zennit.types

Type definitions for convenience.

Classes

`Activation`	Abstract base class that describes activation modules.
`AvgPool`	Abstract base class that describes sum-pooling modules.
`BatchNorm`	Abstract base class that describes batch normalization modules.
`Convolution`	Abstract base class that describes all convolutional modules.
`ConvolutionStandard`	Abstract base class that describes standard (forward) convolutional modules.
`ConvolutionTranspose`	Abstract base class that describes transposed convolutional modules.
`Linear`	Abstract base class that describes linear modules.
`MaxPool`	Abstract base class that describes max-pooling modules.
`SubclassMeta`	Meta class to bundle multiple subclasses.

class zennit.types.Activation[source]

Bases: object

Abstract base class that describes activation modules.

class zennit.types.AvgPool[source]

Bases: object

Abstract base class that describes sum-pooling modules.

class zennit.types.BatchNorm[source]

Bases: object

Abstract base class that describes batch normalization modules.

class zennit.types.Convolution[source]

Bases: object

Abstract base class that describes all convolutional modules.

class zennit.types.ConvolutionStandard[source]

Bases: object

Abstract base class that describes standard (forward) convolutional modules.

class zennit.types.ConvolutionTranspose[source]

Bases: object

Abstract base class that describes transposed convolutional modules.

class zennit.types.Linear[source]

Bases: object

Abstract base class that describes linear modules.

class zennit.types.MaxPool[source]

Bases: object

Abstract base class that describes max-pooling modules.

class zennit.types.SubclassMeta[source]

Bases: type

Meta class to bundle multiple subclasses.