Tensor creation API

Functions
static Derived	empty_like (const Derived &other)
static Derived	zeros_like (const Derived &other)
	Zero tensor like another, i.e. same batch and base shapes, same tensor options, etc.
static Derived	ones_like (const Derived &other)
	Unit tensor like another, i.e. same batch and base shapes, same tensor options, etc.
static Derived	full_like (const Derived &other, Real init)
static Derived	linspace (const Derived &start, const Derived &end, Size nstep, Size dim=0)
	Create a new tensor by adding a new batch dimension with linear spacing between start and end.
static Derived	logspace (const Derived &start, const Derived &end, Size nstep, Size dim=0, Real base=10)
	log-space equivalent of the linspace named constructor

Detailed Description

Function Documentation

empty_like()

template<class Derived>

Derived empty_like ( const Derived & other )

staticnodiscard

Empty tensor like another, i.e. same batch and base shapes, same tensor options, etc.

full_like()

template<class Derived>

Derived full_like ( const Derived & other, Real init )

staticnodiscard

Full tensor like another, i.e. same batch and base shapes, same tensor options, etc., but filled with a different value

linspace()

template<class Derived>

Derived linspace ( const Derived & start, const Derived & end, Size nstep, Size dim = 0 )

staticnodiscard

Create a new tensor by adding a new batch dimension with linear spacing between start and end.

start and end must be broadcastable. The new batch dimension will be added at the user-specified dimension dim which defaults to 0.

For example, if start has shape (3, 2; 5, 5) and end has shape (3, 1; 5, 5), then

linspace(start, end, 100, 1);

will have shape (3, 100, 2; 5, 5), note the location of the new dimension and the broadcasting.

Parameters

start	The starting tensor
end	The ending tensor
nstep	The number of steps with even spacing along the new dimension
dim	Where to insert the new dimension

Returns

Tensor Linearly spaced tensor

logspace()

template<class Derived>

Derived logspace ( const Derived & start, const Derived & end, Size nstep, Size dim = 0, Real base = 10 )

staticnodiscard

log-space equivalent of the linspace named constructor

ones_like()

template<class Derived>

Derived ones_like ( const Derived & other )

staticnodiscard

Unit tensor like another, i.e. same batch and base shapes, same tensor options, etc.

zeros_like()

template<class Derived>

Derived zeros_like ( const Derived & other )

staticnodiscard

Zero tensor like another, i.e. same batch and base shapes, same tensor options, etc.