TString

public interface TString

String type.

This type can be used to store any arbitrary byte sequence of variable length.

Since the size of a tensor is fixed, creating a tensor of this type requires to provide all of its values initially, so TensorFlow can compute and allocate the right amount of memory. Then the data in the tensor is initialized once and cannot be modified afterwards.

Public Methods

abstract NdArray<byte[]>	asBytes()
abstract static TString	scalarOf(String value) Allocates a new tensor for storing a string scalar.
abstract static TString	tensorOf(Shape shape, DataBuffer<String> data) Allocates a new tensor with the given shape and data.
abstract static TString	tensorOf(NdArray<String> src) Allocates a new tensor which is a copy of a given array.
abstract static TString	tensorOf(Charset charset, Shape shape, DataBuffer<String> data) Allocates a new tensor with the given shape and data.
abstract static TString	tensorOf(Charset charset, NdArray<String> src) Allocates a new tensor which is a copy of a given array.
abstract static TString	tensorOfBytes(Shape shape, DataBuffer<byte[]> data) Allocates a new tensor with the given shape and raw bytes.
abstract static TString	tensorOfBytes(NdArray<byte[]> src) Allocates a new tensor which is a copy of a given array of raw bytes.
abstract TString	using(Charset charset) Use a specific charset for decoding data from a string tensor, instead of the default UTF-8.
abstract static TString	vectorOf(String... values) Allocates a new tensor for storing a vector of strings.

Inherited Methods

From interface org.tensorflow.ndarray.NdArray

abstract NdArray<String>	copyTo(NdArray<String> dst) Copy the content of this array to the destination array.
abstract NdArraySequence<? extends NdArray<T>>	elements(int dimensionIdx) Returns a sequence of all elements at a given dimension.
abstract boolean	equals(Object obj) Checks equality between n-dimensional arrays.
abstract NdArray<String>	get(long... coordinates) Returns the N-dimensional element of this array at the given coordinates.
abstract String	getObject(long... coordinates) Returns the value of the scalar found at the given coordinates.
abstract NdArray<String>	read(DataBuffer<String> dst) Read the content of this N-dimensional array into the destination buffer.
abstract NdArraySequence<? extends NdArray<T>>	scalars() Returns a sequence of all scalars in this array.
abstract NdArray<String>	set(NdArray<String> src, long... coordinates) Assigns the value of the N-dimensional element found at the given coordinates.
abstract NdArray<String>	setObject(String value, long... coordinates) Assigns the value of the scalar found at the given coordinates.
abstract NdArray<String>	slice(Index... indices) Creates a multi-dimensional view (or slice) of this array by mapping one or more dimensions to the given index selectors.
abstract NdArray<String>	write(DataBuffer<String> src) Write the content of this N-dimensional array from the source buffer.

From interface org.tensorflow.types.family.TType

abstract void	close() Release resources associated with the Tensor.
abstract DataType	dataType() Returns the DataType of elements stored in the tensor.
abstract long	numBytes() Returns the size, in bytes, of the tensor data.
abstract Class<? extends TType>	type() Returns the type of this tensor as a registered subclass of TType

From interface org.tensorflow.ndarray.Shaped

abstract int	rank()
abstract Shape	shape()
abstract long	size() Computes and returns the total size of this container, in number of values.

From interface org.tensorflow.Tensor

abstract RawTensor	asRawTensor() Returns a raw (untyped) representation of this tensor
abstract void	close() Release resources associated with the Tensor.
abstract DataType	dataType() Returns the DataType of elements stored in the tensor.
abstract long	numBytes() Returns the size, in bytes, of the tensor data.
abstract static <T extends TType> T	of(Class<T> type, Shape shape, long size, Consumer<T> dataInitializer) Allocates a tensor of a given datatype, shape and size.
abstract static <T extends TType> T	of(Class<T> type, Shape shape) Allocates a tensor of a given datatype and shape.
abstract static <T extends TType> T	of(Class<T> type, Shape shape, long size) Allocates a tensor of a given datatype, shape and size.
abstract static <T extends TType> T	of(Class<T> type, Shape shape, Consumer<T> dataInitializer) Allocates and initialize a tensor of a given datatype and shape.
abstract static <T extends TType> T	of(Class<T> type, Shape shape, ByteDataBuffer rawData) Creates a Tensor of any type from the raw data provided by the given buffer.
abstract Shape	shape() Returns the shape of the tensor.

