nz::nodes::io Namespace Reference

This namespace contains standard nodes used in computational graphs for neural networks. More...

Classes
class	InputNode
	Represents an input node in a computational graph. More...
class	OutputNode
	Base class for loss function nodes in a computational graph. More...

Detailed Description

This namespace contains standard nodes used in computational graphs for neural networks.

The nz::nodes::io namespace includes the basic building blocks for a computational graph, such as input nodes and output nodes. These nodes serve as the primary interface for data flow in neural networks.

This namespace includes the following classes:

• InputNode: Represents an input node that provides data to the computational graph.
• OutputNode: A base class for loss function nodes that connects to the output of the computational graph.

These nodes are designed to be used in a larger computational graph where data flows from one node to another during the forward and backward passes. The input node holds the input data and passes it forward, while the output node computes the loss and manages the gradient flow during backpropagation.

Note

• The InputNode does not perform computations, only providing data for the graph.
• The OutputNode serves as the base class for more specific loss functions (e.g., Mean Squared Error, Cross-Entropy).

See also

InputNode for details on providing data to the network.

OutputNode for details on loss function nodes.

Author

Mgepahmge (https://github.com/Mgepahmge)

Date

2024/11/29