DeepWaterParametersV3 (h2o-bindings version 3.18.0.8 API)

java.lang.Object
- water.bindings.pojos.SchemaV3
- - water.bindings.pojos.ModelParametersSchemaV3
  - - water.bindings.pojos.DeepWaterParametersV3

public class DeepWaterParametersV3
extends ModelParametersSchemaV3

Field Summary

Fields
Modifier and Type	Field and Description
`DeepWaterParametersActivation`	`activation` Activation function.
`boolean`	`autoencoder` Auto-Encoder.
`DeepWaterParametersBackend`	`backend` Deep Learning Backend.
`boolean`	`balanceClasses` Balance training data class counts via over/under-sampling (for imbalanced data).
`boolean`	`cacheData` Whether to cache the data in memory (automatically disabled if data size is too large).
`int`	`channels` Number of (color) channels.
`double`	`classificationStop` Stopping criterion for classification error fraction on training data (-1 to disable).
`float[]`	`classSamplingFactors` Desired over/under-sampling ratios per class (in lexicographic order).
`double`	`clipGradient` Clip gradients once their absolute value is larger than this value.
`int[]`	`deviceId` Device IDs (which GPUs to use).
`boolean`	`diagnostics` Enable diagnostics for hidden layers.
`double`	`epochs` How many times the dataset should be iterated (streamed), can be fractional.
`java.lang.String`	`exportNativeParametersPrefix` Path (prefix) where to export the native model parameters after every iteration.
`boolean`	`gpu` Whether to use a GPU (if available).
`int[]`	`hidden` Hidden layer sizes (e.g.
`double[]`	`hiddenDropoutRatios` Hidden layer dropout ratios (can improve generalization), specify one value per hidden layer, defaults to 0.5.
`int[]`	`imageShape` Width and height of image.
`double`	`inputDropoutRatio` Input layer dropout ratio (can improve generalization, try 0.1 or 0.2).
`double`	`learningRate` Learning rate (higher => less stable, lower => slower convergence).
`double`	`learningRateAnnealing` Learning rate annealing: rate / (1 + rate_annealing * samples).
`float`	`maxAfterBalanceSize` Maximum relative size of the training data after balancing class counts (can be less than 1.0).
`int`	`maxConfusionMatrixSize` [Deprecated] Maximum size (# classes) for confusion matrices to be printed in the Logs.
`int`	`maxHitRatioK` Max.
`java.lang.String`	`meanImageFile` Path of file containing the mean image data for data normalization.
`int`	`miniBatchSize` Mini-batch size (smaller leads to better fit, larger can speed up and generalize better).
`double`	`momentumRamp` Number of training samples for which momentum increases.
`double`	`momentumStable` Final momentum after the ramp is over (try 0.99).
`double`	`momentumStart` Initial momentum at the beginning of training (try 0.5).
`DeepWaterParametersNetwork`	`network` Network architecture.
`java.lang.String`	`networkDefinitionFile` Path of file containing network definition (graph, architecture).
`java.lang.String`	`networkParametersFile` Path of file containing network (initial) parameters (weights, biases).
`boolean`	`overwriteWithBestModel` If enabled, override the final model with the best model found during training.
`DeepWaterParametersProblemType`	`problemType` Problem type, auto-detected by default.
`boolean`	`quietMode` Enable quiet mode for less output to standard output.
`double`	`regressionStop` Stopping criterion for regression error (MSE) on training data (-1 to disable).
`boolean`	`replicateTrainingData` Replicate the entire training dataset onto every node for faster training on small datasets.
`double`	`scoreDutyCycle` Maximum duty cycle fraction for scoring (lower: more training, higher: more scoring).
`double`	`scoreInterval` Shortest time interval (in seconds) between model scoring.
`long`	`scoreTrainingSamples` Number of training set samples for scoring (0 for all).
`long`	`scoreValidationSamples` Number of validation set samples for scoring (0 for all).
`long`	`seed` Seed for random numbers (affects sampling) - Note: only reproducible when running single threaded.
`boolean`	`shuffleTrainingData` Enable global shuffling of training data.
`boolean`	`singleNodeMode` Run on a single node for fine-tuning of model parameters.
`boolean`	`sparse` Sparse data handling (more efficient for data with lots of 0 values).
`boolean`	`standardize` If enabled, automatically standardize the data.
`double`	`targetRatioCommToComp` Target ratio of communication overhead to computation.
`long`	`trainSamplesPerIteration` Number of training samples (globally) per MapReduce iteration.
`boolean`	`variableImportances` Compute variable importances for input features (Gedeon method) - can be slow for large networks.

