Class Tensor

java.lang.Object

org.flag4j.arrays.backend.AbstractTensor<Tensor,double[],Double>

org.flag4j.arrays.backend.primitive_arrays.AbstractDoubleTensor<Tensor>

org.flag4j.arrays.backend.primitive_arrays.AbstractDenseDoubleTensor<Tensor>

org.flag4j.arrays.dense.Tensor

All Implemented Interfaces:

Serializable, TensorOverField<Tensor,Tensor,double[],Double>, TensorOverRing<Tensor,Tensor,double[],Double>, TensorOverSemiring<Tensor,Tensor,double[],Double>

public class Tensor
extends AbstractDenseDoubleTensor<Tensor>

A real dense tensor backed by a primitive double array.

A tensor is a multidimensional array. If N indices are required to uniquely identify all elements of a tensor, then the tensor is considered an N-dimensional tensor/array or a rank-N tensor.

The AbstractTensor.data of a Tensor are mutable but the AbstractTensor.shape is fixed.

See Also:

• Serialized Form

Field Summary

Fields inherited from class org.flag4j.arrays.backend.AbstractTensor

data, rank, shape

Constructor Summary

Constructors

Constructor	Description
Tensor(Object nDArray)	Creates a tensor from an nD array.
Tensor(Tensor src)	Constructs a copy of the specified tensor.
Tensor(Shape shape)	Creates a zero tensor with the shape.
Tensor(Shape shape, double fillValue)	Creates a tensor with the specified shape filled with fillValue.
Tensor(Shape shape, double... entries)	Creates a tensor with the specified data and shape.
Tensor(Shape shape, int... entries)	Creates a tensor with the specified data and shape.
Tensor(Shape shape, Double[] entries)	Creates a tensor with the specified data and shape.
Tensor(Shape shape, Integer[] entries)	Creates a tensor with the specified data and shape.

Method Summary

Modifier and Type	Method	Description
CTensor	add(Complex128 b)	Adds a complex-valued scalar value to each entry of this tensor.
CTensor	add(CTensor b)	Computes the element-wise sum between two tensors of the same shape.
CTensor	add(CooCTensor b)	Computes the element-wise sum between two tensors of the same shape.
Tensor	add(CooTensor b)	Computes the element-wise sum between two tensors of the same shape.
CTensor	div(Complex128 b)	Divides each entry of this tensor by a complex scalar value.
CTensor	div(CTensor b)	Computes the element-wise division of two tensors.
CTensor	elemMult(CTensor b)	Computes the element-wise multiplication of two tensors.
CooCTensor	elemMult(CooCTensor b)	Computes the element-wise multiplication between two tensors.
CooTensor	elemMult(CooTensor b)	Computes the element-wise multiplication between two tensors.
boolean	equals(Object object)	Checks if an object is equal to this tensor object.
Tensor	flatten()	Flattens tensor to single dimension while preserving order of data.
Tensor	flatten(int axis)	Flattens a tensor along the specified axis.
int	hashCode()
Tensor	makeLikeTensor(Shape shape, double[] entries)	Constructs a tensor of the same type as this tensor with the given the shape and data.
CTensor	mult(Complex128 b)	Multiplies a complex scalar value to each entry of this tensor.
CTensor	sub(Complex128 b)	Subtracts a complex-valued scalar from each element of this tensor.
CTensor	sub(CTensor b)	Computes the element-wise difference between two tensors of the same shape.
CTensor	sub(CooCTensor b)	Computes the element-wise difference between two tensors of the same shape and stores the result in this tensor.
Tensor	sub(CooTensor b)	Computes the element-wise difference between two tensors of the same shape.
CTensor	toComplex()	Converts this tensor to an equivalent complex valued tensor.
CooTensor	toCoo()	Converts this tensor to an equivalent sparse COO tensor.
CooTensor	toCoo(double estimatedSparsity)	Converts this tensor to an equivalent sparse COO tensor.
Matrix	toMatrix()	Converts this tensor to an equivalent matrix.
Matrix	toMatrix(Shape shape)	Converts this tensor to an equivalent matrix with the specified shape.
String	toString()	Formats this tensor as a human-readable string.
Vector	toVector()	Converts this tensor to an equivalent vector.

