Dtype Type

Represents a storage type for elements of a tensor

Union cases

Union case Description

BFloat16

Full Usage: BFloat16

Store elements as 16-bit floating point numbers (bfloat16 variation)

Bool

Full Usage: Bool

Store elements as booleans

Byte

Full Usage: Byte

Store elements as 8-bit unsigned integers

Float16

Full Usage: Float16

Store elements as 16-bit floating point numbers

Float32

Full Usage: Float32

Store elements as 32-bit floating point numbers

Float64

Full Usage: Float64

Store elements as 64-bit floating point numbers

Int16

Full Usage: Int16

Store elements as 16-bit signed integers

Int32

Full Usage: Int32

Store elements as 32-bit signed integers

Int64

Full Usage: Int64

Store elements as 64-bit signed integers

Int8

Full Usage: Int8

Store elements as 8-bit integers

Instance members

Instance member Description

this.SummationType

Full Usage: this.SummationType

Returns: Dtype

Gets the natural result of the Sum(), SumToSize() and Sum(dim) operation on this dtype

Returns: Dtype

© Copyright 2021, DiffSharp Contributors.