Constructor Summary

Constructors
Constructor and Description

DeepWaterParametersV3()
Public constructor

Constructors
Constructor and Description
`DeepWaterParametersV3()` Public constructor

Method Summary

All Methods Instance Methods Concrete Methods
Modifier and Type Method and Description

java.lang.String toString()
Return the contents of this object as a JSON String.
- Methods inherited from class java.lang.Object
  clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

All Methods Instance Methods Concrete Methods
Modifier and Type	Method and Description
`java.lang.String`	`toString()` Return the contents of this object as a JSON String.

- Field Detail
  - problemType
```
public DeepWaterParametersProblemType problemType
```
    Problem type, auto-detected by default. If set to image, the H2OFrame must contain a string column containing the path (URI or URL) to the images in the first column. If set to text, the H2OFrame must contain a string column containing the text in the first column. If set to dataset, Deep Water behaves just like any other H2O Model and builds a model on the provided H2OFrame (non-String columns).
  - activation
```
public DeepWaterParametersActivation activation
```
    Activation function. Only used if no user-defined network architecture file is provided, and only for problem_type=dataset.
  - hidden
```
public int[] hidden
```
    Hidden layer sizes (e.g. [200, 200]). Only used if no user-defined network architecture file is provided, and only for problem_type=dataset.
  - inputDropoutRatio
```
public double inputDropoutRatio
```
    Input layer dropout ratio (can improve generalization, try 0.1 or 0.2).
  - hiddenDropoutRatios
```
public double[] hiddenDropoutRatios
```
    Hidden layer dropout ratios (can improve generalization), specify one value per hidden layer, defaults to 0.5.
  - maxConfusionMatrixSize
```
public int maxConfusionMatrixSize
```
    [Deprecated] Maximum size (# classes) for confusion matrices to be printed in the Logs.
  - sparse
```
public boolean sparse
```
    Sparse data handling (more efficient for data with lots of 0 values).
  - maxHitRatioK
```
public int maxHitRatioK
```
    Max. number (top K) of predictions to use for hit ratio computation (for multi-class only, 0 to disable).
  - epochs
```
public double epochs
```
    How many times the dataset should be iterated (streamed), can be fractional.
  - trainSamplesPerIteration
```
public long trainSamplesPerIteration
```
    Number of training samples (globally) per MapReduce iteration. Special values are 0: one epoch, -1: all available data (e.g., replicated training data), -2: automatic.
  - targetRatioCommToComp
```
public double targetRatioCommToComp
```
    Target ratio of communication overhead to computation. Only for multi-node operation and train_samples_per_iteration = -2 (auto-tuning).
  - seed
```
public long seed
```
    Seed for random numbers (affects sampling) - Note: only reproducible when running single threaded.
  - learningRate
```
public double learningRate
```
    Learning rate (higher => less stable, lower => slower convergence).
  - learningRateAnnealing
```
public double learningRateAnnealing
```
    Learning rate annealing: rate / (1 + rate_annealing * samples).
  - momentumStart
```
public double momentumStart
```
    Initial momentum at the beginning of training (try 0.5).
  - momentumRamp
```
public double momentumRamp
```
    Number of training samples for which momentum increases.
  - momentumStable
```
public double momentumStable
```
    Final momentum after the ramp is over (try 0.99).
  - scoreInterval
```
public double scoreInterval
```
    Shortest time interval (in seconds) between model scoring.
  - scoreTrainingSamples
```
public long scoreTrainingSamples
```
    Number of training set samples for scoring (0 for all).
  - scoreValidationSamples
```
public long scoreValidationSamples
```
    Number of validation set samples for scoring (0 for all).
  - scoreDutyCycle
```
public double scoreDutyCycle
```
    Maximum duty cycle fraction for scoring (lower: more training, higher: more scoring).
  - classificationStop
```
public double classificationStop
```
    Stopping criterion for classification error fraction on training data (-1 to disable).
  - regressionStop
```
public double regressionStop
```
    Stopping criterion for regression error (MSE) on training data (-1 to disable).
  - quietMode
```
public boolean quietMode
```
    Enable quiet mode for less output to standard output.
