org.deeplearning4j.nn.modelimport.keras

Class KerasModelImport

java.lang.Object

org.deeplearning4j.nn.modelimport.keras.KerasModelImport

public class KerasModelImport
extends java.lang.Object

Reads stored Keras configurations and weights from one of two archives: either (1) a single HDF5 file storing model and training JSON configurations and weights or (2) separate text file storing model JSON configuration and HDF5 file storing weights.

Constructor Summary

Constructors

Constructor and Description
KerasModelImport()

Method Summary

Modifier and Type	Method and Description
static ComputationGraph	importKerasModelAndWeights(java.io.InputStream modelHdf5Stream) Load Keras (Functional API) Model saved using model.save_model(...).
static ComputationGraph	importKerasModelAndWeights(java.io.InputStream modelHdf5Stream, boolean enforceTrainingConfig) Load Keras (Functional API) Model saved using model.save_model(...).
static ComputationGraph	importKerasModelAndWeights(java.lang.String modelHdf5Filename) Load Keras (Functional API) Model saved using model.save_model(...).
static ComputationGraph	importKerasModelAndWeights(java.lang.String modelHdf5Filename, boolean enforceTrainingConfig) Load Keras (Functional API) Model saved using model.save_model(...).
static ComputationGraph	importKerasModelAndWeights(java.lang.String modelJsonFilename, java.lang.String weightsHdf5Filename) Load Keras (Functional API) Model for which the configuration and weights were saved separately using calls to model.to_json() and model.save_weights(...).
static ComputationGraph	importKerasModelAndWeights(java.lang.String modelJsonFilename, java.lang.String weightsHdf5Filename, boolean enforceTrainingConfig) Load Keras (Functional API) Model for which the configuration and weights were saved separately using calls to model.to_json() and model.save_weights(...).
static ComputationGraphConfiguration	importKerasModelConfiguration(java.lang.String modelJsonFilename) Load Keras (Functional API) Model for which the configuration was saved separately using calls to model.to_json() and model.save_weights(...).
static ComputationGraphConfiguration	importKerasModelConfiguration(java.lang.String modelJsonFilename, boolean enforceTrainingConfig) Load Keras (Functional API) Model for which the configuration was saved separately using calls to model.to_json() and model.save_weights(...).
static MultiLayerConfiguration	importKerasSequentialConfiguration(java.lang.String modelJsonFilename) Load Keras Sequential model for which the configuration was saved separately using calls to model.to_json() and model.save_weights(...).
static MultiLayerConfiguration	importKerasSequentialConfiguration(java.lang.String modelJsonFilename, boolean enforceTrainingConfig) Load Keras Sequential model for which the configuration was saved separately using calls to model.to_json() and model.save_weights(...).
static MultiLayerNetwork	importKerasSequentialModelAndWeights(java.io.InputStream modelHdf5Stream) Load Keras Sequential model saved using model.save_model(...).
static MultiLayerNetwork	importKerasSequentialModelAndWeights(java.io.InputStream modelHdf5Stream, boolean enforceTrainingConfig) Load Keras Sequential model saved using model.save_model(...).
static MultiLayerNetwork	importKerasSequentialModelAndWeights(java.lang.String modelHdf5Filename) Load Keras Sequential model saved using model.save_model(...).
static MultiLayerNetwork	importKerasSequentialModelAndWeights(java.lang.String modelHdf5Filename, boolean enforceTrainingConfig) Load Keras Sequential model saved using model.save_model(...).
static MultiLayerNetwork	importKerasSequentialModelAndWeights(java.lang.String modelJsonFilename, java.lang.String weightsHdf5Filename) Load Keras Sequential model for which the configuration and weights were saved separately using calls to model.to_json() and model.save_weights(...).
static MultiLayerNetwork	importKerasSequentialModelAndWeights(java.lang.String modelJsonFilename, java.lang.String weightsHdf5Filename, boolean enforceTrainingConfig) Load Keras Sequential model for which the configuration and weights were saved separately using calls to model.to_json() and model.save_weights(...).

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

KerasModelImport

public KerasModelImport()

Method Detail

importKerasModelAndWeights

public static ComputationGraph importKerasModelAndWeights(java.io.InputStream modelHdf5Stream,
                                                          boolean enforceTrainingConfig)

Load Keras (Functional API) Model saved using model.save_model(...).

