Cytnx v0.9.1
Loading...
Searching...
No Matches
Public Member Functions | Static Public Member Functions | List of all members
cytnx::Tensor Class Reference

an tensor (multi-dimensional array) More...

#include <Tensor.hpp>

Public Member Functions

void Save (const std::string &fname) const
 Save current Tensor to file.
 
void Save (const char *fname) const
 
void Tofile (const std::string &fname) const
 
void Tofile (const char *fname) const
 
void Tofile (std::fstream &f) const
 
void Init (const std::vector< cytnx_uint64 > &shape, const unsigned int &dtype=Type.Double, const int &device=-1, const bool &init_zero=true)
 initialize a Tensor
 
 Tensor (const std::vector< cytnx_uint64 > &shape, const unsigned int &dtype=Type.Double, const int &device=-1, const bool &init_zero=1)
 Construct a new Tensor object.
 
unsigned int dtype () const
 the dtype-id of the Tensor
 
int device () const
 the device-id of the Tensor
 
std::string dtype_str () const
 the dtype (in string) of the Tensor
 
std::string device_str () const
 the device (in string) of the Tensor
 
const std::vector< cytnx_uint64 > & shape () const
 the shape of the Tensor
 
cytnx_uint64 rank () const
 the rank of the Tensor
 
Tensor clone () const
 return a clone of the current Tensor.
 
Tensor to (const int &device) const
 copy a tensor to new device
 
void to_ (const int &device)
 move the current Tensor to the device.
 
const bool & is_contiguous () const
 
Tensor permute_ (const std::vector< cytnx_uint64 > &rnks)
 
Tensor permute (const std::vector< cytnx_uint64 > &rnks) const
 perform tensor permute on the cytnx::Tensor and return a new instance.
 
Tensor contiguous () const
 Make the Tensor contiguous by coalescing the memory (storage).
 
Tensor contiguous_ ()
 Make the Tensor contiguous by coalescing the memory (storage), inplacely.
 
void reshape_ (const std::vector< cytnx_int64 > &new_shape)
 reshape the Tensor, inplacely
 
Tensor reshape (const std::vector< cytnx_int64 > &new_shape) const
 return a new Tensor that is reshaped.
 
Tensor reshape (const std::vector< cytnx_uint64 > &new_shape) const
 
Tensor reshape (const std::initializer_list< cytnx_int64 > &new_shape) const
 
Tensor astype (const int &new_type) const
 return a new Tensor that cast to different dtype.
 
template<class T >
T & at (const std::vector< cytnx_uint64 > &locator)
 [C++ only] get an element at specific location.
 
template<class T >
const T & at (const std::vector< cytnx_uint64 > &locator) const
 
template<class T >
T & item ()
 get an from a rank-0 Tensor
 
Tensor get (const std::vector< cytnx::Accessor > &accessors) const
 get elements using Accessor (C++ API) / slices (python API)
 
void set (const std::vector< cytnx::Accessor > &accessors, const Tensor &rhs)
 set elements with the input Tensor using Accessor (C++ API) / slices (python API)
 
template<class T >
void set (const std::vector< cytnx::Accessor > &accessors, const T &rc)
 set elements with the input constant using Accessor (C++ API) / slices (python API)
 
Storage & storage () const
 return the storage of current Tensor.
 
template<class T >
void fill (const T &val)
 fill all the element of current Tensor with the value.
 
bool equiv (const Tensor &rhs)
 compare the shape of two tensors.
 
Tensor real ()
 return the real part of the tensor.
 
Tensor imag ()
 return the imaginary part of the tensor.
 
template<class T >
Tensor & operator+= (const T &rc)
 addition assignment operator with a Tensor or a scalar.
 
template<class T >
Tensor & operator-= (const T &rc)
 subtraction assignment operator with a Tensor or a scalar.
 
template<class T >
Tensor & operator*= (const T &rc)
 multiplication assignment operator with a Tensor or a scalar.
 
template<class T >
Tensor & operator/= (const T &rc)
 division assignment operator with a Tensor or a scalar.
 
template<class T >
Tensor Add (const T &rhs)
 Addition function with a Tensor or a scalar. Same as cytnx::operator+(const Tensor &self, const T &rhs).
 
template<class T >
Tensor & Add_ (const T &rhs)
 Addition function with a Tensor or a scalar, inplacely. Same as operator+=(const T &rhs).
 
template<class T >
Tensor Sub (const T &rhs)
 Subtraction function with a Tensor or a scalar. Same as cytnx::operator-(const Tensor &self, const T &rhs).
 
template<class T >
Tensor & Sub_ (const T &rhs)
 Subtraction function with a Tensor or a scalar, inplacely. Same as operator-=(const T &rhs).
 
template<class T >
Tensor Mul (const T &rhs)
 Multiplication function with a Tensor or a scalar. Same as cytnx::operator*(const Tensor &self, const T &rhs).
 
template<class T >
Tensor & Mul_ (const T &rhs)
 Multiplication function with a Tensor or a scalar, inplacely. Same as operator*=(const T &rhs).
 
template<class T >
Tensor Div (const T &rhs)
 Division function with a Tensor or a scalar. Same as cytnx::operator/(const Tensor &self, const T &rhs).
 
template<class T >
Tensor & Div_ (const T &rhs)
 Division function with a Tensor or a scalar, inplacely. Same as operator/=(const T &rhs).
 
template<class T >
Tensor Cpr (const T &rhs)
 The comparison function.
 
template<class T >
Tensor Mod (const T &rhs)
 
Tensor operator- ()
 The negation function.
 
Tensor flatten () const
 The flatten function.
 
void flatten_ ()
 The flatten function, inplacely.
 
void append (const Tensor &rhs)
 the append function.
 
void append (const Storage &srhs)
 the append function of the Storage.
 
template<class T >
void append (const T &rhs)
 the append function of the scalar.
 
bool same_data (const Tensor &rhs) const
 
std::vector< Tensor > Svd (const bool &is_UvT=true) const
 the SVD member function. Same as cytnx::linalg::Svd(const Tensor &Tin, const bool &is_UvT) , where Tin is the current Tensor.
 
std::vector< Tensor > Eigh (const bool &is_V=true, const bool &row_v=false) const
 the Eigh member function. Same as cytnx::linalg::Eigh(const Tensor &Tin, const bool &is_V, const bool &row_v) , where Tin is the current Tensor.
 
Tensor & InvM_ ()
 the InvM_ member function. Same as cytnx::linalg::InvM_(Tensor &Tin), where Tin is the current Tensor.
 
Tensor InvM () const
 the InvM member function. Same as cytnx::linalg::InvM(const Tensor &Tin), where Tin is the current Tensor.
 
Tensor & Inv_ (const double &clip)
 the Inv_ member function. Same as cytnx::linalg::Inv_(Tensor &Tin, const double &clip)
 
Tensor Inv (const double &clip) const
 the Inv member function. Same as cytnx::linalg::Inv(const Tensor &Tin, const double &clip)
 
Tensor & Conj_ ()
 the Conj_ member function. Same as cytnx::linalg::Conj_(Tensor &Tin), where Tin is the current Tensor.
 
Tensor Conj () const
 the Conj member function. Same as cytnx::linalg::Conj(const Tensor &Tin), where Tin is the current Tensor.
 
Tensor & Exp_ ()
 the Exp_ member function. Same as linalg::Exp_(Tensor &Tin), where Tin is the current Tensor.
 
Tensor Exp () const
 the Exp member function. Same as linalg::Exp(const Tensor &Tin), where Tin is the current Tensor.
 
Tensor Norm () const
 the Norm member function. Same as linalg::Norm(const Tensor &Tin), where Tin is the current Tensor.
 
Tensor Pow (const cytnx_double &p) const
 the Pow member function. Same as linalg::Pow(const Tensor &Tin, const cytnx_double &p), where Tin is the current Tensor.
 
Tensor & Pow_ (const cytnx_double &p)
 the Pow_ member function. Same as linalg::Pow_(Tensor &Tin, const cytnx_double &p), where Tin is the current Tensor.
 
Tensor Trace (const cytnx_uint64 &a=0, const cytnx_uint64 &b=1) const
 the Trace member function. Same as linalg::Trace(const Tensor &Tin, const cytnx_uint64 &a, const cytnx_uint64 &b), where Tin is the current Tensor.
 
Tensor Abs () const
 the Abs member function. Same as linalg::Abs(const Tensor &Tin), where Tin is the current Tensor.
 
Tensor & Abs_ ()
 the Abs_ member function. Same as linalg::Abs_(Tensor &Tin), where Tin is the current Tensor.
 
Tensor Max () const
 the Max member function. Same as linalg::Max(const Tensor &Tin), where Tin is the current Tensor.
 
