Package io.github.kirstenali.deepj.nn

Class Module

java.lang.Object

io.github.kirstenali.deepj.nn.Module

All Implemented Interfaces:

Layer, Trainable

Direct Known Subclasses:

Activation, Autoencoder, BatchNorm, Conv1D, Conv2D, Dropout, EncoderAttention, Flatten, GlobalAveragePool, Pool1D, Pool2D, Recurrent, RecurrentCell, Reshape, Sequential, TransformerEncoder, TransformerEncoderLayer

public abstract class Module
extends Object
implements Layer

Additive layer base with recursive training mode and identity-deduplicated parameters.

Constructor Summary

Constructors

Constructor	Description
Module()

Method Summary

Modifier and Type	Method	Description
List<Tensor>	buffers()	Persistent non-parameter tensors, in deterministic traversal order.
protected List<Layer>	children()
final void	eval()
final boolean	isTraining()
List<Parameter>	parameters()
final void	setTraining(boolean enabled)
final void	train()

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface io.github.kirstenali.deepj.layers.Layer

forward, forward, supportsIncrementalInference

Methods inherited from interface io.github.kirstenali.deepj.training.Trainable

zeroGrad

Constructor Details

Module

public Module()

Method Details

isTraining

public final boolean isTraining()

train

public final void train()

eval

public final void eval()

setTraining

public final void setTraining(boolean enabled)

children

protected List<Layer> children()

parameters

public List<Parameter> parameters()

Specified by:

parameters in interface Trainable

buffers

public List<Tensor> buffers()

Persistent non-parameter tensors, in deterministic traversal order.