Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance. Project price only 1 $
You can buy this project and download/modify it how often you want.
/*-
*
* * Copyright 2015 Skymind,Inc.
* *
* * Licensed under the Apache License, Version 2.0 (the "License");
* * you may not use this file except in compliance with the License.
* * You may obtain a copy of the License at
* *
* * http://www.apache.org/licenses/LICENSE-2.0
* *
* * Unless required by applicable law or agreed to in writing, software
* * distributed under the License is distributed on an "AS IS" BASIS,
* * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* * See the License for the specific language governing permissions and
* * limitations under the License.
*
*
*/
package org.nd4j.linalg.api.ops;
import org.nd4j.linalg.api.complex.IComplexNumber;
import org.nd4j.linalg.api.ndarray.INDArray;
/**
* An accumulation is an op that given:
* x -> the origin ndarray
* y -> the pairwise ndarray
* n -> the number of times to accumulate
*
*
* Of note here in the extra arguments.
*
* An accumulation (or reduction in some terminology)
* has a concept of a starting value.
*
* The starting value is the initialization of the solution
* to the operation.
*
* An accumulation should always have the extraArgs()
* contain the zero value as the first value.
*
* This allows the architecture to generalize to different backends
* and gives the implementer of a backend a way of hooking in to
* passing parameters to different engines.
*
* Note that Accumulation op implementations should be stateless
* (other than the final result and x/y/z/n arguments) and hence threadsafe,
* such that they may be parallelized using the update, combineSubResults and
* set/getFinalResults methods.
* @author Adam Gibson
*/
public interface Accumulation extends Op {
/**
* Returns the no op version
* of the input
* Basically when a reduce can't happen (eg: sum(0) on a row vector)
* you have a no op state for a given reduction.
* For most accumulations, this should return x
* but certain transformations should return say: the absolute value
*
*
* @return the no op version of the input
*/
INDArray noOp();
/**
* Setter for final transform
* @param applyFinalTransform
*/
void setApplyFinalTransform(boolean applyFinalTransform);
/**
* Whether to apply the final
* transform or not
* @return
*/
boolean applyFinalTransform();
/** Do one accumulation update for a single-argument accumulation, given the
* current accumulation value and another value to be processed/accumulated
* @param accum The current accumulation value
* @param x The next/new value to be processed/accumulated
* @return The updated accumulation value
*/
double update(double accum, double x);
/** Do an accumulation update for a pair-wise (op(x,y)) accumulation, given the
* current accumulation value and a pair of values to be processed/accumulated
* @param accum The current accumulation value
* @param x The next/new x value to be processed/accumulated
* @param y The next/new y value to be processed/accumulated
* @return the updated accumulation value
*/
double update(double accum, double x, double y);
/**@see #update(double, double) */
float update(float accum, float x);
/**@see #update(double, double, double)*/
float update(float accum, float x, float y);
/** Complex update.
* @see #update(double, double)
*/
IComplexNumber update(IComplexNumber accum, double x);
/** Complex update.
* @see #update(double, double, double)
*/
IComplexNumber update(IComplexNumber accum, double x, double y);
/** Complex update.
* @see #update(double, double)
*/
IComplexNumber update(IComplexNumber accum, IComplexNumber x);
/** Complex update.
* @see #update(double, double, double)
*/
IComplexNumber update(IComplexNumber accum, IComplexNumber x, IComplexNumber y);
/** Complex update.
* @see #update(double, double, double)
*/
IComplexNumber update(IComplexNumber accum, IComplexNumber x, double y);
/** Combine sub-results, when the Accumulation operation is split and executed in
* parallel.
* Sometimes this is the same as the update operation (i.e., in sum operation), but often
* it is not (for example, in L2 norm, sqrt(1/n * sum_i x_i^2).
* @param first The first sub-value to be combined
* @param second The second sub-value to be combined
* @return The combined result
*/
double combineSubResults(double first, double second);
/**@see #combineSubResults(double, double) */
float combineSubResults(float first, float second);
/**@see #combineSubResults(double, double) */
IComplexNumber combineSubResults(IComplexNumber first, IComplexNumber second);
/** Get and set the final result.
* The final result is also stored in the Accumulation op itself (and can be retrieved
* by getFinalResult())
* In some Accumulation operations, a final operation must be done on the accumulated
* result. For example, division by n in a mean operation: the accumulated result is
* the sum, whereas the final result is the sum/n.
* @param accum The accumulated result
* @return The final result
*/
double getAndSetFinalResult(double accum);
/** @see #getAndSetFinalResult(double) */
float getAndSetFinalResult(float accum);
/** Complex version of getAndSetFinalResult().
* @see #getAndSetFinalResult(double)
*/
IComplexNumber getAndSetFinalResult(IComplexNumber accum);
/** Get the final result (may return null if getAndSetFinalResult has not
* been called, or for accumulation ops on complex arrays)
*/
Number getFinalResult();
/** Essentially identical to {@link #getFinalResult()}, maintained to avoid breaking
* implementations created before Accumulation ops were refactored.
* May be removed in a future release.
*/
@Deprecated
Number currentResult();
IComplexNumber getFinalResultComplex();
void setFinalResult(Number number);
void setFinalResultComplex(IComplexNumber number);
/** Calculate the final result. unlike {@link #getAndSetFinalResult(double)}, the result is
* merely calculated and returned, not stored in the Accumulation op itself.
* @param accum The accumulated result
* @param n the number of elements accumulated
* @return calculated final result
*/
double calculateFinalResult(double accum, long n);
/** @see #calculateFinalResult(double, long) */
float calculateFinalResult(float accum, long n);
IComplexNumber op(IComplexNumber origin, double other);
IComplexNumber op(IComplexNumber origin, float other);
IComplexNumber op(IComplexNumber origin, IComplexNumber other);
IComplexNumber op(IComplexNumber origin);
/**Initial value (used to initialize the accumulation op)
* @return the initial value
*/
double zeroDouble();
/** Initial value (used to initialize the accumulation op) */
float zeroFloat();
/**
* Initial value for half
* @return
*/
float zeroHalf();
/**Complex initial value
*@return the complex initial value
*/
IComplexNumber zeroComplex();
/**
* This method is only used for Distance functions
* @return
*/
boolean isComplexAccumulation();
}
