sk_resnext_50_32x4d

lucid.models.sk_resnext_50_32x4d(num_classes: int = 1000, **kwargs) SKNet

The sk_resnext_50_32x4d function is a model constructor which creates an instance of the SK-ResNeXt-50 (32x4d) architecture, which is a variant of ResNeXt enhanced with Selective Kernel (SK) mechanisms for dynamic kernel selection.

Total Parameters: 29,274,760

Function Signature

def sk_resnext_50_32x4d(num_classes: int = 1000, **kwargs) -> SKNet

Parameters

  • num_classes (int, optional): The number of output classes for the model. Default is 1000 (for ImageNet classification).

  • kwargs (dict, optional): Additional keyword arguments to configure the model. This allows flexibility in customizing the model architecture or settings.

Returns

  • SKNet: An instance of the SK-ResNeXt-50 (32x4d) model, ready for training or inference.

Key Features

  • ResNeXt Architecture: Incorporates grouped convolutions for efficient computation while maintaining high model capacity.

  • Selective Kernel (SK) Blocks: Dynamically adjusts receptive fields by adaptively combining multiple kernels, enhancing feature representation.

  • Flexible Output Classes: Configurable for different classification tasks by adjusting the num_classes parameter.

Example Usage

The following example demonstrates how to instantiate the sk_resnext_50_32x4d model for a custom task:

from lucid.nn.models import sk_resnext_50_32x4d

# Create the model for a custom task with 1000 classes
model = sk_resnext_50_32x4d(num_classes=1000)

# Print the model summary
print(model)

# Forward pass with a sample input
import lucid
input_tensor = lucid.random.randn(1, 3, 224, 224)  # Batch of one image, 3 channels, 224x224 resolution
output = model(input_tensor)

print("Output shape:", output.shape)