Overview
Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!
Learn how to build a deep convolutional neural network (CNN) image classifier from scratch using Python, TensorFlow, and Keras in this comprehensive 1-hour 25-minute tutorial. Discover techniques for creating a data pipeline, preprocessing images, constructing and training a deep neural network, evaluating performance, and saving the model. Follow along as the instructor guides you through each step, from installing dependencies and collecting data from Google Images to scaling and partitioning the dataset. Gain hands-on experience in building and training the network, plotting model performance, and testing on new data. By the end of this tutorial, you'll have the skills to create your own image classifier using any dataset of your choice.
Syllabus
- Start
- Explainer
- PART 1: Building a Data Pipeline
- Installing Dependencies
- Getting Data from Google Images
- Load Data using Keras Utils
- PART 2: Preprocessing Data
- Scaling Images
- Partitioning the Dataset
- PART 3: Building the Deep Neural Network
- Build the Network
- Training the DNN
- Plotting Model Performance
- PART 4: Evaluating Perofmrnace
- Evaluating on the Test Partition
- Testing on New Data
- PART 5: Saving the Model
- Saving the model as h5 file
- Wrap Up
Taught by
Nicholas Renotte