Class SymbolicGradient

java.lang.Object

org.tensorflow.op.RawOp

org.tensorflow.op.train.SymbolicGradient

All Implemented Interfaces:

Iterable<Operand<TType>>, Op

@Operator(group="train")
public final class SymbolicGradient
extends RawOp
implements Iterable<Operand<TType>>

Computes the gradient function for function f via backpropagation.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	SymbolicGradient.Inputs

Field Summary

Fields

Modifier and Type	Field	Description
static final String	OP_NAME	The name of this op, as known by TensorFlow core engine

Fields inherited from class RawOp

operation

Constructor Summary

Constructors

Constructor	Description
SymbolicGradient(Operation operation)

Method Summary

Modifier and Type	Method	Description
static SymbolicGradient	create(Scope scope, Iterable<Operand<?>> input, List<Class<? extends TType>> Tout, ConcreteFunction f)	Factory method to create a class wrapping a new SymbolicGradient operation.
Iterator<Operand<TType>>	iterator()
List<Output<?>>	output()	Gets output. a list of output tensors of size N;

Methods inherited from class RawOp

equals, hashCode, op, toString

Methods inherited from class Object

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

Methods inherited from interface Iterable

forEach, spliterator

Methods inherited from interface Op

env

Field Details

OP_NAME

public static final String OP_NAME

The name of this op, as known by TensorFlow core engine

See Also:

• Constant Field Values

Constructor Details

SymbolicGradient

public SymbolicGradient(Operation operation)

Method Details

create

@Endpoint(describeByClass=true)
public static SymbolicGradient create(Scope scope,
 Iterable<Operand<?>> input,
 List<Class<? extends TType>> Tout,
 ConcreteFunction f)

Factory method to create a class wrapping a new SymbolicGradient operation.

Parameters:

scope - current scope

input - a list of input tensors of size N + M;

Tout - the type list for the input list.

f - The function we want to compute the gradient for.

The function 'f' must be a numerical function which takes N inputs and produces M outputs. Its gradient function 'g', which is computed by this SymbolicGradient op is a function taking N + M inputs and produces N outputs.

I.e. if we have (y1, y2, ..., y_M) = f(x1, x2, ..., x_N), then, g is (dL/dx1, dL/dx2, ..., dL/dx_N) = g(x1, x2, ..., x_N, dL/dy1, dL/dy2, ..., dL/dy_M),

where L is a scalar-value function of (x1, x2, ..., xN) (e.g., the loss function). dL/dx_i is the partial derivative of L with respect to x_i.

(Needs some math expert to say the comment above better.)

Returns:

a new instance of SymbolicGradient

output

public List<Output<?>> output()

Gets output. a list of output tensors of size N;

Returns:

output.

iterator

public Iterator<Operand<TType>> iterator()

Specified by:

iterator in interface Iterable<Operand<TType>>