public class DeepLearningParameters
extends hex.Model.Parameters
Modifier and Type | Class and Description |
---|---|
static class |
DeepLearningParameters.Activation
Activation functions
|
static class |
DeepLearningParameters.ClassSamplingMethod |
static class |
DeepLearningParameters.InitialWeightDistribution |
static class |
DeepLearningParameters.Loss
Loss functions
Absolute, MeanSquare, Huber for regression
Absolute, MeanSquare, Huber or CrossEntropy for classification
|
static class |
DeepLearningParameters.MissingValuesHandling |
Modifier and Type | Field and Description |
---|---|
DeepLearningParameters.Activation |
_activation
The activation function (non-linearity) to be used the neurons in the hidden layers.
|
boolean |
_adaptive_rate
The implemented adaptive learning rate algorithm (ADADELTA) automatically
combines the benefits of learning rate annealing and momentum
training to avoid slow convergence.
|
boolean |
_autoencoder |
double |
_average_activation |
water.Key |
_checkpoint
A model key associated with a previously trained Deep Learning
model.
|
double |
_classification_stop
The stopping criteria in terms of classification error (1-accuracy) on the
training data scoring dataset.
|
boolean |
_col_major |
boolean |
_diagnostics
Gather diagnostics for hidden layers, such as mean and RMS values of learning
rate, momentum, weights and biases.
|
boolean |
_elastic_averaging |
double |
_elastic_averaging_moving_rate |
double |
_elastic_averaging_regularization |
double |
_epochs
The number of passes over the training dataset to be carried out.
|
double |
_epsilon
The second of two hyper parameters for adaptive learning rate (ADADELTA).
|
boolean |
_export_weights_and_biases |
boolean |
_fast_mode
Enable fast mode (minor approximation in back-propagation), should not affect results significantly.
|
boolean |
_force_load_balance
Increase training speed on small datasets by splitting it into many chunks
to allow utilization of all cores.
|
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.
|
DeepLearningParameters.InitialWeightDistribution |
_initial_weight_distribution
The distribution from which initial weights are to be drawn.
|
double |
_initial_weight_scale
The scale of the distribution function for Uniform or Normal distributions.
|
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).
|
boolean |
_keep_cross_validation_splits |
double |
_l1
A regularization method that constrains the absolute value of the weights and
has the net effect of dropping some weights (setting them to zero) from a model
to reduce complexity and avoid overfitting.
|
double |
_l2
A regularization method that constrdains the sum of the squared
weights.
|
DeepLearningParameters.Loss |
_loss
The loss (error) function to be minimized by the model.
|
int |
_max_categorical_features
Max.
|
float |
_max_w2
A maximum on the sum of the squared incoming weights into
any one neuron.
|
DeepLearningParameters.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.
|
boolean |
_nesterov_accelerated_gradient
The Nesterov accelerated gradient descent method is a modification to
traditional gradient descent for convex functions.
|
boolean |
_overwrite_with_best_model
If enabled, store the best model under the destination key of this model at the end of training.
|
boolean |
_quiet_mode
Enable quiet mode for less output to standard output.
|
double |
_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 |
_rate_annealing
Learning rate annealing reduces the learning rate to "freeze" into
local minima in the optimization landscape.
|
double |
_rate_decay
The learning rate decay parameter controls the change of learning rate across layers.
|
double |
_regression_stop
The stopping criteria in terms of regression error (MSE) on the training
data scoring dataset.
|
boolean |
_replicate_training_data
Replicate the entire training dataset onto every node for faster training on small datasets.
|
boolean |
_reproducible
Force reproducibility on small data (will be slow - only uses 1 thread)
|
double |
_rho
The first of two hyper parameters for adaptive learning rate (ADADELTA).
|
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.
|
DeepLearningParameters.ClassSamplingMethod |
_score_validation_sampling
Method used to sample the validation dataset for scoring, see Score Validation Samples above.
|
long |
_seed
The random seed controls sampling and initialization.
|
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 |
double |
_sparsity_beta |
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 |
boolean |
_variable_importances
Whether to compute variable importances for input features.
|
_balance_classes, _class_sampling_factors, _ignore_const_cols, _ignored_columns, _max_after_balance_size, _max_confusion_matrix_size, _max_hit_ratio_k, _model_id, _offset_column, _response_column, _score_each_iteration, _train, _valid, _weights_column
Constructor and Description |
---|
DeepLearningParameters() |
Modifier and Type | Method and Description |
---|---|
int |
getNumFolds() |
double |
missingColumnsType() |
checksum_impl, defaultDropConsCols, defaultDropNA20Cols, read_lock_frames, read_unlock_frames, train, valid
public boolean _keep_cross_validation_splits
public water.Key _checkpoint
public boolean _overwrite_with_best_model
public boolean _autoencoder
public boolean _use_all_factor_levels
public DeepLearningParameters.Activation _activation
public int[] _hidden
public double _epochs
public long _train_samples_per_iteration
public double _target_ratio_comm_to_comp
public long _seed
public boolean _adaptive_rate
public double _rho
public double _epsilon
public double _rate
public double _rate_annealing
public double _rate_decay
public double _momentum_start
public double _momentum_ramp
public double _momentum_stable
public boolean _nesterov_accelerated_gradient
public double _input_dropout_ratio
public double[] _hidden_dropout_ratios
public double _l1
public double _l2
public float _max_w2
public DeepLearningParameters.InitialWeightDistribution _initial_weight_distribution
public double _initial_weight_scale
public DeepLearningParameters.Loss _loss
public double _score_interval
public long _score_training_samples
public long _score_validation_samples
public double _score_duty_cycle
public double _classification_stop
public double _regression_stop
public boolean _quiet_mode
public DeepLearningParameters.ClassSamplingMethod _score_validation_sampling
public boolean _diagnostics
public boolean _variable_importances
public boolean _fast_mode
public boolean _force_load_balance
public boolean _replicate_training_data
public boolean _single_node_mode
public boolean _shuffle_training_data
public DeepLearningParameters.MissingValuesHandling _missing_values_handling
public boolean _sparse
public boolean _col_major
public double _average_activation
public double _sparsity_beta
public int _max_categorical_features
public boolean _reproducible
public boolean _export_weights_and_biases
public boolean _elastic_averaging
public double _elastic_averaging_moving_rate
public double _elastic_averaging_regularization