Tensor Min () const
 the Min member function. Same as linalg::Min(const Tensor &Tin), where Tin is the current Tensor.
 

Static Public Member Functions

static Tensor Load (const std::string &fname)
 Load current Tensor from file.
 
static Tensor Load (const char *fname)
 
static Tensor Fromfile (const std::string &fname, const unsigned int &dtype, const cytnx_int64 &count=-1)
 
static Tensor Fromfile (const char *fname, const unsigned int &dtype, const cytnx_int64 &count=-1)
 
static Tensor from_storage (const Storage &in)
 

Detailed Description

an tensor (multi-dimensional array)

Constructor & Destructor Documentation

◆ Tensor()

cytnx::Tensor::Tensor ( const std::vector< cytnx_uint64 > &  shape,
const unsigned int &  dtype = Type.Double,
const int &  device = -1,
const bool &  init_zero = 1 
)
inline

Construct a new Tensor object.

This is the constructor of Tensor. It will call cytnx::Tensor::Init() to initialize the Tensor.

Parameters
[in]shapethe shape of tensor
[in]dtypethe dtype of tensor. This can be any of type defined in cytnx::Type.
[in]devicethe device that tensor to be created. This can be cytnx::Device.cpu or cytnx::Device.cuda+<gpuid>, see cytnx::Device for more detail.
[in]init_zeroif true, the content of Tensor will be initialized to zero. If false, the content of Tensor will be un-initialized.
See also
cytnx::Tensor::Init

Member Function Documentation

◆ Abs()

Tensor cytnx::Tensor::Abs ( ) const

the Abs member function. Same as linalg::Abs(const Tensor &Tin), where Tin is the current Tensor.

◆ Abs_()

Tensor & cytnx::Tensor::Abs_ ( )

the Abs_ member function. Same as linalg::Abs_(Tensor &Tin), where Tin is the current Tensor.

◆ Add()

template<class T >
Tensor cytnx::Tensor::Add ( const T &  rhs)
inline

Addition function with a Tensor or a scalar. Same as cytnx::operator+(const Tensor &self, const T &rhs).

Parameters
[in]rhsthe added Tensor or scalar.

◆ Add_()

template<class T >
Tensor & cytnx::Tensor::Add_ ( const T &  rhs)
inline

Addition function with a Tensor or a scalar, inplacely. Same as operator+=(const T &rhs).

Parameters
[in]rhsthe added Tensor or scalar.

◆ append() [1/3]

void cytnx::Tensor::append ( const Storage &  srhs)
inline

the append function of the Storage.

This function is the append function of the Storage. It will same as the append(const Tensor &rhs) function, but the rhs is a Storage.

Parameters
[in]srhsthe appended Storage.
Returns
The appended tensor.
See also
append(const Tensor &rhs)

◆ append() [2/3]

template<class T >
void cytnx::Tensor::append ( const T &  rhs)
inline

the append function of the scalar.

This function is the append function of the scalar. It can only append scalar into rank-1 Tensor.

Parameters
[in]rhsthe appended scalar.
Returns
The appended tensor.
Precondition
  1. The current Tensor must be rank-1. (1D array)
  2. The current Tensor must be contiguous.
  3. rhs must be a scalar.

◆ append() [3/3]

void cytnx::Tensor::append ( const Tensor &  rhs)
inline

the append function.

This function is the append function. It will append the rhs tensor to the current tensor. The rhs tensor must have the same shape as the current tensor, except the first dimension. For example, if the current tensor is \(A(i,j,k)\) and the rhs tensor is \(B(j,k)\), then the output tensor is \(C(i,j,k)\) where

\[ C(i,j,k) = \begin{cases} A(i,j,k) & \text{if } i \neq N \\ B(j,k) & \text{if } i = N \end{cases} \]

where \(N\) is the number of the first dimension of the current tensor.

Parameters
[in]rhsthe appended tensor.
Returns
The appended tensor.
Precondition
  1. The rhs tensor and the current tensor cannot be empty.
  2. The rhs tensor must have the same shape as the current tensor, except the first dimension. Namely, rhs.shape()[i] == this->shape()[i+1] and rhs.shape().size() == this->shape().size()-1.
Note
If the dtype of the rhs is different from the current tensor, the rhs will be casted to the dtype of the current tensor.
See also
append(const Storage &rhs)

◆ astype()

Tensor cytnx::Tensor::astype ( const int &  new_type) const
inline

return a new Tensor that cast to different dtype.

Parameters
[in]new_typethe new dtype. It can be any type defined in cytnx::Type
Returns
[Tensor]
Note
If the new_type is the same as dtype of the current Tensor, return self.

Example:

c++ API:

#include "cytnx.hpp"
#include <iostream>
using namespace cytnx;
using namespace std;
int main() {
Tensor A = zeros({3, 4, 5}, Type.Double);
cout << A;
Tensor B = A.astype(Type.Uint64);
cout << B;
// the new type is the same as current dtype, return self.
Tensor C = A.astype(Type.Double);
cout << is(C, A) << endl; // this should be true.
return 0;
}
cytnx::Tensor
an tensor (multi-dimensional array)
Definition Tensor.hpp:345
cytnx::Tensor::astype
Tensor astype(const int &new_type) const
return a new Tensor that cast to different dtype.
Definition Tensor.hpp:1068
cytnx
Definition Accessor.hpp:12
cytnx::zeros
Tensor zeros(const cytnx_uint64 &Nelem, const unsigned int &dtype=Type.Double, const int &device=Device.cpu)
create an rank-1 Tensor with all the elements are initialized with zero.
Definition Generator.cpp:10
cytnx::Type
Type_class Type
data type
Definition Type.cpp:23

output>

Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]
 [[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]
 [[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]]


Total elem: 60
type  : Uint64
cytnx device: CPU
Shape : (3,4,5)
[[[                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]]
 [[                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]]
 [[                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]]]

1

python API:

import sys
from pathlib import Path
home = str(Path.home())
sys.path.append(home + '/Cytnx_lib')
from cytnx import *
A = zeros([3,4,5],dtype=Type.Double)
print(A)
B = A.astype(Type.Uint64)
print(B)
C = A.astype(Type.Double)
print(C is A)

output>

Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]
 [[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]
 [[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]]




Total elem: 60
type  : Uint64
cytnx device: CPU
Shape : (3,4,5)
[[[                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]]
 [[                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]]
 [[                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]]]



True

◆ at() [1/2]

template<class T >
T & cytnx::Tensor::at ( const std::vector< cytnx_uint64 > &  locator)
inline

[C++ only] get an element at specific location.

Parameters
[in]locatorthe location of the element
Returns
[ref]
Note
  1. This is for C++ API only!
  2. need template instantiation to resolve the type, which should be consist with the dtype of the Tensor. An error will be issued if the template type is inconsist with the current dtype of Tensor.
  3. For python API, use [] directly to get element.

Example:

c++ API:

#include "cytnx.hpp"
#include <iostream>
using namespace cytnx;
using namespace std;
int main() {
Tensor A = arange(30, Type.Float).reshape(2, 3, 5);
cout << A << endl;
// note that type resolver should be consist with the dtype
cout << A.at<cytnx_float>(0, 0, 2) << endl;
// the return is a ref., can be modify directly.
A.at<cytnx_float>(0, 0, 2) = 999;
cout << A.at<cytnx_float>(0, 0, 2) << endl;
// [Note] there are two way to give argument:
// Method 1: more like 'c++' way:
// (alternatively, you can also simply give a std::vector)
A.at<cytnx_float>({0, 0, 2}); // note the braket{}
// Method 2: more like 'python' way:
A.at<cytnx_float>(0, 0, 2);
return 0;
}
cytnx::Tensor::at
T & at(const std::vector< cytnx_uint64 > &locator)
[C++ only] get an element at specific location.
Definition Tensor.hpp:1101
cytnx::Tensor::reshape
Tensor reshape(const std::vector< cytnx_int64 > &new_shape) const
return a new Tensor that is reshaped.
Definition Tensor.hpp:1019
cytnx::cytnx_float
float cytnx_float
Definition Type.hpp:44
cytnx::arange
Tensor arange(const cytnx_int64 &Nelem)
create an rank-1 Tensor with incremental unsigned integer elements start with [0,Nelem)
Definition Generator.cpp:75

output>

 






◆ at() [2/2]








template<class T > 



  

  
      

        

          const T & cytnx::Tensor::at 
          (
          const std::vector< cytnx_uint64 > & 
          locator)
           const
        

      

  		
  
inline  
  






See also
at(const std::vector<cytnx_uint64> &locator) 








◆ clone()








  

  
      

        

          Tensor cytnx::Tensor::clone 
          (
          )
           const
        

      

  		
  
inline  
  







return a clone of the current Tensor. 


