DeepWaterParameters (h2o-algos version 3.12.0.1 API)

java.lang.Object
- water.Iced<hex.Model.Parameters>
- - hex.Model.Parameters
  - - hex.deepwater.DeepWaterParameters

All Implemented Interfaces:

java.io.Externalizable, java.io.Serializable, java.lang.Cloneable, water.Freezable<hex.Model.Parameters>
```
public class DeepWaterParameters
extends hex.Model.Parameters
```
Parameters for a Deep Water image classification model

See Also:
Serialized Form

Nested Class Summary

Nested Classes
Modifier and Type	Class and Description
`static class`	`DeepWaterParameters.Activation` Activation functions
`static class`	`DeepWaterParameters.Backend`
`static class`	`DeepWaterParameters.MissingValuesHandling`
`static class`	`DeepWaterParameters.Network`
`static class`	`DeepWaterParameters.ProblemType`

Nested classes/interfaces inherited from class hex.Model.Parameters
hex.Model.Parameters.CategoricalEncodingScheme, hex.Model.Parameters.FoldAssignmentScheme

Field Summary

Fields
Modifier and Type	Field and Description
`DeepWaterParameters.Activation`	`_activation` The activation function (non-linearity) to be used the neurons in the hidden layers.
`boolean`	`_autoencoder`
`DeepWaterParameters.Backend`	`_backend`
`boolean`	`_cache_data`
`int`	`_channels`
`double`	`_clip_gradient`
`int[]`	`_device_id`
`double`	`_epochs` The number of passes over the training dataset to be carried out.
`java.lang.String`	`_export_native_parameters_prefix`
`boolean`	`_gpu`
`int[]`	`_hidden` The number and size of each hidden layer in the model.
`double[]`	`_hidden_dropout_ratios` A fraction of the inputs for each hidden layer to be omitted from training in order to improve generalization.
`int[]`	`_image_shape`
`double`	`_input_dropout_ratio` A fraction of the features for each training row to be omitted from training in order to improve generalization (dimension sampling).
`double`	`_learning_rate` When adaptive learning rate is disabled, the magnitude of the weight updates are determined by the user specified learning rate (potentially annealed), and are a function of the difference between the predicted value and the target value.
`double`	`_learning_rate_annealing` Learning rate annealing reduces the learning rate to "freeze" into local minima in the optimization landscape.
`java.lang.String`	`_mean_image_file`
`int`	`_mini_batch_size`
`DeepWaterParameters.MissingValuesHandling`	`_missing_values_handling`
`double`	`_momentum_ramp` The momentum_ramp parameter controls the amount of learning for which momentum increases (assuming momentum_stable is larger than momentum_start).
`double`	`_momentum_stable` The momentum_stable parameter controls the final momentum value reached after momentum_ramp training samples.
`double`	`_momentum_start` The momentum_start parameter controls the amount of momentum at the beginning of training.
`DeepWaterParameters.Network`	`_network`
`java.lang.String`	`_network_definition_file`
`java.lang.String`	`_network_parameters_file`
`boolean`	`_overwrite_with_best_model` If enabled, store the best model under the destination key of this model at the end of training.
`DeepWaterParameters.ProblemType`	`_problem_type`
`boolean`	`_quiet_mode` Enable quiet mode for less output to standard output.
`boolean`	`_replicate_training_data` Replicate the entire training dataset onto every node for faster training on small datasets.
`double`	`_score_duty_cycle` Maximum fraction of wall clock time spent on model scoring on training and validation samples, and on diagnostics such as computation of feature importances (i.e., not on training).
`double`	`_score_interval` The minimum time (in seconds) to elapse between model scoring.
`long`	`_score_training_samples` The number of training dataset points to be used for scoring.
`long`	`_score_validation_samples` The number of validation dataset points to be used for scoring.
`boolean`	`_shuffle_training_data` Enable shuffling of training data (on each node).
`boolean`	`_single_node_mode` Run on a single node for fine-tuning of model parameters.
`boolean`	`_sparse`
`boolean`	`_standardize` If enabled, automatically standardize the data.
`double`	`_target_ratio_comm_to_comp`
`long`	`_train_samples_per_iteration` The number of training data rows to be processed per iteration.
`boolean`	`_use_all_factor_levels`

Fields inherited from class hex.Model.Parameters
_auto_rebalance, _balance_classes, _categorical_encoding, _checkpoint, _class_sampling_factors, _distribution, _fold_assignment, _fold_column, _huber_alpha, _ignore_const_cols, _ignored_columns, _is_cv_model, _keep_cross_validation_fold_assignment, _keep_cross_validation_predictions, _max_after_balance_size, _max_categorical_levels, _max_confusion_matrix_size, _max_runtime_secs, _nfolds, _offset_column, _parallelize_cross_validation, _pretrained_autoencoder, _quantile_alpha, _response_column, _score_each_iteration, _seed, _stopping_metric, _stopping_rounds, _stopping_tolerance, _train, _tweedie_power, _valid, _weights_column, MAX_SUPPORTED_LEVELS

