Design a Convolutional Neural Network based on PyTorch for Image Classification on the FashionMNIST Dataset

Introduction

This project was completed as part of a software course(NCUT) under the guidance of Dr. Li Xin.

Links：https://github.com/WeAreCrazyCodingMonsters/Mnist_qt?tab=readme-ov-file

Abstract

The GUI of this software is implemented using PyQt5, featuring an overall dark blue interface. When the user hovers over any button with the mouse, the button changes color. A log output window is set up at the bottom of the software to allow users to conveniently view real-time operation outputs. On the right side of the interface, there is a visualization window that helps users view the visualization of parameters during training and the predicted images.

In addition to implementing image classification on the MNIST dataset, this software also supports classification on the FashionMNIST dataset.

Apart from operating through the visual interface, the software allows users to import their configuration files with one click, saving time on repetitive operations.

The software is highly compatible and can run not only on GPU and CPU environments but has also been tested to work on both Linux and Windows systems.