Returns
[Tensor]


In C++ API, the behavior of assignment operator is designed to have same behavior as python,

 to have a copy of the current tensor, we call clone to return a copy.


Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  Tensor A({3, 4, 5});


 


  Tensor B = A;  // B shares same object with A


  Tensor C = A.clone();  // C is a copy of A


 


  // use is() to check if two variable shares same object


  cout << is(B, A) << endl;


  cout << is(C, A) << endl;


 


  return 0;


}


cytnx::Tensor::clone
Tensor clone() const
return a clone of the current Tensor.
Definition Tensor.hpp:820


 
output>


1
0

 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = Tensor([3,4,5])


 


B = A


C = A.clone()


 


print(B is A)


print(C is A)


 


 
output>


True
False

 






◆ Conj()






      

        

          Tensor cytnx::Tensor::Conj 
          (
          )
           const
        

      





the Conj member function. Same as cytnx::linalg::Conj(const Tensor &Tin), where Tin is the current Tensor. 








◆ Conj_()






      

        

          Tensor & cytnx::Tensor::Conj_ 
          (
          )
          
        

      





the Conj_ member function. Same as cytnx::linalg::Conj_(Tensor &Tin), where Tin is the current Tensor. 








◆ contiguous()








  

  
      

        

          Tensor cytnx::Tensor::contiguous 
          (
          )
           const
        

      

  		
  
inline  
  







Make the Tensor contiguous by coalescing the memory (storage). 


Returns
[Tensor] a new Tensor that is with contiguous memory (storage). 


See also
Tensor::contiguous_() 


Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  Tensor A({3, 4, 5});


  cout << A.shape() << endl;


 


  Tensor B = A.permute({0, 2, 1});


  cout << B.shape() << endl;


 


  //[Note] permute will not actually move the internal memory (storage) layout.


  //       this is called non-contiguous status.


  //       the memory layout will only move when Tensor.contiguous() is called.


  Tensor C = B.contiguous();  // actual moving the memory


  cout << B.is_contiguous() << endl;  // false.


  cout << C.is_contiguous() << endl;  // true.


  cout << C.shape() << endl;


 


  cout << C.same_data(B) << endl;  // false


  cout << B.same_data(A) << endl;  // true


 


  return 0;


}


cytnx::Tensor::same_data
bool same_data(const Tensor &rhs) const
Definition Tensor.cpp:1358


cytnx::Tensor::contiguous
Tensor contiguous() const
Make the Tensor contiguous by coalescing the memory (storage).
Definition Tensor.hpp:932


cytnx::Tensor::is_contiguous
const bool & is_contiguous() const
Definition Tensor.hpp:871


cytnx::Tensor::permute
Tensor permute(const std::vector< cytnx_uint64 > &rnks) const
perform tensor permute on the cytnx::Tensor and return a new instance.
Definition Tensor.hpp:904


cytnx::Tensor::shape
const std::vector< cytnx_uint64 > & shape() const
the shape of the Tensor
Definition Tensor.hpp:795


 
output>


Vector Print:
Total Elements:3
[3, 4, 5]

Vector Print:
Total Elements:3
[3, 5, 4]

0
1
Vector Print:
Total Elements:3
[3, 5, 4]

0
1

 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = Tensor([3,4,5])


print(A.shape())


 


B = A.permute(0,2,1)


print(B.shape())


 


C = B.contiguous()


print(B.is_contiguous()) #false


print(C.is_contiguous()) #true


print(C.shape())


 


 
output>


[3, 4, 5]
[3, 5, 4]
False
True
[3, 5, 4]

 






◆ contiguous_()








  

  
      

        

          Tensor cytnx::Tensor::contiguous_ 
          (
          )
          
        

      

  		
  
inline  
  







Make the Tensor contiguous by coalescing the memory (storage), inplacely. 


See also
Tensor::contiguous() 


Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  Tensor A({3, 4, 5});


  cout << A.shape() << endl;


 


  Tensor B = A.permute({0, 2, 1});


  cout << B.shape() << endl;


 


  //[Note] permute will not actually move the internal memory (storage) layout.


  //       this is called non-contiguous status.


  //       the memory layout will only move when Tensor.contiguous() is called.


  cout << B.is_contiguous() << endl;  // false.


  B.contiguous_();  // actual moving the memory


  cout << B.is_contiguous() << endl;  // true.


 


  return 0;


}


cytnx::Tensor::contiguous_
Tensor contiguous_()
Make the Tensor contiguous by coalescing the memory (storage), inplacely.
Definition Tensor.hpp:952


 
output>


Vector Print:
Total Elements:3
[3, 4, 5]

Vector Print:
Total Elements:3
[3, 5, 4]

0
1

 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = Tensor([3,4,5])


print(A.shape())


 


B = A.permute(0,2,1)


print(B.shape())


 


print(B.is_contiguous()) #false


B.contiguous_()


print(B.is_contiguous()) #true


    


 


 
output>


[3, 4, 5]
[3, 5, 4]
False
True

 






◆ Cpr()








template<class T > 



  

  
      

        

          Tensor cytnx::Tensor::Cpr 
          (
          const T & 
          rhs)
          
        

      

  		
  
inline  
  







The comparison function. 


This function is the comparison function. Same as cytnx::operator==(const Tensor &self, const T &rhs). 
Parameters

  

    
[in]rhsthe compared object. 

  

  










◆ device()








  

  
      

        

          int cytnx::Tensor::device 
          (
          )
           const
        

      

  		
  
inline  
  







the device-id of the Tensor 


See also
cytnx::Device 


Returns
[int] the device_id of the Tensor 








◆ device_str()








  

  
      

        

          std::string cytnx::Tensor::device_str 
          (
          )
           const
        

      

  		
  
inline  
  







the device (in string) of the Tensor 


See also
cytnx::Device, device() const 


Returns
[std::string] the device of the Tensor 








◆ Div()








template<class T > 



  

  
      

        

          Tensor cytnx::Tensor::Div 
          (
          const T & 
          rhs)
          
        

      

  		
  
inline  
  







Division function with a Tensor or a scalar. Same as cytnx::operator/(const Tensor &self, const T &rhs). 


Parameters

  

    
[in]rhsthe divided Tensor or scalar. @attension rhs cannot be zero. 

  

  










◆ Div_()








template<class T > 



  

  
      

        

          Tensor & cytnx::Tensor::Div_ 
          (
          const T & 
          rhs)
          
        

      

  		
  
inline  
  







Division function with a Tensor or a scalar, inplacely. Same as operator/=(const T &rhs). 


Parameters

  

    
[in]rhsthe divided Tensor or scalar. @attension rhs cannot be zero. 

  

  










◆ dtype()








  

  
      

        

          unsigned int cytnx::Tensor::dtype 
          (
          )
           const
        

      

  		
  
inline  
  







the dtype-id of the Tensor 


See also
cytnx::Type 


Returns
[unsigned int] the dtype_id of the Tensor 








◆ dtype_str()








  

  
      

        

          std::string cytnx::Tensor::dtype_str 
          (
          )
           const
        

      

  		
  
inline  
  







the dtype (in string) of the Tensor 


See also
cytnx::Type, dtype() const 


Returns
[std::string] the dtype of the Tensor 








◆ Eigh()






      

        

          std::vector< Tensor > cytnx::Tensor::Eigh 
          (
          const bool & 
          is_V = true, 
        

        

          
          
          const bool & 
          row_v = false 
        

        

          
          )
           const
        

      





the Eigh member function. Same as cytnx::linalg::Eigh(const Tensor &Tin, const bool &is_V, const bool &row_v) , where Tin is the current Tensor. 








◆ equiv()








  

  
      

        

          bool cytnx::Tensor::equiv 
          (
          const Tensor & 
          rhs)
          
        

      

  		
  
inline  
  







compare the shape of two tensors. 


Parameters

  

    
[in]rhsthe tensor to be compared. 

  

  










◆ Exp()






      

        

          Tensor cytnx::Tensor::Exp 
          (
          )
           const
        

      





the Exp member function. Same as linalg::Exp(const Tensor &Tin), where Tin is the current Tensor. 








◆ Exp_()






      

        

          Tensor & cytnx::Tensor::Exp_ 
          (
          )
          
        

      





the Exp_ member function. Same as linalg::Exp_(Tensor &Tin), where Tin is the current Tensor. 








◆ fill()








template<class T > 



  

  
      

        

          void cytnx::Tensor::fill 
          (
          const T & 
          val)
          
        

      

  		
  
inline  
  







fill all the element of current Tensor with the value. 


