TensorFlow Beginner: Basic Image Classification

Just another repository

Project-Based Course Overview Welcome to TensorFlow Beginner: Basic Image Classification. This is a project-based course which should take approximately 2 hours to finish. Before diving into the project, please take a look at the course objectives and structure:

Course Objectives In this course, we are going to focus on three learning objectives:

  1. Learn to create, train and evaluate neural network models with TensorFlow and Keras.
  2. Understand the basics of neural networks.
  3. Learn to solve classification problems with the help of neural networks. By the end of this course, you will be able to create a neural network model which will be able to classify images of hand written digits with a high degree of accuracy.

Project Structure

The hands on project on Basic Image Classification is divided into following tasks:

Task 1:Introduction

  1. Introduction to the basic image classification problem.
  2. What is TensorFlow?
  3. Introduction to the Rhyme interface.

Task 2: The Dataset

  1. Importing the MNIST dataset.
  2. A quick look at the structure of the dataset.
  3. A quick look at the MNIST image examples.

Task 3: One Hot Encoding

  1. What is one hot encoding?
  2. How to encode the labels from the dataset.

Task 4: Neural Networks

  1. Graphical representation of linear equations.
  2. What are neural networks?
  3. What are activation functions?

Task 5: Pre-processing the Examples

  1. Unrolling the input features.
  2. Data normalization with mean and standard deviation.

Task 6: Creating the Model

  1. Creating a sequential model with Keras.
  2. Model architecture - hidden layers and hidden units.
  3. Softmax and ReLU activation functions.
  4. Compiling the model by specifying an optimizer and a loss function.

Task 7: Training the Model

  1. Training the model to fit to training data.
  2. Evaluating the model on the test data.

Task 8: Predictions

  1. Predictions on the test set.
  2. Visualizing the predictions.

Notes