Initializer

abstract class Initializer

Initializer base class: all initializers inherit this class.

Initializers allow you to pre-specify an initialization strategy, encoded in the Initializer object, without knowing the shape and dtype of the variable being initialized.

Constructors

Initializer
Link copied to clipboard
fun Initializer()

Functions

apply
Link copied to clipboard
fun apply(fanIn: Int, fanOut: Int, tf: Ops, input: Operand<Float>, name: String): Assign<Float>

Adds an Assign Op to the graph to initialize a tensorflow variable as specified by the initializer.

initialize
Link copied to clipboard
abstract fun initialize(fanIn: Int, fanOut: Int, tf: Ops, shape: Operand<Int>, name: String): Operand<Float>

Returns a Tensor object initialized as specified by the initializer.

Inheritors

Constant
Link copied to clipboard
Identity
Link copied to clipboard
Ones
Link copied to clipboard
Orthogonal
Link copied to clipboard
ParametrizedTruncatedNormal
Link copied to clipboard
RandomNormal
Link copied to clipboard
RandomUniform
Link copied to clipboard
TruncatedNormal
Link copied to clipboard
VarianceScaling
Link copied to clipboard
Zeros
Link copied to clipboard