DL-mini-project

CIFAR 10 image classification using a modified Residual Network architecture

Abstract

This study examines the application of residual networks in image classification tasks. Various modifications to the ResNet18 model architecture, including different optimizers, learning rates, and the number of residual blocks, will be compared to identify the optimal model. The constraint imposed on the model is that it must have fewer than 5 million parameters, and the objective is to achieve the highest possible accuracy.

Paper

https://docs.google.com/document/d/1Y9xaxNu1u6MfB2mYYDFZGYhLF6YVZQWvj0ym7vsWfL4/edit?usp=sharing

Team Members

  • Abhignya Bhat
  • Anshika Gupta
  • Mudassir Hussain