mobilenet¶
The mobilenet function creates a MobileNet model instance. It supports customizing the width multiplier and the number of output classes for flexibility across various use cases.
Total Parameters: 4,232,008
Function Signature¶
@register_model
def mobilenet(width_multiplier: float = 1.0, num_classes: int = 1000, **kwargs) -> MobileNet
Parameters¶
width_multiplier (float, optional): Scales the width of the network by adjusting the number of channels in each layer. Default is 1.0, which corresponds to the full model size.
num_classes (int, optional): Specifies the number of output classes for classification. Default is 1000, commonly used for ImageNet.
kwargs (dict, optional): Additional arguments passed to the MobileNet constructor for further customization.
Returns¶
MobileNet: An instance of the MobileNet model configured with the specified parameters.
Example¶
import lucid
from lucid.models import mobilenet
# Create a MobileNet model with default parameters
model = mobilenet(width_multiplier=1.0, num_classes=1000)
# Create a sample input tensor
input_tensor = lucid.random.randn(1, 3, 224, 224)
# Perform a forward pass
output = model(input_tensor)
print(output)