This project focuses on developing a Convolutional Neural Network (CNN) to classify images from the CIFAR10 database. A key goal is to achieve over 80% accuracy with less than 50,000 trainable parameters. Notably, the project incorporates various data augmentation techniques to enhance model performance and generalization. The repository contains all training and evaluation codes, the trained model, and a comprehensive report. This report details the data augmentation strategies used and explains the chosen model architecture, providing an in-depth look at the process and decisions involved in this image classification task.