mediapipe_model_maker.model_util.WarmUp

Applies a warmup schedule on a given learning rate decay schedule.

View aliases

Main aliases

mediapipe_model_maker.face_stylizer.face_stylizer.face_stylizer_options.model_opt.loss_functions.model_util.WarmUp, mediapipe_model_maker.face_stylizer.face_stylizer.loss_functions.model_util.WarmUp, mediapipe_model_maker.face_stylizer.face_stylizer.model_opt.loss_functions.model_util.WarmUp, mediapipe_model_maker.face_stylizer.face_stylizer.model_util.WarmUp, mediapipe_model_maker.face_stylizer.face_stylizer_options.model_opt.loss_functions.model_util.WarmUp, mediapipe_model_maker.face_stylizer.model_options.loss_functions.model_util.WarmUp

mediapipe_model_maker.model_util.WarmUp(
    initial_learning_rate: float,
    decay_schedule_fn: Callable[[Any], Any],
    warmup_steps: int,
    name: Optional[str] = None
)

Args
initial_learning_rate	learning rate after the warmup.
decay_schedule_fn	A function maps step to learning rate. Will be applied for values of step larger than 'warmup_steps'.
warmup_steps	Number of steps to do warmup for.
name	TF namescope under which to perform the learning rate calculation.

Methods

from_config

@classmethod
from_config(
    config
)

Instantiates a LearningRateSchedule from its config.

Args
config	Output of get_config().

Returns
A LearningRateSchedule instance.

get_config

View source

get_config() -> Dict[str, Any]

__call__

View source

__call__(
    step: Union[int, tf.Tensor]
) -> tf.Tensor

Call self as a function.