Class and Description 

Layer
A neural network layer.

Class and Description 

AbstractLSTM
LSTM recurrent net, based on Graves: Supervised Sequence Labelling with Recurrent Neural Networks
http://www.cs.toronto.edu/~graves/phd.pdf

AbstractLSTM.Builder 
ActivationLayer
Activation layer is a simple layer that applies the specified activation function to the input activations

AutoEncoder
Autoencoder layer.

AutoEncoder.Builder 
BaseLayer
A neural network layer.

BaseLayer.Builder 
BaseOutputLayer 
BaseOutputLayer.Builder 
BasePretrainNetwork 
BatchNormalization
Batch normalization layer
See: Ioffe and Szegedy, 2014, Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift https://arxiv.org/abs/1502.03167 
BatchNormalization.Builder 
CenterLossOutputLayer
Center loss is similar to triplet loss except that it enforces intraclass consistency and doesn't require feed
forward of multiple examples.

CenterLossOutputLayer.Builder 
ConvolutionLayer
2D Convolution layer (for example, spatial convolution over images).

ConvolutionLayer.BaseConvBuilder 
ConvolutionLayer.Builder 
Deconvolution2D
2D deconvolution layer configuration
Deconvolutions are also known as transpose convolutions or fractionally strided convolutions. 
Deconvolution2D.Builder 
DenseLayer
Dense layer: a standard fully connected feed forward layer

DenseLayer.Builder 
DropoutLayer
Dropout layer.

EmbeddingLayer
Embedding layer: feedforward layer that expects single integers per example as input (class numbers, in range 0 to
numClass1) as input.

EmbeddingLayer.Builder 
FeedForwardLayer
Created by jeffreytang on 7/21/15.

FeedForwardLayer.Builder 
GlobalPoolingLayer
Global pooling layer  used to do pooling over time for RNNs, and 2d pooling for CNNs.

GravesBidirectionalLSTM
Deprecated.
use
Bidirectional instead. With the Bidirectional
layer wrapper you can make any recurrent layer bidirectional, in particular GravesLSTM. Note that this layer adds the
output of both directions, which translates into "ADD" mode in Bidirectional.
Usage: .layer(new Bidirectional(Bidirectional.Mode.ADD, new GravesLSTM.Builder()....build())) 
GravesBidirectionalLSTM.Builder
Deprecated.

GravesLSTM
Deprecated.
Will be eventually removed. Use
LSTM instead, which has similar prediction accuracy, but supports
CuDNN for faster network training on CUDA (Nvidia) GPUs 
GravesLSTM.Builder
Deprecated.

Layer
A neural network layer.

Layer.Builder 
LocalResponseNormalization
Local response normalization layer
See section 3.3 of http://www.cs.toronto.edu/~fritz/absps/imagenet.pdf 
LocalResponseNormalization.Builder 
LossLayer
LossLayer is a flexible output layer that performs a loss function on an input without MLP logic.

LSTM
LSTM recurrent neural network layer without peephole connections.

LSTM.Builder 
OutputLayer
Output layer used for training via backpropagation based on labels and a specified loss function.

OutputLayer.Builder 
PoolingType
Pooling type:
MAX: Max pooling  output is the maximum value of the input values AVG: Average pooling  output is the average value of the input values SUM: Sum pooling  output is the sum of the input values PNORM: Pnorm pooling 
RnnOutputLayer
A version of
OutputLayer for recurrent neural networks. 
RnnOutputLayer.Builder 
SeparableConvolution2D
2D Separable convolution layer configuration.

SeparableConvolution2D.Builder 
SubsamplingLayer
Subsampling layer also referred to as pooling in convolution neural nets
Supports the following pooling types: MAX, AVG, SUM, PNORM

SubsamplingLayer.Builder 
SubsamplingLayer.PoolingType 
Class and Description 

Layer
A neural network layer.

Class and Description 

Layer
A neural network layer.