Constructor Summary

Constructors
Constructor and Description

DeepWaterParameters()

Constructors
Constructor and Description
`DeepWaterParameters()`

Method Summary

Methods
Modifier and Type	Method and Description
`java.lang.String`	`algoName()`
`protected double`	`defaultStoppingTolerance()`
`java.lang.String`	`fullName()`
`java.lang.String`	`javaName()`
`float`	`learningRate(double n)`
`float`	`momentum(double n)`
`long`	`progressUnits()`

Methods inherited from class hex.Model.Parameters
checksum_impl, checksum, defaultDropConsCols, getOrMakeRealSeed, hasCheckpoint, missingColumnsType, read_lock_frames, read_unlock_frames, setTrain, train, valid

Methods inherited from class water.Iced
asBytes, clone, copyOver, frozenType, read, readExternal, readJSON, reloadFromBytes, toJsonString, write, writeExternal, writeJSON

Methods inherited from class java.lang.Object
equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

- Field Detail
  - _clip_gradient
```
public double _clip_gradient
```
  - _gpu
```
public boolean _gpu
```
  - _device_id
```
public int[] _device_id
```
  - _network
```
public DeepWaterParameters.Network _network
```
  - _backend
```
public DeepWaterParameters.Backend _backend
```
  - _network_definition_file
```
public java.lang.String _network_definition_file
```
  - _network_parameters_file
```
public java.lang.String _network_parameters_file
```
  - _export_native_parameters_prefix
```
public java.lang.String _export_native_parameters_prefix
```
  - _problem_type
```
public DeepWaterParameters.ProblemType _problem_type
```
  - _image_shape
```
public int[] _image_shape
```
  - _channels
```
public int _channels
```
  - _mean_image_file
```
public java.lang.String _mean_image_file
```
  - _overwrite_with_best_model
```
public boolean _overwrite_with_best_model
```
    If enabled, store the best model under the destination key of this model at the end of training. Only applicable if training is not cancelled.
  - _autoencoder
```
public boolean _autoencoder
```
  - _sparse
```
public boolean _sparse
```
  - _use_all_factor_levels
```
public boolean _use_all_factor_levels
```
  - _missing_values_handling
```
public DeepWaterParameters.MissingValuesHandling _missing_values_handling
```
  - _standardize
```
public boolean _standardize
```
    If enabled, automatically standardize the data. If disabled, the user must provide properly scaled input data.
  - _epochs
```
public double _epochs
```
    The number of passes over the training dataset to be carried out. It is recommended to start with lower values for initial experiments. This value can be modified during checkpoint restarts and allows continuation of selected models.
  - _activation
```
public DeepWaterParameters.Activation _activation
```
    The activation function (non-linearity) to be used the neurons in the hidden layers. Tanh: Hyperbolic tangent function (same as scaled and shifted sigmoid). Rectifier: Chooses the maximum of (0, x) where x is the input value.
  - _hidden
```
public int[] _hidden
```
    The number and size of each hidden layer in the model. For example, if a user specifies "100,200,100" a model with 3 hidden layers will be produced, and the middle hidden layer will have 200 neurons.
  - _input_dropout_ratio
```
public double _input_dropout_ratio
```
    A fraction of the features for each training row to be omitted from training in order to improve generalization (dimension sampling).
  - _hidden_dropout_ratios
```
public double[] _hidden_dropout_ratios
```
    A fraction of the inputs for each hidden layer to be omitted from training in order to improve generalization. Defaults to 0.5 for each hidden layer if omitted.
  - _train_samples_per_iteration
```
public long _train_samples_per_iteration
```
    The number of training data rows to be processed per iteration. Note that independent of this parameter, each row is used immediately to update the model with (online) stochastic gradient descent. This parameter controls the synchronization period between nodes in a distributed environment and the frequency at which scoring and model cancellation can happen. For example, if it is set to 10,000 on H2O running on 4 nodes, then each node will process 2,500 rows per iteration, sampling randomly from their local data. Then, model averaging between the nodes takes place, and scoring can happen (dependent on scoring interval and duty factor). Special values are 0 for one epoch per iteration, -1 for processing the maximum amount of data per iteration (if **replicate training data** is enabled, N epochs will be trained per iteration on N nodes, otherwise one epoch). Special value of -2 turns on automatic mode (auto-tuning).
  - _target_ratio_comm_to_comp
```
public double _target_ratio_comm_to_comp
```
  - _learning_rate
```
public double _learning_rate
```
    When adaptive learning rate is disabled, the magnitude of the weight updates are determined by the user specified learning rate (potentially annealed), and are a function of the difference between the predicted value and the target value. That difference, generally called delta, is only available at the output layer. To correct the output at each hidden layer, back propagation is used. Momentum modifies back propagation by allowing prior iterations to influence the current update. Using the momentum parameter can aid in avoiding local minima and the associated instability. Too much momentum can lead to instabilities, that's why the momentum is best ramped up slowly. This parameter is only active if adaptive learning rate is disabled.
  - _learning_rate_annealing
```
public double _learning_rate_annealing
```
    Learning rate annealing reduces the learning rate to "freeze" into local minima in the optimization landscape. The annealing rate is the inverse of the number of training samples it takes to cut the learning rate in half (e.g., 1e-6 means that it takes 1e6 training samples to halve the learning rate). This parameter is only active if adaptive learning rate is disabled.
  - _momentum_start
```
public double _momentum_start
```
    The momentum_start parameter controls the amount of momentum at the beginning of training. This parameter is only active if adaptive learning rate is disabled.
  - _momentum_ramp
```
public double _momentum_ramp
```
    The momentum_ramp parameter controls the amount of learning for which momentum increases (assuming momentum_stable is larger than momentum_start). The ramp is measured in the number of training samples. This parameter is only active if adaptive learning rate is disabled.
  - _momentum_stable
```
public double _momentum_stable
```
    The momentum_stable parameter controls the final momentum value reached after momentum_ramp training samples. The momentum used for training will remain the same for training beyond reaching that point. This parameter is only active if adaptive learning rate is disabled.
  - _score_interval
```
public double _score_interval
```
    The minimum time (in seconds) to elapse between model scoring. The actual interval is determined by the number of training samples per iteration and the scoring duty cycle.
  - _score_training_samples
```
public long _score_training_samples
```
    The number of training dataset points to be used for scoring. Will be randomly sampled. Use 0 for selecting the entire training dataset.
  - _score_validation_samples
```
public long _score_validation_samples
```
    The number of validation dataset points to be used for scoring. Can be randomly sampled or stratified (if "balance classes" is set and "score validation sampling" is set to stratify). Use 0 for selecting the entire training dataset.
  - _score_duty_cycle
```
public double _score_duty_cycle
```
    Maximum fraction of wall clock time spent on model scoring on training and validation samples, and on diagnostics such as computation of feature importances (i.e., not on training).
  - _quiet_mode
```
public boolean _quiet_mode
```
    Enable quiet mode for less output to standard output.
  - _replicate_training_data
```
public boolean _replicate_training_data
```
    Replicate the entire training dataset onto every node for faster training on small datasets.
  - _single_node_mode
```
public boolean _single_node_mode
```
    Run on a single node for fine-tuning of model parameters. Can be useful for checkpoint resumes after training on multiple nodes for fast initial convergence.
  - _shuffle_training_data
```
public boolean _shuffle_training_data
```
    Enable shuffling of training data (on each node). This option is recommended if training data is replicated on N nodes, and the number of training samples per iteration is close to N times the dataset size, where all nodes train with (almost) all the data. It is automatically enabled if the number of training samples per iteration is set to -1 (or to N times the dataset size or larger).
  - _mini_batch_size
```
public int _mini_batch_size
```
  - _cache_data
```
public boolean _cache_data
```
- Constructor Detail
  - DeepWaterParameters
```
public DeepWaterParameters()
```
- Method Detail
  - algoName
```
public java.lang.String algoName()
```
    Specified by:
    