Methods inherited from class org.flag4j.arrays.backend.primitive_arrays.AbstractDenseDoubleTensor

add, addEq, allClose, allClose, argmax, argmaxAbs, argmin, argminAbs, div, divEq, elemMult, elemMultEq, get, reshape, round, round, roundToZero, set, sub, subEq, T, T, tensorDot, tensorTr

Methods inherited from class org.flag4j.arrays.backend.primitive_arrays.AbstractDoubleTensor

abs, add, add, addEq, addEq, conj, copy, div, div, divEq, divEq, H, H, isFinite, isInfinite, isNaN, isNeg, isOnes, isPos, isZeros, max, maxAbs, min, minAbs, mult, mult, multEq, multEq, prod, recip, roundToZero, sqrt, sub, sub, subEq, subEq, sum

Methods inherited from class org.flag4j.arrays.backend.AbstractTensor

getData, getRank, getShape, reshape, sameShape, T, totalEntries

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface org.flag4j.arrays.backend.ring_arrays.TensorOverRing

H

Methods inherited from interface org.flag4j.arrays.backend.semiring_arrays.TensorOverSemiring

getData, getRank, getShape, tensorDot, tensorDot, tensorDot, tensorTr

Constructor Details

Tensor

public Tensor(Shape shape)

Creates a zero tensor with the shape.

Parameters:

shape - Shape of this tensor.

Tensor

public Tensor(Object nDArray)

Creates a tensor from an nD array. The tensors shape will be inferred from.

Parameters:

nDArray - Array to construct tensor from. Must be a rectangular array.

Throws:

IllegalArgumentException - If nDArray is not an array or not rectangular.

Tensor

public Tensor(Shape shape,
 double fillValue)

Creates a tensor with the specified shape filled with fillValue.

Parameters:

shape - Shape of this tensor.

fillValue - Value to fill this tensor with.

Tensor

public Tensor(Shape shape,
 double... entries)

Creates a tensor with the specified data and shape.

Parameters:

shape - Shape of this tensor.

entries - Entries of this tensor.

Tensor

public Tensor(Shape shape,
 int... entries)

Creates a tensor with the specified data and shape.

Parameters:

shape - Shape of this tensor.

entries - Entries of this tensor.

Tensor

public Tensor(Tensor src)

Constructs a copy of the specified tensor.

Parameters:

src - The tensor to make a copy of.

Tensor

public Tensor(Shape shape,
 Double[] entries)

Creates a tensor with the specified data and shape.

Parameters:

shape - Shape of this tensor.

entries - Entries of this tensor.

Tensor

public Tensor(Shape shape,
 Integer[] entries)

Creates a tensor with the specified data and shape.

Parameters:

shape - Shape of this tensor.

entries - Entries of this tensor.

Method Details

flatten

public Tensor flatten()

Flattens tensor to single dimension while preserving order of data.

Specified by:

flatten in class AbstractTensor<Tensor,double[],Double>

Returns:

The flattened tensor.

See Also:

• flatten(int)

flatten

public Tensor flatten(int axis)

Flattens a tensor along the specified axis.

Specified by:

flatten in class AbstractTensor<Tensor,double[],Double>

Parameters:

axis - Axis along which to flatten tensor.

Throws:

ArrayIndexOutOfBoundsException - If the axis is not positive or larger than this.{@link #getRank()}-1.

See Also:

• flatten()

makeLikeTensor

public Tensor makeLikeTensor(Shape shape,
 double[] entries)

Constructs a tensor of the same type as this tensor with the given the shape and data.

Specified by:

makeLikeTensor in interface TensorOverSemiring<Tensor,Tensor,double[],Double>

Specified by:

makeLikeTensor in class AbstractTensor<Tensor,double[],Double>

Parameters:

shape - Shape of the tensor to construct.

entries - Entries of the tensor to construct.

Returns:

A tensor of the same type as this tensor with the given the shape and data.

toVector

public Vector toVector()

Converts this tensor to an equivalent vector. If this vector is not rank-1 it will first be flattened then converted to a vector.

Returns:

A vector with data equivalent to this vector.

toMatrix

public Matrix toMatrix()

Converts this tensor to an equivalent matrix. If this matrix is not rank-2 it will first be flattened to a row vector then converted to a matrix.

Returns:

A matrix with data equivalent to this tensor.

toMatrix

public Matrix toMatrix(Shape shape)

Converts this tensor to an equivalent matrix with the specified shape.

Parameters:

shape - New shape for the matrix. Must be rank-2 and broadcastable to this.shape.

Returns:

A matrix with the specified shape and data equivalent to this tensor.

Throws:

IllegalArgumentException - If shape is not broadcastable to this.shape.

TensorShapeException - If shape.getRank() != 2.

toCoo

public CooTensor toCoo()

Converts this tensor to an equivalent sparse COO tensor.

Returns:

A sparse COO tensor that is equivalent to this dense tensor.

See Also:

• toCoo(double)

add

public CTensor add(Complex128 b)

Adds a complex-valued scalar value to each entry of this tensor. If the tensor is sparse, the scalar will only be added to the non-zero data of the tensor.