Class and Description 

ConvolutionLayer.AlgoMode
The "PREFER_FASTEST" mode will pick the fastest algorithm for the specified parameters from the
ConvolutionLayer.FwdAlgo ,
ConvolutionLayer.BwdFilterAlgo , and ConvolutionLayer.BwdDataAlgo lists, but they may be very memory intensive, so if weird errors
occur when using cuDNN, please try the "NO_WORKSPACE" mode. 
Layer
A neural network layer.

Class and Description 

Convolution3D.DataFormat
An optional dataFormat: "NDHWC" or "NCDHW".

Class and Description 

AbstractLSTM
LSTM recurrent net, based on Graves: Supervised Sequence Labelling with Recurrent Neural Networks
http://www.cs.toronto.edu/~graves/phd.pdf

AbstractLSTM.Builder 
ActivationLayer
Activation layer is a simple layer that applies the specified activation function to the input activations

ActivationLayer.Builder 
AutoEncoder
Autoencoder layer.

AutoEncoder.Builder 
BaseLayer
A neural network layer.

BaseLayer.Builder 
BaseOutputLayer 
BaseOutputLayer.Builder 
BasePretrainNetwork 
BasePretrainNetwork.Builder 
BaseRecurrentLayer 
BaseRecurrentLayer.Builder 
BaseUpsamplingLayer
Upsampling base layer

BaseUpsamplingLayer.UpsamplingBuilder 
BatchNormalization
Batch normalization layer
See: Ioffe and Szegedy, 2014, Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift https://arxiv.org/abs/1502.03167 
BatchNormalization.Builder 
CapsuleLayer.Builder 
CapsuleStrengthLayer.Builder 
CenterLossOutputLayer
Center loss is similar to triplet loss except that it enforces intraclass consistency and doesn't require feed
forward of multiple examples.

CenterLossOutputLayer.Builder 
Cnn3DLossLayer
3D Convolutional Neural Network Loss Layer.

Cnn3DLossLayer.Builder 
CnnLossLayer
Convolutional Neural Network Loss Layer.

CnnLossLayer.Builder 
Convolution1DLayer
1D (temporal) convolutional layer.

Convolution1DLayer.Builder 
Convolution3D
3D convolution layer configuration

Convolution3D.Builder 
Convolution3D.DataFormat
An optional dataFormat: "NDHWC" or "NCDHW".

ConvolutionLayer
2D Convolution layer (for example, spatial convolution over images).

ConvolutionLayer.AlgoMode
The "PREFER_FASTEST" mode will pick the fastest algorithm for the specified parameters from the
ConvolutionLayer.FwdAlgo ,
ConvolutionLayer.BwdFilterAlgo , and ConvolutionLayer.BwdDataAlgo lists, but they may be very memory intensive, so if weird errors
occur when using cuDNN, please try the "NO_WORKSPACE" mode. 
ConvolutionLayer.BaseConvBuilder 
ConvolutionLayer.Builder 
ConvolutionLayer.BwdDataAlgo
The backward data algorithm to use when
ConvolutionLayer.AlgoMode is set to "USER_SPECIFIED". 
ConvolutionLayer.BwdFilterAlgo
The backward filter algorithm to use when
ConvolutionLayer.AlgoMode is set to "USER_SPECIFIED". 
ConvolutionLayer.FwdAlgo
The forward algorithm to use when
ConvolutionLayer.AlgoMode is set to "USER_SPECIFIED". 
Deconvolution2D
2D deconvolution layer configuration
Deconvolutions are also known as transpose convolutions or fractionally strided convolutions. 
Deconvolution2D.Builder 
Deconvolution3D
3D deconvolution layer configuration
Deconvolutions are also known as transpose convolutions or fractionally strided convolutions. 
Deconvolution3D.Builder 
DenseLayer
Dense layer: a standard fully connected feed forward layer

DenseLayer.Builder 
DepthwiseConvolution2D
2D depthwise convolution layer configuration.

DepthwiseConvolution2D.Builder 
DropoutLayer
Dropout layer.

EmbeddingLayer
Embedding layer: feedforward layer that expects single integers per example as input (class numbers, in range 0 to
numClass1) as input.