Parameters:

modelHdf5Stream - InputStream containing HDF5 archive storing Keras Model

enforceTrainingConfig - whether to enforce training configuration options

Returns:

ComputationGraph

See Also:

ComputationGraph

importKerasModelAndWeights

public static ComputationGraph importKerasModelAndWeights(java.io.InputStream modelHdf5Stream)

Load Keras (Functional API) Model saved using model.save_model(...).

Parameters:

modelHdf5Stream - InputStream containing HDF5 archive storing Keras Model

Returns:

ComputationGraph

See Also:

ComputationGraph

importKerasSequentialModelAndWeights

public static MultiLayerNetwork importKerasSequentialModelAndWeights(java.io.InputStream modelHdf5Stream,
                                                                     boolean enforceTrainingConfig)
                                                              throws java.io.IOException,
                                                                     InvalidKerasConfigurationException,
                                                                     UnsupportedKerasConfigurationException

Load Keras Sequential model saved using model.save_model(...).

Parameters:

modelHdf5Stream - InputStream containing HDF5 archive storing Keras Sequential model

enforceTrainingConfig - whether to enforce training configuration options

Returns:

ComputationGraph

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

ComputationGraph

importKerasSequentialModelAndWeights

public static MultiLayerNetwork importKerasSequentialModelAndWeights(java.io.InputStream modelHdf5Stream)
                                                              throws java.io.IOException,
                                                                     InvalidKerasConfigurationException,
                                                                     UnsupportedKerasConfigurationException

Load Keras Sequential model saved using model.save_model(...).

Parameters:

modelHdf5Stream - InputStream containing HDF5 archive storing Keras Sequential model

Returns:

ComputationGraph

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

ComputationGraph

importKerasModelAndWeights

public static ComputationGraph importKerasModelAndWeights(java.lang.String modelHdf5Filename,
                                                          boolean enforceTrainingConfig)
                                                   throws java.io.IOException,
                                                          UnsupportedKerasConfigurationException,
                                                          InvalidKerasConfigurationException

Load Keras (Functional API) Model saved using model.save_model(...).

Parameters:

modelHdf5Filename - path to HDF5 archive storing Keras Model

enforceTrainingConfig - whether to enforce training configuration options

Returns:

ComputationGraph

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

ComputationGraph

importKerasModelAndWeights

public static ComputationGraph importKerasModelAndWeights(java.lang.String modelHdf5Filename)
                                                   throws java.io.IOException,
                                                          UnsupportedKerasConfigurationException,
                                                          InvalidKerasConfigurationException

Load Keras (Functional API) Model saved using model.save_model(...).

Parameters:

modelHdf5Filename - path to HDF5 archive storing Keras Model

Returns:

ComputationGraph

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

ComputationGraph

importKerasSequentialModelAndWeights

public static MultiLayerNetwork importKerasSequentialModelAndWeights(java.lang.String modelHdf5Filename,
                                                                     boolean enforceTrainingConfig)
                                                              throws java.io.IOException,
                                                                     InvalidKerasConfigurationException,
                                                                     UnsupportedKerasConfigurationException

Load Keras Sequential model saved using model.save_model(...).

Parameters:

modelHdf5Filename - path to HDF5 archive storing Keras Sequential model

enforceTrainingConfig - whether to enforce training configuration options

Returns:

MultiLayerNetwork

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

MultiLayerNetwork

importKerasSequentialModelAndWeights

public static MultiLayerNetwork importKerasSequentialModelAndWeights(java.lang.String modelHdf5Filename)
                                                              throws java.io.IOException,
                                                                     InvalidKerasConfigurationException,
                                                                     UnsupportedKerasConfigurationException

Load Keras Sequential model saved using model.save_model(...).

Parameters:

modelHdf5Filename - path to HDF5 archive storing Keras Sequential model

Returns:

MultiLayerNetwork

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

MultiLayerNetwork

importKerasModelAndWeights

public static ComputationGraph importKerasModelAndWeights(java.lang.String modelJsonFilename,
                                                          java.lang.String weightsHdf5Filename,
                                                          boolean enforceTrainingConfig)
                                                   throws java.io.IOException,
                                                          InvalidKerasConfigurationException,
                                                          UnsupportedKerasConfigurationException

Load Keras (Functional API) Model for which the configuration and weights were saved separately using calls to model.to_json() and model.save_weights(...).

Parameters:

modelJsonFilename - path to JSON file storing Keras Model configuration

weightsHdf5Filename - path to HDF5 archive storing Keras model weights

enforceTrainingConfig - whether to enforce training configuration options

Returns:

ComputationGraph

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

ComputationGraph

importKerasModelAndWeights

public static ComputationGraph importKerasModelAndWeights(java.lang.String modelJsonFilename,
                                                          java.lang.String weightsHdf5Filename)
                                                   throws java.io.IOException,
                                                          InvalidKerasConfigurationException,
                                                          UnsupportedKerasConfigurationException

Load Keras (Functional API) Model for which the configuration and weights were saved separately using calls to model.to_json() and model.save_weights(...).