Parameters:

b - Scalar value in sum.

Returns:

The sum of this tensor with the scalar b.

add

public Tensor add(CooTensor b)

Computes the element-wise sum between two tensors of the same shape.

Parameters:

b - Second tensor in the element-wise sum.

Returns:

The sum of this tensor with b.

Throws:

TensorShapeException - If this tensor and b do not have the same shape.

add

public CTensor add(CTensor b)

Computes the element-wise sum between two tensors of the same shape.

Parameters:

b - Second tensor in the element-wise sum.

Returns:

The sum of this tensor with b.

Throws:

TensorShapeException - If this tensor and b do not have the same shape.

add

public CTensor add(CooCTensor b)

Computes the element-wise sum between two tensors of the same shape.

Parameters:

b - Second tensor in the element-wise sum.

Returns:

The sum of this tensor with b.

Throws:

TensorShapeException - If this tensor and b do not have the same shape.

sub

public CTensor sub(Complex128 b)

Subtracts a complex-valued scalar from each element of this tensor.

Parameters:

b - Scalar value in vector-scalar difference.

Returns:

The tensor resulting from subtracting b from each entry of this tensor.

sub

public CTensor sub(CooCTensor b)

Computes the element-wise difference between two tensors of the same shape and stores the result in this tensor.

Parameters:

b - Second tensor in the element-wise difference.

Throws:

TensorShapeException - If this tensor and b do not have the same shape.

sub

public Tensor sub(CooTensor b)

Computes the element-wise difference between two tensors of the same shape.

Parameters:

b - Second tensor in the element-wise difference.

Returns:

The difference of this tensor with b.

Throws:

TensorShapeException - If this tensor and b do not have the same shape.

sub

public CTensor sub(CTensor b)

Computes the element-wise difference between two tensors of the same shape.

Parameters:

b - Second tensor in the element-wise difference.

Returns:

The difference of this tensor with b.

Throws:

TensorShapeException - If this tensor and b do not have the same shape.

elemMult

public CooTensor elemMult(CooTensor b)

Computes the element-wise multiplication between two tensors.

Parameters:

b - Second tensor in the element-wise multiplication.

Returns:

The element-wise product of this matrix and b.

elemMult

public CTensor elemMult(CTensor b)

Computes the element-wise multiplication of two tensors.

Parameters:

b - Second tensor in the element-wise product.

Returns:

The element-wise product of this tensor and b.

Throws:

IllegalArgumentException - If !this.shape.equals(b.shape)

elemMult

public CooCTensor elemMult(CooCTensor b)

Computes the element-wise multiplication between two tensors.

Parameters:

b - Second tensor in the element-wise multiplication.

Returns:

The element-wise product of this matrix and b.

toCoo

public CooTensor toCoo(double estimatedSparsity)

Converts this tensor to an equivalent sparse COO tensor.

Parameters:

estimatedSparsity - Estimated sparsity of the tensor. Must be between 0 and 1 inclusive. If this is an accurate estimation it may provide a slight speedup and can reduce unneeded memory consumption. If memory is a concern, it is better to over-estimate the sparsity. If speed is the concern it is better to under-estimate the sparsity.

Returns:

A sparse COO tensor that is equivalent to this dense tensor.

See Also:

• toCoo(double)

toComplex

public CTensor toComplex()

Converts this tensor to an equivalent complex valued tensor.

Returns:

A complex tensor whose real components are the same as the data of this tensor and the imaginary components are zero.

mult

public CTensor mult(Complex128 b)

Multiplies a complex scalar value to each entry of this tensor.

Parameters:

b - Scalar value in product.

Returns:

The product of this tensor with b.

div

public CTensor div(Complex128 b)

Divides each entry of this tensor by a complex scalar value.

Parameters:

b - Scalar value in quotient.

Returns:

The tensor-scalar quotient of this tensor with b.

div

public CTensor div(CTensor b)

Computes the element-wise division of two tensors.

Parameters:

b - The second tensor in the element-wise quotient.

Returns:

The element-wise quotient of this tensor with b.

Throws:

IllegalArgumentException - If !this.shape.equals(b.shape)

equals

public boolean equals(Object object)

Checks if an object is equal to this tensor object.

Overrides:

equals in class Object

Parameters:

object - Object to check equality with this tensor.

Returns:

True if the two tensors have the same shape, are numerically equivalent, and are of type Tensor. False otherwise.

hashCode

public int hashCode()

Overrides:

hashCode in class Object

toString

public String toString()

Formats this tensor as a human-readable string. Specifically, a string containing the shape and flattened data of this tensor.

Overrides:

toString in class Object

Returns:

A human-readable string representing this tensor.