From interface java.lang.AutoCloseable

abstract void

close()

Public Methods

public abstract NdArray<byte[]> asBytes ()

Returns

• the tensor data as a n-dimensional array of raw byte sequences.

public static abstract TString scalarOf (String value)

Allocates a new tensor for storing a string scalar.

The string is encoded into bytes using the UTF-8 charset.

Parameters

value	scalar value to store in the new tensor

Returns

• the new tensor

public static abstract TString tensorOf (Shape shape, DataBuffer<String> data)

Allocates a new tensor with the given shape and data.

The data will be copied from the provided buffer to the tensor after it is allocated. The strings are encoded into bytes using the UTF-8 charset.

Parameters

shape	shape of the tensor
data	buffer of strings to initialize the tensor with

Returns

• the new tensor

public static abstract TString tensorOf (NdArray<String> src)

Allocates a new tensor which is a copy of a given array.

The tensor will have the same shape as the source array and its data will be copied. The strings are encoded into bytes using the UTF-8 charset.

Parameters

src	the source array giving the shape and data to the new tensor

Returns

• the new tensor

public static abstract TString tensorOf (Charset charset, Shape shape, DataBuffer<String> data)

Allocates a new tensor with the given shape and data.

The data will be copied from the provided buffer to the tensor after it is allocated. The strings are encoded into bytes using the charset passed in parameter.

If charset is different than default UTF-8, then it must also be provided explicitly when reading data from the tensor, using using(Charset):

// Given `originalStrings` an initialized buffer of strings
 TString tensor =
    TString.tensorOf(Charsets.UTF_16, Shape.of(originalString.size()), originalStrings);
 ...
 TString tensorStrings = tensor.data().using(Charsets.UTF_16);
 assertEquals(originalStrings.getObject(0), tensorStrings.getObject(0));
 

Parameters

charset	charset to use for encoding the strings into bytes
shape	shape of the tensor
data	buffer of strings to initialize the tensor with

Returns

• the new tensor

public static abstract TString tensorOf (Charset charset, NdArray<String> src)

Allocates a new tensor which is a copy of a given array.

The tensor will have the same shape as the source array and its data will be copied. The strings are encoded into bytes using the charset passed in parameter.

If charset is different than default UTF-8, then it must also be provided explicitly when reading data from the tensor, using using(Charset):

// Given `originalStrings` an initialized vector of strings
 TString tensor = TString.tensorOf(Charsets.UTF_16, originalStrings);
 ...
 TString tensorStrings = tensor.data().using(Charsets.UTF_16);
 assertEquals(originalStrings.getObject(0), tensorStrings.getObject(0));
 

Parameters

charset	charset to use for encoding the strings into bytes
src	the source array giving the shape and data to the new tensor

Returns

• the new tensor

public static abstract TString tensorOfBytes (Shape shape, DataBuffer<byte[]> data)

Allocates a new tensor with the given shape and raw bytes.

The data will be copied from the provided buffer to the tensor after it has been allocated.

If data must be read as raw bytes as well, the user must specify it explicitly by invoking asBytes() on the returned data:

byte[] bytes = tensor.data().asBytes().getObject(0);  // returns first sequence of bytes in the tensor
 

Parameters

shape	shape of the tensor to create
data	the source array giving the shape and data to the new tensor

Returns

• the new tensor

public static abstract TString tensorOfBytes (NdArray<byte[]> src)

Allocates a new tensor which is a copy of a given array of raw bytes.

The tensor will have the same shape as the source array and its data will be copied.

If data must be read as raw bytes as well, the user must specify it explicitly by invoking asBytes() on the returned data:

byte[] bytes = tensor.data().asBytes().getObject(0);  // returns first sequence of bytes in the tensor
 

Parameters

src	the source array giving the shape and data to the new tensor

Returns

• the new tensor

public abstract TString using (Charset charset)

Use a specific charset for decoding data from a string tensor, instead of the default UTF-8.

The charset must match the one used for encoding the string values when the tensor was created. For example:

TString tensor =
    TString.tensorOf(StandardCharsets.UTF_16, NdArrays.scalarOfObject("TensorFlow");

 assertEquals("TensorFlow", tensor.data().using(StandardCharsets.UTF_16).getObject());
 

Parameters

charset	charset to use

Returns

• string tensor data using this charset

public static abstract TString vectorOf (String... values)

Allocates a new tensor for storing a vector of strings.

The strings are encoded into bytes using the UTF-8 charset.

Parameters

values	values to store in the new tensor

Returns

• the new tensor