observe the impact on accuracy.

Project Deliverables:

Python code implementing the neural network model, training loop, and evaluation.

Ensure that your code includes comments for clarity.

Provide a README.txt file that explains how to run each part of your program.

Zip your files into a single file named your$\_$student$\_$id$\_$number,zip. $($Due to wrong naming of

your first homework, there are a lot of guys who will not get the scores they expect... You must be

careful; I explained this a lot of times in detail. The students who join the class regularly did this

appropriately.$)$

Remember, this homework is designed to challenge your understanding of various Python concepts for

Classes as the most important structure for OOP.