vggnet_16¶

lucid.models.vggnet_16(num_classes: int = 1000, **kwargs) → VGGNet¶

The vggnet_16 function constructs a VGGNet-16 model, which is a variant of the VGGNet architecture with 16 layers. It builds a VGGNetConfig preset internally and forwards extra keyword arguments to that config.

Total Parameters: 138,357,544

Function Signature¶

@register_model
def vggnet_16(num_classes: int = 1000, **kwargs) -> VGGNet

Parameters¶

num_classes (int, optional): The number of classes for the classification output. Defaults to 1000.
kwargs: Additional keyword arguments passed to VGGNetConfig, such as in_channels, dropout, or classifier_hidden_features.

Returns¶

VGGNet: A VGGNet-16 model initialized with the specified parameters.

Examples¶

from lucid.models import vggnet_16

# Create a VGGNet-16 model
model = vggnet_16(num_classes=100)

print(model)

model = vggnet_16(
    num_classes=10,
    in_channels=1,
    dropout=0.25,
    classifier_hidden_features=(512, 256),
)

print(model.config)