IKH

Introduction

In this session, you will learn to train CNNs using Python + Keras. compared to the privious session,this session will be more hands-on.You will spend a lot of time reading and modifying Python + Keras code and train your models on GPUs.

To get started with the syntax and the process of building CNNs in Keras, you will first use the MNIST dataset (since you are already familiar with it).You will also learn to compute the number of parameters,output sizes etc. of each layer of a network.

Throughout the rest of the session,you will use theCIFAR-10 dataset which has 60000(32 x 32) colour imagees of 10 classes as shown below.In these exercises, we will also experiments of CNNs.

In this session

You will define your own convolutional layers in Keras and train those layers on the CIFAR-10 dataset . You will implement everything on a GPU.

  • Get familiar with CNNs in Keras: The MNIST dataset
  • Setting up your notebook on a GPU
  • Conduct experiments with the CIFAR-10 dataset:
    • Build a base model using the CIFAR-10 dataset
    • Experiment with hyperparameters and draw observations

Prerequisites

There are no prerequisites for this session other than knowledge of the privious session.

Guidelines for in-module questions

The in-video and in-content questions for this module are not graded. Note that graded questions are given on a separate page labelled ‘Graded Questions’ at the end of this session. The graded questions in this session will adhere to the following guidelines:

 First Attempt MarksSecond Attempt Marks
Question with 2 Attempts105
Question with 1 Attempt100

People you will hear from in this session

Subject Matter Expert:

Gopalakrishnan Srinivasaraghavan

Professor, IIIT-Bangalore

The International Institute of Information Technology, Bangalore, commonly known as IIIT Bangalore, is a premier national graduate school in India. Founded in 1999, it offers Integrated M.Tech., M.Tech., M.S. (Research) and PhD programs in the field of Information Technology.    

Report an error