ai.djl.training.hyperparameter.param.Hyperparameter Maven / Gradle / Ivy
/*
* Copyright 2020 Amazon.com, Inc. or its affiliates. All Rights Reserved.
*
* Licensed under the Apache License, Version 2.0 (the "License"). You may not use this file except in compliance
* with the License. A copy of the License is located at
*
* http://aws.amazon.com/apache2.0/
*
* or in the "license" file accompanying this file. This file 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 ai.djl.training.hyperparameter.param;
import ai.djl.training.hyperparameter.optimizer.HpOptimizer;
/**
* A class representing an input to the network that can't be differentiated.
*
* Some hyperparameters include learning rates, network sizes and shapes, activation choices, and
* model selection. In order to evaluate a set of hyperparameters, the only way is to fully train
* your model using those choices of hyperparameters. So, the full training loop involves training
* the model a number of times using different choices of hyperparameters. This can be mostly
* automated by using a {@link HpOptimizer}.
*
* @param the type of the hyperparameter
*/
public abstract class Hyperparameter {
protected String name;
/**
* Constructs a hyperparameter with the given name.
*
* @param name the name of the hyperparameter
*/
public Hyperparameter(String name) {
this.name = name;
}
/**
* Returns the name of the hyperparameter.
*
* @return the name of the hyperparameter
*/
public String getName() {
return name;
}
/**
* Returns a random value for the hyperparameter for a range of a fixed value if it is a {@link
* HpVal}.
*
* @return a random value for the hyperparameter for a range of a fixed value if it is a {@link
* HpVal}
*/
public abstract T random();
}