Tensor

Tensor (i.e., multi-dimensional array).

Tensors are the main data structure underlying all operations in TensorFlow. They represent multi-dimensional arrays of various data types (e.g., FLOAT32). Operations involving tensors can be of two types:

Eager: Operations directly executed on the tensor arguments, returning a new tensor. For example:

val a = Tensor(2.0, 4.5, 3.0, -1.2)
val b = Tensor(Tensor(0.2, 0.4), Tensor(-2.3, 5.0))
a.reshape(Shape(2, 2)) + b == Tensor(Tensor(2.2, 4.9), Tensor(0.7, 3.8))

Symbolic: Operations that need to be constructed as part of a computational Graph before being executing using a Session. For example:

val a = tf.placeholder(FLOAT64, Shape(4))               // Symbolic placeholder for value of a
val b = tf.placeholder(FLOAT64, Shape(2, 2))            // Symbolic placeholder for the value of b
val add = tf.reshape(a, Shape(2, 2)) + b                // Symbolic representation of the computation
val result = Session.run(
  feeds = Map(
    a -> Tensor(2.0, 4.5, 3.0, -1.2),
    b -> Tensor(Tensor(0.2, 0.4), Tensor(-2.3, 5.0))),
  fetches = add)                                        // Performs the actual computation
result == Tensor(Tensor(2.2, 4.9), Tensor(0.7, 3.8))

// TODO: [OPS] Update doc when we enrich op outputs similarly to tensors.

Linear Supertypes

Serializable, utilities.Closeable, TensorLike[D], AnyRef, Any

Instance Constructors

new Tensor(nativeHandleWrapper: NativeHandleWrapper, closeFn: () ⇒ Unit)

Attributes
protected

Value Members

final def !=(arg0: Any): Boolean

Definition Classes
AnyRef → Any
final def ##(): Int

Definition Classes
AnyRef → Any
final def ==(arg0: Any): Boolean

Definition Classes
AnyRef → Any
def apply(firstIndexer: core.Indexer, otherIndexers: core.Indexer*): Tensor[D]
final def asInstanceOf[T0]: T0

Definition Classes
Any
def clone(): AnyRef

Attributes
protected[java.lang]
Definition Classes
AnyRef
Annotations
@throws( ... )
def close(): Unit

Closes this Tensor and releases any resources associated with it.
Closes this Tensor and releases any resources associated with it. Note that an Tensor is not usable after it has been closed.

Definition Classes
Tensor → Closeable
val closeFn: () ⇒ Unit

Attributes
protected
Definition Classes
Tensor → Closeable
def copyToDevice(device: String): Tensor[D]

Returns a copy of this tensor on the provided device.
Returns a copy of this tensor on the provided device.
device
Device name. For example, "CPU:0", or "GPU:2".
def cpu(): Tensor[D]

Returns a copy of this tensor on the CPU.
val dataType: D

Data type of this tensor.
Data type of this tensor.

Definition Classes
Tensor → TensorLike
val device: String

Device in which the tensor is stored.
Device in which the tensor is stored.

Definition Classes
Tensor → TensorLike
def entriesIterator: Iterator[D.ScalaType]
final def eq(arg0: AnyRef): Boolean

Definition Classes
AnyRef
def equals(that: Any): Boolean

Definition Classes
Tensor → AnyRef → Any
def finalize(): Unit

Attributes
protected[java.lang]
Definition Classes
AnyRef
Annotations
@throws( classOf[java.lang.Throwable] )
final def getClass(): Class[_]

Definition Classes
AnyRef → Any
def gpu(gpuIndex: Int = 0): Tensor[D]

Returns a copy of this tensor on the GPU.
Returns a copy of this tensor on the GPU.
gpuIndex
Index of the GPU to use.
def hashCode(): Int

Definition Classes
Tensor → AnyRef → Any
final def isInstanceOf[T0]: Boolean

Definition Classes
Any
final def ne(arg0: AnyRef): Boolean

Definition Classes
AnyRef
final def notify(): Unit

Definition Classes
AnyRef
final def notifyAll(): Unit

Definition Classes
AnyRef
def rank: Int

Rank of this tensor (i.e., number of dimensions).
def scalar: D.ScalaType

Annotations
@throws( ... )
val shape: core.Shape

Shape of this tensor.
Shape of this tensor.