  - overwriteWithBestModel
```
public boolean overwriteWithBestModel
```
    If enabled, override the final model with the best model found during training.
  - autoencoder
```
public boolean autoencoder
```
    Auto-Encoder.
  - diagnostics
```
public boolean diagnostics
```
    Enable diagnostics for hidden layers.
  - variableImportances
```
public boolean variableImportances
```
    Compute variable importances for input features (Gedeon method) - can be slow for large networks.
  - replicateTrainingData
```
public boolean replicateTrainingData
```
    Replicate the entire training dataset onto every node for faster training on small datasets.
  - singleNodeMode
```
public boolean singleNodeMode
```
    Run on a single node for fine-tuning of model parameters.
  - shuffleTrainingData
```
public boolean shuffleTrainingData
```
    Enable global shuffling of training data.
  - miniBatchSize
```
public int miniBatchSize
```
    Mini-batch size (smaller leads to better fit, larger can speed up and generalize better).
  - clipGradient
```
public double clipGradient
```
    Clip gradients once their absolute value is larger than this value.
  - network
```
public DeepWaterParametersNetwork network
```
    Network architecture.
  - backend
```
public DeepWaterParametersBackend backend
```
    Deep Learning Backend.
  - imageShape
```
public int[] imageShape
```
    Width and height of image.
  - channels
```
public int channels
```
    Number of (color) channels.
  - gpu
```
public boolean gpu
```
    Whether to use a GPU (if available).
  - deviceId
```
public int[] deviceId
```
    Device IDs (which GPUs to use).
  - cacheData
```
public boolean cacheData
```
    Whether to cache the data in memory (automatically disabled if data size is too large).
  - networkDefinitionFile
```
public java.lang.String networkDefinitionFile
```
    Path of file containing network definition (graph, architecture).
  - networkParametersFile
```
public java.lang.String networkParametersFile
```
    Path of file containing network (initial) parameters (weights, biases).
  - meanImageFile
```
public java.lang.String meanImageFile
```
    Path of file containing the mean image data for data normalization.
  - exportNativeParametersPrefix
```
public java.lang.String exportNativeParametersPrefix
```
    Path (prefix) where to export the native model parameters after every iteration.
  - standardize
```
public boolean standardize
```
    If enabled, automatically standardize the data. If disabled, the user must provide properly scaled input data.
  - balanceClasses
```
public boolean balanceClasses
```
    Balance training data class counts via over/under-sampling (for imbalanced data).
  - classSamplingFactors
```
public float[] classSamplingFactors
```
    Desired over/under-sampling ratios per class (in lexicographic order). If not specified, sampling factors will be automatically computed to obtain class balance during training. Requires balance_classes.
  - maxAfterBalanceSize
```
public float maxAfterBalanceSize
```
    Maximum relative size of the training data after balancing class counts (can be less than 1.0). Requires balance_classes.
- Constructor Detail
  - DeepWaterParametersV3
```
public DeepWaterParametersV3()
```
    Public constructor
- Method Detail
  - toString
```
public java.lang.String toString()
```
    Return the contents of this object as a JSON String.
    
    Overrides:
    
    toString in class ModelParametersSchemaV3

Class DeepWaterParametersV3

Field Summary

Fields inherited from class water.bindings.pojos.ModelParametersSchemaV3

Constructor Summary

Method Summary

Methods inherited from class java.lang.Object

Field Detail

problemType

activation

hidden

inputDropoutRatio

hiddenDropoutRatios

maxConfusionMatrixSize

sparse

maxHitRatioK

epochs

trainSamplesPerIteration

targetRatioCommToComp

seed

learningRate

learningRateAnnealing

momentumStart

momentumRamp

momentumStable

scoreInterval

scoreTrainingSamples

scoreValidationSamples

scoreDutyCycle

classificationStop

regressionStop

quietMode

overwriteWithBestModel

autoencoder

diagnostics

variableImportances

replicateTrainingData

singleNodeMode

shuffleTrainingData

miniBatchSize

clipGradient

network

backend

imageShape

channels

gpu

deviceId

cacheData

networkDefinitionFile

networkParametersFile

meanImageFile

exportNativeParametersPrefix

standardize

balanceClasses

classSamplingFactors

maxAfterBalanceSize

Constructor Detail

DeepWaterParametersV3

Method Detail

toString