cytnx::random Namespace Reference

The randomization related functions. More...

Functions
void	normal_ (cytnx::Storage &Sin, const double &mean, const double &std, const unsigned int &seed=__static_random_device())
	Randomize the memory of a Storage with normal distributon.
void	uniform_ (cytnx::Storage &Sin, const double &low=0, const double &high=1, const unsigned int &seed=__static_random_device())
	Randomize the memory of a Storage with uniform distributon.
void	normal_ (cytnx::Tensor &Tin, const double &mean, const double &std, const unsigned int &seed=__static_random_device())
	Randomize the memory of a Tensor with normal distributon.
void	uniform_ (cytnx::Tensor &Tin, const double &low=0, const double &high=1, const unsigned int &seed=__static_random_device())
	Randomize the memory of a Tensor with uniform distributon.
void	normal_ (cytnx::UniTensor &Tin, const double &mean, const double &std, const unsigned int &seed=__static_random_device())
	Randomize the memory of a UniTensor with normal distributon.
void	uniform_ (cytnx::UniTensor &Tin, const double &low=0, const double &high=1, const unsigned int &seed=__static_random_device())
	Randomize the memory of a UniTensor with uniform distributon.
Tensor	normal (const cytnx_uint64 &Nelem, const double &mean, const double &std, const int &device=Device.cpu, const unsigned int &seed=__static_random_device(), const unsigned int &dtype=Type.Double)
	create an Tensor with all the elements are initialized with normal distribution
Tensor	normal (const std::vector< cytnx_uint64 > &Nelem, const double &mean, const double &std, const int &device=Device.cpu, const unsigned int &seed=__static_random_device(), const unsigned int &dtype=Type.Double)
	create an Tensor with all the elements are initialized with normal distribution
Tensor	uniform (const cytnx_uint64 &Nelem, const double &low, const double &high, const int &device=Device.cpu, const unsigned int &seed=__static_random_device(), const unsigned int &dtype=Type.Double)
	create an Tensor with all the elements are initialized with uniform distribution
Tensor	uniform (const std::vector< cytnx_uint64 > &Nelem, const double &low, const double &high, const int &device=Device.cpu, const unsigned int &seed=__static_random_device(), const unsigned int &dtype=Type.Double)
	create an Tensor with all the elements are initialized with uniform distribution

Variables
std::random_device	__static_random_device

Detailed Description

The randomization related functions.

Function Documentation

normal() [1/2]

Tensor cytnx::random::normal	(	const cytnx_uint64 &	Nelem,
		const double &	mean,
		const double &	std,
		const int &	device = Device.cpu,
		const unsigned int &	seed = __static_random_device(),
		const unsigned int &	dtype = Type.Double
	)

create an Tensor with all the elements are initialized with normal distribution

Parameters

[in]	Nelem	the number of elements
[in]	mean	the mean of the normal distribution
[in]	std	the standard deviation \( \sigma \) of a normal distribution.
[in]	device	the device that the Tensor is put on. It can be any device defined in
[in]	seed	the seed for the random generator. [Default] using device entropy
[in]	dtype	the data type of the Tensor. [Default] cytnx::Type.Double cytnx::Device
	seed	the seed for the random generator. [Default] using device entropy

Returns

[Tensor]

normal() [2/2]

Tensor cytnx::random::normal	(	const std::vector< cytnx_uint64 > &	Nelem,
		const double &	mean,
		const double &	std,
		const int &	device = Device.cpu,
		const unsigned int &	seed = __static_random_device(),
		const unsigned int &	dtype = Type.Double
	)

create an Tensor with all the elements are initialized with normal distribution

Parameters

[in]	Nelem	the shape of the Tensor
[in]	mean	the mean of the normal distribution
[in]	std	the standard deviation (sigma) of the normal distribution
[in]	device	the device that the Tensor is put on. It can be any device defined in cytnx::Device
[in]	seed	the seed for the random generator. [Default] using device entropy
[in]	dtype	the data type of the Tensor. [Default] cytnx::Type.Double

Returns

[Tensor]

normal_() [1/3]

void cytnx::random::normal_	(	cytnx::Storage &	Sin,
		const double &	mean,
		const double &	std,
		const unsigned int &	seed = __static_random_device()
	)

Randomize the memory of a Storage with normal distributon.

Parameters

[in]	Sin	a Storage
[in]	mean	the mean of a normal distribution
[in]	std	the standard deviation \( \sigma \) of a normal distribution.
[in]	seed	the seed for the random generator. [Default] Using device entropy.

