ai.djl.nn.recurrent

Class RNN

java.lang.Object

ai.djl.nn.AbstractBlock

ai.djl.nn.recurrent.RecurrentBlock

ai.djl.nn.recurrent.RNN

All Implemented Interfaces:

Block

public class RNN
extends RecurrentBlock

RNN is an implementation of recurrent neural networks which applies a single-gate recurrent layer to input. Two kinds of activation function are supported: ReLU and Tanh.

Current implementation refers the [paper](https://crl.ucsd.edu/~elman/Papers/fsit.pdf), Finding structure in time - Elman, 1988.

The RNN operator is formulated as below:

With ReLU activation function: \(h_t = relu(W_{ih} * x_t + b_{ih} + W_{hh} * h_{(t-1)} + b_{hh})\)

With Tanh activation function: \(h_t = \tanh(W_{ih} * x_t + b_{ih} + W_{hh} * h_{(t-1)} + b_{hh})\)

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	RNN.Activation An enum that enumerates the type of activation.
static class	RNN.Builder The Builder to construct a RNN type of Block.

Nested classes/interfaces inherited from class ai.djl.nn.recurrent.RecurrentBlock

RecurrentBlock.BaseBuilder<T extends RecurrentBlock.BaseBuilder>

Field Summary

Fields inherited from class ai.djl.nn.recurrent.RecurrentBlock

batchFirst, bidirectional, dropRate, gates, hasBiases, numLayers, returnState, stateSize

Fields inherited from class ai.djl.nn.AbstractBlock

children, inputNames, inputShapes, parameters, version

Method Summary

Modifier and Type	Method and Description
static RNN.Builder	builder() Creates a builder to build a RNN.
protected NDList	forwardInternal(ParameterStore parameterStore, NDList inputs, boolean training, ai.djl.util.PairList<java.lang.String,java.lang.Object> params) A helper for Block.forward(ParameterStore, NDList, boolean, PairList) after initialization.

Methods inherited from class ai.djl.nn.recurrent.RecurrentBlock

beforeInitialize, getNumDirections, getOutputShapes, loadMetadata, prepare

Methods inherited from class ai.djl.nn.AbstractBlock

addChildBlock, addParameter, cast, clear, describeInput, forward, forward, forwardInternal, getChildren, getDirectParameters, getParameters, initialize, initializeChildBlocks, isInitialized, loadParameters, readInputShapes, saveInputShapes, saveMetadata, saveParameters, setInitializer, setInitializer, setInitializer, toString

Methods inherited from class java.lang.Object

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

Methods inherited from interface ai.djl.nn.Block

forward, validateLayout

Method Detail

forwardInternal

protected NDList forwardInternal(ParameterStore parameterStore,
                                 NDList inputs,
                                 boolean training,
                                 ai.djl.util.PairList<java.lang.String,java.lang.Object> params)

A helper for Block.forward(ParameterStore, NDList, boolean, PairList) after initialization.

Specified by:

forwardInternal in class AbstractBlock

Parameters:

parameterStore - the parameter store

inputs - the input NDList

training - true for a training forward pass

params - optional parameters

Returns:

the output of the forward pass

builder

public static RNN.Builder builder()

Creates a builder to build a RNN.

Returns:

a new builder