Package com.yahoo.tensor.functions

Class CompositeTensorFunction<NAMETYPE extends Name>

java.lang.Object

com.yahoo.tensor.functions.TensorFunction<NAMETYPE>

com.yahoo.tensor.functions.CompositeTensorFunction<NAMETYPE>

Direct Known Subclasses:

Argmax, Argmin, Diag, Expand, L1Normalize, L2Normalize, Matmul, Random, Range, ReduceJoin, Softmax, XwPlusB

public abstract class CompositeTensorFunction<NAMETYPE extends Name>
extends TensorFunction<NAMETYPE>

A composite tensor function is a tensor function which can be expressed (less tersely) as a tree of primitive tensor functions.

Author:

bratseth

Constructor Summary

Constructors

Constructor	Description
CompositeTensorFunction()

Method Summary

Modifier and Type	Method	Description
Tensor	evaluate(EvaluationContext<NAMETYPE> context)	Evaluates this by first converting it to a primitive function
TensorType	type(TypeContext<NAMETYPE> context)	Finds the type this produces by first converting it to a primitive function

Methods inherited from class com.yahoo.tensor.functions.TensorFunction

arguments, asScalarFunction, evaluate, hashCode, toPrimitive, toString, toString, withArguments, withTransformedFunctions

Methods inherited from class java.lang.Object

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

Constructor Details

CompositeTensorFunction

public CompositeTensorFunction()

Method Details

type

public TensorType type(TypeContext<NAMETYPE> context)

Finds the type this produces by first converting it to a primitive function

Specified by:

type in class TensorFunction<NAMETYPE extends Name>

Parameters:

context - a context which must be passed to all nexted functions when evaluating

evaluate

public Tensor evaluate(EvaluationContext<NAMETYPE> context)

Evaluates this by first converting it to a primitive function

Specified by:

evaluate in class TensorFunction<NAMETYPE extends Name>

Parameters:

context - a context which must be passed to all nested functions when evaluating