Examples

This section show and explain all the examples used in DEATF. These examples are divided in two: the simple cases where predefined network descriptors are evaluated individualy and the custom cases that use custom evaluation functions and network descriptors are combined in order to create more complex networks.

Examples of the predefined networks

These cases are a collection of files where each one is for testing a predefined network. Due to that each file only uses one type of network, these files are ideal for testing. If new mutations are added or any change is done to the Network or NetworkDescriptor these would be the test to verify if everything has been correctly done.

• Auxiliar functions for testing
• Predefined MLP
• Predefined CNN
• Predefined TCNN
• Predefined RNN

Custom examples

These other cases are examples of applications where combinations and more complex models are used. They also define different levels of complexity in the evaluation method: predefined evalautions, creating functions and even creating the step of training.

• Autoencoder
• Convolution Neural Network - Autoencoder
• Convolutional Neural Network
• Generative Adversarial Network
• Multi-Network
• Recurrent Neural Network
• Sequential Network
• Multi Layer Perceptron
• Single Skip
• Weighted Neural Network