Package ai.djl.nn.convolutional

Class Conv2dTranspose

java.lang.Object

ai.djl.nn.AbstractBaseBlock

ai.djl.nn.AbstractBlock

ai.djl.nn.convolutional.Deconvolution

ai.djl.nn.convolutional.Conv2dTranspose

All Implemented Interfaces:

Block

public class Conv2dTranspose
extends Deconvolution

The input to a Conv2dTranspose is an NDList with a single 4-D NDArray. The layout of the NDArray must be "NCHW". The shapes are

• data: (batch_size, channel, height, width)
• weight: (num_filter, channel, kernel[0], kernel[1])
• bias: (num_filter,)
• out: (batch_size, num_filter, out_height, out_width)
out_height = f(height, kernel[0], pad[0], oPad[0], stride[0], dilate[0])
out_width = f(width, kernel[1], pad[1], oPad[1], stride[1], dilate[1])
where f(x, k, p, oP, s, d) = (x-1)*s-2*p+k+oP

Both weight and bias are learn-able parameters.

See Also:

Deconvolution

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	Conv2dTranspose.Builder	The Builder to construct a Conv2dTranspose type of Block.

Nested classes/interfaces inherited from class ai.djl.nn.convolutional.Deconvolution

Deconvolution.DeconvolutionBuilder<T extends Deconvolution.DeconvolutionBuilder>

Field Summary

Fields inherited from class ai.djl.nn.convolutional.Deconvolution

bias, dilation, filters, groups, includeBias, kernelShape, outPadding, padding, stride, weight

Fields inherited from class ai.djl.nn.AbstractBlock

children, parameters

Fields inherited from class ai.djl.nn.AbstractBaseBlock

inputNames, inputShapes, outputDataTypes, version

Method Summary

Modifier and Type	Method	Description
static Conv2dTranspose.Builder	builder()	Creates a builder to build a Conv2dTranspose.
static NDList	conv2dTranspose(NDArray input, NDArray weight)	Applies 2D deconvolution over an input signal composed of several input planes.
static NDList	conv2dTranspose(NDArray input, NDArray weight, NDArray bias)	Applies 2D deconvolution over an input signal composed of several input planes.
static NDList	conv2dTranspose(NDArray input, NDArray weight, NDArray bias, Shape stride)	Applies 2D deconvolution over an input signal composed of several input planes.
static NDList	conv2dTranspose(NDArray input, NDArray weight, NDArray bias, Shape stride, Shape padding)	Applies 2D deconvolution over an input signal composed of several input planes.
static NDList	conv2dTranspose(NDArray input, NDArray weight, NDArray bias, Shape stride, Shape padding, Shape outPadding)	Applies 2D deconvolution over an input signal composed of several input planes.
static NDList	conv2dTranspose(NDArray input, NDArray weight, NDArray bias, Shape stride, Shape padding, Shape outPadding, Shape dilation)	Applies 2D deconvolution over an input signal composed of several input planes.
static NDList	conv2dTranspose(NDArray input, NDArray weight, NDArray bias, Shape stride, Shape padding, Shape outPadding, Shape dilation, int groups)	Applies 2D deconvolution over an input signal composed of several input planes.
protected LayoutType[]	getExpectedLayout()	Returns the expected layout of the input.
protected java.lang.String	getStringLayout()	Returns the string representing the layout of the input.
protected int	numDimensions()	Returns the number of dimensions of the input.

Methods inherited from class ai.djl.nn.convolutional.Deconvolution

beforeInitialize, forwardInternal, getOutputShapes, loadMetadata, prepare

Methods inherited from class ai.djl.nn.AbstractBlock

addChildBlock, addChildBlock, addChildBlockSingleton, addParameter, getChildren, getDirectParameters

Methods inherited from class ai.djl.nn.AbstractBaseBlock

cast, clear, describeInput, forward, forward, forwardInternal, getInputShapes, getOutputDataTypes, 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, freezeParameters, freezeParameters, getOutputShapes

Method Detail

getExpectedLayout

protected LayoutType[] getExpectedLayout()

Returns the expected layout of the input.

Specified by:

getExpectedLayout in class Deconvolution

Returns:

the expected layout of the input

getStringLayout

protected java.lang.String getStringLayout()

Returns the string representing the layout of the input.

Specified by:

getStringLayout in class Deconvolution

Returns:

the string representing the layout of the input

numDimensions

protected int numDimensions()

Returns the number of dimensions of the input.

Specified by:

numDimensions in class Deconvolution

Returns:

the number of dimensions of the input

conv2dTranspose

public static NDList conv2dTranspose(NDArray input,
                                     NDArray weight)

Applies 2D deconvolution over an input signal composed of several input planes.