Parameters

  

    
[in]valthe assigned value

  

  




Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  Tensor A = arange(60).reshape({3, 4, 5});


  cout << A << endl;


 


  A.fill(999);


  cout << A << endl;


 


  return 0;


}


cytnx::Tensor::fill
void fill(const T &val)
fill all the element of current Tensor with the value.
Definition Tensor.hpp:1294


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 ]
  [4.50000e+01 4.60000e+01 4.70000e+01 4.80000e+01 4.90000e+01 ]
  [5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 ]
  [5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]]]



Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]]
 [[9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]]
 [[9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]]]



 
python API


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = arange(60).reshape(3,4,5)


print(A)


 


A.fill(999)


print(A)


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 ]
  [4.50000e+01 4.60000e+01 4.70000e+01 4.80000e+01 4.90000e+01 ]
  [5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 ]
  [5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]]]




Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]]
 [[9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]]
 [[9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 9.99000e+02 ]]]




 






◆ flatten()








  

  
      

        

          Tensor cytnx::Tensor::flatten 
          (
          )
           const
        

      

  		
  
inline  
  







The flatten function. 


This function is the flatten function. It will clone (deep copy) , contiguos the current tensor and reshape it to undetermined rank. 
Note
compare to the flatten_() function, this function will return a new tensor and the current tensor will not be changed. 








◆ flatten_()








  

  
      

        

          void cytnx::Tensor::flatten_ 
          (
          )
          
        

      

  		
  
inline  
  







The flatten function, inplacely. 


This function is the flatten function, inplacely. It will contiguos the current tensor and reshape it to undetermined rank. 
Note
compare to the flatten() function, this is an inplacely function, the current tensor will be changed. 








◆ from_storage()








  

  
      

        

          static Tensor cytnx::Tensor::from_storage 
          (
          const Storage & 
          in)
          
        

      

  		
  
inlinestatic  
  












◆ Fromfile() [1/2]








  

  
      

        

          Tensor cytnx::Tensor::Fromfile 
          (
          const char * 
          fname, 
        

        

          
          
          const unsigned int & 
          dtype, 
        

        

          
          
          const cytnx_int64 & 
          count = -1 
        

        

          
          )
          
        

      

  		
  
static  
  












◆ Fromfile() [2/2]








  

  
      

        

          Tensor cytnx::Tensor::Fromfile 
          (
          const std::string & 
          fname, 
        

        

          
          
          const unsigned int & 
          dtype, 
        

        

          
          
          const cytnx_int64 & 
          count = -1 
        

        

          
          )
          
        

      

  		
  
static  
  












◆ get()








  

  
      

        

          Tensor cytnx::Tensor::get 
          (
          const std::vector< cytnx::Accessor > & 
          accessors)
           const
        

      

  		
  
inline  
  







get elements using Accessor (C++ API) / slices (python API) 


Parameters

  

    
[in]accessorsthe Accessor (C++ API) / slices (python API) to get the elements. 

  

  




Returns
[Tensor] 


See also
Accessor for cordinate with Accessor in C++ API. 


Note
    

  1. the return will be a new Tensor instance, which not share memory with the current Tensor.
    2. 





Equivalently:


One can also using more intruisive way to get the slice using [] operator.


Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  typedef Accessor ac;


 


  /*


      In C++ API, generally, there are two ways you can access Tensor.


      1. Using operator():


          This is more of 'python' way.


 


      2. Using low-level API or operator[]:


          This is more of 'c++' way.


  */


 


  Tensor A = arange(60).reshape({3, 4, 5});


  cout << A << endl;


 


  // Method 1, Using operator():


  //-----------------------------------------


  Tensor B = A(2, ":", "2:5:1");


  cout << B << endl;


 


  /* [Note]


      This is equivalent as:


      > Tensor B = A(2,ac::all(),ac::range(2,5,1));


 


      See also cytnx::Accessor.


  */


 


  // Method 2, Using operator[] or low-level API get():


  //----------------------------------------


  Tensor B2 = A[{ac(2), ac::all(), ac::range(2, 5, 1)}];  // remember the {}braket


  cout << B2 << endl;


 


  /* [Note]


      You can also use the low-level API get() as


      > Tensor B2 = A.get({ac(2),ac::all(),ac::range(2,5,1)});


  */


  return 0;


}


cytnx::Accessor
object that mimic the python slice to access elements in C++ [this is for c++ API only].
Definition Accessor.hpp:16


cytnx::Accessor::all
static Accessor all()
access the whole rank, this is similar to [:] in python
Definition Accessor.hpp:104


cytnx::Accessor::range
static Accessor range(const cytnx_int64 &min, const cytnx_int64 &max, const cytnx_int64 &step=1)
access the range at assigned rank, this is similar to min:max:step in python
Definition Accessor.hpp:123


cytnx::ac
Accessor ac
Definition BlockUniTensor.cpp:20


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 ]
  [4.50000e+01 4.60000e+01 4.70000e+01 4.80000e+01 4.90000e+01 ]
  [5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 ]
  [5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]]]



Total elem: 12
type  : Double (Float64)
cytnx device: CPU
Shape : (4,3)
[[4.20000e+01 4.30000e+01 4.40000e+01 ]
 [4.70000e+01 4.80000e+01 4.90000e+01 ]
 [5.20000e+01 5.30000e+01 5.40000e+01 ]
 [5.70000e+01 5.80000e+01 5.90000e+01 ]]



Total elem: 12
type  : Double (Float64)
cytnx device: CPU
Shape : (4,3)
[[4.20000e+01 4.30000e+01 4.40000e+01 ]
 [4.70000e+01 4.80000e+01 4.90000e+01 ]
 [5.20000e+01 5.30000e+01 5.40000e+01 ]
 [5.70000e+01 5.80000e+01 5.90000e+01 ]]



 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = arange(60).reshape(3,4,5)


print(A)


 


B = A[2,:,2:5:1]


print(B)


 


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 ]
  [4.50000e+01 4.60000e+01 4.70000e+01 4.80000e+01 4.90000e+01 ]
  [5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 ]
  [5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]]]




Total elem: 12
type  : Double (Float64)
cytnx device: CPU
Shape : (4,3)
[[4.20000e+01 4.30000e+01 4.40000e+01 ]
 [4.70000e+01 4.80000e+01 4.90000e+01 ]
 [5.20000e+01 5.30000e+01 5.40000e+01 ]
 [5.70000e+01 5.80000e+01 5.90000e+01 ]]




 






◆ imag()






      

        

          Tensor cytnx::Tensor::imag 
          (
          )
          
        

      





return the imaginary part of the tensor. 


Returns
[Tensor] the imaginary part of the tensor. 


Precondition
the tensor must be complex type (Type.ComplexDouble or Type.ComplexFloat). 


See also
cytnx::Type 








◆ Init()








  

  
      

        

          void cytnx::Tensor::Init 
          (
          const std::vector< cytnx_uint64 > & 
          shape, 
        

        

          
          
          const unsigned int & 
          dtype = Type.Double, 
        

        

          
          
          const int & 
          device = -1, 
        

        

          
          
          const bool & 
          init_zero = true 
        

        

          
          )
          
        

      

  		
  
inline  
  







initialize a Tensor 


Parameters

  

    
[in]shapethe shape of tensor. 

    
[in]dtypethe dtype of tensor. This can be any of type defined in cytnx::Type 

    
[in]devicethe device that tensor to be created. This can be cytnx::Device.cpu or 

    
[in]init_zeroif true, the content of Tensor will be initialized to zero. if false, the content of Tensor will be un-initialize. cytnx::Device.cuda+<gpuid>, see cytnx::Device for more detail.

  

  




Note
The content of Tensor created will be un-initialize! See zeros(), ones()  or arange()  for generating an Tensor.


Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  /*


      1. Create a Tensor with


      shape (3,4,5),


      dtype =Type.Double [default],


      device=Device.cpu [default]


  */


  Tensor A({3, 4, 5});


  cout << A << endl;


 


  /*


      2. Create a Tensor with


      shape (3,4,5),


      dtype =Type.Uint64,


      device=Device.cpu [default],


      [Note] the dtype can be any one of the supported type.


