/Resnet-14-cifar-10-classification

This repo hosts a ResNet-14 implementation for CIFAR-10 image classification. It leverages ResNet's residual connections for training deeper networks. Ideal for educational use and baseline comparisons.

Primary LanguageJupyter Notebook

ResNet-14 CIFAR-10 Classification

Introduction

This project is an implementation of ResNet-14 for image classification on the CIFAR-10 dataset. ResNet, or Residual Network, is a convolutional neural network (CNN) architecture which allows for the training of much deeper networks by utilizing residual connection to jump over some layers.

Dataset

The CIFAR-10 dataset consists of 60000 32x32 colour images in 10 classes, with 6000 images per class. There are 50000 training images and 10000 test images.

Model

The model used in this project is ResNet-14, a variant of the original ResNet model, which has 14 layers. The architecture is designed to solve the problem of vanishing gradients and allows for the training of much deeper networks.

Requirements

  • Python 3.6+
  • PyTorch 1.0+
  • torchvision
  • Numpy
  • Matplotlib

Results

ResNet-14 model achieved a classification accuracy of 83.54% on the CIFAR-10 test set.