Parameters:

input - the input NDArray of shape (batchSize, inputChannel, height, width)

weight - filters NDArray of shape (outChannel, inputChannel/groups, height, width)

Returns:

the output of the conv2dTranspose operation

conv2dTranspose

public static NDList conv2dTranspose(NDArray input,
                                     NDArray weight,
                                     NDArray bias)

Applies 2D deconvolution over an input signal composed of several input planes.

Parameters:

input - the input NDArray of shape (batchSize, inputChannel, height, width)

weight - filters NDArray of shape (outChannel, inputChannel/groups, height, width)

bias - bias NDArray of shape (outChannel)

Returns:

the output of the conv2dTranspose operation

conv2dTranspose

public static NDList conv2dTranspose(NDArray input,
                                     NDArray weight,
                                     NDArray bias,
                                     Shape stride)

Applies 2D deconvolution over an input signal composed of several input planes.

Parameters:

input - the input NDArray of shape (batchSize, inputChannel, height, width)

weight - filters NDArray of shape (outChannel, inputChannel/groups, height, width)

bias - bias NDArray of shape (outChannel)

stride - the stride of the deconvolving kernel: Shape(height, width)

Returns:

the output of the conv2dTranspose operation

conv2dTranspose

public static NDList conv2dTranspose(NDArray input,
                                     NDArray weight,
                                     NDArray bias,
                                     Shape stride,
                                     Shape padding)

Applies 2D deconvolution over an input signal composed of several input planes.

Parameters:

input - the input NDArray of shape (batchSize, inputChannel, height, width)

weight - filters NDArray of shape (outChannel, inputChannel/groups, height, width)

bias - bias NDArray of shape (outChannel)

stride - the stride of the deconvolving kernel: Shape(height, width)

padding - implicit paddings on both sides of the input: Shape(height, width)

Returns:

the output of the conv2dTranspose operation

conv2dTranspose

public static NDList conv2dTranspose(NDArray input,
                                     NDArray weight,
                                     NDArray bias,
                                     Shape stride,
                                     Shape padding,
                                     Shape outPadding)

Applies 2D deconvolution over an input signal composed of several input planes.

Parameters:

input - the input NDArray of shape (batchSize, inputChannel, height, width)

weight - filters NDArray of shape (outChannel, inputChannel/groups, height, width)

bias - bias NDArray of shape (outChannel)

stride - the stride of the deconvolving kernel: Shape(height, width)

padding - implicit paddings on both sides of the input: Shape(height, width)

outPadding - Controls the amount of implicit zero-paddings on both sides of the output for outputPadding number of points for each dimension.

Returns:

the output of the conv2dTranspose operation

conv2dTranspose

public static NDList conv2dTranspose(NDArray input,
                                     NDArray weight,
                                     NDArray bias,
                                     Shape stride,
                                     Shape padding,
                                     Shape outPadding,
                                     Shape dilation)

Applies 2D deconvolution over an input signal composed of several input planes.

Parameters:

input - the input NDArray of shape (batchSize, inputChannel, height, width)

weight - filters NDArray of shape (outChannel, inputChannel/groups, height, width)

bias - bias NDArray of shape (outChannel)

stride - the stride of the deconvolving kernel: Shape(height, width)

padding - implicit paddings on both sides of the input: Shape(height, width)

outPadding - Controls the amount of implicit zero-paddings on both sides of the output for outputPadding number of points for each dimension.

dilation - the spacing between kernel elements: Shape(height, width)

Returns:

the output of the conv2dTranspose operation

conv2dTranspose

public static NDList conv2dTranspose(NDArray input,
                                     NDArray weight,
                                     NDArray bias,
                                     Shape stride,
                                     Shape padding,
                                     Shape outPadding,
                                     Shape dilation,
                                     int groups)

Applies 2D deconvolution over an input signal composed of several input planes.

Parameters:

input - the input NDArray of shape (batchSize, inputChannel, height, width)

weight - filters NDArray of shape (outChannel, inputChannel/groups, height, width)

bias - bias NDArray of shape (outChannel)

stride - the stride of the deconvolving kernel: Shape(height, width)

padding - implicit paddings on both sides of the input: Shape(height, width)

outPadding - Controls the amount of implicit zero-paddings on both sides of the output for outputPadding number of points for each dimension. Shape(height, width)

dilation - the spacing between kernel elements: Shape(height, width)

groups - split input into groups: input channel(input.size(1)) should be divisible by the number of groups

Returns:

the output of the conv2dTranspose operation

builder

public static Conv2dTranspose.Builder builder()

Creates a builder to build a Conv2dTranspose.

Returns:

a new builder