EfficientNetConfig¶

class lucid.models.EfficientNetConfig(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)¶

EfficientNetConfig stores the compound scaling and classifier settings used by lucid.models.EfficientNet. It defines the EfficientNet-B family width, depth, and resolution scale choices together with classifier regularization.

Class Signature¶

@dataclass
class EfficientNetConfig:
    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

Parameters¶

num_classes (int): Number of output classes.
width_coef (float): Compound width scaling coefficient.
depth_coef (float): Compound depth scaling coefficient.
scale (float): Input resolution scale factor applied by the initial upsampling layer.
dropout (float): Dropout probability applied before the classifier.
se_scale (int): Reduction ratio used by squeeze-and-excitation blocks.
stochastic_depth (bool): Whether stochastic depth should be enabled during training.
p (float): Initial survival probability used when stochastic depth is enabled.

Validation¶

num_classes, width_coef, depth_coef, and scale must be greater than 0.
dropout must be in the range [0, 1).
se_scale must be a positive integer.
stochastic_depth must be a boolean.
p must be in the range [0, 1].

Usage¶

import lucid.models as models

config = models.EfficientNetConfig(
    num_classes=10,
    width_coef=1.0,
    depth_coef=1.0,
    scale=1.0,
    stochastic_depth=True,
    p=0.8,
)
model = models.EfficientNet(config)