BlackCat_Tensors
A GPU-supported autograd and linear algebra library, designed for neural network construction
bc::nn::LayerChain< Index, Derived,... > Struct Template Reference

Layer_Chain is an iterator-like object that connects different types of neural-network layers and defines convenient iterator-like methods. More...

#include <layer_chain.h>

Detailed Description

template<class Index, class Derived, class...>
struct bc::nn::LayerChain< Index, Derived,... >

Layer_Chain is an iterator-like object that connects different types of neural-network layers and defines convenient iterator-like methods.

It is tightly coupled with Layer_Manager Layer_Manager expects LayerChain to derive from itself.

This enables Layer_Manager (which inherits from a neural-network layer) to cast-itself to the LayerChain class and than access the other layers of the network.

This enables Layer_Manager's to have acess to all other layer's of the neural network.

The Layer_Chain simply acts as an iterator-like object. The Layer_Manager handles memory and time_indexes (for recurrent layers) for actual neural-network layers.


The documentation for this struct was generated from the following file: