View source on GitHub |
Helper function to compute the action feasibility probability.
tf_agents.bandits.policies.constraints.compute_feasibility_probability(
observation: tf_agents.typing.types.NestedTensor
,
constraints: Iterable[tf_agents.bandits.policies.constraints.BaseConstraint
],
batch_size: tf_agents.typing.types.Int
,
num_actions: int,
action_mask: Optional[types.Tensor] = None
) -> tf_agents.typing.types.Float