Parameters:

modelJsonFilename - path to JSON file storing Keras Model configuration

weightsHdf5Filename - path to HDF5 archive storing Keras model weights

Returns:

ComputationGraph

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

ComputationGraph

importKerasSequentialModelAndWeights

public static MultiLayerNetwork importKerasSequentialModelAndWeights(java.lang.String modelJsonFilename,
                                                                     java.lang.String weightsHdf5Filename,
                                                                     boolean enforceTrainingConfig)
                                                              throws java.io.IOException,
                                                                     InvalidKerasConfigurationException,
                                                                     UnsupportedKerasConfigurationException

Load Keras Sequential model for which the configuration and weights were saved separately using calls to model.to_json() and model.save_weights(...).

Parameters:

modelJsonFilename - path to JSON file storing Keras Sequential model configuration

weightsHdf5Filename - path to HDF5 archive storing Keras model weights

enforceTrainingConfig - whether to enforce training configuration options

Returns:

MultiLayerNetwork

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

MultiLayerNetwork

importKerasSequentialModelAndWeights

public static MultiLayerNetwork importKerasSequentialModelAndWeights(java.lang.String modelJsonFilename,
                                                                     java.lang.String weightsHdf5Filename)
                                                              throws java.io.IOException,
                                                                     InvalidKerasConfigurationException,
                                                                     UnsupportedKerasConfigurationException

Load Keras Sequential model for which the configuration and weights were saved separately using calls to model.to_json() and model.save_weights(...).

Parameters:

modelJsonFilename - path to JSON file storing Keras Sequential model configuration

weightsHdf5Filename - path to HDF5 archive storing Keras model weights

Returns:

MultiLayerNetwork

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

MultiLayerNetwork

importKerasModelConfiguration

public static ComputationGraphConfiguration importKerasModelConfiguration(java.lang.String modelJsonFilename,
                                                                          boolean enforceTrainingConfig)
                                                                   throws java.io.IOException,
                                                                          InvalidKerasConfigurationException,
                                                                          UnsupportedKerasConfigurationException

Load Keras (Functional API) Model for which the configuration was saved separately using calls to model.to_json() and model.save_weights(...).

Parameters:

modelJsonFilename - path to JSON file storing Keras Model configuration

enforceTrainingConfig - whether to enforce training configuration options

Returns:

ComputationGraph

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

ComputationGraph

importKerasModelConfiguration

public static ComputationGraphConfiguration importKerasModelConfiguration(java.lang.String modelJsonFilename)
                                                                   throws java.io.IOException,
                                                                          InvalidKerasConfigurationException,
                                                                          UnsupportedKerasConfigurationException

Load Keras (Functional API) Model for which the configuration was saved separately using calls to model.to_json() and model.save_weights(...).

Parameters:

modelJsonFilename - path to JSON file storing Keras Model configuration

Returns:

ComputationGraph

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

ComputationGraph

importKerasSequentialConfiguration

public static MultiLayerConfiguration importKerasSequentialConfiguration(java.lang.String modelJsonFilename,
                                                                         boolean enforceTrainingConfig)
                                                                  throws java.io.IOException,
                                                                         InvalidKerasConfigurationException,
                                                                         UnsupportedKerasConfigurationException

Load Keras Sequential model for which the configuration was saved separately using calls to model.to_json() and model.save_weights(...).

Parameters:

modelJsonFilename - path to JSON file storing Keras Sequential model configuration

enforceTrainingConfig - whether to enforce training configuration options

Returns:

MultiLayerNetwork

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

MultiLayerNetwork

importKerasSequentialConfiguration

public static MultiLayerConfiguration importKerasSequentialConfiguration(java.lang.String modelJsonFilename)
                                                                  throws java.io.IOException,
                                                                         InvalidKerasConfigurationException,
                                                                         UnsupportedKerasConfigurationException

Load Keras Sequential model for which the configuration was saved separately using calls to model.to_json() and model.save_weights(...).

Parameters:

modelJsonFilename - path to JSON file storing Keras Sequential model configuration

Returns:

MultiLayerNetwork

Throws:

java.io.IOException

InvalidKerasConfigurationException

UnsupportedKerasConfigurationException

See Also:

MultiLayerNetwork