Fully Connected Neural Networks with Keras

Overview

Neural networks, with Keras, bring powerful machine learning to Python applications. They can answer questions like “How much traffic will hit my website tonight?” or answer classification questions like “Will this customer buy our product?” or “Will the stock price go up or down tomorrow?”
In this course, we’ll build a fully connected neural network with Keras. This is the most basic type of neural network you can create, but it’s powerful in application and can jumpstart your exploration of other frameworks.
We’ll start the course by creating the primary network. Then we’ll:
build and configure the network, then evaluate and test the accuracy of each
save the model and learn how to load it and use it to make predictions in the future
expose the model as part of a tiny web application that can be used to make predictions
You don’t need to know a lot of Python for this course, but some basic Python knowledge will be helpful. Enjoy!

Syllabus

Course Introduction: Fully Connected Neural Networks with Keras
Create a Fully Connected TensorFlow Neural Network with Keras
Train a Sequential Keras Model with Sample Data
Separate Training and Validation Data Automatically in Keras with validation_split
Manually Set Validation Data While Training a Keras Model
Evaluate a Keras Model with Test Data
Testing Different Neural Network Topologies
Understand the Structure of a Keras Model by Viewing the Model Summary
Make Predictions on New Data with a Trained Keras Models
Save a Trained Keras Model Weights and Topology to a File
Load and Use a Saved Keras Model
Create a Neural Network for Two Category Classification with Keras
Import Data From a CSV to Use with a Keras Model Using NumPy’s genfromtxt Method
Make Binary Class Predictions with Keras Using predict and predict_classes
Create a Dense Neural Network for Multi Category Classification with Keras
Make Predictions on New Data with a Multi Category Classification Network
Change the Learning Rate of the Adam Optimizer on a Keras Network
Change the Optimizer Learning Rate During Keras Model Training
Continue to Train an Already Trained Keras Model with New Data