EmbeddingLayer.Builder 
EmbeddingSequenceLayer
Embedding layer for sequences: feedforward layer that expects fixedlength number (inputLength) of integers/indices
per example as input, ranged from 0 to numClasses  1.

EmbeddingSequenceLayer.Builder 
FeedForwardLayer
Created by jeffreytang on 7/21/15.

FeedForwardLayer.Builder 
GlobalPoolingLayer
Global pooling layer  used to do pooling over time for RNNs, and 2d pooling for CNNs.

GlobalPoolingLayer.Builder 
GravesBidirectionalLSTM
Deprecated.
use
Bidirectional instead. With the Bidirectional
layer wrapper you can make any recurrent layer bidirectional, in particular GravesLSTM. Note that this layer adds the
output of both directions, which translates into "ADD" mode in Bidirectional.
Usage: .layer(new Bidirectional(Bidirectional.Mode.ADD, new GravesLSTM.Builder()....build())) 
GravesBidirectionalLSTM.Builder
Deprecated.

GravesLSTM
Deprecated.
Will be eventually removed. Use
LSTM instead, which has similar prediction accuracy, but supports
CuDNN for faster network training on CUDA (Nvidia) GPUs 
Layer
A neural network layer.

Layer.Builder 
LearnedSelfAttentionLayer
Implements Dot Product Self Attention with learned queries
Takes in RNN style input in the shape of [batchSize, features, timesteps]
and applies dot product attention using learned queries.

LearnedSelfAttentionLayer.Builder 
LocallyConnected1D
SameDiff version of a 1D locally connected layer.

LocallyConnected1D.Builder 
LocallyConnected2D
SameDiff version of a 2D locally connected layer.

LocallyConnected2D.Builder 
LocalResponseNormalization
Local response normalization layer
See section 3.3 of http://www.cs.toronto.edu/~fritz/absps/imagenet.pdf 
LocalResponseNormalization.Builder 
LossLayer
LossLayer is a flexible output layer that performs a loss function on an input without MLP logic.

LossLayer.Builder 
LSTM
LSTM recurrent neural network layer without peephole connections.

NoParamLayer 
OutputLayer
Output layer used for training via backpropagation based on labels and a specified loss function.

OutputLayer.Builder 
PoolingType
Pooling type:
MAX: Max pooling  output is the maximum value of the input values AVG: Average pooling  output is the average value of the input values SUM: Sum pooling  output is the sum of the input values PNORM: Pnorm pooling 
PReLULayer
Parametrized Rectified Linear Unit (PReLU)
f(x) = alpha * x for x < 0, f(x) = x for x >= 0
alpha has the same shape as x and is a learned parameter. 
PReLULayer.Builder 
PrimaryCapsules.Builder 
RecurrentAttentionLayer
Implements Recurrent Dot Product Attention
Takes in RNN style input in the shape of [batchSize, features, timesteps]
and applies dot product attention using the hidden state as the query and
all time steps as keys/values.

RecurrentAttentionLayer.Builder 
RnnLossLayer
Recurrent Neural Network Loss Layer.

RnnLossLayer.Builder 
RnnOutputLayer
A version of
OutputLayer for recurrent neural networks. 
RnnOutputLayer.Builder 
SelfAttentionLayer
Implements Dot Product Self Attention
Takes in RNN style input in the shape of [batchSize, features, timesteps]
and applies dot product attention using each timestep as the query.

SelfAttentionLayer.Builder 
SeparableConvolution2D
2D Separable convolution layer configuration.

SeparableConvolution2D.Builder 
SpaceToBatchLayer
Space to batch utility layer configuration for convolutional input types.

SpaceToBatchLayer.Builder 
SpaceToDepthLayer
Space to channels utility layer configuration for convolutional input types.

SpaceToDepthLayer.Builder 
SpaceToDepthLayer.DataFormat
Deprecated.
Use
CNN2DFormat instead 
Subsampling1DLayer
1D (temporal) subsampling layer  also known as pooling layer.

