Derived types

The derived types provided by the ATHENA library can be separated into subcategories. The prominent derived type is

network_type

This is the container for a neural network and all its layers.

Layer Types

The following layer types are available:

• Convolution layers
• Fully connected layers (aka dense layers)
• Normalisation layers
• Pooling layers
• Regularisation layers

• Input layers (automated, no page yet)
• Reshaping layers (automatic, no page yet)

Optimiser and associated types

There exist a set of derived types that are used to store the optimisation parameters and procedures, in addition to the training metrics. These types are:

• Base optimiser
• Stochastic Gradient Descent
• RMSprop
• Adam
• AdaGrad

• metric_dict_type