DScalar

interface DScalar : DTensor

A differentiable scalar (float). We represent a number DScalar as either

a FloatScalar, which is a wrapper around a float, or
a ForwardScalar for forward differentiation with a DScalar primal value, and a DTensor tangent, or
a ReverseScalar for reverse mode differentiation.

Functions

get

open operator fun get(index: Int): DTensor

open operator fun get(vararg indices: Int): DTensor

toCodeString

open fun toCodeString(): String

wrap

open override fun wrap(wrapper: Wrapper): DTensor

Wrapper around the tensor

Properties

derivativeID

abstract val derivativeID: DerivativeID

Each derivative is assigned a unique DerivativeID

indices

open val indices: Iterator<IntArray>

An iterator over the indices for the tensor

isScalar

open val isScalar: Boolean

True if the tensor is a scalar.

operations

abstract val operations: Operations

The operations available on a tensor.

primal

abstract override val primal: DScalar

primal points to the actual tensor

rank

open override val rank: Int

rank is the number of dimension of a tensor. scalars always have a rank of 0

shape

open override val shape: Shape

shape indicates the number of dimensions and the length of the dimensions. Since scalars have rank 0, shape always returns Shape()

size

open val size: Int

The total number of elements of this tensor.

Inheritors

FloatScalar

ForwardScalar

GpuFloatScalar

ReverseScalar

TracingScalar

Extensions

basePrimal

fun DScalar.basePrimal(): FloatScalar

Returns the primal value as a FloatScalar.

compareTo

operator fun DScalar.compareTo(right: Float): Int

operator fun DScalar.compareTo(right: DScalar): Int

div

operator fun DScalar.div(right: Int): DScalar

operator fun DScalar.div(right: DScalar): DScalar

operator fun DScalar.div(right: Float): DScalar

infix fun DScalar.eq(other: DScalar): DScalar

infix fun DScalar.eq(other: Float): DScalar

infix fun DScalar.ge(other: DScalar): DScalar

infix fun DScalar.ge(other: Float): DScalar

infix fun DScalar.gt(other: DScalar): DScalar

infix fun DScalar.gt(other: Float): DScalar

infix fun DScalar.le(other: DScalar): DScalar

infix fun DScalar.le(other: Float): DScalar

infix fun DScalar.lt(other: DScalar): DScalar

infix fun DScalar.lt(other: Float): DScalar

minus

operator fun DScalar.minus(right: Float): DScalar

Scalar (floating-point) subtraction.

operator fun DScalar.minus(right: DScalar): DScalar

infix fun DScalar.ne(other: DScalar): DScalar

infix fun DScalar.ne(other: Float): DScalar

plus

operator fun DScalar.plus(right: Float): DScalar

The addition of a floating-point number to a DScalar simply updates the primal with the scaled value.

operator fun DScalar.plus(right: DScalar): DScalar

Scalar (floating-point) addition.

pow

fun DScalar.pow(x: DScalar): DScalar

fun DScalar.pow(x: Float): DScalar

fun DScalar.pow(x: Int): DScalar

primal

fun DScalar.primal(derivativeID: DerivativeID): DScalar

relu

@JvmName(name = "DScalarRelu")

fun DScalar.relu(): DScalar

times

operator fun DScalar.times(right: Float): DScalar

operator fun DScalar.times(right: DScalar): DScalar

operator fun DScalar.times(right: DTensor): DTensor

unaryMinus

operator fun DScalar.unaryMinus(): DScalar