Subsampling1DLayer.Builder 
Subsampling3DLayer
3D subsampling / pooling layer for convolutional neural networks
Supports max and average pooling modes

Subsampling3DLayer.BaseSubsamplingBuilder 
Subsampling3DLayer.Builder 
Subsampling3DLayer.PoolingType 
SubsamplingLayer
Subsampling layer also referred to as pooling in convolution neural nets
Supports the following pooling types: MAX, AVG, SUM, PNORM

SubsamplingLayer.BaseSubsamplingBuilder 
SubsamplingLayer.Builder 
SubsamplingLayer.PoolingType 
Upsampling1D
Upsampling 1D layer
Repeats each step size times along the temporal/sequence axis (dimension 2)For input shape [minibatch, channels, sequenceLength] output has shape [minibatch, channels, size *
sequenceLength] Example: If input (for a single example, with channels down page, and sequence from left to right) is: [ A1, A2, A3] [ B1, B2, B3] Then output with size = 2 is: [ A1, A1, A2, A2, A3, A3] [ B1, B1, B2, B2, B3, B2] 
Upsampling1D.Builder 
Upsampling2D
Upsampling 2D layer
Repeats each value (or rather, set of depth values) in the height and width dimensions by size[0] and size[1] times respectively. 
Upsampling2D.Builder 
Upsampling3D
Upsampling 3D layer
Repeats each value (all channel values for each x/y/z location) by size[0], size[1] and size[2] If input has shape [minibatch, channels, depth, height, width] then output has shape [minibatch, channels, size[0] * depth, size[1] * height, size[2] * width] 
Upsampling3D.Builder 
ZeroPadding1DLayer
Zero padding 1D layer for convolutional neural networks.

ZeroPadding3DLayer
Zero padding 3D layer for convolutional neural networks.

ZeroPaddingLayer
Zero padding layer for convolutional neural networks (2D CNNs).

ZeroPaddingLayer.Builder 
Class and Description 

Layer
A neural network layer.

Layer.Builder 
NoParamLayer 
Class and Description 

BaseLayer
A neural network layer.

BaseLayer.Builder 
FeedForwardLayer
Created by jeffreytang on 7/21/15.

FeedForwardLayer.Builder 
Layer
A neural network layer.

Layer.Builder 
Class and Description 

Layer
A neural network layer.

Layer.Builder 
Class and Description 

BaseLayer
A neural network layer.

BaseLayer.Builder 
BaseRecurrentLayer 
BaseRecurrentLayer.Builder 
FeedForwardLayer
Created by jeffreytang on 7/21/15.

FeedForwardLayer.Builder 
Layer
A neural network layer.

Layer.Builder 
Class and Description 

Layer
A neural network layer.

Layer.Builder 
Class and Description 

Layer
A neural network layer.

Layer.Builder 
NoParamLayer 
Class and Description 

BaseLayer
A neural network layer.

BaseLayer.Builder 
BasePretrainNetwork 
BasePretrainNetwork.Builder 
FeedForwardLayer
Created by jeffreytang on 7/21/15.

FeedForwardLayer.Builder 
Layer
A neural network layer.

Layer.Builder 
Class and Description 

Layer
A neural network layer.

Layer.Builder 
Class and Description 

BaseLayer
A neural network layer.

BaseLayer.Builder 
BaseOutputLayer 
BaseOutputLayer.Builder 
FeedForwardLayer
Created by jeffreytang on 7/21/15.

FeedForwardLayer.Builder 
Layer
A neural network layer.

Layer.Builder 
Class and Description 

Convolution3D.DataFormat
An optional dataFormat: "NDHWC" or "NCDHW".

Class and Description 

BaseLayer
A neural network layer.

BaseOutputLayer 
Layer
A neural network layer.

Class and Description 

BaseOutputLayer 
BasePretrainNetwork 
Layer
A neural network layer.

