CISC7016 Course Paper Implementation Code
The implementation code of this course paper is based on PyTorch framework: https://pytorch.org/
The dataset (CIFAR-10) used in this implementation: https://www.cs.toronto.edu/~kriz/cifar.html
Programming environment: WSL Linux Sub-system of Windows (Debian) & Anaconda
Python version: 3.10.15
Hardware specification: AMD Ryzen 7 6800H with Radeon Graphics CPU (16 GB) & NVIDIA GeForce RTX 3050 Ti GPU (4 GB)
Title: A Comparative Study of Multi-layer Perceptron, Convolutional Neural Network, and Transfer Learning Architectures for CIFAR-10 Image Classification
Abstract: This report explores image classification methods using Deep Learning (DL) on the CIFAR-10 dataset, focusing on three architectures: Multi-layer Perceptron (MLP), Convolutional Neural Network (CNN), and Transfer Learning (TL). Customized models based on these architectures are proposed, implemented, and evaluated. The report concludes with experimental results and a comparative analysis of the model's performance.
The architecture of multi-layer perceptron (MLP):
The architecture of convolutional neural network (CNN):
The architecture of transfer learning (TL):
If you want to reproduce the experiments, please ensure your environment has been configured correctly and execute the following commands step by step:
chmod +x script.sh./script.shContributor:
Yumu Xie (MC451742) [email protected]
Department of Computer and Information Science, Faculty of Science and Technology, University of Macau