    algoName in class hex.Model.Parameters
  - fullName
```
public java.lang.String fullName()
```
    Specified by:
    
    fullName in class hex.Model.Parameters
  - javaName
```
public java.lang.String javaName()
```
    Specified by:
    
    javaName in class hex.Model.Parameters
  - defaultStoppingTolerance
```
protected double defaultStoppingTolerance()
```
    Overrides:
    
    defaultStoppingTolerance in class hex.Model.Parameters
  - progressUnits
```
public long progressUnits()
```
    Specified by:
    
    progressUnits in class hex.Model.Parameters
  - learningRate
```
public float learningRate(double n)
```
  - momentum
```
public final float momentum(double n)
```

Class DeepWaterParameters

Nested Class Summary

Nested classes/interfaces inherited from class hex.Model.Parameters

Field Summary

Fields inherited from class hex.Model.Parameters

Constructor Summary

Method Summary

Methods inherited from class hex.Model.Parameters

Methods inherited from class water.Iced

Methods inherited from class java.lang.Object

Field Detail

_clip_gradient

_gpu

_device_id

_network

_backend

_network_definition_file

_network_parameters_file

_export_native_parameters_prefix

_problem_type

_image_shape

_channels

_mean_image_file

_overwrite_with_best_model

_autoencoder

_sparse

_use_all_factor_levels

_missing_values_handling

_standardize

_epochs

_activation

_hidden

_input_dropout_ratio

_hidden_dropout_ratios

_train_samples_per_iteration

_target_ratio_comm_to_comp

_learning_rate

_learning_rate_annealing

_momentum_start

_momentum_ramp

_momentum_stable

_score_interval

_score_training_samples

_score_validation_samples

_score_duty_cycle

_quiet_mode

_replicate_training_data

_single_node_mode

_shuffle_training_data

_mini_batch_size

_cache_data

Constructor Detail

DeepWaterParameters

Method Detail

algoName

fullName

javaName

defaultStoppingTolerance

progressUnits

learningRate

momentum