Class and Description 

ConvolutionLayer.AlgoMode
The "PREFER_FASTEST" mode will pick the fastest algorithm for the specified parameters from the
ConvolutionLayer.FwdAlgo ,
ConvolutionLayer.BwdFilterAlgo , and ConvolutionLayer.BwdDataAlgo lists, but they may be very memory intensive, so if weird errors
occur when using cuDNN, please try the "NO_WORKSPACE" mode. 
ConvolutionLayer.BwdDataAlgo
The backward data algorithm to use when
ConvolutionLayer.AlgoMode is set to "USER_SPECIFIED". 
ConvolutionLayer.BwdFilterAlgo
The backward filter algorithm to use when
ConvolutionLayer.AlgoMode is set to "USER_SPECIFIED". 
ConvolutionLayer.FwdAlgo
The forward algorithm to use when
ConvolutionLayer.AlgoMode is set to "USER_SPECIFIED". 
Class and Description 

PoolingType
Pooling type:
MAX: Max pooling  output is the maximum value of the input values AVG: Average pooling  output is the average value of the input values SUM: Sum pooling  output is the sum of the input values PNORM: Pnorm pooling 
Class and Description 

ConvolutionLayer.AlgoMode
The "PREFER_FASTEST" mode will pick the fastest algorithm for the specified parameters from the
ConvolutionLayer.FwdAlgo ,
ConvolutionLayer.BwdFilterAlgo , and ConvolutionLayer.BwdDataAlgo lists, but they may be very memory intensive, so if weird errors
occur when using cuDNN, please try the "NO_WORKSPACE" mode. 
ConvolutionLayer.BwdDataAlgo
The backward data algorithm to use when
ConvolutionLayer.AlgoMode is set to "USER_SPECIFIED". 
ConvolutionLayer.BwdFilterAlgo
The backward filter algorithm to use when
ConvolutionLayer.AlgoMode is set to "USER_SPECIFIED". 
ConvolutionLayer.FwdAlgo
The forward algorithm to use when
ConvolutionLayer.AlgoMode is set to "USER_SPECIFIED". 
PoolingType
Pooling type:
MAX: Max pooling  output is the maximum value of the input values AVG: Average pooling  output is the average value of the input values SUM: Sum pooling  output is the sum of the input values PNORM: Pnorm pooling 
Class and Description 

Layer
A neural network layer.

Class and Description 

AbstractLSTM
LSTM recurrent net, based on Graves: Supervised Sequence Labelling with Recurrent Neural Networks
http://www.cs.toronto.edu/~graves/phd.pdf

BaseRecurrentLayer 
FeedForwardLayer
Created by jeffreytang on 7/21/15.

GravesBidirectionalLSTM
Deprecated.
use
Bidirectional instead. With the Bidirectional
layer wrapper you can make any recurrent layer bidirectional, in particular GravesLSTM. Note that this layer adds the
output of both directions, which translates into "ADD" mode in Bidirectional.
Usage: .layer(new Bidirectional(Bidirectional.Mode.ADD, new GravesLSTM.Builder()....build())) 
Class and Description 

Layer
A neural network layer.

Class and Description 

Layer
A neural network layer.

Class and Description 

FeedForwardLayer
Created by jeffreytang on 7/21/15.

Layer
A neural network layer.

Class and Description 

ActivationLayer
Activation layer is a simple layer that applies the specified activation function to the input activations

PReLULayer
Parametrized Rectified Linear Unit (PReLU)
f(x) = alpha * x for x < 0, f(x) = x for x >= 0
alpha has the same shape as x and is a learned parameter. 
Class and Description 

Convolution1DLayer
1D (temporal) convolutional layer.

ConvolutionLayer
2D Convolution layer (for example, spatial convolution over images).

Deconvolution2D
2D deconvolution layer configuration
Deconvolutions are also known as transpose convolutions or fractionally strided convolutions. 
DepthwiseConvolution2D
2D depthwise convolution layer configuration.

SeparableConvolution2D
2D Separable convolution layer configuration.

SpaceToDepthLayer
Space to channels utility layer configuration for convolutional input types.