  */


  Tensor B({3, 4, 5}, Type.Uint64);


  cout << B << endl;


 


  /*


      3. Initialize a Tensor with


      shape (3,4,5),


      dtype =Type.Double,


      device=Device.cuda+0, (on gpu with gpu-id=0)


      [Note] the gpu device can be set with Device.cuda+<gpu-id>


  */


  Tensor C({3, 4, 5}, Type.Double, Device.cuda + 0);


  cout << C << endl;


 


  // 4. Create an empty Tensor, and init later


  Tensor D;


  D.Init({3, 4, 5}, Type.Double, Device.cpu);


 


  return 0;


}


cytnx::Tensor::Init
void Init(const std::vector< cytnx_uint64 > &shape, const unsigned int &dtype=Type.Double, const int &device=-1, const bool &init_zero=true)
initialize a Tensor
Definition Tensor.hpp:709


cytnx::Device
Device_class Device
data on which devices.
Definition Device.cpp:140


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]
 [[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]
 [[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]]



Total elem: 60
type  : Uint64
cytnx device: CPU
Shape : (3,4,5)
[[[                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]]
 [[                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]]
 [[                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]]]



 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


 


#1. Create a Tensor with 


#  shape (3,4,5),


#  dtype =Type.Double [default],


#  device=Device.cpu [default]


A = Tensor([3,4,5])


print(A)


 


#2. Create a Tensor with 


#  shape (3,4,5),


#  dtype =Type.Uint64,


#  device=Device.cpu [default],


#  [Note] the dtype can be any one of the supported type.


B = Tensor([3,4,5],dtype=Type.Uint64)


print(B)


 


#3. Initialize a Tensor with 


#  shape (3,4,5),


#  dtype =Type.Double,


#  device=Device.cuda+0, (on gpu with gpu-id=0)


#  [Note] the gpu device can be set with Device.cuda+<gpu-id>


C = Tensor([3,4,5],dtype=Type.Double,device=Device.cuda+0);


print(C)


 


#4. Create an empty Tensor, and init later


D = Tensor()


D.Init([3,4,5],dtype=Type.Double,device=Device.cpu);


 


 


 


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]
 [[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]
 [[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]]




Total elem: 60
type  : Uint64
cytnx device: CPU
Shape : (3,4,5)
[[[                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]]
 [[                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]]
 [[                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]
  [                  0                   0                   0                   0                   0 ]]]




Total elem: 60
type  : Double (Float64)
cytnx device: CUDA/GPU-id:0
Shape : (3,4,5)
[[[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]
 [[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]
 [[0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00 ]]]




 






◆ Inv()






      

        

          Tensor cytnx::Tensor::Inv 
          (
          const double & 
          clip)
           const
        

      





the Inv member function. Same as cytnx::linalg::Inv(const Tensor &Tin, const double &clip) 








◆ Inv_()






      

        

          Tensor & cytnx::Tensor::Inv_ 
          (
          const double & 
          clip)
          
        

      





the Inv_ member function. Same as cytnx::linalg::Inv_(Tensor &Tin, const double &clip) 








◆ InvM()






      

        

          Tensor cytnx::Tensor::InvM 
          (
          )
           const
        

      





the InvM member function. Same as cytnx::linalg::InvM(const Tensor &Tin), where Tin is the current Tensor. 








◆ InvM_()






      

        

          Tensor & cytnx::Tensor::InvM_ 
          (
          )
          
        

      





the InvM_ member function. Same as cytnx::linalg::InvM_(Tensor &Tin), where Tin is the current Tensor. 








◆ is_contiguous()








  

  
      

        

          const bool & cytnx::Tensor::is_contiguous 
          (
          )
           const
        

      

  		
  
inline  
  












◆ item()








template<class T > 



  

  
      

        

          T & cytnx::Tensor::item 
          (
          )
          
        

      

  		
  
inline  
  







get an from a rank-0 Tensor 


Returns
[T]


Note
    

  1. This can only be called on a rank-0 Tensor (scalar). For C++ API, a template instantiation of type is needed to resolve the type, which should be connsist with the dtype of the Tensor. An error will be issued if the template type if inconsist with the current dtype of Tensor.
    2. 

  2. Although the return is by reference in C++ part, the return in python is not.
    3. 

  3. From 2., We recommend user to use at<T> (C++ API) and [] (python API) to modify the value of the element to have consistant syntax across two languages.
    4. 





Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  Tensor A = ones(1, Type.Uint64);


  cout << A << endl;


 


  // note that type resolver should be consist with the dtype


  cout << A.item<cytnx_uint64>() << endl;


 


  return 0;


}


cytnx::Tensor::item
T & item()
get an from a rank-0 Tensor
Definition Tensor.hpp:1155


cytnx::ones
Tensor ones(const cytnx_uint64 &Nelem, const unsigned int &dtype=Type.Double, const int &device=Device.cpu)
create an rank-1 Tensor with all the elements are initialized with one.
Definition Generator.cpp:25


cytnx::cytnx_uint64
uint64_t cytnx_uint64
Definition Type.hpp:45


 
output>


Total elem: 1
type  : Uint64
cytnx device: CPU
Shape : (1)
[                  1 ]


1

 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = ones(1,Type.Uint64)


print(A)


 


print(A.item())


 


 


 
output>


Total elem: 1
type  : Uint64
cytnx device: CPU
Shape : (1)
[                  1 ]



1

 






◆ Load() [1/2]








  

  
      

        

          Tensor cytnx::Tensor::Load 
          (
          const char * 
          fname)
          
        

      

  		
  
static  
  






See also
Load(const std::string &fname) 








◆ Load() [2/2]








  

  
      

        

          Tensor cytnx::Tensor::Load 
          (
          const std::string & 
          fname)
          
        

      

  		
  
static  
  







Load current Tensor from file. 


Parameters

  

    
fname[in](within postfix ".cytn")

  

  




load the Storage from file with file path specify with input param 'fname' 
Precondition
the file must be saved with postfix ".cytn" 








◆ Max()






      

        

          Tensor cytnx::Tensor::Max 
          (
          )
           const
        

      





the Max member function. Same as linalg::Max(const Tensor &Tin), where Tin is the current Tensor. 








◆ Min()






      

        

          Tensor cytnx::Tensor::Min 
          (
          )
           const
        

      





the Min member function. Same as linalg::Min(const Tensor &Tin), where Tin is the current Tensor. 








◆ Mod()








template<class T > 



  

  
      

        

          Tensor cytnx::Tensor::Mod 
          (
          const T & 
          rhs)
          
        

      

  		
  
inline  
  












◆ Mul()








template<class T > 



  

  
      

        

          Tensor cytnx::Tensor::Mul 
          (
          const T & 
          rhs)
          
        

      

  		
  
inline  
  







Multiplication function with a Tensor or a scalar. Same as cytnx::operator*(const Tensor &self, const T &rhs). 


Parameters

  

    
[in]rhsthe multiplied Tensor or scalar. 

  

  










◆ Mul_()








template<class T > 



  

  
      

        

          Tensor & cytnx::Tensor::Mul_ 
          (
          const T & 
          rhs)
          
        

      

  		
  
inline  
  







Multiplication function with a Tensor or a scalar, inplacely. Same as operator*=(const T &rhs). 


Parameters

  

    
[in]rhsthe multiplied Tensor or scalar. 

  

  










◆ Norm()






      

        

          Tensor cytnx::Tensor::Norm 
          (
          )
           const
        

      





the Norm member function. Same as linalg::Norm(const Tensor &Tin), where Tin is the current Tensor. 








◆ operator*=()








template<class T > 

      

        

          Tensor & cytnx::Tensor::operator*= 
          (
          const T & 
          rc)
          
        

      





multiplication assignment operator with a Tensor or a scalar. 


This function will multiply the template type to the current tensor, inplacely. The template can be either a scalar or a tensor. If the template is a scalar, then the scalar will be multiplied to all the elements of the current tensor. If the template is a tensor, then the shape of the template tensor must be the same as the current tensor. The supported type of the template are Tensor, Scalar or any scalar type (see cytnx_complex128, cytnx_complex64, cytnx_double, cytnx_float, cytnx_int64, cytnx_int32, cytnx_int16, cytnx_uint64, cytnx_uint32, cytnx_uint16, cytnx_bool). 
Parameters

  

    
[in]rcthe multiplied Tensor or scalar. 

  

  




Precondition
If the template type is Tensor, then the shape of the template tensor must be the same as the current tensor. 








◆ operator+=()








template<class T > 

      

        

          Tensor & cytnx::Tensor::operator+= 
          (
          const T & 
          rc)
          
        

      





addition assignment operator with a Tensor or a scalar. 


This function will add the template type to the current tensor, inplacely. The template can be either a scalar or a tensor. If the template is a scalar, then the scalar will be added to all the elements of the current tensor. If the template is a tensor, then the shape of the template tensor must be the same as the current tensor. The supported type of the template are Tensor, Scalar or any scalar type (see cytnx_complex128, cytnx_complex64, cytnx_double, cytnx_float, cytnx_int64, cytnx_int32, cytnx_int16, cytnx_uint64, cytnx_uint32, cytnx_uint16, cytnx_bool). 
Parameters

  

    
[in]rcthe added Tensor or scalar. 

  

  




Precondition
If the template type is Tensor, then the shape of the template tensor must be the same as the current tensor. 








◆ operator-()








  

  
      

        

          Tensor cytnx::Tensor::operator- 
          (
          )
          
        

      

  		
  
inline  
  







The negation function. 


