sk_resnet_34¶

lucid.models.sk_resnet_34(num_classes: int = 1000, **kwargs) → SKNet¶

The sk_resnet_34 function constructs a Selective Kernel ResNet-34 model. This preset uses SKNetConfig(block=”basic”, layers=[3, 4, 6, 3]).

Total Parameters: 45,895,512

Function Signature¶

@register_model
def sk_resnet_34(num_classes: int = 1000, **kwargs) -> SKNet

Parameters¶

num_classes (int, optional): Number of output classes for the final fully connected layer. Default is 1000.
kwargs (dict, optional): Additional keyword arguments forwarded to SKNetConfig, excluding the preset block and layers fields.

Returns¶

SKNet: A Selective Kernel ResNet-34 model instance with the preset (block=”basic”, layers=[3, 4, 6, 3]).