Uses of Interface
org.deeplearning4j.earlystopping.EarlyStoppingModelSaver

Packages that use EarlyStoppingModelSaver

Package	Description
org.deeplearning4j.earlystopping
org.deeplearning4j.earlystopping.saver

Uses of EarlyStoppingModelSaver in org.deeplearning4j.earlystopping

Methods in org.deeplearning4j.earlystopping with parameters of type EarlyStoppingModelSaver

Modifier and Type	Method and Description
EarlyStoppingConfiguration.Builder<T>	EarlyStoppingConfiguration.Builder.modelSaver(EarlyStoppingModelSaver<T> modelSaver) How should models be saved? (Default: in memory)

Uses of EarlyStoppingModelSaver in org.deeplearning4j.earlystopping.saver

Classes in org.deeplearning4j.earlystopping.saver that implement EarlyStoppingModelSaver

Modifier and Type	Class and Description
class	InMemoryModelSaver<T extends Model> Save the best (and latest) models for early stopping training to memory for later retrieval Note: Assumes that network is cloneable via .clone() method
class	LocalFileGraphSaver Save the best (and latest/most recent) ComputationGraphs learned during early stopping training to the local file system. Instances of this class will save 3 files for best (and optionally, latest) models: (a) The network configuration: bestGraphConf.json (b) The network parameters: bestGraphParams.bin (c) The network updater: bestGraphUpdater.bin NOTE: The model updater is an object that contains the internal state for training features such as AdaGrad, Momentum and RMSProp. The updater is not required to use the network at test time; it is saved in case further training is required.
class	LocalFileModelSaver Save the best (and latest/most recent) models learned during early stopping training to the local file system. Instances of this class will save 3 files for best (and optionally, latest) models: (a) The network configuration: bestModelConf.json (b) The network parameters: bestModelParams.bin (c) The network updater: bestModelUpdater.bin NOTE: The model updater is an object that contains the internal state for training features such as AdaGrad, Momentum and RMSProp. The updater is not required to use the network at test time; it is saved in case further training is required.