com.spotify.zoltar.tf

Class JTensor

java.lang.Object

com.spotify.zoltar.tf.JTensor

All Implemented Interfaces:

Serializable

public abstract class JTensor
extends Object
implements Serializable

Wrapper for Tensor that manages memory in JVM heap and closes the underlying Tensor.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
JTensor()

Method Summary

Modifier and Type	Method and Description
static JTensor	create(Tensor<?> tensor) Create a new JTensor instance by extracting data from the underlying Tensor and closing it afterwards.
protected abstract Object	data()
double[]	doubleValue() Double array value of the underlying Tensor, if DataType is DOUBLE.
float[]	floatValue() Float array value of the underlying Tensor, if DataType is FLOAT.
int[]	intValue() Integer array value of the underlying Tensor, if DataType is INT32.
long[]	longValue() Long array value of the underlying Tensor, if DataType is INT64.
String	stringValue() String value of the underlying Tensor, if DataType is STRING and numDimensions() is 0.
<T> T	value() Value of the underlying Tensor, lets the caller take care of typing.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

JTensor

public JTensor()

Method Detail

create

public static JTensor create(Tensor<?> tensor)

Create a new JTensor instance by extracting data from the underlying Tensor and closing it afterwards.

data

protected abstract Object data()

value

public <T> T value()

Value of the underlying Tensor, lets the caller take care of typing.

stringValue

public String stringValue()

String value of the underlying Tensor, if DataType is STRING and numDimensions() is 0.

intValue

public int[] intValue()

Integer array value of the underlying Tensor, if DataType is INT32.

longValue

public long[] longValue()

Long array value of the underlying Tensor, if DataType is INT64.

floatValue

public float[] floatValue()

Float array value of the underlying Tensor, if DataType is FLOAT.

doubleValue

public double[] doubleValue()

Double array value of the underlying Tensor, if DataType is DOUBLE.