Conv2D

class Conv2D(filters: Int, kernelSize: IntArray, strides: IntArray, dilations: IntArray, activation: Activations, kernelInitializer: Initializer, biasInitializer: Initializer, kernelRegularizer: Regularizer?, biasRegularizer: Regularizer?, activityRegularizer: Regularizer?, padding: ConvPadding, useBias: Boolean, name: String) : AbstractConv, TrainableLayer

2D convolution layer (e.g. spatial convolution over images).

This layer creates a convolution kernel that is convolved (actually cross-correlated) with the layer input to produce a tensor of outputs. Finally, the activation is applied to the outputs as well.

It expects input data of size (N, H, W, C) where

N - batch size
H - height
W - width
C - number of channels

Constructors

Conv2D

fun Conv2D(filters: Int = 32, kernelSize: Int = 3, strides: Int = 1, dilations: Int = 1, activation: Activations = Activations.Relu, kernelInitializer: Initializer = HeNormal(), biasInitializer: Initializer = HeUniform(), kernelRegularizer: Regularizer? = null, biasRegularizer: Regularizer? = null, activityRegularizer: Regularizer? = null, padding: ConvPadding = ConvPadding.SAME, useBias: Boolean = true, name: String = "")

Conv2D

fun Conv2D(filters: Int = 32, kernelSize: IntArray = intArrayOf(3, 3), strides: IntArray = intArrayOf(1, 1, 1, 1), dilations: IntArray = intArrayOf(1, 1, 1, 1), activation: Activations = Activations.Relu, kernelInitializer: Initializer = HeNormal(), biasInitializer: Initializer = HeUniform(), kernelRegularizer: Regularizer? = null, biasRegularizer: Regularizer? = null, activityRegularizer: Regularizer? = null, padding: ConvPadding = ConvPadding.SAME, useBias: Boolean = true, name: String = "")

Creates Conv2D object.

Functions

build

open override fun build(tf: Ops, input: Operand<Float>, isTraining: Operand<Boolean>, numberOfLosses: Operand<Float>?): Operand<Float>

open fun build(tf: Ops, input: List<Operand<Float>>, isTraining: Operand<Boolean>, numberOfLosses: Operand<Float>?): Operand<Float>

Extend this function to define variables in the layer and compute layer output.

invoke

operator fun invoke(vararg layers: Layer): Layer

Important part of functional API. It takes layers as input and saves them to the inboundLayers of the given layer.

toString

open override fun toString(): String

Properties

activation

open override val activation: Activations

Activation function.

activityRegularizer

open override val activityRegularizer: Regularizer? = null

Regularizer function applied to the output of the layer (its "activation").

biasInitializer

open override val biasInitializer: Initializer

An initializer for the bias vector.

biasRegularizer

open override val biasRegularizer: Regularizer? = null

Regularizer function applied to the bias vector.

dilations

open override val dilations: IntArray

Four numbers, specifying the dilation rate to use for dilated convolution for each dimension of input tensor.

filters

open override val filters: Int = 32

The dimensionality of the output space (i.e. the number of filters in the convolution).

hasActivation

open override val hasActivation: Boolean

Returns True, if layer has internal activation function.

inboundLayers

var inboundLayers: MutableList<Layer>

Returns inbound layers.

isTrainable

open override var isTrainable: Boolean = true

True, if layer's weights could be changed during training. If false, layer's weights are frozen and could not be changed during training.

kernelInitializer

open override val kernelInitializer: Initializer

An initializer for the convolution kernel

kernelRegularizer

open override val kernelRegularizer: Regularizer? = null

Regularizer function applied to the kernel weights matrix.

kernelSize

open override val kernelSize: IntArray

Two long numbers, specifying the height and width of the 2D convolution window.

name

var name: String

Layer name. A new name is generated during model compilation when provided name is empty.

outboundLayers

var outboundLayers: MutableList<Layer>

Returns outbound layers.

outputShape

lateinit var outputShape: TensorShape

Output data tensor shape.

padding

open override val padding: ConvPadding

The padding method, either 'valid' or 'same' or 'full'.

paramCount

open val paramCount: Int

Number of parameters in this layer.

parentModel

var parentModel: GraphTrainableModel? = null

Model where this layer is used.

strides

open override val strides: IntArray

Four numbers, specifying the strides of the pooling operation for each dimension of input tensor. NOTE: Specifying any stride value != 1 is incompatible with specifying any dilations value != 1.

useBias

open override val useBias: Boolean = true

If true the layer uses a bias vector.

variables

open override val variables: List<KVariable>

Variables used in this layer.