EfficientNet¶

ConvNet Image Classification

class lucid.models.EfficientNet(num_classes: int = 1000, width_coef: float = 1.0, depth_coef: float = 1.0, scale: float = 1.0, dropout: float = 0.2, se_scale: int = 4, stochastic_depth: bool = False, p: float = 0.5)¶

The EfficientNet class implements a scalable and efficient convolutional neural network architecture that can be configured to encompass all EfficientNet-B0 to B7 variants.

Class Signature¶

class EfficientNet(nn.Module):
    def __init__(
        self,
        num_classes: int = 1000,
        width_coef: float = 1.0,
        depth_coef: float = 1.0,
        scale: float = 1.0,
        dropout: float = 0.2,
        se_scale: int = 4,
        stochastic_depth: bool = False,
        p: float = 0.5,
    ) -> None

Parameters¶

num_classes (int, optional): Number of output classes for the final classification layer. Defaults to 1000 (e.g., for ImageNet).
width_coef (float, optional): Coefficient to scale the width (number of channels) of the network. Defaults to 1.0.
depth_coef (float, optional): Coefficient to scale the depth (number of layers) of the network. Defaults to 1.0.
scale (float, optional): Global scaling factor applied to the input resolution. Defaults to 1.0.
dropout (float, optional): Dropout rate applied to the final fully connected layer. Defaults to 0.2.
se_scale (int, optional): Reduction ratio for the squeeze-and-excitation (SE) block. Defaults to 4.
stochastic_depth (bool, optional): Whether to use stochastic depth regularization. Defaults to False.
p (float, optional): Probability for stochastic depth when enabled. Defaults to 0.5.

Configurations¶

The following table summarizes the configurations for EfficientNet variants B0 to B7:

EfficientNet Configurations¶
Variant	Width Coefficient	Depth Coefficient	Input Resolution	Dropout Rate
B0	1.0	1.0	224x224	0.2
B1	1.0	1.1	240x240	0.2
B2	1.1	1.2	260x260	0.3
B3	1.2	1.4	300x300	0.3
B4	1.4	1.8	380x380	0.4
B5	1.6	2.2	456x456	0.4
B6	1.8	2.6	528x528	0.5
B7	2.0	3.1	600x600	0.5

Examples¶

from lucid.models import EfficientNet

# Instantiate EfficientNet-B0
model_b0 = EfficientNet(num_classes=1000, width_coef=1.0, depth_coef=1.0, scale=1.0)

# Forward pass with a random input
input_tensor = lucid.random.randn(1, 3, 224, 224)  # Batch size of 1, ImageNet resolution
output = model_b0(input_tensor)
print(output.shape)  # Output shape: (1, 1000)

# Instantiate EfficientNet-B7
model_b7 = EfficientNet(num_classes=1000, width_coef=2.0, depth_coef=3.1, scale=2.0)

# Forward pass
input_tensor = lucid.random.randn(1, 3, 600, 600)  # Larger resolution for B7
output = model_b7(input_tensor)
print(output.shape)  # Output shape: (1, 1000)