Note

The Storage sould be real floating type or complex type.

normal_() [2/3]

void cytnx::random::normal_	(	cytnx::Tensor &	Tin,
		const double &	mean,
		const double &	std,
		const unsigned int &	seed = __static_random_device()
	)

Randomize the memory of a Tensor with normal distributon.

Parameters

[in]	Tin	a Tensor
[in]	mean	the mean of a normal distribution
[in]	std	the standard deviation \( \sigma \) of a normal distribution.
[in]	seed	the seed for the random generator. [Default] Using device entropy.

Note

The Tensor sould be real floating type or complex type.

normal_() [3/3]

void cytnx::random::normal_	(	cytnx::UniTensor &	Tin,
		const double &	mean,
		const double &	std,
		const unsigned int &	seed = __static_random_device()
	)

Randomize the memory of a UniTensor with normal distributon.

Parameters

[in]	Tin	a UniTensor
[in]	mean	the mean of a normal distribution
[in]	std	the standard deviation \( \sigma \) of a normal distribution.
[in]	seed	the seed for the random generator. [Default] Using device entropy.

Note

The UniTensor sould be real floating type or complex type.

uniform() [1/2]

Tensor cytnx::random::uniform	(	const cytnx_uint64 &	Nelem,
		const double &	low,
		const double &	high,
		const int &	device = Device.cpu,
		const unsigned int &	seed = __static_random_device(),
		const unsigned int &	dtype = Type.Double
	)

create an Tensor with all the elements are initialized with uniform distribution

Parameters

[in]	Nelem	the number of elements
[in]	low	the lower-bound of range
[in]	high	the upper-bound of range
[in]	device	the device that the Tensor is put on. It can be any device defined in cytnx::Device
[in]	seed	the seed for the random generator. [Default] using device entropy
[in]	dtype	the data type of the Tensor. [Default] cytnx::Type.Double

Returns

[Tensor]

uniform() [2/2]

Tensor cytnx::random::uniform	(	const std::vector< cytnx_uint64 > &	Nelem,
		const double &	low,
		const double &	high,
		const int &	device = Device.cpu,
		const unsigned int &	seed = __static_random_device(),
		const unsigned int &	dtype = Type.Double
	)

create an Tensor with all the elements are initialized with uniform distribution

Parameters

[in]	Nelem	the number of elements
[in]	low	the lower-bound of range
[in]	high	the upper-bound of range
[in]	device	the device that the Tensor is put on. It can be any device defined in cytnx::Device
[in]	seed	the seed for the random generator. [Default] using device entropy
[in]	dtype	the data type of the Tensor. [Default] cytnx::Type.Double

Returns

[Tensor]

uniform_() [1/3]

void cytnx::random::uniform_	(	cytnx::Storage &	Sin,
		const double &	low = 0,
		const double &	high = 1,
		const unsigned int &	seed = __static_random_device()
	)

Randomize the memory of a Storage with uniform distributon.

Parameters

[in]	Sin	a Storage
[in]	low	the lower-bound of the uniform distribution
[in]	high	the higher-bound of the uniform distribution
[in]	seed	the seed for the random generator. [Default] Using device entropy.

Note

1. The Storage sould be real floating type or complex type.
2. In cpu, it generate random number in domain [low,high); but in gpu(cuda), it generate random number in domain (low,high]; (cuRandv10)

uniform_() [2/3]

void cytnx::random::uniform_	(	cytnx::Tensor &	Tin,
		const double &	low = 0,
		const double &	high = 1,
		const unsigned int &	seed = __static_random_device()
	)

Randomize the memory of a Tensor with uniform distributon.

Parameters

[in]	Tin	a Tensor
[in]	low	the lower-bound of the uniform distribution
[in]	high	the higher-bound of the uniform distribution
[in]	seed	the seed for the random generator. [Default] Using device entropy.

Note

1. The Tensor sould be real floating type or complex type.
2. In cpu, it generate random number in domain [low,high); but in gpu(cuda), it generate random number in domain (low,high]; (cuRandv10)

uniform_() [3/3]

void cytnx::random::uniform_	(	cytnx::UniTensor &	Tin,
		const double &	low = 0,
		const double &	high = 1,
		const unsigned int &	seed = __static_random_device()
	)

Randomize the memory of a UniTensor with uniform distributon.

Parameters

[in]	Tin	a UniTensor
[in]	low	the lower-bound of the uniform distribution
[in]	high	the higher-bound of the uniform distribution
[in]	seed	the seed for the random generator. [Default] Using device entropy.

Note

1. The UniTensor sould be real floating type or complex type.
2. In cpu, it generate random number in domain [low,high); but in gpu(cuda), it generate random number in domain (low,high]; (cuRandv10)

Variable Documentation

__static_random_device

std::random_device cytnx::random::__static_random_device

extern