This function is the negation function. Namely, if the current tensor is \(A\), then the output tensor is \(-A\). 
Returns
The negation of the current tensor. 








◆ operator-=()








template<class T > 

      

        

          Tensor & cytnx::Tensor::operator-= 
          (
          const T & 
          rc)
          
        

      





subtraction assignment operator with a Tensor or a scalar. 


This function will subtract the template type to the current tensor, inplacely. The template can be either a scalar or a tensor. If the template is a scalar, then the scalar will be subtracted to all the elements of the current tensor. If the template is a tensor, then the shape of the template tensor must be the same as the current tensor. The supported type of the template are Tensor, Scalar or any scalar type (see cytnx_complex128, cytnx_complex64, cytnx_double, cytnx_float, cytnx_int64, cytnx_int32, cytnx_int16, cytnx_uint64, cytnx_uint32, cytnx_uint16, cytnx_bool). 
Parameters

  

    
[in]rcthe subtracted Tensor or scalar. 

  

  




Precondition
If the template type is Tensor, then the shape of the template tensor must be the same as the current tensor. 








◆ operator/=()








template<class T > 

      

        

          Tensor & cytnx::Tensor::operator/= 
          (
          const T & 
          rc)
          
        

      





division assignment operator with a Tensor or a scalar. 


This function will divide the template type to the current tensor, inplacely. The template can be either a scalar or a tensor. If the template is a scalar, then the scalar will be divided to all the elements of the current tensor. If the template is a tensor, then the shape of the template tensor must be the same as the current tensor. The supported type of the template are Tensor, Scalar or any scalar type (see cytnx_complex128, cytnx_complex64, cytnx_double, cytnx_float, cytnx_int64, cytnx_int32, cytnx_int16, cytnx_uint64, cytnx_uint32, cytnx_uint16, cytnx_bool). 
Parameters

  

    
[in]rcthe divided Tensor or scalar. 

  

  




Precondition
    

  1. If the template type is Tensor, then the shape of the template tensor must be the same as the current tensor.
    2. 

  2. rc cannot be zero. 
    3. 











◆ permute()








  

  
      

        

          Tensor cytnx::Tensor::permute 
          (
          const std::vector< cytnx_uint64 > & 
          rnks)
           const
        

      

  		
  
inline  
  







perform tensor permute on the cytnx::Tensor and return a new instance. 


Parameters

  

    
[in]rnksthe permute indices, should have No. of elements equal to the rank of tensor. 

  

  




Returns
[Tensor] a permuted new Tensor 


Precondition
    

  1. The size of input and output Tensor should be the same.
    2. 

  2. rnks cannot contain duplicated elements.
    3. 





Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  Tensor A({3, 4, 5});


  cout << A.shape() << endl;


 


  Tensor B = A.permute({0, 2, 1});


  cout << B.shape() << endl;


 


  cout << is(B, A) << endl;  // this should be false, different object.


 


  cout << B.same_data(A) << endl;  // this should be true, since no new pointer/memory is created.


 


  return 0;


}


 
output>


Vector Print:
Total Elements:3
[3, 4, 5]

Vector Print:
Total Elements:3
[3, 5, 4]

0
1

 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = Tensor([3,4,5])


print(A.shape())


 


B = A.permute(0,2,1)


print(B.shape())


 


print(B is A) #False


 


print(B.same_data(A)) #True


 


 


 
output>


[3, 4, 5]
[3, 5, 4]
False
True

 






◆ permute_()








  

  
      

        

          Tensor cytnx::Tensor::permute_ 
          (
          const std::vector< cytnx_uint64 > & 
          rnks)
          
        

      

  		
  
inline  
  












◆ Pow()






      

        

          Tensor cytnx::Tensor::Pow 
          (
          const cytnx_double & 
          p)
           const
        

      





the Pow member function. Same as linalg::Pow(const Tensor &Tin, const cytnx_double
&p), where Tin is the current Tensor. 








◆ Pow_()






      

        

          Tensor & cytnx::Tensor::Pow_ 
          (
          const cytnx_double & 
          p)
          
        

      





the Pow_ member function. Same as linalg::Pow_(Tensor &Tin, const cytnx_double
&p), where Tin is the current Tensor. 








◆ rank()








  

  
      

        

          cytnx_uint64 cytnx::Tensor::rank 
          (
          )
           const
        

      

  		
  
inline  
  







the rank of the Tensor 


Returns
[cytnx_uint64] the rank of the Tensor 








◆ real()






      

        

          Tensor cytnx::Tensor::real 
          (
          )
          
        

      





return the real part of the tensor. 


Returns
[Tensor] the real part of the tensor. 


Precondition
the tensor must be complex type (Type.ComplexDouble or Type.ComplexFloat). 


See also
cytnx::Type 








◆ reshape() [1/3]








  

  
      

        

          Tensor cytnx::Tensor::reshape 
          (
          const std::initializer_list< cytnx_int64 > & 
          new_shape)
           const
        

      

  		
  
inline  
  






See also
reshape(const std::vector<cytnx_int64> &new_shape) const 








◆ reshape() [2/3]








  

  
      

        

          Tensor cytnx::Tensor::reshape 
          (
          const std::vector< cytnx_int64 > & 
          new_shape)
           const
        

      

  		
  
inline  
  







return a new Tensor that is reshaped. 


Parameters

  

    
[in]new_shapethe new shape of the Tensor. 

  

  




Returns
[Tensor] 


Precondition
    

  1. The size of input and output Tensor should be the same.
    2. 

  2. new_shape cannot be empty. 
    3. 





Note
    

  1. This function will not change the original Tensor.
    2. 

  2. You can use Tensor::reshape_() to reshape the Tensor inplacely.
    3. 

  3. You can set new_shape to -1 if the shape of the Tensor is undetermined. 
    4. 





See also
Tensor::reshape_() 


Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  Tensor A = arange(60);


 


  // there are two ways you can give argument to reshape:


 


  // Method 1: more like 'C++' way:


  Tensor B = A.reshape({5, 12});  // note the braket{}


  cout << A << endl;


  cout << B << endl;


 


  // Method 2: more like 'python' way:


  Tensor B2 = A.reshape(5, 12);


 


  return 0;


}


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (60)
[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 4.50000e+01 4.60000e+01 4.70000e+01 4.80000e+01 4.90000e+01 5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]



Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (5,12)
[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 1.00000e+01 1.10000e+01 ]
 [1.20000e+01 1.30000e+01 1.40000e+01 1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 ]
 [2.40000e+01 2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 3.50000e+01 ]
 [3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 4.50000e+01 4.60000e+01 4.70000e+01 ]
 [4.80000e+01 4.90000e+01 5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]]



 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = arange(60)


 


B = A.reshape(5,12)


print(A)


print(B)


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (60)
[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 4.50000e+01 4.60000e+01 4.70000e+01 4.80000e+01 4.90000e+01 5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]




Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (5,12)
[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 1.00000e+01 1.10000e+01 ]
 [1.20000e+01 1.30000e+01 1.40000e+01 1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 ]
 [2.40000e+01 2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 3.50000e+01 ]
 [3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 4.50000e+01 4.60000e+01 4.70000e+01 ]
 [4.80000e+01 4.90000e+01 5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]]




 






◆ reshape() [3/3]








  

  
      

        

          Tensor cytnx::Tensor::reshape 
          (
          const std::vector< cytnx_uint64 > & 
          new_shape)
           const
        

      

  		
  
inline  
  






See also
reshape(const std::vector<cytnx_int64> &new_shape) const 








◆ reshape_()








  

  
      

        

          void cytnx::Tensor::reshape_ 
          (
          const std::vector< cytnx_int64 > & 
          new_shape)
          
        

      

  		
  
inline  
  







reshape the Tensor, inplacely 


Parameters

  

    
[in]new_shapethe new shape of the Tensor. 

  

  




Precondition
    

  1. The size of input and output Tensor should be the same.
    2. 

  2. new_shape cannot be empty. 
    3. 





See also
Tensor::reshape()  


Note
Compare to reshape(), this function will not create a new Tensor, but reshape the current Tensor inplacely.


Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  Tensor A = arange(60);


  cout << A << endl;


 


  // there are two ways you can assign the argument:


 


  // Method 1: more like 'c++' way:


  A.reshape_({5, 12});  // note the braket{}


  cout << A << endl;


 


  // Method 2: more like 'python' way:


  A.reshape_(5, 4, 3);


  cout << A << endl;


 


  return 0;


}


cytnx::Tensor::reshape_
void reshape_(const std::vector< cytnx_int64 > &new_shape)
reshape the Tensor, inplacely
Definition Tensor.hpp:978


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (60)
[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 4.50000e+01 4.60000e+01 4.70000e+01 4.80000e+01 4.90000e+01 5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]



Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (5,12)
[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 1.00000e+01 1.10000e+01 ]
 [1.20000e+01 1.30000e+01 1.40000e+01 1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 ]
 [2.40000e+01 2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 3.50000e+01 ]
 [3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 4.50000e+01 4.60000e+01 4.70000e+01 ]
 [4.80000e+01 4.90000e+01 5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]]



Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (5,4,3)
[[[0.00000e+00 1.00000e+00 2.00000e+00 ]
  [3.00000e+00 4.00000e+00 5.00000e+00 ]
  [6.00000e+00 7.00000e+00 8.00000e+00 ]
  [9.00000e+00 1.00000e+01 1.10000e+01 ]]
 [[1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 ]
  [1.80000e+01 1.90000e+01 2.00000e+01 ]
  [2.10000e+01 2.20000e+01 2.30000e+01 ]]
 [[2.40000e+01 2.50000e+01 2.60000e+01 ]
  [2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 ]
  [3.30000e+01 3.40000e+01 3.50000e+01 ]]
 [[3.60000e+01 3.70000e+01 3.80000e+01 ]
  [3.90000e+01 4.00000e+01 4.10000e+01 ]
  [4.20000e+01 4.30000e+01 4.40000e+01 ]
  [4.50000e+01 4.60000e+01 4.70000e+01 ]]
 [[4.80000e+01 4.90000e+01 5.00000e+01 ]
  [5.10000e+01 5.20000e+01 5.30000e+01 ]
  [5.40000e+01 5.50000e+01 5.60000e+01 ]
  [5.70000e+01 5.80000e+01 5.90000e+01 ]]]



 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = arange(60)


print(A)


A.reshape_(5,12)


print(A)


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (60)
[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 4.50000e+01 4.60000e+01 4.70000e+01 4.80000e+01 4.90000e+01 5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]




Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (5,12)
[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 1.00000e+01 1.10000e+01 ]
 [1.20000e+01 1.30000e+01 1.40000e+01 1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 ]
 [2.40000e+01 2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 3.50000e+01 ]
 [3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 4.50000e+01 4.60000e+01 4.70000e+01 ]
 [4.80000e+01 4.90000e+01 5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]]




 






◆ same_data()






      

        

          bool cytnx::Tensor::same_data 
          (
          const Tensor & 
          rhs)
           const
        

      










◆ Save() [1/2]






      

        

          void cytnx::Tensor::Save 
          (
          const char * 
          fname)
           const
        

      




See also
Save(const std::string &fname) 








◆ Save() [2/2]






      

        

          void cytnx::Tensor::Save 
          (
          const std::string & 
          fname)
           const
        

      





Save current Tensor to file. 


Parameters

  

    
[in]fnamefile name (without file extension)

  

  




save the Tensor to file with file path specify with input param fname with postfix ".cytn" 
Postcondition
the file will be saved with postfix ".cytn" 


See also
Load(const std::string &fname) 








◆ set() [1/2]








template<class T > 



  

  
      

        

          void cytnx::Tensor::set 
          (
          const std::vector< cytnx::Accessor > & 
          accessors, 
        

        

          
          
          const T & 
          rc 
        

        

          
          )
          
        

      

  		
  
inline  
  







set elements with the input constant using Accessor (C++ API) / slices (python API) 


Parameters

  

    
[in]accessorsthe list(vector) of accessors. 

    
rc[Const]

  

  




See also
Tensor::fill  for filling all elements with assigned constant.


Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  typedef Accessor ac;


 


  // Let's define two Tensor.


  Tensor A = arange(60).reshape({3, 4, 5});


  cout << A << endl;


 


  Tensor B = zeros({4, 3});


  cout << B << endl;


 


  // I. To set part of A with Tesnor B, or constant:


  // [Method 1] Using direct assignment


  //-------------------------------------


  A(2, ":", "2:5:1") = B;


  cout << A << endl;


 


  A(2, ":", "2:5:1") = 999;


  cout << A << endl;


 


  // note this is the same as


  // A(ac(2),ac::all(),ac::range(2,5,1)) = B;


  // A[{ac(2),ac::all(),ac::range(2,5,1)}] = B; // note that braket{}


 


  // [Method 2] Using low-level API set():


  //--------------------------------------


  A.set({ac(2), ac::all(), ac::range(2, 5, 1)}, B);


  cout << A << endl;


 


  A.set({ac(2), ac::all(), ac::range(0, 2, 1)}, 999);


  cout << A << endl;


 


  return 0;


}


cytnx::Tensor::set
void set(const std::vector< cytnx::Accessor > &accessors, const Tensor &rhs)
set elements with the input Tensor using Accessor (C++ API) / slices (python API)
Definition Tensor.hpp:1234


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 ]
  [4.50000e+01 4.60000e+01 4.70000e+01 4.80000e+01 4.90000e+01 ]
  [5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 ]
  [5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]]]



Total elem: 12
type  : Double (Float64)
cytnx device: CPU
Shape : (4,3)
[[0.00000e+00 0.00000e+00 0.00000e+00 ]
 [0.00000e+00 0.00000e+00 0.00000e+00 ]
 [0.00000e+00 0.00000e+00 0.00000e+00 ]
 [0.00000e+00 0.00000e+00 0.00000e+00 ]]



Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [4.50000e+01 4.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.00000e+01 5.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.50000e+01 5.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]]]



Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [4.50000e+01 4.60000e+01 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [5.00000e+01 5.10000e+01 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [5.50000e+01 5.60000e+01 9.99000e+02 9.99000e+02 9.99000e+02 ]]]



Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [4.50000e+01 4.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.00000e+01 5.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.50000e+01 5.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]]]



Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[9.99000e+02 9.99000e+02 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [9.99000e+02 9.99000e+02 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [9.99000e+02 9.99000e+02 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [9.99000e+02 9.99000e+02 0.00000e+00 0.00000e+00 0.00000e+00 ]]]



 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = arange(60).reshape(3,4,5)


print(A)


 


B = zeros([4,3])


print(B)


 


A[2,:,2:5:1] = B


print(A)


 


A[0,:,0:2:1] = 999


print(A)


 


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 ]
  [4.50000e+01 4.60000e+01 4.70000e+01 4.80000e+01 4.90000e+01 ]
  [5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 ]
  [5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]]]




Total elem: 12
type  : Double (Float64)
cytnx device: CPU
Shape : (4,3)
[[0.00000e+00 0.00000e+00 0.00000e+00 ]
 [0.00000e+00 0.00000e+00 0.00000e+00 ]
 [0.00000e+00 0.00000e+00 0.00000e+00 ]
 [0.00000e+00 0.00000e+00 0.00000e+00 ]]




Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [4.50000e+01 4.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.00000e+01 5.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.50000e+01 5.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]]]




Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[9.99000e+02 9.99000e+02 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [9.99000e+02 9.99000e+02 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [9.99000e+02 9.99000e+02 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [9.99000e+02 9.99000e+02 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [4.50000e+01 4.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.00000e+01 5.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.50000e+01 5.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]]]




 






◆ set() [2/2]








  

  
      

        

          void cytnx::Tensor::set 
          (
          const std::vector< cytnx::Accessor > & 
          accessors, 
        

        

          
          
          const Tensor & 
          rhs 
        

        

          
          )
          
        

      

  		
  
inline  
  







set elements with the input Tensor using Accessor (C++ API) / slices (python API) 


Parameters

  

    
[in]accessorsthe list(vector) of accessors. 

    
rhs[Tensor] 

  

  




Note
: the shape of the input Tensor should be the same as the shape that indicated using Accessor. The memory is not shared with the input Tensor.


Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  typedef Accessor ac;


 


  // Let's define two Tensor.


