| Package | Description |
|---|---|
| org.deeplearning4j.arbiter |
| Class and Description |
|---|
| BaseNetworkSpace
This is an abstract ParameterSpace for both MultiLayerNetworks (MultiLayerSpace) and ComputationGraph (ComputationGraphSpace)
Functionality here should match
NeuralNetConfiguration.Builder |
| BaseNetworkSpace.Builder |
| BaseNetworkSpace.LayerConf |
| ComputationGraphSpace
ComputationGraphSpace: Defines the space of valid hyperparameters for a ComputationGraph.
|
| ComputationGraphSpace.Builder |
| ComputationGraphSpace.VertexConf |
| DL4JConfiguration
DL4JConfiguration: simple configuration method that contains the following:
- MultiLayerConfiguration - Early stopping settings, OR number of epochs Note: if early stopping configuration is absent, a fixed number of epochs (default: 1) will be used. |
| GraphConfiguration
Analogous to
DL4JConfiguration, GraphConfiguration includes a configuration for ComputationGraphs, as well
as early stopping (or, optionally numEpochs) fields. |
| MultiLayerSpace |
| MultiLayerSpace.Builder |
Copyright © 2020. All rights reserved.