Module: tf_agents.trajectories.policy_step

Policy Step.

Classes

class CommonFields: Strings which can be used for querying returned PolicyStep.info field.

class PolicyInfo: PolicyInfo(log_probability,)

class PolicyStep: Returned with every call to policy.action() and policy.distribution().

Functions

get_log_probability(...): Gets the CommonFields.LOG_PROBABILITY from info depending on type.

set_log_probability(...): Sets the CommonFields.LOG_PROBABILITY on info to be log_probability.

Type Aliases

ActionType

absolute_import Instance of __future__._Feature
division Instance of __future__._Feature
print_function Instance of __future__._Feature