mediapipe_model_maker.face_stylizer.ModelOptions

Configurable model options for face stylizer models.

mediapipe_model_maker.face_stylizer.ModelOptions(
    swap_layers: Sequence[int] = dataclasses.field(default_factory=lambda : [4, 5, 10, 11]),
    alpha: float = 0.1,
    perception_loss_weight: mediapipe_model_maker.face_stylizer.face_stylizer.loss_functions.PerceptualLossWeight = dataclasses.field(default_factory=_default_perceptual_quality_loss_weight),
    adv_loss_weight: float = 0.2
)

swap_layers The layers of feature to be interpolated between encoding features and StyleGAN input features.
alpha Weighting coefficient of style latent for swapping layer interpolation. Its valid range is [0, 1]. The greater weight means stronger style is applied to the output image. Expect to set it to a small value, i.e. < 0.1.
perception_loss_weight Weighting coefficients of image perception quality loss.
adv_loss_weight Weighting coeffcieint of adversarial loss versus image perceptual quality loss. It expects a small value, i.e. < 0.2.

Methods

__eq__

__eq__(
    other
)

adv_loss_weight 0.2
alpha 0.1