tf_agents.policies.PyTFEagerPolicy

Exposes a numpy API for TF policies in Eager mode.

Inherits From: PyTFEagerPolicyBase, PyPolicy

Used in the notebooks

Used in the tutorials

policy tf_policy.TFPolicy instance to wrap and expose as a py_policy.
time_step_spec A TimeStep ArraySpec of the expected time_steps. Usually provided by the user to the subclass.
action_spec A nest of BoundedArraySpec representing the actions. Usually provided by the user to the subclass.
policy_state_spec A nest of ArraySpec representing the policy state. Provided by the subclass, not directly by the user.
info_spec A nest of ArraySpec representing the policy info. Provided by the subclass, not directly by the user.
use_tf_function Wraps the use of policy.action in a tf.function call which can help speed up execution.
batch_time_steps Wether time_steps should be batched before being passed to the wrapped policy. Leave as True unless you are dealing with a batched environment, in which case you want to skip the batching as that dim will already be present.

action_spec Describes the ArraySpecs of the np.Array returned by action().

action can be a single np.Array, or a nested dict, list or tuple of np.Array.

collect_data_spec Describes the data collected when using this policy with an environment.
info_spec Describes the Arrays emitted as info by action().
observation_and_action_constraint_splitter

policy_state_spec Describes the arrays expected by functions with policy_state as input.
policy_step_spec Describes the output of action().
time_step_spec Describes the TimeStep np.Arrays expected by action(time_step).
trajectory_spec Describes the data collected when using this policy with an environment.

Methods

action

View source

Generates next action given the time_step and policy_state.

Args
time_step A TimeStep tuple corresponding to time_step_spec().
policy_state An optional previous policy_state.
seed Seed to use if action uses sampling (optional).

Returns
A PolicyStep named tuple containing: action: A nest of action Arrays matching the action_spec(). state: A nest of policy states to be fed into the next call to action. info: Optional side information such as action log probabilities.

get_initial_state

View source

Returns an initial state usable by the policy.

Args
batch_size An optional batch size.

Returns
An initial policy state.

variables

View source