tf_agents.bandits.multi_objective.multi_objective_scalarizer.Scalarizer
bookmark_border Stay organized with collections Save and categorize content based on your preferences.

Abstract base class for different Scalarizers.

tf_agents.bandits.multi_objective.multi_objective_scalarizer.Scalarizer(
    num_of_objectives: int
)

The Scalarizer class is a callable that transforms multiple objectives into a single scalar reward.

Args
`num_of_objectives`	A non-negative integer indicating the number of objectives to scalarize.

Raises
`ValueError`	if `not isinstance(num_of_objectives, int)`.
`ValueError`	if `num_of_objectives < 2`.

Methods

@abc.abstractmethod
set_parameters(
    **kwargs
)

Setter method for scalarization parameters.

__call__(
    multi_objectives: tf.Tensor
) -> tf.Tensor

Returns a single reward by scalarizing multiple objectives.

Args
`multi_objectives`	A `Tensor` of shape [batch_size, number_of_objectives], where each column represents an objective.

Returns: A Tensor of shape [batch_size] representing scalarized rewards.

Raises
`ValueError`	if `multi_objectives.shape.rank != 2`.
`ValueError`	if `multi_objectives.shape.dims[1] != self._num_of_objectives`.