  Tensor A = arange(60).reshape({3, 4, 5});


  cout << A << endl;


 


  Tensor B = zeros({4, 3});


  cout << B << endl;


 


  // I. To set part of A with Tesnor B, or constant:


  // [Method 1] Using direct assignment


  //-------------------------------------


  A(2, ":", "2:5:1") = B;


  cout << A << endl;


 


  A(2, ":", "2:5:1") = 999;


  cout << A << endl;


 


  // note this is the same as


  // A(ac(2),ac::all(),ac::range(2,5,1)) = B;


  // A[{ac(2),ac::all(),ac::range(2,5,1)}] = B; // note that braket{}


 


  // [Method 2] Using low-level API set():


  //--------------------------------------


  A.set({ac(2), ac::all(), ac::range(2, 5, 1)}, B);


  cout << A << endl;


 


  A.set({ac(2), ac::all(), ac::range(0, 2, 1)}, 999);


  cout << A << endl;


 


  return 0;


}


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 ]
  [4.50000e+01 4.60000e+01 4.70000e+01 4.80000e+01 4.90000e+01 ]
  [5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 ]
  [5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]]]



Total elem: 12
type  : Double (Float64)
cytnx device: CPU
Shape : (4,3)
[[0.00000e+00 0.00000e+00 0.00000e+00 ]
 [0.00000e+00 0.00000e+00 0.00000e+00 ]
 [0.00000e+00 0.00000e+00 0.00000e+00 ]
 [0.00000e+00 0.00000e+00 0.00000e+00 ]]



Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [4.50000e+01 4.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.00000e+01 5.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.50000e+01 5.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]]]



Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [4.50000e+01 4.60000e+01 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [5.00000e+01 5.10000e+01 9.99000e+02 9.99000e+02 9.99000e+02 ]
  [5.50000e+01 5.60000e+01 9.99000e+02 9.99000e+02 9.99000e+02 ]]]



Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [4.50000e+01 4.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.00000e+01 5.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.50000e+01 5.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]]]



Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[9.99000e+02 9.99000e+02 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [9.99000e+02 9.99000e+02 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [9.99000e+02 9.99000e+02 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [9.99000e+02 9.99000e+02 0.00000e+00 0.00000e+00 0.00000e+00 ]]]



 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = arange(60).reshape(3,4,5)


print(A)


 


B = zeros([4,3])


print(B)


 


A[2,:,2:5:1] = B


print(A)


 


A[0,:,0:2:1] = 999


print(A)


 


 
output>


Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 4.20000e+01 4.30000e+01 4.40000e+01 ]
  [4.50000e+01 4.60000e+01 4.70000e+01 4.80000e+01 4.90000e+01 ]
  [5.00000e+01 5.10000e+01 5.20000e+01 5.30000e+01 5.40000e+01 ]
  [5.50000e+01 5.60000e+01 5.70000e+01 5.80000e+01 5.90000e+01 ]]]




Total elem: 12
type  : Double (Float64)
cytnx device: CPU
Shape : (4,3)
[[0.00000e+00 0.00000e+00 0.00000e+00 ]
 [0.00000e+00 0.00000e+00 0.00000e+00 ]
 [0.00000e+00 0.00000e+00 0.00000e+00 ]
 [0.00000e+00 0.00000e+00 0.00000e+00 ]]




Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[0.00000e+00 1.00000e+00 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [5.00000e+00 6.00000e+00 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [1.00000e+01 1.10000e+01 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [1.50000e+01 1.60000e+01 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [4.50000e+01 4.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.00000e+01 5.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.50000e+01 5.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]]]




Total elem: 60
type  : Double (Float64)
cytnx device: CPU
Shape : (3,4,5)
[[[9.99000e+02 9.99000e+02 2.00000e+00 3.00000e+00 4.00000e+00 ]
  [9.99000e+02 9.99000e+02 7.00000e+00 8.00000e+00 9.00000e+00 ]
  [9.99000e+02 9.99000e+02 1.20000e+01 1.30000e+01 1.40000e+01 ]
  [9.99000e+02 9.99000e+02 1.70000e+01 1.80000e+01 1.90000e+01 ]]
 [[2.00000e+01 2.10000e+01 2.20000e+01 2.30000e+01 2.40000e+01 ]
  [2.50000e+01 2.60000e+01 2.70000e+01 2.80000e+01 2.90000e+01 ]
  [3.00000e+01 3.10000e+01 3.20000e+01 3.30000e+01 3.40000e+01 ]
  [3.50000e+01 3.60000e+01 3.70000e+01 3.80000e+01 3.90000e+01 ]]
 [[4.00000e+01 4.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [4.50000e+01 4.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.00000e+01 5.10000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]
  [5.50000e+01 5.60000e+01 0.00000e+00 0.00000e+00 0.00000e+00 ]]]




 






◆ shape()








  

  
      

        

          const std::vector< cytnx_uint64 > & cytnx::Tensor::shape 
          (
          )
           const
        

      

  		
  
inline  
  







the shape of the Tensor 


Returns
[std::vector<cytnx_uint64>] the shape of the Tensor 








◆ storage()








  

  
      

        

          Storage & cytnx::Tensor::storage 
          (
          )
           const
        

      

  		
  
inline  
  







return the storage of current Tensor. 


Returns
[Storage]


Note
:
    

  1. The return storage shares the same instance of the storage of current Tensor. Use Storage.clone()  to create a new instance of the returned Storage. 
    2. 











◆ Sub()








template<class T > 



  

  
      

        

          Tensor cytnx::Tensor::Sub 
          (
          const T & 
          rhs)
          
        

      

  		
  
inline  
  







Subtraction function with a Tensor or a scalar. Same as cytnx::operator-(const Tensor &self, const T &rhs). 


Parameters

  

    
[in]rhsthe subtracted Tensor or scalar. 

  

  










◆ Sub_()








template<class T > 



  

  
      

        

          Tensor & cytnx::Tensor::Sub_ 
          (
          const T & 
          rhs)
          
        

      

  		
  
inline  
  







Subtraction function with a Tensor or a scalar, inplacely. Same as operator-=(const T &rhs). 


Parameters

  

    
[in]rhsthe subtracted Tensor or scalar. 

  

  










◆ Svd()






      

        

          std::vector< Tensor > cytnx::Tensor::Svd 
          (
          const bool & 
          is_UvT = true)
           const
        

      





the SVD member function. Same as cytnx::linalg::Svd(const Tensor &Tin, const bool &is_UvT) , where Tin is the current Tensor. 








◆ to()








  

  
      

        

          Tensor cytnx::Tensor::to 
          (
          const int & 
          device)
           const
        

      

  		
  
inline  
  







copy a tensor to new device 


Parameters

  

    
[in]devicethe device-id that is moving to. it can be any device defined in cytnx::Device 

  

  




Returns
[Tensor]


description:

 if the device-id is the same as current Tensor's device, then return self.

 otherwise, return a copy of instance that located on the target device. 

 see also: Tensor.to_  

 
Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  Tensor A({3, 4, 5});


 


  // move the tensor to different device by creating a clone object


  Tensor B = A.to(Device.cuda + 0);


  cout << B.device_str() << endl;


  cout << A.device_str() << endl;


 


  return 0;


}


cytnx::Tensor::device_str
std::string device_str() const
the device (in string) of the Tensor
Definition Tensor.hpp:789


cytnx::Tensor::to
Tensor to(const int &device) const
copy a tensor to new device
Definition Tensor.hpp:846


 
output>


 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = Tensor([3,4,5])


B = A.to(Device.cuda+0);


print(B.device_str())


print(A.device_str())


 


 


 
output>


cytnx device: CUDA/GPU-id:0
cytnx device: CPU

 






◆ to_()








  

  
      

        

          void cytnx::Tensor::to_ 
          (
          const int & 
          device)
          
        

      

  		
  
inline  
  







move the current Tensor to the device. 


Parameters

  

    
[in]devicethe device-id that is moving to. it can be any device defined in cytnx::Device

  

  




description:

 see also: Tensor.to 

 
Example:


c++ API:


#include "cytnx.hpp"


#include <iostream>


 


using namespace cytnx;


using namespace std;


int main() {


  Tensor A({3, 4, 5});


 


  // move the instance tensor to different device


  A.to_(Device.cuda + 0);


  cout << A.device_str() << endl;


 


  return 0;


}


cytnx::Tensor::to_
void to_(const int &device)
move the current Tensor to the device.
Definition Tensor.hpp:869


 
output>


 
python API:


import sys


from pathlib import Path


home = str(Path.home())


sys.path.append(home + '/Cytnx_lib')


from cytnx import *


 


A = Tensor([3,4,5])


 


A.to_(Device.cuda+0);


print(A.device_str())


 


 
output>


cytnx device: CUDA/GPU-id:0

 






◆ Tofile() [1/3]






      

        

          void cytnx::Tensor::Tofile 
          (
          const char * 
          fname)
           const
        

      










◆ Tofile() [2/3]






      

        

          void cytnx::Tensor::Tofile 
          (
          const std::string & 
          fname)
           const
        

      










◆ Tofile() [3/3]






      

        

          void cytnx::Tensor::Tofile 
          (
          std::fstream & 
          f)
           const
        

      










◆ Trace()






      

        

          Tensor cytnx::Tensor::Trace 
          (
          const cytnx_uint64 & 
          a = 0, 
        

        

          
          
          const cytnx_uint64 & 
          b = 1 
        

        

          
          )
           const
        

      





the Trace member function. Same as linalg::Trace(const Tensor &Tin, const
cytnx_uint64 &a, const cytnx_uint64 &b), where Tin is the current Tensor. 







The documentation for this class was generated from the following files:





Generated by doxygen 1.9.8