DeepLearningModelInfo (h2o-algos 3.46.0 API)

java.lang.Object
- water.Iced<DeepLearningModelInfo>
- - hex.deeplearning.DeepLearningModelInfo

All Implemented Interfaces:

java.io.Externalizable, java.io.Serializable, java.lang.Cloneable, water.Freezable<DeepLearningModelInfo>
```
public final class DeepLearningModelInfo
extends water.Iced<DeepLearningModelInfo>
```
This class contains the state of the Deep Learning model This will be shared: one per node

See Also:

Serialized Form

Nested Class Summary

Nested Classes
Modifier and Type Class and Description

static class DeepLearningModelInfo.GradientCheck

Nested Classes
Modifier and Type	Class and Description
`static class`	`DeepLearningModelInfo.GradientCheck`

Field Summary

Fields
Modifier and Type	Field and Description
`DataInfo`	`data_info`
`static DeepLearningModelInfo.GradientCheck`	`gradientCheck`
`static DeepLearningModelInfo.GradientCheck`	`gradientCheckBias`
`double[]`	`mean_a`
`DeepLearningModel.DeepLearningParameters`	`parameters`
`double[]`	`rms_weight`
`water.util.TwoDimTable`	`summaryTable`

Constructor Summary

Constructors
Constructor and Description
`DeepLearningModelInfo(DeepLearningModel.DeepLearningParameters params, water.Key model_id, DataInfo dinfo, int nClasses, water.fvec.Frame train, water.fvec.Frame valid)` Main constructor

Method Summary

All Methods Static Methods Instance Methods Concrete Methods
Modifier and Type	Method and Description
`void`	`add_processed_global(long p)`
`void`	`add_processed_local(long p)`
`void`	`add(DeepLearningModelInfo other)` Add another model info into this This will add the weights/biases/learning rate helpers, and the number of processed training samples Note: It will NOT add the elastic averaging helpers, which are always kept constant (they already are the result of a reduction)
`protected long`	`checksum_impl()` Unique identifier for this model's state, based on raw numbers
`void`	`computeStats()` Compute statistics about this model on all nodes
`float[]`	`computeVariableImportances()` Compute Variable Importance, based on GEDEON: DATA MINING OF INPUTS: ANALYSING MAGNITUDE AND FUNCTIONAL MEASURES
`DataInfo`	`data_info()`
`protected void`	`div(double N)` Divide all weights/biases by a real-valued number
`water.Key`	`elasticAverageModelInfoKey()`
`Storage.DenseRowMatrix`	`get_ada_dx_g(int i)`
`Storage.DenseVector`	`get_avg_activations(int i)`
`Storage.DenseVector`	`get_biases_ada_dx_g(int i)`
`Storage.DenseVector`	`get_biases_momenta(int i)`
`Storage.DenseVector`	`get_biases(int i)`
`DeepLearningModel.DeepLearningParameters`	`get_params()`
`long`	`get_processed_global()`
`long`	`get_processed_local()`
`long`	`get_processed_total()`
`Storage.DenseRowMatrix`	`get_weights_momenta(int i)`
`Storage.DenseRowMatrix`	`get_weights(int i)`
`boolean`	`isUnstable()`
`water.Key`	`localModelInfoKey(water.H2ONode node)`
`protected void`	`mult(double N)` Multiply all weights/biases by a real-valued number
`void`	`set_params(DeepLearningModel.DeepLearningParameters p, water.Key<hex.Model> model_id)`
`void`	`set_processed_global(long p)`
`void`	`set_processed_local(long p)`
`void`	`setUnstable()`
`long`	`size()`
`static DeepLearningModelInfo`	`timeAverage(DeepLearningModelInfo nodeAverageModel)` TimeAveraging as part of Elastic Averaging Algorithm Cf.
`java.lang.String`	`toString()` Print a summary table
`java.lang.String`	`toStringAll()` Debugging printout

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

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

Field Detail

summaryTable

public water.util.TwoDimTable summaryTable

data_info
```
public DataInfo data_info
```

parameters

public DeepLearningModel.DeepLearningParameters parameters

rms_weight
```
public double[] rms_weight
```

mean_a
```
public double[] mean_a
```

gradientCheck

public static DeepLearningModelInfo.GradientCheck gradientCheck

gradientCheckBias

public static DeepLearningModelInfo.GradientCheck gradientCheckBias

Constructor Detail

DeepLearningModelInfo

public DeepLearningModelInfo(DeepLearningModel.DeepLearningParameters params,
                             water.Key model_id,
                             DataInfo dinfo,
                             int nClasses,
                             water.fvec.Frame train,
                             water.fvec.Frame valid)

