ai.h2o.automl.modeling.DeepLearningStepsProvider Maven / Gradle / Ivy
The newest version!
package ai.h2o.automl.modeling;
import ai.h2o.automl.*;
import ai.h2o.automl.preprocessing.PreprocessingConfig;
import ai.h2o.automl.preprocessing.TargetEncoding;
import hex.deeplearning.DeepLearningModel;
import hex.deeplearning.DeepLearningModel.DeepLearningParameters;
import java.util.HashMap;
import java.util.Map;
public class DeepLearningStepsProvider
implements ModelingStepsProvider
, ModelParametersProvider {
public static class DeepLearningSteps extends ModelingSteps {
static final String NAME = Algo.DeepLearning.name();
static abstract class DeepLearningModelStep extends ModelingStep.ModelStep {
public DeepLearningModelStep(String id, AutoML autoML) {
super(NAME, Algo.DeepLearning, id, autoML);
}
@Override
protected PreprocessingConfig getPreprocessingConfig() {
//TE useless for DNN
PreprocessingConfig config = super.getPreprocessingConfig();
config.put(TargetEncoding.CONFIG_PREPARE_CV_ONLY, aml().isCVEnabled());
return config;
}
}
static abstract class DeepLearningGridStep extends ModelingStep.GridStep {
DeepLearningGridStep(String id, AutoML autoML) {
super(NAME, Algo.DeepLearning, id, autoML);
}
public DeepLearningParameters prepareModelParameters() {
DeepLearningParameters params = new DeepLearningParameters();
params._epochs = 10000; // early stopping takes care of epochs - no need to tune!
params._adaptive_rate = true;
params._activation = DeepLearningParameters.Activation.RectifierWithDropout;
return params;
}
@Override
protected PreprocessingConfig getPreprocessingConfig() {
//TE useless for DNN
PreprocessingConfig config = super.getPreprocessingConfig();
config.put(TargetEncoding.CONFIG_PREPARE_CV_ONLY, aml().isCVEnabled());
return config;
}
public Map prepareSearchParameters() {
Map searchParams = new HashMap<>();
searchParams.put("_rho", new Double[] { 0.9, 0.95, 0.99 });
searchParams.put("_epsilon", new Double[] { 1e-6, 1e-7, 1e-8, 1e-9 });
searchParams.put("_input_dropout_ratio", new Double[] { 0.0, 0.05, 0.1, 0.15, 0.2 });
return searchParams;
}
}
private final ModelingStep[] defaults = new DeepLearningModelStep[] {
new DeepLearningModelStep("def_1", aml()) {
@Override
public DeepLearningParameters prepareModelParameters() {
DeepLearningParameters params = new DeepLearningParameters(); // don't use common params for default DL
params._hidden = new int[]{ 10, 10, 10 };
return params;
}
},
};
private final ModelingStep[] grids = new DeepLearningGridStep[] {
new DeepLearningGridStep("grid_1", aml()) {
@Override
public Map prepareSearchParameters() {
Map searchParams = super.prepareSearchParameters();
searchParams.put("_hidden", new Integer[][] {
{ 20 },
{ 50 },
{ 100 }
});
searchParams.put("_hidden_dropout_ratios", new Double[][] {
{ 0.0 },
{ 0.1 },
{ 0.2 },
{ 0.3 },
{ 0.4 },
{ 0.5 }
});
return searchParams;
}
},
new DeepLearningGridStep("grid_2", aml()) {
@Override
public Map prepareSearchParameters() {
Map searchParams = super.prepareSearchParameters();
searchParams.put("_hidden", new Integer[][] {
{ 20, 20 },
{ 50, 50 },
{ 100, 100 }
});
searchParams.put("_hidden_dropout_ratios", new Double[][] {
{ 0.0, 0.0 },
{ 0.1, 0.1 },
{ 0.2, 0.2 },
{ 0.3, 0.3 },
{ 0.4, 0.4 },
{ 0.5, 0.5 }
});
return searchParams;
}
},
new DeepLearningGridStep("grid_3", aml()) {
@Override
public Map prepareSearchParameters() {
Map searchParams = super.prepareSearchParameters();
searchParams.put("_hidden", new Integer[][] {
{ 20, 20, 20 },
{ 50, 50, 50 },
{ 100, 100, 100 }
});
searchParams.put("_hidden_dropout_ratios", new Double[][] {
{ 0.0, 0.0, 0.0 },
{ 0.1, 0.1, 0.1 },
{ 0.2, 0.2, 0.2 },
{ 0.3, 0.3, 0.3 },
{ 0.4, 0.4, 0.4 },
{ 0.5, 0.5, 0.5 }
});
return searchParams;
}
},
};
public DeepLearningSteps(AutoML autoML) {
super(autoML);
}
@Override
public String getProvider() {
return NAME;
}
@Override
protected ModelingStep[] getDefaultModels() {
return defaults;
}
@Override
protected ModelingStep[] getGrids() {
return grids;
}
}
@Override
public String getName() {
return DeepLearningSteps.NAME;
}
@Override
public DeepLearningSteps newInstance(AutoML aml) {
return new DeepLearningSteps(aml);
}
@Override
public DeepLearningParameters newDefaultParameters() {
return new DeepLearningParameters();
}
}