Deep Learning for Multi-Class Classification Using CIFAR-10

Project Overview

This project involves designing and implementing multiple Deep Neural Network (DNN) architectures to classify images from the CIFAR-10 dataset into 10 different categories. The objective was to evaluate and fine-tune hyperparameters, including the number of layers, activation functions, and optimizers, to achieve better performance than a baseline model.

Dataset

• 50,000 training images (32x32 color images)
• 10,000 testing images
• 10 classes: airplane, automobile, bird, cat, deer, dog, frog, horse, ship, and truck.

Objective

• Build a baseline DNN model for image classification.
• Design and fine-tune two additional DNN models with optimized hyperparameters for better performance.

Key Outcomes

• Created a baseline DNN model and two fine-tuned models.
• Improved accuracy through hyperparameter tuning and optimization.
• Evaluated model performance using metrics and comparisons.

Tools Used

• Python
• Keras
• TensorFlow
• Jupyter Notebook

View the Code

Click the link below to view the full code and documentation for this project on GitHub:

View on GitHub