Main constructor

Parameters:: params - Model parameters; dinfo - Data Info; nClasses - number of classes (1 for regression, 0 for autoencoder); train - User-given training data frame, prepared by AdaptTestTrain; valid - User-specified validation data frame, prepared by AdaptTestTrain

Method Detail
- data_info
```
public DataInfo data_info()
```
- size
```
public long size()
```
- get_weights
```
public final Storage.DenseRowMatrix get_weights(int i)
```
- get_biases
```
public final Storage.DenseVector get_biases(int i)
```
- get_weights_momenta
```
public final Storage.DenseRowMatrix get_weights_momenta(int i)
```
- get_biases_momenta
```
public final Storage.DenseVector get_biases_momenta(int i)
```
- get_ada_dx_g
```
public final Storage.DenseRowMatrix get_ada_dx_g(int i)
```
- get_biases_ada_dx_g
```
public final Storage.DenseVector get_biases_ada_dx_g(int i)
```
- get_avg_activations
```
public final Storage.DenseVector get_avg_activations(int i)
```
- get_params
```
public final DeepLearningModel.DeepLearningParameters get_params()
```
- set_params
```
public final void set_params(DeepLearningModel.DeepLearningParameters p,
                             water.Key<hex.Model> model_id)
```
- isUnstable
```
public boolean isUnstable()
```
- setUnstable
```
public void setUnstable()
```
- get_processed_global
```
public long get_processed_global()
```
- set_processed_global
```
public void set_processed_global(long p)
```
- add_processed_global
```
public void add_processed_global(long p)
```
- get_processed_local
```
public long get_processed_local()
```
- set_processed_local
```
public void set_processed_local(long p)
```
- add_processed_local
```
public void add_processed_local(long p)
```
- get_processed_total
```
public long get_processed_total()
```
- toString
```
public java.lang.String toString()
```
  Print a summary table
  
  Overrides:
  
  toString in class java.lang.Object
  
  Returns:
  
  String containing ASCII version of summary table
- toStringAll
```
public java.lang.String toStringAll()
```
  Debugging printout
  
  Returns:
  
  String with useful info
- add
```
public void add(DeepLearningModelInfo other)
```
  Add another model info into this This will add the weights/biases/learning rate helpers, and the number of processed training samples Note: It will NOT add the elastic averaging helpers, which are always kept constant (they already are the result of a reduction)
  
  Parameters:
  
  other - Other DeepLearningModelInfo to add into this one
- mult
```
protected void mult(double N)
```
  Multiply all weights/biases by a real-valued number
  
  Parameters:
  
  N - multiplication factor
- div
```
protected void div(double N)
```
  Divide all weights/biases by a real-valued number
  
  Parameters:
  
  N - divisor
- computeVariableImportances
```
public float[] computeVariableImportances()
```
  Compute Variable Importance, based on GEDEON: DATA MINING OF INPUTS: ANALYSING MAGNITUDE AND FUNCTIONAL MEASURES
  
  Returns:
  
  variable importances for input features
- computeStats
```
public void computeStats()
```
  Compute statistics about this model on all nodes
- checksum_impl
```
protected long checksum_impl()
```
  Unique identifier for this model's state, based on raw numbers
- timeAverage
```
public static DeepLearningModelInfo timeAverage(DeepLearningModelInfo nodeAverageModel)
```
  TimeAveraging as part of Elastic Averaging Algorithm Cf. equation 6 of arXiv:1412.6651v5
  
  Parameters:
  
  nodeAverageModel - current average of per-node models
  
  Returns:
  
  Time-average of node-averages (consensus model, "the" model)
- localModelInfoKey
```
public water.Key localModelInfoKey(water.H2ONode node)
```
- elasticAverageModelInfoKey
```
public water.Key elasticAverageModelInfoKey()
```

Class DeepLearningModelInfo

Nested Class Summary

Field Summary

Constructor Summary

Method Summary

Methods inherited from class water.Iced

Methods inherited from class java.lang.Object

Field Detail

summaryTable

data_info

parameters

rms_weight

mean_a

gradientCheck

gradientCheckBias

Constructor Detail

DeepLearningModelInfo

Method Detail

data_info

size

get_weights

get_biases

get_weights_momenta

get_biases_momenta

get_ada_dx_g

get_biases_ada_dx_g

get_avg_activations

get_params

set_params

isUnstable

setUnstable

get_processed_global

set_processed_global

add_processed_global

get_processed_local

set_processed_local

add_processed_local

get_processed_total

toString

toStringAll

add

mult

div

computeVariableImportances

computeStats

checksum_impl

timeAverage

localModelInfoKey

elasticAverageModelInfoKey