Upsampling1D
Upsampling 1D layer
Repeats each step size times along the temporal/sequence axis (dimension 2)For input shape [minibatch, channels, sequenceLength] output has shape [minibatch, channels, size *
sequenceLength] Example: If input (for a single example, with channels down page, and sequence from left to right) is: [ A1, A2, A3] [ B1, B2, B3] Then output with size = 2 is: [ A1, A1, A2, A2, A3, A3] [ B1, B1, B2, B2, B3, B2] 
Upsampling2D
Upsampling 2D layer
Repeats each value (or rather, set of depth values) in the height and width dimensions by size[0] and size[1] times respectively. 
Upsampling3D
Upsampling 3D layer
Repeats each value (all channel values for each x/y/z location) by size[0], size[1] and size[2] If input has shape [minibatch, channels, depth, height, width] then output has shape [minibatch, channels, size[0] * depth, size[1] * height, size[2] * width] 
ZeroPadding1DLayer
Zero padding 1D layer for convolutional neural networks.

ZeroPadding3DLayer
Zero padding 3D layer for convolutional neural networks.

ZeroPaddingLayer
Zero padding layer for convolutional neural networks (2D CNNs).

Class and Description 

ActivationLayer
Activation layer is a simple layer that applies the specified activation function to the input activations

DenseLayer
Dense layer: a standard fully connected feed forward layer

DropoutLayer
Dropout layer.

Class and Description 

LocalResponseNormalization
Local response normalization layer
See section 3.3 of http://www.cs.toronto.edu/~fritz/absps/imagenet.pdf 
Class and Description 

EmbeddingSequenceLayer
Embedding layer for sequences: feedforward layer that expects fixedlength number (inputLength) of integers/indices
per example as input, ranged from 0 to numClasses  1.

Class and Description 

LocallyConnected1D
SameDiff version of a 1D locally connected layer.

LocallyConnected2D
SameDiff version of a 2D locally connected layer.

Class and Description 

DropoutLayer
Dropout layer.

Class and Description 

BatchNormalization
Batch normalization layer
See: Ioffe and Szegedy, 2014, Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift https://arxiv.org/abs/1502.03167 
Class and Description 

GlobalPoolingLayer
Global pooling layer  used to do pooling over time for RNNs, and 2d pooling for CNNs.

PoolingType
Pooling type:
MAX: Max pooling  output is the maximum value of the input values AVG: Average pooling  output is the average value of the input values SUM: Sum pooling  output is the sum of the input values PNORM: Pnorm pooling 
Subsampling1DLayer
1D (temporal) subsampling layer  also known as pooling layer.

Subsampling3DLayer
3D subsampling / pooling layer for convolutional neural networks
Supports max and average pooling modes

SubsamplingLayer
Subsampling layer also referred to as pooling in convolution neural nets
Supports the following pooling types: MAX, AVG, SUM, PNORM

Class and Description 

Layer
A neural network layer.

Class and Description 

Layer
A neural network layer.

Class and Description 

Layer
A neural network layer.

Class and Description 

ConvolutionLayer.AlgoMode
The "PREFER_FASTEST" mode will pick the fastest algorithm for the specified parameters from the
ConvolutionLayer.FwdAlgo ,
ConvolutionLayer.BwdFilterAlgo , and ConvolutionLayer.BwdDataAlgo lists, but they may be very memory intensive, so if weird errors
occur when using cuDNN, please try the "NO_WORKSPACE" mode. 
Layer
A neural network layer.

Class and Description 

Convolution3D.DataFormat
An optional dataFormat: "NDHWC" or "NCDHW".

Layer
A neural network layer.

PoolingType
Pooling type:
MAX: Max pooling  output is the maximum value of the input values AVG: Average pooling  output is the average value of the input values SUM: Sum pooling  output is the sum of the input values PNORM: Pnorm pooling 
Class and Description 

BatchNormalization
Batch normalization layer
See: Ioffe and Szegedy, 2014, Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift https://arxiv.org/abs/1502.03167 
ConvolutionLayer
2D Convolution layer (for example, spatial convolution over images).

DenseLayer
Dense layer: a standard fully connected feed forward layer

SubsamplingLayer
Subsampling layer also referred to as pooling in convolution neural nets
Supports the following pooling types: MAX, AVG, SUM, PNORM

SubsamplingLayer.PoolingType 
Copyright © 2020. All rights reserved.