Package ai.djl.nn
Contains classes to construct neural networks.
The primary construct used to build up the networks is the Block (see for
details). This package contains a number of implementations of blocks as well as helpers for
blocks.
The following subpackages also contain a number of standard neural network operations to use with blocks:
-
Interface Summary Interface Description Block ABlockis a composable function that forms a neural network.BlockFactory Block factory is a component to make standard for block creating and saving procedure.SymbolBlock SymbolBlockis aBlockis used to load models that were exported directly from the engine in its native format. -
Class Summary Class Description AbstractBaseBlock This provides shared functionality for both the DJL-basedAbstractBlocks and the importedAbstractSymbolBlocks.AbstractBlock AbstractBlockis an abstract implementation ofBlock.AbstractSymbolBlock AbstractSymbolBlockis an abstract implementation ofSymbolBlock.Activation Utility class that provides activation functions and blocks.BlockList Represents a set of names and Blocks.Blocks Utility class that provides some useful blocks.IdentityBlockFactory ABlockFactoryclass that creates IdentityBlock.LambdaBlock LambdaBlockis aBlockwith no parameters or children.ParallelBlock ParallelBlockis aBlockwhose children form a parallel branch in the network and are combined to produce a single output.Parameter Parameteris a container class that holds a learnable parameter of a model.Parameter.Builder A Builder to construct aParameter.ParameterList Represents a set of names and Parameters.SequentialBlock SequentialBlockis aBlockwhose children form a chain of blocks with each child block feeding its output to the next. -
Enum Summary Enum Description Parameter.Type Enumerates the types ofParameter. -
Exception Summary Exception Description UninitializedParameterException Thrown to indicate that aParameterwas not initialized.