org.nd4j.linalg.api.ops.impl.transforms.strict.ELU Maven / Gradle / Ivy

Show more of this group Show more artifacts with this name
Show all versions of nd4j-api Show documentation
There is a newer version: 1.0.0-M2.1
/*******************************************************************************
 * Copyright (c) 2015-2018 Skymind, Inc.
 *
 * This program and the accompanying materials are made available under the
 * terms of the Apache License, Version 2.0 which is available at
 * https://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.
 *
 * SPDX-License-Identifier: Apache-2.0
 ******************************************************************************/

package org.nd4j.linalg.api.ops.impl.transforms.strict;

import org.nd4j.autodiff.samediff.SDVariable;
import org.nd4j.autodiff.samediff.SameDiff;
import org.nd4j.base.Preconditions;
import org.nd4j.imports.NoOpNameFoundException;
import org.nd4j.linalg.api.buffer.DataType;
import org.nd4j.linalg.api.ndarray.INDArray;
import org.nd4j.linalg.api.ops.DynamicCustomOp;

import java.util.Collections;
import java.util.List;

/**
 * ELU: Exponential Linear Unit (alpha=1.0)

 * Introduced in paper:

 * Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)

 * Djork-Arné Clevert, Thomas Unterthiner, Sepp Hochreiter (2015)

 * http://arxiv.org/abs/1511.07289
 *
 * @author Alex Black
 */
public class ELU extends DynamicCustomOp {
    public static final double DEFAULT_ALPHA = 1.0;

    protected double alpha;

    public ELU(SameDiff sameDiff, SDVariable i_v) {
        super(sameDiff, new SDVariable[]{i_v});
        this.alpha = DEFAULT_ALPHA;
        addTArgument(alpha);
    }

    public ELU() {
    }

    public ELU(INDArray x, INDArray z) {
        this(x, z, DEFAULT_ALPHA);
    }

    public ELU(INDArray x, INDArray z, double alpha) {
        super(null, wrapOrNull(x), wrapOrNull(z));
        this.alpha = alpha;
        addTArgument(alpha);
    }

    public ELU(INDArray x) {
        this(x, null, DEFAULT_ALPHA);
    }

    @Override
    public String opName() {
        return "elu";
    }

    @Override
    public String onnxName() {
        throw new NoOpNameFoundException("No onnx op opName found for " + opName());
    }

    @Override
    public String tensorflowName() {
        return "Elu";
    }

    @Override
    public List doDiff(List i_v) {
        //ELU: e^x-1 if x<0, x otherwise
        //dL/dIn = dL/Out * dOut/dIn
        return Collections.singletonList(f().eluBp(arg(), i_v.get(0), alpha));
    }

    @Override
    public List calculateOutputDataTypes(List dataTypes) {
        Preconditions.checkState(dataTypes != null && dataTypes.size() == 1, "Expected exactly 1 datatype for ELU, got %s", dataTypes);
        Preconditions.checkState(dataTypes.get(0).isFPType(), "Expected floating point input type for ELU, got %s", dataTypes);

        return dataTypes;
    }
}