Definition Classes
Tensor → TensorLike
def size: Long

Number of elements contained in this tensor.
def slice(firstIndexer: core.Indexer, otherIndexers: core.Indexer*): Tensor[D]
def summarize(maxEntries: Int = 6, flattened: Boolean = false, includeInfo: Boolean = true): String

Returns a summary of the contents of this tensor.
Returns a summary of the contents of this tensor.
maxEntries
Maximum number of entries to show for each axis/dimension. If the size of an axis exceeds maxEntries, the output of that axis will be shortened to the first and last three elements. Defaults to 6. Values below 6 are ignored.
flattened
If true, the summary is flattened to one line. Otherwise, the summary may span multiple lines.
includeInfo
If true, the data type and the shape of the tensor are explicitly included in the summary. Otherwise, they are not.
returns
Tensor summary.
final def synchronized[T0](arg0: ⇒ T0): T0

Definition Classes
AnyRef
def toOutput: ops.Output
def toProto: TensorProto

Converts this object to its corresponding ProtoBuf object.
Converts this object to its corresponding ProtoBuf object.
returns
ProtoBuf object corresponding to this object.

Definition Classes
Tensor → Serializable
def toString(): String

Definition Classes
Tensor → AnyRef → Any
def toTensor: Tensor[D]

Returns this tensor.
Returns this tensor.

Definition Classes
Tensor → TensorLike
def toTensorIndexedSlices: TensorIndexedSlices[D]

Returns an TensorIndexedSlices that has the same value as this TensorLike.
Returns an TensorIndexedSlices that has the same value as this TensorLike.
returns
TensorIndexedSlices that has the same value as this TensorLike.

Definition Classes
Tensor → TensorLike
def toTensorProto: TensorProto

Constructs and returns a TensorProto object that represents this tensor.
Constructs and returns a TensorProto object that represents this tensor.
returns
Constructed TensorProto.
final def wait(): Unit

Definition Classes
AnyRef
Annotations
@throws( ... )
final def wait(arg0: Long, arg1: Int): Unit

Definition Classes
AnyRef
Annotations
@throws( ... )
final def wait(arg0: Long): Unit

Definition Classes
AnyRef
Annotations
@throws( ... )
def writeNPY(file: Path, fortranOrder: Boolean = false): Unit

Writes this tensor to the provided file, using the Numpy (i.e., .npy) file format.
Writes this tensor to the provided file, using the Numpy (i.e., .npy) file format. Note that this method will replace the file, if it already exists.

Related Docs: object Tensor | package tensors

class Tensor[+D <: types.DataType] extends TensorLike[D] with utilities.Closeable with Serializable

Instance Constructors

new Tensor(nativeHandleWrapper: NativeHandleWrapper, closeFn: () ⇒ Unit)

Value Members

final def !=(arg0: Any): Boolean

final def ##(): Int

final def ==(arg0: Any): Boolean

def apply(firstIndexer: core.Indexer, otherIndexers: core.Indexer*): Tensor[D]

final def asInstanceOf[T0]: T0

def clone(): AnyRef

def close(): Unit

val closeFn: () ⇒ Unit

def copyToDevice(device: String): Tensor[D]

def cpu(): Tensor[D]

val dataType: D

val device: String

def entriesIterator: Iterator[D.ScalaType]

final def eq(arg0: AnyRef): Boolean

def equals(that: Any): Boolean

def finalize(): Unit

final def getClass(): Class[_]

def gpu(gpuIndex: Int = 0): Tensor[D]

def hashCode(): Int

final def isInstanceOf[T0]: Boolean

final def ne(arg0: AnyRef): Boolean

final def notify(): Unit

final def notifyAll(): Unit

def rank: Int

def scalar: D.ScalaType

val shape: core.Shape

def size: Long

def slice(firstIndexer: core.Indexer, otherIndexers: core.Indexer*): Tensor[D]

def summarize(maxEntries: Int = 6, flattened: Boolean = false, includeInfo: Boolean = true): String

final def synchronized[T0](arg0: ⇒ T0): T0

def toOutput: ops.Output

def toProto: TensorProto

def toString(): String

def toTensor: Tensor[D]

def toTensorIndexedSlices: TensorIndexedSlices[D]

def toTensorProto: TensorProto

final def wait(): Unit

final def wait(arg0: Long, arg1: Int): Unit

final def wait(arg0: Long): Unit

def writeNPY(file: Path, fortranOrder: Boolean = false): Unit

Inherited from Serializable

Inherited from utilities.Closeable

Inherited from TensorLike[D]

Inherited from AnyRef

Inherited from Any

Ungrouped