com.simiacryptus.mindseye.layers.java.LinearActivationLayer Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of mindseye-java Show documentation
Pure Java Neural Networks Components
There is a newer version: 2.1.0
/*
 * Copyright (c) 2019 by Andrew Charneski.
 *
 * The author licenses this file to you 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 com.simiacryptus.mindseye.layers.java;

import com.google.gson.JsonObject;
import com.simiacryptus.mindseye.lang.*;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import javax.annotation.Nonnull;
import javax.annotation.Nullable;
import java.util.Arrays;
import java.util.List;
import java.util.Map;
import java.util.UUID;
import java.util.stream.IntStream;

/**
 * A tunable linear (y=A*x+B) function, whose parameters can participate in learning. Defaults to y=1*x+0, and is NOT
 * frozen by default.
 */
@SuppressWarnings("serial")
public class LinearActivationLayer extends LayerBase {


  @SuppressWarnings("unused")
  private static final Logger log = LoggerFactory.getLogger(LinearActivationLayer.class);
  @Nullable
  private final Tensor weights;

  /**
   * Instantiates a new Linear activation key.
   */
  public LinearActivationLayer() {
    super();
    weights = new Tensor(2);
    weights.set(0, 1.);
    weights.set(1, 0.);
  }

  /**
   * Instantiates a new Linear activation key.
   *
   * @param json      the json
   * @param resources the resources
   */
  protected LinearActivationLayer(@Nonnull final JsonObject json, Map resources) {
    super(json);
    weights = Tensor.fromJson(json.get("weights"), resources);
  }

  /**
   * From json linear activation key.
   *
   * @param json the json
   * @param rs   the rs
   * @return the linear activation key
   */
  public static LinearActivationLayer fromJson(@Nonnull final JsonObject json, Map rs) {
    return new LinearActivationLayer(json, rs);
  }

  @Override
  protected void _free() {
    weights.freeRef();
    super._free();
  }

  @Nonnull
  @Override
  public Result eval(final Result... inObj) {
    final Result in0 = inObj[0];
    final TensorList inData = in0.getData();
    in0.addRef();
    inData.addRef();
    final int itemCnt = inData.length();
    final double scale = weights.get(0);
    final double bias = weights.get(1);
    weights.addRef();
    return new Result(TensorArray.wrap(IntStream.range(0, itemCnt).mapToObj(dataIndex -> {
      @Nullable final Tensor input = inData.get(dataIndex);
      @Nullable Tensor map = input.map(v -> scale * v + bias);
      input.freeRef();
      return map;
    }).toArray(i -> new Tensor[i])), (@Nonnull final DeltaSet buffer, @Nonnull final TensorList delta) -> {
      if (!isFrozen()) {
        IntStream.range(0, delta.length()).forEach(dataIndex -> {
          @Nullable Tensor deltaT = delta.get(dataIndex);
          @Nullable Tensor inputT = inData.get(dataIndex);
          @Nullable final double[] deltaData = deltaT.getData();
          @Nullable final double[] inputData = inputT.getData();
          @Nonnull final Tensor weightDelta = new Tensor(weights.getDimensions());
          for (int i = 0; i < deltaData.length; i++) {
            weightDelta.add(0, deltaData[i] * inputData[inputData.length == 1 ? 0 : i]);
            weightDelta.add(1, deltaData[i]);
          }
          buffer.get(LinearActivationLayer.this.getId(), weights.getData()).addInPlace(weightDelta.getData()).freeRef();
          inputT.freeRef();
          deltaT.freeRef();
          weightDelta.freeRef();
        });
      }
      if (in0.isAlive()) {
        @Nonnull final TensorList tensorList = TensorArray.wrap(IntStream.range(0, delta.length()).mapToObj(dataIndex -> {
          @Nullable Tensor tensor = delta.get(dataIndex);
          @Nullable final double[] deltaData = tensor.getData();
          @Nonnull final Tensor passback = new Tensor(inData.getDimensions());
          for (int i = 0; i < passback.length(); i++) {
            passback.set(i, deltaData[i] * weights.getData()[0]);
          }
          tensor.freeRef();
          return passback;
        }).toArray(i -> new Tensor[i]));
        in0.accumulate(buffer, tensorList);
      }
    }) {

      @Override
      public boolean isAlive() {
        return in0.isAlive() || !isFrozen();
      }

      @Override
      protected void _free() {
        weights.freeRef();
        inData.freeRef();
        in0.freeRef();
      }

    };
  }

  /**
   * Gets bias.
   *
   * @return the bias
   */
  public double getBias() {
    return weights.get(1);
  }

  /**
   * Sets bias.
   *
   * @param bias the bias
   * @return the bias
   */
  @Nonnull
  public LinearActivationLayer setBias(final double bias) {
    weights.set(1, bias);
    return this;
  }

  @Nonnull
  @Override
  public JsonObject getJson(Map resources, @Nonnull DataSerializer dataSerializer) {
    @Nonnull final JsonObject json = super.getJsonStub();
    json.add("weights", weights.toJson(resources, dataSerializer));
    return json;
  }

  /**
   * Gets scale.
   *
   * @return the scale
   */
  public double getScale() {
    return weights.get(0);
  }

  /**
   * Sets scale.
   *
   * @param scale the scale
   * @return the scale
   */
  @Nonnull
  public LinearActivationLayer setScale(final double scale) {
    weights.set(0, scale);
    return this;
  }

  @Nonnull
  @Override
  public List state() {
    return Arrays.asList(weights.getData());
  }

}