data

LeNetConfig

extendsModelConfig

LeNetConfig(num_classes: int = 10, in_channels: int = 1, activation: str = 'tanh', pooling: str = 'avg')

Configuration for LeNet-5.

activation controls the nonlinearity:

pooling controls the sub-sampling layers:

in_channels defaults to 1 (grayscale). Set to 3 for RGB inputs, though the canonical use-case is MNIST / single-channel images.

Constructors

dunder

→None

__init__(num_classes: int = 10, in_channels: int = 1, activation: str = 'tanh', pooling: str = 'avg')