se_resnet_152

lucid.models.se_resnet_152(num_classes: int = 1000, **kwargs) → SENet

Overview

The se_resnet_152 function constructs an SE-ResNet-152 model, a deep and accurate residual network with SE blocks, suitable for large-scale and complex image classification tasks.

Total Parameters: 66,821,848

Function Signature

@register_model
def se_resnet_152(num_classes: int = 1000, **kwargs) -> SENet:

Parameters

  • num_classes (int, optional): Number of output classes for the classification task. Default is 1000.

  • kwargs: Additional keyword arguments to customize the model.

Returns

  • SENet: An instance of the SE-ResNet-152 model.

Examples

Creating an SE-ResNet-152 model for 1000 classes:

model = se_resnet_152(num_classes=1000)
print(model)

Note

  • SE-ResNet-152 uses a configuration of [3, 8, 36, 3] for its layers.

  • Incorporates _SEResNetBottleneck for SE operations.

  • Initializes weights internally unless specified otherwise through kwargs.