An example of a fully connected neural network in Keras

Getting Started

• Install Python
• Install Tensorflow
• Install Keras
• You may also need h5py if you don't already have it
• Configure Keras to use Tensorflow

Lesson Order

The files in this directory are meant to build on each other in an order - starting with just defining a new neural network, and building towards predictions, saving and loading, and testing and evaluating the models.

Create a Fully Neural Network

create_network.py

Train a Sequential Model

train_sample_data.py

Separate Training and Validation Data

validation_split.py

Manually Set Validation Data

validation_data.py

Evaluate the Model

evaluate_model.py

Changing the Network Topology

deep_network.py

Viewing a Summary of the Model

model_summary.py

Use the Model to Make Predictions

predict.py

Save the Model to a File

save_model.py

Load the Model from a File

load_model.py

Create a Network for Binary Classification (2 Classes)

binary_classification.py

Import and Use CSV Data

import_csv.py

Make Class Predictions

predict_classes.py

Create a Mutli-class Classification Network

multi_class_model.py

Change the Learning Rate of the Network

change_learning_rate.py

Change the Learning Rate During Training

change_lr_and_keep_training.py

Load and Continue Training a Model

continue_training_model.py